Activation and repression of Epstein-Barr Virus and Kaposi's sarcoma-associated herpesvirus lytic cycles by short- and medium-chain fatty acids

Abstract

The lytic cycles of Epstein-Barr virus (EBV) and Kaposi's sarcoma-associated herpesvirus (KSHV) are induced in cell culture by sodium butyrate (NaB), a short-chain fatty acid (SCFA) histone deacetylase (HDAC) inhibitor. Valproic acid (VPA), another SCFA and an HDAC inhibitor, induces the lytic cycle of KSHV but blocks EBV lytic reactivation. To explore the hypothesis that structural differences between NaB and VPA account for their functional effects on the two related viruses, we investigated the capacity of 16 structurally related short- and medium-chain fatty acids to promote or prevent lytic cycle reactivation. SCFAs differentially affected EBV and KSHV reactivation. KSHV was reactivated by all SCFAs that are HDAC inhibitors, including phenylbutyrate. However, several fatty acid HDAC inhibitors, such as isobutyrate and phenylbutyrate, did not reactivate EBV. Reactivation of KSHV lytic transcripts could not be blocked completely by any fatty acid tested. In contrast, several medium-chain fatty acids inhibited lytic activation of EBV. Fatty acids that blocked EBV reactivation were more lipophilic than those that activated EBV. VPA blocked activation of the BZLF1 promoter by NaB but did not block the transcriptional function of ZEBRA. VPA also blocked activation of the DNA damage response that accompanies EBV lytic cycle activation. Properties of SCFAs in addition to their effects on chromatin are likely to explain activation or repression of EBV. We concluded that fatty acids stimulate the two related human gammaherpesviruses to enter the lytic cycle through different pathways. Importance: Lytic reactivation of EBV and KSHV is needed for persistence of these viruses and plays a role in carcinogenesis. Our direct comparison highlights the mechanistic differences in lytic reactivation between related human oncogenic gammaherpesviruses. Our findings have therapeutic implications, as fatty acids are found in the diet and produced by the human microbiota. Small-molecule inducers of the lytic cycle are desired for oncolytic therapy. Inhibition of viral reactivation, alternatively, may prove useful in cancer treatment. Overall, our findings contribute to the understanding of pathways that control the latent-to-lytic switch and identify naturally occurring molecules that may regulate this process.

FIG 1

Structures of fatty acids. SCFAs have 2 to 5 carbon atoms, and MCFAs have 6 to 10 carbon atoms.

FIG 2

Activation and blocking of the EBV lytic cycle by straight-chain fatty acids. EBV-infected HH514-16 cells were treated with a single straight-chain fatty acid alone (A and C) or combined with butyrate (3 mM) (B and D) for 18 h. The concentration of fatty acid was 10 mM unless noted as 3 mM. (A and B) Lytic induction was determined by the relative expression of BZLF1 and BGLF5 mRNAs, measured by RT-qPCR analysis, in triplicate, of RNAs extracted from untreated versus treated cells. Expression is presented relative to stimulation by butyrate (100%). The expression of BZLF1 was induced 11-fold by butyrate compared to that in untreated cells, and that of BGLF5 was induced 150-fold. (C and D) Lytic induction was measured by the level of ZEBRA protein detected on an immunoblot. HDAC inhibition was measured using an anti-acetyl (lysine 9 and lysine 14) H3 rabbit polyclonal antibody (AcH3). Immunoreactive protein levels were normalized to β-actin levels and expressed relative to the amount in the untreated control cells.

FIG 3

Activation and blocking of the KSHV lytic cycle by straight-chain fatty acids. KSHV-infected HH-B2 cells were treated with a single straight-chain fatty acid or with butyrate (3 mM) combined with another straight-chain fatty acid (10 mM) for 12 h. (A) Lytic induction was determined by the relative expression of ORF50 mRNA, measured by RT-qPCR analysis, in triplicate, of RNAs extracted from untreated versus treated cells. Expression is presented relative to stimulation by butyrate (100%). ORF50 mRNA was induced 240-fold by butyrate compared to that in untreated cells. (B and C) Lytic induction was measured by the ORF50 protein levels on an immunoblot probed with anti-ORF50 antibody.

FIG 4

Lytic reactivation of EBV and KSHV by straight-chain fatty acids in dually infected cells. KSHV- and EBV-infected BC-1 cells were treated with a single fatty acid (10 mM unless noted as 3 mM) (A) or with TPA (20 ng/ml) combined with a straight-chain fatty acid (B) for 18 h. The relative expression of EBV BRLF1 (top) and KSHV ORF50 (bottom) was measured by RT-qPCR analysis, in triplicate, of RNAs extracted from untreated versus treated cells.

FIG 5

Activation and blocking of the EBV lytic cycle by branched-chain fatty acids. EBV-infected HH514-16 cells were treated with fatty acids branched at C-2 (A and C), C-3, or C-4 (B and D) in the absence or presence of butyrate for 18 h. The concentration of fatty acid was 10 mM unless noted as 3 mM. (A and B) Lytic induction was determined by the relative expression of BZLF1 and BGLF5, measured by RT-qPCR analysis, in triplicate, of RNAs extracted from untreated versus treated cells. Expression is presented relative to stimulation by butyrate (100%). The expression of BZLF1 was induced 23-fold by butyrate compared to that in untreated cells, and that of BGLF5 was induced 150-fold. (C and D) Lytic induction was measured by an immunoblot probed with anti-ZEBRA antibody. HDAC inhibition was measured using an anti-acetyl (lysine 9 and lysine 14) H3 rabbit polyclonal antibody (AcH3).

