Abstract
Nicotinamide phosphoribosyltransferase (NAMPT) plays an important role in the biosynthesis of nicotinamide adenine dinucleotide (NAD+) via the nicotinamide (NAM) salvage pathway. While the structural biochemistry of eukaryote NAMPT has been well studied, the catalysis mechanism of prokaryote NAMPT at the molecular level remains largely unclear. Here, we demonstrated the NAMPT-mediated salvage pathway is functional in the Gram-negative phytopathogenic bacterium Xanthomonas campestris pv. campestris (Xcc) for the synthesis of NAD+, and the enzyme activity of NAMPT in this bacterium is significantly higher than that of human NAMPT in vitro. Our structural analyses of Xcc NAMPT, both in isolation and in complex with either the substrate NAM or the product nicotinamide mononucleotide (NMN), uncovered significant details of substrate recognition. Specifically, we revealed the presence of a NAM binding tunnel that connects the active site, and this tunnel is essential for both catalysis and inhibitor binding. We further demonstrated that NAM binding in the tunnel has a positive cooperative effect with NAM binding in the catalytic site. Additionally, we discovered that phosphorylation of the His residue at position 229 enhances the substrate binding affinity of Xcc NAMPT and is important for its catalytic activity. This work reveals the importance of NAMPT in bacterial NAD+ synthesis and provides insights into the substrate recognition and the catalytic mechanism of bacterial type II phosphoribosyltransferases.
Similar content being viewed by others
Introduction
NAD+ is a vital coenzyme that is involved in a wide range of biological processes, including cellular redox reactions1, cell signaling2, DNA repair3, and epigenetic regulation4,5. There is evidence to suggest that supplementing with NAD+ precursors can have therapeutic benefits for various neurological disorders associated with NAD+ decline6. Moreover, the impact of NAD+ on aging and longevity has received substantial attention in recent research7. Research has shown that the levels of NAD+ decline with age in animals, and that NAD+ supplementation can prolong animal life span8,9.
NAD+ metabolism is a fundamental process that occurs across all organisms. There are three primary routes in which NAD+ is synthesized: the de novo biosynthetic pathway (with tryptophan or aspartate as initial precursor), the Preiss-handler deamidated salvage pathway (with nicotinic acid [NA] as initial precursor), and the NAM salvage pathway (with nicotinamide [NAM] as initial precursor)9,10,11. Among these pathways, vertebrates predominantly rely on the direct two-step NAM salvage pathway for NAD+ production, through the rate-limiting enzyme, NAM phosphoribosyltransferase (NAMPT)12. Despite its importance in vertebrates, this pathway does not appear to be universally preferred for NAD+ synthesis across organisms, as some fungi, plants, and bacteria lack NAMPT and instead rely on the Preiss-Handler deamidated salvage pathway to produce NAD+ by converting NAM to NA10,13,14. Therefore, the pathways for NAD+ synthesis differ among organisms due to the absence or presence of NAMPT.
Previous studies have established NAMPT as an essential regulator of the NAD+ pool, which is closely associated with longevity in vertebrates7. Importantly, supplementation of NAMPT can enhance NAD+ biosynthesis and extend the health-span of aged mice7. Meanwhile, NAMPT activators can potentially treat diseases associated with NAD+ deficiency15. In mammals, NAMPT-mediated NAD+ biosynthesis is recognized as a major source of NAD+ 12, and cancer cells rely more severely on NAMPT than healthy cells due to their high NAD+ requirement16, making NAMPT an attractive target for anticancer therapy. An NAMPT inhibitor called FK866 has been found to kill carcinoma cells by inhibiting NAD+ synthesis17. However, the sensitivity of cancer cells to NAMPT inhibitors such as FK866 is affected by several factors14,18, such as some mutations in human NAMPT confer resistance to this NAMPT inhibitor19. Thus, NAMPT proteins from different organisms may have varying sensitivity to human NAMPT inhibitors. The bacterial NadV gene is responsible for encoding NAMPT, and it is also been found in Xanthomonas campestris pv. campestris (Xcc)10, a Gram-negative bacterium that causes black root disease in crops. As previously described, Xcc has the ability to produce NAD+ through two pathways, the de novo pathway (from Trp)20 and the salvage pathway21. In addition, Xcc contains a NMN deamidase (PncC)21, which is a common gene in bacteria that can convert NMN to NAMN via the Preiss-Handler pathway, thus linking the two salvage pathways22. However, this gene is absent in Eukarya23. While the enzymatic activities of several bacterial NAMPTs have been reported24,25,26, a comprehensive understanding of their underlying structural mechanisms remains entirely elusive. Studying the structural differences of NAMPTs from various species and how these particular bacteria metabolize NAD+ could provide valuable insights into NAMPT-mediated NAD+ metabolism.
In addition to NAMPT, NA phosphoribosyltransferase (NAPRT) and quinolinic acid phosphoribosyltransferase (QAPRT) are also involved in NAD+ synthesis. All of these enzymes belong to the type II phosphoribosyltransferases (PRTs)27,28. Although bacterial structures of NAPRT and QAPRT have been reported28,29, the lack of available data on bacterial NAMPT structures has impeded mechanistic analyses and comparisons of substrate recognition and catalysis of bacterial type II PRTs. Consequently, it is crucial to determine if the structures and catalytic mechanisms of NAMPTs from lower organisms with different NAD+ metabolic pathways, such as bacteria, differ significantly from those found in vertebrates.
The current research demonstrated that the NAMPT-dependent salvage pathway plays a functional role in NAD+ production in Xcc. Through the use of purified recombinant NAMPT proteins, we confirmed that Xcc NAMPT is capable of effectively catalyzing the production of NMN in vitro. Notably, Xcc NAMPT was found to have significantly greater activity than human NAMPT. Furthermore, the three-dimensional structure of bacterial NAMPT from the plant pathogen Xcc was determined in three states, namely the apo form, the NAM-bound form, and the NMN-bound form. Structural analyses facilitated the identification of NAM binding within a tunnel that was distinct from the catalytic site, and further structural and biochemical studies revealed that this tunnel played important roles in both catalysis and inhibitor function. Meanwhile, we demonstrated that phosphorylation of the His residue at 229 plays a crucial role in enhancing the substrate affinity of Xcc NAMPT and is important for maintaining its enzymatic activity. These exciting structural results, combined with biochemical analyses, have demonstrated the presence of a NAM binding site in the tunnel and established the functional roles of the tunnel in both regulating NAMPT activity and binding inhibitors.
Results
NAMPT-mediated pathway is functional for NAD+ synthesis in Xcc
To investigate the molecular mechanisms of NAD+ biosynthesis in Xcc, we performed a genomic survey for enzymes involved in both de novo and salvage pathways of NAD+ synthesis in Xcc strain 8004. The result displayed that Xcc has all the necessary enzymes for NAD+ biosynthesis, with the exception of aspartate oxidase, quinolinate synthetase, and nicotinamide riboside kinase. This suggests that Xcc most likely produces NAD+ via the de novo pathway from tryptophan, but not from aspartatic acid, and also utilizes NAM and Preiss-Handler salvage pathways to synthesize NAD+ from NA, NAM and nicotinamide riboside (NmR) (Fig. 1a), which is consistent with previous findings20,21.
To test whether the de novo, Preiss-Handler and the NAM salvage pathways are functional in Xcc, we investigated the potential involvement of two homologous enzymes, the NAD+ synthetase (NADS) homologue XC_1067 and the NAMPT homologue XC_0719, in NAD synthesis and bacterial growth. Specifically, XC_1067 is responsible for the last step of NAD+ production in both the de novo and Preiss-Handler pathways, while XC_0719 catalyzes the rate-limiting step of the NAM salvage pathway. To investigate the involvement of these enzymes, we constructed two mutant strains, ΔNads and ΔNampt, by deleting the open reading frames (ORFs) of XC_1067 and XC_0719, respectively, from the Xcc wild-type strain 8004 (Table S1). We then assessed their ability to produce NAD+ and support bacterial growth in the absence of exogenous NAD+. Surprisingly, both mutant strains were able to grow in the minimal medium MMX, similar to the wild-type strain (Fig. 1b). While it was expected that the ΔNampt mutant could grow in the medium due to the possible existence of the de novo and Preiss-Handler pathways, it was unexpected that the ΔNads mutant was also able to grow, as NAM, which is required for the NAMPT-mediated salvage pathway, is not available in the medium. Furthermore, our bioinformatics analysis did not identify any proteins in Xcc homologous to the enzymes responsible for generating NMN from nicotinic acid mononucleotide (NaMN), an intermediate in the de novo and Preiss-Handler pathways (Fig. 1a). These results suggest that there may be unknown enzymes in Xcc, which can convert intermediates from the de novo and Preiss-Handler pathways into NAM or NmR, and thereby synthesize NAD+. Alternatively, there may be additional pathways that are NADS and NAMPT independent for NAD+ synthesize in Xcc. To investigate the presence of additional pathways for NAD+ synthesis and utilization in Xcc, we constructed a double deletion mutant deficient in the two predicted NAD+ synthesis pathways by deleting both ORFs of XC_0719 and XC_1067 (Table S1). This double deletion mutant, named ΔNamptΔNads, was then tested for growth in nutrient-deficient MMX medium. The result revealed that the mutant was only able to grow in MMX supplemented with NAD+, but not in MMX alone (Fig. 1b). This suggests that the synthesis of NAD+ in Xcc is dependent on NADS and NAMPT. Given that there is no evidence suggesting that NAD+ can be absorbed directly by Xcc, we proposed that it might be degraded to NmR, NA, or NAM before being utilized by Xcc cells. However, the double-deficient strain still did not grow in MMX supplemented with any of these additives (Fig. 1b), suggesting that NAD+ might be absorbed directly by Xcc.
