Probing the Reactivity of [4Fe-4S] Fumarate and Nitrate Reduction (FNR) Regulator with O₂ and NO: Increased O₂ Resistance and Relative Specificity for NO of the [4Fe-4S] L28H FNR Cluster

Round 1

Reviewer 1 Report

In this paper, Crack et al. studied iron-sulfur cluster regulator proteins such as FNR discriminate between alternative signaling molecules. FNR [4Fe-4S] cluster regulator acts as switch between anaerobic aerobic respiration in response to O2 availability, whereas FNR also reacts with nitric oxide. The authors show the L28H mutant of FNR is much less reactive towards O2 than wild type FNR. This paper is an interesting approach to characterize FNR. There, however, several concerns that needed to be addressed before it is accepted for publication.

Molecular dynamic simulation using the [4Fe-4S] L28H FNR crystal structure as a starting model show that the formation of His28-Arg184 cation-π interaction is an important factor for the flexibility of Cys loop structure. However, Arg184 and His28 are close to each other but a cation-π interaction is not observed. The MD simulation of 1-ms for [4Fe-4S] L28H suggest the interaction. Why the stable interaction is present as a starting crystal. 

Author Response

In this paper, Crack et al. studied iron-sulfur cluster regulator proteins such as FNR discriminate between alternative signaling molecules. FNR [4Fe-4S] cluster regulator acts as switch between anaerobic aerobic respiration in response to O2 availability, whereas FNR also reacts with nitric oxide. The authors show the L28H mutant of FNR is much less reactive towards O2 than wild type FNR. This paper is an interesting approach to characterize FNR. There, however, several concerns that needed to be addressed before it is accepted for publication.

Molecular dynamic simulation using the [4Fe-4S] L28H FNR crystal structure as a starting model show that the formation of His28-Arg184 cation-π interaction is an important factor for the flexibility of Cys loop structure. However, Arg184 and His28 are close to each other but a cation-π interaction is not observed. The MD simulation of 1-ms for [4Fe-4S] L28H suggest the interaction. Why the stable interaction is present as a starting crystal. 

Response:

The reviewer is presumably asking why the cation-π interaction is not present in the crystal or at the start of the simulation. This is a good question and, we were puzzled by this also. It is important to recognize that the structure, while extremely useful, is only a snapshot, and other factors, such as crystal packing forces, can influence the conformation such that it does not necessarily represent the solution structure in all aspects. Indeed, an examination of crystal contacts indicates that crystal packing forces likely keep the two residues too far apart to interact. That was why the MD simulations were so important to explore whether such a cation-π interaction can occur that might account for the increased order of the cluster loop that is observed in the crystal structure of L28H compared to wild type FNR. We have added some further explanation of this to the discussion section.

Reviewer 2 Report

The main objection of this reviewer raised by the fact that the authors use a NO donor (PROLI-NONOate), which has a very short half-life of 1.5 seconds. So, the question is whether the NO concentration is stable and effective during the course of each experiment since the duration of each titration step and the acquisition of the measurements through spectroscopies and/or stopped-flow are much longer (from several seconds to minutes) than the half-life of the NO donor. Indeed, some of the experiments have been performed using excess of NO, suggesting that the measurable effect is concentration-dependent, or the results are drawn using different NO concentration in different experiments.

Interaction of the titration experiments has as important parameter the molar ratio of the interacting molecules, and in this case the effective concentration of the NO into the medium of the interaction cannot be precisely calculated. All diagrams in Figure 3 (UV-vis, CD, EPR) and Figure 4 (Stopped-flow) are drawn as a function of NO concentration, but this concentration is hard to be measured when the half-time of the NO donor is only 1.5s.

It would be suggested to the authors to repeat at least one set of experiments using a NO donor with a longer half-life.

Additional comments

Under paragraph “Electrophoretic mobility shift assays” in section “2.2. Mass spectrometry” the quantities of reactions, DNA, loading due etc. are provided in liters (L). This amount seems unreasonably high; this is probably a typo. The authors are advised to confirm the units of measure.

In Table 1, samples 1 and 2 are referred to by the same designation, although the table caption indicates they are different samples. The designations need to be revised to describe each sample accordingly.

Figure 2 refers to section 3.1 but is under section 3.2. It needs to be placed under section 3.1.

Figure 8 refers to section 3.6 but is under section 4. It needs to be placed under section 3.6.

Author Response

The main objection of this reviewer raised by the fact that the authors use a NO donor (PROLI-NONOate), which has a very short half-life of 1.5 seconds. So, the question is whether the NO concentration is stable and effective during the course of each experiment since the duration of each titration step and the acquisition of the measurements through spectroscopies and/or stopped-flow are much longer (from several seconds to minutes) than the half-life of the NO donor.

NONOates provide a convenient way of storing nitric oxide. NONOates can be prepared at concentrations above the solubility limit of NO gas, ideal for titrations, and are otherwise stable provided they are stored correctly. Proli-NONOate is the fastest of the NONOates series spontaneously releasing two molecules of NO in a few seconds (Saavedra JE et al. J Med Chem. 1996 39, 4361-5. doi: 10.1021/jm960616s; Keefer LK et al. Methods Enzymol. 1996 268, 281-93. doi: 10.1016/s0076-6879(96)68030-6).  NO is quite stable in solution, provided O₂ is excluded. In the presence of O₂, however, free NO concentrations fall extremely rapidly (Lewis RS et al. Chem Res Toxicol. 1994 7, 568-74. doi: 10.1021/tx00040a013).

