First Crystal Structures of Mycobacterium tuberculosis 6-Oxopurine

Article pubs.acs.org/jmc

First Crystal Structures of Mycobacterium tuberculosis 6‑Oxopurine Phosphoribosyltransferase: Complexes with GMP and Pyrophosphate and with Acyclic Nucleoside Phosphonates Whose Prodrugs Have Antituberculosis Activity Wai Soon Eng,† Dana Hocková,*,‡ Petr Špaček,‡ Zlatko Janeba,‡ Nicholas P. West,† Kyra Woods,† Lieve M. J. Naesens,§ Dianne T. Keough,† and Luke W. Guddat*,† †

The School of Chemistry and Molecular Biosciences, The University of Queensland, Brisbane, 4072 QLD Australia Institute of Organic Chemistry and Biochemistry, Academy of Sciences of the Czech Republic, v.v.i. Flemingovo nam. 2, CZ-166 10 Prague 6, Czech Republic § Rega Institute for Medical Research, KU Leuven − University of Leuven, Minderbroedersstraat 10, B-3000 Leuven, Belgium ‡

S Supporting Information *

ABSTRACT: Human tuberculosis is a chronic infectious disease affecting millions of lives. Because of emerging resistance to current medications, new therapeutic drugs are needed. One potential new target is hypoxanthine-guanine phosphoribosyltransferase (MtHGPRT), a key enzyme of the purine salvage pathway. Here, newly synthesized acyclic nucleoside phosphonates (ANPs) have been shown to be competitive inhibitors of MtHGPRT with Ki values as low as 0.69 μM. Prodrugs of these compounds arrest the growth of a virulent strain of M. tuberculosis with MIC50 values as low as 4.5 μM and possess low cytotoxicity in mammalian cells (CC50 values as high as >300 μM). In addition, the first crystal structures of MtHGPRT (2.03−2.76 Å resolution) have been determined, three of these in complex with novel ANPs and one with GMP and pyrophosphate. These data provide a solid foundation for the further development of ANPs as selective inhibitors of MtHGPRT and as antituberculosis agents.

■

INTRODUCTION Each year, almost 9 million new cases of human tuberculosis (TB) are diagnosed, and 1.5 million deaths are reported worldwide.1 The etiological agent causing TB is Mycobacterium tuberculosis (Mt).1 Therapeutic drugs for this disease include rifampicin, isoniazid, pyrazinamide, and ethambutol, but such treatment regimes can last for up to nine months or longer and produce serious side-effects.2 Thus, together with the emergence of multi-, extensively-, and totally drug resistant M. tuberculosis strains, the discovery of novel anti-TB drug targets is of increasing importance in combating this disease. Griffin et al. have used random transposon mutagenesis to identify the genes essential for growth of M. tuberculosis.3 This study showed that hpt, which codes for hypoxanthine-guanine phosphoribosyltransferase (MtHGPRT), is one such gene, suggesting that the enzyme could be a chemotherapeutic target for M. tuberculosis infections. In addition, we, and others, have also proposed that MtHGPRT is a target for anti-TB drug discovery.4 MtHGPRT belongs to a family of enzymes known as the 6-oxopurine phosphoribosyltransferases (PRTases) that synthesize the purine nucleoside monophosphates essential for DNA/RNA production. These enzymes are found in most organisms including humans, protozoan parasites, and bacteria. © XXXX American Chemical Society

In many of these organisms and, in particular, protozoan parasites, 6-oxopurine PRTase activity has been deemed essential for their survival.4,5 However, in humans, while a germline mutation in the hpt gene can lead to Lesch-Nyhan syndrome or gouty arthritis, somatic mutations allow the continued survival of the mutant cell.6 Thus, partial inhibition of the human enzyme by a therapeutic drug for TB is likely to be well tolerated. The reaction catalyzed by the 6-oxopurine PRTases is shown in Figure 1. The presence of a divalent metal ion (e.g., Mg2+) is required for the reaction to proceed. For the human enzyme, the reaction is ordered with 5-phospho-α-D-ribosyl-1-pyrophosphate (PRib-PP) binding first, followed by the purine base. Catalysis then occurs and pyrophosphate (PPi) dissociates from the complex followed by the nucleoside monophosphate in the rate-limiting step.7 For the orthologous enzymes from Tritrichomanas foetus,8 Schistosoma mansoni,9 and Trypanosoma cruzi,10 the mechanism is also sequential. Acyclic nucleoside phosphonates (ANPs) are a class of compound originally developed as antiviral agents that target Received: April 21, 2015

A

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

structure of MtHGPRT and provide a platform for the rational design of more potent and selective inhibitors.

■

RESULTS AND DISCUSSION Kinetic Constants for Guanine, Hypoxanthine, PRibPP, GMP, and IMP and Comparison with Those of the 6Oxopurine PRTases from Other Organisms. Purified recombinant MtHGPRT has a specific activity for guanine and hypoxanthine of 1.6 and 1.4 μmol min−1 mg−1, respectively. These values are lower than those for human HGPRT7 (34 and 19 μmol min−1 mg−1 of protein, respectively), E. coli HPRT14 (177 μmol min−1 mg−1, hypoxanthine as base) and E. coli XGPRT14 (114 μmol min−1 mg−1, xanthine as base). However, they are comparable to those of the parasite enzymes. Pf HGXPRT15 has a specific activity of 1.5 and 0.77 μmol min−1 mg−1 for guanine and hypoxanthine, respectively while PvHGPRT16 has a specific activity of 5.2 and 2.1 μmol min−1 mg−1 for guanine and hypoxanthine, respectively. Thus, in this regard, MtHGPRT is more similar to the parasite 6-oxopurine PRTases than to the human or E. coli enzymes. The Km of guanine for MtHGPRT is 3.4 μM, ∼2-fold lower than for hypoxanthine (Table 1); xanthine is not a substrate.

Figure 1. Reaction catalyzed by the 6-oxopurine phosphoribosyltransferases (PRTases). The purine bases are guanine (R is −NH2), hypoxanthine (R is −H), and xanthine (R is −OH).

HIV reverse transcriptase and viral DNA polymerases. These ANPs have adenine (e.g., adefovir and tenofovir) or cytosine (e.g., cidofovir) as the base and contain a phosphonate group which has a C−P bond that is chemically stable in vivo.11 We hypothesized that by replacing the pyrimidine or 6-amino purine base with either guanine or hypoxanthine and by modification of the acyclic moiety new ANPs would inhibit the 6-oxopurine PRTases.12 This was found to be the case and such ANPs have been shown to be potent inhibitors of human, Plasmodium falciparum (Pf), P. vivax (Pv), and the two Escherichia coli 6-oxopurine PRTases.4b,5 In most instances the Ki values for the ANPs against enzymes from different organisms are not the equivalent. For example, 8-hydroxy-9-(2(phosphonomethyl)ethyl)guanine has a Ki of 1.2 μM for Pf HGXPRT, but 68 μM for human HGPRT.12 Another class of acyclic phosphonates, the acyclic immucillin phosphonates (AIP), are also inhibitors of human HGPRT and Pf HGXPRT but are up to 592-fold more selective for the Pf enzyme compared to the human counterpart.13 These data suggest that suitable derivatives of the ANPs can be designed as potent inhibitors of MtHGPRT and can also possess selectivity over the human enzyme. For the ANPs to exhibit antimicrobial activity, the negative charges on the phosphonate groups have to be masked to increase cell permeability. This can be accomplished by attaching hydrophobic groups to the phosphonic oxygen atoms by an ester or phosphoramidate bond that, on entry into the cell, can be cleaved to produce the active parent compound. Four such prodrugs of the ANPs have been shown to arrest the growth of a nonvirulent strain of M. tuberculosis in culture.4b With this in mind, active recombinant MtHGPRT was expressed and purified to homogeneity with the goal being to determine if, and how, ANPs inhibited MtHGPRT. Prodrugs of new ANPs were then synthesized to determine the MIC50 values for M. tuberculosis grown in culture. Although assessment of selectivity through measurement of relative Ki values for the M. tuberculosis and human enzyme is one criterion to assess whether a compound is a suitable drug lead it should not be considered as the sole arbiter. A more informative indicator is that potential candidates must possess low cytotoxicity in human cells; thus, the CC50 values for the prodrugs were measured. To begin to develop these compounds into inhibitors that would specifically target MtHGPRT, crystal structures of MtHGPRT were determined. One of these is in complex with GMP and PPi, the two products of the reaction (Figure 1) and the others are in complex with three different ANPs. These ANP inhibitors are specifically designed to mimic the substrates/products of the reaction to limit the ability of the pathogen to mutate and develop resistance. These data provide the first insights into the

Table 1. Kinetic Constants for Recombinant MtHGPRT substrate guanine PRib-PP hypoxanthine PRib-PP

kcat (s−1)

Km (μM)

kcat/Km (μM−1 s−1)

± ± ± ±

4.4 ± 0.4 465 ± 15 8.3 ± 0.7 1443 ± 105

0.1409 0.0012 0.0663 0.0006

0.62 0.55 0.55 0.86

0.02 0.01 0.01 0.03

Human HGPRT also binds guanine with a 2-fold preference compared with hypoxanthine though the Km values are slightly lower (guanine 1.9 μM; hypoxanthine 3.4 μM) than for MtHGPRT.15 MtHGPRT has kcat/Km values of 0.1 μM−1 s−1 for both substrates, making it less efficient than other 6oxopurine PRTases including human HGPRT (kcat/Km values of 4.3 μM −1 s −1 for guanine and 1.5 μM −1 s−1 for hypoxanthine),15 E. coli HPRT (4.9 μM−1 s−1 for hypoxanthine),14 E. coli XGPRT (6.5 μM−1 s−1 for guanine),14 PvHGPRT (0.8 μM−1 s−1 for hypoxanthine; 0.9 μM−1 s−1, guanine),16 or Pf HGXPRT (4.3 μM−1 s−1 for hypoxanthine and 80 μM−1 s−1 for guanine).15 The Km of PRib-PP for MtHGPRT is 3-fold higher when hypoxanthine is the base compared to that when the base is guanine (Table 1). This data agrees with Biazus et al., who also showed that the Km for PRib-PP changes depending on the identity of the purine base.17 This appears to be a unique property of MtHGPRT as the reported Km values for other 6oxopurine PRTases for PRib-PP is mainly independent of the second substrate.7,8,15,18 One possible explanation for this phenomenon is that PRib-PP could bind in a slightly different orientation depending on the identity of the base. The Km for PRib-PP is also significantly higher than for human 6-oxopurine PRTases (65 μM, guanine as base) demonstrating another unique property of this enzyme.7,19 The Ki values for GMP and IMP are 20 μM and 170 μM, respectively, in agreement with the fact that the Km for guanine is lower than for hypoxanthine. These Ki values are significantly higher than those for human HGPRT (5.8 μM, GMP and 5.4 μM, IMP) but lower than those for E. coli HPRT (526 μM, GMP and 247 μM, IMP).12,14 Considering these kinetic constants, MtHGPRT has properties that are different from any B

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Scheme 1. Synthesis of the Double Branched ANPs and Their Prodrugs

Scheme 2. Synthesis of the Double Branched ANPs with Three Phosphonate Groups and Their Prodrugs

C

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Table 2. Comparison of the Ki Values of the ANPs for MtHGPRT and Human HGPRTa

a

Data are given as Ki value ± SD

other 6-oxopurine PRTase though it does have individual characteristics in common with one or other members of this family of enzymes. This data suggests that it is likely that inhibitors selective for MtHGPRT can be made. Chemical Synthesis of ANPs. The synthesis of double branched aza-acyclic nucleoside phosphonate (aza-ANP) bis-

phosphonates 9−12 was based on stepwise construction of the acyclic moiety on the N9-position of a purine base (Scheme 1). At first, selective NH2-alkylation of N-(2-hydroxyalkyl)ethylenediamine with 9-(2-bromoethyl)guanine20a or 9-(2bromoethyl) hypoxanthine afforded compounds 15−17. In the next step, aza-Michael addition of diethyl vinylphosphonate D

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 2. (A) Comparison of the three-dimensional structures of the active site of MtHGPRT in complex with 5 and human HGPRT in complex with 3 (Table 2) and the amino acid sequence alignment of these two enzymes. The colored regions highlight corresponding regions in the two enzymes. The differences in the identity of the amino acid residues in and near the active site of MtHGPRT and human HGPRT are indicated by a “Δ” in the sequence alignment. Comparisons of the 5′-phosphate binding sites, purine ring binding sites and pyrophosphate binding site are shown in B, C, and D, respectively.

tetrahydrofuran to decompose the triphenylphosphoranylidene intermediate rising from the presence of the free amino group. Resulting phosphonates 26 and 27 were transformed to 6oxopurine derivatives 28 and 29 by nucleophilic aromatic substitution in acidic conditions (75% aqueous trifluoroacetic acid). The above prepared key aza-ANP ethyl esters 18, 19, 21, 24, 28, and 29 were used for the synthesis of the target aza-ANPbased HGPRT inhibitors 9−14 and their prodrugs 30−33 (Schemes 1 and 2). To form free phosphonic acids 9−14, all ethylester groups were simultaneously cleaved under standard conditions using Me3SiBr/acetonitrile followed by hydrolysis. These aza-ANPs are polar molecules with very limited ability to cross cell membranes. To improve permeability, we decided to synthesize phosphoramidate prodrugs because they are already established as ANP-based antivirals and nontoxic amino acids are released when they are cleaved.20c A sufficiently lipophilic ethyl ester of phenylalanine was selected and the prodrugs 30−

to the both NH-groups in the main scaffold formed the key bisphosphonate esters 18, 19, and 21. In the case of derivative 16, containing hypoxanthine as a nucleobase, further addition on purine occurred and trisphosphonate 20 was isolated as a main product together with desired bisphosphonate 19. To prevent this side product formation in the synthesis of the elongated hypoxanthine derivative 24, 6-benzyloxy-protected20b starting compound 22 was used and the benzyl group of 23 was cleaved after aza-Michael addition. Synthesis of the second type of double branched aza-ANPs with three phosphonate groups was based on a different approach (Scheme 2).5a The whole acyclic moiety was prearranged by aza-Michael reaction using an excess of diethyl vinylphosphonate with N-(2-hydroxyethyl)ethylenediamine and the resulting trisphosphonate 25 was introduced to the N9-position of 6-chloropurine and 2-amino-6-chloropurine via Mitsunobu reaction. In the case of 2-amino-6-chloropurine, the Mitsunobu reaction had to be followed by heating in water/ E

