Refolding Pathway and Association Intermediates of Glutamate Dehydrogenase from the Hyperthermophile Pyrococcus furiosus

Eur. J. Biochem. 239, 679-685 (1996) 0 FEBS 1996

Refolding pathway and association intermediates of glutamate dehydrogenase from the hyperthermophile Pyrococcus furiosus Valerio CONSALVI I, Roberta CHIARALUCE ’, Stefania MILLEVOI I , Alessandra PASQUO‘, Paola VECCHINI‘, Emilia CHIANCONE’ * and Roberto SCANDURRA

I Dipartimento di Scienze Biochimiche ‘A. Rossi Fanelli’, UniversitB ‘La Sapienza’, Roma, Italy ’ Centro di Biologia Molecolare del Consiglio Nazionale delle Ricerche, Roma, Italy

(Received 12 March/24 April 1996) - EJB 96 0357/4

The denaturation and renaturation processes of the hexameric glutamate dehydrogenase from the hyperthermophilic archaeon Pyrococcus furiosus have been investigated using guanidinium chloride as denaturant. The enzyme is highly stable and the transition midpoint for guanidinium chloride denaturation is 6.1 M. The recovery of enzyme structure occurs after dilution of the denaturant at 20°C through the formation of structured monomers. Concentration of the structured monomers leads to the formation of higher association states with a tertiary structure different from that of the native enzyme. Activity is observed only in the presence of the hexamers, although a heating step at 70°C is required to fully reactivate the hexamer formed at 20°C. The refolding process and the intermediate(s) were studied by activity assay, spectroscopic methods, size-exclusion chromatography, and ultracentrifugation analysis.

Keywords: glutamate dehydrogenase; guanidinium chloride; hyperthermophile ; intermediate ; renatur- ation.

Extremophilic Archaea living in marine superheated areas have aroused significant biotechnological interest during the last decade [ l . 21. The study of the enzymes extracted from these microorganisms may help in understanding the molecular mech- anisms selected to achieve thermal adaptation [ 3 ] .

In the hyperthermophilic archaeon Pyrococcus furiosus [4], 20% of the soluble proteins is represented by a NAD(P)-depen- dent hexameric glutamate dehydrogenase (GDH). GDH, a ubiquitous enzyme placed at the branch point between carbon and nitrogen metabolism, catalyzes the reversible oxidative de- amination of L-glutamate to 2-oxoglutarate and ammonia with the reduction of NAD(P) [5]. All glutamate dehydrogenases ex- tracted from extremophilic Archaea, similarly to what has been reported for the same enzyme from mesophilic sources [6, 71, are hexamers of identical 46-49-kDa subunits [8-111 with the notable exception of the halophilic phenotype [12]. The highly thermostable GDH from R furiosus (half-life of 12 h at 100°C) [9, 131 has been sequenced [14, 151 and crystals diffracting to beyond 0.2 nm have been obtained recently [16]. Since its high resolution three-dimensional structure is known [17] and its pri- mary structure shows 34 % identity with Clostridium symbiosum GDH 1181, Rfuriosus GDH represents a model to investigate the amino acid substitution(s) responsible for increased stability in hyperthermophilic multisubunit enzymes.

Expression of GDH from the hyperthermophile R furiosus [19] and ES4 [lo] in mesophilic systems yielded a recombinant

Correspondence fo V. Consalvi, Dipartimento di Scienze Biochim- iche ‘A. Rossi Fanelli’, UniversitB ‘La Sapienza’, Piazzale Aldo Moro, 5, 1-00185 Roma, Italy

Fux: +39 6 4440062. Abbreviations. GdmC1, guanidinium chloride ; GDH, glutamate de-

hydrogenase ; N, native enzyme ; SEC, size-exclusion chromatography ; U, enzyme unfolded by 8 M GdmCl treatment; V,, elution volume.

Enzymes. Glutamate dehydrogenase (EC 1.4.1.2); glutamate dehy- drogenase [NAD(P) ‘ ] (EC 1.4.1.3); glutamate dehydrogenase (NADP+) (EC 1.4.1.4).

protein with different activity and lower stability to heat expo- sure. The production of a recombinant protein with properties different from the native wild type may be due to several factors difficult to identify in an in vivo system. The study of in vitro refolding of R furiosus GDH may elucidate the role played by temperature and protein concentration in the assembly of an active enzyme starting from its guanidinium chloride (GdmCl) denatured species. This type of investigation is preliminary to a comparative study on the stability of mutant proteins and may represent an interesting system to evaluate the conditions for the acquisition of the native structure in vitro for proteins from extremophilic sources. GDH from the moderate thermophile Ba- cillus acidocaldarius [20] has been reconstituted after GdmCl unfolding. To our knowledge, the effect of temperature and other factors on the refolding of multisubunit enzymes from hyper- thermophilic sources has been studied in detail only for glyceral- dehyde-3-phosphate dehydrogenase from Thermotoga marititna

In this paper, we report the study of the in vitro reconstitu- tion of Rfuriosus GDH after GdmCl denaturation with particu- lar attention to the effect of temperature and protein concentra- tion on the refolding process and on the structural and functional properties of the product(s). The minimum state of association required for catalysis has been identified. The relationship be- tween the state of subunit association and GDH catalytic activity was investigated by activity measurements, spectroscopic meth- ods, size-exclusion chromatography (SEC), and ultracentrifuga- tion analysis.

[21, 221.

