Solution-State Inter-Copper Distribution of Redox Partner-Linked Copper Nitrite Reductases: A Pulsed Electron–Electron Double Resonance Spectroscopy Study

Copper nitrite reductases (CuNiRs) catalyze the reduction of nitrite to form nitric oxide. In recent years, new classes of redox partner linked CuNiRs have been isolated and characterized by crystallographic techniques. Solution-state biophysical studies have shed light on the complex catalytic mechanisms of these enzymes and implied that protein dynamics may play a role in CuNiR catalysis. To investigate the structural, dynamical, and functional relationship of these CuNiRs, we have used protein reverse engineering and pulsed electron–electron double resonance (PELDOR) spectroscopy to determine their solution-state inter-copper distributions. Data show the multidimensional conformational landscape of this family of enzymes and the role of tethering in catalysis. The importance of combining high-resolution crystallographic techniques and low-resolution solution-state approaches in determining the structures and mechanisms of metalloenzymes is emphasized by our approach.


S4
The RpNiR 'cytochrome c' proteins were purified by nickel affinity chromatography using similar methods mentioned above for the RpNiR protein. A copper-loading step was omitted from this purification. Likewise, the N-terminal His-tags were removed from the protein using TEV protease. The RpNiR 'cytochrome c' protein was shown to run as a monomer on gel filtration column, as previously shown. [2] The RpNiR 'cytochrome c' was stored at -80 o C.

Hyphomicrobium denitrificans copper nitrite reductase 'core' expression and purification
The codon optimised recombinant 'core' region of Hyphomicrobium denitrificans copper nitrite reductase (HdNiR 'core') was cloned in pETM-11 between the NcoI and the XhoI sites. N-terminal Histagged HdNiR 'core' was expressed and purified using published protocols for the RpNiR 'core' protein.
Briefly, the pETM-11 plasmid containing the gene for the HdNiR 'core' were transformed into BL21(DE3) E. coli cells and grown in 0.5 L of Terrific Broth media. Protein synthesis was induced when an OD 600 of 0.6 was reached by the addition of 0.3 mM IPTG. 0.1 mM CuSO 4 was added to cultures after reaching the stationary phase to increase the copper loading of the HdNiR 'core' protein. Cells were grown for >12 hours and harvested using centrifugation. Cell pellets were stored at -20 o C before purification.
The HdNiR 'core' proteins were purified by nickel affinity chromatography using similar methods mentioned above for the RpNiR 'core' protein. Likewise, the N-terminal His-tags were removed from the protein using TEV protease. The HdNiR 'core' protein was shown to run as a trimer on a gel filtration column. The HdNiR 'core' was stored at -80 o C.

Tobacco etch virus (TEV) protease expression and purification
The N-terminal His-tag tobacco etch virus (TEV) protease was expressed and purified by following previously published protocols. [5] Electron paramagnetic resonance (EPR) Electron paramagnetic resonance (EPR) measurements were carried out using a Bruker ELEXSYS-500/580 X-band EPR spectrometer operating in both continuous-wave (CW) and pulsed modes, equipped with an Oxford variable-temperature unit and ESR900 cryostat with Super High-Q resonator. EPR measurements were performed on approximately 900 μM of monomeric NiR (300 μM of trimeric NiR/150 μM of hexameric NiR) dissolved in 50 mM potassium phosphate buffers supplemented with 10 % glycerol. To increase the pulsed electron-electron double resonance (PELDOR) measurement window (time-domain, vide infra), both the buffer and the glycerol used for these measurements were deuterated. EPR samples were placed in 4 mm quartz capillary tubes (Wilmad-LabGlass) and frozen in liquid nitrogen. Samples were stored in liquid nitrogen until the measurements were conducted on them.

