Continuous Assay for the dNTP Triphosphohydrolase of Activated SAMHD1

Abstract

Sterile alpha motif and histidine-aspartate domain-containing protein 1 (SAMHD1) is the only member of the triphosphoric monoester hydrolase family in humans $(\text{dNTP}+{\mathrm{H}}_2\mathrm{O}\to \text{dN}+\text{PPPi})$. The dNTPase activity of SAMHD1 inhibits DNA synthesis, resulting in cell-cycle arrest and restricting viral replication. The complex allosteric regulation mechanism of SAMHD1 and a reaction that lacks a direct spectroscopic signal make its kinetic analysis and inhibitor discovery challenging. We describe a continuous assay for monitoring SAMHD1 phosphatase activity in its activated physiological state. The assay uses a sequential assembly to generate the active tetrameric form of the enzyme. Two phosphatases convert inorganic triphosphate (PPPi) to inorganic phosphate (Pi). The released Pi reacts with the 7-methyl-6-thioguanosine and purine nucleoside phosphorylase to provide a sensitive continuous spectrophotometric assay. The assay is suitable for 96-microwell plate formats to provide a continuous measurement of SAMHD1 activity. The assay is benchmarked with inhibitors of SAMHD1. With a Z-prime value > 0.90, the assay can be used for high-throughput screening of inhibitors for SAMHD1 and characterizing the allosteric or catalytic activity of the new inhibitors.

Graphical Abstract

1. Introduction

The balanced equilibrium of cellular deoxyribose nucleotide triphosphates (dNTPs) is critical in the fidelity of DNA during replication and repair processes.^1-5 Actively dividing mammalian cells, maintain relatively high dNTP levels (1-15 μM) throughout the cell cycle. During the G0/G1-phase, the dNTP pool is low, and as cells commit to replicate DNA during the S-phase, biosynthesis of dNTP increases. Once DNA synthesis is complete, dNTP levels in G2-phase decline with concentrations remaining low during mitosis.^{6, 7} SAMHD1 (EC 3.1.5.1) is one of the key enzymes actively involved in the regulation of intracellular dNTP pools. Its unique dNTPase activity adjusts dNTP levels by the formation of cognate 2'-deoxynucleosides and inorganic triphosphate (PPPi).^{8, 9} Substrates of SAMHD1 include all endogenous dNTPs, their epigenetically modified counterparts, and the triphosphates of nucleoside drugs.^10-16 SAMHD1 is a ubiquitous nuclear enzyme that differentially expresses in response to immune activation, viral infections, and cancerous processes.^17-19 Dividing cells express relatively high levels of SAMHD1 in G1-phase. However, as cells enter S-phase, SAMHD1 levels decrease to maintain sufficient deoxynucleotides for DNA replication. In G2-phase, the expression of SAMHD1 increases, likely to ensure appropriate regulation of dNTP balance, facilitating mitotic entry.²⁰ The dNTPase catalytic activity of SAMHD1 is modulated through post-translational modifications, including phosphorylation,²¹ acetylation,²² and disulfide bond formation,²³ as well as two distinct allosteric sites.^{24, 25} SAMHD1 purifies as an inactive monomer. The dNTPase activity of SAMHD1 requires binding of nucleotide triphosphates at two adjacent allosteric sites to form homotetramers prior to hydrolysis of dNTP substrates.²⁶ During the cell cycle, increased dNTPs in the S-phase,²⁷ elevates the proportion of activated SAMHD1 which persists even after the intracellular dNTP pool has declined below the activation threshold.^{25, 28}

Multiple studies have established that high expression of SAMHD1 compromises efficacy of nucleoside drugs.^{29, 30} For example cytarabine (ara-C) is converted to ara-CTP for its anticancer function but is catabolised by SAMHD1. Likewise, the purine nucleoside phosphorylase inhibitor forodesine (aka Mundesine^®)) causes a lethal accumulation of dGTP, specifically in malignant T cells, with expression of SAMHD1 preventing this effect.¹⁹ Therefore, SAMHD1 inhibitors are of interest to boost the efficacy of anticancer and antiviral nucleoside drugs. Most efforts to characterize or screen inhibitors for SAMHD1 have relied on discontinuous assays, which limit the detailed kinetic analysis of this complex target and reduce the ability to observe allosteric regulation. Previous assays for SMAHD1 have relied on malachite green or xanthine oxidase (colorimetric),^31-33 MDCC-PBP (fluorometric),³⁴ or tritiated [³H] substrates (radiometric).^{24, 25, 31} The malachite green colorimetric method is discontinuous, and single endpoint experiments often compromise initial rate kinetic data. Xanthine oxidase provides a continuous redshift UV assay, and is most effective with 6-oxypurine nucleobases.³⁵ Fluorometric assays suffer from high background and are subject to interference by the frequent inclusion of fluorescence-active agents in inhibitor libraries. Specifically, MDCC-PBP has a high affinity for phosphate; hence, it is sensitive to phosphate contamination and requires the removal of phosphate from the assay reagents.³⁶ Radiometric enzyme assays offer high sensitivity and specificity, but they require special handling and often do not address high-throughput capabilities.³⁷ Given that heterocycles are among the most frequently encountered fragments in compound libraries in high-throughput screening (HTS), our goal is to develop a robust continuous colorimetric assay with absorption wavelengths further away from most heterocyclic structures to minimize false positives.³⁸

The continuous assay described here couples the reaction of purine nucleoside phosphorylase (PNP) and 7-methyl-6-thioguanosine (MESG) to detect inorganic phosphate (Pi). This coupled system monitors the redshift of the substrate (MESG) in the presence of inorganic phosphate (Pi) to the product 2-amino-6-mercapto-7-methylpurine base (AMMP) with an increase in absorbance at 360 nm to permit a robust signal-to-noise ratio $(\Delta {\epsilon }_{360\text{nm}}=11,000{\mathrm{M}}^{-1}{\text{cm}}^{-1})$. ³⁹ dNTP reaction with SAMHD1 liberates PPPi, which is hydrolyzed by efficient phosphatases, Ppx1 $(\text{PPPi}+{\mathrm{H}}_2\mathrm{O}\to \text{Pi}+\text{PPi})$ and PPase, and consecutively converts triphosphate into three inorganic phosphate molecules, thereby tripling the yield of the primary reaction. (Scheme 1).

