Samples collected in 2008 and 2009, from 49 turkey flocks of 6 to 43 days in age and presenting clinical signs of enteric disease and high mortality, were tested by polymerase chain reaction and reverse transcription–polymerase chain reaction for the presence of viruses currently associated with enteric disease (ED) syndromes: astrovirus, reovirus, rotavirus, coronavirus, adenovirus, and parvovirus. Turkey astroviruses were found in 83.67% of the cases and turkey astrovirus 2 (TAst-2) in 26.53%. The investigations directly demonstrated the high prevalence of turkey parvovirus (TuPV) in 23 flocks (46.9%) experiencing signs of ED, making this pathogen the second most identified after astroviruses. Phylogenetic analysis on a 527 base pair-long region from the NS1 gene revealed two main clusters, a chicken parvovirus (ChPV) and a TuPV group, but also the presence of a divergent branch of tentatively named “TuPV-like ChPV” strains. The 23 Hungarian TuPV strains were separately positioned in two groups from the American origin sequences in the TuPV cluster. An AvaII-based restriction fragment length polymorphism assay has also been developed for the quick differentiation of TuPV, ChPV, and divergent TuPV-like ChPV strains. As most detected enteric viruses have been directly demonstrated in healthy turkey flocks as well, the epidemiology of this disease complex remains unclear, suggesting that a certain combination of pathogens, environmental factors, or both are necessary for the development of clinical signs.
Poult enteritis complex (PEC) is an enteric disease (ED) syndrome of turkey poults up to 6 wk of age which, in cases of increased mortality, is referred to as poult enteritis and mortality syndrome (PEMS). Morbidity and mortality are variable, and the economic impact of those syndromes is primarily due to poor production, failure of affected birds to grow, an increase in costs of therapy, and poor feed conversion efficiency 2,3,10. The etiology of the disease is not completely understood but is considered multifactorial 21.
Viral and bacterial agents were isolated from flocks with clinical signs of PEMS. Turkey coronaviruses, astroviruses, and reoviruses have been identified in flocks suffering from PEMS 11,12,19 but also in healthy turkey flocks 23,25, suggesting that a certain combination of pathogens and factors can lead to PEMS. Turkey coronavirus (TCV), a member of the group 3 coronaviruses 6, was identified in 1951 as the etiologic agent of a highly contagious enteric disease named “bluecomb disease” 31. Studies based on immunofluorescence and immunoperoxidase staining procedures determined that enterocytes of the jejunum and ileum 1,4,27 and the epithelium of the bursa Fabricius 11 were primary sites of replication for TCV. In recent years, TCV has been increasingly identified as an important cause of ED in turkeys and has been associated with PEMS 2,6; however, studies revealed that TCV was not required for installation of PEMS 5.
Astroviruses have been associated with acute gastroenteritis in mammals and turkeys, as well as with hepatitis in ducks 28,30, and have been detected in birds with PEMS although their exact role remains unclear 13,19,34. Turkey astroviruses (TAstVs) have been isolated from 1 to 3 wk-of-age turkey poults experiencing viral enteritis 29. A TAstV isolate from turkey flocks with severe signs of PEMS was isolated and referred to as TAstV-2 and proven to be molecularly distinct from the original TAstVs 19. Avian nephritis virus (ANV) is known to cause disease in young chickens; it evolves with distinct kidney lesions and enteritis 20 and was recently identified for the first time in commercial turkey flocks 23. Rotaviruses, such as avian rotavirus (AvRV), are a major cause of enteritis in a wide range of mammalian and bird species 15. Avian reoviruses (ARV) have been isolated from turkeys with PEMS 12,14,24,28,29 and also from chickens with runting-stunting syndrome 26. Ever since 1984, parvoviruses have been suspected to have a role in ED 16,18.