Xcc NAMPT has a significantly higher in-vitro enzyme activity than human NAMPT
Our genetic investigation has confirmed the functional involvement of Xcc NAMPT in the NAM salvage pathway for NAD+ synthesis. To further explore its enzymatic activity, we overexpressed and purified Xcc NAMPT, and measured its activity using a fluorometric assay that detects NMN levels30. We then determined and compared the kinetic parameters of Xcc and human NAMPT proteins. Our investigation revealed a significantly higher enzymatic activity of Xcc NAMPT compared to human NAMPT, observed not only at physiological temperatures (28 °C) but also at non-physiological temperatures (37 °C) of Xcc. (Fig. 2a). Consistently, Xcc NAMPT has a significantly higher catalytic efficiency than human NAMPT (Fig. 2a), despite the high sequence similarity between human and Xcc NAMPTs (Fig S1). Specifically, the kcat/Km value of Xcc NAMPT for NAM is 32.28 μM−1s−1, which is approximately 17.93 times higher than that of human NAMPT (1.8 μM−1s−1)31 (Fig. 2b, c). In addition, the kcat/Km value of Xcc NAMPT for PRPP is 0.29 μM−1s−1, which is ~36.25 times higher than that of human NAMPT (0.008 μM−1s−1) (Figs. 2b, d and S2).
Given that ATP hydrolysis is a critical factor for substrate binding and catalysis in human NAMPT31, we sought to investigate the contribution of ATP to NMN synthesis in Xcc NAMPT. To accomplish this, we utilized a NADH-coupled ATP hydrolysis detection system and confirmed that Xcc NAMPT is also capable of hydrolyzing ATP (Fig. S3). Furthermore, our observations revealed that ATP is critical for NAMPT-mediated NMN synthesis (Fig. 2e). Notably, our results suggest that ATP likely functions as a substrate during the catalytic reaction of NMN synthesis, as the Michaelis-Menten saturation curve provides an accurate description of the contribution of ATP in NAMPT-mediated NMN synthesis, with a kcat/Km value of 7.86 mM−1s−1 (Fig. 2b).
Xcc NAMPT employs a homodimeric architecture to establish its active site
To investigate the catalytic mechanism of Xcc NAMPT, we determined the crystal structures of the isolated Xcc NAMPT protein and its complexes with either the reactant NAM or the product NMN.
We obtained two apo structures from two types of crystals, namely the tablet-shaped (1.79 Å) and diamond-shaped (2.08 Å) crystals. These structures had almost identical overall conformations (RMSD, 0.3 Å). However, since the structure determined from the diamond-shaped crystal was incomplete in domain A (Fig. S4a), we focused on describing the more complete apo structure obtained from the tablet-shaped crystal. The NAMPT monomer consists of 19 α-helices and 18 β-strands (Figs. S1 and 3a), which can be divided into two domains, domain A and domain B (Fig. 3a). Domain A is a typical β/α barrel that has a letter-C-like shape, which is formed by a five-stranded β sheet (β6, β7, β8, β11, and β12) wrapped around by 11 α helices (α7-α17). Domain B is consisted of a central β sandwich that is covered on one face by 5 helices (α1, α2, α3, α4, and α5). Notably, α6 serves as a scaffolding helix to connect domain A with domain B.
There are two molecules in an asymmetric unit of the Xcc NAMPT crystal, which assemble into a compact and stable dimer that covers a surface area of 4140 Å2 (Fig. 3b). To investigate the oligomeric state of Xcc NAMPT in solution, purified proteins were analyzed using gel filtration and Sedimentation Velocity Analytical Ultracentrifugation (SV-AUC) (Figs. S5a and 3c). The elution volume of purified NAMPT on gel filtration corresponds to an apparent molecular weight of approximately 107.98 kDa, consistent with that expected for a NAMPT dimer (106.324 kDa) (Fig. S5a). Similarly, SV-AUC analysis revealed that NAMPT proteins behave mainly as dimers in solution (Fig. 3c), regardless of whether the protein is in isolation or with reactant (NAM or PRPP) or with the product (NMN). These findings suggest that our structural analysis has identified a strong candidate dimer interface, which resembles that of human NAMPT (Fig. S6).
The dimeric structure of Xcc NAMPT consists of two protomers arranged head to tail, with domain A on one protomer interacting with domain B on the other protomer (Fig. 3b). In particular, the open side of the C-shape β/α barrel in domain A packs against a four-helix bundle in domain B to create a complete active site. To confirm that the dimeric formation is essential for Xcc NAMPT to function, we propose that Xcc NAMPT during catalysis has a shared active site as that observed structurally. We thus mixed the R373A mutant from domain B with another mutant from domain A (H229K, H229R, R180A or D335S) and tested if heterodimer could be formed to restore enzyme activity. As anticipated, although these mutants had extremely low activity when present alone, they were able to restore a significant amount of enzyme activity after they were pre-incubated for 24 h (Figs. 3d and S5b). Consistent with our notion that the dimer is extremely stable, enzyme activity could not be restored if corresponding pairs of mutants was used without pre-incubation (Figs. 3d and S5c). Collectively, our results strongly suggest that the dimeric status of NAMPT is crucial for its catalytic activity, since both protomers contribute to the formation of the active site.
It should be noted that the active site is not entirely closed but instead exposed to the solvent through two openings: an entrance situated directly above the active site and a separate lateral tunnel (Figs. S7 and 4a).
Xcc NAMPT features two distinct sites for NAM binding
To understand how Xcc NAMPT recognizes substrates, we planned to determine the crystal structures of the NAMPT-NAM and NAMPT-PRPP complexes. Despite numerous attempts, we were only successful in soaking NAM into the diamond-shaped crystal and the structure was determined at a resolution of 2.0 Å. In one of the two active sites of the NAMPT dimer, we observed two distinct NAM molecules with clear electron densities, while the other active site remained empty (Fig. S7). Therefore, we focused our discussion mainly on the active site containing the two NAM molecules. As shown in Figs. S7 and 4a, one NAM molecule binds in the catalytic site of the β/α-barrel center, while the other binds in the lateral tunnel that connects the catalytic site to the bulk solvent. As a result, we referred to these two NAM molecules as the catalytic-site NAM and tunnel NAM, respectively.
The presence of NAM molecules in both the catalytic site and the tunnel of NAMPT is confirmed by well-defined electron densities (Fig. 4a). In the catalytic site, NAM is sandwiched between the aromatic side chains of Phe 177 and Tyr 14 through π-π interactions, and the carboxamide group forms a hydrogen bond with the carboxyl group of Asp 203, which properly aligns the pyridine ring of NAM for subsequent reactions (Fig. 4b). The location of the aromatic rings of Phe 177 and Tyr 14 has been slightly adjusted from the apo structure to accommodate the NAM molecule (Fig. S8). In the tunnel, NAM is located in a hydrophobic area composed mainly of non-polar residues, such as Ile 224, Ala 259, Val 291, Ile 332 and Ala 360, resulting in primarily hydrophobic interactions between NAMPT and NAM through van der Waals contacts (Fig. 4c). To strengthen NAM binding, the imidazole ring of His 175 forms a herringbone (edge-to-face) stacking interaction with the pyridine ring of NAM (Fig. 4c).
Our structural analysis of Xcc NAMPT was supported by our ITC experiments, which showed that there are two separate NAM binding sites in the enzyme (Fig. 4d). The raw data could not be adequately fitted using a single-site binding model; however, it was successfully fitted using a sequential binding model, with stoichiometry values of 2. The fitting revealed Kd values of 1.41 mM and 21 mM for the two NAM binding sites, respectively, with Kd1 < Kd2. This finding aligns with the positive cooperativity demonstrated by these sites32,33. Moreover, we observed a positive cooperative relationship between NAM binding in the tunnel and NAM binding in the catalytic site during the catalytic reaction, as indicated by a hill coefficient of ~1.98 for NAM in Xcc NAMPT (Fig. 4e). To confirm that positive cooperativity is present within a single active pocket rather than between the two active pockets of the enzyme dimer, we evaluated the kinetic parameters of inactive mutant pair (R373A and R180A) that can only possess one functional active pocket. Intriguingly, we found that the mutant pair exhibited a Km value towards NAM similar to that of the wild-type (WT) enzyme. Notably, the hill coefficient for this mutant pair remained greater than 1 (H = 1.64), suggesting the continued presence of positive cooperativity (Fig. S9).
