From amino acid mixtures to peptides in liquid sulphur dioxide on early Earth

The formation of peptide bonds is one of the most important biochemical reaction steps. Without the development of structurally and catalytically active polymers, there would be no life on our planet. However, the formation of large, complex oligomer systems is prevented by the high thermodynamic barrier of peptide condensation in aqueous solution. Liquid sulphur dioxide proves to be a superior alternative for copper-catalyzed peptide condensations. Compared to water, amino acids are activated in sulphur dioxide, leading to the incorporation of all 20 proteinogenic amino acids into proteins. Strikingly, even extremely low initial reactant concentrations of only 50 mM are sufficient for extensive peptide formation, yielding up to 2.9% of dialanine in 7 days. The reactions carried out at room temperature and the successful use of the Hadean mineral covellite (CuS) as a catalyst, suggest a volcanic environment for the formation of the peptide world on early Earth.

The authors show that when mixtures of amino acids are reacted under SPIF conditions with sulfur dioxide as the solvent, dipeptide formation can be observed for a large proportion of the possible amino acid combinations, even at lower concentrations, where practically none are observed in the equivalent reaction in water. This is certainly a nice result, but I have a number of concerns about the manuscript and the framing of the result as something prebiotically useful..
In the discussion, the authors note "formation of functional oligomer chains from single monomer building blocks remains a fundamental challenge" (in many Origin of Life scenarios), and I completely agree. However, the formation of dipeptides does not actually progress us very much further along that road unless larger aggregates can be formed. It is known that longer oligopeptides (certainly 5-, 6-mer and longer) can be formed in dehydrating / SPIF conditions, with the obvious possibility of achieving more complex products and approaching the size where sequences could impart real function. I assume the tri-/ tetra-peptides formed here were in very low concentration, given that there is so little discussion of them and their sequences (just the short paragraph at line 187)? The authors note that they have developed an analysis method capable of separation and sequencing down to quite low concentrations of products, but do not really use it to its full advantage here.
The expression 'complex mixture' of amino acids is used several times, but I do not understand in what way the mixtures of amino acids used is 'complex', nor can I see any analysis here of how the sequences produced under different conditions might differ from one another, which could be of interest.
There is some discussion of differences in reactivity observed in the amino acids when mixed and in the different solvents towards the end of the discussion. This is interesting and could (should) be made more of in a revised manuscript.
Note: the title of the SI is not the same as the manuscript. The version on the main manuscript is better, the products are not 'complex'.

Reviewer #2 (Remarks to the Author):
In the paper by Sauer F et al, the authors have investigated the metal-catalyzed synthesis of short peptide oligomers in sulphur dioxide as an alternative prebiotic solvent for supporting condensation-dehydration reactions between amino acid monomers. The results show that only very short peptides are formed, and with a relatively low yield (few percent) compared to several other reported strategies. However, it seems that the authors have developed a robust and elegant experimental framework that allows them to investigate and characterize reactions involving a very complex mixture (all 20 proteinogenic amino acids). Hence, if several key points will be addressed, I believe that this paper is worthy of publication in Nature Communications.
Major points: 1. In section 3.1 of the Results, are the authors sure that all products are accounted for when measuring the concentration of short peptides (i.e. by also integrating the remaining monomers)? 2. It the 4th paragraph in section 3.1. in Results, it is stated that "Contrary to our expectations, the reaction was favored at lower concentrations, leading to the formation of more tri-and even tetrapeptides at 50 mM after 21 d". However, it is not clear how the authors reached this conclusion. The only quantification observed from Fig. 1 and the relevant SI figures (not referred to in the paper) are of dipeptides. If tripeptides were quantified, then data should be shown. 3. Can you estimate the total percent conversion of amino acid into (any) products? This number will be very useful for comparison with previously known methods.
Minor points: 1. When discussing prebiotic peptide synthesis in the Introduction, it will be good to cite Greenwald J et al. (Angewandte 2016), in which carbonyl sulfide was used to polymerize amino acids into peptides. 2. Did you check if there is any racemization during the reaction (only L-amino acids were used as monomers)? 3. Supplementary figures 2-23 are not called in the main text. 4. It will be good to add a separate SI figure with the individual standards that were detected as a reference (e.g. AA, GG), as well as those that were not detected (GG DKP). 5. The authors state that "Urea has already been used as a condensation agent in prebiotic phosphorylation reactions with nucleosides41,42". Burcar B et al (Angewandte 2016) should also be added as a reference to this sentence. 4. Raw data is not shown for Supplementary Fig. 19. 5. To me, one of the very interesting findings of this paper was that the analysis of complex mixtures reveals effects that would have been overlooked in reactions of single amino acids. While this is mentioned in the Discussion, no possible explanation is offered. I believe it will be good to discuss some possible explanations.

