Author response: Large-scale filament formation inhibits the activity of CTP synthetase

Article Figures and data Abstract eLife digest Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract CTP Synthetase (CtpS) is a universally conserved and essential metabolic enzyme. While many enzymes form small oligomers, CtpS forms large-scale filamentous structures of unknown function in prokaryotes and eukaryotes. By simultaneously monitoring CtpS polymerization and enzymatic activity, we show that polymerization inhibits activity, and CtpS's product, CTP, induces assembly. To understand how assembly inhibits activity, we used electron microscopy to define the structure of CtpS polymers. This structure suggests that polymerization sterically hinders a conformational change necessary for CtpS activity. Structure-guided mutagenesis and mathematical modeling further indicate that coupling activity to polymerization promotes cooperative catalytic regulation. This previously uncharacterized regulatory mechanism is important for cellular function since a mutant that disrupts CtpS polymerization disrupts E. coli growth and metabolic regulation without reducing CTP levels. We propose that regulation by large-scale polymerization enables ultrasensitive control of enzymatic activity while storing an enzyme subpopulation in a conformationally restricted form that is readily activatable. https://doi.org/10.7554/eLife.03638.001 eLife digest Enzymes are proteins that perform reactions that can convert one or more chemicals (the substrates) into others (the products). The rate at which an enzyme produces its product is often carefully regulated. Some molecules slow or stop an enzyme by binding to and blocking the site where its substrates normally bind: its ‘active site’. Other molecules can also bind to sites other than the active site, which can cause the enzyme to become either more or less active. Almost all living things have an enzyme called CTP synthetase that makes one of the building blocks that is used to build DNA and a similar molecule called RNA. This enzyme converts a molecule called uridine triphosphate (or UTP) into another called cytidine triphosphate (CTP): a reaction that is powered by breaking down molecules of adenosine triphosphate (ATP). The amount of CTP synthetase made by a cell is carefully controlled. The enzyme's activity is also regulated by the levels of UTP and CTP, and by another molecule (called GTP) that binds to a site outside of its active site. Four copies of the CTP synthetase protein must work together before this enzyme can turn UTP into CTP. The enzyme also forms much larger aggregates, or polymers; however, it is not clear what causes these polymers to form, or what they do in a cell. Barry et al. have now discovered that CTP synthetase is almost completely inactivated when these polymers are formed. Furthermore, CTP encourages the polymers to form, whilst UTP and ATP cause them to disassemble. Therefore, this enzyme is least active when there is excess product in the cell, and most active when its substrates are plentiful. By determining the three-dimensional structure of a CTP synthetase polymer, Barry et al. reveal that although CTP is bound to the enzymes, their active sites are still freely accessible. However, the enzymes in the polymer appear to be locked into a shape that makes them unable to carry out their function. When Barry et al. then mutated the enzyme so that it was unable to form polymers it was also no longer inactivated in the same way by CTP. Bacterial cells with only these mutant versions of CTP synthetase are unable to properly control their levels of CTP. This suggests that polymer formation is important for regulating this enzyme in response to a build up of its product. Further work is needed to see whether the regulation of CTP synthetase activity by forming polymers is specific to this enzyme or a widespread mechanism that is used to control other enzymes too. https://doi.org/10.7554/eLife.03638.002 Introduction Many enzymes form small-scale oligomers with well-defined subunit numbers, typically ranging from 2 to 12 subunits per oligomer. Recent studies suggest that some enzymes can also form large, higher-order polymers in which dozens to hundreds of subunits assemble into filaments (Barry and Gitai, 2011). For most of these structures, we lack an understanding of both the regulation and functional significance of their polymerization. To address these questions, we focused on the assembly of CTP synthetase (CtpS), an essential and universally conserved metabolic enzyme. CtpS forms large, micron-scale filaments in a wide variety of bacterial and eukaryotic species (Ingerson-Mahar et al., 2010; Liu, 2010; Noree et al., 2010), but the structure of these polymers, what triggers their formation, and the relationship between CtpS polymerization and enzymatic activity were unknown until now. Cellular CTP levels are subject to exquisitely tight homeostatic control, and CtpS is one of the most regulated enzymes in the cell. In both prokaryotes and eukaryotes, CtpS activity is regulated by allosteric control and feedback-inhibition of enzymatic activity, and CtpS levels are regulated by transcriptional and post-translational control (Long and Pardee, 1967; Levitzki and Koshland, 