This is an adaptation of the published procedure for RNAse treatment of Drosophila polytene chromosomes to remove proteins whose binding is RNA dependent, such as the dosage compensation complex member MLE (Richter et al., 1996). In this adaptation, I used differently methylated histone tail peptides to compete Polycomb protein from polytene chromosomes, finding that competition was peptide specific and locus-specific (Ringrose et al., 2004). (I also tried RNAse treatment, but found it to have no effect on Polycomb binding).
Strictly speaking, this is not an in vivo assay, as the salivary gands are removed from larvae just before peptide competition. However, since the assay is performed on unfixed, intact salivary glands, the hope is that it will reflect the in vivo behaviour of chromatin binding proteins to some extent. Another limitation is that the salivary glands represent a rather strange tissue, that will not persist in the adult fly, and whose chromosomes have undergone many rounds of endoreplication. However salivary gland chromosomes are unparalleled for cytological purposes, and thus have many advantages if specific sites of competition of chromatin binding proteins are to be mapped (see Ringrose et al., 2004).
Essentially the idea behind competition with histone tail peptides is that if a protein is anchored to chromatin by an interaction with histone tails, (or by an interaction that uses the same binding surface of the protein), and if this interaction is dynamic, involving an exchange between bound and free protein, then this binding can be competed by incubating freshly prepared, unfixed chromosomes in an excess of histone tail peptide. If your protein of interest is very tightly bound to chromatin, by its histone tail interaction or by other interactions, then you would not expect to see loss of binding upon peptide competition. A good positive control for the assay is to use H3K9 dimethylated peptide to compete HP1 from the pericentric heterochromatin. HP1 is a highly mobile protein whose binding is readily lost upon competition with this binding partner.
Beyond histone tails, the technique may be useful for probing the anchoring of chromatin bound complexes by other interactions, for example, specific DNA sequences or protein domains. I have not tried any of these but they may be worth investigating - as long as the competing molecule is small enough to enter permeabilised salivary glands, it should be suitable for this assay.
The assay has three main steps:
- Preparation of larval salivary glands,
- Permeabilisation and incubation with competitor peptide, and
- Fixation and staining for the protein of interest.
By far the most the most critical step in the assay is permeabilisation: this step can destroy chromosome morphology, which can lead to artifactual apparent loss of binding of the protein of interest. In addition, antibody access to chromosome spreads can vary across a single slide, again leading to apparent "loss of binding". For this reason it is essential to include an internal control antibody for chromosome morphology and antibody
access. For this internal control I use a mouse monoclonal anti-Modulo antibody (Perrin et al., 1998; the antibody was obtained from Jacques Pradel) in double stainings with rabbit anti Polycomb antibody. The Modulo protein is highly enriched in the nucleolus, and also binds the chromosome arms, giving many bands, rather like DAPI staining, (see Figure 1B and 1E). I found this chromosome arm staining to be exquisitely sensitive to chromosome morphology, whereas the anti-Polycomb antibody gave robust staining on chromosomes of even very poor morphology. Therefore, after the permeabilisation and competition, only those spreads that showed good Modulo staining throughout all the arms were analysed for loss of Polycomb binding. I strongly recommend using this anti-Modulo antibody as an internal control, rather than an antibody against an integral chromatin protein such as a histone, which may be more resistant to disruption of morphology. Indeed the initial motivation for using Modulo was as a control for RNAse treatment: it is lost from the nucleolus but not from the chromosome arms upon RNAse treatment of permeabilised salivary glands (Perrin et al., 1999).
Institute of Molecular Biotechnology (IMBA) GmbH
Dr. Bohr-Gasse 3 - 1030 Vienna, Austria