2-(Oxalylamino)-Benzoic Acid Is a General, Competitive Inhibitor of Protein-tyrosine Phosphatases*

Protein-tyrosine phosphatases (PTPs) are critically involved in regulation of signal transduction processes. Members of this class of enzymes are considered attractive therapeutic targets in several disease states, e.g. diabetes, cancer, and inflammation. However, most reported PTP inhibitors have been phosphorus-containing compounds, tight binding inhibitors, and/or inhibitors that covalently modify the enzymes. We therefore embarked on identifying a general, reversible, competitive PTP inhibitor that could be used as a common scaffold for lead optimization for specific PTPs. We here report the identification of 2-(oxalylamino)-benzoic acid (OBA) as a classical competitive inhibitor of several PTPs. X-ray crystallography of PTP1B complexed with OBA and related non-phosphate low molecular weight derivatives reveals that the binding mode of these molecules to a large extent mimics that of the natural substrate including hydrogen bonding to the PTP signature motif. In addition, binding of OBA to the active site of PTP1B creates a unique arrangement involving Asp181, Lys120, and Tyr46. PTP inhibitors are essential tools in elucidating the biological function of specific PTPs and they may eventually be developed into selective drug candidates. The unique enzyme kinetic features and the low molecular weight of OBA makes it an ideal starting point for further optimization.

Reversible tyrosine phosphorylation reactions play pivotal roles in most cellular signaling processes. Protein-tyrosine kinases phosphorylate cellular substrates on tyrosine residues and protein-tyrosine phosphatases (PTPs) 1 remove phosphate from these residues (for reviews, see Refs. 1 and 2). It is generally believed that low molecular weight, selective PTP inhibitors may be useful in the treatment of a variety of diseases such as diabetes, autoimmunity, and cancer (1,2).
Recent studies have provided important insight to some of the basic structural requirements for PTP-substrate and -inhibitor interactions (reviewed in Refs. [3][4][5][6][7][8]. It is particularly encouraging that highly selective, active site-directed PTP1B inhibitors have been reported (9). However, although significant progress has been made toward developing PTP inhibitors, most compounds have features that make them unsuitable as starting points for optimization to orally active drugs (for a recent review, see Ref. 6). As an example, peroxovanadium compounds have contributed significantly to our understanding of insulin signaling (10,11), but appear to be too toxic and unspecific for use as drugs. Bisphosphonates, such as alendronate, have been shown to inhibit PTPs, but their inherent affinity for bone is likely to prevent their general use in other target tissues (12). Furthermore, several of these inhibitors are time-dependent and seem to act through covalent modification of the catalytic cysteine in PTPs (13,14). The most specific inhibitors produced so far have either been phosphonates (9,15) or those based on peptides that are not suited for clinical use due to their lack of oral bioavailability and metabolic instability (reviewed in Ref. 6).
The phosphonate requirements of most of the previously described PTP inhibitors have their natural justification, in that they have all addressed the active site of PTPs and thus mimic substrate binding. This site comprise the phosphate binding P-loop Cys 215 -(X) 5 -Arg 221 in the base of the site, extending from a ␤-strand (␤12) going over into an ␣-helix (␣4) (16). Several other conserved loop residues participate in substrate and inhibitor binding, including the WPD loop residues Asp 181 -Phe 182 , and residues Tyr 46 , Val 49 , Lys 120 , and Gln 262 . At the rim of the active site, Arg 47 and Asp 48 are located giving rise to a total depth of 8 -9 Å for the site (16).
Based on the potential therapeutic usefulness of selective, non-peptide inhibitors we initiated high throughput screening of a diverse compound library using PTP1B and a synthetic 33 P-phosphorylated peptide as substrate. It was our aim to identify compounds that fulfilled a set of selection criteria, which would allow further optimization. We decided to include general PTP inhibitors only. The main reason for this criteria is our wish to develop a set of inhibitors that selectively inhibit different PTPs. Therefore, the common starting point should be a general PTP inhibitor that could be optimized for selectivity and potency using structure-based approaches. To secure such a general mechanism of inhibition, the compounds should be active site-directed reversible inhibitors and show classical competitive inhibition. In particular, we wanted to avoid timedependent inhibitors that might be difficult to optimize for specificity against different PTPs utilizing structural information for the target PTPs. Also, since it is our goal to develop inhibitors that might be orally active drug candidates, only low molecular weight compounds were included. 2-(Oxalylamino)benzoic acid (Fig. 1, compound 1) was identified during these studies and found to be one of the most efficient phenyl phosphate mimetics identified so far. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  ¶ Contributed equally to this report. 1 The abbreviations used are: PTP, protein-tyrosine phosphatase; F 2 Pmp, difluorophosphonomethyl phenylalanine; OBA, 2-(oxalylamino)-benzoic acid; Tyr(P), phosphotyrosine; SH2, Src homology domain 2.
