Dihydrofolate reductase from Bacillus subtilis and its artificial derivatives: expression, purification, and characterization.

The Bacillus subtilis dihydrofolate reductase (DHFR) gene was expressed in Escherichia coli. The gene product was purified to homogeneity by Butyl-Toyopearl, Toyopearl HW55, and DEAE-Toyopearl column chromatographies, and its molecular properties were compared to those of E. coli DHFR. The specific enzyme activity of the B. subtilis DHFR was 240 units/mg under the standard assay conditions, being about four times higher than that of the E. coli DHFR. Km for coenzyme NADPH was 20.7 microM, a value about three times larger than that of E. coli, whereas Km (1.5 microM) for the substrate, dihydrofolate, was similar to that of E. coli DHFR. This seems to reflect the low homology of the amino acid sequence in residues 61-88 of the two DHFRs where one of the NADPH binding sites is located [Bystrof, C. & Kraut, J. (1991) Biochemistry 30, 2227-2239]. Similar to the E. coli DHFR [Iwakura, M. et al. (1992) J. Biochem. 111, 37-45], the extension of amino acid sequences at the C-terminal end of the B. subtilis DHFR could be attained without loss of the enzyme function or decrease of the protein yield. Thus, the DHFR is useful as a carrier protein for expressing small polypeptides, such as leucine enkephalin, bradykinin, and somatostatin.     

Purification and characterization of human-derived Pneumocystis jirovecii dihydrofolate reductase expressed in Sf21 insect cells and in Escherichia coli

Pneumonia caused by Pneumocystis jirovecii is still a major opportunistic infection among patients with AIDS. This opportunitistic pathogen is susceptible to therapy with inhibitors of the enzyme dihydrofolate reductase (DHFR) that target cell growth. Recent studies have shown that recombinant human-derived Pneumocystis DHFR (pDHFR) differs from rat-derived pDHFR by 38% in amino acid sequence. However, characterization of drug susceptibility, kinetics, and the three-dimensional structure of human-derived pDHFR has been hampered by the limited availability of purified material. The present study was undertaken to develop procedures to prepare sufficient enzyme for structure/function studies. Protein yield was limited when human-derived pDHFR was expressed in Escherichia coli using a pET28a(+) vector with an N-terminal His-tag for the 25 kDa protein. Therefore, the protein was expressed in Sf21 insect cells by baculovirus infection. The soluble enzyme was purified from cell lysates via Ni-chelated chromatographic columns, yielding about 5.1 mg of human-derived pDHFR fusion protein per liter of Sf21 culture. The purified protein had the expected mass as determined from Western blots with antibody for the N-terminal His-tag. This His-tagged recombinant DHFR from human-derived Pneumocystis was catalytically active and demonstrated kinetics similar to the recombinant enzyme from rat-derived Pneumocystis. The present studies for production of soluble human-derived pDHFR indicated that the baculovirus expression system supported production of significantly purer catalytically active enzyme in higher yields than that expressed in bacterial cultures. These protocols now make it possible to facilitate screening of antifolates with selectivity for human-derived pDHFR and to determine its three-dimensional structure.     

Improving protein solubility: the use of the Escherichia coli dihydrofolate reductase gene as a fusion reporter

We have devised a strategy for screening mutant libraries for enzyme variants with enhanced solubility. The method is based on the observation that Escherichia coli can become insensitive to the antibiotic trimethoprim (TMP) if dihydrofolate reductase (DHFR) is expressed at an appropriate level. DHFR is a very soluble protein and can be expressed at levels that exceed normally lethal concentrations of TMP. In our approach, the gene encoding an insoluble target protein is placed in a vector so that the translated protein will be fused to DHFR. The resulting fusion protein will form inclusion bodies and inactivate DHFR-the cells will be susceptible to TMP. Mutations to the target protein that make it more soluble will also make the fusion protein more soluble so that DHFR will be at least partially active-the cells will be resistant to TMP. As the solubility of the target protein increases, the cells will become more resistant to TMP. The system was tested with a putative acetyltransferase (ACE) from a strain of the marine bacterium Vibrio fischerii. The gene encoding this protein was of interest since it is part of a mobile gene cassette within an integron array of the strain in question. After multiple rounds of shuffling and selection, ACE mutants were produced that had significantly improved solubility.     

