1-Deoxy-D-Xylulose 5-Phosphate Reductoisomerase
Created by Michelle Odette
1-deoxy-D-xylulose 5-phosphate reductoisomerase (PDB ID = 2y1d), often shortened to DXR or IspC, from Mycobacteria tuberculosis is an essential enzyme to convert isopentenyl diphosphate (IPP) into isoprenoids (1). Isoprenoids have many crucial functions in M. tuberculosis. These functions include a role in the development of the cell wall and a role in the electron transport chain leading to the production of ATP (2). In M. tuberculosis, IPP and subsequently isoprenoids are exclusively synthesized through the non-mevalonate pathway (2). The non-mevalonate pathway starts with pyruvate and D-glyceraldehyde 3-phosphate, using a variety of enzymes that results in IPP product (1). DXR is the enzyme for the second step of the non-mevalonate pathway. It catalyzes the rearrangement and reduction of 1-deoxy-D-xylulose 5-phosphate (DXP) to form 2-C-methyl-D-erythritol 4-phosphate (MEP) (1). Since one pathway forms essential isoprenoids in M. tuberculosis, all the enzymes of the non-mevalonate pathway are critical to the production of isoprenoids. Thus, DXR is a potential target for drugs to inhibit the production of isoprenoids and kill the M. tuberculosis bacterium that cause tuberculosis in humans.
DXR has a molecular weight of 83294.1 Da and an isoelectric point (pI) of 5.59. The pI is low because there are twice the number of negatively charged amino acids, Asp and Glu, as positively charged amino acids, Lys and Arg (3). The crystallized form of the protein of interest is missing residues 1 to 10, but does not affect the function of the protein. As a homodimer, DXR consists of two identical subunits, A and B. Each subunit has four domains: an N-terminal NADPH binding domain, a catalytic domain, a connecting region, and a C-terminal domain. The NADPH-binding domain, located at the N-terminal, consists of residues 11–150 (4). This domain’s tertiary structure is a Rossmann fold, commonly seen in proteins that bind to a nucleotide. The second domain is a central catalytic domain with an α/β-structure that consists of residues 151-275 (4). The catalytic domain contains the active site of the enzyme, which can also be the binding site for drugs to inhibit DXR function (1). The active site also contains the binding site for the metal ligand, Mn2+. Normally, the active site has two different conformations, open and closed. In the open conformation, the flexible active site flap, which consists of residues 198-208, does not cover the active site. This conformation allows the substrate, DXP, to enter and bind to the active site. In the closed conformation, the active site flap covers the active site and catalytic function is activated (1). The inhibitor fosmidomycin does not affect the active site flap change from open to closed. However, in the protein of interest, the α-aryl analogue of fosmidomycin that has a dichlorophenyl ring repulses the indole ring of Trp203 and prevents the active site flap from going into the closed conformation. Thus, this crystallization of mtDXR remains in the open conformation (1). The connecting region, which consists of residues 276-306, is the location in which most of the dimer interface occurs (4). The last domain is a C-terminal domain of a four-helix bundle, consisting of residues 307–389 (4). Each domain contributes to the V-shape of each subunit. The NADPH-binding domain and the C-terminal domain act as the arms of V formation, while the catalytic domain and connecting region are located at the vertex of the V-shape (5). All of these properties of the DXR’s structure allow DXR to function.
When crystallizing a protein, the protein will often lose some of its function in creating the crystal packing. In DXR, the active site flap of the B subunit is locked into an open conformation due to the crystal packing interactions. These interactions cause poor electron density and do not allow the substrate or inhibitor to bind. Thus, no ligand is located in the catalytic domain of the B subunit when DXR is crystallized. The A subunit is used to see the interactions between DXR and its substrate, its inhibitor, or its cofactor in the crystallized form (5).
DXR acts as an oxidoreductase and functions to reduce and rearrange DXP into MEP (1). The mechanism of this reaction occurs in multiple steps, in which the last step is NADPH dependent but not the rate determining step (6). The exact mechanism of the reaction is not known, but three possible mechanisms have been proposed (Figure 1). The first proposed mechanism is that DXR oxidizes the 3-hydroxyl group to its ketone form and transfers 1-hydroxyl-2-phosphoethyl to the second carbon. After this, hydrogens are donated from NAPDH to reduce the ketone into an alcohol. In the second proposed mechanism, DXR oxidizes the 4-hydroxyl group to generate the enol and enolate components of DXP via retroaldolization. The enol and enolate then go through aldolization in a different configuration, in which the intermediate is then reduced to MEP using NAPDH. The last proposed mechanism is that DXR goes through a 1,2-hydride shift and a 1,2-methyl shift, after which the intermediate is reduced to MEP using NAPDH as a source of hydrogen (6).
