Pyridoxal phosphate site in glycogen phosphorylase b: structure in native enzyme and in three derivatives with modified cofactors

The regulation of glycogenolysis in the brain

The key regulatory enzymes of glycogenolysis are phosphorylase kinase, a hetero-oligomer with four different types of subunits, and glycogen phosphorylase, a homodimer. Both enzymes are activated by phosphorylation and small ligands, and both enzymes have distinct isoforms that are predominantly expressed in muscle, liver, or brain; however, whole-transcriptome high-throughput sequencing analyses show that in brain both of these enzymes are likely composed of subunit isoforms representing all three tissues. This Minireview examines the regulatory properties of the isoforms of these two enzymes expressed in the three tissues, focusing on their potential regulatory similarities and differences. Additionally, the activity, structure, and regulation of the remaining enzyme necessary for glycogenolysis, glycogen-debranching enzyme, are also reviewed.

High frequency of missense mutations in glycogen storage disease type VI

Deficiency of liver glycogen phosphorylase in glycogen storage disease (GSD) type VI results in a reduced ability to mobilize glucose from glycogen. Six mutations of the PYGL gene, which encodes the liver isoform of the enzyme, have been identified in the literature. We have characterized eight patients from seven families with GSD type VI and identified 11 novel PYGL gene defects. The majority of the mutations were missense, resulting in the substitution of highly conserved residues. These could be grouped into those that were predicted to affect substrate binding (p.V456M, p.E673K, p.S675L, p.S675T), pyridoxal phosphate binding (p.R491C, p.K681T), or activation of glycogen phosphorylase (p.Q13P) or that had an unknown effect (p.N632I and p.D634H). Two mutations were predicted to result in null alleles, p.R399X and [c.1964_1969inv6;c.1969+1_+4delGTAC]. Only 7 of the 23 (30%) reported PYGL alleles carry nonsense, splice site or frameshift mutations compared to 68-80% of affected alleles of the highly homologous muscle glycogen phosphorylase gene, PYGM, that underlie McArdle disease. There was heterogeneity in the clinical symptoms observed in affected individuals. These varied from hepatomegaly and subclinical hypoglycaemia, to severe hepatomegaly with recurrent severe hypoglycaemia and postprandial lactic acidosis. We conclude that deficiency of liver glycogen phosphorylase is predominantly the result of missense mutations affecting enzyme activity. There are no common mutations and the severity of clinical symptoms varies significantly.

Functional attributes of the phosphate group binding cup of pyridoxal phosphate-dependent enzymes

Twenty-four structures of pyridoxal-5'-phosphate (PLP)-dependent enzymes that represent five different folds are shown to share a common recognition pattern for the phosphate group of their PLP-ligands. All atoms that interact with the phosphate group of PLP in these proteins are organized within a two-layer structure so that the first interacting layer contains from five to seven atoms and parallel with this is a second layer containing from three to seven interacting atoms. In order to identify features of the phosphate-binding site common to PLP-dependent enzymes, a simple procedure is described that assigns relative positions to all interacting atoms unambiguously, such that the networks of interactions for different proteins can be compared. On the basis of these diagrams for 24 enzyme-cofactor complexes, a detailed comparison of the two-layer structures of PLP-dependent enzymes, with both similar and different folds, was made. A majority of the structurally defined PLP-dependent proteins use the same atom types in analogous "key" positions to bind their PLP-ligands. In some instances, proteins use water molecules when a key position is unoccupied. A similar two-layer recognition pattern extends to protein recognition of at least one other, non-PLP ligand, glucosamine 6-phosphate. We refer to this three-dimensional recognition pattern as the phosphate-binding cup. In general, the phosphate-binding cup provides a very stable anchoring point for PLP. When numerous water molecules occur within the cup, however, then the phosphate group of PLP participates directly in the enzymatic reactions with inorganic phosphate replacing the water molecules of the cup. With glucosamine-6-phosphate synthase, the water molecules of the phosphate-binding cup facilitate the entry of substrate and the exit of product.

