Combined effects of intrinsic host gene (TRAPPC9) and extrinsic nutrient (folate) on resistance against S. aureus induced bovine mastitis

Background: Drug-resistance and immunological escape of Staphylococcus aureus and its “superbug”, methicillin-resistant S. aureus (MRSA), have become one of main causes of bacterial infection in both human and animals. In dairy cattle, elimination of bovine mastitis induced by S. aureus is of importance because S. aureus-infected cows normally are culled passively. Methods: Here, we investigated the benecial effects of bovine tracking protein particle complex 9 (TRAPPC9) gene and folic acid supplementation in the control of mastitis induced by S. aureus or MRSA by a series of in vivo and in vitro experiments. Results: The data showed that the genetic mutations and DNA methylation of TRAPPC9 were highly linked with the mastitis resistance of dairy cows. Additionally, knockdown of bovine TRAPPC9 was signicantly involved in the mRNA expression levels of interleukin’s genes (increased IL-1β and IL-6), and down-regulated the protein level of NF-κB-P65 in the mastitis cell model induced by MRSA. Meanwhile, dose-dependent folic acid addition can inhibit the invasion of MRSA into Mac-T cells and improve TRAPPC9 expression in dairy cows. Conclusions: Altogether, our data suggest that an appropriate dose of folic acid can signicantly reduce the inammation caused by MRSA partially through TRAPPC9 mediated NF-κB pathway. These ndings provide new insights to control the drug-resistant pathogens and to restrict the overuse of antibiotics through combined effects of the intrinsic host gene and extrinsic nutrient. R1 Region-Reverse: GGGACACCGCTGATCGTTTAAACAATCCACTCCAATTACTATTACC, and R1 Region-Sequencing: TTTTAAAGGAAAGGAAAGTG. R2 Region-Forward: GAGTTTGGAGTGGTTTTTTTAGG, R2 Region-Reverse: GGGACACCGCTGATCGTTTAAAATAAATTCCCCTTTTACTATC, and R2 Region-Sequencing: GTTTGGAGTGGTTTTTTTAGG.


Introduction
Hosts possess innate and adaptive immune systems that are tailored to counter pathogen challenges effectively. Under general bacterial infections, the nuclear factor-κB (NF-κB) pathway can be activated in response to stimuli [1][2][3]. However, Staphylococcus aureus, one of the rst well-characterized microbes, can modulate or avoid host immune systems and acquire strong resistance to most antibiotics and drugs [4]. Moreover, the "superbug" strain, methicillin-resistant S. aureus (MRSA) is still the common causes of infections and mortality in hospitals [5,6]. Aside from hospital-associated MRSA and communityacquired MRSA, livestock-associated MRSA infection has been rising in frequency since the 1970s [7][8][9][10]. The earliest isolation of livestock-associated MRSA strain was reported in lactation cows with mastitis [7]. Since S. aureusor MRSA-infected cows are associated with the increased use of antimicrobials and rising drug resistance, the cows infected by these bacteria are normally culled as soon as they were detected. Therefore, developing useful new methods to control and reduce the infection of S. aureus and MRSA in dairy farms is imperative.
Bovine milk is the most popular animal food by humans of all ages. Thus, improving bovine udder health is critical to protect public health. Our previous work observed that the bovine tra cking protein particle complex 9 (TRAPPC9) gene is signi cantly associated with milk somatic cell score (SCS, [11]), a usual indicator of the degree of bovine mastitis. The nuclear factor NF-kappaB (NF-κB) pathway has long been proved as a prototypical proin ammatory signaling pathway. Considering that TRAPPC9 plays an important role in the NF-κB signaling pathway and membrane tra c of Golgi [12,13], our previous study proved that it can serve as a candidate biomarker for S. aureus and MRSA infections in dairy cows and bovine mammary epithelial cell lines (Mac-T) [14,15]. Aside from providing hosts with genetic resistance to pathogens, folic acid (a synthetic form of folate), a methyl donor, also alleviates in ammation through reducing the expression of pro-in ammatory factors [16][17][18]. We thus hypothesized that host TRAPPC9 and supplementation of folic acid improve host defense against S. aureus or MRSA infection in dairy cattle and bovine mammary epithelial cells.
In the present study, we designed a series of experiments to investigate the effects of genetic mutations, DNA methylation, and expression of the bovine TRAPPC9 gene on mastitis progress in a new dairy population and to determine the roles of this gene in the defense against S. aureus and MRSA infections in bovine Mac-T cells. We also systematically assessed the in uences of folic acid supplementation on the incidence of mastitis in vivo, the in ammation progress induced by S. aureus and MRSA and the invasion abilities of S. aureus and MRSA into Mac-T cells. These data provide new insights into the protection of food animals and public health from the in ammatory disease induced by zoonotic bacteria.