FIG 6

Activation and blocking of the KSHV lytic cycle by branched-chain fatty acids. KSHV-infected HH-B2 cells were treated with C-2-branched fatty acids (A) or C-3- or C-4-branched fatty acids (10 mM unless noted as 3 mM) (B) in the absence or presence of butyrate (3 mM) for 12 h. Lytic induction was determined by the relative expression of ORF50, measured by RT-qPCR analysis, in triplicate, of RNAs extracted from untreated versus treated cells. Expression is presented relative to stimulation by butyrate (100%). ORF50 mRNA was induced 108-fold by butyrate compared to that in untreated cells. Lytic induction was also measured by an immunoblot probed with anti-ORF50 antibody.

FIG 7

Lytic reactivation of EBV and KSHV by branched-chain fatty acids in dually infected PEL cells. KSHV- and EBV-infected BC-1 cells were treated for 18 h with C-2-branched fatty acids (A and B) or C-3- or C-4-branched fatty acids (C and D) in the absence (A and C) or presence (B and D) of TPA (10 ng/ml). Fatty acids were used at 10 mM unless noted as 3 mM. The relative expression of EBV BRLF1 (top) and KSHV ORF50 (bottom) was measured by RT-qPCR analysis, in triplicate, of RNAs extracted from untreated versus treated cells. In panels B and D, expression is presented relative to stimulation by TPA (100%). TPA induced BRLF1 expression 19- to 78-fold and ORF50 expression 22- to 33-fold in biological triplicate experiments.

FIG 8

Summary comparing activation or blocking of reactivation of the EBV and KSHV lytic cycles by fatty acids. The data from all EBV⁺ cell lines tested (HH514-16, Raji, and BC-1) (A and C) or all KSHV⁺ cell lines tested (HH-B2, BC-3, and BC-1) (B and D) were combined. Activation of EBV (A) or KSHV (B) in cells treated with each fatty acid was compared to that in untreated cells. Repression of EBV (C) or KSHV (D) in cells treated with a fatty acid combined with a known inducing agent (either butyrate or TPA) was compared to that with the known inducing agent alone, set at 100% lytic activation. The values are mean ± standard errors of the means (SEM). **, P < 0.0001; *, P < 0.05. (E) Summary comparing activation (red) or blocking (green) of reactivation of the EBV and KSHV lytic cycles by fatty acids. ¹MCFAs block EBV reactivation in cells induced by butyrate.

FIG 9

Correlation between lipophilicity and activation or blocking of EBV and KSHV lytic cycles. The calculated octanol-water distribution coefficients at pH 7.4 (LogD) (ChEMBL database [90]) of the straight- and branched-chain fatty acids, an assessment of lipophilicity, were plotted against their effects on EBV (A) and KSHV (B) lytic reactivation.

FIG 10

Fatty acids alter activation of the BZLF1 promoter (Zp). (A) Schematic diagram of the EBV BZLF1 promoter (Zp) from positions −221 to +12 relative to the transcription start site, with known response elements labeled. (B) Effects of short-chain fatty acids (10 mM, except for butyrate [3 mM]) on the expression of luciferase regulated by a fragment of Zp from positions −547 to +12. HH514-16 cells were transfected with Zp−547/+12-GL2 by nucleofection and incubated at 37°C for 1 h. Fatty acids were then added to cells from one transfection divided into a 96-well plate for 48 h. (C and D) Effects of butyrate, VPA, or the fatty acids combined on the expression of Zp−221/+12-GL2, either wild type or with inactivating mutations in the ZIIIA/ZIIIB, ZIIR, ZV/ZV′, or ZIIR/ZV/ZV′ elements. The luciferase activities were normalized to the total protein level. The data are averages for at least three separate transfections. RLU, relative light units.

FIG 11

EBV early proteins are expressed in the presence of VPA when ZEBRA is overexpressed. HH514-16 cells were transfected with empty vector or a plasmid that expresses the ZEBRA protein under the control of the CMV IE promoter. Immediately after transfection, the cells were treated with butyrate (NaB; 3 mM), VPA (10 mM), or NaB plus VPA for 24 h. Expression of Rta, EA-D, ZEBRA, and β-actin was measured by immunoblotting. Data are representative of biological triplicate experiments.

FIG 12

Effects of butyrate and VPA on activation of H2AX and p53 at early and late times. EBV-infected HH514-16 cells were treated with butyrate (3 mM) and/or VPA (10 mM) for 4, 6, 8, 15, or 24 h (A) or 0, 2, 4, or 24 h (B). H2AX, γH2AX, p53, and phospho-p53(Ser15) levels in cell lysates were measured by immunoblotting. The ratios of γH2AX to total H2AX and phospho-p53(Ser15) to total p53 are presented relative to those in untreated cells at time zero. Lytic induction was measured by the level of ZEBRA protein. Data are representative of at least triplicate biological experiments.