|Home | About | Journals | Submit | Contact Us | Français|
The hyperthermophilic archaeon Pyrococcus furiosus genome encodes three proteasome component proteins: one α protein (PF1571) and two β proteins (β1-PF1404 and β2-PF0159), as well as an ATPase (PF0115), referred to as proteasome-activating nucleotidase. Transcriptional analysis of the P. furiosus dynamic heat shock response (shift from 90 to 105°C) showed that the β1 gene was up-regulated over twofold within 5 minutes, suggesting a specific role during thermal stress. Consistent with transcriptional data, two-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed that incorporation of the β1 protein relative to β2 into the 20S proteasome (core particle [CP]) increased with increasing temperature for both native and recombinant versions. For the recombinant enzyme, the β2/β1 ratio varied linearly with temperature from 3.8, when assembled at 80°C, to 0.9 at 105°C. The recombinant α+β1+β2 CP assembled at 105°C was more thermostable than either the α+β1+β2 version assembled at 90°C or the α+β2 version assembled at either 90°C or 105°C, based on melting temperature and the biocatalytic inactivation rate at 115°C. The recombinant CP assembled at 105°C was also found to have different catalytic rates and specificity for peptide hydrolysis, compared to the 90°C assembly (measured at 95°C). Combination of the α and β1 proteins neither yielded a large proteasome complex nor demonstrated any significant activity. These results indicate that the β1 subunit in the P. furiosus 20S proteasome plays a thermostabilizing role and influences biocatalytic properties, suggesting that β subunit composition is a factor in archaeal proteasome function during thermal stress, when polypeptide turnover is essential to cell survival.
Proteasomes are heteromultimeric proteases found in all domains of life that play a key role in intracellular protein degradation (14-17, 26, 29-31, 33, 38, 42). In eukaryotes, the composition of the proteasome core particle (CP, or the 20S proteasome) is based on many versions of small α- and β-type proteins (21 to 31 kDa) that assemble into heptameric stacked rings (α7β7β7α7; the set of 14 α and 14 β subunits that comprise the CP assembly in eukaryotes can depend upon cellular status) (16). The Saccharomyces cerevisiae CP can have up to seven different β proteins in its assembly, which contribute to multiple proteolytic activities (1). The genomes of higher eukaryotes, such as Arabidopsis thaliana, encode as many as 13 α-type and 10 β-type proteins, creating the possibility for numerous versions of the CP (12). Many α and β proteins in a particular eukaryote are more than 90% identical at the amino acid sequence level, possibly representing some degree of redundancy (13). CP composition may, in some cases, also be variable in prokaryotes, although the possibilities are much more limited. In bacteria, 20S proteasomes have been identified in Rhodococcus erythropolis (8) as well as in Mycobacterium tuberculosis (25), both of which appear to be based on single α and β proteins. In Haloferax volcanii, the proteasome CP is found in at least two different isoforms, based on combinations of one β protein and two α proteins (7, 21). The genome of Haloarcula marismortui encodes two versions of both the α and β proteins, although it is not known specifically how these contribute to the CP assembly (11). In the hyperthermophilic archaea, available genome sequence data indicate that the CP is based either on one α and one β protein or on two β proteins and one α protein. The role of an additional α or β protein in archaeal proteasome regulation may provide some biochemical and/or biophysical versatility along the lines seen in the eukaryotic proteasome. For example, in H. volcanii, which has two α proteins, it has been proposed that the CP assembly, and association with the proteasome-activating nucleotidases (PanA and PanB), may be growth phase associated (32). Environmental factors may also contribute to the β protein composition in hyperthermophiles with one α and two β proteins encoded in their genomes. N-terminal sequence analysis of the native CP purified from Pyrococcus furiosus grown at 88°C showed that it was based primarily on one α (PF1571) and one β (PF0159, referred to here as β2) protein (2). However, when P. furiosus was exposed to supraoptimal temperatures, transcription of an open reading frame (ORF) encoding the putative β1 protein (PF1404) was up-regulated after 60 min (37). The extent to which β1 is incorporated into the P. furiosus CP and the resulting functional implications have not been examined. Given the relative simplicity of the archaeal proteasome compared to eukaryotic versions, many interesting questions concerning the relationship between CP α and β protein content can be addressed. For example, do hyperthermophilic archaea with genomes encoding a second β protein (e.g., P. furiosus  and Sulfolobus solfataricus ) have any physiological or ecological advantages over those with only one β protein (e.g., Methanococcus jannaschii [27, 28] and Archaeoglobus fulgidus )? To begin to answer this question, we examined the influence of temperature on in vivo and in vitro CP assembly for P. furiosus.
