ATP-, K+-dependent Heptamer Exchange Reaction Produces Hybrids between GroEL and Chaperonin from Thermus thermophilus *

Chaperonin from Thermus thermophilus(Tcpn6014· Tcpn107) splits at the plane between two Tcpn607 rings into two parts in a solution containing ATP and K+ (Ishii, N., Taguchi, H., Sasabe, H., and Yoshida, M. (1995) FEBS Lett. 362, 121–125). WhenEscherichia coli GroEL14 was additionally included in the solution described above, hybrid chaperonins GroEL7·Tcpn607 and GroEL7· Tcpn607·Tcpn107 were formed rapidly (<20 s) at 37 °C. The hybrid was also formed from Tcpn6014 and GroEL14 but not from a mutant GroEL14 lacking ATPase activity. The hybrid formation was saturated at ∼300 μm ATP and ∼300 mmK+. These results imply that GroEL14 also splits and undergoes a heptamer exchange reaction withThermus chaperonin under nearly physiological conditions. Similar to parent chaperonins, the isolated hybrid chaperonins exhibited ATPase activity that was susceptible to inhibition by Tcpn107 or GroES7 and mediated folding of other proteins. Once formed, the hybrid chaperonins were stable, and the parent chaperonins were not regenerated from the isolated hybrids under the same conditions in which the hybrids had been formed. Only under conditions in which GroEL in the hybrids was selectively destroyed, such as incubation at 70 °C, Thermus chaperonin, but not GroEL14, was regenerated from the hybrid. Therefore, the split reaction may not be an obligatory event repeated in each turnover of the chaperonin functional cycles but an event that occurs only when chaperonin is first exposed to ATP/K+.

Chaperonin from Thermus thermophilus (Tcpn60 14 ⅐ Tcpn10 7 ) splits at the plane between two Tcpn60 7 rings into two parts in a solution containing ATP and K ؉ (Ishii, N., Taguchi, H., Sasabe, H., and Yoshida, M. (1995) FEBS Lett. 362, 121-125). When Escherichia coli GroEL 14 was additionally included in the solution described above, hybrid chaperonins GroEL 7 ⅐Tcpn60 7 and GroEL 7 ⅐ Tcpn60 7 ⅐Tcpn10 7 were formed rapidly (<20 s) at 37°C. The hybrid was also formed from Tcpn60 14 and GroEL 14 but not from a mutant GroEL 14 lacking ATPase activity. The hybrid formation was saturated at ϳ300 M ATP and ϳ300 mM K ؉ . These results imply that GroEL 14 also splits and undergoes a heptamer exchange reaction with Thermus chaperonin under nearly physiological conditions. Similar to parent chaperonins, the isolated hybrid chaperonins exhibited ATPase activity that was susceptible to inhibition by Tcpn10 7 or GroES 7 and mediated folding of other proteins. Once formed, the hybrid chaperonins were stable, and the parent chaperonins were not regenerated from the isolated hybrids under the same conditions in which the hybrids had been formed. Only under conditions in which GroEL in the hybrids was selectively destroyed, such as incubation at 70°C, Thermus chaperonin, but not GroEL 14 , was regenerated from the hybrid. Therefore, the split reaction may not be an obligatory event repeated in each turnover of the chaperonin functional cycles but an event that occurs only when chaperonin is first exposed to ATP/K ؉ . GroEL 14 is an Escherichia coli chaperonin that facilitates the proper folding of proteins in an ATP-dependent manner (1-3). As determined by x-ray crystallography, GroEL 14 seems to be a hollow cylinder of 14 identical 57-kDa subunits consisting of 2 heptamer rings stacked back-to-back with a dyad symmetry (GroEL 7 ⅐GroEL 7 ) (4). GroEL 14 functionally cooperates with cochaperonin GroES 7 (5), a single heptameric ring of 10-kDa subunits that binds one end, or two ends in some cases, of the GroEL 14 cylinder (6 -8). From a thermophilic bacterium, Thermus thermophilus, GroEL homolog Tcpn60 1 (Thermus chapero-nin 60) is purified as a complex with GroES homolog Tcpn10. The complex, termed Thermus holo-chaperonin (T.holo-cpn), is composed of two heptameric rings of Tcpn60 and a single ring of Tcpn10 7 (9 -12). In contrast to chaperonins from E. coli and T. thermophilus, several members of the chaperonins, including those from Thermoanaerobacter brockii (13,14) and mitochondria (15,16), are purified as a single heptameric ring. In addition, purified chaperonin from Paracoccus denitrificans is a mixture of a large number of tetradecamers and a small number of heptamers (10). The physiological meaning of such a divergence in the quarternary structure of chaperonin has not been understood.
