Effects of Anions on the Substrate Affinities of the Pyridoxal and Pyridoxamine Forms of Mitochondrial and Supernatant Aspartate Transaminases

Release of pyridoxal 5'-phosphate upon unfolding of mitochondrial aspartate aminotransferase

Dimeric mitochondrial aspartate aminotransferase (mAAT) contains a molecule of pyridoxal 5'-phosphate (PLP) tightly attached to each of its two identical active sites. The presence of this natural reporter allows us to study separately local perturbations in the architecture of this critical region of the molecule during unfolding. Upon unfolding of the enzyme with guanidine hydrochloride (GdnHCl), the coenzyme is completely released from the active site. The transition midpoint for the dissociation of PLP is 1.4+/-0.02 M when determined by size-exclusion chromatography (SEC) and 1.6+/-0.02 M when the protein-bound PLP is estimated by electrospray mass spectrometry (ESI-MS). In both cases the transition midpoint is higher than that of inactivation (1.3+/-0.01 M). On the other hand, the midpoint of the unfolding transition obtained by monitoring changes in ellipticity at 356 nm, which reflects the asymmetric environment of the PLP cofactor at the active site, is 1.19+/-0.011 M guanidine. These results indicate that the unfolding of mAAT is a multi-step process which includes an intermediate containing bound PLP but lacking catalytic activity.

Structure and function of branched chain aminotransferases

Branched chain aminotransferases (BCATs) catalyze transamination of the branched chain amino acids (BCAAs) leucine, isoleucine, and valine. Except for the Escherichia coli and Salmonella proteins, which are homohexamers arranged as a double trimer, the BCATs are homodimers. Structurally, the BCATs belong to the fold type IV class of pyridoxal phosphate (PLP) enzymes. Other members are D-alanine aminotransferase and 4-amino-4-deoxychorismate lyase. Catalysis is on the re face of the PLP cofactor, whereas in other classes, catalysis occurs from the si face of PLP. Crystal structures of the fold type IV proteins show that they are distinct from the fold type I aspartate aminotransferase family and represent a new protein fold. Because the fold type IV enzymes catalyze diverse reactions, it is not surprising that the greatest structural similarities involve residues that participate in PLP binding rather than residues involved in substrate binding. The BCATs are widely distributed in the bacterial kingdom, where they are involved in the synthesis/degradation of the BCAAs. Bacteria contain a single BCAT. In eukaryotes there are two isozymes, one is mitochondrial (BCATm) and the other is cytosolic (BCATc). In mammals, BCATm is in most tissues, and BCATm is thought to be important in body nitrogen metabolism. BCATc is largely restricted to the central nervous system (CNS). Recently, BCATc has been recognized as a target of the neuroactive drug gabapentin. BCATc is involved in excitatory neurotransmitter glutamate synthesis in the CNS. Ongoing structural studies of the BCATs may facilitate the design of therapeutic compounds to treat neurodegenerative disorders involving disturbances of the glutamatergic system.

Overexpression and characterization of the human mitochondrial and cytosolic branched-chain aminotransferases

We have developed overexpression systems for the human branched-chain aminotransferase isoenzymes. The enzymes function as dimers and have substrate specificity comparable with the rat enzymes. The human cytosolic enzyme appears to turn over 2-5 times faster than the mitochondrial enzyme, and there may be anion and cation effects on the kinetics of both enzymes. The two proteins demonstrate similar absorption profiles, and the far UV circular dichroism spectra show that no global structural changes occur when the proteins are converted from the pyridoxal to pyridoxamine form. On the other hand, the near UV circular dichroism spectra suggest differences in the local environment surrounding tyrosines within these proteins. Both enzymes require a reducing environment for maximal activity, but the mitochondrial enzyme can be inhibited by nickel ions in the presence of reducing agents, while the cytosolic enzyme is unaffected. Chemical denaturation profiles of the proteins show that there are differences in structural stability. Titration of -SH groups with 5,5'-dithiobis(2-nitrobenzoic acid) suggests that no disulfide bonds are present in the mitochondrial enzyme and that at least two disulfide bonds are present in the cytosolic enzyme. Two -SH groups are titrated in the native form of the mitochondrial enzyme, leading to complete inhibition of activity, while only one -SH group is titrated in the cytosolic enzyme with no effect on activity. Although these proteins share 58% identity in primary amino acid sequence, the local environment surrounding the active site appears unique for each isoenzyme.