MATERIALS AND METHODS

Chemicals and buffers. Electrophoresis reagents were purchased from Bio-Rad. Trypsin from bovine pancreas (code TRTPCK) was from Worthington Enzyme (Cooper Biomedical). GdmCl (Biochemika microselect) and all the other chemicals

680 Consalvi et al. ( E M J. Biochem. 2391

were from Fluka. Buffer A was 20mM sodium phosphate, pH 7.0, containing 0.4 M Na,SO,.

Enzyme. P furiosus cells were grown on peptone and yeast extract as described elsewhere [23]. GDH from R,furiosus was purified as described in [8]. Protein concentration was deter- mined using &,,,) = 81.038 M-' cm-' calculated according to [24]. Enzyme activity was assayed at 25°C with a Perkin Elmer Lambda 16 computerized spectrophotometer at 340 nm ; the standard assay mixture contained 100 mM sodium phosphate, pH 7.6, 0.75 mM 2-oxoglutarate, 75 mM ammonia, 100 pM NADPH and 0.2-0.4 pg enzyme in 1 ml final volume. Kinetic parameters of the native and refolded enzyme were determined at 25°C and analyzed by ENZFITTER [25] using at least seven data pairs.

Molecular mass determination. The molecular masses of native and refolded R furiosus GDH and analysis of the associa- tion states during enzyme refolding were determined by SEC on a Superose 12 column (Pharmacia) at 20°C eluted with 20 mM sodium phosphate, pH 7.0, containing 0.4 M Na,SO, (buffer A) at a flow rate of 0.4 ml/min using a Dionex gradient pump. Elu- tion was monitored at 280 nm and 226 nm utilizing a Kontron D-430 computerized detector. The column was calibrated with horse spleen apoferritin (440 kDa, elution volume V, 8.2 ml), rabbit muscle aldolase (161 kDa, V, 9.9 ml), horse liver alcohol dehydrogenase (80 kDa, V, 11.2 ml), bovine serum albumin (66 kDa, V, 11.5 ml), ovalbumin (45 kDa, V , 12.2 ml), and cyto- chrome c (12 kDa, V, 14.4 ml).

Ultracentrifugation analysis. Sedimentation velocity ex- periments were carried out in a Beckman Optima XL-A analyti- cal ultracentrifuge at 30000 rpm or 40000 rpm and 20°C. The movement of the protein toward the bottom of the cell was de- termined by absorption scans along the centrifugation radius at a wavelength of 280nm. The protein concentration was 0.35 mg/ml in buffer A. Sedimentation coefficients were cor- rected to s,,,, using standard procedures. The observed sedi- mentation coefficients were correlated to the state of association of the enzyme by means of the equation s, = s,'@ where s,, and s,are the sedimentation coefficients of the n-mer and of the monomer, respectively, and rn is the molecular mass of the mo- nomer [26].

Spectroscopic techniques. Fluorescence and light scattering measurements were performed on a LSSOB Perkin Elmer spec- trofluorimeter using 1 -cm pathlength quartz cuvettes. Fluores- cence emission spectra were recorded at 300-400 nm at 20"C, with the excitation wavelength set at 295 nm. Light scattering was measured at 20°C with both excitation and emission wave- length set at 480 nm and by use of the same spectral bandwidth of 4nm. CD measurements were performed with a JASCO 5-720 spectropolarimeter : far-ultraviolet spectra were recorded i n a 0.2-cm pathlength quartz cuvette at 200-250 nm, near-ul- traviolet measurements were carried out at 250-310 nm in a 1- cm pathlength quartz cuvette. The results are expressed as the mean residue ellipticity (0) assuming a mean residue mass of 110 Da per amino acid residue.

Fluorescence quenching by acrylamide [27] was performed at 50 pg/ml protein concentration in buffer A at 20°C and the acrylamide concentration was varied over 0-210 mM. The exci- tation wavelength was 290 nm and emission spectra were re- corded at 300-400nm 10 min after the addition of the quencher. The values of effective quenching constants were ob- tained from modified Stern-Volmer plots analyzing FJAF versus l/[acrylamide] (15 data pairs) 1281.

Equilibrium transition studies and GdmC1-induced un- folding. I? furiosus GDH (50 pglml) was incubated at increasing GdmCl concentration (0-8 M) in 20 mM sodium phosphate, pH 7.0, at 20°C. After 24 h, fluorescence emission at 343 nm

and ellipticity at 222 nm were measured at 20°C and normalized according to [29]. For refolding experiments unfolding was per- formed by incubating the enzyme (0.1 - I .4 mg/ml) in the pres- ence of 8 M GdmCI, 1 mM dithio-m-threitol and 150 pM EDTA in 20 mM sodium phosphate, pH 7.0 for 24 h at 20°C. Kinetics of unfolding were studied by measuring the decrease in fluorescence intensity at 343 nni (295 nm excitation wavelength) every 2.5 s after 100-fold manual dilution of the protein in 8 M GdmCl under continuous stirring at 18°C. Control mixing of the protein in buffer A did not show any decrease in fluorescence intensity at 343 nm due to sample photobleaching.

