INTRODUCTION
Cryptosporidium spp. is recognized worldwide as one of the most common enteropathogens causing diarrhoeal disease in a broad range of vertebrate hosts. The genus is composed of multiple genetically distinct forms, including host-adapted species/genotypes and others infecting a wide range of mammals, with C. parvum being considered an important zoonotic agent (Xiao and Ryan, Reference Xiao, Ryan, Fayer and Xiao2008). This species is a major pathogen of ruminant livestock, causing serious and economically significant outbreaks of neonatal diarrhoea during calving and lambing (Santín and Trout, Reference Santín, Trout, Fayer and Xiao2008). C. parvum is also one of the 2 major causes of human cryptosporidiosis in developed countries, the other being C. hominis (Xiao and Feng, Reference Xiao and Feng2008).
Molecular surveys have demonstrated age-related changes in the distribution of Cryptosporidium species/genotypes in cattle, concluding that calves less than 2 months old are the major reservoir for C. parvum infections in humans (Santín and Trout, Reference Santín, Trout, Fayer and Xiao2008). However, the role of other ruminant livestock in the transmission of human cryptosporidiosis is not clear. Some studies have reported that C. parvum are not commonly detected in sheep faeces, which more often contain oocysts of the Cryptosporidium cervine genotype and other genotypes (Ryan et al. Reference Ryan, Bath, Robertson, Read, Elliot, McInnes, Traub and Besier2005; Santín et al. Reference Santín, Trout and Fayer2007; Elwin and Chalmers, Reference Elwin and Chalmers2008; Geurden et al. Reference Geurden, Thomas, Casaert, Vercruysse and Claerebout2008), but other molecular analyses have indicated that C. parvum naturally infects lambs and is the major Cryptosporidium species responsible for diarrhoeic outbreaks on small-ruminant farms in some areas (Mueller-Doblies et al. Reference Mueller-Doblies, Giles, Elwin, Smith, Clifton-Hadley and Chalmers2008; Quílez et al. Reference Quílez, Torres, Chalmers, Hadfield, Del Cacho and Sánchez-Acedo2008 a).
Subgenotype analyses at the 60 kDa glycoprotein (GP60) locus have confirmed that human-specific, animal-specific and zoonotic subtypes occur within C. parvum (Cacciò et al. Reference Cacciò, Thompson, McLauchlin and Smith2005). At least 10 GP60 subtype families (IIa to IIj) and several subtypes within each family have been identified among C. parvum isolates from humans and/or animals. Nevertheless, only IId and particularly the most common family IIa are considered zoonotic, with the subtype IIaA15G2R1 being reported worldwide as one of the predominant subtype in calves and the major subtype responsible for zoonotic cryptosporidiosis (Misic and Abe, Reference Misic and Abe2007; Wielinga et al. Reference Wielinga, de Vries, van der Goot, Mank, Mars, Kortbeek and van der Giessen2008; Xiao and Ryan, Reference Xiao, Ryan, Fayer and Xiao2008). The C. parvum subtypes infecting sheep are not well documented, but some surveys indicate that diarrhoeic lambs are also usually infected by zoonotic subtypes families IIa and especially IId, revealing the potential role of lambs as a source for human cryptosporidiosis (Chalmers et al. Reference Chalmers, Ferguson, Cacciò, Gasser, Abs El-Osta, Heijnen, Xiao, Elwin, Hadfield, Sinclair and Stevens2005; Quílez et al. Reference Quílez, Torres, Chalmers, Hadfield, Del Cacho and Sánchez-Acedo2008 a).
In Spain, Cryptosporidium infection is known to be one of the most prevalent enteropathogens in both pre-weaned calves and lambs, although molecular studies at the subtype level are limited. A recent survey indicates that diarrhoeic dairy calves are mostly infected by C. parvum isolates in northern provinces of the country, with the subtype IIaA15G2R1 being the most common and most widely distributed (Quílez et al. Reference Quílez, Torres, Chalmers, Robinson, Del Cacho and Sánchez-Acedo2008 b). C. parvum was also the only species identified in diarrhoeic lambs in Aragón (NE Spain), but most subtypes (98%) belonged to the subtype family IId (Quílez et al. Reference Quílez, Torres, Chalmers, Hadfield, Del Cacho and Sánchez-Acedo2008 a). In order to provide data on the prevalence and zoonotic potential of Cryptosporidium isolates in other geographical areas, we genotyped Cryptosporidium specimens from diarrhoeic pre-weaned calves and lambs in Galicia, an important livestock-rearing area located in north-western Spain.
