An integrated workflow for quantitative analysis of the newly synthesized proteome

The analysis of proteins that are newly synthesized upon a cellular perturbation can provide detailed insight into the proteomic response that is elicited by specific cues. This can be investigated by pulse-labeling of cells with clickable and stable-isotope-coded amino acids for the enrichment and mass spectrometric characterization of newly synthesized proteins (NSPs), however convoluted protocols prohibit their routine application. Here we report the optimization of multiple steps in sample preparation, mass spectrometry and data analysis, and we integrate them into a semi-automated workflow for the quantitative analysis of the newly synthesized proteome (QuaNPA). Reduced input requirements and data-independent acquisition (DIA) enable the analysis of triple-SILAC-labeled NSP samples, with enhanced throughput while featuring high quantitative accuracy. We apply QuaNPA to investigate the time-resolved cellular response to interferon-gamma (IFNg), observing rapid induction of targets 2 h after IFNg treatment. QuaNPA provides a powerful approach for large-scale investigation of NSPs to gain insight into complex cellular processes.

Reviewer #1 (Remarks to the Author): In the manuscript entitled: "An integrated workflow for quantitative analysis of the newly synthesized proteome", Borteçen at al. present a QuaNPA workflow for the quantitative analysis of the newly synthesized proteome (NSP).Although the importance of protein synthesis is well recognized, quantitative analysis of protein synthesis has proven to be challenging.This workflow includes an optimized semiautomated sample preparation protocol and the application of multiplexing DIA-MS to reduce the sample input, increase the proteomic depth, reduce missing values, and improve quantitative accuracy of the analysis.The authors first optimized an automated protocol including the synthesis of magnetic alkyne agarose beads and NSP enrichment and SP3 peptide purification using a Bravo liquid handling system enabling the analysis of 8-96 samples in parallel with a minimized sample input.Then, two benchmark samples were used to compare different MS methods and data processing software for NSP analysis showing a better performance of DIA MS2-based quantification in plexDIA than DDA.Finally, the optimized protocol was applied to study the response to IFNy in HeLa cells.
The main novelty of the manuscript lies in the technology improvement enabling parallel processing of multiple samples and using lower sample input.Furthermore, the NSP analysis depth and quantification benefits from a novel combination with a DIA-based MS method.The manuscript is well written and presents high quality data.My main concern is the novelty of the method application part, since a similar analysis of IFNy signaling has been performed recently (Kleinpenning, 2020).As a benchmark experiment, the analysis of IFNy signaling showed that QuaNPA is able to gain biologically relevant insights by comparing to other omic datasets and annotation databases and lead to identification of potentially novel targets of IFNy signaling.However, to increase the impact of the biological findings of the current study, I would strongly suggest performing follow-up experiments focusing, e.g., on the validation of some of the potentially novel targets of IFNy signaling or extending the study to a more interesting system.
Comments: 1)Page 3, line 84-85: "Although DIA is directly compatible with label-free quantification, it has been scarcely applied in combination with SILAC labelling because of challenges in data analysis 41,42,43.Recently, plexDIA was introduced in the DIA-NN environment, demonstrating deep proteome coverage and quantitative accuracy for the analysis of multiplexed samples with non-isobaric labels (mTRAQ).Conceptually, plexDIA could also be applied to SILAC-labelled samples, in particular for the analysis of NSPs."Another DIA-data analysis software, Spectronaut has been capable of analyzing SILAC-DIA samples for several years, and it has been later optimized for the analysis of pSILAC samples (Salovska et al., 2020) enabling a straightforward analysis of SILAC and pSILAC DIA-MS data.The way this paragraph of the introduction is written gives a false impression that the analysis of such data is problematic and plexDIA offers a unique solution, which is not true.
