ABSTRACT
Cryptosporidium parvum is a protozoan parasite that causes the disease cryptosporidiosis in a variety of mammals, including neonatal calves and humans. Millions of oocysts are shed during acute cryptosporidiosis, and zoonotic transmission is inferred, though not proven, to be a general phenomenon. Very little is known about the degree of strain variation exhibited by bovine and human isolates, though such knowledge would enable the amount of bovine-to-human transmission to be more precisely analyzed. This research was initiated to determine whether variations exist among bovine strains isolated from a localized geographic area, the watershed of the Red River of the North. Sixteen strains were isolated and compared to each other and to two human and two calf strains from Australia by randomly amplified polymorphic DNA PCR. A statistical analysis of the data indicated that the isolates belonged to four different groups of strains.
Cryptosporidium parvum(phylum Apicomplexa) is the causative agent of cryptosporidiosis, a disease that results in a severe, yet self-limiting, diarrhea in immunocompetent humans but a chronic, severely debilitating and sometimes fatal infection in immunocompromised humans (15). Cryptosporidiosis is also a major cause of death in neonatal calves and is responsible for annual economic losses amounting to millions of dollars (9). Although there have been some encouraging pilot studies, C. parvum is not currently treatable with antimicrobial drugs, nor is an effective vaccine available (6, 7,15).
Millions of oocysts, the infectious form of the parasite, are shed from infected animals into the environment during acute cryptosporidiosis and often continue to be shed for days, weeks, or even months after the acute disease has resolved (15, 20). Disease outbreaks often involve water (1), and zoonotic transmission from cattle is inferred. However, except in job-related infections of farmworkers and veterinarians (17), the extent to which cattle act as a reservoir of human disease is unknown; in fact, initial strain typing studies suggest that there are bovine- and human-specific strains (4, 5, 14, 16). However, Morgan et al. (14) determined that 2 of the 14 human isolates they examined were more closely related to the 8 isolates from cattle than to the other human isolates. While it is clear that C. parvum is multiclonal, little is known about strain variability, differences in virulence, and the degree of cross-infectivity between bovine- and human-specific strains.
In this study, we examined the clonality of bovine strains using randomly amplified polymorphic DNA (RAPD) PCR and demonstrated that even oocysts isolated from the relatively localized geographic area of Minnesota and North Dakota are multiclonal. Furthermore, we confirmed that RAPD PCR is a very sensitive monitor of the genetic diversity ofCryptosporidium and an especially useful tool for typing clinical isolates when only limited amounts of material are available. For all of these reasons, we suggest that RAPD PCR is a feasible method of testing the hypothesis that bovine strains of C. parvumare responsible for a significant proportion of human cases of cryptosporidiosis.
Oocysts were purified from fecal samples (provided by Neil Dyer, State Veterinary Diagnostic Laboratory, North Dakota State University, Fargo) by sequential sucrose and CsCl gradients and subsequently washed as described previously (2, 10). DNA was released from purified oocysts during six freeze-thaw cycles. One cycle consisted of 2 min in a dry ice-ethanol bath followed by 2 min in a 98°C dry-heating block. DNA Dipstick (Stratagene, La Jolla, Calif.) was used to quantitate the DNA according to the manufacturer’s instructions, and the DNA was stored at −20°C until used as a PCR template. After DNA Dipstick quantitation, PCR was performed on similar amounts of DNA from all samples by using the primers described by Laxer et al. (11). In addition to detection of isolates positive forCryptosporidium (see below), PCR product intensities were compared to each other. Whereas the DNA Dipstick measured the DNA concentration, PCR amplification with the primers described by Laxer et al. measured the amplifiability of each DNA sample. Since the oocyst DNA was liberated only from the oocyst and was not purified from oocyst debris, we felt that the relative amplifiability was an important empirical measurement, especially for the subsequent RAPD reactions, but in fact, DNA Dipstick quantitation (0.2 to 5.0 ng/PCR) and amplifiability correlated well for most samples. PCRs were carried out in a buffer containing 50 mM KCl, 10 mM Tris-HCl (pH 8.4), 2 mM MgCl2, and a 0.2 mM concentration of each nucleotide in a final volume of 100 μl. Each reaction mixture also contained 2.5 U ofTaq polymerase, the primer(s), and the template. The primer sequences and concentrations are shown in Table1. Sequences within primer 1 and primer 2 were amplified for 35 cycles at 94°C for 1 min, 56°C for 2 min, and 72°C for 3 min followed by a final cycle at 72°C for 9 min. Products of PCR with RAPD primers R2817 and (GAA)5 were generated during 35 cycles at 94°C for 1 min, 37°C for 1 min, and 72°C for 2 min followed by a final cycle at 72°C for 9 min. RAPD primer 1344 was used under conditions identical to those for the other RAPD primers except that the annealing temperature was 42°C. All reactions were carried out in a model 110S Tempcycler II thermocycler (Coy). Agarose gel electrophoresis, photography, Southern transfers, and hybridizations were performed under standard conditions (19).
