Aminopropyltransferases are essential enzymes that form polyamines in eukaryotic and most prokaryotic cells. Spermidine synthase (SpdS) is one of the most well-studied enzymes in this biosynthetic pathway. The enzyme uses decarboxylated S-adenosylmethionine and a short-chain polyamine (putrescine) to make a medium-chain polyamine (spermidine) and 5′-deoxy-5′-methylthioadenosine as a byproduct. Here, we report a new spermidine synthase inhibitor, decarboxylated S-adenosylhomocysteine (dcSAH). The inhibitor was synthesized, and dose-dependent inhibition of human, Thermatoga maritima, and Plasmodium falciparum spermidine synthases, as well as functionally homologous human spermine synthase, was determined. The human SpdS/dcSAH complex structure was determined by X-ray crystallography at 2.0 Å resolution and showed consistent active site positioning and coordination with previously known structures. Isothermal calorimetry binding assays confirmed inhibitor binding to human SpdS with Kd of 1.1 ± 0.3 μM in the absence of putrescine and 3.2 ± 0.1 μM in the presence of putrescine. These results indicate a potential for further inhibitor development based on the dcSAH scaffold.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Polyamines are essential for normal cell growth and development and the polyamine biosynthetic pathway is an important target for the development of therapeutics including cancer chemopreventive agents, cancer therapeutics, and agents for treatment of diseases, such as sleeping sickness, Chagas disease, leishmaniasis, and malaria, caused by parasitic protozoa.1–7 Spermidine is a major polyamine that serves multiple functions in growth including its role as a precursor of hypusine, which is essential for the activity of the eukaryotic initiation factor elF-5A.8–11
Polyamines are formed in eukaryotes and in many, but not all, prokaryotes by the action of aminopropyltransferases. These comprise a class of enzymes that use decarboxylated S-adenosylmethionine (dcAdoMet) as a substrate and transfer the aminopropyl group to an amine acceptor molecule generating 5′-deoxy-5′-methylthioadenosine (MTA) as a byproduct. Some aminopropyltransferases such as human spermidine synthase (SpdS) (which acts upon putrescine) and spermine synthase (SpmS) (acting on spermidine) are highly specific for their amine acceptors,12–14 while others, such as those from acute thermophiles, which contain a variety of polyamines not found in mammals, are less discriminating.12, 13, 15–17 There are now numerous published structures for aminopropyltransferases including those for SpdS from Thermotoga maritima (TmSpdS),16Caenorhabditis elegans,18Plasmodium falciparum (PfSpdS),19Helicobacter pylori,20 human (hSpdS),13Arabidopsis thaliana (PDB code 2Q41), and Trypanosoma cruzi (PDB code 3BWC), aminopropylagmatine/aminopropylcadaverine synthases from Thermus thermophilus (PDB code 1UIR), Pyrococcus horikoshii (PDB code 2ZSU) and Pyrococcus furiosus,15 and SpmS from humans (hSpmS).14
A general mechanism for aminopropyl transfer has been proposed based on kinetic studies of hSpdS, TmSpdS, and hSpmS, their structures with bound substrates and inhibitors, and the results of site-directed mutagenesis of key residues (Fig. 1).13, 16 This mechanism depends on a conserved Asp (residue 173 in hSpdS) in the active site that interacts with the bound amine substrate to deprotonate the attacking nitrogen atom of the amine. This interaction is reinforced by additional interactions of this nitrogen atom with the hydroxyl group of a conserved Tyr (residue 79 in hSpdS) and a backbone carbonyl group (Ser174 in hSpdS). These interactions allow the attack on the methylene carbon atom of the aminopropyl group attached to the sulfonium center of dcAdoMet. Electron transfer to this sulfur atom completes the reaction forming MTA and the polyamine product. The positively charged sulfonium ion is a critical part of this reaction. The dcAdoMet binding site at the active site of the aminopropyltransferases whose structures are known shows a negatively charged binding pocket with highly conserved residues that interact with the purine base, the 6-amino substituent and the ribose hydroxyl groups of the nucleoside substrate as well as the aminopropyl group of the sidechain, but there are no direct interactions with the S+. These findings raise the possibility that an analog of dcAdoMet lacking this positive charge would bind specifically in the active site but would be unable to take part in the displacement reaction.
