Nearly full-length genome characterization of canine parvovirus strains circulating in Nigeria.

Abstract Canine parvovirus type 2 (CPV‐2) emerged suddenly in the late 1970s as pathogen of dogs, causing a severe and often fatal gastroenteric disease. The original CPV‐2 was replaced by three antigenic variants, CPV‐2a, CPV‐2b and CPV‐2c, which to date have gained a worldwide distribution with different relative proportions. All previous studies conducted in Africa were based on partial VP2 gene sequences. The aim of this study was to provide a genome analysis to characterize the CPV strains collected in Nigeria, Africa. Rectal swab samples (n = 320) were collected in 2018 and tested by means of an immunochromatographic assay. Among the 144 positive samples, 59 were selected for further analyses using different molecular assays. The results revealed a high prevalence of CPV‐2c (91.5%) compared to the CPV‐2a variant (8.5%). The VP2 gene sequences showed a divergence from the strains analysed in 2010 in Nigeria and a closer connection with CPV strains of Asian origin. The non‐structural gene analysis evidenced amino acid changes never previously reported. The molecular analysis based on genomic sequences evidenced a geographical pattern of distribution of the analysed strains, suggesting a potential common evolutionary origin with CPV of Asian origin. This study represents the first CPV molecular characterization including all the encoding gene sequences conducted in the African continent and contributes to define the current geographical spread of the CPV variants worldwide.

reading frames (ORFs) encoding for two non-structural (NS1 and NS2) and two structural (VP1 and VP2) proteins through alternative splicing of the same mRNAs (Reed at al., 1988).
In the African continent, CPV has been described in South Africa and Namibia (Dogonyaro, Bosman, Sibeko, Venter, & Vuuren, 2013;Steinel, Venter, Van Vuuren, Parrish, & Truyen, 1998), Zambia (Kapiya et al., 2019), Mozambique (Figuiredo et al., 2017), Ghana (Folitse et al., 2017), Morocco (Amrani et al., 2016), Cape Verde (Costanheira et al., 2014), Nigeria (Chollom et al., 2013) and Tunisia (Touhiri et al., 2009). In Nigeria, only recently the molecular analyses based on the partial VP2 gene sequence of CPV strains described the circulating CPV variants (Apaa, Daly, & Tarlinton, 2016;Dogonyaro et al., 2013;Fagbohun & Omobowale, 2018). Unfortunately, all previous studies conducted in Africa lack a comprehensive sequence analysis including all the viral genome, thus preventing a more in-depth knowledge on the origin and evolution of the circulating CPV strains. The aim of this study was to characterize the CPV strains recently collected in Nigeria, analysing the nearly complete CPV genome sequence and comparing the obtained sequences with worldwide related sequences available in GenBank. 482 m a.s.l). Samples were submitted from eight private veterinary clinics, two from each state, and from kennels/breeders in the same areas. Rectal swabs were tested using an in-clinic assay for detection of CPV antigen (SensPERT ® Canine Parvovirus Test Kit, VetAll Laboratories), according to the manufacturer's instructions. Among the positive samples, 59 rectal swabs were selected and submitted for molecular analyses, where they were stored at −80°C until use.

