DNA Sequence Similarity between California Isolates of Cryptosporidium parvum

Abstract

We evaluated whether nucleic acid amplification with primers specific for Cryptosporidium parvum followed by automated DNA sequence analysis of the PCR amplicons could differentiate between California isolates of C. parvum obtained from livestock, humans, and feral pigs. Almost complete sequence identity existed among the livestock isolates and between the livestock and human isolates. DNA sequences from feral pig isolates differed from those from livestock and humans by 1.0 to 1.2%. The reference sequence obtained by Laxer et al. (M. A. Laxer, B. K. Timblin, and R. J. Patel, Am. J. Trop. Med. Hyg. 45:688–694, 1991.) differed from California isolates of C. parvum by 1.8 to 3.2%. These data suggest that DNA sequence analysis of the amplicon of Laxer et al. does not allow for differentiation between various strains of C. parvum or that our collection of isolates obtained from various hosts from across California was limited to one strain of C. parvum.

Cryptosporidium parvum, a protozoan parasite, is an important etiologic agent of enterocolitis in mammals. C. parvum appears to be infectious for 79 mammals, including domestic and wildlife species (5). Transmission of oocysts can occur by the direct fecal-oral route or through the consumption of food or water contaminated with oocysts (5). With respect to waterborne C. parvum, identifying the mammalian source following a waterborne outbreak has proven to be very difficult (7, 10, 15, 16). The standard approach has been to collect fecal samples from mammalian populations of concern and to determine the proportion with detectable levels of C. parvum oocysts. Those mammalian populations found to be shedding C. parvum are presumed to have been the source of the waterborne outbreak (7, 15, 16). The validity of this approach may be strengthened by utilizing molecular fingerprinting to assess whether C. parvum isolated from the human case(s) is similar to or different from the isolate(s) obtained from the suspect source(s).

Desirable attributes for a DNA fingerprinting method would include the ability to detect minute concentrations of oocysts, high specificity, and the ability to function with fresh or archived fecal and water samples. Isoenzyme typing, immunoblotting, restriction fragment length polymorphism on whole genomes, and field inversion gel electrophoresis typically require large numbers of oocysts and have not been able to differentiate between animal isolates or could only distinguish between animal and human isolates (2, 3, 11–13). Arbitrarily primed PCR requires that contaminating DNA not be present in the sample. An alternative method for differentiating C. parvum from different mammalian sources would be nucleic acid amplification with primers specific for C. parvum followed by automated DNA sequence analysis of the amplicons. The present study was undertaken to determine if such an approach could differentiate C. parvum isolates obtained from different mammalian sources from throughout California.

We utilized a previously developed set of PCR primers (8). Although these primers were not known to target a specific gene (14), they have the test attributes we were seeking, e.g., sensitive for C. parvum and able to function with formalin-fixed samples, and they have been shown not to amplify other microorganisms commonly found in water and feces (6, 8, 9, 14). Additionally, we tested this primer pair against bovine Neospora, Salmonella typhimurium, Escherichia coli, Nematodirus battus, Eimeria zurnii, Eimeria ellipsoidalis, Cryptosporidium muris, Toxoplasma gondii, and bovine Giardia duodenalis and found no cross-reaction.

Fecal samples were collected from dairy calves, beef calves, goats, horses, and feral pigs from throughout California. Secondary treated wastewater samples (presumably of human origin) were collected from one plant. Samples were screened for C. parvum by a direct immunofluorescence assay (MERIFLUOR Cryptosporidium/Giardia; Meridian Diagnostic, Inc., Cincinnati, Ohio). For each isolate, C. parvum oocysts were purified by using a sucrose gradient (17) or low-speed centrifugation (4), followed by bleach sterilization. DNA was extracted by incubating oocysts for 48 h in TES (10 mM Tris HCl [pH 7.5], 1 mM EDTA, 10 mM NaCl) buffer containing 0.8% Sarkosyl (Sigma, St. Louis, Mo.) and 200 μg of proteinase K (Sigma) per ml. DNA was extracted with phenol-chloroform-isoamyl alcohol (24:24:1), precipitated in 100% cold ethanol, dried, and stored at 4°C in Tris-EDTA buffer (pH 7.5). PCR amplifications were performed in a Hot-Start tube with the GeneAmp Core Reagents (Perkin-Elmer, Foster City, Calif.), 64 to 125 ng of DNA, and sterile water. DNA concentrations were determined by the Nucleic Acid Test Instant Quantitation Kit (NBI, Plymouth, Minn.). Primer annealing temperature was 52°C. All PCR products were purified with the Qiaquick PCR purification kit (Qiagen, Inc., Valencia, Calif.). The products were sequenced with the Taq FS Dye terminator mix (Perkin-Elmer), 20 ng of template, and 10 pmol of primer. The sequence was generated on an ABI 377 automated DNA sequencer (Applied Biosystems, Alameda, Calif.). Contiguous sequence was generated from the forward and reverse strands with the Sequencher program (Genecodes, Ann Arbor, Mich.). Each isolate was sequenced three times for the forward and reverse directions.

