COVID-19 pandemic caused by SARS-CoV-2 infection severely threatens global health and economic development. No effective antiviral drug is currently available to treat COVID-19 and any other human coronavirus infections. We report herein that a CFDA-approved macrolide antibiotic, carrimycin, potently inhibited the cytopathic effects (CPE) and reduced the levels of viral protein and RNA in multiple cell types infected by human coronavirus 229E, OC43, and SARS-CoV-2. Time-of-addition and pseudotype virus infection studies indicated that carrimycin inhibited one or multiple post-entry replication events of human coronavirus infection. In support of this notion, metabolic labelling studies showed that carrimycin significantly inhibited the synthesis of viral RNA. Our studies thus strongly suggest that carrimycin is an antiviral agent against a broad-spectrum of human coronaviruses and its therapeutic efficacy to COVID-19 is currently under clinical investigation.

Bitespiramycin (BT) is produced by recombinant spiramycin (SP) producing strain Streptomyces spiramyceticus harboring a heterologous 4″-O-isovaleryltransferase gene (ist). Exogenous l-Leucine (l-Leu) could improve the production of BT. The orf2 gene found from the genomic sequence of S. spiramyceticus encodes a leucine-responsive regulatory protein (Lrp) family regulator named as SSP_Lrp. The functions of SSP_Lrp and l-Leu involved in the biosynthesis of spiramycin (SP) and BT were investigated in S. spiramyceticus. SSP_Lrp was a global regulator directly affecting the expression of three positive regulatory genes, bsm23, bsm42 and acyB2, in SP or BT biosynthesis. Inactivation of SSP_Lrp gene in S. spiramyceticus 1941 caused minor increase of SP production. However, SP production of the ΔSSP_Lrp-SP strain containing an SSP_Lrp deficient of putative l-Leu binding domain was higher than that of S. spiramyceticus 1941 (476.2 ± 3.1 μg/L versus 313.3 ± 25.2 μg/L, respectively), especially SP III increased remarkably. The yield of BT in ΔSSP_Lrp-BT strain was more than twice than that in 1941-BT. The fact that intracellular concentrations of branched-chain amino acids (BCAAs) decreased markedly in the ΔSSP_Lrp-SP demonstrated increasing catabolism of BCAAs provided more precursors for SP biosynthesis. Comparative analysis of transcriptome profiles of the ΔSSP_Lrp-SP and S. spiramyceticus 1941 found 12 genes with obvious differences in expression, including 6 up-regulated genes and 6 down-regulated genes. The up-regulated genes are related to PKS gene for SP biosynthesis, isoprenoid biosynthesis, a Sigma24 family factor, the metabolism of aspartic acid, pyruvate and acyl-CoA; and the down-regulated genes are associated with ribosomal proteins, an AcrR family regulator, and biosynthesis of terpenoid, glutamate and glutamine. SSP_Lrp in S. spiramyceticus was a negative regulator involved in the SP and BT biosynthesis. The deletion of SSP_Lrp putative l-Leu binding domain was advantageous for production of BT and SP, especially their III components.

Glycosylation inactivation is one of the important macrolide resistance mechanisms. The accumulated evidences attributed glycosylation inactivation to a glucosylation modification at the inactivation sites of macrolides. Whether other glycosylation modifications lead to macrolides inactivation is unclear. Herein, we demonstrated that varied glycosylation modifications could cause inactivation of midecamycin, a 16-membered macrolide antibiotic used clinically and agriculturally. Specifically, an actinomycetic glycosyltransferase (GT) OleD was selected for its glycodiversification capacity towards midecamycin. OleD was demonstrated to recognize UDP-D-glucose, UDP-D-xylose, UDP-galactose, UDP-rhamnose and UDP-N-acetylglucosamine to yield corresponding midecamycin 2′-O-glycosides, most of which displayed low yields. Protein engineering of OleD was thus performed to improve its conversions towards sugar donors. Q327F was the most favorable variant with seven times the conversion enhancement towards UDP-N-acetylglucosamine. Likewise, Q327A exhibited 30% conversion enhancement towards UDP-D-xylose. Potent biocatalysts for midecamycin glycosylation were thus obtained through protein engineering. Wild OleD, Q327F and Q327A were used as biocatalysts for scale-up preparation of midecamycin 2′-O-glucopyranoside, midecamycin 2′-O-GlcNAc and midecamycin 2′-O-xylopyranoside. In contrast to midecamycin, these midecamycin 2′-O-glycosides displayed no antimicrobial activities. These evidences suggested that besides glucosylation, other glycosylation patterns also could inactivate midecamycin, providing a new inactivation mechanism for midecamycin resistance. Cumulatively, glycosylation inactivation of midecamycin was independent of the type of attached sugar moieties at its inactivation site.

Thiazinogeldanamycin (2) was identified from Streptomyces hygroscopicus 17997 at the late stage of the fermentation. The pH was firstly proposed as an important factor in the biosynthesis of it. It was verified that 2 was produced by direct chemical reactions between geldanamycin (1, GDM) and cysteine or aminoethanethiol hydrochloride at pH > 7 in vitro. The proposed synthesis pathway for compound 2 was also discussed. Eleven new C-19-modified GDM derivatives, including five stable hydroquinone form derivatives, were synthesized, most of which exhibited desirable properties such as lower cytotoxicity, increased water solubility, and potent antitumor activity. Especially, compounds 5 and 8 showed antitumor activities against HepG2 cell with IC₅₀ values of 2.97–6.61 μM, lower cytotoxicity and at least 15-fold higher water solubility compared with 1 in pH 7.0 phosphate buffer.