DNA segments of the IncX plasmid R485 determining replication and incompatibility with plasmid R6K

The expression of incompatibility properties between the IncX plasmids R6K and R485 of Escherichia coli was examined. For small autonomously replicating derivatives of both plasmid elements, the requirements for incompatibility expression include a functional R485 replicon and an active R6K beta-origin region. Functional R6K alpha and gamma origins are not directly involved in incompatibility expression between R6K and R485. A trans-acting replication system was constructed for plasmid R485. It consists of a 3.2-(kb) DNA fragment of R485 that specifies a product(s) in trans which supports replication from an R485 origin plasmid. A minimal R485 origin region of 591 bp was derived utilizing this trans-acting replication system and the nucleotide sequence of this origin region determined. The most striking feature of the sequence is the presence of six tandem 22-bp nucleotide sequence direct repeats.     

Anin vivo cointegrate of two plasmids from incompatibility group X

Formation of a recombinant plasmid designated pNH603 was observed when two plasmids from incompatibility group X, the multicopy plasmid pNH602 (a higher-copy-number deletion derivative of R6K) and the oligocopy plasmid R485, coexisted in a single Escherichia coli cell. According to its size and its restriction endonuclease cleavage pattern, plasmid pNH603 is a true cointegrate of pNH602 and R485. An insertion-sequence-like element coming from plasmid R485 is supposed to mediate the fusion of both replicons. The pNH603 copy number (1-2 per chromosome) indicates that the mechanism of replication of the low-copy-number plasmid is dominant in this cointegrate. No dissociation of pNH603 to parental plasmids was observed even in E. coli K-12 recA+ cells. On the other hand, deletion derivatives of four size classes originate from pNH603 in both recA+ and recA hosts. A miniplasmid designated pNH604, a representative of the most frequent 7 Mg/mol size class, was found, in a low number of copies per host chromosome.     

Polymorphism in exon 10 of the human coagulation factor V gene in a population at risk for sickle cell disease

In Caucasians, the R506Q mutation in exon 10 of the factor V gene (FV Leiden) confers an increased risk of thromboembolism. We have scanned this region of the gene for possible mutations in 450 subjects from populations at risk for sickle cell disease (SCD). The R506Q mutation was absent in subjects from sub-Saharan Africa, whereas its allelic frequency was 2.5% in the West Indies. Only one other substitution with no functional consequences in vitro (R485K) was found (32.4% allelic frequency) in sub-Saharan Africa. Thus, we found no mutations in exon 10 of the FV gene constituting an additional risk factor for thrombosis in SCD in sub-Saharan Africa. This suggests that the putative selective advantage conferred by R506Q does not exist in these populations, unless R485K has functional consequences in vivo. It further suggests that R506Q in American Africans is of Caucasian origin. Our data are the first to document ethnic variations in the frequency of the R485K polymorphism. Received: 16 December 1996 / Accepted: 16 March 1997     

Determination of very thin pili by the bacterial drug resistance plasmid R485.

The IncX bacterial drug resistance plasmid R485 was found by electron microscopy to determine numerous very thin filaments (designated 485 pili) only 5.0 nm thick. They exhibited a characteristic helical structure (pitch, 4.6 nm), and were able to form large pseudocrystals when detached from the cell. The concomitant transfer of both pili and the sulfonamide resistance determinant of R485 to RecA strains of Escherichia coli confirmed that the pilus determinant was part of the plasmid and had not been mobilized from the chromosome of the host strain. An extensive examination failed to reveal any similar pili on strains carrying the IncX type plasmid R6K.     

Delineating Anopheles gambiae coactivator associated arginine methyltransferase 1 automethylation using top–down high resolution tandem mass spectrometry