This suggests that the binding of NAM molecules in these two sites is positively cooperative, highlighting the importance of the tunnel in NAM binding. In fact, our docking studies showed that the inhibitory effect of FK866 on NAMPT could be in part explained by the possibility that FK866 blocks NAM binding in the tunnel (Fig. S10a). Taken together, our findings suggest that Xcc NAMPT features two distinct NAM binding sites, the catalytic site and the tunnel, and that the binding of NAM molecules in these two sites is positively cooperative.
The NAM binding tunnel of Xcc NAMPT is important for both catalysis and inhibitor binding
To gain insights into the functional significance of the tunnel in NAMPT catalysis, we introduced mutations to amino acids within the internal cavity of the tunnel and investigated their impact on enzyme activity. To completely block the tunnel, which had a large diameter, we made multiple substitutions using bulky aromatic residues simultaneously. This led to the creation of the H175F-I224F-V291F-I332F mutant, where four residues were each replaced by a phenylalanine. The crystal structure of this quadruple mutant was determined (3.11 Å), which confirmed that the tunnel had been successfully blocked. By comparing the structure of the mutant with that of the NAMPT-NAM complex, we discovered that the presence of the aromatic side chain of Phe 224 caused a conflict with the NAM molecule bound within the tunnel. Consequently, the binding of NAM to this site would be either significantly reduced or completely lost in the mutant (Figs. 2b, 4f and S10b).
We then proceeded to investigate whether the introduced mutations altered the substrate specificity of NAMPT. Our findings showed that the enzymatic activity of the quadruple mutant was substantially lower than that of the wild-type enzyme. Specifically, blocking the tunnel caused a 129-fold increase in Km and a 3.35-fold decrease in kcat for the substrate NAM, resulting in a 403-fold decrease in catalytic efficiency (Figs. 2b, 4f and S10c). Although the changes were not that significant for PRPP, blocking the tunnel still resulted in a 2.55-fold decrease in Km and a 15.89-fold decrease in kcat, leading to an overall decrease in catalytic efficiency of about 5.80-fold (Figs. 2b, 4f and S10d). Consistent with the positively cooperative model and the existence of two NAM binding sites, the hill coefficient of the quadruple mutant dropped to 1.00, indicating a loss of positive cooperativity after blocking the tunnel (Fig. 4e). Our ITC results also supported these findings, as the affinity between the quadruple mutant and NAM was undetectable (Fig. S10b), indicating a significant reduction in NAM binding at the catalytic site without binding at the tunnel location.
Notably, our structural analysis revealed that the quadruple mutation of Xcc NAMPT not only blocks binding of NAM in the tunnel, but also effects the binding of FK866 (Fig. 5a). To investigate the role of the tunnel in inhibitor binding and the impact of its blockage on inhibitor inhibition, we performed biochemical experiments to examine the IC50 and binding affinity of FK866 with both the quadruple mutant and wild-type Xcc NAMPT. We found that FK866 effectively inhibited Xcc NAMPT with an IC50 value of 0.23 nM (Fig. 5b), similar to its inhibition of human NAMPT34. However, the quadruple mutant with blocked tunnel was significantly less sensitive to FK866 inhibition, exhibiting an IC50 value of 153.54 nM, which is more than 600-fold higher than that of the wild-type enzyme (Fig. 5c). In addition, according to the ITC results, the quadruple mutant had a weaker binding affinity for FK866, with a Kd value of ~6.47 μM, which is about 78-fold greater than the Kd value of 82.20 nM for the wild-type enzyme (Fig. 5d, e). These findings demonstrate the significance of the NAM binding tunnel for both catalysis and inhibitor function.
Structural basis of NMN recognition by Xcc NAMPT
To reveal comprehensive details for substrate recognition, the crystal structure of Xcc NAMPT complexed with NMN was determined at a resolution of 2.2 Å. In the NAMPT-NMN complex, NMN occupies only one active site, with the other remaining vacant, which is similar to the NAMPT-NAM complex (Fig. 6a). NMN is bound at the interface between the two subunits, with its phosphate-ribose moiety inserted deeply into the β/α-barrel structure from one subunit, and its nicotinamide moiety packed by domain B from the other subunit. The NAM moiety of NMN has a similar conformation to the substrate NAM in the NAMPT-NAM complex, with its nicotinamide group stacking with aromatic rings of Phe 177 and Tyr 14 and its amide group hydrogen-bonding to the carboxyl group of Asp 203. The phosphate-ribose moiety of NMN forms hydrogen-bond with Arg 373 and a heavy density of van der Waals contacts with surrounding residues, including Asp 12, Phe 177, Arg 180, Arg 293, and Gly 364.
To explore the importance of specific residues involved in substrate-binding, various mutants were generated and their functional significance was investigated. Enzyme activity experiments revealed that several NMN-contacting residues of Xcc NAMPT play a critical role in the catalytic process (Fig. 6b). Specifically, asparagine substitution of an aspartate in position 203, which is responsible for amide group recognition, resulted in decreased enzyme activity of Xcc NAMPT, which is consistent with the results from the D219N mutation in human NAMPT35. In addition, mutation of His 229, which involved in NAMPT activation, led to undetectable activity. Phosphate and ribose contacts are also required for efficient catalysis of NAMPT. As anticipated, mutation of Arg 180, which stabilizes Tyr 14 as well as the ribose group (Fig. 6a), resulted in a large reduction in activity (Fig. 6b). Similarly, alteration of the phosphate-stabilizing residue Arg 373 resulted in a loss of enzyme activity. Moreover, replacement of Asp 335 by serine, which plays a role in ribose binding, rendered the NAMPT enzyme almost completely inactivated. When this residue was mutated to asparagine, there was also a significant decrease in enzyme activity. In contrast, the R293A mutant exhibited a level of catalysis comparable to that of the wild type, suggesting that the Arg 293 is dispensable for enzyme activity. These results from mutagenesis studies are consistent with our structural observations for NMN contact.
Phosphorylation modification is required for the activation of Xcc NAMPT
According to a previous report, human NAMPT catalysis could be enhanced by ATP-dependent autophosphorylation of the enzyme31,36. To explore if Xcc NAMPT utilizes a similar mechanism for NMN catalysis, we subjected the enzyme to ATP treatment and subsequently purified it through gel filtration. We then investigated whether phosphorylation occurs during ATP hydrolysis and whether ATP-pretreated Xcc NAMPT retains histidine phosphorylation in the absence of ATP. Western blot analysis confirmed the presence of histidine phosphorylation in Xcc NAMPT during ATP hydrolysis, and this phosphorylation persisted for a certain duration after ATP removal (Fig. S11). Moreover, the phosphorylation of the histidine residue H229 was directly confirmed through liquid chromatography mass spectrometry (LC-MS/MS) analysis following tryptic digestion of the protein (Fig. S12). To assess the role of histidine phosphorylation in enzyme activation, we introduced a mutation (H229A) that abolished the histidine (H229) modification in Xcc NAMPT (Fig. S11). Notably, the H229A mutant retained some level of ATPase activity (Fig. S3); however, it completely lost the ability to produce NMN from NAM (Fig. 6b), underscoring the critical role of the histidine phosphorylation in NMN production.
To support our hypothesis, we performed isothermal titration calorimetry (ITC) experiments to measure the binding affinity of NAM to both ATP-pretreated and untreated Xcc NAMPT. The results demonstrated a significant increase in the binding capability of Xcc NAMPT to NAM following ATP pretreatment (Figs. 4d and S13). Notably, the binding of NAM to Xcc NAMPT followed a sequential binding pattern, with two distinct binding sites characterized by Kd values of 1.41 mM and 21 mM, respectively. After ATP treatment, the binding affinity of Xcc NAMPT to NAM increased for both binding sites, with Kd values of 0.33 mM and 2.99 mM, respectively. For comparison, the binding affinities of NAM to the two sites in the phosphorylated state of Xcc NAMPT were 4.22-fold and 7.02-fold higher than those observed in the unphosphorylated state, respectively (Figs. 4d and S13).
Comparison of Xcc NAMPT with different NAPRTs and QAPRTs reveals structural diversity of type II PRTs
The type II PRTs can be categorized into three main clades: NAPRT, NAMPT and QAPRT. While the structures of QAPRT and NAPRT have been determined in both bacteria and vertebrates, the structure of NAMPT has only been reported in vertebrates until the current crystal structure of Xcc NAMPT. This structure has enabled a systematical analysis of the substrate recognition specificity and working mechanism of the type II PRTs, including the three clade members from both vertebrates and bacteria. Despite low sequence similarity between NAPRT and QAPRT (Figs. 7a, S14a and S15), Xcc NAMPT shares the overall architecture characteristics of type II PRTs (Fig. 7b)37. These proteins are best described as two-domain proteins, consisting of an irregular β/α barrel domain (domain A) and an open-faced α-β sandwich domain (domain B) (Fig. 7b). Compared to other type II PRTs, Xcc NAMPT has relatively longer domains A and B, whereas QAPRTs lack a large portion of domain B in the C-terminal region and have a shorter domain A (Fig. 7a).