Scheme 1:

Coupled enzyme assay to detect SAMHD1 activity. Triphosphohydrolase activity with dTTP releases triphosphate (PPPi), converted to three phosphates by the action of yeast exopolyphosphatase (Ppx1) and inorganic pyrophosphatase (PPase). Purine nucleoside phosphorylase (PNP) couples phosphate release to the phosphorolysis of MESG, production of AMMP, and the spectral signal at ΔA = 360 nm.

Two allosteric sites tightly regulate the dNTPase activity of SAMHD1. Upon binding of both NTPs and dNTPs at the allosteric sites (AS1 and AS2), SAMHD1 undergoes a stochastic conformational search at the subunits interface, leading to adjustments in the size and shape of the substrate-binding pocket for dNTPs.⁴⁰ Here, we describe a detailed methodology and show the importance of the homotetramer for the SAMHD1 dNTPase assay to achieve fully activated catalytic rates. Mixing high concentrations of allosteric site activators (GTP and dNTP) in low volumes forms a quasi-stable homotetramer. Dilution steps into assay mixtures provide coupled enzyme activity while keeping the allosteric activating dNTP level far below the $K_M$ for the active site.

The reported Michaelis-Menten constants $(K_{\mathrm{M}})$ of dNTPs with human SAMHD1 range from 50-300 μM for different dNTPs and depends on status of the allosteric sites.^24-27 For kinetic characterization of SAMHD1, the preferred ligands for AS1 and AS2 are GTP and dATP, respectively.

dTTP provides a convenient substrate as the thymidine product does not interfere with the PNP reaction. With this assay, the catalytic efficiency $(k_{\text{cat}}∕K_{\mathrm{M}})$ with dTTP is 5.8 ± 0.30 × 10³ M⁻¹s⁻¹, consistent with previously described values.⁴¹ Rapid dilution of the activated tetramer allows the measurement of kinetic parameters under well-defined activation and catalytic site conditions. The spectral difference in absorption at 360 nm provides a robust signal-to-noise ratio in the case of substrates that have low micromolar $K_M$ values. Initial rates are readily obtained with small amounts of product formation.

This protocol permits SAMHD1 screening in 96-microwell plate format. Wth a ${\mathrm{Z}}^{\prime }$ of 0.96. Earlier reports characterized α,β-imido-dNTP (dNPNPP) as a ${\mathrm{K}}_{\mathrm{i}}=52\pm 2\text{nM}$ inhibitor. The decreased electrophilic properties of α-phosphorus $({\mathrm{P}}^{\alpha })$ eliminates its substrate activity.⁴² The coupled assay described here is used to characterize additional dNPNPPs as competitive inhibitors as proof-of-concept for the method.

2. Materials and Methods

2.1. General Information

pET-28a (+) plasmids containing the codon-optimized open reading frame constructs for Homo sapiens SAMHD1 (64 kDa, UniProt ID: Q9Y3Z3) and Saccharomyces cerevisiae Ppx1 (47 kDa, UniProt ID: P38698) were synthesized by GenScript Biotech (Piscataway, NJ). Homo sapiens PNP (32 kDa, UniProt ID: P00491) (Plasmid # 64076) was obtained through Addgene (Watertown, MA). BL21(DE3) and DH5α competent E. coli were purchased from New England Biolabs (NEB). Superdex gel filtration column (HiLoad 26/60 200 pg) (Cat. No. 28989336) and HisTrap FF protein purification column (Cat. No. 17525501) were purchased from Cytiva. 4–12% Criterion^™ XT Bis-Tris Protein Gel (Cat. No. 3450123 and 3450112) and Precision Plus Protein Dual Color Standards (Cat. No. 1610374) were purchased from Bio-Rad. Luria-Bertani (LB) Lennox powder (Cat. No. BP-1427), and LB Agar Miller granulated (Cat. No. BP-9724) were purchased from Fisher Scientific. Bio-One UV-Star^™ 96-Well Half-area Microplate (Cat. No. 675801) were purchased from Greiner. Miniprep Kit (Cat. No. 27104) was purchased from Qiagen. Pierce^™ Silver Stain kit (Cat. No. 24612) and S.O.C medium (Cat. No. 15544034), Kanamycin sulfate (Cat. No. 611290050) were purchased from Thermo Fisher Scientific. Ampicillin sodium (Cat. No. AA44831), and carbenicillin disodium salt (Cat. No. AC09915) were purchased from Biosynth. 2’-Deoxythymidine-5’-triphosphate (dTTP, 100 mM solution) (Cat. No. N1004L), 2'-Deoxyadenosine-5'-[(α,β)-imido]triphosphate (Cat. No. NU-443), 2'-Deoxycytidine-5'-[(α,β)-imido]triphosphate (Cat. No. NU-439), 2'-Deoxyguanosine-5'-[(α,β)-imido]triphosphate (Cat. No. NU-440), and 2'-Deoxythymidine-5'-[(α,β)-imido]triphosphate (Cat. No. NU-907) were purchased from Jena Bioscience. 7-methyl-6-thioguanosine (MESG) (Cat. No. PR 3790) was purchased from LGC Genomics. 2′-Deoxyadenosine 5′-triphosphate (dATP) (Cat. No. 32566) was purchased from Cayman Chemical Company. Isopropyl β-D-1-thiogalactopyranoside (IPTG) (Cat. No. P1010) was purchased from Ubiquitin-Proteasome Biotechnologies. Inorganic pyrophosphatase (E. coli) (Cat. No. I5907), guanosine 5′-triphosphate (GTP) (Cat. No. G8877), DNase I (bovine pancreas) (Cat. No. DN25), and all materials to prepare buffers were purchased from Millipore Sigma.

2.2. Protein purification

Heterologous overexpression and purification of Saccharomyces cerevisiae Ppx1 and Homo sapiens PNP were performed as previously described (Figure S1).^{43, 44}

2.2.1. Expression and purification of His-tagged SAMHD1

pET-28a(+)-SAMHD1 containing residues 114-626 was modified by adding 37 amino acids, including a His⁶ tag and a thrombin cleavage site at the N-terminal. The plasmid was transformed into BL21(DE3) E. coli using a heat shock procedure, from which single colonies were grown overnight at 37 °C in 10 ml of LB broth containing 50 μg/ml of kanamycin. The overnight culture was used to inoculate 1 L of LB broth containing 50 μg/ml of kanamycin, which was grown at 37 °C to an OD₆₀₀ of 0.6 and induced by the addition of 0.5 mM IPTG. The temperature was reduced to 18 °C for 20 hours, after which cells were harvested by centrifugation.

