An Investigation of the Redox Regulation of the Deoxynucleotide Triphosphohydrolase Samhd1 and Its Role in Nucleotide Homeostasis

Home , Deoxyribonucleotide, DGTPase

AN INVESTIGATION OF THE REDOX REGULATION OF THE DEOXYNUCLEOTIDE TRIPHOSPHOHYDROLASE SAMHD1 AND ITS ROLE IN NUCLEOTIDE HOMEOSTASIS

CHRISTOPHER H. MAUNEY

A Dissertation Submitted to the Graduate Faculty of

WAKE FOREST UNIVERSITY GRADUATE SCHOOL OF ARTS AND SCIENCES

In Partial Fulfillment of the Requirements

For the Degree of

DOCTOR OF PHILOSOPHY

BIOCHEMISTRY AND MOLECULAR BIOLOGY

May 2017

Winston-Salem, North Carolina

Approved by

Thomas Hollis, PhD - Advisor

Rebecca Alexander, PhD – Committee Chair

W. Todd Lowther, PhD

Doug Lyles, PhD

David Ornelles, PhD

Fred Perrino, PhD

DEDICATION

This work is dedicated to my wife, Jessica, for her love and support throughout this process, and for all she did, big and small, to make this possible.

TABLE OF CONTENTS

1. INTRODUCTION: A CRITICAL REVIEW: THE STATE OF SAMHD1 RESEARCH ...……....4

2. CHAPTER 1: THE SAMHD1 DNTP TRIPHOSPHOHYDROLAS IS CONTROLLED BY A REDOX SWITCH…..……………………………………………………………..……………...……………….40

3. CHAPTER 2: REDOX REGULATION OF SAMHD1 IN SKOV3 CELLS MODULATES NUCLEOTIDE POOLS AND CELL CYCLE PROGRESSION …..…………………….………..……93

4. CHAPTER 3: IDENTIIFICATION OF INHIBITORS OF THE DNTP TRIPHOSPHOHYDROLASE SAMHD1 USING A NOVEL AND DIRECT HIGH THROUGHPUT ASSAY………………………………………………………………………………………122

5. CONCLUSION…………...……………….…………...... 160

6. CURRICULUM VITA…………………………………...…………………………………………..….……...165

INTRODUCTION

A Critical Review: The State of SAMHD1 Research

NUCLEOTIDE METABOLISM OVERVIEW

Proper regulation of intracellular deoxynucleotide triphosphates (dNTPs) is essential for normal cellular metabolism. A balanced supply of each of the four canonical dNTPs is required for accurate genomic and mitochondrial DNA synthesis and repair1–3. As such, nucleotide metabolism is tightly regulated within cells. Organisms have evolved complex and dynamic mechanisms to control the synthesis and degradation of dNTPs (Fig. 1)4. The enzymes involved in these pathways are frequently regulated in a concentration dependent manner by their respective substrates and products5–9. Synthesis of dNTPs is accomplished by way of two distinct pathways; the de novo synthesis and the salvage pathways10,11.

During de novo synthesis, cytosolic ribonucleotide diphosphates are reduced to deoxynucleotide diphosphates (dNDPs) by the enzyme ribonucleotide reductase (RNR)12. dNDPs are then phosphorylated by cytosolic or mitochondrial kinases. This process is dramatically upregulated during S-phase, when the cell requires large amounts of dNTPs to replicate its DNA4. De novo synthesis of nucleotide triphosphates is also important for supplementing intracellular nucleotide pools during other stages of the cell cycle but is

carried out at a reduced rate13–15. The salvage pathway is constitutively active and performed in parallel in both the cytosol and mitochondrial compartments. It is essential for maintaining dNTP pool levels for DNA synthesis and repair in resting cells by recycling cellular deoxynucleosides that are products of nucleic acid degradation, cellular nucleotidases, or imported from the extracellular space4.

Opposing the de novo and salvage anabolic pathways, various enzymes function together to comprise the catabolic pathways that facilitate the degradation of dNTPs. 5’-nucleotidases, nucleoside phosphorylases, and deaminases are responsible for breaking down dNTPs into the constituent deoxynucleoside or nucleobase, at which point they can reenter the salvage pathway or be transported across the cell membrane in order to maintain the overall size and ratio of intracellular dNTPs. The opposing anabolic and catabolic pathways work in competing directions to maintain a dynamic equilibrium of intracellular dNTPs in a process referred to as substrate cycling16. Isotope flow experiments of radiolabeled nucleotides have traced this phenomena and elucidated its critical role as a regulatory mechanism for maintaining homeostatic dNTP levels17–19. It is also important to note that cytosolic dNTP pools do not exist in isolation, but are directly related to mitochondrial dNTP pools. In fact, dNTP precursors transport between the two compartments and the nucleotide metabolism processes of one compartment can influence the other20,21.

Substrate cycling, DNA synthesis and repair, and nucleoside import and export form a dynamic regulatory system that maintains homeostatic dNTP concentrations within the cell.

The levels are finely tuned according to the cell type and stage in the cell cycle. Peculiarly, intracellular dNTP concentrations are not present in equimolar amounts. dGTP is

consistently observed to be least prevalent in eukaryotic cells3,10. While dNTP concentrations are asymmetric, the proper balance of the dNTP pool is crucial to cell survival. Imbalanced dNTP pools, resulting from mutations to the enzymes involved in maintaining dNTP equilibrium, can lead to increased mutation rates and the induction of a mutator phenotype, increased DNA damage and inhibition of genomic repair, aborted DNA synthesis and stalled replication forks, and cell cycle arrest22–29. These molecular perturbations of genomic fidelity manifest themselves as severe pathologies including mitochondrial depletion syndromes, severe immunodeficiency disorders, and cancer30.

Imbalanced dNTP pools may also impact apoptosis and inflammation, as dNTPs have been implicated in the activation of both pathways31,32.

Figure 1. Overview of dNTP metabolism pathways and enzymes (Adapted from Rampazzo et al4). The de novo pathway is outlined in red. R1/R2 and R1/p53R2 represent the different isoforms of ribonucleotide reductase that reduce NDPs to dNDPs. The salvage pathway is outlined in blue and consists of complementary pathways in the cytosol and mitochondria. Cellular kinases (TK1, dCK, TK2, dGK) phosphorylate deoxynucleosides (dN) and deoxynucleotides (dNMP, dNDP). 5’- deoxynucleotidases (cdN, mdN) and phosphorylases (TP, PNP) catalyze the opposing reactions which degrade nucleotides. The reaction catalyzed by SAMHD1 is highlighted in green. Boxed symbols indicate cell cycle regulation. Shaded symbols identify enzymes implicated in disease.

SAMHD1 ACTIVATION, CATALYSIS, AND REGULATION

SAMHD1 Activation and Catalysis: Sterile Alpha Motif and Histidine-Aspartic acid domain containing protein 1 (SAMHD1) is an enzyme that performs an essential function in nucleotide metabolism. It was first termed dendritic cell derived IFN-γ induced protein

(DCIP) following its identification in dendritic cells as the human orthologue of an interferon-γ induced mouse protein33. A subsequent study revealed that DCIP was upregulated in response to TNF-⍺ in an IRF-1 dependent manner, but the precise function of the enzyme remained unclear34. It was not until almost 10 years after its initial discovery that SAMHD1 was implicated in nucleotide metabolism and demonstrated to be a prominent effector of the innate immune response35. Subsequent biochemical analysis reveal SAMHD1 to be a dNTP triphosphohydrolase that catalyzes the hydrolysis of canonical dNTPs into the constituent nucleoside and inorganic tripolyphosphate in the presence of regulatory nucleotide triphosphates and an activating divalent cation36,37 (Fig. 2). Through its dNTPase activity, SAMHD1 maintains dNTP pools within the cell at levels appropriate for

DNA replication and repair but below a potentially mutagenic threshold.

Figure 2. SAMHD1 is a deoxynucleotide triphosphohydrolase. SAMHD1 hydrolyzes dNTPs in the presence of activating nucleotides and divalent metal cations into their cognate nucleoside and inorganic tripolyphosphate. Nucleosides are then degraded further, exported from the cell, or recycled through the nucleotide salvage pathway.

X-ray crystallographic studies have since elucidated the precise structural features and catalytic mechanism of SAMHD1 (Fig. 3). It is a 626 amino acid protein comprised of an N- terminal sterile alpha motif (SAM) and a histidine-aspartic acid containing domain (HD).

While the function of the SAM domain is yet to be determined, SAM domains are commonly involved in protein-protein and protein-DNA/RNA interaction38. A nuclear localization signal precedes the SAM domain and confers the nuclear occupancy observed in most studies39. The HD domain is defined by its characteristic quartet of metal coordinating histidine and aspartic acid residues within the enzyme active site. HD-domain containing proteins represent a superfamily of phosphohydrolases commonly involved in nucleic acid metabolism40. The HD domain of SAMHD1 houses the dNTPase active site, regulatory sites, and the requisite interfaces for enzyme oligomerization. The C-terminus of SAMHD1 consists of a distinct region that is important for stabilizing the oligomeric state of the enzyme and nucleic acid interaction41–43.

SAMHD1 catalysis is regulated in a complex but elegant manner. Inactive apo-SAMHD1 normally exists in a monomer-dimer equilibrium, but tetramerizes in the presence of activating nucleotides in order to form the catalytically competent holoenzyme44. Each

SAMHD1 monomer contains two discrete regulatory sites (RS1 and RS2) and activating nucleotide triphosphates must sequentially bind at each site in order to induce a conformational shift that facilitates tetramerization and subsequent catalytic activation42,43.

The residues which constitute the RS1 pocket are uniquely structured such that only a guanosine triphosphate nucleotide is capable of binding with an estimated KD between 0.1-

0.4µM 36,45–48. Given the 1000-fold excess of GTP over dGTP within cells, it is likely that RS1 is constitutively occupied by GTP as a primary activating nucleotide under physiological

conditions45,46,49. GTP binding at RS1 drives the oligomeric equilibrium toward the dimer as the GTP nucleotide forms contacts with both subunits. Further studies have suggested that constitutive binding of GTP by SAMHD1 can form relaxed tetramers that are primed for full assembly and high-efficiency catalysis upon binding of a second activating nucleotide at

RS250.

In contrast to RS1, RS2 presents a more promiscuous binding site. RS2 can accommodate any of the four canonical dNTPs with apparent KD values between 1-20µM depending on the nucleobase46,48,51–53. These concentrations are physiologically relevant and thus binding of a dNTP in the RS2 of SAMHD1 can occur when intracellular dNTP concentrations are elevated into the activating range21. While RS2 is capable of binding any of the four dNTPs, there is a clear preference for purine nucleotides48,51. Thus, given the asymmetric nature of dNTP pools within the cell, it is likely that dATP is most frequently docked in RS246. While dATP may be the primary nucleotide bound in RS2, recent work has suggested that the specific dNTP docked in RS2 may confer differential stability and substrate specificity to the activated enzyme 52.

The binding event of a dNTP at RS2, which is preceded by docking of GTP in the guanine specific RS1 pocket, drives the oligomeric equilibrium towards tetramerization. Subunit assembly results in the formation of four regulatory clefts comprised of an RS1 and RS2 from adjacent monomers. The two nucleotides bound in each cleft are coordinated by a single Mg2+ ion that stabilizes their phosphate tails. At each of the four occupied regulatory clefts in the activated tetramer, the activating nucleotides form a network of hydrogen bonds, electrostatic, and pi-pi orbital interactions with residues from three different

monomers in order to stabilize subunit assembly. Tetramer assembly is further stabilized by several key structural motifs, specifically protein-protein interfaces formed by an ordering of the C-terminus (454-599), and a long alpha helix (352-373) that significantly contributes to the dimer-dimer interface42,43,47,48. Binding of activating nucleotides in each regulatory site and the subsequent formation of the tetramer results in conformational changes that remodel the active site and facilitate substrate binding and catalysis.

It has been suggested that the catalytically active tetrameric species can persist for extended periods even after dNTP levels have diminished below the effective level for

SAMHD1 activation45,52. This long-lived active state may be important for maintaining cellular dNTP pools at the extremely low concentrations observed in non-cycling cells.

Regardless of the duration of the activity, the elegant and strictly regulated mechanism of

SAMHD1 catalytic activation represents an ordered and sequential process in which dNTPs serve as both substrate and activating ligand. The finely-tuned autoregulatory mechanism, enables SAMHD1 to sense small fluctuations of dNTP concentrations within the cell and respond accordingly by degrading them to non-pathogenic levels.

SAMHD1 catalytic activity is uniquely adapted for hydrolysis of tripolyphosphate from deoxynucleotide triphosphates. As of this review, it represents the only described triphosphohydrolase in human cells. Binding of dNTPs in the active site is coordinated by a series of hydrogen bonds and electrostatic interactions between active site residues and the deoxyribose and phosphate moieties of dNTPs. Steric clashes with active site residues by the 2’-hydroxyl group found on ribonucleotides precludes efficient rNTP hydrolysis.

Figure 3. The sequential binding of regulatory nucleotides drives SAMHD1 tetramerization and catalytic activation. (A, B) SAMHD1 monomer depicting the HD domain major lobe (green) and C-terminal region (blue). The catalytic site and regulatory site 1 (RS1) and 2 (RS2) are indicated with bound dGTP (C) SAMHD1 monomer as in A and B, with paired nucleotides from the tetrameric regulatory cleft. Each cleft contains two regulatory nucleotide binding sites from adjacent monomers that stabilize subunit interactions. Chain A RS1 forms a regulatory cleft with RS2 from Chain D. Chain A RS2 forms a cleft with RS1 from Chain B. (D) Two views of the catalytically active tetrameric holoenzyme of SAMHD1 depicting the primary tetramer interface (left) and the homotetrameric assembly (right). Two of the four catalytic and regulatory clefts are visible and identified. (PDB ID: 4BZC)

Nucleobases are stabilized in the active site through a series of non-specific interactions with water molecules, which likely affords SAMHD1 substrate promiscuity42,43,47,48. In addition to the ability to hydrolyze all four canonical dNTPs and dUTP, the SAMHD1 active site is also able to accommodate base modified substrates54. dNTPs are oriented for chemistry to occur through an interaction between their α-phosphate and a divalent metal cation (commonly Mg2+) coordinated by the His167-His206-Asp207-Asp311 quartet characteristic of HD domains. His210, His233, and Asp218 are believed to be the residues responsible for actual catalysis, which occurs through an in-line nucleophilic attack at the alpha phosphate resulting in tripolyphosphate and the cognate nucleoside as reaction products 55.

Despite the substrates of SAMHD1 also serving as its activators, SAMHD1 displays little evidence of cooperative kinetics. The sequential activation and assembly of the SAMHD1 tetramer occur at dNTP concentrations multiple orders of magnitude below the KM. Thus, while KM values for SAMHD1 vary considerably depending on the study from 50-350 μM, under all the experimental conditions steady-state Michaelis-Menten kinetics apply

45,46,48,53,56,57. Given the steady-state conditions of an activated SAMHD1 tetramer, the turnover rate for each substrate appears to depend solely on its particular affinity for the active site and its intracellular concentration. In multiple biochemical analyses, dCTP and dGTP were consistently observed to have the highest turnover rate, while dATP represented the weakest substrate. Recent evidence however, suggests that this model may not be entirely accurate. In a manner resembling ribonucleotide reductase activation, the

particular dNTP bound in the RS2 may influence substrate specificity and catalytic efficiency53. The precise nature of SAMHD1 activation and substrate specificity in vivo represent an open inquiry that requires further investigation into the underlying molecular mechanism and physiological significance.

SAMHD1 Nucleic Acid Interaction: In addition to its dNTPase activity, SAMHD1 is also a demonstrated nucleic acid interacting protein. SAMHD1 is capable of binding DNA and

DNA:RNA duplexes, but preferentially binds single stranded nucleic acid polymers (ssNAs), with a greater affinity for ssRNA over ssDNA 58–60. The specific affinity of SAMHD1 for ssNAs may be regulated by their secondary structure61. The phenomena of SAMHD1 ssNA interaction has also been reported in cell culture, where multiple SAMHD1 monomers converge on sites of ssRNA or ssDNA62. Consistent with in vivo data, in vitro work has confirmed that it is the monomeric form of SAMHD1 that primarily interacts with nucleic acids, and that SAMHD1 monomers can form multi-subunit complexes in the presence of ssNA58.

Binding to ssNAs requires an intact HD domain and is mediated by a region of the C- terminus (residues 583-626) that can bind nucleic acids as an independent peptide 58–60,62,63.

Interestingly, nucleic acid binding by SAMHD1 monomers inhibits dNTPase activity by obstructing the interfaces responsible for tetramer association. This obstruction can be relieved by increasing the concentration of activating nucleotides63. While the precise function of nucleic acid interaction in cells is yet to be determined, further investigation of the dynamic cycle between ssNA bound inactive monomer-dimer and catalytically active tetramer may reveal important aspects of SAMHD1 biology.

SAMHD1 has also been reported to exhibit exonuclease activity against both ssRNA and ssDNA 61,64,65. The proposed mechanism of catalysis for SAMHD1 exonuclease activity is a phosphorolytic cleavage of the phosphodiester bond in nucleic acid polymers66. These findings are widely contested however, as the majority of groups that have investigated

SAMHD1 have not been able to detect nuclease activity 36,58–60,67.

SAMHD1 regulation by post-translational modification: While multiple biochemical and structural studies have rigorously characterized the mechanism of SAMHD1 activation and catalysis, an emerging area of interest is the effect of known post-translational modifications on SAMHD1 activity. SAMHD1 is phosphorylated at multiple sites, but the most extensively studied is phosphorylation at its C-terminal T592 residue (P-T592).

Multiple groups have reported that SAMHD1 is phosphorylated at this residue by cyclin dependent kinases prior to S-phase entry68–75. The exact effect of phosphorylation on

SAMHD1 activity is widely disputed in the literature, however. Several groups claim that phosphorylation negatively modulates SAMHD1 tetramerization and dNTPase activity76,77.

Several investigations utilizing cell culture techniques correlate SAMHD1 phosphorylation with elevated intracellular dNTP pools70,72,74,78. Other studies find no effect on SAMHD1 catalytic activity, oligomerization equilibrium, and nucleic acid binding52,56,58,69,79,80.

Complicating the model further, phosphorylation may affect SAMHD1 substrate specificity53.

Structural studies attempting to shed light on this discrepancy reveal that P-T592 results in

an increase in the rate of tetramer dissociation and expedited regulatory nucleotide release52,80. P-T592 was also less likely to form the activated tetramer at low concentration of activating nucleotides 53,56. Crystallography data support these findings by identifying structural shifts instigated by a phosphomimetic mutant (T592E) that destabilizes the tetramer interface76. Taken in sum, these data appear to indicate that phosphorylation at

T592 is a mechanism for calibrating SAMHD1 activity by altering the thermodynamics of subunit association. The discrepant results may stem from variations in assay conditions that lack the sensitivity to accurately capture the true effect of phosphorylation. Fine tuning

SAMHD1 activity within the cell by phosphorylation, as opposed to a binary on or off state, may be important as a complementary method of regulation in order to maintain dNTP pools within the narrow window conducive to genomic integrity.

SAMHD1 catalytic activity is also regulated by redox signaling. Redox signaling takes the form of reactive oxygen species that are generated by cellular oxidases in response to growth factors binding and activating cellular receptors81. These ROS act as secondary messengers that can covalently, but reversibly, interact with an enzyme to modify its tertiary structure and subsequent activity. SAMHD1 is catalytically inactivated when treated with the oxidizing agent H2O2 in a dose dependent but reversible manner57.

Oxidation of SAMHD1 was demonstrated to inhibit tetramerization, but this deficiency was ablated in SAMHD1 mutants containing a C522A mutation. Closer inspection of SAMDH1 crystal structures reveal that C522 comprises one member of a cysteine triad - C522, C341, and C350 - that resides adjacent to the regulatory nucleotide binding site RS1. In several crystal structures, a disulfide bond exists between C341 and C350 and results in a conformational shifts that may disrupt activating nucleotide binding and tetramer stability.

Additionally, SAMHD1 was demonstrated to be oxidized in cells in response to proliferative signals57. It was proposed that cells oxidize SAMHD1 in response to proliferative signaling in order to accumulate the dNTPs necessary for DNA replication. This oxidation effect is mediated by the cysteine triad, referred to as a redox switch, that can detect ROS secondary messengers and translate them into structural rearrangements that alter SAMHD1 catalytic activity. Importantly, this phenomenon is reversible and tightly controlled. As with phosphorylation, redox regulation of SAMHD1 represents a layer of control orthogonal to catalytic activation that can be localized spatially and temporally. The meticulously controlled dynamic gradient of SAMHD1 activity through multiple means of regulation likely performs an important role in maintaining dNTP pools within a homeostatic range at each point in the cell cycle (Fig. 4).

Figure 4. SAMHD1 catalytic activity is tightly controlled by regulatory nucleotides and essential for the maintenance of nucleotide homeostasis. Under conditions of low dNTPs, SAMHD1 exists in a monomer-dimer equilibrium. Binding of GTP in RS1 stabilizes the dimer conformation. Elevation of intracellular dNTP concentrations above the activation threshold (1- 20µM) results in dNTPs binding at RS2 and SAMHD1 tetramerization. Tetrameric SAMHD1 is able to catalyze the degradation of dNTPs and thereby prevent accumulation to levels that would be cytotoxic. Phosphorylation is proposed to destabilize tetramer stability with out modifying catalytic efficiency, thereby allowing for an increase in dNTP pools necessary for DNA replication without creating mutagenic dNTP conditions.

SAMHD1 IS AN EFFECTOR OF INNATE IMMUNITY

SAMHD1 has garnered significant attention for its antiviral properties in non-dividing cells of hematopoietic lineage, where it acts as an interferon inducible host restriction factor against viral infection, including HIV60,82–84. SAMHD1 is constitutively expressed in non- dividing nucleated hematopoietic cells such as macrophages, dendritic cells, and resting

CD4+T-cells85–87. SAMHD1 is also transcriptionally upregulated in an interferon dependent manner following viral challenge or in response to Il-12 and Il-1875,88. In these cells it maintains dNTP levels that are 130-250 fold lower than in those cell where it is not expressed or down-regulated, such as in activated cycling human CD4+ T cells85,87. Through its dNTPase functionality, SAMHD1 inhibits productive viral infection by reducing dNTP pools below the effective concentration of viral DNA polymerases84,89. SAMHD1 thereby provides a kinetic block at the critical juncture where viral genomic content is replicated and incorporated into the host’s genome. The kinetic block erected by SAMHD1 is reversible upon treatment with exogenous dNTPs or knockdown of SAMHD183,84,89,90. This strategy of warding off invasive pathogens through depletion of precursor metabolites is not a novel approach as even Eschericia coli expresses a dGTPase that reduces its susceptibility to T7 bacteriophage infection91,92. Therapeutically targeting nucleic acid metabolizing enzymes in order to deplete intracellular dNTP pools is also an approach implemented by modern medicine in order to counter viral infection and cancer.

While SAMHD1 inhibits HIV-1 infection of non-dividing myeloid and resting T-cells, the HIV-

2 subtype is able to actively transduce these cell types. HIV-2 virions are packed with a virulence accessory protein, Vpx, which orchestrates the degradation of nuclear SAMHD193.

The loss of SAMHD1 and subsequent increase in dNTPs permits productive viral infection.

The importance of Vpx mediated SAMHD1 degradation for productive infection was evidenced through experiments in which Vpx was incorporated into HIV-1 subtype virions and enabled them to productively infect previously non-permissive cell lines94. Vpx accomplishes the degradation of SAMHD1 by recruiting SAMHD1 to an E3 ubiquitin ligase where it is ubiquitinylated and marked for proteosomal degradation41,95. The recruitment is facilitated by Vpx forming a complex with the ubiquitin ligase substrate adaptor molecule

DCAF1 and the C-terminal region of SAMHD196.

More recently however, dNTP depletion has come under question as the primary mechanism through which SAMHD1 restricts viral infection. It was demonstrated that tetramerization deficient SAMHD1 enzymes are still able to exert antiviral effects80,97.

Additionally, SAMHD1 requires an intact C-terminus for HIV restriction but not dNTPase activity44,56. Phosphorylation of T592, located on the C-terminus, has been well established as coinciding with the ablation of the viral restriction capacity within cells. As previously described, phosphorylation of SAMHD1 at T592 is an important regulatory post- translational modification and is found in actively cycling cells with elevated dNTP pools.

Multiple studies suggest, however, that while phosphorylation of SAMHD1 directly reduces its viral restriction capabilities, it does not do this by down-regulating its catalytic activity69,98. Cells expressing phosphomimetic T592D or T592E mutations support this model, as they too are unable to restrict viral infection but do not exhibit elevated dNTP pools. Interferon signaling or the terminal differentiation of a cell results in the loss of phosphorylation at Thr592, and the reemergence of a cellular phenotype that is refractory to HIV-1 infection68,69.

Interestingly, of the two subtypes, HIV-2 is the less virulent. Why this is the case is not entirely understood, but it has been hypothesized that by using Vpx to degrade SAMHD1 and enable productive infection of macrophages and dendritic cells, which serve as the sentinel cells of the immune system, HIV-2 may in fact be alerting the host’s natural defenses to its presence. HIV-1 restriction in non-dividing immune cells may enable it to maintain an undetectable reservoir of latent virus, while avoiding activation of the host’s immune systems99–101. Recent evidence supports this model by demonstrating that SAMHD1 restriction of HIV-1 limits the innate and adaptive immune responses by preventing the activation of the cGAS/STING pathway102. Inhibitors of SAMHD1 may therefore represent an important therapeutic tool for facilitating infection of myeloid cells and activating a robust immune response to HIV-1 infection103.

SAMHD1, NUCLEOTIDE METABOLISM, AND CELL CYCLE REGULATION

Compared to its innate immune function, less attention has been given to the role SAMHD1 plays in normal homeostatic cell maintenance. Given that SAMHD1 is nearly ubiquitously expressed in most tissue types however, its role as a regulator of nucleotide metabolism may represent its primary biological function. Multiple studies identify SAMHD1 as a central regulator of dNTP pools within cells89,104–106. SAMHD1 maintains reduced dNTP pools in cells by catalyzing their degradation into deoxynucleoside and inorganic triphosphate. Silencing or degradation of SAMHD1 results in elevated dNTP pools and G1 cell cycle arrest.

Maintenance of dNTP pools at the appropriate concentrations by SAMHD1 is important for several reasons. As cells progress through the cell cycle, their need for dNTPs fluctuates.

During S-phase, cells need large quantities of dNTPs in order to replicate their genomic

DNA. Thus, SAMHD1 catalytic activity directly influences the replicative capacity of the cell.

Up-regulation of SAMHD1 has been demonstrated to be a key effector of maintaining a quiescent or differentiated phenotype in varying cell types105,107. Similarly, overexpression of exogenous SAMHD1 results in reduced proliferation and increased apoptosis in HuT78 cells, a human T-cell lymphoma cell line108. Conversely, in cells where SAMHD1 has been knocked out, elevated dNTP pools coincide with more rapid cell proliferation and reduced apoptosis106.

Maintenance of proper concentrations of each dNTP is also essential for genomic stability.

Dysregulation of nucleotide metabolism can have dangerous mutagenic or cytotoxic consequences, as imbalanced dNTP pools can impede DNA replication and repair. By degrading dNTPs in resting or quiescent cells before they accumulate to levels that impede the efficiency or accuracy of the replication and repair machinery, SAMHD1 performs an essential function as a protector of genomic fidelity. Data from patients expressing inactive mutants of SAMHD1 underscore the importance of SAMHD1 as a sentinel of genome integrity. Cells from these patients exhibit elevated dNTPs, growth arrest in G1, and increased DNA damage109.

Given the importance of nucleotide metabolism, and the central function performed by

SAMHD1 in its regulation, the catalytic activity of SAMHD1 is tightly controlled in a cell cycle dependent manner. Regulation of SAMHD1 is accomplished through both transcriptional and post-translational mechanisms. Each mechanism is coordinated with the metabolic needs of the cell and results in a finely tuned system that supplies the cell with the appropriate level of dNTPs in relation to its cell cycle stage. While the total amount of

SAMHD1 present in cells depends on the specific cell type and cell cycle stage, the regulation of SAMHD1 begins at a transcriptional level, where it is constitutively expressed in the majority of non-dividing human cells110. In actively dividing cells, SAMHD1 is also present, although at reduced levels and often in a post-translationally modified state. Reduced levels of SAMHD1 expression are in part due to down-regulation of the SAMHD1 promoter by methylation, as well as proteasomal degradation111–114.

As previously discussed, SAMHD1 is post-translationally modified through both phosphorylation and oxidation. SAMHD1 is phosphorylated at T592 by cyclin dependent kinases in actively cycling cells. Phosphorylation of SAMHD1 at T592 appears to affect

SAMHD1 dNTPase activity by increasing the kinetics of tetramer disassembly, thereby permitting an increase in dNTP pools prior to S-phase DNA replication. Phosphorylation of

SAMHD1 is not observed in cells in G0/G1 phase. Reversible oxidation of SAMHD1 occurs via redox signaling pathways upregulated in response to mitogenic signals. Oxidation of

SAMHD1 at C522 triggers the formation of a disulfide bond in a region termed the SAMHD1 cysteine switch. It is proposed that this disulfide bond impairs tetramerization and catalytic function, and thereby contributes to increasing intracellular dNTP levels and subsequent DNA replication.

SAMHD1 is a primary modulator of nucleotide pools and cell cycle progression. As such, its expression, activation, and catalysis, are controlled through multiple layers of regulation.

Through these regulatory mechanisms, cells are able generate a spectrum of SAMHD1 activity that can be calibrated to meet their metabolic needs at specific points in the cell cycle. SAMHD1 is part of a broader network of tightly regulated nucleotide metabolizing enzymes that work in parallel to maintain appropriate intracellular dNTP concentrations.

Despite its centrality to cellular homeostasis, relatively little is known about the intracellular regulation of SAMHD1 Future research in these areas will likely yield important insights into the underlying mechanisms that control nucleotide metabolism and cell cycle progression, as well as the pathologies that result from their dysregulation.

SAMHD1 MUTATIONS RESULT IN DISEASE

Autoimmunity. Given the central role SAMHD1 performs in nucleotide metabolism and immunity, mutations to SAMHD1 can result in dangerous pathologies. One of these pathologies is the Type I interferonopathy Aicardi-Goutieres Syndrome (AGS) – a genetically determined autoimmune disease that resembles a congenital viral infection and is linked to aberrant nucleic acid processing115. AGS is an autosomal recessive disorder that mimics the sequelae of a viral infection and retains a significant overlap in phenotypic presentation with that of systemic lupus erythematosus (SLE)116,117. AGS is characterized by encephalopathy, leukodystrophy, calcifications of the basal ganglia, psychomotor retardation, CSF lymphocytosis, and an over-production of interferon-α118. Approximately

40% of the cases result in early childhood death119.

AGS is a genetically heterogeneous disease, with mutations to SAMHD1 accounting for approximately five percent of all documented cases. Its etiology can be traced to mutations in SAMHD1 and several other enzymes involved in nucleic acid processing120. Mutagenic inactivation of these enzymes may result in the cytosolic accumulation of unprocessed nucleic acids which mimic invading pathogens. Unprocessed nucleic acids can then activate endogenous nucleic acid sensing pathways and result in subsequent interferon production.

The common functionality of nucleic acid metabolism which links the causative AGS enzymes has led to the hypothesis that they all play a role in suppressing antigenic nucleic acid by-products. The source of the antigenic nucleic acids however is unknown, and may vary by implicated enzyme. Conjectures as to how SAMHD1 negatively regulates interferon pathways have ranged widely - from suppression of retroelement activation; to restriction of nucleotide induced inflammation; to reduction of ssDNA in cells through the maintenance of proper replication kinetics or the inhibition of futile DNA synthesis121–125.

Interestingly, the phenotypic presentation of SAMHD1 associated AGS has several novel features not observed in other AGS associated genotypes, including cerebral vasculopathy, arthropathy, and mitochondrial DNA deletions126–128. In separate studies, several of these pathologies have also been linked to dysregulated nucleotide metabolism17,129,130. The clinical presentation of SAMHD1 induced AGS may therefore provide clues to the mechanism through which SAMHD1 contributes to nucleotide homeostasis and negatively regulates the innate immunity. Attempts to recapitulate an AGS phenotype in SAMHD1 null mice have proven perplexing however, as SAMHD1 knockout mice exhibit elevated interferon levels but do not develop autoimmune disease131,132.

Cancer. Ample evidence exists to connect dNTP pool dysregulation to cancer30. As discussed in detail above, maintenance of proper dNTP levels is essential for genomic stability. Elevated or imbalanced dNTP pools can result in severely mutagenic conditions

23,24,28. These conditions, termed a mutator phenotype, disrupt genomic integrity and DNA replication and repair133,134. Conversely, dNTP pool depletion can result in genomic instability that progresses to cancer135. Any condition that results in an imbalance to dNTP pools can lead to loss of genomic fidelity and impaired replication efficiency and accuracy

25–27,136.

It is therefore no surprise that recent studies have begun to implicate SAMHD1 in various cancers, including chronic lymphocytic leukemia, lung adenocarcinoma, and colon cancer111,137,138. Mutations to SAMHD1 likely result in an imbalanced dNTP pools and create the prerequisite conditions for mutagenesis and genomic instability139. It is also possible that SAMHD1 prevents mutations by sanitizing nucleotide pools by degrading base modified dNTPs before they are incorporated into DNA140. Given its role in the maintenance of genomic stability, SAMHD1 fulfills an additional function in cells as a tumor suppressor enzyme. It will be important to determine if mutations to SAMHD1 are foundational in that they drive tumorigenicity, or if cancer cells must down-regulate the restrictive effect of

SAMHD1 in order to supply themselves with the metabolites necessary for unconstrained growth.