Type II PRTs recognize different substrates by utilizing distinct sets of residues that are specific for binding and catalyzing the unique functional groups of quinolinic acid (QA), NAM, and NA (Fig. S14a, b). To gain insights into the structural features of these enzymes, a comparison analysis was performed on the overall structure and substrate binding site of bacterial PRTs, based on the structures of StQAPRT, EfNAPRT, and Xcc NAMPT (Fig. 7c). Since the pyridine backbones of the three substrates share the same conformation, PRTs have to distinguish between them by recognizing their different ring substituents (Fig. 7c). Specifically, Xcc NAMPT and EfNAPRT utilize conserved amino acids to form aromatic stacking interactions with the pyridine ring of the substrate, while such interactions are not observed in the QA binding of QAPRT (Figs. 7c and S14a). To accommodate the highly acidic QA substrate, StQAPRT has a more positively charged active site. Although the active sites of Xcc NAMPT and NAPRTs are highly conserved, the catalytic residue Asp 203 of NAMPT is divergent and becomes a serine in NAPRT. It is reasonable to assume that the negatively charged Asp 203 in NAMPT is favorable for NAM binding but not for NA binding. Supporting this, the serine residue at the same position in NAPRT is not involved in NA binding35. These findings contribute to our understanding of the substrate specificity of Type II PRTs and highlight the importance of specific residues in substrate recognition.
The proteins in all three type II PRT clades have canonical dimers, but some QAPRT and NAPRT proteins can also form complex oligomeric states in addition to dimers (Fig. 7a). Xcc NAMPT remains in a dimeric state in both its apo form and its substrate/product bound form, indicating a constant oligomeric state during the reaction (Fig. 3c). Conversely, QAPRT can exist in either hexameric (PDB ID: 1X1O, 4I9A, 4KWV, 3C2R, 2B7N, and 5HUP) or dimeric states (PDB ID: 1O4U, 1QPN, and 1QAP)38,39,40,41,42,43,44,45,46, and NAPRT can form both monomers (PDB ID: 1YBE, 4HL7, 3O54, 1YIR, 1VLP, and 2IM5) and other oligomeric states (PDB ID: 4YUB, 2F7F, and 1YTD) (Fig. 7a)27,29. The NAPRT dimer resembles the NAMPT dimer, in which domain A of one protomer binds to domain B of the other protomer. However, in NAPRT monomers, domain A seems to form intramolecular contacts with domain B in the same molecule to maintain domain stability and active site integrity (Figs. 7e, S14c, d). In typical dimer conformations, the scaffolding α-helix is rigid and bridges domains A and B in PRTs (Fig. 7d), while monomer-type PRTs have an additional “turning loop” sequence in the middle portion of the scaffolding helix that makes it less rigid (Fig. 7d, e), enabling intramolecular contacts between domain A and domain B. Xcc NAMPT protein lacks this “turning loop” sequence and have a rigid scaffolding helix that supports a dimeric state.
Discussion
In the present study, we investigated the role of the NAMPT-mediated salvage pathway in NAD+ biosynthesis in a phytopathogenic bacterium and found that it is functional in Xcc and plays an important role in NAD+ biosynthesis. We presented the prokaryote NAMPT structure and its complex with the substrate NAM or the product NMN. The crystal packing of the NAMPT-NMN or NAMPT-NAM complex showed that only one active pocket in the complex binds NMN or NAM due to the induction of refolding of α11 and α12 (Figs. S4a, b), which inhibits NMN/NAM binding in the proximal active site of a neighboring dimer (Figs. S4c and S16). However, NAM binding at the catalytic site and tunnel area are readily identified and analyzed. The study showed that the NAM binding tunnel plays an important role in the enzymatic activity of Xcc NAMPT, allowing efficient NAM delivery to the catalytic site through the channel and/or positively regulating NAM binding at the catalytic site (Figs. 2b and 4f). Given that the binding of most present NAMPT inhibitors requires the tunnel47,48,49, the tunnel-blocking quadruple mutant described in the study may affect the binding of most inhibitors, similar to its impact on the FK866 inhibitor (Fig. 5c). Our work provides important insights into the functional significance of the tunnel in both NAMPT catalysis and inhibitor binding.
The study found that despite some similarities in primary sequence and three-dimensional structure between human and Xcc NAMPT, the latter enzyme has significantly higher catalytic activity (Fig. 2a). In humans, the NAM salvage pathway is a major source of NAD+ synthesis in humans, and NAMPT has been considered as a promising target for cancer treatment due to its role in providing NAD+ for tumor growth. Inhibiting NAMPT activity can reduce NAD+ levels11, which are crucial for tumor growth, but a deficiency in NAD+ can result in aging-related issues. To address this, researchers have developed NAMPT activators such as SBI-797812 and NAT that almost double the enzymatic activity of human NAMPT15,50. Although the NAM binding tunnel has not been identified in human NAMPT, the binding location of the NAT activator overlaps with the tunnel (Fig. S17), indicating that the tunnel NAM of Xcc NAMPT may also serve as an activator and enhance its activity. As mentioned earlier, our research demonstrated that NAM binding in the tunnel of Xcc NAMPT positively regulates NAM binding in the catalytic site (Figs. 2b, c, 4d–f, S10c and S13). Comparing the sequences and structures of Xcc and vertebrate NAMPT proteins, we noticed that the differences mainly occur in the loops above entrance of tunnel and the catalytic site (Figs. S1, S18a and S18b). In particular, the long loops above the catalytic site of human NAMPT may restrict substrate entry and product release, while variations in the loops above the tunnel entrance may determine NAM binding efficiency in the tunnel, thereby affecting the positive cooperativity of this site for the catalytic site during the reaction (Fig. S18). These findings may explain why Xcc NAMPT has a higher enzyme activity than human NAMPT. It was observed that NAMPTs and NAPRTs have longer domain B compared to QAPRTs (Fig. 7a), suggesting that evolutionary processes led to sequence extension to fulfill new functional requirements51,52. The domain B of NAMPTs and NAPRTs is involved in ATP binding and hydrolysis, which has been reported to be important for autophosphorylation28,36. The present study provides evidence that ATP-dependent autophosphorylation in Xcc NAMPT plays a vital role in NAM binding and enzyme activation, consistent with previous reports in human NAMPT and other bacterial NAMPT enzymes24,31,36. On the other hand, QAPRT proteins lack an intact domain B, and ATP is not required for their enzymatic activity. In addition, comparison studies revealed that type II PRTs have different oligomeric states. QAPRTs and NAMPTs typically function as homo-oligomers, while NAPRTs from lower organisms can exist in a monomeric state. Although the closed monomer has only been observed in the crystal structure and not in solution, the presence of a “turning loop” supports the formation of a complete active site in the monomer. Some studies suggest that homodimerization or homo-oligomerization improves protein stability and substrate specificity53,54,55, compared to the monomeric conformation.
It is generally considered that most bacteria cannot directly take up NAD+, but instead degrade it extracellularly into non-phosphorylated intermediates. However, our study showed that when NAD+ synthesis was entirely blocked in Xcc, the addition of only NAD+ but not its possible degradation products, such as NA, NmR, and NAM, could support bacterial growth in the minimal medium MMX (Fig. 1b). This finding suggests that Xcc may be capable of directly uptaking extracellular NAD+. Similarly, Chlamydia trachomatis, an obligate intracellular Gram-negative bacterial pathogen, has an NAD+ transporter and can directly utilize NAD+ from its host cells56. A homology search revealed a Xcc protein (XC_1236) that shares 19.7% identity with the Chlamydia NAD+ transporter and has a similar structure as predicted by the AlphaFold2 models (Fig. S19). We thus hypothesize that Xcc may use similar or distinct NAD+ transporters to the Chlamydia one to uptake and utilize extracellular NAD+. This intriguing question necessitates further investigation.
In conclusion, this study presents a comprehensive analysis of the three-dimensional structure and activity of Xcc NAMPT. The findings provide novel insights into the binding mode and regulatory mechanism of NAM, the function of the NAM binding tunnel in positive cooperativity and inhibitor binding, and the critical role of ATP in the enzymatic reaction. These valuable discoveries have the potential to facilitate the development and optimization of NAMPT inhibitors and activators in the future.
Methods
Bioinformatic analysis
The bacterial gene sequences were obtained from the KEGG Database57. To collect functional annotations of genes involved in NAD+ metabolism, the BLAST tool58 was employed. Multiple sequence alignment was performed using the ClustalW server (available at https://www.genome.jp/tools-bin/clustalw)59.
Bacterial strains and culture conditions
Escherichia coli (E. coli) strains were cultured at 37 °C in Luria-Bertani (LB) broth medium. The Xcc strains were cultured in nutrient-rich (NYG) or minimal (MMX) media at 28 °C with a rotational speed of 200 rpm. The following antibiotics were added to the medium to facilitate selection: kanamycin (Kan) at 25 μg/ml, rifampicin (Rif) at 50 μg/ml, and tetracycline (Tc) at 5 μg/ml for Xcc strains and 15 μg/ml for E. coli strains.
