Bacterial ghosts – biological particles as delivery systemsfor antigens, nucleic acids and drugsChakameh Azimpour Tabrizi1, Petra Walcher1, Ulrike Beate Mayr1,Thomas Stiedl1,2, Matthias Binder1,2, John McGrath1,3 and Werner Lubitz1,2,�
Despite the exponential rate of discovery of new antigens and
DNA vaccines resulting from modern molecular biology and
proteomics, the lack of effective delivery technology is a
major limiting factor in their application. The bacterial ghost
system represents a platform technology for antigen, nucleic
acid and drug delivery. Bacterial ghosts have significant
advantages over other engineered biological delivery particles,
owing to their intrinsic cellular and tissue tropic abilities,
ease of production and the fact that they can be stored and
processed without the need for refrigeration. These
particles have found both veterinary and medical applications
for the vaccination and treatment of tumors and various
infectious diseases.
Addresses1 Institute of Microbiology and Genetics, Section Microbiology and
Biotechnology, University of Vienna, Althanstrasse 14, UZAII, 2B 522,
A-1090, Vienna, Austria2 BIRD-C GmbH&CoKEG, Schonborngasse 12/12, A-1080 Vienna,
Austria3 Gadi Research Centre, University of Canberra, ACT 2601, Australia�e-mail: [email protected]
Current Opinion in Biotechnology 2004, 15:530–537
This review comes from a themed issue on
Pharmaceutical biotechnology
Edited by Carlos A Guzman and Giora Z Feuerstein
Available online 28th October 2004
0958-1669/$ – see front matter
# 2004 Elsevier Ltd. All rights reserved.
DOI 10.1016/j.copbio.2004.10.004
Abbreviations
APC a
Current
ntigen-presenting cell
DOX d
oxorubicinDDS d
rug delivery systemMHC m
ajor histocompatibility complexOmpA o
uter membrane protein AIntroductionNew generation recombinant protein and DNA vaccines
are generally poorly immunogenic, thus there is an urgent
need to develop improved delivery systems and adjuvant
formulations. This requirement has fueled intense world-
wide research into biological particles as novel delivery
systems for antigens, nucleic acids and drugs. The ambi-
tious aim of future vaccines is to provide maximum
Opinion in Biotechnology 2004, 15:530–537
efficacy with minimum number of doses, delivered safely
and easily [1��] (see Table 1; Box 1). Several biological
and synthetic systems ranging from fusion proteins, lipid
spheres and sugar particles to virus-like particles and
whole-cell bacteria are in use or being investigated as
dual-carrier molecules and adjuvants for antigens, nucleic
acids and drugs (Box 2).
This review will focus on current strategies for antigen,
nucleic acid and drug delivery by biological particles with
an emphasis on the use of bacterial ghosts as delivery
systems.
Non-bacterial delivery systemsVarious non-bacterial biological delivery systems are listed
in Box 2. Of these, live attenuated or inactivated viruses
and virus-derived particles are the best documented deliv-
ery systems, as they are obligate parasites of eukaryotic
cells. So far, live attenuated viruses, virus-like particles
and virosomes have been developed [2]. Viral vectors can
improve the long-term expression of target genes through
the natural integration of the viral genome into that of the
host. Virus-like particles have been engineered using viral
structural proteins and nucleic acids as an alternative deliv-
ery system. One disadvantageofviruses,virus-like particles
and virosomes alike, is that their capacity to encapsulate
foreign antigens or DNA is restricted. Virosomes are immu-
nopotentiating reconstituted influenza virus envelopes of
approximately 150 nm in diameter, which comprise the
influenza surface glycoproteins haemagglutinin and neur-
aminidase (NA) and a mixture of natural and synthetic
phospholipids [3]. Safety concerns owing to the possibility
of incomplete inactivation, immunocompromised vaccine
recipients or revertants further limit the use of virosomes
and live attenuated viruses [4��].
Edible vaccines from transgenic plants offer a safer deliv-
ery system, but unfortunately this system requires strong
adjuvants to be immunogenic [5�] and much work is still
needed to improve the practical aspects of this approach.
Non-organism alternatives are being developed as deliv-
ery vehicles with greater success, especially for drug
delivery. Liposomes are spherical phospholipid bilayers
with an internal space that allows the incorporation of
hydrophilic antigens. To enhance the delivery capacity of
liposomes, different receptor molecules have been
included in the bilayer [6,7,8�,9,10,11��]. ISCOMs
www.sciencedirect.com
Bacterial ghosts Tabrizi et al. 531
Box 2 Particles used for immunization and drug delivery.
Biological particles
Live attenuated or inactivated bacteria [14,15,17,31��,33�,51–53]
Bacterial ghosts [28–30,54]
Live attenuated or inactivated viruses [55–58]
Virus-like particles [5�,36,37,59,60�]
Table 1
Advantages and limitations of biological particles as delivery vehicles.
Biological particle Advantages Limitations
Live attenuated or
inactivated bacteria
Activation of innate immune system via pattern recognition receptors Reversion to virulence
Generation of humoral and cell-mediated immune response Horizontal gene transfer
Used as carrier for antigens Stability of recombinant phenotype
Pre-existing immunity against carrier strain
Antibiotic markers
Live GRAS bacteria Non-pathogenic Poorly immunogenic
Virosomes Biodegradable, contain no preservatives or detergents
and present fewer localized adverse events
Pathogenic for immunocompromised
recipients
Limited capacity to encapsulate foreign
antigens or DNA
Live attenuated or
inactivated viruses
Improved long-term expression of target genes
using viral integration system into host genome
Incomplete inactivation
Revertants to virulence
Edible vaccines from
transgenic plants
Low-cost production, ease of use Requirement for strong adjuvants
Erythrocytes Biocompatibility, complete biodegradability Different serotypes
Hazard for blood transfusion
Bacterial ghosts Non-living carriers Presence of lipopolysaccharides
Carriage of different antigens, DNA and drugs simultaneously
Strong adjuvant properties
Good recognition and uptake by antigen-presenting cells
High loading capacity for DNA
Targeting properties for different tissues
Ideal vaccine characteristics are summarized in Box 2. GRAS, generally recognized as safe.
(immune-stimulating complexes) are nanometer-sized
micellar assemblies composed of saponin, cholesterol
and phospholipids that contain amphiphilic membrane
protein antigens [11��]. Nanoparticles and microparticles
based on liposomal or vesicular (niosomes) systems might
find a broad range of application in the future [12,13].
Bacterial delivery systemsIt is well accepted that live attenuated and inactivated
whole-cell bacteria constitute an effective delivery sys-
tem for recombinant antigens and nucleic acids [14].
The advantage of live or inactivated bacteria is that innate
immune cells recognize highly conserved structures on
microorganisms — termed pathogen-associated molecu-
lar patterns (e.g. lipopolysaccharides, CpG and outer
membrane protein A [OmpA]) — via pattern-recognition
receptors, such as the mannose receptor and Toll-like
Box 1 Ideal vaccine characteristics.
Safe in all ages and in immunocompromised patients
Efficacious in all ages (from young infants to the elderly)
Efficacious single-dose vaccine to be administered soon after birth
Early onset of protection (day 8)
Long-lived protection (years)
Cold-chain independent (i.e. can be stored and processed without
refrigeration)
Needle-free delivery (e.g. mucosal route of immunisation)
Practical, simple formulation (e.g. stable dry powder for non-invasive
administration)
Multiple vaccine combinations
www.sciencedirect.com
receptors [15]. Thus, heterologous antigens aside, carrier
bacteria can generate both humoral and cell-mediated
immunity. As a constituent of the bacteria, the hetero-
logous antigens are not only targeted to appropriate path-
ways of major histocompatibility complex (MHC) class I
and class II antigen processing and presentation, but also
generate an adequate cytokine milieu for promoting
antigen-specific responses [16].
Despite the strong and well-documented immunological
advantages of bacterial delivery vectors, there are safety
Edible vaccines from transgenic plants [5�]
Erythrocytes [43�]
Phage display of antigens [61]
Synthetic particles
Virosomes [3,4��,62]
Liposomes [6,7,8�,9,10,11��,32��]
Peptides and proteins [44�,63]
Proteosomes [11��,64]
Niosomes [12,13]
ISCOM (immune-stimulating complexes) [11��,65,66]
Chitosan microspheres [11��,67,68]
Polylactide/polyglycolide microspheres [69–71]
Cochleates, phospholipids-calcein precipitates [11��]
Current Opinion in Biotechnology 2004, 15:530–537
532 Pharmaceutical biotechnology
Figure 1
Scanning electron micrograph of an E. coli bacterial ghost produced using modified protein E lysis procedure [28]. A large opening enables the
bacterial ghost envelope complex to be revealed.
limitations with the bacterial adjuvant/delivery system.
Traditionally, pathogenic bacteria like Listeria spp.,
Salmonella spp., Yersinia spp., Shigella spp. and Mycobac-teria spp. are the most frequently used bacterial carriers.
The most significant problems in their usage are the
possibility of reversion to virulence, horizontal gene
transfer, the stability of the recombinant phenotype,
pre-existing immunity against the carrier strain and the
presence of antibiotic markers [17]. An alternative
approach is the use of live GRAS (generally recognized
as safe) bacteria, which are non-pathogenic and are inves-
tigated solely as delivery vectors (see Table 1; Box 1) [18].
Bacterial ghosts as delivery systemsBacterial ghosts represent empty non-denaturated envel-
opes derived from Gram-negative bacteria by protein-
E-mediated lysis, which retain all morphological and
structural features of the natural cell (for a review see
[19]). They can be used as vaccine candidates per se with
intrinsic adjuvant properties based on well-known
immune-stimulating compounds such as lipopolysacchar-
ides, lipid A and peptidoglycan (Figures 1 and 2). Alter-
natively, they can be employed as a delivery system for
proteins/antigens, nucleic acids, drugs and soluble com-
pounds for various medical and technical applications
[20–27]. Not only can bacterial ghosts act as delivery
vehicles for inner and/or outer membrane tethered anti-
gens, but they can also deliver water-soluble drugs or
antigens (see Figure 2; Table 2) [28–30]. The results of
several in vivo and in vitro studies confirm the potential of
bacterial ghosts as dual carrier/adjuvant technology for
modern vaccine development (for a review see [28]). The
cellular and tissue tropism of bacterial ghosts in combina-
tion with excellent carrier capacity in several cellular
compartments (Table 2) offers much potential for anti-
Current Opinion in Biotechnology 2004, 15:530–537
gen, nucleic acid and drug delivery. Bacterial ghosts are
taken up very effectively by antigen-presenting cells
(APCs) such as macrophages and dendritic cells and
are particularly suited as vaccines for mucosal adminis-
tration by oral, intranasal or aerogenic routes, resulting in
the induction of humoral and cellular immune responses
[28,30].
Delivery systems for antigensBecause new-generation vaccines based on recombinant
proteins are often less immunogenic than traditional
vaccines, they require specific choices of delivery parti-
cles and adjuvants to improve their presentation and
targeting and to thus induce an appropriate protective
immune response. This is particularly important at the
mucosa, the most effective site for immune stimulation.
The choice of carriers and/or adjuvants and of the antigen
itself have the potential to modulate the immune
response (i.e. predominantly B-cell or T-cell based)
appropriate for a particular pathogen.
The development of efficacious vaccines against intra-
cellular bacteria, parasites and viruses requires the induc-
tion of T-cell mediated responses. One ingenious
delivery system makes use of antigen fusions with the
listeriolysin enzyme of Listeria monocytogenes to help
induce a CD8+ T-cell response against phagosomal and
cytosolic antigens [31��]. Listeriolysin expression disrupts
the phagosomal membrane, releasing the target antigen
into the cytoplasm for MHC I presentation and T-cell
activation. Similarly, other groups have employed the
Escherichia coli a-haemolysin (HlyA) secretion system
for delivery of heterologous antigens and a large number
of hybrid proteins have been generated by gene fusion
with the C-terminal end of HlyA [32��].
www.sciencedirect.com
Bacterial ghosts Tabrizi et al. 533
Figure 2
Outer membrane
Periplasm
Inner membrane and cytoplasm
Phospholipid
Lipopolysaccharide
Peptidoglycan
Mannose
Maltose-binding proteinTarget antigen
S-layerprotein
Omp A
Streptavidin Biotin
DNA Drug
E′
LacI LacOs
Membrane proteinsPorin
Lipoprotein
Pilus
Current Opinion in Biotechnology
L′ Membrane anchor
Schematic of the localisation of different molecules in bacterial ghosts. The outer membrane, the periplasmic space and the inner membrane
facing the cytoplasmic lumen with integrated target antigens and different structural elements are drawn in a cartoon with specific emphasis
on the potential locations of target antigens and their carrier proteins.
Another intracellular antigen delivery system has been
developed with virus-like particles as well as virosomes
[4��]. Peptide vaccination with virosome carriers has been
investigated in several disease models including malaria,
melanoma and hepatitis C [3].
Compared with simple virus-like particle carriers, the
bacterial cell offers several compartments for the delivery
of immunogenic antigens and has a greater capacity.
Expression of an antigen in the cytosol, periplasm or
outer membrane of the carrier bacteria can have a pro-
found impact on the elicited immune response. For
Table 2
Applications of bacterial ghosts as delivery systems.
Display compartment Display of antigens
Outer membrane Surface presentation by OmpA fusion or
through fusion with pili structures
Periplasmic space Presentation of foreign antigens by MalE fusion
Inner membrane Anchoring of foreign proteins specific with
N0-, C0- or N0- and C0- membrane
anchors to the inner membrane
Cytoplasmic space Paracristalline fusion protein sheets,
which remain in the cytoplasmic lumen
after E-mediated lysis of the carrier bacteria
www.sciencedirect.com
example, surface-exposed expression or secretion of anti-
gens leads to a better induction of specific antibodies [14].
Antigenic epitopes have also been inserted into flagellin,
fimbriae or in the outer membrane or periplasmic proteins
MalE, LamB, OmpA and PhoE in different Salmonellastrains [31��]. An alternative to conventional bacterial
delivery system approaches has been reported in the
development of a surface-display system based on the
use of the spore coat of Bacillus subtilis [33�]. This has an
interesting advantage in that bacterial spores can survive
extremes of temperature, desiccation and exposure to
solvents and other noxious chemicals.
Therapeutic proteins
or peptides
Nucleic acids Drugs
Binding of hydrophobic
drugs by affinity to
membranes
Membrane-bound
enzymes
Filling with DNA
plasmids from 4 to
5000 copies/ghost
Sealing bacterial
ghosts for water
soluble drugs
Current Opinion in Biotechnology 2004, 15:530–537
534 Pharmaceutical biotechnology
Bacterial ghosts offer a safe, easy to manipulate and
straightforward to produce alternative to traditional anti-
gen bacterial carrier systems, with all of the advantages of
the latter. Foreign protein localisation within bacterial
ghosts is performed by fusion with specific anchor
sequences for attachment on the inside of the inner
membrane, export into the periplasmic space by fusion
to the MalE signal sequence or attachment to the outer
membrane as fusion proteins with OmpA or pili (Table 2;
Figure 2) [34]. Together with heterologously expressed
S-layer proteins SbsA and SbsB, which form shell-like
self-assembly structures filling the periplasmic and/or the
internal lumen cytoplasmic space, the capacity of ghost
vectors to function as carriers of polypeptides is vastly
extended [35].
The suitability of bacterial ghost technology for designing
an antichlamydial vaccine was evaluated by constructing
a candidate vaccine based on a Vibrio cholerae vector
expressing major outer membrane proteins. The efficacy
of the vaccine was assessed in a murine model of Chla-midia trachomatis genital infection [34]. Intramuscular
delivery of the vaccine candidate induced elevated local
genital mucosal as well as systemic T helper 1 (Th1)
responses. In addition, immune T cells from immunized
mice could transfer partial protection against a C. tracho-matis genital challenge to naıve mice. These results
suggest that V. cholerae ghosts expressing chlamydial
proteins might constitute a suitable subunit vaccine for
inducing an efficient mucosal T-cell response that pro-
tects against C. trachomatis infection.
Importantly for conformation-dependent B-cell epitope
presentation, it has been shown that the enzymatic activ-
ities of membrane-attached b-galactosidase and polyhy-
droxybutyrate synthase in bacterial ghosts is not impaired
by the attachment. This indicates that the membrane
anchors do not interfere with the proper folding of the
target proteins and that self-assembly of subunits (e.g. for
b-galactosidase) is possible [30].
Delivery systems for nucleic acidsAlthough no DNA vaccine has yet been approved for
routine human or veterinary use, the potential of this
vaccination strategy has been repeatedly demonstrated in
experimental animal models. Because of the simplicity
and versatility of these vaccines, various routes and modes
of delivery are used to elicit the desired immune
response; however, the need for large amounts of DNA
and numerous doses for optimal vaccination has led to the
search for delivery systems better able to target cells and
improve currently poor immunogenicity.
