Reaction Center of Rhodobacter Sphaeroides, a Photoactive Protein for pH Sensing: A Theoretical Investigation of Charge Transport Properties
Round 1
Reviewer 1 Report
This manuscript studies the electrical response of bRC within the context of proteotronics, investigating pH and illumination effects on the conductive properties of the molecules. While the main idea is interesting and can help interpret experiments, the presentation and application of the idea in this study are somewhat far from being conclusive. I suggest that the Authors deeply consider the points raised below (some of them are very important in my opinion). The manuscript in its current form is, in my opinion, borderline between being rejected or required a major revision, because of some of the points detailed below. I am suggesting a major revision, but I would reject the manuscript in a later revision, were I involved in it, if the points below were not adequately addressed.
Page 1, line 16
It is written “an even more green trend”. One should add the term of comparison: “more” compared to what.
Page 2, line 67
Typo: “short list” should be changed to “shortlist”.
Page 2, line 68
I would suggest to change to “scenario of protein energy states” to “landscape of protein energy states”.
Page 2, line 86
I think that “effects of tuning these data” should be changed to “effects of tuning these conditions”.
Page 2, line 93
I read “… which should reduce the contribution of thermal fluctuations”. The meaning of this piece of sentence is a bit unclear (in particular, one should specify what the contribution is referred to). I guess that the Authors mean that they are fixing the temperature variable and that they rely that thermal fluctuations will not determine the protein escape from the initial free energy minimum.
Page 3, line 96
It is written “Proteins are affected …”. I would suggest to change “affected” to “influenced”, as “affect” has a somewhat negative connotation that is not necessary in this case.
Page 3
A clear and complete separation of the effects of temperature and pH would also require the analysis of pH effects at different given temperatures. The Authors should notice this fact (maybe as the subject of desirable future investigations) on this page or in the conclusions.
Page 3
The statement “proteotronics”, a recent branch of molecular electronics we have recently developed” needs to be corrected. This branch contains old works of other people too, using other methods.
Page 4, line 156
Typo: “of concerns” to be changed to “of concern”.
Page 4
In particular, it is unclear how a single macromolecule can be “representative of the specific experimental arrangements” of a device.
Page 4, Table 1
The symbols that appear in the header of Table 1 should be explicitly named in the table caption.
It is written that the structures listed in the table describe proteins crystallized at different pH values, both in dark and under illumination, but illumination was used for only one of the proteins in the table. What is the motivation for this choice?
Page 5
There can be no “static” I-V characteristics. Perhaps the Authors mean to say “steady-state”.
Page 5
I read that node pairs are linked if the involved distances are below a certain cut-off. The model should be described in more detail in the manuscript. For example, are the links differentiated according to whether they are between adjacent, covalently bound amino acids, or non-bonded amino acids, charged ones, etc.? From other parts of the manuscript, it seems that only crystal structures (without running any molecular dynamics) are considered and that the links are only electrically characterized. The electrical characterization is not sufficiently discussed, nor its feasibility and value are commented. What is the justification for using the same link resistivity for all amino acids? What supports this approximation? References [9] and [24] do not help: the first one is a book (which cannot be easily accessed by the reader; for example, I could not access it), the second one does not provide sufficient information.
It is also written that the role of the medium is “only indirectly accounted for” and that “its effects are resumed in the protein conformation”. How is this accomplished?
A nonnegligible part of the philosophy behind the model resembles that of the tunneling pathways model, although, at difference from the latter, the analysis is not at an atomistic level and the links are electrically characterized. Indeed, the two models share the problem that the charge (which is characterized by a wave function) does not really transfer this way, and sometimes not even by far so, in biomolecular systems. Yet, the approximation is discussed (and partly mitigated by introducing interferences among tunneling pathways) in the tunneling model. How is the approximation justified or its approximation defined in the present model?
Page 5, lines 199-200
The notation 1/0, instead of {1,0}, is misleading in my opinion.
The meaning of “the presence or any of a link” is unclear.
Page 7, line 262
As above, “static” should be replaced by “steady-state”.
