Role of aspartate 27 in the binding of methotrexate to dihydrofolate reductase from Escherichia coli

From a binding module to essential catalytic activity: how nature stumbled on a good thing

Enzymes are complex macromolecules capable of catalyzing a wide variety of chemical reactions with high efficiency. Nonetheless, biological catalysis can be rudimentary. Here, we describe an enzyme that is built from a simple protein fold. This short protein sequence - almost a peptide - belongs to the ancient SH3 family of binding modules. Surprisingly, this binding module catalyzes the specific reduction of dihydrofolate using NADPH as a reducing cofactor, making this a dihydrofolate reductase. Too small to provide all the required binding and catalytic machinery on its own, it homotetramerizes, thus creating a large, central active site environment. Remarkably, none of the active site residues is essential to the catalytic function. Instead, backbone interactions juxtapose the reducing cofactor proximal to the target imine of the folate substrate, and a specific motion of the substrate promotes formation of the transition state. In this feature article, we describe the features that make this small protein a functional enzyme capable of catalyzing a metabolically essential reaction, highlighting the characteristics that make it a model for the evolution of primitive enzymes from binding modules.

Distal Regions Regulate Dihydrofolate Reductase-Ligand Interactions

Protein motions play a fundamental role in enzyme catalysis and ligand binding. The relationship between protein motion and function has been extensively investigated in the model enzyme dihydrofolate reductase (DHFR). DHFR is an essential enzyme that catalyzes the reduction of dihydrofolate to tetrahydrofolate. Numerous experimental and computational methods have been used to probe the motions of DHFR through the catalytic cycle and to investigate the effect of distal mutations on DHFR motions and ligand binding. These experimental investigations have pushed forward the study of protein motions and their role in protein-ligand interactions. The introduction of mutations distal to the active site has been shown to have profound effects on ligand binding, hydride transfer rates and catalytic efficacy and these changes are captured by enzyme kinetics measurements. Distal mutations have been shown to exert their effects through a network of correlated amino acids and these effects have been investigated by NMR, protein dynamics, and analysis of coupled amino acids. The experimental methods and the findings that are reviewed here have broad implications for our understanding of enzyme mechanisms, ligand binding and for the future design and discovery of enzyme inhibitors.

Mutational analysis confirms the presence of distal inhibitor-selectivity determining residues in B. stearothermophilus dihydrofolate reductase

Many antibacterial and antiparasitic drugs work by competitively inhibiting dihydrofolate reductase (DHFR), a vital enzyme in folate metabolism. The interactions between inhibitors and DHFR active site residues are known in many homologs but the contributions from distal residues are less understood. Identifying distal residues that aid in inhibitor binding can improve targeted drug development programs by accounting for distant influences that may be less conserved and subject to frequent resistance causing mutations. Previously, a novel, homology-based, computational approach that mines ligand inhibition data was used to predict residues involved in inhibitor selectivity in the DHFR family. Expectedly, some inhibitor selectivity determining residue positions were predicted to lie in the active site and coincide with experimentally known inhibitor selectivity determining positions. However, other residues that group spatially in clusters distal to the active site have not been previously investigated. In this study, the effect of introducing amino acid substitutions at one of these predicted clusters (His38-Ala39-Ile40) on the inhibitor selectivity profile in Bacillus stearothermophilus dihydrofolate reductase (Bs DHFR) was investigated. Mutations were introduced into these cluster positions to change sidechain chemistry and size. We determined kcat and KM values and measured KD values at equilibrium for two competitive DHFR inhibitors, trimethoprim (TMP) and pyrimethamine (PYR). Mutations in the His38-Ala39-Ile40 cluster significantly impacted inhibitor binding and TMP/PYR selectivity - seven out of nine mutations resulted in tighter binding to PYR when compared to TMP. These data suggest that the His38-Ala39-Ile40 cluster is a distal inhibitor selectivity determining region that favors PYR binding in Bs DHFR and, possibly, throughout the DHFR family.

Crowders Steal Dihydrofolate Reductase Ligands through Quinary Interactions

Dihydrofolate reductase (DHFR) reduces dihydrofolate (DHF) to tetrahydrofolate using NADPH as a cofactor. Due to its role in one carbon metabolism, chromosomal DHFR is the target of the antibacterial drug, trimethoprim. Resistance to trimethoprim has resulted in a type II DHFR that is not structurally related to the chromosomal enzyme target. Because of its metabolic significance, understanding DHFR kinetics and ligand binding behavior in more cell-like conditions, where the total macromolecule concentration can be as great as 300 mg/mL, is important. The progress-curve kinetics and ligand binding properties of the drug target (chromosomal E. coli DHFR) and the drug resistant (R67 DHFR) enzymes were studied in the presence of macromolecular cosolutes. There were varied effects on NADPH oxidation and binding to the two DHFRs, with some cosolutes increasing affinity and others weakening binding. However, DHF binding and reduction in both DHFRs decreased in the presence of all cosolutes. The decreased binding of ligands is mostly attributed to weak associations with the macromolecules, as opposed to crowder effects on the DHFRs. Computer simulations found weak, transient interactions for both ligands with several proteins. The net charge of protein cosolutes correlated with effects on NADP⁺ binding, with near neutral and positively charged proteins having more detrimental effects on binding. For DHF binding, effects correlated more with the size of binding pockets on the protein crowders. These nonspecific interactions between DHFR ligands and proteins predict that the in vivo efficiency of DHFRs may be much lower than expected from their in vitro rates.

Chemical Approaches to Elucidate Enzymatic Profiles of UDP-Glucose: Glycoprotein Glucosyltransferase

In the endoplasmic reticulum (ER), uridine 5'-diphosphate-glucose: glycoprotein glucosyltransferase 1 (UGGT1) recognizes misfolded glycoproteins and transfers a glucose residue to the specific non-reducing end of high-mannose-type glycans. However, precise molecular mechanism by which UGGT1 senses the folding has not been understood clearly. To address this issue, various model substrates for UGGT1 have been prepared using biological approaches. Recently, we introduced chemical approaches using synthetic glycan probes that were designed for studying N-glycan processing in the ER and Golgi apparatus. Our approach can outfit the homogeneous and functionalized glycan probes. In this review, recent results on functional analysis of UGGT1 are summarized.

