Reply to Ramos-Silva et al.: Regarding coral skeletal proteome.

We thank Ramos-Silva et al. (1) for their thoughtful comments on our work recently published in PNAS (2). We agree that careful cleaning of biomineral samples is indeed necessary for appropriate proteomic analysis. However, we respectfully disagree with their interpretation of our cleaning methods and our decision to include certain proteins in our final list of 36 potential biomineralization proteins. As stated in the Methods of our report (2), Stylophora pistillata skeletal fragments were soaked in 3% (wt/vol) commercial bleach for 4 h, thoroughly rinsed, dried, and finely ground. The 3% (wt/vol) would be∼50% (vol/ vol) of the commercial bleach stock. Although we did notmention it in our report, only powder that passed through a 150-μm sieve was retained [compared with Ramos-Silva et al.’s 200-μm sieving (1)]. This fine powder was again bleached, rinsed, and dried as above. Therefore, by both bleach concentration and skeleton grain size, our cleaning method is more extensive than Ramos-Silva et al.’s complete method, although we acknowledge that they performed the first bleaching (on unground skeleton fragments) for a longer time. Additionally, whereas Ramos-Silva et al. observed “residual soft tissue” in sectionedAcropora millepora skeleton fragments after the first bleaching step (before grinding) (1), we did not observe any tissue or cells in S. pistillata skeleton fragments (Fig. 1) and did not obtain an organic signal by energy dispersive X-ray spectroscopy (EDS) (Fig. 1, Inset). We also used SDS/PAGE to confirm that proteins were not present in SDS-soaked S. pistillata sub– 150-μm skeleton grains (Fig. 2). Ramos-Silva et al. (1) correctly suggest that certain cellular proteins should be considered as potential coral tissue contamination in insufficiently cleaned skeleton samples. Histones, as nuclear proteins, are obvious examples and were observed in “simple bleaching” but not “extended bleaching” samples by those authors (1) but not in our study. However, other cellular proteins, such as actins and tubulins, should not be ruled out as “contamination” without additional research. Although the topic of vesicular transport of Ca and organic matrix macromolecules remains to be resolved in corals (3), inhibition of cytoskeletal (i.e., actin and tubulin) (see Table 1 in ref. 1, proteins 1, 3, and 7) polymerization lowers skeletal organic matrix incorporation into coral skeleton (4), and actins and tubulins are known to be involved in biomineral formation in other organisms [for example, in diatoms (5)]. Finally, although we stand behind the proteins derived from S. pistillata skeletons and sequencedbyLC-MS/MS,weacknowledge that we made editorial judgments in giving names to proteins that showed potential homologs in GenBank. We hope that further study of coral skeletal organic matrix proteins by us, RamosSilva et al. and others will expand the coral biomineralization “toolkit” proteins and elucidate their precise mechanisms by methods other than simply sequence similarity.