Abstract
Enoyl-CoA carboxylases/reductases (ECRs) are the most efficient CO2-fixing enzymes described to date, outcompeting RubisCO, the key enzyme in photosynthesis in catalytic activity by more than an order of magnitude. However, the molecular mechanisms underlying ECR’s extraordinary catalytic activity remain elusive. Here we used different crystallographic approaches, including ambient temperature X-ray Free Electron Laser (XFEL) experiments, to study the dynamic structural organization of the ECR from Kitasatospora setae. K. setae ECR is a homotetramer that differentiates into a dimer of dimers of open- and closed-form subunits in the catalytically active state, suggesting that the enzyme operates with “half-site reactivity” to achieve high catalytic rates. Using structure-based mutagenesis, we show that catalysis is synchronized in K. setae ECR across the pair of dimers by conformational coupling of catalytic domains and within individual dimers by shared substrate binding sites. Our results provide unprecedented insights into the dynamic organization and synchronized inter- and intra-subunit communications of nature’s most efficient CO2-fixing enzyme during catalysis.
INTRODUCTION
The capture and conversion of atmospheric CO2 remains a challenging task for chemistry, resulting in an ever-increasing interest to understand and exploit CO2 fixation mechanisms offered by biology1. The recently described family of enoyl-CoA carboxylases/reductases (ECRs) represent the most efficient CO2-fixing enzymes found in nature to date2,3. ECRs catalyze the reductive carboxylation of a variety of enoyl-CoA thioester substrates at catalytic rates that are up to 20-fold higher than Ribulose-1,5-bisphosphate carboxylase/oxygenase (RubisCO), an enzyme involved in the first carbon fixation step in the Calvin-Benson cycle of photosynthesis1,4.
ECRs catalyze the reduction of α,β-unsaturated enoyl-CoAs using the reduced form of the cofactor nicotinamide adenine dinucleotide phosphate (NADPH). This generates a reactive enolate species, which acts as a nucleophile to attack a CO2 molecule2,3,5. The structural details of the carboxylation reaction have remained elusive, due in part to the lack of high-resolution structures of ECRs containing catalytic intermediates and carboxylated products. Currently, there are five available ECR structures. However, they all have different substrate specificities, ranging from short-(PDB: 3HZZ, 3KRT) to long-chain (4A0S6) and aromatic enoyl-CoA substrates (4Y0K7), and are from different biological backgrounds including primary (i.e. central carbon) metabolism (PDB: 4GI2) and secondary metabolism. Moreover, most of them were co-crystalized with NADPH or NADP+ only and do not contain CO2, enoyl-CoA substrates or acyl-CoA products. This significantly limits our structural understanding of the enzyme’s catalytic mechanism.
The aim of this study was to provide a detailed structural understanding of the carboxylation reaction of ECRs at the level of the oligomeric protein complex. To this end, we chose the ECR from K. setae, which shows high substrate specificity for crotonyl-CoA and superior catalytic efficiency (see Table 1). Using cryogenic X-ray crystallography at synchrotrons and room temperature serial femtosecond X-ray crystallography (SFX) at an XFEL, four high-resolution ECR structures were determined in different conformational states: the apo form and three holo forms, in binary complex with the reduced cofactor NADPH, in ternary complex with NADPH and butyryl-CoA, and in binary complex with the oxidized cofactor NADP+ (Figure 1a).
Here we show that the tetrameric complex assumes a dimer-of-dimers (“a pair of dimers”) configuration during catalysis. The central oligomerization domains of ECR remain largely unchanged, while the peripheral catalytic domains move drastically to provide two sets of active site conformations, open- and closed-form, upon binding of the NADPH cofactor alone or in the presence of substrates. This coordinated motion is enabled by a tight coupling of catalytic domains across the pairs of dimers. Structure based mutagenesis of the interface of the catalytic domains supports this notion and provides compelling evidence that synchronization across the pair of dimers is a crucial factor in K. setae ECR to achieve the high catalytic efficiency. Further kinetic experiments demonstrate that subunit communication within the pair of dimers is important to synchronize open- and closed-states. Altogether, our data unveil a detailed picture of the dynamic structural organization and subunit synchronization of the ECR complexes, providing unprecedented insights into the functional organization of nature’s most efficient CO2-fixing enzyme during catalysis.
RESULTS
Apo ECR is a symmetric homotetramer, readily accessible for NADPH binding
We first determined the apo form of the ECR crystal structure from K. setae at 1.8 Å resolution by using synchrotron X-ray crystallography at cryogenic temperature (Supplementary Table 1). The asymmetric unit contains one homotetramer composed of four subunits arranged in a dimer of dimers geometry (“pair of dimers”) similar to those of the previously reported binary (PDB: 4Y0K) and ternary (PDB: 4A0S) ECR structures. Overall, the tetramer shows a non-crystallographic, close to D2 (dihedral) symmetry (Supplementary Figure S1, top right panel) with four conformationally identical subunits (Supplementary Figures S1&S2, RMSD = 0.1 Å). The tetrameric structure of K. setae ECR is further supported by size-exclusion chromatography which showed that the apo enzyme eluted as a single peak at 205 kDa compared to the expected monomer molecular weight of 51.2 KDa corresponding to a functional complex of four subunits (Supplementary Figure S3).
Each ECR subunit consists of two domains – a larger catalytic domain formed by residues 1-212 and 364-445, and a smaller oligomerization domain formed by residues 212 to 363 (Supplementary Figure S4). The oligomerization domain comprises a Rossmann fold8 with repeating αβ-motifs that forms a 6-stranded β-sheet (β12 to β17). The 6-stranded β-sheets of two neighboring subunits are combined into one 12-stranded β-sheet, forming the core of one dimer, A/C or B/D. Two of these 12-stranded β-sheets then form the core of the tetrameric complex (Supplementary Figure S4).
The catalytic domains of K. setae ECR are located at the periphery of the tetrameric complex. The active site of ECR is formed by helix 8 and surrounding loops at the interface with an adjacent subunit in the tetramer (Supplementary Figure S4). The active site cavities in the apo form are open and accessible for both the cofactors and substrates.
