Allosteric pyruvate kinase-based “logic gate” synergistically senses energy and sugar levels in Mycobacterium tuberculosis

Zhong, Wenhe; Cui, Liang; Goh, Boon Chong; Cai, Qixu; Ho, Peiying; Chionh, Yok Hian; Yuan, Meng; Sahili, Abbas El; Fothergill-Gilmore, Linda A.; Walkinshaw, Malcolm D.; Lescar, Julien; Dedon, Peter C.

doi:10.1038/s41467-017-02086-y

Download PDF

Article
Open access
Published: 07 December 2017

Allosteric pyruvate kinase-based “logic gate” synergistically senses energy and sugar levels in Mycobacterium tuberculosis

Wenhe Zhong^1,2,
Liang Cui¹,
Boon Chong Goh^1,2,
Qixu Cai³,
Peiying Ho¹,
Yok Hian Chionh ORCID: orcid.org/0000-0002-4909-8382¹^nAff7,
Meng Yuan⁴^nAff8,
Abbas El Sahili^2,5,
Linda A. Fothergill-Gilmore⁴,
Malcolm D. Walkinshaw ORCID: orcid.org/0000-0001-5955-9325⁴,
Julien Lescar^2,5 &
…
Peter C. Dedon^1,6

Nature Communications volume 8, Article number: 1986 (2017) Cite this article

5385 Accesses
44 Citations
15 Altmetric
Metrics details

Subjects

Abstract

Pyruvate kinase (PYK) is an essential glycolytic enzyme that controls glycolytic flux and is critical for ATP production in all organisms, with tight regulation by multiple metabolites. Yet the allosteric mechanisms governing PYK activity in bacterial pathogens are poorly understood. Here we report biochemical, structural and metabolomic evidence that Mycobacterium tuberculosis (Mtb) PYK uses AMP and glucose-6-phosphate (G6P) as synergistic allosteric activators that function as a molecular “OR logic gate” to tightly regulate energy and glucose metabolism. G6P was found to bind to a previously unknown site adjacent to the canonical site for AMP. Kinetic data and structural network analysis further show that AMP and G6P work synergistically as allosteric activators. Importantly, metabolome profiling in the Mtb surrogate, Mycobacterium bovis BCG, reveals significant changes in AMP and G6P levels during nutrient deprivation, which provides insights into how a PYK OR gate would function during the stress of Mtb infection.

Dihydroxyacetone phosphate signals glucose availability to mTORC1

Article 27 July 2020

Jose M. Orozco, Patrycja A. Krawczyk, … David M. Sabatini

A rotary mechanism for allostery in bacterial hybrid malic enzymes

Article Open access 23 February 2021

Christopher John Harding, Ian Thomas Cadby, … Andrew Lee Lovering

Near-equilibrium glycolysis supports metabolic homeostasis and energy yield

Article 23 September 2019

Junyoung O. Park, Lukas B. Tanner, … Joshua D. Rabinowitz

Introduction

Mycobacterium tuberculosis (Mtb) is among the deadliest infectious diseases on a global scale, killing more than one-million people annually¹, with emerging antimicrobial drug resistance posing serious challenges to existing diagnosis and treatment programs¹. Although generally considered aerobic, Mtb has successfully adapted to the hypoxic and carbon-poor environment in human macrophages by evolving flexible carbon metabolism and co-catabolism to defend against the stresses posed by the human immune system^2,3,4. In particular, increasing evidence suggests that the metabolic flexibility of central carbon metabolism (CCM: glycolysis, gluconeogenesis, pentose phosphate pathway and TCA pathway) is critical in Mtb physiology and pathogenicity^5,6,7. This is illustrated by the rapid regulation of glycolytic activity in response to changes in ATP levels, which explains the enhanced efficacy of drug combinations that target the electron transport chain⁸. Here we describe a unique feature of the Mtb glycolytic enzyme, pyruvate kinase (PYK), which provides insights into its pivotal role in Mtb CCM⁹ and the metabolic flexibility of Mtb.

PYK (EC 2.7.1.40) forms a tetramer (Fig. 1a) that catalyzes transfer of phosphate from phosphoenolpyruvate (PEP) to ADP to form pyruvate and ATP. Given its central role in controlling glycolytic flux and ATP generation, PYK has been exploited as a drug target in bacterial pathogens¹⁰, parasites¹¹ and cancer¹², and was recently identified as a potential target of the antimalarial drug artemisinin¹³. The activity of most PYKs is tightly controlled by physiological modulators, with the notable exception of the constitutive activity of mammalian muscle isoenzyme M1. However, in spite of many decades of study, the complex allosteric mechanisms governing PYK activity in different organisms have eluded definition. For example, while fructose 2,6-bisphosphate (F26BP) allosterically activates trypanosome PYKs¹⁴, the upstream glycolytic intermediate, fructose 1,6-bisphosphate (F16BP), is the most widely recognised allosteric activator of PYKs in many bacteria^{15, 16}, in yeast¹⁷ and in mammals¹⁸. In general, PYKs from higher organisms have a single essential modulator, with the exception of F16BP and amino-acid regulation of human M2PYK in cancer-cell proliferation^19,20,21. This stands in contrast to the many bacterial PYKs that use ‘non-canonical’ effectors such as AMP and the sugar monophosphates glucose 6-phosphate (G6P) and ribose 5-phosphate (R5P) for allosteric regulation (Supplementary Table 1), including PYKs from important human pathogens such as Mtb⁹, Streptococcus mutans ²², Staphylococcus aureus ²³ and Salmonella typhimurium ¹⁶ (a sequence alignment of selected bacterial PYKs is shown in Supplementary Fig. 1 and pairwise identities in Supplementary Table 2).

One of the reasons that the allosteric mechanisms regulating bacterial PYKs have remained elusive for decades is the lack of structural information about the binding of physiological effectors to identify allosteric sites. Here we report biochemical, structural and computational modelling studies of Mtb PYK that reveal a ‘rock-shape-lock’ allosteric mechanism regulated by the synergistically acting activators AMP and G6P. Stress-induced metabolomic changes in the Mtb surrogate, Mycobacterium bovis BCG, point to AMP and G6P as molecular input signals that position MtbPYK as a unique molecular OR ‘logic gate’ to sense changes in energy and sugar levels during Mtb infection.

Results

Effectors AMP and G6P activate MtbPYK synergistically

We first performed detailed kinetic studies of purified MtbPYK to evaluate AMP and G6P as allosteric activators and to test whether they share a single binding site (Table 1). MtbPYK purified from an Escherichia coli expression system was more active than that reported by Noy et al.⁹ under similar assay conditions, with a k _cat value of 183 ± 1 s⁻¹ versus 63 s⁻¹, respectively (Supplementary Table 1), and it showed hyperbolic kinetics with respect to its substrate ADP, with a K _m value of 0.47 ± 0.02 mM. For the substrate PEP, the enzyme displayed sigmoidal kinetics in the absence of effector, with a S _0.5 value of 0.41 ± 0.01 mM and a Hill coefficient (h) of 1.82 ± 0.05, indicating positive cooperativity. The affinity to substrate PEP was estimated by its S _0.5 value in the presence of saturating ADP. Therefore, both AMP and G6P enhanced the affinity of PEP by ~2-fold and decreased the cooperativity with respect to PEP (h = ~1.2). Inhibition by the product ATP (Supplementary Fig. 2) reduced PEP affinity by 5-fold (S _0.5 =2.20 ± 0.08 mM) and increased the cooperativity to PEP (h = ~2.6). ATP inhibition was reversed by adding the allosteric activators AMP and G6P.

