VOLUME 291 (2016) PAGES 20085–20095
The authors regret to report that they have been unable to repeat the observed effects of glycosylation on dimerization of the IgV domain of CEACAM1. Recently prepared samples show NMR line widths more consistent with dimer molecular weights, and an analytical centrifuge determination of a dissociation constant of 2.5 μm (unpublished, courtesy of John H. Kim). The authors are unsure of the origin of the change in dimerization tendency. The original NMR data were properly acquired and analyzed; the appearance of a monomer in a glycosylated sample was, in fact, backed up with multi-angle light scattering data (not included in the publication at the time). However, the authors have recently observed that glycosylated samples have some variation in the extent of glycosylation and proteolysis of the amino terminus. Whether this type of variation or some other inadvertent chemical modification could have been the cause of dimer inhibition, the authors do not know. The authors believe it important to let the scientific community know about the incorrect statement regarding glycosylation effects on CEACAM1 dimerization. The remainder of the data and conclusions in the original paper, including the determination of the non-glycosylated dimer structure, remain valid.