Identification of African Swine Fever Virus Transcription within Peripheral Blood Mononuclear Cells of Acutely Infected Pigs

African swine fever virus (ASFV) has become widespread in Europe, Asia and elsewhere, thereby causing extensive economic losses. The viral genome includes nearly 200 genes, but their expression within infected pigs has not been well characterized previously. In this study, four pigs were infected with a genotype II strain (ASFV POL/2015/Podlaskie); blood samples were collected before inoculation and at both 3 and 6 days later. During this period, a range of clinical signs of infection became apparent in the pigs. From the blood, peripheral blood mononuclear cells (PBMCs) were isolated. The transcription of the ASFV genes was determined using RNAseq on poly(A)+ mRNAs isolated from these cells. Only very low levels of virus transcription were detected in the PBMCs at 3 days post-inoculation (dpi) but, at 6 dpi, extensive transcription was apparent. This was co-incident with a large increase in the level of ASFV DNA within these cells. The pattern of the virus gene expression was very reproducible between the individual pigs. Many highly expressed genes have undefined roles. Surprisingly, some genes with key roles in virus replication were expressed at only low levels. As the functions of individual genes are identified, information about their expression becomes important for understanding their contribution to virus biology.


Introduction
African swine fever virus (ASFV) is the sole member of the Asfarviridae family. The virus has a large, linear dsDNA genome (ca. 170-190 kbp, depending on the strain) that includes nearly 200 genes (reviewed in [1]). This virus infects domestic pigs together with a range of wildlife species (family Suidae), including bush pigs and warthogs in Africa, while wild boar are important hosts in Europe and Asia [2]. In addition, ASFV can replicate within soft ticks (genus Ornithodoros) and is unique in being the only known DNA arbovirus. A sylvatic cycle involving replication in soft ticks and warthogs occurs in Africa [2]; the infection is largely asymptomatic in the warthogs but becomes apparent when domestic pigs become involved. Outside of Africa, the transmission of the virus is believed to occur mainly by the direct or indirect contact between infected pigs, generally without the involvement of soft ticks; however, some aspects of its transmission are poorly understood [3].
Many (at least 24) different genotypes of the virus exist in Africa; these are distinguished based on the sequence of the VP72 gene [4][5][6][7][8]. In 1957 and 1960, excursions of a genotype I ASFV from Africa into Europe (Portugal) occurred, and the virus (e.g., the Ba71 strain) was present in the Iberian peninsula until the 1990s [9]. In 2007, a genotype II virus entered into Georgia (in the Caucasus region), and, subsequently, African swine fever (ASF) has become widespread within neighboring countries, such as Russia and those in Eastern Europe. It has also spread into Western Europe, including Belgium and, during 2020, Germany [10]. Furthermore, in 2018, essentially the same virus was reported from China, the world's largest pig producer [10,11], and quickly moved into many countries in the vicinity (e.g., Vietnam, Korea and Cambodia) and to the Philippines. In 2021, the virus has been introduced into pigs in the Dominican Republic and Haiti [10]; thus, this virus is a global concern.
Infection with ASFV can result in very high levels of case fatality and, thus, has major economic importance. There are no commercially available approved vaccines or antiviral agents to control the disease, so the control measures rely on the culling of infected animals, restrictions on animal movement and high biosecurity [12,13].
Highly virulent isolates of ASFV often cause a peracute to acute disease progression with high fever (>41 • C) and a range of clinical signs, including anorexia and lethargy, which occur within a few days of infection [14][15][16]. ASFV replicates within the cytoplasm of infected cells and encodes its own RNA polymerase and transcription factors. Genes can be expressed at different stages of the virus life cycle, e.g., early (prior to DNA replication) or late (following DNA replication). The open reading frames (ORFs) are closely spaced on the viral genome and are transcribed from the two different strands of the DNA [17,18]. The mRNA transcripts are capped at their 5 -termini and are post-transcriptionally modified at their 3 -termini to generate a poly(A) tail by a virus encoded capping enzyme and poly(A) polymerase, respectively (reviewed in [1]). An initial analysis of the virus gene expression, using the total RNA extracted from the whole blood of pigs infected with the ASFV Georgia 2007/1 virus, has been performed [19] but showed very variable levels of gene expression between animals. Furthermore, a detailed analysis of the ASFV transcription has been undertaken using the cell culture adapted Ba71V strain (genotype I) of ASFV within Vero cells (derived from the African green monkey kidney) [18]. Analyses of the transcription within porcine peripheral blood macrophages and porcine pulmonary alveolar macrophages (PPAM), infected in vitro with isolates of ASFV, have also been reported previously [20,21].
In the pig, the initial sites of virus replication during a natural infection include the pharyngeal tonsils, and the secondary sites include the spleen, lymph nodes and liver (reviewed in [1]). More specifically, the virus primarily replicates in the cells of the monocyte macrophage lineage [22]. In a virus-infected animal, it is not possible to achieve the synchronous infection of all the cells within the animal (c.f. all the cells within a flask). However, in order to follow the time course of infection, in this study, we have examined the expression of virus genes from within the peripheral blood mononuclear cell (PBMC) population (including lymphocytes, monocytes and macrophages) harvested from individual animals at 3 and 6 days post-inoculation (dpi) with ASFV/POL/2015/Podlaskie. This has allowed the transcription of the ASFV genes to be assessed within key target cells of the natural host animal during the progression of individual pigs from being uninfected to being diseased. As the functions of the ASFV genes become known, then information about their expression should assist in the understanding of their contribution to virus biology.