Coactivator‐associated arginine methyltransferase 1 (CARM1), originally defined as a coactivator for steroid receptors, is a member of the protein arginine methyltransferases. Here, we report the discovery and characterization of an automethylation event by AgCARM1, a CARM1 homologue in the mosquito Anopheles gambiae, using top–down high resolution tandem mass spectrometry, which allows fine mapping of modifications in the intact protein accurately and quantitatively without priori knowledge. Unexpectedly, we found that AgCARM1 has already been predominantly dimethylated during its expression in Escherichia coli. A single arginine methylation site, R485, was identified which is conserved among CARM1 in insects. No methylation was observed in the intact AgCARM1^R485K mutant where R485 is mutated to lysine, which confirms that R485 is the only detectable methylation site. Using AgCARM1 methyltransferase defective mutants, we confirmed that this is an automethylation event and show the automethylation of AgCARM1 occurs intermolecularly. This study represents the first comprehensive characterization of an automethylation event by top–down mass spectrometry. The unexpected high percentage of automethylated recombinant AgCARM1 expressed in E. coli may shed light on other bacterially expressed post‐translational modifying enzymes, which could be modified but overlooked in biochemical and structural studies. Top–down high resolution tandem mass spectrometry thus provides unique opportunities for revealing unexpected protein modification, localizing specific modification to one amino acid, and delineating molecular mechanism of an enzyme.     

Five control systems preventing transfer of Escherichia coli K-12 sex factor F.

The transfer inhibition systems of 28 Fin+ plasmids have been characterized, using Flac mutants insensitive to inhibition by R100 or R62. All F-like plasmids (except R455) and one N group plasmid determined systems analogous to that of R100; this is designated the FinOP system. None of these plasmids could supply a FinP component of the transfer inhibitor able to replace that of F itself. In addition to the FinOP and R62 transfer inhibition systems described previously, new systems were encoded by the F-like plasmid R455, the I-like plasmid JR66a, and the group X plasmid R485. Besides inhibiting F transfer, JR66a also inhibited F pilus formation and surface exclusion, whereas R485 inhibited only pilus formation and R455 inhibited neither. All three R factors inhibited transfer of J-independent Flac elements, indicating that they act directly on one or more genes (or products) of the transfer operon, rather than directly via traJ. The tral products and transfer origin sequences of two Fin+ F-like plasmids, ColB2 and R124, appear to have similar specificities to those of F itself.     

Isolation and characterization of a variant myoblast cell line that is temperature sensitive for differentiation.

A new variant rat myogenic cell line, ts485, was isolated by subcloning the cell line ts3b2 (H. T. Nguyen, R. M. Medford, and B. Nadal-Ginard, Cell 34:281-293, 1983). Unlike the progenitor cell line, ts485 was thermosensitive for differentiation. Experiments with conditioned medium suggested that diffusible extracellular factors were not involved in dictating the differential phenotypes of ts485 cells cultured at the permissive and nonpermissive temperatures. Temperature shift experiments performed on cultures of ts485 cells indicated that the temperature-sensitive lesion was in a factor active during the growth phase and required to trigger a cascade of events leading to terminal differentiation.     

Further characterization of the F fertility inhibition systems of "unusual" Fin+ plasmids.

Flac mutants insensitive to transfer inhibition by R factors. JR66a and R485 were isolated and characterized. Representative mutations were cis dominant and are therefore presumed to be at the sites of action, fisU and fisV, respectively, of the FinU and FinV transfer inhibition systems encoded by JR66a and R485. The mutants were used to confirm that the FinU and FinV fertility inhibition systems are different from each other and from the FinOP, FinQ, and FinW systems of R100, R62, and R455, respectively. Together with traO and fisQ mutants of Flac, the new mutants were also used to investigate the nature of the F fertility inhibition systems encoded by a further group of "unusual" Fin+ plasmids. Of these, two incompatibility group X plasmids were found to carry finO+ genes, and of five incompatibility group I plasmids, three encoded FinQ systems, one the FinU system, and one a new system (FinR). Transfer of a variety of derepressed F-like plasmids was inhibited by the FinQ, FinU, and FinV systems, but a quantitatively very different levels; this emphasizes the differences as well as the similarities between the conjugation systems of F-like plasmids.     

Influence of Eight Unclassified Missense Variants of the MLH1 Gene on Lynch Syndrome Susceptibility

Missense mutations in MLH1 have frequently been detected in patients with Lynch syndrome, but their genetic significance has not been extensively assessed. In this study, we attempt to evaluate the etiological role of eight MLH1 missense variants. The variants were analyzed for their ability to affect MLH1 protein interaction with its partner PMS2 in vivo employing a yeast two-hybrid system. In addition, a SIFT (sorting intolerant from tolerant) algorithm was adopted to predict the effects of amino acid substitutions. Finally, scanning of mutations in a normal Chinese population and assay of the clinical characteristics have all been taken into account. Our results demonstrated that the MLH1 variants D485E and L653R cause functional alterations of the human MutLα complex significantly. The R265C, D304V, A586P, and R755S variants affect partial interaction. The remaining two variants, N38D and L559R, could be nonfunctional polymorphisms or might affect the mismatch repair system through other mechanisms.     

