Study area

The study was conducted in six different districts in the South Kivu province located in the Eastern part of the Democratic Republic of Congo, including Fizi, Kabare, Kalehe, Mwenga, Uvira, and Walungu. The collection points in each district are shown in Fig. 1. This vast region has an area of 66.814 Km2, located between longitudes 26° 10’ 30” and 29° 58’ East and latitudes 00’ 58” North and 4° 51’ 21” South.

As part of a larger survey to elucidate the presence of ASFV in apparently healthy [5], six districts (Kabare, Kalehe, Fizi, Mwenga, Uvira, Walungu) in DRC Congo were revisited after reports that ASFV-infected pigs had recovered from the disease, to assess the relationship of ASFV tolerance with RelA polymorphisms and cytokine levels.

Fig. 1

Map of the South Kivu province showing the study location and collection points of samples (Drawn with Arc-GIS)

The origin, PCR, and ELISA status of the tested samples

In total, 90 samples from the districts mentioned above were subjected to PCR analysis and ELISA tests to assess if the pigs had been or were infected with ASFV along with a clinical assessment, such as high fever, severe recumbency, cyanotic skin on ears, bleeding from bodily orifices, and difficulty of breathing [5]. Any PCR positive sample was considered as infected pigs while an ELISA positive test was considered as previous infection with non-active infection. In addition, samples were selected based on geographical location, and the criteria were that they should be ELISA or PCR positive to proceed in the study. From this number, 60 of the PCR-positive samples (with the highest DNA yield) were amplified for the RelA gene, of which 40 samples resulted in RelA sequences of sufficient quality.

Collection of samples

Whole blood (4 ml) was collected in EDTA tubes from pigs (all local breed) with clinical symptoms of ASF collected in December 2018 (high fever, severe recumbency, cyanotic skin on ears, bleeding from bodily orifices and difficulty of breathing) and from apparently healthy pigs collected in June 2017. Samples were stored at -20 °C.

Genomic DNA isolation

Porcine gDNA was extracted directly from 200 μl of whole blood using a DNeasy® Blood & Tissue kit (Qiagen, Hilden, Germany) according to manufacturer’s instructions. The gDNA concentrations were determined by spectrophotometry using Nanodrop (PCRmax Lamda, UK). The integrity of the DNA was measured on a 1% agarose gel stained with SYBR® Green I Nucleic Acid Gel Stain (10,000x, Invitrogen). The infection status of the animals was determined by polymerase chain reaction (PCR) and indirect enzyme-linked immunosorbent assay (ELISA), respectively, as reported in our previous work [5].

Primer design and PCR amplification of the porcine RelA gene

The genomic sequence of the RelA gene was retrieved from the GenBank database’s Sus scrofa isolate TJ Tabasco Breed Duroc chromosome 2 Sscrofa 11.1 (NC_010444.4). Primer-Blast tool was used to design 6 pairs of primers to amplify complete and overlapping exons and adjacent introns of the RelA gene (Table 1) while the expression level of RelA gene was assessed by primers described in Table 2. Primers were synthesized by Macrogen Europe (Amsterdam, Netherlands).

The porcine RelA gene is located on chromosome 2 between positions 6.594.869 to 6.602.684 with 7816 base pair in length. The genomic sequence includes 10 exons and 9 introns, and additional 5 ’and 3’ untranslated regions (Fig. 2). It comprises 1,662 nucleotides encoding a protein of 554 amino acids in length. In this study, five overlapping primer sets covering all 10 exons were designed to amplify and fully sequence all 10 exons of the RelA gene to identify potential polymorphisms in pigs. Primers were placed in conserved areas, facilitating amplification of any polymorphisms in the 10 exons. It was assumed that nonsynonymous single nucleotide polymorphisms (nsSNP) within the coding sequence (exons) would be of significant value as they could affect protein function.