Hydrophobic Tagged Dihydrofolate Reductase for Creating Misfolded Glycoprotein Mimetics

In the endoplasmic reticulum (ER), nascent glycoproteins that have not acquired the native conformation are either repaired or sorted for degradation by specific quality-control systems composed by various proteins. Among them, UDP-glucose:glycoprotein glucosyltransferase (UGGT) serves as a folding sensor in the ER. However, the molecular mechanism of its recognition remains obscure. This study used pseudo-misfolded glycoproteins, comprising a modified dihydrofolate reductase with artificial pyrene-cysteine moiety on the protein surface (pDHFR) and Man9 GlcNAc2 -methotrexate (M9-MTX). All five M9-MTX/pDHFR complexes, with a pyrene group at different positions, were found to be good substrates of UGGT, irrespective of the site of pyrene modification. These results suggest UGGT's mode of substrate recognition is fuzzy, thus allowing various glycoproteins to be accommodated in the folding cycle.

Exploiting ligand-protein conjugates to monitor ligand-receptor interactions

We introduce three assays for analyzing ligand-receptor interactions based on the specific conjugation of ligands to SNAP-tag fusion proteins. Conjugation of ligands to different SNAP-tag fusions permits the validation of suspected interactions in cell extracts and fixed cells as well as the establishment of high-throughput assays. The different assays allow the analysis of strong and weak interactions. Conversion of ligands into SNAP-tag substrates thus provides access to a powerful toolbox for the analysis of their interactions with proteins.

Evidence for dynamics in proteins as a mechanism for ligand dissociation

Signal transduction, regulatory processes, and pharmaceutical responses are highly dependent upon ligand residence times. Gaining insight into how physical factors influence residence times, or k_off, should enhance our ability to manipulate biological interactions. We report experiments that yield structural insight into k_off for a series of eight 2,4-diaminopyrimidine inhibitors of dihydrofolate reductase that vary by six orders of magnitude in binding affinity. NMR relaxation dispersion experiments revealed a common set of residues near the binding site that undergo a concerted, millisecond-timescale switching event to a previously unidentified conformation. The rate of switching from ground to excited conformations correlates exponentially with K_i and k_off, suggesting that protein dynamics serves as a mechanical initiator of ligand dissociation within this series and potentially for other macromolecule-ligand systems. Although k_conf,forward is faster than k_off, use of the ligand series allowed for connections to be drawn between kinetic events on different timescales.

Identification of endogenous ligands bound to bacterially expressed human and E. coli dihydrofolate reductase by 2D NMR

Dihydrofolate reductase (DHFR) is a well-studied drug target and a paradigm for understanding enzyme catalysis. Preparation of pure DHFR samples, in defined ligand-bound states, is a prerequisite for in vitro studies and drug discovery efforts. We use NMR spectroscopy to monitor ligand content of human and Escherichia coli DHFR (ecDHFR), which bind different co-purifying ligands during expression in bacteria. An alternate purification strategy yields highly pure DHFR complexes, containing only the desired ligands, in the quantities required for structural studies. Interestingly, ecDHFR is bound to endogenous THF while human DHFR is bound to NADP. Consistent with these findings, a designed "humanized" mutant of ecDHFR switches binding specificity in the cell.

Thermodynamics and solvent effects on substrate and cofactor binding in Escherichia coli chromosomal dihydrofolate reductase

Chromosomal dihydrofolate reductase from Escherichia coli catalyzes the reduction of dihydrofolate to tetrahydrofolate using NADPH as a cofactor. The thermodynamics of ligand binding were examined using an isothermal titration calorimetry approach. Using buffers with different heats of ionization, zero to a small, fractional proton release was observed for dihydrofolate binding, while a proton was released upon NADP(+) binding. The role of water in binding was additionally monitored using a number of different osmolytes. Binding of NADP(+) is accompanied by the net release of ∼5-24 water molecules, with a dependence on the identity of the osmolyte. In contrast, binding of dihydrofolate is weakened in the presence of osmolytes, consistent with "water uptake". Different effects are observed depending on the identity of the osmolyte. The net uptake of water upon dihydrofolate binding was previously observed in the nonhomologous R67-encoded dihydrofolate reductase (dfrB or type II enzyme) [Chopra, S., et al. (2008) J. Biol. Chem. 283, 4690-4698]. As R67 dihydrofolate reductase possesses a nonhomologous sequence and forms a tetrameric structure with a single active site pore, the observation of weaker DHF binding in the presence of osmolytes in both enzymes implicates cosolvent effects on free dihydrofolate. Consistent with this analysis, stopped flow experiments find betaine mostly affects DHF binding via changes in k(on), while betaine mostly affects NADPH binding via changes in k(off). Finally, nonadditive enthalpy terms when binary and ternary cofactor binding events are compared suggest the presence of long-lived conformational transitions that are not included in a simple thermodynamic cycle.

