Archaeal-like chaperonins in

Stephen M. Techtmann1 and Frank T. Robb2

aInstitute of Marine and Environmental Technology, Program in the Biology of Model Systems, 701 East Pratt Street, Baltimore, MD 21202 and Department of Microbiology and Immunology, University of Maryland School of Medicine, 685 West Baltimore Street, Baltimore, MD 21201

Edited by Arthur L. Horwich, Yale University School of Medicine, New Haven, CT, and approved September 14, 2010 (received for review April 16, 2010)

Chaperonins (CPN) are ubiquitous oligomeric protein machines that by preventing aggregation of the misfolded substrates (16–18). mediate the ATP-dependent folding of polypeptide chains. These Additionally, the active Iterative Annealing Model proposes that chaperones have not only been assigned stress response and nor- the misfolded protein is refolded by several rounds of binding, mal housekeeping functions but also have a role in certain human release and reengagement (19–21). disease states. A longstanding convention divides CPNs into two While both groups of chaperonins catalyze protein folding via groups that share many conserved sequence motifs but differ in encapsulation, there are key structural and mechanistic differ- both structure and distribution. Group I complexes are the well ences between the two groups. The Group II chaperonins (also known GroEL/ES heat-shock proteins in bacteria, that also occur known as the thermosome or TCP1) are structurally different in some of mesophilic archaea and in the endosymbiotic from Group I, the GroEL/GroES complexes. Both Group I organelles of eukaryotes. Group II CPNs are found only in the and Group II complexes are composed of two stacked rings. cytosol of archaea and eukaryotes. Here we report a third, diver- The Group II CPNs are octo- or nonomeric rings (10), whereas gent group of CPNs found in several species of bacteria. We the GroEL is a homotetradecamer. Group II CPNs are often het- propose to name these Group III CPNs because of their distant erooligomers comprised of a combination of multiple subunits relatedness to both Group I and II CPNs as well as their unique (10). For example many Group II chaperonins from eukaryotes genomic context, within the hsp70 operon. The prototype Group III are heterohexadecamers composed of eight different subunits. CPN, hydrogenoformans chaperonin (Ch-CPN), Additionally, the function of the Group I cochaperone, GroES, is able to refold denatured proteins in an ATP-dependent manner is replaced in Group II CPNs by a helical protrusion in the apical and is structurally similar to the Group II CPNs, forming a 16-mer domain of each subunit. These flexible domains combine to form with each subunit contributing to a flexible lid domain. The Group a “built-in lid” that is essential for encapsulation of the denatured BIOCHEMISTRY III CPN represent a divergent group of bacterial CPNs distinct from protein (22–24). The opening and closing of this lid is ATP- the GroEL/ES CPN found in all bacteria. The Group III lineage may responsive and is allosterically regulated (24, 25). Recent work represent an ancient horizontal gene transfer from an archaeon has also indicated that, while certain CPN domains share striking into an early Firmicute lineage. An analysis of their functional similarity between Group I and Group II, the extent and types and structural characteristics may provide important insights into of structural rearrangements that occur during the CPN folding the early history of this ubiquitous family of proteins. cycle vary greatly (8, 10, 22, 25, 26).