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Table 3. Data Collection and Refinement Statistics for the MtHGPRT in Complex with GMP.PPi and Compounds 5, 9, and 14 GMP.PPi unit cell parameters length a, b, c (Å) 56.4, 85.6, 156.6 angle α, β, γ (deg) 90, 90, 90 space group P212121 diffraction dataa temp (K) 100 resolution range (Å) 47.11−2.76 observations 85818 (12234) unique reflections 20023 (2799) completeness (%) 99.4 (97.4) Rmergeb 0.050 (0.237) Rpimc 0.039 (0.186) ⟨I⟩/⟨σ(I)⟩ 12.3 (3.4) refinement resolution limits (Å) 47.11−2.76 Rwork 0.2250 (0.3491) Rfree 0.2841 (0.4051) rmsd bond lengths (Å) 0.003 rmsd angles (deg) 0.737 clashscored 8.6 components of the asymmetric unit GMP.PPi/inhibitors 4 solvent (water) 44 Mg2+ 8 Ramachandran plot (%) favored 93.3 outliers 0.64

5

9

14

94.7, 94.7, 335.5 90, 90, 120 P65

53.4, 85.7, 88.1 90, 98.77, 90 P21

55.2, 85.5, 152.7 90, 90, 90 P212121

100 47.35−2.57 456856 (38001) 53656 (4590) 99.8 (98.6) 0.074 (0.772) 0.040 (0.421) 14.9 (2.4)

100 48.56−2.03 283004 (19101) 50489 (3527) 99.5 (94.3) 0.118 (0.672) 0.082 (0.476) 10.3 (2.1)

100 46.34−2.32 227675 (20999) 32001 (2982) 99.5 (95.6) 0.111 (0.873) 0.066 (0.524) 9.9 (2.1)

47.35−2.57 0.2329 (0.2576) 0.2983 (0.3186) 0.003 0.607 4.46

48.56−2.03 0.1867 (0.2632) 0.2261 (0.2937) 0.003 0.657 3.7

46.34−2.32 0.1898 (0.2775) 0.2473 (0.3441) 0.007 0.770 3.4

6 84 8

4 552 8

4 237 8

96.23 0.20

97.82 0.00

97.19 0.15

The values in parentheses are for the outer resolution shell. bRmerge = ΣhklΣi|Ii(hkl) − (I(hkl))|/ΣhklΣiIi(hkl). cRp.i.m = Σhkl[1/([N(hkl) − 1])]1/2Σi| Ii(hkl) − (I(hkl))|ΣhklΣiIi(hkl), where Ii(hkl) is the observed intensity and is the average intensity obtained from multiple observations of symmetryrelated reflections. dClashscore is defined as the number of bad overlaps ≥0.4 Å per 1000 atoms. a

33 were prepared by a recently published method.20d In the first step of the one-pot reaction sequence, the cleavage of the phosphonate ethyl esters with Me3 SiBr forms in situ trimethylsilyl esters. In the second step, the reaction with ethyl (L)-phenylalanine in the presence of 2,2′-dithiodipyridine (Aldrithiol) and triphenylphosphine yield the corresponding tetra-amidates 30−31 (Scheme 1) or hexa-amidates 32 and 33 (Scheme 2). Previously known ANPs, 9-[2-(2-phosphonoethoxy)ethyl]guanine (1), 9-[2-(2-phosphonoethoxy)ethyl]hypoxanthine (2), bisphosphonates 3−8 (Table 2), and their prodrugs 34− 37 (Table 4), were prepared according to our recently published procedures.5b,c Inhibition of MtHGPRT by ANPs. Fourteen ANPs were investigated as inhibitors of MtHGPRT and their Ki values measured (Table 2). These compounds are classified into three groups (A, B, and C) depending on whether they have one, two, or three phosphonate moieties attached to the chemical structure. Within each of these groups, the compounds have either guanine or hypoxanthine as the base. In group B, the ANPs have a second phosphonate attached to either the second carbon atom from N9 or at the third atom (via a nitrogen) from the N9 atom in the purine ring. The most potent inhibitors in groups A and B are 3 and 5 (Ki values of 0.69 and 0.77 μM, respectively) with both of these containing guanine as the base (Table 2). These Ki values are ∼28-fold lower than that for GMP (Ki = 20 μM) showing that these modifications increase the affinity for MtHGPRT. Within these two groups, compounds containing guanine as the base

have lower K i values compared to those containing hypoxanthine, though the selectivity ratio (hypoxanthine analogue/guanine analogue) is variable (1.4 to >70). The lowest selectivity with respect to the purine base is for 11 with 12 and 13 with 14 (Table 2). For the other five compounds, the ratio in favor of guanine over hypoxanthine varies from 10 to 70-fold. Thus, it appears that the composition of each of the three moieties (the purine base, the phosphonate groups or the hydroxyl group) contributes to the differences in affinity for these compounds for MtHGPRT. The preference for inhibitors containing guanine is in agreement with the fact that the Ki for GMP is 8.5-fold lower than that for IMP. Comparison of the Affinities of the ANPs for Human HGPRT and MtHGPRT. For 12 of the ANP inhibitors examined here, the Ki values are lower for the human enzyme than for MtHGPRT with ratios varying from 5 to 400. In only two cases, compounds 1 and 5 (when guanine is the base) are the Ki values the same for both enzymes. These inhibitors were designed based on the crystal structures of human HGPRT in complex with ANPs as the structure of MtHGPRT was unknown at the time. Thus, it is reasonable that they should exhibit lower Ki values for this enzyme compared with MtHGPRT. However, with the availability of the crystal structure of MtHGPRT (see later), inhibitors designed to specifically target this enzyme will now be feasible. This is the next step in generating ANPs with higher potency and selectivity for the M. tuberculosis enzyme. As would be expected for enzymes that catalyze the same reaction, a comparison of the amino acid sequence shows that many of the active site F

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 3. Comparison of the tetrameric arrangement for MtHGPRT and human HGPRT. (Left) MtHGPRT in complex with GMP and PPi and (right) human HGPRT in complex with GMP (PDB: 1HMP). GMP and PPi are shown as colored spheres. The large mobile loops are in red. In the MtHGPRT structure, these loops (residues 90−106) are located between the subunits whereas in the human HGPRT structure (residues 101−117), the loops are at the extremities of the tetramer.

gondii22 enzymes. However, it is similar to that observed for the bacterial enzymes, E. coli HPRT14 and T. tengcongensis HGPRT.23 The tetrameric units of MtHGPRT and human HGPRT do have a common structural feature, and this is the association of two of the four subunits with subunits A and B constituting one dimer and subunits C and D the second. Therefore, equivalent dimer interfaces exist between A and B and C and D (Figure 3). Thus, the major difference between the structure of these two enzymes lies in the way in which these two dimer pairs then associate to form the tetramer. The critical outcome of this difference in arrangement of the two dimers in the enzymes is the positioning of the large mobile loop (residues 90−106 in MtHGPRT; 101−117 in human HGPRT; Figure 3). This loop is proposed to close over the active site during the catalytic reaction in the other 6oxopurine PRTases.13 The two crystal structures that demonstrate this phenomenon are human HGPRT and Pf HGXPRT in complex with the transition state analogues, (1S)-1-(9-deazaguanin-9-yl)-1,4-dideoxy-1,4-imino-D-ribitol 5phosphate (ImmGP) or (1S)-1-(9-deazahypoxanthine-9-yl)1,4-dideoxy-1,4-imino-D-ribitol 5-phosphate (ImmHP), respectively, PPi and Mg2+.13 In all the published crystal structures of human HGPRT, the loop is located on the outside of the tetramer and is exposed to solvent; there are no intersubunit interactions that could hinder its movement.5c,21 The flexibility of this loop is demonstrated by the fact that only in complex with the transition state analogs is electron density for all the residues in this loop visible. In contrast, the structure of MtHGPRT in complex with GMP.PPi and with two of the ANPs (5 and 14) shows that the loop is partially buried between the subunits (Figure 3). At the tetrameric interface of MtHGPRT, the main chain and side chain atoms in the mobile loop from subunit A or B (residues 92−110) form hydrogen bonds and hydrophobic interactions with the main chain and side chain atoms in the mobile loop from the adjacent subunit (subunit D or C; Figure S1). Thus, in the tetrameric structure, the highly conserved Y93 in the mobile loop of one subunit is within hydrogen bonding distance of R141 and of the C-terminal residues of the mobile loop present in the adjacent subunit (residues 109−110) forming an antiparallel arrangement (Figure S1). Thus, for the mobile

residues are conserved (Figure 2A). However, the wide disparity in the kinetic constants for the reaction catalyzed by these enzymes together with the Ki values for the inhibitors shows that differences exist that can be exploited for drug design. With hypoxanthine as the base, all the compounds examined have a lower Ki value for the human enzyme compared with those for MtHGPRT (5 to 100-fold). When guanine is the base, this ratio varies from 1 to 400 fold. However, the contribution of the base is different for each particular ANP. For example, for compound 5, the Ki values are the same with guanine as the base but for the same ANP with hypoxanthine as the base, the variability is >12.5. For compound 11, with guanine as the base, the selectivity in favor of human HGPRT is 400-fold but this drops to only 14 when hypoxanthine replaces guanine. Thus, it can be deduced that the nature of the purine base is one of the important factors that can affect selectivity between these two enzymes. The addition of a second phosphonate group (compounds 3−14) does not appear to be as critical in decreasing the Ki value for the M. tuberculosis enzyme (0.69−15 μM; guanine as base) as it is for the human enzyme (0.008−1 μM; guanine as base). The fact that the Ki values are generally higher for the M. tuberculosis enzyme is no doubt due to the Km for PRib-PP being much higher for MtHGPRT than for the human enzyme. This difference is 8-fold when guanine is the base and 24-fold when hypoxanthine is the base. Crystal structures of MtHGPRT were then obtained to provide a rationale for these differences and to provide a platform for improvement for the design of specific MtHGPRT inhibitors. Crystal Structures of M. tuberculosis HGPRT. Four crystal structures of MtHGPRT, one in complex with GMP, PPi, and Mg2+ and three in complex with different ANPs, have been determined. Two of these ANPs are from group B (5 and 9) and one is from group C (14) (Table 2). These structures have been determined at 2.76 (GMP.PPi.Mg2+), 2.57 (5), 2.03 (9), and 2.32 Å (14) resolution (Table 3). In three of the structures, the enzyme crystallized as a tetramer, but in the complex with 9, only dimers were observed. Comparison of the Tetrameric Structure of MtHGPRT with Other 6-Oxopurine PRTases. The arrangement of the four subunits in MtHGPRT (Figure 3, left) is different from that of the human (Figure 3, right),21 P. falciparum13b and T. G

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

loop to close over the active site these intersubunit interactions have to be broken (Figure S1). The reason why MtHGPRT has evolved to form this particular arrangement of subunits is not clear. Given the critical role that this mobile loop plays in protecting the active site from bulk solvent during catalysis, one possible explanation for its location in the M. tuberculosis enzyme is that it could be a protective mechanism against proteolytic attack within the cell. For example, in E. coli XGPRT, the loop can be cleaved by inherent proteases in the cell.4b The importance of maintaining the native structure of this loop is illustrated by a mutagenesis experiment where six residues in the mobile loop of E. coli XGPRT were removed, resulting in an 80-fold decrease in catalytic efficiency.4b Structure of the GMP.PPi Complex Demonstrating the Movement of Amino Acid Residues Surrounding the Active Site as These Two Products Bind. A number of crystal structures of the 6-oxopurine PRTase in complex with a nucleoside monophosphate are available24 but there is only one reported structure in complex with the two products of the reaction.14,21 This is Toxoplasma gondii HGXPRT in complex with XMP and PPi.22 The structure shows both compounds bound in the active site in what appears to be their location immediately prior to catalysis in the reverse reaction (Figure 1). The second 6-oxopurine PRTase structure which contains the nucleoside monophosphate and PPi in the active site is that of MtHGPRT. The difference compared to the T. gondii structure is that GMP has yet to reach its optimum location for catalysis to occur and, so, this structure shows instead GMP entering (or leaving) the active site, exemplifying the changes in the enzyme structure, and/or GMP itself, that have to occur for this to happen (Figure 4A). To obtain the MtHGPRT structure, only GMP was added to the enzyme solution before crystallization. The PPi in the crystal structure has arisen from the Mg2+dependent hydrolysis of PRib-PP which is present in the storage buffers of MtHGPRT.25 An obvious difference between the structures of T. gondii HGXPRT and MtHGPRT is the conformation of the nucleoside monophosphates. In MtHGPRT, there is a ∼57° rotation in the torsion angle around the N9 of the purine base and the C1′ carbon of the ribosyl ring compared to its angle in the T. gondii HGXPRT structure (Figure 4A). In subunits A and C of the MtHGPRT structure, the rotation of the ribose ring results in the formation of a hydrogen bond between the hydroxyl group attached to the C2′ carbon and a phosphate oxygen of PPi. The positioning of the ribose ring results in the 5′-phosphate group of GMP not being found in its predicted location (Figure 2A) but, rather, it points out of the active site (Figure 4A). Role for K66, G67, and R188. PPi is held in place by hydrogen bonds between the phosphoryl oxygen atoms to one or both backbone amides of K66 and G67. A second phosphoryl oxygen forms a hydrogen bond to the side chain of R188 and this oxygen is also coordinated to a magnesium ion (Figure 4B). The peptide bond between L65 and K66 is in the cis conformation as has been observed in other 6-oxopurine PRTases when the product, PPi, is also bound in the active site.13 In subunit D, the side chain of K66 has started to rotate toward the center of the active site, an orientation it usually adopts in the absence of PPi (Figure 4B).26 Role for F175. This side chain forms a π-stacking arrangement with the purine ring in all the 6-oxopurine

Figure 4. (A) Active site of MtHGPRT in complex with GMP and PPi (yellow), with the superposition of XMP and PPi (green) as observed in the Toxoplasma gondii HGXPRT complex.22 The binding mode of GMP differs such that, in MtHGPRT, the 5′-phosphate of GMP is pointing out of the active site instead of binding in the predicted 5′phosphate binding pocket (residues 126−130; blue). In both structures, the purine bases are positioned in the purine binding pocket (pink). PPi binds in the expected PPi binding site (cyan). Magnesium ions are shown as spheres. Yellow, MtHGPRT; green, Toxoplasma gondii HGXPRT. (B) Stereoview of GMP, PPi, and magnesium binding to the four subunits of MtHGPRT. The spheres are magnesium ions. These occupy slightly different places in each subunit.