Cloning, Expression, and Purification-cDNAs encoding the catalytic domains of PTP1B (17), SHP-1 (18), PTP␣ (19), PTP⑀ (19), PTP␤ (19), and CD45 (20) were obtained by polymerase chain reaction using primers with convenient cloning sites and appropriate cDNA templates. The cDNAs were inserted into prokaryotic expression vectors as detailed in Table I. All PTP coding sequences were confirmed by DNA sequencing. The PTP-LAR expression vector was a kind gift from M. Streuli, Boston, MA. The constructs were inserted into pGEX expression vectors (Amersham Pharmacia Biotech) and transformed into Escherichia coli BL21 (Amersham Pharmacia Biotech). Overnight cultures were diluted 1:25 and grown for 3 h at 37°C before addition of isopropyl-␤-D-thiogalactoside to a final concentration of 0.1 mM. The glutathione S-transferase fusion proteins were purified according to the manufacturer's instructions (Amersham Pharmacia Biotech). For x-ray crystallography, a pET11a expression plasmid encoding the first 321 amino acids of PTP1B was transformed into E. coli BL21(DE3). An overnight culture was diluted 1:160 into a total volume of 2 liters of SOB medium (2% (w/v) Bacto-tryptone, 0.5% (w/v) Bacto-yeast extract, 10 mM NaCl, 10 mM MgCl 2 , and 10 mM MgSO 4 ) and grown at 37°C until the A 600 was 0.6. Isopropyl-␤-D-thiogalactoside was added to a final concentration of 0.1 mM, and the incubation was continued at room temperature for 3 h. PTP1B was purified essentially as described previously (21).
High Throughput Screening Assay-A scintillation proximity technology based high-throughput screening assay utilizing recombinant PTP1B (17) and a biotinylated 33 P-labeled peptide substrate was used to screen the Novo Nordisk compound library (to be reported elsewhere).
Determination of Inhibitor Constants, K i -The enzyme reactions were carried out using standard conditions essentially as described by Burke et al. (22). The assay conditions were as follows. Appropriately diluted inhibitors (4 different concentrations: diluted 1-, 3-, 9-, and 27-fold) were added to the reaction mixtures containing different concentrations of the substrate, p-nitrophenyl phosphate (usual range: 0.31 to 20 mM, final assay concentration). The buffer used was 100 mM sodium acetate, pH 5.5, 50 mM sodium chloride, 0.1% (w/v) bovine serum albumin, 5 mM glutathione, and 1 mM EDTA (total volume 100 l). The reaction was started by addition of the enzyme and carried out in microtiter plates at 25°C for 60 min. The reactions were stopped by addition of NaOH. The enzyme activity was determined by measurement of the absorbance at 405 nm with appropriate corrections for absorbance of the compounds and p-nitrophenyl phosphate. The data were analyzed using nonlinear regression hyperbolic fit to classical Michaelis-Menten enzyme kinetic models. Inhibition is expressed as K i values in micromolar.
Time Dependence-The same assay format as that described above for K i determinations was used. The inhibitor, compound 1, was used at different concentrations (62.5, 125, and 250 M), whereas the concentration of substrate, p-nitrophenyl phosphate was kept constant at 2.5 mM. The reaction was started by addition of enzyme and then stopped at 3-min intervals by addition of NaOH. The buffer used in these experiments was 100 mM sodium acetate, pH 5.5, 100 mM NaCl, 5 mM dithiothreitol, and 0.1% (w/v) bovine serum albumin with further addition of 1 mM EDTA or 25 g/ml catalase as indicated.
Co-crystallization of PTP1B with Inhibitors-A 6 -10 mg/ml preparation of PTP1B in 10 mM Tris, pH 7.5, 25 mM NaCl, 0.2 mM EDTA, and 3 mM dithiothreitol, was used for crystallization. Crystals were grown by the sitting, as well as the hanging drop, vapor diffusion methods. A 1:10 (PTP1B:inhibitor) molar ratio mixture was prepared at least 1 h prior to crystallization. Two l of PTP1B-inhibitor solution was mixed with 2 l of reservoir solution consisting of 0.1 M Hepes buffer, pH 7.5, 0.3-0.4 M sodium acetate or magnesium acetate, 12-16% (w/v) polyethylene glycol 8000 and/or 4% (v/v) glycerol. The reservoir volume was 1 ml. Crystals grew to the size of 0.