evolution ∣ protein-folding ∣ thermophilic Results Phylogeny of Group III Chaperonins. Recently, we annotated a diver- olecular chaperones have an essential role in ensuring gent class of chaperonin genes in the genomes of two carboxydo- Mproper folding of cellular proteins. Two primary protein- trophic bacteria, as encoding TCP1-like chaperonins. As shown in folding machines, Hsp70 (DnaK) and the chaperonin (CPN) are Fig. 1, these CPNs are conserved in the genomes of 11 bacteria present in the majority of living organisms (1–6). These two and are monophyletic and distant from all known Group I and systems constitute the ATP-dependent folding functions of both Group II chaperonins. Phylogenetic analysis using multiple algo- nascent polypeptides as well as the salvage of stress-denatured rithms revealed that these proteins form a deeply rooted clade, proteins (7–10). and are highly divergent from the other known chaperonins with Chaperonins are able to positively influence refolding through robust bootstrap support (Fig. 1; individual trees, Fig. S1). Trends sequestration of nonnative proteins. The general mechanism in the divergence of these proteins mirror those seen in the phy- whereby chaperonin affect refolding is through the encapsulation logenetic trees (Table S1 and Fig. S2). There are several deletions of denatured protein within a hydrophilic cavity. While the exact in Group III CPN relative to the other groups of CPNs (Fig. S3). mechanism of CPN-mediated protein folding is still unclear, These indels account for the some of the sequence divergence much is known about key steps in the CPN reaction cycle. The and may influence localized structural elements. This clade repre- most well studied chaperonin homolog is the Escherichia coli sents a previously unknown and unique group of CPNs whose GroEL (8). GroEL forms a barrel structure composed of two physiological role has yet to be determined and which we propose seven-membered rings, with each ring containing a folding cavity. to name as Group III CPNs. The significant divergence from Controversy still prevails over the exact mechanism by which the two known classes of CPNs leads us to suggest that this group GroEL catalyzes proteins folding. According to one model, is a monophyletic subfamily of the Group II CPNs, which may partially folded proteins bind to hydrophobic residues that line represent the remnant of an ancestral chaperonin group. the GroEL cavity (11). ATP binding causes large conformational changes that simultaneously draw the nonnative protein into the Author contributions: S.M.T. and F.T.R. designed research; S.M.T. performed research; cavity, increase the number of polar residues within the cavity, S.M.T. and F.T.R. analyzed data; and S.M.T. and F.T.R. wrote the paper. and trigger rapid binding of the Cochaperone GroES (12). This The authors declare no conflict of interest. sequestration in a hydrophilic cavity forces the protein to refold This article is a PNAS Direct Submission. through forcing the internalization of the hydrophobic residues of 1Present address: Department of Biochemistry and Molecular Biology, Uniformed Services the nonnative protein (13, 14). ATP hydrolysis in the GroES- University for the Health Science, 4301 Jones Bridge Road, Bethesda, MD 20814. bound ring, followed by ATP binding to the opposite ring triggers 2To whom correspondence should be addressed. E-mail: Frobb@som. the release of GroES followed by the encapsulated protein. There umaryland.edu. are several alternative models for GroEL activity (15), including This article contains supporting information online at www.pnas.org/lookup/suppl/ the passive model whereby GroEL assists in refolding simply doi:10.1073/pnas.1004783107/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1004783107 PNAS Early Edition ∣ 1of6 Downloaded by guest on September 25, 2021 Fig. 1. Phylogenetic analysis of Group III chaperonin: Un- rooted minimum evolution phylogenetic tree representing Group I, Group II, and the new Group III chaperonins, indi- cating the great phylogenetic distance between the Group III clade and the conserved Group I and Group II clusters. The same alignment was used to construct trees using multiple algorithms (minimum evolution, maximum parsimony, and neighbor-joining). Nodes present in trees created by different alogorithms and with high bootstrap values are denoted as described. The clade representing Group III CPN contains the following organisms: Carboxydothermus hydrogenoformans, Geobacillus spY412MC, Syntrophomo- nas. wolfei wolfei Goettingen, Desulforudis audaxviator, , Heliobacterium modesticaldum, Ther- mosinus carboxydivorans, acetoxidans, Desulfotomaculum reducens, thermopro- pionicum, and Gloeobacter violaceus.

Genomic Context of Group III Chaperonins. All of the currently unique group of chaperonins we cloned, expressed, and charac- identified Group III chaperonins, with one exception, are found terized a prototype Group III CPN from the extreme thermophile internal to the Hsp70 molecular chaperone operon (Fig. 2A). Carboxydothermus hydrogenoformans (Ch-CPN). C. hydrogenofor- These species all have conventional Group I CPNs encoded in mans is a member of the phylum and is known for its separate operons with their GroES cochaperone. The Group ability to grow on carbon monoxide as its sole carbon and energy III chaperonins are the only CPNs found near or in association source, linking the oxidation of carbon monoxide to hydrogen with the Hsp70 operon. This unique operon structure and the unprecedented association with Hsp70 provide evidence for a functional distinction between Group II and Group III CPNs. Group III CPNs are found between hrcA, the heat-shock responsive transcriptional repressor, and the Hsp70 operon (grpE, dnaK, dnaJ). The location of the Group III chaperonin downstream of the hrcA suggested that Group III CPNs could be HrcA regulated. The Group I CPN and Hsp70s of many bacteria are HrcA regulated (27–30). In silico analysis revealed the presence of a CIRCE consensus site upstream of the HrcA as well as the Group I CPN in these genomes (Fig. 2C). CIRCE (Controling IR of Chaperone Expression) is the DNA sequence found in many heat-shock responsive promoters that is recog- nized by the HrcA transcriptional regulator. The unique operon structure and the apparent coregulation with the Hsp70 argues for a functional relationship between Group III CPN and Hsp70. Fig. 2. Genomic context of the Group III chaperonin: (A) The genomic Unlike the other Group III CPNs, the Group III CPN gene context of all but one of the Group III chaperonins. As shown, the Group in Gloeobacter violaceus is not associated with any other heat- III CPN (cpn) is downstream of hrcA and is adjacent to the Hsp70 operon shock-related genes (Fig. 2B). The G. violaceus CPN also forms (dnaK, dnaJ, grpE) This topology is found in C. hydrogenoformans, Geoba- cillus spY412MC, S. wolfei wolfei Goettingen, D. audaxviator, A. degensii, the deepest branch of the Group III CPN clade. H. modesticaldum, T. carboxydivorans, D. acetoxidans, D. reducens, and P. thermopropionicum. (B) The genomic context of G. violaceus. This CPN Expression of mRNA Encoding the Carboxydothermus hydrogenofor- is not in close association with any other chaperones. (C) The CIRCE sequences mans CPN. To investigate the protein-folding functions of this upstream of the Group I and Group III CPN from C. hydrogenoformans.