Table 1 Kinetic properties of MtbPYK

Full size table

The concentrations of AMP and G6P were then varied to investigate their interplay in enzyme activation (Table 1). The K _a0.5 value of AMP fell up to 3.2-fold (~64 to ~20 µM) as the concentration of G6P was increased. In a similar manner, the apparent affinity of G6P increased up to 3.5-fold (~147 to ~42 µM) with increasing AMP. This positive cooperativity between AMP and G6P suggests two distinct binding sites instead of competitive binding to the same effector site. In addition, the enzyme catalytic efficiency, which is estimated by k _cat/S _0.5 value (Table 1), is also enhanced by the synergy of AMP and G6P. In contrast to previous computational studies in which effectors were docked at a single canonical allosteric site of bacterial PYKs²⁴, our kinetic results demonstrate that AMP and G6P act in synergy to allosterically activate MtbPYK and suggest two cooperative allosteric sites. The latter was borne out in structural studies.

A rigid-body rocking motion in allosteric transition

More than 40 years ago, Waygood et al.²⁵ demonstrated for the first time that AMP and G6P were the activators of PYK using the E. coli enzyme as the model. However, AMP- or G6P-bound PYK structures have been hitherto unavailable to elucidate the corresponding effector sites and allosteric mechanisms. Here, we have successfully determined a series of six high-resolution crystal structures of PYK from M. tuberculosis in both inactive T-state and active R-state, and with or without AMP and G6P (Table 2).

Table 2 Data collection and refinement statistics

Full size table

MtbPYK adopts a tetrameric architecture formed by identical subunits with three domains (A, B and C domains) (Fig. 1a), similar to typical PYK structures. The catalytic site is located in the cleft between the A and B domains (lid domain), while the C domain harbours a canonical AMP allosteric site. The allosteric site binding for G6P (sugar monophosphate site) is adjacent to the AMP site and lies between the A and C domains (Fig. 1b). It is noteworthy that the amino-acid binding site of human M2PYK is also formed by the A and C domains but is located on the opposite pole of the molecule relative to the canonical allosteric site (Fig. 1b).

The allosteric mechanism was analysed at the level of quaternary protein structure by superposition of the inactive T-state tetramer structure (PDB: 5WRP) onto the fully ligated R-state tetramer structure (MtbPYK-OX/AMP/G6P; PDB: 5WSB) excluding the mobile B domains. This gave an RMS fit of 3.2 Å for all C-α atoms. The superposition analysis suggests that each subunit of the MtbPYK tetramer simultaneously undergoes a 9° AC-core (A and C domains) rigid-body rotation concomitant with the T- to R-state transition (Fig. 2a, b; Supplementary Table 3), which is consistent with the ‘rocking motion’ mechanism observed in trypanosomatid PYKs^{14, 26, 27} and human M2PYK¹⁹. We have also compared the rigid-body rotation angles from T-state to different ligated R states of the MtbPYK tetramer (Supplementary Table 3), and observed similar motion angles (<1° variation). Our results provide the detailed structural evidence that bacterial PYKs possess a concerted ‘rocking motion’ allosteric mechanism (Supplementary Movie 1), despite having a distinct repertoire of allosteric effectors.

The interplay between AMP and C-terminal tail loops

We next explored the determinants of this concerted rigid-body rotations of all four chains during the allosteric interconversion of T- and R-states in the MtbPYK tetramer (Fig. 2a, b). Unlike trypanosomatid PYKs that use additional salt bridges across the C–C interface to stabilise the R-state tetramer structure^{14, 26, 27}, the motions of the AC-cores in MtbPYK require the disruption of C–C interface interactions between the AMP loop and the C-terminal tail loop (Thr455…Asp470), as well as the hydrogen bonds (H-bonds) formed between the two adjacent Cα4 structures (Thr422…Asp423) (Fig. 2c, d). The electron density of the AMP loop is poorly identified in the R-state MtbPYK-OX structure (PDB: 5WS8) indicating high flexibility in the absence of effector binding.

The relocation of the AMP loop leaves a free space at the C–C interface and thus allows the reorganisation of the C-terminal tail loop (Fig. 2c). This tail loop flips ~180° to approach the C–C interface and forms H-bonds with the Cβ4 strand from the same protein chain (Asp471…His463/His465) while the hydrophobic residue Val472 (the C-terminal portion of the tail loop) forms a hydrophobic interaction with the side-chain of Ile426 in helix Cα4 (Fig. 2d). This ‘in-and-out’ interplay between the AMP loop and the C-terminal tail loop, in response to the transition between T- and R-states, favours the formation of the R conformation by abolishing interface H-bonds and generating new interactions within each protomer (Supplementary Movie 2). The loss of interface interactions is also in agreement with the reduction in the C–C interface area in the R-state (Supplementary Table 4).

The thermal stability of MtbPYK in all possible ligated states was determined in the presence of a wide variety of ligands by performing fluorescence-based thermal shift assays (Supplementary Fig. 3). When one or more allosteric ligands were mixed with the enzyme in the presence of substrate PEP or oxalate, the thermal vibration decreased (ΔT _m = 1–10 °C), suggesting that the stability of the R conformation is enhanced by a network of interactions between allosteric ligands and each protein subunit, which overcomes the reduction in C–C interface interactions. Kinetic evidence of enzyme activation by AMP and G6P correlates well with the enhanced rigidity of the R-state PYK in the presence of AMP and G6P. The physical AMP and G6P binding sites were established next, with the discovery of the G6P-binding site.

AMP binds at the canonical allosteric site

While AMP binds tightly at the canonical allosteric site of PYK (Fig. 3) that recruits bisphosphate effectors such as F16BP^{17, 18} and F26BP¹⁴ in many organisms, the MtbPYK AMP binding site has several unique features. In MtbPYK, the AMP phospho group is H-bonded to residues Thr374, Gln375, Ser376, Thr379 and AMP-loop residue Gly457, and forms a salt bridge with the side-chain of Arg351 in helix Cα1. By comparison with the AMP-free T-state structure (PDB: 5WRP), the effector loop (residues 451–458) moves further from the dimer–dimer (C–C) interface to wrap around AMP and form H-bonds with it. The adenine-binding site for AMP is relatively hydrophobic and is composed of residues Phe373, Trp398 and Met425. These residues form stacking interactions with the adenine ring of AMP (Supplementary Fig. 4). Interestingly, the binding of Trp398 to AMP requires reorientation of the side chain of this residue to close up the AMP pocket. The fact that Trp398 is not highly conserved in other AMP-activated PYKs (Supplementary Fig. 1) suggests a unique regulatory function in MtbPYK. Hydrophobic residues Phe373 and Met425 are relatively conserved among PYKs that use AMP as activator. Residue Met425 in helix Cα4 is normally replaced by arginine in F16BP- or F26BP-activated PYKs^{14, 17, 18} (Supplementary Fig. 5a), whereby a positively charged side chain is essential to form a salt bridge with the negatively-charged 1′- or 2′-phospho group to hold the effector in place. Moreover, Gln375 in MtbPYK is substituted by a positively charged lysine residue in human M2PYK which interacts with the 1′-phospho group of F16BP via a salt bridge¹⁸ (Supplementary Fig. 5a). Therefore, the hydrophobicity of the adenine-binding site in the AMP pocket is strictly required for recognition. By contrast, in F16BP- or F26BP-activated PYKs, the effector site needs to provide more positive charges (Arg, Lys substitutions) to recognise and lock both negatively charged phospho groups in place.

G6P binds at an allosteric site adjacent to canonical site

The kinetics-based prediction of a unique G6P-binding site was borne out in the structural analyses. The MtbPYK structures in complex with G6P reveal an allosteric site that is distinct from the canonical AMP-binding allosteric site (Fig. 3). The G6P site is located at the boundary between the A and C domains (Fig. 1), and is composed of helices Aα6, Cα1, Cα2 and the G6P-loop (Fig. 3). Two positively-charged residues Arg382 and Arg385 in Cα2 lock the G6P phospho group in place via salt bridges together with three H-bonds provided by three residues in the G6P-loop (His345, Arg348, Thr349). Interestingly, the side-chain of His345 in the G6P loop adopts multiple positions within a tetramer in the absence of G6P binding, but is locked in one position by H-bonding to G6P. The sugar group of G6P makes a H-bond with the main-chain of Asn268 in Aα6. Water molecules further stabilise G6P. In contrast, the corresponding site in non-G6P-regulated PYKs (i.e., trypanosome PYK, human M2PYK) has more negatively-charged residues that could prevent G6P binding (Supplementary Fig. 5b).

