101
UC Berkeley UC Berkeley Electronic Theses and Dissertations Title Investigating the Role of Conserved Coding-Region Regulatory RNA Elements in Modulating the Dengue Viral Life Cycle Permalink https://escholarship.org/uc/item/5m017928 Author Groat Carmona, Anna Maria Publication Date 2011 Peer reviewed|Thesis/dissertation eScholarship.org Powered by the California Digital Library University of California

Dissertation Template - Draft 4

  • Upload
    others

  • View
    4

  • Download
    0

Embed Size (px)

Citation preview

UC BerkeleyUC Berkeley Electronic Theses and Dissertations

TitleInvestigating the Role of Conserved Coding-Region Regulatory RNA Elements in Modulating the Dengue Viral Life Cycle

Permalinkhttps://escholarship.org/uc/item/5m017928

AuthorGroat Carmona, Anna Maria

Publication Date2011 Peer reviewed|Thesis/dissertation

eScholarship.org Powered by the California Digital LibraryUniversity of California

  

Investigating the Role of Conserved Coding‐Region Regulatory RNA Elements in Modulating the Dengue Viral Life Cycle 

by 

Anna Maria Groat Carmona 

 

A dissertation submitted in partial satisfaction of the 

requirements for the degree of 

Doctor of Philosophy 

in 

Infectious Diseases and Immunity 

in the 

Graduate Division 

of the 

University of California, Berkeley 

 

Committee in charge: 

Professor Eva Harris, Chair Professor Britt Glaunsinger Professor Kathleen Collins 

Fall 2011

  

1  

ABSTRACT 

Investigating the Role of Conserved Coding‐Region Regulatory RNA Elements in Modulating the Dengue Viral Life Cycle 

by 

Anna Maria Groat Carmona 

Doctor of Philosophy in Infectious Disease and Immunity 

University of California, Berkeley 

Professor Eva Harris, Chair 

 

Dengue  (DENV)  is  the most  important mosquito‐borne  virus  affecting  humans  and  is  transmitted  by Aedes aegypti and Ae. albopictus mosquitoes.   DENV  is an enveloped virus  in  the  family Flaviviridae, with a positive‐strand RNA genome that contains a single open reading frame, which is flanked by highly structured 5' and 3' untranslated  regions  (UTRs).   Thus  far,  investigations of cis‐acting  regulatory RNA elements have focused on those within the UTRs, though we have recently described an RNA element in the  DENV  coding‐region,  the  capsid‐coding  hairpin,  which  has  been  shown  to  regulate  both  viral translation  and  RNA  synthesis.    Using  sequence  alignments  and  secondary  structure  prediction algorithms, additional coding‐region RNA elements were identified and characterized for their ability to modulate the viral life cycle in mammalian and mosquito cells.  Three candidate RNA structure elements were identified in the capsid‐ and NS5‐coding region, termed capsid‐1 (C‐1), C‐2 and NS5‐2.  Three RNA sequence elements were  identified  in the capsid‐, prM‐ and NS5‐coding region of the genome, termed the  prM‐coding  conserved  region  1  (prMCR1),  the NS5‐coding  conserved  region  1  (NS5CR1)  and  the conserved capsid‐coding region 1 (CCR1).  Mutations were introduced into a DENV2 infectious clone to disrupt either the conserved predicted secondary structures or the primary nucleotide sequence while maintaining  reading  frame, amino acid R groups, adjacent predicted  secondary  structures, and codon usage  bias  in  both  baby  hamster  kidney  (BHK)  and  Ae.  albopictus  mosquito  (C6/36)  cells.    Initial infectivity screens showed that neither C‐1, C‐2, NS5‐2, prMCR1 nor NS5CR1 were critical  for the viral life cycle.  However, changes to the primary nucleotide sequence of CCR1 were shown to decrease viral titer in both BHK and C6/36 cells, though its effects were shown to have a more critical role in mosquito cells.   The defects  in viral  replication  in C6/36 cells were confirmed  in vivo, where mutations  to CCR1 decrease viral replication in Ae. aegypti mosquito bodies, leading to a defect in viral dissemination to the salivary glands.  Furthermore, CCR1 was shown not to regulate viral translation, RNA synthesis or virion retention  in  BHK  or  C6/36  cells  but  rather  to modulate  viral  assembly,  as mutations  resulted  in  the release of non‐infectious  virions.   Whereas most of  the  identified RNA  elements  in DENV have been shown to regulate RNA synthesis or viral translation, our data has implicated CCR1 as an assembly signal whose  function  is  more  important  for  the  mosquito  vector  rather  than  the  mammalian  host.  Understanding  the  role of  cis‐acting  regulatory RNA elements  in  the DENV  coding  region will provide insight into viral replication strategies and may uncover novel anti‐viral drug targets.  

i  

Dedicated to my grandparents who sacrificed everything so that their children and their children's children could have a better life. 

  

Esteban and Cruz Galindo & Eligio and María del Carmen Carmona Gracias por creer en mi. 

   

Eligio Carmona Solano (11/13/1918 – 2/14/2009) You will always be with me. 

ii  

TABLE OF CONTENTS   Chapter 1:  Introduction                  1   The Epidemiology of Dengue                2     The Mosquito Vector                3   The Dengue Virus Life Cycle                4     Host Infection                  4     Vector Infection                5   RNA Elements in the DENV UTRs              6     Protein‐RNA Interactions              7   Coding‐Region RNA Elements                8     Coding‐Region RNA Elements in Other Viruses          8   Novel Antiviral Therapies                10   Objectives                    10   Figures                      11   References                    14   Chapter 2:  Identification and Characterization of Coding‐Region RNA Elements     24   Introduction                    25   Materials and Methods                  26   Results                      28     The Identification of Novel RNA Elements in the DENV Coding‐Region       Based on Sequence and Secondary Structure Conservation    28     Designing Mutant Constructs and Optimizing the Transfection Protocol       on BHK Cells                29     Candidate RNA Elements are Examined for Viral Infectivity in BHK Cells    31     Examining whether the Putative RNA Elements Play a Role in the Viral       Life Cycle in C6/36 Cells             32     CCR1 and NS5‐2 have no Effect on Viral Translation or RNA Synthesis in       BHK Cells                33   Discussion      34   Figures        36   Tables        48   References                    51      

iii  

Chapter 3:  A Novel Coding‐Region RNA Element Modulates Infectious Dengue Virus   Particle Production in Both Mammalian and Mosquito Cells        54   Introduction                    55   Materials and Methods                  56   Results                      59     A Novel RNA Element is Predicted to be Conserved Both in Sequence       and SecondaryStructure in the DENV and TBEV Serogroups    59     CCR1 Modulates the Viral Life Cycle in Both BHK and C6/36 Cells     60     Altering CCR1 Impairs Viral Replication in Ae. aegypti Mosquitoes    61     Mutating CCR1 has no Effect on Viral Translation or RNA Synthesis in       Either BHK or C6/36 Cells            62     Observed Effects of CCR1 Mutant Constructs on Viral Titer are not due       to RNA Instability              63     CCR1 Regulates Infectious Particle Production in BHK and C6/36 Cells    63   Discussion                    64   Figures                      67   Tables                      77   References                    78   Chapter 4:  “Tying Up Loose Ends”                82   “Tying Up Loose Ends”                  83     Identification of Novel Coding‐Region RNA Elements        83     Sequencing vRNA from Infected BHK Cells          83   Future Directions                  85     Determining the Mechanism by which CCR1 Regulates Infectious       Particle Production              85   Conclusions                    86   Materials and Methods                  87   Figures                      88   Tables                      90   References                    91    

iv  

FIGURES   Chapter 1: Introduction   Figure 1.1:  Schematic diagram of the DENV genome.          11   Figure 1.2:  Schematic representation of the DENV intracellular life cycle.    11   Figure 1.3:  The DENV infection cycle in Aedes mosquitoes.        12   Figure 1.4:  The life cycle of the Aedes mosquitoes.          12   Figure 1.5:  Schematic for the synthesis of negative‐strand vRNA during viral     replication.                  13   Chapter 2:  Identification and Characterization of Coding‐Region RNA Elements   Figure 2.1:  Putative RNA sequence elements are identified in the capsid‐, prM‐     And NS5‐coding region based on sequence conservation in the DENV     And JEV serogroups.                36   Figure 2.2:  Putative RNA structural elements are identified in the capsid‐ and     NS5‐coding region based on structure conservation in the DENV, JEV     and TBEV serogroups.                38   Figure 2.3:  Generation of mutant constructs for the putative RNA structure     elements.                  40   Figure 2.4:  Optimizing viral infectivity analysis.            42   Figure 2.5:  Examination of the putative RNA elements in BHK cells.      43   Figure 2.6:  Infection of BHK and C6/36 cells with mutant viruses reveals a role     for CCR1 in C6/36 cells.                45   Figure 2.7:  Examination of the putative RNA elements in C6/36 cells.      46   Figure 2.8:  CCR1 has no effect on viral translation or RNA synthesis in BHK cells.   47   Chapter 3:  A Novel Coding‐Region RNA Element Modulates Infectious Dengue Virus   Particle Production in Both Mammalian and Mosquito Cells   Figure 3.1:  A conserved RNA element is identified in the capsid coding‐region of     the DENV and TBEV serogroups.             67   Figure 3.2:  CCR1 acts as an RNA sequence element in mediating the viral life cycle     in BHK and C6/36 cells.                69   Figure 3.3:  Infecting BHK and C6/36 cells with mutant viruses confirms that CCR1     plays a greater role in C6/36 cells.            71   Figure 3.4:  CCR1 modulates viral replication and dissemination to the salivary     glands in Ae. aegypti mosquitoes.            72   Figure 3.5:  CCR1 has no effect on viral translation, RNA synthesis or RNA stability     in BHK and C6/36 cells.                73   Figure 3.6:  CCR1 mediates infectious viral particle production.        74   Figure 3.7:  CCR1 has no effect on virion retention in BHK and C6/36 cells.    76   Chapter 4:  “Tying Up Loose Ends”   Figure 4.1:  Putative RNA structural elements are identified in the NS5‐coding region     based on structure conservation in the DENV, JEV and TBEV serogroups.   88

v  

TABLES   Chapter 2:  Identification and Characterization of Coding‐Region RNA Elements   Table 2.1:  SOE‐PCR cloning primers for the generation of DNA mutant constructs.  48   Table 2.2:  qRT‐PCR primer/probe sets directed against the 3’ UTR or the viral     protein NS5.                  49   Table 2.3:  Sequences for the WT and mutant C‐1 constructs.        49   Table 2.4:  Sequences for the WT and mutant C‐2 constructs.        49   Table 2.5:  Sequences for the WT and mutant NS5‐2 constructs.        49   Table 2.6:  Sequences for the WT and mutant prMCR1 constructs.      49   Table 2.7:  Sequences for the WT and mutant NS5CR1 constructs.      50   Table 2.8:  Sequences for the WT and mutant 5’/3’ CS controls.        50   Chapter 3:  A Novel Coding‐Region RNA Element Modulates Infectious Dengue Virus   Particle Production in Both Mammalian and Mosquito Cells   Table 3.1:  SOE‐PCR cloning primers for the generation of DNA mutant constructs.  77   Table 3.2:  qRT‐PCR primer/probe set directed against the viral protein NS5.    77   Table 3.3:  Sequence of WT CCR1 and CCR1 mutants.          77   Chapter 4:  “Tying Up Loose Ends”   Table 4.1:  Sequencing results from BHK cells infected with the “CCR1orig mut”     mutant virus at 48 hours post‐infection.           90   Table 4.2:  Sequencing primers used for the generation of cDNA amplicons using     RT‐PCR.                   90     

vi  

ACKNOWLEDGEMENTS     This isn’t in any particular order.  The saying is that “Rome wasn’t built in a day.”  While it’s not something people talk about, Rome was also not built by one person but rather a somewhat randomly linked network of people with a diverse  set of backgrounds, affiliations,  interests and a whole  lot of slaves.  I would like to say that no slaves were harmed in the making of this dissertation.   Berkeley.  Well I’d be lying if I said I had fun.  =)  I think graduate school is designed to be a series of “ups” and “downs.”  How else do you find out what you are capable of?  For the “ups,” I’m thankful; for  the “downs,” give me a  few months  to buy new  rose colored glasses.    I do want  to  thank all  the students (past and present) and faculty in the IDI department, it’s been one hell of a ride!  I’d like to give a special thanks to Dr. Britt Glausinger, Dr. Kathy Collins and Dr. Caroline Kane for agreeing to be on my Qual  and/or  Dissertation  committee  and  for  helping me  through  some  of  the  tougher  times  in my graduate career.   The Harris Lab!  Sondra and Milton Schlesinger, I can’t thank you enough for all your questions, thoughts, ideas and suggestions that you've given me over the years.  You’ve been an invaluable part of my education in the Harris lab.  Dr. Karen Clyde, my (tor)mentor who introduced me to the wondrously difficult world of RNA elements, I don’t think I could thank you enough for what you’ve taught me nor can  I  imagine my  grad  career without  your  advice  and  guidance.    Thank  you  for  everything.    I have nothing  but  happy  thoughts  for  all  “Z”  past  Harris  lab members  but  Dr.  Jennifer  Kyle,  Dr.  Suman Paranjape, Diana Flores and Ritela Gonzalez; you may have moved on but you are not forgotten and  I can’t imagine my life in this lab without having known you.  For my undergrad, Dipti Banerjee, thank you for all your help (and company) this past year and a half, I couldn’t have asked for a better undergrad.  For Kate Williams  (my  fellow  comrade  in  arms),  Susana Orozco, Mary Pohl, Claire Quiner, Dr. Kristie Ibarra, Dr. Molly OhAinle, Dr. Poornima Parameswaran, Dr. Micha Schmid, Dr. Simona Zompi, Carolyn Winfrey, (“Sargent”) Rachel Sargent, Dr. Robert Beatty, Dr. Josefina Coloma and the hoard of undergrads that roam the TC room‐you have all been a riot through and through!  The beauty of this group is that I can talk science, laugh at the general nerdiness, learn German and debate “inappropriate," "taboo," and "controversial” topics while simultaneously pipetting a 96‐well plate at my bench within the span of a standard enzymatic digestion.  Special thanks to my postdoc, Dr. Peter Friebe, for not killing me during the course of his tenure in the lab.  Wouldn’t have been the same without you Pit!   To my mentor, Dr. Eva Harris, thank you for providing me the opportunity to grow as a scientist and teaching me that I am tougher than I thought.   Through the good and through the bad, thank you for not giving up on me.  Your guidance and support throughout these years has been invaluable.   For my family and friends, thank you for putting up with the late night phone calls, random text messages, obscure tales of woe and my complete lack of a social life.  To my brothers, Victor and Carlos, I thank you for reminding me that there is a light at the end of the tunnel (even if we sometimes have to forge a spoon out of clay and dig through the wall of the tunnel to get to the light).  To my parents, who have always been there to support me (even when I was digging for bones out in the dunes behind our house in the vain hope that I would become an archeologist), I cannot express in words how much you mean to me and how grateful I am for all your sacrifices.  As for the “Carmona,” “Galindo” and “Groat” clans, stay crazy and thank you for all your support and encouragement.  I am in your debt.   For my husband, LeRoy, you mean the world to me and  I can’t fathom how  I could have done this without you by my side or imagine my life without you in it.  I am grateful to all my friends around the world and across the continent but I want to give a shout out to my Berkeley crew.  Sara Irene Ruiz, Danielle Kristy Augustin and Alexander Chun‐Hao Chang; you  three have been  the cornerstone  to my sanity and  I  love you dearly for everything.   Thank you for being there through the  laughter, the tears and the homicidal rages; I truly could never have done it without you guys.  Long live Fight Club! 

1  

CHAPTER ONE 

INTRODUCTION

2  

THE EPIDEMIOLOGY OF DENGUE  Dengue  virus  (DENV)  is  the  most  prevalent  arthropod‐borne  viral  illness  affecting  humans  and  is transmitted  by  the  Aedes  aegypti  and  Ae.  albopictus  mosquitoes  (69).    The  four  DENV  serotypes (DENV1‐4) are members of  the  family Flaviviridae, which  includes other major public health concerns such yellow fever virus (YFV), hepatitis C virus (HCV) and West Nile virus (WNV) (18, 69).  Flaviviruses are grouped into eight antigenic complexes based on reactivity in neutralization assays, including the DENV serogroup, the Japanese encephalitis virus (JEV) serogroup and the tick‐borne encephalitis virus (TBEV) serogroup  (21).   DENV  causes  a  spectrum of  clinical disease  ranging  from  an  acute debilitating,  self‐limited  febrile  illness  (dengue  fever,  DF)  to  a  life‐threatening  vascular  leakage  syndrome  (dengue hemorrhagic fever, DHF) as well as hypotensive shock and circulatory failure (dengue shock syndrome, DSS).   Despite being within  the  same  serogroup  and displaying  similar disease phenotypes, DENV1‐4 only  share  62‐67%  amino  acid  homology  (70),  which means  that  they  are  four  genetically  distinct viruses; a fact that must be considered when discussing the epidemiology and  immunopathogenesis of DENV.  Symptomatic illness due to DENV has been described as the tip of an iceberg (97), given that less than 10% of symptomatic cases are reported (2) and 50‐90% of all viral infections are asymptomatic (10, 19, 47, 158).  Clinical descriptions of dengue‐like syndrome have been recorded as far back as 992 A. D. in China, though the first documented dengue‐like epidemics occurred in 1779‐1780 (67), indicating that dengue has been a major public health concern for quite some time.   DENV causes an estimated 50 million new cases of DF and 250,000‐500,000 cases of DHF/DSS annually worldwide, with a case fatality rate as high as 10‐20% depending on availability of supportive care (1, 61).  Apparent disease as a result of DENV infection is often immune mediated and the incidence of DHF/DSS  varies depending on  geographic  region.   The  risk  factors  for DHF/DSS  include a previous infection with a distinct DENV serotype (secondary heterotypic infection) (19, 47, 64, 72, 76, 145, 158), increased  time  between  infections  (host  immunity),  younger  age,  race,  socioeconomic  status,  host genetics, viral  sequence and viral genotype  (66, 75).   Serotype‐specific  immunity  is believed  to  last a lifetime, though complete cross‐protective  immunity only  lasts 1‐2 months after the primary  infection and partial cross‐protective  immunity only  lasts 6‐9 months (143).   Most symptomatic cases occur as a result of a primary or secondary infection, though a secondary heterotypic infection is the greatest risk factor for DHF/DSS (76).   DENV is associated with explosive urban epidemics, and the incidence of DHF/DSS has increased over  500‐fold  in  the  last  50  years,  affecting  over  100  countries  (98),  including  tropical  Asia,  Africa, Australia  and  the Americas, where  the Ae. aegypti  vector  is present  (30, 48, 163, 164).    It has been estimated  that  in order  to  sustain DENV  transmission within a given population,  the mosquito vector needs  access  to  ~10,000  to  1  million  susceptible  people  (96,  150).    Based  on  vector  distribution, approximately 3.5 billion of  the world's population  is  at  risk  for DENV  infection, often  as  a  result of increases  in  the human population,  increased urbanization and greater ease  in  international  shipping and  travel  (66, 68, 98, 138).   Despite  vector  control methods, dengue  remains a major public health concern given the robustness of the mosquito vector, which can recover and adapt to diverse ecological changes in a short period of time, and the emergence of dengue in the continental United States (121).  Since 1980,  there have been  few  cases of  locally acquired dengue  infections  along  the Texas‐Mexico border, which are temporally associated with large outbreaks in the neighboring Mexican cities (22, 23, 25).  Recently, a dengue outbreak was described in Key West (2009‐2010) and represents the first case of dengue acquired outside of the Texas‐Mexico border in the continental United States since 1945 and the first locally acquired outbreak  in Florida since 1934 (24).   Since the Ae. albopictus mosquito  is now endemic  in  the  Southeastern  United  States,  DENV  has  been  recognized  as  a  Category  A  priority pathogen (66, 77) and classified as a domestic re‐emerging disease (73). 

3  

The Mosquito Vector.   Ae. aegypti  is the primary vector responsible for DENV transmission.   Originally from Africa, the spread of the vector to the American continent has been hypothesized to coincide with the slave trade in the 17th to 19th centuries (70, 150).  Ae. aegypti mosquitoes are efficient viral vectors for many reasons, such as  the preference  to  lay eggs  in clean water and  in artificial containers  in and around human habitations,  feeding on humans principally during  the daytime and  remaining  indoors (70).    Since  Ae.  aegypti  mosquitoes  commonly  exhibit  interrupted  feeding  behavior,  most  female mosquitoes  feed  multiple  times  between  egg  laying,  which  contributes  to  the  explosive  nature  of dengue  epidemics  (147,  154).    Furthermore,  Ae.  aegypti  eggs  can withstand  desiccation  for  several months until the next rainfall or flooding induces hatching, but their eggs are not as adapted for colder climates as those of the secondary DENV vector, Ae. albopictus (79, 137) and less easily infected (70, 88, 119, 141).   The  low  susceptibility of Ae. aegypti  to oral  infection  results  in  the  requirement  for high DENV titers in the blood of infected human hosts in order to contract the virus (exceeding 105/mL) (18).  Thus,  it  is  likely that the Ae. aegypti vector  is an  important selection mechanism for DENV strains with higher viremia and possibly for maintenance of virulent strains within a population.  Genetic differences between mosquitoes may  also  account  for  the  varying  susceptibilities  to DENV  infection  (18, 70, 98, 119).  In the absence of Ae. aegypti, Ae. albopictus has been identified as the primary vector in several dengue epidemics (65).   Ae.  albopictus  is  a  peridomestic  tree‐hole  dwelling mosquito, which  serves  as  the  secondary vector  for DENV  and  is  believed  to  be  an  important means  for  carrying  the  virus  through  the  inter‐epidemic period.   Ae. albopictus  is  indigenous to South East Asia and has spread to Africa, the Middle East, Europe,  the Caribbean, as well as South and North America  (65).   Within  the continental United States, Ae. aegypti  is predominantly found  in the southern and southeastern states, however, since  its introduction  in 1985, Ae. albopictus has also spread  throughout  the southeastern states.   Though Ae. albopictus feeds predominantly on mammalian hosts (136), the host‐feeding pattern of this mosquito is a significant limiting factor in its vector potential for arboviral transmission since it predominately feeds on non‐human mammals (65, 136).  While the opportunistic feeding behavior of Ae. albopictus reduces its ability to acquire or transmit DENV, it also allows the mosquito to take advantage of available hosts (136).   Ae. albopictus was  found  to be  responsible  for a dengue outbreak  in Hawaii  in 2001, but  this outbreak most likely occurred because the virus was introduced by a Hawaii resident who had returned from  Tahiti  (44).    Indeed, many  travelers who  are  still  viremic,  return  to  the United  States,  and  the potential to introduce DENV into a community that has the mosquito vector is a major concern (24).   DENV  infections  show a  seasonality  in  tropical and  subtropical  climates, which peaks  in  rainy season,  indicating  that  the  virus must be maintained  in  either  infected mosquitoes or  asymptomatic hosts during the  interepidemic period (52, 70, 74).   Vertical transmission of DENV  in Aedes species has been experimentally documented.   DENV has been  isolated  from  field  collected  larvae of Ae. aegypti and male Ae.  furcifer‐taylori mosquitoes  in West Africa, and vertical  transmission  is believed  to be an important mechanism for overwinter survival for certain mosquito‐ and tick‐borne flaviviruses (18, 139).  Though  described  as  vertical  transmission,  flaviviruses  actually  infect  the  genital  tract  of  the  female mosquitoes  and enter  the  fully developed egg  through  the micropyle  at  the  time of  fertilization  and oviposition (18, 139).  Sexual transmission can also occur in male Aedes species with inherited infections to  susceptible  females, which may  subsequently pass  the virus  to  their progeny  (18, 139)  though Ae. aegypti is not as efficient at vertical transmission as Ae. albopictus (142).  Additionally, there is evidence for a DENV  sylvatic  cycle whereby  jungle primates and mosquito vectors perpetuate  the virus, a well known mechanism  for maintenance  of  YFV within  a  population  (37).    However,  DENV  transmission requires viral adaptation to the vector rather that to the host to change from a sylvatic cycle to an urban cycle (37).   Currently,  there  are  no  effective  antiviral  therapies  or  vaccines  available  for  dengue,  and treatment  is  largely  supportive;  therefore,  decreasing  the mosquito  population  in  order  to  reduce 

4  

transmission is the best defense.  While non are currently available, there are several candidate vaccines in various stages of advanced development, with clinical trials currently in Phase IIb and recruitment for Phase III is underway.  As with YFV, DENV introductions from a sylvatic cycle may require future vaccine programs  continue  indefinitely  (37).   Despite  the worldwide morbidity  and mortality  associated with DENV infection, neither the molecular virology or pathogenesis is completely understood; both of which are necessary to develop effective vaccines and novel targets for antivirals.   

THE DENGUE VIRUS LIFE CYCLE  DENV is a small (~50 nanometer), enveloped virus with a positive‐strand RNA genome of ~11 kilobases that encodes ~3300 amino acids  in one open reading  frame  (ORF)  (Fig. 1.1).    Its genome  is structured much  like a host cellular mRNA, containing a 5'  type 1 7‐methyl‐G cap and an ORF  that  is  flanked by highly  structured  5'  and  3'  untranslated  regions  (UTRs),  though  unlike  cellular mRNAs, DENV  lacks  a poly(A)  tail  (27)  (Fig.  1.1).    The  5'  UTR  is  approximately  100  nucleotides  in  length,  starting  with  a conserved  AG,  while  the  3'  UTR  ends  with  a  conserved  CU  and  is  of  variable  length  (~450‐600 nucleotides)  (Fig.  1.1).    The  ORF  is  translated  as  a  single  polyprotein  that  is  cleaved  co‐  and  post‐translationally  by  both  viral  and  host  proteases  (8).    The  virus  encodes  three  structural  (capsid‐C, premembrane/membrane‐prM/M and envelope‐E) and  seven nonstructural  (NS) proteins  (NS1, NS2A, NS2B, NS3, NS4A, NS4B and NS5) (Fig. 1.1).  Host  Infection.    Infection  begins  when  the  mosquito  vector  takes  a  blood  meal  and  the  virus  is introduced  into the human host.    It  is the general consensus that cells of the mononuclear phagocyte lineage  (monocytes, macrophages and dendritic cells‐DC) are the primary targets of DENV (11, 40, 87, 175),  specifically  Langerhans  cells  (skin‐resident DCs)  (175).   DENV gains entry  into  the  target  cell by receptor‐mediated  endocytosis  (RME)  (102),  and  a  number  of  low‐affinity  receptors  have  been proposed  to  be  involved,  including  DC‐specific  intercellular  adhesion molecule  3  (ICAM‐3)‐grabbing nonintegrin  (DC‐SIGN)  (109,  123,  157)  (Fig.  1.2).   DC‐SIGN  is  an  attachment  factor  that  can  facilitate infection of all four DENV serotypes, and the ectopic expression of DC‐SIGN confers permissiveness to non‐permissive cell  lines  (157).    In a DENV  infection, RME  involves an additional high‐affinity receptor that mediates entry after  the  low‐affinity  receptor captures  the virus at  the cell surface  (34).   Recent work  has  implicated  the  stage  of  the  cell  cycle  as  a  factor  contributing  to  determination  of  cellular tropism (82, 133).   Human hepatocyte‐derived cells have been shown to be more permissive to DENV and  produce  higher  viral  titers when  stalled  in  the G2  phase  of  the  cell  cycle,  resulting  in  increased infectivity (133).  On the other hand, viral assembly in Ae. albopictus mosquito (C6/36) cells is enhanced when the cells are stalled  in the S phase of the cell cycle, causing a 30‐fold  increase  in viral titers (82).  The viral E protein mediates virus attachment, and once inside the endosome, low pH conditions induce the  E  protein  to  undergo  an  irreversible  trimerization  (118)  that  exposes  the  fusion  peptide, which mediates endosomal fusion of the viral envelope and vesicular membranes (80, 81, 118).   After fusion, the nucleocapsid enters the cytoplasm and the viral genome is uncoated (80, 81) (Fig. 1.2).   Much  like with poliovirus (128), flavivirus translation  is coupled to RNA synthesis (91, 92, 128), and viral RNAs must be translated in order to be replicated.  Translation of the input strand takes place (early) until the replication complex has been assembled in association with the endoplasmic recticulum (ER)‐derived membranes and the viral polyprotein is processed by the viral serine protease NS3/NS2B as well as host proteases (8) (Fig. 1.2).  The small viral hydrophobic proteins NS2A, NS4A and NS4B are not well  characterized  but  have  been  implicated  in  localizing  the  replication  complex  to  the  sites  of replication  (103).   According  to  a model  that has been proposed  for  the  flavivirus  Kunjin  (KUN),  the replication complex is initiated during translation when NS3 and NS2A bind to conserved regions in the 

5  

NS5  protein,  the  RNA‐dependent  RNA  polymerase  (RdRp)  (89).    After  translation  takes  place,  the partially assembled replication complex binds to the 3'UTR of the genome via NS2A (110), NS3 and NS5, and  is transported to the replication sites  in the ER  (92).   Electron tomography  (ET) experiments have shown that DENV‐induced membrane structures are a part of the ER‐derived network containing vesicle pores that could enable the release of newly synthesized viral RNA (173).   Given that the vesicle pores were  found  to be directly opposed  to budding DENV particles on ER membranes,  it  is  likely  that  the DENV‐induced modifications  to  the  ER membrane  structure  coordinate  viral  replication  and  efficient assembly (173).  Once at the ER, the structural proteins and NS1 undergo co‐translational translocation and membrane‐associated cleavage, whereas  the  remaining NS proteins  remain  in  the cytoplasm  (49, 113).   Besides the ER‐resident form that co‐localizes with the viral replication complex, NS1 exists as a membrane anchored form and a secreted form (4, 9).  Interestingly, flavivirus translation can occur even in conditions where cellular translation is inhibited (43).   After early  translation and assembly of  the  viral  replication  complex,  the  virus  switches  from translation to synthesis of a negative‐strand  intermediate that serves as a template for the production of positive‐strand viral RNA (vRNA) (Fig. 1.2).   NS3 exhibits the nucleoside triphosphatase and helicase functions required for replication as well as the 5' triphosphatase activity that  is necessary for 5' 7mG capping (12, 13, 100).  NS5 acts as the RdRp as well as the methyl transferase, which is also necessary for 5' capping (45, 127, 155).  The positive‐strand vRNA is produced in excess with respect to the negative‐strand  intermediate  and  serves  as  template  for  negative‐strand  synthesis,  RNA  translation  and  viral packaging.  Successive rounds of translation (late) produce high levels of the viral structural proteins (C, prM/M and E), which are required for the assembly of the virion.  NS3 has been shown to interact with nuclear  receptor  binding  proteins, which  implicates  it  in  the  intracellular  trafficking  of  the  progeny noninfectious virions between the ER and the Golgi compartment (31), where they undergo maturation before being exocytosed via the secretory pathway (80) (Fig. 1.2).  The virion consists of a nucleocapsid core containing the vRNA, surrounded by an ER‐derived lipid bilayer that includes the viral E and prM/M proteins.   During maturation  in  the  trans Golgi compartment,  the prM/M protein  is processed  to  the mature M protein by the host protease furin, a necessary step in order to expose the E receptor binding domain  that  confers  viral  infectivity  (152,  184)  (Fig.  1.2).    As  infection  proceeds,  virus‐induced hypertrophy of  intracellular membranes continues, and vesicle packets that are thought to be sites of viral replication accumulate (173, 174).  Vector  Infection.    Biological  transmission  of  flaviviruses  by  arthropods  depends  on  ingesting  a  blood meal with sufficient virus to establish an infection in the epithelial cells lining the mesenteron (midgut) (18, 140) (Fig. 1.3).  The virus must escape from the midgut epithelium into the hemocele (body cavity) in order to spread to the brain, fat body and salivary glands of the infected mosquito (18, 140) (Fig. 1.3).  Additional replication occurs in the salivary glands, and the virus is secreted in the saliva to be spread to a susceptible host (18, 140) (Fig. 1.3).  Viral transmission to a susceptible host occurs during feeding, and salivary virus is deposited principally in the extravascular tissues of the host (18).   Many flaviviruses exhibit a high degree of specificity in their ability to infect and be transmitted by arthropod species (18, 141), though maintenance of the virus within the mosquito population can be achieved via a number of mechanisms.   Female mosquitoes  lay  their eggs on  the  inner, wet walls of containers that are filled with water, and the larvae hatch when  inundated with water (Fig. 1.4).  Once hatched, larvae begin to feed until they reach sufficient size, at which point metamorphosis is triggered and the larvae transform into pupae (Fig. 1.4).  Pupae do not feed, but gradually molt into the body of an adult mosquito, which emerges from the water after breaking the pupal skin (Fig. 1.4).  The entire life cycle  lasts  8‐10  days  at  room  temperature,  depending  on  the  level  of  feeding.    Unlike mosquitoes infected with alphaviruses, there are no detectable pathologic changes that can be found  in flavivirus‐

6  

infected mosquitoes  (18).   Additionally,  infected mosquitoes  remain  infectious  for  life  (1‐2 weeks  to ~174 days); a fact that contributes to the persistence of dengue within a population (70).   