Pigs
Four male pigs, eight weeks of age, were included in this study. The pigs were obtained from a conventional Spanish swine herd (Landrace × Large White). On arrival at the research facility, one week before the start of the experiment, all pigs were found to be healthy by veterinary inspection. Water and a commercial diet for weaned pigs were provided ad libitum.
Animal care and maintenance, experimental procedures and euthanasia were conducted in accordance with EU legislation on animal experimentation (EU Directive 2010/63/EU). The experiment was performed within high containment facilities at the Centre de Recerca en Sanitat Animal (IRTA-CReSA, Barcelona, Spain).

Challenge Virus
For the experimental infection, ASFV was isolated from spleen material obtained from a dead wild boar in 2015 in the Podlaskie voivodeship (province), Poland, as previously described [16]. This virus is designated here as ASFV POL/2015/Podlaskie, the genome has been sequenced [23] and it is very closely related to the updated ASFV Georgia_2007/1 sequence (GenBank Accession no. FR682468.2, [24]). Briefly, clarified spleen suspension was passaged twice in PPAM and the titer of the second passage was then determined by end-point titration in PPAM [16]. For intranasal inoculation of pigs, the second passage virus was diluted in phosphate buffered saline (PBS) to a final concentration of 4 log 10 50% tissue culture infectious doses (TCID 50 ) per 2 mL, as used previously [16]. Back titration of the inoculum was carried out in PPAM to confirm the administered dose.

Study Design
Upon arrival at the research facility, the four pigs were housed together in a high containment stable unit (BSL-3). After an acclimatization period of one week, the pigs were inoculated intranasally with 2 mL virus suspension containing 4 log 10 TCID 50 of the ASFV POL/2015/Podlaskie (see [16]). The time course of the infection in the pigs was followed from their clinical signs, rectal temperatures and using laboratory analyses as described in the following sections.

Clinical Examination and Euthanasia
Clinical scores and rectal temperatures were recorded from individual pigs on each day. A total clinical score was calculated per day based on a modified system from that described previously [16], omitting food intake as it was available ad libitum; also, the rectal temperatures were recorded separately and were not used as part of the clinical score. The total clinical scores were calculated as the sum of scores given in eight categories (see Table 1). This allowed a maximum total clinical score of 31.
The pigs were euthanized after they reached the humane end-points set in the study, which occurred at 6 days post-infection (dpi), by intravascular injection of Pentobarbital following deep anesthesia.

Sampling from the Inoculated Pigs
EDTA-stabilized blood (EDTA blood) samples were collected prior to inoculation at 0 dpi, at 3 dpi and at 6 dpi, just prior to euthanasia. EDTA blood samples were processed directly for isolation of PBMCs (see below). Paralysis or convulsions This is modified from the system described previously [16] since feed was available ad libitum and rectal temperatures were recorded separately.

PBMC Isolation and Processing
Using fresh EDTA blood samples (4 mL) from each pig, PBMCs were isolated using the Histopaque ® system (Sigma-Aldrich, St. Louis, MO, USA). The PBMC fraction samples were lysed by addition of Trizol TM Reagent (ThermoFisher Scientific, Waltham, MA, USA), and the samples were stored frozen at −80 • C until further processing.

