Based on feedback from ABRF members, ABRF-92SEQ was designed with two particular goals: to provide facilities with enough sample (500 pmol) for sequencing as well as additional analyses, and to incorporate post-translationally modified amino acids. The amino acid sequence of ABRF-92SEQ was:
ARGUAGGSAKIJQQDRHPAVLTNGMGSEQEAKIHNFAK
where U= enzymatically phosphorylated serine and J= hydroxyproline. The effects of some standard problem regions such as doublets (G6G & Ql2Q), a hydrophobic region (A18VL), and a deamidation site (E27QE) on sequence assignment were also evaluated. Alanines were included for repetitive yield calculations. Sites for CNBr, trypsin, endoproteinase V-8 and endoproteinase Asp-N cleavage were also incorporated in the design.
This year the participating facilities (74/225=32.9%) determined longer sequences with a high accuracy of positive calls (93.6%). Accuracy of ABRF92SEQ was comparable to those of STD-1 and ABRF-89SEQ, which were also samples with high peptide content. Interestingly, there was no correlation between the amount of sample used for sequencing and the accuracy of positive calls. The differences in accuracy from different models or makes of sequencer were also statistically insignificant.
The amino acid which was most frequently correctly identified was Ala at residue 5 and 9 (100%). Phosphoserine was the most difficult amino acid to identify with an accuracy of 13.9% (6/43). It was most often miss-identified as Ser, followed by Cys. Phosphoserine was correctly identified by the procedure of Meyers et al [FEBS Lett. 204:61-66 (1986)] or by the ratio of serine to dehydroalanine. Hydroxyproline was less of a problem: nearly 46% (34/74) of the facilities used Edman degradation to identify this residue with 77% accuracy (34/44). It was most often miss-identified as His (8/74).
There were two correct sequence identifications for ABRF-92SEQ. One facility (#178) assigned 36 residues as positive correct calls and residue 4 was a tentative correct call. A second facility (#291) had 35 positive correct calls and residues 4 and 37 were tentative correct assignments. Another facility (#361) assigned 36 residues correctly and assigned residue 4 only as serine. Two other facilities had all correct except for residue 4, they miss-identified it as either serine (#206), or dehydroalanine (#186), but each overcalled the peptide by one residue. The most complete analysis was reported by facility #178, with both sequence analysis and mass spectrometry used to identify the modified amino acids, and reduction/alkylation with 4-vinyl pyridine to rule out cysteine at residue 4.
The extent of additional analytical techniques performed on ABRF-92SEQ was quite low: 34.6%, 26.6%, and 5.6% used one, two or three additional techniques, respectively. These were often amino acid analysis (41 %) or cleavage for the generation of peptides (21%). Cleavage techniques were cyanogen bromide (7), trypsin (4), endoproteinase Asp-N (2), endoproteinase Lys-C (1), and ammonium hydroxide (1). Facilities used either RP-HPLC (15/74) or CZE (2/74) for separation (10/74, 1/74) or analysis (5/74, 1/74, respectively). The availability and use of mass spectrometry increased, with 29.3% of the facilities using this technique. Three mass determinations were within 3 mass units.
To facilitate identifying modified PTH-amino acids, the ABRF is presently compiling data on chromatographically identified modified PTH-amino acids [ABRF News (1992) 3:21. It would benefit all ABRF core facilities to contribute to this project so that sequence identification of modified amino acids becomes the norm, rather than the exception, for protein sequencing core facilities.
Return to the The ABRF Home Page