Biosynthesis of GMGT lipids by a radical SAM enzyme associated with anaerobic archaea and oxygen-deficient environments

Abstract

Archaea possess characteristic membrane-spanning lipids that are thought to contribute to the adaptation to extreme environments. However, the biosynthesis of these lipids is poorly understood. Here, we identify a radical S-adenosyl-L-methionine (SAM) enzyme that synthesizes glycerol monoalkyl glycerol tetraethers (GMGTs). The enzyme, which we name GMGT synthase (Gms), catalyzes the formation of a C(sp³)-C(sp³) linkage between the two isoprenoid chains of glycerol dialkyl glycerol tetraethers (GDGTs). This conclusion is supported by heterologous expression of gene gms from a GMGT-producing species in a methanogen, as well as demonstration of in vitro activity using purified Gms enzyme. Additionally, we show that genes encoding putative Gms homologs are present in obligate anaerobic archaea and in metagenomes obtained from oxygen-deficient environments, and appear to be absent in metagenomes from oxic settings.

Fig. 1. Identification of GMGTs synthase by in vivo and in vitro assays.

a GMGT-0 biosynthesis pathway. b UPLC-MS extracted ion chromatograms of lipid extracts from M. maripaludis with plasmid pMEV4-tes showing production of GDGT-0, and with plasmid pMEV4-gms-tes showing production of GMGT-0. The in vitro biochemical reaction reveals the synthesis of GMGT-0 via incubation of the purified Gms protein with GDGT-0, SAM and DTH at 37 °C for 72 h.

Fig. 2. EPR spectroscopic characterization and AlphaFold2 predicted structure model of Gms.

a, b Domain organization (a) and the overall architecture of Gms (b) showing an N-terminal domain comprised of six helices, a conserved RS domain, a bridging region with two helices, a β-harpin motif, and a C-terminal SPASM domain. c The interaction between the β-hairpin and the pocket involves two hydrogen bonds between F326 and N351, as well as between S334 and E346. d Cavity map for the active site of Gms reveals the hydrophobic pocket. e Representations of the three [4Fe–4S] clusters in the structure model of Gms. Protein is shown as a cartoon, SAM, and [4Fe–4S] clusters are shown in ball-and-stick models. f X-band CW EPR spectra of various Gms samples (i) the wild-type (WT) Gms reduced by DTH; (ii) subsequently adding SAM to DTH-reduced Gms; (iii) the mutant sample of Gms (C438AC370AC372A) containing only radical SAM (RS) cluster reduced by DTH, and incubated with SAM; (iv) the mutant sample of Gms (C438AC123AC127A) containing only AuxI cluster reduced by DTH; (v) the mutant sample of Gms (C123AC127A C370AC372A) containing only AuxII cluster reduced by DTH. Black, blue and green traces are experimental spectra, and the red traces are simulated spectra.

Fig. 3. The distribution of Gms homologs in archaea genomes and environmental metagenomes.

a The BLASTp searches of Gms protein homologs in the NCBI database and the respiration features of archaea are derived from published studies^–. Taxonomic groups that contain at least one species with Gms homologs are designated by black circles. It is important to note that the presence of black circles in a group does not imply that every species within the group possesses Gms homologs. Conversely, taxonomic groups lacking any Gms homologs are represented by white circles. For respiration features, the groups containing only obligate anaerobic archaea are marked in orange, aerobic and facultative archaea are marked in blue, and the groups containing anaerobic and aerobic/facultative archaea are marked in a mix color of orange and blue. b The relative abundances of gms gene homologs are present as RPKM values, and the reads of gms were obtained by tBLASTn searches of metagenomes in the SRA database on NCBI. The sample depths and oxygen concentration profiles are included in the Source Data. The color indicates oxygen levels (blue is oxic, yellow is suboxic, and orange is anoxic). The y-axis represents the approximate depths of water and sediment. c Geographical distribution of analyzed metagenome samples. The map was created using Ocean Data View (version 5.6.5). Source data are provided as a Source Data file.