Protein Scence

Aug. 5, 2005

Dear Dr. Hartley:

I have had your above-referenced manuscript evaluated by two referees whose assessments have been returned. To view the reviews of this manuscript, please visit and your "Author Area". You will find that both reviewers have serious concerns about the work and the manuscript although both are also encouraging that the work is important. The reviews point out weaknesses in the presentation and that analysis of only barstar is too limiting. I am sorry to have to tell you that in the absence of a strong endorsement by either referee it will not be possible to publish your article in Protein Science. In the face of these two assessments I have no alternative but to return your manuscript with the hope that the comments provided by the referees will be of some value in your further development of these studies.

Yours sincerely, Sincerely, Carol Post


Reviewer 1 Comments for the Author ...

Hartley presents an interesting and important study, which is likely to be of interest to many readers of Protein Science. Unfortunately, however, in its present form the manuscript is not suitable for publication. There are two main problems:

(1) The paper summarizes the results of experimental studies, but the experiments themselves are not fully presented. The manuscript summarizes the experiments as they would be discussed in a Review Article. The text does not contain a true "Results" section, nor are any of the data (except the sequences) presented in figures or tables. The author states that most of the methods were presented in a previous paper (sometimes cited as Hartley 2001, other times as Hartley 2004). Previous publication of the methods may indeed obviate the need for a detailed description of methods in the current manuscript, but a detailed description of the Results is necessary. Publication of experimental results in Protein Science requires presentation of the data.

(2) The manuscript, in its current form, ignores much of the earlier work on this topic. Although a few earlier studies (e.g. Axe et al., Gassner et al., and Matsuura & Pluckthun) are cited, most of the seminal work in the field was ignored. For example
(a) Six lines into the introduction, the text speculates on the possible design of new folds not found in nature. This has been achieved, as was described by David Baker and coworkers in a paper published last year in Science.
(b) Further down on the same page, the text discuses the possibility of using libraries of novel protein sequences in which the interior side chains are hydrophobic and the exterior side chains are hydrophilic. Michael Hecht has been doing this for a decade. His work was summarized in a Review Article published last year in Protein Science, and is described in detail in the papers cited in that Review. (Others have used Hecht's approach to select new sequences for natural proteins. For example, see Hilvert's work on chorismate mutase.)
(c) The bottom of the first page suggests using phage display to isolate proteins with novel sequences by holding constant certain surface epitopes. This approach was used by Baker and coworkers in the mid 1990's.
(d) The alternative approach described in this section -- selecting for protease resistance -- was pioneered several years ago by Dek Woolfson, and also by Andreas Pluckthun.
(e) The last page of the text discusses how the mutants in this paper can be analyzed by considering the sizes of the side chains and how they fill space. This kind of analysis - and the overall idea of repacking protein interiors - was pioneered by Wendell Lim and Robert Sauer in 1989 & 1990.

In summary, the work outlined in this manuscript sounds interesting and exciting. However, to merit publication in Protein Science, the presentation of the work must be more complete. This referee is excited by the approach described in this manuscript, and would be pleased to referee a more complete version of the manuscript.


Reviewer 2 Comments for the Author ...

This short manuscript describes experiments to investigate the factors that govern the correct/satisfactory interactions in the non-polar core of the enzyme inhibitor, barstar. It also puts the issue of the non-polar core packing in the context of the number of observed and possible protein folds, which is nice. The method used to screen for proper folding is appropriate, It is the capacity for binding to the cognate enzyme barnase. The research connects to the results on barnase's central non-polar core by the Fersht group in which the constraints on the non-polar buried residues were subjected to similar analysis.

The barstar results are interesting, like those on barnase they demonstrate a degree of plasticity in the substitutions implying in turn that the interference of the non-polar contacts on the main chain fold (providing they are positioned appropriately) is limited and accepting of the influence of the remainder of the barstar sequence. The author points out that the requirement for barnase binding resides in the helix and its adjacent loop, their local conformation is not necessarily going to depend too much on the interactions in the non-polar core. Clearly it is important to investigate the core structures of well-characterised proteins and barstar is a natural candidate for focussing on the influence of the buried non-polar residues on the overall structure. However doing these is just a beginning, I suspect that the structure and interactions of the main chain will in many cases dictate the tightness of the sidechain packing in the non-polar core and that there will be a range of plasticities found in protein cores. While these findings for barstar are not very surprising it is important to know the sequence-folding relationships in proteins in this kind of detail, The volume of the core and the buried non-polar sidechains is usefully highlighted and likely to be a critical parameter.

The analyses the mutation and binding results come up with numbers. Here I have a problem, I just did not follow the arithmetic of the "functional" barstar sequences. There are 20 listed in Table 1. The text states there are 22 core residues and 20 distinct sequences identified. However 33 genes were sequenced,11 appeared twice and 3, three times. I cant make the 20 distinct sequences in Table 1 from this.

I realise I have missed something but I am also unable to follow how the total number of 22aa core sequences of 2.4x10*15 implies 10*8-9 functional inhibitors. These technicalities should be clarified for the inexpert, especially since research on the folding, stability and conformation of proteins is of general consequence.

In conclusion I recommend that the paper should be published since the results are interesting, the experiments have been carried out most competently and they add to the picture presented by barnase; but the author should clarify the numbers he uses and he might also want to consider the role of the rest of the sequence in determining the main chain's structure even in a protein as small as barstar.

Return to "The fold of barstar and the composition of its hydrophobic core."