PRTases.4b,5c,13,14,21 However, in subunit D of the MtHGPRT complex, this side chain is turned away from the active site allowing GMP to exit or enter. Thus, this side chain is flexible and rotates away from the active site to allow the entry/exit of the purine base of the nucleoside monophosphate (Figure 4B). Role for K154. The side chain of this amino acid forms a hydrogen bond with the 6-oxo group in the purine ring and is one of the reasons for the specificity of these enzymes compared with the adenine PRTases.4b,5c,13,14,21 However, it is only in subunits A and B that a hydrogen bond is formed (2.5 Å in subunit A and 3.2 Å in subunit B). In subunit D, it has commenced to move away (4 Å) and, in subunit C, there is no electron density for this side chain. Therefore, this side chain is either moving into position to form this essential bond (entering) or it is moving away (exiting), breaking the interaction with the purine base (exiting; Figure 4B). Role for D126. The carboxylate of this amino acid forms a hydrogen bond with the N7 atom of the purine ring when the transition state analog binds to human HGPRT.13 However, when GMP is bound in the active site of human HGPRT, it is too far away (3.9 Å). In the MtHGPRT complex with GMP, the side chain of D126 is rotated away from the purine ring in subunits A and B (3.6 and 4.1 Å, respectively), while in subunit C, there is no electron density for D126. In subunit D, the side chain of D126 has moved into the 5′-phosphate pocket forming a hydrogen bond with the main chain nitrogen of G128; the H

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 5. Fo-Fc electron density map and structure of the ANPs together with the Connolly surface of MtHGPRT. (A) 5, (B) 9, and (C) 14.

purine ring has rotated by 13° relative to its position in the other subunits (Figure 4B). Thus, comparison of the structures of each of the subunits in the MtHGPRT.GMP.PPi complex illustrates the movement of the amino acid residues surrounding the active site that have to occur to allow GMP and PPi to enter or leave. This structure demonstrates the flexibility of the enzyme and suggests possible movements of these flexible loops that could occur when the inhibitors bind in the active site. Crystal Structure of MtHGPRT in Complex with the ANPs, Compounds 5, 9, and 14. Crystal structures have been determined for MtHGPRT in complex with three different ANPs (Table 3). In all of the subunits, there is strong electron density for the inhibitors identifying unequivocally their location in the active site (Figure 5). Data collection and refinement statistics are shown in Table 3. Though the three ANPs have different chemical structures, their Ki values are quite similar but with a trend of 5 < 14 < 9 (Table 2). Based on the comparison of the amino acid sequence (Figure 2A) and the crystal structures of other 6-oxopurine PRTases, the purine base and the phosphonate group (mimicking the 5′-phosphate group of GMP/IMP) are located in their predicted binding pockets (Figure 5).4b,5c The hydrogen bonds that hold the purine base in position are identical in all the subunits in the complexes with 9 and 14 (except for the interaction of the exocyclic amino group with the carbonyl of D182; compound 9). However, they differ in each of the subunits of the complex with 5 showing that this base can adopt slightly different positions in the active site (Figure 6). The common hydrogen bonds in all three structures are between the NZ atom of K154 and the exocyclic 6-oxo group and between the carbonyl of V176 with the N1 of the purine ring (Figure 6). It would be expected that, if the chemical structures of two ANPs are identical with the only difference being the base, then the inhibitors should bind in exactly the same place in the active site as the only difference is the interaction of the exocyclic amino group with the carbonyl of D182. If this were the case, though, there would only be one extra hydrogen bond (guanine compared with hypoxanthine) and this is unlikely to be sufficient to account for the wide disparity in Ki values when 5 and 7 (guanine) are compared with 6 and 8 (hypoxanthine) (>50 fold between the two bases; Table 2). Thus, it is possible that the purine base (guanine or hypoxanthine) could bind differently in these ANPs. This slight change in orientation could thus influence the location of these ANPs in the active site and their interactions with the amino acid residues that surround them.

Figure 6. Stereoview of the active site of MtHGPRT in complex with three ANP inhibitors. (A) 5, subunit B (B) 9, subunit A (C) 14, subunit D. Spheres represent Mg2+ ions (green) and waters (red).

Compounds 13 and 14, with guanine and hypoxanthine respectively as the base have similar Ki values for MtHGPRT suggesting that, in this instance, these two compounds bind in the same position and the only difference is the extra hydrogen bond(s) due to the exocyclic amino group (Figure 6). This does not affect their Ki as predicted. It is, therefore, deduced that the third phosphonate group in these compounds determines how these compounds bind and this would result in hypoxanthine (14) being found in a similar position as guanine. The purine ring of hypoxanthine in 14 forms the same interactions with active site atoms as does guanine in 5 (cf. Figure 6B,C) suggesting this hypothesis is plausible. There are five amino acid residues that surround the predicted 5′-phosphate binding pocket in MtHGPRT. These are residues D126 to T130, which is a highly conserved region in the amino acid sequence of the 6-oxopurine PRTases (Figure I

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

movement of flexible loops surrounding the active site, together with that of adjacent flexible loops that affects the kinetic constants (Tables 1 and 2). Here we analyze the differences in the structures and amino acid composition in the large mobile loop and three key binding pockets of the active site. To date, there are no structures available for the same ANP bound to human HGPRT and MtHGRPT. Thus, to perform this study, we have superimposed the crystal structure of MtHGPRT in complex with 5 with that of human HGPRT in complex with 3. The difference between these two ANPs is only one extra carbon atom in 5 (Table 2). Mobile Loop (Figure 2A). One major difference between these two structures is the arrangement of the subunits which causes the amino and carboxy terminal ends of one of the longest of the flexible loops (residues 90−106) to be buried at the interface between subunits in MtHGPRT while in human HGPRT it is exposed to solvent (this is described in detail earlier). Within this loop only 4 out of 16 amino acid residues are identical. Among the variations in the sequences is the replacement of K102 (human HGPRT) by S91 (MtHGPRT). In the human enzyme, the side chain of K102 is linked to the side chain of E196 by hydrogen bonds via water. This may help stabilize the flexible loop in human HGPRT, whereas this is not possible in MtHGPRT as it contains serine in this position. 5′-Phosphate Binding Pocket (Figure 2B). One of the common features between these two structures is the hydrogen bond network between the OD2 atom of D126 (D137 in human) and an invariant water molecule which forms a bifurcated hydrogen bond to the ND2 atom of N173 (OD2 atom of D184 in human) and the hydroxyl of T156 (T167 in human). When ImmGP binds to human HGPRT, the side chain of R169 replaces this invariant water molecule. The OD2 atom of D137 instead forms a hydrogen bond directly to the NH1 atom of R169 which then forms a hydrogen bond to the OD1 atom of D184. This network helps to anchor the essential aspartic acid residue (D137) in place for catalysis to occur as it now forms a hydrogen bond to the N7 atom of the purine ring. In MtHGPRT, the arginine residue is replaced by V158 so this bond (with D126) could not be formed. Thus, this critical residue (D126) may not be able to make such a stable bond (as in human HGPRT) with the N7 of the purine base, and this could partially account for the much lower kcat value for MtHGPRT compared with human HGPRT. Formation of this bond is essential for catalysis to occur.7,13a The amino acids that surround the 5′-phosphate binding pocket in MtHGPRT are DSGLT (126−130) compared with DTGKT (137−141) in human. There is another loop surrounding this loop whose role appears to be to keep the 5′-phosphate binding loop anchored in position. This loop constitutes P155 to I164 in MtHGPRT and R166 to P175 in human HGPRT. In MtHGPRT, there are three hydrogen bonds between the carbonyl groups of the primary loop and those amino acids in the loop that encircles it. These are D126 to the amide of A157, G129 and the amide of S132, and T130 to the amide of L134. In human HGPRT, however, the 5′-phosphate binding loop is held more firmly in position by five hydrogen bonds to this second loop: T138 to amide of V171, G139 to the amide of T144 and the hydroxyl of T144, T141, and the amide of L145. The overall 3D structure of this second loop differs in the two enzymes. In MtHGPRT, it forms a partial α-helix while in human HGPRT it is a random coil. These differences must therefore contribute to the affinity of PRib-PP, nucleotide product and ANP binding.

2A). The hydrogen bonds between the phosphonyl oxygens of the ANPs and the main chain atoms in MtHGPRT are with the amide nitrogens of D126, S127, G128, and T130; there are also two hydrogen bonds with the hydroxyl side chains of T130 and S127. However, for each of the three structures, and even within the four subunits of a particular structure, there is a slight variation in this network. The structure (Cα and side chain atoms) of the five amino acid residues is identical, so these small variations are only due to the location of the phosphonyl oxygens of 5, 9, and 14. Though these interactions are important in conferring affinity, it is unlikely that they alone contribute to the small differences in the Ki values for these three ANP inhibitors. In the structure of MtHGPRT in complex with these ANPs, one of the oxygen atoms attached to the phosphorus atom in the second phosphonate group is coordinated to a magnesium ion which in turn is coordinated to the carboxylate of D182 (Figure 6). This mimics that found when PPi, together with the transition state analog, is bound to either human HGPRT or Pf HGPRT.13 Thus, this second phosphonate group is located in the site where one of phosphate groups of PPi would be expected to bind. Only in 14, and then only in subunit D, is the ANP close enough to form interactions with the large mobile loop, and this is a hydrogen bond between the hydroxyl atom of S92 and one of the phosphonyl oxygen atoms present in the third phosphonate group. However, this interaction is not the same as when the transition state binds to human HGPRT or Pf HGXPRT, as it is the amide of the conserved serine residue that forms a hydrogen bond to the terminal oxygen of PPi.13 The hydroxyl group of 9 is flexible and can occupy one, two or three positions within a particular subunit (A, three; B, two; C, two; and D, one). In the primary conformation (the occupancy for this site ranges from 56 to 100% in each of the four subunits), it forms hydrogen bond with a water molecule that bridges to the carbonyl of L65 and the amide of V90. An additional hydrogen bond via this water to the carbonyl of V90 is observed in three subunits B, C and D. However, in subunits A and B, an alternative conformation is also observed where the hydroxyl forms a hydrogen bond with the side chain of S92. Thus, this attachment may assist in moving the mobile loop toward the active site. Thus, in all three complexes, there is a different network of hydrogen bonds holding these ANPs in the active site and orientating them in position. It is therefore proposed that the individual components of the ANPs contribute differently to affinity, and in some instances one or the other component plays a predominant role in binding. The interface areas between the ANPs and the enzyme are 360 Å2 (5), 439 Å2 (9) and 415 Å2 (14). Compound 5 has the smallest interface area but the lowest Ki. Therefore, the hydrogen bonding network would seem to be more influential on the binding than the hydrophobic forces. Comparison of the Amino Acid Sequence and 3D Structure of MtHGPRT and Human HGPRT in Complex with ANPs. There is 30% sequence identity (49% homology) between human HGPRT and MtHGPRT, with most of the key active site amino acids being identical (Figure 2A). This suggests that structural components besides the highly conserved amino acids contribute to the differences in kcat, Km, and Ki values for these enzymes. Human HGPRT exhibits great plasticity when substrates bind, catalysis occurs and products are released.5b It is our contention that it is the J

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Table 4. MIC50 and CC50 Values of the Prodrugs nucleobase: guanine

nucleobase: hypoxanthine

prodrug

parent ANP

MIC50a (μM)

CC50b (μM)

SIc

prodrug

parent ANP

MIC50a (μM)

CC50b (μM)

SIc

34 36 30 32

3 5 9 13

11 9 4.5 4.5

>3005c 107 ± 555c 193 ± 36 >300

>27 >11 >43 >67

35 37 31 33

4 6 10 14

7.5 15 8 4.5

101 ± 175c 90 ± 21 156 ± 28 28 ± 6

>13 >6 >19.5 >6

a Determined from virulent M. tuberculosis strain (H37Rv). bDetermined using human A549 lung carcinoma cells. cSI = Selectivity index, data are given as mean value ± SD. The structures of compounds 30, 31, 32, and 33 are given in Schemes 1 and 2. The attachments to the phosphonate groups are identical for prodrugs 34, 35, 36, and 37.5c

