Silver Binding to Bacterial Glutaredoxins Observed by NMR
, Matthew J. Panzner 6, Leah P. Shriver 7,8 and Thomas C. Leeper 9,*
Round 1
Reviewer 1 Report
The authors described the first 3D structure of a Glutoredoxin (GRX) with its native CPYC active site bound to a metal ion (silverI): GRX adopts a dimeric form in the presence of silver(I), which is further supported by clear complementary approaches ; enzymology, SEC chromatography, and MD simulations. Their results support the fact that GRXs may transfer a variety of metals between proteins and that disruption of those transfers by xenobiotic metals may lead to ROS accumulation. This last point could open original ways to stimulate ROS-mediated antibiotics in the context of the health emergency of antibiotic resistance.
Major points
General comments:
Recent references are clearly missing, especially in the introduction section. Original references should be retained, but supplemented with recent works. Currently, the most recent references cited throughout the article are from 2013.
Results 3.1
Figure 2b: The authors stated that the spectra are identical (if so, move the superimposition into the supplementary data, which allows for a larger view of Figure 2a), but at least two differences are visible (and not labelled). Please clarify this point.
Results 3.2
Table S1 & Figure S3. Shifts are clear for C19. However, they are not clear for C16. Why do the authors state that both C19 and C16 are modified by the metal binding?
“Helix 2 is angled diagonally across the βsheet in the silver liganded structure, whereas, in the non-metalated protein this helix is generally perpendicular to the axis of the β-sheet”. This statement is really difficult to check on Fig3. Please add another orientation highlighting this point, here or as supplementary data.
Results 3.7
Figure7: Please explain what is shown on Fig 7b and 7d
Minor points
Thank you for clarifying that Brucella melitensis is a Gram-negative bacteria, in the abstract and in its first citation.
I am not sure that all abbreviations are clearly indicated (IPTG, HRV, HEDS….). Please check.
Mat & Meth 2.2:
Is there a reason why the authors used two different NMR fields?
Mat & Meth 2.5:
Please check the nomenclature “BCBC(CO)NH”
Ref 25 should be supplemented with recent references
Mat & Meth 2.6:
Put the numbers in subscript for H20 and D20
Ref 27 corresponds to DIANA and not to the more recent CYANA software. Please modify.
Mat & Meth 2.8:
Please add a reference for Bio Tool kit
Fig 1 should appear in the results part (and not in the Mat & met)
Results 3.1
Figure 2a: The authors should highlight the key residues R13 C16 C19 Y18 on the spectrum.
Figure 3: Figure 3 should appear in the 3.3 section (and not 3.2). To make the reading easier, please label the helices. The legend “a) ensemble” should be modified for example into “superimposition of X selected structures, rainbow colored from the N-terminal (in blue) to C-terminal (in red).
Figure 3C does not exist but is cited in the text.
Results 3.8
This figure is a key figure. Please label the residue, which are displayed in stick and blue. And the authors should add the PDB code for the model of the silver metalated dimer structure displayed in grey.
Results 3.10
Please explain why you show a sigmoidal trend (Fig11c) and a parabolic trend (Fig 11d) instead of showing fit curves. It is certainly more rigorous to do this, but it requires some explanation.
Author Response
Reviewer 1:
We thank the reviewer for the helpful comments and suggestions. To address these, we have updated the references with more current literature, have edited figure captions and figures to clarify relevant points, updated abbreviations lists and acronyms, and fixed various editorial concerns. While not addressed in the text, the lower field instrument was used for control titration experiments with other metal ions when the higher field instrument was unavailable. We do observe a significant changes for C-16 Calpha and amide during metalation and are uncertain why C-16 Cbeta doesn’t experience large changes. While the Cbeta would be expected to shift, that residue is known to have lower pKa and hence likely to be in thiolate state that may be more similar to metalated chemical environment than a non-metalated thiol but we are not prepared to include this speculation in this document and may be an area for future research in GRX-metal complexes. Regarding figure 2b, we note that the very subtle difference between copper(I) and (II) metalated forms may be due to broadening from residual copper(II); we would expect much more significant broadening (and/or CSP) if a substantial portion of the copper(II) avoided reduction, and also noticed some very small changes which we attributed to trace amount of copper(II). We have also provided a clarification of the HEDS data as requested. Detailed list follows (author comments in italics):
Major points
General comments:
Recent references are clearly missing, especially in the introduction section. Original references should be retained, but supplemented with recent works. Currently, the most recent references cited throughout the article are from 2013.
Additional literature support provided throughout.
