Nuclear architecture underlying geneexpression in Trypanosoma bruceiMiguel Navarro, Xenia Penate and David Landeira

Instituto de Parasitologıa y Biomedicina Lopez-Neyra, Consejo Superior de Investigaciones Cientıficas (Spanish National Research

Council), Avda. del Conocimiento s/n, 18100 Granada, Spain

Review TRENDS in Microbiology Vol.15 No.6

The influence of nuclear architecture on the regulation ofdevelopmental gene expression has recently become evi-dent in many organisms ranging from yeast to humans.During interphase, chromosomes and nuclear structuresare in constant motion; therefore, correct temporalassociation is needed to meet the requirements of geneexpression. Trypanosoma brucei is an excellent modelsystem in which to analyze nuclear spatial implications inthe regulation of gene expression because the two mainsurface-protein genes (procyclin and VSG) are transcribedby the highly compartmentalized RNA polymerase I andundergo distinct transcriptional activation or downregu-lation during developmental differentiation. Further-more, the infective bloodstream form of the parasiteundergoes antigenic variation, displaying sequentiallydifferent types of VSG by allelic exclusion. Here, we dis-cuss recent advances in understanding the role of chro-mosomal nuclear positioning in the regulation of geneexpression in T. brucei.

Trypanosomes and regulation of gene expressionAfrican trypanosomes display intriguing characteristics atthe transcriptional level that set them apart from mosteukaryoteswith respect to developmentally regulatedgenesin addition to housekeeping genes. On the one hand, devel-opmental-stage-specific surface protein genes are tran-scribed by RNA polymerase I (pol I). On the other hand,housekeeping genes are clustered into long polycistronictranscriptional units and transcribed by RNA polymeraseII (pol II) (reviewed in Ref. [1]). The T. brucei genomeconsists of 11 diploid megabase chromosomes (1–6 Mbp),1–5 intermediate chromosomes (200–900 kbp) and �100minichromosomes (50–150 kbp). The fact that RNA pol IIpromoters for housekeeping genes have proved elusive sofar, togetherwith differential regulation of polycistronicallytranscribed mRNA, led to the conclusion that constitutivegene expression is predominant in trypanosomes [2,3].Recently, a pol II promoter has been described for the spliceleader RNA [4], albeit showing rather weak activity. Thisweak level of activity could explain the difficulty in identify-ing other pol II promoters. Most of the more well-charac-terized promoters are associated with pol I and drive thedevelopmentally regulated robust transcription of variantsurface glycoprotein (VSG) genes andprocyclins, in additionto rDNA (Figure 1). In contrast to diploid housekeeping

Corresponding author: Navarro, M. (miguel.navarro@ipb.csic.es).Available online 4 May 2007.

www.sciencedirect.com 0966-842X/$ – see front matter � 2007 Elsevier Ltd. All rights reserv

genes,VSG genes are subject tomonoallelic expression fromhaploid chromosomal telomeres (reviewed in Refs [5–7]).There are �1000 VSG genes spread throughout differentchromosomes and the copy number and location can varylargely between strains [8,9].

In this review, we discuss recent results concerningtrypanosome nuclear architecture in the context of geneexpression, and search for insights into the mechanismsinvolved in gene regulation. Rather than review the largeamount of information available on antigenic variation orthe trypanosome life cycle, we discuss developmentallyregulated gene expression in the context of chromosomalposition and associations with nuclear structures in aneffort to provide new insights on epigenetic regulation.

Developmentally regulated gene expression ofsurface proteinsThe parasite Trypanosoma brucei alternates between itsmammalian host and tsetse fly vector during a complex lifecycle involving several developmental stages (reviewed inRefs [10,11]). The infective, extracellular bloodstream formis covered by a dense coat that consists mainly of one typeof glycosylphosphatidylinositol (GPI)-anchored VSG. Thesequential expression of VSG genes causes a persistentinfection by enabling bloodstream parasites to elude theimmune response of the host in a process known as anti-genic variation. Upon differentiation to the procyclic form(or tsetse mid-gut stage), the parasite expresses no VSGand, conversely, an invariant glycoprotein named procyclincovers its surface (reviewed in Refs [6,7,10]).

The procyclin gene family (EP and GPEET) is arrangedin tandem copies in internal chromosomal positions. How-ever, the expressed VSG is always located adjacent to atelomere at the end of one out of �20 polycistronic lociknown as expression sites (ESs) (Figure 1a). Only one ES isfully transcribed at a time, yielding a single VSG on thesurface of the infective trypanosome. A set of ES-associatedgenes [12] are polycistronically transcribed together withthe VSG from a promoter located 40–60 kb upstream,depending on the ES. The majority of antigenic variationis achieved by recombination events involving the replace-ment of VSG genes into the active ES [13] (reviewed in Ref.[5]); however, antigenic variation can also occur by thetranscriptional activation and inactivation of ESs in aprocess named in situ switching.

Many models to explain the expression of a singlesubtelomeric VSG-ES have been suggested thus far. Threeof themore important models are: (i) telomeric position, (ii)

ed. doi:10.1016/j.tim.2007.04.004

Figure 1. Genomic organization of RNA polymerase I transcribed loci and expression pattern in two developmental forms of Trypanosoma brucei. Genes are represented as

black boxes, except for genes encoding surface proteins (orange and green, VSG; blue, procyclin family). Flags indicate promoter position; full transcriptional activity is

outlined by red dashed arrows and abortive transcription is indicated by small brown dashed arrows. Telomere repeats are represented by horizontal triangles and other

repetitive sequences are indicated with vertically hatched boxes. Figures are not to scale. (a) The bloodstream form trypanosome eludes the mammalian immune response

by frequently changing its VSG coat, which is transcribed polycistronically from a subtelomeric region called the expression site (ES). T. brucei has �20 ESs but only one is

fully active at any given time. Abortive transcription occurs in the promoter-proximal region of at least some of the �19 inactive ESs [24]. The invariant surface protein or

procyclin genes are not transcribed. rDNA tandem arrays are constitutively transcribed. (b) Upon ingestion by the tsetse fly vector or when triggered in vitro, bloodstream

forms differentiate to the procyclic form. In the procyclic stage, no VSG transcription occurs and the surface coat is covered by the procyclins. This gene family is expressed

from two polycistronic loci located on chromosomes VI and X. rDNA tandem arrays retain constitutive expression.