The intracellular nature of viruses has been exploited as a
tool for DNA delivery with the development of virus-like
particles. The abilities of a non-replicative DNA delivery
system based on parvovirus-like particles to induce cyto-
Current Opinion in Biotechnology 2004, 15:530–537
toxic T lymphocyte responses in the neonatal period has
been shown recently [36]. Results from phase I and phase
II human clinical trials indicated that virus-like particles
are safe, well tolerated and immunogenic when adminis-
tered parenterally [5�,37]. In a more complex approach,
the efficacy of an intranasally administered mumps DNA-
vaccine delivered using cationic virosomes as carrier and
E. coli heat-labile toxin as adjuvant has been demon-
strated in a mouse model [3].
The development of bacterial carriers somewhat extends
the application of DNA vaccines for mucosal immuniza-
tion [28]. Significant humoral and cellular immune
responses against bacterial, viral and tumor antigens have
been induced by in vivo delivery of DNA vaccines in
small-animal models. Encouragingly, results have been
demonstrated with a broad spectrum of Gram-positive
and Gram-negative bacterial vectors, including L. mono-cytogenes [38], Salmonella typhimurium [39��], Salmonellatyphi, Shigella flexneri [39��,40] and invasive E. coli [38].
A delivery system based on bacterial ghosts has also
proven effective for DNA vaccines. In vitro studies
showed that Mannheimia haemolytica ghosts loaded with
a plasmid carrying the gene encoding green fluorescent
protein are efficiently taken up by APCs with high (52–
60%) transfection rates [41]. Subsequent in vivo vaccina-
tion studies in Balb/c mice demonstrated that M. haemo-lytica ghost-mediated DNA delivery by intradermal or
intramuscular route of a eukaryotic expression plasmid
encoding for b-galactosidase under the control of a cyto-
megalovirus promoter (pCMVbeta), stimulated more effi-
cient antigen-specific humoral and cellular (CD4+ and
CD8+) immune responses than naked DNA. It was shown
that the use of bacterial ghosts as DNA carriers allowed
for modulation of the major T-helper cell response (from
a mixed Th1/Th2 to a more dominant Th2 pattern) to the
b-galactosidase gene product compared with the naked
DNA. Moreover, intravenous immunization with dendri-
tic cells loaded ex vivo with pCMVbeta-containing ghosts
elicited b-galactosidase-specific responses [41]. The
results are certainly encouraging considering the primary
role of dendritic cells as APCs. Bacterial ghosts not only
target the DNA vaccine construct to APCs, but also
provide a strong danger signal by acting as natural adju-
vants (being inactivated whole-cell bacteria), thereby
promoting efficient maturation and activation of dendritic
cells. Thus, bacterial ghosts constitute a promising tech-
nology platform for the development of more efficient
DNA vaccines.
More recently, a new delivery system based on bacterial
ghosts has been developed in which, following E-
mediated lysis, DNA is tethered via a DNA-binding
membrane-anchored protein to the ghost inner mem-
brane. This system is the subject of continuing studies
(P Mayrhofer et al., unpublished).
www.sciencedirect.com
Bacterial ghosts Tabrizi et al. 535
Delivery systems for drugsMany diseases and cancers require the systemic admin-
istration of highly aggressive drugs to already immuno-
compromised patients. Deleterious and often severe side
effects result from a lack of cellular and tissue selectivity.
Another major issue is the poor solubility of some drugs
used in cancer treatment. Considering this, the develop-
ment of a safer and more efficient drug delivery system
(DDS) is the priority for future prophylactic treatments.
The DDS has three major goals: enhancement of drug
permeability for crossing physiological barriers; targeting
of drugs to the point of action; and the controlled release
of drugs [42]. Several biological DDSs are currently
employed, although none is ideal for all applications.
The use of erythrocytes in drug targeting has many
advantages, including biocompatibility, complete biode-
gradability and lack of toxic products, longer life-span
compared with synthetic carriers and a relatively inert
intracellular environment. Applications for this system
include intravenous slow drug release (e.g. antineo-
plasms, vitamins and antibiotics), enzyme therapy, target-
ing the reticuloendothelial system (e.g. adriamycin and
bleomycin against hepatic tumors and antileishmanial or
antiamoebial drugs against parasitic disease), and the
improvement of oxygen delivery to tissues [43�]. Another
DDS that exploits host constituents is the use of fusions
with the plasma protein, albumin. Albumin meets several
requirements of a drug carrier and shows accumulation in
tumors and in inflamed joints in patients with rheumatoid
arthritis [44�].
In a similar drug-packaging mechanism to erythrocytes,
liposomes have been utilized as drug delivery vehicles
for cancer treatment [45,46]. An ideal liposome formula-
tion, optimized for stability, will improve drug delivery
by decreasing the required dose and increasing the
efficacy of the entrapped drug at the target organ or
tissue [47].
More recently, several novel DDSs have been developed.
One of these involves nanoscale hepatitis B virus surface
antigen (HBsAg) L particles, which have many properties
that make them useful as in vivo gene transfer vectors and
as a DDS. The display of various cell-binding molecules
on the surface makes L particles particularly useful for
cell- and tissue-specific gene/drug delivery [48].
Bacteria might also offer a solution for drug delivery,
particularly through their tropic capacity. As a naturally
tissue tropic delivery system, bacterial ghosts have shown
early promise as a DDS. Recently, bacterial ghosts made
from the colonic commensal M. haemolytica were used for
the in vitro delivery of doxorubicin (DOX) to human
colorectal adenocarcinoma (Caco-2) cells. Adherence
studies showed that the M. haemolytica ghosts targeted
the Caco-2 cells and released the loaded DOX within the
www.sciencedirect.com
cells. Cytotoxicity assays showed a two-log enhancement
in cytotoxic and antiproliferative activity in cells incu-
bated with DOX-loaded ghosts compared with DOX
directly added to the culture media [49].
Current work with bacterial ghosts lies in the investiga-
tion of the carrier capacity of the cytoplasmic lumen. This
intracellular space of bacterial ghosts can be filled either
with water-soluble substances or emulsions such that the
drug(s) of interest can be coupled to streptavidin
anchored on the inside of the cytoplasmic membrane.
For some purposes, it is advantageous to fill the internal
space of the ghost with a substituted matrix, which then
binds the drug(s) of interest. In model experiments,
biotinylated fluorescence-labelled dextran has been used
to completely fill the internal space of streptavidin-ghosts
[50]. As substituted dextran has a high capacity for bind-
ing peptides, drugs or other substances, therapy and
prevention might yet prove feasible with bacterial ghosts
as tropic carriers. Also, bacterial ghosts can be filled and
sealed for the delivery of fluid, non-anchored substances.
In a recent study, E. coli ghosts were filled with the
reporter substance calcein and were sealed by fusion with
membrane vesicles to maintain inner membrane integr-
ity. Adherence and uptake studies showed that murine
macrophages and human Caco-2 cells took up the bacter-
ial ghosts and calcein was released within the cell [29].
ConclusionsBacterial ghosts are very useful non-living carriers, as they
can carry foreign antigens, nucleic acids and drugs in one
or more cellular locations simultaneously. Their ease of
manufacture, the fact that they can be stored and pro-
cessed without the need for refrigeration and their excel-
lent safety profile — even when administered at high
doses — are important considerations for a broad spec-
trum of applications. The identical surface receptors of
bacterial ghosts and their living counterparts are being
exploited for specific cellular and tissue targeting. Few
other biological delivery systems offer such excellent
carrier qualities in combination with application-based
tropism in humans, animals or plant tissues.
AcknowledgementsThe technical assistance of Beate Bauer for preparing the manuscriptis greatly appreciated. This work was supported by grantGZ 309.049/1-VI/6/2003.
References and recommended readingPapers of particular interest, published within the annual period ofreview, have been highlighted as:
� of special interest�� of outstanding interest
1.��
Sesardic D, Dobbelaer R: European union regulatorydevelopments for new vaccine adjuvants and deliverysystems. Vaccine 2004, 22:2452-2456.
This review covers EU regulations for new vaccine adjuvants, deliverysystems, safety evaluations and existing bottle necks in the field ofvaccine development.
Current Opinion in Biotechnology 2004, 15:530–537
536 Pharmaceutical biotechnology
2. Warfield KL, Swenson DL, Negley DL, Schmaljohn A, Aman MJ,Bavari S: Marburg virus-like particles protect guinea pigs fromlethal Marburg virus infection. Vaccine 2004, 22:3459-3502.
3. Gluck R, Metcalfe IC: Novel approaches in the development ofimmunopotentiating reconstituted influenza virosomes asefficient antigen carrier systems. Vaccine 2003, 21:611-615.
4.��
Gluck R, Metcalfe IC: New technology platforms in thedevelopment of vaccines for the future. Vaccine 2002,20 (Suppl 5):B10-B16.
In this paper, the authors extensively describe virosomes and their use ascytosolic delivery systems for drugs and peptides and genetic vaccines.In addition, other novel approaches including bacterial carriers, recom-binant measles virus vaccine and polysaccharide–protein conjugates arediscussed.
5.�
Warzecha H, Mason HS, Lane C, Tryggvesson A, Rybicki E,Williamson AL, Clements JD, Rose RC: Oral immunogenicity ofhuman papillomavirus-like particles expressed in potato.J Virol 2003, 77:8702-8711.
This study demonstrates the feasibility of producing edible vaccines usingvirus-like particles by introducing a codon-optimized version of humanpapilloma virus type 11 L1 into tobacco and potato.
6. Copland MJ, Baird MA, Rades T, McKenzie JL, Becker B, Reck F,Tyler PC, Davies NM: Liposomal delivery of antigen to humandendritic cells. Vaccine 2003, 21:883-890.
7. Gregoriadis G, Bacon A, Caparros-Wanderley W, McCormack B:A role for liposomes in genetic vaccination. Vaccine 2002,20 (Suppl 5):B1-B9.
8.�
Park JW: Liposome-based drug delivery in breast cancertreatment. Breast Cancer Res 2002, 4:95-99.
This publication focuses on different approaches to treating breastcancer, including a description of different drugs for therapeutic applica-tions and a discussion of targeting systems that are still in development orwhich have already been approved.
9. Wolff AC: Liposomal anthracyclines and new treatmentapproaches for breast cancer. Oncologist 2003,8 (Suppl 2):25-30.
10. Colletier JP, Chaize B, Winterhalter M, Fournier D: Proteinencapsulation in liposomes: efficiency depends oninteractions between protein and phospholipid bilayer.BMC Biotechnol 2002, 2:9.
11.��
Kersten GF, Crommelin DJ: Liposomes and ISCOMs. Vaccine2003, 21:915-920.
This review focuses on developments in the field of liposomes, liposome-like particles as well as ISCOMs (immune-stimulating complexes). A briefoverview on recent developments with emphasis on pharmaceuticalaspects is given.
12. Roco MC: Nanotechnology: convergence with modern biologyand medicine. Curr Opin Biotechnol 2003, 14:337-346.
13. Nasseri B, Florence AT: Microtubules formed by capillaryextrusion and fusion of surfactant vesicles. Int J Pharm 2003,266:91-98.
14. Kang HY, Curtiss R III: Immune responses dependent onantigen location in recombinant attenuated Salmonellatyphimurium vaccines following oral immunization.FEMS Immunol Med Microbiol 2003, 37:99-104.
15. Jeannin P, Magistrelli G, Goetsch L, Haeuw JF, Thieblemont N,Bonnefoy JY, Delneste Y: Outer membrane protein A (OmpA): anew pathogen-associated molecular pattern that interactswith antigen presenting cells-impact on vaccine strategies.Vaccine 2002, 20 (Suppl 4):A23-A27.
16. Beyer T, Herrmann M, Reiser C, Bertling W, Hess J: Bacterialcarriers and virus-like-particles as antigen delivery devices:role of dendritic cells in antigen presentation. Curr Drug TargetsInfect Disord 2001, 1:287-302.
17. Dietrich G, Spreng S, Favre D, Viret JF, Guzman CA: Liveattenuated bacteria as vectors to deliver plasmid DNAvaccines. Curr Opin Mol Ther 2003, 5:10-19.
18. Medina E, Guzman CA: Use of live bacterial vaccine vectors forantigen delivery: potential and limitations. Vaccine 2001,19:1573-1580.
Current Opinion in Biotechnology 2004, 15:530–537
19. Lubitz W, Witte A, Eko FO, Kamal M, Jechlinger W, Brand E,Marchart J, Haidinger W, Huter V, Felnerova D et al.: Extendedrecombinant bacterial ghost system. J Biotechnol 1999,73:261-273.
20. Lubitz W, Szostak M: Carrier-bound recombinant proteins,processes for their production and use as immunogens andvaccines. 1997, US Patent 5,470,573.
21. Lubitz W, Haidinger W: Nucleic acid free ghost preparations.2002, Patent PCT/EP02/07758.
22. Lubitz W: Process for the production of vaccines and their use.2001, US Patent 6,177,083.
23. Lubitz W: Compartmentalization of recombinant polypeptidesin host cells. 2003, US Patent 6,610,517.
24. Lubitz W, Resch S: Secretion of carrier-bound proteins into theperiplasm and into the extracellular space. 2003, US Patent6,596,510.
25. Lubitz W, Huter V: Loaded bacterial ghosts as targeting andcarrier vehicles. 2003, Patent EP 1158966 B1.
26. Lubitz W, Harkness R: Recombinant DNA, process for theproduction thereof and the use thereof. 1991, US Patent5,075,223.
27. Lubitz W, Paukner S: Sealing of bacterial ghosts. 2001,Patent PCT/EP01/00864.
28. Jalava K, Eko FO, Riedmann E, Lubitz W: Bacterial ghosts ascarrier and targeting systems for mucosal antigen delivery.Expert Rev Vaccines 2003, 2:45-51.
29. Paukner S, Kohl G, Jalava K, Lubitz W: Sealed bacterialghosts – novel targeting vehicles for advanced drug delivery ofwater-soluble substances. J Drug Target 2003, 11:151-161.
30. Jechlinger W, Haidinger W, Paukner S, Mayrhofer P, Riedmann E,Marchart J, Mayr U, Haller C, Kohl G, Walcher P et al.: Bacterialghosts as carrier and targeting systems for antigen delivery. InVaccine Delivery Strategies. Edited by Goebel W. HorizonScientific Press; 2002:163-184.
31.��
Samuelson P, Gunneriusson E, Nygren PA, Stahl S: Display ofproteins on bacteria. J Biotechnol 2002, 96:129-154.
This review gives an overview of the basic principles of various bacterialdisplay systems and highlights current uses and possible future trends intheir technological application.
32.��
Dietrich G, Viret JF, Gentschev I: Haemolysin A and listeriolysin –two vaccine delivery tools for the induction of cell-mediatedimmunity. Int J Parasitol 2003, 33:495-505.
This review demonstrates the use of haemolysin A and listeriolysin asvaccine delivery tools and highlights their use for vaccination againstprotozoan parasites by induction of cell-mediated immunity.
33.�
Mauriello EM, Duc le H, Isticato R, Cangiano G, Hong HA,De Felice M, Ricca E, Cutting SM: Display of heterologousantigens on the Bacillus subtilis spore coat using CotCas a fusion partner. Vaccine 2004, 22:1177-1187.
In this paper the authors describe the bacterial spore as an efficientvehicle for mucosal immunization and demonstrate the advantages of thisnovel antigen delivery strategy.
34. Eko FO, Lubitz W, McMillan L, Ramey K, Moore TT, Ananaba GA,Lyn D, Black CM, Igietseme JU: Recombinant Vibrio choleraeghosts as a delivery vehicle for vaccinating against Chlamydiatrachomatis. Vaccine 2003, 21:1694-1703.
35. Riedmann EM, Kyd JM, Smith AM, Gomez-Gallego S, Jalava K,Cripps AW, Lubitz W: Construction of recombinant S-layerproteins (rSbsA) and their expression in bacterial ghosts – adelivery system for the nontypeable Haemophilus influenzaeantigen Omp26. FEMS Immunol Med Microbiol 2003, 37:185-192.
36. Martinez X, Regner M, Kovarik J, Zarei S, Hauser C, Lambert PH,Leclerc C, Siegrist CA: CD4-independent protective cytotoxic Tcells induced in early life by a non-replicative delivery systembased on virus-like particles. Virology 2003, 305:428-435.