Cows
A total of 514 Chinese Holstein cows used for the association study were randomly collected from six farms in the North of China ( Figure 1). The parities of the cows ranged from 1 to 4 and were milked three times a day, the lactation days were among 30-400 days, and SCS varied from 0 to 10. All the dairy cows were fed on the same lactation diet according to the recommendations of lactating Chinese Holsteins.
In addition, 123 healthy perinatal Holstein cows (from a single dairy farm in Hebei Province, China) were randomly selected for the folic acid supplementation experiment. The parities of the cows were 1 or 2, the estimated body weight of each cow was shown in S1 Table, and the expected date of calving of the cows were within 10 days. Accordingly, they were grouped into three experimental groups (groups A, B, and C) supplemented with different doses of folic acid (Group A: 0 mg/day; Group B: body weight (kg) × 0.24 mg/kg/day; Group C: body weight (kg) × 0.48 mg/kg/day). In this study, folic acid was coated by Oriental Kingherb Company (Beijing, China). The coated folic acid was fed to groups B and C for 14 days before calving and 7 days after calving.
Milk samples were aseptically collected and immediately sent to the o cial Dairy Center of China (Beijing, China) for the detection of somatic cell count (SCC). Peripheral blood samples were collected from the cows' tail vein by caudal venipuncture and divided into two tubes. One tube containing coagulant and the other containing potassium ethylenediaminetetraacetic acid (EDTA K3E 15%, 0.12mL; Becton, Dickinson and Company) were used for DNA and RNA extraction respectively. Sera were isolated from the peripheral blood and sent to Beijing Huaying Biotechnology Research Institute for cytokine (NIBP,  NF-κB, TNF-α, IL17, IL4, IFN-γ, IL6, and IL10) tests detected by radioimmune out t, which was performed by R-911-automatic radioimmunoassay count instrument.

Bacterial strains
S. aureus (strain 90-1) was originally isolated from a subclinical case of bovine mastitis by our labs in Beijing, China. MRSA (strain W18), provided by Prof. Xin Wang (Northwest A&F University, Yangling, Shanxi, China), was isolated from a clinical case of bovine mastitis. S.aureus and MRSA (50 μL) were inoculated into 5 mL of tryptone soya broth (Beijing Land Bridged Technology Ltd.), and allowed to grow overnight for 24 h at 37 °C, 200 rpm/min. The concentrations of S. aureus and MRSA were determined by serial dilution and standard plate counting. In Brief, S. aureus and MRSA were diluted six gradients successively with DPBS. Then, 100 μL of the diluent was transferred into plate count agar (PCA, Beijing Land Bridged Technology Ltd) and spread with a glass spreader. Subsequently, the agar plates were incubated at 37 °C for 18-24 h. After the PCA plate culture of the S. aureus and MRSA diluents, the numbers of S. aureus and MRSA colonies were counted. Each diluent PCA was conducted in triplicate.
Finally, the bacteria were diluted in DMEM to obtain 1×10 8 CFU/mL.

Cell culture
The bovine mammary epithelial cell line (Mac-T) was provided by Zhejiang University. Mac-T cells were re-suspended in the warm growth medium of DMEM with Glutamax (Gibco) supplemented with 10% fetal bovine serum and 100 U/mL penicillin and streptomycin (100 mg/mL, Gibco). Mac-T cells were seeded in a 25 cm 2 tissue culture ask at 37 °C in a 5% CO 2 humidi ed incubator, and the medium was changed once every 24 h. After 48 h, cells at 85% con uence were split by adding 1 mL of 0.25% trypsin/EDTA (Gibco) after washing with 2 mL of DPBS (Gibco). Mac-T cells were cultured up to a maximum of three passages to reduce the risk of aberrant expression caused by extended culturing. These cells were centrifuged in 5 mL of DMEM growth medium for 5 min at 1000 rpm/min, seeded at 5×10 5 cells in a 6well cell culture plate (Corning), and then allowed to grow in a growth medium at 37 °C in 5% CO 2 humidi ed incubator.

Total RNA extraction, reverse transcription and quantitative Real-Time PCR
Buffy coat of fresh blood and collected Mac-T cells were placed directly in 1 mL of Trizol reagent (Invitrogen), and total RNA was extracted following the manufacturer's instructions. The mRNA was converted to cDNA with the PrimeScript TM RT reagent kit (Takara). RT-qPCR was performed using the SYBR Green I Master kit (Roche Diagnostics GmbH) on the LightCycler® 480 II (Roche Diagnostics Ltd) following the manufacturer's instructions. RT-qPCR reaction was performed using the following program: 95 °C for 10 min, followed by 45 cycles at 95 °C for 10 s, 60 °C for 10 s and 72 °C for 10 s. At the end of each run, a dissociation melt curve was determined. All melt curves showed a single peak and were consistent with the presence of a single amplicon. Bovine GAPDH was used as the house-keeping gene to normalize the target genes' expression levels. Each cDNA sample was analyzed in duplicate. The relative mRNA expression levels of the target genes were calculated using the 2 -△△Ct method. Primer sequences are as follows (obtained from Tsingke): TRAPPC9-forward: CTGCTCCGCTCGGTGAATGAC, TRAPPC9-Reverse: GCTTTACCGCCAGTTCCACCA; IL-1β-Forward: CTGTCGGACCCATATGAGC, IL-1β-Reverse:  GCTCATGGAGAATATCACTTGTTG; IL-6-Forward: TCCTGAGAAACCTTGAGAAT, IL-6-Reverse:  ATAAGTTGTGTGCCCAGTGG; IL-8-Forward: TGAAGCTGCAGTTCTGTCAAG, IL-8-Reverse: TTCTGCACCCACTTTTCCTTGG; GAPDH-Forward: GGTGCTGAGTATGTGGTGGA, GAPDH-Reverse: GGCATTGCTGACAATCTTGA.

