Phylogenetic Analysis of HA and NA Genes of Swine Influenza Viruses in Serbia in 2016-2018

Abstract Pigs are very important for the epidemiology of influenza A viruses, being commonly infected with the lineages of most adapted H1N1, H3N2, H1N2 swine subtypes. Epidemiological complexity of swine influenza is increasing by a periodic spillover of human or avian viruses in the pig population when genetic shifts can occur. The objectives of this research were to determine the presence of the influenza A virus in nasal and tracheobronchial swabs and lung tissue samples of ill and dead pigs on commercial farms, to determine circulating subtypes and characterize them through the phylogenetic analysis of hemagglutinin (HA) and neuraminidase (NA) genes. A total of 255 samples collected from 13 farms were analyzed by means of real-time RTPCR. The genome of influenza A virus was detected in 24 samples, which represented a 61.5% prevalence at the farms level (influenza A virus was confirmed in 8 out of 13 farms included in this study). Based on HA and NA gene sequences of 8 viruses, the circulation of H1N1 and H3N2 subtypes of influenza A viruses were determined. In addition, one farm exhibited a time separated circulation of H1N1 and H3N2 virus subtypes. Using Influenza Research Database, our viruses of the H1 subtype were classified into 1C.2.1 and 1A.3.3.2. clade. Based on the nucleotide sequences of HA genes, three viruses of the H1N1 subtype belong to the H1N1pdm09 lineage, and the other four to Eurasian “avian-like” H1avN1 lineage; while based on NA genes sequences, these seven viruses belong to Eurasian “avian-like” H1avN1 lineage. Both HA and NA genes of the virus of the H3N2 subtype belonged to the A/swine/ Gent/1/1984-like H3N2 lineage.


INTRODUCTION
Infl uenza is one of the most important zoonotic diseases nowadays. The causative agents belong to the family Orthomyxoviridae, and are assigned to Infl uenza A, B, and C genus according to their genetic and antigenic characteristics. Besides, in studies from 2011 onward, a genetically distinguishable novel virus [1,2] has been described, offi cially named Infl uenza D virus. Infl uenza A viruses (IAVs) through history have earned the biggest attention because they caused multiple epidemics in human and animal populations worldwide, as well as pandemics [3]. Infl uenza A viruses have a broad host range (humans, mammals, birds). The majority of susceptible animals become ill after infection with a certain subtype of infl uenza A virus (IAV) previously adapted to the specifi c family or class of animals. Pigs are very important for the epidemiology of infl uenza, being commonly infected with the lineages of most adapted H1N1, H3N2, H1N2 subtypes [4]. Besides this, pigs can be susceptible to the viruses previously adapted to humans or birds, representing in that way the potential intermediate host where genetic shifts occur. Antigenically distinct viruses can be further transmitted to both humans and bird species [5,6].
Swine infl uenza is an enzootic disease in most areas of the world with dense pig populations [7,8]. Within Europe, a wide study based on the molecular characterization of swine IAVs [6], provided details of the genome of circulating viruses, information about their diversity, pathogenicity, pandemic potential and correlation to the viruses circulating worldwide. The presence of infl uenza A virus in the swine population in Serbia was confi rmed in an investigation conducted on commercial farms in 2011-2012 [9]. The virus was not detected among analyzed wild boars in Serbia [10]. Although the circulation of the virus in domestic pig populations has been confi rmed, the molecular characterization of circulating subtypes has not been conducted yet.
The objectives of this research were to examine samples from sick and dead pigs from commercial farms from the north, central and east parts of Serbia for the presence of IAVs, to determine the circulating subtypes, and to characterize them through the sequence analysis of HA and NA genes.

Samples
Samples for IAV detection were collected during passive surveillance conducted in the period from August 2016 until January 2018 on 13 farms (marked A, B, C, D, E, F, G, H, I, J, K, L, M) located in the north, central and east parts of Serbia (Fig 1). All farms were farrow-to-fi nish farms with 800 to 1500 sows each. At the time of collection the sampled farms did not vaccinate the pigs against infl uenza.
Samples included nasal swabs from two to four months old pigs with infl uenza-like symptoms and lung tissue samples and tracheobronchial swabs from dead pigs of different ages with lung lesions resembling interstitial pneumonia. A total of 255 samples were analyzed for the presence of the virus including: 207 nasal swabs, 32 lung tissue samples, and 16  Collected swab samples were soaked in one milliliter of cell culture medium (DMEM, Gibco, ThermoFisher Scientifi c, USA), with the addition of 100 μg/ml of Gentamicin (Gentamicin, Hemofarm A.D., Serbia) and 2% of Amphotericin B (250 μg/ml) (Gibco™ Amphotericin B, Gibco, ThermoFisher Scientifi c, USA) in accordance to the procedure described in the OIE Terrestrial Manual [11]. The lung tissue samples were homogenized in a 1:10 ratio with the same medium. After the one-hour incubation at 4-8 °C and intense shaking of both swabs and lung tissue homogenates, they were centrifuged at 1700 g for 15 minutes [11]. The supernatant was removed and used for further analysis.