The most conservative mimic of the dcAdoMet substrate that might serve as an inhibitor of spermidine synthase is decarboxylated S-adenosylhomocysteine (dcSAH; SRI6402), a derivative of SAH, which in turn is a key metabolic product of AdoMet.21 The metabolic origin of dcSAH is not clear; however, it has been isolated in trace amounts in rat liver and adrenals.22 In nature, dcSAH is found at the highest levels in the Arius felis (sea catfish) eye tissue as part of its reflective pigment material.23 To date, dcSAH has not been reported to inhibit any enzyme.
We synthesized dcSAH to investigate its ability to inhibit hSpdS and hSpmS, and we determined the crystal structure of its complex with hSpdS (PDB ID 3RW9). X-ray diffraction data for the complex show the binding site occupied by dcSAH in a similar conformation to the previously determined substrate and product complex structures. We did not observe the binding of putrescine upon co-crystallization of hSpdS in the presence of dcSAH and putrescine, suggesting that binding would not require the cellular polyamine concentration. Putrescine effect on dcSAH binding was examined by isothermal calorimetry (ITC). The results demonstrate that dcSAH is a potentially useful first generation inhibitor that may be used to develop more potent and specific inhibitors of aminopropyltransferases.
Inhibition of aminopropyltransferase activity by dcSAH
As shown in Figure 2, dcSAH was a potent dose-dependent inhibitor of both hSpdS and hSpmS with IC50 values 43 μM and 5 μM, respectively. In preliminary studies, we also tested dcSAH as an inhibitor of two other aminopropyltransferases: TmSpdS and PfSpdS. Both of these enzymes were significantly more sensitive than hSpdS with IC50 values of 2 μM and 5 μM, respectively.
Crystal structure of the hSpdS/dcSAH complex
The hSpdS structure in complex with dcSAH was determined [Fig. 3(A)]. The bound inhibitor was present in both monomer chains of the hSpdS dimer [Fig. 3(B)]. The hSpdS dimer consists of an N-terminal β-sheet domain (residues 1–72), a central core domain (residues 73–267), and a C-terminal domain (residues 268–302), as described previously.13 The first 14 residues were disordered in the crystal structure, consistent with previous structural data. Inhibitor complex structure superposition with previously determined hSpdS/dcAdoMet dimer (PDB ID 2O0L) yielded an overall RMS of 0.22 Å.
The dcSAH binding is facilitated by hydrogen bonding with highly conserved active site residues [Fig. 3(B,C)].13 The positively charged amino group of the aminopropyl moiety is positioned in the negatively charged polar binding pocket of Asp104 (absolutely conserved), Asp173 (absolutely conserved), and Gln80. The ribose ring forms hydrogen bonds with the side chains of Glu124 and an absolutely conserved Gln49. The inhibitor is bound in the S form, with N1 of the adenosine moiety interacting with the backbone of Gly156, N7 weakly hydrogen bonding with the backbone of Ala181, and N6 coordinated by Asp155 and the backbone of Pro180.
Comparison of the active site positions with previous structures
The position of the adenosyl moiety in hSpdS/dcSAH closely matches the hSpdS/MTA structure, determined previously (PDB ID 2O07). The largest deviation of 2.4 Å was between the sulfur atoms. The amine group of dcSAH is positioned within 0.7 Å of N1 in spermidine in the product complex structure, in excellent agreement with the expected amine coordination [Fig. 4(A)]. The rest of the spermidine/putrescine binding site cavity is occupied by water molecules, which were consistent between the two chains of the hSpdS/dcSAH dimer.
The gate keeping loop (residues 174–182) that envelops the active site is known to become ordered upon ligand binding.13 In the hSpdS/dcSAH complex structure the gate keeping loop is only partially ordered, with residues 175–178 absent in chain A [Fig. 4(B)] and residues 176–178 absent in chain B. Weak partial density was observed, but the main chain connectivity in the loop could not be determined.