1972b; Yang et al., 1996; Meng et al., 2004). Cells in all kingdoms of life synthesize CTP using CtpS (Long and Pardee, 1967), and its essentiality makes CtpS an attractive chemotherapeutic and antiparasitic target (Williams et al., 1978; Hofer et al., 2001). The CtpS enzyme has two domains connected by an elongated linker: a glutaminase (GATase) domain that deaminates glutamine and a synthetase (ALase) domain that aminates UTP in an ATP-dependent manner to form CTP. CtpS has binding sites for substrates (glutamine, ATP, and UTP), product (CTP), and a proposed binding site for an allosteric modulator (GTP) (Levitzki and Koshland, 1972b). CtpS tetramerization is necessary for its catalytic activity and is controlled by nucleotide availability; ATP, UTP, or CTP can favor tetramer formation (Figure 1A; Levitzki and Koshland, 1972a; Anderson, 1983; Pappas et al., 1998; Endrizzi et al., 2004). Of critical regulatory importance, CtpS activity is also inhibited by CTP (Long and Pardee, 1967). Figure 1 with 7 supplements see all Download asset Open asset CtpS polymerization and enzymatic activity are inversely related. (A) A model of oligomeric regulation of CtpS. Tetramer formation from CtpS dimers is favored by a combination of enzyme concentration as well as nucleotide (substrates ATP and UTP or product CTP) and Mg2+ binding. (B) CtpS was incubated in activity buffer containing all substrates for CTP production. As the enzyme concentration increases, CtpS shows assembly by light scattering and the kcat value (Vobs/[CtpS]) decreases. Error bars = standard error (SE), n = 3–5. (C) Negative stain image of CtpS filaments assembled after CTP synthesis reaction. Smaller particles in the background resemble the X-shaped CtpS tetramer. A single filament is shown at bottom. (D) CtpS polymers formed in activity buffer were ultracentrifuged to pellet polymers. The pellet fraction was resuspended and CTP production was recorded. (E) CtpS assembly and activity were assayed after CtpS was first polymerized, followed by addition of saturating amounts of substrate after 600 s. https://doi.org/10.7554/eLife.03638.003 Here, we determine the function and mechanism of CtpS polymerization. We demonstrate that CtpS polymerization negatively regulates CtpS activity when its CTP product accumulates. We also present the structure of the CtpS polymers and the resulting implications for CtpS inhibition. We confirm the physiological significance of CtpS assembly by demonstrating that polymerization-mediated regulation is essential for the proper growth and metabolism of Escherichia coli. Together, these findings establish CtpS as a model for understanding enzymatic regulation by large-scale polymerization. Finally, we model how coupling CtpS activity to its large-scale assembly can enable cooperative regulation and discuss the implications of polymerization-based regulation for ultrasensitive metabolic control and cytoskeletal evolution. Results CtpS polymerization inhibits enzymatic activity Because CtpS filament formation is conserved between divergent organisms, we hypothesized that CtpS polymerization may regulate its conserved enzymatic function. We therefore designed a system to simultaneously monitor the assembly and activity of purified E. coli CtpS. We used a fluorometer to assay CtpS assembly by right-angle light scattering and CtpS activity by the specific absorbance of its CTP product. CtpS assembly and activity were assayed across a range of enzyme concentrations in activity buffer containing saturating amounts of substrates (UTP, ATP, and glutamine) as well as GTP and Mg2+ (referred to as ‘activity buffer’ throughout the text) (Figure 1B). CtpS protein was first pre-incubated in an incomplete activity buffer without glutamine to favor active tetramer formation. CTP production was then initiated by the addition of glutamine to form a complete activity buffer. The formation of well-ordered filaments was confirmed by negative stain electron microscopy (EM) (Figure 1C). Interestingly, at CtpS levels where robust changes in light scattering are observed (above approximately 1–2 μM), CtpS activity (determined by the rate of CTP production per enzyme) sharply decreases (Figure 1B, Figure 1—figure supplements 1 and 2). This abrupt transition in activity state supports the hypothesis that there is a threshold for polymerization and that polymerization is inhibitory. Noise and nonlinearity in the light scattering data make it difficult to determine an exact critical concentration value. However, based on correlation between light scattering and CTP production changes, we predict the assembly threshold of CtpS to be approximately 1–2 μM. The cellular level of CtpS protein in E. coli grown in minimal media was measured at 2.3 μM (Figure 1—figure supplement 3), indicating that the CtpS polymerization observed in vitro may be physiologically favorable. To determine if polymerization indeed inhibits CtpS activity, we assayed the activity of polymers purified by ultracentrifugation. The polymer-containing pellet was least enzymatically active immediately after centrifugation and CtpS activity increased as the polymers in the pellet disassembled (Figure 1D, Figure 1—figure supplements 4 and 5). CtpS polymers are thus inactive or much less than maximally