Intraspecies Genotype Variability of the Microsporidian Parasite Encephalitozoon hellem

Abstract

Seven isolates of Encephalitozoon hellem from human immunodeficiency virus-positive patients were genotyped through a series of markers: the internal transcribed spacer (ITS) of ribosomal DNA, the polar tube protein (PTP) gene, and two intergenic spacers (IGS-TH and IGS-HZ) whose polymorphism is newly reported. The genome markers were all analyzed at three levels: PCR amplification followed by polyacrylamide gel electrophoresis, single-strand conformation analysis (SSCA), and DNA sequencing. The polymorphisms detected involve insertions/deletions and point mutations. SSCA can distinguish any pair of sequences, even those differing by a single base pair. The different isolates studied fit into the previously described ITS genotype 1A, except one which seems to be a 2A derivative variant (2D). When PTP and the new markers IGS-TH and IGS-HZ were analyzed, most of the isolates displayed different genotypes, demonstrating that E. hellem has a strong intraspecies variability. A set of markers such as those used here may be very useful in genotyping of clinical samples and in the assessment of epidemiological relationships among E. hellem strains.

Microsporidia are ubiquitous, obligate intracellular parasites that infect all animal phyla (3) and that are gaining attention as emerging pathogens, mainly in AIDS patients but also in immunocompetent patients (32, 21). Among the human microsporidian parasites, the species Encephalitozoon hellem is the third most frequently identified as causing clinical disease in humans. Cases of ocular, respiratory, urogenital, and systemic infections have been reported (20). To date, humans are the main host of this microsporidian species, which has also been found in birds (2, 27). This finding contrasts with other common microsporidia, such as Enterocytozoon bieneusi and Encephalitozoon intestinalis, which have been detected infecting a wide range of domestic and wild animals (1, 2, 5, 8, 15).

In the last few years and parallel to the development of molecular tools, several studies have focused on the transmission of human microsporidiosis and its propagation through animal contact (5, 26). These studies have both a phylogenetic and prophylactic interest due to the characteristics of the main patients affected by microsporidiosis. In the case of Encephalitozoon cuniculi, the best known microsporidian species in molecular terms, three genotypes have been described to be related to the internal transcriber spacer (ITS) of the rRNA (9, 10, 11, 25, 28). Their distribution clearly indicated that E. cuniculi of animal origin may be a source of human infection. The zoonotic potential of other species is still under discussion.

Based on the sequence of the ITS of the ribosomal DNA (rDNA) of E. hellem, Mathis et al. (22) described three genotypes differing in small insertions or deletions and some point mutations. However, these authors had already determined that the genotypes had no absolute correspondence with antigen profiles, a finding which suggests that the genetic variability may actually be higher. Accordingly, Delarbre et al. (7) detected that the size of the chromosomes is also subject to genetic diversity in E. hellem.

The detection of new polymorphic genetic markers in E. hellem entails certain difficulties. First, the genomes of the microsporidia in species are extremely small, ranging from 19.5 Mbp to only 2.3 Mbp (18), with short spacers and scarce opportunity for repetitive DNA fractions. In addition, whereas some information about E. cuniculi, whose genome has been sequenced (17), may be useful for this different species, other information is not, since many sequences may be different; for instance, there are species-specific primer pairs for different Encephalitozoon species (4, 31).