Reviewer #3 (Remarks to the Author):
Overview This is a creative study which explores the chemistry of amino acid oligomerization in various solvents. Notably, this study finds that liquid SO2 promotes AA oligomerization, even at low reactant concentration, using prebiotically plausible mineral counterparts.
Major comments: This manuscript presents compelling synthetic chemistry. However, the authors have not adequately established that this synthetic chemistry is also *prebiotic* chemistry, i.e. that it is strongly relevant to early Earth.
My strongest criticism is the plausibility of condensed (liquid) SO2 on early Earth. The boiling point of SO2 is 263 K at 1 atm (https://pubchem.ncbi.nlm.nih.gov/compound/Sulfur-dioxide). Most models of early Earth assume ~1 atm surface pressure and predict mean surface temperatures above freezing in order to explain the zircon evidence. In this region of T-P space, liquid SO2 cannot exist. I do not agree with the offhand claim of this manuscript that liquid SO2 environments are highly *likely* on early Earth. Liquid SO2 may have been *possible* if early Earth did not conform to the "standard" picture, many aspects of which are only weakly justified. For example: • A high surface pressure for early Earth is not predicted from models (e.g. Kadoya et al. 2018), but is also not ruled out. Surface pressures of up to 10 bar are proposed (Kasting et al. 1993). Such conditions might allow liquid SO2, though I emphasize that these are extremal scenarios not consistent with current predictions of early Earth. • A cold early Earth is possible, particularly on a transient basis (Kadoya et al. 2018). Such a planet might have hosted temperatures locally (e.g, at the poles) or globally. The authors must identify the scenarios in which SO2 could be stable, and discuss their plausibility/prevalence in the text. I suggest the authors plot a phase diagram of SO2 and show on this phase diagram the different scenarios under which liquid SO2 would have been stable.
Note I have discussed here only concerns about liquid SO2 from the T/P perspective. There are other concerns as well. For example, SO2 is photochemically unstable. Could pure lakes of it have accumulated? This SO2 would also likely have been mixed with substantial amounts of water, which is also volcanically outgassed and which was also supplied by the hydrological cycle. Does their chemistry also work in mixed SO2-H2O solution?
A more minor concern relates to the authors' use of copper. Copper is geologically rare. The authors invoke Covellite as a plausible Cu(II) source. Can the authors comment on the estimated prevalence of covellite on early Earth?
Minor comments: In exploring SIPF: Can you comment on the implications of high salt concentrations for other parts of the protolife apparatus, like vesicles and nucleic acids.
Lines 50-51: At face value, this sentence implies Cu(II) would not have been available on early Earth, since photochemical oxygen abundances are predicted to exceed 10^-35 atm by many orders of magnitude (e.g., Haqq-Misra, Kasting & Lee 2011).
Line 60, 230-231: There is no consensus that early Earth had higher p than modern, and even some indications that it had lower p than modern (e.g., Som  Lines 65-67: I commend the authors for acknowledging the inevitable complexity and non-ideal conditions that must have been present in prebiotic chemistry.
Line 118-121: The authors have done well in considering the scaling of their reaction with reactant concentration, thus remediating a common criticism of prebiotic chemistry. However, 50 mM of amino acids are not low concentrations. I would describe that as quite concentrated, albeit less so than what is sometimes assumed.
Line 127-132: On what basis is it known that covellite was Hadean? How widespread was it on early Earth?
Lines 130-132: Can the authors please expand on the results. From what is written, it sounds like covellite is much worse than pure Cu(II) as a catalyst (e.g., no tri-or tetra-peptides). How does the yield compare.
Lines 135-148: Bravo to the authors! This is exactly the kind of work that is needed to move prebiotic chemistry from the lab to nature, i.e. convert it from synthetic to prebiotic chemistry.
Lines 197-200: What were the trends in dipeptide formation? Did they, e.g., match what is observed in biology?
On behalf of all the authors, I would like to thank the competent reviewers for providing us with great feedback on our manuscript. We greatly appreciate all the helpful suggestions and valuable comments provided by the reviewers to improve the quality of the manuscript.