NADPH binding induces ECR into a dimer of dimers with distinct open and closed form subunits
To understand how cofactor binding affects the enzyme, we determined the crystal structure of the K. setae ECR-NADPH binary complex at 2.4 Å resolution by using serial femtosecond X-ray crystallography (SFX) at ambient temperature (Figure 2, Supplementary Table 1)9–12. The simple Fo-Fc difference electron density map allowed us to unambiguously place NADPH molecules in all four subunits. NADPH binds with its adenine moiety in the oligomerization domain and spans the catalytic domain, where its nicotinamide moiety is located (Supplementary Figure S5).
Notably, binding of NADPH breaks the dihedral D2 symmetry observed in the apo-form tetramer structure, while symmetry about the y-axis is retained, resulting in a non-crystallographic, almost cyclic C2 symmetry (Supplementary Figure S1 bottom right). In the NADPH-ECR binary complex, the four subunits of ECR differentiate into two forms (A & B and C & D), which are structurally distinct from each other (Supplementary Figure S1, RMSD = 0.5 Å between A & B and C & D, 1.8 Å between A & C, A & D, B & C and B & D respectively) (Figure 2a&b). The A & B subunits show cofactor-binding pockets that are open, referred hereafter as “open-form” state (Figure 2b). On the other hand, in the C & D subunits, the cofactor binding pocket is compressed inwards, which seals the NADPH cofactor within the catalytic domain, resulting in a “closed-form” state (Figure 2b).
The bulk of the NADPH cofactor is bound almost identically in the two closed-form subunits, C & D (Figure 2d,e&f). However, the nicotinamide moiety adopts two alternate conformations in the two open-form subunits A & B (Figure 2e&f), indicating a more flexible cofactor binding than in the closed-form subunits (Figure 2d). Possible conformations of the NADPH cofactor in the open and closed binding cavities and its flexibility were studied with molecular dynamics (MD) simulations in a dimer of subunits A and C. In the closed-form subunit C the cofactor kept its position in the binding site as observed in the X-ray crystal structure. When we placed the NADPH cofactor in the same position in the open subunit A and performed similar MD simulations, NADPH left this initial confirmation in all three 200 ns trajectories and adopted various alternate conformations in the open cavity, including the two that were observed in our high-resolution crystal structure (Supplementary Figure S6). These variable conformations in the open A subunit are allowed by a substrate binding pocket that is more than 5 Å wider than the closed C subunit (Figure 2c).
In summary, binding of the NADPH cofactors to the apo enzyme induces the four subunits of the enzyme to differentiate into open- and closed-form states in both dimers (Figure 2b) thus breaking the dihedral D2 symmetry to cyclic C2 symmetry. This coupled subunit rearrangement of K. setae ECR and the large active site differences within each pair of dimers suggest that catalysis is synchronized between the individual subunits of the complex, which will become clearer in the subsequent analysis sections below.
Ternary complex supports half-site reactivity in ECR catalysis
We next attempted to determine the structure of the K. setae ECR ternary complex crystallized in the presence of spent cofactor NADP+ and the reaction product ethylmalonyl-CoA. The structure of the ternary complex, however, indicated that the carboxylate group was lost during the crystallization process, which resulted in butyryl-CoA, which is in line with the finding that ethylmalonyl-CoA is unstable and tends to decarboxylate at the active site of ECR into butyryl-CoA and CO2 over time2,13(Supplementary Figure S7). Numerous attempts of preserving ethylmalonyl-CoA in the crystal structure proved to be extremely challenging and therefore we co-crystallized ECR with butyryl-CoA and NADPH and determined its structure at 1.7 Å resolution (Figure 3). This structure revealed that two butyryl-CoA molecules are bound at the active sites of the closed-form subunits B & D.
This ternary complex structure is overall very similar to the structure of ECR-NADPH binary complex. It also displays the non-crystallographic, pseudo C2 cyclic symmetry (Figure 3a and Supplementary Figure S1 bottom right panel) and comprises of open- and closed-form subunits that overlay very well with the open- and closed-form subunits of the ECR-NADPH binary complex (Supplementary Figures S1&S2, RMSD = 0.1, 2.1 Å respectively). The NADPH cofactor appears bound to all active sites, however, only the closed-form subunits B & D also contain the completely intact butyryl-CoA thioester (Figure 3a&b). This strongly suggests that the closed-form subunits represent the Michaelis complex in which substrate and cofactor are positioned for catalysis, while the open-form subunits represent catalytically incompetent complexes that are in place to perform the next round of catalysis.
ECR uses an elegant mechanism to align CoA-ester for catalysis in the closed-form subunit pairs. The active site of the closed-form subunits is sealed by the collective motion of loops 37-44, 88-94, 338-350, and helices 6, 7 and 21 of the catalytic domain (Figure 3c), which creates multiple interactions of the protein with the CoA-ester (Figure 3d,e&f). Notably, the CoA-ester extends from the catalytic closed-form domain into the neighboring open-form subunit within the same dimer pair, where Arg352, and Tyr353 interact with the phosphate backbone of CoA. The CoA moiety stretches further into the neighboring open-form subunit, where its adenosine tail interacts with a small binding pocket formed by three residues on the surface of the dimerization domain, Tyr328, Lys296, and Arg303 (Figure 3g).
When we inspected the adenine binding pocket of the closed-form subunits, we also observed electron density of adenine, indicating that the CoA-ester was bound (Figure 3g right inset). The electron density beyond the adenine ring, however, becomes disordered, suggesting that the part of the CoA molecule that extends into the active site of the neighboring open-form subunit remains flexible, which is corroborated by the higher anisotropy of the CoA binding site (Figure 1c). Quantum mechanical/molecular mechanics (QM/MM) simulations on a dimer of subunits A and C were performed to further evaluate the flexibility of the substrate in the open- and closed-form subunits. These simulations showed that the substrate residing in the closed subunit had significantly lower B-factors than the simulated substrate in the open subunit (Supplementary Videos 1a-e, 2a-e). In the open subunit, the acyl moiety was found to have a high degree of flexibility within the active site (Supplementary Figure S8, Supplementary Videos 1&2). Taken together, both the crystallographic analyses and QM/MM simulations are agreeing with the idea that the closed-, but not the open-form subunits represent the catalytically competent subunits. Furthermore, the organization of the ternary K. setae ECR complex into catalytically competent and incompetent subunits, suggests that the enzyme operates with half-site reactivity, in which active sites alternate during catalysis14–16.
Product release returns ECR back into a symmetric homotetramer
Following catalysis, K. setae ECR has to release the product and oxidized cofactor. In order to understand the structural basis for this part of the catalytic cycle, the protein was co-crystallized with NADP+ and its structure was solved at 1.8 Å resolution. The enzyme transitioned back to the D2 symmetry of the apo enzyme, with four conformationally identical subunits (Supplementary Figures S1&S2) all containing bound NADP+ molecule (Figure 4a). Compared to the homotetrameric apo enzyme in which helix α6 and loop 88-93 of the catalytic domain stabilize the phosphate backbone of the NADP+ molecule(Figure 4b), the corresponding helix and loop moved closer to each other (Figure 4a), similar to the ternary complex (Figure 3d), which leaves the homotetramer in an “all-closed” state. Notably, while the NADP+ binding mode is comparable between the ternary complex and the complex with NADP+ alone, the B-factors of the latter are larger than those in the ternary complexes (Figure 1b), indicating that in the NADP+ bound enzyme the atoms become in general more mobile, which may be advantageous for discharging the oxidized, spent cofactor. The configuration of the nicotinamide group of NADP+ in the NADP+ bound structure is similar to those of the NADPH in the cofactor-bound, closed form subunits (Figure 4c).
Swing motion of the peripheral catalytic domain during catalysis
Comparison of the high-resolution structures of apo, NADPH-bound, NADPH/butyryl-CoA-bound and NADP+-bound K. setae tetrameric ECRs suggests that there are coordinated motions of the catalytic domains which are peripheral to the more rigid oligomerization domains (Supplementary videos 3 & 4). The apo-form and NADP+-bound form show 4 equivalent subunits, while NADPH-bound and NADPH/butyryl-CoA-bound forms divide into two groups of open-(i.e., catalytically incompetent) and closed-form (i.e., catalytically competent) subunit pairs (Figure 1a). In order to understand how these global structural changes affect catalysis, a principal component analysis (PCA) was used to extract major structural differences among the four structures.
The PCA revealed 8 major contributions based on their singular values (Supplementary Figure S9a) and the movement of the catalytic domains described above was the strongest, followed by other less significant structural changes. The 8 PCA components were used to analyze the contributions of PC1 to PC8 to the structural changes between each of the 4 structures and the average tetramer structure. This analysis showed that the first three PCA components, PC1-3 (Supplementary Videos 5a,b,&c), can explain more than 50% of the structural changes (Supplementary Videos 5d). PC1 shows that the peripheral catalytic domains are coupled and swing up and down on either side of the central oligomerization domains; PC2 shows that each of the catalytic domains moves away from its partner catalytic domain, and finally the catalytic domains undergo subtle tilt motions in PC3. The deviations of the NADPH-bound and NADPH/butyryl-CoA-bound structures from the average structure is explained mainly by PC1, the NADP+-bound form by PC2, and the apo form by PC3 (Supplementary Figure S9b).
Communication between pairs of dimers promotes catalysis
Given the coordinated motions of the peripheral catalytic domains during the catalysis, how is catalysis synchronized across the enzyme complex? One intriguing aspect of the ECR tetramer structures is that the catalytic domains share a common interface of 1636 Å2 between the pairs of dimers (between the catalytic domains of A and D, and those of B and C, Figure 5a&b) suggesting that they move together as rigid bodies. A comparison of the overall domain movements between the apo-form and NADPH/product-bound ternary complex shows that the enzyme tetramer changes from the homotetrameric apo state to the open and closed-form subunit dimer pairs (Figure 5a). Upon binding of CoA and NADPH, neighboring catalytic domains rotate, which couples the widening of one active site with the compression of the other active site across the pair of dimers (Figure 5a). Thus, it seems to be a direct consequence of the rigid structure of the inter-catalytic domain interface that the enzyme will adopt two distinct conformational states in each dimer when it becomes catalytically active.
What are the molecular determinants that synchronize catalysis across the pair of open/closed form dimers? The inter-catalytic domain interface is mostly hydrophobic, but also features some electrostatic interactions (Figure 5b). Most notable are Asn218 of one subunit of one dimer that forms a hydrogen bond to Asn157 of the adjacent subunit of the other dimer, as well as Glu151 of one subunit of one dimer that forms hydrogen bonds to the main chain nitrogen of Asn133 (and/or Ala134) of the neighboring subunit of the other dimer (Figure 5b). Multiple sequence alignment showed that Glu151, Ans218 and Asn157 are highly conserved in ECRs from primary (i.e., central carbon) metabolism, which show faster CO2-fixation kinetics (average kcat 28 s−1), but not in ECRs from secondary metabolism (Figure 5c), raising an interesting question about their roles in catalysis (average kcat 1.2 s−1).
Mutation of these residues, that are more than 20 Å away from the active site, dramatically affected the kinetic parameters of K. setae ECR (Table 1). In the E151D variant, the kcat value was fivefold decreased, demonstrating that weakening the interaction of catalytic domains has profound effects on the catalytic rate of the enzyme. Mutations that targeted the asparagine interaction network showed also strong effects on the catalytic rate, but did additionally affect KM of substrate binding. Most notable were variants N218E single and E151D/N157E/N218E triple variants that decreased the kcat by more than 25- and 100-fold, respectively, highlighting that communication at the interface of the catalytic domains of the pair of dimers is an important determinant of the catalytic rate in K. setae ECR.
To exclude that the overall structure of the complex was not altered through these mutations, we used gel filtration, as well as native gel analysis to analyze the oligomerization state of the different enzyme variants (Supplementary Figure S3a). Gel filtration assays were performed under the same conditions as our kinetic measurements and showed that all mutant enzyme variants kept their tetrameric form. Only native gel analysis, which was performed under more disruptive conditions, showed slightly increase in the dimer and monomer fractions, indicating that interface interactions are weakened in these variants (Supplementary S3b). Overall, our mutational and kinetic data supports the hypothesis that synchronization of catalytic domains strongly contributes to catalytic rate and is conferred through hydrogen-bond network at the interface of the pair of dimers.