Purine Binding Site (Figure 2C). In the MtHGPRT complex, the 6-oxogroup of the purine ring forms a hydrogen bond to the NZ atom of K154 which in turn is locked into position by a hydrogen bond to the carbonyl of D174. In comparison, in the human complex, the 6-oxo group of the purine ring also forms an additional hydrogen bond, and this is with the carbonyl group of K185. This does not occur for MtHGPRT as the corresponding carbonyl of D174 is too far away (3.7 Å). This extra bond could be a factor that contributes to the decreased affinity of the ANPs for MtHGPRT. There is no difference in the structure surrounding the 2-amino group, and thus, both enzymes favor guanine over hypoxanthine. Further, the aromatic amino acid that forms a π-stacking arrangement with the purine ring is the same in both enzymes and moves over this ring to clamp it in position.5c Valine is at position 125 in MtHGPRT but the equivalent residue in the human enzyme is isoleucine. The side chains of these two residues project toward the purine ring. In MtHGPRT, the carbon atoms of this side chain are 5 Å away from the purine ring of the ANP while, in human HGPRT, they are only 4 Å away from the ANP. Thus, there is a subtle difference in the active sites at this location that could contribute to differences in affinity of the ANPs for the two enzymes. Pyrophosphate Binding Site (Figure 2D). One difference in the vicinity of the PPi binding site is R100 in human HGPRT which is replaced by A89 in MtHGPRT. In the human HGPRT complex, this side chain forms a hydrogen bond with the side chain of D119 (the C-terminal end of the large mobile loop) in the adjacent subunit. Though this aspartic acid residue is conserved in MtHGPRT (D108), the side chain cannot form a bond with A89. In the structure of Pf HGXPRT in complex with the acylic immucillin phosphonates, the side chain of the corresponding arginine residue (R112) forms a hydrogen bond with one of the phosphoryl oxygens of the pyrophosphate group. Thus, it can be deduced that this side chain can also play a part in the affinity of the ANPs and is another factor responsible for the higher Ki values with MtHGPRT. Thus, overall, there are a number of movements of mobile loops that surround the active site and these are not the same between the two enzymes accounting for differences in the kinetic constants and for the selectivity of the ANP inhibitors. Activity of Prodrugs against M. tuberculosis and Their Cytotoxicity in Human Cells. None of the tested parent ANPs possessed antimicrobial activity in the cell-based assays against virulent or avirulent strains of M. tuberculosis. This highlights the difficulty for antibiotics to penetrate the highly impermeable lipid rich cell wall of this organism. This natural barrier represents a significant challenge in the development of any anti-TB compound. To overcome this problem, phosphoramidates of the ANPs were synthesized to mask the charges of the phosphonate moieties20c,d and to facilitate the penetration

into the M. tuberculosis cells. These prodrugs showed anti-TB activity in a virulent strain of M. tuberculosis (H37Rv) with three most active having MIC50 values of 4.5 μM (Table 4). Their cytotoxicity was low with a selectivity index (SI) [CC50/ MIC50] in favor of anti-TB activity, with the best being 32, with an SI > 67 (Table 4).

■

CONCLUSION The crystal structure of MtHGPRT has been determined in complex with GMP and PPi. Intriguingly, the MtHGPRT tetramer crystallizes with a different arrangement of subunits compared to the human counterpart. It would appear that the two enzymes may have evolved from a common dimeric ancestor, but each organism preferentially chose different arrangements to build their respective tetramers. Several ANPs were designed, synthesized, and tested as inhibitors of MtHGPRT. The design of these ANPs is based upon combining the structure of the two products of the reaction, and, as such these compounds would be expected to occupy those sites. With this design strategy in place, the barrier for the M. tuberculosis pathogen to develop resistance to these inhibitors should be high. Two of the ANPs synthesized have Ki values below 1 μM. Prodrugs of the ANPs show good antiTB activity and minimal cytotoxicity. Low cytotoxicity values are a critical factor in drug development. Though some of these compounds are good inhibitors of human HGPRT, the reason for this low cytotoxicity could be that they are not efficiently hydrolyzed to their parent compound in human cells. One of the goals is to improve the drug design so that the prodrugs are maximally hydrolyzed only when they reach the target pathogen. The availability of the crystal structures of MtHGPRT allows compounds with higher potency for MtHGPRT to be designed. Thus, the combination of the anti-TB activity together with low cytotoxicity in mammalian cells and the ability to increase potency and selectivity demonstrate that the ANPs have excellent prospects for development into anti-TB therapeutics.

■

EXPERIMENTAL SECTION

Chemical Synthesis of the Aza-ANPs. Experimental. Unless otherwise stated, solvents were evaporated at 40 °C/2 kPa. NMR spectra were recorded on Bruker Avance 500 (1H at 500 MHz, 13C at 125.8 MHz) and Bruker Avance 400 (1H at 400 MHz, 13C at 100.6 MHz) spectrometers, chemical shifts in ppm with TMS as internal standard or referenced to the residual solvent signal, coupling constants given in Hz. Mass spectra were measured on UPLC-MS (Waters SQD-2). HRMS were measured on LTQ Orbitrap XL (ThermoFisher Scientific) using ESI ionization. The basic chemicals were obtained from commercial sources or prepared according to the published procedures. Preparative HPLC purifications were performed on columns packed with 7 μm C18 reversed phase resin (Waters Delta 600 chromatography column, 17 × 250 mm) in ca. 200 mg batches of K

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

JC,P = 134.8, C−P); 16.15 (m, 4C, Et). MS (ESI): m/z = 610 [M + H]+. Diethyl (2-((hypoxanthin-9-yl)ethyl)(2-((2-(diethoxyphosphoryl)ethyl)(2-hydroxyethyl)amino)ethyl)amino)ethyl)phosphonate (19). Starting from hypoxanthine derivative 16, the desired phosphonate 19 was isolated as minor product, yield 20%. 1H NMR (DMSO-d6): 12.25 (s, 1H, NH); 8.11 (s, 1H) and 8.02 (s, 1H, H-2 and H-8); 4.38 (m, 1H, OH); 4.16 (t, 2H, J1′,2′ = 6.1, H-1′); 3.96 (m, 8H, Et); 3.38 (m, 2H, CH2−O); 2.80 (t, 2H, J2′,1′ = 6.1, H-2′); 2.65 (m, 4H), 2.46 (m, 2H), 2.42 (t, 2H) and 2.32 (t, 2H, CH2−N); 1.80 (m, 4H, CH2−P); 1.21 (m, 12H, Et). 13C NMR (DMSO-d6): 156.51 (C-6); 148.22 (C4); 145.12 (C-2); 140.70 (C-8); 123.67 (C-5); 60.75 (m, 4C, Et); 59.06 (C−OH); 55.42, 52.42, 51.29, 50.82, 47.25, and 47.15 (C−N); 41.64 (C-1′); 22.21 (d, JC,P = 134.4, C−P); 22.06 (d, JC,P = 134.8, C− P); 16.13 (m, 4C, Et). MS (ESI): m/z = 595 [M + H]+. Side Product. Diethyl (2-((2-((2-(diethoxyphosphoryl)ethyl)(2hydroxyethyl)amino)ethyl)(2-(1-(2-(diethoxyphosphoryl)ethyl)-hypoxanthin-9-yl)ethyl)amino)ethyl)phosphonate (20). Starting from hypoxanthine derivative 16, phosphonate 20 was isolated as an undesired main product, yield 40%. 1H NMR (DMSO-d6): 8.35 (s, 1H) and 8.14 (s, 1H, H-2 and H-8); 4.37 (m, 1H, OH); 4.16 (m, 4H, H-1′ and CH2N); 3.98 (m, 12H, Et); 3.38 (m, 2H, CH2−O); 2.79 (t, 2H, J2′,1′ = 6.1, H-2′); 2.66 (m, 4H), 2.48 (m, 2H), 2.41 (m, 2H) and 2.32 (m, 2H, CH2−N); 2.25 (m, 2H) and 1.80 (m, 4H, CH2−P); 1.20 (m, 18H, Et). 13C NMR (DMSO-d6): 155.75 (C-6); 148.03 (C-2); 147.56 (C-4); 141.22 (C-8); 122.88 (C-5); 60.90 (m, 6C, Et); 59.06 (C−OH); 55.38, 52.48, 51.30, 50.80, 47.26, and 47.06 (C−N); 41.48 (C-1′); 40.84 (C−N); 24.43 (d, JC,P = 136.3), 22.09 (d, JC,P = 134.5) and 22.04 (d, JC,P = 134.6, C−P); 16.07 (m, 6C, Et). MS (ESI): m/z = 759 [M + H]+. Diethyl (2-((guanin-9-yl)ethyl)(2-((2-(diethoxyphosphoryl)ethyl)(2-hydroxypropyl)amino)ethyl)amino)ethyl)phosphonate (21). Starting from guanine derivative 17, yield 49%. 1H NMR (DMSOd6): 10.58 (s, 1H, NH); 7.69 (s, 1H, H-8); 6.42 (s, 2H, NH2); 4.42 (m, 1H, OH); 3.95 (m, 10H, Et and H-1′); 3.38 (m, 2H, CH2−O); 2.78 (t, 2H, J2′,1′ = 5.5, H-2′); 2.62 (m, 4H), 2.47 (m, 2H), 2.37 (t, 2H) and 2.28 (t, 2H, CH2−N); 1.80 (m, 4H, CH2−P); 1.46 (m, 2H, H-10′); 1.21 (m, 12H, Et). 13C NMR (DMSO-d6): 156.80 (C-6); 153.37 (C2); 151.01 (C-4); 137.91 (C-8); 116.33 (C-5); 60.82 (m, 4C, Et); 59.02 (C−OH); 52.35, 50.88, 50.82, 49.86, 47.21, and 46.58 (C−N); 41.16 (C-1′); 29.77 (C-10′); 22.25 (d, JC,P = 134.3) and 21.63 (d, JC,P = 134.5, C−P); 16.16 (m, 4C, Et). MS (ESI): m/z = 624 [M + H]+. Diethyl (2-((hypoxanthin-9-yl)ethyl)(2-((2-(diethoxyphosphoryl)ethyl)(2hydroxypropyl)amino)ethyl)amino)ethyl)phosphonate (24). Starting from 6-benzyloxypurine derivative 22, crude intermediate diethyl (2-((2-(6-(benzyloxy)-9H-purin-9-yl)ethyl)(2-((2-(diethoxyphosphoryl) ethyl)(3-hydroxypropyl)amino)ethyl)amino)ethyl)phosphonate (23) [1H NMR (DMSO-d6): 8.54 (s, 1H) and 8.43 (s, 1H, H-2 and H-8); 7.51 (m, 2H) and 7.42 (m, 3H, Ar); 5.63 (s, 2H, CH2Ph); 4.27 (t, 2H, J1′,2′ = 6.0, H-1′); 3.95 (m, 8H, Et); 3.51 (m, 2H, CH2−O); 2.85 (t, 2H, J2′,1′ = 6.1, H-2′); 2.67 (m, 2H), 2.58 (m, 2H), 2.48 (m, 2H), 2.33 (m, 2H) and 2.20 (m, 2H, CH2−N); 1.78 (m, 4H, CH2−P); 1.42 (m, 2H, H-10′); 1.20 (m, 12H, Et). 13C NMR (DMSOd6): 159.62 (C-4); 152.36 (C-6); 151.15 (C-2); 145.55 (C-8); 136.37, 128.47 (2C), 128.35 (2C) and 128.16 (Ar); 120.45 (C-5); 67.70 (CPh); 60.88 (m, 4C, Et); 59.06 (C−OH); 52.20, 50.92, 50.82, 49.86, 47.21, and 46.63 (C−N); 41.83 (C-1′); 29.77 (C-10′); 21.92 (d, JC,P = 135.3) and 20.94 (d, JC,P = 134.0, C−P); 16.26 (m, 4C, Et). MS (ESI): m/z = 699 [M + H]+] was dissolved in dry methanol (10 mL) and was adjusted to pH 2 by the addition of HCl in methanol. The reaction mixture was stirred at room temperature for 3 days, solvent was evaporated and the crude product 24 was purified by chromatography on silica gel (CHCl3-20%MeOH/CHCl3), yield 86%. 1H NMR (DMSO-d6): 12.33 (s, 1H, NH); 8.12 (s, 1H) and 8.03 (s, 1H, H-2 and H-8); 4.17 (m, 3H, OH and H-1′); 3.97 (m, 8H, Et); 3.37 (m, 2H, CH2−O); 2.89 (t, 2H, J2′,1′ = 6.1, H-2′); 2.66 (m, 4H), 2.48 (m, 2H), 2.41 (t, 2H) and 2.33 (m, 2H, CH2−N); 1.82 (m, 4H, CH2−P); 1.47 (m, 2H, H-10′); 1.20 (m, 12H, Et). 13C NMR (DMSO-d6): 156.65 (C-6); 148.37 (C-4); 145.31 (C-2); 140.85 (C-8); 123.80 (C-5); 60.94 (m, 4C, Et); 58.97 (C−OH); 52.50, 50.80, 49.89, 47.21, and