DNA extraction and genotyping of the bovine samples
The genomic DNA was extracted from the whole blood of the dairy cows using the DP-318 Blood DNA kit (Tiangen Biotech Co) following the manufacturer's instructions. DNA was assessed with the NanoDrop TM ND-2000c Spectrophotometer (Thermo Scienti c) and by 1% agarose gel electrophoresis.
The genotyping of the bovine samples consists of two steps. First, SNPs identi cation was conducted by DNA pooling and sequencing. Brie y, a total of 30 samples were randomly selected from 514 cows to construct a DNA pool with an equal DNA concentration of 50 ng/μL for each sample. PCRs were performed in a 25 μL volume containing 12.5 μL of PreMIX (Takara Biotechnology Co), 1 μL of genomic DNA (50 ng), 9.5 μL of ddH2O, and 1μL of each primer (10 pmol/μL). Cycling reaction conditions were as follows: initial denaturation at 95 °C for 5 min; followed by 34 cycles of 30 s at 95 °C, annealing from 52°C to 62 °C for 35 s, and 72 °C for 1 min; and a nal extension at 72 °C for 10 min. The PCR products were sequenced using the ABI3730XL DNA analyzer (Applied Biosystems). In total, four SNPs were identi ed in the new population. For SNP1, SNP2, SNP3 and SNP4, 492, 511, 492 and 501 samples were detected successfully (S2 Table).
And Haploview (Version 4.1) was used for linkage disequilibrium analysis of identi ed SNPs.

DNA bisul te treatment and hot-start PCR
The EZ DNA Methylation Golden kit was used for the sodium bisul te conversion of genomic DNA following the manufacturer's instructions (ZYMO Research). The PCR and sequencing primers of the TRAPPC9 gene used for DNA methylation detection were designed with Oligo 6.0 software and PSQ Assay Design software (Qiagen). Hot start PCR was carried out in a 40 μL system including 20 μL of hotstart PCR premix (ZYMO Research), 1 μL of forward primer (10 pmol/μL), 0.1 μL of reverse primer with the universal tail (10 pmol/μL, [19]), 1 μL of biotin-labeled universal primer (10 pmol/μL), and 4 μL of bisul te-treated DNA. The PCR conditions were 95 °C for 15 min, followed 94 °C for 30 s, 56 °C for 30 s and 72 °C for 30 s for 45 cycles, and 72 °C for 10 min. The PCR products were detected using 2% agarose gel with ethidium bromide [19].

Quantitative DNA methylation analyses by pyrosequencing
The pyrosequencing assays were used to analyze the promoter methylation level of the TRAPPC9 gene quantitatively. Six CpG sites located in the promoter region of the bovine TRAPPC9 gene were tested to analyze the DNA methylation level. The Pyro Q-CpG system (Qiagen) was used to analyze DNA methylation following the manufacturer protocol. In brief, bisul te PCR products were bound with Streptavidin Sepharose High Performance (GE Healthy care). The Sepharose beads grasping the biotinlabeled PCR products were puri ed in 70% ethanol for 5 s, denatured in a denaturing buffer for 5 s, and then washed in washing buffer for 15 s using the Pyrosequencing Vacuum Prep Tool (Qiagen). Subsequently, a 0.5 μM pyrosequencing primer was annealed to the puri ed single-stranded PCR product with annealing buffer (Qiagen). The levels of CpG methylation were expressed as the percentage m C/ ( m C+C), and non-CpG cytosine residues were used as internal controls to verify bisul te conversion. The primers measuring the methylation levels of bovine TRAPPC9 are as follows (obtained from Tsingke): R1 Region-Forward: GGGAAGAGTATAGATAATAGTTAGATAGT, R1 Region-Reverse: GGGACACCGCTGATCGTTTAAACAATCCACTCCAATTACTATTACC, and R1 Region-Sequencing: TTTTAAAGGAAAGGAAAGTG. R2 Region-Forward: GAGTTTGGAGTGGTTTTTTTAGG, R2 Region-Reverse: GGGACACCGCTGATCGTTTAAAATAAATTCCCCTTTTACTATC, and R2 Region-Sequencing: GTTTGGAGTGGTTTTTTTAGG.

Invasion assay of MRSA and S. aureus
The gentamicin protection assay was performed to analyze the invasion ability of MRSA and S. aureus into Mac-T cells as described previously [20]. In brief, Mac-T cells at 85% con uence in 6-well plates were cultured in modi ed growth medium at 37 °C in 5% CO 2 . The monolayers were stimulated with MRSA and S. aureus (1×10 8 CFU/mL) for 3, 6, 8 and 10 hr at a multiplicity of infections of 10 bacteria to 1 Mac-T cell (MOI = 10:1). Then, the Mac-T cells were washed three times with PBS and incubated in DMEM supplemented with 100 μg/mL gentamycin for 2 h in 5% CO 2 at 37 °C without serum. Each experimental treatment was conducted in triplicate. These cells were washed three times in PBS and collected at different time points post-inoculation using 0.25% trypsin/EDTA for RNA extraction. The number of CFU/mL was determined by the qPCR technique.

Challenging Mac-T cells with S. aureus and MRSA strains
Mac-T cells at 85% con uence in 6-well plates were cultured in modi ed growth medium at 37 °C in 5% CO 2 . The monolayers were stimulated with S. aureus (90-1) or MRSA (W18) (1×10 8 CFU/mL) for 3, 6, 12 and 24 h at a multiplicity of infections of 10 bacteria to 1 Mac-T cell (MOI = 10:1). The Mac-T cells infected with bacteria alone and cultured in the same volume of DMEM growth media without any treatment served as controls. Each experimental treatment was conducted in triplicate. These cells were washed three times in PBS and collected at different time points post-inoculation for RNA extraction.