Reviewer #1 (Remarks to the Author):
This work is a very nice result in that they can see dipeptides in sulfur dioxide in most combinations of AAs and at lower concentration compared to the equivalent reaction in water. I think this lays a really important new foundation in the formation of peptides. However, there is a lot of discussion here relating to sequence and function, while the longest they see is trace tetramers in limited cases, so it's probably a bit oversold -> I think they need to frame this as a neat (prebiotic potentially) way of making peptide /bonds/ in an anhydrous environments. Having said that, the work is of significance and very high quality and hence should be published in NCOMMS with suitable reframing.
 We thank the reviewer for acknowledging the great potential of SO2 as alternative solvent for prebiotic peptide formation.

Notes:
The In the discussion, the authors note "formation of functional oligomer chains from single monomer building blocks remains a fundamental challenge" (in many Origin of Life scenarios), and I completely agree. However, the formation of dipeptides does not actually progress us very much further along that road unless larger aggregates can be formed. It is known that longer oligopeptides (certainly 5-, 6-mer and longer) can be formed in dehydrating / SPIF conditions, with the obvious possibility of achieving more complex products and approaching the size where sequences could impart real function. I assume the tri-/ tetra-peptides formed here were in very low concentration, given that there is so little discussion of them and their sequences (just the short paragraph at line 187)? The authors note that they have developed an analysis method capable of separation and sequencing down to quite low concentrations of products, but do not really use it to its full advantage here.
 Plausible prebiotic reactions are characterized by robust reaction pathways that create a wide variety of products under simple conditions with a limited number of reactants. Our study focuses on the development of reasonable reaction conditions in an alternative, prebiotically plausible medium, where the barriers of peptide formation can be overcome. We definitely agree with the reviewer that the formation of dipeptides alone does not advance us in creating more functional peptides. The formation of longer peptides is for sure important, however, the function and capabilities of a specific peptide are also decisively determined by its sequence. Oligopeptides can lack function if they are assembled from the same amino acid. Therefore, complex amino acid mixtures were used in the study, as the successful conversion of these mixtures reinforces the prebiotic significance of the new reaction conditions. We could show that under the established simple environment every proteinogenic amino acid can be incorporated into dipeptides. We believe that this is an equally important aspect than the formation of longer peptides would be. Due to the high effort of the analysis (large amount of products), we limited ourselves to the evaluation of all 400 possible dipeptide sequences. However, tri-and tetrapeptides were also observed on several occasions (in

The expression 'complex mixture' of amino acids is used several times, but I do not understand in what way the mixtures of amino acids used is 'complex', nor can I see any analysis here of how the sequences produced under different conditions might differ from one another, which could be of interest.
 We understand that the expression "complex mixture" can be misleading as we are referring to the number of different reactants and not the functionality of the group. We therefore replaced "complex" with "large".
There is some discussion of differences in reactivity observed in the amino acids when mixed and in the different solvents towards the end of the discussion. This is interesting and could (should) be made more of in a revised manuscript.  We thank the reviewer for the constructive suggestion and we further elaborated the different amino acid activities in the different environments in the Discussion (lines 244-253 and 260-267): "A similar observation could be made for E which showed increased reactivity in the total mixture in SO2 compared to the prebiotic mixture. The results show that there are cooperative effects between different amino acids which could arise from interactions between the different side groups. For example, the hydroxyl group of S, T or Y could form a temporary ester bond with another amino acid, thereby promoting the formation of an amide bond through an energetically favoured ester-amide exchange. This effect has already been observed in the formation of depsipeptides. 1,2 Furthermore, the catalytic activity of especially glycine and histidine in peptide formation of other amino acids has been observed in other studies. 3 In the proposed mechanism, the catalytically active amino acid promotes the formation of a mixed tripeptide and the following cleavage of it leaves a homo-dipeptide of the other amino acid." "The comparison of the resulting dipeptide product distributions showed distinctive differences of amino acid activities in the two solvents. In water, for W, Q, C and H only a low reactivity could be observed. W, Q and N exhibit a low stability in the environment of the SIPF in which Q and N are hydrolysed to the respective acidic amino acids. The corresponding dipeptides could only be observed in a few cases. On the other hand, the acidic amino acid D and V showed a high reactivity in water. In SO2, rather poor reactivity of the acidic amino acids was noted. Furthermore, only traces of Y seem to be soluble in SO2 and accordingly, only few dipeptides of those dipeptides could be detected in the reaction mixtures." Note: the title of the SI is not the same as the manuscript. The version on the main manuscript is better, the products are not 'complex'.  Thank you, we corrected the title of the SI.

