Quick viewing(Text Mode)

NAD Analogs in Aid of Chemical Biology and Medicinal Chemistry

NAD Analogs in Aid of Chemical Biology and Medicinal Chemistry

molecules

Review NAD Analogs in Aid of Chemical Biology and Medicinal

Anais Depaix and Joanna Kowalska * Division of Biophysics, Institute of Experimental Physics, Faculty of Physics, University of Warsaw, Pasteura 5, 02-093 Warsaw, Poland; [email protected] * Correspondence: [email protected]; Tel.: +48-22-55-43774

 Academic Editor: Pawel Kafarski  Received: 2 October 2019; Accepted: 12 November 2019; Published: 19 November 2019 Abstract: dinucleotide (NAD) serves as an essential co-factor and mediator of multiple biological processes. Besides its well-established role in electron transfer reactions, NAD serves as a for other biotransformations, which, at the molecular level, can be classified as post-translational modifications (protein deacylation, mono-, and polyADP-ribosylation) and formation of signaling molecules (e.g., cyclic ADP ). These biochemical reactions control many crucial biological processes, such as cellular signaling and recognition, DNA repair and epigenetic modifications, stress response, immune response, aging and senescence, and many others. However, the links between the biological effects and underlying molecular processes are often poorly understood. Moreover, NAD has recently been found to tag the 50-ends of some cellular , but the function of these NAD-capped RNAs remains largely unrevealed. Synthetic NAD analogs are invaluable molecular tools to detect, monitor, structurally investigate, and modulate activity of NAD-related and biological processes in order to aid their deeper understanding. Here, we review the recent advances in the design and development of NAD analogs as probes for various cellular NAD-related enzymes, enzymatic inhibitors with anticancer or antimicrobial therapeutic potential, and other NAD-related chemical biology tools. We focus on research papers published within the last 10 years.

Keywords: nicotinamide-adenine dinucleotide; probe design; synthetic analogs; inhibitors

1. Introduction NAD is one of the most important and ubiquitous cofactors present both in prokaryotic and eukaryotic organisms. The essential role of NAD is the participation in cellular redox processes maintained by NAD+/NADH interconversion (Figure1A). NAD can be phosphorylated by NAD to NADP, which is another redox utilized in biosynthetic pathways and protection against oxidants [1]. In addition to the role in -catalyzed electron-transfer reactions, NAD is utilized by various enzymes to modify cellular biopolymers or produce signaling molecules (Figure1B) [ 2–4]. For example, Poly (ADP-ribose) (PARPs) (also referred to as ADP-ribosyl , ARTs) use NAD+ to post-translationally modify by addition of poly-ADP-ribose (PAR) chains in response to DNA damage. PARP inhibitors have potential as anticancer and chemo-sensitizing [5,6]. Another class of disease-associated NAD-dependent enzymes are , which utilize NAD+ to remove acyl groups from proteins such as histones and transcription factors, thereby regulating gene expression [7]. Some sirtuins, along with other enzymes, possess mono-ADP-ribosyltransferase activity. Cyclic ADP-ribose (cADPR), one of the key secondary messengers in signaling, is produced from NAD+ by ADP-ribosyl cyclase (CD38 family), which is an ectoenzyme involved, among others, in immune response and adhesion. CD38 has also NAD glycohydrolase activity (referred to also as NAD or NADase), i.e., it catalyzes NAD+ through

Molecules 2019, 24, 4187; doi:10.3390/molecules24224187 www.mdpi.com/journal/molecules Molecules 2019, 24, 4187 2 of 34 glycosidic bond cleavage leading to release of nicotinamide (Nam) and ADP-ribose, thereby additionally contributing to NAD turnover and signaling pathways. NAD is also a target for intracellular , such as bacterial NudC or mammalian NUDT12 , which contribute to the control of endogenous NAD levels [8]. Most recently, NAD has been found to serve as a ‘ cap’ present at the 50 end of some cellular RNAs in various organisms, from to higher mammals [9–17]. The modification is introduced to RNA co-transcriptionally by RNA polymerases or synthesized de novo using nicotinamide mononucleotide (NMN) [18–21]. Although the biological function of NAD-capped RNAs remains largely unclear, an increasing number of literature reports indicate that NAD tag strongly influences RNA stability and is targeted by specific cellular hydrolases [22,23]. Molecules 2019, 24, x 3 of 34

FigureFigure 1. Biological 1. Biological roles roles of NADof NAD in .in eukaryotes. ( AA)) TheThe NAD++/NADH/NADH pair pair commonly commonly involved involved in in cellularcellular redox redox transformations; transformations; atom atom numbering numbering is is shown shown onon NADNAD++; ;(B)B Select) Select enzymatic enzymatic reactions reactions consumingconsuming NAD NAD+ leading+ leading to to miscellaneousmiscellaneous NAD NAD derivatives derivatives and andmetabolites (labelled (labelled in bold). inThe bold). The enzymesenzymes catalyzingcatalyzing NAD-dependen NAD-dependentt reactions reactions are are labelled labelled in ingreen green and and the thebiological biological processes processes involvinginvolving or aff orected affected by theirby their products products are are marked marked in in blue.blue.

2. Fluorescent NAD Analogs for Real-Time Monitoring of NAD-Processing Enzymes Cofactors and cofactor analogs with distinctive optical properties have paved the way for many important applications, including monitoring enzymatic transformations in real-time, developing inhibitor discovery assays, and studying cofactor-related covalent protein modifications in vitro and in vivo. Both NAD+ and NADH absorb in the UV region, but differ in absorptive and emissive properties. NAD+ absorbs with a single maximum around 260 nm, whereas NADH has additionally a second absorption band with maximum around 340 nm corresponding to the reduced form of

Molecules 2019, 24, 4187 3 of 34

This multitude of NAD-dependent molecular processes orchestrates a variety of complex biological functions controlling DNA repair, cell survival and apoptosis, response to stress, circadian rhythm, aging and longevity, immune response, and many others [2,3,24–27]. Consequently, there is a strong demand for NAD-derived molecular tools that could aid the efforts towards deeper understanding of these biological pathways as well as prompt the development of new therapeutic strategies. Indeed, many synthetic NAD analogs have been designed as reporter substrates or modulators for NAD-dependent enzymes with therapeutic potential. For instance, the inhibition of anabolic pathways leading to generation of NAD and NADP is explored as an anti-cancer and antimicrobial strategy, whereas of those pathways may be a viable protective strategy to mitigate oxidative damage and age-related degenerative diseases. Here, we review the recent advances in the development of molecular tools and therapeutic candidates derived from or inspired by the NAD structure. To this end, we focus on five main areas that, in our opinion, have been the driving force for the development of NAD analogs in the past decade:

NAD analogs as fluorescent probes for real-time monitoring of miscellaneous NAD-processing enzymes • NAD analogs for the detection and visualization of polyADPribose activity • Potential therapeutic agents derived from NAD structure • NAD analogs for the development of biorthogonal redox systems • Methods and tools to advance our understanding of NAD-capped RNAs • We conclude our review by providing a short personal perspective on the future challenges and opportunities for the field.

2. Fluorescent NAD Analogs for Real-Time Monitoring of NAD-Processing Enzymes Cofactors and cofactor analogs with distinctive optical properties have paved the way for many important applications, including monitoring enzymatic transformations in real-time, developing inhibitor discovery assays, and studying cofactor-related covalent protein modifications in vitro and in vivo. Both NAD+ and NADH absorb in the UV region, but differ in absorptive and emissive properties. NAD+ absorbs with a single maximum around 260 nm, whereas NADH has additionally a second absorption band with maximum around 340 nm corresponding to the reduced form of nicotinamide riboside. Moreover, NADH has emissive properties with maximum at 460 nm, while NAD+ is virtually non-fluorescent. These differences in spectroscopic properties have been widely applied to monitor the enzymatic activity of NAD-dependent redox enzymes. The changes in fluorescent properties of NADH have also been exploited to study protein binding in vitro and in live cells [28–30]. However, the relatively low quantum yield of NADH (~2%) [31] and lack of emission for NAD+ exclude applications in the investigation of many other NAD-related phenomena. Therefore, there is a strong demand for the development of fluorescent NAD analogs that can act as NAD surrogates in various biological settings. An important class of NAD-derived molecular tools are analogs with photophysical properties responsive to structural changes occurring during miscellaneous enzymatic transformations. These analogs are usually modified by replacing with adenosine analog exhibiting emissive properties. The first analog representing this group, developed as early as the 70s, was ε-NAD+ (1) containing 1,N6-ethenonoadenosine (ε-A, Figure2). However, the analog poorly supported many of the studied NAD-dependent processes (Table1), which was explained by the bulkiness and significantly altered H-bonding pattern of ε-A. Hence, the use of ε-NAD+ 1 was limited to enzymes for which purine ring recognition is not critical for binding at the . During the last decade, a significant surge has been observed in adenosine-modified NAD analogs with properties superior to ε-NAD. Molecules 2019, 24, x 4 of 34 nicotinamide riboside. Moreover, NADH has emissive properties with maximum at 460 nm, while NAD+ is virtually non-fluorescent. These differences in spectroscopic properties have been widely applied to monitor the enzymatic activity of NAD-dependent redox enzymes. The changes in fluorescent properties of NADH have also been exploited to study protein binding in vitro and in live cells [28–30]. However, the relatively low quantum yield of NADH (~2%) [31] and lack of emission for NAD+ exclude applications in the investigation of many other NAD-related phenomena. Therefore,Molecules 2019 there, 24, 4187 is a strong demand for the development of fluorescent NAD analogs that can act4 of as 34 NAD surrogates in various biological settings.

O N O NH2 H N N N N N H2N H2N O O O O N+ N N N+ N N O P O P O O O P O P O O O- O- O- O- O O HO OH HO OH HO OH HO OH 1 : εΟΝΠ1 + 2e

O NH2 O NH2 N H2N N H2N N O O S O O S + + N O P O P O O N N O P O P O O N O- O- O- O- O O HO OH HO OH HO OH HO OH 6 : NtzAD+ 8 :NthAD+

N

-O S O S O 3 N Molecules 2019, 24, x O H2N NH 5 of 34 H O O O S N N+ O P O P O N N + O O strongH forN compoundN 2e (46-fold quantum yield increase upon removalO- Oof- nicotinamide moiety). 2 O O HO OH HO OH Thus, the utilityN+ of this analogO to study NAD-consuming processes was verified in the context of three F O + enzymatic activities: O glycohydrolaseP O N N (from O porcin brain), pyrophosphatase10 : 6-Thio NHD (snake venom O- N N phosphodiesteraseO from Crotalus adamanteusH ), and ADP-ribosyl cyclase (ADPRC from Aplysia OH 9 californica), which also possess N1-cADPR activity. Analog 2e was readily consumed by pyrophosphataseFigure 2.2. RepresentativeRepresentative and glycohydrolase NAD NAD+ analogs+ analogs activities, developed developed which, as fluorescent as as fluorescent expected, probes probes forresulted NAD-consuming for in NAD-consuming a strong enzymes. increase of fluorescenceenzymes.The fluorophore intensity,The fluorophore is marked confirming is in marked purple. its inutility For purp emissive le.as Fora turn-on properties,emissive probe properties, see Table for 1enzymatic see. Table 1. activity monitoring. The behavior of 2e towards ADPRC depended on enzyme concentration. In the presence of low enzymeAnPergolizzi importantconcentration, et al.class [ 32the of] decrease reportedNAD-derived of a fluorescence series molecular of NAD intensity tools analogs are was analogs containing obse rvedwith andphotophysical an aromaticinterpreted fluorogenic properties as a result responsivesubstituentof cyclization to at ofstructural the NAD 8-position analog changes at (Scheme the occurring N1-position,1A, compoundsduring to miscellaneousgive 8-substituted2a–e). The enzymatic substituentsN1-cADPR. transformations. At were higher introduced enzyme These analogsbyconcentrations derivatization are usually after of 8-Br-adenosinemodifiedinitial fluorescence by replacing monophosphate drop, adenosine fluorescence3 (8-Br-AMP)with adenosine increase under analogwas Suzuki–Miyaura observed, exhibiting which emissive aqueous was properties.couplinginterpreted conditions. Theas thefirst hydrolysis analog The dinucleotides representing of N1-cADPR were this then group,into obtained ADPR developedby analog. coupling as Hence,early the correspondingas the the new 70s, NADwas 8-substituted ε -NADanalog+ (2e1) containingAMPallowed morpholidate direct 1,N6-ethenonoadenosine visualization4 with NMN of the in different formamide (ε-A, Figure reaction in 2). the However, pathways presence the ofcatalyzed MgSO analog4 byandpoorly ADPRC. MnCl supported2. many of the studied NAD-dependent processes (Table 1), which was explained by the bulkiness and significantly altered H-bonding pattern of ε-A. Hence, the use of ε-NAD+ 1 was limited to enzymes for which purine ring recognition is not critical for binding at the active site. During the last decade, a significant surge has been observed in adenosine-modified NAD analogs with properties superior to ε-NAD. Pergolizzi et al. [32] reported a series of NAD analogs containing an aromatic fluorogenic substituent at the 8-position (Scheme 1A, compounds 2a–e). The substituents were introduced by derivatization of 8-Br- 3 (8-Br-AMP) under Suzuki–Miyaura aqueous coupling conditions. The dinucleotides were then obtained by coupling the corresponding 8- substituted AMP morpholidate 4 with NMN in formamide in the presence of MgSO4 and MnCl2. The fluorescence properties of these compounds varied depending on the nature of the 8- substituent, but in all cases, the fluorescence quantum yield for NAD analog was lower than that for the corresponding AMP analog, indicating intramolecular quenching effect in NAD caused by stacking of nucleobases, similar to that previously described for ε-NAD. This effect was particularly

Scheme 1.1. ChemicalChemical synthesissynthesis ofof selectselect fluorescentfluorescent NAD NAD++ analogsanalogs[ 32[32,33].,33]. (A)) 8-substituted8-substituted NADNAD++ tztz + analogs reportedreported byby PergolizziPergolizzi etet al. al. andand ( B)N) N AD firstfirst reported reported by Rovira et al. Conditions: a) Br2,, KH22PO4(aq), ,rt; rt; b) b) Na Na22PdClPdCl4,4 TPPTS,, TPPTS, R-B(OH) R-B(OH)22, ,KK2CO2CO3,3 H,H2O,2O, Δ∆, 1–24, 1–24 h; h; c) c) morpholine, morpholine, dithiodipyridine, dithiodipyridine, Ph3P, DMSO, rt, 22 h;h; d)d) NMN,NMN, drydry MgSOMgSO44,, MnClMnCl22, formamide,formamide, rt,rt, 2424 h;h; e)1.e)1. POClPOCl3,, proton sponge, (MeO)(MeO)3PO, 4 °C,◦C, 2 h, 2. hydrolysis; f) NMN imidazolide (synthesized from NMN using CDI, Et33N, DMF at rt), DMF, rt,rt, 44 days.days.

Significant advancements in the field of fluorescent NAD analogs for enzymatic activity monitoring have been achieved by the Tor group by implementation of their unique adenosine analogs, isothiazolo[4,3-d]pyrimidine riboside (abbreviated as tzA, 5, Scheme 1B) and thieno[3,4- d]pyrimidine riboside (abbreviated as thA). These nucleosides have initially been developed as tools for oligonucleotide modification and functional probes for ATP-consuming enzymes and are shown to minimize the structural and functional perturbations arising by replacing adenosine due to very similar size and H-bonding pattern [34–37]. Rovira et al. [33] incorporated tzA moiety into NAD to obtain an isomorphic and isofunctional analog 6, enabling real-time monitoring of various NAD transformations. NtzAD+ (6) was synthesized starting from tzA (5), which was treated with oxychloride (POCl3) in trimethyl to yield tzAMP (7), which, in turn, was coupled with imidazole-activated NMN (obtained by treatment with CDI) in DMF (Scheme 1B). The spectroscopic properties of NtzAD+ were very similar to those of isolated tzA with emission maximum at 411 nm (Table 1). The biocompatibility of NtzAD+ and the sensitivity of fluorescence properties to structural transformations were tested with several enzymes, including , NAD glycohydrolase, and monoADP-ribosyltransferases. NtzAD+ was converted into NtzADH by Saccharomyces cerevisiae alcohol (ADH) with efficiency comparable to that of NAD+. About three-fold decrease in fluorescence intensity was observed during this process,

Molecules 2019, 24, 4187 5 of 34

The fluorescence properties of these compounds varied depending on the nature of the 8-substituent, but in all cases, the fluorescence quantum yield for NAD analog was lower than that for the corresponding AMP analog, indicating intramolecular quenching effect in NAD caused by stacking of nucleobases, similar to that previously described for ε-NAD. This effect was particularly strong for compound 2e (46-fold quantum yield increase upon removal of nicotinamide moiety). Thus, the utility of this analog to study NAD-consuming processes was verified in the context of three enzymatic activities: glycohydrolase (from porcin brain), (snake venom from Crotalus adamanteus), and ADP-ribosyl cyclase (ADPRC from Aplysia californica), which also possess N1-cADPR hydrolase activity. Analog 2e was readily consumed by pyrophosphatase and glycohydrolase activities, which, as expected, resulted in a strong increase of fluorescence intensity, confirming its utility as a turn-on probe for enzymatic activity monitoring. The behavior of 2e towards ADPRC depended on enzyme concentration. In the presence of low enzyme concentration, the decrease of fluorescence intensity was observed and interpreted as a result of cyclization of NAD analog at the N1-position, to give 8-substituted N1-cADPR. At higher enzyme concentrations after initial fluorescence drop, fluorescence increase was observed, which was interpreted as the hydrolysis of N1-cADPR into ADPR analog. Hence, the new NAD analog 2e allowed direct visualization of the different reaction pathways catalyzed by ADPRC. Significant advancements in the field of fluorescent NAD analogs for enzymatic activity monitoring have been achieved by the Tor group by implementation of their unique adenosine analogs, isothiazolo[4,3-d]pyrimidine riboside (abbreviated as tzA, 5, Scheme1B) and thieno[3,4-d]pyrimidine riboside (abbreviated as thA). These nucleosides have initially been developed as tools for oligonucleotide modification and functional probes for ATP-consuming enzymes and are shown to minimize the structural and functional perturbations arising by replacing adenosine due to very similar size and H-bonding pattern [34–37]. Rovira et al. [33] incorporated tzA moiety into NAD to obtain an isomorphic and isofunctional analog 6, enabling real-time monitoring of various NAD transformations. NtzAD+ (6) was synthesized tz starting from A(5), which was treated with phosphorus oxychloride (POCl3) in trimethyl phosphate to yield tzAMP (7), which, in turn, was coupled with imidazole-activated NMN (obtained by treatment with CDI) in DMF (Scheme1B). The spectroscopic properties of NtzAD+ were very similar to those of isolated tzA with emission maximum at 411 nm (Table1). The biocompatibility of N tzAD+ and the sensitivity of fluorescence properties to structural transformations were tested with several enzymes, including dehydrogenases, NAD glycohydrolase, and monoADP-ribosyltransferases. NtzAD+ was converted into NtzADH by Saccharomyces cerevisiae (ADH) with efficiency comparable to that of NAD+. About three-fold decrease in fluorescence intensity was observed during this process, making the properties of this analog complementary to those of NAD+ (for which fluorescence intensity increases upon reduction). The NtzADH process could be almost completely reversed by addition of either acetaldehyde or (LDH) and pyruvate. Both NtzAD+ and NtzADH were hydrolyzed by porcine brain glycohydrolase showing modest (30–40%) increase in fluorescence. Finally, the emissive properties of NtzAD+ were employed to real-time monitoring of protein monoADPribosyltransferase (MART) activity. In this context, two MARTs were tested in the presence of agmantine as a model substrate, human ART5 and cholera toxin subunit A (CTA). It was found that the human enzyme catalyzed almost exclusively hydrolysis of NAD+/NtzAD+ to the corresponding ADPribose, whereas CTA primarily catalyzed ADPribosylation of agmantine. Both processes could be efficiently monitored by emission spectroscopy if NtzAD+ was used as a substrate. Halle et al. [38] aimed to increase the availability and applicability of NtzAD+. To this end, they developed an enzymatic method for the synthesis of NtzAD+ from tzATP and NMN catalyzed by nicotinamide mononucleotide adenylyl 1 (NMNAT-1). Further, they showed that NtzAD+ can be phosphorylated by NAD+ (NADK from B. subtilis) to the corresponding NtzADP+, which in turn serves as a substrate for glucose-6-phosphate dehydrogenase (G6PDH from S. cerevisiae) Molecules 2019, 24, 4187 6 of 34 producing NtzADPH. Finally, NtzADPH can be reoxidated to NtzADP+ by . The emissive properties of the NtzADP+/NtzADPH redox system were very similar to those previously reported for NtzAD+/ NtzADH pair and enabled real-time monitoring by emission spectroscopy. Feldman et al. [39] applied a similar chemoenzymatic approach to obtain a thieno[3,4-d]pyrimidine- based NAD+ analog (NthAD+, 8, Figure2). Synthetic thATP was quantitatively converted into NthAD+ by NMNAT1 in the presence of NMN, albeit at a lower rate than ATP or tzATP.

Table 1. Optical properties of NAD and select adenosine-modified emissive NAD analogs.

Enzymatic Activities Tested with Compound Modified Nucleoside λ a λ a (Φ) Refs. ex em NAD Analog b NAD+ 259 - n/a [31] NADH 340 460 (0.02) (+) NAD glycohydrolase (8%) 1,N6-Ethenoadenosine ε-NAD+ 1 300 415 NAD pyrophosphatase (54%) [40–42] (ε-A) Alcohol dehydrogenase (3%) PARP (-) NAD glycohydrolase (+) 8-(Pyr)NAD+ 8-(2-pyrrolyl)adenosine 300 410 (0.005) pyrophosphatase (+) [32] 2e ADPR cyclase/cADPR hydrolase (+) Alcohol dehydrogenase (+) NAD glycohydrolase (+) Mono-ADP-ribose transferase (ART NtzAD+ 6 isothiazolo[4,3-d] 336 411 (0.038) 5c/CTAd)(+) pyrimidine-riboside NAD+ kinaseEnzymatic synthesis from [33,38,39] (tzA) tzATP by NMNAT (+) PARP (+/-) Lactate dehydrogenase (+) NtzADH 336 412 (0.015) NAD glycohydrolase (+) Enzymatic synthesis from thATP by NMNAT (+) thieno[3,4-d] Alcohol dehydrogenase (+) NthAD+ 8 pyrimidine-riboside 341 431 (0.071) [39] NADase (+) (thA) Mono-ADP-ribose transferase (CTA) (+) PARP (-) a λ values are given in nm; b values in brackets indicate relative activity reported in reference to NAD; “+” indicates that the co-factor supported enzymatic activity comparably to NAD or at an analytically useful level; “-” indicates that the co-factor did not support the enzymatic activity, the reaction was very inefficient or analytically not useful. c Hydrolysis to ADPR was observed as the major process; d ADP-ribosylation was observed as the major process.; n/a–not applicable.