X-ray structure of the ternary MTX.NADPH complex of the anthrax dihydrofolate reductase: a pharmacophore for dual-site inhibitor design

For reasons of bioterrorism and drug resistance, it is imperative to identify and develop new molecular points of intervention against anthrax. Dihydrofolate reductase (DHFR) is a highly conserved enzyme and an established target in a number of species for a variety of chemotherapeutic programs. Recently, the crystal structure of Bacillus anthracis DHFR (baDHFR) in complex with methotrexate (MTX) was determined and, based on the structure, proposals were made for drug design strategies directed against the substrate-binding site. However, little is gleaned about the binding site for NADPH, the cofactor responsible for hydride transfer in the catalytic mechanism. In the present study, X-ray crystallography at 100 K was used to determine the structure of baDHFR in complex with MTX and NADPH. Although the NADPH binding mode is nearly identical to that seen in other DHFR ternary complex structures, the adenine moiety adopts an off-plane tilt of nearly 90 degrees and this orientation is stabilized by hydrogen bonds to functionally conserved Arg residues. A comparison of the binding site, focusing on this region, between baDHFR and the human enzyme is discussed, with an aim at designing species-selective therapeutics. Indeed, the ternary model, refined to 2.3 A resolution, provides an accurate template for testing the feasibility of identifying dual-site inhibitors, compounds that target both the substrate and cofactor-binding site. With the ternary model in hand, using in silico methods, several compounds were identified which could potentially form key bonding contacts in the substrate and cofactor-binding sites. Ultimately, two structurally distinct compounds were verified that inhibit baDHFR at low microM concentrations. The apparent Kd for one of these, (2-(3-(2-(hydroxyimino)-2-(pyridine-4-yl)-6,7-dimethylquinoxalin-2-yl)-1-(pyridine-4-yl)ethanone oxime), was measured by fluorescence spectroscopy to be 5.3 microM.

Synthesis and biological evaluation of novel 2,4-diaminoquinazoline derivatives as SMN2 promoter activators for the potential treatment of spinal muscular atrophy

Proximal spinal muscular atrophy (SMA) is an autosomal recessive disorder characterized by death of motor neurons in the spinal cord that is caused by deletion and/or mutation of the survival motor neuron gene ( SMN1). Adjacent to SMN1 are a variable number of copies of the SMN2 gene. The two genes essentially differ by a single nucleotide, which causes the majority of the RNA transcripts from SMN2 to lack exon 7. Although both SMN1 and SMN2 encode the same Smn protein amino acid sequence, the loss of SMN1 and incorrect splicing of SMN2 have the consequence that Smn protein levels are insufficient for the survival of motor neurons. The therapeutic goal of our medicinal chemistry effort was to identify small-molecule activators of the SMN2 promoter that, by up-regulating gene transcription, would produce greater quantities of full-length Smn protein. Our initial medicinal chemistry effort explored a series of C5 substituted benzyl ether based 2,4-diaminoquinazoline derivatives that were found to be potent activators of the SMN2 promoter; however, inhibition of DHFR was shown to be an off-target activity that was linked to ATP depletion. We used a structure-guided approach to overcome DHFR inhibition while retaining SMN2 promoter activation. A lead compound 11a was identified as having high potency (EC50 = 4 nM) and 2.3-fold induction of the SMN2 promoter. Compound 11a possessed desirable pharmaceutical properties, including excellent brain exposure and long brain half-life following oral dosing to mice. The piperidine compound 11a up-regulated expression of the mouse SMN gene in NSC-34 cells, a mouse motor neuron hybrid cell line. In type 1 SMA patient fibroblasts, compound 11a induced Smn in a dose-dependent manner when analyzed by immunoblotting and increased the number of intranuclear particles called gems. The compound restored gems numbers in type I SMA patient fibroblasts to levels near unaffected genetic carriers of SMA.

Substrate and inhibitor specificity of Mycobacterium avium dihydrofolate reductase

Dihydrofolate reductase (EC 1.5.1.3) is a key enzyme in the folate biosynthetic pathway. Information regarding key residues in the dihydrofolate-binding site of Mycobacterium avium dihydrofolate reductase is lacking. On the basis of previous information, Asp31 and Leu32 were selected as residues that are potentially important in interactions with dihydrofolate and antifolates (e.g. trimethoprim), respectively. Asp31 and Leu32 were modified by site-directed mutagenesis, giving the mutants D31A, D31E, D31Q, D31N and D31L, and L32A, L32F and L32D. Mutated proteins were expressed in Escherichia coli BL21(DE3)pLysS and purified using His-Bind resin; functionality was assessed in comparison with the recombinant wild type by a standard enzyme assay, and growth complementation and kinetic parameters were evaluated. All Asp31 substitutions affected enzyme function; D31E, D31Q and D31N reduced activity by 80-90%, and D31A and D31L by > 90%. All D31 mutants had modified kinetics, ranging from three-fold (D31N) to 283-fold (D31L) increases in K(m) for dihydrofolate, and 12-fold (D31N) to 223 077-fold (D31L) decreases in k(cat)/K(m). Of the Leu32 substitutions, only L32D caused reduced enzyme activity (67%) and kinetic differences from the wild type (seven-fold increase in K(m); 21-fold decrease in k(cat)/K(m)). Only minor variations in the K(m) for NADPH were observed for all substitutions. Whereas the L32F mutant retained similar trimethoprim affinity as the wild type, the L32A mutation resulted in a 12-fold decrease in affinity and the L32D mutation resulted in a seven-fold increase in affinity for trimethoprim. These findings support the hypotheses that Asp31 plays a functional role in binding of the substrate and Leu32 plays a functional role in binding of trimethoprim.

Exploration of oligosaccharide-protein interactions in glycoprotein quality control by synthetic approaches

High-mannose-type oligosaccharides, which are cotranslationally introduced to nascent polypeptides, play important roles in glycoprotein quality control. This process is highly complex, involving a number of lectins, chaperones, and glycan-processing enzymes. For example, calnexin and calreticulin (CRT) are molecular chaperones that recognize monoglucosylated forms of high-mannose-type glycans. UDP-glucose : glycoprotein glucosyltransferase (UGGT) only glucosylates high-mannose-type glycans attached to partially folded proteins. Fbs1 is a component of ubiquitin ligase that recognizes sugar chains. Although recent studies have clarified the properties of these proteins, most of them used oligosaccharides derived from natural sources, which contain structural heterogeneity. In order to gain a more precise understanding, we started our program to comprehensively synthesize high-mannose-type glycans associated with a protein quality control system. Additionally, investigation of artificial glycoproteins led us to the discovery of the first nonpeptidic substrate of UGGT. These synthetic oligosaccharide probes have allowed us to conduct quantitative evaluations of the activity and specificity of CRT, Fbs1, and UGGT.

