A modulator domain controlling thermal stability in the Group II chaperonins of Archaea

Archives of Biochemistry and Biophysics 512 (2011) 111–118

Contents lists available at ScienceDirect

Archives of Biochemistry and Biophysics journal homepage: www.elsevier.com/locate/yabbi

A modulator domain controlling thermal stability in the Group II chaperonins of Archaea Haibin Luo, Frank T. Robb ⇑ Institute of Marine and Environmental Technology, Department of Microbiology and Immunology, University of Maryland School of Medicine, 685 West Baltimore Street, Baltimore, MD 21201, USA

a r t i c l e

i n f o

Article history: Received 6 April 2011 and in revised form 22 April 2011 Available online 7 May 2011 Keywords: Hyperthermophile Psychrophile Archaea Thermosome Hsp60 Cpn ATPase activity Thermal stability Structural stability Ion pair

a b s t r a c t Archaeal Group II chaperonins (Cpns) are strongly conserved, considering that their growth temperatures range from 23 to 122 °C. The C-terminal 15–25 residues are hypervariable, and highly charged in thermophilic species. Our hypothesis is that the C-terminal is a key determinant of stabilization of the Cpn complex. The C-terminus of the Cpn from the hyperthermophile Pyrococcus furiosus was mutated to test this hypothesis. C-terminal deletions and replacement of charged residues resulted in destabilization. The stability of ATPase activity declined in proportion to the reduction in charged residues with Ala or Gly. An EK-rich motif (528EKEKEKEGEK537) proved to be a key domain for stabilization at or near 100 °C. Mutations ‘‘tuned’’ the Cpn for optimal protein folding at lower optimal temperatures, and Glu substitution was more potent than Lys replacement. Pf Cpn stability was enhanced by Ca2+, especially in the mutant Cpn lacking C-terminal Lys residues. This suggests that Glu-Glu interactions between C termini might be mediated by Ca2+. The C-terminal of a Cpn from the psychrophilic archaeon Methanococcoides burtonii was replaced by a domain from the hyperthermophile, resulting in increased thermostability and thermoactivity. We conclude that localized evolutionary variation in the C-terminus modulates the temperature range of archaeal Cpns. Ó 2011 Elsevier Inc. All rights reserved.

Introduction Molecular chaperones are essential multifunctional cellular systems that assist folding and assembly of newly synthesized proteins, translocation of unfolded proteins across membranes, as well as refolding and degradation of misfolded and aggregated proteins [1–3]. Chaperonins (Cpns)1 are ubiquitous, hollow complexes that promote correct folding of a wide range of unfolded, misfolded or partially folded proteins [4–6]. Cpns are divided into two groups: Group I Cpns, represented by GroEL in all bacteria, mitochondria and chloroplasts, and Group II Cpns, occurring in eukaryotes and Archaea [5,7,8]. Group I and Group II Cpns share similar quaternary structures consisting of a double toroid cylinder assembled into two rings of subunits stacked back to back. This arrangement generates two adjacent and identical cavities that induce unfolded proteins to fold correctly [6,9]. The two chambers open and shut turn and turn about, activated by ATP hydrolysis at each subunit [2,5]. Group I Cpns are composed of 14 identical subunits arranged in two heptameric rings and require a co-chaperone, GroES, to facilitate protein folding by shutting the chamber orifices as originally shown by ⇑ Corresponding author. Fax: +1 410 234 8896. E-mail address: [email protected] (F.T. Robb). Abbreviations used: Pf, Pyrococcus furiosus; Hsp, heat shock protein; Cpn, chaperonin; GDH, glutamate dehydrogenase; MDH, malate dehydrogenase. 1

0003-9861/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.abb.2011.04.017

Horwich and colleagues [2,10]. Group II Cpns form double rings with an 8- or 9-fold rotary symmetry and have built-in lids, allowing them to function without a co-chaperone [11]. The Archaea are similar to Eukarya in that they have Group II Cpns, however hyperthermophilic Archaea encode minimal Cpn complexes that simplify the study of these chaperones [12]. By contrast, the thermoacidophilic Archaea Sulfolobus spp. contain three different Group II subunits while Methanosarcina acetivorans encodes five Cpn subunits in both Group I and Group II [8,13]. In species with multiple Cpn homologs, one or two subunits may be up-regulated upon heat shock (e.g. Cpn b-subunit of Thermococcus strain KS-1 (T. KS-1) [14], and the Cpn a- and b-subunit of Sulfolobus shibatae [15]). Heat shock inducible subunits and complexes are typically more stable than non-heat shock regulated subunits from the same organism. The paralogy, heterologous complex formation and variable stability of these Cpns greatly complicates their study. Pf Cpn expression is significant under normal growth conditions and is highly induced upon 105 °C heat shock, probably contributing to the exceptional thermal tolerance of Pyrococcus spp. [16]. In contrast to Archaea that encode two, three or more subunits able to form heterologous Cpn complexes, the model archaeon, Pyrocococcus furiosus (Pf), a hyperthermophile growing optimally at 100 °C, encodes only one Cpn gene [13,17]. Pf Cpn complexes therefore have identical subunits and are minimally complex which renders Pf Cpn an ideal model Cpn for