From both theoretical and practical points of view, the method used in genotype analysis is also important. The nucleotide sequences are obviously the ultimate limit for genetic variability; however, even though in recent years sequencing services are available for many laboratories, routine analyses of many loci and isolates may not be affordable. PCR, followed by gel electrophoresis, distinguishes DNA fragments that differ in length (a minimum of about 20 bp for agarose gels and 1 bp for polyacrylamide gels). Point mutations can be resolved by restriction analysis only if they affect a specific target; thus, in E. cuniculi, two of them could be distinguished from the three genotypes described by Xiao et al. (34) on the basis of PCR, followed by restriction fragment length polymorphism analysis, of part of the polar tube protein (PTP) gene, in addition to a third genotype that had a 78-bp deletion. Alternatively, single-strand conformation analysis (SSCA), when used under accurate conditions, may produce different electrophoretic patterns for two DNA segments differing in just 1 bp (29). In fact, Fedorko et al. (12) used such a method to distinguish several types of microsporidia, including five isolates of E. hellem, from humans; unfortunately, the aim of that study was to achieve species-specific patterns, and no intraspecific variability was reported. Clear evidence of such variability was provided by Xiao et al. (33), who reported six small subunit (SSU)-rRNA genotypes, plus two ITS and four PTP variants, among 24 isolates of E. hellem. In the present study, seven E. hellem isolates of human origin were analyzed for a set of four polymorphic markers, two of them described here for the first time, following PCR, SSCA, and sequencing, to assess the extent of variability in this species.

MATERIALS AND METHODS

Strains.

Seven E. hellem isolates were examined in the present study: EHVS96 (Spain); CDC:V257 (United States); and MiPV4-94, PV5-95, PV6-95, PV7-95, and PV8-95 (Italy). All of these strains were obtained from human immunodeficiency virus (HIV)-positive patients. Each strain was cultured on Vero-E6 cells according to the method described by Visvesvara et al. (30).

DNA isolation.

After three washes with phosphate-buffered saline-EDTA buffer (10 mM phosphate-buffered saline [pH 7.3], 1 mM cold EDTA), DNA was extracted from 500-μl aliquots containing 10⁸ to 10⁹ spores per ml by using the Fast-DNA-Spin kit according to the manufacturer's instructions (Bio 101, Carlsbad, Calif.).

Loci analyzed and PCR.

Four loci were analyzed: (i) the ITS between the 16S and 5.8S rRNA genes, (ii) the IGS-TH located between the tRNA-tryptophanyl transferase gene and the mitochondrial type heat shock protein gene (hsp70), (iii) the IGS-HZ located between the latter and the zinc-finger protein gene, and (iv) the variable region of the PTP gene (PTP). The fragments analyzed for IGS-TH and IGS-HZ also included small segments (ranging from 22 to 54 bp) from the correspondent border genes. The primers used for ITS amplification were described by Hollister et al. (16); the remaining primer pairs are listed in Table 1 and were generated with the program OLIGO 6.23 (Molecular Biology Insights Co.) from the sequences of E. hellem GenBank accession numbers AB056660 and AF044915. Reactions were prepared in a final volume of 25 μl per tube, including 50 ng of template DNA, 0.2 mM concentrations of deoxynucleoside triphosphates, 0.2 μM concentrations of each primer, 1.5 mM MgCl₂, and 1 U of Taq polymerase (Roche Molecular Biochemicals) and its correspondent buffer. PCR was performed in a PTC-100 thermocycler (MJ Research) with the following program: 80°C for 5 min, 98°C for 30 s, annealing temperature (specific for each primer pair [see Table 1]) for 30 s, and 72°C for 1.5 min; this initial cycle was followed by 29 cycles of 30 s at 94°C, 30 s at the annealing temperature, and 1.5 min at 72°C; followed by a final step at 72°C for 9 min. The tubes were then maintained at 4°C until the next protocol. Each isolate was analyzed by a minimum of four PCR per locus. The quality of PCR amplifications were checked in routine 2% agarose gel electrophoresis.

TABLE 1.

