Protocol for Increasing the Sensitivity of MS-Based Protein Detection in Human Chorionic Villi
, Pavel Pogodin 2, Victor Zgoda 3, Olesya Larina 4, Yulia Kisrieva 4, Maria Klimenko 5, Oleg Latyshkevich 5, Peter Klimenko 6, Andrey Lisitsa 1 and Natalia Petushkova 4Round 1
Reviewer 1 Report
In this article, the authors present a novel purification protocol for the identification of low-abundance proteins in chorionic villi. They propose a method (‘1DE-gel concentration’) for the removal of SDS in short electrophoresis run in stacking gel only, without separation of proteins. Following in-gel digestion, they analyze the peptides in MS and present the results of 6 datasets, 3 treatments each of two tissue sources: elective or missed abortion. The lowest abundance proteins, some of which were not previously detected using MS, correspond with Human Protein Atlas data.
Major:
General remarks
1. The English must be reviewed and corrected (e.g., removal of long and unclear sentences)
2. There are no biological replicates for almost all datasets (see Figure 1). This reduces the reliability. This is also relevant to the rest of the analysis. For example, Table 2:I reliably is defined as a number of proteins identified by ≥ 2 peptides in 2/3 or more replicates in at least 2/3 of samples. However, most datasets (no. 2-6) are based on one sample only. Not clear where ‘2/3 samples’ come from.
3. It is not clear why tissue from all patients in both groups (elective and missed abortions) was not used as biological repeats (‘samples’)? This would make the data much stronger statistically.
4. In light of the goal of the paper: Section 3.2 states that the elective abortion samples were used as ‘healthy’ samples, which implies that the authors are looking to compare healthy and non-healthy pregnancy endings; however, this is nowhere stated as a goal of the paper.
5. It was stated that one of the pitfalls of lysis protocols is mechanical lysis, however, the authors use sonication and manual homogenization in their protocols. Please clarify.
6. Line 200: 32 genes involved in female pregnancy seem to be very little, compared to the number of proteins with this annotation in GO (GO:0007565, including hits in ‘child terms’) that reports of ~4850 proteins. Please explain the low number.
Minor:
• The introduction would benefit from a more detailed background, while there are too many details on the protocol and analysis details (e.g., Lines 15, 16, 27)
• Improving writing is needed, See examples (lines 49-52, 108-114 are very long sentences 153 – 155 unclear meaning
• Line 193: Detection of 468 proteins in successful, dataset 2, is described in the table as 460, not 468.
• Remove repetitions: Description of 189-223 is exactly the numbers in Table 2.
• In abstract: A clear hypothesis/problem statement is lacking although it is clearly described in the introduction. Please revise.
• It’s unclear in Figure 4 what is the difference between "Dataset_Sample_Replicate" format and "Technology (sample preparation protocol), sample number, replicate number" distribution, as described in the figure legend.
• It is not clear what the added value of Fig 4 is over Fig 3. Seems redundant
• Line 340-345: How can NSAF be negative? Explain
• Table 3 legend for ** is unclear.
Author Response
Authors’ Reply to the Review Report
The authors agree with the comments and recommendations from Reviewer and have taken them into account while preparing a revised version of the manuscript.
Comments for Authors:
In this article, the authors present a novel purification protocol for the identification of low-abundance proteins in chorionic villi. They propose a method (‘1DE-gel concentration’) for the removal of SDS in short electrophoresis run in stacking gel only, without separation of proteins. Following in-gel digestion, they analyze the peptides in MS and present the results of 6 datasets, 3 treatments each of two tissue sources: elective or missed abortion. The lowest abundance proteins, some of which were not previously detected using MS, correspond with Human Protein Atlas data.
1. General remarks:
Point 1.1: The English must be reviewed and corrected (e.g., removal of long and unclear sentences)
Response 1.1: We have carefully revised the text of our manuscript and tried to correct the English. Special attention was paid for long and unclear sentences, which were removed or rewritten. Some sections (e.g. Abstract, Conclusions) were rechecked with a native English speaker.
Point 1.2: There are no biological replicates for almost all datasets (see Figure 1). This reduces the reliability. This is also relevant to the rest of the analysis. For example, Table 2:Ireliably is defined as a number of proteins identified by ≥ 2 peptides in 2/3 or more replicates in at least 2/3 of samples. However, most datasets (no. 2-6) are based on one sample only. Not clear where ‘2/3 samples’ come from.
Response 1.2: We concur that the absence of biological replicates for the majority of datasets can reduce the reliability. To overcome it, we proposed the preprocessing strategy of protein records about identification and quantification. The preprocessing permitted to extract comparable protein lists for further statistical analysis. We assume that proposed data preprocessing can be helpful for researchers in circumstances of insufficient funding, low number of samples, difficulties in obtaining and collecting human tissues, and so on.
We clarified that the "2/3 or more replicates" filter was applied for dataset 2, 3, 5 and 6, while all replicates were enrolled in preprocessing of dataset 1 and 4; the "in at least 2/3 sample" filter was applied only for dataset 1, and single sample results were used for treatment with dataset 2-5. Corresponding changes have been made in the text (Note of Table 2, Lines 180-188).
Point 1.3: It is not clear why tissue from all patients in both groups (elective and missed abortions) was not used as biological repeats (‘samples’)? This would make the data much stronger statistically.
Response 1.3: We agree that using tissue from all patients in both groups as biological repeats would improve statistical confidence. The same sample preparation methods were applied for elective and missed abortion samples. We preprocessed datasets of different sample preparation methods and sample origins separately in order to check the variability of protein identification and quantification achieved with 1DE gel-concentration. Downstream statistical analysis showed that sample preparation protocol had a stronger impact on results than tissue origin. So the statistical assessment processing (Wilcoxon test, Lines 275-285) was handled by merging elective and missed abortion profiles. Generally, it is a pilot study with the presentation of 1DE gel-concentration approach. Results may be additionally validated with more biological repeats, technical replicates, and application on other tissues and cell cultures.
Point 1.4: In light of the goal of the paper: Section 3.2 states that the elective abortion samples were used as ‘healthy’ samples, which implies that the authors are looking to compare healthy and non-healthy pregnancy endings; however, this is nowhere stated as a goal of the paper.
Response 1.4: Thank you for your important comment. Indeed, the healthy and non-healthy pregnancy endings comparison was not a goal of the paper. The elective and missed abortion samples of chorionic villi data are treated to demonstrate the sensitivity of 1DE gel-concentration approach. We took this note into account and tried to reflect it more clearly in the text. We have also removed the term "healthy" (Subsections 3.2 and 3.5).
Point 1.5: It was stated that one of the pitfalls of lysis protocols is mechanical lysis, however, the authors use sonication and manual homogenization in their protocols. Please clarify.
Response 1.5: Despite the pitfalls of mechanical lysis, it is often used in combination with detergent-based cell lysis to achieve better protein extraction. We added corresponding clarification to the text of Introduction (Lines 77-79).
Point 1.6: Line 200: 32 genes involved in female pregnancy seem to be very little, compared to the number of proteins with this annotation in GO (GO:0007565, including hits in ‘child terms’) that reports of ~4850 proteins. Please explain the low number.
Response 1.6: Yes, the female pregnancy GO annotation reports ~4850 proteins. However, these proteins include not only fetal proteins (proteins of chorionic villi), but also maternal proteins expressed in other parts of placenta. Additionally, not all genes involved in female pregnancy show expression in the first trimester. The post-translational modifications of proteins may also interfere protein identification.
2. Minor:
Point 2.1: The introduction would benefit from a more detailed background, while there are too many details on the protocol and analysis details (e.g., Lines 15, 16, 27)
Response 2.1: We have revised Abstract and Introduction. In the Introduction, we tried to provide more a detailed background, at the same time we eliminated redundant details about the protocol and analysis from Abstract.
Point 2.2: Improving writing is needed, See examples (lines 49-52, 108-114 are very long sentences 153 – 155 unclear meaning)
Response 2.2: We tried to correct the writing language to provide more understandable sentences and clear meaning. The long constructions were revised for conciseness.
Point 2.3: Line 193: Detection of 468 proteins in successful, dataset 2, is described in the table as 460, not 468.
Response 2.3: Thank you, corrected (deleted in the revised version).
Point 2.4: Remove repetitions: Description of 189-223 is exactly the numbers in Table 2.
Response 2.4: The repetitions were removed, and generally, the description of Table 2 was shortened.
Point 2.5: In abstract: A clear hypothesis/problem statement is lacking although it is clearly described in the introduction. Please revise.
Response 2.5: Thank you for your note. We revised abstract and tried to formulate transparent hypothesis/problem statement.
Point 2.6: It’s unclear in Figure 4 what is the difference between "Dataset_Sample_Replicate" format and "Technology (sample preparation protocol), sample number, replicate number" distribution, as described in the figure legend.
Response 2.6: Thank you for pointing this out. There is no difference between "Dataset_Sample_Replicate" format and "Technology (sample preparation protocol), sample number, replicate number" distribution, as was described in the Figure 4 legend. Sorry, the error was made. Now it is Figure 3, we corrected description, Lines 244-248, 269-270.
Point 2.7: It is not clear what the added value of Fig 4 is over Fig 3. Seems redundant
Response 2.7: Agree, Fig 4 and Fig 3 showed overlap. We decided to exclude Figure 3 (heatmaps) and leave Figure 4 (tanglegrams) because we believe that tanglegrams are more straightforward and reader-friendly. Tanglegrams show relations between datasets and, at the same time, severe dependencies between LC-MS/MS analysis results and classification by sample preparation protocol, sample number and replicate number.
Point 2.8: Line 340-345: How can NSAF be negative? Explain
Response 2.8: We centered NSAF values for each dataset, i.e. the null value was moved: NSAF1 + …. + NSAFn = 0. The explanation was added to the text (Line 306).
Point 2.9: Table 3 legend for ** is unclear.
Response 2.9: Agree. The appropriate correction was made (Lines 323-324).
In conclusion, we would like once again to thank Reviewer 1 for throughout review and valuable comments.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Without talking about Urea, SDS-Gel and time-saving method, it is notvery clear from this version about the main goal of the work, if that
deals with a new breakthrough in current protein biochemistry (see
introduction), a new efficient method in analysis of human chorionic
villi (see Abstract) or both. The objective of the work needs to be
very clear in the text; Abstract brings attention on chorionic villi
(which are never described really), while Introduction immediately
addresses the problem of analyzing proteins that are expressed in
limited amounts and therefore escape the detection levels of classical
spectrometric (and immunological assay). This refers to “missing
proteins”, which is a term to be explained by the authors, especially
because this is not a new terminology. A missing protein (MP) is an
unconfirmed genetic sequence for which a protein is not yet detected
and it refers to numerous citations in the literature that should also
be cited here (Understanding missing proteins: a functional
perspective…etc)…..I guess the authors tested a method that may help
to declassify MP, i.e. to bring MP in the detection range, and that
they used human chorionic villi to test the efficiency of their
methods. It maybe relevant for Current Issues in Molecular Biology,
assuming that the text is thoroughly revised for clarity (see comments
before) and the quality of figures is seriously improved for
publication.
A figure 1 showing a clear view of microvilli maybe more illustrative
that the sketch on Figure 1-v1.
The quality of Figure 1-v1 per se is not good enough. We hardly see
which band (if any) was cut, and the denaturing and/or electrophoretic
separation on sample 2 is not optimal really (see blue dye front).
Even without fractionation, it would help to run the samples together
with molecular weight markers. It needs to be very clear about the
advantage of the process here, maybe it just removes high molecular
weights from the protein solution, and in that case an Amicon Filter
will just do the job fine enough. What is the advantage of the
author’s method vs Amicon Filter with protein size exclusion? The
protein solubility (in Tris buffer) is an important factor to be
considered before to assign protein to MP. Usually the buffer in use
for detection method (LC/MS/MS and/or HPLC) is the limiting factor. I
don't think the method proposed here help to circumvent this problem.
Then, it seems to me that non-denaturing conditions will be as
critical as SDS, this would help overcome the loss of proteins due to
dilution of samples and to some notice where most of the molecular
weight of identified proteins are not correlated to its m.wt in
SDS-PAGE.
Number of samples and number of replicates per sample across datasets
is never the same (Table 1). Therefore, it would better help to focus
on one protein gene family, one type of MP, to test the method.
This example of MP would show gene expression as measured by real-time
PCR with no clear visualization of the corresponding protein as
illustrated by current methods. The method described here in villi
clearly shows visualization of the selected MP, and this maybe
applicable to most MPs, which maybe critical for approach of MP
diagnostic of human genetic disease.
The link between figure 1 and figures 2-4 is rather unclear. This
could be due to the lack of comprehension in the use of color code or
figure rather poorly legended. Technology sample replicate?
6_1_3? 4_1_2? The color code in figure 1 is not really helpful to
apprehend the rest of figures. Maybe the same color code should be
used from figure 1 to figure 4, this would help understand which
sample in blue on figure 1 is investigated, analyzed in detail and
compared to other sample in green.
The maths in section 3.8 is rather unsuitable. The whole section
should be revised and/or replaced by one example of MP, attesting that
the method is valid for protein marker and disease association study.
Author Response
Authors’ Reply to the Review Report (Reviewer 2)
We express our gratitude to the Reviewer, we found your comments extremely helpful. The authors agree with the received comments and recommendations and have taken them into account while preparing a revised version of the manuscript.
Comments for Authors:
Without talking about Urea, SDS-Gel and the time-saving method, it is not very clear from this version about the main goal of the work, if that deals with a new breakthrough in current protein biochemistry (see introduction), a new efficient method in analysis of human chorionic villi (see Abstract) or both. The objective of the work needs to be very clear in the text;
Thank you very much for your note. The main goal of our work is to present a new "1DE gel-concentration" approach as a prospective tool for uncovering the low-abundant part of the human proteome. We used human chorionic villi for both approving the efficiency of 1DE gel-concentration, and testing the tissue for missing proteins (MP) search. We have revised Abstract and tried to formulate transparent statement of objective.
Abstract brings attention on chorionic villi (which are never described really), while Introduction immediately addresses the problem of analyzing proteins that are expressed in limited amounts and therefore escape the detection levels of classical spectrometric (and immunological assay). This refers to “missing proteins”, which is a term to be explained by the authors, especially because this is not a new terminology. A missing protein (MP) is an unconfirmed genetic sequence for which a protein is not yet detected and it refers to numerous citations in the literature that should also be cited here (Understanding missing proteins: a functional perspective…etc)
We agree with you, Abstract and Introduction need to be very clear and well-coordinated with each other. Therefore we tried to reflect problems of analysing proteins with limited expression in Abstract and provide more a detailed problem background in Introduction. We added the explanation for the "missing protein" term (Line 39) and addressed it to numerous citations in the literature in concordance with your remark (reference number 2).
I guess the authors tested a method that may help to declassify MP, i.e. to bring MP in the detection range, and that they used human chorionic villi to test the efficiency of their methods. It may be relevant for Current Issues in Molecular Biology, assuming that the text is thoroughly revised for clarity (see comments before) and the quality of figures is seriously improved for publication. A figure 1 showing a clear view of microvilli may be more illustrative than the sketch on Figure 1-v1. The quality of Figure 1-v1 per se is not good enough. We hardly see which band (if any) was cut, and the denaturing and/or electrophoretic separation on sample 2 is not optimal really (see blue dye front).
Thank you for your throughout review. We agree that the quality of Figure 1 is not good enough. So the figure was edited and we tried to create a more informative sketch. We added the clear view of microvilli per se and provided the photo of electrophoresis with better quality of each single band (which was further cut) on each electrophoretic run.
Even without fractionation, it would help to run the samples together with molecular weight markers. It needs to be very clear about the advantage of the process here, maybe it just removes high molecular weights from the protein solution, and in that case an Amicon Filter will just do the job fine enough. What is the advantage of the author's method vs Amicon Filter with protein size exclusion?
We concur that even without fractionation, it would help to run samples with molecular weight markers to ensure that high molecular weights are not removed. Amicon purification can provide a high number of particles in the concentrate; however, its implementation may be connected with the loss of proteins [https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5457435/]. Another problem arises due to the importance of choosing the proper size cut-off for centrifugal filters [https://bitesizebio.com/44738/spinning-around-tips-and-tricks-for-using-centrifugal-filters/]. We have matched top 15 proteins with the highest molecular weights in the lists of identifications obtained by different sample preparation protocols (based on SDS or urea-thiourea). The average masses were comparable, and the maximum value was equal to 630 kDa in case of SDS + 1DE gel-concentration, as well as protocol with urea that did not include electrophoresis (i.e. no problems connected to the possible removal of high molecular weights during electrophoresis).
The protein solubility (in Tris buffer) is an important factor to be considered before to assign protein to MP. Usually the buffer in usefor detection method (LC/MS/MS and/or HPLC) is the limiting factor. I don't think the method proposed here help to circumvent this problem. Then, it seems to me that non-denaturing conditions will be as critical as SDS, this would help overcome the loss of proteins due to dilution of samples and to some notice where most of the molecular weight of identified proteins are not correlated to its m.wt in SDS-PAGE.
We agree that protein solubility is not essential to assign protein to MP. However, if the missing proteins are defined according to the global neXtProt database, it was not possible to detect them for the last 15 years via LC-MS/MS. Exactly in that case we proposed to fish out missing proteins employing protein solubility (in Tris buffer or in other buffers). We are very grateful for the comment regarding the discordance of the molecular weight of identified proteins and their molecular weight in SDS-PAGE. We have emphasised that in Conclusions.
Number of samples and number of replicates per sample across datasets
is never the same (Table 1). Therefore, it would better help to focus on one protein gene family, one type of MP, to test the method. This example of MP would show gene expression as measured by real-time PCR with no clear visualization of the corresponding protein as illustrated by current methods. The method described here in villi clearly shows visualization of the selected MP, and this may be applicable to most MPs, which may be critical for approach of MP diagnostic of human genetic disease.
Thank you for your valuable suggestion. In the revised version, we paid special attention on the PSG subfamily. The human PSGs are a group of molecules almost exclusively expressed by chorionic villi during pregnancy. PSG levels have been found to correlate well with placental function and fetal well-being. Low PSG levels are associated with poor pregnancy outcome. We were able to detect all PSG members, including dubious or uncertain protein PSG7. The evidence at the protein level of PSG7 protein had not previously been demonstrated by mass spectrometry. We highlighted the applicability of the described method for the visualization of most MPs in Conclusions.
The link between figure 1 and figures 2-4 is rather unclear. This could be due to the lack of comprehension in the use of color code or figure rather poorly legended. Technology sample replicate?6_1_3? 4_1_2? The color code in figure 1 is not really helpful to apprehend the rest of figures. Maybe the same color code should be used from figure 1 to figure 4, this would help understand which sample in blue on figure 1 is investigated, analyzed in detail and compared to other sample in green.
Thank you for catching our error. The colour codes were not used to trace the connections between figures. We paid attention on this note while revising Figure 1.
The maths in section 3.8 is rather unsuitable. The whole section should be revised and/or replaced by one example of MP, attesting that the method is valid for protein marker and disease association study.
Agree. We revised section 3.8 and removed the math.
Thank you once again for your insightful comments.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
- The paper, including the abstract, conclusions, and results section now better reflects the author's logic and goals.
- Various points were greatly improved in the current manuscript.
- English editing is still required.
Author Response
Authors’ Reply to the Review Report
We express our gratitude to the Reviewer, we found your comments extremely helpful. The authors agree with the received comments and recommendations and have taken them into account while preparing a revised version of the manuscript.
Comments and Suggestions for Authors:
The paper, including the abstract, conclusions, and results section now better reflects the author's logic and goals.
Various points were greatly improved in the current manuscript.
Point 1.1: English editing is still required.
Response 1.1: We took into account your comment and sent the text to English editing (by MDPI Language Editing Service).
Point 1.2: The main comment remains the issue of biological repeats. The authors explain that the statistically significant clustering together of the samples according to protocols demonstrates that the physiological source of the samples is of lesser importance than the sample preparation protocol. However, in their explanation, the authors do not take into account that each sample is taken from a different woman (n=7) according to the Methods section. If I correctly understand, the number of biological samples in total is 8, which may also affect clustering. This notion and the limitations of the presented data should be reflected in the text.
Response 1.2: You have indeed correctly understood that the number of biological samples in total is 8. Samples are taken from 8 different woman, five were obtained from the patients after elective abortions and three were from the patients with missed abortion. Thus the number of biological samples and number of patients is equal. Thank you very much for catching our typo in Materials and Methods section.
A technical point: There is no figure 3, the numbering of legends and references to figures 4 and onwards need to be corrected accordingly.
Response for a technical point: The authors apologize for the figure numeration error. However, by the suggestion of Reviewer 2, the Figure 1 was splitted onto two new figures —- Figure 1 (human embryo and the clear view of chorionic villi) and Figure 2 (the scheme of the 1DE gel-concentration protocol). Thus, the current version includes 4 figures, and the Figure 4 numeration is now correct. The text was double checked accordingly to avoid similar mistakes.
In conclusion, we would like once again to thank Reviewer for the detailed examination of our work.
Author Response File: 
Author Response.docx
Reviewer 2 Report
Minor revisions:
1° The points of comments below should be used to draw a clear point of assessment and discussion in text. This should include the references suggested by the authors or perhaps more references re. protein solubility and Amicon filters efficiency/stability.
Even without fractionation, it would help to run the samples together with molecular weight markers. It needs to be very clear about the advantage of the process here, maybe it just removes high molecular weights from the protein solution, and in that case an Amicon Filter will just do the job fine enough. What is the advantage of the author's method vs Amicon Filter with protein size exclusion?
We concur that even without fractionation, it would help to run samples with molecular weight markers to ensure that high molecular weights are not removed. Amicon purification can provide a high number of particles in the concentrate; however, its implementation may be connected with the loss of proteins [https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5457435/]. Another problem arises due to the importance of choosing the proper size cut-off for centrifugal filters [https://bitesizebio.com/44738/spinning-around-tips-and-tricks-for-using-centrifugal-filters/]. We have matched top 15 proteins with the highest molecular weights in the lists of identifications obtained by different sample preparation protocols (based on SDS or urea-thiourea). The average masses were comparable, and the maximum value was equal to 630 kDa in case of SDS + 1DE gel-concentration, as well as protocol with urea that did not include electrophoresis (i.e. no problems connected to the possible removal of high molecular weights during electrophoresis).
The protein solubility (in Tris buffer) is an important factor to be considered before to assign protein to MP. Usually the buffer in usefor detection method (LC/MS/MS and/or HPLC) is the limiting factor. I don't think the method proposed here help to circumvent this problem. Then, it seems to me that non-denaturing conditions will be as critical as SDS, this would help overcome the loss of proteins due to dilution of samples and to some notice where most of the molecular weight of identified proteins are not correlated to its m.wt in SDS-PAGE.
2° I still believe that there is a figure 1 missing. Maybe enlarged illsutrations about human embryo and microvilli and move them to Figure 1. Figure 2 now looks much better. I am sure there is a way to remove the two too flashy white lines on sample n. Please move embryo and microvilli to Figure 1. Start Figure 2 with Eppis. Enlarge Figure 2. Then, there is one figure missing (Figure 4 -reduce the character size- follows figure 2, presently no figure 3. Text should be double checked accordingly.
