Cyclic 5-membered disulfides are not selective substrates of thioredoxin reductase, but are opened nonspecifically

The cyclic five-membered disulfide 1,2-dithiolane has been widely used in chemical biology and in redox probes. Contradictory reports have described it either as nonspecifically reduced in cells, or else as a highly specific substrate for thioredoxin reductase (TrxR). Here we show that 1,2-dithiolane probes, such as “TRFS” probes, are nonspecifically reduced by thiol reductants and redox-active proteins, and their cellular performance is barely affected by TrxR inhibition or knockout. Therefore, results of cellular imaging or inhibitor screening using 1,2-dithiolanes should not be interpreted as reflecting TrxR activity, and previous studies may need re-evaluation. To understand 1,2-dithiolanes’ complex behaviour, probe localisation, environment-dependent fluorescence, reduction-independent ring-opening polymerisation, and thiol-dependent cellular uptake must all be considered; particular caution is needed when co-applying thiophilic inhibitors. We present a general approach controlling against assay misinterpretation with reducible probes, to ensure future TrxR-targeted designs are robustly evaluated for selectivity, and to better orient future research.

However, the fluorescence microscopy ( Figure 4c), flow-cytometry (Figure 4d), and the fluorescence imaging of zebrafish (Figure 4e) don't add much to the story and may even be a distraction to some readers. Additionally, the animal study is complicated because of the expression pattern of TrxR1 at various life stages of the animal. Moreover, the determination of the mechanism by which AF inhibits the interaction between TrxR1 and the probe is critical and will add great value to this work. Whether or not AF's binding to membrane thiols actually affects the cellular entrance of the probe can be investigated by designing a similar probe in which the fluorophore is always turned on and remains attached to the 1,2dithiolane moiety irrespective of its redox status. Also, the author should consider the possibility that when the probe enters the cell and gets reduced to the thiol form, it can react in the reduced form with AF, which, as the author mentioned, is a "potent thiol reactive species". Expectedly, the progress of the reaction can be monitored using mass spectrometry or some other method.
(3) We also introduce new comparisons between several probes to show how assay interpretation requires combining independent experiments in order to advance; these settle any remaining questions about our statements in regard to previously published reports. For example: (3a) 1,2-dithiolane SS50-PQ is unaffected by cellular TrxR knockout (new Fig 5, new Fig S8): so it is not significantly cellularly activated by TrxR: yet its signal is mildly suppressed by cellular treatment with chalcophilic S N Ar-based electrophiles such as TRi-1 and TRi-3 (new Fig 5), and strongly suppressed by cellular treatment with the lipophilic Au (I)-based Lewis acid auranofin (AF), which has more than 20 attested targets (new Fig S10-15). The conclusion is that treating cells with these thiol/selenol-affine electrophiles and observing this suppression of cellular signal from a dithiolane probe, cannot be cited as a proof that that 1,2-dithiolane-based probe is a selective reporter of TrxR, since SS50-PQ provides a clear counterexample. We also must take into account e.g. Matile's studies, that showed that 1,2-dithiolanes substantially rely on free exofacial thiols for cellular uptake. This uptake is inhibited by general thiol-reactive species of all tested chemotypes (including even other disulfides) typically even suppressing uptake to only 10% of normal. This provides a coherent explanation that matches classical organic chemistry: these lipophilic electrophiles can suppress cellular 1,2-dithiolane signal generation by reacting with thiols on the cell surface, so substantially blocking the otherwise strain-promoted enhanced cellular uptake that dithiolanes can experience. (3b) Now we examine the electrophile treatment data for 1,2-dithiolane TRFS-green. This shows almost identical signal suppression as SS50-PQ (new Fig 5, new Fig S11). We have excluded that the electrophile assay tests TrxR selectivity: but, supported by literature, we would expect that these electrophiles should very similarly inhibit strain-promoted cellular uptake of the dithiolane TRFS-green as for the dithiolane SS50-PQ; and the observation that the level of inhibition is so similar between SS50-PQ and TRFS-green again suggests that their chemical behaviour (rooted in their 1,2-dithiolane) is the same, i.e. that the signal of TRFS-green is likewise being inhibited by electrophilic blocking of cellular thiols, not from the effects those electrophiles also have by partially reacting with cellular TrxR (which we stress, is only one of their many cellular targets). (3c) To counter-test these results, we now bring in the linear disulfide probe SS00-PQ. This probe should not benefit from strain-promoted thiol-mediated uptake enhancement (which can be suppressed by electrophiles); and we as well as others have shown that linear disulfide probes are not selective for any particular cellular reductant -while they are reducible by TrxR, they are also reducible by the vastly more concentrated GSH, Trx, Grx, etc. The electrophile assay results are perfectly coherent with this expectation (new Fig 5, new Fig S12): the electrophiles do not suppress signal from SS00-PQ, since the signal from this non-dithiolane probe is not limited by strain-promoted uptake.