Lastly, SAMHD1 plays an important role in cancer therapy. Nucleotide analogue therapeutics are a common tool used in treatment of cancer and viral infections141.

Following phosphorylation by intracellular kinases, these analogues are structurally similar

to the endogenous dNTP substrate of SAMHD1. Several of these nucleotide analogues have been demonstrated to serve as substrates for SAMHD1. Most notably, Ara-C, a first line therapeutic regimen against acute myelogenous leukemia (AML), is degraded by SAMHD1 in cells142,143. This minimizes the efficacy of the treatment to such an extent that SAMHD1 expression levels were negatively correlated with Ara-C treatment success in individuals with AML. Based on these finding, the development of a robust SAMHD1 inhibitor that can potentiate nucleotide analogue therapeutic regimens should become a priority for SAMHD1 researchers. Several groups have developed high throughput assays, but as of this writing, no potent inhibitor of SAMHD1 has been identified144,145.

CONCEPTUAL FRAMEWORK OF THIS STUDY

Since its discovery as an HIV restriction factor, SAMHD1 has been the target of significant research efforts. The excellent science that has resulted from these efforts has extensively detailed the fundamental biochemistry and cell biology of SAMHD1. Important questions related to SAMHD1 biology remain, however. Further research into the catalytic activity and mechanisms of regulation in vivo will be important for a clear understanding of the enzyme’s operation in the intricate and dynamic pathway of nucleotide metabolism.

Additionally, the role of the SAM domain and SAMHD1 interactions with other proteins or nucleic acid remain open lines of inquiry. In the context of disease, the comprehensive methods of viral restriction appears to be incompletely described, as does the precise manner in which SAMHD1 acts as a negative regulator of interferon signaling. Research into the role of SAMHD1 in disease, specifically cancer and autoimmunity, is at its nascent

stages, but each new discovery presents exciting translational opportunities.

In a broader context, the centrality of SAMHD1 to cellular homeostasis is becoming apparent. SAMHD1 is a piece of a much larger puzzle that is essential for maintaining genomic integrity and regulating progression through the cell cycle. Thus, in this study, my goal was to investigate several of the outstanding questions listed above within the more expansive paradigm of SAMHD1 as a central regulator of cellular homeostasis. Specifically, I focused on the intracellular regulation of SAMHD1 and the opportunities for therapeutically targeting its catalytic activity.

Study Objective. The purpose of this study was to investigate the regulation of SAMHD1 and its relation to nucleotide metabolism and cell cycle progression. Phosphorylation of

SAMHD1 is the most extensively documented means of regulation. While questions still remain regarding regulation of SAMHD1 by phosphorylation, I chose to investigate an alternative mechanism of regulation. A major signaling pathway that has not been investigated with regards to SAMHD1 regulation is redox biology. Redox biology is emerging as a central regulator in myriad cellular processes including cell cycle progression, cellular metabolism, and cell signaling. Given the role that nucleotide metabolism plays in these processes, this study was undertaken to test the hypothesis that

SAMHD1 is a redox regulated enzyme and that redox signaling is important for modulating

SAMHD1 activity in vivo.

Based on this hypothesis, this study sought to address the following questions:

1. Does redox signaling represent an alternative mechanism of regulation of SAMHD1?

2. How does redox signaling exert its effect on SAMHD1 function?

3. What effect does redox regulation of SAMHD1 have on nucleotide metabolism and

homeostasis?

4. Given the importance of SAMHD1 to multiple cellular processes and its implication

in several disease states, can SAMHD1 be therapeutically targeted?

Methodological Approach. All work was performed in the laboratory of Dr. Thomas

Hollis, faculty member at the Wake Forest University School of Medicine and the Wake

Forest Center for Structural Biology. A combination of biophysical, biochemical, and cell culture techniques were utilized to conduct experiments aimed at answering the questions outlined in the study objective. Individual techniques and methods are explained in detail in each chapter.

Significance of the Study. Through the research efforts documented in this dissertation, I describe a novel mechanism of regulation for SAMHD1. This discovery advances the understanding of SAMHD1 catalytic activity and intracellular control. Perhaps more importantly however, it directly links redox signaling pathways and nucleotide metabolism.

This finding underscores the importance of redox biology as an essential mediator of cellular processes. Lastly, the discovery and optimization of a high throughput screen, as well as the identification of several small molecule inhibitors of SAMHD1, has important translational significance in the effort to modulate SAMHD1 activity for improved health outcomes.

Organization of the Study. The body of work presented in this study is divided into three main chapters. Chapters 1 describes a biochemical and cell culture based investigation of

SAMHD1 as a redox regulated enzyme. It is comprised of a manuscript published in

Antioxidants and Redox Signaling (February 2017). Chapter 2 is an examination of the importance of SAMHD1 redox regulation for cellular homeostasis. It is a preliminary study that will form the foundation of a manuscript to be submitted pending validation of the observed effects in multiple cell lines. It also serves to detail the development of several methods used to examine the redox regulation of SAMHD1 in cells. Chapter 3 details development of a high throughput screen and subsequent drug discovery research related to SAMHD1. It will be submitted as a manuscript for publication pending minor modifications.

REFERNCES

1. Reichard P. Interactions Between Deoxyribonucleotide And DNA Synthesis. Annu Rev Biochem. 1988;57(1):349-374. doi:10.1146/annurev.biochem.57.1.349.

2. Mathews CK. Deoxyribonucleotides as genetic and metabolic regulators. FASEB J. 2014;28(9):3832-3840. doi:10.1096/fj.14-251249.

3. Mathews CK. DNA precursor metabolism and genomic stability. FASEB J. 2006;20(9):1300-1314. doi:10.1096/fj.06-5730rev.

4. Rampazzo C, Miazzi C, Franzolin E, et al. Regulation by degradation, a cellular defense against deoxyribonucleotide pool imbalances. Mutat Res - Genet Toxicol Environ Mutagen. 2010;703(1):2-10. doi:10.1016/j.mrgentox.2010.06.002.

5. Hofer A, Crona M, Logan DT, Sjöberg B-M. DNA building blocks: keeping control of manufacture. Crit Rev Biochem Mol Biol. 2012;47(1):50-63. doi:10.3109/10409238.2011.630372.

6. Walldén K, Nordlund P. Structural basis for the allosteric regulation and substrate recognition of human cytosolic 5???-nucleotidase II. J Mol Biol. 2011;408(4):684-696. doi:10.1016/j.jmb.2011.02.059.

7. Johansson K, Ramaswamy S, Ljungcrantz C, et al. Structural basis for substrate specificities of cellular deoxyribonucleoside kinases. Nat Struct Biol. 2001;8(7):616- 620. doi:10.1038/89661.

8. Hunsucker SA, Mitchell BS, Spychala J. The 5???-nucleotidases as regulators of nucleotide and drug metabolism. Pharmacol Ther. 2005;107(1):1-30. doi:10.1016/j.pharmthera.2005.01.003.

9. Bianchi V, Spychala J. Mammalian 5???-Nucleotidases. J Biol Chem. 2003;278(47):46195-46198. doi:10.1074/jbc.R300032200.

10. Reichard P. Interactions between deoxyribonucleotide and DNA synthesis. Annu Rev Biochem. 1988;57(1):349-374. doi:10.1146/annurev.biochem.57.1.349.

11. Mathews CK. DNA precursor metabolism and genomic stability. Faseb J. 2006;20(9):1300-1314. doi:10.1096/fj.06-5730rev.

12. Nordlund P, Reichard P. Ribonucleotide Reductases. Annu Rev Biochem. 2006;75(1):681-706. doi:10.1146/annurev.biochem.75.103004.142443.

13. Håkansson P, Hofer A, Thelander L. Regulation of mammalian ribonucleotide reduction and dNTP pools after DNA damage and in resting cells. J Biol Chem. 2006;281(12):7834-7841. doi:10.1074/jbc.M512894200.

14. Pontarin G, Ferraro P, Bee L, Reichard P, Bianchi V. Mammalian ribonucleotide reductase subunit p53R2 is required for mitochondrial DNA replication and DNA repair in quiescent cells. Proc Natl Acad Sci U S A. 2012;109(33):13302-13307. doi:10.1073/pnas.1211289109.

15. Pontarin G, Ferraro P, Rampazzo C, et al. Deoxyribonucleotide metabolism in cycling and resting human fibroblasts with a missense mutation in p53R2, a subunit of ribonucleotide reductase. J Biol Chem. 2011;286(13):11132-11140. doi:10.1074/jbc.M110.202283.

16. Nicander B, Reichard P. Evidence for the involvement of substrate cycles in the regulation of deoxyribonucleoside triphosphate pools in 3T6 cells. J Biol Chem. 1985;260(16):9216-9222. http://www.ncbi.nlm.nih.gov/pubmed/3926765.

17. Bianchi V, Pontis E, Reichard P. Dynamics of the dATP pool in cultured mammalian cells. Exp Cell Res. 1992;199(1):120-128. doi:10.1016/0014-4827(92)90469-O.

18. Spyrou G, Reichard P. Dynamcis of the thymidine triphosphate pool during the cell cycle of synchronized 3T3 mouse fibroblasts. Mutat Res - Fundam Mol Mech Mutagen. 1988;200(1-2):37-43. doi:10.1016/0027-5107(88)90069-3.

19. Leanza L, Ferraro P, Reichard P, Bianchi V. Metabolic interrelations within guanine deoxynucleotide pools for mitochondrial and nuclear DNA maintenance. J Biol Chem. 2008;283(24):16437-16445. doi:10.1074/jbc.M801572200.

20. Rampazzo C, Ferraro P, Pontarin G, Fabris S, Reichard P, Bianchi V. Mitochondrial Deoxyribonucleotides, Pool Sizes, Synthesis, and Regulation. J Biol Chem. 2004;279(17):17019-17026. doi:10.1074/jbc.M313957200.

21. Gandhi V V, Samuels DC. A review comparing deoxyribonucleoside triphosphate (dNTP) concentrations in the mitochondrial and cytoplasmic compartments of normal and transformed cells. Nucleosides Nucleotides Nucleic Acids. 2011;30(5):317- 339. doi:10.1080/15257770.2011.586955.

22. Weinberg G, Ullman B, Martin DW. Mutator phenotypes. 1981;78(4):2447-2451.

23. Kunz BA. Mutagenesis and deoxyribonucleotide pool imbalance. Mutat Res - Fundam Mol Mech Mutagen. 1988;200(1-2):133-147. doi:10.1016/0027-5107(88)90076-0.

24. Meuth M. The molecular basis of mutations induced by deoxyribonucleoside triphosphate pool imbalances in mammalian cells. Exp Cell Res. 1989;181(2):305- 316. doi:10.1016/0014-4827(89)90090-6.

25. Kumar D, Viberg J, Nilsson AK, Chabes A. Highly mutagenic and severely imbalanced dNTP pools can escape detection by the S-phase checkpoint. Nucleic Acids Res. 2010;38(12):3975-3983. doi:10.1093/nar/gkq128.

26. Buckland RJ, Watt DL, Chittoor B, Nilsson AK, Kunkel TA, Chabes A. Increased and Imbalanced dNTP Pools Symmetrically Promote Both Leading and Lagging Strand Replication Infidelity. PLoS Genet. 2014;10(12). doi:10.1371/journal.pgen.1004846.

27. Watt DL, Buckland RJ, Lujan SA, Kunkel TA, Chabes A. Genome-wide analysis of the specificity and mechanisms of replication infidelity driven by imbalanced dNTP pools. Nucleic Acids Res. 2015;44(4):1669-1680. doi:10.1093/nar/gkv1298.

28. Chabes A, Stillman B. Constitutively high dNTP concentration inhibits cell cycle progression and the DNA damage checkpoint in yeast Saccharomyces cerevisiae. Proc Natl Acad Sci U S A. 2007;104(4):1183-1188. doi:10.1073/pnas.0610585104.

29. Davidson MB, Katou Y, Keszthelyi A, et al. Endogenous DNA replication stress results in expansion of dNTP pools and a mutator phenotype. EMBO J. 2012;31(4):895-907. doi:10.1038/emboj.2011.485.

30. Mathews CK. Deoxyribonucleotide metabolism, mutagenesis and cancer. Nat Rev Cancer. 2015;15(9):528-539. doi:10.1038/nrc3981.

31. Davis BK, Wen H, Ting JP-Y. The Inflammasome NLRs in Immunity, Inflammation, and Associated Diseases. Annu Rev Immunol. 2011;29(1):707-735. doi:10.1146/annurev- immunol-031210-101405.

32. Riedl SJ, Salvesen GS. The apoptosome: signalling platform of cell death. Nat Rev Mol Cell Biol. 2007;8(5):405-413. doi:10.1038/nrm2153.

33. Li N, Zhang W, Cao X. Identification of human homologue of mouse IFN-γ induced protein from human dendritic cells. Immunol Lett. 2000;74(3):221-224. doi:10.1016/S0165-2478(00)00276-5.

34. Liao W, Bao Z, Cheng C, Mok Y-K, Wong WSF. Dendritic cell-derived interferon-γ- induced protein mediates tumor necrosis factor-α stimulation of human lung fibroblasts. Proteomics. 2008;8(13):2640-2650. doi:10.1002/pmic.200700954.

35. Rice GI, Bond J, Asipu A, et al. Mutations involved in Aicardi-Goutières syndrome implicate SAMHD1 as regulator of the innate immune response. Nat Genet. 2009;41(7):829-832. doi:10.1038/ng.373.

36. Powell RD, Holland PJ, Hollis T, Perrino FW. Aicardi-Goutieres Syndrome Gene and HIV-1 Restriction Factor SAMHD1 Is a dGTP-regulated Deoxynucleotide Triphosphohydrolase. J Biol Chem. 2011;286(51):43596-43600. doi:10.1074/jbc.C111.317628.

37. Goldstone DC, Ennis-Adeniran V, Hedden JJ, et al. HIV-1 restriction factor SAMHD1 is a deoxynucleoside triphosphate triphosphohydrolase. Nature. 2011;480(7377):379- 382. doi:10.1038/nature10623.

38. Qiao F, Bowie JU. The Many Faces of SAM. Sci Signal. 2005;2005(286):re7-re7. doi:10.1126/stke.2862005re7.

39. Brandariz-Nuñez A, Valle-Casuso J, White TE, et al. Role of SAMHD1 nuclear localization in restriction of HIV-1 and SIVmac. Retrovirology. 2012;9(1):49. doi:10.1186/1742-4690-9-49.

40. Aravind L, Koonin E V. The HD domain defines a new superfamily of metal- dependent phosphohydrolases. Trends Biochem Sci. 1998;23(12):469-472. doi:10.1016/S0968-0004(98)01293-6.

41. DeLucia M, Mehrens J, Wu Y, Ahn J. HIV-2 and SIVmac accessory virulence factor Vpx down-regulates SAMHD1 enzyme catalysis prior to proteasome-dependent degradation. J Biol Chem. 2013;288(26):19116-19126. doi:10.1074/jbc.M113.469007.

42. Zhu C, Gao W, Zhao K, et al. Structural insight into dGTP-dependent activation of tetrameric SAMHD1 deoxynucleoside triphosphate triphosphohydrolase. Nat

Commun. 2013;4:2722. doi:10.1038/ncomms3722.

43. Ji X, Wu Y, Yan J, et al. Mechanism of allosteric activation of SAMHD1 by dGTP. Nat Struct Mol Biol. 2013;20(11):1304-1309. doi:10.1038/nsmb.2692.

44. Yan J, Kaur S, DeLucia M, et al. Tetramerization of SAMHD1 is required for biological activity and inhibition of HIV infection. J Biol Chem. 2013;288(15):10406-10417. doi:10.1074/jbc.M112.443796.

45. Hansen EC, Seamon KJ, Cravens SL, Stivers JT. GTP activator and dNTP substrates of HIV-1 restriction factor SAMHD1 generate a long-lived activated state. Proc Natl Acad Sci. 2014;111(18):E1843-E1851. doi:10.1073/pnas.1401706111.

46. Miazzi C, Ferraro P, Pontarin G, Rampazzo C, Reichard P, Bianchi V. Allosteric Regulation of the Human and Mouse Deoxyribonucleotide Triphosphohydrolase Sterile -Motif/Histidine-Aspartate Domain-containing Protein 1 (SAMHD1). J Biol Chem. 2014;289(26):18339-18346. doi:10.1074/jbc.M114.571091.

47. Zhu C-F, Wei W, Peng X, Dong Y-H, Gong Y, Yu X-F. The mechanism of substrate- controlled allosteric regulation of SAMHD1 activated by GTP. Acta Crystallogr Sect D Biol Crystallogr. 2015;71(3):516-524. doi:10.1107/S1399004714027527.

48. Ji X, Tang C, Zhao Q, Wang W, Xiong Y. Structural basis of cellular dNTP regulation by SAMHD1. Proc Natl Acad Sci. 2014;111(41):E4305-E4314. doi:10.1073/pnas.1412289111.

49. Amie SM, Bambara RA, Kim B. GTP Is the Primary Activator of the Anti-HIV Restriction Factor SAMHD1. J Biol Chem. 2013;288(35):25001-25006. doi:10.1074/jbc.C113.493619.

50. Li Y, Kong J, Peng X, Hou W, Qin X, Yu X-F. Structural insights into the high-efficiency catalytic mechanism of the sterile α-motif/histidineaspartate domain-containing protein. J Biol Chem. 2015;290(2):jbc.M115.663658. doi:10.1074/jbc.M115.663658.

51. Koharudin LMI, Wu Y, DeLucia M, Mehrens J, Gronenborn AM, Ahn J. Structural Basis of Allosteric Activation of Sterile Motif and Histidine-Aspartate Domain-containing Protein 1 (SAMHD1) by Nucleoside Triphosphates. J Biol Chem. 2014;289(47):32617- 32627. doi:10.1074/jbc.M114.591958.

52. Wang Z, Bhattacharya A, Villacorta J, Diaz-Griffero F, Ivanov DN. Allosteric activation of SAMHD1 protein by deoxynucleotide triphosphate (dNTP)-dependent tetramerization requires dNTP concentrations that are similar to dNTP concentrations observed in cycling T cells. J Biol Chem. 2016;291(41):21407-21413. doi:10.1074/jbc.C116.751446.

53. Jang S, Zhou X, Ahn J. Substrate Specificity of SAMHD1 Triphosphohydrolase Activity Is Controlled by Deoxyribonucleoside Triphosphates and Phosphorylation at Thr592. Biochemistry. 2016;55(39):5635-5646. doi:10.1021/acs.biochem.6b00627.

54. Amie SM, Daly MB, Noble E, Schinazi RF, Bambara RA, Kim B. Anti-HIV host factor SAMHD1 regulates viral sensitivity to nucleoside reverse transcriptase inhibitors via modulation of cellular deoxyribonucleoside triphosphate (dNTP) levels. J Biol Chem.

2013;288(28):20683-20691. doi:10.1074/jbc.M113.472159.

55. Seamon KJ, Hansen EC, Kadina AP, et al. Small molecule inhibition of SAMHD1 dNTPase by tetramer destabilization. J Am Chem Soc. 2014;136(28):9822-9825. doi:10.1021/ja5035717.

56. Arnold LH, Groom HCT, Kunzelmann S, et al. Phospho-dependent Regulation of SAMHD1 Oligomerisation Couples Catalysis and Restriction. Hope TJ, ed. PLOS Pathog. 2015;11(10):e1005194. doi:10.1371/journal.ppat.1005194.

57. Mauney C, Rogers L, Harris R, et al. The SAMHD1 dNTP triphosphohydrolase is controlled by a redox switch. Antioxid Redox Signal. 2017;(336):ars.2016.6888. doi:10.1089/ars.2016.6888.

58. Seamon KJ, Sun Z, Shlyakhtenko LS, Lyubchenko YL, Stivers JT. SAMHD1 is a single- stranded nucleic acid binding protein with no active site-associated nuclease activity. Nucleic Acids Res. 2015;43(13):6486-6499. doi:10.1093/nar/gkv633.

59. Goncalves A, Karayel E, Rice GI, et al. SAMHD1 is a nucleic-acid binding protein that is mislocalized due to aicardi-goutières syndrome-associated mutations. Hum Mutat. 2012;33(7):1116-1122. doi:10.1002/humu.22087.

60. White TE, Brandariz-Nuñez A, Valle-Casuso JC, et al. Contribution of SAM and HD domains to retroviral restriction mediated by human SAMHD1. Virology. 2013;436(1):81-90. doi:10.1016/j.virol.2012.10.029.

61. Beloglazova N, Flick R, Tchigvintsev A, et al. Nuclease activity of the human SAMHD1 protein implicated in the aicardi-gouti??res syndrome and HIV-1 restriction. J Biol Chem. 2013;288(12):8101-8110. doi:10.1074/jbc.M112.431148.

62. Tüngler V, Staroske W, Kind B, et al. Single-stranded nucleic acids promote SAMHD1 complex formation. J Mol Med. 2013;91(6):759-770. doi:10.1007/s00109-013-0995- 3.

63. Seamon KJ, Bumpus NN, Stivers JT. Single-Stranded Nucleic Acids Bind to the Tetramer Interface of SAMHD1 and Prevent Formation of the Catalytic Homotetramer. Biochemistry. 2016;55(44):6087-6099. doi:10.1021/acs.biochem.6b00986.

64. Ryoo J, Choi J, Oh C, et al. The ribonuclease activity of SAMHD1 is required for HIV-1 restriction. Nat Med. 2014;20(8):936-941. doi:10.1038/nm.3626.

65. Choi J, Ryoo J, Oh C, Hwang S, Ahn K. SAMHD1 specifically restricts retroviruses through its RNase activity. Retrovirology. 2015;12(1):46. doi:10.1186/s12977-015- 0174-4.

66. Ryoo J, Hwang SY, Choi J, Oh C, Ahn K. SAMHD1, the Aicardi-Gouti??res syndrome gene and retroviral restriction factor, is a phosphorolytic ribonuclease rather than a hydrolytic ribonuclease. Biochem Biophys Res Commun. 2016;477(4):977-981. doi:10.1016/j.bbrc.2016.07.013.

67. Antonucci JM, St Gelais C, de Silva S, et al. SAMHD1-mediated HIV-1 restriction in cells does not involve ribonuclease activity. Nat Publ Gr. 2016;22(10).

doi:10.1038/nm.4163.

68. Cribier A, Descours B, Valadão A, Laguette N, Benkirane M. Phosphorylation of SAMHD1 by Cyclin A2/CDK1 Regulates Its Restriction Activity toward HIV-1. Cell Rep. 2013;3(4):1036-1043. doi:10.1016/j.celrep.2013.03.017.

69. White TE, Brandariz-Nuñez A, Valle-Casuso JC, et al. The Retroviral Restriction Ability of SAMHD1, but Not Its Deoxynucleotide Triphosphohydrolase Activity, Is Regulated by Phosphorylation. Cell Host Microbe. 2013;13(4):441-451. doi:10.1016/j.chom.2013.03.005.

70. Wittmann S, Behrendt R, Eissmann K, et al. Phosphorylation of murine SAMHD1 regulates its antiretroviral activity. Retrovirology. 2015;12(1):103. doi:10.1186/s12977-015-0229-6.

71. St. Gelais C, de Silva S, Hach JC, et al. Identification of Cellular Proteins Interacting with the Retroviral Restriction Factor SAMHD1. J Virol. 2014;88(13):7689-7689. doi:10.1128/JVI.01089-14.

72. Pauls E, Badia R, Torres-Torronteras J, et al. Palbociclib, a selective inhibitor of cyclin-dependent kinase4/6, blocks HIV-1 reverse transcription through the control of sterile alpha motif and HD domain-containing protein-1 (SAMHD1) activity. Aids. 2014;28(15):2213-2222. doi:10.1097/QAD.0000000000000399.

73. Pauls E, Ruiz A, Riveira-Muñoz E, et al. p21 regulates the HIV-1 restriction factor SAMHD1. Proc Natl Acad Sci U S A. 2014;111(14):E1322-4. doi:10.1073/pnas.1322059111.

74. Ruiz A, Pauls E, Badia R, et al. Cyclin D3-dependent control of the dNTP pool and HIV- 1 replication in human macrophages. Cell Cycle. 2015;14(11):1657-1665. doi:10.1080/15384101.2015.1030558.

75. Pauls E, Jimenez E, Ruiz A, et al. Restriction of HIV-1 replication in primary macrophages by IL-12 and IL-18 through the upregulation of SAMHD1. J Immunol. 2013;190(9):4736-4741. doi:10.4049/jimmunol.1203226.

76. Tang C, Ji X, Wu L, Xiong Y. Impaired dNTPase activity of SAMHD1 by phosphomimetic mutation of Thr-592. J Biol Chem. 2015;290(44):26352-26359. doi:10.1074/jbc.M115.677435.

77. Yan J, Hao C, DeLucia M, et al. CyclinA2-Cyclin-dependent Kinase Regulates SAMHD1 Protein Phosphohydrolase Domain. J Biol Chem. 2015;290(21):13279-13292. doi:10.1074/jbc.M115.646588.

78. Pauls E, Ruiz A, Badia R, et al. Cell Cycle Control and HIV-1 Susceptibility Are Linked by CDK6-Dependent CDK2 Phosphorylation of SAMHD1 in Myeloid and Lymphoid Cells. J Immunol. 2014;193(4):1988-1997. doi:10.4049/jimmunol.1400873.

79. Welbourn S, Dutta SM, Semmes OJ, Strebel K. Restriction of Virus Infection but Not Catalytic dNTPase Activity Is Regulated by Phosphorylation of SAMHD1. J Virol. 2013;87(21):11516-11524. doi:10.1128/JVI.01642-13.

80. Bhattacharya A, Wang Z, White T, et al. Effects of T592 phosphomimetic mutations on

tetramer stability and dNTPase activity of SAMHD1 can not explain the retroviral restriction defect. Sci Rep. 2016;6(April):31353. doi:10.1038/srep31353.

81. Poole LB. The basics of thiols and cysteines in redox biology and chemistry. Free Radic Biol Med. 2015;80:148-157. doi:10.1016/j.freeradbiomed.2014.11.013.

82. Gramberg T, Kahle T, Bloch N, et al. Restriction of diverse retroviruses by SAMHD1. Retrovirology. 2013;10(1):26. doi:10.1186/1742-4690-10-26.

83. Laguette N, Sobhian B, Casartelli N, et al. SAMHD1 is the dendritic- and myeloid-cell- specific HIV-1 restriction factor counteracted by Vpx. Nature. 2011;474(7353):654- 657. doi:10.1038/nature10117.

84. Lahouassa H, Daddacha W, Hofmann H, et al. SAMHD1 restricts the replication of human immunodeficiency virus type 1 by depleting the intracellular pool of deoxynucleoside triphosphates. Nat Immunol. 2012;13(3):223-228. doi:10.1038/ni.2236.

85. Baldauf H-M, Pan X, Erikson E, et al. SAMHD1 restricts HIV-1 infection in resting CD4+ T cells. Nat Med. 2012;18(11):1682-1689. doi:10.1038/nm.2964.

86. Diamond TL, Roshal M, Jamburuthugoda VK, et al. Macrophage tropism of HIV-1 depends on efficient cellular dNTP utilization by reverse transcriptase. J Biol Chem. 2004;279(49):51545-51553. doi:10.1074/jbc.M408573200.

87. Descours B, Cribier A, Chable-Bessia C, et al. SAMHD1 restricts HIV-1 reverse transcription in quiescent CD4+ T-cells. Retrovirology. 2012;9(1):87. doi:10.1186/1742-4690-9-87.

88. Riess M, Fuchs N V., Idica A, et al. Interferons Induce Expression of SAMHD1 in Monocytes through Down-regulation of miR-181a and miR-30a. J Biol Chem. 2017;292(1):264-277. doi:10.1074/jbc.M116.752584.

89. Kim B, Nguyen LA, Daddacha W, Hollenbaugh JA. Tight Interplay among SAMHD1 Protein Level, Cellular dNTP Levels, and HIV-1 Proviral DNA Synthesis Kinetics in Human Primary Monocyte-derived Macrophages. J Biol Chem. 2012;287(26):21570- 21574. doi:10.1074/jbc.C112.374843.

90. Hrecka K, Hao C, Gierszewska M, et al. Vpx relieves inhibition of HIV-1 infection of macrophages mediated by the SAMHD1 protein. Nature. 2011;474(7353):658-661. doi:10.1038/nature10195.

91. Amie SM, Noble E, Kim B. Intracellular nucleotide levels and the control of retroviral infections. Virology. 2013;436(2):247-254. doi:10.1016/j.virol.2012.11.010.

92. Beauchamp BB, Richardson CC. A unique deoxyguanosine triphosphatase is responsible for the optA1 phenotype of Escherichia coli. Proc Natl Acad Sci U S A. 1988;85(8):2563-2567. http://www.ncbi.nlm.nih.gov/pubmed/2833745%5Cnhttp://www.pubmedcentral. nih.gov/articlerender.fcgi?artid=PMC280037.

93. Hofmann H, Logue EC, Bloch N, et al. The Vpx Lentiviral Accessory Protein Targets SAMHD1 for Degradation in the Nucleus. J Virol. 2012;86(23):12552-12560.

doi:10.1128/JVI.01657-12.

94. Pertel T, Reinhard C, Luban J. Vpx rescues HIV-1 transduction of dendritic cells from the antiviral state established by type 1 interferon. Retrovirology. 2011;8(1):49. doi:1742-4690-8-49 [pii]\r10.1186/1742-4690-8-49 [doi].

95. Ahn J, Hao C, Yan J, et al. HIV/Simian Immunodeficiency Virus (SIV) accessory virulence factor Vpx loads the host cell restriction factor SAMHD1 onto the E3 ubiquitin ligase complex CRL4 DCAF1. J Biol Chem. 2012;287(15):12550-12558. doi:10.1074/jbc.M112.340711.

96. Schwefel D, Groom HCT, Boucherit VC, et al. Structural basis of lentiviral subversion of a cellular protein degradation pathway. Nature. 2014;505(7482):234-238. doi:10.1038/nature12815.

97. Brandariz-Nuñez A, Valle-Casuso JC, White TE, et al. Contribution of oligomerization to the anti-HIV-1 properties of SAMHD1. Retrovirology. 2013;10(1):131. doi:10.1186/1742-4690-10-131.

98. Welbourn S, Strebel K. Low dNTP levels are necessary but may not be sufficient for lentiviral restriction by SAMHD1. Virology. 2016;488:271-277. doi:10.1016/j.virol.2015.11.022.

99. Ayinde D, Bruel T, Cardinaud S, et al. SAMHD1 limits HIV-1 antigen presentation by monocyte-derived dendritic cells. J Virol. 2015;89(April):JVI.00069-15. doi:10.1128/JVI.00069-15.

100. Brégnard C, Benkirane M, Laguette N. DNA damage repair machinery and HIV escape from innate immune sensing. Front Microbiol. 2014;5(APR):1-10. doi:10.3389/fmicb.2014.00176.

101. Manel N, Hogstad B, Wang Y, Levy DE, Unutmaz D, Littman DR. A cryptic sensor for HIV-1 activates antiviral innate immunity in dendritic cells. Nature. 2010;467(7312):214-217. doi:10.1038/nature09337.

102. Maelfait J, Bridgeman A, Benlahrech A, Cursi C, Rehwinkel J. Restriction by SAMHD1 Limits cGAS/STING-Dependent Innate and Adaptive Immune Responses to HIV-1. Cell Rep. 2016;16(6):1492-1501. doi:10.1016/j.celrep.2016.07.002.

103. van Montfoort N, Olagnier D, Hiscott J. Unmasking immune sensing of retroviruses: Interplay between innate sensors and host effectors. Cytokine Growth Factor Rev. 2014;25(6):657-668. doi:10.1016/j.cytogfr.2014.08.006.

104. Hollenbaugh JA, Tao S, Lenzi GM, et al. dNTP pool modulation dynamics by SAMHD1 protein in monocyte-derived macrophages. Retrovirology. 2014;11(1):63. doi:10.1186/s12977-014-0063-2.

105. Franzolin E, Pontarin G, Rampazzo C, et al. The deoxynucleotide triphosphohydrolase SAMHD1 is a major regulator of DNA precursor pools in mammalian cells. Proc Natl Acad Sci. 2013;110(35):14272-14277. doi:10.1073/pnas.1312033110.

106. Bonifati S, Daly MB, St. Gelais C, et al. SAMHD1 controls cell cycle status, apoptosis and HIV-1 infection in monocytic THP-1 cells. Virology. 2016;495:92-100.

doi:10.1016/j.virol.2016.05.002.

107. Dragin L, Munir-Matloob S, Froehlich J, et al. Evidence that HIV-1 restriction factor SAMHD1 facilitates differentiation of myeloid THP-1 cells. Virol J. 2015;12(1):201. doi:10.1186/s12985-015-0425-y.

108. Kodigepalli KM, Li M, Liu S-L, Wu L. Exogenous Expression of SAMHD1 Inhibits Proliferation and Induces Apoptosis in Cutaneous T-cell Lymphoma-derived HuT78 Cells. Cell Cycle. 2016;4101(January 2017):00-00. doi:10.1080/15384101.2016.1261226.

109. Kretschmer S, Wolf C, Konig N, et al. SAMHD1 prevents autoimmunity by maintaining genome stability. Ann Rheum Dis. 2015;74(3):e17-e17. doi:10.1136/annrheumdis- 2013-204845.

110. Schmidt S, Schenkova K, Adam T, et al. SAMHD1’s protein expression profile in humans. J Leukoc Biol. 2015;98(1):5-14. doi:10.1189/jlb.4HI0714-338RR.

111. Wang J, Lu F, Shen X-Y, Wu Y, Zhao L. SAMHD1 is down regulated in lung cancer by methylation and inhibits tumor cell proliferation. Biochem Biophys Res Commun. 2014;455(3-4):229-233. doi:10.1016/j.bbrc.2014.10.153.