2of6 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1004783107 Techtmann and Robb Downloaded by guest on September 25, 2021 production (31, 32). The genome sequence of C. hydrogenofor- The reactions were incubated at 42 °C for 30 min with aliquots mans provided the prototype gene encoding the Group III CPN removed at intervals and assayed for MDH activity. (32). It is important to note that genomes encoding Group III No spontaneous refolding of denatured MDH was observed at CPNs each contain a single copy of the Group III and Group I 42 °C. Additionally, MDH was not refolded in reactions where CPNs. Interestingly, seven of the eleven organisms encoding Ch-CPN alone was added. A similar phenomenon was seen with Group III CPNs have the ability to carry out anaerobic CO oxi- the GroEL/ES from archaeon Methanosarcina mazeii, which, by dation, like C. hydrogenoformans. The first step in understanding itself was unable to refold MDH as noted by Figueiredo et al. the physiological significance of the Group III CPN was to (36). This inability was overcome by the addition of 0.5 M ascertain whether this ORF is expressed under normal growth ammonium sulfate to the refolding mixtures, allowing the Ma conditions. GroEL/GroES to refold MDH efficiently. The addition of 0.5 M RNA was extracted from C. hydrogenoformans cells grown ammonium sulfate to the Ch-CPN reaction mixture resulted in under normal growth conditions. RT-PCR was performed with CPN-dependent refolding of denatured MDH (Fig. 3C). Because primer pairs specific to both the Group I CPN (GroEL) and the MDH activity requires formation of the homodimer, it is possible Group III CPN from C. hydrogenoformans. RNA for both Group that ammonium sulfate promotes efficient refolding and dimer- I and Group III genes was detected under normal growth condi- ization of MDH. In these experiments, MDH refolding occurred tions (Fig. 3A), demonstrating that both the Group I and Group in a GroES-independent manner similar to other Group II CPN III CPN are expressed in vivo during normal growth at 72 °C. (37), establishing a distinct difference between Ch-CPN and other known bacterial chaperonins (Group I). Characterization of the Chaperonin Activity of the C.h.-CPN. Recom- To further confirm chaperone activity, E. coli survival assays binant Ch-CPN assembled into a highly thermostable hexade- were performed. E. coli BL21(DE3) expressing Ch-CPN were camer and was purified to homogeneity. Enzyme protection, heated to 50 °C. After 2 h the viable cell count in the empty vector refolding, and E. coli survivability assays were applied to deter- control cultures was four orders of magnitude lower than at mine whether Ch-CPN was an active chaperone. As shown in t ¼ 0. The E. coli expressing Ch-CPN declined in viability by less Fig. 3B, salvage activity of Ch-CPN was tested using glutamate than one order of magnitude after 2 h. This data demonstrates dehydrogenase (GDH) protection assays (33–35). Bovine GDH that Ch-CPN is able to protect live E. coli cells from heat killing is extremely heat-labile and is rapidly inactivated at 50 °C. Native at 50 °C (Fig. 3D). These three experiments confirm the annota- bovine GDH was mixed with the Ch-CPN and ATP and heated tion of the Ch-CPN as an ATP-dependent chaperonin and de-

at both 42 °C and 50 °C for 2 h. GDH alone was largely inacti- monstrate that Ch-CPN is capable of protecting and refolding BIOCHEMISTRY vated after 1 h at 42 °C and after 30 min at 50 °C (Fig. 3B). In proteins in a GroES-independent manner. mixtures of GDH plus Ch-CPN at 42 °C the Ch-CPN was able to protect the GDH completely for 2 h, whereas the activity in the Determination of Key Structural Features of Group III CPN. To deter- control declined to 5% of the original value. At 50 °C, GDH with mine the mechanism whereby Ch-CPN is capable of refolding Ch-CPN retained about 40% of activity after 2 h whereas the con- substrates in the absence of GroES we undertook a series of ex- trol had no detectable activity after 30 min of incubation. A con- periments to identify whether a built-in lid was present. Structural trol without ATP confirmed that the ability of Ch-CPN to protect modeling of the Ch-CPN with the Group II chaperonin from GDH was greatly decreased with the omission of ATP. These re- Thermoplasma acidophilum revealed the presence of a putative sults demonstrate the ability of Ch-CPN to act as a molecular cha- helical protrusion (Fig. 4A and Fig. S4). To confirm the modeling perone and to salvage bovine GDH following heat denaturation. data, we examined the mobility of Ch-CPN on native polyacryla- Many chaperonins are able to protect native proteins, and mide gels. Quaite-Randall et al. (38) showed that pure Group II also to refold nonnative proteins de novo. To test the refolding Sulfolobus shibatae chaperonin (Ss-CPN), when resolved on a activity of Ch-CPN, guanidine denatured porcine heart malate native polyacrylamide gel, separated into two bands. These two dehydrogenase (MDH) was mixed with Buffer A (25 mM Hepes bands were purified from one another and electron microscopy pH 7.2, 300 mM NaCl, 1 mM MgCl2) with and without Ch-CPN. was used to study the conformational differences. Quaite-Randall