Enzyme renaturation and characterization of the re- folded enzyme. Refolding after complete denaturation in 8 M GdmCl was started by 20-fold dilution of the unfolding mixture in buffer A at 20°C. The final GdmCl concentration in the rena- turation mixture was 0.4 M, while the protein concentration was in the range 5-70 pg/ml. For equilibrium studies, the catalytic activity at 25"C, fluorescence, and CD spectra at 20°C were measured 0.5 -24 h after dilution of the denaturant. Reactivation kinetics were studied by 20-fold dilution of the unfolding mix- ture in buffer A at 25, SO, 60, 65, 70, and 75°C at 15 pg/ml final protein concentration, by measurement of the catalytic activity at 25°C on aliquots (0.2 pg) withdrawn at increasing time in- tervals.

Characterization of the refolded enzyme after 24 h incuba- tion at 30-70 pg/ml at 70°C in buffer A was performed after ultracentrifugation at 50000Xg for 30 min at 4°C in a Beckman L7-55 ultracentrifuge. Thermal stability was studied by incuba- tion of the enzyme (50 pg/ml) in buffer A at 108°C in glass tubes sealed under vacuum; at various intervals, the enzyme was cooled at 20°C and catalytic activity was assayed on 0.2 pg.

Stopped-flow measurements. Rapid mixing experiments to measure the rate of structured monomers formation were per- formed with a SFA-20 rapid mixer (Hi-tech, Salisbury, England) connected to a cuvette designed by the manufacturer to meet the optical beam geometries of a LSSOB Perkin Elmer spectrofluo- rimeter. Recovery of fluorescence emission at 343 nm (295-nm excitation wavelength) was monitored at 18 "C after a jump from 8 M GdmCl to 1.6 M GdmCl achieved by using an asymmetric mixing ratio (1 : 5 ) . Emission changes were monitored every 100 ms. Possible mixing artifacts were checked using a 30 pM N-acetyltryptophanamide solution in 8 M GdmC1, which was di- luted as the protein unfolding mixture. Apparent mixing artifacts and heat effect because of GdmCl dilution were less than 5 % of the signal amplitude.

RESULTS

Unfolding. Incubation of P f ~ ~ r i o s u s GDH at increasing concen- trations of GdmCl resulted in a progressive change of intrinsic fluorescence emission and loss of secondary structure as moni- tored by far-ultraviolet CD (Fig. 1) with the same transition mid- point at 6.1 M GdmC1. The unfolding profiles are identical and indicate an apparent two-state transition from the native (N) to the unfolded (U) state with no intermediate(s) detectable at equi- librium [29]. A plot of the shift in the maximum fluorescence emission wavelength as a function of GdmCl concentration (data not shown) was identical to the denaturation profile. Above 7.5 M GdmC1, the ellipticity at 222 nm and the fluorescence emission at 343 nm were 5 % and 40% with respect to the values characteristic of the native enzyme while the maximum fluores- cence emission wavelength was red-shifted from 343 nm (N) to 358 nm (U).

In the kinetic studies the unfolding reaction of the enzyme was induced by a GdmCl concentration jump at 18°C from 0 M

Consalvi et al. ( E m J. Biochem. 239) 681

100

80

60

hx 40

I

M d

20

0

0 2 4 6 8 [GdmCl] (M)

Fig. 1. GdmCl induced denaturation of Z? furiosus glutamate dehy- drogenase. Fluorescence emission at 343 nm (295 nm excitation wave- length) (0) and ellipticity at 222 nm (.) were recorded at 20°C in the presence of increasing concentrations of GdmCl at 50 pg/ml protein con- centration. Data were normalized according to [29].

160 I 1

I 0 5000 10000 15000 20000 25000

Time (s)

- 4 "m Fig. 2. Kinetics of unfolding of I? furiosus glutamate dehydrogenase. Process monitored by following the decrease in fluorescence emission intensity at 343 nm (295 nm excitation wavelength). A concentration jump from 0 M to 8 M GdmCl in 20 mM sodium phosphate, pH 7.0, at 18OC was performed by manual mixing under continuous stirring after 100-fold dilution of the protein. The final protein concentration was 40 pg/ml. The lower panel vhows residuals (difference between the data and the fitted function).

to 8 M and subsequent time-dependent changes in fluorescence intensity were recorded at 343 nm. The kinetic amplitude ob- served in the unfolding reaction was identical to the static ampli- tude measured at equilibrium between N and U. Fig. 2 shows a progress curve for unfolding analyzed by non-linear fitting as a sum of two first-order decay processes, with half-lives of 38 min and 93 min. The amplitude of the first fast phase accounted for 85% of the observed change. The end point of the process mea- sured by the fluorescence change at 343 nm was also verified by far-ultraviolet CD spectroscopy of the protein denatured in 8 M GdmC1.

Refolding. Effect of protein concentrutiorz on the refolding at 20°C. After 20-fold dilution of the denaturant, the enzyme recovered the secondary structure (far-ultraviolet CD) and the native fluorescence properties (data not shown) in less than 2 min, but was totally inactive. The state of subunit association monitored by SEC, as expected, was dependent on the protein

260 I 1

- 120 B 140 I/ lq , , , , , 1 60

0 50 100 150 200 250 300 Time (s)

6 0

-6

Fig. 3. Kinetics of refolding of Z? furiosus glutamate dehydrogenase. The process was monitored by fluorescence emission recovery at 343 nm (295 nm excitation wavelength) after stopped-flow dilution of the un- folding mixture from 8 M to 1.6 M GdmCl in buffer A at 18°C. Protein concentration in the refolding mixture was 20 pg/ml. The lower panel shows the residuals (difference between the data and the fitted function).

concentration in the refolding mixture and on the time elapsed after dilution of the denaturant. When the unfolding mixture was diluted 20-fold by direct injection in the loop of the SEC col- umn, only one protein peak was observed in the elution profile at V, 12.1 ml, and this was independent of the initial protein concentration. The molecular mass of this species corresponded to a monomer of 47 kDa, based on the calibration reported in the Materials and Methods section.