Fig. 2

The genomic structure of the porcine RelA gene. Rectangles represent exons and are drawn on a scale. Black rectangles represent exons that compose the Rel-homology domain, while blue rectangles represent exons of the transactivation domain. Green rectangles show the untranslated regions found in the RelA cDNA. P1 represents the primers location for the sequenced exons to P6. Ex: exons, P: primer

For amplification of the target sequence, 1 μl of gDNA (50 ng/μl) was mixed with 10 pmol of each primer into 12.5 μl of PCR master mix (2xAccuPower Tap PCR Master Mix, Bioneer) containing 1U of TaqDNA polymerase (Bioneer, USA), 250 μM of dNTPs each, 1x reaction buffer with 1.5 mM MgCl2 and trace of tracking dye. The final reaction volume was brought to 25 μl with nuclease-free water.

The PCR cycling conditions used comprised an initial denaturation step of 5 min at 95 °C followed by 35 cycles comprising denaturation at 95 °C for 30 s, primer annealing at 55 °C for 15 s, and extension at 72 °C for 1 min. The final extension was done at 72 °C for 5 min. The PCR programs for all the primers were optimized and were identical for each target sequence. PCR products were resolved on a 1.2% agarose gel stained with SYBR® Green I Nucleic Acid Gel Stain (10,000x, Invitrogen), run in an electrophoresis chamber at 60 volts/cm for 40 min, and visualized using UV light. Amplicons of the expected size were purified using QIAquick PCR purification Kit (Qiagen, Germany) following the manufacturer’s recommendation and were sent for sequencing (Macrogen Europe Inc., Amsterdam, Netherlands).

Table 1 Primers for characterization of the porcine RelA geneTable 2 Primers for quantification of mRNA expressionPolymorphisms and prediction of their effect on protein function

The amplicon sequences were trimmed, and quality checks were done using CLC main Work Bench version 7.8.1 software (https://digitalinsights.qiagen.com). The nucleotide sequences and the translated amino acid sequences from surviving pigs were aligned and compared with sequences from symptomatic (sick) pigs by using the ClustalW (http://www.ebi.ac.uk/Tools/msa/clustalo/) tool in the MEGA 7 software. Potential polymorphism sites were detected by sequence comparisons using the DNAstar software (DNAstar Inc., Madison, WI, USA) and the DNA Sequence Polymorphism (DnaSP) version 6.12.03 (Universita de Barcelona).

In silico analyses were performed to predict the functional effects of amino acid changes using the software: Sorting Intolerant From Tolerant (SIFT) [30] and Polymorphism Phenotyping V2 (Polyphen v2) [31]. SIFT is a sequence homology-based prediction tool used to identify potential amino-acid substitutions that may affect biological functions through protein structural modifications. From SIFT analyses, scores between 0.00 and 0.05 were considered damaging, while scores beyond the threshold fixed at 0.05 were predicted to be neutral. PolyPhen-2, a physical and evolutionary comparative tool, was used to predict the effect of amino acid substitutions on protein structure and function. The scores were classified as probably damaging (≥ 0.85), possibly damaging (0.5–0.84), and benign (< 0.5). In this study, we assumed that any SNP with scores “probably damaging” and “possibly damaging” would affect protein functions.

Prediction of disease-related amino acid substitutions and their effect on protein stability

The web-based tools, MutPred (http://mutpred.mutdb.org/) and PredictSNP http://loschmidt.chemi.muni.cz/predictsnp are online server tools which integrate genetic and molecular data to predict the detrimental effect of amino acid substitutions in a mutant protein [32, 33]. The outputs from these two tools were combined to improve the prediction accuracy. For MutPred, scores with g-value (probability for pathogenic amino acid substitutions) > 0.50 and p-value < 0.05 were considered actionable hypotheses (a given amino acid change with pathogenic effect). In contrast, the scores with g-value > 0.70 and p-value < 0.05 are referred to as confident hypotheses (with no pathogenic effect). However, for Predict SNP, scores with p-values (probability for deleterious of given amino acid substitution) <-1 to 0: neutral; p-value: 0 to + 1: deleterious.