The C terminus of mouse ornithine decarboxylase confers rapid degradation on dihydrofolate reductase. Support for the pest hypothesis

Several years ago, we proposed that polypeptide regions rich in proline (P), glutamic acid (E), serine (S), and threonine (T) (PEST) target intracellular proteins for destruction (Rogers, S., Wells, R., and Rechsteiner, M. (1986) Science 234, 364-368). To test the PEST hypothesis, we have produced chimeric proteins in which the N or C terminus of mouse dihydrofolate reductase is extended by the PEST-containing C terminus of mouse ornithine decarboxylase. Oligonucleotides encoding the 37 C-terminal residues of mouse ornithine decarboxylase (mODC) or equivalent lengths of dissimilar amino acids were inserted at appropriate sites in a dihydrofolate reductase (DHFR) expression vector. The various fusion proteins were expressed in Escherichia coli and purified to homogeneity by enzyme affinity chromatography. All purified fusion proteins exhibited similar abilities to convert dihydrofolate to tetrahydrofolate, thereby demonstrating that the attachment of peptide extensions to either terminus did not prevent the proper folding of DHFR. Metabolic stabilities of the radioiodinated fusion proteins were assayed in rabbit reticulocyte lysate or Xenopus egg extract. Proteolysis was found to be energy-dependent with mODC-DHFR fusion proteins being degraded from 2 to almost 40-fold faster than the parental DHFR molecule or DHFR fusion proteins bearing non-PEST extensions. Deletion of most of the PEST region from the mODC extension resulted in a significantly more stable fusion protein. Rapid proteolysis of DHFR proteins containing intact mODC extensions provides support for the PEST hypothesis.     

Import of a DHFR hybrid protein into glycosomes in vivo is not inhibited by the folate-analogue aminopterin

Dihydrofolate reductase fusion proteins have been widely used to study conformational properties of polypeptides translocated across membranes. We have studied the import of dihydrofolate reductase fusion proteins into glycosomes and mitochondria of Trypanosoma brucei. As signal sequences we used the last 22 carboxy-terminal amino acids of glycosomal phosphoglycerate kinase for glycosomes, and the cleavable presequences of yeast cytochrome b2 or cytochrome oxidase subunit IV for mitochondria. Upon addition of aminopterin, a folate analogue that stabilizes the dihydrofolate reductase moiety, import of the fusion protein targeted to glycosomes was not inhibited, although the results of protease protection assays showed that the fusion protein could bind the drug. Under the same conditions, import of a DHFR fusion protein targeted to mitochondria was inhibited by aminopterin. When DHFR fusion proteins targeted simultaneously to both glycosomes and mitochondria were expressed, import into mitochondria was inhibited by aminopterin, whereas uptake of the same proteins into glycosomes was either unaffected or slightly increased. These findings suggest that the glycosomes possess either a strong unfolding activity or an unusually large or flexible translocation channel.     

Ligand binding modulates the mechanical stability of dihydrofolate reductase

We use single-molecule force spectroscopy to demonstrate that the mechanical stability of the enzyme dihydrofolate reductase (DHFR) is modulated by ligand binding. In the absence of bound ligands, DHFR extends at very low forces, averaging 27 pN, without any characteristic mechanical fingerprint. By contrast, in the presence of micromolar concentrations of the ligands methotrexate, nicotinamide adenine dihydrogen phosphate, or dihydrofolate, much higher forces are required (82 +/- 18 pN, 98 +/- 15 pN, and 83 +/- 16 pN, respectively) and a characteristic fingerprint is observed in the force-extension curves. The increased mechanical stability triggered by these ligands is not additive. Our results explain the large reduction in the degradation rate of DHFR, in the presence of its ligands. Our observations support the view that the rate-limiting step in protein degradation by adenosine triphosphate-dependent proteases is the mechanical unfolding of the target protein.     

Testing the role of chain connectivity on the stability and structure of dihydrofolate reductase from E. coli: fragment complementation and circular permutation reveal stable, alternatively folded forms

The effects of chain cleavage and circular permutation on the structure, stability, and activity of dihydrofolate reductase (DHFR) from Escherichia coli were investigated by various spectroscopic and biochemical methods. Cleavage of the backbone after position 86 resulted in two fragments, (1--86) and (87--159) each of which are poorly structured and enzymatically inactive. When combined in a 1 : 1 molar ratio, however, the fragments formed a high-affinity (K(a) = 2.6 x 10(7) M(-1)) complex that displays a weakly cooperative urea-induced unfolding transition at micromolar concentrations. The retention of about 15% of the enzymatic activity of full-length DHFR is surprising, considering that the secondary structure in the complex is substantially reduced from its wild-type counterpart. In contrast, a circularly permuted form with its N-terminus at position 86 has similar overall stability to full-length DHFR, about 50% of its activity, substantial secondary structure, altered side-chain packing in the adenosine binding domain, and unfolds via an equilibrium intermediate not observed in the wild-type protein. After addition of ligand or the tight-binding inhibitor methotrexate, both the fragment complex and the circular permutant adopt more native-like secondary and tertiary structures. These results show that changes in the backbone connectivity can produce alternatively folded forms and highlight the importance of protein-ligand interactions in stabilizing the active site architecture of DHFR.     

