Mammalian Lipopolysaccharide Receptors Incorporated into the Retroviral Envelope Augment Virus Transmission

Abstract

The orally transmitted retrovirus mouse mammary tumor virus (MMTV) requires the intestinal microbiota for persistence. Virion-associated lipopolysaccharide (LPS) activates Toll-like receptor 4 (TLR4), stimulating production of the immunosuppressive cytokine IL-10 and MMTV evasion of host immunity. However, the mechanisms by which MMTV associates with LPS remain unknown. We find that the viral envelope contains the mammalian LPS-binding factors CD14, TLR4, and MD-2, which, in conjunction with LPS-binding protein (LBP), bind LPS to the virus and augment transmission. MMTV isolated from infected mice lacking these LBPs cannot engage LPS or stimulate TLR4 and have a transmission defect. Furthermore, MMTV incorporation of a weak agonist LPS from Bacteroides, a prevalent LPS source in the gut, significantly enhances the ability of this LPS to stimulate TLR4, suggesting that MMTV intensifies these immunostimulatory properties. Thus, an orally transmitted retrovirus can capture, modify, and exploit mammalian receptors for bacterial ligands to ensure successful transmission.

Figure 1. Mammalian LPS-binding receptors are present in MMTV virions

(A) MMTV virions isolated from the milk of B6 wild-type (WT) and TLR4^−/− mice were captured either by an anti-mouse TLR4 or anti-gp36Env Abs and detected with biotinylated anti-gp36Env Abs. Error bars represent the SE of three independent experiments. Significance was calculated using a two-sample t-test and asterisks indicate degree of significance (* = P<0.03). (B) Western blot of proteins from virions purified from the milk of B6 (WT) and CD14^−/− females. The same density fraction isolated from uninfected milk (MMTV⁻) was used as a negative control. Anti-mouse CD14 Abs and anti-gp52Env Abs were used to detect CD14 and viral Env, respectively. Mouse CD14 recombinant protein was used as a positive control for anti-CD14 Abs. (C) Immunogold labeling of MMTV virions. Top panel, MMTV virions purified from B6 (WT) (a, b, c, d) or CD14^−/− (e, f) mouse milk were stained with anti-gp52Env Abs followed by anti-mouse Ab coupled to gold particles (b, f) or anti-mouse Ab alone (a) or anti-CD14 Ab followed by anti-rat Ab coupled to gold particles (d, e) or anti-rat Abs alone (c). Bottom panel, Quantification of three independent experiments with over 50 virions examined per condition in each experiment. Error bars represent the SE. (D) MMTV virions isolated from the milk of mice lacking CD14, LBP, MD-2, TLR4 and from animals deficient in all four factors were captured with anti-gp36Env Abs bound to plastic. The ELISA was developed either with biotinylated LPS (E. Coli, serotype EH100) (top panel), or anti-gp36Env Abs (bottom panel). Error bars represent SE of three independent experiments. Significance was calculated using a two-sample t-test and asterisks indicate degree of significance (** =P<0.001).

Figure 2. LPS binding proteins carried by the virus are essential for immune activation and efficient transmission

(A) MMTV was isolated from the milk of the B6 (WT), MD-2^−/− (MD-2⁻), CD14^−/− (CD14⁻) and all LPS binding factors (All⁻) deficient females. Viral isolates with LPS content normalized to 1 ng/ml via the LAL assay were added to either WT or CD14^−/− splenocytes. IL-6 in tissue culture supernatants was quantified by ELISA 16 h later. Error bars represent SE of three independent experiments. Significance was calculated using a two-sample t-test, and asterisks indicate degree of significance (** = P<0.001). LPS, 1 ng/ml of LPS from E. Coli, serotype 055:B5, (B) Top panel, B6 (WT) mice were fostered on either WT or MD-2^−/− (MD-2⁻) viremic females for 48h. WT pups fostered on uninfected milk were used as a negative control. The deletion of SAg-cognate Vβ6⁺CD4⁺ T cells among CD4⁺ T cells was used as a read out for viral infection. X, animals confirmed as uninfected by qPCR. Bottom panel, Splenic DNA only from infected mice (not marked with X in top panel) was analyzed for integrated proviruses by qPCR. Data are presented as delta between the cycle threshhold (CT) obtained with MMTV-specific primers and IFN-gamma-specific primers. Significance was calculated using a two-sample t-test and asterisks indicate degree of significance (* = P<0.03, ** = P<0.001). (C) B6 mice were fostered on either BALB/cJ (WT) or BALB/cJ.LBP^−/− (LBP⁻) viremic females for 48h. B6 pups fostered on uninfected BALB/cJ milk were used as a negative control. The deletion of SAg-cognate Vβ6⁺CD4⁺ T cells among CD4⁺ T cells was used to assess whether mice were infected. Error bars represent SE. Significance was calculated using a two-sample t-test and asterisks indicate degree of significance (*** = P<0.0001).

