Sequence Diversity of the Trypanosoma cruzi Complement Regulatory Protein Family

Compilation of websites utilized for sequence analysis.BLASTP (Blosum62 matrix) at www.ncbi.nlm.nih.gov/ was utilized to search GenBank for homologous copies of CRP. ClustalW at www.ebi.ac.uk/tools/clustalw/ was used to align paralogous copies of CRP. The SAPS tool at www.ebi.ac.uk/saps/index.html was used to determine the spacing between cysteine residues within individual proteins. BLAST 2 sequences at www.ncbi.nlm.nih.gov/blast/bl2seq/wblast2.cgi were used to determine sequence identity between CRP and RCA family members and CRP and glycoprotein C of herpes simplex viruses 1 and 2. The prediction algorithm NetNGlyc (version 1.0; Center for Biological Sequence Analysis, Technical University of Denmark [www.cbs.dtu.dk/services/NetNGlyc/ ]) was used to predict the potential for N-linked glycosylation among CRP paralogs and to determine the conservation of these sites between family members.

Parasites. T. cruzi trypomastigotes (Y strain) were passed in BALB/c mice as described previously (18) and used to initiate an infection in NIH 3T3 cells at a multiplicity of infection of 5 in Dulbecco's modified Eagle's medium buffered with 10 mM HEPES (pH 7.4) and supplemented with 5% fetal bovine serum, 5 mM l-glutamine, 0.2 mM sodium pyruvate, and 50 μg/ml gentamicin, all from Invitrogen. Tissue culture-derived trypomastigotes were recovered after 7 to 9 days from supernatants of infected cultures and used immediately for labeling.

Biosynthetic labeling and membrane preparation.Tissue culture-derived trypomastigotes were recovered from the supernatant of infected cells, washed twice in phosphate-buffered saline (PBS)-1% glucose, and resuspended at 10⁸ cells per milliliter in labeling medium (ICN Biochemicals). [³⁵S]methionine (Trans-Label; ICN Biochemicals) was added at 50 μCi/ml, and cells were incubated for 1 h at 37°C. After labeling, the cells were washed twice at 4°C in PBS-1% glucose. The final pellet was used to prepare trypomastigote membrane extracts as described previously (18).

C3b preparation and affinity purification of CRP.Human C3 was purified from fresh plasma, and C3b was prepared and coupled to Affi-Gel 10 (Bio-Rad Laboratories, Richmond, CA) as described previously (19). CRP was affinity purified from trypomastigote membrane extracts (19). A total of 50 mM NaHCO₃ (pH 10.5) was used to strip protein from beads by gentle agitation at room temperature for 20 min. Eluate was neutralized with a 1/10 volume of 1 M NaPO₄ (pH 6.8).

Immunoprecipitation of CRP following C3b affinity purification, sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and fluorography.Sera from patients with chronic Chagas' disease were used at 1:20 for immunoprecipitation assays. A total of 60 μl of affinity-purified material (as described above) was used for each immunoprecipitation. Immunoprecipitations were carried out as described previously (19). Samples were analyzed by SDS-PAGE (8.5%), followed by fluorography as described previously (18).

Cloning of crp genes from high-similarity group (HSG) and low-similarity group (LSG).Multiple different copies of the crp gene were obtained by PCR. T. cruzi Y strain chromosomal DNA was used as the template, and oligonucleotides CRP-F (AACATGTCCCGTCATGTGTTT) and CRP-R (TCACACCTCAGCAGAGAACC) were used to target the crp gene family. The annealing temperature was 59°C, and extension by Platinum Taq high fidelity (Invitrogen Corporation, Carlsbad, CA) was for 3 min at 72°C. The MgSO₄ concentration was 2 mM. A 3-kbp band was amplified, and these DNAs were cloned into pTOPO-XL (Invitrogen Corporation, Carlsbad, CA). Sequence analysis confirmed the isolation of several different crp genes.

Subcloning and sequence modifications of crp isolates.Two isolates designated CRP1.5 and CRP1.6 had 58% and 83% identity, respectively, to AAB49414 and were selected for further analysis. Eight amino acid codons at the 5′ end of each gene were replaced with eight amino acid codons of the rat preproinsulin signal sequence, as described previously (2), and the native GPI signal sequence was replaced with the GPI signal sequence from human DAF, as described previously (2). These constructs were subcloned into the eukaryotic expression vector pcDNA (Invitrogen Corporation, Carlsbad, CA) for further analysis.

