(by mlsulliv from wisc.edu)
Tue Jul 3 13:12:35 EST 2007
On Jul 2, 2007, at 11:59 AM, Michael Sullivan wrote:
> I've prepared a rabbit polyclonal antiserum against a protein of
> interest. From immunoblots, it appears that the antiserum is
> crossreacting with the large subunit of Rubisco to make a big fat
> band. Since the antigen was protein that was produced in bacteria,
> I think any cross reactivity must just be coincidental. Also, the
> pre-immune serum, did not have any problem with rubisco
> crossreactivity. I haven't found any obvious stretches of sequence
> similarity between rubisco and my protein of interest to account
> for the crossreactivity, but I think it can be hard to make
> predictions about antibody cross-reactivity based on sequence
> alone. Unfortunately, the protein I would like to detect with the
> antiserum is very similar in size to Rubisco. I'm worried I won't
> be able to detect "real" signal from my protein of interest because
> of the big fat Rubisco band.
> Has anybody had a similar problem, and if so, are there any
> solutions? Might different blocking agents help (I'm currently
> using 5% NFDM in TBS)? Can one do a more "stringent" wash to
> eliminate unwanted signals. Currently, I'm thinking the best
> strategy might be to express Rubisco in bacteria and use the lysate
> to preadsorb the antiserum. Has anybody here tried that strategy
> Thanks for any input.
Thanks to everybody who offered solutions to my antibody problem
I was asked to post the replies, since often those who reply do so
directly to the poster, and not to the group.
Apparently, lots of you have experienced this and have found various
ways to dealing with it. There were also differing opinions about
whether the Rubisco cross-reacting antibodies represent low affinity
antibodies that react with the very abundant rubisco protein or
whether they represent antibodies that are raised against rubisco in
the rabbit's diet because of the heightened immune response during
immunization (due to adjuvant). I had considered the latter
possibility, but ruled it out because I had never experienced it
myself. I'm reconsidering this though, since of course something
could have changed since the last time I had antiserum prepared.
The solutions tended to fall into 5 categories:
1. Affinity purify the antibodies, either with a column or using
antigen bound to a membrane. (I will probably try this first)
2. Deplete the serum of rubisco reactive antibodies using a crude
3. Alter blocking/washing conditions.
4. Try a different electorphoresis approach, e.g. native or 2-D.
5. Deplete samples of rubisco or try using non-green tissues.
Here are the replies as they came to me.
> Try 0.1 to 1% tween (or other non ionic detergent) in TBS. Also look
> for competition with the peptide to find out if this is specific/
> nonspecific signal. Try also native vs SDS gel, one might work better
> than the other.
>> I've prepared a rabbit polyclonal antiserum against a plant
>> protein of
>> interest. From immunoblots, it appears that the antiserum is
>> crossreacting with the large subunit of Rubisco to make a big fat
>> Might different blocking agents help (I'm currently using
>> 5% NFDM in TBS)? Can one do a more "stringent" wash to eliminate
>> unwanted signals.
> You might as well try BSA as your block, and using a carbonate
> instead of
> tris-glycine buffer for the transfer, but these are usually tricks to
> reduce nonspecific binding, rather than a cross-reaction. I've never
> mucked about with stringency of washes, but more detergent
> apparently does
> the trick; i would imagine higher salt might too. Again, though, if
> have a specific cross-reaction, i'd guess this probably won't help.
>> Currently, I'm thinking the best strategy might be to express
>> Rubisco in
>> bacteria and use the lysate to preadsorb the antiserum.
> Sadly, it might be. Or if you happened to have plant cells that didn't
> express your protein, you could pre-adsorb against lysate from those,
> which would avoid having to make rubisco in bugs.
> I guess making a monoclonal, and picking one that doesn't cross-
> react, is
> also an option, but a rather laborious one.
> Alternatively, if you're currently running denaturing PAGE, maybe
> try a
> native gel; if you're lucky, the epitope that's being recognised on
> rubisco will be hidden, and the one on your protein won't. It
> should also
> shift rubisco up away from your band, since it's a monstrous huge
> Of course, if your protein is also in a complex, this may not be so
> Hi -- a fellow (current) Vierstra lab resident here. Most
> rabbits are fed alfalfa or other green leafy vegetables. After
> initial injection with your antigen in Freund's, a major immune
> response occurs and all kinds of antibodies to foreign substances
> are made, especially those that are found in high quantities (e.g.
