The mechanism of flaviviral suppression of vitamin A metabolism

Mosquito-borne viruses, primarily including families of Flaviviridae (dengue, Zika etc.) and Togaviridae (Chikungunya etc.), infect roughly ~400 million people worldwide. By a blood meal, a naïve female mosquito may acquire an infectious agent (e.g., virus) from an infected person, and pass it on to next person. Therefore, the host-seeking and blood-feeding behavior of female mosquitoes is a key determinant of virus circulation in nature. The mammalian body odors contain many volatile chemical compounds that influence mosquitoes' behaviors. Some of these are well-characterized mosquito attractants. These cues vary from a person to another person, and with the physiological conditions of the same person. However, it remains unknown if mosquito- transmitted viruses can manipulate their host odors to influence mosquito behaviors. In a recent study, we found that dengue (DENV) and Zika (ZIKV) can manipulate their mammalian host skin microbiota to produce a potent mosquito attractant, acetophenone. This was because viral infection significantly increased the abundance of skin Bacillus bacteria, the major producer of acetophenone as a volatile metabolite. DENV/ZIKV downregulated (~25-fold) the expression of a resistin-like bactericidal antimicrobial peptide (AMP), RELMα, in the mouse skin. These results demonstrate that DENV/ZIKV can enhance their host attractiveness to mosquitoes, thus facilitate their transmission. However, how these viruses inhibit RELMα expression is unknown. Constitutive RELMα expression (not any other skin AMPs) requires dietary retinols (a.k.a vitamin A), suggesting that DENV/ZIKV likely intervenes retinol signaling. Retinols are bound by retinol binding protein 4 (RBP4) and transported by a cell surface receptor STRA6, converted into retinaldehydes by retinol dehydrogenase 1 (RDH1), then to retinoic acids by retinaldehyde dehydrogenase 1A (RALDH1A). Retinoic acids bind retinoic acid receptors (RARs), which enter the nucleus and activate gene transcription, including RELMα and RARs. Intriguingly, oral administration of a retinoic acid, isotretinoin, restored normal RELMα expression and skin microbial homeostasis to DENV/ZIKV-infected mice, suggesting that DENV/ZIKV target a step between retinol and retinoic acid. Indeed, our new data showed that DENV/ZIKV replicated productively and suppressed RELMα and RAR expression induced by retinol, but not by retinaldehyde or isotretinoin in mouse primary skin epidermal keratinocytes, the primary epidermal cell type and producer of AMPs. Of note, the Chikungunya virus did not inhibit retinol signaling. Therefore, we hypothesize that DENV/ZIKV interferes with a step between retinol and retinaldehyde of the retinol pathway. To this end, we will pinpoint the step of the retinol pathway targeted by DENV/ZIKV, and identify the viral protein (s) that suppress the retinol pathway. We will address how these flaviviruses suppress the retinol pathway and RELMα expression in the skin. These results could be applicable to other mosquito-borne flaviviruses and lay a foundation for in-depth mechanistic studies.