The NifZ accessory protein has an equivalent function in maturation of both nitrogenase MoFe protein P-clusters.

The Mo-dependent nitrogenase comprises two interacting components called the Fe protein and the MoFe protein. The MoFe protein is an α2ß2 heterotetramer that harbors two types of complex metalloclusters, both of which are necessary for N2 reduction. One type is a 7Fe-9S-Mo-C-homocitrate species designated FeMo-cofactor, which provides the N2-binding catalytic site, and the other is an 8Fe-7S species designated the P-cluster, involved in mediating intercomponent electron transfer to FeMo-cofactor. The MoFe protein's catalytic partner, Fe protein, is also required for both FeMo-cofactor formation and the conversion of an immature form of P-clusters to the mature species. This latter process involves several assembly factors, NafH, NifW, and NifZ, and precedes FeMo-cofactor insertion. Here, using various protein affinity-based purification methods as well as in vivo, EPR spectroscopy, and MALDI measurements, we show that several MoFe protein species accumulate in a NifZ-deficient background of the nitrogen-fixing microbe Azotobacter vinelandii These included fully active MoFe protein replete with FeMo-cofactor and mature P-cluster, inactive MoFe protein having no FeMo-cofactor and only immature P-cluster, and partially active MoFe protein having one αß-unit with a FeMo-cofactor and mature P-cluster and the other αß-unit with no FeMo-cofactor and immature P-cluster. Also, NifW could associate with MoFe protein having immature P-clusters and became dissociated upon P-cluster maturation. Furthermore, both P-clusters could mature in vitro without NifZ. These findings indicate that NifZ has an equivalent, although not essential, function in the maturation of both P-clusters contained within the MoFe protein.

Light-enhanced bioaccumulation of molybdenum by nitrogen-deprived recombinant anoxygenic photosynthetic bacterium Rhodopseudomonas palustris.

As molybdenum (Mo) is an indispensable metal for plant nitrogen metabolisms, accumulation of dissolved Mo into bacterial cells may connect to the development of bacterial fertilizers that promote plant growth. In order to enhance Mo bioaccumulation, nitrogen removal and light illumination were examined in anoxygenic photosynthetic bacteria (APB) because APB possess Mo nitrogenase whose synthesis is strictly regulated by ammonium ion concentration. In addition, an APB, Rhodopseudomonas palustris, transformed with a gene encoding Mo-responsive transcriptional regulator ModE was constructed. Mo content was most markedly enhanced by the removal of ammonium ion from medium and light illumination while their effects on other metal contents were limited. Increases in contents of trace metals including Mo by the genetic modification were observed. Thus, these results demonstrated an effective way to enrich Mo in the bacterial cells by the culture conditions and genetic modification.

Another role for CO with nitrogenase? CO stimulates hydrogen evolution catalyzed by variant Azotobacter vinelandii Mo-nitrogenases.

A likely entry/exit path for nitrogenase substrates, products, and/or protons involves residues α277(Arg), α192(Ser), and α356(Gly), all of which are highly conserved among MoFe proteins from different organisms. The α192(Ser) and α277(Arg) residues form part of a hydrogen-bonded network that also involves α195(His), which interacts with a FeMo cofactor-based sulfide. The terminal amino groups of α277(Arg) are also hydrogen-bonded directly to α281(Tyr), which resides at the surface of the MoFe protein. Individual amino acid substitutions placed at position α277 or α192 resulted in a variety of effects on the catalytic and/or spectroscopic properties of the resulting variant MoFe protein. Of particular interest was the effect of CO on H2 evolution catalyzed by three MoFe protein variants, α277(Cys), α192(Asp), and α192(Glu). All three variants exhibited CO stimulation of H2 evolution under high-electron flux conditions but not under low-electron flux conditions. This observation is best explained by these variants being redox-compromised but only at the most reduced redox states of the MoFe protein. Normally, these states are accessed and operational only under high-electron flux conditions, and the effect of added CO is to prevent access to these most reduced redox states, resulting in a normal rate of catalysis. Furthermore, via correlation of the effect of pH changes on H2 evolution activity for both the wild type and the α277(Cys) MoFe protein variant under argon, with or without 10% CO present, likely pathways for the delivery of a proton to the FeMo cofactor were identified.

P+ state of nitrogenase p-cluster exhibits electronic structure of a [Fe4S4]+ cluster.