Results 3.1
Figure 2b: The authors stated that the spectra are identical (if so, move the superimposition into the supplementary data, which allows for a larger view of Figure 2a), but at least two differences are visible (and not labelled). Please clarify this point.
Copper (II) is paramagnetic so those two residues that look different between the 2 spectra may be a function of line broadening from some residual/incomplete reduction. We have updated the figure caption (line 240) and text (lines 233 to 234) to reflect this observation.
Results 3.2
Table S1 & Figure S3. Shifts are clear for C19. However, they are not clear for C16. Why do the authors state that both C19 and C16 are modified by the metal binding?
The shift for the Ca on C16 is not as dramatic as the shift for the same carbon in C19, but we still observe a shift from 56.315 (apo) to 55.24 (silver).
“Helix 2 is angled diagonally across the βsheet in the silver liganded structure, whereas, in the non-metalated protein this helix is generally perpendicular to the axis of the β-sheet”. This statement is really difficult to check on Fig3. Please add another orientation highlighting this point, here or as supplementary data.
We have updated figure 3 and in the revised figure altered the angle slightly to look down the helix in metalated structure highlighting how the helix angle shifts in apo form.
Results 3.7
Figure7: Please explain what is shown on Fig 7b and 7d
We have edited the figure caption to read solvent exposed residues protected from broadening (lines 381 to 383).
Minor points
Thank you for clarifying that Brucella melitensis is a Gram-negative bacteria, in the abstract and in its first citation.
I am not sure that all abbreviations are clearly indicated (IPTG, HRV, HEDS….). Please check.
Missing abbreviation definitions have been addressed (throughout)
Mat & Meth 2.2:
Is there a reason why the authors used two different NMR fields?
The low field instrument was used for alternative metals, nickel, cobalt, and mercury since the 750 MHz was unavailable and these were essentially ancillary controls relative to copper and silver which we were more keenly interested in studying. This protein is sufficiently small to have excellent HSQC spectra even at 500 MHz.
Mat & Meth 2.5:
Please check the nomenclature “BCBC(CO)NH”
We have corrected this to CBCA(CO)NH (line 133)
Ref 25 should be supplemented with recent references
We made the citation with the reference for the version of CCPNMR that this work was performed under. We have indicated in the text that this work was performed with version 2 (line 135).
Mat & Meth 2.6:
Put the numbers in subscript for H20 and D20
We have made this correction (line 139).
Ref 27 corresponds to DIANA and not to the more recent CYANA software. Please modify.
We have added the more recent CYANA citation from 2004 (line 143)
Mat & Meth 2.8:
Please add a reference for Bio Tool kit
Bio Tool Kit is a plugin for commercial software PeakView®. The supplier, Ab Sciex, is mentioned (line 175).
Fig 1 should appear in the results part (and not in the Mat & met)
Fixed.
Results 3.1
Figure 2a: The authors should highlight the key residues R13 C16 C19 Y18 on the spectrum.
Red boxes added to figure 2 to highlight these residues.
Figure 3: Figure 3 should appear in the 3.3 section (and not 3.2). To make the reading easier, please label the helices. The legend “a) ensemble” should be modified for example into “superimposition of X selected structures, rainbow colored from the N-terminal (in blue) to C-terminal (in red).
Figures moved.
Figure 3C does not exist but is cited in the text.
Added.
Results 3.8
This figure is a key figure. Please label the residue, which are displayed in stick and blue. And the authors should add the PDB code for the model of the silver metalated dimer structure displayed in grey.
The residue has been added to the caption (line 406). The model presented in Figure 2, while based upon the metaled NMR structure (Figure 3c), is a dimer model generated in MUSTANG based upon known dimer structures of iron-sulfur cluster dimers; so not appropriate to be deposited to the pdb.
Results 3.10
Please explain why you show a sigmoidal trend (Fig11c) and a parabolic trend (Fig 11d) instead of showing fit curves. It is certainly more rigorous to do this, but it requires some explanation.
We edited the text in this results section (lines 457 to 467) to better clarify how these plots justify the dimeric inhibition model. It is not our intention to offer binding constants based on these rate data, so we felt fitting curves to these plots would not be an appropriate exercise. The text now clarifies that the sigmoidal rate data is a step along the way to producing the parabolic trend data. Then, the Dixon plots are described how they support dimeric inhibition in a qualitative manner. We hope this satisfies the reviewer without additional data manipulation.
Reviewer 2 Report
The manuscript presents a NMR-based interaction between a bacterial glutaredoxin and a silver ion. This metal complex has also been characterized using various methods. Generally, the manuscript is well-written and concise, and the figures appropriate.