This new intercomparison of results across multiple probes before reaching conclusions (new Fig S13) is, we feel, a strength of our paper and one which was not present in previously published reports; and given the unity of results between our releasing probe SS50-PQ and the alternative releasing design TRFS-green, we consider that this delivers solid evidence for the generality of the nonselectivity and effects that we report.
Secondly, we now supply cell-free experiments with the non-releasing 1,2-dithiolane-based probe Fast-TRFS (Fang et al. 2019Nat. Commun., 10.1038. We reached important conclusions only by our comparison to its novel linear disulfide analogue "Linear-TRFS" of which we also give the first report and characterisation (we include much of the following logic as expanded discussion in Supplementary Note 3). Fast-TRFS was intended to be fluorescence-quenched in its oxidised cyclic 1,2-dithiolane state, and was reported to be "a specific and superfast fluorogenic probe of mammalian thioredoxin reductase" that should act in cells by increasing in fluorescence upon "reduction [to Dithiol-TRFS]". We now show that there are at least three hitherto unreported confounding aspects to this Fast-TRFS probe system, and discuss how these combine with each other to give misleading results: and in doing so we support our thesis that 1,2-dithiolane is not a viable redox sensor for selective probes.
Firstly, as we discuss in relation to new Fig 6, in our hands, Fast-TRFS performs with poor reproducibility that matches entirely to Whitesides' characterisation of polymerisation-prone dithiolanes, and to the observations we report about polymerisation-prone SS50-PQ and its precursors.
Secondly, we show that a ring-opened species resulting from strain-promoted thiol attack on the Fast-TRFS 1,2-dithiolane [model compound Linear-TRFS] will be just as fluorescent, as the fully-reduced Dithiol-TRFS (e.g. on the dose-response panel in Fig 6d, at 9 mg/mL lecithin, ca. 36 fluorescence units). This means that any of several likely TrxR-independent ring-opening mechanisms will generate TrxRindependent fluorescence signal from Fast-TRFS (Fig 6, Fig S16). These mechanisms include but are not limited to: (a) strain-promoted oligomerisation, that should be particularly relevant if high local concentrations of Fast-TRFS are created, resulting in all ring-opened probe molecules becoming fluorescent (see below and shown in Fig 6c); (b) strain-promoted cellular uptake is via the known mechanism of exofacial thiol attack on the dithiolane and so will likewise activate fluorescence before it ever reaches the intracellular environment; (c) opening of the dithiolane by any cellular thiol.
Thirdly, we show that the fluorescence signal from Fast-TRFS and from its oligomerisation and its reduction products, also have strong environment dependence which creates additional problems. For example, oxidised Fast-TRFS fluorescence is instantly ca. 3-fold enhanced just by moving from aqueous to apolar environment (at the first minute of Fig 6d, the fluorescence intensity without vesicles is 0.99, but rises to ca. 3.4 as the lipid concentration increases -i.e. already 1/10 of the total potential fluorescence signal)and this is before non-reductive strain-promoted ring-opening has significantly taken place (which it then does, reaching 50% of completion within the next 29 minutes). We also show that the environmentdependence of the Fast-TRFS system entirely rules the fluorescence behaviour of the model compound Linear-TRFS, which is up to 17-fold enhanced just by exposure to lipids (Fig 6d: from 2.0 to 35) but is almost entirely unaffected by reduction (irrespective of which environment the probe is in -see TCEP spike at the end of its incubation, Fig S16). Therefore, the faster that a Fast-TRFS probe of any redox status accumulates into (intracellular) lipid environments, the more quickly that the cellular fluorescence will rise.
