Protonation/deprotonation-driven switch for the redox stability of low-potential [4Fe-4S] ferredoxin

Reviewer #1 (Public review):

Summary:

The authors introduced neutron crystallography coupled with room temperature X-ray crystallography to exam the redox properties of the BtFt [4Fe-4S] cluster expressed in E. coli. Neutron structure allowed the authors to exam the influence of Asp64 on the redox properties of the [4Fe-4S] cluster. The neutron structure also allowed for the identification of the hydrogen network around the [4Fe-4S] structure. This work was followed by density functional theory calculation to examine different redox states which also pointed to the role of Asp64 in affecting or dictating redox function of the [4Fe-4S] cluster. Based on the DFT work the authors examine the redox properties under oxic and anoxic conditions in wild type enzymes and in a D64N mutant again showing the role of Asp64 on the redox kinetics and redox potential of the [4Fe-4S] cluster. Lastly, the authors examined similar [4Fe-4S] ferredoxins from several organisms and with a Asp64 or Glu64 observed a similar role of Asp64 on the low potential state of the [4Fe-4S] cluster. The major conclusion of the study was to identify the role of specific amino acids, in this case Asp64, in controlling the redox state and kinetics of [4Fe-4S] clusters. The authors also demonstrate the strength of neutron crystallography when combined with classical X-ray crystallography and classical spectral/redox studies.

Strengths:

In general, the experimental design is logical and the results are convincing demonstrating the role of Asp64 on the redox properties of [4Fe-4S] clusters in ferredoxins.

Weaknesses:

The role(s) of coordinating amino acids on the redox properties of a functional group is not surprising, this reviewer believes this is a novel result in ferredoxins and does make a nice contribution to the field.

Reviewer #2 (Public review):

In this study, Wada et al. investigate the low potential ferredoxin from Bacillus thermoproteolyticus (BtFd) using a combination of neutron crystallography, x-ray crystallography, DFT and spectroscopy to determine the influence of hydrogen bonding networks on the redox potential of ferredoxin's 4Fe-4S cluster. The use of neutron diffraction allowed the authors to probe the precise location of hydrogens around the 4Fe-4S cluster, which was not possible from prior studies, even with the previously reported high-resolution (0.92 Å) structure of BtFd. This allowed the authors to revise prior models of the proposed H bonding network theorized from earlier x-ray crystallography studies ( for example, showing that there is not in fact a H bond formed between the Thr63-O𝛾1 and the [4Fe-4S]-S4 atoms). With this newly described H-bonding network established, the electronic structure of the 4Fe-4S cluster was then investigated using DFT methodology, revealing a startling role of the deprotonated surface residue Asp64, which bears substantial electronic density in the LUMO which is otherwise localized to the 4Fe-4S cluster. While aspartate is usually deprotonated at physiological pH, the authors provide compelling evidence that this aspartate has a much higher pKa than is usual, and is able to act as a protonation-dependent switch which controls the stability of the reduced state of the 4Fe-4S cluster, and thus the redox potential.

The findings of this study and the conclusions drawn from them are well supported by the data and computational work. Their findings have implications for similar control mechanisms in other, non-ferredoxin 4Fe-4S bearing electron transport proteins which have yet to be explored, providing great value to the metalloprotein community. One change that the authors may consider to enhance the clarity of the manuscript regards the nomenclature used for the varying models discussed (CM, CMNA, CMH and so forth). It would be beneficial to the reader if the nomenclature included the redox state (ox. vs red.) of the model in the model's name.

Comments on revisions:

I'm satisfied with their revisions, it looks good to me.

Author response:

The following is the authors’ response to the original reviews.

Reviewer #1:

Reviewer #1 was very appreciative of our results and commented “This is a novel result in ferredoxin and a significant contribution to the field”. We are very honored and pleased.

Reviewer #2:

(1) Changing the nomenclature of the models investigated to include the oxidation state being discussed. As they are now (CM, CMNA, etc), multiple re-reads were required to ascertain which redox state was being discussed for a particular model in a given section of the text. Appending "Ox" or "Red" for oxidized or reduced would be sufficient.

As you indicated there are several nomenclatures to distinguish the model systems in the text. On the other hand, the main issue discussed in the text is the ionization potential (IP), which is calculated by the difference in energies between oxidized and reduced states for each model. In other words, a discussion of the IP value on each model includes both the “Ox” and “Red” energies. In order to clarify the relationship between the nomenclature of models and redox states, we added sentences below.

“Note that the IP value is obtained for each model by calculating both the Ox and Red state energies of the model.” (lines 195-196).

On the other hand, we must specify the charge state when the geometry optimization is performed for CM and CMH models. Therefore, we revised the sentence as follows.

“The decrease in |IP| value indicates that the relative stability of the Red state is suppressed compared with the CMH but is significantly larger than the CM, suggesting the importance of the protonation of Asp64 (Fig. S2B).

To consider the effect of the structural change caused by the redox on the IP, geometrical optimization of the 4Fe-4S core was performed for the CM (Red) and CMH (Red) models using the same level of theory to the single-point calculations. The optimized Cartesian coordinates are summarized in Table S3. As illustrated in Fig. S2A, the IP values of CM and CMH change from –3.27 to –2.38 eV (|DIP| = 0.89 eV), and from –1.06 to –0.19 eV (|DIP| = 0.87 eV), respectively, before and after the geometrical optimization.” (lines 224-232)

(2) In addition to the very thorough DFT investigation of the different spin and charge combinations, did the authors try a broken-symmetry calculation to obtain the ground state description of the FeS cluster? Given the ubiquity of this approach in other FeS cluster studies, it was surprising that this approach was not taken here. Granted, the DFT investigation of each possible combination is sufficiently thorough and need not be redone.

Thank you for your comments. A term “spin-unrestricted method”, which is used in the manuscript in the text is synonym of “broken-symmetry method”. In order to emphasize this, we revised the manuscript as follows.

“All calculations were performed by using the spin-unrestricted (broken-symmetry) hybrid DFT method with the B3LYP functional set. As the basis set, 6-31G* and 6-31+G* were used for [Fe, C, N, O, H] and [S] atoms, respectively, for the IP calculations.” (Line 451)

(3) Line 161 "an" to "a"

We corrected the mistake. Thank you so much. (Line 161)

(4) Figure 4A seems a bit odd. Why do the traces eclipse the y-axis? And the traces between 330 and 370 nm are much noisier and appear thicker than the rest of the plot. Is this an issue with the monochromator grating used in wavelength selection? Reducing the thickness of the individual traces may help the data presentation in this figure. Also, the arrows on the plot have an opaque white background. Can this be removed so that the arrows do not eclipse the traces in the plot?

The spectrum in the Fig.4A seemed to be odd. The spectral figure has been revised to improve its appearance. (We have also corrected E53A in Figure 5B.) This reviewer also pointed out that “the traces between 330 and 370 nm are much noisier”. We are struggling with the noise caused by the grating (or the motor malfunction) of the monochromator as you pointed out. Once the monochromator is repaired and a smooth spectrum is obtained, we will upload further revisions.

(5) Figure S9 is a very nice schematic illustrating the general findings of the study. Can this be moved to the main text?

Thank you for your helpful comment. Accordingly, the Fig.9S and its legend are moved to the main text. (Lines 675-680)