112. De Silva S, Hoy H, Hake TS, Wong HK, Porcu P, Wu L. Promoter methylation regulates SAMHD1 gene expression in human CD4 + T cells. J Biol Chem. 2013;288(13):9284- 9292. doi:10.1074/jbc.M112.447201.

113. De Silva S, Wang F, Hake TS, Porcu P, Wong HK, Wu L. Downregulation of SAMHD1 Expression Correlates with Promoter DNA Methylation in Sézary Syndrome Patients. J Invest Dermatol. 2013;134(August):562-565. doi:10.1038/jid.2013.311.

114. Kyei GB, Cheng X, Ramani R, Ratner L. Cyclin L2 is a critical HIV dependency factor in macrophages that controls samhd1 abundance. Cell Host Microbe. 2015;17(1):98- 106. doi:10.1016/j.chom.2014.11.009.

115. Crow YJ, Manel N. Aicardi–Goutières syndrome and the type I interferonopathies. Nat Rev Immunol. 2015;15(7):429-440. doi:10.1038/nri3850.

116. Giroud M, Gouyon JB, Chaumet F, et al. A case of progressive familial encephalopathy in infancy with calcification of the basal ganglia and chronic cerebrospinal fluid lymphocytosis. Child’s Nerv Syst. 1986;2(1):47-48. doi:10.1007/BF00274035.

117. Goutières F, Aicardi J, Barth PG, Lebon P. Aicardi-Goutières syndrome: an update and results of interferon-alpha studies. Ann Neurol. 1998;44(6):900-907. doi:10.1002/ana.410440608.

118. Crow YJ, Rehwinkel J. Aicardi-Goutie’res syndrome and related phenotypes: Linking nucleic acid metabolism with autoimmunity. Hum Mol Genet. 2009;18(R2):130-136. doi:10.1093/hmg/ddp293.

119. Ravenscroft JC, Suri M, Rice GI, Szynkiewicz M, Crow YJ. Autosomal dominant inheritance of a heterozygous mutation in SAMHD1 causing familial chilblain lupus. Am J Med Genet Part A. 2011;155(1):235-237. doi:10.1002/ajmg.a.33778.

120. Crow YJ, Chase DS, Schmidt JL, et al. Characterization of Human Disease Phenotypes Associated with Mutations in TR E X1 , RNAS E H2A ,. Am J Med Genet Part A. 2015:296-312.

121. Kunkel T a. Evolving Views of DNA Replication ( In ) Fidelity Evolving Views of DNA Replication ( In ) Fidelity OVERLAPPING FUNCTIONS. 2009;LXXIV:91-101. doi:10.1101/sqb.2009.74.027.

122. Zhao K, Du J, Han X, et al. Modulation of LINE-1 and Alu/SVA Retrotransposition by Aicardi-Gouti??res Syndrome-Related SAMHD1. Cell Rep. 2013;4(6):1108-1115. doi:10.1016/j.celrep.2013.08.019.

123. Vitiello L, Gorini S, Rosano G, La Sala A. Immunoregulation through extracellular nucleotides. Blood. 2012;120(3):511-518. doi:10.1182/blood-2012-01-406496.

124. Stetson DB. Endogenous retroelements and autoimmune disease. Curr Opin Immunol. 2012;24(6):692-697. doi:10.1016/j.coi.2012.09.007.

125. Gorini S, Gatta L, Pontecorvo L, Vitiello L, la Sala A. Regulation of innate immunity by extracellular nucleotides. Am J Blood Res. 2013;3(1):14-28. http://www.ncbi.nlm.nih.gov/pubmed/23358447%5Cnhttp://www.pubmedcentral .nih.gov/articlerender.fcgi?artid=PMC3555188.

126. Xin B, Jones S, Puffenberger EG, et al. Homozygous mutation in SAMHD1 gene causes cerebral vasculopathy and early onset stroke. Proc Natl Acad Sci U S A. 2011;108(13):5372-5377. doi:10.1073/pnas.1014265108.

127. Ramesh V, Bernardi B, Stafa A, et al. Intracerebral large artery disease in Aicardi- Gouti??res syndrome implicates SAMHD1 in vascular homeostasis. Dev Med Child Neurol. 2010;52(8):725-732. doi:10.1111/j.1469-8749.2010.03727.x.

128. Thiele H, Du Moulin M, Barczyk K, et al. Cerebral arterial stenoses and stroke: novel features of Aicardi-Goutières syndrome caused by the Arg164X mutation in SAMHD1 are associated with altered cytokine expression. Hum Mutat. 2010;31(11):1836- 1850. doi:10.1002/humu.21357.

129. Pontarin G, Ferraro P, Valentino ML, Hirano M, Reichard P, Bianchi V. Mitochondrial DNA depletion and thymidine phosphate pool dynamics in a cellular model of mitochondrial neurogastrointestinal encephalomyopathy. J Biol Chem. 2006;281(32):22720-22728. doi:10.1074/jbc.M604498200.

130. Bourdon A, Minai L, Serre V, et al. Mutation of RRM2B, encoding p53-controlled ribonucleotide reductase (p53R2), causes severe mitochondrial DNA depletion. Nat Genet. 2007;39(6):776-780. doi:10.1038/ng2040.

131. Roesch F, Schwartz O. The SAMHD1 knockout mouse model : in vivo veritas ? EMBO J. 2013;32(18):2427-2429. doi:10.1038/emboj.2013.190.

132. Rehwinkel J, Maelfait J, Bridgeman A, et al. SAMHD1-dependent retroviral control and escape in mice. EMBO J. 2013;32(18):2454-2462. doi:10.1038/emboj.2013.163.

133. Sohl CD, Ray S, Sweasy JB. Pools and Pols: Mechanism of a mutator phenotype. Proc Natl Acad Sci. 2015;112(19):201505169. doi:10.1073/pnas.1505169112.

134. Williams LN, Marjavaara L, Knowels GM, et al. dNTP pool levels modulate mutator phenotypes of error-prone DNA polymerase ε variants. Proc Natl Acad Sci U S A. 2015;112(19):E2457-66. doi:10.1073/pnas.1422948112.

135. Bester AC, Roniger M, Oren YS, et al. Nucleotide deficiency promotes genomic instability in early stages of cancer development. Cell. 2011;145(3):435-446. doi:10.1016/j.cell.2011.03.044.

136. Pajalunga D, Franzolin E, Stevanoni M, et al. A defective dNTP pool hinders DNA replication in cell cycle-reactivated terminally differentiated muscle cells. Cell Death Differ. 2017:1-11. doi:10.1038/cdd.2017.4.

137. Rentoft M, Lindell K, Tran P, et al. Heterozygous colon cancer-associated mutations of SAMHD1 have functional significance. Proc Natl Acad Sci. 2016;113(17):4723-4728. doi:10.1073/pnas.1519128113.

138. Clifford R, Louis T, Robbe P, et al. SAMHD1 is mutated recurrently in chronic lymphocytic leukemia and is involved in response to DNA damage. Blood. 2014;123(7):1021-1031. doi:10.1182/blood-2013-04-490847.

139. Kohnken R, Kodigepalli KM, Wu L. Regulation of deoxynucleotide metabolism in cancer: novel mechanisms and therapeutic implications. Mol Cancer. 2015;14:176. doi:10.1186/s12943-015-0446-6.

140. Mathews CK. Oxidized deoxyribonucleotides, mutagenesis, and cancer. FASEB J. 2017;31(1):11-13. doi:10.1096/fj.201601100.

141. Parker WB. Enzymology of Purine and Pyrimidine Antimetabolites Used in the Treatment of Cancer. Chem Rev. 2009;109(7):2880-2893. doi:10.1021/cr900028p.

142. Schneider C, Oellerich T, Baldauf H-M, et al. SAMHD1 is a biomarker for cytarabine response and a therapeutic target in acute myeloid leukemia. Nat Med. 2016;(i). doi:10.1038/nm.4255.

143. Herold N, Rudd SG, Ljungblad L, et al. Targeting SAMHD1 with the Vpx protein to improve cytarabine therapy for hematological malignancies. Nat Med. 2017;23(2):256-263. doi:10.1038/nm.4265.

144. Seamon KJ, Stivers JT. A High-Throughput Enzyme-Coupled Assay for SAMHD1 dNTPase. J Biomol Screen. 2015;20(6):801-809. doi:10.1177/1087057115575150.

145. Arnold LH, Kunzelmann S, Webb MR, Taylor IA. A Continuous Enzyme-Coupled Assay for Triphosphohydrolase Activity of HIV-1 Restriction Factor SAMHD1. Antimicrob Agents Chemother. 2015;59(1):186-192. doi:10.1128/AAC.03903-14.

CHAPTER 1

The SAMHD1 dNTP triphosphohydrolase is controlled by a redox switch

Christopher H. Mauney1, LeAnn C. Rogers1, Reuben S. Harris4, Larry W. Daniel1,2, Nelmi O. Devarie-Baez3, Hanzhi Wu3, Cristina M. Furdui2,3, Leslie B. Poole1,2, Fred W. Perrino1, Thomas Hollis1,2*

1Department of Biochemistry, Center for Structural Biology, 2Center for Molecular Communication and Signaling, 3Dept of Internal Medicine, Section on Molecular Medicine, Wake Forest School of Medicine, Winston-Salem, NC 4 Howard Hughes Medical Institute, Department of Biochemistry, Molecular Biology and Biophysics, Masonic Cancer Center, and Institute for Molecular Virology, University of Minnesota, Minneapolis, MN

* Correspondence should be addressed to: [email protected] ph: 336-716-0768

Published in Antioxidants and Redox Signaling, Feb 2017. doi: 10.1089/ars.2016.6888

Keywords: SAMHD1, protein oxidation, HIV restriction, Aicardi-Goutieres syndrome, redox switch, enzyme catalysis.

ABSTRACT

Aims: Proliferative signaling involves reversible posttranslational oxidation of proteins.

However, relatively few molecular targets of these modifications have been identified. We investigate the role of protein oxidation in regulation of SAMHD1 catalysis.

Results: Here we report that SAMHD1 is a major target for redox regulation of nucleotide metabolism and cell cycle control. SAMHD1 is a triphosphate hydrolase, whose function involves regulation of dNTPs pools. We demonstrate that the redox state of SAMHD1 regulates its catalytic activity. We have identified three cysteine residues that constitute an intrachain disulfide bond ‘redox switch’ that reversibly inhibits protein tetramerization and catalysis. We show that proliferative signals lead to SAMHD1 oxidation in cells and oxidized

SAMHD1 is localized outside of the nucleus.

Innovation and Conclusions: SAMHD1 catalytic activity is reversibly regulated by protein oxidation. These data identify a previously unknown mechanism for regulation of nucleotide metabolism by SAMHD1.

INTRODUCTION

Regulation of cellular dNTP levels is essential for proper DNA synthesis and cell cycle progression. SAMHD1 hydrolyzes deoxynucleotide triphosphates (dNTPs)1,2, to control dNTP pool concentrations 3,4. Dysregulation of nucleotide metabolism can have dangerous biological consequences, and mutations in the SAMHD1 gene result in the development of familial chilblain lupus and Aicardi-Goutieres syndrome, a severe neurological autoimmune disease that resembles a congenital viral infection 5–7. SAMHD1 mutations have also been identified as a contributing factor to several types of cancers 8–10. SAMHD1 has garnered significant attention for its antiviral properties in non-dividing hematopoietic cells, where it acts as an interferon inducible host restriction factor against viral infections, including HIV

11–19. SAMHD1 is proposed to inhibit productive viral infection by maintaining dNTP levels below the effective concentration for viral DNA polymerases thereby restricting viral DNA synthesis 20–22. However, SAMHD1 does not restrict HIV in actively dividing cells 23–25, implying a strong connection between its catalytic activity and cell cycle progression, but the mechanisms of SAMHD1 catalytic regulation have remained an open question.

Structural and biochemical studies of SAMHD1 have elucidated the precise mechanism of catalysis. Binding of deoxynucleotides in a regulatory site induces conformational rearrangements of the enzyme that result in tetramerization and catalytic activation 26–30.

This finely tuned autoregulation, in which substrate also acts as an activating ligand, enables SAMHD1 to sense fluctuating dNTP concentrations and respond according to the specific metabolic requirements of the cell. Efficient and accurate DNA synthesis and repair requires strict control of dNTP concentrations, which peak during S-phase when large quantities of dNTPs are required for DNA replication 31. As a central effector of nucleotide

concentration SAMHD1 is likely responsive to regulatory pathways that govern cell cycle progression 32. Elucidating the mechanisms that regulate SAMHD1 activity is critical for defining its function at the interface of nucleotide metabolism, viral replication, and autoimmunity.

Protein oxidation is emerging as a crucial signaling mechanism in cellular metabolism, cell signaling, and cell cycle progression 33–35. Here we identify three critical cysteine residues of

SAMHD1 (Cys341, Cys350, and Cys522) that create a ‘redox switch’ through the formation of intrachain disulfide bonds to reversibly inhibit SAMHD1 tetramerization and dNTPase activity. We also demonstrate that SAMHD1 is oxidized in cells in response to proliferative signals and colocalizes with sites of protein oxidation outside of the nucleus. Lastly, our data support a model in which Cys522 acts as the primary sensor of redox signals and forms a disulfide linkage with either Cys341 or Cys350 that destabilizes the binding of activating nucleotides, thereby inhibiting tetramerization and dNTP hydrolase activity.

RESULTS

The dNTPase activity of SAMHD1 is reversibly inhibited by oxidation. We investigated the impact of oxidation and reduction on SAMHD1 dNTP triphosphohydrolase activity.

Oxidation of SAMHD1 by H2O2 results in a significant decrease in SAMHD1 activity by 10μM

H2O2 and greater than 90% decrease in activity by 100μM (Fig. 1A, Supplemental Fig 1). The addition of an excess of reducing agents (10mM DTT/2mM TCEP) to the samples restores

Fig 1. SAMHD1 is reversibly inhibited by H2O2 oxidation. A) SAMHD1 deoxynucleotide hydrolase activity is decreased in the presence of increasing H2O2, but is rescued after reduction of the protein with DTT/TCEP. B, C) Kinetics of inhibition by H2O2 and Hanes-Woolf plot show a decrease in Vmax with relatively constant Km most consistent with a non-competitive mechanism of inhibition. D) Cellular reducing agents glutathione, cysteine, and thioredoxin-thioredoxin reductase are able to reactivate SAMHD1 after oxidation. Error values for activity measurements (A & D) represent the standard deviation (*p<0.05, **p<0.01, ***p<0.001) and kinetic plots (B & C) are standard error of the mean.

activity with no significant difference between the untreated sample and any of the concentrations of H2O2 tested.

A more detailed investigation of the effect of oxidation on the kinetics of SAMHD1 dNTPase activity reveals a similar effect. SAMHD1 was incubated with increasing concentrations of

H2O2, followed by addition of 100μM dATPαS and GTP. The nucleotides dATPαS and GTP induce the activated tetramer, but are not substrates for SAMHD1 and are not consumed in the reaction. Instead, they serve to maintain a steady state of enzymatic activation in the samples. Using this approach, SAMHD1 exhibits standard Michaelis-Menten kinetics, with a progressive decrease in turnover number in response to increasing H2O2 concentrations

(Fig 1B). An increase in H2O2 concentration results in a decrease of Vmax while the Km is relatively unaffected. These data indicate that the reaction is inhibited by H2O2 in a manner consistent with non-competitive inhibition (Fig 1C). Taken together, these results reveal that H2O2 inhibits the catalytic capability of SAMHD1 through a mechanism that is both reversible and non-competitive, which is consistent with oxidation of the enzyme.

We next tested the ability of common cellular reductants to rescue SAMHD1 activity following oxidation. Glutathione and cysteine restored the dNTPase activity of oxidized

SAMHD1 (Fig 1D). In addition, thioredoxin / thioredoxin reductase rescued SAMHD1 dNTPase activity (Fig 1D) demonstrating cellular reduction systems are capable of reducing oxidized SAMHD1.

Oxidation of SAMHD1 inhibits dNTPase activity by preventing subunit oligomerization. The triphosphate hydrolase activity of SAMHD1 is dependent on the

Fig 2. Oxidation of SAMHD1 prevents protein oligomerization. A) Incubation of SAMHD1 with GTP-dATPαS to form the tetramer protects the enzyme activity from inactivation by oxidation (*p<0.05, **p<0.01, ***p<0.001 compared to respective 0μM H2O2). B) Dynamic light scattering measurements reveal SAMHD1 remains a monomer/dimer when oxidized, but can form tetramer after being reduced. C) Preincubation with activating nucleotides prior to addition of H2O2 stimulates the formation of SAMHD1 tetramer. D) dNTPase activity of SAMHD1 proteins used under conditions of the light scattering experiments show oxidized protein is unable to form a tetramer and has reduced activity. E) Estimated hydrodynamic radii and molecular weights from dynamic light scattering experiments in B and C.

ability of the protein to form a tetramer 28,36,37. To interrogate the mechanism of oxidative inactivation of SAMHD1 activity we looked at the effects of oxidation on protein oligomerization. SAMHD1 was preincubated with either GTP alone or GTP and dATPαS. dNTPase activity was then measured following SAMHD1 oxidation. GTP and dATPαS together will induce tetramer formation but will not be consumed in the reaction, while GTP alone will bind the guanine specific regulatory site but not induce tetramerization 26,38. Both

GTP and GTP-dATPαS samples display significantly increased dNTPase activity when compared to samples not preincubated with activating nucleotides (Fig 2A). However, while

GTP-only samples appear primed for dNTPase activity resulting in an increase in activity compared to the control, they exhibit a similar oxidative inactivation as untreated samples.

Contrary to untreated and GTP-only samples, samples preincubated with GTP-dATPαS, show no loss of activity even at the highest level of protein oxidation tested. This implies that nucleotide binding protects SAMHD1 from the effects of oxidation, either by shielding critical residues or through the formation of the tetrameric species.

Next we used dynamic light scattering to evaluate the formation of SAMHD1 tetramers as a function of protein oxidation. Dynamic light scattering provides a measure of hydrodynamic diameter (DH) and is able to differentiate the oligomeric states of SAMHD1. The average DH of untreated SAMHD1 measures 8.9 nm (Fig 2B, 2E), which translates to a weight-averaged molecular weight of 110 kDa, and represents the dynamic monomer-dimer equilibrium of

SAMHD1 observed in solution without activating nucleotides present (SAMHD1 monomer =

72.2 kDa). As expected, the addition of activating nucleotides results in the formation of the tetrameric species (DH = 13.5 nm, Mw = 292 kDa). Oxidation of SAMHD1 inhibits formation of the tetrameric species in the presence of activating nucleotides (DH = 9.9 nm, Mw = 142 kDa), and instead appears to sequester the enzyme in the dimeric state. The addition of reducing agent prior to protein oxidation mitigates its inhibitory effect and allows for enzyme tetramerization (DH = 13.1 nm, Mw = 274 kDa). Preincubation with activating nucleotides prior to protein oxidation also facilitates enzyme tetramerization supporting the hypothesis that tetramerization shields SAMHD1 from the inhibitory effects of oxidation

(Fig 2C). The phosphohydrolase activities of samples used in the light scattering experiments were tested and show results similar to those in Fig 2A correlating with the

ability of SAMHD1 to form tetramers (Fig 2D). Alternative experimental approaches investigating the oligomeric state of SAMHD1 under conditions similar to those used in the

DLS experiments returned consistent results. Both glutaraldehyde cross linking resolved by western blotting, and analytical size exclusion chromotography linked to multiangle light scattering detectors (SEC-MALS) reveal a reduction of SAMHD1 tetramerization in the presence of H2O2 (Supplemental Fig 2).

SAMHD1 contains multiple redox sensitive cysteines that form disulfide bonds when oxidized. ROS mediated enzymatic inactivation often occurs through the oxidation of specific cysteinyl thiolates to sulfenic acid (R-SOH), which can further form various reversible or irreversible chemical modifications to the enzyme 39. The reversible nature of

SAMHD1’s catalytic inhibition by micromolar levels of H2O2 indicated the potential for a similar cysteine mediated mechanism for enzymatic inactivation. SAMHD1 contains thirteen cysteine residues, and a review of SAMHD1 structures available in the Protein Data Base reveals that three cysteines (Cys 341, 350, and 522) are in proximity to each other and adjacent to the regulatory nucleotide binding pockets (Fig 3A). Residue Cys522 is in a flexible and often unstructured loop while Cys341 and Cys350 form a disulfide bond in at least four structures (5AO4, 4RXO, 4MZ7, 3U1N) 2,26,37,40, which is absent in the remainder of published structures. Additionally, in a primary sequence alignment of SAMHD1 from various vertebrates, these cysteine residues are all highly conserved (Supplemental Fig 3).

Analysis of residue specific redox sensitivity was carried out using mass spectrometry 41.

Protein S-sulfenylation, the conversion of a cysteinyl thiolate to sulfenic acid, is a common first step in the oxidative signaling mechanism that leads to the formation of a disulfide bond or a sulfinic (R-SO2H) or sulfonic (R-SO3H) acid chemical moiety 42,43. The reagent 5,5-

Fig 3. Identification of SAMHD1 redox sensitive cysteines. A) Structure of SAMHD1 monomer showing position of C341, C350, and C522 adjacent to the regulatory nucleotide binding site. (pdbid: 4MZ7)37 B) Mass spectrometry results summarizing chemical modifications to indicated SAMHD1 cysteine residues under varying redox conditions (IAA; iodoacetamide). C) Mass Matrix heat map representations of probable disulfide pairs from MS/MS data.

dimethyl-1,3-cyclohexanedione (dimedone) covalently reacts with cysteine sulfenic acids, and provides a means of identifying cysteine residues sensitive to redox signals. When

SAMHD1 is oxidized in the presence of dimedone, multiple labeled cysteine residues are identified indicating the formation of sulfenic acid (Fig 3B, Supplemental Fig 4).

Additionally, SO2 and SO3 chemical moieties are observed on these three residues under all conditions indicating their sensitivity to redox chemistry.

Tandem mass spectrometry search algorithms were also used to identify the disulfide bonds present in SAMHD1 samples oxidized in the absence of dimedone. Computational algorithms (Mass Matrix 44 and p-linkSS 45) were used to search the experimental mass

spectra of SAMHD1 following oxidizing treatments. Both identified Cys341-Cys350 as the predominant intrachain disulfide bond formed by SAMHD1 (Fig 3C, Supplemental Fig 5).

Several other disulfide linkages were identified with lower scores. Many of these also involved at least one cysteine from the Cys341, Cys350, Cys522 triad, providing further evidence for the redox sensitivity of these particular residues, and hinting at a possible dynamic thiol-disulfide exchange mechanism.

Cysteines 341, 350, and 522 are the primary regulators of redox sensitivity and are important for catalytic function. We examined the role of identified cysteine residues in modulating the redox sensitivity of SAMHD1. Cysteine to alanine mutants of SAMHD1 were tested in vitro for redox sensitivity. We also tested an N-terminal deletion of SAMHD1 that contains only the catalytic HD domain (residues 120-626) (HD1Δ). The C522A mutant

SAMHD1 shows no significant inhibition by oxidation (Fig 4A). Consistent with this, dynamic light scattering experiments using the C522A mutant SAMHD1 protein show that oxidation does not inhibit tetramerization, unlike the wild-type protein (Fig 4B), implying that C522 plays a critical role in sensing redox signals and translating them into a structural message that inhibits tetramerization and subsequent SAMHD1 catalysis. The other mutants tested are inactivated by oxidation, suggesting that alone they are not responsible for the mechanism of inactivation of SAMHD1 (Fig 4A). While the inhibitory effect of oxidation on the HD1Δ enzyme is not as dramatic as in wild-type, presumably due to the enhanced catalytic capacity of HD1Δ, the inhibition of the HD1Δ enzyme is significant as it localizes the primary mechanism for oxidative inactivation to the cysteine residues in the catalytic HD domain.

Fig 4. Cysteines 341, 350, and 522 are the primary regulators of redox sensitivity. A) Mutational analysis of SAMHD1 shows C522A mutant protein is resistant to H2O2 oxidation (*p<0.05, **p<0.01, ***p<0.001). B) Dynamic light scattering reveals C522A SAMHD1 protein can form a tetramer after H2O2 treatment. C) SAMHD1 proteins separated by non-reducing SDS-PAGE show the formation of a disulfide species after H2O2 exposure. The disulfide bond is removed by reducing agents (DTT/TCEP) or blocked by preincubation with activating nucleotides. The C522A SAMHD1 and C341A,C350A SAMHD1 show no disulfide species.

One commonly observed manifestation of redox signals translated into structural motifs that alter enzyme activity is the formation of an intrachain disulfide bond which can be observed by differential migration patterns using non-reducing denaturing gel electrophoresis 42,46. SAMHD1 wild type and Cys mutants were exposed to either increasing concentrations of H2O2, or some combination of H2O2, reducing agent, and activating nucleotides. The reaction products separated on non-reducing SDS-PAGE clearly show the emergence of a disulfide-containing band in response to increasing H2O2 in the wild-type protein (Fig 4C). In the presence of reducing agent or preincubation with activating nucleotides, the oxidized SAMHD1 band is not observed, implying that either of these agents can prevent formation of the disulfide-containing species. C341A and C350A SAMHD1 proteins also form a disulfide bond upon oxidation, with the C350A mutant appearing to be more susceptible to disulfide formation than the wild-type. These results are interesting because they reveal that the disulfide linkage is not strictly between Cys341-Cys350 as is

observed in crystal structures. The C522A and C341-350A mutants provide insight into a potential molecular mechanism. In neither of these mutant proteins is a disulfide species observed, indicating that the disulfide linkage does indeed consist of some arrangement of the three Cys residues (341, 350, and 522), and that Cys522 can form a disulfide linkage with either Cys341 or Cys350. In the 500μM H2O2 sample, the C341-C350A monomer is less pronounced, with much of the protein forming a high molecular weight aggregate that does not migrate into the gel. This is consistent with our experience that the C341-350A mutant rapidly aggregates in the absence of reducing agent (size exclusion chromatography data not shown). Circular dichroism analysis suggests that this may in part be related to altered structural stability stemming from a reduction in alpha-helix content (Supplemental Fig 6).

This underscores the importance of Cys341 and Cys350 for protein stability while hinting at a protective function against nonspecific Cys522 oxidation. As a control the C320A mutant, which is structurally distant from the Cys522, Cys341, C350 triad, shows the formation of disulfide bonds similar to wild-type.

SAMHD1 is oxidized in vivo in response to proliferative signals. Next, we looked for evidence of SAMHD1 oxidation in cells. We employed an established system of treating PC3 prostate cancer cells, RWPE-1 prostate epithelial cells, and WI-38 lung fibroblast cells with lysophosphatidic acid (LPA), R1881, and PDGF, respectively, to elicit intracellular production of ROS 47–49. Oxidized proteins were labeled with a dimedone-like probe conjugated to biotin (DCP-Bio1) and affinity captured using Streptavidin beads 50. Labeled proteins and the respective total lysate were immunoblotted and probed with an anti-

SAMHD1 antibody. Following addition of the proliferative agent, an oxidized species of

SAMHD1 is observed, demonstrating that SAMHD1 is indeed oxidized in intact cells (Fig

5A). Interestingly, the oxidation of SAMHD1 follows a specific time course in each cell

Fig 5. SAMHD1 is oxidized in cells in response to growth stimuli. A) Oxidized proteins in PC3, RWPE, and WI-38 cells stimulated with LPA, R1881, and PDGF, respectively, were labeled with DCP- Bio1 and isolated with streptavidin beads. Captured protein was resolved by SDS-PAGE, blotted onto nitrocellulose and probed with α-SAMHD1 antibodies. SAMHD1 oxidation increases post stimulation. Oxidized SAMHD1 and total SAMHD1 in the lysate are indicated. AhpC was used as a procedural control. B) Molecular co-localization of SAMHD1 and protein oxidation was visualized using a proximity ligation assay as described. The oxidized species of SAMHD1 (red signal) is present outside the nucleus. C) Upon stimulation with LPA, total SAMHD1, imaged by immunofluorescence, moves outside the nucleus over time and colocalizes with areas of oxidized proteins labeled by DCP-Rho1. The mean fluorescence intensity of DAPI, cytoplasmic SAMHD1, and DCP-Rho1 was measured in about 100 cells at each time point, showing an increase in cytoplasmic SAMHD1. Manders’ correlation coefficient representing the fraction of total cytoplasmic DCP-Rho1 that colocalizes with pixels containing SAMHD1 signal indicates an increasing colocalization of SAMHD1 with DCP-Rho1. Error bars represent standard error of the mean.

type. The signal initially increases following stimulation before the oxidation intermediate disappears around 60-90 minutes post stimulation. The total amount of SAMHD1 in whole cell lysate does not change. Further investigation of SAMHD1 activity in vivo provides additional evidence for the importance of oxidation as a means for catalytic regulation. We directly measured SAMHD1 catalytic activity from fibroblast cells transfected with SAMHD1 wt, SAMHD1 C522A, or empty vector and ± PDGF treatment, a known activator of ROS production. SAMHD1 activity in cells expressing wt protein is diminished after treatment with PDGF when compared to no PDGF (Supplemental Fig 7). However, the C522A expressing cells did not show any reduction in dNTPase activity following PDGF treatment.

These data indicate that SAMHD1 oxidation in cells reduces its catalytic activity and that the

C522A mutation is resistant to oxidation and therefore maintains catalytic activity following growth factor treatment.

We also performed fluorescence microscopy experiments to investigate the oxidation of

SAMHD1 in cells. These experiments provided visual evidence of the spatial localization of

SAMHD1 and oxidized proteins following the application of ROS-inducing growth factors.

PC3 cells treated with LPA and labeled with DCP-Bio1 were imaged using a proximity ligation assay (PLA). In the PLA assay fixed cells are incubated with anti-SAMHD1 antibodies and anti-biotin antibodies. Secondary antibodies covalently linked to half of a reporter construct are then added and the signal is amplified. A fluorescent signal is observed in sites where both SAMHD1 and labeled cysteines are in proximity, implying that the oxidized protein is SAMHD1. Following LPA incubation in PC3 cells, the fluorescence intensity increases 1.4x over the control samples at the ten minute time point, and gradually diminishes through the 60 minute time point (Fig 5B). These results provide further evidence of SAMHD1 oxidation in vivo, as well as localizes the oxidation of protein to the

cytoplasm. Although, it is unclear if oxidation of SAMHD1 is the cause of protein localization or if it happens in the cytoplasm after relocalization.

A similar experiment was performed using immunofluorescent microscopy to visualize

SAMHD1 colocalization with sites of oxidation. PC3 cells were treated as indicated, and cysteine-SOH were labeled with a dimedone-based probe conjugated to the fluorophore rhodamine (DCP-Rho1). Cells were fixed and probed with an anti-SAMHD1 antibody, and stained with DAPI. Statistical analysis of these cells using colocalization software reveals a correlation between SAMHD1 and DCP-Rho1 (a marker for cysteine sulfenic acid) that peaks at 30 minutes (Fig 5C). Additionally, we noticed a peculiar trend in SAMHD1 localization at the different time points. While SAMHD1 has previously been reported to be a nuclear protein 51, these experiments show the distribution of SAMHD1 is varied, implying that an intracellular trafficking event occurs. SAMHD1 is initially observed in the nucleus in the control samples, but following LPA stimulation it translocates to the cytosol where it overlaps with sites of oxidation, before eventually returning to the nucleus at the 60 minute time point. The trafficking of SAMHD1 within the cell follows a similar time course to the oxidation of SAMHD1 in response to a proliferative signal. Taken together, the cellular imaging experiments localize SAMHD1 and oxidized proteins, both temporally and spatially, to occupying similar regions in the cytosol following growth factor stimulation.

DISCUSSION

The connection between SAMHD1 and viral infection has driven much of the initial research efforts into understanding the enzyme’s unique structure and dNTPase activity. Given the importance of nucleic acid metabolism to cellular homeostasis throughout the cell cycle, the

role SAMHD1 plays in non-pathophysiological cell processes may comprise a more primary biological function. Here we investigate the relationship between SAMHD1 dNTPase activity, and the emerging field of redox signaling, suggesting oxidation status as an essential regulator of cell cycle progression.

Our data reveal that SAMHD1 catalytic activity is sensitive to reversible redox regulation in vitro. Oxidation of SAMHD1 interferes with its ability to oligomerize and form an activated tetrameric state. We also demonstrate that SAMHD1 is oxidized in different cell types in response to proliferative signals for each cell type. While the present study and previously reported structural data suggest that oxidation results in a disulfide bond between Cys341 and Cys350, ultimately the ability to respond to redox signals is controlled by Cys522.

Mutation of Cys522 to alanine renders SAMHD1 resistant to oxidative inactivation.

Based on these data we propose a model of oxidative inactivation of SAMHD1 that utilizes

Cys522 as a “cysteine switch”. Cysteine switches can adopt different forms, but functionally exists as a mechanism that can translate redox signals into binary on/off structural configurations within enzymes 34,52. Within the SAMHD1 monomer/dimer Cys522 resides on a peripheral flexible loop adjacent to the regulatory nucleotide binding site (Fig 6A).