Pelleted cells were resuspended in lysis buffer containing 50 mM HEPES pH 8.0, 500 mM NaCl, 5 mM MgCl₂, 0.5 mM TCEP, 0.1% Triton X-100, 10 % glycerol, lysed using an EmulsiFlex-C3 cell homogenizer. Insoluble materials were discarded after centrifugation for 20 min at 18,000 rpm, and soluble protein was loaded onto a 5 ml HisTrap FF column. Fractions were eluted from the column using 5 column volumes of 50 mM HEPES pH 8.0, 200 mM NaCl, 5 mM MgCl₂, 0.5 mM TCEP, and stepwise increments of 5 to 500 mM imidazole. The desired enzyme fractions were pooled and concentrated to 5 mL, further purified by size-exclusion chromatography (SEC) on an ÄKTA pure system using a Superdex gel filtration column (HiLoad 26/60 200 pg) with the buffer 50 mM HEPES pH 8.0, 200 mM NaCl, 5 mM MgCl₂, 0.5 mM TCEP at 4 °C. The desired fractions were pooled and concentrated to 5.0 mg/mL using a 10 kDa MW cutoff Amicon Ultra concentrator at 4 °C, supplemented with 10% glycerol, and stored at −80 °C (Figure S2). A yield of 15 mg/L of media was obtained. SAMHD1 concentration was determined using an extinction coefficient ${\epsilon }_{280}=61.77{\text{mM}}^{-1}{\text{cm}}^{-1}$

Plasmids used for expression of SAMHD1 and Ppx1 will be deposited to Addgene to permit ready access.

2.3. Optimization of assay condition

Assays were conducted in triplicate, subtracting the background absorbance using an intraplate control containing no substrate. The initial rates were converted from ΔA₃₆₀ to μM dTTP consumed per second as described below. All initial rates were monitored between 200 - 1200s reads were performed at 25 °C on a SpectraMax M5 unless otherwise stated. All data fitting of enzyme kinetics was performed using GraphPad Prism 10.4.1.

2.3.1. The experimental extinction coefficient of AMMP at 360 nm

Standard curves with varying concentrations of MESG with and without PNP and a saturating concentration of sodium phosphate (1 M) were generated to determine the apparent extinction coefficient $({{\epsilon }_{360}}^{\text{app}})$ to account for the shorter path length in the well of the half-area plate. HPLC confirmed the complete conversion of MESG to AMMP with PNP in the presence of Pi (data not shown) (Figure S3A). Upon reaction of MESG and Pi with PNP, absorbance changes as a function of AMMP concentration multiplied by the difference in the apparent molar absorptivities of the two species $({\epsilon }_1\text{-}{\epsilon }_2)$ (Figure S3B). The enzymatic rate from the linear fit of the initial rate data was reported as consumption of dTTP in μM s⁻¹, accounting for the 3-fold boost in signal due to the generation of 3 Pi per dTTP. (Figure S3C)

2.3.2. Optimizing SAMHD1 concentration

SAMHD1 was varied from 1000 nM to 40 nM and activated in an assay mixture containing 50 mM HEPES pH 8.0, GTP, dATP, dTTP, MESG, MgCl₂, TCEP, Ppx1, and PPase were 10 μM, 500 nM, 500 μM, 500 μM, 5 mM, 500 μM, 10 nM, and 10 nM, respectively. At these enzyme concentrations, consumption rates of dTTP were monitored (Figure 1).

Figure 1.

Optimization of assay condition for SAMHD1. A – Traces of dNTPase activity of SAMHD1 titration, 1 μM (cyan), 500 nM (coral), 250 nM (blue), 100 nM (green), 40 nM (magenta) at saturating concentration of dTTP (500 μM). Data points and error bars represent mean ± SD from three independent experiments. B – Consumption rate of dTTP at varied concentrations of SAMHD1in 100 μL reaction volume. Data points and error bars represent mean ± SD from three independent experiments.

2.4. Allosteric activation of SAMHD1

2.4.1. Assembly of tetramer in SAMHD1

Purified SAMHD1 (2.5 μL of 40 μM) was mixed with combinations of GTP (2.5 μL of 4.0 mM) and dATP (5.0 μL of 100 μM dATP). The mixtures were diluted 100-fold with or without dTTP (100 μM) into 50 mM HEPES buffer at pH 8.0, followed by crosslinking with glutaraldehyde (10 μL, 5 M) for 15 min and quenching with Tris buffer (100 μL, 1 M). 35 μL of the mixtures were separated by the SDS-gel. (Figure 4C&D). Images were taken by Bio-Rad ChemiDoc XRS+ System. Using Image Lab 6.1 software, the band intensities were measured by integrating the pixel intensity of each lane. The band intensity plots were overlaid and plotted together to visualize the relative differences between the time points. All measurements were conducted under unsaturated conditions and normalized to the zero time point.

Figure 4.

Stability of tetramer complex. A – Silver staining of SAMHD1 oligomeric state promoted with different nucleotides. The gel is representative of three independent experiments. B – Relative initial rate relative to the time zero sample after 10-fold dilution of the activated complex of SAMHD1 reacting with dTTP (100 μM) at different time points. Data points and error bars represent mean ± SD from three independent experiments, dashed lines represent the fits to eq.1.C – Silver staining of the 10-fold diluted complex at different time points. The gel is representative of three independent experiments. D – Dissociation of SAMHD1 complex quantified from pixel intensity of each lane normalized to the zero time point in C.

2.4.2. Activation of SAMHD1

The hydrolysis of dTTP was evaluated as the first allosteric site (AS1) was filled with GTP concentrations (100 μM – 10 mM, ratios to SAMHD1 were 2.5 - 250) with 100 nM SAMHD1, 100 μM dATP, and varying dTTP between 8.4 - 500 μM. dTTP hydrolysis was evaluated as the second allosteric site (AS2) was filled with dATP (25 −1000 μM, ratios to SAMHD1 were 1.25 - 50) with 100 nM SAMHD1, 4 mM GTP, and varying dTTP between 8.4 - 500 μM.

