Many Ig domains can form beta strand-swapped dimers which don't resolve on SDS page - so the sizes you describe may be about right for that. Sometimes it is worth extending the ends of your construct to produce a more stable monomer.

Hi,

possible explanations could be:

- pure artifact of SDS-PAGE, either covalent dimeric species (disulfide-linked and not sufficiently cleaved by reducing agent in your loading buffer, can happen when buffer too old especially with DTT or not heated long enough before loading) or just aggregated, it's more likely to happen with membrane proteins, but you never know.

-frameshift product, has been shown with motif "AAAAAAG" (present in pDEST-vectors (Gateway), see Belfield et al. (2007), PMID: 17272299), we saw that with 4 different constructs in our own expressions, but usually it's rather unlikely to result in such a big shift, a stop codon usually appears somewhere downstream of that much earlier.

-low pI, i.e. many negative charges can lead to significant shifts, usually not as much as 10 kDa in a 15 kDa protein

-could it be CAT protein (significant expression band at ~23kDa with vectors containing chloramphenicol resistance) and your target protein expressed in much lower amounts?

Hope that helps.

Good luck.

Hüseyin

I agree with Joan-send both bands for MS and only worry about it if the upper band is not your protein. Depending on the buffers and gel systems you use you can often see partly denatured proteins running much larger (or smaller) than anticipated or even dimers, trimers etc. of your protein. If your anti-his Westerns are normally clean then you may just have one of those proteins that run anomalously on SDS-PAGE.

IF MS data comes back to say you have the wrong protein you may just have to add in another purification step, SEC or IEX unless it is an E.coli chaperone (not sure which ones you might have at that size) that may be more difficult to separate from your protein.

Good luck.

to be sure that the 25kDa product is expressed by your construct, analyze by anti-his blot the lysate of your untrasformed E.coli cell line.

Please try again Your interesting experimental research. If the results are identical, hence it is necessary to explain. If the results do not match, change the methodology.

Sounds like your stop codon is wobbly and not completely active. I would suggest cloning the PCR product again with at least two stop codons in tandem. To perform an in silico test, change your sequence so that the stop codon is converted to one that encodes for an amino acid and drop the sequence (whole plasmid) into an analysis tool like NEBcutter or other bioinformatics tool that provides ORF data. If the size of the "mutated" protein using the next "random" stop codon present in the sequence correlates with that seen in your gel, it may confirm what I wrote above.

Agree with Joan on identifying both the proteins. But are you sure about the MW of the protein on SDS-PAGE??. It is not unusual to find your protein migrating anomalously at higher MW on PAGE than calculated MW.. Some correlated this to the % proline residues in the sequence.....

Second suggestion is that it could be some E.coli protein (25 kDa) that Co purify in IMAC.. Since your protein of interest is a domain, if not completely stable sequence, it might be solvent exposed and be bound by some chaperones. Some reviews list a Protein "CA" (Acc. No: NP-414668), which shows 25 kDa molecular weight, for instance . So it could possibly be one of those..

Hope it helps ..Best wishes

Sai

Good luck

It might be due to contamination. I suggest to do Two step purification and MS analysis.

Thanks all.

@ Joan: it doesn't show with same expression vector and cells, other inserts so we thought it "shouldn't" be (!) a contaminant. But yes, we're thinking about PMF now.

@ Deborah: I haven't heard of this. Coincidentally we have marker proteins of 15 and 25 kDa and our product doesn't seem big enough to be a dimer by comparison. Any idea what we might do to resolve beta strand-swapped dimers in order to check?

@ Hüseyin: a bunch of ideas, thanks. pI is 5.0, I agree, this seems like a very big increase. No CAT involved - amp vector. I've looked again at the seq and we have an AAAAG motif straddling stop codon but next stops in alternate frames are 3 aa's or 33 aa's (~3.5 kDa) away, or 4 aa's in same frame. Thanks for reference, it's useful.