S5
The X-band EPR tubes were then transferred into the EPR probehead, which was pre-cooled to 20 K. The low-temperature cw-EPR spectra were measured at 20 K. A microwave power of 30 dB (0.2 mW) and modulation of 5 G appear to be optimal for recording the EPR spectra of full-length (3-domain) and 'core' regions of the RpNiR and HdNiR proteins. The low temperature EPR spectra were acquired using the following conditions: sweep time of 84 s, microwave power of 0.2 mW, time constant of 41 ms, average microwave frequency of 9.386 GHz and modulation amplitude of 5 G.
X-band pulsed-EPR and PELDOR measurements were performed at 10 K, using the pulsedmode with a dielectric resonator (ER4118X-MD-5) and a closed cryostat system (Bruker). Echodetected field-swept spectra were measured using the following pulse sequence: t p −τ−2t p −τ−echo with t p (/2) = 16 ns, τ = 0.2 μs, pulse repetition rate of 5 ms. The PELDOR experiments were performed using the pulse sequences, as indicated in Scheme S1: Scheme S1. Pulse sequences for PELDOR experiments carried at 10 K. The positions of the primary (PE) and refocused (RE) echoes are marked in the pulse sequences.
The π/2 and π pulses of the detection sequence had lengths of 16 and 32 ns, and the pump pulse was 14 ns long. The π/2 pulse was phase-cycled to eliminate the receiver offsets. The ∆ν (ν det -ν pump ) was  100 MHz. The pulse separations, τ 1 and τ 3 , were 200 and 100 ns, respectively, and the echo signal was integrated using a video amplifier bandwidth of 20 MHz. The τ 2 for the various constructs were 2.6 μs for the full-length HdNiR and the HdNiR 'core' region, 1.6 μs and 2.2 μs for the full-length RpNiR, and 2.8 μs for the RpNiR 'core'. The pump pulse was stepped out by 4 ns for a given τ 2 in timedomain axis. The PELDOR data were analysed using DeerAnalysis2022 [6] and DEERNet Spinach SVN Rev 5662. [7,8] Analysis of the cw-EPR spectra was performed using the EasySpin toolbox (5.2.35) for the Matlab program package. [9] The raw-PELDOR data and validation of distance distribution for various constructs of RpNiR and HdNiR proteins are given below (Figures S9-S21). When conducting PELDOR analysis, background deconvolution is not exact and can introduce error. This error can often dominate the error in the resulting distance distributions. To overcome this, the background starting time (500-1500 ns), background dimensionalities (between 2-3) and the white noise (0.003-1.5) were varied systematically within a limit. The resultant 'n' form factors obtained were subjected to Tikhonov regularisation and the distance distributions obtained (lower/upper limit and mean value) are plotted in

S6
Figures S9-S21. The PELDOR data displayed in the main text (HdNiR full-length enzyme was analysed using DEERNet in the Spinach 2.5.5449 distribution) 7 were also analysed using DEERNet (neural network), where the errors due to the user-defined parameters are no longer present. It is also noteworthy to bring it to the attention of the readers that the PELDOR time traces for the "core" and "full-length" proteins (RpNiR and Hd-NiR) were stretched as long as possible (Cu-Cu distances range from 3.0 -6.1 nm for both Rp-NiR and Hd-NiR proteins) in an effort to minimise the error in the background deconvolutions, which often introduce error in PELDOR analysis.