active and polymerization is readily reversible. We directly demonstrated the reversibility of CtpS assembly and inactivation by first allowing CtpS to polymerize in activity buffer (with all substrates present) and then adding 1 mM UTP and ATP. Upon addition of these substrate nucleotides, we observed a sharp decrease in light scattering that corresponded to a sharp increase in CtpS activity. This transition was followed by a gradual increase in light scattering and corresponding decrease in activity back to the initial residual level (Figure 1E). Control experiments confirmed that the decrease in CtpS polymerization was not due to mechanical disruption by substrate addition (Figure 1—figure supplement 6). The correlation between the decrease in light scattering and the initiation of CTP production at the time of substrate addition indicates that substrate addition leads to rapid depolymerization and subsequent enzyme reactivation. Immediately after this point, we observed an increase in both CTP levels and polymerization. We therefore conclude that polymerized CtpS enzymes are inactive and must disassociate from the polymer to resume normal enzymatic activity. Despite the fact that polymerization occurs in a buffer containing substrates, polymerization only occurs with CTP production, suggesting that polymerization is triggered not by the initial substrates, but rather by the accumulation of CTP product. CtpS polymerization is induced by its product and repressed by its substrate In order to identify the factors that control CtpS inhibition by assembly, we first confirmed that none of the substrates alone induced polymerization (Figure 2—figure supplement 1). We then directly tested our hypothesis that CtpS's product, CTP, a known inhibitor of CtpS activity, stimulates CtpS polymerization. In the absence of substrates (UTP, ATP, and glutamine), incubation with CTP caused CtpS to polymerize (Figure 2A). The threshold concentration for robust changes in light scattering by CtpS with saturating CTP (1–2 μM CtpS; Figure 2—figure supplement 2) agrees with the threshold concentration in the presence of substrates (1–2 μM CtpS; Figure 1—figure supplement 1). This result suggests that CTP alone is sufficient to influence polymerization and that the substrates and any other products of the enzymatic reaction are not necessary. To confirm that CTP stimulates CtpS assembly, we used ultracentrifugation as an independent assembly assay. Titrating with increasing amounts of CTP caused an increase in the amount of CtpS found in the pellet with respect to the 0 mM CTP condition (Figure 2B, Figure 2—figure supplement 3). Figure 2 with 5 supplements see all Download asset Open asset CTP is sufficient and necessary to stimulate CtpS polymerization. (A) CtpS levels were titrated in buffer containing 1 mM CTP (with no substrates present). Polymerization was observed in the same range of protein concentrations as in activity buffer. Error bars = SE, n = 3. (B) CtpS was allowed to polymerize at different CTP concentrations (with no substrates present). The polymers were collected by ultracentrifugation and changes in CtpS pellet abundance were quantified by immunoblot. Error bars = SE, n = 2. (C) Purified CtpSE155K, which is defective in CTP binding, showed no obvious changes in light scattering during the normal conditions of wild-type polymer assembly in activity buffer. Initial light scattering values were normalized to 1 to place wild-type CtpS and CtpSE155K on the same scale. Error bars = SE, n = 3. (D) CtpS Filaments of wild-type and mutants by negative stain electron microscopy. There were very few filaments observed in the absence of CTP (top row). Upon the addition of nucleotide and MgCl2, filaments were only observed in the wild-type sample (first column). Micrographs were all taken at 55,000X magnification. (E) CtpS was incubated in the inhibitor DON and 1 mM CTP and allowed to polymerize. Addition of ATP and UTP depolymerized the sample. Polymers did not reform. https://doi.org/10.7554/eLife.03638.011 We further demonstrated that CTP binding is necessary for polymerization by showing that a CtpSE155K mutant defective for CTP-binding feedback inhibition (reviewed in Endrizzi et al., 2005) (Trudel et al., 1984; Ostrander et al., 1998) fails to polymerize under the same CTP-producing conditions in which wild-type enzyme polymerizes (Figure 2C). Furthermore, electron microscopy confirmed that, unlike wild-type CtpS, CtpSE155K does not polymerize in the presence of CTP (Figure 2D). Together, our data indicate that within our studied range of enzyme concentrations, CtpS's product, CTP, is both necessary and sufficient to induce CtpS polymerization. The CtpS crystal structure suggests that the enzyme's UTP and CTP binding sites partially overlap (Endrizzi et al., 2005), raising the question of whether CtpS assembly is controlled by the absolute level of CTP or the relative product/substrate levels. 