Substrate Specificity Analysis of Endoplasmic Reticulum Glucosidase II Using Synthetic High Mannose-type Glycans

Glucosidase II (Glc'ase II) is a glycan-processing enzyme that trims two alpha1,3-linked Glc residues in succession from the glycoprotein oligosaccharide Glc2Man9GlcNAc2 to give Glc1Man9GlcNAc2 and Man9GlcNAc2 in the endoplasmic reticulum (ER). Monoglucosylated glycans, such as Glc1-Man9GlcNAc2, generated by this process play a key role in glycoprotein quality control in the ER, because they are primary ligands for the lectin chaperones calnexin (CNX) and calreticulin (CRT). A precise analysis of the substrate specificity of Glc'ase II is expected to further our understanding of the molecular basis to glycoprotein quality control, because Glc'ase II potentially competes with CNX/CRT for the same glycans, Glc1Man7-9GlcNAc2. In this study, a quantitative analysis of the specificity of Glc'ase II using a series of structurally defined synthetic glycans was carried out. In the presence of CRT, Glc'ase II-mediated trimming from Glc2Man9GlcNAc2 stopped at Glc1Man9GlcNAc2, supporting the notion that the glycan structure delivered to the CNX/CRT cycle is Glc1Man9GlcNAc2. Unexpectedly, our experiments showed that Glc1Man8(B)GlcNAc2 had nearly the same reactivity as Glc1Man9GlcNAc2, which was markedly greater than that of its positional isomer Glc1Man8(C)GlcNAc2. An analysis with glycoprotein-like probes revealed the stepwise formation of Glc1Man9GlcNAc2 and Man9GlcNAc2 from Glc2Man9GlcNAc2, even in the presence of CRT. It was also shown that Glc1Man8(B)GlcNAc2 had even greater reactivity than Glc1Man9GlcNAc2 at the glycoprotein level. Moreover, inhibitory activities by nonglucosylated glycans suggested that Glc'ase II recognized the C arm (Manalpha1, 2Manalpha1, 6Man-) of high mannose-type glycans.

High-mannose-type glycan modifications of dihydrofolate reductase using glycan–methotrexate conjugates

Various high-mannose-type glycan modifications of dihydrofolate reductase (DHFR) were achieved by ligand-based approach using glycan-methotrexate (MTX) conjugates as tight binding glycan bearing ligands for DHFR. The resulting glycan-MTX conjugates and the corresponding artificial glycoproteins could be useful as oligosaccharide- and glycoprotein-probes to perform quantitative analysis of glycan recognizing protein such as lectins, glycosyltransferases or glycosidases. Moreover, artificial glycoproteins having two different high-mannose-type glycans were developed for the first time by a combination of two different types of glycan modification strategies.

Dissecting virulence pathways of Mycobacterium tuberculosis through protein-protein association

The sudden increase in information derived from the completed Mycobacterium tuberculosis (Mtb) genome sequences has revealed the need for approaches capable of converting raw genome sequence data into functional information. To date, an experimental system for studying protein-protein association in mycobacteria is not available. We have developed a simple system, termed mycobacterial protein fragment complementation (M-PFC), that is based upon the functional reconstitution of two small murine dihydrofolate reductase domains independently fused to two interacting proteins. Using M-PFC, we have successfully demonstrated dimerization of yeast GCN4, interaction between Mtb KdpD and KdpE, and association between Esat-6 and Cfp-10. We established the association between the sensor kinase, DevS, and response regulator, DevR, thereby demonstrating the potential of M-PFC to study protein associations in the mycobacterial membrane. To validate our system, we screened an Mtb library for proteins that associate with the secreted antigen Cfp-10 and consistently identified Esat-6 in our screens. Additional proteins that specifically associate with Cfp-10 include Rv0686 and Rv2151c (FtsQ), a component and substrate, respectively, of the evolutionary conserved signal recognition pathway; and Rv3596c (ClpC1), an AAA-ATPase chaperone involved in protein translocation and quality control. Our results provide empirical evidence that directly links the Mtb specialized secretion pathway with the evolutionary conserved signal recognition and SecA/SecYEG pathways, suggesting they share secretory components. We anticipate that M-PFC will be a major contributor to the systematic assembly of mycobacterial protein interaction maps that will lead to the development of better strategies for the control of tuberculosis.

Identification of amino acids required for the functional up-regulation of human dihydrofolate reductase protein in response to antifolate Treatment

Human dihydrofolate reductase (DHFR) protein levels rapidly increase upon exposure to methotrexate, a potent inhibitor of this enzyme. A model to explain this increase proposes that DHFR inhibits its own translation by binding to its cognate mRNA and that methotrexate disrupts the DHFR protein-mRNA complex allowing its translation to resume. In the present study, Chinese hamster ovary cells lacking DHFR were transfected with wild type and mutants of human DHFR to identify amino acids that are essential for increases in DHFR in response to methotrexate. Glu-30, Leu-22, and Ser-118 were involved in the up-regulation of DHFR protein levels by methotrexate and certain other antifolates. Cells transfected with E30A, L22R, and S118A mutants that did not respond to methotrexate up-regulation had higher basal levels of DHFR, consistent with the model, i.e. lack of feedback regulation of these enzymes. Although cells containing the S118A mutant enzyme had higher levels of DHFR and had catalytic activity similar to that of wild type DHFR, they had the same sensitivity to the cytotoxicity of methotrexate, as were cells with wild type DHFR. This finding provides evidence that the adaptive up-regulation of DHFR by methotrexate contributes to the decreased sensitivity to this drug. Based on these observations, a new model is proposed whereby DHFR exists in two conformations, one bound to DHFR mRNA and the other bound to NADPH. The mutants that are not up-regulated by methotrexate are unable to bind their cognate mRNA.