Effects of five-tryptophan mutations on structure, stability and function of Escherichia coli dihydrofolate reductase

To elucidate the roles of tryptophan residues in the structure, stability, and function of Escherichia coli dihydrofolate reductase (DHFR), its five tryptophan residues were replaced by site-directed mutagenesis with leucine, phenylalanine or valine (W22F, W22L, W30L, W47L, W74F, W74L, W133F, and W133V). Far-ultraviolet circular dichroism (CD) spectra of these mutants reveal that exciton coupling between Trp47 and Trp74 strongly affects the peptide CD of wild-type DHFR, and that Trp133 also contributes appreciably. No additivity was observed in the contributions of individual tryptophan residues to the fluorescence spectrum of wild-type DHFR, Trp74 having a dominant effect. These single-tryptophan mutations induce large changes in the free energy of urea unfolding, which showed values of 1.79-7.14 kcal/mol, compared with the value for wild-type DHFR of 6.08 kcal/mol. Analysis of CD and fluorescence spectra suggests that thermal unfolding involves an intermediate with the native-like secondary structure, the disrupted Trp47-Trp74 exciton coupling, and the solvent-exposed Trp30 and Trp47 side chains. All the mutants except W22L (13%) retain more than 50% of the enzyme activity of wild-type DHFR. These results demonstrate that the five tryptophan residues of DHFR play important roles in its structure and stability but do not crucially affect its enzymatic function.     

Construction of a synthetic gene for an R-plasmid-encoded dihydrofolate reductase and studies on the role of the N-terminus in the protein

R67 dihydrofolate reductase (DHFR) is a novel protein that provides clinical resistance to the antibacterial drug trimethoprim. The crystal structure of a dimeric form of R67 DHFR indicates the first 16 amino acids are disordered [Matthews et al. (1986) Biochemistry 25, 4194-4204]. To investigate whether these amino acids are necessary for protein function, the first 16 N-terminal residues have been cleaved off by chymotrypsin. The truncated protein is fully active with kcat = 1.3 s-1, Km(NADPH) = 3.0 microM, and Km(dihydrofolate) = 5.8 microM. This result suggests the functional core of the protein resides in the beta-barrel structure defined by residues 27-78. To study this protein further, synthetic genes coding for full-length and truncated R67 DHFRs were constructed. Surprisingly, the gene coding for truncated R67 DHFR does not produce protein in vivo or confer trimethoprim resistance upon Escherichia coli. Therefore, the relative stabilities of native and truncated R67 DHFR were investigated by equilibrium unfolding studies. Unfolding of dimeric native R67 DHFR is protein concentration dependent and can be described by a two-state model involving native dimer and unfolded monomer. Using absorbance, fluorescence, and circular dichroism techniques, an average delta GH2O of 13.9 kcal mol-1 is found for native R67 DHFR. In contrast, an average delta GH2O of 11.3 kcal mol-1 is observed for truncated R67 DHFR. These results indicate native R67 DHFR is 2.6 kcal mol-1 more stable than truncated protein. This stability difference may be part of the reason why protein from the truncated gene is not found in vivo in E. coli.     

Nucleotide sequence of the dihydrofolate reductase gene of Saccharomyces cerevisiae

The nucleotide sequence of a DNA fragment that contained the Saccharomyces cerevisiae gene DFR coding for dihydrofolate reductase (DHFR) was determined. The DHFR was encoded by a 633-bp open reading frame, which specified an Mr24264 protein. The polypeptide was significantly related to the DHFRs of chicken liver and Escherichia coli. The yeast enzyme shared 60 amino acid (aa) residues with the avian enzyme and 51 aa residues with the bacterial enzyme. DHFR was overproduced about 40-fold in S. cerevisiae when the cloned gene was present in the vector YEp24. As isolated from the Saccharomyces library, the DFR gene was not expressed in E. coli. When the gene was present on a 1.8-kb BamHI-SalI fragment subcloned into the E. coli vector, pUC18, weak expression in E. coli was observed.     