The ionization states of the 5'-phosphate group in the various coenzyme forms bound to mitochondrial aspartate aminotransferase

We have carried out a Fourier transform infrared spectroscopic study of mitochondrial aspartate aminotransferase in the spectral region where phosphate monoesters give rise to absorption. Infrared spectra in the above-mentioned region are dominated by protein absorption. Yet, below 1020 cm-1 protein interferences are minor, permitting the detection of the band arising from the symmetric stretching of dianionic phosphate monoesters [T. Shimanouchi, M. Tsuboi, and Y. Kyogoku (1964) Adv. Chem. Phys. 8, 435-498]. The integrated intensity of this band in several enzyme forms (pyridoxal phosphate, pyridoxamine phosphate, and sodium borohydride-reduced, pyridoxyl phosphate form) does not change with pH in the range 5-9. This behavior contrasts that of free pyridoxal phosphate (PLP) and pyridoxamine phosphate (PMP) in solution, where the dependence of the same infrared band intensity with pH can be correlated to the known pK values for the 5'-phosphate ester in solution. The integrated intensity value of this infrared band for the PLP enzyme form before and after reduction with sodium borohydride is close to that given by free PLP at pH 8-9. These results are taken as evidence that in the active site of mitochondrial aspartate aminotransferase the 5'-phosphate group of PLP remains mostly dianionic even at a pH near 5. Thus, it is suggested that the chemical shift changes associated with pH titrations of various PLP forms reported in a previous 31P NMR study of this enzyme [M. E. Mattingly, J. R. Mattingly, and M. Martinez-Carrion (1982) J. Biol. Chem. 257, 8872] are due to the fact that the phosphorus chemical shift senses the O-P-O bond distortions induced by the ionization of a nearby residue. Since no chemical shift changes were observed in pH titrations of the PMP forms (lacking an ionizable internal aldimine) of this isozyme, the Schiff base between PLP and Lys-258 at the active site is the most likely candidate for the ionizing group influencing the phosphorus chemical shift in this enzyme.

Aspartate aminotransferase from Eleusine coracana, a C4 plant: purification, characterization, and preparation of antibody

Aspartate aminotransferase (AspAT) isozymes from Eleusine coracana (an NAD-malic enzyme type C4 plant) were examined. Three groups of isoenzymes were identified (AspAT-1, AspAT-2, and AspAT-3). AspAT-1 (localized in the mesophyll cells) and AspAT-3 (localized in the bundle sheath cells), both of which are considered to function in the C4 acid pathway, were purified and their kinetic and physical properties studied. Both isoenzymes had a molecular mass of 80 kDa and were shown to consist of two identical 40-kDa monomers. Except for the higher affinity for aspartate and the lower activity for the forward direction (Asp----OAA) at lower pH exhibited by AspAT-3 compared with AspAT-1, the isozymes had similar kinetic properties. However they had quite different isoelectric points. Polyclonal antibodies raised against AspAT-3 preferentially cross-reacted with AspAT-3 but did show some cross-reactivity with AspAT-1.

Aspartate aminotransferase isoenzymes

Aspartate aminotransferase (AST, EC 2.6.1.1) exists in human tissues as two distinct isoenzymes, one located in the cytoplasm (c-AST), and the other in mitochondria (m-AST). Striated muscle, myocardium, and liver tissues are the main sources of AST. A growing body of information suggests that determination of AST isoenzymes in human serum is useful in evaluating damage to some of these organs. In hepatic disease, the test is used to assess liver necrosis and for determining prognosis. It may also assist in identifying patients with active alcoholic liver disease. In patients with acute myocardial infarction, measurement of AST isoenzymes provides diagnostic information that differs from that obtained by determination of total creatine kinase and lactate dehydrogenase enzymes, and their isoenzymes.

Kinetic studies of chloride inhibition in aspartate aminotransferase activity

The inhibitive effects of chloride anion on the activity of mitochondrial aspartate aminotransferase (L-aspartate: 2-oxoglutarate-aminotransferase EC. 2.6.1.1.) from chicken (Gallus domesticus) and turkey (Maleagris gallopavo) were studied. Steady-state velocities were obtained from a wide range of chloride concentrations. The data were fitted by rational functions of 0:2 and 1:2 for chloride, using a non-linear regression program which guaranteed the fit. The goodness of fit was improved by the use of a computer program that combined model discrimination, parameter refinement and sequential design. It was concluded that chloride aspartate aminotransferase inhibition requires a minimum velocity equation of 1:2 with regard to chloride, and a plausible kinetic mechanism for this experimental result was proposed.