37. Koutsky LA, Ault KA, Wheeler CM, Brown DR, Barr E, Alvarez FB,Chiacchierini LM, Jansen KU: A controlled trial of a humanpapillomavirus type 16 vaccine. N Engl J Med 2002,347:1645-1651.
www.sciencedirect.com
Bacterial ghosts Tabrizi et al. 537
38. Gentschev I, Dietrich G, Spreng S, Kolb-Maurer A, Brinkmann V,Grode L, Hess J, Kaufmann SH, Goebel W: Recombinantattenuated bacteria for the delivery of subunit vaccines.Vaccine 2001, 19:2621-2628.
39.��
Weiss J: Transfer of eukaryotic expression plasmids tomammalian host cells by salmonella spp. Int J Med Microbiol2003, 293:95-106.
This comprehensive review addresses the principles and application ofbacteria-mediated gene transfer. The paper highlights the ability ofattenuated intracellular bacterial, such as S. typhimurium, S. typhi andS. flexneri, L. monocytogenes, to transfer plasmid DNA into the host cell.Recombinant E. coli and Agrobacterium tumeophaciens baterial ghostsare described in this context. In addition, the author thoroughly evaluatesthe effectiveness and potential of Salmonella-mediated and mucosalDNA vaccination.
40. Xu F, Hong M, Ulmer JB: Immunogenicity of an HIV-gag DNAvaccine carried by attenuated Shigella. Vaccine 2003,21:644-648.
41. Ebensen T, Paukner S, Link C, Kudela P, de Domenico C, Lubitz W,Guzman CA: Bacterial ghosts are an efficient delivery systemfor DNA vaccines. J Immunol 2004, 172:6858-6865.
42. Yasukawa T, Ogura Y, Tabata Y, Kimura H, Wiedemann P,Honda Y: Drug delivery systems for vitreoretinal diseases.Prog Retin Eye Res 2004, 23:253-281.
43.�
Hamidi M, Tajerzadeh H: Carrier erythrocytes: an overview.Drug Deliv 2003, 10:9-20.
An excellent review covering most aspects of erythrocytes as carriervehicles. Besides the descriptions of various carrier applications it givesan overview of methods for encapsulation of drugs and bioactive agents,advantages of erythrocytes for drug delivery, in vitro characteristics andstorage of carrier erythrocytes and information about drug release.
44.�
Wunder A, Muller-Ladner U, Stelzer EH, Funk J, Neumann E,Stehle G, Pap T, Sinn H, Gay S, Fiehn C: Albumin-based drugdelivery as novel therapeutic approach for rheumatoidarthritis. J Immunol 2003, 170:4793-4801.
An interesting publication concerning a new approach in the therapy ofrheumatoid arthritis using methothrexat coupled to human serum albu-min. Results of both in vitro and in vivo experiments are shown.
45. Pignatello R, Puleo A, Puglisi G, Vicari L, Messina A: Effect ofliposomal delivery on in vitro antitumor activity of lipophilicconjugates of methotrexate with lipoamino acids.Drug Deliv 2003, 10:95-100.
46. Mandal AK, Sinha J, Mandal S, Mukhopadhyay S, Das N:Targeting of liposomal flavonoid to liver in combatinghepatocellular oxidative damage. Drug Deliv 2002, 9:181-185.
47. Anderson M, Omri A: The effect of different lipid components onthe in vitro stability and release kinetics of liposomeformulations. Drug Deliv 2004, 11:33-39.
48. Yamada T, Ueda M, Seno M, Kondo A, Tanizawa K, Kuroda S:Novel tissue and cell type-specific gene/drug delivery systemusing surface engineered hepatitis B virus nano-particles.Curr Drug Targets Infect Disord 2004, 4:163-167.
49. Paukner S, Kohl G, Lubitz W: Bacterial ghosts as noveladvanced drug delivery systems: antiproliferative activity ofloaded doxorubicin in human Caco-2 cells. J Control Release2004, 94:63-74.
50. Huter V, Szostak MP, Gampfer J, Prethaler S, Wanner G, Gabor F,Lubitz W: Bacterial ghosts as drug carrier and targetingvehicles. J Control Release 1999, 61:51-63.
51. Dietrich G, Kolb-Maurer A, Spreng S, Schartl M, Goebel W,Gentschev I: Gram-positive and Gram-negative bacteria ascarrier systems for DNA vaccines. Vaccine 2001, 19:2506-2512.
52. Weiss S: Transfer of eukaryotic expression plasmids tomammalian hosts by attenuated Salmonella spp. Int J MedMicrobiol 2003, 293:95-106.
53. Spreng S, Dietrich G, Goebel W, Gentschev I: Protection againstmurine listeriosis by oral vaccination with recombinantSalmonella expressing protective listerial epitopes within asurface-exposed loop of the TolC-protein. Vaccine 2003,21:746-752.
www.sciencedirect.com
54. Riedmann EM, Kyd JM, Cripps AW, Lubitz W: Adjuvantproperties of bacterial ghosts. 2004, in press.
55. Rayner JO, Dryga SA, Kamrud KI: Alphavirus vectors andvaccination. Rev Med Virol 2002, 12:279-296.
56. Lundstrom K: Alphavirus vectors for vaccine production andgene therapy. Expert Rev Vaccines 2003, 2:447-459.
57. Balasuriya UB, Heidner HW, Davis NL, Wagner HM, Hullinger PJ,Hedges JF, Williams JC, Johnston RE, David Wilson W, Liu IKet al.: Alphavirus replicon particles expressing the two majorenvelope proteins of equine arthritis virus induce high levelprotection against challenge with virulent virus in vaccinatedhorses. Vaccine 2002, 20:1609-1617.
58. Bramson J, Dayball K, Evelegh C, Wan YH, Page D, Smith A:Enabling topical immunization via microporation: anovel method for pain-free and needle-free deliveryof adenovirus-based vaccines. Gene Ther 2003,10:251-260.
59. Yao Q, Bu Z, Vzorov A, Yang C, Compans RW: Virus-like particleand DNA-based candidate AIDS vaccines. Vaccine 2003,21:638-643.
60.�
Reed SG, Campos-Neto A: Vaccines for parasitic and bacterialdiseases. Curr Opin Immunol 2003, 15:456-460.
This paper outlines recent developments in vaccine strategies againstcommon parasitic diseases such as malaria and leishmaniasis, as well asbacterial diseases like tuberculosis and meningitis.
61. McGrath S, Fitzgerald GF, van Sinderen D: The impact ofbacteriophage genomics. Curr Opin Biotechnol 2004,15:94-99.
62. Cusi MG, Zurbriggen R, Valassina M, Bianchi S, Durrer P,Valensin PE, Donati M, Gluck R: Intranasal immunization withmumps virus DNA vaccine delivered by influenza virosomeselicits mucosal and systemic immunity. Virology 2000,277:111-118.
63. Tung CH, Weissleder R: Arginine containing peptides asdelivery vectors. Adv Drug Deliv Rev 2003, 55:281-294.
64. Jones T, Allard F, Cyr SL, Tran SP, Plante M, Gauthier J,Bellerose N, Lowell GH, Burt DS: A nasal proteosome influenzavaccine containing baculovirus-derived hemagglutinininduces protective mucosal and systemic immunity.Vaccine 2003, 21:3706-3712.
65. Agrawal L, Haq W, Hanson CV, Rao DN: Generating neutralizingantibodies, Th1 response and MHC non-restrictedimmunogenicity of HIV-I env and gag peptides in liposomesand ISCOMs with in-built adjuvanticity. J Immune Based TherVaccines 2003, 1:5.
66. Demana PH, Davies NM, Berger B, Rades T: Incorporation ofovalbumin into ISCOMs and related colloidal particlesprepared by the lipid film hydration method. Int J Pharm 2004,278:263-274.
67. Sinha VR, Singla AK, Wadhawan S, Kaushik R, Kumria R, Bansal K,Dhawan S: Chitosan microspheres as a potential carrier fordrugs. Int J Pharm 2004, 274:1-33.
68. Jiang HL, Park IK, Shin NR, Yoo HS, Akaike T, Cho CS: Controlledrelease of Bordetella Bronchiseptica Dermonecrotoxin (BBD)vaccine from BBD-loaded chitosan microspheres in vitro.Arch Pharm Res 2004, 27:346-350.
69. Peyre M, Fleck R, Hockley D, Gander B, Sesardic D: In vivouptake of an experimental microencapsulated diphtheriavaccine following sub-cutaneous immunisation. Vaccine 2004,22:2430-2437.
70. Stanley AC, Buxton D, Innes EA, Huntley JF: Intranasalimmunisation with Toxoplasma gondii tachyzoite antigenencapsulated into PLG microspheres induces humoraland cell-mediated immunity in sheep. Vaccine 2004,22:3929-3941.
71. Keegan ME, Whittum-Hudson JA, Mark Saltzman W: Biomimeticdesign in microparticulate vaccines. Biomaterials 2003,24:4435-4443.
Current Opinion in Biotechnology 2004, 15:530–537
www.elsevier.com/locate/jconrel NE
DELIV
ER
Y
Journal of Controlled Releas
GE
Immobilization of plasmid DNA in bacterial ghosts
Peter Mayrhofera,b,c, Chakameh Azimpour Tabrizia, Petra Walchera,b,
Wolfgang Haidingera,b, Wolfgang Jechlingera,b,c,*, Werner Lubitza,b
aInstitute of Microbiology and Genetics, Section Microbiology and Biotechnology, University of Vienna, UZA II, 2B522,
Althanstrasse 14, A-1090 Wien, AustriabBIRD-C GmbH and CoKEG, Schonborngasse 12, A-1080 Vienna, Austria
cMayrhofer and Jechlinger OEG, Strozzigasse 38/12, A-1080 Vienna, Austria
Received 8 July 2004; accepted 21 October 2004
Available online 14 November 2004
Abstract
The development of novel delivery vehicles is crucial for the improvement of DNA vaccine efficiency. In this report, we
describe a new platform technology, which is based on the immobilization of plasmid DNA in the cytoplasmic membrane of a
bacterial carrier. This technology retains plasmid DNA (Self–Immobilizing Plasmid, pSIP) in the host envelope complex due to
a specific protein/DNA interaction during and after protein E-mediated lysis. The resulting bacterial ghosts (empty bacterial
envelopes) loaded with pDNA were analyzed in detail by real time PCR assays. We could verify that pSIP plasmids were
retained in the pellets of lysed Escherichia coli cultures indicating that they are efficiently anchored in the inner membrane of
bacterial ghosts. In contrast, a high percentage of control plasmids that lack essential features of the self-immobilization system
were expelled in the culture broth during the lysis process. We believe that the combination of this plasmid immobilization
procedure and the protein E-mediated lysis technology represents an efficient in vivo technique for the production of non-living
DNA carrier vehicles. In conclusion, we present a bself-loadingQ, non-living bacterial DNA delivery vector for vaccination
endowed with intrinsic adjuvant properties of the Gram-negative bacterial cell envelope.
D 2004 Elsevier B.V. All rights reserved.
Keywords: Bacterial ghosts; Self-immobilization; DNA carrier vehicle; DNA delivery system; Gene transfer
0168-3659/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jconrel.2004.10.026
* Corresponding author. Institute of Bacteriology, Mycology
and Hygiene, Department of Pathobiology, University of Veterinary
Medicine, Veterin7rplatz 1, A-1210, Vienna, Austria. Tel.: +43 1
25077 2104; fax: +43 1 25077 2190.
E-mail address: [email protected]
(W. Jechlinger).
1. Introduction
The ability of DNA vaccination to induce immune
response has been shown in a range of infectious
disease models [1]. However, using naked DNA, it
has become apparent that high doses and/or multiple
immunizations are required to induce immune
responses in larger animals and humans [2,3]. There-
fore much effort is now focused on increasing the
e 102 (2005) 725–735
P. Mayrhofer et al. / Journal of Controlled Release 102 (2005) 725–735726
GEN
ED
ELIV
ER
Y
efficiency of DNA vaccines. The various approaches
being taken to improve the potential of this emerging
technology are: (i) modification of vector DNA to
increase expression levels or to affect the type of
immune response [4], (ii) use of adjuvants to enhance
the immune response elicited by DNA vaccines [5,6],
and (iii) development of new delivery systems for
specific targeting and/or better DNA uptake [7–9]. As
the major role of antigen presenting cells (APCs) in
vaccination strategies became clear, the latter
approach especially deals with the recruitment of
APC’s to the application site or with the increase of
the transfection efficiency [10,11].
Delivery of vaccine antigens using bacterial
carriers has resulted in the elicitation of effective
humoral and cellular immune response due to the
adjuvant properties of these vector systems [7]. In this
report, we describe the development of a new, non-
living bacterial DNA delivery system based on the
bacterial ghost technology. Bacterial ghosts are empty
envelopes of Gram-negative bacteria, which are
produced by controlled expression of the cloned lysis
gene E of bacteriophage fX174 [12]. Expression of
gene E results in the formation of a protein E specific
transmembrane tunnel structure [13]. During this
process the cytoplasmic content is expelled through
this lysis tunnel due to the osmotic pressure difference
between the cytoplasm and the culture medium [13].
The bacterial ghosts retain the structural integrity of
native cell envelopes therefore representing excellent
vaccine candidates endowed with the intrinsic adju-
vant properties. In the extended bacterial ghost
system, heterologous proteins are either mixed with
ghosts, exported into the periplasmic space, or they
are attached to the cytoplasmic membrane via specific
hydrophobic membrane anchor-peptides [14–16]. In
addition to proteins, this system can be used for the
packaging of drugs, nucleic acids or other compounds
for numerous applications [9,16,17].
In this study we developed a novel technique to
immobilize plasmid DNA to the inner membrane of
bacterial ghosts. This plasmid DNA (the Self-Immo-
bilizing Plasmid, pSIP) carries a tandem repeat of a
modified lactose operator sequence (the lacOs sites),
which is recognized by a fusion protein composed of
the repressor of the lactose operon (LacI) and a
hydrophobic membrane anchoring sequence (LV)[14,15] derived from the lysis protein of phage
MS2. The LacI-LVfusion protein produced from pSIP
is immobilized in the cytoplasmic membrane of
Escherichia coli via the hydrophobic sequences of
the truncated lysis protein (LV), whereby the LacI
repressor domain simultaneously binds to the lacOs
elements on the pSIP. During the lysis process most of
the cytoplasmic proteins and nucleic acids are
expelled through the E-specific lysis tunnel but the
anchored plasmid DNA is retained in the bacterial cell
envelopes.
This novel platform technology is designed to
combine three essential steps in production of non-
living bacterial vaccine vectors for gene delivery: (i)
the preparation of a non-living bacterial delivery
vehicle with adjuvant properties, (ii) the amplification
of plasmid DNA, and (iii) the loading of this non-
living bacterial vector with plasmid DNA.
Using the SIP technology these three steps can be
performed in vivo in a single cost-efficient process
using the appropriate bacterial carrier co-transformed
with an E-specific lysis vector and the self-immobi-
lizing plasmid. The proof of principle of the SIP
technology has been demonstrated in E. coli by
comparing the supernatant and pellet fractions of
lysed bacterial cultures, which were transformed
either with the pSIP or two control plasmids, each
lacking an essential feature of the immobilization
system. Quantitative analysis using real-time PCR
reveals that only a negligible percentage of the SIP-
DNA is expelled during the lysis process, whereas
about 50% of control plasmids can be found in the
supernatants of the bacterial lysates.
2. Materials and methods
2.1. Bacterial strains, plasmids and growth conditions
Plasmids pAWJ [18], pBAD24 [19], pPHB-LV[20]and pBBR122 [21] as well as E. coli strains MC4100
[22] and MG1655 [23] have been described. Plasmid
pREP-4 was purchased from QIAgen (Hilden,
Germany). Bacteria were grown in Luria broth
(LB) supplemented with ampicillin (200 Ag/ml),
and kanamycin (50 Ag/ml) as required. For gene
expression from vectors derived from pBAD24 the
medium was supplemented with 0.5% l-arabinose
final concentration. Expression of the lysis gene E
P. Mayrhofer et al. / Journal of Controlled Release 102 (2005) 725–735 727
GEN
ED
ELIV
ER
Y
from vector pKLys36 was achieved by a temperature
shift from 35 8C to 42 8C. Growth and lysis of
bacterial cells was monitored by measuring the
optical density at 600 nm (OD600 nm). To determine
the E-mediated inactivation of bacterial cultures,
colony forming unit counts were performed as
previously described [24].
2.2. DNA manipulations
Preparation of plasmid DNA and isolation of DNA
fragments was carried out using kits from QIAgen
(PCR Purification Kit and Gel Extraction Kit) and
Peqlab (E.Z.N.A. Plasmid Miniprep Kit I; PEQ-
LAB—Biotechnologie, Erlangen, Germany). Trans-
formation of bacterial strains and electrophoresis of
DNA were performed as described previously [25]. If
not otherwise stated, restriction enzymes, DNA-
modifying enzymes and nucleotides were obtained
from New England Biolabs (Frankfurt am Main,
Germany) or Roche Diganostics (Vienna, Austria).