NthAD+ showed slightly red-shifted emission and almost 2-fold higher brightness compared to NtzAD+. The compound served as a substrate for alcohol dehydrogenase (from S. cerevisiae) producing NthADH, but the process was accompanied by only 8% decrease in fluorescence. The analog was also a substrate for porcine brain glycohydrolase producing thADPR. The kinetics for this analog was again slower than in the case of NtzAD+, yet provided about four-fold higher sensitivity. Qualitatively similar observations were made for CTA-catalyzed ADP-ribosylation of agmantine. Finally, both NtzAD+ and NthAD+ were tested as substrates for human polyADPribosyltransferase (PARP1) in auto-polyADPribosylation assay. It turned out that NthAD+ is a much better substrate for PARP1 than NtzAD+, but none of the analogs enabled real-time monitoring of the process. All in all, the results indicated that despite higher quantum yield and responsiveness of NthAD+ relative to NtzAD+, the lack of N7-nitrogen atom causes reduced compatibility with biological processes. The optical and biological properties of select fluorescent NAD analogs are summarized in Table1. Other types of fluorescent NAD analogs, usually tailored for a single specific application, have also been described. A rhodamine-labelled NMN analog (compound 9, Figure2) that was employed as a probe for covalent labelling and real-time dynamics monitoring of the CD38 (which possesses NAD glycohydrolase activity) has been reported by Schrimp et al. [43]. Moreau et al. [44] Molecules 2019, 24, 4187 7 of 34

+ synthesized a nicotinamide 6-mercaptopurine 50-dinucleotide (6-Thio NHD , 10, Figure2) and found that it is accepted as a substrate by ADP-ribosyl cyclase to form a fluorescent N1-cADPR analog with emission maximum at 415 nm. Fluorescence methods have also been widely explored to study protein mono- and polyADP ribosylation, including real-time visualization [45]. Yet, in this context, the post-enzymatic fluorescent labelling of modified PAR-protein, obtained using an appropriately functionalized NAD analog, is the commonly adopted approach; therefore, we will review this topic separately in the next paragraph.

3. NAD Analogs to Study Protein ADP-Ribosylation ADP-ribosylation is a post-translational protein modification catalyzed by ADP-ribosyltransferases (ARTs), which regulates DNA repair and gene expression. The reaction comprises of the transfer of either single (mono-ADPribosylation) or multiple (polyADPribosylation or PARylation) moieties on specific amino residues in the target protein, catalyzed by MARTs and PARPs, respectively. Different amino with nucleophilic groups in the side chains have been identified as modification sites, including glutamate, aspartate, arginine, lysine, serine, and cysteine. The polyADPribose chains are added onto the target proteins upon DNA damage to trigger repair mechanisms. The growing polymer consists of either linear or occasionally branched chains consisting of up to a few hundred ADP ribose subunits [46], which makes analysis of PARylated proteins a challenging task. Mono-ADPribosylation usually to the inactivation of the target protein and is catalyzed not only by intracellular enzymes, but also by bacterial toxins [47]. One of the approaches to study protein targets for mono- and poly-ADP-ribosylation is by taking advantage of chemically modified substrates that are utilized by enzymes similarly as NAD+ and are either already labelled or amenable to post-enzymatic labelling with fluorescent dyes or affinity tags (Figure3A). A ground-laying work for this field was reported by Zhang [ 48], who developed NAD+ analogs labelled with and tigitoxin at positions 6 or 8 using NHS chemistry. The most promising analog (6-Bio-17-NAD 11; Figure3B) was initially validated by labelling of Elongation Factor 2 through diphtheria toxin-catalyzed ADP-ribosylation and affinity purification of nitric -enhanced ADP-ribosylated protein [48], and later was employed by others to study PARP activation in cells and tissues under oxidative stress or diabetic dysfunction [49]. More recently, NAD+ analogs modified with an alkyne or azide group have been introduced and can be easily functionalized using azide-alkyne cycloaddition (AAC) click reaction. Du et al. [50] investigated mono ADP-ribosylation catalyzed by sirtuins using alkyne functionalized NAD+ analogs. To this end, two NAD+ analogs, containing a propargyl moiety either at the position C8 or N6 position (Figure3, compounds 12a and 12b), were synthesized using diphenyl phosphoryl chloride activation for the bond formation and tested as substrates for catalyzed-ADP ribosylation in cell extracts and in vitro. The monoADP-ribosylated products were detected and isolated by reacting them with azido-functionalized and biotin-coupled rhodamine. The results indicated that the tested analogs of NAD+ are substrates for sirtuins, which was independently confirmed by 32P-NAD+. However, the ADP-ribosylation process had slow kinetics, which led the authors to believe that it is a non-physiological artifact, rather than a physiologically relevant process. Jiang et al. [51] used the same set of NAD+ analogs to identify substrate proteins of PARP-1 (known also as ARTD1) and Tankyrase-1. To that end, they first investigated substrate properties of the compounds with recombinant enzymes, followed by experiments in cell extracts. Both PARP-1 and Tankyrase-1 underwent PARylation in the presence of these analogs, as shown using SDS-PAGE analysis. However, the NAD+ analog labelled at the N6 position (12b), with kcat/KM value 12- (PARP-1) and 4-fold (Tankyrase-1) lower compared to NAD+, was a much better substrate that the one labeled at the C8-position (12a). 6-alkyne NAD+ was then incubated in cell extracts, followed by with azido-biotin and protein isolation on streptavidin beads, which led to the isolation of 79 ADP-ribosylated proteins. These included known PARP-1 targets, thereby validating the method, as well as potential novel ones. Proteins polyADPribosylated with clickable NAD+ analogs are also often analyzed by electrophoresis combined Molecules 2019, 24, 4187 8 of 34 with fluorescent labelling of PAR moieties by catalyzed AAC (CuAAC) for visualization. However, the differences in the composition of the synthesized PAR chains, which may differ in length and branching pattern, result in variable migration and pose difficulty in the analysis of complex Moleculesmixtures 2019 of, PAR-proteins.24, x 10 of 34

+ Figure 3. NAD+ analogsanalogs used used as as substrates for protein ADP-ribosylationADP-ribosylation studies.studies. A (A) )The The idea idea of of + + chemoenzymaticchemoenzymatic ADPR-ribose labellinglabelling usingusing modifiedmodified NADNAD+ analogs;analogs; (B)) NAD + modificationmodification sites (in purple, purple, orange, orange, and and green) green) expl exploredored in in the the literature literature so sofar; far; C) (overviewC) overview of analogs of analogs tested tested in this in application.this application. The Marx’s group has proposed several strategies to overcome this difficulty, without compromising the robustness of the PARylation analysis. Wang et al. [52] designed NAD+ analogs that carried an alkyne moiety at the C2 position of adenosine to enable derivatization by CuAAC and optionally were modified at the ribose moiety of adenosine to make them substrates that acted as chain terminators for polyADPribosylation reaction (Figure3, compounds 13a–d). The synthesis of the analogs 13a–d was achieved from appropriately modified 2-iodoadenosine derivatives, followed by a Sonogashira coupling with trimethylsilylacetylene, with POCl3, and coupling with CDI-activated NMN. The compounds were then tested as substrates for ARTD1 in auto-ADP-ribosylation reaction as well as in trans-ADP-ribosylation of histone H1.2. As expected, compounds 13b and 13d lacking the 20’-hydroxyl group or both 20’- and 30’-hydroxyls, respectively, were accepted as substrates for ARTD-1, but produced much shorter products, compared to compound 13a with unmodified ribose moiety. Importantly, the products obtained with analogs 13b and 13d migrated as a single band

Figure 4. Real-time monitoring of PAR turnover in living cells using NAD+ analog 24, labelled with tetramethylrodamine (TMR). A) Experiment overview. B) Time-dependent snapshots of cell nucleus obtained by confocal microscopy. The site-specific DNA damage was induced at 0 sec by microirradation at 775 nm. C) Quantitative data obtained from multiple experiments. D) Structure of

Molecules 2019, 24, 4187 9 of 34

during electrophoretic analysis. The compound lacking only the 30’-hydroxyl (13c) resulted in more efficient and heterogeneous PAR formation, albeit still less efficient than the ribose-unmodified analog 13a. Moreover, it was demonstrated that the dideoxy compound 13d efficiently competed with NAD+ in trans-ADP-ribosylation reaction, but not for auto-ADP-ribosylation. The proteins modified with these analogs were not only visualized by azido-Cy5 but also affinity tagged with azido-biotin, highlighting the possibility of application in proteomic analyses. Wallrodt et al. [53] aimed to apply functionalized NAD+ to intracellular imaging of ARTD-1 catalyzed PARylation. To that end, they explored both known and new NAD+ analogs creating a set of compounds varying in the position of modification (N6-, and C2 or C7) and the type enabling labelling through different bioorthogonal reactions (copper catalyzed or strain promoted azide-alkyne cycloaddition, CuAAC and SPAAC, respectively, and inverse electron demand Diels-Alder cycloaddition, IEDDA) (Figure 3, compounds 14–20). First, the variously functionalized analogs were tested in the previously established in vitro assay using histone H1.2, to gain insight into the substrate scope of ARTD-1. It was found that in contrast to modifications at the positions N6 and C2 of adenosine, the presence of the ethynyl substituent at the C7 position of C7-deazaadenosine completely abolished PAR formation. The analogs modified at the N6-position with bulkier substituents (15 and 18) supported PAR formation, but only in the presence of NAD+, indicating that they are not good initiators of this process. The C2 position of adenosine was found to be most compatible with ARDT-1 as even analogs carrying bulky substituent enabling SPAAC (16) or IEDDA (19, 20) efficiently supported PAR formation both in the presence and absence of NAD+. Finally, it was found that the modified PAR chains can be degraded by PARG (poly(ADP-ribose) glycohydrolase), the major PAR degrading enzyme, which leads to formation of mono-ADPribosylated proteins, thereby simplifying their analysis. These analogs were next tested in cultured cells, to select compounds suitable for visualization of nuclear protein + PARylation occurring after oxidative DNA damage induced by H2O2 treatment. The NAD analogs were incubated with transiently permeabilized cells, followed by cell fixing, labeled with dyes containing complementary functional groups, and analysis by confocal microscopy. Among the tested analogs, the highest signal-to-noise ratios, indicating for the highest specificity, were obtained for compounds 13a and 16 functionalized with alkyne at the C2-position that are amenable to labelling with azide-functionalized dyes using CuAAC and SPAAC. In the final experiment, it was demonstrated that by the simultaneous use of analogs 13a and 16, two-color labelling of PAR moieties in cells is possible, revealing potential of these compounds for applications in more advanced experimental techniques. Fluorescent NAD+ analogs as probes for real-time monitoring of PAR formation in cells have been developed by Buntz et al. [45] (Figure4). To that end, two NAD + analogs, labelled at the N6 or C2 position with TMR, were synthesized. In vitro evaluation revealed, similarly as in other studies, that the C2 substituted compound 24 (Figure4D) was more competitive and e fficient substrate for ARTD-1. The analogue 24 was successfully used in real-time monitoring of PAR synthesis and degradation in living cells for the first time. (Figure4A–C). Moreover, using cells expressing GFP-fused ARTD1 and FLIM-FRET microscopy, the interaction of modified PAR polymer and specific protein in living cells was demonstrated. Recently, another group has for the first time explored the modifications of nicotinamide riboside for protein PARylation studies [54]. To that end, Zhang et al. reported a set of three NAD+ analogs modified with either an alkyne or azide group at the 30-position of nicotinamide riboside (Figure3, 21–23) and their 20-counterparts (not shown), which are accessible by either chemical or chemoenzymatic + methods. Preliminary characterization identified compound 23 (30-N3-NAD ) as the most active and highly specific substrate of PARP1 and PARP2 that has kinetic parameters superior to previously reported adenine-modified compounds and is amenable to post-enzymatic labelling with biotin or fluorescent tags using CuAAC. The utility of compound 23 was demonstrated by in vitro and cell lysate PARylation assays, followed by visualization and labeling of mitochondrial PARylated proteins in fixed HeLa cells. Molecules 2019, 24, x 10 of 34

Figure 3. NAD+ analogs used as substrates for protein ADP-ribosylation studies. A) The idea of chemoenzymatic ADPR-ribose labelling using modified NAD+ analogs; B) NAD+ modification sites Molecules(in 2019purple,, 24, 4187orange, and green) explored in the literature so far; C) overview of analogs tested in this10 of 34 application.

FigureFigure 4.4. Real-timeReal-time monitoring monitoring of of PAR PAR turnover turnover in inliving living cells cells using using NAD NAD+ analog+ analog 24, labelled24, labelled with withtetramethylrodamine tetramethylrodamine (TMR). (TMR). A) Experiment (A) Experiment overview. overview. B) Time-dependent (B) Time-dependent snapshots snapshots of cell nucleus of cell nucleusobtained obtained by confocal by confocal microscopy. microscopy. The site-specific The site-specific DNA DNA damage damage was was induced induced at at0 0sec sec byby microirradationmicroirradation at at 775 775 nm. nm. (C) Quantitative datadata obtainedobtained fromfrom multiplemultiple experiments. experiments. ( D)) StructureStructure ofof compound 24. A–C are reproduced with permission from Buntz et al., Angew. Chem. Int. Ed. 55, 11256;

published by John Wiley and SONS, 2016 [45].

4. NAD-Derived Inhibitors of Biologically and Therapeutically Relevant Processes Cancer and infectious diseases such as tuberculosis or malaria are responsible for millions of deaths every year. An increasing problem in treating those diseases is that the long-term use of drugs leads to resistance. This necessitates a continuous search for new active compounds. The development of new candidates is often initiated by the design of analogs of natural compounds that could target specific cellular pathways. Since disruption in NAD can to serious consequences, NAD-dependent enzymes are potential antimicrobial or anticancer drug targets. In the next paragraphs, we will review recent efforts to develop NAD-derived inhibitors, which offer new functional and mechanistic insights, as well as NAD-inspired drug candidates, which provide perspectives for fighting disease.

4.1. NAD Analogs Targeting Enzymes that Attack the Glyosidic Bond

4.1.1. CD38 Ligands and Inhibitors CD38 is a multifunctional glycoprotein expressed on the surface on many types of cells, including immune cells, which regulates various signaling pathways. CD38 acts as a promiscuous ectoenzyme that catalyzes NAD+ conversion to cADPR, the hydrolysis of cADPR to ADPR, and the deamidation of NADP+ to nicotinic acid adenine dinucleotide phosphate (NaADP), thereby participating in Ca2+-dependent signaling. It has been connected to many diseases, and CD38-targeting antibody has been approved as a therapeutic for multiple myeloma [55]. Development of small molecules targeting CD38 has also been pursued to deepen the understanding of it biological functions as well as develop drug candidates. As a natural substrate for CD38, NAD+ has been often used as a starting point for inhibitor design. In 2011, Dong et al. [56] designed a set of 14 analogs, based on a nicotinamide moiety attached to an aromatic substituent via an ether linkage. The linker was introduced in replacement of the pyrophosphate bond present in the natural NAD+ substrate, to ensure better bioavailability. The susceptibility of analogs to inhibit NAD glycohydrolase (NADase) activity of CD38 was assessed using an in vitro assay. All of the compounds had only weak inhibitory activity (2 mM < IC50 < 10 mM). Molecules 2019, 24, x 11 of 34

compound 24. A–C are reproduced with permission from Buntz et al., Angew. Chem. Int. Ed. 55, 11256; published by John Wiley and SONS, 2016 [45].

4. NAD-Derived Inhibitors of Biologically and Therapeutically Relevant Processes Cancer and infectious diseases such as tuberculosis or malaria are responsible for millions of deaths every year. An increasing problem in treating those diseases is that the long-term use of drugs leads to resistance. This necessitates a continuous search for new active compounds. The development of new drug candidates is often initiated by the design of analogs of natural compounds that could target specific cellular pathways. Since disruption in NAD metabolism can lead to serious consequences, NAD-dependent enzymes are potential antimicrobial or anticancer drug targets. In the next paragraphs, we will review recent efforts to develop NAD-derived inhibitors, which offer new functional and mechanistic insights, as well as NAD-inspired drug candidates, which provide perspectives for fighting disease.

4.1. NAD Analogs Targeting Enzymes that Attack the Glyosidic Bond

4.1.1. CD38 Ligands and Inhibitors CD38 is a multifunctional glycoprotein expressed on the surface on many types of cells, including immune cells, which regulates various signaling pathways. CD38 acts as a promiscuous ectoenzyme that catalyzes NAD+ conversion to cADPR, the hydrolysis of cADPR to ADPR, and the deamidation of NADP+ to nicotinic acid adenine dinucleotide phosphate (NaADP), thereby participating in Ca2+-dependent signaling. It has been connected to many diseases, and CD38- targeting antibody has been approved as a therapeutic for multiple myeloma [55]. Development of small molecules targeting CD38 has also been pursued to deepen the understanding of it biological functions as well as develop drug candidates. As a natural substrate for CD38, NAD+ has been often used as a starting point for inhibitor design. In 2011, Dong et al. [56] designed a set of 14 analogs, based on a nicotinamide moiety attached to an aromatic substituent via an ether linkage. The linker was introduced in replacement of the + Moleculespyrophosphate2019, 24, 4187 bond present in the natural NAD substrate, to ensure better bioavailability.11 ofThe 34 susceptibility of analogs to inhibit NAD glycohydrolase (NADase) activity of CD38 was assessed using an in vitro assay. All of the compounds had only weak inhibitory activity (2 mM< IC50 <10 mM). Nonetheless, an observation waswas made that the presencepresence of an aromatic groupgroup atat thethe distal end of the nicotinamide moiety and a longer aliphatic chain increaseincrease the eefficiencyfficiency ofof thethe inhibition,inhibition, whereaswhereas the presence of an electron withdrawing group such as CF leads to the loss of the affinity for CD38. the presence of an electron withdrawing group such as CF3 leads to the loss of the affinity for CD38. ToTo verifyverify thethe biologicalbiological activityactivity ofof thesethese compounds,compounds, thethe authorsauthors assessedassessed their physiological eeffectsffects 2+ on musclemuscle relaxation.relaxation. Indeed Indeed,, it has been shownshown thatthat CaCa2+ increaseincrease is required for airwayairway smoothsmooth muscle contraction,contraction, whichwhich cancan be be induced induced by by the the presence presence of of acetylcholine acetylcholine (Ach). (Ach). This This is mediatedis mediated by CD38,by CD38, which which can, can, thereby, thereby, be used be used as a targetas a target for investigating for investigating its physiological its physiological role in cellsrole in and cells tissues. and Usingtissues. the Using tracheal the tracheal trips from trips rat from/guinea rat/guinea pig, the pi authorsg, the authors induced induced contraction contraction of this muscle of this bymuscle Ach treatmentby Ach treatment and subsequently and subsequently introduced introduced the inhibitors the describedinhibitors above. described Among above. the testedAmong compounds, the tested twocompounds, (25 and 26 two, Figure (25 and5) have26, Figure led to 5) a have significant led to relaxationa significant of relaxation muscle, in of a muscle, concentration in a concentration dependent way.dependent Compound way. Compound25 was also 25 crystallized was also incrystallized complex within complex CD38,confirming with CD38, that confirming its nicotinamide that its + moietynicotinamide binds inmoiety the active binds site in inthe a active similar site way in as a insimilar NAD way, providing as in NAD a guide+, providing for the designa guide of for more the potentdesign inhibitorsof more potent of CD38. inhibitors of CD38.

Figure 5. Inhibitors of CD38 NADase activity showing muscle relaxation eeffectffect [[56].56].

In 2012, Kwong et al. [57] synthesized stable NAD+ analogs and studied their inhibitory properties towards NADase activity of CD38. As the starting point for the design of inhibitors, they chose ara-20F-NMN 27 (IC50 0.5 µM, Figure6), which was known as one of the most potent inhibitors, but was unstable against activity. The first modifications that were tested to potentially overcome this difficulty were phosphate substitutions with two groups (Figure6A, R3 and R4), but they led to the loss of inhibition. Therefore, in the next iteration, only one alkyl group was introduced, which led to only modest loss of inhibition. Sugar moiety modifications were also explored in this 1 study. The change in orientation of 20-F(R ) did not lead to a significant change in inhibitory effect, 2 whereas introducing 30-F(R ) instead of 20-F led to dramatic increase in the IC50 value (IC50 > 0.25 mM). Replacing the fluorine atom by either Cl or N3 also increased IC50 (IC50 = 11 µM and IC50> 1 mM, respectively). Finally, by replacing phosphate with (Figure6A, X = S) or by introducing bulkier aromatic groups onto the phosphate, they obtained compounds with IC50 only slightly higher than ara-20F-NMN 27 (IC50 < 3 µM). Compound 28 (IC50 2 µM, Figure6B) was crystallized with CD38 (Figure7A), confirming that the negatively charged phosphate is crucial for the interaction with this protein. Overall, the study provided new insights into SAR of ara-20F-NMN and emphasized the importance of the negative charge on phosphate group for inhibition of the NADase activity of CD38. Importantly, the presence of thiophosphate or monosubstituted phosphate reduced the susceptibility of compounds to , thereby increasing their metabolic stability and enabling more advanced biological studies. Hence, the effects of those compounds toward two cADPR-related processes were then studied. cADPR level increases concomitantly with CD38 expression when inducing HL-60 cells differentiation by retinoic acid. Thus, the authors have tested the inhibition effect of compound 28 in cell model and found a concentration-dependent response, leading to the decrease of cADPR levels. It has also been shown that lowering the amount of cADPR in rat lacrimal cells results in impairment in aortic contraction. After confirming that the compound 29 can inhibit rat and human CD38 (IC50 2 µM), it was found that this compound inhibits the contraction of rat mesenteries, previously induced by the addition of phenylephrine, with a half maximal effective concentration of 30 µM. These results confirmed that properly designed CD38 inhibitors are useful tools for studying CD38/cADPR signaling pathways in cells and tissues. Molecules 2019, 24, x 12 of 34

In 2012, Kwong et al. [57] synthesized stable NAD+ analogs and studied their inhibitory properties towards NADase activity of CD38. As the starting point for the design of inhibitors, they chose ara-2′F-NMN 27 (IC50 0.5 µM, Figure 6), which was known as one of the most potent inhibitors, but was unstable against phosphatase activity. The first modifications that were tested to potentially overcome this difficulty were phosphate substitutions with two alkyl groups (Figure 6A, R3 and R4), but they led to the loss of inhibition. Therefore, in the next iteration, only one alkyl group was introduced, which led to only modest loss of inhibition. Sugar moiety modifications were also explored in this study. The change in orientation of 2′-F(R1) did not lead to a significant change in inhibitory effect, whereas introducing 3′-F(R2) instead of 2′-F led to dramatic increase in the IC50 value (IC50 > 0.25 mM). Replacing the fluorine atom by either Cl or N3 also increased IC50 (IC50 = 11 µM and IC50> 1 mM, respectively). Finally, by replacing phosphate with thiophosphate (Figure 6A, X = S) or by introducing bulkier aromatic groups onto the phosphate, they obtained compounds with IC50 only slightly higher than ara-2′F-NMN 27 (IC50 < 3 µM). Compound 28 (IC50 2 µM, Figure 6B) was crystallized with CD38 (Figure 7A), confirming that the negatively charged phosphate is crucial for the interaction with this protein. Overall, the study provided new insights into SAR of ara-2′F-NMN and emphasized the importance of the negative charge on phosphate group for inhibition of the NADase activity of CD38. Importantly, the presence of thiophosphate or monosubstituted phosphate reduced the susceptibility of compounds to phosphatases, thereby increasing their metabolic stability and enabling more advanced biological studies. Hence, the effects of those compounds toward two cADPR-related processes were then studied. cADPR level increases concomitantly with CD38 expression when inducing HL-60 cells differentiation by retinoic acid. Thus, the authors have tested the inhibition effect of compound 28 in cell model and found a concentration-dependent response, leading to the decrease of cADPR levels. It has also been shown that lowering the amount of cADPR in rat lacrimal cells results in impairment in aortic contraction. After confirming that the compound 29 can inhibit rat and human CD38 (IC50 2 µM), it was found that this compound inhibits the contraction of rat mesenteries, previously induced by the addition of phenylephrine, with a half Moleculesmaximal2019, 24 effective, 4187 concentration of 30 µM. These results confirmed that properly designed CD3812 of 34 inhibitors are useful tools for studying CD38/cADPR signaling pathways in cells and tissues.

