To qualify as a transcription factor, a protein must possess two qualities. 1) Ability to bind to DNA and 2) recruit RNA polymerase/alter transcription of a gene. Chromatin Immunoprecipitation with your protein of interest (if no antibody is available, a tagged version of the protein overexpressed in the host cells) could be used to pull down DNA fragments. Compared to a known control that does not bind to DNA (ex. Actin) you can evaluate if your protein has ability to bind to DNA. Note that pulled down DNA could be amplified by attaching PCR adapters to it through ligation. So you need not know a priori the sequence of the TF binding site. To evaluate if it can activate transcription you can either use a reporter fusion where you have added one of the pulled down DNA sequence to a reporter gene (such as Luciferase or GFP) and assess if the DNA fragment (TF binding site) in the presence of your protein and a generic promoter element (ex. hsp70 in Drosophila) leads to the expression of the reporter or You can verify if your protein is able to activate transcription by fusing it with the DNA binding domain of yeast Gal4. This would be similar to an yeast one hybrid assay. Possible pitfalls to this approach is that your protein might be a chromatin modifier (or other non-TF DNA binding protein). It would be positive for DNA binding through ChIP, but will yield no significant consensus in DNA sequence for a binding site.  Your protein might be a repressive transcription factor. In which case you will have to have the reporter under an active promoter to evaluate its ability to repress the reporter expression.  

You can use the InterPro Domain scan from the EBI to determine whether the gene product's amino acid sequence bears similarity to any known protein domain, including those of known transcription factors:

http://www.ebi.ac.uk/interpro/

Alternatively, if you can use the PHYRE algorithm to predict secondary structure of your protein sequence and compare it to other known protein structures.

http://www.sbg.bio.ic.ac.uk/phyre2/html/page.cgi?id=index

Have you blasted the protein or DNA sequence to see if you can find any homologies? Otherwise you can use prediction tools for protein domains and see, what the outcome is. Interpro would be one possibility.

You can try the SUPERFAMILY domain database, which has a sequence search facility and associated with it the dcGO function prediction tool. You can start here:

http://supfam.cs.bris.ac.uk/SUPERFAMILY/cgi-bin/dcpredictormain.cgi

then look for transcription factor or related functions in your results.

I don't think, that a single gene can act as transcription factor,because the transcription factors in the developing embryo are complex molecules consisting of a mRNA and HOX proteins- mRNP particles- morphogens. They unlock selectively new genes and are responsible for cell differentiation.

@Marlene: I don't agree. There is quite a number of single genes that act as transcription factors - even in developing embryos. They are usually part of bigger, but the protein is the key part for transcriptional regulation. The microphthalmia-associated transcription factor is such a factor.

Hi Christian, there are transcription factors for differentiation. They cannot be proteins, because proteins cannot read DNA sequences to unlock selectively a new gene. I thin k that the transcription factors you are talking about are stimuli, activating transcription of a differentiated gene leading to function. So function and differentiation are two com pletely different events.Once a cel lis differentiated ( forinstance a salivary gland cell), it can be activated for function zilion tlmes by different proteins or other stimulating or transcription factors, but after transcription seases, the cell remains differentiated.

@Marlene: No, clearly not. Mitf is defnitely involved in differentiation of cells. But you have to decide between the general transcription factors (like EF2 and so on) which assemble the transcription machinery at each site of transcription which is open for transcription and specialized ones, which generally regulate these gene.

@Marlene: I think that there is a confusion here about the term of 'transcription factor'. It's clear that in most cases transcription factors do not act in isolation, but in concert and in interaction with other proteins (other transcription factors bound to DNA, cofactors) or RNA's. However, transcription factors, in the usual definition, are the DNA binding proteins and product of single genes and not the complex within which they act.

Thanks a lot for your interesting answers and arguments.....as mentioned above that most of the TFs would need to assemble their machinery to initiate activation of a gene which can involve various cofactors etc. Some of the genes also need to get phosphorylated to act as a TF....i guess the better example would be STAT family TFs......i will try the above methods and will let u know the results.

Between i have another question concerning the above methods: The gene in question which i am working with does not have its original sequence in databases and has only sequence inferred by homology from a mammalian model organism as it is only published once that the gene expresses in human but by using degenerate primers. We identified the gene with normal primers and have the gene product now. Can we get the original sequence using the gene product by any method/tool/company??

To qualify as a transcription factor, a protein must possess two qualities. 1) Ability to bind to DNA and 2) recruit RNA polymerase/alter transcription of a gene. Chromatin Immunoprecipitation with your protein of interest (if no antibody is available, a tagged version of the protein overexpressed in the host cells) could be used to pull down DNA fragments. Compared to a known control that does not bind to DNA (ex. Actin) you can evaluate if your protein has ability to bind to DNA. Note that pulled down DNA could be amplified by attaching PCR adapters to it through ligation. So you need not know a priori the sequence of the TF binding site. To evaluate if it can activate transcription you can either use a reporter fusion where you have added one of the pulled down DNA sequence to a reporter gene (such as Luciferase or GFP) and assess if the DNA fragment (TF binding site) in the presence of your protein and a generic promoter element (ex. hsp70 in Drosophila) leads to the expression of the reporter or You can verify if your protein is able to activate transcription by fusing it with the DNA binding domain of yeast Gal4. This would be similar to an yeast one hybrid assay. Possible pitfalls to this approach is that your protein might be a chromatin modifier (or other non-TF DNA binding protein). It would be positive for DNA binding through ChIP, but will yield no significant consensus in DNA sequence for a binding site.  Your protein might be a repressive transcription factor. In which case you will have to have the reporter under an active promoter to evaluate its ability to repress the reporter expression.