The structure of SpoT reveals evolutionary tuning of enzymatic output through constraint of the conformational landscape

31 Stringent factors orchestrate bacterial cell reprogramming through increasing the level of the 32 alarmones (p)ppGpp. In Betaand Gammaproteobacteria, SpoT hydrolyses (p)ppGpp to 33 counteract the synthetase activity of RelA. However, structural information about how SpoT 34 controls the levels of (p)ppGpp is missing. Here we present the crystal structure of the 35 hydrolase-only SpoT from Acinetobacter baumannii and uncover the mechanism of intra36 molecular regulation of “long”-RSH factors. In contrast to ribosome-associated Rel/RelA that 37 adopt an elongated structure, SpoT assumes a compact τ-shaped structure in which the 38 regulatory domains wrap around a Core subdomain that controls the conformational state of the 39 enzyme. The Core is key to the specialisation of long-RSHs towards either synthesis or 40 hydrolysis: while the short and structured Core of SpoT stabilises the τ-state priming the HD 41 domain for (p)ppGpp hydrolysis, the longer, more dynamic Core of RelA destabilises the τ42 state precluding (p)ppGpp hydrolysis and priming RelA for synthesis. 43 44


Introduction
conformation. In these complexes the regulatory C-terminal domain region (CTD: TGS, HEL, 82 ZFD and RRM domains) is highly structured, while the N-terminal catalytic region (NTD: HD 83 and SYNTH domains) and the interdomain linker regions are highly dynamic and unresolved 84 in some structures [18][19][20][21] . Off the ribosome, our structural understanding of long RSHs relies on 85 the structures of isolated NTDs of several Rel representatives [21][22][23][24] . While the physiological role 86 of SpoT as a key virulence and stress tolerance factor is well established 25,26 , structural insights 87 into SpoT are lacking altogether. This limits our ability to interpret the physiological and 88 microbiological studies on the molecular level. Obtaining full-length structures of 89 Rel/RelA/SpoT is essential for understating how the auto-regulation signal transferred from the 90 CTD to the NTD. 91 The structural and biochemical data presented here provide the long-missing structural 92 insight into the molecular mechanism of SpoT. We show that A. baumannii SpoT (SpoTAb) is 93 indeed, a monofunctional (p)ppGpp hydrolase and uncover how its CTD is an allosteric 94 activator of the HD hydrolase function. The structures of the full-length HD-active SpoTAb 95 complexed with the ppGpp substrate reveal a compact monomeric conformation in which all 96 the regulatory domains wrap around a Core subdomain that connects the pseudoSYNTH and 97 TGS domains. The Core is one of the intrinsically disordered regions (IDR) present in Rel and 98 RelA when in the active synthetase state. In SpoTAb, Core and TGS cooperate to align and 99 activate the hydrolase domain active site while translating allosteric feedback from the other 100 regulatory domains to modulate the HD output. Finally, we propose a unifying conceptual 101 framework that rationalises the relative balance between HD vs SYNTH activities of long RSHs 102 Rel, RelA and SpoT, fine-tuned through the entropic force produced by intrinsically disordered 103 regions that function as conformational gatekeepers of the enzyme. 104

A. baumannii SpoTAb is a monofunctional hydrolase long RSH 106
Lack of conservation of active site residues critical for SYNTH activity suggest that 107 Moraxallaceae SpoT enzymes have -like RelA -undergone subfunctionalisation to become 108 monofunctional long RSHs ( Fig. 1a and b) of SpoTAb ( Fig. 1c and Supplemental Data), demonstrating that SpoTAb is HD-active in the 120 surrogate E. coli host. 121 Next, we used our dual plasmid co-expression system to probe the (p)ppGpp synthetase 122 activity of SpoT RSHs. ppGpp 0 E. coli is auxotrophic for eleven amino acids, and (p)ppGpp 123 synthetase activity of SpoTEc is essential for growth of DrelA E. coli on minimal medium 12 . 124 Unlike the SYNTH-active SpoTEc, SpoTAb failed to promote the growth of ppGpp 0 E. coli on 125 M9 minimal medium (Fig. 1d), confirming that SpoTAb is SYNTH-inactive. Taken together, 126 these results demonstrate that SpoTAb is a specialised monofunctional long RSH that lacks the 127 ability to synthesise (p)ppGpp. 128 129