Recently, we found that when T.holo-cpn is incubated with ATP and K ϩ , it splits into two parts at the equator plane between the two rings of Tcpn60 7 , producing cone-shaped particles (Tcpn60 7 ⅐Tcpn10 7 ) and ring-shaped particles probably corresponding to Tcpn60 7 (17). Then we observed that the products of the split reaction can reassociate to form T.holo-cpn under the appropriate conditions (18). In contrast to the split reaction, Todd et al. (14) reported that the single-ring T. brockii cpn60 7 dimerizes to form a double-ring structure in the presence of T. brockii cpn10 7 and ATP. These results raise the question of whether GroEL 14 also undergoes the tetradecamerheptamer transition.
Here, we report that when GroEL 14 and T.holo-cpn are incubated with ATP and KCl, hybrid chaperonins such as GroEL 7 ⅐Tcpn60 7 ⅐Tcpn10 7 are formed as a result of the heptamer exchange reaction. This suggests that ATP/K ϩ -dependent transient dissociation of a tetradecamer into heptamers occurs not only in Thermus chaperonin but also in GroEL.

EXPERIMENTAL PROCEDURES
Proteins and Materials-Isopropylmalate dehydrogenase (IPMDH) from T. thermophilus strain HB8 was a kind gift from Dr. T. Oshima and his colleagues (Tokyo University of Pharmacy and Life Science, Hachioji, Japan) (19). (2R*,3S*)-3-Isopropylmalic acid, a substrate of IPMDH, was purchased from Wako Pure Chemical Corp. T.holo-cpn was purified as described previously (9,20). Tcpn60 14 expressed in E. coli was purified using modifications of procedures described previously (12). The lysate of E. coli cells containing expressed Tcpn60 14 was heated at 70°C for 20 min. The supernatant containing Tcpn60 14 was recovered by centrifugation and applied to a hydrophobic interaction chromatography (Butyl-Toyopearl). The fractions containing Tcpn60 14 were pooled and concentrated by ammonium sulfate precipitation. The concentrated protein solution was applied to a gel permeation HPLC column (G3000SWXL; Tosoh) equilibrated with 25 mM Tris-HCl, pH 7.0, 100 mM Na 2 SO 4 , and 20% (v/v) methanol. The fractions containing Tcpn60 14 were further purified by a DEAE-5PW HPLC column. Tcpn10 7 expressed in E. coli was purified as described previously (12). GroEL 14 and GroES 7 were purified as described previously (21) from * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Formation of Hybrid Chaperonin-T.holo-cpn or Tcpn60 14 (5 g) was mixed (final volume, 10 l) with GroEL 14 (5 g) and incubated at 37°C for 10 min in Buffer A (25 mM Tris-HCl, pH 7.5, 300 mM KCl, and 5 mM MgCl 2 ) containing 1 mM ATP, unless otherwise indicated. The sample solutions were applied to nondenaturing polyacrylamide gel electrophoresis (6% acrylamide), and electrophoresis was continued for about double the duration of the period required for the leading dye (bromphenol blue) to reach the front of the gel. The protein bands were stained by Coomassie Brilliant Blue. Only the regions of the chaperonin protein bands are shown in the figures.
Isolation of Hybrid Chaperonin-The ammonium sulfate precipitate of the mixture of parent and hybrid chaperonins was solubilized in a minimum volume of Buffer B (25 mM Tris-HCl, pH 7.5, and 5 mM MgCl 2 ) and applied to a Sephadex G-25 (Pharmacia) column to remove the excessive salts. The eluted protein solution was applied to a DEAE-5PW (Tosoh) column equilibrated with Buffer B and eluted with a 0 -1.0 M NaCl gradient at 1 ml/min. The chromatography was monitored by absorbance at 280 nm.
