The Cystathionine-β-synthase Domains on the Guanosine 5′-Monophosphate Reductase and Inosine 5′-Monophosphate Dehydrogenase Enzymes from Leishmania Regulate Enzymatic Activity in Response to Guanylate and Adenylate Nucleotide Levels

Summary

The Leishmania guanosine 5′-monophosphate reductase (GMPR) and inosine 5′-monophosphate dehydrogenase (IMPDH) are purine metabolic enzymes that function maintaining the cellular adenylate and guanylate nucleotide. Interestingly, both enzymes contain a cystathionine-β-synthase domain (CBS). To investigate this metabolic regulation, the Leishmania GMPR was cloned and shown to be sufficient to complement the guaC (GMPR), but not the guaB (IMPDH), mutation in E. coli. Kinetic studies confirmed that the Leishmania GMPR catalyzed a strict NADPH-dependent reductive deamination of GMP to produce IMP. Addition of GTP or high levels of GMP induced a marked increase in activity without altering the K_m values for the substrates. In contrast, the binding of ATP decreased the GMPR activity and increased the GMP K_m value 10-fold. These kinetic changes were correlated with changes in the GMPR quaternary structure, induced by the binding of GMP, GTP, or ATP to the GMPR CBS domain. The capacity of these CBS domains to mediate the catalytic activity of the IMPDH and GMPR provides a regulatory mechanism for balancing the intracellular adenylate and guanylate pools.

Introduction

Protozoan parasites of the Leishmania genus are the etiologic agent of leishmaniasis, a spectrum of disease states, ranging from disfiguring but self-healing cutaneous lesions to a visceralizing form of the disease that effects predominantly the liver, spleen and bone marrow (Desjeux, 1996). Leishmania are digenetic parasites that exist as highly motile, flagellated promastigotes in the alimentary tract of the phlebotomine sandfly vector and as non-motile, intracellular amastigotes that reside in the phagolysosomal compartment of mammalian host macrophages. Like all other protozoan parasites investigated to date, Leishmania lack the biosynthetic machinery for de novo synthesis of the purine ring (Marr et al., 1978). Consequently, these parasites obligatorily rely on purine salvage from the host milieu in order to survive and proliferate (Marr et al., 1978, Hassan & Coombs, 1985, LaFon et al., 1982). Recent studies have demonstrated that Leishmania are capable of sensing the levels of purine in the extracellular environment (de Koning et al., 2000, Carter et al., 2010). In response to purine starvation, L. donovani rapidly upregulates purine salvage machinery including the plasma membrane nucleoside transporters, LdNT1 and LdNT2, and the purine salvage enzymes hypoxanthine-guanine phosphoribosyltransferase (HGPRT), xanthine phosphoribosyltransferase (XPRT), and adenine phosphoribosyltransferase (APRT) as a strategy to maintain intracellular purine nucleotide levels (de Koning et al., 2000, Carter et al., 2010).

Metabolic flux (LaFon et al., 1982, Marr et al., 1978) and biochemical studies in Leishmania have underscored four enzymes, HGPRT, XPRT, APRT, and adenosine kinase (AK), capable of salvaging preformed extracellular purines into the parasite nucleotide pool (Hwang & Ullman, 1997, Iovannisci et al., 1984, Kidder & Nolan, 1982, Hwang et al., 1996, Jardim et al., 1999, Iovannisci & Ullman, 1984) (Fig. 1). Genetic analyses of mutant Δhgprt, Δxprt, Δaprt, and AK-deficient L. donovani cell lines revealed that no single purine salvage enzyme is, in itself, absolutely critical for survival or growth (Boitz & Ullman, 2006a). Furthermore, the construction and characterization of a L. donovani Δhgprt/Δxprt double knockout cell line has established that a functional HGPRT or XPRT is indispensable for purine salvage in both the promastigote and amastigote stages of the parasite, while APRT and AK are functionally superfluous (Boitz & Ullman, 2006b, Boitz et al., 2012). These genetic studies also demonstrated that adenylosuccinate lyase, the enzyme required for AMP synthesis is crucial for survival of both life cycle stages of L. donovani (Boitz et al., 2013) and represent a potential therapeutic target (Fig. 1).

Fig. 1.

Purine metabolism in Leishmania. The diagram illustrates the enzymatic machinery in Leishmania parasites involved in the salvage of 6-oxopurines and 6-aminopurines and the metabolic pathways that produce the adenylate and guanylate monophosphate nucleotide pools. 1, GMP reductase; 2, inosine monophosphate dehydrogenase; 3, GMP synthase; 4, xanthine phosphoribosyltransferase; 5, guanine deaminase; 6, hypoxanthine-guanine phosphoribosyltransferase; 7, adenylosuccinate synthetase; 8, adenylosuccinate lyase; 9, AMP deaminase; 10, adenine phosphoribosyltransferase; 11, adenine aminohydrolase; IMP, inosine monophosphate; Ade, adenine; Hyp, hypoxanthine; XMP, xanthine monophosphate; Xan, xanthine; Gua, guanine.

Two pivotal nucleotide interconverting enzymes that are postulated to regulate cellular guanylate and adenylate pools in Leishmania are inosine monophosphate dehydrogenase (IMDPH) and guanosine monophosphate reductase (GMPR), two homologous enzymes that catalyze substantially different reactions (Fig. 1). The recombinant L. donovani IMPDH (LdIMPDH) enzyme has been purified to homogeneity and was shown to be susceptible to inhibition by mycophenolic acid (MPA) (Dobie et al., 2007), a well-characterized inhibitor of an assortment of phylogenetically diverse IMPDH enzymes (Umejiego et al., 2004, Weber et al., 1992, Hedstrom, 2009, Gan et al., 2002, Wilson et al., 1991, Sintchak et al., 1996, Digits & Hedstrom, 2000). MPA is also a potent inhibitor of L. donovani promastigote growth in medium, in which either adenine or hypoxanthine is the purine source, but does not impact growth in medium supplemented with xanthine or guanine, either of which can circumvent the pharmacologic block after phosphoribosylation (Fig. 1). A similar growth phenotype was observed when promastigotes of a Δldimpdh null mutant were exposed to media in which the purine source was varied (Fulwiler et al., 2011). The Δldimpdh parasites elicited parasite burdens in mice that were similar to those of wild type parasites, intimating that the parasite can extract a source of xanthine or guanine/guanosine from the phagolysosome. These data suggest that the Leishmania GMPR, which is essential for shunting xanthine and guanine into adenylate nucleotides (Fig. 1), is likely an important checkpoint for regulating the intracellular adenylate and guanylate nucleotides levels.

GMPR catalyzes an NADPH-dependent deamination of GMP to IMP. Several crystal structures have been determined for the human GMPR bound to ligands within the active site (Li et al., 2006, Patton et al., 2011). These structures establish GMPR as a member of the α/β TIM barrel protein family. Moreover, kinetic characterization of the human and Artemia GMPRs revealed that these enzymes exhibit a bimodal substrate saturation curve that is highlighted by a notable increase in catalytic activity in the presence of GMP (Renart et al., 1976, Spector et al., 1979). This kinetic behavior has been attributed to a heteromeric complex composed of two different GMPR isoforms (Zhang et al., 2003, Deng et al., 2002). LdGMPR has been suggested to be potential antileishmanial chemotherapeutic target as pyrazolopyrimidine nucleotide monophosphates were potent GMPR inhibitors that induced a misregulation of the intracellular adenylate and guanylate pools (Looker et al., 1986, Spector & Jones, 1982).