Functional characterization of mutations in inherited human cPLA₂ deficiency

Group IVA cytosolic phospholipase A(2) (cPLA(2)α) catalyzes the first step in the arachidonic acid cascade leading to the synthesis of important lipid mediators, the prostaglandins and leukotrienes. We previously described a patient deficient in cPLA(2)α activity, which was associated with mutations in both alleles encoding the enzyme. In this paper, we describe the biochemical characterization of each of these mutations. Using saturating concentrations of calcium, we showed that the R485H mutant was nearly devoid of any catalytic activity, that the S111P mutation did not affect the enzyme activity, and that the known K651R polymorphism was associated with activity slightly higher than that of the wild type. Using MDCK cells, we showed that translocation to the Golgi in response to serum activation was impaired for the S111P mutant but not for the other mutants. Using immortalized mouse lung fibroblasts lacking endogenous cPLA(2)α activity, we showed that both mutations S111P and R485H/K651R caused a profound defect in the enzyme catalytic activity in response to cell stimulation with serum. Taken together, our results show that the S111P mutation hampers calcium binding and membrane translocation without affecting the catalytic activity, and that the mutation R485H does not affect membrane translocation but blocks catalytic activity that leads to inactivation of the enzyme. Interestingly, our results show that the common K651R polymorphism confers slightly higher activity to the enzyme, suggesting a role of this residue in favoring a catalytically active conformation of cPLA(2)α. Our results define how the mutations negatively influence cPLA(2)α function and explain the inability of the proband to release arachidonic acid for eicosanoid production.     

Ca(2+)-binding site in Rhodobacter sphaeroides cytochrome C oxidase

Cytochrome c oxidase (COX) from R. sphaeroides contains one Ca(2+) ion per enzyme that is not removed by dialysis versus EGTA. This is similar to COX from Paracoccus denitrificans [Pfitzner, U., Kirichenko, A., Konstantinov, A. A., Mertens, M., Wittershagen, A., Kolbesen, B. O., Steffens, G. C. M., Harrenga, A., Michel, H., and Ludwig, B. (1999) FEBS Lett. 456, 365-369] and is in contrast to the bovine oxidase, which binds Ca(2+) reversibly. A series of R. sphaeroides mutants with replacements of the E54, Q61, and D485 residues, which form the Ca(2+) coordination sphere in subunit I, has been generated. The substitutions for the E54 residue do not assemble normally. Mutants with the Q61 replacements are active and retain the tightly bound Ca(2+); their spectra are not perturbed by added Ca(2+) or EGTA. The D485A mutant is active, binds to Ca(2+) reversibly, like the mitochondrial oxidase, and exhibits the red shift in the heme a absorption spectrum upon Ca(2+) binding for both reduced and oxidized states of heme a. The K(d) value of 6 nM determined by equilibrium titrations is much lower than that reported for the homologous D477A mutant of Paracoccus denitrificans or for bovine COX (K(d) = 1-3 microM). The rate of Ca(2+) binding with the D485A oxidase (k(on) = 5 x 10(3) M(-1) s(-1)) is comparable to that observed earlier for bovine COX, but the off-rate is extremely slow (approximately 10(-3) s(-1)) and highly temperature-dependent. The k(off) /k(on) ratio (190 nM) is about 30-fold higher than the equilibrium K(d) of 6 nM, indicating that formation of the Ca(2+)-adduct may involve more than one step. Sodium ions reverse the Ca(2+)-induced red shift of heme a and dramatically decrease the rate of Ca(2+) binding to the D485A mutant COX. With the D485A mutant, 1 Ca(2+) competes with 1 Na(+) for the binding site, whereas 2 Na(+) compete with 1 Ca(2+) for binding to the bovine oxidase. This finding indicates that the aspartic residue D442 (a homologue of R. sphaeroides D485) may be the second Na(+) binding site in bovine COX. No effect of Ca(2+) binding to the D485A mutant is evident on either the steady-state enzymatic activity or several time-resolved partial steps of the catalytic cycle. It is proposed that the tightly bound Ca(2+) plays a structural role in the bacterial oxidases while the reversible binding with the mammalian enzyme may be involved in the regulation of mitochondrial function.     

