Mechanism of Oxygenase Pathway Reactions Catalysed by Rubisco

Subscriber access provided by ECU Libraries

B: Biophysics; Physical Chemistry of Biological Systems and Biomolecules

Mechanism of Oxygenase Pathway Reactions Catalysed by Rubisco from Large Scale Kohn-Sham Density Functional Calculations Babu Kannappan, Peter L. Cummins, and Jill E. Gready J. Phys. Chem. B, Just Accepted Manuscript • DOI: 10.1021/acs.jpcb.9b00518 • Publication Date (Web): 07 Mar 2019 Downloaded from http://pubs.acs.org on March 9, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Mechanism of Oxygenase Pathway Reactions Catalysed by Rubisco from Large Scale Kohn-Sham Density Functional Calculations

Babu Kannappan,* Peter L. Cummins and Jill E. Gready John Curtin School of Medical Research, The Australian National University, Canberra ACT 0200, Australia

*

To whom correspondence should be addressed. E-mail: [email protected]

ACS Paragon Plus Environment

1

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 38

Abstract Ribulose 1,5-bisphosphate carboxylase-oxygenase (Rubisco) is the primary carbon-fixing enzyme in photosynthesis, fixing CO2 to a 5-carbon sugar, RuBP, in a series of 5 reactions. However, it also catalyses an oxygenase reaction by O2 addition to the same enolized RuBP substrate in an analogous reaction series in the same active site, producing a waste product and loss of photosynthetic efficiency. Starting from RuBP, the reactions are: enolization to the enediolate form; addition of CO2 or O2 to form the carboxy or peroxo adduct; hydration to a form a gemdiolate; scission of the C2-C3 bond of the original RuBP; and stereospecific or nonstereospecific protonation to form 2 molecules of 3-carbon PGA product, or one molecule of PGA, one of 2-carbon PG (waste product) and one water molecule. Reducing the loss of efficiency from the oxygenase reaction is an attractive means to increasing crop productivity. However, lack of understanding of key aspects of the catalytic mechanisms for both the carboxylase and oxygenase reactions, particularly those involving proton exchanges and roles of water molecules, have stymied efforts at re-engineering Rubisco to reduce losses from the oxygenation reaction. As the stable form of molecular oxygen is the triplet biradical state (3O2), its reaction with near-universal singlet-state molecules is formally spin forbidden. Although in oxygenase enzymes, 3O2 activation is usually achieved by one-electron transfers using transition metal ions or organic cofactors, recently, cofactor-less oxygenases in which the substrate itself is the source of the electron for 3O2 activation have been identified but in all such cases an aromatic ring stabilizes the substrate's negative charge. Here we present the first large-scale Kohn-Sham DFT study of the reaction mechanism of the Rubisco oxygenase pathway. First, we show that the enediolate substrate complexed to Mg2+ and its ligands extends the region for charge delocalisation and stabilization of its negative charge to allow formation of a caged biradical enediolate-O2 complex. Thus, Rubisco is a unique type of oxygenase without precedent in the literature. Second, for the O2 addition to proceed to the singlet peroxo-adduct intermediate, the system must undergo an intersystem crossing (ISC). We found that the presence of protonated LYS334 is required to stabilize this intermediate, and that both factors (strongly stabilized anion and protonated LYS334) facilitate a barrier-less activation of 3O2. This finding supports our recent proposal that deoxygenation, i.e. reversal of gas binding, is possible. Third, as neither CO2 nor O2 bind to the enzyme, our findings support the proposal from our recent carboxylase study that the observed KC or KO (Michaelis-Menten constants) in the steady-state kinetics reflect the


2

Page 3 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


respective adducts, carboxy or peroxo. Fourth, after computing hydration pathways with water addition both syn and anti to C3, we found, in contrast to the results of our carboxylation study indicating anti addition, that in the oxygenation reaction only syn hydration is capable of producing a stable gemdiolate that facilitates the rate-limiting C2-C3 bond scission to final products. Fifth, we propose that an excess proton we previously found was required in the carboxylation reaction for activating the C2-C3 bond scission is utilized in the oxygenation reaction for the required elimination of a water molecule. In summary: despite its oxygenase handicap, Rubisco's success in directing 75% of its substrate through the carboxylation pathway can be considered impressively effective. Although native C3 Rubiscos are in a fix with unwanted activity of 3O2 hampering its primary carboxylase function, mechanistic differences presented here with findings in our recent carboxylase study for both the gas-addition and subsequent reactions provide some clues as to how creative Rubisco re-engineering may offer a solution to reducing the oxygenase activity.


3


Page 4 of 38

Introduction Ribulose 1,5-bisphosphate carboxylase-oxygenase (EC 4.1.1.39, Rubisco) underpins life on earth; it is the primary carbon-fixing enzyme present in photosynthetic organisms and, thus, the gateway for transfer of carbon from inorganic CO2 in the atmosphere into the biosphere.1-2 However, Rubisco’s evolution over geological time has been complicated by huge changes in the absolute and relative concentrations of atmospheric CO2 and O2 and by the complex and difficult chemistry in Rubisco’s multi-step mechanism. Its catalytic turnover is only ~3 CO2 fixations/sec in C3 plants. Consequently, a very large amount of Rubisco is required to maintain adequate photosynthesis; it constitutes up to 50% of soluble protein in leaves and is the most abundant protein on earth.1, 3 Rubisco efficiency is, thus, a major factor limiting photosynthetic potential and crop productivity.4 Engineering a better Rubisco and implementing it in a plant showing improved photosynthetic productivity is a Holy Grail in plant biology.5 But despite much effort over the last three decades, there are no literature reports of success although we have achieved promising results based on a patented Rubisco re-engineering method6 and follow-up assessment of transgenic plants. The root of Rubisco’s catalytic dilemma lies at its origins in an anoxic world and its apparently limited capacity to adapt to an O2-containing atmosphere – which, paradoxically, was produced by photosynthesis. Development of the O2-containing atmosphere, thus, presented Rubisco with a novel gas absent at its birth that is a competitive substrate to CO2 in an oxygen-fixation reaction. Unfortunately, the resultant products need to be recycled by photorespiration, which wastes carbon and energy. In the current atmosphere of 408 ppm CO2 (0.041%, https://www.esrl.noaa.gov/gmd/ccgg/trends/weekly.html) and 21% O2, photorespiration can inhibit photosynthesis of C3 plants by more than 30% at warmer temperatures (> 30°C).7 The


4

Page 5 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


complexity of the reaction sequence catalysed by Rubisco, which constrains the active-site sequence space available for mutation6, is a likely major factor for this evolutionary block. In addition, several other nuclear-encoded enzymes are involved in assembling and maintaining the functional form of Rubisco, which is a hexadecameric complex of 8 large and 8 small subunits (L8S8). The importance of interactions of Rubisco with these enzymes also limits the sequence space available for Rubisco to re-tune its catalysis evolutionarily.