Construction and complementation of deletion mutants
Nampt and Nads deletion mutants were constructed by double-crossover homologous recombination method using the suicide plasmid pK18mobsacB (Table S1). All primers utilized for the construction of recombinant plasmids were designed based on the nucleotide sequences of Nampt and Nads (Table S2). The upstream sequence (545 bp) and downstream sequence (540 bp) of Nampt were initially amplified from the genomic DNA of the Xcc wild-type strain 8004 by using the primer sets Nampt-LF/Nampt-LR and Nampt-RF/Nampt-RR (Table S2), respectively. Thereafter, the PCR-amplified fragments were ligated into the suicide plasmid pK18mobsacB, resulting the recombinant plasmid pK18mobsacB-Nampt, which was then transferred into E. coli DH5α cells. Following this, pK18mobsacB-Nampt was transferred into strain 8004 by triparental conjugation and the transconjugants were screened on NYG plate supplemented with Rif and 10% sucrose. The obtained Nampt deletion mutant was further confirmed by PCR and named ∆Nampt (Table S1). To complement the Nampt deletion mutant, a 1561-bp fragment containing the promoter and ORF of Nampt was amplified using the primers CNampt-F/CNampt-R (Table S2). The fragment was subsequently ligated into the vector pLAFR3 to generate the recombinant plasmid pNampt, which was then introduced into the mutant strain ∆Nampt via triparental conjugation to produce the complemented strain C∆Nampt (Table S1).
To construct a plasmid for generation of Nads deletion mutant, the upstream sequence (562 bp) and downstream sequence (529 bp) of the Nads gene were PCR-amplified using the genomic DNA of strain 8004 and the primer sets Nads-LF/Nads-LR and Nads-RF/Nads-RR (Table S2), and cloned together into the vector pK18mobsacB. The resulting recombinant plasmid, pK18mobsacB-Nads, was then transferred into strain 8004 to construct the Nads deletion mutant ∆Nads (Table S1) by the procedure described above. To make a Nampt/Nads double deletion mutant, pK18mobsacB-Nads was introduced into C∆Nampt by triparental conjugation and the transconjugants were screened on NYG plate supplemented with rifampicin, tetracycline and 10% sucrose. The obtained transconjugants were incubated in NYG with 1 mM NAD+ at 28 °C for 16 h and then 37 °C for 12 h to remove the intracellular plasmid pNampt (Table S1). The resulting Nampt/Nads double deletion mutant ΔNamptΔNads was tested for Tc sensitive. All the deletion mutants were confirmed by PCR (Fig. S20).
Measurement of bacterial growth
The bacterium Xcc was grown in NYG medium at 28 °C and shaked at 200 rpm for 16–18 h. Once the cells had grown to a point where OD600 reached 1.0, they were harvested via centrifugation and washed twice with MMX medium. Following this, a volume of 0.1 ml of the cell suspension was added to 5 ml of fresh MMX medium. The NAD+, NA, NAM or NmR stock solution was adjusted to pH 7.0 before added to the medium. The culture was grown at 28 °C and the OD600 values were measured every 12 h. The experiments were repeated at least three times to ensure consistency.
Protein expression and purification
The Xcc NAMPT and human NAMPT genes were cloned into the pRSFDuet1 vector (Table S1). Detailed cloning information can be found in Table S3, and primer information is provided in Table S4. The proteins were overexpressed in E. coli BL21(DE3) cells that were cultured in LB medium with 50 μg/ml kanamycin at 37 °C until the OD600 reached 0.6. Protein expression was then induced using 0.5 mM isopropyl-β-D-thiogalactopyranoside (IPTG) at 16 °C overnight. The cells were harvested and resuspended in a lysis buffer containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl and 10 mM MgCl2, and then homogenized using a low-temperature ultra-high-pressure cell disrupter (JNBIO, Guangzhou, China). Supernatant was collected after centrifugation for 40 min at 12,000 rpm and added to a Ni-NTA resin (Qiagen). The proteins were washed and eluted with an elution buffer containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 300 mM imidazole. The proteins were further purified by anion exchange chromatography and gel filtration, followed by concentration to 30 mg/ml and flash-frozing for storage at −80 °C. In all experiments conducted in this study, the protein samples were analyzed with their His tag intact, without removal.
Proteins pretreated with ATP were prepared by incubating 1 μM protein and 1 mM ATP at 25 °C for 30 min. The incubation buffer contained 20 mM Tris-HCl (pH 7.5), 150 mM NaCl and 10 mM MgCl2. The ATP was subsequently removed by gel filtration using the Superdex 200 Increase 10/300 GL column. The final concentrations of proteins were measured by Nanodrop.
Site-specific mutagenesis
The mutant clones were constructed through a PCR-based strategy using the Fast Mutagenesis System (TransGen Biotech, Beijing, China). The specific primers employed in this process are listed in Table S4. The mutant plasmids were subsequently transformed into competent E. coli BL21(DE3) cells for expression. The mutant proteins were purified using the same procedure to that used for the wild-type protein.
Measurement of NAMPT enzyme activity
To ensure the reliability of the results, all assays were performed multiple times. NAMPT enzyme activity was assessed using a chemical approach that converted NMN into a fluorescent derivative, in accordance with established protocols15,30. To determinate the kinetic constants for NAM, NMN production was measured by incubating 0.03 nM His-tagged Xcc NAMPT at 37 °C with varying concentrations of NAM (0.1–4 μM, and a control group with no NAM), 50 μM PRPP, and 2.5 mM ATP in a reaction buffer containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, and 10 mM MgCl2. Similarly, the kinetic constants for PRPP were evaluated by incubating 0.3 nM His-tagged Xcc NAMPT with 10 μM NAM, varying concentrations of PRPP (10–120 μM), and 2.5 mM ATP in the same reaction buffer. When appropriate, ATP concentrations in the reaction buffer were adjusted from 0 to 3 mM while maintaining a constant concentration of 0.3 nM His-tagged Xcc NAMPT, 50 μM PRPP and 10 μM NAM.
To determine the kinetic parameters of human NAMPT, a concentration of 30 nM His-tagged human NAMPT was used. The PRPP concentrations in the reaction buffer were varied from 3 to 50 μM, while maintaining a constant concentration of 2.5 mM ATP and 10 μM NAM. The remaining conditions were kept the same as those employed in the Xcc NAMPT reaction system.
To evaluate the complementation of inactive mutant pairs, the pairs of mutants were mixed and incubated for 24 h at 4 °C or left untreated prior to activity assessment. For the reaction, NAMPT variants at a concentration of 25 nM were incubated with 10 μM NAM, 50 μM PRPP, and 2.5 mM ATP.
The enzymatic reactions were carried out for 15 min at 37 °C, and the quantification of NMN was achieved using the fluorescence method. Briefly, the sample containing NMN (25 μl) was mixed with 10 μl of 20% acetophenone (200 μl acetophenone dissolved in 800 μl DMSO) and 10 μl of 2 M KOH. The mixture was then incubated on ice for 10 minutes before adding 45 μl of 100% formic acid. The resulting mixture was incubated at 37 °C for 15 min. Samples (40 μl) were then transferred to a 384-well plate, and fluorescence intensity was measured using the Victor Nivo Multimode plate reader (Perkin Elmer), with excitation and emission wavelengths set at 355/40 and 435/20 nm, respectively.
To determine the kinetic parameters of NAM, the volume of NMN added was increased from 25 to 40 μl. Correspondingly, the additions of 20% acetophenone, 2 M KOH, and 100% formic acid were adjusted proportionally. The resulting final samples (120 μl) were transferred to 96-well plates, and the fluorescence values were measured following the same procedure as described earlier.
The kinetic data was analyzed using Graphpad Prism 8. To determine the hill coefficients, substrate-velocity curves were analyzed by nonlinear regression with an allosteric sigmoidal model using the equation v = Vmax Sh/(Kmh + Sh).
ATPase activity assay
ATP hydrolysis was quantified using the NADH-coupled enzyme system60. In this system, ATP is hydrolyzed by NAMPT to ADP, which is then converted back to ATP through the action of pyruvate kinase (PK) and phosphoenolpyruvate (PEP). The pyruvate produced in process is converted to lactate by L-lactate dehydrogenase (LDH), which consumes NADH. Therefore, the reduction in NADH is proportional to the rate of ATP hydrolysis. The reaction was performed at 25 °C for a specific period, and the absorbance at 340 nm was measured using a Victor Nivo Multimode plate reader (Perkin Elmer). The reaction mixture (100 μl) contained 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 10 mM MgCl2, 200 μM NADH, 1 mM PEP, 1 mM ATP, 20 units/ml LDH, 15 units/ml PK and 10 μM protein.
