Ensemble cryo-EM structures demonstrate human IMPDH2 filament assembly tunes allosteric regulation

Inosine monophosphate dehydrogenase (IMPDH) mediates the first committed step in guanine nucleotide biosynthesis and plays important roles in cellular proliferation and the immune response. The enzyme is heavily regulated to maintain balance between guanine and adenine nucleotide pools. IMPDH reversibly polymerizes in cells and tissues in response to changes in metabolic demand, providing an additional layer of regulatory control associated with increased flux through the guanine synthesis pathway. Here, we report a series of human IMPDH2 cryo-EM structures in active and inactive conformations, and show that the filament resists inhibition by guanine nucleotides. The structures define the mechanism of filament assembly, and reveal how assembly interactions tune the response to guanine inhibition. Filament-dependent allosteric regulation of IMPDH2 makes the enzyme less sensitive to feedback inhibition, explaining why assembly occurs under physiological conditions, like stem cell proliferation and T-cell activation, that require expansion of guanine nucleotide pools.


Introduction
Ribonucleotides play a central role in cellular physiology, and complex regulatory networks maintain 21 optimal nucleotide levels according to the variable metabolic state of the cell (Lane & Fan 2015). Under 22 most conditions, cells rely on salvage pathways to regenerate degradation products and maintain 23 nucleotide pools. However when nucleotide demand is high, for example during cellular proliferation, flux 24 through de novo nucleotide biosynthesis pathways is up-regulated.

36
IMPDH reversibly assembles into filaments in vertebrate cells and tissues, providing an additional layer 37 of regulation. Guanine deprivation induces assembly of IMPDH into micron-scale ultrastructures 38 composed of bundled filaments, which disassemble once homeostasis is restored (Labesse et al. 2013; three of which reached high resolution: the filament assembly interface, and two different reconstructions 119 of octameric filament segments (Table 1). We resolved many conformations of filament segments, 120 however for all segments the filament assembly interface was identical. Therefore the filament assembly 121 interface structure is a dataset consensus structure, an average of every segment included in the dataset.

122
We did not observe any conformational cooperativity in the compressed/extended conformational 123 equilibrium between IMPDH2 octamers sharing a filament assembly interface.

124
We determined the structure of the ATP/IMP/NAD + consensus filament assembly interface at 3.0 Å 125 resolution (Figs. 3C, S2B-C). This D4 symmetric region is composed of two symmetrically opposed 126 catalytic domain tetramers. The interface between tetramers is formed by the 12 amino-terminal residues 127 of eight protomers, which each extend from the core of the molecule to bind one catalytic domain on the 128 opposite tetramer, in a shallow surface groove formed by a short helix (476-485), two beta strands (51-enzyme filaments were actively turning over when flash frozen, we have likely captured a mixture of substrate-, intermediate-, and product-bound states. Attempts at focused classification of the active site 155 to structurally isolate these states were unsuccessful.

156
The extended conformation is an ensemble of flexible states 157 We resolved multiple conformations of filament segments in the ATP/IMP/NAD + dataset (Fig. S2D) To understand how IMP and GTP allosterically influence filament assembly and disassembly of ATPbound IMPDH2, we acquired cryo-EM data of the enzyme in two ligand states: 1) ATP/GTP/IMP, and 2) 189 ATP/GTP. To ensure morphological consistency, we sought to saturate the enzyme with GTP. For the 190 ATP/GTP dataset, we used 2 mM GTP, which for both our initial negative stain experiments and cryo-191 EM preparations resulted in complete disassembly of filaments into free octamers (Fig. S4C). However, 192 under saturating IMP concentrations, 2 mM GTP resulted in filaments that were often bent (Fig. S4D).

193
This should not be possible if GTP were saturating all sites; our structures above suggest bent filaments 194 must contain some extended protomers whose GTP-binding allosteric site 3 is disrupted. For the 195 ATP/GTP/IMP cryo-EM dataset we therefore used a much higher GTP concentration (20 mM), which 196 resulted in filaments with fully compressed segments (Fig. S4E). We explore this apparent difference in 197 GTP affinity in more detail below.