CP peptidase activity was determined by endpoint assay in 50 mM sodium phosphate buffer (SPB), pH 7.2, and 95°C (unless otherwise noted), with a microtiter plate reader (model HTS 7000 Plus Bio Assay reader; Perkin-Elmer, Wellesley, MA) by detection of 7-amino-4-methylcoumarin (MCA) released from the carboxyl terminus of N-terminally blocked peptides (Sigma-Aldrich, St. Louis, MO) (2). Negative controls (no enzyme) were run in triplicate to account for thermal degradation of substrates. Kinetic constants were determined using a least squares fit of the appropriate model to the initial velocity (U/μg) as a function of substrate concentration (0.01 mM to 0.50 mM) at 95°C. One unit of protease activity was defined as the amount of enzyme required to release 1 pmol of MCA per min. Total protein concentrations were determined using the Coomassie blue dye-binding method (4) (Bio-Rad, Hercules, CA) in microtiter plates with bovine serum albumin (Sigma-Aldrich, St. Louis, MO) as the standard.
P. furiosus (DSM 3638) was grown on sea salts medium (40 g/liter sea salts [Sigma-Aldrich, St. Louis, MO], 3.3 g/liter piperazine-N,N-bis(2-ethanesulfonic acid) buffer, 1 ml/liter trace elements, 5 g/liter tryptone, 1 g/liter yeast extract, 120 g total sulfur). The culture (12-liter working volume) was grown in a 14-liter fermentor (New Brunswick, Edison, NJ) at an agitation rate of 600 rpm, pressure of 0.5 bar, and sparge rate (N2) of 0.5 liter/min. The cells grew at 80°C (pH 6.2) for 16.0 h and then were shifted to 90.0°C and held at this temperature for 1 h. Culture samples of 2.4 liters were taken at 16 h and 17 h; the resulting cell pellet was stored at −80°C.
Purification of the proteasome from cells grown at 80°C and 90°C was done using the same purification scheme. The cell pellets were resuspended in 20 ml of 20 mM Tris, pH 8.0, passed (four times) through a French pressure cell at 16,000 lb/in2, and centrifuged (10,000 × g, 4°C) for 20 min. The soluble protein fraction was applied to a 40-ml Q-Sepharose XK 26/20 column (GE Life Sciences, Piscataway, NJ) and eluted between 0.5 and 0.7 M NaCl. Many of the smaller contaminating proteins were cleared from the resulting VKM-MCA active pool using a Microcon centrifugal concentrator of 100,000 molecular weight cutoff (MWCO). This pool was then applied to a hydroxyapatite (Calbiochem, San Diego, CA) XK 16/30 column. Elution occurred between 220 and 265 mM SPB during a linear gradient of 0.05 to 0.5 M SPB, pH 8.0. After concentration, this active pool was applied to a HiPrep Sephacryl S-300 high-resolution XK 16/40 column calibrated with a high-molecular-weight calibration kit (GE Life Sciences, Piscataway, NJ). All CP VKM-MCA activity was eluted in the first peak, which corresponded to a protein of approximately 660 kDa; these fractions were then concentrated using a 30,000 MWCO Centriplus centrifugal filter device (Millipore, Billerica, MD).