Western blotting using 1-pHis antibody
To perform Western blotting analysis, 1 μg of proteins was subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE) for separation. The separated proteins were then transferred onto a PVDF membrane using a running buffer containing 25 mM Tris, 192 mM glycine, and 20% methanol, at a constant current of 100 mA for 60 min. The transfer was confirmed using prestained markers. Subsequently, the PVDF membrane was blocked with 5% BSA in TBST buffer (50 mM Tris, 150 mM NaCl, 0.1% Tween-20, pH 7.5) for 2 h at room temperature (RT). The membrane was incubated overnight at 4 °C with an affinity-purified anti-N1-phosphohistidine (1-pHis) antibody (Sigma-Aldrich), diluted 1:10,000 in TBST buffer. Following antibody incubation, the membrane was washed six times for 5 min each with TBST buffer. Then, the membrane was incubated with a goat anti-rabbit secondary antibody (Abcam), diluted 1:10,000 in TBST buffer, for 1 h at 4 °C. After washing the membrane six times with a wash buffer, it was drained and exposed to an ECL chemiluminescence substrate solution (Beyotime Biotechnology, Shanghai, China) for approximately 1 min at RT. The chemiluminescence emitted from the membrane was captured using an Amersham Imager 600 (GE Healthcare Life Sciences).
LC-MS/MS analysis
Gel strips containing Xcc NAMPT proteins were excised and subjected to reduction using 10 mM DTT, and alkylation with 55 mM iodoacetamide (Thermo Fisher Scientific). In-gel digestion was then carried out by incubating the gel strips with a working solution of trypsin (Sigma-Aldrich) at a concentration of 10 ng/μL. The incubation was performed at 37 °C for 12–16 h. Following digestion, the proteins were extracted twice using a 50% aqueous solution of acetonitrile for 10 min each time.
The samples were subjected to analysis using an Easy-nLC 1000 nanoliter liquid chromatography system (Thermo Fisher Scientific), coupled with an LTQ-Orbitrap Elite high-resolution mass spectrometer (Thermo Fisher Scientific). Peptide separation was accomplished using an analytical capillary column (50 µm × 15 cm) packed with 2 µm, 100 Å C18 resin (Thermo Fisher Scientific). A linear gradient of solution A (2% acetonitrile/0.1% formic acid) and solution B (98% acetonitrile/0.1% formic acid) was employed at a flow rate of 250 nL/min for a total duration of 65 min.
The mass spectrometer was operated in ESI positive ion scan mode, and the peptides were ionized using a spray voltage of 2.0 kV. Full-scan MS spectra were acquired in the Orbitrap at a resolution of 120,000, covering the m/z range of 300–1800. Fragment ions were acquired at a resolution of 30,000. MS/MS analysis was performed using data-dependent acquisition (DDA) in MS2 mode, where the precursor ions were ranked based on their response signals, from strongest to weakest. The top 15 precursor ions were selected for high-energy collisional dissociation (HCD) and electron-transfer dissociation (ETD) alternately. The ETD reaction time was automatically determined based on the charge number and mass-to-charge ratio of the peptides. The normalized collision energy for HCD was set at 33%. A dynamic exclusion time of 30 seconds was implemented to prevent repeated selection of previously analyzed precursor ions.
The SEQUEST search engine, available in Proteome Discoverer 2.1 software61, was employed to search and analyze the spectra obtained from NanoLCMS. The amino acid sequence of the target protein was imported into the software for analysis. Subsequently, the collected spectra from mass spectrometry were compared with the theoretical sequence spectra of the proteins. Computational scoring algorithms were applied to determine high-confidence protein amino acid sequences and identify any modification sites.
The search criteria used were as follows: trypsin specificity was required for peptide cleavage, allowing for up to two missed cleavages. The mass deviation range for MS precursor ions was set at ±12 ppm, while for MS/MS fragment ions, it was set at ±0.02 Da for HCD and ±1.2 Da for ETD. Phosphorylation was set as a dynamic modification. A high-confidence score filter with a threshold of ≥99% was applied, and a false discovery rate (FDR) of <0.01 was used to ensure the reliability of the identified results.
Isothermal titration calorimetry (ITC)
Thermodynamic parameters and affinities of FK866 and NAM binding to NAMPT were measured using a MicroCal PEAQ-ITC calorimeter (Malvern Panalytical). The proteins and small molecules were diluted in PBS buffer (pH 7.5) containing 140 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4, and 10 mM MgCl2. The small molecules were injected into the proteins at 30 °C, with each titration consisting of 19 consecutive injections at 120 s intervals. The first injection used 0.5 μl of small molecules, with subsequent injections using 2 μl. The instrument reference power was set at 10 μCal/s. After each titration experiment, a control experiment was performed under identical conditions, but with buffer instead of protein solution. The data were analyzed using MicroCal PEAQ-ITC analysis software.
Determination of IC50 value for FK866
The assay was performed using a previously described method30 with slightly modifications. Specifically, 20 μl of reaction solution containing NAMPT and 0.5 μl of FK866 (containing 20% DMSO) were mixed and incubated at 37 °C for 5 min. After that, 4.5 μl of NAM was added to initiate the enzyme reaction. The final reaction system consisted of 0.4% DMSO, 20 mM Tris-HCl (pH 7.5), 10 mM MgCl2, 2.5 mM ATP, 50 μM PRPP, and 15 μM NAM. For the Xcc NAMPT reaction system, the final enzyme concentration was 0.4 nM, while for the mutant reaction system, the final enzyme concentration was 3.5 nM. The enzyme reaction was carried out for 15 min, and the NMN content was immediately determined using the same chemical method as described in the enzyme activity assay. The relative enzyme activity (Activity%) was calculated according to the equation: Activity% = NMNvar/NMN100%. Where NMN100% represents the amount of NMN produced in the system in the absence of FK866 but with 0.4% DMSO. NMNvar represents the amount of NMN produced under various concentrations of FK866. The experiment was performed at least three times. The IC50 values, which determine the concentration of FK866 that inhibits 50% of the enzyme activity, were calculated by nonlinear fitting of the data to the four-parameter IC50 logistic equation using Graphpad Prism v8.3.0.
Protein crystallization and data collection
To obtain crystals of NAMPT, the protein was concentrated to 10 mg/ml after exchanging it into a buffer containing 20 mM Tris-HCl (pH 8.0) and 150 mM NaCl. NAMPT crystals were obtained using the hanging drop method by mixing 1 μl of the protein with an equal volume of reservoir solution at 16 °C. Two different types of NAMPT crystals were produced, including plate-shaped crystals and diamond-shaped crystals.
The optimal growth condition for the plate-shaped crystals was achieved by mixing 1 μl of NAMPT with 0.833 μl of crystallization solution (0.04 M Citric acid, 0.06 M BIS-TRIS propane/pH 6.4 and 20% w/v Polyethylene glycol 3,350), followed by adding 0.167 μl 0.01 M GSH (L-Glutathione reduced), 0.01 M GSSG (L-Glutathione oxidized), and 0.2 μl 0.5% w/v 1,2,3-Heptanetriol. On the other hand, the best growth condition for the diamond-shaped crystals was achieved by mixing 1 μl of NAMPT (containing 200 mM nicotinamide) with 0.833 μl of reservoir solution (0.96 M ammonium tartrate, 0.4 M Ammonium sulfate, 0.02 M Citric acid pH 3.5), and then adding 0.167 μl N, N-Bis(3-D-gluconamidopropyl) cholamide and 0.2 μl N,N-Dimethyldecylamine-N-oxide.
To obtain crystals of the NAMPT-NMN complex, the diamond-shaped crystals were soaked in the reservoir solution containing 77 mM NMN for 30 min. To obtain crystals of the NAMPT-NAM complex, the diamond-shaped crystals were soaked in the reservoir solution containing 7.5 mM NAM and 7.5 mM PRPP for 30 min.
The optimal growth condition for quadruple mutant crystals was achieved by mixing 1 μl of mutant NAMPT with 0.833 μl of reservoir solution (0.96 M ammonium tartrate, 0.4 M Ammonium sulfate, 0.02 M Sodium acetate trihydrate), and then adding 0.167 μl N, N-Bis(3-D-gluconamidopropyl) cholamide and 0.2 μl N,N-Dimethyldecylamine-N-oxide.
All crystals were transferred to a crystallization buffer containing the reservoir solution plus 10–20% glycerol and flash-freezed using liquid nitrogen. X-Ray diffraction data were collected at the Shanghai Synchrotron Radiation Facility (SSRF) at beamline BL17U1 and BL19U1. The diffraction images were indexed, integrated, and scaled using the program XDS package62 or 3dii package63. The data processing statistics are summarized in Table 1.
Structural determination and refinement
The structure of Xcc NAMPT was determined by the molecular replacement methods using the program Phaser-MR in Phenix64. The search model used was the structure with PDB code 6E68, which shared 44.19% sequence identity. The initial model without ligands was rebuilt in COOT65 and refined in the Refinement program in Phenix66. The ligands were built in COOT using the 2mFo-DFc map and refined in Phenix. The structural figures were generated using the program PyMOL (http://pymol.sourceforge.net/). Possible interactions between Xcc NAMPT and NMN or NAM were drawn schematically with the aid of LIGPLOT67.
