N. S. Scrutton, Manchester Interdisciplinary Biocentre and Faculty of Life Sciences, University of Manchester, 131 Princess Street, Manchester M1 7DN, UK Fax: +44 161 306 8918 Tel: +44 161 306 5153 E-mail: email@example.com
Human methionine synthase reductase (MSR), a diflavin enzyme, restores the activity of human methionine synthase through reductive methylation of methionine synthase (MS)-bound cob(II)alamin. Recently, it was also reported that MSR enhances uptake of cobalamin by apo-MS, a role associated with the MSR-catalysed reduction of exogenous aquacob(III)alamin to cob(II)alamin [Yamada K, Gravel RA, TorayaT & Matthews RG (2006) Proc Natl Acad Sci USA103, 9476–9481]. Here, we report the expression and purification of human methionine synthase from Pichia pastoris. This has enabled us to assess the ability of human MSR and two other structurally related diflavin reductase enzymes (cytochrome P450 reductase and the reductase domain of neuronal nitric oxide synthase) to: (a) stimulate formation of holo-MS from aquacob(III)alamin and the apo-form of MS; and (b) reactivate the inert cob(II)alamin form of MS that accumulates during enzyme catalysis. Of the three diflavin reductases studied, cytochrome P450 reductase had the highest turnover rate (55.5 s−1) for aquacob(III)alamin reduction, and the reductase domain of neuronal nitric oxide synthase elicited the highest specificity (kcat/Km of 1.5 × 105m−1·s−1) and MSR had the lowest Km (6.6 μm) for the cofactor. Despite the ability of all three enzymes to reduce aquacob(III)alamin, only MSR (the full-length form or the isolated FMN domain) enhanced the uptake of cobalamin by apo-MS. MSR was also the only diflavin reductase to reactivate the inert cob(II)alamin form of purified human MS (Kact of 107 nm) isolated from Pichia pastoris. Our work shows that reactivation of cob(II)alamin MS and incorporation of cobalamin into apo-MS is enhanced through specific protein–protein interactions between the MSR FMN domain and MS.
reductase domain of neuronal nitric oxide synthase
Human methionine synthase (EC 184.108.40.206; hMS) – an essential cellular housekeeping enzyme – produces methionine (through the methylation of homocysteine) and tetrahydrofolate (H4-folate) from the demethylation of methyltetrahydrofolate (CH3-H4-folate) (Fig. 1). Cobalamin serves as an intermediary in methyl transfer reactions, and it cycles between the methylcob(III)alamin and cob(I)alamin forms . Cob(I)alamin is a powerful nucleophile that extracts a relatively inert methyl group from the tertiary amine of CH3-H4-folate. The reactive nature of cob(I)alamin makes it susceptible to oxidation [conversion to cob(II)alamin], an event that occurs every 200–1000 catalytic turnovers of hMS . Regeneration of hMS activity involves reductive methylation of cob(II)alamin to form methylcob(III)alamin, a process that couples transfer of an electron from methionine synthase reductase (MSR) with methyl transfer from AdoMet .
Most structural and functional information on hMS is derived by comparison with Escherichia coli cobalamin-dependent methionine synthase (MetH), which shares 55% sequence identity with hMS. There are four functional modules in hMS, arranged linearly and separated by interdomain connectors. By analogy with E. coli MetH, the N-terminal region of hMS comprises two closely packed (βα)8 barrels that bind homocysteine and CH3-H4-folate [4,5]. The cobalamin-binding module is located in the centre of the polypeptide. A crystal structure exists for the C-terminal region of hMS . This contains the ‘activation domain’ (AD) that binds AdoMet and MSR [7,8].
The mechanisms of reactivation of MetH and hMS are distinct. MetH is reactivated by the transfer of reducing equivalents from NADPH to MetH, catalysed by FAD-dependent ferredoxin-NADP+ reductase (FNR) and mediated by flavodoxin (Fld) . MSR is a natural fusion of FNR and Fld [3,9]. It is therefore a member of the cytochrome P450 reductase (CPR) family , which also includes the reductase module of nitric oxide synthase (nNOSred) [11,12] and a novel oxidoreductase 1 of unknown physiological function . These proteins catalyse NADPH oxidation and transfer electrons from the enzyme-bound FAD to the FMN centre, and ultimately to an acceptor redox protein or domain. Although the bacterial FNR/Fld and mammalian MSR are not interchangeable in reactivating MetH and hMS, respectively , human novel oxidoreductase 1 is able to reactivate hMS, but the functional significance of this is unknown .