| MATERIAL S AND ME THODS
Details are reported in Table S1.
Swab homogenates were obtained as previously described . Viral DNA was extracted from 200 µl of swab homogenate using the DNeasy Blood & Tissue Kit (Qiagen S.p.A.), according to the manufacturer's instructions. The presence of CPV DNA was confirmed using a primer pair (Table S2) in a PCR protocol amplifying a 700-bp fragment of the VP2 gene (Touihri et al., 2009), using the commercial kit GoTaq ® G2 DNA Polymerase (Promega Italia s.r.l.), as previously described (Mira, Dowgier, et al., 2018).
Amplicons were checked after electrophoresis on a 3% agarose gel supplemented with ethidium bromide.
Ten microlitres of each amplicon was digested with 5 units (U) of restriction endonuclease MboII (New England BioLabs ® Inc.) in a 50-µl reaction mix consisting of 5 µl of NEBuffer and 34 µl of nuclease-free water, under the following reaction conditions: incubation at 37°C for 2 hr and inactivation at 65°C for 20 min. The profile was determined by electrophoresis on a 3% agarose gel supplemented with ethidium bromide.
Specimens from each city of collection (n = 4 from Makurdi; n = 11 from Jos; n = 8 from Abuja and n = 5 from Lafia), representing dogs with different age, vaccinal and clinical status, were selected to determine the VP2 gene sequence (Table 1). The nearly complete VP2 gene sequence was determined using a primer pair (Table S2), which allows for amplification of a 1,745-bp fragment (Battilani et al., 2001;Mokizuki et al., 1996), using the commercial kit GoTaq ® G2 DNA Polymerase (Promega Italia s.r.l.), as previously described  with minor modifications (thermal conditions: 1 min for the annealing step). After electrophoresis on agarose gel supplemented with ethidium bromide, amplicons were purified with Illustra™ GFX™ PCR DNA and Gel Band Purification Kit (GE Healthcare Life Sciences) and submitted to BMR Genomics srl for direct Sanger sequencing with 6.4 pmol of the reverse primer used for amplification and of two additional internal primers (Table S2). Sequences were assembled and analysed using BioEdit software (Hall, 1999).
By excluding the VP2 gene sequences with complete nt identities, 8 CPV DNAs from different cities of collection were further selected for a further sequence analysis by amplifying a long genome sequence encompassing both major ORFs, without the 5′ and 3′ flanking sequences (Table 1). Sequence analyses were carried out using primer pairs described by Pérez et al. (2014), using the commercial kit GoTaq ® G2 DNA Polymerase (Promega Italia s.r.l.), as previously described (Mira, Dowgier, et al., 2018) with minor modifications (Mira, Canuti, et al., 2019). After electrophoresis on agarose gel, amplicons were purified and submitted for direct Sanger sequencing with a set of primers, as previously described (Mira, Purpari, Lorusso, et al., 2018). Sequences were assembled, analysed and submitted to nBLAST program (Zhang, Schwartz, Wagner, & Miller, 2000) to search related sequences into public domain databases. These sequence data were submitted to the DDBJ/EMBL/ GenBank databases under accession numbers MK895483-90.
Phylogenetic analyses based on the full-length NS1 and VP2 gene sequences and on the whole genome encompassing the two ORFs were conducted using the best-fit model of nt substitution with MEGA version X software (Kumar, Stecher, Li, Knyaz, & Tamura, 2018), inferred with the maximum-likelihood method based on the Tamura 3-parameter (T92) and Hasegawa-Kishino-Yano (HKY) models, with discrete gamma distribution (five rate categories) (G) and invariant sites (I) (bootstrap 1,000 replicates), the best-fitting models after the model test analyses (VP2: T92 + G; NS1: HKY + G; whole genome: HKY + G+I).

| RE SULTS
Among the collected 320 samples, 144 rectal swabs tested positive for CPV by in-clinic assay (45%). The prevalence of the positive samples for each city of collection is reported in Table 2. The presence of CPV DNA was confirmed in the selected samples (n = 59) using the conventional PCR assay. The same samples tested negative for CDV, CAdVs, CCoV and CRoV by gel-based or real-time (RT)-PCR assays. Based on the RFLP analysis, 54 CPV-positive TA B L E 1 Identification code, origin, age, vaccination and clinical status, strain and sequence information of the dogs selected for molecular investigation Amino acid change I60V in NS1 also lies at the same residue in the NS2-encoding sequence. Additional four amino acid changes in the NS2-encoding sequences were observed: D93E, T94A, D151N and M152V (Table 5). These changes resulted in silent mutations in the corresponding encoded NS1 protein.
The aa change K116R in the VP1 gene sequence is added to the aa changes of the VP2 gene sequence lying in the corresponding en-

