molecules

Review Flavin-Dependent Thymidylate Synthase as a New Antibiotic Target

Michael Choi, Kalani Karunaratne and Amnon Kohen *

Department of Chemistry, The University of Iowa, Iowa City, IA 52242-1727, USA; [email protected] (M.C.); [email protected] (K.K.) * Correspondence: [email protected]; Tel.: +1-319-335-0234

Academic Editor: Jean Jacques Vanden Eynde Received: 6 April 2016; Accepted: 13 May 2016; Published: 20 May 2016

Abstract: In humans de novo synthesis of 21-deoxythymidine-51-monophosphate (dTMP), an essential building block of DNA, utilizes an enzymatic pathway requiring thymidylate synthase (TSase) and (DHFR). The flavin-dependent thymidylate synthase (FDTS) represents an alternative enzymatic pathway to synthesize dTMP, which is not present in human cells. A number of pathogenic bacteria, however, depend on this enzyme in lieu of or in conjunction with the analogous human pathway. Thus, inhibitors of this enzyme may serve as antibiotics. Here, we review the similarities and differences of FDTS vs. TSase including aspects of their structure and chemical mechanism. In addition, we review current progress in the search for inhibitors of flavin dependent thymidylate synthase as potential novel therapeutics.

Keywords: flavin; enzyme; mechanism; thymidylate synthase; antibiotic

1. Introduction Replication of all cells requires thymidine, a pyrimidine that makes up DNA. Thymidine is synthesized in human cell de novo. The last step of the synthesis is catalyzed by the enzyme thymidylate synthase (TSase) coded by the thyA gene. TSase catalyzes the conversion of 21-deoxyuridine-51-monophosphate (dUMP) to 21-deoxythymidine-51-monophosphate (dTMP or 5 10 thymidylate) using N ,N -methylene-5,6,7,8-tetrahydrofolate (MTHF or CH2H4folate) in a reductive reaction (Scheme1). Through this process, dTMP is made, which can then be used in DNA synthesis. Also as a result of the activity of TSase, MTHF is converted to dihydrofolate (DHF or H2folate). Since the availability of derivatives is a rate-limiting factor in cell replication, DHF must be recycled back into MTHF. The recycling of DHF is catalyzed by the folA encoded enzyme dihydrofolate reductase (DHFR), which synthesizes tetrahydrofolate (THF or H4folate) (Scheme1), which is then converted to MTHF by serine hydroxymethyltransferase [1]. The recycling of DHF and the involved in this cycle are clinically important. Pharmaceuticals which target TSase, such as 5-fluorouracil and [2], are important chemotherapeutics in skin, colon, ovarian, and other cancers [3]. The enzyme DHFR is also targeted by drugs like methotrexate for cancer, autoimmune disorders, and other illnesses [4,5]. De novo dTMP production is not only important for humans. The antibiotic trimethoprim is a bacteriostatic inhibitor of DHFR that has both Gram positive and Gram negative coverage [6]. Because of the importance of dTMP to the synthesis of DNA, it was thought that TSase and the enzymes required to regenerate MTHF were essential. However, in 2002, it was reported that a number of organisms did not have these systems in place [7]. Some of those organisms did not have the Tdk gene (encoding for thymidine ), thus could not even scavenge for thymidine in media or host.

Molecules 2016, 21, 654; doi:10.3390/molecules21050654 www.mdpi.com/journal/molecules Molecules 2016, 21, 654 2 of 10 Molecules 2016, 21, 654 2 of 9