6-Diazo-5-oxo-L-norleucine (DON) is a glutamine analog that covalently binds glutaminase active sites and irreversibly inactivates enzymatic activity (Chakraborty and Hurlbert, 1961). When added to activity buffer, DON abolishes both CTP production and CtpS polymerization (Figure 2—figure supplement 4). However, DON-treated CtpS can still polymerize when CTP is added to the solution (Figure 2E). Polymers formed in the presence of CTP and DON disassemble upon the addition of substrates but do not reform after substrate addition (Figure 2E), presumably because the DON-inhibited CtpS cannot produce additional CTP. DON treatment has no effect on CtpS polymerization when the enzyme is incubated with saturating CTP (Figure 2—figure supplements 1 and 5). These results suggest that competition between substrate (UTP) and product (CTP) binding controls the polymerization equilibrium of CtpS. The dependence of polymerization on CTP levels may explain why DON treatment abolishes in vivo CtpS assembly in some cellular contexts (Ingerson-Mahar et al., 2010) but not others (Chen et al., 2011). The structure of the CtpS polymer suggests a mechanism for enzymatic inhibition To better understand the mechanism of enzymatic inhibition by polymerization, we determined the structure of the CtpS filament by cryo-electron microscopy at 8.4 Å resolution (Figure 3—figure supplement 1). The repeating subunits of the filament are X-shaped CtpS tetramers (Figure 3A). The helical symmetry of the filament results in CtpS tetramers stacked atop one another with the arms of the adjacent Xs interdigitated. The 222 point group symmetry of the tetramer is maintained within the filament, resulting in overall twofold symmetry both along and perpendicular to the helical axis. A significant effect of this unusual symmetry is that, unlike many biological polymers, CtpS filaments are apolar. Figure 3 with 2 supplements see all Download asset Open asset Cryo-EM structure of CtpS filaments at 8.4 Å resolution. (A) A segment of the reconstructed filament, colored by helical subunit. (B) The E. coli CtpS crystal structure monomer fit into the cryo-EM density. Each domain was fit as a separate rigid body. (C) Novel filament assembly contacts between the linker domains. (D) Novel assembly contacts between the GATase domains. https://doi.org/10.7554/eLife.03638.017 To create an atomic model of the CtpS filament, we fit a monomer of the E. coli CtpS crystal structure into the cryo-EM structure as three rigid bodies (ALase domain, GATase domain, and the linker region) (Figure 3B). There is a slight rotation between the GATase and ALase domains, similar to the variation seen across crystal structures of full length CtpS (Figure 3—figure supplement 2A). There is a strong density for CTP bound at the inhibitory site, and no density in the predicted UTP active site (Figure 3—figure supplement 2B), confirming the biochemical data that CTP binding favors assembly. Weaker density is also observed for ADP, but there is no density in the predicted GTP allosteric regulatory site (Figure 3—figure supplement 2C,D). There is a minor rearrangement of the tetramerization interface in the filament relative to the crystal structure that results in a compression of the tetramer by about 3 Å along the length of the filament axis (Figure 4). Figure 4 Download asset Open asset Rearrangement of the CtpS tetramerization interface within the filament. (A) Superposition of the E. coli crystallographic tetramer (gray) with the atomic model from the cryo-EM structure (color), shows a rearrangement of the tetramerization contacts, primarily a compression of the tetramer along the filament axis. (B) Rearrangements of the tetramerization contacts shift the relative positions of helices near bound CTP (gray: crystal structure; color cryo-EM structure). https://doi.org/10.7554/eLife.03638.020 The cryo-EM structure of the CtpS filament offers insight into the mechanism of enzymatic regulation. All of the enzyme active sites are solvent accessible, suggesting that UTP, ATP, and glutamine can freely diffuse into the filament (Figure 5A). This observation rules out occlusion of active sites as a regulatory mechanism. An alternative mechanism of CtpS inhibition is blocking the transfer of ammonia between the GATase and ALase active sites, which are separated by ∼25 Å. The detailed mechanism of ammonia transfer is unknown, but likely involves a conformational rearrangement in the vicinity of a putative channel that connects the two domains (Endrizzi et al., 2004; Goto et al., 2004). One prediction is that a conformational change, induced by UTP and ATP binding, rotates the GATase domain toward the ALase domain to create a shorter channel between the active sites (Goto et al., 2004). Such a large-scale rotation would be unattainable in the steric environment of the filament, as it would lead to clashing of the moving GATase domain with an adjacent CtpS tetramer (Figure 5B,C). Regardless of the specific changes involved, quaternary constraints imposed by the filament structure likely provide the mechanism for inhibition of the synthesis reaction. Figure 5 Download asset Open asset Implications of the CtpS filament structure for the mechanism of enzyme inhibition. (A) The binding sites for ATP, CTP, and glutamine are all solvent accessible in the filament, suggesting that they are freely exchangeable in the filament form. (B) The approximate direction of the putative rotation of the glutaminase domain toward the amidoligase domain (arrow), which is predicted to create a shorter channel for ammonia diffusion. (C) In the filament structure, such