264 Review TRENDS in Microbiology Vol.15 No.6

differential elongation–RNAmaturation, and (iii) a uniquenuclear transcription site. Here, we discuss these modelsand consider them within the framework of nuclear archi-tecture. Furthermore, these models are probably notmutually exclusive; instead, the mechanisms involvedmight work together or at different times during theregulation of VSG expression.

The VSG-ES telomeric positionThe telomeric position of VSG genes in VSG-ESs andminichromosomes has recently been the focus of muchattention (reviewed in Refs [5,14]). These studies haveled to the conclusion that the telomeric position of VSGgenes facilitates the major mechanism of antigenic switch-ing: recombination involving duplicative gene conversionand reciprocal telomere translocation. A search for themachinery involved in these events has proven difficult[15]; however, RAD51 T. brucei mutants did show areduced ability to undergo VSG switching [16]. Otherproteins such as Ku and TERT have been shown to beinvolved in telomere maintenance but do not significantlyalter switching rates [17–20].

Furthermore, antigenic switching can also occur byVSG-ES in situ transcriptional switching by an epigeneticmechanism. No sequence alterations, DNA rearrange-ments or chromatin changes that correlate with an in situtranscriptional switch have been detected in the promoterregion, although chromatin differences have beendescribed in the active telomeric VSG (reviewed in Ref.[7]). The position of VSG at chromosome ends has ledresearchers to hypothesize that a mechanism similar totelomeric silencing might operate on inactive ESs in blood-stream form trypanosomes. However, if a silencing mech-anism were involved, it is probably unrelated to yeast

www.sciencedirect.com

telomeric silencing for the following reasons: (i) in a nullmutant of the trypanosome ortholog of yeast silent infor-mation regulator SIR2 (TbSIR2P), no effect on antigenicswitching rates was observed [21]; (ii) two ESs cannot befully active at the same time [22]; and (iii) no associationwith the nuclear envelope has been detected for an inactiveVSG-ES promoter [23]. Taken together, these resultssuggest that a classical telomeric silencing mechanismdoes not operate on either inactive telomeric VSG-ESsor in situ transcriptional switching (reviewed in Refs[7,24]). In contrast to trypanosomes, Plasmodium anti-genic variation of var genes seems to be dependent onPfSIR2 for the silencing of inactive var genes [25,26].Importantly, inactive var genes tend to be associated withthe nuclear periphery [27].

The role of RNA polymerase I subnuclearcompartments in the regulation of VSG expressionThe recent recognition of subnuclear compartments asmodulators of transcription in eukaryotes has placedemphasis on research involving transcriptional regulationat the level of nuclear architecture (reviewed in Refs[28,29]. In contrast to pol II, pol I localizes exclusively inthe nucleolus of eukaryotic cells where it transcribes rDNAloci (reviewed in Ref. [30]). Because such compartmenta-lization occurs for pol I,much attention has been focused oninvestigation of the nuclear spatial organization of rDNA,procyclin and VSG loci transcribed by this machinery in T.brucei.

Early studies focused on the nuclear position of telomericand minichromosome sequences using fluorescence in situhybridization (FISH) analysis [31]. The authors concludedthat the majority of these sequences are found together inapproximately ten clusters, with a tendency to be localized

Review TRENDS in Microbiology Vol.15 No.6 265

in the periphery of the nucleus in the procyclic form and in amore central position in bloodstream form nuclei. Lateranalyses confirmed these results and investigated theimplications in the dynamics of chromosome segregation[32,33].

It was not until years later that a series of elegantstudies by the Borst group re-established an interest inthe investigation of nuclear position-dependent expres-sion regulation. Chaves et al. [34] demonstrated for thefirst time that active VSG-ES transcription does not occurin the nucleolus. FISH analysis using the 50 bp repeatsequences upstream of every VSG-ES promoter showedthat inactive ESs are widespread in the nucleoplasmand do not cluster towards the nuclear envelope in thebloodstream form [34]. Furthermore, double selection oftwo different ES promoters tagged with selectable mar-kers showed that the distance between the two nascentRNAs is less than half the distance of pol II control loci[22], hinting at the existence of a unique nuclear sub-compartment that allows full ES transcription. Theunstable nature of dual VSG expressers and, therefore,a possible intermediate switching state [22], togetherwith the inability to obtain triple-marked ES expressers

Figure 2. Subnuclear location of RNA pol I promoters as revealed by GFP tagging of chro

are represented as black boxes, except for genes encoding surface proteins (orange

transcriptional activity is outlined by red dashed arrows. Telomere repeats are repre

vertically hatched boxes. Insertion of lac-operator repeats upstream of the promoter is

from different GFP-lacI-tagged cell lines labeled by double immunofluorescence with a

DAPI staining (blue) [54]. Merged signals of maximum intensity projections from three

stained area, where the bulk of labeled pol I is located. Scale bar = 1 mm. (a) In bloodstre

is associated with an extranucleolar pol I-containing nuclear site, named the expression s

perinucleolar position in 96% of the nuclei (n = 100), instead of associating with the ESB

detected and the active procyclin promoter (arrowhead) is located in the nucleolar perip

(arrowhead).

www.sciencedirect.com

[35], suggest that this unique site only allows expressionof a single ES.