Cell viability assay
Cell viability was measured using the One Solution Cell Proliferation assay (Promega Corporation) to investigate the effect of folic acid on cell growth. In brief, the Mac-T cells were seeded at 5×10 4 cells/mL in 96-well plates at 37 °C in 5% CO 2 for 24 h, and the cell medium included different concentrations of folic acid (0, 5, 10, 20, and 40 μg/mL). After 24 h, the cells were treated with 20 μL of MTS solutions/well at 37 °C with 5% CO 2 for 4 hr. The Mac-T cells cultured in the medium alone and the medium without cells in sextuplicate served as zero adjustment controls. Optical density was measured at 490 nm on a microplate spectrophotometer (Tecan).

Mac-T cells treated with folic acid and followed bacterial infection
Mac-T cells at 85% con uence in 6-well plates were cultured in modi ed growth medium at 37 °C in 5% CO 2 . The monolayers were treated with folic acid at different doses (5, 10, 20, and 40 μg/mL) for 2 hr, and then stimulated with S. aureus (90-1) or MRSA (W18) (1×10 8 CFU/mL) for 6 hr at a multiplicity of infections of 10 bacteria to 1 Mac-T cell (MOI = 10:1). The Mac-T cells infected with S. aureus or MRSA alone and cultured in the same volume of DMEM growth media without any treatment served as controls. Each experimental treatment was conducted in triplicate. These cells were washed three times in PBS and collected at 6 hr post-inoculation for RNA extraction.

Cell transfection and RNA interference (RNAi)
For cell-based functional assays, cell transfections were conducted with Lipofectamine 2000 (Invitrogen) on the RNA interference method following the manufacturer's manual. In brief, the Mac-T cells were seeded at 5×10 5 cells/mL in 6-well plates for 24 hr, and then treated with DMEM containing extra folic acid (5 μg/mL) (treated group) or only DMEM (control group) for 2 h. For RNAi, cells were co-transfected with synthetic small interfering RNA (siRNA, 100 pmol), Lipofectamine 2000 (5 μL) supplemented with warm DMEM containing 10% FBS in a humidi ed atmosphere with 5% CO 2 at 37 °C at 2 h postinoculation. The Mac-T cells were treated with folic acid and then RNAi (TRAPPC9), because the purpose of this study was to investigate the protective effect of folic acid on the expression of in ammatory cytokines (indicators of in ammation). The cells were stimulated with S. aureus or MRSA for 6 hr with MOI = 10:1 at 48 h post RNAi transfection. Each experimental treatment was performed in triplicate. The Mac-T cells were washed three times in PBS and collected at 6 hr post-inoculation for RNA extraction. The primers used for bovine RNAi are as follows (designed and synthesized by GenePharma): TRAPPC9 siRNA-Forward: GCAUGGAAGCGUCGGAAUUTT, and TRAPPC9 siRNA-Reverse: AAUUCCGACGCUUCCAUGCTT. Negative control siRNA-Forward: UUCUCCGAACGUGUCACGUTT, and Negative control siRNA-Reverse: ACGUGACACGUUCGGAGAATT.

Total protein extraction and Western Blot analysis
Mac-T cells were stimulated for 6 hr with S. aureus at 72 hr post-transfection and 2 hr pre-treated with folic acid (5 μg/mL). Finally, these cells were harvested with 500 μL of 0.25% trypsin/EDTA and washed twice in PBS for protein extraction. Total proteins from the cells were extracted by mammalian protein extraction reagent. The cellular proteins were separated by 10% sodium dodecyl sulfate -polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto a polyvinylidene uoride membrane (Millipore, Bedford, USA). The membrane was blocked for 2 hr with 5% non-fat dry milk in Tris-buffered saline-Tween (TBST) at room temperature on a rotary shaker and then washed three times (10 min each) with TBST. The membrane was blotted with 1:500 speci c primary anti-bovine antibodies followed by 1:5000 secondary antibody conjugated with horseradish peroxidase at room temperature for 1 hr. The membrane was again washed three times (10 min each), and blots were generated with the ECL Plus Western Blotting Detection System (Amersham Life Science). Rabbit anti-GAPDH polyclonal antibody (Good Here) was used as the loading control.

Statistical analysis
The number (n) of individual animals used per group is described in each gure panel or gure legend. The replicates of experiments are described in each gure legend. The data were analyzed using SAS 9.2. For a single comparison between two groups, t-test was used, and for multiple groups comparison, ANOVA followed by Tukey's studentized range multiple tests (Tukey) was used. Graphs show mean ± SEM. The upper threshold for statistical signi cance for all experiments was set at P < 0.05. In summary, ANOVA followed by Tukey's studentized range multiple tests (Tukey) were applied in the data of Figures  1A, 1B, 4E, 5B, 5C, 5E and S1C; and t-tests were applied in Figures 1C (left panel), 1F, 2C, 3B, 3D, 3E, 4A, 4C, 4D, S1A.
Association study of these four SNPs in TRAPPC9 with milk somatic cell score (SCS) and the concentration of NF-κB, IL-17, IL-6, IL-4, TNF-α and IFN-γ were analyzed using general linear model (SAS 9.0): y = μ + hys + p + m + h + g + e where y is the phenotype value, μ is the overall mean, hys is the effect of herd-year-season, p is the effect of parity, m is the effect of lactation stages, h is the effect of healthy status, g is the effect of SNP, and e is random residuals.