Molecules 2019, 24, x 13 of 34

addition to the 2′-ara-F, or the replacement of adenine with nicotinamide resulted in lower inhibitory Figure 6. A) General structures of the different NMN derivatives synthesized as CD38 inhibitors; B) Figure 6.activity.(A) General This revealed structures the importance of the di ffoferent the pyrophosphate NMN derivatives bridge synthesized and the purine as CD38ring for inhibitors; structures of ara-2′F-NMN 27 and analogs 28 and 29 [57]. (B) structuresrecognition, of ara-2 albeit0F-NMN the latter27 showingand analogs some flexibility28 and towards29 [57]. modification. Wang et al. [58] approached the development of CD38 inhibitors by exploring various modifications of all three structural parts present in NAD+, namely the ribose, the pyrophosphate, and the nucleobase. After in vitro evaluation as CD38 NADase inhibitors, three most active compounds were selected (30a–c, Figure 8). Interestingly, all of them possess a 2′-ara fluorine atom on the nicotinamide moiety as well as purine modification. Other modifications such as replacement of the phosphate neighboring the nicotinamide with thiophosphate, a second 2′-substituent in

Figure 7. Crystallographic structures of representative NAD+ or analogs in complexes with target Figure 7. Crystallographic structures of representative NAD+ or analogs in complexes with target proteins. A) NAD+ analog 28 bound to CD38 (enzyme labelled in blue) PDB code: 3ROK; B) S-NAD+ + + proteins. (A)39 NAD bound toanalog SIRT2 (enzyme28 bound labelled to CD38in green), (enzyme PDB code: labelled 6EDR; C) in carba-NAD blue) PDB+ 33 bound code: to 3ROK; SIRT3 (B) S-NAD 39 bound to(enzyme SIRT2 (enzymelabelled in green), labelled PDB incode: green), 4FVT; D PDB) carba-NAD code:+ 6EDR;33 bound ( Cto) SIRT5 carba-NAD (enzyme labelled+ 33 bound in to SIRT3 + (enzyme labelledgreen), inPDB green), code: 4G1C; PDB E code:) NAD 4FVT;bound to (D h)IMPDH2 carba-NAD (enzyme+ labelled33 bound in light to SIRT5blue), PDB (enzyme code: labelled in 1NFB; F) MAD 44 bound to hIMPDH2 (enzyme labelled in light blue, labelled in dark blue), + green), PDBPDB code: code: 4G1C; 1NF7. General (E) NAD labels boundfor all the to structures:hIMPDH2 carbon (enzyme from NAD labelled+ analog in inpurple, light carbon blue), PDB code: 1NFB; (F) MADfrom amino44 bound acid side to chainshIMPDH2 in grey, (enzyme in red, labelled nitrogen in in blue, light phosphorus blue, ligand in orange labelled and in dark blue), PDB code: 1NF7.in yellow. General labels for all the structures: carbon from NAD+ analog in purple, carbon from amino acid side chains in grey, oxygen in red, nitrogen in blue, phosphorus in orange and sulfur in yellow.

Molecules 2019, 24, 4187 13 of 34

Wang et al. [58] approached the development of CD38 inhibitors by exploring various modifications of all three structural parts present in NAD+, namely the ribose, the pyrophosphate, and the nucleobase. After in vitro evaluation as CD38 NADase inhibitors, three most active compounds were selected (30a–c, Figure8). Interestingly, all of them possess a 2 0-ara fluorine atom on the nicotinamide moiety as well as purine modification. Other modifications such as replacement of the phosphate neighboring the nicotinamide with thiophosphate, a second 20-substituent in addition to the 20-ara-F, or the replacement of adenine with nicotinamide resulted in lower inhibitory activity. This revealed the importance of the pyrophosphate bridge and the purine ring for recognition, albeit the latter showing some flexibility MoleculesMoleculestowards 2019 2019, modification. 24,, x24 , x 14 of14 34of 34

FigureFigure 8. Best 8. BestBest potent potent potent inhibitors inhibitors inhibitors of NADase of of NADase NADase activity activity activity of CD38 of CD38 described described by Wangby Wang et al. et [58]. al. [[58].58 ]. 4.1.2. NAD Analogs Targeting Sirtuins and ADPribosyltransferases 4.1.2.4.1.2. NAD NAD Analogs Analogs Targeting Targeting Sirtuins Sirtuins and and ADPribosyltransferases ADPribosyltransferases Sirtuins are NAD-dependent enzymes, which induce the removal of acyl groups of different SirtuinsSirtuins are are NAD-dependent NAD-dependent enzymes, enzymes, which which indu induce cethe the removal removal of acylof acyl groups groups of differentof different lengths (from C2 up to C14) from amino acid residues in various protein substrates, with deacetylation lengthslengths (from (from C2 C2up upto toC14) C14) from from amino amino acid acid residues residues in invarious various protein protein substrates, substrates, with with of acetyllysine being the most thoroughly studied process [59]. Human sirtuins can be classified into deacetylationdeacetylation of acetyllysineof acetyllysine being being the the most most thoroughly thoroughly studied studied process process [59]. [59]. Human Human sirtuins sirtuins can can be be seven families (SIRT 1–7) and are involved in different cellular processes, including aging, transcription, classifiedclassified into into seven seven families families (SIRT (SIRT 1–7) 1–7) and and are are involved involved in differentin different cellular cellular processes, processes, including including and apoptosis. Some sirtuins have been identified as molecular targets for disease treatment, including aging,aging, transcription, transcription, and and apoptosis. apoptosis. Some Some sirtuins sirtuins have have been been identified identified as asmolecular molecular targets targets for for cancer, diabetes, and age-related diseases. diseasedisease treatment, treatment, including including cancer, cancer, diabetes, diabetes, and and age-related age-related diseases. diseases. In 2011, a set of 8-substituted NAD+ analogs as potential sirtuin inhibitors were designed by Pesnot In In2011, 2011, a set a setof 8-substitutedof 8-substituted NAD NAD+ analogs+ analogs as aspotential potential sirtuin sirtuin inhibitors inhibitors were were designed designed by by et al. [60]. To avoid the most problematic step in NAD+ analog synthesis, which is pyrophosphate PesnotPesnot et etal. al.[60]. [60]. To Toavoid avoid the the most most problematic problematic step step in inNAD NAD+ analog+ analog synthesis, synthesis, which which is is bond formation, the synthetic strategy was based on a one-pot 2-step sequence enabling direct pyrophosphatepyrophosphate bond bond formation, formation, the the synthetic synthetic strategy strategy was was based based on ona one-pota one-pot 2-step 2-step sequence sequence functionalization of NAD+ (Scheme2). The first step consisted of 8-selective bromination of NAD + enablingenabling direct direct func functionalizationtionalization of ofNAD NAD+ (Scheme+ (Scheme 2). 2).The The first first step step consisted consisted of of8-selective 8-selective using saturated bromine water (orange path, Scheme2) or neat bromine (blue path, Scheme2), a ffording brominationbromination of NADof NAD+ using+ using saturated saturated bromine bromine water water (orange (orange path, path, Scheme Scheme 2) or2) neator neat bromine bromine (blue (blue the intermediate 31, followed by a key step of Suzuki-Miyaura cross-coupling reaction to introduce path,path, Scheme Scheme 2), 2),affording affording the the intermediate intermediate 31 , 31followed, followed by bya keya key step step of ofSuzuki-Miyaura Suzuki-Miyaura cross- cross- different substituents at the C8-position of adenine. The synthesized NAD+ analogs 32a–g were then couplingcoupling reaction reaction to introduceto introduce different different substituents substituents at theat the C8-position C8-position of adenine.of adenine. The The synthesized synthesized studied as inhibitors of SIRT1 and SIRT2. NADNAD+ analogs+ analogs 32a 32a–g –wereg were then then studied studied as inhibitorsas inhibitors of SIRT1of SIRT1 and and SIRT2. SIRT2.

Scheme 2. Synthesis of 8-substituted NAD+ analogs designed as inhibitors of sirtuins [60]. Conditions: SchemeScheme 2. Synthesis 2. Synthesis of 8-substituted of 8-substituted NAD NAD+ analogs+ analogs designed designed as inhibitors as inhibitors of sirtuins of sirtuins [60]. [60]. Conditions: Conditions: a) (Br2)aq, (NaOAc)aq, pH 4, rt, 1.75 h; b) Na2Cl4Pd, TXPTS, R-B(OH)2,K2CO3,H2O, 40 ◦C, 20–35 min; a) (Bra) 2(Br)aq,2 )(NaOAc)aq, (NaOAc)aq, pHaq, pH 4, rt, 4, 1.75rt, 1.75 h; b) h; Na b) 2NaCl42Pd,Cl4Pd, TXPTS, TXPTS, R-B(OH) R-B(OH)2, K22,CO K2CO3, H32,O, H 240O, °C,40 °C,20–35 20–35 min; min; c) c) c) i. neat Br2, (NaOAc)aq, pH 4, rt, 30 min; ii. Na2Cl4Pd, TXPTS, R-B(OH)2,K2CO3,H2O, 40 ◦C, i. neati. neat Br2 ,Br (NaOAc)2, (NaOAc)aq, pHaq, pH 4, rt, 4, 30rt, min;30 min; ii. Na ii. 2NaCl42Pd,Cl4Pd, TXPTS, TXPTS, R-B(OH) R-B(OH)2, K22,CO K2CO3, H32,O, H 240O, °C,40 °C,20–30 20–30 min. min. 20–30 min. AllAll the the 8-substituted 8-substituted NAD NAD+ analogs+ analogs showed showed inhibitory inhibitory activity activity at atlow low micro-molar micro-molar concentrationsconcentrations toward toward SIRT2, SIRT2, but but only only moderate moderate activity activity toward toward SIRT1. SIRT1. These These first first results results suggested suggested thatthat the the modification modification at theat the position position 8 of 8 NADof NAD+ is+ generallyis generally well well accepted accepted by bythe the SIRT1 SIRT1 enzyme. enzyme. In In furtherfurther studies, studies, only only adenosine adenosine or oradenosine adenosine mo monophosphatenophosphate scaffolds, scaffolds, with with or orwithout without the the 8- 8- modification,modification, were were tested tested toward toward the the two two sirtuins. sirtuins. It resulted It resulted in eitherin either a complete a complete loss loss of activity,of activity, or or in similarin similar activity activity towards towards both both enzymes, enzymes, showing showing the the importance importance of theof the complete complete NAD NAD+ pattern+ pattern for for enzymeenzyme selectivity. selectivity. NMR-based NMR-based experiments experiments direct directed edat determiningat determining the the conformation conformation of theof the NAD NAD+ + analogsanalogs in solutionin solution provided provided hints hints about about probable probable reasons reasons for for selectivity. selectivity. The The chemical chemical shifts shifts of theof the H2H2′’ atoms′’ atoms indicated indicated that that all allof theof the 8-substituted 8-substituted an alogsanalogs adopt adopt preferentially preferentially the the “syn” “syn” conformation, conformation, whereaswhereas sirtuins sirtuins would would require require the the “anti” “anti” conformati conformation onin thein the active active site, site, suggesting suggesting that that the the SIRT2 SIRT2 is moreis more capable capable of inducingof inducing this this conformational conformational change change than than SIRT1. SIRT1.

Molecules 2019, 24, x 15 of 34 Molecules 2019, 24, 4187 14 of 34 In 2012, Szczepankiewicz et al. [61] synthesized 4′-carba-NAD+ 33, a stable and unreactive mimic of NAD+, in order to determine the ternary structures of SIRT3 and SIRT5 in cofactor-analog bound + state.All Because the 8-substituted the natural NAD NAD+ analogsforms a showedshort-lived inhibitory complex activity with atthese low sirtuins, micro-molar using concentrations this 4′-carba- towardNAD+ was SIRT2, of great but interest only moderate as it could activity not be towardsubjected SIRT1. to nicotinamide These first displacement. results suggested that the + modificationAlthough at the the synthesis position 8of of 4′-carba-NAD NAD is generally+ was described well accepted earlier by [62], the a SIRT1 new improved enzyme. Inroute further was studies,developed only here. adenosine The synthesis or adenosine was achieved monophosphate in six steps, scaff withoutolds, with requiring or without any theprotecting 8-modification, groups were(Scheme tested 3). towardAfter the the dihydrox two sirtuins.ylation It resulted and acidic in either methanol a complete treatment loss ofof activity,the lactam or in 34 similar, the resulting activity towardsaminoester both 35 enzymes, was transformed showing in the the importance amino derivative of the complete 36. Then NAD the glycosidic+ pattern for bond enzyme was formed selectivity. by NMR-basedreaction with experiments the derivative directed 37, affording at determining the 4′-carba the conformation NMN 38. Finally, of the the NAD latter+ analogs was coupled in solution with providedAMP morpholidate hints about in probable the presence reasons of pyridinium for selectivity. tosy Thelate chemical as a buffering shiftsof agent the H2”in order atoms toindicated form the thattarget all dinucleotide of the 8-substituted 33. analogs adopt preferentially the “syn” conformation, whereas sirtuins wouldThe require compound the “anti” was conformationthen successfully in the used active for site, the suggestingdetermination that of the the SIRT2 X-ray is crystal more capable structure of inducingof cofactor-analog this conformational bound SIRT3 change in complex than SIRT1. with its substrate protein, acetyl-CoA enzyme + 2 (ACS2).In 2012, In this Szczepankiewicz structure, a well-defined et al. [61] synthesized electron density 40-carba-NAD from 4′-carba-NAD33, a stable+ was and visible, unreactive which mimicmade multipoint of NAD+, ininteractions order to determine with SIRT3 the (Figure ternary 7C). structures Interestingly, of SIRT3 the andconformation SIRT5 in cofactor-analog of the 4′-carba- + boundNAD+ in state. this Because complex the was natural very NADsimilarforms to the a short-livedone in a previously complex withdetermined these sirtuins, ternary using complex this + 4between0-carba-NAD sirtuinswas of greatyHst2, interest histone as itH4 could notbe and subjected 4′-carba-NAD to nicotinamide+. The X-ray displacement. structure of the + complexAlthough of SIRT5 the with synthesis 4′-carba-NAD of 40-carba-NAD+ and isocitratewas described dehydrogenase earlier [262 (IDH2),], a new which improved is known route to was be developeddessuccinylated here. by The SIRT5, synthesis revealed was a achieved very close in similarity six steps, between without complexes requiring any with protecting SIRT3 and groups SIRT5 (Scheme(Figure 7D).3). After One thedifference dihydroxylation is that 4′-carba-NAD and acidic methanol+ formed treatmentmore interactions of the lactam with SIRT534, the than resulting with aminoesterSIRT3, which35 iswas in agreement transformed with in the the amino observation derivative from36. authorsThen thethat glycosidic the Km for bondthe natural was formed co-factor by reactionis lower for with SIRT5 thederivative than for SIRT3.37, aff Byording comparing the 40-carba the structure NMN 38of. this Finally, complex the latterwith the was one coupled previously with AMPreported morpholidate between SIRT5, in the succinyl presence H3 of peptide, pyridinium and tosylateNAD+, they as a have buffering shown agent that in both order natural to form NAD the+ targetand carba-NAD dinucleotide+ 3333 adopt. the same position in the different complexes.

++ Scheme 3.3. AA syntheticsynthetic pathwaypathway toto 440-carba-NAD′-carba-NAD ((33)[) [61].61]. Conditions:Conditions: a) i.i. OsOOsO44, NMO,NMO, tt-pentyl-pentyl alcohol, H 2O,O, 70 70 °C;◦C; b) b) MeOH, MeOH, HCl, HCl, 25 25 °C;◦C; c) c)i. LiEt i. LiEt3BH,3BH, THF, THF, 0 °C; 0 ◦ ii.C; HF ii.(aq) HF, 0(aq) °C;, 0iii.◦C; LiOH iii.(aq) LiOH, 25 (aq)°C;, 25d) ◦37C;, NaOAc, d) 37, NaOAc, MeOH, MeOH,25 °C; e) 25 POCl◦C; e)3, (MeO) POCl33,PO, (MeO) 0 °C;3PO, f) 0MnCl◦C;2 f), AMP-morpholidate, MnCl2, AMP-morpholidate, p-TsOH, pformamide.-TsOH, formamide.

TheIn 2018, compound Dai et al was. [63] then reported successfully the chemoe usednzymatic for the determination synthesis of 4 of′-S-NAD the X-ray+ 39 crystal – an NAD structure+ analog of cofactor-analogthat, similar to 4 bound′-carba-NAD SIRT3+, in is complexstable against with itsenzymes substrate that protein, cleave the acetyl-CoA C1′-N1 glycosidic enzyme synthase bond, but 2 + (ACS2).features Inribose this structure, geometry a well-definedand electrostatic electron properties density from more 40 -carba-NADsimilar to thewas native visible, cofactor. which made The + multipointsynthesis of interactions the 4′-S-NAD with+ has SIRT3 been (Figure performed7C). Interestingly,starting fromthe D-gluconic conformation acid γ of-lactone the 4 0 -carba-NAD40, leading in ineight this steps complex to the was thioribose very similar 41 [64] to the(Scheme one in 4). a Then, previously the introducti determinedon of ternary the nicotinamide complex between nucleobase yeast + sirtuinsafforded yHst2, the protected histone compound H4 peptide 42 and, followed 40-carba-NAD by the removal. The X-ray of all structure the protecting of the complexgroups leading of SIRT5 to + withthe nicotinamide 40-carba-NAD 4′-thioriboseand isocitrate 43. dehydrogenaseFinally, a two-step 2 (IDH2), enzymatic which isconversion known to using be dessuccinylated nicotinamide byriboside SIRT5, kinase revealed 1 (NRK1) a very and close NMNAT similarity led to between 4′-S-NAD complexes+. with SIRT3 and SIRT5 (Figure7D). + One difference is that 40-carba-NAD formed more interactions with SIRT5 than with SIRT3, which is

Molecules 2019, 24, 4187 15 of 34

in agreement with the observation from authors that the Km for the natural co-factor is lower for SIRT5 than for SIRT3. By comparing the structure of this complex with the one previously reported between SIRT5, succinyl H3 peptide, and NAD+, they have shown that both natural NAD+ and carba-NAD+ 33 adopt the same position in the different complexes. + + In 2018, Dai et al. [63] reported the chemoenzymatic synthesis of 40-S-NAD 39–an NAD + analog that, similar to 40-carba-NAD , is stable against enzymes that cleave the C10-N1 glycosidic bond, but features ribose geometry and electrostatic properties more similar to the native cofactor. + The synthesis of the 40-S-NAD has been performed starting from d-gluconic acid γ-lactone 40, leading in eight steps to the thioribose 41 [64] (Scheme4). Then, the introduction of the nicotinamide nucleobase afforded the protected compound 42, followed by the removal of all the protecting groups leading to the nicotinamide 40-thioribose 43. Finally, a two-step enzymatic conversion using nicotinamide + ribosideMolecules 2019 kinase, 24, x 1 (NRK1) and NMNAT led to 40-S-NAD . 16 of 34

SchemeScheme 4.4. ChemoenzymaticChemoenzymatic pathwaypathway leadingleading toto thethe S-NAD S-NAD++ 39 [[63].63]. The enzymatic transformationstransformations

areare markedmarked in in blue. blue. Conditions: Conditions: a) HBr,a) HBr, toluene, toluene, 0 ◦C, 0 4 h;°C, b) 4 nicotinamide, h; b) nicotinami MeCN,de, rt,MeCN, 18 h; c)rt, TFA 18/ Hh;2 O,c) 0TFA/H◦C, 6 h;2O, d) 0 NH°C, 36/ MeOH,h; d) NH 0 3◦/MeOH,C, 48 h. 0 °C, 48 h.