112

H. Luo, F.T. Robb / Archives of Biochemistry and Biophysics 512 (2011) 111–118

protein structure analysis and mutational dissection [13,17]. We reported previously that the recombinant Pf Cpn was exceptionally stable in vitro [13], resisting 4.5 M guanidine hydrochloride and 105 °C, however the mechanism of its exceptional stability remains unclear. In all crystal structures of Group II Cpns, the C- and N-terminal regions together form an equatorial domain, but about 20 amino acid residues of the C-terminus remained unresolved in all available structures [18–21]. The C-terminal residues are localized in the equatorial domain close to the inter-ring interface and are thought to project into the inside of the ring and to be involved in intra/inter-ring contacts [20,22]. Current insights into the biological function of the C-terminus stem mainly from structural studies of bacterial Group I Cpn (GroEL) [18]. For example, C-terminally truncated Escherichia coli GroEL retains oligomeric structure and protein folding activity, but has reduced thermostability [21,23,24]. The GGM repeat in the GroEL C -terminus is closely involved in modulating the rate of ATP hydrolysis [24,25]. A hydrophilic segment ‘‘KNDAAD’’ in the GroEL C-terminus was found to be critical for substrate protein folding within the central cavity [26]. However, the C-termini of most Group II Cpns lack these characterized domains and are divergent from GroEL. There is limited information of the role of the C-terminus of Group II Cpns. The thermostability difference between a-Cpn and b-Cpn from T. KS-1 is thought to arise mainly from the differences in their C termini [21]. By comparing archaeal Group II Cpns from organisms with different growth temperatures we observed that they are conserved and collinear in the internal domains but have highly divergent C-terminal regions, which increase in the proportion of charged residues in proportion with growth temperature. Hyperthermophiles have clustered charged residues whereas the psychrophilic Archaea have few or no charged residues (Table 1). Because ionic interactions have been considered to contribute significantly to the thermostability of hyperthermophilic proteins [27–30], we formed the hypothesis that Cpn thermostability is determined largely by the variable charges of their C-termini. To examine this hypothesis, we carried out targeted mutation studies in order to produce Pf Cpn complexes with the Cterminal charge densities corresponding to Cpns from Archaea with lower growth temperatures. We also replaced the C-terminal domain of a Cpn from a psychrophilic archaeon with the C-terminus from the P. furiosus Cpn, producing a chimera with enhanced thermal stability and increased optimal temperature. Materials and methods Chemicals, enzymes and reagents Escherichia coli strains used in this study were DH5a for the initial cloning of the pET expression plasmid and BL21(DE3) for

recombinant protein expression. Restriction enzymes, polymerase and ligase were purchased from New England Biolabs (Beverley, MA). All other chemicals were of analytical grade from Sigma–Aldrich (St. Louis, MO). All solutions were made up in ultra-pure water. The native gel protein marker was NativeMark™ unstained molecular weight protein standard from Invitrogen (Carlsbad, CA). Glutamate, ATP, EDTA, DTT, b-NADH and NAD were purchased from Sigma–Aldrich (St. Louis, MO). Bovine glutamate dehydrogenase (GDH) and porcine heart malate dehydrogenase (MDH) were purchased from Amresco (Solon, OH). Cloning, expression and purification of Pf Cpn wild-type and mutants Pf Cpn was prepared as previously reported [13]. Genes for Pf Cpn mutants (CD1, CD2, EKD, 0E, 0K, 4E, 3E, 2E, 1E and MA) were amplified by PCR with above sense primer and the corresponding antisense primers listed in Table 2. Genes of Pf Cpn ND was amplified by the antisense primer of Pf Cpn and the following sense primer: 50 -AATTCCATGGGCATGCTCGTTGATAGCC-30 . The mutants were cloned, expressed and lyzed using the protocol for Pf Cpn [13] with modification. The supernatant extracts were heated at 80 °C (WT, CD1, CD2, EKD, 4E, 3E, 2E, 1E and ND) or 65 °C (0E, 0K and MA) for 30 min, then they were purified to homogeneity by two successive rounds of anion exchange: HiTrap™ Q HP cartridge from Biorad (Hercules, CA) and Bio-Scale™ macro-prep high Q cartridge from GE healthcare (Uppsala, Sweden). Cloning, expression and purification of Mb Cpn and MbPf Cpn Methanococcoides burtonii (Mb) DSM 6242 Cpn gene (GenBank Accession No. CP000300.1) was synthesized by Integrated Device Technology (Coralville, IA). It was amplified by PCR with following primers with digestion sites (underlined): sense, 50 -GGCCCATATGATGGCAGGACAGATGTC-30 ; antisense, 50 -GGCCGGATCCTTACATCATTGGAGGCATTC-30 . The gene for MbPf Cpn was produced by the sense primer and the following antisense primer: 50 -GGCCGGATCCTCAGTCTAGAT CACTGCTGAAGTCCTCGCTTCCTCCTCCCTTCTCACCTTCTTTCTCCTTCT CTTTCTCGAGCTTGCTGGCCCTGAGGATCATTACTG-30 . Mb Cpn and MbPf Cpn were cloned using the procedure for Pf Cpn [13] with modification. During their induced expressions, IPTG was added into the culture at OD600 of 0.7 to a final concentration of 1 mM. Induction was allowed to proceed at 20 °C for 12 h. Harvested cells were suspended in extraction buffer (25 mM Hepes–KOH, pH 7.2, 100 mM KCl) lyzed by French Press and centrifuged at 25,000g for 30 min. The supernatants were purified by Protino Ni-TED 1000 Kit (Macherey–Nagel, Düren).

Table 1 Comparison of the C-termini of Group II Cpns.

a

Species

C-termini of Group II Cpns

Opt./Max. Temp. (°C)

Negatively charged residues

Pyrodictium occultum Pyrobaculum aerophilum str.IM2ba Pyrobaculum islandicum DSM 4184 Methanopyrus kandleri Pyrococcus furiosus Pyrobaculum arsenaticum DSM 13514 Pyrococcus horikoshii OT3 Archaeoglobus fulgidusba Desulfurococcus Thermococcus kodakarensis Kod1ba Sulfolobus shibatae Thermoplasma acidophilum DSM 1728 Methanococcus burtonii DSM 6242 2146a

EEKEEKEKEKKEEGEE KREEKGKKKEGEEGEEKKEETKFD EKEKEEKKGEEEKKEEKKEFD SKEEEEEEEEGGSSEF EKEKEKEGEKGGGSEDFSSDLD EKEEKGEKKEEKKEEFD EKEKEGEKGGGSEEFSGSSDLD EKEKGPEGESGGEEDSEE EKDKEDKGGSNDFGSDLD EKDKEGGKGGSEDFGSDLD GGSEPGGKKEKEEKSSED SSSSSSNPPKSGSSSESSED APPMPDGGMGGMPPMM

105/110 100/104 100/103 98/110 100/103 95/103 95/100 83/95 85/95 65/95 75/80 60/65 18/23

10 10 11 9 9 9 8 9 7 7 6 3 1

Indicated Cpn subunit.

