Message-Id: <9701031914.AA9649@pho903.sbphrd.com>
From: Roland S Annan <Roland_S_Annan@SBPHRD.com>
Date: 3 Jan 97 11:13:08 EDT
Subject: microsequencing by MS/MS
To: Recipients of ABRF List <abrf@aecom.yu.edu>
Ken Williams asked about very low subpicomole sequencing of SDS PAGE separated
proteins by tandem MS.
1. It's difficult and probably not wise to speak for others, but my guess is
that there are relatively few laboratories that can really do what you asked
about, that is sequence peptides from "unknown" SDS PAGE derived proteins at
the very low/subpicomole level. I doubt whether any contract labs could even
come close. While the acquisition of the data can be straightforward, it is
certainly not "relatively easy" If this were true, many more groups would be
doing it. I believe that most, but not all, core protein labs do not have
either the required instrumentation or the necessary mass spectrometry
experience.
The group at EMBL takes these types of projects on a collaborative basis. Our
group here at SmithKline will also take these types of projects on a
collaborative basis. Because we are a small group, and obviously have other
commitments, we have to be somewhat choosy about the projects. There are
perhaps 2 or 3 other protein oriented mass spectrometry labs that could do
this, but at the risk of offending anyone, they should speak up for themselves.
The following comments relate to our experience in this area. We prefer to
work from colloidal Coomassie stained gels. Based on comparative 2D gels, we
believe the faintest colloidal Coomassie spots are about midpoint of silver
stain sensitivity. In our hands, this is the limit of our ability. If we can
even faintly see it on a colloidal Coomassie stained gel, we can get some
internal sequence. Could we get enough sequence to construct a probe? At that
level, I doubt, but we have identified a dozen "unknown" proteins at that level
using database searches of MS derived sequence tags. How much protein is
this? We don't know! For this low level, we add 25 ng of Promega trypsin for
the digest, and see one autolysis peak that is anywhere from 1-3 times the
intensity of the protein peptides. Since we doubt that we get quantitative
cleavage for this trypsin peptide, we estimate we have no more than 25 ng of
protein on the gel. I should also point out, that at this level, we observe
only a few peptides, even for a 60-80 kDa protein, although if we do a
precursor scan for m/z 175 (y1 ion for c-term Arg) or m/z 86 (immonium ion for
Leu or Ile) we see a few more.
We do this work on an ES triple quad (I'm not sure the brand matters these
days) using the nanoES source developed at EMBL. We use MS derived sequence
tags for database searching. Sequence tags can usually be generated in
minutes. In fact we usually do the first sequence tag search while the second
peptide from the digest is being sequenced. De novo sequencing (as in the case
of no database hits) is hard work, takes considerable experience, and almost
always benefits from having several spectra of the same sequence (for instance
two collision energies, derivatized peptide, or MS/MS on low m/z bn or yn
ions).
There will probably always be some need for complete de novo sequencing. For
instance, novel peptide ligands such as small growth factors and phermones,
with MWs under 4000 Da. These would probably not be accessible by genomics.
Ken, if you have any specific questions you can contact me directly or via
this forum.
Roland S. Annan
Research Mass Spectrometry Laboratory
SmithKline Beecham
(Did Murray Hackett really imply that those of us in corporate laboratories
aren't bright enough or motavated enough to do top shelf work? Or did he mean
that being paid a good salary would be a disincentive to doing good work.)