Transposition of phage Mu is catalyzed by an extremely stable transposase-DNA complex. Once recombination is complete, the Escherichia coli ClpX protein, a member of the Clp/Hspl00 chaperone family, initiates disassembly of the complex for phage DNA replication to commence. To understand how the transition between recombination and replication is controlled, we investigated how transposase-DNA complexes are recognized by CIpX. We find that a 10-amino-acid peptide from the carboxy-terminal domain of transposase is required for its recognition by ClpX. This short, positively charged peptide is also sufficient to convert a heterologous protein into a ClpX substrate. The region of transposase that interacts with the transposition activator, MuB protein, is also defined further and found to overlap with that recognized by ClpX. As a consequence, MuB inhibits disassembly of several transposase-DNA complexes that are intermediates in recombination. This ability of MuB to block access to transposase suggests a mechanism for restricting ClpX-mediated remodeling to the proper stage during replicative transposition. We propose that overlap of sequences involved in subunit interactions and those that target a protein for remodeling or destruction may be a useful design for proteins that function in pathways where remodeling or degradation must be regulated.

activate the normal transposition pathway, further indicating that ATP plays critical regulatory roles during transposition. These mutant proteins fall into two classes: class I mutants are defective in target DNA binding, whereas class II mutants bind target DNA, deliver it to transposase, but fail to promote recombination with this DNA. Based on these studies, we propose that the switch from the ATP-to ADP-bound form allows MuB to release the target DNA while maintaining its stimulatory interaction with transposase. Thus, ATP-hydrolysis by MuB appears to function as a molecular switch controlling how target DNA is delivered to the core transposition machinery.

Using modified nuclear lysis and binding conditions, we have examined the binding of an embryonal carcinoma (EC) cell factor, binding factor A, to a stem cell-specific silencer which acts at the DNA level and overlaps the Moloney murine leukemia virus (M-MuLV) proline primer binding site (PBS). Following our protocol, we found that in vitro binding of factor A correlated with the in vivo activity of the M-MuLV silencer. Factor A bound specifically to the wild-type silencer element at room temperature and 30؇C, but not at 4؇C, and bound 10-fold better to the full-length silencer than to a minimal silencer core element. The factor was enriched in nuclear compared with cytosolic extracts and in undifferentiated EC cells compared with differentiated cells in which the silencer is nonfunctional. Salt and ion requirements for factor A binding were investigated, and partial purification steps indicated the factor to be a heparin-Sepharose-binding moiety of greater than 100 kDa. To examine possible relationships between silencer and PBS activities, sequences representing phenylalanine, isoleucine, lysine-1,2, lysine-3, methionine, and tryptophan PBS DNA fragments were tested in vivo for stem cell-specific repression of M-MuLV expression and in vitro in DNA binding assays. Of these PBS elements, only the lysine-1,2 PBS DNA fragment showed consistently high levels of repression. Interestingly, the lysine-1,2 PBS DNA fragment also formed a complex with an EC cell factor with characteristics similar to those of factor A. However, the two factors did not cross-compete in binding studies, suggesting that they may be different but related factors. Our results suggest that expression of Mason-Pfizer monkey virus, visna virus, and spumavirus, which use the lysine-1,2 PBS, may be inhibited in undifferentiated stem cells.