PELDOR orientation selectivity
The dipolar coupling frequency of two interacting electron spins, S A and S B is described in Equation S1 : Eq. S1 where  0 is the vacuum permeability,  B is the Bohr magneton, g A and g B are the g-values of the interacting electron spins (S A and S B ), is the reduced Planck constant, r AB is the distance between the ℏ two spins, and  AB is the angle between the external magnetic field and the inter-spin vector (r AB ). The dipolar coupling frequency is dependent on both r AB and  AB . In the case of flexible nitroxides, the relative orientations of the interacting spins are often randomized. Therefore, it is assumed that the microwave pulses in the PELDOR experiments uniformly excite most of the  AB at X-band. A well-known frequency distribution, called Pake pattern is produced from these experiments. [10] However, due to the large g-anisotropy of both type I (T1Cu) and type II (T2Cu) Cu II centres present in RpNiR and HdNiR proteins, the Cu II -Cu II PELDOR experiments performed in this study fall under the category of 'orientation selection'. This leads to selective excitation of certain g-tensor components of the interacting Cu II centres and a breakdown of the approximation of a Pake pattern as a frequency response. [10] Therefore, the Tikhonov method can no longer be applied in these investigations due to selective excitation of  AB by the applied microwave pulses. [6,11] In this case, it is often required to collect the PELDOR data at multiple magnetic fields across the EPR envelope to reliably extract the distance distribution between the interacting spins. As reported previously, to overcome this 'orientation selectivity', PELDOR data were collected in our investigation for the RpNiR and HdNiR constructs at frequencies close to (3310-3350 G), where contributions from molecules with a wide range of  orientations are overlapped. This reduces the orientation selection at this magnetic fields and allows Tikhonov regularisation to be employed in this study. [12][13][14] S9 Figure S3. CW-EPR spectra of the full-length and 'core' regions of RpNiR measured at 20 K (top) and

X-band continuous-wave EPR and Simulations of the RpNiR-WT and its 'core'
its simulation (middle and bottom) as reported previously. [15] The EPR signal observed at 2100 G in the full-length RpNiR, indicated by the black asterisk mark is due to the heme c cofactor, which is absent   RpNiR core), respectively. The EPR signal observed at 2250 G, indicated by the black asterisk mark is due to the ferric heme c species, which is absent in the 'core' region of RpNiR, consistent with the cw-EPR results. The PELDOR data collected at different pumping and detecting fields for the RpNiR full length sample did not produce additional distances corresponding to longer ferric heme-Cu II distances.
EPR spectrometer conditions: π-pulse length = 32 ns, τ = 0. 2 s, short repetition rate = 5 ms, average microwave frequency = 9.674 GHz. (bottom) Simulation of the RpNiR 'core' using the spin-Hamiltonian parameters as reported previously. [15] The relative weight of the T2Cu(II) to that of the T1Cu(II) centre is ~ 26%. The discrepancy in weight between cw-and pulsed-EPR method is currently not known and is likely associated with the relaxation effects of the T1Cu and T2Cu centres.  Figures S5 and S6. From the modelling of the pulsed EPR spectra, it is clear that the relative weight of the T2Cu(II) to that of the T1Cu(II) centre is ~ 25-30%.

S14
Four-pulse PELDOR spectrum and distance distribution of the RpNiR 'core' Figure S9. Four-pulse PELDOR spectrum of RpNiR 'core'. The pulse separations and the lengths of the 'detection/pump' microwave pulses were as described in scheme S1. The spectrum was recorded at 3310 G with  detection = 9.5779 GHz and  pump = 9.6778 GHz; The ∆ν (ν det -ν pump ) was --100 MHz, whereas in all other constructs, the ∆ν was  100 MHz. The DEER trace has been averaged over '8780' scans with the shot repetition time of 3060 s. The validation of a distance distribution is given in Figure   S10 (linked to main text 2C)  Figure 2C).

S15
Four-pulse PELDOR spectrum and distance distribution of the HdNiR 'core' Figure S11. Four-pulse PELDOR spectrum of HdNiR 'core'. The pulse separations and the lengths of the 'detection/pump' microwave pulses were as described in scheme S1. The spectrum was recorded at 3310 G with  detection = 9.7742 GHz and  pump = 9.6739 GHz; the ∆ν (ν det -ν pump ) was +100 MHz.
The DEER trace has been averaged over '21681' scans with the shot repetition time of 5000 s. The validation of a distance distribution is given in Figure S12 (linked to main Figure 2D).