I have though a number of comments that should be addressed:
Abstract
L16: “The CPYC active site of GRX is a plausible metal binding site but was previously theorized not to bind metals due to its cis-proline configuration.”
This is not really wrong but it has to be rephrased differently. At least two publications show that these CPYC-glutaredoxins can incorporate an iron-sulfur cluster (Ceylan et al., 2010 & Roret et al., 2021). The proline residue would rather prevent dimerization around an iron-sulfur cluster than the incorporation of a single atom.
Introduction
L30: “GRX’s redox centers consist of either a dithiol active site (CPYC) or a monothiol active site (CGFS) [1].”
This is too brief and this suggests to the reader that there are only 2 different types of glutaredoxins. Maybe the authors should mention the existence of all classes and put that class I is more of the type CxxC/S, CGFS for class II, CCxx for class III, …
L42: “One hypothesis for this specificity is that the proline in the GRX active site prevents metal binding [10].”
This is not entirely true. The key role of this proline is unclear. In some cases, it seems that the proline prevents the iron-sulfur cluster binding. However, some mutagenesis experiments (Hoffmann et al., 2011) and some true CPYC-glutaredoxins (Trypanosoma brucei Grx1 and Chlamydomonas reinhardtii Grx2) show that despite the presence of proline a cluster can be incorporated. As some experiments were not done in a glovebox, it is possible that other CPYC-glutaredoxins are capable of it.
Materials and Methods
L66: “The C70S mutation was transformed in to BL21(DE3) E. coli cells …”
It is not the C70S mutation but the Brucella melitensis glutaredoxin C70S variant which was transformed into E. coli.
L77: “under inert gas using a stirred cell”
If the experiments were carried out in a glove box, it must be specified.
L86: “CSP studies …”
The meaning of CSP should be added the first time “Chemical Shift Perturbation”.
L125: “A dimer model of the PDB coordinates of the final output from CYANA were created by aligning a duplicate for each of the twenty calculated structures onto the Fe2S2 induced dimer of E. coli glutaredoxin (PDB 2WCI) using the MUSTANG algorithm”
I wondered why you used a CGFS-Grx (class II) as a model while you study a CPYC-Grx (class I)? Classes I and II are known to have different dimerization modes around a [Fe2-S2]-cluster. I think it is better to use a PDB entry like 2E7P, 2HT9, 3RHC or 5J3R as a template.
L131: “distances between the dimer 16C were calculated”
At the first reading I wondered what 16C was. It must be indicated that it is the catalytic cysteine numbered 16.
Results
L171: “HSQC of apo GRX (red) and 0.5 equivalents of silver (blue).”
You have to change the colors because it's not red.
L209: “Overlay of Cu(I) (green) and Cu(II) (red) shows identical chemical shift perturbations upon ligand binding.”
Is it possible to use the same scale for all 1H-15N HSQC?
L230: “superposition of apo (green) and Ag-metalated protein (blue) based on an alignment of C-alphas.”
You show the structure of the Ag-metalated protein but you show neither the Ag nor the residues involved in the coordination.
L253: “Figure 3c”
There is no figure 3C.
L256: “The NMR CSP experiments indicate that a dimer forms around silver …”
To be rephrased according to results in section 3.4. As both apo- and TCEP-reduced glutaredoxin exist as a dimer (as shown by SEC), there is no formation of a dimer around the silver atom but rather its incorporation into the dimer.
L269: “but also a nonspherical shape, like that of a prolate spheroid …”
There are numerous examples of SEC performed on glutaredoxin holodimers in the literature. References could be added here.
L431:” Table 1. Metals Tested for GRX Binding.”
There are no titles in the table for the columns.
Author Response
Reviewer 2:
We thank the reviewer for the helpful comments and suggestions. To address these, we have updated the references with more current literature, have edited figure captions and figures to clarify relevant points, revised the background section to clarify assertions in the literature regarding GRXs (i.e. flexibility of role of CPYC vs CGFS vs CXXC and impact of proline), and fixed various editorial concerns. We note in the text our rationale for choosing 2wci rather than three of the templates suggested (6j3r was not present at the time). Detailed list follows (author comments in italics):
Abstract
L16: “The CPYC active site of GRX is a plausible metal binding site but was previously theorized not to bind metals due to its cis-proline configuration.”
This is not really wrong but it has to be rephrased differently. At least two publications show that these CPYC-glutaredoxins can incorporate an iron-sulfur cluster (Ceylan et al., 2010 & Roret et al., 2021). The proline residue would rather prevent dimerization around an iron-sulfur cluster than the incorporation of a single atom.
We have added addition text to clarify this notion (lines 47 to 58).