SchemeScheme 1. Reactions 1. Reactions catalyzed catalyzed by by TSase TSaseand and DHFRDHFR ( toptop),), and and FDTS FDTS (bottom (bottom). Highlighted). Highlighted are the are the reducing hydrogen in the TSase reaction (green), methylene (blue). R, 2′-deoxyribose-5′- phosphate; reducing hydrogen in the TSase reaction (green), methylene (blue). R, 21-deoxyribose-51- phosphate; R′, (p-aminobenzoyl) glutamate; NADP+, nicotinamide adenine dinucleotide phosphate; NADPH, R1,(p-aminobenzoyl) glutamate; NADP+, nicotinamide adenine dinucleotide phosphate; NADPH, reduced form of NADP+. To demonstrate that these enzymes are not structurally related, the crystal reduced form of NADP+. To demonstrate that these enzymes are not structurally related, the crystal structure of each protein is presented above its name (PDB IDs: 2KCE, 1RX2, and 1O26, for TSase, structureDHFR, of and each FDTS, protein respectively). is presented above its name (PDB IDs: 2KCE, 1RX2, and 1O26, for TSase, DHFR, and FDTS, respectively). Instead, they appeared to have a different enzyme, which is encoded by the gene thyX. This newInstead, enzyme they was appeared found to to contain have aflavin different adenine enzyme, dinucleotide which (FAD) is encoded and use by nicotinamide the gene thyX adenine. This new enzymedinucleotide was found phosphate to contain (NADPH) flavin as adeninethe reducing dinucleotide agent. It catalyzes (FAD) and the useconversion nicotinamide of dUMP adenine to dinucleotidedTMP, and phosphate thus named (NADPH) flavin dependent as the thymidylate reducing agent. synthase Itcatalyzes (FDTS). There the conversionwas also no analog of dUMP of to dTMP,the andMTHF thus recycling named system flavin dependentfound in some thymidylate of these prokaryotes. synthase (FDTS). Altogether There it is was estimated also no that analog of theapproximately MTHF recycling 40% of system prokaryotes found carry in some the thyX of thesegene, though prokaryotes. some of Altogether these prokaryotes it is estimated also have that approximatelythe thyA and 40% folAof genes prokaryotes for TSase carryand DHFR, the thyX respectively.gene, though some of these prokaryotes also have Comparing TSase and DHFR to FDTS shows no sequence or structural homology [8–10]. the thyA and folA genes for TSase and DHFR, respectively. Importantly, a number of human pathogens have FDTS, including Bacillus anthracis, Mycobacterium Comparing TSase and DHFR to FDTS shows no sequence or structural homology [8–10]. tuberculosis, Helicobacter pylori, and all Rickettsia species (see Table 1) [7,11,12]. The lack of this Importantly,enzyme in a humans number and of humanthe important pathogens role that have thymidylate FDTS, including synthesis Bacillusplays could anthracis make ,FDTSMycobacterium a new tuberculosisantibiotic, Helicobacter target. Antibiotic pylori, andresistance all Rickettsia is increasingspecies in (seetuberculosis Table1)[ and7,11 emerging,12]. The in lack H. ofpylori this and enzyme in humansB. anthracis and, thewhich important make the role search that for thymidylate new classes synthesisof antibiotics plays more could pressing make [13–15]. FDTS a new antibiotic target. Antibiotic resistance is increasing in tuberculosis and emerging in H. pylori and B. anthracis, which makeTable the 1. Table search of forpathogenic new classes bacterial of species antibiotics containi moreng ThyX pressing as well [as13 whether–15]. (Tdk) and TSase (ThyA) or DHFR (FolA) are present [16,17]. Table 1. TableBacterial of pathogenic Species bacterial speciesAssociated containing ThyXDisease as (s) well as whetherTdk thymidineThyA FolA kinase (Tdk)Bacillus and TSase anthracis (ThyA) or DHFR (FolA) are present [anthrax16,17]. + + + Borrelia burgdorferi Lyme disease + − − CampylobacterBacterial jejuni Species Associateddiarrhea Disease (s) Tdk− ThyA− FolA − Chlamydia trachomatis trachoma − − + Bacillus anthracis anthrax + + + Chlamydia pneumoniae pneumonia − − + Borrelia burgdorferi Lyme disease + ´ ´ Clostridium botulinum botulism + + + Campylobacter jejuni diarrhea ´ ´ ´ Helicobacter pylori stomach ulcer, gastric cancer − − − Chlamydia trachomatis trachoma ´ ´ + Mycobacterium leprae leprosy − + + Chlamydia pneumoniae pneumonia ´ ´ + Mycobacterium tuberculosis tuberculosis − + + Clostridium botulinum botulism + + + Rickettsia prowazeki typhus − − − Helicobacter pylori stomach ulcer, gastric cancer ´ ´ ´ Rickettsia rickettsii spotted fever − − − Mycobacterium leprae leprosy ´ + + Treponema pallidum syphilis − − − Mycobacterium tuberculosis tuberculosis ´ + + Rickettsia prowazeki typhus ´ ´ ´ Rickettsia rickettsii spotted fever ´ ´ ´ Treponema pallidum syphilis ´ ´ ´ Molecules 2016, 21, 654 3 of 10

Molecules 2016, 21, 654 3 of 9 2. Thymidylate Synthase (TSase) vs. Flavin Dependent Thymidylate Synthase (FDTS) 2. Thymidylate Synthase (TSase) vs. Flavin Dependent Thymidylate Synthase (FDTS) Structurally, TSase is a homodimer (Scheme1)[ 18] with one in each monomer. FDTS Structurally, TSase is a homodimer (Scheme 1) [18] with one active site in each monomer. FDTS on the other hand is a homotetramer (Scheme1)[ 10,19] with four active sites, each one at the interface on the other hand is a homotetramer (Scheme 1) [10,19] with four active sites, each one at the interface of three of the monomers (Figure1). Crystallization of FDTS with different folate derivatives, FAD, of three of the monomers (Figure 1). Crystallization of FDTS with different folate derivatives, FAD, andand dUMP dUMP indicate indicate that that the the substrates substrates areare stackedstacked withinwithin the active active site site (Figure (Figure 1).1). The The FAD FAD molecule molecule is sandwichedis sandwiched between between the the dUMP dUMP and and folatefolate [[19,2019,20].]. Since this this arrangement arrangement places places the the methylene methylene donordonor far far away away from from the the dUMP dUMP acceptor, acceptor, it suggests it suggests an unusualan unusual mechanism mechanism with with respect respect to the to TSasethe reaction,TSase reaction, where methylene where methylene is transferred is transfer directlyred directly from folate from tofolate . to nucleotide.