Primer pairs used in this study

Primer	Sequence (5′-3′)	Annealing temp (°C)	Source or reference
A	ttgtacacaccgcccgtcg	55.0	Hollister et al. (16)
B	ccgataatgccaatcaatcc	55.0	Hollister et al. (16)
IGS-TH (upper)	agtttcgtctgtagacacca	51.0	This work
IGS-TH (lower)	cccagatctattccaattatg	51.0	This work
IGS-HZ (upper)	gaggggtgaagacttcgac	51.2	This work
IGS-HZ (lower)	tctctcgaagcatcttgactt	51.2	This work
PTP-h (upper)	cagttggttcctgggaatact	63.1	This work
PTP-h (lower)	tttgaggaacagagacgca	63.1	This work

Polyacrylamide gel electrophoresis (PAGE) and SSCA.

Polyacrylamide gels (8%; acrylamide-bisacrylamide [19:1]) were prepared in TBE buffer (45 mM Tris-borate, 0.1 mM EDTA [pH 8.3]) in vertical cuvettes (Hoeffer SE400; Amersham Pharmacia Biotech), by using 175 μl of 0.1% ammonium persulfate and 8.8 μl of N,N,N′,N′-tetramethylethylenediamine (TEMED) as activators. Bands were resolved by electrophoresis for 23 h at 130 V at room temperature. SSCA analysis was performed according to the method of Myers et al. (23), with some modifications: the PCR products were denatured by mixing with 95% deionized formamide and 5% gel loading buffer (GenSura Laboratories, San Diego, Calif.) and incubated for 2 min at 95°C, followed by rapid cooling in ice bath. TBE-8% polyacrylamide gels were prepared in a Hoefer SE600 apparatus, at both 99:1 and 49:1 acrylamide-bisacrylamide proportions, by using 8.8 μl of TEMED and 175 μl of 0.1% ammonium persulfate as activators. The electrophoretic analyses were performed for 15 h at 10°C and 300 V. The gels were revealed by using the following silver staining procedure: (i) 20-min fixation in 45 ml of distilled water plus 5 ml of methanol and 250 μl of acetic acid; (ii) staining with 0.1 g of AgNO₃ in 50 ml of H₂O; (iii) after two washes with 50 ml of water, the bands were revealed by a 10-min treatment with 0.75 N NaOH-0.26% formaldehyde; and (iv) fixation in 50% methanol 10% acetic acid for 20 min. The gels were then digitally scanned and finally stored in a 50% methanol-10% glycerol-water solution.

DNA sequencing.

PCR products were prepared according to the manufacturer's instructions by using the Bioclean Columns kit (21.016 Biotools). Sequencing was carried out in both directions by the sequencing service of the Centro de Investigaciones Biológicas, Madrid, Spain. For the new alleles detected, two or three different PCR products were also sequenced in both directions.

RESULTS

ITS.

As reported in a preliminary study (6), all of the isolates described here fit into genotype 1 except for strain PV7-95. The alignment of this sequence (GenBank accession no. AY171241) indicates that it does not fit with any of the six sequences described (33) and suggests that PV7-95 has a genotype probably derived from genotype 2A (Xiao et al. [33]) that is equivalent to genotype 2 of Mathis et al. (22). Since the differences involve both point mutations and small insertion or deletions, the variants were detected both after PAGE or sequencing. SSCA also allowed us to distinguish the variants 1 and 2D found in our material (Fig. 1). The ITS sequence obtained for EHVS96 was also identical to genotype 1; however, this strain has a point difference in the SSU-rRNA gene fragment included in the segment amplified, which yields a slightly different SSCA pattern (arrowed in Fig. 1).

FIG. 1.

PAGE of single-stranded DNA fragments of the ITS. A 208-bp fragment, including the 46-bp ITS, was amplified by PCR, and the products were denaturated. Migration was from top to bottom. Each band corresponds to a different conformation of a single strand: small differences in sequence (even a single nucleotide) lead to different conformations and hence to different migrations. All patterns are similar (genotype 1A), except for that of PV7/95 (genotype 2D, double-headed arrow). EHVS96 also shows a slightly different pattern, with an additional band (arrowed). However, the ITS sequence is still that of 1A, and the difference comes from a point mutation in the SSU-rDNA fragment included in the PCR-amplified segment.