The combination of these three effects provided us a unified model rationalising the published "superfast" fluorescence turn-on of Fast-TRFS, depending on the situation under testing.
We propose that totally TrxR-independent, but simply local-concentration-dependent ring-openingoligomerisation can be activated by partitioning the 1,2-dithiolane Fast-TRFS from a relatively large volume of aqueous medium into a relatively smaller volume of apolar environment (ideally, of high surface area for rapidity). Then Fast-TRFS fluorescence would rapidly rise as increasing amounts of the more fluorescent oligomer would be formed. We predicted that this could allow even maximal theoretical fluorescence signal to be reached either with catalytic, or even zero, added reductants. We tested this experimentally with a series of vesicle assays (Fig 6 and Fig S16), simply sonicating commercial soybean lecithin (Sigma, P5638, ca. 60% phospholipids and 35% oils; zero content of TrxR or NADPH) diluted to <1%wt in distilled water to form the lipid vesicles, and applying Fast-TRFS to various concentrations of these vesicles. Surpassing our expectations, in just one hour, the Fast-TRFS signal in 0.9%wt vesicle mixture without any reductants reached full maximal signal plateau, corresponding to that seen with full TCEP reduction to Dithiol-TRFS. This can only be understood as local-concentration-dependent oligomerisation to the poly-(disulfido-TRFS), that is just as fluorescent as Linear-TRFS or Dithiol-TRFS when they are compared in the vesicle system (Fig S16). This shows the confounding influence of redox-independent, strain-promoted processes on 1,2-dithiolane chemical behaviour: and illustrates how the environment-dependent readout of the Fast-TRFS compounds expands this complexity by generating a false positive signal.
Thirdly, as the aminonaphthilimide core of TRFS-green is also a classic environment-dependent fluorophore, and since its slow anilide elimination suggested that it could only perform fluorogenically in cells by harnessing environment-dependent effects that might well be independent of reduction (new discussions at Fig S3) we looked in the literature to see if environment-dependency might be shown as a similarly confounding issue elsewhere. We now added a discussion (Supplementary Note 4) to this effect. For example, we note that recently published results from the Fang group (10.1021/acssensors.1c00049) show that their naphthilimide-based compound S1, which is similar to TRFS-green, experiences an instantaneous 45-fold enhancement of fluorescence intensity upon leaving all-aqueous environment and noncovalently associating to albumin in a cell-free experiment (their Figure 3; note too that their aminocoumarin compound S3, which is similar to Fast-TRFS, has a ca. 4-fold fluorescence enhancement in the same experiment, matching our vesicle results with Fast-TRFS).
Paralleling our experimental investigation of TrxR-independent signal generation in the Fast-TRFS system, this literature report of environment-dependent-signal highlights how interpreting fluorescence increases with the TRFS-green probe may similarly not be straightforward since (1) in any context, even just leaving the aqueous (extracellular) environment will trigger a fluorescence increase, that can be entirely independent of any reaction on its 1,2-dithiolane motif; (2) in the cellular context, all TrxR-independent strainpromoted thiol-mediated uptake at the cell surface will covalently associate the TRFS-green probe onto membrane proteins and into membranes, thereby giving cell-uptake-driven signal independent of molecular encounter of the TRFS-green probe with intracellular TrxR, let alone cyclisation-driven release of the cargo. It is therefore consistent with these hypotheses that the same factors as we advanced for the 1,2-dithiolane Fast-TRFS (concentration into membranes aided by membrane-thiol-based opening of the strained dithiolane, which gives an environment-dependent signal turn-on) will apply to TRFS-green, permitting it generate signal based on cellular exofacial thiol status, without even encountering TrxR. These suggestions should be considered in light of the demonstration (see above) that cellular signal from TRFS-green is manifestly independent of TrxR.