In addition to electron transfer activity, MSR also has putative chaperone-like activity; it promotes the stability of hMS by facilitating uptake of cobalamin by the apo-form of hMS . The enhanced cofactor binding is thought to result from MSR-catalysed reduction of exogenous aquacob(III)alamin (AqCbl) to form cob(II)alamin. Reduction of the Co centre promotes the dissociation of the lower dimethylbenzimidazole base of the cofactor. Consistent with this, the crystal structure of the cobalamin-binding domain of MetH reveals that the dimethylbenzimidazole base is buried within the protein scaffold, well removed from the corrin ring, suggesting that the lower-coordinated Co state preferentially binds to hMS .
Herein, we report for the first time the development of an expression and purification system for hMS based on the expression host Pichia pastoris. This has enabled us to investigate: (a) the potential for cobalamin incorporation mediated by other mammalian diflavin reductases and also subdomains of MSR; (b) the extent of reductive remethylation of hMS catalysed by the different redox states of MSR; and (c) the ability of structurally related diflavin reducatases to reactivate hMS. These studies have enabled us to refine the chaperone-like role of MSR. We show that specific protein–protein interactions between hMS and MSR (over and above the need to catalyse the reductive chemistry) are required to promote the insertion of the cobalamin into hMS. We also demonstrate that the chaperone-like role is orchestrated entirely through the FMN domain of MSR and is not linked to MSR-catalysed reduction of exogenous AqCbl to form cob(II)alamin as previously proposed .
Results and Discussion
Purification of hMS
The ability to express hMS in a recombinant and functional form has been a major limitation in studies of the hMS and MSR redox system. However, we found that recombinant hMS is expressed as an apoenzyme in P. pastoris at levels that enable purification of sufficient quantities for functional analysis (Table 1). A clear advantage of using Pichia as a heterologous host, as opposed to other eukaryotic expression systems, is the capacity to grow large-scale cultures on relatively inexpensive media. The fact that the enzyme is expressed in the apo-form is consistent with yeast being unable to synthesize cobalamin or transport it across the cell membrane . We purified hMS using two steps, employing ion exchange chromatography followed by cobalamin affinity chromatography (Table 1). The affinity chromatography step conveniently converts the apo-form of hMS into the holoenzyme. The activity of hMS through all purification steps was determined using a nonradioactive spectrophotometric assay (see Experimental procedures). Recombinant hMS was found to be homogeneous after cobalamin affinity chromatography, as judged by SDS/PAGE analysis (Fig. 2, inset). The absorption spectrum of the purified enzyme was typical of the hydroxycobalamin form of the enzyme (Fig. 2). The recovery of the activity was ∼ 10%, and the enzyme was purified 3669-fold. The specific activity and yield of purified hMS were similar to the values obtained using the baculovirus expression system .
Table 1. Purification of hMS from an expressing strain of P. pastoris. The crude extract was generated from 103 g of wet Pichiapastoris cell pellet containing the integrated pPICZMS plasmid. hMS activity was measured using the discontinuous spectroscopic assay outlined in Experimental procedures.
Total protein (mg)
Total activity (nmol·min−1)
Specific activity (nmol·min−1·mg−1)
Reactivation of hMS by MSR
Reductive activation of hMS by MSR was measured by following the incorporation of 14CH3 into methionine from 14CH3-H4-folate. The rate of 14CH3 incorporation was found to saturate with respect to MSR concentration (Fig. 3A). The parameter Kact defines the MSR concentration that defines 0.5 of the total recoverable activity of hMS, and was calculated to be 107 ± 14 nm; the maximal recoverable activity at saturation (kcat) was 1.5 μmol·min−1·mg−1, which is similar to previously reported values for nonrecombinant forms of hMS [3,14]. Reactivation of hMS was not observed when MSR was replaced by nNOSred or CPR, highlighting the need for specific protein–protein interactions between MSR and hMS. Reactivation of hMS was found to be dependent on NADPH concentration in a hyperbolic manner (Fig. 3B), yielding an apparent Km for NADPH of 23.2 ± 3.4 μm. This value is approximately 10-fold higher than that reported previously for purified porcine methionine synthase (MS), but we emphasize that studies with the porcine enzyme were conducted under different assay conditions . Previously, we measured, by isothermal thermal calorimetry and product inhibition studies, an apparent Kd of 37 μm for the MSR–NADP+ complex , and, by stopped-flow experiments, an apparent Kd of ∼ 50 μm for NADPH for the MSR–NADPH complex . These values are in reasonable agreement with the apparent Km for NADPH measured in our reactivation assays.