RNA Purification
Total RNA was extracted from the PBMCs in Trizol TM Reagent, (ThermoFisher Scientific) using the Direct-zol TM RNA MiniPrep kit (Zymo Research, Irvine, CA, USA). This purification system includes a DNAse I digestion to remove host and viral DNA. Analysis of the RNA transcripts was performed using poly(A)+ selected mRNAs. These samples include both viral and host mRNAs, but the selection removes most ribosomal RNA, which were then sequenced (following reverse transcription using random primers, second strand synthesis, adaptor ligation and PCR-amplification) by BGI Europe Genome Center (Copenhagen, Denmark) (termed RNA-T on DNBseq with ca. 40 million reads per sample).

ASFV DNA Detection by Quantitative Real-Time Polymerase Chain Reaction (qPCR)
Following chloroform-mediated phase separation of the remaining volume of PBMCs lysed in Trizol TM Reagent (ThermoFisher Scientific), DNA purification was performed on the interphase material using the MagNA Pure 96 system (Roche, Basel, Switzerland). The presence of ASFV DNA was determined by qPCR assays employing 45 cycles [16,25]. Results are presented as viral genome copy numbers/per mL EDTA blood calculated by reference to a standard curve based on a 10-fold dilution series of a pVP72 plasmid (prepared from a cloned PCR product amplified with primers dCCCGGTCC-GAAGCGCGCTTTCCCGGGATGGCATCAGGAGGAGCTTTTTG and dCGAAAGCGGC-CGCGGGATCGACTAGTCTATTAGGTACTGTAACGCAGCAC; the sequences in italics match exactly to the ASFV p72 coding sequence from GenBank Accession no. MH681419.1) using as the template DNA extracted from a spleen sample of a pig infected with ASFV POL/2015/Podlaskie [16]).

Data Analysis
Mapping of Sequence Reads to the Pig Genome and to the ASFV Genome Sequence reads (27-47 million per sample) were initially mapped using STAR v. 2.7.0 [26] to the USMARCv1.0 pig genome assembly (Accession no. PRJNA392765), which was derived from a male pig within a population that was approximately one-half Landrace, one-quarter Duroc and one-quarter Yorkshire (see [27]). The analysis of changes in expression of the pig genes within these infected pigs will be reported separately. The unmapped reads were then mapped to the updated ASFV Georgia_2007/1 genome (Gen-Bank Accession no. FR682468.2) using BWA v.0.7.10 [28]; note that one correction to the annotation for the D205R gene was made (using the coding sequence as nt 138482 to 139234). Reads were mapped to individual virus genes using featureCounts, which is a part of the Subread package, v.2.0.3 [29]. The counts were then standardized, when indicated, to take account of the library size and the length of each gene from each sample (gene length corrected trimmed mean of M-values (GeTMM)) [30].
The reads that did not map to the pig genome were also examined using Kaiju [31] to determine the origin of the reads. The parameters used were: minimum match length (11), minimum match score (75), allowed mismatches (5) and maximum E-value (0.01).

Course of Infection in the Inoculated Pigs
Following intranasal inoculation on day 0, three out of the four pigs (numbers 10, 11 and 12), at 4 dpi, had high fever (rectal temperature above 41 • C; see Figure 1A). Furthermore, at 5 and 6 dpi, all four pigs had high fever. Clinical signs of infection also became apparent and included depression, anorexia, mildly labored breathing, hyperemia of the skin and cyanosis on the ears and distal limbs, plus blood in feces (pig 10). At 6 dpi, all the inoculated pigs were euthanized since pigs 10, 11 and 12 had reached the predetermined humane endpoint and the remaining animal, pig 9, was euthanized to avoid having a solitary animal. The clinical scores for the four inoculated pigs through the course of the infection are depicted in Figure 1B. All four pigs showed at least some clinical signs of disease, in addition to the elevated temperature by 5 dpi, which had become much more apparent in pigs 10, 11 and 12 at 6 dpi, just prior to euthanasia. were also scored (B). Blood samples were collected at 0 dpi (prior to inoculation), at 3 dpi and at 6 dpi; the animals were euthanized at 6 dpi. PBMCs were purified from the blood samples (4 mL). The presence of ASFV DNA in the PBMCs (C) was quantified by qPCR and values converted to gene copy numbers/mL blood by reference to a standard curve. Levels below 10 3 ASFV genomes/mL (indicated by dashed line) were outside of the linear range of the assay.