Data Collection-All diffraction data collections were performed at 100 K. The following cryo conditions were used: to the hanging or sitting drop 3 l of 50% (v/v) glycerol (containing 0.5 mmol of inhibitor) were added. The crystal was removed from the drop after 5-30 min and transferred to 50% glycerol (containing 0.5 mmol inhibitor) and flash frozen. Data were collected using a Mar345 image plate detector either at the MAX-lab synchrotron facilities at Lund University or in house on a rotating anode (RU300, CuK␣ 50kV/80 mA) equipped with Osmic multilayer mirror system. Typically a 1°oscillation per image was used for 60 images. Data sets in the resolution range 2.7-1.8 Å were obtained. The space group was determined to be P3 1 21 for all crystals used. Data processing was performed using Denzo, Scalepack, and the CCP4 program suite (23,24).
Refinements-As P3 1 21 contains a polar axis and, thus, possesses more than one indexing possibility, a molecular replacement solution using Amore (24,25) was determined prior to the refinements. A high resolution PTP1B structure was used as a starting model (PDB file: 1PTV (26)), with ligand and water molecules omitted from the structure. All refinements were performed with Xplor. version 3.851 (Molecular Simulations Inc., MSI). Interchanging cycles of model building using Quanta X-build (MSI) and refinement were performed. The 2F o Ϫ F c maps were inspected by the use of X-ligand (MSI) at a 1.3 level for densities that could correspond to the structures of the inhibitors. In all cases, a well suited inhibitor electron density was identified in the active site pocket (see below). No other densities were identified to fit the inhibitors. Water molecules were inserted using the X-solvate program (MSI) based on 1.5 2F o Ϫ F c electron density maps. For further details see Table II.
Compound Synthesis-Substituted anthranilic acids or esters were treated with ethyl oxalyl chloride in THF yielding 2-(oxalylamino)benzoic acid mono-or diesters. Base hydrolysis of the ester group(s) gave after treatment with aqueous HCl the desired 2-(oxalylamino)benzoic acids 1 and 5 (Fig. 1). Using the same procedure, the naphthalene analogue 2 was prepared. The indole-based compounds 3 and 4 were synthesized applying the same procedure as described for 1 using either 6-amino-1H-indole-5-carboxylic acid ethyl ester or 6-amino-1Hindole-7-carboxylic acid ethyl ester as starting material prepared as described by Showalter and co-workers (27). Further details regarding synthesis and SAR of this family of compounds will be published elsewhere.

High Throughput Screening for General, Competitive PTP
Inhibitors-Our goal was to develop selective low molecular weight PTPase inhibitors that potentially could be used for treatment of diverse disease states, such as diabetes, inflammation, and cancer. As a starting point, we therefore desired a lead compound with the following characteristics. It should (i) be a general inhibitor mimicking the binding of tyrosine phosphate (Tyr(P)); (ii) be a competitive, reversible active site inhibitor; and (iii) have a molecular weight below 300 to leave room for further optimization for potency and selectivity.
Using a scintillation proximity technology-based high throughput screening assay, several inhibitors were identified, some being time-dependent inhibitors, and some were co-

2-(Oxalylamino)-Benzoic Acid 1 Is a Competitive, Reversible, Active Site Inhibitor of PTP1B-
The enzyme kinetic behavior of OBA was found to be a textbook example of a classical competitive inhibitor with a K i value ϳ20 M against PTP1B at pH 5.5 (Fig. 2). It appears that OBA behaves as a non-cleavable tyrosine phosphate mimetic. While the enzyme kinetic behavior of OBA as a classical, competitive inhibitor is fully retained at neutral pH, the K i value obtained shows pH dependence with a 10-fold lower potency at pH 7.0 (K i ϭ 200 M, not shown). It is of specific interest, however, that the observed affinity of OBA for PTP1B, even at neutral pH, is about 10-fold higher than the affinity of the non-hydrolyzable analog of the natural substrate, ␣,␣-difluorobenzyl phosphonic acid (K i ϭ 2.5 mM) (28).