Fig. 3. Chaperone activity: (A)RT-PCRof the Group I and Group III chaperonins from C. hydrogenoformans.(B) GDH protection assays with Ch-CPN at both 42 °C (closed symbols) and 50 °C (open symbols). One No-ATP control was performed at 50 °C (open triangles). (C) Malate dehydrogen- ase refolding assays. Denatured MDH protein was mixed with Ch-CPN and ATP to determine Ch-CPN’s ability to refold de- natured proteins. The original (predena- turation) activity was used to determine 100%. No MDH refolding was observed with Ch-CPN alone (open symbols). Refold- ing was observed when 0.5 M AS was added (closed symbols). (D) E. coli survival assays. E. coli BL21 expressing Ch-CPN (open symbols) and E. coli BL21 with empty vector (closed symbols) were heated to 50 °C for 2 h. Viable counts were taken at time intervals.

Techtmann and Robb PNAS Early Edition ∣ 3of6 Downloaded by guest on September 25, 2021 et al. (38) identified the two bands as conformers of the complex Discussion in which the built-in lids were either opened or closed. The com- In this study we performed a biochemical characterization of a plexes were all converted to the open (slow) form upon addition prototype Group III CPN. This characterization revealed that of ATP. while there is great sequence divergence from Group II CPNs The Ch-CPN separated into two bands when run on a native as a whole, many of the structural and functional properties ap- polyacrylamide gel, resembling the data for Ss-CPN (Fig. 3B). pear to be shared. Group III CPN are able to refold denatured Ch-CPN complexes were incubated with different components proteins in a GroES-independent manner, a feature not seen of buffer A and ATP at 65 °C for 50 min and applied to a native before in any bacterial CPNs. Native PAGE experiments, struc- polyacrylamide gel to separate the closed (fast) and open (slow) tural modeling, and negative-stained EM image analysis reveal forms of the chaperonin (Fig. 4B). Two bands are present in all a typical articulating built-in lid, which is a characteristic feature of the lanes without ATP and a single slow band was observed of Group II CPNs. This structure explains the ability of Group III in lanes 4, 6, 7, and 8, after incubation with ATP, similar to the CPN to refold proteins in the absence of GroES. Additionally, results seen by Quaite-Randall et al. (38). This finding demon- the Group III CPNs are composed of two stacked octomeric strates an ATP-dependent conformational change typical of rings, unlike any bacterial CPNs, which have seven-membered Group II chaperonins. The presence of a built-in lid accounts rings. Despite extensive sequence divergence, the structural and for the ability of Ch-CPN to refold denatured proteins in a functional properties of this unique group of CPNs are similar to GroES-independent manner. Group II CPN. Electron microscopy and image averaging was used to further The occurrence of this group of CPNs in bacteria suggests that validate the presence of a built-in lid. A 20 μM solution of ATP- they might shed light on the evolution of chaperonin systems. incubated Ch-CPN was adsorbed to glow discharged carbon Based on phylogenetic analysis alone, the group III CPN appear coated grids and stained with 1% uranyl acetate. These grids were to be a monophyletic group of CPNs distinct from extant Group I imaged on a Tecnai T12 TEM at 120 kV at a magnification of and Group II proteins. A phylogenetic analysis of all known CPN 52;000×. Images were analyzed using the EMAN software pack- sequences performed by Williams et al. (2010) (40) led these age (39). Approximately 12,000 particles were picked and aver- authors to suggest that Group III CPN formed as a subgroup aged. Class averaged images reveal the presence of a built-in lid of the Group II CPN. Williams et al. proposed that these bacterial (Fig. 4C white arrows). This lid domain is a key feature shared Group II CPNs resulted from a recent horizontal gene transfer between the Group II and these Group III CPNs. into bacteria from the methanoarchaea. Our findings on the func- Another key difference between the Group I and Group II tional and structural similarities of the Group III and Group II CPN is the multimeric structure of the ring complex. The CPN do support the conclusion that the unique group is more class-averages from top-views of CPN indicated that the Ch-CPN closely related to Group II CPNs than to Group I CPNs. How- was composed of two octameric rings. This finding was confirmed ever, the unique genomic context of the Group III suggests that by sizing the Ch-CPN using multiangle light scattering with the the conclusion may need to be reevaluated. MiniDawn Treos instrument. The size of the Ch-CPN complex Our analysis of the genomic context of the Group III CPN was 937.3 kDa 1.8 kDa. The oligomer was determined to be revealed that the Group III CPN are found in operon with the composed of sixteen subunits, based on a calculated monomeric DnaK and its cochaperones (DnaJ and GrpE), except for one molecular mass of 56.6 kDa, thus indicating a composition of of the organisms, Gloeobacter violaceus. It is important to note double-stacked octomeric rings. This finding demonstrates addi- that while all bacteria and eukaryotes have DnaK homologs, they tional structural similarities between Ch-CPN and the Group II are absent in many archaea, including most of the hyperthermo- chaperonins. philes (2). In order to further clarify the coevolution of the Group