| D ISCUSS I ON
Canine parvovirus has still been playing a main role in inducing severe and often fatal gastroenteritis in young or non-immunized dogs. During years, CPV spread and evolution have been well documented in North and South America, Europe and Asia (Miranda & Thompson, 2016;Zhou, Zeng, Zhang, & Li, 2017). More recently, data about its spread were also obtained from Australia and Africa (Amrani et al., 2016;Castanheira et al., 2014;Chollom et al., 2013;Dogonyaro et al., 2013;Figuiredo et al., 2017;Folitse et al., 2017;Kapiya et al., 2019;Touhiri et al., 2009;Woolford et al., 2017). Most of these studies were based on the partial or complete VP2 gene sequence, due to the involvement of the VP2 capsid protein in host switch and to its fast evolutionary rate Nelson, Palermo, Hafenstein, & Parrish, 2007;Shackelton, Parrish, Truyen, & Holmes, 2005), with limited information on other CPV encoding  Fagbohun & Omobowale, 2018). These analyses conducted in Nigeria evidenced firstly the circulation of CPV-2a strains (Apaa et al., 2016;Dogonyaro et al., 2013) and only recently of CPV-2b/2c types (Fagbohun & Omobowale, 2018  Moreover, the VP2 aa change was due to a nt change in the second base of the codon (VP2 c14g), but this change was common only to CPV strains of Asian origin Mira, Purpari, Lorusso, et al., 2018;Wang et al., 2016;Zhuang et al., 2019). The potential biological relevance of these changes has not been described yet and needs to be assessed in further studies.
As most recent Asian CPVs, the Nigerian strains displayed other three aa substitutions in the VP2 sequence (F267Y, Y324I and Q370R). While change at aa residue 324 is predominant in all three CPV variants in Asia (Geng et al., 2015;Yi, Tong, Cheng, Song, & Cheng, 2016;Zhao et al., 2017;Zhou et al., 2017), the other changes have been less frequently observed, mainly in China since 2013, and change Q370R has been detected only in CPV-2c strains (Geng et al., 2015;Guo et al., 2013;Mira, Purpari, Lorusso, et al., 2018;Wang et al., 2016;Zhuang et al., 2019). These aa substitutions are located in the greatest variable VP2 GH loop, comprised between aa 267 and 498, but while residue 267 is not exposed on the capsid surface (Chiang, Wu, Chiou, Chang, & Lin, 2016) and may not affect the antigenicity of CPV (Xu et al., 2015), residues 324 and 370 could have immunological implications or biological relevance. Indeed, residue 324 is subject to positive selection (Hoelzer, Shackelton, Parrish, & Holmes, 2008) and is adjacent to residue 323, which affects binding to the canine transferrin receptor . Residue 370 is close to residues associated with the ability of CPV to haemagglutinate, altering the pH dependence of haemagglutination or affecting the canine transferrin receptor (TfR) binding that determines the canine host range (Guo et al., 2013;Kaelber et al., 2012;Tsao et al., 1991).
Among the synonymous substitutions observed in the CPV-2a VP2 gene sequences, nt change a1275g has been previously described in CPV-2c strains (Amrani et al., 2016;Decaro, Desario, Amorisco, et al., 2013;Decaro et al., 2009). This change, observed in the strains UV1 and UV6, was detected in the binding region of the type-2a and type-2c specific probes of the minor groove binder (MGB) probe assay (Decaro et al., 2005. Although this change potentially accounts for the absence of VIC fluorescence in the 2a/2b and 2b/2c assays, specific additional studies are necessary to evaluate its real implication in the characterization of this CPV-2a mutant by the MGB probe assay, as previously done for the same substitution in the CPV-2c mutants (Decaro, Desario, Billi, et al., 2013). The in-clinic assay used in this study was able to detect the CPV-2a showing this substitution, as previously observed for another rapid assay used to test also the CPV-2c mutants (Decaro, Desario, Billi, et al., 2013).
Limited studies are available on the CPV non-structural genes (Hoelzer et al., 2008;Pérez et al., 2014), and, only recently, the analysis of the NS1 gene sequence was included in the CPV phylogenies from several countries (Canuti, Rodrigues, Whitney, & Lang, 2017;Grecco et al., 2018;Mira, Canuti, et al., 2019;Zhuang et al., 2019). In this study, sequence analysis revealed aa changes previously described mainly in NS1/NS2 gene sequences of CPV-2a/2c strains of Asian origin. Additional changes were also evident, some previously reported in South/North America and others never previously observed. This divergence may suggest the same ancestral origin with the CPV strains of Asian origin but a separate evolution, as well as a continuous adaptive process of the virus in separate environments. Indeed, some of these changes lay in the potential encoding sequence of functional domains (Mira, Canuti, et al., 2019) and, particularly, residues 351, 517 and 545 are located between the α5-and α6-helices, between the β5-and α11-helices and just close to the α11-helix of the helicase domain protein sequence, respectively, as illustrated in Niskanen, Ihalainen, Kalliolinna, Häkkinen, and Vihinen-Ranta (2010). Therefore, their role needs to be further evaluated.
The molecular analysis based on long genome sequences evidenced the geographical origin of the analysed strains rather than the clustering based only on the CPV antigenic variant. Therefore, this study supports further studies aimed to track the viral spread and elucidate the CPV evolution. Indeed, the recent evidence of specific aa changes, as well as the divergence from previous circulating strains, does not allow to rule out the possible introduction of these strains from other countries, highlighting the need for further studies on CPV whole genome in different geographical areas. This suggestion is supported by the evidence of genetic signatures typical of CPV or other canine viruses with different origins (Decaro, Campolo, et al., 2007;Martella et al., 2006;Mira, Purpari, Lorusso, et al., 2018), probably connected with the trading and transport of dogs between countries and continents.
In this study, with the aim to investigate the prevalence of the most frequent canine enteric viruses, all collected samples were analysed for selected pathogens. CPV was the only enteric viral pathogen detected in this study and this result confirmed the correlation between CPV infection and development of enteric signs, with a limited role of other viral enteric pathogens (Dowgier et al., 2017).
Nevertheless, further studies, based on a wider sampling also including the wild potential susceptible species, are necessary to better elucidate the effective spread of the other viruses in Nigeria.
This study represents the first CPV molecular characterization including all the encoding gene sequences conducted in the African continent and contributes to define the current geographical spread of the CPV variants worldwide. The evidence of mutations that have not been detected before suggests the need for further investigations in order to determine any biological consequences and underlines the continue evolution of CPV.

ACK N OWLED G EM ENTS
The authors would like to thank Dr. Ijeoma Chekwube Chukwudi, Dr. Pam Dachung Luka, Dr. Emmanuel Tumininu Obishakin and Dr.
Ukamaka Uchenna Eze for their skilful technical assistance and also practitioners, private veterinary clinics and dog breeders for sample collection.

CO N FLI C T O F I NTE R E S T
The authors of this manuscript declare that there are no conflicts of interest.