Overexpression of wild-type and catalytically inactive forms of GRK2 and GRK6 fails to alter the agonist-induced phosphorylation of the C5a receptor (CD88): evidence that GRK6 is autophosphorylated in COS-7 cells.

The G protein-coupled receptor kinase family comprises six members (GRK1 to GRK6) that phosphorylate and desensitize a number of agonist-occupied G protein-coupled receptors. Overexpression of the dominant negative mutant GRK2-K220R is often accompanied by an inhibition of the agonist-mediated phosphorylation of G protein-coupled receptors. In the case of the C5a receptor (C5aR), the overexpression of wild-type GRK2 or GRK6 as well as of catalytically inactive forms of these kinases (GRK2-K220R and GRK6-K215R) failed to increase or to inhibit the agonist-mediated phosphorylation of C5aR, respectively. Replacement of Lys215 by an arginine residue in GRK6 yielded a protein with a relative molecular mass of 63 kDa, whereas wild-type GRK6 had a relative molecular mass of 66 kDa on polyacrylamide gel. The mutations S484D and T485D in the catalytically inactive mutant GRK6-K215R resulted in a protein (GRK6-RDD) with the same electrophoretic mobility as wild-type GRK6. Furthermore, in the absence of phosphatase inhibitors, GRK6 was rapidly converted into the 63 kDa species, whereas GRK6-RDD was not. Overepression of GRK6-RDD failed to alter the agonist-mediated phosphorylation of C5aR. Taken together, the results suggest that C5aR is not a substrate for either GRK2 or GRK6 and that GRK6 is very likely autophosphorylated on Ser484 and Thr485 in vivo.     

Isolation and characterization of a higher-copy-number mutant of plasmid R6K

A stable copy-number mutant (pNH601) of plasmid R6K was isolated by selection for increased resistance to ampicillin determined by this plasmid. The size of the mutant plasmid was found to be unchanged (26 Mg/mol) but it is present in 27 copies of pNH601 perE. coli K-12 chromosome which represents a two-fold increase of R6K copy number value. The following genetic properties of pNH601 are reported and compared with those of R6K: conjugative transfer, fertility inhibition of plasmids belonging to other incompatibility groups, incompatibility with plasmid R485 under both non-selective and selective conditions and the integrative suppression of thednaA ts mutation. The mutant plasmid pNH601 was found to be different from the original R6K in most of these properties.     

Activation of cardiac AMP-activated protein kinase by LKB1 expression or chemical hypoxia is blunted by increased Akt activity

AMP-activated protein kinase (AMPK) plays a major role in the regulation of cardiac energy substrate utilization and can be negatively regulated by Akt activation in the heart. It has recently been shown that Akt directly phosphorylates AMPKalpha(1)/alpha(2) on Ser(485/491) in vitro and prevents the AMPK kinase (AMPKK) LKB1 from phosphorylating AMPKalpha at its primary activation site, Thr(172) (S Horman, D Vertommen, R Heath, D Neumann, V Mouton, A Woods, U Schlattner, T Wallimann, D Carling, L Hue, and MH Rider. J Biol Chem 281: 5335-5340, 2006). To determine whether this is also the case in the cardiac myocyte, neonatal rat cardiac myocytes (NRCM) were infected with a recombinant adenovirus expressing a constitutively active mutant of Akt1 (myrAkt1) and then with or without adenoviruses expressing the active LKB1 complex. Expression of myrAkt1 blunted LKB1-induced phosphorylation of AMPKalpha at Thr(172), which resulted in a dramatic decrease in phosphorylation of AMPK's target, acetyl CoA-carboxylase. This decrease in AMPK activity was associated with prior Akt1-dependent phosphorylation of AMPKalpha(1)/alpha(2) at Ser(485/491). To investigate whether Akt1 activation was also able to prevent other AMPKKs from phosphorylating AMPKalpha, we subjected NRCM to chemical hypoxia and noted a marked increase in phosphorylation of AMPKalpha at Thr(172), despite no change in LKB1 activity. NRCM expressing myrAkt1 demonstrated increased phosphorylation of AMPKalpha(1)/alpha(2) at Ser(485/491) and a complete inhibition of chemical hypoxia-induced phosphorylation of AMPKalpha at Thr(172). Taken together, our data show that activation of Akt1 is able to prevent activation of cardiac AMPK by LKB1 and at least one other AMPKK, likely by prior phosphorylation of AMPKalpha(1)/alpha(2) at Ser(485/491).     

