T cell-mediated adaptive immunity requires naïve, unstimulated T cells to transition from a quiescent metabolic state into a highly proliferative state upon T cell receptor engagement. This complex process depends on transcriptional changes mediated by Ca2+-dependent NFAT signaling, mTOR-mediated signaling and increased activity of the guanine nucleotide biosynthetic inosine-5′-monophosphate (IMP) dehydrogenase 1 and 2 enzymes (IMPDH1 and IMPDH2, hereafter IMPDH). Inhibitors of these pathways serve as potent immunosuppressants. Unexpectedly, we discovered that all three pathways converge to promote the assembly of IMPDH protein into micron-scale macromolecular filamentous structures in response to T cell activation. Assembly is post-transcriptionally controlled by mTOR and the Ca2+ influx regulator STIM1. Furthermore, IMPDH assembly and catalytic activity were negatively regulated by guanine nucleotide levels, suggesting a negative feedback loop that limits biosynthesis of guanine nucleotides. Filamentous IMPDH may be more resistant to this inhibition, facilitating accumulation of the higher GTP levels required for T cell proliferation.
T cell receptor (TCR) stimulation initiates events leading to T cell activation, proliferation and differentiation. This core process of adaptive immunity is associated with dramatic metabolic alterations required for T cell proliferation (MacIver et al., 2013; Buck et al., 2015; Almeida et al., 2016; Bantug et al., 2018). Indeed, several immunosuppressive agents target these metabolic dependencies. For example, inhibitors targeting mechanistic target of rapamycin (mTOR) and an inhibitor of the inosine-5′-monophosphate (IMP) dehydrogenase 1 and 2 (IMPDH1 and IMPDH2, hereafter IMPDH) enzymes mycophenolic acid (MPA) are used to suppress organ transplant rejection, although how they function is not fully understood (Allison and Eugui, 1996; Thomson et al., 2009; Nguyen le et al., 2015). However, restoring guanine nucleotide levels in MPA-treated T cells rescues their proliferation (Quemeneur et al., 2003), demonstrating that production of guanine nucleotides is the primary role of IMPDH in T cell activation. Intriguingly, mTOR also regulates guanine nucleotide synthesis (Ben-Sahra et al., 2016; Valvezan et al., 2017).
An early event in T cell activation is an increase in cytosolic Ca2+. TCR-triggered release of endoplasmic reticulum (ER) Ca2+ is sensed by stromal interaction molecule 1 (STIM1), a protein that mediates store-operated Ca2+ entry and is critical for sustained cytosolic Ca2+ (Hogan et al., 2003; Gwack et al., 2007; Oh-Hora et al., 2008; Srikanth and Gwack, 2013). Cytosolic Ca2+ activates calcineurin, which dephosphorylates nuclear factor of activated T cells (NFAT) leading to nuclear import. Indeed, calcineurin inhibitors are potent immunosuppressants (Martinez-Martinez and Redondo, 2004). Importantly, the mTOR, IMPDH and STIM1/calcineurin pathways are thought to mediate the immune response via non-overlapping mechanisms. Here, we present evidence that these pathways converge to promote the TCR-mediated assembly of a macromolecular structure containing IMPDH.
IMPDH is the rate-limiting enzyme in guanine nucleotide biosynthesis and is dramatically upregulated by T cell activation (Dayton et al., 1994). IMPDH assembles into micron-scale structures in cultured cells, generally in response to pharmacological perturbations (Ji et al., 2006; Calise et al., 2014, 2016; Carcamo et al., 2014; Keppeke et al., 2015; Zhao et al., 2015). The physiological relevance of these filamentous assemblies is unknown, although they are reported to be catalytically active (Chang et al., 2015; Zhao et al., 2015; Anthony et al., 2017). We report that IMPDH assembles into filaments in T cells both in vivo and ex vivo in response to TCR engagement. Unexpectedly, we found that assembly, but not upregulated expression, of IMPDH was dependent on STIM1 and mTOR. Thus, IMPDH regulation is a common thread linking the pathways targeted by three major classes of immunosuppressive drugs, suggesting that IMPDH assembly serves an essential function in T cell activation by supporting guanine nucleotide production.