+ TheThe 440′-S-NAD-S-NAD+ waswas subjected subjected to to functional characterization in various NAD-related processes. AsAs expected, the the compound compound showed showed resistance resistance to th toe theN-glycosidic N-glycosidic bond bond cleavage cleavage catalysed catalysed by CD38, by CD38,inhibited inhibited this enzyme, this enzyme, and was and successfully was successfully applied appliedfor X-ray for structure X-ray structure determination determination of the complex of the + complex(Figure 7B). (Figure 4′-S-NAD7B). 4 0-S-NAD+ also inhibitedalso inhibited activity activityof SIRT2. of SIRT2.On the On other the otherhand, hand, due to due structural to structural and andelectronic electronic resemblance resemblance to NAD to NAD+, the+, the analogue analogue readily readily supported supported redox redox reactions, reactions, opening thethe opportunityopportunity forfor dissectingdissecting variousvarious NAD-dependentNAD-dependent processesprocesses inin complexcomplex biologicalbiological systems.systems. NADNAD++ analogsanalogs have have also also been beenapplied applied for the inhibi for thetion inhibition of ADP-ribosyltransferases. of ADP-ribosyltransferases. Langelier et Langelieral. [65] used et al.an [65unhydrolyzable] used an unhydrolyzable benzamide adenin benzamidee dinucleotide adenine (BAD) dinucleotide to study (BAD) the auto-inhibition to study the auto-inhibitionmechanism of mechanismPARP1. Functional of PARP1. and Functional structural and studies structural using studies this compound using this compoundshowed that showed auto- thatinhibition auto-inhibition results from results a selective from a selectivedisruption disruption of NAD+ ofbinding NAD+ andbinding revealed and the revealed existence the existenceof a reverse of aallosteric reverse allostericregulatory regulatory mechanism. mechanism.

4.2.4.2. NAD Analogs TargetingTargeting NAD-DependentNAD-Dependent MetabolicMetabolic PathwaysPathways NAD-dependentNAD-dependent redoxredox reactionsreactions areare keykey stepssteps inin manymany metabolicmetabolic pathways,pathways, foundfound acrossacross allall livingliving organisms.organisms. However,However, thesethese pathwayspathways maymay didifffferer betweenbetween humanshumans andand pathogens,pathogens, asas wellwell asas betweenbetween healthyhealthy andand cancerouscancerous humanhuman cells,cells, oofferifferingng thethe chancechance forfor thethe developmentdevelopment ofof selectiveselective + inhibitioninhibition strategies.strategies. SomeSome ofof thethe crucialcrucial pathwayspathways thatthat havehave beenbeen targetedtargeted by by NAD NAD+ analogsanalogs includeinclude thethe biosynthesisbiosynthesis ofof nucleicnucleic acids,acids, fattyfatty acids,acids, proteins,proteins, asas wellwell asas thethe biosynthesisbiosynthesis andand transformationstransformations + ofof NADNAD+ itself.itself.

4.2.1. NAD Analogs Targeting Purine Pathway Inosine-5′-monophosphate dehydrogenase (IMPDH) is an NAD-dependent enzyme responsible for guanine nucleotide de novo synthesis, necessary for the growth of cells and, thereby, is a target for the treatment of cancer, autoimmune diseases, and bacterial infections. There are two isoforms of human IMPDH, the hIMPDH1 and hIMPDH2, which share 84% sequence similarity and have similar activity, but differ in expression levels among cells and tissues. hIMPDH1 functions in angiogenesis, DNA-binding, and translation regulation. hIMPDH2 is usually up-regulated in proliferating cells, including cancer cells. Some IMPDH inhibitors have been tested in or have already entered the clinic (Figure 9). IMPDH is also necessary for de novo synthesis of guanine in bacteria and, hence, constitutes an attractive antimicrobial target. Pankiewicz and co-workers have developed a significant number of NAD+ analogs as potent inhibitors of IMPDH. This team has been working on IMPDH inhibitors for a few decades, and have published comprehensive reviews on the topic [66,67]; here, we will focus on their most recent works. Generally, NAD+ analogs can be designed to interact with three different subsites of the IMPDH cofactor binding domain: the nicotinamide binding subsite (N-subsite), the adenosine binding subsite

Molecules 2019, 24, 4187 16 of 34 Molecules 2019, 24, x 17 of 34

(A-subsite), and the pyrophosphate binding subsite (P-subsite). Pankiewicz and coworkers have set 4.2.1. NAD Analogs Targeting Purine Biosynthesis Pathway out their work based on the structure of two of the four known IMPDH inhibitors, which are used in clinicInosine-5 (Figure 0-monophosphate9) [66]. The first dehydrogenaseone is the mycophenolic (IMPDH) is anacid NAD-dependent (MPA), a nanomolar responsible of forhIMPDH1, guanine which nucleotide blocks de tumor-induced novo synthesis, angiogenesis necessary for in thevivo growth by interaction of cells and,with thereby,the N-subsite, is a target and foris used the treatmentas an immunosuppressant. of cancer, autoimmune This has diseases, led to andthe synthesis bacterial infections.of MAD 44 There as a potent are two inhibitor. isoforms The of humansecond IMPDH,is tiazofurin, the h biotransformedIMPDH1 and hIMPDH2, in cells in which thiazole share 4-carboxamide 84% sequence adenine similarity dinucleotide and have similar(TAD, activity,45), which but is di oneffer of in expressionthe most potent levels inhibitors among cells of h andIMPDH2 tissues. interactinghIMPDH1 with functions all three in angiogenesis, subsites and DNA-binding,which is used in and the translation treatment of regulation. chronic myelogenhIMPDH2ous is leukemia usually up-regulated (CML). It has in to proliferating be noted that cells, the includingability of cancerNAD+ cells.derivatives Some IMPDH to cross inhibitors cell membranes have been istested required in or for have in alreadyvivo activity. entered Here, the clinic the (Figurereplacement9). IMPDH of the phosphate is also necessary bridge forby dea methylen novo synthesise (MAD, of 44) guanine or its absence nucleotides (tiazofurin) in bacteria increases and, hence,permeability constitutes of the an derivatives. attractive antimicrobial target.

Figure 9. Some currently known inhibitors of IMPDH used in clinic and theirtheir derivativesderivatives [[66].66].

+ PankiewiczIn 2010, following and co-workers their previous have developed work on a significantthe synthesis number of ofNAD NAD+ analogsanalogs containing as potent inhibitorsmycophenolic of IMPDH. moiety This and/or team triazole has been link working (Figure on 10), IMPDH Chen inhibitorset al. [68] forpublished a few decades, a set of and triazole- have publishedcontaining comprehensive NAD+ analogs reviews and focused on the topicon their [66,67 inhibition]; here, we activity will focus toward on their the most two recent isoforms works. of + Generally,hIMPDH as NAD well analogsas Mycobacterium can be designed tuberculosis to interact IMPDH with (Mtb three IMPDH), different a potential subsites oftarget the IMPDHfor the cofactortreatment binding of tuberculosis. domain: the They nicotinamide synthesized binding four subsite compounds (N-subsite), (46a– theb and adenosine 47a–b, bindingFigure subsite10) by (A-subsite),attaching mycophenolic and the pyrophosphate moiety, as a bindingreplacement subsite of nicotinamide, (P-subsite). Pankiewicz to adenine and moiety coworkers through have a 1,2,3- set outtriazole their linker. work basedThe latter on the has structure been introduced of two of either the four directly known at IMPDH the 5′-position inhibitors, of adenosine which are used(46a– inb) clinicor via (Figurean oxymethylene9)[ 66]. The linkage first one (47a is the–b). mycophenolic The mycophenolic acid (MPA), moiety a has nanomolar been linked inhibitor to the of triazolehIMPDH1, via whicheither a blocks C2 or tumor-induceda C4 chain. angiogenesis in vivo by interaction with the N-subsite, and is used as an immunosuppressant. This has led to the synthesis of MAD 44 as a potent inhibitor. The second is tiazofurin, biotransformed in cells in thiazole 4-carboxamide adenine dinucleotide (TAD, 45), which is one of the most potent inhibitors of hIMPDH2 interacting with all three subsites and which is used in the treatment of chronic myelogenous leukemia (CML). It has to be noted that the ability of NAD+ derivatives to cross cell membranes is required for in vivo activity. Here, the replacement of the phosphate bridge by a methylene (MAD, 44) or its absence (tiazofurin) increases permeability of the derivatives. In 2010, following their previous work on the synthesis of NAD+ analogs containing mycophenolic moiety and/or triazole link (Figure 10), Chen et al. [68] published a set of triazole-containing NAD+ analogs and focused on their inhibition activity toward the two isoforms of hIMPDH as well as

Mycobacterium tuberculosis IMPDH (Mtb IMPDH), a potential target for the treatment of tuberculosis. They synthesizedFigure 10. four Triazol compounds containing (46a NAD–b+and analogs47a –synthesizedb, Figure 10 as) potent by attaching inhibitors mycophenolic 46–47 [68]. moiety, as a replacement of nicotinamide, to adenine moiety through a 1,2,3-triazole linker. The latter has The authors assessed the inhibitory activity toward hIMPDH1, hIMPDH2, and MtbIMPDH. A been introduced either directly at the 50-position of adenosine (46a–b) or via an oxymethylene linkage (better47a–b activity). The mycophenolic towards human moiety enzymes has been was linkedobserved to the when triazole the longer via either carbon a C2 chain or a C4was chain. present (R = b), which has a similar length as the natural NAD+ substrate. The two compounds containing the C4 chain (46b and 47b) showed similar nanomolar inhibition towards both hIMPDH1 and hIMPDH2, while the compounds with shorter carbon chain (46a and 47a) displayed only micromolar activity. Interestingly, only the compound with a longer chain and containing the oxymethylene linker (47b)

Molecules 2019, 24, x 17 of 34

(A-subsite), and the pyrophosphate binding subsite (P-subsite). Pankiewicz and coworkers have set out their work based on the structure of two of the four known IMPDH inhibitors, which are used in clinic (Figure 9) [66]. The first one is the mycophenolic acid (MPA), a nanomolar inhibitor of hIMPDH1, which blocks tumor-induced angiogenesis in vivo by interaction with the N-subsite, and is used as an immunosuppressant. This has led to the synthesis of MAD 44 as a potent inhibitor. The second is tiazofurin, biotransformed in cells in thiazole 4-carboxamide adenine dinucleotide (TAD, 45), which is one of the most potent inhibitors of hIMPDH2 interacting with all three subsites and which is used in the treatment of chronic myelogenous leukemia (CML). It has to be noted that the ability of NAD+ derivatives to cross cell membranes is required for in vivo activity. Here, the replacement of the phosphate bridge by a methylene (MAD, 44) or its absence (tiazofurin) increases permeability of the derivatives.

Figure 9. Some currently known inhibitors of IMPDH used in clinic and their derivatives [66].

In 2010, following their previous work on the synthesis of NAD+ analogs containing mycophenolic moiety and/or triazole link (Figure 10), Chen et al. [68] published a set of triazole- containing NAD+ analogs and focused on their inhibition activity toward the two isoforms of hIMPDH as well as Mycobacterium tuberculosis IMPDH (Mtb IMPDH), a potential target for the treatment of tuberculosis. They synthesized four compounds (46a–b and 47a–b, Figure 10) by attaching mycophenolic moiety, as a replacement of nicotinamide, to adenine moiety through a 1,2,3- triazole linker. The latter has been introduced either directly at the 5′-position of adenosine (46a–b) orMolecules via an2019 oxymethylene, 24, 4187 linkage (47a–b). The mycophenolic moiety has been linked to the triazole17 of via 34 either a C2 or a C4 chain.

Figure 10. Triazol containing NAD + analogsanalogs synthesized synthesized as as potent potent inhibitors inhibitors 4646––4747 [68].[68].

The authors authors assessed assessed the the inhibitory inhibitory activity activity toward toward hIMPDH1,hIMPDH1, hIMPDH2,hIMPDH2, and and MtbMtbIMPDH.IMPDH. A Abetter better activity activity towards towards human human enzymes enzymes was was observed observed when when the the longer longer carbon carbon chain chain was was present present (R + (=R b),= bwhich), which has has a similar a similar length length as as the the natural natural NAD NAD+ substrate.substrate. The The two two compounds compounds containing containing the C4 chain ( 46b and 47b47b)) showed showed similar similar nanomolar nanomolar inhibition inhibition towards towards both both hhIMPDH1IMPDH1 and and hhIMPDH2,IMPDH2, while the compounds with shorter carbon carbon chain chain ( (46a46a andand 47a47a)) displayed displayed only only micromolar micromolar activity. activity. Interestingly, onlyonly the the compound compound with with a longera longer chain chain and and containing containing the the oxymethylene oxymethylene linker linker (47b ()47b had) a significant inhibition activity towards MtbIMPDH. In order to understand the differences in activity, the authors used a previously obtained NAD+ analogue (MAD 44, Figure9) as a model for binding in the presence of inosine monophosphate (IMP), with the three IMPDHs. Computational studies combined with the X-ray crystal structure of the complex of NAD+ analog/IMPDH/IMP (hIMPDH2, Figure7F) revealed aromatic interactions between adenine ring and enzyme for both human IMPDH, whereas no aromatic stacking was observed for MtbIMPDH. Moreover, a bond was present in the N-subsite binding domain only with human IMPDH, where an unfavorable binding was calculated for MtbIMPDH. These differences in binding therefore explain the inhibition selectivity toward human IMPDH. Interestingly, similar computations performed using the new synthesized NAD+ analog (47b) revealed new favorable interactions in the P-subsite domain in addition to the absence of unfavorable ones, oppositely to the observations obtained when using the previous NAD+ analog. The lack of interactions in the A-subsite domain was, therefore, overcome thanks to these new compounds, and led to the first potent inhibitor of MtbIMPDH. Later, a new set of NAD+ analogs that are potential anticancer agents was described [69]. The authors mainly focused on attaching mycophenolic acid moiety to adenosine through a variety of linkers to improve the cellular permeability of compounds compared to those carrying a pyrophosphate (Figure 11). The resulting analogs were tested toward hIMPDH 1 and 2 and as inhibitors of leukemia K562 cells proliferation. When the pyrophosphate of 2-ethyl TAD 48a was replaced by a difluoromethylenebisphosphonate (48b), it preserved the low nanomolar inhibition of hIMPDH 1 and 2, as well as increased the ability to penetrate cell membrane, resulting in micromolar inhibition of K562 cells proliferation (IC50 = 4.7 µM). In a similar way, 2-ethyl MAD analog with the same pyrophosphate modification (49a), improved its inhibitory activity toward hIMPDH1 and 2 (IC50 = 5 and 24 µM, respectively) compared to unmodified MAD 44, and displayed a potent inhibition of K562 cells proliferation (IC50 = 0.45 µM), which made it the most potent inhibitor of K562 cells proliferation described so far. Further development of ethylene bis(phosphonate) analogs to improve cell permeability led to even better inhibition of K562 cell growth (compounds 49b and 49c, IC50 = 4 and 10.5 µM, respectively). Phosphonophosphate (X = OCH2) containing compounds, despite being potent inhibitors of hIMPDH in vitro, showed low inhibitory effects on cells proliferation. Other modifications such as bis(sulfonamide), phosphonoacetamide, thioethylenephosphonate or triazole linkers also showed inhibitory activity, revealing that neither the tetrahedral geometry of phosphate nor the presence of phosphorus atom itself are necessary to maintain the inhibitory activity. Overall, the study showed Molecules 2019, 24, x 18 of 34 had a significant inhibition activity towards MtbIMPDH. In order to understand the differences in activity, the authors used a previously obtained NAD+ analogue (MAD 44, Figure 9) as a model for binding in the presence of inosine monophosphate (IMP), with the three IMPDHs. Computational studies combined with the X-ray crystal structure of the complex of NAD+ analog/IMPDH/IMP (hIMPDH2, Figure 7F) revealed aromatic interactions between adenine ring and enzyme for both human IMPDH, whereas no aromatic stacking was observed for MtbIMPDH. Moreover, a hydrogen bond was present in the N-subsite binding domain only with human IMPDH, where an unfavorable binding was calculated for MtbIMPDH. These differences in binding therefore explain the inhibition selectivity toward human IMPDH. Interestingly, similar computations performed using the new synthesized NAD+ analog (47b) revealed new favorable interactions in the P-subsite domain in addition to the absence of unfavorable ones, oppositely to the observations obtained when using the previous NAD+ analog. The lack of interactions in the A-subsite domain was, therefore, overcome thanks to these new compounds, and led to the first potent inhibitor of MtbIMPDH. Later, a new set of NAD+ analogs that are potential anticancer agents was described [69]. The authors mainly focused on attaching mycophenolic acid moiety to adenosine through a variety of linkers to improve the cellular permeability of compounds compared to those carrying a pyrophosphate (Figure 11). The resulting analogs were tested toward hIMPDH 1 and 2 and as inhibitors of leukemia K562 cells proliferation. When the pyrophosphate of 2-ethyl TAD 48a was replaced by a difluoromethylenebisphosphonate (48b), it preserved the low nanomolar inhibition of hIMPDH 1 and 2, as well as increased the ability to penetrate cell membrane, resulting in micromolar inhibition of K562 cells proliferation (IC50 = 4.7 µM). In a similar way, 2-ethyl MAD analog with the Molecules 2019, 24, 4187 18 of 34 same pyrophosphate modification (49a), improved its inhibitory activity toward hIMPDH1 and 2 (IC50 = 5 and 24 µM, respectively) compared to unmodified MAD 44, and displayed a potent thatinhibition the P-subsite of K562 cells is a promisingproliferation domain (IC50 = to0.45 introduce µM), which modifications, made it the whichmost potent can a ffinhibitorord potent of K562 new inhibitorscells proliferation of IMPDH described in vitro so and far. in vivo.

Figure 11. PotentPotent inhibitors inhibitors of IMPDH IMPDH and and K562 K562 cell cell pro proliferationliferation obtained by pyrophosphate pyrophosphate linkage modificationmodification in TAD andand MADMAD analogsanalogs [[69].69].

InFurther 2011, thedevelopment same group of focused ethylene on bis(phosphonate) the synthesis of NAD analogs+ analogs to impr withove pyridone cell permeability modification led [70 to]. Theeven study better was inhibition inspired, of among K562 others,cell growth by the fact(compounds that in the 49b 1970s, and 4-pyridone-3-carboxamide 49c, IC50 = 4 and 10.5 µM, was foundrespectively). in Phosphonophosphate from patients suffering of(X chronic= OCH myelogenous2) containing leukemia compounds, (CML), despite becoming being a potential potent markerinhibitors for of this h diseaseIMPDH [71in]. vitro, Likewise, showed the corresponding low inhibitory triphosphate effects on (4-pyridone-3-carboxamidecells proliferation. Other 5modifications0-triphosphate) such was as later bis(sulfon foundamide), in high phosphonoacetamide, concentration in erythrocytes thioethylenephosphonat of patients with chronice or triazole renal failurelinkers also (CRF). showed Therefore, inhibitory the authorsactivity, focusedrevealing on that the neither synthesis the tetrahedral of NAD+ analogs, geometry which of phosphate contain anor pyridone the presence modification of phosphorus at diff erentatom positionsitself are nece of thessary nicotinamide to maintainmoiety, the inhibitory as well activity. as a methylene Overall, bisphosphonate,the study showed instead that the of aP-subsite pyrophosphate is a promisin to improveg domain compound to introduce stability modifications, and cell permeability. which can affordTo potent that end, new three inhibitors NAD derivatives of IMPDH (Scheme in vitro5 and; 2-pyridone in vivo. 50a, 4-pyridone 50b and 6-pyridone 50c) wereIn synthesized 2011, the same through group a coupling focused reactionon the synthesis between of 50 NAD-adenosine+ analogs methylenebisphosphonate with pyridone modification and an[70]. appropriate The study isopropylidene-protectedwas inspired, among others, nucleoside by the fact in the that presence in the 1970s, of N,N 4-pyridone-3-carboxamide-diisopropylcarbodiimide (DIC).was found The synthesisin urine wasfrom first patients performed suffering following of chronic the pathway myelogenous B as represented leukemia (CML), with the becoming 4-pyridone a 50bpotential(Scheme marker5B, blue for pathway),this disease from [71]. the Likewise, corresponding the corresponding pyridone 51. Unfortunately,triphosphate (4-pyridone-3- when applied tocarboxamide the carboxyamide 5′-triphosphate) derivatives, was thelater coupling found in step high led concen to thetration cyano in derivatives erythrocytes (52b of inpatients the case with of 4-pyridone)chronic renal as failure a by-product. (CRF). Therefore, It was regrettably the authors impossible focused on to the separate synthesis the of by-product NAD+ analogs,53 from which the desiredcontain compounda pyridone50b modificationby HPLC. Therefore,at different they positions developed of the another nicotinamide synthetic pathwaymoiety, as (Scheme well as5C, a orange pathway) from the carboxylic acid derivative 54, a strategy which has also been applied to the 2-pyridone derivative 50a. Here, after the deprotection of acetyl groups and a methylation step affording the derivative 55, the 20,30-hydroxyls were protected with an isopropylidene. The resulting compound 56 was subsequently coupled with the adenosine moiety leading to the protected NAD analog 57 before a last step of removal of the 20,30-protecting group. The 6-pyridone-containing analog was obtained following the first pathway, because in this case, the by-product and the desired compound could be separated by HPLC. The compounds were subsequently tested towards human and Mycobacterium IMPDH and NAD kinase. Although no inhibition was found toward IMPDH, 6-pyridone NAD 50c showed equal inhibition toward both human and MtB NAD kinase (same IC50 = 0.5 mM) and 2-pyridone 50a was 8 times more potent inhibitor of hNADK than MtbNADK. Interestingly, they also found that 4-pyridone-3-carboxamide-adenine dinucleotide phosphate is an inhibitor of Toxoplasma gondi FNR enzyme (Ki = 30 µM). In 2014, the same group described three new classes of potent inhibitors of IMPDH [72], all based on previously described MAD 44 (Figure9)[ 68]. The exploration of various adenosine modifications and replacements of the pyrophosphate moiety with more lipophilic groups led to the development of a second generation of IMPDH inhibitors (Figure 12). Molecules 2019, 24, x 19 of 34

methylene bisphosphonate, instead of a pyrophosphate to improve compound stability and cell permeability. To that end, three NAD derivatives (Scheme 5; 2-pyridone 50a, 4-pyridone 50b and 6-pyridone 50c) were synthesized through a coupling reaction between 5′-adenosine methylenebisphosphonate and an appropriate isopropylidene-protected nucleoside in the presence of N,N- diisopropylcarbodiimide (DIC). The synthesis was first performed following the pathway B as represented with the 4-pyridone 50b (Scheme 5B, blue pathway), from the corresponding pyridone 51. Unfortunately, when applied to the carboxyamide derivatives, the coupling step led to the cyano derivatives (52b in the case of 4-pyridone) as a by-product. It was regrettably impossible to separate the by-product 53 from the desired compound 50b by HPLC. Therefore, they developed another synthetic pathway (Scheme 5C, orange pathway) from the carboxylic acid derivative 54, a strategy which has also been applied to the 2-pyridone derivative 50a. Here, after the deprotection of acetyl groups and a methylation step affording the derivative 55, the 2′,3′-hydroxyls were protected with an isopropylidene. The resulting compound 56 was subsequently coupled with the adenosine moiety leading to the protected NAD analog 57 before a last step of removal of the 2′,3′-protecting group. MoleculesThe 6-pyridone-containing2019, 24, 4187 analog was obtained following the first pathway, because in this case,19 of the 34 by-product and the desired compound could be separated by HPLC.

