{"paper_id":"31add491-c253-421c-899e-be25548568ad","body_text":"1 \n \n \nConformational modulation of a mobile loop controls catalysis \nin the (ba)8-barrel enzyme of histidine biosynthesis HisF \n \nEnrico Hupfeld,1, # Sandra Schlee,1,# Jan Philip Wurm,1 Chitra Rajendran,1 Dariia Yehorova,2 \nEva Vos,2 Dinesh Ravindra Raju,2 Shina Caroline Lynn Kamerlin,2* Remco Sprangers1* and \nReinhard Sterner1* \n \n1Institute of Biophysics and Physical Biochemistry, Regensburg Center for Biochemistry, \nUniversity of Regensburg, Universitätsstrasse 31, 93053 Regensburg, Germany \n \n2School of Chemistry and Biochemistry, Georgia Institute of Technology, 901 Atlantic Drive \nNW, Atlanta, GA 30318. \n \n#These two authors contributed equally \n*Corresponding authors: \nskamerlin3@gatech.edu  \nRemco.Sprangers@ur.de \nReinhard.sterner@ur.de \n \n \n \n \n \n \n \n \n \n \n \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n2 \n \nABSTRACT \nThe overall significance of loop motions for enzymatic activity is generally accepted. However, \nit has largely remained unclear whether and how such motions can control different steps of \ncatalysis. We have studied this problem on the example of the mobile active site β 1α1-loop \n(loop1) of the ( ba)8-barrel enzyme HisF, which is the cyclase subunit of imidazole glycerol \nphosphate synthase. Loop1 variants containing single mutations of conserved amino acids \nshowed drastically reduced rates for the turnover of the substrates N´-[(5´-phosphoribulosyl) \nformimino]-5-aminoimidazole-4-carboxamide ribonucleotide  (PrFAR) and ammonia to the \nproducts imidazole glycerol phosphate (ImGP) and 5-aminoimidazole-4-carboxamide-ribotide \n(AICAR). A comprehensive mechanistic analysis including stopped -flow kinetics , X -ray \ncrystallography, NMR spectroscopy, and molecular dynamics simulations detected three \nconformations of loop1 (open, detached, closed) whose populations differed between wild-type \nHisF and functionally affected loop1 variants . T ransient stopped -flow kinetic experiments \ndemonstrated that wt -HisF binds PrFAR by an induced -fit mechanism whereas catalytically \nimpaired loop1 variants bind PrFAR by a simple two -state mechanism. Our findings suggest \nthat PrFAR-induced formation of the closed conformation of loop1 brings active site residues \nin a productive orientation for chemical turnover, which we show to be the rate-limiting step of \nHisF catalysis. After the cyclase reaction, the closed loop conformation is destabilized, which \nfavors the formation of detached and open conformations and hence facilitates the release of \nthe products ImGP and AICAR. Our data demonstrate how different conformations of active \nsite loops contribute to different catalytic steps, a finding that is presumably of broad relevance \nfor the reaction mechanisms of (ba)8-barrel enzymes and beyond. \n \n \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n3 \n \nINTRODUCTION \nEnzymes perform reactions with remarkable catalytic efficiency, selectivity and specificity and \ntheir function is closely linked to their molecular motions. 1-3 Along these lines, the process of \ncatalysis typically involves movements of residues in the active site of the enzyme during \nsubstrate binding and product release. These steps include motions of single residues as well \nas opening and closing of loop regions or entire lid domains. 4-5 Such ligand -driven \nconformational changes are very well documented6-7 and are grouped under the umbrella term \n“induced-fit motions”, stating that binding of the substrate leads to the transition of non -\nproductive and often poorly defined active site conformations into a single well -defined \nconformation that is complementary  to the reaction transition state. 8-9 Moreover, it is \nincreasingly recognized that enzyme conformational fluctuations enable the sampling of high-\nenergy intermediates or conformational sub -states along the enzyme reaction coordinate. \nExperimental evidence continues to indicate that in many ca ses, the catalytic efficiency of \nenzymes is directly tied to the rate of conformational transitions into such sub -states.10-13 \nNevertheless, the direct role of enzyme motions in accelerating the individual states of the \ncatalytic reaction is still under debate.14-17  \nThe importance of mobile loops as critical participants in substrate binding and regulation \nof enzyme activity and specificity is reflected in natural enzyme evolution, where sequence \nchanges are frequently localized at loop regions and the associated modi fications of loop \nconformational plasticity have contributed to diversification of various enzyme families. 18-19 In \naddition, it has become increasingly clear that loop mobility needs to be considered in protein \nengineering approaches aiming at the development of more powerful enzyme catalysts.20-22  \nThe (βα)8- or TIM-barrel fold is the most abundant and most versatile fold of enzymes in \nnature. Around 10 % of all structurally characterized proteins contain at least one domain of \nthis fold.23 TIM-barrels catalyse a wide variety of unrelated reactions, covering 5 of the 7 EC \nclasses.24 The fold consists of eight alternations of β-strands and α-helices, the strands forming \na central β-barrel, which is surrounded by the α-helices. On the C-terminal face of the barrel, \nthe connecting βnαn-loops often contain residues involved in substrate binding and catalysis, \nwhile the αnβn+1-loops on the opposite N-terminal face of the barrel mainly contribute to protein \nstability.25 This separation of function and stability is probably one source of the folds versatility \nand has distinguished it as promising protein scaffold for enzyme design.26-28 As the βnαn-loops \ncan be easily modified or exchanged without compromising stability of the protein core, (β/α)8-\nbarrel enzymes are highly suitable for studying the relationship between loop dynamics and \ncatalysis.29-30  \n(βα)8-barrel enzymes, where ligand-induced loop motion has been shown to be a critical \ncomponent for catalytic activity, include triose phosphate isomerase (TIM) and a number of \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n4 \n \n(βα)8-barrel enzymes involved in tryptophan and histidine biosynthesis, namely TrpF, TrpC, \nHisA, and PriA.31-33 Both the position and length of the loops involved in catalysis, as well as \nthe nature of loop motions and their role in the catalytic mechanism, differ in the various (βα)8-\nbarrel enzymes. To extend the spectrum of loop conformational changes and their relation to \nthe catalytic mechanism in (βα)8-barrel enzymes more broadly, we have focussed on another \nenzyme of the histidine biosynthetic pathway, the cyclase subunit HisF of imidazole glycerol \nphosphate synthase (ImGPS) from the hyperthermophilic bacteriu m Thermotoga maritima. \nHisF catalyzes the conversion of N´-[(5´-phosphoribulosyl)formimino]-5-aminoimidazole-4-\ncarboxamide ribonucleotide (PrFAR) and ammonia that is provided by the glutaminase subunit \nHisH into imidazole glycerol phosphate (ImGP) and 5-aminoimidazole-4-carboxamide-ribotide \n(AICAR) (Figure 1A).34-37  \n \n \nFigure 1: Reaction catalyzed by HisF and crystal structures of HisF from Thermotoga maritima \nwith open/detached/closed loop1 conformations. (A) HisF catalyzes the conversion of PrFAR and \nammonia into ImGP and AICAR. The substrate analogue ProFAR binds to the active site but is not \ncleaved. (B)  In the apo state loop1 (residues R16-G30, orange) adopts an open conformation (PDB \nentry 1vh738). The catalytic residues D11 (general base, located within β-strand β1) and D130 (general \nacid, located within β-strand β5) are shown as blue sticks. (C) In some structures of HisF, loop1 is not \nresolved and presumably detached from the HisF core (PDB entry 3zr439). (C) In the presence of HisH \nand bound substrate analogue ProFAR (green sticks), loop1 adopts a closed conformation (PDB entry \n7ac840, chain E) and forms a β-sheet.  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n5 \n \nWhile ImGP is further processed to histidine, the second product AICAR is salvaged in \npurine biosynthesis. Prokaryotic ImGPS enzymes (including the ImGPS from Thermotoga \nmaritima) form heterodimeric bi-enzyme complexes that consist of the cyclase HisF and the \nglutaminase HisH subunit, which supplies ammonia by glutamine hydrolysis. 41-42 ImGPS is a \nwell-known model system for studies of allostery.40, 43-47 Binding of the substrate PrFAR or its \nanalogue, N´-[(5´-phosphoribosyl)formimino]-5-aminoimidazole-4-carboxamide-ribonucleotide \n(ProFAR)39 (Figure 1A ), results in the allosteric stimulation of the glutaminase reaction, \ninvolving a drastically increased rate of glutamine turnover by HisH. Importantly, under in vitro \nconditions the cyclase subunit HisF is able to catalyze the cyclase reaction in the absence of \nHisH, using externally added ammonium salts at basic pH values.37 The active site of HisF is \nlocated at the C-terminal face of the central β-barrel, where two conserved aspartate residues, \nD11 and D130, catalyse the cyclase reaction. 34 Numerous crystal structures of the isolated \nsubunit HisF38, 48 and of the HisH-HisF heterodimer in absence and presence of ligands39-40, 49 \nhave been determined. Based on the conformation of the loop that connects strand β1 with \nhelix α1 (loop1), these structures can be separated into an open conformation, where loop1 is \nflipped toward the outer a-helical barrel ring ( Figure 1B), a detached conformation, where \nloop1 is not visible in the electron density and presumably flexible ( Figure 1C) and a closed \nconformation, where loop1 closes over the active site (Figure 1D).   \nHere, we aimed at linking the different loop conformations in HisF with function. To that \nend we combined an extensive mutational analysis with steady-state and stopped-flow enzyme \nkinetics, NMR spectroscopy, X -ray crystallography, and molecular dynamics simulations. \nBased on that, we establish a model in which the function of loop1 is to close around the \nsubstrate to stabilize it  in the active site pocket such that efficient catalysis can take place.  \nMutants that fail to form a stably closed loop1 conformation are consequently considerably \nimpaired. Further, we demonstrate that mutations alter the distribution of open -detached-\nclosed conformational states between wild-type and mutated HisF variants, in agreement with \nprior computational work on triosephosphate isomerase 50, HisF/TrpF/PriA 32, and protein \ntyrosine phosphatases13, 51. The wide -spread adoption of such conformational fine -tuning of \nloop dynamics across unrelated enzymes suggests that such evolutionary conformational \nmodulation is a feature not just of (β/α)8-barrel enzymes, but of loopy enzymes more broadly.  \nRESULTS \nSubstitution of conserved amino acids in loop1 decrease catalytic activity \nThe influence of loop1 sequence on HisF function was assessed by mutational analysis. First, \na multiple sequence alignment (MSA) was compiled which revealed that most residues within \nloop1 are highly conserved ( Figure 2A ), indicating a function of this loop in the catalytic \nmechanism. To test this hypothesis, conserved residues were replaced by either alanine, \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n6 \n \nproline, or glycine. Whereas alanine substitutions should uncover effects based on \nelectrostatic or hydrophobic interactions, proline or glycine substitutions were introduced to \nreveal effects related to loop mobility. The assumption was that introduction of proline residues \nwould render loop1 more rigid  in the detached state, whereas inclusion of glycine residues \nwould increase loop1 mobility.  The resulting HisF loop1 variants were expressed in E. coli, \npurified, and characterized by steady-state enzyme kinetics. The determined turnover numbers \n(kcat) and Michaelis constants for PrFAR (KMPrFAR) are listed in Table 1.  \n \n \nFigure 2: Sequence conservation and mutational analysis of loop1. (A) Sequence logo (generated \nwith WebLogo3.6) based on a multiple sequence alignment (MSA) of about 1300 HisF sequences. \nResidues are numbered according to HisF from T. maritima. Mutated residues are marked with white \narrows. Residues whose mutation to Ala, Pro, or Gly resulted in a significant reduction of catalytic activity \nare marked with orange arrows. ( B) Detailed view of the open loop1 conformation (PDB ID: 1VH7 38). \nFunctionally important residues within loop1 are shown as orange sticks or spheres. Residue F38 is \nmarked in yellow sticks, the catalytic residues D11 and D130 are shown as blue sticks. (C) Detail view \nof the closed loop1 conformation (PDB ID: 7AC840, chain E). The bound substrate analogue ProFAR is \nshown in stick representation (colored by element).  \n \nTable 1: Steady state kinetic parameters of wt-HisF and loop1 variants at 25 °C. \n \n \nkcat \n(s-1) \nKMPrFAR \n(µM) \nkcat/KMPrFAR \n(M-1 s-1) \nwt 2.4 ± 0.2 4.5 ± 0.5 5.3 x 105 \nK19A 1.1 ± 0.1 6.1 ± 1.8 1.8 x 105 \nG20A 2.2 ± 0.1 x 10-2 2.1 ± 0.2 1.0 x 104 \nG20P n.d. n.d. - \nT21G 1.8 ± 0.1 x 10-2 5.0 ± 0.8 3.6 x 103 \nT21P n.d. n.d. - \nN22A 2.9 ± 0.2 x 10-1 8.4 ± 1.7 3.4 x 104 \nF23A 5.5 ± 0.4 x 10-3 6.8 ± 1.3 8.1 x 102 \nE24P n.d. n.d. - \nL26A 1.1 ± 0.03 2.0 ± 0.3 5.5 x 105 \nD28A 2.4 ± 0.1 4.5 ± 0.9 5.3 x 105 \nG30A 1.0 ± 0.1 x 10-1 3.7 ± 1.3 2.7 x 104 \nG30P n.d. n.d. - \nF38A 2.3 ± 0.1 3.5 ± 0.6 6.6 x 105 \nwt CouA 1.5 ± 0.1 4.1 ± 0.9 3.7 x 105 \nn. d.: no activity detectable. \nValues ± SE for kcat and KMPrFAR were determined by fitting with equations 1 and 2 of the mean for \ntechnical triplicates.  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n7 \n \nWhile some importance has been attributed to the residue corresponding to K19 in the \nhomologous yeast enzyme His752, we did not observe significant loss of activity for the K19A \nvariant. Likewise, the amino acid substitutions L26A and D28A did not affect catalytic activity. \nHowever, most of the substitutions (G20A, N22A, F23A, G30A, T21G) resulted in a significant \ndecrease of the kcat value, whereas the KMPrFAR values did not differ by more than two -fold in \ncomparison to wt-HisF. The most severe effects were observed for the proline substitutions \n(G20P, T21P, E24P and G30P) which caused a drop of catalytic activity belo w the detection \nlimit. The relatively constant KMPrFAR values and the dramatically decreased kcat values imply \nthat loop1 does not contribute significantly to the energetics of substrate binding, but rather \nplays a role for catalysis. As there are no indications that loop1 residues are directly involved \nin acid-base catalysis, loop1 must play an indirect role in substrate turnover. To obtain insights \ninto this role we have concentrated on three of the identified HisF variants that likely modulate \nthe conformational landscape of loop1. First, the HisF-F23A variant was selected to enhance \nthe flexibility of loop1. This variant will likely destabilize both the closed and open \nconformations as F23 stacks onto the PrFAR ligand in the closed state and interacts with F38A \nto form the open state (Figure 2B, C).  Second, the HisF-G20P variant was selected to restrict \nconformational flexibility of loop1 in the detached state . At the same time, this variant will \ndestabilize the closed conformation as residue 20 is part of a β -strand in that state ( Figure \n2C). Finally, the HisF-F38A variant was selected. It contains a mutation outside loop1 and is \nintended to destabilize the open conformation without effecting the closed conformation. \nWhereas the G20P and F23A substitutions decrease the kcat of wt-HisF by several orders of \nmagnitude, the F38A substitution has no effect on the steady-state catalytic parameters (Table \n1).  \nAmino acid substitutions shift the populations of the loop1 conformations \nTo assess whether the mutations have an influence on the conformation of  loop1 we \ndetermined the structures of the HisF-F23A and HisF-G20P variants by X-ray crystallography. \nIn the crystal, the HisF -G20P variant was found in the open conformation, similar as the wt -\nHisF protein (Figure S1A). For the HisF-F23A variant the electron density for residues 20-24 \nin loop 1 was lacking, indicating that the conformation of loop1 shifted from the open towards \nthe detached state (Figure S1B). \nTo complement these static crystal structures, we subjected the wt -HisF, as well as the \nvariants HisF -F23A, HisF -G20P, and HisF -F38A, to a limited proteolysis analysis. This \nexperiment should provide insights into the conformational mobility of the proteins, since \nprotease cleavage rates depend on the accessibility of the respective target 53-54 and it has \nbeen shown previously, that trypsin specifically cleaves HisF after R27 in loop1. 34 In our \nexperiments we observed that wt-HisF and the variant HisF-F38A are cleaved at similar rates. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n8 \n \nThe HisF-F23A variant, on the other hand, was cleaved faster, whereas the HisF-G20P variant \ndisplayed a reduced cleavage rate (Figure 3A). These results are in accordance with the static \nstructures that we solved and that suggested a shift towards the mobile detached state for the \nHisF-F23A variant and a stably formed open conformation for the HisF-G20P variant.  \nTo obtain direct information on the flexibility of loop1, we exploited NMR experiments. First, \nwe made use of heteronuclear NOE (hetNOE) measurements that probe structural fluctuations \non the ps-ns timescale.55 These fast motions result in {1H}-15N hetNOE values below 0.7. For \nthe wt-HisF protein we found that loop1 is the most dynamic loop in the protein. This implies \nthat loop1 predominantly occupies the detached state in solution. It should, however, be noted \nthat the open state of loop1 is also sampled as deletion of loo p1 results in chemical shift \nperturbations in residues that interact with loop1 in the open state. To assess the effect of the \nmutations on the conformation of loop1 we compared {1H}-15N hetNOE values of wt-HisF with \nthose of the variants HisF-F38A, HisF-F23A, and HisF -G20P (Figure 3B-D). This revealed \nthat the structural flexibility of loop1 is increased in HisF-F23A variant and, to a small degree, \nin variant HisF -F38A. These findings confirm that loop1 spends more time in the detached \nstate when the open state is destabilized by the F23A and F38A mutations. By contrast, the \nps-ns dynamics of loop1 in the HisF -G20P variant is slightly reduced, in agreement with an \nincreased stability of the open state and thus a shift in the conformation away from the \ndetached state.  \n \nFigure 3: Amino acid substitutions change flexibility and ps -ns dynamics of loop1. (A) Limited \nproteolysis assays monitoring the rates of trypsin cleavage at loop1 residue R27 for wt-HisF, HisF-F38A, \nHisF-F23A, and His-G20P. Cleavage patterns observed immediately after addition of trypsin (0 min) and \nafter incubation at 25°C for 20 min and  200 min are visualized by SDS polyacrylamide gel \nelectrophoresis (PAGE) analysis. (B, C, D) {1H}-15N hetNOE values of wt-HisF (blue) in comparison to \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n9 \n \nloop1 variants (orange) HisF -F38A (B), HisF-F23A (C) and HisF -G20P (D). Decreased values in the \nloop1 region (residues 19-30) of HisF-F38A and HisF-F23A in comparison to wt-HisF reveal increased \ndynamics on the ps to  ns timescale. The loop dynamics of the HisF -G20P is slightly decreased \ncompared to the wt-HisF. \n \nWe further supplemented our analysis with molecular dynamics (MD) simulations of wt -\nHisF and the HisF-G20P, HisF-F23A and HisF-F38A variants, in both the unliganded state and \nin complex with PrFAR. In the case of the unliganded enzyme, as there is no exper imental \nevidence for loop closure in this state, we initiated trajectories only from the loop1 -open \nconformation of the enzyme. However, in the case of the PrFAR-bound enzymes, we initiated \nsimulations from both the open and closed states of loop1 for completeness. \nFigure S2 shows the root mean square fluctuations (RMSF) of all Cα-atoms of HisF during \nMD simulations of the different systems studied. These reflect the flexibility of loop1 (residues \n19-30) in wt -HisF and how this is impacted by the mutations. This figure shows  only subtle \ndifferences in loop1 flexibility among the loop variants: however, given that the loop is  highly \nflexible in all variants, the relative flexibility of the loop will not necessarily change, although \nthere may be shifts within that ensemble between open, detached , and closed states . We \nfurther note that the large absolute value of the loop1 RMSF obtained in the simulations of the \nPrFAR-bound enzymes initiated from the loop1 closed conformation ( Figure S2B) is due to \nconformational adjustment of the loop to a new (but still closed) conformation (Figure S3), \nlikely due to the change in ligand from ProFAR present in the crystal structure ( Figure 1) to \nthe substrate PrFAR (see Materials and Methods). \n In order to explore the impact of mutations on loop1 flexibility in the different loop states in \nmore detail, we examined the relative mobilities of loop1 based on this RMSF analysis (Figure \n4). The mobility data is supplemented by a projection of loop1 motion in wt-HisF along the first \nprincipal component, PC1, from principal component analysis (PCA) of these MD simulations \nto illustrate the dominant dynamic motif. From this data, it can be seen that the relative mobility \nprofiles vary depending on enzyme variant in simulations initiated from the open conformation \nof loop1 (in both the unliganded and PrFAR bound states of the enzyme, see Figures 4A and \nB), with much more subtle differences in simulations initiated from the PrFAR -bound loop-\nclosed conformation (Figure 4C). This is due to the high mobility of loop1 in all variants, as our \nsimulations shift the loop towards a new closed conformation. We note also that in simulations \ninitiated from the closed state of loop1, we observe a clear monomodal distribution of mobilities \nin all variants, peaking towards the center of the loop. In contrast, loop mobility is more complex \n(and variant dependent) in simulations initiated from the loop1 open conformations, likely due \nto the loop changing shape as it samples both open and detached conformations. \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n10 \n \n \nFigure 4: Relative mobility of loop 1 during MD simulations. The mobilities were calculated from the \nRMSF of the loop1 C α-atoms, as outlined in the Supplemental Methods. Shown here are data from \nanalysis of simulations of the (A) unliganded simulations initialized from the loop1 open conformation, \nand of simulations of the PrFAR -bound enzymes initialized from the loop1 ( B) open and (C) closed \nconformations. For comparison, panels ( D-F) show projections of the first principal component, PC1, \nfrom principal component (PCA) analysis of these simulations (performed as described in the \nSupplemental Methods) onto representative structures of the ( D) open unliganded, (E) open PrFAR \nbound, and (F) closed PrFAR bound states of wt -HisF. The color gradient indicates the transition of \nloop1 along this principal component. \n \nFinally, to further analyze the flexibility of loop1, we constructed 2D histograms of  loop1 \nmotion as a function of the root mean square deviations (RMSD) of the C α-atoms of loop1 \nrelative to the closed conformation observed in the crystal structure of wild -type HisF/HisH in \ncomplex with ProFAR ( PDB ID: 7ac840, chain E and F ), and the  distance RMSD of all non-\ncovalent interactions in the loop1 open conformation of the loop (PDB ID: 1THF48) projected as \na single vector. The corresponding data is shown in Figures 5, alongside snapshots illustrating \nthe conformational space sampled by loop1 in each set of simulations, colored by C α-atom \nRMSF of loop1. From this data, it can be seen that in both unliganded and liganded simulations \n(Figure 5), we sample both open and detached state (the latter show up as a “smear” on the \nhistograms, as this state is very mobile). The relative population of these states is then shifted \nby the introduction of point mutations on the loop. In the case of the F38A and F23A variants, \nwe see a clear shift towards more detached states dominating our simulations, but not in the \ncase of the G20P variant. This shift is also illustrated in the enlarged 1D histograms of the \ndRMSD from the open state contacts (y -axis of the  2D plot), where the histogram of low \ndRMSD values is decreased for F23A and F38A and increased for G20P.  This is in agreement \nwith (and confirming) the observations from our {1H}-15N hetNOE experiments (Figure 3). \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n11 \n \nFigure 5: Conformational ensemble of loop1 during molecular dynamics simulations of \nunliganded and PrFAR-bound wt-HisF, HisF-F38A, HisF-F23A, and HisF-G20P. Shown in \nthe middle of the figure are 2D-histograms of the root mean square deviations (RMSD) of the Cα-atoms \nof loop1 relative to the crystal closed structure of wild-type and the distance RMSD of all non-covalent \ninteractions in the loop1 that stabilize open liganded conformation during simulations of unliganded (left \nside) and PrFAR-bound systems (right side). Regions corresponding to closed and open conformations \nare indicated with a circle. Enlarged 1D histograms of the distance RMSD are included along with the \n2D-histograms for the simulations without a ligand. For details of how the distance RMSD values were \ncalculated, see the Supplemental Methods. The panel on the left shows, from top to bottom, snapshots \nof loop1 motion in wt-HisF, HisF-F38A, HisF-F23A, and HisF-G20P during the unliganded simulations, \ncolored by the Cα-atom RMSF of loop1 . The panel on the right shows the analogous data from our \ncorresponding PrFAR-bound MD simulations (variants presented in the same order). \n \nFurthermore, in our simulations of the liganded enzyme (Figure 5), where we also included \nthe closed state of the loop in our simulations, we observe only sparse sampling of this closed \nstate in the F23A and G20P variants compared to the corresponding sampling of the closed \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n12 \n \nstate in the F38A variant and the wild-type enzyme. A comparison of Ramachandran plots for \nglycine and proline in wt-HisF and HisF-G20P (Figure S4) shows that angles of G20 sampled \nin our simulations of wild-type closed or closed active conformations are forbidden by proline \nRamachandran plot, explaining why the closed state is so destabilized for the HisF -G20P \nvariant. The impaired sampling of a catalytically competent closed conformation in these \nvariants helps rationalize the diminished/abolished activity observed for these variants in the \nkinetic data (Table 1). \nIn summary, our crystallography, proteolysis, NMR, and simulation data demonstrate that \nthe HisF-G20P and HisF-F23A variants have opposing effects on the conformation of loop1. \nIn both the HisF -G20P and HisF-F23A variants, there is a shift away from the closed \nconformations in our simulations, with preferred sampling of detached or open states of the \nloop. However, whereas loop1 primarily samples the open conformation in the HisF -G20P \nvariant, the loop1 p opulation shifts towards the highly flexible detached conformation in the \nHisF-F23A variant. As these variants both strongly reduce catalytic turnover (Table 1), it is not \npossible to link the population of open and detached conformations of loop 1 with the rate-\nlimiting step ( kcat) in the turnover reaction.  