MATERIALS AND METHODS
Sample collection
Faecal specimens from diarrhoeic calves (n=61) and lambs (n=127) up to 21 days old were collected over a 1-year period (March 2008–April 2009) from farms in Galicia (NW Spain). Fresh faeces were collected directly from the rectum of calves originating from 27 cattle farms across the province of Lugo. Rectal cotton swabs were used to collect samples from lambs, originating from 28 sheep farms in the provinces of Lugo (n=12), Ourense (n=9), A Coruña (n=5) and Pontevedra (n=2). From each farm, between 1 and 15 animals were sampled throughout the period of study (mean: 3·6±3·2). Both fresh faeces and cotton swabs were examined for the presence of Cryptosporidium oocysts by microscopy of carbol fuchsin negative-stained smears (Heine, Reference Heine1982), and positive specimens were selected for molecular examination. A farm was considered infected if at least 1 positive sample was detected.
DNA extraction
Oocyst suspensions from fresh faecal specimens were purified by saturated NaCl solution as previously described (Elwin et al. Reference Elwin, Chalmers, Roberts, Guy and Casemore2001), resuspended in distilled water and stored at 4°C prior to use. Total DNA was extracted from 200 μl of oocyst suspensions by 3 cycles of freezing with liquid nitrogen (1 min) and thawing at 100°C (5 min), followed by incubation at 56°C for 30 min in lysis buffer containing proteinase K and purification over a spin column according to the manufacturer's instructions (Ultraprep Cell DNA Kit, AHN Biotechnologie GmbH). Oocysts were removed from rectal swabs by immersing and vigorously rubbing the head of the swab against the wall of a 2-ml microcentrifuge tube containing 1 ml of distilled water. An aliquot of 200 μl of each suspension was used for DNA extraction as described above. DNA extracts were stored at −20°C.
Cryptosporidium species/genotypes differentiation
Cryptosporidium species/genotypes were determined by nested PCR of a small-subunit (SSU) rRNA gene fragment and restriction fragment length polymorphism (RFLP) analysis with the endonucleases SspI, VspI and MboII, using primers previously described (Xiao et al. Reference Xiao, Singh, Limor, Graczyk, Gradus and Lal2001). PCR products and restriction fragments were subjected to electrophoresis in 1% or 2% agarose gels, respectively, and visualized by staining with ethidium bromide. Species/genotype assignment was made by comparing RFLP profiles to known profiles reported in the literature (Feng et al. Reference Feng, Ortega, He, Das, Xu, Zhang, Fayer, Gatei, Cama and Xiao2007). DNA sequence analysis was performed on samples that produced a banding pattern different from that of C. parvum, for confirmation of RFLP results for those samples.
Subtyping of Cryptosporidium parvum
Subtyping of C. parvum isolates was performed using a nested PCR to amplify a fragment of the GP60 gene (800 to 850 bp) as described previously (Alves et al. Reference Alves, Xiao, Sulaiman, Lal, Matos and Antunes2003). Selected SSU rRNA products as well as all GP60 products were purified and sequenced in both directions with the forward and reverse primers used in the secondary PCRs on a MegaBACE 500 sequencer (Amersham Biosciences) according to the manufacturer's instructions. The consensus sequences of both the sense and anti-sense strands were analysed using a BLAST search on the NCBI databases. Subtypes were named according to nomenclature described by Sulaiman et al. (Reference Sulaiman, Hira, Zhou, Al-Ali, Al-Shelahi, Shweiki, Iqbal, Khalid and Xiao2005). Unique partial SSU rRNA and GP60 sequences of Cryptosporidium identified in this study were deposited in GenBank under Accession numbers GU111577 to GU111580.