All of the mechanisms require that the cofactor NAPDH in order for the reaction to occur (6). In addition, a divalent cation ligand, such as Mn2+, Co2+, or Mg2+, must be present on DXR in order for the enzyme to function (4). The divalent cation helps DXR by forming electrostatic interactions with DXP, specifically with the two oxygens on the non-phosphate side of the DXP (6). This allows any of the three proposed mechanisms to occur easily. However, the different divalent cations have different binding affinities (Km) to the substrate. With a lower Km indicating a higher binding affinity to DXP from M. tuberculosis in DXR, Mn2+ has a Km of 4.0 ± 0.3 μM, Co2+ has a Km of 42 ± 7 μM, and Mg2+ has a Km of 100 ± 4 μM (2). Thus, Mn2+ has the highest binding affinity to DXP and will allow the reaction to occur at the fastest rate.
The ligands bound to DXR include two Mn2+ ions and a fosmidomycin analogue called (1S)-1-(3,4-dichlorophenyl)-3-[formyl(hydroxy)amino]propylphosphonic acid (Compund 9a). Compound 9a has a similar function to fosmidomycin in DXR, in which both inhibit DXR’s function. However, as previously stated, Compound 9a does not allow the flexibility of the active site flap in subunit A, so DXR remains in the open conformation when crystallized. Compound 9a has a similar structure to DXP, DXR’s natural substrate, and can fit into the same pocket that DXP usually fills (6). Thus, Compound 9a is a competitive inhibitor and will reduce the amount of substrate that is converted to MEP. The residues that interact with Compound 9a in this DXR crystallization are Asp151, Glu153, Glu222, Ser152, Ser177, Ser 213, Asn 218, and Lys219. Two Mn2+ ions are also found in the crystallization of DXR (PBD ID = 2y1d) on both subunit A and B (1). The Mn2+ ions can form electrostatic, or ionic, bonds with the two oxygens in DXP. Thus, Mn2+ serves two purposes in the reaction of DXP to MEP. Mn2+ promotes the ionization of the substrate hydroxyl group and assists in the formation of intermediates by stabilizing the alkoxide form of the substrate and intermediate hydroxyl groups (6). On DXR, the Mn2+ ion associates with Asp151, Glu153, and Glu222 by electrostatic interactions (5).
The residues that make up the active site, the divalent cation site, and the NADPH site fit that respective ligand and allow the reaction to occur. The active site, which is made up of residues Ser177, Ser213, Asn218, and Lys219, is a positively charged pocket that interacts with the negatively charged substrate DXP by hydrogen bonds and some electrostatic interactions. The active site has an active site flap of residues 198 to 208, but Trp203 causes the active site flap in this DXR crystallization to remain open due to repulsions with Compound 9a (1). Asp151, Glu153, and Glu222 interact with Mn2+, but still allowing Mn2+ to form additional electrostatic bonds with the substrate or with the inhibitor (5). Asp151, Glu153, Glu222, Ser152, Ser177 , Ser 213, Asn 218, and Lys219 interact with inhibitor 9a, which includes all of the active site residues and the residues that interact with the Mn2+ (1). NADPH-cofactor binding region is located in the N-terminal domain and NADPH interacts specifically with residues Thr21, Gly22, Ser23, Ile24, Gly47, Gly48, Ala49, Glu129, Gly206, and Asn209 (5).
The secondary structure of DXR is 42% alpha helices, 14 % beta sheets, and 42% random coils. There are 20 alpha helices, which is approximately 171 residues, and 13 beta strands of about 58 residues in DXR (1). The N-terminal domain that binds to NADPH has the secondary structure of one beta sheet, containing seven parallel beta strands and seven alpha helices surrounding the beta sheet. The catalytic domain has a beta sheet of four beta strands in which the third strand in the sheet is antiparallel to the other three strands. The beta sheet of the catalytic domain adheres to the C-terminal domain and the N-terminal domain with the use of a layer of helices. In addition, the connecting region has one alpha helix, whereas the C-terminal domain contains a four alpha helix bundle (4). The two subunits of DXR interact at the interface of the catalytic domain with the use of the connecting region. The core of the dimer interface is an imperfect beta barrel of ten beta strands. The four-stranded beta sheet from each subunit’s catalytic domain come together to form a twisted eight-stranded beta sheet and the two beta strands from the two connecting regions add to the ends of the twisted beta sheet, which creates the imperfect beta barrel (4).