The genes for the three proteins associated with the proteasome, psmA (α, PF1571; proteasome, subunit alpha), psmB-1 (β1, PF1404; proteasome, subunit beta), and psmB-2 (β2, PF0159; proteasome, subunit beta), were separately cloned into the pET-24d(+) vector (Novagen, Madison, WI). The α gene was amplified using the following primers: forward, 5′-TGAACGCCATGGCATTTGTTCCACCTCA-3′; reverse, 5′-ATAAAAATTGGATCCAAGTCAGTAGTTGCTATCCA-3′. The β2 gene was amplified with forward primer 5′-TTAGGTGGTGCTCATGAAGAAAAAGACTGGAA-3′ and reverse primer 5′-TAAGGAAGCCTGGATCCTTCATACTACAAACTCTT-3′. The β1 gene was cloned using an ORF that started with the fourth amino acid from the reported amino terminus (based on locations of the start codon and likely ribosomal binding site). N-terminal sequencing confirmed that the unprocessed β1 subunit was expressed correctly. The primers used for β1 gene amplification were forward, 5′-TGTTGCCCATGGAAGAGAAACTTAAGGGAA-3′, and reverse, 5′-AAATTGTCGGATCCTTGGACTACTTTAACATTTT-3′.
The α gene was expressed in E. coli BL21(DE3), while the β1 and β2 genes were separately expressed in E. coli BL21-CodonPlus(DE3)-RIL (Stratagene, La Jolla, CA). Expression was induced with 0.4 mM isopropyl-β-d-thiogalactopyranoside (IPTG) (optical density at 595 nm, 0.60); cells were harvested 3 to 5 h after induction (37°C). The resuspended cell pellets were treated with lysozyme, sonicated (Misonix, Inc., Farmingdale, NY), and centrifuged (18,000 × g, 4°C) for 30 min. Two 20-min heat treatments of the soluble protein fractions for α and β1 were performed, the first at 85°C and the second at 90°C, to remove residual E. coli protein. Each treatment was followed by cooling on ice for 30 min and centrifugation (18,000 × g, 4°C) for 30 min to remove insoluble protein. The β2 protein preparation required only one 20-min heat treatment at 85°C.
To assemble the CP, the α and β proteins were combined in equimolar ratios to a final total protein concentration of 0.5 to 0.7 mg/ml. This mixture was then incubated at the indicated temperature for 1 h and cooled on ice for 1 h, and precipitated material was removed through centrifugation (16,000 × g, 4°C) for 30 min. The CP soluble protein was then purified by a gel filtration step to remove unincorporated α and β proteins, using the approach described above for the native CP. Fractions were concentrated using a 30,000 MWCO membrane filter, as described above for the native CP.
The PAN gene (PF0115) was cloned into a pET-21b(+) vector (Novagen, Madison, WI) using the following primers: forward, 5′-GGTGATACATATGAGTGAGGACGAAGCTCAATTT3′; reverse, 5′-TAAAAATTAGGATCCTCAGCCGTAAATGACTTCA 3′. PAN was expressed using BL21-CodonPlus(DE3)-RIL (Stratagene) with induction by 0.4 mM IPTG (optical density at 595 nm, 0.60); cells were harvested 3 to 5 h after induction (37°C). Cells were resuspended in 20 mM Tris pH 8.0 plus 0.5% 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS), 1 mM dithiothreitol (DTT), 1 mg/ml lysozyme and sonicated (Misonix, Inc., Farmingdale, NY). The samples were centrifuged for 30 min (18,000 × g, 4°C), and the supernatant was removed. The pellet was then resuspended in 20 mM Tris pH 8.0 plus 0.5% CHAPS, 1 mM DTT and heat treated at 85°C for 20 min. The sample was then cooled and centrifuged, and the pellet was then resuspended in the same manner as before and heat treated at 90°C for 20 min. The resulting suspension was centrifuged to remove precipitated debris, with the supernatant containing pure PAN.