In conclusion: through the two new time-intensive biochemical experimental series with TRFS-green and Fast-TRFS, we consider that we have now gone far enough beyond our original scope of demonstrating the general liabilities of 1,2-dithiolane, fully responding to reviewer recommendations to illuminate how such entirely general chemical problems of 1,2-dithiolane can be amplified, according to chemical design, into strongly confounded readouts that may easily give misleading chemical interpretations. In doing so, we have tackled both cargo-releasing as well as non-cargo-releasing 1,2-dithiolane probe types. We feel that these results build in one paper, a larger and more convincing body of evidence than has hitherto been accessible even by combining multiple literature sources, and which matches both chemical logic as well as prior literature in a manner we feel is both convincing and coherent. There are still other literature-reported 1,2-dithiolane probes and even prodrugs, but our aim is to talk about strained disulfide 1,2-dithiolane probes in general, and not about specific probes or reports; and as we believe this now builds a very strong general case we consider this the right moment to stop.
Please note: "the authors (apparently?) just hypothesized this" refers to Fang et al., not to our paper. Given the greater emphasis we introduced on Fast-TRFS and TRFS-green and on cellular studies, our treatment of catalytic, cell-and-enzyme-free, auranofin-mediated signal suppression by strain-promoted oligomerisation has had to take a smaller position in the revised manuscript. Nothing has been removed, and many months of experiments have been added (new Fig S9-S15 and Supplementary Note 2: Auranofin): but we have consolidated all this data, discussion and references in the Supporting Information, with only a short summary and link to it in the Main Text.
We have tested the catalytic suppression of reducibility in several additional confirmatory experimental rounds, which supported our previous hypothesis that AF treatment can likely catalyse forming nonreducible, likely aggregated/precipitated oligo-SS50-PQ (new Fig S9). We appreciate the referee's interest in AF and its role in previous studies; we also think this is an engaging feature that we can bring to the community's attention. However, we cannot go assay-by-assay-style through previous studies or speculate about the controls they chose. What we instead did is to run the cell-free AF assay on unstrained linear SS00-PQ and unstrained cyclic SS66C-PQ as comparison species, showing as expected that there was no inhibition of signal generation for either of these compounds, which supports the strain-promoted mechanism (new Fig S9a). With this added result, we feel that our expanded discussion, particularly in light of the extensive generalised investigations into the previously published 1,2-dithiolane probes detailed above, provides sufficient mechanistic insight to demonstrate a convincing case to the reader about the actual performance of our and previously published probes regardless of what controls were or were not used in other works; and to draw a line under it we refer the reader to the further references related to AF-phosphine dissociation (Supplementary Note 2), which we hope addresses this point satisfactorily.
3. There are other published TrxR probes that have other seemingly nonspecific functional groups such as linear disulfides, including relatively simple commercial kits (Ellman's reagent?). Why would these other reagents not be commented upon by the authors herein?
We didn't comment, because our scope is showing the unsuitability of 1,2-dithiolane in cells. Therefore, our literature summary treatment (Fig S2) focuses on 1,2-dithiolane probes. We are also very interested developing our own novel, robust, cellularly selective probes, such as the RX1 probe for TrxR (10.33774/chemrxiv-2021-52kwx); but the proper place to deal with alternative TrxR probes is in that paper not here. Also: with a cap at 70 references, it would not be possible to tackle even a fraction of the different systems that have been claimed as relating to TrxR in this paper. And as most in the field are aware, many of those compounds like Ellman's reagent are clearly nonsense in the cellular context too -but it would not bring our probe research forward to discuss them here. Still, we appreciate the suggestion; we ask the referee's understanding that this paper is not the right place to lose focus on 1,2-dithiolane.
We thank Reviewer 1 for their close reading and carefully considered opinions on the paper, which challenged and encouraged us to dive a couple of levels deeper into 1,2-dithiolane biochemistry, and which we are sure now makes for a very much stronger paper.
Reviewer #2: The goal of this work was to verify whether the 1,2-dithiolane-based probes, claimed to be specific inhibitors of mammalian thioredoxin reductase, are indeed specific. To this end, the authors have designed a new probe based on previous work. In vitro assays clearly shows that monothiols such as cysteine (Cys), Nacetylcysteine (NAC), N,N-dimethyl-cysteamine (MEDA), and cysteamine (CA) and the reductive enzymatic systems -Trx/TrxR and Grx/GSH/GR are excellent reductant of the 1,2-dithiolane moiety and generate rapid fluorescence response. In other words, the probe is not specific.
