INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

Entamoeba histolytica is a protozoan parasite that causes amebic colitis and liver abscess in developing countries such as Mexico, India, and Bangladesh (6, 17, 24, 32). The HM-1 strain of E. histolytica, which was isolated from a dysenteric patient in Mexico more than 30 years ago, still causes disease in experimental animals and has been used for nearly all immunological, biochemical, and molecular biological studies of amebae (14). E. histolytica is morphologically indistinguishable from Entamoeba dispar, which remains in the colonic lumen and so does not cause disease (8, 39, 44). Sequences of E. histolytica and E. dispar small-subunit rRNA genes differ by 1.7%, suggesting that the parasites diverged from each other tens of millions of years ago (9, 29, 40).

The haploid genome of E. histolytica contains 14 chromosomes and totals ~20 Mb (47). Homologous chromosomes vary in length and number (from one to four), suggesting that parasites may be polyploid (47). E. histolytica coding regions are AT rich, contain few introns, and are separated by relatively short intergenic regions (4, 22, 43). 5' untranslated regions of amebic genes contain conserved TATA, CAAT, GAAC, and initiation sequences (5, 41). Two palindromic copies of each rRNA gene are present on 24-kb episomal plasmids, which are present in 100 to 200 copies per nucleus (40).

The relationships between different clinical isolates of E. histolytica are not known. No monoclonal antibodies have been able to distinguish isolates of E. histolytica (17). Isoenzyme groups or zymodemes have not proven useful for differentiating strains of E. histolytica, as the majority of strains fall into two main zymodeme groups (3, 39). In addition, sequences of the genes encoding hexokinase and phosphoglucomutase, which were used to distinguish parasites, failed to reveal any heterogeneity among E. histolytica and E. dispar isolates, respectively (30, 31). Although sequences of internal transcribed spacers (ITS) between rRNA genes discriminate bacterial isolates and strains of Leishmania spp., ITS sequences failed to distinguish isolates of E. histolytica or isolates of E. dispar from Mexico City, Mexico (11, 21, 28). In contrast, axenic strains of E. histolytica have been distinguished by restriction fragment length polymorphisms of PCR products of the genes encoding the Ser-rich E. histolytica protein, also known as the K2 protein (10, 20, 42). The Ser-rich protein, which is present on the surface of amebae, is composed of an N-terminal signal sequence and a hydrophobic C-terminal anchor, surrounding a series of tetrapeptide and octapeptide repeats of hydrophilic and acidic amino acids. The Ser-rich protein is an important amebic vaccine candidate, and antibodies to the Ser-rich protein correlate with infection (27, 46, 49). The amebic chitinase, which is expressed only by cysts, also contains a series of acidic and antigenic repeats between a putative N-terminal lectin domain and a C-terminal half catalytic domain (12).

While little is known about the molecular epidemiology of amebae, much has been learned from numerous excellent molecular epidemiological studies of Plasmodium falciparum, the cause of severe malaria. First, there was apparently a recent bottleneck in the evolution of P. falciparum infecting humans, such that parasites from all over the world are identical in sequences of genes encoding housekeeping proteins (34). This bottleneck may have occurred as recently as 7,000 years ago. Second, malaria genes encoding repeat antigens on the parasite surface (e.g., circumsporozoite protein, merozoite surface protein 2, or merozoite S antigen) are extraordinarily polymorphic (2, 16, 26). Third, most of the diversity of circumsporozoite protein genes is secondary to DNA slippage rather than meiotic recombination, while maintenance of variation is likely caused by immune selection (33).

To better understand the genetics and molecular epidemiology of E. histolytica and E. dispar, we asked three questions in these studies. First, what is the ploidy of E. histolytica? Second, can intergenic sequences between superoxide dismutase and actin 3 (sod-actin) genes be used to distinguish isolates of E. histolytica and E. dispar from Mexico City, San Diego, Calif., and Calcutta, India? Third, can isolates of E. histolytica and E. dispar be distinguished by sequences of genes that encode chitinase and the Ser-rich protein?