Purified samples of the proteasome (45 μg) were precipitated in a 10% trichloroacetic acid solution on ice for 1 h. The resulting pellet was washed three times with 150 μl of ice-cold acetone (−20°C) and then dried for 5 min at 60°C. The protein was resuspended in 125 μl of rehydration buffer (8 M urea, 2% CHAPS, 50 mM DTT, 0.2% Bio-Lyte ampholytes [Bio-Rad, Hercules, CA]) and applied to a 7.0-cm pH 4 to 7 isoelectric focusing (IEF) strip (Bio-Rad). The strip was subjected to active rehydration (50 V) for 16 h. The conditions used for focusing were as follows: 250 V, linear ramp of 20 min; 4,000 V, linear ramp of 2 h; 4,000 V, rapid ramp of 10,000 V-h. After IEF, the strips were incubated in equilibration buffer I (6 M urea, 0.375 M Tris-HCl, pH 8.8, 2% sodium dodecyl sulfate [SDS], 20% glycerol, 2% DTT) for 10 min at room temperature, followed by another 10 min incubation in equilibration buffer II (6 M urea, 0.375 M Tris-HCl, pH 8.8, 2% SDS, 20% glycerol, 2.5% iodoacetamide). For the second dimension, the IEF strip was then placed on top of a 12% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) gel and covered with a two-dimensional (2D) agarose overlay gel (0.5% low-melting-point agarose, Tris base [2.9 g/liter], glycine [14.4 g/liter], SDS [1.0 g/liter]). The gels were stained with GelCode Blue staining reagent (Pierce, Rockford, IL) and analyzed on a GS-710 calibrated imaging densitometer (Bio-Rad).
The melting temperatures of all expressed proteins were determined using a CSC Nano differential scanning calorimeter (DSC; Calorimetry Sciences Corp., American Fork, UT). All samples were dialyzed against 50 mM SPB, pH 7.2, which was the buffer used to generate the baseline scan. Samples (0.21 mg/ml) were degassed and scanned from 25 to 125°C using a scan rate of 0.5°C/min for two heating and cooling cycles. Heat capacity-versus-temperature curves were generated using the software program accompanying the DSC instrument to determine melting temperatures. After each sample was analyzed on the DSC, activity assays and native gels were used to determine if complete or irreversible denaturation had occurred. Samples were centrifuged (16,000 × g, 4°C) to remove aggregates, total protein concentrations were determined, and activity assays were run simultaneously against the corresponding initial samples to obtain relative loss of activity.
High-temperature incubation of the active recombinant proteasome forms (α+β2 and α+β1+β2 assembled at 90°C and 105°C) was used to compare their stabilities. Each assembly was adjusted to a baseline concentration of 0.15 mg/ml and incubated at 115°C in an oil bath for up to 12 h. Aliquots were taken at time points from 0 to 12 h and stored on ice until the end of the incubation period. The standard VKM-MCA microtiter plate assay was then used to compare the activities of the mixtures (300 ng enzyme, based on preincubation concentration, was mixed with 5 μM VKM-MCA and heated to 95°C for 15 min). The resulting fluorescence scores, with average background values subtracted, were determined and used to determine first-order decay constants.
A whole-genome cDNA microarray including 2,065 ORFs was printed, following protocols described previously (5). The array was used to determine the transient-transcriptional response after a temperature shift from 90°C to 105°C. P. furiosus (DSM 3638) was cultured anaerobically at 90°C on sea salts medium (SSM), as described previously (37). Tryptone (Sigma, St. Louis, MO) was added to SSM (final concentration, 3.28 g/liter) as a carbon source prior to inoculation, along with elemental sulfur (10 g/liter). A 60-ml batch culture was used to inoculate 500 ml of SSM supplemented with 3.28 g/liter tryptone and 10 g/liter sulfur in a 1-liter pyrex bottle. A 250-ml aliquot of this culture was added to 12 liters of medium in a 14-liter fermentor (New Brunswick Scientific, Edison, NJ). The fermentor contained an internal temperature controller, and the pH was maintained by a Chemcadet controller (Cole Parmer, Vernon Hills, IL). High-purity N2 was used to reduce the medium and to sparge during inoculation. The culture was grown to mid-log phase at 90°C, after which a sample was collected. The temperature set point was then shifted to 105°C, with the culture taking approximately 2 min to reach the set point temperature. Once the culture reached 105°C, samples were taken at 0, 5, 10, 60, and 90 min. Approximately 20 ml of culture was collected prior to sampling at each time point to eliminate preexisting fluid in the sampling lines. At each time point, 500 ml of culture was withdrawn and immediately put on ice until it was processed for RNA extraction. One ml of sample was removed for cell density enumeration by epifluorescence microscopy with acridine orange stain (20).