Rubisco’s unwanted ability to catalyse addition of O2 instead of CO2 arises from the qualitative similarity of the electrostatic potential (ESP) of CO2 and O2 in its normal triplet state (the ESP for singlet O2 is quite different), making it difficult for Rubisco to differentiate between the two small gaseous substrates.8 On the productive pathway, Rubisco catalyses a series of five reactions starting from D-ribulose 1,5-bisphosphate (RuBP) to form two molecules of 3phospho-D-glycerate (3PGA) as products. The first reaction intermediate formed in the reaction sequence, the enediolate of RuBP, is unstable and susceptible to several alternate chemical fates,9 as shown in Scheme 1. To continue on the productive reaction pathway it must react with CO2. However, the negative charge localized on the enediolate necessary to promote nucleophilic addition to CO2 allows several alternative reactions: reprotonation on the wrong face of the molecular plane leading to the inhibitor xylulose-1,5-bisphosphate; phosphate elimination leading to a wasteful pentosephosphate side product; or, dominantly, oxygenation leading finally to one molecule of 3PGA and one molecule of a wasteful side product (2-phosphoglycolate (2PG)). Metabolism of the latter in photorespiration results in loss of carbon fixed previously by Rubisco, as well as energy in the form of ATP and NADPH. Given Rubisco’s hard task of minimizing alternate reactions while catalysing five reaction steps on the productive pathway


5


Page 6 of 38

within the same active site using and re-using the same set of amino acid residues, the typical performance of a C3 Rubisco in successfully directing 75% of the substrate through the carboxylation pathway10 can be considered as impressively effective. Despite being labelled for a long time as a notoriously slow enzyme with confused specificity, recent analysis of kinetic data from a large enzyme data base suggests Rubisco's catalytic rate is not atypical.11-12 However, in interpreting this result consideration needs to be given to an enzyme's metabolic context and the strength of selection pressure for improved efficiency. The latter is very strong for Rubisco but less so for most of the other slowish enzymes reported. The conundrums of Rubisco's slow turnover and significant oxygenase activity still require further explanation.

Hence, despite Rubisco’s fundamental importance in photosynthesis, there is still a lack of a detailed understanding of the details of the chemical mechanisms of the carboxylase and oxygenase reactions at a molecular level. Such knowledge would shed light on the above conundrums. This is partly due to the complexity of Rubisco catalysis, involving multiple steps and intermediates. As there are no reliable X-ray structures for complexes of Rubisco with most of the intermediates other than the analogue for carboxylated intermediate, information on the protonation states of these intermediates and H-bond networks are not directly inferable from experimental data. Thus, computational methods for studying large biomolecular systems, which have witnessed significant progress in recent decades,13-14 can play a significant role in unravelling details of the mechanism of Rubisco catalysis. Increase in computational power over the years have gradually allowed us to increase understanding of the mechanistic details of the carboxylase reaction mechanism.8, 15-17


6

Page 7 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


As the stable form of molecular oxygen is a biradical species that exists in the triplet electronic state (3O2), its reaction with the singlet-state molecules that comprise most chemical species on earth is a spin-forbidden process. Thus, although the oxidation of organic substrates with 3O2 is thermodynamically feasible, it is kinetically sluggish.18 The relative kinetic inertness of 3O2, necessitates its activation, usually via a series of one-electron transfer reactions.19 Nature has achieved this by recruiting transition metal ions, iron and copper in particular, or organic cofactors for enzymes catalysing 3O2 addition reactions. Transition metal ions exist in multiple oxidation states and can activate 3O2 by a single electron transfer. Transition metal ions also have large spin orbit coupling and, thus, the probability of spin-forbidden transitions in complexes with transition metal ions are nearly as high as spin-allowed transitions.20-21 On the other hand, several oxygenase enzymes use redox-active heteroaromatic cofactors such as flavins, pterins or quinones, and their derivatives, as redox catalysts for 3O2 activation.22-24 Additional or unpaired electrons in these species are highly delocalised promoting stability of the chemical species, which then activate 3O2 by one electron donation.

In recent decades, many enzymes have been identified to be cofactor-less oxygenases.22-23 In these cases the substrate itself is the source of the electron for 3O2 activation. However, in all reported examples, the substrate has an aromatic ring to stabilize the negative charge on the substrate prior to 3O2 activation. On the other hand, in Rubisco the substrate for the O2 addition step is an enediolate (Scheme 2). A comparison of 3O2 activation by a flavin-containing enzyme and substrates of some cofactor-less oxygenases is shown in Figure 1. The apparent potential for delocalization of the negative charge on the substrate is substantial in examples A to D, compared with what might be expected for the enediolate, shown in E. However, in Rubisco,


7


Page 8 of 38

coordination of O2 and O3 of the enediolate substrate to Mg2+ extends the potential region across the metal coordination sphere for charge delocalisation and modulation of its negative charge. This has the possibility to produce a species with the right mix of stability and redox potential to be capable of transferring an electron to 3O2. If so, Rubisco is a unique type of oxygenase as there is no precedent in the literature for this chemistry.

In this work we present a Kohn-Sham density functional theory computational study of the mechanism of reactions in the oxygenase pathway of Rubisco, including consideration of the potential for the formation of a caged biradical O2 complex as in Figure 1(E). Based on these results, together with the results of our previous computational study of the carboxylation reaction, we put forward a comprehensive mechanistic description of the reactions catalysed by Rubisco.

Methods

Active-Site Models. We started by working to find as large a fragment model as possible that would allow QM calculations with reasonable accuracy and also allow the use of TS search algorithms. After initial attempts using 469-atom and 339-atom models used in our earlier study of the carboxylase reaction (without TS optimisation),17 we found we needed to scale down the size of the fragment model to 248 atoms (fragment model FM23, illustrated in Figure S1 with residue composition shown in Table S1 in Supporting Information). With the FM23 model it was possible to compute analytical Hessians and, hence, a TS optimization using the G16 program.25


8

Page 9 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


This was a critical factor due to the number of degrees of freedom as well as the number of alternate stationary points that needed to be computed for both minima and TSs. Starting structures were generated from the crystal structure 8RUC (activated spinach Rubisco with Mg2+ and 2-carboxyarabinitol 1,5-bisphosphate - 2CABP).26 Model FM23 included all active-site residues that directly interact with CABP in the crystal structure.

Computational Methods. Constrained energy-minimization calculations were carried out for all the stationary points along the reaction pathway using the NWCHEM program.27 For minimumenergy structures, only the link atoms and oxygen atoms of some crystallographic waters were fixed (see Table S1 in Supporting Information for details). Subsequent frequency calculations were performed using the Gaussian16 program.25 For transition states, starting structures were optimized with the NWCHEM program by additionally constraining the distance between atoms involved in bond formation/breaking. Once a reasonable guess geometry was obtained by taking steps along the reaction coordinate, TS optimization was carried out with the G16 program by calculating force constants at every step with the Berny algorithm using GEDIIS.28 All geometry optimizations were carried out at the B3LYP/6-31G(d,p) level of theory.

Single-point

calculations were also carried out at the B3LYP/6-31G(d) level. Although B3LYP/6-31G(d) is widely used, its shortcomings are well known.29-30 However, Kruze et al.31 have shown a simple way to calculate and apply a correction that removes two major deficiencies: missing London dispersion effects and basis set superposition error. The correction scheme, named gCP-D3, has been shown to provide improvements in both reaction energies and barrier heights for the B3LYP/6-31G(d) functional. As gCP-D3 corrections were not available at B3LYP/631G(d,p)


9


Page 10 of 38

level, we applied the corrections on single-point energies at B3LYP/6-31G(d) level computed for geometries optimized at B3LYP/6-31G(d,p) level.