Numerous studies have found that IMPDH and GMPR from phylogenetically diverse organisms favor the formation of tetrameric structures (Labesse et al., 2013, Li et al., 2006, Patton et al., 2011, Martinelli et al., 2011, Luecke et al., 1997, Dobie et al., 2007, Zhang et al., 1999, Wang et al., 1996). Another common structural feature of many IMPDH and GMPR enzymes is a ~131 amino acid segment containing two tandemly arranged cystathionine-β-synthase (CBS) motifs that form a CBS domain, also known as a Bateman domain (Scott et al., 2004, Ignoul & Eggermont, 2005, Pimkin & Markham, 2008, Zhang et al., 1999, Sintchak et al., 1996). Bioinformatic analysis of the Leishmania genomes has revealed that the GMPR and IMPDH enzymes both possess CBS domains. Bateman domains have been ascribed a regulatory role for controlling protein function in response to adenylate nucleotide levels (Baykov et al., 2011, Ignoul & Eggermont, 2005, Scott et al., 2004, Hardie, 2003). Although CBS domains have been detected in a wide array of proteins which include; protein kinases, magnesium transporters, voltage gated channels, and most GMPR and IMPDH enzymes (Hardie, 2003, Scott et al., 2004, Pimkin et al., 2009, Baykov et al., 2011, Ignoul & Eggermont, 2005), these domains do not appear to be essential for IMPDH activity, as the absence of this domain in the B. burgdorferi or C. parvum IMPDH does not impact catalytic activity (Nimmesgern et al., 1999, Zhou et al., 1997, Hedstrom, 2009). Crystal structures of IMPDH and GMPR from evolutionarily diverse organisms indicate that CBS domains are typically located distal to the active site of these enzymes (Labesse et al., 2013, Zhang et al., 1999, Sintchak et al., 1996, Makowska-Grzyska et al., 2015). However, the retention of the CBS domains in the Leishmania IMPDH and GMPR suggest that these elements could function in a regulatory capacity to control the activity of these enzymes (Labesse et al., 2013, Pimkin et al., 2009).

Here we demonstrate that the enzymatic activities of the LdIMPDH and LmGMPR are regulated by the differential binding of allosteric effectors ATP and GTP to the CBS domains of their respective enzymes. The binding of GTP caused a reduction in LdIMDPH activity which was accompanied by a structural changes favoring tetramer formation, whereas the binding of ATP promoted tetramerization failed to alter the LdIMPDH enzymatic activity. In contrast, the binding of GTP or GMP to LmGMPR stabilized the assembly of an active homotetramer; while ATP promoted the formation of a monomer/dimer structure with reduced catalytic activity. We also demonstrate that CBS domains in LdIMPDH and LmGMPR bind both adenylate and guanylate nucleotides and potentially modulate enzymatic activity.

Results

Isolation of the LmGMPR

The human GMPR isoform 2 was used to identify two homologous ORFs, LmjF19.1560 and LmjF17.0725, located on chromosome 19 and chromosome 17, respectively. Both LmjF19.1560 and LmjF17.0725 were annotated as IMPDH genes in the L. major genome database. A prior biochemical investigation with LdIMPDH (Dobie et al., 2007) effectively confirmed that the Lmj19.1560 ORF encoded the L. major IMPDH. A pairwise alignment of the L. donovani (LdIMPDH) and Lmj17.0725 (LmGMPR) proteins demonstrated that the two encoded proteins shared 34% amino acid sequence identity. Similar sequence alignments revealed that the predicted LmGMPR shared 30–41% sequence identity with IMPDHs from humans, E. coli, and Borrelia burgdorferi, but only 24–26% sequence identity with the E. coli and human GMPRs (Fig. 2), establishing that the LmGMPR is more related in terms of primary structure to IMPDH than GMPR enzymes.

Fig. 2.

Multiple sequence alignment of GMP reductases and IMP dehydrogenases. Protein sequences for L. major (LmGMPR; Q4QEB3), L. donovani (LdGMPR; E9BDA8), E. coli (EcGMPR; EDV65100), Homo sapiens (HsGMPR2; CAG33437) guanosine monophosphate reductases, L. donovani (LdIMPDH; AAA29253), Homo sapiens isoform 1 (HsIMPDH1; P20839), E. coli (EcIMPDH; P0ADG8), and Tritrichomonas foetus (TfIMPDH; AAB01581) inosine monophosphate dehydrogenases were aligned using the computer program CLUSTAL X. Signature sequences containing key amino acids that, i) encode the cystathionine β synthtase domain (CBS), ii) mediate interactions with the adenine base of the nicotinamide adenine dinucleotide (phosphate) (NAD(P)H binding site (shaded dark grey) (Riera et al., 2008)), iii) loop containing the catalytic cysteine (catalytic cys, boldface print), vi) form the binding sites for the IMP/GMP ribose moiety (IMP/GMP ribose, double underline & boldface print), v) hydrogen bond with the IMP/GMP 5′-phosphate binding site (IMP/GMP 5′ PO_4, wavey underline & boldface print), vi) hydrogen bond with the 6-exocyclic oxygen atom of the IMP or GMP purine ring (IMP/GMP O6, dashed underline & boldface print), vii) residues stabilizing the binding of mycophenolic acid/NAD⁺ (Sintchak et al., 1996, Riera et al., 2008) are indicated (underlined and shaded grey) on the human IMPDH1 sequence.

Bioinformatic and structural comparisons with the T. foetus IMPDH and human GMPR (Patton et al., 2011, Prosise & Luecke, 2003) permitted the identification of a number of functional motifs in LmGMPR. These include: i) the highly conserved IMPDH/GMPR signature sequence containing an active site catalytic cysteine that forms an enzyme-XMP or enzyme-IMP covalent intermediate (Hedstrom, 2012, Gan et al., 2002), ii) an NADPH binding site that predominantly forms interactions with the adenine moiety of the cofactor, iii) an IMP/GMP binding cassette that forms contacts with the ribose-5′-phosphate, or the 6-exocyclic oxygen constituent of the nucleotide substrate, iv) a low affinity mycophenolic acid binding motif, and v) a C-terminal tripeptide (Ser-Lys-Leu), a type 1 topogenic signal that targets proteins to the glycosome (Riera et al., 2008, Dobie et al., 2007, Gan et al., 2002, Jardim et al., 2000, Umejiego et al., 2004, Zarella-Boitz et al., 2004, Sintchak et al., 1996), a peroxisomal-like microbody present in Leishmania and related trypanosomatids (Michels et al., 2005) (Fig. 2). A unique structural feature that distinguishes LmGMPR from the human and E. coli counterparts is a 131 amino acid insertion encoding two tandem CBS domains. A number of studies have demonstrated that CBS domains bind adenylate nucleotides (Thomas et al., 2012, Scott et al., 2004, Labesse et al., 2013, Hirata et al., 2014) and are elements conserved among most IMPDHs, with the exception of the Borrelia burgdorferi and Cryptosporidium parvum IMPDH. Interestingly, although the LdIMPDH and LmGMPR polypeptides both contain putative CBS domains, the CBS domains in the two proteins only exhibit ~24% amino acid identity.