The putrescine binding site is occupied by water molecules
No putrescine binding was observed in the hSpdS/dcSAH complex structure (Fig. 4). Superposition with the known hSpdS/MTA/putrescine complex (PDB ID 2O06) shows a water molecule occupying the N2 binding pocket of putrescine (difference of 0.8 Å). A second water molecule is hydrogen bonded to the side chains of Gln70 and Gln206, which are found pointing towards the putrescine binding cavity with an average of 1.0 Å deviation compared to the putrescine complex structure (PDB code 2O06, Fig. 4).
Presence of excess putrescine lowers dcSAH binding affinity to hSpdS
ITC assays were done in the presence and absence of putrescine at 25°C to determine the binding affinity of dcSAH to hSpdS (Fig. 5). In the absence of putrescine, Kd = 1.1 ± 0.3 μM, ΔH = −10.8 ± 0.5 kcal/mol, and ΔS = −9.1 cal/mol/deg, which results in the ΔG of −8.1 kcal/mol. In the presence of putrescine, equilibrated with the hSpdS solution at the ratio of 10:1 to the enzyme, the resulting binding was determined to have Kd = 3.2 ± 0.3 μM, ΔH = −23.9 ± 0.4 kcal/mol, and ΔS = −55 cal/mol/deg, which gives the ΔG of −7.5 kcal/mol. The difference in Kd values was determined to be statistically significant within the accuracy of the experiments at the 95% confidence level.
Typically, SpdS forms a homodimer, and our findings are consistent with the expected oligomerization state.12 Here we report a complex structure of hSpdS with the novel inhibitor dcSAH bound to both monomers of hSpdS in the crystal structure. Electron density in the putrescine binding site is indicative of only the water molecules at the charged amine coordination sites, even with the condition of 5 mM putrescine in the crystallization buffer. Binding of dcSAH in the active site is consistent with previously determined substrate, product, or inhibitor SpdS structures, indicating a conserved mechanism of binding and catalysis. The gate keeping loop on one side of the active site that is typically ordered around an occupied active site and disordered in apo structures is found mostly disordered in our structure with only weak, partial density in the connecting region [Fig. 4(B)].
Inhibitor binding was tested for SpdS from several species. IC50 values ranged from 2 μM in TmSpdS, 5 μM in PfSpdS, to 43 μM in hSpdS. The dcSAH is also effective in hSpmS inhibition with an IC50 of 5 μM. This distribution of activity places hSpdS as the weakest binding partner for dcSAH of the constructs tested. Additionally, due to close structural similarity to dcAdoMet, it is not surprising that we observed significant inhibition of hSpmS as well.
Inhibitor binding affinity to hSpdS was determined to be in the low micromolar range by ITC. There was a slight increase of binding affinity in the absence of putrescine, which suggests a preference for an empty putrescine binding pocket observed in the crystal structure. The gate keeping loop has been observed in the disordered state previously in the hSpdS/dcAdoMet substrate complex without putrescine,13 so the current structure is consistent with putrescine absence in the active site. The ordering of the gate keeping loop is predicted to be involved in the positioning of putrescine for the transferase reaction, and the opening of the gate keeping loop is necessary for the release of the products.13 Previously, putrescine was crystallized only with the product MTA present,13 suggesting that putrescine orientation and binding is sensitive to the overall active site occupancy and the extended environment including the orientation of the gate keeping loop.
The necessity of polyamines in cellular development and viability make SpdS inhibition an important study for therapeutic purposes. The necessary production of higher polyamines like spermidine or spermine in eukaryotes requires the action of SpdS; therefore, there have been several inhibitors developed for this enzyme. The strongest inhibitor to date is a transition-state mimic adenosylspermidine [3-(R,S)-(5′-deoxy-5′-carbaadenos-6′yl)-spermidine], with an IC50 of ∼ 14 nM.24 While adenosylspermidine, like several other inhibitors, is not exclusive to SpdS and also inhibits SpmS, these families of structures exemplify the mode of action of aminopropyltransferases, as well as direct the drug design efforts.