mixtures using a MeOH/H2O gradient as the eluent. The purity of the synthesized compounds was determined with HPLC (H2O−CH3CN, linear gradient) and the combustion analysis (C, H, N) and was higher than 95%. Synthesis of Hydroxyalkyl Derivatives 15−17 and 22. General Procedure. 9-(2-Bromoethyl)purine derivative (3 mmol) and N-(2-hydroxyethyl)ethylenediamine or N-(2-hydroxypropyl)ethylenediamine (12 mmol) were stirred in dimethylformamide (10 mL) overnight. The solvent was evaporated in vacuo, and the mixture was coevaporated with toluene/ethanol (10 mL each). The product was purified by chromatography on silica gel (gradient isopropanol− isopropanol/H2O/NH3) and/or by preparative HPLC. 9-(2-((2-((2-Hydroxyethyl)amino)ethyl)amino)ethyl)guanine (15). Starting from 9-(2-bromoethyl)guanine20a and N-(2-hydroxyethyl)ethylenediamine, yield 57%. 1H NMR (DMSO-d6): 7.65 (s, 1H, H-8); 6.51 (s, 2H, NH2); 4.20 (t, 1H, J = 6.0, OH); 3.96 (t, 2H, J1′,2′ = 6.3, H-1′); 3.41 (t, 2H, J6′,5′ = 5.7, H-6′); 2.82 (t, 2H, J2′,1′ = 6.3, H-2′); 2.55 (m, 6H, H-3′, H-4′ and H-5′). 13C NMR (DMSO-d6): 157.02 (C-6); 153.02 (C-2); 151.09 (C-4); 137.69 (C-8); 116.36 (C-5); 60.32 (C-6′); 51.54 (C-2′); 48.86, 48.57, and 48.38 (C-3′, C-4′, and C-5′); 42.78 (C-1′). MS (ESI): m/z = 282 [M + H]+. 9 - ( 2 - (( 2 - ( (2 - H y d r ox y e t h y l ) am i n o )e th y l ) am i n o ) e th y l ) hypoxanthine (16). Starting from 9-(2-bromoethyl)hypoxanthine20a and N-(2-hydroxyethyl)ethylenediamine, yield 92%. 1H NMR (DMSO-d6): 8.05 (s, 1H) and 8.02 (s, 1H, H-2 and H-8); 4.16 (t, 2H, J1′,2′ = 6.2, H-1′); 3.41 (m, 2H, H-6′); 2.89 (t, 2H, J2′,1′ = 6.2, H2′); 2.52 (m, 6H, H-3′, H-4′, H-5′). 13C NMR (DMSO-d6): 156.76 (C-6); 148.32 (C-4); 145.31 (C-2); 140.46 (C-8); 123.72 (C-5); 60.29 (C-6′); 51.51 (C-2′); 48.79, 48.42, and 48.37 (C-3′, C-4′, and C-5′); 43.33 (C-1′). MS (ESI): m/z = 267 [M + H]+. 9-(2-((2-((2-Hydroxypropyl)amino)ethyl)amino)ethyl)guanine (17). Starting from 9-(2-bromoethyl)guanine 20a and N-(2hydroxypropyl)ethylenediamine, yield 56%. 1H NMR (DMSO-d6): 7.65 (s, 1H, H-8); 6.55 (s, 2H, NH2); 4.20 (t, 1H, J = 6.0, OH); 3.96 (t, 2H, J1′,2′ = 6.2, H-1′); 3.44 (m, 2H, H-7′); 2.82 (t, 2H, J2′,1′ = 6.2, H-2′); 2.53 (m, 6H, H-3′, H-4′ and H-5′); 1.52 (m, 2H, H-6′). 13C NMR (DMSO-d6): 156.93 (C-6); 153.48 (C-2); 151.09 (C-4); 137.71 (C-8); 116.36 (C-5); 59.43 (C-7′); 48.91 (C-2′); 48.27 (2C) and 46.55 (C-3′, C-4′, and C-5′); 42.76 (C-1′); 32.45 (C-6′). MS (ESI): m/z = 296 [M + H]+. 3-((2-((2-(6-(Benzyloxy)-9H-purin-9-yl)ethyl)amino)ethyl)amino)propan-1-ol (22). Starting from 6-benzyloxy-9-(2-bromoethyl)purine20b and N-(2-hydroxypropyl)ethylenediamine, yield 75%. 1H NMR (DMSO-d6): 8.53 (s, 1H) and 8.36 (s, 1H, H-2 and H-8); 7.50 (m, 2H) and 7.40 (m, 2H, Ar); 4.27 (t, 2H, J1′,2′ = 6.0, H-1′); 3.41 (t, 2H, J7′,6′ = 6.3, H-7′); 2.93 (t, 2H, J2′,1′ = 6.1, H-2′); 2.56 (t, 2H) and 2.47 (m, 4H, H-3′, H-4′ and H-5′); 1.48 (m, 2H, H_6′). 13C NMR (DMSO-d6): 159.51 (C-6); 152.29 (C-4); 151.04 (C-2); 144.27 (C8); 136.30, 128.36 (2C), 128.22 (2C) and 128.03 (Ar); 120.40 (C-5); 67.57 (C-Ph); 59.45 (C-7′); 48.91, 48.22, 47.94, and 46.56 (C-2′, C3′, C-4′, and C-5′); 43.42 (C-1′); 32.55 (C-6′). MS (ESI): m/z = 371 [M + H]+. Michael Addition of Diethyl Vinylphosphonate. General Procedure. The 9-substituted purine derivative 15−17 or 22 (2 mmol) and diethyl vinylphosphonate (8 mmol) were stirred in water (20 mL) for 48 h, and then additional diethyl vinylphosphonate (4 mmol) was and reaction was stirred for 24 h. The solvent was evaporated in vacuo, and the mixture was coevaporated with toluene/ ethanol (10 mL each). The product was purified by chromatography on silica gel (gradient CHCl3 − 20%MeOH/CHCl3). Diethyl (2-((guanin-9-yl)ethyl)(2-((2-(diethoxyphosphoryl)ethyl)(2-hydroxyethyl)amino) ethyl)amino)ethyl)phosphonate (18). Starting from guanine derivative 15, yield 42%. 1H NMR (DMSO-d6): 10.53 (s, 1H, NH); 7.70 (s, 1H, H-8); 6.40 (s, 2H, NH2); 4.41 (m, 1H, OH); 3.96 (m, 10H, Et and H-1′); 3.39 (m, 2H, CH2−O); 2.82 (t, 2H), 2.66 (m, 4H), 2.44 (m, 4H) and 2.36 (m, 2H, CH2−N); 1.81 (m, 4H, CH2−P); 1.21 (m, 12H, Et). 13C NMR (DMSO-d6): 156.68 (C6); 153.29 (C-2); 150.98 (C-4); 137.91 (C-8); 116.33 (C-5); 60.77 (m, 4C, Et); 59.09 (C−OH); 55.40, 52.34, 51.33, 50.86, 47.27, and 47.21 (C−N); 41.11 (C-1′); 22.24 (d, JC,P = 134.3, C−P); 22.03 (d, L

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Diethyl (2-((2-((2-(hypoxanthin-9-yl)ethyl)(2(diethoxyphosphoryl)ethyl)amino)ethyl)(2-(diethoxyphosphoryl)ethyl)amino)ethyl)phosphonate (29). Starting from 6-chloropurine derivative 27, yield 79%. 1H NMR (DMSO-d6): 12.26 (s, 1H, NH); 8.11 (s, 1H) and 8.02 (s, 1H, H-2 and H-8); 4.16 (t, 2H, J1′,2′ = 6.1, H1′); 3.96 (m, 12H, Et); 2.80 (t, 2H, J2′,1′ = 6.1, H-2′); 2.67 (m, 2H), 2.60 (m, 4H), 2.47 (m, 2H) and 2.31 (t, 2H, CH2−N); 1.80 (m, 6H, CH2−P); 1.21 (m, 18H, Et). 13C NMR (DMSO-d6): 156.51 (C-6); 148.24 (C-4); 145.13 (C-2); 140.68 (C-8); 123.67 (C-5); 60.75 (m, 6C, Et); 52.41, 50.65, 50.01, 47.19, and 46.06 (2C) C−N); 41.60 (C1′); 22.23 (d, JC,P = 134.3, C−P); 22.08 (d, 2C, JC,P = 134.9, C−P); 16.14 (m, 6C, Et). MS (ESI): m/z = 715 [M + H]+. Synthesis of Free Phosphonic Acids. General Procedure. A mixture of the corresponding ester (0.5 mmol), acetonitrile (10 mL), and BrSiMe3 (0.5 mL) was stirred for 24 h at room temperature. After evaporation and codistillation with acetonitrile (2 × 10 mL), the residue was treated with aqueous methanol (2:1, 10 mL) for 30 min and evaporated to dryness. The residue was purified by preparative HPLC. (2-((2-(Guanin-9-yl)ethyl)(2-((2-hydroxyethyl)(2phosphonoethyl)amino)ethyl)amino)ethyl)phosphonic acid (9). Starting from ester 18, yield 45%. 1H NMR (D2O): 8.10 (s, 1H, H8); 4.50 (t, 2H, J1′,2′ = 5.8, H-1′); 3.92 (m, 2H, CH2−O); 3.63 (m, 6H), 3.45 (m, 4H), and 3.41 (m, 2H, CH2−N); 2.06 (m, 4H, CH2− P). 13C NMR (D2O): 159.28 (C-6); 156.10 (C-2); 152.96 (C-4); 140.70 (C-8); 115.15 (C-5); 57.02 (C−OH); 56.69, 53.89, 52.06, 50.65, 49.74, and 49.23 (C−N); 41.43 (C-1′); 24.59 (d, JC,P = 129.6, C−P); 24.34 (d, JC,P = 129.0, C−P). MS (ESI): m/z = 498 [M + H]+. HRMS (ESI) C15H28O8N7P2 [M-H]− calcd. 496.14801, found 496.14725. (2-((2-(Hypoxanthin-9-yl)ethyl)(2-((2-hydroxyethyl)(2phosphonoethyl)amino)ethyl)amino)ethyl)phosphonic acid (10). Starting from ester 19, yield 75%. 1H NMR (D2O): 8.28 (s, 1H) and 8.26 (s, 1H, H-2 and H-8); 4.77 (m, 2H, H-1′); 3.95 (m, 2H, CH2−O); 3.79 (t, 2H, J2′,1′ = 6.4, H-2′); 3.75 (m, 4H), 3.54 (t, 4H) and 3.45 (t, 2H, CH2−N); 2.09 (m, 4H, CH2−P). 13C NMR (D2O): 157.23 (C-6); 148.12 (C-4); 145.75 (C-2); 141.02 (C-8); 122.0 (C5); 54.44 (C−OH); 54.24, 51.17, 49.67, 48.64, 46.78, and 46.42 (C− N); 38.65 (C-1′); 21.76 (d, JC,P = 129.4, C−P); 21.64 (d, JC,P = 129.0, C−P). MS (ESI): m/z = 483 [M + H]+. HRMS (ESI) C15H27O8N6P2 [M-H]− calcd. 481.13711, found 481.13672. (2-((2-(Guanin-9-yl)ethyl)(2-((2-hydroxypropyl)(2phosphonoethyl)amino)ethyl)amino)ethyl)phosphonic acid (11). Starting from ester 21, yield 65%. 1H NMR (D2O): 8.11 (s, 1H, H8); 4.45 (t, 2H, J1′,2′ = 5.9, H-1′); 3.66 (t, 2H, J11′, 10′ = 5.8, CH2−O); 3.51 (m, 6H), 3.41 (m, 4H) and 3.28 (m, 2H, CH2−N); 2.01 (m, 4H, CH2−P); 1.92 (m, 2H, H-10′). 13C NMR (D2O): 160.05 (C-6); 156.96 (C-2); 153.74 (C-4); 141.59 (C-8); 115.83 (C-5); 61.14 (C− OH); 54.76, 53.71, 52.10, 51.23, 50.70, and 50.12 (C−N); 42.53 (C1′); 28.23 (C-10′); 25.59 (d, JC,P = 129.2, C−P); 25.12 (d, JC,P = 129.0, C−P). MS (ESI): m/z = 510 [M-H]−. HRMS (ESI) C16H30O8N7P2 [M-H]− calcd. 510.16366, found 510.16269. (2-((2-(Hypoxanthin-9-yl)ethyl)(2-((2-hydroxypropyl)(2phosphonoethyl)amino)ethyl)amino)ethyl)phosphonic acid (12). Starting from ester 24, yield 64%. 1H NMR (D2O): 8.19 (s, 2H, H2 and H-8); 4.58 (t, 2H, J1′,2′ = 6.4, H-1′); 3.64 (t, 2H, J11′,10′ = 5.8, CH2−O); 3.50 (t, 2H, J2′,1′ = 6.4, H-2′); 3.45 (m, 6H) and 3.27 (m, 4H, CH2−N); 1.97 (m, 4H, CH2−P); 1.89 (m, 2H, H-10′). 13C NMR (D2O): 158.14 (C-6); 148.82 (C-4); 146.17 (C-2); 142.00 (C-8); 122.93 (C-5); 58.46 (C−OH); 51.88, 50.90, 49.34, 48.47, 48.21, and 47.19 (2C) C−N); 40.04 (C-1′); 25.49 (C-10′); 22.85 (d, JC,P = 129.5, C−P); 22.41 (d, JC,P = 129.0, C−P). MS (ESI): m/z = 495 [M-H]−. HRMS (ESI) C16H29O8N6P2 [M-H]− calcd. 495.15276, found 495.15227. (2-((2-((2-(Guanine-9-yl)ethyl)(2-phosphonoethyl)amino)ethyl)(2-phosphonoethyl)amino)ethyl)phosphonic acid (13). Starting from ester 28, yield 90%. 1H NMR (D2O): 8.65 (s, 1H, H-8); 4.66 (t, 2H, J1′,2′ = 6.1, H-1′); 3.72 (t, 2H, J2′,1′ = 6.1, H-2′); 3.66 (m, 4H), 3.54 (m, 2H) and 3.47 (m, 4H, CH2−N); 2.08 (m, 6H, CH2−P). 13C NMR (D2O): 155.01 (C-6); 153.98 (C-2); 149.58 (C-4); 137.07 (C-