RESULTS AND DISCUSSION
TCR stimulation promotes IMPDH assembly in T lymphocytes
Murine splenic T cells were isolated and activated using antibodies against the TCR co-receptors CD3ε and CD28 (Fig. 1A). Strikingly, IMPDH assembled into linear assemblies and toroids in the vast majority of T cells within 24 h (Fig. 1A,B). Refinement of the T cell population into CD4+ and CD8+ subsets by fluorescence-activated cell sorting (FACS) revealed IMPDH filaments in both subsets (Fig. S1A). Filament assembly was accompanied by a dramatic increase in IMPDH protein levels (Fig. 1C,D) demonstrating that increased IMPDH expression and filament assembly are direct downstream consequences of TCR activation and establishing a system to analyze these processes in vitro.
To confirm the in vivo relevance, we investigated T cells in the natural context of lymphocytic choriomeningitis virus (LCMV) infection. In LCMV-infected mice, it is known that T cells recognizing LCMV antigens become activated and proliferate. Following the resolution of infection, ∼95% of activated T cells undergo apoptosis and surviving memory T cells confer protection against future LCMV infection (Murali-Krishna et al., 1998). We infected mice with LCMV for 7 days, a time of peak anti-viral CD8+ T cell cytotoxicity (Hassett et al., 2000; Knipe and Howley, 2013), and isolated splenic T cells. Immunostaining revealed IMPDH filaments that were absent in cells from uninfected mice (Fig. 1E). Western blotting revealed a 3-fold increase in IMPDH protein levels in total splenic T cells from LCMV-challenged versus control mice (Fig. 1F). To ask whether IMPDH filaments persist in memory T cells, CD69+ T cells (representing a mixed population of both memory T cells and activated T cells) were isolated by FACS at 30 days post-infection. No IMPDH filaments were observed in these CD69+ T cells (Fig. S1B), demonstrating that the transient IMPDH filament assembly during initial activation does not persist in quiescent memory cells.
STIM1 and mTOR regulate IMPDH filament assembly
To elucidate signaling mechanisms controlling IMPDH assembly, we compared IMPDH filament formation and expression in splenic T cells isolated from mice with a T cell-specific knockout of STIM1 (STIM1fl/fl/Cd4-Cre) and control mice (STIM1fl/fl) (Oh-Hora et al., 2008). There was a strong reduction of IMPDH filament formation in STIM1-deficient cells compared to control (Fig. 2A,B). STIM1 loss had only moderate effects on IMPDH protein expression, indicating that STIM1 is not primarily responsible for IMPDH upregulation (Fig. 2C); this demonstrates that IMPDH assembly is not a simple consequence of IMPDH upregulation, implying that assembly relies on a post-translational Ca2+-dependent regulatory mechanism. Interestingly, IMPDH expression and assembly occurred normally in T cells isolated from STIM2-deficient mice (Fig. S2), demonstrating a critical role for STIM1 specifically in IMPDH filament assembly.
mTOR is a master regulator of diverse metabolic pathways during T cell activation (Chi, 2012; MacIver et al., 2013). Recently, mTOR was shown to promote purine biosynthesis (Ben-Sahra et al., 2016), in part to support ribosomal biogenesis (Valvezan et al., 2017). Conversely, purine levels regulate mTORC1 activity (Emmanuel et al., 2017; Hoxhaj et al., 2017), highlighting an intimate relationship between mTOR and purine nucleotides. Furthermore, mTOR acts downstream of store-operated Ca2+ entry to promote metabolic alterations required for T cell activation (Vaeth et al., 2017). We therefore asked whether mTOR activity was important for IMPDH filament assembly.
Because mTOR plays roles in both early and late T cell activation, we stimulated T cells overnight to establish IMPDH filaments and then utilized small-molecule inhibitors to acutely inhibit mTOR (Fig. 3A). A 1 h treatment with allosteric mTOR inhibitors (rapamycin, everolimus or temsirolimus) or the ATP-competitive inhibitor AZD8055 led to the almost completely disassembly of IMPDH filaments (Fig. 3B,C). Rapid disassembly of IMPDH filaments after treatment with mTOR inhibitors implies that continued mTOR signaling is required for their maintenance. Furthermore, filament disassembly was not due to changes in IMPDH protein expression (Fig. 3D), supporting the existence of an mTOR-dependent post-translational regulatory mechanism.