ATPase Assay-ATPase activities were assayed by measuring the amount of produced inorganic phosphate (23). Typically, the reaction was started by the addition of ATP (final concentration, 1 mM) to Buffer A containing 0.88 M 2 GroEL 14 or Tcpn60 14 and, when indicated, 1.3 M GroES 7 or Tcpn10 7 . The assay solution was preincubated for 10 min at 37°C before the addition of ATP. The reactions were terminated by the addition of perchloric acid after incubations at 37°C for 5, 10, 15, and 20 min. The solution was treated with a malachite green reagent, and the absorbance at 630 nm was measured. One unit of activity is defined as the activity that hydrolyzes 1 mol of ATP/min.
Chaperonin-promoted IPMDH Folding-IPMDH (16.2 M) denatured in 6.4 M guanidine HCl was diluted 25-fold at 37°C in Buffer A containing the components indicated in the figure legends. After a 20-min incubation at 37°C, an aliquot was withdrawn, and the reactivated IPMDH activity was determined as described previously (9).
Other Methods-Protein concentration was determined by the method of Bradford with bovine serum albumin as a standard (24). Proteins were analyzed by polyacrylamide gel electrophoresis either on a 10% polyacrylamide gel in the presence of SDS (SDS-PAGE) or on 6% polyacrylamide gels without SDS (native PAGE) (25). To obtain higher resolution on the native PAGE, electrophoresis was continued for about double the duration of the period required for the leading dye (bromphenol blue) to reach the front of the gel. Gels were stained by Coomassie Brilliant Blue R-250. 14 and Thermus Chaperonin-After we found the T.holo-cpn split (17), we attempted to identify the heptameric state of GroEL under various conditions but had no success. If GroEL 14 splits only transiently in its ATPase cycle, we could not detect the heptamer GroEL by the usual methods. Then we used Thermus chaperonin as a trap for GroEL 7 , that is, we incubated GroEL 14 with ATP/K ϩ in the presence of T.holo-cpn and examined whether the hybrid chaperonin containing GroEL 7 and Tcpn60 7 was formed. We took advantage of the different electrophoretic mobility of GroEL 14 and T.holo-cpn in native PAGE (Fig. 1A, lanes 1-5) to identify the hybrid. As shown in Fig. 1A, lane 6, two closely moving bands appeared between T.holo-cpn and GroEL 14 when they were incubated with ATP/K ϩ . The NH 2 -terminal amino acid sequences of the two bands confirmed that the upper band contained GroEL, Tcpn60, and Tcpn10 and that the lower band contained GroEL and Tcpn60 (data not shown). Yields of phenylthiohydantoin derivatives from GroEL, Tcpn60, and Tcpn10 (upper band) were generally close to each other, indicating that the upper and lower bands corresponded to GroEL 7 ⅐Tcpn60 7 ⅐Tcpn10 7 and GroEL 7 ⅐Tcpn60 7 , respectively. As described later, an analysis of isolated hybrid chaperonins supported the structures described above. For these hybrids to be formed, heptamer exchange reactions should occur, that is, both T.holo-cpn and GroEL 14 should split into heptamers that then rebind each other with a random combination into tetradecamers. Hereafter, we term the hybrid chaperonins as follows: GroEL 7 ⅐Tcpn60 7 ⅐Tcpn10 7 , Hybrid (EL-60 -10); and GroEL 7 ⅐Tcpn60 7 , Hybrid (EL-60). 3 The conditions required for the formation of hybrid chaperonins were the same as those required for the split reaction of T.holo-cpn (17); formation was absolutely dependent on ATP and K ϩ , and other combinations such as adenosine 5Ј-(␤, ␥-imino) triphosphate ϩ K ϩ (Fig. 1A, lane 7), adenosine 5Ј-O-(thiotriphosphate) ϩ K ϩ (data not shown), ADP ϩ K ϩ (lane 8), and ATP ϩ Na ϩ (lane 10) were not effective in generating hybrids. When free Mg 2ϩ was removed by trans-1,2-diaminocyclohexanetetraacetic acid, no hybrid was formed (lane 9). The relatively high concentrations of ATP and K ϩ were necessary. The hybrid formation was half-maximal at ϳ50 M ATP and ϳ50 mM K ϩ and saturated at ϳ300 M ATP and ϳ300 mM K ϩ (Fig.  1, B and C). The addition of excess Tcpn10 7 to the reaction mixture resulted in an increased yield of hybrid chaperonins with a simultaneous decrease of T.holo-cpn (Fig. 1D, lane 2), but the addition of GroES 7 had only little, if any (lane 3), effect. The hybrids were formed rapidly. In 1 min (Fig. 1E, lane 5), hybrids with an amount similar to that formed in 10 min (lane 6) were detected, and a significant amount of hybrids was formed even in 20 s (lane 4). That is as fast as a single ATPase turnover in the GroEL catalytic cycle (6).