Results

Encounter complex. Two minimum energy structures (A and B) were obtained on optimization of the triplet-state complex of enzyme-bound enediolate and O2. The main difference between the two structures was the orientation and distance of the O2 molecule relative to the enediolate, with the distances between O2 and the C2 carbon of the enediolate being 3.05 Å and 2.78 Å, respectively, for structures A and B. Structure B also exhibits a longer O-O bond length (1.24 Å vs 1.22 Å), greater negative Mulliken charge on O2 (-0.17 vs -0.07) and lower sum of atomic spin densities on O2 (1.76 vs 1.88) compared with structure A, showing that there is greater partial electron transfer from the enediolate to the O2 molecule in structure B. Structure B is also lower in energy by 8 kcal/mol. This preferred structure is henceforth referred to as structure I in the oxygenase reaction pathway. Figure 2 shows an isosurface of electron spin density of structure I (Figure 2A) together with the substrate…Oxygen complexes for two other cofactor-free oxygenases (Figures 2B and 2C) and a 4-carbon mimic of the enediolate (Figure 2D). Analysis of the spin density distribution as well as the charge on O2 and other geometric parameters clearly show that I has biradical characteristics with a partial electron transferred from the enediolate to 3O2. The electron transfer appears to proceed without an energy barrier, similarly to the case of activation of 3O2 by the FADH2 cofactor in glucose oxidase (GO) reported by Prabhakar et al.32-33; the spin density plot for the cofactor in GO is shown in Figure 2B. Prabhakar et al.33 suggested that the barrier-less energy transfer in GO is facilitated by a nearby


10

Page 11 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


conserved protonated histidine. Our earlier carboxylation study8 and the analysis of Cleland et al.34 suggest that in Rubisco LYS334 stabilizes the TS and product for the CO2 addition reaction; a similar role can be speculated for O2 addition. Our recent work17 also proposed an additional proton channelled from elsewhere to the active site via GLU60 may serve this role assisted by LYS334. Comparison of interatomic distances between carboxylation and oxygenation reaction species using the FM23 model are shown in Table S2 in Supporting Information.

We also explored this initial oxygenation step using an additional proton in the vicinity by protonating GLU60. This fragment model is, henceforth, referred to as model FM23b. In the resulting optimised complex the O-O bond length is 1.33Å (characteristic of bond order 1), the sum of atomic spin densities on O2 is 1.0 and the Mulliken charge on O2 is -0.61. Taken together, all of this points to a complete transfer of an electron from the enediolate to O2 and the formation of a fully-fledged biradical complex. Comparison of atomic spin density plots of FM23 and FM23b is shown in Figure S2 in Supporting Information.

Addition of O2. For the O2 addition reaction, the system must undergo an intersystem crossing (ISC) before reaching the subsequent singlet product. The C2-peroxo product is achieved by direct coupling of the two radical species and the ISC likely occurs at the TS structure. Accurate prediction of the structure for the triplet to singlet spin conversion would require use of a multiconfiguration SCF (MCSCF) method and complete active-space calculation. However, MCSCF calculations are infeasible for a system of the required size. Consequently, we used the utility program from Harvey’s research group35 to compute the minimum energy crossing point (MECP) between the triplet and singlet potential energy surfaces (PES). The location of the


11


Page 12 of 38

MECP on the PES is given in Figure 3 together with the complete reaction-energy profile of the oxygenase reactions. The MECP geometry can qualitatively be taken as the geometry at which the ISC occurs. Attempts to optimize a triplet transition state, to explore the direct-attack mechanism, failed. Optimization of the MECP structure in the triplet state led back to the encounter complex, while optimization in the singlet state led to the peroxo intermediate (III). Optimized structures for the enediolate intermediate (I), MECP (II) and peroxo intermediate (III) are provided in Figure S3 in Supporting Information.

The MECP is 10 kcal/mol higher in energy than the preceding complex for model FM3 and 6.6 kcal/mol higher when an additional proton is provided through GLU60 in model FM23b. Interestingly, the geometrical parameters for the MECP (II) are not significantly different from that in structure I, both in the presence and absence of an additional proton. In II, the distance between O2 and C2 decreases from 2.78 Å (structure I) to 2.12 Å, which is much shorter than the C(CO2)…C2 distance (2.45 Å) for the TS in the CO2 addition. However, for model FM23b, where there is an additional proton near O2, the distance between O2 and C2 decreases from 2.98 Å (in I) to 2.46 Å (in II). At the MECP (II) the O―O bond elongates only slightly from 1.24 Å in I to 1.31 Å (1.33 Å in I to 1.35 Å in the model FM23b). Thus, some of the geometrical changes that one would expect to be achieved in the TS structure for O2 addition are already in place in the biradical complex (I) and the structural changes required for taking the singlet and triplet PESs to nearly equal energy are minimal. This accounts for the modest activation energy of 7-10 kcal/mol for the triplet to singlet transition.


12

Page 13 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


The geometry for the peroxo intermediate (PKABP) was obtained by optimizing II with decreasing C2…OS1 distance (see OS1 and OS2 labelling in Figure S3 in Supporting Information. With the model FM23 both CO2 and O2 addition appear to proceed without any proton exchange leading to a charged carboxylate/peroxo group which forms H-bonds with the Mg-coordinated water molecule (H2O[Mg]) and H+-LYS334. There is no significant difference in the optimized structures of PKABP and CKABP. Our earlier work (unpublished) using several small fragment models (Figure S4 in Supporting Information) showed that the charge on LYS334 has significant impact on the initial electron transfer to O2, but negligible impact on the energetics of O2 addition itself (Figure 4), whereas for CO2 addition the charge on LYS334 has significant impact on the activation barrier as well as the reaction energy for the reaction step. For carboxylation, as the charge on LYS334 increases from 0 to 1 (through the fragment models in the order of increasing charge from FM18, FM17, FM20 and to FM16) the reaction energy changes from 11 kcal/mol to -32 kcal/mol, while for oxygen addition the reaction energy only varies between -19 kcal/mol and -22 kcal/mol (Figure 4A). Similarly the activation barrier for CO2 addition changes from 15 kcal/mol to 4 kcal/mol, whereas for O2 addition the activation barrier shifts only by about 3 kcal/mol. The largest of these models had only 77 atoms. The current fragment model FM23 with 248 atoms is more representative of the active site of Rubisco. The reaction energy for oxygen addition with the FM23 model is -12 kcal/mol whereas it is -31 kcal/mol for the FM23b model (Figure 3).

Thus in summary, the O2 addition involves firstly a single electron transfer step forming a biradical complex (I), followed by C-O covalent bond formation via a MECP (II) to yield the (singlet) peroxo intermediate (III).


13


Page 14 of 38

Hydration. Our recent work for the carboxylase pathway17 showed that the water for hydration more likely comes from the direction anti to that of approach of the gas substrate, with KCX201 acting as a base to deprotonate the water. We also found that the introduction of an additional proton on the syn-side results in a large reduction in the energy barrier of the (later) C2-C3 bond scission reaction. In this work we also explored the potential energy surfaces for the reaction of the hydrating water molecule approaching from both the syn- and anti-directions to the peroxo group, with (H+-syn and H+-anti pathways using model FM23b) and without (syn- and antipathways using model FM23) an additional proton on the peroxo group prior to hydration.

Syn-hydration. In contrast to our results for the carboxylase pathway,17 the activation barrier for syn hydration of the peroxo product is only 9.5 kcal/mol (Figure 3). The negatively charged peroxo group acts as the base for syn hydration. This requires the peroxo group to rotate about the C2-OS1 bond and for OS2 to break away from H bonds with LYS334 and ASN123 (C2-OS1OS2…H). In the TS structure, the Ow atom moves out of Mg-coordination to a distance of 3.2 Å from Mg2+. The subsequent structure along the reaction pathway, the gemdiolate intermediate (V), is highly stable (-8.6 kcal/mol) relative to the peroxo intermediate (Figure 3).

Anti-hydration. For the anti-hydration pathway an additional solvent water was introduced near KCX201 in the peroxo product complex (III) and optimized. A local minimum was obtained with the added water molecule at 3.2 Å from the C3-carbon and forming an H bond with KCX201. Introduction of this additional water molecule causes subtle changes in H- and coordination bonds, which in turn result in the re-orientation of the peroxo group and the proton


14

Page 15 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


from H2O[Mg] being transferred to the peroxo group even before C3-hydration. The transferred proton is 1.4 Å from the hydroxide ion which remains coordinated to Mg2+. The H bond between the peroxo group and LYS334 is disrupted in this structure. Geometry optimization following the reaction coordinate (Ow(solvent)…C3 distance) leads to a transition state at a Ow…C3 distance of 1.8 Å, a much shorter length than the Ow…C3 distance in the syn-hydration TS (2.3 Å). The energy barrier for the anti-hydration is 19.6 kcal/mol (Figure 3). Continuing further along the reaction coordinate leads to the gemdiolate intermediate, which is lower in energy than the peroxo intermediate by 5.8 kcal/mol.

