If one examines the predicted signal sequence cleavage using the in-frame Met that is encoded nine amino acids upstream, SignalP now predicts cleavage between the Ser and Gln that we observed in our studies (Figure S3).