Full-length SpoTAb has a compact mushroom-like τ-shaped structure 130
To gain insight into the molecular workings of SpoT, we solved an X-ray structure of full-131 length catalytically-active SpoTAb in a ppGpp-bound state at 2.9Å resolution. The structure 132 revealed a multi-domain architecture strikingly different to that observed earlier for ribosome-133 bound long RSHs Rel and RelA 18-21 (Fig. 2a-c and Supplementary Table 1). The HD, 134 SYNTH, TGS, HEL, ZFD and RRM domains of SpoTAb form a mushroom-like tau (τ)-shaped 135 quaternary structure (Fig. 2a-c). In this arrangement, pseudo-SYNTH, TGS, HEL, ZFD and 136 RRM domains all lie in a single plane and form a compact disc-like structure that forms the 137 "cap" of the "mushroom" (Fig. 2b). A helix-turn-helix sub-domain (residues 334 to 379) that 138 provides the transition between the NTD and CTD regions, lies at the "Core" of the "cap" and 139 seemingly mediates interactions among all domains of the enzyme. Such an arrangement 140 suggests that the Core -which is disordered in Rel/RelA structures -stabilises the disc-like 141 "cap" of SpoT (Fig. 2c). Moreover, the Core provides the HD domain with a physical link to 142 each domain of SpoTAb. Finally, the HD protrudes from the plane of the "cap" in the opposite 143 direction of the C-terminal RRM domain, forming the "stem" of the protein structure ( Fig. 2b-144   c). 145 The τ-shaped structure of SpoTAb suggests a possible structural mechanism for the auto-146 inhibition of SYNTH activity by the regulatory CTD both in Rel 27,28 and RelA 29,30 . While the 147 SYNTH and TGS domains are sequestered in the "cap", the HD hydrolase stands out 148 unconfined and primed for (p)ppGpp hydrolysis. The TGS domain, which in the case of amino 149 acid starvation sensors Rel and RelA specifically engages the deacylated tRNA CCA-3′ end at 150 the A site 19-21,31 , in the case of SpoTAb is partially trapped between the HD, HEL and ZFD 151 domains. While we do detect a mild inhibitory effect of tRNA on SpoTAb hydrolysis activity, 152 the effect is insensitive to tRNA aminoacylation status, i.e. non-specific (Fig. 2d). This is in 153 contrast to the HD activity of bifunctional E. coli SpoT (SpoTEc), which was specifically 154 inhibited by deacylated, but not aminoacylated tRNA 32 . 155 Our structure reveals that the sites from the ZFD and RRM domains that mediate rRNA 156 recognition in Rel/RelA 18-21,31 are held in by the Core subdomain, suggesting that in the τ-157 shaped conformation the hydrolytically active (HD ON ) SpoTAb is incompatible with ribosome 158 binding. In good agreement with this structural prediction, while the ribosome strongly 159 suppresses the HD activity of Bacillus subtilis Rel (RelBs) 28 , the addition of E. coli 70S has no 160 effect on the hydrolysis activity of SpoTAb (Fig. 2d) Fig. 1a) has posed an experimental challenge for structural studies 18-21 . The 168 molecular function of these flexible regions, unresolved in the structures, is unknown. 169 Comparison between the well-structured SpoTAb in τ-state and partially unstructured ribosome-170 bound RelA/Rel suggests that the unfolding of Core and HEL domains constitutes part of the 171 conformational switch that positions TGS, ZFD and RRM domains to stimulate the synthesis 172 activity of Rel/RelA upon recruitment to the ribosome (Supplementary Fig. 1b). 173 The length of these disordered or flexible regions is on average shorter in 174 monofunctional SpoT and much longer in the monofunctional RelA. Bifunctional Rels have 175 interdomain IDRs of sizes between both monofunctional enzymes (Supplementary Fig. 1c). 176 The α6-α7 loop of the HD domain of SpoT[Hs] in particular is a third of the size of that of 177 RelA, which, in turn, is twice longer than that of bifunctional Rel (Supplementary Fig. 1c). 178 The same pattern is observed for the other two IDRs: the Core subdomain and the region 179 connecting HEL and ZFD domains. This is consistent with the significantly lower disordered 180 propensity of the Core of SpoTAb compared to RelAAb (Supplementary Fig. 1d- It was shown earlier that both Rel and RelA are prone to dimerization via the CTD, which 187 would potentially serve to regulate their enzymatic activity 21,33-35 . This idea is a subject of 188 debate, with both genetic 30 and mass photometry 28 experiments suggesting that the 189 dimerization is unlikely to take place at physiologically relevant concentrations. Therefore, we 190 used small-angle X-ray scattering (SAXS) coupled to size exclusion chromatography (SEC) to 191 probe the conformation and oligomeric state of SpoTAb in solution ( Fig. 2e-f ). 