Reactivation of hMS by different redox states of MSR and the isolated FMN domain
There is a significant thermodynamic barrier to electron transfer from either the MSR FMN semiquinone (FMNsq) or FMN hydroquinone (FMNhq) to the hMS-bound cob(II)alamin . Specifically, the midpoint potential values for FMNox/sq and FMNsq/hq are respectively 380 and 270 mV more electropositive than the putative midpoint potential of the cob(II)alamin/cob(I)alamin couple (determined for MetH ), which equates to free energy changes of 36 and 26 kJ·mol−1, respectively [21–23], for electron transfer between the two cofactors.
We examined whether reductive methylation of hMS–cob(II)alamin requires full or partial reduction of MSR [i.e. whether electron transfer to cob(II)alamin occurs from FMNsq or FMNhq]. We reduced MSR or the isolated FMN domain under anaerobic conditions by titration with dithionite to the desired redox state, and then mixed prereduced enzyme with the remaining reaction components (see Experimental procedures). In an anaerobic reaction mixture, hMS was able to catalytically turn over in the absence of a reactivation partner (Table 2). This is at first sight a puzzling result, as hMS was isolated in the inactive form with the cofactor in the Co3+ oxidation state (i.e. AqCbl). Activity may arise from: (a) a small amount of enzyme present in the active methylcob(III)alamin (MeCbl) form; or (b) the presence of a reducing agent (e.g. thiols) converting AqCbl to cob(II/I)alamin, an event that is more feasible in an anaerobic environment . We found that the addition of oxidized FMN domain to the reaction mixture resulted in an ∼ 3-fold increase in hMS turnover, despite the FMN cofactor being preoxidized by FeCN. The presence of a reducing agent (e.g. natural light or thiols; see Table 2 footnote) in the reaction mixture may reduce a proportion of the FMN domain, converting some of the enzyme to the active form. The ∼ 3-fold increase in activity may arise from the binding of the FMN domain to hMS facilitating binding of AdoMet and/or methyl transfer, although this has not been formally shown. It is known that the isolated FMN domain in the oxidized form does bind to the hMS AD . We found that FMNsq and FMNhq increased hMS product yield by 8- and 11-fold, respectively. This indicates that the isolated FMN domain participates in the reductive remethylation of hMS, which by necessity involves endergonic electron transfer from FMNsq to cobalamin. This energetically unfavourable electron transfer is tightly coupled to methyl group transfer from AdoMet (a highly exothermic reaction), which drives the net reaction forwards.
Table 2. Anaerobic reactivation of hMS by different redox forms of MSR and the isolated MSR FMN domain. Various redox forms of MSR, or the FMN domain (40 μm), were added to an assay mixture containing 0.2 m potassium phosphate buffer (pH 7.2), 100 μm AdoMet, 1 mm homocysteine, 250 μm14CH3-H4-folate (1200 d.p.m. per nmol), and hMS, in a total volume of 250 μL. The reaction was incubated for 10 min at 37 °C, and quenched and analysed following the protocol for the radioactive hMS activity assay described in Experimental procedures.
MS activity (nmol·min−1)
a The low level of hMS activity seen in the absence of flavoprotein may arise from a small fraction of hMS in the MeCbl form following purification. b The increase in hMS activity shown in the presence of the oxidized FMN domain may be due to photoreduction of the FMN cofactor by natural light, in particular during gel filtration to remove excess FeCN . Alternatively, a small amount of reducing agent (e.g. thiols) present in the assay may be reducing FMN and/or the cobalamin. The source of the reducing agent is unknown, but it may originate from dithionite on the surface of the gloves in the anaerobic glove box or surface-exposed thiols on the proteins themselves.