Pages 7-8
D is a model parameter. From Figures 3 and 4, as well as their discussion, one cannot deduce which value of D better describes the system. The fact that D = 10.5 produces the most marked dependence of the resistance on the pH does not mean that this is the right value to be used. The resistance dependence on pH could be smaller than that obtained for D = 10.5 and better correspond to another D value. In other words, the model should provide a recipe for choosing an optimal D value either based on the physical grounds or based on a suitable comparison with experiments. The mentioned “resolution” has no meaning in my view, as one wants to see the pH dependence of the resistance R and not what parameter of the model would provide the largest change in R, irrespective of whether this is the real one. Such resolution might be meaningful only at a very qualitative level, if one assumes that the pH influences the charge conduction and wants to see if a pH change determines an increase or decrease in R. However, even in this case, it is difficult to understand how much it is real or just an artefact of the model. The comparison with experiments could help, but it is essentially missing in this part of the manuscript. The sentence “Also in this case, the maximal resolution shows up for a cut-off value D=10.5 Å, with a reduction of resistance in fair agreement with experiments [12]” is by far insufficient.
Page 9, line 337
The term “writes” should be changed to “reads”.
Page 10
The Authors should comment on whether and why the pH influences shown in Figure 6 are reliable based on the accuracy of the method. What would happen by using another D value?
Pages 10-11
The observed dependence of charge transfer on the energy barrier is nothing new or surprising. Inspecting the current in dark conditions for the different barrier values would have been necessary to safely assert what is written at the end of page 10. For example, how do I know that the onset of higher current for Phi = 0.206 eV does not occur in a significantly different way in dark?
Pages 11-12
The choice of rho looks just empirical: just to fit the experimental curves in Figure 8. It should be motivated (at least as for the ordering of the values) based on the kinds of experimental setups. I can think of other (very different) models that can reproduce the experimental curves in Figure 8. So, it is difficult for me to accept Figure 8 as a demonstration of the appropriateness of the method, the way it is presented. In fact, the conclusions on page 11 are not so clear to me and partly do not look ok for what I can understand from the text (for example, the match between Fermi level and frontier orbital energy should increase the current indeed).
Author Response
Answers to REFEREE1
The authors acknowledge the constructive comments of the Referee 1, in the following the answers to her/his criticisms
This manuscript studies the electrical response of bRC within the context of proteotronics, investigating pH and illumination effects on the conductive properties of the molecules. While the main idea is interesting and can help interpret experiments, the presentation and application of the idea in this study are somewhat far from being conclusive. I suggest that the Authors deeply consider the points raised below (some of them are very important in my opinion). The manuscript in its current form is, in my opinion, borderline between being rejected or required a major revision, because of some of the points detailed below. I am suggesting a major revision, but I would reject the manuscript in a later revision, were I involved in it, if the points below were not adequately addressed.
R1:Page 1, line 16
:It is written “an even more green trend”. One should add the term of comparison: “more” compared to what.
A1: The sentence has been modified as follows: “In the perspective of an increasing attention to ecological aspects of science and technology..”
R2:Page 2, line 67
R2:Typo: “short list” should be changed to “shortlist”.
A2: ok
R3:Page 2, line 68
I would suggest to change to “scenario of protein energy states” to “landscape of protein energy states”.
A3: ok
R4:Page 2, line 86
R:I think that “effects of tuning these data” should be changed to “effects of tuning these conditions”.
A4: ok
R5:Page 2, line 93
I read “… which should reduce the contribution of thermal fluctuations”. The meaning of this piece of sentence is a bit unclear (in particular, one should specify what the contribution is referred to). I guess that the Authors mean that they are fixing the temperature variable and that they rely that thermal fluctuations will not determine the protein escape from the initial free energy minimum.
A5: The sentence has been appropriately modified
R6:Page 3, line 96
It is written “Proteins are affected …”. I would suggest to change “affected” to “influenced”, as “affect” has a somewhat negative connotation that is not necessary in this case.
A6: ok
R7:Page 3
A clear and complete separation of the effects of temperature and pH would also require the analysis of pH effects at different given temperatures. The Authors should notice this fact (maybe as the subject of desirable future investigations) on this page or in the conclusions.
A7: We agree with the referee and this point has been included in the conclusions.
R8:Page 3
The statement “proteotronics”, a recent branch of molecular electronics we have recently developed” needs to be corrected. This branch contains old works of other people too, using other methods.
A8: We agree with the referee and deleted “recently”, and “we”, otherwise, the methodology and term were developed inside our research group .
R9:Typo: “of concerns” to be changed to “of concern”.:Page 4, line 156
A9: ok
R10:Page 4
In particular, it is unclear how a single macromolecule can be “representative of the specific experimental arrangements” of a device.