Conformational changes in the active site loops of dihydrofolate reductase during the catalytic cycle

Escherichia coli dihydrofolate reductase (DHFR) has several flexible loops surrounding the active site that play a functional role in substrate and cofactor binding and in catalysis. We have used heteronuclear NMR methods to probe the loop conformations in solution in complexes of DHFR formed during the catalytic cycle. To facilitate the NMR analysis, the enzyme was labeled selectively with [(15)N]alanine. The 13 alanine resonances provide a fingerprint of the protein structure and report on the active site loop conformations and binding of substrate, product, and cofactor. Spectra were recorded for binary and ternary complexes of wild-type DHFR bound to the substrate dihydrofolate (DHF), the product tetrahydrofolate (THF), the pseudosubstrate folate, reduced and oxidized NADPH cofactor, and the inactive cofactor analogue 5,6-dihydroNADPH. The data show that DHFR exists in solution in two dominant conformational states, with the active site loops adopting conformations that closely approximate the occluded or closed conformations identified in earlier X-ray crystallographic analyses. A minor population of a third conformer of unknown structure was observed for the apoenzyme and for the disordered binary complex with 5,6-dihydroNADPH. The reactive Michaelis complex, with both DHF and NADPH bound to the enzyme, could not be studied directly but was modeled by the ternary folate:NADP(+) and dihydrofolate:NADP(+) complexes. From the NMR data, we are able to characterize the active site loop conformation and the occupancy of the substrate and cofactor binding sites in all intermediates formed in the extended catalytic cycle. In the dominant kinetic pathway under steady-state conditions, only the holoenzyme (the binary NADPH complex) and the Michaelis complex adopt the closed loop conformation, and all product complexes are occluded. The catalytic cycle thus involves obligatory conformational transitions between the closed and occluded states. Parallel studies on the catalytically impaired G121V mutant DHFR show that formation of the closed state, in which the nicotinamide ring of the cofactor is inserted into the active site, is energetically disfavored. The G121V mutation, at a position distant from the active site, interferes with coupled loop movements and appears to impair catalysis by destabilizing the closed Michaelis complex and introducing an extra step into the kinetic pathway.     

Effects of single-tryptophan mutations on R67 dihydrofolate reductase

R67 dihydrofolate reductase (DHFR) is an R-plasmid-encoded enzyme that confers clinical resistance to the antibacterial drug trimethoprim. This enzyme shows no sequence or structural homology to the chromosomal DHFRs. The active form of the protein is a homotetramer possessing D(2) symmetry and a single active-site pore. Two tryptophans occur per monomer: W38 and its symmetry-related residues (W138, W238, and W338) occur at the dimer-dimer interfaces, while W45 and its symmetry-related partners (W145, W245, and W345) occur at the monomer-monomer interfaces. Two single-tryptophan mutant genes were constructed to determine the structural and functional consequences of four mutations per tetramer. The W45F mutant retains full enzyme activity and the fluorescence environment of the unmutated W38 residues clearly monitors ligand binding and a pH dependent tetramer right harpoon over left harpoon 2 dimers equilibrium. In contrast, four simultaneous W38F mutations at the dimer-dimer interfaces result in tetramer destabilization. The ensuing dimer is relatively inactive, as is dimeric wild-type R67 DHFR. A comparison of emission spectra indicates the fluorescent signal of wild-type R67 DHFR is dominated by the contribution from W38. Equilibrium unfolding/folding curves at pH 5.0, where all protein variants are dimeric, indicate the environment monitored by the W38 residue is slightly less stable than the environment monitored by the W45 residue.     

Atypical transient state kinetics of recombinant human dihydrofolate reductase produced by hysteretic behavior. Comparison with dihydrofolate reductases from other sources