Cell lines and medium.Freestyle 293 cells were purchased from Invitrogen Corporation, Carlsbad, CA. Cells were grown in Wheaton double-side-arm stirrer flasks on a magnetic stir plate at 121 rotations per minute in Freestyle 293 expression medium (Invitrogen Corporation, Carlsbad, CA) and maintained in a humidified atmosphere of 5% CO₂ at 37°C.

Transfections, phosphatidylinositol-specific phospholipase C treatment of cells, and sample concentration.Transfections were carried out according to the manufacturer's instructions by using 55 ml of medium, 10⁶ cells/ml, and 293fectin (Invitrogen Corporation, Carlsbad, CA) according to the manufacturer's instructions. Forty-two hours after the start of the transfection, cells were placed into fresh medium containing 700 μU phosphatidylinositol-specific phospholipase C (Molecular Probes, Inc., Eugene, OR) and incubated for 6 h at 37°C. At the completion of the incubation, cultures were spun at 300 × g in a Beckman Coulter Allegra 25R centrifuge for 5 min and supernatants were transferred to fresh tubes. A total of 15 ml of supernatant from each transfection was concentrated to approximately 0.5 ml in iCON concentrators (7 ml/20 kDa; Pierce, Rockford, IL) at 4°C according to the manufacturer's recommendation. Buffer exchange was achieved by reconstituting samples twice to the original volume in 10 mM phosphate 25 mM NaCl and concentrating to 0.5 ml. Concentrated samples were stored at 4°C until used.

DAA.Alternative pathway C3 convertase decay acceleration activity (DAA) was determined by enzyme-linked immunosorbent assay as described previously (10), with the following modifications: sample concentrations were 500 μg/ml, triplicate wells were averaged, and the DAA was performed for 20 min. C3b and factors B, D, and H (positive control) were purchased from Complement Technology, Inc., Tyler, TX. Goat anti-factor B and horseradish peroxidase-conjugated rabbit anti-goat immunoglobulin were purchased from Calbiochem, San Diego, CA.

Comparative sequence analysis to determine the approximate number of crp genes in the T. cruzi genome.Previously, we cloned and sequenced a full-length copy of a T. cruzi crp (20) and demonstrated that this gene encodes a functional CRP (2). We used this CRP (GenBank accession number AAB49414) as the reference sequence to analyze the T. cruzi genome for paralogs.

The CL-Brener strain is a hybrid strain of T. cruzi that is well characterized experimentally and was used for the T. cruzi genome-sequencing project (9), and protein and DNA sequences from this project are accessible through GenBank. A GenBank BLASTP search, with CRP copy AAB49414 derived from T. cruzi Y strain, revealed 14 full-length CRP paralogs and 6 partial-length paralogs from the CL-Brener strain as well as 1 full-length paralog and 2 partial-length paralogs from the CL strain. All paralogs had E values of 0, indicating that there is no chance that these sequences were randomly related to AAB49414. Table 1 shows the sequence identities and similarities of the CL-Brener proteins relative to AAB49414. A full-length copy of AAB49414 was not present in this strain; instead, two partial protein sequences (XP_804086 and XP_805135) with 99% identity to AAB49414 were identified. Overlapping sequence between these partial protein sequences showed that they were unique proteins, suggesting that at least two very similar copies of this gene exist in the CL-Brener strain. Because the CL-Brener strain is a heterozygote hybrid at most loci (9), these two proteins might represent heterozygote copies of the same locus. AAA30196, CAA50255, and CAA50289 represent three protein sequences identified from the CL strain; these proteins were 100%, 98%, and 93% identical, respectively, to full-length proteins in the CL-Brener strain. Additionally, of the six CL-Brener partial protein sequences with E values of 0, three were amino terminus, partial-length sequences and three were carboxy terminus, partial-length sequences, but overlapping sequence showed they were six unique proteins. An additional partial-length protein sequence (XP_804894), with a significant E value, but not equal to 0, was too short to determine whether it was unique or the amino terminus of another partial-length CRP. Whether any of these partial-length protein sequences represent true partial-gene copies or result from the inherent difficulty in sequencing this highly repetitive genome will require experimental confirmation. Nevertheless, our data suggest that 14 to 20 copies of the crp gene (representing 7 to 10 loci) are present in a single strain of T. cruzi and several paralogs are retained between strains.

