Entamoeba histolytica Lectins Contain Unique 6-Cys or 8-Cys Chitin-Binding Domains

Abstract

The Jacob lectin, the most abundant glycoprotein in the cyst wall of Entamoeba invadens, contains five unique 6-Cys chitin-binding domains (CBDs). We identified Entamoeba histolytica and Entamoeba dispar genes encoding Jacob homologues, each of which contains two predicted 6-Cys CBDs. A unique 8-Cys CBD was found at the N termini of the E. histolytica chitinase and three other predicted lectins, called Jessie 1 to Jessie 3. The CBDs of four E. histolytica lectins (Jacob, chitinase, and Jessies 2 and 3) were expressed in secretory vesicles of transfected amebae and shown to bind to particulate chitin.

The diagnostic and infectious stage of Entamoeba histolytica is the quadranucleate cyst, which has a wall composed in part of chitin (2). While E. histolytica causes dysentery and liver abscess, Entamoeba dispar (formerly known as nonpathogenic E. histolytica) is morphologically identical and does not cause disease (7). Because E. histolytica does not encyst in axenic culture, the reptilian pathogen Entamoeba invadens is used to model amebic encystation (10). The most abundant E. invadens cyst wall protein (Jacob) is a unique lectin composed of five tandemly arranged chitin-binding domains (CBDs), each of which has six Cys residues and conserved aromatic amino acids (11). The five CBDs in the amebic Jacob lectin presumably cross-link and stabilize chitin fibrils in the amebic cyst wall in the same way that CBDs of insect peritrophins cross-link chitin fibrils in the wall that surrounds the insect blood meal (11, 24). Although CBDs of peritrophins and insect chitinases also contain six Cys residues and are related to each other by a common ancestry, the 6-Cys CBDs of the E. invadens Jacob lectin are not related to them by a common ancestry (convergent evolution) (9, 11, 24).

Although neither chitin synthase nor chitinase is present in E. invadens trophozoites, both enzymes are expressed by encysting amebae (5, 28). Amebic chitinases have a series of hydrophilic heptapeptide repeats between an N-terminal signal sequence and a C-terminal glycohydrolase domain (6, 14, 16). The chitinase heptapeptide repeats, which vary from 4 to 18 repeats in different isolates of E. histolytica and Entamoeba dispar, have an amino acid composition similar to that of spacers or hinges present between CBDs of the Jacob lectin (11, 13). This similarity led us to identify a putative 8-Cys CBD near the N terminus of the E. histolytica and E. invadens chitinases. Plant chitinases also have 8-Cys CBDs at their N termini, which presumably help the chitinases bind to the chitin fibrils that the glycohydrolase domain will degrade (3).

We expressed the 8-Cys CBD of E. histolytica chitinase in trophozoites and showed that it does bind to chitin. A survey of shotgun sequences of the E. histolytica genome showed that this 8-Cys CBD is also present at the N termini of three hypothetical proteins (Jessie 1 to Jessie 3). The CBDs of Jessie 2 and Jessie 3 were expressed and shown to bind chitin. In this study, we also identified homologues of the E. invadens Jacob lectin in E. histolytica and E. dispar and showed that the E. histolytica Jacob is a chitin-binding protein.

A predicted E. histolytica Jacob lectin has two 6-Cys, chitin-binding domains.