The Pfu ploymerase used for PCR reactions was
purchased from Promega (Mannheim, Germany) and
used as specified by the manufacturer. For real time
PCR analysis, Dynazyme polymerase (Finnzymes Oy,
Espoo, Finland) and SYBR-Green I (Molecular
probes; Invitrogen, Lofer, Austria) were used accord-
ing to the instructions of the manufacturers.
2.3. Construction of plasmid pLacOs
A 306 bp PCR fragment containing an unrelated
spacer sequence derived from the archea phage
fCH1 [26] flanked by lacOs sites (underlined
sequence in primers lacos5 and lacos3) was
generated by PCR amplification using plasmid
pQE32a/h (a kind gift of Dr. Angela Witte;
unpublished data) as template and primers lacos5
(5V-CAGCAGATCGATAATTGTGAGCGCTCA-
CAATTGGAACTCAATACGACGGC-3V) and lacos3
(5V-CTGCTGATCGATAATTGTGAGCGCTCA-
CAATTAGCCGTGCCGGAGTA-3V) to introduce
ClaI restriction sites at the termini. The PCR reaction
containing 0.01 Ag/Al primer DNA, 0.2 mM dNTPs,
template DNA, 0.05 U/Al Pfu polymerase in Pfu
polymerase buffer was subjected to the following
conditions: 3 min 94 8C pre-denaturation, 30 cycles:
30 s 94 8C, 30 s 55 8C, 1 min 72 8C. The PCR
fragment was digested with ClaI and subsequently
cloned in the corresponding single site of the plasmid
pBAD24 resulting in the vector pLacOs.
2.4. Construction of plasmid pLacOsI
Using pREP-4 as template a 1118 bp PCR fragment
containing the lacI gene was obtained by PCR
amplification. Oligonucleotides laci5 (5V-CAGCAGC-CATGGGTAAACCAGTAAC GTATACGATGTC-3V)and laci3 (5V-TGCTGCCTGCAGCTGCTGTCATC-TAGACTGCC CGCTTTCCA-3V) containing NcoI
(laci5) and XbaI/PstI (laci3) as terminal restriction
sites were used as primers. As the NcoI site was used,
the lacI sequence starts with ATG as the first codon
thereby removing the original lacI start codon GGT.
The laci3 primer contains a stop codon in between the
XbaI and PstI sites. Amplifying the lacI sequence with
these primers resulted in a sequence with an additional
serine and arginine codon (derived from the XbaI site)
at the 3V-end. The PCR reaction was performed using
the Expand Long Template PCR System from Roche
according to the instruction of the manufacturer. The
reaction mixture was subjected to the following
conditions: 2 min 94 8C pre-denaturation, 30 cycles:
30 s 94 8C, 30 s 53 8C, 2 min 68 8C and finally 5 min
elongation at 68 8C. The PCR fragment was digested
with NcoI and PstI and subsequently cloned in the
corresponding single sites of plasmid pLacOs resulting
in vector pLacOsI. The vector pLacOsI contains the
lacI gene under the control of the araB promoter.
2.5. Construction of plasmid pSIP
A 196 kb fragment containing the membrane
anchoring sequence, encoding the 56 C-terminal amino
acids of the lysis protein L derived from the phage
MS2, was generated by PCR amplification. Plasmid
pPHB-LVwas used as template and primers ms2l5 and
ms2l3 to introduce terminal XbaI and PstI restriction
sites, respectively: ms215 (5V-CAG CAGTCTA-
GAGGGCCATTCAAACATGA-3V); ms2l3 (5V-TGCTGCCTGCAGTTAAGTA AGCAATTGCTG-
TAAAGTC-3V). PCR was performed as described
above (see Construction of plasmid pLacOs) except
for the annealing temperature, which was 53 8C. Theresulting PCR fragment was digested with XbaI and
PstI and subsequently cloned in the corresponding
P. Mayrhofer et al. / Journal of Controlled Release 102 (2005) 725–735728
GEN
ED
ELIV
ER
Y
single sites of plasmid pLacOsI resulting in the vector
pSIP. The vector pSIP contains the lactose repressor
fused to 56 C-terminal amino acids of the lysis protein
L under control of the araB promoter.
2.6. Construction of plasmid pDSIP
The vector pDSIP was constructed by digesting
pSIP with ClaI. The vector was eluted from an agarose
gel to remove the fragment containing the lacOs sites
and the spacer sequence. The eluted plasmid fragment
was religated resulting in the vector pDSIP.
2.7. Construction of plasmid pKLys36
A 1.98 kb fragment containing an E-specific lysis
cassette (cI857 repressor-EpRmut-gene E) of plasmid
pAWJwasgeneratedbyPCRamplification(948C,1min;
55 8C, 1 min; 72 8C 2 min; 30 cycles). The following
primers were used to introduce NcoI restriction sites:
sense (cI-Pvu) 5V-TTCCCCCCATGGCGATCGGCT-CAATTGTTATCAGC–3V; anti-sense (E-Pvu) 5V-TTAAAAACCATGGCGATCGCTCGGTACGGT-
CAGGC–3V. The PCR fragment was cloned in the
corresponding restriction site of the broad-host range
low copy number plasmid pBBR122. The resulting
plasmid pKLys36 contains the E-specific lysis cassette,
a broad-host rangeoriginof replication andakanamycin
resistancegene.
2.8. DNA precipitation from culture broth and pellet
of lysed E. coli cultures
Twenty-milliliter samples were taken at each time
point (onset and end of lysis). These samples were
centrifuged at 10,000�g for 15 min. The supernatants
of the lysed culture were collected in a SS-34
centrifuge tube, whereas the pellets were washed three
times with 500 Al of 10 mM Tris–Cl; pH 8.00. After
washing the pellet, the supernatant of the washing step
was added to the previously collected culture super-
natant samples. Plasmid DNA from the washed pellets
was isolated using the Peqlab E.Z.N.A Plasmid
Miniprep Kit I and plasmid DNA of the supernatant
sample were extracted using the CTAB method as
previously described [27] with some modifications.
Briefly, the supernatant was filtered using 0.22-Amfilter to avoid contamination with cellular debris
before 2.5 ml of 5% CTAB (w/v in 0.5 N NaCl) was
added to 20 ml supernatant of the lysates. After
centrifugation at 10,000�g for 15 min at 4 8C, thepellet was resuspended in 400 Al 1.2 N NaCl and
transferred to a 1.5-ml tube. The DNAwas precipitated
with 2.5 volumes of absolute Ethanol, washed with 1
ml 70% EtOH and air-dried. To remove the bulk of
chromosomal DNA, the preparation was further
purified with QIAgen Extraction kit according to the
instructions of the manufacturer.
The amount of DNA recovered from the super-
natant using the CTAB method was evaluated.
Duplicates of six DNA standards in 20 ml LB were
extracted using the CTAB method and quantified by
real-time PCR in duplicates. This revealed that
93F3.2% of the DNA could be recovered using this
CTAB protocol. The evaluation of the Peqlab mini
preparation kit was performed as well by parallel
extractions of six DNA standards in duplicates
followed by quantification in duplicates. This revealed
that 84.6F12.7 of the loaded DNA was extracted.
2.9. Quantification of plasmid DNA using real time
PCR
A real time PCR approach was optimized to
quantify the amount of pKLys36, pSIP, pDSIP and
pLacOsI in the pellet as well as in the supernatant of
lysed cultures. The template DNA was prepared from
supernatants and pellets of lysed bacterial cultures as
described above. Real time PCR was performed in
200 Al reaction tubes containing 200 AM dNTP, 1
AM of each primer, 2.5 Al of polymerase buffer
(10�), 5 Al of 1:100 diluted extracted DNA as
template, 0.25 Al SYBR-Green I and 0.25 AlDynazyme polymerase (2 U/Al) in a final volume
of 25 Al. Primers binding to the ampicillin resistance
gene (fwd: 5V-ATGAGTATTCAACATTTCCGTGTC-3V; rev: 5V-TTACCAATGCTTAATCAGTGAGG-3V)of plasmids pSIP, pDSIP and pLacOsI, respectively,
were used to amplify a 860 bp PCR fragment. To
quantify the plasmid pKLys36, a pair of primers was
designed to amplify gene E as follows: Fwd: 5VGCTGGACTTGGGATAC-3V, rev: 5V-GACATTACAT-CACTCCTTCTGC-3V. Tenfold serial dilutions
(10�2–10�6) of the plasmid pLacOsI, which was
prepared using a DNA minipreparation kit (Peqlab)
were used as the standard for reactions amplifying
P. Mayrhofer et al. / Journal of Controlled Release 102 (2005) 725–735 729
GEN
ED
ELIV
ER
Y
the ampicillin resistance gene. As with the pLacOsI
standard, tenfold serial dilutions (10�2–10�6) of the
plasmid pKLys36 were used as standard for reactions
amplifying gene E. For all PCR reactions the
following conditions were applied: 1 min 94 8C, 1min 60 8C and 1 min 72 8C (30 cycles). The data
shown in Fig. 4 represents an average of 3 measure-
ments for each construct. Samples for real time PCR
analysis were taken at the first time-point when the
OD600 of the cultures was decreasing (onset of lysis)
or at the last time-point measured (end of lysis). The
Rotor Gene 2000, real time Cycler (Corbett
Research, Mortlake, Australia) was used as specified
by the manufacturer. Data had been analysed with
special software, supplied by the manufacturer.
2.10. SDS-PAGE and Western blotting
Protein samples were heated to 95 8C for 5 min in
sample buffer and separated on a 12% SDS-PAGE
gel according to Laemmli [28]. Proteins were trans-
ferred to nitrocellulose membranes by semidry
electro blotting. Membranes were blocked for 1 h
in 5% low fat milk powder in TBS. Proteins were
detected either with polyclonal rabbit antiserum to
the Lac Repressor or to the MS2 lysis protein L. The
a-LacI antiserum was purchased from Stratagene
(Strategene Europe, Amsterdam, The Netherlands),
the a-MS2 serum was a kind gift from Joachim-
Volker Hfltje (Max-Planck-Institute for Evolutionary
Biology in Tqbingen, Germany). Anti Lac Repressor
serum was diluted 1:3000, anti L serum was diluted
1:350 in TBS, 0.5% BSA, 0.05% NaN3. Alkaline
phosphatase coupled to goat anti-rabbit antibodies
from Sigma was used in a 1:5000 dilution. The
membranes were incubated for 1 h with the
respective antiserum dilution. To stain the antigen–
antibody complex, BCIP and NTB from Roche in
alkaline phosphatase buffer were used as recom-
mended by the supplier. Antisera were pre-treated
with acetone powder from the appropriate E. coli
strain as described previously to avoid unspecific
binding [29].
2.11. Preparation of membrane fractions from E. coli
Five-hundred-milliliter cultures of E. coli MC4100
transformed either with pLacOsI or pSIP were grown
at 28 8C. Expression of the repressor protein LacI
from plasmid pLacOsI as well as the LacI-LV fusionprotein from plasmid pSIP was induced with l-
arabinose at an OD600 of 0.3. Cells were harvested
after 2 h of induction. Membrane fractions were
prepared according to Schnaitman [30,31] with a
slight modification. Briefly, prior to cell disruption,
the buffer was supplemented with the complete mini
protease inhibitor cocktail from Roche, according to
the manufacturer’s instructions. The cells were
opened by two times passing through the French
Press. The membranes were subsequently collected
by high-speed centrifugation (1 h, 4 8C, 105,000�g).
The proteins in the supernatant fraction of the
centrifugation step were considered as the soluble
cytoplasmic protein fraction. The pellet after high-
speed centrifugation, the membrane fraction, was
further treated according to Schnaitman to separate
the inner (cytoplasmic) and outer membrane fraction.
After separation a 1-ml aliquot of the inner mem-
brane band in the sucrose gradient was collected for
further analysis.
3. Results
3.1. Expression and membrane anchoring of protein
LacI-LV
Vectors pLacOsI and pSIP (Fig. 1) were analyzed
for their ability to express the cloned LacI or the LacI-
LVfusion proteins in E. coli MC4100, respectively. To
determine whether the fusion protein is anchored in
the cytoplasmic membrane, membrane fractions were
prepared after protein expression and subjected to
SDS-PAGE and Western Blot analysis. Aliquots of the
inner membrane fraction, the cytoplasmic protein
fraction as well as whole bacterial cells were
analyzed. The expressed proteins were detected with
both, antibodies to the LacI protein as well as
antibodies to the LV-peptide part of the fusion protein.
There is no detectable signal when proteins expressed
from E. coli MC4100 (pLacOsI) were analyzed using
anti-LV antibodies (Fig. 2A, lanes 1–3). Analysis of
samples prepared from E. coli MC4100 (pSIP)
revealed a signal at the molecular weight expected
for the LacI-LVhybrid protein. This 45 kDa protein
can be detected in whole bacterial cells, in the inner
Fig. 1. The Self-Immobilizing Plasmid pSIP (A) and the control plasmids each lacking essential features for immobilization pDSIP (B) and
pLacOsI (C), respectively. Ap, ampicillin resistance gene; ColE1, origin of replication; lacI, lac repressor gene; LV, amino acids 21–75 of MS2 L
protein; PBAD, arabinose-inducible promoter; araC, repressor/inducer of the PBAD promoter; Os, lac operator sites with a tenfold higher binding
affinity for the lac repressor than the wild-type operator sites; spacer, unrelated spacer sequence derived from the archea phage fCH1.
P. Mayrhofer et al. / Journal of Controlled Release 102 (2005) 725–735730
GEN
ED
ELIV
ER
Y
membrane fraction as well as in the cytoplasmic
fraction (see Fig. 2A, lanes 4–6).
Aliquots of the same samples were applied for
protein detection with anti-LacI antiserum (Fig. 2B).
The LacI protein expressed from MC4100 (pLacOsI)
can be detected at the expected molecular weight of 38
kd in whole bacterial cells (Fig 2B, lane 1), in the
cytoplasmic fraction (Fig. 2B, lane 3) but not in the
inner membrane fraction (Fig. 2B, lane 2). As expected
Fig. 2. Western blot analysis for the detection of LacI-LVwith anti-L antiser
MC4100 (pLacOsI) whole cells; lane 2, E. coli MC4100 (pLacOsI) inner
fraction; lane 4, E. coli MC4100 (pSIP) whole cells; lane 5, E. coli MC41
cytoplasmic fraction.
analysis of the LacI-LVsamples derived from E. coli
MC4100 (pSIP) using anti-LacI antiserum revealed the
same signals as when detection was performed with
anti-LV-antiserum (Fig. 2A, lanes 4–6).
The Western blot analyses show that the LacI-LVfusion protein, but not the lactose repressor alone, can
be found in the membrane fraction, indicating that the
fusion protein is associated with the cytoplasmic
membrane via the LVsequence.
um (A) and with anti-LacI antiserum (B), respectively: lane 1, E. coli
membrane fraction; lane 3, E. coli MC4100 (pLacOsI) cytoplasmic
00 (pSIP) inner membrane fraction; lane 6, E. coli MC4100 (pSIP)
P. Mayrhofer et al. / Journal of Controlled Release 102 (2005) 725–735 731
GEN
ED
ELIV
ER
Y
3.2. Self-immobilization of plasmid pSIP and
quantification of DNA by real time PCR
A schematic system description is provided in
Fig. 3A. Co-transformed E. coli MC4100 harboring
the lysis plasmid pKLys36 and the vectors pLacOsI,
pDSIP or pSIP, respectively, were monitored for the
lysis and immobilization behavior. Control plasmids
pDSIP and pLacOsI should not be immobilized in the
cytoplasmic membrane as they lack the lacI recog-
nition sites or the LV anchor sequence, respectively.
Protein expression from the araBAD promoter was
induced with l-arabinose at an optical density of
about OD600 0.1. Induced cells were grown to an
OD600 of about 0.3 and subsequently shifted to 42
8C to induce expression of the lysis protein (Fig.
3B). Expression of the LacI-LVhybrid protein for 75
min in MC4100 (pSIP) and MC4100 (pDSIP) does
not significantly influence the lysis behavior com-
pared to MC4100 (pLacOsI) (Fig. 3B). Analyzing
Fig. 3. Schematic system description (A) and growth/lysis experiments with
pKLys36 (B). (A) Cells transformed with pSIP and the lysis plasmid pK
promoter. The hybrid protein integrates in the bacterial membrane and bind
inactivate the cI857 repressor for lysis induction. The cytoplasmic content i
the bacterial ghost. (B) Growth and lysis of E. coli MC4100 (pKLys36) co-
by arrows cells were grown at 35 8C to an OD600 of approximately 0.1 fo
shifted to 42 8C for induction of lysis. Open symbols indicate the ghost f
CFU counts revealed that at least 99.9% of bacteria
were killed by protein E mediated lysis.