Chemical dimerizers and three-hybrid systems: scanning the proteome for targets of organic small molecules

The integration of technological advances in areas as diverse as chemical biology, proteomics, genomics, automation, and bioinformatics has led to the emergence of novel screening paradigms for analyzing the molecular basis of drug action. This review summarizes recent advances in three-hybrid technologies and their application to the characterization of small molecule-protein interactions and proteome-wide identification of drug receptors.

Tight binding ligand approach to oligosaccharide-grafted protein

A novel type of artificial glycoprotein was developed, by using dihydrofolate reductase (DHFR) and methotrexate (MTX) as a protein-ligand pair. Various oligosaccharides linked to MTX were shown to bind tightly with DHFR and afforded oligosaccharide-grafted protein, which could be isolated easily by lectin beads.

Chemically induced dimerization of dihydrofolate reductase by a homobifunctional dimer of methotrexate

BACKGROUND

Chemically induced dimerization (CID) can be used to manipulate cellular regulatory pathways from signal transduction to transcription, and to create model systems for study of the specific interactions between proteins and small-molecule chemical ligands. However, few CID systems are currently available. The properties of, and interactions between, Escherichia coli dihydrofolate reductase (DHFR) and the ligand methotrexate (MTX) meet many of the desired criteria for the development of a new CID system.

RESULTS

BisMTX, a homobifunctional version of MTX, was synthesized and tested for its ability to induce dimerization of DHFR. Gel-filtration analysis of purified DHFR confirmed that, in vitro, the protein was a monomer in the absence of dimerizer drug; in the presence of bisMTX, a complex of twice the monomeric molecular weight was observed. Furthermore, the off-rate was found to be 0.0002 s(-1), approximately 100 times slower than that reported for DHFR-MTX. Interestingly, the addition of excess bisMTX did not result in formation of the binary complex (1 protein:1 dimerizer) over the ternary complex (2 proteins:1 dimerizer), which suggests cooperative binding interactions (affinity modulation) between the two DHFR molecules in the bisMTX:DHFR(2) ternary complex.

CONCLUSIONS

The combination of DHFR and bisMTX provides a new CID system with properties that could be useful for applications in vivo. Formation of the bisMTX:DHFR(2) ternary complex in vitro is promoted over a wide range of dimerizer concentrations, consistent with the idea that formation of the ternary complex recruits energetically favorable interactions between the DHFR monomers in the complex.

Dihydrofolate reductase from Kaposi's sarcoma-associated herpesvirus

Kaposi's sarcoma-associated herpesvirus (KSHV) is the first human virus known to encode dihydrofolate reductase (DHFR), an enzyme required for nucleotide and methionine biosynthesis. We have studied the purified KSHV-DHFR enzyme in vitro and analyzed its expression in cultured B-cell lines derived from primary effusion lymphoma (PEL), an AIDS-associated malignancy. The amino acid sequence of KSHV-DHFR is most similar to human DHFR (hDHFR), but the viral enzyme contains an additional 23 amino acids at the carboxyl-terminus. The viral DHFR, overexpressed and purified from E. coli, was catalytically active in vitro. The K(m) of KSHV-DHFR for dihydrofolate (FH(2)) was 2.4 microM, which is significantly higher than the K(m) of recombinant hDHFR (rhDHFR) for FH(2) (390 nM). K(m) values for NADPH were similar for the two enzymes, about 1 microM. KSHV-DHFR was inhibited by folate antagonists such as methotrexate (K(i): 200 pM), aminopterin (K(i): 610 pM), pyrimethamine (K(i): 29 nM), trimethoprim (K(i): 2.3 microM), and piritrexim (K(i): 3.9 nM). In all cases, K(i) values for these folate antagonists were higher for KSHV-DHFR than for rhDHFR. The viral enzyme was expressed at levels two- to tenfold higher than hDHFR in PEL cell lines as an early lytic cycle gene. KSHV-DHFR mRNA and protein appeared from 6 to 24 h after chemical induction of the KSHV lytic cycle. Epitope-tagged KSHV-DHFR and rhDHFR both localized to the nucleus of transfected cells, while other KSHV nucleotide metabolism genes localized to the cytoplasm. DHFR activity was not essential for viral replication in cultured PEL cells. Since hDHFR was not detectable in peripheral blood mononuclear cells (PBMCs), KSHV-DHFR may function to provide increased DHFR activity in vivo in infected cells that have little or none of their own enzyme.

Comparison of the protection of cells from antifolates by transduced human dihydrofolate reductase mutants

Retroviral transduction of antifolate-resistant variants of human dihydrofolate reductase (hDHFR) into cells can increase their resistance to the cytotoxic effects of these drugs. We evaluated the ability of wild-type hDHFR and 20 mutant enzymes (13 with single-amino acid substitutions, 7 with two substitutions) to prevent growth inhibition in antifolate-treated CCRF-CEM cells. The wild-type enzyme and all of the variants significantly protected transduced cells from trimetrexate (TMTX)-induced growth inhibition. However, only half of the variants conferred more protection than does the wild-type enzyme. For the variants tested, the observed protective effect was higher for TMTX than for methotrexate (< or =7.5-fold increased resistance), piritrexim (< or =16-fold), and edatrexate (negligible). Transduction of the variants L22Y-F31S and L22Y-F31R led to the greatest protection against TMTX (approximately 200-fold). Protection from loss of cell viability was similar to protection from growth inhibition. The protection associated with a particular mutant hDHFR did not result from the level of expression: Efficient protection resulted from low affinity of the variant for antifolates, reasonable catalytic activity, and good thermal stability. Clones isolated from a polyclonal population of transduced cells varied by as much as 30-fold in their resistance to TMTX, the resistance differences depending on hDHFR expression levels.