H. Luo, F.T. Robb / Archives of Biochemistry and Biophysics 512 (2011) 111–118

113

Table 2 Antisense primers for mutants of Pf Cpn. Mutants

Anti-sense primers

CD1

50 -AATTGGATCCTCAGGCAGCGATGACATCG-30

CD2

50 -GGCCGGATCCTCACTTCTCACCTTCTTTCT-30

EKD

50 -GGCCGGATCCTCAGTCTAGATCACTGCTGAAGTCCTCGCTTCCTCCTCCCTTCGCACCTGCTTTCGCCTTCGCTTTCGCGAG-30

0E

50 -GGCCGGATCCTCAGTCTAGATCACTGCTGAAGTCCTCGCTTCCTCCTCCCTTCGCACCTGCTTTCGCCTTCGCTTTCGCGAG-30

0K

50 -GGCCGGATCCTCAGTCTAGATCACTGCTGAAGTCCTCGCTTCCTCCTCCCGCCTCCGCTTCTGCCTCCGCCTCTGCCTCGAG-30

4E

50 -CCGGGGATCCTCAGTCTAGATCACTGCTGAAGTCCTCGCTTCCTCCTCCCTTCTCACCTTCTTTCTCCTTCTCTTTGCCGAGC-30

3E

50 -GGCCGGATCCTCAGTCTAGATCACTGCTGAAGTCCTCGCTTCCTCCTCCCTTCTCACCTTCTTTCTCCTTGCCTTTGCCGAGC-30

2E

50 -GGCCGGATCCTCAGTCTAGATCACTGCTGAAGTCCTCGCTTCCTCCTCCCTTCTCACCTTCTTTGCCCTTGCCTTTGCCGAGC-30

1E

50 -GGCCGGATCCTCAGTCTAGATCACTGCTGAAGTCCTCGCTTCCTCCTCCCTTCTCACCTCCTTTGCCCTTGCCTTTGCCGAGC-30

MA

50 -AATTGGATCCTCAGCCTAGGCCACTGCTGAAGCCGCCGCTTCCTCCTCCCTTGCCACCTCCTTTGCCCTTGCCTTTGCCGAGC-30

ATP hydrolysis activity and half life determination The standard reaction mixture contained 25 mM Hepes–KOH, pH 7.2, 300 mM KCl, 1 mM MgCl2, 0.1 mg/ml Cpn and 100 lM ATP in a final volume of 90 ll. The reaction mixture was incubated at series of temperatures ranging from 20 to 105 °C for 3 min before adding ATP to initiate ATP hydrolysis exactly as reported previously [13]. The reaction was performed for 15 min at different temperature in thermocycler, followed by 2% perchloric acid addition to quench the reaction. The liberated Pi was determined by the Malachite Green assay at 630 nm [22,31,32]. Reactions at temperatures higher than 105 °C were performed in silicon oil bath in heating block. Light mineral oil was added to the top of reaction solution for avoiding liquid evaporation. To determine half life at indicated temperature, protein samples were prepared at 1 mg/ml in 25 mM Hepes–KOH, pH 8.0, 300 mM KCl. Samples were incubated at 100 or 70 °C for indicated time before taking out for ATPase activity. Deamidation of Asn and Gln of mutants may happen at high temperatures, thus we used the same procedure to handle the wild-type and mutants. Considering the mutants are similar in most of their domains, deamidation should happen at same sites with similar effects, which should validate the comparison of these mutants. The ATPase activities of the samples were determined at their optimal temperature. Light mineral oil was used to minimize evaporation. In plotting activity versus time, errors were calculated from assays repeatedly three times and the graphs were enacted without special regression analysis. CD spectra and thermal denaturation scan CD spectra were recorded in a 0.1 cm cuvette with a Jasco-810 CD spectrometer (Jasco, Easton, MD) at a protein concentration of 0.20 mg/ml and scanning rate of 20 nm/min; the background spectra were subtracted. Thermal denaturation was performed from 20 to 95 °C using proteins (0.20 mg/ml) prepared freshly in 5 mM Hepes buffer, pH 8.0, in temperature controlled water-circulated quartz cell (1-mm path length) by monitoring the ellipticity at 222 nm with a scan rate of 1.0 °C/min. Bovine glutamate dehydrogenase (GDH) protection assays Chaperone activities of Pf Cpn WT and mutants were compared by their protection of bovine GDH at 50 °C. GDH samples (2.5 mg/ ml) were incubated in buffer A (25 mM Hepes–KOH, pH 8.0, 300 mM KCl and 1 mM MgCl2) at 50 °C in the presence or absence of Cpn (0.2 mg/ml) and 4 mM ATP. Thermostability was assayed by