DNA was precipitated from the culture super-
natant as described in Materials and methods. The
control vectors pDSIP and pLacOsI could be precipi-
tated from the culture supernatant after lysis, whereas
no visible band was found when a corresponding
volume of the supernatant preparation derived from a
lysed MC4100 (pSIP) culture was applied on the
agarose gel (data not shown). For the initial experi-
ments E. coli MC4100, which is a D(arg-lac)U169
strain, was chosen to avoid interference with the
endogenous lac operon. Performing the same experi-
ments with an E. coli wild type strain (MG1655) led
to the same results as with MC4100 suggesting that
the principle of self-immobilization is independent
from the genetic background of the host. These
findings indicate that the plasmid pSIP is immobilized
in the cytoplasmic membrane of the bacterial ghosts,
whereas the control plasmids are expelled in the
E. coli MC4100 harbouring pSIP or control plasmids together with
Lys36 are grown at 35 8C and LacI-LVis expressed from the PBADs to lacOs sites on the pSIP vector. The culture is shifted to 42 8C to
ncluding the lysis plasmid is expelled. The pSIP vector is retained in
transformed with pLacOsI (E), pSIP (n) or pDSIP (x). As indicatedr induction of protein expression with l-arabinose and subsequently
ormation.
P. Mayrhofer et al. / Journal of Controlled Release 102 (2005) 725–735732
GEN
ED
ELIV
ER
Y
culture broth during the lysis process. However, to
give clear evidence of the self-immobilization these
DNA preparations were further analyzed. The amount
of pSIP DNA and control plasmids in the pellet and in
the supernatant of bacterial lysates were quantified by
real time PCR. To analyze the general ratio of pDNA
expelled to the culture broth and plasmids unspecifi-
cally retained after protein E mediated lysis, real-time
PCR was performed with samples derived from E.
coli MC4100 (pKLys36) at first. As it is shown in Fig.
4, row 4, by onset of the lysis process, only low
amounts of plasmid pKLys36 are detectable, whereas
at the end of lysis process around 50% of the lysis
plasmid were expelled. The same experiments were
performed with E. coli MC4100, which harbored the
lysis plasmid together with either pSIP (Fig. 4, row 1)
or pLacOsI (Fig. 4, row 2) or pDSIP (Fig. 4, row 3).
Comparable to pKLys36, about 45% of the control
plasmid pDSIP and pLacOsI were lost during the lysis
process, whereas the loss of immobilized plasmid
Fig. 4. Quantification of plasmid DNA in the pellet and in the
supernatant of lysed E. coli MC4100 cultures by real time PCR.
100% represents the total amount of plasmid DNA measured in the
pellet and supernatant of the corresponding lysates. Row 1, E. coli
MC4100 (pSIP, pKLys36); row 2, E. coli MC4100 (pLacOsI,
pKLys36); row 3, E. coli MC4100 (pDSIP, pKLys36); row 4, E. coli
MC4100 (pKLys36). Dark-gray bars represent the percentage of
plasmid DNA measured in the pellets of lysed cells, whereas white
bars indicate the percentage of plasmid DNA detected in the
supernatants of lysates. Samples for real time PCR analysis were
taken at the first time-point when the OD600 of the cultures was
decreasing (onset of lysis) or at the last time-point measured (end of
lysis) (Fig. 3). The error bars indicate the standard deviation of three
individual experiments.
pSIP was on average 7%. We calculated that roughly
10 Ag pSIP DNAwas immobilized per mg of bacterial
ghosts.
4. Discussion
Immobilization of the target sequence in the
cytoplasmic membrane of a bacterial delivery vector
is the key element of this novel platform technology.
The combination of the SIP technology and protein E
mediated lysis results in a bself-loadingQ, non-livingbacterial DNA carrier system. Using the SIP, the target
DNA can be amplified and retained in the bacterial
ghosts, while the bulk of the cytoplasmic content is
expelled into the culture broth during the lysis
process.
The central element of the SIP technology is the
LacI-LV hybrid protein. Results indicate that the
fusion of the highly hydrophobic membrane span-
ning sequence (56 amino acids) used in this study
does not impair the DNA binding functionality of the
lactose repressor protein. The results of the Western
Blot analyses show that LacI-LV is associated with
the inner membrane solely due to its hydrophobic
sequence. Since the LacI-LVfusion protein proves to
be functional, we conclude that the membrane
anchoring fragment does not interfere with the
DNA binding properties of the repressor protein
and vice versa (e.g. by inducing conformational
changes in the DNA binding domain due to
improper folding or by inhibiting the essential
oligomerization).
The LacI-LVhybrid has also been detected in the
cytoplasmic fraction after high-speed centrifugation,
indicating that the hybrid is still soluble although
the repressor protein was fused to a highly hydro-
phobic sequence. This might result from the fact
that a 38.5 kDa protein was fused to a 6.4 kDa
peptide and that the larger protein simply plays the
dominant role in terms of physico-chemical charac-
teristics including solubility. The harsh mechanical
disruption of bacterial cells by two times passing
through the French Press for membrane fraction
preparation (described in Sections 2.11. and 3.1)
completely destroys the integrity of the membrane
system, thereby releasing associated proteins in the
cytoplasmic fraction. Apart from the formation of a
P. Mayrhofer et al. / Journal of Controlled Release 102 (2005) 725–735 733
GEN
ED
ELIV
ER
Y
trans-membrane tunnel, the protein E mediated lysis
process does not alter the bacterial envelope and will
therefore not release that much, if any immobilized
hybrid proteins.
The repressor of the lactose operon was chosen
because the structural and functional stability of this
molecule has been demonstrated in various studies,
e.g. (i) it has been shown that fusions to the C-
terminus of LacI do not alter the DNA binding
properties of the protein even if the fusion protein is
as large as h-galactosidase [32], (ii) several mod-
ifications of this repressor had been performed
without significant loss of the DNA binding
capability [33–35] and (iii) it has been demonstrated
that plasmid expressed LacI fusion proteins bind to
the vector encoding them, when the lactose operator
sequence is integrated in this construct [36,37].
Apart from its functional stability, the high stability
of the protein/DNA complex is another advanta-
geous feature of the lactose operator/repressor
system. The repressor tetramer has two DNA
binding sites and exhibits strong cooperative bind-
ing to DNA molecules containing two suitably
spaced operator sequences, thereby forcing the
intervening DNA into a loop structure [38,39]. This
complex is further stabilized using a symmetric
variant of the wild type operator with a tenfold
higher affinity, the lacOs site [40,41].
Pellet as well as supernatant DNA preparations
corresponding to an identical volume of culture broth
was subjected to real-time PCR analysis. The DNA
content of both fractions calculated on basis of the
real-time PCR assays was summarized and defined as
100%. Therefore the calculations made are based on
isolated DNA only, i.e. losses during the preparation
are not considered. Furthermore, different methods
have been applied to isolate DNA from the culture
broth (CTAB method) and the lysed bacterial pellet
(Peqlab kit). Therefore, the calculated ratio of plasmid
DNA might not properly reflect the exact ratio in the
culture broth and the pellets, even though the
comparison of the methods revealed very similar
yields if they are used to isolate DNA from the
supernatant or the bacterial ghost pellet, respectively.
Although the bacterial ghost pellets have been
washed several times, it cannot be excluded that
there is still plasmid DNA unspecifically associated
with these cell envelopes. Evaluation of the lysis
efficiency revealed only 0.1% unlysed cells indicating
only a negligible contribution of these survivors to
the overall result. Based on these findings it is
probably not possible to exactly determine the
percentage of control plasmid DNA (pLacOSI,
pDSIP) retained in bacterial ghosts compared to a
DNA content of 100%, which is the number of
plasmids in intact (unlysed) bacterial cells. However,
the experimental setup for this study and the
calculations based on it were designed to answer
the question if there is more pSIP retained in bacterial
ghosts than control plasmids. As the conditions
mentioned above are similar for each investigated
plasmid, the results for the different control plasmids
and the pSIP can naturally be compared to each other.
The significant higher level of retained pSIP clearly
demonstrates the efficiency of the self-immobilization
system. It has been shown that there is still residual
genomic DNA as well as unspecifically associated
plasmid DNA in ghost preparations, if not treated
with a nuclease [24]. However, many immunization
studies have been performed with bacterial ghosts
[9,42–44]. In none of these studies there has been
evidence for a negative effect of the residual DNA on
the immune response.
Recently it has been demonstrated that bacterial
ghosts are excellent vehicles for the delivery of in
vitro loaded plasmid DNA [9]. They showed high
efficiency in the transfection of macrophages and
primary dendritic cells (52% to 60%). Ghost-mediated
DNA delivery resulted in the elicitation of more
efficient humoral and cellular immune responses than
using equal amounts of naked DNA [9]. At present
there is no evidence whether the strong interaction
between the membrane-anchored LacI protein and the
plasmid DNA hampers or even improves immune
responses. Immunization studies are planned which
will compare ghosts mediated immune responses
against target antigens encoded on pSIP or on
conventional in vitro loaded plasmids. To formulate
a non-living DNA delivery system, normally the
production of the carrier, the pDNA as well as the
loading of the carrier with the pDNA are separate,
laborious and therefore cost intensive processes. In
this report we demonstrated that the SIP system is able
to combine these essential steps in a single in vivo
process, resulting in a non-living bacterial DNA
delivery vehicle.
P. Mayrhofer et al. / Journal of Controlled Release 102 (2005) 725–735734
GEN
ED
ELIV
ER
Y
Appendix A. Supplementary data
Supplementary data associated with this article can
be found, in the online version, at doi:10.1016/
j.jconrel.2004.10.026.
References
[1] W.C. Lai, M. Bennett, DNA vaccines, Crit. Rev. Immunol. 18
(1998) 449–484.
[2] S. van Drunen Littel-van den Hurk, V. Gerdts, B.I. Loehr, R.
Pontarollo, R. Rankin, R. Uwiera, L.A. Babiuk, Recent
advances in the use of DNA vaccines for the treatment of
diseases of farmed animals, Adv. Drug Deliv. Rev. 43 (2000)
13–28.
[3] S.A. Johnston, A.M. Talaat, M.J. McGuire, Genetic immuni-
zation: what’s in a name? Arch. Med. Res. 33 (2002)
325–329.
[4] S. Manoj, L.A. Babiuk, S. van Drunen Littel-van den Hurk,
Approaches to enhance the efficacy of DNA vaccines, Crit.
Rev. Clin. Lab. Sci. 41 (2004) 1–39.
[5] J.Y. Scheerlinck, Genetic adjuvants for DNA vaccines,
Vaccine 19 (2001) 2647–2656.
[6] A.D. Cohen, J.D. Boyer, D.B. Weiner, Modulating the immune
response to genetic immunization, FASEB J. 12 (1998)
1611–1626.
[7] E. Medina, C.A. Guzman, Use of live bacterial vaccine vectors
for antigen delivery: potential and limitations, Vaccine 19
(2001) 1573–1580.
[8] M.D. Brown, A.G. Sch7tzlein, I.F. Uchegbu, Gene delivery
with synthetic (non viral) carriers, Int. J. Pharm. 229 (2001)
1–21.
[9] T. Ebensen, S. Paukner, C. Link, P. Kudela, C. de Domenico,
W. Lubitz, C.A. Guzman, Bacterial ghosts are an efficient
delivery system for DNA vaccines, J. Immunol. 172 (2004)
6858–6865.
[10] J.J. Donnelly, M.A. Liu, J.B. Ulmer, Antigen presentation and
DNA vaccines, Am. J. Respir. Crit. Care Med. 162 (2000)
S190–S193.
[11] J.B. Ulmer, G.R. Otten, Priming of CTL responses by DNA
vaccines: direct transfection of antigen presenting cells versus
cross-priming, Dev. Biol. (Basel) 104 (2000) 9–14.
[12] M.P. Szostak, A. Hensel, F.O. Eko, R. Klein, T. Auer, H.
Mader, A. Haslberger, S. Bunka, G. Wanner, W. Lubitz,
Bacterial ghosts: non-living candidate vaccines, J. Biotechnol.
44 (1996) 161–170.
[13] A. Witte, G. Wanner, U. Bl7si, G. Halfmann, M. Szostak, W.
Lubitz, Endogenous transmembrane tunnel formation medi-
ated by phi X174 lysis protein E, J. Bacteriol. 172 (1990)
4109–4114.
[14] M. Szostak, G. Wanner, W. Lubitz, Recombinant bacterial
ghosts as vaccines, Res. Microbiol. 141 (1990) 1005–1007.
[15] M.P. Szostak, T. Auer, W. Lubitz, Immune response against
recombinant bacterial ghosts carrying HIV-1 reverse tran-
scriptase, in: F. Brown, R.M. Chanock, H.S. Ginsberg, R.A.
Lerner (Eds.), Modern Approaches to New Vaccines Including
Prevention of AIDS, Cold Spring Harbor Laboratory Press,
New York, 1993, pp. 419–425.
[16] W. Jechlinger, W. Haidinger, S. Paukner, P. Mayrhofer, E.
Riedmann, J. Marchart, U. Mayr, C. Haller, G. Kohl, P.
Walcher, P. Kudela, J. Bizik, D. Felnerova, E.M.B. Denner, A.
Indra, A. Haslberger, M. Szostak, S. Resch, F. Eko, T.
Schukovskaya, V. Kutyrev, A. Hensel, S. Friederichs, T.
Schlapp, W. Lubitz, Bacterial ghosts as carrier and targeting
systems for antigen delivery, in: G. Dietrich, W. Goebel (Eds.),
Vaccine Delivery Strategies, Horizon Scientific Press,
Wymondham, 2002, pp. 163–184.
[17] S. Paukner, G. Kohl, W. Lubitz, Bacterial ghosts as novel
advanced drug delivery systems: antiproliferative activity of
loaded doxorubicin in human Caco-2 cells, J. Control. Release
94 (2004) 63–74.
[18] W. Jechlinger, M.P. Szostak, A. Witte, W. Lubitz, Altered
temperature induction sensitivity of the lambda pR/cI857
system for controlled gene E expression in Escherichia coli,
FEMS Microbiol. Lett. 173 (1999) 347–352.
[19] L.M. Guzman, D. Belin, M.J. Carson, J. Beckwith, Tight
regulation, modulation, and high-level expression by vectors
containing the arabinose PBAD promoter, J. Bacteriol. 177
(1995) 4121–4130.
[20] S. Kalousek, Molekularbiologische Untersuchungen zur
rekombinanten Expression des PHB-Operons von Alcaligenes
Eutrophus in Escherichia coli, PhD Thesis, Institute of
Microbiology and Genetics, University of Vienna, Vienna,
1993.
[21] R. Antoine, C. Locht, Isolation and molecular character-
ization of a novel broad-host-range plasmid from Borde-
tella bronchiseptica with sequence similarities to plasmids
from gram-positive organisms, Mol. Microbiol. 6 (1992)
1785–1799.
[22] T.J. Silhavy, M.L. Berman, L.W. Enquist, Experiments with
Gene Fusions, Cold Spring Habor Laboratory, Cold Spring
Habor, NY, 1984.
[23] B.J. Bachmann, Derivations and genotypes of some mutant
derivatives of Escherichia coli K12, in: J.L. Ingarahan, K.B.
Low, B. Magasanik, M. Schaechter, H.E. Umbarger (Eds.),
Escherichia coli and Samonella thyphimurium: Cellular and
Molecular Biology, American Society for Micorbiology,
Washington, DC, 1987, pp. 1191–1219.
[24] W. Haidinger, U.B. Mayr, M.P. Szostak, S. Resch, W. Lubitz,
Escherichia coli ghost production by expression of lysis gene
E and Staphylococcal nuclease, Appl. Environ. Microbiol. 69
(2003) 6106–6113.
[25] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning: a
Laboratory Manual, Cold Spring Harbor Press, 1989.
[26] R. Klein, U. Baranyi, N. Rfssler, B. Greineder, H. Scholz, A.Witte, Natrialba magadii virus phiCh1: first complete
nucleotide sequence and functional organization of a virus
infecting a haloalkaliphilic archaeon, Mol. Microbiol. 45
(2002) 851–863.
[27] G. Del Sal, G. Manfioletti, C. Schneider, The CTAB-DNA
precipitation method: a common mini-scale preparation of
P. Mayrhofer et al. / Journal of Controlled Release 102 (2005) 725–735 735
GEN
ED
ELIV
ER
Y
template DNA from phagemids, phages or plasmids suitable
for sequencing, BioTechniques 7 (1989) 514–520.