Figure 3. MMTV binds the lipid A component of LPS from diverse species and enables agonist LPS to resist inhibition by an antagonist LPS

(A) LPS-free MMTV from GF MyD88/TLR4 double-deficient mice was incubated with 40 ng/ml of lipid A (E. coli, serotype R515) (E.c.) or R. sphaeroides (R.s.) and pelleted through 30% sucrose cushion. The amount of bound endotoxin was quantified by the LAL assay (left panel). The ability of virus-bound or free lipid A (normalized to 1 ng/ml via the LAL assay) to stimulate IL-6 production by B6 (WT) splenocytes was measured by ELISA (right panel). Neither E.c. or R.s. lipid A elicit IL-6 in MD-2^−/− splenocytes. Error bars represent SE of three independent experiments. Significance was calculated using a two-sample t-test. NS - non-significant. (B) LPS from E. coli serotype 055:B5 (at 5, 50 and 100 ng/ml) or intact SPF-MMTV virions (at 5 ng/ml of endotoxin) was added to B6 splenocytes alone or together with increasing concentration of R.s. lipid A. IL-6 in tissue culture supernatants was quantified by ELISA 16 h later. X, is the factor by which R.s. lipid A concentration was increased relative to E.c. LPS. Error bars represent SE of three independent experiments. Significance was calculated using a paired t-test and asterisks indicate degree of significance (* = P<0.05)

Figure 4. MMTV augments the immunopotency of commensal LPS

(A) The immunostimulatory properties of LPS from B. theta and E. coli (serotype 055:B5) were compared by IL-6 ELISA after addition to B6 (WT) splenocytes. MD-2⁻, MD-2^−/− splenocytes. Error bars represent SE of three independent experiments. Significance was calculated using a two-sample t-test and asterisks indicate degree of significance (** = P<0.001). (B) Binding of biotinylated B. theta and E. coli (serotype O26) LPS by SPF-MMTV virions captured to plastic with anti-gp36Env Abs was measured in serial dilutions. Error bars represent SE of three independent experiments. Significance was calculated using a two-way ANOVA. NS, non-significant. (C) B. theta or E. coli (serotype O26) LPS (at 40ng/ml of endotoxin) were incubated with or without GF-MMTV and spun down via 30% sucrose cushion. Endotoxin levels were measured in pelleted fractions using LAL assay. Fractions with no virus added had no detectable LPS (not shown). Virus-LPS pelleted fractions were added to B6 splenocytes at an endotoxin concentration of 71 pg/ml. The amount of secreted IL-6 was measured in tissue culture supernatants 16 h later. In separate cultures, virus-free B. theta and E. Coli LPS were added to B6 splenocytes at the same endotoxin concentration. Error bars represent SE of three independent experiments. Significance was calculated using a paired t-test and asterisks indicate degree of significance (*= P<0.03 and ** = P<0.001). (D) LPS species isolated from SPF-MMTV virions were added to B6 splenocytes as free form (‘viral’ LPS) or bound to SPF-MMTV virions at 35 pg/ml. The amount of secreted IL-6 was used as a read-out for TLR4 activation. Error bars represent SE of three independent experiments. Significance was calculated using a paired t-test and asterisks indicate degree of significance (* = P<0.03). (E) IL-6 secretion elicited by SPF-MMTV virions in B6 (WT), Caspase 1/4^−/− (Caspase 1/4⁻), or MD-2^−/− (MD-2⁻) splenocytes or by E. coli (serotype 055:B5) LPS from Sigma (S) or Enzo (E) was measured by ELISA. Endotoxin concentrations of 1 ng/ml in all samples were verified by the LAL assay. Error bars represent SE of three independent experiments. Significance was calculated using a paired t-test and asterisks indicate degree of significance (** = P<0.001).