Virus Derived RNA Transcript Analysis from PBMCs of ASFV-Inoculated Pigs
Blood samples (12 in total) were collected separately from the four pigs on day 0 (prior to inoculation), at 3 dpi and 6 dpi (on the day of euthanasia). The PBMC fraction was isolated from the EDTA blood samples, and the level of ASFV DNA in the nucleic acids isolated from the aliquots of the PBMCs was determined by qPCR ( Figure 1C). Only low levels of ASFV DNA were detectable in the PBMCs at 3 dpi, but much higher levels (>1000-fold) were present at 6 dpi in all four pigs ( Figure 1C). The total RNA was also extracted from the PBMC samples and freed of most of the DNA from the host and the virus. The poly(A)+ RNA (mRNAs) was then selected, reverse transcribed and sequenced (as described in Materials and Methods). The reads (ca. 27-47 million per sample) were initially mapped to the pig genome (resulting in 81-92% of the reads being uniquely mapped, see Table 2), and the unmapped reads were then mapped to the ASFV Georgia_2007/1 sequence (see Table 2). As expected, no reads mapping to the ASFV genome were present in any of the samples collected on day 0, prior to inoculation. At 3 dpi, the PBMCs from pigs 10, 11 and 12 each only generated between 600-850 reads (less than 0.0025% of the total reads) mapping to the ASFV genome, with no ASFV-derived reads detected in pig 9 (Table 2). However, in contrast, at 6 dpi, between 871,681 and 1,945,879 reads (ca. 2-4% of the total reads) from the PBMCs mapped to the ASFV genome from each of the four pigs (Table 2); the lowest numbers of ASFV reads were found in pigs 9 and 10. For pigs 11 and 12, about 40% of the reads that did not map to the pig genome did map to the ASFV genome. Thus, in parallel with the appearance of clinical signs between 3 and 6 dpi and the large increase in the levels of ASFV DNA in the PBMCs of the inoculated pigs (see Figure 1), a huge increase (>1000-fold) in the level of ASFV RNA transcripts within the PBMCs was apparent during this time period.