Another interesting result is that deletion of the activeES promoter leads to the stable activation of a differenttelomeric ES, resulting in an in situ switching event thatnormally occurs with low frequency [36]. This study indi-cated that activation and inactivation are coupled pro-cesses, and a model was hypothesized whereby theactive ES is located in a unique transcriptional site, adisruption of which leads to occupancy by another ES[36]. Working with this model, the search for a putativenuclear site led to the identification of a coherent andtranscriptionally active extra-nucleolar body, named theexpression site body (ESB), which associates with theactive VSG-ES promoter (Figure 2a) [23]. Thus, a modelwas proposed to explain VSG monoallelic expressionwhereby ESB-dependent VSG-ES recruitment leads tothe activation of a single VSG-ES, while inactive ESsare excluded [23]. It seems that the architecture of theESB might define the singularity of the active ES, therebyenabling a high level of both transcriptional elongation andRNA processing by recruiting transcriptional and RNAprocessing machinery [24].

mosomes. Left: A schematic overview of the GFP-lacI-tagged pol I promoter. Genes

and green, VSG; blue, procyclin family). Flags indicate promoter position and

sented by horizontal triangles and other repetitive sequences are indicated with

marked with green circles. Schemes are not to scale. Right: micrographs of nuclei

nti-pol I polyclonal antibodies (red), anti-GFP monoclonal antibodies (green) and

-channel 3D datasets are shown. The nucleolus is defined as the less intense DAPI

am form trypanosomes, the fully expressed active expression site (ES) (arrowhead)

ite body (ESB) (arrow). (b) Conversely, partially active ESs (arrowhead) localize in a

(arrow). (c) In procyclic form trypanosomes, no extranucleolar pol I signal can be

hery. (d) A similar perinucleolar location is observed for an active rDNA promoter

266 Review TRENDS in Microbiology Vol.15 No.6

Differential transcriptional elongation and RNAprocessingThe existence of a unique nuclear site is in agreement withthe efficient processing of transcripts from the active ES[37] and includes the model for ES regulation based on thedifferential control of transcriptional elongation and RNAprocessing proposed by Pays and coworkers [6,24]. Thismodel points to monoallelic expression of a single VSG-ESthrough differential transcriptional elongation and proces-sing of the RNA in the active ES [6,24], based on thedetection of transcription in inactive ES proximal regionsby RT-PCR [37–39]. Furthermore, a significant transcrip-tion level has been detected from inactive ES promotersthrough the insertion of selectable marker genes down-stream of inactive promoters [35,40–42] (reviewed in Ref.[7]). Thus, an important question emerges as to exactlywhere a partially active ES might localize in the nucleus,challenging the model of allelic exclusion based on ESB-dependent recruitment of a single ES.

Nuclear position analysis of the GFP-tagged (Box 1)inactive 121ES selected for partial activation shows thatit is not associated with the ESB but rather with thenucleolus (Figure 2b; D. Landeira and M. Navarro, unpub-lished), and can still be distinguished from the fully activeES because the parasites still display 221VSG on the mem-brane. This result suggests that the partially active ESpromoter in the nucleolus allows the ES to be accessibleto thepol Imachinery.However, thenucleolusprobablydoesnot contain the proper transcriptional elongation and pro-cessing machinery for VSG mRNA production, which islocated exclusively in the ESB. Thus, partial transcriptionof inactiveES sequencesmight bemediated byESpromoterpositioning in the nucleolus instead of the ESB.

Previous results show that active VSG-ES promoterreplacement with an rDNA promoter did not affect ESregulation [43]. This suggests that themachinery recruited

Box 1. Tetracycline-induced GFP–lacI tagging of

chromosomes in Trypanosoma brucei

To obtain a precise positional analysis of a particular locus in the

nucleus, in vivo GFP tagging of chromosomes [66,67] has been

applied to the two main proliferative trypanosome developmental

stages [23,54]. Classical FISH can alter nuclear architecture [68],

whereas GFP tagging is better at maintaining nuclear structure.

By using a tetracycline-inducible system [69] and GFP fused to the

lacI repressor, it is possible to localize a particular DNA sequence in

the nucleus at any given time, as visualized by GFP–lacI binding to

the lac operator sequences inserted into the chromosome. Stable

transformants in Trypanosoma brucei occur by homologous

recombination, which enables the insertion of a tagging cassette

containing 256 lac operators through a single crossover event. For

the purpose of localization of a particular locus, lac operator repeats

can be inserted a few Kb upstream of the promoter of interest. The

use of tetracycline-induced expression of GFP–lacI also enables

consideration of whether a chromatin region is accessible to the

binding of the lacI repressor or not at any given time.

Analysis of the position of a GFP-tagged promoter can be carried

out by in vivo microscopy in the procyclic form. 3D deconvolution

microscopy can be applied to both developmental forms upon

double indirect immunofluorescence using anti-GFP antibodies to

provide high-resolution nuclear position analysis. Overall, this

system is a powerful new tool in the study of chromosomal

positioning in T. brucei.

www.sciencedirect.com

by both the rDNA promoter (in the nucleolus) and theVSG-ES promoter (in the extra-nucleolar ESB) is similar.However, later nuclear position analysis showed that anES driven by an rDNA promoter was extra-nucleolar [34]and presumably located in the ESB, suggesting that othercis- or trans-acting sequences different from the promoterare responsible for ES recruitment to the ESB.