Figure 1. Structure of the active site of FDTS from T. maritima with (magenta), FAD Figure 1. Structure of the active site of FDTS from T. maritima with folinic acid (magenta), FAD (yellow), (yellow), and dUMP (blue) (PDB ID 4GTA) [20]. The three monomers that combine this active site are and dUMP (blue) (PDB ID 4GTA) [20]. The three monomers that combine this active site are in green, in green, light blue and orange, and the carbon that is to be transferred is marked in black. light blue and orange, and the carbon that is to be transferred is marked in black. Because of the cycling of the flavin between reduced and oxidized states, both the oxidativeBecause and of reductive the cycling half of reactions the flavin should cofactor be addressed. between NADPH reduced and and other oxidized reducing states, agents both can the oxidativebe used andby FDTS reductive for reducing half reactions equivalents should during be addressed. the reductive NADPH half-reaction and other [16]. reducing Formation agents of the can beproduct used by dTMP FDTS foroccurs reducing during equivalents the oxidative during half reaction, the reductive and following half-reaction the oxidation [16]. Formation state of ofthe the productflavin dTMPspectroscopically occurs during indicates the oxidative that, it gets half oxid reaction,ized during and followingthe formation the of oxidation dTMP [21,22]. state ofLike the with flavin spectroscopicallyTSase, 14C radiolabeling indicates studies that, it done gets oxidizedon the methylene during thecarbon formation of MTHF of dTMP(C11, Scheme [21,22]. 2) Like demonstrate with TSase, 14Cthe radiolabeling incorporation studies of 14C into done the on dTMP the methylene product. However, carbon of unlike MTHF classical (C11, SchemeTSase, FDTS2) demonstrate does not rely the incorporationon MTHF for of the14C reducing into the hydride dTMP product.equivalent, However, thus no DHF unlike is classicalproduced, TSase, and no FDTS DHFR does is needed not rely in on MTHFFDTS for dependent the reducing organisms. hydride equivalent,Studies carried thus noout DHF with is [6- produced,3H]-MTHF and did no DHFRnot show is needed any tritium in FDTS dependentincorporation organisms. into the Studies dTMP carried product; out this with is [6-3H]-MTHFan evidence against did not FDTS show serving any tritium as a bifunctional incorporation intoenzyme the dTMP with product; both TSase this and is an DHFR evidence activities against [9]. FDTS This serving observation as abifunctional together with enzyme the structural with both TSasedifferences and DHFR and the activities presence [9]. of ThisFAD observationsuggested that together TSase and with FDTS the structural may have differencesdifferent catalytic and the presencemechanisms, of FAD which suggested prompted that mechanistic TSase and studies FDTS mayto compare have different these two catalytic enzymes. mechanisms, which promptedThe mechanistic mechanism studiesof TSase to (Scheme compare 2a) these has been two ex enzymes.tensively studied [1,23,24]. In this mechanism dUMPThe mechanismis activated by of an TSase enzymatic (Scheme nucleophile2a) has been cysteine extensively via Michael studied addition [ 1,23 to,24 the]. InC6 this of the mechanism dUMP (steps 1 and 2). The methylene is subsequently transferred to the C5 position of the covalently-activated dUMP is activated by an enzymatic nucleophile cysteine via Michael addition to the C6 of the dUMP dUMP (steps 3 and 4) followed by reduction (step 5), yielding DHF and dTMP. (steps 1 and 2). The methylene is subsequently transferred to the C5 position of the covalently-activated In the TSase reaction MTHF provides both the methylene group as well as the reducing hydride dUMP (steps 3 and 4) followed by reduction (step 5), yielding DHF and dTMP. to synthesize dTMP [1]. The chemotherapeutic 5- is a mechanism based inhibitor. In the In the TSase reaction MTHF provides both the methylene group as well as the reducing hydride cell it is metabolized to 5-fluoro dUMP, which reacts similarly as the enzyme’s native substrate, to synthesizedUMP, by being dTMP attacked [1]. The by chemotherapeutic the nucleophilic enzyme 5-fluorouracil side-chain, is a but mechanism the fluorine based on the inhibitor. C5 position In the cellis it unable is metabolized to leave (step to 5-fluoro 4, Scheme dUMP, 2a) resulting which reacts in a dead-end similarly asinhibitor the enzyme’s for the enzyme native substrate, [3]. dUMP, by beingBecause attacked the by first the chemical nucleophilic modifi enzymecation side-chain,of dUMP in but the the classical fluorine TSase on themechanism C5 position involves is unable a to leaveMichael (step addition 4, Scheme with2 ana) resultingenzymatic in nucleophile, a dead-end early-proposed inhibitor for the mechanisms enzyme [3 ].of FDTS included an analogousBecause scheme the first (Scheme chemical 2b). modification FDTS does not of have dUMP a conserved in the classical cysteine TSase residue mechanism within its active involves site a Michaellike TSase; addition instead, with a anconserved enzymatic serine nucleophile, residue was early-proposed proposed as the mechanisms Michael nucleophile. of FDTS includedHowever, an

Molecules 2016, 21, 654 4 of 10

analogousMolecules 2016 scheme, 21, 654 (Scheme 2b). FDTS does not have a conserved cysteine residue within its active4 of site9 like TSase; instead, a conserved serine residue was proposed as the Michael nucleophile. However, whenwhen that that serine serine was was mutated mutated to to alanine, alanine, inin vitro assays showed showed that that the the enzyme enzyme was was still still active active [19]. [19 ]. TheThe observation observation that that FDTS FDTS doesdoes notnot useuse anan enzymaticenzymatic nucleophile nucleophile further further suggested suggested a different a different mechanismmechanism from from that that of of TSase. TSase.

CH2H4folate H4folate H2folate H H H N N N H H H H2N N N 2 H N N N R' H2N N N H2N N N 2 6 H H HN HN N N R' R' HN

N H HN N HN N N

a) 11 O H N N O H H O HN OH2C N (1) CH2 O H O H H R' O R' (3) O CH2 (4) O (5) O H H CH CH HN (2) HN 5 HN HN 2 HN 3 6 HS S O N S O N S O N O N S O N dUMP R Cys 146 R Cys 146 R Cys 146 R Cys 146 dTMP R 146 Cys

CH2H4folate H4folate FADH2 FAD H H H H H R'' H R'' H2N N N H2N N N H N N N H N N N O N N O N N H 2 2 HN 6 H HN R' R' R' b) N N N HN HN HN HN 11 H H N N N N N N OH C N (1) O H H H H 2 CH2 O O H O H O H O R' O (3) O CH2 (4) O (5) O (6) O H H CH CH2 CH3 HN (2) HN 5 HN HN 2 HN HN 6 HO O N O O N O O N O O N O O N O O N R R Ser 88 R Ser 88 R R R dUMP Ser 88 Ser 88 Ser 88 dTMP 88 Ser

H FADH2 H R'' H H2N N N H N N N O N N R' 2 HN R' N N HN c) HN H H N N O O N O H H CH CH CH2 O HN 2 (4) HN 3 O O H (1) O (2) O CH2 (3) H H O N H O N HN HN 5 HN dTMP 6 FAD H4folate R R O N O N H O N H CH2H4folate R dUMP R R

