WORKSHOPS ON OPPORTUNJSTIC PROTISTS 563

PCR-Mediated Recombination between Cryptosporidium spp. of Lizards and Snakes

LING ZHOU,"," CHUNFU YANG" and LIHUA XIAO" "Centers for Diseuse Control and Prevention, Atlunta, GA, and

"Atlanta Research and Education Foundation, DecatuK GA

ABSTRACT. The presence or absence of genetic recombination has often been used as one of the criteria for Cryptosporidium species designation and population structure delineation. During a recent study of cryptosporidiosis in reptiles that were housed in the same room, 4 lizards were found to have concurrent infections of C. serpentis ( a gastric parasite) and C. saurophilum (an intestinal parasite), and 6 snakes were concurrently infected with C. serpentis, C. sattrophitmm and a new Cryptosporidium as indicated by PCR-RFLP analysis of the SSU rRNA gene. DNA sequence analysis of cloned PCR products confirmed the diagnosis of mixed infections. Surprisingly, it appeared that 1 1 of the 22 clones (8 and 14 clones from a lizard and a snake, respectively) had chimeric sequences of two Cryptosporidium spp. BootScan analysis indicated the existence of recombinants among the cloned sequences and detection of the informative sites confirmed the BootScan results. Because the probability for genetic recombination between gastric and intestinal parasites is small, these hybrid sequences were likely results of PCR artifacts due to the presence of multiple templates. This was confirmed by PCR-sequencing analysis of single-copy templates using diluted DNA samples. Direct sequencing of 69 PCR products from 100- to 1.000-fold diluted DNAs from the same snake and lizard produced only sequences of C. serpentis, C. saurophilum and the unnamed Cryntosnoridium SP. Thus, care should be taken to eliminate PCR artifacts when determining the presence of genetic recombination or interpreting _ . . results of population genetic studies.

Recombination is an important evolutionary process and plays a major role in contributing to and maintaining genetic diversity in populations. Cryptosporidium, as with most protists, diverged relatively early in the eukaryotic lineage. This parasite maintains a haploid genome over most of its life cycle and undergo multiple cycles of asexual reproduction, which should greatly facilitate genetic stability. In recent years, the population structure of Cryptosporidium has attracted some attention. Recombination in Cryptcisporidium spp. hdS been speculated (2,5], and intra-specific recombination has been observed in C. parvurn [1,3]. Thus, even though inter-specific recombination has never been convincingly demonstrated, intra- specific recombination in Cryptosporidiurn is possible.

There is extensive genetic diversity in Cryptosporidium spp. of reptiles. Two common species have been found in both snakes and lizards: C. serpentis, and C. suurophilum (61. Six other species and genotypes of C,-y/~tosporidiunz have also been found, including C. muris, C. parvurn, the Cryptospnridium mouse genotype, a C. serpentis-like genotype, and two new Cryptosporidium spp. 161. Cryptosporidium serpentis differs from C. saurophilium in oocyst morphology and infection sites; the former has larger oocysts and infects the stomach, whereas the latter has smaller oocysts and infects the intestine. Recently, we identified a group of 10 snakes and lizards with concurrent infections of C. serpentis, C. saurophilium and a new intestine Cryptosporidium sp. [6]. Surprisingly, hybrid sequences were obtained from cloned PCR products. To confirm these hybrids are either true recombinants or experimental artifacts, we undertook the current investigation.

MATERIALS AND METHODS Isolates and DNA extraction. Two reptiles concurrently infected

with different Cryptosoridium spp. were chosen for this study. One snake saniple (#938) concurrently infected with C. serpentis, C. suuropkilum, and a new Cryptosporidium sp. and one lizard sample (#944) concurrently infected with C. serpentis and C. saurophilum. Detailed description of housing of these reptiles and DNA extraction were provided previously (61.

