Comparative Sequence Analysis of Genes Encoding Outer Proteins of African Swine Fever Virus Isolates from Different Regions of Russian Federation and Armenia

African swine fever (ASF) is an economically important disease of domestic pigs. ASF is endemic in most of sub-Saharan Africa, including Madagascar Island; the highest incidence of disease being recorded from the West and East Africa to the Southern Africa. Disease outbreaks have also occurred in Europe, South America and the Caribbean. In 2007 it was introduced into Georgia, and has since spread throughout the Caucasus and into southern Russia. There is no vaccine or treatment available to control ASF v irus. Therefore t imely ASFV detection and characterization are crit ical to understand and contain its spread. In this report, we describe the nucleotide structure analysis of genes E183L, KP177R and O61R encoding for proteins p54, p22 and p12, respectively, for different ASFV isolates collected within three years from South European regions of Russia and Armenia. The comparative analysis of these demonstrated variability in p12 sequences of the isolates from d ifferent geographic regions and hosts, whereas p54 and p22 sequences were conserved among the isolates. However, hydropathy profile analyses did not reveal any structural variat ions for all three proteins. It suggests that p12 genes, but not p54 and p22 genes, is under strong selective pressure and can be a valuable genomic marker for studying of evolutionary pathways and genetic diversity of ASFV isolates.


Introduction
The causative agent of African swine fever (ASF) is a large desoxyribovirus that has been assigned to Asfarviridae fa mily [1]. Dixon L.K. et al. noted, that although initially ASFV was classified as an Iridovirus based largely on virion morphology, increasing molecular bio logical data led to its reclassification as the sole member of a new DNA virus fa mily, Asfarviridae (A s f a r -A frican swine fever and related viruses). African swine fever virus (ASFV) replicates in cytoplasm and has variable virulence to domestic pigs, with infect ions ranging from h ighly lethal to subclinical. ASFV genes encoding proteins which modulate host immune response, viral v iru lence to domestic swine, and the ability of ASFV to replicate and spread in its tick vector. ASFV is the only known DNA-arbovirus.
Do mestic and wild pigs are susceptible. African bush pigs and warthogs infected with ASFV are generally asymptomat ic, with low level viremia. ASFV persistently infects ticks of the genus Ornithodoros fro m wh ich ASFV can be isolated 5-10 years post infection [2]. ASFV infection of domestic swine results in several d isease forms, ranging fro m h ighly lethal to subclinical depending on contributing viral and host factors. African swine fever is characterized by oedemas, ascites and haemorrhages, virus replication and spread within mononuclear-macrophage system, long-term carrier state, nonvirulent strains of v irus induce latent infection. Its virion diameter is ~200 n m, contains more than 50 proteins and consists of several concentric layers enclosing an electron-dense nucleoid, containing a double stranded DNA genome of appro ximately 190 kilobase pairs (kbp). The core is enwrapped by an inner lipid envelope that lies beneath the icosahedral capsid [3,4]. Ext racellular part icles possess an additional envelope derived fro m the plasma me mb rane [4].
Analyses of cell ext racts revealed the presence of more than 100 viral proteins synthesized on different phases of ASFV life cycle [5][6]. ASFV p 54 is an externally located of African Swine Fever Virus Isolates from Different Regions of Russian Federation and Armenia viral structural protein of 25-27 kDa, encoded by the virus gene -the open reading frame (ORF) E183L [7][8].
ASFV protein p 54 is involved in the adsorption of the virion on susceptible cells and the early steps of v iral infection [9].
Published data have shown that ASFV p54, p30, p22 and p12 are essential v irus proteins involved in early events of the replication [10,11].
The protein encoded by the gene -ORF KP177R, is an early structural protein of apparent molecu lar weight 22 kDa, located externally in the viral part icle. The protein contains a hydrophobic region at the N-terminus with the characteristics of a signal peptide and seems to appear transiently in the plasma memb rane early after ASF v irus infection [12].
The ASFV virus protein p12 (ORF O61R) is involved in virus attachment to the host cell and located within the outer envelope of the virions. This protein is not synthesized during early phases of infect ion and undergoes post translational modification [13]. The protein p 12 has trans-me mb rane domain and cysteine-rich region wh ich is responsible for the dimerization at the C-tail.
Multimers of the protein with an apparent mo lecular mass of 17 kDa in the absence of 2-mercaptoethanol were detected. Labelling e xperiments with[35S] methionine and immunoprecipitation with specific antibodies directed against protein p12 indicated that the gene is expressed during the infection of swine macrophages with all of the viruses tested. The nucleotide sequence of a DNA frag ment containing the gene encoding for protein p12 in several virus strains has shown that the 5' flanking region is conserved in all the virus isolates sequenced, whereas the intergenic reg ion downstream of the gene varies among different isolates [14].
Currently, there is no vaccine available against ASF and the disease is controlled by animal quarantine and slaughter. It is believed that the lack of efficient ASFV vaccines might be due to unique molecular and b iological propert ies of ASFV proteins responsible for virus-cell interactions: p54, p22 and p12.
It is believed that extensive studies of ASFV structure, genes, immune mechanis ms which affect viral replication, virus-host interactions, and virulence will help to effectively control this dangerous agent.
This article is devoted to the investigation of the nucleotide structure of genes encoding outer proteins p54, p22 and p12 of African swine fever virus isolates from the different regions of Russian Federation and Armenia.