Shared substrate binding within dimers is important for catalysis
While our study on the interface between catalytic domains explained how communication is conferred between different dimers through the strong coupling of the two catalytic domains, each from two dimers (inter-dimer interaction, Figure 5a), it did not explain how catalysis is synchronized between the open- and closed-form subunits within the same dimer. We turned our attention back to the fact that CoA substrate binding is shared between the open- and closed-form subunits in each dimer through the adenine binding pocket (intra-dimer interaction, Figure 5a).
To understand the role of substrate adenine binding in catalysis, we characterized the kinetics of different single, double and triple mutant variants of the adenine binding site (Figure 3g and Table 2). Mutations in the adenine binding pocket, and in particular of Arg303, strongly increased the KM of the CoA substrate as expected, but also decreased the apparent kcat of the enzyme by a factor of two to three. Notably, a comparable decrease in kcat was also observed in the wild-type enzyme when we used crotonyl-panthetheine, a truncated substrate that lacks the adenosine moiety and cannot bind to the adenine binding pocket. This indicated that shared cofactor binding between neighboring subunit is important for efficient catalysis, but did not provide a conclusive answer, how catalysis is synchronized across the subunits.
We noticed that the substrate adenine binding pocket is directly followed by a loop that carries a lysine residue (Lys332), which interacts with the active site of the neighboring subunit. Lys332 residue from the open-form subunit engages in a hydrogen bonding network with the nicotine amide group of the NADPH cofactor bound to the closed-form subunit through Gln165 and His365 of the neighboring subunit (Figure 5d). These interactions are not observed in the active site of the open-form subunit (Figure 5d), raising the question whether the hydrogen bonding network connected to the adenine binding pocket might be important for catalysis. In K332A and Q165A variants, kcat was decreased two-to three-fold (Table 2). When we tested these variants with crotonyl-panthetheine, we saw much to our surprise that catalytic activity in the K332A variant was reduced by more than two orders of magnitude, leaving us with the suggestion that adenine binding together with the loop carrying Lys332 are important to synchronize catalysis between the two subunits within the dimer. Together with the inter-domain coupling, this intra-dimer synchronization drive fast CO2-fixation by K. setae ECR.
CONCLUSIONS
Our structural studies of K. setae ECR revealed unprecedented details on the functional organization of nature’s most efficient and fastest CO2-fixing enzyme. During catalysis, the enzyme complex differentiates into distinct functional subunits. Binding of NADPH cofactor and substrates forces the homotetrameric apo enzyme into a dimer of dimers in which each dimer is constituted of an open- and a closed-form subunits. In the closed-form subunits the NADPH cofactor and CoA substrate are aligned with each other, suggesting that this is the catalytically competent state. The open-form subunits bind cofactor and the adenine rings of the substrates but the rest of the acyl-CoA substrate remains flexible and invisible in the active site. Thus, the open-subunit active sites seem to represent a catalytically incompetent state that is pre-organized for a next round of catalysis. Altogether, this structural reorganization of ECR strongly supports the idea that the enzyme operates with “half-site reactivity”, according to which catalysis is synchronized across the enzyme tetramer and alters between the open- and closed-form subunits to increase the overall catalytic efficiency of the complex14–17.
Interaction of the catalytic domains of neighboring subunits is crucial for efficient catalysis in K. setae ECR. Especially important is the interaction of catalytic domains between the pairs of dimers. As soon as this interaction is disturbed, the catalytic rate of the enzyme is severely diminished. This observation is consistent with theoretical and experimental data on half-site reactivity. Synchronization of the distant catalytic subunits can enhance the catalytic rate of enzymes several-fold18,19. Mutation of a single amino acid coupling the two catalytic sites of heptose isomerase GmhA reduced catalytic rate of GmhA to 6% of wild-type activity20. Escherichia coli thymidylate synthase is another example for an enzyme showing half-site reactivity21. Disturbing the interaction network in E. coli thymidylate synthase leads to a 400-fold decrease in kcat22,23, demonstrating that domain interactions are important factors in promoting enzyme catalysis24.
Besides the inter-dimer domain interaction, our study on ECR also suggests joint substrate-binding between neighboring subunits as another potentially important mechanism of fast synchronized catalysis. The binding of the adenine end of the CoA ester into a pocket in the neighboring subunit seems to be connected back via a hydrogen-bonding network to the active site of the subunit where the CoA ester originated. This provides the missing link of how catalysis might be synchronized between the open- and closed-form subunits within one dimer. Taken together an attractive model of continuous turnover scheme emerges explaining the overall fast catalytic cycle of ECR; two consecutive reaction cycles alternate aided by the coupled inter-dimer catalytic domain motions (Figure 5e). In the first cycle (right half of Figure 5e), the open-form subunits A and B receive two sets of substrate and cofactor molecules, while the closed subunits C and D finish the previous reaction cycle, and release the products and NADP+. As a result, the subunits A and B become closed and the subunits C and D switch to the open-subunit state. In the second cycle (left half of Figure 5e), subunits A and B perform the reaction and release the products and NADP+ becoming open subunits, and the C and D subunits switch to closed state by acquiring a new set of substrate and NADPH.
While structural and biochemical data indicate that K. setae ECR achieves high catalytic rates by synchronizing active sites, this might not necessarily be true for other ECRs. A differentiation into dimers of dimers was not observed in NADPH-bound or ternary structures of other ECRs so far (e.g., PDB: 4Y0K and 4A0S respectively, which share substantial amino acid identity) (Supplementary Figure S10). Another reason might be that not all ECRs might perform synchronized catalysis. Note that ECRs fall into two different classes. Primary ECRs that operate in central carbon metabolism and secondary ECRs that serve in secondary metabolism, where they provide extender units for the synthesis of polyketides. Whereas primary ECRs are under strong evolutionary pressure and show on average kcat values of 28 s−1 25, secondary ECRs are not selected for high catalytic rates, which is also reflected by the fact that they show an average kcat value of 1.2 s−1 25. Thus, it might be tempting to speculate that secondary ECRs are not selected for high turnover rates during catalysis and thus might not display synchronized “half-site reactivity”.