Because mTOR acts downstream of store-operated Ca2+ entry in T cells (Vaeth et al., 2017), we considered whether the failure of IMPDH to assemble in STIM1-deficient T cells was due to loss of mTOR activity. However, phosphorylation of the mTOR target S6 ribosomal protein was similar in STIM1-deficient and control cells (Fig. 2C). This suggests that STIM1 and mTOR contribute independently to promoting IMPDH assembly.
IMPDH filaments in T cells are regulated by guanine nucleotide levels
IMPDH contains allosteric regulatory binding sites for certain guanine and adenine nucleotides. ATP promotes the assembly of recombinant IMPDH octamers into linear polymers (Labesse et al. 2013; Anthony et al., 2017). These protofilaments are assumed to be precursors in the assembly of the larger IMPDH assemblies observed in cells. Within ATP-IMPDH filaments, individual octamers can adopt either expanded or compressed conformations (Anthony et al., 2017). By contrast, binding of GDP or GTP to partially overlapping sites specifically promotes the catalytically inactive collapsed form. Whether the collapsed form affects the stability of IMPDH polymers is unknown. Thus, relative cellular purine levels are thought to regulate both IMPDH polymer conformation and catalytic activity. To examine IMPDH filament responsiveness to nucleotides in T cells, we stimulated cells overnight to induce IMPDH assembly, and then supplemented the medium with individual nucleosides for 1 h. Following uptake, nucleosides are readily converted into the corresponding mono-, di- and tri-phosphorylated nucleotides, bypassing de novo synthesis pathways (Fig. 4A). Guanosine addition triggered rapid disassembly of IMPDH filaments whereas filament number and morphology were unchanged by other nucleosides (Fig. 4B,C). Furthermore, disassembly of IMPDH filaments by guanosine was not associated with alterations in IMPDH protein levels (Fig. 4D). Thus, guanosine and/or its downstream metabolites promote IMPDH filament disassembly.
To investigate the effect of guanosine addition on de novo purine nucleotide biosynthesis, we monitored the incorporation of isotopically labeled glycine into purine nucleotides after guanosine treatment of activated splenic T cells. Glycine is an early precursor in de novo purine biosynthesis (Fig. 4A) and is incorporated into the purine ring. T cells stimulated overnight or left unstimulated were incubated with [13C2,15N]-glycine and both total and [13C2,15N]-glycine-labeled IMP, AMP and GMP levels were quantified. IMP is the last common biosynthetic precursor of adenine and guanine nucleotides and AMP and GMP were chosen as representative nucleotides of each pathway.
TCR stimulation led to a dramatic upregulation of total IMP, AMP and GMP pools (Fig. 4E), as expected (Fairbanks et al., 1995). This was at least partially due to increased de novo purine biosynthesis, since a corresponding increase was observed in the fraction of each nucleotide containing the [13C2,15N]-glycine-derived isotopomer in stimulated versus unstimulated cells (Fig. 4F); however, it should be noted that stimulated cells exhibit higher uptake of exogenous glycine (Fig. S3). Guanosine treatment led to a further ∼2-fold increase in total GMP pool size (Fig. 4E). Strikingly, this was associated with almost complete inhibition of [13C2,15N]-glycine incorporation into GMP, demonstrating that although guanosine addition increased the total GMP pool, it suppressed de novo GMP synthesis (Fig. 4F). These observations are best explained by the conversion of exogenous guanosine to form unlabeled GMP, GDP and GTP. GMP directly inhibits IMPDH catalytic activity competitively with the substrate IMP (Gilbert et al., 1979), while GDP and GTP inhibit IMPDH allosterically by promoting the catalytically inactive, collapsed conformation (Anthony et al., 2017; Buey et al., 2017). Guanine nucleotides are also known to exert feedback control at two additional points in purine biosynthesis; GMP inhibits the synthesis of phosphoribosyl pyrophosphate in the rate-limiting first step of purine biosynthesis, and GTP is a co-factor required by adenylosuccinate synthase activity to divert IMP toward adenine nucleotide biosynthesis (Fig. 4A). Consistent with this, we found that guanosine addition decreased total IMP pools by half and suppressed incorporation of [13C2,15N]-glycine into IMP (Fig. 4F). We also observed a slight, non-significant decrease in total AMP, perhaps due to the decreased availability of IMP (Fig. 4E). However, the incorporation of [13C2,15N]-glycine into AMP was increased (Fig. 4F), suggesting that any IMP synthesized under conditions of excess guanosine is shunted to adenine nucleotide biosynthesis in the absence of IMPDH activity.
