Dihydrofolate Reductase (1RF7) from Escherichia coli 
Created by: Nana Bosomtwe

Dihydrofolate reductase (DHFR) from Escherichia coli (PDB ID: 1RF7) is a housekeeping enzyme that maintains the intracellular levels of tetrahydrofolate. Tetrahydrofolate is a precursor of cofactors necessary for the biosynthesis of purines, pyrimidines, and several amino acids. Dihydrofolate reductase, which is the sole source of tetrahydrofolate, catalyzes the reduction of 7,8-dihydrofolate (DHF) to 5,6,7,8-tetrahydrofolate (THF), using nicotinamide adenine dinucleotide phosphate (NADPH) as an electron donor (1, 8). Being the only source of THF, DHFR is an Achilles’ heel of rapidly proliferating cells such as cancer cells, making it an attractive target of several anticancer and antimicrobial drugs (1). These drugs, such as methotrexate, trimethoprim, and pyrimethamine, inhibit DHFR by competitively binding to its active site (2). Inhibition affects the production of THF and the synthesis of nucleotides and several amino acids; thus, it slows the rate of cell proliferation. Since cancer cells are often the most rapidly reproducing cells, the drugs will have the strongest effect on the cancer cells. Researchers are interested in DHFR as a target in the fight against cancer and are interested to use DHFR to a therapeutic advantage (1).

Escherichian coli (E. coli) DHFR is a small protein with a molecular weight around 18 kDa and isoelectric point of 4.76 (1, 2). Crystallographic studies using x-ray diffraction show that it is a monomeric enzyme with many secondary structural elements. These include a central eight-stranded β-sheet labeled βA through βH; seven of these strands are parallel and one runs antiparallel. Four α-helices, labeled B through F, are connected to the β-sheets by random coils (1). DHFR has two subdomains, the adenosine-binding subdomain and the loop subdomain. The adenosine-binding subdomain is the smaller of the two and binds the adenosine moiety of NADPH. The adenosine-binding subdomain consists of βB (residues 39 – 43), βC (residues 58 – 63), βD (residues 73 – 75), and βE (residues 89 – 95), as well as α-helices C (residues 43 – 50), E (residues 77 – 86), and F (residues 96 – 104) (12, 1). The loop subdomain contains three loops, the Met-20 loop (residues 10 – 24), the F-G loop (residues 117-131), and the G-H loop (residues 146 – 148). Between the two subdomains is the active site, a long groove, a hydrophobic cleft where the substrate DHF and the cofactor NADPH bind (5). The size of the active site is regulated by the movements of the two subdomains. Hinge-bending is a result of stored elastic energy in α-helices, which drives the proximal movement between the two subdomains. Lys-38 and Val-88contribute to the hinge-bending so that NADPH is brought closer to the active site (1).

DHFR has three conformations: open, closed, and occluded. The catalytic function of DHFR is related to the dynamic fluctuation between the three conformations (3). The reaction pathway can be split into five kinetic intermediates: a holoenzyme form bound to NADPH, a Michaelis complex with NADPH and DHF, a transition state complex, a THF binary complex, and a THF-NADPH complex (7). The first three kinetic intermediates correspond to the closed conformation, in which the Met-20 loop covers the active site, protecting the substrate from solvent interaction. DHF fits into the para-aminobenzoyl glutamate (pABG) binding pocket of DHFR, and DHF initially binds to the residues Ile-5, Ala-7, Met-20, Trp-22, Asp-27, Leu-28, Phe-31, Lys-32, Arg-52, Leu-54 and Arg-57 in order to form a binary complex, which is in its open conformation (8). NADPH binds to DHFR while in its extended conformation, along the 5 β-strands at the conserved residues Ala-7, Ile-14, His-45, Thr-46, Leu-62, Ser-63, Ser-64, Lys-76, Ile-94, Gly-95 and Arg-98. The center of the Met-20 loop (residues 16 – 19) form a type III β-hairpin turn.Asn-18(N-eta-1) forms a hydrogen bond with His-45(O) of helix C, the N terminal of the Met-20 loop forms two hydrogen bonds with the F-G loop, between Gly-15(O) and Asp-122(N), and Glu-17(N) and Asp-122(O-eta-2), stabilizing this intermediate structure. Ile-14(O) of the Met-20 loop forms a hydrogen bond with the nicotinamide carboxamide of NADPH. The van der Waals interactions between Asn-18, Ala-19, and Met-20 with the ribose of NADPH further stabilize this conformation (1).