The kinetics of structure recovery by stopped-flow dilution of the unfolding mixture from 8 M to 1.6 M GdmCl at 18°C were followed by fluorescence emission at 343 nm. Nonlinear regression analysis of the progress curves indicated a nionopha- sic exponential process with k,,,, of 1.6X10-* s-' (Fig. 3). The static amplitude of the refolding process measured at equilibrium after manual dilution of the unfolding mixture was comparable with that measured during stopped-flow refolding kinetics. Ki- netic parameters for formation of the monomer did not change by increasing protein concentration from 15 pg/ml to 70 Fg/ml. Moreover, upon increasing the denaturation time from 7 h to 48 h at 18"C, no variation in the unimolecular refolding rate constant and in the amplitude of the fluorescence signal were observed, and no additional refolding phase occurred.

Higher states of association were observed following the time course of reconstitution at protein concentrations in the range 28-70 pg/ml at 20°C. The refolding mixtures were in- jected onto a SEC column after increasing incubation time. Analysis of the elution profiles revealed the progressive decrease of the monomer peak at V, 12.1 ml and the concomitant appear- ance of a peak at V, 10.1 ml, which corresponds to the molecular mass of a trimer, assuming a spherical shape of the protein spe- cies. N o species between the monomer and the trimer peak were detectable. However, it should be noticed that a trimer and an elongated dimer could have a similar elution behaviour. The rate of increase of the reassociation intermediate peak area versus time was dependent on protein concentration. For the three ini- tial concentrations of 70, 40, and 28 pg/ml (Fig. 4A), the frac- tional amount of the intermediate increased with time and fol- lowed apparent third-order kinetics, with a rate constant of 7x10' M-' s-'. This finding does not imply the occurrence of a termolecular reaction, which is unlikely in solution, but is consistent with the formation of a trimeric intermediate (see Dis-

682 Consalvi et al. ( E M J. Biochern. 239)

50

4 5 1 A J 100

, , , , , , I n

0 30 60 90 120 150 180

Time (min)

h

B si L1 N Y

l 10 c -I

0 400 800 1200 1600

Time (min)

80

60

40 P

20

0 0 4 8 12 16 20 24

Volume (d)

h

Y k4

0 400 800 1200 1600

Time (min)

Fig. 4. Time course of f! furiosus glutamate dehydrogenase trimer and hexamer formation and reactivation. (A) Trimer assembly from structured monomers. The Area refers to the amount of trimer expressed as a percentage of the monomer plus trimer area. After various reconstitution times at 20°C aliquots of the refolding mixtures at 70 pg/ml (0) and 40 pg/ml (U) in buffer A were injected onto Superose 12 and eluted with the same buffer. Elution was monitored at 226 nm. (B) Hexarneric structure assembly from trimers as monitored by SEC at 0.75 mglml at increasing refolding time. The labeled peaks refer to hexamer (H) and trimer (T). (C) Integration of the hexamer peak areas eluted from SEC column at increasing time of reconstitution at 0.24 mg/ml (A) and 0.62 mg/ml (V) protein concentrations. The area refers to the amount of hexamer expressed as a percentage of the trimer plus hexamer area. (D) Specific activity measurements were performed at various time intervals on the refolding mixture incubated at 70°C at 30 pg/ml ( ) and at 20°C at 0.62 mg/ml (*).

cussion). No hexamer formation was observed even after pro- longed refolding time up to protein concentrations of 70 pglml.

Under the above mentioned conditions, association among structured monomers never exceeded 45% of the relative amount of trimer peak area and at protein concentrations of the refolding mixture higher than 100 pg/ml aggregation occurred as revealed by light scattering and SEC. However, aggregation was prevented if the refolding mixture was kept for 2 m i n at 20°C at a protein concentration below 100 yg/ml to allow the formation of structured monomer before increasing the protein concentration up to 0.2 -1 mg/ml by a rapid (G 10 min) ultra- filtration step in Centricon-30 tubes. The refolding products were analysed at various time intervals and light scattering as well as SEC did not reveal aspecific aggregation. Peak analysis by SEC revealed the progressive formation of hexamer (V, 8.9 ml) accompanied by a decrease of the trimer peak area (V, 10.1 ml) (Fig. 4B). Hexamer assembly followed a second-order reaction (Fig. 4C). Half-lives for hexamer formation were 870, 310, and 270 min at 0.24, 0.62, and 0.75 mg/ml, respectively. The refolding products analyzed in the analytical ultracentrifu- gation sedimented as single peaks characterized by s2(), values of 7.5 S and 10.4 S, which correspond to trimeric and hexameric species assuming a spherical shape of the molecules. It should be noticed that the sedimentation profile of the native enzyme

showed the presence of 10-15% of trimers in all the prepara- tions analysed (data not shown). These appeared as a slight skewness in the elution profile of SEC.