Consideration of the pH-dependent inhibition of dihydrofolate reductase by methotrexate

Poisson-Boltzmann calculations were used to determine the pKa of protein functional groups in the unliganded dihydrofolate reductase enzyme, and the pKa of protein and ligand groups in methotrexate-enzyme complexes. The results reported here are in conflict with two fundamental tenets of dihydrofolate reductase inhibition by methotrexate: (1) Asp27 is not expected to be protonated near pH 6.5 in the apoenzyme as previously proposed based on fitting of empirical equations to binding data, and (2) the calculated pKa for the pteridine N1 of the inhibitor while bound to the protein is significantly lower than that estimated for this group from interpretation of NMR data (>10). In fact, the electrostatic calculations and complementary quantum chemical calculations indicate that Asp27 is likely protonated when methotrexate is bound, resulting in a neutral dipole-dipole interaction rather than a salt-bridge between the enzyme and the inhibitor. Reasons for this discrepancy with the experimental data are discussed. Furthermore, His45 and Glu17 in the Escherichia coli enzyme are proposed to be in part responsible for the pH dependence of the conformational degeneracy in the inhibitor-enzyme complex.

Slow-binding inhibition: the general case

Two basic kinetic mechanisms have been described to account for the slow-binding inhibition of enzyme-catalyzed reactions. One mechanism involves the slow interaction of an inhibitor with enzyme (Mechanism A), while the other involves the rapid formation of an enzyme-inhibitor complex that undergoes a slow isomerization reaction (Mechanism B). But the initial interaction of enzyme and inhibitor may not necessarily be fast so that the free enzyme and the two forms of enzyme inhibitor complex are in steady-state equilibrium. This assumption would give rise to a more general form of Mechanism B. The present study has been concerned with attempts to determine whether it might be possible to distinguish between the three possible inhibition mechanisms by steady-state kinetic techniques. The approach to the investigation has been to derive theoretical data for the most general mechanism by using three different ratios for the two rate constants that determine which mechanism applies. The progress curve data were then fitted to the rate equations that describe the other two mechanisms. The results draw attention to the difficulties of deducing that experimental data conform to the most general mechanism. They also show how the values for the kinetic parameters, as determined from fits of the data to the equations that describe Mechanisms A and B, can be considerably in error.

Expression and characterization of the Trypanosoma cruzi dihydrofolate reductase domain

We have cloned and expressed in Escherichia coli a 702-base pair gene coding for the dihydrofolate reductase (DHFR) domain of the bifunctional dihydrofolate reductase-thymidylate synthase (DHFR-TS) from Trypanosoma cruzi. The DHFR domain was purified to homogeneity by methotrexate-Sepharose chromatography followed by an anion-exchange chromatography step in a mono Q column, and displayed a single 27-kDa band on SDS-PAGE. Gel filtration showed that the catalytic domain was expressed as a monomer. Kinetic parameters were similar to those reported for the wild-type bifunctional enzyme with Km values of 0.75 microM for dihydrofolate and 16 microM for NADPH and a kcat value of 16.5 s-1. T. cruzi DHFR is poorly inhibited by trimethoprim and pyrimethamine and the inhibition constants were always lower for the bifunctional enzyme. The binding of methotrexate was characteristic of a class of inhibitors that form an initial complex which isomerizes slowly to a tighter complex and are referred to as 'slow, tight-binding' inhibitors. While the slow-binding step of inhibition was apparently unaffected in the individually expressed DHFR domain, the overall inhibition constant was two-fold higher as a consequence of the superior inhibition constant value obtained for the initial inhibitory complex.

Methotrexate-resistant variants of human dihydrofolate reductase with substitutions of leucine 22. Kinetics, crystallography, and potential as selectable markers

Although substitution of tyrosine, phenylalanine, tryptophan, or arginine for leucine 22 in human dihydrofolate reductase greatly slows hydride transfer, there is little loss in overall activity (kcat) at pH 7.65 (except for the arginine 22 variant), but Km for dihydrofolate and NADPH are increased significantly. The greatest effect, decreased binding of methotrexate to the enzyme-NADPH complex by 740- to 28,000-fold due to a large increase in the rate of methotrexate dissociation, makes these variants suitable to act as selectable markers. Affinities for four other inhibitors are also greatly decreased. Binding of methotrexate to apoenzyme is decreased much less (decreases as much as 120-fold), binding of tetrahydrofolate is decreased as much as 23-fold, and binding of dihydrofolate is decreased little or increased. Crystal structures of ternary complexes of three of the variants show that the mutations cause little perturbation of the protein backbone, of side chains of other active site residues, or of bound inhibitor. The largest structural deviations occur in the ternary complex of the arginine variant at residues 21-27 and in the orientation of the methotrexate. Tyrosine 22 and arginine 22 relieve short contacts to methotrexate and NADPH by occupying low probability conformations, but this is unnecessary for phenylalanine 22 in the piritrexim complex.

A kinetic mechanism of the allosteric control of enzyme-coenzyme binding: glutamate dehydrogenase-NADPH-phosphate-acetate-hydrogen ion interactions

We have previously characterized the thermodynamic relationships which govern the dissociation of NADPH from bovine liver glutamate dehydrogenase and the allosteric control of that mechanically and physiologically important process by a variety of effectors. We have found that the cooperative occupancy of a specific anion binding, while the occupancy of a second allosteric acetate binding site disrupts that anion binding site and opposes those effects (Singh & Fisher, 1994). We report here the results of transient-state studies on the kinetics of the various processes involved in this complex equilibrium. We find that the only intrinsically slow steps are those of NADPH binding and dissociation, that the complex kinetic behavior of the overall system is due solely to very rapid equilibrium binding processes involving phosphate, acetate, and hydrogen ions, and that these ions exert their various effects on the kinetics of the binding process by altering the equilibrium concentrations of the two kinetically significant reactive species, E and E-NADPH. The slow intrinsic rates of NADPH association and dissociation are ascribed to a ligand-induced conformational change involving a major alteration in the degree of closure of the enzyme's active-site cleft.