The proximity of NADPH and DHF promotes the reduction reaction.  The type III β-hairpin within the central Met-20 loop flips out so that the residues Asn-18 and Met-20 collapse down, bringing the NADPH and DHF close together. Hydrogen bonding between the pyrophosphate oxygen atoms of NADPH and N-terminal amide nitrogen atoms of helix B stabilizes the close interaction of the nicotinamide and pteridine rings. The active site holds NADPH and DHF within a van der Waals radius and as the hydride is transferred, the side chain of Met-20 chain moves away from N-5 of DHF, allowing a water molecule to enter the active site. The Met-20 loop helps stabilize the nicotinamide ring of the NADPH to promote the transfer of the hydride from NADPH to DHF and protonation of N5 of DHF (6). In the end, DHF is reduced to THF and NADPH is oxidized to NAD. The high flexibility of Met-20 and other loops near the active site play a role in promoting the release of the product, THF (3, 7). 

The occluded conformation corresponds to the last two kinetic intermediates of the reaction pathway. This conformation is marked by the Met-20 loop forming a -helix.  The Met-20 loop extends into the nicotinamide binding pocket, disrupting the binding of DHFR to NADPH. Residues 17 – 20 of the Met-20 loop form a - helix, and Met-16 engages in van der Waals interactions with Thr-46 and Ser-49 of helix C. The hydrogen bonds between the N-terminal of the Met-20 loop and the F-G loop are disrupted, and the C-terminal of the Met-20 loop forms a pair of hydrogen bonds with the G-H loop, between Asn-23(O) and Ser-148(N), and Asn-23(N) and Ser-148(O-gamma). This sterically hinders the nicotinamide ring, preventing it from being held within the active site. The enzyme’s conformational changes cause the ternary complex to reach a higher energy than that of its holoenzyme form.  The long range motion between the adenosine binding loop and distant regions causes a relaxation of the ternary complex, which releases the THF (7).

The PDB entry for DHFR of E. coli shows that the enzyme has four ligands and prosthetic groups: NAD, manganese (Mn(II)), β-mercaptoethanol and folic acid. Mn(II) and β-mercaptoethanol were used to produce isomorphous protein crystal structures, and do not play a role in the biological function of DHFR (1). NAD is the oxidized form of NADPH, the cofactor that serves as the hydride donor to the substrate, DHF (1,2). Folic acid is the structural precursor to DHF, which has two additional hydrogen atoms. The adenosine binding subdomain binds the cofactor NADPH, and the loop subdomain contains the substrate, DHF, and these two subdomains facilitate hydride transfer from C-4 of NADPH to C-6 of DHF and protonation of N-5 of DHF (6, 1).

Methotrexate (MTX) is a competitive inhibitor of DHFR, and binds tightly in the pABG pocket that normally fits DHF. The protein-drug complex (PDB ID: 1RX3) is similar to the protein-substrate complex, as MTX and DHF are structurally similar. DHF cannot form hydrogen bonds with the carbonyl oxygen atoms of Leu-5(O) and Ile-94(O), but MTX can do so. This partly contributes to the higher binding affinity of MTX to DHFR (8). As a competitive inhibitor, MTX binds to the active site of DHFR with higher affinity, blocks DHF from binding to the active site, and prevents the conversion of DHF into THF. THF is essential to the synthesis of DNA; this inhibition by MTX indirectly disrupt the cell’s ability to synthesize DNA, and subsequently cells are unable to reproduce. Since cancer cells are often the most rapidly reproducing cells in a patient, the drugs will have the strongest effect on the cancer cells. And thus MTX is used as an anticancer drug.

DHFR from Bacillus anthracis (PDB ID: 3E0B) is structurally and functionally homologous to DHFR from E. coli; it has a calculated PSI-BLAST e value of 4e-88, which corresponds to a high degree of amino acid sequence similarity. 39% of overall sequence and 62% sequence of the active site region of DHFR from Bacillus anthracis (B. anthracis) matches that of DHFR from E. coli (9). It also has a Dali z score of 26.6, which indicates a high degree of tertiary structural similarity (10). The proteins are overwhelmingly similar, as demonstrated in the superimposition of the two enzymes. Both enzymes have a secondary structure comprised of four α-helices and an eight-stranded β-sheet, and both bind NADPH and DHF in similar binding pockets (14). The α-helices are interspersed along the outside of the globule. Both protein forms have an active site within the protein core, surrounded by a looping strand of amino acids (10). Compared to the DHFR from E. coli, the hydrophobic pABG binding pocket of DHFR from B. anthracis contains Asn-47 in place of Thr-46 (in E. coli), Ala-50 in place of Ser-49, and conserves Ile-51 and Leu-55 (14). The changes in the binding domain make the B. anthracis DHFR susceptible to folate analog inhibitors that are not specific for DHFR from other species (4). DHFR from B. anthracis is sensitive to the nonselective inhibitor MTX, but resistant to trimethoprim (TMP), a selective DHFR inhibitor for E. coli and Staphylococcus aureus (PDB ID: 3M08) (11). Finding a selective DHFR inhibitor for B. anthracis could reduce the threat of anthrax as a biological weapon.