Positional analysis by FISH using probes common to allESs showed that inactive ESs are randomly located in thenucleus [34,44]. Differential nuclear positioning of VSG-ESs could help to explain the different degrees of tran-scriptional initiation detected from inactive ES promoters[37], both in the nucleolus (predominant RT-PCR product,or partially active ES) and in the nucleoplasma (minor RT-PCR product). As discussed previously [7], partially activeESs do not result in a higher frequency of in situ activation,suggesting that they are not involved in switching. How-ever, transcription of partially active ESs might provideadvantageous expression of different transferring receptortypes located downstream of all ES promoters.

Nuclear architecture and subnuclearcompartmentalizationVickerman’s early work on trypanosome nuclearultrastructure using electron microscopy (EM) describesmany nuclear components including nuclear pores anddense chromatin along the internal side of a mitotic-per-sistent nuclear envelope (NE) [45]. Chromatin structure inT. brucei is different from that of other eukaryotes becausea 30 nm fiber is not formed and mitotic chromosome con-densation is also absent [45]. This could be explained bythe divergence of core histones in this parasite [46]. Tran-scriptional regulation by the ’histone code’ in T. bruceimight have a crucial role, as suggested by recent data[47–50]. In T. brucei, histone methylation and acetylationhave been described [50], different histone modifyingenzymes have been identified [46,49,51], and there areat least two interesting histone variants [47,48]. One ofthese, H2AZ, contrary to yeast, does not co-localize with polII transcriptional foci in T. brucei but rather associateswith repetitive DNA [47], suggesting a different role in thisparasite.

A combination of FISH andEMprovides amore detailedpicture of T. brucei mitotic and interphase nuclei [32]. Theauthors show electron-dense chromatin aggregates thatform clusters close to the NE and disperse in the nucleo-plasm, and a prominent nucleolus that does not disinteg-rate during mitosis. The trypanosome nucleolus as viewedby EM is not as clearly structured into different concentriccomponents as in higher eukaryotes [52], although areas ofgranular and fibrillar appearance have been described[32]. Additional structural analyses by EM suggest differ-ences between isolated nuclei in the two developmentalstages [53]. Procyclic formnuclei appear to be roundedwitha clear NE, a large nucleolus and heterochromatinarranged into small regions, whereas the nucleus in thebloodstream form is irregular in shape and smaller, with aless distinct NE, a smaller nucleolus and more contiguousheterochromatin [53]. Structural components of the mainsubnuclear compartments in T. brucei have been purifiedand analyzed [53]; however, protein function has been

Review TRENDS in Microbiology Vol.15 No.6 267

inferred by subnuclear localization only. For instance,NUP-1 localizes to the nuclear envelope [32] and mightserve as a component of the trypanosome nuclear lamina[53]. Lamina proteins are believed to have a major role inreformation of the NE during late telophase in eukaryotes.In yeast, for example, nuclei do not disassemble their NEsduring mitosis and lack lamina proteins (reviewed in Ref.[29]). Because trypanosomes maintain their nuclear envel-ope during mitosis, the presence of the putative nuclearlamina protein NUP-1 suggests a function other than post-mitotic NE reformation (reviewed in Ref. [33]). Lamin Amight influence RNA polymerase II activity: as a negativeregulator of transcription that localizes to the NE, it mightfavor the incorporation of target genes into a repressiveenvironment [29]. Thus, a possible role for trypanosomeNUP-1 might be involvement in the organization of hetero-chromatin associated with the nuclear envelope in trypa-nosomes.

Labeling of nascent RNA with Br-UTP in the nuclei ofpermeabilized cells has facilitated the localization of tran-scriptionally competent areas in the nucleus of trypano-somes.Many transcriptional foci were detected throughoutthe nucleoplasm, however, these foci were absent towardsthe nuclear envelope in both developmental stages[23,39,54]. This suggests that the spatial organization ofchromosomes in the trypanosome nucleus might serve tofacilitate or to impede transcription, and thus nuclearrepositioning could regulate gene expression.

Procyclin chromosome locus positioning suggests afunctionally compartmentalized nucleolusThe nuclear localization of the highly transcribed procyclinchromosome loci in the procyclic form of trypanosomes hasbeen the subject of considerable study. Heterologous genestranscribed from the procyclin locus generate mRNAs thatare localized either to the nucleolus [55,56] or to the nucleo-plasm [34], as assessed by RNA-FISH. However, no infor-mation regarding the nuclear position of the chromosomeloci was available until recently [54]. Analysis by 3Dmicro-scopy indicates that the GFP–lacI tagged procyclin locuslocalizes to the nucleolar periphery (Figure 2c), similar tothe position of rDNA loci [54] (Figure 2d).

Furthermore, pol I is subcompartmentalized in thenucleolar periphery in a horseshoe-shaped manner [54].Pol I transcriptional activity is detected in the nucleolarperiphery in the procyclic form, consistent with the local-ization of the procyclin promoter and pol I [54]. Thisperinucleolar distribution of transcriptional activity isrelated to the subnuclear location of the exosome complex,which is involved in rRNA processing and mRNA degra-dation in trypanosomes [57]. The exosome localizes tospeckles in the nucleoplasm but it is enriched at theperiphery of the nucleolus, revealing a perinucleolar ri-ng-like pattern [58].

The confinement of pol I-transcribed DNAs (rDNA andprocyclin) to the nucleolar periphery, similar to the distri-bution of pol I and transcriptional activity, suggests thatthe structure of the nucleolus in T. brucei is different fromthat of other eukaryotes. Mammalian nucleoli show aconserved and concentric arrangement of three elementsdetectable by electron microscopy: the fibrillar center, the

www.sciencedirect.com

dense fibrillar component and the granular component.Each element corresponds to a different step in ribosomebiogenesis [52]. The peripheral organization of the trypa-nosome nucleolus suggests a model whereby transcriptionand splicing of procyclin mRNA take place towards thenucleoplasm, where coding RNAmaturation factors can befound.