192 The SAXS data revealed that in solution SpoTAb has an oblate shape compatible with 193 the structure determined by X-ray. Both SAXS and SEC consistently support the monomeric 194 nature of SpoTAb, even at concentrations as high as 8 mg/mL. Both the molecular weight of ≈90 195 kDa by SEC as well the estimates of Mw of ≈85 kDa and Rg of 34.9Å by SAXS (Fig. 2e-f) 196 agree with the 80 kDa theoretical molecular weight of monomeric SpoTAb. Furthermore, the 197 analysis of the normalised Kratky plot derived from the scattering curve lends further support 198 for a compact monomeric structure of SpoTAb in solution (Fig. 2f), and the ab initio envelope 199 calculated from the experimental SAXS data (Fig. 2g) is compatible with the τ-shaped structure 200 of SpoTAb determined by X-ray. Collectively these results demonstrate that in solution the 201 monomeric SpoTAb adopts a conformation that is very similar to the τ-shaped conformation 202 observed in the crystal with the HD domain protruding from the disc-shaped enzyme. 203

204
The enzymatically-inactive pseudo-SYNTH of SpoTAb is a regulatory domain 205 In the monofunctional stringent factor RelA, the enzymatically inactive pseudo-HD domain has 206 evolved into a regulatory domain controlling catalysis via an intra-NTD allosteric regulatory 207 mechanism 36,37 . This is also the case with the specialisation of SpoTAb as a monofunctional 208 hydrolase where the pseudo-SYNTH domain has evolved into a strictly regulatory/structural 209 domain. Superposition of the SYNTH domain from RelTt onto the pseudo-SYNTH domain of 210 SpoTAb reveals extensive reorganisation of the vestigial catalytic domain in SpoTAb, consistent 211 with differential conservation patterns in the G-loop and the ATP recognition motif 212 ( Supplementary Fig. 2a). These involve the residues that coordinate adenosine and guanosine 213 (R249 to N241, R277 to E267 and Y329 to N304) and the majority of phosphate-coordinating 214 groups. Crucially, the catalytic residues D272 and Q347 are substituted for S263 and T321, 215 respectively. These substitutions essentially impede the deprotonation and activation of the 3′-216 OH of GD(T)P, and Mg 2+ binding, precluding the nucleophilic attack on the β-phosphate of 217 ATP. We directly probed GDP binding by SpoTAb NTD and RelAAb NTD by ITC. As expected, 218 while SpoTAb does not bind GDP, RelAAb binds GDP with an affinity of 62 μM, which is similar 219 to our earlier estimates for RelAEc NTD and RelBs NTD 28,36 (Supplementary Fig. 2b-c). To directly validate the lack of pppGpp-mediated regulation in SpoTAb, we characterised 234 the interaction between pppGpp and SpoTAb NTD by ITC. As expected, SpoTAb NTD does not bind 235 pppGpp allosterically (Supplementary Fig. 2d-e). As observed earlier for RelTt NTD 23 , the hydrolase active site of SpoTAb displays a dipolar 255 charge distribution with a highly basic half mediating the stabilization of the 5′-and 3′-256 polyphosphate groups of the substrate and the other highly acidic half mediating the 3′-257 pyrophosphate hydrolysis ( Fig. 3a-b). Closer inspection of the complex reveals the crucial role 258 of Y51 and the 82 ED 83 active site motifs as they work together with the Mn 2+ cofactor to 259 coordinate and stabilise a network of water molecules near the sugar-phosphate moiety during 260 hydrolysis ( Fig. 3b-c). Indeed, substitutions of Y51, E82, D83 or N147 render SpoTAb HD-261 inactive in our enzymatic assays ( Fig. 3d). At the positively charged side active site the 5′-262 polyphosphate is loosely coordinated and exposed to the bulk solvent. By contrast K140 and 263 R144 hold the 3′-pyrophosphate in place during hydrolysis and Ala substitutions of these 264 residues decrease the activity of the enzyme between 5-and 10-fold suggesting these are key 265 residues that orient the scissile bond. 266 267 Mn 2+ ion organizes the HD active site of SpoTAb 268 The essential role of the divalent manganese ion Mn 2+ in (p)ppGpp pyrophosphate hydrolysis 269 is well documented for both Rel 22,28,38,39 and SpoTEc 40 . Our isothermal titration calorimetry 270 (ITC) measurements demonstrate that unliganded, metal-free SpoTAb NTD binds Mn 2+ with a KD 271 of 35.3 μM (Fig. 3e). Furthermore, while metal-free full-length SpoTAb is completely HD-272 inactive, the HD activity is readily restored upon addition of Mn 2+ (Fig. 3f). 