RESULTS
Prevalence and Cryptosporidium species/genotypes identification
Overall, 30 calves (49·2%) and 39 lambs (30·7%) were found to be positive for Cryptosporidium by microscopy after negative staining with carbol fuschin. The herd-level prevalence was 63% (17/27) and 60·7% (17/28) for cattle and sheep farms, respectively. PCR products of the SSU rRNA locus were obtained for 27 Cryptosporidium-positive isolates from calves and 23 of those from lambs, belonging to 16 and 10 farms, respectively. Restriction analyses generated profiles of C. parvum in all 27 calf isolates and 14 lamb isolates from 5 farms, whereas the remaining 9 isolates from lambs yielded a banding pattern indicative of the Cryptosporidium cervine genotype (Feltus et al. Reference Feltus, Giddings, Schneck, Monson, Warshauer and McEvoy2006). Three types of Cryptosporidium cervine genotype sequences were identified. Five isolates from 2 farms were identical to the cervine genotype 1 sequence (AF442484), but 2 isolates from 2 additional sheep farms exhibited a single nucleotide transition at position 425 (nucleotide G instead of nucleotide A) when compared with cervine genotype 2 (EF362480) and 1 isolate from another additional flock had an AAT deletion in addition to the latter nucleotide polymorphism. The sequence of the remaining isolate was unclear because of the presence of underlying signals. All farms where 2 or more animals were analysed had only 1 Cryptosporidium species/genotype, except for 1 sheep flock from Lugo where both C. parvum and a cervine genotype were concurrently identified.
GP60 subtypes
All 41 C. parvum isolates were analysed at the GP60 locus. PCR amplification was obtained for all 27 calf isolates and 12/14 isolates from lambs. All these products were sequenced successfully, except for 2 of the 12 lamb isolates, which had underlying signals in the electropherogram that prevented the accurate readout of sequences. Alignment of the sequences obtained with reference sequences downloaded from GenBank revealed that all isolates belonged to the C. parvum family IIa and divided them into 3 subtypes using the nomenclature system proposed by Sulaiman et al. (Reference Sulaiman, Hira, Zhou, Al-Ali, Al-Shelahi, Shweiki, Iqbal, Khalid and Xiao2005). Subtype IIaA15G2R1 was the most common and widely distributed in calves, identified in 26 (96·3%) isolates from 15 cattle farms, whereas the remaining calf isolate originated from a different farm and was identified as the novel subtype IIaA13G1R1. Two different subtypes were identified in sheep flocks including IIaA16G3R1, which was seen in 7 lamb isolates from a single farm and subtype IIaA15G2R1, identified in 3 isolates from 2 farms. All of these IIa subtypes were identical to each other in the non-repeat region, except for 1 lamb isolate subtyped as IIaA15G2R1, which differed from the remaining sequences by a 3-nucleotide (GGT) insertion downstream of the tri-nucleotide repeats, resulting in the addition of a Gly amino acid. Likewise, the novel subtype IIaA13G1R1 exhibited a single nucleotide transition (G instead of A) downstream of the repeat region, resulting in a change from Asp to Gly. A single subtype was identified in all farms where 2 or more animals were analysed.
DISCUSSION
The current study has revealed that Cryptosporidium spp. is a prevalent parasite in livestock farms in north-western Spain since more than 60% of cattle and sheep farms with diarrhoeic outbreaks harboured infected animals. The percentage of infected calves (49·2%) correlates well with cross-sectional studies conducted in central and other northern areas of the country, recording prevalence rates ranging from 52·5% to 64·7% in diarrhoeic pre-weaned calves by using negative-staining methods (Quílez et al. Reference Quílez, Sánchez-Acedo, Del Cacho, Clavel and Causapé1996, Reference Quílez, Torres, Chalmers, Robinson, Del Cacho and Sánchez-Acedo2008 b; De la Fuente et al. Reference De la Fuente, García, Ruíz-Santa-Quiteria, Luzón, Cid, García, Orden and Gómez-Bautista1998; Castro-Hermida et al. Reference Castro-Hermida, González-Losada and Ares-Mazás2002). The individual prevalence in lambs (30·7%) was similar to that previously reported in asymptomatic lambs in Galicia (31%) (Castro-Hermida et al. Reference Castro-Hermida, González-Warleta and Mezo2007) but much lower than that seen in central (45·2%) and especially the northeast of Spain (79·4%), where Cryptosporidium is considered one of the major agents in the aetiology of neonatal diarrhoea in lambs (Muñoz et al. Reference Muñoz, Álvarez, Lanza and Cármenes1996; Causapé et al. Reference Causapé, Quílez, Sánchez-Acedo, Del Cacho and López-Bernad2002).