OBA Acts as a Time-independent Inhibitor-Other PTP inhibitors have been shown to be very sensitive to assay constituents such as EDTA and dithiothreitol. As an example, in a recent, thorough study it was shown that addition of EDTA or catalase prevented alendronate from inhibiting PTP1B and CD45, suggesting that a combination of alendronate, trace metal ions, and H 2 O 2 was responsible for the observed timedependent PTP inhibition of alendronate (14). We, therefore, analyzed if OBA inhibited PTP1B in a time-dependent manner. Fig. 3 shows that (a) there is no time dependence and (b) the inhibition is not influenced by the addition of EDTA or catalase. Taken together, our kinetic analyses clearly show that OBA acts as classical, time-independent, active site-directed, reversible competitive inhibitor that does not covalently modify PTP1B.
OBA Can Be Optimized-Previous studies have shown that naphthyl phosphonates are considerably more potent as PTP inhibitors than their phenyl phosphonate counterparts (29). The structural basis for the increased potency of inhibitors with an additional phenyl ring has been provided by x-ray crystallography showing additional hydrophobic interactions between ligand and enzyme (22,30). Therefore, as a first attempt to improve the potency of OBA we made compounds 2, 3, and 4 (Fig. 1). In agreement with the predictions, both the naphthyl-(2) and the indole-(3 and 4) based compounds showed increased affinity for PTP1B (Table III), although not nearly as dramatic as that observed for naphthyl phosphonate. Since part of the increased affinity most likely arises from hydrophobic interactions between the second phenyl ring of the inhibitors and hydrophobic side chains of PTP1B we also made compound 5. As shown in Table III, significant changes in the inhibitor profiles against the different PTPs are observed for these compounds. As an example, compound 3 (in comparison with OBA) shows about 30-fold increase in potency against SHP-1, but only a 2-fold increase against PTP1B. These differences are particularly noteworthy since all compounds, due to their small size, most likely address the active site binding pocket only.
OBA Is a General PTP Inhibitor-OBA was tested against a diverse set of catalytic domains representing 6 different PTP families. Table III shows that OBA inhibits most PTPs with PTP-LAR as a notable exception. Importantly, compound 2 shows considerable affinity for PTP-LAR, thus providing support for the notion that an aromatic carboxylic acid substituted with an ortho-positioned oxamic acid acts as a general PTP inhibitor.
X-ray Crystallography: Co-crystallization of PTP1B with OBA-Although the enzyme kinetic analyses clearly indicate that OBA and compounds 2-5 act as active site inhibitors, it was impossible from these experiments to map the exact binding mode, i.e. the information needed for future structurebased design and optimization. Therefore, we undertook cocrystallization studies of PTP1B and compound 1. A well suited density was identified in the active site of PTP1B. The 1 2F o Ϫ F c omit map of compound 1 in the active site is shown in Fig.  4, A and B. No other densities were identified to fit compound 1. The electron density allows an unambiguous fitting of 1 with the oxamic acid and the o-carboxy group clearly discernible.
Superimposing all equivalent atoms of PTP1B complexed with compound 1 and PTP1B C215S complexed with Tyr(P) (26), a root mean square deviation between C␣ atoms of 0.42 Å is observed. Hence no alterations in secondary and tertiary structure are detected. Minor side chain differences are ob- ͉F o ͉ Data from 6 Å to the high resolution limit were used for each data set in the R factor calculation.
where T is a test set containing a random 5% of the observations omitted from the refinement.

2-(Oxalylamino)-Benzoic Acid Is a General PTP Inhibitor
served, and those related to the active site binding pocket will be discussed below.
The overall ligand conformation of compound 1 is almost planar with both carboxylic acid groups possessing small twists out of the plane (these torsion angle twists are consistently found for all ligand structures determined, see Figs. 4 -7).
Binding to the P-loop-The PTP signature motif Cys 215 -(X) 5 -Arg 221 , the P-loop, is one excellent example of natures design of a highly efficient binding pocket for phosphate, forming a halfcircle with Cys 215 almost in the center. The main chain amides are pointing toward Cys 215 and together with the side chain of Arg 221 they form 6 hydrogen bonds and two salt bridges with the three distal oxygens of Tyr(P) (16,26). A similar binding pattern is observed for the PTP1B-orthovanadate complex (8). As seen in Fig. 5, the oxalylamino part of compound 1 share several of the above Tyr(P) interaction points within the Cys 215 -(X) 5 -Arg 221 motif, with the lack of hydrogen bonding to the main chain amide of Ala 217 (not shown) and Gly 218 as noticeable exceptions. The carboxy group of oxamic acid is positioned 2.9 -3.0 Å from the guanidinium group of Arg 221 forming a salt bridge, as well as a hydrogen bond with the main chain amide of Arg 221 and Ser 216 , and the carbonyl forms a hydrogen bond with the main chain amide of Gly 220 . Thus, it can be concluded that compound 1 behaves as a phosphate mimetic for binding in the active site of PTP1B and probably all PTPs, in accordance with our original selection criteria.