273 To directly reveal the structural role of Mn 2+ we determined the X-ray structure of 274 SpoTAb NTD in the metal-free state ( Fig. 3g and Supplementary Table 1). Comparison with the 275 structure of the SpoTAb-ppGpp complex provides a structural explanation for the essentiality of 276 Mn 2+ for catalysis: in addition to its role in hydrolysis, by connecting α3, α4 and α8, Mn 2+ 277 coordination brings together the two halves of the HD domain and provides structural support 278 to the active site ( Fig. 3g-h). While the overall topology of the SpoTAb HD domain is similar to 279 that of Mn 2+ -liganded RelTt NTD 23 , the removal of the metal ion has a profound effect on the 280 local conformation of the active site of SpoTAb NTD . The catalytic 78 HD 79 and 82 ED 83 motifs are 281 largely misaligned, loops S110-Y117 and A153-K158 that are involved in the 3′-and 5′-282 phosphate coordination are disordered, and the guanine-coordinating loop T44-Y51 assumes a 283 conformation incompatible with the base coordination ( Fig. 3h). Importantly, all of these While SpoTAb variants lacking the RRM or the RRM and ZFD domains retained wild-306 type ability to sustain the bacterial growth grow -i.e. could efficiently degrade (p)ppGpp 307 synthesised by RelA -further C-terminal truncations compromised the in vivo HD 308 functionality, as evidenced from pronounced growth defects (Fig. 4a). Biochemical assays are 309 in agreement with the in vivo data (Fig. 4b). Truncation of the RRM and ZFD decreases the 310 HD activity 5-fold. Further deletion of the TGS-HEL domains leads to a dramatic 42-fold 311 decrease in activity. Truncations beyond the TGS compromised the activity by 70-fold or more 312 and isolated HD domain was nearly inactive. Collectively, our results suggest that the CTD 313 region functions as an allosteric activator of the hydrolase function of SpoTAb. Next, we set out 314 to dissect the molecular mechanism of the CTD-mediated NTD control and assign the 315 The Core domain is a linchpin that controls the τ-state 318 Both the overall structural arrangement of SpoTAb and our sequential domain truncation 319 experiments ( Fig. 4a-b) suggest that the Core-mediated allosteric crosstalk between the HD 320 and rest of the domains of the enzyme is essential for enzyme's functionality. To specifically 321 assess the role of the individual interdomain interactions we introduced single point 322 substitutions at each of the interfaces of the Core with regulatory CTD domains (Fig. 4c) and 323 measured the hydrolase activity of the SpoTAb variants (Fig. 4d). An intact HD:Core:TGS 324 interface -the structure involved in scaffolding the HD active site -is crucial for HD activity, 325 as the Y375G substitution at the HD:Core:TGS resulted in a 5-fold decrease in activity 326 compared to the wild type. While substitutions at the ZFD (L373G / D374G) and RRM 327 (A351K) domain interfaces also resulted in a pronounced defect (19 and 3-fold decrease, 328 respectively), perturbations at the Core:pseudo-SYNTH domain interface (A348R) had only a 329 minor effect on hydrolysis. Finally, decoupling the contacts of HD from the τ-cap via the L356D 330 substitution, located at the interface between Core domain and α6-α7 motif of HD 23 , has a 331 dramatic 35-fold decrease in HD activity, suggestive of an allosteric signal transduction path 332 between the cap and stem regions of the enzyme. When we monitored the thermodynamic 333 stability of these Core variants of SpoTAb we observed they all have lower stability and loss of 334 structure compared to the wild type (Supplementary Fig. 3a-e). This suggests that an increase 335 in the configurational entropy of the Core has a global effect in the dynamics and compactness 336 of the enzyme. The existence of an allosteric relay mediating a CTD-dependent activation of 337 HD via the Core is further supported by the consistent decrease in hydrolysis associated with 338 the aforementioned C-terminal truncations that affect the feedback of the Core to the HD (Fig.  339 4a-b), as well as by the observation that the deletion of domains HEL and TGS results in a 50-340 fold decrease in activity despite the presence of the other regulatory domains (pseudo-SYNTH, 341 ZFD and RRM) (Fig. 4d). 