LmGMPR complements the ΔguaC mutation in E. coli

To validate the enzymatic activity LmGMPR a complementation analysis was performed using the E. coli strains H724 and H1174 which lack the bacterial IMPDH (ΔguaB) or GMPR (ΔguaC), respectively. These E. coli strains also have an additional mutation that blocks the de novo synthesis of purines (De Haan et al., 1969, Nijkamp & De Haan, 1967). E. coli H1174 transformed with the pQE80-LmGMPR plasmid showed robust growth in all purines including guanine, a result confirming that LmGMPR was a bona fide GMPR (Fig. 3A). In contrast, E. coli H1174 transformed with the pQE80 control vector grew in media containing hypoxanthine or adenine, nucleobases that do not require a functional GMPR for bacterial growth, but not in media containing only guanine. Since the E. coli GMPR lacks a CBS domain (Fig. 2), we next examined the capacity of the lmgmprΔCBS, a deletion mutant lacking the CBS domain to complement the growth of E. coli H1174. Surprisingly, E. coli H1174 expressing lmgmprΔCBS showed low levels of growth in guanine containing media (Fig. 3A). These data indicate that loss of the CBS domain may impair the catalytic activity of lmgmprΔCBS and/or cause a disruption of the protein structure.

Fig. 3.

Complementation of E. coli mutants with the LmGMPR. E. coli strains H1174 (ΔguaC) (A) and H724 (ΔguaB) (B) deficient in the bacterial GMPR or IMPDH activities, respectively, were transformed with the pQE80L, pQE80L-LmGMPR, or the pQE80L-lmgmprΔCBS construct. Clones were then cultivated in chemically defined media containing either no purines or 100 μM each of hypoxanthine, guanine, or adenine. Cultures were incubated with vigorous shaking for 16 h and the culture densities were measured at an OD₆₀₀. Experiments were performed in triplicate and each experiment was repeated three times.

Transformation of the E. coli H724, a strain lacking IMPDH activity with the pQE80-LmGMPR plasmid showed no significant growth in media containing hypoxanthine or adenine (Fig. 3B). These data established that LmGMPR failed to convert IMP to XMP in bacteria. In contrast, E. coli H724 transformed with the pQE80-LdIMPDH exhibited vigorous growth in hypoxanthine and adenine (data not shown). No growth of E. coli H1174 and H724 was detected in media lacking an exogenous purine source (Fig. 3).

Steady state kinetic analysis of LmGMPR

Kinetic analysis of LmGMPR was performed on recombinant enzyme expressed in E. coli strain H1174 and purified to apparent homogeneity using Cibacron blue affinity chromatography (Fig. S1). At a fixed NADPH concentration, plotting the initial velocity of LmGMPR as a function of GMP concentration resulted in a response with a characteristic sigmoidal curve (Fig. 4A) that was analyzed by fitting the data to the Hill equation (equation 1). K_s and V_max values for GMP (Table I) were calculated for LmGMPR in the presence of a saturating amount of NADPH. Further analysis of these data using the Hill plot (log (v/V_max-v) versus log [GMP]) revealed a Hill coefficient (n) of 2.3, a value that was diagnostic of GMP inducing a positive cooperativity effect. Interestingly, addition of the divalent cation Mg²⁺ to the reaction buffer had no significant affect on enzymatic activity.

Fig. 4.

Steady state kinetic analysis of LmGMPR. (A) The K_m value for GMP was determined at a fixed saturating concentration of NADPH while varying the concentration of GMP. The initial velocity data was plotted as a function of GMP concentration to examine the Michaelis-Menten relationship. (B), The K_m value for NADPH was determined at a fixed GMP concentration while varying the NADPH concentration. The Michaelis-Menten plot was generated by plotting the initial rate data vs NADPH concentration. Hanes-Woolf plot analysis of the initial rate data (inset) shows a linear Michaelis-Menten response with respect to the NADPH concentration.

Table I.

Kinetic parameters of LmGMPR^a

Substrate	Ks (Km) (μM)	Vmax (nmol min−1)
GMP	48 + 4.0	3.0 ± 0.3
NADPH	9 ± 2.0	2.2 ± 0.2
GMP + GTP	16 ± 3.2	23.4 ± 0.4
GMP + ATP	535 ± 35	0.46 ± 0.1
NADPH + GTP	10 ± 2.3	5.8 ± 0.4
NADPH +ATP	9 ± 1.9	6.2 ± 0.6

^asee (Dobie et al., 2007) for comparable kinetic values for LdIMPDH

Analysis of the LmGMPR kinetics with respect to NADPH showed initial velocities that exhibited a hyperbolic response as a function of NAPDH concentration (Fig. 4B). K_m and V_max were calculated from Hanes-Woolf plot analysis of the initial rate data (Fig. 4B, inset & Table I). Comparable kinetic studies performed with NADH (up to 5.0 mM) failed to induce the reductive deamination of GMP, indicating that LmGMPR exhibits a strict requirement for NADPH as the hydride donor.

To further investigate the regulatory properties of LmGMPR and the capacity of ATP and GTP to modulate enzymatic activity, kinetic studies were performed in the presence of increasing concentrations of these nucleotide triphosphates. Interestingly, addition of 10 μM GTP resulted in a ~8-fold increase in the V_max value and concomitantly decreased the GMP K_s value ~3-fold (Fig. 5A). Higher concentrations of GTP (>10 μM) resulted in only modest increases in GMPR activity. Although LmGMPR bound GTP, steady state kinetic analysis showed that this nucleotide triphosphate was not utilized as a substrate.

Fig. 5.

ATP and GTP modulate LmGMPR activity. The effect of the nucleotide triphosphates (A) no GTP (■), 10 μM GTP (□), 50 μM GTP (●), 100 μM GTP (○), 200 μM GTP (◆), and 1000 μM GTP (⋄) or (B) no ATP (■), 10 μM ATP (□), 25 μM GTP (●), 50 μM GTP (○), 200 μM GTP (◆), and 1000 μM GTP (⋄)ATP on LmGMPR activity and the GMP K_m value was assessed in the presence of increasing GTP or ATP at a fixed concentration of NADPH. The effect of (C) no GTP (■), 25 μM GTP (□), 50 μM GTP (●), 100 μM GTP (○), and 200 μM GTP (◆) or (D) no ATP (■), 25 μM ATP (□), 50 μM ATP (●), 100 μM ATP (○), and 200 μM ATP (◆)on GMPR activity and the NADPH K_m value was assessed in the presence of increasing GTP or ATP at a fixed GMP concentration. (E) The effect of IMP on the GMPR deamination GMP was evaluated in the presence (●) or absence (■) of IMP. (F) To evaluate the capacity of mycophenolic acid to inhibit LmGMPR activity, steady state kinetic analysis was performed using a fixed GMP concentration while varying NADPH in the presence of increasing concentrations of mycophenolic acid. Mycophenolic acid inhibition was analyzed using the Dixon plot.

Surprisingly, ATP had a marked impact on the LmGMPR kinetic parameters causing an ~10-fold increase in the GMP K_s value and a ~4-fold decrease in the V_max value in the presence of 1.0 mM ATP (Fig. 5B).

A similar analysis using saturating concentrations of GMP showed that GTP and ATP induced only a modest ~2-fold increase in the NADPH K_m value (Fig. 5C & 5D). Collectively, these data imply that ATP and GTP did not bind in the LmGMPR NAPDH binding pocket. Inhibition studies with IMP showed, not surprisingly, that this nucleotide was a competitive inhibitor of GMP and had a marked inhibition of the LmGMPR activity at 1.0 mM (Fig. 5E) which is in agreement with a K_i value of 14 μM previously reported by Spector and Jones (Spector & Jones, 1982). No production of NADPH or GMP was detected via high performance liquid chromatography or spectrophotometric assays in enzymatic reactions containing high concentrations of LmGMPR (400 nM), IMP (1.0 mM), NH₄⁺ (10 mM) and NADP⁺ (100 μM), even after a 20 h incubation at 25 °C, confirming that the back conversion of IMP to GMP was highly unfavorable.