RNA was extracted from each 500-ml sample culture as described previously (37). The 500-ml samples from the fermentor were centrifuged for 20 min (10,000 × g, 4°C). After treatment with RNA lysis buffer, the samples were stored at −70°C. Extractions proceeded with ethanol precipitation and purification using Ambion RNAqueous kits. Concentrations and degree of purity were determined by optical density at 260 nm and 280 nm, as well as with gel electrophoresis (1% agarose gel, 60 V). Procedures for reverse transcription reactions, aminoallyl labeling with Cy3 and Cy5, and hybridization reactions are reported elsewhere (5).
A loop experimental design incorporated reciprocal labeling of time point samples with both Cy3 and Cy5. Mixed model analysis was used to evaluate differential expression data using approaches presented elsewhere (5). Briefly, least squares estimates of gene-specific treatment effects, corrected for global and gene-specific sources of error, were used to construct pair-wise contrasts analogous to changes for each gene between all pairs of conditions. The statistical significance of these changes was determined, and a Bonferroni correction was used to establish an experiment-wide false-positive rate of α = 0.05 by dividing α by 2,821, the number of comparisons performed for all genes over all possible treatment pairs. The corrected false positive rate was 1.77 × 10−5 [corresponding to a −log10(P) of >4.8]. Least squares estimates of gene-specific treatment effects were also used to perform hierarchical clustering in JMP 5.0 (SAS Institute, Cary, NC).
We previously reported using a targeted cDNA microarray to examine the transcriptional response of P. furiosus grown quiescently in serum bottles, comparing before and 60 min after a temperature shift from 90 to 105°C (37). Table Table11 shows more recent efforts focused on tracking transcriptional transients for P. furiosus grown in a 14-liter agitated fermentor for the same thermal shift. The larger culture volume allowed for multiple time point RNA samples for microarray interrogation. Also, agitation effected rapid equilibration of the entire culture volume to the higher temperature. At the earliest sampling time (0+), which was 2 min or less following temperature shift, the thermosome (PF1974) and small heat shock protein (PF1883) were up-regulated 8.6-fold and 34.3-fold, respectively, indicative of a heat shock response. The β2 protein and PAN (PF0115) were not responsive at the 0+ sampling time, while the α protein was down-regulated twofold. The β1 protein was up-regulated twofold at the earliest sampling time (0+). In fact, transcript levels for this gene were maintained at normal or above-normal levels throughout the 90-min tracking period, in contrast to the β2 and α proteins, which both dropped off considerably by 90 min. The gene encoding PAN rose twofold at 30 min but ultimately dropped off more than threefold by 60 min.
Some differences in transcript levels were noted at the 60-min point between the quiescent (37) and agitated cultures for certain ORFs (data included in Table Table1).1). Agitation seemed to accelerate the dynamics of the heat shock response; note that transcript levels at 30 min for the agitated culture were comparable to the 60-min levels for the quiescent cultures. Inspection of cells using epifluorescence microscopy with acridine orange stain showed that significantly more cell lysis was occurring at the 60-min point in the agitated culture than in the quiescent culture, perhaps reflecting the adverse impact of increased shear sensitivity on cellular function at supraoptimal temperatures.