Transcription of Individual ASFV Genes
In general, there was a very good agreement between the pattern of expression of the individual ASFV genes obtained for the PBMCs from each of the different pigs, although the actual levels varied to some extent (see Supplementary Table S1). In particular, the highest levels of virus gene expression were observed in pigs 11 and 12, and the levels of ASFV transcripts present in these cells from pigs 9 and 10 were, on average, about 50% lower. The most highly expressed genes were C312R, CP204L, MGF 100-1L and A151R (see Table 3), while some genes were expressed at a much lower level, i.e., less than 1% of these most highly expressed genes (see Table 4 and Supplementary Table S1). Interestingly, all the identified ASFV coding genes, with ORFs > 180 nt in length, were expressed to some extent (see Figure 2 and Supplementary Tables S1 and S2). Note that there are some very short genes, designated pNG1-7, that were recently identified [18] from the Ba71V genome that were not specifically included here, but several of them correspond to features annotated in the ASFV Georgia_2007/1 genome as ASFV G ACD 00xx0 (see [18]). Some of these genes (e.g., ASFV G ACD 00350 and ASFV G ACD 00600) were expressed (see Supplementary  Table S1). The gene designated as pNG4 (in the Ba71V genome) does not seem to have a corresponding gene in the ASFV Georgia_2007/1 genome. Some annotated features within the ASFV Georgia/2007 genome were present within the reads derived from other annotated gene transcripts (e.g., "polyC regions" (nt 14225 to 14237 (within the MGF 110-10-L-MGF110-14L fusion ORF) and nt 15666 to 15682 (within the MGF 110-13Lb ORF) plus a "polyG region" (nt 19993 to 20008, within the ASFV G ACD 00350 ORF)). However, another "polyG region" (nt 17624 to 17632) was expressed at quite high levels (even in the list of total reads not adjusted for the length of the transcript, see Supplementary Table S1). This sequence is present within a very short open reading frame (encoding a 15-residue peptide, MFDLSSILIRGGGPY) that was not annotated previously. The two genes flanking this polyG region were detected at lower levels (see Supplementary Table S1) and, hence, it does not seem possible to account for these reads as read-through from adjacent genes. Three so-called "hypothetical" genes (nt 19411 to 19506, nt 51223 to 51337 and nt 182044 to 182151, see Supplementary Table S1) were detected with many reads. However, the genes annotated as "ASFV G ACD 01760 no indication" (nt 175922 to 176006) and "DP63R no indication" (nt 178506 to 178652) were present within the transcripts from the I177L and MGF 360-16R genes, respectively, and are not listed separately.  CP312R  128277  129200  66398  78462  138910  144010  141460  CP204L  125783  126367  59137  50204  128247  133890  131069  MGF 100-1L  180479  180904  50566  30183  108077  98680  103379  A151R  49652  50107  45198  45603  96565  89945  93255  K205R  64174  64791  29906  34834  64872  65805  65339  I73R  173088  173306  23887  24501  59167  54809  56988  MGF 360-15R  50346  51215  24275  23647  47350  49844  48597  MGF 110-5L-6L  9490  10107  18740  20840  42820  46048  44434  A240L  48633  49343  18353  15318  41882  45882  43882  MGF 110-7L  10314  10727  20780  20790  43090  42632  42861  E165R  167468  167965  17095  20759  40914  41374  41144  MGF 100-3L  181269  181577  15716  18157  36029  42209  39119  F334L  56956  57960  17153  17920  38103  39044  38574  MGF 110-3L  8239  8613  16201  22633  35842  34045  34944  285L  11042  11326  17175  14716  34762  32868  33815  I215L  174794  175432  16341  14199  32361  32736  32549  MGF 505-3R  35760  36602  13836  11627  25567  30317  27942  DP96R  185339  185629  12107  10550  27516  28141  27829  ASFV G ACD  00600  48000  48152  9902  7709  22675  22563  22619   K196R  65113  65703  9812  14044  22163  22183  22173 1 : Average number of reads from pigs 11 and 12 at 6 dpi were used to place genes into order. The values given are the total number of sequence reads mapped per annotated region (not standardized for gene length or total number of reads) in the updated ASFV Georgia_2007/1 genome sequence (GenBank Acc. No. FR682468.2). Note that the start and end of the CDS are indicated by their position in the genome independently of the orientation of the gene. Potentially, some sequence reads could be derived from residual ASFV DNA in the samples, even after the DNAseI treatment and purification of the poly(A)+ RNA. However, it appears that the number of such reads should be very low since a number of genes are only included in a small number of reads (<50, see Table 4); this should be compared to >100,000 reads for some of the highly expressed genes ( Table 3). The reads derived from genomic DNA should be derived similarly from all the regions of the genome.
We have also "standardized" the number of reads to take into account the size differences between the libraries generated prior to sequencing and the lengths of the different genes. This process had little overall effect on the pattern of highly expressed genes (see Supplementary Table S2). The most highly expressed genes, based on the total reads as listed in Table 3, are also high on the list of the ASFV genes expressed that is shown in Supplementary Table S2, following standardization, e.g., CP312R, CP204L, MGF 100-1L, A151R, K205R and I73R (see also Table 5). The distribution of RNA transcripts across the genome, derived from the two different strands of the genomic DNA, is shown in Figure 2. Most of the highly expressed transcripts are derived from genes that are well-separated across the genome, and the highly expressed transcripts are copied from each of the strands of the genome. As may be expected, the different strands of the genome are either read in one direction or the other; this prevents the production of mRNAs that are complementary to each other and would form dsRNAs. Two genes, A151R and MGF360-15R, which are highly expressed (see Table 3) are adjacent to each other and transcribed in the same direction (see [17]). However, it is noteworthy that, after the standardization of the number of reads, to take account of the gene length, the apparent level of MGF360-15R expression is less markedly high (Table 5) and does not feature among the most highly expressed genes shown in Figure 2.
The small number of ASFV reads detected in the pigs at 3 dpi (Supplementary Table S1) precludes the accurate quantification of their relative gene expression. However, it is apparent that the few reads that were observed at 3 dpi corresponded to transcripts that were highly expressed at 6 dpi (e.g., see Table 5). It seems most likely that the viral RNA reads observed in the PBMCs at 3 dpi represent the infection of a very small proportion of the cells, consistent with the low levels of ASFV DNA present within the PBMCs at this stage of the infection. It is not possible to differentiate "early" and "late" transcription within the pigs (c.f. in cell culture, [18]) as it cannot be expected that a synchronous state of infection can be achieved. In the RNAseq analysis performed here, the use of alternative transcription start sites has not yet been explored. 1 : The total numbers of reads per gene were standardized according to the length of the open reading frame and for the size of the library. Only genes with a coding sequence > 180 nt (encoding > 60 amino acids) were included. The genes are listed according to the mean value for the reads from pigs 11 and 12 at 6 dpi. For comparison, the number of reads generated from the PBMCs at 3 dpi are also indicated for pigs 11 and 12 (for pigs 9 and 10, there were zero (0) or between 0 and 8 reads for each of these genes on this day, respectively). A complete listing of the ASFV genes expressed at 3 and 6 dpi is given in the Supplementary Tables S1 and S2. Gene product properties are from Dixon et al. [17] and Jaing et al. [19].