Sedimentation velocity analytical ultracentrifugation
Analytical ultracentrifugation experiments were carried out at 20 °C in an XL-I analytical ultracentrifuge (Beckman Coulter) using an An50Ti machine. The samples were prepared in the buffer containing 20 mM Tris-HCl (pH 7.5),150 mM NaCl and 10 mM MgCl2. For complex groups, the substrate or product was mixed with the protein at a molar ratio of 2.5:1. All samples were scanned for absorbance at 280 nm at 6-min intervals. Then, 400 μl of sample and an equal amount of reference buffer were loaded onto double-sector aluminum centerpieces and centrifuged at 50,000 rpm for 8 h. Data analysis was conducted using the software SEDFIT68 and SEDPHAT.
Gel filtration
Analytical gel filtration was performed to determine the oligomeric state of Xcc NAMPT using a Superdex 200 Increase 10/300 GL column at 16 °C. The elution buffer contained 20 mM Tris-HCl (pH 8.0) and 150 mM NaCl. Protein standards were purchased from Sigma-Aldrich (#69385), including bovine thyroglobulin (670 kDa), γ-globulins from bovine blood (150 kDa), chicken egg grade VI albumin (44.3 kDa), ribonuclease A (13.7 kDa), and a low molecular marker p-aminobenzoic acid (137 Da).
Molecular docking
FK866 was docked into the active site of Xcc NAMPT using the UCSF DOCK 6.9 program (http://dock.compbio.ucsf.edu/). The Xcc NAMPT structure was used to generate the molecular surface file. The ligand structure was retrieved from the Protein Date Bank (ID: 2GVJ, FK866). The ligand and protein were both prepared for docking by adding hydrogen atoms and charges using Chimera69. The best binding pose of FK866 corresponded to its lowest binding energy conformation in Xcc NAMPT.
Phylogenetic analysis
Multiple sequence alignments (MSA) were performed using ClustalW59. The phylogenetic tree was constructed using the Maximum Likelihood method with the Jones-Taylor-Thornton (JTT) model and 1000 bootstrap replicates in MEGA6.070. The resulting tree was displayed using FigTree 1.4.2 software available at http://tree.bio.ed.ac.uk/software/figtree, as shown in the figure.
Statistics and reproducibility
All the numbers of independent measurements are described in the figure legends. For AUC experiment, only one set of data was obtained. For structures analysis, only one dataset of each crystal condition was solved and refine. LC-MS/MS was performed at least twice. All other remaining experiments were repeated independently at least three times. Replicate experiments were successful.
Reporting summary
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
Data availability
All data are available in the main text or the supplementary information. The source data for all graphs and all uncropped images in the manuscript are in the Supplementary Data file. Coordinates have been deposited in Protein Data Bank with accession codes 7YQR (NAMPT-NAM complex), 7YQQ (NAMPT-NMN complex), 7YQO (apo NAMPT from tablet-shaped crystals), 7YQP (apo NAMPT from diamond-shaped crystals) and 8IGZ (quadruple mutant). Any additional information is available from the corresponding authors upon reasonable request.
References
Xiao, W., Wang, R.-S., Handy, D. E., Loscalzo, J. J. A. & Signaling, R. NAD (H) and NADP (H) redox couples and cellular energy metabolism. Antioxid. Redox Signal 28, 251–272 (2018).
Kochnolte, F., Haag, F., Guse, A. H., Lund, F. E. & Ziegler, M. Emerging Roles of NAD+ and its metabolites in cell signaling. Sci. Signal 2, mr1 (2009).
Fang, E. F. et al. NAD+ Replenishment Improves Lifespan and Healthspan in Ataxia Telangiectasia Models via Mitophagy and DNA Repair. Cell Metab. 24, 566–581 (2016).
Carbone, F. et al. Regulation and function of extracellular nicotinamide phosphoribosyltransferase/visfatin. Compr. Physiol. 7, 603–621 (2017).
Ryall, J. G. et al. The NAD(+)-dependent SIRT1 deacetylase translates a metabolic switch into regulatory epigenetics in skeletal muscle stem cells. Cell Stem Cell 16, 171–183 (2015).
Hou, Y. et al. NAD+ supplementation normalizes key Alzheimer’s features and DNA damage responses in a new AD mouse model with introduced DNA repair deficiency. Proc. Natl Acad. Sci. USA 115, E1876–E1885 (2018).
Yoshida, M. et al. Extracellular vesicle-contained eNAMPT delays aging and extends lifespan in mice. Cell Metab. 30, 329–342 (2019).
Chowdhry, S. et al. NAD metabolic dependency in cancer is shaped by gene amplification and enhancer remodelling. Nature 569, 570–575 (2019).
Verdin, E. NAD+ in aging, metabolism, and neurodegeneration. Science 350, 1208–1213 (2015).
Gazzaniga, F. S., Stebbins, R., Chang, S. Z., Mcpeek, M. A. & Brenner, C. Microbial NAD metabolism: lessons from comparative genomics. Microbiol. Mol. Biol. Rev. 73, 529–541 (2009).
Garten, A., Petzold, S., Körner, A., Imai, S. & Kiess, W. Nampt: linking NAD biology, metabolism and cancer. Trends Endocrinol. Metab. 20, 130–138 (2009).
Revollo, J. R., Grimm, A. A. & Imai, S. The NAD biosynthesis pathway mediated by nicotinamide phosphoribosyltransferase regulates Sir2 activity in mammalian cells. J. Biol. Chem. 279, 50754–50763 (2004).
Bockwoldt, M. et al. Identification of evolutionary and kinetic drivers of NAD-dependent signaling. Proc. Natl Acad. Sci. USA 116, 15957–15966 (2019).
Shats, I. et al. Bacteria boost mammalian host NAD metabolism by engaging the deamidated biosynthesis pathway. Cell Metab. 31, 564 (2020).
Gardell, S. J. et al. Boosting NAD(+) with a small molecule that activates NAMPT. Nat. Commun. 10, 3241 (2019).
Galli, U. et al. Medicinal chemistry of nicotinamide phosphoribosyltransferase (NAMPT) inhibitors. J. Med. Chem. 56, 6279–6296 (2013).
Desautels, D. N., Bourrier, M., Peltier, C., Banerji, V. & Banerji, S. O. Effect of FK866, NAMPT inhibitor, on small cell lung cancer cell lines. J. Clin. Oncol. 33, e18562–e18562 (2015).
Grozio, A. et al. CD73 protein as a source of extracellular precursors for sustained NAD+ biosynthesis in FK866-treated tumor cells. J. Biol. Chem. 288, 25938–25949 (2013).
Wang, W. et al. Structural basis for resistance to diverse classes of NAMPT inhibitors. PLoS ONE 9, e109366 (2014).
Lima, W. C., Varani, A. M. & Menck, C. F. M. NAD biosynthesis evolution in bacteria: lateral gene transfer of kynurenine pathway in Xanthomonadales and Flavobacteriales. Mol. Biol. Evol. 26, 399–406 (2008).
Huang, N. et al. Bifunctional NMN adenylyltransferase/ADP-ribose pyrophosphatase: structure and function in bacterial NAD metabolism. Structure 16, 196–209 (2008).
Foster, J. W., Kinney, D. M. & Moat, A. G. Pyridine nucleotide cycle of Salmonella typhimurium: isolation and characterization of pncA, pncB, and pncC mutants and utilization of exogenous nicotinamide adenine dinucleotide. J. Bacteriol. 137, 1165–1175 (1979).
Galeazzi, L. et al. Identification of nicotinamide mononucleotide deamidase of the bacterial pyridine nucleotide cycle reveals a novel broadly conserved amidohydrolase family. J. Biol. Chem. 286, 40365–40375 (2011).
Sorci, L. et al. Genomics-driven reconstruction of acinetobacter NAD metabolism: insights for antibacterial target selection. J. Biol. Chem. 285, 39490–39499 (2010).
Amici, A. et al. Synthesis and degradation of adenosine 5′-tetraphosphate by nicotinamide and nicotinate phosphoribosyltransferases. Cell Chem. Biol. 24, 553–564 (2017).
Gerdes, S. Y. et al. Comparative genomics of NAD biosynthesis in cyanobacteria. J. Bacteriol. 188, 3012–3023 (2006).
Marletta, A. S. et al. Crystal structure of human nicotinic acid phosphoribosyltransferase. FEBS Open Bio 5, 419–428 (2015).
Chappie, J. S. et al. The structure of a eukaryotic nicotinic acid phosphoribosyltransferase reveals structural heterogeneity among type II PRTases. Structure 13, 1385–1396 (2005).
Shin, D. H. et al. Crystal structure of a nicotinate phosphoribosyltransferase from Thermoplasma acidophilum. J. Biol. Chem. 280, 18326–18335 (2005).
Zhang, R. Y. et al. A fluorometric assay for high-throughput screening targeting nicotinamide phosphoribosyltransferase. Anal. Biochem. 412, 18–25 (2011).
Burgos, E. S. & Schramm, V. L. Weak coupling of ATP hydrolysis to the chemical equilibrium of human nicotinamide phosphoribosyltransferase. Biochemistry 47, 11086–11096 (2008).