SchemeScheme 5. 5.( AA)) StructureStructure of of the the three three NAD NAD analogs analogs carrying carrying a pyridone a pyridone moiety; moiety; (B,C) B synthetic) and C) pathways synthetic

developedpathways fordeveloped the pyridone for the containing pyridone NAD containing analogs NAD [70]. Conditions:analogs [70]. a) Conditions: i. AMPCH2 P,a) pyridine, i. AMPCH DIC,2P, rt,pyridine, overnight, DIC, ii. rt, nucleoside overnight,51 ,ii. 70 nucleoside h, 65 ◦C, iii. 51, H 702O, h,TEA, 65 °C, 6 iii. h, 65H2◦O,C; TEA, b) 80% 6 h, TFA, 65 °C; 3 h, b) 0 ◦80%C; c) TFA, i. NH 33 h,in 0 MeOH,°C; c) i. 48NH h,3 rt;in ii.MeOH, MeOH, 48 TMSCH h, rt; ii.2 MeOH,N2 then TMSCH AcOH, 152N2 min, then rt; AcOH, d) dimethoxypropane, 15 min, rt; d) dimethoxypropane, acetone, p-TsOH, Molecules3acetone, h, 2019 rt; e., 24 NHp-, xTsOH, 3 in MeOH, 3 h, rt; 3e. days,NH3 in rt; MeOH, f) HCO 23H days, 50%, rt; overnight, f) HCO2H rt. 50%, overnight, rt. 20 of 34

The compounds were subsequently tested towards human and Mycobacterium IMPDH and NAD kinase. Although no inhibition was found toward IMPDH, 6-pyridone NAD 50c showed equal inhibition toward both human and MtB NAD kinase (same IC50 = 0.5 mM) and 2-pyridone 50a was 8 times more potent inhibitor of hNADK than MtbNADK. Interestingly, they also found that 4- pyridone-3-carboxamide-adenine dinucleotide phosphate is an inhibitor of Toxoplasma gondi FNR enzyme (Ki = 30 µM). In 2014, the same group described three new classes of potent inhibitors of IMPDH [72], all based on previously described MAD 44 (Figure 9) [68]. The exploration of various adenosine modifications and replacements of the pyrophosphate moiety with more lipophilic groups led to the development of a second generation of IMPDH inhibitors (Figure 12).

FigureFigure 12. 12. ThreeThree classes classes of of MPA MPA containing containing NAD NAD analogs synthesized by Felczak et al. [72]. [72].

TheThe first first group of analogs carried a substi substituenttuent at the C2 position of adenosine ( 58a–f). A A set set of of sixsix corresponding corresponding MAD MAD derivatives derivatives was was obtained obtained by by coupling coupling different different 2-substituted 2-substituted adenosines and and mycophenolicmycophenolic bis(phosphonate).bis(phosphonate).In In vitro vitroevaluation evaluation revealed revealed compounds compounds with highwith selectivityhigh selectivity toward towardhIMPDH hIMPDH 1. In particular, 1. In particular, the 4-pyridyl the 4-pyridyl analog analog58c was 58c 23 was times 23 times more more selective selective for hIMPDH for hIMPDH 1 than 1 than2 and 2 wasand sowas far so the far most the potentmost potent inhibitor inhibitor of hIMPDH1 of hIMPDH1 (Ki 0.6 (K nM).i 0.6 ThenM). compounds The compounds showed showed also a alsosignificantly a significantly improved improved antiproliferatory antiproliferatory activity towards activity cancer towards cells cancer (leukemia cells K562, (leukemia Hela cervicalK562, Hela cells, cervicaland colon cells, cancer and HT29colon cancer cells). TheHT29 crystal cells). structure The crystal of NADstructure+ bound of NAD to h+IMPDH2 bound to (FigurehIMPDH27E), (Figure which 7E),suggested which asuggested preference a preference for a strong for aromatic a strong system aromatic in Asystem sub-domain, in A sub-domain, led to another led to MAD another derivative MAD derivativewherein adenosine wherein wasadenosine replaced was by replaced 4-amino-benzimidazole. by 4-amino-benzimidazole. As expected, As this expected, compound this hadcompound a slight hadpreference a slight for preferencehIMPDH for 2, buthIMPDH with only 2, but three with times only morethree potenttimes more inhibition potent activity. inhibition The activity. second classThe secondof MAD-derived class of MAD-derived inhibitors was inhi obtainedbitors was by replacingobtained theby replacin pyrophosphateg the pyrophosphate with various esterswith various linkers. estersThe new linkers. MAD The derivatives new MAD were deriva synthesizedtives were both synthesized from l and bothd–adenosine from L and (59a D–b –). adenosine They both ( showed59a–b). They both showed an inhibition activity toward hIMPDH 1 and 2, but the D- was relatively stable toward , whereas the L-ester was quickly cleaved (few minutes). An attempt to replace the ester by amide, expected to be resistant to esterases, led to potent inhibitor of IMPDH. Unfortunately, the compound did not show any anti-proliferative activity, likely due to its incapacity to cross cell membranes. The last class of MAD derivatives explored in this work was amino acid diamide derivatives, in which an amino acid is placed between the ‘adenosine amide’ moiety and mycophenolic acid moiety, via an extra amide link (60a–h). Both the final compounds containing different amino acids as well as all the precursors of the synthetic pathway (free acids, methyl and adenosine derivatives) were tested as inhibitors of hIMPDH 1 and 2 and anti-proliferative agents. In general, the final MAD derivatives had nanomolar activity in vitro and were more potent than the corresponding precursors, but they did not show anti-proliferative activity, probably because they were not able to penetrate the cells. Overall, the results underlined how even minor structural alterations may dramatically influence the biological activity of the studied inhibitors.

4.2.2. NADK Targeting Nucleotides and Pronucleotides The NMN/Nicotinic acid mononucleotide (NaMN) adenylyl transferase and the NAD kinase are key enzymes involved in NAD metabolism. The first class of enzyme transforms NMN/NaMN into the corresponding NAD/NaAD using ATP. It is present in humans as three different isoforms, hNMNAT 1, 2 and 3, the first one is located in nucleus, whereas isoforms 2 and 3 are present in the Golgi complex and in mitochondria, respectively. The NAD kinase (NADK) is responsible for the 2′- O-phosphorylation of NAD(H), leading to NADP(H), using ATP or other sources of active phosphate. This enzymatic reaction is the only known pathway for the formation of NADP(H) in both eukaryotic and prokaryotic cells. Three different classes of NADK activity can be distinguished regarding its substrate specificity. The first one is represented by the human ATP-NAD kinase, which utilizes only ATP as a phosphoryl donor. The second one, found in gram positive bacteria, is ATP/ NAD kinase, where inorganic polyphosphate can be used as an alternative to ATP. The third one is NADK activity from gram-negative bacteria and single cellular eukaryotes,

Molecules 2019, 24, 4187 20 of 34 an inhibition activity toward hIMPDH 1 and 2, but the d-ester was relatively stable toward esterases, whereas the l-ester was quickly cleaved (few minutes). An attempt to replace the ester by amide, expected to be resistant to esterases, led to potent inhibitor of IMPDH. Unfortunately, the compound did not show any anti-proliferative activity, likely due to its incapacity to cross cell membranes. The last class of MAD derivatives explored in this work was amino acid diamide derivatives, in which an amino acid is placed between the ‘adenosine amide’ moiety and mycophenolic acid moiety, via an extra amide link (60a–h). Both the final compounds containing different amino acids as well as all the precursors of the synthetic pathway (free acids, methyl esters and adenosine derivatives) were tested as inhibitors of hIMPDH 1 and 2 and anti-proliferative agents. In general, the final MAD derivatives had nanomolar activity in vitro and were more potent than the corresponding precursors, but they did not show anti-proliferative activity, probably because they were not able to penetrate the cells. Overall, the results underlined how even minor structural alterations may dramatically influence the biological activity of the studied inhibitors.

4.2.2. NADK Targeting Nucleotides and Pronucleotides The NMN/Nicotinic acid mononucleotide (NaMN) adenylyl transferase and the NAD kinase are key enzymes involved in NAD metabolism. The first class of enzyme transforms NMN/NaMN into the corresponding NAD/NaAD using ATP. It is present in humans as three different isoforms, hNMNAT 1, 2 and 3, the first one is located in nucleus, whereas isoforms 2 and 3 are present in the Golgi complex and in mitochondria, respectively. The NAD kinase (NADK) is responsible for the 20-O-phosphorylation of NAD(H), leading to NADP(H), using ATP or other sources of active phosphate. This enzymatic reaction is the only known pathway for the formation of NADP(H) in both eukaryotic and prokaryotic cells. Three different classes of NADK activity can be distinguished regarding its substrate specificity. The first one is represented by the human ATP-NAD kinase, which utilizes only ATP as a phosphoryl donor. The second one, found in gram positive bacteria, is ATP/polyphosphate NAD kinase, where inorganic polyphosphate can be used as an alternative to ATP. The third one is NADK activity from gram-negative bacteria and single cellular eukaryotes, named as NTP-NADK because of its ability to use as a substrate NTPs other than ATP. The discovery of bacterial NADK activities, in particular, the enzymes from Mycobacterium tuberculosis and Listeria monocytogenes [73] made this NAD kinase an attractive antibiotic target. Moreover, human NADK is also a therapeutic target for cancer [74]. The therapeutic potential of NMNAT and NADK inhibitors has been reviewed in 2011 [75]. The first identified inhibitors of NADK are NdAD (61a, Figure 13) and NdADH [75], which both lack the 20-hydroxyl group, making the phosphorylation step impossible. Based on this fact, Petrelli et al. proposed other ribose 20-modifications, namely an inversion of configuration of the hydroxyl (61b) or the replacement of the hydroxyl by a fluorine substituent, in both arabino (61c) and ribo (61d) configuration [76]. The simple inversion of configuration was enough to block the phosphorylation step; however, all of these modifications led to rather weak inhibitory activity of human and Mtb NADK. A known NAD analog, the benzamide adenine dinucleotide (BAD), was also assessed and found to be a potent inhibitor of hNADK (Ki 90 µM). Based on the earlier findings of Poncet-Montagne et al. [73], who published the crystal structure of Listeria monocytogenes NADK in both apo and NAD(P) bound state, and described di-50-thio-adenosine (DTA 62a) as a potent inhibitor (Ki 20 µM), the authors also tested this compound against human and Mtb NADK. This analog 62a was a moderate inhibitor of these enzymes, with an IC50 of 45 µM and 87 µM, respectively. Consequently, they performed the synthesis of disulfide NAD analogs. Based on the structure of DTA and tiazofurin, which is a known inhibitor of IMPDH that is transformed in cells into TAD 45, the adenosine was replaced by tiazofurin to yield ASST (63) and TSST. While the first one showed quite potent inhibition of human and Mtb NADK (Ki= 110 µM and 80 µM, respectively), the di-tiazofurin compound was inactive, revealing the importance of at least one adenosine moiety. Molecules 2019, 24, x 21 of 34 named as NTP-NADK because of its ability to use as a substrate NTPs other than ATP. The discovery of bacterial NADK activities, in particular, the enzymes from Mycobacterium tuberculosis and Listeria monocytogenes [73] made this NAD kinase an attractive antibiotic target. Moreover, human NADK is also a therapeutic target for cancer [74]. The therapeutic potential of NMNAT and NADK inhibitors has been reviewed in 2011 [75]. The first identified inhibitors of NADK are NdAD (61a, Figure 13) and NdADH [75], which both lack the 2′-hydroxyl group, making the phosphorylation step impossible. Based on this fact, Petrelli et al. proposed other ribose 2′-modifications, namely an inversion of configuration of the hydroxyl (61b) or the replacement of the hydroxyl by a fluorine substituent, in both arabino (61c) and ribo (61d) configuration [76]. The simple inversion of configuration was enough to block the phosphorylation step; however, all of these modifications led to rather weak inhibitory activity of human and Mtb NADK. A known NAD analog, the benzamide adenine dinucleotide (BAD), was also assessed and found to be a potent inhibitor of hNADK (Ki 90 µM). Based on the earlier findings of Poncet-Montagne et al. [73], who published the crystal structure of Listeria monocytogenes NADK in both apo and NAD(P) bound state, and described di-5′-thio-adenosine (DTA 62a) as a potent inhibitor (Ki 20 µM), the authors also tested this compound against human and Mtb NADK. This analog 62a was a moderate inhibitor of these enzymes, with an IC50 of 45 µM and 87 µM, respectively. Consequently, they performed the synthesis of disulfide NAD analogs. Based on the structure of DTA and tiazofurin, which is a known inhibitor of IMPDH that is transformed in cells into TAD 45, the adenosine was replaced by tiazofurin to yield ASST (63) and TSST. While the first one showed quite potentMolecules inhibition2019, 24, 4187 of human and Mtb NADK (Ki= 110 µM and 80 µM, respectively), the di-tiazofurin21 of 34 compound was inactive, revealing the importance of at least one adenosine moiety.

Figure 13. NAD-derived inhibitors of NADK [76]. Figure 13. NAD-derived inhibitors of NADK [76]. Following these results, they next introduced a bromine atom on the 8 position of adenosine of Following these results, they next introduced a bromine atom on the 8 position of adenosine of DTA, forming either mono (62b) or di-substituted (62c) derivatives. Indeed, it has been described DTA, forming either mono (62b) or di-substituted (62c) derivatives. Indeed, it has been described that that this substituent can block the “syn” conformation of the corresponding compound, which could this substituent can block the “syn” conformation of the corresponding compound, which could increase the inhibitory potency if preferred by the enzyme. In this case, both mono and dibromo increase the inhibitory potency if preferred by the enzyme. In this case, both mono and dibromo compounds were found to inhibit hNADK (IC50 = 6 µM for both) and MtbNADK (IC50 = 19 µM and compounds were found to inhibit hNADK (IC50 = 6 µM for both) and MtbNADK (IC50 = 19 µM and 14 14 µM, respectively) in a similarly potent way. Thereby, it can be assumed that this enzyme has indeed a µM, respectively) in a similarly potent way. Thereby, it can be assumed that this enzyme has indeed preference for the syn conformation. On the other hand, compounds modified by introducing a methyl a preference for the syn conformation. On the other hand, compounds modified by introducing a group at either the 20- or 30-position on both ribose moieties in order to fix the sugar conformation at either the 2′- or 3′-position on both ribose moieties in order to fix the sugar showed a weaker inhibition of both human and MtbNADK than the unmodified parent compound conformation showed a weaker inhibition of both human and MtbNADK than the unmodified parent (DTA, 62a) or even no activity at all (30-methyl derivative). compound (DTA, 62a) or even no activity at all (3′-methyl derivative). The group of Bertino and coworkers explored the anticancer properties of NADK inhibitors. The group of Bertino and coworkers explored the anticancer properties of NADK inhibitors. This was in contrast to previous studies that, except for a few examples [75,76], focused mainly on This was in contrast to previous studies that, except for a few examples [75,76], focused mainly on infectious diseases. Since NADPH plays a crucial role in the synthesis of nucleic acids and lipids, infectious diseases. Since NADPH plays a crucial role in the synthesis of nucleic acids and lipids, cancer cells require it for growth and proliferation. The deficiency of NAD(P)H can result in inhibition cancer cells require it for growth and proliferation. The deficiency of NAD(P)H can result in of tumor cells growth or even induce their death. NADPH, as a reducing agent, is also responsible inhibition of tumor cells growth or even induce their death. NADPH, as a reducing agent, is also for the regulation of reactive oxygen species (ROS) levels, wherein moderate levels of these species responsible for the regulation of reactive oxygen species (ROS) levels, wherein moderate levels of are beneficial for cancer cells, while excess induces apoptosis. In 2013, Hsieh et al. [77] reported the these species are beneficial for cancer cells, while excess induces apoptosis. In 2013, Hsieh et al. [77] degradation of (DHFR) as a result of NADK inhibition by a newly identified reported the degradation of dihydrofolate reductase (DHFR) as a result of NADK inhibition by a NADK inhibitor, thionicotinamide adenine dinucleotide phosphate (NADPS). The DHFR is an NADP+ newly identified NADK inhibitor, thionicotinamide adenine dinucleotide phosphate (NADPS). The dependent enzyme playing a crucial role in DNA and RNA synthesis. Thus, the inhibition of DHFR DHFR is an NADP+ dependent enzyme playing a crucial role in DNA and RNA synthesis. Thus, the leads to the inhibition of the nucleic acids synthesis, and thereby to cell death, making it an important therapeutic target for cancer. Notably, DHFR is the molecular target for methotrexate (MTX)—a drug widely used in cancer treatment. Because NADK is the only enzyme producing cellular NADP(H) from NAD(H), the authors envisaged that its inhibition should result in decrease of the NADPH pool, necessary for DHFR stability, and consequently inducing DHFR degradation. The authors assessed 14 different potential inhibitors of human cytosolic NADK starting with benzamide riboside. Although this compound showed cNADK inhibition in vitro, and therefore decreased NADPH levels, toxicity in animals was revealed by in vivo tests. Among the other 13 NAD(P) analogs tested, only the two compounds, thionicotinamide adenine dinucleotide and its 30-phosphate (NADS 64 and NADPS 65, respectively, Figure 14) were potent inhibitors of NADK (ED50 1 µM). The equal effectiveness of both compounds was explained by occurrence of 30-phosphorylation of NADS to NADPS by NADK, designating in both cases the phosphorylated form as the active one. The inhibitors were next tested on two different cell lines, a metastatic human colon cancer cell line (C85 cells), and T-cell lymphoblastic leukemia cells. First, it was found that the inhibition of NADK by NADPH is nicotinamide dose dependent in a reverse manner. Next, by comparing C85 cells and normal cells, it was confirmed that NADPS exhibits a selective activity toward cancer cells. Additionally, by testing the effect of NADPS on three different dehydrogenases (GADPH, IMPDH and ALDH1a) in C85 cells, they observed that there was no change in levels of these enzymes, confirming its specificity for NADK/DHFR. Further, the molecular mechanism behind the decrease in DHFR level induced by NADPS was studied. Neither DHFR transcription nor translational upregulation was involved in DHFR decrease, confirming that lowering NADPH pool by NADK inhibition is responsible Molecules 2019, 24, x 22 of 34

inhibition of DHFR leads to the inhibition of the nucleic acids synthesis, and thereby to cell death, making it an important therapeutic target for cancer. Notably, DHFR is the molecular target for methotrexate (MTX)—a drug widely used in cancer treatment. Because NADK is the only enzyme producing cellular NADP(H) from NAD(H), the authors envisaged that its inhibition should result Moleculesin decrease2019, 24of, 4187the NADPH pool, necessary for DHFR stability, and consequently inducing DHFR22 of 34 degradation. The authors assessed 14 different potential inhibitors of human cytosolic NADK starting with benzamide riboside. Although this compound showed cNADK inhibition in vitro, and therefore fordecreased this observation. NADPH levels, Indeed, toxicity DHFR in is animals stabilized was by revealed NADPH by and in depletion vivo tests. of Among NADPH the makes other this 13 proteinNAD(P) more analogs susceptible tested, only to degradation. the two compounds, Finally, athionicotinamide synergistic cell death adenine effect dinucleotide was observed and whenits 3′- NADPSphosphate was (NADS combined 64 and with NADPS methotrexate, 65, respectively, suggesting Figure that NADK 14) were inhibition potent inhibitors could be a of promising NADK (ED way50 + to1 overcomeµM). The methotrexate-resistance. equal effectiveness of A modelboth forcompounds NAD(P) complexwas explained with NADK by wasoccurrence also developed, of 3′- + + + revealingphosphorylation that NADP of NADSshows to NADP less stableS by NADK, binding designating than NAD in, but both the cases binding the phosphorylated of NADPS is moreform favorableas the active to aone. higher number of stabilizing van der Waals interactions.

Figure 14.14. NAD analogs studied as NADK inhibitors inducinginducing DHFR degradation [[77]77] and G6PD inhibition [[78].78].