Fig. 4. Evidence for a helical protrusion in the Group III chaperonin. (A) Structural model of C. hydrogeno- formans CPN based on the structure of the T. acidophi- lum Group II chaperonin, constructed using the Wurst program. The structure is colored based on conserva- tion to all CPN in the alignment. Molecular graphics images were produced using the UCSF Chimera pack- age from the Resource for Biocomputing, Visualiza- tion, and Informatics at the University of California, San Francisco (49). (B) Native PAGE of Ch-CPN incu- bated at 65 °C with various components of buffer A and ATP. The two bands representing two forms of Ch-CPN (open and closed). Upon addition of ATP only the open form is observed. (C) Representative class averaged images of side-views of Ch-CPN elucidating the presence of built-in lid (indicated by white arrows). These side views are averages of 411, 353, and 298 images respectively. Top view is the average of 245 images. All of the classes generated during image averaging are in (Fig. S5).

4of6 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1004783107 Techtmann and Robb Downloaded by guest on September 25, 2021 III CPNs and their associated DnaK homologs, they were aligned The most parsimonious model for the acquisition of the Group with representative bacterial and archaeal DnaK proteins. The III CPN is one in which they originated in this group of Firmicute maximum likelihood phylogenetic analysis of this collection of bacteria as a result of a horizontal gene transfer from an ancient homologs indicates that the DnaK genes associated with the archaeon into the ancestor of this group of Firmicutes. As the Group III CPN gene forms a distinct clade together with the DnaK group diverged, it was lost in some members, yet was maintained genes from closely related Firmicutes (Fig. 5 and Fig. S6). The in several species. For example, all but one member of the DnaK tree topology mimics the tree topology for the GroELs from sequenced contain Group III CPN. Genomic the Group III-containing organisms. As previously described, the rearrangement in the Group III-containing organisms resulted archaeal sequences cluster into three distinct groups (41), none of in the placement of the Group III CPN in operon with DnaK which is closely related to the specialized DnaK homologs asso- chaperone. This gene arrangement would allow for the DnaK ciated with Group III CPNs. and Group III CPN to be coordinately expressed from the same It is generally accepted that the DnaK in the archaea origi- promoters and may provide a selective advantage for these nated via horizontal gene transfer from the bacteria (6, 41). Gen- organisms. erally DnaK genes occur en bloc with genes encoding the Of the 11 Group III CPN currently described, Gloeobacter violaceus is the only exception in which the Group II CPN is cochaperones GrpE and DnaJ. Therefore, if a Group III CPN not found in operon with any other molecular chaperone, the gene cluster was recently transferred from an archaeon into a gene topology of most archaeal Group II CPNs. Additionally, Firmicute bacterium, we would expect that the gene order in the DnaK and GroEL from G. violaceus are highly divergent from the Firmicutes should be present in some archaea. However, the DnaK and GroEL of the other Group III CPN containing the gene order found in the Group III CPN clusters has not been organisms. G. violaceus may represent an example of a recent found to date in any archaeal genomes. Because all bacteria horizontal gene transfer of a Group II chaperonin from an encode DnaKs, if the cluster containing a Group III CPN was archaeon. recently acquired from the archaea, we would expect there to All Group III-containing organisms have both Group I and be multiple DnaK encoding genes in the Group III CPN contain- Group III CPNs, implying that there is a selective advantage to ing organisms. However, Group III CPN containing bacteria do maintaining both groups. Williams et al. (2010) (40) suggest that not typically have multiple DnaK genes. Seven of the eleven the Group III CPNs may have a specialized group of target pro- Group III CPN containing genomes encode single DnaK genes teins. Because the majority of the Group III CPN have the ability within the Group III CPN clusters. If this operon were trans- to perform anaerobic carbon monoxide oxidation, it is possible BIOCHEMISTRY ferred from the archaea, it would require the loss of DnaKs in that the Group III CPNs are involved in folding proteins involved seven of the eleven members of this group. This argues strongly in anaerobic carbon monoxide oxidation. Hirtreiter et al. (2009) against a recent horizontal gene transfer of this operon from the (42) analyzed the substrate specificity of the Group I and Group archaea into the bacteria. II CPN in the archaeon Methanosarcina mazeii. This study sug- gests that the two classes of CPN in M. mazeii each fold a distinct set of client proteins. Hirtreiter et al. (42) suggest that the distinc- tion between the Group I and Group II CPN in these archaea may be in the metabolic proteins that they fold, but more likely the cochaperones with which the CPNs associate. This finding suggests the following hypothesis for the specialized roles of the Group III system. We propose that the distinction between the Group I and Group III CPN is the association of the Group III CPN with the DnaK. There is prior evidence that the eukar- yotic CPN interacts directly with the DnaK homolog Hsc70 (43). Like the eukaryotic homolog, the Group III CPN may be core- gulated and function directly with the Hsp70 homolog. In eukar- yotes the CPN-Hsc70 association allows for more efficient transfer of substrate proteins from the Hsp70 to the CPN. This direct association may also be the case in these bacteria and these systems might be a prokaryotic model system to study the asso- ciation of the Hsp70 and CPN. In conclusion, this study indicates that the unique group of CPNs are distantly related to the Group II CPN, representing an ancient lateral gene transfer from the archaea into an ances- tral bacterium, and appear to have a close association to the Hsp70 unlike any known prokaryotic CPNs. Methods Evolutionary Analyses and Construction of Phylogenetic Trees. Multiple se- quence alignment was done using various HSP60s in Clustal X. This alignment was imported into MEGA 4. Evolutionary analyses were performed in Mega 4, which was used to process Clustal X alignments and to construct and render trees. Minimum evolution trees were calculated with Poisson correc- tion, complete deletions in gap-containing columns. Tree topology was ver- ified by bootstrap analysis using 500 iterations and checked for consistency. Fig. 5. Phylogenetic tree of DnaK homologs from Group III CPN-containing Maximum Parsimony and Neighbor-joining trees were constructed in MEGA organisms. A maximum liklihood tree was constructed using the PhyML pro- 4 using similar conditions and bootstrapped with 500 replicates. Evolutionary gram. DnaK sequences from the Group III CPN gene cluster were aligned with distances in Figs. S1–S6 and SI Table S1 were calculated in Mega 4 using all archaeal DnaK homologs and those of some bacteria. Clades composed of Poisson correction and the pairwise deletion option, meaning all positions all bacteria are colored blued. Clades comprised of archaea are colored red. containing alignment gaps and missing data were eliminated only in pairwise The Group III CPN-associated DnaKs cluster along with other Firmicutes in a sequence comparisons. Maximum likelihood trees were constructed using clade distinct from any archaeal DnaKs. (Full tree is in Fig. S6) the PhyML program (44).