A10: In the specific case the model correlates the sensing effect to the conformational change of the single protein structure, i.e. no conformational change no sensing effect. Of course, this represent a first attempts to provide a microscopic model and to this purpose only relative change in the response of the protein to an external agent are considered. The comparison with experiments provides a qualitative agreement which validates the available experimental evidence, especially for the case of the bacteriorhodopsin, an other photosensitive protein, that was experimentally analyzed up to the single protein level (Casuso, I., Fumagalli, L., Samitier, J., Padrós, E., Reggiani, L., Akimov, V., & Gomila, G. (2007). Nanoscale electrical conductivity of the purple membrane monolayer. Physical Review E, 76(4), 041919.). Finally, in the present analysis, we compare our simulations with experiments performed on samples of few proteins (about 10) which should be well represented by a single protein -ref 12-
R11:Page 4, Table 1
The symbols that appear in the header of Table 1 should be explicitly named in the table caption.
It is written that the structures listed in the table describe proteins crystallized at different pH values, both in dark and under illumination, but illumination was used for only one of the proteins in the table. What is the motivation for this choice?
A11: The table caption has been modified according to the referee suggestion. Concerning the choice of studying illumination for only one of the pH values, it was detailed in figure 7 and related comments. Specifically, our calculations lead to a modest increase of current for the structures in light (independently of the pH values). Therefore, we think that the large current increase, observed in experiments, is due to the electrons injected in the protein and correlate this effect to the barrier height .
R12:Page 5
There can be no “static” I-V characteristics. Perhaps the Authors mean to say “steady-state”.
A12: ok
R13:Page 5
I read that node pairs are linked if the involved distances are below a certain cut-off. The model should be described in more detail in the manuscript. For example, are the links differentiated according to whether they are between adjacent, covalently bound amino acids, or non-bonded amino acids, charged ones, etc.? From other parts of the manuscript, it seems that only crystal structures (without running any molecular dynamics) are considered and that the links are only electrically characterized. The electrical characterization is not sufficiently discussed, nor its feasibility and value are commented. What is the justification for using the same link resistivity for all amino acids? What supports this approximation? References [9] and [24] do not help: the first one is a book (which cannot be easily accessed by the reader; for example, I could not access it), the second one does not provide sufficient information.
A13: The links are drawn between nodes that are geometrically close each other so that the graph which emerges has a topological valence: it declares that the protein works like a whole or fully connected system. Concerning the comments about refs. [9] and [24]: we added refs. [14,25] which also contain many details on the model.
The electrical features are collected in Eqs. 1-3, the choice of using the same value of resistivity for all the links is a simplified assumption, since no information is available about amino acid resistivity. Furthermore, the present model correlates the sensing effect due to the conformational change directly with the change of the geometrical structure of the single protein which is a characteristic associated with the PDBs available from the literature, When some information will become available (as it happens for nucleobases) a different choice can be followed. A schematic flowchart of calculations is added
R14:It is also written that the role of the medium is “only indirectly accounted for” and that “its effects are resumed in the protein conformation”. How is this accomplished?
A14: In our speculation, the medium, identified with the solution in which the protein is merged, is the main responsible of the protein conformational change, in particular, the transformation that the pH variation from 6.5 to 10 in dark produces on the structure is much larger than that induced by illumination at each pH value (in figure 6, this is observed for the dark-light structure couple at pH 6.5) In the revised version, the sentence has been slightly modified to make it more understandable.
A nonnegligible part of the philosophy behind the model resembles that of the tunneling pathways model, although, at difference from the latter, the analysis is not at an atomistic level and the links are electrically characterized. Indeed, the two models share the problem that the charge (which is characterized by a wave function) does not really transfer this way, and sometimes not even by far so, in biomolecular systems. Yet, the approximation is discussed (and partly mitigated by introducing interferences among tunneling pathways) in the tunneling model. How is the approximation justified or its approximation defined in the present model?
A15: We agree with the referee that our model has many contact points with the pathways model, often citated in previous papers. Otherwise it is only a conceptual analogy which enforces our approach, which is , in some sense, more phenomenological. We start from the observation that the protein is a connected system based on interactions among its components, amino acids at this level of refinement. Each of these interactions is assumed to be a channel of charge transfer. just like is made, for example, when modeling the electrical properties of axons. This connected system is analyzed in terms of a classical linear impedance network that is solved using the Kirchhoff laws plus, in the case of a non-linear I-V characteristics, a stochastic procedure which mimics the tunneling mechanism of charge transfer between two nodes.
