Fig 2: Screening for H2 production mutants excess to hpm91 a Outline of the screening method. b Phenotype confirmation of the hpm91-derived mutants. H2-producing capability of the selected mutants was determined with a GC-2014 gas chromatographer (Shimadzu; Japan) at 5 days of sulfur deprivation. Standard deviations were estimated from 3 biological replicates

Fig 3: P. atrosepticum HyfK can activate E. coli Hyd-3, and E. coli HycI can activate P. atrosepticum Hyd-4. (a) E. coli strains MC4100 (FHL-1+) and MGE1dI (?hycI) were transformed separately with plasmids harbouring hycIEc, hyfKPba or a vector control (pQE80). Strains were grown anaerobically in LB medium supplemented with 0.2 % (w/v) formate, ampicillin and 1 mM IPTG where indicated (+) for 24 h at 37 °C. (b) P. atrosepticum strains PH002 (?hybC, FHL-2+) and PH006 (?hybC, ?hyfK) were transformed with plasmids containing hycIEc, hyfKPba or pQE80. Strains were grown anaerobically in low-salt LB (LSLB) medium supplemented with ampicillin and 1 mM IPTG where indicated (+) for 48 h at 24 °C. (c) E. coli strains MGE1dI (?hycI), FTE003 (?hyaB, ?hybC, ?hycE, PT5::hyfA) and FTE004 (?hyaB, ?hybC, ?hycE, PT5::hyfA, ?hyfG::hycE) were transformed with plasmids containing hycIEc, hyfKPba or a vector control (pQE80). Strains were grown anaerobically in LB medium supplemented with 0.8 % (w/v) glucose, ampicillin and 1 mM IPTG for 16 h at 37 °C. In all cases, H2 headspace samples were extracted and analysed by gas chromatography (Shimadzu GC2014 using a 5A molecular packed column with thermal conductivity detection). Data was normalized by OD600 and culture volume. Error bars represent sd (n=3).