In summary, this work provides the first overall picture of the organization of the ECR homotetrameric complex. The observation of the differentiation of the apo tetramer into open and closed form subunits upon binding of NADPH seemed to have been made possible by room temperature data collection using the XFEL beam at SACLA, highlighting the power of ambient temperature crystallography to study larger scale motions in macromolecular crystals26,27. Further experiments using time-resolved X-ray crystallography at room temperature and mixing jets will be helpful to obtain a fully dynamic picture of the ECR complex during catalysis, which will be important to fill the gaps in the mechanistic understanding of nature’s most efficient CO2-fixing principle.
Data Availability
Coordinates of the four ECR structures have been deposited in the Protein Data Bank under accession codes, 6NA3 (apo), 6NA4 (Butryl-CoA/NADPH bound), 6NA5 (NADP+ bound), and 6NA6 (NADPH-bound).
Online Methods
Amplification & cloning of K. setae ECR
The K. setae enoyl-CoA carboxylase reductase (ECR) coding sequence was codon optimized using the E. coli codon frequency table, and synthesis constraints were removed using the Build Optimization Software Tools (BOOST) developed by DOE-Joint Genome Institute (JGI), USA1. Overlapping synthetic DNA fragments were obtained from Thermo Fisher Scientific and cloned into the NdeI site of the pET16b vector (Novagen) by using the Gibson Assembly HiFi kit (SGI-DNA). The resulted colonies were sequence verified by the PacBio sequencing platform.
Site-directed Mutagenesis of K. setae ECR
Mutations are introduced by the similar methods as described in the previous section. Two fragments flanking the mutagenesis site were amplified and the Gibson assembly was performed as described above. Below is the FASTA sequence of the ECR protein and list of primers that we used to introduce catalytic site single mutations Y328F, R303K, and K296A, and the triple mutant K296A/R303K/Y328F.
tr|E4N096|E4N096_KITSK Putative crotonyl-CoA reductase OS= Kitasatospora setae
MQEILDAILSGDAASADYAALALPESYRAVTLHKGEERMFDGLASRDKDPRKSLHLDDVP LPELGPGEALVAVMASSVNYNTVWSSIFEPVSTFGFLERYGRLSPLTARHDLPYHVLGSD LAGVVLRTGAGVNAWKPGDEVVAHCLSVELESPDGHNDTMMDPEQRIWGFETNFGGLAQL ALVKTNQLLPKPKHLTWEEAASPGLVNSTAYRQLVSRNGAGLKQGDNVLIWGASGGLGSY ATQYALAGGATPICVVSSPRKADICRAMGAEAIIDRSAEGYRFWKDEHHQDPREWKRLGG KIREFTGGEDVDIVFEHPGRETFGASVYVTRKGGTIVTCASTSGYMHQYDNRYLWMSLKR IVGSHFANYREAFEANRLVAKGKIHPTLSKVYALEETGQAALDVHHNKHQGKVGVLCLAP REGLGVTDPELRSKHLTKINAFRNV
Single mutations were introduced with the following *_F&*_R primer pairs:
E4N096_DMP_064_Y328F_F: catccgtgttcgtgacccgcaaaggtggcactatcg
E4N096_DMP_064_Y328F_R: gcgggtcacgaacacggatgcaccgaaggtttcgcg
E4N096_DMP_064_R303K_F: ggtggcaaaatcaaggaattcaccggtggggaagacgtgg
E4N096_DMP_064_R303K_R: aattccttgattttgccacccagacgtttccactcacg
E4N096_DMP_064_K296A_F: agtgggcccgtctgggtggcaaaatccgtgaattcaccg
E4N096_DMP_064_K296A_R: ccagacgggcccactcacgcgggtcttggtggtgttcg
The double and triple mutants were introduced in the following order: By using Y328F plasmid we introduced R303K mutation to generate double mutant Y328F/R303K
For the triple mutant we used the following special primers pair
E4N096_DMP_064_triple_F: agtgggcccgtctgggtggcaaaatcaaggaattcaccg
E4N096_DMP_064_triple_R: ccagacgggcccactcacgcgggtcttggtggtgttcg
Mutagenesis of K. setae ECR subunit interface residues
Two fragments franking the mutagenesis site were amplified and the Gibson assembly was performed as with the list of primers that we used to introduce various combinations of subunit interface single mutations N157E, N218E, E151D, E151R, E151K, E151L, E151I
Seven single mutants were introduced with the following *_F&*_R primer pairs:
ECR_N157E_F: ccggacggtcacgaagacactatgatggacccagagcagc
ECR_N157E_R: catcatagtgtcttcgtgaccgtccggagattccagttcaacagacagg
ECR_N218E_F: gctggtgtctcgtgaaggcgccggcctgaaacagggtgacaacg
ECR_N218E_R: caggccggcgccttcacgagacaccagctgacgataagcggtagagttaacg
ECR_E151D_F: gtctgttgaactggattctccggacggtcacaacgacactatgatgg
ECR_E151D_R: gaccgtccggagaatccagttcaacagacaggcagtgagcaaccacctcgtcacc
ECR_E151R_F: gtctgttgaactgaggtctccggacggtcacaacgacactatgatgg
ECR_E151R_R: gaccgtccggagacctcagttcaacagacaggcagtgagcaaccacctcgtcacc
ECR_E151K_F: gtctgttgaactgaagtctccggacggtcacaacgacactatgatgg
ECR_E151K_R: gaccgtccggagacttcagttcaacagacaggcagtgagcaaccacctcgtcacc
ECR_E151L_F: gtctgttgaactgctgtctccggacggtcacaacgacactatgatgg
ECR_E151L_R: gaccgtccggagacagcagttcaacagacaggcagtgagcaaccacctcgtcacc
ECR_E151I_F: gtctgttgaactgatctctccggacggtcacaacgacactatgatgg
ECR_E151I_R: gaccgtccggagagatcagttcaacagacaggcagtgagcaaccacctcgtcacc
Five triple mutants were obtained with the following *_F&*_R primer pairs respectively (third mutation varies):
Asn157Glu, Asn218Glu, Glu151Asp
ECR_N157E_N218E_E151D_F:
ctgtctgttgaactggattctccggacggtcacgaagacactatgatggacccagagcagcgcatctgg
ECR_N157E_N218E_E151D_R:
gtccatcatagtgtcttcgtgaccgtccggagaatccagttcaacagacaggcagtgagcaaccacctcg
Asn157Glu, Asn218Glu, Glu151Arg