We agree, but, particularly given the new results shown in Fig 5-6, we would rather rephrase it that the 1,2-dithiolane motif itself is not selective, and that the probes we have tested this motif in, merely reveal this nonselectivity. We are more consistent about this now in the text. Additionally, we have now expanded our 1,2-dithiolane assessment with TRFS-green and Fast-TRFS, as well as key non-dithiolane comparison probes RX1, SS00-PQ, Linear-TRFS, and SS66C-PQ (see major discussions above). Through this combination of tests we are now sure that we demonstrate the general unsuitability of 1,2-dithiolane for "enzyme-specific" interpretation in the cellular context. We feel that these experiments showing the quality of our PQ probe performance are a central part of our story (responding to goal b: suitability of general probe design). We have now rephrased passages to make this clearer. Our scope has been to show the non-selectivity of 1,2-dithiolanes (goal a), by using an unimpeachably robust and easily-interpreted environment-independent probe system that can be extended to any arbitrary redox sensor trigger units, and can deliver valuable information by FACS, live animal imaging thanks to cellular retention, etc (goal b). We now clarified this in the text, also by citing our two more recent papers that rely on the cellular and in vivo performance that we have established and benchmarked in this work (10.1021/jacs.1c03234, 10.33774/chemrxiv-2021-52kwx). We did not intend the animal to complicate anything -we now clarified that since the 1,2-dithiolane probes are anyway not selective for any one reductant, the signal they provide cannot and absolutely should not be interpreted as being related to TrxR1 (which is expressed and active inside all somatic cells since it is needed for Trx maintenance and for DNA synthesis, etc). With these changes we feel that the images and proof of concept animal application clearly form a part of the coherent investigation.
Moreover, the determination of the mechanism by which AF inhibits the interaction between TrxR1 and the probe is critical and will add great value to this work. Whether or not AF's binding to membrane thiols actually affects the cellular entrance of the probe can be investigated by designing a similar probe in which the fluorophore is always turned on and remains attached to the 1,2-dithiolane moiety irrespective of its redox status. Also, the author should consider the possibility that when the probe enters the cell and gets reduced to the thiol form, it can react in the reduced form with AF, which, as the author mentioned, is a "potent thiol-reactive species". Expectedly, the progress of the reaction can be monitored using mass spectrometry or some other method.
We thank the reviewer for highlighting this. Indeed, we feel that bringing this incompatibility of AF with reducible strained disulfide probes to the attention of the redox community is an important feature arising from our work: even though as we wrote above, we have had to concentrate all the AF experiments into the Supporting Information to cope with the scope expansion in the Main Text.
We feel that now with the Supporting Information treatment of the AF experiments, expanded to stretch over Fig S9-S15 and the expanded accompanying Supporting Note 2, this makes a self-complete demonstration to alert the community to this problem. We also now cite the relevant paper (10.1002/ anie.201502358) at several relevant junctures, since it has employed exactly such a compound as the referee proposed (permanently-fluorescent xanthene attached to 1,2-dithiolanes: compounds 3 and 4 in that paper) for a similar purpose, and demonstrated similar dependency of net cellular fluorescence upon the redox status of exofacial thiols (which requires assuming a thiol-dependent uptake mechanism to be the ruling feature of cell entry for 1,2-dithiolanes, which in turn suggests just as we do, that 1,2-dithiolane cannot be interpreted as a TrxR-selective motif). We think in these ways that we have highlighted, and proposed rational explanations for, key problematic features that have never been raised before in this community. For example, the ligand exchange that permits the AF phosphine to initiate reduction of one dithiolane, which had been observed in coordination chemistry work, but had never been cited as an issue for redox biology (Supporting Note 2).
Like the reviewer we are interested to know what AF does. However, we do think that resolving the spectrum of things AF actually does, and under what circumstances, is a daunting topic even split over many research groups -it is no accident that after thousands of papers using it, a typical summary of AF's effects remains "we still do not understand the scope and consequences of AF's cellular targets". AF is known to be a dirty drug, and the sooner our group stops investigating its drawbacks to focus exclusively on reducible probes that are used in clean assays and with clean inhibition strategies, the better. Deceptively simple though such an experiment might appear, the complexity of the cellular membrane composition, lack of good species-resolved analysis techniques, and especially the lack of a good simplified in vitro model make this a tricky challenge. Furthermore, we only treat auranofin as an example of one way in which lack of cross-disciplinary communication and appropriate controls, has probably misled previous investigatorsand we have to orient our investigations around 1,2-dithiolane in this paper.