Techtmann and Robb PNAS Early Edition ∣ 5of6 Downloaded by guest on September 25, 2021 Glutamate Dehydrogenase Protection Assays. Bovine GDH (Sigma Aldrich) was E. coli BL21 Survival. A single colony of E. coli BL21 containing the Ch-CPN used in protection assays. Experiments were carried out as described in Lak- expression plasmid was inoculated into auto-induction medium and grown sanalamai et al. (34) with the following changes. The final concentration of overnight at 37 °C shaking 200 rpm. The same was done for E. coli BL21 Ch-CPN was 0.05 mg∕mL. These experiments were carried out with a molar containing the empty pET 19b plasmid. The optical density 600 nm was re- ratio of GDH hexamers to Ch-CPN complexes of 3∶2, theoretically allowing corded. Cell-free extracts were resolved on SDS PAGE to confirm expression each GDH to be sequestered in one Ch-CPN complex. Experiments were done of Ch-CPN, which was approximately 10% of the soluble protein. Both cul- at both 42 °C and 50 °C. A control experiment was done at 50 °C with no ATP tures were diluted to an OD 600 nm of 0.8 with sterile auto-induction med- to determine the ATP-independent protection of GDH. In the ATP experiment ium. The diluted cultures were aliquoted into three borosilicate glass tubes the final concentration of ATP was 2 mM and the final concentration of and rapidly equilibrated to 50 °C. Samples were removed at 0, 30, 60, and MgCl2 was 1 mM. 120 min. Samples from each time point were diluted and plated. These viable counts were used to determine the average percent of survivability relative Malate Dehydrogenase Refolding Assays. Porcine heart MDH (Amresco) was to the initial colony forming units. The log of percent survivability vs. time is denatured in Buffer A (25 mM Hepes pH 7.2, 300 mM NaCl, 1 mM MgCl2) shown, as a measure of cell death relative to the initial colony count. containing 4 M Guanidine-HCl and 5 mM DTT at 37 °C for 1 h. Denatured Detailed procedures for RT-PCR, cloning and expression of Ch-CPN, MDH was diluted 100-fold into Buffer A supplemented with 1 mM ATP, with structural modeling, and electron microscopy image analysis are provided or without 0.5 M ammonium sulfate (AS) at 42 °C in the presence and absence in the SI Methods. of 69 nM Ch-CPN. The concentration of denatured MDH was 500 nM. The ratio of Ch CPN to denatured MDH was approximately 1∶10. ACKNOWLEDGMENTS. We thank Dr. Ru-Ching Hsia for assistance with The reactions were incubated at 42 °C for 30 min with samples being operation of the Tecnai T12 Electron microscope. Dr. Travis Gallagher and removed at intervals. MDH activity was measured as described in (36, 46), Dr. Jane Ladner for assistance with Multiangle light scattering. We also thank at 25 °C in an assay mixture containing 90 mM Hepes, pH 8.0, 0.22 mM NADH Dr. Alberto Marcario and Dr. Everly Conway de Marcario for careful reading (Sigma), and 0.55 mM oxaloacetate (Sigma). The time dependent oxidation of the manuscript. This work was funded by Grants NSF EAR 65132 and of NADH was measured by observing the rate of decrease of absorbance AFOSR FA9550-07-1-0022 from the National Science Foundation and the at 340 nm. Air Force Office of Scientific Research.