2) Page 4, paragraph starting at line 132: Please briefly clarify in the text and related figure description how the MS data for this optimization step were acquired and analyzed.It is not clear from the manuscript until the next results section.3) Furthermore, in Figure 3B-D and 3E-F, it is not clear whether the analysis is comparing the same set of overlapping ids identified across all conditions or every boxplot/violin plot represents all protein ids identified in each condition (which would be fair).For Figure 3C, also please clarify whether these are raw or normalized SILAC ratios.In the Method section (line 597) it is mentioned that the normalized ratios were used.The authors might also consider presenting peptide-precursor level identification and ratios, since collapsing the peptides into protein level H/M ratios might remove a lot of quantitative noise.4) Page 6, Line 178-180: In the recent paper, plexDIA has been used to analyze multiplexed samples using mTRAQ labeling in which peptide precursor with C-terminal arginine generate unlabeled y-ions during fragmentation.SILAC labeling presents a similar challenge when using MS2 quantification due to the lack of labeling of the majority of the b-ions and the co-isolation of the labeling pairs in the same DIA window.Since this is important for the SILAC MS2 quantification, could you please briefly clarify how DIA-NN handles these unlabeled b-ions?5) Figure 4C -in addition to protein level identification it might be also informative to also show the number of identified peptides.6) Figure 4D-E -Please, clarify whether the presented analysis is based on a set of overlapping protein ids between all the methods used or all ids identified in the respective experiments.It might be informative to evaluate the precision and accuracy for the set of overlapping ids and the "extra" ids identified by DIA and also show the precursor level result when possible.Are the observed differences statistically significant?Please also add a brief clarification in the methods section how DIA-NN performs the MS1 and MS2-based quantification, is there any fragment ion or precursor ion level filtering?Especially considering the b-ion interference mentioned above?7) Page 7, line 236-243 (Figure 5A): The depiction of the AHA and SILAC labeling time points in Figure 5A is not very clear, and it is also not described in the relevant section of the manuscript.Perhaps a different visualization using a timeline would be better to directly understand that for the later timepoints, the labeling was only performed in the last 6 hours?8) Page 8, line 248-249, "Shorter labeling times did not lead to noticeable reduction in the number of quantified proteins or precision, etc..": While the good identification numbers in the early time points might be explained by the enrichment, the analysis of all samples together can also affect the identification result in individual samples, e.g., due to the MBR function.To make this conclusion, have you also tried to process the samples separately and are the identification results consistent? 9) Page 8, line 258-263, Figure S4, comparison to PhosID.When relating to the PhosID dataset, the supplementary figure shows more differentially regulated proteins after 4 hours in the PhosID dataset compared to QuanNPA.It might be fair to comment on this observation too.10) Page 8, line 276-277, Figure 6: The two log2FC cutoffs mentioned in the manuscript and shown in the Figure 6 legend are a bit confusing.I would suggest also adding the p < 0.05 and |log2FC| > 1 cutoff into the Figure 6, for example in the grey box next to the label "differentially expressed proteins".Moreover, in the legend of Figure 6, it is a bit confusing why all subgroups are depicted with an asterisk while the asterisk indicate known link to the IFNy treatment as described in the legend text.11) Page 11, line 364-373: The authors hypothesize that plexDIA might have difficulties with quantifying the "extreme" ratios in mix1 due to the fact that DIA-NN does not offer the re-quantify function that is available in MaxQuant.It would be relevant in this section to discuss the fact that another DIA-data processing software, Spectronaut, offers the "inverted spike-in workflow" that has been shown to recover even more "extreme" ratios of H/L pairs, up to 1:16 in a benchmark sample (Salovska 2020 and 2021).12) Page 15, line 512 and Page 17, line 590 -What is meant by the "global proteome sample"?13) Page 17, line 594-597 -It is not clear in which analyses the MBR function was enabled and disabled.Maybe add a link to specific figures here?As I mentioned above, it is also unclear to me in which analyses the raw and normalized SILAC ratios were used, as reported by MaxQuant, and in which the authors used the corresponding light and heavy signals and assembled the proteins using MaxLFQ.14) Page 26, Figures 5 and 6  Reviewer #2 (Remarks to the Author): In the manuscript "An integrated workflow for quantitative analysis of the newly synthesized proteome", Borteçen et al. present an optimized workflow to study newly synthesized proteins (NSPs) which tackles several challenges such as high sample inputs, lack of automation and sufficient throughput, long sample preparation and measurement times, and limited proteome coverage.To support the validity and benefits of their workflow, they apply it to an already well-studied biological system analyzing NSPs in response to IFNg and compare their results with existing data.The authors present a well-written and insightful manuscript with clear figures.The single sample preparation steps follow previously described protocols, but the authors assemble them in a coherent workflow, which they also describe in detail in the method section and which thus appears easy to follow.Given the downscaling and increased throughput of this optimized workflow, this study will likely raise some interest in the proteomics field.However, it remains mainly a method paper with limited new biological insight.