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Fluorescence in situ hybridization (FISH) of E. histolytica trophozoites. The HM-1 strain of E. histolytica was grown axenically at 37°C in TYI medium. For production of parasites with condensed chromosomes, amebae were incubated in 7 mg of colchicine per ml for 12 h, fixed in methanol-acetic acid (3:1), treated with 20 µg of RNase per ml for 30 min, and stained with 0.3 µg of propidium iodide per ml in Antifade (Oncor). For FISH, parasites were washed in phosphate-buffered saline, swelled in 70 mM potassium chloride, fixed in methanol-acetic acid, and dropped onto uncoated slides. DNA was denatured by dipping slides in 70% formamide at 70°C. E. histolytica genes encoding chitinase, pyruvate:ferredoxin oxidoreductase, nicotinamide nucleotide transhydrogenase, plasma membrane calcium-transporting ATPase, cysteine proteinase 5, and P-glycoprotein 6 genes were labeled by nick translation with biotin using an Oncor kit (12, 13, 15, 18, 37, 48). A negative control was pBluescript without an insert. Hybridizations were performed with 4 µg of each biotinylated probe per ml in 30% formamide and 2× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate) at 37°C for 16 h (25). Slides were washed in the same buffer, 2× SSC, and phosphate buffer plus detergent (Oncor). Probes were detected with fluorescein isothiocyanate-avidin and amplified with antiavidin antibody, followed by a second incubation with fluorescein isothiocyanate-avidin. Parasite nuclei were counterstained with propidium iodide, and slides were examined with a Leitz Orthoplan epifluorescence microscope or a Leica confocal microscope.

Isolation of amebic DNA from clinical isolates of E. histolytica and E. dispar. Stools containing amebae as shown by light microscopy came from patients presenting to (i) the Hospital Infantil in Mexico City or a neighborhood clinic in Netzahualcoyotl, which is a barrio of 100,000 persons outside Mexico City; (ii) the Kothari Medical Center in Calcutta; or (iii) the University of California at San Diego. Other parasite DNA came from axenized E. histolytica (HM-1, HK-9, and NIH-200) and E. dispar (SAW 760) strains. For the most part, amebae from clinical isolates were cultured in Robinson's medium, concentrated by low-speed centrifugation, washed in phosphate-buffered saline, and lysed in 1% sodium dodecyl sulfate-50 mM EDTA-50 mM Tris, pH 8 (1, 35). Amebic DNA was then extracted with glass milk (Elutip; Schleicher and Schuell) (36). On some occasions, DNA was isolated directly from stool parasites, which were enriched by low-speed centrifugation and lysed in the same buffer (19). Identification of isolates as E. histolytica or E. dispar was made by restriction fragment length polymorphism analysis of PCR products using primers that spanned the ITS between rRNA genes (28). The E. histolytica rRNA ITS PCR product was cut with RsaI, while E. dispar rRNA ITS PCR products were cut with EcoRV.

PCR amplification and sequencing of sod-actin, chitinase, and Ser-rich protein genes. The intergenic regions between amebic sod-actin genes of E. histolytica and E. dispar were amplified using PCR and the same set of primers (see Fig. 1) (4). A sense PCR primer (TTGGTGGAATGTAGTCAACTG) was located at the 3' end of the coding region of the superoxide dismutase gene. An antisense primer (AAATCCGGCTTTACACATTCC) bound to a sequence located at the 5' end of the coding region of the actin 3 gene. The amebic chitinase gene repeats were amplified using PCR and the same antisense primer (TCTGTATTGTGCCCAATT) for both E. histolytica and E. dispar (12). An E. histolytica chitinase-specific sense primer was GGAACACCAGGTAAATGTATA. An E. dispar chitinase-specific sense primer was GGAACACCAGGTAAATGCCTT. The amebic Ser-rich protein gene repeats were amplified using PCR and primers specific for E. histolytica and E. dispar, respectively. An E. histolytica Ser-rich protein-specific sense primer was GCTAGTCCTGAAAAGCTTGAAGAAGCTG, while an E. histolytica Ser-rich protein-specific antisense primer was GGACTTGATGCAGCATCAAGGT (20, 42). An E. dispar Ser-rich protein-specific sense primer was AGATACTAAGATTTCAGTC, while an E. dispar-specific Ser-rich protein antisense primer was CATAATGAAAGCAAAGAG (20).

Methods for alignments of Ser-rich protein and chitinase gene repeats. Common algorithms for aligning sequences are not adequate for sequences containing numerous degenerate repeats. Therefore, amebic chitinase and Ser-rich protein repeat sequences were coded by methods used to compare P. falciparum circumsporozoite gene sequences (33). Briefly, each repeat that predicted a unique amino acid sequence was given a number. Silent changes in nucleotide sequences of each repeat were identified, and the numbers (names) of each repeat were modified by a unique underbar. Sequences were then assembled from these repeats, using two assumptions. First, only identical or nearly identical repeats were aligned. Second, gaps (indicated by dashes) were added wherever needed.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

Condensed chromosomes of E. histolytica trophozoites number 28, suggesting that amebae are diploid. E. histolytica trophozoites form condensed chromosomes, which are rare in untreated amebae and frequent in amebae treated with 7 mg of colchicine per ml (Fig. 1). Each trophozoite averaged 22 ± 5, with a maximum of 28 and a minimum of 15. Assuming 14 chromosomes in haploid parasites (47) and assuming some inefficiency in counting closely opposed or small chromosomes, E. histolytica trophozoites appear to be diploid. Diploidy is consistent with the presence of two E. histolytica Ser-rich protein genes and two E. dispar Ser-rich protein genes in cDNA libraries of each parasite and the presence of two Ser-rich protein PCR products in numerous axenized strains of E. histolytica (10, 20). In contrast, the presence of as many as four bands within a linkage group on Southern blots of pulsed-field gels of E. histolytica chromosomes suggests the possibility that amebae are tetraploid for some chromosomes (47).

View larger version (78K): [in this window] [in a new window]   FIG. 1.   Fluorescence micrograph of a colchicine-treated E. histolytica trophozoite stained with propidium iodide. Condensed chromosomes number 28, which is twice the number of chromosomes (14) identified on pulsed-field gels.

E. histolytica is also diploid for five "single-copy" genes. To determine the ploidy of amebae by an independent method, we performed FISH with an arbitrary set of amebic genes, which appear to be present in a single copy as shown by Southern blotting or by repetitive probing of cDNA or genomic DNA libraries. These E. histolytica genes encoded a diverse set of proteins involved in cyst wall destruction (chitinase), fermentation (pyruvate:ferredoxin oxidoreductase), exchange of electrons (nicotinamide nucleotide transhydrogenase), ion transport (plasma membrane calcium-transporting ATPase), or host tissue destruction (cysteine proteinase 5) (12, 15, 18, 37, 48). With all five E. histolytica genes, FISH showed a mixture of diploid and tetraploid parasites (Fig. 2A through E). Diploid parasites were presumably in G₁ prior to DNA synthesis, while tetraploid parasites were presumably in G₂ prior to mitosis. Negative controls with vector sequences alone showed no staining. FISH was also performed with E. histolytica p-glycoprotein genes, which are present in at least six copies and encode proteins associated with emetine resistance (13). As expected, p-glycoprotein gene probes bound numerous times to amebic trophozoites (Fig. 2F). These results suggest that the basic ploidy of E. histolytica is diploid, although these results do not rule out the possibility that some amebic genes will be present in more than two copies (if genes or portions of chromosomes are duplicated) or once (if a copy of the gene is deleted).

View larger version (47K): [in this window] [in a new window]   FIG. 2.   Confocal micrographs of FISH of E. histolytica with single-copy genes encoding chitinase (A), pyruvate:ferredoxin oxidoreductase (B), nicotinamide nucleotide transhydrogenase (C), plasma membrane calcium-transporting ATPase (D), and cysteine proteinase 5 (E), which bind twice (yellow) to each nucleus. In contrast, FISH with p-glycoprotein genes (F), which are in multiple copies, shows numerous spots.