PCR analysis of the SSU rRNA gene. Identification of Crypto- sporidium spp. was based on PCR-RFLP analysis of the SSU rRNA gene [6,7]. Species were differentiated from each other by banding patterns in restriction digestions with enzymes Ssp I (New England BioLabs, Beverly, Mass) and Vsp I (Promega, Madison, WI). Each sample was examined three times by independent PCR-RFLP analyses.

Corresponding author: L. Xiao. Telephone: 770-488-4840; Fax: 770-488- 4454; Email: Ixiao@cdc.gov

Cloning of PCR products. Secondary PCR products of samples #938 (from a snake, with triple infections) and #944 (from a lizard, with double infections) were cloned into the pGEMB-T Easy Vector System (Promega, Madison, WI), and transformed into E. coli DH5a competent cells. Positive clones were sequenced in both directions on an AB13 100 sequencer (Applied Biosystems, Foster City, CA) using MI3 primers. Nucleotide sequences obtained were aligned using the program ClustalX (http://inn-prot.weizmann.ac.il/software/ClustalX. html). A neighbor-joining tree was constructed from the alignment using the program Treecon (http://www.psb.rug.ac.be/bioinformatics/ psb/Userman/treeconw .html) based on genetic distances calculated using the Kimura 2- parameter model. Nucleotide sequences of the clones were submitted to the GenBank database under accession number AY 12091 3, and AY382 I6 1 -AY382173.

Single-template PCR-sequencing analysis. DNA from isolates 938 and 944 were diluted 1 : lOO to 1:100,000, and used in single- template PCR analysis. The single template nature of DNA was confirmed by less than 50% PCR amplification of diluted DNA and by RFLP analysis of all PCR products. All PCR products were sequenced directly, and the sequences generated were analyzed.

Recombinant analyses. BootScan analysis was used to detect whether these cloned sequences were recombinants using the SimPlot software using the default settings (http://sray.med.som.jhmi. edu/RaySoft/). To further confirm the recombinant breakpoints, phylogenetic analysis of the sub-segments on either sides of the breakpoint was carried out and informative site analysis was also implemented.

RESULTS AND DISCUSSION Recombination phenomenon in PCR-clone analysis among

Cryptosporidium spp. in snakes and lizards. As reported before, of the 10 samples from the group of reptiles with concurrent infec- tions in this study, four (from four lizards) had C. serpentis and C. .saurophiZum, and six (from six snakes) had C. serpentis, C. saurophilurn and a new Cryptosporidium sp. by PCR-RFLP analysis of the SSU rRNA gene. We chose isolate #938 from a snake and isolate #944 from a lizard to further investigate whether or not recombination occurred. Of the 22 clones sequenced (8 and 14 clones from isolates #944 and #93X, respectively), 12 appeared to have chimeric sequences of different Crypfo.sp(Jridium spp. at the SSU rRNA locus by BootScan analysis as shown in Fig. 1. These sequences also formed separate branches within the three species groups in a phylogenetic analysis (Fig. 2A). More importantly, in- formative site analysis further confirmed that these were truly re- combinants, despite that phylogenetic analysis of the sub-segments on either sites of the breakpoint were not informative due to the

564

loo 90

80

70

I 60

50 D. % 40

30

2 0 .

A @ - - - - - , * - - - - .

\ I 4; \ / - I I

“:I 9448 . . . . . . . . . . . . . ’r

i: I

-.

J. EUKARYOT. MICROBIOL., 2003

B

10 7

f 60

50 n

t.....

- 938H

9448 . . . . . . . . . . . . .

