Ravi Kaul and Wanda M. Wenman

Department of Pediatrics, Section of Infectious Diseases, 403 Neurosciences Building, School of Medicine, 1515 Newton Court, University of California, Davis, CA 95616

Received 2/16/ 98 Accepted 2/20/98

5. FUNCTIONAL ROLE(S) OF EUKARYOTIC-LIKE HISTONES AND THEIR TERMINI

Condensation of chlamydial nucleoid occurs late in its life cycle, concomitant with the expression of Hc1 and Hc2 and accompanied by down regulation of transcription and metabolic processes. In order to examine the role of eukaryotic-like histones in macromolecular confirmation of DNA Barry et al (28) expressed C. trachomatis Hc1 in E. coli . E. coli expressing recombinant Hc1 revealed a condensed nucleoid structure similar to that of chlamydiae when examined by light and electron microscopy. Further, Hc1 was shown to co-sediment with purified recombinant E. coli nucleoid. These results strongly support a role for Hc1 in condensation of the chlamydial nucleoid. Additional evidence invoking the role of Hc1 in DNA condensation was derived from in vitro studies showing purified Hc1 complexed to double stranded DNA, leading to the formation of large aggregates, often in the form of condensed spherical bodies (29, 30). Expression of Hc1 in E. coli was shown to down regulate translation, transcription and replication at concentrations similar to those observed in chlamydial EB (31). These authors also demonstrated that low level expression of Hc1 in E. coli results in net relaxation of chromosomal DNA - suggesting a bifunctional role for Hc1 depending upon its concentration. These observations agreed with the earlier findings demonstrating that highly supercoiled DNA is associated with the EB stage of the developmental cycle (32). Pedersen et al (33) have shown that purified Hc1 inhibits transcription and translation in vitro by interacting directly with DNA and RNA. The tight coupling of Hc1 gene expression with DNA replication accounts for its role in cell growth and division. Interestingly, expression of Hc2 in E. coli induces compaction of bacterial chromatin distinct from that mediated by Hc1, suggesting differential DNA-binding modes for Hc1 and Hc2 (31). Pedersen et al (33), on the other hand, concluded that that DNA-condensation is not the principal function of Hc2 because of their failure to observe nucleoid condensation among E. coli expressing Hc2. Further, Hc2 immobilized on nitrocellulose displayed a higher affinity for single stranded DNA and RNA than for double stranded DNA (30).

A first step towards understanding the role of chlamydial Hc1 is to delineate the regions responsible for its characteristics. We have recently subcloned segments of the Hc1 gene corresponding to its amino terminal portion (amino acids 2-65, designated H1N) and carboxyl terminal portion (amino acids 68-125, designated H1C) (34). Expression of these subunits in E. coli has identified peptides with estimated molecular weights of 14,000 for the carboxyl and 8,000 for the amino terminal portion of histone H1. Again, the highly basic nature of the carboxyl terminal domain (pI 13.2) may explain the discrepancy between its estimated (14,000 daltons) and calculated (7,100 daltons) molecular weights. Polyclonal antibodies raised against either amino or carboxyl termini of Hc1 react with recombinant E. coli as well as to C. trachomatis EB, suggestive of antigenic conservation. Subsequently, we examined the binding of double stranded DNA to Hc1 and its two terminal portions using Southwestern blotting. Hc1 and its carboxyl terminal portion were found to bind DNA; however, no binding was observed between the amino terminus and DNA. E. coli expressing either whole Hc1 or its carboxyl terminal peptide was observed to condense its own DNA. The condensation ranged from highly electron dense particles in cells expressing whole Hc1 to loose condensation in the case of carboxyl terminal expression. No condensation was observed in cells expressing the amino terminal portion. Structurally, cells expressing the amino terminus appeared similar to controls. These results clearly imply a role for the carboxyl terminal portion in DNA:protein interaction, a function similar to its eukaryotic counterpart. However, this interaction alone does not appear sufficient to condense DNA tightly; intramolecular protein:protein interactions seem necessary to mediate that effect. Given the sequence conservation at its amino terminus it is tempting to speculate its role in protein-protein interaction. Recently Pedersen et al (35) presented experimental evidence to show a potential dimerization site at the N-terminal domain of Hc1. Based on these results, the authors concluded that the amino terminus is functionally important in protein-protein interactions despite the fact that it is not involved in any appreciable interaction with DNA.These results may also help to explain why Hc1 forms higher order complexes with DNA.

Evidence is mounting that Hc1 serves as a non specific yet carefully regulated transcriptional repressor in a manner similar to eukaryotic histone H1 and prokaryotic histone H1-like protein H-NS (30-31, 36-38). In eukaryotes, selective removal of histone H1 renders some genes transcriptionally active, while addition of H1 protein has been shown to repress in vitro transcription (39). In vivo histone H1 is more prevalent on chromatin of transcriptionally inactive genes than on transcriptionally active genes (40). Although it seems certain that all kinds of DNA will bind histone H1, there are reports that some sequences bind H1 better than others (41-42). However, the association of Hc1 with transcriptionally active genes during the chlamydial growth cycle has not been examined. To identify sequence specific and sequence independent interactions between Hc1 and chlamydial DNA we developed a cross-linking immunoprecipitation protocol to immune precipitate chlamydial Hc1 cross linked to DNA (43). The DNA thus purified was used to probe Southern blots. Our data clearly indicate the presence of sequence specific binding sites on the chlamydial plasmid and Hc1 gene upstream of its open reading frame, in addition to other sequence-independent sites. The sequence-specific high affinity binding sites on the chlamydial plasmid and Hc1 gene were mapped to 24-bp regions that were 70% identical. No intrinsic curvature(s) was detected within these high affinity sites arguing against the role of DNA bends in Hc1 binding. We also ruled out the possibility that such interaction was mediated through a high A+T ratio of the 24-bp fragment. Control fragments with similar A+T ratio failed to associate under similar experimental conditions. More experiments are required to address what mediates sequence specific binding. The observation that Hc1 preferentially binds to only one strand of plasmid DNA is intriguing. Mathews and Sriprakash (44) reported earlier a strand specific endonucleolytic activity in high salt extracts of C. trachomatis. Coincidentally, this activity was specific for a region that lies adjacent to the primary Hc1 binding site on chlamydial plasmid. Whether an interplay between endonucleolytic activity and Hc1 induced superhelicity exists remains to be explored. The site specific affinity of Hc1 was further demonstrated by atomic force microscope (AFM) data images. Hc1 binding was always followed by coiling, shrinking and aggregation of the affected DNA (43).