Characterization of reassociation intermediates. The refold- ing mixture did not show catalytic activity unless it contained at least 40% hexamer (Fig. 4D). However, even after 24 h at 20"C, when only the hexameric species was detectable in solution by SEC, the specific activity was fivefold lower than that of the native enzyme. Prolonged incubation of this partially active hex- amer up to 200 h at 20°C did not increase the specific activity; incubation at 70°C was necessary to obtain a higher yield of reactivation (Fig. 4D).

A spectroscopic study of the hexamer formed at 20°C re- vealed the same fluorescence properties as N, a similar far-ultra- violet CD spectrum (data not shown) and a significantly dif- ferent near-ultraviolet CD spectrum (Fig. 5) indicative of differ- ences in the tertiary structure of the two forms. The hexamer formed at 20°C has an increased positive intensity at 282 nm and 275 nm in a region diagnostic of tyrosine and at 255- 270 nm typical of phenylalanine with a less pronounced defini- tion of the 286-nrn and 292-nm bands of tryptophan. These re- sults suggest that the contribution of the ten tryptophans [14, 151 overlaps with an increased positive ellipticity due to constraints of tyrosines and/or to a new environment of these residues.

Consalvi et al. (Eur: J . Biochem. 239) 683

260 270 280 290 300 310 Wavelength (nm)

Fig. 5. Near-ultraviolet CD spectra of 19 furiosus glutamate dehydro- genase. Spectra of the native enzyme (N) (A), of the hexamer formed at 20°C prior reactivation at 70°C (B) and after a 40-min partial reacti- vation at 70°C (C) were performed at 1.44 mg/ml.

However, the main feature is a less pronounced definition of all the individual aromatic peaks indicating that tertiary contacts are different, and possibly less precisely formed, with respect to N, the native form of the enzyme. Near-ultraviolet CD spectra mon- itored during the 24-h assembly from a predominantly trimeric protein to the complete formation of the partially active hexamer at 20"C, did not show any appreciable spectral change (data not shown). Difference spectroscopy between N and the hexamer formed at 20 "C gave complementary information and showed two main difference peaks at 286 nm and 292 nm indicating that tryptophans are mainly responsible for the observed changes be- cause of their number and higher absorption coefficient. A com- parison of the difference spectrum between N and the hexamer formed at 20°C with that obtained between N and the structured monomer formed within 2 min after dilution of the denaturant, indicates that most of the tertiary changes associated with poly- mer formation are related to trimer assembly (data not shown). A difference spectrum between N and a predominantly trimeric protein was similar to N minus the hexamer. These results and the similarity of the near-ultraviolet CD spectra monitored dur- ing hexamer assembly from the trimeric intermediate( s) indicate that most of the spectral changes occur only upon trimer assem- bly whereas hexamer assembly does not modify tertiary contacts significantly.

The hexamer formed at 20°C sediments with essentially the same xz0, value (10.4 S) of the native enzyme (1 1 .O S) although the peak is slightly more diffused (data not shown), whereas the SEC elution profiles of the hexamer and N are indistinguishable.

The uncharged quencher acrylamide was used to probe the accessibility of the hydrophobic core and the dynamic properties of the protein unfolded in 8 M GdmCl and of the hexamer formed at 20°C compared to those of the native enzyme. Effec- tive acrylamide quenching constants obtained from modified Stern-Vollmer plots [28] were 5.3, 10.7, and 8.0 M-' for N, U, and hexamer, respectively. These data indicate that the transi- tions from N to U and from U to hexamer alter the solvent accessibility of tryptophans and the global dynamics of the pro- tein. A quantitative interpretation is not possible because l? furi- osus GDH is a heterogeneously emitting system although it is clear that in the hexamer more tryptophanyl residues are accessi- ble to the quencher than in N.

Effect of temperature on the enzyme reconstitution. The reac- tivation during the assembly of the hexamer at 20°C shows that lower state(s) of association are not competent for catalysis and that the complete formation of the hexamer is necessary albeit

70

60

50

40

5 30 m

h

v be

.d &

a * 20

10

0

4 I

0 300 600 900 1200 1500 Time (min)

Fig.6. Temperature effect on the time course of reactivation of 19 furiosus glutamate dehydrogenase. Enzyme activity was measured at various time intervals on the refolding mixture incubated at 50 (O), 60 (W), 70 (A), and 75°C (T). The protein concentration was 15 pg/ml.

not sufficient for the complete recovery of activity. A study of the effect of temperature on the reactivation of the hexamer formed at 20°C was possible since light scattering measure- ments and SEC did not indicate significant aggregation and since dilution of the reconstituted hexamers did not lead to dissociation into subunits. Fig. 4 D depicts the increase in spe- cific activity during incubation of the hexamer(s) at 70°C at 30 pg/ml in comparison to that obtained at 20°C and 0.62 mg/ ml. The incubation at 70°C yields an enzyme with the same specific activity of the native form and the total protein recovery was 75 % after ultracentrifugation. The tertiary structure of the fully reactivated enzyme is identical to the native form as well the sedimentation profile (data not shown). The conditions for reactivation were very stringent (30 pg/ml, 70"C, 24 h) since lower (5-20 pg/ml) or higher (80-100 pg/ml) protein concen- trations as well as temperatures (60°C and 80°C) yielded recov- eries of active protein lower than 40%.