Previous studies demonstrated that mycophenolic acid is a potent inhibitor of LdIMPDH that blocked Leishmania growth when parasites were grown in media containing hypoxanthine as the sole purine (Dobie et al., 2007, Wilson et al., 1991). In contrast, parasites supplemented with the either guanine or xanthine refractory to MPA and exhibited robust growth (Fig. S2). Since LmGMPR also contains a number of conserved residues that have been implicated in the MPA binding domain (Fig. 2) (Sintchak et al., 1996, Prosise & Luecke, 2003, Gan et al., 2002), we investigated the capacity of MPA to inhibit LmGMPR. Kinetic analysis of LmGMPR in the presence of increasing levels of MPA resulted in a set of parallel lines on a Hanes-Woolf plot; a pattern consistent with MPA being a competitive inhibitor of NADPH (Fig. S3). Dixon plot analysis resulted in a K_i of ~ 20 μM for MPA (Fig. 5F).

Binding of nucleotides to LmGMPR

LmGMPR contains a single tryptophan residue (Trp121) centrally located in the CBS domain (Fig. 2). Changes in the intrinsic fluorescence intensity and emission λ_max of Trp121 were used to monitor the binding of GMP, GTP, ATP, or NADPH to LmGMPR. The fluorescence spectrum of exhaustively dialyzed LmGMPR revealed, that despite containing a single tryptophan, two emission maxima centered at 335 and 350 nm were observed (Fig. 6A). This pattern suggested that LmGMPR subunits may adopt conformations with Trp121 differentially exposed to the aqueous environment (Lakowicz, 2006). Addition of high concentrations of GMP caused a diminution in fluorescence intensity and a single λ_max centered at 335 nm. These changes suggest that GMP binding induced a shift in Trp121 to a more nonpolar environment (Fig. 6A). Similarly, addition of GTP to LmGMPR resulted in an emission spectrum with a single λ_max at 335 nm. However, the fluorescence quenching was less pronounced when compared to GMP (Fig. 6A). The binding of ATP resulted in an emission spectrum with a λ_max centered at 350 nm (Fig. 6A) indicative of a conformational change that position Trp121 into a more solvent exposed environment.

Fig. 6.

Analysis of nucleotide binding to full length LmGMPR or the LmGMPR CBS domain. (A) The intrinsic fluorescence emission spectra of LmGMPR (26 μM) was recorded at an excitation of 290 nm in the absence or presence of GMP, GTP, or ATP. (B) Binding of GMP, GTP and ATP to LmGMPR in the absence of the second substrate NADPH was measured by monitoring changes in fluorescence intensity at 335 nm. (C) The binding of the nucleotides to GMP, GTP, ATP, and IMP to the LmGMPR CBS domain was measured by monitoring changes in fluorescence intensity at 335 nm. (D) The binding of NADPH to LmGMPR, in the absence of GMP, was measured by titrating with NADPH monitoring changes in λ_max NADPH (inset) at an excitation wavelength of 340 nm. (E) For fluorescence energy transfer experiments, solutions containing NADPH or a mixture of LmGMPR + NADPH were excited at 290 or 340 nm and the emission spectra recorded from 300–550 nm.

Titration of LmGMPR with GMP induced a biphasic fluorescence intensity change (Fig. 6B). A pattern that is suggestive of two sites with differential binding affinities as illustrated by the binding model (Prinz, 2009),

$$\mathrm{E}+\mathrm{L}\stackrel{{\mathrm{K}}_{\mathrm{d}1}}{\leftrightarrow }\text{EL}+\mathrm{L}\stackrel{{\mathrm{K}}_{\mathrm{d}2}}{\leftrightarrow }\text{LEL}$$

where E and L correspond to the concentration of LmGMPR and GMP, respectively and K_d1 and K_d2 are the two dissociation constants for GMP. K_d for GMP were calculated by curve fitting the fluorescence intensity data to equation 2 (see Experimental Procedures). Dissociation constants of 20 ± 2 μM (K_d1) and 261 ± 38 μM (K_d2) were calculated for GMP. A similar titration performed with GTP exhibited in a comparable biphasic response and fitting the fluorescence data to equation 2 resulted in two K_d1 and K_d2 values of 16 ± 1 μM and 369 ± 18 μM, respectively (Fig. 6B & Table II). It is possible that the two affinities observed with GMP and GTP may also arise from the binding of these nucleotides to LmGMPR subunits with different conformational states. In contrast, addition of ATP to LmGMPR caused a decrease in fluorescence intensity at the λ_max of 335 nm and fitting the data to the Hill equation resulted in a K_d value of 87 ± 24 μM (Fig. 6B).

Table II.

Nucleotide binding to LmGMPR

Ligand	Kd1 (μM)	Kd2 (μM)
GMP	20 ± 2	261 ± 38
GTP	16 ± 1	369 ± 18
ATP	87 ± 24	-

To validate the hypothesis that the 131 amino acid fragment encompassing the LmGMPR CBS domain (His₆-gmpr92-222) formed a nucleotide binding domain this fragment was expressed and nucleotide binding was monitoring changes in the intrinsic fluorescence intensity of Trp121. In contrast to the full length LmGMPR, His₆-gmpr92-222 exhibited a single phase binding isotherm with K_d values of 188 ± 19 and 333 ± 27 μM for GTP and GMP, respectively (Fig. 6C). Binding studies with ATP and IMP showed that these nucleotides associated with His₆-gmpr92-222; however the interaction affinities were notably lower and saturation binding was not attained even at a concentration of 3.0 mM (Fig. 6C). The higher K_d values observed with the CBS domain are likely due the increased conformation flexibility of this smaller polypeptide domain. Similar findings have been previously reported for the human IMPDH2, chloride channel CLC2, and AMP-activated protein kinase γ-subunit CBS domains (Scott et al., 2004, Labesse et al., 2013, Hardie, 2003).

NADPH binding to LmGMPR

The binding of NADPH to LmGMPR in the absence of GMP was assessed using the extrinsic fluorescence signal of NADPH. In aqueous solutions NADPH has an emission λ_max centered at 464 nm at an excitation wavelength of 340 nm. However, on binding LmGMPR, the emission λ_max shifted to 436 nm and was accompanied by an increase in the fluorescence intensity (Fig. 6D, inset). Titration of LmGMPR with NADPH caused a shift in the emission λ_max that was proportional to the NADPH concentration. Fitting the fluorescence data to equation 3 (see Experimental Procedures) resulted in a K_d value of 27 ± 4 μM for NADPH (Fig. 6D).

NAD⁺ has been reported to bind in the CBS domains (Baykov et al., 2011). Here we examined NADPH binding to the LmGMPR CBS domain using Forster Resonance Energy Transfer (FRET) experiment, a distance-dependent fluorescence technique that permits energy transfer between a donor (Trp121) and an acceptor (NADPH) if these two moieties are in close proximity. Excitation of NADPH free in solution at wavelength of 290 nm showed no significant fluorescence between 300–550 nm; whereas excitation at 340 nm resulted in a diagnostic NADPH emission λ_max at 464 nm (Fig. 6E). In contrast, when the LmGMPR:NADPH complex was excited at 290 nm it showed a strong fluorescence signal at 335 nm arising from Trp121. However, no notable emission for the bound NADPH was detected at ~400–470 nm. The absence of a fluorescence resonance energy transfer signal suggests that unlike GMP, GTP, or ATP, NADPH binds to a site that is distal to Trp121 the CBS domain (Fig. 6E).