Control – no hMS
0.24 ± 0.02
FMN domain oxidizedb
0.66 ± 0.05
FMN domain 1e−
1.86 ± 0.17
FMN domain 2e−
2.65 ± 0.18
0.03 ± 0.02
1.11 ± 0.11
1.19 ± 0.09
1.61 ± 0.11
Oxidized MSR does not reactivate hMS (Table 2). In fact, the activity of hMS was found to be less than in the absence of any flavoprotein, which may be attributed to a tendency of oxidized MSR to withdraw reducing equivalents from hMS, thereby inhibiting the reactivation process. Alternatively, the binding of oxidized MSR to hMS may prevent an exogenous reducing agent from reducing hMS and returning it to the catalytic cycle. The turnover of hMS increases as MSR is reduced to the one-electron, two-electron and four-electron reduced states, reflecting a higher concentration of reducing equivalents needed to return hMS to the catalytic cycle.
Reduction of free AqCbl by diflavin reductases
MSR was shown previously to reduce AqCbl to cob(II)alamin, and this activity is thought to facilitate the uptake of cobalamin by the apo-form of hMS . We studied the ability of other diflavin reductase enzymes to reduce AqCbl and facilitate uptake of cobalamin by apo-MS. We found that CPR, nNOSred and MSR reduced AqCbl to cob(II)alamin (Table 3). CPR has the highest turnover number for NADPH-catalysed reduction of AqCbl (55.5 s−1), > 20-fold that of MSR (2.7 s−1) and ∼ 6-fold that of nNOSred (9.0 s−1). Calculated values for specificity constants (kcat/Km) reveal that CPR has the greatest specificity (10.6 × 105m−1·s−1) for AqCbl, with MSR (4.1 × 105m−1·s−1) and nNOSred (1.5 × 105m−1·s−1) working less effectively with this substrate. We demonstrated that the AqCbl reductase activity is dependent on the FMN domain, because the isolated NADP(H)/FAD domain of MSR was unable to reduce this cofactor directly. Thus, AqCbl can be likened to cytochrome c3+, in that it serves as a nonphysiological electron acceptor of diflavin reductases, but in doing so it takes electrons only from the FMN domain.
Table 3. Kinetic parameters obtained for the NADPH-catalysed reduction of AqCbl to cob(II)alamin. The rate of NADPH-catalysed reduction of AqCbl by MSR, CPR and nNOSred was measured by following the decrease in absorbance at 525 nm. Reactions were performed in 50 mm potassium phosphate buffer (pH 7.2), 85 μm NADPH, 5–20 × 1012 mol of enzyme, and variable concentrations of AqCbl, in a total volume of 1 mL, at 37 °C.
kcat/Km (× 105m−1·s−1)
2.7 ± 0.1
6.6 ± 0.4
4.1 ± 0.3
55.5 ± 1.3
52.4 ± 2.7
10.6 ± 0.6
9.0 ± 0.3
60.2 ± 4.0
1.5 ± 0.1
Holoenzyme synthase activity of diflavin reductases
The fact that CPR and nNOSred can enzymatically reduce AqCbl to cob(II)alamin prompted us to investigate whether these reductases can mimic MSR  by enhancing the uptake of cobalamin by apo-MS. It was previously shown that apo-MS generated by expression in insect cells or purified from rat liver is unstable at 37 °C [14,25]. In our studies, hMS activity dropped to 0.01 nmol·min−1 in Pichia cell extracts expressing apo-MS which were incubated at 37 °C for 70 min in the absence of cobalamin (Table 4). The addition of MSR, nNOSred or CPR to samples for which cobalamin was omitted had a negligible affect on activity (0.02–0.04 nmol·min−1; data not shown). The addition of AqCbl or MeCbl to the crude extract resulted in a small increase in activity (∼ 13 and ∼ 6-fold, respectively), which was not greatly affected by the addition of CPR or nNOSred. In contrast, the presence of MSR along with AqCbl or MeCbl caused a dramatic increase in hMS activity (7-fold for AqCbl and 20-fold for MeCbl) as compared to the cofactor alone. Similarly, the addition of the FMN domain of MSR along with AqCbl and MeCbl caused a 6-fold and 20-fold stimulation of hMS activity. The presence of NADPH in the preincubation mixture along with MSR and MeCbl/AqCbl did not have a significant effect in stimulating hMS activity.