198
The compressed filament dataset resulted in three high-resolution reconstructions: the consensus 199 filament assembly interface, a fully compressed octameric filament segment, and a fully compressed

207
Compared to the uninhibited filament datasets, these GTP-saturated, +IMP filaments were shorter in 208 length. As a result, in addition to filament segments, we identified many filament ends: octamers in which 209 one tetramer does not have an assembled interface. The best-resolved structure of these filament ends 210 (3.3 Å) is conformationally similar to the filament segments, being fully compressed, with well-resolved 211 IMP-bound active sites in the inhibited conformation; however, without the filament assembly interface, 212 the N-terminus is only partially resolved (Figs. 4D, S5I-L).

213
From the free octamer dataset containing ATP/GTP, we used a similar symmetry-expansion and 214 classification strategy to that used for the filament datasets ( Fig S6A). This scheme confirmed that 215 virtually all the particles were symmetric, fully compressed free octamers in which both potential filament 216 assembly interfaces are unbound, with some poorly resolved minority classes of larger oligomers (  ATP/GTP-bound IMPDH octamer types are fully compressed, with ligand density in all regulatory sites 225 including the critical GTP/GDP-specific third site, which "staples" IMPDH octamers in the fully 226 compressed, inhibited conformation. But we observed two key structural differences that correlated with 227 assembly state: the conformation of the N-terminus, and the relative orientation of protomers within each 228 tetramer. The filament interface of the inhibited segments is unchanged from the uninhibited filaments 229 (Fig. 4F). However, at the free filament ends the N-terminus is only partially unresolved (a.a. 1-11), with 230 the resolvable portion rotated ~30° degrees compared to the bound interface, such that Val13 inserts into 231 the shallow hydrophobic pocket formed by A307, A308, and Lys311 of the neighboring protomer, very 232 similar to its position in the GTP-bound free octamer crystal structure (PDB 6i0o) (

236
We noticed a striking difference between the conformation of the catalytic core assembly in free versus 237 assembled states. Each protomer of the free octamer and at filament ends is tilted ~5° relative to the 238 four-fold symmetry axis, such that the tetramer becomes more "bowed" than the "flat" tetramers at 239 filament assembly interfaces (Fig. 4I, Video S1). The tetramers found in the GTP-bound free octamer

252
These three structures of inhibited IMPDH2 conformations provide a model by which filament assembly 253 is influenced by IMP and GTP through tetramer bowing and flattening ( Fig 4K). In the absence of 254 substrate, GTP induces both compression of filament segments and tetramer bowing, with the latter 255 resulting in filament disassembly. But when IMP is bound, the disordered active site loops become 256 ordered and rigid, buttressing intra-tetramer contacts as well as forming a pseudo-beta-barrel between 257 opposing tetramers in the GTP-bound state. These increased contacts work to resist tetramer bowing 258 and more readily sample the flat tetramer conformation, which promotes, and is stabilized by, the filament assembly interface. When IMP levels are low, GTP promotes filament disassembly, but high IMP levels 260 shift the equilibrium towards filament assembly.

IMPDH2 filaments resist GTP inhibition
262 Based on this model, during quiescence, when the salvage pathways supply ample GTP and IMP 263 production is downregulated, intracellular IMPDH2 will be in the fully compressed, fully inhibited, 264 ATP/GTP-bound free octamer state. Without increased IMP production, guanine depletion will result in 265 transient octamer extension and filament assembly, both of which will reverse as the resulting increase

272
To probe whether filament assembly influences the regulatory effects of GTP on IMPDH2 activity we 273 compared enzyme kinetics of the wildtype enzyme with the non-assembling mutant Y12A. We measured 274 the GTP dose-response of IMPDH2 pre-incubated with varying levels of ATP, and found that wild-type 275 enzyme is less sensitive to GTP inhibition, compared with Y12A (Figs. S8A-C). Depending on ATP 276 concentration, the apparent GTP IC50 of the wild-type was roughly two-fold lower than for Y12A. At 277 higher ATP levels, the apparent GTP IC50 of WT and the non-assembly mutant both increase. We 278 attribute this to competition between ATP and GTP at the first and second Bateman sites, suggesting 279 that independent of filament assembly, GTP inhibition of IMPDH2 is affected by ATP. Notably, the range 280 of GTP in which the substrate-saturated filaments resist GTP inhibition is within the upper range of in vivo 281 concentrations (Traut 1994). One function, then, of IMPDH2 filaments is to resist inhibition by GTP.