Lysine 2,3-aminomutase from Clostridium subterminale SB4: mass spectral characterization of cyanogen bromide-treated peptides and cloning, sequencing, and expression of the gene kamA in Escherichia coli

Lysine 2,3-aminomutase (KAM, EC 5.4.3.2.) catalyzes the interconversion of L-lysine and L-beta-lysine, the first step in lysine degradation in Clostridium subterminale SB4. KAM requires S-adenosylmethionine (SAM), which mediates hydrogen transfer in a mechanism analogous to adenosylcobalamin-dependent reactions. KAM also contains an iron-sulfur cluster and requires pyridoxal 5'-phosphate (PLP) for activity. In the present work, we report the cloning and nucleotide sequencing of the gene kamA for C. subterminale SB4 KAM and conditions for its expression in Escherichia coli. The cyanogen bromide peptides were isolated and characterized by mass spectral analysis and, for selected peptides, amino acid and N-terminal amino acid sequence analysis. PCR was performed with degenerate oligonucleotide primers and C. subterminale SB4 chromosomal DNA to produce a portion of kamA containing 1,029 base pairs of the gene. The complete gene was obtained from a genomic library of C. subterminale SB4 chromosomal DNA by use of DNA probe analysis based on the 1,029-base pair fragment. The full-length gene consisted of 1,251 base pairs specifying a protein of 47,030 Da, in reasonable agreement with 47, 173 Da obtained by electrospray mass spectrometry of the purified enzyme. N- and C-terminal amino acid analysis of KAM and its cyanogen bromide peptides firmly correlated its amino acid sequence with the nucleotide sequence of kamA. A survey of bacterial genome databases identified seven homologs with 31 to 72% sequence identity to KAM, none of which were known enzymes. An E. coli expression system consisting of pET 23a(+) plus kamA yielded unsatisfactory expression and bacterial growth. Codon usage in kamA includes the use of AGA for all 29 arginine residues. AGA is rarely used in E. coli, and arginine clusters at positions 4 and 5, 25 and 27, and 134, 135, and 136 apparently compound the barrier to expression. Coexpression of E. coli argU dramatically enhanced both cell growth and expression of KAM. Purified recombinant KAM is equivalent to that purified from C. subterminale SB4.

Common structural elements in the architecture of the cofactor-binding domains in unrelated families of pyridoxal phosphate-dependent enzymes

A detailed comparison of the structures of aspartate aminotransferase, alanine race-mase, the beta subunit of tryptophan synthase, D-amino acid aminotransferase and glycogen phosphorylase has revealed more extensive structural similarities among pyridoxal phosphate (PLP)-binding domains in these enzymes than was observed previously. These similarities consist of seven common structural segments of the polypeptide chain, which form an extensive common structural organization of the backbone chain responsible for the appropriate disposition of key residues, some from the aligned fragments and some from variable loops joined to these fragments, interacting with PLPs in these enzymes. This common structural organization contains an analogous hydrophobic minicore formed from four amino acid side chains present in the two most conserved structural elements. In addition, equivalent alpha-beta-alpha-beta supersecondary structures are formed by these seven fragments in three of the five structures: alanine racemase, tryptophan synthase and glycogen phosphorylase. Despite these similarities, it is generally accepted that these proteins do not share a common heritage, but have arisen on five separate occasions. The common and contiguous alpha-beta-alpha-beta structure accounts for only 28 residues and all five enzymes differ greatly in both the orientation of the PLP pyridoxal rings and their contacts with residues close to the common structural elements.

Inhibitors of glycogen phosphorylase b: synthesis, biochemical screening, and molecular modeling studies of novel analogues of hydantocidin

The synthesis and biochemical screening of four novel spironucleosides 1-4 against rabbit liver glycogen phosphorylase b (Gpb), along with molecular modeling studies on compound 2 and its 4-hydroxy analogue VII, have been presented. Gpb is a key enzyme of glycogen metabolism, and is known to be involved in the control of diabetes mellitus. The general strategy for synthesis involved base-catalyzed condensation of diethyl 2,4-dioxoimidazolidine-5-phosphonate (5) with either 2-deoxy-D-ribose or D-ribose, followed by sequential reactions involving ring-closure with phenylselenenyl chloride and reduction with tri-n-butyltin hydride catalyzed by azobisisobutyronitrile. Compounds 2 and 4 were found to be weak competitive inhibitors of Gpb, whereas 1 and 3 were inactive.

