No abstract

Hemagglutinin (HA), a glycoprotein of Influenza A viruses and its proton channel M2 are site-specifically modified with fatty acids. Whereas two cysteines in the short cytoplasmic tail of HA contain only palmitate, stearate is exclusively attached to one cysteine located at the cytoplasmic border of the transmembrane region (TMR). M2 is palmitoylated at a cysteine positioned in an amphiphilic helix near the TMR. The enzymes catalyzing acylation of HA and M2 have not been identified, but zinc finger DHHC domain-containing (ZDHHC) palmitoyltransferases are candidates. We used a siRNA library to knockdown expression of each of the 23 human ZDHHCs in HA-expressing HeLa cells. siRNAs against ZDHHC2 and 8 had the strongest effect on acylation of HA as demonstrated by Acyl-RAC and confirmed by 3H-palmitate labeling. CRISPR/Cas9 knockout of ZDHHC2 and 8 in HAP1 cells, but also of the phylogenetically related ZDHHCs 15 and 20 strongly reduced acylation of group 1 and group 2 HAs and of M2, but individual ZDHHCs exhibit slightly different substrate preferences. These ZDHHCs co-localize with HA at membranes of the exocytic pathway in a human lung cell line. ZDHHC2, 8, 15 and 20 are not required for acylation of the HA-esterase-fusion protein of Influenza C virus that contains only stearate at one transmembrane cysteine. Knockout of these ZDHHCs also did not compromise acylation of HA of Influenza B virus that contains two palmitoylated cysteines in its cytoplasmic tail. Results are discussed with respect to the acyl preferences and possible substrate recognition features of the identified ZDHHCs.

Lipid modification of viral proteins with fatty acids of different lengths (S-acylation) is crucial for virus pathogenesis. The reaction is catalyzed by members of the DHHC family and proceeds in two steps: the autoacylation is followed by the acyl chain transfer onto protein substrates. The crystal structure of human DHHC20 (hDHHC20), an enzyme involved in the acylation of S-protein of SARS-CoV-2, revealed that the acyl chain may be inserted into a hydrophobic cavity formed by four transmembrane (TM) α-helices. To test this model, we used molecular dynamics of membrane-embedded hDHHC20 and its mutants either in the absence or presence of various acyl-CoAs. We found that among a range of acyl chain lengths probed only C16 adopts a conformation suitable for hDHHC20 autoacylation. This specificity is altered if the small or bulky residues at the cavity’s ceiling are exchanged, e.g., the V185G mutant obtains strong preferences for binding C18. Surprisingly, an unusual hydrophilic ridge was found in TM helix 4 of hDHHC20, and the responsive hydrophilic patch supposedly involved in association was found in the 3D model of the S-protein TM-domain trimer. Finally, the exchange of critical Thr and Ser residues in the spike led to a significant decrease in its S-acylation. Our data allow further development of peptide/lipid-based inhibitors of hDHHC20 that might impede replication of Corona- and other enveloped viruses.