RNA ELEMENTS IN THE DENV UTRS  As  with  other  positive‐strand  RNA  viruses,  the  5'  and  3'  UTRs  play  a  key  role  in  regulating  viral translation and RNA synthesis (102).   Both the 5’ and 3’ UTRs contain several conserved RNA elements (sequences and secondary structures) that have various regulatory roles in the DENV viral life cycle.  The 5'  UTR  nucleotide  sequence  is  homologous  among  DENV1‐4  and  contains  the  conserved  secondary structures stem loop (SL) A (SLA) and SLB, the latter of which terminates in the first translation initiator AUG  (Fig.  1.1).    SLB  contains  within  it  the  5'  upstream  AUG  region  (UAR)  sequence,  which  has  a complementary sequence,  termed  the 3’ UAR,  located within  the 3’ UTR  (Fig. 1.1).   While  the 5’ UAR sequence  is  required  for  RNA  synthesis,  the  SLB  structure  is  not  required  and may  be  conserved  in DENV1‐4 because the structure  is conserved (6).   The 5' conserved sequence (CS)  is  located within the capsid‐coding sequence and also has a complementary sequence in the 3’ UTR, called the 3’ CS (Fig. 1.1).  Recently, an additional set of complementary sequences, termed the 5' downstream AUG region (DAR) was  identified  just  downstream  of  the  initiating  AUG  within  the  capsid‐coding  region  with  a complementary sequence in the 3' UTR, termed the 3’ DAR (55, 56) (Fig. 1.1).  Despite the fact that the 5’/3’ CS and 5’/3’ DAR contain a 5’ sequence located within the coding‐region, they are still considered when discussing the RNA elements in the UTRs.   The  5'/3' DAR  acts  together with  the  5’/3’  CS  and  5'/3' UAR  in  order  to  circularize  the  viral genome, which is a requirement for efficient RNA synthesis but is not involved in translation (5‐7, 29, 55, 56, 90) (Fig. 1.5).  It is known that many viral RNA genomes circularize in order to coordinate the switch between the initiation of translation and RNA synthesis at the 5' and 3' ends of the genome mediated by protein‐RNA and/or RNA‐RNA  interactions, though  in DENV, replication  is mediated solely by RNA‐RNA interactions  (42,  90)  (Fig.  1.5).    Complementarity  between  the  5’/3’  CS,  5’/3’  UAR  and  5'/3'  DAR  is required but is not sequence‐specific since changing the primary nucleotide sequence while maintaining base pairing rescues defects in RNA synthesis though not always to wild type (WT) levels (6, 7, 55, 90).  Recently,  it has been  shown  that  viral  replication  requires  a balance between  the  linear  and  circular conformations since shifting the equilibrium towards either the linear or circular conformation results in decreased viral  infectivity  (168, 169).   The oscillation between alternative conformations may act as a mechanism for organizing multiple functions of the viral genome, although to date, no mechanism for coordinating  the molecular  switch  between  viral  translation  and  RNA  synthesis  has  been  described (168).   The SLA stem structure rather than the primary nucleotide sequence has been shown to be vital for RNA synthesis (20, 51) (Fig. 1.1 and 1.5).  Two helical regions were identified in the SLA, a side stem loop,  a  top  loop  and  a U bulge  (108).   While  the  SLA  loop  sequence  and  a  conserved oligo(U)  track present downstream of SLA were shown to be  important  for RNA synthesis, the disruption of the SLA side  loop, or bulge, had no effect  (20, 51, 108).    It  is believed that when the genome circularizes as a result of RNA‐RNA interactions (5’/3’ CS, 5’/3’ UAR and 5'/3' DAR), the SLA acts as the promoter element by providing the RdRp access to the 3' end after it binds to the SLA secondary structure and directs the synthesis of the negative‐strand intermediate (7, 51) (Fig. 1.5).  In WNV, an RNA secondary structure at the  3’  end  of  the  negative‐strand  has  been  shown  to  be  important  in  the  subsequent  asymmetric amplification of vRNA from the negative‐strand template (46, 148).   RNA  elements  in  the  3’ UTR  are  involved  in modulating  5’‐end  dependent  initiation  of  RNA translation (5, 29, 84, 171) and RNA synthesis (5, 41) (Fig. 1.1).   The 3’ UTR also contains a number of RNA elements, including the 3’ stem loop (SL), dumbbell structures (DB) 1 and 2, the variable region (VR) 

7  

as well as the aforementioned 3’ CS, 3’ UAR and 3' DAR  (Fig. 1.1).    In the context of a DENV replicon, deleting the 3' UTR altogether had no effect on early translation events (5, 41) despite the role of the 3’ SL  in enhancement of translation  in reporters and replicons (84, 85).   More recently,  it was found that RNA elements in the 3’ UTR can positively and negatively regulate translation, potentially explaining the differences observed between these studies (171).  Additional RNA elements in the 3' UTR include DB1 and DB2, which contain the sequences CS2 and repeated CS2 (RCS2) respectively (Fig. 1.1).  Also present is  the  VR, which  is  not  conserved  among  flaviviruses  or  even  among  viruses within  the  same DENV serotype and varies in sequence and length, unlike the CS2 and RCS2 (Fig. 1.1).  Deleting DB1, DB2 and the VR significantly decreases DENV translational activity in both C6/36 and baby hamster kidney (BHK‐21, clone 15; hereafter BHK) cells  (5, 29).   Additionally, deleting  the 3' UTR, DB1, DB2 or both DB1/2 abolishes RNA synthesis in C6/36 and BHK cells (5).  However, the functions of the dumbbells are most likely redundant since the deletion of both DB1 and DB2 has a more deleterious effect than the deletion of  either  structure  alone  (5).    The DB  structures have  the  ability  to  form  two potential pseudoknots between  the  identical  five‐nucleotide  terminal  loops 1 and 2  (TL1 and TL2) and  their  complementary pseudoknot motifs (PK2 and PK1), which are proximally  located  in the 3’ UTR (112).   All four motifs as well as and CS2 are important for viral replication but only TL2 has a modest effect on viral translation (112).  Since viral translation is only reduced by  60% when TL1 and TL2 are mutated, it is possible that TL1 exhibits a cooperative synergy with TL2, however, complementarity between TL1/PK2 and TL2/PK1 can maintain WT  levels of  viral  translation  even when noncanonical  translation  initiation  is  inhibited (112).  Recently, small structured non‐coding RNAs derived from the flavivira WNVl 3’UTR (sfRNAs) were shown  to  accumulate  within  the  cell  (101,  111,  134)  and  have  been  implicated  in  virus‐induced cytopathic effects  in  cell  culture and  in  influencing viral pathogenicity  in mice  (134).   Overall,  studies have demonstrated that mutations in the DENV 5' and 3' UTRs reduce viral replication in cell culture and mosquitoes, neurovirulence in mice and viremia in monkeys (20, 116, 182).   Interestingly,  deleting  the  VR  reduces  RNA  synthesis  in  BHK  cells  but  slightly  amplifies  RNA synthesis in C6/36 cells, indicating that this sequence has differential effects depending on the cell type (5).   Similarly, changes  to  the bottom  stem  structure of  the 3’ SL were also  shown  to have cell  type‐associated differential effects, resulting in WT replication levels in BHK cells but not in C6/36 cells (182).  To  date,  the  VR  and  the  bottom  portion  of  the  3’  SL  are  the  only  known  RNA  elements  that  have differential effects on the viral  life cycle  in a cell‐type dependent manner.   Elucidating of host‐specific protein  interactions  is  vital  in  understanding  how  DENV  differentially modulates  its  life  cycle  in  the human host and  the mosquito vector,  though more  research  is necessary  in order  to elucidate  these global molecular regulatory mechanisms.  Protein‐RNA  Interactions.   While  the  flavivirus  5'  UTRs  are  longer  and more  structured  than many cellular capped 5' UTRs, they are shorter and less structured than the uncapped internal ribosome entry site  (IRES)‐containing  5' UTRs  characteristic  of  the  other members  of  the  Flaviviridae  family  like  the hepaci‐ or pestiviruses.  It is presumed that most flavivirus translation is cap‐dependent and the DENV 5' UTR  does  not  appear  to  contain  IRES  activity,  although  DENV  can  replicate  under  noncanonical mechanisms  when  cellular  cap‐dependent  translation  is  inhibited  (42,  43).    In  cap‐dependent translation, the m7G cap associates with elongation  initiation factor 4F (eIF4F); the cap‐binding protein eIF4E, the scaffold protein eIF4G, and eIF4A, which functions as a helicase in conjunction with eIF4B (59, 144, 156).   Several other mammalian proteins that have been shown to  interact with the DENV 3’ UTR include  the poly(A) binding protein  (PABP), poly‐pyrimidine  tract binding protein  (PTB), Y‐box binding protein 1 (YB‐1), and heterogenous nuclear ribonucleoproteins (hnRNP A1, hnRNP A2/B1 and hnRNP Q) (14, 15, 35, 36, 131, 135, 178, 179).   YB‐1  binds  to  the  terminal  loop  of  the  3’  SL  and  has  an  anti‐viral  effect  on DENV  infection, mediated  in part by  inhibiting  input strand translation (131).   YB‐1‐mediated translational repression  is 

8  

not observed with reporter constructs in vitro or in cultured cells (131) and is not affected by providing DENV proteins  in trans,  implying that translational repression requires genomic sequences and/or viral proteins provided in cis.  The mosquito and human La autoantigen (La) has been shown to interact with both the DENV UTRs (60, 178, 179), as well as with the negative‐strand 3’UTR, suggesting it might play a role in RNA synthesis (36, 60, 178, 179).  The T‐cell intracellular antigen‐related (TIAR) protein and T‐cell intracellular antigen‐1  (TIA‐1) have been  shown  to  interact with  the 3’ SL on  the WNV RNA negative‐strand (46, 148) and the TIAR binding sites have been mapped to short AU sequences (UAAUU) located in two internal loops of the 3' SL RNA structure (46).  Based on studies with WNV, KUN and JEV, the 3' SL has been shown to mediate binding of NS5, NS3 (28), NS2A (110), eLF1A (15), and other host proteins (14).   Binding of NS5 and cellular GAPDH  to  the 3’ end of  the negative‐strand has been shown  in  JEV (176).  All in all, much is known about the roles of RNA elements in the UTRs in regulating the viral life cycle by either RNA‐RNA or RNA‐protein interactions, but there is a great deal less information available as to the roles of putative RNA elements located within the DENV coding‐region.   

CODING‐REGION RNA ELEMENTS  The DENV start codon is located in a poor initiation context and is conserved among DENV1‐3, as well as members of the JEV serogroup.  To compensate for the poor initiation context, it was hypothesized that a downstream (~14 nts) RNA secondary structure stalled the ribosome in order to enhance the selection of the  first AUG  (94, 95).   This structure, termed  the capsid‐coding region hairpin  (cHP), was  found to enhance selection of the first AUG in a position‐dependent, sequence‐independent manner proportional to  its  stability  (33)  (Fig.  1.1).    Further  work  has  shown  that  the  cHP  additionally  functions  in  RNA synthesis and that these effects are  independent of sequence for both DENV and WNV (32).   Thus far, little  is  known  about  the  existence  of  potential  coding‐region  RNA  elements  in  DENV.    Based  on structure prediction algorithms that have been published, the DENV coding‐sequence has been reported to be somewhat unstructured (83, 159), although no functional proof exists that there are no additional putative coding‐region RNA sequence elements.  Coding‐Region RNA Elements  in Other Viruses.   There are a number of animal and plant viruses with known coding‐region RNA elements that modulate various aspects of the viral  life cycle  including viral translation,  RNA  synthesis  and  viral  assembly  as  well  as  other  less  well  characterized  replication intermediates (106, 117).   With respect to mediating viral replication, one of the most well characterized  is the cis‐acting replicating element  (CRE), which  is a 61‐nt  stem‐loop  found within  the coding  sequence of poliovirus (PV) protein  2C  (63).   Within  the  Picornaviridae  family,  the CREs  can  vary  in  length  though  they  are preferentially  located within  the  coding‐region,  the only exception being  the  foot‐and‐mouth disease virus  (FMDV)  in  which  the  CRE  is  located  in  the  5’  nontranslated  region  (NTR)  (114,  132,  153).  Adenosines  in  the  loop of  the CRE RNA structure  function as  the  template  for  the uridylylation of  the viral protein VPg, which in turn serves as the primer for the PV RdRp.  It has been shown that inhibiting CRE mediated uridylylation can  inhibit the synthesis of the positive‐strand but not the negative‐strand vRNA  (120, 122).   HCV contains several evolutionarily conserved  secondary structures within  the viral coding‐region,  including  the SLV and SLVI within  the core protein  that are  thought  to be  required  for translation and replication (115, 166), as well as the SL9266 within the NS5B‐coding sequence, which is involved in RNA synthesis (54, 99, 180, 181).  It has been proposed that the SLV and SLVI might stimulate HCV IRES function by reducing inhibitory interactions between the 5’ NTR and the core region (86, 170) though an  interaction between SLVI and the 24‐38 nt of the 5’ NTR are predicted to be detrimental to IRES dependent  translation  (93).   The SL9266 has been  implicated  in a kissing  loop  interaction with a 

9  

conserved SL in the 3' NTR (54, 180) as well as with an upstream RNA sequence located at position 9110 in the NS5B coding‐region, which is involved in viral replication (38).  This cooperative binding of SL9266 with both its 5’ and 3’ sequences increases its stability by creating a pseudoknot, though it is presently unknown if the functions of these kissing loops are connected (38).   A  34‐nt  sequence  located  within  the  ORF7  of  the  porcine  reproductive  syndrome  virus  is required for negative‐strand synthesis and while  it  is predicted to form a hairpin (167) the kissing  loop interaction with the 7‐nt sequence in the loop of this structure with the loop of a hairpin in the 3’ NTR is more  important  for  its  function  (167).    In  cardioviruses, Theiler’s murine encephalomyelitis virus and mengovirus encode a 9‐nt RNA sequence element  located  in a bulge structure within  the VP2 capsid‐coding  sequence, which  has  been  implicated  in  RNA  synthesis  (107).    The  RNA  sequence  and  bulge structure were both  found  to be conserved amongst various cardioviruses and potentially  throughout the  Picornaviridae  family,  although  its  location  within  the  genome  varies  (107).    In  the  bovine coronavirus  (BCoV),  there  are  two  stem  loops  (SLV  and  SLVI)  in  the  nsp1‐coding  sequence  that  are required  for replication of  the  full‐length viral genome  (17, 71).   The CRE element of hepatitis A virus was recently identified, consisting of a large 53‐nt stem‐loop structure located near the end of the 3Dpol coding‐region, which appears to be vital for the replication of a subgenomic (sg) replicon (177).  Likewise in mediating  synthesis  of  a  sgRNA,  in  the  alphavirus  Sindbis  (SIN)  coding‐region,  there  lies  an  RNA element in the “junction region” of the genomic RNA preceding and including the beginning of the 26S sgRNA.   However,  its functions as the promoter for the synthesis of the sgRNA  is  in the context of the negative‐strand (129).  The promoter for the sgRNA on the negative‐strand of the rubella virus is at the junction UTR  (J‐UTR)  separating  the  two ORFs  in  the viral genome  (162).   A  stem‐loop  located  in  the negative‐strand just 6‐nt upstream from the transcription site of the sgRNA  in all caliciviruses acts as a promoter element for the synthesis of the sgRNA, functioning in a sequence‐independent but structure‐dependent manner (149).   Translation of different overlapping ORFs can also be mediated by coding‐region RNA elements.  Consider  that  the  two  replicase  proteins  of  the  SARS‐CoV  are  produced  by  ribosomal  frameshifting mediated by a heptanucleotide sequence  (UUUAAAC) and a closely spaced pseudoknot  located  in  the replicase ORF  (16, 39).   ORF1b  translation  in arteriviruses  requires a  ribosomal  frameshift  just before termination of ORF1a and the sequence (GUUAAC) and downstream pseudoknot structure that promote this function are located within the overlap region (62, 151).   The  functional  role of coding‐region RNA elements  is not  limited  to viral  translation and RNA synthesis,  but  rather  can  be  involved  in  cellular  localization  of  the  vRNA  as  well  as mediate  viral assembly.   Tombusviruses code for two overlapping replication proteins, the p33 auxiliary protein and the p92 polymerase, but in the cucumber necrosis virus, p33 recruitment of the positive‐ and negative‐strand  to peroxisomal membranes  is mediated via binding  to a coding‐region RNA element  (130).    In flock  house  viruses  an  element  in  RNA1  (position  68‐205)  is  predicted  to  form  two  stem‐loops with nearly  identical  sequences  that  are  required  both  for  the  recruitment  of  RNA1  to  the  outer mitochondrial membranes and  for positive‐ and negative‐strand RNA synthesis  (165).    In  retroviruses, the packaging signal (designated Ψ) is usually located in the 5’ end between the 5’ UTR and the Gag ORF (26, 105).  However, in the case of SIN, the 132‐nt long segment of the nsp1‐coding sequence (position 944‐1076)  is predicted to form four stem  loop structures (58, 104, 172) but only two purine rich  loops were found to be essential for vRNA encapsidationl (104).  In the ORF1b of group 2 CoV mouse hepatitis virus  there  is  a 69‐nt bulged  stem‐loop  that  is  required  for  viral packaging  (53).    In nodaviruses,  the packaging signal for RNA2 is located within a 32‐nt region of RNA2 (nt 186‐217) (183).   Other  less conventional aspects of  the viral  life cycle  that are mediated by coding‐region RNA elements can be observed in coxsackie A viruses (78).  It has been observed that PV is more resistant to degradation by RNase L, a  latent endoribonucleoase  in an  interferon‐regulated, double stranded RNA‐activated  pathway,  compared  to  other  positive‐strand  RNA  viruses  like  HCV.    Inhibition  of  RNase  L 

10  

degradation  is mediated by stem‐loops 1  (position 5742–5824) and 4  (position 5906–5967)  located  in the 3C coding‐region of the genome (78, 160), which are predicted to be involved in a putative “kissing loop”  interaction  (160).    RNA  elements  can  also  explain  cell  type‐specific  interactions, which  are  of particular  importance  when  discussing  differential  regulation  of  the  arboviral  life  cycle  within  the mammalian  host  and  the  viral  vector.   Within  the  SIN  genome,  there  exists  a  51‐nt  long  conserved sequence element (CSE)  in the nsP1‐coding sequence (position 155‐205) that  is predicted to form two smaller stem‐loop structures (SL3 and SL4).  These structures act as replication enhancers though these effects are more important in mosquito cells than in mammalian cells (50, 57, 126).   Studies such as these provide evidence that RNA elements located within the coding‐region can have  diverse  regulatory  roles  in  the  viral  life  cycle.    Thus  far,  there  have  not  been many  efforts  to elucidate the roles of coding‐region RNA elements within the DENV genome.   

NOVEL ANTIVIRAL THERAPIES  Conventional drugs  are often  specifically directed  against  the  viral proteins  though  for most  viruses, resistances  can  develop  quickly.    Therefore  it  is  important  to  expand  our  knowledge  of  possible therapeutic  compounds  and  approaches  by  targeting  the  vRNA  and  potentially  using  known  RNA elements  to our advantage.   RNA  interference based  strategies have been  successfully developed  for many  viruses,  including  DENV  (125,  161).    The  antiviral  use  of  phosphorodiamidate  morpholino oligomers  and  ribozyme  in  knocking  down DENV  replication  has  been  demonstrated  (85,  124).    The effective use of aptamers as antivirals has been demonstrated to  interfere with HCV replication (146).  Alternatively,  RNA  decoys  or  compounds  that  bind  to  viral  RNA  element  can  out‐compete  natural binding partners  and have been  implicated  as useful  antiviral  tools  (3,  146).    Since most RNA based antiviral therapies depend on a high complementarity to the viral target sequence, coding‐region RNA elements provide ideal targets because conserved motifs are also required for amino acid conservation.  If  there  are  no  additional  restrictions  responsible  for  the  conservation  of  the  primary  nucleotide sequence, viral escape mutants are  likely  to emerge, however, coding‐region RNA elements minimize this  possibility.    Thus  far,  only  the  5’  CS,  5’  DAR  and  the  cHP  have  been  identified  as  known  RNA elements within  the DENV coding‐region, but  in  this study we hope  to  identify other putative coding‐region RNA regulatory elements that modulate the viral life cycle.   

OBJECTIVES  Despite  the  fact  that a number  coding‐region RNA  regulatory elements have been  identified  in other RNA viruses, the cHP is the only known DENV coding‐region RNA element ‐‐ given that current research has focused on those elements located in the UTRs.  As discussed in Chapter 2, the goal of this study is to  identify and  characterize additional  cis‐acting RNA  regulatory elements  in  the DENV  coding‐region that regulate the viral life cycle in mammalian and mosquito cells.  In Chapter 3, we describe the role of a novel coding‐region regulatory RNA element, termed the conserved capsid‐coding region 1 (CCR1), in modulating infectious particle production in both mammalian and mosquito cells, though its effects are shown to be more dramatic  in the mosquito vector both  in vitro and  in vivo.    In Chapter 4, we discuss the possible mechanism of action  through which CCR1 might act as an assembly  signal  for DENV and propose how  future  studies might be pursued.   The  identification  and  characterization of novel RNA regulatory elements in the DENV coding region will improve our understanding of the viral life cycle and provide  possible  targets  for  novel  therapeutics,  since  the  evolution  of  such  sequences  towards resistance is constrained by both their RNA regulatory function and amino acid coding capacity.

11  

FIGURES   

  Figure  1.1:    Schematic  diagram  of  the DENV  genome.   Outline  of  the DENV  RNA  genome,  including  the  viral proteins  encoded  in  the  single  open  reading  frame  (ORF)  and  the  known  cis‐acting  regulatory  RNA  elements located in the 5’ and 3’ untranslated region (UTR), as well as in the capsid‐coding sequence.  SLA, stem‐loop A; SLB, stem‐loop B; UAR, upstream AUG  region; CS,  conserved  sequence; DAR, downstream AUG  region;  cHP,  capsid‐coding hairpin; VR, variable region; DB, dumbbell; 3’ SL, 3’ stem‐loop; CS2, conserved sequence 2; RCS2, repeated conserved sequence 2; C, capsid protein; prM/M, premembrane/membrane protein; E, envelope protein; NS1‐5, nonstructural proteins 1‐5.  Adapted from Dr. Charlotta Polacek (Dr. Eva Harris’ Laboratory, 2008).   

  Figure 1.2:   Schematic  representation of  the DENV  intracellular  life  cycle  (34).   DENV  (1) binds and  (2) enters target  cells by  receptor‐mediated  endocytosis  through  an  as  yet unknown  receptor(s).    (3) Acidification of  the endosome allows  for a conformational change  in  the viral E protein, which allows  for  fusion and  release of  the nucleocapsid core into the cytoplasm.  Once uncoated, the vRNA is trafficked to the ER where (4) the viral genome is translated and the viral replication complex is assembled.  (5) DENV replication occurs through a negative‐strand intermediate and (6) subsequent rounds of viral translation and RNA synthesis lead to the assembly of immature virions.  (7) Virion maturation occurs during transit through the trans Golgi compartment where prM/M is cleaved by  the  host  protease  furin  and mature  infectious  virions  are  secreted  from  the  cell  via  the  exocytic  pathway.  Copyright 2006 by American Society for Microbiology; reprinted with permission.   

 

 Figure 1.3:(1) the  inglining  the hemocele undergo ain the salivc/spip.php  

 Figure 1.4terrestrial hatch whefor severalmetamorpform until breaking th  

:  The DENV ingestion of a blomesenteron  (m(body cavity) tnother round va to be transmp?article80. 

:   The  life cyclform.  Female en the eggs arel days and aftephosis  is  triggethe body of thhe pupal skin.  

nfection cycle iood meal whemidgut).   The to eventually (of replication mitted to a susc

le of Aedes mmosquitoes (1e inundated wer three separaered and  the  lahe adult, flyingReprinted fro

n Aedes mosqere sufficient qvirus must  (2)(4) disseminatin the salivary ceptible host.  

osquitoes.   Th1) lay their eggith water.  Larate moltings, tarvae  transfor mosquito is fom CDC, http://

12 

quitoes.  Aedesquantities of v) escape  from e to the brain,glands so thatAdapted from

he Aedes mosqgs on the wet wrvae feed on mthey grow fromm  into pupaeormed.  The ne/www.cdc.gov/

s mosquitoes bvirus establish the midgut e, fat body andt (6) sufficient Pasteur Instit

quito  life cyclewalls of water‐microorganismsm first to fourt.   Pupae do noewly formed a/dengue/ ento

 

become infectean  infection  inpithelium and salivary gland amounts of vute, http://ww

 

 has two stagefilled containes and particulath instars.  In tot  feed and  indult emerges fomologyEcolog

ed with DENV dn the epitheliad  (3)  replicate ds.   (5) The virirus can be secww.institutpast

es, an aquatic ers and (2) the ate inorganic mthe fourth instnstead  (4) chafrom the wategy/m_lifecycle.

during al cells in  the us will creted teur.n 

and a larvae matter tar, (3) nge  in r after html. 

13  

  Figure 1.5:  Schematic diagram of the synthesis of negative‐strand vRNA during DENV replication (51).  The DENV genome circularizes due to RNA‐RNA interactions, independent of protein binding.  The process of circularization is mediated by the basepairing interaction of the 5’/3’ UAR, 5’/3’ CS and 5’/3’ DAR.  The viral protein NS5 is the RNA‐dependent  RNA  polymerase  (RdRp), which  binds  to  SLA  and  gains  access  to  the  3’  end  of  the  genome  during circularization to  initiate synthesis of the negative‐strand  intermediate, which  in turn serves as the  template  for synthesis  of  the  positive‐strand  RNA.    Copyright  2006  by  Cold  Spring Harbor  Laboratory  Press;  reprinted with permission.      

14  

REFERENCES   1.  1997.  Dengue  Haemorrhagic  Fever:  Diagnosis,  treatment,  prevention,  and  control.    Second 

Edition. World Health Organization. 2.  2000.  Strengthening  implementation  of  the  global  strategy  for  dengue  fever/dengue 

hemorrhagic fever prevention and control. World Health Organization. 3.  Akkina, R., A. Banerjea, J. Bai, J. Anderson, M. J. Li, and J. Rossi. 2003. siRNAs, ribozymes and 

RNA decoys  in modeling  stem  cell‐based gene  therapy  for HIV/AIDS. Anticancer Res 23:1997‐2005. 

4.  Alcon‐LePoder, S., M. T. Drouet, P. Roux, M. P. Frenkiel, M. Arborio, A. M. Durand‐Schneider, M. Maurice,  I.  Le Blanc,  J. Gruenberg, and M.  Flamand. 2005. The  secreted  form of dengue virus  nonstructural  protein  NS1  is  endocytosed  by  hepatocytes  and  accumulates  in  late endosomes: implications for viral infectivity. J Virol 79:11403‐11411. 

5.  Alvarez, D. E., A. L. De Lella Ezcurra, S. Fucito, and A. V. Gamarnik. 2005. Role of RNA structures present  at  the  3'UTR  of  dengue  virus  on  translation,  RNA  synthesis,  and  viral  replication. Virology 339:200‐212. 

6.  Alvarez, D. E., C. V. Filomatori, and A. V. Gamarnik. 2008. Functional analysis of dengue virus cyclization sequences located at the 5� and 3�UTRs. Virology 375:223‐235. 

7.  Alvarez, D. E., M. F. Lodeiro, S. J. Ludueña, L. I. Pietrasanta, and A. G. Gamarnik. 2005. Long‐range RNA‐RNA interactions circularize the dengue virus genome. J Virol 79:6631‐6643. 

8.  Arias, C. F., F. Preugschat, and J. H. Strauss. 1993. Dengue 2 virus NS2B and NS3 form a stable complex that can cleave NS3 within the helicase domain. Virology 193:888‐899. 

9.  Avirutnan, P., L. Zhang, N. Punyadee, A. Manuyakorn, C. Puttikhunt, W. Kasinrerk, P. Malasit, J. P. Atkinson, and M. S. Diamond. 2007. Secreted NS1 of dengue virus attaches to the surface of cells via interactions with heparan sulfate and chondroitin sulfate E. PLoS Pathog 3:e183. 

10.  Balmaseda, A., S. N. Hammond, Y. Tellez, L. Imhoff, Y. Rodriguez, S. Saborio, J. C. Mercado, L. Perez, E. Videa, E. Almanza, G. Kuan, M. Reyes, L. Saenz, J. J. Amador, and E. Harris. 2006. High seroprevalence of antibodies against dengue virus  in a prospective  study of  schoolchildren  in Managua, Nicaragua. Trop Med Intl Health 11:935‐942. 

11.  Balsitis, S.  J.,  J. Coloma, G. Castro, A. Alava, D. Flores,  J. H. McKerrow, P. R. Beatty, and E. Harris. 2009. Tropism of dengue virus in mice and humans defined by viral nonstructural protein 3‐specific immunostaining. Am J Trop Med Hyg 80:416‐424. 

12.  Barthelma, G., and R. Padmanabhan. 2002. Expression, purification, and characterization of the RNA 5'‐triphosphatase activity of dengue virus type 2 nonstructural protein 3. Virology 299:122‐132. 

13.  Benarroch, D., B. Selisko, G. A. Locatelli, G. Maga, J. L. Romette, and B. Canard. 2004. The RNA helicase,  nucleotide  5'‐triphosphatase,  and  RNA  5'‐triphosphatase  activities  of  Dengue  virus protein  NS3  are  Mg2+‐dependent  and  require  a  functional  Walker  B  motif  in  the  helicase catalytic core. Virology 328:208‐218. 

14.  Blackwell,  J.  L.,  and M.  A.  Brinton.  1995.  BHK  cell  proteins  that  bind  to  the  3'  stem‐loop structure of the West Nile Virus genome RNA. J Virol 69:5650‐5658. 

15.  Blackwell, J. L., and M. A. Brinton. 1997. Translation elongation factor‐1 alpha interacts with the 3' stem‐loop region of West Nile Virus genomic RNA. J Virol 71:6433‐6444. 

16.  Brierley, I., M. E. Boursnell, M. M. Binns, B. Bilimoria, V. C. Blok, T. D. Brown, and S. C. Inglis. 1987.  An  efficient  ribosomal  frame‐shifting  signal  in  the  polymerase‐encoding  region  of  the coronavirus IBV. EMBO 6:3779‐3785. 

15  

17.  Brown, C. G., K. S. Nixon, S. D. Senanayake, and D. A. Brian. 2007. An RNA stem‐loop within the bovine  coronavirus  nsp1  coding  region  is  a  cis‐acting  element  in  defective  interfering  RNA replication. J Virol 81:7716‐7724. 

18.  Burke, D.  S.,  and  T.  P. Monath.  2001.  Flaviviruses,  p.  1043‐1126.  In D. M.  Knipe  and  P. M. Howley (ed.), Fields Virology, 4th ed, vol. 1. Lippincott Williams and Wilkins, Philadelphia. 

19.  Burke, D. S., A. Nisalak, D. E.  Johnson, and R. M. Scott. 1988. A prospective study of dengue infections in Bangkok. Am J Trop Med Hyg 38:172‐180. 

20.  Cahour,  A.,  A.  Pletnev, M.  Vazeille‐Flacoz,  L.  Rosen,  and  C.‐J.  Lai.  1995.  Growth‐restricted dengue  virus mutants  containing  deletions  in  the  5'  noncoding  region  of  the  RNA  genome. Virology 207:68‐76. 

21.  Calisher, C. H., N. Karabatsos, J. M. Dalrymple, R. E. Shope, J. S. Porterfield, E. G. Westaway, and W. E. Brandt. 1989. Antigenic  relationships between Flaviviruses as determined by cross‐neutralization tests with polyclonal antisera. J Gen Virol 70:37‐43. 

22.  CDC. 1996. Dengue fever at the US‐Mexico border, 1995‐1996. MMWR 45:841‐844. 23.  CDC. 2007. Dengue hemorrhagic fever‐‐‐US‐Mexico border, 2005. MMWR 56:785‐789. 24.  CDC. 2010. Locally acquired dengue‐‐‐Key West, Florida, 2009‐2010. MMWR 59:577‐581. 25.  CDC. 2001. Underdiagnosis of dengue‐‐‐Laredo, Texas, 1999. MMWR 50:57‐59. 26.  Chadwick,  D.  R.,  and  A. M.  Lever.  2000.  Antisense  RNA  sequences  targeting  the  5'  leader 

packaging  signal  region  of  human  immunodeficiency  virus  type‐1  inhibits  viral  replication  at post‐transcriptional stages of the life cycle. Gene Ther 7:1362‐1368. 

27.  Chambers, T.  J., C. S. Hahn, R. Galler, and C. M. Rice. 1990.  Flavivirus genome organization, expression, and replication. Ann Rev Microbiol 44:649‐688. 

28.  Chen,  C.  J., M.  D.  Ku,  L.  J.  Chien,  S.  L. Hsu,  Y. M. Wang,  and  J. H.  Lin.  1997.  RNA‐protein interactions:  involvement of NS3, NS5, and 3' noncoding regions of Japanese encephalitis virus genomic RNA. J Virol 71:3466‐3473. 

29.  Chiu, W. W., R. M. Kinney, and T. W. Dreher. 2005. Control of  translation by  the 5'‐ and 3'‐terminal regions of the dengue virus genome. J Virol 79:8303‐8315. 

30.  Chow, V. T. K., Y. C. Chan, R. Yong, K. M. Lee, L. K. Lim, Y. K. Chung, S. G. Lam‐Phua, and B. T. Tan.  1998. Monitoring  of  dengue  viruses  in  field‐caught Aedes  aegypti  and Aedes  albopictus mosquitoes by a type‐specific polymerase chain reaction and cycle sequencing. Am J Trop Med Hyg 58:578‐586. 

31.  Chua, J. J., M. M. Ng, and V. T. Chow. 2004. The non‐structural 3 (NS3) protein of dengue virus type 2 interacts with human nuclear receptor binding protein and is associated with alterations in membrane structure. Virus Res 102:151‐163. 

32.  Clyde, K.,  J. Barrera, and E. Harris. 2008. The capsid‐coding  region hairpin element  (cHP)  is a critical determinant of dengue virus and West Nile virus RNA synthesis. . Virology 379:314‐323. 

33.  Clyde, K., and E. Harris. 2006. RNA  secondary  structure  in  the  coding  region of dengue  virus type  2  directs  translation  start  codon  selection  and  is  required  for  viral  replication.  J  Virol 80:2170‐2182. 

34.  Clyde,  K.,  J.  L.  Kyle,  and  E.  Harris.  2006.  Recent  advances  in  deciphering  viral  and  host determinants of dengue virus replication and pathogenesis. J Virol 80:11418‐11431. 

35.  Davis, W. G., J. L. Blackwell, P. Y. Shi, and M. A. Brinton. 2007. Interaction between the cellular protein eEF1A  and  the 3'‐terminal  stem‐loop of West Nile  virus  genomic RNA  facilitates  viral minus‐strand RNA synthesis. J Virol 81:10172‐10187. 

36.  De Nova‐Ocampo, M., N. Villegas‐Sepulveda, and R. M. del Angel. 2002. Translation elongation factor‐1alpha,  La,  and  PTB  interact with  the  3'  untranslated  region  of  dengue  4  virus  RNA. Virology 295:337‐347. 

16  

37.  Diallo, M., A. A. Sall, A. C. Moncayo, Y. Ba, Z. Fernandez, D. Ortiz, L. L. Coffey, C. Mathiot, R. B. Tesh, and S. C. Weaver. 2005. Potential role of sylvatic and domestic African mosquito species in dengue emergence. Am J Trop Med Hyg 73:445‐449. 

38.  Diviney, S., A. Tuplin, M. Struthers, V. Armstrong, R. M. Elliott, P. Simmonds, and D. J. Evans. 2008. A hepatitis C virus cis‐acting replication element forms a long‐range RNA‐RNA interaction with upstream RNA sequences in NS5B. 82 18. 

39.  Dos  Ramos,  F.,  M.  Carrasco,  T.  Doyle,  and  I.  Brierley.  2004.  Programmed  ‐1  ribosomal frameshifting in the SARS coronavirus. Biochem Soc Trans 32:1081‐1083. 

40.  Durbin, A., M. J. Vargas, B. Thumar, S. N. Hammond, G. Gordon, C. Rocha, A. Balmaseda, and E. Harris. 2008. Phenotyping of peripheral blood mononuclear cells during acute dengue illness demonstrates  infection  and  increased  activation  of monocytes  in  severe  cases  compared  to classic dengue fever. Virology 376:429‐435. 

41.  Edgil,  D., M.  S.  Diamond,  K.  L.  Holden,  S. M.  Paranjape,  and  E.  Harris.  2003.  Translation efficiency  determines  differences  in  cellular  infection  among  dengue  virus  type  2  strains. Virology 317:275‐290. 

42.  Edgil, D., and E. Harris. 2006. End‐to‐end communication in the modulation of mammalian RNA virus translation. Virus Res 119:43‐51. 

43.  Edgil, D., C. Polacek, and E. Harris. 2006. Dengue virus utilizes a novel strategy for translation initiation when cap‐dependent translation is inhibited. J Virol 80:2976‐2986. 

44.  Effler, P. V., I. Pang, and P. Kitsutani. 2005. Dengue fever, Hawaii, 2001‐2002. Emerg Infect Dis 11:742‐749. 

45.  Egloff,  M.  P.,  D.  Benarroch,  B.  Selisko,  J.  L.  Romette,  and  B.  Canard.  2002.  An  RNA  cap (nucleoside‐2'‐O‐)‐methyltransferase in the flavivirus RNA polymerase NS5: crystal structure and functional characterization. EMBO J 21:2757‐2768. 

46.  Emara, M. M., H. Liu, W. G. Davis, and M. A. Brinton. 2008. Mutation of mapped TIA‐1/TIAR binding sites  in the 3' terminal stem‐loop of West Nile virus minus‐strand RNA  in an  infectious clone negatively affects genomic RNA amplification. J Virol 82:10657‐10670. 

47.  Endy, T. P., S. Chunsuttiwat, A. Nisalak, D. H. Libraty, S. Green, A. L. Rothman, D. W. Vaughn, and  F.  A.  Ennis.  2002.  Epidemiology  of  inapparent  and  symptomatic  acute  dengue  virus infection:   A prospective  study of primary  school children  in Kamphaeng Phet, Thailand. Am  J Epidemiol 156:40‐51. 

48.  Failloux,  A.  B.,  M.  Vazeille,  and  M.  A.  Rodgers.  2002.  Geographic  genetic  variation  in populations of the dengue virus vector Aedes aegypti. J Mol Evol 55:653‐663. 

49.  Falgout, B., and L. Markoff. 1995. Evidence that flavivirus NS1‐NS2A cleavage  is mediated by a membrane‐bound host protease in the endoplasmic reticulum. J Virol 69:7232‐7243. 

50.  Fayzulin, R., and I. Frolov. 2004. Changes of the secondary structure of the 5' end of the Sindbis virus  genome  inhibit  virus  growth  in  mosquito  cells  and  lead  to  accumulation  of  adaptive mutations. J Virol 78:4953‐4964. 

51.  Filomatori, C. V., M. F. Lodeiro, D. E. Alvarez, M. M. Samsa, L. Pietrasanta, and A. V. Gamarnik. 2006. A 5' RNA element promotes dengue virus RNA synthesis on a circular genome. Genes Dev 20:2238‐2249. 

52.  Focks, D., and R. Barrera. 2007. Dengue transmission dynamics: assessment and implications for control. World Health Organization. 

53.  Fosmire,  J.  A.,  K.  Hwang,  and  S.  Makino.  1992.  Identification  and  characterization  of  a coronavirus packaging signal. J Virol 66:3522‐3530. 

54.  Friebe, P., J. Boudet, J. P. Simorre, and R. Bartenschlager. 2005. Kissing‐loop interaction in the 3' end of the hepatitis C virus genome essential for RNA replication. J Virol 79:380‐392. 

17  

55.  Friebe, P., and E. Harris. 2010.  Interplay of RNA elements  in  the dengue virus 5' and 3' ends required for viral RNA replication. J Virol 84:6103‐6118. 

56.  Friebe, P., P. Y. Shi, and E. Harris. 2011. The 5' and 3' downstream AUG  region elements are required for mosquito‐borne flavivirus RNA replication. J Virol 85:1900‐1905. 

57.  Frolov, I., R. Hardy, and C. M. Rice. 2001. Cis‐acting RNA elements at the 5' end of Sindbis virus genome RNA regulate minus‐ and plus‐strand RNA synthesis. RNA 7:1638‐1651. 

58.  Frolova, E., I. Frolov, and S. Schlesinger. 1997. Packaging signals in alphaviruses. J Virol 71:248‐258. 

59.  Gale, M.,  S.‐L.  Tan,  and M. G.  Katze.  2000.  Translational  control  of  viral  gene  expression  in eukaryotes. Microbiol. Mol Biol Reviews 64:239‐280. 

60.  Garcia‐Montalvo, B. M., F. Medina, and R. M. Del Angel. 2004. La protein binds to NS5 and NS3 and to the 5' and 3' ends of Dengue 4 virus RNA. Virus Res 102:141‐150. 

61.  Gibbons, R. V., and D. W. Vaughn. 2002. Dengue: an escalating problem. BMJ 324:1563‐1566. 62.  Giedroc,  D.  P.,  and  P.  V.  Cornish.  2009.  Frameshifting  RNA  pseudoknots:  structure  and 

mechanism. Virus Res 139:193‐208. 63.  Goodfellow,  I., Y. Chaudhry, A. Richardson,  J. Meredith,  J. W. Almond, W. Barclay, and D.  J. 

Evans.  2000.  Identification  of  a  cis‐acting  replication  element  within  the  poliovirus  coding region. J Virol 74:4590–4600. 

64.  Graham,  R.  R., M.  Juffrie,  R.  Tan,  C. G. Hayes,  I.  Laksono,  C. Ma'roef,  Erlin,  Sutaryo,  K.  R. Porter, and S. B. Halstead. 1999. A prospective seroepidemiologic study on dengue  in children four to nine years of age  in Yogyakarta,  Indonesia  I. Studies  in 1995‐1996. Am J Trop Med Hyg 61:412‐419. 