Upon exposure to a cellular oxidant, like H2O2, we propose that Cys522 undergoes S- sulfenylation. The flexibility of the loop could allow the oxidized Cys522 to react with the thiol group of either Cys341 or Cys350 to form a disulfide linkage. The formation of the disulfide bond between Cys522 and Cys341 or Cys350 results in a structural rearrangement at the regulatory nucleotide binding site, in which Lys523, is not able to stabilize the γ- phosphate groups of the activating nucleotides. Additionally, oxidation of SAMHD1 appears to result in subtle structural shifts within the regulatory region of the enzyme, specifically in

Fig 6. Model for SAMHD1 regulation by oxidation and cell proliferation. A) Cys522 is positioned in a flexible loop region adjacent to the regulatory nucleotide binding pocket. Flexibility of the loop might allow oxidized Cys522 to interact with Cys341 or Cys 350 to form a disulfide linkage, followed by a resolution to a Cys341-Cys350 disulfide. It is likely this conformation prevents proper tetramerization of SAMHD1 and therefore inhibits catalytic activity. B) Proposed model for regulation of SAMHD1 activity by oxidation to regulate cell cycle progression.

relation to α-helices 12 and 13 and the β-sheets on which Cys341 and Cys350 reside. These structural events likely render SAMHD1 incapable of binding the activating nucleotides essential for SAMHD1 tetramerization and subsequent catalytic activation. Oxidation does not appear to affect the dimer interface of SAMHD1 or the enzyme’s ability to form a dimer, which aligns with structural observations. Given the preponderance of data that identify

Cys341-Cys350 as the primary disulfide following oxidation, our model proposes a further thiol-disulfide exchange to generate Cys341-Cys350 and reduce Cys522. Formation of the

Cys341-Cys350 disulfide appears to maintain a conformational change in α-helices 12 and

13, which comprise parts of the regulatory nucleotide binding site and tetramer interface, and likely stabilizes the inactive species of SAMHD1. Reduction of the Cys341-Cys350 disulfide by cellular reductants could regenerate the enzyme to complete a multi-step regulatory cycle.

Reversible protein oxidation is emerging as a crucial regulatory pathway for essential cellular processes such as cell cycle progression and cell proliferation 53,54. The data presented in this study implicate nucleotide metabolism as another integral process subject to changes in the cellular redox state, and SAMHD1 as a key arbiter between changes in redox state and nucleotide concentrations. We demonstrate the presence of oxidized

SAMHD1 in cells following proliferative signals, which likely results in an increase in the intracellular dNTP pools (Fig 6b). This is a prerequisite step for DNA replication during S- phase in which high concentrations of dNTPs are necessary 31. We also observe a translocation of SAMHD1 from the nucleus to the cytosol following stimulation with growth factor and colocalization with sites of oxidation. This raises the possibility that oxidation not only affects catalytic activity, but also cellular localization, either directly, or indirectly by priming the enzyme for additional post-translational modifications. An additional line of

inquiry revealed by this study will be related to how the redox status of SAMHD1 affects its activity in states of pathophysiological stress. Cancer cells maintain higher levels of oxidation 55,56 than normal cells and certain innate immune signals can initiate an oxidative signaling burst similar to proliferative signals 49,57. Thus, given the important role SAMHD1 plays in innate immunity and its connection to certain cancers, this represents a rich field of potential research.

INNOVATION

Here we present a previously unknown mechanism for regulation of SAMHD1 catalytic activity through protein oxidation. This new information directly ties redox biology to regulation of deoxynucleotide pools, and presents another connection between oxidative signaling and cell proliferation.

MATERIALS AND METHODS

Reagents and antibodies. Primary antibodies for Western blots to SAMHD1 were from

Sigma-Aldrich. Biotin-HRP antibody was purchased from Cell Signaling Technology.

Antibody to Salmonella typhimurium AhpC was purified from rabbit serum. AlexaFluor fluorescent secondary antibodies, and RPMI 1640 medium were from Invitrogen. Fetal bovine serum was from Lonza. Nitrocellulose membranes were from Schleicher and Schuell and Super Signal chemiluminescence reagent was from Pierce. Alkyl-linked 18:1 lysophosphatidic acid [1-(9Z-octadecenyl)-2-hydroxy-sn-glycero-3-phosphate (ammonium salt)] was from Avanti Polar Lipids, Inc. DCP-Bio1 and DCP-Rho1 were obtained from Dr.

Bruce King (Wake Forest University, Dept. of Chemistry).

SAMHD1 expression and purification. The full length human SAMHD1 gene was amplified using PCR and cloned into a modified pET28 expression vector (pLM303-SAMHD1) that contained an N-terminal MBP tag and an intervening rhinovirus 3C protease cleavage site.

Cysteine to alanine mutations (C320A, C341A, C350A, C522A, and C341-350A) were synthesized and sequenced by GenScript, using the pLM-SAMHD1 expression vector as a template. Expression constructs were transformed into E. coli BL21* and grown in Luria-

Bertani medium at 37 °C with shaking to an OD600 = 0.6. Cultures were induced with 0.3mM isopropyl-β-D-thio-galactoside (IPTG), rapidly cooled on ice to 20 °C, and then allowed to express for 16-18 hours at 16 °C. Harvested cells were resuspended in Amylose Column

Buffer (ACB) (50mM Tris-HCl pH 7.5, 200mM NaCl, 5mM MgCl2, 0.1mM EDTA, 2mM DTT, and 5% glycerol) and lysed using an Avestin Emulsiflex-C5 cell homogenizer. Cell debris was cleared by centrifuging at 18.5K r.p.m. for 30 minutes at 4 °C, and the cleared lysate was passed over amylose high flow resin (New England Biolabs), and washed with 3 column volumes of ACB+2M NaCl to remove residual nucleic acid. Bound MBP-SAMHD1 was eluted over a linear gradient with ACB+20mM Maltose. The desired fractions were pooled and dialyzed overnight against Heparin Column Buffer (HCB) (50mM Tris-HCl pH 7.5, 100mM

NaCl, 5mM MgCl2, 0.1mM EDTA, 2mM DTT, and 5% glycerol). PreScission Protease (GE

Biosciences) was added (100:1) to the MBP-SAMHD1 dialysate to cleave the MBP tag. The

SAMHD1 containing sample was separated from the cleaved MBP using a 5ml Heparin

HiTrap (GE Healthcare Life Sciences) and eluted using a linear gradient of HCB+2M NaCl.

SAMHD1 containing fractions were again pooled and further purified using Superdex 200 size exclusion chromatography column equilibrated with 50mM Tris-HCl pH 7.5, 250mM

NaCl, 5mM MgCl2, 0.1mM EDTA, 2mM DTT, and 5% glycerol. The protein was concentrated and flash frozen in individual aliquots using liquid nitrogen, and stored at -80 °C until use.

SAMHD1 deoxynucleotide triphosphohydrolase activity assay. Individual protein aliquots were thawed and passed through a microspin column packed with BioGel-P6 (BioRad) and equilibrated with 50mM NaH2PO4 pH 7.5, 200mM NaCl, 1mM MgCl2, 0.1 mM EDTA, and 2% glycerol, in order to remove all reducing agents from the sample. The concentration of

SAMHD1 containing samples was then determined using a Nanodrop 2000 (Thermo

Scientific) and the extinction coefficient of fully reduced SAMHD1 (ε =75750 M-1 cm-1,

A280). In assays investigating the reversible effect of oxidation on SAMHD1 dNTPase activity, unless preincubation with activating nucleotides (100μM dATPαS-GTP) is indicated, SAMHD1 was first incubated with H2O2 at the specified concentrations for 30 minutes at 21 °C. Where indicated, catalase (100 units) was added to scavenge H2O2.

Reducing agents were then added to the designated samples and allowed to incubate at 21

°C for 30 minutes (DTT/TCEP – 10/2mM or 2/0.4 mM; reduced Glutathione – 2mM; L-

Cysteine – 2mM; E.Coli Thioredoxin/Thioredoxin Reductase/NADPH – 10/0.1/100μM). DTT and TCEP were used simultaneously to insure reducing conditions were maintained.

Following reduction, treated enzyme samples were mixed with a reaction buffer (20mM

Tris pH 7.5, 5mM MgCl2, 1μM EDTA), and the dNTPase assays were initiated by the addition of 500 μM dATP and 50μM GTP. Formation of dA was used to measure the progress of the reaction. SAMHD1 concentration in all reactions was 500nM. Reactions were quenched using EDTA to a final concentration of 10mM after 10 minutes. dNTPase reaction products were analyzed using ion pair reverse phase chromatography on a Waters HPLC system. A CAPCell PAK C18 column (Shiseido Fine Chemicals) was equilibrated with 20mM NaH2PO4 pH 7.0, 5mM tetra n-butylammonium phosphate, and 5% methanol. Reactants and products were eluted with a linear gradient of methanol from 5-

50%. A254 was used to measure elutant peaks, and quantification was performed using the

Empower Software to integrate the area under each reaction component peak. Data analysis

and curve fitting was performed using Sigmaplot v13 (Systat Software, Inc.). The Hanes-

Woolf plot was generated by fitting data to the equation:

[푑퐺푇푃] 1 퐾 = [푑퐺푇푃] + 푚 . 휈 푉푚푎푥 푉푚푎푥

SAMHD1 kinetic analysis. SAMHD1 kinetic analyses were performed in a manner similar to dNTPase activity assays with several notable exceptions. SAMHD1 concentration in kinetic assays was reduced to 250nM, and the enzyme was pre-activated with 100 μM GTP-dATPαS before the addition of dGTP substrate. This ensured that steady state conditions were achieved even at lower concentrations of dGTP. dGTP was chosen in place of dATP as a substrate due to its distinct HPLC peak observed under the specific analytical conditions.

Dynamic Light Scattering. SAMHD1 particle size distributions as a function of hydrodynamic diameter (DH) were determined under varying oxidative states as well as in the presence of activating nucleotides using a Malvern Nano-S Zetasizer (Malvern Instruments,

Worcestershire, UK). Reducing agent was removed from recombinant protein using a microspin column packed with BioGel-P6 and equilibrated with the DLS buffer (50mM

NaH2PO4 pH 7.5, 250mM NaCl, 5mM MgCl2). Samples were prepared by incubating SAMHD1 with activating nucleotides (100μM dATPαS-GTP), H2O2 (100 molar equivalents), and a reducing agent(s) (10mM DTT/2mM TCEP), or an indicated combination of the three, in a specific order. SAMHD1 final concentration in the samples was 0.2 mg/ml. Raw intensity weighted data was converted to volume % distribution in order to mitigate the impact of small amounts of aggregated protein.

Circular Dichroism. SAMHD1 was transferred to CD buffer (20mM NaH2PO4 pH 7.0, 100mM

NaCl, 1mM MgCl2) using a microspin column packed with BioGel-P6 and diluted to 0.2 mg/ml. CD spectra were acquired using a Jasco J-720 Circular Dichroic Spectrometer.

Samples were scanned at 21 °C from 190nm to 250nm, at pitch of 1nm, and a speed of

2nm/min.

Mass Spectrometry. SAMHD1 was incubated with TCEP immobilized reducing gel (Pierce) at

21 °C for 1 hour prior to sample preparation in order to reduce any disulfide bonds formed during protein purification. SAMHD1 was transferred into the reaction buffer (50mM Tris pH7.5, 250mM NaCl, 1mM MgCl2, 0.1mM EDTA) using a microspin column packed with

BioGel-P6. 100μl reactions were prepared by mixing SAMHD1 with dimedone (5mM),

DTT/TCEP (1/0.2mM) when designated, and the indicated concentration of H2O2. In experiments investigating the presence of disulfide bonds, dimedone was omitted from the reaction. Reactions were allowed to proceed for 180 minutes (dimedone labeling) or 30 minutes (disulfide formation) before the protein was partially denatured using 0.1% SDS and free thiols alkylated using 5mM iodoacetamide. The alkylation reaction was quenched after 60 minutes by passing the samples through a microspin column packed with Biogel P-

6 and equilibrated with a volatile buffer (40mM NH4HCO3 pH 8.1, 1mM CaCl2, 10%

Acetonitrile). Protein samples were digested with modified trypsin (sequencing grade,

Thermo; 1:100) overnight at 37 °C. The tryptic peptides were dried using a speedvac and resuspended in 5% acetonitrile and 1% formic acid. 5μg of protein were analyzed using an

LC-MS/MS system that consisted of a Q Exactive HF Hybrid Quadrupole-Orbitrap Mass

Spectrometer (Thermo Scientific, Rockford, IL) and a Dionex Ultimate-3000 nano-UPLC system (Thermo Scientific, Rockford, IL) employing a Nanospray Flex Ion Source (Thermo

Scientific, Rockford, IL). An Acclaim PepMap 100 (C18, 5 μm, 100 Å, 100 μm x 2 cm) trap column and an Acclaim PepMap RSLC (C18, 2 μm, 100 Å, 75 μm x 15 cm) analytical column were used for the stationary phase. Good chromatographic separation was observed with a

90 min linear gradient consisting of mobile phases A (5% acetonitrile with 0.1% formic acid) and B (80% acetonitrile with 0.1% formic acid). MS Spectra were searched employing

SEQUEST search algorithm with Proteome Discoverer v2.1 (Thermo Scientific, Rockford,

IL). Search parameters were as follow; FT-trap instrument, parent mass error tolerance of

10 ppm, fragment mass error tolerance of 0.02 Da (monoisotopic), variable modifications on methionine of oxidation, cysteine of alkylation, oxidation and dimedone. MS Spectra were searched for the presence of disulfide bonds using Mass Matrix 44 and pLink-SS45 software.

Cell culture and treatments. PC3 cells (from ATCC stocks) were grown, maintained, and treated at 37 °C with 5% CO2 in RPMI 1640 medium supplemented with 10% fetal bovine serum, L-glutamine, penicillin, and streptomycin. RWPE-1 (from ATCC stocks) cells were grown in Kerotinocyte Serum Free Media supplemented with 0.05 mg/ml Bovine Pituitary

Extract and 5 ng/ml Epidermal Growth Factor. WI-38 and CCD-1070SK cells (from ATCC stocks) were grown in EMEM containing 10% fetal bovine serum, L-glutamine, penicillin, and streptomycin. LPA (Avanti), supplied in chloroform, was dried under a stream of nitrogen, resuspended to a concentration of 1 mM in phosphate buffered saline (PBS) containing 1% fatty acid-free bovine serum albumin (BSA), and then diluted into culture medium to the indicated concentrations. Full length human SAMHD1 wild type and C522A containing an N-terminal HA tag was cloned into the eukaryotic expression vector pCDNA3.1 Synthesis and sequencing were performed by GenScript.

Western blotting. For Western blotting, cells were plated at 5×105 cells per dish in 100-mm dishes, treated or not with pharmacological agents, washed with cold, calcium-free PBS, scraped into lysis buffer (50 mM Tris-HCl, 100 mM NaCl, 2 mM EDTA, 0.1% SDS, 0.5% sodium deoxycholate, 1 mM PMSF, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 50 mM NaF, and

1 mM sodium vanadate), and centrifuged to remove cell debris after one freeze/thaw cycle.

Protein concentration was measured (Pierce BCA protein assay) and samples (typically 40

μg protein/lane) were resolved on SDS polyacrylamide gels, then transferred to nitrocellulose membranes, probed with protein-specific antibodies and visualized using

Super Signal chemiluminescence reagent.

Non-reducing Western blotting. Samples were prepared by adding in specific order, fill buffer (50mM NaH2PO4, pH 7.2, 150mM NaCl, 5mM MgCl2, 2% glycerol), activating nucleotides (dATPαS-GTP) as indicated, enzyme as indicated (0.02 mg/ ml), and then incubated at RT for 5 min in order to form the SAMHD1 tetramer in samples with dNTPs added. H2O2 was titrated in as indicated (1-100 μM) followed immediately by reducing agent as indicated (10mM DTT/2mM TCEP). Samples were incubated at RT for 20 minutes before adding 100mM NEM to block unreacted cysteines. SAMHD1 protein was diluted to

0.1 µg using a non-reducing SDS sample buffer and resolved by 7% SDS-PAGE, transferred to nitrocellulose, probed for SAMHD1 and visualized using Super Signal chemiluminescence reagent.

DCP-Bio1 labeling and affinity capture. Cells (~5×105) were grown in 100-mm plates for 48 h as described above, and after incubation of 100 nM alkyl-LPA with the cells for indicated time points, DCP-Bio1 was added with the lysis buffer to chemically trap sulfenic acid proteins as described by Klomsiri, et al 50 with slight modifications. Lysis buffer was freshly

prepared and supplemented with 1 mM DCP-Bio1, 10 mM NEM, 10 mM iodoacetamide, and

200 U/ml catalase. Use of NEM and iodoacetamide together provided the best blocking of free thiols. Lysis buffer (150 µL/plate) was added to each plate, and then cells were scraped from the plates, transferred to Eppendorf tubes, incubated on ice for 30 min before storing at −80 °C. For affinity capture and elution of the labeled proteins, unreacted DCP-Bio1 was removed immediately on thawing using a BioGel P6 spin column. Samples were then assayed for protein concentration, diluted to 1.0 mg/ml protein into 400 µl buffer containing a final concentration of 2 M urea, and supplemented with prebiotinylated AhpC

(0.5 µg/400 µg of total lysate) to control for the efficiency of the affinity capture and elution procedures. Samples were precleared with Sepharose CL-4B beads (Sigma), applied to columns containing high capacity streptavidin-agarose beads from Pierce, and then incubated overnight at 4 °C. Multiple stringent washes of the beads were performed (at least 4 column volumes and two washes each) using, in series, 1% SDS, 4 M urea, 1 M NaCl,

10 mM DTT, 50 mM ammonium bicarbonate and water, before elution with Laemmli sample buffer50. Samples were stored at −80 °C as needed and analyzed by Western blot as described above. Alternatively, samples lysed in the presence of DCP-Bio1 were immunoprecipitated using SAMHD1 antibody and Pierce Protein A/G Magnetic Beads according to the manufacturer’s protocol. Proteins were eluted from the beads using SDS- sample buffer containing beta-mercaptoethanol, and detected by Western blotting.

Immunofluorescence. For immunofluorescence studies, 4×104 cells were plated in 0.5 ml of media in 4-well chambered coverglass and incubated for 24 h, and then treated (or not) with LPA for indicated time points, DCP-Rho1 at 10 µM was added for the final 10 min of

LPA stimulation47, and then cells were washed extensively with RPMI containing 10% fetal bovine serum before fixing with 10% formalin for 15 min. Cells were washed 3 times with

0.1 mM glycine in PBS containing 2% fetal bovine serum, and then permeabilized for 10 min with 0.1% Triton-X100 in PBS-T. After washing 3 times with PBS-T, samples were blocked with 2% BSA in PBS-T for 1 h, and then incubated with primary antibodies (1:200 dilution) overnight at 4 °C in 2 % BSA in PBS-T. After washing with PBS-T, AlexaFluor-labeled secondary antibodies diluted 1:1000 were added and incubated for 2 h at room temperature, and washed again with PBS-T. Chambers were removed and coverslips were mounted using ProLong Gold Antifade Reagent with DAPI (Molecular Probes), incubated 18 hours and then imaged. Labeled cells were visualized by confocal microscopy using an

Olympus FV 1200 Spectral Laser Scanning Confocal Microscope. A goat anti-rabbit

AlexaFluor488 secondary antibody with excitation at 488 nm and emission at 505–550 nm was used for imaging of SAMHD1. Excitation at 561 nm and emission at 566–703 nm was used for the DCP-Rho1 signal. Intensity values were kept in the linear range, and the absence of crossover and bleed through between channels was confirmed as described in

Klomsiri, et al 47. Image analysis was performed using Fiji image processing software and the coloc2 plugin 58. An intensity threshold was manually determined on a control image and identically applied to all images to correct for background signal. The cytoplasmic fraction was isolated by subtracting the DAPI signal from images converted to gray scale for analysis. Manders’ coefficients59 represent the fraction of the total summed intensities (i) for a fluorescent probe, R, that colocalizes with pixels containing intensities for a second probe, G, as described by the equation:

∑ 𝑖푅푖,푐표푙표푐푎푙 퐺 푀퐶퐶 = , 푤ℎ푒푛 퐺푖 > 0 푅푡표푡푎푙

Proximity Ligation Assay (PLA). PC3 cells (~5×105 per well) were plated on a 4 well chambered slide, and incubated 48 hours before stimulating with LPA as indicated. DCP-

Bio1 was added at a concentration of 10 µM for the final 10 minutes of LPA stimulation.

Cells were rinsed with growth media three times and one time with PBS before incubating

10 minutes in 10% formalin containing 10 mM NEM and 10 mM iodoacetamide. After fixation, cells were rinsed three times with PBS-T containing 2% FBS and 0.1M glycine. Cells were permeablized with 0.1% Trition X-100 in PBS-T for 10 minutes, washed with PBS-T and then blocked for 1 hour with 2% BSA in PBS-T. Primary antibodies against SAMHD1 and Biotin (1:200 dilution) was added overnight in blocking buffer. After washing in PBS-T,

PLA probes from the Duolink In Situ Red Starter Kit (Sigma-Aldrich) were added to cells and the ligation reaction was done according to manufacturer’s instructions. Amplification of the PLA signal was done using the supplied red amplification buffer and polymerase per instructions. Cells were mounted in Duolink In Situ Mounting Medium containing DAPI and visualized by confocal microscopy using an Olympus FV 1200 Spectral Laser Scanning

Confocal Microscope. Excitation at 561 nm and emission at 566–703 nm was used for the

PLA signal. Intensity values were kept in the linear range, and the absence of crossover and bleed through between channels was confirmed as described in Klomsiri, et al 47. ImageJ software was used to quantify mean fluorescence intensity of at least 100 cells per timepoint.

Cell Lysate dNTPase Assay: CCD-1070SK fibroblasts (0.25x106 cells) were plated in 100mm dishes in 10ml of media. 24 hours after seeding, cells were transiently transfected using

FuGENE® HD Transfection Reagent (Promega) with pCDNA3.1 vectors containing either wild type SAMHD1, C522A mutant SAMHD1, or an empty vector control. Prior to experimental treatment, cells were incubated in serum free media for 24 hours. 72 hours

post transfection, PDGF (or an equivalent volume of PBS) was added to the dishes to a final concentration of 20nM. Treated and control cells were washed with cold PBS 60 minutes post treatment, harvested by trypsinization, pelleted, and lysed in 100μl of lysis buffer

(50mM Tris pH 7.5, 150mM NaCl, 2mM MgCl2, 1mM EDTA, 10% glycerol, 0.05% NP-40, 1%

Triton-X, 100μM PMSF, 200 U/ml catalase, 10ng/ml Aprotinin, and 10ng/ml Lupeptin). Cell debris was removed by centrifugation at 10000 rcf for 10 minutes and protein content for each sample determined using a BCA assay. 100μg of protein was then used to conduct a dNTPase assay as described above, varying only by substrate (1mM dGTP) and time (5 hours) given the relatively small amount of SAMHD1 within each sample. Reactions were quenched using EDTA to a final concentration of 25mM, boiled for 1 min, and centrifuged at

15000 rcf for 10 min to removed protein. The supernatant was then analyzed by HPLC, and reaction progress determined by the relative integrated areas under the dG and dGTP peaks.

Size Exclusion Chromatography - Multi-Angle Light Scattering (SEC-MALS). Purified recombinant SAMHD1 was passed through a microspin column containing BioGel-P6

(BioRad) equilibrated with Analytical Size Exclusion Buffer (ASEC) (50mM HEPES pH 7.5,

200mM NaCl, 1mM MgCl2, 100μM EDTA) in order to remove all reducing agents from the protein. SAMHD1 at 2.0 mg/ml was treated with H2O2 (250μM) for 30 minutes, H2O2 and then Reducing Agent (10mM DTT/2.5mM TCEP), or ASEC buffer alone, before tetramerization was induced by the addition of dGTPαS (250μM). 100μl of each sample were then injected onto an analytical size exclusion TSKgel column (7.8mm x 30 cm, 8μm particle size; Tosoh Bioscience) equilibrated with SEC buffer and run at 0.5 ml/min. Elution profiles were monitored by UV280 absorbance (2998 Photodiode Array Detector; Waters) and molecular masses determined by multi-angle light scattering (HELEOS II; Wyatt

Technology) with linked refractive index determination (Optilab rEX; Wyatt Technology).

Astra 6.1 software (Wyatt Technology) was used to analyze experimental data and generate figures.

Glutaraldehyde Crosslinking. Recombinant SAMHD1 wild type or C522A mutant was passed through a microspin column containing BioGel-P6 (BioRad) equilibrated with crosslinking buffer (50mM NaH2PO4 pH 7.5, 250mM NaCl, 5mM MgCl2, 100μM EDTA, 2% glycerol).

100μl reactions were set up containing 2.5μg of protein sequentially treated with the indicated combination of activating nucleotides (200μM dGTPαS), H2O2 for 10, 20, or 30 minutes, and reducing agent (10mM DTT/2mM TCEP). Glutaraldehyde was then added to a final concentration of 2.5mM and the crosslinking allowed to proceed for 10 minutes at 21

°C before quenching with 100μl of 0.5M Tris/0.5M Glycine. Protein samples were separated using 4-12% SDS-PAGE, and imaged by Western Blot. Densitometry analysis of the relative intensity of bands corresponding to the monomer, dimer, and tetramer species was performed using Fiji software.

Data and Statistical Analysis: Charts and graphs were generated using SigmaPlot v13.0, which was also used for statistical analysis, t-test comparisons, and nonlinear regression.

Statistical significance was determined at three levels of significance; p<0.05 (*), p<0.01

(**), p<0.001 (***). Molecular representations were generated using the program PyMol 60.

Author Contributions

Conceptualization- C.H.M., L.C.R., R.S.H., L.W.D., F.W.P., T.H.; Methodology- C.H.M., L.C.R.,

C.M.F., L.B.P., H.W., T.H., Investigation- C.H.M, L.C.R., N.O.D-B. H.W., Writing-original draft-

C.H.M, L.C.R, T.H., Writing-review & editing- C.H.M, L.C.R, R.S.H, L.W.D, L.B.P, F.W.P., T.H.,

Supervision- T.H, F.W.P., R.S.H., C.M.F, Funding Acquisition- T.H, F.W.P., R.S.H., C.M.F, L.W.D,

L.B.P.

Acknowledgments

We thank Dr. Jingyun Lee, and the staff of the Metabolomics and Proteomics Shared

Resource of the Wake Forest School of Medicine for their assistance in the performance and analysis of mass spectrometry experiments, and Amy Molan and Michael Carpenter for comments on the manuscript. This work has been supported through funding from the NIH

(RO1 GM108827 to TH, RO1 CA142838 to LWD, R01 GM06692 to FWP, and R33 CA177461 to LBP and T32 GM095440 support for CHM), Alliance for Lupus Research to FWP, and the

Comprehensive Cancer Center of Wake Forest University National Cancer Institute Cancer

Center Support Grant P30CA012197. RSH is an Investigator of the Howard Hughes Medical

Institute.

Author Disclosures

All of the authors declare they have no conflicts of interest, financial or otherwise, that influenced the data collection, interpretation, or publication of this work.

Abbreviations

AGS - Aicardi-Goutieres syndrome

BSA - bovine serum albumin

CD - circular dichroism

Cys - cysteine

DH - hydrodynamic diameter

DAPI - 4',6-diamidino-2-phenylindole dATPαS - deoxyadenosine triphosphate alpha S

DCP-Bio1 - dimedone coupled probe to biotin

DCP-Rho1 - dimedone coupled probe to rhodamine dNTP - deoxynucleotide triphosphate

DTT - dithiothreitol

EDTA - ethylenediaminetetraacetic acid

GTP - guanosine triphosphate

H2O2 - hydrogen peroxide

IPTG - isopropyl-β-D-thio-galactoside

LPA - lysophosphatidic acid

NEM - N-ethylmaleimide

PBS - phosphate buffered saline

PBS-T - phosphate buffered saline, 0.1% Triton X100

PDGF - platelet derived growth factor

PLA - proximity ligation assay

PMSF - phenylmethylsulfonyl fluoride

R-SOH - sulfenic acid

R-SO2H - sulfinic acid

R-SO3H - sulfonic acid

SDS - sodium dodecyl sulfate

SDS-PAGE - sodium dodecyl sulfate polyacrylamide gel electrophoresis

SEC-MALS - size exclusion chromatography - multi-angle light scattering

TCEP - Tris(2-carboxyethyl)phosphine

References

1. Powell RD, Holland PJ, Hollis T, Perrino FW. Aicardi-Goutieres Syndrome Gene and HIV-1 Restriction Factor SAMHD1 Is a dGTP-regulated Deoxynucleotide Triphosphohydrolase. J Biol Chem. 2011;286(51):43596-43600. doi:10.1074/jbc.C111.317628. 2. Goldstone DC, Ennis-Adeniran V, Hedden JJ, et al. HIV-1 restriction factor SAMHD1 is a deoxynucleoside triphosphate triphosphohydrolase. Nature. 2011;480(7377):379- 382. doi:10.1038/nature10623. 3. Franzolin E, Pontarin G, Rampazzo C, et al. The deoxynucleotide triphosphohydrolase SAMHD1 is a major regulator of DNA precursor pools in mammalian cells. Proc Natl Acad Sci. 2013;110(35):14272-14277. doi:10.1073/pnas.1312033110. 4. Kretschmer S, Wolf C, Konig N, et al. SAMHD1 prevents autoimmunity by maintaining genome stability. Ann Rheum Dis. 2015;74(3):e17-e17. doi:10.1136/annrheumdis- 2013-204845. 5. Ravenscroft JC, Suri M, Rice GI, Szynkiewicz M, Crow YJ. Autosomal dominant inheritance of a heterozygous mutation in SAMHD1 causing familial chilblain lupus. Am J Med Genet Part A. 2011;155(1):235-237. doi:10.1002/ajmg.a.33778. 6. Rice GI, Bond J, Asipu A, et al. Mutations involved in Aicardi-Goutières syndrome implicate SAMHD1 as regulator of the innate immune response. Nat Genet. 2009;41(7):829-832. doi:10.1038/ng.373. 7. Crow YJ, Rehwinkel J. Aicardi-Goutieres syndrome and related phenotypes: linking nucleic acid metabolism with autoimmunity. Hum Mol Genet. 2009;18(R2):R130- R136. doi:10.1093/hmg/ddp293. 8. Clifford R, Louis T, Robbe P, et al. SAMHD1 is mutated recurrently in chronic lymphocytic leukemia and is involved in response to DNA damage. Blood. 2014;123(7):1021-1031. doi:10.1182/blood-2013-04-490847. 9. Rentoft M, Lindell K, Tran P, et al. Heterozygous colon cancer-associated mutations of SAMHD1 have functional significance. Proc Natl Acad Sci. 2016;113(17):4723-4728. doi:10.1073/pnas.1519128113. 10. Wang J, Lu F, Shen X-Y, Wu Y, Zhao L. SAMHD1 is down regulated in lung cancer by methylation and inhibits tumor cell proliferation. Biochem Biophys Res Commun. 2014;455(3-4):229-233. doi:10.1016/j.bbrc.2014.10.153. 11. Baldauf H-M, Pan X, Erikson E, et al. SAMHD1 restricts HIV-1 infection in resting CD4+ T cells. Nat Med. 2012;18(11):1682-1689. doi:10.1038/nm.2964. 12. Gramberg T, Kahle T, Bloch N, et al. Restriction of diverse retroviruses by SAMHD1. Retrovirology. 2013;10(1):26. doi:10.1186/1742-4690-10-26. 13. Hollenbaugh JA, Gee P, Baker J, et al. Host Factor SAMHD1 Restricts DNA Viruses in Non-Dividing Myeloid Cells. Ross SR, ed. PLoS Pathog. 2013;9(6):e1003481. doi:10.1371/journal.ppat.1003481. 14. Hrecka K, Hao C, Gierszewska M, et al. Vpx relieves inhibition of HIV-1 infection of macrophages mediated by the SAMHD1 protein. Nature. 2011;474(7353):658-661. doi:10.1038/nature10195. 15. Laguette N, Sobhian B, Casartelli N, et al. SAMHD1 is the dendritic- and myeloid-cell-

specific HIV-1 restriction factor counteracted by Vpx. Nature. 2011;474(7353):654- 657. doi:10.1038/nature10117. 16. Li N, Zhang W, Cao X. Identification of human homologue of mouse IFN-γ induced protein from human dendritic cells. Immunol Lett. 2000;74(3):221-224. doi:10.1016/S0165-2478(00)00276-5. 17. St Gelais C, de Silva S, Amie SM, et al. SAMHD1 restricts HIV-1 infection in dendritic cells (DCs) by dNTP depletion, but its expression in DCs and primary CD4+ T- lymphocytes cannot be upregulated by interferons. Retrovirology. 2012;9(1):105. doi:10.1186/1742-4690-9-105. 18. White TE, Brandariz-Nuñez A, Valle-Casuso JC, et al. Contribution of SAM and HD domains to retroviral restriction mediated by human SAMHD1. Virology. 2013;436(1):81-90. doi:10.1016/j.virol.2012.10.029. 19. Bonifati S, Daly MB, St. Gelais C, et al. SAMHD1 controls cell cycle status, apoptosis and HIV-1 infection in monocytic THP-1 cells. Virology. 2016;495:92-100. doi:10.1016/j.virol.2016.05.002. 20. Lahouassa H, Daddacha W, Hofmann H, et al. SAMHD1 restricts the replication of human immunodeficiency virus type 1 by depleting the intracellular pool of deoxynucleoside triphosphates. Nat Immunol. 2012;13(3):223-228. doi:10.1038/ni.2236. 21. Kim B, Nguyen LA, Daddacha W, Hollenbaugh JA. Tight Interplay among SAMHD1 Protein Level, Cellular dNTP Levels, and HIV-1 Proviral DNA Synthesis Kinetics in Human Primary Monocyte-derived Macrophages. J Biol Chem. 2012;287(26):21570- 21574. doi:10.1074/jbc.C112.374843. 22. Welbourn S, Strebel K. Low dNTP levels are necessary but may not be sufficient for lentiviral restriction by SAMHD1. Virology. 2016;488:271-277. doi:10.1016/j.virol.2015.11.022. 23. Cribier A, Descours B, Valadão A, Laguette N, Benkirane M. Phosphorylation of SAMHD1 by Cyclin A2/CDK1 Regulates Its Restriction Activity toward HIV-1. Cell Rep. 2013;3(4):1036-1043. doi:10.1016/j.celrep.2013.03.017. 24. Welbourn S, Dutta SM, Semmes OJ, Strebel K. Restriction of Virus Infection but Not Catalytic dNTPase Activity Is Regulated by Phosphorylation of SAMHD1. J Virol. 2013;87(21):11516-11524. doi:10.1128/JVI.01642-13. 25. White TE, Brandariz-Nuñez A, Valle-Casuso JC, et al. The Retroviral Restriction Ability of SAMHD1, but Not Its Deoxynucleotide Triphosphohydrolase Activity, Is Regulated by Phosphorylation. Cell Host Microbe. 2013;13(4):441-451. doi:10.1016/j.chom.2013.03.005. 26. Zhu C., Wei W, Peng X, Dong Y., Gong Y, Yu XF. The mechanism of substrate- controlled allosteric regulation of SAMHD1 activated by GTP. Acta Crystallogr,SectD. 2015;71:516-524. 27. Koharudin LMI, Wu Y, DeLucia M, Mehrens J, Gronenborn AM, Ahn J. Structural Basis of Allosteric Activation of Sterile Motif and Histidine-Aspartate Domain-containing Protein 1 (SAMHD1) by Nucleoside Triphosphates. J Biol Chem. 2014;289(47):32617- 32627. doi:10.1074/jbc.M114.591958. 28. Yan J, Kaur S, DeLucia M, et al. Tetramerization of SAMHD1 is required for biological activity and inhibition of HIV infection. J Biol Chem. 2013;288(15):10406-10417.