Table 4. Holoenzyme synthase (chaperone) activity of diflavin reductases. Crude extract (150 μL) of P. pastoris expressing recombinant apo-MS was preincubated at 37 °C for 70 min, with and without the components indicated, in a total volume of 200 μL. Following incubation, the activity of hMS was measured by the radioactive assay described in Experimental procedures, using the AqCbl/dithiothreitol reducing system. The concentrations of the components were 200 nm MSR, MSR FMN domain, CPR or nNOSred, 50 μm MeCbl or AqCbl, and 200 μm NADPH.
hMS activity (nmol·min−1) 70 min
Without NADPH or MSR
0.01 ± 0.01
0.02 ± 0.01
With NADPH and MSR
0.02 ± 0.01
Without MSR or NADPH
0.13 ± 0.01
0.91 ± 0.14
With FMN domain
0.75 ± 0.02
With NADPH and MSR
1.12 ± 0.04
With NADPH and CPR
0.09 ± 0.02
With NADPH and nNOSred
0.18 ± 0.02
Without MSR or NADPH
0.06 ± 0.01
1.23 ± 0.15
With FMN domain
1.20 ± 0.10
With NADPH and MSR
1.19 ± 0.03
With NADPH and CPR
0.20 ± 0.01
with NADPH and nNOSred
0.19 ± 0.02
These results indicate that although uptake of cobalamin by apo-MS is not entirely dependent on MSR, the enzyme does greatly enhance the stability of the apoenzyme. During purification of hMS from Pichia, we measured hMS activity in crude extract by the AqCbl/dithiothreitol assay (see Experimental procedures), which does not contain MSR. Thus, the cofactor can be incorporated into apo-MS by a diffusive process. However, it is clear from Table 4 that MSR enhances the stability of apo-MS, and that this stabilization effect is strictly dependent on the presence of AqCbl or MeCbl. It is possible to infer from these data that MSR is eliciting a ‘holosynthase-like’ function. Our studies show that the mechanism by which MSR serves as a putative molecular chaperone for hMS does not rely on NADPH-catalysed reduction of exogenous AqCbl to cob(II)alamin: this follows because (a) hMS activity is not stimulated by the addition of NADPH to the preincubation mixture, (b) the FMN domain has a similar effect to that of full-length MSR in improving hMS stability, (c) MSR enhances hMS stability with both AqCbl and MeCbl, and (d) CPR and nNOSred are unable to affect hMS stability, despite having AqCbl reductase activity. Therefore, the incorporation of cobalamin mediated by MSR requires specific interaction between MSR and hMS, and in particular contact through the FMN domain, analogous to that for the hMS–MSR reactivation complex.
The sequestering of cobalamin between two partner proteins has been observed for in vitro formation of adenosylcobalamin by MSR and ATP:cobalamin adenosyltransferase (ATR) . In this system, MSR and ATR form a complex to sequester the highly reactive cob(I)alamin intermediate that is formed in the MSR-catalysed reduction of cob(II)alamin. The containment of the B12 cofactor within a protein complex potentially facilitates effective adenosylation of cob(I)alamin by ATR to form adenosylcobalamin.