342 We next used SEC-SAXS to directly probe the role of each contact at the interface of 343 the Core with the different domains of SpoTAb on stabilisation of the τ-sate. The L356D 344 Fig. 4c and Supplementary Table 2) results in the segregation of 345 the population into two conformational states with major differences in RG (radius of gyration) 346 and particle dimensions (DMAX). In SpoTAb L356D one state is the compact τ-shape observed in 347 the crystal structure (Fig. 4e), while the other state is more relaxed (RG = 41Å, DMAX = 130Å) 348 with dimensions reminiscent of that of the less compacted Rel and RelA -but not quite as 349 elongated as in the ribosome-bound state (Fig. 4f). In this relaxed state the Core and HEL 350 domains appear to have transitioned to a more disordered state that is consistent with the 351 conformational states of these regions in the fully elongated state observed in Rel/RelA ( with that of the relaxed state of SpoTAb L356D , whereas RelBs is populated by both the relaxed and 356 τ-states ( Fig. 4i-k and Supplementary Table 2). 357 Collectively, our results suggest that the Core domain functions as an allosteric relay 358 that conveys signals from the CTD to the HD. At the structural level the composition of the 359 Core is the key to the conformational state of the enzyme as defined by the three major 360 conformations observed in SpoT, Rel and RelA (Fig. 4l) The α6-α7 element plays a crucial role in the allosteric regulation of the opposing activities of 372 bifunctional RelTt 23 . In RelTt, α6-α7 of projects away from the HD catalytic centre to 373 accommodate the 3′ and 5′ polyphosphate groups as well as allowing the catalytic 82 ED 83 motif 374 to get in position, close to the 3′ phosphates, priming the enzyme for hydrolysis. In SpoTAb the 375 outward-pointing conformation of α6-α7 is further stabilised by the N-terminal region of the 376 TGS and the Core domains which function as a clamp to keep α6-α7 in the HD-compatible 377 position, with the HEL domain providing an additional support via the Core (Fig. 5a). The 378 dramatic drop in the activity of the SpoTAb variant lacking the TGS and HEL domains (Fig. 4d)  379 substantiates the functional importance of this stabilising effect. 380 At the HD:TGS interface the β-hairpin of the TGS -the very element which is involved 381 in tRNA recognition in Rel 21,28,42 and RelA 19,20,31 -is buried and stacking directly the α6-α7 382 element via a small hydrophobic interface formed by W382, Y384, L390 and the R124-E392 383 salt bridge (Fig. 5a). Disrupting this interface with the E379K / W382K substitutions 384 (SpoTAb E379K/W382K ) led to a 17-fold decrease in the hydrolase activity of the enzyme (Fig. 4d) 385 suggesting that the HD:TGS interface constitutes as an important allosteric signal transduction 386 pathway. This scaffolding role is complemented by the Core that wraps tightly around α7 thus 387 preventing the recoil of α6-α7 away of the HD active site, which, as we observed earlier in RelTt 388 23 , induces the opening of the NTD . Indeed, substitutions at the Core:α6-α7 interface such as 389 the aforementioned Y375G also affected hydrolysis (Fig. 4c-d). Interestingly, despite the 390 strongly attenuated HD activity of SpoTAb E379K / W382K , SAXS showed SpoTAb E379K / W382K 391 remains in the τ-state (RG = 35Å, DMAX = 104Å), suggesting an allosteric communication via 392 the HD:Core:TGS axis ( Fig. 5b and Supplementary Table 2). 393 Given that SpoTAb is SYNTH-inactive and is not specifically regulated by tRNA or 394 ribosomes (Fig. 2d), it is not surprising that TGS residues involved in tRNA recognition -such 395 as the crucial His residue involved in the recognition of the 3′ CCA end by Rel 21,28,42 and RelA 396 19,31 (S407 in SpoTAb) -are lost in the monofunctional SpoTAb (but are present in bifunctional 397 SpoTEc 1 ). Moreover, the τ-state is sterically incompatible with the potential recognition of 398 tRNA by TGS due to sequestration the β-hairpin and α-helical elements. All these observations 399 suggest that in SpoTAb the TGS has been repurposed as a scaffolding domain crucial to sustain 400 hydrolysis, with both TGS and Core cooperating to lock the α6-α7 in place, stabilising the HD 401 active site. This contrasts with its crucial function of recognition of uncharged tRNA in 402 Rel/RelA 19-21,28,31 . 403 404