Regarding probe-AF interaction inside cells, we were pleased to get this remark. We now added to our Supporting Information discussion (end of Supporting Note 2) that we think that the intracellular background thiol concentration (5 mM GSH, up to 50 mM protein thiols) is so much higher than the instantaneous reduced intracellular probe reduction dithiol intermediate is ever likely to be (particularly for the rapidly-cyclising SS50-PQ), that since we see no basis for preferential AF reaction of the reduced dithiol intermediate compared to any other monothiol, we do not imagine that intracellular reaction can be significant to suppression of signal generation.
In conclusion, by bringing this, as well as the polypharmacological targets of AF to light in this community, we feel that we have delivered substantial advances that will suitably caution or inspire redox biologists in future studies. As we feel that our several figures of AF results are already proportionally of greater weight than they should be given that our paper focuses on 1,2-dithiolane chemistry and on probe performance criteria, we fear that diving more into the polypharmacological AF's effects would explode the scope of this paper and distract from the chemical achievements that we have anyway shown: nonselectivity of reductants, and non-reduction-based signal generation, in cellular and in cell-free tests. Therefore, we trust that the revised manuscript now addresses these issues at a suitable level for publication. -New Figure S1 has been provided (summary overview), to assist the reader in comparing the results in this work which pertain to TrxR selectivity or lack thereof.

Reviewer #1 (Remarks to the Author):
This revised manuscript includes additional experiments to augment the authors' claims that certain cyclic 1,2-dithiolanes in general, and the known dithiolane TRFS probes specifically, are not selective for TrxR and should not be used as probes for TrxR. The paper is focused on proving significant interpretations of prior, peer-reviewed work are wrong.
To address the prior review comments, the authors synthesized and studied TRFS probes that they are criticizing. This is commendable. However, the newly reported findings are still too speculative for consideration for publication in a top-tier journal.
For example, one proposal is that the cyclic dithiolanes polymerize and that this causes nonspecific fluorescence that interferes with TrxR signaling. However, the experiments were not performed in cells containing TrxR, and, though a signal was generated in vesicles, there is no direct proof of any polymerization event given, despite the experiment in (more tractable) artificial media. The knockout experiments show that a potentially troubling non-specific signal is formed by the TRFS probe ion cell media. However, without the presence of TrxR, this does not take into account, or directly demonstrate, that the cellular reaction with TrxR may have superior kinetics compared to whatever mechanism is causing the signal under knockout conditions. While the authors' suppositions may have validity, that reactivity with other biomolecules apart from TrxR does not allow the TRFS and related probes to function properly, they have not provided strong enough evidence to make such claims.
In the prior review, the suggestion was made to consider an analogy with covalent drugs. I will be more specific here. Fosfomycin, for example, contains an electrophilic epoxide. Worse, Zanubrutinib contains an acrylate moiety! Using the authors' rationale, none of these (or the numerous other covalent inhibitors that are now major prescribed pharmaceuticals), should have ever been marketed, as it is so easy to show that their functionality exhibits well-known covalent reaction promiscuity, that these authors would argue ensures that none should never exist long enough to reach the desired protein target.
One could argue, as the researchers here have done, for example, via extensive chemical literature citations about basic chemistry of such functionality, and one could readily design several experiments in non-natural model conditions, to "prove" that such drugs do not perform as claimed, because they react with other biomolecules besides their targets.
Finally, the paper and the responses are relatively dense. This is not meant as a criticism of the paper or the writing style. The broader issue is that a non-specialist reader of a multidisciplinary journal can likely be overwhelmed by the extensive arguments and citations that can obscure the lack of substantive experimental proof and the lack of a balanced assessment of the prior probe modalities criticized herein.
In summary, the authors afford ample proof that the compounds and functional groups they are questioning are indeed reactive with species besides TrxR. However, this is already a well-known fact, that these functional groups are highly reactive. Moreover, they have not shown compelling, direct substantive evidence to overturn peer-reviewed results showing the efficacy of the TRFS and related TrxR probes.