Ab initio studies of 2,4-diamino triazine and its complexes with ligands: a model for inhibitor-active site interactions of dihydrofolate reductase

The protonation energies of 2,4-diamino triazine, an inhibitor of the therapeutic target dihydrofolate reductase, has been calculated using ab initio (Hartree-Fock) calculations. It is found that N1 (see Fig. 1) exhibits the highest proton affinity (261.6 kcal/mol) by comparison with other inhibitor protonation sites. The energies of binding of the formate ion and formamide (as models for the amino acid residues in the active site of dihydrofolate reductase) to neutral and protonated 2,4-diamino triazine are also obtained. The highest binding energies are featured by the complex formed from a formate attached to the N4 and N1 protonated forms of the triazine. However, as N4 has a comparatively low proton affinity (195.0 kcal/mol), it is unlikely that an interaction of this nature would prevail. On the other hand, the formate-protonated N1 interaction is similar to the structures identified by X-ray crystallography of enzyme-triazine complexes.

Collapse of parallel folding channels in dihydrofolate reductase from Escherichia coli by site-directed mutagenesis

The rate-limiting steps in the folding of dihydrofolate reductase from Escherichia coli have been shown to involve the conversion of a set of four intermediates to a corresponding set of native conformers via four parallel channels [Jennings et al. (1993) Biochemistry 32, 3783-3789]. Fluorescence and absorbance studies of the unfolding and refolding of the C85S/C152E double mutant at various final urea concentrations reveal two slow folding reactions, two fewer than observed in the wild-type protein. Refolding in the presence of substoichiometric levels of the inhibitor methotrexate shows that the two remaining slow reactions correspond to two parallel channels which lead to a pair of native conformers capable of binding the inhibitor. A combination of stopped-flow circular dichroism and cofactor binding studies confirms that the four parallel channels observed in the wild-type protein have collapsed into two channels in the mutant. Kinetic and equilibrium studies of the single cysteine mutants suggest that replacements of Cysteine-85 which perturb the hydrophobic core containing this side chain are responsible for the simplification of the kinetic mechanism. These results demonstrate that at least two of the parallel folding channels in dihydrofolate reductase arise when tertiary structure develops and are not dependent upon cis/trans isomerization at prolyl peptide bonds.

A strong protein unfolding activity is associated with the binding of precursor chloroplast proteins to chloroplast envelopes

Protein conformational changes related to transport into chloroplasts have been studied. Two chimaeric proteins carrying the transit peptide of either ferredoxin or plastocyanin linked to the mouse cytosolic enzyme dihydrofolate reductase (EC 1.5.1.3.) were employed. In contrast to observations in mitochondria, we found in chloroplasts that transport of a purified ferredoxin-dihydrofolate reductase fusion protein is not blocked by the presence of methotrexate, a folate analogue that stabilizes the structural conformation of dihydrofolate reductase. It is shown that transport competence of this protein in the presence of methotrexate is not a consequence of alteration of the folding characteristics or methotrexate binding properties of dihydrofolate reductase by fusion to the ferredoxin transit peptide. Binding of dihydrofolate reductase fusion proteins to chloroplast envelopes is not inhibited by low temperature and it is only partially diminished by methotrexate. It is demonstrated that the dihydrofolate reductase fusion proteins unfold, despite the presence of methotrexate, on binding to the chloroplast envelopes. We propose the existence of a strong protein unfolding activity associated to the chloroplast envelopes.

Mobility of the spin-labeled side chains of some novel antifolate inhibitors in their complexes with dihydrofolate reductase

Four spin-labeled inhibitors of dihydrofolate reductase (DHFR) have been synthesized, each of which has the 2,2,6,6-tetramethyl-1-piperidinyloxy (TEMPO) reporting group at a different distance from the 2,4-diaminopyrimidine moiety by which the inhibitors are anchored and oriented in the active site. Inhibitors in which the TEMPO group is attached by a short side chain are weakly bound to DHFR from bacteria (Streptococcus faecium and Lactobacillus casei), to the bovine enzyme and to recombinant human DHFR. However, binding is sufficiently tight, especially in the ternary complexes with NADPH, for recording of the EPR spectra of the bound ligands. The spectra indicate that when these inhibitors are bound to the enzyme the TEMPO group is highly immobilized with correlation time, tau c, 4-20ns. Inhibitors that have the reporter group attached to the glutamate moiety of methotrexate bind to all four DHFRs more tightly than the inhibitors with shorter side chains by factors of up to 10(6). However, in most complexes formed by the inhibitors with longer side chains immobilization of the TEMPO group is slight (tau c 0.2-4 ns). These results are in general agreement with predictions from X-ray crystallographic results including thermal factors but there are some unanticipated differences between some results for bacterial and eukaryotic enzymes. Three of the splin-labeled inhibitors would provide good probes for distance measurements in and around the active site of mammalian DHFR.

Evidence for two interconverting protein isomers in the methotrexate complex of dihydrofolate reductase from Escherichia coli

Two-dimensional 1H NMR methods and a knowledge of the X-ray crystal structure have been used to make resonance assignments for the amino acid side chains of dihydrofolate reductase from Escherichia coli complexed with methotrexate. The H7 proton on the pteridine ring of methotrexate was found to have NOEs to the methyl protons of Leu-28 which were assigned by using the L28F mutant. These NOEs indicated that the orientation of the methotrexate pteridine ring is similar in both solution and crystal structures. During the initial assignment process, it became evident that many of the resonances in this complex, unlike those of the folate complex, are severely broadened or doubled. The observation of two distinct sets of resonances in a ratio of approximately 2:1 was attributed to the presence of two protein isomers. At 303 K, NOESY spectra with mixing times of 100 ms did not show interconversion between these isomers. However, exchange cross-peaks were observed in a 700-ms NOESY spectrum at 323 K which demonstrated that these isomers are interconverting slowly on the NMR time scale. Many of the side chains with clearly doubled resonances were located in the beta-sheet and the active site. Preliminary studies on the apoprotein also revealed doubled resonances in the absence of the inhibitor, indicating the existence of the protein isomers prior to methotrexate binding. In contrast to the methotrexate complex, the binary complex with folate and the ternary MTX-NADPH-DHFR complex presented a single enzyme form. These results are proposed to reflect the ability of folate and NADPH to bind predominantly to one protein isomer.