Formation of Hybrid Chaperonin from GroEL
The ATPase activity of Thermus chaperonin at 37°C, the temperature at which hybrid formation was observed, is very low (see Fig. 5); hydrolysis of a single ATP molecule by Tcpn60 14 takes ϳ15 s. Nevertheless, hybrid chaperonin was formed within 20 s. This result means that only the hydrolysis of a single ATP by Tcpn60 14 is sufficient to form the hybrid chaperonins. This rapid formation of hybrid might be related to the observation that the formation of a tetradecamer from T. brockii cpn60 7 is also very rapid, occurring before all the cpn60 subunits could hydrolyze ATP (14). Although the real reason why the hybrid was formed in such a short period is not known, one of the possible explanations is that the initial single turnover by one cpn60 in the tetradecamer might induce a quarternary structural change in the double ring of cpn60 7 in the presence of a high concentration of K ϩ .
Cpn10 Is Not Required for Hybrid Formation-To know whether cpn10 is required for hybrid formation or not, next we used Tcpn60 14 instead of T.holo-cpn as one of the parent chaperonins. Tcpn60 14 was isolated from recombinant E. coli (12), and we confirmed that Tcpn60 14 also split in the presence of ATP/K ϩ . 4 The hybrid chaperonin was formed between Tcpn60 14 and GroEL 14 in the presence of ATP/K ϩ (Fig. 2, lane  6). Unlike the experiment using T.holo-cpn, only a single band appeared between the parent chaperonins. As expected, this band was indeed Hybrid (EL-60), because NH 2 -terminal amino acid sequencing showed that the band contained an almost equal amount of Tcpn60 and GroEL. It is likely that in the experiments described in Fig. 1, Hybrid (EL-60) was formed at first, and Hybrid (EL-60 -10) was generated next by attaching Tcpn10 7 to Hybrid (EL-60).
ATP Hydrolysis by GroEL 14 Is Essential for Hybrid Formation-In spite of the fact that the condition required for hybrid formation as described above was the same condition required for the split reaction of Thermus chaperonin (17,18), there was no direct evidence of a requirement for ATP hydrolysis by GroEL 14 for hybrid formation. To address the question, we used a GroEL mutant called GroELAEX instead of the wildtype GroEL as a parent chaperonin (21). GroELAEX is a mutant in which apical and equatorial domains in the same GroEL subunit can be cross-linked in a reversible manner (apicalequatorial cross (X)-link) (21). In the presence of a reducing reagent, GroELAEX 14 retains normal functional activity as a chaperonin. In contrast, oxidized GroELAEX 14 , which is locked in a "closed" conformation by an interdomain disulfide bond, can bind but not hydrolyze ATP (21). Therefore, the requirement for ATP hydrolysis of GroEL 14 in the hybrid formation would be tested in the presence or absence of a reducing reagent. Note that the Thermus chaperonins have no cysteine residue (12). Just like wild-type GroEL 14 , the hybrid chaperonin was formed from GroELAEX 14 and Tcpn60 14 in an ATP/ K ϩ -dependent manner under reducing conditions (Fig. 3, lane  6). However, the hybrid was not formed under the oxidizing condition (lane 2), whereas the ATPase activity of GroELAEX 14 was completely blocked (21). Wild-type GroEL 14 was able to form the hybrid with Tcpn60 14 , irrespective of reducing or oxidizing conditions (lanes 4 and 8). The inability of oxidized GroELAEX 14 to form hybrid chaperonin indicates that ATP binding is not sufficient for hybrid formation and that the occurrence of ATP hydrolysis on GroEL 14 is essential for hybrid formation.