The PCR has been found to be more sensitive than microscopic methods such as IFA for detecting Cryptosporidium oocysts (Santín et al. Reference Santín, Trout, Fayer and Xiao2008). However, a high proportion of microscopy-positive lamb specimens failed in the PCR amplification in the current study (16/39). This finding was probably associated with insufficient sample volumes and processing procedures used to analyse faecal specimens from lambs. The low number of oocysts seen in the microscopic analysis of all these PCR-negative specimens also suggests that a poor recovery was reached by immersing the rectal swabs in microtubes with distilled water. This relatively low sensitivity of the PCR in specimens with few parasites has previously been reported. Fayer et al. (Reference Fayer, Santín and Trout2007) and Mueller-Doblies et al. (Reference Mueller-Doblies, Giles, Elwin, Smith, Clifton-Hadley and Chalmers2008) found that most of the microscopy-positive samples that contained less than 10 oocysts per gram of faeces failed in the PCR amplification.
Restriction analysis of PCR products at the SSU rRNA locus confirmed the presence of only C. parvum in faecal specimens from calves, revealing that this is the major species responsible for diarrhoeic outbreaks on cattle farms in this geographical area. This situation is similar to that described in pre-weaned calves in Europe and North America, where Cryptosporidium species other than C. parvum have been reported to play a secondary role in calves under 30 days of age (Santín et al. Reference Santín, Trout, Xiao, Zhou, Greiner and Fayer2004, Trotz-Williams et al. Reference Trotz-Williams, Martin, Gatei, Cama, Peregrine, Martin, Nydam, Jamieson and Xiao2006; Geurden et al. Reference Geurden, Berkvens, Martens, Casaert, Vercruysse and Claerebout2007; Langkjær et al. Reference Langkjær, Vigre, Enemark and Maddox-Hyttel2007; Plutzer and Karanis, Reference Plutzer and Karanis2007; Thompson et al. Reference Thompson, Dooley, Kenny, McCoy, Lowery, Moore and Xiao2007; Broglia et al. Reference Broglia, Reckinger, Cacciò and Nöckler2008; Quílez et al. Reference Quílez, Torres, Chalmers, Robinson, Del Cacho and Sánchez-Acedo2008 b; Brook et al. Reference Brook, Hart, French and Christley2009; Coklin et al. Reference Coklin, Uehlinger, Farber, Barkema, O'Handley and Dixon2009), whereas C. bovis and C. ryanae (formerly known as the Cryptosporidium deer-like genotype) are dominant in post-weaned calves and heifers and C. andersoni is the most prevalent species in mature cattle (Fayer et al. Reference Fayer, Santín and Trout2007, Reference Fayer, Santín and Trout2008; Feltus et al. Reference Feltus, Giddings, Khaitsa and McEvoy2008).
Previous molecular studies in sheep have revealed geographical differences in the prevalence of zoonotic and non-zoonotic species/genotypes but, in contrast to findings in cattle, no globally consistent age-related variations in the distribution of different Cryptosporidium spp. have been detected. C. parvum was the most prevalent species in both sheep flocks and lambs younger than 6 weeks with either clinical cryptosporidiosis or subclinical infection in the United Kingdom (Mueller-Doblies et al. Reference Mueller-Doblies, Giles, Elwin, Smith, Clifton-Hadley and Chalmers2008; Pritchard et al. Reference Pritchard, Marshall, Giles, Chalmers and Marshall2007, Reference Pritchard, Marshall, Giles, Mueller-Doblies, Sayers, Marshall, Elwin and Chalmers2008). Similarly, 26 lambs aged from 2 weeks to 3 months from 5 farms with histories of diarrhoeic episodes in Italy only harboured C. parvum (Paoletti et al. Reference Paoletti, Giangaspero, Gatti, Iorio, Cembalo, Milillo and Traversa2009). By contrast, the cervine genotype was the most prevalent Cryptosporidium spp. in both ewes and their pre-weaned lambs in the United States as well as 4-month-old lambs from a flock in Scotland (Santín et al. Reference Santín, Trout and Fayer2007; Elwin and Chalmers, Reference Elwin and Chalmers2008). The cervine genotype was also seen in 9 out of 10 lambs aged 1 day to 10 weeks from 3 farms in Belgium and was the predominant genotype in post-weaned (>10 weeks old) lambs in Australia (Ryan et al. Reference Ryan, Bath, Robertson, Read, Elliot, McInnes, Traub and Besier2005; Geurden et al. Reference Geurden, Thomas, Casaert, Vercruysse and Claerebout2008).