InThe 2015, inhibitors Tedeschi were et al.next [78 tested] studied on two the edifferenffect of thionicotinamidet cell lines, a metastatic (TN 66 human, Figure colon 14) on cancer cytosolic cell NADPHline (C85 pool, cells), expecting and T-cell it to lymphoblastic both inhibit NAD-dependent leukemia cells. metabolicFirst, it was pathways found andthatimpair the inhibition the control of ofNADK ROS by species, NADPH leading is nicotinamide to cell death. dose They dependent showed in that a reverse TN acted manner. as a pronucleotide Next, by comparing penetrating C85 ecellsfficiently and normal into cells cells, and it was being confirmed transformed that NADPS into NADS exhibits and a NADPS, selective which activity inhibited toward cancer NADK cells. and glucoseAdditionally, 6-phosphate by testing dehydrogenase the effect of (G6PD).NADPS They on three also different showed that dehydrogenases nicotinamide (GADPH, lowers the IMPDH toxicity ofand TN ALDH1a) in C85 cells, in indicatingC85 cells, thatthey TN observed selectively that targets there NAD-relatedwas no change pathways. in levels Treatment of these of enzymes, C85 cells withconfirming 100 µM its TN specificity resulted in for NADK NADK/DHFR. inhibition andFurther, decrease the molecular in NAD(P)H mechanism pool overtime behind (up the to 60–70% decrease in lessin DHFR in 24 h). level Additionally, induced by a significant NADPS depletionwas studied. of fatty Neither acids andDHFR inhibition transcription of protein nor synthesis translational were observed.upregulation TN alonewas involved had no significant in DHFR influence decrease, on confirming ROS levels, that but inlowering combination NADPH with pool oxidative by NADK stress (Hinhibition2O2), the is ROS responsible levels drastically for this observation. increased, suggesting Indeed, DHFR loss ofis controlstabilized on ROSby NADPH levels. Similarand depletion results wereof NADPH obtained makes when this using protein TN more simultaneously susceptible to with degradation. either a chemotherapeutic Finally, a synergistic drug cell (gemcitabine, death effect docetaxel,was observed irinotecan) when orNADPS menadione, was combined all known with to generate methotrexate, ROS. TN suggesting did not show that neurotoxicity NADK inhibition in mice xenograftcould be a model.promising It is way an improvement to overcome comparedmethotrexate-resistance. to 6-aminonicotinamide, A model for a knownNAD(P) G6PD+ complex inhibitor, with whichNADK showed was also severe developed, neurotoxicity revealing during that anti-cancer NADP+ shows clinical less trial. stable This studybinding further thana NADffirmed+, but NADK the asbinding a promising of NADPS target+ is for more cancer favorable treatment. to a higher number of stabilizing van der Waals interactions. In 2015, Tedeschi et al. [78] studied the effect of thionicotinamide (TN 66, Figure 14) on cytosolic 4.2.3. Miscellaneous and Unidentified Molecular Targets NADPH pool, expecting it to both inhibit NAD-dependent metabolic pathways and impair the controlNAD of analogsROS species, with biological leading to activity cell death. havealso They been showed synthesized that TN for otheracted molecularas a pronucleotide targets or withoutpenetrating a specific efficiently target into in mind.cells and A series being of transformed adducts of aninto anti-turberculosis NADS and NADPS, drug, which isoniazid inhibited (INH, FigureNADK 15 and), with glucose truncated 6-phosphate NAD moietydehydrogenase modified (G6PD). with a lipophilic They also hydrocarbon showed that chainnicotinamide was developed lowers bythe Delainetoxicityet of al. TN [79 ].in Indeed,C85 cells, it has indicating been reported that TN earlier selectively that the targets active formNAD-related of isoniazid pathways. creates INH-NADTreatment of adduct, C85 cells which with inhibits 100 µM the TN enoyl-ACP resulted in reductase NADK inhibition (InhA) enzyme and decrease from Mtb in NAD(P)H, leading to pool the blockovertime of fatty (up acidto 60–70% elongation. in less Thereby in 24 the h). biosynthesis Additionally, of thea significant mycolic acid, depletion essential of compound fatty acids of and the mycobacteriainhibition of protein envelop synthesis is stopped. were However, observed. resistance TN alone to INH had hasno significant been increasingly influence observed, on ROS and levels, can bebut explained in combination either bywith a mutationoxidative instressMtb ,(H which2O2), the decreases ROS levels the ability drastically of INH increased, to form asuggesting complex with loss NADof control or by on a mutationROS levels. of Similar the enzyme, results which were interacts obtained with when diphosphate using TN simultaneously moiety of NADH. with Therefore, either a thechemotherapeutic authors designed drug and (gemcitabine, synthesized adocetaxel, set of compounds irinotecan) with or modifiedmenadione, and all truncated known to INH-NAD generate moiety,ROS. TN to did overcome not show the neurotoxicity resistance coming in mice fromxenogr theaft diphosphate model. It is an moiety. improvement The first compared generation to of6- truncated INH-NAD analogs reported in a preceding work did not afford sufficient inhibition of InhA. Hence in this work, in order to enhance the affinity and the selectivity for InhA catalytic site, a lipophilic chain was added to the nicotinamide pattern using Suzuki-Miayura cross-coupling reaction. To that end, three different classes of nicotinamide derivatives (pyridines 67, pyridinums 68 and Molecules 2019, 24, x 23 of 34 aminonicotinamide, a known G6PD inhibitor, which showed severe neurotoxicity during anti-cancer clinical trial. This study further affirmed NADK as a promising target for cancer treatment.

4.2.3. Miscellaneous and Unidentified Molecular Targets NAD analogs with biological activity have also been synthesized for other molecular targets or without a specific target in mind. A series of adducts of an anti-turberculosis drug, isoniazid (INH, Figure 15), with truncated NAD moiety modified with a lipophilic hydrocarbon chain was developed by Delaine et al. [79]. Indeed, it has been reported earlier that the active form of isoniazid creates INH-NAD adduct, which inhibits the enoyl-ACP reductase (InhA) enzyme from Mtb, leading to the block of fatty acid elongation. Thereby the biosynthesis of the mycolic acid, essential compound of the mycobacteria envelop is stopped. However, resistance to INH has been increasingly observed, and can be explained either by a mutation in Mtb, which decreases the ability of INH to form a complex with NAD or by a mutation of the enzyme, which interacts with diphosphate moiety of NADH. Therefore, the authors designed and synthesized a set of compounds with modified and truncated INH-NAD moiety, to overcome the resistance coming from the diphosphate moiety. The first generation of truncated INH-NAD analogs reported in a preceding work did not afford sufficient inhibition of InhA. Hence in this work, in order to enhance the affinity and the selectivity for InhA Moleculescatalytic2019 site,, 24 a, 4187lipophilic chain was added to the nicotinamide pattern using Suzuki-Miayura 23cross- of 34 coupling reaction. To that end, three different classes of nicotinamide derivatives (pyridines 67, pyridinums 68 and dihydropyridine 69, Figure 15) were prepared and assessed as InhA and bacterial dihydropyridinegrowth inhibitors.69 As, Figure a result, 15) five were of prepared these compounds and assessed (67a, as 68a InhA–c, 69, and Figure bacterial 14) growthwere identified inhibitors. as Aspotent a result, InhA five inhibitors. of these compounds (67a, 68a–c, 69, Figure 14) were identified as potent InhA inhibitors.

FigureFigure 15.15. Three classesclasses ofof isoniazid-nicotinamideisoniazid-nicotinamide adduct analogs developed as inhibitors of InhAInhA fromfrom MycobacteriumMycobacterium tuberculosis tuberculosis:: A A - pyridine,- pyridine, B B - pyridinium,- pyridinium, C -C dihydropyridine; - dihydropyridine; and and structure structure of the of mostthe most relevant relevant inhibitors inhibitors [79, 80[79,80].].

InIn theirtheir nextnext paper,paper, anotheranother setset ofof thosethose derivativesderivatives (compounds(compounds 67a–c, 68a, 69)) werewere testedtested againstagainst Corynebacterium glutamicum glutamicum, ,which which does does have have InhA InhA activity, activity, EscherichiaEscherichia Coli Coli (as(as possessing possessing an anInhA InhA homolog, homolog, it would it would test mycobacteria test mycobacteria selectivity) selectivity) and Plasmodium and Plasmodium Falciparum Falciparum (malaria pathogen(malaria pathogenagent) [80]. agent) The pyridine [80]. The 67a pyridine inhibited67a specificallyinhibited specificallyMtb/ MycobacteriaMtb/ Mycobacteria smegmatis (Ms) smegmatis and not(Ms) C. andglutamicum not C. glutamicum, in the same, in way the as same INH, way indicating as INH, that indicating biological that activity biological of this activity compound of this arises compound solely arisesfrom InhA solely inhibition. from InhA In inhibition.contrast, two In other contrast, compounds two other (68a compounds and 69) inhibited (68a and both69) Mtb inhibited/Ms and both C. Mtbglutamicum/Ms and, C.proving glutamicum that InhA, proving enzyme that is InhA not enzymethe only isone not targeted the only by one the targeted compounds. by the compounds. InIn 2009,2009, aa newnew methodmethod toto synthesizesynthesize nicotinamide-containingnicotinamide-containing phosphodiesterphosphodiester analogsanalogs waswas developeddeveloped by by Liu Liu et et al. al. [81]. [81 They]. They designed designed a set aof set fourteen of fourteen NMN NMNderivatives derivatives (Figure (Figure16) to provide 16) to providetools for toolsbiological for biological studies of studies NAD-dependent of NAD-dependent processes. processes. The analogs The were analogs functionalized were functionalized within the withinphosphate the phosphategroup with group an alkane with or an an alkane aromatic or an substituent aromatic substituentlinked to the linked nicotinamide to the nicotinamide moiety by a moietyphosphodiester by a phosphodiester bond. The synthesis bond. Thewas synthesisbased on wasa key based reaction on a between key reaction the nicotinamide between the nicotinamidemonophosphate monophosphate and the corresponding and the corresponding alcohol. These alcohol. products These were products additionally were additionally2′- and 3′-O 2acetylated0- and 30-O to acetylated afford better to aff ordsolubility better solubilityand higher and reactivity. higher reactivity. Among Amongall the reagents all the reagents tested, tested,2,4,6- Molecules2,4,6-triisopropylbenzenesulfonyltriisopropylbenzenesulfonyl 2019, 24, x chloride chloride (TIPS-Cl) (TIPS-Cl) was the was most the mosteffective effective as a ascondensation a condensation agent24 agent of and 34 andwas wasused used under under optimized optimized conditions. conditions. Preliminary Preliminary biological biological assays assays showed showed that thatcompound compound 70 (10070 µM)(100 hadµM) a had beneficial a beneficial effect eonffect the on growth the growth of Eschericia of Eschericia Coli (DH5 Coliα(DH5) and αSaccaromyces) and Saccaromyces cerevisiae cerevisiae (ATCC 26108).(ATCC 26108).The compound The compound has no effect has no on e fftheect activity on the activityof S. cerevisiae of S. cerevisiae alcohol alcoholdehydrogenase. dehydrogenase.

Figure 16. PhosphodiesterPhosphodiester analogs analogs tested tested as as bacterial bacterial an andd yeast yeast growth growth enhancers, enhancers, including the most potent compound 70 [[81].81].

2+ NaADP is actingacting asas aa secondsecond messenger messenger to to stimulate stimulate Ca Ca2+release. release. This This process process has has been been proved proved in inmany many human human systems, systems, but thebut details the details of molecular of mechanismcular mechanism behind itbehind are still it unraveled. are still unraveled. Therefore, 2+ Therefore,to study the to NaADP-mediatedstudy the NaADP-mediated control of Ca controlrelease, of Ca the2+ release, need of the NaADP need analogsof NaADP is required analogs for is requiredfurther biological for further studies biological as affi nitystudies binding as affinity experiments binding or experiments labelling through or labelling a click reaction.through a To click this reaction.end, the substitutionTo this end, the of the substitution 4 and 5-position of the 4 ofan thed 5-position nicotinic of acid the was nicotinic explored acid bywas Jain explored et al. [ 82by]. Jain et al. [82]. They revealed that small 5-substituents as methyl, ethyl, azido or amino were tolerated, whereas the 4-substitution led to agonists with low potency. In a more recent work, Trabbic et al. [83] designed a set of 11 new analogs, carrying modification at the 5 position of the nicotinic acid and/or at the 8-position of adenosine (Figure 17). After the assessment of spectroscopic properties, they evaluated ability of the compounds to affect three different processes from Strongylocentrotus purpuratus (sea urchin egg, a model of NaADP-mediated Ca2+ release in mammals). Firstly, they checked their aptitude of releasing Ca2+ using a fluorometric assay where the Ca2+ release is linked to an increase in fluorescence. A competition ligand binding assay revealed that the analog efficiently competed with 32P-NaADP. Finally, an experiment evaluating the ability of the compounds to induce subthreshold receptor desensitization was performed. As a result, the 5-substituted derivatives with small substituents were identified as highly potent agonists (71a–c, Figure 17). Moreover, they showed that 8-subsituted compounds, in particular the 8-azido 71d and the disubstituted derivative 71e also displayed potent agonist activity at low concentration.

Figure 17. The most potent agonist-type NAD+ analogs developed by Trabbic et al. [83].

5. NAD Analogs as Alternative Redox Co-Factors

NAD+ analogs can also be considered as alternative or even biorthogonal co-factors for cellular redox processes. Such analogs can aid better understanding of enzymatic reactions and provide new hydride donors/acceptors useful for enzymatic or chemical transformations. This topic has been quite recently reviewed by Paul and co-workers [84]; therefore, here we will only highlight a few select examples.

Molecules 2019, 24, x 24 of 34

µM) had a beneficial effect on the growth of Eschericia Coli (DH5α) and Saccaromyces cerevisiae (ATCC 26108). The compound has no effect on the activity of S. cerevisiae alcohol dehydrogenase.

Figure 16. Phosphodiester analogs tested as bacterial and yeast growth enhancers, including the most potent compound 70 [81].

NaADP is acting as a second messenger to stimulate Ca2+ release. This process has been proved in many human systems, but the details of molecular mechanism behind it are still unraveled. Therefore, to study the NaADP-mediated control of Ca2+ release, the need of NaADP analogs is Moleculesrequired2019 for, 24 further, 4187 biological studies as affinity binding experiments or labelling through a24 click of 34 reaction. To this end, the substitution of the 4 and 5-position of the nicotinic acid was explored by Jain et al. [82]. They revealed that small 5-substituents as methyl, ethyl, azido or amino were tolerated, Theywhereas revealed the 4-substitution that small 5-substituents led to agonists as with methyl, low ethyl, potency. azido or amino were tolerated, whereas the 4-substitutionIn a more ledrecent to agonists work, Trabbic with low et al. potency. [83] designed a set of 11 new analogs, carrying modification at theIn 5 a position more recent of the work, nicotinic Trabbic acid et al. and/or [83] designed at the 8-position a set of 11 of new adenosin analogs,e (Figure carrying 17). modification After the atassessment the 5 position of spectroscopic of the nicotinic properties, acid and they/or ateval theuated 8-position ability of of adenosine the compounds (Figure to17 ).affect After three the assessmentdifferent processes of spectroscopic from Strongylocentrotus properties, they purpuratus evaluated ability (sea urchin of the egg, compounds a model to of a ffNaADP-mediatedect three different 2+ processesCa2+ release from in mammals).Strongylocentrotus Firstly, they purpuratus checked(sea their urchin aptitude egg, of a releasing model of Ca NaADP-mediated2+ using a fluorometric Ca 2+ releaseassay where in mammals). the Ca2+ Firstly,release theyis linked checked to an their increase aptitude in fluorescence. of releasing Ca A competitionusing a fluorometric ligand binding assay 2+ whereassay therevealed Ca releasethat the is linkedanalog to efficiently an increase competed in fluorescence. with 32 AP-NaADP. competition Finally, ligand an binding experiment assay 32 revealedevaluating that the the ability analog of effi theciently compounds competed to with induP-NaADP.ce subthreshold Finally, receptor an experiment desensitization evaluating was the abilityperformed. of the As compounds a result, the to 5-substituted induce subthreshold derivatives receptor with small desensitization substituents was were performed. identified As as a highly result, thepotent 5-substituted agonists (71a derivatives–c, Figure with 17). Moreover, small substituents they showed were that identified 8-subsituted as highly compounds, potent agonists in particular (71a–c, Figurethe 8-azido 17). Moreover, 71d and the they disubstituted showed that 8-subsitutedderivative 71e compounds, also displayed in particular potent theagonist 8-azido activity71d and at low the disubstitutedconcentration. derivative 71e also displayed potent agonist activity at low concentration.

FigureFigure 17.17. The most potent agonist-type NAD + analogsanalogs developed developed by by Trabbic Trabbic et al. [83]. [83].

5. NAD Analogs as Alternative Redox Co-Factors + NAD+ analogsanalogs can can also also be be considered considered as alternative or even biorthogonal co-factors for cellular redox processes. Such analogs can aid better unde understandingrstanding of enzymatic reactions and provide new hydride donors/acceptors donors/acceptors useful useful for for enzymatic enzymatic or orchem chemicalical transformations. transformations. This Thistopic topic has been has quite been quiterecently recently reviewed reviewed by Paul by and Paul co-workers and co-workers [84]; therefore, [84]; therefore, here we here will we only will highlight only highlight a few aselect few selectexamples. examples. In 2011, Hou et al. [85] designed 13 NAD+ analogs, where the adenosine moiety was replaced by a 1,2,3-triazole moiety (Figure 18A). The ability of the compounds to be used as cofactors was then tested and revealed activity in the presence of malic enzyme and alcohol dehydrogenase, suggesting flexibility in the NAD-binding pockets of those enzymes. Later, by screening those analogs against different mutants of malic enzyme, they identified a mutant (ME-L30K/L404S) characterized by over 1000-fold + preference for analog 72 over NAD [86]. Thus, the study laid foundation for the development of an orthogonalMolecules 2019 redox, 24, x system with potential for engineering. 25 of 34

FigureFigure 18.18. NAD + analogsanalogs developed developed as cofactors for orthogonal redoxredox systems.systems. (AA)) Triazole Triazole derivative 7272 [[85,86];85,86]; (BB)) NCD NCD and and analogs analogs [87–89]; [87–89]; C (C) )Nicotinamide Nicotinamide modified modified compounds compounds [90]. [90].

In 2011, Hou et al. [85] designed 13 NAD+ analogs, where the adenosine moiety was replaced by a 1,2,3-triazole moiety (Figure 18A). The ability of the compounds to be used as cofactors was then tested and revealed activity in the presence of malic enzyme and alcohol dehydrogenase, suggesting flexibility in the NAD-binding pockets of those enzymes. Later, by screening those analogs against different mutants of malic enzyme, they identified a mutant (ME-L30K/L404S) characterized by over 1000-fold preference for analog 72 over NAD+ [86]. Thus, the study laid foundation for the development of an orthogonal redox system with potential for metabolic pathway engineering. Ji et al. [87] focused on the development of NCD and analogs (Figure 18B, compounds 73a–e). NCD 73a and NFCD 73b were found to display high activity as cofactors for a particular mutant of malic enzyme (ME-L310R/G401C). They later developed other analogs, NClCD 73c, NBrCD 73d and NMeCD 73e [88], which in the same way showed good orthogonality for the natural pair ME/NAD+, except for the last one (NMeCD). Recently [89], NAD+ analog 73a was also found to be used as a non- natural cofactor by the phosphite dehydrogenase (Pdh) mutants, revealing a preference for NCD over NAD. Another class of analogs was developed by Lee et al. [90], where the amide group of the nicotinamide moiety of NAD was replaced by different substituents (Figure 18C). These derivatives were used as artificial electron carriers for photo-enzymatic synthesis under visible light. The study revealed that the compounds 74a and 74b were the most efficient reducing agents.

6. NAD-Derived Chemical Tools to Study NAD-Capped RNAs

In eukaryotic cells, the 5′ end of mRNA is protected with 7-methyl guanosine (m7G) via a 5′-5- triphosphate bridge. More recently, it has been discovered that both in bacteria and in higher organisms, including yeast, mammalian cells, and plants, NAD+ is present at the 5′-end of some RNAs [9,12,14,15,17,91,92]. However, the biogenesis and biological functions of this RNA modification are still unraveled. To better understand its role, several groups have worked on this newly discovered structure, to provide a way for the chemical synthesis of NAD-RNA models [93], to create methods for detection and quantification of NAD-capped RNAs in cells [9,94,95], and to modulate its susceptibility to different enzymes [96].

6.1. Chemoenzymatic Methods of NAD-RNA Detection Since the discovery of NAD-capped (or NAD-linked) RNA, several groups have been working on new ways to detect it in cells, since the initially established MS-based method is time-consuming and requires large amounts of RNA [17]. A widely-adopted chemo-enzymatic method for identification of NAD-capped RNAs taking advantage of the ADP-ribosyl cyclase (ADPRC) activity was developed by Cahova et al. [9]. The method involves a substitution of the terminal nicotinamide moiety in NAD-RNA by an alcohol with terminal alkyne group through ADPRC action. Only RNAs terminated with 5′-NAD+ undergo this transformation. The resulting alkyne-conjugated RNA is then

Molecules 2019, 24, 4187 25 of 34

Ji et al. [87] focused on the development of NCD and analogs (Figure 18B, compounds 73a–e). NCD 73a and NFCD 73b were found to display high activity as cofactors for a particular mutant of malic enzyme (ME-L310R/G401C). They later developed other analogs, NClCD 73c, NBrCD 73d and NMeCD 73e [88], which in the same way showed good orthogonality for the natural pair ME/NAD+, except for the last one (NMeCD). Recently [89], NAD+ analog 73a was also found to be used as a non-natural cofactor by the phosphite dehydrogenase (Pdh) mutants, revealing a preference for NCD over NAD. Another class of analogs was developed by Lee et al. [90], where the amide group of the nicotinamide moiety of NAD was replaced by different substituents (Figure 18C). These derivatives were used as artificial electron carriers for photo-enzymatic synthesis under visible light. The study revealed that the compounds 74a and 74b were the most efficient reducing agents.

6. NAD-Derived Chemical Tools to Study NAD-Capped RNAs

7 In eukaryotic cells, the 50 end of mRNA is protected with 7-methyl guanosine (m G) via a 50-5-triphosphate bridge. More recently, it has been discovered that both in bacteria and in higher + organisms, including yeast, mammalian cells, and plants, NAD is present at the 50-end of some RNAs [9,12,14,15,17,91,92]. However, the biogenesis and biological functions of this RNA modification are still unraveled. To better understand its role, several groups have worked on this newly discovered structure, to provide a way for the chemical synthesis of NAD-RNA models [93], to create methods for detection and quantification of NAD-capped RNAs in cells [9,94,95], and to modulate its susceptibility to different enzymes [96].

6.1. Chemoenzymatic Methods of NAD-RNA Detection Since the discovery of NAD-capped (or NAD-linked) RNA, several groups have been working on new ways to detect it in cells, since the initially established MS-based method is time-consuming and requires large amounts of RNA [17]. A widely-adopted chemo-enzymatic method for identification of NAD-capped RNAs taking advantage of the ADP-ribosyl cyclase (ADPRC) activity was developed by Cahova et al. [9]. The method involves a substitution of the terminal nicotinamide moiety in NAD-RNA by an alcohol with terminal alkyne group through ADPRC action. Only RNAs terminated with + 50-NAD undergo this transformation. The resulting alkyne-conjugated RNA is then subjected to a click reaction with biotin azide. The last step consists of isolation of biotinylated RNAs on streptavidin beads and the consequent analysis of the RNA by high throughput sequencing. More recently, Vvedenskaya and co-workers [94] developed CapZyme-Seq method to detect and quantify not only NAD-capped RNA but also other non-canonical initiating nucleotides (NCIN), as FAD- and dpCoA- capped RNAs. The first step is the processing of the RNA using NudC or Rai1 enzyme, both releasing 50-P-RNA by either a pyrophosphatase or a phosphodiesterase activity, respectively. Then a high throughput sequencing is used to quantify the capped-RNA. NAD-capQ is another method for the detection and quantification of NAD-RNAs developed by Grudzien-Nogalska et al. [95]. NAD-capQ is a two-step procedure, which involves enzymatic treatment of the RNA with P1, resulting in the phosphodiester bonds cleavage, followed + by the quantification of the released 50-NAD through a commercially available colorimetric assay. Thus, the method does not require any specialized laboratory equipment (except for a simple spectrophotometer) and is rapid in comparison to the previous method, but does not provide information on the sequences of the quantified RNAs.