Samples included in this study and the PCR results for the targeted pathogens.A
Molecular biologic studies positioned the newly involved pathogen in the Parvoviridae family 17. Recent studies revealed the presence of parvovirus in commercial chicken and turkey flocks from the United States 35 and Hungary 22. The phylogenetic analysis comparing nonstructural gene (NS1) gene segments revealed a strong similarity between the chicken and turkey parvoviruses from Hungary 22 and also that the chicken and turkey parvovirus isolates formed distinct phylogenetic groups 35. The complete genome analysis of chicken parvovirus (ChPV) and the closely related turkey parvovirus (TuPV) revealed that these viruses are members of a distinct group of the Parvovirinae subfamily 9. In general, parvoviruses have a linear, single-stranded DNA, between 4 and 6 kilobase pairs, with two major genes; an NS1 that is conserved within studied parvoviruses and is used as a target for PCR-based diagnosis, and a structural viral protein gene 7. A previous study determined that the ChPV NS1-deduced amino acid sequence contains highly conserved motifs important for the initiation of parvovirus replication 36. The phylogenetic analysis of a fragment from the NS1 gene revealed that ChPVs and TuPVs cluster in separate groups 35. A similar finding was encountered in the case of Hungarian parvovirus isolates collected from flocks with clinical signs of ED and increased mortality 22.
The scope of this study was to determine, by direct demonstration, the incidence of the following enteric viral agents associated with PEC and PEMS in Hungarian commercial turkey flocks: astrovirus, coronavirus, reovirus, rotavirus, adenovirus and parvovirus; and to phylogenetically characterize the ChPV and TuPV strains circulating in Hungary. An additional aim of the present study was to find a quick and reliable way to differentiate strains belonging to the previously described clusters without the need for time-consuming nucleic acid sequencing.
MATERIALS AND METHODS
The 49 cases considered in this study were collected between January 2008 and December 2009 from Hungarian turkey flocks experiencing signs of PEC and PEMS. The age of the turkeys was below 43 days (Table 1). Pooled intestinal tissue was used; one sample corresponds to one flock or house and contains pooled intestinal tissue from 5 bird carcasses that succumbed to clinical signs of PEC or PEMS. Fresh intestinal tissue was homogenized in 5 ml sterile phosphate-buffered saline, the viral DNA and RNA was purified from the supernatants using High Pure Viral Nucleic Acid Kit (Roche, Basel, Switzerland), and amplifications were performed according to previously described protocols (Table 1) in a PCR Sprint Thermal Cycler SPRT001 (Hybaid, Basingstoke, U.K.). Following PCR, the amplicons were electrophoresed in a 0.5X Tris borate-EDTA-agarose gel (1.2% SeaKem® LE Agarose, Cambrex, East Rutherford, NJ) and stained with GR safe nucleic acid stain (InnoVita, Pleasant View, UT) at 80 V for 80 min.
Product sizes were determined with reference to 50-base pair (bp) molecular weight markers (Fermentas, Vilnius, Lithuania). Positive and negative controls were added to each run. Following electrophoresis, the TuPV amplicons were cut out from the gel and DNA was extracted with the BigDye kit (Qiagen, Hilden, Germany). Fluorescence-based direct sequencings were performed in both directions on the amplicons at Biogon Kft (Budapest, Hungary) employing an ABI 3100 genetic analyzer (Applied Biosystems, Carlsbad, CA). Nucleotide sequences were identified by BLAST search ( http://www.ncbi.nlm.nih.gov/BLAST/) against GenBank databases.
Nucleotide sequences were compiled and aligned using the Align Plus 4 software (Scientific & Educational Software, Cary, NC and phylogenetic analysis was performed employing the ClustalX program (University College, Dublin, Ireland; www:clustal.org). A phylogenetic tree of the nucleic acid sequences was constructed using sequence data of the 23 Hungarian TuPV strains from this study, 7 sequences retrieved from GenBank, and 20 sequences kindly provided by L. Zsak. Blocks of sequence data leading to 527 nucleotides for TuPV and 524 nucleotides for ChPV were used for the phylogenetic analysis. The phylogenetic tree was constructed by neighbor joining with a two-parameter distance matrix using the online Phylip program ( http://www.evolution.genetics.washington.edu/phylip). Goose parvovirus strain HG5 was used as outgroup. In order to establish whether there was any significant connection and correlation in the incidence of the investigated pathogens, statistical analysis was performed using the correlation testing and P-values were determined. The incidence of the pathogens was compared in pairs and type I error rate can increase; therefore, P-values were corrected using the Holm correction.