Structurally, Cα1 bridges the AMP and G6P sites, where Arg351 in Cα1 forms an ionic pair with AMP and Thr349 (next to Cα1) interacts with the G6P phospho group by H-bonding. The presence of this prominent structure network of AMP…Cα1…G6P correlates well with the kinetic synergy between AMP and G6P. These results suggest the existence of two physical allosteric pathways linking the AMP and G6P binding sites with the active site ~40 Å away, which was established next.

Dual allosteric pathways with synergistic cooperation

To explore the physical basis for this dual allostery, we performed a molecular dynamics (MD) simulation on the tetramer of MtbPYK with bound OX/AMP/G6P in order to identify potential allosteric pathways between the catalytic site (OX site) and the two allosteric sites (AMP and G6P sites). The allosteric pathways between oxalate and AMP/G6P were then extracted from the MD simulation using the Weighted Implementation of Suboptimal Pathways (WISP)²⁸. Dominant allosteric pathways connecting oxalate and AMP (OX-AMP), as well as connecting oxalate and G6P (OX-G6P), are shown in Fig. 4a, b. An allosteric pathway was considered dominant when it consisted of a set of residues with high frequency of occurrence in the histogram shown in Fig. 4c. The two allosteric pathways propagate through three identical residues en route to the catalytic site, namely Ala237, Ala217 and Lys218, which suggests that synergistic cooperation exists, perhaps at the level of individual MtbPYK monomers. Further community analysis revealed that AMP and G6P-binding sites belong to the same group of highly correlated residues (Supplementary Fig. 9), suggesting binding of either activator could allosterically regulate the catalytic binding site. This was explored next.

A synergistic allostery beyond rock-and-lock

Studies of PYKs from trypanosomatids¹⁴ and humans¹⁹ have established an allosteric ‘rock-and-lock’ regulation mechanism involving rigid-body rotations of each subunit in the tetramer. We have demonstrated this in MtbPYK (Fig. 2). However, the conformational flexibility within each subunit of the tetramer has not yet been reported, except for the flexible effector loop and the mobile B domain²⁷ (Supplementary Fig. 6; Supplementary Table 5). To study this type of structural plasticity and to probe the mobile secondary structural elements (α helix, β strand, loops), we superposed monomers (instead of tetramers) of all available R-state MtbPYK structures onto the T-state monomer structure. The mobile B domain and AMP loop were excluded from the superposition. The distances between Cα-carbons following superposition ranged from 0.7 to 1.1 Å (Supplementary Table 3), which indicates only a subtle change in structure. We then further analysed the superposition data by plotting a heat-map using the Cα distance of each pair of superposed atoms (Supplementary Fig. 7). The heat-map clearly reveals correlated local structural rearrangements (shown in red) in response to the concerted rigid-body motion and the AMP/G6P synergistic binding. This indicates the ‘plastic’ nature of a MtbPYK monomer, as illustrated in Fig. 5 and Supplementary Movie 3.

AMP/G6P kinetic synergy results from a synergistic allostery

A more granular analysis of the allosteric site reshaping upon AMP and G6P binding revealed a unique mechanism underlying the AMP/G6P kinetic synergy. Cα1 and Cα4 are essential α helices at the two allosteric sites of MtbPYK that regulate the binding of G6P and AMP, respectively (Fig. 3). Both exhibit obvious movements during the T- and R-state transition (Fig. 5; Supplementary Fig. 7). This structural plasticity of allosteric sites in response to effector binding has not been reported in any other PYK. We then further analysed the movements of superposed structures and found that Cα1 and Cα4 both rotated and shifted by ~6° from the T-state (yellow) to the fully ligated R-state (cyan) (Fig. 5). In contrast, other non-fully ligated MtbPYK R-state structures exhibit less obvious but similar Cα1/Cα4 motion during the T- and R-state interconversion. Briefly, as shown in Fig. 5, only a subtle movement of Cα1 is observed between the T- (yellow) and oxalate-bound R-state (MtbPYK-OX; orange), with Cα4 remaining static. AMP binding (purple) triggers nearly maximum Cα4 movement (~6°) but modest Cα1 movement. By contrast, G6P binding (magenta) causes Cα1 to shift nearly maximally (~6°), but with only a subtle movement of Cα4. To further confirm the structural plasticity of MtbPYK allosteric sites, we soaked AMP and G6P into MtbPYK-OX crystals resulting in nearly identical allosteric sites (Cα RMS fit: 0.17 Å) compared with that of the fully ligated R-state structure determined from the crystal using a co-crystallization method. Both fully ligated R-state structures, obtained from distinct crystallization methods, display the same degree of Cα1/Cα4 movement (~6°) during the T- and R-state transition, further demonstrating the dramatic plasticity of the allosteric-site structures regulated by effectors.

Thus, the displacements of helices Cα1 and Cα4 resulting in allosteric-site reshaping are not substantially regulated by rigid-body rotation, but instead are correlated with the presence of AMP and G6P. This unique ‘synergistic allostery’ suggests that the binding of either activator is able to initiate the movements of both Cα1 and Cα4 to a certain extent and consequently reshape the allosteric sites to favour binding of the other activator. The synergy of AMP and G6P shown in structural rearrangements agrees with the findings in enzyme kinetics where the presence of one activator promotes the activation efficiency of the other (Table 1). Furthermore, the binding of activators shows a remarkable increase in the thermal stability of R-state MtbPYK with a ΔT _m value of 4–9 °C (Fig. 6d; Supplementary Fig. 3), suggesting that allosteric-site reshaping enhances the stability of the R-state conformation.

It is noteworthy that the helix bundle Aα6′–Aα6 that is located at the A–A interface exhibits a ~6° rotation around the pivot near Asn268 during T- and R-state inter-conversion. No further movement is observed between R-state structures of different ligated states, suggesting that the conformational rearrangement of Aα6′–Aα6 only correlates with the concerted rocking motions, and is required in any R-state structure. The movement of Aα6′-Aα6 plays a role in forming an ‘active’ catalytic site for the binding of substrate and its cofactor (Fig. 5), in agreement with our previous findings in Trypanosoma brucei PYK, where we showed the reorientations of side chains of active-site residues for substrate binding²⁷. In addition, another pair of A–A interface helices Aα7′ and Aα7, which display modest movement shown in the heat-map, form a H-bond across the A–A interface with the adjacent helix pair Aα6′ and Aα6 via the side-chain rotation of Arg290 (Supplementary Fig. 8). One consequence of the T and R transition is therefore to reorganise the positions of Aα6′–Aα6 and Aα7′–Aα7 to transform the catalytic site between its ‘active’ and ‘inactive’ conformations.

A molecular logic gate senses metabolic changes in vivo

The preceding biochemical and structural results are all consistent with the idea that MtbPYK functions as an ‘OR logic gate’ to tightly regulate metabolism based on integration of AMP and G6P levels during the large metabolic shifts that occur during the stress of infection. The ‘rock-and-lock’ model for trypanosome PYK allostery²⁶ and the ‘dock-rock-lock’ model for human M2PYK allostery¹⁹ are based largely on structural evidence. Until now, a lack of bacterial PYK structures with native effectors bound precluded any comparisons between prokaryotic and eukaryotic enzymes. The MtbPYK structures presented here reveal atomic-level details of a unique AMP/G6P synergism controlling PYK allostery. We propose a three-step ‘rock-shape-lock’ model to describe the synergistic allostery of MtbPYK: (1) Subunits of the MtbPYK tetramer rock 9° simultaneously from the T-state to the R-state (Supplementary Fig. 8a, b) in response to substrate binding, movements that are coupled with interface-loop interplay and catalytic-site formation via the movement of helices Aα6′ and Aα6. (2) AMP and G6P binding reshapes the allosteric sites by local conformational changes (Supplementary Fig. 8c–e), with the two allosteric pathways working independently or in positive cooperativity. (3) A network of interactions within and between allosteric sites is then established, which ‘double locks’ the PYK tetramer in the most catalytically efficient active state termed the ‘synergistically activated R-state’ (Supplementary Fig. 8c).