Time-resolved diffraction studies on glycogen phosphorylase b

Glycogen phosphorylase catalyses the reversible phosphorylation of glycogen to give glucose-1-phosphate in a reaction mechanism promoted by the 5'-phosphate of the cofactor pyridoxal phosphate. The reaction with the small substrate heptenitol has been probed using Laue diffraction at the Synchrotron Radiation Source, Daresbury. The reaction was initiated following photolysis from a caged phosphate compound 3,5-dinitrophenylphosphate (DNPP). In measurements on photolysis in the crystal using a diode array spectrophotometer approximately 7 mM cage (and hence phosphate) was released from a 21 mM solution with five flashes from a xenon flash lamp. In an experiment with the home source it was shown that DNPP is stable in the crystal under conditions of X-ray measurements and that on flashing sufficient phosphate is released to promote catalysis within 24 h. In a similar experiment with the synchrotron and Laue diffraction, data were recorded before and then 3 min, 15 min and 1 h after initiation of the reaction. Theoretical analysis of the point spread function arising from partial data-sets, numerical calculations with ideal data and the experimental results have shown the importance of low-resolution terms for the interpretation of Laue difference maps. Inclusion of terms obtained from unscrambling the wavelength harmonic overlaps led to significant improvement. The maps showed heptenitol bound at the catalytic site but no evidence for catalysis under these conditions. A rational for the lack of reaction and suggestions for future experiments with improved data are outlined.

Activator anion binding site in pyridoxal phosphorylase b: the binding of phosphite, phosphate, and fluorophosphate in the crystal

It has been established that phosphate analogues can activate glycogen phosphorylase reconstituted with pyridoxal in place of the natural cofactor pyridoxal 5'-phosphate (Change YC. McCalmont T, Graves DJ. 1983. Biochemistry 22:4987-4993). Pyridoxal phosphorylase b has been studied by kinetic, ultracentrifugation, and X-ray crystallographic experiments. In solution, the catalytically active species of pyridoxal phosphorylase b adopts a conformation that is more R-state-like than that of native phosphorylase b, but an inactive dimeric species of the enzyme can be stabilized by activator phosphite in combination with the T-state inhibitor glucose. Co-crystals of pyridoxal phosphorylase b complexed with either phosphite, phosphate, or fluorophosphate, the inhibitor glucose, and the weak activator IMP were grown in space group P4(3)2(1)2, with native-like unit cell dimensions, and the structures of the complexes have been refined to give crystallographic R factors of 18.5-19.2%, for data between 8 and 2.4 A resolution. The anions bind tightly at the catalytic site in a similar but not identical position to that occupied by the cofactor 5'-phosphate group in the native enzyme (phosphorus to phosphorus atoms distance = 1.2 A). The structural results show that the structures of the pyridoxal phosphorylase b-anion-glucose-IMP complexes are overall similar to the glucose complex of native T-state phosphorylase b. Structural comparisons suggest that the bound anions, in the position observed in the crystal, might have a structural role for effective catalysis.

Phosphonate and alpha-fluorophosphonate analogue probes of the ionization state of pyridoxal 5'-phosphate (PLP) in glycogen phosphorylase

To investigate the role of the essential cofactor pyridoxal phosphate in rabbit muscle glycogen phosphorylase catalysis, two phosphonate analogues of pyridoxal phosphate, 5'-deoxypyridoxal 5'-methylenephosphonic acid and 5'-deoxypyridoxal 5'-difluoromethylenephosphonic acid, have been prepared and reconstituted into apophosphorylase b. UV/Vis spectroscopic and 31P and 19F NMR studies confirmed the successful reconstitution and revealed significant changes in phosphate environment upon nucleotide activation. Both such reconstituted enzymes had activities of approximately 25%-30% of that observed in the native enzyme, while K(m) values were similar to those of the native enzyme. Very similar dependences upon pH of Vmax, K(m), and Vmax/K(m) were found for the two reconstituted enzyme derivatives and the native enzyme despite the considerable difference in phosphonic acid pKa values. These results suggest that pyridoxal phosphate does not function as an essential acid/base catalyst in glycogen phosphorylase; rather they suggest that the cofactor phosphate moiety remain dianionic throughout catalysis and functions as an essential dianion. Mechanistic implications of these findings are discussed.