Hi,

This sounds like a dimerization. If the protein you are expressing contains cysteines, it is possible that dimerization may occur because of this. Make sure you are using a reducing agent like DTT and 2-mercaptoethanol, and that this agent is freshly made, or in the case of DTT, comes from a stock that is made at time of use or that has been frozen.

Do you elute from IMAC with a gradient of imidazole?

If it is a contamination this protein should/ could elute earlier than the 6xHis tagged proteins.

The marker proteins for the SDS-PAGE are optimal proteins, so the protein is globular and didn't interact with the gel-matrix.

If your protein is not globular or interact with the matrix the retention-factor is different and so the protein looks like it is bigger than expected.

A dimer. Try to boil the samples, use b-mercaptoethanol and suspend the sample in 4x gel loading buffer and see if you increase the smaller band.

Hi Gerard:

Following up on all the suggestions, I would suggest N-terminal sequencing as well. Since this is an IgG like domain, do you know if your protein is expected to bind either Protein A or G ( based on the size, I am assuming you don't have Fc, so this may not work)? You may want to test on a small scale and if the 25kDa is a contaminant, this may be one way to purify your protien of interest i.e. if it is stable at pH 3 (elution buffer for Protein A/G).

Also, you could try size exclusion chromatography (superdex 75) will seperate the two if the 25kDa protein is real and not an artifact of SDS PAGE.

Good Luck!!

You could try adding urea and DTT. If it is locked in a strand-swap by a disulphide, it needs to unfold sufficiently to allow the reducing agent access. I'll see if I can find a protocol we used...

Looks like it could be a dimer, try boiling the sample and add DTT in buffer.

I am not sure if recloning in to another vector will be useful too.

Lots of good suggestions, thanks. I'll try to address all I can:

* we've repeated the expt (a number of time), always the same

* we haven't used an untransformed cell line but have expressed various other proteins in this vector+strain, never see this "additional" product

* we've only one stop (usually use two) but seq anal suggests that read-through would only add 3-4 amino acids before another stop appears, and another soon after

* chaperones a nice idea given the domain; a chaperone that randomly comes up in an anti-his blot too though? Will look up Protein CA

* some form of dimerisation looks a good option; will try to boil in b-mercaptoethanol, urea, DTT and see if larger product converts into smaller one; hopefully will address any lack of globularity (?) too

* don't have an Fc so can't try this

* currently elute in imidazole conc steps, comes off at 100 mM (I think); could try gradient to see if it helps - but we're unsure that we want to use the "correct" protein if we can't figure out what's happening in the expression; similar issue for SEC.

Mass spec of larger band and better denaturation prior to SDS-PAGE look to me like best options for now ...

I guess a dimerization is happening. I would use a loading buffer that contains b-mercaptoethanol and/or DTT and boil the samples before loading on the SDS-page. I do not think changing the cloning vector would help, although you could try this finally. Good Luck

Gerard - A lot depends on the protein itself. Since it was not stated, I assume that you are eluting your protein in non-denaturing conditions, and in such conditions the expressed protein might be strongly interacting with one of the E.coli proteins itself. An example of such would be the proteins involved in transcription or translation, check the functions of your protein if it is involved in any multi-molecular mediated biological mechanism and check if there are any proteins in E.coli that could interact with it. In any case, increasing the salt concentration during washing might help, but higher salt concentration might release the protein itself from IMAC column. Good luck.

Is this a denaturing gel? Have you considered inclusion bodies ?

If DTT doesn't help try treating your protein with better reducing agents like TCEP or sodium borohydride.

I think you need to re-calculate the protein molecular weight. including also 6 histaq

It's true, mass spectrometry of a larger band is needed now. The next logical step - to the functional analysis. And then compare the differences in the properties of a "classical" prototype in E. coli.