Sequence diversity suggests two distinct groups of CRPs.The CRP paralogs fell into two distinct groups. Seven full-length copies with more than 80% identity and more than 86% similarity to AAB49414 were designated the HSG. Seven full-length copies with sequence identity between 54 to 62% and similarity between 64 to 71% to CRP (AAB49414) were designated the LSG. A ClustalW alignment of the HSG paralogs revealed shared sequence identity and similarity of approximately 60% and 84%, respectively, along the length of the polypeptides (Fig. 1). A similar alignment of the paralogs in the LSG showed strong identity at the amino termini, with shared identity and similarity of approximately 48% and 72%, respectively, and limited sequence identity (7%) and low similarity (33%) at the carboxy termini. Additionally, of the seven partial copies, three fell in the HSG and four were in the LSG. Interestingly, the bulk of the insertions or deletions were located in the carboxy termini of these proteins.

• Open in new tab
• Download powerpoint

FIG. 1.

Sequence identity/similarity (I/S) of the HSG and the LSG of CRPs relative to other members of each group. Relative to the AAB49414 sequence, the sialidase domain is 400 amino acids, the central region is 275 amino acids, and the carboxy-terminal region is 279 amino acids.

Among the CRPs, cysteine residues are conserved and localized to the amino termini.The spacing between conserved cysteines in the prototype CRP sequence, AAB49414, and the 14 full-length CRP paralogs from the CL-Brener strains showed a unique spacing pattern between conserved cysteines (Fig. 2). Of 10 cysteines in AAB49414, 5 were absolutely conserved in all paralogs, with two additional cysteines conserved among the HSG paralogs. No cysteines were found in the carboxy-terminal third of AAB49414, and few cysteines were located in this region for any of the 14 paralogs.

• Open in new tab
• Download powerpoint

FIG. 2.

Spacing between conserved cysteines is unique among TS superfamily subfamilies. Representative subfamilies of the TS superfamily show spacing between conserved cysteine residues. Bold reference sequences in the CRP family represent proteins from the LSG.

We also analyzed the primary amino acid sequences of several representative proteins from the various subfamilies of the TS superfamily. A unique pattern of spacing between conserved cysteine residues was found for each subfamily (Fig. 2). The TS and CRP subfamilies had the most similar pattern of spacing between conserved cysteines, and this result may indicate that these two subfamilies are most closely related.

Relationship of T. cruzi CRP to mammalian RCA and viral inhibitors of complement.CRP is functionally equivalent to DAF. Sequence alignment of CRP to DAF and other RCA family members (factor H, C4bp, MCP, CR1, and CR2) showed that CRP lacks primary sequence identity with the RCA. RCA proteins are composed primarily of CCP modules of about 60 amino acids in length. The spacing between cysteine residues in CCP modules is highly conserved and is probably an essential element in the structure of these modules. No CCP modules have been identified for T. cruzi CRP. Nevertheless, several cysteines are conserved among the putative CRP paralogs and the spacing between them is unique to this TS-like subfamily. CCP modules also share a conserved tryptophan (W) residue and well-conserved glycine (G), tyrosine (Y), and phenylalanine (F) residues. W is the only amino acid that is statistically more conserved within a protein family than C is (8). A ClustalW alignment of all 15 full-length CRP paralogs revealed that 87% of the W, 76% of the F, 60% of the Y, and 35% of the G amino acids were conserved throughout the family (see the supplemental material). In comparison, conservation values for methionines, arginines and prolines were 33%, 32% and 19%, respectively. Additionally C, W, F, and Y were absent or underrepresented in the carboxy-terminal third of these proteins. Other hydrophobic amino acids, such as isoleucine, leucine, and valine, are well represented in the carboxy termini of the CRPs, and the significance of this representation is not understood.

Viral homologs from the Poxviridae and Herpesviridae as well as from glycoprotein C of herpes simplex viruses 1 and 2 also lacked sequence identity or similarity to the CRP (data not shown).