> rubisco in food). No easy way around this except to immuno-purify
> your Ab. Two ways to do this relatively easily. If your antigen
> is soluble, crosslink it to some beads (cyanogen-bromide activated
> agarose from anyone, or Affigel from BioRad) and then pass your
> serum over. Antibodies to the antigen should bind, wash thoroughly
> and elute off with .1M glycine pH 2.7 or so. If your protein is
> insoluble, you can run an SDS-PAGE gel with either a prep-comb (one
> large well all the way across) or a few larger wells to load as
> much protein as possible (250-500 ug). Run the gel and transfer to
> a membrane as for a Western blot. After staining the membrane
> (with Ponceau S?), cut out your antigen. You then use this strip
> in a modified western-blot protocol (block with milk for a hour,
> incubate your antisera (2.5 mls plus 280 ul 10x PBS or TBS) on the
> membrane for a few hours, wash extensively and elute off the bound
> antibody with low pH elution buffer as above. Be sure to
> neutralize your low pH with some Tris, pH 8m so as to not damage
> the Ab proteins. You could also try your method, but I think the
> two mentioned above would be better.
> Can share more specific details if necessary and pass along
> protocols (can campus mail copies or send over scans). Hope this
> helps and if you have any further questions, email me directly or
> give a call
> Hi Mike,
> A postdoc of mine used rubisco purified from grocery-bought spinach
> (I believe) to preadsorb the serum (or IgG). It worked well. I'll
> see if he can send you details.
> all the best,
> Mike--there's a company that has already made antisera against
> Rubisco for use in cleaning up proteins for 2-D gel analysis. Don't
> remember who off-hand but you could probably find out with some
> Google sleuthing.--Eric
> Good luck with your antibody. Immunopurifying monospecific
> antibodies may be the best option but of course it is not always
> technically straightforward; so many antibodies just do not behave
> well re: binding, elution etc. But maybe an alternative strategy
> might be to remove Rubisco from your samples, if you are pretty
> sure that is the problem (is your protein of interest present in
> roots or etiolated tissue, for comparison?). I seem to recall there
> were a few kits out there that proported to do this, one is below
> (pricey), maybe there are others?
> we have observed a similar cross-reaction with anti-sera prepared
> to bacterial expressed plant proteins. I think that there is just
> so much rubisco and it must be "sticky".
> It seems to help are using some tween in your blocking buffer and
> wash solutions and including BSA in the block and wash solutions
> (See attached for recipe for HNAT).
> Another solution seems to be affinity purification of the antisera.
> These antibodies seem to be more selective. Probably all the high
> affinity antibodies do not come off the NC readily and those may be
> the ones sticking to rubisco. Or maybe it is the low affinity ones
> that stick to rubisco and this protocol eliminates those.....The
> protocol is attached. Good luck, hope it helps.
> Our department chair, who's on my thesis committee, forwarded me
> this question, since I have had problems with rubisco cross-
> reactivity on my westerns as well. What my lab normally does for
> Westerns is a high-salt protocol that we got from the Vierstra lab
> (so I assume you have it, too; just in case you don't, it's
> attached). Prior to incubation with the primary antibody, we block
> for 1 hour in TBS-T with 10% milk. That worked beautifully for one
> of our proteins but hasn't been so helpful for two others; I'm not
> sure why. One other thing that has worked very well for us in
> eliminating unwanted bands is diluting the antibodies. In our
> current trials for a western we're optimizing, for example, we are
> using 20uL of 1:100 primary antibody in 25mL of HST. How large is
> your protein? Could you run your gels long enough to resolve it
> from rubisco?
> I would love it if you posted your replies to this query back to
> the newsgroup. I have struggled with the same problem for several
> I recently came across a commercial column that removes Rubisco
> from extracts. It is expensive but certainly looks like it works
> well. I will forward you the email I was sent in a following message.
> Hi Barb,
> Thank you very much for your interest in our product.
> The antibody targets both small and large subunits of the Rubisco.
> Please, find attached data and data-sheets for different formats.
> All columns are multiple use (100 times use). These are the
> columns available in 3 formats:
> Part# 28-288-21253-SC. Spin Column kit (includes 2 spin columns and
> and collection tubes) - each column used 100 times. Each run with each
> column removes 0.8 mg of Rubisco. Capacity of each column is 0.8 mg of
> Rubisco per run. Price - $1,950 for the whole 2 column kit.
> Part# 28-288-21253-LC2. HPLC LC-2 column (2 mL bed volume). One
> HPLC column
> and buffers. Can be used 100 times. Capacity about 2 mg of Rubisco
> per run.
> Price - $4,200/kit.
> Part# 28-288-21253-LC10. HPLC LC-10 column (10 mL bed volume). One
> column and buffers. Can be used 100 times. Capacity about 10 mg of
> per run. Price - $14,500/kit.