Isolation and Characterization of Hybrid Chaperonins-The hybrid between GroEL 14 and Thermus chaperonins was separated from the parent chaperonins with anion-exchange HPLC (Fig. 4, A and B). The hybrid chaperonin fraction contained Hybrid (EL-60 -10) and Hybrid (EL-60) (see Fig. 7A, lane 5). In a similar manner, Hybrid (EL-60) was also purified (see Fig.  7A, lane 4). The relative staining intensities of the GroEL band and the Tcpn60 band in SDS-PAGE (Fig. 4, inset, lanes 1 and 2) were almost the same, again confirming that the hybrid chaperonins consisted of equal molar amounts of each chaperonin subunit. The molecular sizes of the isolated hybrid chaperonins were the same or very close to those of the parent chaperonins, because they were eluted from a gel-permeation HPLC column at the same retention time as that of GroEL 14 (data not shown). When hybrid chaperonins formed from GroEL 14 and T.holo-cpn were examined by electron micrograph, two kinds of particles, GroEL 14 -like rectangular particles and bullet-shaped particles similar to T.holo-cpn, were observed (data not shown). Hybrid (EL-60) hydrolyzed ATP at 0.09 unit/mg Ϫ1 at 37°C (Fig. 5). Because T.holo-cpn and Tcpn60 14 hydrolyzed ATP very slowly 3 The term Hybrid (EL-60 -10) does not necessarily mean that Tcpn10 7 is attached to the Tcpn60 7 ring in the hybrid. Although several lines of preliminary results, such as the protease sensitivity of each ring in the hybrid and the relatively weak affinity of Tcpn10 7 to GroEL 14 , suggest that Tcpn10 7 is at the Tcpn60 7 ring side, we cannot exclude the possibility that some Tcpn10 7 is at the GroEL 7 ring side. 4 Unpublished observations.
Chaperone Activity of Isolated Hybrid Chaperonins-We examined the effect of hybrid chaperonins on the folding of IP-MDH from T. thermophilus under a condition in which the yield of spontaneous folding was only ϳ10% (Fig. 6). The following experiments were carried out in the presence of ATP. Under this condition, GroEL 14 alone hardly promoted reactivation (less than 10% reactivation of IPMDH activity), and GroES 7 was required for effective GroEL-promoted folding. Tcpn10 7 was as effective as GroES 7 in this GroEL-promoted folding assay. In contrast to GroEL 14 , Tcpn60 14 alone was able to promote folding of IPMDH (ϳ35%). Further addition of Tcpn10 7 increased the yield of reactivation about twice, whereas the effect of GroES 7 was only marginal. Similar to GroEL 14 , Hybrid (EL-60) alone had almost no effect on folding (less than 10%), and the addition of GroES 7 or Tcpn10 7 was required for effective folding (80 -90%). In the case of the mixture of Hybrid (EL-60 -10) and Hybrid (EL-60), the folding of IPMDH was promoted moderately (ϳ35%) even in the absence of GroES 7 or Tcpn10 7 , probably due to the endogenous presence of Tcpn10 7 . The inclusion of GroES 7 or Tcpn10 7 in the solution caused additional promotion of folding (50 -60%). Thus, it is clear that both Hybrid (EL-60) and Hybrid (EL-60 -10) are active in promoting protein folding.
Parent Chaperonins Were Not Regenerated from Isolated Hybrid Chaperonins-The isolated hybrid chaperonins were very stable. After storage at 4°C for 3 weeks, about 90% of the hybrid chaperonins were still in the hybrid forms (data not shown). To investigate whether the two heptamer rings of hybrid chaperonins could reexchange each other in the presence of ATP/K ϩ , we incubated the hybrid chaperonins with ATP/K ϩ and analyzed them by native PAGE. As shown in Fig.  7B, regeneration of the parent chaperonins, GroEL 14 and T.holo-cpn (or Tcpn60 14 ), was not observed, irrespective of whether Hybrid (EL-60) or Hybrid (EL-60 -10) was used as a starting hybrid chaperonin. Further addition of either GroES 7 or Tcpn10 7 did not change the result (data not shown). This result was unexpected, because if the hybrid chaperonin split in the presence of ATP/K ϩ , as observed for Thermus chaperonin, parent chaperonins should be regenerated more or less as a result of random reassociation of heptamers. Then we examined the effects of heat (70°C for 10 min) or proteinase K treatment on the stability of the hybrid chaperonins. As shown in Fig. 7, C and D, Thermus chaperonin was resistant to both treatments under the conditions tested, whereas GroEL was destroyed. After the isolated hybrid chaperonins were incu-  1, 2, 5, and 6) or GroEL 14 (lanes 3, 4, 7, and 8) were incubated for 10 min at 37°C in Buffer A in the absence (left four lanes) or presence (right four lanes) of 5 mM dithiothreitol. When indicated, 5 g of Tcpn60 14 and 1 mM ATP were included. After the incubation, sample solutions were analyzed with 6% native PAGE as described in the legend to Fig. 1. Note that GroELAEX 14 is electrophoresed at a slightly faster mobility than GroEL 14 under oxidizing conditions (21).

