Molecular biology and plasmid construction
Plasmid pBC001 is a derivative of pIY107, which contains the Alteromonas macleodii “Deep ecotype” (DSM 17117) [29] hydrogenase operon driven by four TRC promoters (Additional file 1: Table S3 and Additional file 1: Figure S2). In pIY107, the C-terminus of HynS is modified with a strep-tag [8, 30, 31] and the N-terminus of HynL is modified with a His6-tag. Details of the construction of pIY107 will be published elsewhere, but we provide its sequence in the Additional file 1. Plasmid pBC001 is a further modification of pIY107 where a sequence containing an AvrII site within the “orf2” gene of pIY107 was silently mutated, and a new AvrII site was added immediately before the terminator-promoter unit preceding the hynS ORF. With unique AvrII and AgeI restriction sites flanking the hynS gene, this modification facilitated site-directed mutagenesis of hynS.
To construct pBC001, two amplicons of pIY107 were generated using Q5 PCR kit (New England Biolabs): An “upstream” amplicon spanning the region from the AvrII site in the middle of orf2 to the pTRC cassette, and a “downstream” amplicon spanning the pTRC cassette to the AgeI site at the leading end of the hynL gene. The upstream amplicon was generated by primers BC000AvrF and BC001AvrR; the downstream amplicon was generated by primers BC002AvrF and IY171HynSR [8]; sequences and descriptions of primers are provided in Additional file 1: Table S4. A three-piece Gibson isothermal assembly was then performed using both amplicons and pIY107 doubly digested with AgeI and AvrII. The sequence of the inserted region of pBC001 was confirmed by Sanger DNA sequencing (Operon).
Plasmids pBC002 – pBC049 were generated using amplicons generated by forward primer BC002AvrF and a reverse primer bearing the appropriate DNA mutations to effect a substitution, as well as a complementary forward primer and the reverse primer IY171HynSR. A full list of primers used is provided in Additional file 1: Table S4. These amplicons were subjected to Gibson isothermal assembly with plasmid pBC001 doubly digested with the AvrII and AgeI restriction enzymes. E. coli transformation was initially performed in NEB-5α (New England Biolabs) strain but then switched to Epi300 (Clontech). The DNA sequences of the inserted regions spanning the hynS ORF were confirmed by Sanger DNA sequencing (Operon).
Doubly substituted construct plasmids were generated by amplification of regions bearing the 4 best-performing distal and 2 best-performing medial cluster aspartic acid substitutions, followed by adjustment to match concentration, and Gibson isothermal assembly as described above. Sixteen colonies of the assembly library were picked, screened, and sequenced, fortuitously resulting in a full sampling of all 8 combinations of substitutions (likelihood ~25%) and an additional 2 triply-substituted constructions (mechanism of assembly unknown).
Crude whole cell hydrogenase screening assay
To quickly ascertain the effect of an amino acid substitution, a crude screen for hydrogen evolution activity was employed. After transforming the plasmid bearing the hydrogenase into Escherichia coli strain BL21ΔH4 cells [8, 32] and overnight colony outgrowth, individual colonies were picked and used to inoculate 1.7 mL of autoinduction media [33] in sterile 10 mL scintillation vials sealed with sterile natural rubber septa (Aldrich). Cultures were grown overnight (~24 hours) at 30°C, 200 rpm rotation. Following growth, 0.1 mL of 40 mg mL−1 methyl viologen (Aldrich) and 0.1 mL of 0.5 M potassium phosphate solution, pH 7.0, and 0.01 mL of 10% (w/v) Triton X-100 (Aldrich) were anaerobically added from a nitrogen-sparged master solution. Finally, 0.1 mL of 2 M sodium dithionite was anaerobically added and the sealed vial was incubated for 2–4 hours at 30°C. Total hydrogen evolved was assessed using gas chromatography (6890 N, Agilent) using a Fused Silica Molsieve 5A column (CP7537, Varian) of 250 μL samples taken from the vial headspace. Activity was normalized to the activity of pBC001-bearing cultures prepared in parallel.
Bacterial lysate activity assay
Bacterial lysate activity was measured as previously described [8], except experiments were performed in 10 mL vials; sparging was conducted under nitrogen; 250 μL samples were taken from the vial headspace; and different chromatography apparatus was used (6890 N, Agilent).