> Currently, Rubisco columns are on the back order until the end of
> April. If
> you order by the end of next week, I will put you on the waiting
> list for
> the next delivery in the end of the month.
>> I've prepared a rabbit polyclonal antiserum against a plant protein
>> of interest. From immunoblots, it appears that the antiserum is
>> crossreacting with the large subunit of Rubisco to make a big fat
>> band. As many of you probably know, Rubisco is very prominent in
>> extracts derived from green plant tissues and can constitute up to
>> 50% of the protein in such a sample. Since the antigen was protein
>> that was produced in bacteria, I think any cross reactivity must just
>> be coincidental.
> You have antibody raised against a peptide ? or just a part of the
> recombinant protein? Either way you could try affinity purification
> of your antibody with the protein it was raised against.... you could
> have a cleaner antibody.
> Also, the pre-immune serum, did not have any problem
>> with rubisco crossreactivity. I haven't found any obvious stretches
>> of sequence similarity between rubisco and my protein of interest to
>> account for the crossreactivity, but I think it can be hard to make
>> predictions about antibody cross-reactivity based on sequence alone.
>> Unfortunately, the protein I would like to detect with the antiserum
>> is very similar in size to Rubisco. I'm worried I won't be able to
>> detect "real" signal from my protein of interest because of the big
>> fat Rubisco band.
> Another way would be to preadsorb the serum/ antibody with a bacterial
> "powder".... Molecular cloning methods by Sambrook et al or methodss
> in molecular biology or even the manual "Antibodies" would give you a
> protocol to prepare the bacterial "powder". You don't have to express
> the Rubisco there just yet, I imagine.
> Mike, I don't have a solution, just thnking aloud. A guy in our lab
> had the same problem - except that his protein strictly co-migrated
> with Rubisco on gels... I suggested 2D gels to him, he never got
> them going but I still have a feeling that it can solve most of the
> Another alternative is native blue gels - if you are lucky, your
> protein will ran differently than Rubisco.
> Also, IMHO, instead of expressing Rubisco in bacteria, the fastest
> solution (in a protein purification--friendly lab) would be just
> native Rubisco for making preabsorption columns.
> Here are some suggestions that may help you out:
> 1. Try doing westerns with extracts prepared from non-
> photosynthetic organs, where RUBISCO is not expressed - i.e.,
> roots; this is assuming that you gene is expressed there.
> 2. Try using different dilutions of your primary antibodies - try
> a series
> 3. You can certainly try washing your westerns with higher
> stringency - by increasing that salt concentration and the use of
> non-ionic detergents (e.g., NP40 or Trition X100).
> 4. You could also add recombinant protein into your plant extracts
> - this would work as a spiked internal standard that would confirm
> if your protein of interest is same/different from RUBISCO.
> 5. You can also change your SDS-PAGE system, so as to get some
> separation between RUBISCO and your protein of interest.
> 6. Some combination of the above.
> I need to dig around on some old CDs at home to pull up my notes.
> But, the technique was really straight-forward. Using an old
> Rubisco purification protocol, I made Rubisco from spinach. It
> didn't require a lot in the way of special equipment or column
> Next, I linked the Rubisco to a column resin (Pierce AminoLink, if
> I remember correctly). Next, I simply passed aliquots of my
> antiserum over the column. Pre- and post-column western blots
> showed that I'd cleaned up the problem bands.
> I was writing this up for a BioTechniques article, but never got
> around to finishing it. I'll see if I can find the write-up on
> some old CDs at home and send it over to you.
> I'd recommend two things
> 1. Affinity purify your antibody. Link your antigen to a column
> (kits from Pierce or Biorad are good) and purify the antibodies
> you want to keep. If you want and the cloning is easy, you can split
> your expressed protein in half, and purify antibodies that bind to the
> N-term and the C-term separately on two different columns.
> 2. Compare extracts from different subcellular fractions with the
> purified antibody - unless you expect your protein is also in the
> chloroplast stroma. Do you expect it soluble, membrane, or in the
> cytosol? What about in roots? Hard to recommend without knowing
> about your expectations. If your protein is similar in size to
> you need to get Rubisco out of the way anyway to see what's going on.
> Spend your time trying to see if what you want is in your antiserum-
> looking in cellular fractions depleted of rubisco.
> As for blocks, 5% milk is a lot. I usually use 2% in TBS or TBST.
> I've always gotten what I need with either the milk block or Peggy's
> HST recipes that were used in the Vierstra lab.
Michael L. Sullivan
Plant Research Molecular Geneticist
US Dairy Forage Research Center
1925 Linden Drive West
Madison, WI 53706
(608) 890-0046 (Phone)
(608) 890-0076 (FAX)
More information about the Methods