Reviewer #2 (Remarks to the Author):
In  We examined our reactions for side products commonly occurring in peptide condensation reactions (DKPs, degradation products of amino acids), however, as already mentioned no such compounds were found. Furthermore, we investigated the conversion of the amino acids Ala and Gly at different reaction times in SO2 by integrating the corresponding signals of CE measurements. The results showed a conversion of up to 10% for Ala and a 30% conversion for Gly. The conversion of Ala agrees very well with the observed dipeptide yields and shows that a targeted synthesis of dipeptides occurs during the reaction. However, the conversion of Gly is higher than the corresponding dipeptide yields would suggest. Therefore, minor side reactions seem to occur for Gly but not for Ala. Similar observations were also made in H2O for Trp, Gln, Asn. 4 We agree, that these findings are interesting, as they show an acceptable stability of the reactants in the established reaction conditions. However, the differing amounts of conversion of the two amino acids also show, that yields cannot be determined through the analysis of conversion rates alone. This study focuses on dipeptide formation of complex mixtures and on the yields which can be achieved in the new SO2-environment. In our opinion, the comparison with conversion rates of different methods does not add further insight into this topic.

It the 4th paragraph in section 3.1. in Results, it is stated that "Contrary to our expectations, the reaction was favored at lower concentrations, leading to the formation of more tri-and even tetrapeptides at 50 mM after 21 d". However, it is not clear how the authors reached this conclusion.
The only quantification observed from Fig. 1 Fig. 41-44).
Minor points:

When discussing prebiotic peptide synthesis in the Introduction, it will be good to cite Greenwald J et al. (Angewandte 2016), in which carbonyl sulfide was used to polymerize amino acids into peptides.
 Thank you very much for bringing this publication to our attention. We have added the reference.

Did you check if there is any racemization during the reaction (only L-amino acids were used as monomers)?
 Indeed, only L-amino acids were used in the reactions. So far, we didn't check for racemization, but we will extend this in further studies, which is an interesting point in the context of symmetry breaking.

Supplementary figures 2-23 are not called in the main text.
 Figures 19-36 (dipeptide product spectra of the different reactions) are referred to in line 200. Figures 2-18 show the calibration plots and the analysis runs for the quantification of dipeptide formation in SO2. We added a corresponding reference in the methods section.

It will be good to add a separate SI figure with the individual standards that were detected as a reference (e.g. AA, GG), as well as those that were not detected (GG DKP).
 The EIEs of the reference peptides GG, AG/GA, AA, GGG and AGG are already included in the SI (Supplementary Figure 37). We further added a CE-MS run of the DKPs of GG and AA (Supplementary Figure 38).

The authors state that "Urea has already been used as a condensation agent in prebiotic phosphorylation reactions with nucleosides41,42". Burcar B et al (Angewandte 2016) should also be added as a reference to this sentence.
 We included the requested reference. Supplementary Fig. 19.  The raw data was already shown in Supplementary Figures 103 and 104.

To me, one of the very interesting findings of this paper was that the analysis of complex mixtures reveals effects that would have been overlooked in reactions of single amino acids.
While this is mentioned in the Discussion, no possible explanation is offered. I believe it will be good to discuss some possible explanations.  We thank the reviewer for the suggestion and added some possible explanations for the observed effects in the Discussion (lines 244-253): "A similar observation could be made for E which showed increased reactivity in the total mixture in SO2 compared to the prebiotic mixture. The results show that there are cooperative effects between different amino acids which could arise from interactions between the different side groups. For example, the hydroxyl group of S, T or Y could form a temporary ester bond with another amino acid, thereby promoting the formation of an amide bond through an energetically favoured ester-amide exchange. This effect has already been observed in the formation of depsipeptides. 1,2 Furthermore, the catalytic activity of especially glycine and histidine in peptide formation of other amino acids has been observed in other studies. 3 In the proposed mechanism, the catalytically active amino acid promotes the formation of a mixed tripeptide and the following cleavage of it leaves a homo-dipeptide of the other amino acid."

