Evolutionary Applications Evolutionary Applications ISSN 1752-4571 ORIGINAL ARTICLE A new heterogeneous family of telomerically encoded

Cryptosporidiosis is predominantly caused by two closely related species of protozoan parasites the zoonotic Cryptosporidium parvum and anthroponotic Cryptosporidium hominis which diverge phenotypically in respect to host range and virulence. Using comparative genomics we identified two genes displaying overt heterogeneity between species. Although initial work suggested both were species specific, Cops-1 for C. parvum and Chos-1 for C. hominis, subsequent study identified an abridged ortholog of Cops-1 in C. hominis. Cops-1 and Chos-1 showed limited, but significant, similarity to each other and share common features: (i) telomeric location: Cops-1 is the last gene on chromosome 2, whilst Chos-1 is the first gene on chromosome 5, (ii) encode circa 50-kDa secreted proteins with isoelectric points above 10, (iii) are serine rich, and (iv) contain internal nucleotide repeats. Importantly, Cops-1 sequence contains specific SNPs with good discriminatory power useful epidemiologically. C. parvum-infected patient sera recognized a 50-kDa protein in antigen preparations of C. parvum but not C. hominis, consistent with Cops-1 being antigenic for patients. Interestingly, anti-Cops-1 monoclonal antibody (9E1) stained oocyst content and sporozoite surface of C. parvum only. This study provides a new example of protozoan telomeres as rapidly evolving contingency loci encoding putative virulence factors.


Introduction
Pathogens can be defined as microbes capable of causing damage to the host, and virulence factors defined as mediators which enable that damage to occur (Casadevall and Pirofski 2001). By these definitions, determinants of host specificity are not only encoded by contingency genes adapted through rapid, red-queen evolution for the particular biological niche represented by their host, but are virulence factors which specifically enable colonization, infection and pathogenesis.
Cryptosporidiosis is one of the commonest parasitic causes of diarrhoea and a sometimes fatal disease in the immune-suppressed (Hunter and Nichols 2002;Karanis et al. 2007) with chronic sequelae postacute infection (Hunter et al. 2004). Of some eight species which infect humans (Xiao and Fayer 2008), Cryptosporidium parvum and Cryptosporidium hominis are the main species of public health importance, causing the majority of human cases as both sporadic and outbreak-related cases. In a survey of over 4000 isolates in the UK, C. parvum and C. hominis were responsible of 38.5% and 57.3% of the cases, respectively (Chalmers et al. 2010). Although closely related C. hominis mainly infects humans, while C. parvum exhibits a broader host range including humans, livestock and rodents (Xiao et al. 2004).
The genetic determinants driving the host preference of Cryptosporidium species are not yet elucidated, but in other protozoan parasites, such genes are frequently found clustered close to the telomeres, can be hypervariable with regard to sequence and are often part of heterogenous multigene families (Barry et al. 2003). The proteins encoded are frequently surface, shed or secreted glycoproteins and for closely related protozoan parasite isolates these can be the only loci which vary dramatically in terms of allele number and sequence (Franzen et al. 2011).For Cryptosporidium, it had been speculated that genome sequencing would assist in uncovering mechanisms of host specificity and virulence; however, this has been hindered by the high sequence similarity (95-97%) between the genome sequences of C. parvum and C. hominis (Widmer et al. 2002).It has been concluded that the genomes of C. parvum and C. hominis are highly similar, and the phenotypic differences, which relate mainly to host preference and pathogenicity, are caused by polymorphisms in coding regions and differences in gene regulation (Xu et al. 2004;Pain et al. 2005). This similarity has also largely frustrated reverse diagnostic approaches to the provision of improved tests for robust molecular and serological discrimination of Cryptosporidium infections. Nevertheless, detailed assessment of the genetic variation between C. hominis and C. parvum remains the most tractable approach to the provision of valuable epidemiological markers and a gene discovery vehicle for previously unnoticed genetic determinants of host specificity and virulence. To date, only one protein GP60 has shown large considerable heterogeneity between Cryptosporidium species and strains. This protein whose encoding gene has been used widely as a basis for strain subtyping is believed to be a surface glycoprotein subject to direct selection pressure and with a role in hostparasite interaction (Widmer and Lee 2010).
In a study by Kuo and Kissinger (2008), comparative genomic analysis of apicomplexan parasites (Plasmodium, Theileria and Cryptosporidium) revealed that as many as 45% of Cryptosporidium genes could be considered genus specific. Ordinarily, it is expected that a restriction in host range from a generalist such as C. parvum to a specialist such as C. hominis may be accompanied by selective loss of alleless; however, this study identified 334 putatively C. hominis-specific genes and 178 C. parvum-specific genes. Nevertheless, it seemed likely that this subset of genes would include the genetic determinants of phenotypic differences such as host range.
In a previous study, we also used a comparative genomic approach to identify putative species-specific genes as possible markers for host adaptation of C. hominis and C. parvum (Bouzid et al. 2010a). Our in silico analysis gave similar numbers of putatively species-specific genes to Kuo and Kissinger: 93 and 211 genes for C. hominis and C. parvum, respectively (Bouzid et al. 2010a). Unexpectedly, though, when tested experimentally, the vast majority of the predicted species-specific genes were common to C. parvum and C. hominis, in part, owing to the incomplete status of the C. hominis genome. Nevertheless, amongst this group of genes, experimental evidence was found for one C. parvum-specific gene (Cops-1) and one C. hominis-specific gene (Chos-1; Bouzid et al. 2010a).
Interestingly, these two genes share common characteristics. Here, we describe evidence that they are founder members of a novel family of secreted telomeric contingency proteins, which may play a role in facilitating the interaction between Cryptosporidium and mammalian intestinal epithelium.