100

20

10

0 50 100150200 250300 350400450 500550600650700

944-b2 744-b3 944-b8 944-b4 944-bS 944-bll 9 4 4 4 6 944-al3

;:% Y44-bI3 C. saurophilum 944.~8 944-a9 944-b7

Position

77 Y38-a29 944-a14 938-a21 944-a4

944-al5 938-a31 938-a3 738-bl 938-a2 938-1114 738-b2

738-bS 9.18-117

938-a13 C. serpentis

3

Fig. 1. Recombination breakpoint determination by BootScan analysis. The putative sites of recombination, or breakpoints were analyzed by BootScan analysis of the segment from clone together with three parental sequences: Crypto.spporidium sp. (938H), C. saurophilum (944C), and C. serpentis (944B). Results of the analysis revealed the occurrence of putative “recombinations” between Cryptosporidium sp. and C. serpentis in clone 9385 (A), and between C. serpentis and C. saurophilum in clone 944E (B). The arrows indicate points where the putative “recombinations” occurred.

smaller fragments of the sequences. In general, the breakpoints resided within 200 bp of the 5‘ or the 3’ end of a specific amplified region (Table I ) ; six recombinants were at the 3’ end, 3 were at the 5’ end, and 2 were in region shortly after the 5’ end.

A 938D

59J9386 j

100 I L.938p

Failed detection of recombination in concurrent infection by single-template PCR analysis. A dilution method was used to facilitate PCR analysis of a single copy of DNA template. At a given dilution point, if 2 to 5 out of 10 diluted templates (less than 50%)

0.02 substitutiondsite

H , Cryptosporidium sp.

parental sequence

9

B

9446 L ion r-

C. saurophilum Parental sequence

L 100

C. serpentis parental sequence

9 3 8 4 9

938-a23 938-a22 L 91R-nl 9 3 8 4 0 Cryptosporidium sp.

Fig. 2. SSU rRNA sequence analysis of isolates #938 and #944. A. Presence of heterogeneous sequences within C. saurophilum, C. .serpentis or Cryptosporidium sp. as revealed by a neighbor-joining analysis of the sequences from PCR. Parental sequences for C. saurophilum and C. .serpentis are labeled, and clones with recombinant sequences are underlined. B. Absence of recombination in SSU rRNA sequences from single-template PCR of #938 and #944. A total of 69 PCR products from 100- to 1,000-fold diluted DNA templates were sequenced directly without cloning, and phylogenetic relationship were inferred by neighbor-joining analysis. No recombination was seen by the phylogenetic analysis.

WORKSHOPS ON OPPORTUNISTIC PROTISTS 565

Table 1. Patterns of mosaic sequences (“recombination”) in PCR clones of isolates 938 and 944.

Clone 5’ sequence nem 5’ sequence 3’ sequence Breakpoint

938J Cryptosporidium sp. C. serpentis Cryptosporidium sp. 5’ end and near 5’ end 938E C. serpentis C. serpentis Cryptosporidium sp. 3’ end 938K Cryptosporidium sp. Cryptosporidium sp. C. serpentis 3‘ end 938N Cryptosporidium sp. Cryptosporidium sp. C. serpentis 3’ end 938Q Cryptosporidium sp. Cryptosporidium sp. C. serpentis 3’ end 938C Clyptosporidium sp. Cryptosporidium sp. C. serpentis 3’ end 938P Cryptosporidium sp. Cryptosporidium sp. C. saurophilum 3’ end 944A C. saurophilum C. serpentis 944E C. serpentis 944D C. serpentis 944G C. saurophilum

C. saurophilum C. saurophilum C. serpentis

yielded PCR products, that particular dilution point was considered an optimal dilution factor, and the PCR products were considered to be more likely a homogenous population derived from a single copy of DNA template. In this study, 100- to 1,000-fold-diluted DNA was chosen as optimal dilution factors for samples #938 and #944. R E P analysis of PCR products from diluted templates confirmed the presence of only a single Cryptosporidium in all PCR products. Furthermore, 69 single-template PCR products of the SSU rRNA gene were sequenced. Sequence alignment and phylogenetic analysis revealed the absence of chimeric sequences (Fig. 2B). None of the sequences from the directly sequenced PCR products derived from single-template DNA amplifications showed any evidence of re- combination between different Cryptosporidium spp.