The analysis of all nucleic acid sequences available from the GenBank (and of those determined in the present study) revealed that strains clustered in the TuPV, ChPV, and TuPV-like ChPV groups have different Ava II cleavage patterns on the segment amplified by the diagnostic primer pair. Enzymatic digestion using AvaII of products amplified by these primers was performed according to the following protocol: 5 µl of the PCR amplicons were mixed with 3.5 µl ddH2O, 1 µl of AvaII enzyme (New England BioLabs, Ipswich, MA), and 1 µl of 10XNEB4 buffer (New England Biolabs). The mixture was heated to 37 C for 60 min and mixed at 10.62 × g at intervals of 30 sec in a Thermomixer Comfort device (Eppendorf, Hamburg, Germany). Product sizes were determined following gel electrophoresis performed according to the previously described protocol.
The poults presented higher than normal daily mortality (15% to 45%), stunted growth, diarrhea, dehydration, and high variation in weight amongst the individuals of one flock. The number of affected poults in one flock varied from 25% to 70%. At necropsy, the small intestine was partially filled with fluid-mucoid content and a large amount of gas. Dilation of the intestinal blood vessels and catarrhal enteritis were identified in the jejunum and ileum, and atrophy of the bursa of Fabricius was observed. Results of the PCR and RT-PCR amplifications are presented in Table 1. Hemorrhagic enteritis virus (HEV) was not identified in any of the samples, and only 14.28% positivity was found for TCV and ARV, respectively. TAstVs were found in 83.67% of the cases and TAstV-2 was found in 26.53% of the cases. The statistical analysis revealed that, with only a few exceptions, there were no significant correlations between the incidences of the investigated pathogens. The only statistically significant negative values were observed between the incidence of ARV and TuPV (corrected P-value = 0.02) and between TCV and TAstVs (corrected P-value = 0), while the only positive “almost” significant correlation was between the incidence of TAstV-2 and ARV (corrected P-value = 0.057).
Sequence data of 527-bp long were obtained for TuPV strains. The phylogenetic tree, constructed based on the nucleotide sequence of the analyzed NS1 gene segment, revealed an evident clustering of the virus strains of different species origin ChPV and TuPV (Fig. 1). The Hungarian turkey stains clustered in two groups separately from the American strains. The analysis of the deduced amino acid sequence of the ChPV strains resulted in 174 amino acid- (aa) long sequences and 175 aa-long in the case of TuPV strains (Fig. 2). In the case of the turkey samples, similar constant substitutions were found at different positions when compared with the reference ChPV strain (EU 304808). At the aa level, the identity between the relevant strains considered for comparison varied from 82.3% to 99.6%, being the lowest when comparing sample B70/09 with HM208289 and the highest between samples B33/10 and B31/6/09. Two samples presented a unique substitution at position 499 (B70/09 and B364/09), and a second substitution at position 523 was shared with 11 Hungarian samples. Three other samples presented a unique sequence at position 531 (B584/08, B307/2/09, B164/1/09). Positions are according to the considered reference strain (EU304808). Compared with the reference strain (EU304808), the lowest level of identity was found for samples B70/09 (Fig. 2). Overall, the levels of identity in cases of aa comparison were lower than for the nucleotide sequence comparisons, reflecting a very low prevalence of silent mutations.
Following the AvaII-based restriction fragment length polymorphism (RFLP; Fig. 3), products of various sizes were obtained in cases of strains belonging to the previously mentioned groups. In the case of TuPV strains, the enzymatic digestion resulted in clearly differentiable bands at the predicted sizes of 323 bp and 238 bp. In the case of ChPV strains, the sizes were 415 bp and 146 bp, while in the case of TuPV-like ChPV strains the enzymatic digestion resulted in three bands at different levels: 92 bp, 146 bp, and 323 bp (Fig. 4).