Movement of the WPD Loop-Binding of Tyr(P), tyrosinephosphorylated peptide substrates (26), bis-(para-phosphophenyl)methane (31), [1,1-difluoro-1-(2-naphthalenyl)methyl]phosphonic acid (22), and vanadate (8)   by aromatic-aromatic interactions between the phenyl ring of Tyr(P) and the side chain of Phe 182 . A similar movement of the WPD loop is observed when compound 1 binds to PTP1B. However, in addition to the backbone motion we also observe a change in the torsion angle of Phe 182 which brings the phenyl ring into a more favorable position for aromatic-aromatic interaction with the phenyl ring of 1 (see Fig. 5B). It should be noted that the aromatic cores of compound 1 and Tyr(P) are only partly overlapping (Fig. 5A).
The closure of the WPD loop brings the conserved Asp 181 into an apparently unfavorable position, i.e. only 2.9 Å from the o-carboxy group of compound 1 (see Fig. 6), which might lead to repulsion. However, it should be remembered that numerous biochemical studies have shown that Asp 181 functions as a general acid (Ref. 33; reviewed in Ref. 6) taking part in the hydrolysis of Tyr(P). In other words, Asp 181 acts as hydrogen donor in the first step of Tyr(P) hydrolysis and must, therefore, be protonated in a significant proportion of the molecules. Hence, Asp 181 can make hydrogen-bonding interaction with the o-carboxy group of compound 1. The proportion of PTP1B molecules with a protonated Asp 181 will be relatively low at neutral pH in comparison with pH maximum. In accordance with this, compound 1 was found to be about 6 -10-fold less potent at neutral pH than at pH 5.5 when analyzed with different PTPs including PTP1B (not shown). The reduced potency at neutral pH is then believed to be due to both loss of attractive forces and introduction of repulsive forces between Asp 181 and compound 1.
Interaction with Lys 120 and Tyr 46 -In comparison with previous x-ray crystallographic analyses of PTP1B (26), Lys 120 has moved approximately 1 Å to be in about 2.8 Å distance from the o-carboxy group of compound 1. To our knowledge, this is the first description of a significant conformational change of this highly conserved amino acid to accommodate a ligand. The pK a values for compound 1 have been determined to be 3.8 and 4.8, respectively (not shown). Thus, the o-carboxy group of compound 1 is likely to be fully deprotonated at neutral pH. This will allow the formation of a salt bridge between Lys 120 and compound 1. In addition, the o-carboxy group is within hydrogen bonding distance from the side chain of Tyr 46 . Clearly the o-carboxy group of compound 1 provides additional points of interactions in the active site binding cavity as compared with Tyr(P) (Fig. 6).
The Catalytic Water Molecule-As in PTP1B complexed with peptide substrate (26), we observe a water molecule similarly positioned and trapped under the WPD loop.
PTP1B Co-crystallized with Compounds 2, 3, and 5-All complexes determined have 100% ligand occupancy in the active site binding pocket. As seen for compound 1, the omit maps for compounds 2, 3 and 5 allow an easy ligand fitting in the electron density maps (Fig. 4).

DISCUSSION
Several attempts to make PTP inhibitors by replacing the phosphate functionality of Tyr(P) containing peptide substrates with non-hydrolyzable phosphate mimetics have been reported. Initial experiments included sulfotyrosyl (34 -38), phosphonates (15,39), and O-malonyltyrosine (40,41). In most studies, the Tyr(P) mimetics were incorporated in peptides known to be substrates for PTPases. In particular, one peptide derived from the epidermal growth factor receptor, EGF-R 922-993 (Ac-DADEXL-NH 2 , where X denotes Tyr(P) or Tyr(P) mimetics), has been useful (6). When difluorophosphonomethyl phenylalanine (F 2 Pmp) is introduced in the EGF-R 922-993 peptide, a potent inhibitor of PTP1B is obtained with K i ϭ 180 nM (15). Since the "minimal unit" of F 2 Pmp, ␣,␣-difluorobenzyl phosphonic acid, shows a K i of 2.5 mM for PTP1B (28), this suggests that the peptide part of the inhibitor contributes significantly to the overall binding affinity. Incorporation of Omalonyltyrosine and fluoro-O-malonyltyrosine in the Ac-DA-DEXL-NH 2 peptide leads to PTP1B inhibitors with IC 50 values of 10 and 1 M, respectively (42,43). It is difficult to make a direct comparison of OBA with the above phenyl phosphate/ Tyr(P) mimetics. Nonetheless, given the fact that OBA at neutral pH (without being incorporated in a peptide) displays a K i of 200 M for PTP1B would indicate that this compound is one of the most potent minimal unit active site PTP1B inhibitors identified so far.