Generation and analyses of R8L barttin knockin mouse

Barttin, a gene product of BSND, is one of four genes responsible for Bartter syndrome. Coexpression of barttin with ClC-K chloride channels dramatically induces the expression of ClC-K current via insertion of ClC-K-barttin complexes into plasma membranes. We previously showed that stably expressed R8L barttin, a disease-causing missense mutant, is retained in the endoplasmic reticulum (ER) of Madin-Darby canine kidney (MDCK) cells, with the barttin β-subunit remaining bound to ClC-K α-subunits (Hayama A, Rai T, Sasaki S, Uchida S. Histochem Cell Biol 119: 485-493, 2003). However, transient expression of R8L barttin in MDCK cells was reported to impair ClC-K channel function without affecting its subcellular localization. To investigate the pathogenesis in vivo, we generated a knockin mouse model of Bartter syndrome that carries the R8L mutation. These mice display disease-like phenotypes (hypokalemia, metabolic alkalosis, and decreased NaCl reabsorption in distal tubules) under a low-salt diet. Immunofluorescence and immunoelectron microscopy revealed that the plasma membrane localization of both R8L barttin and the ClC-K channel was impaired in these mice, and transepithelial chloride transport in the thin ascending limb of Henle's loop (tAL) as well as thiazide-sensitive chloride clearance were significantly reduced. This reduction in transepithelial chloride transport in tAL, which is totally dependent on ClC-K1/barttin, correlated well with the reduction in the amount of R8L barttin localized to plasma membranes. These results suggest that the major cause of Bartter syndrome type IV caused by R8L barttin mutation is its aberrant intracellular localization.     

Autophosphorylation of threonine 485 in the activation loop is essential for attaining eIF2alpha kinase activity of HRI

In heme deficiency, protein synthesis is inhibited by the activation of the heme-regulated eIF2alpha kinase (HRI) through its multiple autophosphorylation. Autophosphorylation sites in HRI were identified in order to investigate their functions. We found that there were eight major tryptic phosphopeptides of HRI activated in heme deficiency. In this report we focused on the role of autophosphorylation at Thr483 and Thr485 in the activation loop of HRI. Disruption of the autophosphorylation of Thr485, but not Thr483, resulted in a lower autokinase activity and locked Thr485Ala HRI in a hypophosphorylated state. Most importantly, autophosphorylation of Thr485, but not Thr483, was essential for attaining eIF2alpha kinase activity of HRI. In addition, autophosphorylation of Thr485 was necessary for arsenite-induced activation of the eIF2alpha kinase activity of HRI, while autophosphorylation at Thr483 was not required for activation by arsenite. The function of Thr490, another conserved Thr residue in the activation loop of HRI, was also investigated. Mutations of Thr490 to either Ala or Asp resulted in reduced autokinase activity and loss of eIF2alpha kinase activity in heme deficiency or upon arsenite treatment. Since Thr490 was not identified as an autophosphorylated site, it is likely that Thr490 itself might be critical for the catalytic activity of HRI. Importantly, Thr485 was very poorly phosphorylated in Thr490 mutant HRI. Collectively, our results demonstrate that autophosphorylation of Thr485 is essential for the hyperphosphorylation and activation of HRI and is required for the acquisition of the eIF2alpha kinase activity.     

Uptake, transport and regulation of JBP485 by PEPT1 in vitro and in vivo

Cyclo-trans-4-l-hydroxyprolyl-l-serine (JBP485) is a dipeptide with anti-hepatitis activity that has been chemically synthesized. Previous experiments in rats showed that JBP485 was well absorbed by the intestine after oral administration. The human peptide transporter (PEPT1) is expressed in the intestine and recognizes compounds such as dipeptides and tripeptides. The purposes of this study were to determine if JBP485 acted as a substrate for intestinal PEPT1, and to investigate the characteristics of JBP485 uptake and transepithelial transport by PEPT1. The uptake of JBP485 was pH dependent in human intestinal epithelial cells Caco-2. And JBP485 uptake was also significantly inhibited by glycylsarcosine (Gly-Sar, a typical substrate for PEPT1 transporters), JBP923 (a derivative of JBP485), and cephalexin (CEX, a β-lactam antibiotic and a known substrate of PEPT1) in Caco-2 cells. The rate of apical-to-basolateral transepithelial transport of JBP485 was 1.84 times higher than that for basolateral-to-apical transport. JBP485 transport was obviously inhibited by Gly-Sar, JBP923 and CEX in Caco-2 cells. The uptake of JBP485 was increased by verapamil but not by cyclosporin A (CsA) and inhibited by the presence of Zn²⁺ or the toxic metabolite of ethanol, acetaldehyde (AcH) in Caco-2 cells. The in vivo uptake of JBP485 was increased by verapamil and decreased by ethanol in vivo, which was consisted with the in vitro study. PEPT1 mRNA levels were enhanced after exposure of the cells to JBP485 for 24 h, compared to control. In conclusion, JBP485 was actively transported by the intestinal oligopeptide transporter PEPT1. This mechanism is likely to contribute to the rapid absorption of JBP485 by the gastrointestinal tract after oral administration.     