I believe that addressing the following points will enhance the quality and impact of the presented manuscript further: 1.The authors enrich and quantify the newly synthesized proteome and conclude that regulations of these NSP result from a change in protein synthesis.This appears reasonable given e.g. the biology of their IFNg experiment, but this conclusion can still be misleading.Allow me the following thought experiment: Let us assume that newly synthesized and preexisting proteins are degraded at identical rates, i.e. the cell cannot distinguish between a rather new and a rather old protein.This is an assumption frequently made in pulsed SILAC/turnover experiments and a prerequisite of curve fitting to time course pulsed SILAC experiments.Further, let us assume that the synthesis rate of protein A remains the same after a treatment, but its degradation rate increases.In this case, a larger fraction of what is synthesized newly will be degraded at any given timepoint and we would observe a reduced amount of newly synthesized protein A. Following the claims in the paper, however, this situation would always be interpreted as a sloweddown synthesis of protein A. Likewise, if the degradation of protein B was slower and its synthesis rate remained the same, the amount of newly synthesized protein B would increase because a smaller fraction is degraded, which would be interpreted falsely as a faster synthesis.It is a common misconception that pulse experiments can distinguish between synthesis and degradation, although we can only measure the combined effects of both in a pulsed SILAC experiment.I believe this misconception should be addressed.I hope the authors will follow my argument and add a section to their manuscript which should state that, even though they do enrich for newly synthesized proteins, observed changes in NSPs can also come from a change in their degradation.2. In supplementary figure 2, combined (labeled + unlabeled) intensities are used as a proxy for total protein abundance (i.e. protein copies).While there is some correlation between measured protein intensities and actual abundance, the intensities are also markedly affected by protein length (i.e.number of potentially detectable peptides).Please use iBAQ or similar as a proxy for absolute protein abundances to establish that 'stickiness' does not depend on protein abundance.3. I was intrigued by the authors' comparison of the two different DIA methods m1 and m2.Do the authors have any suggestions as to why m2 yields consistently more quantified proteins than m1 for both MS1-and MS2-based quantification despite longer cycle times and fewer MS2 scans?This appears counterintuitive given the notion in the DIA field that more MS2 scans and faster cycle times are better.4.An interesting observation is also that DDA yields >10% fewer quantified proteins when the light channel is not the most abundant one.This may indicate a long-existing bias in how the data is processed in MaxQuant.Perhaps it would be worth explicitly mentioning this in the manuscript.5. Finally, in line 203 and following, the authors state that "the gained proteins in the DIA data compared to DDA are likely to be in the lower abundance range".I suggest that the authors support this statement with actual analyses.For example, what is the overlap in proteins between DDA and DIA? Are the proteins exclusive to DIA lower in intensity than the ones shared with DDA?
Minor edits/comments: 1. Please consistently capitalize the first word following the numbering in Figure 1 (e.g."1.Preparing magnetic…") 2. Please define AFA. 3. Line 255, 266, 277: Closing brackets are missing.4. Line 391: "lysates from" instead of "lysates form" 5. Supplementary Figure 3 needs to be referenced in the main text.6. Line 474, 526, 527, 549, 957, and 958: "Thermo Fisher" instead of "Thermo Fischer" Reviewer #3 (Remarks to the Author): Bortecen et al. introduce a method called QuaNPA, which stands for quantitative analysis of the newly synthesized proteome, that enables the analysis of newly synthesized proteins (NSPs) in response to cellular perturbations.QuaNPA involves pulse-labeling cells with clickable and stable-isotope-coded amino acids, and uses mass spectrometry and data analysis to quantitatively analyze the NSPs.The study found that QuaNPA provides a powerful approach for large-scale investigation of NSPs, and was able to successfully investigate the time-resolved cellular response to interferon-gamma (IFNg).The main mass spectrometry related principles used in this manuscript have already been introduced significantly earlier and many of them pioneered by the Krijsveld group themselves.Nevertheless, I do think that the authors present very impactful improvement relative to most methods available to study changes in protein production at a global scale and at relatively short timescales.The main advantages are that new kind of magnetic beads are presented with a higher capacity to capture newly produced proteins, an automated experimental workflow to isolate NSPs at high efficiency and also that the authors took advantage of the plexDIA workflow to significantly increase protein coverage without increasing mass spec measurement time.Therefore, this highly optimized workflow to measure protein production changes across many conditions, with a significantly lower sample input amount and at higher coverage and throughput, will be extremely useful to the community (admittedly, not all labs will have a robot available to apply the automated workflow, but many will be able to still use the majority of improvements with standard lab equipment).In addition, the manuscript is generally well written and the experiments very thorough.There are only a few (mostly minor) points that I think the authors should address before acceptance.