1. Laksanalamai P, Robb FT (2004) Small heat shock proteins from extremophiles: 27. Ballard SA, Go M, Segers RP, Adler B (1998) Molecular analysis of the dnaK locus of a review. Extremophiles 8:1–11. Leptospira interrogans serovar Copenhageni. Gene 216:21–29. 2. Laksanalamai P, Whitehead TA, Robb FT (2004) Minimal protein-folding systems in 28. Grandvalet C, Rapoport G, Mazodier P (1998) hrcA, encoding the repressor of the hyperthermophilic archaea. Nat Rev Microbiol 2:315–324. groEL genes in Streptomyces albus G, is associated with a second dnaJ gene. J Bacteriol 3. Large AT, Goldberg MD, Lund PA (2009) Chaperones and protein folding in the 180:5129–5134. archaea. Biochem Soc Trans 37:46–51. 29. Schulz A, Schumann W (1996) hrcA, the first gene of the Bacillus subtilis dnaK operon 4. Large AT, Lund PA (2009) Archaeal chaperonins. Front Biosci 14:1304–1324. encodes a negative regulator of class I heat shock genes. J Bacteriol 178:1088–1093. 5. Macario AJ, Conway De Macario E (2001) The molecular chaperone system and other 30. Wilson AC, Tan M (2002) Functional analysis of the heat shock regulator HrcA of anti-stress mechanisms in archaea. Front Biosci 6:D262–283. Chlamydia trachomatis. J Bacteriol 184:6566–6571. 6. Macario AJ, Malz M, Conway de Macario E (2004) Evolution of assisted protein folding: 31. Svetlitchnyi V, et al. (1991) Carboxydothermus hydrogenoformans gen. nov., sp. nov., a the distribution of the main chaperoning systems within the phylogenetic domain CO-utilizing thermophilic anaerobic bacterium from hydrothermal environments of – archaea. Front Biosci 9:1318–1332. Kunashir Island. Syst Appl Microbiol 14:254 260. 7. Hartl FU, Hayer-Hartl M (2002) Molecular chaperones in the cytosol: from nascent 32. Wu M, et al. (2005) Life in hot carbon monoxide: the complete genome sequence of chain to folded protein. Science 295:1852–1858. Carboxydothermus hydrogenoformans Z-2901. PLoS Genet 1:e65. 8. Saibil HR (2008) Chaperone machines in action. Curr Opin Struct Biol 18:35–42. 33. Laksanalamai P, Maeder DL, Robb FT (2001) Regulation and mechanism of action 9. Saibil HR, Ranson NA (2002) The chaperonin folding machine. Trends Biochem Sci of the small heat shock protein from the hyperthermophilic archaeon Pyrococcus – 27:627–632. furiosus. J Bacteriol 183:5198 5202. 10. Spiess C, Meyer AS, Reissmann S, Frydman J (2004) Mechanism of the eukaryotic 34. Laksanalamai P, Narayan S, Luo H, Robb FT (2009) Chaperone action of a versatile chaperonin: protein folding in the chamber of secrets. Trends Cell Biol 14:598–604. small heat shock protein from Methanococcoides burtonii, a cold adapted archaeon. Proteins 75:275–281. 11. Farr GW, et al. (2000) Multivalent binding of nonnative substrate proteins by the 35. Luo H, Laksanalamai P, Robb FT (2009) An exceptionallystable Group II chaperonin chaperonin GroEL. Cell 100:561–573. from the hyperthermophile Pyrococcus furiosus. Arch Biochem Biophys 486:12–18. 12. Horwich AL, Fenton WA, Chapman E, Farr GW (2007) Two families of chaperonin: 36. Figueiredo L, et al. (2004) Functional characterization of an archaeal GroEL/GroES cha- physiology and mechanism. Annu Rev Cell Dev Biol 23:115–145. peronin system: significance of substrate encapsulation. J Biol Chem 279:1090–1099. 13. Brinker A, et al. (2001) Dual function of protein confinement in chaperonin-assisted 37. Hongo K, et al. (2006) A novel ATP/ADP hydrolysis activity of hyperthermostable Group protein folding. Cell 107:223–233. II chaperonin in the presence of cobalt or manganese ion. FEBS Lett 580:34–40. 14. Lin Z, Rye HS (2006) GroEL-mediated protein folding: making the impossible, possible. 