Cells and Viruses
Primary porcine blood macrophage cell (PBMC) or leucocyte cultures were prepared from swine blood. Briefly, heparin-treated swine blood was incubated at 37°C for 1 h to allo w sedimentation of the erythrocyte fraction. Mononuclear leukocytes were separated by flotation over a Ficoll-Paque (Pharmacia). The monocyte-macrophage cell fraction was cultured in Du lbecco's modified Eagle's medium containing 5% fetal bovine serum (FBS, Sig ma) [15].

He mads or pti on (HAD) Test
The hemadsorption test [16] is based on the ability of pig erythrocytes to adhere to the surface of pig monocyte or macrophage cells infected with ASFV. Three days old leucocyte culture in 96-well microplates was inoculated with 10-fo ld dilutions (200µl/ well) of treated sterile samp le of med iu m or virus (at least four wells per dilution).
After that 20 µl of freshly prepared 1% suspension of pig erythrocytes in the physiologic buffered solution (PBS) were added to each well. The inoculated cells were incubated at 37°, 5% CO2 during 7 days and the plates were checked for the presence of HAD.
The DNA templates used for the amp lification react ions were obtained either fro m peripheral blood of infected pigs, for the wild-type viruses, or fro m supernatants of ASFV infected cell cultures.
The DNA samples used in the PCR were heated at 95°C for 5 min before addition to the reaction mixture and PCR-amplification PCR frag ments consisted of a single band that varied in size among the virus genes when analyzed by 1,5% agarose gel electrophoresis [14,17]. PCR products were cloned into plas mids pTZ57R/T (Fermentas, Republic of Lithuania) and pAL-TA (ZAO Evrogen, Russia) [18]. DNA samples, containing full size copies of genes encoding p54, p22 and p 12 (E183L, KP177R and O61R) were obtained by PCR amplificat ion fro m purified DNA of different ASFV isolates using the specific primers. Reco mbinant clones were generated by inserting PCR-products of p54 genes and p22 genes into pTZ57R/T or PCR-p roducts of p12 genes into pAL-TA. DNA samp les of reco mb inant plasmids, containing p54 genes, p22 genes and p12 genes, were sequenced using the same primer pairs as for PCR.
The sequences were obtained for at least two independent clones of each virus isolate.

DNA Sequencing and Computer Analysis
Cloned p54, p 22 and p12 genes were sequenced by the dideoxynucleotide chain terminator method, according to standard procedures [19].
Co mparative analysis of DNA sequences of E183L, KP177R and O61R genes of different ASFV field isolates and strains was performed using the package Bio Ed it 6.0. Nucleotide

Hydropathy Profiles
Hydropathy profiles of ASFV proteins p54, p 22 and p12 were obtained following the procedure of Hoop & Woods [20].

Phylogenetic Analysis
Bio Ed it 6.0 software packages was used for the phylogenetic analysis. Phylogenetic trees were constructed using neighbour-joining algorith m [21]. A 603 bp, 534 bp, 138 bp frag ments were used to construct phylogenetic trees for p54, p22 and p12 genes, respectively.

He mads or pti on Test
ASFV protein CD2v is responsible for the attachment of erythrocytes to infected cells and this ASF' v irus protein is the most variable protein [22,23]; we examined all Russian isolates for the presence of the phenomenon hemadsorption during their reproduction.
HAD analyses demonstrated that all tested ASFV isolates fro m different regions of Russian Federation and Armen ia Since the character of hemadsorption during the virus reproduction with these isolates was identical, indicating a reasonable stability of the gene structure, we considered it is not necessary to analyze the gene encoding this protein.
Therefore, the ORF E183L locates at the 5' end negative-chain DNA of the Georg ia 2007/ 1 isolate genome (162222-162776 nt) and codes for a late induced structural glycoprotein p54 of 25 kDa (Fig. 2).
The ORF KP177R locates at the 5' end of the positive-chain DNA of Geo rgia 2007/ 1 isolate genome (3212 -3781 nt) and codes for an early induced external viral structural protein p22 of 22 kDa (Fig. 2) The ORF O61R locates at the positive-chain DNA central left part of Georgia 2007/1 isolate genome (128803-128988 nt) and codes for an early induced structural protein p12 of 12 kDa (Fig. 2). African ASFV isolates had lower the p54 gene nucleotide sequence identity with that of European and American iso-of African Swine Fever Virus Isolates from Different Regions of Russian Federation and Armenia lates, indicating greater nucleotide heterogeneity among viruses from different geographic regions and indicating a common origin for non-African isolates [24].  A few mutations in the sequence of p12 gene were observed between 170 to 265 nucleotides when different isolates were co mpared, whereas identical sequences were found when European group isolates were analysed (Lisbon 57, Espana 70 and OURT 88/ 3) ( Figure 5).