IGS-TH.

While a single PCR product of 193 bp was obtained for each of the seven strains, SSCA revealed three different patterns: isolates EHVS96, MiPV4-94, PV5-95, PV6-95, and PV8-95 share the same pattern, but isolates CDC:V257 and PV7-95 each showed a different pattern (Fig. 2). The sequencing of the correspondent fragments showed that the three alleles are due to point mutations: CDC:V257 revealed a sequence identical to that of GenBank accession AB056660 (allele 1), whereas allele 2 (AY171238) showed a single difference and allele 3 (AY171239) showed two differences. The second polymorphic site shown in Fig. 2c is located outside the IGS within the hsp70 gene.

FIG. 2.

Analysis of IGS-TH marker. (a) PAGE of native PCR products. There are no differences in length, and a single 193-bp fragment was obtained from each isolate. MW, molecular weight marker (200 bp). (b) SSCA reveals three different patterns (1, 2, and 3) corresponding to the three sequences shown in panel c. (c) Allele sequences. The dots denote sequence identity with allele 1, and dashes depict nucleotide deletions. Note that the single alleles 1 and 2, as well as alleles 2 and 3, differ in one base pair, which is enough to produce clearly distinguishable SSCA patterns. Underlined sequences are not included in the spacer belonging to the hsp70 gene.

IGS-HZ.

A 230-bp PCR product was obtained for each isolate (Fig. 3a), so there are no length differences for this spacer. However, as shown in Fig. 3b, the SSCA clearly revealed three distinguishable patterns: (i) EHVS96 and CDC:V257; (ii) MI-PV4-94, PV5-95, PV6-95, and PV8-95; and (iii) PV7-95. As in the previous case, these patterns are the result of point differences in the correspondent nucleotide sequences, which are located at positions 21, 31, 37, and 55 of the intergenic spacer (IGS) (Fig. 3c). The GenBank accession numbers are AY171235 (allele 1), AY171237 (allele 2), and AY171236 (allele 3). The three variants share an adenine at position 97, which does not seem to exist in the sequence for AB056660, so this would correspond to a fourth allelic variant.

FIG. 3.

Analysis of IGS-HZ marker. (a) PAGE of the native PCR product of 230 bp, identical to all of the isolates. MW, molecular weight marker(200 bp). (b) SSCA. (c) Allele sequenes. Clear interstrain variability was observed with three patterns, namely, 1, 2, and 3, and the DNA sequences corresponding to these variants are shown. Dots denote sequence identity with allele 1. The same fragment size obtained for all of the isolates in panel a is resolved into different SSCA patterns due to one (allele 1 versus allele 2) or three (allele 3 versus allele 1 or 2) point differences. Allele 4 was not found among our isolates and corresponds to GenBank accession no. AB056660.

PTP.

The polymorphism in the gene of the PTP is both due to the different number of repeats of an oligonucleotide unit and to several point mutations. Four variants were detected after PAGE (Fig. 4a). After SSCA, the same previous variants were obtained, but the electrophoretic patterns were too complex to distinguish all of the possible point mutations (results not shown). The fragment analyzed here has 490 to 620 bp, which is much longer than the maximum 300-bp length preferred for an accurate SSCA (13, 14). The basic unit repeated in PTP alleles may have either 60 or 66 bp, both involving different variants due to point nucleotide mutations (33). Isolates MiPV4-94, PV5-95, PV6-95, and PV8 share allele 1C and CDC:V257 has allele 1A, whereas the sequence obtained for EHVS96 (AY024342) corresponds to allele 1B; these variants have eight, seven, and six repeats of the 60-bp unit, respectively (Fig. 4b). Finally, PV7-95 has seven repeats involving two units of 60 bp and five of 66 bp; this new allele could be named 2C (accession no. AY171240).

FIG. 4.