65.  Gratz, N. G. 2004. Critical  review of  the  vector  status of Aedes albopictus. Med Vet Entomol 18:215‐227. 

66.  Gubler, D. J. 1998. Dengue and dengue hemorrhagic fever. Clin Microbiol Reviews 11:480‐496. 67.  Gubler, D. J. 1997. Dengue and dengue hemorrhagic fever: Its history and resurgence as a global 

public  health  problem.  In  D.  J.  Gubler  and  G.  Kuno  (ed.),  Dengue  and  Dengue  Hemorrhagic Fever. CAB International, New York. 

68.  Gubler, D.  J. 2002. Epidemic dengue/dengue hemorrhagic  fever as a public health,  social and economic problem in the 21st century. Trends Microbiol 10:100‐103. 

69.  Gubler,  D.  J.  2002.  The  global  emergence/resurgence  of  arboviral  diseases  as  public  health problems. Arch Med Res 33:330‐342. 

70.  Gubler, D. J., and G. Kuno. 1997. Dengue and dengue hemorrhagic fever. CAB International. 71.  Gustin,  K.  M.,  B.  J.  Guan,  A.  SDziduszko,  and  D.  A.  Brian.  2009.  Bovine  coronavirus 

nonstructural  protein  1  (p28)  is  an  RNA  binding  protein  that  binds  terminal  genomic  cis‐replication elements. J Virol 83:6087‐6097. 

72.  Guzman, M. G., G. Kouri, L. Valdes, J. Bravo, S. Vazquez, and S. B. Halstead. 2002. Enhanced severity  of  secondary  dengue‐2  infections:  death  rates  in  1981  and  1997  Cuban  outbreaks. Revista Panamericana de Salud Publica 11:223‐227. 

73.  Halstead, S. B. 2007. Dengue. Lancet 370:1644‐1652. 74.  Halstead, S. B. 2008. Dengue virus‐mosquito interactions. Annu Rev Entomol 53:273‐291. 75.  Halstead, S. B. 1997. Epidemiology of dengue and dengue hemorrhagic fever. In D. J. Gubler and 

G. Kuno (ed.), Dengue and Dengue Hemorrhagic Fever. CAB International, New York. 76.  Halstead,  S.  B.  1970. Observations  related  to  pathogensis  of  dengue  hemorrhagic  fever.  VI. 

Hypotheses and discussion. Yale J Biol Med 42:350‐362. 77.  Halstead, S. B.,  J. A. Suaya, and D. S. Shepard. 2007. The burden of dengue  infection. Lancet 

369:1410‐1411. 

18  

78.  Han, J. Q., H. L. Townsend, B. K. Jha, J. M. Paranjape, R. H. Silverman, and D. J. Barton. 2007. A phylogenetically  conserved  RNA  structure  in  the  poliovirus  open  reading  frame  inhibits  the antiviral endoribonuclease RNase L. J Virol 81:5561–5572. 

79.  Hawley, W.  A.,  P.  Reiter,  R.  S.  Copeland,  C.  B.  Pumpuni,  and  G.  B.  J.  Craig.  1987.  Aedes albopictus  in North America: probable  introduction  in used  tires  from northern Asia.  Science 236:1114‐1116. 

80.  Heinz, F., G. Auer, K. Stiasny, H. Holzmann, C. Mandl, F. Guirakhoo, and C. Kunz. 1994. The interactions of  the  flavivirus envelope proteins:  implications  for  virus entry  and  release. Arch Virol 9(S):339‐348. 

81.  Heinz, F., K. Stiasny, G. Puschner‐Auer, H. Holzmann, S. Allison, C. Mandl, and C. Kunz. 1994. Structural changes and functional control of the tick‐borne encephalitis virus glycoprotein E by the heterodimeric association with the protein prM. Virology 198:109‐117. 

82.  Helt, A.‐M., and E. Harris. 2005. S‐phase‐dependent enhancement of dengue virus 2 replication in mosquito cells, but not in human cells. J Virol 79:7291‐7299. 

83.  Hofacker, I. L., W. Fontana, P. F. Stadler, S. Bonhoeffer, M. Tacker, and P. Schuster. 1994. Fast folding and comparison of RNA secondary structures. Monatsh Chem 125:167‐188. 

84.  Holden,  K.  L.,  and  E.  Harris.  2004.  Enhancement  of  dengue  virus  translation:  Role  of  the 3’untranslated region and the terminal 3’ stem‐loop domain. Virology 329:119‐133. 

85.  Holden,  K.  L.,  D.  Stein,  T.  C.  Pierson,  A.  Ahmed,  K.  Clyde,  P.  Iverson,  and  E.  Harris.  2006. Inhibition of dengue virus translation and RNA synthesis by a morpholino oligomer to the top of the 3’ stem‐loop structure. Virology 344:439‐452. 

86.  Honda, M., M. R. Beard, L. H. Ping, and S. M. Lemon. 1999. A phylogenetically conserved stem‐loop structure at the 5' border of the internal ribosome entry site of hepatitis C virus is required for cap‐independent viral translation. J Virol 73:1165‐1174. 

87.  Jessie, K., M. Y. Fong, S. Devi, S. K. Lam, and K. T. Wong. 2004. Localization of dengue virus in naturally infected human tissues, by immunohistochemistry and in situ hybridization. J Infect Dis 189:1411‐1418. 

88.  Jumali, Sunarto, D. J. Gubler, S. Nalim, S. Eram, and J. Sulianti Saroso. 1979. Epidemic dengue hemorrhagic fever in rural Indonesia. III. Entomological studies. Am J Trop Med Hyg 28:717‐724. 

89.  Kapoor, M., L. Zhang, M. Ramachandra, J. Kusukawa, K. E. Ebner, and R. Padmanabhan. 1995. Association between NS3 and NS5 proteins of dengue virus type 2 in the putative RNA replicase is linked to differential phosphorylation of NS5 J Biol Chem 270:19100‐19106. 

90.  Khromykh, A. A., H. Meka, K. J. Guyatt, and E. G. Westway. 2001. Essential role of cyclization domains in flavivirus RNA replication. J Virol 75:6719‐6728. 

91.  Khromykh, A. A.,  P.  L.  Sedlak,  and  E. G. Westaway.  2000.  cis‐  and  trans‐acting  elements  in flavivirus RNA replication. J Virol 74:3253‐3263. 

92.  Khromykh, A. A., P. L. Sedlak, and E. G. Westaway. 1999.  trans‐Complementation analysis of the  flavivirus Kunjin ns5 gene  reveals an essential  role  for  translation of  its N‐terminal half  in RNA replication. J Virol 73:9247‐9255. 

93.  Kim, Y. K., S. H. Lee, C. S. Kim, S. K. Seol, and S. K. Jang. 2003. Long‐range RNA‐RNA interaction between  the  5'  nontranslated  region  and  the  core‐coding  sequences  of  hepatitis  C  virus modulates the IRES‐dependent translation. RNA 9:599‐606. 

94.  Kozak, M. 1990. Downstream secondary structure facilitates recognition of  initiator codons by eukaryotic ribosomes. Proc Natl Acad Sci USA 87:8301‐8305. 

95.  Kozak,  M.  1991.  Structural  features  in  eukaryotic  mRNAs  that  modulate  the  initiation  of translation. J Biol Chem 266:19867‐19870. 

96.  Kuno, G. 1997. Factors  influencing  the  tranmsission of dengue viruses.  In D.  J. Gubler and G. Kuno (ed.), Dengue and Dengue Hemorrhagic Fever. CAB International, New York. 

19  

97.  Kurane,  I., and F. A. Ennis. 1992.  Immunity and  immunopathology  in dengue virus  infections. Sem Immunol 4:121‐127. 

98.  Kyle,  J. L., and E. Harris. 2008. Global persistence and spread of dengue. Annu Rev Microbiol 62:71‐92. 

99.  Lee, H., H. Shin, E. Wimmer, and A. V. Paul. 2004. cis‐acting RNA signals in the NS5B C‐terminal coding sequence of the hepatitis C virus genome. J Virol 78:10865‐10877. 

100.  Li, H., S. Clum, S. You, K. E. Ebner, and R. Padmanabhan. 1999. The serine protease and RNA‐stimulated nucleoside triphosphatase and RNA helicase functional domains of dengue virus type 2 NS3 converge within a region of 20 amino acids. J Virol 73:3108‐3116. 

101.  Lin, K. C., H. L. Chang, and R. Y. Chang. 2004. Accumulation of a 3'‐terminal genome fragment in Japanese encephalitis virus‐infected mammalian and mosquito cells. J Virol 78:5133‐5138. 

102.  Lindenbach, B. D., and C. M. Rice. 2001. Flaviviridae: The viruses and their replication, p. 991‐1041.  In D. M. Knipe and P. M. Howley  (ed.), Fields Virology, 4th ed.  Lippincott Williams and Wilkins, Philadelphia. 

103.  Lindenbach, B. D., and C. M. Rice. 2003. Molecular biology of flaviviruses. Adv Virus Res 59:23‐61. 

104.  Linger,  B.  R.,  L.  Kunovska,  R.  J.  Kuhn,  and  B.  L.  Golden.  2004.  Sindbis  virus  nucleocapsid assembly: RNA folding promotes capsid protein dimerization. RNA 10:128‐138. 

105.  Linial, M.  L.,  and  A.  D. Miller.  1990.  Retroviral  RNA  packaging:  sequence  requirements  and implications. Curr Top Microbiol Immunol 157:125‐152. 

106.  Liu, Y., E. Wimmer, and A. V. Paul. 2009. Cis‐acting RNA elements  in human and animal plus‐strand RNA viruses. Biochim Biophys Acta 1789:495‐517. 

107.  Lobert,  P.  E., N.  Escriou,  J. Ruelle,  and  T. Michiels.  1999. A  coding  RNA  sequence  acts  as  a replication signal in cardioviruses. Proc Natl Acad Sci USA 96:11560–11565. 

108.  Lodeiro, M. F., C. V. Filomatori, and A. V. Gamarnik. 2009. Structural and functional studies of the promoter element for dengue virus RNA replication. J Virol 83:993‐1008. 

109.  Lozach, P. Y., L. Burleigh, I. Staropoli, E. Navarro‐Sanchez, J. Harriague, J. L. Virelizier, F. A. Rey, P.  Despres,  F.  Arenzana‐Seisdedos,  and  A.  Amara.  2005.  Dendritic  cell‐specific  intercellular adhesion molecule 3‐grabbing non‐integrin  (DC‐SIGN)‐mediated enhancement of dengue virus infection is independent of DC‐SIGN internalization signals. J Biol Chem 280. 

110.  Mackenzie,  J.  M.,  A.  A.  Khromykh,  M.  K.  Jones,  and  E.  G.  Westaway.  1998.  Subcellular localization  and  some  biochemical  properties  of  the  flavivirus  Kunjin  nonstructural  proteins NS2A and NS4A. Virology 245:203‐215. 

111.  Maeda, A., J. Maeda, H. Takagi, and I. Kurane. 2008. Detection of small RNAs containing the 5'‐ and the 3'‐end sequences of viral genome during West Nile virus replication. Virology 371:130‐138. 

112.  Manzano,  M.,  E.  D.  Reichert,  S.  Polo,  B.  Falgout,  W.  Kasprzak,  B.  A.  Shapiro,  and  R. Padmanabhan. 2011.  Identification of cis‐acting elements  in  the 3'‐untranslated  region of  the dengue  virus  type  2  RNA  that modulate  translation  and  replication.  J  Biol  Chem  286:22521‐22534. 

113.  Markoff,  L., A. Chang,  and B.  Falgout. 1994. Processing of  flavivirus  structural  glycoproteins: stable membrane  insertion  of  premembrane  requires  the  envelope  signal  pepetide.  Virology 204:526‐540. 

114.  Mason, P. W., S. V. Bezborodova, and T. M. Henry. 2002. Identification and characterization of a cis‐acting  replication element  (cre) adjacent  to  the  internal  ribosome entry site of  foot‐and‐mouth disease virus. J Virol 76:9686‐9694. 

20  

115.  McMullan, L. K., A. Grakoui, M. J. Evans, K. Mihalik, M. Puig, A. D. Branch, S. M. Feinstone, and C. M. Rice. 2007. Evidence for a functional RNA element in the hepatitis C virus core gene. Proc Natl Acad Sci USA 104:2879‐2884. 

116.  Men, R., M. Bray, D. Clark, R. M. Chanock, and C.‐J. Lai. 1996. Dengue  type 4 virus mutants containing  deletions  in  the  3'  noncoding  region  of  the  RNA  genome:  analysis  of  growth restriction in cell culture and altered viremia pattern and immunogenicity in Rhesus monkeys. J Virol 70:3930‐3937. 

117.  Miller, W. A., and K. A. White. 2006. Long‐distance RNA‐RNA  interactions  in plant virus gene expression and replication. Annu Rev Phytopathol 44:447‐467. 

118.  Modis,  Y.,  S.  Ogata,  D.  Clements,  and  S.  C.  Harrison.  2004.  Structure  of  the  dengue  virus envelope protein after membrane fusion. Nature 427:313‐319. 

119.  Moncayo, A. C., Z. Fernandez, D. Ortiz, M. Diallo, A. Sall, S. Hartman, C. T. Davis, L. Coffey, C. C. Mathiot,  R.  B.  Tesh,  and  S.  C.  Weaver.  2004.  Dengue  emergence  and  adaptation  to peridomestic mosquitoes. Emerg Infect Dis 10:1790‐1796. 

120.  Morasco, B.  J., N.  Sharma,  J. Parilla,  and  J. B.  Flanegan.  2003.  Poliovirus  cre(2C)‐dependent synthesis of VPgpUpU  is  required  for positive‐ but not negative‐strand RNA  synthesis.  J Virol 77:5136–5144. 

121.  Morens, D. M.,  and A.  S.  Fauci.  2008. Dengue  and  hemorrhagic  fever:  a  potential  threat  to public health in the United States. JAMA 299:214‐216. 

122.  Murray, K. E., and D. J. Barton. 2003. Poliovirus CRE‐dependent VPg uridylylation is required for positive‐strand RNA synthesis but not for negative‐strand RNA synthesis. J Virol 77:4739–4750. 

123.  Navarro‐Sanchez,  E.,  R.  Altmeyer,  A.  Amara,  O.  Schwartz,  F.  Fieschi,  J.  L.  Vierelizier,  R. Arenazana‐Seisdedos,  and  P.  Despres.  2003.  Dendritic‐cell‐specific  ICAM3‐grabbing  non‐integrin  is  essential  for  the  productive  infection  of  human  dendritic  cells  by mosquito‐cell‐derived dengue viruses. EMBO Reports 4:723‐729. 

124.  Nawtaisong,  P.,  J.  Keith,  T.  Fraser,  V.  Balaraman,  A.  Kolokoltsov,  R.  A.  Davey,  S. Higgs,  A. Mohammed, Y. Rongsriyam, N. Komalamisra, and M. J. Fraser, Jr. 2009. Effective suppression of Dengue  fever  virus  in mosquito  cell  cultures using  retroviral  transduction  of hammerhead ribozymes targeting the viral genome. Virol J 6:73. 

125.  Ng, C. Y., F. Gu, W. Y. Phong, Y. L. Chen, S. P. Lim, A. Davidson, and S. G. Vasudevan. 2007. Construction and characterization of a stable subgenomic dengue virus type 2 replicon system for antiviral compound and siRNA testing. Antiviral Res 76:222‐231. 

126.  Niesters, H. G., and J. H. Strauss. 1990. Mutagenesis of the conserved 51‐nucleotide region of Sindbis virus. J Virol 64:1639‐1647. 

127.  Nomaguchi, M., M. Ackermann, C. Yon, S. You, and R. Padmanabhan. 2003. De novo synthesis of negative‐strand RNA by Dengue virus RNA‐dependent RNA polymerase  in vitro: nucleotide, primer, and template parameters. J Virol 77:8831‐8842. 

128.  Novak, J. E., and K. Kirkegaard. 1994. Coupling between genome translation and replication  in an RNA virus. Genes Dev 8:1726‐1737. 

129.  Ou,  J. H., C. M. Rice, L. Dalgarno, E. G. Strauss, and  J. H. Strauss. 1982. Sequence  studies of several alphavirus genomic RNAs in the region containing the start of the subgenomic RNA. Proc Natl Acad Sci USA 79:5235‐5239. 

130.  Panavas, T., C. M. Hawkins, Z. Panaviene, and P. D. Nagy. 2005. The role of the p33:p33/p92 interaction domain  in RNA replication and  intracellular  localization of p33 and p92 proteins of Cucumber necrosis tombusvirus. Virology 338:81‐95. 

131.  Paranjape,  S. M.,  and  E. Harris.  2007.  Y‐box‐binding  protein‐1  binds  to  the  dengue  virus  3'‐untranslated region and mediates antiviral effects. J Biol Chem. 282:30497‐30508. 

21  

132.  Paul, A. V. 2002. Possible unifying mechanism of picornavirus genome replication. ASM Press, Washington DC. 

133.  Phoolcharoen, W.,  and  D.  R.  Smith.  2004.  Internalization  of  the  dengue  virus  is  cell  cycle modulated in HepG2, but not vero cells. J Med Virol 74. 

134.  Pijlman, G.  P., A.  Funk, N.  Kondratieva,  J.  Leung,  S.  Torres,  L.  van  der Aa, W.  J.  Liu, A.  C. Palmenberg,  P.  Y.  Shi, R. A. Hall,  and A. A. Khromykh.  2008. A  highly  structured,  nuclease‐resistant,  noncoding  RNA  produced  by  flaviviruses  is  required  for  pathogenicity.  Cell  Host Microbe 4:579‐591. 

135.  Polacek,  C.,  P.  Friebe,  and  E.  Harris.  2009.  Poly(A)‐binding  protein  binds  to  the  non‐polyadenylated 3' untranslated  region of dengue virus and modulates  translation efficiency.  J Gen Virol 90:687‐692. 

136.  Richards, S.  L.,  L. Ponnusamy, T. R. Unnasch, H. K. Hassan, and C. S. Apperson. 2006. Host‐feeding patterns of Aedes albopictus (Diptera:Culicidae) in relation to availability of human and domestic animals in suburban landscapes of central North Carolina. J Med Entomol 43:543‐551. 

137.  Romi, R. 1995. History and updating on  the spread of Aedes albopictus  in  Italy. Parassitologia 37:99‐103. 

138.  Rosen,  L.  1977.  The  Emperor's New  Clothes  revisited,  or  reflections  on  the  pathogenesis  of dengue hemorrhagic fever. Am J Trop Med Hyg 26:337‐343. 

139.  Rosen, L. 1987. Sexual transmission of dengue viruses by Aedes albopictus. Am J Trop Med Hyg 37:398‐402. 

140.  Rosen, L., and D. Gubler. 1974. The use of mosquitoes to detect and propagate dengue viruses. Am J Trop Med Hyg 23:1153‐1160. 

141.  Rosen,  L.,  L.  E.  Roseboom, D.  J. Gubler,  J.  C.  Lien,  and  B. N.  Chaniotis.  1985.  Comparative susceptibility of mosquito species and strains to oral and parenteral  infection with dengue and Japanese encephalitis viruses. Am J Trop Med Hyg 34:603‐615. 

142.  Rosen, L., D. A. Shroyer, R. B. Tesh, J. E. Freier, and J. C. Lien. 1983. Transovarial transmission of dengue  viruses  by  mosquitoes:  Aedes  albopictus  and  Aedes  aegypti.  Am  J  Trop  Med  Hyg 32:1108‐1119. 

143.  Sabin, A. B. 1950. The dengue group of viruses and its family relationships. Bacteriol Rev 14:225‐232. 

144.  Sachs, A. 2000. Physical and  functional  interactions between  the mRNA cap structure and  the poly(A)  tail,  p.  447‐465.  In  N.  Sonenberg,  J.  W.  B.  Hershey,  and  M.  B.  Matthews  (ed.), Translational  Control  of Gene  Expression,  2nd  ed.  Cold  Spring Harbor  Laboratory  Press,  Cold Spring Harbor, NY. 

145.  Sangkawibha, N.,  S. Rojanasuphot,  S. Ahandrik,  S. Viriyapongse,  S.  Jatanasen, V.  Salitul, B. Phanthumachinda,  and  S.  B.  Halstead.  1984.  Risk  factors  in  dengue  shock  syndrome:  a prospective  epidemiologic  study  in  Rayong,  Thailand.  I.  The  1980  outbreak.  Am  J  Epidemiol 120:653‐669. 

146.  Scherer, L., J. J. Rossi, and M. S. Weinberg. 2007. Progress and prospects: RNA‐based therapies for treatment of HIV infection. Gene Ther 14:1057‐64. 

147.  Scott, T. W., A. Naksathit, J. F. Day, P. Kittayapong, and J. D. Edman. 1997. A fitness advantage for Aedes aegypti and  the viruses  it  transmits when  females  feed only on human blood. Am  J Trop Med Hyg 57:235‐239. 

148.  Shi, P. Y., W. Li, and M. A. Brinton. 1996. Cell proteins bind specifically to West Nile virus minus‐strand 3' stem‐loop RNA. J Virol 70:6278‐6287. 

149.  Simmonds, P., I. Karakasiliotis, D. Bailey, Y. Chaudhry, D. J. Evans, and I. G. Goodfellow. 2008. Bioinformatic  and  functional  analysis  of  RNA  secondary  structure  elements  among  different genera of human and animal caliciviruses. Nucleic Acids Res 36:2530‐2546. 

22  

150.  Smith, C. E. 1956. The history of dengue  in  tropical Asia and  its probable  relationship  to  the mosquito Aedes aegypti. J Trop Med Hyg 59:243‐251. 

151.  Snijder, E.  J., and  J.  J. Meulenberg. 1998. The molecular biology of  arteriviruses.  J Gen Virol 79:961‐979. 

152.  Stadler, K., S. L. Allison, J. Schalich, and F. X. Heinz. 1997. Proteolytic activation of tick‐borne encephalitis virus by furin. J Virol 71:8475–8481. 

153.  Steil,  B.  P.,  and  D.  J.  Barton.  2008.  Poliovirus  cis‐acting  replication  element‐dependent  VPg Uridylylation lowers the Km of the initiating nucleoside triphosphate for viral RNA replication. J Virol 82:9400‐9408. 

154.  Strickman, D.,  and  P.  Kittayapong.  1993.  Loboratory  demonstration  of  oviposition  by  Aedes aegypti (Diptera: Culicidae) in covered water jars. J Med Entomol 30:947‐949. 

155.  Tan, B.‐H., J. Fu, R. J. Sugrue, E.‐H. Yap, Y.‐C. Chan, and Y. H. Tan. 1996. Recombinant dengue type 1 virus NS5 protein expressed in Escherichia coli exhibits RNA‐dependent RNA polymerase activity. Virology 216:317‐325. 

156.  Tarun, S. Z., and A. B. Sachs. 1996. Association of  the yeast poly(A)  tail binding protein with translation initiation factor eIF4G. EMBO Journal 15:7168‐7177. 

157.  Tassaneetrithep, B., T. H. Burgess, A. Granelli‐Piperno, C. Trumpfheller, J. Finke, W. Sun, M. A. Eller, K. Pattanapanyasat, S. Sarasombath, D. L. Birx, R. M. Steinman, S. Schlesinger, and M. A. Marovich. 2003. DC‐SIGN  (CD209) mediates dengue virus  infection of human dendritic cells.  J Exp Med 197:823‐829. 

158.  Thein, S., M. M. Aung, T. N. Shwe, M. Aye, A. Zaw, K. Aye, K. M. Aye, and J. Aaskov. 1997. Risk factors in dengue shock syndrome. Am J Trop Med Hyg 56:566‐572. 

159.  Thurner,  C.,  C. Witwer,  I.  L.  Hofacker,  and  P.  F.  Stadler.  2004.  Conserved  RNA  secondary structures in Flaviviridae genomes. J Gen Virol 85:1113‐1124. 

160.  Townsend, H. L., B. K. Jha, J. Q. Han, N. K. Maluf, R. H. Silverman, and D. J. Barton. 2008. A viral RNA  competitively  inhibits  the  antiviral  endoribonuclease  domain  of  RNase  L.  RNA  14:1026‐1036. 

161.  Travanty, E. A., Z. N. Adelman, A. W. Franz, K. M. Keene, B. J. Beaty, C. D. Blair, A. A. James, and K. E. Olson. 2004. Using RNA interference to develop dengue virus resistance in genetically modified Aedes aegypti. Insect Biochem Mol Biol 34:607‐613. 

162.  Tzeng, W.  P.,  and  T.  K.  Frey.  2002. Mapping  the  rubella  virus  subgenomic  promoter.  J Virol 76:3189‐3201. 

163.  Urdaneta‐Marquez, L., and A. B. Failloux. 2010. Population genetic structure of Aedes aegypti, the principal vector of dengue viruses. Infect Genet Evol 11:253‐261. 

164.  Urdaneta,  L.,  F.  Herrera, M.  Pernalete,  N.  Zoghbi,  Y.  Rubio‐Palis,  R.  Barrios,  J.  Rivero,  G. Comach, M. Jimenez, and M. Salcedo. 2005. Detection of dengue viruses in field‐caught Aedes aegypti  (Dipterra: Culicidae)  in Maracay, Aragua  state, Venezuela by  type‐specific polymerase chain reaction. Infect Genet Evol 5:177‐184. 

165.  Van  Wynsberghe,  P.  M.,  and  P.  Ahlguist.  2009.  5'  cis  elements  direct  nodavirus  RNA1 recruitment to mitochondrial sites of replication complex formation. J Virol 83:2976‐2988. 

166.  Vassilaki, N., P. Friebe, P. Meuleman, S. Kallis, A. Kaul, G. Paranhos‐Baccala, G. Leroux‐Roels, P. Mavromara, and R. Bartenschlager. 2008. Role of the hepatitis C virus core+1 open reading frame  and  core  cis‐acting  RNA  elements  in  viral  RNA  translation  and  replication.  J  Virol 82:11503‐11515. 

167.  Verheije, M. H., R. C. Olsthoorn, M. V. Kroese, P. J. Rottier, and J. J. LMeulenberg. 2002. Kissing interaction between 3' noncoding and coding sequences is essential for porcine arterivirus RNA replication. J Virol 76:1521‐1526. 

23  

168.  Villordo, S. M., D. E. Alvarez, and A. V. Gamarnik. 2010. A balance between circular and linear forms of the dengue virus genome is crucial for viral replication. RNA 16:2325‐2335. 

169.  Villordo,  S. M., and A. V. Gamarnik. 2009. Genome  cyclization  as  strategy  for  flavivirus RNA replication. Virus Res 139:230‐239. 

170.  Wang, T. H., R. C. Rijnbrand, and S. M. Lemon. 2000. Core protein‐coding  sequence, but not core protein, modulates  the efficiency of cap‐independent  translation directed by  the  internal ribosome entry site of hepatitis C virus. J Virol 74:11347‐11358. 

171.  Wei, Y., C. Qin, T. Jiang, X. Li, H. Zhao, Z. Liu, Y. Deng, R. Liu, S. Chen, M. Yu, and E. Qin. 2009. Translational regulation by the 3' untranslated region of the dengue type 2 virus genome. Am J Trop Med Hyg 81:817‐824. 

172.  Weiss, B., U. Geigenmuller‐Gnirke, and S. Schlesinger. 1994. Interactions between Sindbis virus RNAs  and  a  68  amino  acid  derivative  of  the  viral  capsid  protein  further  defines  the  capsid binding site. Nucleic Acids Res 22:780‐786. 

173.  Welsch, S., S. Miller, I. Romero‐Brey, A. Merz, C. K. Bleck, P. Walther, S. D. Fuller, C. Antony, J. Krijnse‐Locker, and R. Bartenschlager. 2009. Composition and  three‐dimensional architecture of the dengue virus replication and assembly sites. Cell Host Microbe 5:365‐375. 

174.  Westaway,  E.  G.,  A.  A.  Khromykh,  and  J.  M.  Mackenzie.  1999.  Nascent  flavivirus  RNA colocalized in situ with double‐stranded RNA in stable replication complexes. Virology 258:108‐117. 

175.  Wu, S. J., G. Grouard‐Vogel, W. Sun, J. R. Mascola, E. Brachtel, R. Putvatana, M. K. Louder, L. Filgueira, M. A. Marovich, H. K. Wong, A. Blauvelt, G. S. Murphy, M. L. Robb, B. L. Innes, D. L. Birx, C. G. Hayes, and S. S. Frankel. 2000. Human skin Langerhans cells are  targets of dengue virus infection. Nature Med 6:816‐820. 

176.  Yang, S. H., M. L. Liu, C. F. Tien, S. J. Chou, and R. Y. Chang. 2009. Glyceraldehyde‐3‐phosphate dehydrogenase  (GAPDH)  interaction  with  3'  ends  of  Japanese  encephalitis  virus  RNA  and colocalization with the viral NS5 protein. J Biomed Sci 16:40. 

177.  Yang, Y., M. Yi, D. J. Evans, P. Simmonds, and S. M. Lemon. 2008. Identification of a conserved RNA  replication  element  (cre)  within  the  3Dpol‐coding  sequence  of  hepatoviruses.  J  Virol 82:10118‐10128. 

178.  Yocupicio‐Monroy, M., R. Padmanabhan, F. Medina, and R. M. del Angel. 2007. Mosquito La protein binds  to  the 3' untranslated  region of  the positive and negative polarity dengue virus RNAs and relocates to the cytoplasm of infected cells. Virology 357:29‐40. 

179.  Yocupicio‐Monroy, R. M.,  F. Medina,  J. Reyes‐del Valle, and R. M. del Angel. 2003. Cellular proteins  from human monocytes bind  to dengue  4  virus minus‐strand 3' untranslated  region RNA. J Virol 77:3067‐3076. 

180.  You, S., and C. M. Rice. 2008. 3' RNA elements  in hepatitis C virus replication: kissing partners and long poly(U). J Virol 82:184‐195. 

181.  You, S., D. D. Stump, A. D. Branch, and C. M. Rice. 2004. A cis‐acting replication element in the sequence encoding the NS5B RNA‐dependent RNA polymerase  is required for hepatitis C virus RNA replication. J Virol 78:1352‐1366. 

182.  Zeng, L., B. Falgout, and L. Markoff. 1998. Identification of specific nucleotide sequences within conserved  3'‐SL  in  the dengue  type  2  virus  genome  required  for  replication.  J Virol  72:7510‐7522. 

183.  Zhong, W., R. Dasgupta, and R. Rueckert. 1992. Evidence that the packaging signal for nodaviral RNA2 is a bulged stem‐loop. Proc Natl Acad Sci USA 89:11146‐11150. 

184.  Zybert,  I.  A.,  H.  van  der  Ende‐Metselaar,  J.  Wilschut,  and  J.  M.  Smit.  2008.  Functional importance of dengue virus maturation:  infectious properties of  immature virions.  J Gen Virol 89:3047‐3051. 

24  

CHAPTER TWO 

IDENTIFICATION AND CHARACTERIZATION OF CODING‐REGION RNA ELEMENTS

25  

INTRODUCTION  Dengue virus  (DENV)  is a member of  the  family Flaviviridae  (5, 21) and  causes a  spectrum of  clinical disease, ranging from an acute febrile  illness known as dengue fever (DF) to the more severe forms of the illness, dengue hemorrhagic fever (DHF) and dengue shock syndrome (DSS).  DENV is an enveloped positive‐sense RNA virus whose genome is ~11 kilobases and encodes 10 viral proteins in a single open reading frame (ORF) that is flanked by highly structured 5' and 3' untranslated regions (UTRs) (7, 11, 32).  While  the DENV  genome  contains  a  5'  type  1  7‐methyl‐G  cap,  unlike  cellular mRNAs, DENV  lacks  a poly(A) tail (7).  The ORF is translated as a single polyprotein that is cleaved co‐ and post‐translationally by both viral and host proteases  (4)  into  three  structural  (capsid‐C, premembrane/membrane‐prM/M and envelope‐E) and seven nonstructural (NS) proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B and NS5).   As with other positive‐sense RNA viruses,  the 5'  (16) and 3' UTRs play a key role  in regulating viral translation (1, 8, 14, 24, 47) and RNA synthesis (1‐3).   In the 5’ UTR, stem‐loop A (SLA) acts as the promoter  element  for  synthesis  of  the  negative‐strand  by  providing  the  RNA‐dependent  RNA polymerase (RdRp) access to the 3' end of the genome during circularization (2, 3, 16).  The panhandle structure that forms during viral replication is mediated by three pairs of circularization sequences, the 5’/3’  conserved  sequence  (CS),  5’/3’  upstream AUG  region  (UAR)  and  5'/3'  downstream AUG  region (DAR), which are required for RNA synthesis though they have no effect on viral translation (2, 3, 8, 18, 19, 27).   Recently  it has been  shown  that viral  replication  requires a balance between  the  linear and circular conformations (45, 46), which could act as a mechanism for coordinating the molecular switch between viral translation and RNA synthesis (45).  Both the 3’ stem‐loop (SL) and the 3’ UTR are required for efficient RNA synthesis (1, 14, 15), and RNA elements  in the 3’ UTR have been shown to positively and negatively  regulate  viral  translation  (8, 14, 24, 47).   Recently,  small  structured non‐coding RNAs derived  from  the  flaviviral 3’UTR  (sfRNAs) were  shown  to  accumulate  in  cells  (31, 39),  causing  virus‐induced  cytopathic  effects  in  cell  culture  and  contribute  to  viral  pathogenicity  in  mice  (39).    The importance of the regulatory RNA elements present within the UTRs  is best demonstrated by the fact that mutations  in either  the 5' and 3' UTR can  reduce viral  replication  in cell culture and mosquitoes, neurovirulence in mice, and viremia in monkeys (6, 36).   Published structure prediction algorithms have shown that the DENV coding‐sequence is rather unstructured  (23, 43),  though  there has been no  functional proof  that  there are no putative  coding‐region RNA sequence elements.  To date, only one RNA element has been identified in the DENV coding‐region, an RNA secondary structure known as the capsid‐coding hairpin (cHP), which regulates both viral translation  and  RNA  synthesis  (9,  10).    Though  the  cHP  is  the  only  known DENV  coding‐region  RNA element, there are several coding‐region RNA elements in other animal and plant viruses that modulate various  aspects  of  the  viral  life  cycle  including  viral  translation,  RNA  synthesis,  vRNA  intracellular trafficking, and viral assembly, as well as other less well characterized stages in viral replication (33, 37).  The most well  characterized  coding‐region  RNA  element  is  the  cis‐acting  replicating  element  (CRE), which  is  a 61‐nt  stem‐loop  found within  the 2C‐coding  sequence of poliovirus  (PV)  (20).   Within  the Picornaviridae family, the CREs can vary in length, and with the exception of the foot‐and‐mouth disease virus, the CREs are preferentially located within the coding‐region (35, 38, 41).  Hepatitis C virus contains several conserved secondary structures within the viral coding‐region,  including the SL9266 within the NS5B‐coding sequence, which is involved in RNA synthesis (17, 30, 48, 49) and has been implicated in a kissing  loop  interaction with a  conserved  SL  in  the 3' UTR  (17, 48) as well as with an upstream RNA sequence  located at position 9110  in the NS5B coding‐region (13).   The cooperative binding of SL9266 with both its 5’ and 3’ complementary sequences increases genomic stability by creating a pseudoknot, though it is presently unknown if the functions of these two kissing loops are connected (13).  Thus far, there have not been many efforts to elucidate the roles of coding‐region RNA elements within the DENV 

26  

genome,  though  evidence  from  other  positive‐strand  RNA  viruses  suggests  that  coding‐region  RNA elements can have diverse regulatory roles in the viral life cycle.   This study sought  to  identify whether additional RNA elements were  located within  the DENV coding  region with  the  goal  of  expanding  our  knowledge  of  how  the  intracellular  viral  life  cycle  is regulated.   Using sequence alignments and secondary structure conservation algorithms as a platform for  the  identification  of  novel  RNA  elements,  three  candidate  RNA  structural  elements  and  three candidate RNA sequence elements were  identified  in the DENV capsid‐, prM‐ and NS5‐coding regions.  Initial  screening  was  conducted  in  mammalian  and  mosquito  cells  in  order  to  assess  the  mutant constructs for their impact on viral infectivity, though only one of the original six putative RNA elements was shown to modulate the viral  life cycle.   The novel RNA sequence element  identified  in these early screens was called the conserved capsid‐coding region 1 (CCR1), and preliminary evidence indicates that minor  nucleotide  changes  to  CCR1  impact  viral  spread  in mammalian  cells  to  some  degree  but  can significantly delay viral replication in mosquito cells.  Studies such as these are important for improving our understanding of cis‐acting determinants that differentially regulate the viral  life cycle  in both the human host and mosquito vector, which could help in the design of attenuated vaccine strains or serve as candidate targets for antiviral agents.   