6.2. Synthesis of Molecular Tools to Study NAD-Capped RNAs Since this area of research has emerged quite recently, not many NAD-derived tools to study NAD-RNAs have been developed so far. The first chemical synthesis of NAD capped RNA was described by Hofer et al. [93], as an alternative to enzymatic synthesis using in vitro transcription (IVT) [97]. The IVT method requires the use of NAD+ in large excess and leads to a mixture of NAD- as Molecules 2019, 24, x 26 of 34 subjected to a click reaction with biotin azide. The last step consists of isolation of biotinylated RNAs on streptavidin beads and the consequent analysis of the RNA by high throughput sequencing. More recently, Vvedenskaya and co-workers [94] developed CapZyme-Seq method to detect and quantify not only NAD-capped RNA but also other non-canonical initiating nucleotides (NCIN), as FAD- and dpCoA- capped RNAs. The first step is the processing of the RNA using NudC or Rai1 enzyme, both releasing 5′-P-RNA by either a pyrophosphatase or a phosphodiesterase activity, respectively. Then a high throughput sequencing is used to quantify the capped-RNA. NAD-capQ is another method for the detection and quantification of NAD-RNAs developed by Grudzien-Nogalska et al. [95]. NAD-capQ is a two-step procedure, which involves enzymatic treatment of the RNA with nuclease P1, resulting in the phosphodiester bonds cleavage, followed by theMolecules quantification2019, 24, 4187 of the released 5′-NAD+ through a commercially available colorimetric assay. 26Thus, of 34 the method does not require any specialized laboratory equipment (except for a simple spectrophotometer) and is rapid in comparison to the previous method, but does not provide well as 5 -triphosphate RNA, which are difficult to separate. Hofer et al. [93] used either an 5 -P-RNA information0 on the sequences of the quantified RNAs. 0 20 mer obtained by solid phase synthesis or 38 mer prepared from in vitro transcription, followed by 6.2.pyrophosphatase Synthesis of Molecular treatment Tools to atoff ordStudy the NAD-Capped 50-monophosphate RNAs from 50-triphosphate (Figure 19).

Figure 19. NAD-RNANAD-RNA synthesis synthesis strategy strategy according according to Hofer et al. [93]. [93].

SinceTo add this NMN area moiety of research to these has RNAs, emerged they quite synthesized recently, nicotinamide not many NAD-derived 50-phosphorimidazolide tools to study by NAD-RNAsactivating NMN have using been CDIdeveloped in DMF so followed far. The byfirst a carbonatechemical hydrolysissynthesis of step. NAD In capped the last RNA step, theywas describedused 1000 byexcess Hofer of et the al. resulting [93], as an NMN-Im alternative in coupling to enzymatic reaction synthesis with 50-P-RNA using in in vitro an aqueous transcription media × 2+ (IVT)containing [97]. The Mg IVT. This method step ledrequires to the the mixture use of of NAD NAD-RNA+ in large and excess unreacted and leads 50-P-RNA. to a mixture This latter of NAD- was ashowever well as easily 5′-triphosphate removed by RNA, treatment which with are difficult Xrn-1, a nucleaseto separate. which Hofer uses et only al. [93] 50-P-RNA used either as substrates. an 5′-P- RNAThe authors 20 mer usedobtained the sameby solid method phase to synthesis afford the or nicotinamide 38 mer prepared guanosine from (NGD),in vitro uridine transcription, (NUD) followedand cytidine by pyrophosphatase (NCD) dinucleotide-capped treatment to afford RNAs. the They 5′-monophosphate also showed that from the 5′ synthetic-triphosphate NAD-RNA, (Figure 19).similarly as the one obtained by IVT, is accepted by the deNADding enzyme, NudC–a Nudix hydrolase responsibleTo add forNMN removal moiety of to the these NAD RNAs,+ moiety they from synthesized RNA in bacteria.nicotinamide 5′-phosphorimidazolide by activatingIn 2018, NMN our using team CDI reported in DMF on thefollowed synthesis by a of carbonate novel NAD hydrolysis+ analogs step. and In their the last incorporation step, they usedinto RNA1000× [excess96]. The of the goal resulting of the study NMN-Im was in to coupling provide NAD-RNAreaction with analogs 5′-P-RNA that in are an lessaqueous susceptible media containingto so-far-identified Mg2+. This deNADding step led to the enzymes mixture than of NAD-RNA the native and structure. unreacted Several 5′-P-RNA. enzymes This capablelatter was of + howeverremoving easily the 5 0removed-NAD from by treatment RNA were with identified. Xrn-1, a nuclease Generally, which they uses can beonly divided 5′-P-RNA into as two substrates. different Theclasses: authors (i) deNADding used the same enzymes method with to afford pyrophosphatase the nicotinamide activity guanosine (e.g., NudC (NGD), and uridinehNudt12) (NUD) and and (ii) cytidinedeNADding (NCD) enzymes dinucleotide-capped with phosphodiesterase RNAs. activity They (e.g.,also Raishowed or DXO) that [10 ].the The synthetic first group NAD-RNA, cleaves the similarlypyrophosphate as the bond one inobtained NAD-RNA, by IVT, releasing is acce thepted nicotinamide by the deNADding 50-monophosphate enzyme, andNudC the – 5 0a-P-RNA, Nudix + hydrolasewhereas the responsible second group for removal removes of the the entire NAD NAD+ moietymoiety from fromRNA RNAin bacteria. 50 also revealing 50-P-RNA. InTaking 2018, into our account team reported the cleavage on the sites synthesis of these of enzymes, novel NAD we designed+ analogs aand set their of NAD incorporation+ analogs–either into RNAdi- or [96]. trinucleotides–containing The goal of the study was modifications to provide on NA theD-RNA pyrophosphate analogs that bridge, are less 20-position susceptible of adenosine to so-far- identifiedand/or within deNADding the phosphodiester enzymes than bondthe native between structure. NAD Several+ and theenzymes third capable nucleotide of removing (Figure the20). 5The′-NAD synthesis+ from wasRNA performed were identified. using a Generally, ZnCl2-mediated they ca couplingn be divided reaction into between two different NMN-Im classes: and an(i) deNADdingappropriate adenosine-containing enzymes with pyrophosphatase mono- or dinucleotide. activity (e.g., The NudC compounds and hNudt12) have later and been (ii) deNADding incorporated enzymesinto RNA with via IVT phosphodiesterase using polymerase activity T7 and (e.g., a 10-fold Rai excessor DXO) of NAD [10].+ analog,The first followed group thecleaves analysis the pyrophosphateby polyacrylamide bond gel in electrophoresisNAD-RNA, releasing to determine the nicotinamide RNA quality 5′-monophosphate and NAD+ analog and the incorporation 5′-P-RNA, whereasefficiency. the Remarkably,second group it removes was found the entire that trinucleotide NAD+ moiety NAD from+ RNAanalogs 5′ also76a-f revealingare very 5′-P-RNA. efficiently incorporated into RNA compared to NAD+ or dinucleotide NAD+ analogs 75a-c. The unmodified trinucleotide (NADpG, 76a) is therefore a superior reagent for generation of NAD-capped RNAs by

IVT compared to NAD+. The resulting NAD-RNAs were subjected to three different deNADding enzymes, which led to the identification of compounds showing resistance only to deNADding pyrophosphatases (NAD50S, 75c), only to phosphodiesterase (NRppAmpG, 76d), or both types of activities (NRppSAmpG D1 76e and NRppSAmpG D2 76f). The corresponding NAD-RNA analogs are expected to be more stable under cellular conditions and therefore have a potential to become useful molecular tools for the studies on its biological role of NAD-RNA. Molecules 2019, 24, x 27 of 34

Taking into account the cleavage sites of these enzymes, we designed a set of NAD+ analogs – either di- or trinucleotides – containing modifications on the pyrophosphate bridge, 2′-position of adenosine and/or within the phosphodiester bond between NAD+ and the third nucleotide (Figure 20). The synthesis was performed using a ZnCl2-mediated coupling reaction between NMN-Im and an appropriate adenosine-containing mono- or dinucleotide. The compounds have later been incorporated into RNA via IVT using polymerase T7 and a 10-fold excess of NAD+ analog, followed the analysis by polyacrylamide gel electrophoresis to determine RNA quality and NAD+ analog incorporation efficiency. Remarkably, it was found that trinucleotide NAD+ analogs 76a-f are very efficiently incorporated into RNA compared to NAD+ or dinucleotide NAD+ analogs 75a-c. The unmodified trinucleotide (NADpG, 76a) is therefore a superior reagent for generation of NAD- capped RNAs by IVT compared to NAD+. The resulting NAD-RNAs were subjected to three different deNADding enzymes, which led to the identification of compounds showing resistance only to deNADding pyrophosphatases (NAD5′S, 75c), only to phosphodiesterase (NRppAmpG, 76d), or both types of activities (NRppSAmpG D1 76e and NRppSAmpG D2 76f). The corresponding NAD-RNA analogs are expected to be more stable Moleculesunder cellular2019, 24, 4187conditions and therefore have a potential to become useful molecular tools for27 ofthe 34 studies on its biological role of NAD-RNA.

+ Figure 20. Hydrolysis-resistant NAD + analogsanalogs designed designed to to study study the the corresponding corresponding NAD-RNAs [[96].96]. 7. Conclusions and Perspectives for the Future 7. Conclusions and Perspectives for the Future Here, we reviewed the recent progress in the design of NAD analogs as molecular probes and Here, we reviewed the recent progress in the design of NAD analogs as molecular probes and biologically active compounds. The last decade has brought many significant advancements in the biologically active compounds. The last decade has brought many significant advancements in the design of fluorescent probes enabling real-time monitoring of NAD-dependent processes via fluorescent design of fluorescent probes enabling real-time monitoring of NAD-dependent processes via methods. Although most of these probes have been utilized only in vitro or in cell extracts, the first fluorescent methods. Although most of these probes have been utilized only in vitro or in cell extracts, impressive applications in living cells have also been achieved [45]. We envisage that the field will the first impressive applications in living cells have also been achieved [45]. We envisage that the progress further in this this direction, especially if universal solutions for efficient intracellular delivery field will progress further in this this direction, especially if universal solutions for efficient of NAD analogs are developed. A step in this direction was reported recently by Wang et al. [98], intracellular delivery of NAD analogs are developed. A step in this direction was reported recently who developed an Escherichia coli strain expressing an efficient NAD importer and showing limited by Wang et al. [98], who developed an Escherichia coli strain expressing an efficient NAD importer extracellular degradation to facilitate characterization of NAD analogs in vivo. If similar solutions are and showing limited extracellular degradation to facilitate characterization of NAD analogs in vivo. also devised for eukaryotic cells, they will greatly advance NAD-related research in complex biological If similar solutions are also devised for eukaryotic cells, they will greatly advance NAD-related settings. These solutions could also facilitate probe and inhibitor design by helping dissect issues research in complex biological settings. These solutions could also facilitate probe and inhibitor with compound permeability form other problems with in vivo activity. NAD-derived inhibitors design by helping dissect issues with compound permeability form other problems with in vivo have undoubtedly tremendous potential for the treatment of cancer, bacterial infections, age-related activity. NAD-derived inhibitors have undoubtedly tremendous potential for the treatment of cancer, disorders (the latter via targeting of sirtuins), and other diseases. However, in this context, the selectivity bacterial infections, age-related disorders (the latter via targeting of sirtuins), and other diseases. of the NAD-derived inhibitors is a very important, yet often concealed issue. Many structurally similar However, in this context, the selectivity of the NAD-derived inhibitors is a very important, yet often compounds have been designed with completely different molecular targets in mind and showed concealed issue. Many structurally similar compounds have been designed with completely different activity in quite unrelated biological systems. It is difficult to point out substitution sites that are molecular targets in mind and showed activity in quite unrelated biological systems. It is difficult to selectively preferred by particular enzyme classes as for most of them systematic structure-activity point out substitution sites that are selectively preferred by particular enzyme classes as for most of relationship studies have not yet been completed. One of the exceptions are ADPribosyltransferases them systematic structure-activity relationship studies have not yet been completed. One of the (e.g., PARP) for which it has been determined that the preferred modification sites are N6 and C2 exceptions are ADPribosyltransferases (e.g., PARP) for which it has been determined that the positions of adenine and 30-O position of nicotinamide riboside. In the case of other NAD-related enzymes, adenine modifications at the N6 and C8 positions have been commonly explored (among others) and often proved to be biologically active. Taking into account the overlapping artificial substrate and inhibitor scopes for many enzymes, it is likely that some of the observed discrepancies between the results from in vitro and in vivo evaluation arise from off-target effects. Standardization of inhibitor evaluation procedures, preceded by identification of enzyme families that have similar structural requirements for NAD recognition, could help addressing the issue of selectivity at early stages of inhibitor development. Finally, the discovery of NAD-capped RNA has opened completely new horizons for NAD analog development. Here, of great value may be compounds and methods enabling assaying the activity of enzymes responsible for NAD-RNA biosynthesis and degradation in vitro and in vivo, compounds enabling capture and isolation of specifically interacting proteins from biological mixtures, and inhibitors selectively targeting proteins that bind or process NAD-capped RNA.

Funding: This research and APC were funded by the National Science Centre, Poland, grant number 2015/18/E/ST5/00555. Molecules 2019, 24, 4187 28 of 34

Acknowledgments: We thank Mikolaj Chrominski, Marcin Warminski, and Jacek Jemielity (University of Warsaw) for critical reading of the manuscript. Conflicts of Interest: The authors declare no conflict of interest.

Abbreviations

AAC Azide alkyne cycloaddition Ach Acetylcholine ACS Acetyl-coA enzyme synthase ADH Alcohol dehydrogenase ADPRC ADP-ribosyl cyclase ALDH dehydogenase Aq aqueous ART ADP-ribosyl transferase ASST Tiazofurin adenosine disulfide BAD Benzamide adenine dinucleotide cADPR Cyclic ADP-ribose CDI 1,10-carbonyldiimidazole CML Chronic myleogenous leukemia CRF Chronic renal failure CTA Cholera toxin subunit A CuAAC Copper catalysed AAC dpCoA Dephospho DHFR dehydrofolate reductase DIC N,N-diidopropylcarbodiimide DMF Dimethylformamide DMSO Dimethylsulfoxide DTA di-50-thioadenosine dinucleotide FAD Flavine adenine dinucleotide FLIM-FRET Fluorescence lifetime imaging microscopy–Förster resonance energy transfer G6PDH Glucose-6-phosphate dehydrogenase GADPH Glyceraldehyde phosphate dehydrogenase GFP Green fluorescent protein IDH IEDDA Inverse electron demand Diels-Alder cycloaddition IMP(DH) Inosine-50-monophosphate (dehydrogenase) INH Isoniazid InhA Enoyl-ACP reductase IVT in vitro transcription LDH Lactate dehydrogenase MAD Mycophenolic acid adenine dinucleotide MART MonoADP-ribosyl transferase MeCN Acetonitrile MPA Mycophenolic acid Mtb Mycobacterium MTX Methotrexate NaAD Nicotinic acid adenine dinucleotide NAD Nicotinamide adenine dinucleotide NADase NAD glycohydrolase NADK NAD kinase NAD(P)S Thionicotinamide adenine dinucleotide (phosphate) Nam Nicotinamide NaMN Nicotinic acid mononucleotide NaMNAT NaMN adenylyl transferase Molecules 2019, 24, 4187 29 of 34

NCIN Non-canonical initiating nucleotides NCD Nicotinamide cytidine dinucleotide NdAD Nicotinamide-20-deoxy adenosine dinucleotide NGD Nicotinamide guanosine dinucleotide NMN Nicotinamide mononucleotide NMNAT NMN adenylyl transferase NMO 4-methylmorpholine N-oxide NRK Nicotinamide riboside kinase NUD Nicotinamide uridine dinucleotide PAR Poly(ADP-ribose) PARG Poly(ADP-ribose) glycohydrolase PARP/ARTD Poly(ADP-ribose) polymerase Pdh Phosphite dehydrogenase PPi Pyrophosphate ROS Reactive oxygen species Rt Room temperature SIRT Sirtuin SPAAC Strain promoted AAC TFA Trifluoroacetic acid TIPS-Cl 2,4,6-triisopropylbenzenesulfonyl chloride thA Thieno[4,3-d]pyrimidine riboside TMR Tetramethylrhodamine TN Thionicotinamide TPPTS 3,30,3”-Phosphanetriyltris(benzenesulfonic acid) trisodium salt TSST Ditiazofurin disulfide TXPTS Tris(2,4-dimethyl-5-sulfophenyl)phosphine trisodium salt tzA Isothiazolo[4,3-d]pyrimidine riboside

References

1. Agledal, L.; Niere, M.; Ziegler, M. The phosphate makes a difference: Cellular functions of NADP. Redox Rep. 2010, 15, 2–10. [CrossRef] 2. Di Stefano, M.; Conforti, L. Diversification of NAD biological role: The importance of location. FEBS J. 2013, 280, 4711–4728. [CrossRef] 3. Katsyuba, E.; Auwerx, J. Modulating NAD+ metabolism, from bench to bedside. EMBO J. 2017, 36, 2670–2683. [CrossRef] 4. Stromland, O.; Niere, M.; Nikiforov, A.A.; VanLinden, M.R.; Heiland, I.; Ziegler, M. Keeping the balance in NAD metabolism. Biochem. Soc. Trans. 2019, 47, 119–130. [CrossRef] 5. Davar, D.; Beumer, J.H.; Hamieh, L.; Tawbi, H. Role of PARP inhibitors in cancer biology and therapy. Curr. Med. Chem. 2012, 19, 3907–3921. [CrossRef] 6. Pant, S.; Maitra, A.; Yap, T.A. PARP inhibition—Opportunities in pancreatic cancer. Nat. Rev. Clin. Oncol. 2019, 16, 595–596. [CrossRef] 7. David Adams, J.; Klaidman, L.K. Sirtuins, Nicotinamide and Aging: A Critical Review. Lett. Drug Des. Discov. 2007, 4, 44–48. [CrossRef] 8. Abdelraheim, S.R.; Spiller, D.G.; McLennan, A.G. Mammalian NADH diphosphatases of the Nudix family: Cloning and characterization of the human peroxisomal NUDT12 protein. Biochem. J. 2003, 374, 329–335. [CrossRef] 9. Cahova, H.; Winz, M.L.; Hofer, K.; Nubel, G.; Jaschke, A. NAD captureSeq indicates NAD as a bacterial cap for a subset of regulatory RNAs. Nature 2015, 519, 374–377. [CrossRef] + 10. Kiledjian, M. Eukaryotic RNA 50-End NAD Capping and DeNADding. Trends Cell Biol. 2018, 28, 454–464. [CrossRef] 11. Jaschke, A.; Hofer, K.; Nubel, G.; Frindert, J. Cap-like structures in bacterial RNA and epitranscriptomic modification. Curr. Opin. Microbiol. 2016, 30, 44–49. [CrossRef] Molecules 2019, 24, 4187 30 of 34

12. Walters, R.W.; Matheny, T.; Mizoue, L.S.; Rao, B.S.; Muhlrad, D.; Parker, R. Identification of NAD+ capped mRNAs in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. USA 2017, 114, 480–485. [CrossRef] 13. Julius, C.; Yuzenkova, Y. Bacterial RNA polymerase caps RNA with various cofactors and cell wall precursors. Nucleic Acids Res. 2017, 45, 8282–8290. [CrossRef]

14. Jiao, X.; Doamekpor, S.K.; Bird, J.G.; , B.E.; Tong, L.; Hart, R.P.; Kiledjian, M. 50 End Nicotinamide Adenine Dinucleotide Cap in Human Cells Promotes RNA Decay through DXO-Mediated deNADding. Cell 2017, 168, 1015–1027. [CrossRef] 15. Wang, Y.; Li, S.; Zhao, Y.; You, C.; Le, B.; Gong, Z.; Mo, B.; Xia, Y.; Chen, X. NAD+-capped RNAs are widespread in the Arabidopsis transcriptome and can probably be translated. Proc. Natl. Acad. Sci. USA 2019, 116, 12094–12102. [CrossRef] 16. Frindert, J.; Zhang, Y.; Nubel, G.; Kahloon, M.; Kolmar, L.; Hotz-Wagenblatt, A.; Burhenne, J.; Haefeli, W.E.; Jaschke, A. Identification, Biosynthesis, and Decapping of NAD-Capped RNAs in B. subtilis. Cell Rep. 2018, 24, 1890–1901. [CrossRef] 17. Chen, Y.G.; Kowtoniuk, W.E.; Agarwal, I.; Shen, Y.; Liu, D.R. LC/MS analysis of cellular RNA reveals NAD-linked RNA. Nat. Chem. Biol. 2009, 5, 879–881. [CrossRef] 18. Bird, J.G.; Zhang, Y.; Tian, Y.; Panova, N.; Barvik, I.; Greene, L.; Liu, M.; Buckley, B.; Krasny, L.; Lee, J.K.; et al. + The mechanism of RNA 50 capping with NAD , NADH and desphospho-CoA. Nature 2016, 535, 444–447. [CrossRef] 19. Julius, C.; Riaz-Bradley, A.; Yuzenkova, Y. RNA capping by mitochondrial and multi-subunit RNA polymerases. Transcription 2018, 9, 292–297. [CrossRef] 20. Barvik, I.; Rejman, D.; Panova, N.; Sanderova, H.; Krasny, L. Non-canonical transcription initiation: The expanding universe of transcription initiating substrates. FEMS Microbiol. Rev. 2017, 41, 131–138. [CrossRef] 21. Bird, J.G.; Basu, U.; Kuster, D.; Ramachandran, A.; Grudzien-Nogalska, E.; Towheed, A.; Wallace, D.C.; + Kiledjian, M.; Temiakov, D.; Patel, S.S.; et al. Highly efficient 50 capping of mitochondrial RNA with NAD and NADH by yeast and human mitochondrial RNA polymerase. Elife 2018, 7, e42179. [CrossRef][PubMed] 22. Grudzien-Nogalska, E.; Wu, Y.; Jiao, X.; Cui, H.; Mateyak, M.K.; Hart, R.P.; Tong, L.; Kiledjian, M. Structural and mechanistic basis of mammalian Nudt12 RNA deNADding. Nat. Chem. Biol. 2019, 15, 575–582. [CrossRef][PubMed] 23. Hofer, K.; Li, S.; Abele, F.; Frindert, J.; Schlotthauer, J.; Grawenhoff, J.; Du, J.; Patel, D.J.; Jaschke, A. Structure and function of the bacterial decapping enzyme NudC. Nat. Chem. Biol. 2016, 12, 730–734. [CrossRef] [PubMed] 24. Imai, S. “Clocks” in the NAD World: NAD as a metabolic oscillator for the regulation of metabolism and aging. Biochim. Biophys. Acta 2010, 1804, 1584–1590. [CrossRef][PubMed] 25. Lin, H. Nicotinamide adenine dinucleotide: Beyond a redox coenzyme. Org. Biomol. Chem. 2007, 5, 2541–2554. [CrossRef][PubMed] 26. Lin, S.-J.; Guarente, L. Nicotinamide adenine dinucleotide, a metabolic regulator of transcription, longevity and disease. Curr. Opin. Cell Biol. 2003, 15, 241–246. [CrossRef] 27. Chini, C.C.S.; Tarrago, M.G.; Chini, E.N. NAD and the aging process: Role in life, death and everything in between. Mol. Cell Endocrinol. 2017, 455, 62–74. [CrossRef] 28. Lakowicz, J.R.; Szmacinski, H.; Nowaczyk, K.; Johnson, M.L. Fluorescence lifetime imaging of free and protein-bound NADH. Proc. Natl. Acad. Sci. USA 1992, 89, 1271–1275. [CrossRef] 29. Skala, M.C.; Riching, K.M.; Bird, D.K.; Gendron-Fitzpatrick, A.; Eickhoff, J.; Eliceiri, K.W.; Keely, P.J.; Ramanujam, N. In vivo multiphoton fluorescence lifetime imaging of protein-bound and free nicotinamide adenine dinucleotide in normal and precancerous epithelia. J. Biomed. Opt. 2007, 12, 024014. [CrossRef] 30. Blacker, T.S.; Mann, Z.F.; Gale, J.E.; Ziegler, M.; Bain, A.J.; Szabadkai, G.; Duchen, M.R. Separating NADH and NADPH fluorescence in live cells and tissues using FLIM. Nat. Commun. 2014, 5, 3936. [CrossRef] 31. Gordon Scott, T.; Spencer, R.D.; Leonard, N.J.; Weber, G. Emission Properties of NADH. Studies of Fluorescence Lifetimes and Quantum Efficiencies of NADH, AcPyADH, and Simplified Synthetic Models. J. Am. Chem. Soc. 1970, 92, 687–695. [CrossRef] 32. Pergolizzi, G.; Butt, J.N.; Bowater, R.P.; Wagner, G.K. A novel fluorescent probe for NAD-consuming enzymes. Chem. Commun. 2011, 47, 12655–12657. [CrossRef][PubMed] Molecules 2019, 24, 4187 31 of 34