Brown, A. Analysis of cooperativity by isothermal titration calorimetry. Int. J. Mol. Sci. 10, 3457–3477 (2009).
MicroCal PEAQ-ITC Analysis Software User Manual (Malvern Instruments Ltd., 2015).
Hasmann, M. & Schemainda, I. FK866, a highly specific noncompetitive inhibitor of nicotinamide phosphoribosyltransferase, represents a novel mechanism for induction of tumor cell apoptosis. Cancer Res. 63, 7436–7442 (2003).
Khan, J. A., Tao, X. & Tong, L. Molecular basis for the inhibition of human NMPRTase, a novel target for anticancer agents. Nat. Struct. Mol. Biol. 13, 582–588 (2006).
Burgos, E. S., Ho, M.-C., Almo, S. C. & Schramm, V. L. A phosphoenzyme mimic, overlapping catalytic sites and reaction coordinate motion for human NAMPT. Proc. Natl Acad. Sci. USA 106, 13748–13753 (2009).
Cao, H., Pietrak, B. L. & Grubmeyer, C. Quinolinate phosphoribosyltransferase: kinetic mechanism for a type II PRTase. Biochemistry 41, 3520–3528 (2002).
Malik, S. S., Patterson, D. N., Ncube, Z. & Toth, E. A. The crystal structure of human quinolinic acid phosphoribosyltransferase in complex with its inhibitor phthalic acid. Proteins: Struct., Funct., Bioinforma. 82, 405–414 (2014).
Youn, H. S. et al. Crystal structure of Sus scrofa quinolinate phosphoribosyltransferase in complex with nicotinate mononucleotide. PLoS ONE 8, e62027 (2013).
Luccio, E. D. & Wilson, D. K. Comprehensive X-ray structural studies of the quinolinate phosphoribosyl transferase (BNA6) from Saccharomyces cerevisiae. Biochemistry 47, 4039–4050 (2008).
Kim, M. K., Im, Y. J., Lee, J. H. & Eom, S. H. Crystal structure of quinolinic acid phosphoribosyltransferase from Helicobacter pylori. Proteins 63, 252–255 (2006).
Booth, W. T. et al. Streptococcus pyogenes quinolinate-salvage pathway-structural and functional studies of quinolinate phosphoribosyl transferase and NH3 -dependent NAD+ synthetase. FEBS J. 284, 2425–2441 (2017).
Liu, H. et al. Structural and kinetic characterization of quinolinate phosphoribosyltransferase (hQPRTase) from homo sapiens. J. Mol. Biol. 373, 755–763 (2007).
Schwarzenbacher, R. et al. Crystal structure of a type II quinolic acid phosphoribosyltransferase (TM1645) from Thermotoga maritima at 2.50 A resolution. Proteins 55, 768–771 (2004).
Sharma, V., Grubmeyer, C. & Sacchettini, J. C. Crystal structure of quinolinic acid phosphoribosyltransferase from Mmycobacterium tuberculosis: a potential TB drug target. Structure 6, 1587–1599 (1998).
Eads, J. C., Ozturk, D., Wexler, T. B., Grubmeyer, C. & Sacchettini, J. C. A new function for a common fold: the crystal structure of quinolinic acid phosphoribosyltransferase. Structure 5, 47–58 (1997).
Galli, U. et al. Recent advances in NAMPT inhibitors: a novel immunotherapic strategy. Front. Pharm. 11, 656 (2020).
Zhang, S. L. et al. Crystal structure-based comparison of two NAMPT inhibitors. Acta Pharmacologica Sin. 39, 294–301 (2018).
Wei, Y., Xiang, H. & Zhang, W. Review of various NAMPT inhibitors for the treatment of cancer. Front. Pharm. 13, 970553 (2022).
Yao, H. et al. Discovery of small-molecule activators of nicotinamide phosphoribosyltransferase (NAMPT) and their preclinical neuroprotective activity. Cell Res. 32, 1–15 (2022).
Xu, L. et al. Average gene length is highly conserved in prokaryotes and eukaryotes and diverges only between the two kingdoms. Mol. Biol. Evol. 23, 1107–1108 (2006).
Scopes, R. K. Longer is better—random elongation mutagenesis. Nat. Biotechnol. 17, 21 (1999). -21.
André, I., Strauss, C. E. M., Kaplan, D. B., Bradley, P. & Baker, D. Emergence of symmetry in homooligomeric biological assemblies. Proc. Natl Acad. Sci. USA 105, 16148–16152 (2008).
Marianayagam, N. J., Sunde, M. & Matthews, J. M. The power of two: protein dimerization in biology. Trends Biochem. Sci. 29, 618–625 (2004).
Larsen, T. A., Olson, A. J. & Goodsell, D. S. Morphology of protein–protein interfaces. Structure 6, 421–427 (1998).
Fisher, D. J., Fernández, R. E. & Maurelli, A. T. Chlamydia trachomatis transports NAD via the Npt1 ATP/ADP translocase. J. Bacteriol. 195, 3381–3386 (2013).
Kanehisa, M. The KEGG database. Novartis Found. Symp. 247, 91–101, discussion 101–103, 119–128, 244–152 (2002).
Sayers, E. W. et al. Database resources of the national center for biotechnology information. Nucleic Acids Res. 50, D20–D26 (2022).
Thompson, J. D., Higgins, D. G. & Gibson, T. J. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res. 22, 4673–4680 (1994).
Roskoski, R. in Methods in Enzymology, Vol. 99, 3–6 (Academic Press, 1983).
Orsburn, B. C. Proteome discoverer—a community enhanced data processing suite for protein informatics. Proteomes 9, 15 (2021).
Kabsch, W. XDS. Acta Crystallogr. Sect. D., Biol. Crystallogr. 66, 125–132 (2010).
Winter, G. xia2: an expert system for macromolecular crystallography data reduction. J. Appl. Crystallogr. 43, 186–190 (2010).
Zwart, P. H. et al. Automated structure solution with the PHENIX suite. Methods Mol. Biol. 426, 419–435 (2008).
Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr. Sect. D., Biol. Crystallogr. 60, 2126–2132 (2004).
Murshudov, G. N., Vagin, A. A. & Dodson, E. J. Refinement of macromolecular structures by the maximum-likelihood method. Acta Crystallogr. Sect. D.-Biol. Crystallogr. 53, 240–255 (1997).
Wallace, A. C., Laskowski, R. A. & Thornton, J. M. LIGPLOT: a program to generate schematic diagrams of protein-ligand interactions. Protein Eng. 8, 127–134 (1995).
Schuck, P., Perugini, M. A., Gonzales, N. R., Howlett, G. J. & Schubert, D. Size-distribution analysis of proteins by analytical ultracentrifugation: strategies and application to model systems. Biophys. J. 82, 1096–1111 (2002).
Allen, W. J. et al. DOCK 6: Impact of new features and current docking performance. J. Comput. Chem. 36, 1132–1156 (2015).
Tamura, K., Stecher, G., Peterson, D., Filipski, A. & Kumar, S. MEGA6: molecular evolutionary genetics analysis version 6.0. Mol. Biol. Evol. 30, 2725–2729 (2013).
Acknowledgements
We thank the staff members of Shanghai Synchrotron Radiation Facility, Dr. Bing Zhang from ShanghaiTech University, and Dr. Liming Yan from Tsinghua University for assistance during data collection. We also thank Dr. Qing Chang and Dr. Wendan Chu (Tsinghua University Technology Centre for Protein Research) for assistance with AUC data analysis. This work was supported by grants from the National Natural Science Foundation of China (32160064), Guangxi Natural Science Foundation (2020GXNSFFA297007), Ba-Gui Scholar Program of Guangxi (To Z.G.H), and China Scholarship Council (CSC) (202108455001).
Author information
Authors and Affiliations
Contributions
Conceptualization: Z.H.M. and J.L.T. Methodology: G.L.X., J.X.M., Q.F., Q.P., F.M.L., and Z.H.M. Investigation: G.L.X., J.X.M., D.J.T., Q.P., X.Q.S., and W.H. Visualization: G.L.X., J.X.M., Q.F., Q.P., X.Q.S., X.J. Q.Q.Z., W.H., and Y.Q.K. Supervision: Z.H.M. and J.L.T. Writing—original draft: G.L.X., J.X.M. Writing—review and editing: G.L.X., J.X.M., Q.F., Z.H.M., and J.L.T.
Corresponding authors
Ethics declarations
Competing interests
The authors declare no competing interests.
Peer review
Peer review information
This manuscript has been previously reviewed at another Nature Portfolio journal. The manuscript was considered suitable for publication without further review at Communications Biology. Primary Handling Editors: Tobias Goris. A peer review file is available.
Additional information
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
About this article
Cite this article
Xu, G., Ma, J., Fang, Q. et al. Structural insights into Xanthomonas campestris pv. campestris NAD+ biosynthesis via the NAM salvage pathway. Commun Biol 7, 255 (2024). https://doi.org/10.1038/s42003-024-05921-3
Received:
Accepted:
Published:
DOI: https://doi.org/10.1038/s42003-024-05921-3
- Springer Nature Limited