Mo nitrogenase consists of two component proteins: the Fe protein, which contains a [Fe(4)S(4)] cluster, and the MoFe protein, which contains two different classes of metal cluster: P-cluster ([Fe(8)S(7)]) and FeMoco ([MoFe(7)S(9)C·homocitrate]). The P-cluster is believed to mediate the electron transfer between the Fe protein and the MoFe protein via interconversions between its various oxidation states, such as the all-ferrous state (P(N)) and the one- (P(+)) and two-electron (P(2+)) oxidized states. While the structural and electronic properties of P(N) and P(2+) states have been well characterized, little is known about the electronic structure of the P(+) state. Here, a mutant strain of Azotobacter vinelandii (DJ1193) was used to facilitate the characterization of the P(+) state of P-cluster. This strain expresses a MoFe protein variant (designated ΔnifB ß-188(Cys) MoFe protein) that accumulates the P(+) form of P-cluster in the resting state. Magnetic circular dichroism (MCD) spectrum of the P-cluster in the oxidized ΔnifB ß-188(Cys) MoFe protein closely resembles that of the P(2+) state in the oxidized wild-type MoFe protein, except for the absence of a major charge-transfer band centered at 823 nm. Moreover, magnetization curves of ΔnifB ß-188(Cys) and wild-type MoFe proteins suggest that the P(2+) species in both proteins have the same spin state. MCD spectrum of the P(+) state in the ΔnifB ß-188(Cys) MoFe protein, on the other hand, is associated with a classic [Fe(4)S(4)](+) cluster, suggesting that the P-cluster could be viewed as two coupled 4Fe clusters and that it could donate either one or two electrons to FeMoco by using one or both of its 4Fe halves. Such a mode of action of P-cluster could provide energetic and kinetic advantages to nitrogenase in the complex mechanism of N(2) reduction.

ATP-uncoupled, six-electron photoreduction of hydrogen cyanide to methane by the molybdenum-iron protein.

A detailed study of the eight-electron/eight-proton catalytic reaction of nitrogenase has been hampered by the fact that electron and proton flow in this system is controlled by ATP-dependent protein-protein interactions. Recent studies have shown that it is possible to circumvent the dependence on ATP through the use of potent small-molecule reductants or light-driven electron injection, but success has been limited to two-electron reductions of hydrazine, acetylene, or protons. Here we show that a variant of the molybdenum-iron protein labeled with a Ru-photosensitizer can support the light-driven, six-electron catalytic reduction of hydrogen cyanide into methane and likely also ammonia. Our findings suggest that the efficiency of this light-driven system is limited by the initial one- or two-electron reduction of the catalytic cofactor (FeMoco) to enable substrate binding, but the subsequent electron-transfer steps into the FeMoco-bound substrate proceed efficiently.

Catalytic activities of NifEN: implications for nitrogenase evolution and mechanism.

NifEN is a key player in the biosynthesis of nitrogenase MoFe protein. It not only shares a considerable degree of sequence homology with the MoFe protein, but also contains clusters that are homologous to those found in the MoFe protein. Here we present an investigation of the catalytic activities of NifEN. Our data show that NifEN is catalytically competent in acetylene (C(2)H(2)) and azide (N(3)(-)) reduction, yet unable to reduce dinitrogen (N(2)) or evolve hydrogen (H(2)). Upon turnover, C(2)H(2) gives rise to an additional S = 1/2 signal, whereas N(3)(-) perturbs the signal originating from the NifEN-associated FeMoco homolog. Combined biochemical and spectroscopic studies reveal that N(3)(-) can act as either an inhibitor or an activator for the binding and/or reduction of C(2)H(2), while carbon monoxide (CO) is a potent inhibitor for the binding and/or reduction of both N(3)(-) and C(2)H(2). Taken together, our results suggest that NifEN is a catalytic homolog of MoFe protein; however, it is only a "skeleton" version of the MoFe protein, as its associated clusters are simpler in structure and less versatile in function, which, in turn, may account for its narrower range of substrates and lower activities of substrate reduction. The resemblance of NifEN to MoFe protein in catalysis points to a plausible, sequential appearance of the two proteins in nitrogenase evolution. More importantly, the discrepancy between the two systems may provide useful insights into nitrogenase mechanism and allow reconstruction of a fully functional nitrogenase from the "skeleton" enzyme, NifEN.

