Resolving the structure of the E1 state of Mo nitrogenase through Mo and Fe K-edge EXAFS and QM/MM calculations

The FeMoco cluster of Mo nitrogenase undergoes minor distortions upon reduction to E1, supporting iron-based reduction and belt sulfide protonation.


S1. EPR of the resting and low-flux turnover states of MoFe
EPR measurements were performed to establish the relative degree of reduction of the S = 3 /2 signal of the E0 resting state of MoFe N2ase in turnover samples.
These measurements were performed on the exact samples which were used for further XAS measurements. All continuous wave (CW) X-band EPR measurements were recorded using a Bruker E500 ELEXSYS spectrometer equipped with a Bruker dual-mode cavity (ER4116DM) and an Oxford Instruments helium flow cryostat (ESR 900). All measurements were performed at 10 K. A high-sensitivity Bruker Super-X (ER-049X) bridge with integrated microwave frequency counter was employed as the microwave unit. A magnetic field controller (ER032T) was externally calibrated with a Bruker NMR field probe (ER035M). X-band measurements utilized a 2 mW microwave power and 0.746 mT/100 kHz modulation. Spectra were analyzed in both the local software package esim_gfit (available from Eckhard Bill) and the software package EasySpin (version 5.1.9) as implemented in Matlab. 1 An average of 50% reduction in the S = 3 /2 E0 signal was found for turnover samples, based upon both relative intensity of the g1 feature at g = 4.3 and spin-integration area ( Figure S1-1).
Figure S1-1. X-band EPR spectra of the resting (E0, black) and turnover (E0+E1, red) states. The large radical signal ~ g = 2 arises from the sample holder itself and is unaffiliated with the MoFe sample.

S2. Acquiring and adjusting degeneracy "N"
In EXAFS fitting, N represents the degeneracy of a given scattering path. This can be interpreted as the average "coordination number", or how many of a given scatterer surround the absorbing atom. In a more technical sense, it is the number of identical distinct ways per absorbing atom that the scattering defined by a given path can occur.   We can further examine whether these Fe-Fe paths need to be further differentiated. Looking to Figure S2

S3. The feasibility of fitting the Fe-C scatterer
It is now well established that the FeMoco cluster contains a light atom at its core which has been identified as a carbide. [3][4][5][6] The knowledge of its existence naturally has implications for any structural modelling of the FeMoco. However, considering C is a very light atom (and thus a weak scatterer) which is coordinated to only 6 of 15 unique irons (giving a degeneracy of N = 0.4), is inclusion of the Fe-C scattering path really required?

S4. Why is E 0 important?
The parameter E0 (not be confused with the enzyme resting state E0) is often referred to as the edge energy, reference energy, or even as the inner potential of the absorber of interest. In a formal sense, this parameter functions as the energy which can eject a photoelectron with zero final momentum from a given electronic shell (in the case of K-edge XAS, the 1s). Since numerous electronic processes can occur at the K-edge, including shake-up and shake-down transitions, this parameter becomes difficult to precisely determine. Therefore, it is usually necessary to fit E0 when modeling EXAFS. E0 functions to define at what point the EXAFS region begins, as can be seen from equation 1: where ℏ is the reduced Planck's constant, E is a given incident energy, and : is the mass of an electron. Since E0 determines the beginning of the EXAFS region, it should also be intuitive that the value of E0 lies in neither the pre-edge region or EXAFS regions; instead, it should correspond to some point along the edge, and possibly up to the white-line region.
In the case of a single absorbing atom, there can only be one E0 for all scattering paths. This is because E0 is defined as a characteristic of the absorber itself, and is independent of the scattering path. If a different E0 is provided for each scattering path, a change in phase of these paths relative to one another will occur. While doing this may provide fits that appear statistically reasonable, the determined distances are unphysical and can carry major errors. This is demonstrated in Figure S4-1/     The pre-edge feature arises from the 1s ® 3d transition, which, while formally dipole forbidden, gains intensity through Td-symmetry allowed 3d-4p mixing. [7][8][9] Upon reduction, decreases in intensity at the pre-edge and rising edge are followed by an increased white-line intensity.          both the MS = 1 and MS = 2 broken symmetry state solutions were initially probed ( Figure S6.1-1).

S6.2. Determination of the homocitrate protonation state
The protonation state of the hydroxyl group of homocitrate has direct implications on the structural environment of Mo in the FeMoco cluster. Previous studies have supported that this group is protonated in the E0 state, and is held hydrogen-bonded to one of the uncoordinated carboxylate groups of the homocitrate ( Figure S6.2-1). 10,16 However, it is unknown whether this remains true for any of the En states (n > 0), and necessitates further investigation in the case of E1.

S6.3.1 Protonated cluster models of E1 in a polarizable continuum
To test the influence of the surrounding secondary environment on the energy of belt-sulfide protonated models of E1, these models were calculated using a simple minimal cluster model of FeMoco with a surrounding polarizable continuum (ϵ = 4) described by CPCM. 17 In these models the MM region and all secondary amino acid residues surrounding FeMoco are absent (homocitrate and the sidechains of His442 and Cys275 included). The resulting relative energies for these models, as well as the Fe-hydride model, is shown in Figure S6.5-1 for the BS-346 electronic state. The relative energies show that removal of the secondary coordination environment makes the bridging sulfides almost equally favorable as protonation sites.

S6.3.2 Direction of protonation in QM/MM models of E1
It has been previously found that the direction of protonation of the FeMoco cluster embedded in a QM/MM environment can result in different relative energies.
The orientations of protonation selected for each of the models in the present study was partially based on previous studies investigating the favorability of these different orientations for protonation. 18 The protonation of the S5A position was performed such           Unweighted root-mean-square deviations for the calculated changes in distances (DR calc ) of the computational models of the E1 state. Calculation of the unweighted RMSDs is described in the Statistical Analysis section of the Experimental details. Experimental changes in distances (DR exp ) used in these calculations were acquired from E1-E0.