measuring the enzyme activity after different incubation time intervals (0, 10, 20 and 30 min). GDH activity was determined by oxidative deamination measuring the glutamate-dependent reduction of NADP at 340 nm as previously described [33]. The reaction mixture contained 90 mM EPPS, pH 8.0, 2 mM L-glutamate, 0.5 mM NADP and enzyme in a total volume of 325 ll. Malate dehydrogenase (MDH) refolding assays Porcine heart MDH was denatured for 20 min in buffer A (25 mM Hepes–KOH, pH 8.0, 300 mM KCl, 1 mM MgCl2) containing 2 M guanidine HCl, 5 mM dithiothreitol (DTT) at 37 °C and diluted 100-fold into buffer A containing 0.5 M ammonium sulfate (AS) and 5 mM ATP, in the absence or presence of 70 nM (oligomer) Pf Cpn mutants (as specified in the figure legend). The concentration of denatured MDH was 70 nM. This was 1:1 ratio of Cpn to denatured MDH. The refolding was carried out at 37 °C for 30 min with aliquots being removed at the time indicated. MDH enzyme activities were measured immediately as described [34] with modification at 25 °C in assay mixture (90 mM Hepes–KOH, pH 8.0, 0.22 mM b-NADH (Sigma–Aldrich, St. Louis, MO), 0.55 mM oxaloacetate (Sigma–Aldrich). The time dependent oxidation of b-NADH was measured by observation of the decrease of absorbance at 340 nm. Proteinase K digestion assay The proteinase K digestion assays were carried out according to the reported procedure [35] with the following modification. Briefly, purified Pf Cpn WT, 0E and 0K were in 25 mM Hepes– KOH, pH 8.0, 300 mM KCl. After adding proteinase K (0.13– 0.5 lg/ml) and further incubation for 10 min at 25 °C, PMSF was added at a final concentration of 5 mM to inhibit protease activity. The reaction was then incubated on ice for 10 min with loading buffer and analyzed by SDS–PAGE. Results Mutagenesis of the Pf Cpn C-terminus The C-terminal domain of the Pf Cpn was subject to systematic mutagenesis. In the C-terminal 22 amino acids (528EKEKEKEGEKGGGSEDFSSDLD549), Pf Cpn contains six glutamic acids, three aspartic acids and four lysines. The region 528EKEKEKEGEK537 which we will refer to as the EK-rich motif contains five of the Glu residues and all but one of the Lys residues. The other four charged residues are located in an adjacent domain

114

H. Luo, F.T. Robb / Archives of Biochemistry and Biophysics 512 (2011) 111–118

‘‘542EDFSSDLD549’’. These two domains are connected by a linker ‘‘538GGGS541’’ with predicted high flexibility. To test the hypothesis that the charged residues in the C-terminus form determinants of thermal stability, mutations that progressively reduced the charge density of the C-terminus of PfCpn were constructed and the complexes were expressed and purified. The C-terminal mutant sequences are summarized in Table 3. The optimal temperatures for ATPase activity and half-life at high temperature for the wildtype and these mutants were determined as described previously [13]. To separate the contributions of the C-terminal sub-domains, deletion mutants were constructed, expressed and analyzed, as shown in Table 3. Initially, a C-terminal deletion mutant (CD1) lacking 22 amino acids was produced and compared to the wild-type (WT). As indicated by Table 3, CD1 has an optimal ATPase activity at 79 °C, 13 °C lower than that of WT (92 °C). The half life (t1/2) at 100 °C of CD1 was 28.5 min, reduced 2.5-fold compared to that of WT (71.2 min). As shown in Fig. 1A, CD1 retains double-ring structure comparable to WT, however its oligomeric stability decreased significantly. CD1 and WT were comparable in protecting GDH against heat-inactivation and in refolding chemical-denatured MDH (Fig. 1B). The C-terminus is apparently not essential for chaperone functions or complex assembly but does determine stability at temperatures above 90 °C. Two less extensive C-terminal deletion mutants, EKD and CD2, both showed reduced optimal temperatures and thermal stabilities (Table 3). The deletion of the EK-rich motif (‘‘EKD’’) greatly destabilized the protein, having a similar half-life (t1/2 = 30.5 min) to that of the longest C-terminal deletion (t1/2 = 28.5 min). We interpret this to suggest that the stabilization effect of the C-terminus as a whole stems mainly from the EK-rich motif. To determine whether Glu or Lys residues in the EK-rich motif (528EKEKEKEGEK537) contribute equally to protein stability, mutants were constructed in which Glu and Lys residues were replaced by Ala. The mutant ‘‘0E’’ (528AKAKAKAGAK537) was depleted in Glu and mutant ‘‘0K’’ (528EAEAEAEGEA537) was depleted in Lys. The 0E construct had a remarkably low optimal temperature of 58.3 °C and an extremely short half life (t1/2 < 1 min) at 100 °C. Mutant 0K also had reduced optimal temperature for ATPase activity (73.2 °C), but retained significant thermal stability at 100 °C (t1/2 = 53.6 min), indicating that Glu residues in the EK-rich motif contribute more to protein stability than Lys residues. Curiously, the mutants 0E and 0K had even lower optimal temperatures for ATPase activity compared to CD1. Mutation of all Glu residues impaired protein stability more than replacement of all Lys residues. To examine the contribution of each Glu residue in the EK rich motif, a progressive mutation experiment was carried out, in which mutants with 4 Glu residues (4E), 3 Glu residues

Fig. 1. Comparison of Pf Cpn WT and CD1. (A) Thermal dissociation of oligomers of Pf Cpn WT and CD1. Samples (1 mg/ml in 25 mM Hepes–KOH, pH 8.0, 300 mM KCl) were incubated at the indicated temperature for 30 min and took out to centrifuge to remove precipitates. The supernatants were analyzed by 3–6% PAGE and 10% SDS PAGE. Each lane contains 20 lg of protein. (B) Protection of GDH (2.5 mg/ml) against inactivation and refolding of MDH by Pf Cpn WT and CD1.

(3E), 2 Glu residues (2E) and 1 Glu residue (1E), were produced and analyzed. As described in Table 3, stability decreased progressively with the ordered replacement of Glu residues: WT > 4E > 3E > 2E > 1E > 0E. There are four additional negatively charged residues (E542, D543, D547 and D549) in the adjacent domain (542EDFSSDLD549). Accordingly all negatively charged residues in the C-terminus were mutated to more flexible Gly residues to produce the MA (‘‘mutate all’’) variant (Table 3). MA had a very low optimal temperature (40.1 °C) and an extremely short half life at 100 °C (t1/2 < 1 min). MA is also less stable than 0E, having a shorter half life at 70 °C (Table 3). This corroborates the observation implicit in the results shown in Table 3

Table 3 Summary of Pf Cpn mutants. Mutants

Description

ATPase (mmol/min/mg)

Opt. Temp. (°C)

t1/2 at 100 °C (min)