Do you target your expression to the periplasm ? If so your protein might not be cleaved of secretory peptide, which was the case when I expressed scFv. Otherwise the suggestion of dimers might be the truth. Igg tend to dimerize frequently, try to use urea ind the sample buffer.

Which E.coli strain are you using for production? Many carry supE or supF suppressors, so that the particular stop codons you have *are* important. With those suppressors, readthrough may happen and hence you have a longer protein which is like yours (anti his blot, IMAC purification). I would suggest checking the suppressors of your strain, then changing your stop codon to prevent readthrough. If I remember correctly (but please check), supE and supF suppressors do readthrough of amber and ochre stop codons, so changing to an opal stop codon by site directed mutagenesis should solve the problem. Another approach would be to put several in-frame stop codons one after the other, as that should prevent readthrough completely (the reibosome pauses and releases your mRNA).

Regarding the potential dimerization it may be useful to use thioglycolic acid as DTT and 2-ME are not charged and thus dont run into the gel where your protein may re-oxidise (i.e. form interchain disulfide bridges). You may also want to ensure your buffers used for extraction and purification are fresh to avoid covalent cross-links. Adding low concentration of 2-ME (1mM) or Na2S2O5 (up to 20 mM) may also be useful to reduce oxidative crosslinking. If you small scale lab-experiments be also aware that large surface areas (such as in a drop or when decanting supernatants) expose your protein to air (=oxygen), thus promoting crosslinks. On larger scale you may be able to keep oxygen away from your samples.

I consulted with a professional in this field. This is Vladislav KHRUSTALEV. He advised to carefully check the nucleotide sequence of the gene, which is introduced into the plasmid. We must remember that a poly-histidine tract is saved as one or the other protein. Perhaps at some point the nucleotide sequence of the ribosome "jumps" to the correct reading frame. Such jumps are possible on polynucleotide tracts (row when there are 4 or more identical nucleotides).

Digest the long band (and if possible the 15-kDa band, too) with trypsin and analyse the peptide mixture you get by mass-spec. If the 'large' protein is just a dimer of your Ig-like domain (sounds unlikely to me, but who knows?), you should get only peptides from this protein. If the long product is the result of a ribosome readthrough, you should get aberrant peptides recognisable as derived from translation of the 3'UTR. If your long band contains parts of another E.coli protein or of another plasmid-emcoded protein, you might be able to identify these, too.

I would increase the concentration of DTT in the sample buffer and try again.

I encountered a similar problem when i expressed NKG2D which is about 16KDa. So i tried expressing it as inclsions bodies by increasing the tempreature and rpm, and reducing the time for fermentation. This time round i had two bands with one a little above the other but both around required 16KDa in anti-his blot. I denatured the inclusion bodies and did purification. Afterwards SDS page showed only one band with the right size. I therefore went ahead and did renaturation. So if the above suggestions fail, you can try expressing it as an inclusion bodies (although difficult and time consuming). All the best.

We have observed such a trend with some proteins, especially ones with low pI values (around 5), where a protein of abt 15 kDa runs at a size of around 25-30 kDa. We thought that the protein could be forming dimers, but when we did a gel filtration chromatography, the size was determined to be 15-16 kDa.

I agree with Valencio, we have the same problems, but our proteins have a very high pI value, at round about 11.5. One protein is calculated with 9kDa but the protein shows retention times round about 16kDa, so too small for a dimer. If we do posttranslational modifications on the fusionprotein, which should reduce the mass only by 1Da (Mass is checked by Mass-spec) the band in SDS-page is round about 5kDa higher than expected and the pI changed by the posttranslational modification from 9.8 to 6.5.

For my understanding SDS-Page is not the best method to determine a mass of a protein.

I too am having a problem similar to this. Does anyone know a good paper that discusses inclusion body identification so I know if this is the reason for my increased bands? A further question, I am getting this large mass on a MALDI as well. Would inclusion bodies fly on a MALDI at the increased weight as they run on a SDS page gel? Thanks