2.5. Steady-state kinetics of SAMHD1

Sequential activation of SAMHD₁ followed the procedure outlined in Scheme 2, where the specific reagents were added at each individual step to activate and/or provide coupled assay components. In step one (pre-activation step), 2.5 μL purified SAMHD₁ at a concentration of 40 μM was mixed with 2.5 μL of 4.0 mM GTP to give 5.0 μL SAMHD1 and GTP at 20 μM and 2.0 mM μM, respectively. In step two (tetramerization step), the SAMHD1/GTP mixture was mixed with 5.0 μL of 100 μM dATP to give to give 10 μL of SAMHD1, GTP, and dATP at 10 μM, 1 mM, and 50 μM, respectively. In step three (first dilution step), the SAMHD1/GTP/dATP mixture was diluted to 100 μL with the reaction buffer. In step four (assay cocktail step), the mixture was mixed with 800 μL of the assay mixture (625 μM MESG, 375 μM PNP, 12.5 nM Ppx1, and 12.5 nM PPase). This mixture was distributed at 90 μL together with 10 μL of dTTP to give the desired dTTP concentration. Reaction mixtures were monitored for 20 min at 360 nm. The final concentrations in the assay mixtures for SAMHD1, GTP, dATP, MESG, PNP, Ppx1, and PPase were 100 nM, 10 μM, 500 nM, 500 μM, 300 μM, 10 nM, and 10 nM, respectively, with dTTP varied between 8.4 - 500 μM. Throughout the process the buffer was 50 mM HEPES pH 8.0, 200 mM NaCl, 5 mM MgCl₂, 0.5 mM TCEP, and 10% glycerol. Initial reaction rates were plotted as a function of substrate concentration (dTTP) and fitted to the Michaelis-Menten equation to obtain the $k_{\text{cat}}$ and $K_{\mathrm{M}}$ values (Eq. 1),

$$V=(V_{\mathrm{Max}}[\mathrm{S}])∕(K_{\mathrm{M}}+[\mathrm{S}])$$ (eq. 1)

where $V_{\mathrm{Max}}$ is the maximal velocity, $[\mathrm{S}]$ is the concentration of the variable substate. $k_{\text{cat}}$ was determined by dividing the constant $V_{\mathrm{Max}}$ by the SAMHD1 concentration.

Scheme 2:

Schematic of the ordered assembly of catalytically active SAMHD1oligomer. Activation of allosteric site 1 (AS1) with GTP, followed by activation of allosteric site 2 (AS2) with dATP forms a homotetramer. First dilution with buffer was used to lower the concentration of residual dATP from the tetramerization step, then second dilution into the assay mixture (MESG, Ppx1, PPase, PNP), followed by the addition of the substrate (dTTP) initiated the catalytic reaction.

2.6. Stability of tetramer complex

2.6.1. Chemical cross-liking of the tetramer

Purified SAMHD1 (2.5 μL of 40 μM) was mixed with 2.5 μL of 4.0 mM GTP, mixed with 5.0 μL of 100 μM dATP, diluted 10-fold with 50 mM HEPES buffer at pH 8, and incubated at room temperature. Timed samples were diluted with 900 uL of 50 mM HEPES buffer at pH 8.0 and crosslinked with glutaraldehyde (10 μL, 5 M) for 15 min and quenching with Tris buffer (100 μL, 1M). 35 μL of the mixtures were separated by the SDS-gel. (Figure 4C&D).

2.6.2. Enzymatic rate of the tetramer complex

Purified SAMHD₁ (2.5 μL of 40 μM) was mixed with 2.5 μL of 4.0 mM GTP, followed by 5.0 μL of 100 μM dATP and a 10-fold dilution with 50 mM HEPES buffer. At timed intervals (0 to 120 mins), the solution was mixed with 800 μL of assay mixture (625 μM MESG , 375 μM PNP, 12.5 nM Ppx1, 12.5 nM PPase). Reactions were initiated with 100 μL dTTP (100 μM) and monitored for 20 min at 360 nm.

2.7. Inhibition Assay

2.7.1. SAMHD1 Inhibition Assay

Purified SAMHD₁ (2.5 μL of 40 μM) was mixed with 2.5 μL of 4.0 mM GTP, followed by 5.0 μL of 100 μM dATP and a 10-fold dilution with 50 mM HEPES buffer. The enzyme mixture was added to 800 μL of the assay cocktail (625 μM MESG, 375 μM PNP, 12.5 nM Ppx1, 12.5 nM PPase). The inhibitor (dNPNPP) was aliquoted into the dTTP substrate. The final concentrations in the assay mixture for SAMHD1, GTP, dATP, dTTP, MESG, MgCl₂, PNP, Ppx1, and PPase were 100 nM, 10 μM, 500 nM, 100 μM, 500 μM, 5 mM, 300 μM, 10 nM, and 10 nM, respectively. To 90 μL of the assay mixture in each well, reactions were initiated with 10 μL mixture of inhibitor and dTTP at the desired concentrations, and the reaction mixture was monitored for 20 min at 360 nm. Corrected Reaction rates were calculated as in section 2.3.1. IC₅₀ is calculated from inhibited rates relative to an uninhibited control (Eq. 2).

$$\text{Relative Velocity}=((\text{Top}-\text{Bottom})∕(1+([\mathrm{I}]∕{\text{IC}}_{50})))$$ (eq. 2)

where top and bottom are defined as the maximal and minimal reaction rates for the titration, respectively. Inhibition constant $(K_{\mathrm{i}})$ values were determined from IC₅₀ values using the Cheng-Prusoff equation (Eq. 3). ⁴⁵

$$K_{\mathrm{i}}=({\text{IC}}_{50}∕(1+([\mathrm{S}]∕K_{\mathrm{M}}))$$ (eq. 3)

where $[\mathrm{S}]$ and $K_{\mathrm{M}}$ are the substrate concentration and Michaelis constant for the substrate, respectively.

2.7.2. Tight-binding inhibitors of SAMHD1

Similar to section 2.7.1, all reganets were mixed, and the substrate (dTTP) concentration was increased to 500 μM. Tight-binding inhibition was calculated through the Morrison equation relative to an uninhibited control (Eq. 4 & 5).

$$V=V_0(1\text{-}(((({[\mathrm{E}]}_{\mathrm{t}}+[\mathrm{I}]+K_{\mathrm{i}}^{\ast })\text{-}({(({[\mathrm{E}]}_{\mathrm{t}}+{[\mathrm{I}]}_{\mathrm{t}}+K_{\mathrm{i}}^{\ast })}^{\wedge 2})\text{-}{4{[\mathrm{E}]}_{\mathrm{t}}{[\mathrm{I}]}_{\mathrm{t}})}^{\wedge 0.5}))∕(2{[\mathrm{E}]}_{\mathrm{t}})))$$ (eq. 4)

$$K_{\mathrm{i}}^{\ast }=(K_{\mathrm{i}}(1+([\mathrm{S}]∕K_{\mathrm{M}})))$$ (eq. 5)

where $V$, $V_0$, ${[\mathrm{E}]}_{\mathrm{t}}$, ${[\mathrm{I}]}_{\mathrm{t}}$, and $K_{\mathrm{i}}^{\ast }$ are the observed rate, uninhibited rate, total enzyme concentration, total inhibitor concentration, and apparent inhibition constant in the presence of substrate, respectively.