This dynamic model is consistent with the idea of an OR logic gate that enables MtbPYK to sense changes in both AMP and G6P levels and appropriately modulate enzyme activity during stress-induced metabolic shifts. This logic gate regulatory activity of PYK can be understood in the context of the metabolic shift that occurs during mycobacterial nutrient deprivation mimicking the tubercular granuloma. M. bovis BCG is a well-established experimental surrogate of Mtb for studying metabolism. In particular, the PYK protein sequence is identical for the two. Here we used LC-MS to quantify changes in the level of glycolytic intermediates in M. bovis BCG subjected to nutrient deprivation. As shown in Fig. 6a, starvation caused significant reductions (up to 30-fold) of G6P compared with logarithmic growth, which should lead to downregulation of PYK. In contrast, AMP levels increased (up to 4-fold) at 4 and 10 days of starvation (S4 and S10 in Fig. 6). This should upregulate PYK activity early in nutrient deprivation. Additionally, the PYK substrates ADP and PEP accumulate at S4 and S10, where PEP is a toxic glycolytic intermediate that inhibits isocitrate dehydrogenase in the TCA cycle⁹. These shifts in PYK activity as a function of AMP and G6P levels thus satisfy the criteria for an OR logic gate (inset table in Fig. 6b) that enables PYK to sense the stress-induced reduction of G6P and increase of AMP (Fig. 6a), and to continually keep PYK primed to facilitate potential carbon co-catabolism and prevent further accumulation of toxic metabolites during the stress-induced metabolic shifts (output in Fig. 6b).

Discussion

This flexibility of CCM is essential for mycobacterial physiology and pathogenicity, particularly under the stresses of infection². Allosteric mechanisms of key CCM enzymes enable bacteria to efficiently sense the changes in metabolite levels and react immediately by allosterically regulating mechanisms to maintain homoeostasis and defence against environmental threats⁵. The results from our biochemical, structural and metabolic studies uncover a unique ‘rock-shape-lock’ allostery, which endows PYK as a molecular OR logic gate to simultaneously sense the changes of AMP and G6P levels in a synergistic manner (Fig. 6a, b). This OR gate uses AMP (the low-energy-state signal) and G6P (essential for M. tuberculosis pathogenesis^{29, 30}) as molecular inputs to promote PYK activity (output), where either input at a certain concentration is able to fully activate the enzyme activity for the needs of the bacterium (Fig. 6c). Our findings also shed light on the importance of enzyme evolution providing M. tuberculosis with metabolic flexibility and the ability to adapt to challenging environmental changes. The conservation of key residues in AMP and G6P binding suggests that the above allosteric model is likely to apply to other bacterial PYKs. Importantly, our results provide both structural foundation and allosteric mechanism for designing MtbPYK inhibitors that could potentially provide a fast-killing therapy in combination with respiratory ATPase inhibitors like bedaquiline⁸ to completely block ATP production.

Methods

Cloning and protein production

A series of gene codon-optimisation procedures from Bio Basic Inc. were applied to improve the efficiency of soluble expression of MtbPYK in E. coli (Supplementary Table 6). The codon-optimised gene was synthesised and cloned into a pYUB28b-TEV vector by Bio Basic Inc. The vector pYUB28b-TEV is modified from pYUB28b (Addgene #37277) that was provided by Dr. Ghader Bashiri from The University of Auckland. The verified construct (pYUB28b-TEV_MtbPYK) allows the expression of N-terminal 6xhistidine-tagged MtbPYK with subsequent tag removal by TEV protease digestion.

MtbPYK was overexpressed in E. coli BL21(DE3) cells (Novagen, Merck Millipore) and purified as described previously²⁷ with some modifications. Briefly, 10 ml of overnight culture was inoculated in 500 ml of 2xTY medium with 50 μg ml⁻¹ hygromycin B and grown to OD₆₀₀ = 0.6–0.8 at 37 °C. MtbPYK protein overexpression was induced by 0.5 mM IPTG after a 30 min incubation at 4 °C. The cells were further cultured for another 24 h at 16 °C before harvesting.

Cell pellets were lysed in ice-cold buffer (50 mM triethanolamine-HCl (TEA) pH 7.2, 300 mM KCl, 20 mM imidazole, 10% glycerol) using a Constant Systems Cell Disruptor (Panda) (set to 850 bar). The soluble fraction was obtained by centrifugation at 39,191 × g for 45 min. The supernatant was filtered through a 0.22 μm syringe filter then applied onto an ÄKTA system fitted with 1 ml HisTrap HP IMAC column (GE Healthcare) (prepacked with Ni Sepharose). The column was maintained at a constant flow rate of 1.0 ml min⁻¹ throughout the whole-purification process. Following loading of the clarified lysate, the column was washed with 20 column volumes of cell-lysis buffer. The target enzyme MtbPYK was eluted from the column by applying 100% elution buffer (50 mM TEA pH 7.2, 300 mM KCl, 500 mM imidazole, 10% glycerol) and digested with TEV protease at 4 °C for 12 h to remove the 6xhistidine tag. The digested sample was dialysed against lysis buffer and loaded onto a HisTrap HP IMAC column (pre-equilibrated in cell-lysis buffer). The untagged enzyme MtbPYK was collected from the flow-through fractions and concentrated to 5 ml using a Vivaspin column (GE Healthcare, 100 kDa cutoff).

Concentrated proteins were then loaded onto a HiLoad 16/600 Superdex 200 prep grade gel-filtration column with a flow rate of 1.0 ml min⁻¹ in a buffer containing 20 mM TEA pH 7.2, 50 mM KCl, 10 mM MgCl₂, 20% glycerol. MtbPYK proteins eluted in a peak at ~63 ml and were further analysed on SDS–PAGE. Pure proteins (>95% purity) were concentrated to at least 20 mg ml⁻¹ and were stored at −80 °C. The protein concentration was determined by measuring the absorbance at 280 nm with an extinction coefficient for MtbPYK of 25,440 M⁻¹ cm⁻¹.

Enzyme activity assay and kinetic studies

The activity and kinetics of MtbPYK were determined as described for T. brucei PYK previously²⁷ with some modifications. Briefly, PYK activity was measured by following the decrease in NADH absorbance at 340 nm using a BioTek Synergy 4 microplate reader and the data were analysed by the software Graphpad Prism 7. The assay was performed at 25 °C in 100 μl reaction mixtures consisting of 1x assay buffer (50 mM TEA pH 7.2, 100 mM KCl, 10 mM MgCl₂), 0.5 mM NADH, 3.2 U LDH and 1 μg ml⁻¹ MtbPYK. The specific activity of MtbPYK was calculated from the corresponding kinetic curve. One activity unit is the reaction of 1 μmol substrate per minute under the condition used in this assay. The turnover value (k _cat) of MtbPYK was calculated from the specific activity divided by the subunit molar mass of 50,950.4 g mol⁻¹.

Enzyme kinetics with respect to ADP were studied with saturated PEP (8 mM) and at variable concentrations of ADP (from 0 to 4 mM). Enzyme kinetics with respect to PEP were studied with saturated ADP (4 mM) and at variable concentrations of PEP (from 0 to 8 mM) in the presence or absence of 1 mM activator AMP or G6P, or 2 mM inhibitor ATP. The combinations of effectors were also studied: 1 mM AMP plus 1 mM G6P with or without 2 mM ATP. The hyperbolic kinetics with respect to ADP and sigmoidal kinetics with respect to PEP were analysed using Graphpad Prism 7, respectively.