[28] U.K. Laemmli, Cleavage of structural proteins during the
assembly of the head of bacteriophage T4, Nature 227 (1970)
680–685.
[29] E. Harlow, D. Lane, Antibodies, A Laboratory Manual, Cold
Spring Harbor Press, 1988.
[30] C.A. Schnaitman, Examination of the protein composition of
the cell envelope of Escherichia coli by polyacrylamide gel
electrophoresis, J. Bacteriol. 104 (1970) 882–889.
[31] C.A. Schnaitman, Protein composition of the cell wall and
cytoplasmic membrane of Escherichia coli, J. Bacteriol. 104
(1970) 890–901.
[32] B. Mqller-Hill, J. Kania, Lac repressor can be fused to beta-
galactosidase, Nature 249 (1974) 561–563.
[33] M.C. Hu, N. Davidson, Targeting the Escherichia coli lac
repressor to the mammalian cell nucleus, Gene 99 (1991)
141–150.
[34] M.A. Labow, S.B. Baim, T. Shenk, A.J. Levine, Conversion of
the lac repressor into an allosterically regulated transcriptional
activator for mammalian cells, Mol. Cell. Biol. 10 (1990)
3343–3356.
[35] N. Panayotatos, A. Fontaine, S. Backman, Biosynthesis of a
repressor/nuclease hybrid protein, J. Biol. Chem. 264 (1989)
15066–15069.
[36] M.G. Cull, J.F. Miller, P.J. Schatz, Screening for receptor
ligands using large libraries of peptides linked to the C
terminus of the lac repressor, Proc. Natl. Acad. Sci. U. S. A. 89
(1992) 1865–1869.
[37] P.J. Schatz, Use of peptide libraries to map the substrate
specificity of a peptide-modifying enzyme: a 13 residue
consensus peptide specifies biotinylation in Escherichia coli,
Biotechnology (NY) 11 (1993) 1138–1143.
[38] H. Kramer, M. Amouyal, A. Nordheim, B. Mqller-Hill, DNAsupercoiling changes the spacing requirement of two lac
operators for DNA loop formation with lac repressor, EMBO
J. 7 (1988) 547–556.
[39] H. Kramer, M. Niemoller, M. Amouyal, B. Revet, B. von
Wilcken-Bergmann, B. Mqller-Hill, Lac repressor forms loops
with linear DNA carrying two suitably spaced lac operators,
EMBO J. 6 (1987) 1481–1491.
[40] J.R. Sadler, H. Sasmor, J.L. Betz, A perfectly symmetric lac
operator binds the lac repressor very tightly, Proc. Natl. Acad.
Sci. U. S. A. 80 (1983) 6785–6789.
[41] A. Simons, D. Tils, B. von Wilcken-Bergmann, B. Mqller-Hill,Possible ideal lac operator: Escherichia coli lac operator-like
sequences from eukaryotic genomes lack the central G X C
pair, Proc. Natl. Acad. Sci. U. S. A. 81 (1984) 1624–1628.
[42] F.O. Eko, W. Lubitz, L. McMillan, K. Ramey, T.T. Moore,
G.A. Ananaba, D. Lyn, C.M. Black, J.U. Igietseme, Recombi-
nant Vibrio cholerae ghosts as a delivery vehicle for
vaccinating against Chlamydia trachomatis , Vaccine 21
(2003) 1694–1703.
[43] J. Marchart, M. Rehagen, G. Dropmann, M.P. Szostak, S.
Alldinger, S. Lechleitner, T. Schlapp, S. Resch, W. Lubitz,
Protective immunity against pasteurellosis in cattle, induced
by Pasteurella haemolytica ghosts, Vaccine 21 (2003)
1415–1422.
[44] K. Panthel, W. Jechlinger, A. Matis, M. Rohde, M. Szostak,
W. Lubitz, R. Haas, Generation of Helicobacter pylori ghosts
by PhiX protein E-mediated inactivation and their evaluation
as vaccine candidates, Infect. Immun. 71 (2003) 109–116.
ARTICLE IN PRESS
1
2
34
5
6
7
8
9
101112131415161718
19
20
212223242526
www.elsevier.com/locate/addr
DTD 5
.
Advanced Drug Delivery Rev
OFBacterial ghosts as antigen delivery vehicles
Ulrike Beate Mayra, Petra Walchera, Chakameh Azimpoura, Eva Riedmanna,b,
Christoph Hallera, Werner Lubitza,b,TaInstitute of Microbiology and Genetics, University of Vienna, A-1090 Vienna, Austria
bBIRD-C GmbH and CoKEG, Schoenborngasse 12, A-1080 Vienna, Austria
Received 31 January 2004; accepted 25 January 2005
OORRECTED PR
Abstract
The bacterial ghost system is a novel vaccine delivery system unusual in that it combines excellent natural intrinsic adjuvant
properties with versatile carrier functions for foreign antigens. The efficient tropism of bacterial ghosts (BG) for antigen
presenting cells promotes the generation of both cellular and humoral responses to heterologous antigens and carrier envelope
structures. The simplicity of both BG production and packaging of (multiple) target antigens makes them particularly suitable
for use as combination vaccines. Further advantages of BG vaccines include a long shelf-life without the need of cold-chain
storage due to their freeze-dried status, they are safe as they do not involve host DNA or live organisms, they exhibit improved
potency with regard to target antigens compared to conventional approaches, they are versatile with regards to DNA or protein
antigen choice and size, and as a delivery system they offer high bioavailability.
D 2005 Elsevier B.V. All rights reserved.
Keywords: Target antigen; Gram-negative bacterial envelope; Particle presentation technology; DNA vaccine; Adjuvant; Delivery system;
Bacterial ghosts
Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
2. Production of bacterial ghosts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
3. Bacterial ghosts as candidate vaccines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
3.1. Parenteral immunization with bacterial ghosts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
U0169-409X/$ - s
doi:10.1016/j.ad
Abbreviation
dendritic cells;
membrane; OMP
target antigen; T
T Correspondi
+43 1 406 50 9
E-mail addr
iews xx (2005) xxx–xxx
NC
ee front matter D 2005 Elsevier B.V. All rights reserved.
dr.2005.01.027
s: APC, antigen presenting cells; APP, Actinobacillus pleuropneumoniae; BG, bacterial ghosts; CPS, cytoplasmic space; DC,
GFP, green fluorescent protein; IM, inner membrane; LPS, lipopolysaccharide; MBP, maltose binding protein; OM, outer
, outer membrane protein; PBMC, peripheral blood derived monocytic cells; PPS, periplasmic space; StrpA, streptavidin; TA,
CP, toxin-co-regulated pili; VCG, Vibrio cholerae ghosts.
ng author. Institute of Microbiology and Genetics, University of Vienna, A-1090 Vienna, Austria. Tel.: +43 1 4277 54670; fax:
3.
esses: [email protected], [email protected] (W. Lubitz).
ADR-11331; No of Pages 11
ARTICLE IN PRESS
2728293031
32
33
3435363738394041424344454647484950515253545556575859606162636465666768697071
U.B. Mayr et al. / Advanced Drug Delivery Reviews xx (2005) xxx–xxx2
3.2. Induction of cytokines by bacterial ghosts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
3.3. Mucosal immunizations with bacterial ghosts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
4. Bacterial ghost system as carrier of foreign target antigens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0
T
72737475767778798081828384
85
8687888990919293949596979899100101102103104105106107108
UNCORREC
1. Introduction
Subunit vaccines composed of purified compo-
nents can be produced from many microorganisms;
however, they are often poorly immunogenic neces-
sitating an appropriate adjuvant in the vaccine
formulation. Similarly, for DNA vaccines to reach
their full potential, new vaccine delivery systems need
to be developed which better activate mucosal
immune responses.
The bacterial ghost system is one such vaccine
delivery system, combining targeting of antigen
components to antigen presenting cells (APC) and
providing the required adjuvant activity without the
need for further additions. BG are produced by protein
E-mediated lysis of Gram-negative bacteria. They are
non-living bacterial envelopes, which maintain the
cellular morphology and native surface antigenic
structures including bioadhesive properties of the
natural cell. Lipopolysaccharide (LPS) present in the
outer membrane (OM) does not limit the use of BG as
vaccine candidates because of the minimal toxicity of
cell-associated LPS compared to free LPS. The
intrinsic adjuvant properties of BG enhances T-cell
activation and systemic, mucosal and cellular immun-
ity to target antigens.
E-mediated lysis has been achieved in various
Gram-negative bacteria, including Escherichia coli
K12, enterohaemorrhagic (EHEC) and entertoxigenic
(ETEC) strains, Actinobacillus pleuropneumoniae,
Bordetella bronchiseptica, Erwinia cypripedii, Heli-
cobacter pylori, Klebsiella pneumoniae, Mannheimia
haemolytica, Pasteurella multocida, Pseudomonas
putida, Ralstonia eutropha, Salmonella typhimurium
and enteritidis strains, and Vibrio cholerae. This broad
spectrum of bacteria shows that E-mediated lysis most
probably works in every Gram-negative bacterium,
provided that the E specific lysis cassette can be
introduced into the new recipient by an appropriate
vector allowing tight repression and induction control
PROOFof lethal gene E. Although BG have been used as
vaccine candidates against their own envelope struc-
tures, a more practical use remains versatile carrier and
adjuvant vehicles for foreign target antigens of
bacterial or viral origin. As described in more detail
in the following sections, BG have the capacity for
pre-lysis localization of target antigens in or a
combination of the OM, the inner membrane (IM),
the periplasmic space (PPS) and the internal lumen of
the cytoplasmic space (CPS). The choice of antigen
compartmentalization gives the bacterial ghost system
significant potential for the challenges of constructing
subunit or DNA human and veterinary vaccines.
ED2. Production of bacterial ghosts
BG are produced by expression of cloned gene E
from bacteriophage PhiX174 resulting cell lysis in
Gram-negative bacteria. Expression of gene E can be
placed under transcriptional control of either the
thermosensitive EpL/pR-cI857 promoter, or under
chemical inducible promoter repressor systems, like
lacPO or the tol expression system [1–3]. Mutations to
the EpR promoter/operator regions have resulted in
new expression systems, which stably repress gene E
expression at temperatures of up to 37 8C, but stillallowed induction of cell lysis at a temperature range
of 39–42 8C [4]. Alternatively, by combining the EpRpromoter/cI repressor system with the lacI/lacPO for
control of gene E expression, a cold-sensitive system
for ghost formation by lowering the growth temper-
ature of the bacteria from 37 8C or higher to 28 8C or
lower has been obtained [5].
Gene E codes for a membrane protein of 91 amino
acids, which is able to fuse inner and outer mem-
branes of Gram-negative bacteria [6,7], forming an E-
specific lysis tunnel through which the cytoplasmic
content is expelled [8]. The remaining empty CPS
(internal lumen) of the bacteria is largely devoid of
T
ARTICLE IN PRESS
109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156
157158159160
161
162
163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200
U.B. Mayr et al. / Advanced Drug Delivery Reviews xx (2005) xxx–xxx 3
UNCORREC
nucleic acids, ribosomes or other constituents,
whereas the IM and OM structures of the BG are
well preserved [6,9]. The IM remains intact during
expulsion of cytoplasmic material and electron micro-
graphs clearly show a sealed PPS [6,8]. The protein E-
induced occurrence of lysophosphatidylethanol-amine
in the host cell membrane most probably facilitates
the IM and OM fusion [10]. Although E-mediated
lysis is dependent on activities of the autolytic system
of the bacteria peptidoglycan, sacculi prepared from
E-lysed cells remain intact emphasizing that the
overall composition of this rigid layer is not changed
by the E-mediated process. The diameter of the E-
specific transmembrane tunnel structure ranges
between 40 and 200 nm and is determined by the
mesh size of the surrounding murein [11]. Electron
microscopic studies emphasize that the E-specific
transmembrane tunnel structure is not randomly
distributed over the cell envelope but is restricted to
areas of potential division sites, predominantly in the
middle of the cell or at polar sites [6,8]. Analysis of E-
mediated lysis in bacterial mutant strains with defects
in cell division suggest that initiation of cell division
rather than specific functions of the septosome plays
an essential role in protein E-mediated lysis [9,11].
The E-specific membrane fusion process can be
divided into three phases including the integration of
protein E into the IM, followed by a conformational
change of protein E and assembly into multimers at
potential cell division sites [12]. The mechanism for
the conformational change is most probably a cis–
trans isomerization of the proline 21 residue within
the first membrane-embedded a-helix of protein E
[13]. The local fusion of the IM and OM is achieved
by a transfer of the C-terminal domain of protein E
towards the surface of the OM of the bacterium.
BG have been developed for envelope and/or
heterologous antigen presentation from a range of
important Gram-negative bacterial pathogens includ-
ing Francisella tularensis, Brucella melitensis, enter-
otoxigenic and enterohemaorrhagic E. coli (EHEC,
ETEC), and V. cholerae. To date, immune responses
against P. multocida, Mannheimia (former Pasteur-
ella) haemolytica, A. pleuropneumoniae and V.
cholerae have been assessed in several animal models
for parenteral, oral and aerogenic modes of delivery,
in view of human and veterinary applications. The BG
particle presentation technology for target antigens to
induce an immune response against the target antigens
has also been studied extensively by our group
(reviews in [14–17]) and promising results of recent
studies will be presented below.
ED PROOF
3. Bacterial ghosts as candidate vaccines
3.1. Parenteral immunization with bacterial ghosts
Bovine pneumonic pasteurellosis caused by M.
haemolytica is a serious disease leading to death in
cattle if it remains untreated. Pilot subcutaneous BG
immunization studies of mice and rabbits with either
P. multocida- or M. haemolytica-ghosts induced
antibodies cross-reactive to heterologous Pasteurella
strains. The number of proteins in Pasteurella whole-
cell protein extracts recognized by the sera constantly
increased during the observation period of 50 days.
More importantly, dose-dependent protection against
homologous nasal challenge was observed in mice
immunized with P. multocida ghosts [18].
Following on from this work, M. haemolytica
ghosts cattle immunization studies were performed
using a cattle lung challenge model. Protective
immunity of cattle against homologous challenge
was induced by alum-adjuvanted M. haemolytica
ghosts. It is important to note that BG do not need
additional adjuvants to induce protective immunity.
However, alum was added to the ghost vaccine
preparation in this study to compare its antigenicity
with a commercially available vaccine [19].
Bacterial ghosts have been tested as a vaccine
against swine pleuropneumonia, a disease with a high
mortality rate in pigs. Intramuscular immunization of
pigs with A. pleuropneumoniae (APP) ghosts or
formalin-inactivated APP whole-cell bacteria pro-
tected for clinical disease in both vaccination groups.
The protective efficacy was evaluated by clinical,
bacteriological, serological and post-mortem exami-
nations. Immunization with BG did not cause clinical
side-effects. After aerosol challenge, the control group
of pigs developed fever and pleuropneumonia. In both
vaccination groups, animals were fully protected
against clinical disease and lung lesions, whereas
colonization of the respiratory tract with APP was
prevented by BG immunization alone. The induction
of specific mucosal antibodies as detected in the
T
ARTICLE IN PRESS
201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231
232
233234235236237238239240241242243244245246
247248249250251252253254255256
257
258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292
U.B. Mayr et al. / Advanced Drug Delivery Reviews xx (2005) xxx–xxx4
UNCORREC
bronchoalveolar lavage suggests that immunization
with BG induces antibody populations specific to
non-denaturated surface antigens. In this study at
least, APP-BG are more efficacious in protecting pigs
against colonization and infection than the inactivated
whole-cell vaccine [20]. Indications for a cross-
protective potential of the ghost vaccine were sup-
ported by studies on rabbit hyperimmune sera [21].
Cholera is a significant cause of morbidity and
mortality in humans and a vaccine is very much
needed. To this end, V. cholerae ghosts were produced
and assessed in a rabbit model. Rabbits were immu-
nized s.c./i.m. with ghosts prepared from V. cholerae
strains of O1 or O139 serogroup following growth
under culture conditions which favour or repress the
production of toxin-co-regulated pili (TCP). Immuno-
blotting confirmed the TCP status of V. cholerae
ghosts (VCG), which retained the cellular morphol-
ogy and surface component profile of viable bacteria.
Sera from immunized rabbits was assayed for anti-
bodies to lipopolysaccharide (LPS) and to TCP.
Regardless of the TCP status of the VCG preparations
used for immunization, all animals produced anti-
bodies to LPS as demonstrated in bactericidal assays.