Functional expression of the nitrogenase Fe protein in transgenic rice.

Engineering cereals to express functional nitrogenase is a long-term goal of plant biotechnology and would permit partial or total replacement of synthetic N fertilizers by metabolization of atmospheric N2. Developing this technology is hindered by the genetic and biochemical complexity of nitrogenase biosynthesis. Nitrogenase and many of the accessory proteins involved in its assembly and function are O2 sensitive and only sparingly soluble in non-native hosts. We generated transgenic rice plants expressing the nitrogenase structural component, Fe protein (NifH), which carries a [4Fe-4S] cluster in its active form. NifH from Hydrogenobacter thermophilus was targeted to mitochondria together with the putative peptidyl prolyl cis-trans isomerase NifM from Azotobacter vinelandii to assist in NifH polypeptide folding. The isolated NifH was partially active in electron transfer to the MoFe protein nitrogenase component (NifDK) and in the biosynthesis of the nitrogenase iron-molybdenum cofactor (FeMo-co), two fundamental roles for NifH in N2 fixation. NifH functionality was, however, limited by poor [4Fe-4S] cluster occupancy, highlighting the importance of in vivo [Fe-S] cluster insertion and stability to achieve biological N2 fixation in planta. Nevertheless, the expression and activity of a nitrogenase component in rice plants represents the first major step to engineer functional nitrogenase in cereal crops.

Differential accumulation of nif structural gene mRNA in Azotobacter vinelandii.

Northern analysis was employed to investigate mRNA produced by mutant strains of Azotobacter vinelandii with defined deletions in the nif structural genes and in the intergenic noncoding regions. The results indicate that intergenic RNA secondary structures effect the differential accumulation of transcripts, supporting the high Fe protein-to-MoFe protein ratio required for optimal diazotrophic growth.

Molybdenum cofactor deficiency: Mutations in GPHN, MOCS1, and MOCS2.

All molybdenum-containing enzymes other than the bacterial nitrogenase share an identical molybdenum cofactor (MoCo), which is synthesized via a conserved pathway in all organisms and therefore also is called "universal molybdenum cofactor." In humans, four molybdoenzymes are known: aldehyde oxidase, mitochondrial amidoxime reducing component (mARC), xanthine oxidoreductase, and sulfite oxidase. Mutations in the genes encoding the biosynthetic MoCo pathway enzymes abrogate the activities of all molybdoenzymes and result in the "combined" form of MoCo deficiency, which is clinically very similar to isolated sulfite oxidase deficiency, caused by mutations in the gene for the corresponding apoenzyme. Both deficiencies are inherited as an autosomal-recessive disease and result in progressive neurological damage and early childhood death in most cases. The majority of mutations leading to MoCo deficiency have been identified in the genes MOCS1 (type A deficiency), MOCS2 (type B deficiency), with one reported in GPHN. For type A deficiency an effective substitution therapy has been described recently.

Chronological changes of the amplitude-integrated EEG in a neonate with molybdenum cofactor deficiency.

Molybdenum cofactor (Moco) deficiency is a rare neurometabolic disorder, characterized by neurological impairment and refractive seizures, due to toxic accumulation of sulfite in the brain. Earlier it was suggested that in Moco-deficient humans maternal clearance of neurotoxic metabolites prevents prenatal brain damage. However, limited data are available about the time profile in which neurophysiologic deterioration occurs after birth. The amplitude-integrated electroencephalography (aEEG) is a bedside method in neonates to monitor cerebral recovery after hypoxic-ischemic insults, detect epileptic activity, and evaluate antiepileptic drug treatment. We describe a chronological series of changes in aEEG tracings in a neonate with Moco deficiency. He presented with myoclonic spasms and hypertonicity a few hours after birth, however, the aEEG pattern was still normal. Within 2 days, the aEEG rapidly changed into a burst suppression pattern with repetitive seizures. After antiepileptic treatment, the aEEG remained abnormal. In this patient, the normal aEEG pattern at birth may have been due to maternal clearance of sulfite in utero. After birth, accumulation of sulfite causes progressive brain damage, reflected by the progressive depression of the aEEG tracings. This is in agreement with the results from a Moco-deficient mouse model, suggesting that maternal sulfite clearance suppresses prenatal brain damage. To our knowledge, this is the first case report describing the chronological changes in the aEEG pattern in a Moco-deficient patient. Insight into the time profile in which neurologic deterioration in Moco-deficient humans occurs is essential, especially when potential treatment strategies are being evaluated.