Instead, the G20P and F23A substitutions likely \ninfluence turnover via alterations in the closed conformation of loop1.  \nAmino acid substitutions in loop1 have a limited effect on substrate binding affinities \nSince different loop1 conformations in the apo state cannot explain the higher catalytic \nactivities of wt -HisF and HisF -F38A compared to HisF -G20P and HisF -F23A, it was next \nanalyzed whether these amino acid substitutions have consequences for substrate or product \nbinding. To study the thermodynamics and kinetics of PrFAR binding to HisF, fluorescence \nequilibrium titrations and transient fluorescence kinetic measurements were performed. \nAlthough intrinsic fluorescence of the single tryptophan residue 156 of  HisF has previously \nbeen used as spectroscopic signal transmitter in ligand binding studies45, it proved unsuitable \nfor kinetic measurements because of the unspecific fluorescence quenching upon addition of \nthe substrate PrFAR. We sought to avoid this effect by the introduction of an alternative \nfluorescent probe. The unnatural amino acid L -(7-hydroxycoumarin-4-yl)ethylglycine (CouA) \nwas applied because this probe is relatively small, has good spectroscopic properties and can \neasily be introduced by genetic code extension. 56-57 CouA has been used extensively as \nprotein-based fluorescent sensor that reports on protein-ligand interactions56, 58-60 and enzyme-\nsubstrate binding.61-62 As the 7-hydroxycoumarin moiety can exist in a number of tautomeric \nforms in the ground state, absorption/emission maxima are strongly influenced by \nenvironmental factors such as dielectric constant, hydration and pH.63 For our purposes CouA \nwas incorporated into HisF in place of a lysine at position 132, a position that is not conserved \nin HisF sequences and has a distance of ~ 15 Å to the ligand binding site ( Figure 6A). HisF-\n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n13 \n \nK132CouA variants (wt, F38A, F23A, G20P) were purified with reasonable yields and the \nincorporation of CouA was verified spectroscopically ( Figure S 5). Steady -state kinetic \nparameters of CouA -labeled HisF were virtually identical to those of its non -labelled \ncounterpart (Table 1), corroborating that enzymatic activity is not affected by incorporation of \nthe fluorophore. Ligand binding is associated with a decrease of CouA fluorescence emission. \nEquilibrium titrations with the substrate PrFAR were done in the absence of ammonia to allow \nfor observation of the binding separately from the turnover reaction (Figure 6B).  \n \n \nFigure 6: Ligand binding monitored by equilibrium titrations with CouA-labelled HisF. (A) Site of \nCouA incorporation. The structure of HisF is shown with the open loop1 conformation (orange, PDB \nentry 1vh738) and an overlay of the closed loop1 conformation (beige, PDB entry 7ac8 40). CouA was \nmodelled into the structure and is shown at position 132 (within β -strand 5) as cyan sticks, the bound \nsubstrate precursor ProFAR and the catalytic residues D11 (within β-strand 1) and D130 (within β-strand \n5) are shown as sticks. (B) Equilibrium titrations of HisF-CouA variants at 25°C. Binding of the substrate \nPrFAR to the variants (0.2 µM) resulted in a decrease in CouA fluorescence ( lex = 370 nm, lem = 452 \nnm). Relative emission intensity was plotted vs. PrFAR concentration. Lines represent hyperbolic fits of \nthe data. (C) Apparent KD values obtained in equilibrium titrations for the binding of the substrate PrFAR \nor the product molecules ImGP and AICAR to the HisF -CouA variants in absence or presence of the \nsecond ligand (AICAR or ImGP), respectively. The associated numerical values are listed in Table S1. \nKD values ± SE were determined by fitting the mean ± SEM for at least two technical replicates w ith \nequation 3.   \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n14 \n \nThe KD-values determined in the equilibrium titrations show that the G20P and F23A \nsubstitutions in loop1 slightly weaken the affinity between HisF and the substrate PrFAR or the \nproducts ImGP or AICAR  (Figure 6C, Table S 1). For example, compared between wt -HisF \nand HisF-F23A the dissociation constant for the substrate PrFAR is increased 1.8 -fold, for \nImGP 1.2-fold, and for AICAR 1.4 -fold. Furthermore, we noticed that the apparent affinity of \nAICAR is slightly increased in the presence of the second ligand ImGP and vice versa, which \nindicates a higher formation propensit y of the ternary complex (HisF*ImGP*AICAR) in \ncomparison to the respective binary complexes (HisF*AICAR) and (HisF*ImGP). The \nstabilization effect due to formation of the ternary complex is similar for wt -HisF and HisF -\nG20P, HisF-F23A, and HisF -F38A, indicating that this is a general feature. Looking at the \nfluorescence changes upon titration of the dimeric or ternary complexes, another difference \nbetween wt-HisF/His-F38A and HisF-F23A/HisF-G20P is noticeable. While the fluorescence \namplitudes in the case o f HisF-wt and HisF -F38A are higher when the ternary complex is \nformed than when the binary complexes are formed, the opposite is true for the variants HisF-\nF23A and HisF -G20P ( Table S 2). This is a first indication that the environment of the \nfluorophore CouA in the ternary complex for the active variants wt-HisF and HisF-F38A differs \nfrom the environment in the inactive variants HisF-F23A and HisF-G20P.      \nAn induced-fit movement during PrFAR binding is exclusively observed for wt-HisF and \nHisF-F38A \nThe kinetics of the PrFAR binding reaction were studied in stopped-flow experiments, whereby \nthe CouA fluorescence decrease was recorded after rapidly mixing the respective HisF-CouA \nvariant with a molar excess of PrFAR. For an assessment of ligand binding  kinetics, the \nobserved binding transients were fitted with exponential functions. The number of exponential \nfunctions required to describe the transients allows conclusions to be drawn about the number \nof reaction steps in the binding reaction. In additio n, the secondary plots derived from \nexponential fitting, e.g. kobs as function of the PrFAR concentration, provide initial clues to the \nbinding mechanism. In general, time traces (Figure S6) for wt-HisF and HisF-F38A resemble \neach other, whereas time traces associated with the loop1 variants HisF-F23A and HisF-G20P \nshowed notable differences. In the case of wt -HisF ( Figure S 6A) and variant HisF -F38A \n(Figure S6B) time traces are biphasic  (sum of two exponential terms ). A fast fluorescence \ndecrease is followed by a slow phase with a very small signal amplitude. In the case of loop \nvariants HisF-F23A (Figure S6C) and HisF-G20P (Figure S6D), single exponential functions \nwere adequate to describe the time traces. The overall fluorescence changes associated with \nPrFAR binding were smaller, which resulted in lower signal-to-noise ratios. A plot of the first-\norder rates (kobs) for the binding reaction as a function of PrFAR concentration provides insights \ninto potential differences in the binding mechanisms of the different variants. In the case of wt-\nHisF and HisF -F38A, the turnover rate ( kobs1) depends on the substrate concentration in a  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n15 \n \nhyperbolic manner (Figure S6E), indicating that binding takes place via an induced fit or \nconformational selection mechanism.64-65 In contrast, in the case of loop1 variants HisF-F23A \nand HisF-G20P, kobs, increased linearly with increasing concentrations of PrFAR (Figure S6F), \nwhich is indicative of a simple binding process without involvement of conformational changes. \nThis finding is in accordance with a model where the wt-HisF and HisF-F38A proteins bind the \nligand when the protein is in the open or detached conformation, after which loop1 stably \ncloses over the li gand to form the closed conformation. The HisF -G20P and HisF -F23A \nvariants on the other hand are unable to form a stably closed conformation and prefer to remain \nin the open or detached conformation even in the presence of the ligand, as also observed in \nour simulations (Figures 5).  \nTo directly assess if the formation of the closed state is impaired in the HisF -G20P and \nHisF-F23A variants we again turned to NMR titration experiments. To that end, we added \nProFAR (a stable PrFAR analogue; see Figure 1A), to 15N labelled HisF and followed the \ninduced chemical shift perturbations (CSPs). For all HisF proteins we observed CSPs that \ndirectly report on the interactions between HisF and the ligand.  Interestingly, we observed a \nnew set of signals that likely reports on the closed conformation of loop 1, as F23 is one of the \nresidues that displays a novel conformation upon PrFAR binding (Figure 7, circles). This new \nset of signals thus reports on the formation of the closed state of loop1 in the presence of the \nligand analogue. This stable set of signals does not appear in the HisF-G20P and HisF-F23A \nvariants, proving that the closed conformation is not stably adopted in those cases. This agrees \nwell with simulation data presented in Figure 5. \n \nFigure 7: ProFAR binding induces a conformational change of loop1 only in wt -HisF and HisF-\nF38A. NMR titration experiments recorded in 1H-15N TROSY spectra showing apo HisF (blue) and HisF \nin the presence of saturating amounts of ProFAR (red). The large CSP of F23 upon ProFAR binding is \nshown by a black arrow. The position of several signals with large CSPs is indicated by black circles. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n16 \n \nLarge chemical shift perturbations associated with a substantial conformational change are only \nobserved in the spectra of wt-HisF and HisF-F38A. \n \nOur combined NMR and stopped flow measurements thus are indicative of a two-step binding \nmechanism in the wt-HisF protein, in which the ligand first interacts with the open or detached \nconformations of HisF after which loop1 closes to facilitate catalysis (Scheme 1, top). \n \n \n \nScheme 1. Binding of PrFAR to wt-HisF and loop1 variants:  Induced fit model versus two-state model  \n \nTo obtain insights into the rates that are associated with this two-step binding process we \nfitted the stopped-flow experiments that were performed under pseudo -first order conditions \nfor HisF-CouA (excess of HisF-CouA over PrFAR) to the induced-fit model in Scheme 1. These \nhyperbolic fits allowed for the determination of the KD1 (= k-1/k1), kconf and k-conf for wt-HisF and \nHisF-F38A (Table S3). This shows that the equilibrium of the conformational change is on the \nclosed side and that a stable closed conformation is thus efficiently formed which subsequently \nfacilitates efficient substrate turnover. In contrast, PrFAR binding kinetics for the loop1 variants \nHisF-F23A and HisF-G20P are compatible with a simple one -step binding reaction (Scheme \n1, bottom). The k1 and k-1 values shown in Table S3 result from the slope and intercept with \nthe y-axis of a linear fit (Figure S6F).  \nIn summary, the NMR and stopped flow experiments, as well as molecular dynamics \nsimulations, establish that substrate binding to HisF occurs via an induced fit mechanism for \nwt-HisF and for the HisF-F38A variant. The HisF-G20P and HisF-F23A variants on the other \nhand interact with the substrate via a one step binding mechanism as loop1 is, in those cases, \nunable to close properly over the substrate. Interestingly, these variants still interact efficiently \nwith PrFAR, indicating that loop 1 does not contribu te considerably to the binding energy of \nthe substrate.  \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n17 \n \nProduct release from wt-HisF and HisF-F38A is more complex than for HisF-F23A and \nHisF-G20P  \nNext, we aimed to obtain insights into the release of the AICAR and ImGP products. \nEquilibrium titration measurements showed that both ligands, AICAR and ImGP, can bind \nindependently to wt -HisF and all loop1 variants, causing a decrease of CouA fluorescence \n(Figure 6 ). Exclusively for wt -HisF and HisF -F38A, we noted a significantly higher \nfluorescence change when the ternary complex was formed than when the binary complexes \nwere formed (Table S2), combined with an increase in apparent binding affinity upon formation \nof the ternary complex (Table S1). This suggests that the interaction of either product (ImGP \nor AIRCAR) does not result in a conformational change in the enzyme, whereas the interaction \nwith both products at the same time does result in the closing of loop1.  \nTo obtain the rates that are associated with product release from  wt-HisF and the loop1 \nvariants we made use of stopped-flow measurements. Representative time traces for wt-HisF \nare shown in Figure S7A, the time traces for HisF-F38A resemble those of wt-HisF (data not \nshown). As the transient kinetic measurements show, formation of the binary HisF*AICAR and \nHisF*ImGP complexes is completed within the dead time of the stopped flow device. Based \non the used enzyme and substrate concentrations and an instrument dead -time of ~2.0 ms, \nan observed rate constant kobs of greater than 1000 s-1 is required to obscure all evidence of \nassociation, suggesting that association rate constants for binary complex formation must be \n≥ 106 M-1 s-1. In contrast, when monitoring the formation of the ternary complex, fluorescence \nchanges with rate constants kobs in the range of 50 s-1 were observed. This is visible from the \nexponential fluorescence decrease in the stopped -flow transients when the free enzyme \ninteracts with a mixture of both ligands or when the preformed binary complexes are mixed \nwith the second ligand (Figure S7A: HisF + ImGP/AICAR, HisF*ImGP + AICAR, HisF*AICAR \n+ ImGP). These data thus agree with the equilibrium titrations ( Figure 6 ) that revealed a \nsynergistic effect when AICAR plus Im GP bind to the enzyme and with the notion that the \ninteraction with both ligands is associated with a conformational change in the enzyme. In \ncontrast, in the case of the loop variants F23A and G20P, both the binary and ternary \ncomplexes were formed within  the instrument dead -time in stopped -flow measurements \n(Figure S7B), which confirms that interaction with both ligands does not lead to loop1 closure \nin these variants.  \nTo obtain rate constants for association and dissociation kinetics of the reaction products \nAICAR and ImGP a dataset of 16 time traces was recorded by mixing excess of the ligand with \nlimiting concentrations of HisF CouA or the binary complexes (HisFCouA*I mGP and His \nCouA*AICAR). A kinetic model describing ImGP/AICAR binding to HisF (Figure S7C) includes \nassociation and dissociation of the two ligands to the apo enzyme and to the respective binary \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n18 \n \ncomplexes ( k1 and k-1 for ImGP binding as well as k2 and k-2 for AICAR binding) and a \nconformational change (the closing of loop 1) to stabilize the ternary complex.  \nThe stopped-flow datasets for wt -HisF (Figure S8) and variant HisF -F38A (Figure S9) \nwere subjected to a global fitting analysis according to this kinetic model. The curves resulting \nfrom global fitting analysis are indicated by dashed lines. The determined values for the rate \nconstants are summarized in Table S4. The rate constants obtained for the HisF-F38A variant \nin the global fitting analysis resemble those for wt -HisF. Importantly, the KD values for the \nbinding reactions that were calculated from the global fitting parameters roughly match the KD \nvalues obtained in equilibrium titrations (cf. Table S1 and Table S4). It should be emphasised \nthat the binding model is a minimal model that accounts for key features of the experimental \ndata. It could well be that the rate constants for binding of AICAR and ImGP to apo HisF and \nthe HisF*ImGP/HisF*AICAR complex, respectiv ely, differ. However, this cannot be better \nresolved with the stopped-flow datasets, as the binary enzyme-ligand complexes form within \nthe dead time of the stopped-flow instrument. Importantly, the rates for the loop opening (k-conf) \nare higher for the ImGP:AICAR complex than for the PrFAR complex, which indicates that \nloop1 opens after the reaction to allow for product release. \nThe motions of loop1 are not rate limiting in the kinetic mechanism of HisF \nTo discern which step in the catalytic mechanism is rate -determining for wt -HisF and HisF -\nG20P, HisF-F23A, and HisF-F38A, turnover kinetics under multiple turnover conditions were \ncompared with turnover rates obtained with single turnover conditions. In the multiple turnover \nmode HisF was mixed with an excess of the substrate PrFAR and the turnover of PrFAR was \nmonitored based on the decrease of absorption at 300 nm. Catalytic turnover of PrFAR by \nHisF occurs only in the presence of the second substrate ammonia. Therefore, we compared \nturnover traces in the presence of ammonia with control traces obtained in the absence of \nammonia to discriminate absorption changes accompanying PrFAR turnover from signals \nstemming from binding or mixing reactions. For the re action of wt -HisF a representative \nmultiple turnover trace and the associated control trace are shown in Figure 8A . The \ncorresponding data for the HisF variants are shown in Figure S10A (HisF-F38A), Figure S11A \n(HisF-F23A), and Figure S12A (HisF-G20P).  \nThe time traces obtained under multiple turnover conditions showed a linear steady -state \nphase that is preceded by an exponential burst phase. The burst phase was observed also in \nthe control curve in absence of ammonia. We attribute this burst phase to a mixing artifact of \nthe stopped-flow instrument and this phase was not analyzed any further. Turnover velocities \nwere deduced from a linear fit of the steady-state phase and were plotted as a function of the \nPrFAR concentration to obtain the kcat and KMPrFAR values for wt-HisF (Figure 8B) and HisF-\nF38A, HisF-F23A, and HisF-G20P (Figures S10B – S12B). \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n19 \n \nFor measurements under single -turnover conditions the substrate PrFAR was saturated \nwith enzyme so that all PrFAR molecules participate in the single turnover. The rate of turnover \nrate in that case is unaffected by product release and can be determined by fitting the change \nin the fluorescence over time to a single-exponential function. A representative single-turnover \ntransient for wt-HisF is shown in Figure 8C. The corresponding data for the HisF variants are \nshown in Figure S1 0C (HisF-F38A), Figure S1 1C (HisF-F23A), and Figure S1 2C (HisF-\nG20P). Single-turnover rates kobs determined from these exponential fits were independent of \nthe applied PrFAR concentrations, both for wt-HisF (Figure 8D) and the variants HisF-F38A, \nHisF-F23A, and HisF-G20P (Figure S10D- S12D). The kinetic constants determined for the \nmultiple and single turnover measurements are summarized in Table S5.  \n \nFigure 8: Multiple - and single -turnover kinetics of the wt -HisF reaction.  (A) Representative \ntransient monitoring PrFAR conversion in multiple turnover mode at 25°C after mixing 0.1 µM HisF with \n10.0 µM PrFAR (final concentrations) in the presence of 100 mM ammonium acetate (turnover curve, \nblue line). A linear approximation of the steady-state phase (dashed line) yielded a turnover velocity of \nv = 0.192 µM s -1. The control curve (light blue line) shows the progress of the reaction in absence of \nammonium acetate. (B) Plot of the turnover velocity v vs. the respective PrFAR concentration in multiple \nturnover experiments. kcat and KMPrFAR-values were obtained by fitting to the Michaelis-Menten equation. \n(C) Representative transient monitoring PrFAR conversion in single turnover mode after mixing an \nexcess of HisF (20 µM) with 10 µM PrFAR in the presence of 100 mM ammonium acetate (turnover \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n20 \n \ncurve, red line). The turnover curve was fit with a single exponential decay function (dashed line, 𝑦 =\n𝑎 ∗ 𝑒!\"!\"#∗$ + 𝑐). The control curve (orange line) shows the progress of the reaction in the absence of \nammonium acetate. (D) Plot of the turnover rates, kobs, observed under single turnover conditions, vs. \nthe respective PrFAR concentration. kcat-, KM- and kobs-values are summarized in Table S5. \n \nIn summary, turnover rate measurements confirm that variant HisF-F38A is catalytically as \nactive as wt-HisF, whereas the activities of the two loop1 variants HisF-F23A and HisF-G20P \nare significantly reduced. This deterioration of catalytic activity manifests mainly in kcat values \nand single -turnover rates, which are reduced by three orders of magnitude, but is also \nexpressed in a 2 to 5 -fold increase of the KM values. Remarkably, rate constants obtained in \nmultiple and single turnover measurements have the same order of magnitude, implying that \nproduct release and associated conformational changes are not rate determining in the \ncatalytic mechanism. Hence, it is concluded that the chemical step is rate -determining for \ncatalysis by HisF. This is in contrast to other enzymes in this pathway, such as HisA and PriA, \nwhere loop motion is likely rate determining.32  \n \nCONCLUSIONS \nDue to their catalytic versatility and adaptability (βα)8-barrel enzymes have been successfully \nharnessed as scaffolds for enzyme design. 28, 66 -67 It is already becoming apparent that the \ninclusion of loop engineering into design strategies has an immense potential for the targeted \nengineering of substrate selectivity and catalytic activity.68-70 To do so, a comprehension of the \nconformational states of active-site loops and their significance for the catalytic mechanism is \ncritical. Here, we have studied the importance of the flexible active -site loop1 for the kinetic \nmechanism of the (βα )8-barrel enzyme HisF. In several crystal structures, loop1 adopts a \ndefined open conformation in the absence of substrates or a defined closed conformation when \nthe binding partner HisH and substrates are bound in the active site.40 Furthermore, the NMR \nmeasurements presented here show that loop1, in the absence of substrates, adopts a highly \nflexible ensemble of detached conformations, which appear to be the predominant \nconformations in solution, and our molecular dynamics simulatio ns indicate that the loop is \nconformationally plastic and capable of taking a range of conformational states, depending \nboth on loop sequence and whether a ligand is bound to the active site or not (Figures 5). \nTo address the importance of the different loop conformations for catalytic turnover we \nshifted the conformation of loop1 through single point mutations. Subsequently we assessed \nthe binding properties and activity of these variants through a combination of steady state and \nstopped-flow kinetics, X -ray crystallography, NMR spectroscopy and molecular dynamics \nsimulations.  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n21 \n \nWe established that loop1 in unliganded wt -HisF adopts both the open and detached \nconformations, where the substrate binding site is accessible, but without fully being able to \naccess a catalytically competent closed conformation similar to that observed in the HisF/HisH \ncomplex (PDB ID: 7AC840). After recruitment of the substrate, however, loop1 remodels and \ncloses over the substrate binding pocket. In this closed conformation F23 in loop1 stacks onto \nthe substrate. The formation of the catalytically important enzyme:substrate complex thus \ntakes place in two steps: binding of substrate, followed by the closing of loop1 over the \nsubstrate.  \nIn the HisF -F38A variant the open conformation of loop1 was slightly destabilized by \nremoving an aromatic contact between this loop and the core of the enzyme. In the unliganded \nstate, this led to a small shift from the open conformation towards the detached conformation. \nIn the substrate-bound state, the HisF-F38A variant properly formed the closed conformation. \nIn binding and activity assays, this variant was indistinguishable from wt -HisF. Based on that \nthe equilibrium between the open and detached confo rmations does not play a rate limiting \nrole in the catalytic cycle of the enzyme.  \nIn the unliganded form of the HisF -G20P and HisF -F23A variants loop1 was slightly \nstabilized in the open conformation (G20P) or considerably shifted towards the detached \nconformation (F23A). As both variants interact with the substrate with a similar affin ity this \nimplies that the loop1 conformation in the apo state (open or detached) does not influence \nsubstrate recruitment. Nevertheless, both variants display a slightly reduced substrate binding \naffinity compared to the apo enzyme and, importantly, bind the substrate in a simple one step \nbinding mechanism. In addition, NMR experiments reveal that these mutations impair the \nformation of the closed conformation. Consequently, the activity of these variants is reduced \nby three orders of magnitude compared to the wt-HisF protein.  \nTaken together, our data reveal a clear model that correlates conformational changes in \nloop1 with substrate turnover ( Figure 9 ). In this model the formation of the closed loop1 \nconformation takes place after substrate recruitment and is essential for substrate turnover. \nAfter the cyclase reaction, which is the rate limiting step in the catalytic cycle, the closed loop \nconformation is destabilized, which facilitates product release. \n \n \nFigure 9: Model of the conformational changes of loop1 and their importance for the catalytic reaction.  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n22 \n \nIt has been shown that ligand-gated loop motions in (βα) 8-barrel enzymes differ in their \nmagnitude (e.g., number of loops involved) and dynamics, and affect different reaction steps \nof the catalytic mechanism. In the prototypical and well -studied TIM, which catalyzes the \nisomerization of glyceraldehyde-3-phosphate and dihydroxy-acetone phosphate in glycolysis, \nloop6 is a phosphate -gripper loop that moves ~7 Å from a catalytically inactive open \nconformation to a catalytically active closed conformation upon substrate -binding.71 The \nconformational change occurs in concert with substantial internal rearrangement of the \nadjacent loop β7α7.50, 72 The most important consequence of the conformational change is the \nexclusion of solvent from the active site, reducing the dielectric constant in the surrounding of \na catalytic glutamate residue and shifting its pKa value in such a way that it can act as a general \nbase.73-74 The movement of loop6 has long been interpreted as a rigid body movement, with \nthe loop moving as a lid attached to two hinges. 33, 75-76 However, more recent computational \nwork indicates that loop6 is highly flexible, sampling multiple open distinct conformations that \ninterconvert between each other, whereas the closed conformation falls in a very narrowly \ndefined energy basin, and any dev iation from this conformation has negative impact on \ncatalytic turnover. 50 Product release was identified as the rate -determining step in the \nbiologically relevant reaction (conversion of dihydroxyacetone phosphate to D-glyceraldehyde \n3-phosphate) and loop6 movement is necessary, among others, to release the product from  \nthe active-site.76-77  \nA similar role is played by loop movements in the catalytic mechanism of the (βα) 8-barrel \nenzyme indole-3-glycerol phosphate synthase (IGPS, TrpC), which catalyzes the indole ring \nclosure reaction during tryptophan biosynthesis. 78-80 In IGPS, dynamics of the loop1, which \nhouses a catalytically important Lys residue, are governed by competing interactions on the \nN- and C -terminal sides of the loop. Disrupting these interactions through amino acid \nsubstitutions quenches loop dynamics on the microsecond to millisecond timescales and slows \ndown the dehydration step in the catalytic reaction.81-82 It seems that loop1 is maintained in a \nstructurally dynamic state by the competing interactions, whereby the extent of loop mobility \ncorrelates with the rate-limiting step of the catalytic reaction and product release is rate-limiting \nat ambient temperature.83  \nIn the case of HisA, PriA and TrpF, (βα) 8-barrel enzymes that catalyze isomerization \nreaction in histidine and tryptophan biosynthesis, multiple active site loops undergo substantial \nligand-gated conformational changes. 84-86 Loop dynamics in HisA , PriA and TrpF is highly \ncomplex, with loop motion being at least partially rate limiting for substrate binding and being \nlinked to substrate selectivity.32  \nL23oop1 motion in HisF differs from previous examples in so far as the chemical \nconversion itself, rather than substrate binding or product release, is rate -determining for the \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n23 \n \noverall turnover reaction. A plausible reaction mechanism for the chemical conversion \ncatalyzed by HisF has been proposed previously 36 and involves the formation of two imine \nintermediates during acid-base catalysis. In the absence of structures of HisF bound to PrFAR \nor reaction intermediates, we can only speculate about how the induced fit facilitates the \nconversion reaction. It is reasonable to assume that the pKa of the catalytic acid D11, which is \nvery close to loop1, is increased by the generation of a hydrophobic environment stabilizing \nthe protonated form of the aspartate side chain. When loop1 is in the closed conformation, F23 \ncomes into close proximity of V 18 and I52 and thereby create a hydrophobic cavity for the \ncatalytic acid and shield it from solvent. A similar effect, the shielding of a catalytic aspartate \nresidue through the closure of an active site loop has been described in TIM.87 In future work, \nthis hypothesis may be substantiated by analyzing the protonation states of the catalytic \nresidues D11 and D130 in HisF.  \n \n \nMATERIALS AND METHODS \nSite-directed Mutagenesis  \nPoint mutations were introduced into pET28a_HisF88 with a modified version of the protocol of \nthe Phusion site -directed mutagenesis kit (Thermo Fisher Scientific) with HPSF -purified \nprimers (BioSynth). To facilitate phosphorylation of the PCR product, T4 polynucleotide kinase \nwas added during ligation. Mu tagenesis was confirmed by Sanger sequencing \n(MicrosynthSeqlab). For the incorporation of the unnatural amino acid L-(7-hydroxycoumarin-\n4-yl)ethylglycine (CouA) at position 132, an amber stop codon mutation (TAG) was introduced \ninto pET28a_HisF according to the protocol described above (pET28a_HisF_TAG). \nProtein Expression and Purification \nAll experiments were performed with Thermotoga maritima HisF (Uniprot ID: Q9X0C6) or HisF \nloop1 variants. Genes were expressed from modified pET vectors, encoding an N -terminal \nHis6-tag followed by a TEV cleavage site, in E. coli  BL21Gold (DE3) cells (Agilent \nTechnologies). Expression was performed at 30°C overnight after induction with 1 mM \nIsopropyl β-D-1-thiogalactopyranoside (IPTG) at an OD600 of 0.6-0.8. Cells were harvested by \ncentrifugation, resuspended in 50 mM Tris-HCl pH 7.5, 300 mM NaCl, 10 mM imidazole, and \nlysed by sonication. E. coli proteins were precipitated by a heat shock (15 min, 60°C) and \nremoved by centrifugation. The supernatant was subjected to Ni -immobilized metal affinity \nchromatography (IMAC) (HisTrap FF C rude column, 5 ml, GE Healthcare). Proteins were \neluted with a linear gradient of imidazole (10 -500 mM). Fractions containing the protein of \ninterest were identified by sodium dodecyl sulfate -polyacrylamide gel electrophoresis (SDS -\n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n24 \n \nPAGE) and pooled. Eluted proteins were digested with TEV protease at room temperature \nover night during dialysis against 50 mM Tris -HCl pH 7.5. TEV protease and non -cleaved \nprotein was removed by IMAC (HisTrap FF Crude column, 5 ml, GE Healthcare) with a linear \ngradient of imidazole (0 -500 mM). Fractions at low imidazole concentration containing the \nproteins of interest were identified by SDS-PAGE analysis, pooled, and further purified with a \nsize-exclusion chromatography (SEC) column (Superdex 75 HiLoad26/ 260, GE Healthcare) \nby using 50 mM Tris-HCl pH 7.5 as the running buffer. Eluted protein fractions were checked \nby SDS-PAGE for > 90% purity, pooled, concentrated, and dripped into liquid nitrogen for \nstorage at -80 °C.  \nFor expression of HisF containing the unnatural amino acid CouA, pET28a_HisF_TAG was \nco-transformed with pEVOL_CouA, carrying the gene for the modified tyrosyl aminoacyl-tRNA \nsynthethase from M. janaschii57 into E. coli BL21Gold (DE3). Cells were grown at 37°C in 6 L \nof LB medium until the OD 600 reached 0.6-0.8. Cells were harvested by centrifugation and \nresuspended in 600 ml terrific broth (TB) medium. Bacterial growth at 37 °C was continued up \nto an OD 600 of 10 and incorporation was induced by addition of 0.45 mM CouA and 0.02 % \narabinose. Gene expression was induced by addition of 1 mM IPTG. Cultures were incubated \novernight at 30 °C and the proteins were purified by nickel -affinity chromatography as \ndescribed above.  \nThe auxiliary enzymes HisA and HisE/IG from T. maritima were purified by standard \nmethods from E. coli BL21-Gold cells (Agilent Technologies) that overexpressed the respective \nproteins. \nProFAR/PrFAR Synthesis \nThe HisF ligands were synthesized enzymatically from 5 -phospho-D-ribosyl α-1-\npyrophosphate and adenosine triphosphate using the purified enzymes HisE/IG. 89 The \nprogress of the reaction was traced spectrophotometrically and the ProFAR product was \npurified using ion -exchange chromatography (POROS column; HQ 20, 10 ml, Applied \nBiosystems) using a linear gradient of 50 mM to 1 M ammonium acetate. ProFAR purity w as \nexamined through the absorbance ratio A 290/A260 and the concentration was determined at a \nwavelength of 300 nm (e300 = 6069 M-1 cm-1). PrFAR was synthesized from ProFAR with HisA \nfrom T. maritima. The product was purified using ion-exchange chromatography as described \nfor ProFAR. Highly concentrated and >95% pure (A 290/A260 = 1.1-1.2) fractions were unified, \nflash frozen in liquid nitrogen, and stored at -80°C. \nLimited Proteolysis \nProteolytic stability was tested at room temperature by incubating 10 µM HisF with 64 nM \ntrypsin in 50 mM Tris/HCl pH 7.5. The reaction was stopped after different time intervals by \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n25 \n \nadding one volume of 2x SDS -PAGE sample buffer and heating for 5 min at 95°C. The time \ncourse of proteolysis was followed on SDS-PAGE. \nSteady-state Enzyme Kinetics \nThe ammonia -dependent activity of HisF was measured by recording PrFAR turnover \ncontinuously at 300 nm [ De300(PrFAR-AICAR) = 5637 M-1 cm-1] in 50 mM Tris/acetate pH 8.5 \nat 25 °C with a Jasco V650 UV-vis spectrophotometer. To determine KMPrFAR, 0.1-0.3 µM of wt-\nHisF or HisF loop1 variants were saturated with ammonia by adding 100 mM ammonium \nacetate (corresponding to 14.4 mM NH3 at pH 8.5). PrFAR (1-40 µM) was synthesized in situ \nfrom ProFAR, using a molar excess (0.5 µM) of HisA from T. maritima and converted by HisF \nto ImGP and AICAR. To ensure that ProFAR is completely turned over to PrFAR, the reaction \nmixture was incubated for at least 2 min before addition of wt -HisF or HisF loop1 variants. \nEnzyme activity was deduced from the initial slopes of the transition curves. Michaelis-Menten \nconstants KM and kcat were determined by plotting the measured mean activity values and their \nstandard error of the mean (SEM) of at least two technical replicates against the ProFAR \nconcentration and fitting the data with the Michaelis-Menten equation (eqn. 1, 2): \n \n𝑣 =\n%$%&['()*+,]\n.'/['()*+,]                 (1)       \nkcat = vmax/[HisF]                (2) \n \nEquilibrium Ligand Titrations \nFluorescence titrations of CouA-labeled HisF were performed at 25 °C in 50 mM Tris/acetate \npH 8.5 in a Jasco FP-6500 spectrometer. CouA fluorescence emission was monitored at 451 \nnm with excitation at 367 nm. The substrate PrFAR was added stepwise from stock solutions \nin small volumes under constant stirring to a solution containing 0.2 µM CouA -labeled HisF \nand fluorescence emission was determined for each ligand concentration. Similarly, the \nproduct molecules AICAR or ImGP were titrated to 1.0 µM CouA-HisF (1.0 µM CouA-HisF/4.0 \nmM AICAR or 1.0 µM CouA -HisF/0.5 mM ImGP, respectively). Fluorescence values were \ncorrected for dilution effects and the intrinsic fluorescence of ImGP. Fluorescence changes \n(ΔF) were plotted as a function of the ligand concentration  and plots were fit to hyperbolic \nequations using SigmaPlot to obtain apparent KD values (eqn. 3):  \n \n∆𝐹 =\n0*$%&∗[123456]\n.(/[123456]                (3) \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n26 \n \nGiven KD values represent the average and standard error of at least two technical replicates. \nStopped-flow Turnover and Ligand Binding Kinetics \nStopped-flow studies were performed at 25 °C using the SX20 stopped -flow instrument \n(Applied Photophysics). The instrument was equipped with a LED300 light source for \nabsorbance measurements and a Me -Xe-Arc lamp for fluorescence measurements. At least \nfive individual traces were recorded at each condition and averaged. Concentrations refer to \nfinal concentrations in the observation cell, unless otherwise specified. \nMultiple- and single-turnover measurements were performed in 50 mM Tris/acetate (pH \n8.5) and 100 mM ammonium acetate (turnover curves) or without ammonium acetate (control \ncurves). For multiple-turnover measurements the A300 [De300(PrFAR-AICAR) = 5637 M-1 cm-1] \nwas recorded over time after mixing a constant concentration of HisF ( -wt/-F38A: 0.1 µM, -\nF23A: 0.5 µM or -G20P: 1.5 µM) with a molar excess of PrFAR (0.5 -50 µM) in a 1:1 volume \nratio. Slopes were obtained by linear approximation of the steady-state part of the curves. The \nobtained turnover velocities ( v = slope/(1 cm * 0.005637 µM -1 cm-1)) were replotted as a \nfunction of the PrFAR concentration and fitted with the Michaelis -Menten equation to obtain \nkcat and KM-values. Single-turnover concentration series were measured by mixing excess HisF \n(20 µM) with different concentrations of PrFAR (2.5, 5.0 and 10.0 µM). Traces were fit with \nexponential decay functions ( 𝑦 = 𝑎7 ∗ 𝑒!