Chang et al. Vol. 138, No. 5, September 8, 1997. Pages 1149–1157.

Due to a typographical error, “aa 54–200” was mistakenly printed as “aa 154–200” in the second paragraph of Results on page 1151. The correct sentence appears below in its entirety.

In a yeast two-hybrid assay, both the original partial cDNA (aa 54–200) and the full-length cDNA exhibited comparable interaction with the β1 integrin cytoplasmic domain but not with unrelated baits (Table I).