Bioinformatics
Cops-1 and Chos-1 sequences were analysed using online software tools for protein analysis, primarily accessed via the Swiss Institute for Bioinformatics' EXPASY site (http:// www.expasy.org/) and included kalign (http://www.ebi.ac. uk/Tools/msa/kalign/), GENESTREAM (Pearson et al. 1997 To test experimentally the predicted specificity of the selected genes and determine the genetic polymorphism of Cops-1 gene, a panel of purified oocysts and genomic DNAs from predominantly clinical Cryptosporidium isolates of C. parvum, C. hominis and C. cuniculus isolates was made available by the Cryptosporidium Reference Unit (Swansea, UK). Details of epidemiological and genotyping data from these isolates were previously described (Bouzid et al. 2010a). The clinical strains tested were C. parvum (Cp2, Cp3 and Cp4), anthroponotic C. parvum (W7265, W7266, W7267 and W7270), C. hominis (Ch2, Ch3 and Ch4) and C. cuniculus (W17330, W18455, W17525 and W17435). In addition, commercial strains Iowa, Moredun and TU502 were also tested.
PCR screening, product sequencing and cloning and expression of the Cops-1 gene Primers were designed, PCR performed and PCR products sequenced using standard methods as previously described (Bouzid et al. 2010b). Full-length CpCops-1 gene was amplified using the expand High Fidelity PCR system (Roche, Hertfordshire, UK) and Cgd2_4380_FF and FR primers (Table S1). Cloning of PCR product into cloning vector pET100/D-TOPO was achieved by topoisomerase I action allowing directional cloning. Then, the vector was transfected into a variety of competent Escherichia coli expression strains including DH5a, BL21 Star TM (DE3), C41 (DE3), C43 (DE3), BL21-Codon Plus-RP and Rosetta TM 2 (DE3) for expression purposes.