Remarkable genetic and epigenetic effects of bovine TRAPPC9 on mastitis resistance
We had assessed that the genetic effect of mutations in the bovine TRAPPC9 gene was signi cantly associated with bovine mastitis resistance (three signi cant SNPs, SNP2, SNP3 and SNP4 within TRAPPC9 were identi ed, at P values 1.24E-10, 3.29E-08 and 3.64E-08, respectively, Fig. 1D) on the genome-wide level with Illumina Bovine SNP50 BeadChip (n = 2093, [11]). In the present study, a new Holstein population was hired to screen and validate the genetic effects of the gene on mastitis resistance.
First, to dissect the in uence of TRAPPC9 on mastitis resistance, the 514 experimental cows were divided into three categories (health, subclinical mastitis, and clinical mastitis, Fig. 1A). We observed that the transcriptional levels of the gene in peripheral lymphocyte cells decreased along with the degree of mastitis (left panel, P = 0.003, Fig. 1A). Similar trends were observed for the concentration of TRAPPC9 coding protein (NIBP) (middle panel, Fig. 1A, P = 0.008) and NIBP-regulated protein NF-κB in serum (right panel, Fig. 1A, P = 0.012).
Next, four screened SNPs in bovine TRAPPC9 gene with pooled DNA samples (n = 30) were genotyped in the new dairy cattle population (n = 514, S2 Table), of which SNP2, SNP3 and SNP4 were detected in our initial work [11], and SNP1 was a novel mutation. We observed that SNP1 was associated with SCS (a widely used mastitis-related trait) at P-value of 0.06, SNP2 was signi cantly associated with the concentrations of TNF-α and IFN-γ in the serum (P < 0.05), while SNP3 and SNP4 were not signi cantly associated with any mastitis-related traits (S3 Table). In particular, the investigation of the in uence of the four SNPs on the gene's transcriptional levels also clari ed that only SNP2 had a signi cant effect on the expression level of the gene (P = 0.016, n = 37, Fig. 1B). The results indicated the importance of the SNP2 at the TRAPPC9 gene on dairy mastitis resistance.
On the bovine genome, copy number variation (CNV) is another type of genetic variation. A previous study observed that human TRAPPC9 is a multi-copy gene and its CNVs are associated with intellectual disability [21]. We further measured the CNVs of bovine TRAPPC9 (2 ~ 8 copies) and assessed their association with the traits of mastitis resistance. The data revealed that the CNVs were remarkably associated with the incidence of mastitis, i.e. the subclinical/clinical mastitis cows (SCC > 20⊆10 4 /mL) had signi cantly higher CNVs than the healthy ones (SCC < 20⊆10 4 /mL) (left panel, P = 0.006, Fig. 1C). The population-speci c median copy numbers of the gene ( ve-copies) can serve as a switch point (red arrow, Fig. 1C) for mastitis incidence, because the cows possessing less than or more than ve-copies had signi cantly lower or higher risks of mastitis (right panel, Fig. 1C) Aside from genetics, epigenetic modi cation is also critical for disease prevention. A previous study reported that the change of average DNA methylation in TRAPPC9 is related to the regrowth of the Michigan Cancer Foundation-7 (MCF-7) breast cancer cell line [22]. In the present study, we analyzed the connection between the promoter methylation level of bovine TRAPPC9 and four immune-related traits in the dairy cattle population (Fig. 1F and S1A Figure). On the promoter region of bovine TRAPPC9, two CpG regions (R1 and R2) were chosen to examine the methylation levels quantitatively by bisul te sequencing and analyze their relationship with different immune-related traits separately (Figs. 1D-1F). Data showed that the methylation levels at the CpG 1-1 and CpG 1-2 sites in the R1 region were 100% in all the examined cattle (Fig. 1E). Using DNA sequencing, we found that the cytosine in the CpG 1-3 is on the site of SNP3 (C to T mutation). Therefore, the non-methylation of the CpG on this site could be attributed to the C→T SNP. Notably, the methylation status of the three CpG sites in the R2 region can be divided into two statuses. One is non-methylation (0%) and another is methylation (3.10% ~ 15.16%) for the tested cows (Fig. 1F). In addition, a signi cant association was found between the methylation status and the concentration of NF-κB in serum for the CpG 2-2 site (P = 0.006, Fig. 1F). Moreover, the three CpG sites in the R2 region were remarkably related to the concentrations of serum TNF-α, IL-17, and IFN-γ (S1A Figure).
The above results explained that bovine mastitis-related status traits were closely linked to the transcriptional levels, CNVs, and SNPs at the TRAPPC9 gene and that the promoter methylation levels of the gene could signi cantly in uence the immune responses of the dairy cows.