Therefore, no sign of the wav-particle dualism is present in the model, instead, the network of connections is treated like an electrical network in which the transfer of electron is simply described in terms of impedances. The introduction of a tunneling pathways model is made at a phenomenological stochastic-level to interpret the super linear I-V characteristics, following previous results obtained for the case of bacteriorhodopsin (bR), where the experimental I-V curves indicated the presence of tunneling mechanisms (including direct and Fowler-Nordheim types). Here we applied the same approach, even if the deviation from linearity is weaker than the case of bR because of the lower range of applied voltages values.
Page 5, lines 199-200
The notation 1/0, instead of {1,0}, is misleading in my opinion.
A15: the bracket notation has been introduced in the revised version
The meaning of “the presence or any of a link” is unclear.
A16: the sentence has been modified
Page 7, line 262
As above, “static” should be replaced by “steady-state”.
A17: ok
R18a Pages 7-8
D is a model parameter. From Figures 3 and 4, as well as their discussion, one cannot deduce which value of D better describes the system. The fact that D = 10.5 produces the most marked dependence of the resistance on the pH does not mean that this is the right value to be used.
A18a We agree, in principle it is right. On the other side, the analysis of the relative resistance as a function of D shown in Figs 2, 3 is meaningful because it reveals both a trend, which is generally more significant than a single point, and the critical distance which magnifies the differences. This is the D value that here and in previous investigations we used for the analysis. Comparison with experiments confirm our choice is correct, i.e. it finely reproduce them.
R18b:The resistance dependence on pH could be smaller than that obtained for D = 10.5 and better correspond to another D value. In other words, the model should provide a recipe for choosing an optimal D value either based on the physical grounds or based on a suitable comparison with experiments. The mentioned “resolution” has no meaning in my view, as one wants to see the pH dependence of the resistance R and not what parameter of the model would provide the largest change in R, irrespective of whether this is the real one. Such resolution might be meaningful only at a very qualitative level, if one assumes that the pH influences the charge conduction and wants to see if a pH change determines an increase or decrease in R. However, even in this case, it is difficult to understand how much it is real or just an artefact of the model. The comparison with experiments could help, but it is essentially missing in this part of the manuscript. The sentence “Also in this case, the maximal resolution shows up for a cut-off value D=10.5 Å, with a reduction of resistance in fair agreement with experiments [12]” is by far insufficient.
R18b.We appreciate the constructive criticism of the referee: the term “resolution” is used in the flattest meaning of “the smallest difference that can be detected within the uncertainty associated with the numerical uncertainty”. Now, for each structure considered in Table 1 we calculate the resistance associated with the corresponding network obtained from the given PDB as function of the cut-off distance D. The values of the resistance so found is by definition a property of the structure for the given D. Then Figures 3 and 4 report the ratio of the resistances pertaining to two networks that differ only from their PDB as function of D. Accordingly, Figs 3 and 4 provide the information about the resolution in the sensing of the pH or of the light. In particular for D= 10.5 A it is obtained the maximum sensitivity. At present only qualitative agreement with experiments of Ref. 12 is claimed. When the experiments will become quantitative, then we could discuss on the value of D that better interprets experiments. We already carried out in the recent past this kind of investigation for the case of several proteins, as reported in the references.
R19.Page 9, line 337
The term “writes” should be changed to “reads”.
A19: ok
R20 Page 10
R20 The Authors should comment on whether and why the pH influences shown in Figure 6 are reliable based on the accuracy of the method. What would happen by using another D value?
A20: By changing the value of D the network changes, because it become more/less connected and, in turn, this changes the maximal ohmic resistance (see eq.1 with zero frequency) and accordingly shifts the I-V characteristics . In the comparison among different pH values, you can tune to zero these differences or magnify them according to Figs 3 an 4.
R21 Pages 10-11
The observed dependence of charge transfer on the energy barrier is nothing new or surprising. Inspecting the current in dark conditions for the different barrier values would have been necessary to safely assert what is written at the end of page 10. For example, how do I know that the onset of higher current for Phi = 0.206 eV does not occur in a significantly different way in dark?
A21: We agree that it not surprising that the reduction of an energy barrier allows a larger charge transfer, but this is not the point. The point, instead, is that even a very small reduction of the energy barrier leads to a significant increase of charge transfer probably due to the specific parameters of the model. A sentence has been added to highlight this point.