doi:10.1074/jbc.M112.443796. 29. Amie SM, Bambara RA, Kim B. GTP Is the Primary Activator of the Anti-HIV Restriction Factor SAMHD1. J Biol Chem. 2013;288(35):25001-25006. doi:10.1074/jbc.C113.493619. 30. Hansen EC, Seamon KJ, Cravens SL, Stivers JT. GTP activator and dNTP substrates of HIV-1 restriction factor SAMHD1 generate a long-lived activated state. Proc Natl Acad Sci. 2014;111(18):E1843-E1851. doi:10.1073/pnas.1401706111. 31. Reichard P. Interactions Between Deoxyribonucleotide And DNA Synthesis. Annu Rev Biochem. 1988;57(1):349-374. doi:10.1146/annurev.biochem.57.1.349. 32. Pauls E, Ruiz A, Badia R, et al. Cell Cycle Control and HIV-1 Susceptibility Are Linked by CDK6-Dependent CDK2 Phosphorylation of SAMHD1 in Myeloid and Lymphoid Cells. J Immunol. 2014;193(4):1988-1997. doi:10.4049/jimmunol.1400873. 33. Finkel T. Signal transduction by reactive oxygen species. J Cell Biol. 2011;194(1):7- 15. doi:10.1083/jcb.201102095. 34. Go Y-M, Jones DP. The Redox Proteome. J Biol Chem. 2013;288(37):26512-26520. doi:10.1074/jbc.R113.464131. 35. Go Y-M, Chandler JD, Jones DP. The cysteine proteome. Free Radic Biol Med. 2015;84:227-245. doi:10.1016/j.freeradbiomed.2015.03.022. 36. Ji X, Wu Y, Yan J, et al. Mechanism of allosteric activation of SAMHD1 by dGTP. Nat Struct Mol Biol. 2013;20(11):1304-1309. doi:10.1038/nsmb.2692. 37. Zhu C, Gao W, Zhao K, et al. Structural insight into dGTP-dependent activation of tetrameric SAMHD1 deoxynucleoside triphosphate triphosphohydrolase. Nat Commun. 2013;4:2722. 38. Ji X, Tang C, Zhao Q, Wang W, Xiong Y. Structural basis of cellular dNTP regulation by SAMHD1. Proc Natl Acad Sci. 2014;111(41):E4305-E4314. doi:10.1073/pnas.1412289111. 39. Poole LB. The basics of thiols and cysteines in redox biology and chemistry. Free Radic Biol Med. 2015;80:148-157. doi:10.1016/j.freeradbiomed.2014.11.013. 40. Arnold L., Groom HC., Kunzelmann S, et al. Phospho-Dependent Regulation of Samhd1 Oligomerisation Couples Catalysis and Restriction. Plos Pathog. 2015;11:5194. 41. Furdui CM, Poole LB. Chemical approaches to detect and analyze protein sulfenic acids. Mass Spectrom Rev. 2014;33(2):126-146. doi:10.1002/mas.21384.Chemical. 42. Paulsen CE, Carroll KS. Cysteine-Mediated Redox Signaling: Chemistry, Biology, and Tools for Discovery. Chem Rev. 2013;113(7):4633-4679. doi:10.1021/cr300163e. 43. Winterbourn CC, Hampton MB. Thiol chemistry and specificity in redox signaling. Free Radic Biol Med. 2008;45(5):549-561. doi:10.1016/j.freeradbiomed.2008.05.004. 44. Xu H, Hsu P-H, Zhang L, Tsai M-D, Freitas M a. Database search algorithm for identification of intact cross-links in proteins and peptides using tandem mass spectrometry. J Proteome Res. 2010;9(7):3384-3393. doi:10.1021/pr100369y. 45. Lu S, Fan S-B, Yang B, et al. Mapping native disulfide bonds at a proteome scale. Nat Methods. 2015;12(4):329-331. doi:10.1038/nmeth.3283. 46. Cremers CM, Jakob U. Oxidant sensing by reversible disulfide bond formation. J Biol

Chem. 2013;288(37):26489-26496. doi:10.1074/jbc.R113.462929. 47. Klomsiri C, Rogers LC, Soito L, et al. Endosomal H2O2 production leads to localized cysteine sulfenic acid formation on proteins during lysophosphatidic acid-mediated cell signaling. Free Radic Biol Med. 2014;71:49-60. doi:10.1016/j.freeradbiomed.2014.03.017. 48. Lu JP, Monardo L, Bryskin I, et al. Androgens induce oxidative stress and radiation resistance in prostate cancer cells though NADPH oxidase. Prostate Cancer Prostatic Dis. 2010;13(1):39-46. doi:10.1038/pcan.2009.24. 49. Sundaresan M, Yu ZX, Ferrans VJ, et al. Regulation of reactive-oxygen-species generation in fibroblasts by Rac1. Biochem J. 1996;318 ( Pt 2:379-382. http://www.ncbi.nlm.nih.gov/pubmed/8809022. 50. Klomsiri C, Nelson KJ, Bechtold E, et al. Use of Dimedone-Based Chemical Probes for Sulfenic Acid Detection. Thiol Redox Transitions Cell Signaling, Part A. 2010;473(10):77-94. doi:10.1016/S0076-6879(10)73003-2. 51. Brandariz-Nuñez A, Valle-Casuso J, White TE, et al. Role of SAMHD1 nuclear localization in restriction of HIV-1 and SIVmac. Retrovirology. 2012;9(1):49. doi:10.1186/1742-4690-9-49. 52. Klomsiri C, Karplus PA, Poole LB. Cysteine-Based Redox Switches in Enzymes. Antioxid Redox Signal. 2011;14(6):1065-1077. doi:10.1089/ars.2010.3376. 53. Jones DP. Redox sensing: orthogonal control in cell cycle and apoptosis signalling. J Intern Med. 2010;268(5):432-448. doi:10.1111/j.1365-2796.2010.02268.x. 54. Burhans WC, Heintz NH. The cell cycle is a redox cycle: Linking phase-specific targets to cell fate. Free Radic Biol Med. 2009;47(9):1282-1293. doi:10.1016/j.freeradbiomed.2009.05.026. 55. Szatrowski TP, Nathan CF. Production of large amounts of hydrogen peroxide by human tumor cells. Cancer Res. 1991;51(3):794-798. http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt= Citation&list_uids=1846317%5Cnhttp://cancerres.aacrjournals.org/content/51/3/7 94.full.pdf. 56. Policastro L, Molinari B, Larcher F, et al. Imbalance of antioxidant enzymes in tumor cells and inhibition of proliferation and malignant features by scavenging hydrogen peroxide. Mol Carcinog. 2004;39(2):103-113. doi:10.1002/mc.20001. 57. Radeke HH, Meier B, Topley N, Flöge J, Habermehl GG, Resch K. Interleukin 1-α and tumor necrosis factor-α induce oxygen radical production in mesangial cells. Kidney Int. 1990;37(2):767-775. doi:10.1038/ki.1990.44. 58. Schindelin J, Arganda-Carreras I, Frise E, et al. Fiji: an open-source platform for biological-image analysis. Nat Methods. 2012;9(7):676-682. doi:10.1038/nmeth.2019. 59. Dunn KW, Kamocka MM, McDonald JH. A practical guide to evaluating colocalization in biological microscopy. AJP Cell Physiol. 2011;300(4):C723-C742. doi:10.1152/ajpcell.00462.2010. 60. Schrödinger, LLC. The {PyMOL} Molecular Graphics System, Version~1.8.; 2015. 61. Mathews CK. Deoxyribonucleotides as genetic and metabolic regulators. FASEB J. 2014;28(9):3832-3840. doi:10.1096/fj.14-251249.

62. Bebenek K, Roberts JD, Kunkel TA. The effects of dNTP pool imbalances on frameshift fidelity during DNA replication. J Biol Chem. 1992;267(6):3589-3596. 63. Chabes A, Stillman B. Constitutively high dNTP concentration inhibits cell cycle progression and the DNA damage checkpoint in yeast Saccharomyces cerevisiae. Proc Natl Acad Sci U S A. 2007;104(4):1183-1188. doi:10.1073/pnas.0610585104. 64. Buckland RJ, Watt DL, Chittoor B, Nilsson AK, Kunkel TA, Chabes A. Increased and Imbalanced dNTP Pools Symmetrically Promote Both Leading and Lagging Strand Replication Infidelity. PLoS Genet. 2014;10(12). doi:10.1371/journal.pgen.1004846. 65. Kumar D, Viberg J, Nilsson AK, Chabes A. Highly mutagenic and severely imbalanced dNTP pools can escape detection by the S-phase checkpoint. Nucleic Acids Res. 2010;38(12):3975-3983. doi:10.1093/nar/gkq128. 66. Bester AC, Roniger M, Oren YS, et al. Nucleotide deficiency promotes genomic instability in early stages of cancer development. Cell. 2011;145(3):435-446. doi:10.1016/j.cell.2011.03.044. 67. Pajalunga D, Franzolin E, Stevanoni M, et al. A defective dNTP pool hinders DNA replication in cell cycle-reactivated terminally differentiated muscle cells. Cell Death Differ. 2017:1-11. doi:10.1038/cdd.2017.4. 68. Rampazzo C, Miazzi C, Franzolin E, et al. Regulation by degradation, a cellular defense against deoxyribonucleotide pool imbalances. Mutat Res - Genet Toxicol Environ Mutagen. 2010;703(1):2-10. doi:10.1016/j.mrgentox.2010.06.002. 69. Goldstone DC, Ennis-Adeniran V, Hedden JJ, et al. HIV-1 restriction factor SAMHD1 is a deoxynucleoside triphosphate triphosphohydrolase. Nature. 2011;480(7377):379- 382. doi:10.1038/nature10623. 70. Zhu C-F, Wei W, Peng X, Dong Y-H, Gong Y, Yu X-F. The mechanism of substrate- controlled allosteric regulation of SAMHD1 activated by GTP. Acta Crystallogr Sect D Biol Crystallogr. 2015;71(3):516-524. doi:10.1107/S1399004714027527. 71. Hollenbaugh JA, Tao S, Lenzi GM, et al. dNTP pool modulation dynamics by SAMHD1 protein in monocyte-derived macrophages. Retrovirology. 2014;11(1):63. doi:10.1186/s12977-014-0063-2. 72. Clifford R, Louis T, Robbe P, et al. SAMHD1 is mutated recurrently in chronic lymphocytic leukemia and is involved in response to DNA damage. Blood. 2014;123(7):1021-1031. doi:10.1182/blood-2013-04-490847. 73. Schmidt S, Schenkova K, Adam T, et al. SAMHD1’s protein expression profile in humans. J Leukoc Biol. 2015;98(1):5-14. doi:10.1189/jlb.4HI0714-338RR. 74. Bhattacharya A, Wang Z, White T, et al. Effects of T592 phosphomimetic mutations on tetramer stability and dNTPase activity of SAMHD1 can not explain the retroviral restriction defect. Sci Rep. 2016;6(April):31353. doi:10.1038/srep31353. 75. Tang C, Ji X, Wu L, Xiong Y. Impaired dNTPase activity of SAMHD1 by phosphomimetic mutation of Thr-592. J Biol Chem. 2015;290(44):26352-26359. doi:10.1074/jbc.M115.677435. 76. Wang Z, Bhattacharya A, Villacorta J, Diaz-Griffero F, Ivanov DN. Allosteric activation of SAMHD1 protein by deoxynucleotide triphosphate (dNTP)-dependent tetramerization requires dNTP concentrations that are similar to dNTP concentrations observed in cycling T cells. J Biol Chem. 2016;291(41):21407-21413.

doi:10.1074/jbc.C116.751446. 77. Wittmann S, Behrendt R, Eissmann K, et al. Phosphorylation of murine SAMHD1 regulates its antiretroviral activity. Retrovirology. 2015;12(1):103. doi:10.1186/s12977-015-0229-6. 78. Mauney C, Rogers L, Harris R, et al. The SAMHD1 dNTP triphosphohydrolase is controlled by a redox switch. Antioxid Redox Signal. 2017;(336):ars.2016.6888. doi:10.1089/ars.2016.6888. 79. Hao F, Tan M, Xu X, et al. Lysophosphatidic acid induces prostate cancer PC3 cell migration via activation of LPA1, p42 and p38α. Biochim Biophys Acta - Mol Cell Biol Lipids. 2007;1771(7):883-892. doi:10.1016/j.bbalip.2007.04.010. 80. Guo R, Kasbohm EA, Arora P, et al. Expression and Function of Lysophosphatidic Acid LPA 1 Receptor in Prostate Cancer Cells. Endocrinology. 2006;147(10):4883-4892. doi:10.1210/en.2005-1635. 81. Lennicke C, Rahn J, Lichtenfels R, Wessjohann L a, Seliger B. Hydrogen peroxide - production, fate and role in redox signaling of tumor cells. Cell Commun Signal. 2015;13:39. doi:10.1186/s12964-015-0118-6. 82. O’Brien V, Brown R. Signalling cell cycle arrest and cell death through the MMR System. Carcinogenesis. 2006;27(4):682-692. doi:10.1093/carcin/bgi298. 83. Qu J, Sun W, Zhong J, et al. Phosphoglycerate mutase 1 regulates dNTP pool and promotes homologous recombination repair in cancer cells. J Cell Biol. 2017;216(2):409-424. 84. Paull TT. Mechanisms of ATM activation. Annu Rev Biochem. 2015;84:711-738. doi:10.1146/annurev-biochem-060614-034335. 85. Meuth M. The molecular basis of mutations induced by deoxyribonucleoside triphosphate pool imbalances in mammalian cells. Exp Cell Res. 1989;181(2):305- 316. doi:10.1016/0014-4827(89)90090-6. 86. Mathews CK. Deoxyribonucleotide metabolism, mutagenesis and cancer. Nat Rev Cancer. 2015;15(9):528-539. doi:10.1038/nrc3981. 87. Pai C, Kearsey S. A Critical Balance: dNTPs and the Maintenance of Genome Stability. Genes (Basel). 2017;8(2):57. doi:10.3390/genes8020057. 88. Watt DL, Buckland RJ, Lujan SA, Kunkel TA, Chabes A. Genome-wide analysis of the specificity and mechanisms of replication infidelity driven by imbalanced dNTP pools. Nucleic Acids Res. 2015;44(4):1669-1680. doi:10.1093/nar/gkv1298. 89. Kunz BA. Mutagenesis and deoxyribonucleotide pool imbalance. Mutat Res - Fundam Mol Mech Mutagen. 1988;200(1-2):133-147. doi:10.1016/0027-5107(88)90076-0. 90. Clifford R, Louis T, Robbe P, et al. SAMHD1 is mutated recurrently in chronic lymphocytic leukemia and is involved in response to DNA damage. Blood. 2014;123(7):1021-1031. doi:10.1182/blood-2013-04-490847. 91. Rentoft M, Lindell K, Tran P, et al. Heterozygous colon cancer-associated mutations of SAMHD1 have functional significance. Proc Natl Acad Sci. 2016;113(17):4723-4728. doi:10.1073/pnas.1519128113. 92. Kohnken R, Kodigepalli KM, Wu L. Regulation of deoxynucleotide metabolism in cancer: novel mechanisms and therapeutic implications. Mol Cancer. 2015;14:176. doi:10.1186/s12943-015-0446-6.

93. Aye Y, Li M, Long MJC, Weiss RS. Ribonucleotide reductase and cancer: biological mechanisms and targeted therapies. Oncogene. 2014;(April 2014):2011-2021. doi:10.1038/onc.2014.155. 94. Parker WB. Enzymology of Purine and Pyrimidine Antimetabolites Used in the Treatment of Cancer. Chem Rev. 2009;109(7):2880-2893. doi:10.1021/cr900028p. 95. Rampazzo C, Tozzi MG, Dumontet C, Jordheim LP. The druggability of intracellular nucleotide-degrading enzymes. Cancer Chemother Pharmacol. 2016;77(5):883-893. doi:10.1007/s00280-015-2921-6. 96. Amie SM, Daly MB, Noble E, Schinazi RF, Bambara RA, Kim B. Anti-HIV host factor SAMHD1 regulates viral sensitivity to nucleoside reverse transcriptase inhibitors via modulation of cellular deoxyribonucleoside triphosphate (dNTP) levels. J Biol Chem. 2013;288(28):20683-20691. doi:10.1074/jbc.M113.472159. 97. Hollenbaugh JA, Schader SM, Schinazi RF, Kim B. Differential regulatory activities of viral protein X for anti-viral efficacy of nucleos(t)ide reverse transcriptase inhibitors in monocyte-derived macrophages and activated CD4+ T cells. Virology. 2015;485:313-321. doi:10.1016/j.virol.2015.08.006. 98. Ballana E, Badia R, Terradas G, et al. SAMHD1 specifically affects the antiviral potency of thymidine analog HIV reverse transcriptase inhibitors. Antimicrob Agents Chemother. 2014;58(8):4804-4813. doi:10.1128/AAC.03145-14. 99. Ordonez P, Kunzelmann S, Groom HCT, et al. SAMHD1 enhances nucleoside-analogue efficacy against HIV-1 in myeloid cells. Sci Rep. 2017;7(February):42824. doi:10.1038/srep42824. 100. Hollenbaugh JA, Shelton J, Tao S, et al. Substrates and Inhibitors of SAMHD1. PLoS One. 2017;12(1):e0169052. doi:10.1371/journal.pone.0169052. 101. Schneider C, Oellerich T, Baldauf H-M, et al. SAMHD1 is a biomarker for cytarabine response and a therapeutic target in acute myeloid leukemia. Nat Med. 2016;(i). doi:10.1038/nm.4255. 102. Herold N, Rudd SG, Ljungblad L, et al. Targeting SAMHD1 with the Vpx protein to improve cytarabine therapy for hematological malignancies. Nat Med. 2017;23(2):256-263. doi:10.1038/nm.4265. 103. Jones RM, Petermann E. Replication fork dynamics and the DNA damage response. Biochem J. 2012;443(1):13-26. doi:10.1042/BJ20112100. 104. Bartkova J, Horejsí Z, Koed K, et al. DNA damage response as a candidate anti-cancer barrier in early human tumorigenesis. Nature. 2005;434(7035):864-870. doi:10.1038/nature03482. 105. van Montfoort N, Olagnier D, Hiscott J. Unmasking immune sensing of retroviruses: Interplay between innate sensors and host effectors. Cytokine Growth Factor Rev. 2014;25(6):657-668. doi:10.1016/j.cytogfr.2014.08.006. 106. Maelfait J, Bridgeman A, Benlahrech A, Cursi C, Rehwinkel J. Restriction by SAMHD1 Limits cGAS/STING-Dependent Innate and Adaptive Immune Responses to HIV-1. Cell Rep. 2016;16(6):1492-1501. doi:10.1016/j.celrep.2016.07.002. 107. Manel N, Hogstad B, Wang Y, Levy DE, Unutmaz D, Littman DR. A cryptic sensor for HIV-1 activates antiviral innate immunity in dendritic cells. Nature. 2010;467(7312):214-217. doi:10.1038/nature09337.

108. Brégnard C, Benkirane M, Laguette N. DNA damage repair machinery and HIV escape from innate immune sensing. Front Microbiol. 2014;5(APR):1-10. doi:10.3389/fmicb.2014.00176. 109. Arnold LH, Kunzelmann S, Webb MR, Taylor IA. A Continuous Enzyme-Coupled Assay for Triphosphohydrolase Activity of HIV-1 Restriction Factor SAMHD1. Antimicrob Agents Chemother. 2015;59(1):186-192. doi:10.1128/AAC.03903-14. 110. Seamon KJ, Stivers JT. A High-Throughput Enzyme-Coupled Assay for SAMHD1 dNTPase. J Biomol Screen. 2015;20(6):801-809. doi:10.1177/1087057115575150. 111. Huynh TN, Luo S, Pensinger D, et al. An HD-domain phosphodiesterase mediates cooperative hydrolysis of c-di-AMP to affect bacterial growth and virulence. Proc Natl Acad Sci U S A. 2014;112(7):1-10. doi:10.1073/pnas.1416485112. 112. Rao F, Qi Y, Murugan E, Pasunooti S, Ji Q. 2’,3’-cAMP hydrolysis by metal-dependent phosphodiesterases containing DHH, EAL, and HD domains is non-specific: Implications for PDE screening. Biochem Biophys Res Commun. 2010;398(3):500-505. doi:10.1016/j.bbrc.2010.06.107. 113. Meyer-klaucke W, Marchfelder A. Metal Requirements and Phosphodiesterase Activity of tRNase Z Enzymes †. Biochemistry. 2007;46(2007):14742-14750. 114. Keppetipola N, Shuman S. A Phosphate-binding Histidine of Binuclear Metallophosphodiesterase Enzymes Is a Determinant of 2’,3’-Cyclic Nucleotide Phosphodiesterase Activity. J Biol Chem. 2008;283(45):30942-30949. doi:10.1074/jbc.M805064200. 115. Keppetipola N, Shuman S. Characterization of the 2???,3??? cyclic phosphodiesterase activities of Clostridium thermocellum polynucleotide kinase-phosphatase and bacteriophage ?? phosphatase. Nucleic Acids Res. 2007;35(22):7721-7732. doi:10.1093/nar/gkm868. 116. Feng BY, Shoichet BK. A detergent-based assay for the detection of promiscuous inhibitors. Nat Protoc. 2006;1(2):550-553. doi:10.1038/nprot.2006.77. 117. Shoichet BK. Interpreting steep dose-response curves in early inhibitor discovery. J Med Chem. 2006;49(25):7274-7277. doi:10.1021/jm061103g. 118. Prinz H. Hill coefficients, dose-response curves and allosteric mechanisms. J Chem Biol. 2010;3(1):37-44. doi:10.1007/s12154-009-0029-3. 119. Baell JB, Holloway GA. New substructure filters for removal of pan assay interference compounds (PAINS) from screening libraries and for their exclusion in bioassays. J Med Chem. 2010;53(7):2719-2740. doi:10.1021/jm901137j. 120. Dahlin JL, Nissink JWM, Strasser JM, et al. PAINS in the assay: Chemical mechanisms of assay interference and promiscuous enzymatic inhibition observed during a sulfhydryl-scavenging HTS. J Med Chem. 2015;58(5):2091-2113. doi:10.1021/jm5019093. 121. Schorpp K, Rothenaigner I, Salmina E, et al. Identification of Small-Molecule Frequent Hitters from AlphaScreen High-Throughput Screens. J Biomol Screen. 2014;19(5):715-726. doi:10.1177/1087057113516861. 122. Baell J, Walters MA. Chemical con artists foil drug discovery. Nature. 2014;513:481- 483. doi:10.1038/513481a. 123. Pery E, Sheehy A, Miranda Nebane N, et al. Redoxal, an inhibitor of de novo

pyrimidine biosynthesis, augments APOBEC3G antiviral activity against human immunodeficiency virus type 1. Virology. 2015;484:276-287. doi:10.1016/j.virol.2015.06.014. 124. Usach I, Melis V, Peris J-E. Non-nucleoside reverse transcriptase inhibitors: a review on pharmacokinetics, pharmacodynamics, safety and tolerability. J Int AIDS Soc. 2013;16(1). doi:10.7448/IAS.16.1.18567. 125. Trott O, Olson AJ. AutoDock Vina: Improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem. 2009:NA-NA. doi:10.1002/jcc.21334. 126. Pettersen EF, Goddard TD, Huang CC, et al. UCSF Chimera?A visualization system for exploratory research and analysis. J Comput Chem. 2004;25(13):1605-1612. doi:10.1002/jcc.20084.

SUPPLEMENTAL FIGURES

Supplemental Figure 1. The SAMHD1 dNTPase assay is linear under the specified conditions. A) A schematic detailing the basic protocol followed for dNTPase assay described in this study. B) HPLC UV254 chromatographs from a representative dNTPase experiment demonstrate the formation of the dA product and the depletion of 500μM dATP substrate in the presence of 50μM GTP over a period of 30 minutes. C) The dNTPase reaction is linear through 20 minutes in the presence of varying H2O2 concentrations indicating the reactions described in this study were performed under steady state conditions.

Supplemental Figure 2. SAMHD1 tetramerization is inhibited by oxidation. A) Chemical crosslinking of recombinant wild type and C522A SAMHD1 demonstrate a progressive reduction in the ability of SAMHD1 to tetramerize following increasing exposure to H2O2. Pre-incubation with activating nucleotides, the addition of reducing agent, or the mutation of Cys522 to Ala mitigates the inhibitory effect of H2O2. B) Quantitation of crosslinked SAMHD1 (as in panel A) showing averages from three independent crosslinking experiments. C) SEC-MALS experiments demonstrate a similar trend in which H2O2 reversibly inhibits SAMHD1 tetramerization. Molar Mass is denoted above the respective peaks which represent UV280 absorbance traces of protein elution from an analytical size exclusion chromatography column. D) Average percent of SAMHD1 as a tetramer under indicated conditions from three independent SEC-MALS experiments.

Supplemental Figure 3. Sequence alignment of metazoan SAMHD1 sequences reveals multiple conserved cysteine residues. Conserved amino acid residues are indicated by asterisks and the conserved cysteines are highlighted.

Supplemental Figure 4. LC-MS/MS identifies the dimedone adduct on cysteine 341, 350, and 522 under oxidizing conditions.

Supplemental Figure 5. The C341-350 disulfide bond is the predominant disulfide bond identified under oxidizing conditions. LC-MS/MS analysis of tryptic peptides derived from SAMHD1 wild type exposed to varying concentrations of H2O2 show that by 1μM H2O2 50% of the disulfide bonds identified by pLink-SS software were between Cys341 and Cys350.

Supplemental Figure 6. SAMHD1 Cys to Ala mutants are structurally similar with the exception of C341-350A. Molar ellipticity at 222nm indicates consistent folding patterns and alpha- helix content for single Cys to Ala mutants in SAMHD1. The C341-350A double mutant demonstrates reduced alpha helical content suggesting an altered folding pattern or reduced structural integrity.

Supplemental Figure 7. The SAMHD1 C522A mutant retains dNTPase capacity within cells following treatment with PDGF. Serum starved fibroblasts transiently expressing SAMHD1 wild type, C522A, or an empty vector control, were treated with 20nM PDGF or an equivalent volume of PBS and harvested after 60 minutes, and the cell lysates tested for dNTPase activity. In contrast to both the empty vector and wild type samples, the C522A samples exhibited no reduced dNTPase capacity.

Supplemental Figure 8. DCP-Bio1 pull down protocol is specific to oxidized SAMHD1. Western blot analysis of cell lysates in which no DCP-Bio1 is included show no protein band, indicating that the interaction between SAMHD1 and the streptavidin beads utilized during the extraction protocol is specific to DCP-Bio1 labeled SAMHD1.

Supplemental Figure 9. Protein oxidation and capture by dimedone based probes. The horizontal pathway represents a simplified schematic of protein oxidation on cysteine residues. ROS generated by a proliferative signal can oxidize the cysteine sulfhydryl group to form the transient sulfenic acid, which can then be irreversibly oxidized to sulfinic or sulfonic acid. Alternatively, sulfenic acid can condense to form an intra or interchain disulfide, which can be utilized to confer functional changes to the oxidized protein. The vertical pathway represents the capture of the transient sulfenic acid species, and the available analysis methods for the labeled protein once is has covalently bound the dimedone based probe.

Supplemental Figure 10. Uncropped western blots.

CHAPTER 2

Redox regulation of SAMHD1 in SKOV3 cells modulates nucleotide pools and cell cycle progression

Christopher H. Mauney1, LeAnn C. Rogers1, Reuben S. Harris4, Larry W. Daniel1,2, Cristina M. Furdui2,3, Fred W. Perrino1, Thomas Hollis1,2*

ABSTRACT

SAMHD1 is a dNTP triphosphohydrolase that performs an essential function in the control of nucleotide metabolism. Given the importance of nucleotide metabolism to proper DNA replication and repair, and in a broader context, genomic integrity, the regulation of

SAMHD1 activity is tightly controlled. SAMHD1 is a redox regulated enzyme that is oxidized in cells in response to proliferative signaling. Oxidation and reduction of SAMHD1 at a redox switch proximal to the regulatory nucleotide binding site cycles the enzyme between active and inactive conformations. In this study we investigate the physiological importance of SAMHD1 redox regulation for proper nucleotide metabolism and cell cycle progression in SKOV3 cells. We demonstrate that cells stably expressing a redox insensitive mutant of SAMHD1 (C522A) are less susceptible to mitogenic signaling induced ROS that inhibits cellular dNTPase capacity, have diminished dNTP pools, and proliferate at a reduced rate characterized by an accumulation in G2/M phase. These findings underscore the importance of SAMHD1 redox regulation for normal cellular functioning.

INTRODUCTION

Proper nucleotide metabolism is essential for genomic integrity1,2. DNA replication and repair are two central processes that maintain genomic integrity while ensuring accurate transfer of genetic information to progeny cells. In order to achieve the precision required of these processes, deoxynucleotide triphosphate (dNTP) precursors must be available to the replication and repair machinery at appropriate levels. Too many or too few dNTPs can result in a deleterious effect on genomic fidelity and have harmful consequences for cellular homeostasis and cell cycle progression3–8. The catabolic and anabolic pathways of dNTP metabolism are themselves meticulously regulated processes9. dNTP pools expand just prior to S-phase when cells require large quantities for DNA replication. In non-dividing cells, dNTP pools must be present in order to supply basal repair processes, but maintained at reduced levels to avoid potentially mutagenic conditions1.

SAMHD1 is a dNTP triphosphohydrolase that performs an important function in nucleotide metabolism. It catalyzes the reaction in which a dNTP is hydrolyzed into its cognate nucleoside and inorganic triphosphate10,11. SAMHD1 accomplishes its catalytic function by first binding two activating nucleotide triphosphates at distinct regulatory sites, which in turn result in the formation of the catalytically active tetramer12,13. Through its dNTPase activity SAMHD1 is able to maintain reduced dNTP levels within the cell14,15. The importance of SAMHD1 in regulating nucleotide pools is made evident by the fact that mutations to SAMHD1 result in disease phenotypes indicative of aberrant nucleotide metabolism, such as autoimmunity and cancer16–19.

SAMHD1 is broadly expressed in most tissue types, but its expression levels and catalytic activity vary by stage in the cell cycle20. Non-dividing cells in G0/G1 exhibit elevated levels of SAMHD1 presumably to maintain reduced levels of dNTPs21. Phosphorylation of

SAMHD1 at Thr592 by cyclin dependent kinases prior to S-phase destabilize the catalytically active tetramer and enable the accumulation of dNTPs required for DNA synthesis22–27. Recently, our group has discovered that SAMHD1 activity is further regulated through a redox switch28,29. This switch is comprised of three Cysteine residues (Cys522,

Cys341, and Cys350) in close proximity to the regulatory nucleotide binding sites that can form, and subsequently resolve, a disulfide bond that cycles the enzyme between an active and inactive conformation. This redox switch is oxidized in cells by reactive oxygen species generated in response to proliferative signaling. We proposed a model in which reactive oxygen species (ROS) acting as second messengers produced by cellular oxidases are detected by SAMHD1 at Cys522. Oxidation at Cys522 is then translated into structural changes via the redox switch that inactivates the enzyme and enables an increase in dNTP pools required for DNA replication.

This study represents a preliminary investigation of the importance of SAMHD1 redox regulation for proper nucleotide metabolism and cell proliferation. We accomplish this by utilizing SKOV3 cells engineered to stably overexpress either wild type SAMHD1 or a redox insensitive mutant, C522A. SKOV3 cells are an ovarian epithelial cancer cell line that was selected due the extensively studied redox signaling pathways30. When treated with lysophosphatidic acid (LPA), a signaling molecule that activates mitogenic pathways by binding LPA receptors31,32, SKOV3 cells generate a robust redox signaling cascade

characterized by high levels of ROS. Here we show that SAMHD1 is oxidized in SKOV3 cells in response to LPA. Redox insensitive SAMHD1-C522A is not susceptible to the downstream effects of redox signaling, however, and cells expressing C522A retain enhanced dNTPase activity, reduced dNTP pools, and impeded proliferative capacity characterized by an accumulation in G2/M phase. These findings underscore the importance of SAMHD1 regulation by redox signaling for proper cellular functioning, and lay the groundwork for more rigorous experimental investigation.

RESULTS

SAMHD1 is oxidized in SKOV3 cells in response to LPA. In previous work, we demonstrated that SAMHD1 is oxidized in cells in response to various growth factors.