WT CD1 CD2 EKD 0E

—SKLEKEKEKEGEKGGGSEDFSSDLD —SKL———————————————————— —SKLEKEKEKEGEK——————————— —SKL————————GGGSEDFSSDLD

0.06311 0.04913 0.05434 0.05114 0.07036

92.4 78.8 84.3 83.5 58.3

71.2 28.5 50.1 30.5 <1 (6.1a)

0K

—SKLEAEAEAEGEAGGGSEDFSSDLD

0.06350

73.2

53.6

4E

—SKLGKEKEKEGEKGGGSEDFSSDLD

0.06256

86.2

49.6

—SKLAKAKAKAGAKGGGSEDFSSDLD

3E

—SKLGKGKEKEGEKGGGSEDFSSDLD

0.05960

73.2

45.9

2E

—SKLGKGKGKEGEKGGGSEDFSSDLD

0.07278

74.2

43.5

1E

—SKLGKGKGKGGEKGGGSEDFSSDLD

0.05161

71.4

36.8

MA

—SKLGKGKGKGGGKGGGSGGFSSGLG N-terminal 50 AA deletion

0.07560

40.1

<1 (4.9a)

0.01310

91.6

>100

ND

Mutated amino acids are underlined. a Half life was obtained at 70 °C.

H. Luo, F.T. Robb / Archives of Biochemistry and Biophysics 512 (2011) 111–118

115

from C-terminal deletions, suggesting that C-terminal charge is a key stabilizing factor. Comparison of WT, 1E, 0E and MA on oligomer stability, structural stability and chaperone activity at the low temperature range To study the effects of replacement of the negatively charged residues in the C-terminus on oligomeric stability, structural stability and chaperone activity at low temperature, three mutants, namely, 1E, 0E and MA, were chosen for comparison. As shown in Fig. 2A, an oligomeric stability order as WT > 1E > 0E > MA was obtained, indicating that the number of negatively charged residues is also proportional to oligomeric stability. It is noted that 0E and MA formed unstable one-ring structures, which suggested that the replacement of anionic residues by flexible residues weakens inter-ring interactions, aborting double ring formation to a greater extent than C-terminal deletions. These proteins were also compared by circular dichroism (CD) scans as shown in Fig 2. The inset of Fig. 2B revealed that they had identical CD spectra at 20 °C, suggesting that their overall secondary structures were very similar. Thermal unfolding scans at 222 nm (from 20 to 95 °C) produced no clear transitions for WT and 1E, whereas the mutants with decreased C-terminal Glu residues showed transitions at 70–90 °C. The relative stabilities of the a-helical components of their secondary structures at high temperature decreased in the order WT > 1E > 0E > MA. Chaperone activities of WT, 1E and MA at low temperature were compared via protection of bovine GDH against 50 °C heat inactivation and refolding capability on chemically denatured porcine MDH at 37 °C. In GDH protection assays (Fig. 3A), MA provided enhanced protection to GDH at 50 °C compared to the more stable variants. Mutant 1E provided more protection than WT but less than MA. In MDH refolding assays (Fig. 3B), MA also showed significantly improved refolding efficiency, restoring 35.3% of the initial activity of MDH at 30 min. WT was essentially inactive in refolding activity at 37 °C and mutant 1E was intermediate. Taken together, these data suggest that the replacement of the negatively charged residues in C-terminus by flexible residues can modulate the oligomer mobility, enhancing chaperone efficiency at low temperature. The C-terminus from Pf Cpn confers thermostability and raises the Topt for ATPase activity on a psychrophilic Cpn We performed a reciprocal switch of C-terminal domains, to evaluate whether the charged residues could upgrade the stability of a cold-adapted Cpn. The M. burtonii (Mb) is an Antarctic methanoarchaeon with an optimal grow temperature of 23 °C [33]. The MbCpn was cloned and expressed and a chimera Cpn (MbPf Cpn) was constructed with the C-terminus of Pf Cpn. MbPf Cpn showed higher optimal temperature for ATPase activity (Fig. 4) than Mb Cpn, shifting the range of the activity vs temperature profile significantly upwards and eliminating cold-adapted activity. Unaltered Mb Cpn was highly active at 25–40 °C and was inactivated rapidly at temperatures higher than 60 °C, whereas the chimera MbPf Cpn was inactive below 40 °C and retained the majority of ATPase activity at 70 °C. These data confirmed that the highly charged C-terminus increased the thermal tolerance and cooperative ATPase activity of Mb Cpn significantly. A possible mechanism for C-terminal stabilization of Pf Cpn The results of C-terminal mutagenesis beg the question as to how the clustered charged residues in C-terminus stabilize the 16-mer complex. Although the C-terminus has not been resolved in current crystal structures of Group II Cpns, it is thought to project from the equatorial domain into the inside of the cavity

Fig. 2. Comparison of oligomeric and structural stabilities of WT, 1E, 0E and MA (A) Thermal dissociation of oligomers of Pf Cpn WT, 1E, 0E and MA. Samples were incubated at the indicated temperature for 30 min and were centrifuge to remove precipitates. The supernatants were analyzed by 3–6% PAGE and 10% SDS PAGE. Each lane contains 20 lg of protein. (B) Thermal denaturation curves for WT, 1E, 0E and MA in Hepes–KOH, pH 8.0. The inlet showed the identical CD spectrum of WT, 1E, 0E and MA. All proteins were prepared in 5 mM Hepes–KOH, pH 8.0 at 0.2 mg/ ml.