282
To correlate structural differences in IMPDH2 filaments as a function of GTP concentration, we collected 283 negative stain EM data directly from three reaction volumes used for the GTP inhibition experiments, 284 corresponding to uninhibited filaments (no GTP), ~10% inhibited filaments (2.5 mM GTP), and fully 285 inhibited filaments (20 mM GTP) (Fig. 5A). As in the cryo-EM datasets, we found that uninhibited filaments 286 were often extended, and fully inhibited filaments were universally compressed. However, we noted that 287 the partially inhibited filaments contained a more heterogeneous mix of extended, bent, and compressed 288 segments.

289
Next, we compared NAD + kinetics of uninhibited (0 mM GTP) and partially inhibited (2 mM GTP) filaments 290 with saturating IMP concentrations (Figs. 5B-C). As expected, in the absence of GTP WT and the non-291 assembling mutant have similar apparent Michelis-Menton kinetics (Anthony et al. 2017). Inhibition of is strongly inhibited. In the absence of saturating IMP, these same GTP levels result in complete octamer compression and filament disassembly, providing a possible mechanism by which filament assembly 295 alters GTP inhibition: by resisting the fully compressed state.

296
To better understand this partially inhibited state, we collected cryo-EM data of IMPDH2 with saturating 297 substrates, 0.5 mM ATP, and 2 mM GTP (Figs. S8D). Under these conditions we observed the non-298 assembly mutant Y12A was 84% inhibited but WT enzyme was only 16% inhibited. We resolved to high 299 resolution structures of not only the filament assembly interface and several distinct filament segments, 300 but surprisingly, two different free octamers; a canonical free octamer (a dimer of tetramers bound by 301 Bateman domains), and also a small class of free interfacial octamers (a dimer of tetramers bound by 302 the filament assembly interface) (Figs. S8E, S9A, Table 6).

303
The 3.1 Å resolution consensus assembly interface map is identical to the uninhibited and fully inhibited 304 consensus interfaces, including a well-resolved active site with strong density for both substrates; as with 305 the ATP/IMP/NAD+ structures these filaments are actively turning over and we have likely captured many 306 states, which we have modelled simply as IMP/NAD + (Figs S9B-C, S12A). As expected, the filament 307 segments exhibited a range of conformations (Fig. S9D). The best resolved of these were a fully

Filament-specific IMPDH2 conformations reduce GTP affinity and promote activity
From our different cryo-EM datasets combined, we have now determined structures of canonical IMPDH2 330 octamers bound to allosteric effectors and both substrates, in six distinct conformations (Fig. 5D). From 331 the uninhibited ATP/IMP/NAD + dataset we resolved a fully extended filament segment, and a bent 332 segment in which for each tetramer, 3 protomers were extended and 1 was compressed. From the 333 partially inhibited GTP/ATP/IMP/NAD + dataset, we resolved 2:2, 3:1, and fully compressed filament

384
We propose a model that describes the regulatory role of human IMPDH2 filament assembly, in which 385 assembly reduces feedback inhibition of enzyme activity in a substrate-dependant manner, increasing 386 flux through the de novo guanine nucleotide synthesis pathway in response to proliferative signalling. In 387 the absence of either IMP or guanine ligands, IMPDH2 is conformationally dynamic (Fig. 6A). Apo 388 IMPDH2 forms stable tetramers, which freely sample both the "bowed" and "flat" tetramer conformations, 389 with the latter resulting in release of the N-terminus and assembly into stable interfacial octamers.