a conformational change would be sterically hindered by contacts with adjacent filament subunits. https://doi.org/10.7554/eLife.03638.021 A CtpS polymerization interface mutant disrupts feedback regulation To validate the filament structure and its mechanistic implications, we generated structure-guided mutants in the CtpS polymerization interface. Two discrete segments constitute the novel filament assembly contacts: the linker region α-helix 274–284, and the short α-helix of the GATase domain (Figure the exact at the assembly are not well relative to the of CtpS, both sites many or for polymerization across species (Figure supplement 1). We previously demonstrated that in E. an the filamentous of CtpS assayed by (Ingerson-Mahar et al., As an initial for CtpS assembly, we therefore in the linker region α-helix and and into (Figure All polymerization interface mutants a diffuse rather than filaments (Figure Figure with 1 supplement see all Download asset Open asset form a polymerization interface. (A) The positions of the polymerization mutants in the model of the filament assembly interface. (B) mutants were into an and upon in E. coli. = 3 forms filaments while mutant show diffuse The of filamentous does not the that the polymerization interface mutants form small filaments that cannot be by light microscopy. to determine if the diffuse in vivo a polymerization we purified one of the linker region and its polymerization by light scattering and did not polymerize in activity buffer, and no filaments be by (Figure Figure supplement confirming that cannot properly polymerize. We the slight increase in light scattering with increasing concentration of to the increase in protein Figure 7 with 4 supplements see all Download asset Open asset polymerization and cause a growth (A) The CTP production activity of titrated levels of a small decrease in enzymatic activity as enzyme concentration when to wild-type Error bars = n = (B) Purified does not polymerize in the presence of CTP. For both wild-type and CtpS, there were very few filaments observed in the absence of CTP (top row). Upon the addition of nucleotide and MgCl2, filaments were only observed in the wild-type sample (first column). (C) wild-type and cells in defective growth when to cells with wild-type CtpS. were grown and into to the defective growth of mutant E. coli in minimal mutants defective Error bars = SE, n = We determined the of the polymerization interface on CtpS activity. the protein concentration CTP production of to the protein (Figure To determine if the polymerization of was due to large-scale assembly or CTP production, we used to its polymerization in the presence of saturating CTP levels. did not polymerize in the presence of levels of CTP (Figure We thus conclude that polymerization of its effect on activity. was in its activity at enzyme concentrations, a much concentration at which kcat is one of its due to polymerization (the to wild-type CtpS = μM = Furthermore, the concentration dependence of kcat was less than with of its activity at the enzyme concentration tested (Figure This was in to wild-type CtpS, activity to of its at enzyme concentrations, activity than while at enzyme concentrations activity was than that of One for the decrease in activity as a function of enzyme concentration is that produces CTP, which at CtpS concentrations can and CtpS activity, resulting in a slight activity However, this mutant the in CtpS activity by large-scale assembly into As predicted from the to polymerize also leads to bind CTP less with a value than the wild-type enzyme μM μM at Figure supplement 2). These data are thus with the model that CtpS is negatively regulated in two CTP inhibits UTP binding, and large-scale assembly sterically hinders a conformational change for CtpS activity. The between and activity suggest that large-scale assembly rapid and inhibition of enzymatic activity. The polymerization interface mutant disrupts E. coli growth and metabolism To determine the of on cell we wild-type CtpS with at its in E. coli. This defective growth to in (Figure and minimal media (Figure supplement 3). time was while the time was in confirmed that was at similar levels to wild-type CtpS (Figure supplement 4). One for the growth is that not produce CTP to robust However, CTP as measured by are not in the (Figure supplement 1). In CTP levels are in the mutant than in cells Because CTP levels are in these likely does not growth due to CTP production. the CTP levels and the observation that growth at the hypothesis that the mutant is defective in regulating CTP levels when to changes in the cellular wild-type CtpS with also levels of other and their or (Figure Figure supplement 1). For the amount of the was 2.3 in the with the that is and CTP production at the of its Together, these data indicate that the CtpS polymerization interface does not CtpS or CTP. we that E. coli growth by nucleotide metabolism in a manner with CtpS by a negative regulatory mechanism. These data are with the observation that at the cellular concentration of CtpS, is more active than the wild-type enzyme. Figure with 3 supplements see all Download asset Open asset of polymerization interface disrupts CTP in (A) of a

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.