FADH CH2H4folate 2 H R'' H H H H N N N H N N N O N N H2N N N 2 2 H R' R' 6 H HN HN HN HN N N N N N O O N H H H H 11 (1) O O O H CH2 CH3 OH2C N CH2 O CH2 (3) O (4) HN (5) HN d) O O (2) CH R' HN H HN 2 O N H O N H dTMP HN HN 5 H4folate FAD R R O N O N dUMP O N 6 O N R R R R Enz-FADH H 2 H R'' CH H folate H Enz-FADH 2 4 H2N N N O N N 2 H R'' H H2N N N R' H4folate O N N H2N N N R' HN HN HN N N N N e) R' N H H HN (1) HN N H H O (4a) O H N N H (2) OO CH (3) O CH2 O O H 2 O H CH2 H HN CH2 O O HN Enz-FADH HN Enz-FAD 2 H HN H O N H O N O N O N 6 HN 5 Enz-FAD R H O O N R R O N H N N dUMP R R N H R'' Enz-FAD (5) (4 b ) R = 2'-deoxyribose-5'-phosphate H folate R' = (p-aminobenzoyl)-glutamate O 4 O Enz-FAD R'' = adenosine-5'-pyrophosphate-ribityl CH3 CH HN HN 2 O N O N H R (6) dTMP R

SchemeScheme 2. 2. ProposedProposed chemical chemical mechanisms mechanisms of of(a) (aclassical) classical TSase; TSase; (b) FDTS (b) FDTS where where an active an activesite sitenucleophile nucleophile is activating is activating the dUMP the dUMP ring; (c) ring; FDTS ( cwhere) FDTS the wherereduced the flavin reduced acting flavinas the nucleophile; acting as the nucleophile;(d) FDTS mechanism (d) FDTS where mechanism dUMP whereis activated dUMP by enzy is activatedmatic polarization by enzymatic rather polarization than Michael rather addition than and (e) mechanism of FDTS which combines mechanisms (c,d) to explain substrate activation by the Michael addition and (e) mechanism of FDTS which combines mechanisms (c,d) to explain substrate reduced enzyme. R = 2′-deoxyribose-5′-phosphate, R′ = (p-aminobenzoyl)-glutamate, R″ = adenosine-5′- activation by the reduced enzyme. R = 21-deoxyribose-51-phosphate, R1 = (p-aminobenzoyl)-glutamate, pyrophosphate-ribityl [1,9,22,25,26]. R” = adenosine-51-pyrophosphate-ribityl [1,9,22,25,26].

Furthermore, it was observed that when the FDTS reaction was conducted in D2O at low Furthermore, it was observed that when the FDTS reaction was conducted in D2O at low temperature, 60% of the product, dTMP, had deuterium at position C6 [19]. Since in D2O the reduced temperature, 60% of the product, dTMP, had deuterium at position C6 [19]. Since in D O the reduced flavin is deuterated, it was proposed that the nucleophile activating the aromatic ring2 of dUMP by flavin is deuterated, it was proposed that the nucleophile activating the aromatic ring of dUMP by Michael addition might be the hydride from the FADH2 (Scheme 2c). In that proposed mechanism, Michaelfirst, the addition hydride might from bethe the reduced hydride flavin from does the a FADH Michael2 (Scheme attack on2c). C6 In of that dUMP proposed followed mechanism, by the first,methylene the hydride transfer from to the the C5 reduced position. flavin Finally, does either a Michaela 1,3-hydride attack shift on or C6 proton of dUMP addition-elimination followed by the methylenemechanism transfer is needed to the to C5 yield position. dTMP. Finally, The possibility either a 1,3-hydrideof N5 of the shift flavin or acting proton as addition-elimination a nucleophile for the Michael addition at C6 was refuted because it was reported that FDTS reconstituted with

Molecules 2016, 21, 654 5 of 10 mechanism is needed to yield dTMP. The possibility of N5 of the flavin acting as a nucleophile for the Michael addition at C6 was refuted because it was reported that FDTS reconstituted with 5-deazaflavin was active [19]. This is because the analogous carbon position in 5-deazaflavin cannot act as a nucleophile. However, spectrophotometric experiments following the redox state of the flavin along with radiolabeling have demonstrated that the flavin oxidation occurs late in the reaction [21,22]. This observation lead to another proposed mechanism, one that involves no Michael addition at all. Instead, dUMP is activated by polarization (Scheme2d) [ 20,22,26]. In such mechanism, the accumulated negative charge on C5 of the polarized dUMP nucleophilically attacks the Schiff base (step 2 in Scheme2d) followed by deprotonation and elimination of the THF. The exocyclic methylene dUMP is then reduced by the flavin (step 4) and undergo a 1,3-hydride shift to form dTMP (step 5) [22]. In order to further understand the mechanism of FDTS, intermediates of the oxidative half reaction (where flavin is oxidized and dUMP is converted to dTMP) were trapped under single-turnover conditions. FDTS single turnover experiments quenched by acid showed the accumulation of 7-hydroxymethyl-dUMP which is generated from water addition after the methylene transfer to C5 of dUMP [21]. Although only one intermediate was trapped, the broad peak of accumulation of the acid-quenched intermediate suggested that more than one intermediate was trapped in acid as 7-hydroxymethyl-dUMP [26]. Independent experiments have also suggested two intermediates due to the observation of a spectral flavin change as the reaction progresses followed by the transient disappearance of reactant and product [22]. Subsequent experiments [26] demonstrated that incubation of dUMP with reduced FDTS in D2O exchanges the hydrogen at the C5 position. This does not occur with oxidized FDTS. Also, when the reaction is done in D2O, dTMP becomes monodeuterated in the C6 or C7 position, but the acid-trapped intermediate was found to have no deuterium. A new mechanism (Scheme2e) was proposed to reconcile this new data. Hydride transfer from the N5 of the reduced flavin occurs at the C6 position prior to methylene transfer to C5. It is then proposed that elimination occurs in step 3 by FAD. Steps 4 through 6 describe hydride transfer and elimination of H4folate either through a step-wise or concerted mechanism followed by a 1,3 hydride shift to yield dTMP [26]. Mechanism 2e also accords with the stopped-flow experiment where flavin oxidation happen close to dTMP formation rather than early in the reaction [21,22].

3. The Latest Proposed Mechanism for FDTS (April 2016) Recently, new evidence regarding the mechanism of FDTS was acquired through quenching the reaction in base instead of acid, which yielded a different intermediate. Mass spectrometry showed the base-trapped intermediate to be [M ´ H]´ 705.1212 Dalton. Isotopic labeling and NMR studies elucidated the structure to be dUMP with a methylene bridge to a derivative of the isoalloxazine ring from the FAD (Scheme3a). The trapping of the labeled methylene carbon from methylene tetrahydrofolate suggests that the isoalloxazine ring of the flavin may play a part as a methylene relay system from the MTHF to the dUMP. Given the finding that the isoalloxazine ring appears to act as a methylene relay system, and based on the structure presented in Scheme3a, a new mechanism has been proposed (Scheme3b). This new mechanism involves an initial Schiff base formation of the MTHF followed by methylene transfer to the N5 of the flavin (steps 1 and 2). Polarization of the dUMP substrate and transfer of the methylene group to C5 of the dUMP leads to a bridged methylene intermediate, I1 (steps 2 and 3). A derivative of this intermediate is the bridged intermediate isolated from base quenching of FDTS (Scheme3a). Methylene transfer to dUMP (steps 4 and 5) followed by flavin oxidation and I 2 reduction to dTMP (step 6) and oxidized flavin. NADPH then reduces the flavin during the reductive half reactions, and the cycle continues. The relay of the methylene group from folate to flavin and ultimately to the dUMP accords well with the crystal structure where the flavin cofactor is between folate and dUMP (Scheme3b) [ 17,20]. Why was this mechanism not proposed in 2012 when the structure was solved (Figure1). Because of the report that FDTS with FAD substituted for 5-deaza-FAD Molecules 2016, 21, 654 6 of 10 is still active, and thus N5 of FAD cannot serve in the role proposed in Scheme3b (step 1). Extensive recent efforts [17] found that actually FDTS with 5-deaza-FAD is not active, and the residual activity Molecules 2016, 21, 654 6 of 9 reported in [19] was from residual FAD left in the protein. Step 6 indicates reduction of I2 at C6, toisomer non-productive if dTMP. isomerThis principal if dTMP. mechanism This principal while mechanism not very detailed, while notis in very accordance detailed, with is in all accordance current withobservations, all current but observations, one: the deuterium but one: found the deuterium at C6 of dTMP found at atlow C6 temperature of dTMP at [19]. low That temperature observation, [19 ]. Thathowever, observation, might result however, from might reversible result but from non-productive reversible but reduction non-productive of I2 at C6, reduction followed of by I2 atthe C6, followedproductive by theirreversible productive reduction irreversible at C7, leading reduction to the at product C7, leading dTMP. to the product dTMP.

(a)

(b)

Scheme 3. (a) Trapped intermediate derivative from quench-flow experiments of FDTS with base Scheme 3. (a) Trapped intermediate derivative from quench-flow experiments of FDTS with base identified through NMR, MS, and labeling experiments. The derivative is composed of dUMP (U), a identified through NMR, MS, and labeling experiments. The derivative is composed of dUMP (U), methylene from MTHF (7) and a flavin derivative (F); (b) New proposed mechanism of FDTS (see a methylene from MTHF (7) and a flavin derivative (F); (b) New proposed mechanism of FDTS (see text). I1 is the intermediate before base-induced breakdown of the isoalloxazine ring, and I2 is the text). I1 is the intermediate before base-induced breakdown of the isoalloxazine ring, and I2 is the intermediate that is reduced to dTMP upon quenching with base. Both I1 and I2 undergo water intermediate that is reduced to dTMP upon quenching with base. Both I and I undergo water addition addition under acidic conditions to form 7-hydroxymethyl-dUMP. Modified1 from2 reference [17]. under acidic conditions to form 7-hydroxymethyl-dUMP. Modified from reference [17]. In the reductive half-reaction of the enzyme, NADPH reacts with the bound FAD of the enzyme to Inreduce the reductive the cofactor. half-reaction The initial ofbinding the enzyme, of dUMP NADPH to the reactsenzyme with greatly the boundenhances FAD the of rate the of enzyme the toreduction reduce the while cofactor. the presence The initial of MTHF binding [27] of or dUMP its analogue, to the enzymefolinic acid greatly [20], enhancesreduces the the rate rate of ofthe the reductionreductive while half thereaction, presence indicating of MTHF an overlapping [27] or its analogue, folinic of NADPH acid [20 ],and reduces . the This rate also of the reductivesuggests halfan ordered reaction, binding indicating mechanism an overlapping where the enzyme binding first site binds of NADPHto dUMP, andthe NADPH folates. reduces This also suggeststhe flavin, an orderedfollowed binding by NADP mechanism+ leaving with where thethe MTHF enzyme binding first in binds the vacant to dUMP, site. the NADPH reduces the flavin, followed by NADP+ leaving with the MTHF binding in the vacant site. 4. FDTS Specific Inhibitors 4. FDTS Specific Inhibitors Many inhibitors of TSase are in laboratory and clinical use, but specific and efficient inhibitors of FDTSMany (Ki inhibitorsat the low ofnM TSase range) are are in not laboratory yet reported and clinicalnor in clinical use, but use. specific Attractive and classes efficient of inhibitorsantibiotic of FDTSdrugs (K iagainstat the lowFDTS nM dependent range) are orga notnisms yet reported need to nor exhibit in clinical high specificity use. Attractive toward classes FDTS of and antibiotic low drugsspecificity against toward FDTS human dependent TSase organismsin order to reduce need to toxicity. exhibit High high throughput specificity screening toward FDTS studies and and low synthesis of inhibitors based on substrate scaffolds have been attempted, and some of these inhibitors achieve an IC50 in the 100 nM range against FDTS (Table 2) [28–31].