Evidence for bottlenecks or demographic sweeps in populations of E. histolytica and E. dispar. Previously, we found no differences among ITS among rRNA genes of 10 E. histolytica isolates (28). Here the intergenic sequences between superoxide dismutase and actin 3 genes of E. histolytica were sequenced, because these sequences are single copy and are longer (380 nucleotides) than ITS between rRNA genes (165 nucleotides total) (4, 28). The sod-actin intergenic sequences of three isolates of E. histolytica from Mexico City, San Diego, and Calcutta were each the same as that of the HM-1 strain (Fig. 3), even though these E. histolytica isolates differed at chitinase or Ser-rich protein loci (see below). The sod-actin intergenic sequences of five isolates of E. dispar, which came from two continents and differed from each other at chitinase or Ser-rich protein loci, were also identical to each other. The E. dispar sod-actin intergenic sequences differed in 83 nucleotides (22%) from those of E. histolytica. Conserved in the E. histolytica and E. dispar sequences was a TATA-like box upstream from the actin 3 coding region (5, 41). These results, which are consistent with the idea of a bottleneck or demographic sweep like that described previously for P. falciparum (34), indicate that human infections with each Entamoeba species derived from a single organism or from an identical group of organisms. Although these amebic bottlenecks were recent relative to the divergence of E. histolytica and E. dispar, how recent is not clear. With a great deal more sequence data available for P. falciparum, a bottleneck as recent as 7,000 years ago has been argued for (34).

View larger version (24K): [in this window] [in a new window]   FIG. 3.   Alignment of the intergenic sequences between superoxide dismutase and actin 3 genes of E. histolytica and E. dispar. Periods indicate identity of E. dispar with E. histolytica, and dashes indicate gaps. Unshaded boxes indicate superoxide dismutase and actin gene coding regions, and the dotted box indicates the TATA-like sequence upstream of the start codon of actin 3, while arrows indicate locations of PCR primers. The E. histolytica sequence was identical to that reported previously with GenBank accession no. X70852 (40).

chitinase alleles tentatively identify E. histolytica HM-1 strain amebae in Mexico City and NIH-200 strain amebae in Calcutta. Primers flanking the E. histolytica chitinase gene repeats produced a single PCR product from each clinical isolate (data not shown). This result suggests that the amebic chitinase genes are homozygous, as parasites are diploid at this locus (Fig. 2A). The chitinase PCR products were sequenced without cloning, and the sequences of the repeats were coded by methods used to compare P. falciparum circumsporozoite protein gene sequences (Fig. 4) (33). The E. histolytica chitinase gene repeats ranged from 84 to 252 nucleotides and so predicted chitinases with four heptapeptide repeats (28 amino acids) to 12 heptapeptide repeats (84 amino acids). The 168-nucleotide chitinase gene repeat of a San Diego isolate (Eh SD1) was identical to that of the HM-1 strain, while the 84-nucleotide chitinase gene repeat of a Calcutta isolate (Eh K1) was the same as those of NIH-200 amebae (12). These results, which suggest that the original axenized strains of E. histolytica may still be located in the New and Old Worlds, should be reassuring to bench scientists, who study these amebae. Because the amebae for the present studies came from individuals with amebic cysts in their stools, it is likely that E. histolytica chitinases with varying numbers of heptapeptide repeats are equally functional. This conclusion is consistent with the idea that heptapeptide repeats are spacers between lectin and catalytic domains of amebic chitinases (12).

View larger version (21K): [in this window] [in a new window]   FIG. 4.   Patterns of E. histolytica and E. dispar chitinase repeats. (A) Building blocks of chitinase repeats. Each 21-nucleotide sequence, which encodes a unique heptapeptide repeat, is given a number. The numbers are marked to indicate silent nucleotide changes, which are underlined. (B) Patterns of chitinase repeats. Shown are the chitinase PCR products from clinical isolates of E. histolytica (Eh) and E. dispar (Ed) in Mexico City (HI), San Diego (SD), and Calcutta (K). Each PCR product is coded using the numbers in panel A, while gaps are indicated by dashes.