Structural motifs for pyridoxal-5'-phosphate binding in decarboxylases: an analysis based on the crystal structure of the Lactobacillus 30a ornithine decarboxylase

Two of the five domains in the structure of the ornithine decarboxylase (OrnDC) from Lactobacillus 30a share similar structural folds around the pyridoxal-5'-phosphate (PLP)-binding pocket with the aspartate aminotransferases (AspATs). Sequence comparisons focusing on conserved residues of the aligned structures reveal that this structural motif is also present in a number of other PLP-dependent enzymes including the histidine, dopa, tryptophan, glutamate, and glycine decarboxylases as well as tryptophanase and serine-hydroxymethyl transferase. However, this motif is not present in eukaryotic OrnDCs, the diaminopimelate decarboxylases, nor the Escherichia coli or oat arginine decarboxylases. The identification and comparison of residues involved in defining the different classes are discussed.

Control of phosphorylase b conformation by a modified cofactor: crystallographic studies on R-state glycogen phosphorylase reconstituted with pyridoxal 5'-diphosphate

Previous crystallographic studies on glycogen phosphorylase have described the different conformational states of the protein (T and R) that represent the allosteric transition and have shown how the properties of the 5'-phosphate group of the cofactor pyridoxal phosphate are influenced by these conformational states. The present work reports a study on glycogen phosphorylase b (GPb) complexed with a modified cofactor, pyridoxal 5'-diphosphate (PLPP), in place of the natural cofactor. Solution studies (Withers, S.G., Madsen, N.B., & Sykes, B.D., 1982, Biochemistry 21, 6716-6722) have shown that PLPP promotes R-state properties of the enzyme indicating that the cofactor can influence the conformational state of the protein. GPb complexed with pyridoxal 5'-diphosphate (PLPP) has been crystallized in the presence of IMP and ammonium sulfate in the monoclinic R-state crystal form and the structure refined from X-ray data to 2.8 A resolution to a crystallographic R value of 0.21. The global tertiary and quaternary structure in the vicinity of the Ser 14 and the IMP sites are nearly identical to those observed for the R-state GPb-AMP complex. At the catalytic site the second phosphate of PLPP is accommodated with essentially no change in structure from the R-state structure and is involved in interactions with the side chains of two lysine residues (Lys 568 and Lys 574) and the main chain nitrogen of Arg 569. Superposition of the T-state structure shows that were the PLPP to be incorporated into the T-state structure there would be a close contact with the 280s loop (residues 282-285) that would encourage the T to R allosteric transition. The second phosphate of the PLPP occupies a site that is distinct from other dianionic binding sites that have been observed for glucose-1-phosphate and sulfate (in the R state) and for heptulose-2-phosphate (in the T state). The results indicate mobility in the dianion recognition site, and the precise position is dependent on other linkages to the dianion. In the modified cofactor the second phosphate site is constrained by the covalent link to the first phosphate of PLPP. The observed position in the crystal suggests that it is too far from the substrate site to represent a site for catalysis.

The molecular mechanism for the tetrameric association of glycogen phosphorylase promoted by protein phosphorylation

The allosteric transition of glycogen phosphorylase promoted by protein phosphorylation is accompanied by the association of a pair of functional dimers to form a tetramer. The conformational changes within the dimer that lead to the creation of a protein recognition surface have been analyzed from a comparison of the crystal structures of T-state dimeric phosphorylase b and R-state tetrameric phosphorylase a. Regions of the structure that participate in the tetramer interface are situated within structural subdomains. These include the glycogen storage subdomain, the C-terminal subdomain and the tower helix. The subdomains undergo concerted conformational transitions on conversion from the T to the R state (overall r.m.s. shifts between 1 and 1.7 A) and, together with the quaternary conformational change within the functional dimer, create the tetramer interface. The glycogen storage subdomain and the C-terminal subdomain are distinct from those regions that contribute to the dimer interface, but shifts in the subdomains are correlated with the allosteric transitions that are mediated by the dimer interface. The structural properties of the tetramer interface are atypical of an oligomeric protein interface and are more similar to protein recognition surfaces observed in protease inhibitors and antibody-protein antigen complexes. There is a preponderance of polar and charged residues at the tetramer interface and a high number of H-bonds per surface area (one H-bond per 130 A2). In addition, the surface area made inaccessible at the interface is relatively small (1,142 A2 per subunit on dimer to tetramer association compared with 2,217 A2 per subunit on monomer-to-dimer association).

