Interaction between heat shock proteins and antimicrobial peptides

Drosocin, pyrrhocoricin, and apidaecin, representing the short (18-20 amino acid residues) proline-rich antibacterial peptide family, originally isola ted from insects, were shown to act on a target bacterial protein in a ster eospecific manner. Native pyrrhocoricin and one of its analogues designed f or this purpose protect mice from bacterial challenge and, therefore, may r epresent alternatives to existing antimicrobial drugs. Furthermore, this mo de of action can be a basis for the design of a completely novel set of ant ibacterial compounds, peptidic or peptidomimetic, if the interacting bacter ial biopolymers are known. Recently, apidaecin was shown to enter Escherich ia coli and subsequently kill bacteria through sequential interactions with diverse target macromolecules. In this paper report, we used biotin- and f luorescein-labeled pyrrhocoricin, drosocin, and apidaecin analogues to iden tify biopolymers that bind to these peptides and are potentially involved i n the above-mentioned multistep killing process. Through Use of a biotin-la beled pyrrhocoricin analogue, we isolated two interacting proteins from E. coli. According to mass spectrometry, Western blot, and fluorescence polari zation, the short, proline-rich peptides bound to DnaK, the 70-kDa bacteria l heat shock protein, both in solution and on the solid-phase. GroEL, the 6 0-kDa chaperonin, also bound in solution. Control experiments with an unrel ated labeled peptide showed that while binding to DnaK was specific for the antibacterial peptides, binding to GroEL was not specific for these insect sequences. The killing of bacteria and DnaK binding are related events, as an inactive pyrrhocoricin analogue made of all-D-amino acids failed to bin d. The pharmaceutical potential of the insect antibacterial peptides is und erscored by the fact that pyrrhocoricin did not bind to Hsp70, the human eq uivalent of DnaK. Competition assay with unlabeled pyrrhocoricin indicated differences in GroEL and DnaK binding and a probable two-site interaction w ith DnaK. In addition, all three antibacterial peptides strongly interacted with two bacterial lipopolysaccharide (LPS) preparations in solution, indi cating that the initial step of the bacterial killing cascade proceeds thro ugh LPS-mediated cell entry.