VTVH-MCD study of the Delta nifB Delta nifZ MoFe protein from Azotobacter vinelandii.

NifZ is a member of a series of proteins associated with the maturation of the nitrogenase MoFe protein. An MCD spectroscopic study was undertaken on the Delta nifB Delta nifZ MoFe protein generated in the absence of both NifZ and NifB (deletion of NifB generates an apo-MoFe protein lacking the FeMo cofactor). Results presented here show that, in the absence of NifZ, only one of the two P-clusters of the MoFe protein is matured to the ultimate [8Fe-7S] structure. The other P-cluster site in the protein contains a [4Fe-4S] cluster pair, representing a P-cluster precursor that is electronically identical to the analogous clusters observed in the Delta nifH MoFe protein. These results suggest that the MoFe protein is synthesized in a stepwise fashion where NifZ is specifically required for the formation of the second P-cluster.

Optimization of FeMoco maturation on NifEN.

Mo-nitrogenase catalyzes the reduction of dinitrogen to ammonia at the cofactor (i.e., FeMoco) site of its MoFe protein component. Biosynthesis of FeMoco involves NifEN, a scaffold protein that hosts the maturation of a precursor to a mature FeMoco before it is delivered to the target location in the MoFe protein. Previously, we have shown that the NifEN-bound precursor could be converted in vitro to a fully complemented "FeMoco" in the presence of 2 mM dithionite. However, such a conversion was incomplete, and Mo was only loosely associated with the NifEN-bound "FeMoco". Here we report the optimized maturation of the NifEN-associated precursor in 20 mM dithionite. Activity analyses show that upon the optimal conversion of precursor to "FeMoco", NifEN is capable of activating a FeMoco-deficient form of MoFe protein to the same extent as the isolated FeMoco. Furthermore, EPR and XAS/EXAFS analyses reveal the presence of a tightly organized Mo site in NifEN-bound "FeMoco", which allows the observation of a FeMoco-like S = 3/2 EPR signal and the modeling of a NifEN-bound "FeMoco" that adopts a conformation very similar to that of the MoFe protein-associated FeMoco. The sensitivity of FeMoco maturation to dithionite concentration suggests an essential role of redox chemistry in this process, and the optimal potential of dithionite solution could serve as a guideline for future identification of in vivo electron donors for FeMoco maturation.

A substrate channel in the nitrogenase MoFe protein.

Nitrogenase catalyzes the six electron/six proton reduction of N2 to two ammonia molecules at a complex organometallocluster called "FeMo cofactor." This cofactor is buried within the alpha-subunit of the MoFe protein, with no obvious access for substrates. Examination of high-resolution X-ray crystal structures of MoFe proteins from several organisms has revealed the existence of a water-filled channel that extends from the solvent-exposed surface to a specific face of FeMo cofactor. This channel could provide a pathway for substrate and product access to the active site. In the present work, we examine this possibility by substituting four different amino acids that line the channel with other residues and analyze the impact of these substitutions on substrate reduction kinetic parameters. Each of the MoFe protein variants was purified and kinetic parameters were established for the reduction of the substrates N2, acetylene, azide, and propyne. For each MoFe protein, V (max) values for the different substrates were found to be nearly unchanged when compared with the values for the wild-type MoFe protein, indicating that electron delivery to the active site is not compromised by the various substitutions. In contrast, the K(m) values for these substrates were found to increase significantly (up to 22-fold) in some of the MoFe protein variants compared with the wild-type MoFe protein values. Given that each of the amino acids that were substituted is remote from the active site, these results are consistent with the water-filled channel functioning as a substrate channel in the MoFe protein.

Alkyne substrate interaction within the nitrogenase MoFe protein.

