Towards a generic prototyping approach for therapeutically-relevant peptides and proteins in a cell-free translation system

Advances in peptide and protein therapeutics increased the need for rapid and cost-effective polypeptide prototyping. While in vitro translation systems are well suited for fast and multiplexed polypeptide prototyping, they suffer from misfolding, aggregation and disulfide-bond scrambling of the translated products. Here we propose that efficient folding of in vitro produced disulfide-rich peptides and proteins can be achieved if performed in an aggregation-free and thermodynamically controlled folding environment. To this end, we modify an E. coli-based in vitro translation system to allow co-translational capture of translated products by affinity matrix. This process reduces protein aggregation and enables productive oxidative folding and recycling of misfolded states under thermodynamic control. In this study we show that the developed approach is likely to be generally applicable for prototyping of a wide variety of disulfide-constrained peptides, macrocyclic peptides with non-native bonds and antibody fragments in amounts sufficient for interaction analysis and biological activity assessment.

Reviewer #3: Remarks to the Author: Cell-free peptide and protein synthesis, based on in vitro transcription and translation, is attracting a lot of attention right now. Many laboratories are using this technology because it enables a broad variety of bioengineering applications. There is, currently, a considerable effort put in this technology for biomanufacturing high-value peptides or proteins. In their article 'Towards a generic prototyping approach for therapeutically-relevant peptides and proteins in a cell-free translation system', Wu et al show that by using affinity matrix directly in cell-free reactions, one can produce greater amounts of bioactive peptides and proteins (such as antimicrobials and antibodies) compared to standard batch reactions. The authors show that with the affinity matrix, the level of aggregation of the synthesized peptides or proteins is reduced and that one can perform some folding biochemistry. The work covers a good set of peptides and proteins, including some related to the covid-19 virus. The experiments focus mostly on peptides and proteins with disulfide bonds. The study also includes the use of non-canonical amino acids. My comments are grouped into major and minor points.
Major points: 1. The use of an affinity matrix is relatively new and brings some originality to the current cell-free expression technology capabilities. The authors show that attaching a peptide or protein using an affinity tag onto a matrix simultaneously with its synthesis does improve the production of highvalue biologics. In that sense, the work grasps the current trend very well, especially when it is about antimicrobial peptides and antibodies. Whether the set of proteins and peptides tested is enough to claim generalization of the approach is not clear, as in bioengineering, and any biologyrelated field, one can always ask for more tests. I personally think that the peptides and proteins assayed in this work form a reasonable set of tests. The authors support their observations with some good references. Especially, their analysis of the thermodynamics and kinetics of folding is interesting and they certainly provide some valuable insights into a complex problem. 2. The claim that this approach is simple is still not convincing. The work is very technical and requires substantial preparation. Claiming that it can be easily adopted by non-specialized labs is hard to believe. Cell-free expression systems are never easy to work with at the entry-level. It does not diminish the good observations made in the study. It just does not seem that easy for non-specialized labs. The preparation of the matrix, for instance, is facile when one knows how to do it, certainly less straightforward for non-specialists. Related to this, there are a lot of data, and it is hard to get clear messages from it. The number of abbreviations is very large, and it is often hard to follow, and some abbreviations are not spelled out. I suggest making a list of abbreviations. Also, it would be good to summarize the method in a single figure or a protocol-like page. If it is that easy, it should be possible to summarize the approach concisely so that others can use it. 3. The results obtained in this work could have been discussed or compared with the current stateof-the-art capabilities of the field. For instance, some companies are selling cell-free systems specifically for producing peptides and proteins with disulfide bonds (e.g., SHuffle kit from the company NEB). Some of the results obtained with a lysate-based E. coli cell-free system are compared to the PURE system which is good. Another area that could have been included in the discussion is the current trend for fast biomanufacturing at the point of care. For example, several groups have shown how to prepare lyophilized cell-free proteins synthesis systems, removing the need for a cold chain and enabling biomanufacturing at the point of care. How does the approach described in this article position itself with respect to that? If the approach is not yet adapted for point-of-care applications, just state it, and explain how it could be adapted for point-of-care applications (if it is possible).
Minor points: • Using cell-free translation all over the place is confusing, as it appears that the work is cell-free transcription and translation for the most part.
• Methods: should be reviewed and completed. For example, what is the composition of buffer S30B?
Reviewer #3 (Remarks to the Author): Cell-free peptide and protein synthesis, based on in vitro transcription and translation, is attracting a lot of attention right now. Many laboratories are using this technology because it enables a broad variety of bioengineering applications. There is, currently, a considerable effort put in this technology for biomanufacturing high-value peptides or proteins. In their article "Towards a generic prototyping approach for therapeutically-relevant peptides and proteins in a cell-free translation system", Wu et al show that by using affinity matrix directly in cell-free reactions, one can produce greater amounts of bioactive peptides and proteins (such as antimicrobials and antibodies) compared to standard batch reactions. The authors show that with the affinity matrix, the level of aggregation of the synthesized peptides or proteins is reduced and that one can perform some folding biochemistry. The work covers a good set of peptides and proteins, including some related to the covid-19 virus. The experiments focus mostly on peptides and proteins with disulfide bonds. The study also includes the use of non-canonical amino acids. My comments are grouped into major and minor points.
Major points: Point 1. The use of an affinity matrix is relatively new and brings some originality to the current cell-free expression technology capabilities. The authors show that attaching a peptide or protein using an affinity tag onto a matrix simultaneously with its synthesis does improve the production of high-value biologics. In that sense, the work grasps the current trend very well, especially when it is about antimicrobial peptides and antibodies. Whether the set of proteins and peptides tested is enough to claim generalization of the approach is not clear, as in bioengineering, and any biology-related field, one can always ask for more tests. I personally think that the peptides and proteins assayed in this work form a reasonable set of tests. The authors support their observations with some good references. Especially, their analysis of the thermodynamics and kinetics of folding is interesting and they certainly provide some valuable insights into a complex problem.
Response to point 1: We thank the Reviewer for the positive assessment of our work. Point 2. The claim that this approach is simple is still not convincing. The work is very technical and requires substantial preparation. Claiming that it can be easily adopted by non-specialized labs is hard to believe. Cell-free expression systems are never easy to work with at the entrylevel. It does not diminish the good observations made in the study. It just does not seem that easy for non-specialized labs. The preparation of the matrix, for instance, is facile when one knows how to do it, certainly less straightforward for non-specialists. Related to this, there are a lot of data, and it is hard to get clear messages from it. The number of abbreviations is very large, and it is often hard to follow, and some abbreviations are not spelled out. I suggest making a list of abbreviations. Also, it would be good to summarize the method in a single figure or a protocol-like page. If it is that easy, it should be possible to summarize the approach concisely so that others can use it.