Enzyme kinetics with respect to the synergistic activators AMP and G6P were studied at 4 mM ADP, 0.2 mM PEP and variable concentrations of AMP and G6P. The synergistic activation was studied by adding G6P or AMP in AMP titration or G6P titration. The synergistic activation by AMP was studied in the presence of G6P (25 or 50 µM) and variable concentrations of AMP (from 0 to 3.2 mM). The synergistic activation by G6P was similarly studied in the presence of AMP (12.5 or 25 µM) and variable concentrations of G6P (from 0 to 3.2 mM). The enzyme kinetics with respect to allosteric effectors were analysed with Graphpad Prism 7 using an allosteric sigmoidal model. All kinetic results are summarised in Table 1. A 3D bar illustration was plotted by MATLAB showing the relation of effector concentrations and MtbPYK enzyme activity, where x/y-axis represents the effector concentration and the height (z-axis) is the percentage (%) of enzyme’s maximum rate (V _max).

Thermal shift assay

The thermal stability analysis of MtbPYK was determined as described for T. brucei PYK previously²⁷. Briefly, the assay was performed in a 96-well PCR plate (Bio-Rad) with 50 μl per reaction containing 5× SYPRO Orange dye (Invitrogen), 4 μM MtbPYK enzyme and 10 mM of the test ligand(s) (PEP, oxalate, pyruvate, ATP, AMP, G6P, F16BP, R5P, PO₄). The assay buffer (50 mM TEA pH 7.2, 100 mM KCl, 10 mM MgCl₂) was added instead of the test ligand as a negative control. To start the experiment, the temperature was increased from 25 to 95 °C in an i-Cycler iQ5 real-time PCR system (Bio-Rad). The thermal stability curve was analysed using the Bio-Rad iQ5 software, then the temperature midpoint T _m for the protein-unfolding transition was calculated.

Crystallization and data collection

The crystallisation experiments were performed by the vapour-diffusion method using the hanging-drop technique at 4 °C. The drops were equilibrated against a reservoir filled with 1 ml well solution. To crystallise MtbPYK in the inactive state (T-state), drops were formed by mixing 1.5 μl well solution with 1.0 μl protein solution (20 mg ml⁻¹). To co-crystallise the enzyme with the PEP analogue oxalate (OX), product ATP or synergistic activators G6P and AMP, 1.0 μl protein solution was mixed with 0.5 μl ligand solution (20 mM) and incubated at room temperature for 1–2 min. Then 1.5 μl well solution was added to the mixture for crystallisation. Oxalate is a structural analogue of the enolate form of pyruvate and has been generally used in crystallisation conditions to stabilise PYK in the active R-state^{26, 27}. The well solution consisted of 6–20% PEG 8000, 10-20% glycerol, 50 mM TEA buffer pH 7.2, 100 mM KCl, 50 mM MgCl₂. Crystals of T-state MtbPYK, R-states MtbPYK-OX, MtbPYK-OX/ATP/AMP and MtbPYK-OX/AMP/G6P were grown in conditions consisting of 11–16% PEG 8000 plus 10–20% glycerol. Although ATP was also added to the crystallisation conditions of the crystals MtbPYK-OX and MtbPYK-OX/AMP/G6P, electron density of ATP was not discovered in either structure. Crystals of MtbPYK-OX/G6P and MtbPYK-OX/AMP/G6P (different from the co-crystallised one) were obtained by soaking MtbPYK-OX crystals with 5 mM G6P and 5 mM AMP/G6P mixture, respectively. The soaking solution contained 5 mM for each ligand, 20% PEG 8000 and 20% glycerol.

X-ray intensity data for crystals of T-state MtbPYK and MtbPYK-OX/ATP/AMP were collected at the Swiss Light source (SLS, Switzerland) while the intensity data for crystals MtbPYK-OX, MtbPYK-OX/G6P, MtbPYK-OX/AMP/G6P and MtbPYK-OX/AMP/G6P (soak) were collected at the Australian Synchrotron (Australia). Each data set was from a single crystal flash-cooled in liquid nitrogen at 100 K. Data were then processed with MOSFLM³¹ and scaled with AIMLESS^{32, 33}. The data-collection and processing statistics are summarised in Table 2.

Structure determination

The T-state MtbPYK structure was solved by molecular replacement using the program Phaser³⁴. The initial search model (MtbPYK monomer) for the molecular-replacement experiment was obtained from the I-TASSER server³⁵. The structure was manually adjusted using Coot³⁶ followed by several cycles of restrained refinement in REFMAC³⁷. Where appropriate, water molecules and ligands were added to the structure and TLS refinement was applied at later stage of refinement. The mobile B-domain of chain C has poor density that could not be improved after refinement. Thus, this B-domain (residues 72–164) was deleted from the final refined structure.

The diffraction data sets of the R-state MtbPYK structures [MtbPYK-OX, MtbPYK-OX/ATP/AMP, MtbPYK-OX/G6P, MtbPYK-OX/AMP/G6P and MtbPYK-OX/AMP/G6P (soak)] were scaled to the space group of P6₄22 with reasonable statistics assessed by the program AIMLESS^{32, 33}. The refined structure of the T-state MtbPYK monomer was used as the search model for molecular replacement by Phaser³⁴. After refinement using REFMAC³⁷ and manual model building in Coot³⁶, the R/R _free remained high (both values >0.40 for data sets at 1.9–2.8 Å resolution). The L-test results indicated merohedral twinning had occurred in data sets, although there is no twin law in the space group of P6₄22. The data sets were then scaled to all possible space groups in lower symmetry with possible twin laws. The results of molecular replacement and refinement indicate the data sets scaled in P3₁ have the most reasonable R/R _free values (0.15–0.20) and density qualities. The solved structures determined in space group P3₁ were further refined by following the procedures used for the T-state structure refinement.

The quality of the structures was assessed using the MOLPROBITY server³⁸, and the figures were generated using PyMOL³⁹. The data processing and refinement statistics are summarised in Table 2. The structure factors and coordinates for T-state MtbPYK, MtbPYK-OX, MtbPYK-OX/ATP/AMP, MtbPYK-OX/G6P, MtbPYK-OX/AMP/G6P and MtbPYK-OX/AMP/G6P (soak) have been deposited in the RCSB Protein Data Bank as PDB entries 5WRP, 5WS8, 5WS9, 5WSA, 5WSB and 5WSC, respectively.

Structure analysis

The program Superpose⁴⁰ in the CCP4⁴¹ suite was used to calculate the allosteric rigid-body rotations from the superposition of T-state and R-state tetramers as described previously²⁶. Both RMS differences and rotation matrices were calculated in the superposition process²⁶.

The equivalent Cα atom distances of superposed structures are analysed by heat-map analysis to highlight the conformational changes within structures. AC-cores (A and C domains, residues 1–70 and 168–472) from each R-state MtbPYK structure were superposed (Cα superposition) onto the AC-core of the T-state structure using program Superpose⁴⁰. Values of equivalent Cα distances (Dist.) from each pair of superposition were used to generate heat maps by GraphPad Prism 7. The increase of distance is represented as a blue to red gradient.

Molecular simulations and structural network analysis

Three MtbPYK crystal structures with OX/G6P, OX/AMP, OX/AMP/G6P bound, were prepared using Visual Molecular Dynamics (VMD)⁴². Each tetrameric structure was solvated in a sufficiently large water box with salinity set to 150 mM NaCl. The systems, typically involving 220,000 atoms, were then subjected to all-atom MD simulations for 2 ns with positional restraints to the backbone atoms and a further 150 ns without any restraints. The simulations were performed using NAMD 2.10⁴³ assuming the CHARMM36 force field for the protein⁴⁴ and assuming the TIP3P model for water molecules⁴⁵. The CHARMM parameters for oxalate, AMP and G6P were obtained from cgenff^{46, 47}. Network analysis was performed to investigate the dynamic coupling between oxalate and AMP/G6P. We calculated 1000 highly correlated paths of residues between ligand sites of interest for every MtbPYK monomer by employing the weighted implementation of suboptimal paths (WISP) algorithm^{28, 48}. Therefore, a total of 4000 suboptimal paths were collected for the MtbPYK tetramer. To perform community network analysis⁴⁹, a weighted network was first constructed where each residue represents a node and the weight of the connection between nodes represents their respective correlation values. Communities were then identified such that each community contains a group of highly correlated residues⁵⁰. Note that the simulation has converged after 20 ns (Supplementary Fig. 10) and all analyses were done on the last 100 ns of the MD simulation trajectory with 20 ps interval.