Sites for asparagine-linked glycosylation are conserved among putative CRP paralogs.Previous work in our laboratory suggested that asparagine (N)-linked glycosylation may be important for CRP binding to C3b and C4b (19). Among the HSG members, four N-linked glycosylation sites were conserved, and two of these were also conserved among the LSG members (Fig. 3). Conserved sites for the addition of N-linked glycans were located exclusively within the amino-terminal two-thirds of these proteins, intimating a role of N-linked glycosylation in the structure and or function in this part of the molecules.

• Open in new tab
• Download powerpoint

FIG. 3.

Conserved N-linked glycosylation sites between LSG and HSG of CRP and regions of highest and lowest sequence identity between group members.

Relationship of CRP to active trans-sialidase. CRP (AAB49414) and the active TS (BAA09334) share 31% primary sequence identity.A ClustalW alignment of the 15 full-length CRP paralogs to this T. cruzi TS molecule revealed that 7 of 21 TS putative active-site residues (25) were conserved among all CRP paralogs. An alignment of CRP AAB49414 to the active TS (BAA09334) (Fig. 4) illustrates these conserved residues. The conserved putative active-site residues are shown for both sequences. Aspartic acid (Asp) boxes (SXDXGXTW) are common features among bacterial sialidases, trypanosomal sialidases, and trans-sialidases (7, 22). Asp boxes are often found among enzymes that have polysaccharides as substrates, and it has been suggested that they are important structural elements of these proteins (5). TS Asp boxes are shown in Fig. 4. No Asp boxes were found among CRP paralogs, supporting previous experimental data that state CRP does not have trans-sialidase activity (K. A. Norris, unpublished data).

• Open in new tab
• Download powerpoint

FIG. 4.

Alignment of BAA09334 (active TS) and AAB49414 (functional CRP). Asp boxes and FRIP boxes are highlighted in gray. Putative TS active-site residues are bold and underlined. Conserved tryptophan and cysteine residues are bold and italicized. Asterisks indicate identity, colons indicate conservative substitution, and periods indicate semiconservative substitution between aligned proteins.

CRPs from both HSG and LSG retain complement regulatory activity.An enzyme-linked immunosorbent assay-based DAA was used to determine AP complement regulatory activity for the recombinant CRPs expressed in F293 cells. The assay was performed in triplicate in multiple different experiments. Figure 5 shows a representative experiment with different CRP paralogs, including representative clones from the HSG and the LSG. All paralogs were capable of accelerating the decay of the AP C3 convertase compared to that of protein derived from vector-transfected cells. A two-tailed Student's t test that assumed nonpaired equal variance was used to compare each test group to the negative control. P values were 0.011, 0.009, 0.0038, and 0.00024 for CRP, CRP1.5, CRP1.6, and factor H, respectively.

• Open in new tab
• Download powerpoint

FIG. 5.

Decay acceleration of the AP convertase. CRP and CRP1.6 are members of the HSG, and CRP1.5 is a member of the LSG. Factor H serves as a positive control.

Multiple different CRP molecules are synthesized by a single parasite strain and differentially recognized by Chagasic patient sera.Although multiple copies of the complete crp gene are contained in the T. cruzi genome, it is not clear how many functional copies of the protein are produced simultaneously. To address this question, [³⁵S]methionine and cysteine-labeled CRP from a clonal line of T. cruzi (Y strain) was affinity purified with human complement protein C3b as described previously (20). Purified CRP was subjected to immunoprecipitation with several different chronic Chagasic sera, and proteins were analyzed by SDS-PAGE, followed by fluorography as described previously (21). Figure 6 shows the distinct migration patterns observed among the CRPs precipitated with different patient sera. These results suggests that multiple variants of the CRP are synthesized by a single strain, that these variants were capable of binding human C3b, and that different patients had predominant antibody reactivity to different variants.

• Open in new tab
• Download powerpoint

FIG. 6.

Immunoprecipitation of C3b affinity-purified CRP in Chagasic patient sera reveals multiple variants expressed by a clonal strain of T. cruzi. CRP was affinity purified from a [³⁵S]methionine- and cysteine-labeled TCT lysate and was immunoprecipitated with normal human sera (lane 1) or different Chagasic patient sera (lanes 2 to 8).