MATERIALS AND METHODS  Sequence  conservation  and  secondary  structure  prediction  algorithms.    For  conserved  sequence prediction,  sequences  were  aligned  using  the  Clustal  W2  web  software  (42)  from  the  European Bioinformatics Institute.   Portions of the Clustal W2 alignments were processed by the RNAalifold web server (22) from the Institute of Theoretical Chemistry at the University of Vienna in order to generate the phylogenetically  conserved  secondary  structure predictions.    Individual RNA  secondary  structures were predicted using  the mfold 3 web server  (50)  from  the RNA  Institute at  the University at Albany.  Reported ΔG  values  reflect  enf  refinement  (the mfold  free‐energy  computation  incorporating  coaxial stacking and the Jacobson‐Stockmeyer theory for multibranched loops) (50).  Construction of mutant DNA contructs.   Mutations were  introduced  into either the capsid‐, prM‐ and NS5‐coding region or the 3’ UTR of an infectious clone (IC) of the DENV2 Thai strain 16681 (pD2/IC‐30P, hereafter pD2/IC; a gift of R. Kinney, Centers for Disease Control and Prevention, Fort Collins CO) using overlap  extension‐PCR  (SOE‐PCR).    For mutations  introduced  into  either  the  capsid‐  or  prM‐coding sequence, the T7 promoter and the first 1,391 nucleotides (nts) of DENV2 were amplified, whereas for mutations  introduced  into  the NS5‐coding  sequence or  the 3’ UTR,  the  last 1,867 nts of DENV2 were amplified.  Resulting products were either digested with SacI/SphI (for mutations in capsid and prM) and ligated into a SacI/SphI‐digested pD2/IC as previously described (9, 10) or digested with BsrGI/XbaI (for mutations  in  NS5  and  the  3’  UTR)  and  ligated  into  a  BsrGI/XbaI‐digested  pD2/IC.    For  the  Renilla luciferase reporter constructs, the T7 promoter and the first 167 nts (for mutations in the first 72 nts of capsid)  were  amplified  and  resulting  products  were  digested  with  NotI/MluI  and  ligated  into  a NotI/MluI‐digested pDRrep as previously described  (9).   For mutations  in NS5 and the 3’ UTR, the  last 5,298 nts of the mutant IC were digested with XhoI/XbaI and ligated into a XhoI/XbaI‐digested pDRrep.  Primer sequences are provided in Table 2.1.  Generation  of  RNA  templates.    For  all  DNA  templates  (pD2/IC,  pDRrep  and  pDRrep‐RdRpmu  (9)), linearized  templates  were  generated  by  digestion  with  XbaI  and  purified  using  a  25:24:1 phenol/chloroform/isoamyl  alcohol extraction  (pH 6.7 + 0.2,  Fischer  Scientific).   RNA  templates were generated via in vitro transcription using the RiboMax Large Scale RNA Production System (T7, Promega) 

27  

with the following modifications to the manufacturer’s protocol: 5 mM each GTP, CTP and UTP; 1 mM ATP; 5 mM m7G(5’)ppp(5’)A cap analog  (New England Biolabs) were  incubated  for 4 hours  (h) at 30oC with the addition of 2 mM ATP after 30 minutes (min).  RNAs were subsequently treated with 80 U/mL TURBO DNase (Ambion) for 15 min at 37oC and unincorporated nucleotides were removed from DENV RNAs by size exclusion chromatography using Micro Bio‐Spin P‐30 Tris columns (BioRad Laboratories).  Cell  culture.    Baby  hamster  kidney  (BHK‐21,  clone  15;  hereafter  BHK)  cells were  grown  in minimal essential medium‐alpha  (MEMα, Gibco) with 5% fetal bovine serum (FBS; HyClone) at 37oC  in 5% CO2.  For  infections and preliminary  transfections  in Ae. albopictus mosquito  (C6/36, ATCC #CRL‐1660) cells were grown  in Leibovitz’s medium (L‐15, Gibco) with 10% FBS at 28oC without CO2.   For more efficient transfections  in  C6/36  cells,  a  high  temperature  (HT)  variant  was  utilized  (gift  from  A.  Garmanik, Fundación  Instituto  Leloir, Buenos Aires Argentina) and grown  in  L‐15 medium with 10% FBS at 33oC without CO2.  All media was supplemented with 10 mM HEPES (pH 7.5), 100 units/mL penicillin and 100 µg/mL streptomycin.  Virus  titration by plaque  assay.   Plaque  assays using BHK  cells  for  virus  titration were  conducted  as described previously  (9, 10, 12).   Viral  titer was calculated as plaque  forming units  (pfu) per mL.   For transfections, viral titers were normalized to either  the amount of viral RNA (vRNA)  in the cells at 2 h post‐transfection  (hpt,  relative  transfection  efficiency)  or  at  the  time  that  the  virus‐containing supernatants were harvested, as determined by quantitative reverse transcription (qRT)‐PCR.   p‐values relative to the WT control were derived by Wicoxon rank sum test in Mstat.  RNA transfections.  BHK cells were seeded in 24‐well plates, grown to 50% confluence, and transfected with 0.5 µg IC or replicon RNA using 4X Lipofectamine 2000 (Invitrogen), followed by a 2‐h incubation at 37oC in 5% CO2.  BHK cells were washed 3X in MEMα medium and incubated at 37oC with 5% CO2 until harvest at  indicated  timepoints.   C6/36 cells were seeded  in 12‐well plates, grown  to 70% confluence and  transfected with 1 µg  IC RNA using 8X TransMessenger  (Qiagen),  followed by a 3‐h  incubation at 28oC without CO2.   C6/36 cells were washed 1X  in L‐15 media and  incubated at 28oC without CO2 until harvest  at  indicated  timepoints.    C6/36  HT  cells  were  seeded  in  12‐well  plates,  grown  to  95% confluence,  and  transfected with  1  µg  IC  or  replicon  RNA  using  6X  Lipofectamine  2000  (Invitrogen), followed by a 3‐h  incubation at 33oC without CO2.   C6/36 HT cells were washed 1X  in L‐15 media and incubated  at  33oC without  CO2  until  harvest  at  indicated  timepoints.    Virus‐containing  supernatants were collected at  the  indicated  timepoints and viral  titer was assessed by plaque assay.   Cellular RNA was  harvested  from  a  duplicate  well  at  either  2  hpt  or  at  the  same  time  as  the  virus‐containing supernatants  and  purified  using  the  Mini  RNA  Isolation  II  Kit  (Zymo  Research)  so  that  vRNA concentrations could be determined using qRT‐PCR.  Viral infections.  Virus‐containing supernatants were harvested at 72 hpt from IC‐transfected BHK cells.  Virus‐containing  supernatants were  concentrated by ultracentrifugation at 89,500 g  for 2 h, and viral titer was assessed by plaque assay.  BHK and C6/36 cell monolayers were infected at an MOI of 0.01, as determined by plaque assay.  Infections were incubated at 37oC with 5% CO2 (BHK) or 28

oC without CO2 (C6/36), and at 2 (BHK) or 3 (C6/36) hours post infection (hpi), cells were washed 1X with MEMα (BHK) or  L‐15  (C6/36) medium and  incubated at 37oC with 5% CO2  (BHK) or 28

oC without CO2  (C6/36) until harvest  at  indicated  timepoints.    Virus‐containing  supernatants  were  collected,  and  viral  titer  was assessed by plaque assay.  Luciferase assay.  BHK cells transfected transfected with the Renilla luciferase reporter constructs were harvested  at  the  indicated  timepoints,  and  samples were  analyzed with  the  Renilla  Luciferase Assay 

28  

System  (Promega)  according  to  the  manufacturer’s  instructions  on  a  GloMax‐96  Microplate Luminometer (Promega) with the following modifications to the manufacturer’s instructions: 50 µL cell lysate and 40 µL of 1X Renilla Luciferase Assay Substrate was used for each sample.   Translation  levels were normalized  to  the  relative  transfection efficiency at 2 hpt as determined by qRT‐PCR.   p‐values relative to the WT control were derived by Wicoxon rank sum test in Mstat.  Quantitative  RT‐PCR.    Cellular  RNA was  harvested  at  indicated  timepoints  from  duplicate wells  and purified  using  the Mini  RNA  Isolation  II  Kit  (Zymo  Research).    qRT‐PCR was  first  conducted  using  a modification to the method of Houng et al. (26), which uses a FAM‐TAMRA labeled probe and primer set that  is  directed  against  the  3’  UTR  (Table  2.2).    Due  to  the  discovery  of  sfRNAs  (39),  the  qRT‐PCR procedure was modified using the Laue et al. (29) method, which employs a FAM‐TAMRA labeled probe and primer set that is directed against the viral protein NS5 (Table 2.2).  qRT‐PCR was conducted using the TaqMan One‐Step Master Mix (Applied Biosystems) system with the following modifications to the protocol: the RT step was conducted at 48oC for 30 min, one denaturation cycle was included at 95oC for 10 min, annealing  temperature was set at 60oC, and  the reaction volume was  fixed at 20 µL.   For  the purposes of determining transfection efficiency (2 hpt or at the time of virus‐containing supernatants), target vRNA was  first normalized  to  the  cellular 18S RNA using  the TaqMan VIC‐MGB Primer  Limited Eukaryotic  18S  rRNA  Endogenous  Control  (Applied  Biosystems)  in  a  parallel  reaction.    qRT‐PCR was conducted on a Sequence Detection System 7300 (Applied Biosystems).  Statistical  analysis.   Wilcoxon  rank  sum  test were  performed  in Mstat.    Statistical  significance was defined as p<0.05.   

RESULTS  The  identification  of  novel  RNA  elements  in  the  DENV  coding‐region  based  on  sequence  and secondary structure conservation.   The  selection of candidate coding‐region RNA  sequence elements was based on an alignment of sequences from the representative genomes of DENV1‐4, generated using Clustal W2  (42).   Putative RNA  sequence elements were defined by comparing  the  level of homology within  a  given  stretch  to  that  of  the  background  sequences, which were  defined  as  30  nucleotides upstream and downstream of  the putative element,  requiring at  least 70% homology  in  the  region of interest.   As  a  control,  the 5’  and 3’  conserved  sequence  (CS,  Fig. 2.1A  and  2.1B) were  analyzed  for comparison, demonstrating 87% sequence homology in the 5’ region (~67% background, Fig. 2.1A) and 100% homology in its 3’ complementary sequence (~90% background, Fig. 2.1B).  Based on the DENV1‐4 sequence alignment, two novel RNA sequence elements were identified.  In the prM‐coding sequence, a region  termed  the  prM‐coding  conserved  region  1  (prMCR1)  was  identified  with  73%  sequence homology (~59% background, Fig. 2.1C), whereas in the NS5‐coding sequence, a region called the NS5‐coding conserved region 1  (NS5CR1) was  found with 83% sequence homology  (~58% background, Fig. 2.1D).   When  an  alignment of  sequences  from  the  representative  genomes of  viruses  in  the  closely‐related  Japanese  encephalitis  virus  (JEV)  serogroup was  examined  using  Clustal W2  (42),  the  same candidate  RNA  sequence  elements  were  identified,  though  the  levels  of  sequence  homology  were reduced even among  the control  sequences.    In  the  JEV  serogroup alignment,  the 5’ CS  showed 73% homology  (~32%  background,  Fig.  2.1E), while  the  3’  CS  had  56%  homology  (~42%  background,  Fig. 2.1F).    Likewise,  the  candidate  RNA  sequence  element  prMCR1  displayed  60%  homology  (~40% background,  Fig.  2.1G),  which  was  only  slightly  lower  than  the  candidate  RNA  sequence  element NS5CR1,  which  showed  69%  homology  (~45%  background,  Fig.  2.1H).    Although  the  putative  RNA sequence elements were  less conserved  in  the  JEV serogroup  than  they were  in  the DENV serogroup, 

29  

they had comparable levels of homology when compared to the 5’ and 3’ CS control sequences.  Based on the high degree of sequence conservation identified in both the DENV and JEV serogroups, prMCR1 and NS5CR1 were chosen as candidates RNA sequence elements (Fig. 2.1C and 2.1D).   Since  the  function  of  viral  RNA  elements  can  be  structure‐dependent  rather  than  sequence‐dependent, analyzing  the sequence alignments was not sufficient  to  identify novel coding‐region RNA elements.    To  this  end,  the  sequence  alignments  were  modified  and  examined  for  structural conservation using a prediction algorithm called RNAalifold (22).  The closely‐related DENV, JEV and tick‐borne encephalitis virus (TBEV) serogroup alignments were modified so as to  include either the 5’ UTR and  a  portion  of  the  capsid‐coding  region  or  a  portion  of  the NS5‐coding  region  and  the  3’ UTR  to generate  phylogenetically‐conserved  structure  predictions  using  RNAalifold  (22).    The  conserved secondary structure predictions showed that within the DENV serogroup, there were two candidate RNA structural elements in the capsid‐coding region, termed capsid‐1 (C‐1) and C‐2 (Fig. 2.2A).  Interestingly, both C‐1 and C‐2 were  found to be conserved amongst members of the  JEV  (Fig. 2.2B) and TBEV  (Fig. 2.2C)  serogroups.   Much  like  with  the  known  RNA  structure  element  cHP,  the  primary  nucleotide sequence of C‐1 and C‐2  is not well conserved  in either  the DENV or  JEV serogroups, despite  the  fact that the predicted secondary structures are conserved (data not shown).  There was one candidate RNA structural element identified in the NS5‐coding region, called NS5‐2 (Fig. 2.2D), though it was not found to  be  conserved  in  either  the  JEV  or  TBEV  serogroups  (data  not  shown).    The  predicted  secondary structures  of  C‐1,  C‐2  and  NS5‐2  showed  the  same  level  of  covariance  as  the  predicted  secondary structures of cHP and to a lesser extent than the SLA (Fig. 2.2).  Amino acid (aa) conservation alone is not sufficient  to explain  the observed sequence and structure conservation of all  the  identified candidate RNA elements  in  the DENV,  JEV and TBE  serogroups.   Taken  together,  it  is  likely  that  the novel RNA elements  that were  identified are conserved  in either  their  sequence or  secondary  structure because they may have a regulatory role in mediating the viral life cycle.  Designing mutant  constructs  and  optimizing  the  transfection  protocol  in  BHK  cells.    To  determine whether  the  chosen  candidate  RNA  elements  regulate  the  viral  life  cycle,  various  mutations  were introduced  into an  infectious clone (IC) of the DENV2 Thai strain 16681 (28).   To disrupt the predicted secondary  structure of C‐1 with minimal changes  to  the primary nucleotide  sequence, nine  sequence changes were introduced into either side of the predicted stem structure, which created a side bulge as predicted by  the mfold prediction software  (50), however, these mutations also  introduced conserved aa changes (“C1aa mut,” Fig. 2.3A, Table 2.3).   Since even conserved aa changes will alter the protein‐coding sequence,  it  is possible that changes to the aa sequence could mask an effect  that  the altered secondary  structure may  have  on  the  viral  life  cycle.    Therefore,  eight  nucleotide mutations  were introduced into C‐1 so as to restore the wild‐type (WT) stem‐loop structure and maintain the aa changes that were  introduced  into  the parental “C1aa mut”  construct  (“C1aa cmut,” Fig. 2.3A, Table 2.3).   By restoring the WT structure but maintaining the mutant aa sequence, it was possible to complement the structural defect to determine if the aa changes in "C1aa mut" are important for the viral life cycle.  To determine whether the predicted secondary structure rather than the aa sequence of C‐1  is  important for  the  viral  life  cycle,  silent  point  mutations  were  introduced  to  disrupt  the  predicted  secondary structure  (“C1sil  cmut”  and  “C1sil mut,”  Fig.  2.3A,  Table  2.3).   However,  one mutant  construct was designed  so  it would  have  the  same mutant  structure  as  the  parental  “C1aa mut”  construct  (“C1sil cmut”)  in order to complement the aa changes  in the original mutant "C1aa mut" while examining the structural changes for their impact on the viral life cycle.  The second structure mutant was designed so as to determine whether different structural changes in C‐1 would have different impacts on its role, if any, on the viral life cycle ("C1sil mut").   To  disrupt  the  predicted  secondary  structure  of  C‐2  with  minimal  changes  to  the  primary nucleotide sequence, ten mutations were introduced to minimize the predicted stem‐structure, though 

30  

these sequence changes also introduced conserved aa changes (“C2aa mut,” Fig. 2.3B, Table 2.4).  Since C‐2  is directly downstream of another predicted stem‐loop, the "C2aa mut" construct was designed so that  the  mutations  would  not  affect  neighboring  RNA  elements  but  would  stably  disrupt  the  C‐2 structure as predicted by mfold (50).  For NS5‐2, thirteen nucleotide changes were introduced to disrupt its predicted secondary structure by creating two smaller stem‐loops that share a common stem, though these changes also  introduced conserved aa changes (“NS5‐2aa mut,” Fig. 2.3C, Table 2.5).   Much  like with  C‐1,  to  determine whether  the  aa  sequence  rather  than  the  predicted  structure  of NS5‐2 was important  for  the viral  life  cycle, eleven nucleotide  changes were made  in a  complementary mutant, which  restored  the WT  structure  while maintaining  the  aa  changes  that  were  introduced  into  the parental  “NS5‐2aa mut”  construct  (“NS5‐2aa  cmut,”  Fig.  2.3C,  Table  2.5).    To  examine whether  the predicted secondary structure of NS5‐2  is  important  for  the viral  life cycle,  five silent point mutations were introduced to disrupt the predicted secondary structure without affecting the aa sequence (“NS5‐2sil mut,” Fig. 2.3C, Table 2.5).   For the putative RNA sequence element, prMCR1, three nucleotide changes were introduced so as to disrupt the primary nucleotide sequence while  introducing conserved aa changes (“prMaa mut,” Table 2.6).   However, to examine the effect that conserved aa changes would have on the prM‐coding sequence, an additional mutant construct was designed with nine silent point mutations (“prMsil mut,” Table 2.6).  A similar approach was taken with the putative RNA sequence element NS5CR1.  However, there  were  eleven  sequence  changes  introduced  in  the  “NS5aa  mut”  construct,  which  contained conserved  aa  changes  (Table  2.7)  and  thirteen  silent  point mutations  in  the  “NS5sil mut”  construct (Table 2.7).   A series of control constructs were designed for these studies,  introducing mutations  into either the 5' or 3' CS.  While three nucleotide changes were intended to be introduced into the 5’ CS so as to disrupt the 5' and 3' CS interaction (“5’CSmut,” Table 2.8), due to technical difficulties, two of these mutations occurred  in a portion of  the  capsid‐coding  sequence  that  is  just downstream of  the 5' CS.  Since none of the mutations affected the aa sequence of capsid, a mutant construct was designed with one nucleotide change in the 3' CS, which would restore basepair complementarity with the “5’CS mut” (“3’CSmut,” Table 2.8).   A third mutant construct was created that contained the mutations from both the  "5'CSmut"  and  "3'CSmut"  constructs, which would  restore  the  5’/3’  CS‐mediated  circularization interaction  (“CSmut  comp,”  Table  2.8).    An  additional mutant  was  created  that  contained  the  two nucleotide  changes  that  were  introduced  downstream  of  the  5'  CS  in  the  “5'CSmut”  construct ("CCR1orig mut," Table 2.8).   For all the designed mutant constructs used  in this study, reading  frame and  codon usage bias  in baby hamster  kidney  (BHK)  and Ae. albopictus mosquito  (C6/36)  cells were maintained while preserving the surrounding secondary structures as predicted by the mfold (50).   RNA transcripts derived from the WT  IC were transfected  into BHK cell monolayers, and virus‐containing supernatants were harvested at 24, 48 and 72 hours post‐transfection (hpt) to be assayed for viral  infectivity using plaque assays.    Initially,  the viral  titers  (pfu/mL) generated  in  the plaque assays were normalized to the relative transfection efficiency by extracting cellular RNA from BHK cells at 2 hpt.  The cellular RNA extract was then purified and the amount of viral RNA (vRNA) in the cells quantified via qRT‐PCR using a primer/probe set directed against the 3’ UTR (1‐3, 9, 10).  However, recent studies have shown  that  small  structures RNAs derived  from  the 3’UTR  (small  flaviviral RNAs,  sfRNAs)  accumulate over time, which can lead to cytopathic effects in cells (39).  While vRNA samples are harvested at 2 hpt, which  is before RNA  synthesis occurs  (1),  to avoid  the possibility  that normalizing  viral  titers using a primer/probe  set  directed  against  the  3’  UTR  could  skew  results,  a  pilot  study was  conducted  that compared the 3' UTR primer/probe set (26) with a primer/probe set directed against the viral NS5 gene (29).   We  determined  that  there  is  no  difference  in  viral  titers  that  have  been  normalized  using  a primer/probe set directed against either the 3’ UTR or the viral protein NS5 at 2 hpt (Fig. 2.4A), which implies that either primer/probe set could be used to normalize the transfection data in this study.   

31  

  During the course of the normalization study, we found that there were minor inconsistencies in the calculated viral titer in independent experiments when normalizing the plaque assays to the relative transfection efficiency at 2 hpt regardless of which primer/probe set was utilized (Fig. 2.4B).  While the variation was minor when compared to the raw viral titers (Fig. 2.4B), it is possible that small significant differences  in viral titer could be obscured by the variability  introduced by  the normalization method.  At  the  time of  this study,  there were  four possible methods  to normalize  transfection experiments of this nature,  including normalizing to the transfection efficiency at 2 hpt, normalizing to the amount of vRNA  in  the cells at  the  time  that  the virus‐containing supernatant  is harvested, normalizing  the viral titer to the amount of vRNA in the supernatant at the time of harvest, or making comparisons based off the raw viral titers.   Raw viral titers were used as a standard ("Raw Titer") with which to compare the variability  inherent to each of the different normalization techniques.   We chose to compare raw viral titer to the standard normalization technique ("2 hpt") as well as to the technique that examined how much virus is actively being replicated by the cells at the time of supernatant harvest ("vRNA in Cells").  The  alternate  method  of  normalizing  the  viral  titer  to  the  amount  vRNA  in  the  virus‐containing supernatant  at  the  time  of  harvest  is  used  to  calculate  a  particle‐to‐plaque  forming  unit  (pfu)  ratio, which is often used to determine how much of the viral supernatant is infectious rather than reflect the amount  virus  being  produced.    For  this  reason,  this  technique  was  not  considered  as  a  possible normalization  technique.   While normalizing  viral  titer  to  the  transfection  efficiency  at 2 hpt did not display  a  significant  difference  compared  to  the  other  normalization method  tested,  it  did  slightly amplify the viral titers at all the timepoints tested.  On the other hand, normalizing the viral titers to the amount of vRNA in the cells at the time that the viral supernatants are harvested showed less variability across all experiments and was therefore chosen as the new standard (Fig. 2.4B).  Candidate RNA elements are examined for viral infectivity in BHK cells.  RNA transcripts for the WT and mutant  ICs  were  transfected  into  BHK  cell  monolayers,  and  virus‐containing  supernatants  were harvested at 24, 48 and 72 hpt to be assessed for viral  infectivity using plaque assays.   All mutant C‐1 constructs  that  had  conserved  aa  changes  (“C1aa mut”  and  “C1aa  cmut”)  had  either  a  no‐plaque phenotype or a viral titer that was barely above the  limit of detection (Fig. 2.5A).   On the other hand, both mutant constructs where  the predicted secondary structure was disrupted without affecting  the protein‐coding sequence (“C1sil cmut” and “C1sil mut”) resulted  in WT  levels of replication (Fig. 2.5A), indicating that the predicted secondary structure of C‐1  is not  important  for the viral  life cycle  in BHK cells.  However, these results indicate that even minor conserved aa changes in this region of the capsid‐coding sequence can have drastic consequences on the viral life cycle.  In contrast, changes to both the aa  sequence  and  the predicted  secondary  structure of C‐2  (“C2aa mut”) had no  effect on  viral  titer, which  implies that while the C‐2 stem structure  is not a putative RNA element and that this portion of the capsid‐coding sequence is not as sensitive to aa changes as C‐1 (Fig. 2.5B).  NS5‐2 behaved much like C‐1,  in  that  the protein‐coding  sequence of NS5‐2 was  shown  to be more  important  for  the viral  life cycle  than  its  predicted  secondary  structure,  since  both  the  “NS5‐2aa  mut”  and  “NS5‐2aa  cmut” constructs had a no‐plaque phenotype and while  the  “NS5‐2sil mut”  construct had no effect on viral titer  (Fig.  2.5C).    Similar  results were  observed when  aa  changes were  introduced  into  the  prMCR1 (“prMaa mut,” Fig. 2.5D) and NS5CR1 (“NS5aa mut,” Fig. 2.5E) regions of the genome, where mutations to the aa sequence resulted in a no‐plaque phenotype and this defect in viral titer was restored to WT levels when silent point mutations were  introduced  instead (“prMsil mut,” Fig. 2.5D; “NS5sil mut,” Fig. 2.5E), indicating that neither prMCR1 or NS5CR1 are putative RNA sequence elements in BHK cells.   Changes to the 5’ CS had the expected result of producing a no‐plaque phenotype (“5’CSmut,” Fig. 2.5F), though the changes to the 3’ CS were not sufficient to completely diminish viral titer  in BHK cells (“3’CSmut,” Fig. 2.5F), which supports the hypothesis that the 3’ circularization sequence  is more forgiving to sequence changes than  its 5’ complementary sequence  (2, 3, 19, 27).   As expected, when 

32  

the  5’  and  3’ mutant  CS  sequences  were  combined  in  one  complementary mutant  construct  that restored base‐pair complementarity (“CSmut comp”), the viral titer was restored to close to WT  levels (Fig. 2.5F)  though unexpectantly,  the  resulting virus had a  small plaque phenotype  (Fig. 2.5G).   Given that the 5’/3’ CS have been studied extensively (1, 27) and none of the associated mutants displayed a small  plaque  phenotype,  it  was  thought  that  the  diminished  plaque  size  could  be  due  to  the  two sequence changes that were  introduced  into the region  just downstream of the 5’ CS.   Analysis of the region  just  downstream  of  the  5'  CS  showed  a  high  degree  of  sequence  and  secondary  structure conservation  in both  the DENV and TBEV  serogroups  (addressed  in Chapter 3),  so  it was  termed  the conserved capsid‐coding region 1  (CCR1).   The  two nucleotide changes  that were outside of  the 5’ CS region  in the “5’CSmut” construct were  isolated  in separate mutant construct, termed “CCR1orig mut” (Table 2.8).  While the “CCR1orig mut” construct had no effect on viral titer (Fig. 2.5F), it demonstrated the same small plaque phenotype observed with the “CSmut comp” construct (Fig. 2.5G), indicating that CCR1 might be a putative RNA sequence element, though these small sequence changes only affected plaque morphology in BHK cells.  Examining whether  the putative RNA elements play a  role  in  the  viral  life  cycle  in C6/36  cells.   As transfections bypass key steps  in viral replication, such as entry and uncoating, the mutant constructs were examined in the context of viral infections.  Since the "C2aa mut" construct did not affect the viral titer, despite having aa  changes  in  the  capsid‐coding  region,  it was not used  for  further  investigation since  it  was  not  likely  to  be  a  putative  RNA  sequence  element.    Instead,  only  the  candidate  RNA structure  elements  (C‐1  and  NS5‐2)  that  had  mutant  constructs  with  silent  point  mutations  were examined.  The "CCR1orig mut" construct was included in these studies since it was the only candidate RNA sequence element that showed a distinct phenotype  in  IC‐transfected BHK cells.   Virus‐containing supernatants were harvested from IC‐transfected BHK cells at 72 hpt, concentrated, titered and used to infect BHK cell monolayers at an MOI of 0.01.  Virus‐containing supernatants were harvested at 48 hours post‐infection  (hpi)  before  being  assayed  for  viral  infectivity  by  plaque  assay.    Given  the  technical difficulties  that existed at  the  time with  respect  to  transfecting C6/36 cells,  the  infection experiments provided an opportunity  to examine  the effect of  the putative RNA elements on  the viral  life cycle  in mosquito  cells.   As DENV  replication  kinetics  are  slower  in C6/36  cells,  virus‐containing  supernatants derived  from  infected C6/36  cells were assessed  for viral  infectivity at 120 hpi.   While  the  “CCR1orig mut”  virus maintains  a  small plaque phenotype  in  infected BHK  cells, none of  the  remaining mutant viruses showed a significantly  lower viral  titer as compared  to  the WT  in  infected BHK cells  (Fig. 2.6), implying  that neither C‐1 or NS5‐2 plays a  role  in entry and uncoating.   The  infections  in C6/36  cells confirm that NS5‐2 is not important for the viral life cycle in any of the tested cell lines, though the C‐1 results were conflicting (Fig. 2.6).   The “C1sil cmut” virus had a significantly  lower viral titer  in  infected C6/36 cells but the “C1sil mut” mutant virus had WT  levels of replication (Fig. 2.6),  indicating that this RNA element may have minor effects on viral titer in C6/36 cells.  Since both mutant constructs (“C1sil mut” and “C1sil cmut”) have distinct structural changes but vary slightly  in  their ability  to  replicate  in C6/36 cells, C‐1 may have a sequence‐dependent rather than structure‐dependent effect on early steps in the viral life cycle.  In contrast, the “CCR1orig mut” mutant virus displayed a small plaque phenotype and a significantly  lower viral titer that was barely above the  limit of detection  in  infected C6/36 cells, implying  that  this  region of  the  capsid‐coding  sequence has a  far more  significant effect  in mosquito cells (Fig. 2.6).   As stated earlier, C‐1, NS5‐2, and CCR1 were examined for their effect on the viral  life cycle  in mosquito cells using  infections due to technical difficulties associated with conducting transfections  in C6/36 cells.   During the course of this study, the procedure for transfecting BHK cells was successfully adapted for use in C6/36 cells (Fig. 2.7A), though this initial procedure was inconsistent and detectable virus could only be measured after 96 hpt.  Infected C6/36 cells produced detectable viral titers at 48 hpi 

33  

though the levels are only slightly higher than the limit of detection at this time (data not shown), so the transfected C6/36  should produce detectable  virus by  48 or  72 hpt.    Furthermore,  viral  titers  in  the transfected C6/36 cells were ~2‐3 logs lower than the titers achieved during viral infections at the same timepoint (data not shown), indicating that viral production after the procedure was inefficient.  A high temperature (HT) variant of the C6/36 cell line was shown to be more easily and efficiently transfected with  ICs  (40),  though  the  transfection  reagent caused massive cell death at 24 hpt, which  resulted  in inability  to consistently reproduce  the  transfections.   After additional modifications,  the C6/36 HT cell line was  transfected with  the WT  IC, producing detectable high quantities of viral  titer at 72 hpt  (Fig. 2.7A).  Once the protocol was more efficient and reproducible, the remaining mutant constructs, “prMsil mut” and “NS5sil mut,” could be examined for their effect on the viral life cycle in C6/36 cells since they had not been  tested  in  infected mosquito  cells.    prMCR1  and NS5CR1 were  examined  alongside  the “CCR1orig mut”  construct, which  had  showed  promising  results  in  the  infected  C6/36  cells.   While neither  the  “prMsil mut”  and  “NS5sil mut”  constructs  had  an  effect  on  viral  titer  in  the  transfected C6/36 HT cells, the “CCR1 orig mut” construct showed a significantly  lower viral titer at all  timepoints tested (Fig. 2.7B) as well as a small plaque phenotype whenever detectable virus was measured.  These results confirmed the  infection experiments,  indicating that CCR1  is a putative RNA sequence element though it effects are more significant in C6/36 cells.  CCR1 and NS5‐2 have no effect on viral translation or RNA synthesis in BHK cells.  The majority of the known DENV RNA elements  regulate viral  translation  (1, 8, 10) or RNA  synthesis  (2, 3, 9, 16, 18, 19).  Other than the 5'/3' CS controls, CCR1 was the only candidate RNA element that was shown to modulate the viral life cycle, so it was of interest to determine whether its effect on plaque phenotype was due to defects  in viral translation or RNA synthesis.   Since the C‐1, C‐2 and prMCR1 regions are not found on the currently available Renilla luciferase reporter construct (9), only CCR1, the 5'/3' CS controls and the NS5‐2 mutant construct that had no aa changes in the protein‐coding sequence were subcloned into the replicons.  Neither the "NS5‐2sil mut" not the "NS5sil mut" construct were shown to modulate viral titer in BHK or C6/36 cells, so only one ("NS5‐2sil mut") was examined for its effects on viral translation and RNA synthesis, serving as an  internal control.   Another negative control  that was utilized  in  this assay was the pDRrep‐RdRpmut construct (“GVD”), which is unable to replicate due to a mutation in the active site of  the viral RdRp  (9).   As with other DENV and WNV replicons,  the replication kinetics are slower compared to the infectious clone (1, 25, 34, 44).   RNA transcripts were transfected into BHK cell monolayers and assayed for luciferase activity at 4, 8, 12, 24, 48 and 72 hpt.   All constructs had WT  levels of  luciferase counts at early  timpoints post‐transfection, indicating that input viral translation is not affected by the introduced mutations (Fig. 2.8).  Aside from the GVD negative control, the only mutant construct that had a significantly lower luciferase count at later timepoints was the “5’CSmut” construct, indicating that this construct has a defect in RNA synthesis.  This is not unexpected given that the "5'CSmut" mutant construct demonstrated a no‐plaque phenotype in IC‐transfected BHK cells (Fig. 2.8) and it is know that disruption of the 5'/3' CS interaction can inhibit RNA synthesis (1, 27).  The "3'CSmut," “CSmut comp” and “NS5‐2sil mut” constructs behaved as expected, displaying WT levels of luciferase counts throughout the assay, indicating that they have no significant  effect  on  viral  translation  or  RNA  synthesis  (Fig.  2.8).    Interestingly,  the  “CCR1orig mut” construct also had no significant difference in luciferase counts as compared to the WT at all timepoints tested,  implying  that CCR1 has no effect on viral  translation and RNA  synthesis  in BHK cells  (Fig. 2.8) despite the effect on plaque morphology in IC‐transfected BHK cells.    

34  

DISCUSSION  Sequence and structure conservation was used as a platform  for  identifying putative RNA elements  in the DENV coding‐region, since RNA elements are more likely to perform a function in the viral life cycle if  they are conserved.   All  three  candidate RNA  structure elements  (C‐1, C‐2 and NS5‐2) and  the  two candidate RNA sequence elements (prMCR1 and NS5CR1) were shown not to play a role in the viral life cycle  in either BHK or C6/36 cells.   However, this does not exclude the possibility  that additional RNA elements  exist within  the DENV  coding‐region.   Continuation of  the  current  study  should  include  re‐examining  the  sequence  alignments  and  phylogenetically  conserved  structure  predictions  and identifying additional candidate RNA elements that could be tested using the tools outlined in this study (discussed in Chapter 4), though it would be best if all mutant constructs were designed with silent point mutations  since altering  the protein‐coding  sequence  can have drastic effects on  viral  replication.    It should be noted that while the C‐1, NS5‐2, prMCR1 and NS5CR1 protein‐coding sequences were shown to  be  important  for  virus  production,  the  C‐2  aa  sequence  is  dispensable.    This  is  important when considering the effect that sequence changes may have on protein folding, implying that this portion of the  capsid‐coding  sequence  can  tolerate  aa  changes  without  affecting  the  role  of  capsid  in  viral replication.   The  “C1sil mut” and  “C1sil  cmut”  constructs had distinct  structural  changes  to C‐1, however, they  varied  slightly  in  their  ability  to  replicate  in  C6/36  cells.    Differences  in  viral  replication  could indicate  that  C‐1  has  a  sequence‐dependent  rather  than  structure‐dependent  effect  on  the  viral  life cycle in C6/36 cells.  However, given the lack of sequence homology in this region of the genome in both the DENV and JEV serogroups (discussed in Chapter 4), it is unlikely that C‐1 would be a candidate RNA sequence element.    It would be of  interest  to  test  this  theory by  introducing more  specific  sequence changes  in  C‐1,  focusing  on  altering  the  primary  nucleotide  sequence  rather  than  the  predicted secondary structure and testing its effects on the viral life cycle in C6/36 cells.   While it has been shown that C‐1, C‐2, NS5‐2, prMCR1 and NS5CR1 are not important cis‐acting regulatory RNA elements, aa changes introduced into C‐1, NS5‐2, prMCR1 and NS5CR1 did impact viral titer, which provides some insight into the requirement for their protein coding‐sequence in the viral life cycle, though these observations are outside the scope of this study.  However, initial testing of the 5’/3’ CS  control mutant  constructs  identified  a  putative  RNA  sequence  element,  termed  CCR1.    CCR1  is further analyzed  in Chapter 3; however, preliminary characterization of  the “CCR1orig mut” construct indicates that CCR1 may act as a cell‐type specific RNA element since it has a more dramatic phenotype in  C6/36  cells.    Considering  that  DENV  is  an  arbovirus  with  distinct  replication  strategies  in  the mammalian  host  and mosquito  vector,  cell‐type  dependent  cis‐acting  determinants  can  help  explain how viral replication is differentially regulated.  However, while two nucleotide changes were enough to significantly  reduce  viral  titer  in  C6/36  cells,  the  “CCR1orig mut”  construct  only  influenced  plaque morphology in BHK cells.  It is possible that CCR1 is important for viral replication in BHK cells, but that additional  sequence changes need  to be  introduced  to observe a more dramatic phenotype.   Further investigation should explore which aspects of the primary nucleotide sequence for CCR1 are important for  the viral  life cycle  in C6/36 cells, as well as  further characterize  the effect of CCR1 on mammalian cells.   It  is known  from other positive‐strand RNA viruses  that multiple coding‐region RNA elements can have  important  regulatory  roles  in modulating various aspects of  the viral  life cycle  (33).   Studies such as  these can help  identify additional RNA elements  located within  the coding‐region, which may provide insight into novel replication strategies, help elucidate the regulation of other stages in the viral life cycle besides viral translation and RNA synthesis, and potentially explain the differential regulation of  the  viral  life  cycle  in  the mammalian host  and mosquito  vector.    Thus  far,  few  efforts have been undertaken to elucidate the roles of coding‐region RNA elements within the DENV genome, which limits 

35  

our understanding of DENV replication.  However, the potential role of CCR1 in modulating the viral life cycle  in mosquito  cells might provide  insight  into  the  viral  replication  strategies within  the mosquito vector. 