Extraction of nucleic acid, PCR, and sequencing
The viral RNA was extracted using Cador Pathogen Mini Kit (Qiagen, Germany) according to the protocol for purifi cation of pathogen nucleic acids from fl uid samples. Real-time reverse transcriptase-polymerase chain reaction (RRT-PCR) was performed using VERSO-1STEP qRT-PCR ROX Kit (Thermo Scientifi c, USA), and previously published primers and probes that amplify the part of the M gene of the IAVs [12,13]. The samples producing the sigmoid curve in the RRT-PCR were considered as positive. Eight samples with the Ct values below 30 were sent for whole genome sequencing to the OIE reference laboratory for swine infl uenza (IZSLER). The next-generation sequencing (NGS) was performed using Illumina Miseq platform. The sequence assembly and editing were done with CLC Genomic Workbench v.11 software.

Phylogenetic analyses
The obtained sequences of complete HA and NA genes were compared with available IAV sequences using Basic Local Alignment Search Tool (BLAST). Phylogenetic and evolutionary analyses HA and NA genes of Serbian swine IAVs were performed using software MEGA 7.0. The phylogenetic trees were constructed using the Maximum Likelihood method with 1000 bootstrap replicates [14].

Ethical approval
The conducted research is not related to the use of animals. No ethical approval was obtained because this study did not involve laboratory animals and involved only noninvasive procedures.

RRT-PCR results
Out of 255 tested, the IAV genome was detected in 24 samples. At the farm level, IAV was confi rmed in 8 out of the 13 farms included in this study (61.5%). Regarding the sample type, out of 207 nasal swab samples, 17 were positive, and out of the 32 lung tissue samples, the IAV genome was detected in 7. However, IAVs were not detected in any of the 16 tracheobronchial swabs (Table 1).   (Fig. 2).
Regarding the NA gene, 7 swine IAV strains showed the highest similarity with the sequences of NA gene of European "avian-like" H1N1 swine strains [6]. The NA gene of the H3N2 virus showed the highest similarity with the A/swine/Gent/1/1984-like H3N2 European isolates [6] (Fig. 3). Analysis of the amino acid sequence of the HA gene of H1N1 viruses from farms A, B and K using H3 amino acid numbering scheme (IRD) revealed the presence of glutamic acid (E) at position 77, aspartic acid (D) at position 190 and 225, and NTT (asparagine-threonine-threonine) glycosylation site at the position 278-280 (Table 3). Viruses from farms F, G, I and J in the amino acid sequence at the cleavage site of the HA gene possessed the PSIQSR motif, and D190/E225 combination within receptorbinding site ( Table 3).
Virus of the H3 subtype from farm B displayed the 226 leucine (L)/228 serine (S) combination in its amino acid sequence within the receptor-binding site (Table 3).