Analysis of PTP marker. (a) PAGE revealing four different length alleles, named according to the study by Xiao et al. (33). MW, molecular weight marker. The left lane shows a 20-bp ladder above and below the darker 500-bp band, and the right lane shows a 603-bp marker. (b) Sequence alignment of the repeated units. Dots denote sequence identity to the first repeated block of allele 2B, and dashes depict nucleotide deletions. The number of each repeat reflects the corresponding order within the allelic variant. The alleles of PTP are classified into two families depending on whether they have any 66-bp repeat (family 2) or not (family 1). Note that alleles 1A, 1B, and 1C involve similar but not identical types of a 60-bp block, which are repeated in different orders and numbers; the same picture emerges from 2B and 2C, which involve both the 60-bp block and an additional 66-bp block. The sequence differences that involve an amino acid change (from that above) are shaded.

The genotypes of the seven strains are summarized in Table 2.

TABLE 2.

Genotypes obtained for the isolates of E. hellem

Marker	Genotype of E. hellum isolate:
	EHVS96	CDC:V257	MiPV4-94	PV5-95	PV6-95	PV7-95	PV8-95
ITS	1A	1A	1A	1A	1A	2D	1A
IGS-TH	2	1	2	2	2	3	2
IGS-HZ	1	1	2	2	2	3	2
PTP	1B	1A	1C	1C	1C	2C	1C

DISCUSSION

Genotyping E. hellem strains.

Since many mutations may be lethal or severely affect the normal gene products, functional genes are expected to show lower levels of variability than noncoding regions. The spore wall protein gene of E. cuniculi and that of the PTP in both E. cuniculi and E. hellem are the two only known polymorphic genes in these species (24, 33, 34). Although it is still not clear which physiological mechanisms are involved, such polymorphisms may have an adaptive significance since these proteins are important candidates for early interaction with the host cell (24).

Among the noncoding regions, the ITS of rDNA is usually one of the first genome segments analyzed in a number of materials, including microsporidia, from a molecular-variability point of view. This is due to the fact that rRNA genes may be conserved from one species to another and so primers for PCR can easily be established, even in materials for which the molecular data thus far are limited. A similar strategy is followed for other new sequences: once their organization in the genome is available, the IGSs may be the subject of a systematic search for polymorphisms. At this point, E. cuniculi, for which a draft of the complete genome already exists (17), is far ahead of E. hellem. It is noteworthy that in the case of either rDNA (6) or IGS-TH (Fig. 2), some polymorphic sites are actually included within the expressed sequences and thus are not in the spacers.

For practical reasons, the nature of the polymorphism is important to consider. Allelic variants that differ in size due to the number of repeats of a given unit of small insertions or deletions may be distinguished by PCR, followed by electrophoresis in either agarose or polyacrylamide gels (depending on the absolute difference in the number of nucleotides). Thus, genotyping for PTP and ITS alleles is easy to perform by quite simple methods (33), although for our material the latter proved more difficult to distinguish since alleles 1 and 2D differ by only 2 bp in length. Nevertheless, in many instances some alleles differ in point nucleotide changes. Strategies such as restriction enzyme digestion of PCR products, used to resolve alleles I and II of PTP in E. cuniculi (34), require that the variability affect a restriction site, and thus its range of application is narrow.

Under appropriate conditions, SSCA may detect virtually all point changes. All of the DNA fragments studied here that have different nucleotide sequences-even when a single base pair is involved-can also be distinguished after SSCA. Thus, the uniform size observed for IGS-TH and IGS-HZ segregates into three different SSCA patterns, reflecting the corresponding different sequences (Fig. 2 and 3). One disadvantage is that the technique has to be empirically optimized for each new segment of DNA to be analyzed. Also, given a different pattern of bands in the electrophoresis, a different sequence is known to exist but, a priori, it does not indicate how deep that difference is. For instance, in Fig. 3b, the difference between the pattern of CDC:V257 and that of MiPV4-94 is due to a single base pair, but that of PV7-95 is the result of three and four different base pairs with respect to the other two (Fig. 3c).