2.7.3. Lineweaver-Burk plot and mechanism of inhibition

Purified SAMHD₁ (2.5 μL of 40 μM) was mixed with 2.5 μL of 4.0 mM GTP, followed by 5.0 μL of 100 μM dATP and a 10-fold dilution with 50 mM HEPES buffer. The enzyme mixture was added to 800 μL of the assay cocktail (625 μM MESG, 375 μM PNP, 12.5 nM Ppx1, 12.5 nM PPase). The varied inhibitor concentration of dGPNPP (up to 4 μM) was aliquoted into the varied concentration of dTTP (0.625 - 10 mM). The final concentrations in the assay mixture for SAMHD1, GTP, dATP, MESG, MgCl₂, PNP, Ppx1, and PPase were 100 nM, 10 μM, 500 nM, 500 μM, 5 mM, 300 μM, 10 nM, and 10 nM, respectively. To 90 μL of the assay mixture in each well, reactions were initiated with 10 μL mixture of dGPNPP inhibitor and dTTP at the desired concentrations, and the reactions were monitored. Initial rates of SAMHD1 with varied concentrations of dTTP and dGPNPP were plotted to obtain the inhibition pattern for dGPNPP. Data was fit to Eq. 6 for competitive inhibition.

$$V=(V_{\mathrm{Max}}[\mathrm{S}])∕(K_{\mathrm{M}}(1+1∕K_{\text{is}})+[\mathrm{S}])$$ (eq. 6)

where $V$, $V_{\mathrm{Max}}$, $[\mathrm{S}]$, $K_{\mathrm{M}}$, and $K_{\text{is}}$ are observed rate, uninhibited rate, substrate concentration, Michaelis constant for the substrate, and inhibition constant for the inhibitor, respectively.

2.7.4. Z prime ($Z^{\prime }$) factor

To evaluate the assay robustness and high-throughput screening (HTS) suitability, the screening window coefficient (${\mathrm{Z}}^{\prime }$) was determined from initial rates in the first 20 min of assays, comparing twenty-four nosubstrate conditions for negative control (neg) and twenty-four 100 μM dTTP conditions as positive control (pos), and the ${\mathrm{Z}}^{\prime }$ equation (Eq. 7).⁴⁶

$${\mathrm{Z}}^{\prime }=1-((3{\sigma }_{\text{pos}}-3{\sigma }_{\text{neg}})∕(\mid {\mu }_{\text{pos}}\mid +\mid {\mu }_{\text{neg}}\mid ))$$ (eq. 7)

Where $\mu$ is the mean value and $\sigma$ is the standard deviation of three conditions. ($3\sigma$, corresponds to a 99.73% confidence interval). ${\mathrm{Z}}^{\prime }$-factor is a commonly cited statistic in HTS that evaluates the dynamic range of the signal and data variability of positive and negative controls, and describes the extent to which an assay can distinguish true signal from background. A ${\mathrm{Z}}^{\prime }$ value between 0.5 and 1.0 indicates an excellent assay with low variability and a broad separation band, suggesting high reproducibility and reliability for screening purposes (Figure S5).

2.7.5. SAMHD1 Inhibition Assay Coupled Enzyme Activity Controls

Purified SAMHD₁ (2.5 μL of 40 μM) was mixed with 2.5 μL of 4.0 mM GTP, followed by 5.0 μL of 100 μM dATP and a 10-fold dilution with 50 mM HEPES buffer. The enzyme mixture was added to 800 μL of the assay cocktail (625 μM MESG, 375 μM PNP, 12.5 nM Ppx1, 12.5 nM PPase). Approximately a concentration of $10\times K_{\mathrm{i}}$ of each inhibitor (dNPNPP) was added into 1 mM dTTP substrate. The final concentrations in the assay mixture for SAMHD1, GTP, dATP, MESG, MgCl₂, PNP, Ppx1, and PPase were 100 nM, 10 μM, 500 nM, 500 μM, 5 mM, 300 μM, 10 nM, and 10 nM, respectively. To 90 μL of the assay mixture in each well, reactions were initiated with 10 μL mixture of inhibitor and dTTP and the reaction mixture was monitored for 20 min at 360 nm. Enzymatic rates were calculated as in section 2.3.1.

3. Result and discussion

3.1. Optimization of Assay Conditions

3.1.1. Principals of the coupled enzyme assay

For SAMHD1 dNTPase, the hydrolytic reaction $(k_1)$ is irreversible, and providing high levels of the substrate (dTTP) provides apparent zero-order initial rate kinetics (Scheme 3). In other words, The concentration of dTTP does not change significantly during initial rate measurements. Each of the coupled enzyme reactions (Ppx1, PPase, and PNP) is also required to be irreversible and a first-order reaction with respect to its own substrates. The concentration of the PPPi as SAMHD1 product must be kept to a near-zero concentration to achieve valid initial rate observations. While the concentration of substrates for the coupling enzymes must be much lower than their $K_{\mathrm{M}}$, it is also required to have saturating concentrations of the second substrates for the coupled enzymes to achieve these goals (e.g., H₂O for Ppx1 and PPase and MESG for PNP).⁴⁷ The rates $k_2\text{-}{k}_4$ are large enough to keep the concentrations of the intermediates near zero by the coupled enzymes. Low concentrations also serve to keep all of the reactions functionally irreversible.⁴⁸ The assay conditions gave a linear response to SAMHD1 concentration with a near-saturating concentration of dTTP (500 μM). Under these conditions, initial rates were linear over the first 1200 sec, even with 250 nM SAMHD1 (Figure 1A).

Scheme 3:

Rate constants of SAMHD1 and the coupled enzymes in the assay. $k_1$ (SAMHD1), $k_2$ (Ppx1), $k_3$ (pyrophosphatase) and $k_4$ (PNP).