E. dispar chitinase gene polymorphisms suggest transcontinental spread of parasites. Primers flanking the E. dispar chitinase gene repeats produced a single PCR product from each clinical isolate (data not shown). Eleven different E. dispar chitinase gene alleles were found in 25 clinical isolates of E. dispar from Mexico City, San Diego, and Calcutta (Fig. 4). The average number of E. dispar chitinase gene repeats (16 ± 2) was greater than those of E. histolytica (8 ± 4; P < 0.01). This result suggests that forces which cause diversifying selection at the chitinase locus are at least as active against E. dispar parasites. Five different chitinase gene repeats, which predicted 12 to 18 heptapeptides, were identified among PCR products of 10 E. dispar isolates from Mexico City. Interestingly, one of these Mexican E. dispar chitinase alleles (Ed HI1) was also identified in an E. dispar isolate from San Diego, the E. dispar cDNA library made by Egbert Tannich (SAW 760), and the strain of E. dispar axenized by Graham Clark (SAW 1734) (7, 44). These results suggest that this E. dispar strain may have traveled between the New World and the Old World, as both SAW 760 and SAW 1734 were isolated from Ethiopian Jews in Israel (David Mirelman, personal communication). A second E. dispar Mexican chitinase gene allele (Ed HI5) was identified six times in PCR products of 10 E. dispar isolates from Calcutta (Ed K1), suggesting the transcontinental movement of this E. dispar strain as well. These studies, however, cannot determine the direction of travel of the parasites.

Ser-rich protein gene polymorphisms among clinical isolates of E. histolytica suggest the persistence of the early axenized strain in Mexico City. Primers flanking the E. histolytica and E. dispar Ser-rich protein gene repeats frequently produced two PCR products on agarose gels from one clinical isolate (data not shown). This result suggests that the amebae are often heterozygous at the Ser-rich protein locus, as has been shown previously for axenic strains of E. histolytica and the SAW 760 strain used to make the E. dispar cDNA library (10, 20). Because only one Ser-rich protein PCR product was usually obtained from each clinical isolate (indicating that one Ser-rich protein PCR product was lost), we were unable to determine the linkage between chitinase gene and Ser-rich protein gene polymorphisms. The Ser-rich protein PCR product from one Mexican E. histolytica isolate (Eh HI1) matched one of the cloned Ser-rich protein genes of HM-1 parasites (Fig. 5). Two other isolates of E. histolytica from stools of children in Mexico City showed Ser-rich protein alleles, which encoded proteins with few (four to six) tetrapeptide and octapeptide repeats.

View larger version (20K): [in this window] [in a new window]   FIG. 5.   Patterns of E. histolytica and E. dispar Ser-rich protein repeats. (A) Building blocks of Ser-rich protein repeats. Each 24-nucleotide sequence, which encodes a unique octapeptide repeat, or 12-nucleotide sequence, which encodes a unique tetrapeptide repeat, is given a number. The numbers are marked to indicate silent nucleotide changes, which are underlined. (B) Patterns of Ser-rich protein repeats. Shown are Ser-rich protein PCR products from clinical isolates of E. histolytica (Eh) and E. dispar (Ed) in Mexico City (HI) and Calcutta (K). Each PCR product is coded using the numbers in panel A, while gaps are indicated by dashes.

Implications of these data for genetic and molecular epidemiological studies of amebae. The data here are too fragmentary for us to make any strong conclusions concerning populations of amebae which infect humans. This is particularly the case for E. histolytica clinical isolates. Still, a few implications of this data are worth noting. First, the E. histolytica genome appears to be diploid like those of most other eukaryotes. Whether amebae have a cryptic haploid stage and sex or reproduce clonally remains to be determined (45). Second, the HM-1 strain, which is the E. histolytica genome sequencing project strain, is likely identical in most aspects to other E. histolytica strains (sod-actin data). Although there is no guarantee that genes have not been lost from the HM-1 strain, it is reassuring that these parasites still cause lesions in animal models (23, 49). A possible explanation for the demographic sweep or bottleneck in human populations of E. histolytica is that amebae, like P. falciparum, predominantly infect humans and do not infect other mammalian hosts. Third, it appears that Ser-rich protein and chitinase alleles may be used to discriminate isolates of E. histolytica or E. dispar, at least in big cities, where these studies were performed (38). As parasites were homozygous at the chitinase locus, chitinase alleles may be easier to work with than Ser-rich protein alleles. Fourth, it appears that the HM-1 strain of E. histolytica, which was isolated more than 30 years ago from Mexico City, is still there and in San Diego (chitinase and Ser-rich protein data) (14). Previously, the pattern of the Ser-rich protein repeats was shown to remain constant during 30-plus years of culture (10). Fifth, chitinase gene alleles suggest that some E. dispar parasites have spread between the Old and New Worlds, although the direction of spread cannot be determined.