There are at least 3 non-exclusive reasons why you see multiple bands: (1) there are multiple isoforms of the protein due to different mRNA splicing and/or post-translational modifications, (2) the protein has suffered proteolytic degradation, (3) the antibody is not specific.

Is the highest molecular weight band at the expected size of the protein? If so, it is probable that the other bands are proteolytic fragments. You may be able to reduce this problem by adding a protease inhibitor cocktail during extraction of the sample.

Is the lowest molecular weight band at the expected molecular weight of the protein? If so, the protein may be glycosylated or have other post-translational modifications. You can try treating the sample with PNGase F to remove N-linked glycosylation. If you are extracting the protein from cell culture, you can try pre-treating the cells with an inhibitor of N-linked glycosylation, such as tunicamycin, to prevent glycosylation.

If you have a sample that lacks the protein but is otherwise the same, you can pre-adsorb the antibody with that sample to remove non-specific reactivity.

There are at least 3 non-exclusive reasons why you see multiple bands: (1) there are multiple isoforms of the protein due to different mRNA splicing and/or post-translational modifications, (2) the protein has suffered proteolytic degradation, (3) the antibody is not specific.

Is the highest molecular weight band at the expected size of the protein? If so, it is probable that the other bands are proteolytic fragments. You may be able to reduce this problem by adding a protease inhibitor cocktail during extraction of the sample.

Is the lowest molecular weight band at the expected molecular weight of the protein? If so, the protein may be glycosylated or have other post-translational modifications. You can try treating the sample with PNGase F to remove N-linked glycosylation. If you are extracting the protein from cell culture, you can try pre-treating the cells with an inhibitor of N-linked glycosylation, such as tunicamycin, to prevent glycosylation.

If you have a sample that lacks the protein but is otherwise the same, you can pre-adsorb the antibody with that sample to remove non-specific reactivity.

Look on the Pubmed GenBank if the protein has more than 1 isoform. Contact a lab that has published something using the same antibody you have and ask them if they experienced something similar, usually they are happy to help. Ask your PI to write the email, they are usually answered faster this way. Sometimes trying different blocking methods helps (milk, casein, BSA, 1h, 2h, overnight) etc...

Dear Asa, If you have only one uncharacterized antibody (more than one antibody sometimes helps in a democratic decision!!) and no additional controls (eg. Transfected lysates, knockdown or knockout cells), it is very hard to know whether which are the non-specific bands and which ones are real isoforms of your target. If I am getting into a new subject, I will always prefer to use a previously published antibody, so that even if there are non-specific bands, at least the real band will be there somewhere.  

To get a first impression, u should certainly look in the genbank for reported protein sequences as Luiza suggested above. If you have major bands at the right sizes, U may  still have to verify the identity of the other bands by general (eg; siRNA ) or target protein specific approaches (eg. From human protein atlas, u can get information regarding the cells / tissues which expresses a protein strongly or not express them and these samples can be used as controls).

Regarding some of your specific queries, it is very rare chance that you can have interacting complexes present in SDS-PAGE. But there are proteins which really are naughty in their migration patterns (Theoretical MW and the apparent MW in the gel differing a lot), sometimes due to sequence features or posttranslational modifications.

If it is getting really complicated, please get help from people with experience with the protein under question. Good Luck

Thank you all. Coud you please explain me a bit more on how to use genbank?  When I go to genbank page. Should I select protein or nucleotide(default) on searchbox? In case I select protein. it will show list of protein with its aa (#of amino acid). And What should I do next from here?

Dear Asa,

There are several ways to use NCBI to search for isoforms. The quickest way is to go to Pubmed (http://www.ncbi.nlm.nih.gov/pubmed) and select "protein" from the dropdown menu next to "pubmed" search bar. Type your protein of interest either spelled out or gene symbol and then type the species you are interested in. This will give a list of database entries for that protein and you will be able to see their aa length. Some of the shortest entries will be incomplete sequences and not isoforms. Click on an entry and then look at the references to see if any isoforms have been published.

A more comprehensive way is to select "gene" from the dropdown menu instead of "pubmed" and type in the gene symbol/abbreviation of interest plus the species. This will give me more information than you want but there is a section about halfway down the page called NCBI reference sequences, which details mRNA entries to the database. I would recommend searching those mRNAs to see if any are truly isoform variants. These could again be incomplete sequences, so you will have to read the literature provided to be sure.

Good luck,

If you have your positive control or you can buy from commercial you can compare the molecular weight with your sample.

If I see unspecific bands, I will try to block with 5% blocking at 4 C overnight,  washing longer like like 3 times 20 min with large volume TBST. Also, if you have other batch of primary antibody tried to test from different company or borrow your colleague if someone has. (คนไทยรึปลาว ครับ เห็นอยู่ มหิดล :D )