The procyclin family of surface protein genes isconstitutively transcribed at a similar level for all allelicvariants [59] and such genes localize to the nucleolarperiphery [54]. By contrast, the monoallelically expressedVSG-ES promoter was found to be segregated to the ESB,which serves as a unique recruitment transcription site[23]. A model based on the recruitment of a single VSG-ESto the ESB is in agreement with the model described forodorant receptor monoallelic expression, where a singletrans-acting DNA element allows the activation of only oneodorant receptor allele in a neuron at a time [60]. Intrypanosomes, the unique transcription site (the ESB)was first described because extra-nucleolar pol I is easilydetectable. Whether this unique ESB is the result of asingle enhancer trans-acting DNA element that is respon-sible for its nucleation remains to be shown, although thisis a testable hypothesis.

VSG-ES promoter chromatin repositioning to thenuclear envelope during developmental silencingDuring B lymphocyte development, selective activation fortranscription and rearrangement of immunoglobulin locicorrelate with localization away from the nuclear periph-ery [61]. InSaccharomyces cerevisiae, transcriptional silen-cing of genes adjacent to telomeres is associated withnuclear envelope positioning (reviewed in Ref. [29]). Whenbloodstream forms are ingested by the tsetse fly, theparasite replaces its VSG coat with procyclin and no VSGsare expressed (Figure 1b). To investigate possible reposi-tioning of a chromosomal site during differentiation frombloodstream to procyclic form, GFP tagging of chromo-somes (Box 1) has been successfully employed in trypano-somes [54]. To perform these complex transgenicexperiments, a culture-adapted strain was used. Althoughthis strain does not reach the short stumpy bloodstreamstage, parasites had efficiently undergone differentiationto the procyclic stage after 24 h, as assessed by the largepercentage of cells expressing procyclin on the surface(88%). The active VSG-ES promoter selectively relocatesto the nuclear envelope 5 h after differentiation [54]. At thesame time, the extra-nucleolar pol I ESB is no longerdetected (Figure 3). An eventual repositioning of all inac-tive ES promoters to the nuclear envelope also occurs, asshown by FISH experiments [44] and by GFP tagging of asingle ES promoter in established procyclic trypanosomes[54]. Such nuclear repositioning upon early differentiationto the procyclic form precedes full VSG downregulationbecause VSG mRNA is still detected in this strain, asdescribed elsewhere [62]. During the transformation ofdividing slender forms into non-replicative short stumpybloodstream forms, an analysis of VSG-ES transcriptionshowed transcripts from many inactive ESs, suggestingthe loss of monoallelic control of the active ES [39]. How-ever, nascent nuclear RNA labeling suggested that short

Figure 3. Integrated model for the developmental repositioning of chromosomal sites in the nucleus during in vitro differentiation from bloodstream to procyclic form

trypanosomes. In the bloodstream form, the active expression site (ES) (orange sphere) is associated with an extranucleolar pol I-containing body named the expression

site body (ESB) (black sphere), whereas the inactive ES (brown sphere) remains in the nucleoplasm. Furthermore, abortive transcription occurs in a partially active ES

(yellow sphere) in the nucleolus (black horseshoe). The inactive procyclin locus (blue sphere) is partially repressed by a chromatin-mediated mechanism, and is most likely

to be located near the nuclear periphery, where no transcription is detected. The question mark indicates the putative position of the procyclin locus in the bloodstream

form of the parasite. Early on in differentiation to the procyclic form, when both VSG and procyclin mRNAs are detectable and parasites display mixed surface coats, the

active ES (orange sphere) selectively repositions to the nuclear envelope and the ESB is disassembled. Subsequently, a chromatin-mediated mechanism silences the active

ES. Dashed arrows indicate changes with respect to the previous point in development. At a similar time point, the procyclin locus (blue sphere) should be repositioned to a

transcriptionally active perinucleolar position. In established procyclic forms, both the inactive (brown sphere) and partially active (yellow sphere) ESs eventually reposition

to the nuclear envelope, where they undergo chromatin condensation.

268 Review TRENDS in Microbiology Vol.15 No.6

stumpy transformation results in an overall reduction oftranscriptional activity. The authors conclude from run-ondata that pol I is still associated with the promoter-prox-imal region of the active ES, thus establishing a decreasingtranscriptional gradient. Selective repositioning of theactive ES promoter to the nuclear envelope (in contrastto inactive ESs at early differentiation [54]) could explainthe loss of monoallelic control (Figure 3).

Using GFP tagging of chromosomes, it was alsoobserved that after differentiation only 8% of GFP-positivenuclei showed a detectable GFP–lacI dot, whereas detec-tion in control loci was unaffected [54]. This suggests thatGFP–lacI binding might reflect changes in chromatin con-densation. This supports previous findings regarding theaccessibility of ES chromatin in the bloodstream form andits subsequent inaccessibility upon differentiation [36].Recently, a chromatin decondensation event upon geneactivation was detected in mammalian cells using theGFP–lacI system [63].