Since transcriptional analysis indicated that β1 protein transcripts were up-regulated upon heat shock, an effort was made to determine whether growth temperature impacted the β2/β1 protein ratio in the CP. Native versions of the proteasome were purified from P. furiosus cells grown in the agitated 14-liter fermentor at 80 and 90°C. Significant cell lysis at 105°C, the temperature used to elicit a heat shock response reported in Table Table1,1, made purification of the native proteasome at this temperature problematic. Densitometry on CP proteasome proteins, separated by two-dimensional gel electrophoresis, from the two growth temperatures showed that the ratio of β2/β1 dropped from 3.0 at 80°C to 1.8 at 90°C (Fig. (Fig.11).
At 90°C, combinations of recombinant versions of α and β1 proteins led to nothing more than α and β1 proteins by themselves as viewed by native PAGE (data not shown). However, the α+β2 and α+β1+β2 combinations at 90 and 105°C led to active and fully assembled (~660-kDa) proteasomes. Furthermore, recombinant versions of the P. furiosus CP reflected the β2/β1 ratio observed for the native CP; at 80°C the β2/β1 was 3.8, which dropped to 1.9 at 90°C. The effect of assembly temperature on recombinant CP β protein composition at 100°C (β2/β1 = 1.4) and 105°C (β2/β1 = 0.9) reinforced the trend of increasing β1 protein content with increasing temperature (Fig. (Fig.11).
Recombinant CP assemblies were screened for activity against a number of peptide substrates. For those peptides for which hydrolysis was noted, relative activities of the 105°C-assembled proteasome were as follows: VKM LLL > LLE > LLVY > AAF. These results were consistent with a previous characterization of the native CP from P. furiosus (2). VKM was used as the comparative basis to track the activities of various assemblies of the recombinant 20S complexes. Each individual protein (α, β1, and β2) was inactive against VKM-MCA when initially expressed and stored at 4°C or when heated at 90°C, 98°C, or 105°C under the assembly conditions used here. When β1 and β2 proteins were combined in a 1:1 molar ratio, they were not active against VKM-MCA after incubation at temperatures of 4°C, 85°C, 90°C, or 105°C. Compared to the native proteasome, the combinations α plus β2 (assembled from a 1:1 molar ratio) and α plus β1 plus β2 (assembled from a 1:1:1 molar ratio) were active after incubation at 90, 98, and 105°C. The α+β1 (1:1 molar ratio) combination was essentially inactive at any assembly temperature. When higher levels of β1 were added to the assembly mixture so that α/β1 molar ratios were 1:5 and 1:10, activity increased slightly with increased β1 but was still extremely low in comparison to the native form and the other recombinant forms of the CP.
The effect of assembly temperature (and hence, β protein composition) was examined for recombinant versions of the CP. Table Table22 shows kinetic parameters determined at 95°C for recombinant versions of the P. furiosus CP assembled at 90°C and 105°C on three different peptide substrates; note that a standard Michaelis-Menten model was used for LLVY and AAF, while substrate inhibition was included in the model for VKM (39). Km values measured were all between 20 and 70 μM, which are similar to values for CPs from Rhodococcus and Mycobacterium species (25). Vmax values obtained were also consistent with previously reported values for other prokaryotic CPs (25, 42) and consistent with the native P. furiosus CP (2). Based on the information in Fig. Fig.1,1, the CP assembly at 105°C had twice the β1 protein content than that at 90°C and comprised approximately half of the CP, compared to one-third of the CP at 90°C. Thus, β1 protein content impacted the kinetic parameters for the three substrates shown in Table Table2.2. In all cases, the catalytic efficiency (kcat/Km), measured at 95°C, was higher for the 105°C assembly than for the 90°C assembly. The kcat/Km was 2.4-, 3.6-, and 15-fold higher for the 105°C CP assembly than the 90°C version on VKM, LLVY, and AAF, respectively.