1. Generally, some of the optimization/characterization of QuaNPA as shown in Figure 3, should actually be at least partly repeated by doing DIA measurements instead of DDA as DIA does indeed show significantly higher sensitivity and therefore some conclusions drawn from the DDA measurements might not be translated 1:1 to DIA measurements.The most important one is the conclusion the authors draw about the protein input amount that is needed for improving protein coverage.Based on the DDA measurements not more than 100ug total protein input is needed for NSP enrichment as additional amount does not lead to higher coverage (Figure 3F).However, due to the increased sensitivity of DIA relative to DDA this might not hold true for the final QuaNPA approach.Also, for example non-enriched samples show fewer quantified protein groups (Figure 3A) might not be 100% true with DIA.It would be great if this could be assessed by the authors.
2. For nearly all comparisons where sensitivity is assessed also peptide level identification numbers should be provided.Especially for the DDA and DIA comparison as DIA-NN is a bit more "aggressive" about the protein grouping, meaning separating proteins already on less stringent criteria into separate groups than other programs.4D and 4E: for the CV comparison and log2 ratio comparison between DIA and DDA, I would suggest to also separately look only at the protein groups that overlap between all measurements.This is probably a bit fairer as DDA measured proteins are probably on average higher expressed and potentially provide a better SNR.

Figure
4. I would suggest that when the authors introduce the data for Figure 3, they explicitly mention that the samples were measured by DDA.This gets clear in the next section when plexDIA is introduced, but nevertheless it would already be helpful at this point in the manuscript.
5. For the IFNg experiment: the authors mention that different labeling times were applied.It would be good to maybe have the details about the length of the labeling pulse already in the main text and not just Materials and Methods.
6. Figure legend 3A and 3B -the legend seems to be swapped relative to the figure.Also, in line 346 the sentence seems to refer to Figure 3A not 3B.
7. I am a bit nitpicking here, but the authors use throughout the manuscript the term "protein translation".This term, although often used, does actually not really make sense -it should be either "mRNA translation" or "protein synthesis/production".

REVIEWER COMMENTS
Reviewer #1 (Remarks to the Author): In the manuscript entitled: "An integrated workflow for quantitative analysis of the newly synthesized proteome", Borteçen at al. present a QuaNPA workflow for the quantitative analysis of the newly synthesized proteome (NSP).Although the importance of protein synthesis is well recognized, quantitative analysis of protein synthesis has proven to be challenging.This workflow includes an optimized semi-automated sample preparation protocol and the application of multiplexing DIA-MS to reduce the sample input, increase the proteomic depth, reduce missing values, and improve quantitative accuracy of the analysis.The authors first optimized an automated protocol including the synthesis of magnetic alkyne agarose beads and NSP enrichment and SP3 peptide purification using a Bravo liquid handling system enabling the analysis of 8-96 samples in parallel with a minimized sample input.Then, two benchmark samples were used to compare different MS methods and data processing software for NSP analysis showing a better performance of DIA MS2-based quantification in plexDIA than DDA.Finally, the optimized protocol was applied to study the response to IFNy in HeLa cells.
The main novelty of the manuscript lies in the technology improvement enabling parallel processing of multiple samples and using lower sample input.Furthermore, the NSP analysis depth and quantification benefits from a novel combination with a DIA-based MS method.The manuscript is well written and presents high quality data.My main concern is the novelty of the method application part, since a similar analysis of IFNy signaling has been performed recently (Kleinpenning, 2020).As a benchmark experiment, the analysis of IFNy signaling showed that QuaNPA is able to gain biologically relevant insights by comparing to other omic datasets and annotation databases and lead to identification of potentially novel targets of IFNy signaling.However, to increase the impact of the biological findings of the current study, I would strongly suggest performing follow-up experiments focusing, e.g., on the validation of some of the potentially novel targets of IFNy signaling or extending the study to a more interesting system.