Previously, we have shown that the addition of the hMS AD to the FMN domain or full-length MSR results in a quenching of the intrinsic flavin fluorescence, suggesting that the flavin chromophore is shielded from the solvent in the protein–protein complex . From the fluorescence titration assays, an apparent dissociation constant (Kd of 4.5 μm) for the complex was determined for the hMS AD–MSR complex, which closely mimics that of the MetH–Fld system (see Fig. S1 and Doc. S1) . Titration of CPR or nNOSred with the hMS AD did not result in a quenching of flavin fluorescence, confirming that these two proteins do not interact with hMS. We have compared the electrostatic potentials for the surface of the CPR FMN domain (in the region of the solvent-exposed FMN) with that of a homology model of the MSR FMN domain (based on the structure of the CPR FMN domain; Protein Data Bank: 1b1c) (see Doc. S1 and Fig. S2). The surface corresponding to the binding region for hMS AD is considerably less negatively charged in MSR than in the corresponding region of CPR. The electrostatic surface potential of the hMS AD (Protein Data Bank code: 202K) also contains relatively few charged groups near the S-adenosylmethionine-binding site (see Fig. S3 and Doc. S1). Thus, the propensity of hydrophobic residues on the putative binding interface of the MSR FMN domain suggests less emphasis on electrostatic interactions mediating hMS–MSR complex formation as compared to that of the CPR–P450 redox pair.
In conclusion, both FMNsq and FMNhq of MSR can participate in reductive methylation of hMS. MSR has a second important physiological function in facilitating uptake of cobalamin by hMS, a role that necessitates formation of an hMS–MSR complex. The latter finding is potentially important for future investigations into how polymorphic or clinical mutants of MSR manifest in disease states, such as hyperhomocysteinemia or megaloblastic anemia.
Hydroxycobalamin, MeCbl, AdoMet and homocysteine thiolactone were obtained from Sigma Chemical Company (Poole, UK). Restriction enzymes and T4 DNA ligase were from New England Biolabs (Hitchin, UK). Pfu Turbo DNA polymerase and XL1-blue competent cells were purchased from Stratagene (La Jolla, CA, USA). CH3-H4-folate and 5-[14C]methyl-H4-folate were obtained from Schircks Laboratories (Jona, Switzerland) and Amersham Biosciences UK Ltd (Chalfont St Giles, UK), respectively. Oligonucleotides were supplied by Invitrogen (Paisley, UK).
Heterologous expression of hMS in P. pastoris
The cloning and mutagenesis of the cDNA for hMS is described in Doc. S1. The sequences of the oligonucleotides used for cloning and mutagenesis of the hMS cDNA are listed in Tables S1–S2. The pPICZMS plasmid was digested with Pme1, and the linearized plasmid was transformed into P. pastoris strain SMD1168 by electrophoration, using the protocol outlined in the manual supplied by the commercial supplier of the strain (Invitrogen). Transformed colonies were selected on YPDS [1% (w/v) yeast extract, 2% (w/v) peptone, 1 m sorbitol, 2% (w/v) dextrose] plates containing 100 μg·mL−1 zeocin. Several transformed colonies were streaked onto plates containing 1000 μg·mL−1 zeocin to select for colonies containing multiple copies of the integrated cDNA for hMS. The fermentative growth of Pichia was adapted from the Invitrogen protocol (Pichia Fermentation Growth Guidelines; Invitrogen). Expression of recombinant hMS was obtained by first inoculating 5 mL of BMGY medium [1% (w/v) yeast extract, 0.5% (w/v) peptone, 100 mm potassium phosphate, pH 6.0, 1.34% yeast nitrogen base, 0.4 μg·mL−1 biotin, and 1% (w/v) glycerol] with a single transformed colony, and incubating the culture for 8 h at 30 °C with gentle aeration. The 5 mL culture was then used to inoculate 200 mL of BMGY medium, which was subsequently incubated for 16 h at 30 °C with gentle aeration. Fermentation of the Pichia culture was performed in a 7.5 L Bioflo 110 benchtop fermenter (New Brunswick Scientific, Edison, NJ, USA) equipped with microprocessor control of pH, dissolved oxygen, agitation, temperature, and nutrient feed, and with electronic foam control. The vessel contained 3.5 L of media comprising 0.93 g·L−1 CaSO4, 18.2 g·L−1 K2SO4, 14.9 g·L−1 MgSO4.7H2O, 26.7 mL of 85% phosphoric acid, 4.13 g of KOH, and 40 g·L−1 glycerol, along with 4.25 mL·L−1 of trace salts (PTM1; Invitrogen). The fermentation medium was inoculated with 200 mL of starter culture. Throughout growth, the temperature was maintained at 29 °C, and agitation was constant at 900 r.p.m. A pH of 5.0 was maintained using 14% (w/v) ammonium hydroxide. The glycerol batch phase was run until glycerol was completely consumed (∼ 22 h). During the second phase of growth (the ‘methanol–glycerol mix feed phase’), glycerol (containing 12 mL of PTM1 trace salts per litre) was added to the culture at 3.6 mL·h−1·L−1 of initial fermentation volume. After 1 h, methanol (containing 12 mL of PTM1 trace salts per litre) was added to the culture at 1.2 mL·h−1·L−1 of initial fermentation volume. After an additional 1 h, the methanol flow rate was increased to 2.4 mL·h−1·L−1 and the glycerol feed rate was decreased to 2.4 mL·h−1·L−1. Finally, after another 2 h, the methanol flow rate was increased to 3.6 mL·h−1·L−1 of initial fermentation volume, and the glycerol feed was terminated. The cells continued to grow on methanol supplied at 3.6 mL·h−1·L−1 of initial fermentation volume for a further 24 h. The cells were centrifuged at 3000 g for 10 min, and the wet cell pellet was frozen at −80 °C.