46.68 (2C) C−N); 41.74 (C-1′); 29.73 (C-10′); 22.27 (d, JC,P = 134.2, C−P); 21.60 (d, JC,P = 132.3, C−P); 16.13 (m, 4C, Et). MS (ESI): m/ z = 609 [M + H]+. Tetraethyl (((2-((2-(diethoxyphosphoryl)ethyl)(2-hydroxyethyl)amino)ethyl)azanediyl)bis(ethane-2,1-diyl))bis(phosphonate) (25). N-(2-Hydroxyethyl)ethylenediamine (2 mL, 19 mmol) and diethyl vinylphosphonate (12.5 mL, 76 mmol) were stirred in water (30 mL) for 48 h, and then additional diethyl vinylphosphonate (4 mmol) was added, and the reaction was stirred for 24 h. Water was evaporated in vacuo, and the mixture was coevaporated with toluene (2 × 10 mL). The product was purified by chromatography on silica gel (gradient CHCl3 − 20%MeOH/CHCl3), yield 10.47 g, 92%, colorless oil. 1H NMR (DMSO-d6): 4.44 (s, 1H, OH); 3.98 (m, 12H, Et); 3.41 (m, 2H, CH2−O), 2.67 (m, 6H), and 2.46 (m, 6H, CH2−N); 1.86 (m, 6H, CH2−P); 1.22 (m, 18H (Et)). 13C NMR (DMSO-d6): 60.71 (m, 6C (Et)); 59.15 (C−OH), 55.41, 51.21, 50.17, 47.38, and 46.08, (2C) C− N); 22.17 (d, 3C, JC,P = 134.8, C−P); 16.12 (m, 6C, Et). MS (ESI): m/z = 597 [M + H]+. Mitsunobu Reaction. General Procedure. To the solution of triphenylphosphine (2.2 g, 8.4 mmol) in dry THF (50 mL) cooled to −30 °C under argon atmosphere was added slowly diisopropylazadicarboxylate (DIAD, 1.65 mL, 8.4 mmol). The mixture was stirred for 30 min and this preformed complex was added to the mixture of corresponding purine base (6.1 mmol) and hydroxy derivative 25 (2.85 g, 4.77 mmol) in dry THF (50 mL) at −30 °C under argon. The resulting mixture was slowly warmed to room temperature and stirred for 3 days. Then, in the case of 26, water (10 mL) was added, and the mixture was heated at 80 °C for 30 h. Solvent was evaporated and the crude mixture was purified by chromatography on silica gel (MeOH− CHCl3). The pure product was obtained as yellowish foam. Diethyl (2-((2-((2-(2-amino-6-chloropurin-9-yl)ethyl)(2(diethoxyphosphoryl)ethyl)amino)ethyl)(2-(diethoxyphosphoryl)ethyl)amino)ethyl)phosphonate (26). Starting from 2-amino-6chloropurine, yield 34%. 1H NMR (DMSO-d6): 8.16 (s, 1H, H-8); 6.88 (s, 2H, NH2); 4.07 (t, 2H, J1′,2′ = 5.6, H-1′); 3.98 (m, 12H, Et); 2.78 (t, 2H, J2′,1′ = 5.6, H-2′); 2.60 (m, 6H), 2.47 (m, 2H), and 2.29 (t, 2H, J = 6.6, CH2−N); 1.82 (m, 6H, CH2−P); 1.21 (m, 18H, Et). 13C NMR (DMSO-d6): 159.57 (C-2); 154.00 (C-4); 149.07 (C-6); 143.77 (C-8); 123.15 (C-5); 60.79 (m, 6C, Et); 52.07, 51.89, 50.64, 50.10, 47.15, and 45.98 (2C) C−N); 41.37 (C-1′); 22.23 (d, JC,P = 134.1, C−P) and 21.99 (d, 2C, JC,P = 134.8, C−P); 16.15 (m, 6C, Et). MS (ESI): m/z = 748 [M + H]+. Diethyl (2-((2-((2-(6-chloropurin-9-yl)ethyl)(2(diethoxyphosphoryl)ethyl)amino)ethyl)(2-(diethoxyphosphoryl)ethyl)amino)ethyl)phosphonate (27). Starting from 6-chloropurine, yield 42%. 1H NMR (DMSO-d6): 8.77 (s, 1H) and 8.75 (s, 1H, H-2 and H-8); 4.33 (t, 2H, J1′,2′ = 5.8, H-1′); 3.97 (m, 12H, Et); 2.86 (t, 2H, J2′,1′ = 5.8, H-2′); 2.67 (m, 2H), 2.54 (m, 4H), 2.45 (m, 4H) and 2.21 (t, 2H, CH2−N); 1.78 (m, 6H, CH2−P); 1.22 (m, 18H, Et). 13C NMR (DMSO-d6): 151.93 (C-4); 151.16 (C-2); 148.73 (C-6); 148.02 (C-8); 130.61 (C-5); 60.78 (m, 6C, Et); 52.58, 50.32, 49.98, 47.03, and 45.91 (2C) C−N); 42.03 (C-1′); 21.93 (d, JC,P = 130.6, 3C, C− P); 16.15 (m, 6C, Et). MS (ESI): m/z = 733 [M + H]+. Synthesis of Triphosphonates 28 and 29 -. General Procedure. The corresponding 6-chloropurine or 2-amino-6-chloropurine derivative (2 mmol) was dissolved in trifluoroacetic acid (aqueous, 75%, 15 mL) and stirred overnight. The solvent was evaporated, and the residue was codistilled with water (3 × 10 mL) and ethanol (20 mL). After purification by preparative HPLC, the pure products were obtained as colorless foams. Diethyl (2-((2-((2-(guanin-9-yl)ethyl)(2-(diethoxyphosphoryl)ethyl)amino)ethyl)(2-(diethoxyphosphoryl)ethyl)amino)ethyl)phosphonate (28). Starting from 2-amino-6-chloropurine derivative 26, yield 50%. 1H NMR (DMSO-d6): 10.51 (s, 1H, NH); 7.70 (s, 1H, H-8); 6.40 (s, 2H, NH2); 3.96 (m, 14H, Et and H-1′); 2.74 (m, 2H), 2.63 (m, 8H), and 2.35 (m, 2H, CH2−N); 1.82 (m, 6H, CH2−P); 1.20 (m, 18H, Et). 13C NMR (DMSO-d6): 156.67 (C-6); 153.31 (C-2); 151.00 (C-4); 137.84 (C-8); 116.32 (C-5); 60.83 (m, 6C, Et); 52.35, 50.68, 47.21, and 46.02 (3C, C−N); 40.51 (C-1′); 22.24 (d, JC,P = 136.6, C−P); 22.02 (d, 2C, JC,P = 131.3, C−P); 16.15 (m, 6C, Et). MS (ESI): m/z = 730 [M + H]+. M

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

18H, Et). 13C NMR (DMSO-d6): 172. 98 (m, 6C, CO); 156.81 (C-6); 153.46 (C-2); 151.13 (C-4); 137.47 (C-8); 137.14 (m, 6C), 129.30 (m, 12C), 128.04 (m, 12C) and 126.38 (m, 6C, Ar); 116.30 (C-5); 60.28 (m, 6C, Et); 54.00 (m, 6C, NHCH); 52.42, 50.27, 49.87, 47.71, and 46.47 (2C) C−N); 40.36 (C-1′); 26.01 (d, JP,C = 111.6, C−P); 25.52 (d, 2C, JP,C = 110.1, C−P); 13.86 (m, 6C, Et). MS (ESI): m/z = 1613 [M + H]+. HRMS (ESI) C81H109O16N13P3[M + H]+ calcd. 1612.73226, found 1612.73286. Tetra-(L-phenylalanine ethyl ester) Prodrug of (2-((2-((2-(Hypoxanthine-9-yl)ethyl)(2-phosphonoethyl)amino)ethyl)(2phosphonoethyl)amino)ethyl)phosphonic acid (33). Starting from ester 29, yield 60%. 1H NMR (DMSO-d6): 12.29 (s, 1H, NH); 8.05 (m, 2H, H-2 and H-8); 7.20 (m, 30H, Ar); 4.52 (m, 3H) and 4.27 (m, 3H, NH); 4.08 (m, 2H, H-1′); 3.98 (m, 18H, Et and CHNH); 2.84 (m, 12H, CH2Ph), 2.62 (m, 2H, H-2′); 2.45 (m, 2H), 2.32 (m, 6H) and 2.08 (m, 2H, CH2−N); 1.35 (m, 6H, CH2−P); 1.05 (m, 18H, Et). 13 C NMR (DMSO-d6): 173.08 (m, 6C, CO); 156.62 (C-6); 148.27 (C-4); 145.39 (C-2); 140.48 (C-8); 137.15 (m, 6C), 129.30 (m, 12C), 128.05 (m, 12C) and 126.40 (m, 6C, Ar); 123.73 (C-5); 60.22 (m, 6C, Et); 53.98 (m, 6C, NHCH); 52.53, 50.43, 49.85, 47.57, and 46.45 (2C) (C−N); 41.23 (C-1′); 25.93 (d, JP,C = 111.0, C−P); 25.47 (d, 2C, JP,C = 111.0, C−P); 13.85 (m, 6C, Et). MS (ESI): m/z = 1598 [M + H]+. HRMS (ESI) C81H108O16N12P3 [M + H]+ calcd. 1597.72488, found 1597.72196. Expression and Purification of MtHGPRT. The cDNA coding for MtHGPRT was cloned into the expression vector pET-22b(+) by GeneArt to produce a recombinant N-terminal hexa-histidine tagged protein when expressed. The construct was then transformed into BL21(DE3) E. coli cells. These were subsequently grown in LB media at 37 °C until an OD600 of 0.7 was reached. Protein expression was induced by the addition of 1 mM isopropyl-β-D-thiogalactoside (IPTG) for 4 h at 25 °C. Cells were harvested by centrifugation at 5000 g for 10 min at 4 °C. The cell pellet was resuspended in 0.1 M Tris-HCl, 12 mM MgCl2, 300 μM PRib-PP, pH 7.4 and stored at −70 °C. The cells were lysed by the addition of lysozyme (2−4 mg per mL of cell suspension) and by freeze/thawing in liquid nitrogen three times.14 The enzyme was purified by immobilized metal affinity chromatography (IMAC). The elution buffer was 0.1 M Tris-HCl, 12 mM MgCl2, 0.4 M imidazole, pH 7.4. The enzyme was then dialyzed into 0.1 M Tris-HCl, 12 mM MgCl2, pH 7.4 at 4 °C to remove salts. The enzyme was concentrated and PRib-PP added to give a final concentration of 200 μM and stored in 25 or 100 μL aliquots at −70 °C. Under these conditions, there was no loss of activity for at least four months. Assays for Kinetic Constant Determination. The kinetic constants, kcat and K m(app) were measured by a continuous spectrophotometric assay in 0.1 M Tris-HCl, 12 mM MgCl2, pH 7.4, at 25 °C.15 For the determination of the kinetic constants using guanine or hypoxanthine, the concentration of PRib-PP was 2 mM. For the determination of the kinetic constants for PRib-PP, the concentration of the purine bases were at saturating concentrations (range of concentration of PRib-PP was 100−2000 μM). The reaction was initiated by the addition of enzyme to give a final concentration of 200 nM. The reaction was followed at 257.5 nm for guanine and 245 nm for hypoxanthine over 60 s, 0.05 absorbance range. The Δε values are 5816.5 M−1 cm−1 for guanine and 2439 M−1 cm−1 for hypoxanthine. The Km and kcat values were calculated using Prism6 (GraphPad Software, Inc., La Jolla, CA). Inhibition Studies. The Ki values for MtHGPRT were determined in 0.1 M Tris-HCl, 12 mM MgCl2, pH 7.4, at 25 °C.12 The concentration of guanine was fixed at 67.5 μM, and the concentration of the variable substrate, PRib-PP, was in the range 20−1500 μM, depending on the value for Km(app) at that concentration of inhibitor. The concentration of inhibitor in the assay ranged from 1.4 μM (for the best inhibitor) to 56 μM (for the weakest inhibitor). The reaction was initiated by the addition of enzyme to give a final concentration of 200 nM. The reaction was followed at 257.5 nm for 60 s with a total change in absorbance of 0.05 units. The Km(app), Vmax, and Ki values were calculated using Prism6 (GraphPad Software, Inc., La Jolla, CA). Hanes’ plots for compounds 5, 9, and 13 are shown in Figure S2.