The o-carboxylic acid of OBA occupies a different part of the active site when compared with the Tyr(P) substrate. A unique arrangement is thus formed with a salt bridge connecting the o-carboxylic acid and Lys 120 . Furthermore, the o-carboxylic acid is in hydrogen bond distance with Tyr 46 and Asp 181 . This arrangement is critical for the observed affinity for PTP1B since removal of the o-carboxy group from compound 1 increases the K i value to Ͼ2000 M (not shown). In this context, it is of interest that Burke and co-workers (41) recently showed that introduction of a o-carboxy group into 4-(O-carboxymethyl)-L-tyrosine increased the potency more than 100-fold when incorporated in the above Ac-DADEXL-NH 2 peptide.
Only one x-ray structure of PTP1B in complex with a nonphosphonate PTP inhibitor, a fluoromalonyl tyrosine-based cyclic peptide, has been reported (30). Although the fluoromalo-nyl tyrosine peptide binds to the PTP signature motif of the active site of PTP1B, this was not accompanied by closure of the WPD loop. In contrast, we found that the WPD loop closes in all structures reported here. This motion brings (i) the side chain of Phe 182 into a favorable position for aromatic-aromatic interaction with OBA and its derivatives, (ii) the side chain of Asp 181 close to the o-carboxylic acid of OBA. The optimal positioning of both of the carboxy groups of OBA is likely to contribute significantly to the potency of this class of compounds.
When comparing with the complexes of PTP1B and Tyr(P) containing peptides (26), several structural differences have been identified in the present study. Thus, in addition to the above movement of Lys 120 , a minor rotation of the Phe 182 side chain is observed (Fig. 5B), most likely due to the shift seen in the aromatic cores between OBA and Tyr(P) (Fig. 5A). Similarly, the different orientation of the Asp 181 side chain might be due partly to the shift in the aromatic core, partly to interaction with the o-carboxylic acid group. This demonstrates that the active site of PTP1B, although fairly rigid, possesses the capacity to accommodate other ligands that have no or limited similarity to the natural substrate. This accommodation process is observed in the surface exposed loop regions, whereas no differences in the PTP loop are observed between the Tyr(P) and OBA complex structures. As indicated above, the potency of OBA for PTP1B (and possibly most other PTPs) appears to arise from (i) OBA's ability to closely mimic the Tyr(P) substrate (phosphate and aromatic ring) and (ii) as well as the formation of a novel set of contacts with residues Asp 181 , Lys 120 , and Tyr 46 .
The general interest in phosphate recognition sites as drug targets (e.g. in serine and tyrosine phosphatases, SH2 domains, PTB domains, and also kinases) points to the importance of identification of novel biologically active phosphate mimetics. Compounds containing one or more carboxylic acids have been particularly useful in the design of PTP inhibitors (6). However, although significant progress has been made in this area, the current study emphasizes the need for correctly positioning the carboxy groups in the active site. It appears that the unique combination of a phosphate mimetic (the oxamic acid part), the o-carboxy group, and an aromatic ring in the OBA structure provides a significant improvement over other known carboxylic acid-based PTP inhibitors. Interestingly, Beaulieu and coworkers (44) recently reported the replacement of phosphate with the non-hydrolyzable oxamic acid group in peptide-based SH2 domain ligands. Although 20 times less potent than phosphate, these findings add to the notion of oxamic acid as a phosphate mimetic.
In conclusion, we have demonstrated that OBA is a (i) novel, (ii) general, (iii) low molecular weight, (iv) classical, competitive, and (v) active site inhibitor of PTPs. Detailed enzyme kinetic analyses in combination with extensive x-ray crystallographic evaluation of OBA and three derivatives have shown that this class of inhibitors, in addition to utilizing many of the interaction points of the natural substrate Tyr(P), also create a unique binding mode. We anticipate that OBA will be an essential tool in future explorations of the structure and function of PTPs. Furthermore, the structural insight provided here should make OBA an ideal starting point in our search for potent and selective PTP inhibitors.