The reactivation of hexamer required a long heating step probably because of trapping of the refolded hexamer(s) in the wrong conformation(s), which requires the proper activation en- ergy to form the native enzyme. To increase the rate of enzyme reconstitution, the effect of temperature on the refolding process was tested by 20-fold dilution of U at increasing temperatures (50-75 "C) in buffer A at 0.4 M residual GdmCl and at a protein concentration of 15 pg/ml to prevent aggregation. A screening of the effect of protein concentration (5-50 pg/ml) on the re- folding process at 70°C indicated 15 pg/ml as the optimal con- centration. The restoration of enzyme activity increased with time following first-order exponential processes for all the tem- peratures tested; the corresponding half-times for reactivation decreased with increasing temperature (Fig. 6). The percentage of total active protein recovery was always lower than that ob- tained after reactivation of hexamer.

The fully reactivated enzyme, obtained after 24 h incubation at 70"C, was compared to N in terms of catalytic parameters and thermal stability. K,, and V,,,, values for all the substrates of the refolded enzyme are the same as those determined for the native P. furiosus GDH. The thermal stability of the refolded enzyme was studied by incubation at 50 pg/ml at 108°C in buffer A. At increasing time intervals, the residual catalytic ac- tivity was measured and compared to the native enzyme incu- bated under the same conditions. The half-life for thermal inacti- vation at 108 "C of both the native and refolded P. furiosus GDH was 4 h.

684 Consalvi et al. (EUK J. Biochern. 239)

DISCUSSION

P.,furiosus GDH is a hexameric enzyme that binds one NADP/monomer. Homology-based modelling with the Clostrid- ium symbiosum counterpart [30] reveals 35% identity and 413 equivalent positions. The recently published X-ray structure of the R furiosus [ 171 enzyme indicates that the hexainer is a dimer of trimers (32 symmetry) with a 47-kDa monomer composed of two interacting domains. In accordance with the 32 symmetry arrangement for the six subunits shown by the X-ray structure, the present reconstitution experiments point to the formation of a trimeric intermediate as a preliminary step for the assembly of hexamer. Catalytic activity was observed only in the presence of the hexameric form of the enzyme, thus indicating that a lower state of association is not competent for catalysis.

Equilibrium and kinetic experiments indicate P. furiosus GDH as one of the most resistant dehydrogenases ever de- scribed. The resistance to physical stress (f,,, at 100°C was 12 h) [8] is paralleled by a high resistance to GdmCl denaturation. The denaturation profiles in GdmCl detected by two different physical techniques overlap completely and depolymerization and loss of secondary structure coincide. This result is paralleled by the observation that the monomer and all the intermediate(s) observed during reconstitution show the same far-ultraviolet el- lipticity and the same intrinsic fluorescence of the native en- zyme. The lack of evidence for significantly populated interme- diate state(s) during GdmCl-induced unfolding does not exclude their involvement in the denaturation process, since they might well be present without becoming detectable.

A general prospect of the refolding pathway of R furiosus GDH is summarized in Scheme 1 where the rate constants for each step are indicated. Dilution of denaturant rapidly causes shielding of the tryptophan residues from the solvent and re- stores the secondary structure of the monomer. The refolding rate constant for the formation of structured monomers is inde- pendent of protein concentration indicating that intrachain in- teractions, in the absence of significant interchain interactions, are sufficient to stabilize a compact monomer with secondary structure elements. Refolding rate constants for the formation of the monomer do not change upon increase of the unfolding time, thus indicating that proline isomerization is not a rate-limiting step for the formation of structured monomer(s).

The formation of higher states of association requires high enzyme concentrations during reconstitution. Concentration of the structured monomers by ultrafiltration up to 1 mg/ml never caused aggregation. In contrast, aggregation was observed when refolding was promoted by simple dilution of the unfolding mix- ture at concentrations higher than 100 pg/ml. These differences in behaviour indicate that folding intermediate(s) that do not lead to structured monomers are responsible for aggregation and that i n the structured monomers and/or the reconstitution inter- mediate(s) most of the hydrophobic groups are internaliaed.

Assembly of reconstitution intermediates is a relatively fast reaction in comparison with hexamer( s) formation and reactiva- tion indicating that in the structured monomer(s) all the essential

interacting portions necessary for polymer assembly are formed. The best fits for the data indicate that structured monomers asso- ciate to form trimers. An apparent termolecular process is un- likely in solution and the assembly of trimers may be repre- sented by the reversible formation of a dimer followed by the addition of a monomer to form the trimer in an irreversible reac- tion, in analogy to what was proposed for the trimer assembly of aspartate transcarbamoylase [31].

Near-ultraviolet CD and difference spectroscopy clearly in- dicate that all tertiary changes occur upon assembly of mono- mer(s) to the trimer state, whereas formation of hexamer(s) is accompanied by spectroscopically less evident modifications of tertiary contacts.

The product of refolding at 20°C is a hexamer with a defined quaternary structure, with tertiary contacts differently arranged and less active with respect to the native enzyme as indicated by the sedimentation velocity and spectroscopic characteristics. A comparative analysis of primary structures between meso- philic and thermophilic GDH enzymes indicates the hinge and the interface between the two domains of the 47-kDa subunit as a region for significant changes which can modify the dynamics of interdomain interaction. The refolding at 20°C may lock the protein in an alternatively folded conformation because of increased rigidity at the hinge region in the hyperthermophilic protein [30] whose native conformation can be restored by heat- ing to 70°C. The analysis of enzyme assembly at 20°C indicates that high temperatures are not required for reaching a quaternary structure closely similar to the native enzyme. The decreased flexibility of the protein and the internalization of hydrophobic residues may be responsible for the alternatively folded state adopted by the enzyme during refolding at 20OC.