8); 116.41 (C-5); 50.62, 48.24 (2C), 48.04, 46.55, and 46.17 (C−N); 38.96 (C-1′); 21.43 (d, JC,P = 133.5, C−P). MS (ESI): m/z = 560 [MH]−. HRMS (ESI) C15H31O10N7P3 [M-H]− calcd. 562.13398, found 562.13416. (2-((2-((2-(Hypoxanthine-9-yl)ethyl)(2-phosphonoethyl)amino)ethyl)(2-phosphonoethyl)amino)ethyl)phosphonic acid (14). Starting from ester 29, yield 89%. 1H NMR (D2O): 8.45 (s, 1H) and 8.28 (s, 1H, H-2 and H-8); 4.75 (m, 2H, H-1′); 3.79 (t, 2H, J2′, 1′ = 6.5, H2′); 3.71 (m, 4H) and 3.48 (m, 6H, CH2−N); 2.09 (m, 6H, CH2−P). 13 C NMR (D2O): 156.41 (C-6); 147.74 (C-4); 146.18 (C-2); 140.58 (C-8); 120.61 (C-5); 50.79, 48.35 (2C), 46.34 and 46.11 (C−N); 38.63 (C-1′); 21.58 (d, 2C, JC,P = 129.9, C−P); 21.35 (d, JC,P = 129.9, C−P). MS (ESI): m/z = 545 [M-H]−. HRMS (ESI) C15H28O10N6P3 [M-H]− calcd. 545.10852, found 545.10795. Synthesis of Phosphoramidate Prodrugs. General Procedure. A mixture of corresponding ester (0.5 mmol), dry pyridine (8 mL), and BrSiMe3 (1 mL) was stirred overnight at room temperature under argon. After evaporation and codistillation with pyridine (2 × 5 mL) under argon atmosphere, the residue was dissolved in dry pyridine (10 mL) and ethyl (L)-phenylalanine hydrochloride (1.75 g, 7.1 mmol) and triethylamine (3.1 mL) were added. The mixture was heated to 70 °C under argon atmosphere and then a solution of Aldrithiol (2.31 g, 10.5 mmol) and triphenylphosphine (2.75 g, 10.5 mmol) in dry pyridine (8 mL) was added. The reaction mixture was heated at 70 °C for 3 days, the solvent was evaporated and the residue was coevaporated with toluene (2 × 10 mL) and purified by column chromatography on silica gel (gradient CHCl3-MeOH) and the crude product further purified by preparative HPLC. The phosphoramidate prodrug was obtained as foam. Tetra-(L-phenylalanine ethyl ester) Prodrug of (2-((2-(Guanin-9yl)ethyl)(2-((2-hydroxyethyl)(2-phosphonoethyl)amino)ethyl)amino)ethyl)phosphonic acid (30). Starting from ester 18, yield 59%. 1 H NMR (DMSO-d6): 10.65 (s, 1H, NH); 7.67 (s, 1H, H-8); 7.20 (m, 20H, Ar); 6.55 (s, 2H, NH2); 5.33 (s, 1H, OH); 4.62 (m, 4H, NH); 3.99 (m, 14H, Et, CHNH and H-1′); 3.66 (m, 2H, H-10′); 2.89 (m, 8H, CH2Ph), 2.80 (m, 4H), 2.67 (m, 4H), 2.54 (m, 2H) and 2.46 (m, 2H, CH2−N); 1.7 (m, 4H, CH2−P); 1.09 (m, 12H, Et). 13C NMR (DMSO-d6): 172. 99 m, 4 C (CO); 156.72 (C-6); 153.49 (C-2); 151.02 (C-4); 137.41 (C-8); 137.12 (m, 4C), 129.29 (m, 8C), 128.04 (m, 8C) and 126.40 (m, 4C, Ar); 116.36 (C-5); 60.30 (m, 4C, Et); 60.21 (C−OH); 53.98 (m, 4C, NHCH); 55.94, 50.48, 50.22, 49.59, and 47.95 (2C) C−N); 41.49 (C-1′); 27.67 (d, JP,C = 279.6, C−P); 27.60 (d, JP,C = 271.4, C−P); 13.87 (m, 4C, Et). MS (ESI): m/z = 1199 [M + H]+. HRMS (ESI) C59H82O12N11P2 [M + H]+ calcd. 1198.56142, found 1198.56209. Tetra-(L-phenylalanine ethyl ester) Prodrug of (2-((2-(Hypoxanthin-9-yl)ethyl)(2-((2-hydroxyethyl)(2-phosphonoethyl)amino)ethyl)amino)ethyl)phosphonic acid (31). Starting from ester 19, yield 37%. 1H NMR (DMSO-d6): 12.33 (s, 1H, NH); 8.08 (s, 1H) and 8.04 (s, 1H, H-2 and H-8); 7.20 (m, 16H) and 7.12 (m, 4H, Ar); 5.32 (s, 1H, OH); 4.60 (m, 4H, NH); 4.16 (m, 2H, H-1′); 4.00 (m, 10H, Et and CHNH); 3.88 (m, 2H, CHNH); 3.66 (m, 2H, H-10′); 2.89 (m, 8H, CH2Ph), 2.81 (m, 4H), 2.67 (m, 2H), 2.55 (m, 2H), 2.45 (m, 2H) and 2.33 (m, 2H, CH2−N); 1.61 (m, 4H, CH2−P); 1.09 (m, 12H, Et). 13 C NMR (DMSO-d6): 172.89 (m, 4C, CO); 156.58 (C-6); 148.22 (C-4); 145.43 (C-2); 140.44 (C-8); 137.10 (m, 4C), 129.27 (m, 8C), 128.05 (m, 8C) and 126.42 (m, 4C, Ar); 123.77 (C-5); 60.27 (m, 4C, Et); 60.13 (C−OH); 53.93 (m, 4C, NHCH); 57.03, 56.23, 52.13, 49.30, 47.95, and 45.45 (C−N); 41.22 (C-1′); 27.59 (d, JP,C = 266.6, C−P); 27.55 (d, JP,C = 267.1 (C−P); 13.82 (m, 4C, Et). MS (ESI): m/ z = 1184 [M + H]+. HRMS (ESI) C59H81O12N10P2 [M + H]+ calcd. 1183.55052, found 1183.55129. Tetra-(L-phenylalanine ethyl ester) Prodrug of (2-((2-((2-(Guanine-9-yl)ethyl)(2-phosphonoethyl)amino)ethyl)(2phosphonoethyl)amino)ethyl)phosphonic acid (32). Starting from ester 28, yield 57%. 1H NMR (DMSO-d6): 10.53 (s, 1H, NH); 7.64 (s, 1H, H-8); 7.20 (m, 30H, Ar); 6.53 (s, 2H, NH2); 4.55 (m, 3H) and 4.28 (m, 3H, NH); 3.94 (m, 20H, Et, CHNH and H-1′); 2.83 (m, 12H, CH2Ph), 2.55 (m, 2H, H-2′); 2.45 (m, 2H), 2.37 (m, 4H), 2.31 (m, 2H) and 2.14 (m, 2H, CH2−N); 1.40 (m, 6H, CH2−P); 1.06 (m, N

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

in a humidified incubator prior to the addition of 30 μL of a 0.02% resazurin solution and 12.5 μL of 20% Tween 80 to each well. Sample fluorescence was measured 30h later on a FluoroStar Omega fluorescent plate reader (BMG) with an excitation wavelength of 530 nm and emission read at 580 nm. Changes in fluorescence relative to positive control wells (H37Rv with no inhibitor) minus negative control wells (no H37Rv) were plotted for determination of MIC50. Evaluation of in Vitro Cytotoxicity of ANP Prodrugs. The inhibitory effect of the test compounds on cell proliferation was determined in A549 human lung carcinoma cells.34 To determine the cytostatic effect of the test compounds, the cells were seeded in 96well plates at 7500 cells per well, and 24 h later, the compounds were added at serial dilutions. After 4 days incubation at 37 °C, the cells were trypsinized and counted with a Coulter Counter apparatus. The CC50 values, or compound concentrations at which cell proliferation was 50% compared to that in untreated cells, were calculated by extrapolation. Data presented are the mean ± SEM of two or three independent tests. The selectivity indices (SI) were determined as a ratio of CC50 in human A549 cell lines vs the MIC50 values in M. tuberculosis (H37Rv).

These compounds contain guanine as the base. 5 is from group B where the second phosphonate group branches off a carbon atom, 9 is from group B where the second phosphonate group branches off the nitrogen and 13 from group C contains three phosphonate groups (Table 2). Crystallization and Structure Determination of MtHGPRT. MtHGPRT, at a concentration of 8.3 mg/mL (0.36 mM), was incubated with the ligands (0.45 mM for GMP, 2.7 mM for 5, 1.6 mM for 9 and 4.1 mM for 14) on ice for ∼5 min. For crystallization, the hanging drop method was used where 1 μL of MtHGPRT in complex with the inhibitor was added to 1 μL of well solution. The trays were then stored at 18 °C. The reservoir solution for GMP, 9, and 14 was 20% PEG8000, 0.1 M Tris-HCl pH 8.5, 0.2 M MgCl2. For 5, the well solution was 0.1 M potassium sodium tartrate, 0.1 M imidazole pH 8.0, 0.7 M NaCl. Prior to data collection, crystals of the MtHGPRT-5 complex were transferred to a cryoprotectant solution that contained well solution, 2 mM inhibitor, and 25% glycerol. For the GMP.PPi complex, 10% glycerol was required for cryoprotection. No cryoprotectant was used for the complexes with 9 and 14. Crystals were cryocooled in liquid nitrogen for shipment to the Australian Synchrotron where they were robotically placed in a cryostream (100 K) on beamline MX2. X-ray data were collected remotely by BLUICE.27 All data sets were scaled and merged with XDS.28 The structures were solved by molecular replacement using the program PHASER,29 within PHENIX 1.7.3.30 The protein coordinates of subunit A from the IMP bound EcHPRT complex was the starting model (PDB code 1G9S) to solve the crystal structure for MtHGPRT in complex with 9. The other three structures were determined using this structure as the starting model. Subsequent refinement and model building was with PHENIX 1.7.330 and COOT 0.7,31 respectively. The structural restraints file for the inhibitors were generated by PRODRG2.32 Visible residues in the electron density for MtHGPRT in complex with GMP.PPi are 22−49, 56−94, 101−199 (subunit A), 21−50, 57−93, 102−200 (subunit B), 24−49, 54−93, 101−158, 160− 196 (subunit C), 22−48, 57−96, 101−157, 161−198 (subunit D); for MtHGPRT in complex with 5 are 14−50, 54−93, 100−202 (subunit A), 16−93, 101−158, 161−202 (subunit B), 18−93, 101−199 (subunit C), 21−50, 55−93, 103−199 (subunit D), 17−49, 55−92, 105−202 (subunit E), 17−49, 56−93, 102−202 (subunit F); for MtHGPRT in complex with 9, 17−94, 99−201 (subunit A), 16−49, 52−92, 103−202 (subunit B), 16−92, 103−159, 161−202 (subunit C), 18−92, 103−201 (subunit D); for MtHGPRT in complex with 14, 17−50, 56−93, 102−201 (subunit A), 17−49, 56−94, 101−201 (subunit B), 17−50, 54−93, 100−202 (subunit C), 18−49, 53−95, 100−202 (subunit D). Residues where only alanine could be modeled for the side-chains are for MtHGPRT in complex with GMP.PPi, E31, E42, R136, D197, R199 (subunit A); K24, E31, Q34, R111, R136, N142, N161, D184, D197, R199 (subunit B); K24, E31, Q34, E39, E42, R49, T54, Q57, D126, R136, L152, R153, K154, V158, N161, I171, N173, D174, L196 (subunit C); K24, E31, Q34, I37, E39, E42, S96, H114, D117, S127, R147, L152, R153, E163, D170, N173 (subunit D); for MtHGPRT in complex with 5 are R199 (subunit A); R49, N161 (subunit B); E50, L51, E163, R199 (subunit C); R199 (subunit D); E17 (subunit E); MtHGPRT in complex with 14 are R199 (subunit A); G31, E163, R199 (subunit B); G31, D156, R199 (subunit C); G31, R49, E163 (subunit D). M. tuberculosis Inhibition Assay. M. tuberculosis H37Rv (ATCC 25177) was grown in Middlebrook 7H9 broth medium supplemented with OADC (oleic acid dextrose catalase), 0.5% glycerol, and 0.02% tyloxapol. Freshly seeded cultures were grown at 37 °C, for approximately 14 days, to midexponential phase (OD600 0.4−0.8) for use in inhibition assays. The potency of the inhibitors was measured by a resazurin reduction microplate assay, as previously described,33 with some alterations. M. tuberculosis, grown to midexponential phase (OD600 0.4−0.8), was diluted to OD600 0.001 in 7H9S media (Middlebrook 7H9 with OADC, 0.5% glycerol, 0.75% tween-80, 1% tryptone) containing 0.5% DMSO. 96-well microtiter plates were setup with 100 μL of inhibitors, serially diluted into 7H9S media. 100 μL of diluted M. tuberculosis, representing ∼2 × 104 CFU/ mL was added to each well. Plates were incubated for 5 days at 37 °C

■

ASSOCIATED CONTENT

S Supporting Information *

The atomic coordinates and structure factors of M. tuberculosis HGPRT in complex with the three inhibitors and GMP.PPi.Mg2+ have been deposited in the Protein Data Bank. For the GMP.PPi.Mg2+ complex, phosphonic acid (5) complex, {2-((2-(guanin-9-yl)ethyl)(2-((2-hydroxyethyl)(2phosphonoethyl)amino)ethyl)amino) ethyl} phosphonic acid (9) complex, and {2-((2-((2-(hypoxanthine-9-yl)ethyl)(2phosphonoethyl)amino)ethyl)(2-phosphonoethyl)amino)ethyl}phosphonic acid (14) complex, the accession codes are 4RHT, 4RHU, 4RHX, and 4RHY, respectively. The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jmedchem.5b00611.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The initial crystallographic conditions were determined using the Mosquito and Rock Imager facilities at the University of Queensland Remote-Operation Crystallization and X-ray Diffraction facility (UQROCX). Measurements were made at the MX2 beamline, Australian Synchrotron, Clayton, Victoria with the assistance of Tom Caradoc-Davies. The views expressed here are those of the authors and not necessarily those of the Australian Synchrotron. This work was supported by funds from the National Health and Medical Research Council (Grant No. 1030353), by the subvention for development of research organization (Institute of Organic Chemistry and Biochemistry, RVO 61388963), by the Grant Agency of the Czech Republic (Grant No. P207/11/0108), MVCR VG20102015046, and by Gilead Sciences (Foster City, CA).