H+-syn-hydration In the next computational experiment the peroxo group in structure III was protonated and re-optimized. The added proton forms a short H bond with GLU60 and causes weakening of the H bonds with LYS334…Os2 and (HIS294)H…O3. In the syn-hydration with the additional proton GLU204 acts as the base to activate H2O[Mg] for hydration. The activation barrier and reaction energy for H+-syn-hydration are 18.4 and 12.4 kcal/mol, respectively.

H+-anti-hydration Similar to the case of syn-hydration with an additional proton, in the starting geometry for anti-hydration the additional proton (H+-anti) on the peroxo group forms a short H bond with GLU60 and causes subtle perturbations to other interactions in the vicinity. The TS is formed when the Ow atom is at 1.8 Å from the C3 carbon atom and the minimum energy structure for the gemdiolate at an Ow…C3 distance of 1.47 Å, which is slightly lower than the distance in the anti-hydration pathway. The activation barrier and reaction energy for H+-antihydration are 14.1 and 5.9 kcal/mol, respectively. The optimized hydration TS structures for all the four pathways are provided in Figure 5 for comparison.


15


Page 16 of 38

Coupled C2-C3 and O-O bond breaking Following on from the gemdiolate intermediate complexes for each of the four hydration pathways we obtained the respective transition states for C2-C3 bond scission and the final reaction products. In all four pathways the C2-C3 bond scission is coupled with breaking of the peroxo bond. In contrast to the carboxylase reactions where the last step in the reaction is the stereospecific protonation of the C2 carbon, the OS2 atom needs to be protonated (non-stereospecific) to form a water molecule for the last step to succeed. In all pathways except for the H+-syn-pathway this proton is sourced from LYS334. In the H+syn-pathway there are two concerted proton transfers, one from GLU204 (which was protonated by the hydrating water molecule) to the Ow atom and another from the Ow atom to the OS2 atom. In the anti-pathway the H2O[Mg] is short of a proton at the completion of the reaction sequence with OH- remaining bound to Mg2+. This causes KCX201 to disengage from Mg-coordination. The activation barriers are 23.0, 25.4, 24.1 and 13.5 kcal/mol, respectively, for syn, anti, H+-syn and H+-anti pathways. In all four scenarios the C2-C3 bond scission is coupled with the breaking of the peroxo bond. This allows the negative charge present on the C2-carbon in the carboxylase pathway to be transferred to the OS2 atom of the peroxide group. The partial negative charge on the OH– group breaking away from the peroxide is stabilised by LYS334 in all scenarios except for the syn-pathway, in which case it is stabilised by LYS177. In the H+-syn-pathway, the C2-C3 distance is unusually short (1.74 Å) in the TS. The geometry for the final products in each of the pathways reflects the configuration in their respective TSs after abstracting a proton from the H+ donor. The reaction energies for the bond scission step are -104.2, -100.7, -36.5 and -50.3 kcal/mol for syn, H+-syn, anti and H+-anti pathways, respectively.


16

Page 17 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Basis Set Superposition and Dispersion Corrections to B3LYP/6-31G(d) Corrected energies are shown in Figure S6 and Table S3 in Supporting Information. The gCP-D3 corrected energy barriers computed at B3LYP/6-31G(d)// B3LYP/6-31G(d,p) are provided in Table 1. The gCPD3 corrections increase activation barriers for hydration in all pathways, but increases activation barrier only for H+-anti for the bond scission step. As a result, after applying the gCP-D3 corrections the energy barriers are higher for hydration, compared with bond scission, to varying extents for all except the syn pathway. The gCP-D3 correction increases the relative energy of the MECP by 3.2 and 0.6 kcal/mol for the FM23 and FM23b models, respectively.

Thus in summary, feeding an additional proton via GLU60 lowers the activation barrier for O2 addition and significantly stabilizes the peroxo product, but has a complex effect on the hydration and bond-scission reactions. It causes a large increase in the energy barrier for hydration in the syn-pathway, but lowers the barrier for both hydration and bond-scission reactions in the anti-pathway. More importantly, for FM23b model, the gemdiolate intermediate is thermodynamically unstable compared with the peroxo product, thereby increasing the energy required for catalysis. Thus, the energy barrier for the combined hydration and bond-scission steps ("hydolysis"), for the FM23b model for both the syn- and antipathways are given by the energy of the bond-scission TS relative to the energy of the gemdiolate intermediate. Whereas, for the FM23 model, which forms a thermodynamically stable gemdiolate intermediate, the largest of the energy barriers for the hydration and bond-scission steps (i.e. hydrolysis) determines the catalytic rate. The gCP-D3 correction has a non-uniform effect on the energetics and, thus, is critical for interpretation of the preferred reaction pathway. In addition to the


17


Page 18 of 38

energetics of the hydrolyis reactions, the energetics relevant to the gas addition step are also provided in Table 1.

Discussion

Oxygen addition. It is postulated in the literature that the enzyme-catalysed reaction between a negatively charged substrate and triplet oxygen can proceed via: (a) a one electron transfer from the substrate anion to 3O2 to form a 3[superoxide radical…substrate radical] pair (i.e. a biradical),36-38 which undergoes spin inversion and recombination to form the singlet peroxide species; (b) direct attack of triplet oxygen to form a triplet intermediate, which undergoes spin inversion to yield the singlet peroxide intermediate;39-40 or (c) enzyme-assisted creation of a substrate radical which reacts with triplet oxygen.41 Recently Hernandez-Ortega et al.39 proposed that oxygenation in 1-H-3-hydroxy-4-oxoquinaldine-2,4-dioxygenase (HOD) proceeds via a direct attack of 3O2 leading to a triplet intermediate. However, in a subsequent computational work Silva20 has shown that oxygenation by HOD proceeds via single-electron transfer from the substrate anion to 3O2 followed by radical recombination. More recently, Wei et al.38 also proposed a similar mechanism (electron transfer to 3O2 followed by radical recombination) for another cofactor-free oxygenase, uricase oxidase. In summary, the majority of examples of oxygenation reactions catalysed by cofactor-free oxygenases in the literature are reported to proceed via substrate-initiated electron transfer to 3O2, i.e. (a).

Selected interatomic distances for optimized structures of the complex of triplet oxygen with the enediolate substrate for different fragment models are provided in Figure S5 in Supporting


18

Page 19 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Information. It is apparent from this comparison that there is spontaneous transfer of an electron from the enediolate if there is sufficient stabilization of the superoxide radical through H bonds, i.e. option (a) above. Spin electron density plots (Figure 2) and the sum of Mulliken atomic charges on O2 in the encounter complexes also confirm the spontaneous electron transfer. Comparison of isolated substrate + O2 complexes (Figure 2B, 2C and 2D) show that the enediolate by itself has greater propensity to transfer the electron to 3O2 (Figure 2D). However, the negative charge on the substrate (in Figure 2D) is not delocalized. Inclusion of the Mg2+ ion as well as other active-site residues in the FM23 model delocalizes the charge on the enediolate and slightly reduces its propensity to donate an electron; this brings it in line with the aromatic substrates in other enzyme examples (Figure 2B and 2C). Addition of another proton in the vicinity of GLU60 in the FM23b model again increases the propensity for formation of the superoxide radical. This is similar to glucose oxidase (GO)-catalysed oxygenation, where a protonated histidine increases the electron affinity of oxygen enabling the electron transfer. Thus, a combination of adequate negative charge on the enediolate substrate and a positively charged moiety near O2, to stabilize the superoxide radical, are necessary for spontaneous transfer of a complete electron from substrate to 3O2. Our recent work has shown that these same stabilisation factors also facilitate the carboxylation pathway reactions.17 Negative charge on the enediolate is essential for nucleophilic attack on CO2, while protonation of the newly formed carboxylate by a positively charged moiety stabilizes the intermediate as well as makes the subsequent hydration thermodynamically feasible.

