Reviewer 2's Review

“This paper reports the discovery and characterization of an important reagent-namely a peptide toxin that specifically blocks sustained mechanically-gated currents in mammalian DRG neurons. It will be of interest to physiologists studying somatosensation as well as to researchers interested in selective pharmacological tools for interrogating the contribution of mechanical nociception to pain sensation and somatosensory function. The peptide, which the authors are calling NMB-1, blocks mechanically-gated currents with an apparent affinity of 1µM and may also block mechanotransduction channels in vertebrate hair cells. NMB-1 appears to act specifically on sustained mechanically-gated currents in DRG neurons and the authors claim that it has no effect on voltage-gated channels, TRPV1 channels, or ASIC1 channels.

As written, the manuscript lacks focus and some important experimental methods are missing. Assuming these deficiencies can be addressed, I am very enthusiastic about the paper and the new reagent presented in it. The main problems are:

1. The description of the data presented in Fig. 7 regarding the frequency-dependence of RA and SA current activation is very difficult to follow. Also, it is unclear how this analysis of stimulus encoding is related to remainder of the manuscript which concerns the affinity, site of action, and specificity of NMB-1. Even it the link could be made, the authors stretch the interpretation of these data a bit too far. For example, on p. 9, they claim that the apparent velocity independence implies that "ion channels go to a stretch dependent low conductance state or flicker open at lower rates." Lower rates than what? Also, it is not really possible to draw inferences about single-channel behavior from these macroscopic data. Moreover, it is entirely possible that SA currents are velocity-dependent, but that the velocities interrogated simply saturate the channel activation mechanism for SA currents but not for RA currents. Thus, although this analysis is intriguing, it is incomplete and unrelated to the rest of the manuscript. For this reason, I think that it should be deleted in its entirety and the authors should consider completing this analysis and publishing it as a separate story.

2. Though the authors claim that NMB-1 has little or no effect on voltage-gated Na, Ca and K channels, capsacin-evoked TRPV1 currents, low pH-evoked ASIC1 currents (p. 8) and present summary data in Table 1, they provide no information about how these tests were conducted. Did the authors investigate such currents in cultured DRG neurons? If so, how were each of these ion currents isolated from other currents present in the cells? Or, did the authors test NMB-1 on particular isoforms expressed in heterologous cells? If so, what isoforms and what cells? In the absence of this information it is impossible for the reader to assess the validity of the author's claim of specificity.

3. The authors claim to have tested more than 500 peptide and related molecules in their screen (p. 5). How was the library constructed? Is it commercially available? What is the nature of the "related molecules" and how were they chosen? Additional detail should be provided on how the screen was conducted and, ideally, the identity of the peptides in the pools that failed to block mechanically-gated currents. The description of the screen should include answers to the following questions: How many cells were tested with each peptide pool? Where they classified according the RA, IA, SA scheme used by the authors to subdivide mechanosensitive DRG neurons? If so, how many cells of each type were tested? NMB-1 is described as the most efficient antagonist (p. 5). How was efficiency defined and determined?”

N.B. These are the general comments made by the reviewer when reviewing this paper in light of which the manuscript was revised. Specific points addressed during revision of the paper are not shown.