FIGURES

DEN1 ACGGGTCGACCGTCTTTCAATATGCTGAAACGCGCGAGAAACCGCGTGTCAACTGTTTCADEN3 ACGGGAAAACCGTCTATCAATATGCTGAAACGCGTGAGAAACCGTGTGTCAACTGGATCADEN2 GCGAAAAACACGCCTTTCAATATGCTGAAACGCGAGAGAAACCGCGTGTCGACTGTGCAADEN4 GTGGTTAGACCACCTTTCAATATGCTGAAACGCGAGAGAAACCGCGTATCAACCCCTCAA

* * ** ****************** ********* ** ** ** *

A

B

DEN1 CCCCGGATCACTGAGACGGAACCAGATGACGTTGACTGTTGGTGCAATGCCACGGAGACAC

DEN1 GGTGGTAAGGACTAGAGGTTAGAGGAGACCCCCCGCACAACAACAAACAGCATATTGACGDEN3 CTTGCAAAGGACTAGAGGTTAGAGGAGACCCCCCGCA-AATAA-AAACAGCATATTGACGDEN2 CGCTGGAAGGACTAGAGGTTAGAGGAGACCCCCCCGA-AACAAAAAACAGCATATTGACGDEN4 TGGTGGAAGGACTAGAGGTTAGAGGAGACCCCCCCAA-CACAA-AAACAGCATATTGACG

**************************** * * ** ****************

DEN3 CCCCACATTACCGAAGTGGAACCTGAAGACATTGACTGCTGGTGCAACCTTACATCAACADEN2 CCCCTTCTCAGGCAGAATGAGCCAGAAGACATAGACTGTTGGTGCAACTCTACGTCCACGDEN4 CCCCTACTGGTCAATACCGAACCTGAAGACATTGATTGCTGGTGCAACCTCACGTCTACC

**** * * ** ** ** *** * ** ** ******** ** **

DEN1 CAAACTTGGATACATACTCAGAGACATATCAAAGATTCCAGGGGGAAATATGTATGCAGADEN3 CAAGCTAGGCTACATATTAAGGGACATTTCCAAGATACCCGGAGGAGCCATGTATGCTGA

2

D

DEN2 CAAGCTAGGTTACATTCTAAGAGACGTGAGCAAGAAAGAGGGAGGAGCAATGTATGCCGADEN4 CAGATTGGGATATATCCTGGAGGAGATAGACAAGAAGGATGGAGACCTAATGTATGCTGA

** * ** ** ** * ** * **** ** * ******** **

DEN1 TGACACAGCCGGATGGGACACAAGAATAACAGAGGATGATCTTCAGAATGAGGCCAAAATDEN3 TGACACAGCTGGTTGGGACACAAGAATAACAGAAGATGACCTGCACAATGAGGAAAAGATDEN2 TGACACCGCAGGATGGGATACAAGAATCACACTAGAAGACCTAAAAAATGAAGAAATGGTDEN4 TGACACAGCAGGCTGGGACACAAGAATCACTGAGGATGACCTTCAAAATGAGGAACTGATDEN4 TGACACAGCAGGCTGGGACACAAGAATCACTGAGGATGACCTTCAAAATGAGGAACTGAT

****** ** ** ***** ******** ** ** ** ** * ***** * *

JEV C------CAGGAGGGCCC--GGTAAAAACCGGGCTATCAATATGCTGAAACGCGGCCTACMVEV C------CAGGAGGACCC--GGGAAGCCCCGGGTCGTCAATATGCTAAAACGCGGCATACWNV C------CAGGAGGGCCC--GGTAAAAACCGGGCTGTCAATATGCTAAAACGCGGTATGCKUN C------CAGGAGGGCCC--GGCAAAAGCCGGGCTGTCAATATGCTAAAACGCGGAATGCSLEV C------CAGGAAAACCC--GGTAGAAACCGGGTTGTCAATATGCTAAAACGCGGCGTAT

E

SLEV C CAGGAAAACCC GGTAGAAACCGGGTTGTCAATATGCTAAAACGCGGCGTATYFV TGTCTGGTCGTAAAGCTCAGGGAAAAACCCTGGGCGTCAATATGGTACGACGAGGAGTTC

* * * * ** * ** ** ******** * *** ** *

36

JEV GGAGACCCCGCATTT GCATCA AACA GCATATTGACACCTGGGAATAGACTF

G

JEV GGAGACCCCGCATTT----GCATCA-AACA-----GCATATTGACACCTGGGAATAGACTMVEV GGAGACCCCACTCTCAAAAGCATCA-AACAA--CAGCATATTGACACCTGGGAAAAGACTWNV GGAGACCCCGTGCCA-AAAACACCAAAAGAA-ACAGCATATTGACACCTGGGA-TAGACTKUN GGAGACCCCGTGCCGCAAAACACCACAACAACACAGCATATTGACACCTGGGA-TAGACTSLEV GGAGACCCCTTGCCG-TTAACGCAA--ACAA--CAGCATATTGACACCTGGAA—-AGAC-YFV GGAGACCCTCCAGGG-AACAAATAGTGGGAC-----CATATTGACGCCAGGGAAAGACCG

******** * ********* ** ** * *

JEV AAGCTCACCATGGGCAATGATCCAGAGGATGTGGATTGCTGGTG---TGACAACCAAGAAMVEV AAATTGGAAAGTGGAAATGACCCTGAAGACATTGACTGCTGGTG---TGACAAACAAGCTWNV GTCCTAGCTGCTGGAAATGACCCTGAAGACATTGACTGCTGGTG---CACGAAATCATCTKUN GTGTTGTCGGCCGGAAATGATCCAGAAGACATTGACTGCTGGTG---CACGAAGTTAGCASLEV GTGCTTTCAGCGGGAAATGATCCCGAGGACATTGACTGTTGGTG---TGACGTCGAAGAGYFV AATCTCAGTCCAAGAGAGGAGCCAGATGACATTGATTGCTGGTGCTATGGGGTGGAAAAC

* * * ** ** ** ** * ** ** ***** *

G

JEV TCCAAAAGCTGGGATACATCCTCCGTGACATAGCAGGAAAGCAAGGAGGGAAAATGTACGMVEV TTCAGAAGCTGGGATACATCTTGAGAGATGTGGCTCAAAAGCCTGGAGGGAAAATTTATGWNV TGCAGAAGCTCGGGTACATCTTGAAGGAAGTTGGAACAAAGCCTGGAGGAAAGGTTTACGKUN TCCAGAAATTAGGCTACATCCTGCGTGAAGTTGGCACCCGACCCGGAGGCAGAATCTACGSLEV TCCAGAAACTTGGATACATCCTGCAAGAAATCTCCCAAATTCCAGGAGGAAAGATGTACGYFV TACAATACCTAGGATATGTGATCAGAGACCTGGCTGCAATGGATGGTGGTGGATTCTACG

H

YFV TACAATACCTAGGATATGTGATCAGAGACCTGGCTGCAATGGATGGTGGTGGATTCTACG* ** * * ** ** * * ** * ** ** * ** *

JEV CTGATGACACCGCCGGATGGGACACTAGAATTACCAGAACTGATTTAGAAAATGAAGCCAMVEV CCGATGACACAGCTGGTTGGGACACCCGCATCACACAAGCTGACCTTGAGAACGAAGCTAWNV CTGATGATACCGCAGGCTGGGACACACGCATCACCAAAGCTGACCTCGAGAATGAAGCGAKUN CTGATGACACAGCCGGTTGGGACACCCGCATCACAAGAGCTGACCTGGAGAATGAAGCCASLEV CAGATGACACAGCCGGCTGGGACACCCGGATCACAAAAGAAGACTTGAAAAATGAAGCAAYFV CGGATGACACCGCTGGATGGGACACGCGCATCACAGAGGCAGACCTTGATGATGAACAGG

* ***** ** ** ** ******** * ** ** ** * * * ***

Figure 2.1: Putative RNA sequence elements identified in the capsid‐, prM/M‐ and NS5‐coding region basedon sequence conservation in the DENV and JEV serogroups. Partial alignment of the representative genomesfor the DENV serogroup (accession codes DENV1 U88535, DENV2 NC_001474, DENV3 AY099336 and DENV4GU289913) for the (A) 5’ CS (position 118‐177, numbering based on DENV2 Thai strain 16681), (B) 3’ CS) ( ) (p , g ), ( )(position 10572‐10630), (C) prMCR1 (position 598‐657) and (D) NS5CR1 (position 9108‐9227). Partial alignmentof the representative genomes for the JEV serogroup (accession codes JEV M18370, Murray Valley encephalitisvirus AF161266, WNV NC_001563, Kunjin virus AY274505, St. Louis encephalitis virus DQ525916 and YFVU17067) for the (E) 5’ CS (position 108‐159, numbering based on JEV), (F) 3’ CS (position 10840‐10889), (G)prMCR1 (position 639‐695) and (H) NS5CR1 (position 9217‐9336). Alignments were generated using Clustal W2(42). Positions of sequence homology are indicated by asterisks; the 5’ CS is underlined with a solid line; the 3’CS is underlined with a dashed line and the candidate RNA sequence elements prMCR1 and NS5CR1 are boxed.

37

A

C‐1SLA

5’3’

SLA B

5’ 3’cHP

C 2

CCR1

cHP

C‐2

C‐1

C‐2

C

’3’

5’

C‐1

SLA

cHP

C‐2

CCR1

38

D

5’3’

NS5‐2

3’ SL DB1

DB2

VR

Figure 2.2: Putative RNA structural elements identified in the capsid‐ and NS5‐coding region based onstructure conservation in the DENV, JEV and TBEV serogroups. Phylogenetic consensus structure based on (A)the 5’ UTR and the entire capsid‐coding sequence of DENV1‐4, (B) the 5’ UTR and portions of the capsid‐codingsequence from the JEV serogroup (position 1‐273, numbering based on JEV) or (C) the 5’ UTR and portions ofthe capsid‐coding sequence from the TBEV serogroup (position 1‐302, numbering based on TBEV; accessiond d k h h ) ( ) h lcodes TBEV U27495, Powassan virus NC_003687 and Omsk hemorrhagic virus AY193805). (D) Phylogenetic

consensus structure based on the 3’ UTR and portions of the NS5‐coding sequence from the DENV serogroup(position 9647‐10159, numbering based on DENV2 Thai strain 16681). Consensus structures were computedwith RNAalifold (22). Covariance is depicted as a colored spectrum ranging from red, to dark pink, lightpink, yellow and then green. Major known RNA elements are indicated as follows: stem‐loop A (SLA), capsid‐coding hairpin (cHP), conserved capsid‐coding region 1 (CCR1), variable region (VR), dumbbells 1 and 2 (DB1/2)and 3’ stem‐loop (3’ SL). Putative RNA structure elements C‐1 (red), C‐2 (black) and NS5‐2 (blue) are indicatedby colored circles.

39

A WTΔG = ‐24.6

C1sil mutΔG = ‐20.4

C1aa mutΔG = ‐13.9ΔG    24.6 ΔG    20.4ΔG    13.9

C1aa cmut C1sil cmutC1aa cmutΔG = ‐16.1

C1sil cmutΔG = ‐15.1

B WTΔG = ‐13.3

C2aa mutΔG = ‐6.5

C‐2CCR1 CCR1

C‐2

40

C WTΔG = 20

NS5‐2aa mutΔG = 18 11ΔG = ‐20 ΔG = ‐18.11

NS5‐2sil mutΔG = ‐23

NS5‐2aa cmutΔG = ‐18.5

Figure 2.3: Generation of mutant constructs for the putative RNA structural elements. mfold predictions (50)of the mutant constructs designed to disrupt the predicted secondary structures of putative RNA structuralelements (A) C‐1, (B) C‐2 and (C) NS5‐2. The stem‐loop structure of C‐1 was first mutated with conservedamino acid changes (“C1aa mut”) and then restored with a complementary mutant that had a WT structure butmutant amino acid sequence (“C1aa cmut”). Two additional mutants were designed with silent pointq ( ) g pmutations that disrupted the predicted structure of C‐1 (“C1sil cmut” and “C1sil mut”). The predicted stem‐loop of C‐2 was disrupted with conserved amino acid changes (“C2aa mut”), as was the stem loop of NS5‐2(“NS5‐2aa mut”). The amino acid changes in “NS5‐2aa mut” were maintained in a complementary mutant thatrestored the WT structure of NS5‐2 (“NS5‐2aa cmut”). The predicted secondary structure of NS5‐2 was alsomutated with silent point mutations (“NS5‐2sil mut”). The positions of introduced mutations are indicated byboxes. Major known RNA elements are indicated as follows: CCR1, conserved capsid‐coding region 1.

41

6 5

A B

3

4

5

6

g(pfu/mL)

3' UTR Primer/Probe NS5 Primer/Probe

4.5

5

/mL)

Raw Titer 2 hpt vRNA in Cells

0

1

2

24 48 72

Log

H P t T f ti

3.5

4

24 48 72

Log(pfu/

H P t T f tiHours Post‐Transfection Hours Post‐Transfection

Figure 2.4: Optimization of viral infectivity analysis. Virus‐containing supernatants from IC‐transfected BHKcells were assessed for viral titer (pfu/mL) by plaque assay. (A) The normalization protocol was optimized byharvesting vRNA from IC‐transfected cells at 2 hpt and analyzed using qRT‐PCR with primer/probe sets directedagainst either the 3’ UTR or the viral protein NS5. (B) The normalization protocol was modified to examinewhether viral titers are better reflected by the transfection efficiency at 2 hpt ("2 hpt") or the amount of vRNAin the cells at the time that the virus containing supernatants were harvested ("vRNA in Cells") as they comparein the cells at the time that the virus‐containing supernatants were harvested ( vRNA in Cells ) as they compareto the raw viral titers ("Raw Titer"). Results were obtained from 3 independent experiments, conducted induplicate. The limit of detection is shown by a dashed line (1.9 pfu/mL). *p<0.05 relative to the WT. Errorsbars indicate standard deviation.

42

A B6

3

4

5

6

(pfu/m

L)

3

4

5

6

7

(pfu/m

L)

WT C1aa mutC1aa cmut C1sil cmutC1sil mut

0

1

2

WT C2aa mut

Log(

0

1

2

3

24 48 72

Log

H P t T f ti

* * **

C

6

7 WT NS5‐2aa mut

NS5‐2aa cmut NS5‐2sil mut

Hours Post‐Transfection

6

7WT prMaa mut prMsil mut

D

2

3

4

5

Log(pfu/mL)

* **2

3

4

5

Log(pfu/mL)

* **

0

1

24 48 72Hours Post‐Transfection

0

1

24 48 72Hours Post‐Transdection

43

E F

3

4

5

6

7

(pfu/m

L)

WT NS5aa mut NS5sil mut

3

4

5

6

7

g(pfu/mL)

WT 5'CSmut3'CSmut CSmut compCCR1orig mut

SP*SP*

0

1

2

3

24 48 72

Log(

Hours Post Transfection

* **

0

1

2

3

24 48 72

Log

Hours Post Transfection

* **

Hours Post Transfection Hours Post Transfection

G

WT SmallPlaque (SP)

Figure 2.5: Examination of the putative RNA elements in BHK cells. Virus‐containing supernatants from IC‐/transfected BHK cell monolayers were assessed for viral titer (pfu/mL) by plaque assay. The role of (A) C‐1, (B)

C‐2, (C) NS5‐2, (D) prMCR1, (E) NS5CR1, and (F) the control mutant constructs 5’/3’ CS were examined bycomparing the WT and mutant IC constructs. (G) Photographic representation of the small plaque phenotypeobserved with the “CSmut comp” and “CCR1orig mut” mutant constructs. Viral titers were normalized to theamount of vRNA present in the cells at the time of harvest as measured by qRT‐PCR using the primer/probe setdirected against the viral NS5 gene. Results were obtained from 3 independent experiments, conducted induplicate. Plaque phenotypes are indicated (SP, small plaque). The limit of detection is shown by a dashed line(1.9 pfu/mL). *p<0.05 relative to the WT. Errors bars indicate standard deviation. Note: prMCR1 and NS5CR1transfections in BHK cells were conducted by Dipti Banerjee UC Berkeley Senior Thesis 2011transfections in BHK cells were conducted by Dipti Banerjee, UC Berkeley Senior Thesis 2011.

44

7WT CCR1orig mut C1sil cmut C1sil mut NS5‐2sil mut

2

3

4

5

6

Log(pfu/mL) *SP

SP*

0

1

2

BHKs ‐ 48 hrs C6/36 ‐ 120 hrs

Figure 2.6: Infection of BHK and C6/36 cells with mutant viruses demonstrates a role for CCR1 in C6/36 cells.Mutant viruses were harvested from IC‐transfected BHK monolayers, concentrated, titered and used to infectBHK and C6/36 cell monolayers at an MOI of 0.01, and the viral titer was assessed by plaque assay at 48 (BHKcells) and 120 (C6/36 cells) hours post‐infection. Results were obtained from 3 independentexperiments, conducted in duplicate. Plaque phenotypes are indicated (SP, small plaque). The limit ofdetection is shown by a dashed line (1.9 pfu/mL). *p<0.05 relative to the WT. Errors bars indicate standarddeviation.

45

A B

5

6

7

8

9

10

pfu/mL)

WT prMsil mut

NS5sil mut CCR1orig mut

5

6

7

8

9

10

(pfu/m

L)

C6/36 C6/36 HTs

0

1

2

3

4

72 96 120 144

Log(p

Hours Post‐Transfection

0

1

2

3

4

72 96 120 144

Log

Hours Post‐Transfection

*

SP*

**

Hours Post‐TransfectionHours Post‐Transfection

Figure 2.7: Examination of the putative RNA elements in C6/36 cells. (A) C6/36 and C6/36 HT cells wereanalyzed for their ability to be transfected with ICs. (B) Virus‐containing supernatants from IC‐transfectedC6/36 HT cell monolayers were assessed for viral titer (pfu/mL) by plaque assay in order to determine the roleof CCR1, prMCR1 and NS5CR1 in the viral life cycle. Viral titers were normalized to the amount of vRNA presentin the cells at the time of harvest as measured by qRT‐PCR using the primer/probe set directed against the viralprotein NS5 Results were obtained from 3 independent experiments conducted in duplicate Plaqueprotein NS5. Results were obtained from 3 independent experiments, conducted in duplicate. Plaquephenotypes are indicated (SP, small plaque). The limit of detection is shown by a dashed line (1.9 pfu/mL).*p<0.05 relative to the WT. Errors bars indicate standard deviation. Note: prMCR1 and NS5CR1 transfectionsin C6/36 HT cells were conducted by Dipti Banerjee, UC Berkeley Senior Thesis 2011.

46

12

WT GVD 5'Csmut 3'Csmut Csmut comp CCR1orig mut NS5‐2sil mut

4

6

8

10

Log(rLUC)

* **

*

0

2

0 10 20 30 40 50 60 70 80Hours Post‐Transfection

**

Figure 2.8: CCR1 shows no effect on viral translation or RNA synthesis in BHK cells. Mutations weresubcloned into Renilla luciferase replicons and the mutant replicons were transfected into BHK cell monolayers.Cellular lysates were harvested at the indicated timepoints and assayed for luciferase activity. Results wereobtained from 3 independent experiments, conducted in duplicate. *p<0.05 relative to the WT replicon. Errorsbars indicate standard deviation.

47

48  

TABLES   Table 2.1:  SOE‐PCR cloning primers for the generation of DNA mutant constructs.  For cloning into the IC,  the mutant  forward  (FWD)  primers  (listed) were  used  in  conjunction with  the WT  reverse  (REV) primer  to  generate  the  “A”  fragment.    The mutant REV  primers  are  the  reverse  complement  of  the mutant FWD primer, and used  in conjunction with the WT FWD primer to generate the “B” fragment.  The “A” and “B” fragments were spliced together using the WT FWD and REV primers (“C” fragment), which was used  to generate  the  inserts.   Certain mutant constructs  required  two  rounds of SOE‐PCR; primers used in the second round of SOE‐PCR are indicated by asterisks.  Codons where mutations were introduced are in lower case.  WT flanking primers are underlined. Backbone  Construct  Primer Sequence 

IC 

F1 (FWD)  5’-AACGACAGGAGCACGATCATGC-3’

C1aa mut (FWD)  5’-GATTCTCACTTGGAATGCTGaacGGACGAGGACCAatcAAACTGTTCATGGCCCTGG-3’

C1aa mut (FWD)* 5’-GACCAatcAAACTGTTCATGggcatcggcGCGTTCCTTCGTTTCCTAAC-3’

C1aa cmut (FWD)  5’-GATTCTCACTTGGAATGCTGaatGGACGAGGACCAataAAACTGTTCATGGCCCTGG-3’

C1aa cmut (FWD)* 5’-GACCAataAAACTGTTCATGggcatcggaGCGTTCCTTCGTTTCCTAAC-3’

C1sil cmut (FWD)  5’-GATTCTCACTTGGAATGCTGcagGGACGAGGACCActcAAACTGTTCATGGCCCTGG-3’

C1sil cmut (FWD)* 5’-GACCAcucAAACTGTTCATGgcgcuggucGCGTTCCTTCGTTTCCTAAC-3’

C1sil mut (FWD)  5’-CTTGGAATGCTGCAGGGAagaggcCCActgaagCTGTTCATGGCCCTGGTG-3’

C1sil mut (FWD)* 5’-GAagaggcCCActgaagCTGtttATGgctttggtagctTTCCTTCGTTTCCTAACAATC-3’

C2aa mut (FWD)  5’-CGCGAGAGAAACCGCGTGacctccGTGCAAaacatctccAAGAGATTCTCACTTGG-3’

5’CSmut (FWD)  5’-GCGAAAAACACGCCTTTCaacATGCTGaagcggGAGAGAAACCGCGTGTCG-3’

CCR1orig mut (FWD)  5’-GCCTTTCAATATGCTGaagcggGAGAGAAACCGCGTGTCG-3’

prMaa mut (FWD)  5’-CTTCTCAGGCAGAATGAGcccGAAgaaATAgaaTGTTGGTGCAACTCTACG-3’

prMsil mut (FWD)  5’-CCCCTTCTCAGGCAGAATgaacccgaggatatcgatTGTTGGTGCAACTCTACG-3’

prMsil mut (FWD)* 5’-gaacccgaggatatcgattgcTGGtgtaatTCTACGTCCACGTGGG-3’

ESphIR (REV)  5’-TTTCCGACTGCATGCTCTTC-3’

BsrGI (FWD)  5’-GAGCTGGTTGACAAGGAAAG-3’

NS5‐2aa mut (FWD)  5’-GAGACGGCCTGTTTGGGGcggaccTACGCCaacATGTGGAGCTTGATGTACTTC-3’

NS5‐2aa mut (FWD)*  5’-cTACGCCaacATGTGGAGCatcATGTACTTCCACaagaagGACCTCAGGCTGGCGGC-3’

NS5‐2aa cmut (FWD)  5’-GAGACGGCCTGTTTGGGGaggacuTACGCCaacATGTGGAGCTTGATGTACTTC-3’

NS5‐2aa cmut (FWD)*  5’-ctTACGCCaacATGTGGAGCataATGTACTTCCACaagaagGACCTCAGGCTGGCGGC-3’

NS5‐2sil mut (FWD)  5’-CTTTGCGGGAGACGGCCtgcctgggcAAGTCTTACGCCcagATGTGGAGCTTGATGTAC-3’

NS5‐2sil mut (FWD)*  5’-GCTTGATGTACTTCCACAGAcggGACCTCAGGCTGGCGGC-3’

NS5aa mut (FWD)  5’-GAAAGAGGGAGGAGCAATGtacggcgacgatACCggcGGATGGGATACAAGAATC-3’

NS5aa mut (FWD)*  5’-cggcgacgatACCggcGGATGGGATtccaggATCtccCTAGAAGACCTAAAAAATG-3’

NS5sil mut (FWD)  5’-GAAAGAGGGAGGAGCAATGtacgctgacgatactgccGGATGGGATACAAGAATC-3’

NS5sil mut (FWD)*  5’-GtacgctgacgatactgccggcTGGgacactcgcattaccCTAGAAGACCTAAAAAATG-3’

3’CSmut (FWD)  5’-CGAAACAAAAAACAGCATgTTGACGCTGGGAAAGACC-3’

XbaIR (REV)  5’-CGACTCTAGAGAACCTG-3’

Replicon SacIT7RepF (FWD)  5’-CGAGCTCGCGGCCGCAAATTTAATACGACTCAC-3’

MluICrepR (REV)  5’-CATACGCGTGGACACGCGGTTTCTCTCGCGTTTC-3’

CCR1orig mutR (REV)  5’-CATACGCGTGGACACGCGGTTTCTCTCccgcttC-3’

     

49  

Table 2.2:  qRT‐PCR primer/probe sets directed against the 3’ UTR or the viral NS5 gene. Directed Against…  Construct  Primer or Probe Sequence 

3’ UTR DV10621F (FWD)  5’-CATATTGACGCTGGGAAAGA-3’ DV10723R (REV)  5’-AGAACCTGTTGATTCAACAGCACC-3’

DV‐3UTR‐F/T (Probe)  5’-FAM-CTGTCTCCTCAGCATCATTCCAGGCA-TAMRA-3’

NS5 DV9944F (FWD)  5’-ACAAGTCGAACAACCTGGTCCAT-3’ DV10121R (REV)  5’-GCCGCACCATTGGTCTTCTC-3’

DV‐NS5‐F/T(Probe)  5’-FAM-TGGGATTTCCTCCCATGATTCCACTGG-TAMRA-3’

  Table 2.3:   Sequences for the WT and mutant C‐1 constructs.   Mutations are  in  lower case and amino acid changes are underlined. 

Construct  Nucleotide Sequence  Amino Acid Sequence 

WT 5’–CTTGGAATGCTGCAGGGACGAGGACCATTA AAACTGTTCATGGCCCTGGTGGCGTTCCTT–3’ 5’-LGMLQGRGPLKLFMALVAFL-3’

C1aa mut 5’-CTTGGAATGCTGaAcGGACGAGGACCAaTc AAACTGTTCATGGgCaTcGgcGCGTTCCTT-3’ 5’-LGMLNGRGPIKLFMGIGAFL-3’

C1aa cmut 5’-CTTGGAATGCTGaAtGGACGAGGACCAaTA AAACTGTTCATGGgCaTcGgaGCGTTCCTT-3’ 5’-LGMLNGRGPIKLFMGIGAFL-3’

C1sil cmut 5’-CTTGGAATGCTGCAGGGACGAGGACCAcTc AAACTGTTCATGGCgCTGGTcGCGTTCCTT-3’ 5’-LGMLQGRGPLKLFMALVAFL-3’

C1sil mut 5’-CTTGGAATGCTGCAGGGAaGAGGcCCAcTg AAgCTGTTtATGGCttTGGTaGCtTTCCTT-3’ 5’-LGMLQGRGPLKLFMALVAFL-3’

  Table 2.4:   Sequences for the WT and mutant C‐2 constructs.   Mutations are  in  lower case and amino acid changes are underlined. 

Construct  Nucleotide Sequence  Amino Acid Sequence WT  5’-GTGTCGACTGTGCAACAGCTGACAAAG-3’ 5’-VSTVQQLTK-3’

C2aa mut  5’-GTGaCctCcGTGCAAaAcaTctCcAAG-3’ 5’-VTSVQNISK-3’

  Table 2.5:  Sequences for the WT and mutant NS5‐2 constructs.  Mutations are in lower case and amino acid changes are underlined. 

Construct  Nucleotide Sequence  Amino Acid Sequence 

WT 5’-ACGGCCTGTTTGGGGAAGTCTTACGCCCAAATG TGGAGCTTGATGTACTTCCACAGACGCGACCTC-3’ 5’-TACLGKSYAQMWSLMYFHRRDL-3’

NS5‐2aa mut 5’-ACGGCCTGTTTGGGGcgGaCcTACGCCaAcATG TGGAGCaTcATGTACTTCCACAagaagGACCTC-3’ 5’-TACLGRTYANMWSIMYFHKKDL-3’

NS5‐2aa cmut 5’-ACGGCCTGTTTGGGGAgGaCTTACGCCaAcATG TGGAGCaTaATGTACTTCCACAagaagGACCTC-3’ 5’-TACLGRTYANMWSIMYFHKKDL-3’

NS5‐2sil mut 5’-ACGGCCTGccTGGGcAAGTCTTACGCCCAgATG TGGAGCTTGATGTACTTCCACAGACGgGACCTC-3’ 5’-TACLGKSYAQMWSLMYFHRRDL-3’

  Table 2.6:   Sequences  for  the WT and mutant prMCR1 constructs.   Mutations are  in  lower case and amino acid changes are underlined. 

Construct  Nucleotide Sequence  Amino Acid Sequence WT  5’-GAGCCAGAAGACATAGACTGTTGGTGCAAC-3’ 5’-EPEDIDCWCN-3’

prMaa mut  5’-GAGCCcGAAGAaATAGAaTGTTGGTGCAAC-3’ 5’-EPEEIECWCN-3’ prMsil mut  5’-GAaCCcGAgGAtATcGAtTGcTGGTGtAAt-3’ 5’-EPEDIDCWCN-3’

50  

Table 2.7:   Sequences  for  the WT and mutant NS5CR1 constructs.   Mutations are  in  lower case and amino acid changes are underlined. 

Construct  Nucleotide Sequence  Amino Acid Sequence 

WT 5’-ATGTATGCCGATGACACCGCA GGATGGGATACAAGAATCACA-3’ 5’-MYADDTAGWDTRIT-3’

NS5aa mut 5’-ATGTAcGgCGAcGAtACCGgc GGATGGGATtCcAGgATCtCc-3’ 5’-MYGDDTGGWDSRIS-3’

NS5sil mut 5’-ATGTAcGCtGAcGAtACtGCc GGcTGGGAcACtcGcATtACc-3’ 5’-MYADDTAGWDTRIT-3’

  Table  2.8:    Sequences  for  the WT  and mutant  5’/3’  CS  controls,  and  the  putative  RNA  sequence element CCR1.  The 5’ CS is highlighted.  The "CSmut comp" construct contains the mutations that were introduced  in  the  "5'CSmut" and  "3'CSmut"  constructs.   Mutations are  in  lower  case and amino acid changes are underlined. 

  

 

Construct  Nucleotide Sequence  Amino Acid Sequence WT 5’ CS  5’-TTCAATATGCTGAAACGCGAG-3’ 5’-FNMLKRE-3’ 5’CSmut  5’-TTCAAcATGCTGAAgCGgGAG-3’ 5’-FNMLKRE-3’

WT 3’ CS  5’-AGCATATTGA-3’ 3’CSmut  5’-AGCATgTTGA-3’

CSmut comp 5’-TTCAAcATGCTGAAgCGgGAG-3’ 5’-FNMLKRE-3’

5’-AGCATgTTGA-3’

CCR1orig mut  5’-TTCAATATGCTGAAgCGgGAG-3’ 5’-FNMLKRE-3’

51  

REFERENCES   1.  Alvarez, D. E., A. L. De Lella Ezcurra, S. Fucito, and A. V. Gamarnik. 2005. Role of RNA structures 

present  at  the  3'UTR  of  dengue  virus  on  translation,  RNA  synthesis,  and  viral  replication. Virology 339:200‐212  

2.  Alvarez, D. E., C. V. Filomatori, and A. V. Gamarnik. 2008. Functional analysis of dengue virus cyclization sequences located at the 5′ and 3′UTRs. Virology 375:223‐235. 

3.  Alvarez, D. E., M. F. Lodeiro, S. J. Ludueña, L. I. Pietrasanta, and A. G. Gamarnik. 2005. Long‐range RNA‐RNA interactions circularize the dengue virus genome. J Virol 79:6631‐6643. 

4.  Arias, C. F., F. Preugschat, and J. H. Strauss. 1993. Dengue 2 virus NS2B and NS3 form a stable complex that can cleave NS3 within the helicase domain. Virology 193:888‐899. 

5.  Burke, D.  S.,  and  T.  P. Monath.  2001.  Flaviviruses,  p.  1043‐1126.  In D. M.  Knipe  and  P. M. Howley (ed.), Fields Virology, 4th ed, vol. 1. Lippincott Williams and Wilkins, Philadelphia. 

6.  Cahour,  A.,  A.  Pletnev, M.  Vazeille‐Flacoz,  L.  Rosen,  and  C.‐J.  Lai.  1995.  Growth‐restricted dengue  virus mutants  containing  deletions  in  the  5'  noncoding  region  of  the  RNA  genome. Virology 207:68‐76. 

7.  Chambers, T.  J., C. S. Hahn, R. Galler, and C. M. Rice. 1990.  Flavivirus genome organization, expression, and replication. Ann Rev Microbiol 44:649‐688. 

8.  Chiu, W. W., R. M. Kinney, and T. W. Dreher. 2005. Control of  translation by  the 5'‐ and 3'‐terminal regions of the dengue virus genome. J Virol 79:8303‐8315. 

9.  Clyde, K.,  J. Barrera, and E. Harris. 2008. The capsid‐coding  region hairpin element  (cHP)  is a critical determinant of dengue virus and West Nile virus RNA synthesis. . Virology 379:314‐323. 

10.  Clyde, K., and E. Harris. 2006. RNA  secondary  structure  in  the  coding  region of dengue  virus type  2  directs  translation  start  codon  selection  and  is  required  for  viral  replication.  J  Virol 80:2170‐2182. 

11.  Clyde,  K.,  J.  L.  Kyle,  and  E.  Harris.  2006.  Recent  advances  in  deciphering  viral  and  host determinants of dengue virus replication and pathogenesis. J Virol 80:11418‐11431. 

12.  Diamond, M. S., D. Edgil, T. G. Roberts, B. Lu, and E. Harris. 2000. Infection of human cells by dengue virus is modulated by different cell types and viral strains. J Virol 74:7814‐7823. 

13.  Diviney, S., A. Tuplin, M. Struthers, V. Armstrong, R. M. Elliott, P. Simmonds, and D. J. Evans. 2008. A hepatitis C virus cis‐acting replication element forms a long‐range RNA‐RNA interaction with upstream RNA sequences in NS5B. 82 18. 

14.  Edgil,  D., M.  S.  Diamond,  K.  L.  Holden,  S. M.  Paranjape,  and  E.  Harris.  2003.  Translation efficiency  determines  differences  in  cellular  infection  among  dengue  virus  type  2  strains. Virology 317:275‐290. 

15.  Edgil, D., and E. Harris. 2006. End‐to‐end communication in the modulation of mammalian RNA virus translation. Virus Res 119:43‐51. 

16.  Filomatori, C. V., M. F. Lodeiro, D. E. Alvarez, M. M. Samsa, L. Pietrasanta, and A. V. Gamarnik. 2006. A 5' RNA element promotes dengue virus RNA synthesis on a circular genome. Genes Dev 20:2238‐2249. 

17.  Friebe, P., J. Boudet, J. P. Simorre, and R. Bartenschlager. 2005. Kissing‐loop interaction in the 3' end of the hepatitis C virus genome essential for RNA replication. J Virol 79:380‐392. 

18.  Friebe, P., and E. Harris. 2010.  Interplay of RNA elements  in  the dengue virus 5' and 3' ends required for viral RNA replication. J Virol 84:6103‐6118. 

19.  Friebe, P., P. Y. Shi, and E. Harris. 2011. The 5' and 3' downstream AUG  region elements are required for mosquito‐borne flavivirus RNA replication. J Virol 85:1900‐1905. 

52  

20.  Goodfellow,  I., Y. Chaudhry, A. Richardson,  J. Meredith,  J. W. Almond, W. Barclay, and D.  J. Evans.  2000.  Identification  of  a  cis‐acting  replication  element  within  the  poliovirus  coding region. J Virol 74:4590–4600. 

21.  Gubler,  D.  J.  2002.  The  global  emergence/resurgence  of  arboviral  diseases  as  public  health problems. Arch Med Res 33:330‐342. 

22.  Hofacker,  I. L., M. Fekete, and P. F. Stadler. 2002. Secondary structure prediction  for aligned RNA sequences. J Mol Biol 319:1059‐1066. 

23.  Hofacker, I. L., W. Fontana, P. F. Stadler, S. Bonhoeffer, M. Tacker, and P. Schuster. 1994. Fast folding and comparison of RNA secondary structures. Monatsh Chem 125:167‐188. 

24.  Holden,  K.  L.,  and  E.  Harris.  2004.  Enhancement  of  dengue  virus  translation:  Role  of  the 3’untranslated region and the terminal 3’ stem‐loop domain. Virology 329:119‐133. 

25.  Holden,  K.  L.,  D.  Stein,  T.  C.  Pierson,  A.  Ahmed,  K.  Clyde,  P.  Iverson,  and  E.  Harris.  2006. Inhibition of dengue virus translation and RNA synthesis by a morpholino oligomer to the top of the 3’ stem‐loop structure. Virology 344:439‐452. 

26.  Houng, H. H., D. Hritz, and N. Kanesa‐thasan. 2000. Quantitative detection of dengue 2 virus using fluorogenic RT‐PCR based on 3'‐noncoding sequence. J Virol Methods 86:1‐11. 

27.  Khromykh, A. A., H. Meka, K. J. Guyatt, and E. G. Westway. 2001. Essential role of cyclization domains in flavivirus RNA replication. J Virol 75:6719‐6728. 

28.  Kinney, R. M., S. Butrapet, G. J. Chang, K. R. Tsuchiya, J. T. Roehrig, N. Bhamarapravati, and D. J. Gubler. 1997. Construction of  infectious cDNA clones for dengue 2 virus strain 16681 and  its attenuated vaccine derivative, strain PDK‐53. Virology 230:300‐308. 

29.  Laue, T., P. Emmerich, and H. Schmitz. 1999. Detection of dengue virus RNA  in patients after primary or secondary dengue infection by using the TaqMan automated amplification system. J Clin Microbiol 37:2543‐2547. 

30.  Lee, H., H. Shin, E. Wimmer, and A. V. Paul. 2004. cis‐acting RNA signals in the NS5B C‐terminal coding sequence of the hepatitis C virus genome. J Virol 78:10865‐10877. 

31.  Lin, K. C., H. L. Chang, and R. Y. Chang. 2004. Accumulation of a 3'‐terminal genome fragment in Japanese encephalitis virus‐infected mammalian and mosquito cells. J Virol 78:5133‐5138. 