Are NADP-dependent isocitrate dehydrogenases and ferredoxin-dependent glutamate synthase co-regulated by the same photoreceptors?

Activities of NADP-dependent isocitrate dehydrogenases (cytosolic and plastidic isoforms, ICDH1 and ICDH2; EC 1.1.1.42) and ferredoxin-dependent glutamate synthase (Fd-GOGAT; EC 1.4.7.1) in turions of Spirodela polyrhiza were all stimulated by light. Single or repeated red light (R) pulses induced the activity of the enzymes and this effect was reverted by subsequent far-red light (FR) pulses. The enzymes are, therefore, co-regulated by the low-fluence response of phytochrome. For ICDH, this is reported here for the first time. Neither an effect of the very low-fluence response nor of the FR-mediated high-irradiance response was detectable. Irradiance with continuous R resulted in enhanced enzyme activities and protein levels (Western analysis using polyclonal antibodies against ICDH1 and Fd-GOGAT). These additional effects of continuous R (called a non-induction effect) could be inhibited for ICDH1 and ICDH2 by the inhibitor of photosynthetic electron transport, 3-(3,4-dichlorophenyl)-1,1-dimethylurea, and are therefore related to the effect of photosynthesis. In contrast, the non-induction effect of Fd-GOGAT was resistant against this inhibitor. Moreover, hourly R pulses did not replace the effect of continuous R. The non-induction effect of light on the activity and protein level of Fd-GOGAT was therefore tentatively classified as an R-mediated high-irradiance response. The activity of Fd-GOGAT but not that of ICDHs was additionally regulated by a specific blue-light receptor. It can be concluded that the levels of ICDHs and Fd-GOGAT were coordinated by light but were not co-regulated by the same photoreceptors. Nitrate is necessary for the light regulation of both enzymes, contributing to the coordinated expression of the relevant genes.Abbreviations DCMU 3-(3,4-Dichlorophenyl)-1,1-dimethylurea - Fd-GOGAT Ferredoxin-dependent glutamate synthase - FR Far-red light - HIR High-irradiance response - ICDH NADP-dependent isocitrate dehydrogenase - ICDH1 Cytosolic ICDH - ICDH2 Chloroplastic ICDH - LFR Low-fluence response - R Red light - SDS–PAGE Denaturing polyacrylamide gel electrophoresis - VLFR Very low-fluence response     

New site(s) of methylglyoxal-modified human serum albumin, identified by multiple reaction monitoring, alter warfarin binding and prostaglandin metabolism

Methylglyoxal (MG) is a biologically reactive byproduct of glucose metabolism, levels of which increase in diabetes. MG modification of protein generates neutral hydroimidazolone adducts on arginine residues which can alter functional active sites. We investigated the site-specificity of MG adduction to human serum albumin (HSA) using multiple reaction monitoring (MRM) of 13 MG-modified tryptic peptides, each containing an internal arginine. Seven new sites for MG modification (R257>R209>R222>R81>R485>R472>R10) are described. Analysis of MG-treated HSA showed substantial R257 and R410 modification, with MG-modified R257 (at 100μM MG) in drug site I causing significant inhibition of prostaglandin catalysis. The MG hydroimidazolone (MG-H1) adduct was modeled at R257, and molecular dynamics simulations and affinity docking revealed a decrease of 12.8-16.5kcal/mol (S and R isomers, respectively) for warfarin binding in drug site I. Taken together, these results suggest that R257 is a likely site for MG modification in vivo, which may have functional consequences for prostaglandin metabolism and drug bioavailability.