Our structure of hSpdS with the novel inhibitor dcSAH adds a new scaffold for SpdS inhibitor development using a substrate mimic. The simple, yet effective, change results from removal of the methyl group and, consequently, the positive charge at the reactive sulfonium center. This change creates an inactive substrate mimic (dcAdoMet-like small molecule) that binds to the SpdS active site. The structure presented here shows the dcSAH position in the active site and enables further efforts of inhibitor improvement while using the hSpdS/dcSAH complex structure as the starting point in inhibitor design and optimization.
To approximately 250 mL of liquid ammonia in a dry ice bath was added 1.12 g (5.14 mmol) of 3-benzylthio-1-aminopropane hydrochloride.25 This solution was stirred while metallic sodium was added until the blue color persisted for about 10 min. After the color dispelled, 5′-O-p-toluenesulfonyladenosine (2.43 g, 5.75 mmol) was added slowly and the solution was stirred for about 10 min. Then the solution was kept sufficiently cool to allow the ammonia to slowly evaporate over a period of hours. The residue was taken up in water, brought to approximately pH 2, and then washed with CHCl3. The aqueous layer was applied to a column of Dowex 50W-X4 (NH4+ form, 50–100 mesh), and eluted with a gradient of aqueous NH4OH beginning at 1N and ending with 10N. Product fractions were combined, evaporated to dryness, taken up in 1N H2SO4, and filtered. Dilution with two volumes of ethanol resulted in the formation of the desired product as a white solid, which was filtered off and dried, yielding 2.09 g (90%) of dcSAH, m.p. 178–180°C (dec), literature m.p. 180–183°C.25 Microanalytical data confirmed a composition of dcSAH·H2SO4·0.9 H2O.
Overexpression and purification of hSpdS
hSpdS was overexpressed in Escherichia coli BL21(DE3) containing the pET-28a(+) (Novagen), which encodes an N-terminal His6 tag followed by a tobacco etch virus (TEV) protease cleavage site. Cells were grown in LB medium with 100 mg/mL kanamycin with agitation and induced at OD600 = 0.6 with 1 mM isopropyl-1-thio-β-D-thiogalactoside (IPTG) at 15°C. Cells were harvested after 20–24 h with centrifugation at 8000g for 15 min, frozen overnight at −80°C and resuspended in the lysis buffer (50 mM Tris-HCl, pH 8.0, 100 mM sodium chloride, 5% v/v glycerol, 0.1 mM phenylmethylsulphonyl fluoride, and 2.5 mM putrescine). Cells were lysed using a French pressure cell, the lysate was cleared by centrifugation, and the cell supernatant was purified using a Ni-NTA affinity chromatography column (Qiagen) in the buffer above with 10 mM imidazole, eluting the desired protein with 250 mM immidazole. The His6 tag was cleaved overnight via dialysis in the presence of TEV protease against a buffer of 20 mM Tris-HCl, pH 8.0, 100 mM sodium chloride, 5 mM calcium chloride, 2.5 mM magnesium chloride, 5 mM β-mercaptoethanol, and 1% v/v glycerol. Cleaved protein sample was eluted through a Ni-NTA affinity column and further purified using a Superdex 75 size-exclusion column (GE Healthcare) equilibrated in 20 mM Tris-HCl pH 8.0, 50 mM sodium chloride. The final protein sample was desalted using a HiPrep 26/10 desalting column (GE Healthcare) into the final buffer of 20 mM Tris-HCl, pH 8.0 and concentrated to 14 mg/mL in a 10 kDa cutoff microcon concentrator (Amicon). The protein was judged to be greater than 95% pure by SDS-PAGE gel analysis (data not shown). The yield of the purified protein was 10 mg/L.
Crystallization was carried out using the hanging drop vapor diffusion method at 22°C, each drop containing 1.5 μL reservoir solution and 1.5 μL protein solution. Initial conditions were obtained from sparse matrix screens (Hampton Research, Emerald Biostructures). Optimized crystals grew in 100 mM Tris-HCl, pH 8.4, 200 mM sodium sulfate, and 20% w/v polyethylene glycol 3350. To obtain the ligand-complexed structure, 5 mM putrescine and 5 mM dcSAH were added to the protein solution and left to equilibrate for up to 6 h on ice prior to crystallization. The resulting plate-like crystals were cryoprotected with 10% v/v glycerol and flash frozen in liquid nitrogen for X-ray intensity measurements.