3.2. Allosteric activation of SAMHD1

The apo-form of human SAMHD1 exists in an equilibrium of monomeric and dimeric states. The addition of guanine-containing nucleotides activates at the first allosteric site (AS1), stabilizing formation of the dimer. Subsequently, any dNTPs bind to the second nonspecific allosteric site (AS2), causing assembly of the catalytically active homotetramer. In mammalian cells, NTPs are present at higher concentrations than dNTPs.⁴⁹ In macrophages, GTP is almost 1000-fold the dGTP content.²⁵ When the AS2 activator (dATP) is constant, the dTTP hydrolysis increases as GTP concentration increases. Activation is primarily in $k_{\text{cat}}$ with little change in the $K_{\mathrm{M}}$ of dTTP, consistent with V-type allosteric behavior. (Figure 2A & B),

Figure 2.

Activation of SAMHD1 at two allosteric sites. A – GTP concentration dependence at allosteric site 1 (AS1) effect on dTTP hydrolysis rate. B – V-type allostery at different ratios of GTP:SAMHD1 as depicted by increasing rate with higher GTP presence. C – dATP concentration dependence at allosteric site 2 (AS2) effect on dTTP hydrolysis rate. D – V-type allostery at different ratios of dATP:SAMHD1 as depicted by increasing rate with higher dATP presence. Data points and error bars represent mean ± SD from three independent experiments, solid lines represent the fits to eq.1.

The AS2 prefers dNTPs, and occupancy of this site supports the dimer-dimer interactions required for homotetramer formation. The AS2 is known to have a preference for dGTP > dATP > dCTP > dTTP.⁵⁰ Building upon this preference, we devised a sequential approach to activate SAMHD1. In the initial pre-activation step, a dimer is formed by adding GTP to SAMHD1 at 2 mM (ratio of 100 to SAMHD1). Subsequently, because of its high affinity for AS2, dATP was introduced to the mixture at 50 μM (ratio of 5 to SAMHD1). The SAMHD1 tetramer remains stable after filling AS2. Filling the AS2 with dATP increased reaction rates consistent with V-type allosteric behavior. The sub-micromolar concentration of dATP used to activate at AS2 showed no apparent interference with dTTP binding at the catalytic site (Figure 2C&D).

3.3. Substrate characterization for SAMHD1

Activated SAMHD1 is reported to exhibit active site selectivity with dCTP ~ dGTP > dTTP > dATP.⁵¹ The SAMHD1 HD domain contains a bimetallic Fe²⁺-Mg²⁺ center, which facilitates the activation of the nucleophilic water molecule, with another Mg²⁺ aiding in binding of the dNTP. The activated water attacks ${\mathrm{P}}^{\alpha }$ of the dNTP, leading to the cleavage of the ${\mathrm{P}}^{\alpha }\text{-}\mathrm{O}{5}^{’}$ phosphoester bond and the subsequent release of 2'-deoxynucleoside and triphosphate products.^{42, 52, 53} This assay converts triphosphate to Pi via Ppx1 and PPase, and catalyzes phosphorolysis of MESG by PNP to generate ribose-1-phosphate and the AMMP base, resulting in an increase in absorbance at 360 nm $(\Delta {\epsilon }_{360\text{nm}}=11,000{\mathrm{M}}^{-1}{\text{cm}}^{-1})$. We use dTTP as the substrate because deoxythymidine does not compete with MESG for reaction with human PNP.⁴⁷ The assay concentrations of MESG, PNP, Ppx1 and PPase were optimized as 500 μM, 300 μM, 10 nM and 10 nM, respectively. PPase and Ppx1 act to convert triphosphate into three molecules of inorganic phosphate (Scheme 1). Assay volume was 100 μL, and conditions were optimized for 20 min initial rate measurements at 25 °C using a SpectraMax M5 microplate reader at 360 nm. Initial rates were converted from ΔA₃₆₀ to consumption of dTTP μM s⁻¹ divided by SAMHD1 concentration in μM (Figure 3 & Figure S4).

Figure 3.

Steady-state kinetics of SAMHD1 with dTTP. A – Traces of dNTPase reaction against dTTP at varied concentrations (8.4 μM to 500 μM) with 100 nM SAMHD1. Data points are representative of three independent experiments. B – Michaelis-Menten plot of dTTP substrate activity, initial rates are reported after dividing by enzyme concentration. Data points and error bars represent mean ± SD from three independent experiments, solid lines represent the fits to eq.1.

3.4. Stability of tetramer complex

The oligomeric state of SAMHD1 was evaluated by crosslinking, gel electrophoresis, and silver staining. Consistent with literature, guanine nucleotides are required for tetramerization (lanes 1, 2, and 4, Figure 4A). Thus, in the absence of GTP, no tetramer formed (lanes 5-8, Figure 4A). Activation studies demonstrate that dNTPs, in this case, dATP, promote the formation of homotetramer after GTP treatment without the presence of dTTP substrate (lane 2, Figure 4A). However, in the absence of dATP, the homotetramers do not form upon dilution into buffer. (lane 3, Figure 4A). Treatment with GTP alone followed by 100-fold dilution into substrate dTTP is insufficient to complete tetramer formation (lane 4, Figure 4A). Therefore, the presence of AS2 activator prior to reaction initiation is important to fully activate SAMHD1. Maintenance of fully activated SAMHD1 into the assay is accomplished by adding substrate dTTP to the fully homotetramer complex. Adding SAMHD1 tetramer in a concentrated solution to dTTP substrate keeps the enzyme in the homotetrameric form during the initial rate measurements.

Tetramer stability of SAMHD1 was measured as a function of time, and the diluted homotetramer complex exhibits a half-life of > 80 mins (Figure 4C&D). The quantified densities in Figure 4D correspond to the areas of interest assigned to each band in Figure 4C. As the diluted tetramer complex dissociates over time, the initial rate declines correspondingly (Figure 4B). Under these conditions, the tetramer population decreases to 20% over 2 hr, likely due to dNTPase activity at the 10-fold dilution stage, causing loss of the dATP activator. With this reduced tetramer population, the addition of dTTP (100 μM) gives 33% of maximal activity. Under these conditions, the addition of dTTP substrate does not reestablish the complete population of SAMHD1 homotetramer. However, in the absence of dATP as the AS2 activator, the addition of dTTP into the assay sample partly re-establishes the active homotetramer (lane 4, Figure 4A).