[18–21,36]. Pf Cpn shares over 80% sequence similarity with T. KS-1 a-Cpn, thus it is likely that they share similar structures. In the crystal structure of T. KS-1 a-Cpn, the N-terminal residues 20 through 40 are located roughly alongside the C-terminus and constitute a long a-helix (a1) with four positively charged residues and three negatively charged residues (data not shown), raising the possibility that the C- and N-termini form a coordinated ion pair network. Accordingly, a1-helix of Pf Cpn was deleted and expressed as a mutant named ND (Table 3) with the anticipation that the deletion would result in drastic destabilization. However, the ND mutation did not alter the optimal temperature of Pf Cpn. It still hydrolyzed ATP optimally at 91.6 °C, similar to WT (92.4 °C). The Cterminal stabilization effect is not therefore mediated in combination with the N-terminal 1-helix. Another possibility is that E–K ion pairs form between the EKrich motifs from different subunits during complex formation, since Glu and Lys are oppositely charged residues at neutral pH. Since we had found that the C-terminus is not an essential domain

116

H. Luo, F.T. Robb / Archives of Biochemistry and Biophysics 512 (2011) 111–118

Fig. 3. Comparison of chaperone activities of WT, 1E and MA. (A) Protection of GDH (2.5 mg/ml) against inactivation by Pf Cpn wild-type and mutants (0.2 mg/ml). (B) Refolding of MDH by Pf Cpn and mutants. Chemically denatured MDH (7 lM) was diluted 100-fold into buffer A in the presence of Cpn (70 nM) at 37 °C. Refolding reaction was initiated by the addition of 5 mM ATP. The MDH activity is shown as a percentage of the activity obtained with 70 nM MDH.

for double ring formation, the E–K ion pairs possibly crosslink the adjacent subunits to enhance the existing inter-subunit associations and this maintain the structure during the extensive conformational changes accompanying folding activity. In this case, mutation of Glu or Lys to flexible Ala residues should render the entire Cpn complex more flexible and active at lower temperatures. To test this, proteinase K digestion assay was carried out on mutants 0E, 0K and WT. As shown in Fig. 5A, mutant 0E and 0K were more accessible to proteinase K digestion than WT, confirmed that Glu-Ala and Lys-Ala mutations decreased complex tightness. Accordingly, Glu- and Lys-mutation in the C-terminus should render the double-ring complex less stable. As we found above, extensive Glu to Ala mutations caused defects in double ring assembly, resulting in single rings as shown in Fig. 2A. The introduction of flexible Ala residues increases the flexibility of the C-terminus, possibly counteracting the inter-ring interactions from other domains. Single rings were also detected on the mutant 0K, however double rings were retained at the same time. The side chain of Lys is longer than that of Glu, which might cause Lys (in 0E) to generate stronger repulsive forces than Glu (in 0K). However, there are five Glu residues in the 0K (four Lys residues in 0E), which could produce substantial repulsive forces as well. The ability of the mutant 0K to form double ring complexes is anonymously in light of the protein stability literature, proteins containing clusters of negatively charged residues show enhanced thermal stability or oligomeric stability through coordination with divalent cations, especially Ca2+ [37–43]. Also, several hyperthermophiles with growth optima exceeding P. furiosus have Cpns have homo-Glu tracts and no Lys residues in their C-terminis (Table 1). To explore whether Cal2+ could mediate E to E interactions, we incubated the 0K Cpn mutant with Ca2+ before analysis by gradient native gel. As indicated in Fig. 5B, Ca2+ converted single rings of 0K into double rings, whereas Ca2+ did not alter 0E and WT. Furthermore, after 0K complexes were pre-incubated with Ca2+, this mutant increased in optimal temperature (95.4 °C) and exhibited enhanced thermal stability (t1/2 = 67.6 min at 100 °C). As a control to test whether Ca2+ enhancement was mediated by the C-terminus and not other domains, the ATPase activities of WT and CD1 (C-terminal-truncated) mutant were expressed and purified with and without Ca2+. As shown in Fig. 6, the optimal temperature of ATPase activity of Pf Cpn WT was increased from 92.4 to 102 °C by the presence of Ca2+, whereas the optimal temperature of Pf Cpn CD1 improved marginally with Ca2+.

Discussion

Fig. 4. Temperature profile of ATPase activities of Mb Cpn and MbPf Cpn. The standard reaction mixture contained 25 mM Hepes–KOH, pH 7.2, 300 mM KCl, 10 mM MgCl2, 0.5 mg/ml Cpn and 100 lM ATP in a final volume of 90 ll. The reaction mixture was incubated at indicated temperature for 3 min before adding ATP to initiate ATP hydrolysis.

We compared Group II Cpns from 13 Archaea, including psychrophiles, mesophiles, thermophiles and hyperthermophiles, confirming that Group II Cpns are conserved, except for their hypervariable carboxyl termini, and somewhat variable N-termini. We hypothesized that Group II Cpns from different Archaea are adapted to their optimal growth temperatures by modulation of the composition of their C-termini (Supplementary Fig. 1). To address our hypothesis and clarify the stabilization mechanism of the C-terminus, we constructed a set of C-terminal mutants of Pf Cpn and examined their stabilities and activities. The removal of the C-terminus did not alter double ring formation and protein folding functions, and caused moderate destabilization, indicating that the C-terminus is a protein stabilizing domain. The EK-rich motif (528EKEKEKEGEK537) in the C-terminus was found to be a key domain for stabilization and setting the optimal temperature for activity. At first, it seemed that inter-subunit ion pairs (E–K) forming among complementary EK-rich motifs might be stabilizing Pf Cpn, because E–K ion pairs are widely reported to be a stabilizing factor in many proteins [27,44–48]

H. Luo, F.T. Robb / Archives of Biochemistry and Biophysics 512 (2011) 111–118

Fig. 5. (A) Proteinase K digestion of Pf Cpn WT, 0K and 0E. The indicated amount of proteinase K was incubated with proteins (1 mg/ml) at 25 °C for 10 min. PMSF was added to stop the digestion reaction before samples were analyzed by SDS–PAGE. Each lane contains 15 lg proteins. (B) Oligomeric status analysis of Pf Cpn WT, 0E and 0K by native gradient PAGE. Each lane contains 10 lg protein samples. Treated 0K complexes were incubated with 1 mM CaCl2 for 30 min before SDS–PAGE analysis.

Fig. 6. Temperature profiles of ATPase activity for Pf Cpn WT and CD with and without CaCl2 incubation.