The thermostability of the individual recombinant CP proteins was assessed by differential scanning microcalorimetry. The α protein was found to be very thermostable, even relative to other P. furiosus proteins; no thermal transitions were noted for scans up to 120 to 125°C (data not shown), the upper limit that could be tested by the instrument used here. Melting temperatures for the β1 and β2 subunits were 104.4°C and 93.1°C, respectively (Fig. (Fig.2a).2a). In the absence of the α subunit, the propeptide region of six to seven N-terminal residues upstream of the putative active site Thr within the “TTT” tripeptide in each of the expressed β subunits was retained, as determined by N-terminal sequencing. It is not known whether the presence or absence of the N-terminal region impacts the thermostability of the β proteins. The P. furiosus version of PAN melted at 94.2°C (Fig. (Fig.2b2b).
Figure Figure33 shows the melting curves for the α+β2 and α+β1+β2 assembly mixture forms at 90°C and 105°C. The α+β1 combination was not tested because significant amounts of assembled CP were never obtained. A thermal transition was noted for all cases within 110.5 to 112°C. For the α+β1+β2 form assembled at 90°C (curve C), a transition at 104.5°C was noted, likely corresponding to unincorporated β1 subunit, which melts at that temperature (Fig. (Fig.2a).2a). No transition at this temperature was observed for the α+β1+β2 form assembled at 105°C (curve D). Furthermore, analysis of samples twice scanned to 125°C and cooled back to ambient temperatures showed that at least 25% of initial activity on VKM-MCA remained. The fully assembled CP, without evidence of denatured subunits, could also be viewed on SDS-PAGE (Coomassie stain) following thermal scans to 125°C (data not shown). This was consistent with what appeared to be a thermal transition (curve D) which was beginning at 120°C.
The thermal inactivation of the α+β2 and α+β1+β2 assemblies is shown in Fig. Fig.4.4. The calculated kobs values for α+β2 (90°C assembly temperature), α+β2 (105°C assembly temperature), and α+β1+β2 (90°C assembly temperature) were all relatively close, ranging from 0.15 to 0.19 h−1. In contrast, the calculated kobs for α+β1+β2 assembled at 105°C was significantly lower (0.025 h−1). When samples remaining after the 12-hour incubation period were viewed on a 10% native gel, only the α+β1+β2 form assembled at 105°C was visible by Coomassie staining.
The P. furiosus 20S proteasome β protein complement differs from several other archaeal proteasomes that have been characterized to date (A. fulgidus , T. acidophilum , M. thermophila , M. jannaschii [9, 10, 42], and H. volcanii ), in that it encodes two distinct β protein homologs (Table (Table3).3). Original isolation of the proteasome from H. marismortui showed a single α and β composition(11); however, later genome sequencing has revealed the coding possibility for two α and two β proteins. Unlike the case in eukaryotes, in which some multiple versions of α and β proteins have been observed to share 90% identical protein sequences (13), the P. furiosus β proteins are only 48% identical at the amino acid level. This suggests that the β subunits can play distinct roles in P. furiosus CP structure and function. Recombinant versions of the biocatalytically active P. furiosus CP could not be generated based solely on α and β1 proteins but could be obtained using combinations of α and β2 or all three proteins. Although the combinations of α and β2, lacking β1, were fully active, combinations that included β1 had enhanced catalytic efficiency at 95°C. Versions containing α, β1, and β2 proteins that had been assembled at 105°C were also significantly more thermostable. Thus, this leads to the conclusion that the β1 subunit, while not essential for catalytic activity, plays a role in stabilizing and activating the P. furiosus CP assembly, particularly at supraoptimal temperatures, such as those encountered during thermal stress events. Assembly of the recombinant CP from M. jannaschii was found to require the unfolding and refolding of the α and β proteins to obtain activity levels comparable to the native version (9, 10). In comparison, the in vitro CP assembly for T. acidophilum, H. volcanii, and M. thermophila did not require unfolding and refolding to produce a recombinant protein with biochemical characteristics analogous to those of the respective native enzymes (27, 41, 43). This indicates that the assembly protocol needed for a fully functioning proteasome may vary depending on the individual proteasome being investigated. Here, the native and active recombinant versions (α+β2 and α+β1+β2) of the P. furiosus CP were comparable with respect to relative activity of peptide substrates, such that the functional properties of the recombinant CP did not seem to be impacted significantly by assembly protocol. Certainly, the relationship between temperature and β-subunit composition was similar for native and recombinant assemblies (Fig. (Fig.1).1). This may be the result of the fact that each recombinant protein underwent heat treatment at 85 to 90°C for at least 20 min to facilitate purification prior to assembly. The melting temperatures determined for the recombinant CP proteins (Fig. (Fig.2)2) also appear to be consistent with previous reports on the thermostability of a range of P. furiosus proteins. However, hints of thermal transitions beginning at temperatures of 115°C and higher (Fig. (Fig.3)3) raise the prospect that multiple versions of a functional CP can exist when two β subunits are involved.