38. Quaite-Randall E, Trent JD, Josephs R, Joachimiak A (1995) Conformational cycle Crit Rev Biochem Mol Biol 41:211–239. of the archaeosome, a TCP1-like chaperonin from Sulfolobus shibatae. J Biol Chem 15. Jewett AI, Shea JE (2009) Reconciling theories of chaperonin accelerated folding with 270:28818–28823. experimental evidence. Cell Mol Life Sci 67:255–276. 39. Ludtke SJ, Baldwin PR, Chiu W (1999) EMAN: semiautomated software for high- 16. Agard DA (1993) To fold or not to fold. Science 260:1903–1904. resolution single-particle reconstructions. J Struct Biol 128:82–97. 17. Ellis RJ (2003) Protein folding: importance of the Anfinsen cage. Curr Biol 13:R881–883. 40. Williams TA, Codoner FM, Toft C, Fares MA (2010) Two chaperonin systems in bacterial 18. Horwich AL, Apetri AC, Fenton WA (2009) The GroEL/GroES cis cavity as a passive genomes with distinct ecological roles. Trends Genet 26:47–51. – anti-aggregation device. FEBS Lett 583:2654 2662. 41. Macario AJ, Brocchieri L, Shenoy AR, de Macario EC (2006) Evolution of a protein- 19. Betancourt MR, Thirumalai D (1999) Exploring the kinetic requirements for enhance- folding machine: genomic and evolutionary analyses reveal three lineages of the – ment of protein folding rates in the GroEL cavity. J Mol Biol 287:627 644. archaeal hsp70(dnaK) gene. J Mol Evol 63:74–86. 20. Sfatos CD, Gutin AM, Abkevich VI, Shakhnovich EI (1996) Simulations of chaperone- 42. Hirtreiter AM, et al. (2009) Differential substrate specificity of Group I and Group II – assisted folding. Biochemistry 35:334 339. chaperonins in the archaeon Methanosarcina mazei. Mol Microbiol 74:1152–1168. 21. Weissman JS, Kashi Y, Fenton WA, Horwich AL (1994) GroEL-mediated protein 43. Cuellar J, et al. (2008) The structure of CCT-Hsc70 NBD suggests a mechanism for Hsp70 folding proceeds by multiple rounds of binding and release of nonnative forms. Cell delivery of substrates to the chaperonin. Nat Struct Mol Biol 15:858–864. – 78:693 702. 44. Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm to estimate large 22. Booth CR, et al. (2008) Mechanism of lid closure in the eukaryotic chaperonin TRiC/CCT. phylogenies by maximum likelihood. Syst Biol 52:696–704. Nat Struct Mol Biol 15:746–753. 45. Studier FW (2005) Protein production by auto-induction in high density shaking 23. Ditzel L, et al. (1998) Crystal structure of the thermosome, the archaeal chaperonin cultures. Protein Expres Purif 41:207–234. and homolog of CCT. Cell 93:125–138. 46. Hayer-Hartl M () Chaperonin protocols: methods in molecular biology. (Humana Press, 24. Reissmann S, Parnot C, Booth CR, Chiu W, Frydman J (2007) Essential function of Totowa, NJ), Vol 140, pp 127–132. the built-in lid in the allosteric regulation of eukaryotic and archaeal chaperonins. 47. Schagger H (2001) Blue-native gels to isolate protein complexes from mitochondria. Nat Struct Mol Biol 14:432–440. Methods Cell Biol 65:231–244. 25. Meyer AS, et al. (2003) Closing the folding chamber of the eukaryotic chaperonin 48. Schagger H, von Jagow G (1991) Blue native electrophoresis for isolation of membrane requires the transition state of ATP hydrolysis. Cell 113:369–381. protein complexes in enzymatically active form. Anal Biochem 199:223–231. 26. Zhang J, et al. (2010) Mechanism of folding chamber closure in a Group II chaperonin. 49. Pettersen EF, et al. (2004) UCSF Chimera–a visualization system for exploratory Nature 463:379–383. research and analysis. J Comput Chem 25:1605–1612.

6of6 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1004783107 Techtmann and Robb Downloaded by guest on September 25, 2021