■

ABBREVIATION USED HGPRT, hypoxanthine-guanine phosphoribosyltransferase; HGXPRT, hypoxanthine-guanine-xanthine phosphoribosyltransferase; PRTase, phosphoribosyltransferase; PRib-PP, 5O

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

(13) (a) Shi, W. X.; Li, C. M.; Tyler, P. C.; Furneaux, R. H.; Grubmeyer, C.; Schramm, V. L.; Almo, S. C. The 2.0 Å structure of human hypoxanthine guanine phosphoribosyltransferase in complex with a transition-state analog inhibitor. Nat. Struct. Biol. 1999, 6 (6), 588−593. (b) Shi, W.; Li, C. M.; Tyler, P. C.; Furneaux, R. H.; Cahill, S. M.; Girvin, M. E.; Grubmeyer, C.; Schramm, V. L.; Almo, S. C. The 2.0 Å structure of malarial purine phosphoribosyltransferase in complex with a transition-state analogue inhibitor. Biochemistry 1999, 38 (31), 9872−9880. (14) Guddat, L. W.; Vos, S.; Martin, J. L.; Keough, D. T.; de Jersey, J. Crystal structures of free, IMP-, and GMP-bound Escherichia coli hypoxanthine phosphoribosyltransferase. Protein Sci. 2002, 11 (7), 1626−1638. (15) Keough, D. T.; Skinner-Adams, T.; Jones, M. K.; Ng, A.-L.; Brereton, I. M.; Guddat, L. W.; de Jersey, J. Lead compounds for antimalarial chemotherapy: Purine base analogs discriminate between human and P. falciparum 6-oxopurine phosphoribosyltransferases. J. Med. Chem. 2006, 49 (25), 7479−7486. (16) Keough, D. T.; Hocková, D.; Krečmerová, M.; Č esnek, M.; Holý, A.; Naesens, L.; Brereton, I. M.; Winzor, D. J.; de Jersey, J.; Guddat, L. W. Plasmodium vivax hypoxanthine-guanine phosphoribosyltransferase: A target for anti-malarial chemotherapy. Mol. Biochem. Parasitol. 2010, 173 (2), 165−169. (17) Biazus, G.; Schneider, C. Z.; Palma, M. S.; Basso, L. A.; Santos, D. S. Hypoxanthine−guanine phosphoribosyltransferase from Mycobacterium tuberculosis H37Rv: Cloning, expression, and biochemical characterization. Protein Expr. Purif. 2009, 66 (2), 185−190. (18) Wenck, M. A.; Medrano, F. J.; Eakin, A. E.; Craig, S. P. Steadystate kinetics of the hypoxanthine phosphoribosyltransferase from Trypanosoma cruzi. Biochim. Biophys. Acta 2004, 1700 (1), 11−18. (19) Keough, D.; Ng, A.-L.; Winzor, D.; Emmerson, B.; de Jersey, J. Purification and characterization of Plasmodium falciparum hypoxanthine−guanine−xanthine phosphoribosyltransferase and comparison with the human enzyme. Mol. Biochem. Parasitol. 1999, 98 (1), 29−41. (20) (a) Legros, V.; Hamon, F.; Violeau, B.; Turpin, F.; DjedainiPilard, F.; Desire, J.; Len, C. Toward the Supramolecular Cyclodextrin Dimers Using Nucleobase Pairs. Synthesis 2011, 2, 235−242. (b) Ramzaeva, N.; Goldberg, Yu.; Alksnis, E.; Lidaks, M. Synthesis and N-Alkylation of 6-Benzyloxypurine Under Phase-Transfer Conditions. Synth. Commun. 1989, 19, 1669−1676. (c) Pertusati, F.; Serpi, M.; McGuigan, C. Medicinal chemistry of nucleoside phosphonate prodrugs for antiviral therapy. Antiviral Chem. Chemother. 2012, 22 (5), 181−203. (d) Jansa, P.; Baszczynski, O.; Dracinsky, M.; Votruba, I.; Zidek, Z.; Bahador, G.; Stepan, G.; Cihlar, T.; Mackman, R.; Holy, A.; Janeba, Z. A novel and efficient one-pot synthesis of symmetrical diamide (bis-amidate) prodrugs of acyclic nucleoside phosphonates and evaluation of their biological activities. Eur. J. Med. Chem. 2011, 46 (9), 3748−3754. (21) Eads, J. C.; Scapin, G.; Xu, Y.; Grubmeyer, C.; Sacchettini, J. C. The crystal structure of human hypoxanthine-guanine phosphoribosyltransferase with bound GMP. Cell 1994, 78 (2), 325−334. (22) Heroux, A.; White, E. L.; Ross, L. J.; Davis, R. L.; Borhani, D. W. Crystal structure of Toxoplasma gondii hypoxanthine-guanine phosphoribosyltransferase with XMP, pyrophosphate, and two Mg2+ ions bound: Insights into the catalytic mechanism. Biochemistry 1999, 38 (44), 14495−14506. (23) Chen, Q.; You, D.; Liang, Y.; Su, X.; Gu, X.; Luo, M.; Zheng, X. Crystal structure of Thermoanaerobacter tengcongensis hypoxanthineguanine phosphoribosyl transferase L160I mutant − insights into inhibitor design. Febs J. 2007, 274 (17), 4408−4415. (24) de Jersey, J.; Holý, A.; Hocková, D.; Naesens, L.; Keough, D. T.; Guddat, L. W. 6-oxopurine phosphoribosyltransferase: A target for the development of antimalarial drugs. Curr. Top Med. Chem. 2011, 11 (16), 2085−2102. (25) Meola, M.; Yamen, B.; Weaver, K.; Sandwick, R. K. The catalytic effect of Mg2+ and imidazole on the decomposition of 5phosphoribosyl-α-1-pyrophosphate in aqueous solution. J. Inorg. Biochem. 2003, 93 (3−4), 235−242.

phospho-α-D-ribosyl-1-pyrophosphate; Mt, Mycobacterium tuberculosis; Pf , Plasmodium falciparum; SI, selectivity index; azaANP, aza-acyclic nucleoside phosphonate; ANP, acyclic nucleoside phosphonate; PPi, pyrophosphate

■

REFERENCES

(1) World Health Organization, Global Tuberculosis Report. http:// www.who.int/tb/publications/global_report/en/; accessed October 20, 2014. 2012. (2) Günther, G. Multidrug-resistant and extensively drug-resistant tuberculosis: a review of current concepts and future challenges. Clin. Med. 2014, 14 (3), 279−285. (3) Griffin, J. E.; Gawronski, J. D.; DeJesus, M. A.; Ioerger, T. R.; Akerley, B. J.; Sassetti, C. M., High-resolution phenotypic profiling defines genes essential for mycobacterial growth and cholesterol catabolism. PLoS Pathog. 2011, 7 (9). (4) (a) Ducati, R. G.; Breda, A.; Basso, L. A.; Santos, D. S. Purine salvage pathway in Mycobacterium tuberculosis. Curr. Med. Chem. 2011, 18 (9), 1258−1275. (b) Keough, D. T.; Hocková, D.; Rejman, D.; Špaček, P.; Vrbková, S.; Krečmerová, M.; Eng, W. S.; Jans, H.; West, N. P.; Naesens, L. M. J.; de Jersey, J.; Guddat, L. W. Inhibition of the Escherichia coli 6-oxopurine phosphoribosyltransferases by nucleoside phosphonates: Potential for new antibacterial agents. J. Med. Chem. 2013, 56 (17), 6967−6984. (5) (a) Hocková, D.; Keough, D. T.; Janeba, Z.; Wang, T.-H.; de Jersey, J.; Guddat, L. W. Synthesis of novel N-branched acyclic nucleoside phosphonates as potent and selective inhibitors of human, Plasmodium falciparum and Plasmodium vivax 6-oxopurine phosphoribosyltransferases. J. Med. Chem. 2012, 55 (13), 6209−6223. (b) Keough, D. T.; Hocková, D.; Janeba, Z.; Wang, T.-H.; Naesens, L.; Edstein, M. D.; Chavchich, M.; Guddat, L. W. Aza-acyclic nucleoside phosphonates containing a second phosphonate group as inhibitors of the human, Plasmodium falciparum and vivax 6-oxopurine phosphoribosyltransferases and their prodrugs as anti-malarial agents. J. Med. Chem. 2015, 58 (2), 827−846. (c) Keough, D. T.; Špaček, P.; Hocková, D.; Tichý, T.; Vrbková, S.; Slavětínská, L.; Janeba, Z.; Naesens, L.; Edstein, M. D.; Chavchich, M.; Wang, T. H.; de Jersey, J.; Guddat, L. W. Acyclic nucleoside phosphonates containing a second phosphonate group are potent inhibitors of 6-oxopurine phosphoribosyltransferases and have antimalarial activity. J. Med. Chem. 2013, 56 (6), 2513−2526. (6) Burkhart-Schultz, K. J.; Thompson, C. L.; Jones, I. M. Spectrum of somatic mutation at the hypoxanthine phosphoribosyltransferase (hprt) gene of healthy people. Carcinogenesis 1996, 17 (9), 1871− 1883. (7) Xu, Y.; Eads, J.; Sacchettini, J. C.; Grubmeyer, C. Kinetic mechanism of human hypoxanthine−guanine phosphoribosyltransferase: Rapid phosphoribosyl transfer chemistry. Biochemistry 1997, 36 (12), 3700−3712. (8) Munagala, N. R.; Chin, M. S.; Wang, C. C. Steady-state kinetics of the hypoxanthine-guanine-xanthine phosphoribosyltransferase from Tritrichomonas foetus: the role of threonine-47. Biochemistry 1998, 37 (12), 4045−4051. (9) Yuan, L.; Craig, S. P., 3rd; McKerrow, J. H.; Wang, C. C. Steadystate kinetics of the schistosomal hypoxanthine-guanine phosphoribosyltransferase. Biochemistry 1992, 31 (3), 806−10. (10) Wenck, M. A.; Medrano, F. J.; Eakin, A. E.; Craig, S. P. Steadystate kinetics of the hypoxanthine phosphoribosyltransferase from Trypanosoma cruzi. Biochim. Biophys. Acta 2004, 1700 (1), 11−18. (11) De Clercq, E.; Holý, A. Acyclic nucleoside phosphonates: A key class of antiviral drugs. Nat. Rev. Drug Discovery 2005, 4 (11), 928− 940. (12) Keough, D. T.; Hocková, D.; Holý, A.; Naesens, L. M. J.; Skinner-Adams, T. S.; de Jersey, J.; Guddat, L. W. Inhibition of hypoxanthine-guanine phosphoribosyltransferase by acyclic nucleoside phosphonates: A new class of antimalarial therapeutics. J. Med. Chem. 2009, 52 (14), 4391−4399. P

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

(26) Keough, D. T.; Brereton, I. M.; de Jersey, J.; Guddat, L. W. The crystal structure of free human hypoxanthine-guanine phosphoribosyltransferase reveals extensive conformational plasticity throughout the catalytic cycle. J. Mol. Biol. 2005, 351 (1), 170−181. (27) McPhillips, T. M.; McPhillips, S. E.; Chiu, H.-J.; Cohen, A. E.; Deacon, A. M.; Ellis, P. J.; Garman, E.; Gonzalez, A.; Sauter, N. K.; Phizackerley, R. P.; Soltis, S. M.; Kuhn, P. Blu-Ice and the Distributed Control System: Software for data acquisition and instrument control at macromolecular crystallography beamlines. J. Synchrotron. Radiat. 2002, 9 (6), 401−406. (28) Kabsch, W. Xds. Acta Crystallogr. D 2010, 66, 125−132. (29) McCoy, A. J.; Grosse-Kunstleve, R. W.; Adams, P. D.; Winn, M. D.; Storoni, L. C.; Read, R. J. Phaser crystallographic software. J. Appl. Crystallogr. 2007, 40, 658−674. (30) Adams, P. D.; Afonine, P. V.; Bunkoczi, G.; Chen, V. B.; Davis, I. W.; Echols, N.; Headd, J. J.; Hung, L. W.; Kapral, G. J.; GrosseKunstleve, R. W.; McCoy, A. J.; Moriarty, N. W.; Oeffner, R.; Read, R. J.; Richardson, D. C.; Richardson, J. S.; Terwilliger, T. C.; Zwart, P. H. PHENIX: a comprehensive Python-based system for macromolecular structure solution. Acta Crystallogr. D 2010, 66, 213−221. (31) Emsley, P.; Lohkamp, B.; Scott, W. G.; Cowtan, K. Features and development of Coot. Acta Crystallogr. D 2010, 66, 486−501. (32) Schüttelkopf, A. W.; van Aalten, D. M. F. PRODRG: a tool for high-throughput crystallography of protein-ligand complexes. Acta Crystallogr. D 2004, 60 (8), 1355−1363. (33) (a) Taneja, N. K.; Tyagi, J. S. Resazurin reduction assays for screening of anti-tubercular compounds against dormant and actively growing Mycobacterium tuberculosis, Mycobacterium bovis BCG and Mycobacterium smegmatis. J. Antimicrob. Chemother. 2007, 60 (2), 288− 293. (b) West, N. P.; Cergol, K. M.; Xue, M.; Randall, E. J.; Britton, W. J.; Payne, R. J. Inhibitors of an essential mycobacterial cell wall lipase (Rv3802c) as tuberculosis drug leads. Chem. Commun. 2011, 47 (18), 5166−5168. (34) Chen, T. R. Chromosome changes in 6-TG-resistant mutant strains derived from a karyotypically stable human line, C32. Cytogenet. Cell Genet. 1983, 35 (3), 181−189.

Q

DOI: 10.1021/acs.jmedchem.5b00611 J. Med. Chem. XXXX, XXX, XXX−XXX