In addition, MUPro [34] and I-Mutant 3.0 (http://gpcr2.biocomp.unibo.it/cgi/predictors/ are support vector machine-based tools that were used for predicting the effect of nonsynonymous amino acids substitutions on protein stability. MUPro predicts the energy change value and yields a confidence score between − 1 and 1 to be used for calculating the confidence of the prediction. Scores < 0 suggest that the amino acid change decreases protein stability, whereas scores > 0 indicate increased protein stability [35]. Moreover, the outputs of the I-Mutant prediction method of protein stability changes are based on the value of free energy change: largely destabilizing ( < − 0.5 Kcal mol − 1), largely stabilizing (> 0.5 Kcal mol − 1), or weakly stabilizing or destabilizing (− 0.5 ≤ Delta Delta Energy (DDG) ≤ 0.5 Kcal mol − 1).

Relative quantification of IFNα, IL10, and TNF-α gene expression by real-time PC

RNA was extracted from blood samples using PureLink™ RNA Mini Kit (Thermo Fisher, Ambion Life Technology, California) following the manufacturer’s recommendations. The RNA content and purity were estimated using the Nanodrop (PCRmax Lambda) spectrophotometer and the 260/280 nm ratio, respectively. cDNA was done using RevertAid First Strand cDNA Synthesis Kit (Fermentas, Thermo Scientific, #1622) following the manufacture’s protocol. Briefly, 10 μg of the total RNA and 1 μl of oligo (dT) primer were added to 8 μl of RevertAid master mix, and the volume was brought up to 20 μl. The reactions were incubated for 5 min at 25 °C followed by 60 min at 42 °C. The reaction was heated at 70 °C for 5 min in a thermocycler machine (ProFlex, PCR system, Applied Biosystem). The obtained cDNA samples were stored at − 20 °C until further use. For quantitative PCR analysis, expressions of all the genes (IL-10, IFN-α, and TNF-α) were quantified using specific primers, as presented in Table 3. Values were normalized to 18 S rRNA, a housekeeping gene, which is a common choice as a reference gene (Table 2).

Table 3 Primer sequences used in this study to quantify the mRNA expression of the selected cytokines

The specificity of the qPCR was assessed by the melting curves generated after amplification. All the samples were run in triplicates. The relative expression of each sample was calculated using the method suggested by Livak and Schmittger [38]. The qPCR mix comprised 10 μl of 2X Luna Universal qPCR Master Mix (New England, BioLabs Inc.), 2 μl of cDNA, and 0.5 μl containing 10 pmol of each forward and reverse primer were added; the volume was topped up to 25 μl with nuclease-free water. Amplification of ASFV was accomplished in a LightCycler ® 96 (LifeScience, Roche) with the following conditions: 50 °C for 2 min, one cycle (uracil N-deglycosylase digest); 95 °C for 1 min, 95 °C for 15 s, 62 °C for 60 s, 40 cycles. For specificity of the PCR, information on melting curves was collected continuously from 65 °C to 95 °C.

Quantification of IL-10, IFN-α, and TNF-α levels in serum samples by enzyme-linked immunosorbent assay (ELISA)

The level of IFN-α, TNF-α, and IL-10 cytokines in symptomatic, surviving, and healthy pigs was evaluated by quantitative sandwich-type enzyme-linked immunosorbent assay (ELISA) kits (Eagle Biosciences, Inc., Nashua NH, USA) from heparin-free serum samples following the manufacturer’s instructions. Results were expressed as values in pg/ml for all tested cytokines for each naturally infected pig and the control group; each sample (serum) was tested neat in duplicate. Healthy pigs were collected from pig farms which had never reported ASFV infection in the Walungu district and, which were ASFV negative by both ELISA and PCR tests.

Statistical analysis

The relative expression of the target genes for each sample was calculated as described in the previous section and presented as fold changes. One-way analysis of variance (ANOVA) was used to compare the relative concentrations of cytokine proteins and mRNA expression of IFN-α, TNF-α, and IL-10 among groups. Correlation between the mRNA level for each cytokine and the expression of RelA mRNA was assessed by the Pearson correlation coefficient test. All statistical analysis was performed using SAS version 9.4 software (SAS Institute Inc., Cary, USA), and a confidence level of 95% was used in all tests to determine the statistical significance between groups. A p-value of < 0.05 was considered significant. Visualization and graphical presentations were done using GraphPad Prism 7 software (Diego, CA, USA).