\"!\"#∗$ + 𝑐). In the replot, kobs-values were plotted \nagainst associated PrFAR concentrations. \nTo study ligand binding kinetics, the change in fluorescence emission intensity of CouA -\nlabeled HisF upon ligand binding was recorded over time in 50 mM Tris/acetate, pH 8.5 with \nan excitation wavelength of 367 nm and a 420 nm cut -off filter. For analysis of the PrFAR \nbinding reaction, a constant concentration of CouA-labeled HisF (0.1 µM) was mixed with an \nexcess of PrFAR (0.5 -40 µM) in a 1:1 volume ratio to observe the binding reaction. Traces \ncorresponding with wt-HisF and the HisF-F38A variant were fit to the sum of two exponential \nfunctions (𝑦 = 𝐴𝑚𝑝7 ∗ 𝑒!\"!\"#7∗$ + 𝐴𝑚𝑝8 ∗ 𝑒!\"!\"#8∗$ + 𝑐), corresponding traces of variants HisF-\nF23A and HisF-G20P were fit to single exponential functions (𝑦 = 𝐴𝑚𝑝 ∗ 𝑒!\"!\"#7∗$ + 𝑐). \nTo analyse the binding kinetics of the reaction products AICAR and ImGP, CouA -labeled \nHisF (0.05 µM) was mixed with an excess of AICAR (0.25 mM, 1.0 mM), an excess of ImGP \n(0.1 mM, 0.25 mM), or a mixture of the two molecules. In addition, the preformed bin ary \ncomplexes (0.05 µM HisF -CouA/1.25 mM AICAR, 0.05 µM HisF -CouA/0.4 mM ImGP) were \nmixed with the respective second product molecule (0.1-0.25 mM ImGP, 0.25-1.0 mM AICAR) \nto observe formation and dissociation of the ternary complex.  \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n27 \n \nGlobal Fitting Analysis \nSets of primary kinetic traces associated with the binding of ImGP and AICAR to HisF were fit \nglobally to kinetic models using DynaFit (BioKin) 90, which utilizes direct numerical integration \nto simulate experimental results. The script file for the global analysis of the binding reaction \nof wt -HisF is shown in the Supplemental Methods . Rate constants and the associated \nresponse coefficients were optimized iteratively in the global analysis. DynaFit features an \nerror analysis functionality and model discrimination analysis, which was utilized to compare \nvarious kinetic models and to evaluate the quality of the fits.  \nProtein Crystallization, X-ray Data Collection, and Structure Determination \nFor crystallization HisF was concentrated to 25 mg/ml and mixed 1:1 with the respective \nreservoir solution for hanging drop vapour diffusion crystallization. Crystals of wt -HisF were \ngrown in previously determined conditions using Qiagen EasyXtal 15 well plates.91 HisF-F23A \nwas crystallized in 1.2 M ammonium phosphate, using wt -HisF crystals for micro seeding. \nCrystals of HisF -G20P were obtained using a Morpheus II Screen (Molecular dimensions). \nCrystals were mounted on a nylon loop and flash frozen in liquid nitrogen without addition of \ncryoprotectants. Data sets were collected  using synchrotron radiation from the Swiss Light \nSource (SLS), Switzerland at beamline PXIII and PXI. Data collection was done at cryogenic \ntemperature (see Table S6 for data collection and refinement statistics). Data were processed \nusing XDS92, and the data quality was assessed using the program PHENIX93. Structures were \ndetermined by molecular replacement with MOLREP and programs within the CCP4isuite 94 \nusing PDB entry 1THF 48 as the search model. Initial refinement was performed using \nREFMAC95. The model was further improved in several refinement rounds using automated \nrestrained refinement with the program PHENIX93 and interactive modelling with Coot96. \nNMR Measurements \nHisF used for ProFAR titration experiments was 15N-labeled. HisF used for backbone \nassignment and {1H}-15N hetNOE experiments were 2H, 13C, 15N-labeled. Isotope labelling was \nachieved by expression in E. coli BL21-CodonPlus (DE3) cells in M9 minimal medium.  The \nM9 medium was H 2O based and contained 0.5 g/L 15NH4Cl for expression of 15N-labeled \nprotein and 2 g/l 13C -glucose for 13C-labeled samples. The M9 medium for the expression of \n2H, 13C, 15N-labeled protein was D2O based and contained 0.5 g/L 15NH4Cl and 2 g/L 2H/13C-\nglucose. Protein expression was induced at an OD 600 of 0.8 by addition of 1 mM IPTG to the \nmedium and proteins were expressed overnight at 25 °C. Cells were harvested by \ncentrifugation and lysed by sonication. E. coli proteins were precipitated by a heat shock (20 \nmin, 70 °C). Precipitated proteins and cell debris were  removed by centrifugation. Isotope -\nlabeled HisF was purified by IMAC as described for non-labelled HisF. After TEV cleavage the \n2H, 13C, 15N-labeled HisF was unfolded and refolded to exchange the  2H from the expression \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n28 \n \nmedium to 1H for the amide groups in the protein core. This was achieved by dialysis overnight \nat room temperature against 50 mM Arg/Glu pH 7.3, 25 mM HEPES, 5 M guanidinium chloride, \n2 mM DTT for unfolding and subsequent dialysis overnight at room temperature agains t 50 \nmM Arg/Glu pH 7.3, 25 mM HEPES, 50 mM NaCl, 2 mM DTT. Afterwards the proteins were \nsubjected to reverse IMAC. The flow through of this column was concentrated and purified by \nSEC (Superdex 75 HiLoad26/260, GE Healthcare) using NMR buffer (20 mM HEPES pH 7.3, \n50 mM NaCl, 1 mM DTT) as running buffer.  \nNMR experiments were conducted at 30 °C in NMR buffer supplemented with 5 % (v/v) \nD2O on 600 and 800 MHz Bruker Avance Neo spectrometers equipped with N 2 (600 MHz) or \nhelium (800 MHz) cooled cryoprobes. NMR samples contained 100-200 µM 15N-labeled HisF \nfor ProFAR titrations or 300 -600 µM 2H, 13C, 15N-labeled HisF for { 1H}-15N hetNOE and \nbackbone assignment experiments. The previously published backbone assignments of wt -\nHis97 that were obtained under different buffer conditions were transferred to our measurement \nconditions based on TROSY variants of 3D-HNCACB, 3D-HN(CO)CACB, 3D-HN(CA)CO and \n3D-HNCO experiments.98 The same set of experiments was used to transfer the assignment \nfrom wt-HisF to the loop1 variants HisF-F23A, HisF-G20P, and HisF-F38A.  \nDue to the instability of ProFAR, which prevents the use of triple resonance spectra for \nbackbone assignment, a selective unlabeling strategy was used for the assignment of F23 in \nthe ProFAR-bound state (Figure S13): A 15N-labeled sample of HisF with non-labeled Tyr/Phe \nresidues and a 15N13C-labeled sample with non -labeled Asn residues were prepared by \naddition of non -labeled amino acids (100 mg/L each) to 15N or 15N/13C H 2O-M9 medium. \nUnlabeling of Tyr in addition to Phe was chosen due to isotope scramblin g between the two \namino acids. As F23 is the only Phe/Tyr succeeding an Asn it can be assigned by comparing \n1H15N-TROSY spectra of the Tyr/Phe unlabeled sample and 2D 1H15N-HNCO spectra of the \nAsn unlabeled sample with fully labeled samples. Comparison of the 1H15N-TROSY spectra \nreveals all Tyr/Phe signals, whereas the comparison of the 2D 1H15N-HNCO spectra reveals \nall signals succeeding an Asn. Only the signal of F23 is missing in both spectra. As both, 1H15N-\nTROSY and 1H15N-HNCO spectra, can be recorded in a couple of hours this allows the \nunambiguous assignment of F23 even in the unstable ProFAR sample. \n{1H}-15N hetNOE experiments were recorded using the pulse sequence from Lakomek et \nal.99 with a recovery delay of 1 s and a proton saturation time of 9 s. Spectra were processed \nwith Topspin 4.0.2 or NMRPipe 9.6100. Spectra were analyzed with CARA101and integrated with \nNMRPipe100. \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n29 \n \n \nMolecular Dynamics Simulations and Analysis \nMolecular dynamics simulations with HisF were performed in the loop1 open and closed states, \nboth with and without PrFAR bound to the active site. Due to the lack of a structure of HisF \nisolated from HisH in a loop closed conformation, all loop-closed simulations were performed \nby extracting wt-HisF coordinates from the crystal structure of the HisF/HisH complex (PDB \nID: 7AC8 40), with substrate PrFAR aligned with and replacing the crystallized substrate \nanalogue ProFAR in the HisF active site. Loop open simulations were initiated from the loop -\nopen structure of wt-HisF (PDB ID: 1THF48), with the introduction of a S21T reversion to match \nthe other crystal structures used in this work. In both loop open and loop closed systems, \nstarting structures of the HisF -F23A and HisF-F38A variants for simulation were constructed \nbased on the corresponding wt-HisH crystal structure. In the case of the HisF -G20P variant, \nstarting structures for loop closed simulations of this variant were generated based on the \ncorresponding wt-HisF crystal structure, whereas the loop open simulations were initiated from \nthe corresponding crystal structure of this HisF variant (PDB ID: 8S8R, this work). All manually \ngenerated mutant structures were created using PyMOL 102 applying the “Mutagenesis” \nfunction. Rotamers were selected from the backbone -dependent rotamer library such as to \neliminate structural clashes.  \nThe resulting crystal structures were then prepared for simulations and equilibrated \nfollowing a standard equilibration procedure, as described in detail in the Supplemental \nMethods. Once equilibrated, ten 1 µs production runs were performed for each system in an \nNPT ensemble (1 atm pressure and 300 K), resulting in 30 µs cumulative simulation time per \nvariant (initiated from loop open and closed conformations for liganded systems but just open \nfor unliganded ones ), and 1 20 µs cumulative simulation time acr oss all enzyme variants.  \nConvergence of the simulations is shown in Figures S1 4 – S16. Hydrogen atoms in all \nproduction simulations were scaled using hydrogen mass repartitioning 103 allowing for a 4 fs \nsimulation time step. Temperature and pressure were regulated using Langevin temperature \ncontrol (collision frequency 1 ps−1), and a Berendsen barostat (1 ps pressure relaxation time). \nAll simulations were performed using the AMBER ff14SB force field 104, and the TIP3P water \nmodel105, using the CUDA-accelerated version of the Amber22 simulation package106. Further \ndetails of simulation setup, equilibration and analysis are provided as Supplemental Methods, \nand a data package containing simulation starting structures, snapshots from trajectories, \nrepresentative input files and any non-standard simulation parameters has been uploaded to \nZenodo for reproducibility and is available for download under a CC -BY l icense at DOI: \n10.5281/zenodo.12211377. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n30 \n \nASSOCIATED CONTENT \n \nSupporting Information \nSupplemental Figures S1-S16, Supplemental Tables S1-S9, Supplemental Methods, \nSupplemental References. \n \nAUTHOR INFORMATION \n \nCorresponding Authors \nShina Caroline Lynn Kamerlin – School of Chemistry and Biochemistry, Georgia Institute \nof Technology, Atlanta, GA-30332, USA \norcid.org/0000-0002-3190-1173  \nE-mail: skamerlin3@gatech.edu  \n \nRemco Sprangers – Institute of Biophysics and Physical Biochemistry, Regensburg Center \nfor Biochemistry, University of Regensburg, 93053 Regensburg, Germany \norcid.org/0000-0001-7323-6047 \nE-mail: remco.sprangers@ur.de \n \nReinhard Sterner – Institute of Biophysics and Physical Biochemistry, Regensburg Center \nfor Biochemistry, University of Regensburg, 93053 Regensburg, Germany \norcid.org/0000-0001-8177-8460;  \nE-mail: reinhard.sterner@ur.de \n \nAuthors \nEnrico Hupfeld – Technical University of Munich, Campus Straubing for Biotechnology \nand Sustainability. Chair of Chemistry of Biogenic Resources, 94315 Straubing, Germany \n Sandra Schlee – Institute of Biophysics and Physical Biochemistry, Regensburg Center \nfor Biochemistry, University of Regensburg, 93053 Regensburg, Germany \norcid.org/0009-0005-7728-9134 \n Jan-Philip Wurm, Bruker Biospin, Rudolf-Plank-Str.12, 76275 Ettlingen, Germany \n Chitra Rajendran – Institute of Biophysics and Physical Biochemistry, Regensburg \nCenter for Biochemistry, University of Regensburg, 93053 Regensburg, Germany \n Dariia Yehorova – School of Chemistry and Biochemistry, Georgia Institute of \nTechnology, Atlanta, GA-30332, USA \n Eva Vos – School of Chemistry and Biochemistry, Georgia Institute of Technology, \nAtlanta, GA-30332, USA \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n31 \n \n Dinesh Ravindra Raju – School of Chemistry and Biochemistry, Georgia Institute of \nTechnology, Atlanta, GA-30332, USA \n  \nAuthor Contributions \nE. H. conceptualized the project and performed enzymatic measurements and X -ray \nexperiments. S. S. performed and analyzed transient kinetic measurements. J.P.W. \nperformed, analyzed, and interpreted NMR experiments. C. R. performed and analyzed X-ray \nexperiments. D.Y., E.V., and D. R. R. performed, analyzed , and interpreted MD simulations. \nE. H., S. S., and J. P. W. wrote the original draft. S.C.L.K., R.Sp. and R. St. supervised the \nproject, acquired funding, and revised and edited the manuscript.  \n \nFunding \nThis work was supported by a grant of the Deutsche Forschungsgemeinschaft (STE 891/11 -\n2), and by the Swedish Research Council (grant number 2019 -03499). The computational \nsimulations and data handling were enabled by resources provided by the National Academic \nInfrastructure for Supercomputing in Sweden (NAISS) at Chalmers Centre for Computational \nScience and Engineering (C3SE), High Performance Computing  Center North (HPC2N) and \nUppsala Multidisciplinary Center for Advanced Computational Science (UPPMAX) p artially \nfunded by the Swedish Research Council through grant agreement no. 2022 -06725 (SNIC \n2022/3-2 and NAISS 2023/3 -5). Additionally, this work used the Hive cluster, which \nis supported by the National Science Foundation under grant number 1828187 and w as \nsupported in part through research cyberinfrastructure resources and services provided by the \nPartnership for an Advanced Computing Environment (PACE) at the Georgia Institute of \nTechnology, Atlanta, Georgia, USA. Further simulations were performed on the Theta cluster \nat the Argonne Leadership Computing Facility (ALCF), through a Director’s Discretionary \naward. \n \nNotes \nThe authors declare no competing financial interest.  \n \nACKNOWLEDGMENTS \nThe authors thank Jeannette Ueckert and Sabine Laberer  for excellent technical assistance, \nJohanna Stoefl for support in the production of proteins for NMR studies,  as well as Frank \nRaushel, Matthias Wilmanns, and Sihyun Sung for critical reading of the manuscript and for \nfruitful discussions. \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n32 \n \nABBREVIATIONS \nAICAR, 5-aminoimidazol-4-carboxamidribotide; HisF, cyclase subunit of ImGPS (used for HisF \nfrom T. maritima); HisH, glutaminase subunit of ImGPS; IGPS, indole glycerol phosphate \nsynthase; ImGP, imidazole glycerol phosphate; ImGPS, imidazole glycerol phosph ate \nsynthase, PDB, Protein Data Bank; PrFAR, N´-[(5´-phosphoribulosyl)formimino]-5-amino-\nimidazole-4-carboxamide ribonucleotide (HisF substrate); ProFAR, N´-[(5´-\nphosphoribosyl)formimino]-5-amino-imidazole-4-carboxamide ribonucleotide (HisF substrate \nanalogue); TIM, triose phosphate isomerase; wt, wild type. \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n33 \n \nREFERENCES \n1. Bhabha, G.; Biel, J. T.; Fraser, J. S., Keep On Moving: Discovering and Perturbing the \nConformational Dynamics of Enzymes. Acc. Chem. Res. 2015, 48, 423-430. \n2. Hammes, G. G.; Benkovic, S. J.; Hammes -Schiffer, S., Flexibility, Diversity, and \nCooperativity: Pillars of Enzyme Catalysis. Biochemistry 2011, 50, 10422-10430. \n3. Hammes-Schiffer, S.; Benkovic, S. J., Relating Protein Motion to Catalysis. Annu. Rev. \nBiochem. 2006, 75, 519-541. \n4. Berendsen, H. J.; Hayward, S., Collective Protein Dynamics in Relation to Function. \nCurr. Opin. Struct. Biol. 2000, 10, 165-169. \n5. Sawaya, M. R.; Kraut, J., Loop and Subdomain Movements in the Mechanism of \nEscherichia coli Dihydrofolate Reductase: Crystallographic Evidence. Biochemistry 1997, 36, \n586-603. \n6. Changeux, J. P.; Edelstein, S., Conformational Selection Or Induced Fit? 50 Years of \nDebate Resolved. F1000 Biol. Rep. 2011, 3, 19. \n7. Stank, A.; Kokh, D. B.; Fuller, J. C.; Wade, R. C., Protein Binding Pocket Dynamics. \nAcc. Chem. Res. 2016, 49, 809-815. \n8. Koshland, D. E., Application of a Theory of Enzyme Specificity to Protein Synthesis. \nProc. Natl. Acad. Sci. USA 1958, 44, 98-104. \n9. Richard, J. P., Protein Flexibility and Stiffness Enable Efficient Enzymatic Catalysis. J. \nAm. Chem. Soc. 2019, 141, 3320-3331. \n10. Agarwal, P. K., A Biophysical Perspective on Enzyme Catalysis. Biochemistry 2019, \n58, 438-449. \n11. Agarwal, P. K.; Bernard, D. N.; Bafna, K.; Doucet, N., Enzyme Dynamics: Looking \nBeyond a Single Structure. ChemCatChem 2020, 12, 4704-4720. \n12. Agarwal, P. K.; Doucet, N.; Chennubhotla, C.; Ramanathan, A.; Narayanan, C., \nConformational Sub-states and Populations in Enzyme Catalysis. Methods Enzymol. 2016, \n578, 273-297. \n13. Shen, R.; Crean, R. M.; Olsen, K. J.; Corbella, M.; Calixto, A. R.; Richan, T.; Brandao, \nT. A. S.; Berry, R. D.; Tolman, A.; Loria, J. P.; Johnson, S. J.; Kamerlin, S. C. L.; Hengge, A. \nC., Insights Into the Importance of WPD-loop Sequence for Activity and Structure in Protein \nTyrosine Phosphatases. Chem Sci 2022, 13, 13524-13540. \n14. Benkovic, S. J.; Hammes, G. G.; Hammes -Schiffer, S., Free -Energy Landscape of \nEnzyme Catalysis. Biochemistry 2008, 47, 3317-3321. \n15. Kamerlin, S. C.; Warshel, A., At the Dawn of the 21st Century: Is Dynamics the Missing \nLink for Understanding Enzyme Catalysis? Proteins 2010, 78, 1339-1375. \n16. Kohen, A., Role of Dynamics in Enzyme Catalysis: Substantial Versus Semantic \nControversies. Acc. Chem. Res. 2015, 48, 466-473. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n34 \n \n17. Nagel, Z. D.; Klinman, J. P., A 21st Century Revisionist's View at a Turning Point in \nEnzymology. Nat. Chem. Biol. 2009, 5, 543-550. \n18. Furnham, N.; Sillitoe, I.; Holliday, G. L.; Cuff, A. L.; Laskowski, R. A.; Orengo, C. A.; \nThornton, J. M., Exploring the Evolution of Novel Enzyme Functions within Structurally Defined \nProtein Superfamilies. PLoS Comput. Biol. 2012, 8, e1002403. \n19. Newton, M. S.; Guo, X.; Soderholm, A.; Nasvall, J.; Lundstrom, P.; Andersson, D. I.; \nSelmer, M.; Patrick, W. M., Structural and Functional Innovations in the Real-Time Evolution \nof New (betaalpha)8 Barrel Enzymes. Proc. Natl. Acad. Sci. USA 2017, 114, 4727-4732. \n20. Crean, R. M.; Gardner, J. M.; Kamerlin, S. C. L., Harnessing Conformational Plasticity \nto Generate Designer Enzymes. J. Am. Chem. Soc. 2020, 142, 11324-11342. \n21. Nestl, B. M. H., B., Engineering of Flexible Loops in Enzymes. ACS Catal. 2014, 4, \n3201-3211. \n22. Schenkmayerova, A.; Pinto, G. P.; Toul, M.; Marek, M.; Hernychova, L.; Planas -\nIglesias, J.; Daniel Liskova, V.; Pluskal, D.; Vasina, M.; Emond, S.; Dorr, M.; Chaloupkova, R.; \nBednar, D.; Prokop, Z.; Hollfelder, F.; Bornscheuer, U. T.; Damborsky, J., En gineering the \nProtein Dynamics of an Ancestral Luciferase. Nat. Commun. 2021, 12, 3616. \n23. Sterner, R.; Höcker, B., Catalytic Versatility, Stability, and Evolution of the (betaalpha)8-\nBarrel Enzyme Fold. Chem. Rev. 2005, 105, 4038-4055. \n24. Nagano, N.; Orengo, C. A.; Thornton, J. M., One Fold With Many Functions: The \nEvolutionary Relationships Between TIM Barrel Families Based on Their Sequences, \nStructures and Functions. J. Mol. Biol. 2002, 321, 741-765. \n25. Wierenga, R. K., The TIM -Barrel Fold: A Versatile Framework for Efficient Enzymes. \nFEBS Lett. 2001, 492, 193-198. \n26. Claren, J.; Malisi, C.; Höcker, B.; Sterner, R., Establishing Wild-Type Levels of Catalytic \nActivity on Natural and Artificial (beta alpha)8-Barrel Protein Scaffolds. Proc. Natl. Acad. Sci. \nUSA 2009, 106, 3704-3709. \n27. Huang, P. S.; Feldmeier, K.; Parmeggiani, F.; Fernandez Velasco, D. A.; Höcker, B.; \nBaker, D., De Novo Design of a Four-Fold Symmetric TIM-Barrel Protein With Atomic-Level \nAccuracy. Nat. Chem. Biol. 2015, 12, 29-34. \n28. Röthlisberger, D.; Khersonsky, O.; Wollacott, A. M.; Jiang, L.; DeChancie, J.; Betker, \nJ.; Gallaher, J. L.; Althoff, E. A.; Zanghellini, A.; Dym, O.; Albeck, S.; Houk, K. N.; Tawfik, D. \nS.; Baker, D., Kemp Elimination Catalysts by Computational Enzyme Design. Nature 2008, \n453, 190-195. \n29. Ochoa-Leyva, A.; Barona-Gómez, F.; Saab-Rincón, G.; Verdel-Aranda, K.; Sánchez, \nF.; Soberón, X., Exploring the Structure-Function Loop Adaptability of a (beta/alpha)8-Barrel \nEnzyme Through Loop Swapping and Hinge Variability. J. Mol. Biol. 2011, 411, 143-157. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n35 \n \n30. Ochoa-Leyva, A.; Soberón, X.; Sánchez, F.; Argüello, M.; Montero -Morán, G.; Saab-\nRincón, G., Protein Design Through Systematic Catalytic Loop Exchange in the (beta/alpha)8 \nFold. J. Mol. Biol. 2009, 387, 949-964. \n31. Corbella, M.; Pinto, G. P.; Kamerlin, S. C. L., Loop Dynamics and the Evolution of \nEnzyme Activity. Nat Rev Chem 2023, 7, 536-547. \n32. Romero-Rivera, A.; Corbella, M.; Parracino, A.; Patrick, W. M.; Kamerlin, S. C. L., \nComplex Loop Dynamics Underpin Activity, Specificity, and Evolvability in the (betaalpha)8 \nBarrel Enzymes of Histidine and Tryptophan Biosynthesis. JACS Au 2022, 2, 943-960. \n33. Williams, J. C.; McDermott, A. E., Dynamics of the Flexible Loop of Triosephosphate \nIsomerase: The Loop Motion is Not Ligand Gated. Biochemistry 1995, 34, 8309-8319. \n34. Beismann-Driemeyer, S.; Sterner, R., Imidazole Glycerol Phosphate Synthase From \nThermotoga maritima. Quaternary Structure, Steady-State Kinetics, and Reaction Mechanism \nof the Bienzyme Complex. J Biol Chem 2001, 276, 20387-20396. \n35. Chaudhuri, B. N.; Lange, S. C.; Myers, R. S.; Chittur, S. V.; Davisson, V. J.; Smith, J. \nL., Crystal Structure of Imidazole Glycerol Phosphate Synthase: A Tunnel Through a \n(beta/alpha)8 Barrel Joins Two Active Sites. Structure 2001, 9, 987-997. \n36. Chaudhuri, B. N.; Lange, S. C.; Myers, R. S.; Davisson, V. J.; Smith, J. L., Toward \nUnderstanding the Mechanism of the Complex Cyclization Reaction Catalyzed by Imidazole \nGlycerolphosphate Synthase: Crystal Structures of a Ternary Complex and the Free Enzyme. \nBiochemistry 2003, 42, 7003-7012. \n37. Klem, T. J.; Davisson, V. J., Imidazole Glycerol Phosphate Synthase: The Glutamine \nAmidotransferase in Histidine Biosynthesis. Biochemistry 1993, 32, 5177-5186. \n38. Badger, J.; Sauder, J. M.; Adams, J. M.; Antonysamy, S.; Bain, K.; Bergseid, M. G.; \nBuchanan, S. G.; Buchanan, M. D.; Batiyenko, Y.; Christopher, J. A.; Emtage, S.; Eroshkina, \nA.; Feil, I.; Furlong, E. B.; Gajiwala, K. S.; Gao, X.; He, D.; Hendle, J.; Huber, A.; Hoda, K.; \nKearins, P.; Kissinger, C.; Laubert, B.; Lewis, H. A.; Lin, J.; Loomis, K.; Lorimer, D.; Louie, G.; \nMaletic, M.; Marsh, C. D.; Miller, I.; Molinari, J.; Muller -Dieckmann, H. J.; Newman, J. M.; \nNoland, B. W.; Pagarigan, B.; Park, F.; Pe at, T. S.; Post, K. W.; Radojicic, S.; Ramos, A.; \nRomero, R.; Rutter, M. E.; Sanderson, W. E.; Schwinn, K. D.; Tresser, J.; Winhoven, J.; Wright, \nT. A.; Wu, L.; Xu, J.; Harris, T. J., Structural Analysis of a Set of Proteins Resulting From a \nBacterial Genomics Project. Proteins 2005, 60, 787-796. \n39. List, F.; Vega, M. C.; Razeto, A.; Hager, M. C.; Sterner, R.; Wilmanns, M., Catalysis \nUncoupling in a Glutamine Amidotransferase Bienzyme by Unblocking the Glutaminase Active \nsite. Chem. Biol. 2012, 19, 1589-1599. \n40. Wurm, J. P.; Sung, S.; Kneuttinger, A. C.; Hupfeld, E.; Sterner, R.; Wilmanns, M.; \nSprangers, R., Molecular Basis for the Allosteric Activation Mechanism of the Heterodimeric \nImidazole Glycerol Phosphate Synthase Complex. Nat. Commun. 2021, 12, 2748. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n36 \n \n41. Massiere, F.; Badet -Denisot, M. A., The Mechanism of Glutamine-Dependent \nAmidotransferases. Cell. Mol. Life Sci. 1998, 54, 205-222. \n42. Raushel, F. M.; Thoden, J. B.; Holden, H. M., The Amidotransferase Family of \nEnzymes: Molecular Machines for the Production and Delivery of Ammonia. Biochemistry \n1999, 38, 7891-7899. \n43. Klem, T. J.; Chen, Y.; Davisson, V. J., Subunit Interactions and Glutamine Utilization \nby Escherichia coli Imidazole Glycerol Phosphate Synthase. J. Bacteriol. 2001, 183, 989-996. \n44. Lisi, G. P.; Currier, A. A.; Loria, J. P., Glutamine Hydrolysis by Imidazole Glycerol \nPhosphate Synthase Displays Temperature Dependent Allosteric Activation. Front. Mol. \nBiosci. 2018, 5, 4. \n45. Lisi, G. P.; East, K. W.; Batista, V. S.; Loria, J. P., Altering the Allosteric Pathway in \nIGPS Suppresses Millisecond Motions and Catalytic Activity. Proc. Natl. Acad. Sci. USA 2017, \n114, E3414-E3423. \n46. Lisi, G. P.; Manley, G. A.; Hendrickson, H.; Rivalta, I.; Batista, V. S.; Loria, J. P., \nDissecting Dynamic Allosteric Pathways Using Chemically Related Small-Molecule Activators. \nStructure 2016, 24, 1155-1166. \n47. Negre, C. F. A.; Morzan, U. N.; Hendrickson, H. P.; Pal, R.; Lisi, G. P.; Loria, J. P.; \nRivalta, I.; Ho, J.; Batista, V. S., Eigenvector Centrality For Characterization of Protein \nAllosteric Pathways. Proc. Natl. Acad. Sci. USA 2018, 115, E12201-E12208. \n48. Lang, D.; Thoma, R.; Henn-Sax, M.; Sterner, R.; Wilmanns, M., Structural Evidence for \nEvolution of the Beta/Alpha Barrel Scaffold by Gene Duplication and Fusion. Science 2000, \n289, 1546-1550. \n49. Douangamath, A.; Walker, M.; Beismann -Driemeyer, S.; Vega -Fernandez, M. C.; \nSterner, R.; Wilmanns, M., Structural Evidence for Ammonia Tunneling Across the (beta \nalpha)(8) Barrel of the Imidazole Glycerol Phosphate Synthase Bienzyme Complex. Structure \n2002, 10, 185-193. \n50. Liao, Q.; Kulkarni, Y.; Sengupta, U.; Petrovic, D.; Mulholland, A. J.; van der Kamp, M. \nW.; Strodel, B.; Kamerlin, S. C. L., Loop Motion in Triosephosphate Isomerase Is Not a Simple \nOpen and Shut Case. J. Am. Chem. Soc. 2018, 140, 15889-15903. \n51. Crean, R. M.; Biler, M.; van der Kamp, M. W.; Hengge, A. C.; Kamerlin, S. C. L., Loop \nDynamics and Enzyme Catalysis in Protein Tyrosine Phosphatases. J. Am. Chem. Soc. 2021, \n143, 3830-3845. \n52. Amaro, R. E.; Sethi, A.; Myers, R. S.; Davisson, V. J.; Luthey-Schulten, Z. A., A Network \nof Conserved Interactions Regulates the Allosteric Signal in a Glutamine Amidotransferase. \nBiochemistry 2007, 46, 2156-2173. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n37 \n \n53. Hubbard, S. J.; Beynon, R. J.; Thornton, J. M., Assessment of Conformational \nParameters as Predictors of Limited Proteolytic Sites in Native Protein Structures. Prot. Eng. \n1998, 11, 349-359. \n54. Kayode, O.; Wang, R.; Pendlebury, D. F.; Cohen, I.; Henin, R. D.; Hockla, A.; Soares, \nA. S.; Papo, N.; Caulfield, T. R.; Radisky, E. S., An Acrobatic Substrate Metamorphosis \nReveals a Requirement for Substrate Conformational Dynamics in Trypsin Proteolysis. J. Biol. \nChem. 2016, 291, 26304-26319. \n55. Jarymowycz, V. A.; Stone, M. J., Fast Time Scale Dynamics of Protein Backbones: \nNMR Relaxation Methods, Applications, and Functional Consequences. Chem Rev 2006, 106, \n1624-1671. \n56. Amaro, M.; Brezovsky, J.; Kovacova, S.; Sykora, J.; Bednar, D.; Nemec, V.; Liskova, \nV.; Kurumbang, N. P.; Beerens, K.; Chaloupkova, R.; Paruch, K.; Hof, M.; Damborsky, J., Site-\nSpecific Analysis of Protein Hydration Based on Unnatural Amino Acid Fluore scence. J. Am. \nChem. Soc. 2015, 137, 4988-4992. \n57. Wang, J.; Xie, J.; Schultz, P. G., A Genetically Encoded Fluorescent Amino Acid. J. \nAm. Chem. Soc. 2006, 128, 8738-8739. \n58. Gleason, P. R.; Kelly, P. I.; Grisingher, D. W.; Mills, J. H., An Intrinsic FRET Sensor of \nProtein-Ligand Interactions. Org. Biomol. Chem. 2020, 18, 4079-4084. \n59. Gleason, P. R.; Kolbaba-Kartchner, B.; Henderson, J. N.; Stahl, E. P.; Simmons, C. R.; \nMills, J. H., Structural Origins of Altered Spectroscopic Properties upon Ligand Binding in \nProteins Containing a Fluorescent Noncanonical Amino Acid. Biochemistry 2021, 60, 2577-\n2585. \n60. Ko, W.; Kim, S.; Lee, H. S., Engineering a Periplasmic Binding Protein For Amino Acid \nSensors With Improved Binding Properties. Org. Biomol. Chem. 2017, 15, 8761-8769. \n61. Dean, S. F.; Whalen, K. L.; Spies, M. A., Biosynthesis of a Novel Glutamate Racemase \nContaining a Site -Specific 7 -Hydroxycoumarin Amino Acid: Enzyme -Ligand Promiscuity \nRevealed at the Atomistic Level. ACS Central Sci. 2015, 1, 364-373. \n62. Mendes, K. R.; Martinez, J. A.; Kantrowitz, E. R., Asymmetric Allosteric Signaling in \nAspartate Transcarbamoylase. ACS Chem Biol 2010, 5, 499-506. \n63. Henderson, J. N.; Simmons, C. R.; Fahmi, N. E.; Jeffs, J. W.; Borges, C. R.; Mills, J. \nH., Structural Insights into How Protein Environments Tune the Spectroscopic Properties of a \nNoncanonical Amino Acid Fluorophore. Biochemistry 2020, 59, 3401-3410. \n64. Guengerich, F. P.; Child, S. A.; Barckhausen, I. R.; Goldfarb, M. H., Kinetic Evidence \nfor an Induced -Fit Mechanism in the Binding of the Substrate Camphor by Cytochrome \nP450(cam). ACS Catal. 2021, 11, 639-649. \n65. Vogt, A. D.; Di Cera, E., Conformational Selection or Induced Fit? A Critical Appraisal \nof the Kinetic Mechanism. Biochemistry 2012, 51, 5894-5902. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n38 \n \n66. Althoff, E. A.; Wang, L.; Jiang, L.; Giger, L.; Lassila, J. K.; Wang, Z.; Smith, M.; Hari, \nS.; Kast, P.; Herschlag, D.; Hilvert, D.; Baker, D., Robust Design and Optimization of Retroaldol \nEnzymes. Protein Sci. 2012, 21, 717-726. \n67. Jiang, L.; Althoff, E. A.; Clemente, F. R.; Doyle, L.; Röthlisberger, D.; Zanghellini, A.; \nGallaher, J. L.; Betker, J. L.; Tanaka, F.; Barbas, C. F., 3rd; Hilvert, D.; Houk, K. N.; Stoddard, \nB. L.; Baker, D., De Novo Computational Design of Retro-Aldol Enzymes. Science 2008, 319, \n1387-1391. \n68. Cheng, F.; Yang, J. H.; Bocola, M.; Schwaneberg, U.; Zhu, L. L., Loop Engineering \nReveals the Importance of Active-Site-Decorating Loops and Gating Residue in Substrate \nAffinity Modulation of Arginine Deiminase (An Anti-Tumor Enzyme). Biochem. Biophys. Res. \nCommun. 2018, 499, 233-238. \n69. Liu, B. B.; Qu, G.; Li, J. K.; Fan, W. C.; Ma, J. A.; Xu, Y.; Nie, Y.; Sun, Z. T., \nConformational Dynamics-Guided Loop Engineering of an Alcohol Dehydrogenase: Capture, \nTurnover and Enantioselective Transformation of Difficult -to-Reduce Ketones. Adv. Synth. \nCatal. 2019, 361, 3182-3190. \n70. Park, H. S.; Nam, S. H.; Lee, J. K.; Yoon, C. N.; Mannervik, B.; Benkovic, S. J.; Kim, H. \nS., Design and Evolution of New Catalytic Activity with an Existing Protein Scaffold. Science \n2006, 311, 535-538. \n71. Jogl, G.; Rozovsky, S.; McDermott, A. E.; Tong, L., Optimal Alignment for Enzymatic \nProton Transfer: Structure of the Michaelis Complex of Triosephosphate Isomerase at 1.2-A \nResolution. Proc. Natl. Acad. Sci. USA 2003, 100, 50-55. \n72. Katebi, A. R.; Jernigan, R. L., The Critical Role of the Loops of Triosephosphate \nIsomerase for its Oligomerization, Dynamics, and Functionality. Prot. Sci. 2014, 23, 213-228. \n73. Richard, J. P.; Amyes, T. L.; Goryanova, B.; Zhai, X., Enzyme Architecture: On the \nImportance of Being in a Protein Cage. Curr. Opin. Chem. Biol. 2014, 21, 1-10. \n74. Richard, J. P.; Amyes, T. L.; Malabanan, M. M.; Zhai, X.; Kim, K. J.; Reinhardt, C. J.; \nWierenga, R. K.; Drake, E. J.; Gulick, A. M., Structure -Function Studies of Hydrophobic \nResidues That Clamp a Basic Glutamate Side Chain during Catalysis by Triosep hosphate \nIsomerase. Biochemistry 2016, 55, 3036-3047. \n75. Joseph, D.; Petsko, G. A.; Karplus, M., Anatomy of a Conformational Change: Hinged \n\"Lid\" Motion of the Triosephosphate Isomerase Loop. Science 1990, 249, 1425-1428. \n76. Rozovsky, S.; Jogl, G.; Tong, L.; McDermott, A. E., Solution-state NMR Investigations \nof Triosephosphate Isomerase Active Site Loop Motion: Ligand Release in Relation to Active \nSite Loop Dynamics. J. Mol. Biol. 2001, 310, 271-280. \n77. Albery, W. J.; Knowles, J. R., Free -Energy Profile for Reaction Catalyzed by \nTriosephosphate Isomerase. Biochemistry 1976, 15, 5627-5631. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n39 \n \n78. Hennig, M.; Darimont, B.; Sterner, R.; Kirschner, K.; Jansonius, J. N., 2.0 Å Structure \nof Indole-3-Glycerol Phosphate Synthase from the Hyperthermophile Sulfolobus solfataricus: \nPossible Determinants of Protein Stability. Structure 1995, 3, 1295-1306. \n79. Hennig, M.; Darimont, B. D.; Jansonius, J. N.; Kirschner, K., The Catalytic Mechanism \nof Indole-3-Glycerol Phosphate Synthase: Crystal Structures of Complexes of the Enzyme \nfrom Sulfolobus solfataricus with Substrate Analogue, Substrate, and Product. J. Mol. Biol. \n2002, 319, 757-766. \n80. Schlee, S.; Dietrich, S.; Kurćon, T.; Delaney, P.; Goodey, N. M.; Sterner, R., Kinetic \nMechanism of Indole-3-Glycerol Phosphate Synthase. Biochemistry 2013, 52, 132-142. \n81. O'Rourke, K. F.; Jelowicki, A. M.; Boehr, D. D., Controlling Active Site Loop Dynamics \nin the (beta/alpha)8 Barrel Enzyme Indole-3-Glycerol Phosphate Synthase. Catalysts 2016, 6, \n129-142. \n82. Zaccardi, M. J.; O'Rourke, K. F.; Yezdimer, E. M.; Loggia, L. J.; Woldt, S.; Boehr, D. \nD., Loop-Loop Interactions Govern Multiple Steps in Indole-3-Glycerol Phosphate Synthase \nCatalysis. Protein Sci. 2014, 23, 302-311. \n83. Schlee, S.; Klein, T.; Schumacher, M.; Nazet, J.; Merkl, R.; Steinhoff, H. J.; Sterner, R., \nRelationship of Catalysis and Active Site Loop Dynamics in the (betaalpha)8 -Barrel Enzyme \nIndole-3-glycerol Phosphate Synthase. Biochemistry 2018, 57, 3265-3277. \n84. Due, A. V.; Kuper, J.; Geerlof, A.; von Kries, J. P.; Wilmanns, M., Bisubstrate Specificity \nin Histidine/Tryptophan Biosynthesis Isomerase from Mycobacterium tuberculosis by Active \nSite Metamorphosis. Proc. Natl. Acad. Sci. USA 2011, 108, 3554-3559. \n85. Henn-Sax, M.; Thoma, R.; Schmidt, S.; Hennig, M.; Kirschner, K.; Sterner, R., Two \n(betaalpha)(8)-Barrel Enzymes of  Histidine and Tryptophan Biosynthesis Have Similar \nReaction Mechanisms and Common Strategies for Protecting Their Labile Substrates. \nBiochemistry 2002, 41, 12032-12042. \n86. Soderholm, A.; Guo, X. H.; Newton, M. S.; Evans, G. B.; Nasvall, J.; Patrick, W. M.; \nSelmer, M., Two -Step Ligand Binding in a (beta alpha)(8) Barrel Enzyme: Substrate -bound \nStructures Shed New Light on the Catalytic Cycle of HisA. J. Biol. Chem. 2015, 290, 24657-\n24668. \n87. Malabanan, M. M.; Amyes, T. L.; Richard, J. P., A Role for Flexible Loops in Enzyme \nCatalysis. Curr. Opin. Struct. Biol. 2010, 20, 702-710. \n88. Kneuttinger, A. C.; Straub, K.; Bittner, P.; Simeth, N. A.; Bruckmann, A.; Busch, F.; \nRajendran, C.; Hupfeld, E.; Wysocki, V. H.; Horinek, D.; Konig, B.; Merkl, R.; Sterner, R., Light \nRegulation of Enzyme Allostery through Photo-responsive Unnatural Amino Acids. Cell Chem. \nBiol. 2019, 26, 1501-1514. \n89. Davisson, V. J.; Deras, I. L.; Hamilton, S. E.; Moore, L. L., A Plasmid-Based Approach \nfor the Synthesis of a Histidine Biosynthetic Intermediate. J. Org. Chem. 1994, 59, 137-143. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n40 \n \n90. Kuzmic, P., DynaFit--A Software Package for Enzymology. Methods Enz. 2009, 467, \n247-280. \n91. Thoma, R.; Obmolova, G.; Lang, D. A.; Schwander, M.; Jeno, P.; Sterner, R.; \nWilmanns, M., Efficient Expression, Purification and Crystallisation of Two Hyperthermostable \nEnzymes of Histidine Biosynthesis. FEBS Lett 1999, 454, 1-6. \n92. Kabsch, W., Automatic Processing of Rotation Diffraction Data from Crystals of Initially \nUnknown Symmetry and Cell Constants. J. Appl. Crystallogr. 1993, 26, 795-800. \n93. Adams, P. D.; Grosse -Kunstleve, R. W.; Hung, L. W.; Ioerger, T. R.; McCoy, A. J.; \nMoriarty, N. W.; Read, R. J.; Sacchettini, J. C.; Sauter, N. K.; Terwilliger, T. C., PHENIX: \nBuilding New Software for Automated Crystallographic Structure Determination. Acta \nCrystallogr. D 2002, 58, 1948-1954. \n94. Potterton, L.; McNicholas, S.; Krissinel, E.; Gruber, J.; Cowtan, K.; Emsley, P.; \nMurshudov, G. N.; Cohen, S.; Perrakis, A.; Noble, M., Developments in the CCP4 Molecular-\nGraphics Project. Acta Crystallogr D 2004, 60, 2288-2294. \n95. Murshudov, G. N.; Vagin, A. A.; Dodson, E. J., Refinement of Macromolecular \nStructures by the Maximum-Likelihood Method. Acta Crystallogr D 1997, 53, 240-255. \n96. Emsley, P.; Cowtan, K., Coot: Model-Building Tools for Molecular Graphics. Acta \nCrystallogr. D 2004, 60, 2126-2132. \n97. Lipchock, J. M.; Loria, J. P., 1H, 15N and 13C Resonance Assignment of Imidazole \nGlycerol Phosphate (IGP) Synthase Protein HisF from Thermotoga maritima. Biomol. NMR \nAssign. 2008, 2, 219-221. \n98. Sattler, M.; Schleucher, J.; Griesinger, C., Heteronuclear Multidimensional NMR \nExperiments for the Structure Determination of Proteins in Solution Employing Pulsed Field \nGradients. Prog. Nucl. Mag. Res. Sp. 1999, 34, 93-158. \n99. Lakomek, N. A.; Ying, J. F.; Bax, A., Measurement of N -15 Relaxation Rates In \nPerdeuterated Proteins by TROSY-Based Methods. J. Biomol. NMR 2012, 53, 209-221. \n100. Delaglio, F.; Grzesiek, S.; Vuister, G. W.; Zhu, G.; Pfeifer, J.; Bax, A., N MRpipe - A \nMultidimensional Spectral Processing System Based on Unix Pipes. J. Biomol. NMR 1995, 6, \n277-293. \n101. Keller, R., The Computer Aided Resonance Assignment Tutorial. Cantina Verlag: 2004. \n102. Schrodinger, LLC, The PyMOL Molecular Graphics System, Version 1.8. 2015. \n103. Hopkins, C. W.; Le Grand, S.; Walker, R. C.; Roitberg, A. E., Long-Time-Step Molecular \nDynamics through Hydrogen Mass Repartitioning. J. Chem. Theory Comput. 2015, 11, 1864-\n1874. \n104. Maier, J. A.; Martinez, C.; Kasavajhala, K.; Wickstrom, L.; Hauser, K. E.; Simmerling, \nC., ff14SB: Improving the Accuracy of Protein Side Chain and Backbone Parameters from \nff99SB. J. Chem. Theory Comput. 2015, 11, 3696-3713. \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n41 \n \n105. Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L., \nComparison of simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 1983, \n79, 926-935. \n106. Case, D. A.; Aktulga, H. M.; Belfon, K.; Ben-Shalom, I. Y.; Berryman, J. T.; Brozell, S. \nR.; Cerutti, D. S.; Cheatham, I., T. E.; Cisneros, G. A.; Cruzeiro, V. W. D.; Darden, T. A.; \nForouzesh, N.; Giambasu, G.; Giese, T.; Gilson, M. K.; Gohlke, H.; Go etz, A. W.; Harris, J.; \nIzadi, S.; Izmailov, S. A.; Kasavajhala, K.; Kaymak, M. C.; King, E.; Kovalenko, A.; Kurtzman, \nT.; Lee, T. S.; Li, P.; Lin, C.; Liu, J.; Luchko, T.; Luo, R.; Machado, M.; Man, V.; Manathunga, \nM.; Merz, K. M.; Miao, Y.; Mikhailovskii, O.; Monard, G.; Nguyen, H.; O´Hearn, K. A.; Onufriev, \nA.; Pan, F.; Pantano, S.; Qi, R.; Rahnamoun, A.; Roe, D. R.; Roitberg, A.; Sagui, C.; Schott -\nVerdugo, S.; Shajan, A.; Shen, C. L.; Simmerling, C. L.; Skrynnikov, N. R.; Smith, J.; Swails, \nJ.; Walker, R. C.; Wang, J.; Wang, J.; Wei, H.; Wu, X.; Wu, Y.; Xiong, Y.; Xue, Y.; York, D. M.; \nZhao, S.; Zhu, Q.; Kollman, P. A., Amber 2023. University of California, San Francisco: 2023. \n \n  \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint \n\n42 \n \nTOC graphic \n \n  \n \n \n.CC-BY 4.0 International licensemade available under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is \nThe copyright holder for this preprintthis version posted June 22, 2024. ; https://doi.org/10.1101/2024.06.21.600150doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}