Further work will be necessary to elucidate whether T.bruceimechanisms of ES nuclear repositioning are a causeor a consequence of developmental chromatin silencing.Recent work by the Gasser laboratory has shown that SIR-mediated repression can occur independently from nuclearenvelope repositioning in a yKu mutant. However, it can-not be ruled out that temporal nuclear repositioning isrequired to establish silencing [64]. In this context andsimilar to yeast, TbKu80-deficient trypanosomes areunable to halt VSG-ES developmental transcriptional

www.sciencedirect.com

silencing [18]. However, no information on possible nuclearrepositioning is available for this mutant. TbSIR2RP1localizes to the nuclear periphery together with telomericsequences [51]. However, transcription of a reporter genelocated downstream from the ES promoter was unaffectedby TbSIR2 disruption upon differentiation to the procyclicform, indicating that TbSIR2 is unlikely to have a directrole in the developmental control of the VSG-ES (M. Hoek,PhD thesis, Rockefeller University, 2001). These datasuggest that trypanosome pol I-mediated transcriptionof the VSG-ES is silenced by a TbSIR2-independent mech-anism. In addition, ES promoters in large episomes thatlack telomeres are effectively downregulated, indicatingthat silencing of ES promoters in the insect stage of T.brucei, although influenced by the genomic context, is notdependent upon telomere sequences [65]. The importantnew conclusion from these studies is that the rapid anddramatic repositioning rate of the active ES promoter tothe NE can be detected early in the differentiation process,suggesting that temporal nuclear periphery repositioningcontributes to the establishment of VSG-ES silencing [54](Figure 3).

Concluding remarks and future perspectivesNuclear repositioning of chromosomes has an importantrole in accomplishing proper control of gene expression ineukaryotes [28]. Correct timing in the positioning ofchromosome sites and association with particular nuclearstructures is necessary to coordinate and achieve gene

Review TRENDS in Microbiology Vol.15 No.6 269

regulation. Recent findings in trypanosome nuclearchromosome dynamics suggest that these phenomenaare extremely conserved because T. brucei is in a distalposition in the eukaryotic cell linage. It is clear that NErelocation affects RNA pol II transcription and, further-more, this relocation also influences pol I transcription inT. brucei [54]. This suggests that nuclear repositioning hasa global role in the regulation of transcription.

It now seems clear that procyclin gene transcription of allallele variants is unnecessary for association with a singleextra-nucleolar body [54], in contrast to the monoallelicexpression of the VSG-ES segregated in a single ES body[23]. Upon differentiation from the bloodstream to the pro-cyclic stage, the activeVSG-ES promoter undergoes a rapidrepositioning to thenuclearperipherywith concomitant lossof the ES body. After repositioning, the active ES promoterundergoes chromatin condensation (Figure 3). We favor amodel based on current data that suggests the existence oftwo mechanisms for VSG-ES regulation influenced bynuclear position. Firstly, a single ES is activated by theESbody ina recruitment-dependentmanner,which enablesmonoallelic expression and transcriptional switching by ESreplacement with a transcriptionally competent inactiveES. Secondly, during differentiation, the ES promoter repo-sitions to the NE and this repositioning is followed bychromatin condensation.

We have discussed several models for VSG-ESregulation, however, these do not represent mutuallyexclusive mechanisms; rather, they might exist concomi-tantly. Notwithstanding recent findings, many questionsremain to be answered, some of which are listed in Box 2.The combination of GFP chromosome tagging withhigh-resolution fluorescence microscopy and functional

Box 2. Unanswered questions

Are there any cis- or trans-active sequences required for mono-

allelic expression?

Certain sequences enable a single VSG-ES to be expressed either in

cis (linked to the active VSG-ES), or in trans (a unique enhancing

sequence located in a different chromosome). The latter is similar to

the situation described for olfactory receptors [60].

Where are the individual telomeres containing VSG-ESs located in

the nucleus?

To answer this question, GFP tagging of chromosomal telomeric

sites will probably be required.

Is there any difference between the RNA pol I machinery that isinvolved in rDNA and the one involved in VSG-ES and/or procyclin

transcription?

It would be necessary to identify specific transcription factors that

promote VSG and/or procyclin transcription, in addition to those

that regulate the RNA pol I core complex involved in rRNA

production.

Are there any VSG-ES-specific factors?

These would include ESB-specific transcription factors (for initia-

tion, elongation or RNA maturation) and ESB structural components

(if any) that enable the body to be structurally coherent and singular.

How does VSG-ES developmental silencing occur in Trypanosoma

brucei?

Do nuclear envelope repositioning and chromatin condensation

involve novel proteins that are different to those described in yeast?

www.sciencedirect.com

analysis is likely to provide new insights concerning themechanisms that underlie transcriptional regulation andantigenic variation in T. brucei.

AcknowledgementsWe apologize to colleagues whose work could not be cited owing tospace limitations. We thank Markus Engstler for trypanosome 3Dmicroscopy advice. We thank George A.M. Cross for sharingunpublished results and Daria Van Tyne for critical reading of thismanuscript. M.N. is a Howard Hughes Medical Institute InternationalResearch Scholar. X.P. is the recipient of a FPI fellowship funded bythe Spanish Ministerio de Educacion y Ciencia (MEC). This work wasfunded by HHMI-55005525, MEC SAF2005-00657 and Junta deAndalucia P05 CVI-00908 grants.

References1 Barry, D. et al. (2007) Trypanosomes: After The Genome. Horizon

Scientific Press2 Clayton, C.E. (2002) Life without transcriptional control? From fly to

man and back again. EMBO J. 21, 1881–18883 Palenchar, J.B. and Bellofatto, V. (2006) Gene transcription in

trypanosomes. Mol. Biochem. Parasitol. 146, 135–1414 Gilinger, G. and Bellofatto, V. (2001) Trypanosome spliced leader RNA

genes contain the first identified RNA polymerase II gene promoter inthese organisms. Nucleic Acids Res. 29, 1556–1564

5 Horn, D. and Barry, J.D. (2005) The central roles of telomeres andsubtelomeres in antigenic variation in African trypanosomes.Chromosome Res. 13, 525–533

6 Pays, E. et al. (2004) Antigenic variation in Trypanosoma brucei: facts,challenges and mysteries. Curr. Opin. Microbiol. 7, 369–374