Anti-LPS antibodies were likely responsible for
conferring passive immunity in the infant mouse
cholera model to challenge with the homologous
O139 strain. Cross-protective anti-TCP antibody was
generated only in rabbits immunized with TCP-
positive VCG. This sera induced protection against
heterologous challenge [22].
3.2. Induction of cytokines by bacterial ghosts
To investigate the activation of APC by BG we
studied the in vitro uptake of VCG and E. coli BG in
dendritic cells (DC) and RAW macrophages and the
induction of inflammatory mediators in the THP-1
human macrophage cell line. The synthesis of inflam-
matory mediators such as TNF-a in the THP-1 cell line
was stimulated by a hundred-fold higher dose of VCG
than the corresponding amount of free LPS [23,24].
These results support in vivo experiments in rabbits
with intravenous administration of E. coli BG. Below
a threshold dose, no toxic effects of BG administration
could be detected whilst the doses used stimulated
significant humoral immune responses [25]. Signifi-
cant production of IL-12 in DC was induced by E. coli
ED PROOF
BG. Secretion in DC of cytokines TNFa and IL-12
was increased 37 and 18-fold, respectively, whereas in
peripheral blood monocytes the secretion of TNFa and
IL-12 increased only twofold. These results suggest
that BG stimulate the activation of cellular Th1
immune responses. In addition, maturation of DC is
a prerequisite for efficient stimulation of T cells and
exposure of DC to BG resulted in a marked increase in
their ability to activate T cells. Thus, BG are promising
carrier and adjuvants for target antigens.
3.3. Mucosal immunizations with bacterial ghosts
Different routes of mucosal immunizations with
BG (aerogen, oral, intranasal, intravaginal, intraocular
or rectal) have been assessed in various animal models.
Binding and uptake of BG into APC is dependent on
surface structures of the envelope being recognized by
toll like receptors on human or animal cells.
Inhalation and deposition of BG within the airways
are the initial steps preceding adherence of the vaccine
candidates to the respiratory tract. Once BG are
deposited in the lung lining fluids (mucosa), they
are rapidly cleared by alveolar macrophages and
translocation of deposited particles in the mucus also
lead to clearance via the gastrointestinal tract [26].
APP-BG, after evaluation in the pig lung infection
model was then further assessed in an aerosol
immunization model [27]. The model utilized com-
puter-controlled standardized inhalation conditions for
the recipient pigs. APP-BG aerosol immunization has
been shown to induce complete protection against
pleuropneumonia in pigs [28].
The capability of BG to induce a T-cell-mediated
immune response was studied following uptake of
APP ghosts by primary APC of pigs. Specific T-cell
responses were detected after in vitro re-stimulation of
primed blood T cells with APP ghosts. In addition, we
investigated uptake of APP BG by DC and subse-
quent DC activation. DC are known to be phagocytic
in specific immature stages of development. Follow-
ing the internalization and processing of the antigens,
increased expression of MHC class II molecules in
APC was shown 12 h after their exposure to BG.
Together with the specific T-cell response to the
antigen processed by the APC, it could be demon-
strated that porcine APC have the capacity to
stimulate antigen-specific T cells after internalization
T
ARTICLE IN PRESS
293294295296297298299300301302303304305306307308309310311312313314315316317318319320321
322323
324325326327328329330331332333334335336337
338339340341342343344345346347348349350351352353354355356357358359360361362363364365366
U.B. Mayr et al. / Advanced Drug Delivery Reviews xx (2005) xxx–xxx 5
C
and processing of the antigen. The data suggest that
BG effectively stimulate monocytes and macrophages
to induce TH1-type cytokine directed immune
responses. DC stimulated by BG can be used for
active immunization and immunotherapy in situ [24].
The immunological and protective efficacy of V.
cholerae ghosts (VCG) expressing TCP (VCG-TCP)
from V. cholerae serogroups O1 and O139 has been
investigated in the reversible intestinal tie adult rabbit
diarrhea (RITARD) model. Rabbits were immunized 3
times intragastrically with a mixture of lyophilized
VCG-TCP from serogroup O1 and serogroup O139
and were challenged 30 days after the first immuniza-
tion with virulent V. cholerae O1 and V. cholerae
O139 strains.
Serum vibriocidal antibodies were observed in all
immunized animals and it could be shown that adult
rabbits were protected against diarrhea and death
following intralumen challenge with fully virulent V.
cholerae serogroups O1 and O139 [29]. Animal
models indicate that VCG induce humoral and cellular
immune responses against cell envelope constituents
including protective immunity against challenge
infections. All oral ghost vaccination experiments
were carried out with freeze-dried ghosts resuspended
in saline without the addition of adjuvants, stabilizers
or other substances [29]. VCG have the advantage of
ease of production by simple fermentation under
conditions, which favour the expression of TCP.
367368369370371372373374375376377378379380381382383384385
UNCORRE4. Bacterial ghost system as carrier of foreign
target antigens
Foreign target antigens can be tethered to the OM or
IM, exported into the PPS or can be expressed as S-
layer fusion proteins, which form shell-like self
assembly structures filling either the PPS or CPS
(Fig. 1). The OM (Fig. 1) is an asymmetric lipid bilayer
with LPS in the outer leaflet and phospholipids in the
inner leaflet. The polysaccharide moieties of LPS,
filaments and pili extend from the OM to the environ-
ment. The role of TCP to confer cross-protective
immunity in VCG has been mentioned earlier.
Outer membrane target antigen expression exploits
outer membrane proteins (OMP), which can be
modified to incorporate unrelated sequences [30]. In
a recent study, hepatitis B virus core 149 antigen was
ED PROOF
incorporated into OMP-A as fusion protein and
displayed on the surface of E. coli BG. These ghosts
induced a significant immune response against the
HBC 149 core antigen in mice [31].
Localization of target antigens in the PPS offers
several advantages. Target antigens exported to this
compartment are not only protected from external
degradation processes but are also immersed in a sugar-
rich environment of membrane-derived oligosacchar-
ides, which protect TA during lyophilization. Further-
more, soluble target antigens can be expressed in the
PPS of BG as the E-lysis tunnel seals the IM and OM.
MalE fusion proteins have been constructed which
secrete the target antigens into the PPS either as a
soluble protein or as part of a S-layer self-assembly
structure (Fig. 1). Site-directed mutagenesis of the S-
layer genes sbsA and sbsB and structural/functional
analysis of S-layer domains essential for intra- and/or
inter-molecular interactions [32–34] revealed flexible
surface loops in both proteins that accept foreign
target antigens sequences coding for up to 600 aa [35].
Such recombinant S-layer fusion proteins within a BG
consist of several hundred thousand monomers per
cell and because of their ability to assemble into a
superstructure, they do not form inclusion bodies.
Depending on the specific aim, multiple presenta-
tion of target antigens within the S-layer structure
could have beneficial effects compared to the soluble
form of the corresponding antigen (Fig. 1).
Electron microscopic pictures show sheet like self-
assembly structures of recombinant SbsA–Omp26
subunits in the PPS of E. coli ghosts [36]. The
Omp26 of non-typeable Haemophilus influenzae
(NTHi) was carried within the superstructure and E.
coli BG harboring this construct were highly immu-
nogenic for Omp26 when administered intraperito-
neally to mice [33].
The potential of E. coli ghosts carrying MalE–
Omp26 or MalE–SbsA–Omp26 fusion proteins in the
PPS was assessed as a delivery system for mucosal
immunization in a rat model and different routes of
immunization were evaluated. Animals were muco-
sally immunized targeting either gut only or gut and
lung mucosal sites. In the gut/lung regime, two initial
gut targeted inoculations with BG were followed by an
intratracheal (IT) boost with purified Omp26. The gut
only immunization regime showed a moderate
enhancement of bacterial clearance following pulmo-
UNCORRECTED PROOF
ARTICLE IN PRESS
OMP A
Pori
n
LPS
L
L
OM IM
PPS
CPS
OM
PPS
IM
CPS CPS
IM
IM
OM
Fig. 1. Bacterial ghost as carriers of autologous or foreign target antigens. Target antigens are compartmentalized into the outer membrane (OM),
the periplasmic space (PPS), the inner membrane (IM) and the cytoplasmic space (CPS). Target antigens (TA) can be BG components themselves
(e.g. pili, LPS, OMP, inner membrane proteins (IMP) and flagella). Foreign target antigens are displayed on the BG surface as a fusion protein
with OMP-A (upper left corner), or are exported to the sealed PPS as maltose binding protein (malE) or malE-sbsA/sbsB S-layer fusion proteins
(upper right corner). Target antigens can be anchored to the inner membrane via EV, LV or EV and LV anchor sequences. Membrane anchored StrpA
(EV-StrpA) can bind any biotinylated target antigen to the inner membrane. DNA carrying a lac operator sequence can bind to a membrane
anchored lacI (LacI-LV) repressor molecule (lower left and right corners). Recombinant S-layer proteins carrying foreign target antigens can fill
up the CPS of the BG. By loading the bacterial lumen with cccDNA plasmids, BG can act as carrier for DNA vaccines (lower left corner) .
Pilus Maltose binding protein S-layer
Outer membrane protein Outer membrane protein A carrying a target antigen Inner membrane protein
LPS Target antigen Porin
Phospholipid E’-Anchor L’-Anchor
Penicillin binding protein Flagellum E’anchored StrpA
L’anchored LacI repressor cccDNA ccc-DNA carrying thelac operator site
Biotin Biotinylated target antigen Bacterial ghost
Peptidoglycan
U.B. Mayr et al. / Advanced Drug Delivery Reviews xx (2005) xxx–xxx6
T
ARTICLE IN PRESS
386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433
434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481
U.B. Mayr et al. / Advanced Drug Delivery Reviews xx (2005) xxx–xxx 7
UNCORREC
nary challenge whereas the gut/lung immunization
regime resulted in significantly enhanced pulmonary
clearance of NTHi. Both immunization regimes
induced high levels of Omp26 specific antibodies in
the serum of immunized rats, with higher levels in the
groups that received the IT boost with purified Omp26.
Analysis of IgG isotypes present in serum suggest that
the immune response was predominantly of a T-helper1
type. Additionally, immunization induced a significant
cellular immune response with lymphocytes from
animals vaccinated using the gut/lung regime respond-
ing significantly to Omp26 when compared to control
groups. In summary, mucosal immunization with
recombinant Omp26 in E. coli ghosts followed by a
boost with purified Omp26 induced a specific and
protective immune response [37].
Bacterial ghosts have also been produced to
express target antigens in the CPS. Expression of
SbsA or SbsB fusion proteins in the CPS followed by
E-mediated lysis of the bacteria results in crystalline
planar arrays of S-layer proteins not released to the
surrounding medium (Fig. 1).
The CPS of BG can be filled either with water
soluble subunit antigens or emulsions such that the
target antigen itself or a matrix can be coupled to
appropriate anchors on the inside of the IM of BG
(Fig. 2). For example, BG with streptavidin anchored
on the inside of the IM can be filled by resuspending
lyophilized BG in solutions carrying biotinylated TA
[38].
For membrane anchoring of target antigens or of
acceptor proteins like streptavidin to the cytoplasmic
side of the IM, a membrane targeting system was
developed [39]. By cloning foreign DNA sequences
into the membrane targeting vector pMTV5, any gene
of interest can be expressed as a hybrid protein with
N-, C- or N-/C-terminal (EV-, LV-, EV-LV; Figs. 1 and 2)
membrane anchors directing and attaching the fusion
proteins to the cytoplasmic side of the IM of the
bacteria prior to E-mediated lysis. The current list of
membrane anchored target proteins comprises various
viral core or envelope proteins and bacterial target
antigens or enzymes. For the latter, it could be shown
that the enzymatic activities of h-galactosidase, PHB-synthase or alkaline phosphatase were not impaired
indicating that the membrane anchors do not interfere
with the proper folding of the target proteins and that
clustering and self-assembly (for example for h-
ED PROOF
galactosidase) is possible. The IM anchored HIV1-
RT and HIV1-gp41 target antigens carried by BG
induced humoral as well as a cellular immune
responses in animal models [40,41].
Any BG can be used as carrier for foreign antigens.
In a recent study [42], VCG have been successful used
to immunize against Chlamydia trachomatis. In
accordance with the new paradigm for vaccine design,
an efficacious anti-chlamydial vaccine should elicit a
genital mucosal Th1 response. To design a candidate
vaccine against Chlamydia based on the BGS, the
gene encoding the major OMP, omp1, of C. tracho-
matis was expressed in V. cholerae, as an IM-anchored
protein. Intranasal and intramuscular immunization of
naive mice with V. cholerae ghosts expressing OMP1
induced a strong Th1 immune response in the genital
mucosa. The ability of this vaccine delivery system to
protect susceptible animals from chlamydial infection
offers potential for the future development of effica-
cious vaccines capable of protecting human against
pathogens causing intracellular infections. In addition,
immune T cells from immunized mice could transfer
partial protection against a C. trachomatis genital
challenge to naRve mice. These results suggest that
VCG expressing chlamydial proteins may constitute a
suitable subunit vaccine for inducing an efficient
mucosal T-cell response that protects against C.
trachomatis infection [42]. In this example, VCG
offer the opportunity for designing TA vaccines within
the context of a cell envelope which is also able to
induce protective immunity against cholera.
Bacterial ghosts have been more recently developed
for delivery of antigens such as DNA. The internal
space of BG can be filled with a substituted matrix, e.g.
biotinylated dextran or polylysine which then binds the
target antigens of interest (Fig. 2a). For DNAvaccines,
it has been shown that plasmid DNA complexed with
polylysine can be efficiently packaged into BG [38]. If
the lac repressor proteins (LacI) is membrane anchored
(Fig. 1), it is still able to bind lac operator sequences
carried on plasmid DNA. Plasmids bound to the
membrane by this specific interaction are retained in
BG and are not expelled to the culture medium
following induction of E-mediated lysis.
It has also been observed that plasmid DNA
associates unspecifically with the inside of the IM.
Purified covalent closed circular DNA (cccDNA) can
be loaded to BG by resuspension of freeze dried BG
UNCORRECTED PROOF
ARTICLE IN PRESS
StrpA
BiotinylatedTA
StrpA
BiotinylatedPolymer
a b
c d
e f
ETTS
Fig. 2. Targeting membrane vesicles on top of the E-specific transmembrane tunnel (ETTS) structure of bacterial ghosts. (a) BG with
streptavidin-biotin coupled target antigens (TA) or biotinylated polymer in the CPS with open E-specific transmembrane tunnel structure. (b)
Sealing of inside-out vesicles of Gram-negative bacteria to the E-specific transmembrane tunnel structure of BG. Protein E fusion proteins (c) in
vivo biotinylated ( ) or (d) extended with streptavidin ( ) using the specific biotin–streptavidin interactions ( ) to position the membrane
vesicle of the E-specific transmembrane tunnel structure. (e) Multimers of streptavidin–biotin molecules (for simplicity only one construct is
shown) can form chimney like structures between the BG and the targeted vesicle. (f) Soluble target antigens of other substances ( ) can be
carried in the CPS of BG.
U.B. Mayr et al. / Advanced Drug Delivery Reviews xx (2005) xxx–xxx8
T
ARTICLE IN PRESS
482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529
530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559
560
561562563564565
566567568
569570
571
572
573
U.B. Mayr et al. / Advanced Drug Delivery Reviews xx (2005) xxx–xxx 9
UNCORREC
in DNA solutions. Depending on the DNA concen-
tration roughly 2000–4000 plasmids per single BG
can be attached to the IM.
When BG carrying plasmids encoding the green
fluorescent protein (GFP) were exposed under tissue
culture conditions to Caco-2 cells, DC or macro-
phages the cells expressed high levels of GFP [43].
As the latter two cell types are well known for their
high capacity to degrade phagocytosed material, it is
even more astonishing that GFP plasmids delivered by
BG transferred the DNA from the endosome/lyso-
some to the nucleus without any help of endo-
somolytic agents or of nuclear localization signals.
In an initial immunization study, plasmids encoding
the TA LacZ loaded in BG revealed strong humoral as
well as a Th1 help-mediated immune responses
against h-galactosidase in mice [44].
The feasibility of plugging the E-lysis tunnel of BG
to entrap target antigens in the CPS following loss of
cellular cytoplasmic constituents has been assessed.
Using a vesicle-to-ghost membrane fusion system,
BG can be plugged in order to use BG as carrier and
adjuvant systems for soluble, non-attached, hydro-
philic TA. The sealing process of ghosts requires
inside-out vesicles of gram-negative bacteria and
fuses the vesicles to the inner membrane at the edges
of the lysis tunnel of the ghost carrier. Ortho-
nitrophenyl-galactoside (ONPG), calcein and fluores-
cein-labeled DNAwere used as reporter substances to
test that BG can be sealed by restoring membrane
integrity (Fig. 2b) [45].