Quantitative description of absorption spectra of a pyridoxal phosphate-dependent enzyme using lognormal distribution curves

Ultraviolet-visible absorption spectra of cytosolic aspartate aminotransferase of pig hearts have been analyzed by resolution with lognormal distribution curves. These have been compared with spectra of reference Schiff bases of pyridoxal 5'-phosphate. Spectra of the free enzyme in two different states of protonation and of complexes with monoanions, dicarboxylates, the substrates L-glutamate, L-aspartate, and L-erythro-3-hydroxyaspartate, and the quasi-substrate 2-methylaspartate have been analyzed. Relative amounts of three tautomeric species have been estimated, as have amounts of various enzyme-substrate intermediates. Bandshape parameters which can be used as a guide to analysis of spectra of other pyridoxal phosphate-dependent enzymes are tabulated. Some formation constants and pKa values, which were evaluated at the same time as the spectra of the complexes, are also reported.

Intramitochondrial location and some characteristics of chicken liver aspartate aminotransferase

Chicken liver mitochondrial aspartate aminotransferase was found to be located in the intermembrane space and bound to the inner mitochondrial membrane. Purification of two mitochondrial fractions containing aspartate aminotransferase activity was performed. Both fractions showed similar chromatographic behaviour and identical isoelectric point and molecular weight values. There were no significant differences in the general kinetic mechanism, Km values, substrates inhibition and effect of various anions on the activity of mitochondrial aspartate aminotransferase purified from both fractions.

Effect of phosphate and other inorganic anions on the activity of chicken liver cytosolic aspartate aminotransferase

Chicken liver aspartate aminotransferase was inhibited by several inorganic anions. The inhibitory effect of the anions was related to their chaotropic character. Apparent Km (2-oxoglutarate) and Km (L-aspartate) values depended on the molarity of the buffer. The profile of the curves obtained did not depend on the nature of the enzyme sample assayed. Phosphate slightly inhibited the holoaspartate aminotransferase and was a strong inhibitor of apoaspartate aminotransferase with respect to pyridoxal phosphate.

Measurement of aminotransferases: Part 1. Aspartate aminotransferase

Aminotransferases are ubiquitous enzymes of mammalian cells and several are of important diagnostic use. The application of aspartate aminotransferase activity measurements in serum from individuals suffering from myocardial infarction brought about a new dimension in clinical laboratory testing in the 1950s. This review focuses on measurement techniques for aspartate aminotransferase and their application (a subsequent article will review other aminotransferases). Assay techniques measuring enzyme activity are direct spectrophotometric measurements, manometric techniques, assays using dye substances, coupled enzyme techniques, and radiometric procedures. Of these procedures, the one employing malate dehydrogenase and NADH is the most important and is covered in particular detail. The estimation of the mitochondrial isoenzyme of aspartate aminotransferase is also of clinical interest, in particular for estimating severity of disease or in specific applications (e.g., chronic alcoholism). Methods reviewed for estimation of this enzyme are electrophoresis, chromatography, differential kinetic behavior, and immunochemical separation. Determination of the enzyme protein by techniques independent of its catalytic activity are also reviewed.

Inhibition of glutamate-aspartate transaminase by beta-methylene-DL-aspartate

beta-Methylene-DL-aspartate, a new beta, gamma-unsaturated amino acid, is an irreversible inhibitor of soluble pig heart glutamate-aspartate transaminase (Ki approximately 3 mM with respect to the L-form; limiting rate constant for inactivation approximately 0.4 min-1). The new amino acid is the most specific inhibitor of glutamate-aspartate transaminase thus far studied. It does not inactivate pig heart glutamate-alanine transaminase, soluble rat kidney glutamine transaminase K, gamma-aminobutyrate transaminase (from Pseudomonas fluorescens), glutamate decarboxylase (Escherichia coli), snake venom L-amino acid oxidase, or hog kidney D-amino acid oxidase. In addition, the following enzymes were not inhibited by beta-methylene-DL-aspartate in rat tissue homogenates: gamma-aminobutyrate transaminase (brain), tyrosine transaminase (liver), glutamine transaminase L (liver), asparagine, transaminase (liver), ornithine transaminase (liver) or branch-chain transaminase(s) (kidney). Intraperitoneal injection of beta-methylene-DL-aspartate into mice decreased kidney and liver glutamate-aspartate transaminase activities but had no effect on liver glutamate-alanine transaminase activity.