7 Borst, P. and Ulbert, S. (2001) Control of VSG gene expression sites.Mol. Biochem. Parasitol. 114, 17–27

8 Berriman, M. et al. (2005) The genome of the African trypanosomeTrypanosoma brucei. Science 309, 416–422

9 Callejas, S. et al. (2006) Hemizygous subtelomeres of an Africantrypanosome chromosome may account for over 75% of chromosomelength. Genome Res. 16, 1109–1118

10 Barry, J.D. and McCulloch, R. (2001) Antigenic variation intrypanosomes: enhanced phenotypic variation in a eukaryoticparasite. Adv. Parasitol. 49, 1–70

11 Matthews, K.R. (2005) The developmental cell biology of Trypanosomabrucei. J. Cell Sci. 118, 283–290

12 Pays, E. et al. (2001) The VSG expression sites of Trypanosoma brucei:multipurpose tools for the adaptation of the parasite to mammalianhosts. Mol. Biochem. Parasitol. 114, 1–16

13 Conway, C. et al. (2002) Two pathways of homologous recombination inTrypanosoma brucei. Mol. Microbiol. 45, 1687–1700

14 Dreesen, O. et al. (2007) Telomere structure and function intrypanosomes: a proposal. Nat. Rev. Microbiol. 5, 70–75

15 Robinson, N.P. et al. (2002) Inactivation of Mre11 does not affect VSGgene duplication mediated by homologous recombination inTrypanosoma brucei. J. Biol. Chem. 277, 26185–26193

16 McCulloch, R. and Barry, J.D. (1999) A role for RAD51 and homologousrecombination in Trypanosoma brucei antigenic variation. Genes Dev.13, 2875–2888

17 Conway, C. et al. (2002) Ku is important for telomere maintenance, butnot for differential expression of telomeric VSG genes, in Africantrypanosomes. J. Biol. Chem. 277, 21269–21277

18 Janzen, C.J. et al. (2004) Telomere length regulation andtranscriptional silencing in KU80-deficient Trypanosoma brucei.Nucleic Acids Res. 32, 6575–6584

19 Dreesen, O. and Cross, G.A. (2006) Consequences of telomereshortening at an active VSG expression site in telomerase-deficientTrypanosoma brucei. Eukaryot. Cell 5, 2114–2119

20 Dreesen, O. and Cross, G.A. (2006) Telomerase-independent stabili-zation of short telomeres in Trypanosoma brucei. Mol. Cell. Biol. 26,4911–4919

21 Alsford, S. et al. (2007) A sirtuin in the African trypanosome is involvedin both DNA repair and telomeric gene silencing but is not required forantigenic variation. Mol. Microbiol. 63, 724–736

22 Chaves, I. et al. (1999) Control of variant surface glycoprotein gene-expression sites in Trypanosoma brucei. EMBO J. 18, 4846–4855

270 Review TRENDS in Microbiology Vol.15 No.6

23 Navarro, M. and Gull, K. (2001) A pol I transcriptional body associatedwith VSG mono-allelic expression in Trypanosoma brucei. Nature 414,759–763

24 Pays, E. (2005) Regulation of antigen gene expression in Trypanosomabrucei. Trends Parasitol. 21, 517–520

25 Freitas-Junior, L.H. et al. (2005) Telomeric heterochromatinpropagation and histone acetylation control mutually exclusiveexpression of antigenic variation genes in malaria parasites. Cell121, 25–36

26 Duraisingh, M.T. et al. (2005) Heterochromatin silencing and locusrepositioning linked to regulation of virulence genes in Plasmodiumfalciparum. Cell 121, 13–24

27 Ralph, S.A. et al. (2005) Antigenic variation in Plasmodium falciparumis associated with movement of var loci between subnuclear locations.Proc. Natl. Acad. Sci. U. S. A. 102, 5414–5419

28 Spector, D.L. (2003) The dynamics of chromosome organization andgene regulation. Annu. Rev. Biochem. 72, 573–608

29 Taddei, A. et al. (2004) The function of nuclear architecture: a geneticapproach. Annu. Rev. Genet. 38, 305–345

30 Scheer, U. and Hock, R. (1999) Structure and function of the nucleolus.Curr. Opin. Cell Biol. 11, 385–390

31 Chung, H-MM. et al. (1990) Architectural organization in theinterphase nucleus of the protozoan Trypanosoma brucei: Locationof telomeres and mini-chromosomes. EMBO J. 9, 2611–2619

32 Ogbadoyi, E. et al. (2000) Architecture of the Trypanosoma bruceinucleus during interphase and mitosis. Chromosoma 108, 501–513

33 Ersfeld, K. et al. (1999) Nuclear and genome organization ofTrypanosoma brucei. Parasitol. Today 15, 58–63

34 Chaves, I. et al. (1998) Subnuclear localization of the active variantsurface glycoprotein gene expression site in Trypanosoma brucei. Proc.Natl. Acad. Sci. U. S. A. 95, 12328–12333

35 Ulbert, S. et al. (2002) Expression site activation in Trypanosomabrucei with three marked variant surface glycoprotein geneexpression sites. Mol. Biochem. Parasitol. 120, 225–235

36 Navarro, M. et al. (1999) Trypanosoma brucei variant surfaceglycoprotein regulation involves coupled activation/inactivationand chromatin remodeling of expression sites. EMBO J. 18, 2265–2272

37 Vanhamme, L. et al. (2000) Differential RNA elongation controls thevariant surface glycoprotein gene expression sites of Trypanosomabrucei. Mol. Microbiol. 36, 328–340

38 Rudenko, G. et al. (1994) VSG gene expression site control in insectform Trypanosoma brucei. EMBO J. 13, 5470–5482

39 Amiguet-Vercher, A. et al. (2004) Loss of the mono-allelic control of theVSG expression sites during the development ofTrypanosoma brucei inthe bloodstream. Mol. Microbiol. 51, 1577–1588