Hyperthermophilic E–K ion pairs have been reported in subunit interfaces, for example in the dimer interface of hyperthermophilic citrate synthase [27,45], in human a1-antitrypsin [47] and in hyperthermophilic glutamate dehydrogenase [30]. Our further results suggested that Glu residues could form ion pairs with Lys residues or with Glu residues through Ca2+, the latter actually contributes stronger stabilization. Lys residues are not essential for high temperature stability in the presence of Ca2+. C-terminal negatively charged residues were also determinants to oligomeric

117

stability and structural tightness in the presence of Ca2+. Divalent cations were found to enhance protein oligomerization and stability through binding to negatively charged residues in many disparate proteins [37–43]. For instance, E130 in human soluble calcium activated nucleotidase (CAN) from paired, adjacent subunits interact with Ca2+ to provoke dimer formation [42]; divalent cations binding to the C-terminal five anionic residues in EF-hand binding loop significantly enhanced the thermal stability of polcalcin [37,39,40,42]. Therefore, we propose that the negative residues in the C-terminus of Pf Cpn crosslink subunits from adjacent rings through Ca2+. The number of Glu residues in the EK-rich motif is proportional to protein thermal stability, both in our experiments and in other archaeal species (Table 1). Interestingly; C-termini of Group II Cpn from high growth-temperature organisms have more negatively charged residues than those from psychrophilic and moderately thermophilic Archaea (Table 1). For example, Cpn from Pyrodictium occultum which grows optimally at 105 °C has 10 negatively charged residues in its C-terminus; b-subunit of Cpn of Thermoplasma acidophilum (Topt = 60 °C) has 3; Cpn of Methanoccoccus burtonii (Topt = 23 °C) has only 1. It is interesting that positively charged residues are usually less abundant than negatively charged residues in the C-terminal of hyperthermophilic Group II Cpns, especially in those species growing above 100 °C (Table 1), indicating that ion pairs forming between oppositely charged residues do not occur on every Group II Cpn. Moreover, the Cpn from Methanopyrus kandleri (with an extremely high growth temperature of 110–122 °C) has a run of eight Glu residues in its C-terminus, and no Lys residues. This supports our conclusion that negatively charged residues in the C-terminus are of great importance to thermal stability of Group II Cpns and positively charged residues are relatively less important. The Yohda group recently published a study on a single mutation in the built-in lid, K323R, of the Cpn from T. KS-1 which resulted in higher activity at relatively low temperatures [49]. This mutation induced higher mobility of the built-in lid of the Cpn to activate protein folding at lower temperature, but the authors also stated that the cold adaptation of K323R is not sufficient for practical applications. We believe that our finding on the C-terminus provides a potential way for the rational design of cold-adapted Cpns with full activity at low temperatures. Finally, our work suggests that thermal stabilization results from adaptive mutation in a small region of the protein, distinct from the catalytic domains and intersubunit interfaces. Considering that this chaperone must form functional megadalton complexes consisting, of nonidentical subunits in many species, the evolutionary constraints on sequence divergence must be similar to those in ribosomal proteins. Indeed, the Cpn subunits show exceptional conservation of sequence in both the Archaea and Eukarya. The mechanism we describe in this paper allows the temperature range of the Cpn complexes to be adjusted across a wide temperature range by minimal evolutionary divergence of a compact electrostatic ‘‘switch’’ domain in the C-terminus. Since the C-terminus is missing from the current solved Group II Cpn structures, crystallization of full-length Pf Cpn aiming to obtain structural information of the C-terminus is under way in our lab. We note that if may be possible to study the interactions of peptides representing the C-terminal regions in vitro, or by molecular dynamic simulation. Acknowledgments The authors thank Dr. Ilia V. Baskakov and Valeriy G. Ostapchenko for their technical assistance on CD experiments. This research was funded by Air Force Grants (AFOSR 03-S-28900 and 9550-10-1-0272) and NSF Grant BES 06-24224.

118

H. Luo, F.T. Robb / Archives of Biochemistry and Biophysics 512 (2011) 111–118

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.abb.2011.04.017.

References [1] J.C. Young, V.R. Agashe, K. Siegers, F.U. Hartl, Nat. Rev. Mol. Cell Biol. 5 (2004) 781–791. [2] B. Bukau, A.L. Horwich, Cell 92 (1998) 351–366. [3] M. Okochi, H. Matsuzaki, T. Nomura, N. Ishii, M. Yohda, Extremophiles 9 (2005) 127–134. [4] N.A. Ranson, H.E. White, H.R. Saibil, Biochem. J. 333 (1998) 233–242. [5] A.J. Macario, M. Lange, B.K. Ahring, M.E. Conway de, Microbiol. Mol. Biol. Rev. 63 (1999) 923–967. [6] H. Saibil, Curr. Opin. Struct. Biol. 10 (2000) 251–258. [7] I. Gutsche, L.O. Essen, W. Baumeister, J. Mol. Biol. 293 (1999) 295–312. [8] H. Kubota, G. Hynes, K. Willison, Eur. J. Biochem. 230 (1995) 3–16. [9] N.R. Douglas, S. Reissmann, J. Zhang, B. Chen, J. Jakana, R. Kumar, W. Chiu, J. Frydman, Cell 144 (2011) 240–252. [10] F.U. Hartlm, M. Hayer-Hartl, Nat. Struct. Mol. Biol. 16 (2009) 574–581. [11] R. Iizuka, T. Yoshida, Y. Shomura, K. Miki, T. Maruyama, M. Odaka, M. Yohda, J. Biol. Chem. 278 (2003) 44959–44965. [12] P. Laksanalamai, T.A. Whitehead, F.T. Robb, Nat. Rev. Microbiol. 2 (2004) 315– 324. [13] H. Luo, P. Laksanalamai, F.T. Robb, Arch. Biochem. Biophys. 486 (2009) 12–18. [14] T. Yoshida, A. Ideno, S. Hiyamuta, M. Yohda, T. Maruyama, Mol. Microbiol. 39 (2001) 1406–1413. [15] H.K. Kagawa, T. Yaoi, L. Brocchieri, R.A. McMillan, T. Alton, J.D. Trent, Mol. Microbiol. 48 (2003) 143–156. [16] K.R. Shockley, D.E. Ward, S.R. Chhabra, S.B. Conners, C.I. Montero, R.M. Kelly, Appl. Environ. Microbiol. 69 (2003) 2365–2371. [17] F.T. Robb, D.L. Maeder, J.R. Brown, J. DiRuggiero, M.D. Stump, R.K. Yeh, R.B. Weiss, D.M. Dunn, Methods Enzymol. 330 (2001) 134–157. [18] K. Braig, Z. Otwinowski, R. Hegde, D.C. Boisvert, A. Joachimiak, A.L. Horwich, P.B. Sigler, Nature 371 (1994) 578–586. [19] L. Ditzel, J. Lowe, D. Stock, K.O. Stetter, H. Huber, R. Huber, S. Steinbacher, Cell 93 (1998) 125–138. [20] Y. Shomura, T. Yoshida, R. Iizuka, T. Maruyama, M. Yohda, K. Miki, J. Mol. Biol. 335 (2004) 1265–1278. [21] T. Yoshida, T. Kanzaki, R. Iizuka, T. Komada, T. Zako, R. Suzuki, T. Maruyama, M. Yohda, Extremophiles 10 (2006) 451–459.