We thank the reviewer for the assessment and comments.Indeed, the main focus of the manuscript lies on introducing and combining the novel technologies and methods.In fact, we applied QuaNPA to IFNg stimulation because it is such a well-characterized system.To follow the suggestions of the reviewer we have performed a targeted proteomic analysis via label-free parallel reaction monitoring (PRM), to validate whether the observed changes in NSP levels of the potential novel targets of IFNg lead to changes in protein abundance upon IFNg stimulation.Using PRM analysis, we were able to confirm differential expression of novel targets BICD2 and SIN3B, following IFNg treatment.Interestingly, and in line with our NSP data (Fig 5E and 6A), IFNg treatment led to an increase in the abundance of SIN3B, while BICD2 was decreased, indicating that QuaNPA can detect expression changes in both directions.Furthermore, we could highlight the quantitative analysis of NSPs is able to measure significant upregulation of canonical IFNg targets such as ICAM1, TAP1 and STAT1 <= 4h, whereas increase in overall protein abundance of these 3 proteins as determined by PRM only reached significance at later time points.These data are included in the revised manuscripts as new Figures 6b and Supplementary   Comments: 1)Page 3, line 84-85: "Although DIA is directly compatible with label-free quantification, it has been scarcely applied in combination with SILAC labelling because of challenges in data analysis 41,42,43.Recently, plexDIA was introduced in the DIA-NN environment, demonstrating deep proteome coverage and quantitative accuracy for the analysis of multiplexed samples with non-isobaric labels (mTRAQ).Conceptually, plexDIA could also be applied to SILAC-labelled samples, in particular for the analysis of NSPs."Another DIA-data analysis software, Spectronaut has been capable of analyzing SILAC-DIA samples for several years, and it has been later optimized for the analysis of pSILAC samples (Salovska et al., 2020) enabling a straightforward analysis of SILAC and pSILAC DIA-MS data.The way this paragraph of the introduction is written gives a false impression that the analysis of such data is problematic and plexDIA offers a unique solution, which is not true.
We thank the reviewer for the comment.Indeed, Spectronaut does enable the analysis of DIA data with SILAC labels.However, we believe that our approach is the first to benchmark the approach, showing that SILAC DIA analysis not only resulted in an increase of protein identifications, but also in quantification accuracy that is comparable to DDA.The study by Salovska et al does not include a benchmark with known SILAC ratios to assess the accuracy of the SILAC ratio quantification by Specronaut, instead carrying out a comparison of MS1 and MS2-based quantification in which they comment on the reduced standard deviation of the MS2-based ratios in pulse SILAC samples with 3 different labelling times.In our hands, data processed with Spectronaut (version 15) resulted in relatively noisy quantification with many outlier ratios that strongly deviate from the known SILAC mix ratios.Yet, we have changed the paragraph to clarify that the use of plexDIA is not the first or only approach for DIA analysis of SILAC samples.
2) Page 4, paragraph starting at line 132: Please briefly clarify in the text and related figure description how the MS data for this optimization step were acquired and analyzed.It is not clear from the manuscript until the next results section.
We thank the reviewer for pointing out the need to clarify the description.The respective paragraph has been edited to include this information.
3) Furthermore, in Figure 3B-D and 3E-F, it is not clear whether the analysis is comparing the same set of overlapping ids identified across all conditions or every boxplot/violin plot represents all protein ids identified in each condition (which would be fair).For Figure 3C, also please clarify whether these are raw or normalized SILAC ratios.In the Method section (line 597) it is mentioned that the normalized ratios were used.The authors might also consider presenting peptide-precursor level identification and ratios, since collapsing the peptides into protein level H/M ratios might remove a lot of quantitative noise.
We thank the reviewer for pointing out the need to clarify the description.Normalized SILAC ratios have been used, indeed as indicated in the Methods section.It is our understanding that MaxQuant performs median normalization of all precursor SILAC ratios for the protein group ratio, thus we do not expect any different results from aggregation of the peptide level data.Indeed, analysis at the precursor level (Review Figure 1) delivers nearly identical results as those at the protein level (Figure 4D, E).Upon suggestion from 1.5x the interquartile range.Values outside this range are plotted as dots and represent outliers.D) Principal component analysis (PCA) of the NSP samples.9) Page 8, line 258-263, Figure S4, comparison to PhosID.When relating to the PhosID dataset, the supplementary figure shows more differentially regulated proteins after 4 hours in the PhosID dataset compared to QuanNPA.It might be fair to comment on this observation too.