The transient state kinetics of catalysis for dihydrofolate reductase (DHFR) from several enzyme sources including highly purified recombinant human enzyme (rHDHFR) have been examined. Like DHFR from Escherichia coli, the enzyme from Lactobacillus casei, and isoenzyme 2 from Streptococcus faecium exhibit a slow increase in activity upon addition of substrates to enzyme. No slow hysteresis of this type was detected with recombinant human DHFR (rHDHFR) or DHFR from chicken or bovine liver or L1210 mouse leukemia cells (MDHFR). In contrast, both rHDHFR and MDHFR exhibited a very rapid decrease in activity (t1/2 = 30 and 20 ms, respectively) during a phase that occurred after the first turnover of the enzyme but before establishment of the steady state. This intermediate phase was not observed for the bacterial enzymes or the avian enzyme, nor was it observed with a mutant of rHDHFR in which Phe-31 has been replaced by leucine. For rHDHFR the intermediate phase is not a consequence of product inhibition, substrate depletion, or enzyme instability. It may therefore be concluded that this unusual transient state kinetic behavior results from the existence of two conformers of the enzyme, one of which has a higher turnover number than the other with the equilibrium shifting in favor of the less active conformer during the course of catalysis. The equilibrium is particularly favorable for the less active conformer when NADP is present in the active site of rHDHFR, whereas bound tetrahydrofolate favors the more active conformer. The more active conformer has a 6-fold higher Km for dihydrofolate than does the less active conformer. The existence of these conformers is likely to produce cooperative behavior by rHDHFR in vivo.     

Gammaherpesviruses encode functional dihydrofolate reductase activity

We overexpressed and purified from Escherichia coli the dihydrofolate reductase (DHFR) of the gammaherpesviruses human herpesvirus 8 (HHV-8), herpesvirus saimiri (HVS), and rhesus rhadinovirus (RRV). All three enzymes proved catalytically active. The K(m) value of HHV-8 DHFR for dihydrofolate (DHF) was 2.02+/-0.44 microM, that of HVS DHFR was 4.31+/-0.56 microM, and that of RRV DHFR is 7.09+/-0.11 microM. These values are approximately 5-15-fold higher than the K(m) value reported for the human DHFR. The K(m) value of HHV-8 DHFR for NADPH was 1.31+/-0.23 microM, that of HVS DHFR was 3.78+/-0.61 microM, and that of RRV DHFR was 7.47+/-0.59 microM. These values are similar or slightly higher than the corresponding K(m) value of the human enzyme. Methotrexate, aminopterin, trimethoprim, pyrimethamine, and N(alpha)-(4-amino-4-deoxypteroyl)-N(delta)-hemiphthaloyl-L-ornithine (PT523), all well-known folate antagonists, inhibited the DHFR activity of the three gammaherpesviruses competitively with respect to DHF but proved markedly less inhibitory to the viral than towards the human enzyme.     

Sequential conformation change and activation of chicken liver dihydrofolate reductase in low concentration of guanidine hydrochloride

The conformation changes of dihydrofolate reductase (DHFR) from chicken liver in guanidine hy-drochloride were monitored by protein intrinsic fluorescence, hydrophobic fluorescence probe TNS and limited proteol-ysis by proteinase K. The kinetics of the enzyme denaturation were also studied and compared with its activity changes. It was indicated by the enhanced fluorescence of 2-p-toluidinylnaphthalene (TNS) that a subtle conforma-tional change of the enzyme in dilute GuHCl parallels GuHCl-induced activation. At GuHCl concentration higher than 0.75 mol/L, the conformational change can be detected by increased susceptibility of the enzyme to proteinase K, but no significant gross conformational change of the enzyme molecule is observed by intrinsic fluorescence up to a GuHCl concentration of 1.2 mol/L. The results suggest that the denaturation of DHFR by GuHCl does not follow strictly the two-state model. The enzyme seems to open up sequentially with increasing concentrations of denaturants, mainly at th     

Pivotal role of Gly 121 in dihydrofolate reductase from Escherichia coli: the altered structure of a mutant enzyme may form the basis of its diminished catalytic performance

The structure and folding of dihydrofolate reductase (DHFR) from Escherichia coli and the mutant G121V-DHFR, in which glycine 121 in the exterior FG loop was replaced with valine, were studied by molecular dynamics simulations and CD and fluorescence spectroscopy. The importance of residue 121 for the chemical step during DHFR catalysis had been demonstrated previously. High-temperature MD simulations indicated that while DHFR and G121V-DHFR followed similar unfolding pathways, the strong contacts between the M20 loop and the FG loop in DHFR were less stable in the mutant. These contacts have been proposed to be involved in a coupled network of interactions that influence the protein dynamics and promote catalysis [Benkovic, S. J., and Hammes-Schiffer, S. (2003) Science 301, 1196-1202]. CD spectroscopy of DHFR and G121V-DHFR indicated that the two proteins existed in different conformations at room temperature. While the thermally induced unfolding of DHFR was highly cooperative with a midpoint at 51.6 +/- 0.7 degrees C, G121V-DHFR exhibited a gradual decrease in its level of secondary structure without a clear melting temperature. Temperature-induced unfolding and renaturation from the urea-denatured state revealed that both proteins folded via highly fluorescent intermediates. The formation of these intermediates occurred with relaxation times of 149 +/- 4.5 and 256 +/- 13 ms for DHFR and G121V-DHFR, respectively. The fluorescence intensity for the intermediates formed during refolding of G121V-DHFR was approximately twice that of the wild-type. While the fluorescence intensity then slowly decayed for DHFR toward a state representing the native protein, G121V-DHFR appeared to be trapped in a highly fluorescent state. These results suggest that the reduced catalytic activity of G121V-DHFR is the consequence of nonlocal structural effects that may result in a perturbation of the network of promoting motions.     