Molecules 2016, 21, 654 7 of 10 specificity toward human TSase in order to reduce toxicity. High throughput screening studies and synthesis of inhibitors based on substrate scaffolds have been attempted, and some of these inhibitors achieve an IC50 in the 100 nM range against FDTS (Table2)[28–31]. Molecules 2016, 21, 654 7 of 10 MoleculesMolecules 20162016,, 2121,, 654654 77 ofof 1010 Table 2. Current inhibitors and their IC50 against FDTS and TSase [28,29,31]. Table 2. Current inhibitors and their IC50 against FDTS and TSase [28,29,31]. TableTable 2.2. CurrentCurrent inhibitorsinhibitors andand theirtheir ICIC5050 againstagainst FDTSFDTS andand TSaseTSase [28,29,31].[28,29,31]. ThyA IC50 EntryEntry StructureStructure ThyX ThyX IC50 (µM) IC50 (µM)ThyAThyA ICIC ThyA5050 IC50 (µM) EntryEntry StructureStructure ThyX ThyX ICIC5050 (µM)(µM) (µM) (µM)(µM)

1 1 0.91 0.91>50 >50 11 0.910.91 >50>50

43% inhibition at 50 2 43%43% inhibitioninhibition atat 5050 22 µM µ 2 43% inhibitionµMµM at 50 M

3 0.13 3 33 0.130.13 0.13

4 0.057 4 44 0.0570.057 0.057

Encouragingly, lipophilic derivatives of compound 1 in Table 2 have been shown to be inhibitory Encouragingly,Encouragingly, lipophiliclipophilic derivativesderivatives ofof compoundcompound 11 inin TableTable 22 havehave beenbeen shownshown toto bebe inhibitoryinhibitory to the growth of Mycobacterium species as well as herpes viruses, and its inhibition of classical TSase toto thethe growthgrowth ofof MycobacteriumMycobacterium speciesspecies asas wellwell asas herpesherpes viruses,viruses, andand itsits inhibitioninhibition ofof classicalclassical TSaseTSase Encouragingly,was much weaker lipophilic than for derivativesFDTS [30]. Compound of compound 3’s inhibitory1 in activity Table against2 have FDTS been is reversed shown by to be inhibitory waswas muchmuch weakerweaker thanthan forfor FDTSFDTS [30].[30]. CompoundCompound 33’s’s inhibitoryinhibitory activityactivity againstagainst FDTSFDTS isis reversedreversed byby to the growthexcess of dUMP,Mycobacterium but compoundspecies 4’s is not as [31]. well However, as herpes for many viruses, of the and proposed its inhibition inhibitors, the of classical TSase excessexcess dUMP,dUMP, butbut compoundcompound 44’s’s isis notnot [31].[31]. However,However, forfor manymany ofof thethe proposedproposed inhibitors,inhibitors, thethe specificity towards FDTS over TSase, the efficacy of these compounds against bacterial pathogens, was muchspecificityspecificity weaker towardstowards than for FDTSFDTS FDTS overover TSase,TSase, [30]. thethe Compound efficacyefficacy ofof thesethese3’s compoundscompounds inhibitory againstagainst activity bacterialbacterial against pathogens,pathogens, FDTS is reversed and the toxicity towards human cells remain to be further tested. Possible inhibitors to investigate in andand thethe toxicitytoxicity towardstowards humanhuman cellscells remainremain toto bebe fufurtherrther tested.tested. PossiblePossible inhibitorsinhibitors toto investigateinvestigate inin by excessthe dUMP, future butmay include compound folate based4’s ismolecules not [31 that]. ta However,ke into account for the many nucleophilicity of the proposedof the flavin inhibitors, the thethe futurefuture maymay includeinclude folatefolate basedbased moleculesmolecules thatthat tatakeke intointo accountaccount thethe nucleophilicitynucleophilicity ofof thethe flavinflavin specificitycofactor towards towards FDTS the overmethylene TSase, of MTHF the efficacy to covalently of these trap its compounds N5. For example, against electrophilic bacterial pathogens, cofactorcofactor towardstowards thethe methylenemethylene ofof MTHFMTHF toto covalentlycovalently traptrap itsits N5.N5. ForFor example,example, electrophilicelectrophilic substituents like epoxides at N5 of folate may be able to create dead-end folate-flavin adducts [17]. and the toxicitysubstituentssubstituents towards likelike epoxidesepoxides human atat N5N5 cells ofof folatefolate remain maymay bebe to ableable be toto further createcreate dead-enddead-end tested. folate-flavinfolate-flavin Possible adductsadducts inhibitors [17].[17]. to investigate An alternative strategy may involve a similarly nucleophilic “warhead” on C5 of dUMP, but those AnAn alternativealternative strategystrategy maymay involveinvolve aa similarlysimilarly nunucleophiliccleophilic “warhead”“warhead” onon C5C5 ofof dUMP,dUMP, butbut thosethose in the futurestudies may are still include in their folateinfancy. based molecules that take into account the nucleophilicity of the studiesstudies areare stillstill inin theirtheir infancy.infancy. flavin cofactorstudies towards are still in their the methyleneinfancy. of MTHF to covalently trap its N5. For example, electrophilic 5. Conclusions substituents5.5. ConclusionsConclusions like epoxides at N5 of folate may be able to create dead-end folate-flavin adducts [17].

An alternative strategy may involve a similarly nucleophilic “warhead” on C5 of dUMP, but those studies are still in their infancy.