The refolding by direct dilution of the unfolding mixture at 70°C is a faster, but less efficient first-order process if compared to that obtained when the unfolded protein is allowed to assem- ble at 20°C and is then reactivated at 70°C. This result suggests that some folding intermediates may not be temperature resistant and raises important questions about the in vivo folding of this protein which may need assistance. Similarly, the hexamer formed at 20°C appears less thermostable with respect to the native enzyme because it cannot be reactivated at temperatures higher than 70 "C suggesting a less temperature-stable structure for this intermediate despite its hexameric structure. The overall increased hydrophobicity of the pyrococcal protein relative to the mesophilic counterpart [15] may lead to aggregation during the 70°C rearrangement necessary to give a native hexameric structure. It is noteworthy that reactivation at high temperature is a first-order process when initiated from dilution of U or after formation of the hexamer, which suggests a monomolecular re- arrangement prior to formation of the active enzyme.

The enzyme obtained at the end of the process is identical to the native form as indicated by the kinetic and structural prop- erties and by the thermal stability. In general terms, it is of inter- est that a fully active hyperthermophilic enzyme can be obtained at temperatures significantly lower than those required for the

Scheme 1. General scheme for the refolding pathway of P.furiosus GDH.

U - Structuredmonomers - Trimers Hexamers

I I

I I

V Aggregates Active hexamers

Consalvi et al. (Eui: J. Biochem. 23Y) 685

growth of the microorganism (70°C versus 100°C), even though at room temperature an energy barrier prevents complete reacti- vation. The availability of a refolding procedure for a hyperther- mophilic GDH whose X-ray structure is known provides new possibilities for testing the crucial role of residues involved in conferring stability. In turn, the role of the primary structure in determining the attainment of the native structure can be studied in combination with the role of extrinsic factors like temper- ature.

We thank Drs N. Raven and R. Sharp at the Centre for Applied Microbiology and Research Division of Biotechnology, Porton Down, Salisbury, England, for providing cell paste of P: ,furiosus. This work was supported by the BIOTECH-Programme of the EC and by the Italian Minister0 UniversitZ & Ricerce Scientifica & Tecnologica (MURST). Part of this work will be submitted in partial fulfillment of the re- quirements of the Ph.D. degree at the University ‘La Sapienza’, Rome (by S. M.).

REFERENCES 1. Stetter, K. 0 (1986) Diversity of extremely thermophilic archaebac-

teria, in Thermophiles: general, molecular, and applied microbi- ology (Brock, T. D., ed.) pp. 39-74, John Wiley, New York.

2. Neuner, A,, Jannasch, H. W., Belkin, S. & Stetter, K. 0. (1990) Thermococcus Eitorulis sp. nov. A new species of extremely ther- mophilic marine Archaebacteria, Arch. Microbiol. Rev. 153, 205- 207.

3. Jaenicke, R. (1991) Protein stability and molecular adaptation to extreme conditions, EUK J. Biochem. 202, 715-728.

4. Fiala, G. & Stetter, K. 0. (1986) Pyrococcus,furiosus sp. nov. repre- sents a novel genus of marine heterotrophic archaebacteria grow- ing optimally at 100°C, Arch. Microbiol. 145, 56-61.

5. Smith, E. L., Austen, B. M., Blumenthal, K. M. & Nyc, J. F. (1975) Glutamate dehydrogenase, in The enzymes (Boyer, P. D., ed.) 3rd edn, vol. 11, pp. 293-367, Academic Press, New York.

6. McPherson, M. J. & Wootton, J. C. (1983) Complete nucleotide sequence of the Escherichia coli gdhA gene, Nucleic Acids Res.

7. Rice, D. W., Hornby, D. P. & Engel, P. C. (1985) Crystallization of an NAD’-dependent glutamate dehydrogenase from Clostridium symbiosum, J. Mol. Biol. 181, 1479-1491.

8. Consalvi, V., Chiaraluce, R., Politi, L., Vaccaro, R., De Rosa, M. & Scandurra, R. (1 991) Extremely thermostable glutamate dehydro- genase from the hyperthermophilic archaebacterium Pyrococcus furiosus, Eur. .I. Biochem. 202, 1 189- 11 96.

9. Consalvi, V., Chiaraluce, R., Politi, L., Gambacorta, A., De Rosa, M. & Scandurra, R. (1991) Glutamate dehydrogenase from the thermoacidophilic archabacterium Su1folohu.s .solfataricus, Eui: J. Biochem. 196,459-467.

10. Di Ruggiero, J., Robb, F. T., Jagus, R., Klump, H., Borges, K., Kes- sel, M., Mai, X . & Adams, M. W. W. (1993) Characterization, cloning, and in vitro expression of the extremely thermostable glutamate dehydrogenase from the hyperthermophilic archaeon ES4, J. Biol. Chem. 268, 17 767 - 17 714.

aton and properties of ex- tremely thermostable glutamate dehydrogenases from two hyper- thermophilic archaebacteria, Pyrococcus woesei and Pyrococci~~ furiosus, Biosci. Biotechnol. Blochem. 57, 945 - 95 1.

12. Leicht, W., Werber, M. M. & Eisenberg, H. (1978) Purification and characterization of glutamate dehydrogenase from Hulobacterium of the Dead Sea, Biochemistry 17, 4004-4010.