R22:Pages 11-12
The choice of rho looks just empirical: just to fit the experimental curves in Figure 8. It should be motivated (at least as for the ordering of the values) based on the kinds of experimental setups. I can think of other (very different) models that can reproduce the experimental curves in Figure 8. So, it is difficult for me to accept Figure 8 as a demonstration of the appropriateness of the method, the way it is presented. In fact, the conclusions on page 11 are not so clear to me and partly do not look ok for what I can understand from the text (for example, the match between Fermi level and frontier orbital energy should increase the current indeed).
A22: We agree that different models may be designed to fit the same set of data. In the past, this approach has given satisfactory results, but we do not pretend to be exhaustive and the criticism of the referee is part of the game.
The value of rho_max was chosen to fit the experiments at the lowest fields, linear or Ohmic region, while the rho_min value and the barrier height were chosen to fit the super-linear region of the I-V characteristics. These three parameters represent the minimum number able to fit experiments.
We also thanks the referee for noticing the oversights concerning the “reduced” conductivity. It has been fixed in the revised version.
Reviewer 2 Report
1) The abstract should be improved, highlighting the relevance of the research 2) The authors should describe the equipment and software used both for the characterization of proteins and for the study of properties with changes in pH 3) For curves I and V Were any diode preparation done? What setup did you have? What is the relevance of the studies? 4) There should be a section of conclusions that describes the relevance of the research according to the results obtained 5) The discussion section should be more extensive or, failing that, results and discussion section should be included
Author Response
Answers to REFEREE2
The authors acknowledge the constructive comments of the Referee 1, in the following the answers to her/his criticisms
R1 The abstract should be improved, highlighting the relevance of the research 2) The authors should describe the equipment and software used both for the characterization of proteins and for the study of properties with changes in pH 3) For curves I and V Were any diode preparation done? What setup did you have? What is the relevance of the studies? 4) There should be a section of conclusions that describes the relevance of the research according to the results obtained 5) The discussion section should be more extensive or, failing that, results and discussion section should be included
A1: Abstract has been extended accordingly with Referee’s comment. We reported major comments and discussions in the section Discussion and conclusions which was extended as proposed by the Referee.
Reviewer 3 Report
The authors chose the reaction center of Rhodobacter Sphaeroides, a photoactive protein to investigate the effect of pH on charge transport properties. Increasing the pH, the protein deforms from bottom to up, thus becoming more tight toward the periplasmic side. The protein undergoes a resistance reduction when the pH changes from weak acid (pH = 6.5) to alkaline conditions. I recommend to be published after minor revision listed below:
- How does the photon activate an electron transfer in protein cell ? Please present a diagram of excited electrons transfer ?
- The manuscript discusses photosensitive proteins. The effect of pH on charge transport of proteins should combine photon.
- P113, the presentation of “temperature of 100 °K” had better revised as “temperature of 100 oC” or “temperature of 100 K”.
Comments for author File: 
Comments.doc
Author Response
Answers to REFEREE3
The authors acknowledge the constructive comments of the Referee 1, in the following the answers to her/his criticisms
The authors chose the reaction center of Rhodobacter Sphaeroides, a photoactive protein to investigate the effect of pH on charge transport properties. Increasing the pH, the protein deforms from bottom to up, thus becoming more tight toward the periplasmic side. The protein undergoes a resistance reduction when the pH changes from weak acid (pH = 6.5) to alkaline conditions. I recommend to be published after minor revision listed below:
- How does the photon activate an electron transfer in protein cell ? Please present a diagram of excited electrons transfer ?
A1: The process of electron transfer in the Reaction Center is widely present in the literature, see, for example, ref[7] and also Wikipedia https://en.wikipedia.org/wiki/Photosynthetic_reaction_centre
The model does not describe the mechanism of e-phonon interaction, rather it analyzes the effects that a change of the pH and also illumination conditions produce on the protein structure and, in turn, on its electrical response.
The manuscript discusses photosensitive proteins. The effect of pH on charge transport of proteins should combine photon.
A2: Not necessarily the effect of pH on charge transport is related to the presence of light. As for the A1 the protein in the native state can undergo a conformational change because of a change of the pH of the crystal solution or of the presence of light, or of both pH and light. In all the cases the electric response of the protein can be modified. Within the present model we evaluate this change of the electrical properties starting from the knowledge of the protein tertiary structure provided by the PDB in all the conditions.
A2. P113, the presentation of “temperature of 100 °K” had better revised as “temperature of 100 oC” or “temperature of 100 K”.
A3: OK –
Round 2
Reviewer 2 Report
I have verified the corrections made to the article.
