Assimilatory sulfate reduction in the marine methanogen Methanothermococcus thermolithotrophicus

Methanothermococcus thermolithotrophicus is the only known methanogen that grows on sulfate as its sole sulfur source, uniquely uniting methanogenesis and sulfate reduction. Here we use physiological, biochemical and structural analyses to provide a snapshot of the complete sulfate reduction pathway of this methanogenic archaeon. We find that later steps in this pathway are catalysed by atypical enzymes. PAPS (3′-phosphoadenosine 5′-phosphosulfate) released by APS kinase is converted into sulfite and 3′-phosphoadenosine 5′-phosphate (PAP) by a PAPS reductase that is similar to the APS reductases of dissimilatory sulfate reduction. A non-canonical PAP phosphatase then hydrolyses PAP. Finally, the F420-dependent sulfite reductase converts sulfite to sulfide for cellular assimilation. While metagenomic and metatranscriptomic studies suggest that the sulfate reduction pathway is present in several methanogens, the sulfate assimilation pathway in M. thermolithotrophicus is distinct. We propose that this pathway was ‘mix-and-matched’ through the acquisition of assimilatory and dissimilatory enzymes from other microorganisms and then repurposed to fill a unique metabolic role.


Enzymatic rates.
The specific activities for the coupled enzymes ATPS-APSK are low but within the range described for other organisms (i.e.Zea mays ATPS has a specific enzyme activity of 0.000246 µmol/min/mg; Arabidopsis thaliana ATPS has a specific enzyme activity of ~0.145 µmol/min/mg; Saccharomyces cerevisiae ATPS has a specific enzyme activity of 0.69-140 µmol/min/mg according to the Brenda database).
MtATPS, MtAPSK and MtPAPSR are assimilating enzymes.Therefore, their turnover may not be as high as dissimilatory ones.They may even be tightly regulated (i.e.competitive inhibition by end products, as shown by the retro-inhibition of PAP in the absence of PAPP) to avoid excessive cellular energy consumption.
We cannot exclude that our experimental set-up had a negative effect on the enzymatic rates, as we could not use high salt concentrations (high KPO4 2-concentrations have been shown to increase enzymatic activity for certain hydrogenotrophic methanogens 3 , but cannot be used here as PO4 2-is a reaction product that would disturb the assay equilibrium), nor could we work at physiological temperature (65 °C) because we used enzymes from mesophiles (i.e.pyrophosphatase from E. coli) in the coupled assays.In addition, the artificial electron donor methyl viologen is a surrogate, and the physiological electron donor could significantly favour the reaction towards sulfite production, thus triggering the equilibrium.
Our structural and phylogenetic analyses suggest that MtATPS and MtAPSK are closely related to the thermophilic gram-negative bacterium Thermus thermophilus and marine archaeon Aeropyrum pernix, respectively.Since ATPS and APSK coding genes have only been found in less than 40 methanogens (Supplementary Fig. 8), it would argue for a lateral gene transfer rather than an ancestral origin common to all methanogens followed by the loss of these genes 4 .Retrieving the original donors to the different methanogen species would require deeper analyses and a larger set of sequences.
The tree suggests that MtPAPSR is of bacterial origin and was probably acquired by horizontal gene transfer from a bacterium (possibly from Thermincola potens or within its clade).
The archaeal PAP phosphatases form their own clade.Once more, only a limited number of methanogen genomes harbour a PAPP coding gene (Supplementary Fig. 8).Therefore, the genes might have been horizontally transferred from an archaeon (e.g.belonging to the Thermococcales).Since archaeal PAPphosphatases belong to the DHH family, we propose that this new class probably evolved from an ancestor containing the DHH motif.
It has been proposed that methanogenesis and sulfate reduction may have been intertwined pathways for more than 3.4 Gyr. 5 Sulfate assimilation might have been more prevalent in ancient methanogens, and while most of them lost the ability to assimilate SO4 2-, some may have retained the required enzymes and adapted them to serve a hitherto unknown sulfur trafficking function.The pathway presented in our work would require a set of genes that only a few methanogenic genomes encode.If sulfate reduction was used in ancient methanogens, then it is unlikely that it was the one described in M. thermolithotrophicus.