Response to point 2:
We agree with the Reviewer that for the research groups being new to cell-free translation the described approach would not seem trivial. However, E. coli-based cellfree system has an important advantage over the other cell-free systems especially of eukaryotic origin in that it is not very capricious owing to a less strict translational control. We also agree that the abundance of data obscures the technicalities of the method described. To this end, following the reviewer"s suggestion, we provide a step-by-step workflow describing the preparation of affinity clamp-coated resin, preparation of S30 extract, reaction assembly and refolding conditions for peptides, disulfide-constrained proteins and also for proteins free of disulfide bonds. The protocol page is now provided under the Supplementary methods section. We believe that ordering of the technical steps into a straight and cohesive procedure would make for a great complement to this work. We also included the list of abbreviations as Supplementary Table 8.
Below we provide the screenshot of the Supplementary methods section: Point 3. The results obtained in this work could have been discussed or compared with the current state-of-the-art capabilities of the field. For instance, some companies are selling cellfree systems specifically for producing peptides and proteins with disulfide bonds (e.g., SHuffle kit from the company NEB). Some of the results obtained with a lysate-based E. coli cell-free system are compared to the PURE system which is good. Another area that could have been included in the discussion is the current trend for fast biomanufacturing at the point of care. For example, several groups have shown how to prepare lyophilized cell-free proteins synthesis systems, removing the need for a cold chain and enabling biomanufacturing at the point of care. How does the approach described in this article position itself with respect to that? If the approach is not yet adapted for point-of-care applications, just state it, and explain how it could be adapted for point-of-care applications (if it is possible).