Nucleotide binding to the LdIMPDH

Multiple protein sequence alignments revealed that LdIMPDH also contained a CBS domain (Fig. 2). Ligand binding to the full length LdIMPDH was assessed by monitoring changes in intrinsic fluorescence of tyrosine and tryptophan following the addition of GMP or GTP (Fig. 7A). Plotting the fluorescence change as a function of nucleotide concentration produced a single phase binding isotherm that, upon curve fitting, produced K_d values of 66 ± 7, and 67 ± 5 μM for GMP and GTP, respectively (Fig. 7A). Interestingly, the K_d value measured for GMP was notably lower than the inhibition constant (K_i) of ~200 μM previously reported for this nucleotide (Dobie et al., 2007). Similar measurements performed with ATP and IMP showed that these ligands had lower binding affinities for LdIMPDH with K_d values of 130 ± 11 and 243 ± 31 μM, respectively (Fig. 7A). Titrations performed with hypoxanthine showed that this nucleobase exhibited no significant binding to LdIMPDH even at concentration of 2.0 mM (Fig. 7A).

Fig. 7.

Analysis of nucleotide binding to LdIMPDH. (A) Full length LdIMPDH or (B) the LdIMPDH CBS domain (His₆-ldimpdh112-223) were titrated with hypoxanthine (Hyp), XMP, IMP, GTP, GMP, or ATP and ligand binding was monitored by changes in intrinsic fluorescence at an emission wavelength of 307 nm.

To evaluate ligand binding capacity of the LdIMPDH CBS domain, a fragment encompassing residues 112-233 (His₆-impdh112-223) was expressed and the binding assessed using intrinsic fluorescence changes. As with the full length LdIMPDH, the CBS domain showed comparable K_d values of 92 ± 6, and 89 ± 11 for GMP and GTP, respectively (Fig. 7B). The K_d value for ATP binding to the LdIMPDH CBS domain was 148 ± 19 μM, a value similar to that measured for the full length LdIMPDH. In contrast, no saturable binding was observed with IMP or hypoxanthine (Fig. 7B). These data are consistent with the LdIMPDH CBS domain binding both guanylate and adenylate nucleotides.

Nucleotide binding alters LmGMPR and LdIMPDH quaternary structure

To assess if the potential allosteric responses observed with LmGMPR in the presence of GTP or high levels of GMP were associated with quaternary structure changes, rate zonal centrifugation experiments were performed. LmGMPR in the absence of nucleotides migrated as two populations with molecular masses of 52 and 220 kDa, sizes that correspond to monomeric and tetrameric forms, respectively (Fig. 8A). This mixed population was consistent with the fluorescence experiments showing, that in the absence of nucleotide ligands, Trp121 exhibited two emission λ_max values (Fig. 6A). Addition of saturating amounts of GMP or GTP stabilized the formation of the LmGMPR tetramer (Fig. 8A). Surprisingly, in the presence of saturating ATP concentrations, LmGMPR migrated as two populations with apparent molecular masses of 52 and 120 kDa, masses that correspond to monomeric and dimeric forms of the enzyme, respectively (Fig. 8A). These data indicate that the binding of guanylate and adenylate nucleotides is sufficient to induce quaternary structure changes that influence LmGMPR catalytic activity.

Fig. 8.

Structural analysis of LdIMPDH and LmGMPR. The quaternary structure of (A) LmGMPR or (B) LdIMPDH in the absence or presence of GMP, GTP, or ATP was determined by rate zonal centrifugation on a glycerol gradient. Gradients were fractionated and the distribution of LmGMPR or LdIMPDH was visualized on a Coomassie Blue stained SDS-PAGE gel. Glycerol gradients were calibrated using a protein mixture containing bovine serum albumin, bovine IgG, and catalase.

Rate zonal ultracentrifugation analysis of exhaustively dialyzed LdIMPDH showed that in the absence of ligands LdIMPDH migrated as two populations with molecular masses of ~120 and 240 kDa, respectively, implying that LdIMPDH forms dimeric and tetrameric structures. However, in the presence of saturating amounts GMP, AMP, or ATP, LdIMPDH migrated as a ~240 kDa tetrameric structure (Fig. 8B).

Discussion

Leishmania express a battery of purine interconverting enzymes that include HGPRT, XPRT and the enzyme adenine aminohydrolase which are vital for salvage and growth of Leishmania on any single purine nucleobase (Fig. 1) (Boitz et al., 2012). Central to purine metabolism are the Leishmania IMPDH and GMPR which play a critical role in shunting GMP to AMP via an IMP intermediate (Jardim et al., 1999, Boitz et al., 2013, Boitz & Ullman, 2006a). We postulate that GMPR has a role in maintaining an optimal balance in guanylate and adenylate pools; particularly with the salvage of xanthine from the extracellular environment. Recent genetic studies have shown that L. donovani Δgmpr null mutant parasites failed to proliferate when cultivated in media containing guanine or xanthine as the sole purine source (Boitz, et al, manuscript in preparation).

Kinetic studies showed that LmGMPR catalyzes an NADPH-dependent reductive deamination of GMP to IMP, a reaction that involves the GMPR-XMP* covalent intermediate which involves the formation of a thioether with active site residue Cys³¹⁹ (Fig. 2), a linkage that is cleaved hydride transfer to IMP (Martinelli et al., 2011). The nucleotide specificity of LmGMPR was verified in vivo by showing that LmGMPR complemented the growth of the E. coli ΔguaC mutant, but not ΔguaB mutant, indicating that the Leishmania enzyme exhibited strict GMPR activity. In addition, we demonstrated that catalysis of the back reaction by LmGMPR was unfavorable as no detectable levels of GMP were observed in the presence of IMP, NADP⁺, and NH₄Cl. This contrasts to experiments with the E. coli GMPR which showed that the stable E-XMP* covalent intermediate was cleaved in the presence of ammonium ions to generate GMP (Patton et al., 2011). Although IMP bound to LmGMPR, it is unlikely to function as a competitive inhibitor in vivo since the intracellular levels of IMP are low (LaFon et al., 1982).

Interestingly, Leishmania promastigote growth in media containing guanine or xanthine was highly refractory to MPA, a response that was attributed to a weak association of this drug with the Leishmania GMPR (Dobie et al., 2007) and the direct salvage of these nucleobases through HGPRT and XPRT (Fig. 1). In contrast the majority of eukaryotic IMPDH enzymes, including human and Leishmania, form an extremely tight interaction with MPA (K_i values of 4–20 nM). Two key residues identified as critical for stabilizing contacts with MPA are Arg³²² and Gln⁴⁴1 (human IMPDH1 numbering) (Gan et al., 2002, Sintchak et al., 1996, Prosise & Luecke, 2003) which are situated on loops with the motifs GLR³²²VGM and AQ⁴⁴¹GVSG, respectively (Fig. 2). In the Leishmania GMPR, Borrelia, Tritrichomonas, and most bacterial IMPDH enzymes, the corresponding loops have a lysine and glutamine, respectively, at these key locations which accounts for the low binding affinity of MPA (K_i of ~20 μM). This hypothesis is strengthened by the studies of Gan et al (Gan et al., 2002) which showed that incorporating a Lys to Arg and Glu to Gln substitutions into the Tritrichomonas IMPDH increased the MPA binding affinity ~20-fold. This increased resistance to MPA inhibition, together with phylogentic analysis, makes it tempting to speculate that the Leishmania GMPR may have arisen as a result of a gene transfer of an ancestral prokaryotic IMPDH.