Because there are other ligands that can be bound to DXR, there a couple of alternate conformations of DXR from M. tuberculosis (mtDXR) besides the protein of interest. One confirmation of mtDXR (PDB ID = 2JCZ) has fosmidomycin, Mn2+, and NADPH as ligands. In this conformation, the active site flap is in the closed conformation due to the presence of fosmidomycin instead of Compund 9a. Fosmidomycin does not interfere with Trp203 on the active site flap like Compound 9a does, so the active site flap is able to close in this DXR conformation (5). In addition, this conformation of DXR has 22 alpha helices and 13 beta strands (5). The number of beta strands of both proteins is the same, but there are a few more alpha helices in this closed conformation of DXR than protein of interest, which has 20 alpha helices. Another conformation of DXR (PDB ID = 1K5H) is an apoprotein, which means no that there are no prosthetics, or ligands, bound to it. The apoprotein DXR makes a super-open conformation of the active site flap, in which the active site flap is displaced away from its normal open conformation (4). Yet, another conformation of DXR (PDB ID = 2Y1C) is one in which the only ligand in Mn2+. Similar to the protein of interest, this conformation of DXR adopts an open conformation. However, the method of hydrogen bonding in the active site changes such that two water molecules fill the active site and form three hydrogen bond interactions, leaving the Ser213 side chain to rotate around and form an interaction with residue 209 (1).
DXR can be compared to proteins of similar tertiary structure using the DALI server (7) and compared to proteins of similar primary structure using the Psi-BLAST (8). The DALI Server compared the secondary structure of the protein of interest to a wide assortment of other proteins and gave a Z-score to indicate how well the comparison protein matched to the protein of interest, with a Z-score greater than 2 indicating a good match. The Psi-BLAST server allowed the protein of interest to be compared to other proteins by primary structure, in which an E-score close to zero is significant. From the results of the two servers, two comparison proteins were chosen, DXR from Yersinia pestis (PDB ID = 3IIE) and DXR from Zymomonas mobilis (PDB ID = 1R0K).
The first comparison protein, DXR from Yersinia pestis (PDB ID = 3IIE), receives a Z-score of 45.9 (7) and an E value of 5e-178 (8) with the protein of interest, which indicates that the protein has a high similarity to the protein of interest in both primary and tertiary structure. The comparison protein DXR is made of 47% alpha helices, consisting of 19 alpha helices of 189 residues, and 14% beta sheets, consisting of 13 beta strands of 60 residues (9). Thus, the comparison protein has one fewer alpha helix than the protein of interest. Also, the ligands in the comparison protein vary from the ligands in the protein of interest. The comparison protein has 1,2-ethanediol and Mn2+ as its associated ligands. The Mn2+ ion is bound to Asp151 and Glu153, which correlates with the protein of interest, but, instead of being bound to Glu222, the Mn2+ ion from the comparison protein is bound to an alternate glutamate, Glu 231, which indicates a change in that binding region. 1,2-ethanediol does not act as either the substrate or the inhibitor, but rather exists on the exterior surface of the protein to help maintain the structure and does not correlate with any ligand present in the protein of interest (9). In addition, the comparison protein has 86 negatively charged residues (Asp and Glu) and 72 positively charged residues (Arg and Lys), which has significantly more positively charged amino acids than the protein of interest mtDXR (3). Despite these differences in primary structure, the essential residues and structure of the catalytic active site and the active site flap are conserved in both proteins, which allow both proteins to fulfill the same function of rearranging and reducing DXP into MEP.
DXR from Zymomonas mobilis (PDB ID = 1R0K) receives a Z-score of 47.3 (7) and an E value of 8e-168 (8) with mtDXR. However, this comparison protein, which is an apoprotein, dimerizes with itself and crystallizes into two homodimers with four identical subunits (10). In comparison, the protein of interest does not dimerize with itself and only contains the homodimer. Because the protein of interest does contain ligands and the comparison protein does not, these differences in the crystallization do not explicitly affect the two proteins’ functions (1). In addition, other differences include different residues in the NADPH binding site, specifically in the recognition of the adenine ring of NADPH (10). Although the proteins have differences in the primary structure and the crystallization methods, the two proteins share similar secondary structure, including a conserved active site to bind to the substrate DXP. The comparison protein consists of 49% alpha helices with 23 alpha helices of 191 residues and 16% beta sheet with 13 beta strands of 63 residues (10). From these, the secondary structures of the two proteins are very similar due to serving the same function in different organisms of converting DXP to MEP.