The ZFD and RRM domains finetune the hydrolytic activity of SpoTAb 405
With ZFD and RRM positioned close to the disc-shaped cap and connecting with the pseudo-406 SYNTH domain, the resulting inter-domain interfaces are likely to play a role in the stability 407 the τ-state as well as to allosterically control of HD via the HD:pseudo-SYNTH relay. In 408 agreement with this hypothesis, disruptive substitutions at the Core:HD (L356D), Core:pseudo-409 SYNTH:RRM (A351K) and Core:ZFD (L373G / D374G) that decreased the stability of the τ-410 state (Supplementary Fig. 3b-e) also decreased the HD activity of the enzyme by 35-, 3-and 411 22-fold, respectively (Fig. 4d). Therefore, we reasoned that substitutions stabilising the 412 Core:pseudo-SYNTH:RRM and Core:ZFD interfaces would, conversely, trigger an allosteric 413 activation of hydrolysis. 414 To probe this hypothesis, we introduced substitutions that would increase the contacts 415 of RRM with pseudo-SYNTH via hydrogen bonds, I637D / R641D, and the Core with the ZDF, 416 D374R (Fig. 5c). Denaturation experiments showed SpoTAb D374R and SpoTAb I637D / R641D have 417 higher stability and compactness than the WT (Supplementary Fig 4a-c) and SAXS 418 measurements on SpoTAb I637D / R641D confirmed this variant retained the τ-state ( Fig. 5d and  419 Supplementary Table 2). As expected, the HD turnover of both enzyme variants increased (by 420 2.1-and 1.6-fold, respectively, Fig. 4d), and both behave like wild-type SpoTAb in vivo (Fig.  421   5e). 422 Collectively, our results establish that HD activity is coupled to the stability of the τ-423 state, with the Core domain working as an allosteric transducer that allows the catalytic HD to 424 communicate with all the regulatory domains. Substitutions or interactions that stabilise the τ-425 state increase hydrolysis, whereas τ-state-destabilising substitutions lower the HD activity. 426 427

An intact τ-shaped SpoTAb is required for virulence of A. baumannii 428
Functional (p)ppGpp-mediated signalling plays a crucial role in antibiotic tolerance and 429 virulence of A. baumannii 15,43 . We used the wax moth G. mellonella larvae infection model to 430 assess the functionality of mutant spoTAb variants in supporting virulence of A. baumannii 431 AB5075 (Fig. 5f). Only the strain with wild type-like virulence was the one expressing SpoTAb 432 D374R variant with a HD activity slightly higher than that of the WT SpoT. The spoTAb D374R 433 strain has rapidly killed 100% of the larvae within the first two days whereas 60% of the larvae 434 survived 6 days of infection with the (p)ppGpp 0 ∆relA strain. Infection with A. baumannii 435 expressing the ΔRRM-truncated enzyme SpoTAb 1-614 resulted in 25% survival rate of larvae 436 after 6 days. Notably, the RRM-truncated SpoTAb 1-614 had 6-fold lower hydrolase activity as 437 compared to wild type (Fig. 4d), and the strain displays no growth defects when grown on LB 438 plates (Fig. 4a). The defect in virulence becomes more prominent with truncations beyond the 439 TGS domain: SpoTAb 1-454 and SpoTAb 1-339 . The strong decrease in HD activity associated with 440 the A. baumannii strains expressing these SpoT variants results in 100% survival of the infected 441 larvae (Fig. 5f). Collectively our results suggest that while a basal level of the HD hydrolase 442 activity is sufficient to sustain bacterial growth in non-stressed conditions (e.g. on a plate and 443 in liquid culture), the pathogen requires fully functional CTD-and Core-mediated control of 444 SpoTAb to tune the HD activity and efficiently establish a successful infection. We propose a unifying scheme that rationalises the evolution of the enzymatic output 464 of long RSHs through fine-tuning of the conformational