5. Conclusions Mechanistic and structural studies of FDTS have demonstrated a catalytic path that is very different from that of classical TSase. Most notably, the use of flavin as the reducing equivalent and as relay system for both the electrons and methylene, and the lack of an enzymatic nucleophile makes the Molecules 2016, 21, 654 8 of 10 chemistry of FDTS dramatically different from human TSase. The fact that several human pathogens depend on FDTS and the differences between its mechanism to that of human TSase and DHFR, makes it an excellent target for new classes of antibiotics. Pathogens depending on FDTS include organisms that have developed antibiotic resistance, such as multi-drug-resistance tuberculosis (MDR TB). The unprecedented mechanism of FDTS gives hope that specific mechanism-based inhibitors might serve as leads toward synthesis of non-toxic antibiotic drugs. Designing such inhibitors require extensive work in both computational and experimental fields. Additional studies on the mechanistic details of FDTS will provide a basis for the rational design of mechanism-based-inhibitors that may serve as leads to a new class of antibiotics. Finally, the unique chemistry that has so far been elucidated will also provide insight into alternative mechanisms as well as understanding the reactivity of flavins.

Acknowledgments: The work has been funded by NIH RO1 GM110775 to AK, the Iowa Center for Biocatalysis and Bioprocessing Training grant to KK, NIHT32-GM008365 and the University of Iowa Medical Scientist Training Program support NIH 5 T32 GM007337 to MC. Conflicts of Interest: The authors declare no conflicts of interest.

Abbreviations The following abbreviations are used in this manuscript: dTMP 21-deoxythymidine-51-monophosphate dUMP 21-deoxyuridine-51-monophosphate TSase Thymidylate synthase DHFR Dihydrofolate reductase FDTS Flavin dependent thymidylate synthase 5 10 MTHF, CH2H4folate N ,N -methylene-5,6,7,8,-tetrahydrofolate DHF, H2folate Dihydrofolate THF, H4folate Tetrahydrofolate FAD Flavin adenine dinucleotide NADPH Nicotinamide adenine dinucleotide phosphate

References

1. Carreras, C.W.; Santi, D.V. The catalytic mechanism and structure of thymidylate synthase. Annu. Rev. Biochem. 1995, 64, 721–762. [CrossRef][PubMed] 2. Kelly, C.; Bhuva, N.; Harrison, M.; Buckley, A.; Saunders, M. Use of raltitrexed as an alternative to 5-fluorouracil and in cancer patients with cardiac history. Eur. J. Cancer 2013, 49, 2303–2310. [CrossRef][PubMed] 3. Longley, D.B.; Harkin, D.P.; Johnston, P.G. 5-fluorouracil: Mechanisms of action and clinical strategies. Nat. Rev. Cancer 2003, 3, 330–338. [CrossRef][PubMed] 4. Jolivet, J.; Cowan, K.H.; Curt, G.A.; Clendeninn, N.J.; Chabner, B.A. The pharmacology and clinical use of methotrexate. N. Engl. J. Med. 1983, 309, 1094–1104. [CrossRef][PubMed] 5. Cronstein, B.N. Low-dose methotrexate: A mainstay in the treatment of rheumatoid arthritis. Pharmacol. Rev. 2005, 57, 163–172. [CrossRef][PubMed] 6. Kielhofner, M.A. Trimethoprim- sulfamethoxazole: Pharmacokinetics, clinical uses, and adverse reactions. Tex. Heart Inst. J. 1990, 17, 86–93. [PubMed] 7. Myllykallio, H.; Lipowski, G.; Leduc, D.; Filee, J.; Forterre, P.; Liebl, U. An alternative flavin-dependent mechanism for thymidylate synthesis. Science 2002, 297, 105–107. [CrossRef][PubMed] 8. Giladi, M.; Bitan-Banin, G.; Mevarech, M.; Ortenberg, R. Genetic evidence for a novel thymidylate synthase in the halophilic archaeon Halobacterium salinarum and in Campylobacter jejuni. FEMS Microbiol Lett 2002, 216, 105–109. [CrossRef][PubMed] 9. Agrawal, N.; Lesley, S.A.; Kuhn, P.; Kohen, A. Mechanistic studies of a flavin-dependent thymidylate synthase. Biochemistry 2004, 43, 10295–10301. [CrossRef][PubMed] 10. Kuhn, P.; Lesley, S.A.; Mathews, I.I.; Canaves, J.M.; Brinen, L.S.; Dai, X.; Deacon, A.M.; Elsliger, M.A.; Eshaghi, S.; Floyd, R.; et al. Crystal structure of thy1, a thymidylate synthase complementing protein from Thermotoga maritima at 2.25 A resolution. Proteins 2002, 49, 142–145. [CrossRef][PubMed] Molecules 2016, 21, 654 9 of 10