The viruses identified in the flocks included in this study are similar with previous reports; however, a broader range of pathogens was included, along with the scarcely known TuPV, for a better understanding of the epidemiology of ED in turkeys. The agent most recently suspected of contributing to ED is a new member of the Parvoviridae family 9,16,17,18. Our investigations directly demonstrate the presence of TuPV in 23 out of 49 Hungarian turkey flocks (46.9%) experiencing clinical signs of PEC or PEMS, making this pathogen the second most identified after astroviruses. The wide distribution of TuPV in American commercial flocks was previously reported 36; even so, the investigated flocks did not present any clinical signs or unsatisfactory feed conversion efficiency. There are no data regarding the distribution of this pathogen in the European poultry population. Recent studies have also shown the presence of TuPV and ChPV in Hungarian turkey and chicken flocks causing increased mortality and clinical signs of ED. To our knowledge, there are no data on the role of TuPV in the ED of turkeys; however, the results of the present survey accentuate their potential involvement, as suggested earlier 22. Astroviruses were the most-identified viral pathogens in the investigated flocks. Nevertheless, previous reports have shown a high prevalence of astroviruses in healthy turkey flocks 23,25, making this finding difficult to interpret. ANV, a pathogen known to infect and cause disease in young chickens, was found in 2 flocks, but no data are available concerning any possible involvement of this pathogen in PEMS, perhaps because it was only recently directly identified for the first time in healthy commercial turkey flocks 23. Positivity for ANV was confirmed by direct nucleotide sequencing (data not shown). Nucleotide sequences were identified by BLAST ( http://www.ncbi.nlm.nih.gov/BLAST/) search against GenBank. The eight flocks found to be free of astroviruses were infected with TuPV, TCV, or both. It was not unexpected that HEV was not detected in any of the flocks because the disease is well controlled and prevented by the extensive use of vaccination. The statistical analysis results are suggestive of a negative correlation between ARV and TuPV, and TCV and TAstVs, respectively; hence, if one of the viruses is present it is less likely for the second one to appear. The only “almost” positive correlation was found between TAstV-2 and ARV, as they were found most frequently together. Nevertheless, this type of conclusion should be cautiously formulated and interpreted and should always be reinforced and sustained by comprehensive experimental infections.
The sequencing of the amplicons of the TuPV/ChPV diagnostic PCR resulted in 527-bp long nucleotide sequences, as previously reported for TuPV on this NS1 segment 22,35,36. The phylogenetic tree constructed, based on the nucleotide sequence of the analyzed NS1 gene segment, revealed an evident clustering of the virus strains of different species origin, the ChPV group and the TuPV group. Hungarian TuPV strains clustered separately from all the American strains, in two groups; however, no time-connected grouping could be observed. Interestingly, samples B23/1/08 and B23/3/08 (Table 1) were collected from the same flock at the same time but from different houses; even so, they did not cluster together as expected. The same situation happened for samples B369/4/08 and B369/9/08. The poults from the same flock but different houses could have gotten infected with two different strains at the same time, or in the short period of time, dramatic changes occurred with the viral strains; this hypothesis is unlikely for parvoviruses. The low level of identity of the deduced amino acid sequences of the Hungarian strains is explicable by the identified substitutions.
The differences between the AvaII digestion pattern of parvovirus strains belonging to the TuPV, ChPV, and TuPV-like ChPV groups seem to provide a quick and reliable differentiation of all these strains without the need for nucleic acid sequencing. In the future, the newly described protocol could be used to obtain valuable epidemiologic data more rapidly and will turn out to be even more practical in cases of potential differences in the pathogenicity of these strains. Both possibilities should be revealed by future studies.
In conclusion, the present study provides strong evidence that there is a high prevalence of TuPV in Hungarian turkey flocks experiencing ED; this finding could imply a potential major role of the pathogen in the development of this complex syndrome. However, due to the considerable genetic diversity of all circulating virus strains, the limitation of current diagnostic techniques, and the fact that most enteric viruses incriminated in PEC and PEMS have been directly demonstrated in healthy turkey flocks as well, conclusions regarding the potential implication and role of these viruses in the pathogenesis of ED should be carefully formulated and sustained by planned experimental infections, combined with extensive phylogenetic analysis, and corroborated with epidemiologic field studies.
We are grateful to Laszlo Zsak for providing the 20 nucleotide sequences of American origin. This study was partially supported by NKB grant 15934/2010.