Fig. 4. Adenine nucleotide-dependent binding of E. coli and chloroplast cpn60 to immobilized cpn21. Plasmid pCK28 encodes an N-terminal fusion between six histidine residues and cpn21. The His-cpn21 protein was synthesized in E. coli, bound to nickel (Ni-NTA) columns and washed free of contaminating proteins. Purified E. coli cpn60, or soluble spinach chloroplast proteins, were incubated with the His-cpn21 resin ± adenine nucleotide, washed, eluted with imidazole, and analyzed by SDS-PAGE. Lane 1, JM105(pCK28) total preinduced proteins; lane 2, total induced JM105(pCK28) proteins; lane 3, His-cpn21 bound and eluted from Ni-NTA column (the arrow for lane 3 marks the position of His-cpn21); lanes 4 and 5 show protein eluted from column following passage of E. coli cpn60 -ADP (lane 4) and + 1 mM ADP (lane 5) (the arrow for lane 5 marks the position of cpn60 and the star the position of His-cpn21); lanes 6-9 show chloroplast extracts where 6 is the column preload, sample 7 is cpn60 bound to His-cpn21 in presence of 1 mM ATP, sample 8 is cpn60 bound in presence of 1 mM ADP, and sample 9 is pretreated with apyrase before column loading in the absence of either ADP or ATP. The two bands marked by arrows in lanes 6-9 are the a (top) and /3 (bottom) of spinach cpn60. Lanes 1-5 were stained with Coomassie brilliant blue, and lanes 6-9 immunoblotted, reacted with rabbit anti-cpn60, and goat anti-rabbit IgG conjugated to alkaline phosphatase (Baneyx et al., 1995).

arranged with rotational symmetry around an axis through the center of the toroid (Baneyx et al., 1995; see Fig. 5).

Because each domain of cpn21 is functional, this raises the possibility that the a and ¡3 cpn60 subunits in chloroplasts require different domain interactions for maximal activity, thus implying that the two cpn21 domains have preferred binding sites either on different cpn60 molecules

Fig. 5. Electron micrographs of purified spinach cpn21 synthesized in E. coli. The protein was fixed with 1% glutaraldehyde and negatively stained with 1% uranyl acetate. Two magnifications are shown; in (A) the bar represents 100 nm and in (B) the bar is equivalent to 20 nm. Electron micrographs were kindly provided by Jan van Breemen (Baneyx et al., 1995).

or on different faces of the same molecule. Perhaps the fused-domain structure of the chloroplast cpn21 satisfies this requirement by presenting two differentially active surfaces to cpn60 in the target polypeptide discharge reaction. Binding experiments with histidine-tagged cpn21 reveal that both the a and the /3 subunits are present in cpn60 molecules that interact with cpn21 in the presence of ADP or ATP, but binding does not occur in the absence of ADP or ATP (Baneyx et al., 1995; see Fig. 4). Irrespective of whether the two cpnlO-like domains of cpn21 have different specificities in their interactions with cpn60 a or why should the chloroplast cpn21 maintain two functional cpnlO domains fused together, rather than using separate cpnlO subunits? Many other proteins have apparently also evolved by domain duplication and fusion (McLach-lan, 1987). For example, domains are autonomous cooperative folding units, and it has been found that domain fusion enhances the rate of folding and can improve stability by reducing the entropy of the unfolded state (Liang et al., 1993). This improvement occurs because the two chains are no longer independent of each other, thus reducing the transla-tional and rotational entropy of the unfolded polypeptide. Perhaps the most compelling reason for domain fusion is that it provides a mechanism for the correct association of dissimilar, but related, subunits (Tang et al., 1979). As each domain collapses to its native-like structure it is tethered to the other domain by a peptide linker, thus ensuring a high local concentration of the fused subunits with a fixed polarity. This polarity may be important in the overall assembly of the cpn21 oligomer to achieve close packing in a sevenfold rotationally symmetrical molecule. This hypothesis implies a protein folding mechanism that requires selectivity in the binding of distinct surfaces of cpn21 to the a and ¡i forms of cpn60, to give optimum interactions and efficiency in the polypeptide discharge reaction. Clearly, a better understanding of the manner in which chloroplast cpn21 interacts with chloroplast cpn60 will require additional structural information, together with the isolation and analysis of mutated plastid cpn21. Such studies may provide an explanation for the unique double-domain cpnlO proteins found in chloroplasts.

0 0

Post a comment