Mycobacterium bovis BCG culture and nutrient starvation

Isogenic freezer stocks of Mycobacterium bovis bacille Calmette Guérin (BCG) Pasteur strain 1172P2 were used for starter cultures in 50 ml conical tubes, which were subsequently passaged into roller bottles with supplemented 7H9 medium⁵¹. To obtain nutrient-deprived cultures, BCG were grown to OD₆₀₀ of 0.8–1.0 and washed twice in PBS (pH 7.4) containing 137 mM NaCl, 2.7 mM KCl, 10 mM NaH₂PO₄, 1.8 mM KH₂PO₄, then re-suspended in PBS with 0.05% Tyloxapol to an OD₆₀₀ of ~1.0 in roller bottles for up to 22 days. At 4, 10 and 22 d post-nutrient deprivation (termed S4, S10 and S22, respectively), samples from 4 independent cultures were taken for metabolic assessment. The remaining samples at S22 were re-inoculated into nutrient replete 7H9 at a starting OD₆₀₀ of 0.1. The metabolites in these cultures (termed R6) are assessed 6 d post inoculation. CFUs for BCG were determined from 10-fold serial dilution plating on 7H10 agar at 37 °C for 3 w.

Metabolite extraction and targeted metabolomics

Metabolite extraction followed the published report⁵². Briefly, cell cultures were harvested at given time points, rapidly quenched and spun down. Cell pellets were resuspended in acetonitrile:methanol:water (2:2:1) and lysed mechanically with 0.1-mm silica beads by using Qiagen TissuelyserII. The lysates were collected and evaporated to dryness in a vacuum evaporator, and the dry extracts were redissolved in 100 µl of 98:2 water/methanol for liquid chromatography-mass spectrometry (LC-MS) analysis.

The targeted LC-MS/MS analysis was performed with Agilent 1290 ultrahigh pressure liquid chromatography system coupled to a 6460 Triple Quadrupole mass spectrometer equipped with a dual-spray electrospray ionization source with Jet Stream^TM (Agilent Technologies, Santa Clara, CA). Chromatographic separation was achieved by using Phenomenex (Torrance, CA) Rezex^TM ROA-Organic Acid H + (8%) column (2.1 × 100 mm, 3 µm) and the compounds were eluted at 40 °C with an isocratic flow rate of 0.3 ml min⁻¹ of 0.1% formic acid in water. The auto-sampler was cooled at 4 °C and an injection volume of 5 μl was used. Electrospray ionisation was performed in negative ion mode with the following source parameters: drying gas temperature 300 °C with a flow of 10 l min⁻¹, nebulizer gas pressure 40 psi, sheath gas temperature 350 °C with a flow of 11 l min⁻¹, capillary voltage 3000 V and nozzle voltage 500 V. Compounds were quantified in multiple reaction monitoring (MRM) mode with the following transitions: m/z 426 > 124 and m/z 426 > 79 for adenosine diphosphate (ADP), m/z 259 > 199 and m/z 259 > 138.9 for glucose 6-phosphate (G6P), m/z 87 > 43.1 and m/z 87 > 32.1 for pyruvate, m/z 167 > 79 and m/z 167 > 63 for phosphoenolpyruvic acid (PEP), m/z 229 > 169.1 and m/z 229 > 138.9 for ribose 5-phosphate (R6P), m/z 339 > 96.9 and m/z 339 > 79 for fructose 1,6-bisphosphate (F16BP). Adenosine monophosphate (AMP) was analysed using Agilent 1290 ultrahigh pressure liquid chromatography system equipped with a 6520 QTOF mass spectrometer equipped with a dual-spray electrospray ionization source (Agilent Technologies, Santa Clara, CA). The column used for the separation was an Agilent rapid resolution HT Zorbax SB-C18 (2.1×100 mm, 1.8 mm). The oven temperature was set at 45 °C and the flow rate was set at 0.4 ml min⁻¹. The gradient mobile phase consisted of 0.1% formic acid in water (A) and 0.1% formic acid in methanol (B). Metabolites were eluted with the following gradient: 0 to 7 min, 2–70% B; 7 to 8 min, 70–100% B; 8–11 min, hold at 100% B, 11–11.1 min, 100-2% B. The auto-sampler was cooled at 4 °C and a 5 μl injection volume was used. Mass spectrometry was performed at positive ion electrospray ionization mode and the mass data were collected between m/z 100 and 600 Da at a scan rate of 2 spectra per second. The ion spray voltage and the heated capillary temperature was set at 4000 V and 350 °C, respectively The drying gas and nebulizer nitrogen gas flow rates were set to 12.0 l min⁻¹ and 50 psi, respectively. Two reference masses were continuously infused to the system to ensure mass accuracy during the run: m/z 121.0509 (C₅H₄N₄) and m/z 922.0098 (C₁₈H₁₈O₆N₃P₃F₂₄). The MassHunter software (Agilent Technologies, USA) was used to collect and process the data. Residual protein content was determined to normalise samples to cell biomass (BCA protein assay kit; Thermo Scientific). The metabolic abundance data were normalised to protein concentration.

Data availability

Coordinates and structure factors for the T-state structure MtbPYK, R-state complexes MtbPYK-OX, MtbPYK-OX/AMP/ATP, MtbPYK-OX/G6P, MtbPYK-OX/AMP/G6P, MtbPYK-OX/AMP/G6P (soaking method) have been deposited in the Protein Data Bank under accession codes 5WRP, 5WS8, 5WS9, 5WSA, 5WSB and 5WSC, respectively. The authors declare that the data supporting the findings of this study are available within the article and its Supplementary Information Files, or from the corresponding authors upon reasonable request.