We thank the reviewer for the comment.Indeed, a larger number of differentially regulated proteins is reported in the 4 h timepoint of the PhosID dataset.We have mentioned this fact in the edited section of the manuscript, as follows: "However, fewer differentially regulated proteins were detected at the 4 h IFNg treatment time point, compared to data produced with the PhosID workflow."10) Page 8, line 276-277, Figure 6: The two log2FC cutoffs mentioned in the manuscript and shown in the Figure 6 legend are a bit confusing.I would suggest also adding the p < 0.05 and |log2FC| > 1 cutoff into the Figure 6, for example in the grey box next to the label "differentially expressed proteins".Moreover, in the legend of Figure 6, it is a bit confusing why all subgroups are depicted with an asterisk while the asterisk indicate known link to the IFNy treatment as described in the legend text.
We thank the reviewer for the helpful suggestions, and for spotting the error with the asterisk.We have changed Figure 6 accordingly.11) Page 11, line 364-373: The authors hypothesize that plexDIA might have difficulties with quantifying the "extreme" ratios in mix1 due to the fact that DIA-NN does not offer the re-quantify function that is available in MaxQuant.It would be relevant in this section to discuss the fact that another DIA-data processing software, Spectronaut, offers the "inverted spike-in workflow" that has been shown to recover even more "extreme" ratios of H/L pairs, up to 1:16 in a bechmark sample (Salovska 2020 and 2021).
We thank the reviewer for referencing the "inverted spike-in" workflow in Spectronaut.We have included a reference to this proposed solution in the manuscript.
12) Page 15, line 512 and Page 17, line 590 -What is meant by the "global proteome sample"?
We thank the reviewer for pointing out the unclear term.The section of the manuscript has been edited and we now prefer to describe the samples as "proteomics samples without NSP enrichment".
13) Page 17, line 594-597 -It is not clear in which analyses the MBR function was enabled and disabled.Maybe add a link to specific figures here?As I mentioned above, it is also unclear to me in which analyses the raw and normalized SILAC ratios were used, as reported by MaxQuant, and in which the authors used the corresponding light and heavy signals and assembled the proteins using MaxLFQ.
We thank the reviewer for pointing out the need to add further clarifications.Individual links to the respective figures and sections have been added.
For all DDA data, except for the benchmark data (in which unnormalized ratios were compared for all methods), only normalized SILAC ratios as produced by Maxquant are used.For the analysis of DIA data, we apply MaxLFQ normalization across the different SILAC channels, using the "translated quantities" produced by DIA-NN.MaxLFQ normalization was also applied in the original plexDIA publication, but in general produces similar results, to simply averaging precursor SILAC ratios.A modified empirical Bayes moderated t-test, adjusting t-statistic and p-values with precursor counts, was carried out with the "spectraCounteBayes" function of the DEqMS R package.This has now been clarified in the figure legends and in the methods section.
15) Page 27, Figure S5, please add statistical significance of the correlations to the correlation plots.
We have now added p-values to the Pearson correlation coefficient values in the scatter plots.
Reviewer #2 (Remarks to the Author): In the manuscript "An integrated workflow for quantitative analysis of the newly synthesized proteome", Borteçen et al. present an optimized workflow to study newly synthesized proteins (NSPs) which tackles several challenges such as high sample inputs, lack of automation and sufficient throughput, long sample preparation and measurement times, and limited proteome coverage.To support the validity and benefits of their workflow, they apply it to an already well-studied biological system analyzing NSPs in response to IFNg and compare their results with existing data.The authors present a well-written and insightful manuscript with clear figures.The single sample preparation steps follow previously described protocols, but the authors assemble them in a coherent workflow, which they also describe in detail in the method section and which thus appears easy to follow.Given the downscaling and increased throughput of this optimized workflow, this study will likely raise some interest in the proteomics field.However, it remains mainly a method paper with limited new biological insight.