40 Navarro, M. and Cross, G.A. (1996) DNA rearrangements associatedwith multiple consecutive directed antigenic switches in Trypanosomabrucei. Mol. Cell. Biol. 16, 3615–3625

41 Navarro, M. and Cross, G.A. (1998) In situ analysis of a variant surfaceglycoprotein expression-site promoter region in Trypanosoma brucei.Mol. Biochem. Parasitol. 94, 53–66

42 Ansorge, I. et al. (1999) Transcription of ‘inactive’ expression sites inAfrican trypanosomes leads to expression of multiple transferrinreceptor RNAs in bloodstream forms. Mol. Biochem. Parasitol. 101,81–94

43 Rudenko, G. et al. (1995) A ribosomal DNA promoter replacing thepromoter of a telomeric VSG gene expression site can be efficientlyswitched on and off in T. brucei. Cell 83, 547–553

44 Perez-Morga, D. et al. (2001) Organization of telomeres during the celland life cycles of Trypanosoma brucei. J. Eukaryot. Microbiol. 48, 221–226

45 Vickerman, K. and Preston, T.M. (1970) Spindle microtubules in thedividing nuclei of trypanosomes. J. Cell Sci. 6, 365–383

46 Alsford, S. and Horn, D. (2004) Trypanosomatid histones. Mol.Microbiol. 53, 365–372

www.sciencedirect.com

47 Lowell, J.E. et al. (2005) Histone H2AZ dimerizes with a novel variantH2B and is enriched at repetitive DNA in Trypanosoma brucei. J. CellSci. 118, 5721–5730

48 Lowell, J.E. and Cross, G.A. (2004) A variant histone H3 is enriched attelomeres in Trypanosoma brucei. J. Cell Sci. 117, 5937–5947

49 Janzen, C.J. et al. (2006) Unusual histone modifications inTrypanosoma brucei. FEBS Lett. 580, 2306–2310

50 Janzen, C.J. et al. (2006) Selective di- or trimethylation of histone H3lysine 76 by twoDOT1 homologs is important for cell cycle regulation inTrypanosoma brucei. Mol. Cell 23, 497–507

51 Garcia-Salcedo, J.A. et al. (2003) A chromosomal SIR2 homologue withboth histone NAD-dependent ADP-ribosyltransferase and deacetylaseactivities is involved in DNA repair in Trypanosoma brucei. EMBO J.22, 5851–5862

52 Raska, I. et al. (2006) Structure and function of the nucleolus in thespotlight. Curr. Opin. Cell. Biol. 18, 325–334

53 Rout, M.P. and Field, M.C. (2001) Isolation and characterization ofsubnuclear compartments from Trypanosoma brucei. Identification ofa major repetitive nuclear lamina component. J. Biol. Chem. 276,38261–38271

54 Landeira, D. and Navarro, M. (2007) Nuclear repositioning of the VSGpromoter during developmental silencing in Trypanosoma brucei. J.Cell Biol. 176, 133–139

55 Rudenko, G. et al. (1991) RNA polymerase I can mediate expression ofCAT and neo protein-coding genes in Trypanosoma brucei. EMBO J.10, 3387–3397

56 Chung, H.M. et al. (1992) RNA polymerase I-mediated protein-codinggene expression in Trypanosoma brucei. Parasitol. Today 8, 414–418

57 Estevez, A.M. et al. (2001) The exosome of Trypanosoma brucei. EMBOJ. 20, 3831–3839

58 Haile, S. et al. (2007) The subcellular localisation of trypanosome RRP6and its association with the exosome.Mol. Biochem. Parasitol. 151, 52–58

59 Acosta-Serrano, A. et al. (1999) The procyclin repertoire ofTrypanosoma brucei. Identification and structural characterizationof the Glu-Pro-rich polypeptides. J. Biol. Chem. 274, 29763–29771

60 Lomvardas, S. et al. (2006) Interchromosomal interactions andolfactory receptor choice. Cell 126, 403–413

61 Kosak, S.T. et al. (2002) Subnuclear compartmentalization ofimmunoglobulin loci during lymphocyte development. Science 296,158–162

62 Janzen, C.J. et al. (2006) Expression site silencing and life-cycleprogression appear normal in Argonaute1-deficient Trypanosomabrucei. Mol. Biochem. Parasitol. 149, 102–107

63 Dietzel, S. et al. (2004) Differential large-scale chromatin compactionand intranuclear positioning of transcribed versus non-transcribedtransgene arrays containing b-globin regulatory sequences. J. CellSci. 117, 4603–4614

64 Gartenberg, M.R. et al. (2004) Sir-mediated repression can occurindependently of chromosomal and subnuclear contexts. Cell 119,955–967

65 te Vruchte, D. et al. (2003) Downregulation ofTrypanosoma bruceiVSGexpression site promoters on circular bacterial artificial chromosomes.Mol. Biochem. Parasitol. 128, 123–133

66 Robinett, C.C. et al. (1996) In vivo localization of DNA sequences andvisualization of large-scale chromatin organization using lac operator/repressor recognition. J. Cell Biol. 135, 1685–1700

67 Straight, A.F. et al. (1996) GFP tagging of budding yeast chromosomesreveals that protein–protein interactions can mediate sister chromatidcohesion. Curr. Biol. 6, 1599–1608

68 Azuara, V. et al. (2003) Heritable gene silencing in lymphocytes delayschromatid resolution without affecting the timing of DNA replication.Nat. Cell Biol. 5, 668–674

69 Wirtz, E. et al. (1999) A tightly regulated inducible expression systemfor conditional gene knock-outs and dominant-negative genetics inTrypanosoma brucei. Mol. Biochem. Parasitol. 99, 89–101