Nitrogenase catalyzes the biological reduction of N(2) to ammonia (nitrogen fixation), as well as the two-electron reduction of the non-physiological alkyne substrate acetylene (HC triple bond CH). A complex metallo-organic species called FeMo-cofactor provides the site of substrate reduction within the MoFe protein, but exactly where and how substrates interact with FeMo-cofactor remains unknown. Recent results have shown that the MoFe protein alpha-70(Val) residue, whose side chain approaches one Fe-S face of FeMo-cofactor, plays a significant role in defining substrate access to the active site. For example, substitution of alpha-70(Val) by alanine results in an increased capacity for the reduction of the larger alkyne propyne (HC triple bond C-CH(3)), whereas, substitution by isoleucine at this position nearly eliminates the capacity for the reduction of acetylene. These and complementary spectroscopic studies led us to propose that binding of short chain alkynes occurs with side-on binding to Fe atom 6 within FeMo-cofactor. In the present work, the alpha-70(Val) residue was substituted by glycine and this MoFe protein variant shows an increased capacity for reduction of the terminal alkyne, 1-butyne (HC triple bond C-CH(2)-CH(3)). This protein shows no detectable reduction of the internal alkyne 2-butyne (H(3)C-C triple bond C-CH(3)). In contrast, substitution of the nearby alpha-191(Gln) residue by alanine, in combination with the alpha-70(Ala) substitution, does result in significant reduction of 2-butyne, with the exclusive product being 2-cis-butene. These results indicate that the reduction of alkynes by nitrogenases involves side-on binding of the alkyne to Fe6 within FeMo-cofactor, and that a terminal acidic proton is not required for reduction. The successful design of amino acid substitutions that permit the targeted accommodation of an alkyne that otherwise is not a nitrogenase substrate provides evidence to support the current model for alkyne interaction within the nitrogenase MoFe protein.

Molecular insights into nitrogenase FeMoco insertion--the role of His 274 and His 451 of MoFe protein alpha subunit.

The final step of FeMo cofactor (FeMoco) assembly involves the insertion of FeMoco into its binding site in the molybdenum-iron (MoFe) protein of nitrogenase. Here we examine the role of His alpha274 and His alpha451 of Azotobacter vinelandii MoFe protein in this process. Our results from combined metal, activity, EPR, stability and insertion analyses show that mutations of His alpha274 and/or His alpha451, two of the histidines that belong to a so-called His triad, to small uncharged Ala specifically reduce the accumulation of FeMoco in MoFe protein. This observation indicates that the enrichment of histidines at the His triad is important for FeMoco insertion and that the His triad potentially serves as an intermediate docking point for FeMoco through transitory ligand coordination and/or electrostatic interaction.

Molybdenum cofactor and isolated sulphite oxidase deficiencies: Clinical and molecular spectrum among Egyptian patients.

AIM: Molybdenum cofactor deficiency (MoCD) and Sulfite oxidase deficiency (SOD) are rare autosomal recessive conditions of sulfur-containing amino acid metabolism with overlapping clinical features and emerging therapies. The clinical phenotype is indistinguishable and they can only be differentiated biochemically. MOCS1, MOCS2, MOCS3, and GPRN genes contribute to the synthesis of molybdenum cofactor, and SUOX gene encodes sulfite oxidase. The aim of this study was to elucidate the clinical, radiological, biochemical and molecular findings in patients with SOD and MoCD. METHODS: Detailed clinical and radiological assessment of 9 cases referred for neonatal encephalopathy with hypotonia, microcephaly, and epilepsy led to a consideration of disorders of sulfur-containing amino acid metabolism. The diagnosis of six with MoCD and three with SOD was confirmed by biochemical tests, targeted sequencing, and whole exome sequencing where suspicion of disease was lower. RESULTS: Novel SUOX mutations were detected in 3 SOD cases and a novel MOCS2 mutation in 1 MoCD case. Most patients presented in the first 3 months of life with intractable tonic-clonic seizures, axial hypotonia, limb hypertonia, exaggerated startle response, feeding difficulties, and progressive cystic encephalomalacia on brain imaging. A single patient with MoCD had hypertrophic cardiomyopathy, hitherto unreported with these diseases. INTERPRETATION: Our results emphasize that intractable neonatal seizures, spasticity, and feeding difficulties can be important early signs for these disorders. Progressive microcephaly, intellectual disability and specific brain imaging findings in the first year were additional diagnostic aids. These clinical cues can be used to minimize delays in diagnosis, especially since promising treatments are emerging for MoCD type A.

Analysis of the nifHDK operon and structure of the NifH protein from the unicellular, diazotrophic cyanobacterium, Cyanothece strain sp. ATCC 51142(1).