Data from the current study show that C. parvum and the cervine genotype were identified in a similar number of farms (5 and 6 farms, respectively) and indicate that both are involved in the aetiology of diarrhoeic outbreaks in sheep flocks in Galicia (NW Spain), although other enteropathogens were not excluded. These findings contrast with results of a previous study in Aragón (NE Spain), where C. parvum was the only species identified within 139 isolates from diarrhoeic lambs up to 3 weeks of age (Quílez et al. Reference Quílez, Torres, Chalmers, Hadfield, Del Cacho and Sánchez-Acedo2008 a). It is noteworthy that concurrent infections by other Cryptosporidium spp. may not be excluded in either our study or most of the molecular studies mentioned above since they were based on SSU rRNA-based PCR tools which selectively amplify the dominant Cryptosporidium populations (Cama et al. Reference Cama, Gilman, Vivar, Ticona, Ortega, Bern and Xiao2006; Feng et al. Reference Feng, Ortega, He, Das, Xu, Zhang, Fayer, Gatei, Cama and Xiao2007). In Australian faecal specimens from 477 pre-weaned lambs screened at the SSU rRNA locus the presence of C. bovis (n=52), the cervine genotype (n=10) and C. parvum (n=2) was revealed, but 51 additional C. parvum isolates were identified by using a C. parvum and C. hominis-specific qPCR at a unique Cryptosporidium-specific protein-coding locus, suggesting that C. parvum-specific primers are necessary to determine the true prevalence of this species (Yang et al. Reference Yang, Jacobson, Gordon and Ryan2009).
Subtyping analyses at the GP60 locus confirmed that all but 1 isolate from calves belonged to the allele IIaA15G2R1, previously reported as one of the major subtypes responsible for zoonotic cryptosporidiosis (Glaberman et al. Reference Glaberman, Moore, Lowery, Chalmers, Sulaiman, Elwin, Rooney, Millar, Dooley, Lal and Xiao2002; Chalmers et al. Reference Chalmers, Ferguson, Cacciò, Gasser, Abs El-Osta, Heijnen, Xiao, Elwin, Hadfield, Sinclair and Stevens2005; Sulaiman et al. Reference Sulaiman, Hira, Zhou, Al-Ali, Al-Shelahi, Shweiki, Iqbal, Khalid and Xiao2005; Alves et al. Reference Alves, Xiao, Antunes and Matos2006) and one of the most prevalent C. parvum subtypes in calves in Europe and North America (Peng et al. Reference Peng, Wilson, Holland, Meshnick, Lal and Xiao2003; Alves et al. Reference Alves, Xiao, Antunes and Matos2006; Trotz-Williams et al. Reference Trotz-Williams, Martin, Gatei, Cama, Peregrine, Martin, Nydam, Jamieson and Xiao2006; Geurden et al. Reference Geurden, Berkvens, Martens, Casaert, Vercruysse and Claerebout2007; Xiao et al. Reference Xiao, Zhou, Santín, Yang and Fayer2007; Broglia et al. Reference Broglia, Reckinger, Cacciò and Nöckler2008; O'Brien et al. Reference O'Brien, McInnes and Ryan2008; Quílez et al. Reference Quílez, Torres, Chalmers, Robinson, Del Cacho and Sánchez-Acedo2008 b; Wielinga et al. Reference Wielinga, de Vries, van der Goot, Mank, Mars, Kortbeek and van der Giessen2008; Brook et al. Reference Brook, Hart, French and Christley2009). This finding suggests that the genetic heterogeneity of C. parvum isolates at this locus among the cattle farms in this geographical area is much more modest than that seen in an earlier study in the north of the country, which revealed the presence of 7 different subtypes within 149 isolates from diarrhoeic pre-weaned calves; these differences, however, may be due to the smaller number of specimens analysed in the present study (Quílez et al. Reference Quílez, Torres, Chalmers, Robinson, Del Cacho and Sánchez-Acedo2008 b). The remaining isolate identified as the novel subtype IIaA13G1R1 originated from the northernmost farm studied and is geographically distant from the other farms.