Expression of TRAPPC9 and interleukin genes of bovine mammary epithelial cells were modulated by S. aureus and MRSA
Given the condition that genetic mutations and DNA methylations of bovine TRAPPC9 are related to the in ammatory progression of the bovine mammary gland, we asked this question "whether the gene contributes to the in ammation pathway of the mammary gland epithelial cells or not". To answer this question, we established a bovine mammary epithelial cell model with the Mac-T cell line induced by S. aureus, a major cause of bovine mastitis. Similarly, the MRSA-induced mastitis cell model was also established. Both strains of S. aureus (90 − 1) and MRSA (W18) used in the current study were separated from fresh bovine milk ( Fig. 2A and S2 Figure).
We measured the transcriptional expression patterns of TRAPPC9 and interleukin genes (IL-1β, IL-6, and IL-8) that are involved in the NF-κB signaling pathway in the S. aureusand MRSA-induced mastitis cell models (Figs. 2B and 2C). We observed that the mRNA expression levels of TRAPPC9 gradually reduced from 3 h to 24 h after the infection of S. aureus or MRSA (Fig. 2C). The expression of the interleukin genes initially increased and then decreased after S. aureus/MRSA infection (Fig. 2C), which are similar to other results [23]. Similar expression trends in Mac-T cells were observed for the four genes after S. aureus or MRSA infection. Remarkably, compared with uninfected cells (control), TRAPPC9 signi cantly decreased at 6 h post-infection, whereas interleukin genes IL-1β, IL-6, and IL-8 signi cantly increased at this time point. On the other hand, a western blot revealed that the protein level of TRAPPC9 (NIBP) at 6 h post-S. aureus infection is consistent with the variation in mRNA expression levels; moreover, the level of NF-κB-p65 slightly decreased in the S. aureus infected cells compared with the control cells (Fig. 2D).
We further carried out the knockdown of TRAPPC9 in the cell model and followed by 6 h infection of S. aureus or MRSA. Figure 3A shows the procedures of the RNAi assay of TRAPPC9 and bacterial infection. The results of RT-qPCR and Western blot indicated that the RNAi targeting TRAPPC9 was effective (Figs. 3B and 3C). The data showed that knockdown of bovine TRAPPC9 without bacterial treatment (Control) correlated with the signi cant upregulation IL-6 expression (P < 0.001) (the left histograms, Fig. 3D). In the cell model, compared with the TRAPPC9-non-knockdown cells (NC), knockdown of the gene followed by S. aureus infection can only signi cantly elevate the expression of IL-1β (the middle histograms, Fig. 3D, P < 0.05), whereas knockdown of the gene followed by the infection of the MRSA strain (W18) can signi cantly up-regulate the expression of IL-1β (P < 0.05) (the right histograms, Fig. 3D). In addition, comparing with the control cells without bacterial treatment (the white histograms in Control, upper panel, Fig. 3D), we observed that the expression of IL-1β in the Lipofectamine 2000-treated Mac-T cells followed by S. aureus infection was signi cantly down-regulated (P < 0.05). It was contradictory with the results in  Fig. 3E), respectively. These two time points were individually chosen for the following functional validation of TRAPPC9 for the two types of bacteria. The knockdown of TRAPPC9 did not in uence the invasion number of S. aureus strain (middle panel, Fig. 3E) but slightly decreased that of MRSA strain (lower panel, Fig. 3E).
Collectively, these data suggest that the expression of TRAPPC9, NIBP, interleukins genes and NF-κB-p65 can be modulated by both S. aureus and MRSA strains. Knockdown of TRAPPC9 can promote the progress of in ammation in vitro by upregulating the expression of pro-in ammatory genes IL-1β and IL-6.

The appropriate dose of folic acid increases TRAPPC9 expression and decreases the invasion of the MRSA strain in vitro
The extent of in ammation can be reduced by dose-dependent folic acid through regulating the expression of key genes [16][17][18]26]. Thus, to determine whether folic acid can effectively in uence the expression of TRAPPC9 and interleukin genes (IL-1β, IL-6 and IL-8), we investigated the effects of folic acid at ve doses on the expression of the four genes. The invasion abilities of the two strains of S. aureus (90 − 1) and MRSA (W18) into Mac-T cells were also measured at the effective dose of folic acid.
We rst evaluated the viability of the Mac-T cells under different doses of folic acid (0, 5, 10, 20, and 40 µg/mL FA) with MTS assay. Results showed that the viability of the folate-treated cells was not affected compared with the untreated cells (0 µg/mL, Fig. 4A). Thus, these ve doses of folic acid were used for the next assays (Fig. 4B). We assessed the mRNA expression levels of TRAPPC9 and the three interleukin genes via RT-qPCR. Folic acid treatment at 5 µg/mL remarkably increased the mRNA expression level of TRAPPC9 and reduced the expression of the interleukin genes in the strains of S. aureus (90 − 1) or MRSA (W18) infected Mac-T cells (Fig. 4C, P < 0.05). Thus, 5 µg/mL folic acid was the suitable dose for restricting the infection of S. aureus and MRSA at least by increasing the expression of TRAPPC9 and reducing the expression of IL-1β, IL-6 and IL-8.
Folic acid treatment of Mac-T cells followed by TRAPPC9 knockdown was performed to determine whether folic acid can mediate the expression of the interleukins by regulating TRAPPC9. In the Mac-T cells added with 5 µg/mL folic acid, knockdown of TRAPPC9 can signi cantly reduce the mRNA expression of IL-6 and IL-1β after the MRSA strain (W18) infection (upper and middle panels in Fig. 4D), and moderately decreased that of IL-8 (lower panel in Fig. 4D). No apparent changes were observed in the S. aureus (strain 90 − 1)-infected cells for the three genes (Fig. 4D). This nding suggests that either TRAPPC9 or folic acid can down-regulate the in ammation factors of MRSA-induced cells.
Finally, the in uence of folic acid in the invasion abilities of the two strains of S. aureus and MRSA was investigated. Results showed that 5 µg/mL folic acid can signi cantly decrease the invasion number of MRSA (W18) into Mac-T cells (right panel, Fig. 4E, P < 0.05) but did not affect the invasion of S. aureus (left panel, Fig. 4E). Hence, the above results imply that a suitable dose of folic acid has a positive impact on defense against MRSA-induced mastitis.