32.  Lindenbach,  B.  D.,  H.  J.  Thiel,  and  C.  M.  Rice.  2007.  Flaviviridae:  the  viruses  and  their replication,  p.  1101‐1152.  In  D.  M.  Knipe  and  P.  M.  Howley  (ed.),  Fields  Virology,  5th  ed. Lippincott‐Raven Publishers, Philadelphia. 

33.  Liu, Y., E. Wimmer, and A. V. Paul. 2009. Cis‐acting RNA elements  in human and animal plus‐strand RNA viruses. Biochim Biophys Acta 1789:495‐517. 

34.  Lo, M., M. Tilgner, K. A. Bernard, and P.‐Y. Shi. 2003. Functional analysis of mosquito‐borne flavivirus conserved sequence elements within 3’ untranslated region of West Nile Virus by use of a reporting replicon that differentiates between viral translation and RNA replication. J Virol 77:10004‐10014. 

35.  Mason, P. W., S. V. Bezborodova, and T. M. Henry. 2002. Identification and characterization of a cis‐acting  replication element  (cre) adjacent  to  the  internal  ribosome entry site of  foot‐and‐mouth disease virus. J Virol 76:9686‐9694. 

36.  Men, R., M. Bray, D. Clark, R. M. Chanock, and C.‐J. Lai. 1996. Dengue  type 4 virus mutants containing  deletions  in  the  3'  noncoding  region  of  the  RNA  genome:  analysis  of  growth restriction in cell culture and altered viremia pattern and immunogenicity in Rhesus monkeys. J Virol 70:3930‐3937. 

37.  Miller, W. A., and K. A. White. 2006. Long‐distance RNA‐RNA  interactions  in plant virus gene expression and replication. Annu Rev Phytopathol 44:447‐467. 

38.  Paul, A. V. 2002. Possible unifying mechanism of picornavirus genome replication. ASM Press, Washington DC. 

53  

39.  Pijlman, G.  P., A.  Funk, N.  Kondratieva,  J.  Leung,  S.  Torres,  L.  van  der Aa, W.  J.  Liu, A.  C. Palmenberg,  P.  Y.  Shi, R. A. Hall,  and A. A. Khromykh.  2008. A  highly  structured,  nuclease‐resistant,  noncoding  RNA  produced  by  flaviviruses  is  required  for  pathogenicity.  Cell  Host Microbe 4:579‐591. 

40.  Samsa, M. M., J. A. Mondotte, N. G. Iglesias, I. Assuncao‐Miranda, G. Barbosa‐Lima, A. T. Da Poian, P. T. Bozza, and A. V. Gamarnik. 2009. Dengue virus capsid protein usurps lipid droplets for viral particle formation. PLoS Pathog 5:e1000632. 

41.  Steil,  B.  P.,  and  D.  J.  Barton.  2008.  Poliovirus  cis‐acting  replication  element‐dependent  VPg Uridylylation lowers the Km of the initiating nucleoside triphosphate for viral RNA replication. J Virol 82:9400‐9408. 

42.  Thompson,  J.  D.,  T.  J.  Gibson,  F.  Plewniak,  F.  Jeanmougin,  and  D.  G.  Higgins.  1997.  The CLUSTAL_X Windows  interface:  flexible  strategies  for multiple  sequence  alignment  aided  by quality analysis tools. Nucleic Acids Res 25:4876‐4882. 

43.  Thurner,  C.,  C. Witwer,  I.  L.  Hofacker,  and  P.  F.  Stadler.  2004.  Conserved  RNA  secondary structures in Flaviviridae genomes. J Gen Virol 85:1113‐11124. 

44.  Tilgner, M., and P. Y. Shi. 2004. Structure and function of the 3' terminal six nucleotides of the West Nile virus genome in viral replication. J Virol 78:8159‐8171. 

45.  Villordo, S. M., D. E. Alvarez, and A. V. Gamarnik. 2010. A balance between circular and linear forms of the dengue virus genome is crucial for viral replication. RNA 16:2325‐2335. 

46.  Villordo,  S. M., and A. V. Gamarnik. 2009. Genome  cyclization  as  strategy  for  flavivirus RNA replication. Virus Res 139:230‐239. 

47.  Welsch, S., S. Miller, I. Romero‐Brey, A. Merz, C. K. Bleck, P. Walther, S. D. Fuller, C. Antony, J. Krijnse‐Locker, and R. Bartenschlager. 2009. Composition and  three‐dimensional architecture of the dengue virus replication and assembly sites. Cell Host Microbe 5:365‐375. 

48.  You, S., and C. M. Rice. 2008. 3' RNA elements  in hepatitis C virus replication: kissing partners and long poly(U). J Virol 82:184‐195. 

49.  You, S., D. D. Stump, A. D. Branch, and C. M. Rice. 2004. A cis‐acting replication element in the sequence encoding the NS5B RNA‐dependent RNA polymerase  is required for hepatitis C virus RNA replication. J Virol 78:1352‐1366. 

50.  Zuker, M. 2003. Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res 31:3406‐3415. 

  

54  

CHAPTER THREE 

A NOVEL CODING‐REGION RNA ELEMENT MODULATES INFECTIOUS DENGUE VIRUS PARTICLE 

PRODUCTION IN BOTH MAMMALIAN AND MOSQUITO CELLS 

                  

A. M. Groat Carmona, S. Orozco, A. Payne, L. Kramer and E. Harris.  A Novel Coding‐Region RNA Element Modulates Infectious Dengue Virus Particle Production in both Mammalian and Mosquito Cells.  (manuscript submitted).  Permission to include Co‐Authored Material granted by UC Berkeley 2011. 

55  

INTRODUCTION  Dengue virus (DENV) is a small, enveloped positive‐sense RNA virus whose 5' and 3' untranslated regions (UTR) contain several conserved RNA elements that regulate various aspects of viral translation and RNA synthesis.  During the later stages of viral replication, large quantities of the viral structural proteins (C, prM/M and E) and  the viral RNA  (vRNA) are  synthesized and  subsequently assembled  into  immature noninfectious virions.  The fully assembled flavivirus virion consists of a nucleocapsid core containing the vRNA, surrounded by a  lipid bilayer derived  from the endoplasmic recticulum  (ER), which  includes the viral  E  and  prM/M  proteins  (36).    Flavivirus  virions  are  assembled  in  association  with  intracellular membranes (42), and mature virions are first observed  in the ER  lumen  in vitro (20, 27, 31, 35, 40, 44, 47, 51) and in vivo (21, 52).  However, while virions accumulate within membrane‐bound vesicles clear budding  intermediates and naked nucleocapsids have not been observed  (3, 58),  indicating  that viral assembly  is  rapid.   Assembly  is  localized  to  the membranous  sites of  replication by  the hydrophobic segments of the anchored capsid protein (46), and nucleocapsids acquire an envelope by budding  into the ER lumen.   The  viral  NS3  protein  has  been  implicated  in  the  intracellular  trafficking  of  the  progeny noninfectious virions from the ER through the Golgi compartment (6).   During virion maturation  in the trans Golgi compartment, the prM/M protein is processed to the mature M protein by the host protease furin, which exposes the E receptor binding domain that confers viral  infectivity (53, 66).    Intracellular M‐containing virions have not been detected,  suggesting  that prM/M cleavage occurs  just before  the release of mature virions (36, 43, 60).   Virion transport from the sites of replication to the cell surface involves vesicles derived  from  the ER  (9) and other components of  the exocytic pathway  (20, 22, 51).  Despite the information available on viral assembly, the mechanism by which the vRNA is recognized by the capsid protein to form the nucleocapsid precursor is still unknown.  Thus far, no candidate assembly signals have been  identified within the DENV genome, and the ability of subgenomic replicons  lacking the structural protein‐coding sequence to be assembled  in trans (28, 29) suggests that the signal does not reside in this area of the genome.   Much of what is known about RNA elements located in the UTRs has been shown to be equally important for viral replication in mammalian cells and mosquito cells, with few exceptions.  Deleting the variable region (VR), which is located within the 3’ UTR, reduces RNA synthesis in the mammalian baby hamster kidney cell line (BHK), but slightly amplifies RNA synthesis in an Ae. albopictus mosquito cell line (C6/36) (1).  On the other hand, changes to the bottom stem structure of the 3’ stem‐loop (SL), located at the 3’ end of the viral genome, were shown to have wild‐type (WT) replication levels in BHK cells but not  in C6/36 cells  (61).   To date, these are the only regions within the DENV genome  that display cell type‐associated differential effects on the viral life cycle, likely due to interactions with host‐dependent factors.    Indeed,  host‐specific  protein  interactions  are  vital  for  our  understanding  of  how  the  virus differentially modulates its life cycle in the human host and the mosquito vector, though more research is necessary to elucidate these molecular regulatory mechanisms.   There  is  little  information  available with  respect  to  RNA  elements  located within  the  coding region, though recently an RNA secondary structure, termed the capsid‐coding hairpin (cHP), was found to regulate both viral translation and RNA synthesis (7, 8).  Based on the structure prediction algorithms that have been published, the DENV coding sequence has been reported to be rather unstructured (25, 55),  though no  functional proof exists  that  the  coding‐region  lacks putative RNA  sequence elements.  While none have been identified in DENV, coding‐region RNA elements in a number of animal and plant viruses modulate  various  aspects  of  the  viral  life  cycle  besides  viral  translation  and  RNA  synthesis, including trafficking the vRNA to the membrane sites of replication, viral assembly, and other  less well characterized steps in viral replication (38, 41). 

56  

  This  study  sought  to  identify additional RNA elements  in  the DENV  coding  region  in order  to expand  our  knowledge  of  how  the  viral  life  cycle  is  regulated.   We  describe  a  novel  coding‐region regulatory RNA element termed the conserved capsid‐coding region 1 (CCR1), which regulates the viral life cycle in both mammalian and mosquito cells, though it displays a more prominent role in mosquito cells.  Consistent with our in vitro data, CCR1 mutant viruses do not replicate as well as WT DENV2 in Ae. aegypti  mosquitoes,  resulting  in  a  reduction  of  viral  transmission  to  the  salivary  glands.    We demonstrate that CCR1 plays a role in post‐RNA replication events, possibly acting as an assembly signal for  DENV  in  a  sequence‐dependent  and  cell  type‐specific manner.    These  studies  are  important  for improving our understanding of cis‐acting determinants that differentially regulate the viral life cycle in the human host and mosquito vector, potentially serving in the design of attenuated vaccine strains or as candidate targets for antiviral agents.   

MATERIALS AND METHODS  Sequence  conservation  and  secondary  structure  prediction  algorithms.    For  conserved  sequence prediction,  sequences  were  aligned  using  the  Clustal  W2  web  software  (54)  from  the  European Bioinformatics Institute.   Portions of the Clustal W2 alignments were processed by the RNAalifold web server  (24)  from  the  Institute  of  Theoretical  Chemistry  at  the  University  of  Vienna  to  generate  the phylogenetically conserved secondary structure predictions.   Individual RNA secondary structures were predicted  using  the  mfold  3  web  server  (65)  from  the  RNA  Institute  at  the  University  at  Albany.  Reported ΔG  values  reflect  enf  refinement  (the mfold  free‐energy  computation  incorporating  coaxial stacking and the Jacobson‐Stockmeyer theory for multibranched loops) (65).  Construction of mutant DNA constructs.   Mutations were  introduced  into the capsid‐coding region of an  infectious  clone  (IC) of  the DENV2 Thai  strain 16681  (pD2/IC‐30P, hereafter pD2/IC; a gift  from R. Kinney, Centers  for Disease Control and Prevention, Fort Collins, CO) using splicing overlap extension‐PCR  (SOE‐PCR), where  the T7 promoter and  the  first 1,391 nucleotides  (nts) of DENV2 were amplified and  cloned  into  a  pCR®2.1‐TOPO  vector  (Invitrogen).    Resulting  DNA  constructs were  digested with SacI/SphI and the inserts purified with NucleoSpin Extract II (Macherey‐Nagel) before being ligated into a SacI/SphI‐digested pD2/IC or a SacI/SphI‐digested pD2/IC‐RdRpmut  (“GVD”), as previously described (7, 8).   For the Renilla luciferase reporter constructs, the T7 promoter and the first 167 nts from the IC constructs  were  amplified,  and  resulting  products  were  digested  with  NotI/MluI  and  ligated  into  a NotI/MluI‐digested pDRrep  as previously described  (7).    For  the  “delta”  replicon, CCR1 was  removed from pDRrep using PCR.  Primer sequences are listed in Table 3.1.  Generation of RNA templates.   For all DNA templates (pD2/IC, pD2/IC‐RdRpmut, pDRrep and pDRrep‐RdRpmut),  linearized  templates were  generated  by  digestion with  XbaI  and  purified  using  a  25:24:1 phenol/chloroform/isoamyl  alcohol extraction  (pH 6.7 + 0.2,  Fischer  Scientific).   RNA  templates were generated via in vitro transcription using the RiboMax Large Scale RNA Production System (T7, Promega) with the following modifications to the manufacturer’s protocol: 5 mM each GTP, CTP and UTP; 1 mM ATP; 5 mM m7G(5’)ppp(5’)A cap analog  (New England Biolabs) were  incubated  for 4 hours  (h) at 30oC with the addition of 2 mM ATP after 30 minutes (min).  RNAs were subsequently treated with 80 U/mL TURBO DNase (Ambion) for 15 min at 37oC, and unincorporated nucleotides were removed from DENV RNAs by size exclusion chromatography using Micro Bio‐Spin P‐30 Tris columns (BioRad Laboratories).  Cell  culture.    Baby  hamster  kidney  cells  (BHK‐21,  clone  15;  hereafter  BHK) were  grown  in minimal essential medium‐alpha  (MEMα, Gibco) with 5% fetal bovine serum (FBS; HyClone) at 37oC  in 5% CO2.  

57  

For infections in Ae. albopictus mosquito cells (C6/36, ATCC #CRL‐1660), cells were grown in Leibovitz’s medium  (L‐15,  Gibco)  with  10%  FBS  at  28oC  without  CO2.    For  transfections  in  C6/36  cells,  a  high temperature  (HT) variant was utilized  (gift from A. Garmanik, Fundación  Instituto Leloir, Buenos Aires, Argentina) and grown in L‐15 medium with 10% FBS at 33oC without CO2.  All media was supplemented with 10 mM HEPES (pH 7.5), 100 units/mL penicillin, and 100 µg/mL streptomycin.  Virus  titration by plaque  assay.   Plaque  assays using BHK  cells  for  virus  titration were  conducted  as described previously (7, 8, 10).  The viral titers in Ae. aegypti mosquito bodies and salivary glands were determined by plaque assay on African green monkey kidney cells  (Vero, ATCC #CCL‐81) as previously described (48).  Viral titer was calculated as plaque forming units (pfu) per mL, and the limit of detection was defined as the lowest titer of virus that could be accurately measured in BHK (1.9 pfu/mL) and Vero (1.4 pfu/mL) cells.   For transfections, viral titers were normalized to the amount of viral RNA (vRNA)  in the cells at the time that the virus‐containing supernatants were harvested, as determined by qRT‐PCR.  p‐values relative to the WT control were derived by Wilcoxon rank sum test in Mstat.  RNA transfections.  BHK cells were seeded in 24‐well plates, grown to 50% confluence, and transfected with 0.5 µg IC or replicon RNA using 4X Lipofectamine 2000 (Invitrogen), followed by a 2‐h incubation at 37oC in 5% CO2.  BHK cells were washed 3X in MEMα medium and incubated at 37oC with 5% CO2 until harvest  at  indicated  timepoints.    C6/36  HT  cells  were  seeded  in  12‐well  plates,  grown  to  95% confluence, and transfected with 1 µg IC or replicon RNA using 6X Lipofectamine 2000, followed by a 3‐h incubation at 33oC without CO2.   C6/36 HT cells were washed 1X  in L‐15 media and  incubated at 33oC without CO2 until harvest at indicated timepoints.  Virus‐containing supernatants and intracellular virus were  collected,  and  viral  titer  was  assessed  by  plaque  assay.    Cellular  RNA  was  harvested  from  a duplicate well, purified using  the Mini RNA  Isolation  II Kit  (Zymo Research), and vRNA concentrations were  determined  using  quantitative  reverse  transcriptase  (qRT)‐PCR.    vRNA  in  the  virus‐containing supernatants was isolated using the ZR Viral RNA kit (Zymo Research) and quantified using qRT‐PCR.  Viral infections.  Virus‐containing supernatants were harvested at 72‐h post‐transfection (hpt) from IC‐transfected  BHK  cells.    Virus‐containing  supernatants  were  concentrated  by  ultracentrifugation  at 89,500 g for 2‐h using a 5% cold sucrose column, and viral titer was assessed by plaque assay.  vRNA in the concentrated virus‐containing supernatants was  isolated using the ZR Viral RNA kit and quantified using qRT‐PCR.   BHK  and C6/36  cell monolayers were  infected  at  an MOI of  0.01,  as determined by plaque  assay  or with  approximately  2.5  genome  equivalents  (G‐Eq)/cell,  as  determined  by  qRT‐PCR.  Infections were  incubated at 37oC with 5% CO2 (BHK) or 28

oC without CO2 (C6/36) and at 2 (BHK) or 3 (C6/36) h post‐infection  (hpi),  cells were washed  1X with MEMα  (BHK) or  L‐15  (C6/36) medium  and incubated at 37oC with 5% CO2 (BHK) or 28

oC without CO2 (C6/36) until harvest at indicated timepoints.  Virus‐containing supernatants were collected, and viral titer was assessed by plaque assay.  The vRNA in the virus‐containing supernatants was isolated using the ZR Viral RNA kit and quantified using qRT‐PCR.  Ae. aegypti mosquitoes.   The Ae. aegypti colony originated from eggs collected in Panama by J. Loaiza and were  received  at  the Arbovirus  Laboratory  (Wadsworth  Center)  in  2007.    Colonized mosquitoes were maintained on defibrinated rabbit blood with 10% sucrose (for egg laying) and given 10% sucrose ad  libitum.    Larvae were  reared  and  adults maintained  at  27°C with  70%  relative humidity  (RH)  and under a 16:8 light:dark diurnal cycle in 30.5‐cm3 cages.  Infections in Ae. aegypti.   Virus‐containing supernatants were harvested at 72 hpt from IC‐transfected BHK cells.   Virus‐containing supernatants were concentrated by ultracentrifugation at 89,500 g for 2‐h using a 5% cold sucrose column, and viral titer was assessed by plaque assay.  Viruses were diluted to a 

58  

titer of 104 pfu/mL in mosquito diluent (MD; 20% heat‐inactivated FBS in PBS (Dulbecco), supplemented with 50 µg/mL penicillin/streptomycin, 50 µg/mL gentamicin,  and 2.5 µg/mL  fungizone).   Two‐to‐five day‐old mosquitoes were inoculated intrathoracically (IT) with 0.1‐1 pfu under CO2 anesthesia using the FemtoJet microinjection system (Eppendorf) and held at 30°C with a 16:8 light:dark photoperiod for up to 20 days post‐inoculation (dpi).  Mosquito bodies were placed into 1 mL MD, and samples were stored at −70°C until they were homogenized  in a mixer mill (Qiagen) and clarified by centrifugation at 1,200 rpm  for  3  min.    Body  titer  was  evaluated  at  indicated  timepoints,  and  the  viral  load  in  infected mosquitoes was determined by plaque assay.  Salivary glands were dissected from infected mosquitoes at day 10 post‐inoculation and placed  into 0.3 mL MD, while the mosquito remnants were placed  in 1 mL MD; samples were stored at −70°C.  Viral load in the salivary glands was determined by plaque assay.  Luciferase assay.  BHK cells transfected transfected with the Renilla luciferase reporter constructs were harvested  at  the  indicated  timepoints,  and  samples were  analyzed with  the  Renilla  Luciferase Assay System  (Promega)  according  to  the  manufacturer’s  instructions  on  a  GloMax‐96  Microplate Luminometer (Promega) with the following modifications to the manufacturer’s instructions: 50 µL cell lysate and 40 µL of 1X Renilla Luciferase Assay Substrate was used for each sample.   Translation  levels were normalized  to  the  relative  transfection efficiency at 2 hpt as determined by qRT‐PCR.   p‐values relative to the WT control were derived by Wilcoxon rank sum test in Mstat.  Quantitative  RT‐PCR.    Cellular  RNA was  harvested  at  indicated  timepoints  from  duplicate wells  and purified  using  the Mini  RNA  Isolation  II  Kit,  and  the  vRNA  in  the  virus‐containing  supernatants was isolated using the ZR Viral RNA kit.  qRT‐PCR was performed using a modification to the method of Laue et al.  (34), which uses  a  FAM‐TAMRA  labeled probe  and primer  set  that  is directed  against  the  viral protein NS5 (Table 3.2).  The TaqMan One‐Step Master Mix (Applied Biosystems) system was used with the  following modifications  to  the protocol:  the RT step was conducted at 48oC  for 30 min,  the  initial denaturation step was performed at 95oC for 10 min, the annealing temperature was set at 60oC, and the reaction volume was fixed at 20 µL.  For the purposes of determining transfection efficiency, target vRNA was first normalized to the cellular 18S RNA using the TaqMan VIC‐MGB Primer Limited Eukaryotic 18S rRNA Endogenous Control (Applied Biosystems) in a parallel reaction.  qRT‐PCR was conducted on a Sequence Detection System 7300 (Applied Biosystems).  Determination  of  vRNA  half‐life.    Total  cellular  RNA was  harvested  from  BHK  cells  transfected with pD2/IC‐RdRp constructs ("GVD") at 2, 4, 8, 24, 48 and 72 hpt, treated with DNase and quantified by NS5 qRT‐PCR as described above.  Values were graphed, and an exponential trendline was fit to the data in Microsoft  Excel  2007.    The  half‐life  of  each  variant  in  each  experiment  was  determined  from  the exponential trendline equation, averaged and shown  in Fig. 3.5C.   p‐values relative to the GVD control were derived by Wilcoxon rank sum test in Mstat.  Intracellular virus extraction.   Virus‐containing  supernatants were harvested  from  IC‐transfected BHK and C6/36 HT cell monolayers at  indicated timepoints and assessed by plaque assay to determine the extracellular viral  titers.   To obtain  the  intracellular viral  titers, cells were washed 1X  in PBS and  then incubated  for  3  min  at  4oC  with  an  alkaline/high‐salt  solution  of  1  M  NaCl  plus  50  mM  sodium bicarbonate, pH 9.5, to remove the surface‐bound virus (11, 23).   Cells were washed 2X  in PBS before the  addition of 0.25% Trypsin‐EDTA  (Gibco)  to detach  cells  from  the plates.   Cells were harvested  in MEMα  (BHK) or L‐15  (C6/36) medium and  lysed during a single freeze‐thaw cycle.   Cellular debris was pelleted by centrifugation at 3,200 g  for 5 min, and  the  intracellular virus‐containing supernatant was removed and assessed for viral titer by plaque assay.   Extra‐ vs.  intracellular viral concentrations were calculated  as  ratios,  where  a  ratio  greater  than  one  indicates  that  there  was  no  accumulation  of 

59  

infectious  virus within  the  cell  and  ratios  less  than  one  indicated  that  there was more  intracellular infectious virus than what was being released into the supernatant.  p‐values relative to the WT control were derived by Wilcoxon rank sum test in Mstat.  Statistical  analysis.   Wilcoxon  rank  sum  test were  performed  in Mstat.    Statistical  significance was defined as p<0.05.   

RESULTS  A novel RNA element  is predicted to be conserved both  in sequence and secondary structure  in the DENV  and  TBEV  serogroups.    Candidate  selection  of  coding‐region  RNA  elements was  based  on  an alignment of sequences from the representative genomes of DENV1‐4, generated using Clustal W2 (54).  Putative RNA  sequence elements were  identified by  comparing  the  level of homology within a given stretch  to  that  of  the  surrounding  sequences,  which  was  defined  as  30  nucleotides  upstream  and downstream of the putative element.  Putative RNA sequence elements required at least 70% homology in the region of interest as compared to the background sequences.  Preliminary screening (discussed in Chapter  2)  and  further  investigation  of  the  DENV1‐4  sequence  alignment  revealed  that  within  the capsid‐coding  region,  the  5’  conserved  sequence  (CS)  is  located  at  the  5’  end  of  a  short  stretch  of sequence  that  contains a greater degree of  sequence homology  (~89%)  than  that of  the  surrounding sequences  (~64%)  (Fig.  3.1A).   While  the  5’  CS  is  a  highly  conserved  sequence  in  the  capsid‐coding region,  it was  included  in  the  calculation  of  the  surrounding  background  sequence  for  the  putative region of interest, though this did not change the observed sequence homology in the region of interest.  When  an  alignment  of  sequences  from  the  representative  genomes  of  viruses  in  the  tick‐borne encephalitis virus (TBEV) serogroup was examined using Clustal W2 (54), the same stretch of sequence was  identified  downstream  from  the  initiator  AUG,  which  contains  a  greater  degree  of  sequence homology  (100%)  than  that  of  the  surrounding  sequences  (65%)  (Fig.  3.1B).    However,  neither  the sequence nor the predicted secondary structure of CCR1 could be identified in members of the Japanese encephalitis virus (JEV) serogroup (data not shown).  Based on this high level of sequence conservation, the putative RNA sequence element was termed the conserved capsid‐coding region 1 (CCR1).   Since the function of viral RNA elements can be dependent on the primary nucleotide sequence as well as secondary structure, CCR1 was examined  for structural conservation  that might explain  the high degree of sequence homology observed in both the DENV and TBEV serogroups.  The DENV1‐4 and TBEV serogroup alignments were modified so as to include the 5’ UTR as well as the entire capsid‐coding region  to  generate  phylogenetically‐conserved  structure  predictions  using  RNAalifold  (24).    These secondary  structure  predictions  showed  that  the  putative  RNA  element  CCR1  is  contained  entirely within a small hairpin structure that is maintained by covariation in both the DENV and TBEV serogroups (Fig.  3.1C  and  3.1D).    While  the  predicted  secondary  structure  of  CCR1  shows  a  high  degree  of covariance, it is not as highly conserved as the predicted secondary structures of other known elements such as  the cHP  (Fig. 3.1C and 3.1D),  indicating  that  the predicted secondary structure may not be as important as the primary nucleotide sequence.  Since the nucleotide sequence of the 5’ CS and CCR1 is highly conserved, the amino acid sequence in this region of capsid is also conserved in DENV1‐4 though amino  acid  conservation  alone  is  not  sufficient  to  explain  the  observed  sequence  and  structure conservation of CCR1  in both the DENV and TBEV serogroups.   Taken together,  it  is  likely that CCR1  is conserved both in sequence and secondary structure because it plays a putative role in modulating the viral life cycle.  

60  

CCR1 modulates the viral life cycle in both BHK and C6/36 cells.  To determine whether CCR1 regulates the viral  life cycle, silent point mutations were  introduced  into an  infectious clone of  the DENV2 Thai strain 16681, termed pD2/IC (30).  To disrupt the predicted secondary structure with minimal changes to the  primary  nucleotide  sequence,  two  nucleotide  changes  were  introduced  into  either  side  of  the predicted  stem  structure  (“5’/3’  flank mut,” Fig. 3.2A, Table 3.3), which  linearized  the  stem‐loop and eliminated  competing  structure predictions  in  the mfold prediction  software  (65).   Next,  the primary nucleotide sequence was mutated without affecting  the predicted secondary structure by  introducing six nucleotide changes into CCR1 (“seq mut,” Fig. 3.2A, Table 3.3) in such a way that the predicted free energy  is similar  to  that of  the wild‐type  (WT).   Thus, despite minor changes  to  the stem,  the overall structure of CCR1  is maintained  in  this mutant construct, as predicted by mfold  (65).   A  third mutant, termed the “center mut” construct, contains only the nucleotide changes that were  introduced  in  the predicted loop of the “seq mut” construct (Fig. 3.2A, Table 3.3).  As a complement to the “center mut” mutant construct, the mutations that were  introduced  in the 5’ and 3’ portions of the predicted stem structure of the “seq mut” mutant construct were also isolated (“5’ flank mut” and “3’ flank mut,” Table 3.3).  For all designed mutant constructs, amino acid coding‐sequence, reading frame, and codon usage bias  in  baby  hamster  kidney  (BHK)  and  Ae.  albopictus mosquito  (C6/36)  cells was maintained while preserving the surrounding secondary structures as predicted by the mfold (65) (Fig. 3.2A).   RNA  transcripts  were  transfected  into  BHK  or  C6/36  cell  monolayers,  and  virus‐containing supernatants were  harvested  at  various  timepoints  (indicated  as  hours  post‐transfection,  hpt)  to  be assayed  for viral  infectivity via plaque assays.   As DENV  replication kinetics are  slower  in C6/36  cells, virus‐containing supernatants were assayed  for viral  infectivity at  later times post‐transfection than  in BHK cells (BHK: 24, 48 and 72 hpt; C6/36: 72, 96, 120 and 144 hpt).  Interestingly, all mutant constructs demonstrated a small plaque phenotype (Fig. 3.2B), but only the mutations that disrupted the primary nucleotide  sequence  (“seq mut”)  rather  than  the  predicted  secondary  structure  (“5’/3’  flank mut”) significantly  reduced  the viral  titer  in BHK cells at all  timepoints  tested  (Fig. 3.2C).   Since  the primary nucleotide  sequence of CCR1  is  important  in BHK  cells, an additional mutant  construct was designed with nine sequence changes, termed “seq mut v2” (Table 3.3).   When transfected  into BHK cells, both sequence mutants  (“seq mut” and  “seq mut v2”) displayed a no‐plaque phenotype at 24 hpt, and at later timepoints both showed a significantly lower viral titer in comparison to the WT (Fig. 3.2C).  While the “seq mut” construct showed a small plaque phenotype, the “seq mut v2” construct demonstrated an  extra  small  plaque  phenotype  (Fig.  3.2B)  at  all  timepoints  yielding  detectable  virus.   Altering  the primary  nucleotide  sequence  decreased  viral  titer  in  BHK  cells, which was  influenced  in  part  by  the central sequences since the “center mut” construct had a significantly lower viral titer in comparison to the WT  at  all  timepoints  tested  and  the  “5’  flank mut”  construct diplayed WT  replication  levels  (Fig. 3.2C).   In contrast to BHK cells, changes to the primary nucleotide sequence (“seq mut” and “seq mut v2”) as well as changes to the predicted secondary structure (“5’/3’ flank mut”) significantly decreased the viral titer in C6/36 cells at all timepoints tested (Fig. 3.2D).  The “seq mut” construct did not result in any  infectious virus at early  timepoints, and viral  titers barely exceeded  the  limit of detection at  later times  post‐transfection, while  the  “seq mut  v2”  construct maintained  a  no‐plaque  phenotype  at  all timepoints tested  (Fig. 3.2D).   Unlike  in BHK cells, the effect on viral titer  in C6/36 cells appears to be primarily mediated by  the 5’ and 3’  flanking  sequences  since  the  “center mut”  construct had  slightly higher titers at  later timepoints than the “5’/3’  flank mut” and “5’  flank mut” constructs although the “center mut”  construct displayed a  significantly  lower  titer  than  the WT at all  timepoints  tested  (Fig. 3.2D).  In transfected C6/36 cells, all mutant constructs tested that produced detectable infectious virus displayed  a  small  plaque  phenotype.    Taken  together,  these  results  indicate  that  CCR1  functions primarily  as  an  RNA  sequence  element  in  both  BHK  and  C6/36  cells,  though  these  effects  are more dramatic  in  C6/36  cells.    Furthermore,  the  structure  of  CCR1  may  also  be  required  for  efficient 

61  

replication in C6/36 cells since the “5’/3’ flank mut” construct replicated less efficiently than the “center mut” construct.   As transfections bypass key steps  in viral replication, such as entry and uncoating, we assayed the mutant constructs  in  the context of viral  infections.   To determine  if CCR1 has an effect on early steps  in the viral  life cycle, virus‐containing supernatants were harvested from transfected BHK cells at 72 hpt,  concentrated,  titered,  and used  to  infect BHK  and C6/36  cell monolayers  at  an MOI of 0.01.  Virus‐containing supernatants from these infected cells were harvested at various timepoints (indicated as hours post‐infection, hpi; BHK: 24, 48 and 72 hpi; C6/36: 72, 96, 120 and 144 hpi) and assayed  for viral infectivity by plaque assay.  In BHK cells, all mutant viruses displayed a significantly lower viral titer at 24 hpi,  though  the  “5’/3’  flank mut,”  “5’  flank mut” and  “seq mut v2” mutant viruses acquire WT levels by 72 hpi (Fig. 3.3A), implying that these viruses can accumulate after serial passage in BHK cells.  In contrast, the “seq mut” and “center mut” mutant viruses maintained a significantly lower titer in BHK cells  at  all  timepoints  tested  (Fig.  3.3A), which  supports  the  conclusion  that  the  primary  nucleotide sequence rather than the predicted secondary structure of CCR1 is important for replication in BHK cells.  Despite displaying WT levels of replication in the transfected BHK cells and at later timepoints during the infections  in  BHK  cells,  the  “5’/3’  flank mut” mutant  virus  has  a  low  viral  titer  at  24  hpi  (Fig.  3.3A), possibly  indicating  that  there  is  an  underlying  replication  defect  that  delays  viral  output  early  in infection.   Alternatively,  it  is possible that the observed  increase  in viral titer  in all mutant viruses over the  course of  the assay  reflects  the buildup of  viral  revertants.   The  likelihood  that  serial passage of mutant viruses in BHK cells can lead to the accumulation of reversions is supported by the observation that the “center mut” mutant virus demonstrates an  intermediate plaque phenotype (Fig. 3.2B)  in the infected  BHK  cells  whereas  the  remaining mutant  viruses  continue  to  demonstrate  a  small  plaque phenotype.   In contrast, the infected C6/36 cells displayed the same growth kinetics and plaque phenotypes as  the  transfected C6/36 cells  (Fig. 3.2D and 3.3B).   The “seq mut” and “seq mut v2” mutant viruses never exceeded  the  limit of detection at all  timepoints  tested, whereas  the  “5’/3’  flank mut” mutant virus displayed a no‐plaque phenotype at early timepoints and a significantly lower viral titer throughout the assay  (Fig. 3.3B).   The “center mut” construct, however, displayed an  intermediate growth curve, consistent  with  the  results  obtained  in  the  transfection  experiments  in  C6/36  cells  (Fig.  3.3B).    To determine whether normalizing  the  infections using plaque  forming units  (pfu) could skew viral  titers, the infections were repeated by normalizing to the amount of vRNA (measured as genome equivalents, G‐Eq) in the supernatant at the time of infection, as determined by qRT‐PCR.  While normalizing the BHK infections using G‐Eq did not alter  the phenotypes observed by normalizing  the  infections  to pfu  (Fig. 3.3C),  the  delayed  replication  kinetics observed  in  infected C6/36  cells were  even more pronounced when normalizing  to G‐Eq  (Fig. 3.3D).   Since  the  transfected vRNA behaved similarly  to  the  infectious virus in both BHK and C6/36 cells, it is unlikely that early steps of the viral life cycle (i.e. entry/uncoating) are mediated by CCR1.  Altering CCR1 impairs viral replication in Ae. aegypti mosquitoes.  Since changes to the CCR1 primary nucleotide  sequence were  shown  to have more drastic effects  in mosquito  cells  (Fig. 3.2D, 3.3B  and 3.3D), Ae. aegypti mosquitoes were infected with CCR1 mutant viruses to determine whether the results obtained in the in vitro experiments could be observed in an in vivo model of DENV infection.  Biological transmission of  flaviviruses by arthropods depends on  ingesting a blood meal  that  contains  sufficient virus to establish an infection in the epithelial cells lining the mesenteron (midgut) (49).  However, since the  mutant  viruses  were  shown  to  have  a  delayed  replication  rate  in  infected  C6/36  cells  when compared  to WT DENV2, Ae. aegypti mosquitoes were  inoculated  intrathoracically  (IT)  so  that all  the viruses would begin their replication cycle at the same time  in the parenteral tissues.   Virus‐containing supernatants were  collected  in  the  same manner as  for  the  in vitro  infections and used  to  infect Ae. 