FIG. 4. The isolation of hybrid chaperonins with anion-exchange HPLC.
A, T.holo-cpn and GroEL 14 were mixed and incubated at 37°C for 10 min in Buffer A containing 1 mM ATP. The chaperonins were separated by HPLC. A DEAE-5PW (Tosoh) column was equilibrated with 25 mM Tris-HCl (pH 7.5) and 5 mM MgCl 2 , and a linear gradient from 300 -500 mM NaCl was applied. Bars, scales of absorbance at 280 nm. B, the fraction of the middle peak in trace A (indicated by a bar) was pooled and then rechromatographed as described in A. Inset, the isolated hybrid chaperonins (Hybrid (EL-60), lane 1; Hybrid (EL-60 -10), lane 2) were analyzed with 10% SDS-PAGE. Note that we used a 10% polyacrylamide gel to separate GroEL 14 from Tcpn60 14 where a Tcpn10 7 band was not seen in this electrophoresis. The presence of a Tcpn10 7 band in the sample solution used in lane 2 was confirmed separately in 15% SDS-PAGE (data not shown). bated at 70°C for 10 min, the hybrid chaperonins disappeared completely, and Thermus chaperonin but not GroEL 14 appeared (Fig. 7C, lanes 4 and 5). Treatment with proteinase K also gave the same results (Fig. 7D, lanes 4 and 5). These results indicate that only when the GroEL 7 moiety of the hybrid is denatured or proteolyzed, survived heptamer rings of Tcpn60 can reassociate to form Tcpn60 14 or T.holo-cpn.

DISCUSSION
Hybrids Are Formed as a Result of Heptamer Exchange-In this report, we demonstrated ATP/K ϩ -dependent formation of the hybrid between GroEL 14 and Thermus chaperonin. This is a result of the heptamer exchange reaction between both chaperonins. One can argue the possibility that the hybrids are made up from three stacked heptamers (GroEL 14 ⅐Tcpn60 7 ), that is, that GroEL 14 simply binds Tcpn60 7 as a substrate protein. This possibility was excluded from the analysis of molecular size with gel-permeation HPLC, the observation of molecular shape by electron micrograph, and the estimation of the molar ratio of GroEL:T.cpn60 from phenylthiohydantoin derivatives recovered in Edman degradation and from the staining intensity of the bands in SDS-PAGE (Fig. 4, inset). The random incorporation of each cpn60 monomer into the tetradecamers is also very unlikely. If it really happened, numerous protein bands corresponding to the complexes with various combinations of parent chaperonin monomers should have appeared between the bands of parent chaperonins in native PAGE. However, only a single band appeared between parent chaperonins when the hybrid was formed from GroEL 14 and Tcpn60 14 (Fig. 2). In addition, under the experiment conditions, bands of monomeric GroEL and monomeric Tcpn60 with meaningful staining intensity were not observed in native PAGE (data not shown). Therefore, there is little possibility, if any, that dissociation into monomers occurs before formation of the hybrid.
Hybrid formation is not restricted to the combination of Thermus chaperonins and GroEL and is also observed between Tcpn60 14 and chaperonin from P. denitrificans (10) in the pres-ence of ATP and K ϩ . 4 In addition, Burston et al. (27) reported that the hybrid between wild-type GroEL 14 and mutant GroEL 14 , called MR1 (mixed ring), is formed at 42°C in the presence of ATP and K ϩ . These observations suggest that hybrid formation is not an exceptional event but a rather common reaction among chaperonins from several species. Because the conditions for hybrid formation are nearly physiological, hybrids can be formed even in a living cell. Indeed, when we expressed Tcpn60 14 in E. coli cells, a part of the chaperonin was purified as a form of hybrid chaperonin that was separated from Tcpn60 14 by anion-exchange HPLC (see "Experimental Procedures"). 4 Furthermore, the observation that single-ring T. brockii cpn60 dimerizes to a tetradecamer in the presence of both adenine nucleotides and cpn10 (14) also implies that there is a heptamer exchange in the bacterium.
