- Email: [email protected]
Monofunctional domains of formiminotransferase-cyclodeaminase retain similar conformational stabilities outside the bifunctional octamer
Biochimica et Biophysica Acta 1338 Ž1997. 223–232
Monofunctional domains of formiminotransferase-cyclodeaminase retain similar conformational stabilities outside the bifunctional octamer Laura Lea Murley, Robert E. MacKenzie
)
Department of Biochemistry, and Joint Centre for Structural Biology, 3655 Drummond St., Montreal, Que. H3G 1Y6, Canada Received 16 August 1996; revised 7 November 1996; accepted 11 November 1996
Abstract Each identical subunit of octameric formiminotransferase cyclodeaminase consists of a transferase and a deaminase domain connected by a short linker sequence. Both domains can be independently expressed in Escherichia coli as monofunctional dimers and show no indication of associating, suggesting that the linker mediates the only substantial interaction between the transferase and deaminase domains. To better understand the benefits arising from octamer formation, we have used equilibrium unfolding methods to examine the properties of the transferase and deaminase domains independently and within the octamer. Each isolated dimeric domain undergoes an apparent change in tertiary structure at low concentrations of urea Ž- 2 molrl. which results in the concurrent loss of intrinsic fluorescence and catalytic activity. The full length octameric enzyme also undergoes inactivation and a loss of intrinsic fluorescence over this concentration range, without apparent change in secondary or quaternary structure. Between 2 and 2.5 M urea the isolated transferase and deaminase domains dissociate to monomers. However, only one of the subunit interfaces in the octamer is disrupted at this urea concentration and dissociation of the second interface occurs between 3.5 and 5 M urea. While each domain shows similar stability to denaturation within and outside of the octamer, one type of subunit interface achieves increased stability within the full length enzyme. Keywords: Conformational stability; Subunit interface; Domain interaction; Equilibrium unfolding; Octamer; Bifunctional enzyme
1. Introduction
Abbreviations: FTCD, formiminotransferase cyclodeaminase; FTCDH 6 , hexahistidine-tagged formiminotransferase cyclodeaminase; FTH 6 , hexahistidine-tagged formiminotransferase; CDH 6 , hexahistidine-tagged cyclodeaminase; CD, circular dichroism; l max , wavelength of maximum fluorescence emission; SEC, size-exclusion chromatography; BSA, bovine serum albumin; Ni-NTA, nickel-chelated nitrilotriacetic acid matrix; H 4 PteGlu, tetrahydrofolate; DTT, dithiothreitol; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis. ) Corresponding author. Fax: q1 Ž514. 3987384; E-mail: [email protected]
The bifunctional enzyme formiminotransferase cyclodeaminase Ž FTCD. catalyzes two sequential reactions in the histidine degradation pathway. The transferase activity transfers a formimino group from the histidine catabolite form im inoglutam ate to H 4 PteGlu n . The cyclodeaminase catalyzes the deamination of the formimino intermediate producing 5,10-methenylH 4 PteGlu n and NHq 4 . With monoglutamylated folates Ž n s 1., the two activities function independently, however, polyglutamylated intermediate Ž n G 4. can be efficiently channelled from the
0167-4838r97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 6 7 - 4 8 3 8 Ž 9 6 . 0 0 2 0 8 - 7
224
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
transferase to the deaminase active site w1,2x. FTCD preferentially channels pentaglutamates and MacKenzie and coworkers w1,2x have proposed that the polyglutamate tail acts as a noncovalently bound swinging arm which anchors the intermediate to the enzyme while allowing the pteroyl moiety to move between the two active sites. FTCD is composed of 8 identical subunits arranged to form a planar, ring-shaped tetramer of dimers w3,4x. Denaturationrrenaturation w5,6x and deletion analysis studies w7x indicate that the octamer is the smallest functional unit of FTCD to retain both catalytic activities and the ability to channel substrates. The FTCD octamer contains 4 high-affinity polyglutamate binding sites w2x, and Findlay and MacKenzie w5,6x have proposed that the integrity of alternating subunit interfaces within the native octamer is a prerequisite for expression of both catalytic activities. While this suggests that the active sites may be situated at the subunit interfaces, exact numbers of catalytic sites and their relative location within the octamer are not known. We have recently determined that each subunit of FTCD consists of an N-terminal transferase domain and a C-terminal deaminase domain separated by a short linker sequence w8x. The polyglutamate binding site maps to the deaminase domain. These domains can be independently expressed in E. coli and show no indication of interacting with each other once purified. This suggests that the linker mediates the only substantial interaction between the transferase and deaminase domains within each subunit of FTCD. Both domains contain sequences governing subunit association, and exist as monofunctional dimers in solution w8x. It is possible that the octameric structure has evolved from the fusion of two different dimeric monofunctional proteins, such that the subunit interfaces formed by identical domains are necessary for the maintenance of domain structure and function, but the hetero-domain interaction within the subunit is relatively unimportant. If this model were accurate, the two types of domains should behave similarly within and outside of the octamer. The ability of FTCD to channel substrate and the 50% decrease in k cat values observed for each isolated domain w8x argue that the transferase and deaminase domains do not function completely autonomously of each other within the octamer.