Effect of inorganic anions of the inhibition of trehalase activity by mercuric chloride

Monovalent inorganic anions showed an unexpected effect on the inhibition of trehalase (alpha, alpha-trehalose glucohydrolase, EC 3.2.1.28) by SH inhibitors. This phenomeon (deinhibition) was caused by monovalent anions, Cl-, Br-, I- and SCN- . F- and ClO4- showed partial deinhibition. Deinhibition was not caused by NO2- and SO4-. The effectiveness of the "active anions' in causing deinhibition was highly dependent on the anion size. Trehalase in the presence of mercuric chloride was "activated' by Cl-, and the activation was saturable. From the results of Dixon plots for trehalase at different concentrations of the "activator' (deinhibitor) and a constant concentration of the substrate, it can be seen that the activator and the inhibitor competed with each other. Thus, it is suggested that the activator and the inhibitor share a common binding site or bind very near each other. The Ki value for mercuric chloride was increased with increasing concentration of NaCl. Therefore, it might be essential to remove the "active anions' in order to determine the inhibitory effect and the Ki value of trehalase for SH inhibitors.

The development and evaluation of a homogeneous immunoassay for the isoenzymes of aspartate aminotransferase

Purified isoenzymes of aspartate aminotransferase (AST) from human liver (mitochondrial) and erythrocytes (cytoplasmic) were used to elicit antisera in rabbits. Each antiserum was characterized for titer and specificity. Complexes formed upon addition of each isoenzyme to its specific antiserum were demonstrated to be catalytically inactive. Results obtained when either filtration or centrifugation was used for separating the complexes (heterogeneous assay) were comparable to those obtained when the complexes were not separated (homogeneous assay) from the mixture before assay. A quality control system was designed to monitor specificity in addition to the usual parameters. The precision for the inhibition of cytoplasmic (CV 4.8%) and mitochondrial (CV 3.5%) isoenzymes was within that of the enzymatic assay. Several parametric conditions of the enzyme-antibody reactions were examined, and the assay was adapted for semi-automation. The homogeneous assay was evaluated with a series of pseudo specimens containing known mixtures of pure isoenzymes to determine the extent of recovery (99.8%) for a particular isoenzyme in the presence of varying concentrations of the other isoenzyme. In addition, sera from patients having elevated AST concentrations were examined for isoenzyme contributions to total AST activity. A mean recovery of 96% was obtained for these specimens.

Catalytic activity of aspartate aminotransferase in the crystal. Equilibrium and kinetic analysis

Natural substrates and analogs rapidly diffuse through crystals of pig heart mitochondrial aspartate aminotransferase and react at the active sites causing spectral changes that can be measured by single-crystal microspectrophotometry. Dissociation constants for natural substrates and rate constants of transamination for slowly reacting substrates have been determined. A comparison between the data obtained in the crystal and in solution shows that the crystalline enzyme is catalytically competent and that events occurring in the crystal essentially parallel those occurring in solution, even though minor differences have been detected.

Purification and characterisation of D-amino acid aminotransferase from Rhizobium japonicum

Rhizobium japonicum has D-amino acid aminotransferase and alanine racemase activities. The D-amino-acid aminotransferase has been partially purified and characterized. This enzyme has a broad specificity and is very active with D-alpha-aminobutyrate and D-aspartate as well as D-alanine and D-glutamate. The stereospecificity of the enzyme for D-amino acids was apparently absolute with respect to product inhibition, pyridoxamine formation as well as catalytic activity. The apparent molecular weight was 58,000 and the pH optimum was 7.8-7.9. The equilibrium constant in the direction of D-glutamate formation was 1.9. Initial-velocity kinetic studies indicate the enzyme acts by a ping-pong mechanism. The dissociation constant for pyridoxal phosphate and the Michaelis constants (+/- standard errors) for D-alanine and 2-oxoglutarate were determined to be 0.51 +/- 0.06 micrometer, and 2.13 +/- 0.18 and 0.058 +/- 0.005 mM respectively. The enzyme is moderately inhibited (30%) by 4 mM p-chloromercuribenzoate.