DISCUSSION
This study, conducted from 2016 to 2018, reveals the presence of the IAVs in Serbian swine on 8 of 13 examined farms, representing a prevalence of 61.5% of IAVs at the farm-level. However, considering that point mutations in the virus genome may affect the result of RRT-PCR [17,18], as well as the short time of virus excretion [19,20], the possibility of virus circulation in the remaining fi ve farms cannot be excluded. Although the nasal swab is considered the most preferable sample for IAV detection in swine [21], the IAV genome was most often detected in the lung tissue samples  [22,23], who reported the highest affi nity of swine infl uenza viruses for epithelial cells in the lungs. Furthermore, the results of our study showed that this type of sample is also most suitable for NGS of the IAV because of its high viral concentration in the lungs. The IAV genome was detected in only 8.21% of nasal swab samples that represented the majority of the analyzed samples.
Phylogenetic analysis of HA and NA gene sequences, revealed the circulation of H1N1 and H3N2 that represent two most prevalent swine subtypes of IAV [24,6,25,21]. During the survey in 2016-2018, six farms experienced the infection with H1N1 subtype of IAV, whereas in one farm two subtypes, H1N1 and H3N2, were detected in time-separate sampling. This is in line with the reports from neighboring countries, where these two subtypes were revealed in Croatia, Romania in 2009, Albania in the period 2007 to 2009, and in Hungary from 2010 to 2013 [26,6,27,28]. Viruses of the H1N2 subtype are enzootic in pig populations of several European countries such as Belgium, Denmark, Germany, France, and Italy [6]. However, this subtype has not been confi rmed in our study. Nevertheless, considering the small number of farms included in the study, the circulation of this subtype cannot be excluded.
Phylogenetic analyses of the HA gene sequence of three viruses of H1N1 subtype including A/swine/Serbia/2/2017(H1N1), A/swine/Serbia/3/2016(H1N1) and A/ swine/Serbia/6/2017(H1N1) originating from farms A, B, and K respectively, showed the highest nucleotide similarity with the sequences of HA genes of human origin viruses isolated during 2009/2010 infl uenza pandemic, indicating that the HA gene of these viruses may have been derived from H1N1pdm09 human strain [30].
Further, HA genes of H1N1 viruses from farms A and B showed high nucleotide similarity (98.25%). Considering the proximity of the farms (Fig. 1) and years of sampling, the possibility of virus transmission between farms cannot be excluded. On the other side, the HA gene sequence from farm K diverges from sequences obtained from farms A and B (93.86% and 94.27% of nucleotide identity respectively) and is located in a separate branch within H1pdm09 clade indicating a different origin (Figure 2).
Analysis of amino acid sequences of the HA gene of these strains showed that they have identical amino acids at the certain position which are specifi c for pandemic H1N1 strains. Thus, using H3 numbering scheme, the presence of glutamic acid (E) at the position 77, corresponding to the conserved spot in the HA amino acid sequence of H1N1pdm09 strains, was revealed in all three sequences [31,32]. Furthermore, the presence of aspartic acid (D) at positions 190 and 225 (instead of glycine) which are essential for receptor preference (high affi nity to α2-6 sialylated glycan receptors dominant in human respiratory epithelium), confi rms the pandemic origin of these strains [33,34]. In addition, the viruses possessed glycosylation site (NTT) at the position 278-280 in HA amino acid sequence, which is also characteristic of pandemic H1pdm09 strains [31] (Table 3).
Hemagglutinin gene of H1N1 viruses originating from farms F (A/swine/ Serbia/1/2017(H1N1)), G (A/swine/Serbia/5/2016(H1N1)), I (A/swine/ Serbia/7/2017(H1N1)) and J (A/swine/Serbia/8/2017(H1N1)), showed the highest similarity with the HA gene of Eurasian "avian-like" swine H1avN1 viruses [6,35]. Viruses from these farms showed the highest nucleotide identity, 93-95%, with European swine infl uenza viruses isolated in the period 2003-2011 ( Figure 2). The avian origin of the HA gene of these viruses was additionally confi rmed by the presence of PSIQSR motif in the cleavage site that is a trait of all low pathogenic avian infl uenza strains [36,37]. Moreover, the presence of aspartic acid at position 190 and glutamic acid at position 225 suggested the preference for human-type receptors (α2-6 sialylated glycan receptors) [38] ( Table 3). The phylogenetic tree of HA gene indicated the same origin of the viruses from farms F and G. However, the nucleotide identity of 93% resulted from the viral evolution during the period 2016-2017. Interestingly, the other two viruses, sharing in-between and to the viruses from farms F and G between 90.66% and 89.66% of nucleotides, had a different origin (Figure 2).  (Table 3). Also, the L226/ S228 combination determinates the virus preference for human-type receptors [38].
Analyses of the NA gene sequences revealed the circulation of two NA subtypes -N1 and N2. The N1 genes showed the highest similarity with the sequences of European swine strains isolated after 1998, which were characterized as the Eurasian avian-like H1avN1 lineage [35,6]. By the phylogenetic analysis of the N1 NA gene sequences, Serbian isolates were grouped into two clades. One clade was composed of Serbian isolates only (A/swine/Serbia/2/2017, A/swine/Serbia/3/2016, A/swine/ Serbia/7/2017, A/swine/Serbia/8/2017), and the other clade, along with Serbian isolates includes the IAV strains from Switzerland (human origin) and Denmark ( Figure 3).
In addition, the A/swine/Serbia/6/2017 isolate formed a separate group (Figure 3). The phylogenetic tree of NA genes derived from IAV isolates originating from farms A, B and J and F and G shows that they most likely have a common origin, indicating the introduction of animals from the same source to these farms. Based on the analysis of HA and NA genes of A/swine/Serbia/2/2017, A/swine/Serbia/3/2016 and A/ swine/Serbia/6/2017 isolates, who showed the highest similarity in the HA gene with the human A(H1N1)pdm09 lineages, and NA gene with the Eurasian avian-like H1avN1 swine lineages, we concluded that reassortment of human and swine IAVs occurred [39].
The sequence of NA genes of N2 virus originating from farm B belonged to the A/ swine/Gent/1/1984-like H3N2 lineage together with European swine viruses isolated between 2011 and 2013( Figure 3) [31].
In conclusion, research confi rmed the circulation of H1N1 and H3N2 IAV subtypes in the Serbian swine population. Although H1N2 subtype has not been detected, its presence cannot be excluded. Based on the phylogenetic analysis of HA genes, Serbian viruses belong to the Eurasian avian-like H1avN1, A(H1N1)pdm09 and A/ swine/Gent/1/1984-like H3N2 lineages, and by sequence analysis of NA genes, they were grouped in the Eurasian avian-like H1avN1 and A/swine/Gent/1/1984-like H3N2 lineages. Furthermore, the reassortment between human and swine viruses was confi rmed. Additional analyses of the internal gene cassette are needed to provide detailed insight into the molecular evolution, origin, pandemic potential and additional reassortment events. The detection of the reassorted genotype underlines the need of active surveillance implementation, aiming the public health protection.
of Swine Infl uenza A viruses. We also acknowledge the veterinarian colleagues who helped with the collection of the samples, and all scientists and technicians that helped with the vlaboratory work.
This work was supported and fi nanced by the Ministry of Education, Science and Technological Development (grant numbers TR31084, TR 37015, III46009) and the Institute of Veterinary Medicine of Serbia, Republic of Serbia.