ECR_N157E_N218E_E151R_F:
ctgtctgttgaactgaggtctccggacggtcacgaagacactatgatggacccagagcagcgcatctgg
ECR_N157E_N218E_E151R_R:
gtccatcatagtgtcttcgtgaccgtccggagacctcagttcaacagacaggcagtgagcaaccacctcg
Asn157Glu, Asn218Glu, Glu151Lys
ECR_N157E_N218E_E151K_F:
ctgtctgttgaactgaagtctccggacggtcacgaagacactatgatggacccagagcagcgcatctgg
ECR_N157E_N218E_E151K_R:
gtccatcatagtgtcttcgtgaccgtccggagacttcagttcaacagacaggcagtgagcaaccacctcg
Asn157Glu, Asn218Glu, Glu151Leu
ECR_N157E_N218E_E151L_F:
ctgtctgttgaactgctgtctccggacggtcacgaagacactatgatggacccagagcagcgcatctgg
ECR_N157E_N218E_E151L_R:
gtccatcatagtgtcttcgtgaccgtccggagacagcagttcaacagacaggcagtgagcaaccacctcg
Asn157Glu, Asn218Glu, Glu151Ile
ECR_N157E_N218E_E151I_F:
ctgtctgttgaactgatctctccggacggtcacgaagacactatgatggacccagagcagcgcatctgg
ECR_N157E_N218E_E151I_R:
gtccatcatagtgtcttcgtgaccgtccggagagatcagttcaacagacaggcagtgagcaaccacctcg
Mutagenesis of Adenine binding residues and intra-dimer communication residues
Variants of the K. Setae Ecr were generated with the QuikChange® Site-Directed Mutagenesis Kit (Stratagene, La Jolla, USA) using 60 ng of template plasmid and the following forward and reverse primer Pairs:
Cell lysis, protein purification, and characterization
The cells were harvested by centrifugation (3000 rpm, 30 min) and the cell pellet was pooled. The pellet was resuspended in a lysis buffer containing 50 mm Tris-HCl pH 8.5, 1 M NaCl, 5% glycerol Supplemented with 100µl Triton x100 per 100ml of final buffer volume (Sigma-Aldrich). The suspension was sonicated at 50% amplitude for 30 seconds three times. Immediately after the lysis, the suspension was ultra-centrifuged at 33,000 rpm for 40 minutes at 4°C.
The soluble fraction was pooled and was applied to a 10 ml Ni-NTA column and purified using an AKTA prime FPLC setup. The column was washed with 2 column volumes of HisA loading buffer (50 mM Tris-HCl pH 8.5, 300 mM NaCl, 10 mM imidazole) for equilibration. Preliminary attempts of His-tag purification were unsuccessful since the protein would precipitate out of solution during application to the column. This was remedied by adding 1 M L-proline (Sigma-Aldrich) to the lysis and HisB elution buffers to ensure the protein remains soluble. The soluble portion was then applied to Ni-NTA column, and then eluted using HisB elution buffer containing 50 mM Tris-HCl pH 8.5, 300 mM NaCl, 500 mM imidazole. The eluted fractions were collected on a fraction collector, and their purities were analyzed by SDS-PAGE, and pure fractions were pooled and concentrated to 10 mg/ml using MilliPore Amicon Ultra 30KDa molecular-weight cutoff concentrators.
Determination of the oligomeric state of KsCcr
Oligomeric state of K. setae ECR was determined by analytical size-exclusion chromatography. 260 µl containing 500 µg of purified protein were injected into a pre-equilibrated S200 INCREASE 10-300GL (GE Healthcare) column. Runs were performed using a 100 mM KH2PO4 pH=8.0 buffer at a flow of 0.75 ml/min. Protein size was determined by comparing the obtained retention volumes (RV) with a Gel filtration standard protein mixture (BioRad).
Spectrophotometric Enzyme assays
Assays were performed on a Cary-60 UV/Vis spectrophotometer (Agilent) at 30°C using quartz cuvettes (1 or 10 mm path length; Hellma). Reactions contained 20 µg/ml carbonic anhydrase and were performed in 100 mM K2HPO4 pH = 8.0. Kinetic parameters for one substrate were determined by varying its concentration while the others were kept constant at 10 times their KM value. Reaction procedure was monitored by following the oxidation of NADPH at 365 nm (εNADPH,365nm = 3.33 M−1 cm−1). Each concentration was measured in triplicates and the obtained curves were fit using GraphPad Prism 8. Hyperbolic curves were fit to the Michaelis-Menten equation to obtain apparent kcat and KM values. For mutants revealing substrate inhibition, the data was fit to v0= (VMax [S])/(KM+ [S] ((1+[S])/Ki))).
Chemical Synthesis of CoA-esters
Crotonic Anhydride, Carbonic anhydrase from bovine erythrocytes, 1,1-Carbonyldiimidazole (CDI) and 4-dimethylaminopyridine (DMAP) were purchased from Sigma Aldrich AG, Coenzyme A trilithium from Roche Diagnostics, NADPH Na4 (98%) and pyridine from Carl Roth GmbH. Solvents and salts were all analytical grade or better. Crotonyl-CoA was synthesized as previously reported2. Briefly 200 mg of CoA trilithium salt were dissolved in 4 ml of 0.4 M KHCO3 and stirred on ice for 45 min. After addition of 64 μl of crotonic anhydride the reaction procedure was tested by mixing 5 μl of reaction mixture with 20 μl of an aqueous solution of DTNB (5,5’-dithio-bis-[2-nitrobenzoic acid]). Crotonyl-CoA was purified by preparative RFLC/MS over a Gemini 10 μm NX-C18 110 Å, 100 x 21.2 mm, AXIA packed column (phenomenex) using a methanol gradient from 5% to 35% over 15 min with 25 mM ammonium formate pH = 8.1 (Buffer 8.1) as the aqueous phase. Fractions containing the product were pooled, lyophilized and stored at −20°C.