[22] T. Yoshida, M. Yohda, T. Iida, T. Maruyama, H. Taguchi, K. Yazaki, T. Ohta, M. Odaka, I. Endo, Y. Kagawa, J. Mol. Biol. 273 (1997) 635–645. [23] T. Langer, G. Pfeifer, J. Martin, W. Baumeister, F.U. Hartl, EMBO J. 11 (1992) 4757–4765. [24] N.F. McLennan, A.S. Girshovich, N.M. Lissin, Y. Charters, M. Masters, Mol. Microbiol. 7 (1993) 49–58. [25] G.W. Farr, W.A. Fenton, A.L. Horwich, Proc. Natl. Acad. Sci. USA 104 (2007) 5342–5347. [26] K. Machida, A. Kono-Okada, K. Hongo, T. Mizobata, Y. Kawata, J. Biol. Chem. 283 (2008) 6886–6896. [27] S. Kumar, B. Ma, C.J. Tsai, R. Nussinov, Proteins 38 (2000) 368–383. [28] F.T. Robb, D.S. Clark, J. Mol. Microbiol. Biotechnol. 1 (1999) 101–105. [29] G. Saelensminde, O. Halskau Jr., I. Jonassen, Extremophiles 13 (2009) 11–20. [30] C. Vetriani, D.L. Maeder, N. Tolliday, K.S. Yip, T.J. Stillman, K.L. Britton, D.W. Rice, H.H. Klump, F.T. Robb, Proc. Natl. Acad. Sci. USA 95 (1998) 12300–12305. [31] A.A. Baykov, V.N. Kasho, S.M. Avaeva, Anal. Biochem. 171 (1988) 271–276. [32] T.P. Geladopoulos, T.G. Sotiroudis, A.E. Evangelopoulos, Anal. Biochem. 192 (1991) 112–116. [33] P. Laksanalamai, S. Narayan, H. Luo, F.T. Robb, Proteins 75 (2009) 275–281. [34] L. Figueiredo, D. Klunker, D. Ang, D.J. Naylor, M.J. Kerner, C. Georgopoulos, F.U. Hartl, M. Hayer-Hartl, J. Biol. Chem. 279 (2004) 1090–1099. [35] S. Reissmann, C. Parnot, C.R. Booth, W. Chiu, J. Frydman, Nat. Struct. Mol. Biol. 14 (2007) 432–440. [36] J. Zhang, M.L. Baker, G.F. Schroder, N.R. Douglas, S. Reissmann, J. Jakana, M. Dougherty, C.J. Fu, M. Levitt, S.J. Ludtke, J. Frydman, W. Chiu, Nature 463 (2010) 379–383. [37] S.I. Belli, A. Sali, Biochem. J. 304 (Pt. 1) (1994) 75–80. [38] G. Bujacz, J. Alexandratos, A. Wlodawer, G. Merkel, M. Andrake, R.A. Katz, A.M. Skalka, J. Biol. Chem. 272 (1997) 18161–18168. [39] P.J. Gotwals, G. Chi-Rosso, S.T. Ryan, I. Sizing, M. Zafari, C. Benjamin, J. Singh, S.Y. Venyaminov, R.B. Pepinsky, V. Koteliansky, Biochemistry 38 (1999) 8280– 8288. [40] M.T. Henzl, M.E. Davis, A. Tan, Biochemistry 49 (2010) 2256–2268. [41] J. Song, K. Talbot, J. Hewitt, R.T. MacGillivray, E.L. Pryzdial, Biochem. J. 422 (2009) 257–264. [42] M. Yang, K. Horii, A.B. Herr, T.L. Kirley, J. Biol. Chem. 281 (2006) 28307–28317. [43] A.R. Zeibdawi, J.E. Grundy, B. Lasia, E.L. Pryzdial, Biochem. J. 377 (2004) 141– 148. [44] C.D. Waldburger, J.F. Schildbach, R.T. Sauer, Nat. Struct. Biol. 2 (1995) 122–128. [45] S. Kumar, R. Nussinov, ChemBioChem 3 (2002) 604–617. [46] S. Marqusee, R.L. Baldwin, Proc. Natl. Acad. Sci. USA 84 (1987) 8898–8902. [47] R.C. Foreman, FEBS Lett. 216 (1987) 79–82. [48] K.D. Walker, T.P. Causgrove, J. Mol. Model. 15 (2009) 1213–1219. [49] T. Kanzaki, S. Ushioku, A. Nakagawa, T. Oka, K. Takahashi, T. Nakamura, K. Kuwajima, A. Yamagishi, M. Yohda, Protein Eng. Des. Sel. 23 (2010) 393–402.