The binding of D-gluconohydroximo-1,5-lactone to glycogen phosphorylase. Kinetic, ultracentrifugation and crystallographic studies

Combined kinetic, ultracentrifugation and X-ray-crystallographic studies have characterized the effect of the beta-glucosidase inhibitor gluconohydroximo-1,5-lactone on the catalytic and structural properties of glycogen phosphorylase. In the direction of glycogen synthesis, gluconohydroximo-1,5-lactone was found to competitively inhibit both the b (Ki 0.92 mM) and the alpha form of the enzyme (Ki 0.76 mM) with respect to glucose 1-phosphate in synergism with caffeine. In the direction of glycogen breakdown, gluconohydroximo-1,5-lactone was found to inhibit phosphorylase b in a non-competitive mode with respect to phosphate, and no synergism with caffeine could be demonstrated. Ultracentrifugation and crystallization experiments demonstrated that gluconohydroximo-1,5-lactone was able to induce dissociation of tetrameric phosphorylase alpha and stabilization of the dimeric T-state conformation. A crystallographic binding study with 100 mM-gluconohydroximo-1,5-lactone at 0.24 nm (2.4 A) resolution showed a major peak at the catalytic site, and no significant conformational changes were observed. Analysis of the electron-density map indicated that the ligand adopts a chair conformation. The results are discussed with reference to the ability of the catalytic site of the enzyme to distinguish between two or more conformations of the glucopyranose ring.

Measurement of active-site homology between potato and rabbit muscle alpha-glucan phosphorylases through use of a linear free energy relationship

The Michaelis-Menten parameters (Vmax and Km) for turnover of an extensive series of deoxy and deoxyfluoro derivatives of alpha-D-glucopyranosyl phosphate by the alpha-glucan phosphorylase from potato tuber have been determined. Very large rate reductions are observed as a consequence of each substitution, primarily due to losses in specific binding interactions, most likely hydrogen bonding, at the enzymic transition state. Comparison of the Vmax/Km values so determined with those measured for rabbit muscle alpha-glucan phosphorylase [Street et al. (1989) Biochemistry 28, 1581] reveals an astonishingly similar specificity, especially in light of the phylogenetic separation of their host organisms. This indicates that very similar hydrogen-bonding interactions between the enzyme and the substrate must be present at the transition states for the two enzymic reactions; therefore, they have very similar active sites. Quantitation of this similarity is achieved by plotting the logarithm of the Vmax/Km value for each substrate analogue with the potato enzyme against the same parameter for the muscle enzyme, yielding straight lines (p = 0.998 and 0.999) of slope 1.0 and 1.2 for the deoxy and deoxyfluoro substrates, respectively. Since the correlation coefficient of such plots is a direct measure of the similarity of the two transition-state complexes, thus of the enzyme active sites, it can be used as a measure of active-site homology between the two enzymes. The extremely high homology observed in this case is consistent with the observed sequence homology at the active site.

Refined crystal structure of the phosphorylase-heptulose 2-phosphate-oligosaccharide-AMP complex

The crystal structure of phosphorylase b-heptulose 2-phosphate complex with oligosaccharide and AMP bound has been refined by molecular dynamics and crystallographic least-squares with the program XPLOR. Shifts in atomic positions of up to 4 A from the native enzyme structure were correctly determined by the program without manual intervention. The final crystallographic R value for data between 8 and 2.86 A resolution is 0.201, and the overall root-mean-square difference between the native and complexed structure is 0.58 A for all protein atoms. The results confirm the previous observation that there is a direct hydrogen bond between the phosphate of heptulose 2-phosphate and the pyridoxal phosphate 5'-phosphate group. The close proximity of the two phosphates is stabilized by an arginine residue, Arg569, which shifts from a site buried in the protein to a position where it can make contact with the product phosphate. There is a mutual interchange in position between the arginine and an acidic group, Asp283. These movements represent the first stage of the allosteric response which converts the catalytic site from a low to a high-affinity binding site. Communication of these changes to other sites is prevented in the crystal by the lattice forces, which also form the subunit interface. The constellation of groups in the phosphorylase transition state analogue complex provides a structural basis for understanding the catalytic mechanism in which the cofactor pyridoxal phosphate 5'-phosphate group functions as a general acid to promote attack by the substrate phosphate on the glycosidic bond when the reaction proceeds in the direction of glycogen degradation. In the direction of glycogen synthesis, stereoelectronic effects contribute to the cleavage of the C-1-O-1 bond. In both reactions the substrate phosphate plays a key role in transition state stabilization. The details of the oligosaccharide, maltoheptaose, interactions with the enzyme at the glycogen storage site are also described.