While the redox signaling pathways of SKOV3 cells have been extensively studied, it was important to confirm that SAMHD1 was oxidized in SKOV3 cells, as well. Therefore, we performed a DCP-Bio1 affinity capture assay on SKOV3 cells treated with LPA for 10 or 30 minutes, or an equivalent volume of PBS as control. The DCP-Bio1 probe used in this assay covalently binds oxidized cysteine residues and can be extracted using the linked biotin moiety. The biotin pull-down is then blotted using a protein specific antibody. This technique revealed that oxidized SAMHD1 increases over the time course examined relative to total SAMHD1, which remains constant at each time point (Fig. 1). This indicates that

SAMHD1 is oxidized in SKOV3 cells in response to LPA signaling.

Figure 1. SAMHD1 is oxidized in SKOV3 cells in response to LPA treatment. DCP-Bio1 affinity capture of oxidized proteins in SKOV3 cells labels SAMHD1. Captured protein was resolved by SDS- PAGE, blotted onto nitrocellulose and probed with α-SAMHD1 antibodies. SAMHD1 oxidation increases post LPA stimulation. Oxidized SAMHD1 and total SAMHD1 in the lysate are indicated.

The C522A mutant SAMHD1 retains enhanced dNTPase capacity compared to parental or wild type cells. Following confirmation that SAMHD1 is oxidized in cells, we next interrogated the effect of redox regulation on SAMHD1 and its ability to modulate dNTPase activity in cells. Asynchronous parental SKOV3 cells or SKOV3 cells stably expressing wild type or C522A SAMHD1 were harvested and the dNTPase capacity of the cell lysate was tested. Given that SAMHD1 is the only known dNTP triphosphohydrolase in humans, any observed dNTPase activity was presumed to originate from SAMHD1 contained within the cell lysate. Additionally, in an example of autocrine signaling, SKOV3 cells make their own LPA to stimulate proliferation pathways. Asynchronous cells that are actively proliferating should therefore be generating and responding to LPA in the culture

media. SKOV3 cells stably expressing the C522A mutant demonstrated a two-fold increase in the dNTPase capacity of the cell lysate compared to the wild type or the parental cell lines

(Fig. 2). This effect was observed even though SAMHD1 was expressed at equivalent levels in the wild type and C522A samples. Interestingly, the dNTPase capacity of the cell lysate from wild type and parental cells was comparable despite significantly less endogenous

Figure 2. C522A mutants retain enhanced catalytic activity. Results depict the dNTPase activity of SKOV3 cell lysates stably expressing wild type SAMHD1, C522A, or parental cells. Data was normalized to parental samples. SAMHD1 samples from cell lysates run equivalent to recombinant SAMHD1 by SDS-PAGE/Western Blot analysis and are present at similar concentrations in stably transfected cells.

SAMHD1 observed in the parental samples. These data support the idea that redox insensitive C522A is refractory to LPA induced oxidation and subsequent catalytic inactivation. As a corollary, they also suggest that redox regulation is an important mechanism for controlling SAMHD1 activity in cells regardless of protein levels.

Redox regulation of SAMHD1 by oxidation controls intracellular dNTP concentration.

In order to connect dNTPase activity levels of SAMHD1 to altered physiological outcomes in cells, we next sought to quantify intracellular dNTP levels in SKOV3 cells stably expressing wild type SAMHD1 or the C522A mutant. For this experiment, asynchronous SKOV3 cells were harvested and the dNTPs extracted using methanol. Mass spectrometry was then used to quantify intracellular dNTPs. Picomoles of dATP and dCTP per million cells were determined as a representative purine and pyrimidine classes of nucleotides. The mass spectrometry data reveals a depletion of dNTPs in both the wild type and C522A samples relative to parental cells (Fig. 3). Most striking is the reduction in dCTP concentrations.

This data aligns with biochemical data that suggests that dCTP is a preferred substrate of

SAMHD1 given its ability to form an extensive network of weak interactions in the active site. These results agree with the observed cell lysate dNTPase activities, and provide additional evidence for the redox regulation of SAMHD1 activity, as well as the importance of redox signaling for modulating intracellular dNTP pools.

100

Figure 3. Expression of C522A mutants results in reduced intracellular dNTP pools. Single clones of SKOV3 cells stably expressing wild type or C522A were harvested and dNTPs extracted for quantification by mass spectrometry. dATP and dCTP were measured as representatives of purine and pyrimidines nucleotide classes. Wild type and C522A both exhibit reduced dNTP pools compared to parental SKOV3 cells.

Proliferation rate is controlled by SAMHD1 redox regulation. Several groups have reported that levels of wild type SAMHD1 expression are inversely correlated with proliferation rates. We hypothesized that overexpression or constitutive activation of

SAMHD1 would result in a reduced growth rate due to depleted intracellular dNTP pools.

In order to test our hypothesis we performed a series of proliferation assays using the MTS reagent. Equal numbers of parental, wild type, or C522A expressing SKOV3 cells were plated in 96 well plates, and their proliferation measured over a 96 hour time course. The data supported our hypothesis, as C522A expressing cells grew at approximately 60% the rate of parental cells (Fig. 4). Wild type SAMHD1 expressing cells also demonstrated a minimal, but statistically significant, reduction in their fold growth over a 96 hour period compared to parental cells. These data imply the dNTPase function of SAMHD1 is an important mediator of cell cycle progression, and that redox regulation of SAMHD1 is a key effector for controlling this mechanism.

101

Figure 4. C522A mutants inhibit cell proliferation. Cell proliferation assessed using the MTS assay demonstrates that (A) SKOV3 cells stably expressing the redox insensitive SAMHD1 mutant C522A proliferate at a reduced rate compared to wild type or parental cells. (B) Quantification of fold growth at 96 hours post seeding reveal 40% reduction in growth normalized to the initial time point by C522A cells relative to parental cells. Wild type cells demonstrate a modest, but significant growth reduction as well (* p<0.05).

Redox regulation of SAMHD1 is required to properly progress through the cell cycle.

In order to further investigate the effect of SAMHD1 redox regulation on cell cycle progression, we performed a series of cell cycle analyses using flow cytometry and propidium iodide staining. We projected that due to the reduced dNTP pools cells would be

102

Figure 5. C522A mutants accumulate in G2/M phase. (A) Cell cycle analysis of SKOV3 cells using flow cytometry and propidium iodide staining of DNA content. Representative histograms consist of cells gated to preclude doublets and at least 10,000 cell counts. Green-G1, Purple-S phase, Orange- G2/M phase. (B) Quantification of flow cytometry results from A reveal that cells expressing C522A, and to a lesser extent wild type, accumulate in G2/M phase

arrested in G1 or S-phase as they would not be able to complete the replication of their genomic DNA during cell division. While we observed that C522A expressing SKOV3 cells did indeed demonstrate altered cell cycle progression, it was not aligned with our hypothesis of G1/S phase arrest. C522A cells instead demonstrated an inhibition of cell cycle progression in G2/M phase. Approximately 35% more cells were observed in G2/M phase in C522A samples than in parental SKOV3 cells. Wild type SAMHD1 cells also

103

demonstrated a modest 20% increase in G2/M phase cells, although compared to parental cells the effect was not statistically significant. These data further support the importance of redox regulation of SAMHD1 for proper progression through the cell cycle. Further, the cell cycle arrest during G2/M phase in cells overexpressing wild type or C522A implies that the G2/M phase DNA damage checkpoint is activated as a function of dNTP depletion.

Phosphorylation makes SAMHD1 more susceptible to oxidation. SAMHD1 regulation in cells is a dynamic process with multiple pathways that finely tune its dNTPase activity.

The most thoroughly investigated regulatory mechanism of SAMHD1 is phosphorylation.

SAMHD1 is phosphorylated at T592 by cyclin dependent kinases in a cell cycle dependent manner22,33. While the precise effect of phosphorylated SAMDH1 is still the subject of debate, the phosphorylation status of SAMHD1 is important for both cell cycle progression and innate immune function. Through previous work and the data presented in this study, we have identified redox regulation as another important signaling pathway for linking

SAMHD1 dNTPase activity to cell cycle progression. In order to asses if these overlapping signaling pathways are able influence each other we performed a dNTPase assay in the presence of increasing concentrations of H2O2 using recombinant wild type SAMHD1,

C522A, and a phosphomimetic mutant T592E (Fig. 6). Wild type SAMHD1 demonstrated the expected dose response where dNTPase activity was correspondingly inactivated by increasing concentrations of H2O2. Under these conditions, 50% of wildtype activity was inhibited at 65µM H2O2. C522A also demonstrated the expected response and was generally refractory to H2O2 mediated inhibition within a physiological range (<100µM).

104

Figure 6. The SAMHD1 phosphomimetic mutant T592E is more susceptible to oxidation than wild type. SAMHD1 wild type, C522A, and T592E were treated with increasing concentrations of H2O2 and then assayed for dNTPase activity. All data was normalized to the 0µM H2O2 condition which represented maximum activity equal to 100%. Dose response curves were fit and the concentration at which SAMHD1 was 50% inactivated (Kin) was determined. Under these conditions, T592E was revealed to be more susceptible to oxidation than wild type or C522A.

T592E displayed a dose dependent inactivation of dNTPase activity similar to wild type, but exhibited 50% inactivation at approximately half the H2O2 concentration required for wild type. This finding suggests that phosphorylated SAMHD1 is more susceptible to oxidation than unphosphorylated SAMHD1. This data aligns with studies that demonstrate phosphorylated SAMHD1 results in a less stable tetramer, and that tetramerization is able to protect SAMHD1 dNTPase activity from H2O2 mediated inactivation. In a physiological context, these results suggests that redox and phosphorylation pathways may not act in isolation, but instead function synergistically to regulate the dNTPase activity of SAMHD1.

105

DISCUSSION

This study is intended as a preliminary investigation into the importance of redox regulation of SAMHD1 in cells. It builds upon published work that identifies SAMHD1 as a redox regulated enzyme that is oxidized in cells in response to proliferative signals28. In order to test our hypothesis that redox regulation of SAMHD1 is a major determinant of nucleotide homeostasis and cell cycle progression, we utilize SKOV3 cells stably overexpressing SAMHD1 wild type or the C522A mutant, which is a SAMHD1 mutant that is resistant to ROS induced catalytic inactivation. Using these tools, we demonstrate that cells expressing the redox refractory C522A mutant exhibit increased catalytic activity, reduced dNTP pools, and impeded proliferation characterized by G2/M phase arrest. These data support our model of redox regulation as an essential means of controlling SAMHD1 activity in a cell cycle dependent manner.

While this is a preliminary study, several aspects of the data nevertheless stand out. Cancer cells, such as SKOV3, are characterized by their uncontrolled proliferative capacity.

Accelerated growth requires an excess of precursor metabolites, like dNTPs, in order to rapidly synthesize cellular macromolecules. It stands to reason then that cancer cells must down regulate the activity of SAMHD1 in order to supply themselves with the requisite building blocks for rapid growth. The ability of SKOV3 cells to restrict over-expressed wild type dNTPase activity, but not that of C522A, in the absence of exogenously applied LPA, suggests that SKOV3 cells inherently utilize redox signaling to accomplish SAMHD1 downregulation. When coupled with evidence that cancer cells maintain an increased oxidizing

106

environment due to upregulated growth factor signaling and inefficient aerobic respiration, this data hints at a broader growth strategy employed by cancer cells34,35. It is possible that one mechanism by which cancer cells supply themselves with the metabolites for rapid proliferation is by maintaining an altered redox state that down-regulates catabolic pathways such as dNTP degradation by SAMHD1. Further investigation of this phenomena in normal and cancer cells will yield important insights about the role of SAMHD1 in cell cycle regulation and the ability of cancer cells to circumvent its restriction of uncontrolled growth.

The cell cycle analysis by flow cytometry presents another intriguing set of data. Originally we had hypothesized that the redox insensitive C522A mutant would inhibit cell proliferation via dNTP depletion, and that this effect would arrest the cell cycle in G1 or S phase. While the C522A mutant did modify the ability of cells to proceed through the cell cycle, it was not in the manner we anticipated. Instead, cells overexpressing C522A, and to a lesser extent wild type, accumulated in G2/M phase. G2/M phase accumulation implied activation of the G2/M phase checkpoint which halts entry to mitosis in the presence of DNA damage7,8,36,37. This outcome was a symptom of dNTP depletion resulting from the presence of redox insensitive C522A variant of SAMHD1, or simply the overexpression of

SAMHD1 wild type that overwhelmed extant cellular redox systems. Imbalanced or depleted dNTP pools result in an activation of the ATM/ATR DNA damage response pathway through a series of genomic malfunctions including altered kinetics of DNA replication, stalled replication forks, increased DNA breakage, and reduced replication fidelity and proofreading efficiency38,39. Although the precise mechanism is unknown, our data suggest that accumulation of SKOV3 in G2/M phase is indicative of DNA damage

107

instigated by dNTP depletion. This further supports the central role of redox regulation of

SAMHD1 for proper cell growth and establishes a rationale for a more thorough inquiry into the downstream effects of SAMHD1 regulation and its relationship to genomic fidelity. It is also interesting to consider that while our study measured dNTP depletion by constitutively active SAMHD1, AGS linked mutations that inactivate SAMHD1 and result in imbalanced dNTP pools may have a similar effect on the DNA replication and repair machinery.

Lastly, our data demonstrates that the phosphomimetic mutant of SAMHD1 is more susceptible to oxidation. This finding is important for understanding the ways in which converging regulatory pathways communicate in order to modulate enzyme function. Both phosphorylation and redox signaling are important means of cellular communication that regulate SAMHD1 activity in relation to cell cycle stage. Dissecting the precise spatiotemporal features of SAMHD1 modifications and the integrated effect they exert will be important for a complete understanding of the functioning of SAMHD1 in cells.

Additional biochemical and cell culture studies that elucidate the channels through which these orthogonal pathways influence each other in relation to SAMHD1 activity will be critical going forward.

The observations presented in this study offer further evidence that SAMHD1 is a redox regulated enzyme, and that redox control of SAMHD1 activity is an essential mechanism of maintaining dNTP pools for proper cell cycle progression. While this study was limited in scope, it lays the groundwork for a more intensive examination of the pathways through which cells regulate SAMHD1 and the downstream effect of SAMHD1 posttranslational modifications on nucleotide homeostasis.

108

MATERIALS AND METHODS

Reagents and Antibodies. Primary antibodies for Western blots to SAMHD1 were from

Sigma- Aldrich. Biotin-HRP and HA antibodies were purchased from Cell Signaling

Technology. AlexaFluor fluorescent secondary antibodies, and RPMI 1640 medium were from Invitrogen. Fetal bovine serum was from Lonza. Super Signal chemiluminescence reagent was from Pierce. Alkyl-linked 18:1 lysophosphatidic acid [1-(9Z- octadecenyl)-2- hydroxy-sn-glycero-3-phosphate (ammonium salt)] was from Avanti Polar Lipids, Inc. DCP-

Bio1 and DCP-Rho1 were obtained from Dr. Bruce King (Wake Forest University, Dept. of

Chemistry). Stable isotope labeled dNTPs were purchased from Sigma-Aldrich.

Cell Culture and treatments. SKOV3 cells (from ATCC stocks) were grown, maintained, and, unless otherwise indicated, treated at 37oC with 5% CO2 in RPMI 1640 medium supplemented with 10% fetal bovine serum, L-glutamine, penicillin, and streptomycin. LPA, supplied in chloroform, was dried under a stream of nitrogen, resuspended to a concentration of 1mM in phosphate buffered saline (PBS) containing 1% fatty acid-free bovine serum albumin (BSA), and then diluted into culture medium to the indicated concentrations.

Generation of stably transfected SKOV3 cells. Full length human SAMHD1 wild type and

C522A containing an N-terminal HA tag and hygromycin B resistance was cloned into the eukaryotic expression vector pCDNA3.1 Synthesis and sequencing were performed by

GenScript. Transfection was performed using a lipid based transfection reagent (Fugene,

109

Promega) and colonies were grown in selection media supplemented with hygromycin B.

Western Blotting. Cells were washed with cold, calcium-free PBS, scraped into lysis buffer

(50 mM Tris-HCl, 100 mM NaCl, 2 mM EDTA, 0.1% SDS, 0.5% sodium deoxycholate, 1 mM

PMSF, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 50 mM NaF, and 1 mM sodium vanadate), and centrifuged to remove cell debris after one freeze/thaw cycle. Protein concentration was measured (Pierce BCA protein assay) and samples (typically 10 μg protein/lane) were resolved on SDS polyacrylamide gels, then transferred to nitrocellulose membranes, probed with protein-specific antibodies and visualized using Super Signal chemiluminescence reagent.

DCP-Bio1 labeling and affinity capture. Cells (~5×105) were grown in 100-mm plates for 48 hours as described above, and after incubation of 100 nM alkyl-LPA with the cells for indicated time points, DCP-Bio1 was added with the lysis buffer to chemically trap sulfenic acid proteins as described by Klomsiri, et al40 with slight modifications. Lysis buffer was freshly prepared and supplemented with 1 mM DCP-Bio1, 10 mM NEM, 10 mM iodoacetamide, and 200 U/ml catalase. Use of NEM and iodoacetamide together provided the best blocking of free thiols. Lysis buffer (150 µl/plate) was added to each plate, and then cells were scraped from the plates, transferred to Eppendorf tubes, incubated on ice for 30 min before storing at −80 °C. For affinity capture and elution of the labeled proteins, unreacted DCP-Bio1 was removed immediately on thawing using a BioGel P6 spin column.

Samples were then assayed for protein concentration, diluted to 1.0 mg/ml protein into 400

µl buffer containing a final concentration of 2 M urea, and supplemented with prebiotinylated AhpC (0.5 µg/400 µg of total lysate) to control for the efficiency of the

110

affinity capture and elution procedures. Samples were precleared with Sepharose CL-4B beads (Sigma), applied to columns containing high capacity streptavidin- agarose beads from Pierce, and then incubated overnight at 4 °C. Multiple stringent washes of the beads were performed (at least 4 column volumes and two washes each) using, in series, 1% SDS,

4 M urea, 1 M NaCl, 10 mM DTT, 50 mM ammonium bicarbonate and water, before elution with Laemmli sample buffer(27). Samples were stored at −80 °C as needed and analyzed by

Western blot as described above.

Cell Lysate dNTPase Assay. Stably transfected (Wild Type and C522A) and parental cells

(0.25x106 cells) were plated in 100mm dishes in 10ml of media. 48 hours after seeding, cells were washed with cold PBS 60 minutes post treatment, harvested by trypsinization, pelleted, and lysed in 100μl of lysis buffer (50mM Tris pH 7.5, 150mM NaCl, 2mM MgCl2,

1mM EDTA, 10% glycerol, 0.05% NP-40, 1% Triton-X, 100μM PMSF, 200 U/ml catalase,

10ng/ml Aprotinin, and 10ng/ml Lupeptin). Cell debris was removed by centrifugation at

10,000 r.c.f. for 10 minutes and protein content for each sample determined using a BCA assay. 100μg of protein was then used to conduct a dNTPase assay. Cell lysate was mixed with reaction buffer (final concentration 50mM Tris pH 7.5, 100mM NaCl, 5mM activating

MgCl2, 0.1mM EDTA) and the reaction initiated by the addition of dGTP to a final concentration of 1mM. Reactions were quenched using EDTA to a final concentration of

25mM after 6 hours, heated to 70C for 5 min, and centrifuged at 15000 r.c.f. for 10 min to remove protein. The supernatant was then analyzed by HPLC, using ion pair reverse phase chromatography on a Waters HPLC system. A CAPCell PAK C18 column (Shiseido Fine

Chemicals) was equilibrated with 20mM NaH2PO4 pH 7.0, 5mM tetra n-butylammonium phosphate, and 5% methanol. Reactants and products were eluted with a linear gradient of

111

methanol from 5-50%. A254 was used to measure eluting peaks, and quantification of the relative integrated areas under the dG and dGTP peaks was performed using the Empower

Software.

SAMHD1 deoxynucleotide triphosphohydrolase and H2O2 dose response assay. Individual protein aliquots were thawed and fully reduced by adding 10mM DTT and 2mM TCEP and allowed to incubate at 21oC for 20 minutes. Following incubation with reducing cocktail, enzyme samples were passed through a microspin column packed with BioGel-P6 (BioRad) and equilibrated with 50mM NaH2PO4 pH 7.5, 200mM NaCl, 1mM MgCl2, 0.1 mM EDTA, and

2% glycerol, in order to remove all reducing agents from the sample. The concentration of

SAMHD1 containing samples was then determined using a Nanodrop 2000 (Thermo

Scientific) and the extinction coefficient of fully reduced SAMHD1 (ε =75750 M-1 cm-1,

A280). SAMHD1 was then diluted into a reaction buffer (20mM Tris pH 7.5, 5mM MgCl2,

1μM EDTA) and incubated with H2O2 at the specified concentrations for 30 minutes at 21 °C.

Reactions were initiated by the addition of 500 μM dATP and 50μM GTP and 50 units of catalase to scavenge H2O2. Formation of dA was used to measure the progress of the reaction. SAMHD1 concentration in all reactions was 500nM. Reactions were quenched using EDTA to a final concentration of 20mM after 15 minutes and analyzed by HPLC.

Cell Proliferation MTS assay. MTS assays (CellTiter 96® AQueous One Solution Cell

Proliferation Assay (MTS), Promega) were carried out according to the manufacturer’s instructions. Briefly, 2000 cells were plated in a 96 well plate in triplicate for each time point. At the indicated time point, 40µl of the MTS reagent was added to each well, and the plate returned to the 37oC incubator for 2 hours. Following incubation, absorbance readings

112

were taken at 490nm on a Tecan Infinite M1000Pro Plate Reader. A 690nm absorbance reading was also taken for each well and subtracted as a baseline control. Absorbance values at each time point were normalized to the initial reading (24 hours post seeding) in order to determine fold growth.

Cell Cycle analysis using Flow Cytometry. Asynchronous SKOV3 cell were harvested by trypsinization, pelleted, and washed in ice cold PBS + 1% fetal bovine serum. Cells were counted and 1.0x106 cells transferred to clean 15ml conical tube before centrifuging again.

The resulting pellet was resuspended in 1ml cold PBS + 1% fetal bovine serum, and added dropwise to 4ml of ice-cold ethanol while slowly vortexing in order to fix and permeabilize cells. Cells suspensions were stored at -20oC until further analysis. When preparing cells for analysis, the samples were thawed at 21oC for 10 minutes before gently pelleting the cells.

The ethanol was removed and the cells washed in 2ml PBS + 1% fetal bovine serum. The cells were again gently pelleted, before resuspending in a Propidium Iodide/RNAse solution

(36mM Trisodium Citrate, 100mM NaCl, 0.6% NP-40, 50µg/ml propidium iodide, 0.4 mg/ml

RNAse) and incubating at 37oC for 30 minutes in the dark. Samples were analyzed using a

FACSCanto II Flow Cytometer (Becton Dickinson). Data analysis was performed on BD

FACSDiva™ software and appropriate gating controls were implemented to ensure that doublets were not included in the analysis.

Quantification of intracellular dNTPs. SKOV3 cells were harvested by trypsinization, pelleted, and washed by resuspension in PBS+1% fetal bovine serum. Cells were counted, and 1.0x106 cells removed and centrifuged again. The resulting pellet was lysed in 500µl

65% ice-cold methanol. 10 picomoles of two stable isotope labeled nucleotides (dGTP-

113

13C10,15N5, TTP-13C10 ,15N2) were added to the methanol-cell lysate solution to serve as an internal control for data normalization and extraction efficiency. Samples were then stored at -20oC for at least 1 hour to fully extract nucleotides. Following extraction, samples were vortexed for 30 seconds, heated at 95oC for 3 minutes, and then centrifuged at 15K r.c.f. to remove precipitated protein and residual cell debris. The supernatant was transferred to a clean eppendorf tube and evaporated under a vacuum and using a SpeedVac centrifuge.

The dried pellet was resuspended in 50µl of Mobile Phase A (3mM hexylamine, 5mM ammonium bicarbonate) for analysis. A Shimadzu LCMS-8050 triple-quadrupole mass spectrometer running in negative mode, combined with Shimadzu Prominence UPLC system was used to quantitate dNTPs. 10µl of each sample was injected onto a Waters

Xterra MS C18 (3.5 µm, 1.0 x 100 mm) column equilibrated with Mobile Phase A, and eluted at 0.5 ml/min using a 0-50% linear gradient of Acetonitrile over the course of 10 minutes.

Data analysis was performed by integrating the area under the peak of each respective dNTP identified by its unique ionization signature and normalizing to the internal standards.

Data and Statistical Analysis: Charts and graphs were generated using GraphPad Prism v7.00 for Windows (GraphPad Software, La Jolla California USA, www.graphpad.com), which was also used for statistical analysis and nonlinear regression. Statistical significance was determined at p<0.05 (*), and all error bars represent the standard deviation of three independent replicates.

114

Author Contributions

Conceptualization- C.H.M., L.C.R., T.H.; Methodology- C.H.M., L.C.R., T.H., Investigation-

C.H.M, L.C.R. Writing-original draft- C.H.M Writing-review & editing- C.H.M, L.C.R, T.H.,

Supervision- T.H Funding Acquisition- T.H, F.W.P., R.S.H., L.W.D

Acknowledgments

We thank Dr. Jingyun Lee and Anderson Cox of the Metabolomics and Proteomics Shared

Resource of the Wake Forest School of Medicine for their assistance in the performance and analysis of mass spectrometry experiments. We would like to acknowledge Dr. David

Ornelles for his assistance with flow cytometry experiments. This work has been supported through funding from the NIH (RO1 GM108827 to TH, RO1 CA142838 to LWD,

R01 GM06692 to FWP, and R33 CA177461 to LBP and T32 GM095440 support for CHM).

Abbreviations

AGS - Aicardi-Goutieres syndrome

BSA - bovine serum albumin

ROS - reactive oxygen species dATPαS - deoxyadenosine triphosphate alpha S

DCP-Bio1 - dimedone coupled probe to biotin

DCP-Rho1 - dimedone coupled probe to rhodamine

115

dNTP - deoxynucleotide triphosphate

DTT - dithiothreitol

EDTA - ethylenediaminetetraacetic acid

GTP - guanosine triphosphate

H2O2 - hydrogen peroxide

IPTG - isopropyl-β-D-thio-galactoside

LPA - lysophosphatidic acid

NEM - N-ethylmaleimide

PBS - phosphate buffered saline

PBS-T - phosphate buffered saline, 0.1% Triton X100

PDGF - platelet derived growth factor

PMSF - phenylmethylsulfonyl fluoride

R-SOH - sulfenic acid

R-SO2H - sulfinic acid

R-SO3H - sulfonic acid

SDS-PAGE - sodium dodecyl sulfate polyacrylamide gel electrophoresis

TCEP - Tris(2-carboxyethyl)phosphine

Kin - [H2O2] at which point SAMHD1 is 50% inactive

116

REFERENCES

1. Reichard P. Interactions Between Deoxyribonucleotide And DNA Synthesis. Annu Rev Biochem. 1988;57(1):349-374. doi:10.1146/annurev.biochem.57.1.349.

2. Mathews CK. Deoxyribonucleotides as genetic and metabolic regulators. FASEB J. 2014;28(9):3832-3840. doi:10.1096/fj.14-251249.

3. Bebenek K, Roberts JD, Kunkel TA. The effects of dNTP pool imbalances on frameshift fidelity during DNA replication. J Biol Chem. 1992;267(6):3589-3596.

4. Chabes A, Stillman B. Constitutively high dNTP concentration inhibits cell cycle progression and the DNA damage checkpoint in yeast Saccharomyces cerevisiae. Proc Natl Acad Sci U S A. 2007;104(4):1183-1188. doi:10.1073/pnas.0610585104.

5. Buckland RJ, Watt DL, Chittoor B, Nilsson AK, Kunkel TA, Chabes A. Increased and Imbalanced dNTP Pools Symmetrically Promote Both Leading and Lagging Strand Replication Infidelity. PLoS Genet. 2014;10(12). doi:10.1371/journal.pgen.1004846.

6. Kumar D, Viberg J, Nilsson AK, Chabes A. Highly mutagenic and severely imbalanced dNTP pools can escape detection by the S-phase checkpoint. Nucleic Acids Res. 2010;38(12):3975-3983. doi:10.1093/nar/gkq128.

7. Bester AC, Roniger M, Oren YS, et al. Nucleotide deficiency promotes genomic instability in early stages of cancer development. Cell. 2011;145(3):435-446. doi:10.1016/j.cell.2011.03.044.

8. Pajalunga D, Franzolin E, Stevanoni M, et al. A defective dNTP pool hinders DNA replication in cell cycle-reactivated terminally differentiated muscle cells. Cell Death Differ. 2017:1-11. doi:10.1038/cdd.2017.4.

9. Rampazzo C, Miazzi C, Franzolin E, et al. Regulation by degradation, a cellular defense against deoxyribonucleotide pool imbalances. Mutat Res - Genet Toxicol Environ Mutagen. 2010;703(1):2-10. doi:10.1016/j.mrgentox.2010.06.002.

10. Powell RD, Holland PJ, Hollis T, Perrino FW. Aicardi-Goutieres Syndrome Gene and

117

HIV-1 Restriction Factor SAMHD1 Is a dGTP-regulated Deoxynucleotide Triphosphohydrolase. J Biol Chem. 2011;286(51):43596-43600. doi:10.1074/jbc.C111.317628.

11. Goldstone DC, Ennis-Adeniran V, Hedden JJ, et al. HIV-1 restriction factor SAMHD1 is a deoxynucleoside triphosphate triphosphohydrolase. Nature. 2011;480(7377):379- 382. doi:10.1038/nature10623.

12. Ji X, Tang C, Zhao Q, Wang W, Xiong Y. Structural basis of cellular dNTP regulation by SAMHD1. Proc Natl Acad Sci. 2014;111(41):E4305-E4314. doi:10.1073/pnas.1412289111.

13. Zhu C-F, Wei W, Peng X, Dong Y-H, Gong Y, Yu X-F. The mechanism of substrate- controlled allosteric regulation of SAMHD1 activated by GTP. Acta Crystallogr Sect D Biol Crystallogr. 2015;71(3):516-524. doi:10.1107/S1399004714027527.

14. Kim B, Nguyen LA, Daddacha W, Hollenbaugh JA. Tight Interplay among SAMHD1 Protein Level, Cellular dNTP Levels, and HIV-1 Proviral DNA Synthesis Kinetics in Human Primary Monocyte-derived Macrophages. J Biol Chem. 2012;287(26):21570- 21574. doi:10.1074/jbc.C112.374843.

15. Hollenbaugh JA, Tao S, Lenzi GM, et al. dNTP pool modulation dynamics by SAMHD1 protein in monocyte-derived macrophages. Retrovirology. 2014;11(1):63. doi:10.1186/s12977-014-0063-2.

16. Rice GI, Bond J, Asipu A, et al. Mutations involved in Aicardi-Goutières syndrome implicate SAMHD1 as regulator of the innate immune response. Nat Genet. 2009;41(7):829-832. doi:10.1038/ng.373.

17. Wang J, Lu F, Shen X-Y, Wu Y, Zhao L. SAMHD1 is down regulated in lung cancer by methylation and inhibits tumor cell proliferation. Biochem Biophys Res Commun. 2014;455(3-4):229-233. doi:10.1016/j.bbrc.2014.10.153.

18. Rentoft M, Lindell K, Tran P, et al. Heterozygous colon cancer-associated mutations of SAMHD1 have functional significance. Proc Natl Acad Sci. 2016;113(17):4723-4728. doi:10.1073/pnas.1519128113.

19. Clifford R, Louis T, Robbe P, et al. SAMHD1 is mutated recurrently in chronic

118

lymphocytic leukemia and is involved in response to DNA damage. Blood. 2014;123(7):1021-1031. doi:10.1182/blood-2013-04-490847.

20. Schmidt S, Schenkova K, Adam T, et al. SAMHD1’s protein expression profile in humans. J Leukoc Biol. 2015;98(1):5-14. doi:10.1189/jlb.4HI0714-338RR.

21. Franzolin E, Pontarin G, Rampazzo C, et al. The deoxynucleotide triphosphohydrolase SAMHD1 is a major regulator of DNA precursor pools in mammalian cells. Proc Natl Acad Sci. 2013;110(35):14272-14277. doi:10.1073/pnas.1312033110.

22. Cribier A, Descours B, Valadão A, Laguette N, Benkirane M. Phosphorylation of SAMHD1 by Cyclin A2/CDK1 Regulates Its Restriction Activity toward HIV-1. Cell Rep. 2013;3(4):1036-1043. doi:10.1016/j.celrep.2013.03.017.

23. White TE, Brandariz-Nuñez A, Valle-Casuso JC, et al. The Retroviral Restriction Ability of SAMHD1, but Not Its Deoxynucleotide Triphosphohydrolase Activity, Is Regulated by Phosphorylation. Cell Host Microbe. 2013;13(4):441-451. doi:10.1016/j.chom.2013.03.005.

24. Bhattacharya A, Wang Z, White T, et al. Effects of T592 phosphomimetic mutations on tetramer stability and dNTPase activity of SAMHD1 can not explain the retroviral restriction defect. Sci Rep. 2016;6(April):31353. doi:10.1038/srep31353.

25. Tang C, Ji X, Wu L, Xiong Y. Impaired dNTPase activity of SAMHD1 by phosphomimetic mutation of Thr-592. J Biol Chem. 2015;290(44):26352-26359. doi:10.1074/jbc.M115.677435.

26. Wang Z, Bhattacharya A, Villacorta J, Diaz-Griffero F, Ivanov DN. Allosteric activation of SAMHD1 protein by deoxynucleotide triphosphate (dNTP)-dependent tetramerization requires dNTP concentrations that are similar to dNTP concentrations observed in cycling T cells. J Biol Chem. 2016;291(41):21407-21413. doi:10.1074/jbc.C116.751446.