Synthesis of crotonyl-pantetheine (3) was performed according to scheme 1 as previously reported2.
Pantetheine 1 (0.50 g, 1.57 mmol), DMAP (0.02 g, 0.19 mmol) and crotonic anhydride (0.50 ml, 3.37 mmol) in pyridine (12.5 ml) were stirred for 15 h at 23° C then 1 h at 50 °C. Pyridine was removed under reduced pressure and the product dissolved in saturated aqueous NaHCO3 (1 ml) and water (1 ml). The aqueous phase was extracted with CH2Cl2 (3× 5 ml), dried over MgSO4, filtered and the solvent removed under reduced pressure. The obtained product was purified over FC (SiO2; EtOAc/hexane, 2:3 → EtOAc) to afford 2. 2 then stirred in Water/EtOH/FA, 1:1:1 for 30 min at 23 °C. After completion the solution was lyophilized, the solid dissolved in 0.5 % aqueous TFA and then purified with HPLC (C18; 25mM Ammonium formate pH = 8.1/MeOH, 5% → 95%) and lyophilized to yield 3 as a transparent thick oil. For use in assays the compound was resuspended in water and stored at −20°C if not used.
Analysis of Ethylmalonyl-CoA stability
Reactions to measure the stability of ethylmalonyl-CoA under crystallization conditions were performed in 200 mM TrisHCl pH = 7.5, 20% Polyacrylic acid sodium salt 5100 at 19°C. Reactions contained 600 μM Ethylmalonyl-CoA, NADP+ µM and 1 µM KsECR WT. Samples were quenched at different time points using 50% formic acid and spinned at 17’000 g for 10 min to precipitate the protein. The reaction was diluted 10 times into 5% methanol/Buffer 8.1 and analyzed by UHPLC over a Sonoma C18(2), 3 µm 100 Å, 100 x 2.1 mm using a 5 to 45% methanol gradient over 14.5 min.
Crystallization of K.setae ECR complexes
72-well sitting-drop crystallization trays (Terasaki) were set up and screened against a library of various crystallization conditions (Molecular Dimensions, Hampton). Each crystallization well contained 0.77 µl of 10 mg/ml K. setae ECR protein kept in 500 mM Imidazole, 300mM NaCl, 1M proline and TRIS-HCl pH 8.5 mixed with 0.77 µL of the various crystallization buffers. Each well was sealed with 16.6 µL of 100% paraffin oil (Hampton Research) to slow the crystallization process. Crystals of apo ECR protein were observed in various morphologies after 24 hours of incubation. The initial crystallization conditions were from various MIDAS, Crystal Screen, and PGA-LM screening conditions (Molecular Dimensions, Hampton Research). The apo ECR was crystallied from a solution containing 100 mM TRIS pH 8.0 and 20% w/v poly (acrylic acid sodium salt) 5100 and resulted in 30-micron plate-like crystals. It is important to note that all structures were solved using this condition as basis, with addition reagents as needed. The binary and ternary ECR complexes were co-crystallized with final concentration of 5 mM of each respective ligand and cofactor with a protein concentration of 10 mg/ml. Alternative crystallization conditions were used either to obtain larger crystals for higher resolution synchrotron structures or higher microcrystal density for SFX experiments. For the crystallization of K. setae ECR-butyrylCoA-NADPH ternary complex, the crystallization condition contained 17% w/v PEG 10000, 100 mM BisTris-HCl pH 7.5, 100 mM ammonium acetate and resulted in 50-micron plate-like crystals. For K. setae ECR-NADP+ binary complex, the crystallization solution contained 0.2 M ammonium formate, 10% (w/v) polyvinylpyrrolidone, 20% (w/v) PEG 4000 and resulted in 50-micron plate-like crystals. No further seeding was required for any of the synchrotron structures, and crystals were harvested after 30 minutes incubation with 30% (v/v) glycerol as a cryoprotectant. For the SFX experiments, neither of the synchrotron crystallization conditions of crystals would be sufficient due to size limitations and an optimal crystal density of 109 to 1011 crystals/ml could not be obtained. To test various crystal conditions, a batch method was employed with equal parts of protein and crystal condition to see if increased crystal densities could be achieved in 15 mL Corning conical falcon tubes. Initial tests were total volume of 1 ml (0.5 ml 10 mg/ml protein and 0.5 ml crystal condition) and incubated for 48 hours. From the tubes with crystals present, the 1 ml crystal slurry was used to seed a 10 ml total crystallization solution. The best crystals were obtained in final 10 ml sample solution consisted of the 1 ml seed crystal slurry solution, 4.5 ml of 10 mg/ml protein solution and 4.5 ml of crystallization buffer containing 0.03M Magnesium chloride hexahydrate, 0.03M Calcium chloride dihydrate, 0.05M imidazole, 0.05M MES-KOH pH 6.5, 15% v/v glycerol and 15% v/v PEG. Prior to sample injection, the crystals were filtered using a 20-micron nylon mesh filter to separate the contaminant of large crystals from the smaller ones (Millipore).
