In the present manuscripts the authors aim at characterising the chaperone trigger factor (TF) regarding the dimerisation of TF and its implication on interaction with substrates. If the topic is interesting and trying to answer biologically relevant questions, the overall quality of the paper appears poor and too preliminary for publication.

1. There is a lengthy discussion about TF assignments. As presented here, RDB and PPI domains were assigned previously by the authors, SBD and TF113-432 by Dyson’s group and the full length by Kalodimos and Hiller groups. It thus seems to me that all this constructs were assigned before. Please clarify this section and state explicitly your contribution to the field.

2. For characterising the dimer interface and dynamics, the authors use a combination of NMR line-widths analysis and EPR. This is an interesting approach, however the obtained data remain low resolution and do not allow for a clear description of the dimer properties. This section is thus not very conclusive and as the authors have all the tools in hand to label with MTSL each domain of TF, I am wondering why they do not prepare mixture of isotopically labelled TF with each of the MTSL constructs successively to answer which domains of TF are in fact interacting or in close proximity. Extra EPR measurements could be used to derive informations about the domain even if more distant than the PRE distance range. 

3. In the section regarding interaction with peptides, the authors aims at verifying the quality of a theoretical model model regarding TF-target interactions. This is an interesting question however the current approach lack rigorous testing. In fact the selection of the 5 constructs remains unclear to me. For exemple why is the site around 290 not chosen? It seems from the figure to be a potentially better site than 5. Also no testing on a site which is predicted to not or poorly interact is done.

4. In the comparison of the interaction between IcdH2 and IcdH3, in the actual format it is almost impossible to assess the quality and relevance of the data and analysis. The PRE figures are extremely hard to read. Please adjust those figures, possibly with a multiple panel organisation so that the data readable and easily comparable between the different considered systems/probes. Please also indicate where each domain is in the sequence. For exemple, when the authors indicate that “the loss of PRE was much more pronounced for IcdH3 compared to that of IcdH2”, I couldn’t find any quantitative comparison or direct comparison of the experimental data. The conclusion of this section regarding TF dimerisation seems to me quite speculative regarding the current data. It might be possible however to better reach those conclusions if the data where more adequately presented.

5. In the cross-linking section, the authors aims at distinguishing the effect of TF dimerisation on substrate recognition. In Fig 6, the gel used to control the cross-linking data show a band at 50kDa for TF in absence of BS3, while in the presence of BS3 the band is a blurry broad band at very high molecular weight (>130kDa and much higher). This indicate that the form obtained by cross-linking is not a dimer but probably a set of oligomers of TF (with 3, 4 or more TF units). To test the role of dimerisation in the interaction process the cross-linking should have given a dimer and not an oligomer. To me, this render this part of the study inconclusive. 

Considering the points above, I consider it difficult to obtain strong conclusions from this study. In the current state I do not see the added value of the current study for the state of the art of the research in the field. The last figure is quite striking to me, only the (b) part seems to be a novel contribution from the group which I believe is lagging behind current structural studies of TF already existing. I would thus not recommend this paper for publication.

Additional extra minor points

1. For readability I would strongly encourage the authors to homogenise the nomenclature regarding TF constructs and domains.

2. Figure S3 is missing in the file I could download

3. Figure 3 could be moved to the SI

4. Some references are incomplete regarding, e.g. doi numbers

In this study, the authors have measured the structural plasticity of TF domains by carrying out the ESR studies with spin-label placed at different domains in the protein. T2 relaxation is also measured to highlight the domain dynamics. Fluorescence polarization studies were done to measure the apparent binding constant of TF with five peptides labeled with FITC. The binding affinities were correlated with TF binding scores. PREs were measured on various TF variants bound to two peptides. Crosslinked dimer of TF was tested for multiple site binding to urea-denatured MBP to propose that dimerization leads to sequestering of binding sites available in SBD. Although the manuscript discusses an important aspect of protein folding, and the work is carried out meticulously, but it lacks an explanation of various points mentioned below and I, therefore, recommend the manuscript for major revision.

Major comments:

1. The basis for choosing the positions, where four amino acids (which ones?) are mutated to Cysteine, in TF for the spin-labeling is not mentioned in the manuscript. Also, are there any structural perturbations due to Cysteine mutations?
2. Figure 2b shows FP data of TF binding to five peptides. Some of the binding curves do not go all the way to saturation. How reliable is such a fitting to estimate Kd values?
3. What are the structures of the peptides used for binding? Is there any correlation between structure of the peptide and the binding affinity to TF? It is important to look at this aspect as it will shed light on selective recognition of client proteins by TF.
4. It is mentioned in the manuscript that an ideal TF binding motif should be at least eight residues long, and rich in aromatic residues and positively charged lysine or arginine. But that combination may give rise to a gigantic number of peptide sequences. What is the basis for choosing the peptide sequences used in the study?
5. What is the structural evidence that the crosslinked-dimer of TF sequesters the binding sites within SBD and the change in Kd (with respect to Native TF) is not due to loss of binding sites due to some other perturbation in structure originated due to crosslinking?
6. Authors proposed a model where a long nascent chain can be occupied by multiple TF molecules, however, this would be strongly dependent on peptide sequence as the number of favorable binding residues, the 3D structure of the peptide chain, the folding kinetics would all dictate the binding to TF and is dependent on primary sequence. No experimental data has been provided to support the claim and the statement is highly speculative.

Minor comments:

1. A domain map for the protein (in SI or in main) would have been helpful in understanding the individual domain lengths and relative positions.
2. Line 27: Spelling mistake for ‘isomerization’
3. Line 47: Sentence needs to be corrected. It reads: ….corrected sorted by….
4. Line 112: Missing word after ‘eight-channel’.
5. No gap between units and corresponding numbers at several instances in the manuscript.

This work studies the chaperone trigger factor (TF) by NMR and EPR spectroscopy. The study addresses two topics. (1) What is the structure of the trigger factor dimer? (2) How do certain peptides interact with TF? Both these questions are of high biological / biophysical impact and have been previously addressed by other studies, as correctly cited by the authors. While the present study is following those previous works, it has nonetheless the potential to contribute valuable information on both topics and should therefore eventually be published. It requires however major revisions.

On the one hand, I support the technical issues raised by the other two referees, without wanting to rephrase them here. These should be addressed.

On the other hand, the authors should substantially strengthen the interpretation and discussion part of their data such that additional insights can be gained. Contrasts and communalities to prior work are to be spelled out explicitly.

Regarding topic 1, the structure of the TF dimer has been studied previously by two independent studies (Morgado et al. Nat Comm 2017 and Saio et al. eLife 2018). Interestingly, while both studies identified the same global arrangements of domains, they came to opposite results regarding the dynamics of the complex. The Morgado study comes up with the finding that the dimer forms a multi-conformational complex. The Saio study finds that TF dimer is a single conformer. The data presented here can contribute to distinguish between the two scenarios. The authors should revise their manuscript to introduce this question in detail and to come up with an analysis as to which scenario is better (or completely) supported by their data.

Regarding point 2, it seems that the peptides bind in a multi-conformational ensemble at multiple sites, as evidenced by the PRE data in Figures 4 and 5. This finding should be discussed with regard to the functionality of the chaperone and contrasted more clearly to the study Saio et al 2014, where other peptides bind in a single conformation.

Minor point:

• The first abstract of the discussion is essentially an introduction. It should be moved to and merged with the introduction.
• Figure 7 is unsystematic, mixing affinities and lifetimes. Also, the present study adds no new insight to this Figure. It should be removed.