Thank you for a detailed answer on my comments of last week.
I have noted with great interest that all of us aggree that sample cleanu=
p is a
major issue in order to obtain acceptable data once we break the pmol lev=
el. I
think that this is extemely important to investigate this step in a more
systematic manner. I have myself used the technique pionereed by Scott Pa=
tterson
using a miniaturized RP-18 column with some success although the eluent,
containing high organic, have to be dilluted or dried out to be loaded on=
to a
RP-capillary columm. This is of course not ideal. By the way, it is inter=
essant
to note that all sample clean-up methods that I have seen for this small
protein/peptide amounts were designed exclusively for MALDI-Tof or
nanoelectrospray ... Theoretically, in these small amounts, a RP capillar=
y
column should (I say here SHOULD) be more sensitive because the fraction =
where
the peak elutes is smaller in volume (for a 100 =B5m column i.d., a pepti=
de elutes
in approx. 250 nl total, with the 50% of the peak concentrated probably i=
n less
than 100 nl) that it can be retrieved from a step elution fraction. Howev=
er,
capillary LC-ESI-MS remains to be shown to be as consistent as off-line M=
S
technique. I am not despairing yet.
Another word concerning the raging discussion between peptide mass mappin=
g and
MS/MS correlation methods. One issue to be remembered is that these metho=
ds were
developed for different kind of data acquired on different types of instr=
ument.
Peptide mass mapping is especially fitted to take advantage of the high m=
ass
accurancy of MALDi-Tof instruments, at least in the kind of instruments t=
hat use
delayed extraction or the similar. If you can get your job done just with
masses, in particular when you are looking at high throughput and organis=
ms that
are likely to be present in the database, then you should go for it. This=
makes
a lot of sense since acquiring PSD data from peptides is far from trivial=
on
most of these instruments. However, when you dispose of a triple-quadrupo=
le
instrument, like me, it makes a lot of sense to obtain MS/MS data by CID =
since
the mass accuracy of such instruments in full scan mode is probably in th=
e 500
ppm range (you can get higher accurancy, though, but then you have to wor=
k in
the mg range ...). Try to do peptide mass mapping with that ... What I wo=
uld
emphasize here is that each method is more adapted for each kind of instr=
uments.
If your target is contained in the database (not including EST, though, b=
ut
Scott Patterson might correct me on this), I believe that there is an equ=
al
chance to find it by either methods. You just use the one method that fit=
s your
data best.
However, if your target is not in the database, or when you have to look =
at
post-translational modifications or alternate cleavage products, it is
imperative in my opinion to obtain at least a partial amino acid sequence=
, such
as obtained by MS/MS. Here I think that MALDI-Tof are less convenient si=
nce
their PSD capacity is not as efficient as CID in a high pressure collisio=
n cell.
Because my work is focused on protein modifications, I prefer to use a tr=
iple
quadrupole instrument even if this means more trouble and more time when =
I have
to do protein identification of "known" protein. If I start from the assu=
mption
that the sample I am working on is not known, then I want to put all the =
chances
on my side and that means, in particular, that I will try to systematical=
ly
obtain MS/MS data for every peptide that elutes from my column. Now, if t=
he
protein is "known" this might look as "overshooting" it. But if it is not=
, I now
also have, for the same amount of work, MS/MS data that will help me to d=
iscover
what it was. This is for me an acceptable situation since I am not leadin=
g a
core facility, so that, most of the time, I am really dealing with unknow=
ns.
However, when hundreds of sample are streaming in per week, you want to b=
e as
effficient as possible and, if you can eliminate 50% of these samples by =
a rapid
MALDI-Tof peptide mass mapping, because these proteins were known, it is =
worth
your time.
As Roland mentionned earlier last week, I wished sometimes that we had al=
l
possible instruments available that fits best the problem I have to solve=
. For
obvious reason, I only dispose of so much (and not even of a protein sequ=
encer).
So, even when this is not the best possible method, you try it in an alte=
rnate
way, and sometimes you get lucky.
Axel
-------------------------------------------------------------------------=
-------
Axel Ducret, Ph.D.
Senior Research Biologist
Merck-Frosst Canada Inc.
Dept. Biochemistry and Molecular Biology
P.O. Box 1005
Pointe-Claire-Dorval PQ H9R 4P8
Canada
tel. + (514) 428-3428
fax + (514) 428-4900