27. Wittmann S, Behrendt R, Eissmann K, et al. Phosphorylation of murine SAMHD1 regulates its antiretroviral activity. Retrovirology. 2015;12(1):103. doi:10.1186/s12977-015-0229-6.

28. Mauney C, Rogers L, Harris R, et al. The SAMHD1 dNTP triphosphohydrolase is

119

controlled by a redox switch. Antioxid Redox Signal. 2017;(336):ars.2016.6888. doi:10.1089/ars.2016.6888.

29. Klomsiri C, Karplus PA, Poole LB. Cysteine-Based Redox Switches in Enzymes. Antioxid Redox Signal. 2011;14(6):1065-1077. doi:10.1089/ars.2010.3376.

30. Klomsiri C, Rogers LC, Soito L, et al. Endosomal H2O2 production leads to localized cysteine sulfenic acid formation on proteins during lysophosphatidic acid-mediated cell signaling. Free Radic Biol Med. 2014;71:49-60. doi:10.1016/j.freeradbiomed.2014.03.017.

31. Hao F, Tan M, Xu X, et al. Lysophosphatidic acid induces prostate cancer PC3 cell migration via activation of LPA1, p42 and p38α. Biochim Biophys Acta - Mol Cell Biol Lipids. 2007;1771(7):883-892. doi:10.1016/j.bbalip.2007.04.010.

32. Guo R, Kasbohm EA, Arora P, et al. Expression and Function of Lysophosphatidic Acid LPA 1 Receptor in Prostate Cancer Cells. Endocrinology. 2006;147(10):4883-4892. doi:10.1210/en.2005-1635.

33. Welbourn S, Dutta SM, Semmes OJ, Strebel K. Restriction of Virus Infection but Not Catalytic dNTPase Activity Is Regulated by Phosphorylation of SAMHD1. J Virol. 2013;87(21):11516-11524. doi:10.1128/JVI.01642-13.

34. Szatrowski TP, Nathan CF. Production of large amounts of hydrogen peroxide by human tumor cells. Cancer Res. 1991;51(3):794-798. http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?cmd=Retrieve&db=PubMed&dopt= Citation&list_uids=1846317%5Cnhttp://cancerres.aacrjournals.org/content/51/3/7 94.full.pdf.

35. Lennicke C, Rahn J, Lichtenfels R, Wessjohann L a, Seliger B. Hydrogen peroxide - production, fate and role in redox signaling of tumor cells. Cell Commun Signal. 2015;13:39. doi:10.1186/s12964-015-0118-6.

36. O’Brien V, Brown R. Signalling cell cycle arrest and cell death through the MMR System. Carcinogenesis. 2006;27(4):682-692. doi:10.1093/carcin/bgi298.

37. Qu J, Sun W, Zhong J, et al. Phosphoglycerate mutase 1 regulates dNTP pool and promotes homologous recombination repair in cancer cells. J Cell Biol.

120

2017;216(2):409-424.

38. Paull TT. Mechanisms of ATM activation. Annu Rev Biochem. 2015;84:711-738. doi:10.1146/annurev-biochem-060614-034335.

39. Meuth M. The molecular basis of mutations induced by deoxyribonucleoside triphosphate pool imbalances in mammalian cells. Exp Cell Res. 1989;181(2):305- 316. doi:10.1016/0014-4827(89)90090-6.

40. Klomsiri C, Nelson KJ, Bechtold E, et al. Use of Dimedone-Based Chemical Probes for Sulfenic Acid Detection. Thiol Redox Transitions Cell Signaling, Part A. 2010;473(10):77-94. doi:10.1016/S0076-6879(10)73003-2.

121

CHAPTER 3

Identification of inhibitors of the dNTP triphosphohydrolase SAMHD1 using a novel and direct high-throughput assay

ABSTRACT

The dNTP triphosphohydrolase SAMHD1 is a regulator of dNTP pools within the cell. Given its central role in nucleotide metabolism, SAMHD1 performs important functions in cellular homeostasis, cell cycle regulation, and innate immunity. It therefore represents a high profile target for small molecule drug design. SAMHD1 has a complex mechanism of catalytic activation that makes the design of an activating compound challenging. However, an inhibitor of SAMHD1 could serve multiple therapeutic roles including the potentiation of antiviral and anticancer drug regimens. The lack of high-throughput screens that directly measure SAMHD1 catalytic activity has impeded efforts to identify inhibitors of SAMHD1.

Here we describe a novel high-throughput screen that directly measures SAMHD1 catalytic activity. This assay results in a colorimetric endpoint that can be read spectrophotometrically, and utilizes bis-(4-nitrophenyl) phosphate as the substrate and

Mn2+ as the activating cation that facilitates catalysis. When used to screen a library of FDA approved drugs, this HTS identified multiple novel compounds that inhibited SAMHD1 dNTPase activity at micromolar concentrations.

122

INTRODUCTION

Deoxynucleotide triphosphates (dNTPs) are the building blocks of DNA. As such, the pathways controlling their synthesis and degradation are strictly regulated1,2.

Perturbations in dNTP homeostasis can interfere with proper DNA replication and repair which ultimately manifests as disease phenotype such as autoimmunity or cancer3,4. The enzyme SAMHD1 (Sterile Alpha Motif and Histidine and Aspartic Acid containing protein 1) is a deoxynucleotide triphosphohydrolase that belongs to a class of HD enzymes that perform divalent cation dependent phosphohydrolase activity. SAMHD1 catalyzes the hydrolysis of the triphosphate group of deoxynucleotide triphosphates (dNTPs) producing the cognate nucleoside and inorganic polyphosphate as reaction products5,6. Its catalytic activity is regulated through a complex mechanism of sequential nucleotide triphosphate binding in two regulatory sites on each SAMHD1 monomer. Binding of a guanosine nucleotide triphosphate at regulatory site 1 (RS1), and any of the four canonical dNTPs at regulatory site 2 (RS2) induces a conformational shift that enables SAMHD1 to form the catalytically active tetramer7–9. In this way, SAMHD1 activity is effectively regulated by its own substrate.

SAMHD1 is widely expressed in different cell types10. Its expression and activity profile varies by cell cycle stage, however, peaking during G0/G1-phase11. SAMDH1 catalytic activity is strictly regulated through post-translational modifications, including phosphorylation and oxidation12,13. Tight regulation of SAMHD1’s dNTPase activity enables cells to retain sufficient levels of dNTPs for DNA replication and repair without transitioning to potentially mutagenic levels14. Dysregulation of dNTP pools in any capacity,

123

either through an imbalance of the four canonical dNTPs, or an elevation or depletion of dNTPs, can have harmful consequences for the cell15–18. Accordingly, mutations to SAMHD1 are implicated in an autoimmune disorder related to aberrant nucleic acid metabolism,

Aicardi Goutiers Syndrome19. SAMDH1 mutations have also been identified in several different kinds of cancer20–23. In addition to mediated nucleotide homeostasis, SAMHD1 performs an important secondary function as an effector of innate immunity. In non-cycling myeloid and lymphoid cells, where it is expressed at high levels, it has been shown to act as a restriction factor against HIV-1 and other viral infections24–28. By depleting dNTP levels below a threshold required by viral DNA polymerases, SAMHD1 is able to prevent productive viral infection of hematopoietic cells29.

Given its central role in the processes of nucleotide metabolism and immunity, SAMHD1 represents an intriguing therapeutic target. Targeting nucleotide metabolism is a common therapeutic strategy for anticancer and antiviral therapeutic regimens30–32. Already,

SAMHD1 activity has been demonstrated to alter the efficacy of HIV reverse transcriptase inhibitors in cells and their ability to prevent productive HIV infection33–36. While upregulation of the activity of SAMHD1 may represent a gold standard for SAMHD1 drug discovery in relation to viral infection, the complex mechanism of activation makes this a challenging task. However, the rationale for the more straightforward design of a robust

SAMHD1 inhibitor may be just as strong (Fig. 1). Recently, it has been reported that several classes of nucleotide analogue therapeutics (NAT) may act as substrates for SAMHD137.

Specifically, the phosphorylated species of arabinose derived NATs, which lack the axial 2’- hydroxyl group on the sugar moiety, have been identified as targets for degradation by

SAMHD1 in vitro and in vivo. In patients with Acute Myeloid Leukemia (AML) the efficacy of the standard of care therapeutic agent Ara-C was inversely related to levels of SAMHD1

124

expression38. A small molecule inhibitor of SAMHD1 delivered as a complementary therapy may enhance the therapeutic window by potentiating the efficacy of NATs and result in improved patient outcomes39.

SAMHD1 might also be effective as a complementary therapy to DNA damaging agents.

Elevated or asymmetric dNTP pools are reported to enhance DNA damage and upregulate the DNA damage response40–42. In cells with a functioning DNA damage response, SAMHD1 inhibition and a subsequent increase in dNTPs might make cells more susceptible to DNA damaging agents and inhibit proliferation by arresting cells at the G2 checkpoint43.

Lastly, inhibition of SAMHD1 could have potential as an antiretroviral therapy.

Retroviruses like HIV establish latent infections even in the presence of antiretroviral therapy. It is proposed that by reducing the efficiency of replication in differentiated myeloid cells that have been infected, enzymes like SAMHD1 and TREX1 suppress a full immune response to chronic viral infection44–47. Inhibition of SAMHD1 may enhance productive viral infection and thereby initiate a robust immune response that unmasks the subclinical infection. This approach might also be useful as an adjuvant for live attenuated vaccines.

For these reasons, a small molecule that potently inhibits the catalytic activity of SAMHD1 would be a valuable tool for clinicians and researchers, alike. In this study we identify and validate a direct high-throughput screening (HTS) assay to detect inhibitors of SAMHD1 activity. While several HTS assays have been reported, they rely on enzyme-coupled protocols to detect the accumulation of inorganic phosphate48,49. Our assay is a direct colorimetric measure of SAMHD1 catalytic activity based on the ability of SAMHD1 to

125

Figure 1. Schematic depicting the potential therapeutic and research applications of a small molecule inhibitor of SAMHD1.

hydrolyze the phosphodiester model compound bis-(4-nitrophenyl) phosphate (b4NPP) in the presence of Mn2+. It represents a more direct and scalable approach for the rapid screening of large libraries of compounds as inhibitors of SAMHD1 triphosphohydrolase activity. Using this technique we identify and validate several compounds that inhibit

SAMHD1 from an FDA approved drug library. These novel inhibitors represent unique molecular scaffolds for further chemical optimization towards the design of a robust small molecule inhibition of SAMHD1.

126

RESULTS

SAMHD1 hydrolyzes b4NPP in the presence of Manganese and no activating nucleotide. Our initial objective was to develop a direct and quantitative assay of SAMDH1 catalytic activity that could be rapidly scaled up for high-throughput screening. This led us to investigate several chromogenic compounds that contain a phosphodiester or a phosphomonoester bond that structurally resembles those found in dNTPs; para- nitrophenyl thymidine monophosphate (pNP-TMP), bis-(4-nitrophenyl) phosphate

(b4NPP), and 4-nitrophenyl phosphate (4NPP) (Fig. 2A). These compounds are commonly used as surrogate substrates for enzymes with phosphatase, phosphohydrolase, or phosphodiesterase functions similar to SAMHD150–54. When hydrolyzed, these compounds produce yellow 4-nitrophenol as a reaction product. Its accumulation is stoichiometrically proportional to the amount of parent compound hydrolyzed and can be determined spectrophotometrically by measuring the absorbance at 410 nm.

We conducted several screens probing the catalytic activity of SAMHD1 towards the surrogate compounds under varying conditions (with and without activating nucleotides, varying enzyme and substrate concentrations, buffers, salts, pH levels, and divalent metal cations). While limited activity was observed under any conditions for 4NPP and pNP-TMP, a robust signal was generated when SAMHD1 was incubated with b4NPP in the presence of

Mn2+, and to a lesser extent Co2+ (Fig. 2B, Supplemental Fig. 1A). This effect held to be true even in the absence of activating nucleotides. This presented an unexpected result, as the canonical method for activating SAMHD1 is incubation with activating nucleotides to form the catalytically competent tetramer. The metal Mg2+ is also predominantly cited in

SAMHD1 literature as the cation coordinated by the active site histidine and aspartic acid

127

residues that facilitates substrate binding and triphosphohydrolase activity.

Seeking to further assess the Mn2+ dependent activation of SAMHD1, we conducted a series of dNTPase experiments in which SAMHD1 was activated using Mg2+, Mn2+, or Co2+. dATP was added to the reaction as the substrate with or without GTP. As expected, in the sample

Figure 2. Mn2+ activates SAMHD1 catalytic activity against dNTPs and phosphodiester containing compounds. (A) Chemical structures of compounds tested as substrates of SAMHD1. (B) SAMHD1 hydrolyzes b4NPP in the presence of Mn2+, and to a lesser extent Co2+, at a steady rate over the course of 30 minutes. (C) SAMHD1 is able to hydrolyze dATP when activated by Mn2+ in the absence of GTP (*p<0.01). (D) DLS experiment depicting the weight averaged distribution of SAMHD1 under different conditions. Mn2+ was used as the activating metal in all conditions. Mn2+ or b4NPP do not induce or inhibit tetramerization.

128

containing Mg2+, dATP was readily hydrolyzed when GTP was present, but not in its absence.

However, in samples containing Mn2+, SAMHD1 was able to hydrolyze dATP without GTP present in the reaction, albeit with a reduced efficiency (Fig. 2C). Dynamic light scattering, which measures the weight averaged particle size distribution as a function of hydrodynamic diameter, was utilized to examine the oligomeric state of SAMHD1 under the experimental conditions where Mn2+ was used as the activating cation. The data reveal that Mn2+ or b4NPP do not promote tetramerization on their own, but that SAMHD1 is capable of tetramerization in the presence of both Mn2+ and b4NPP (Fig. 2D). When considered together, then DLS data and Mn2+ induced hydrolysis of b4NPP and dATP, suggest that Mn2+ is capable of catalytically activating monomeric SAMHD1.

Validating the HTS as a specific measure of SAMHD1 catalytic activity. Having identified an assay that could be taken to a larger scale, our next step was to validate and optimize the specific conditions for screening. Kinetic analyses of b4NPP hydrolysis by

SAMHD1 under varying conditions were performed in order to identify the ideal reaction conditions. 4mM b4NPP represented the upper limit of these analyses as b4NPP exhibited diminished solubility above this concentration. As expected for a true enzyme-substrate relationship, changes in the enzyme concentration resulted in equivalent changes to the reaction velocity but did not significantly alter the KM (Fig. 3A, 3C). In conditions where either no SAMHD1, or a catalytically inactive mutant (H206A), was added to the reaction, b4NPP hydrolysis was not observed. These data provide additional evidence for SAMHD1 mediated b4NPP hydrolysis utilizing the active site chemistry.

Similar kinetic analyses were conducted in the presence of activating nucleotides in an effort to determine the effect of SAMHD1 tetramerization on b4NPP hydrolysis (Fig. 3B, 3C).

129

GTP and nonhydrolyzable dATPαS were chosen as activating nucleotides. Together they are able to induce SAMHD1 tetramerization, but are not substrates for the enzyme. Increasing concentrations of GTP and dATPαS were inversely correlated to reaction velocity. The precise mechanism of inhibition by activating nucleotides is unclear, although the decreased

Vmax at the highest concentration of activating nucleotides suggests that tetramerization, and the subsequent conformational shifts that occur, may decrease the binding affinity of b4NPP at the active site.

It is also possible that increasing levels of nucleotides are able to out-compete b4NPP for access to the active site. Indeed, increasing concentrations of dGTP completely inhibited

SAMHD1 mediated b4NPP hydrolysis in the low micromolar range (Fig. 3D). This implies that in the absence of the natural substrate of SAMHD1, b4NPP is able to bind in the active site. Computational docking software (Autodock Vina) was used to further investigate the binding of b4NPP to SAMHD1. Simulations place b4NPP in a similar orientation within the active site as the dGTP molecule found in the holoenzyme crystal structure (PDB 4BZC_A) and estimated a KD of 2.1μM (Fig. 3E). In this simulation, the b4NPP phosphate group is properly oriented for chemistry to occur, as it is aligned with the alpha-phosphate of dGTP - the bond normally hydrolytically cleaved by SAMHD1 catalysis. As a control, a dGTP binding in the active site of SAMHD1 was also simulated (Supplemental Fig. 2). The docking software modeled dGTP in an almost identical position as is found in the crystal structure, with an estimated KD of 0.14μM. This provides confidence in the model of b4NPP docking in the active site, as well as supports the model of dNTPs outcompeting b4NPP as the active site. Taken in sum, these kinetic and binding analyses suggest mixed inhibition of b4NPP by activating nucleotides, while strongly supporting SAMHD1 mediated hydrolysis of b4NPP at its active site.

130

Given that the tetrameric SAMHD1 holoenzyme hydrolyzed b4NPP less efficiently than the apoenzyme, it was important to determine that inhibition of the monomer would correspond to a similar effect on the biologically relevant tetrameric species. A single plate was selected from the FDA approved drug library and used to test the ability of SAMHD1 to hydrolyze b4NPP in the presence of potential inhibitors with and without 25/100μM

GTP/dATPαS in the reaction. While the reaction with activating nucleotides proceeded at a slower rate, a regression analysis of the percent SAMHD1 activity normalized to the internal

Figure 3: The b4NPP-SAMHD1 assay specifically measures SAMHD1 catalytic activity. (A) Kinetic analysis of varying concentrations of SAMHD1 on b4NPP hydrolysis. H206A is a catalytically inactive mutant. (B) Kinetic analysis of SAMHD1 mediated b4NPP hydrolysis in the presence of varying concentrations of activating nucleotides, and the (C) calculated kinetic parameters for both sets of experiments. (D) dGTP is able to inhibit SAMHD1 b4NPP hydrolysis at low micromolar concentrations. SAMHD1 activity was normalized to the condition with no dGTP present. (E) Computational docking simulation of b4NPP (green) in the SAMHD1 active site aligns it with dGTP (translucent blue) from the crystal structure (PDB: 4BZC_A). (F) A regression analysis of SAMHD1 %Activity in the presence or absence of activating nucleotides, following incubation with experimental compounds with varying capacities to inhibit SAMHD1. A strong correlation is observed between the two sets of conditions (R2 = 0.818) with minimal false negatives (red) or positives (green).

131

positive control (SAMHD1 + DMSO only) revealed a strong correlation between the inhibition of SAMHD1 activity by a specific compound under both sets of conditions (R2 =

0.818) (Fig. 3F). This demonstrated that inhibition of the Mn2+ activated monomer represents a good model for inhibition of tetrameric SAMHD1, with relatively few false positive or false negative readings identified within these experimental conditions.

HTS of FDA approved drugs identified multiple potential inhibitors of SAMHD1. The

HTS assay validation and optimization process allowed us to identify the optimal conditions for screening a larger library of compounds. This screen was performed using 500nM

SAMHD1 activated with 5mM MnCl2 and in the absence of any activating nucleotides. 2mM b4NPP was used as a final substrate concentration. These conditions generated a robust and easily detectable signal, but were well within the linear range of the assay

(Supplemental Fig. 3).

The optimized conditions were used to assay a library of 640 FDA approved drugs for the ability to inhibit SAMHD1 mediated b4NPP hydrolysis. While this library encompassed only a small number of compounds relative to larger HTS libraries, all compounds were FDA approved chemical entities with broad structural diversity. 2μl of each compound (2 mg/ml in DMSO) was added to the enzyme solution 10 minutes prior to initiating with b4NPP. As expected, the majority of compounds exhibited limited or no effect on SAMHD1 catalytic activity relative to the average of the positive control samples (SAMHD1 + DMSO only) (Fig.

4A). However, approximately 60 compounds inhibited SAMHD1 catalysis of b4NPP by at least 4 standard deviations (SAMHD1 + DMSO only samples fraction activity normalized to

1.0, σ = 0.12) which equated to roughly 50 percent inhibition (Fig. 4B).

132

The prevalence of false positives in screens of this nature led us to investigate the top 60 inhibitors from the HTS using a secondary dNTPase assay. These reactions were prepared in a similar manner, with the exception that MgCl2 was used instead of MnCl2 as the activating cation. dGTP was added to a final concentration of 500μM to initiate the reaction, and served as both activator and substrate. The formation of dG was measured to quantitate reaction progress. This assay represents the established method for investigating SAMHD1 triphosphohydrolase catalytic activity. A regression analysis of the effect of each putative inhibitor on SAMHD1 activity in the dNTPase and HTS assays revealed a strong correlation (R2 = 0.627) when normalized by the molar concentration of the compound in the reaction (Fig. 4C). These results further supported the utility of the

HTS and identify multiple compounds for further analysis of SAMHD1 inhibition.

We also used a modified version of the dNTPase assay to test 5 compounds that appeared to enhance SAMHD1 activity in the HTS. SAMHD1 and either 100μM or 1mM of the experimental compound were combined in solution, and then 500μM of dATP was added to initiate the reaction. No SAMHD1 catalytic activity was observed using any of these compounds (data not shown), and we therefore concluded that the observed absorbance increase in the HTS for these represented an artifact effect of the experimental conditions.

Validation of top hits reveal compounds with micromolar inhibition of SAMHD1 dNTPase activity. In order to determine true inhibitors of SAMHD1 and quantify their ability to restrict enzyme catalysis a more rigorous dose response assay was performed. A persistent challenge encountered with high-throughput screens is the identification of non- specific inhibitors that inactivate enzymes by aggregating into colloidal particles within the

133

Fig 4. HTS results for inhibitors of SAMHD1 from the FDA Approved Drug Library. (A) Results from the HTS of FDA approved compounds. The red line is the mean of the positive control samples (SAMHD1 + DMSO only) from all plates and represents full catalytic activity (Fraction activity = 1.0). The blue line is the mean of the no enzyme control. Dashed lines represent +/- 4 standard deviations of the positive control (σ=0.12). (B) HTS results depicted as a histogram and highlighting compounds selected as top inhibitors for further dNTPase screening. (C) A regression analysis comparing SAMHD1 b4NPP hydrolysis and dNTPase activity when incubated with the top 60 inhibitors identified by the HTS. Percent activity was normalized to the molar concentration of each compound in the reaction. The strong correlation (R2 = 0.627) observed between the two assays supports the predictive capacity of the HTS.

134

Fig 5. Top hits inhibit SAMHD1 dNTPase activity at micromolar concentrations. (Top) Dose response curves of SAMHD1 activity the presence of increasing concentrations of inhibitor. (Bottom) Lowest free energy pose of each compound in the computationally docked in the active site and estimated ΔG and KD values.

reaction55. This effect was mitigated in the SAMHD1 dose response assay by the addition of a small amount of non-ionic detergent (0.025% Tween20) to the reaction buffer.

Additionally, 10x inhibitor stocks were made by serial dilution in 50% DMSO to enhance compound solubility. The final reaction concentration of 6% DMSO did not alter SAMHD1 activity. The top 20 hits identified by the HTS and dNTPase assays, were tested for their ability to inhibit SAMHD1 catalysis across concentrations ranging from 0.125μM to 128μM

(Supplemental Table. 1). 150μM dGTP was used to initiate the reactions for its dual ability to activate and serve as substrate. This concentration of dGTP is in close proximity to the

SAMHD1 KM but still above the threshold at which our analysis protocol enabled us to accurately measure the initial velocity of the reaction.

Under these conditions the inhibitory capacity of several experimental compounds was either severely diminished or abolished completely (Supplemental Fig. 4). Additionally,

135

several other top hits exhibited very shallow or steep dose response curves (Hill Coefficient

< 0.5 or > 2), or non-zero baselines, all of which may be indicative of non-specific enzyme inhibition56,57. Further still, several compounds were filtered out based on a chemoinformatic screen that identifies indiscriminate inhibitors referred to as pan-assay interference compounds (PAINS)58. Five compounds - Lomofungin, Ethacrynic Acid,

Ceftazidime, Troglitazone, and Delavirdine - inhibited SAMHD1 dNTPase activity with IC50s less than 50μM (Fig. 5A). Docking simulations were performed for each compound in order to estimate how they may bind in the active site and a corresponding ΔG was calculated computationally (Fig. 5B). These compounds represent previously unidentified small molecule inhibitors of SAMHD1 activity in vitro. Additionally, their discovery validates the HTS approach we have developed, as well as presents multiple unique chemical scaffolds for further study and optimization.

DISCUSSION

In this study we report the development and validation of a direct and simple HTS assay to measure SAMHD1 catalytic activity using the compound b4NPP and Mn2+ as the activating divalent cation. b4NPP contains a phosphodiester bond that mimics the natural substrate of SAMHD1, and when cleaved results in a colorimetric transformation that can be read spectrophotometrically. b4NPP has been extensively reported as a substrate for phosphodiesterases, nucleases, and HD domain containing enzymes similar to SAMHD1 when activated with Mn2+. Perhaps one of the more intriguing discoveries presented in this study is the ability of SAMHD1 to be catalytically active as a monomer when Mn2+ is the activating cation. While this finding was an important factor in optimizing our HTS, it also

136

generates myriad questions about the nature of the true metal cofactor utilized by SAMHD1 in vivo. If SAMHD1 tetramerization is not required for dNTP hydrolysis there are major implications for our understanding of SAMHD1’s role in nucleotide homeostasis and innate immunity.

Following the validation of our HTS, we subsequently screened a small library of FDA approved compounds for inhibition of SAMHD1 activity. From this screen, we were able to identify 60 compounds that inhibited SAMHD1 activity by at least 50%. A secondary screen of the top inhibitors utilizing the innate dNTPase activity of SAMHD1 allowed us to further refine the list of potential inhibitors.

The presence of indiscriminate inhibitors in high-throughput libraries is a major concern for drug discovery efforts. These broad spectrum inhibitors (PAINS) can interfere with enzyme activity through a variety of nonspecific mechanisms such as aggregation, redox cycling, metal chelation, and covalent modification59–61. Therefore we performed a series of dose response assays on the top 20 inhibitors from the secondary dNTPase assay in the presence of a small amount of non-ionic detergent that can mitigate the effect of compound aggregation. Compounds were also subjected to a chemoinformatic filter to remove known

PAINS or identify potentially promiscuous pharmacophores. Given that competitive inhibition at a single active site invariably results in a Hill Coefficient of 1, compounds exhibiting shallow (HC<0.5), steep (HC>2), or non-zero baseline, dose response curves were also excluded from the final list of top inhibitors. It is worth mentioning however, that the complex mechanism of SAMHD1 activation, in which two activating nucleotides must bind at distinct regulatory sites before catalysis is possible, does not enable us to immediately preclude compounds with Hill coefficients that are not approximately 1 from being

137

legitimate SAMHD1 inhibitors. Compounds may be interacting with either or both of the regulatory sites, or there may be a cooperative effect to inhibiting a single active site on a

SAMHD1 tetramer. Any of these effects would make the interpretation of a Hill coefficient difficult.

Five compounds satisfied the above criteria and demonstrated SAMHD1 inhibition with

IC50s less than 50μM - Lomofungin, Ethacrynic Acid, Ceftazidime, Troglitazone, and

Delavirdine. In general, these compounds contain heterocyclic aromatic ring structures that resemble the nitrogenous bases of purine and pyrimidines. As demonstrated in the docking simulations, additional functional groups and pharmacophores likely stabilize the active site interaction through an enhanced network of weak interactions. Interestingly, Lomofungin and Delavirdine have been reported to interfere in various aspects of nucleic acid metabolism62,63. An analogue to Ceftazidime, a β-lactam antibiotic, was identified as a weak inhibitor of SAMHD1 in a previously describe HTS utilizing an alternative methodology49.

These five compounds represent previously unidentified inhibitors of SAMHD1. They are structurally distinct as determined by their 2D molecular fingerprints. No pair of compounds has a Tanimoto coefficient - a measure of similarity that ranges from 0 (no similarity) to 1 (identical) - greater than 0.61 (Ceftazidime and Delavirdine). The discovery of these compounds as inhibitors of SAMHD1 is an encouraging step towards the realization of a robust small molecule inhibitor of SAMHD1, which will have important therapeutic and research applications. However, much work remains to be done, both in terms of validating these top hits, as well as screening larger chemical libraries. The HTS described in this study will serve as a useful tool for additional screening of large libraries and selecting lead compounds for chemical optimization as potent inhibitors of SAMHD1.

138

MATERIALS AND METHODS

Reagents. Unless otherwise stated, all reagents were purchased from Sigma. dGTP was purchased from Promega (100mM solution). 2'-Deoxyadenosine-5'-O-(1-Thiotriphosphate)

(dATPαS) was from TriLink Biotechnologies. The HTS library (Screen-Well FDA Approved

Drug Library Version 1.5, Enzo Life Sciences, Inc.) was a generous gift from Dr. Leslie Poole.

SAMHD1 expression and purification. Recombinant SAMHD1 WT and H206A was overexpressed and purified as described12. Briefly, the full-length human SAMHD1 gene was cloned into a modified pET28 expression vector (pLM303-SAMHD1) that contained an

N-terminal MBP tag and an intervening rhinovirus 3C protease cleavage site. Expression constructs were transformed into E. coli BL21* and grown in Luria-Bertani medium at 37 °C with shaking to an OD600 = 0.6. Cultures were induced with 0.3mM isopropyl-β-D-thio- galactoside (IPTG), rapidly cooled on ice to 20 °C, and then allowed to express for 16-18 hours at 16 °C. Harvested cells were resuspended in Amylose Column Buffer (ACB) (50mM

Tris-HCl pH 7.5, 200mM NaCl, 5mM MgCl2, 0.1mM EDTA, 2mM DTT, and 5% glycerol) and lysed using an Avestin Emulsiflex-C5 cell homogenizer. Cell debris was cleared by centrifuging at 18.5K r.p.m. for 30 minutes at 4 °C, and the cleared lysate was passed over amylose high flow resin (New England Biolabs), and washed with 3 column volumes of

ACB+1M NaCl to remove residual nucleic acid. Bound MBP-SAMHD1 was eluted over a linear gradient with ACB+20mM Maltose. The desired fractions were pooled and dialyzed overnight against Heparin Column Buffer (HCB) (50mM Tris-HCl pH 7.5, 100mM NaCl, 5mM

MgCl2, 0.1mM EDTA, 2mM DTT, and 5% glycerol). PreScission Protease (GE Biosciences) was added (100:1) to the MBP-SAMHD1 dialysate to cleave the MBP tag. The SAMHD1 containing sample was separated from the cleaved MBP using a 5ml Heparin HiTrap (GE

139

Healthcare Life Sciences) and eluted using a linear gradient of HCB+2M NaCl. SAMHD1 containing fractions were again pooled and further purified using Superdex 200 size exclusion chromatography column equilibrated with 50mM Tris-HCl pH 7.5, 250mM NaCl,

5mM MgCl2, 0.1mM EDTA, 2mM DTT, and 5% glycerol. The protein was concentrated and flash frozen in individual aliquots using liquid nitrogen, and stored at -80 °C until use.

SAMHD1 deoxynucleotide triphosphohydrolase (dNTPase) activity assay. Individual protein aliquots were thawed and mixed with reaction buffer (final concentration 50mM Tris pH

7.5, 100mM NaCl, 5mM activating cation (MgCl2, MnCl2, or CoCl2), 0.1mM EDTA). Aliquots were then transferred to a clear flat-bottom 96 well polystyrene microassay plate.

Reactions were initiated by adding either 500μM dATP and 50μM GTP, or 500μM dGTP alone, to a final reaction volume of 100μl. Formation of the deoxynucleoside was used to measure the progress of the reaction. SAMHD1 concentration in all reactions was 500nM.

Reactions were quenched using EDTA to a final concentration of 20mM after 10 or 20 minutes, depending on the experiment. In experiments investigating the inhibitory effects of top hits from the HTS, MgCl2 was used as the activating cation, and 2μl of each compound was added to the enzyme mixture 10 minutes prior to the addition of dGTP.

dNTPase reaction products were analyzed using ion pair reverse phase chromatography on a Waters HPLC system. A CAPCell PAK C18 column (Shiseido Fine Chemicals) was equilibrated with 20mM NaH2PO4 pH 7.0, 5mM tetra n-butylammonium phosphate, and 5% methanol. Reactants and products were eluted with a linear gradient of methanol from 5-

50%. A254 was used to measure eluent peaks, and quantification was performed using the

Empower Software to integrate the area under each reaction component peak.

140

SAMHD1 bis-(4-nitrophenyl) phosphate (b4NPP) assay. Frozen aliquots of SAMHD1 were thawed and diluted into HTS reaction buffer (HTS-RB) (final concentration 50mM Tris pH7.5, 100mM NaCl, 5mM activating cation, 2% glycerol, 0.5% DMSO, 0.5mM TCEP) and transferred to a clear flat- bottom 96 well polystyrene microassay plate. If present in the reaction, activating nucleotides were added at the indicated concentrations. 64mM bis(4- nitrophenyl) phosphate stocks were made from powder in ultra pure water and stored at -

20C until use. The reaction was initiated by adding b4NPP that had been further diluted in water to a final concentration of 2mM and a final volume of 100μl. The hydrolysis of b4NPP to p-nitrophenol and p-nitrophenyl phosphate was assessed over a period of 45 minutes by measuring the increase in absorbance at 410nm on a Tecan Infinite M1000Pro Plate Reader.

Reactions with no enzyme were used as baseline controls. Reaction slopes were calculated from the first 20 minutes of the reaction to ensure accurate determination of initial reaction velocities.

Dynamic Light Scattering. SAMHD1 particle size distributions as a function of hydrodynamic diameter (DH) were determined as describe previously12 with several minor changes.