Data collection and processing
The X-ray diffraction data were collected at the Northeastern Collaborative Access Team (NE-CAT) 24-ID-C beamline at the Advanced Photon Source (APS). Data were collected over 1–360° range with a 1 s exposure time and a 0.5° oscillation range using an ADSC Quantum 315 detector at 0.97918 Å and crystal to detector distance of 250 mm. The data were indexed, integrated, and scaled using the HKL2000 program suite.26 The data collection statistics are summarized in Table I.
Table I. Data Collection Statistics for hSpdS Complex with dcSAH
The hSpdS/dcSAH structure was determined and refined using the published structure of hSpdS (PDB ID 2O07) at the same space group as the starting model, upon removal of the water molecules and the bound product complex. The model building was carried out using the program COOT 0.6.27 The refinement process involved rounds of simulated annealing, energy minimization, B-factor refinement, calculation of difference Fourier maps, and manual model optimization using the program PHENIX release 1.5–2.28 A difference Fo − Fc Fourier map was used to identify the ligands. dcSAH was successfully positioned in both chains present in the asymmetric unit; however, no density was seen for the putrescine molecule, and the putrescine binding site was occupied by water molecules. The final refinement statistics are given in Table II, where the Ramachandran plot statistics were calculated using PROCHECK29 and showed the partially solvent-exposed Glu208 in each chain to be in the disallowed region.
Table II. Refinement Statistics for the hSpdS/dcSAH Complex
R-factor = ∑hkl||Fobs| − k|Fcal||/∑hkl|Fobs|, where Fobs and Fcal are observed and calculated structure factors, respectively.
R-free is the sum is extended over a subset of reflections (5%) that were excluded from all stages of refinement.
The Glu208 residue in both chains of the dimer is in the disallowed region of the Ramachandran plot.
Recombinant hSpmS and TmSpdS were produced and purified as previously described.14, 16
Assay of aminopropyltransferase activity
Activity was measured by following the production of [35S]MTA from [35S]dcAdoMet in 100 mM sodium phosphate buffer (pH 7.5), in the presence of spermidine or spermine as the amine acceptor.30 Reactions were run in triplicate for 1 h with an amount of enzyme that gave a linear rate of MTA production at 37°C and concentrations of inhibitor that gave from 10 to 90% inhibition. The IC50 values were derived from plots of the residual activity against inhibitor concentration.
Isothermal titration calorimetry assays
ITC titrations were performed using a VP-ITC MicroCalorimeter (MicroCal, Inc.) equilibrated at 25°C. Prior to the each measurement all solutions were degassed under vacuum for ∼ 30 min in a ThermoVac accessory at 22°C. The same stock buffer of 20 mM sodium phosphate pH 7.5, 50 mM NaCl was used in preparing all solutions. Protein solution at 37 μM was placed in the 1.6-mL calorimeter cell. Stock dcSAH inhibitor solution was prepared once for all ITC experiments at the 743 μM concentration. Experiments in the presence of putrescine were performed after incubation with hSpdS for ∼ 10 h in ice with 10:1 putrescine:hSpdS. Same stock solution of putrescine was added to the inhibitor solution in order to maintain the same putrescine concentration during the titration. Control experiments were performed with remaining inhibitor solutions injected against degassed buffer with and without putrescine in the calorimeter cell. Inhibitor was injected over 25 injections (22 in the experiment without putrescine) of 10 μL over 20 s (first injection of 5 μL over 10 s) with 600 s delay between injections. Due to residual heats in the last injections of the experiment without putrescine that were not accounted for by the control experiment, the bottom graph was manually offset to reach zero kcal/mol during data processing.
The statistical significance of the difference in binding affinity within the accuracy of the experiment was determined at the 95% confidence level by applying the standard equivalency test: |KdA – KdB| = 2(SDA2 + SDB2)1/2, where SD is the standard deviation of the measurement.
Access to the MicroCal ITC was graciously provided by Dr. William Horne laboratory at the Cornell Veterinary School.