Response to point 3:
As for the first point raised by the Reviewer regarding the comparison of our system with the current state of the field, we feel that we previously provided a brief summary in the introduction subchapter in which we discussed the general trend and associated limitations in the field. We also extrapolated to how the currently proposed strategy could address them. We argue that at physiological conditions it is generally hard for translated polypeptides to avoid the kinetics traps resulting in misfolding and aggregation in particular in the cell-free system due to possible loss of concerted chaperon activity. Therefore, majority of studies trying to achieve the productive tradeoff between folding and aggregation through combining the oxidizing environment for the accelerated closure of disulfide bridges with complex chaperon cocktail. The Shuffle strain was developed on purpose to support the disulfide bond closure in bacteria cytoplasm. We indirectly mention the Shuffle T7 strain when discussing the previous attempts to produce Pn3a, the neurotoxin, also belonging to our test set, as we cite the work by Sharma et al. where a panel of strains was screened including the Shuffle T7. Although this strain favorably compared with the rest of expression strains in terms of supporting high yields of soluble productthe peptide of interest was remaining in a largely misfolded state. This is consistent with the fact that despite the notion that disulfide bond closure directs folding by providing the constraints many studies show that accelerated formation of disulfide bonds results in randomly cross-linked intermediates. Therefore, we argue that the success rate of the attempts to achieve this tradeoff would still hinge on individual features of translated sequences while thermodynamically controlled conditions would be universally applicable.
"However, the rest, partially undergoing "collapsed" folding 34, demanded a downstream refolding at optimized redox conditions despite their expression as fusions with MBP in E. coli periplasm or in the engineered E. coli strain promoting oxidative folding47, 48." As for the suitability of the proposed strategy for biomanufacturing at the point of carein our previous response to the Reviewer we acknowledged the importance of this subject as it projected for us to the question of scalability performance of the proposed strategy in general. We concluded that the proposed approach in its current format is limited by the use of the protein-coated affinity matrix which can be recycled up to seven times only with a certain loss of binding capacity following each recycling step. Therefore, we concluded that at its current format the approach would be mostly amenable for prototyping rather than for biomanufacturing and provided the statement in the discussion subchapter: "Although the proposed strategy is likely to benefit from the integration with homeostatic energy regeneration 6 and the use of engineered strains 8, 36 it is important to note that in its current format it is rather applicable for small-scale prototyping due to limited reuse of the proteincoated resin." Since the proposed strategy is not applicable for biomanufacturing in its present format we believe that the same would be true for the point of care application since biomanufacturing scale would be a critical factor. On another hand in the same paragraph, we suggested a potential route to tailor the approach towards biomanufacturing scale and outlined the possible issues on this way: "However, it can be potentially tailored to biomanufacturing via utilizing an inexpensive and recyclable metal affinity resin instead. In contrast to protein-coated resin the latter would give an advantage of using elevated denaturant concentrations and prolonged incubation period to ensure complete resolution and recycling of all inter-and intramolecular aggregates. It may come, however, at a cost of compromised product purity42, 54 and the need of readjusting the reductant concentration at multiple stages to avoid metal reduction." While not being suitable to biomanufacturing the proposed strategy, offering a standardized procedure for prototyping a verity of polypeptides, can be of advantage for the synthetic biology in streamlining the design-build-test cycles, rapid iteration through which is crucial for the synthetic biology progresswe also provided a brief account of this in the discussion chapter: "Offering a standardized procedure (Supplementary methods) for prototyping of a wide variety of (poly)peptides the proposed strategy can aid in protein/peptide engineering by streamlining design-build-test cycles, a rapid iteration through which remains one of the major bottlenecks of synthetic biology." Unfortunately, we cannot further elaborate on the point of biomanufacturing at the point of care as the size of the manuscript is currently hitting its limit and we already had to reduce it by 700 words.

Minor points:
Minor point 1: Using cell-free translation all over the place is confusing, as it appears that the work is cell-free transcription and translation for the most part.
Response: Following the reviewer"s concern raised during the first round of reviewing process we already converted "cell-free translation" to "cell-free transcription-translation" at five instances in the manuscript whenever we refer to the technical side (in Results (2) and Methods (3) sections) while for the introduction and discussion subchapters we reserve a more general term "cell-free translation" (occurs at three instances) as we discuss the concerns linked to translation and we feel that the additional use of "transcription" term would not be practical in this light.
Minor point 2: Methods: should be reviewed and completed. For example, what is the composition of buffer S30B?

Response:
We now provided a detailed workflow for the assembly of resin-assisted translation reaction where all buffer compositions are provided in a table format: