In Vivo Quantitative Estimation of DNA-Dependent Interaction of Sox2 and Oct4 Using BirA-Catalyzed Site-Specific Biotinylation

Round 1

Reviewer 1 Report

The study by Kulyyassov and Ogryzko is based on the application of  the proximity utilizing biotinylation (PUB) method to analyze interactions of Sox2 and Oct4 in living cells. The authors already proved the usefulness of the PUB approach combined with MS/MS quantitation in their previous reports (including those cited in the manuscript). In the current manuscript they implemented their expertise to quantify interactions between the two transcription factors that play a key role in cell reprogramming. Although the presented manuscript is in general well written and the obtained data are nicely presented, the major message of the whole story is a bit blurred. One may have an impression that there’s no dramatic methodological improvement when comparing to e.g. Kulyyassov et al. 2011 (ref. 20). On the other hand, much is known about direct DNA-dependent interactions of Sox2 and Oct4. Thus, the authors should emphasize the novelty of their work and discuss it in the appropriate context. Going into details, there are a few more specific issues that need consideration:

It seems that the authors analyzed only nuclear fractions obtained after cell lysis. Did the authors checked the purity of these fractions? Would that be possible to estimate nuclear vs cytoplasmic fractions (and their ratios) for all the proteins studied? This might be an issue particularly while considering a transcription factor fused with a quite large (35 kDa) protein. It could be a nice control to see that the proteins would not interact while staying in the cytoplasm. Going further in this direction, it might be that control experiments with GFP are not sufficient, as this protein localizes mostly in the cytoplasm (see e.g. ref. 20) and therefore BAP-GFP would be less efficiently biotinylated by a BirA-fused proteins. More unambiguous results could be obtained with some nuclear proteins (or even transcription factors) used as a control. Was the combination of BAP-Oct4 and BirA-Sox2 also studied via mass spectrometry? Figure 4 shows that it was indeed checked via WB, but Materials&Methods section is not precise in this respect. What proteins could be identified in MS analysis to bind unspecifically (?) to BAP-GFP? The data require some follow-up and comments. Similarly, the list of proteins identified in tables S2 and S4 show proteins with higher number of matches than BAP-Sox2….

Some minor issues are as follows:

Some abbreviations are not expanded by their first use (e.g. BAP) or not defined at all (e.g. FA) Although the Materials&Methods section is in general precisely written and clearly organized, some parts need clarification (e.g. does it make sense to include access path to a local hard drive?; grammar tense rules should always be kept etc.)      

Author Response

It seems that the authors analyzed only nuclear fractions obtained after cell lysis. Did the authors checked the purity of these fractions?

Yes, we analyzed only nuclear fraction using standard protocol utilizing Cytoskeleton (CSK) buffer [1]with 0.5% Triton[2]. The purity of the nuclear fraction can be monitored by light microscope.

Would that be possible to estimate nuclear vs cytoplasmic fractions (and their ratios) for all the proteins studied?

Yes, it is possible also to keep cytoplasmatic fractions and analyze biotinylation level of Biotin Acceptor Peptide (BAP) both in nuclear and cytoplasm.

More unambiguous results could be obtained with some nuclear proteins (or even transcription factors) used as a control.

We agree that alternative control with some nuclear proteins will be more correct. We have already finished experiments to express other nuclear proteins instead of BAP-GFP from available plasmids in our laboratory and prepared nuclear and cytoplasmic fractions. Western blot results of these experiments will be included soon to the manuscript.

Was the combination of BAP-Oct4 and BirA-Sox2 also studied via mass spectrometry?

No, we did not study this combination by mass spectrometry.

What proteins could be identified in MS analysis to bind unspecifically (?) to BAP-GFP?

The lists of proteins in tables S1-S4 are abundant and nonspecifically bound to Ni-Sepharose beads proteins from HEK293T nuclei.

Similarly, the list of proteins identified in tables S2 and S4 show proteins with higher number of matches than BAP-Sox2

In order to identify proteins which bound to Ni-Sepharose we used Shotgun analysis of tryptic peptides using Mascot search engine http://www.matrixscience.com/help/mis_help.html. For samples derived from cells expressing BAP-GFP we used taxonomy filter “All entries”, and for samples BAP-Sox2 taxonomy filter “Homo sapiens (human)”. The goal of this part of work was approximate estimation of relative amount of recombinant proteins in comparison with abundant and nonspecifically bound proteins. If we use more washing steps with Guanidin hydrochloride (Gu∙HCl) we decrease the level of nonspecific binding to Ni-Sepharose. But as we demonstrated two times of Gu∙HCl washing is sufficient to get clear extracted ion chromatograms (EIC) in multiple reaction monitoring (MRM) experiments for estimation of biotinylation level of BAP. The corresponding LC-MS/MS experimental data (mgf.files) are available on open access PRIDE repositorium (identifier PXD015756).

 

 

Saijo, M.; Hirai, T.; Ogawa, A.; Kobayashi, A.; Kamiuchi, S.; Tanaka, K., Functional TFIIH is required for UV-induced translocation of CSA to the nuclear matrix. Molecular and Cellular Biology 2007, 27, (7), 2538-2547. Hymer, W. C.; Kuff, E. L., ISOLATION OF NUCLEI FROM MAMMALIAN TISSUES THROUGH THE USE OF TRITON X-100. J Histochem Cytochem 1964, 12, 359-63.

 

Reviewer 2 Report

The manuscript from Kulyyassov and Ogryzko is an interesting work dealing with the implementation of a method to detect protein-protein proximity utilizing biotinylation.

The manuscript is properly written, techniques used are suited for the study, experiments follow a logical order, and the conclusions are mostly in agreement with the results obtained. Therefore, this study is interesting and suitable for publication since there is an increasing interest in deciphering the complex interactions between macromolecules that determine or modulate their functions.

However, I have some concerns about this work that authors should address before:

 

- Line 19, BAP is not defined.

- At the end of the introduction authors should include the specific objectives of this work related to the biotin assay.

- Line 91. The meaning of DMEM is put at this line but this acronym is used before. It should be put at line 80.

. Line 330. It should be written “Design and features of the system principle”.

- Footnote figure 2. This footnote could be rewritten, or the figure re-designed since as it is now it is not easy to understand, it is not intuitive.

- Figure 4. Numbers in graphs of part b should be explained.

- Table S5. “St error” should be “St. error of the mean”.

- Is it possible to evaluate the range of distance at which the method works? If not, could the authors propose any experiment to evaluate that range? For example, with control proteins with a known distance interaction.

- Have the authors any idea of why there are so many proteins bound to Ni-agarose beads after purification as determined by LC/MS (supplementary tables 1-4). In this sense, is there any reason to discard the use of imidazole to purify his-tag proteins with Ni-agarose beads?

Author Response

Is it possible to evaluate the range of distance at which the method works? If not, could the authors propose any experiment to evaluate that range? For example, with control proteins with a known distance interaction.

We used humanized wild type biotin ligase BirA fusions in all experiments. The wild-type BirA uses biotin and ATP to generate biotinoyl-AMP[1, 2]. BirA holds on to that reactive biotin molecule until it is covalently attached to a very specific substrate called biotin acceptor peptide[3]. Thus biotinylation is a result of direct contact of BirA and BAP parts of recombinant proteins which occurs in case of protein-protein interaction or random collision in vivo.

 

Have the authors any idea of why there are so many proteins bound to Ni-agarose beads after purification as determined by LC/MS (supplementary tables 1-4). In this sense, is there any reason to discard the use of imidazole to purify his-tag proteins with Ni-agarose beads?

The lists of proteins in tables S1-S4 are abundant and nonspecifically bound to Ni-Sepharose beads proteins from HEK293T nuclei. Standard proteomic sample preparation protocol after affinity purification consists of many steps including gel electrophoresis with increased number of fractions for analysis of each excised gel band. On bead digestion is used to reduce sample preparation steps for LC-MS/MS analysis since it is coupled to direct proteolytic digestion of the bound protein content. If we use more washing steps with Guanidin hydrochloride (Gu∙HCl) we decrease the level of nonspecific binding to Ni-Sepharose. But as we demonstrated two times of Gu∙HCl washing is sufficient to get clear extracted ion chromatograms (EIC) in multiple reaction monitoring (MRM) experiments for estimation of biotinylation level of BAP. So we can discard using imidazole to elute His-tagged proteins from beads since it is a redundant step. Moreover, the presence of imidazole will make the tryptic solution more basic and we will need to add more TFA on the next step to adjust pH at 4.

 

Kwon, K.; Streaker, E. D.; Beckett, D., Binding specificity and the ligand dissociation process in the E. coli biotin holoenzyme synthetase. Protein Sci 2002, 11, (3), 558-70. Kwon, K.; Beckett, D., Function of a conserved sequence motif in biotin holoenzyme synthetases. Protein Science 2000, 9, (8), 1530-1539. Beckett, D.; Kovaleva, E.; Schatz, P. J., A minimal peptide substrate in biotin holoenzyme synthetase-catalyzed biotinylation. Protein Sci 1999, 8, (4), 921-9.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

I would like to thank the authors for their revision. The manuscript improved a lot and could be published in its present form.