Reviewer #3 (Remarks to the Author):
Overview

This is a creative study which explores the chemistry of amino acid oligomerization in various solvents.
Notably, this study finds that liquid SO2 promotes AA oligomerization, even at low reactant concentration, using prebiotically plausible mineral counterparts. This SO2 would also likely have been mixed with substantial amounts of water, which is also volcanically outgassed and which was also supplied by the hydrological cycle. Does their chemistry also work in mixed SO2-H2O solution? Line 60, 230-231: There is no consensus that early Earth had higher p than modern, and even some indications that it had lower p than modern (e.g., Som et al. 2012Som et al. , 2016Gebauer, Grenfell, Lammer. et al. 2020)  We agree with the reviewer -since this is the first time pure liquid SO2 is recognized as prebiotic solvent -that the existence of it on early Earth has to be verified. Please note that at no point in the manuscript we are making offhand claims but attempted to substantiate the statements made with appropriate studies from experts in this field. However, there is no doubt that the question about the reaction conditions at the time of the origin of life is highly complex and to this day not solved. Almost no rock samples from the Hadean have survived, but even if they had, it would be difficult to determine the atmospheric pressure from them. Zahnle et al. assume very high levels of CO2 (approximately 100 bar) and additional 2-3 bar N2, others estimate 10-20 bar CO2. [5][6][7][8] Of course, the pressure decreases with the formation of carbonates, however, it does so in an unknown manner and period of time. Haqq-Misra et al assume a 1-2.8 bar surface pressure for the period of 3-4 Ga for their models. 9 Also about the prevailing temperature no certain prediction can be made at the present time, since it depends substantially on the composition of the atmosphere. Liquid water deposits are considered to be ensured, however, subzero temperatures would have dominated without the presence of greenhouse gases because of the weaker sun. So there are large variations in the assumptions made about the early surface pressure and temperature, however, all of the mentioned conditions would support liquid SO2. We included a phase diagram of SO2 below. 10 Unfortunately, we could not find the reference of Kadoya et al. from 2018, so we assumed the reviewer meant either one from 2019. 11,12 . As the reviewer already mentioned, high pressures as well as low temperatures are not ruled out for the early Earth in these studies. Som et al. and Gebauer et al. acknowledge the difficulties in estimating past pressures and state that the results of studies vary strongly. [13][14][15] However, the assumptions made by Som et al. refer to the time of 2.7 Ga, when life probably already existed on earth. 16 Apart from the general assumptions about the conditions during the origin of life (which is not a fixed point in time but a range of 200-800 Ma), the possibility of different, location-dependent conditions must be considered. Furthermore, seasons and different microhabitats similar to today's earth are conceivable. We therefore think that there are no "standard" prebiotic reaction conditions but instead a wide range of different conditions has to be considered in order to explain the emergence of life. According to the current state of knowledge, liquid SO2 can by no means be ruled out as an alternative solvent for prebiotic reactions since the above discussed scenarios would all allow for its existence.  As described by Kasting et al. the photochemical degradation of SO2 mainly takes place in the atmosphere with the formation of a sulphur layer. 17 Dry and wet depositions are discussed in the same context. We could not find studies regarding the UV stability of SO2-lakes. However, the UV problem affects not only SO2 but also other organic molecules.  We did not investigate peptide formation in defined H2O-SO2-mixtures. However, the SO2 was not dried prior to usage and water originating from the condensation reactions was not separated from the reaction mixture. Consequently, all reaction mixtures contained small amounts of water.
A more minor concern relates to the authors' use of copper. Copper is geologically rare.  Because of the low levels of oxygen in the atmosphere and especially in the aqueous, mineralforming environments of the prebiotic Earth, the number of available copper minerals was far less than it is today. Hazen assumes an oxygen fugacity close to that of the hematite-magnetite buffer (fO2 ~ 10 -72 ), which renders the formation of copper oxides, silicates, carbonates, sulfates, arsenates and phosphates impossible. However, Hazen also states that the formation of the copper-sulfur bond does not depend primarily on the fugacity of oxygen, but on that of sulfur. Since this was sufficiently high (fS2 > 10 -44 ), sulfides and sulfosalts were available on the prebiotic Earth at any fO2 < 10 -52 . Therefore, the number of different copper minerals is estimated to be 13 (including covellite), which constituted 2.1% of the prebiotic mineral inventory. 18 Therefore, especially local deposits of copper minerals cannot be ruled out. However, we are aware of the limited availability of copper(II) and we are working on more abundant alternatives in ongoing investigations.