References

World Health Organization. Global Tuberculosis Report 2015. http://www.who.int/tb/publications/global_report/gtbr15_main_text.pdf (2015).
Schnappinger, D. et al. Transcriptional adaptation of Mycobacterium tuberculosis within macrophages: insights into the phagosomal environment. J. Exp. Med. 198, 693–704 (2003).
Article CAS PubMed PubMed Central Google Scholar
Rhee, K. Y. et al. Central carbon metabolism in Mycobacterium tuberculosis: an unexpected frontier. Trends Microbiol. 19, 307–314 (2011).
Article CAS PubMed PubMed Central Google Scholar
de Carvalho, L. P. S. et al. Metabolomics of Mycobacterium tuberculosis reveals compartmentalized co-catabolism of carbon substrates. Chem. Biol. 17, 1122–1131 (2010).
Article PubMed Google Scholar
Cumming, B. M. & Steyn, A. J. C. Metabolic plasticity of central carbon metabolism protects mycobacteria. Proc. Natl Acad. Sci. USA 112, 13135–13136 (2015).
Article ADS CAS PubMed PubMed Central Google Scholar
Wagner, T., Bellinzoni, M., Wehenkel, A., O’Hare, H. M. & Alzari, P. M. Functional plasticity and allosteric regulation of α-ketoglutarate decarboxylase in central mycobacterial metabolism. Chem. Biol. 18, 1011–1020 (2011).
Article CAS PubMed Google Scholar
Maksymiuk, C., Balakrishnan, A., Bryk, R., Rhee, K. Y. & Nathan, C. F. E1 of α-ketoglutarate dehydrogenase defends Mycobacterium tuberculosis against glutamate anaplerosis and nitroxidative stress. Proc. Natl Acad. Sci. USA 112, E5834–E5843 (2015).
Article ADS CAS PubMed PubMed Central Google Scholar
Lamprecht, D. A. et al. Turning the respiratory flexibility of Mycobacterium tuberculosis against itself. Nat. Commun. 7, 12393 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Noy, T. et al. Central role of pyruvate kinase in carbon co-catabolism of Mycobacterium tuberculosis. J. Biol. Chem. 291, 7060–7069 (2016).
Article CAS PubMed PubMed Central Google Scholar
Zoraghi, R. et al. Methicillin-resistant Staphylococcus aureus (MRSA) pyruvate kinase as a target for bis-indole alkaloids with antibacterial activities. J. Biol. Chem. 286, 44716–44725 (2011).
Article CAS PubMed PubMed Central Google Scholar
Morgan, H. P. et al. The trypanocidal drug suramin and other trypan blue mimetics are inhibitors of pyruvate kinases and bind to the adenosine site. J. Biol. Chem. 286, 31232–31240 (2011).
Article CAS PubMed PubMed Central Google Scholar
Chaneton, B. & Gottlieb, E. Rocking cell metabolism: revised functions of the key glycolytic regulator PKM2 in cancer. Trends Biochem. Sci. 37, 309–316 (2012).
Article CAS PubMed Google Scholar
Wang, J. et al. Haem-activated promiscuous targeting of artemisinin in Plasmodium falciparum. Nat. Commun. 6, 10111 (2015).
Article ADS CAS PubMed PubMed Central Google Scholar
Morgan, H. P. et al. Structures of pyruvate kinases display evolutionarily divergent allosteric strategies. R. Soc. Open Sci. 1, 140120–140120 (2014).
Article PubMed PubMed Central Google Scholar
Waygood, E. B. & Sanwal, B. D. The control of pyruvate kinases of Escherichia coli. I. Physicochemical and regulatory properties of the enzyme activated by fructose 1,6-diphosphate. J. Biol. Chem. 249, 265–274 (1974).
CAS PubMed Google Scholar
Garcia-Olalla, C. & Garrido-Pertierra, A. Purification and kinetic properties of pyruvate kinase isoenzymes of Salmonella typhimurium. Biochem. J. 241, 573–581 (1987).
Article CAS PubMed PubMed Central Google Scholar
Jurica, M. S. et al. The allosteric regulation of pyruvate kinase by fructose-1,6-bisphosphate. Structure 6, 195–210 (1998).
Article CAS PubMed Google Scholar
Dombrauckas, J. D., Santarsiero, B. D. & Mesecar, A. D. Structural basis for tumor pyruvate kinase M2 allosteric regulation and catalysis. Biochemistry 44, 9417–9429 (2005).
Article CAS PubMed Google Scholar
Morgan, H. P. et al. M2 pyruvate kinase provides a mechanism for nutrient sensing and regulation of cell proliferation. Proc. Natl Acad. Sci. USA 110, 5881–5886 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Chaneton, B. et al. Serine is a natural ligand and allosteric activator of pyruvate kinase M2. Nature 491, 458–462 (2012).
Article ADS CAS PubMed PubMed Central Google Scholar
Anastasiou, D. et al. Pyruvate kinase M2 activators promote tetramer formation and suppress tumorigenesis. Nat. Chem. Biol. 8, 839–847 (2012).
Article CAS PubMed PubMed Central Google Scholar
Abbe, K. & Yamada, T. Purification and properties of pyruvate kinase from Streptococcus mutans. J. Bacteriol. 149, 299–305 (1982).
CAS PubMed PubMed Central Google Scholar
Zoraghi, R. et al. Functional analysis, overexpression, and kinetic characterization of pyruvate kinase from methicillin-resistant Staphylococcus aureus. Biochemistry 49, 7733–7747 (2010).
Article CAS PubMed Google Scholar
Veith, N. et al. Organism-adapted specificity of the allosteric regulation of pyruvate kinase in lactic acid bacteria. PLoS Comput. Biol. 9, e1003159 (2013).
Article CAS PubMed PubMed Central Google Scholar
Waygood, E. B., Rayman, M. K. & Sanwal, B. D. The control of pyruvate kinases of Escherichia coli. II. Effectors and regulatory properties of the enzyme activated by ribose 5-phosphate. Can. J. Biochem. 53, 444–454 (1975).
Article CAS PubMed Google Scholar
Morgan, H. P. et al. Allosteric mechanism of pyruvate kinase from Leishmania mexicana uses a rock and lock model. J. Biol. Chem. 285, 12892–12898 (2010).
Article CAS PubMed PubMed Central Google Scholar
Zhong, W. et al. ‘In crystallo’ substrate binding triggers major domain movements and reveals magnesium as a co-activator of Trypanosoma brucei pyruvate kinase. Acta Crystallogr. D. Biol. Crystallogr. 69, 1768–1779 (2013).
Article CAS PubMed Google Scholar
Van Wart, A. T., Durrant, J., Votapka, L. & Amaro, R. E. Weighted implementation of suboptimal paths (WISP): an optimized algorithm and tool for dynamical network analysis. J. Chem. Theory Comput. 10, 511–517 (2014).
Article PubMed PubMed Central Google Scholar
Marrero, J., Trujillo, C., Rhee, K. Y. & Ehrt, S. Glucose phosphorylation is required for Mycobacterium tuberculosis persistence in mice. PLoS Pathog. 9, e1003116 (2013).
Article CAS PubMed PubMed Central Google Scholar
Hasan, M. R., Rahman, M., Jaques, S., Purwantini, E. & Daniels, L. Glucose 6-phosphate accumulation in mycobacteria: implications for a novel F420-dependent anti-oxidant defense system. J. Biol. Chem. 285, 19135–19144 (2010).
Article CAS PubMed PubMed Central Google Scholar
Battye, T. G. G., Kontogiannis, L., Johnson, O., Powell, H. R. & Leslie, A. G. W. iMOSFLM: a new graphical interface for diffraction-image processing with MOSFLM. Acta Crystallogr. D. Biol. Crystallogr. 67, 271–281 (2011).
Article CAS PubMed PubMed Central Google Scholar
Evans, P. IUCr. Scaling and assessment of data quality. Acta Crystallogr. D. Biol. Crystallogr. 62, 72–82 (2006).
Article PubMed Google Scholar
Evans, P. R. & Murshudov, G. N. IUCr. How good are my data and what is the resolution? Acta Crystallogr. D. Biol. Crystallogr. 69, 1204–1214 (2013).
Article CAS PubMed PubMed Central Google Scholar
McCoy, A. J. et al. Phaser crystallographic software. J. Appl. Crystallogr. 40, 658–674 (2007).
Article CAS PubMed PubMed Central Google Scholar
Zhang, Y. I-TASSER server for protein 3D structure prediction. BMC Bioinforma. 9, 40 (2008).
Article Google Scholar
Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics. Acta Crystallogr. D. Biol. Crystallogr. 60, 2126–2132 (2004).
Article PubMed Google Scholar
Murshudov, G. N. et al. REFMAC5 for the refinement of macromolecular crystal structures. Acta Crystallogr. D. Biol. Crystallogr. 67, 355–367 (2011).
Article CAS PubMed PubMed Central Google Scholar
Davis, I. W. et al. MolProbity: all-atom contacts and structure validation for proteins and nucleic acids. Nucleic Acids Res. 35, W375–W383 (2007).
Article ADS PubMed PubMed Central Google Scholar
DeLano, W. L. The PyMOL Molecular Graphics System (DeLano Scientific, San Carlos, CA, USA, 2002).
Krissinel, E. & Henrick, K. Secondary-structure matching (SSM), a new tool for fast protein structure alignment in three dimensions. Acta Crystallogr. D. Biol. Crystallogr. 60, 2256–2268 (2004).
Article CAS PubMed Google Scholar
Winn, M. D. et al. Overview of the CCP4 suite and current developments. Acta Crystallogr. D. Biol. Crystallogr. 67, 235–242 (2011).
Article CAS PubMed PubMed Central Google Scholar
Humphrey, W., Dalke, A. & Schulten, K. VMD: visual molecular dynamics. J. Mol. Graph. 14, 33–39 (1996). 27-8.
Article CAS PubMed Google Scholar
Phillips, J. C. et al. Scalable molecular dynamics with NAMD. J. Comput. Chem. 26, 1781–1802 (2005).
Article CAS PubMed PubMed Central Google Scholar
Best, R. B. et al. Optimization of the additive CHARMM all-atom protein force field targeting improved sampling of the backbone φ, ψ and side-chain χ(1) and χ(2) dihedral angles. J. Chem. Theory Comput. 8, 3257–3273 (2012).
Article CAS PubMed PubMed Central Google Scholar
Jorgensen, W. L., Chandrasekhar, J., Madura, J. D., Impey, R. W. & Klein, M. L. Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 79, 926–935 (1983).
Article ADS CAS Google Scholar
Vanommeslaeghe, K. & MacKerell, A. D. Jr. Automation of the CHARMM General Force Field (CGenFF) I: Bond perception and atom typing. J. Chem. Inform. Model. 52, 3144–3154 (2012).
Article CAS Google Scholar
Vanommeslaeghe, K., Raman, E. P. & MacKerell, A. D. Jr. Automation of the CHARMM General Force Field (CGenFF) II: Assignment of bonded parameters and partial atomic charges. J. Chem. Inform. Model. 52, 3155–3168 (2012).
Article CAS Google Scholar
Heyrana, K. J. et al. Contributions of charged residues in structurally dynamic capsid surface loops to rous sarcoma virus assembly. J. Virol. 90, 5700–5714 (2016).
Article CAS PubMed PubMed Central Google Scholar
Yang, J. et al. Synergistic allosteric mechanism of fructose-1,6-bisphosphate and serine for pyruvate kinase M2 via dynamics fluctuation network analysis. J. Chem. Inf. Model. 56, 1184–1192 (2016).
Article CAS PubMed PubMed Central Google Scholar
Sethi, A., Eargle, J., Black, A. A. & Luthey-Schulten, Z. Dynamical networks in tRNA:protein complexes. Proc. Natl Acad. Sci. USA 106, 6620–6625 (2009).
Article ADS CAS PubMed PubMed Central Google Scholar
Chionh, Y. H. et al. tRNA-mediated codon-biased translation in mycobacterial hypoxic persistence. Nat. Commun. 7, 13302 (2016).
Article ADS CAS PubMed PubMed Central Google Scholar
Berney, M. et al. Essential roles of methionine and S-adenosylmethionine in the autarkic lifestyle of Mycobacterium tuberculosis. Proc. Natl Acad. Sci. USA 112, 10008–10013 (2015).
Article ADS CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