Briefly, SAMHD1 was transferred to DLS buffer (50mM NaH2PO4 pH 7.5, 150mM NaCl, 5mM

MnCl2) using a microspin column packed with BioGel-P6 resin (BioRad). 200μl samples were prepared by incubating SAMHD1 with the indicated concentrations of b4NPP (2mM) and activating nucleotides (25/100μM GTP/dATPαS). SAMHD1 final concentration in the samples was 0.2 mg/ml. SAMHD1 solutions were analyzed using a Malvern Nano-S

Zetasizer (Malvern Instruments, Worcestershire, UK). Raw intensity weighted data was converted to volume % distribution in order to mitigate the impact of small amounts of aggregated protein.

141

High-throughput screen. High-throughput screens were conducted similarly to the b4NPP assay. SAMHD1 was diluted into HTS-RB to a final reaction concentration of 500nM and transferred to a clear flat-bottom 96 well polystyrene microassay plate. 2μl of each compound (2mg/ml in DMSO) from the Screen-Well FDA Approved Drug Library Version

1.5 was added to the SAMHD1 solution and allowed to sit at 21oC for 10 minutes. The reaction was initiated by the addition of b4NPP to a final concentration of 2mM at a final reaction volume of 100μl. All reactions were prepared using a Gryphon high-throughput liquid handling platform (Art Robbins Instruments).

The ability of SAMHD1 to hydrolyze b4NPP to p-nitrophenol and p-nitrophenyl phosphate in the presence of the inhibitor was assessed over a period of 45 minutes by measuring the increase in absorbance at 410nm on a Tecan Infinite M1000Pro Plate Reader. A matched reaction for each compound in which no SAMHD1 was added served as a baseline control for nonspecific compound-substrate interactions and was subtracted from the absorbance reading for the enzyme containing reaction. In each plate, wells containing SAMHD1 and

DMSO only served as positive controls to which experimental wells were normalized.

Reaction velocities were calculated from the slope of the absorbance increase over the first

20 minutes of the reaction, which was within the linear range of the assay. Percent activity was calculated from the reaction slopes as follows:

(푆퐴푀퐻퐷1&푐표푚푝표푢푛푑) − (푐표푚푝표푢푛푑 표푛푙푦) %퐴푐푡𝑖푣𝑖푡푦 = (푆퐴푀퐻퐷1&퐷푀푆푂) − (퐷푀푆푂 표푛푙푦)

Dose Response Curves and IC50 determination. Dose response reactions were prepared in a similar manner as dNTPase reactions described above, with several notable changes.

SAMHD1 was diluted into reaction buffer (final concentration 50mM Tris pH 7.5, 100mM

142

NaCl, 5mM MgCl2, 0.1mM EDTA, 0.1mM TCEP, 1% DMSO, 2% glycerol, 0.025%Tween20) before transferring to a clear flat-bottom 96 well polystyrene microassay plate. 10x stocks of inhibitory compounds were prepared by 2:1 serial dilution in 50% DMSO 50% dH2O.

Compounds were added to the SAMHD1 solution and allowed to incubate for 10 minutes at

21oC before initiating the reaction by the addition of dGTP. A final concentration of 150μM dGTP was selected as a concentration similar to the reported SAMHD1 KM value that results in reproducible peaks using our HPLC analysis method. Final SAMHD1 concentration in the reaction was 250nM. Reactions were quenched after 5 minutes by the addition of EDTA to a final concentration of 20mM in order to ensure that no more than 20% of dGTP was converted to dG. Reaction contents were analyzed using the HPLC method described above.

Data was fit to four parameter non-linear regression with variable Hill slopes in Graphpad

Prism 7.00 using the equation:

(푀푎푥 − 푀𝑖푛) %퐴푐푡𝑖푣𝑖푡푦 = 푀𝑖푛 + 1 + 10(퐿표푔퐸퐶50−푋)∗퐻푖푙푙 푆푙표푝푒

In silico docking simulations. Docking simulations were performed using the Autodock

Vina64 plugin for Chimera molecular visualization software. The SAMHD1 crystal structure

PDB 4BZC_A was used as the docking receptor and ligands were docked into the active site

(Center X 15.75, Y 27.5, Z 42.5; Size X 22.5, Y 21.5, Z 18.5).

Data and Statistical Analysis: Charts and graphs were generated using SigmaPlot v13.0

(Systat Software, Inc.) or GraphPad Prism v7.00 for Windows (GraphPad Software, La Jolla

California USA, www.graphpad.com), which were also used for statistical analysis, t-test comparisons, and linear and nonlinear regression. Statistical significance was determined at p<0.01 (*). Molecular representations were generated using the program Chimera65.

143

Author Contributions

Conceptualization, Methodology - C.H.M., T.H.; Investigation- C.H.M; Writing-original draft-

C.H.M; Writing-review & editing- C.H.M, T.H., Supervision- T.H, Funding Acquisition- T.H,

F.W.P.

Acknowledgments

This work has been supported through funding from the NIH (RO1 GM108827 to TH, R01

GM06692 to FWP, and T32 GM095440 support for CHM). We would like to thank Dr. Fred

Salsbury for his assistance with computational simulations.

Abbreviations

B4NPP - bis-(4-nitrophenyl) phosphate

HTS – high-throughput screen

Ara-C – cytabarine

NAT – Nucleotide analogue therapeutic

DH - hydrodynamic diameter dATPαS - deoxyadenosine triphosphate alpha S dNTP - deoxynucleotide triphosphate

DTT - dithiothreitol

EDTA - ethylenediaminetetraacetic acid

144

GTP - guanosine triphosphate

H2O2 - hydrogen peroxide

IPTG - isopropyl-β-D-thio-galactoside

TCEP - Tris(2-carboxyethyl)phosphine

PAINS – Pan-assay interference compounds

145

SUPPLEMENTAL INFORMATION

Supplemental Figure 1. Screening of 4-nitrophenol containing compounds as SAMHD1 substrates. (A-D) Experiments investigate the catalytic capacity of SAMHD1 to hydrolyze phosphodiester or phosphomonoester containing 4-nitrophenol compounds in the presence of different activating metals (Mg2+ or Mn2+) or activating nucleotides (25/100μM GTP/dATPαS). Only Mn2+ activated hydrolysis of b4NPP resulted in a robust signal that could be scaled up for high- throughput screening.

146

Supplemental Figure 2. dGTP computationally docked in the active site using AutoDock Vina.

147

Supplemental Figure 3. The SAMHD1-b4NPP assay is linear under the specified conditions for the duration of the experiment. Absorbance data from representative SAMHD1-b4NPP experiments utilizing varying conditions. (A) 2mM b4NPP and 0, 250nM, or 500nM of SAMHD1. H206A is a catalytically inactive mutant of SAMHD1. (B) Changes in the concentration of b4NPP. (C) SAMHD1 b4NPP hydrolysis in the presence of activating nucleotides. (D) SAMHD1 hydrolysis of b4NPP compared to pNP-TMP hydrolysis by TREX1 and a generic Phosphodiesterase. Trex1 and the Phosphodiesterase can hydrolyze pNP-TMP but not b4NPP, indicating that the effect is specific to SAMHD1.

148

Supplemental Table 1. Table of Top20 HTS results

149

150

151

Supplemental Figure 4. Dose response curves of top20 inhibitors of SAMHD1 dNTPase activity.

152

153

154

REFERENCES

1. Rampazzo C, Miazzi C, Franzolin E, et al. Regulation by degradation, a cellular defense against deoxyribonucleotide pool imbalances. Mutat Res - Genet Toxicol Environ Mutagen. 2010;703(1):2-10. doi:10.1016/j.mrgentox.2010.06.002. 2. Mathews CK. Deoxyribonucleotides as genetic and metabolic regulators. FASEB J. 2014;28(9):3832-3840. doi:10.1096/fj.14-251249. 3. Mathews CK. Deoxyribonucleotide metabolism, mutagenesis and cancer. Nat Rev Cancer. 2015;15(9):528-539. doi:10.1038/nrc3981. 4. Pai C, Kearsey S. A Critical Balance: dNTPs and the Maintenance of Genome Stability. Genes (Basel). 2017;8(2):57. doi:10.3390/genes8020057. 5. Powell RD, Holland PJ, Hollis T, Perrino FW. Aicardi-Goutieres Syndrome Gene and HIV-1 Restriction Factor SAMHD1 Is a dGTP-regulated Deoxynucleotide Triphosphohydrolase. J Biol Chem. 2011;286(51):43596-43600. doi:10.1074/jbc.C111.317628. 6. Goldstone DC, Ennis-Adeniran V, Hedden JJ, et al. HIV-1 restriction factor SAMHD1 is a deoxynucleoside triphosphate triphosphohydrolase. Nature. 2011;480(7377):379- 382. doi:10.1038/nature10623. 7. Zhu C-F, Wei W, Peng X, Dong Y-H, Gong Y, Yu X-F. The mechanism of substrate- controlled allosteric regulation of SAMHD1 activated by GTP. Acta Crystallogr Sect D Biol Crystallogr. 2015;71(3):516-524. doi:10.1107/S1399004714027527. 8. Ji X, Tang C, Zhao Q, Wang W, Xiong Y. Structural basis of cellular dNTP regulation by SAMHD1. Proc Natl Acad Sci. 2014;111(41):E4305-E4314. doi:10.1073/pnas.1412289111. 9. Hansen EC, Seamon KJ, Cravens SL, Stivers JT. GTP activator and dNTP substrates of HIV-1 restriction factor SAMHD1 generate a long-lived activated state. Proc Natl Acad Sci. 2014;111(18):E1843-E1851. doi:10.1073/pnas.1401706111. 10. Schmidt S, Schenkova K, Adam T, et al. SAMHD1’s protein expression profile in humans. J Leukoc Biol. 2015;98(1):5-14. doi:10.1189/jlb.4HI0714-338RR. 11. Franzolin E, Pontarin G, Rampazzo C, et al. The deoxynucleotide triphosphohydrolase SAMHD1 is a major regulator of DNA precursor pools in mammalian cells. Proc Natl Acad Sci. 2013;110(35):14272-14277. doi:10.1073/pnas.1312033110. 12. Mauney C, Rogers L, Harris R, et al. The SAMHD1 dNTP triphosphohydrolase is controlled by a redox switch. Antioxid Redox Signal. 2017;(336):ars.2016.6888. doi:10.1089/ars.2016.6888. 13. Cribier A, Descours B, Valadão A, Laguette N, Benkirane M. Phosphorylation of SAMHD1 by Cyclin A2/CDK1 Regulates Its Restriction Activity toward HIV-1. Cell Rep. 2013;3(4):1036-1043. doi:10.1016/j.celrep.2013.03.017. 14. Kretschmer S, Wolf C, Konig N, et al. SAMHD1 prevents autoimmunity by maintaining genome stability. Ann Rheum Dis. 2015;74(3):e17-e17. doi:10.1136/annrheumdis- 2013-204845. 15. Watt DL, Buckland RJ, Lujan SA, Kunkel TA, Chabes A. Genome-wide analysis of the

155

specificity and mechanisms of replication infidelity driven by imbalanced dNTP pools. Nucleic Acids Res. 2015;44(4):1669-1680. doi:10.1093/nar/gkv1298. 16. Pajalunga D, Franzolin E, Stevanoni M, et al. A defective dNTP pool hinders DNA replication in cell cycle-reactivated terminally differentiated muscle cells. Cell Death Differ. 2017:1-11. doi:10.1038/cdd.2017.4. 17. Kunz BA. Mutagenesis and deoxyribonucleotide pool imbalance. Mutat Res - Fundam Mol Mech Mutagen. 1988;200(1-2):133-147. doi:10.1016/0027-5107(88)90076-0. 18. Meuth M. The molecular basis of mutations induced by deoxyribonucleoside triphosphate pool imbalances in mammalian cells. Exp Cell Res. 1989;181(2):305- 316. doi:10.1016/0014-4827(89)90090-6. 19. Rice GI, Bond J, Asipu A, et al. Mutations involved in Aicardi-Goutières syndrome implicate SAMHD1 as regulator of the innate immune response. Nat Genet. 2009;41(7):829-832. doi:10.1038/ng.373. 20. Clifford R, Louis T, Robbe P, et al. SAMHD1 is mutated recurrently in chronic lymphocytic leukemia and is involved in response to DNA damage. Blood. 2014;123(7):1021-1031. doi:10.1182/blood-2013-04-490847. 21. Rentoft M, Lindell K, Tran P, et al. Heterozygous colon cancer-associated mutations of SAMHD1 have functional significance. Proc Natl Acad Sci. 2016;113(17):4723-4728. doi:10.1073/pnas.1519128113. 22. Kohnken R, Kodigepalli KM, Wu L. Regulation of deoxynucleotide metabolism in cancer: novel mechanisms and therapeutic implications. Mol Cancer. 2015;14:176. doi:10.1186/s12943-015-0446-6. 23. Wang J, Lu F, Shen X-Y, Wu Y, Zhao L. SAMHD1 is down regulated in lung cancer by methylation and inhibits tumor cell proliferation. Biochem Biophys Res Commun. 2014;455(3-4):229-233. doi:10.1016/j.bbrc.2014.10.153. 24. Lahouassa H, Daddacha W, Hofmann H, et al. SAMHD1 restricts the replication of human immunodeficiency virus type 1 by depleting the intracellular pool of deoxynucleoside triphosphates. Nat Immunol. 2012;13(3):223-228. doi:10.1038/ni.2236. 25. Laguette N, Sobhian B, Casartelli N, et al. SAMHD1 is the dendritic- and myeloid-cell- specific HIV-1 restriction factor counteracted by Vpx. Nature. 2011;474(7353):654- 657. doi:10.1038/nature10117. 26. Hrecka K, Hao C, Gierszewska M, et al. Vpx relieves inhibition of HIV-1 infection of macrophages mediated by the SAMHD1 protein. Nature. 2011;474(7353):658-661. doi:10.1038/nature10195. 27. Gramberg T, Kahle T, Bloch N, et al. Restriction of diverse retroviruses by SAMHD1. Retrovirology. 2013;10(1):26. doi:10.1186/1742-4690-10-26. 28. Hollenbaugh JA, Gee P, Baker J, et al. Host Factor SAMHD1 Restricts DNA Viruses in Non-Dividing Myeloid Cells. Ross SR, ed. PLoS Pathog. 2013;9(6):e1003481. doi:10.1371/journal.ppat.1003481. 29. Kim B, Nguyen LA, Daddacha W, Hollenbaugh JA. Tight Interplay among SAMHD1 Protein Level, Cellular dNTP Levels, and HIV-1 Proviral DNA Synthesis Kinetics in Human Primary Monocyte-derived Macrophages. J Biol Chem. 2012;287(26):21570- 21574. doi:10.1074/jbc.C112.374843.

156

30. Aye Y, Li M, Long MJC, Weiss RS. Ribonucleotide reductase and cancer: biological mechanisms and targeted therapies. Oncogene. 2014;(April 2014):2011-2021. doi:10.1038/onc.2014.155. 31. Parker WB. Enzymology of Purine and Pyrimidine Antimetabolites Used in the Treatment of Cancer. Chem Rev. 2009;109(7):2880-2893. doi:10.1021/cr900028p. 32. Rampazzo C, Tozzi MG, Dumontet C, Jordheim LP. The druggability of intracellular nucleotide-degrading enzymes. Cancer Chemother Pharmacol. 2016;77(5):883-893. doi:10.1007/s00280-015-2921-6. 33. Amie SM, Daly MB, Noble E, Schinazi RF, Bambara RA, Kim B. Anti-HIV host factor SAMHD1 regulates viral sensitivity to nucleoside reverse transcriptase inhibitors via modulation of cellular deoxyribonucleoside triphosphate (dNTP) levels. J Biol Chem. 2013;288(28):20683-20691. doi:10.1074/jbc.M113.472159. 34. Hollenbaugh JA, Schader SM, Schinazi RF, Kim B. Differential regulatory activities of viral protein X for anti-viral efficacy of nucleos(t)ide reverse transcriptase inhibitors in monocyte-derived macrophages and activated CD4+ T cells. Virology. 2015;485:313-321. doi:10.1016/j.virol.2015.08.006. 35. Ballana E, Badia R, Terradas G, et al. SAMHD1 specifically affects the antiviral potency of thymidine analog HIV reverse transcriptase inhibitors. Antimicrob Agents Chemother. 2014;58(8):4804-4813. doi:10.1128/AAC.03145-14. 36. Ordonez P, Kunzelmann S, Groom HCT, et al. SAMHD1 enhances nucleoside-analogue efficacy against HIV-1 in myeloid cells. Sci Rep. 2017;7(February):42824. doi:10.1038/srep42824. 37. Hollenbaugh JA, Shelton J, Tao S, et al. Substrates and Inhibitors of SAMHD1. PLoS One. 2017;12(1):e0169052. doi:10.1371/journal.pone.0169052. 38. Schneider C, Oellerich T, Baldauf H-M, et al. SAMHD1 is a biomarker for cytarabine response and a therapeutic target in acute myeloid leukemia. Nat Med. 2016;(i). doi:10.1038/nm.4255. 39. Herold N, Rudd SG, Ljungblad L, et al. Targeting SAMHD1 with the Vpx protein to improve cytarabine therapy for hematological malignancies. Nat Med. 2017;23(2):256-263. doi:10.1038/nm.4265. 40. Bester AC, Roniger M, Oren YS, et al. Nucleotide deficiency promotes genomic instability in early stages of cancer development. Cell. 2011;145(3):435-446. doi:10.1016/j.cell.2011.03.044. 41. Jones RM, Petermann E. Replication fork dynamics and the DNA damage response. Biochem J. 2012;443(1):13-26. doi:10.1042/BJ20112100. 42. Buckland RJ, Watt DL, Chittoor B, Nilsson AK, Kunkel TA, Chabes A. Increased and Imbalanced dNTP Pools Symmetrically Promote Both Leading and Lagging Strand Replication Infidelity. PLoS Genet. 2014;10(12). doi:10.1371/journal.pgen.1004846. 43. Bartkova J, Horejsí Z, Koed K, et al. DNA damage response as a candidate anti-cancer barrier in early human tumorigenesis. Nature. 2005;434(7035):864-870. doi:10.1038/nature03482. 44. van Montfoort N, Olagnier D, Hiscott J. Unmasking immune sensing of retroviruses: Interplay between innate sensors and host effectors. Cytokine Growth Factor Rev. 2014;25(6):657-668. doi:10.1016/j.cytogfr.2014.08.006.

157

45. Maelfait J, Bridgeman A, Benlahrech A, Cursi C, Rehwinkel J. Restriction by SAMHD1 Limits cGAS/STING-Dependent Innate and Adaptive Immune Responses to HIV-1. Cell Rep. 2016;16(6):1492-1501. doi:10.1016/j.celrep.2016.07.002. 46. Manel N, Hogstad B, Wang Y, Levy DE, Unutmaz D, Littman DR. A cryptic sensor for HIV-1 activates antiviral innate immunity in dendritic cells. Nature. 2010;467(7312):214-217. doi:10.1038/nature09337. 47. Brégnard C, Benkirane M, Laguette N. DNA damage repair machinery and HIV escape from innate immune sensing. Front Microbiol. 2014;5(APR):1-10. doi:10.3389/fmicb.2014.00176. 48. Arnold LH, Kunzelmann S, Webb MR, Taylor IA. A Continuous Enzyme-Coupled Assay for Triphosphohydrolase Activity of HIV-1 Restriction Factor SAMHD1. Antimicrob Agents Chemother. 2015;59(1):186-192. doi:10.1128/AAC.03903-14. 49. Seamon KJ, Stivers JT. A High-Throughput Enzyme-Coupled Assay for SAMHD1 dNTPase. J Biomol Screen. 2015;20(6):801-809. doi:10.1177/1087057115575150. 50. Huynh TN, Luo S, Pensinger D, et al. An HD-domain phosphodiesterase mediates cooperative hydrolysis of c-di-AMP to affect bacterial growth and virulence. Proc Natl Acad Sci U S A. 2014;112(7):1-10. doi:10.1073/pnas.1416485112. 51. Rao F, Qi Y, Murugan E, Pasunooti S, Ji Q. 2’,3’-cAMP hydrolysis by metal-dependent phosphodiesterases containing DHH, EAL, and HD domains is non-specific: Implications for PDE screening. Biochem Biophys Res Commun. 2010;398(3):500-505. doi:10.1016/j.bbrc.2010.06.107. 52. Meyer-klaucke W, Marchfelder A. Metal Requirements and Phosphodiesterase Activity of tRNase Z Enzymes †. Biochemistry. 2007;46(2007):14742-14750. 53. Keppetipola N, Shuman S. A Phosphate-binding Histidine of Binuclear Metallophosphodiesterase Enzymes Is a Determinant of 2’,3’-Cyclic Nucleotide Phosphodiesterase Activity. J Biol Chem. 2008;283(45):30942-30949. doi:10.1074/jbc.M805064200. 54. Keppetipola N, Shuman S. Characterization of the 2???,3??? cyclic phosphodiesterase activities of Clostridium thermocellum polynucleotide kinase-phosphatase and bacteriophage ?? phosphatase. Nucleic Acids Res. 2007;35(22):7721-7732. doi:10.1093/nar/gkm868. 55. Feng BY, Shoichet BK. A detergent-based assay for the detection of promiscuous inhibitors. Nat Protoc. 2006;1(2):550-553. doi:10.1038/nprot.2006.77. 56. Shoichet BK. Interpreting steep dose-response curves in early inhibitor discovery. J Med Chem. 2006;49(25):7274-7277. doi:10.1021/jm061103g. 57. Prinz H. Hill coefficients, dose-response curves and allosteric mechanisms. J Chem Biol. 2010;3(1):37-44. doi:10.1007/s12154-009-0029-3. 58. Baell JB, Holloway GA. New substructure filters for removal of pan assay interference compounds (PAINS) from screening libraries and for their exclusion in bioassays. J Med Chem. 2010;53(7):2719-2740. doi:10.1021/jm901137j. 59. Dahlin JL, Nissink JWM, Strasser JM, et al. PAINS in the assay: Chemical mechanisms of assay interference and promiscuous enzymatic inhibition observed during a sulfhydryl-scavenging HTS. J Med Chem. 2015;58(5):2091-2113. doi:10.1021/jm5019093.

158

60. Schorpp K, Rothenaigner I, Salmina E, et al. Identification of Small-Molecule Frequent Hitters from AlphaScreen High-Throughput Screens. J Biomol Screen. 2014;19(5):715-726. doi:10.1177/1087057113516861. 61. Baell J, Walters MA. Chemical con artists foil drug discovery. Nature. 2014;513:481- 483. doi:10.1038/513481a. 62. Pery E, Sheehy A, Miranda Nebane N, et al. Redoxal, an inhibitor of de novo pyrimidine biosynthesis, augments APOBEC3G antiviral activity against human immunodeficiency virus type 1. Virology. 2015;484:276-287. doi:10.1016/j.virol.2015.06.014. 63. Usach I, Melis V, Peris J-E. Non-nucleoside reverse transcriptase inhibitors: a review on pharmacokinetics, pharmacodynamics, safety and tolerability. J Int AIDS Soc. 2013;16(1). doi:10.7448/IAS.16.1.18567. 64. Trott O, Olson AJ. AutoDock Vina: Improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem. 2009:NA-NA. doi:10.1002/jcc.21334. 65. Pettersen EF, Goddard TD, Huang CC, et al. UCSF Chimera?A visualization system for exploratory research and analysis. J Comput Chem. 2004;25(13):1605-1612. doi:10.1002/jcc.20084.

159

CONCLUSION

SAMHD1 was first identified as an interferon stimulated gene that functioned as a negative modulator of the innate immune response. Interest in SAMHD1 greatly increased however, with the discovery of its role as a retroviral, and specifically HIV, restriction factor. At the onset of this investigation, SAMHD1 had just been revealed to be a dGTP dependent deoxynucleotide triphosphohydrolase. Little was known in regards to the mechanism of catalysis, catalytic activation, and cellular regulation. During the period in which this study was ongoing, many of these questions concerning the basic biochemistry and cell biology of

SAMHD1 were answered by the myriad of laboratories working to gain an in depth understanding of SAMHD1 function. Those efforts have been thoroughly reviewed in the introduction and proved to be instrumental in guiding this study. Building on these efforts, the SAMHD1 field is now beginning to focus on the intracellular pathways that regulate

SAMHD1 and how they calibrate the enzyme’s activity to align with the specific metabolic needs of the cell at each stage in the cell cycle.

The primary objective of this research was to investigate the mechanisms of regulation that modulate SAMHD1 activity in cells. A secondary objective was to determine the effect of

160

SAMHD1 regulation on broader cellular processes like nucleotide metabolism and cell cycle progression. Given that redox signaling is an essential regulatory pathway in multiple cellular processes I hypothesized that SAMHD1 is a redox regulated enzyme and that redox signaling is important for modulating SAMHD1 activity in vivo.

In order to test this hypothesis, a variety of biophysical, biochemical, and cell culture techniques were utilized. The data reveal that SAMHD1 is indeed a redox regulated enzyme, as treatment with H2O2 reversibly inhibits tetramerization and catalytic activity. The redox sensitivity of SAMHD1 is mediated by a cysteine triad located adjacent the a regulatory nucleotide binding site. This cysteine triad, termed the cysteine switch, can form a disulfide bond when oxidized that results in structural changes that modify the dNTPase capability of

SAMHD1. This effect is not limited to in vitro studies. SAMHD1 is demonstrated to be oxidized in cells and shown to localize to sites of oxidation following proliferative signaling induced ROS generation. Further, redox regulation of SAMHD1 is demonstrated to be an important mechanism of regulating its activity in cells. Redox insensitive SAMHD1 mutants exhibited increased catalytic activity, reduced intracellular dNTP pools, and inhibited growth rates characterized by G2/M arrest.

The results of this study advance the field of SAMHD1 research in several important ways. First, it identifies a previously undescribed mechanism of regulating SAMHD1 catalytic activity. Second, by demonstrating oxidation of SAMHD1 in vivo and linking it to alterations in physiological processes, this study provides both a model and rationale for the oxidation of SAMHD1. This model will serve as a guide to hypothesis generation for future research. Lastly, the findings from this research add further evidence to the concept

161

of redox regulation as a fundamental pathway controlling cellular processes. By depicting a clear interdependence between proliferative signaling, ROS generation, and SAMHD1 activity, this work unambiguously couples redox signaling to nucleotide metabolism. This connection advances our understanding of both fields and lays the groundwork for significant research efforts that further characterize the relationship.

Future research that builds off the work presented in this study should focus on both biophysical/biochemical and cell based inquiries. Empirical evidence from biochemical and structural studies is required to dissect the precise reaction mechanism through which oxidation inactivates SAMHD1. It will also be important to determining if oxidation facilitates other post-translational modifications such as phosphorylation, glutathionylation, sumoylation, or ubiquitylation. Cell based studies should initially focus on validating the results from SKOV3 studies in multiple cell lines. If the results are consistent across cell types, more in depth investigations into the spatiotemporal nature of SAMHD1 oxidation, the pathways controlling it, and the downstream effects will be useful illuminating the complete schema of SAMHD1 redox regulation. A detailed investigation of how intracellular signaling pathways converge on SAMHD1 and are integrated by the enzyme in order to calibrate its activity will also be important in order to fully elucidate the ways in which cells modulate nucleotide pools to maintain genome stability. Lastly, exploring the effect of oxidation of SAMHD1 on disease state represents a potentially fascinating line of basic research inquiry with large implications for translational science. Do cancer cells circumvent metabolite restriction by oxidizing SAMHD1? Does cytokine signaling, which can also generate a ROS burst in cells, enhance or impede

SAMHD1’s viral restriction capacity? Do certain AGS mutations make SAMHD1 more

162

susceptible to oxidation. Questions such as these will require concerted research efforts to answer them. The answers however, may yield a large payoff both in our understanding of

SAMHD1 and its role in disease and in potential applications for improving health outcomes.

A third objective of this study was to identify small molecule compounds that could modulate SAMHD1 activity. Discovery of an activator or inhibitor of SAMHD1 would have multiple potential therapeutic and research applications. In order to achieve this objective, a novel high throughput screen was developed and validated. Following optimization of the assay, a small chemical library of FDA approved compounds was screened for inhibition of SAMHD1 dNTPase activity. Several top hits were further validated and five small molecule inhibitors of SAMHD1 with IC50s in the micromolar range were identified.

This work advances efforts to target SAMHD1 therapeutically. The HTS represents an improvement over other published SAMHD1 screening assays as it directly measures catalytic activity with an easy to read colorimetric endpoint. Additionally, in the search for a robust inhibitor of SAMHD1, the identification of multiple chemically diverse inhibitors represent starting points for compound optimization. Future research efforts should continue to validate these top hits using structural and cell culture approaches. In parallel, it will also be necessary to collaborate with a medicinal chemist in order to optimize the molecular scaffold and create a robust and patentable new chemical entity. Lastly, it will be important to continue to screen larger libraries of compounds to identify additional compounds or pharmacophores that can inform the development of SAMHD1 inhibitors.

163

In closing, the research described in this study has advanced our understanding of SAMHD1 biology and the associated regulatory pathways and downstream processes. Despite the predominant focus thus far by the research community on SAMHD1 as an effector of the innate immune response, a larger, more complex picture of SAMHD1 is coming into focus. SAMHD1 is a dynamic and multifaceted enzyme that performs important functions in several processes essential to normal cellular homeostasis (Fig. 1). Continued research into its structure, function, and regulation will yield important insights into the nature of nucleotide metabolism, genome fidelity, and innate immunity, and the pathologies associated with their dysregulation.

Figure 1. SAMHD1 is a key cog in the cellular machinery that maintains homeostasis, genomic integrity, and a robust immune response.

164

CURRICULUM VITAE

Christopher H. Mauney, MS, MBA, PhD

SUMMARY

I am a PhD/MBA with cross disciplinary training and project management experience. My unique skill set enables me to problem solve from multiple perspectives, add value across all phases of the biomedical product development spectrum, and effectively communicate with stakeholders from diverse backgrounds. I am passionate about the translation of rigorous science to implementable technologies. I am seeking a position where I can leverage my scientific and business training to make meaningful contributions to biomedical innovation that improves the health outcomes of individuals and communities.

EDUCATION

Wake Forest University, School of Medicine 2011 – 2017 Doctor of Philosophy (PhD Candidate), Biochemistry and Molecular Biology Certificate in Structural and Computational Biology Research focus: Enzymatic regulation and therapeutic targeting of nucleotide metabolism ● Herbert C. Cheung, PhD Award for demonstrated excellence in the field of Biochemistry, Wake Forest University Graduate School, 2016 ● Cowgill Fellowship, Wake Forest Dept of Biochemistry and Molecular Biology, 2015 ● Artom Fellowship, Wake Forest Dept of Biochemistry and Molecular Biology, 2014 ● NIH T32 Predoctoral Fellowship in Structural and Computational Biophysics, 2012-2014

Wake Forest University, School of Business 2013-2015 Masters of Business Administration ● Beta Gamma, International Business Honors Society, 2015 ● Boston Scientific Healthcare Business Strategy Competition, 2nd place, 2015

Johns Hopkins University 2007 - 2009 Masters of Science, Advanced Biotechnology ● Training focused on applied biotechnology product development, regulatory affairs, biostatistics and epidemiology

Wake Forest University 1999-2003 Bachelor of Arts, English Literature, Minor(s): Biology, Political Science

165

SELECT PROFESSIONAL EXPERIENCE

IntraHealth International Chapel Hill, NC Infectious Disease Program Officer 2010 – 2011 ● Managed technical and programmatic implementation of HIV prevention, care, and treatment programs in Central America and Africa which resulted in delivery of high quality services and measurable increases in key health indicators ● Collaborated with international teams to develop clinical and laboratory capacity for partner country Ministries of Health systems strengthening initiatives ● Forecast, developed, and reported annual and quarterly budget and project pipelines

United States Agency for International Development (USAID) Washington, DC Global Health Research, Technology, and Utilization Division 2008 – 2010 Biomedical Research Analyst ● Provided technical and programmatic guidance to advance lead products/compounds from a global health focused clinical and pre-clinical research portfolio. ● Collaborated with Chief Technical Officers to administer over $45 million in annual funding for global health research by creating scopes of work, budgets, reporting plans, timelines and deliverables, strategic directions, and research priorities ● Represented USAID at international scientific meetings during travel to Africa and Asia ● Supervised and trained new employees within the Bureau for Global Health

United States Agency for International Development (USAID) Washington, DC Global Health Research, Technology, and Utilization Division 2005 – 2008 Program Assistant ● Managed budget, programmatic, and administrative details for three USAID health technology research cooperative agreements. ● Acted as an in-country consultant for USAID Missions in Tbilisi, Georgia and Windhoek, Namibia to assist in the preparation and submission of annual operational plans

The World Technology Network London, United Kingdom Network Assistant 2003 – 2004 ● Coordinated the logistics of the joint WTN-UNESCO World Energy Technology Summit, focusing on sustainable global energy development using innovative technology

166

SELECT PUBLICATIONS AND PRESENTATIONS

● Mauney C, et al. The SAMHD1 dNTP triphosphohydrolase is controlled by a redox switch. Antioxid Redox Signal. 2017; (336): ars.2016.6888. doi:10.1089/ars.2016.6888. ● “The structural and functional regulatory mechanisms of the dNTP triphosphohydrolase SAMHD1 and opportunities for drug discovery.” Wake Forest University Dept of Biochemistry and Molecular Biology, Departmental Seminar(s), 2012-2015 ● Aziz, E., Penders, C., Chani, K., Hight, V., Mauney, C., “Child Mortality in Namibia: Before and After Changes in HIV Testing and Treatment Guidelines.” IAS 2010 Poster ● Mauney, C., Claypool, L. Non-Human Primate Models for Microbicide R&D. The Microbicide Quarterly 6.2 (2008): 13-15 ● Invited presenter for USAID Mini University, 2007: “Microbicides in Human Trials”

167