On the determination of isozyme levels in preparations containing cytoplasmic and mitochondrial aspartate aminotransferase

A spectrophotometric assay has been developed for the determination of the content of each isozyme of aspartate transaminase (L-aspartate:2-oxoglutarate aminotransferase, EC 2.6.1.1) in physiological fluids or tissue extracts. The methods relies on the ability of adipate, at low pH and ionic strength to inhibit the cytoplasmic isozyme but not the one from mitochondria. Two assays are necessary, one at pH 8.0 which measures the content of both isozymes and another at low pH which measures primarily the amount of mitochondrial isozyme. Results obtained by this simple procedure match those in which each isozyme is inhibited by its antibody. The validity of the results obtained by the new method was tested at different ratios of cytoplasmic:mitochondrial isozyme and with tissue extracts. Since the amounts of each isozyme determined by radial immunodiffusion match those values gathered by following enzymatic activity, it is concluded that the quantity of each isozyme obtained from its respective catalytic activity must represent the total protein content of each isozyme in a given sample.

The phosphoenol-pyruvate branchpoint in adult Hymenolepis diminuta (Cestoda): a study of pyruvate kinase and phosphoenol-pyruvate carboxykinase

The properties of pyruvate kinase (PK) and phosphoenol pyruvate carboxykinase (PEP CK), two enzymes that determine the preferrential accumulation of either succinate or lactate as endproducts of carbohydrate metabolism, are described in adult Hymenolepis diminuta. PK activity at Vmax and Km levels of PEP was unaffected by ATP, alanine, FDP4, OR H+ ions, but was inhibited by 50% at 6.3 mM L-lactate and 30 mM HCO3. The addition of 30 mM HCO3 increased the Km(PEP) by 6-fold but did not alter the Vmax. The inhibition of PK by HCO3 cannot be explained entirely by an effect of ionic strength, but probably represents a specific modulator-enzyme interaction. Under similar conditions PEP CK was maximally activated. Although L-lactate inhibited PEP CK (Ki(lac) = 1.8 mM), this effector may play a minor role in regulation of PEP flux. These results implicate the poise of the HCO3-:CO2 system as a major determiner of endproduct accumulation in H. diminuta.

Fluorine-19 nuclear magnetic resonance studies of effects of ligands on trifluoroacetonylated supernatant aspartate transaminase

The selective reaction of Cys-45 and -82, on the one hand, and Cys-390, on the other, with 3-bromo-1,1,1-trifluoropropanone allows for the probing of these regions of aspartate transaminase in the absence and in the presence of enzymatic ligands by 19F nuclear magnetic resonance (NMR). The 19F chemical shifts of the resonance lines differ for the three cysteines and so does their behavior with pH changes. The resonance signals with chemical shifts at 615 and 800 Hz upfield from trifluoroacetic acid correspond to modified cysteine-82 and -45 and have tentatively been assigned in this order. The 615-Hz resonance is affected by pH changes that fit best the influence of a single ionizing residue. On the 800-Hz line, the pH changes appear to be the influence of a minimum of two ionizing residues. The 19F resonance from modified Cys-390 is pH independent in the pH range 5-9 for the pyridoxal phosphate, pyridoxamine phosphate, and apoenzyme forms of the enzyme. Occupation of the active site by a quasi-enzyme-substrate complex, trifluoromethionine pyridoxyl phosphate, affects the 19F chemical shift of modified Cys-390, making it pH dependent with a pK value of 8.4. The 19F NMR properties of the pyridoxal form of Cys-390-modified enzyme can be used to monitor some ligand interactions with the active-center region. Addition of alpha-ketoglutarate or succinate to the ketone labeled enzyme causes a decrease in the resonance line width, and titrations show that this procedure is a good method with which to study the affinity of the enzyme for these ligands. The interpretation of the chemical shift and line-width characteristics of the 19F resonance arising from Cys-390 are most consistent with a model in which the region around this residue seems to be affected by conformational changes arising from substrate binding to the active-center subsites in productive (covalent) manner. Nonproductive complexes which possess fast ligand-protein exchange, such as those between alpha-ketoglutarate or succinate with the pyridoxal phosphate form of the enzyme, may result only in a greater degree of freedom for Cys-390.