This research was supported by the National Research Foundation of Singapore through the Singapore-MIT-Alliance for Research and Technology Infectious Disease research program, and the Singapore-MIT Alliance for Research and Technology postdoctoral fellowship. During the course of this study, J.L. lab was supported by grant NMRC/CBRG/0073/2014. The MD simulations were performed on ASPIRE 1 of the National Supercomputing Centre (NSCC) Singapore. We are grateful to Dr. Ghader Bashiri from The University of Auckland (New Zealand) for the gift of the vector pYUB28b.

Author information

Yok Hian Chionh
Present address: Tychan Private Ltd, 80 Robinson Road, #17-02, Singapore, 068898, Singapore
Meng Yuan
Present address: Department of Integrative Structural and Computational Biology, The Scripps Research Institute, La Jolla, CA, 92037, USA

Authors and Affiliations

Infectious Disease Interdisciplinary Research Group, Singapore-MIT Alliance for Research and Technology, 1 CREATE Way, Singapore, 138602, Singapore
Wenhe Zhong, Liang Cui, Boon Chong Goh, Peiying Ho, Yok Hian Chionh & Peter C. Dedon
NTU Institute of Structural Biology, Nanyang Technological University, Singapore, 636921, Singapore
Wenhe Zhong, Boon Chong Goh, Abbas El Sahili & Julien Lescar
Division of Life Science, State Key Laboratory of Molecular Neuroscience, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China
Qixu Cai
Institute of Quantitative Biology, Biochemistry and Biotechnology, University of Edinburgh, King’s Buildings, Edinburgh, EH9 3BF, UK
Meng Yuan, Linda A. Fothergill-Gilmore & Malcolm D. Walkinshaw
School of Biological Sciences, Nanyang Technological University, 60 Nanyang Drive, Singapore, 637551, Singapore
Abbas El Sahili & Julien Lescar
Department of Biological Engineering, Massachusetts Institute of Technology, Cambridge, MA, 02139, USA
Peter C. Dedon

Authors

Wenhe Zhong
View author publications
You can also search for this author in PubMed Google Scholar
Liang Cui
View author publications
You can also search for this author in PubMed Google Scholar
Boon Chong Goh
View author publications
You can also search for this author in PubMed Google Scholar
Qixu Cai
View author publications
You can also search for this author in PubMed Google Scholar
Peiying Ho
View author publications
You can also search for this author in PubMed Google Scholar
Yok Hian Chionh
View author publications
You can also search for this author in PubMed Google Scholar
Meng Yuan
View author publications
You can also search for this author in PubMed Google Scholar
Abbas El Sahili
View author publications
You can also search for this author in PubMed Google Scholar
Linda A. Fothergill-Gilmore
View author publications
You can also search for this author in PubMed Google Scholar
Malcolm D. Walkinshaw
View author publications
You can also search for this author in PubMed Google Scholar
Julien Lescar
View author publications
You can also search for this author in PubMed Google Scholar
Peter C. Dedon
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

W.Z. carried out the laboratory work and crystal structure determination, participated in the design of the study and drafted the manuscript; L.C., P.H. and Y.H.C. performed targeted metabolomics and drafted the manuscript; B.C.G. carried out the MD simulations and helped prepare the manuscript; A.E.S. collected the diffraction data and W.Z., Q.C. and M.Y. helped to carry out analysis of diffraction data and to refine structures; W.Z., L.A.F.-G., M.D.W., J.L. and P.C.D. conceived of the study, coordinated the study and helped draft the manuscript.

Corresponding authors

Correspondence to Julien Lescar or Peter C. Dedon.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Additional information

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Electronic supplementary material

Supplementary Information

Description of Additional Supplementary Files

Supplementary Movie 1

Supplementary Movie 2

Supplementary Movie 3

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commonslicense, unless indicated otherwise in a credit line to the material. If material is not included in the article’sCreative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Zhong, W., Cui, L., Goh, B.C. et al. Allosteric pyruvate kinase-based “logic gate” synergistically senses energy and sugar levels in Mycobacterium tuberculosis . Nat Commun 8, 1986 (2017). https://doi.org/10.1038/s41467-017-02086-y

Download citation

Received: 10 July 2017
Accepted: 06 November 2017
Published: 07 December 2017
DOI: https://doi.org/10.1038/s41467-017-02086-y

This article is cited by

A three-level regulatory mechanism of the aldo-keto reductase subfamily AKR12D
- Zhihong Xiao
- Jinyin Zha
- Shaobo Dai
Nature Communications (2024)
Bactericidal bissulfone B7 targets bacterial pyruvate kinase to impair bacterial biology and pathogenicity in plants
- Awei Zhang
- Haizhen Zhang
- Runjiang Song
Science China Life Sciences (2024)
Programming conformational cooperativity to regulate allosteric protein-oligonucleotide signal transduction
- Yuan Liang
- Yunkai Qie
- Cheng Zhang
Nature Communications (2023)
PKM2 deficiency exacerbates gram-negative sepsis-induced cardiomyopathy via disrupting cardiac calcium homeostasis
- Le Ni
- Bowen Lin
- Dan Shi
Cell Death Discovery (2022)
Diethyl phthalate (DEP) perturbs nitrogen metabolism in Saccharomyces cerevisiae
- Corinna Jie Hui Goh
- Liang Cui
- Prakash Arumugam
Scientific Reports (2022)

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.