3° There is this (60) in bold should not be in bold (line 384). The whole text should be double checked for logical order and typo mistakes. See line 385, 369-370, 366, 328-336, etc
4° I would delete the references from Conclusions section. The points with references should be provided in Discussion.
Author Response
Authors’ Reply to the Review Report
We express our deep gratitude for Reviewer for the attention to our manuscript and for valuable comments. The authors agree with the comments and recommendations of the Reviewer and have taken them into account while preparing a revised version of the manuscript.
Minor revisions:
Point 1.1: The points of comments below should be used to draw a clear point of assessment and discussion in text. This should include the references suggested by the authors or perhaps more references re. protein solubility and Amicon filters efficiency/stability.
Even without fractionation, it would help to run the samples together with molecular weight markers. It needs to be very clear about the advantage of the process here, maybe it just removes high molecular weights from the protein solution, and in that case an Amicon Filter will just do the job fine enough. What is the advantage of the author's method vs Amicon Filter with protein size exclusion?
We concur that even without fractionation, it would help to run samples with molecular weight markers to ensure that high molecular weights are not removed. Amicon purification can provide a high number of particles in the concentrate; however, its implementation may be connected with the loss of proteins [https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5457435/]. Another problem arises due to the importance of choosing the proper size cut-off for centrifugal filters [https://bitesizebio.com/44738/spinning-around-tips-and-tricks-for-using-centrifugal-filters/]. We have matched top 15 proteins with the highest molecular weights in the lists of identifications obtained by different sample preparation protocols (based on SDS or urea-thiourea). The average masses were comparable, and the maximum value was equal to 630 kDa in case of SDS + 1DE gel-concentration, as well as protocol with urea that did not include electrophoresis (i.e. no problems connected to the possible removal of high molecular weights during electrophoresis).
The protein solubility (in Tris buffer) is an important factor to be considered before to assign protein to MP. Usually the buffer in usefor detection method (LC/MS/MS and/or HPLC) is the limiting factor. I don't think the method proposed here help to circumvent this problem. Then, it seems to me that non-denaturing conditions will be as critical as SDS, this would help overcome the loss of proteins due to dilution of samples and to some notice where most of the molecular weight of identified proteins are not correlated to its m.wt in SDS-PAGE.
Response 1.1: Thank you very much for this suggestion. We used the proposed comments and references for the discussion in the revised manuscript (Lines 317-328).
Point 1.2: I still believe that there is a figure 1 missing. Maybe enlarged illsutrations about human embryo and microvilli and move them to Figure 1. Figure 2 now looks much better. I am sure there is a way to remove the two too flashy white lines on sample n. Please move embryo and microvilli to Figure 1. Start Figure 2 with Eppis. Enlarge Figure 2. Then, there is one figure missing (Figure 4 -reduce the character size- follows figure 2, presently no figure 3. Text should be double checked accordingly.
Response 1.2: We agree that it is better to split the Figure 1 and have taken your advice. The Figure 1 is now an enlarged illsutration of human embryo and microvilli. The Figure 2 starts with Eppis. The revised manuscript includes Figure 3 (intersection, Figure 2 in previous version), and the Figure 4 numeration is now correct. We checked the text for similar mistakes and font problems.
Point 1.3: There is this (60) in bold should not be in bold (line 384). The whole text should be double checked for logical order and typo mistakes. See line 385, 369-370, 366, 328-336, etc
Response 1.3: Thank you for pointing this out, authors apologize for the technical errors. We double checked the text to avoid typo mistakes. The (60) reference is now not in bold. We have changed the font of protein symbols according to the National Center for Biotechnology Style Guide.
Point 1.4: I would delete the references from Conclusions section. The points with references should be provided in Discussion.
Response 1.4: Agree. We excluded references (3,4), and moved the reference (73) to the Results and Discussion section (Lines 288-290).
In conclusion, we would like once again to thank Reviewer for the throughout review and valuable comments.
Author Response File: Author Response.docx