Anti-Cops-1 monoclonal antibody
The 9E1 IgG1 mouse monoclonal antibody to Cops-1 protein was produced by Cambridge Research Biochemicals (www.crbdiscovery.com). Predicted Cops-1 protein sequence (477 aa) as retrieved from NCBI (XM_626615) was submitted for peptide design, and a peptide sequence (Tyle-2) [C]-RSRPPLPTRKPYSGS-amide (position 297-311) was selected corresponding to high immunogenicity. The selected sequence was tested using Basic Localized Alignment Tools (BLASTp selected for short nearly identical oligopeptides and tBLASTn) to ensure that no sequence similarity was present for another Cryptosporidium species or indeed for species other than Cryptosporidium. This 15 amino acid sequence showed no sequence similarity. The monoclonal antibody produced was used for immunofluorescence (IFA) staining of Cryptosporidium oocysts. For some assays, the 9E1 was directly conjugated using Zenon ® Alexa Fluor ® 594 MouseIgG1 Labelling kit (Molecular Probes, Invitrogen, Paisley, UK). The specificity and staining pattern of 9E1 was tested by IFA using two C. parvum (Iowa, Moredun) and two C. hominis (Ch2, Ch4) strains.

Antigen preparation
For Western blotting, native antigen preparation was performed from 2 9 10 6 oocysts, concentrated by centrifugation 13 000 rpm for 10 min at 4°C. The oocysts were frozen in a dry ice/ethanol bath and immediately thawed in a water bath at 42°C. This cycle was repeated four times to ensure the breakage of the oocyst wall. Then, 2 lL of protease inhibitors was added.

Immunofluorescence and blocking assay
For IFA, approximately 10 3 parasites were air-dried onto single-welled microscope. Directly conjugated 9E1 antibody was incubated for 30 min at room temperature in a humidified staining chamber. After three washes in PBS, crypto-cell antibody (TCS BioSciences Ltd., Buckingham, UK) was added to the slide and incubated at 37°C for 15 min. The antibody was then carefully removed, and the slide was stained using 0.5 lg/mL 4′,6-diamidino-2-phenylindole (DAPI) for 2 min. When excystation of the oocysts was required, it was performed as previously described (Choudhry et al. 2008).
The blocking assay was conducted essentially as described by Nishikawa et al. (2000). Briefly 1 9 10 6 Cryptosporidium oocysts from a purified suspension were preincubated with antibody (9E1) at 37°C for 90 min. As well as 9E1, a matched IgG1 isotype control (AbDSerotec; Mor-phoSys UK Ltd, Oxford, UK) was used. The mixture was then used to infect confluent Caco-2 cell monolayers for 3 h. The cells were washed and further cultured for 24 h before being fixed and stained with Giemsa. The effect of the pre-incubation with the antibody was assessed by parasite count for each condition from 20 randomly chosen fields.

Patient's sera
Anonymised sera were obtained from laboratory-confirmed cryptosporidiosis cases under a study approved in

(Chorley and Preston NHS Trusts Local Research Ethics Committee/Public Health Laboratory Service Ethics Committee for Investigations Involving Human Subjects)
where patients gave written, informed consent for investigation of serological responses to Cryptosporidium infection. Patients were screened for patent Cryptosporidium infection by immunoblot. Positive sera showed high level of expression of the 15/17 and 27-kDa Cryptosporidium immunodominant proteins (R. Chalmers, Unpublished). Subsequently, Cryptosporidium isolates were genotyped so it was possible to differentiate between C. parvum and C. hominis serological responses as described previously (Chalmers et al. 2009). The sera used in these studies was subsequently registered, reviewed and approved (Velindre NHS Trust Research Risk Review Committee) in 2008. A negative Cryptosporidium patient's serum was obtained from a previous study (Elwin et al. 2007). These sera were used to test the reactivity of native C. parvum and C. hominis antigenic preparations using Western blotting, according to the standard methods essentially as previously described (Tyler et al. 2009).

Results
Predicted properties of the Cops-1 and Chos-1 gene products Despite the high number of genes predicted by comparative genomics to be species specific (Kuo and Kissinger 2008;Bouzid et al. 2010a), secondary screening by PCR found that the vast majority were common to C. parvum and C. hominis (Bouzid et al. 2010a). Only two genes emerged from the secondary screen as potentially species specific; the C. parvum-specific gene Cops-1 (Cgd2_4380) and the C. hominis-specific gene Chos-1 (Chro.50011).
Strikingly these two genes have shared features. Both genes were telomeric, Cops-1 is the last gene annotated on chromosome 2 by the C. parvum genome project, while Chos-1 is the first gene annotated on Chromosome 5 by the C. hominis genome project (Fig. 1A). Both apparently encode secreted glycoproteins with very basic core peptides of roughly 50 kDa, and both proteins contain clear but distinctive internal repeats (Fig. 1B). Remarkably, the genes show limited (18.6% sequence identity), but noticeable, sequence similarity to each other along their entire length (Fig. 2) and identified each other as the top ranked and sole significant 'hits' on reciprocal blasts (using tBLASTn at cryptoDB.org) against the two genomes with P values of 0.001 in each case. This, taken together with the, lack of any such significant 'hit' against any of the other completed apicomplexan genomes identifies Cops-1 and Chos-1 as founder members of a novel Cryptosporidium-specific gene family. The initial annotation of the Cops-1 and Chos-1 genes provided by the C. parvum and C. hominis genome projects shows no apparent orthologs in other genomes, including C. muris. The limited annotation provided for Cops-1 describes a gene encoding a serine-rich protein containing repeated motifs, with an N-terminal secretory peptide, situated proximal to the telomeric repeats of chromosome 2. The telomeric location of these genes may have hindered their detection in other Cryptosporidium species as telomeric regions are highly repetitive and renowned to be difficult to assemble. In C. hominis and C. muris, there are no contigs covering the Cops-1 genomic region. In addition, the contigs in the assembled C. hominis and C. muris termini of chromosome 2 show no regions of obvious sequence similarity to the Cops-1 gene. Similarly, the termini of chromosome 5 in C. parvum and C. muris showed no regions of sequence similarity to the Chos-1 gene. With the exception of each other, no clear relatives for these proteins were found using sequence similaritybased searches, implying that this is a family of proteins specific to genus Cryptosporidium. However, interestingly, both Cops-1 and Chos-1 did show evidence for distant homology to the proteophosphoglycans of Leishmania.
Cops-1 and Chos-1 species-specific amplification from a panel of strains Consistent with our bioinformatics-based predictions, when screened against clinical samples (6) and reference strains (3), the full-length gene or a 665-bp diagnostic fragment of Cops-1 could be amplified from C. parvum DNA, but not from C. hominis DNA (Fig. 3A). Similarly, the full-length gene or a 287-bp diagnostic fragment of Chos-1 could be amplified from C. hominis isolates but not from C. parvum DNA (Fig. 3B). C. cuniculus, which is known to be closely related to C. hominis, amplified Chos-1 but not Cops-1, and C. meleagridis amplified Cops-1 but not Chos-1; no other Cryptosporidium spp. which we tested (C. andersoni, C. felis, C. ubiquitumand C. baileyi) gave amplification of either gene (not shown). All PCR product sequences (665 bp for C. parvum and C. meleagridis and 200 bp for C. hominis isolates) were deposited in GenBank [HQ667112-HQ667125]. In addition, when western blots of sera from C. parvum-infected patients were used to probe C. parvum and C. hominis protein extracts, they revealed a C. parvum-specific band of the expected 50 kDa in size (Fig. 3C), supporting the supposition of Cops-1 gene expression in C. parvum but not C. hominis. The primers we designed for diagnostic purposes to robustly amplify the 655-bp cops-1 gene fragment also amplified two minor bands from C. parvum DNA of 450 and 200 bp (Fig. 3A). This 200-bp band was also present in PCR products from C. hominis. Usefully, where this PCR is used diagnostically, this band can serve as an internal control for the presence of human infective Cryptosporidium. Interestingly, for C. meleagridis, PCR amplified a 665-bp product but no 450-or 200-bp bands. Thus, this assay, as it stands, confirms the presence of human infective Cryptosporidium and discriminates C. hominis from C. parvum and C. meleagridis.
In the light of the repeats contained within the Cops-1 gene, we considered whether the 200-bp band amplified might have arisen from alternative binding sites in the Cops-1 gene. Consistent with this, we found that the sequence obtained from both C. hominis and C. parvum showed complete (or highly similar for C. hominis) sequence identity to part of the Cops-1 665-bp PCR 18.6% identity Figure 2 Cops-1 and Chos-1 show some sequence similarity. Although the internal repeats present in each protein differ, clear similarities are present across the entire length of these proteins (E value), suggesting that they are both members of a novel family of proteins peculiar to the Cryptosporidium genus. product, indicating alternative binding sites in the Cops-1 gene for one of the primers and the likely presence of an orthologous gene in C. hominis. The gene was named CpCops-1 and ChCops-1 in C. parvum and C. hominis, respectively.
An ortholog of Cops-1 is present in C. hominis As the primers used to amplify full-length Cops-1 from C. parvum were unable to amplify the C. hominis ortholog, we tested with internal primers to determine which of the full-length primers was failing and found that the 5′ primer was able to amplify with an internal 3′ primer, but that the full-length 3′ primer could not amplify with internal 5′ primers. To amplify the full-length C. hominis gene, we designed a primer from the flanking region of the C. parvum gene. This primer, which was designed to telomeric repetitive sequence, was successful in amplifying the fulllength gene from the DNA of one of the C. hominis clinical isolates. The full length of ChCops-1 gene was determined using a primer walking approach (details of primers are shown in Table S1), and the full ChCops-1 sequence is available online (GenBank HQ667126). The full-length CpCops-1 and ChCops-1 are 1434 and 1263 bp, respectively. The two sequences exhibit 78.8% sequence identity. The difference in size between CpCops-1 and ChCops-1 corresponds to a truncation of approximately 170 bp in C. hominis. The alignment revealed 101 SNPs. This corresponds to an average of one SNP every 13 nucleotides. CpCops-1 and ChCops-1 encoded proteins are predicted to be 477 and 420 aa, respectively. The alignment of protein sequences is 70% identical (Fig. 4), with the N' terminus secretory signal peptide, and the vast majority of the putative glycosylation sites being conserved, thus maintaining the protein's general features. Based on the protein sequences alignment, 79 aminoacid substitutions were detected meaning 78.3% of SNPs are nonsynonymous.

Cops-1 contains sequence variation of potential taxonomic utility
The observation of sequence heterogeneity between C. parvum and C. hominis at the Cops-1 locus suggested potential taxonomic utility for the gene in subtyping of strains. Cops-1 PCR sequences from nine C. parvum, four C. hominis and one C. meleagridis isolates were analysed to assess the potential of this gene as a genotyping marker. The discriminatory power of Cops-1 sequence divergence was compared to genotyping using a standard C. parvum marker (Gp60) and a multi-locus subtyping approach (Bouzid et al. 2010a). Sequences from C. parvum isolates (Cp2, Cp3, Cp4, Iowa, Moredun) were identical to the published Cgd2_4380 gene sequence. Within C. parvum isolates, the anthroponotic subgroup (Gp60 subtype IIc) differed by four nonsynonymous SNPs. The 200-bp C. hominis PCR products, despite being shorter, showed high sequence similarity to PCR products from C. parvum isolates. Nevertheless, the alignment revealed four C. hominis-specific SNPs. Of the C. hominis SNPs, two were synonymous and two nonsynonymous. Overall, 6/8 (75%) of the SNPs detected were nonsynonymous. Sequence analysis of C. meleagridis 665-bpPCR product showed only one SNP difference comparing to C. parvum (nonanthroponotic) sequence and that this change was nonsynonymous. The 665-bp PCR product sequences from C. parvum were used to build a neighbour-joining (NJ) tree, which showed good discrimination of C. parvum and C. parvum anthroponotic subtype (Fig. 5A). The sequences of the 200-bp PCR product were also used to construct a NJ tree (Fig. 5B). This tree showed that the genetic polymorphism associated with this short fragment has sufficient discriminatory power to distinguish Cryptosporidium genotypes and subtypes.