equilibrium of the τ, relaxed and 465 ribosome-bound states of these enzyme (Fig. 6) and relaxed states as part of the conformational spectrum of these enzymes (Fig. 6a-b). While 468 the τ-state primes Rel/SpoT for efficient (p)ppGpp hydrolysis, the more elongated relaxed state 469 sets the enzyme for low-efficiency (p)ppGpp synthesis. To fully activate its SYNTH activity, 470 the enzyme needs to be further stimulated by starved ribosomes to attain the highly elongated 471 ribosome-bound state; this transition is possible for amino acid starvation sensor Rel[HS], but 472 not for SpoT, which is not under allosteric control by starved ribosomes and ppGpp 36 (Fig. 6a).  (Fig. 6c-d). Compared to 476 SpoT[HS], in SpoT[Hs] the equilibrium is further shifted towards the HD-active τ-state 477 required for hydrolysis (Fig. 6c). In contrast, in RelA[hS] the τ-state becomes inaccessible, and 478 the enzyme is primed for ribosomal recruitment upon which it is stabilised in the highly 479 elongated ribosome-bound SYNTH-active state (Fig. 6d). 480 Expansion/contraction of the disordered regions is the likely molecular driver of the 481 fine-tuning of the enzymatic output in long RSHs through the restriction of the conformational 482 space (Supplementary Fig. 1c-e)     (l) Cartoon representation of experimentally observed conformational states as well as particle dimensions of long RSH enzymes. The Core domain of SpoT transduces the allosteric signal from the regulatory CTD and pseudo-SNTH to the enzymatic HD domain. (a, b) Cartoon representation of the interactions stabilising the α6-α7 motif of the HD active site (A). While the Core wraps around α7, the TGS β-hairpin forms a small hydrophobic patch that stabilises α6. These interactions preclude the movement of α6-α7 and maintain SpoTAb in a constitutive hydrolase-primed state. Key interface residues are shown as sticks and labelled.  The enzymatic output of subfunctionalised RelA and SpoT RSH enzymes is evolutionarily tuned through constrains of the conformational landscape. (a) Control of the enzymatic output of the ancestral bifunctional Rel [HS]. Upon amino acid starvation Rel is recruited to starved ribosomal complexes. The ribosome-bound Rel assumes an extended conformation in which the auto-inhibitory effect of the CTD region on the SYNTH activity is released. The full activation of SYNTH activity is achieved upon binding of (p)ppGpp to an 23 allosteric site within the NTD and release of the SYNTH inhibition by the HD domain. Conversely, off the ribosome the enzyme assumes the τ-state. In this conformation locking of the α6-α7 motif by the CTD organises the HD active site residues to promote the HD activity. This, in turn, strongly inhibits the SYNTH activity via inter-NTD regulation. The full activation of either SYNTH or HD requires allosteric signalling from CTD to NTD enzymatic domains. (b, c) Evolution of SpoT as a predominantly dedicated hydrolase involved the loss of the allosteric control of the NTD by (p)ppGpp as well as by the ribosome. In bifunctional SpoT[HS] present in the majority of Gamma-and Betaproteobacteria, while the equilibrium is strongly shifted towards the HD-active τ-state, the enzyme is capable of ine cient (p)ppGpp synthesis in the relaxed state (B). Subfunctionalisation of SpoT in Moraxellaceae has resulted in the monofunctional hydrolase SpoT[Hs], which naturally populates only the compact τ-state and is SYNTH-inactive. (d) Subfunctionalisation of Gamma-and Betaproteobacterial RelA[hS] constitutes the other extreme case of evolutionary restriction of the conformational dynamics of the ancestral Rel [HS]. While losing its HD activity, RelA retains all the allosteric regulatory elements of Rel. Being a dedicated (p)ppGpp synthetase enzyme, off the ribosome RelA does not assume the τ-state.
Instead, it predominantly populates the functionally frustrated resting state equivalent to the relaxed state of Rel, primed to assume the elongated ribosome-associated state triggered by the 70S ribosome, uncharged tRNA and alarmones during stringency. Red circles represent inhibited catalytic centres, green circles represent fully activated catalytic centres, and dashed green circles represent idling catalytic centres.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.