Low genetic variability was also reported within the C. parvum isolates from lambs which displayed only 2 subtypes within the family IIa, including the widely distributed subtype IIA15G2R1 and subtype IIaA16G3R1. This latter subtype had not previously been reported in sheep and was identified in 7 isolates from lambs on a single farm. This subtype is not common in cattle and seems to play a minor zoonotic role, since it has only been identified in modest numbers of specimens from calves and only 1 case of human infection has been recorded (Strong et al. Reference Strong, Gut and Nelson2000; Peng et al. Reference Peng, Wilson, Holland, Meshnick, Lal and Xiao2003; Trotz-Williams et al. Reference Trotz-Williams, Martin, Gatei, Cama, Peregrine, Martin, Nydam, Jamieson and Xiao2006; Thompson et al. Reference Thompson, Dooley, Kenny, McCoy, Lowery, Moore and Xiao2007; Wielinga et al. Reference Wielinga, de Vries, van der Goot, Mank, Mars, Kortbeek and van der Giessen2008; Brook et al. Reference Brook, Hart, French and Christley2009). Nevertheless, the previous finding of subtype IIaA16G3R1 as the second most prevalent allele in cattle farms in the north of Spain suggests that it could have a wide distribution in ruminant farms in some areas of the country (Quílez et al. Reference Quílez, Torres, Chalmers, Robinson, Del Cacho and Sánchez-Acedo2008 b).
Previous studies on the variability of C. parvum subtypes in sheep are limited and have been applied to modest numbers of specimens. GP60 typing results have been published to date for only 5 isolates from lambs in the UK, 2 specimens from sheep in Portugal and 1 specimen in Belgium, revealing the presence of subtypes IIaA15G2R1 and IIdA21G1 (Chalmers et al. Reference Chalmers, Ferguson, Cacciò, Gasser, Abs El-Osta, Heijnen, Xiao, Elwin, Hadfield, Sinclair and Stevens2005; Alves et al. Reference Alves, Xiao, Antunes and Matos2006; Geurden et al. Reference Geurden, Thomas, Casaert, Vercruysse and Claerebout2008). However, a large sample size study conducted in north-eastern Spain concluded that the C. parvum subtype family IId is strongly adapted to both lambs and goat kids. That study identified 11 different subtypes on 137 C. parvum specimens from diarrhoeic pre-weaned lambs, although most subtypes (9) and isolates (98%) belonged to the family IId and only 3 isolates displayed subtypes within the family IIa, namely IIaA15G2R1 and IIaA18G3R1 (Quílez et al. Reference Quílez, Torres, Chalmers, Hadfield, Del Cacho and Sánchez-Acedo2008 a). Further research is needed to understand why there are such dramatic differences in both the diversity and identity of C. parvum subtypes in lambs within a limited geographical area, although it could probably be related to differences in management practices.
In conclusion, this study indicates that both pre-weaned calves and lambs are a reservoir for zoonotic C. parvum subtypes in north-western Spain, although the Cryptosporidium cervine genotype is also involved in the aetiology of diarrhoeic outbreaks in some sheep flocks. This genotype has occasionally been reported in humans and some authors have suggested that it could emerge as an important human pathogen, but it is not at present regarded as a public health risk. This is the first report of the Cryptosporidium cervine genotype from Spain.
ACKNOWLEDGEMENTS
We would like to thank Miguel Viña, Sara Cienfuegos, Diego Fariñas and María Soilán for their help with collecting the samples and the UK Cryptosporidium Reference Unit staff Stephen Hadfield, Guy Robinson and Kristin Elwin for scientific support.
FINANCIAL SUPPORT
This work was supported by a Post-doctoral Research Grant (Xunta de Galicia, Spain) to P. Díaz, and funds from Spanish (AGL2009-10590) and regional (DGA-B82) research programmes.