Even though archaeal 20S proteasomes are based on a limited number of α and β proteins, the results here support the prospect that distinct specialized roles for specific proteins can exist even in prokaryotes. The roles of the two α proteins in H. volcanii appear to be different, with separate proteasome structures being assembled based on different stoichiometric ratios of subunits per structure (21, 42). In fact, H. volcanii synthesizes at least two native versions of the CP, α1+β and α1+α2+β, which may recognize and degrade different types of substrates (21). Certainly in many eukaryotic forms of the proteasome (3, 6, 12), α and β subunits can play specific and distinctive roles. Yeast CPs are comprised of up to seven different β subunits, with certain β subunits responsible for separate proteolytic specificities (19). In fact, yeast proteasome β proteins interact to form active sites for proteolysis that nonetheless resemble archaeal CP active sites which are based on a single β protein (1).
Proteasome function was found to be essential for cell survival during heat shock but not under normal conditions for T. acidophilum (35), which has single α and β subunit types. The relationship between heat shock and CP α/β protein content has not been examined yet to any extent in eukaryotes or prokaryotes with multiple subunit forms. Certainly, temperature was found to be a factor in the composition of other complex multimeric archaeal proteins in thermophilic archaea, e.g., the hetero-oligomeric rosettasome in Sulfolobus shibatae has different α/β/γ subunit composition depending upon growth temperature (22, 23). In mammalian murine RM cells, even though heat shock led to decreased proteasome RNA levels and inhibited formation of the protease complex, no drastic alteration of CP α/β protein content was noted (24). In yeast, a heat shock transcription factor was found to coordinate expression of proteasome proteins to control proteasome synthesis under thermal stress, although the impact on this on CP constitution was not determined (18). Recent results from our lab showed that when S. solfataricus was shifted from optimal (80°C) to supraoptimal (90°C) temperatures, the transcription of ORFs encoding CP proteins (α, β1, and β2) was unaffected throughout the 60-min period following the temperature shift (40). It was also noted that the transcriptional levels of CP proteins in S. solfataricus were much higher than in P. furiosus under normal or stressed conditions, although the ORF encoding S. solfataricus PAN decreased significantly following the temperature shift. For the hyperthermophilic archaea, it is yet to be determined whether CPs with two β subunits confer any special advantages upon specific microorganisms with respect to thermal stress response. Perhaps those hyperthermophiles whose genomes encode two β proteins experience frequent temperature excursions in their natural habitats. In any case, the results here suggest that the impact of CP β subunit content on function at suboptimal, optimal, and supraoptimal temperatures merits further examination, with an eye towards insights that relate to eukaryotic CPs. Such efforts are currently under way.
This work was supported in part by grants from the NSF Biotechnology Program and the DOE Energy Biosciences Program. J.K.M. acknowledges support from an NIH Biotechnology Traineeship. L.S.M. acknowledges support from a GAANN Biotechnology Fellowship.
Published ahead of print on 17 November 2006.