The allosteric transition of glycogen phosphorylase

The crystal structure of R-state glycogen phosphorylase b has been determined at 2.9 A resolution. A comparison of T-state and R-state structures of the enzyme explains its cooperative behaviour on ligand binding and the allosteric regulation of its activity. Communication between catalytic sites of the dimer is provided by a change in packing geometry of two helices linking each site with the subunit interface. Activation by AMP or by phosphorylation results in a quaternary conformational change that switches these two helices into the R-state conformation.

Fluorinated and deoxygenated substrates as probes of transition-state structure in glycogen phosphorylase

A series of deoxyfluoro- and deoxy-alpha-D-glucopyranosyl phosphates have been tested as substrates of rabbit muscle glycogen phosphorylase b. All are found to be utilized by the enzyme, but at substantially reduced rates. Values of Vm/Km for these analogues range from 10(2) to 10(5) times lower than that for the parent substrate. The large rate reductions are suggested to arise from a combination of intrinsic electronic effects and poorer binding of these substrates at the transition state. The data provide substantial evidence for an oxocarbonium-ion-like transition state. They also provide estimates of the strengths of hydrogen bonds to individual sugar hydroxyls at the transition state of the reaction. Further, comparison of such data with those obtained for glucose analogues binding as inhibitors to T-state phosphorylase suggests that these two glucose subsites are essentially identical; thus, the glucose pocket remains intact during the conformational transition associated with activation of the enzyme.

Structural changes in glycogen phosphorylase induced by phosphorylation

A comparison of the refined crystal structures of dimeric glycogen phosphorylase b and a reveals structural changes that represent the first step in the activation of the enzyme. On phosphorylation of serine-14, the N-terminus of each subunit assumes an ordered helical conformation and binds to the surface of the dimer. The consequent structural changes at the N- and C-terminal regions lead to strengthened interactions between subunits and alter the binding sites for allosteric effectors and substrates.

Uridine(5')diphospho(1)-alpha-D-glucose. A binding study to glycogen phosphorylase b in the crystal

UDP-glucose is an R-state inhibitor of glycogen phosphorylase b, competitive with the substrate, glucose 1-phosphate and noncompetitive with the allosteric activator, AMP. Diffusion of 100 mM UDP-glucose into crystals of phosphorylase b resulted in a difference Fourier synthesis at 0.3-nm resolution that showed two peaks: (a) binding at the allosteric site and (b) binding at the catalytic site. At the allosteric site the whole of the UDP-glucose molecule can be located. It is in a well defined folded conformation with its uracil portion in a similar position to that observed for the adenine of AMP. The uracil and the glucose moieties stack against the aromatic side chains of Tyr-75 and Phe-196, respectively. The phosphates of the pyrophosphate component interact with Arg-242, Arg-309 and Arg-310. At the catalytic site, the glucose-1-P component of UDP-glucose is firmly bound in a position similar to that observed for glucose 1-phosphate. The pyrophosphate is also well located with the glucose phosphate interacting with the main-chain NH groups at the start of the glycine-loop alpha helix and the uridine phosphate interacting through a water molecule with the 5'-phosphate of the cofactor pyridoxal phosphate and with the side chains of residues Tyr-573, Lys-574 and probably Arg-569. However the position of the uridine cannot be located although analysis by thin-layer chromatography showed that no degradation had taken place. Binding of UDP-glucose to the catalytic site promotes extensive conformational changes. The loop 279-288 which links the catalytic site to the nucleoside inhibitor site is displaced and becomes mobile. Concomitant movements of residues His-571, Arg-569, and the loop 378-383, together with the major loop displacement, result in an open channel to the catalytic site. Comparison with other structural results shows that these changes form an essential feature of the T to R transition. They allow formation of the phosphate recognition site at the catalytic site and destroy the nucleoside inhibitor site. Kinetic experiments demonstrate that UDP-glucose activates the enzyme in the presence of high concentrations of the weak activator IMP, because of its ability to decrease the affinity of IMP for the inhibitor site.