Structural analysis of phylogenetically diverse IMPDH and GMPR proteins suggest that this family of homologous enzymes predominantly form homomeric structures arranged in a square geometry that places the catalytic site near the subunit interfaces. This orientation allows amino acids from an adjacent subunit to contribute critical residues required for substrate binding (Colby et al., 1999, Hedstrom, 2012, Labesse et al., 2013, Patton et al., 2011, Zhang et al., 1999, Nimmesgern et al., 1996, Makowska-Grzyska et al., 2015). In the case of IMPDH, these inter-subunit contacts accelerate catalysis by forming interactions with a potassium ion which facilitates closure of a flap that mediates cleavage of the E-XMP* covalent intermediate (Riera et al., 2011, Hedstrom, 2012). Crystal structures of the human GMPR revealed that the binding of NADPH is stabilized by contacts between the 2′-PO₄ and the sidechains of Ser²², Arg²⁷, and the 6-NH₂ moiety of adenine, and the sidechains of Ser³¹⁴, Thr³¹⁷, and Tyr³¹⁸; interactions that are donated by a neighboring subunit (Patton et al., 2011). Since many of the analogous residues (Fig. 2, Ser³⁰, Arg³¹, Ser⁴⁴⁷, Ser⁴⁵⁰, and Tyr⁴⁵¹) are conserved in the LmGMPR, it is tempting to speculate that a similar inter-subunit network stabilizes NADPH binding. This may explain why the tetrameric form of LmGMPR exhibits increased catalytic activity.

A distinguishing feature of both the Leishmania GMPR and IMPDH enzymes is the presence of a CBS domain (Fig. 2). The absence of this domain in the human and E. coli GMPRs suggests that it is not required for enzymatic activity (Patton et al., 2011, Li et al., 2006, Deng et al., 2002). Surprisingly, removal of the LmGMPR CBS domain resulted in a mutant enzyme (lmgmprΔCBS) that failed to complement the guaC mutation in E. coli. Why the Leishmania GMPR, in contrast to the mammalian and bacterial counterparts, has retained a CBS domain is not clear. It is possible that this regulatory element may be associated with the localization of the LdIMPDH and LdGMPR to the glycosome (Hassan & Coombs, 1985, Fulwiler et al., 2011).

Previous biochemical studies of CBS domains suggested that this regulatory element selectively binds the adenylate ligands; AMP, ADP, ATP, S-adenosylmethionine and NADH (Lucas et al., 2010, Scott et al., 2004). In contrast, the LmGMPR and LdIMPDH or fragments encompassing the CBS domains, displayed a broader specificity and bind GTP, GMP, and ATP; with guanylate ligands exhibiting a higher binding affinity. The differential binding of GTP and ATP would provide a mechanism for rapidly regulating the activities of the Leishmania GMPR and IMPDH which would be required to maintain optimal intracellular adenylate and guanylate levels, regardless of the extracellular purine bases salvaged by the parasite. We propose that the CBS domain plays a role in regulating Leishmania GMPR wherein accumulating GTP levels triggers activation of GMPR and formation of a tetrameric complex which accelerates the conversion of GMP to IMP. In contrast, increased levels of ATP favor the formation of monomeric/dimeric structures that have reduced catalytic activity. This differential regulation allows Leishmania to grow on any single purine and maintain an optimal adenylate:guanylate balance (LaFon et al., 1982). This would be important particularly in guanine or xanthine rich environments, such as those that might be encountered in the macrophage phagolysosome (Barankiewicz & Cohen, 1985), where these nucleobases would be salvaged primarily through HGPRT and XPRT (LaFon et al., 1982) and incorporated into the guanylate nucleotide pools.

Structural analyses have shown that despite sharing limited sequence homology (15–40%), CBS repeats retain a conserved core β-α-β-β-α secondary structure that allows the CBS domain to adopt a clamshell-like shape with a ligand binding cleft between the CBS repeats (Baykov et al., 2011, Scott et al., 2004). Multiple sequence alignments of CBS domains showed that Asp150 and Asp212 (Leishmania GMPR numbering) are strictly conserved in the CBS repeats of phylogenetically diverse organisms (Fig. 9). These negatively charged residues have been implicated in the formation of contacts with hydroxyl groups of the ribose moiety and stabilizing nucleotide (Baykov et al., 2011, Scott et al., 2004, Labesse et al., 2013, Lucas et al., 2010). Point mutations at the corresponding residues in the human IMPDH are known to not only disrupt ATP binding but have also been linked to the hereditary disease retinitis pigmentosa (Labesse et al., 2013, Bowne et al., 2002). Although the LmGMPR and LmIMPDH CBS domains show differential binding of ATP, GMP, and GTP, the molecular basis for this discrimination is unknown. Several structures for CBS domains containing bound adenylate ligands indicate that the nucleobase ring is stabilized by a series of base stacking interactions with nonpolar residues that line cleft that is formed between two CBS repeats (Lucas et al., 2010). Fluorescence spectroscopy experiments indicated that GTP and ATP induce conformational changes in the LmGMPR CBS domain, as demonstrated by alteration in the exposure of Trp121 to the aqueous environment.

Fig. 9.

Multiple protein sequence alignment of CBS motifs. CBS domain from Methanocaldococcus jannaschii AMP activated kinase γ (MjAMPK_g), Homo sapiens chloride channel CLC5 (HsCLc5), Homo sapiens AMP activated protein kinase γ1 (HsAMPK_g1), Homo sapiens AMP activated protein kinase γ2 (HsAMPK_g2), Leishmania donovani IMPDH (LdIMPDH), Homo sapiens IMPDH-II (HsIMPDHII), Leishmania major GMPR (LmGMPR) were aligned using the CLUSTAL X program. In the consensus sequence, semiconserved residues are designated by lower case letters and strictly conserved residues are shown as upper case letters.

Experimental Procedures

Chemicals and reagents

All restriction endonucleases and buffers were purchased from Invitrogen. Mycophenolic acid (MPA) was purchased from MP Biomedicals. All purines and and nucleotide monophosphate and nucleotide triphosphates were obtained as the sodium salt form from Sigma-Aldrich (Oakville, ON). Other reagents were of the highest quality commercially available. E. coli H1174 and H724 strains were obtained from the E. coli Genetic Stock Center (CGSC, Yale University, New Haven, CT).

Cloning and expression of the LmGMPR and LdIMPDH

The LmGMPR ORF was amplified by PCR from L. major genomic DNA with the primer pair, 5′-GAATTCATTAAAGCATGGCAGCCCTAGGCAGTC-3′ and 5′-GTTCGAATTACAGCTTCGAGATATCGTG-3′, containing EcoRI and HindIII restriction endonuclease sites (underlined), respectively, using Pfx DNA polymerase (Invitrogen) and 30 cycles of denaturation at 94 °C for 30 s, annealing at 52 °C for 30 s, and extension at 72 °C for 90 s. The PCR fragment was cloned into the EcoRI and HindIII sites of the pQE80 vector (Qiagen, Toronto, ON) to generate the pQE80-LmGMPR expression construct. The integrity of this construct was confirmed by DNA sequence analysis. The coding sequence reported in the TriTryp genome database for the LmGMPR encompasses an N-terminal extension of ~53 extra amino acids which is absent from other trypanosome and Leishmania GMPRs. Therefore this N-terminal extension was omitted from the construct used to produce recombinant LmGMPR.