Cyanothece sp. ATCC 51142 is a unicellular, diazotrophic cyanobacterium that demonstrates diurnal rhythms for photosynthesis and N(2) fixation, with peaks of O(2) evolution and nitrogenase activity approximately 12 h out of phase. We cloned and sequenced the nifHDK operon, and determined that the amino acid sequences of all three proteins were highly conserved relative to those of other cyanobacteria and bacteria. However, the Fe-protein, encoded by the nifH gene, demonstrated two differences from the related protein in Azotobacter vinelandii, for which a 3-D structure has been determined. First, the Cyanothece Fe-protein contained a 37 amino acid extension at the N-terminus. This approximately 4 kDa addition to the protein appeared to fold as a separate domain, but remained a part of the active protein, as was verified by migration on acrylamide gels. In addition, the Cyanothece Fe-protein had amino acid differences at positions involved in formation of the Fe-protein dimer-dimer contacts in A. vinelandii nitrogenase. There were also changes in residues involved with interaction between the Fe-protein and the MoFe-protein when compared with A. vinelandii. Since the Cyanothece Fe-protein is quickly degraded after activity, it is suggested that the extension and the amino acid alterations were somehow involved in this degradative process.

Mapping the interaction of the [2Fe-2S] Clostridium pasteurianum ferredoxin with nitrogenase MoFe protein.

The [2Fe-2S] ferredoxin from Clostridium pasteurianum had previously been shown to interact specifically with the nitrogenase MoFe protein, and electrostatic forces were found to be important contributors to the interaction. This phenomenon has now been analyzed in detail by using ferredoxin variants in which charge inversions or cancellations were introduced on all charged residues. The mutated forms of the ferredoxin were covalently cross-linked to the MoFe protein. The reaction products were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and their nitrogenase activity was measured. The latter displayed a consistent inverse correlation with the amount of cross-linked MoFe protein. The data allowed an unambiguous identification of the ferredoxin residues (glutamates 31, 34, 38, 39, 84, 85) that are involved in the interaction with the MoFe protein. Furthermore, whereas the wild-type ferredoxin yielded approximately equal amounts of cross-linked products with the alpha and beta subunits of the MoFe protein, some of its molecular variants displayed a differential decrease of reactivity towards one or the other of these subunits. The positions on the ferredoxin molecule of the residues interacting with the MoFe protein were determined using the recently elucidated crystal structure of the homologous [2Fe-2S] ferredoxin from Aquifex aeolicus.

Interaction with magnesium and ADP stabilizes both components of nitrogenase from Klebsiella pneumoniae against urea denaturation.

The nitrogenase enzyme of Klebsiella pneumoniae consists of two separable proteins, each with multiple subunits and one or more oxygen sensitive metallocenters. The wild-type nitrogenase proteins are stable to electrophoresis in high concentrations of urea under anaerobic conditions. Addition of Mg+2 and ADP greatly increases the stability of the smaller Fe protein (from <4 to >6 M for full unfolding), an effect directly analogous to stabilization in p21ras induced by Mg+2 and GDP. Stabilization by Mg+2 is slight for the holo MoFe protein (from approximately 1.5 to approximately 2.4 M) but more dramatic for the apo protein form of the MoFe protein accumulated by certain Fe protein (nifH gene) mutants. The potent product inhibitor of nitrogenase function, MgADP, increases stability of the MoFe protein more than Mg+2 alone, to approximately 3.6 M, showing that nucleotides interact with the MoFe protein. Mutations of the nifM gene result in slower accumulation of less stable Fe protein, indicating that NifM is involved in correct folding of the Fe protein. Mutationally altered proteins are often difficult to purify for study because of their inherent instability, low expression level, or oxygen lability. Crude extracts of 11 different mutants of Fe protein (nifH gene) were examined by transverse urea gradient gels to rapidly screen for stabilizing interactions in the presence or absence of substrate or inhibitor analogs. Amino acid alterations D44N and R188C, at the interface of the dimer, in the vicinity of the nucleotide binding site(s), have significantly lower stability than the wild-type enzyme in the absence of Mg+2 but comparable stability in its presence, showing the importance of Mg+2 in the subunit interactions. Mutations N163S and E266K, in which residues normally involved in hydrogen bonding far from the active site were altered, are more labile than the wild-type even with Mg+2 added. Seven other mutants, though nonfunctional, did not appear altered in stability compared to the wild-type.