11. Liu, X.Q.; Yang, J. Bacterial thymidylate synthase with intein, group II Intron, and distinctive ThyX motifs. J. Bacteriol. 2004, 186, 6316–6319. [CrossRef][PubMed] 12. Fivian-Hughes, A.S.; Houghton, J.; Davis, E.O. Mycobacterium tuberculosis thymidylate synthase gene thyX is essential and potentially bifunctional, while thyA deletion confers resistance to p-aminosalicylic acid. Microbiology 2012, 158, 308–318. [CrossRef][PubMed] 13. Gandhi, N.R.; Moll, A.; Sturm, A.W.; Pawinski, R.; Govender, T.; Lalloo, U.; Zeller, K.; Andrews, J.; Friedland, G. Extensively drug-resistant tuberculosis as a cause of death in patients co-infected with tuberculosis and HIV in a rural area of South Africa. Lancet 2006, 368, 1575–1580. [CrossRef] 14. Graham, D.Y.; Fischbach, L. Helicobacter pylori treatment in the era of increasing antibiotic resistance. Gut 2010, 59, 1143–1153. [CrossRef][PubMed] 15. Athamna, A.; Athamna, M.; Abu-Rashed, N.; Medlej, B.; Bast, D.J.; Rubinstein, E. Selection of Bacillus anthracis isolates resistant to antibiotics. J. Antimicrob. Chemother. 2004, 54, 424–428. [CrossRef][PubMed] 16. Leduc, D.; Graziani, S.; Meslet-Cladiere, L.; Sodolescu, A.; Liebl, U.; Myllykallio, H. Two distinct pathways for thymidylate (dTMP) synthesis in (hyper)thermophilic Bacteria and Archaea. Biochem. Soc. Trans. 2004, 32, 231–235. [CrossRef][PubMed] 17. Mishanina, T.V.; Yu, L.; Karunaratne, K.; Mondal, D.; Corcoran, J.M.; Choi, M.A.; Kohen, A. An unprecedented mechanism of nucleotide methylation in organisms containing thyX. Science 2016, 351, 507–510. [CrossRef][PubMed] 18. Hardy, L.W.; Finer-Moore, J.S.; Montfort, W.R.; Jones, M.O.; Santi, D.V.; Stroud, R.M. Atomic structure of thymidylate synthase: target for rational drug design. Science 1987, 235, 448–455. [CrossRef][PubMed] 19. Koehn, E.M.; Fleischmann, T.; Conrad, J.A.; Palfey, B.A.; Lesley, S.A.; Mathews, I.I.; Kohen, A. An unusual mechanism of thymidylate in organisms containing the thyX gene. Nature 2009, 458, 919–923. [CrossRef][PubMed] 20. Koehn, E.M.; Perissinotti, L.L.; Moghram, S.; Prabhakar, A.; Lesley, S.A.; Mathews, I.I.; Kohen, A. Folate binding site of flavin-dependent thymidylate synthase. Proc. Natl. Acad. Sci. USA 2012, 109, 15722–15727. [CrossRef][PubMed] 21. Mishanina, T.V.; Koehn, E.M.; Conrad, J.A.; Palfey, B.A.; Lesley, S.A.; Kohen, A. Trapping of an intermediate in the reaction catalyzed by flavin-dependent thymidylate synthase. J. Am. Chem. Soc. 2012, 134, 4442–4448. [CrossRef][PubMed] 22. Conrad, J.A.; Ortiz-Maldonado, M.; Hoppe, S.W.; Palfey, B.A. Detection of intermediates in the oxidative half-reaction of the FAD-dependent thymidylate synthase from Thermotoga maritima: Carbon transfer without covalent pyrimidine activation. Biochemistry 2014, 53, 5199–5207. [CrossRef][PubMed] 23. Stroud, R.M.; Finer-Moore, J.S. Conformational dynamics along an enzymatic reaction pathway: Thymidylate synthase, “the Movie”. Biochemistry 2003, 42, 239–247. [CrossRef][PubMed] 24. Finer-Moore, J.S.; Santi, D.V.; Stroud, R.M. Lessons and conclusions from dissecting the mechanism of a bisubstrate enzyme: Thymidylate synthase mutagenesis, function and structure. Biochemistry 2003, 42, 248–256. [CrossRef][PubMed] 25. Koehn, E.M.; Kohen, A. Flavin-dependent thymidylate synthase: A novel pathway towards thymine. Arch. Biochem. Biophys. 2010, 493, 96–102. [CrossRef][PubMed] 26. Mishanina, T.V.; Corcoran, J.M.; Kohen, A. Substrate activation in flavin-dependent thymidylate synthase. J. Am. Chem. Soc. 2014, 136, 10597–10600. [CrossRef][PubMed] 27. Graziani, S.; Xia, Y.; Gurnon, J.R.; van Etten, J.L.; Leduc, D.; Skouloubris, S.; Myllykallio, H.; Liebl, U. Functional analysis of FAD-dependent thymidylate synthase ThyX from Paramecium bursaria Chlorella virus-1. J. Biol. Chem. 2004, 279, 54340–54347. [CrossRef][PubMed] 28. Parchina, A.; Froeyen, M.; Margamuljana, L.; Rozenski, J.; De Jonghe, S.; Briers, Y.; Lavigne, R.; Herdewijn, P.; Lescrinier, E. Discovery of an acyclic nucleoside phosphonate that inhibits Mycobacterium tuberculosis ThyX based on the binding mode of a 5-alkynyl substrate analogue. Chem. Med. Chem. 2013, 8, 1373–1383. [CrossRef][PubMed] 29. Kogler, M.; Vanderhoydonck, B.; de Jonghe, S.; Rozenski, J.; van Belle, K.; Herman, J.; Louat, T.; Parchina, A.; Sibley, C.; Lescrinier, E.; et al. Synthesis and evaluation of 5-substituted 21-deoxyuridine monophosphate analogues as inhibitors of flavin-dependent thymidylate synthase in Mycobacterium tuberculosis. J. Med. Chem. 2011, 54, 4847–4862. [CrossRef][PubMed] Molecules 2016, 21, 654 10 of 10

30. McGuigan, C.; Derudas, M.; Gonczy, B.; Hinsinger, K.; Kandil, S.; Pertusati, F.; Serpi, M.; Snoeck, R.; Andrei, G.; Balzarini, J.; et al. ProTides of N-(3-(5-(21-deoxyuridine))prop-2-ynyl)octanamide as potential anti-tubercular and anti-viral agents. Bioorg. Med. Chem. 2014, 22, 2816–2824. [CrossRef][PubMed] 31. Esra Onen, F.; Boum, Y.; Jacquement, C.; Spanedda, M.V.; Jaber, N.; Scherman, D.; Myllykallio, H.; Herscovici, J. Design, synthesis and evaluation of potent thymidylate synthase X inhibitors. Bioorg. Med. Chem. Lett. 2008, 18, 3628–3631. [CrossRef][PubMed]

© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).