Regulation of KHNYN antiviral activity by the extended di-KH domain and nucleo-cytoplasmic trafficking

The zinc finger antiviral protein (ZAP) restricts a broad range of viruses by binding CpG dinucleotides in viral RNA to target it for degradation and inhibit its translation. KHNYN was recently identified as an antiviral protein required for ZAP to inhibit retroviral replication, though little is known about its functional determinants. KHNYN contains an N-terminal extended di-KH-like domain, a PIN endoribonuclease domain and a C-terminal CUBAN domain that binds NEDD8 and ubiquitin. We show that deletion of the extended di-KH domain reduces its antiviral activity. However, despite its similarity to RNA binding KH domains, the extended di-KH domain in KHNYN does not appear to bind RNA. Mutation of residues in the CUBAN domain that bind NEDD8 increase KHNYN abundance but do not alter its antiviral activity, suggesting that this interaction regulates KHNYN homeostatic turnover. In contrast, a CRM1-dependent nuclear export signal (NES) at the C-terminus of the CUBAN domain is required for antiviral activity. Deletion of this signal retains KHNYN in the nucleus and inhibits its interaction with ZAP. Interestingly, this NES appeared in the KHNYN lineage at a similar time as when ZAP evolved in tetrapods, indicating that these proteins may have co-evolved to restrict viral replication. AUTHOR SUMMARYAntiviral proteins restrict viral replication in many different ways, including inhibiting viral gene expression. ZAP is an antiviral RNA binding protein that must interact with other cellular proteins to inhibit viral protein synthesis. KHNYN is a ZAP cofactor that is required for it to inhibit retroviral replication. Because little is known about how KHNYN functions in this role, we have analyzed how two of its domains regulate its antiviral activity. We first show that the extended di-KH domain in KHNYN is required for its antiviral activity. While it is related to di-KH domains in RNA binding proteins, it appears to have lost its ability to bind RNA and KHNYN likely acts in the restriction pathway after ZAP binds a target viral RNA. Second, we show that the KHNYN CUBAN domain regulates both its protein abundance and trafficking within the cell. The CUBAN domain contains a nuclear export signal and, when this signal is mutated, KHNYN is sequestered in the nucleus, has substantially reduced antiviral activity and does not interact with ZAP. Overall, we show that the extended di-KH and CUBAN domains in KHNYN are required for it to act as a cofactor for ZAP to restrict viral replication.