Ribosome display for selection of active dihydrofolate reductase mutants using immobilized methotrexate on agarose beads

Ribosome display was applied to the selection of an enzyme. As a model, we selected and amplified the dihydrofolate reductase (DHFR) gene by ribosome display utilizing a wheat germ cell-free protein synthesis system based on binding affinity to its substrate analog, methotrexate, immobilized on agarose beads. After three rounds of selection, the DHFR gene could be effectively selected and preferentially amplified from a small proportion in a mixture also containing competitive genes. Active enzymes were expressed and amplified and by sequence analysis, four mutants of DHFR were identified. These mutants showed as much activity as the wild-type enzyme.     

The coupling of structural fluctuations to hydride transfer in dihydrofolate reductase

The energy barrier for hydride transfer in wild-type G121V and G121S variants of Escherichia coli dihydrofolate reductase (DHFR) fluctuates in a time-dependent manner. This fluctuation may be attributed to structural changes in the protein that modulate the site of chemistry. Despite being far from the active site, mutations at position 121 of DHFR reduce the hydride transfer rate of the enzyme. This occurrence has been suggested to arise from modifications to the conformational ensemble of the protein. We elucidate the effects of the G121S and G121V mutations on the hydride transfer barrier by identifying structural changes in the protein that correlate with lowered barriers. The effect of these structural parameters on the hydride transfer barrier may be rationalized by simple considerations of the geometric constraints of the hydride transfer reaction. Fluctuations of these properties are associated with specific backbone dihedral angles of residues within the Methione-20 (M20) loop. The dihedral angle preferences are mediated by interactions with the region of the enzyme in the vicinity of residue 121 and are translated into distinct ligand conformations. We predict mutations within the M20 loop that may alter the conformational space explored by DHFR. Such mutational changes are anticipated to adjust the hydride transfer efficacy of DHFR by modifying equilibrium distributions of hydride transfer barriers found in the enzyme.     

Site-directed deletion mutants of a carboxyl-terminal region of human dihydrofolate reductase.

Substrate and inhibitor binding to dihydrofolate reductase (DHFR) primarily involves residues in the amino-terminal half of the enzyme; however, antibody binding studies performed in this laboratory suggested that the loop region located in the carboxyl terminus of human DHFR (hDHFR; residues 140-186) is involved in conformational changes that occur upon ligand binding and affect enzyme function (Ratnam, M., Tan, X., Prendergast, N.J., Smith, P.L. & Freisheim, J.H. (1988) Biochemistry 27, 4800-4804). To investigate this observation further, site-directed mutagenesis was used to construct deletion mutants of hDHFR missing 1 (del-1), 2 (del-2), 4 (del-4), and 6 (del-6) residues from loops in the carboxyl terminus of the enzyme. The del-1 mutant enzyme has a two-amino acid substitution in addition to the one-amino acid deletion. Deletion of only one amino acid resulted in a 35% decrease in the specific activity of the enzyme. The del-6 mutant enzyme was inactive. Surprisingly, the del-4 mutant enzyme retained a specific activity almost 33% that of the wild type. The specific activity of the del-2 mutant enzyme was slightly higher (38% wild-type activity) than that of the del-4 mutant. All three active deletion mutants were much less stable than the wild-type enzyme, and all three showed at least a 10-fold increase in Km values for both substrates. The del-1 and del-2 mutants exhibited a similar increase in KD values for both substrate and cofactor. The three active deletion mutants lost activity at concentrations of activating agents such as KCl, urea, and p-hydroxymercuribenzoate that continued to stimulate the wild-type enzyme. Antibody binding studies revealed conformational differences between the wild-type and mutant enzymes both in the absence and presence of bound folate. Thus, although the loops near the carboxyl terminus are far removed from the active site, small deletions of this region significantly affect DHFR function, indicating that the loop structure in mammalian DHFR plays an important functional role in its conformation and catalysis.