The peptide sequence of the E. invadens Jacob lectin and TBLASTN were used to search 49,000 E. histolytica HM-1:IMSS strain genomic DNA sequences, which were deposited in the Genome Survey Sequences database of the National Center for Biotechnology Information by one of the coauthors (B. Loftus) (1, 11). A predicted E. histolytica Jacob lectin, which was 151 amino acids long with a predicted 21-amino-acid signal sequence at its N terminus, showed 23% amino acid identity with a corresponding region of the E. invadens Jacob lectin (Fig. 1A) (11, 22, 26). The two predicted CBDs of the E. histolytica Jacob each contained six Cys residues, which might form three disulfide bonds as described for CBDs of insect peritrophins and insect, nematode, and fungal chitinases (18, 24, 27). Like the E. invadens Jacob lectin, the predicted E. histolytica Jacob lectin did not appear to contain any transmembrane domains, which might anchor the peptide in the plasma membrane (25). The predicted E. histolytica Jacob CBDs did contain numerous aromatic amino acids (Tyr, Phe, and Trp), which have been implicated in carbohydrate binding by plant lectins (3, 29). There was a predicted N-linked glycosylation site in the second CBD of the E. histolytica Jacob lectin, as well as two predicted N-linked glycosylation sites in the hydrophilic spacer between the two CBDs (17). These N-linked glycosylation sites suggest that the E. histolytica Jacob lectin may be a glycoprotein like the E. invadens Jacob lectin (11). Northern blot analysis suggested that mRNAs for the E. histolytica Jacob lectin are not transcribed by E. histolytica trophozoites, which do not encyst in axenic culture (data not shown). We were unable to obtain fresh E. histolytica cysts for Northern blot analysis. Jacob mRNAs are also absent from E. invadens trophozoites but are present in E. invadens cysts (11).

FIG. 1.

(A) Primary structure in single letter code of the putative E. histolytica (Eh) Jacob lectin aligned with homologous regions of the E. dispar (Ed) and E. invadens (Ei) Jacob lectins (residues 107 to 233). Sequence differences between E. histolytica and E. dispar Jacob lectins are shown in purple, while signal sequences, identified with SignalP (22), are shown in gray. Conserved Cys residues in the putative CBDs are shown in red, while other residues identical in all Jacob lectins are shown in blue. Spacers between putative CBDs are shown in green, while potential sites of N-linked glycosylation are shown in light orange. (B) Confocal micrograph of the E. histolytica Jacob expressed with a myc tag in transfected E. histolytica and identified with anti-myc monoclonal antibodies and Alexa Fluor 488-labeled anti-mouse IgG antibodies (staining methods are described in reference 14). (C) Nuclear staining is seen when the cells in panel B are stained with the Alexa Fluor488-labeled anti-mouse IgG antibodies alone. (D) Coomassie blue-stained gel of total proteins from nontransfected E. histolytica trophozoites, trophozoite proteins bound to chitin beads, and trophozoite proteins that did not bind to chitin beads. (E) Western blot analysis of chitin binding by E. histolytica Jacob, detected with anti-myc antibodies and chemiluminescence.

To test the function of the putative 6-Cys CBDs of the E. histolytica Jacob lectin, E. histolytica trophozoites, which do not express the Jacob lectin (above), were transfected with a vector that contained a modified E. histolytica Jacob gene between 5′ and 3′ untranslated regions of the E. histolytica actin gene. The vector and methods were the same as those used previously to express and localize intact and modified chitinases in E. histolytica trophozoites (14). The modified Jacob lectin, which was detected with a monoclonal antibody to a c-myc epitope tag placed at its C terminus, was present in small vesicles of transfected amebae (Fig. 1B). These small vesicles resemble secretory vesicles, which contain the Jacob lectin in encysting E. invadens (11). When transfected trophozoites were lysed in 0.1% Triton X-100, the myc-tagged Jacob lectin bound tightly to particulate chitin and could be eluted only by boiling in 1% sodium dodecyl sulfate-5% 2-mercaptoethanol (Fig. 1E). In contrast, Coomassie blue-stained gels showed that the vast majority of trophozoite proteins do not bind to chitin (Fig. 1D). These results demonstrate that the E. histolytica Jacob lectin binds chitin, most likely through both of its 6-Cys CBDs (although we cannot rule out the possibility that one of the 6-Cys CBDs is not functional). While the E. invadens Jacob lectin has five CBDs and could theoretically cross-link five chitin fibrils, the E. histolytica and E. dispar (next section) Jacob lectins each have two CBDs and could cross-link two chitin fibrils. In the same way, the number of CBDs in insect peritrophins, which cross-link chitin fibrils in the wall around a blood meal, have also been shown to vary between species (24).

An E. dispar Jacob lectin also has two putative 6-Cys, chitin-binding domains.