I still have some concerns, however, about the sometimes overly-speculative nature (relevance) of some, not all, of their conclusions, as stated in my prior review(s). Unfortunately, the authors have misrepresented some points made previously: It is not a choice of mine to believe certain facts and not others, (see "rebuttal" file, "null hypothesis" comment). The dithiolane probes are indeed generally chemically non-selective. Anyone who does not see that is scientifically blind. I never doubted that. I never claimed that in the prior review cycles.
The point in my critique(s) was whether they are non-selective enough under the exact same conditions (to be fair to the prior research groups), used and reported by the prior researchers whose work was being questioned. Too many papers challenge prior art but only after working in a different experimental context than was initially intended or used by the original researchers, and inappropriate conclusions are made. Moreover, no probes or drugs have perfect selectivity. Selectivity is condition and matrix dependent, among many other factors, especially in biological milieu, of course. It is with these issues in mind that I continue to address/have to clarify the points made previously.
To challenge (especially prior peer-reviewed) work without repeating the prior work, and in a manner that is very clearly performed and presented as a very rigorous replicate to the reader is not preferable.
For example, I was very happy to see the authors, after the first review round, decide to actually synthesize and use the same probe as was reported previously by those they are questioning. This went a long way in validating comparisons. I had suggested this in my initial review.
However, instead, the rebuttal claims "SS50PQ and TRFSgreen share the same reductable motif.......shows same results......had been doubted by the reviewer." To use just two probes to conclude that only the dithiolane group, and not the protein etc binding of the rest of the probes, is significant for selectivity, was a reach.
In addition, there was only doubt because the authors had tried to discount the prior work via initially making and using only a different probe than the earlier researchers had used. This was a suggestion of mine to only help strengthen the paper (i.e., to use an actual TRFS probe as a control), which it indeed did, and not due to "doubt" as has been attributed here. I did not discount the fact that the probe lit up in the knockout case. I did not misinterpret this as stated in the "rebuttal". What I stated as a caveat, to add rigor, is that one must consider that the potentially interfering optical turn-on effects might not be kinetically competitive with the TrxR-derived signal. It is not due to some misunderstanding on my part. If that point is already in the paper or the rebuttal, it is not clear to me. Bottom line: one can't claim precisely a "60-100 % non-selectivity" without such a caveat.
"Perhaps this compelling evidence has caused a shock" is neither a helpful nor a professional comment. In fact, I noted in my critique that very many issues about selectivity of dithiolanes were already known (and published). That this fact was needed for background, but not new at all new and not surprising (please see my actual written critique to accurately represent and respond to my suggestions/concerns, thank you).
Again, the fact that the non-selectivity of this functional group is already well-known under other conditions in the prior art, was never the main issue for me.
The problem is that the exact experimental context of the TRFS etc dithiolane probe papers being criticized were not accounted for in a clear enough manner in the submission. Instead, carefully obtained, but general evidence for non-selectivity has been obtained and included.
Despite the current claims that the paper is not a critique of the prior work of others, that's exactly what it is (and was previously), but its at least good that the authors now at the very least try to claim that its only about the dithiolane.
Again, to be really clear a main point is that the authors need to show that, under the exact conditions of the prior art probe studies, very clear proof that they could not repeat the prior results and come up with the same conclusions.
This was only partially demonstrated, and/or not very clearly described in the manuscript. Now, however, the authors of the original prior art have shown that their prior dithiolane probe is indeed non-selective, and that their conclusions were wrong, under their conditions. So it turns out that the authors of this manuscript were correct, though the way they "reproduced" (or didnt reproduce) the prior art using dithiolane probes, (much beyond making the same probe, but only did so upon the first revision), to make their point was, from the beginning, in question.
Apologies for the repetitive nature of aspects of this review, but I am doing my very best to be clear and understood here, since ts clear to me after reading the rebuttal that several issues raised previously were neither understood nor even carefully read. And I am still not certain of this, especially upon seeing the tone of some of the rebuttal comments.
However, I sincerely hope the comments (which are largely clarifications of my prior critiques) are helpful to the authors.
Taking all of the aforementioned into account as a whole, I overall support publication of this paper with minor concerns. I congratulate the authors for an excellent and courageous paper.
Reviewer #3 was asked to act as an arbitrator and see if reviewer #1's concerns have been addressed by the authors. They made comments to the editor and support publication.