Re: High Sensitivity Protein Identification

David Arnott (arnott@gene.COM)
Fri, 23 Jan 1998 16:20:58 -0700


Ken Williams asked about high sensitivity MS or MS/MS for protein
identification at or below the amol level. The sensitivity question
aside, peptide maps from either MALDI-TOF or ESI-quadrupole or ion trap
spectra can be used to search sequence databases. Tandem MS by CID on
triple quads or ion traps, or PSD on TOF instruments provides
structural information which greatly improves search specificity. When
dealing with proteins from "sequenced" genomes, peptide masses may be
all you'll ever need. In other cases, especially if mixtures of
proteins may be present, some MS/MS data will probably be needed for an
important percentage of samples.

Now the hard part... Most people in the buisness are of the opinion
that MALDI-TOF is the most sensitive approach to getting peptide mass
maps, with good spectra fairly easy to get from low fmol amounts of
peptide deposited on a probe. I would suggest that micro-scale ESI
coupled to capillary LC or CE has comparable sensitivity, although
with admittedly more effort. In fact MS/MS using an ion trap can
generate excellent spectra from a few hundred amol (perhaps less) of
peptide injected on a micro capillary column. A few (out of many)
references are:
McCormick et al. Anal. Chem. 1997 (69) 767-76.
Wilm et al. Nature 1996 (379) 466-69.
Figeys et al. Anal. Chem. 1996 (68) 1822-28
Gevaert et al. Electrophoresis 1996 (17) 918-24.
Cox et al. Science 1994 (264) 716-19

Now the harder part... Many discussions on sensitivity focus on the
amount of sample introduced to the mass spectrometer. When someone
says that they identified a protein from x amol of sample this is
almost certainly what he or she is talking about. The important
figure, however, is the amount of protein, presumably starting in a
complex mixture in some precious source, which is needed for the whole
procedure, start to finish. My impression is that many labs now work
with proteins starting with a few pmols in a polyacrylamide gel.
Several labs, including some cited above, have reported routine
identifications at or below the 1 pmol level. My experience is that
samples in the hundreds of fmols are indeed tractable. I'm not aware
of anyone, however, who claims to identify proteins at the truly low or
sub-femtomole level on a day-in, day-out basis. It seems to me (and
may occur to grant reviewers also) that mass spectrometers are not
currently the limiting factor for sensitivity, and that the biggest
gains are to be made in the area of protein isolation, detection, and
sample handling.

I hope this helps, and good luck

David Arnott
Protein Chemistry Dept.
Genentech, Inc.
arnott@gene.com
(650) 225-1240