The Red Queen Hypothesis (Van Valen 1973
) suggests a requirement for rapid co-evolution of the gene sets encoding the proteins most intimately involved in interacting with host cells: matched alleles undergoing an antagonistic co-evolution which serves to drive genetic polymorphism at key loci. We report for the first time the discovery and initial characterization of a family of telomeric proteins from C. hominis
and C. parvum
. The encoding genes were identified using comparative genomic analysis for divergent genes. Interestingly, the identified proteins shared common characteristics and appear to form a family of proteins with distinct properties. It is remarkable, but not altogether surprising, that this family of proteins is telomerically situated.
The telomeric location indicates that these genes are prone to higher recombination rates and are likely to encode contingency proteins. Contingency proteins are often secreted or external glycoproteins and are frequently encoded telomerically (Barry et al. 2003
). Such proteins have been shown to be involved in host–pathogen interaction and parasite survival in the host with examples including the variant surface glycoprotein of Trypanosma brucei,
which undergoes antigenic variation to evade the host immune responses and allow parasite survival (Barry et al. 2003
; Yang et al. 2009
), the var genes of Plasmodium falciparum
(Kyes et al. 2001
), the trans-sialidases of Trypanosoma cruzi
(Kim et al. 2005
), the major surface glycoproteins of Pneumocystis carnii
(Benfield and Lundgren 1998
) and the recently described subtelomeric variable secreted proteins of Theileria
(Schmuckli-Maurer et al. 2009
). The telomeres are prime sites for genes which are interacting with the host and evolving quickly because they are themselves dynamic and subject to novel forms of (epigenetic) regulation (Bhattacharyya and Lustig 2006
; Tonkin et al. 2009
; Yang et al. 2009
In terms of size and structure, Cops-1 and Chos-1
show similarities which suggests that they may be functionally related. As well as showing sequence similarity to each other, they show some similarity to the phosphoproteoglycans of Leishmania
which are secreted glycoproteins that have been shown to interact with the host immune system (Ilg et al. 1995
; Piani et al. 1999
Our comparative genomic approach aimed to identify coding loci with species-specific characteristics with a view to their exploitation for diagnosis and discrimination of human infective Cryptosporidium species. Given the considerable similarity of the vast majority of coding sequences between C. parvum and C. hominis, the genetic heterogeneity displayed by these proteins is itself an indicator that these genes are likely to be under direct selective pressure for adaptation through interaction of the proteins they encode with the host cells. The adaptation of these proteins most likely reflects the characteristics of host–pathogen interaction as a preferred niche and thus may contribute directly to the parasite's ability to colonize and infect particular hosts.
In the case of Cops-1, in silico prediction of species specificity was rejected experimentally when our analysis suggested the presence of an abridged ortholog is present in C. hominis (ChCops-1). Sequencing of a small PCR product from C. hominis showed high sequence similarity to Cops-1 providing evidence for a previously undiscovered ortholog of this gene in C. hominis (ChCops-1 has eluded C. hominis genome project). The exact genomic location of this gene is unknown, although it is likely to be telomeric as the primer which was used to amplify the full-length gene was designed to hybridize telomeric repeats. The two genes were named CpCops-1 and ChCops-1, for C. parvum and C. hominis, respectively. Comparative Southern blot analysis is likely to prove the definitive approach to assessing the presence, absence and positioning of orthologs of these genes in different Cryptosporidium species and strains. This approach is, however, hindered, particularly for C. hominis, by the difficulties in producing large amounts of genomic DNA from organisms which do not proliferate readily or substantially using current culture protocols.
Our PCR for Cops-1
is useful as a diagnostic test, specifically amplifying a 655-bp product from C. parvum
isolates and a 200-bp product from both C. hominis
and C. parvum
DNA. Sequencing of the 200-bp fragment enabled limited subtyping; indeed, a phylogenetic tree drawn from the sequence variation observed in this short fragment had a good discriminatory power and allowed discrimination of Cryptosporidium
genotypes and subtypes, which is consistent with the previous multi-locus analysis, therefore suggesting comparable polymorphism levels between this fragment and other genetic loci (Bouzid et al. 2010a
). We successfully expressed the recombinant CpCops-1 protein raising the prospect for its use for serodiagnosis.
Perhaps unsurprisingly, our anti-CpCops-1 monoclonal antibody (9E1) seems to be a nonblocking and nonneutralizing antibody, and this may simply imply that Tyle-2 epitope is not located in a region interacting with the intestinal cell receptors. When used for immuno-localization studies in faecal samples, the high background staining and relatively low intensity of the C. parvum-specific staining with 9E1 also meant that the antibody was unlikely to serve routinely as a useful diagnostic test to discriminate C. parvum from C. hominis. On purified C. parvum oocysts and sporozoites, though, the monoclonal antibody clearly recognized the contents of the oocyst when permeabilized and was able to stain free sporozoites without permeablization, demonstrating the association of CpCops-1 with the sporozoite surface. Conversely, on C. hominis oocysts and sporozoites, 9E1 showed little or no staining.
We have been unable to find evidence for expression of ChCops-1 by C. hominis either experimentally or in published databases. At point of publication it has not been reported in any of the several transcriptomic and proteomic data sets available for C. hominis. The inability of our antibody to detect the protein in C. hominis may reflect lack of expression but could also be due to differences in antigenicity between the orthologs. Our immunoblot of native C. parvum and C. hominis antigenic preparations probed with C. parvum-specific patient sera revealed an apparently C. parvum-specific antigen of the expected size for cpCops-1 but not for ChCops-1. If this antigen-specific band does accord with CpCops-1, it suggests that the C. hominis ortholog is either sufficiently dissimilar antigenically likewise not to be recognized by the patient sera or alternatively simply not expressed. Irrespective of the relative expression of these family members, the novel characteristics, localization and structural characteristics of this Cryptosporidium-specific family of proteins make Cops-1 and Chos-1 not only diagnostically and taxonomically useful, but plausible candidates as mediators of host specificity and virulence.