The DNA sequences obtained for the 12 C. parvum isolates had various amounts of polymorphism when compared against each other (Fig. 1). Boldface and single-underlined nucleotides indicate substitutions, black dots indicate deletions, and double-underlined nucleotides on the Laxer sequence correspond to diagnostic probe 127, a StyI restriction site, and diagnostic probe 325 (4, 8). The degree of sequence polymorphism (substitutions, deletions, insertions) ranged from 0.0 to 0.5% (Table 1) among the livestock isolates and between the livestock isolates and the secondary treated wastewater isolate, presumably of human origin. The majority of livestock hosts were located on watersheds which were hydrologically independent from each other and had no known common source of feed, personnel, or other such factors which could serve as a means of cross-transmission of C. parvum. The DNA sequences obtained from feral pig isolates differed from those of livestock and human C. parvum by 1.0 to 1.2%, at least twice the proportion compared to the amount of polymorphism among livestock isolates (0.0 to 0.5%) and at least four times the proportion of polymorphism between livestock and human C. parvum (0.0 to 0.2%). These pigs were trapped from a remote site that had no known cohabitating livestock and minimal human contact (1). The DNA sequence published by Laxer et al. (8) had the greatest amount of polymorphism (2.0 to 3.2%) compared against all of the California isolates. The Laxer sequence has been utilized by various researchers to develop PCR-based diagnostic tests for C. parvum (4, 6, 9). Most of the polymorphism between the sequence of Laxer et al. and California isolates did not occur at critical internal diagnostic sites (probe or restriction enzyme), but all three feral pigs had a substitution at probe 127 (A to T) and all 12 California isolates had a deletion at probe 325 compared to the Laxer sequence.

FIG. 1.

Sequence alignment of PCR amplicons from 12 isolates of C. parvum from throughout California and the original sequence obtained by Laxer et al. (8). Feral pigs 2 and 3 have a shorter sequence (201 bp) as a consequence of having to use nested PCR (4) due to low oocyst concentrations in the two fecal samples. Boldface and single-underlined nucleotides indicate substitutions, black dots indicate deletions, and double-underlined nucleotides on the sequence of Laxer et al. correspond to diagnostic probe 127, a StyI restriction site, and diagnostic probe 325.

TABLE 1.

Levels of sequence similarity based on alignment of PCR amplicons from 12 isolates of C. parvum from throughout California and the original sequence obtained by Laxer et al. (8)

Isolate	% Sequence similaritya
	Goat 1	Goat 2	Foal 1	Foal 2	Beef 1	Beef 2	Dairy 1	Dairy 2	STWW	Feral pig 1	Feral pig 2	Feral pig 3	Laxer reference
Goat 1	100	99.8	99.8	100	100	99.8	99.8	99.8	100	98.8	99.0	99.0	98.2
Goat 2		100	99.5	99.8	99.7	99.5	99.5	99.5	99.8	98.8	99.0	99.0	97.8
Foal 1			100	99.8	99.8	100	100	99.5	99.8	99.0	99.0	99.0	97.8
Foal 2				100	100	99.8	99.8	99.8	100	98.8	99.0	99.0	98.2
Beef 1					100	99.8	99.8	99.8	100	98.8	99.0	99.0	98.0
Beef 2						100	100	99.5	99.8	99.0	99.0	99.0	97.9
Dairy 1							100	99.5	99.8	99.0	99.0	99.0	97.8
Dairy 2								100	99.8	98.8	99.0	99.0	97.8
STWW									100	98.8	99.0	99.0	98.0
Feral Pig 1										100	98.5	98.5	96.8
Feral Pig 2b											100	100	98.0
Feral Pig 3b												100	98.0
Laxer													100

^aThe sequences compared are the PCR amplicons from primers designed by Laxer et al. (8) or Balatbat et al. (4). Percent similarities were calculated such that nucleotide differences included substitutions, insertions, and deletions. 

^bFeral pigs 2 and 3 have a shorter sequence (201 bp) as a consequence of our having to use nested PCR (4) due to low oocyst concentrations in the two fecal samples.  

Although the primers of Laxer et al. (8) have many of the test attributes that are desirable for amplifying a specific section of genomic DNA of C. parvum, there does not appear to be sufficient polymorphism within the amplicon to allow for reliable discrimination between C. parvum isolates obtained within California. Alternatively, the marked similarity between the DNA sequences obtained from each of the 12 isolates may indicate that infection with this protozoal parasite was limited to one strain of C. parvum.