Minor comments:
In  The amount of dipeptides formed in the 7-day period using covellite as a catalyst is smaller compared to the amounts obtained with synthetic CuCl2 in the same period of time. Therefore, the corresponding yields cannot be quantified by CE. A rough estimation can be made by comparing the respective dipeptide peak areas of the CE-MS measurements. After 7 days, yields obtained with synthetic CuCl2 are approximately ten times as high as with covellite. Nevertheless, the successful application of the prebiotic mineral in low amounts (21 mmol amino acids per gram mineral as opposed to the often used 0.2 mmol amino acids per gram mineral) 20 demonstrates the great potential of the established reaction conditions in SO2 for peptide formation. On a prebiotic timescale larger amounts of dipeptides will be formed.
Lines 65-67: I commend the authors for acknowledging the inevitable complexity and non-ideal conditions that must have been present in prebiotic chemistry.
Lines 135-148: Bravo to the authors! This is exactly the kind of work that is needed to move prebiotic chemistry from the lab to nature, i.e. convert it from synthetic to prebiotic chemistry.  We are very happy the reviewer acknowledges one of the key features of the study and shares our opinion on the necessity of expanding prebiotic systems.
Lines 197-200: What were the trends in dipeptide formation? Did they, e.g., match what is observed in biology?  We could observe different activities of the amino acids depending on the solvent and mixture the amino acid is used in. For example, we observed an increased formation of proline peptides in SO2 which are known for their catalytic activity. Please also refer to the Discussion of the revised manuscript where we further elaborated these effects (lines 244-253 and 260-267): "A similar observation could be made for E which showed increased reactivity in the total mixture in SO2 compared to the prebiotic mixture. The results show that there are cooperative effects between different amino acids which could arise from interactions between the different side groups. For example, the hydroxyl group of S, T or Y could form a temporary ester bond with another amino acid, thereby promoting the formation of an amide bond through an energetically favoured ester-amide exchange. This effect has already been observed in the formation of depsipeptides. 1,2 Furthermore, the catalytic activity of especially glycine and histidine in peptide formation of other amino acids has been observed in other studies. 3 In the proposed mechanism, the catalytically active amino acid promotes the formation of a mixed tripeptide and the following cleavage of it leaves a homo-dipeptide of the other amino acid." "The comparison of the resulting dipeptide product distributions showed distinctive differences of amino acid activities in the two solvents. In water, for W, Q, C and H only a low reactivity could be observed. W, Q and N exhibit a low stability in the environment of the SIPF in which Q and N are hydrolysed to the respective acidic amino acids. The corresponding dipeptides could only be observed in a few cases. On the other hand, the acidic amino acid D and V showed a high reactivity in water. In SO2, rather poor reactivity of the acidic amino acids was noted. Furthermore, only traces of Y seem to be soluble in SO2 and accordingly, only few dipeptides of those dipeptides could be detected in the reaction mixtures." Lines 260-262: please quantify effect on yields.  We quantified the obtained peptide products in SO2 for a large range of initial reactant concentrations and different reaction times while using a small set of two amino acids (Fig. 1). We did not determine yields of the dipeptides produced from the complex mixtures as the many ions influence each other during mass analysis which renders the precise quantification very difficult. Instead we compared the resulting dipeptide product spectra of SO2 to the ones obtained in H2O (Fig. 5). Especially in the total mixture, the potential of SO2 is on full display, since even at low concentrations a significantly larger amount of dipeptides is formed than in H2O.