Hydropathy Analysis
Since we found single nucleotide substitutions in the p22 and p12 genes, and for the p12 gene nucleotide substitution occurred at the N-terminus with a net positive charge. It was necessary to check to see whether they affect the functional properties of encoded by these genes proteins. Hydropathy analysis makes possible recognition of protein function modifications when respective gene sequences vary. It was performed by using predicted amino acid sequences of ASFV p 54, p22 and p12 p roteins.
The hydrophilicity profile of p54 ( Figure 6) revealed the presence of a highly hydrophobic stretch, formed by 37 amino acid residues (a.a.) of p54 and a long hydrophilic C-terminus formed by 130 a.a. residues. There was also a highly hydrophobic stretch, consisted of 26 amino acid residues and a long hydrophilic C-terminus formed by 140 a.a. residues in hydrophilicity profile of p22 (Figure 7).    The deduced amino acid sequences of the p12 shows a stretch of 22 hydrophobic residues that functions as to anchor the protein in the external virus envelope. A hydrophilic C-terminus formed by 13 a.a. residues (Figure 8).
Despite few mutations present in the sequences of p12 gene of Russian and Armen ian isolates within regions from 195 to 248 and fro m 296 to 320 nucleotides, we did not observe changes of the structure hydrophobic trans me mbrane segment, the hydrophilic C -terminus formed by 13 a.a. residues. The p22 gene phylogenetic analysis demonstrated that Russian isolates and Georgia 2007/1 isolate cluster together of African Swine Fever Virus Isolates from Different Regions of Russian Federation and Armenia and are closer related to European isolates than to Southern African isolates (Figure 9b).