Data collection, processing and structure determination
For the apo, ternary, and NADP+ binary complex structures, the crystals were flash cooled in liquid nitrogen. The apo (1.8 Å) and NADP+ complex (1.75 Å) diffraction datasets were collected at 100 K on Beamline 23ID-B, the Advanced Photon Source, Argonne National Laboratory (Argonne, Illinois, USA), equipped with an Eiger 16M detector. The butyryl-CoA ternary complex (1.7 Å) diffraction dataset was collected at 100 K on Beamline 12-2 at the Stanford Synchrotron Radiation Lightsource, SLAC National Accelerator Laboratory (Menlo Park, California, USA), equipped with a Dectris Pilatus 6M detector. The K. setae apo crystals belonged to the space group P212121 with unit cell dimensions a= 78.1 Å, b= 153.0Å, c= 202.7 Å and α=β=γ=90°. The K. setae ternary complex crystals belonged to the space group P21 with unit cell dimensions of a = 109.3 Å, b = 78.8 Å, c = 138.8 Å and α=90°, β=108.1°, γ=90°. The K. setae NADP+ complex crystals belonged to the space group P212121 with unit cell dimensions of a = 77.0Å, b = 146.7 Å, c = 200.2 Å and α=β= γ=90°. The K. setae ECR-NADPH binary complex was determined using serial femtosecond X-ray crystallography (SFX) at an X-ray Free Electron Laser (XFEL) and was carried out on May 2017 at SACLA beamline 3 (Hyogo, Japan) (Proposal number 2017A8055)3. The SALCA beam had a pulse duration of 10 fs. The photon energy was 10kEV. The in air concentric Electrokinetic Microfluidic Sample Holder (coMESH) injector4 installed at DAPHNIS5 chamber was used to introduce samples suspended in mother liquor to the 10 fs-long X-ray pulses. X-ray diffraction data was recorded by using the multiport CCD (MPCCD)6 detector. Data analysis was performed on the SACLA High Performance Computing Cluster consisting of several steps of parameter optimization. Diffraction images were collected with consistent experimental parameters (attenuation, transmission, detector distance etc.) during one 12-hour shift. Crystal hits were identified with the program Cheetah7. The raw data were processed with CrystFEL’s indexamajig against given cell parameters of the K. setae NADPH (XFEL) complex microcrystals belonging to the space group P21 with unit cell dimensions of a = 109.8 Å, b = 78.1 Å, c = 138.9 Å and α=90°, β=107.8°, γ=90°.
The data processing for synchrotron structures were carried out using autoXDS and scaling was done with XSCALE8,9. A set of 5% of randomly chosen reflections were set aside for the calculation of the free R factor (Rfree). The apo structure was solved using by PHENIX10,11 and PHASER12,13 molecular replacement program. Initial search model for molecular replacement is generated by using SWISS-MODEL14 server against an unpublished CCR structure of a putative crotonyl-CoA carboxylase/reductase (PDB code 4GI2, deposited by S. Weidenweber, T.J. Erb, U. Ermler). The K. setae apo structure served as the model for solving the binary and ternary-complex synchrotron structures and also SFX structure. This resulted in four monomers in the asymmetric unit. The refinement was carried out using PHENIX refinement, utilizing automatically generated TLS groups based on the structure and ordered solvent to place the water molecules15,16. Following the first round of refinement, the structure was manually adjusted to the electron density and waters were added using COOT at one sigma cutoff 17,18. The K. setae/NADP+ complex also shares the same space group as the apo form, and was solved directly using Phenix molecular replacement10,11. The NADP+ structure and restraint files were taken from previously solved CCR/NADP+ complexes (PDB: 4Y0K).
MD and QM/MM studies of the open and closed subunits of ECR
The flexibility of NADPH in the open and closed cavity were studied with a binary complex using the CHARMM22 force field19,20 and parameters from Pavelites et al. for NADPH21 in 200 ns explicit solvent simulations (TIP3P) at 298 K and 1 atm with the AMBER16 software package22 (dt = 2 fs, tau = 1 ps, PME cutoff = 8.0 Å, SHAKE). Additionally, in three independent 200 ns simulations NADPH was positioned in the open cavity in the conformation observed in the closed cavity to test if these conformations are also visited in the open form.
To study crotonyl-CoA binding in the ternary complex we first extracted a dimer with one open- and another closed-subunits (subunit A & C) from the butyryl-CoA/NADPH ECR ternary X-Ray structure. We added the unresolved butyryl-CoA molecule to the open subunit aligning the protein chains from the closed subunit on the open one and shifting the butyryl-CoA coordinates from the closed subunit. The butyryl-CoA molecules were then modified to obtain the substrate crotonyl-CoA deleting the two hydrogen atoms. The resulting dimer consisted of one closed and one open subunit each with NADPH and one crotonyl-CoA molecule. The system was solvated and equilibrated (500ps NVT, 5ns NPT, 100ns NVT) as described above and substrate and NADPH were restrained to their initial configuration to relax the protein and the solvent. From these equilibrated configurations five structures were randomly extracted to study the behavior of the cofactor and substrate molecules. Five trajectories of 2 ns each for the closed and open cavity were performed with the QM/MM method using the DFTB3 Hamiltonian23 and the 3ob parameter24,25 set to describe NADPH and the crotonyl fragment. An electrostatic embedding using the link atom method at 298K and 1 atm was used together with a time step of 1 fs in AMBER16 software package. Parameters for the CoA fragment of the substrate were taken from Aleksandrov et al26.
Acknowledgements
Authors acknowledge Takanori Nakane from University of Tokyo for his help with calibration of XFEL data from SACLA, RIKEN, Japan. We would like to thank Eriko Nango, Rie Tanaka and RIKEN SPring-8 Center for their help with data collection at SACLA, RIKEN, Japan. HD acknowledges support from NSF Science and Technology Center grant NSF-1231306 (Biology with X-ray Lasers, BioXFEL). The XFEL experiments were performed at BL3 of SACLA with the approval of the Japan Synchrotron Radiation Research Institute (JASRI) (Proposal No. 2017A8055). The authors thank the beamline staff of Structural Molecular Biology Group, SSRL, SLAC and GM/CA CAT, Advance Photon Source, ANL for assistance on data collection. YR, RGS, MSH, and BH were supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences under Contract No. DE-AC02-76SF00515. TJE and GS received support from the Max Planck Society, the European Research Council (ERC 637675 ‘SYBORG’), and the U.S. Department of Energy Joint Genome Institute a DOE Office of Science User Facility under Contract No. DE-AC02-05CH11231. DAS and EVM thank the Max-Planck Society for funding as a Max-Planck-Partner group. HD, and SW are supported by DOE Office of Science, Biological Environmental Research, and National Institute of Health, NIGMS. CG and SW were supported by National Science Foundation, Major Research Instrument grant. HD and SW’s work was partially supported by Stanford PRECOURT Institute. Gregory M. Stewart of SLAC and Moe Wakatsuki for graphics work.