Purification of hMS
Human MS was purified by ion exchange and cobalamin affinity chromatography, following a modified protocol of Yamada et al. . Cobalamin–agarose was prepared according to the method of Sato et al. . All purification steps were performed on ice or at 4 °C, unless otherwise stated. Cells (103 g wet weight) were suspended in 250 mL of 50 mm potassium phosphate buffer (pH 7.2), containing 1 mm phenylmethanesulfonyl fluoride and two protease inhibitor tablets (Roche Products Ltd, Welwyn Garden City, UK). The cells were disrupted by passing the cell suspension twice through a cell disrupter (T-series Cabinet; Constant Systems, Daventry, UK) at 40 000 lb in−2. The cell debris was centrifuged at 40 000 g for 45 min. The supernatant was applied to a Q-Sepharose Fast Flow column (5 × 11 cm) equilibrated with 50 mm potassium phosphate buffer (pH 7.2). The protein was eluted with a linear gradient (0–0.5 m NaCl) at 2 mL·min−1. Fractions containing hMS activity were pooled and mixed with the cobalamin–agarose (12 mL) at 22 °C for 1 h. The mixture was then packed into a column (1.5 cm in diameter). The resin was washed with 50 mm potassium phosphate buffer (pH 7.2), followed by 50 mm potassium phosphate buffer (pH 7.2) containing 1 m NaCl, and then equilibrated in 10 mm Tris/HCl (pH 7.2). The resin, as a 50% slurry, was placed into a 25-mL beaker and exposed to light (halogen lamp; SCHOOT-KL1500 LCD set at 3300 K for 15 min on ice). The slurry was then loaded into a column (1.5 cm in diameter), and the resin was washed with 10 mm Tris/HCl (pH 7.2). Human MS was eluted with 10 mm Tris/HCl (pH 7.2) containing 0.5 m NaCl, and then dialysed against 50 mm potassium phosphate buffer (pH 7.2) for 16 h. Purified hMS was concentrated and stored at −80 °C. Protein concentrations at various purification steps were determined using the Bio-Rad (Hercules, CA, USA) protein assay kit, using BSA as a standard. Purification of human CPR , rat nNOSred  and human MSR  followed previously published protocols.
Nonradioactive hMS activity assay
The activity of hMS during purification was measured using a discontinuous spectrophotometric assay in which the product CH3-H4-folate is converted to CH+=H4-folate . The reaction mixture contained 0.1 m potassium phosphate buffer (pH 7.2), 0.1 m KCl, 250 μm CH3-H4-folate, 1 mm homocysteine, 100 μm AdoMet, 25 mm dithiothreitol, and 50 μm AqCbl. The total reaction volume was 0.8 mL. The reactions were initiated by the addition of homocysteine following 5 min of incubation at 37 °C of the enzyme with all other components. Following 10 min of incubation at 37 °C, the reaction was quenched by the addition of 0.2 mL of a solution containing 11 m formic acid and 5 m HCl, and heated to 90 °C for 10 min. The acidification of the reaction mixture quantitatively converts CH3-H4-folate to CH+=H4-folate, which absorbs strongly at 350 nm (Δε = 26 500 m−1 cm−1).