Briefly, bacteria were lysed using a probe sonicator (Bransonic), and 0.2 mL lysate, cleared by centrifugation (16,000 × g, 4°C), was added to a solution containing methyl viologen and potassium phosphate solutions in proportions as described above and were then sparged with nitrogen gas. Sodium dithionite was added in proportions as described above, and incubated for 2 hours at 30°C. Total hydrogen evolved was assessed using gas chromatography and activity was normalized to total protein content of the lysates as measured by Bradford assay.
Tandem IMAC/strep-tactin hydrogenase preparation
Purified hydrogenases were expressed in 100 mL of autoinduction media; containing 0.5% α-lactose and 0.01% glucose; instead of lysis buffer, NP2 buffer (50 mM Na3PO4, 100 mM NaCl, 1 mM 2-mercaptoethanol, pH 7.0) was used; sonication was performed in two batches; and sonicated cell matter was centrifuged for 20 minutes instead of 10.
An IMAC spin column was prepared by applying 200 μL of TALON cobalt resin (Clontech) to an empty micro bio-spin column (Bio-Rad) and rinsing twice with deionized water and once with NP2 buffer. All spins except as noted were performed at 27 × g for 10 s at 4°C. The resulting cleared lysate was applied in two batches to the spin column; each batch was run through the column three times using 20 s spins. The columns were further washed using NP2 buffer (1 × 1 mL), NP2 buffer + 0.01% (w/v) SDS (1 × 1 mL), NP2 buffer + 0.05% (v/v) tween-20 (5 × 1 mL). Elution was achieved using three applications of 200 μL of LP2 buffer (50 mM Na3PO4, 100 mM NaCl, 1 mM 2-mercaptoethanol, pH 5.0). Each 100 mL expression was split into two batches processed in parallel, with the elutions combined afterwards.
pH exchange was conducted by applying IMAC eluate in three rounds (15 m at 13,000 × g, 4°C) to a 30 kDa microcon spin membrane (YM-30, Millipore) followed by one round after adding 500 μL of NP2 buffer. Retentate was collected by inverting the cartridge and centrifuging (3 m at 3,000 × g), and rinsing with an additional 500 μL of NP2 buffer.
Streptactin purification was conducted by applying this total retentate (~600 μL) to 100 μL of strep-tactin magnetic beads (Qiagen) in a 1.5 mL Eppendorf tube (Denville), followed by 1 h incubation with end-over-end agitation at 4°C. Beads were immobilized using magnetic separation and exchanged with NP2 buffer (1 mL) four times, and eluted using 100 μL NP2B (NP2 + 10 mM biotin), followed by a second round of 50 μL NP2B buffer. Protein content was determined by Bradford assay (Bio-Rad) and adjusted to 0.02 mg mL−1 for all samples. Hydrogenase assay was measured as above, except 20 μL of purified enzyme was diluted to 0.2 mL in NP2; and hydrogen evolution was conducted over 20 hours.
SDS-PAGE analyses
Protein samples of crude, IMAC-purified, and tandem-purified samples were adjusted to matching protein content (0.1 mg mL−1, 0.1 mg mL−1, and 0.02 mg mL−1 respectively), supplemented with 5 × SDS-page loading buffer, and boiled for 5 minutes. These samples were then loaded onto a 10% NuPAGE Bis-Tris gel with the NuPAGE MOPS-SDS running buffer system (Invitrogen), and run on ice at 150 V for 2 or 4 hours.
For western blot analysis of crude samples from the “bacterial lysate assay”, blots were prepared as described previously [8].
For PAGE analysis of the samples from the “tandem IMAC/strep-tactin preparation”, one gel containing Crude/IMAC/tandem samples was subjected to SYPRO ruby staining (Invitrogen) and imaged using a Typhoon fluorescence scanning imager (GE). A second gel containing 3 × replicates of tandem samples was subjected to western blot as described previously [8]. One set of replicates was curved on the gel, complicating the densitometry boxing procedure, so it was removed from analysis, although qualitatively it presented similar results.
Supporting data
The data set supporting the results of this article is included within the article (and its additional file).