3.5. Inhibition assay and assay validation

3.5.1. Inhibition of SAMHD1 by dNPNPP

Inhibitor screens for SAMHD1 are most sensitive near the $K_{\mathrm{M}}$ of dTTP (102 ± 9.4 μM) and require a robust signal-to-noise. An assay condition of 100 μM dTTP gave a robust Z prime value of > 0.90, indicating a high-quality assay to measure changes in dTTP triphosphate hydrolysis activity (Figure S5). Inhibition of SAMHD1 with dNPNPPs can serve to benchmark the assays (Figure 5A).⁴² Relative activities of inhibited and uninhibited controls were fitted to initial rate curves to determine IC₅₀ values (Figure 5B). dGPNPP and dCPNPP are effective inhibitors of dTTP triphosphate hydrolysis, giving IC₅₀ of 57 ± 2.7 nM, and 83 ± 8.6 nM, respectively, and $K_{\mathrm{i}}$ values of 31 ± 2.3 nM and 45 ± 5.3 nM (eq. 3). dAPNPP and dTPNPP gave higher $K_{\mathrm{i}}$ values as 207 ± 21 nM and 145 ± 23 nM, respectively. The inhibition profile is similar to previous literature on the inhibition of SAMHD1 activity.⁴²

Figure 5.

A – Chemical structure of dNTP analogs. B –Inhibition of 100 nM SAMHD1 with dNTP analogs. Initial rates relative to uninhibited controls is plotted against the log of inhibitor concentration. C – Inhibition constant $(K_i)$ for each dNTP analog is calculated from the IC₅₀ in 100 nM SAMHD1 and 100 μM dTTP. dTTP $K_M$ is 102 ± 9.4 μM. Data points and error bars represent mean ± SD from three independent esperiments, and solid lines represent either fitting to eq. 1.

3.5.2. Assay for tight binding inhibition of SAMHD1 by dCPNPP and dGPNPP

The Morrison equation models the fraction of the active enzyme available for binding to the inhibitor when inhibitor and enzyme concentrations are similar. Using higher substrate concentration pushes the enzyme toward zero-order kinetics with respect to substrate, so the velocity mainly depends on enzyme and inhibitor binding and not the substrate concentration. We increased the concentration of dTTP to $5\times K_{\mathrm{M}}=500\mu \mathrm{M}$, to reach $V_0$, and using the Morrison equation (Eq. 4 & 5), the inhibition constants for dCPNPP and dGPNPP were $K_{\mathrm{i}}=67\pm 8.0\text{nM}$ and $K_{\mathrm{i}}=25\pm 7.0\text{nM}$, respectively, relatively tight-binding inhibitors. (Figure 6A-C).

Figure 6.

Binding and mode of inhibition of dNPNPPs. A – Activity of SAMHD1 as a function of dCPNPP concentration at low (100 μM, blue) and high concentration (500 μM, red) of dTTP. B – Activity of SAMHD1 as a function of dGPNPP concentration at low (100 μM, blue) and high concentration (500 μM, red) of dTTP. Initial rates relative to uninhibited controls plotted against the log of inhibitor concentration, low substrate concentration was fitted to Eq. 2 & 3, and high substrate concentration was fitted to Eq. 4 & 5. C – Change in $K_i$ by increasing substrate (dTTP) concentration from 100 μM to 500 μM. D – Initial rates of SAMHD1 with varied concentrations of dTTP as the substrate and dGPNPP as the inhibitor. E – Lineweaver-Burke plot transformed from C demonstrating competitive inhibition by dGPNPP with $K_{\text{is}}=53\pm 7.0\text{nM}$. Data points and error bars represent mean ± SD from three independent experiments, and solid lines represent either fitting to Eq. 1 or a global fit of all data points to Eq. 5 for (C), and (D), respectively.

3.5.3. Mode of inhibition of SAMHD1 by dGPNPP

Previous kinetic and structural studies have shown that non-hydrolyzable dNPNPPs can act as competitive inhibitors of SAMHD1.⁴² Initial rate data of SAMHD1 inhibition with dGPNPP at varied concentrations of dTTP gave a double-reciprocal Lineweaver-Burk plot consistent with a competitive inhibition mechanism and a global inhibition constant of $K_{\text{is}}=53\pm 7.0\text{nM}$ (Figure 6 D,E).

3.5.4. Counter-assay validation for Ppx1, PPase and PNP

Assay validation experiments were designed to determine if inhibitors (eg. dTPNPP) interfere with the coupled enzyme system. Single concentrations of each dNPNPP near $K_{\mathrm{i}}$ were added together with 100 μM dTTP. Increasing coupled enzyme concentrations causes no change in the rates of dTTP conversion through the coupled enzyme reactions, therefore dNPNPPs do not inhibit the coupled enzymes (Figure S6).

4. Conclusions

dNTPase activity of SAMHD1 relies on the combination of distinct allosteric and catalytic binding sites. Activation of SAMHD1 to its tetrameric state requires the formation of the homotetramer. Guanine-containing nucleotides (GTP), by binding to the AS1, shift the equilibrium towards dimers through a hydrogen-bond network of guanine and electrostatic interactions of the triphosphate group to the opposing protomer.⁵⁴ Next, any dNTPs that can bind to AS2 at intersubunit interfaces shift the equilibria towards the homotetramer. Cryo-EM imaging has been reported to capture high-resolution images of these conformational states.⁴⁰

Ligands in allosteric sites can function as V-type or K-type activators or inhibitors.⁵⁵ SAMHD1 allosteric effectors promote the formation of the homotetramer, induce the closed conformation in the HD domain, and activate the catalytic activity.⁵⁶ Certain molecules, such as dNTP substrates serve as both allosteric effectors and substrates of SAMHD1. Therefore, their mode of action can be convoluted unless controlled by sequential activation.^{42, 57}

The kinetic analysis described here used SAMHD1 truncated to its HD and C-terminal domains (CtD). Truncation of the SAM domain has no effect on the dNTPase activity of human SAMHD1.²⁸ Our assay quantitates catalytic activity by coupling the production of PPPi to inorganic phosphate with Ppx1 and PPase. The phosphatases provide efficient production of Pi, to be coupled to the well-known PNP enzyme. Triphosphate processing provides a threefold spectral readout, permitting active quantitation even when the substrate is in the low micromolar range. Through the selection of coupled enzymes, activators and substrates, it is possible to optimize assay conditions to avoid complications from using high concentrations substrates and activators with multiple actions at substrate and allosteric sites. For one example, the family I pyrophosphatase enzymes have shown slow promiscuous phosphohydrolase activity against high concentrations of organic triphosphates.⁵⁸