13. Robb, F. T., Park, J. B. & Adams, M. W. W. (1992) Characterization of an extremely thermostable glutamate dehydrogenase: a key en- zyme in the primary metabolism of the hyperthermophilic archae- bacterium Pyrococcus furiosus, Biochim. Biophys. Acta 1120, 267-272.

14. Eggen, R. L., Geerling, A. C. M., Waldkotter, K., Antranikian, G. & de Vos, W. M. (1993) The glutamate dehydrogenase-encoding

11, 5257-5266.

11. Ohshima, T. & Nishida, N. (1993) Pur

gene of the hyperthermophilic archaeon Pyrococcu.~ ,firriosus: se- quence, transcription and analysis of the deduced amino acid se- quence, Gene (Amst.) 132, 143-148.

IS. Maras, B., Valiante, S., Chiaraluce, R., Consalvi, V., Politi, L., De Rosa, M., Bossa, F., Scandurra, R. & Barra, D. (1994) The amino acid sequence of glutamate dehydrogenase from Pyrococcus furi- osus, a hyperthermophilic archaebacterium, J . Protein Chern. 13,

6. Yip, K. S. P., Baker, P. J., Britton, L., Engel, P. C., Rice, D. W., Sedelnikova, S. E., Stillman, T. J., Pasquo, A,, Chiaraluce, R., Consalvi, V. & Scandurra, R. (1995) Crystallization of the NAD(P)-dependent glutamate dehydrogenase from the hyperther- mophile Pyrocvccus furiosus, Acta Crystallogi: D51, 240-242.

7. Yip, K. S. P., Stillman, T. J., Britton, L., Artymiuk, P. J., Baker, P. J., Sedelnikova, S. E., Engel, P. C., Pasquo, A,, Chiaraluce, R., Consalvi, V., Scandurra, R. & Rice, D. W. (1995) The structure of Pyrococcus fkriosus glutamate dehydrogenase reveals a key role for ion-pair networks in inantaining enzyme stability at ex- treme temperatures, Structure 3, 1 147- 1 158.

18. Baker, P. J., Britton, K. L., Engel, P. C., Farrants, G. W., Lilley, K. S., Rice, D. W. & Stillman, T. J. (1992) Subunit assembly and active site location in the structure of glutamate dehydrogenase, Proteins Struct. Funct. Genet. 12, 75- 86.

19. Di Ruggiero, J . & Robb, F. T. (1995) Expression and in vitro assem- bly of recombinant glutamate dehydrogenase from the hyperther- mophi lic archaeon Pyrococcit~ ,furiosus, Appl. Environ. Microbiol.

20. Consalvi, V., M;!levoi, S., Chiaraluce, R., De Rosa, M. & Scandurra, R. (1 995) Refolding of glutamate dehydrogenase from Bacillus acidocaldarius after guanidinium chloride-induced unfolding, Biochem. Mol. Biol. Int. 35, 397-407.

21. Rehaber, V. & Jaenicke, R. (1992) Stability and reconstitution of D- glyceraldehyde-3-phosphate dehydrogenase from the hyperther- mophilic eubacterium Thermotoga maritinm, J . Bid. Chem. 267,

22. Rehaber, V. & Jaenicke, R. (1993) The low-temperature folding in- termediate of hyperthermophilic ~-glyceraldehyde-3-phosphate dehydrogenase from Thermotoga maritima shows a native-like cooperative unfolding transition, FEBS Lett. 31 7, 163 - 166.

23. Raven, N., Ladwa, N., Cossar, D. & Sharp, R. (1992) Continuous culture of the hyperthermophilic archaeum Pyrococcus fiiriosus, Appl. Microfiiol. Biotrchnol. 38, 263 -267.

24. Gill, S. C. & von Hippel, P. H. (1989) Calculation of protein extinc- tion coefficients from amino acid sequence data, Anal. Biochem.

25. Leatherbarrow, R. J. (1987) ENZFITTER. A non linear regression data analysis of program for the IBM-PC, Elsevier-BIOSOFT, Cambridge, England.

26. Gilbert, G. A. (1959) Sedimentation and electrophoresis of interact- ing substances. I. Idealized boundary shape for a single substance aggregating reversibly, Proc. Roy. Snc. Lond. B Biol. Sci. 250, 377-388.

27. Eftink, M. R. & Ghiron, C. A. (1981) Fluorescence quenching studies with proteins, Anal. Biochem. 114, 199-227.

28. Lehrer, S. S. (1971) Solute perturbation of protein fluorescence. The quenching of the tryptophyl fluorescence of model compounds and of lysozyme by iodide ion, Biochemistry 10, 3254-3263.

29. Pace, C. N. (1986) Determination and analysis of urea and guanidine hydrochloride denaturation curves, Methods Enzyrnol. 131, 266- 280.

30. Britton, L. K., Baker, P. J., Borges, K. M. M., Engel, P. C., Pasquo. A,, Rice, D. W., Robb, F. T., Scandurra, R., Stillman, T. J. & Yip, K. S. P. (1995) Insights into thermal stability from a comparison of the glutamate dehydrogenases from Pyrococcus ,furiosus and Thermococcus litoralis, Eur. J. Biochem. 229, 688-695.

31. Burns, D. L. & Schachman, H. K. (1982) Assembly of the catalytic trimers of aspartate transcarbamoylase from folded monomers, J . Biol. Chem. 257, 8638-8647.

253-259.

61, 159-164.

10999-11 006.

182, 219-326.