Heptamer Exchange Necessitates the Split of GroEL 14 -Unless GroEL 14 and Tcpn60 14 split into GroEL 7 and Tcpn60 7 , hybrid formation is impossible. The ATP/K ϩ -dependent split of Thermus chaperonin has been established, and Tcpn60 7 can be readily detected and isolated (17). A stable heptameric form of cpn60 has also been isolated from T. brockii (13,14) and from mitochondria (15,16). A possible heptameric form of GroEL has been reported by two groups. Mendoza et al. (28) suggested that a GroEL species, possibly heptamers, is detected in the presence of 3 M urea after the addition of unfolded rhodanese. Mizobata and Kawata (29) observed a GroEL species exhibiting decreased light scattering in the presence of less than 1 M guanidine HCl, and they speculated that the species is a heptamer of GroEL (29). We also observed a decrease in light scattering of GroEL 14 in response to the addition of ATP. 4 The fact that the hybrid is formed between wild-type GroEL 14 and mutant MR1-GroEL 14 under the appropriate conditions implies that not only mutant MR1-GroEL 14 but also wild-type GroEL 14 splits into heptamers (27). Therefore, GroEL 14 most likely splits, although the final conclusion on the presence of GroEL 7 should be reserved until success in isolating the heptameric form of GroEL has been achieved.
However, the Split May Not Be an Obligatory Step in the Functional Reaction Cycle of Chaperonin-Once the hybrid is formed, it is very stable. Parent chaperonins are not regenerated from hybrid chaperonin even in the presence of ATP/K ϩ (Fig. 7). If the split into heptamers is an obligatory step in the functional reaction cycle of the chaperonin, parent chaperonin The column labeled spontaneous represents the reaction without chaperonin. After dilution, ATP was added to the solution (final concentration, 1 mM) to initiate the folding reaction. When added, the concentration of co-chaperonins (GroES 7 and Tcpn10 7 ) was 1.7 M each (as an oligomer). After a 20-min incubation in the presence of ATP, recovered IPMDH activity was measured. An activity of the same amount of native IPMDH was taken as 100%.
FIG. 7. The stability of hybrid chaperonins under various conditions. A and B, isolated hybrid chaperonins (4 g of each) were incubated at 37°C for 10 min in Buffer A in the absence (A) or presence (B) of 1 mM ATP. C, isolated hybrid chaperonins (4 g of each) were incubated at 70°C for 10 min in Buffer A. D, isolated hybrid chaperonins (4 g of each) were treated with 0.2 g of proteinase K at 25°C in Buffer A. After a 30-min incubation, phenylmethylsulfonyl fluoride (final concentration, 5 mM) was added to stop the proteolysis. Sample solutions were analyzed with 6% native PAGE as described in the legend to Fig. 1. should be formed as a result of a heptamer exchange reaction. Two explanations are possible: (a) split reaction occurs only once before the chaperonin starts the first turnover of the functional cycle; or (b) split reaction occurs in each of the reaction cycles, but reassociation always happens to heterologous combinations (GroEL 7 -Tcpn60 7 ) rather than homologous combinations (GroEL 7 -GroEL 7 or Tcpn60 7 -Tcpn60 7 ), thus producing the hybrid again. However, the latter possibility is unlikely, if not impossible, because the parent GroEL 14 is not regenerated from the hybrid GroEL 7 ⅐MR1-GroEL 7 in the presence of ATP/K ϩ (27), and it is not easy to assume that wild-type GroEL 7 has a much higher affinity to mutant MR1-GroEL 7 than it does to wild-type GroEL 7 .
A requirement for high concentrations of ATP and K ϩ is also contradictory to the notion that the split is one of obligatory steps in the chaperonin functional cycle. Steady-state ATPase activity of GroEL 14 is inversely dependent on K ϩ ; it is saturated at ϳ5 mM K ϩ in the presence of 50 M ATP and at ϳ300 mM K ϩ in the presence of 2 M ATP (26,30). On the contrary, hybrids were formed only when concentrations of both K ϩ and ATP were high, and the yield of hybrids was saturated at ϳ300 mM K ϩ and ϳ300 M ATP (Fig. 1, C and D). If either one of the concentrations was reduced, the yield of hybrids decreased, and no hybrid formation was observed at 5 mM K ϩ /1 mM ATP or at 300 mM K ϩ /2 M ATP. Therefore, the requirement of K ϩ for hybrid formation is a different phenomenon than the K ϩ requirement for steady-state ATPase activity. Although the occurrence of the heptamer exchange reaction has been established, understanding of its functional and physiological significance awaits further study.