The parasitic microorganism Trypanosoma brucei evades recognition by its host's immune system by repeatedly changing its surface coat. The switch in coat follows a risky route, though: DNA break and repair.
Like many other single-celled pathogens, the protozoan Trypanosoma brucei, which causes African sleeping sickness in humans, undergoes antigenic variation — that is, it periodically switches its variant surface glycoprotein (VSG), the molecule targeted by host antibodies. But how switching is triggered has remained largely elusive. On page 278 of this issue, Boothroyd et al.1 show that a DNA double-strand break (DSB) upstream of the T. brucei VSG gene is the likely primary event in this process. Their results add to the few, albeit crucial, cases in which DSBs trigger developmental processes: these include mating-type switching in yeast, rearrangements of immune-system genes in humans and meiotic cell division to produce sex germ cells2.
Antigenic switching can occur through several genetic strategies, the most common being the differential activation of an archive of silent genes and pseudogenes. Although only one gene is transcribed, from a specialized expression site, switching occurs when silent genes, or their fragments, are duplicated in the expression site by a gene-conversion process, replacing all or part of the expressed gene. In some pathogens, the expressed gene can be constructed as a mosaic from several archival pseudogenes; such a combinatorial strategy expands the scale of variation enormously, with, for example, five pseudogenes giving rise to hundreds of combinations3.
Trypanosoma brucei has evolved an even more staggeringly complex system. It, too, transcribes a single VSG gene, but the sources of sequences that contribute to switching are large and diverse. It has several inactive expression sites, and its archive contains up to 200 VSG genes that lie at the ends (telomeres) of a set of mini-chromosomes, as well as a further 1,600 silent genes — of which two-thirds are pseudogenes — on the main chromosomes4. The potential for mosaic variation therefore seems beyond estimation. Intact archival genes are duplicated starting from an upstream set of repeat sequences each 70 base pairs (bp) long5, all the way to sequences at the downstream end of the coding sequence, or, in the case of silent telomeric genes, perhaps to the nearby end of the chromosome. As gene conversion in other organisms is initiated by a DSB in the conversion site, such a break has been proposed also to occur in the T. brucei VSG expression site, in its long set of 70-bp repeats.
Boothroyd et al.1 tested this hypothesis by creating a unique target site for the yeast endonuclease enzyme I-SceI in the expression site of the T. brucei VSG gene, adjacent to the 70-bp repeats (Fig. 1). Inducing this enzyme to become active, which caused DSBs in some 1% of trypanosomes, led to a dramatic increase in antigen switching. The switches involve typical conversions by telomeric archival VSG genes, stretching from the 70-bp repeats to, possibly, the chromosome end. To ascertain that the conversion was not merely repair in response to the artificial introduction of a break, the authors demonstrate that DSBs also occur naturally in the repeat regions of the transcribed gene, but seldom in another, inactive, expression site.
These findings suggest a model for switching in which natural breaks occur in the active expression site, precipitating conversion repair from another locus that contains a distinct but inactive VSG gene. Any model raises questions, and in this case two must be addressed.
One question is how the breaks occur. There are several possible mechanisms. First, they might be caused by an endonuclease, but such an enzyme would have to be strictly regulated to prevent lethally extensive cleaving of the many available 70-bp repeats in the genome. Second, a DNA-modification repair process might occur, similar to that mediated by the AID enzyme in human immunoglobulin-class switching6. But such a process seems too complex for the requirements of the VSG gene.
A third possibility is transcription-associated breakage. Indeed, the actively transcribed VSG expression site displays single-strand DNA sequences7, which could lead to DSBs during experimental DNA isolation; such artefactual breaks, however, are unlikely in the elegant technique Boothroyd et al. used. Alternatively, transcription might induce instability among the 70-bp repeats, triggering repair processes that cause DNA breakage8.
Finally, the structurally unstable repeats could also stall DNA replication, creating DSB-like free DNA ends that prompt repair through recombinational mechanisms9. The simplicity of this last mechanism is attractive, and could explain the abundance of 70-bp repeats in the expression site — to favour such accidents. Many bacterial phase-variation systems, which switch between alternative virulence states, do so through strategically located accident black-spots of unstable DNA-sequence tracts10.
Another, broader question arising from the model based on Boothroyd and colleagues' observations will probably be answered only in the longer term: can initiation of VSG switching by DSB formation explain the complexity and hierarchy of antigenic variation? Hierarchical gene expression — a key element of antigenic variation — arises from variations in the probability that different donor VSG genes are activated, and this is thought to relate to locus type, flanking sequences and, in the case of mosaic genes, the presence of related silent genes. Do these various types of VSG switching all involve DSBs?
Of necessity, the authors have studied a laboratory-adapted trypanosome strain, in which antigenic variation is impaired both quantitatively and qualitatively — recombination events are less frequent and less specific with regard to 70-bp repeats. This compromise might explain an anomaly in their observations: virtually all sequence donors to the switch were inactive expression sites, rather than mini-chromosomal sequences, which are favoured during natural infections (Fig. 1). The presence of DSBs in the low-switcher laboratory strain implies that either the high-switcher natural strains incur many more such breaks, or an essential downstream step, possibly one involved in repair, can become defective during laboratory adaptation, leading to less frequent switching.
It is now crucial to determine whether and how DSBs yield hierarchy in natural strains. Boothroyd and colleagues' findings — which provide a testable model for phenotype-switching systems in other organisms — should also prompt researchers to investigate the molecular players involved in switching.
Boothroyd, C. E. et al. Nature 459, 278–281 (2009).
Shrivastav, M. et al. Cell Res. 18, 134–147 (2008).
Futse, J. E. et al. Mol. Microbiol. 57, 212–221 (2005).
Berriman, M. et al. Science 309, 416–422 (2005).
Aline, R. Jr et al. Nucl. Acids Res. 13, 3161–3177 (1985).
Stavnezer, J. et al. Annu. Rev. Immunol. 26, 261–292 (2008).
Greaves, D. R. & Borst, P. J. Mol. Biol. 197, 471–483 (1987).
Lin, Y., Hubert, L. Jr & Wilson, J. H. Mol. Carcinog. 48, 350–361 (2009).
Labib, K. & Hodgson, B. EMBO Rep. 8, 346–353 (2007).
Moxon, R. et al. Annu. Rev. Genet. 40, 307–333 (2006).
About this article
Mapping replication dynamics in Trypanosoma brucei reveals a link with telomere transcription and antigenic variation
Molecular Microbiology (2011)
Current Opinion in Microbiology (2010)
Nature Chemical Biology (2010)
Trends in Parasitology (2010)