Discussion
The intranasal inoculation of pigs with ASFV is an efficient and consistent means of initiating infection (as seen previously [16]). Each of the inoculated pigs became infected and followed a similar course of disease. Pig 9 was slightly delayed in showing clinical signs, and both pigs 9 and 10 had slightly lower levels of the ASFV derived transcripts within their PBMCs, but it is expected that there will be some differences between animals.
Overall, we have obtained very consistent results among the four pigs infected with the ASFV POL/2015/Podlaskie strain for the expression of each ASFV gene within the PBMCs; this cell population includes the major target cell types (monocytes and macrophages) for this virus. The virus strain used here is very similar to the other closely related genotype II viruses currently circulating in Europe and Asia.
A complete listing of the standardized gene expression data is provided in Supplementary Table S2. In studies described previously by Jaing et al. [19], three pigs were infected (by intranasal instillation) with the ASFV Georgia 2007/1 strain using 10 4 TCID 50 . The pigs were euthanized on days 7-10 after infection; each of the pigs were shown to have ASFV DNA in the blood from 7 dpi and displayed clinical signs of infection from that time. Total RNA was extracted from whole blood and used for the transcriptomic analysis; only about 0.1% of the reads mapped to the ASFV genome. This is a much lower level than observed here (up to 4.2%, see Table 1); presumably, the much higher proportion of ASFV derived reads that we obtained is largely due to the selection of poly(A)+ mRNAs from the PBMC fraction prior to the sequencing (thus removing many reads derived from other RNAs (e.g., ribosomal RNA). About 1-2 million reads that mapped to the ASFV genome were analyzed here from each of the inoculated pigs at 6 dpi. Another advantage obtained from selecting the poly(A)+ mRNA is that reads derived from residual ASFV DNA should be further diminished beyond that achieved by DNAseI treatment alone, which may not be completely effective [32]. However, some reads may be derived from residual ASFV DNA (there are only short regions of the genome that are not included in any reads), along with other incompletely digested host DNA, e.g., pig mitochondrial DNA. In addition, it is worth noting that <1% of the reads (about 250,000 per sample) do not map to the pig genome nor to ASFV but map to bacterial genome sequences, as determined using Kaiju [31]. It seems likely that this results from bacteria being engulfed by the porcine macrophages and then residual DNA being present within these cells.
A key feature of the data presented here is the close correspondence between the results observed for each of the four pigs. Thus, the most highly expressed ASFV genes observed in one pig were also highly expressed in the other three pigs (Table 3). Indeed, the relative order of expression of the top 20 genes was very similar in each of the four pigs. Furthermore, genes that were expressed in one pig at a low level were also expressed at a low level in the other animals (Table 4). This consistency in the gene expression contrasts with the much more variable results reported previously by Jaing et al. [19]. For example, surprisingly, for nine of the seventeen genes indicated in their study as being highly expressed overall, in one of the three pigs examined, apparently no reads corresponding to these highly expressed genes were detected. Indeed, for the most abundantly expressed transcript overall, encoding MGF-360-15R (an inhibitor of interferon-β induction, [33]), the fragments per kilobase per million mapped fragments (FPKM) values obtained for the three pigs were: 1,190,000; 0 (zero) and 19,979, respectively, showing great variation. This gene was the sixth or seventh most highly expressed gene in each of the four pigs studied here (see Table 3), with the total reads ranging from 23,647 to 49,844 (see Supplementary Table S1). Following the standardization (to account for differences in library size and for the lengths of the genes), this gene was fifteenth in the list of highly expressed genes (Table 5) and, thus, it does not feature among the most highly expressed genes (see Figure 2).