Expression of recombinant CpCops-1
In expressing recombinant CpCops-1 for the purposes of antibody production and serodiagnosis, we found that 11/ 17 (64.7%) codons had at least 20% difference in the average codon usage between C. parvum and E. coli (Fig. S1), thus hindering the successful expression of the recombinant protein. Consequently, we compared expression of recombinant Cops-1 in specialized bacterial strains, finding only two of the bacterial strains tested: BL21-Codon 70% identity Figure 4 Cops-1 has an ortholog in Cryptosporidium hominis. In spite of being absent from the C. hominis genome sequence, primer walking allowed for amplification of a completed Cops-1 gene from C. hominis -chCops-1. This ortholog showed 70% identity with cpCops-1at an amino acid level but was shorter because of the loss of approximately 70 amino acids from the third of the protein's internal repeats. Colour coding of the amino acid sequence matches that used for Cops-1 in Fig. 1B. Plus-RP and RosettaTM2 allowed expression of the recombinant protein. Both strains have tRNAs for rare codons. The recombinant His-tagged protein observed was approximately 53 kDa, consistent with successful fusion to the N' terminal tag (Fig. S1).
The anti-Cops-1 monoclonal antibody 9E1 does not block epithelial cell invasion by C. parvum The potential role of Cops-1 protein in host-cell attachment and invasion was investigated in vitro using a blocking assay. Purified excysted Cryptosporidium oocysts were cultured on Caco-2 cell monolayers, and several intracellular multiplication stages were visible from stained slides after 24 h (Fig. 6A). For the blocking assay, oocysts were pre-incubated with our anti-CpCops-1 peptide IgG1 mouse monoclonal antibody -9E1. Control oocysts were incubated with PBS or nonspecific isotype-matched mouse IgG1 antibody. In each assay, parasite numbers were counted from 20 random fields and the mean number of parasites per field was determined (Fig. 6B). Pre-incubation of Cryptosporidium oocysts with 9E1 did not influence Caco-2 cells invasion. The number of parasites detected in cell monolayers following incubation of the parasites in the presence of 0.1 or 1 mg/mL 9E1 dissolved in PBS was not reduced relative to what was observed when the oocysts were incubated with PBS alone or with anisotype control (Fig. 6B).

Cops-1 is present inside the oocyst and associated with the surface of excysting sporozoites
The 9E1 anti-Cops-1 antibody was also used for immunolocalization of CpCops-1 protein. For IFA, a FITC-conjugated antibody that stains the oocyst cell wall of Cryptosporidium species and DAPI staining of the sporozoite nuclei were used as counter stains. When oocysts were permeabilized with methanol, 9E1 antibody enabled staining of oocyst contents both intact and broken oocysts with sporozoites still inside. However, no 9E1 staining of empty oocysts was observed (Fig. 7A, left). Subsequently, we looked at staining pattern on nonpermeabilized paraformaldehyde (PFA)-fixed oocysts and found that 9E1 did not stain intact or empty oocysts. However, staining was associated with sporozoites which were emergent from the oocyst or still inside broken ones (Fig. 7A, middle) and with some free sporozoites (Fig. 7A, right). Taken together these studies suggest that Cops-1 is a secreted protein associated with the sporozoite surface and the oocyst contents.
The 9E1 monoclonal antibody shows selective staining of purified C. parvum oocysts Directly conjugated 9E1 monoclonal antibody was also used to compare staining of C. hominis purified oocysts with C. parvum purified oocysts. In contrast to C. parvum, the contents of C. hominis oocysts stained weakly or not at all (Fig. 7B). It is important to note, however, that while the antibody can discern a clear difference between C. hominis and C. parvum on highly purified permeabilized oocyst preparations, the high level of background and nonspecific staining from this monoclonal antibody on faecal samples (not shown) means that it is unlikely to be immediately useful clinically for discriminating C. parvum from C. hominis infections.

Discussion
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.