33. Rovira, A.R.; Fin, A.; Tor, Y. Emissive Synthetic Cofactors: An Isomorphic, Isofunctional, and Responsive NAD+ Analogue. J. Am. Chem. Soc. 2017, 139, 15556–15559. [CrossRef][PubMed] 34. Shin, D.; Sinkeldam, R.W.; Tor, Y. Emissive RNA alphabet. J. Am. Chem. Soc. 2011, 133, 14912–14915. [CrossRef][PubMed] 35. Noe, M.S.; Sinkeldam, R.W.; Tor, Y. Oligodeoxynucleotides containing multiple thiophene-modified isomorphic fluorescent nucleosides. J. Org. Chem. 2013, 78, 8123–8128. [CrossRef] 36. Rovira, A.R.; Fin, A.; Tor, Y. Chemical Mutagenesis of an Emissive RNA Alphabet. J. Am. Chem. Soc. 2015, 137, 14602–14605. [CrossRef] 37. Mizrahi, R.A.; Shin, D.; Sinkeldam, R.W.; Phelps, K.J.; Fin, A.; Tantillo, D.J.; Tor, Y.; Beal, P.A. A Fluorescent Adenosine Analogue as a Substrate for an A-to-I RNA Editing Enzyme. Angew. Chem. Int. Ed. 2015, 54, 8713–8716. [CrossRef] 38. Halle, F.; Fin, A.; Rovira, A.R.; Tor, Y. Emissive Synthetic Cofactors: Enzymatic Interconversions of tzA Analogues of ATP, NAD+, NADH, NADP+, and NADPH. Angew. Chem. Int. Ed. 2018, 57, 1087–1090. [CrossRef] 39. Feldmann, J.; Li, Y.; Tor, Y. Emissive Synthetic Cofactors: A Highly Responsive NAD+ Analogue Reveals Biomolecular Recognition Features. Chem. Eur. J. 2019, 25, 4379–4389. [CrossRef] 40. Dieter, H.; Koberstein, R.; Sund, H. Nicotinamide 1, N6-ethenoadenine dinucleotide, a coenzyme for glutamate dehydrogenase. FEBS Lett. 1974, 47, 90–93. [CrossRef] 41. Roberts, J.E.; Aizono, Y.; Sonenberg, M.; Swilocki, N.I. Further chemical characterization and biological activities of fluorescent 1, N6-ethenoadenosine derivatives. Bioorg. Chem. 1975, 4, 181–187. [CrossRef] 42. Oei, S.L.; Griesenbeck, J.; Buchlow, G.; Jorcke, D.; Mayer-Kuckuk, P.; Wons, T.; Ziegler, M. NAD+ analogs substituted in the purine base as substrates for poly(ADP-ribosyl) transferase. FEBS Lett. 1996, 397, 17–21. [CrossRef] 43. Shrimp, J.H.; Hu, J.; Dong, M.; Wang, B.S.; MacDonald, R.; Jiang, H.; Hao, Q.; Yen, A.; Lin, H. Revealing CD38 cellular localization using a cell permeable, mechanism-based fluorescent small-molecule probe. J. Am. Chem. Soc. 2014, 136, 5656–5663. [CrossRef][PubMed] 44. Moreau, C.; Kirchberger, T.; Zhang, B.; Thomas, M.P.; Weber, K.; Guse, A.H.; Potter, B.V.L. Aberrant

Cyclization Affords a C-6 Modified Cyclic Adenosine 50-Diphosphoribose Analogue with Biological Activity in Jurkat T Cells. J. Med. Chem. 2012, 55, 1478–1489. [CrossRef][PubMed] 45. Buntz, A.; Wallrodt, S.; Gwosch, E.; Schmalz, M.; Beneke, S.; Ferrando-May, E.; Marx, A.; Zumbusch, A. Real-Time Cellular Imaging of Protein Poly(ADP-ribos) ylation. Angew. Chem. Int. Ed. 2016, 55, 11256–11260. [CrossRef] 46. Leung, A.K. Poly(ADP-ribose): An organizer of cellular architecture. J. Cell. Biol. 2014, 205, 613–619. [CrossRef] 47. Corda, D.; Di Girolamo, M. Functional aspects of protein mono-ADP-ribosylation. EMBO J. 2003, 22, 1953–1958. [CrossRef] 48. Zhang, J. Use of biotinylated NAD to label and purify ADP-ribosylated proteins. Methods Enzymol. 1997, 280, 255–265. 49. Garcia Soriano, F.; Virag, L.; Jagtap, P.; Szabo, E.; Mabley,J.G.; Liaudet, L.; Marton, A.; Hoyt, D.G.; Murthy,K.G.; Salzman, A.L.; et al. Diabetic endothelial dysfunction: The role of poly(ADP-ribose) polymerase activation. Nat. Med. 2001, 7, 108–113. [CrossRef] 50. Du, J.; Jiang, H.; Lin, H. Investigating the ADP-ribosyltransferase activity of sirtuins with NAD analogues and 32P-NAD. Biochemistry 2009, 48, 2878–2890. [CrossRef] 51. Jiang, H.; Kim, J.H.; Frizzell, K.M.; Kraus, W.L.; Lin, H. Clickable NAD analogues for labeling substrate proteins of poly(ADP-ribose) polymerases. J. Am. Chem. Soc. 2010, 132, 9363–9372. [CrossRef][PubMed] 52. Wang, Y.; Rosner, D.; Grzywa, M.; Marx, A. Chain-terminating and clickable NAD+ analogues for labeling the target proteins of ADP-ribosyltransferases. Angew. Chem. Int. Ed. 2014, 53, 8159–8162. [CrossRef] [PubMed] 53. Wallrodt, S.; Buntz, A.; Wang, Y.; Zumbusch, A.; Marx, A. Bioorthogonally Functionalized NAD+ Analogues for In-Cell Visualization of Poly(ADP-Ribose) Formation. Angew. Chem. Int. Ed. 2016, 55, 7660–7664. [CrossRef][PubMed] Molecules 2019, 24, 4187 32 of 34

54. Zhang, X.N.; Cheng, Q.; Chen, J.; Lam, A.T.; Lu, Y.; Dai, Z.; Pei, H.; Evdokimov, N.M.; Louie, S.G.; Zhang, Y. A ribose-functionalized NAD+ with unexpected high activity and selectivity for protein poly-ADP-ribosylation. Nat. Commun. 2019, 10, 4196. [CrossRef][PubMed] 55. Xia, C.; Ribeiro, M.; Scott, S.; Lonial, S. Daratumumab: Monoclonal antibody therapy to treat multiple myeloma. Drugs Today 2016, 52, 551–560. [CrossRef][PubMed] 56. Dong, M.; Si, Y.Q.; Sun, S.Y.; Pu, X.P.; Yang, Z.J.; Zhang, L.R.; Zhang, L.H.; Leung, F.P.; Lam, C.M.; Kwong, A.K.; et al. Design, synthesis and biological characterization of novel inhibitors of CD38. Org. Biomol. Chem. 2011, 9, 3246–3257. [CrossRef][PubMed] 57. Kwong, A.K.; Chen, Z.; Zhang, H.; Leung, F.P.; Lam, C.M.; Ting, K.Y.; Zhang, L.; Hao, Q.; Zhang, L.H.; Lee, H.C. -based inhibitors of the calcium signaling function of CD38. Biochemistry 2012, 51, 555–564. [CrossRef] 58. Wang, S.; Zhu, W.; Wang, X.; Li, J.; Zhang, K.; Zhang, L.; Zhao, Y.J.; Lee, H.C.; Zhang, L. Design, synthesis and SAR studies of NAD analogues as potent inhibitors towards CD38 NADase. Molecules 2014, 19, 15754–15767. [CrossRef] 59. Smith, B.C.; Hallows, W.C.; Denu, J.M. Mechanisms and molecular probes of sirtuins. Chem. Biol. 2008, 15, 1002–1013. [CrossRef] 60. Pesnot, T.; Kempter, J.; Schemies, J.; Pergolizzi, G.; Uciechowska, U.; Rumpf, T.; Sippl, W.; Jung, M.; Wagner, G.K. Two-step synthesis of novel, bioactive derivatives of the ubiquitous cofactor nicotinamide adenine dinucleotide (NAD). J. Med. Chem. 2011, 54, 3492–3499. [CrossRef] 61. Szczepankiewicz, B.G.; Dai, H.; Koppetsch, K.J.; Qian, D.; Jiang, F.; Mao, C.; Perni, R.B. Synthesis of carba-NAD and the structures of its ternary complexes with SIRT3 and SIRT5. J. Org. Chem. 2012, 77, 7319–7329. [CrossRef][PubMed] 62. Slama, J.T.; Simmons, A.M. Carbanicotinamide adenine dinucleotide: Synthesis and enzymological properties of a carbocyclic analogue of oxidized nicotinamide adenine dinucleotide. Biochemistry 1988, 27, 183–193. [CrossRef][PubMed] 63. Dai, Z.; Zhang, X.N.; Nasertorabi, F.; Cheng, Q.; Pei, H.; Louie, S.G.; Stevens, R.C.; Zhang, Y. Facile chemoenzymatic synthesis of a novel stable mimic of NAD. Chem. Sci. 2018, 9, 8337–8342. [CrossRef] [PubMed] 64. Jeong, L.S.; Lee, H.W.; Jacobson, K.A.; Kim, H.O.; Shin, D.H.; Lee, J.A.; Gao, Z.G.; Lu, C.; Duong, H.T.;

Gunaga, P.; et al. Structure-activity relationships of 2-chloro-N6-substituted-40-thioadenosine-50-uronamides as highly potent and selective agonists at the human A3 adenosine receptor. J. Med. Chem. 2006, 49, 273–281. [CrossRef] 65. Langelier, M.F.; Zandarashvili, L.; Aguiar, P.M.; Black, B.E.; Pascal, J.M. NAD+ analog reveals PARP-1 substrate-blocking mechanism and allosteric communication from catalytic center to DNA-binding domains. Nat. Commun. 2018, 9, 844. [CrossRef] 66. Felczak, K.; Pankiewicz, K.W. Rehab of NAD(P)-dependent enzymes with NAD(P)-based inhibitors. Curr. Med. Chem. 2011, 18, 1891–1908. [CrossRef] 67. Pankiewicz, K.W.; Felczak, K. From ribavirin to NAD analogues and back to ribavirin in search for anticancer agents. Heterocycl. Commun. 2015, 21, 249–257. [CrossRef] 68. Chen, L.; Wilson, D.J.; Xu, Y.; Aldrich, C.C.; Felczak, K.; Sham, Y.Y.; Pankiewicz, K.W. Triazole-linked inhibitors of inosine monophosphate dehydrogenase from human and Mycobacterium tuberculosis. J. Med. Chem. 2010, 53, 4768–4778. [CrossRef] 69. Felczak, K.; Chen, L.; Wilson, D.; Williams, J.; Vince, R.; Petrelli, R.; Jayaram, H.N.; Kusumanchi, P.; Kumar, M.; Pankiewicz, K.W. Cofactor-type inhibitors of inosine monophosphate dehydrogenase via modular approach: Targeting the pyrophosphate binding sub-domain. Bioorg. Med. Chem. 2011, 19, 1594–1605. [CrossRef] 70. Felczak, K.; Pankiewicz, K.W. Synthesis of methylenebis(phosphonate) analogues of 2-, 4-, and 6-pyridones of nicotinamide adenine dinucleotide. Nucleosides Nucleotides Nucleic Acids 2011, 30, 512–523. [CrossRef] 71. Dutta, S.P.; Crain, P.F.; McCloskey, J.A.; Chheda, G.B. Isolation and characterization of 1-beta-D-ribofuranosylpyridin-4-one-3-carboxamide from human urine. Life Sci. 1979, 24, 1381–1388. [CrossRef] 72. Felczak, K.; Vince, R.; Pankiewicz, K.W. NAD-based inhibitors with anticancer potential. Bioorg. Med. Chem. Lett. 2014, 24, 332–336. [CrossRef][PubMed] Molecules 2019, 24, 4187 33 of 34

73. Poncet-Montange, G.; Assairi, L.; Arold, S.; Pochet, S.; Labesse, G. NAD kinases use substrate-assisted catalysis for specific recognition of NAD. J. Biol. Chem. 2007, 282, 33925–33934. [CrossRef][PubMed] 74. Tedeschi, P.M.; Bansal, N.; Kerrigan, J.E.; Abali, E.E.; Scotto, K.W.; Bertino, J.R. NAD+ Kinase as a Therapeutic Target in Cancer. Clin. Cancer Res. 2016, 22, 5189–5195. [CrossRef] 75. Petrelli, R.; Felczak, K.; Cappellacci, L. NMN/NaMN adenylyltransferase (NMNAT) and NAD kinase (NADK) inhibitors: Chemistry and potential therapeutic applications. Curr. Med. Chem. 2011, 18, 1973–1992. [CrossRef] 76. Petrelli, R.; Sham, Y.Y.; Chen, L.; Felczak, K.; Bennett, E.; Wilson, D.; Aldrich, C.; Yu, J.S.; Cappellacci, L.; Franchetti, P.; et al. Selective inhibition of nicotinamide adenine dinucleotide kinases by dinucleoside disulfide mimics of nicotinamide adenine dinucleotide analogues. Bioorg. Med. Chem. 2009, 17, 5656–5664. [CrossRef] 77. Hsieh, Y.C.; Tedeschi, P.; Adebisi Lawal, R.; Banerjee, D.; Scotto, K.; Kerrigan, J.E.; Lee, K.C.; Johnson-Farley,N.; Bertino, J.R.; Abali, E.E. Enhanced degradation of dihydrofolate reductase through inhibition of NAD kinase by nicotinamide analogs. Mol. Pharm. 2013, 83, 339–353. [CrossRef] 78. Tedeschi, P.M.; Lin, H.; Gounder, M.; Kerrigan, J.E.; Abali, E.E.; Scotto, K.; Bertino, J.R. Suppression of Cytosolic NADPH Pool by Thionicotinamide Increases Oxidative Stress and Synergizes with Chemotherapy. Mol. Pharm. 2015, 88, 720–727. [CrossRef] 79. Delaine, T.; Bernardes-Genisson, V.; Quemard, A.; Constant, P.; Meunier, B.; Bernadou, J. Development of isoniazid-NAD truncated adducts embedding a lipophilic fragment as potential bi-substrate InhA inhibitors and antimycobacterial agents. Eur. J. Med. Chem. 2010, 45, 4554–4561. [CrossRef] 80. Delaine, T.; Bernardes-Genisson, V.; Quemard, A.; Constant, P.; Cosledan, F.; Meunier, B.; Bernadou, J. Preliminary investigations of the effect of lipophilic analogues of the active metabolite of isoniazid toward bacterial and plasmodial strains. Chem. Biol. Drug Des. 2012, 79, 1001–1006. [CrossRef] 81. Liu, W.; Wu, S.; Hou, S.; Zhao, Z. Synthesis of phosphodiester-type nicotinamide adenine dinucleotide analogs. Tetrahedron 2009, 65, 8378–8383. [CrossRef] 82. Jain, P.; Slama, J.T.; Perez-Haddock, L.A.; Walseth, T.F. Nicotinic acid adenine dinucleotide phosphate analogues containing substituted nicotinic acid: Effect of modification on Ca2+ release. J. Med. Chem. 2010, 53, 7599–7612. [CrossRef][PubMed] 83. Trabbic, C.J.; Zhang, F.; Walseth, T.F.; Slama, J.T. Nicotinic Acid Adenine Dinucleotide Phosphate Analogues Substituted on the Nicotinic Acid and Adenine Ribosides. Effects on ReceptorMediated Ca2+ Release. J. Med. Chem. 2015, 58, 3593–3610. [CrossRef][PubMed] 84. Paul, C.E.; Arends, I.W.C.E.; Hollmann, F. Is Simpler Better? Synthetic Nicotinamide Cofactor Analogues for Redox Chemistry. ACS Catal. 2014, 4, 788–797. [CrossRef] 85. Hou, S.; Liu, W.; Ji, D.; Wang, Q.; Zhao, Z.K. Synthesis of 1,2,3-triazole moiety-containing NAD analogs and their potential as redox cofactors. Tetrahedron Lett. 2011, 52, 5855–5857. [CrossRef] 86. Hou, S.; Ji, D.; Liu, W.; Wang, L.; Zhao, Z.K. Identification of malic enzyme mutants depending on 1,2,3-triazole moiety-containing nicotinamide adenine dinucleotide analogs. Bioorg. Med. Chem. Lett. 2014, 24, 1307–1309. [CrossRef] 87. Ji, D.; Wang, L.; Hou, S.; Liu, W.; Wang, J.; Wang, Q.; Zhao, Z.K. Creation of bioorthogonal redox systems depending on nicotinamide flucytosine dinucleotide. J. Am. Chem. Soc. 2011, 133, 20857–20862. [CrossRef] 88. Ji, D.; Wang, L.; Liu, W.; Hou, S.; Zhao, K.Z. Synthesis of NAD analogs to develop bioorthogonal redox system. Sci. China Chem. 2013, 56, 296–300. [CrossRef] 89. Liu, Y.; Feng, Y.; Wang, L.; Guo, X.; Liu, W.; Li, Q.; Wang, X.; Xue, S.; Zhao, Z.K. Structural Insights into Phosphite Dehydrogenase Variants Favoring a Non-natural Redox Cofactor. ACS Catal. 2019, 9, 1883–1887. [CrossRef] 90. Lee, S.H.; Lee, H.J.; Won, K.; Park, C.B. Artificial electron carriers for photoenzymatic synthesis under visible light. Chem. Eur. J. 2012, 18, 5490–5495. [CrossRef] 91. Zhang, H.; Zhong, H.; Zhang, S.; Shao, X.; Ni, M.; Cai, Z.; Chen, X.; Xia, Y. NAD tagSeq reveals that NAD+-capped RNAs are mostly produced from a large number of protein-coding genes in Arabidopsis. Proc. Natl. Acad. Sci. USA 2019, 116, 12072–12077. [CrossRef][PubMed] 92. Wang, J.; Alvin Chew, B.L.; Lai, Y.; Dong, H.; Xu, L.; Balamkundu, S.; Cai, W.M.; Cui, L.; Liu, C.F.; Fu, X.Y.; et al. Quantifying the RNA cap reveals novel caps in cellular and viral RNA. Nucleic Acids Res. 2019.[CrossRef][PubMed] Molecules 2019, 24, 4187 34 of 34

93. Hofer, K.; Abele, F.; Schlotthauer, J.; Jaschke, A. Synthesis of 50-NAD-Capped RNA. Bioconjug. Chem. 2016, 27, 874–877. [CrossRef][PubMed] 94. Vvedenskaya, I.O.; Bird, J.G.; Zhang, Y.; Zhang, Y.; Jiao, X.; Barvik, I.; Krasny, L.; Kiledjian, M.; Taylor, D.M.;

Ebright, R.H.; et al. CapZyme-Seq Comprehensively Defines Promoter-Sequence Determinants for RNA 50 Capping with NAD+. Mol. Cell 2018, 70, 553–564. [CrossRef][PubMed] 95. Grudzien-Nogalska, E.; Bird, J.G.; Nickels, B.E.; Kiledjian, M. “NAD-capQ” detection and quantitation of NAD caps. RNA 2018, 24, 1418–1425. [CrossRef][PubMed] 96. Mlynarska-Cieslak, A.; Depaix, A.; Grudzien-Nogalska, E.; Sikorski, P.J.; Warminski, M.; Kiledjian, M.; Jemielity, J.; Kowalska, J. Nicotinamide-Containing Di- and Trinucleotides as Chemical Tools for Studies of NAD-Capped RNAs. Org. Lett. 2018, 20, 7650–7655. [CrossRef][PubMed] 97. Huang, F. Efficient incorporation of CoA, NAD and FAD into RNA by in vitro transcription. Nucleic Acids Res. 2003, 31, e8. [CrossRef][PubMed] 98. Wang, L.; Liu, B.; Liu, Y.; Sun, Y.; Liu, W.; Yu, D.; Zhao, Z.K. Escherichia coli Strain Designed for Characterizing In Vivo Functions of Nicotinamide Adenine Dinucleotide Analogues. Org. Lett. 2019, 21, 3218–3222. [CrossRef]

© 2019 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/).

Top View