Possible source of false recombinant sequences. PCR-mediated recombination has been reported before in other parasites [4]. In this study, recombinant sequences were only obtained from cloned PCR products of mixed DNA templates of Cryptosporidium spp. Thus, it is very likely that the recombinants obtained were due to PCR artifacts. Because DNA used in this study had 2-3 different Cryptosporidium spp., PCR aniplified concurrently the SSU rRNA gene of multiple species of Cryptosporidiurn. The high density of DNA templates in the PCR reaction increased the possibility for the Taq DNA polymerase to switch templates during the elongation process, producing sequences of hybrid genotypes.

Results of this study indicate that a care should be taken when interpreting data of recombination or population studies. Results of this study indicate that recombinants or new alleles can be generated by PCR when DNA used for the PCR templates has two or more heterogeneous population of different species or strains. In a recent multilocus study of Cryptosporidiurn population structure, 22% of C. parvum isolates had mixed subtypes and 12% of human samples had C. parviirn and C. korninis [3]. These samples, nevertheless, were included in the inference of Cryptosporidium population structure. It would be difficult to exclude the possibility that different Cryptospo- ridium species or subtypes were amplified at different loci and no PCR

C. serpentis C. saurophilum C. serpentis C. saurophilum

5’ end 5‘ end near 5’ end near 5’ end

artifacts were generated. Thus, samples with mixed Cryptosporidiurn spp. are best to be excluded from the analysis in popuhtion studies.

ACKNOWLEDGMENTS We thank Dr. Thaddeus Graczyk of the Johns Hopkins University

for providing samples. This study was supported in part by funds from the AWWA Research Foundation.

LITERATURE CITED 1. Feng, X., Rich, S.M., Tzipori, S. & Widmer, G. 2002. Experimental

evidence for genetic recombination in the opportunistic pathogen Crypto- sporidium parvum. Mol. Biochem. Parasitol., 1195-62.

2. Leav, B.A., Mackay, M.R., Anyanwu, A,, O’Connor, R.M., Cevallos, A.M., Kindra, G., Rollins, N.C., Bennish, M.L., Nelson, R.G. & Ward, H.D. 2002. Analysis of sequence diversity at the highly polymorphic Cpgp40/15 locus among Cryptosporidiurn isolates from human immunodeficiency virus- infected children in South Africa. fzject. fmmun., 70:388 1-3890.

3. Mallon, M., MacLeod, A,, Wastling, J., Smith, H., Reilly, B. & Tait, A. 2003. Population structures and the role of genetic exchange in the zoonotic pathogen Cryptosporidiurn parvum. J. Mol. Evol., 56:407417.

4. Tanabe, K., Sakihama, N., Farnert, A., Rooth, I., Bjorkman, A,, Walliker, D. & Ranford-Cartwright, L. 2002. In vitro recombination during PCR of Plasmodium falciparum DNA: a potential pitfall in molecular population genetic analysis. Mol. Biochem. Parasitol., 122:211-216.

5. Widmer, G., Tchack, L., Chappell, C.L. & Tzipori, S. 1998. Sequence polymorphism in the beta-tubulin gene reveals heterogeneous and variable population structures in Cryptosporidium parvum. Appl. Environ. Micro- biol., 6444774481.

6. Xiao, L., Ryan, U.M., Graczyk, T.K., Limor, J . , Li, L., Kombert, M., Junge, R., Sulaiman, I.M., Zhon, L., Arrowood, M.J., Koudela, B., Modry, D. & Lal, A. A. 2003. Genetic diversity of Cryptosporidium spp. in captive reptiles. Appl. Environ. Microbiol. (In press).

7. Xiao, L., Sulaiman, I.M., Ryan, U.M., Zhou, L., Atwill, E.R., Tischler, M.L., Zhang, X., Fayer, R. & Lal., A.A. 2002. Host adaptation and host- parasite co-evolution in Cryptosporidium: implications for taxonomy and public health. fnt. J. Parasitol., 32: 1773-1785.