Crystal structure of unliganded Escherichia coli dihydrofolate reductase. Ligand-induced conformational changes and cooperativity in binding

The crystal structure of unliganded dihydrofolate reductase (DHFR) from Escherichia coli has been solved and refined to an R factor of 19% at 2.3-A resolution in a crystal form that is nonisomorphous with each of the previously reported E. coli DHFR crystal structures [Bolin, J. T., Filman, D. J., Matthews, D. A., Hamlin, B. C., & Kraut, J. (1982) J. Biol. Chem. 257, 13650-13662; Bystroff, C., Oatley, S. J., & Kraut, J. (1990) Biochemistry 29, 3263-3277]. Significant conformational changes occur between the apoenzyme and each of the complexes: the NADP+ holoenzyme, the folate-NADP+ ternary complex, and the methotrexate (MTX) binary complex. The changes are small, with the largest about 3 A and most of them less than 1 A. For simplicity a two-domain description is adopted in which one domain contains the NADP+ 2'-phosphate binding site and the binding sites for the rest of the coenzyme and for the substrate lie between the two domains. Binding of either NADP+ or MTX induces a closing of the PABG-binding cleft and realignment of alpha-helices C and F which bind the pyrophosphate of the coenzyme. Formation of the ternary complex from the holoenzyme does not involve further relative domain shifts but does involve a shift of alpha-helix B and a floppy loop (the Met-20 loop) that precedes alpha B. These observations suggest a mechanism for cooperativity in binding between substrate and coenzyme wherein the greatest degree of cooperativity is expressed in the transition-state complex. We explore the idea that the MTX binary complex in some ways resembles the transition-state complex.

Role of the conserved active site residue tryptophan-24 of human dihydrofolate reductase as revealed by mutagenesis

The active sites of all bacterial and vertebrate dihydrofolate reductases that have been examined have a tryptophan residue near the binding sites for NADPH and dihydrofolate. In cases where the three-dimensional structure has been determined by X-ray crystallography, this conserved tryptophan residue makes hydrophobic and van der Waals interactions with the nicotinamide moiety of bound NADPH, and its indole nitrogen interacts with the C4 oxygen of bound folate through a bridge provided by a bound water molecule. We have addressed the question of why even the very conservative replacement of this tryptophan by phenylalanine does not seem to occur naturally. Human dihydrofolate reductase with this replacement of tryptophan by phenylalanine has increased rate constants for dissociation of substrates and products and a considerably decreased rate of hydride transfer. These cause some changes in steady-state kinetic behavior, including substantial increases in Michaelis constants for NADPH and dihydrofolate, but there is also a 3-fold increase in kcat. The branched mechanistic pathway for this enzyme has been completely defined and is sufficiently different from that of wild-type enzyme to cause changes in some transient-state kinetics. The most critical changes resulting from the amino acid substitution appear to be a 50% decrease in stability and a decrease in efficiency from 69% to 21% under intracellular conditions.

Crystal structures of recombinant human dihydrofolate reductase complexed with folate and 5-deazafolate

The 2.3-A crystal structure of recombinant human dihydrofolate reductase (EC 1.5.1.3, DHFR) has been solved as a binary complex with folate (a poor substrate at neutral pH) and also as a binary complex with an inhibitor, 5-deazafolate. The inhibitor appears to be protonated at N8 on binding, whereas folate is not. Rotation of the peptide plane joining I7 and V8 from its position in the folate complex permits hydrogen bonding of 5-deazafolate's protonated N8 to the backbone carbonyl of I7, thus contributing to the enzyme's greater affinity for 5-deazafolate than for folate. In this respect it is likely that bound 5-deazafolate furnishes a model for 7,8-dihydrofolate binding and, in addition, resembles the transition state for folate reduction. A hypothetical transition-state model for folate reduction, generated by superposition of the DHFR binary complexes human.5-deazafolate and chicken liver.NADPH, reveals a 1-A overlap of the binding sites for folate's pteridine ring and the dihydronicotinamide ring of NADPH. It is proposed that this binding-site overlap accelerates the reduction of both folate and 7,8-dihydrofolate by simultaneously binding substrate and cofactor with a sub van der Waals separation that is optimal for hydride transfer.

A second-site mutation at phenylalanine-137 that increases catalytic efficiency in the mutant aspartate-27----serine Escherichia coli dihydrofolate reductase

The adaptability of Escherichia coli dihydrofolate reductase (DHFR) is being explored by identifying second-site mutations that can partially suppress the deleterious effect associated with removal of the active-site proton donor aspartic acid-27. The Asp27----serine mutant DHFR (D27S) was previously characterized and the catalytic activity found to be greatly decreased at pH 7.0 [Howell et al. (1986) Science 231, 1123-1128]. Using resistance to trimethoprim (a DHFR inhibitor) in a genetic selection procedure, we have isolated a double-mutant DHFR gene containing Asp27----Ser and Phe137----Ser mutations (D27S+F137S). The presence of the F137S mutation increases kcat approximately 3-fold and decreases Km(DHF) approximately 2-fold over D27S DHFR values. The overall effect on kcat/Km(DHF) is a 7-fold increase. The D27S+F137S double-mutant DHFR is still 500-fold less active than wild-type DHFR at pH 7. Surprisingly, Phe137 is approximately 15 A from residue 27 in the active site and is part of a beta-bulge. We propose the F137S mutation likely causes its catalytic effect by slightly altering the conformation of D27S DHFR. This supposition is supported by the observation that the F137S mutation does not have the same kinetic effect when introduced into the wild-type and D27S DHFRs, by the altered distribution of two conformers of free enzyme [see Dunn et al. (1990)] and by a preliminary difference Fourier map comparing the D27S and D27S+F137S DHFR crystal structures.