The E. histolytica jacob gene was used to identify an E. dispar jacob gene from a genomic DNA library, which was a generous gift of Michael Duchene (23). The predicted E. dispar Jacob lectin, which was 150 amino acids long with a predicted 21-amino-acid signal sequence, showed 85% amino acid identity with the E. histolytica Jacob lectin (Fig. 1A). Most of the amino acid differences between the E. histolytica and E. dispar Jacob lectins were clustered in the first CBD, while two of three predicted N-linked glycosylation sites were conserved in the E. dispar Jacob lectin. No other homologues of the Jacob lectins were identified in GenBank NR, dBEST, Genome Survey Sequences, and unfinished microbial databases, using BLASTP or TBLASTN (1). These results suggest the Entamoeba Jacob lectins are unique, although they serve the same function (cross-linking chitin fibrils) as insect peritrophins (24). Whether monoclonal antibodies to Jacob lectins may be used to discriminate cysts of E. histolytica from those of E. dispar in patients' stools remains to be determined.

The E. histolytica chitinase has a unique 8-Cys chitin-binding domain at its N terminus.

An alignment of the N termini of E. histolytica and E. invadens chitinases suggested the presence of a putative CBD containing eight Cys residues, which might form four disulfide bonds as in plant lectins (Fig. 2A) (3, 6, 19, 29). Although the signal sequences and hydrophilic repeats of the E. histolytica and E. invadens chitinases were not alignable, the putative CBDs shared 42 of 64 amino acids (66%) (6, 22). Further, the CBDs have eight positionally identical aromatic amino acids (Tyr, Phe, and Trp), which might be involved in carbohydrate binding (3, 29). Although plant lectins and chitinases also have 8-Cys CBDs, the amebic chitinase 8-Cys CBDs were not related to them by a common ancestry (1, 3, 19, 29). This is another example of convergent evolution (9), as has been argued for the 6-Cys CBDs of the Jacob lectin and insect peritrophins (11, 24). The heptapeptide repeats of the E. histolytica and E. dispar chitinases, which vary so much from one clinical isolate to the next (13), are likely spacers or hinges between the CBD and the catalytic domain.

FIG. 2.

(A) Primary structures in single letter code of the N-terminal regions of E. histolytica (Eh) and E. invadens (Ei) chitinases, as well as three E. histolytica hypothetical lectins (Jessie 1 to Jessie 3). The complete sequences of Jessie 1 and Jessie 2 are marked by asterisks, while C-terminal sequences of chitinases and Jessie 3 have been omitted. Signal sequences, identified with SignalP (22), are shown in gray, while conserved Cys residues in CBDs are shown in red. Other amino acids identical in all five CBDs are shown in blue, while residues identical in at least three of five CBDs are shown in dark red. Hydrophilic spacers are shown in green, while potential N-linked glycosylation sites in Jessie 1 and Jessie 2 are shown in light orange. (B) Confocal micrograph of the E. histolytica chitinase CBD expressed in transfected E. histolytica trophozoites and identified with antichitinase antibodies and Texas Red-labeled anti-rabbit IgG antibodies, using methods described in reference 14. (C to E) Confocal micrographs of the myc-tagged Jessie 1 (C), Jessie 2 (D), and Jessie 3 CBD (E) expressed in transfected E. histolytica trophozoites and identified with anti-myc antibodies and Alexa Fluor 488-labeled anti-mouse IgG antibodies. (F) Western blot analysis of chitin binding by the full-length E. histolytica chitinase, the E. histolytica chitinase CBD, Jessie 1, Jessie 2, and the Jessie 3 CBD, detected with antichitinase or anti-myc antibodies and chemiluminescence.

To test the function of the putative chitinase 8-Cys CBD, E. histolytica trophozoites, which do not express chitinase (6), were transfected with an E. histolytica gene encoding a truncated chitinase, which included the signal sequence, putative CBD, and heptapeptide repeats. The vector and methods were the same as those used previously to express and localize intact and modified chitinases in E. histolytica trophozoites (14). The truncated chitinase, which was detected with polyclonal rabbit antibodies to the heptapeptide repeats, was present in small secretory vesicles that were similar in appearance to those associated with intact chitinase (14) (Fig. 2B). When transfected trophozoites were lysed in 0.1% Triton X-100, the truncated chitinase bound tightly to particulate chitin and could be eluted only by boiling it in 1% sodium dodecyl sulfate-5% 2-mercaptoethanol (Fig. 2F). A positive control for chitin binding was the full-length chitinase protein (Fig. 2F). A negative control for chitin binding was a modified amebic Fe-hydrogenase, which was epitope tagged with the chitinase heptapeptide repeats (data not shown) (12). The truncated chitinase did not bind to cellulose or GlcNAc conjugated to agarose (data not shown), suggesting that the chitinase 8-Cys CBD is specific for chitin. It is possible that the 8-Cys CBD increases the activity of the amebic chitinase, in the same way that an 8-Cys CBD increases the activity of a plant chitinase (30).