62  

aegypti mosquitoes  IT with  approximately  104  pfu/mL.    All  inoculated mosquitoes  showed  75‐100% infection by day 20 post‐inoculation by plaque assay (data not shown),  indicating that all viruses were able to replicate in the parenteral tissues of Ae. aegypti mosquitoes after IT inoculation.  The mosquito bodies were assessed for infectious virus at various timepoints (indicated as days post‐inoculation, dpi; days 5, 10 and 20) by plaque assays.  All mutant viruses displayed a significantly lower viral titer than the WT  at  all  timepoints  tested,  though  the  “mut  seq  v2”  and  “5’/3’  flank mut” mutant  viruses  did  not replicate as well as  the “seq mut” and “center mut” mutant viruses  (Fig. 3.4A).   Thus,  the  replication defects observed in the in vitro experiments are reflected in reduced replication in the mosquito vector for all mutant viruses at all timepoints tested.  Overall, the in vivo infections in Ae. aegypti mosquitoes support the  in vitro results that CCR1 functions predominately as an RNA sequence element, mediated in part by the 5’ and 3’ flanking sequences.   During viral replication  in the orally  infected mosquito, the virus must escape from the midgut epithelium  into the hemocele  (body cavity)  in order to disseminate to the brain, fat body and salivary glands of the infected mosquito (49).  Once in the salivary glands, the flaviviruses must undergo another round  of  replication  to  generate  enough  virus  to  be  secreted  in  the  saliva  and  transmitted  to  a susceptible  host  (49).    To  assess whether  the  replication  defects  observed  in  vivo  could  affect  viral transmission  in  the mosquito vector,  the  salivary glands were dissected  from  infected mosquitoes on day 10, and viral  titer was assessed by plaque assay.   Only 30% of  the mosquitoes  infected with  the “5’/3’ flank mut” mutant virus were capable of transmitting the virus, as compared to 100% of the WT‐infected mosquitoes  (Fig.  3.4B).    In  addition,  there was  a  2.6‐fold  decrease  in  the  amount  of  virus present  in  the  salivary  glands  of  the  “5’/3’  flank  mut”‐infected  mosquitoes  as  compared  to  the mosquitoes infected with the WT virus (Fig. 3.4B).  This implies that changes to the CCR1 sequence not only  affect  viral  replication  in  the  mosquito  vector  for  up  to  20  dpi,  but  that  the  effect  on  viral replication could have a major impact on viral transmission by day 10 post‐inoculation.  Despite the fact that both  the  in vitro and  in vivo experiments  support  the conclusion  that CCR1  is an  important RNA sequence element, the question remains as to how it regulates the viral life cycle.  Mutating CCR1 has no effect on viral translation or RNA synthesis in either BHK or C6/36 cells.  Most of the known RNA elements in DENV modulate the viral life cycle by affecting translation (1, 5, 8) or RNA synthesis (1, 2, 7, 14‐16).  To determine whether the decreased viral titers in CCR1 mutants were due to effects on either viral translation or RNA synthesis, all mutations were subcloned into a Renilla luciferase reporter  construct,  termed pDRrep  (7).    Since  the  first 72 nucleotides of  the  capsid‐coding  sequence present in the DENV2 reporter constructs contain the CCR1 sequence, it was not necessary to introduce additional changes  into pDRrep for this study.   The pDRrep‐RdRpmut construct was used as a negative control (“GVD”) because this replicon cannot replicate due to a mutation  in the active site of the RNA‐dependent RNA polymerase (RdRp) (7).  As with other DENV and WNV replicons, the replication kinetics are slower compared to the infectious clone (1, 26, 39, 56).   RNA transcripts were transfected into BHK cell monolayers and assayed for luciferase activity at 4,  8,  12,  24,  48,  72  and  96  hpt.    In  BHK  cells,  none  of  the mutant  constructs  showed  a  significant difference in luciferase counts at early timepoints post‐transfection when compared to the WT and GVD controls,  indicating  that CCR1 does not  influence viral  translation  (Fig. 3.5A).    Likewise,  there was no significant difference  in  luciferase  levels at  later timepoints when compared to the WT,  indicating that there  is  also  no  effect  on  RNA  synthesis  in  BHK  cells  (Fig.  3.5A).    Deleting  CCR1  from  the  Renilla luciferase  replicon  (“delta”)  also had no  significant  effect on  luciferase  counts  at  either  early or  late timepoints, which  confirms  the observation  that CCR1  is dispensable  for viral  replication  in BHK  cells (Fig. 3.5A).  Unfortunately, the full panel of mutant constructs could not be tested in C6/36 cells due to technical  difficulties  with  transfecting  this  particular  cell  line.    Nonetheless,  the  “5’/3’  flank  mut” construct was transfected into C6/36 cell monolayers and assayed for luciferase activity at 4, 8, 72, 96, 

63  

120 and 144 hpt.   While the WT did not behave as expected at  later timepoints post‐transfection, the preliminary experiments showed that the “5’/3’ flank mut” construct displayed luciferase counts similar to the WT and GVD constructs at early timepoints and significantly higher luciferase levels compared to the GVD replicon later in the assay (Fig. 3.5B).  Taken together, altering the predicted stem structure of CCR1  has  no  effect  on  viral  translation  or  RNA  synthesis  in  C6/36  cells,  despite  displaying  delayed replication kinetics in transfected and infected C6/36 cells.  These results suggest that CCR1 influences a stage of the viral life cycle that occurs after viral translation and RNA synthesis in both mammalian and mosquito cells, which has not yet been observed with the other DENV RNA elements identified thus far.  Observed effects of CCR1 mutant constructs on viral titer are not due to RNA instability.  To determine whether  the  introduction  of  silent  point  mutations  into  the  primary  nucleotide  sequence  of  CCR1 destabilized the RNA, all mutations were subcloned  into a GVD variant of the  infectious clone, termed pD2/IC‐RdRpmut (7), which is unable to replicate due to a mutation in the active site of the RdRp.  RNA transcripts derived from this backbone will persist within the cell after transfection, and the rate of RNA degradation for each construct can be measured using qRT‐PCR to derive the RNA half‐life, as compared to the GVD control.   For all constructs tested, none of the  introduced changes significantly altered the RNA half‐life  in comparison  to  the GVD control  (Fig. 3.5C), which  implies  that RNA  instability  is not a cause for the observed effects of CCR1 on viral titer in BHK cells.  CCR1  regulates  infectious  particle  production  in  BHK  and  C6/36  cells.    Though  infectious  virus production can be measured by plaque assay, it does not account for non‐infectious viral particles that are routinely released during infection or infectious particles that do not plaque.  In order to determine the amount of assembled viral particles  in  the virus‐containing  supernatants  from  IC‐transfected BHK and C6/36 cells, the amount of vRNA was quantified using qRT‐PCR.  For all the mutant constructs, there was no  significant decrease  in  the amount of vRNA present  in  the  supernatant  for all  the  timepoints tested  in BHK  cells  (Fig. 3.6A).    In C6/36  cells,  the  “seq mut”  and  “seq mut  v2”  constructs displayed significantly  lower amounts of vRNA as compared  to  the WT  (Fig. 3.6B),  though  these  levels  (~106 G‐Eq/mL)  were  still  high  considering  that  they  had  a  no‐plaque  phenotype  during  transfections  and infections in C6/36 cells (Fig. 3.2D and 3.3B).  The “5’ flank mut” construct also showed significantly low levels of  vRNA  as  compared  to  the WT, however,  at  the  early  timepoints  this mutant  construct  also displayed a no‐plaque phenotype in transfected and infected C6/36 cells (Fig. 3.2D and 3.3B).  However, the amount of vRNA released from these IC‐transfected C6/36 cells was significantly higher than a GVD control  (data not  shown).   This  result  indicates  that  there are  similar amounts of viral particles being released from WT and mutant  IC‐transfected BHK and C6/36 cells but they differ  in their  infectivity as determined by plaque assay.   To determine the efficiency of infectious particle production during transfections, the amount of vRNA present within the virus‐containing supernatants was measured using qRT‐PCR and compared to the viral titer, generating a particle‐to‐pfu ratio for each mutant construct.  In BHK cells, the “5’/3’ flank mut” construct displayed  the same particle:pfu ratio as  the WT whereas  the “seq mut,” "seq mut v2" and “center mut” constructs displayed a significantly greater particle:pfu ratio at all timepoints tested (Fig. 3.6C), indicating that more noninfectious viral particles are being released into the supernatant for these constructs.  Interestingly, all the mutant constructs that showed lower replication kinetics in BHK cells had a significantly higher particle:pfu ratio (Fig. 3.2C and 3.6C).  Similarly, in transfected C6/36 cells, both the “seq mut” and “seq mut v2” constructs have an increased particle:pfu ratio as compared to the WT, demonstrating that mutating the primary sequence of CCR1 impacts infectious particle production in both BHK and C6/36 cells (Fig. 3.6D).  The “5’/3’ flank mut” and “5’ flank mut” constructs also showed a  significantly higher particle:pfu  ratio  than  the WT,  indicating  that  the 5’/3’  flanking  sequences may contribute more to the observed phenotype in transfected C6/36 cells (Fig. 3.6D). 

64  

  In  infected  BHK  cells,  the  particle:pfu  ratio  for  all mutant  constructs  displayed  no  significant difference  from  the WT  at  all  timepoints  tested  (Fig. 3.6E),  supporting  the  conclusion  that  sustained passage of the mutant viruses in this cell line could generate revertant viruses.  The only exception was the “5’ flank mut” construct at 24 hpi (Fig. 3.6E).  In infected C6/ 36 cells, the “5’/3’ flank mut” and “seq mut” constructs  showed a  significantly higher particle:pfu  ratio  than  the WT,  indicating  that  the 5’/3’ flanking sequences contribute substantially to the observed phenotype (Fig. 3.6F).  This is confirmed by the observation that the “center mut” construct demonstrated an intermediate though significant effect on  particle:pfu  ratio  at  all  timepoints  tested  in  infected  C6/36  cells  (Fig.  3.6F),  consistent  with  its influence on viral titer in infected C6/36 cells (Fig. 3.3B).  Given that the “5’ flank mut” construct showed a  lower particle:pfu  level at  later timepoints post‐infection  implies that the 5’ flanking sequences may not be as important than the 3’ flanking sequences (Fig. 3.6F).  The overall trend does not change if the infections  are  normalized  using G‐Eq  in  both  BHK  (Fig.  3.6G)  and  C6/36  cells  (Fig.  3.6H)  though  the differences between the mutants and the WT become more pronounced in the infected C6/36 cells (Fig. 3H).  Taken together, this evidence suggests that specific portions of the CCR1 sequence have a cell‐type dependent  effect on  a post‐replication  stage of  the  viral  life  cycle  that modulates  infectious particle production.   Thus  far,  experimental  evidence  suggests  that mutating CCR1  interferes with post‐replication events such as viral assembly, virion processing/maturation or virion release.  To examine virion release, the amount of  infectious particles outside the cell was compared to the amount of  infectious particles within cells  in both BHK and C6/36 cells.   At various  timepoints post‐transfection  (BHK: 24, 48 and 72 hpt; C6/36: 72, 96, 120 and 144 hpt), cell monolayers were stripped of extracellular virus using a high‐pH,  high‐salt  solution  and  subjected  to  a  single  freeze/thaw  cycle  in  order  to  release  intracellular infectious  virus,  which  was  then  assayed  via  plaque  assay  alongside  supernatants  containing extracellular virus.   None of the mutant constructs showed a significant effect on the ratio of extra‐ to intracellular virus in BHK cells, except for the “seq mut” and “seq mut v2” constructs at 24 hpt, though at  this  timepoint  both  constructs  displayed  a  no‐plaque  phenotype  (Fig.  3.7A).    Similar  observations were  seen  in  C6/36  cells,  where  those  constructs  capable  of  producing  detectable  infectious  virus showed no significant retention of intracellular infectious virus as compared to the WT (Fig. 3.7B).  Since the decreases in viral titer are not due to accumulation of infectious virus within the cell, it is likely that mutating CCR1 results in a virion assembly or processing/maturation defect.   

DISCUSSION  Sequence and structure conservation was used as a platform  for  identifying putative RNA elements  in the DENV coding‐region, as RNA elements are more  likely to perform a function  in the viral life cycle if they are conserved.  CCR1 was identified as a putative RNA element since it is conserved both in primary nucleotide sequence and predicted secondary structure in both the DENV and TBEV serogroups.  In this study, we have shown that CCR1 functions as an RNA sequence element in BHK and C6/36 cells, though its effects on viral titer and infectious virus production are more pronounced in mosquito cells after both transfection  and  infection.    Furthermore, mutations  to  CCR1  reduce  viral  replication  in  infected  Ae. aegypti mosquitoes and decrease viral dissemination to the salivary glands.  Additionally, CCR1 functions primarily through  its central sequences  in BHK cells, whereas  in C6/36 cells  its effects are mediated by both the 5’ and 3’ flanking sequences as well as the central sequences.   RNA  elements  that  act  in  cell‐type  specific manner  can  help  explain  how  viral  replication  is differentially  regulated,  which  is  of  particular  relevance  to  understanding  differences  in  arboviral replication within the mammalian host and the mosquito vector.  For example, within the Sindbis (SIN) viral genome, there is an 51‐nucleotide long conserved sequence element in the nsP1‐coding sequence 

65  

that  is  predicted  to  form  two  smaller  stem‐loop  structures  that  act  as  replication  enhancers whose effects  are more  important  in mosquito  cells  than  in mammalian  cells  (13,  17,  45).    The  fact  that different  portions  of  the  CCR1  sequence  are  implicated  in mediating  viral  titer  and  infectious  virus production  in BHK and C6/36 cells  implies  the  involvement of cell‐type specific  factors  in  its  function.  Though CCR1 has been shown  to  function as an RNA sequence element,  the  fact  that  the “5’/3’  flank mut” construct had a more severe phenotype in C6/36 cells compared to the BHK cells could imply that the predicted  secondary  structure of CCR1  is more  important  to  the viral  life  cycle  in mosquito  cells, possibly due to the lower temperature at which these cells are grown in vitro and in vivo.  Alternatively, it  is possible  that CCR1 also  functions predominantly as an RNA sequence element  in C6/36 cells and that  the  5’  and  3’  flanking  sequences  are more  important  irrespective  of  the  predicted  secondary structure.   Infection of C6/36 cells confirms the observed phenotypes obtained with transfected C6/36 cells though the  infection of BHK cells raises the possibility of revertant viruses.   Alternatively,  it  is possible that mutant  viruses  derived  from  transfected  BHK  cells may  contain  defects  that  result  in  low  viral infectivity  at  early  timepoints  post‐infection  and  that  these  effects  can  be  compensated  over  time.  However, whether the increased viral titer in the “center mut” and “seq mut v2” constructs over time in infected BHK cells  is mediated by revertant viruses  is addressed  in Chapter 4.   The ability  to generate mutant viruses in BHK cells provided a unique opportunity to determine whether the defects observed in  the  in  vitro  experiments  reflect  the  complexity  of  an  in  vivo  infection  by  inoculating  Ae.  aegypti mosquitoes.   Many  flaviviruses  exhibit  a  high  degree  of  specificity  in  their  ability  to  infect  and  be transmitted by arthropod species (36, 50); however, infected mosquitoes remain infectious for life (1‐2 weeks  to ~174 days)  (19).   We show  that mutating CCR1 not only affects viral  replication  in vitro but significantly lowers viral replication in the predominant DENV mosquito vector Ae. aegypti, providing an in vivo model to support our in vitro data.  Sustained flavivirus transmission in arthropods requires that sufficient  amounts  of  virus  are  ingested  to  establish  an  infection  in  the  epithelial  cells  lining  the mesenteron (midgut) (49).  To spread to secondary amplification tissues such as the brain, fat body and salivary glands of  the  infected mosquito,  the virus must  then escape  from  the midgut epithelium  into the body cavity (49).   Changes to the 5’ and 3’ flanking sequences of CCR1 significantly decreased viral dissemination  to  the  salivary  glands, with  a  2.6‐fold  decrease  in  the  amount  of  virus  that  could  be potentially  transmitted by  the mosquito vector as  compared  to  the WT by day 10.   This  implies  that altering CCR1 could have dramatic effects on viral transmission by the arthropod vector.  While the data indicate  that  the mutant  viruses do not  replicate  at  the  same  rate  as  the WT once within  the  body cavity, it would be of interest to investigate how well the mutant viruses can disseminate in Ae. aegypti mosquitoes when they are infected orally and virus must infect and escape from the midgut.   Unlike other known RNA elements  in DENV, CCR1 has been  shown  to have no effect on viral translation, RNA synthesis or stability  in BHK cells.   While not shown,  it  is unlikely that RNA transcripts that are stable in one cell line would be unstable in another, so it is improbable that the effects on viral titer and infectious particle production observed in C6/36 cells are due to RNA instability.  Altering either the primary nucleotide sequence or  the predicted secondary structure of CCR1 had no effect on viral translation and RNA synthesis in BHK cells.  Due to technical difficulties with the transfection of replicons in C6/36 cells, the entire series of mutant replicons could not be completed  in this cell  line.   However, preliminary evidence suggests that changes to the predicted secondary structure of CCR1  (“5’/3’  flank mut”) have no effect on viral translation and RNA synthesis  in C6/36 cells despite decreasing viral titer and  increasing  the  particle:pfu  ratio  in  this  cell  line.    Thus,  it  is  likely  that  viral  translation  and  RNA synthesis are not regulated by CCR1 in C6/36 cells, similar to what was observed in BHK cells.   It  is  possible  that  CCR1  could  serve  as  a  putative  assembly  signal  for DENV;  for  instance,  in retroviruses  (4, 33) and SIN  (18, 37, 57),  the packaging  signal  is  located close  to or within  the coding region.   Since  infection of BHK and C6/36 cells confirmed the results obtained by transfection of these 

66  

cells, it is unlikely that early stages of the viral life cycle, such as entry and uncoating, are influenced by changes in CCR1 in either cell line.  Likewise, CCR1 does not regulate early steps in the replication cycle such as viral translation and RNA synthesis.   We have shown that altering CCR1 results  in an  increased particle:pfu ratio in BHK and C6/36 cells, demonstrating a defect in production of infectious virions.  The effect  of  CCR1  on  the  particle:pfu  ratio  in  C6/36  cells  can  be  observed  in  both  transfections  and infections, and normalizing the infections by G‐Eq rather than by pfu magnifies the differences observed in both cell  lines.   Furthermore,  the same amount of vRNA  is released  from WT and CCR1 mutants  in BHK and C6/36 cells, but much less of the CCR1 mutants are in the form of infectious particles.  Lastly, the  amount  of  extracellular  infectious  virus  is  equivalent  or  greater  to  the  amount  of  intracellular infectious virus  in  the WT and all CCR1 mutants,  indicating  that mutating CCR1 does not  result  in  the retention of virus  inside the cell.   Together, the data suggests that CCR1 plays a role  in viral assembly and/or virion processing/maturation.   Flavivirus  infectious  particle  production  is  achieved  during  maturation  in  the  trans  Golgi compartment, where the prM/M protein is cleaved into the mature M form.  Although inhibiting prM/M cleavage does not impair virion release, studies on prM‐containing particles suggest that furin cleavage (53, 59) or a major structural alteration in prM/M (12) is required to generate highly infectious virus.  It would be of interest to determine whether the observed effects of mutations to the CCR1 sequence on noninfectious  particle  production  could  result  from  a  lack  of  prM/M  cleavage  due  to  decreased accessibility of prM/M to furin during virion maturation, which could explain the overproduction of non‐infectious virions in both cell lines tested.   Many  studies using electron microscopy have  shown  that unlike alphaviruses, neither mature (32, 63) nor immature (63) flavivirus particles appear to have a capsid structure surrounding the vRNA, resulting from a lack of icosahedral symmetry in the nucleocapsid core (64).  The fact that there is little specific interaction between the glycoprotein ectodomain and the capsid protein molecules implies that flavivirus particles are assembled as  they acquire  their envelopes  from  the ER membrane  (62, 63).    If CCR1 acts as an assembly signal, CCR1 mutations may lead to disruptions in the nuclocapsid core, which nonetheless permit  low  levels of  viral particle production  since  flaviviruses do not have  a  structured symmetrical nucleocapsid core but  increase the number of assembled noninfectious particles released (64).  Unfortunately, direct testing of CCR1 interaction with the DENV capsid protein by electrophoretic mobility shift assays is complicated by the fact that capsid associates nonspecifically with RNA and also spontaneously forms protein aggregates  in vitro.   Nonetheless, sequence changes to CCR1 may render more viral particles noninfectious but not necessarily halt the assembly of infectious viral particles given that the nucleocapsid core does not have a defined structure.   Here we have identified a novel RNA sequence element located within the DENV capsid‐coding sequence,  which  functions  in  a  sequence‐dependent,  cell‐type  specific  manner  to  regulate  post‐replication production of  infectious viral particles.   Whether CCR1 serves as a putative assembly signal for DENV  is  still  under  investigation  but  studies  such  as  these  provide  evidence  that  RNA  elements located  within  the  coding‐region  can  have  diverse  roles  in  regulating  the  viral  life  cycle  in  the mammalian host and mosquito vector.  In addition, the identification of mutations affecting late steps in DENV replication will yield valuable  information about the final stages  in the viral  life cycle, which are still poorly understood.   The observation  that CCR1  is more  important  in  the mosquito vector both  in vitro  and  in  vivo  allows  for  exploration  into  viral  replication  strategies  in  mosquito  cells  and  the requirements  for  efficient  viral  transmission.    Finally,  conserved  RNA  elements  in  the  coding‐region could  serve  as  targets  for novel  therapeutics  since development of  resistance  is  constrained by both their RNA regulatory function and amino acid coding capacity. 

A

FIGURES

ADEN1 ACGGGTCGACCGTCTTTCAATATGCTGAAACGCGCGAGAAACCGCGTGTCAACTGTTTCADEN3 ACGGGAAAACCGTCTATCAATATGCTGAAACGCGTGAGAAACCGTGTGTCAACTGGATCADEN2 GCGAAAAACACGCCTTTCAATATGCTGAAACGCGAGAGAAACCGCGTGTCGACTGTGCAADEN4 GTGGTTAGACCACCTTTCAATATGCTGAAACGCGAGAGAAACCGCGTATCAACCCCTCAA

* * ** ****************** ********* ** ** ** *

B

C

TBEV CAAGAGCTGGGGATGGTCAAGAAGGCCATCCTGAAAGGTAAGGGGGGCGGTCCCCCTCGAOmsk CAAGAGCTGGGGATGGCCGGGAAGGCCATTCTGAAAGGAAAGGGGGGCGGTCCCCCTCGAPowassan CAAGAGCTGGGAGTGGTTATGATGACCACTTCTAAAGGAAAGGGGGGCGGTCCCCCTAGG

*********** *** ** * *** ***** ****************** *

SLA

5’3’

cHP

CCR1

D

C‐2

5’3’

SLA

D

cHPCCR1

C‐2CCR1

67

Figure 3.1: A conserved RNA element is identified in the capsid coding‐region of the DENV and TBEVserogroups. Partial alignment of the representative genomes for the (A) DENV serogroup (position 118‐g p g p g ( ) g p (p177, numbering based on DENV2 Thai strain 16681; accession codes DENV1 U88535, DENV2NC_001474, DENV3 AY099336 and DENV4 GU289913) and (B) TBEV serogroup, including the TBEV, Powassanand Omsk hemorrhagic fever viruses (position 121‐180, numbering based on TBEV; accession codes TBEVU27495, Powassan NC_003687 and Omsk AY193805), generated using Clustal W2 (63). Positions of sequencehomology are indicated by asterisks. The 5’ CS is underlined in the DENV serogroup, the initiator AUG isunderlined in the TBEV serogroup, and the putative RNA sequence element CCR1 is boxed. Phylogeneticconsensus structure based on the (C) DENV serogroup (position 1‐274, numbering based off DENV2) and (D)TBEV serogroup (position 1‐302, numbering based on TBEV) as computed by RNAalifold (31), where covarianceis depicted as a colored spectrum ranging from purple (0) to red (1). CCR1 and other known RNAelements, such as the capsid‐coding hairpin (cHP), capsid‐2 (C‐2) and stem‐loop A (SLA), are indicated.

68

WTΔG 21 4

5’/3’ flank mutΔG 16 1

AΔG = ‐21.4

cHP

ΔG = ‐16.1

cHP

B

CCR1 CCR1 IntermediatePlaque (IP)

WTC‐2

C‐2

center mutΔG = ‐21.4

cHP

seq mutΔG = ‐20.4

cHP SmallPlaque (SP)

Extra SmallPlaque (ESP)

CCR1CCR1C‐2 C‐2

8

10

)

WT 5'/3' flank mutcenter mut seq mut5' flank mut seq mut v2

5

6

7

)

WT 5'/3' flank mutcenter mut seq mut5' flank mut seq mut v2

C D

SPSP

0

2

4

6

Log(pfu/mL)

0

1

2

3

4

Log(pfu/mL)

****

*

******

ESP*SP*

*

SP*

*

SP*

*** *

SP*SP*

SP

*SP*

0

72 96 120 144Hours Post‐Transfection

24 48 72Hours Post‐Transfection

69

Figure 3.2: CCR1 acts as an RNA sequence element mediating the viral life cycle in BHK and C6/36 cells. (A)mfold predictions of the mutant constructs designed to disrupt the primary nucleotide sequence (“seq mut”) orf p g p p y q ( q )the predicted secondary structure (“5’/3’ flank mut”). The sequence changes introduced into the predictedloop region of the “seq mut” construct were isolated in the “center mut” without altering the predictedsecondary structure of CCR1 as predicted bymfold (74). The positions of introduced mutations are indicated byboxes. Known RNA elements are indicated (cHP, capsid‐coding hairpin; C‐2, capsid‐2). (B) Plaque phenotype inBHK cells obtained by transfecting BHK or C6/36 cell monolayers with the described mutant constructs(SP, small plaque; IP, intermediate plaque; ESP, extra small plaque). (C) BHK and (D) C6/36 cells weretransfected with IC‐based mutant constructs and assayed for viral titer (pfu/mL) by plaque assay. Viral titerswere normalized to the amount of vRNA present in the cells at the time of harvest as measured by qRT‐PCR.Results were obtained from 4 independent experiments, conducted in duplicate. Plaque phenotypes areindicated for each construct. The limit of detection is shown by a dashed line (1.9 pfu/mL). *p<0.05 relative tothe WT. Error bars indicate standard deviation.

70

7 WT 5'/3' fl k t6 WT 5'/3' fl k

A B

3

4

5

6

7

g(pfu/mL)

WT 5'/3' flank mutcenter mut seq mut5' flank mut seq mut v2

3

4

5

6

g(pfu/mL)

WT 5'/3' flank mutcenter mut seq mut5' flank mut seq mut v2

ESP/SPSP*IP*

** *

SP*

SP

0

1

2

3

72 96 120 144

Log

Hours Post‐Infection

0

1

2

24 48 72

Log

Hours Post‐Infection

*** SP*

** * * *

**

*SP*

C D

5

6

7WT 5'/3' flank mut center mut

4

5

6WT 5'/3' flank mut center mut

1

2

3

4

Log(pfu/mL)

1

2

3

4

Log(pfu/mL)

IP

*

SP

** * *

**

SP*SP*

0

24 48 72Hours Post‐Infection

0

72 96 120 144Hours Post‐Infection

Figure 3.3: Infecting BHK and C6/36 cells with mutant viruses confirms that CCR1 plays a greater role inC6/36 cells. Mutant viruses were harvested from IC‐transfected BHK monolayers and used to infect (A) BHKand (B) C6/36 cell monolayers at an MOI of 0 01 and the viral titer was assessed by plaque assay Theand (B) C6/36 cell monolayers at an MOI of 0.01, and the viral titer was assessed by plaque assay. Theinfections in (C) BHK and (D) C6/36 cells were then normalized to the amount of vRNA (G‐Eq) in the mutantvirus stock, and the viral titer was assessed by plaque assay. Results were obtained from 6 (pfu normalization)or 3 (G‐Eq normalization) independent experiments, performed in duplicate. Plaque phenotypes are indicatedfor each construct (SP, small plaque; IP, intermediate plaque; ESP, extra small plaque). The limit of detection isshown by a dashed line (1.9 pfu/mL). *p<0.05 relative to the WT. Error bars indicate standard deviation.

71

A B

4

5

6

7

8

og(pfu/m

L)

WT 5'/3' flank mutcenter mut seq mutseq mut v2

* 60

80

100

120

nfected

4.5Log(pfu/mL)

** *** **

*

0

1

2

3

Day 5 Day 10 Day 20

Lo

Days Post‐Innoculation

0

20

40

WT 5'/3' flank mut

% In

1.7*Log(pfu/mL)

*

* *

**

ays ost ocu at o

Figure 3.4: CCR1 modulates viral replication and dissemination to the salivary glands in Ae. aegyptimosquitoes. (A) Colonized Panamanian Ae. aegypti mosquitoes were inoculated intrathoracically with mutantviruses and infected mosquitoes were processed at the indicated timepoints to determined viral titer usingplaque assays. (B) Salivary glands from mosquitoes infected with WT and “5’/3’ flank mut” mutant viruses weredissected on day 10 and viral titer was assessed by plaque assay. Results were obtained from 3 independentmosquito infections. The limit of detection is shown by a dashed line (1.4 pfu/mL). *p<0.05 relative to the WT.mosquito infections. The limit of detection is shown by a dashed line (1.4 pfu/mL). p<0.05 relative to the WT.Error bars indicate standard deviation.

72

A B

8

10

12

14

16

Log(rLUC)

WT GVD5'/3' flank mut center mutseq mut 5' flank mut3' flank mut seq mut v2delta mut

** 4

5

6

7

8

og(rLU

C)

WT GVD 5'/3' flank mut

* * *

‐2

0

2

4

6

0 20 40 60 80 100Hours Post‐Transfection

0

1

2

3

0 50 100 150

Lo

Hours Post‐Transfection

* * * *

* * * *

Hours Post Transfection

10

12

14

C

Hours Post Transfection

2

4

6

8

10

Half Life

 (h)

0

2

GVD 5'/3' flank mut center mut seq mut

Figure 3.5: CCR1 has no effect on viral translation, RNA synthesis or RNA stability in BHK and C6/36 cells.Mutations were subcloned into Renilla luciferase replicons, and the mutant replicons were transfected into (A)BHK and (B) C6/36 cell monolayers Cellular lysates were harvested at the indicated timepoints and assayed forBHK and (B) C6/36 cell monolayers. Cellular lysates were harvested at the indicated timepoints and assayed forluciferase activity. An additional mutant, termed “delta,” was created where the entire CCR1 sequence wasremoved from the replicon. *p<0.05 relative to the WT replicon. (C) Mutations were subcloned into a GVDvariant of the IC and transfected into BHK cells. Cells were harvested at 4, 8, 24, 48 and 72 hpt, and theamount of vRNA present was quantified using qRT‐PCR in order to calculate the RNA half‐life in BHK cells.Results were obtained from 3 independent experiments, conducted in duplicate. Error bars indicate standarddeviation.

73

A B

8

10

12

14

Eqin Sup

e)

WT 5'/3' flank mutcenter mut seq mut5' flank mut seq mut v2

8

10

12

14

Eqin Sup

e)

WT 5'/3' flankmutcenter mut seq mutseq mut v2

* **

* * *

0

2

4

6

72 96 120 144

Log(G‐E

H P t T f ti

0

2

4

6

24 48 72

Log(G‐E

H P t T f ti

* * **

8

9

10 WT 5'/3' flank mutcenter mut seq mut5' flank mut seq mut v2

Hours Post‐Transfection

8

9

10 WT 5'/3' flankmutcenter mut seq mutmut seq v2

C

Hours Post‐Transfection

D

2

3

4

5

6

7

Log(G‐Eq/pfu)

2

3

4

5

6

7

Log(G‐Eq/pfu) * *

**

*** **

*

** **

* *

*****

0

1

72 96 120 144Hours Post‐Transfection

0

1

24 48 72Hours Post‐Transfection

74

E F

4

5

6

7

8

G‐Eq/pfu)

WT 5'/3' flank mutcenter mut seq mut5' flank mut seq mut v2

4

5

6

7

8

(G‐Eq/pfu)

WT 5'/3' flank mutcenter mut seq mut5' flank mut

****** *

*** **

**

0

1

2

3

24 48 72

Log(

Hours Post‐Infection

0

1

2

3

72 96 120 144

Log(

Hours Post‐Infection

*

G H

ou s ost ect o ou s ost ect o

5

6

7WT 5'/3' flank mut center mut

6

7

8WT 5'/3' flank mut center mut

1

2

3

4

5

Log(G‐Eq/pfu)

1

2

3

4

5

Log(G‐Eq/pfu) * * * **

* **

0

24 48 72Hours Post‐Infection

0

1

72 96 120 144Hours Post‐Infection

Figure 3.6: CCR1 mediates infectious viral particle production. vRNA found in virus‐containing supernatantsfrom IC‐transfected (A) BHK or (B) C6/36 cell monolayers was extracted at the indicated timepoints andmeasured by qRT‐PCR. Particle:pfu ratios were generated by comparing vRNA concentrations (G‐Eq/mL) in thesupernatant to the viral titers (pfu/mL) obtained at the same timepoint in IC‐transfected (C) BHK or (D) C6/36cell monolayers. Particle:pfu ratios were also determined for the pfu‐normalized infections in (E) BHK and (F)C6/36 cells as well as the G‐Eq‐normalized infections in (G) BHK and (H) C6/36 cells. Results were obtainedfrom 4 (transfections), 6 (pfu‐normalized infections) or 3 (G‐Eq‐normalized infections) independentexperiments, conducted in duplicate. *p<0.05 relative to the WT. Error bars indicate standard deviation.

75

A B

1.5

2

2.5

3

xtra vs. In

tra pfu/mL)

WT 5'/3' flank mutcenter mut seq mut5' flank mut seq mut v2

1.5

2

2.5

3

xtra vs. In

tra pfu/mL)

WT 5'/3' flank mutcenter mut seq mut5' flank mut seq mut v2

0

0.5

1

24 48 72

Fold Cha

nge (E

Hours Post‐Transfection

0

0.5

1

72 96 120 144

Fold Cha

nge (Ex

Hours Post‐Transfection

** * * ** * * * *

Hours Post Transfection

Figure 3.7: CCR1 has no effect on virion retention in BHK and C6/36 cells. The amount of extracellularinfectious virus was compared to the amount of intracellular infectious virus from IC‐transfected (A) BHK and(B) C6/36 cell monolayers at the indicated timepoints. A ratio of extra‐ to intracellular virus greater than 1 wasused to indicate that more virus was released from the cells, whereas a ratio of less than 1 implies that moreinfectious virus was located within the cell. Results were obtained from 3 independentexperiments, conducted in duplicate. A ratio of extra‐ to intracellular virus of 1 is indicated by a dashed line.experiments, conducted in duplicate. A ratio of extra to intracellular virus of 1 is indicated by a dashed line.*p<0.05 relative to the WT. Error bars indicate standard deviation.

76

77  

TABLES   Table 3.1:  SOE‐PCR cloning primers for the generation of DNA mutant constructs.  For cloning into the IC,  the mutant  forward  (FWD)  primers  (listed) were  used  in  conjunction with  the WT  reverse  (REV) primer  to  generate  the  “A”  fragment.    The mutant REV  primers  are  the  reverse  complement  of  the mutant FWD primer, and used  in conjunction with the WT FWD primer to generate the “B” fragment.  The “A” and “B” fragments were spliced together using the WT FWD and REV primers (“C” fragment), which was used to generate the  inserts.   Codons where mutations were  introduced are  in  lower case.  WT flanking primers are underlined. Backbone  Construct  Primer Sequence 

IC 

F1 (FWD)  5’-AACGACAGGAGCACGATCATGC-3’

5’/3’ flank mut (FWD)  5’-GCCTTTCAATATGCTGAAAcgaGAGAGAAACcgaGTGTCGACTGTGCAACAGC-3’

seq mut (FWD)  5’-GCCTTTCAATATGCTGAAAcgagaacggaatcgtGTGTCGACTGTGCAACAG-3’

5’ flank mut (FWD)  5’-CCTTTCAATATGCTGAAAcgaGAGAGAAACCGCGTGTCG-3’

center mut (FWD)  5’-CTTTCAATATGCTGAAACGCgaacggAACCGCGTGTCGACTGTG-3’

3’ flank mut (FWD)  5’-CAATATGCTGAAACGCGAGAGAaatcgtGTGTCGACTGTGCAACAG-3’

seq mut v2 (FWD)  5’-CACGCCTTTCAATATGCTGaagagggaacgtaatagaGTGTCGACTGTGCAACAG-3’

ESphIR (REV)  5’-TTTCCGACTGCATGCTCTTC-3’

Replicon 

SacIT7RepF (FWD)  5’-CGAGCTCGCGGCCGCAAATTTAATACGACTCAC-3’

MluICrepR (REV)  5’-CATACGCGTGGACACGCGGTTTCTCTCGCGTTTC-3’

5’/3’ flank mutR (REV)  5’-CATACGCGTGGACACtcgGTTTCTCTCtcgTTTC-3’

seq mutR (REV)  5’-CATACGCGTGGACACacgattccgttctcgTTTC-3’

5’ flank mutR (REV)  5’-CATACGCGTGGACACGCGGTTTCTCTCtcgTTTC-3’

center mutR (REV)  5’-CATACGCGTGGACACGCGGTTccgttcGCGTTTC-3’

3’ flank mutR (REV)  5’-CATACGCGTGGACACacgattTCTCTCGCGTTTC-3’

seq mut v2R (REV)  5’-CATACGCGTGGACACtctattacgttccctcttC-3’

deltaR (REV)  5’-CATacgcgtgcgtTTCAGCATATTGAAAGGCGTG-3’

  Table 3.2:  qRT‐PCR primer/probe set directed against the viral NS5 gene. 

Construct  Primer or Probe Sequence DV9944F (FWD)  5’-ACAAGTCGAACAACCTGGTCCAT-3’ DV10121R (REV)  5’-GCCGCACCATTGGTCTTCTC-3’

DV‐NS5‐F/T(Probe)  5’-FAM-TGGGATTTCCTCCCATGATTCCACTGG-TAMRA-3’

  Table 3.3:  Sequence of WT CCR1 and CCR1 mutants.  Mutations are in lower case. 

Construct  Nucleotide Sequence WT  5’ – AAACGCGAGAGAAACCGCGTG - 3’

5’/3’ flank mut  5’ - AAACGaGAGAGAAACCGaGTG - 3’ seq mut  5’ - AAACGaGAacGgAAtCGtGTG - 3’

5’ flank mut  5’ - AAACGaGAGAGAAACCGCGTG - 3’ center mut  5’ - AAACGCGAacGgAACCGCGTG - 3’ 3’ flank mut  5’ - AAACGCGAGAGAAAtCGtGTG - 3’ seq mut v2  5’ - AAgaGgGAacGtAAtaGaGTG - 3’

 

78  

REFERENCES   1.  Alvarez, D. E., A. L. De Lella Ezcurra, S. Fucito, and A. V. Gamarnik. 2005. Role of RNA structures 

present  at  the  3'UTR  of  dengue  virus  on  translation,  RNA  synthesis,  and  viral  replication. Virology 339:200‐212  

2.  Alvarez, D. E., M. F. Lodeiro, S. J. Ludueña, L. I. Pietrasanta, and A. G. Gamarnik. 2005. Long‐range RNA‐RNA interactions circularize the dengue virus genome. J Virol 79:6631‐6643. 