For expression of the LmGMPR, H1174 (F-, thr-20, guaC23, fhuA2, proA35, lacY1, tsx-70, glnV44(AS), gal-6, LAM-, trpC45, his-68, tyrA2, rpsL125(strR), malT1(LamR), xyl-7, mtlA2, thi-1, purH57, ilv-635) E. coli, a strain lacking a functional GMPR, was transformed with pQE80-LmGMPR and grown in Luria Broth (LB) containing 100 μg/ml ampicillin at 37 °C to an OD₆₀₀ of 0.7, and protein expression was induced with 0.7 mM isopropyl-β-D-thiogalactopyranoside (IPTG) at 20 °C for 5 h. Bacterial cultures were harvested and the cell pellet was resuspended in 20 ml of 20 mM Tris pH 8.0 then lysed with two passes through a French press. Clarified lysates were made up to 20% saturation with (NH₄)₂SO₄ and incubated for 1 h at 0 °C. Precipitated proteins were removed by centrifugation and the supernatant was made up to 35% saturation with (NH₄)₂SO₄ and incubated at 0 °C for 1 h to precipitate the LmGMPR. The protein pellet was dissolved in 25 ml of 50 mM Tris pH 7.6 with 2 mM EDTA and 10 mM β-mercaptoethanol (TEM) and applied to a Cibracron blue Sephadex column (1.5 × 2 cm) (Sigma-Aldrich, Oakville, ON) equilibrated with TEM. The column was washed with 50 ml of 0.2 M NaCl in TEM and LmGMPR was eluted with a 0.3 – 1.0 M NaCl step gradient in TEM. Fractions containing a 52 kDa protein were pooled and precipitated with 35% (NH₄)₂SO₄. Precipitated LmGMPR was dissolved in a minimal volume of TEM and dialyzed against 2 l of 75 mM sodium phosphate buffer pH 6.9, 1 mM DTT, 1 mM EDTA (PDE) at 4 °C (Fig. S1). The concentration of the purified protein was determined spectrophotometrically using a ε_280nm of 9970 M⁻¹cm⁻¹ (Gill & von Hippel, 1989). To insure that LmGMPR preparation very low or non-detectable levels of NADPH oxidase activity, control experiments were performed using reaction mixtures containing saturating levels of NADPH in the absence of GMP. LmGMPR stored at −80 °C in PDE showed no loss in enzymatic activity after 2 months. LdIMPDH was expressed in the E. coli strain H724, and recombinant LdIMPDH protein was purified as previously described (Dobie et al., 2007).

The 131 amino acid fragment encoding the LmGMPR CBS domain (residues 92-222) was amplified from pQE80-LmGMPR template DNA using the primer pair; 5′-GGTCATATGGTGCGCAAGGTGAAGC-3′ and 5′-ACTGGCAACCTCAACGCCAC-3′, harboring NdeI and BamHI restriction sites (underlined), respectively, and the Pfx DNA polymerase (Invitrogen) after 30 cycles of denaturation at 94 °C for 30 s, annealing at 60 °C for 30 s, and extension at 72 °C for 90 s. The PCR product encoding the CBS fragment was cloned into the NdeI/BamHI sites of the pET15b vector to generate the expression construct pET15b-gmpr92-222. E. coli strain ER2566 (New England Biolabs, Ipswich, MA) transformed with pET15b-ldgmpr92-222 was grown to 0.6 OD₆₀₀ at 37 °C, and His₆-ldgmpr92-222 was induced with 0.25 mM IPTG for 5 h at 20 °C. Bacteria pellets were resuspended in PBS, lysed by French press, and His₆-ldgmpr92-222 purified on Ni²⁺-NTA affinity resin using standard protocols (Qiagen, Toronto, ON). The bound protein was eluted with an 80-400 mM imidazole step gradient in PBS containing 10 mM β–mercaptoethanol, concentrated using a Millipore 3K MWCO centrifugal filter unit, and dialyzed against 1.0 l of PBS, 10 mM β-ME.

A DNA fragment encoding the 122 amino acid of the LdIMPDH CBS domain (residues 112-233) was PCR-amplified using Pfx DNA polymerase (Invitrogen, Burlington, ON), the sense primer 5′–GGAATTCCATATGGAGCGGCAAGTGGAGATG–3′ containing an NdeI restriction site (underlined), and the antisense primer 5′–CGGGGATCCTTACTTGTCCAGTGTGCTGTGC–3′ containing a BamHI restriction site (underlined) from L. donovani genomic DNA as the template with 30 cycles of denaturation at 94 °C for 30 s, annealing at 60 °C for 30 s, and extension at 72 °C for 90 s. The PCR product was cloned into the corresponding restriction sites of the pET15b vector to generate the pET15b-ldimpdh112-233 expression construct and transformed into E. coli strain ER2566 (New England Biolabs). Bacterial cultures were grown to an OD₆₀₀ of 0.6 at 37 °C in LB medium containing 100 μg/ml ampicillin then shifted to 20 °C and protein expression was induced overnight with 0.7 mM IPTG. Cell pellets were resuspended in 20 ml of PBS containing EDTA-free protease inhibitor mixture and lysed by two passes through a French press. Clarified lysate was loaded onto a Ni²⁺-NTA column, washed extensively with PBS, and the bound His₆-impdh112-233 polypeptide was eluted using an 80–400 mM imidazole step gradient in PBS containing 10 mM β-ME, concentrated using a Millipore 3K MWCO centrifugal filter unit (EMD Millipore, Etibcoke, ON) and dialyzed against 1.0 l of PBS, 10 mM β-ME. The cloning and overexpression of the full length LdIMPDH in E. coli and the purification of the recombinant protein has been previously reported (Dobie et al., 2007).

An lmgmpr mutant lacking the CBS domain that encompasses residues 90-217 (lmgmprΔCBS) was generated by PCR mutagenesis using pQE80-LmGMPR as a template, the primer pair; 5′-GACCATCGCGCACTGTTCCTC-3′ and 5′-GTGGCGTTGAGGTTGCCAGT-3′, and Pfx DNA polymerase (Life Technologies) with 20 cycles of denaturation at 94 °C for 60 s, annealing at 64 °C for 60 s, and extension at 68 °C for 14 min. The pQE80-lmgmprΔCBS construct was confirmed by DNA sequence analysis and then used to transform E. coli strain H1174 for complementation analysis.

Complementation of ΔguaC lesion in E. coli by LmGMPR

The E. coli strain H724 (F-, fhuA2, lacY1, tsx-70, glnV44(AS), gal-6, LAM-, e14-, hns-1, trpC45, his-68, rfbC1, purC50, guaB24, tyrA2, Δ(gltB-gltF)500, rpsL125(strR), malT1(LamR), xyl-7, mtlA2, thi-1) which is deficient in IMPDH activity (guaB24) (Nijkamp & De Haan, 1967) and E. coli strain H1174 (thr-20,guaC23, fhuA2::IS2, proA35, lacY1, tsx-70, glnX44(AS), gal-6, λ, trpC45, his-68, tyrA2, rpsL125(strR), malT1(λ^R), xyl-7, mtlA2, thiE1, purH57, ilv-635) which is deficient in bacterial GMPR (guaC) were transformed with pQE80-LmGMPR or pQE80-lmgmprΔCBS and grown on Luria Broth (LB) plates containing 50 μg/ml ampicillin and 50 μg/ml streptomycin sulfate. A colony of each transformant was resuspended in 100 μl of M9 medium supplemented with 100 μM tyrosine, tryptophan, proline, glutamine, threonine, and histidine, 0.2% glucose, 5 μg/ml nicotinic acid, 5 μg/ml thiamine, 0.5 mM IPTG, 50 μg/ml ampicillin, and 50 μg/ml streptomycin and a 20 μl aliquot of each bacterial suspension added to 3.0 ml of this medium containing either no purine, or 100 μM adenine, guanine, or hypoxanthine. Cultures were grown at 37 °C with vigorous shaking for 16 h and the culture density (OD₆₀₀) measured. These experiments were performed in triplicate.

