PROTEIN-PROTEIN INTERACTIONS IN COLICIN E9 DNASE-IMMUNITY PROTEIN COMPLEXES .2. COGNATE AND NONCOGNATE INTERACTIONS THAT SPAN THE MILLIMOLAR TO FEMTOMOLAR AFFINITY RANGE

The in vivo and in vitro cross-binding of the colicin endonuclease-spe cific immunity proteins toward the DNase domain of colicin E9 is descr ibed. In vivo cross-protection was tested by toxin plate assays in whi ch bacterial cells overexpressing each immunity (Im2, Im7, Im8, and Im 9) were challenged with the Co1E9 toxin. Im9, the cognate immunity pro tein, renders cells completely resistant toward very high concentratio ns of the toxin (> 1 mg/mL), whereas the noncognate immunities display a spectrum of weaker cross-reactivities (< 0.01 mg/mL). The order of biological protection in this assay was Im9 much greater than Im2 > Im 8, with Im7 providing no colicin E9 resistance. in vitro binding betwe en the immunity proteins and the E9 DNase was analyzed by determining the dissociation constants for E9 DNase-Im protein complexes at pH 7.0 in the presence of 200 mM salt and at 25 degrees C. Stopped-flow fluo rescence experiments suggest that both Im2 and Im8 associate with the E9 DNase by a two-step mechanism, in which the rate constants for both the bimolecular association (k(1) = similar to 6 x 10(7) M(-1) s(-1)) and the subsequent conformational change (k(2) + k(-2) = 4-5 s(-1)) a re very similar to Im9 binding under the same conditions. Fluorescence chase experiments defined the dissociation rate constants for Im2 and Im8. The estimated values are 10(6)- and 10(8)-fold, respectively, fa ster than the off-rate for the Im9 protein. The ratio of the associati on and dissociation rate constants gives K-d values of approximately 1 0(-8) and 10(-6) M for Im2 and Im8, respectively, whereas the K-d for Im9 under these conditions is 10(-14) M. The activity of the E9 DNase was also analyzed by an in vitro plasmid nicking assay. All three nonc ognate immunity proteins inhibit the E9 DNase. The K-i's for Im2 and I m8 agree closely with the stopped-flow experiments, and the K-i for Im 7 was determined to be 10(-4) M. The order of immunity protein affinit y for the E9 DNase (Im9 much greater than Im2 > Im8 > Im7) is the same as that for the in vivo toxin protection experiments, implying a corr elation between biological specificity and in vitro binding. Our resul ts show that the specificity of E9 DNase-Im protein interactions are s pread over an affinity range that spans more than 10 orders of magnitu de and so offers new opportunities for understanding the basis for spe cificity in protein-protein interactions.