Sulphydryl group modification of aspartate aminotransferase with 3-bromo-1,1,1-trifluoropropanone during catalysis

After protection of cysteine-45 and -82 with iodoacetamide or N-ethylmaleimide, and in the presence of saturating concentrations of substrates, the supernatant isozyme of pig heart aspartate transaminase has been covalently modified at cysteine-390 with 3-bromo-1,1,1-trifluoropropanone. The modified enzyme retains 60-70% of the initial specific activity and is similar to native enzyme in pH and temperature stability. After tagging cysteine-390 with the fluorinated compound, the enzyme retains substrate and inhibitor binding abilities; as shown by direct spectrophotometric titration of the active-site chromophores. The 19F NMR spectrum of the modified enzyme has been obtained by a Fourier transform NMR method. Although the transaminase is a dimeric enzyme, 19F bound at each subunit's cysteine-390 gives rise to only a single 19F resonance upfield from that of trifluoroacetic acid. The fact that the chemical shifts of the 19F probe differ in native and guanidine hydrochloride (Gdn-HCl) denatured enzyme is interpreted as the effect of the native protein groups on the probe. The discordance between the changes induced by varying concentrations of Gdn-HCl on the 19F resonance parameters, on the one hand, and the changes in enzyme activity and prosthetic group absorbance, on the other, suggests that, in aspartate transaminase, cysteine-390 lies in an environment dissimilar from that of the active-site components.

Kinetics of the interaction between aspartic aminotransferase and anions

The kinetics of the interaction between deionized supernatant aspartic aminotransferase and various anions (cacodylate, phosphate and chloride) were studied by the temperature-jump technique. The anion concentration in the range covered by our experiments does not affect the transamination rate. On the other hand the conformational transition, recently observed at the active site of the enzyme, is hindered by an excess of anions. A single relaxation effect was observed at the enzyme chromophore wavelength in systems containing the aldimine form of the enzyme and the above anions. It is shown that this effect corresponds to the protonation of the chromophore. The relaxation times were of about 10 mus with phosphate, 20-100 mus with cacodylate and 1-2 ms with chloride. The pH and concentration dependence of this effect were studied. The fits of experimental data to a rate equations for various models were tested by a chi2 analysis. The best fit was obtained with models where anions bind rapidly to a site close to the chromophore, so that the pK of the chromophore is affected by anions binding. The rate of the observed relaxation considerably increased when the anion has buffering capacities; this indicates, in the case of cacodylate and phosphate, that the acidic component of the buffer directly exchanges a proton with the enzyme chromophore.

Conformational properties of pig-heart cytoplasmic aspartate aminotransferase. Circular-dichroism and absorption-spectroscopic study of dicarboxylate binding

1. The interaction between aspartate aminotransferase and dicarboxylates of various chain lengths and geometries has been studied from pH 6.5 TO 8.5 by circular dichroism (CD) and absorption spectroscopy. Liganding causes protonation of the pyridoxal phosphate-enzyme Schiff's base complex; the consequent changes in optical properties deltaAlambda, deltaCDlambda at the coenzyme maxima (lambda = 363 or 430 nm) are analysed for binding constants and the degree of perturbation of the coenzyme protonic dissociation constant, pKa. 2. Aliphate dicarboxylates follow linear binding functions for all optical parameters; in contrast, m and p-phthalates follow non-linear binding functions for both deltaAlambda and deltaCDlambda, implying that successive phthalate ligands bind with decreasing affinity. The ratio detlaCDlambda is effectively constant for a given ligand and the characteristic values for aromatic ligands indicate a changed environment for the coenzyme. 3. Inspection of the non-linear process for phthalates suggests that initially, binding occurs with high affinity, but with characteristically small effects on pKa. It is inferred that alipathic and aromatic dicarboxylates bind at different subsites in the active site region, perturbing the coenzyme pKa by an indirect protein-mediated mechanism. 4. Non-linearity of binding could derive from multiple binding to an individual subunit. Alternatively, different single sites may exist on adjacent subunits of the dimer, implying non-equivalence between otherwise identical subunits, expressed in properties involving groups close to the active site.