Sequences and concentrations of primers used in this study
Preliminary characterizations.In order to determine if the isolated oocysts were indeed C. parvum, PCR was performed with the primers described by Laxer et al. (11). These primers specifically amplify a 452-bp sequence from C. parvum, but not from other Cryptosporidium species, other parasites, or bacteria (18). All of the amplified samples produced this band (data not shown), indicating that they all contained C. parvum DNA. Primers specific forEscherichia coli and Prevotella(Bacteroides) ruminicola were used to amplifyCryptosporidium DNA since these organisms are common ruminal and intestinal inhabitants and would be likely contaminants if bacteria were present in the oocyst samples in our study (23). When the E. coli primers were used to amplify C. parvum DNA samples, no products were detected. However, when thePrevotella (ribosomal DNA) primers were used, a 389-bp fragment was amplified in four samples (strains 2674, 2894, 2896, and 3296), indicating that they were contaminated with P. ruminicola. In order to determine if the amount of contaminatingPrevotella DNA was enough to generatePrevotella-specific amplification products, a pure culture of Prevotella (ATCC 19188) was amplified by each RAPD primer used to radioactively probe the Cryptosporidium RAPD products from the same primers. Results of even the P. ruminicola-contaminated reaction mixtures showed that there were no Prevotella-specific RAPD products (21).
RAPD PCR.RAPD PCR was used to determine the relatedness of the different Cryptosporidium isolates. Initially, we tried the four primers that had been successfully used by Morgan et al. (14): R2817, R2936, GACA, and (GAA)5. Only R2817 and (GAA)5 resulted in reproducible DNA patterns in our hands. About two dozen additional primers were tested and found to be unsuitable either because they produced no PCR products or because they produced products even in the absence of a template (21). Two suitable primers (1344 and 1640) were eventually found. Ultimately, primer 1344 was used in conjunction with primers (GAA)5 and R2817 for our study.
The RAPD PCR analysis included 20 different isolates: 16 from the upper Midwest in the United States and 4 from Australia. Purified oocysts of human strains 49-H and 53-H and calf strains 248-C and 8700-C were provided by Una Morgan, School of Veterinarian Studies, Murdoch University, Murdoch, Australia. RAPD PCRs were carried out at up to three template concentrations 25-fold different from one another for each DNA sample and showed little product variation.
A total of 81 bands were scored: 30 from primer (GAA)5, 34 from primer R2817, and 17 from primer 1344. Similarity coefficients were determined by using the correlation-principal component analysis option in Microsoft Excel/xlSTAT (Table2). Spearman’s correlation coefficients were determined, and dendrograms were made by utilizing the hierarchical ascending clustering function (cluster analysis 2, or Ward’s clustering technique). When the data for the three primers were combined, four different groups were seen (Fig.1): (i) two human strains from Australia, (ii) most (13 of 16) of the strains isolated during this study (strains from North Dakota or Minnesota [ND/MN strains]), (iii) two Australian calf strains and one ND/MN strain (385), and (iv) two ND/MN strains (2896 and 3296) (Table 3).
Similarity coefficients of the combined data for primers (GAA)5, R2817, and 1344
Dendrogram grouping of C. parvum strains.
Dendrogram grouping of strains used in this study
This study is the first to compare C. parvum isolates from within a specific geographic region. With limited starting material, and neither the means nor the desire to cycle oocysts through animals, we were able to perform our analysis using the small amount of DNA required to perform RAPD PCR. Our observation that most of the isolates we studied belong to one major group was expected. Finding three strains (385, 2896, and 3296) that are quite different from each other and from the major group is somewhat surprising, but these results are similar to the results of Morgan et al. (14), who grouped one human isolate by itself and another with the bovine isolates they studied. However, those strains were isolated from samples from across the entire continent of Australia, while ours were derived from a more localized geographic area. Three possible explanations for our results are that (i) multiple, different strains have been brought into this geographic region, (ii) strains already present have continued to evolve, and (iii) DNA contamination produced apparent strain differences. While we cannot disprove the first possibility, examples of the second are well documented. Microevolution (also referred to as substrain shuffling) is commonly seen in recurrent Candida albicans vaginal infections (12) and is also acknowledged to occur during persistent infection with methicillin-resistant Staphylococcus aureus (13). A few of the strains we examined were contaminated with enoughPrevotella DNA to generate a ribosomal DNA PCR product; even though the amount of contaminating DNA was insufficient to generate any RAPD product, we cannot completely rule out the third possibility—that the strain differences we observed resulted from bacterial contamination. Thus, the heterogeneity of theseCryptosporidium strains needs to be corroborated by an additional taxonomic technique, such as ribotyping.
It is currently acknowledged that there are human and bovine strain types of C. parvum (5, 14). Whether or not these strain types are cross-infective in vivo, or how strain variation relates to disease transmission, is unknown. Having a reliable and comprehensive strain typing system would help address these issues and might ultimately establish whether shedding of oocysts by cattle and oocyst migration into surface waters during seasonal runoffs constitute a significant source of human disease outbreaks, such as those chronicled in Milwaukee, Wis., and Las Vegas, Nev.
ACKNOWLEDGMENTS
This work was supported by a grant from the Grand Forks Water Treatment Plant, Grand Forks, N. Dak.
FOOTNOTES
- Received 15 December 1997.
- Accepted 19 March 1998.
- Copyright © 1998 American Society for Microbiology