Folic acid dose-dependently prevents mastitis and increases TRAPPC9 expression in vivo
A dairy cattle experiment was designed to investigate the effective dose of folic acid to prevent mastitis.
A total of 123 Holstein cows during the perinatal period were divided into three groups (Figs. 5A). After coated folic acid supplementation for 14 days before calving and 7 days after calving, the cows supplemented with body weight (kg) × 0.24 mg/day of folic acid (group B) showed apparently higher expression of TRAPPC9 than the controls (P > 0.05) (Fig. 5B, left panel). Especially, after folic acid supplementation, the expression levels of TRAPPC9 in the group B cows were signi cantly increased compared with before supplementation (P < 0.05) (Fig. 5B, right panel). Meanwhile, the cows fed with body weight (kg) × 0.24 mg/day (group B) of folic acid exhibited signi cantly higher expression of IL-6 than the controls (Fig. 5C, P < 0.01). After delivery, we recorded the incidence of subclinical mastitis of the dairy cattle. The body weight (kg) × 0.24 mg/day of folic acid supplementation can effectively decrease the incidence of bovine subclinical mastitis (Fig. 5D). Moreover, the cows in groups B and C added with folic acid showed moderately lower SCCs than the control cows (group A) at the second, third and fourth months after folic acid supplementation (Fig. 5E).
The above data indicate that folic acid can dose-dependently increase the expression of bovine TRAPPC9 and prevent bovine mastitis.

Discussion
Page 13/30 S. aureus and its "superbug" strain MRSA, have strong abilities of immunodepression and antibioticresistance [27]. In the present study, we designed a series of experiments to clarify the favorable roles of bovine TRAPPC9, a candidate gene of mastitis resistance, and folic acid in the prevention of mastitis and resistance to S. aureus and MRSA infections in dairy cows. The consequences of these examinations condensed three major ndings.
The rst is that the candidate gene TRAPPC9 can be considered as a biomarker for bovine mastitis because all the investigated factors, including mRNA levels, protein expressions, SNP, CNVs, and DNA methylation of the gene, were apparently related to mastitis progress in the dairy population. TRAPPC9 was disclosed as a bovine mastitis candidate gene in our previous GWAS study [11], which encodes the novel protein called NIK-and IKKβ-binding protein (NIBP) in the NF-κB pathway [12]. A new Holstein population was hired to investigate the effects of the gene on bovine mastitis-related traits. Not only the mRNA expression of the gene but also its protein (NIBP) level in serum were signi cantly correlated with mastitis progression. In speci c, higher expression of the gene indicated stronger mastitis resistance of the cows. Notably, linkage disequilibrium analysis did not show close linkage disequilibrium among the identi ed four SNPs (S1B Figure). Among the previously signi cant SNPs (SNP2, SNP3 and SNP4, designed in the Illumina 50K SNP-chip), only SNP2 was associated with mastitis-related traits (serum TNF-α and IFN-γ) in the present study. The others (SNP3 and SNP4) and the novel detected SNP (SNP1) did not show a signi cant effect on the seven studied traits (S3 Table). This nding reminds us that for low-heritability traits ( such as bovine mastitis), SNPs effects among different population may vary, thus searching for the causal mutations is necessary. In addition, abnormal CNV of human TRAPPC9 is correlated with hereditary breast cancer [28]. Our results showed that the cows possessing ve or more copies of TRAPPC9 had a higher incidence of mastitis (Fig. 1C). However, no signi cant differences were found between the CNVs and the mRNA expression of TRAPPC9 in the cows (S1C Figure), which are similar to the results of Kaufman's studies in human [29]. The underlying reason could be that the samples chosen in the expression assay were peripheral blood tissue, while the expression level of the gene varied largely among different types of tissues [12]. Thus, choosing more types of tissues or speci c blood cell types may guarantee clearer results in the future.
Genetic mutants and epigenetic modi cations regulate immune gene expression [30]. Kuhmann et al. (2011) reported that the methylation changes of TRAPPC9 are related to the regrowth of MCF-7 (human breast adenocarcinoma cell line) cells [22]. In the present study, we provided evidence that the unmethylated TRAPPC9 promoter in region R2 is signi cantly correlated with higher concentration of NF-κB and cytokines in cow serum compared to the methylated cows. TRAPPC9 has been reported to be associated with some human diseases (liver disease and breast/colon cancer) [31]. Based on our results in vivo and in vitro, mRNA expression and methylation status of TRAPPC9 can be used to measure the healthy status of bovine udder, and CNVs and SNP2 in this gene can be regarded as the genetic basis of marker assistant selection for bovine mastitis resistance.
The second major nding is that in vitro and in vivo data indicated the preventive effect of folic acid on in ammation induced by S. aureus and MRSA. Decreasing the effectiveness of antibiotic drugs for zoonotic S. aureus and MRSA is a critical issue facing modern medical treatment. Increasing evidence indicates that the epidemiology of livestock-associated MRSA has raised concerns about its existence in food animals [32]. Up to date, the transmission of livestock-associated MRSA from farms to human is reported in several countries [9,33,34]. Folic acid is an essential water-soluble B vitamin with multiple biological functions, such as antibacterial, antibiotic, anti-apoptotic, anticancer, and anti-in ammatory [35][36][37][38][39]. However, the effect of folic acid on in ammation induced by S. aureus or MRSA was rarely reported. In the current study, S. aureusand MRSA-induced mastitis cell models were formulated to investigate the effects of folic acid on mastitis resistance and prevention. In our results, the expression patterns of the interleukin genes (IL-1β, IL-6, and IL-8) in Mac-T cells are similar to other results [40,41], i.e., the mRNA expression of IL-1β, IL-6, and IL-8 initially increased and then decreased (Fig. 2C), which may be related to immunosuppression mediated by S. aureus [28,42]. These results indicated the success of the formulated S. aureusand MRSA-induced mastitis cell models. Folic acid plays a positive role in preventing in ammation, oxidative stress, and some diseases possibly by mediating the expression of cytokines (IL-1β, IL-6, TNF-α and others) [17,19,43,44]. During female pregnancy and lactation, appropriate folate intake reduced the expression of IL-6 and TNF-α of dams and their offspring [45]. Our in vitro experiments showed similar results that appropriately supplementary folic acid (5 µg/mL) reduced the in ammation of the mammary epithelial cells induced by either S. aureus or MRSA may through in uencing the expression of interleukins genes (IL-1β, IL-6, and IL-8). Lower incidence of mastitis was also observed in the cows with folic acid supplementation compared with the control cows.
It is worth noting that contradiction exists between the in vivo and in vitro studies regarding the effect of folic acid on IL-6. This may be due to the two different experimental materials in vivo (peripheral blood) and in vitro (Mac-T cells), and context-dependent pro-and anti-in ammatory properties of IL-6 [46]. In the future, effect of folic acid on the IL-6 expression level in vivo and in vitro will be further investigated.
The third major nding is that folic acid prevents the in ammation progresses induced by MRSA partially through a TRAPPC9 mediated NF-κB pathway. Folic acid can alleviate in ammation by regulating some in ammation-related genes (e.g., interleukin genes) in the NF-κB pathway [17,47,48]. NF-κB is a ubiquitous transcriptional factor and plays a pivotal role in many pathophysiological processes including in ammation, immune response, aberrant cell growth or apoptosis [49][50][51]. NF-κB activation is widely proved to be the central initiating cellular affair of host responses to invasion by bacteria. Under normal conditions, NF-κB remains inactive by the inhibitor of NF-κB (IκB), and it can be activated by IκB kinase (IKK) or NF-κB inducing kinase (NIK) [1]. In the current study, NF-κB-P65 and NIK-and IKKβ-binding protein NIBP (coded by TRAPPC9 gene) decreased at 6 h post-stimulation with S. aureus ( Fig. 2C and 2D). Opposite expression trends between TRAPPC9 and interleukin genes (IL-1β and IL-6) in Mac-T cells infected with S. aureus or MRSA were veri ed in RNAi assay ( Fig. 2C and 3D). These data are consistent with the previous report that high expression of IL-1β corresponds to the peak SCC in cows with mastitis [52]. Based on our results, IL-1β and IL-6 in these three cytokines are more likely to be regulated by TRAPPC9. The importance of NF-κB in the process of bovine mastitis has been widely reported [53,54], thus as the upstream regulatory factor of NF-κB, TRAPPC9 coding NIBP may possess greater potential to promote bovine mastitis resistance and can be regarded as a more valuable biomarker for bovine mastitis. Thus, a deeper study of this potential effect is warranted to be conducted.
Furthermore, in dairy cattle, the expression levels of TRAPPC9 and NF-κB reduced with the increased in ammation of mammary glands (Fig. 1A). However, the expression level of TRAPPC9 was only moderately increased in the cows after feeding extra folic acid during the periparturient period, which may be due to the existence of the above-mentioned dosage effect of folic acid or insu cient feeding time at the individual level. The transcriptional expression levels of RELA and RELB in the NF-κB family and MAP3K14 and IKBKB in the NF-κB pathway were positively correlated with that of TRAPPC9 in a healthy dairy population post folic acid supplementation ( Figure S3, P < 0.05). It is reported that infections caused by methicillin-resistant strains of S. aureus (MRSA) have a higher mortality rate than infections caused by methicillin-susceptible strains [55]. Some researchers veri ed that MRSA has a high capacity to destroy the function of innate immune cells and complement, evade neutrophil killing and survive after phagocytosis [56]. Consistently, our study showed that MRSA led to more severe mastitis in ammation responses than S. aureus in dairy mammary epithelial cells, which could adhere and invade quickly and strongly to Mac-T cells. Moreover, folic acid can markedly restrict the invasion of MRSA into bovine Mac-T cells, and reduce the in ammation induced by MRSA. However, the invasion of S. aureus cannot be decreased. More doses and time points of folic acid addition for the prevention of S. aureus infection should be explored later. Overall, combined above data with previous studies results, we propose that folic acid can prevent MRSA-induced mastitis partially through a TRAPPC9-mediated NF-κB pathway (Fig. 6).