For stopped-flow experiments, and prior to use, ProliNONOate was allowed to decay to aqueous NO inside a gas tight (with zero headspace) syringe in an anaerobic cabinet. Without, decaying the NONOate prior to use, the decay reaction of the NONOate would be the rate-limiting step in all subsequent reactions. The stopped-flow system was flushed with copious amounts of anaerobic buffer prior to use and wild type FNR was used to ‘detect’ residual oxygen prior to commencing NO experiments. We have added this information to the experimental section.

For optical titrations, samples were measured using anaerobic cuvettes and manipulated in an anaerobic cabinet. During the titration, each ProliNONOate aliquot passed through a ~100 half-lives prior to measurement. Replacing ProliNONOate with a slower releasing NONOate would simply increase the incubation time required prior to each measurement.

Indeed, some of the experiments have been performed using excess of NO, suggesting that the measurable effect is concentration-dependent, or the results are drawn using different NO concentration in different experiments.

The optical titrations provide information about the products, reactants and stoichiometry or the reactions. For stopped-flow, excess NO was used to ensure the experimental system was operating under pseudo first order kinetics, and to determine the dependence of rate (constants) on NO concentration. The end points of the stopped-flow reactions were comparable to the species observed at the end of the optical titrations.

Interaction of the titration experiments has as important parameter the molar ratio of the interacting molecules, and in this case the effective concentration of the NO into the medium of the interaction cannot be precisely calculated. All diagrams in Figure 3 (UV-vis, CD, EPR) and Figure 4 (Stopped-flow) are drawn as a function of NO concentration, but this concentration is hard to be measured when the half-time of the NO donor is only 1.5s.

The NO concentrations stated are the initial NO concentrations (routinely calibrated), before any chemistry has taken place with the FeS cluster.

It would be suggested to the authors to repeat at least one set of experiments using a NO donor with a longer half-life.

Our view is that this would be an unnecessary and fruitless exercise. A slower NO-releasing NONOate would simply increase the incubation time required prior to each measurement. A fast-releasing NO reagent, which is allowed to fully decay in a gas tight syringe in an anaerobic cabinet, is equivalent to using a solution of dissolved NO gas.

Additional comments

Under paragraph “Electrophoretic mobility shift assays” in section “2.2. Mass spectrometry” the quantities of reactions, DNA, loading due etc. are provided in liters (L). This amount seems unreasonably high; this is probably a typo. The authors are advised to confirm the units of measure.

This appears to be an issue with journal conversion of the original manuscript. We used ‘µL’ for microliter in the Electrophoretic sections, this was converted to a @ character.

In Table 1, samples 1 and 2 are referred to by the same designation, although the table caption indicates they are different samples. The designations need to be revised to describe each sample accordingly.

Thanks - we have removed the protein designation as they are all L28H samples.

Figure 2 refers to section 3.1 but is under section 3.2. It needs to be placed under section 3.1.

Figure 8 refers to section 3.6 but is under section 4. It needs to be placed under section 3.6.

Figure placement is determined by the journal type-setters. We placed figures at the first convenient place after first mention.

Reviewer 3 Report

Fontecilla-Camps and Brun investigate the response of FNR from E. coli to O₂ or NO. FNR is an iron-sulfur protein whose “active site” [4Fe-4S] cluster falls apart under aerobic conditions, triggering structural changes that impede FNR binding to DNA. Additionally, FNR reacts with NO in a less well understood mechanism. Compared to wild-type FNR, the authors demonstrate now how EcFNR variant L28H is more O₂-resistant but equally sensitive to NO. This allows understanding the reaction with O₂ and NO. Therefore, the crystal structure of FNR variant L28H from Aliivibrio fischeri was solved (80% sequence identity with EcFNR) and analyzed in molecular dynamics simulations. Fontecilla-Camps and Brun conclude that structural changes in the outer coordination sphere of the variant limit access of O₂ to the “active site” but barely impede NO reacting with the [4Fe-4S] cluster directly.

This is a brilliant manuscript. Well written, interesting data, and a clear story. In fact, I will suggest publishing as-is. My only comment is related to Figure 3. I have a few suggestions that may improve the presentation.

Panel A: There is sufficient space for an inset highlighting the low energy regime from 500-700 nm, as discussed in the text. The arrows should be labelled with the central energy.

Panel B: Please add a legend.

Panel C: The arrows should point at the observed bands (or isosbestic points?) and be labelled with the respective energy.

Panel D: While I would recommend using color in panels (A) and (C) to make the figures look more appealing, the usage of color in panel (D) seems actually necessary to follow the transition (add a legend, too). Can a reference for the PROLI-NONOate protocol of NO release be provided? Reference (30) explains only the quantification of NO.

Thank you.

Author Response

We thank the reviewer for their enthusiastic review of the manuscript!

My only comment is related to Figure 3. I have a few suggestions that may improve the presentation.

Panel A: There is sufficient space for an inset highlighting the low energy regime from 500-700 nm, as discussed in the text. The arrows should be labelled with the central energy.

We have added an inset showing the lower energy region, and added the specific wavelength at centre of bands. As stated in the legend, the arrows only indicate the direction of change of bands.

 Panel B: Please add a legend.

We are not sure what the reviewer is referring to here – there is a legend for (B).

Panel C: The arrows should point at the observed bands (or isosbestic points?) and be labelled with the respective energy.

The arrows now refer to specific wavelengths (energies) and indicate direction of change.

Panel D: While I would recommend using color in panels (A) and (C) to make the figures look more appealing, the usage of color in panel (D) seems actually necessary to follow the transition (add a legend, too). Can a reference for the PROLI-NONOate protocol of NO release be provided? Reference (30) explains only the quantification of NO.

We have included a colour version of (D), and added references to the Experimental section that cover the release of NO from the NONOate. Again, there is a legend for part D, so we're not sure what this comment refers to.

Round 2

Reviewer 1 Report

The revised manuscript has been read as a well  modified form.