The highly expressed genes of ASFV are mainly located in different regions across the genome and are derived from each strand. There can be a concern that, if the transcription termination signals are not 100% efficient, then read-through into the adjacent genes could affect the apparent level of transcription of the genes adjacent to the highly expressed genes. For the most highly expressed genes that we have analyzed, this could only be an issue for the A151R gene (fourth most highly expressed) and the adjacent MGF360-15R gene (sixth or seventh most highly expressed, see Table 3) since these genes are both transcribed in the same direction. However, after the normalization of the gene reads to take into account the length of the ORF, the expression of the MGF360-15R gene does not appear to be very high (see Table 5 and Figure 2). It seems unlikely that the reads corresponding to the MGF 360-15R gene were greatly affected by the read-through from the higher number of reads derived from the A151R gene unless the transcription termination process was very inefficient. Furthermore, the studies by Cackett et al. [18] indicated a good correlation beyond the transcription measured by RNA-seq (as used here) and the 5 -end cap analysis gene expression sequencing (CAGE-seq). The latter methodology is independent of any read-through transcription. Finally, there was a good correspondence between the genes found to be highly expressed in the pigs, as shown here (see Tables 3 and 5 and Supplementary Table S1), and those found to be highly expressed during both the early (at 5 h post-infection, prior to DNA replication) and late (at 16 h, after DNA replication) stages of infection in Ba71V-infected Vero cells [18]. Thus, the genes CP312R, CP204L, A151R, K205R and I73R were all found to be highly expressed in both studies. There was also good overall correspondence to the results from Jaing et al. [19] within the pigs, although the higher degree of variability of the signal (as mentioned above) makes for less precise quantification. Many of the ASFV genes are poorly characterized [17], and this applies to the highly expressed genes as well as those expressed at a lower level.
As the functions of the various ASFV genes become known, information about the expression of the genes in different cell types will be important for understanding the biology of the virus. Interestingly, most of the major components of ASF virions (as described previously [34]) are not translated from among the most highly expressed genes. Only the dUTPase component of the virion (from the E165R gene) is produced from among the top 20 expressed transcripts (see Table 3). The genes encoding the major capsid protein (p72 expressed from B646L), the outer envelope protein CD2v (from E402R) plus the two polyproteins pp220 (from CP2475L) and pp65 (from CP530R) do not appear in the list of the top 20 expressed genes. It is noteworthy that the S273R gene, which expresses a protease that processes the two polyproteins (pp220 and pp65) to major components of the virion [34], is expressed at very low levels (Table 4). Clearly, the level of protein expression does not only reflect the level of the mRNA; it will also depend on the stability of the proteins and the translational efficiency of the individual mRNAs. It can be expected that proteins with catalytic functions, e.g., the ASFV protease (S273R), may be required at a lower level than the structural protein precursor substrates.
Overall, it seems that all open reading frames encoding products longer than 60 amino acids (i.e., >180 nt, as annotated in Accession no. FR682468.2) are expressed within the pigs. It could have been possible that some ASFV genes were only expressed in soft ticks as part of its sylvatic cycle, but this does not appear to be the case. However, for the genes that are expressed at very low levels (see Table 3), it is formally possible that these reads are due to the presence of residual ASFV DNA. Clearly, there could be quantitative differences in the pattern of gene expression from the virus when it replicates in the ticks compared to that observed in pigs.
It is apparent that the ASFV must have replicated somewhere in the pigs during the time period from inoculation until after 3 dpi since the virus must have spread to the PBMCs to enable their infection. Following the intranasal infection, it is believed that the primary ASFV replication occurs within the tonsil and regional lymph nodes before being spread, through the lymphatic system and the blood, to other replication sites (e.g., the spleen and liver (see [22]). Further studies should focus on these primary sites of infection and also other sites of subsequent infection. Furthermore, the host responses to the ASFV infection can also be addressed using the PBMCs (as here) and other cells that become infected.