Three hypothetical E. histolytica lectins (Jessie 1 to Jessie 3) contain 8-Cys CBDs like those of chitinases.

The hypothetical lectins Jessie 1 to Jessie 3, which were identified by searching the E. histolytica shotgun sequences with the chitinase CBD and TBLASTN, also contained an 8-Cys putative CBD just distal to a predicted signal sequence (Fig. 2A) (1). Twenty amino acids (31%), including the eight Cys residues, were strictly conserved in the 64-amino-acid CBDs of chitinases and Jessies (26). An additional 24 amino acids (38%), including numerous aromatics, were conserved in the majority of chitinase and Jessie CBDs. Jessie 1 and Jessie 2, which were 90 and 97 amino acids long, respectively, each contained a single site for N-linked glycosylation and little else C-terminal to the CBD. Jessie 3, which was 621 amino acids long, contained a series of hydrophilic amino acids C-prime to the 8-Cys CBD, similar to the hydrophilic repeats that precede the catalytic domain of amebic chitinases (6). However, the C-terminal domain of Jessie 3, which was 469 amino acids long, showed no homology to the catalytic domains of chitinases or any other proteins in GenBank (1). Northern blots with probes specific to each hypothetical Jessie gene suggested that mRNAs for all of the Jessie lectins are absent from E. histolytica trophozoites (data not shown).

To test the function of the 8-Cys CBDs of the Jessie lectins, the entire Jessie 1 and Jessie 2 lectins and the putative CBD of Jessie 3 were expressed with a c-myc epitope tag in transfected E. histolytica trophozoites by methods described above for the Jacob lectin. While all three Jessie lectins were present in small vesicles of transfected amebae (Fig. 2C to E), only Jessie 2 and the Jessie 3 CBD bound to particulate chitin (Fig. 2F). It is not clear why Jessie 1 failed to bind chitin.

The presence of the same 8-Cys CBD in at least four different proteins (chitinase and three Jessie lectins) suggests the gene encoding this unique motif has been duplicated and reinserted into the E. histolytica genome at least three times. Similarly, the 8-Cys CBDs of plants are present in chitinases and lectins (3, 19, 29) and the 6-Cys CBDs of insects are present in chitinases and peritrophins (24), while a Cys-rich sialic acid-binding domain is conserved in the P. falciparum 175-kDa erythrocyte binding protein and other homologous proteins in Plasmodia (15).

Summary.

We have identified Entamoeba genes encoding proteins that contain unique 6-Cys or 8-Cys CBDs. A model showing the 6-Cys CBDs of Entamoeba Jacob lectins and the 8-Cys CBDs of chitinase and Jessie lectins is shown in Fig. 3. The CBDs of Entamoeba are similar to other known CBDs in that they have conserved Cys and aromatic residues, but they are not related by a common ancestry (convergent evolution) (3, 9, 11, 24, 29). The 6-Cys and 8-Cys CBDs are much easier to define than are the carbohydrate recognition domains of the large and intermediate subunits of the E. histolytica Gal/GalNAc lectin (4, 8, 20, 21). Future studies beyond the scope of the present experiments will attempt to determine whether Jacob and Jessie lectins are located in the E. histolytica cyst wall.

FIG. 3.

Cartoons of Entamoeba lectins with unique 6-Cys (Jacob) and 8-Cys (chitinase and Jessie) CBDs.

Nucleotide sequence accession numbers.

The sequences of E. histolytica and E. dispar Jacob lectins have been deposited in GenBank under accession numbers AF401984 and AF401985, respectively. The sequences of the E. histolytica Jessie lectins have been deposited in GenBank under accession numbers AF401986 to AF401988.