S16
Four-pulse PELDOR spectrum and distance distribution of the full-length RpNiR Figure S13. Four-pulse PELDOR spectrum of RpNiR. The pulse separations and the lengths of the 'detection/pump' microwave pulses were as described in Scheme S1. The spectrum was recorded at 3310 G with  detection = 9.7728 GHz and  pump = 9.6728 GHz; the ∆ν (ν det -ν pump ) was +100 MHz and the validation of a distance distribution is given in Figure S14 (linked to main Figure 3A). The DEER trace has been averaged over '11930' scans with the shot repetition time of 3060 s. S17 Figure S15. Four-pulse PELDOR spectrum of WT-RpNiR. The pulse separations and the lengths of the 'detection/pump' microwave pulses were as described in scheme S1. The spectrum was recorded at 3364 G with  detection = 9.5757 GHz and  pump = 9.6757 GHz; the ∆ν (ν det -ν pump ) was -100 MHz and the analysis performed using DeerAnalysis2015 is given in Figure S16. The PELDOR trace has been averaged over '11626' scans with the shot repetition time of 3060 s. S18 Figure S17. Four-pulse PELDOR spectrum of RpNiR. The pulse separations and the lengths of the 'detection/pump' microwave pulses were as described in scheme S1. The spectrum was recorded at 3364 G with  detection = 9.7730 GHz and  pump = 9.6728 GHz; The ∆ν (ν det -ν pump ) was -100 MHz and the distance distribution analysis performed using DeerAnalysis2015 is given in Figure S18. The PELDOR trace has been averaged over '23000' scans with the shot repetition time of 3000 s. Four-pulse PELDOR spectrum and distance distribution of the full-length HdNiR Figure S20. Four-pulse PELDOR spectrum of HdNiR-WT. The pulse separations and the lengths of the 'detection/pump' microwave pulses were as described in scheme S1. The spectrum was recorded at 3335 G with  detection = 9.7752 GHz and  pump = 9.6752 GHz; the ∆ν (ν det -ν pump ) was +100 MHz.
The DEER trace has been averaged over '15100' scans with the shot repetition time of 5000 s. The validation of a distance distribution is given in Figure S21 (linked to main Figure 3B).

S21
Relaxation times -T1-inversion recovery and phase-memory times of the 'core' and full-length (WT) proteins Figure S22. T 1 -inversion recovery (top) and phase-memory time, T 2 (bottom) measured on the HdNiR 'core' and full-length (WT) protein samples at 10 K. The magnetic fields at which T 1 and T 2 were measured are 3320 G and 3355 G for the HdNiR 'core' and full-length protein samples, respectively. Figure S23. T 1 -inversion recovery (top; 20 K) and phase-memory time, T 2 (bottom; 10 K) measured on the RpNiR 'core' (black traces) and full-length (FL; red traces) protein samples at 3310 G. in Figure S3 and Figure S7, respectively.

S25
Four-pulse PELDOR spectrum and distance distribution of the one-electron reduced RpNiR-"core" Figure S26. Four-pulse PELDOR spectrum of the one-electron reduced, RpNiR-core (top panel). The pulse separations and the lengths of the 'detection/pump' microwave pulses were as described in scheme S1. The spectrum was recorded at 3310 G (top) with  detection = 9.5757 GHz and  pump = 9.6757 GHz; the ∆ν (ν det -ν pump ) was -100 MHz and the validation of a distance distribution is given in the bottom panel. The DEER trace has been averaged over "12884" scans with the shot repetition time of 3060 s. It is noteworthy to mention that the distance distribution peak at 4.12 nm is no longer observedwhich was assigned to the T1Cu-T1Cu and one of the T1Cu-T2Cu distances in the RpNiR 'core'.

Author Contributions
TMH conceived and designed experiments, prepared samples, collated, analyzed and interpreted data, wrote the manuscript, and managed the project. AII prepared samples and supported data interpretation. DC prepared samples. DH secured funds and was involved in project management. MS conceived and designed experiments, performed EPR measurements, analysed data and helped finalize the manuscript (prepare figures and edited the final manuscript). NSS secured funds and directed the project.