Steady-state kinetic analysis of LmGMPR

Steady-state kinetics for LmGMPR were performed at 25 °C on a temperature-controlled Beckman-Coulter DU640 spectrophotometer in 75 mM sodium phosphate pH 6.9, 100 mM KCl, 1 mM DTT, 1 mM EDTA (PKDE buffer) containing 200 μM NADPH while varying the GMP concentration (1–800 μM) to determine the K_m value for GMP. To determine the K_m value for NADPH, the GMP concentration was fixed at 300 μM and the NADPH concentration varied (5–200 μM). Reactions were initiated with 46 nM LmGMPR, and the oxidation of NADPH was monitored at 340 nm. The initial reaction rates were calculated using the extinction coefficient 6200 M⁻¹cm⁻¹ for NADPH and the initial velocity data was analyzed by fitting the data to the Hill equation (equation 1);

$$v={\mathrm{V}}_{max}({[\text{GMP}]}^{\mathrm{n}}/({{\mathrm{K}}_{\mathrm{s}}}^{\mathrm{n}}+{[\text{GMP}]}^{\mathrm{n}})$$ (eqn 1)

where v is the initial velocity, V_max is the maximal velocity, n is the Hill coefficient, and K_s is the dissociation constant for GMP.

To examine the effect of GTP and ATP on LmGMPR catalysis, the kinetic parameters for GMP and NADPH were measured in reactions containing 10–1000 μM concentrations of each of the two nucleotide triphosphates. MPA inhibition of LmGMPR was determined using a fixed GMP concentration of 100 μM, while varying the NADPH concentration (10–200 μM), in the presence of increasing levels of MPA (0–50 μM).

GMPR and IMPDH ligand binding analysis

Dissociation constants (K_d) were determined by titrating LmGMPR (26 μM) in PKDE buffer against 0–1000 μM GMP, GTP, ATP, or NADPH and monitoring changes in the intrinsic fluorescence intensity and emission λ_max of tryptophan 121 (Trp121) for LmGMPR using an excitation wavelength of 290 nm and recording the emission spectra from 305–450 nm. K_d values for GMP and GTP in the absence of NADPH were calculated by curve fitting the fluorescence intensity data with the Origins 7.0 graphing software to a biphasic curve described by equation 2,

$$\mathrm{\Delta }\mathrm{F}={\mathrm{F}}_{min}+((\mathrm{\Delta }{\mathrm{F}}_1\ast {\mathrm{L}}^{\text{nH}1})/{{\mathrm{K}}_{\mathrm{d}1}}^{\text{nH}1}+{\mathrm{L}}^{\text{nH}1})+((\mathrm{\Delta }{\mathrm{F}}_2\ast {\mathrm{L}}^{\text{nH}2})/{{\mathrm{K}}_{\mathrm{d}2}}^{\text{nH}2}+{\mathrm{L}}^{\text{nH}2})$$ (eqn 2)

where ΔF is the observed change in fluorescence intensity, F_min is the fluorescence minima between the two binding phases of the curves, ΔF1 and ΔF2 are changes in fluorescence induced by GMP binding to the first and second sites, L is the GMP concentration, n_H1 and n_H2 are the Hill coefficients for the slopes corresponding to the two components of the biphasic curve, and K_d1 and K_d2 are GMP or GTP concentrations that produce 50% saturation of sites 1 and 2, respectively. To control for dilution effects, LmGMPR was titrated with buffer and changes in fluorescence intensity measured. The correction for the inner filter effect was performed graphically as previously described (Mertens & Kagi, 1979). Ligand binding to the LmGMPR CBS domain was monitored as a function of changes in intrinsic fluorescence using 35 μM His₆-lmgmpr92-222 in PBS with 10 mM β-ME. The K_d value for NADPH was calculated by titrating LmGMPR (26 μM) with NADPH (0–100 μM) or buffer, and monitoring changes in the extrinsic fluorescence of NADPH bound to LmGMPR. Fluorescence resonance energy transfer experiments were performed with mixtures containing NADPH, LmGMPR, or an equimolar NADPH/LmGMPR mixture (26 μM) at an excitation wavelength of 290 or 340 nm and recording the emission spectra from 300–550 nm. The formation of the LmGMPR:NADPH complex can be described by equation 3;

$$\mathrm{\Delta }{\mathrm{\lambda }}_{\text{obs}}/\mathrm{\Delta }{\mathrm{\lambda }}_{\text{tot}}=[\mathrm{L}]/({\mathrm{K}}_{\mathrm{d}}+[\mathrm{L}])$$ (eqn 3)

where Δλ_obs is the observed change in the emission wavelength at a given concentration of NADPH and Δλ_tot is the total change in the emission λ_max.

Ligand binding to the LdIMPDH CBS domain (His₆-ldimpdh112-233) was performed as described for the full length LdIMPDH (Dobie et al., 2007). LdIMPDH, rather than L. major IMDPH, was used for the ligand binding studies since the former enzyme had been previously cloned and biochemically characterized (Dobie et al., 2007). It should be noted that IMPDH polypeptides from these two Leishmania species are ~98 % identical. Nucleotide binding to the LdIMPDH was monitored as a function of changes in intrinsic fluorescence at an excitation wavelength of 278 nm and recording the emission spectra from 290–400 nm. LdIMPDH (20 μM) was titrated with 0–2.0 mM ATP, GMP, GTP, IMP, XMP or the nucleobase hypoxanthine. Ligand-dependent fluorescence quenching was determined at an emission wavelength of 305 nm. All fluorescence experiments were performed on a Cary Eclipse using an excitation and emission slit width of 5 nm. The correction factor for the inner filter effect arising from the nonspecific fluorescence quenching associated with the UV/Vis absorption by the nucleotide ligands was determined graphically as previously described (Mertens & Kagi, 1979).

Quaternary structure of the LmGMPR and LdIMPDH

The quaternary structure of LmGMPR and LdIMPDH in the absence or presence of either 1.0 mM GMP, GTP, or ATP was determined by rate zonal centrifugation on a 5–40% linear glycerol gradient in PKDE buffer. Gradients were overlaid with 160 μg of LmGMPR in 140 μl of PKDE and sedimentation was performed by centrifugation at 34,000 rpm on an SW-41 rotor for 16 h at 4 °C. Changes in sedimentation rates were calibrated using 50 μg of bovine hemoglobin as an internal standard in the LmGMPR and LdIMPDH binding reactions. Gradients were fractionated (0.5 ml fractions) and proteins precipitated with trichloroacetic acid prior to analysis by SDS-PAGE. Protein masses were determined using a protein mixture containing 100 μg each of catalase (245 kDa), bovine IgG (160 kDa), bovine serum albumin (66 kDa) and bovine hemoglobin (64 kDa) as standards.

Multiple sequence alignments

Protein sequences for the Leishmania donovani, Homo sapiens, E. coli, Tritrichomonas foetus IMPDHs and the L. major, E. coli, and Homo sapiens GMPRs were obtained from the National Center for Biotechnology Information, and multiple protein sequence alignments were performed using the ClustalX program.

The abbreviations used are

GMPR

guanosine 5′-monophosphate reductase

IMPDH

inosine 5′-monophosphate dehydrogenase

CBS

cystathionine-β-synthase

LmGMPR

L. major GMPR

LdIMPDH

L. donovani IMPDH

HGPRT

hypoxanthine-guanine phosphoribosylatransferase

XPRT

xanthine phosphoribosyltransferase

APRT

adenine phosphoribosyltransferase

AK

adenosine kinase

MPA

mycophenolic acid

IPTG

isopropyl-β-D-thiogalactopyranoside

TEM

Tris-EDTA-β-mercaptoethanol

PDE

sodium phosphate-DTT-EDTA

PKDE

sodium phosphate-KCl-DTT-EDTA