390
Adenine nucleotides bind with high affinity to the Bateman domain, resulting in stable filaments in which 391 the active site loops remain unconstrained and the enzyme is active. The ATP concentration required to 392 induce filament assembly in vitro is far below the expected in vivo levels; we thus predict the apo state to 393 be rare in human cells (Traut 1994). Without bound guanine nucleotides, the Bateman domains of 394 individual protomers freely compress and extend (Fig. 6B). However, the flat tetramer conformation found 395 in assembled IMPDH2 filaments is resistant to full compression of octameric filament segments. Binding 396 of GTP to the Bateman domain stabilizes the compressed state, leading to lattice strain that is relieved 397 by disassembly of the filament interface and tetramer bowing. In this way high intracellular guanine levels 398 disassemble IMPDH2 filaments into stable free octamers whose activity is inhibited. Binding of IMP to 399 the active site stabilizes the flexible active site loops, and saturation with both IMP and GTP results in 400 compressed filaments in which the active sites of opposing tetramers interlock into a stable network. This 401 rigidifies the octameric filament segment, which now resists the lattice strain brought on by compression, 402 blocking GTP-induced filament disassembly. consistent with the states in which IMPDH filaments are observed in cells (Fig. 6C). Under homeostatic 405 conditions, the salvage pathways provide sufficient purine nucleotides, and IMP production is low. Under 406 these conditions IMPDH2 is bound to both adenine and guanine nucleotides, but not IMP, forming free 407 octamers rather than filaments. In vivo, filaments are typically not observed in quiescent cells; rather, as 408 in our model, their assembly is associated with increased intracellular IMP and decreased intracellular

776
Additionally, due to the flexibility of the filaments, and the tendency of filament ends to adhere to the air-777 water interface, many filaments were tilted out of plane, with neighboring segments overlapping in 778 projection.

779
To improve resolution, partial signal subtraction was performed at this stage, using a mask that left only 780 the central eight catalytic domains of the filament assembly interface, subtracting the poorly resolved 781 Bateman domains and neighboring segments, which served to improve resolution after subsequent auto-782 refinement. Per-particle defocus and per-micrograph astigmatism were then optimized using CTF 783 refinement, which improved resolution further. The resulting consensus refinements of the filament 784 assembly interfaces were well-resolved, however data on Bateman domain conformation was missing, 785 with these regions very poorly resolved when subtracted regions were restored to the reconstructions by 786 reversion to original non-subtracted particles (data not shown). To resolve the different Bateman domain 787 conformations, we applied particle symmetry expansion (D4 to C1) and classified particles without 788 additional alignment. Because at this stage the reconstructions were centered on the filament assembly 789 interface, each boxed "particle" contained elements of two different neighboring octamers. The potential 790 conformational space was reduced by applying a mask enclosing only one of these two octamers. By 791 hierarchical focused classification of the off-origin octamers we were able to classify multiple 792 conformations of the octameric filament segments, as well as incomplete segments and filament ends.

793
Symmetry expansion was also applied to the non-filament octamer datasets, with a mask including the 794 entire particle, which allowed classification of the most symmetric and well-resolved classes. To further 795 improve resolution of the varying symmetry-expanded segment classes, the reconstruction symmetry 796 origins were moved from the filament assembly interface to the canonical octamers by re-extraction with 797 re-centering.For each class, symmetry was then collapsed by removing redundant overlapping particles, octameric segments from the asymmetric symmetry-expanded classifications exhibited some apparent 800 symmetry, with fully extended or fully compressed octamers appearing D4 symmetric, and some bent 801 classes apparently D1 symmetric. We therefore applied these symmetries during subsequent refinement 802 and classification of these new octamer-centered classes. As before, signal subtraction of neighboring 803 filament segments improved resolution considerably. Additional rounds of CTF refinement and 3D 804 classification identified the best-resolved particles from each of these conformational classes.

838
Video S1. Comparison between the "flat" tetramer of assembled filaments and the "bowed"