Conclusion
In summary, we have revealed that folic acid has a benignant effect on bovine MRSA mastitis by upregulating the expression of TRAPPC9 in a dose-dependent manner in Mac-T cells (5 µg/mL) and in dairy cows (body weight (kg) × 0.24 mg/day). Our study highlighted the importance of internal bovine TRAPPC9 mediated NF-κB pathway and external folic acid supplementation in treating and preventing zoonotic S. aureus and MRSA infections, which may reduce the overuse of antibiotics in livestock effectively and improve public health.

Availability of data and material
The raw data for the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no con ict of interest.

Funding
The study was supported by the NSFC-PSF joint project (31961143009), Beijing Natural Science     The data are shown as mean ± SEM. *, ** or #, ## indicate a P-value less than 0.05 and 0.01 determined by t-test, respectively.

Figure 5
In uence of folate supplement in vivo. (A) Folate supplement for Holstein cows during the perinatal period. Group A: 0 mg/day feeding of folate, group B: (body weight (kg) × 0.24mg/day, and group C: (body weight (kg) × 0.48mg/day). (B) Relative mRNA expression of TRAPPC9 in the three groups after 21 days of folate supplementation. (C) The concentration of IL-6 in the serum in the three groups after 21 days of folic acid supplementation. (D) Incidence of bovine subclinical mastitis during this experiment within one month after folate supplementation. The red number in brackets means the number of cows with subclinical mastitis, and black number means the total number of cows used in each group. Fisher's exact test was used to analyze this data (P > 0.05). (E) SCC in the three groups at second, third, and fourth months after calving. Figure 6