3.  Alvisi,  G.,  V. Madan,  and  R.  Bartenshlager.  2011.  Hepatitis  C  virus  and  host  cell  lipids:  an intimate connection. RNA Biol 8:258‐269. 

4.  Chadwick,  D.  R.,  and  A. M.  Lever.  2000.  Antisense  RNA  sequences  targeting  the  5'  leader packaging  signal  region  of  human  immunodeficiency  virus  type‐1  inhibits  viral  replication  at post‐transcriptional stages of the life cycle. Gene Ther 7:1362‐1368. 

5.  Chiu, W. W., R. M. Kinney, and T. W. Dreher. 2005. Control of  translation by  the 5'‐ and 3'‐terminal regions of the dengue virus genome. J Virol 79:8303‐8315. 

6.  Chua, J. J., M. M. Ng, and V. T. Chow. 2004. The non‐structural 3 (NS3) protein of dengue virus type 2 interacts with human nuclear receptor binding protein and is associated with alterations in membrane structure. Virus Res 102:151‐163. 

7.  Clyde, K.,  J. Barrera, and E. Harris. 2008. The capsid‐coding  region hairpin element  (cHP)  is a critical determinant of dengue virus and West Nile virus RNA synthesis. . Virology 379:314‐323. 

8.  Clyde, K., and E. Harris. 2006. RNA  secondary  structure  in  the  coding  region of dengue  virus type  2  directs  translation  start  codon  selection  and  is  required  for  viral  replication.  J  Virol 80:2170‐2182. 

9.  Deubel V., and D.  J. P. 1981. Morphogenesis of  yellow  fever  virus  in Aedes  aegypti  cultured cells. I. Isolation of different cellular clones and the study of their susceptibility to infection with the virus. Am J Trop Med Hyg 30:1060‐1070. 

10.  Diamond, M. S., D. Edgil, T. G. Roberts, B. Lu, and E. Harris. 2000. Infection of human cells by dengue virus is modulated by different cell types and viral strains. J Virol 74:7814‐7823. 

11.  Diamond, M. S., and E. Harris. 2001.  Interferon  inhibits dengue virus  infection by preventing translation of viral RNA through a PKR‐independent mechanism. Virology 289:297‐311. 

12.  Elshuber,  S.,  and  C. W. Mandl.  2005.  Resuscitating mutations  in  a  furin  cleavage‐deficient mutant of the flavivirus tick‐borne encephalitis virus. J Virol 79:11813‐11823. 

13.  Fayzulin, R., and I. Frolov. 2004. Changes of the secondary structure of the 5' end of the Sindbis virus  genome  inhibit  virus  growth  in  mosquito  cells  and  lead  to  accumulation  of  adaptive mutations. J Virol 78:4953‐4964. 

14.  Filomatori, C. V., M. F. Lodeiro, D. E. Alvarez, M. M. Samsa, L. Pietrasanta, and A. V. Gamarnik. 2006. A 5' RNA element promotes dengue virus RNA synthesis on a circular genome. Genes Dev 20:2238‐2249. 

15.  Friebe, P., and E. Harris. 2010.  Interplay of RNA elements  in  the dengue virus 5' and 3' ends required for viral RNA replication. J Virol 84:6103‐6118. 

16.  Friebe, P., P. Y. Shi, and E. Harris. 2011. The 5' and 3' downstream AUG  region elements are required for mosquito‐borne flavivirus RNA replication. J Virol 85:1900‐1905. 

17.  Frolov, I., R. Hardy, and C. M. Rice. 2001. Cis‐acting RNA elements at the 5' end of Sindbis virus genome RNA regulate minus‐ and plus‐strand RNA synthesis. RNA 7:1638‐1651. 

18.  Frolova, E., I. Frolov, and S. Schlesinger. 1997. Packaging signals in alphaviruses. J Virol 71:248‐258. 

19.  Gubler, D. J., and G. Kuno. 1997. Dengue and dengue hemorrhagic fever. CAB International. 

79  

20.  Hase, T., P. L. Summers, K. H. Eckels, and W. B. Baze. 1987. An electron and  immunoelectron microscopic  study  of  dengue‐2  virus  infection  of  cultured mosquito  cells: maturation  events. Arch Virol 92:273‐291. 

21.  Hase, T., P. L. Summers, K. H. Eckels, and W. B. Baze. 1987. Maturation process of  Japanese encephalitis  virus  in  cultured mosquito  cells  in vitro and mouse brain  cells  in vivo. Arch Virol 96:135‐151. 

22.  Heinz, F., G. Auer, K. Stiasny, H. Holzmann, C. Mandl, F. Guirakhoo, and C. Kunz. 1994. The interactions of  the  flavivirus envelope proteins:  implications  for  virus entry  and  release. Arch Virol 9(S):339‐348. 

23.  Helt, A.‐M., and E. Harris. 2005. S‐phase‐dependent enhancement of dengue virus 2 replication in mosquito cells, but not in human cells. JVirol 79:7291‐7299. 

24.  Hofacker,  I. L., M. Fekete, and P. F. Stadler. 2002. Secondary structure prediction  for aligned RNA sequences. J Mol Biol 319:1059‐1066. 

25.  Hofacker, I. L., W. Fontana, P. F. Stadler, S. Bonhoeffer, M. Tacker, and P. Schuster. 1994. Fast folding and comparison of RNA secondary structures. Monatsh Chem 125:167‐188. 

26.  Holden,  K.  L.,  D.  Stein,  T.  C.  Pierson,  A.  Ahmed,  K.  Clyde,  P.  Iverson,  and  E.  Harris.  2006. Inhibition of dengue virus translation and RNA synthesis by a morpholino oligomer to the top of the 3’ stem‐loop structure. Virology 344:439‐452. 

27.  Ishak, R., D. G. Tovey, and C. R. Howard. 1988. Morphogenesis of  yellow  fever  virus 17D  in infected cell cultures. J Gen Virol 69:325‐335. 

28.  Jones, C. T., C. G. Patkar, and R.  J. Kuhn. 2005. Construction and application of yellow  fever virus replicons. Virology 331:247‐259. 

29.  Khromykh, A. A., A. N. Varnavski, and E. G. Westaway. 1998. Encapsidation of  the  flavivirus kunjin  replicon  RNA  by  using  a  complementation  system  providing  Kunjin  virus  structural proteins in trans. J Virol 72:5967‐5977. 

30.  Kinney, R. M., S. Butrapet, G. J. Chang, K. R. Tsuchiya, J. T. Roehrig, N. Bhamarapravati, and D. J. Gubler. 1997. Construction of  infectious cDNA clones for dengue 2 virus strain 16681 and  its attenuated vaccine derivative, strain PDK‐53. Virology 230:300‐308. 

31.  Ko, K. K., A. Igarashi, and K. Fukai. 1979. Electron microscopic observations on Aedes albopictus cells infected with dengue viruses. Arch Virol 62:41‐52. 

32.  Kuhn, R.  J., W.  Zhang, M. G. Rossmann,  S. V. Pletnev,  J. Corver,  E.  Lenches, C.  T.  Jones,  S. Mukhopadhyay, P. R. Chipman, E. G. Strauss, T. S. Baker, and J. H. Strauss. 2002. Structure of dengue virus: implications for flavivirus organization, maturation, and fusion. Cell 108:717‐25. 

33.  Kuno, G. 1997. Factors  influencing  the  tranmsission of dengue viruses.  In D.  J. Gubler and G. Kuno (ed.), Dengue and Dengue Hemorrhagic Fever. CAB International, New York. 

34.  Laue, T., P. Emmerich, and H. Schmitz. 1999. Detection of dengue virus RNA  in patients after primary or secondary dengue infection by using the TaqMan automated amplification system. J Clin Microbiol 37:2543‐2547. 

35.  Leary, K., and C. D. Blair. 1980. Sequential events in the morphogenesis of japanese encephalitis virus. J Ultrastruct Res 72:123‐129. 

36.  Lindenbach,  B.  D.,  H.  J.  Thiel,  and  C.  M.  Rice.  2007.  Flaviviridae:  the  viruses  and  their replication,  p.  1101‐1152.  In  D.  M.  Knipe  and  P.  M.  Howley  (ed.),  Fields  Virology,  5th  ed. Lippincott‐Raven Publishers, Philadelphia. 

37.  Linger,  B.  R.,  L.  Kunovska,  R.  J.  Kuhn,  and  B.  L.  Golden.  2004.  Sindbis  virus  nucleocapsid assembly: RNA folding promotes capsid protein dimerization. RNA 10:128‐138. 

38.  Liu, Y., E. Wimmer, and A. V. Paul. 2009. Cis‐acting RNA elements  in human and animal plus‐strand RNA viruses. Biochim Biophys Acta 1789:495‐517. 

80  

39.  Lo, M., M. Tilgner, K. A. Bernard, and P.‐Y. Shi. 2003. Functional analysis of mosquito‐borne flavivirus conserved sequence elements within 3’ untranslated region of West Nile Virus by use of a reporting replicon that differentiates between viral translation and RNA replication. J Virol 77:10004‐10014. 

40.  Matsumara, T., K. Shiraki, T. Sashikata, and S. Hotta. 1977. Morphogenesis of dengue‐1 virus in cultures of a human leukemic leukocyte line (J‐111). Microbiol Immunol 21:329‐334. 

41.  Miller, W. A., and K. A. White. 2006. Long‐distance RNA‐RNA  interactions  in plant virus gene expression and replication. Annu Rev Phytopathol 44:447‐467. 

42.  Murphy, F. A. 1980. Togavirus morphology and morphogenesis, p. 241‐316. In R. W. Schlesinger (ed.), The togaviruses: biology, structure, replication. New York: Academic. 

43.  Nelson, S., C. A. Jost, Q. Xu, J. Ess, J. E. Martin, T. Oliphant, S. S. Whitehead, A. P. Durbin, B. S. Graham, M.  S. Diamond,  and  T.  C.  Pierson.  2008. Maturation  of West Nile  virus modulates sensitivity to antibody‐mediated neutralization. PLoS Pathog 4:e1000060. 

44.  Ng, M. L. 1987. Ultrastructural studies of Kunjin virus‐infected Aedes albopictus cells. J Gen Virol 68:577‐582. 

45.  Niesters, H. G., and J. H. Strauss. 1990. Mutagenesis of the conserved 51‐nucleotide region of Sindbis virus. J Virol 64:1639‐1647. 

46.  Nowak, T., P. M. Farber, G. Wengler, and G. Wengler. 1989. Analyses of the terminal sequences of West Nile virus structural proteins and of the  in vitro translation of these proteins allow the proposal of a complete scheme of the proteolytic cleavages involved in their synthesis. Virology 169:365‐376. 

47.  Ohyama,  A.,  T.  Ito,  E.  Tanimura,  S.  C.  Huang,  and  J.  Hsue.  1977.  Electron  microscopic observation of the budding maturation of group B arboviruses. Microbiol Immunol 21:535‐538. 

48.  Payne, A.  F.,  I. Binduga‐Gajewska,  E. B. Kauffman,  and  L. D.  Kramer.  2006. Quantitation  of flaviviruses by fluorescent focus assay. J Virol Meth 134:183‐189. 

49.  Rosen, L., and D. Gubler. 1974. The use of mosquitoes to detect and propagate dengue viruses. Am J Trop Med Hyg 23:1153‐1160. 

50.  Rosen,  L.,  L.  E.  Roseboom, D.  J. Gubler,  J.  C.  Lien,  and  B. N.  Chaniotis.  1985.  Comparative susceptibility of mosquito species and strains to oral and parenteral  infection with dengue and Japanese encephalitis viruses. Am J Trop Med Hyg 34:603‐615. 

51.  Sriurairatna, S., and N. Bhamarapravati. 1977. Replication of dengue‐2 virus in Aedes albopictus mosquitoes. An electron microscopic study. Am J Trop Med Hyg 26:1199‐1205. 

52.  Sriurairatna,  S., N.  Bhamarapravati,  and O.  Phalavadhtana.  1973. Dengue  virus  infection  of mice: Morphology and morphogenesis of dengue type 2 virus in suckling mouse neurons. Infect Immun 8:1017‐1029. 

53.  Stadler, K., S. L. Allison, J. Schalich, and F. X. Heinz. 1997. Proteolytic activation of tick‐borne encephalitis virus by furin. J Virol 71:8475–8481. 

54.  Thompson,  J.  D.,  T.  J.  Gibson,  F.  Plewniak,  F.  Jeanmougin,  and  D.  G.  Higgins.  1997.  The CLUSTAL_X Windows  interface:  flexible  strategies  for multiple  sequence  alignment  aided  by quality analysis tools. Nucleic Acids Res 25:4876‐4882. 

55.  Thurner,  C.,  C. Witwer,  I.  L.  Hofacker,  and  P.  F.  Stadler.  2004.  Conserved  RNA  secondary structures in Flaviviridae genomes. J Gen Virol 85:1113‐1124. 

56.  Tilgner, M., and P. Y. Shi. 2004. Structure and function of the 3' terminal six nucleotides of the West Nile virus genome in viral replication. J Virol 78:8159‐8171. 

57.  Weiss, B., U. Geigenmuller‐Gnirke, and S. Schlesinger. 1994. Interactions between Sindbis virus RNAs  and  a  68  amino  acid  derivative  of  the  viral  capsid  protein  further  defines  the  capsid binding site. Nucleic Acids Res 22:780‐786. 

81  

58.  Welsch, S., S. Miller, I. Romero‐Brey, A. Merz, C. K. Bleck, P. Walther, S. D. Fuller, C. Antony, J. Krijnse‐Locker, and R. Bartenschlager. 2009. Composition and  three‐dimensional architecture of the dengue virus replication and assembly sites. Cell Host Microbe 5:365‐375. 

59.  Wengler, G., and G. Wengler. 1989. Cell‐associated West Nile flavivirus is covered with E+pre‐M protein heterodimers which are destroyed and reorganized by proteolytic cleavage during virus release. J Virol 63:2521‐2526. 

60.  Yu,  I. M., H. A. Holdaway,  P. R.  Chipman, R.  J. Kuhn, M. G. Rossmann,  and  J.  Chen.  2009. Association of  the pr peptides with dengue virus at acidic pH blocks membrane  fusion.  J Virol 83:12101‐12107. 

61.  Zeng, L., B. Falgout, and L. Markoff. 1998. Identification of specific nucleotide sequences within conserved  3'‐SL  in  the dengue  type  2  virus  genome  required  for  replication.  J Virol  72:7510‐7522. 

62.  Zhang, W., P. R. Chipman, J. Corver, P. R. Jonhnson, Y. Zhang, S. Mukhopadhyay, T. S. Baker, J. H. Strauss, M. G. Rossmann, and R. J. Kuhn. 2003. Visualization of membrane protein domains by cryo‐electron microscopy of dengue virus. Nat Struct Biol 10:907‐912. 

63.  Zhang, Y., J. Corver, P. R. Chipman, W. Zhang, S. V. Pletnev, D. Sedlak, T. S. Baker, J. H. Straus, R.  J.  Kuhn,  and M. G.  Rossmann.  2003.  Structures  of  immature  flavivirus  particles.  EMBO  J 22:2604‐2613. 

64.  Zhang,  Y.,  V.  A.  Kostyuchenko,  and  M.  G.  Rossmann.  2007.  Structural  analysis  of  viral nucleocapsids by subtraction of partial projections. J Struct Biol 157:356‐364. 

65.  Zuker, M. 2003. Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res 31:3406‐3415. 

66.  Zybert,  I.  A.,  H.  van  der  Ende‐Metselaar,  J.  Wilschut,  and  J.  M.  Smit.  2008.  Functional importance of dengue virus maturation:  infectious properties of  immature virions.  J Gen Virol 89:3047‐3051. 

  

82  

CHAPTER FOUR 

“TYING UP LOOSE ENDS” 

83  

“TYING UP LOOSE ENDS”  Identification of novel coding‐region RNA elements.   The goal of this study was to  identify additional RNA elements located in the DENV coding‐region that regulate the intracellular viral life cycle.  Based on sequence  and  secondary  structure  conservation  within  the  DENV  serogroup,  three  candidate  RNA structural elements (C‐1 C‐2 and NS5‐2) and three candidate RNA sequence elements (prMCR1, NS5CR1 and CCR1) were identified and characterized for their effect on viral replication (discussed in Chapter 2).  The initial screens eliminated all but one candidate RNA sequence element, the conserved capsid‐coding region 1 (CCR1), which was shown to regulate the viral life cycle in both baby hamster kidney (BHK) and Aedes  albopictus  mosquito  (C6/36)  cells,  though  its  effects  are  more  dramatic  in  mosquito  cells (discussed in Chapter 3).  Altering the primary nucleotide sequence of CCR1 not only reduced viral titer in mosquito cells, but additionally, CCR1 mutant viruses did not replicate to wild‐type (WT) levels in Ae. aegypti mosquitoes.   Decreased viral replication  in the mosquito body cavity  led to a reduction  in viral dissemination  to  the  salivary  glands, which  could  have  dramatic  implications  for  viral  transmission.  Furthermore, CCR1 was shown to mediate post‐RNA replication events, possibly acting as a sequence‐dependent assembly signal, which has yet to be identified in DENV.   C‐1  is a putative RNA  structure element  that was eliminated as a  candidate during  the  initial screens when it displayed WT replication in BHK cells but had inconclusive results in infected C6/36 cells.  The “C1sil mut” and “C1sil cmut” constructs had distinct structural changes  to C‐1, but varied  in  their ability  to  replicate  in  infected C6/36 cells,  implying  that C‐1 might have a sequence‐dependent  rather than a  structure‐dependent effect on  the viral  life  cycle  in  this  cell  line.   However, C‐1 only has 44% sequence  homology  (~52%  background,  data  not  shown)  in  the DENV  serogroup  and  32%  sequence homology (~22% background, data not shown) in the Japanese encephalitis virus (JEV) serogroup, which indicates that C‐1 is not likely to be a candidate RNA sequence element.  However, given the preliminary evidence  in  infected C6/36 cells, additional silent point mutations could be  introduced so as to disrupt the entire C‐1 nucleotide sequence irrespective of structural changes and the resulting variants screened for their effects on the viral life cycle in C6/36 cells.   As evidenced  in other RNA  viruses  (2),  it  is possible  that  additional  regulatory RNA elements might  exist  within  the  DENV  coding‐region.    In  the  search  for  additional  candidate  RNA  structural elements,  the phylogenetically conserved  structure predictions  for  the NS5‐coding  sequence were  re‐examined.    In  brief,  the  DENV,  JEV  and  tick‐borne  encephalitis  virus  (TBEV)  serogroup  sequence alignments that were generated using Clustal W2 (3) were modified so that they would  include a  large portion  of  the  NS5‐coding  region,  which  was  subsequently  analyzed  using  RNAalifold  (1).    In  this investigation,  two  additional  candidate  RNA  structure  elements  were  identified  in  the  NS5‐coding region, NS5‐1 and NS5‐2, which were shown to be conserved in the DENV (Fig. 4.1A), JEV (Fig. 4.1B) and TBEV (Fig. 4.1C) serogroups; however, NS5‐3 is not as well conserved as NS5‐1 in the JEV serogroup (Fig. 4.1B).    Continuation  of  this  study  should  include  introducing  silent  point mutations  that  disrupt  the predicted  secondary  structure  of  NS5‐1  and  NS5‐3,  and  the  resulting mutant  constructs  should  be screened  for  viral  infectivity  in  mammalian  and  mosquito  cells.    The  identification  of  additional candidate RNA elements  is an  important  first  step  in  the characterization of  the  intracellular viral  life cycle since there are still a number of stages that are not well understood.  Sequencing  vRNA  from  infected BHK  cells.   As  discussed  in  Chapter  3,  the  infections  in  C6/36  cells confirmed  that CCR1 has  a more drastic phenotype  in  this  cell  line while  the  infections  in BHK  cells raised the possibility that revertant viruses were accumulating after serial passage.  Briefly, to determine if CCR1 had an effect on early stages of the viral life cycle, virus‐containing supernatants were collected from transfected BHK cells, concentrated, and used to  infect BHK and C6/36 cells at an MOI of 0.01 or with ~2.5 genome equivalents  (G‐Eq) per cell.   Virus‐containing supernatants from these  infected cells 

84  

were then assessed for viral infectivity by plaque assay.  When the infections were normalized by pfu, all the mutant viruses displayed a significantly lower viral titer than the WT at 24 hours post‐infection (hpi) in BHK cells, the “5’/3’ flank mut,” “5’ flank mut” and “seq mut v2” mutant viruses acquired WT levels by the end of the assay (Fig. 3.3A).  Similar results were observed when the infections were normalized by G‐Eq, where at early  timepoints post‐infection  the “5’/3’  flank mut” and “center mut” constructs had significantly  lower viral  titers  compared  to  the WT  though by  the end of  the assay  they attained WT levels (Fig. 3.3C).  The fact that the mutant viruses accumulated to WT levels after serial passage in BHK cells regardless of whether the infections were normalized by pfu or G‐Eq could indicate the buildup of viral revertants.  The likelihood that the mutant viruses undergo enough selective pressure to generate reversions  in BHK cells  is supported by the observation that the “center mut” mutant virus displays an intermediate plaque phenotype while in previous assays it demonstrated a small plaque phenotype (Fig. 3.2D).   The possibility that viral revertants caused the accumulation of mutant viruses  in  infected BHK cells was confirmed with the observation that the mutant viruses also displayed a WT particle:pfu ratio in BHK cells at all timepoints tested (Fig. 3.6E and 3.6G).   In  an  effort  to  determine whether  viral  revertants were  the  cause  for  the  accumulation  of mutant viruses in infected BHK cells, viral RNA (vRNA) was harvested from virus‐containing supernatants produced from infected BHK cells at 48 hpi.  The vRNA was subjected to a reverse transcriptase (RT)‐PCR reaction using primers designed to amplify a cDNA fragment that contained the first 602 nucleotides of the DENV genome, which was then sequenced directly.  Hence, each sequence in Table 4.1 represents a single  sequencing  reaction  derived  from  the  amplified  cDNA  fragments.    Preliminary  tests  were conducted on the “CCR1orig mut” mutant viruses that were discussed in Chapter 2.  The vast majority of the sequencing  results were  identical  to  the  input virus  (“CCR1orig mut”),  though  in one  instance  the sequence was  that  of  the WT  virus  (“WT”)  or  half  of  the mutations  reverted  to  the WT  sequence (“CCR1orig mut like,” Table 4.1).  In another instance, the sequence still contained the original mutations but  two additional  changes were  found  in  the 3’  flanking  sequence  (“CCR1orig mut plus,” Table 4.1).  The mixture of mutant and WT sequences in the cDNA amplicons implies that there is sufficient pressure in  infected  BHK  cells  to  cause minor  sequence  changes  in  the  CCR1  sequence  as  early  as  48  hpi.  However,  this  preliminary method  of  generating  cDNA  amplicons  is  not  ideal  given  the  incidence  of sequences resembling other mutant constructs, which shows that common DNA surface contaminants can be  amplified  at  random.   An  example of  this  can be observed  in  the  three  instances where  the sequenced cDNA was shown to be identical or similar to the “seq mut v2” construct (“seq mut v2” and “seq  mut  v2  like”,  Table  4.1),  which  happened  to  be  in  the  process  of  being  cloned  while  these experiments were being conducted.  While one cDNA was shown to have the same sequence as the “5’ flank mut” construct, it is not likely that this is the result of a contamination since this construct had not yet been designed when these experiments were in progress.  Interestingly, this result implies that this position  in the CCR1 sequence has no  impact on viral  infectivity, which  is consistent with the fact that the “5’ flank mut” construct had WT levels of replication in transfected and infected BHK cells (discussed in Chapter 3).   In order to overcome the technical difficulties of this procedure, a collaboration with the Broad Institute (Cambridge MA) was established whereby the extracted vRNA samples would be sent for full genome  single‐molecule  sequencing.    Full  genome  coverage  would  allow  for  the  identification  of reversions in CCR1 as well as in additional portions of the genome, which might indicate the presence of a secondary RNA element.  Current studies are directed towards optomizing a high‐fidelity one‐step RT‐PCR  protocol  that  produces  four  sets  of  overlapping  amplicons  that  span  the  entire DENV  genome, which will be sent to the Broad  Institute for deep sequencing.   While  initial studies only examined the “CCR1orig  mut”  mutant  virus  at  48  hpi,  the  collaboration  with  the  Broad  Institute  includes  the investigation of vRNA  isolated  from the  full panel of mutant viruses at early and  late timepoints post‐

85  

infection in both BHK and C6/36 cells, as well as sequencing the input virus that was used to infect both cell types.   

FUTURE DIRECTIONS  Determining the mechanism by which CCR1 regulates  infectious particle production.   As discussed  in Chapter 3, CCR1 was implicated as an assembly signal for DENV that regulates the amount of infectious particle production in both BHK and C6/36 cells, though the precise mechanism by which this occurs is still under  investigation.   Virus particles within the virus‐containing supernatants from transfected BHK and C6/36 cells can be assessed by comparing the amount of vRNA to the amount of viral E or prM/M proteins  present  as  determined  by Western  blot  to  calculate  the  vRNA:particle  ratio.    If  there  is  no defect in viral assembly then the vRNA:particle ratio will be equivalent to that of the WT.  An assembly defect might result in less mutant viral particles being detected in the supernatant as compared to the WT,  and  confirm  the  observation  that  the  same  amount  of  vRNA  is  being  released  from  the  cells (Fig.3.6A and 3.6B).  Additionally, virus‐containing supernatants can be examined using sucrose gradient centrifugation and the amount of vRNA that is contained within assembled particles can be determined by comparing vRNA levels for both the WT and mutant viruses in E‐containing fractions.   If assembly is modulated by CCR1, then mutant vRNA will be  isolated from fractions that contain a disproportionate amount of the viral proteins capsid, prM/M or E while the WT vRNA will mostly associate with all three.  The best approach to determining whether the accumulation of noninfectious virions  is the result of a gross  structural defect  is  to  section BHK  and C6/36  cells  that  have been  infected with mutant CCR1 viruses and examine the samples using transmission electron microscopy (TEM).  TEM would show any structural differences  in  the assembled particles produced by cells  infected with  the different mutant viruses as compared  to  the WT, which might explain how altering CCR1 can have different effects on infectious particle production depending on the cell  line being tested.   As discussed, virion maturation could also be affected as a result of a defect  in viral assembly,  leading to a decrease  in the amount of infectious particles produced.  To detect this, virus‐containing supernatants would be subjected to furin treatment in vitro, and the resulting viruses used to infect a fresh monolayer of BHK cells.  Infection can be  assayed  at  24  hpi  using  immunoflourescence  to  determine  if  treatment with  furin  increases  the number of cells that could become infected, which would imply that insufficient cleavage by furin is the cause for the production of noninfectious particles.   While the manner in which CCR1 decreases the number of assembled particles that can undergo cell‐to‐cell spread is still under investigation, whether CCR1 regulation is dependent on RNA‐RNA and/or RNA‐protein interactions has yet to be determined.  To ascertain whether CCR1 interacts with a protein‐binding partner, electrophoretic mobility shift assays (EMSAs) can be carried out using radiolabelled WT and mutant CCR1 RNAs that are incubated with infected cell lysates.  When these reactions are analyzed by agarose gel electrophoresis, proteins that bind to the WT RNA will cause a shift to occur that should not  be  observed when  the mutant  CCR1  RNA  is  utilized.    For  these  experiments,  uninfected  lysates would be used as a negative control, and no differences in the mobility shift pattern are expected when using uninfected lysates as compared to the infected cell lysates.   If a difference  in the mobility shift pattern  is observed  in the preliminary experiments, an RNA affinity column will be employed using either the WT or mutant CCR1 RNAs.  Infected cell lysates will be passed over  the  column, and proteins bound with high affinity will be eluted using a high molar  salt solution and analyzed using polyacrylamide gel electrophoresis (PAGE).  Eluates that display differences by PAGE analysis will be examined using multidimensional protein  identification  technology  (MudPIT), which can be used  to  identify candidate protein‐binding partners.   Candidate proteins can be  isolated while bound to the WT CCR1 sequence using RNA immunoprecipitation (RIP) assays, and specificity can 

86  

be  assessed  with  quantitative  real‐time  (qRT)‐PCR  targeting  CCR1.    If  the  candidate  protein  is commercially available, protein‐binding  can also be  assessed using EMSAs whereby either  the WT or mutant  CCR1  RNAs  are  radioactively  labeled  and  incubated  with  increasing  concentrations  of  the candidate  protein  before  being  analyzed  by  agarose  gel  electrophoresis.    The  specificity  of  the interaction can be determined by adding increasing amounts of the unlabelled WT sequence to compete away  the protein bound  to  the  labeled mutant CCR1 RNAs.   An RNA protection assay will be used  to determine the binding site of the candidate protein along the CCR1 sequence.  Functional assays would then be needed  to establish  the biological  relevance of  the  interaction.   Cells will need  to be  treated with  small  interfering  RNAs  (siRNAs)  directed  against  the  candidate  protein  to  knock  down  its expression.  If viral infection kinetics are altered when these treated cells are infected with WT virus, this would  indicate  that  CCR1  is  involved  in  the  proposed  protein‐RNA  interaction  with  the  candidate protein.  These studies will help elucidate whether there are candidate protein(s) that bind to CCR1 and modulate  its  role on  infectious particle production as well as provide some  indication of how specific that interaction is.   If  CCR1  functions  through  an  RNA‐RNA  interaction,  this  will  be  addressed  by  studying  the incidence of viral revertants during vRNA sequencing of the virus‐containing supernatants derived from infected BHK and C6/36 cells.  The preliminary sequencing results described earlier, give some indication that there is no alternate RNA element in the DENV genome since reversions have been observed in the region of interest though these results are not conclusive.  Full‐genome sequencing of extracted vRNAs will help  elucidate whether or not  additional unknown RNA  elements  are  implicated  since  individual cDNAs  can be  compared  from  a  single  sample  rather  than  examining  a  consensus  sequence  derived from multiple cDNAs in the same sample.  If vRNA sequencing yields information regarding whether or not  there  is an additional RNA element  in  the DENV genome,  the  interaction can be confirmed using EMSAs whereby CCR1 is radiolabelled and increasing concentrations of the secondary RNA element are added.  If CCR1 is involved in an RNA‐RNA interaction then a mobility shift should be observed when the secondary RNA element is added and this shift will be abolished when either CCR1 or the secondary RNA element is mutated.  Functional assays should include mutating the secondary RNA element to see if the same effect on noninfectious particle production is achieved as observed with the CCR1 mutants.  Taken together,  these  assays  would  help  determine  a  mechanism  of  action  by  which  CCR1  modulates infectious viral particle production in both BHK and C6/36 cells.   

CONCLUSIONS  Overall,  these  studies  are  important  for  improving our understanding of  cis‐acting determinants  that differentially regulate the viral life cycle in the human host and mosquito vector.  The goal of this study was to  identify and characterize a cis‐acting coding‐region regulatory RNA element that modulates the viral life cycle since the predominant focus of research thus far has centered on those elements located in  the  UTRs.    Using  sequence  conservation  and  secondary  structure  predictions,  CCR1  has  been identified as a novel coding‐region RNA sequence element that regulates infectious particle production in  BHK  and  C6/36  cells,  possibly  serving  as  an  assembly  signal  for  DENV,  which  has  not  yet  been identified.    Furthermore,  the  critical  role  of  CCR1  in  C6/36  cells  has  been  shown  in  vivo,  whereby mutations  to  CCR1  reduce  viral  replication  in Ae.  aegypti mosquitoes  and  viral  dissemination  to  the salivary glands, providing  insight  into viral replication strategies  in the mosquito vector.   Coding‐region RNA regulatory elements have been described  in other RNA viruses and may serve as excellent targets for  novel  therapeutics  since  their  evolution  towards  resistance  is  constrained  by  both  their  RNA regulatory  function  and  amino  acid  coding  capacity.    Thus,  studies  that  examine  the  regulatory mechanisms  of  coding‐region  RNA  elements  are  critical,  because  these  determinants  play  important 

87  

roles in the viral life cycle and could potentially serve in the design of attenuated vaccine strains or act as candidate targets for antiviral agents.   

MATERIALS AND METHODS  Sequence  conservation  and  secondary  structure  prediction  algorithms.    For  conserved  sequence prediction,  sequences  were  aligned  using  the  Clustal  W2  web  software  (3)  from  the  European Bioinformatics Institute.   Portions of the Clustal W2 alignments were processed by the RNAalifold web server  (1)  from the  Institute of Theoretical Chemistry at the University of Vienna  in order to generate the phylogenetic conserved secondary structure predictions.  RT‐PCR.    vRNA  in  the  virus‐containing  supernatants was  isolated  using  the  ZR  Viral  RNA  kit  (Zymo Research).   Viral cDNA was  first produced using  the One‐Step RT‐PCR System  (Qiagen), purified using NucleoSpin  Extract  II  (Macherey‐Nagel)  and  sequenced  at  the  UC  Berkeley  Sequencing  Facility.  Subsequent  analysis  included producing  viral  cDNA using  the  SuperScript  III One‐Step RT‐PCR  System with Platinum Taq High Fidelity (Invitrogen), which was purified using NucleoSpin Extract  II and will be sequenced at the Broad Institute (Cambridge MA).  Primers sequences for the reverse transcriptase (RT)‐PCR are available in Table 4.2. 

AFIGURES

5’3’

NS5‐3

NS5‐1

BB

5’3’

C

NS5‐1

NS5‐3C

5’3’

NS5‐1

NS5‐3

88

Figure 4.1: Putative RNA structural elements are identified in the NS5‐coding region based on structureconservation in the DENV, JEV and TBEV serogroups. Phylogenetic consensus structure based on a portion of, g p y g pthe NS5‐coding sequence from the (A) DENV serogroup (position 9647‐10159, numbering based on DENV2 Thaistrain 16681; accession codes DENV1 U88535, DENV2 NC_001474, DENV3 AY099336 and DENV4 GU289913),(B) JEV serogroup (position 9767‐10280, numbering based on JEV; accession codes JEV M18370, Murray Valleyencephalitis virus AF161266, WNV NC_001563, KUN AY274505, St. Louis encephalitis virus DQ525916 and YFVU17067) and (C) TBEV serogroups (position 9678‐10218, numbering based on TBEV; accession codes TBEVU27495, Powassan virus NC_003687 and Omsk hemorrhagic virus AY193805). Consensus structures werecomputed with RNAalifold (1). Covariance is depicted as a colored spectrum ranging from red, to dark pink,light pink, yellow and then green. Putative RNA structure elements NS5‐1 and NS5‐3 are indicated by circles orboxes.

89

90  

TABLES   Table 4.1:   Sequencing  results  from BHK  cells  infected with  the “CCR1orig mut” mutant virus at 48 hours post‐infection.   CCR1 sequence  is highlighted  in yellow.   Mutations are  indicated by  lower case.  Capitalized nucleotides  that have been bolded  indicate  reversion  to  the WT  sequence whereas  lower case nucleotides that have been bolded indicate mutations that were not engineered. 

vRNA Sequencing  Incidence  Sequence Match 5’-CTGAAACGCGAGAGAAACCGCGTG-3’ 1  WT 5’-CTGAAgCGgGAGAGAAACCGCGTG-3’ 18  CCR1orig mut 5’-CTGAAgCGgGAGAGAAACCaCGcG-3’ 1  CCR1orig mut plus 5’-CcGAAgaGgGAacGtAAtaGaGTG-3’ 1  seq mut v2 like 5’-CTGAAACGaGAGAGAAACCGCGTG-3’ 1  5’ flank mut 5’-CTGAAgaGgGAacGtAAtaGaGTG-3’ 2  seq mut v2 5’-CTGAAACGgGAGAGAAACCGCGTG-3’ 1  CCR1orig mut like 

  Table 4.2:  Sequencing primers used for the generation of cDNA amplicons using RT‐PCR. Sequencing Facility  Primer  Sequence 

UC Berkeley F1 (FWD)  5’-AACGACAGGAGCACGATCATGC-3’ D2‐2 (REV)  5’-CAATCACGTACAAGTGTCCCC-3’

Broad Institute 

broad_D2_F_11_b (FWD)  5’-CTACGTGGACCGACAAAGACAG-3’ broad_D2_R_3669 (REV)  5’-CCATRCCTATGTCATCYGTCAT-3’ broad_D2_F_2411 (FWD)  5’-ATGGTGCARGCYGATAGTGGTT-3’ broad_R_6183 (REV)  5’-TCTGTCTGCRTAGTTGATRCCTTC-3’

broad_D2_F_4919 (FWD)  5’-CTGGRACGTCAGGATCTCCAAT-3’ broad_D2b_R_8618 (REV)  5’-CGCTGTTGTCCRAAWGGAGT-3’ broad_D2_F_7293 (FWD)  5’-CTAGAWCCAATACCYTATGATCC-3’ broad_D2_R_10477 (REV)  5’-CCATGCGTACAGCTTCCATGGT-3’

 

91  

REFERENCES   1.  Hofacker,  I. L., M. Fekete, and P. F. Stadler. 2002. Secondary structure prediction  for aligned 

RNA sequences. J Mol Biol 319:1059‐1066. 2.  Liu, Y., E. Wimmer, and A. V. Paul. 2009. Cis‐acting RNA elements  in human and animal plus‐

strand RNA viruses. Biochim Biophys Acta 1789:495‐517. 3.  Thompson,  J.  D.,  T.  J.  Gibson,  F.  Plewniak,  F.  Jeanmougin,  and  D.  G.  Higgins.  1997.  The 

CLUSTAL_X Windows  interface:  flexible  strategies  for multiple  sequence  alignment  aided  by quality analysis tools. Nucleic Acids Res 25:4876‐4882.