AmericanChemicalSociety.com

During the reporting period of this grant the mechanism of a unique sulfur reductase was extensively characterized, including the structure, the nature of the sulfur species which the enzyme reacts with, and much of the complex mechanism of the reduction of the newly discovered persulfide substrate.  We had previously shown that polysulfide is a substrate for the enzyme, and our structure showed the importance of the thiol of a bound coenzyme A in the mechanism of sulfur reduction.  We have shown that the likely oxidizing substrate of this enzyme during the reduction of sulfur in the environment is not elemental sulfur but rather the persulfides of small thiols such as coenzyme A and glutathione; the enzyme is remarkably promiscuous with regard to the identity of the persulfide donor.  Site-directed mutagenesis was performed to mutate the cysteines that we believed to be essential in the mechanism (C43S, C531S and C43,531S mutants were created), and it was shown that cysteine 43 is absolutely required for the catalysis of any S-S cleavage, while cysteine 531 is absolutely required for the reduction of R-S-S-H substrates but not for the reduction of disulfide (R-S-S-R) substrates.  Titrations of the enzyme with NADH, dithionite, titanium(III) or TCEP (tris-(2-carboxyethyl)phosphine) demonstrated the presence of a mixed-disulfide between C43 and a tightly bound CoA, and structures of the C43 and C43,531S mutants confirmed that this coenzyme A remained tightly bound to the enzyme in the absence of a C43-CoA S-S bond.  

The structure of Npsr suggested a likely site for binding and reaction with the persulfide or polysulfide substrates on the surface of the protein at the location of C531 – this cysteine is proposed to attack the persulfide sulfur, which could be held in place in an obvious binding pocket adjacent to the Cys531 sulfur, which is occupied by a chloride ion in our structure.  An obvious “tunnel” leads from cysteine 531 and the persulfide binding pocket on the surface to the main active site deep in the interior of the protein, where C43, the FAD cofactor, and the NADH binding site are all located.  These results are consistent with a mechanism in which the long pantothenate portion of the coenzyme A molecule acts as a delivery arm, shuttling sulfur from its initial binding site on the surface to the buried active site via the observed tunnel, where the persulfide substrate is reduced to the sulfide product.  It is particularly interesting to note how the tightly bound coenzyme A thiol of this enzyme appears to fulfill the role played by the second active site cysteine present in the closely related glutathione reductase family of enzymes, turning the apparently single-cysteine active site of this sulfur/persulfide reductase into an active site more analogous to that of the two-cysteine active site of related enzymes.

From the structure of the enzyme we have proposed a likely reaction cycle for NADH oxidation and sulfur reduction in this system.  In order to test this mechanism and to determine the microscopic rate constants for the overall reaction we determined the primary kinetic isotope effect on the steady-state kinetic parameters for the enzyme using NADH deuterated at the position from which hydride transfer occurs.  An isotope effect of 1.6 on V (the apparent k_cat) was observed, suggesting that NADH oxidation was at least partially rate-limiting in the overall reaction.  The NADH-dependent reductive half reaction of the enzyme was further examined via stopped-flow spectroscopy, which showed that reduction occurs in at least three phases, corresponding to 1) NADH binding, 2) formation of an initial non-productive complex via a pre-equilibrium (a step which shows a small normal to large inverse primary kinetic isotope effect, depending on the [NADH]) and 3) hydride transfer from NADH to FAD.  As expected, the 3^rd phase showed a normal primary kinetic isotope effect of 4.5 with ²H NADH, decreasing from 27.0 s^-1 to 6.5 s^-1, consistent with the results observed in steady-state experiments, where the k_cat of 10.8 s^-1 with the NADH substrate, in which a 27.0 s^-1 reductive step would not be rate-limiting, is reduced to an apparent k_cat of 6.8 s^-1 with ²H-NADH, with the reductive step becoming completely rate-limiting.

In order to more fully characterize the reductive half-reaction, and especially the second phase of the reduction, in which an apparent pre-equilibrium occurs between a non-productive and productive ES complex, an additional mutant of the enzyme was constructed.  Phenylalanine residue 159, which has a large phenyl side chain, was replaced with an alanine, which has a much smaller methyl side chain.  F159 is located in the NADH binding pocket and is believed to be important in positioning the NADH nicotinamide ring in the proper configuration next to the FAD to allow for facile electron transfer.  Much to our surprise, the F159A mutant enzyme showed only a very small decrease in k_cat  from 10.8 s^-1 to 9.0 s^-1, and a modest increase in the K_m of 10.5 to 31.3 uM.  When the isotope effect of ²H-NADH on V determined, it had increased from the value of 1.6 with the wild-type enzyme to a value of 4.0 on the mutant enzyme, consistent with the reductive half-reaction being fully rate limiting on the F159A mutant enzyme.  This observation was confirmed by pre-steady state studies, which showed that the observed rate of the hydride transfer step was significantly decreased on the mutant enzyme.  We are currently using kinetic models to determine a reasonable mechanistic explanation for the large inverse isotope effect observed on the second phase of the reductive half reaction with both the wild-type and F159A mutants of the enzyme.