Therefore, it is important to note that SSCA may be applied to a large number of samples: if the patterns obtained are identical to those previously known, one may assume that the corresponding sequences are also identical. Only a new pattern would make a sequencing procedure interesting in order to establish the molecular nature of the new allele to which it corresponds.

Variability in E. hellem.

The isolates analyzed here all proceeded from HIV-positive patients. On the basis of the former three-genotype ITS classification, all of these strains fit into genotype 1A, except for PV7-95, which seems to be a 2A-derivative variant (6), and can be named as 2D. Once the new markers are analyzed, EHVS-96, CDC:V257, and PV4-94, PV5-95, PV6-95, and PV8-95 may no longer be grouped together, clearly indicating that E. hellem is highly polymorphic. These results agree with Xiao et al. (33), who demonstrated additional variability when SSU-rRNA and the PTP gene analyses were introduced. It is noteworthy that these authors described four PTP alleles among 25 isolates of E. hellem whereas, from the seven strains analyzed here, a fifth variant (2C) was obtained. This cannot result from different methodological approaches, since the PTP alleles usually differ consistently in length and, in each case, sequencing was also performed. Different sources of E. hellem isolates could be important to finding new variants, and the genotyping of other isolates of different origins may be important to the evaluation of the actual variability in this species. Unfortunately, the diagnosis for microsporidia is still far from being routine practice in many hospitals.

The polymorphism for PTP gene is particularly relevant for the study of E. hellem strain evolution. First, depending on having just the 60pb-type repeat or the 60 plus 66bp-type, there are two allelic families: 1 (1A, 1C, or 1D) and 2 (2B or 2C), respectively (Fig. 4b). The 2B variant was not found among our isolates and corresponds to strain CDC:V216 obtained from an HIV-positive patient from Switzerland (33). Second, Fig. 4 also shows that different repeats of the same length may include different point mutations but that other mutations may be shared by two repeats of a given allele. Thus, although we have no data about the rates of substitution for this gene of E. hellem, it is clear that different duplications have been separated by periods long enough to fix point mutations. In addition, comparison of the different repeats (both intra- and interallele) also indicates that the variants are not hierarchically organized either. For instance, the alternative GGC/AAT triplet (shaded in Fig. 4b) is found in all possible combinations on other sites. Therefore, in the absence of recombination, the same mutations must have occurred independently on more than one occasion. This argues in favor of a physiological role of some substitutions at the PTP, as suggested by Keohane et al. (19) and Peuvel et al. (24). The classification of alleles into two families proposed by Xiao et al. (33) is useful for genotyping, since the presence or absence of 66-bp units is easy to determine. However, the existence of different subtypes of repeated units (both within the 66- and the 60-bp blocks) and the apparent tendency of entire blocks to the reorganize may make it difficult to infer phylogenetic distances from the simple alignment of sequences.

Bearing in mind that, as far as we know, this species has an exclusive clonal mode of reproduction, new variants must arise throughout the evolution of particular strains. The similarity of genotypes (Table 2) suggests evolutionary routes for the origin of the different strains, and it becomes clear that PV7-95 differs greatly from the rest: not only is there no allelic coincidence for any of the markers, but its IGS-HZ variant implies three or four point differences compared to the other two alleles, which differ in a single base pair (Fig. 3c). The epidemiological origin of PV7-95 remains unknown; it shows a considerable divergence from the remainder isolates, although all of them proceed from HIV-positive patients. Since, as mentioned above, this strain is the only representative of a different ITS genotype, we may expect that once strains of ITS (1B, 1C, 2A, and 2C) are analyzed with these other markers, new alleles will be found.

Our results point out that a series of markers such as those used here should be part of routine genotyping to assess the epidemiological and evolutionary relationships among E. hellem isolates.