Introduction
L30: “GRX’s redox centers consist of either a dithiol active site (CPYC) or a monothiol active site (CGFS) [1].”
This is too brief and this suggests to the reader that there are only 2 different types of glutaredoxins. Maybe the authors should mention the existence of all classes and put that class I is more of the type CxxC/S, CGFS for class II, CCxx for class III, …
We have added addition text to clarify the issues relating to these classification (lines 33 to 35 and 53 to 56).
L42: “One hypothesis for this specificity is that the proline in the GRX active site prevents metal binding [10].”
This is not entirely true. The key role of this proline is unclear. In some cases, it seems that the proline prevents the iron-sulfur cluster binding. However, some mutagenesis experiments (Hoffmann et al., 2011) and some true CPYC-glutaredoxins (Trypanosoma brucei Grx1 and Chlamydomonas reinhardtii Grx2) show that despite the presence of proline a cluster can be incorporated. As some experiments were not done in a glovebox, it is possible that other CPYC-glutaredoxins are capable of it.
We have added addition text related to these concern (lines 49 to 58 & 93 to 94). We had actually performed the reduction and removal of chemical reductant via dialysis in a glove bag (under argon) initially and found no apparent difference between stirred cell exchange and traditional dialysis in both apo and metalated HSQC spectra. We migrated to the stirred cell approach since it saved significant time, steps, and reagents. While other CPYC GRXs may incorporate metals, in fact we have unpublished data that all other bacterial CPYC-type GRXs that we have access to (4?) do metalate with silver et al., we do not think that this particular GRX metalation is a procedural artifact. We are not ready to publish or speculate upon the generality of bacterial GRX metalation yet…
Materials and Methods
L66: “The C70S mutation was transformed in to BL21(DE3) E. coli cells …”
It is not the C70S mutation but the Brucella melitensis glutaredoxin C70S variant which was transformed into E. coli.
Clarified (lines 77 and 78).
L77: “under inert gas using a stirred cell”
If the experiments were carried out in a glove box, it must be specified.
We have corrected the text to indicate that nitrogen (and/or argon?) gas was used in the stirred cell (lines 93 & 94). See above explanation relating to switch from glove bag to stirred cell exchange.
L86: “CSP studies …”
The meaning of CSP should be added the first time “Chemical Shift Perturbation”.
We have made this correction to define CSP in the text (line 101).
L125: “A dimer model of the PDB coordinates of the final output from CYANA were created by aligning a duplicate for each of the twenty calculated structures onto the Fe2S2 induced dimer of E. coli glutaredoxin (PDB 2WCI) using the MUSTANG algorithm”
I wondered why you used a CGFS-Grx (class II) as a model while you study a CPYC-Grx (class I)? Classes I and II are known to have different dimerization modes around a [Fe2-S2]-cluster. I think it is better to use a PDB entry like 2E7P, 2HT9, 3RHC or 5J3R as a template.
We have provided a short rationale (lines 150 & 151) for this choice: quality of X-ray structure and % identity. I don’t think 5j3r was available at the time but using those criteria, 2wci would still be chosen.
L131: “distances between the dimer 16C were calculated”
At the first reading I wondered what 16C was. It must be indicated that it is the catalytic cysteine numbered 16.
Reworded (line 155 & 156) to clarify.
Results
L171: “HSQC of apo GRX (red) and 0.5 equivalents of silver (blue).”
You have to change the colors because it's not red.
We have changed the figure heading to read maroon (line 212 & 213).
L209: “Overlay of Cu(I) (green) and Cu(II) (red) shows identical chemical shift perturbations upon ligand binding.”
Is it possible to use the same scale for all 1H-15N HSQC?
Figure 2b updated.
L230: “superposition of apo (green) and Ag-metalated protein (blue) based on an alignment of C-alphas.”
You show the structure of the Ag-metalated protein but you show neither the Ag nor the residues involved in the coordination.
We show this in figure 8.
L253: “Figure 3c”
There is no figure 3C.
Fixed.
L256: “The NMR CSP experiments indicate that a dimer forms around silver …”
To be rephrased according to results in section 3.4. As both apo- and TCEP-reduced glutaredoxin exist as a dimer (as shown by SEC), there is no formation of a dimer around the silver atom but rather its incorporation into the dimer.
Clarified this idea (line 290 & 291).
L269: “but also a nonspherical shape, like that of a prolate spheroid …”
There are numerous examples of SEC performed on glutaredoxin holodimers in the literature. References could be added here.
See line 306.
L431:” Table 1. Metals Tested for GRX Binding.”
There are no titles in the table for the columns.
There are titles. They are “Metal, Binds?, Binding residues, and exchange”