I believe that addressing the following points will enhance the quality and impact of the presented manuscript further: 1.The authors enrich and quantify the newly synthesized proteome and conclude that regulations of these NSP result from a change in protein synthesis.This appears reasonable given e.g. the biology of their IFNg experiment, but this conclusion can still be misleading.Allow me the following thought experiment: Let us assume that newly synthesized and preexisting proteins are degraded at identical rates, i.e. the cell cannot distinguish between a rather new and a rather old protein.This is an assumption frequently made in pulsed SILAC/turnover experiments and a prerequisite of curve fitting to time course pulsed SILAC experiments.Further, let us assume that the synthesis rate of protein A remains the same after a treatment, but its degradation rate increases.In this case, a larger fraction of what is synthesized newly will be degraded at any given timepoint and we would observe a reduced amount of newly synthesized protein A. Following the claims in the paper, however, this situation would always be interpreted as a slowed-down synthesis of protein A. Likewise, if the degradation of protein B was slower and its synthesis rate remained the same, the amount of newly synthesized protein B would increase because a smaller fraction is degraded, which would be interpreted falsely as a faster synthesis.It is a common misconception that pulse experiments can distinguish between synthesis and degradation, although we can only measure the combined effects of both in a pulsed SILAC experiment.I believe this misconception should be addressed.I hope the authors will follow my argument and add a section to their manuscript which should state that, even though they do enrich for newly synthesized proteins, observed changes in NSPs can also come from a change in their degradation.
Bortecen et al. introduce a method called QuaNPA, which stands for quantitative analysis of the newly synthesized proteome, that enables the analysis of newly synthesized proteins (NSPs) in response to cellular perturbations.QuaNPA involves pulse-labeling cells with clickable and stable-isotope-coded amino acids, and uses mass spectrometry and data analysis to quantitatively analyze the NSPs.The study found that QuaNPA provides a powerful approach for large-scale investigation of NSPs, and was able to successfully investigate the time-resolved cellular response to interferon-gamma (IFNg).The main mass spectrometry related principles used in this manuscript have already been introduced significantly earlier and many of them pioneered by the Krijsveld group themselves.Nevertheless, I do think that the authors present very impactful improvement relative to most methods available to study changes in protein production at a global scale and at relatively short timescales.The main advantages are that new kind of magnetic beads are presented with a higher capacity to capture newly produced proteins, an automated experimental workflow to isolate NSPs at high efficiency and also that the authors took advantage of the plexDIA workflow to significantly increase protein coverage without increasing mass spec measurement time.Therefore, this highly optimized workflow to measure protein production changes across many conditions, with a significantly lower sample input amount and at higher coverage and throughput, will be extremely useful to the community (admittedly, not all labs will have a robot available to apply the automated workflow, but many will be able to still use the majority of improvements with standard lab equipment).In addition, the manuscript is generally well written and the experiments very thorough.There are only a few (mostly minor) points that I think the authors should address before acceptance.
We thank the reviewer for the assessment and comments.
1. Generally, some of the optimization/characterization of QuaNPA as shown in Figure 3, should actually be at least partly repeated by doing DIA measurements instead of DDA as DIA does indeed show significantly higher sensitivity and therefore some conclusions drawn from the DDA measurements might not be translated 1:1 to DIA measurements.The most important one is the conclusion the authors draw about the protein input amount that is needed for improving protein coverage.Based on the DDA measurements not more than 100ug total protein input is needed for NSP enrichment as additional amount does not lead to higher coverage (Figure 3F).However, due to the increased sensitivity of DIA relative to DDA this might not hold true for the final QuaNPA approach.Also, for example non-enriched samples show fewer quantified protein groups (Figure 3A) might not be 100% true with DIA.It would be great if this could be assessed by the authors.
We thank the reviewer for the suggestion.We have repeated the protein input dilution series experiment using a DIA method (Review Figure 9).This figure is now included as Suppl.Figure 6 of the revised manuscript.Indeed, the greater sensitivity of plexDIA allows for relatively high numbers of identifications, even with 25 µg protein input for the enrichment of samples.However, similarly to the DDA data, no major gains in the number of quantified proteins and identified precursors is noticeable beyond 100 µg protein input.Indeed, as the reviewer suggests, our DIA benchmark data (Figure 4) show that samples resembling non-enriched samples do not result in an increase in protein identifications to the same extent as samples that mostly contain labelled proteins (i.e.NSPs).
We thank the reviewer for the comment and agree on the importance of precise language.We have changed the sentence to mRNA translation.
I would like to thank the authors for addressing all my comments and suggestions sufficiently.However, I still have a minor point of disagreement with the following claims in the response letter.