We were not able to locate a triplet TS for the formation of the C2-O(O2) covalent bond. Our results show that oxygenation by Rubisco proceeds via a MECP (minimum energy crossing


19


Page 20 of 38

point) that lies 10 kcal/mol (13 kcal/mol with gCP-D3 correction; Table 1) above the triplet state minimum in the PES (potential energy surface) (Figure 3A). The barrier for deoxygenation of the peroxo intermediate is 22 kcal/mol (23.5 kcal/mol with gCP-D3 correction). Our earlier unpublished work for a range of smaller fragment models showed that the variations in activation barrier for the gas-addition step due to variation in the charge (proton donor ability) on LYS334 is significantly larger for CO2 addition compared with O2 addition (Figure 4).

Similarly,

although the energy released on formation of the vdW complex of O2 with the active site varies significantly with the charge on LYS334, the change in reaction energy of the O2 addition step itself is relatively small. By contrast, the reaction energy for CO2 addition varies significantly with the charge on LYS334. With the current model (FM23) inclusion of an additional proton through GLU60 (i.e FM23b model) lowers both the barrier and reaction energy for oxygenation by 3.4 kcal/mol and 15.6 kcal/mol, respectively (Figure 3B). Thus, the pKa of LYS334 affects the energy released on transfer of an electron from the enediolate to 3O2 as well as the energy barrier for CO2 addition but has reduced impact on the energy barrier of the O2 addition reaction (Figure 3).

Hydration and C-C bond scission. The four scenarios explored for hydration have activation barriers (with gCP-D3 correction at B3LYP/6-31G(d,p) level in brackets) of 9.5 (10.5), 18.4 (23.3), 19.6 (27.7) and 14.1 (16.5) kcal/mol, and reaction energies of -8.6 (-4.6), 12.4 (17.7), 5.8 (3.2) and 5.9 (8.7) kcal/mol for syn, H+-syn, anti and H+-anti, respectively. The presence of an additional proton in both the syn- and anti-hydration pathways makes the reaction endothermic and increases the activation barrier significantly for syn-hydration. The gCP-D3 correction also makes the anti-hydration endothermic leaving only the syn-pathway with a stable


20

Page 21 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


gemdiolate intermediate. For both syn- and anti-hydration without an additional proton the peroxo group obtains a proton from the water molecule co-ordinated to Mg2+ (H2O[Mg]). In the anti-pathway, with or without an additional proton, KCX201 acts as the base for activating the water molecule, whereas in the H+-syn-pathway GLU204 acts as the base. In contrast to the carboxylase reactions where the primary rate-determining step involves only the breaking of the C2-C3 covalent bond followed by a subsequent C2-protonation, in the oxygenase pathway all the four hydration scenarios explored involve coupled breaking of the C2-C3 and OS1-OS2 bonds in a single step. Interestingly, the overall energy barrier for the synpathway (scission: 23 kcal/mol, 20.7 kcal/mol with gCP-D3 correction) is lower for the oxygenase reactions compared with the largest barrier of 25.5 kcal/mol for syn-hydration in the carboxylase pathway (Figure S7 in Supporting Information). Although in reference to the syn– pathway the energy barrier for the C2-C3 bond-scission step is comparable for the H+-syn (21.6 kcal/mol with gCP-D3 correction) and lower for the H+-anti (14.8 kcal/mol with gCP-D3 correction) pathways, the gemdiolate in both these cases are thermodynamically unstable and, thus, the effective barriers for combined hydration and bond-scission are higher for both pathways (Table 1); hence, the syn-pathway has the lowest overall energy barrier. The oxygenase-pathway reactions are also highly exothermic, to varying extents in the four different hydration pathways, relative to the carboxylase reactions, consistent with the measured reaction enthalpy changes in Rubisco.42

Proposed mechanism of the Rubisco reactions Based on the computational results obtained here for oxygenation and our previous study17 of the carboxylation reaction, in Scheme 2 we propose a holistic mechanistic picture of the Rubisco


21


Page 22 of 38

reactions. Summarizing the carboxylation results,17 we found that protonation of the bound CO2 moiety is absolutely essential for carboxylation, as without it scission of the C2-C3 bond and, hence, formation of product is energetically impossible. While such a proton is conveniently derived from a water molecule hydrating C3 syn to CO2 the resulting gemdiolate is unstable. Gemdiolate stability, however, can be achieved via anti-hydration (with KCX201 as the base that accepts a proton from the hydrating water molecule). However, as there is no obvious proton donor in the active site, this necessitates postulating the operation of an appropriate transport mechanism43-44 to deliver the required proton to bound CO2. Although it is usually assumed that LYS175 is responsible for the stereospecific protonation of C2 in carboxylation, it is also conceivable that the 1-phosphate is protonated on binding the active site and can provide the proton with the assistance of adjacent26 water molecules.

In contrast, oxygenation does not involve stereospecific protonation, and syn hydration at C3 in the peroxo intermediate (Figure 3A) provides the most efficient way to protonate bound O2 and form a stable gemdiolate intermediate, which is then able to promote bond scission of the C2-C3 bond. An extra proton transported into the active site may, thus, not be required for O2 protonation and could then become available to react with the OH- group formed during C2-C3 scission to form water.

Interpretation of Rubisco Kinetic Parameters The results of our computational studies can also assist greatly in the interpretation of in vitro kinetic parameters in terms of rate constants associated with gas binding and subsequent catalytic steps that has recently become a contentious issue in the Rubisco literature.45-46 As neither CO2


22

Page 23 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


nor O2 bind to the enzyme, the observed KC or KO (Michaelis constants) in the steady-state kinetics presumably reflects binding of the gases to the first substrate, the enediolate of RuBP. It has been argued that deoxygenation, i.e. reversal of gas binding, in particular, cannot occur due to the fact that the reaction is spin forbidden.46 However, Figure 3A reveals the existence of two stable intermediates on the oxygenation path. It can be argued that the peroxo form is the Michaelis-Menten-like complex corresponding to the binding component of KO as subsequent hydration involves a third substrate, a water molecule, and further entails substantial chemical changes to the substrate. Thus, the activation barriers of O2 dissociation from the MichaelisMenten-like complex (PKABP) and C2-C3 bond scission (the rate-determining step) in the syn pathway are quite similar (Table 1). Moreover, this interpretation is also supported by the statistical analysis of kinetic parameters which favours significant deoxygenation.45 Considering the gemdiolate to be the Michaelis-Menten complex precludes the possibility of deoxygenation, which is not consistent with the analysis of kinetic data. Although it has been suggested that carboxylation may be concerted with hydration34, 46 or hydration may stabilize the carboxylate to form a gemdiolate Michaelis complex,17 our results for the carboxylation reaction using the FM23 model (Figure S7 in Supporting Information) suggest the formation of a stable carboxylate Michaelis complex, CKABP (i.e. prior to hydration), with a decarboxylation barrier of about 18 kcal/mol which is quite close to the experimental carboxylation activation barrier of 16-17 kcal/mol.47 Thus, the similarity in the respective gas-dissociation and catalytic barriers found in the computational models used here, allow for the distinct possibility that both deoxygenation and decarboxylation are important factors in determining the kinetic behaviour of Rubisco.45

Conclusion


23


Page 24 of 38

Rubisco appears to be a unique case with two factors unlocking its potential as an oxygenase enzyme. First, it needs to stabilize its anionic RuBP-enediolate substrate (first reaction intermediate) very strongly to minimize side reactions while at the same time maintaining its nucleophilicity for subsequent reaction. Second, the productive pathway requires the presence of a protonated LYS (residue 334) to stabilize its second reaction intermediate, the carboxylated product, CKABP. We have shown that both factors – the strongly stabilized anion and the protonated LYS334 – set Rubisco up to facilitate the barrier-less activation of 3O2. As discussed, modifying the positive charge on LYS334 would reduce the rate of the CO2-addition reaction but affect only the lifetime of the ES∙O2 complex and moderately impact the O2-addition reaction. Thus, Rubisco is in a fix with 3O2: it has acquired an unwanted capability that hampers its primary biological function, and there is no obvious way to get rid of it without further damaging the function that Rubisco has evolved to do. However, creative Rubisco re-engineering may offer a solution.

Incorporating findings from our earlier work on carboxylase-pathway reactions we propose a detailed mechanistic picture of both carboxylation and oxygenation reactions. The principal mechanistic difference between the two reactions catalysed by Rubisco is that hydration by a water molecule approaching C3 syn to the C2-peroxo group is likely the preferred pathway for oxygenation, whereas hydration anti to the C2-carboxyl group is the more feasible for carboxylation. It is proposed this mechanistic variation determines the different fates of a proton presumed transported into the active site. Validation of the proposed mechanism (Scheme 2) requires further investigation into the options and feasibility of proton transport into the reaction


24

Page 25 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


sphere, in addition to resolving the source of the proton for stereospecific protonation in the carboxylation reaction.

Supporting Information Available Additional details about fragment models, geometry parameters of optimized structures, energetics from previous unpublished results for carboxylase pathway and energetics with gCPD3 corrections.

Acknowledgment: This research was undertaken with the assistance of resources and services from the National Computational Infrastructure (NCI), which is supported by the Australian Government. We thank the Reviewer for helpful comments.


25


Page 26 of 38

References 1. Wildman, S. G., Along the Trail from Fraction I Protein to Rubisco (Ribulose Bisphosphate Carboxylase-Oxygenase). Photosynth. Res. 2002, 73 (1), 243-250. 2. Andersson, I., Catalysis and Regulation in Rubisco. J. Exp. Bot. 2008, 59 (7), 1555-1568. 3. Raven, J. A., Rubisco: Still the Most Abundant Protein of Earth? New Phytol. 2013, 198 (1), 1-3. 4. Long, S. P.; Zhu, X.-G.; Naidu, S. L.; Ort, D. R., Can Improvement in Photosynthesis Increase Crop Yields? Plant, Cell Environ. 2006, 29 (3), 315-330. 5. Mann, C. C., Genetic Engineers Aim to Soup Up Crop Photosynthesis. Science 1999, 283 (5400), 314-316. 6. J. E. Gready, B. Kannappan, (2009-2017) Process for Generation of Protein and Uses Thereof. US Patent 12/422,190. [filed 2009; granted 2017]. 7. Sage, R. F., The Evolution of C4 photosynthesis. New Phytol. 2004, 161 (2), 341-370. 8. Kannappan, B.; Gready, J. E., Redefinition of Rubisco Carboxylase Reaction Reveals Origin of Water for Hydration and New Roles for Active-site Residues. J. Am. Chem. Soc. 2008, 130 (45), 15063-15080. 9. Pearce, F. G., Catalytic By-product Formation and Ligand Binding by Ribulose Bisphosphate Carboxylases from Different Phylogenies. Biochem. J. 2006, 399 (3), 525-534. 10. Gutteridge, S.; Pierce, J., A Unified Theory for the Basis of the Limitations of the Primary Reaction of Photosynthetic CO2 Fixation: Was Dr. Pangloss Right? Proc. Natl. Acad. Sci. USA 2006, 103 (19), 7203-7204. 11. Bar-Even, A.; Noor, E.; Savir, Y.; Liebermeister, W.; Davidi, D.; Tawfik, D. S.; Milo, R., The Moderately Efficient Enzyme: Evolutionary and Physicochemical Trends Shaping Enzyme Parameters. Biochemistry 2011, 50 (21), 4402-4410. 12. Bathellier, C.; Tcherkez, G.; Lorimer, G. H.; Farquhar, G. D., Rubisco is Not Really so Bad. Plant, Cell Environ. 2018, 41 (4), 705-716. 13. Brunk, E.; Rothlisberger, U., Mixed Quantum Mechanical/Molecular Mechanical Molecular Dynamics Simulations of Biological Systems in Ground and Electronically Excited States. Chem. Rev. 2015, 115 (12), 6217-6263. 14. Sousa, S. F.; Ribeiro, A. J. M.; Neves, R. P. P.; Brás, N. F.; Cerqueira, N. M. F. S. A.; Fernandes, P. A.; Ramos, M. J., Application of Quantum Mechanics/Molecular Mechanics Methods in the Study of Enzymatic Reaction Mechanisms. WIREs Comput Mol Sci 2017, 7 (2), e1281. 15. Mauser, H.; King, W. A.; Gready, J. E.; Andrews, T. J., CO2 Fixation by Rubisco: Computational Dissection of the Key Steps of Carboxylation, Hydration, and C−C Bond Cleavage. J. Am. Chem. Soc. 2001, 123 (44), 10821-10829. 16. King, W. A.; Gready, J. E.; Andrews, T. J., Quantum Chemical Analysis of the Enolization of Ribulose Bisphosphate: The First Hurdle in the Fixation of CO2 by Rubisco. Biochemistry 1998, 37 (44), 15414-15422. 17. Cummins, P. L.; Kannappan, B.; Gready, J. E., Revised Mechanism of Carboxylation of Ribulose-1,5-biphosphate by Rubisco from Large Scale Quantum Chemical Calculations. J. Comput. Chem. 2018, 39 (21), 1656-1665. 18. Klinman, J. P., How Do Enzymes Activate Oxygen without Inactivating Themselves? Acc. Chem. Res. 2007, 40 (5), 325-333.


26

Page 27 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


19. Bugg, T. D. H., Dioxygenase Enzymes: Catalytic Mechanisms and Chemical Models. Tetrahedron 2003, 59 (36), 7075-7101. 20. Silva, P. J., Refining the Reaction Mechanism of O-2 Towards its Co-substrate in Cofactor-free Dioxygenases. PeerJ 2016, 4:e2805. 21. Marian, C. M., Spin‐Orbit Coupling in Molecules. In Rev. Comput. Chem., Lipkowitz, K. B.; Boyd, D. B., Eds. Wiley-VCH: Weinheim, 2001; Vol. 17, pp 99– 204. 22. Fetzner, S., Oxygenases Without Requirement for Cofactors or Metal Ions. Appl. Microbiol. Biotechnol. 2002, 60 (3), 243-257. 23. Fetzner, S.; Steiner, R. A., Cofactor-independent Oxidases and Oxygenases. Appl. Microbiol. Biotechnol. 2010, 86 (3), 791-804. 24. Romero, E.; Gómez Castellanos, J. R.; Gadda, G.; Fraaije, M. W.; Mattevi, A., Same Substrate, Many Reactions: Oxygen Activation in Flavoenzymes. Chem. Rev. 2018, 118 (4), 1742-1769. 25. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Petersson, G. A.; Nakatsuji, H., et al. Gaussian 16 Rev. B.01, Gaussian, Inc.: Wallingford, CT, 2016. 26. Andersson, I., Large Structures at High Resolution: The 1.6 Å Crystal Structure of Spinach Ribulose-1,5- Bisphosphate Carboxylase/Oxygenase Complexed with 2Carboxyarabinitol Bisphosphate. J. Mol. Biol. 1996, 259 (1), 160-174. 27. Valiev, M.; Bylaska, E. J.; Govind, N.; Kowalski, K.; Straatsma, T. P.; Van Dam, H. J. J.; Wang, D.; Nieplocha, J.; Apra, E.; Windus, T. L., et al., NWChem: A Comprehensive and Scalable Open-source Solution for Large Scale Molecular Simulations. Comput. Phys. Commun. 2010, 181 (9), 1477-1489. 28. Li, X.; Frisch, M. J., Energy-Represented Direct Inversion in the Iterative Subspace within a Hybrid Geometry Optimization Method. J. Chem. Theory Comput. 2006, 2 (3), 835-839. 29. Goerigk, L.; Grimme, S., A Thorough Benchmark of Density Functional Methods for General Main Group Thermochemistry, Kinetics, and Noncovalent Interactions. Phys. Chem. Chem. Phys. 2011, 13 (14), 6670-6688. 30. Zhao, Y.; Truhlar, D. G., Density Functionals with Broad Applicability in Chemistry. Acc. Chem. Res. 2008, 41 (2), 157-167. 31. Kruse, H.; Goerigk, L.; Grimme, S., Why the Standard B3LYP/6-31G*Model Chemistry Should Not Be Used in DFT Calculations of Molecular Thermochemistry: Understanding and Correcting the Problem. J. Org. Chem. 2012, 77 (23), 10824-10834. 32. Prabhakar, R.; Siegbahn, P. E. M.; Minaev, B. F., A Theoretical Study of the Dioxygen Activation by Glucose Oxidase and Copper Amine Oxidase. Biochim. Biophys. Acta 2003, 1647 (1-2), 173-178. 33. Prabhakar, R.; Siegbahn, P. E. M.; Minaev, B. F.; Agren, H., Activation of Triplet Dioxygen by Glucose Oxidase: Spin-orbit Coupling in the Superoxide Ion. J. Phys. Chem. B 2002, 106 (14), 3742-3750. 34. Cleland, W. W.; Andrews, T. J.; Gutteridge, S.; Hartman, F. C.; Lorimer, G. H., Mechanism of Rubisco: The Carbamate as General Base. Chem. Rev. 1998, 98 (2), 549-562. 35. Harvey, J. N.; Aschi, M.; Schwarz, H.; Koch, W., The Singlet and Triplet States of Phenyl Cation. A Hybrid Approach for Locating Minimum Energy Crossing Points Between Non-interacting Potential Energy Surfaces. Theor. Chem. Acc. 1998, 99 (2), 95-99.


27


Page 28 of 38

36. Thierbach, S.; Bui, N.; Zapp, J.; Chhabra, S. R.; Kappl, R.; Fetzner, S., SubstrateAssisted O-2 Activation in a Cofactor-Independent Dioxygenase. Chem. Biol. 2014, 21 (2), 217225. 37. Machovina, M. M.; Usselman, R. J.; DuBois, J. L., Monooxygenase Substrates Mimic Flavin to Catalyze Cofactorless Oxygenations. J. Biol. Chem. 2016, 291 (34), 17816-17828. 38. Wei, D. H.; Huang, X. Q.; Qiao, Y.; Rao, J. J.; Wang, L.; Liao, F.; Zhan, C. G., Catalytic Mechanisms for Cofactor-Free Oxidase-Catalyzed Reactions: Reaction Pathways of UricaseCatalyzed Oxidation and Hydration of Uric Acid. ACS Catal. 2017, 7 (7), 4623-4636. 39. Hernandez-Ortega, A.; Quesne, M. G.; Bui, S.; Heyes, D. J.; Steiner, R. A.; Scrutton, N. S.; de Visser, S. P., Catalytic Mechanism of Cofactor-Free Dioxygenases and How They Circumvent Spin-Forbidden Oxygenation of Their Substrates. J. Am. Chem. Soc. 2015, 137 (23), 7474-7487. 40. Bui, S.; Steiner, R. A., New insight into Cofactor-free Oxygenation from Combined Experimental and Computational Approaches. Curr. Opin. Struct. Biol. 2016, 41, 109-118. 41. Gabison, L.; Chopard, C.; Colloc'h, N.; Peyrot, F.; Castro, B.; El Hajji, M.; Altarsha, M.; Monard, G.; Chiadmi, M.; Prange, T., X-ray, ESR, and Quantum Mechanics Studies Unravel a Spin Well in the Cofactor-less Urate Oxidase. Proteins: Struct. Funct. Bioinform. 2011, 79 (6), 1964-1976. 42. Frank, J.; Kositza, M. J.; Vater, J.; Holzwarth, J. F., Microcalorimetric Determination of theRreaction Enthalpy Changes Associated with the Carboxylase and Oxygenase Reactions Catalysed by Ribulose 1,5-bisphosphate Carboxylase/Oxygenase (RUBISCO). Phys. Chem. Chem. Phys. 2000, 2 (6), 1301-1304. 43. Bekçioğlu, G.; Allolio, C.; Sebastiani, D., Water Wires in Aqueous Solutions from FirstPrinciples Calculations. J. Phys. Chem. B 2015, 119 (10), 4053-4060. 44. Kaila, V. R. I.; Hummer, G., Energetics and Dynamics of Proton Transfer Reactions Along Short Water Wires. Phys. Chem. Chem. Phys. 2011, 13 (29), 13207-13215. 45. Cummins, P. L.; Kannappan, B.; Gready, J. E., Directions for Optimization of Photosynthetic Carbon Fixation: RuBisCO's Efficiency May Not Be So Constrained After All. Front. Plant Sci. 2018, 9:183. 46. Tcherkez, G. G.; Bathellier, C.; Farquhar, G. D.; Lorimer, G. H., Commentary: Directions for Optimization of Photosynthetic Carbon Fixation: RuBisCO's Efficiency May Not Be So Constrained After All. Front. Plant Sci. 2018, 9:929. 47. Tcherkez, G. G. B.; Bathellier, C.; Stuart-Williams, H.; Whitney, S.; Gout, E.; Bligny, R.; Badger, M.; Farquhar, G. D., D2O Solvent Isotope Effects Suggest Uniform Energy Barriers in Ribulose-1,5-bisphosphate Carboxylase/Oxygenase Catalysis. Biochemistry 2013, 52 (5), 869877.


28

Page 29 of 38

Release

[2 x 3PGA]

of products

Activated Rubisco

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


C2protonation

C2-C3 bond scission

Hydration

[CKABP] CO2 addition

RuBP binding

[RuBP]

Enolization

[Enediolate] O2 addition

Release of products

[3PGA + 2PG]

C2-C3 bond scission

Hydration

Tautomerization

[XuBP]

Phosphate elimination

[DPDP] [CTBP]

[PKABP]

peroxide [PDBP] elimination

Scheme 1. Sequence of reactions catalyzed by Rubisco. The reaction starts after the binding of RuBP to activated Rubisco. In the productive pathway Rubisco catalyses a sequence of 5 reactions shown by blue arrows. Apart from unwanted oxygenase activity, reactions at the Rubisco active site can also lead to many side products (shown in red), many of which are tightbinding inhibitors that block the Rubisco active site from being reused until removed by the chaperone, Rubisco activase. On completion of the carboxylase and oxygenase reactions the products are released from the active site returning the enzyme to the activated state ready to bind a new RuBP molecule. Abbreviations: RuBP, D-ribulose 1,5-bisphosphate; CKABP, 2carboxy-3-keto-D-arabinitol 1,5-; PKABP, 2-peroxo-3-keto-D-arabinitol 1,5-bisphosphate; 3PGA, 3-phospho-D-glycerate; 2PG, 2-phosphoglycolate; XuBP, xylulose 1,5-bisphosphate; DPDP, 1-deoxy-D-glycero-2,3-pentodiulose 5-phosphate; PDBP, D-glycero-2,3-pentodiulose 1,5-bisphosphate; CTBP, 2-carboxytetritol 1,4-bisphosphate.


29


?

?

PGA

HO

+

HOOC

OH

HO HOOC

H

OPO3

OPO32-

Carbanion

+ OH

PGA

OPO32-

HOOC HO

OH -

O HO

PGA

OPO32-

-

OOC HO

OH O

OPO3

O

H

H 2O

OPO32-

H+

2-

Hydration (anti)

C2-protonation

HOOC

+

C2-C3 bond scission

COOH OPO 23

H

2-

H

OPO32-

KCX201 -2 x PGA

OPO32-

OPO32O

O

+

RuBP

OH

H OPO32-

OPO32-

H294

2-

?

O2 H+ OH

O

OPO32-

2H OPO3

PG

+ OH

PGA OPO3

2-

O

C2-C3 bond scission

HOOC HOOC

OH

H

-(PGA + PG)

H

HO

O

KCX201 OPO3

OPO32-

OH HO

HO

OH -

-HOH

O

OPO32-

Hydration (syn)

HO

Enolization

CO2

Rubisco

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 30 of 38

OO-

H

HO

O

H

OH O

HO

H 2O OPO32-

OPO32-

Scheme 2. Catalytic mechanisms of Rubisco carboxylase and oxygenase reactions. Atoms of the gaseous substrates CO2 and O2 are shown in red, while that of the hydrating water molecule is shown in blue. In the case of carboxylase where the hydration takes place anti to the C2carboxylate two additional protons are required to complete the reaction sequence successfully, whereas in the oxygenation pathway one proton is derived from the syn-hydration water molecule and an additional proton is required to complete the reaction sequence.


30

Page 31 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


A

R N

NH

N H

B

R N

O + O 2

N

O N H

OH

HO

+ O2 S CoA

O

O

R + O2

OH

+ O2

HO

OH

O HO

S CoA H O O

OH

O

CH2OPO3

+ O2

CH2OPO3

2

O COOH

COOH

HO O

O3

O

+H

O H

2

O O R

OH

S CoA

H

HO2

N H

O

COOH

E

O

H

+ O2

+ H2O2

O

N H

H

OH O

NH

N

O

R

O

N

O + O2

+ O2

O

D

+H

O

O

C

R N

NH

N H

O

O

N

+ O2

CH2OPO3

H O OH

2

OH CH2OPO32

+H

+ O2 CH2OPO3

HO

O

O 2

CH2OPO32

Figure 1. Examples of enzyme-catalyzed reactions involving activation of 3O2 by: (A) a flavin-derived cofactor in glucose oxidase,33 (B) substrate in 1H-3-hydroxy-4oxoquinaldine 2,4-dioxygenase,39 (C) substrate in DpgC, a cofactor-less dioxygenase,23 (D) substrate in ActVA-Orf6 monooxygenase23 and (E) the enediolate substrate in


31


Page 32 of 38

Rubisco. Relevant atoms from the reaction centre and O2 are shown in red. In examples (B) to (E) an adjacent base from the enzyme abstracts a proton to produce the activator molecule shown at the beginning of each reaction.

A

B

C

a = -1, b = -0.24, c = 1.73

D

a = -1, b = -0.14, c = 1.86

a = -1, b = -0.48, c = 1.46

a = -1, b = -0.17, c = 1.76

Figure 2. Electron spin density plot (isosurface of value 0.001 a.u.) of the vdW complex of O2 with (A) the FM23 model of the enzyme-bound RuBP-enediolate substrate in comparison with (B) the flavin-derived cofactor in glucose oxidase,33 (C) the substrate of DpgC, a cofactor-less dioxygenase,23 and (D) a four-carbon model for the enediolate. The lobes at the top encompass the O2 molecule. a, Net charge on the complex; b, Mulliken charge on O2; c, Sum of atomic spin densities on O2 (a.u.).


32

Page 33 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60



33


Figure 3. B3LYP/6-31G(d,p) Potential energy surface of the oxygenase reaction pathway starting from the enediolate substrate and comparing syn- and anti-hydration for models FM23 (A) and FM23b (B). In the FM23b model there is an additional proton provided via GLU60 (see text for details). The energies for the rate-limiting step (bond scission) shown on the RHS are referenced to the most stable previous intermediate, i.e gemdiolate for A and peroxo product for B.

A

10 0

-10

-60 -70 -80 -90 -100

FM16 (CO2) FM18 (O2) FM17 (O2) FM20 (O2) FM16(O2)

gas adduct

-50

FM20 (CO2)

TS

-40

FM17(CO2)

ES.CO2/O2

-30

FM18 (CO2)

ES + CO2/O2

-20

relative energy in kcal/mol

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 34 of 38

-110 -120

reaction coordinate


34

Page 35 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4 (A) Reaction energy profiles (in kcal/mol) for the CO2 (blue) and O2 (red) addition reactions at the Rubisco active site computed using fragment models FM18, FM17, FM20 and FM16 (models in order of increasing proton donor ability at LYS334). The first species on the reaction coordinate is the isolated Enzyme∙Substrate (ES) complex and CO2 or O2 molecule, which is followed by the reaction encounter complex ES∙CO2 or ES∙O2 (species 2), the TS or MECP (species 3) and the gas adduct (species 4). The inset shows the energy profiles between species 2 - 4 on the reaction coordinate using ES∙CO2/O2 as the reference. For FM16 (CO2) the TS was not located and an arbitrary number has been used in the plot. (B) Comparison of the potential energy surfaces (PES) for reactions on the carboxylase (blue) and oxygenase (red) pathways computed using fragment model FM20 at the B3LYP/6-31G(d,p) level.8 The starting point on the reaction coordinate is the isolated CO2 or O2 and ES, which is followed by a vdW complex between these two components. Numbers in brackets show the activation energy Ea (in kcal/mol) for the forward and reverse reactions, respectively. T0 and S0 are the triplet and singlet electronic ground-state surfaces, respectively. MECP is the minimum energy crossing point between the T0 and S0 surfaces.


35


Page 36 of 38

Figure 5. Optimized structures of the transition states for hydration for the (A) syn, (B) H+-syn, (C) anti and (D) H+-anti pathways. The oxygen atom of water molecules used for hydration are labelled in red, for anti-pathways the hydrating water molecule is a solvent water molecule


36

Page 37 of 38 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


shown added near KCX201 (see text for details). For H+-syn and H+-anti pathways the additional proton near GLU60 is highlighted by a red box.

Table 1. Activation barriers for oxygenation and hydrolysis. Oxygenation Hydrolysisa (O2 binding/dissociation) (hydration and scission) Method on off syn† anti† on off syn anti (H+) (H+) (H+) (H+) B3LYP/6-31G(d,p) 10.0 6.6 22.0 37.6 23.0 36.5 25.4 19.4 9.8 6.7 21.5 37.9 23.4 36.7 24.8 20.4 B3LYP/6-31G(d) 13.0 7.3 23.5 36.3 20.7 39.2 27.7 23.5 B3LYP-gCP-D3/6-31G(d) a For the model without an additional proton (FM23) the hydrolysis barrier corresponds to the energy barrier for the C2-C3 bond scission step whereas for the protonated model (FM23b) the hydrolysis barrier is the difference in energy between the TS for the C2-C3 bond scission and the peroxo product as the hydration step itself is endothermic; see Figure 3. Further discussion of the definition of "hydrolysis"

is given in the text.


37


Page 38 of 38

TOC Graphic


38

Mechanism of Oxygenase Pathway Reactions Catalysed by Rubisco

Recommend Documents