Radioactive hMS activity assay
A radioactive hMS assay that monitors the transfer of the [14C]methyl group from CH3-H4-folate to the product methionine was adapted from a published protocol . The assay mixture comprised 0.2 m potassium phosphate buffer (pH 7.2), 100 μm AdoMet, 1 mm homocysteine, 50 μm AqCbl, 25 mm dithiothreitol, and 250 μm14CH3-H4-folate (1200 d.p.m. per nmol), in a total volume of 250 μL. All the components were mixed and incubated at 37 °C for 3 min. The reaction was initiated by the addition of 1 mm homocysteine, and then further incubated at 37 °C for 10 min. The reaction was quenched at 90 °C for 3 min, and the reaction mixture was then cooled to room temperature before being applied to a 2 mL (0.8 × 4 cm) AG-1 (Bio-Rad) column. The column was washed with 2 × 1 mL of water. The combined radioactivity in the flowthrough and wash fractions was quantitated by scintillation counting. For MSR-catalysed assays (used for measuring the Km of NADPH and the Kact of MSR), AqCbl and dithiothreitol were replaced by varying concentrations of MSR and NADPH.
Anaerobic hMS reactivation assay
The anaerobic hMS assays were performed in a Belle Technology glove box (O2 ≪ 5 p.p.m.) equipped with an Hitatchi U-1800 spectrophotometer. All buffers and reaction mixtures were extensively bubbled with nitrogen prior to introduction into the glove box. A concentrated sample of MSR or the isolated FMN domain was introduced into the glove box, and FeCN was added to the concentrated protein stock to fully oxidize the flavin cofactors. To remove excess FeCN and O2, MSR and the isolated FMN domain were gel filtered using a 10 mL Econo-pack column (Bio-Rad) equilibrated with anaerobic buffer (10 mm potassium phosphate, pH 7.2). The enzymes were reduced to the 1, 2 and 4 (in the case of full-length MSR) reduced states by titration with dithionite. The UV–visible spectra of the flavoproteins were recorded with sequential addition of dithionite. The various reduced forms of MSR, or the FMN domain (40 μm), were added to an assay mixture containing 0.2 m potassium phosphate buffer (pH 7.2), 100 μm AdoMet, 1 mm homocysteine, 250 μm14CH3-H4-folate (1200 d.p.m. per nmol), and hMS, in a total volume of 250 μL. The reaction was incubated for 10 min at 37 °C, and quenched and analysed following the protocol for the radioactive hMS activity assay. The concentration of MSR and the FMN domain were determined by the absorbance value at 450 nm, using extinction coefficients of 25 600 and 14 700 m−1·cm−1, respectively .
Measurement of cobalamin reductase activity
The rates of NADPH-catalysed reduction of AqCbl by MSR, human CPR and nNOSred were measured by following the decrease in absorbance at 525 nm using a difference extinction coefficient of 5.57 × 10−3m−1·cm−1  on a Cary50 spectrophotometer. Reactions were performed in 50 mm potassium phosphate buffer (pH 7.2), 85 μm NADPH and variable concentrations of AqCbl, at 37 °C, in a total volume of 1 mL. The reaction was initiated by adding 5–20 × 1012 mol of enzyme. The concentrations of CPR and nNOSred were determined by the absorbance value at 450 nm, using extinction coefficients of 22 000 and 21 600 m−1·cm−1, respectively [29,32].
Measurement of holo-MS synthase activity
Holo-MS synthase activity was measured following a previously published protocol . Pichia cells expressing recombinant hMS were disrupted, and the crude extract (2 mL) was applied to a 10 mL gel filtration column to remove small molecules. The filtered extract was then incubated for 70 min at 37 °C in the presence or absence (as noted) of NADPH, MSR, FMN domain of MSR, CPR, nNOSred, AqCbl and MeCbl. The activity of the holo-MS was then measured by the AqCbl/dithiothreitol radioactive assay described above.
This study was funded by the UK Biotechnology and Biological Sciences Research Council. N.S. Scrutton is a BBSRC Professorial Research Fellow. We thank K. Marshall for assistance with early parts of the cloning work reported in the article.