The technique of loosely closing BG is under
optimization and, as can be seen in Fig. 2c–f, antigen
carriers can be obtained by targeting a vesicle on top
of the E-specific transmembrane tunnel. In the most
simple model, vesicles can be targeted to the E-
specific transmembrane tunnel by specific interaction
of biotinylated protein E with membrane anchored
streptavidin on the surface of inside out vesicles (Fig.
2c) or vice versa by using E-streptavidin fusion
proteins for creation of the E-specific transmembrane
tunnel and inside out vesicles with membrane
anchored biotinylated receptor sequences (Fig. 2d).
In an alternative model, both receptor sequences on
the BG as well as on the inside out vesicle display
streptavidin on the surface and free biotin is used as
coupling agent (Fig. 2e). This traps the vesicle on top
of the E-specific transmembrane tunnel and can be
ED PROOF
used to construct BG carrying back packed envelope
fragments from other bacteria or viruses being either
biotinylated or modified with streptavidin.
With such constructs soluble drugs or antigens filled
into BG may be able to leak out through the tiny cleft
between the vesicle and the BG carrier (Fig. 2f). The
release rate of the enclosed substances can be regulated
by the distance between the BG carrier and vesicle
attached by adding various amounts of free biotin and
streptavidin, forming chimney-like structures which
can be constructed with different release properties.
Target antigens embedded in BG can be regarded
as subunit vaccine candidates fully equipped with a
whole bacterial cell adjuvant for better uptake by APC
by pattern recognition. The bacterial ghost antigen
presentation technology combines target antigen(s) on
a carrier, which elicits an efficient immune response.
The formulation of the target antigens packaged into
the ghost envelope structures is dependant on their
own physiochemical properties and it may be of
benefit to either loosely package them into the inner
space of the envelopes or to fix them to a matrix.
Clearly, different combinations of substances may
need to be located simultaneously in various compart-
ments of the BGS for optimal formulation. The
bacterial ghost technology warrants further investiga-
tion as it has great strategic potential in areas of
vaccine development against viral and bacterial
threats for which conventional vaccines do not exist
or are not sufficiently efficient.
Acknowledgements
This work was supported by grant GZ 309.049/1-
VI/6/2003 from the Austrian Ministry of Science. The
technical assistance of Beate Bauer, Alisa Lajta,
Roland N. Leitner and John McGrath for preparing
the manuscript is highly appreciated.
References
[1] M.P. Szostak, A. Hensel, F.O. Eko, R. Klein, T. Auer, H.
Mader, A. Haslberger, S. Bunka, G. Wanner, W. Lubitz,
Bacterial ghosts: non-living candidate vaccines, J. Biotechnol.
44 (1996) 161–170.
[2] M.C. Ronchel, A. Molina, W. Witte, S. Lubitz, J.L. Molin, C.
Ramos, Characterization of cell lysis in Pseudomonas putida
T
ARTICLE IN PRESS
574
575
576
577578
579
580581
582
583
584585
586
587
588589
590
591592
593
594
595596
597
598
599600
601
602
603604
605
606607
608
609
610611
612
613
614615
616
617618
619
620
621622
623
624
625626
627
628629
630
631
632
633
634635
636
637638
639
640
641642
643
644
645646
647
648649
650
651
652653
654
655
656657
658
659
660661
662
663664
665
666
667668
669
670
671672
673
674675
676
677
678679
680
681
682683
684
685686
687
U.B. Mayr et al. / Advanced Drug Delivery Reviews xx (2005) xxx–xxx10
UNCORREC
induced upon expression of heterologous killing genes, Appl.
Environ. Microbiol. 64 (1998) 4904–4911.
[3] D.U. Kloos, M. Stratz, A. Guttler, R.J. Steffan, K.N. Timmis,
Inducible cell lysis system for the study of natural trans-
formation and environmental fate of DNA released by cell
death, J. Bacteriol. 176 (1994) 7352–7361.
[4] W. Jechlinger, M. Szostak, A. Witte, W. Lubitz, Altered
temperature induction sensitivity of the lambda PR/cI857
system for controlled gene E-expression in Escherichia coli,
FEMS Microbiol. Lett. 173 (1999) 347–352.
[5] W. Jechlinger, M. Szostak, W. Lubitz, Cold-sensitive E-lysis
systems, Gene 218 (1998) 1–7.
[6] A. Witte, G. Wanner, U. Bl7si, G. Halfmann, M. Szostak, W.
Lubitz, Endogenous transmembrane tunnel formation medi-
ated by PhiX174 lysis protein E, J. Bacteriol. 172 (1990)
4109–4114.
[7] A. Witte, U. Bl7si, G. Halfmann, M. Szostak, G. Wanner, W.
Lubitz, PhiX174 protein E mediated lysis of Escherichia coli,
Biochimie 72 (1990) 191–200.
[8] A. Witte, G. Wanner, M. Sulzner, W. Lubitz, Dynamics of
PhiX174 protein E-mediated lysis of Escherichia coli, Arch.
Microbiol. 157 (1992) 381–388.
[9] A. Witte, E. Brand, G. Schrot, W. Lubitz, Pathway of
PhiX174 Protein E Mediated Lysis of Escherichia coli, in:
M.A. dePedro, J.-V. Hfltje, W. Lfffelhardt (Eds.), Bacte-
rial Growth and Lysis, Plenum Press, New York, 1993,
pp. 277–283.
[10] W. Lubitz, A.P. Pugsley, Changes in host cell phospholipid
composition of PhiX174 gene E product, FEMS Microbiol.
Lett. 30 (1985) 171–175.
[11] A. Witte, E. Brand, P. Mayrhofer, F. Narendja, W. Lubitz,
Dependence of PhiX174 protein E-mediated lysis on cell
division activities of Escherichia coli, Arch. Microbiol. 170
(1998) 259–268.
[12] P. Schfn, G. Schrot, G. Wanner, W. Lubitz, A. Witte, Two-
stage model for integration of the lysis protein E of PhiX174
into the cell envelope of Escherichia coli, FEMS Microbiol.
Rev. 17 (1995) 207–212.
[13] A. Witte, G. Schrot, P. Schfn, W. Lubitz, Proline 21, a residue
within the a-helical domain of a�174 lysis protein E, is
required for its function in Escherichia coli, Mol. Microbiol.
26 (1997) 337–346.
[14] F.O. Eko, A. Witte, V. Huter, B. Kuen, S. Fqrst-Ladani, A.Haslberger, A. Katinger, A. Hensel, M.P. Szostak, S. Resch, H.
Mader, P. Raza, E. Brand, J. Marchart, W. Jechlinger, W.
Haidinger, W. Lubitz, New strategies for combination vaccines
based on the extended recombinant bacterial ghost system,
Vaccine 17 (1999) 1643–1649.
[15] W.W. Jechlinger, W. Haidinger, S. Paukner, P. Mayrhofer, E.
Riedmann, J. Marchart, U. Mayr, C. Haller, G. Kohl, P.
Walcher, P. Kudela, J. Bizik, D. Felnerova, E.M.B. Denner, A.
Indra, A. Haslberger, M. Szostak, S. Resch, F. Eko, T.
Schukovskaya, V. Kutyrev, A. Hensel, S. Friederichs, T.
Schlapp, W. Lubitz, Bacterial ghosts as carrier and targeting
systems for antigen delivery, in: Guido Dietrich, Werner
Goebel (Eds.), Vaccine Delivery Strategies, Horizin Scientific
Press, Wymondham, UK, 2002, pp. 163–184.
ED PROOF
[16] K. Jalava, A. Hensel, M. Szostak, S. Resch, W. Lubitz,
Bacterial ghosts as vaccine candidates for veterinary applica-
tions, J. Control. Release 85 (2002) 17–25.
[17] K. Jalava, F.O. Eko, E. Riedmann, W. Lubitz, Bacterial ghosts
as carrier and targeting systems for mucosal antigen delivery,
Expert Rev. Vaccines 2 (2003) 45–51.
[18] J. Marchart, G. Dropmann, S. Lechleitner, T. Schlapp, G.
Wanner, M.P. Szostak, W. Lubitz, Pasteurella multocida- and
Pasteurella haemolytica-ghosts: new vaccine candidates,
Vaccine 21 (2003) 3988–3997.
[19] J. Marchart, M. Rehagen, G. Dropmann, M.P. Szostak, S.
Alldinger, S. Lechleitner, T. Schlapp, S. Resch, W. Lubitz,
Protective immunity against Pasteurellosis in cattle, induced
by Pasteurella haemolytica ghosts, Vaccine 21 (2003)
1415–1422.
[20] A. Hensel, V. Huter, A. Katinger, P. Raza, C. Strnistschie, U.
Roessler, E. Brand, W. Lubitz, Intramuscular immunization
with genetically inactivated (ghosts) Actinobacillus pleuro-
pneumoniae serotype 9 protects pigs against homologous
aerosol challenge and prevents carrier state, Vaccine 18 (2000)
2945–2955.
[21] V. Huter, A. Hensel, E. Brand, W. Lubitz, Improved protection
against lung colonization by Actinobacillus pleuropneumo-
niae ghosts: characterization of a genetically inactivated
vaccine, J. Biotechnol. 83 (2000) 161–172.
[22] F.O. Eko, U.B. Mayr, S.R. Attridge, W. Lubitz, Character-
ization and immunogenicity of Vibrio cholerae ghosts
expressing toxin-coregulated pili, J. Biotechnol. 83 (2000)
115–123.
[23] A.G. Haslberger, H.J. Mader, M. Schmalnauer, G. Kohl, P.
Messner, U.B. Sleytr, G. Wanner, S. Fqrst-Ladani, W. Lubitz,
Bacterial cell envelops (ghosts) and LPS but not bacterial S-
layers induce synthesis of immune-mediators in mouse
macrophages involving CD14, J. Endotoxin Res. 4 (1997)
431–441.
[24] A. Haslberger, G. Kohl, D. Felnerova, U.B. Mayr, S. Fqrst-Ladani, W. Lubitz, Activation, stimulation and uptake of
bacterial ghosts in antigen presenting cells, J. Biotechnol. 83
(2000) 57–66.
[25] H.J. Mader, M.P. Szostak, A. Hensel, W. Lubitz, A.G.
Haslberger, Endotoxicity does not limit the use of bacterial
ghosts as candidate vaccine, Vaccine 15 (1997) 195–202.
[26] A. Hensel, W. Lubitz, Vaccination by aerosols: modulation of
clearance mechanisms in the lung, Behring-Inst.-Mitt. 98
(1997) 212–219.
[27] A. Hensel, L.A.G. van Leengoed, M. Szostak, H. Windt, H.
Weissenbfck, N. Stockhofe-Zurwieden, A. Katinger, M.
Stadler, M. Ganter, S. Bunka, R. Pabst, W. Lubitz, Induction
of protective immunity by aerosol or oral application of
candidate vaccines in a dose-controlled pig aerosol infection
model, J. Biotechnol. 44 (1996) 171–181.
[28] A. Katinger, W. Lubitz, M.P. Szostak, M. Stadler, R. Klein,
A. Indra, Pigs aerogenously immunized with genetically
inactivated (ghosts) or irradiated Actinobacillus pleuropneu-
moniae are protected against a homologous aerosol challenge
despite differing in pulmonary cellular and antibody
responses, J. Biotechnol. 73 (1999) 251–260.
T
ARTICLE IN PRESS
688
689
690
691692
693
694695
696
697
698699
700
701
702703
704
705706
707
708
709710
711
712
713714
715
716
717718
719
720721
722
723
724
725
726727
728
729730
731
732
733734
735
736
737738
739
740741
742
743
744745
746
747
748749
750
751
752753
754
755756
757
758
U.B. Mayr et al. / Advanced Drug Delivery Reviews xx (2005) xxx–xxx 11
C
[29] F.O. Eko, T. Schukovskaya, E.Y. Lotzmanova, V.V. Firstova,
N.V. Emalyanova, S.N. Klueva, A.L. Kravtzov, L.F. Livanova,
V.V. Kutyrev, J.U. Igietseme, W. Lubitz, Evaluation of the
protective efficacy of Vibrio cholerae ghost (VCG) candidate
vaccines in rabbits, Vaccine 21 (2003) 3663–3674.
[30] G. Hobom, N. Arnold, A. Ruppert, Omp A fusion proteins for
presentation of foreign antigens on the bacterial outer
membrane, Dev. Biol. Stand. 84 (1995) 255–262.
[31] W. Jechlinger, C. Haller, S. Resch, A. Hofmann, M.P. Szostak,
W. Lubitz, Comparative immunogenicity of the hepatitis B
virus core 149 antigen displayed on the inner and outer
membrane of bacterial ghosts, Vaccine, in press.
[32] M. Truppe, S. Howorka, G. Schroll, S. Lechleitner, B. Kuen,
S. Resch, W. Lubitz, Biotechnological applications of
recombinant Slayer proteins rSbsA and rSbsB from Bacillus
stearothermophilus PV72, FEMS Microbiol. Rev. 20 (1997)
47–98.
[33] B. Kuen, M. Sara, W. Lubitz, Heterologous expression and
self assembly of the S-layer protein SbsA of Bacillus
stearothermophilus in Escherichia coli, Mol. Microbiol. 19
(1995) 495–503.
[34] B. Kuen, A. Koch, W. Asenbauer, M. Sara, W. Lubitz,
Sequence analysis of the sbsA gene encoding the 130-kDa
surface-layer protein of Bacillus stearothermophilus PV72,
Gene 145 (1994) 115–120.
[35] M. Truppe, S. Howorka, G. Schroll, S. Lechleitner, B. Kuen, S.
Resch, W. Lubitz, Biotechnological applications of recombi-
nant slayer proteins rSbsA and rSbsB from Bacillus stear-
othermophilus PV72, FEMSMicrobiol. Rev. 20 (1997) 47–98.
[36] E. Riedmann, J. Kyd, A. Smith, S. Gomez-Gallego, K. Jalava,
A. Cripps, W. Lubitz, Construction of recombinant S-layer
proteins (rSbsA) and their expression in bacterial ghosts—a
delivery system for the nontypeable Haemophilus influenzae
antigen Omp26, FEMS Immunol. Med. Microbiol. 37 (2003)
185–192.
UNCORRE
ED PROOF
[37] E. Riedmann, W. Lubitz, A.M. Smith, S. Gomez-Gallego, J.
McGrath, A.W. Cripps, J.M. Kyd, Immune responses to the
nontypeable Haemophilus influenzae antigen Omp26 deliv-
ered mucosally in bacterial ghosts, Vaccine, in press.
[38] V. Huter, M.P. Szostak, J. Gampfer, S. Prethaler, G. Wanner, F.
Gabor, W. Lubitz, Bacterial ghosts as drug carrier and
targeting vehicles, J. Control. Release 61 (1999) 51–63.
[39] M.P. Szostak, W. Lubitz, Recombinant bacterial ghosts as
multivaccine vehicles, in: R.M. Chanock, et al., (Eds.),
Modern Approaches to New Vaccines Including Prevention
of AIDS, Vaccines, vol. 91, Cold Spring Harbor Laboratory
Press, New York, 1991, pp. 409–414.
[40] M.P. Szostak, T. Auer, W. Lubitz, Immune response against
recombinant bacterial ghosts carrying HIV-1 reverse tran-
scriptase, Vaccine 93 (1993) 419–425.
[41] M.P. Szostak, A. Hensel, F.O. Eko, R. Klein, T. Auer, H.
Mader, A. Haselberger, S. Bunka, G. Wanner, W. Lubitz,
Bacterial ghosts, non-living candidate vaccines, J. Biotechnol.
44 (1996) 161–170.
[42] F.O. Eko, W. Lubitz, L. Mcmillan, K. Ramey, T.T. Moore,
G.A. Ananaba, D. Lyn, C.M. Black, J.U. Igietseme, Recombi-
nant Vibrio cholerae ghosts as a delivery vehicle for
vaccinating against Chlamydia trachomatis , Vaccine 21
(2003) 1694–1703.
[43] S. Paukner, G. Kohl, W. Lubitz, Bacterial ghosts as novel
advanced drug delivery systems: antiproliferative activity of
loaded doxorubicin in human Caco-2 cells, J. Control. Release
94 (2004) 63–74.
[44] T. Ebensen, S. Paukner, C. Link, P. Kudela, C. de Domenico,
W. Lubitz, C.A. Guzman, Bacterial ghosts are an efficient
delivery system for DNA vaccines, J. Immunol. 172 (2004)
6858–6865.
[45] S. Paukner, G. Kohl, K. Jalava, W. Lubitz, Sealed bacterial
ghosts—novel targeting vehicles for advanced drug delivery of
water-soluble substances, J. Drug Target. 11 (2003) 151–161.