Specific activations of the AS1 and AS2 are accomplished by timed dilutions to permit measurement of $K_{\mathrm{M}}$ and $k_{\text{cat}}$ values for SAMHD1 in specifically activated states. We have developed conditions where the homotetramer is relatively stable in the assay condition, and the subsequent addition of the substrate shows sustained initial rates for dNTP hydrolysis. This assay uses specific dilutions to separate the binding events of dNTP to the allosteric site from the subsequent interaction of dNTP with the active site of SAMHD1. Thus, it provides a reliable method to distinguish between modulators of each individual step. We anticipate that this assay can also define the impact of post-translational SAMHD1 modifications, including phosphorylation of Thr592 and oxidation of Cys522, on each step of SAMHD1 activity.^{21, 23}

In cells, activation of SAMHD1 is driven by high levels of dNTPs during S-phase, and this meets the needs to generate the active homotetramer. Activation of the concentrated enzyme before dilution for the assay forms the homotetramer, and it remains stable within in initial rate assay timeframe, desirable for screening and characterizing inhibitors. Without full activation, assays have the potential to reflect dual targeting of allosteric and active sites. Inhibitors identified without an activation step may act to disrupt the oligomerization of SAMHD1.⁵⁹ For physiological targeting of SAMHD1 catalytic activity, it may be possible to generate allosteric site or catalytic site inhibitory agents. For example inhibitors binding at allosteric sites could prevent tetramer formation or could bind at interfaces to prevent oligomer formation or could act at catalytic sites. The assay described here is specific for catalytic site screens.

The use of allosteric effectors as inhibitors of SAMHD1 remains a possibility. One report indicated that tetramers exist as inactive intermediates until allosteric sites can be occupied. dNTPase activity relies on the equilibrium of the inactive and active tetrameric state, which is modulated by the binding and release of allosteric dNTP ligands on the homotetramer.⁵⁰

Inhibitory data with dNPNPPs serves as proof-of-concept for inhibitor analysis. Since tight-binding inhibitors bind so strongly to the target enzyme, they significantly deplete the free inhibitor pool, and they violate the assumption behind classic Michaelis-Menten kinetics. In this situation, solving the quadratic equation of binding equilibrium gives an accurate estimate of $K_{\mathrm{i}}$. The Lineweaver–Burke analysis demonstrated competitive inhibition by the tight-binding inhibitor (dGPNPP). The counter-assay validation by increasing coupled enzyme concentrations provides support for the specificity of the inhibitor against SAMHD1 target without inhibiting the coupled enzyme. Despite the relatively powerful inhibition by anionic α,β-imidotriphosphates,their biological application is limited by cell permeability. Inhibition of SAMHD1 changes nucleotide metabolism by preventing the triphosphohydrolytic catabolism of dNTPs.

In one application, inhibition of SAMHD1 enhances anti-cancer agents including cytarabine and decitabine which are administered as the nucleoside prodrugs and converted to triphosphates as their active forms. Active SAMHD1 has the potential to degrade and decrease the efficacy of these therapies.^{29, 30} In the case of nucleoside reverse transcriptase inhibitors (NRTIs) such as zidovudine, tenofovir, and lamivudine, that structurally differ from endogenous dNTPs, SAMHD1 inhibition does not affect their triphosphate levels directly.^{60, 61} In anticancer treatment, the active triphosphate forms of ara-cytarabine (ara-C), and decitabine are substrates for SAMHD1. Therefore, SAMHD1 inhibition allows increased intracellular level of the active triphosphate nucleoside drugs, enhancing their effects.^12-16 Elevated dNTP DNA precursor levels may outcompete nucleoside analogs and potentially diminish their anticancer and antiviral effects. This dual effect accentuates the need to examine the outcome of SAMHD1 inhibition on both the triphosphate form of nucleoside drugs and the levels of endogenous dNTPs, ensuring the development of optimized therapeutic strategies.

SAMHD1 is highly expressed in immune cells of the myeloid lineage and maintains dNTP pools at very low concentrations (20-70 nM).⁶ However SAMHD1 expression is different in the lymphoid lineage, depending on the activation state of T cells, and their need for DNA synthesis during proliferation.⁶² The anti-T cell leukemia agent forodesine, impairs malignant T-cell proliferation by preventing catabolism of the nucleoside deoxyguanosine (dG), thereby inducing a toxic accumulation of dGTP specifically in rapidly dividing T cells.⁶³ Nevertheless, high expression of SAMHD1 in leukemic T cells counteracts the increase in dGTP, making the drug effective only in patients with low SAMHD1 expression.¹⁹ Therefore, SAMHD1 inhibition is expected to increase the efficacy of the PNP inhibitor in all patients with T-cell leukemia, thus increasing the response to forodesine.⁶⁴

The SAMHD1 assay described here is applicable to any enzyme generating (poly)phosphate as a product. This assay has been developed with a 96-microwell plate format and showed a high $Z^{\prime }$ factor that underlines its potential to be adapted for 384-microwell plates, amenable to high-throughput screening. We envision that the multistep process of activation and controlled dilution steps will differentiate the modes of action of similar enzymes, facilitating the identification of compounds as activators, substrates, or inhibitors.

Highlights.

• The complex allosteric regulation mechanism of SAMHD1 and a reaction that lacks a direct spectroscopic signal make its kinetic analysis and inhibitor discovery challenging

• The assay uses a sequential assembly to generate the active tetrameric form of the enzyme

• Two phosphatases convert inorganic triphosphate to inorganic phosphate, which reacts with MESG and purine nucleoside phosphorylase to provide a continuous spectrophotometric assay

• High Z-prime value allowed testing inhibitors against SAMHD1 and showed their tight binding and competitive mode of action

Abbreviations

AMMP

2-amino-6-mercapto-7-methylpurine base

AS

Allosteric site

dNTP

2-Deoxy ribose-5-triphosphate

MESG

7-methyl-6-thioguanosine

PPase

Pyrophosphatase

Ppx1

exopolyphosphatase

PNP

purine nucleoside phosphate

PPPi

inorganic triphopshate

Data availability

Data will be made available on request.