Advice needed for Western blotting

[Dr Engelbert Buxbaum]

Am 17.06.2008, 07:06 Uhr, schrieb Peter Ellis <pjie2 from cam.ac.uk>:

> We have peptide antibodies prepared (or in the works) for both proteins  
> - the first step is obviously to characterise the antibodies.  We're  
> using peptide antibodies rather than antibodies to recombinant protein  
> because there are many related genes, and we want to avoid  
> cross-reactivity.

That is fine, just remember that the epitope may be hidden in native  
proteins, thus the AB may not work e.g. in immunoprecipitation. In Western  
you deal with denatured proteins, so that is not a problem.

> We have ORFs cloned for each of the genes and their close relatives.   
> The aim is to use in vitro transcription/translation to generate pure  
> proteins in order to check the antibody specificity.  Once we know that  
> the antibodies do actually detect our proteins of interest (and not the  
> relatives), we can move on to other more interesting questions!

In addition to testing against individual proteins you should test against  
a crude tissue lysate. You should see only a single band. Even more  
telling would be a blot from a 2D-gel (SDS-PAGE after IEF).

>  At this point I'm looking for advice / sanity checking on how to  
> extract my proteins and what sort of gels to use for the Western blots.

The most common method is SDS-PAGE according to Laemmli (Laemmli: Nature  
227 (1970) 680 and Laemmli & Favre, J. Mol. Biol. 80 (1973) 575-599).  
Separation is by molecular mass. For more specific detection, this is  
preceeded by isoelectric focussing, a method that separates by isoelectric  
point. See J. Klose and U. Kobalz: Electrophoresis 16 (1995) 1034-1059

> On looking around in the literature, I'm finding myself a little  
> bewildered by buffer compositions - many papers refer to "standard RIPA  
> buffer" or "standard Laemmli sample buffer", and yet there seem to be as  
> many different recipes as there are laboratories.  Secondly, it looks  
> like the most common buffers simply won't work for my proteins!

It's not as bad as that. The Laemmli buffer is a fairly good bet.

> This one is a novel histone.  I've looked at the QIAGEN QProteome  
> Nuclear kit, but I'm loath to commit to using that as it doesn't tell  
> you what any of the buffers are.  I can't even tell if it's an acid  
> extraction or a salt extraction!  Also, it's very expensive for the  
> amount of tissue you can process with it.

Histones have a high positive charge, to get a reasonable estimate of  
their molecular mass from PAGE, use the positive detergent CTAB rather  
than the negatively charged SDS (E. Buxbaum: Anal. Biochem. 314 {2003}  
70-76)

> So I'd prefer to do my own histone extraction if possible.  I believe  
> that means I want to prepare purified nuclei and then do an acid  
> extraction. Previously I've done nuclear preps when preparing high  
> molecular weight gDNA - will the same buffers be appropriate for protein  
> work?  Protocol is as follows:
>
> Homogenise tissue at low speed in lysis buffer [0.05M TrisCl pH 7.5, 1mM  
> EDTA, 5mM MgCl2, 50mM NaCl, 5% glycerol, 1% Triton X-100, 1% B-ME],  
> using 100 mg tissue / ml buffer.

I'd probably use KCl rather than NaCl, as the cell interior is high in K  
and low in Na. I'd also reduce [Mg] to 1 mM or so to discourage  
metal-dependent proteases. The Triton I'd leave out at this step and  
reduce bME to 1 mM (or replace it with DTT, which has a more appropriate  
redox potential). You'll need a protease inhibitor coctail (EDTA, PMSF,  
Pepstatin, Leupeptin, e-aminocaproic acid).

> Once I've got the purified nuclei, the histone prep looks admirably  
> straightforward.  According to Abcam, you just resuspend the nuclear  
> pellet in 0.2M HCl and leave it overnight at 4 degrees C.  Centrifuge to  
> spin down debris and keep the supernatant.  However, other protocols  
> seem to use different acid concentrations, or H2SO4 instead of HCl, and  
> may even include B-ME in the histone extraction buffer.

The acid is quite irrelevant, the idea is simply to increase the  
solubility of the highly charged histons and to pellet as many of the  
other proteins as possible. The bME protects cysteine -SH groups from  
oxidation by air. However, this extraction procedure is too selective for  
your purposes, you will not know whether there are cross-reacting proteins  
in the pellet.
>
> One protocol I Googled called for you to neutralise the preparation with  
> KOH afterwards - is that really called for?  Wouldn't it just drop the  
> proteins straight back out of solution again?

No. But you need a pH near neutral and as low an ion concentration as  
possible for the electrophoresis step. Therefore solubilisation in sample  
buffer rather than HCl is better. That also gives you an idea about the  
cross-reactions of your antibody.

>  For a histone (size ~17 kDa), it'll be quite a high percentage gel.

I find gradient gels (5-15 or 20%) most convenient

>
> Protein 1:   mgclh
> ------------------
>
> RIPA-insoluble, so it's likely mgclh will also be RIPA-insoluble.   
> Protein size is expected to be around 55 kDa.

Again the use of CTAB electrophoresis may help, as CTAB can solubilise  
proteins very efficiently.

> The original literature on Gmcl1 said they resuspended the  
> RIPA-insoluble pellet in "SDS-polyacrylamide gel electrophoresis sample  
> buffer", but doesn't give the composition.  I would assume this is 1x  
> Laemmli sample buffer.  Does this sound plausible?  If so, what recipe  
> for Laemmli buffer would you recommend?  I've found half a dozen  
> different recipes, and the only common factor between them is ~60-100mM  
> Tris pH 6.8 and 2% SDS. There are variants with and without DTT, with  
> and without B-ME, with and without loading dye, using sucrose or  
> glycerol for increased density!

The buffer is actually double concentrated, once you mix it with an equal  
volume of sample you get the working strenght. DTT and bME reduce  
disulphide bonds and thus lead to more complete unfolding of the proteins.  
DTT (Clelands reagent) smells less and has a better standard redox  
potential for this purpose. However, it is more expensive than bME. Wether  
you use glycerol or sucrose to increase density really doesn't matter, you  
just need either for proper gel loading, its physics, not chemistry. The  
dye is important not only to aid in loading, but even more as front  
marker, without it you can't calculate Rf-values which you need for  
molecular mass determination.

> I would imagine I'll need to add B-ME to reduce the proteins before  
> running on the gel, but should I add the B-ME at the point I resuspend  
> the pellet, or only when I'm about to run an aliquot on a gel?  Will I  
> need to add protease inhibitors, or will the Laemmli buffer itself take  
> care of that?

Reducing agent is part of the 2x concentrate, usually the sample/buffer  
mixture is heated to get more efficient reduction. Many people do 2 min in  
a boiling water bath and for your purposes that is probably fine. With  
large proteins I found that 10 min at 60 degrees works better. Very few  
proteases work in hot SDS under reducing conditions, but there have been  
rare cases reported where that was a problem. The protease inhibitors  
should of course be present during the homogenisation of the cells and the  
centrifugation steps. Many (like PMSF) actually chemically inactivate  
proteases, solving the problem permanently anyway.