"However, we believe that our approach is the first to benchmark the approach, showing that SILAC DIA analysis not only resulted in an increase of protein identifications, but also in quantification accuracy that is comparable to DDA.The study by Salovska et al does not include a benchmark with known SILAC ratios to assess the accuracy of the SILAC ratio quantification by Spectronaut, instead carrying out a comparison of MS1 and MS2-based quantification in which they comment on the reduced standard deviation of the MS2-based ratios in pulse SILAC samples with 3 different labelling times.In our hands, data processed with Spectronaut (version 15) resulted in relatively noisy quantification with many outlier ratios that strongly deviate from the known SILAC mix ratios." While I acknowledge that the presented manuscript performed the DIA to DDA benchmark in a very vigorous way, it is not true that this manuscript is the first one to do so.A previously published study by Pino et al., 2021 already showed that DIA was better in quantitative accuracy compared to DDA, so this specific argument does not really support the novelty of the data analysis approach.
Furthermore, I agree it has been shown that some of the recent versions of Spectronaut do provide a less accurate quantification than DIA-NN (Lou et all, Nat Comm, 2023), but despite imperfections, the software has been used to provide important biological insights in several pulsed SILAC studies.Thus, this argument is also not the best supporting the novelty of the approach and why the previous ones should not be specifically mentioned.
Reviewer #3 (Remarks to the Author): The authors have fully addressed all my concerns and I fully support publication of this manuscript.

REVIEWER COMMENTS
Revision #2: Reviewer #1 (Remarks to the Author): I would like to thank the authors for addressing all my comments and suggestions sufficiently.However, I still have a minor point of disagreement with the following claims in the response letter.
"However, we believe that our approach is the first to benchmark the approach, showing that SILAC DIA analysis not only resulted in an increase of protein identifications, but also in quantification accuracy that is comparable to DDA.The study by Salovska et al does not include a benchmark with known SILAC ratios to assess the accuracy of the SILAC ratio quantification by Spectronaut, instead carrying out a comparison of MS1 and MS2-based quantification in which they comment on the reduced standard deviation of the MS2-based ratios in pulse SILAC samples with 3 different labelling times.In our hands, data processed with Spectronaut (version 15) resulted in relatively noisy quantification with many outlier ratios that strongly deviate from the known SILAC mix ratios." While I acknowledge that the presented manuscript performed the DIA to DDA benchmark in a very vigorous way, it is not true that this manuscript is the first one to do so.A previously published study by Pino et al., 2021 already showed that DIA was better in quantitative accuracy compared to DDA, so this specific argument does not really support the novelty of the data analysis approach.
Furthermore, I agree it has been shown that some of the recent versions of Spectronaut do provide a less accurate quantification than DIA-NN (Lou et all, Nat Comm, 2023), but despite imperfections, the software has been used to provide important biological insights in several pulsed SILAC studies.Thus, this argument is also not the best supporting the novelty of the approach and why the previous ones should not be specifically mentioned.
We thank the reviewer for pointing out the inaccurate phrasing in the first response letter.Indeed, improved quantification in terms of dynamic range was shown in the paper by Pino et al., 2021.We did not mean to claim that our benchmark, and the use of the plexDIA features of DIA-NN, marks the first case of improved SILAC quantification accuracy using DIA data.However, we believe that it is the first reported case of SILAC DIA data with high quantification accuracy, precision and significantly increased peptide/protein identification.
As the reviewer indicates, this concerned a statement in the response letter, which we hopefully addressed now.Yet, in the main text we verified we did not make such claims.Therefore, we believe that the sections describing and discussing SILAC plexDIA in the main text of the manuscript do not give the impression that we claim novelty in regard to improved quantification accuracy using SILAC DIA analysis.This includes citation of papers using other software tools for analysis of DIA SILAC data.
We fully agree with the reviewer, that Spectronaut is a very powerful and useful software, which has helped with numerous important biological findings, including the pSILAC-based studies.We did not mean to include a software comparison in this paper, but are keen on exploring future tests with more recent versions of the software.
legends: Please indicate the test used for the statistical analysis of the results depicted in these figures in the figure legend.Please, also add this information to Figure S4 and S5.15) Page 27, Figure S5, please add statistical significance of the correlations to the correlation plots.

14)
Page 26, Figures 5 and 6 legends: Please indicate the test used for the statistical analysis of the results depicted in these figures in the figure legend.Please, also add this information to Figure S4 and S5.