In conjunction with the immunofluorescence data, these results demonstrate for the first time that IMPDH filaments in a primary tissue are sensitive to alterations in guanine nucleotide levels. Increased guanine nucleotides are associated with the disassembly of IMPDH filaments and an inhibition of IMPDH catalytic activity. To test whether these two events are mechanistically connected, we examined whether GTP destabilizes IMPDH polymers formed in vitro. Purified, recombinant IMPDH2 was incubated with ATP to promote assembly (Labesse et al., 2013; Anthony et al., 2017) and samples were analyzed by electron microscopy (EM) before and after the addition of GTP for various times. GTP triggered disassembly of IMPDH polymers (Fig. 4G,H), suggesting that dissolution of IMPDH filaments in T cells treated with guanosine is mediated by the direct action of guanine nucleotides on IMPDH and is associated with the inhibition of IMPDH catalytic activity.
Our results suggest that IMPDH filaments in stimulated T cells are catalytically active, consistent with prior observations that recombinant IMPDH polymers are catalytically active (Chang et al., 2015; Anthony et al., 2017). Excess guanine nucleotides suppress IMPDH assembly and activity, shunting IMP into adenosine nucleotide biosynthesis, presumably to restore balance to the adenine and guanine nucleotide pools. Since guanosine rescues proliferation of T cells treated with the IMPDH inhibitor MPA (Quemeneur et al., 2003), guanine nucleotide production is the primary role of IMPDH in supporting T cell proliferation. Our data suggest that filament assembly may support IMPDH activity in this context.
While IMPDH protein expression is highly upregulated following TCR stimulation, increased IMPDH expression is not sufficient to promote filament assembly. Rather, IMPDH assembly is governed by post-translational mechanisms including guanine nucleotide levels and downstream mediators of TCR signaling, including mTOR and STIM1. More broadly, our findings introduce an experimentally tractable system to analyze IMPDH filament function and regulation in a physiologically relevant cell type.
Previous work has demonstrated that polymerization per se does not enhance IMPDH catalytic activity (Anthony et al., 2017). However, our finding that guanosine treatment promotes both IMPDH filament dissolution and inhibition of IMPDH catalytic activity suggests the interesting possibility that filament assembly might serve to modulate IMPDH sensitivity to negative feedback regulation by guanine nucleotides. For example, IMPDH assemblies in T cells may stabilize the active open conformation of IMPDH octamers, rendering them more resistant to octamer collapse and inactivation by GDP/GTP compared to free octamers. As a result, cells with filamentous IMPDH would be able to maintain higher steady-state levels of guanine nucleotides needed to support T cell activation and proliferation. In contrast to this model, a prior study found that polymerization of recombinant IMPDH in vitro did not protect against GTP-mediated catalytic inhibition (Anthony et al., 2017). However, the linear polymers assembled in vitro may not reflect the structure, composition, or activity of the larger IMPDH assemblies in vivo.
Unexpectedly, we found that three pathways essential for T cell activation, STIM1/calcineurin-mediated NFAT activation, signaling by mTOR and increased IMPDH activity all converge to promote the assembly of IMPDH-containing filaments on TCR engagement. This observation may point to a critical role for filament assembly in T cell function. Indeed, we found that immunosuppressive mTOR inhibitors disrupt IMPDH filament assembly in this context. We expect that calcineurin-targeting agents would produce a similar effect based on the phenotype of STIM1-deficient T cells. In contrast, the IMPDH-targeted immunosuppressant MPA promotes IMPDH assembly in a variety of cell types (Ji et al., 2006; Calise et al., 2014; Carcamo et al., 2014; Keppeke et al., 2015). This could be a homeostatic response to reduced GTP levels in these cells or a direct consequence of MPA binding to IMPDH. Indeed, we found that MPA treatment also promoted IMPDH assembly even in unstimulated T cells without altering basal IMPDH protein levels (Fig. S4). This interesting observation demonstrates that increased IMPDH protein expression is not essential for filament assembly, consistent with a post-translational mechanism. More importantly, IMPDH assembly could serve as a pharmacodynamic biomarker in transplant patients treated with MPA.
MATERIALS AND METHODS
The Institutional Animal Care and Use Committees at Fox Chase Cancer Center and Temple University School of Medicine approved all animal procedures. Mice that were 8 to 12 weeks old were used for all experiments. Wild-type C57BL/6 male mice were either obtained from breeding colonies at the Fox Chase Cancer Center (FCCC) Laboratory Animal Facility or purchased from Taconic Biosciences. The generation of loxP-flanked STIM1 and STIM2 genes on the C57BL/6 background and the production of mice with T cells deficient in STIM1 and STIM2 has been previously described (Oh-Hora et al., 2008). For LCMV infection studies, immunocompetent C57BL/6 mice were infected with 2×105 plaque-forming units of LCMV-Armstrong as previously described (Matullo et al., 2011) for either 7 or 30 days. All mice were euthanized by CO2 asphyxiation.
Isolation of splenic T cells
Spleens were dissected from mice, crushed using the blunt end of a sterile syringe plunger, and passed through a 100 µm mesh filter incubating in T cell medium [RPMI-1640 supplemented with 10% heat-inactivated fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM L-glutamine, 1 mM sodium pyruvate, 50 µM β-mercaptoethanol, and non-essential amino acids (Gibco)]. Splenocytes were harvested and resuspended in ammonium chloride potassium (ACK) lysis buffer (155 mM ammonium chloride, 10 mM potassium bicarbonate, and 0.1 mM EDTA) to lyse red blood cells. Lymphocytes were harvested and subsequently resuspended in HBSS supplemented with 1% bovine serum albumin (BSA). B cells were depleted by negative selection by binding to mouse CD45R (B220) microbeads (Miltenyi Biotec) and applying the cells to LS columns (Miltenyi Biotec) according to manufacturer's instructions.
For FACS isolation of cell populations, splenic T cells were isolated as described above and incubated with allophycocyanin (APC)-conjugated anti-CD3ε (1:50, BD Biosciences 561826), APC-Cy7-conjugated anti-CD4 (1:100, BD Biosciences 561830), BV605-conjugated anti-CD8a (1:50, Biolegend 100743), phycoerythrin (PE)-Cy7-conjugated anti-CD25 (1:100, BD Biosciences 561780), PE-conjugated anti-CD69 (1:100, BD Biosciences 561932), and Alexa Fluor 488-conjugated anti-IMPDH2 antibodies (1:50, Abcam ab200770). Cell sorting was carried out on a 15-color BD FACSAria II instrument outfitted with 407, 488, 532 and 594 nm lasers. Data were collected with BD FACS Diva software.
Ex vivo stimulation
Except where otherwise indicated, 24-well tissue culture plates were coated with 500 µl of a 10 µg/ml solution of anti-mouse CD3ε clone 145-2C11 (Biolegend 100359) solution for 4 h at 37°C. Coating solution was removed and wells were washed with T cell medium to remove excess anti-CD3ε antibody. 1.5×106 cells in 1 ml were applied to each well and anti-mouse CD28 antibody (clone 37.51, Biolegend 102112) was added to each well at a final concentration of 2 µg/ml. Stimulation was carried out for 24 h in a humidified incubator at 37°C, 5% CO2. If used, mTOR inhibitors or guanosine were added for the last hour of stimulation directly to the medium. Guanosine (Sigma) was used at the indicated concentrations while rapamycin (LC Laboratories) was used at 20 nM, everolimus (Sigma) was used at 1 µM, temsirolimus (Sigma) was used at 10 µM and AZD8055 (LC Laboratories) was used at 100 nM.
Immunofluorescence and quantification
Cells were transferred to 8-well Lab-Tek II CC2 chamber slides (Thermo Scientific) and incubated at 37°C and 5% CO2 for 1 h to allow cells to adhere to the surface. Cells were then fixed in 10% formalin diluted into phosphate-buffered saline (PBS). Cells were subsequently washed with PBS, permeabilized with 0.5% Triton X-100 in PBS, and washed with 0.1% Triton X-100. Non-specific binding was blocked with PBS supplemented with 0.1% Triton X-100, 2% BSA and 0.1% sodium azide. Anti-IMPDH (1:500, Abcam ab129165) diluted in blocking buffer was used to stain cells. Alexa Fluor 488-conjugated goat anti-rabbit secondary antibody (ThermoFisher Scientific A-11001) was used to detect primary antibody binding and DAPI was used to stain nuclei. Vectashield mounting medium (Vector) was applied to the cells and 24×50 mm coverglasses (Fisher Scientific) were mounted on slides and sealed using clear nail polish.
Imaging was performed using the 63× oil immersion objective of a Leica Microsystems TCS SP8 Advanced confocal laser scanning microscope. Images were acquired using Leica Application Suite Advanced Fluorescence software at 400 Hz with a zoom factor of 2.0. For each image, z-stacks of focal planes of 0.5 µm depth were collected and presented as maximum projections. The number of cells containing filaments under each treatment condition was quantified as described in the figure legends and all statistics were computed in Graphpad Prism.
Cells were washed with PBS, resuspended and lysed in 30 µl sample buffer (62.5 mM Tris-HCl pH 6.8, 2% SDS, 10% glycerol, 100 mM dithiothreitol and 0.15 mM Bromophenol Blue) per 106 cells, sonicated in a water bath for 5 min, and heated at 95°C for 90 s. Samples were run on SDS-PAGE gels and then transferred to nitrocellulose membranes using the Amersham TE 70 semidry transfer apparatus (GE Healthcare). Nitrocellulose membranes were blocked in a solution of 5% nonfat dry milk in Tris-buffered saline with 0.5% Tween-20. Anti-IMPDH2 (Abcam ab129165), anti-GAPDH (Santa Cruz Biotechnology sc-32233), anti-phospho-S6 ribosomal protein (Ser 235/236) (Cell Signaling Technology 4858), and anti-S6 ribosomal protein (Cell Signaling Technology 2217) were used as primary antibodies. Either horseradish peroxidase (HRP)-conjugated goat anti-rabbit or goat anti-mouse IgG (ThermoFisher Scientific 31460 or 31430) were used as secondary antibodies. Blots were visualized using Western Blotting Luminol Reagent (Santa Cruz Biotechnology) and exposed to film in a dark room.
Glycine incorporation assay and nucleotide measurements
Splenic T cells were either left unstimulated or stimulated ex vivo as above except that 10-cm tissue culture plates were coated with 10 ml of the anti-mouse CD3ε solution and 30×106 cells in 20 ml were stimulated per plate. Cells were either left untreated or treated with the indicated concentrations of guanosine 20 h after stimulation. After 3 h, [13C2,15N]-glycine was added at a concentration of 35 mg/l. After a further hour, cells were harvested (250 g for 5 min) and washed twice with Dulbecco's PBS.
Samples were processed as described previously (Laourdakis et al., 2014). Briefly, cells were resuspended in 70 µl of a solution of 0.5 M perchloric acid, 200 µM [13C9,15N3]-CTP, vortexed for 10 s, and incubated on ice for 20 min. Lysates were then neutralized with 7 µl 5 M potassium hydroxide, vortexed for 10 s, and incubated on ice for 20 min. Debris was harvested at 11,000 g for 10 min on a microcentrifuge. Soluble lysates were further filtered using Amicon Microcon 0.5 ml YM-100 centrifugal filters as per the manufacturer's instructions.
Measurement of nucleotides was carried out on an Acquity I Class (Waters) ultra-performance liquid chromatography (UPLC) instrument coupled to a Xevo TQ-S micro triple quadrupole mass spectrometer (Waters). A total of 5 µl of each sample was loaded onto an Acquity UPLC HSS T3 column (1.8 mm, 2.1 mm×50 mm; Waters) with an in-line column filter. Quantitative analysis was performed using the multiple reaction monitoring mode and MassLynx v. 4.1 software. Graphpad Prism software was used to create final plots and compute statistics.
Glycine uptake assay
Splenic T cells were either left unstimulated or stimulated ex vivo as above. To 1.5×106 cells in 1 ml of medium, a mixture of 35 µg of cold glycine and 10 µCi of [2-3H]-glycine was added. After a 1 h incubation, cells were harvested, washed two times with PBS, and lysed with 1% SDS. Glycine uptake was assessed using a liquid scintillation counter (Beckman).
Negative stain EM of recombinant IMPDH2 and quantification
Recombinant human IMPDH2 protein was expressed and purified as previously described (Anthony et al., 2017). Briefly, IMPDH2 was cloned into pSMT3-Kan and the resultant plasmid was used to transform BL21(DE3) competent E. coli cells. Cells were grown in LB medium at 37°C to an OD600 of 0.8 and then cooled on ice for 5 min. Protein expression was then induced with 1 mM IPTG for 4 h at 30°C. Cells were harvested and resuspended in a buffer containing 50 mM potassium phosphate, 300 mM potassium chloride, 20 mM imidazole, 0.8 M urea and Benzonase nuclease (Sigma) at pH 8. Cells were lysed by homogenization using the Emulsiflex-C3 (Avestin). His-SUMO-tagged IMPDH2 was purified by nickel affinity chromatography using Ni-NTA Agarose (Qiagen) and an elution buffer containing 50 mM potassium phosphate, 300 mM potassium chloride, and 500 mM imidazole at pH 8.0. The His-SUMO tag was cleaved using 1 mg ULP1 protease (Mossessova and Lima, 2000) per 100 mg of protein overnight at 4°C. The next day, 1 mM dithiothreitol and 0.8 M urea were added and the protein was concentrated using a 30,000 MWCO filter (Millipore). IMPDH2 was purified by gel filtration chromatography using either a Superdex 200 or a Superose 6 gel filtration column in a buffer containing 50 mM Tris-HCl, 100 mM potassium chloride and 1 mM dithiothreitol at pH 7.4. Purified IMPDH2 was stored at −80°C as one-time-use aliquots.
Negative stain EM was carried out as previously described (Anthony et al., 2017). IMPDH2 was diluted to 1 µM in assembly buffer (50 mM Tris-HCl, 100 mM potassium chloride, 1 mM dithiothreitol and 1 mM ATP at pH 7.4) and incubated for 15 min at room temperature. 1 mM GTP was then added. At the indicated times, samples were applied to glow-discharged continuous carbon film EM grids and negative-stained with 1% uranyl formate. Transmission EM was performed using a FEI Tecnai G2 Spirit at 120 kV and a Gatan Ultrascan 4000 CCD interfaced with the Leginon software package (Suloway et al., 2005). Image processing was performed using Appion and RELION software (Lander et al., 2009; Scheres, 2012). Particles were picked manually from 59 electron micrographs, and classified into 16 classes using Relion two-dimensional classification. These 16 classes were pooled into three qualitative morphological groupings, corresponding to octamers that were either polymerized, at the ends of polymers or non-polymeric.
We thank Alana O'Reilly for helpful comments on the manuscript.
Conceptualization: K.C.D., J.M.K., J.S., G.F.R., J.R.P.; Methodology: K.C.D., Y.K., M.C.J., J.M.C., J.M.K., G.F.R., A.J.A., J.R.P.; Formal analysis: K.C.D., Y.K., M.C.J., J.M.C.; Investigation: K.C.D., Y.K., M.C.J., G.F.R.; Resources: J.S., G.F.R.; Data curation: K.C.D., Y.K., M.C.J., J.M.C.; Writing - original draft: K.C.D., J.R.P.; Writing - review & editing: K.C.D., Y.K., J.M.K., J.S., G.F.R., A.J.A., J.R.P.; Supervision: J.M.K., J.S., A.J.A., J.R.P.; Project administration: J.R.P.; Funding acquisition: K.C.D., J.M.K., J.S., G.F.R., A.J.A., J.R.P.
This work was supported by National Institutes of Health grants (GM083025 to J.R.P., T32 CA009035 to K.C.D., GM117907 to J.S., GM102503 to A.J.A. and GM118396 to J.M.K.), funds from the Pennsylvania Department of Health to J.R.P. and the F.M. Kirby Foundation to G.F.R., and a Fox Chase Cancer Center Board of Associates Postdoctoral Fellowship Award to K.C.D. Deposited in PMC for release after 12 months.
The authors declare no competing or financial interests.