Phylogenetic Anal ysis of AS FV Isolates
All Russian and Armen ian isolates had specific mutations in O61R gene. The phylogenetic analysis for p12 gene also demonstrated that Russian isolates and Armenia 2007 isolate formed a group separate from European and African isolates. Within this group Russian and Armenian isolates formed several subgroups with Stavropol and Armenia closest to each other and Elbrus, Orenburg and Georg ia sharing higher homology with each other (Figure 9c). While Volgograd 2010 isolate formed another subgroup of its own.
Overall, Stavropolian and Armenian isolates share high homology between each other and with Georgia isolate suggesting that all Russian isolates may share a co mmon origin with Georgia.
Previous results of Carrascosa A.L. et.al. showed that the deduced amino acid sequence of p12 protein included a trans-membrane do main, cysteine rich C-terminal region and a stretch of 22 hydrophobic residues, which functions as to anchor the protein in the external virus envelope [4]. The absence of cleavable N-terminal signal sequences in the predicted amino acid sequence of p12 protein suggests that the polypeptide is inserted into the membrane through the mechanis m, wh ich has been proposed for other proteins of low mo lecular mass [30].
The hydropathy profile o f p 54 reveals the presence of a very hydrophobic stretch, formed by 21 amino acid residues, within the N-terminal reg ion (residues 33 to 53) of the protein, wh ich most likely represents a transmembrane doma in [31].
The analyses of hydropathy profile allows to determine the functional properties of protein structure [32], in our case hydropathy profile of ASFV proteins p54, p 22 and p12 corresponded to the results of a similar analysis for those proteins, but did not show significant differences of their structures.
We observed that the ASF virus, which has a h igh degree of variability, changes in the process of passaging in vivo. Prior to phenotypic changes, ASFV isolates accumulate silent genomic mutations specific for different isolates [33]. Therefore, we chose the outer ASFV proteins that are under selective pressure and involved in the processes of attachment and penetration into the cell.
Bastos A.D. et al. [34] analy zed strains of African swine fever virus fro m different geographic zones by partial p72 gene characterization and the phylogenetic analyses. Genotyping distributed all ASFV isolates to 10 clusters: genotype I included all Eu ropean strains and some isolates fro m West Africa (fro m Angola to Senegal). Remaining 9 genotypes come fro m South, Central and East Africa and Madagascar.
The present genotyping strategy includes a two-step genetic characterisation approach whereby p72 gene sequencing is used to delineate genotypes, prior to intra-genotypic resolution of viral relationships by central variable region (CVR) characterisation of the 9RL ORF classify the ASFV isolates relative into the 22 currently known genotypes [35].
Malogolovkin A. et.al. [36] co mpared sequences of genes encoding p72 and showed that gene B646L of all Russian isolates was highly conserved with no substitutions.
The serotype-based classification of ASFV strains previously developed by Balyshev V.et al. [37] d istributed all ASFV strains in ten groups including eight serotype-based groups, one group of non-serotyped strains and one group consisting of heterogeneous isolates. This classification characterizes strains into different seroimmunotypes based on the results of in vivo the cross-challenge and in vitro the hemadsorption inhibition test [38]. We found out that the immunodominant ASFV p54 is one of the virus proteins responsible for serotype specificity [24].
Phylogenetic analysis based on gene p54 conducted earlier in our laboratory allowed to div ide all tested ASFV strains and isolates into 8 groups consistent with serotype-based classification [24]. A ll Russian, Georg ian and Armenian isolates belonged to VIII seroimmunotype together with Rhodesia 1984 strain. Co mparat ive sequence analysis of p54 genes of these isolates showed that they were identical, thus, we concluded that the isolates belonging to the same serotype have identical genes encoding p54.
Earlier data of Angulo A. et al. [14] demonstrated that the p12 5'-flanking region sequence was conserved among the isolates, whereas sequences downstream of this gene were highly variab le in length and contained direct repeats in tandem. However, the main properties of the p12 protein were not altered in the original isolates and viruses adapted to grow in established cell lines. As noted by Angulo A., et al. [39], there is no correlation between the length of this sequence and any of the properties of these viruses, such as pathogenicity or capacity to grow in d ifferent cell types. Rowlands  The phylogenetic analysis of 123 concatenated genes separated the viruses into two majo r clusters that correlate with their geographical distribution [41].
Gallardo C. et.al. compared sequences of genes encoding p72, p54, p30 and central variab le region (CVR) to increase resolution of additional loci and to study geographic distribution of ASFV isolates [42]. The CVR within the ORF B602L had been found to be the most useful locus for d ifferentiation of closely related isolates and identification of p72-based virus subgroups. Subsequent analysis of the genetic relatedness of Russian and Armenian ASFV isolates based on p12 gene intergenic region demonstrated some differences: Russian and Armenian isolates originate fro m Geo rgia, but contained 9 d ifferent nucleotide substitutions and 5 insertions.

Conclusions
Our assumption that the gene p12 (in particular, its intergenic region) analysis can provide reliable information about the phylogenetic relationship between different isolates was fully confirmed in this study phylogenetic and comparative sequence analyses. Between nucleotides 218 and 255 we found insertions typical for these isolates and Georgia 2007/ 1 strain. A characteristic feature of all investigated genomes of African isolates was the presence of the shorter insertions in the same region. Unfortunately, we did not find more information in Gene Bank and had no data on the sequences of the p12 gene for the Mozamb ique 1979 and Rhodesia 1984 isolates, which would allo w us to exactly determine the feasibility of using phylogenetic analysis for the specified reg ion of ASF virus genome. Nevertheless, our findings suggest there is a need to analyze this region, because we found differences in the isolates with different passage history and of different orig in.
The p12 based phylogenetic analysis allowed to define subgroups among Russian isolates while showed the same relatedness with Eu ropean and African isolates as p22 and p54 based phylogenetic analyses.
The differences between p12 gene sequences may be reflective of d ifferent transmission pathways of the isolates. It is possible that the Volgograd 2010 isolated from a wild boar accumulated numerous genomic changes due to higher transmission rate in these hosts or that a diverse ASFV strain pool circulating in the wild boar population. While ASFV isolates from do mestic pigs in Krasnodar, Stavropol and Orenburg reg ions are highly ho mologous between each other and their spread was likely mediated by humans.
Because our results have confirmed the stability of p54 genes in isolates belonging to one serotype, we can recommend analysis of this geno mic reg ion to identify changes in circu lating virus' serotype, similar to that of long time epizootic in Spain and Portugal, when the virus I serotype was gradually supplanted by the IV serotype virus.
Overall, co mparat ive analysis of the genes encoding p54, p22 and p12 proteins presented in this study indicates that p12 gene, but not p22 and p54 genes, is under strong selective pressure and can be a valuable genomic marker for studying of evolutionary pathways and genetic diversity of ASFV isolates.

Nucleotide Sequence Accession Numbers
The DNA sequence data in this report have been submit-