Substrate channelling is one of the more obvious benefits of octamer formation. We would like to establish if fusion of the transferase and deaminase domains might result in other advantages, such as increased stability of the domains or subunit interfaces. In this paper we examine the urea-induced denaturation of the transferase and deaminase domains independently and as part of the FTCD octamer.
2. Materials and methods 2.1. Materials Folic acid and formimino-L-glutamic acid were from Sigma. Urea solutions ŽUltra Pure urea, ICN. were prepared immediately prior to each experiment to minimize the formation of isocyanate. Ni-NTA resin was purchased from Qiagen. The Superose 6 HR 10r30 column was from Pharmacia. 5-formiminoH 4 PteGlu was prepared enzymatically as described previously w2x except that purified recombinant transferase domain w8x was used to catalyze its synthesis. Acidified 5-formiminoH 4 PteGlu was centrifuged in a microfuge for 10 min at 48C to pellet precipitated enzyme, instead of passing the solution through Centriflo cones ŽAmicon.. 2.2. Enzyme purification Recombinant histidine-tagged FTCDH 6 , FTH 6 and CDH 6 were expressed in E. coli and purified as previously described w8x. Enzyme preparations had specific transferase activities of approx. 30 m molrmin per mg for FTCDH 6 , and 22–24 m molrmin per mg FTH 6 , and specific deaminase activities of 18–21 m molrmin per mg for FTCDH 6 and 16–20 m molrmin per mg for CDH 6 when assayed with 0.1 mM 5-formiminoH 4 PteGlu. Single bands were observed upon SDS PAGE. Triton X 100 was replaced with Tween 20 by rebinding the enzyme to Ni-NTA resin and washing and eluting with buffers containing 0.05% Tween 20 in place of Triton X 100. Enzyme was concentrated to approx. 4 mgrml ŽCentriprep, Amicon. and stored at y208C in buffer containing 0.1 M potassium phosphate ŽpH 7.8., 0.05% Tween 20, 40% glycerol and 1 mM DTT.
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
Protein concentrations Ž as monomer. were determined using the calculated extinction coefficients w9x E280 s 37 920 My1 cmy1 for FTCDH 6 , 28 150 My1 cmy1 for FTH 6 and 9770 My1 cmy1 for CDH 6 . 2.3. Unfolding studies Denaturation experiments were performed in Buffer A Ž0.1 M potassium phosphate, pH 7.3, 0.05% Tween 20 and 1 mM DTT. containing the appropriate concentration of urea Ž0 to 8 M. . Initial timecourse experiments indicated that recombinant FTCDH 6 and the isolated domains undergo almost all of their fluorescence and activity loss within the first 100 minutes of incubation at room temperature with a much slower loss thereafter. This is consistent with a previous observation that FTCD purified from pig liver undergoes significant inactivation within the first 2 hours of urea-induced denaturation w5x. In the following experiments, proteins were incubated for 150 minutes at room temperature prior to analysis, a time sufficient to achieve equilibrium, as monitored by intrinsic fluorescence and enzyme activity.
225
priate amount of urea Ž0 to 8 M. , and the emission spectra Žexcitation at 290 nm. were recorded between 320 and 380 nm. The excitation bandpass was set at 3 nm and the emission bandpass at 5 nm. The spectrum of a buffer blank was subtracted from each sample spectrum.
2.6. Circular dichroism The far-UV CD spectra of samples containing 10 m M protein, incubated for 150 minutes at room temperature in Buffer A containing urea Ž0 to 7.5 M., were recorded in a Jasco J-710 model with a cylindrical cuvette of 0.05 cm path length. Scan speed was set at 100 nmrmin and measurements are reported from 5 accumulations. The spectrum of a buffer blank was subtracted from each sample spectrum and molar ellipticity was determined. Dynode voltages at 222 nm remained at or below 300 volts, even at the higher urea concentrations.
2.7. Size exclusion chromatography 2.4. Enzyme actiÕity Proteins Ž1 m M. were incubated for 150 minutes at room temperature. Transferase activity was assayed w10,11x in 0.5 ml incubation volumes for 3 minutes at 308C. The deaminase activity was determined by time drive assay at 308C in assay mix containing 0.1 M potassium phosphate ŽpH 7.3., 0.1 mM 5-formiminoH 4 PteGlu and 35 mM 2-mercaptoethanol, and the formation of the product 5,10methenylH 4 PteGlu was measured at A 355 on a Beckman DU640 spectrophotometer. All activity measurements are the average of assays done in duplicate or triplicate. 2.5. Fluorescence emission The intrinsic fluorescence of FTCDH 6 , FTH 6 and CDH 6 was measured using a Hitachi F3010 fluorescence spectrophotometer equipped with a thermostatically controlled sample holder at 258C. Samples Ž10 nM to 2 m M. were incubated for 150 minutes at room temperature in Buffer A containing the appro-
FTCDH 6 , FTH 6 and CDH 6 Ž1 mgrml. were incubated at room temperature for 150 minutes in Buffer A containing the appropriate concentration of urea Ž0 to 5 M.. 200 m l aliquots were injected onto a Superose 6 HR 10r30 column equilibrated in Buffer A containing the appropriate concentration of urea. Samples were chromatographed at a flow rate of 0.31 mlrmin, at 48C and the absorbance of the eluate was monitored at OD 280 . Ferritin Ž440 kDa., catalase Ž232 kDa., aldolase Ž158 kDa., bovine serum albumin Ž67 kDa., ovalbumin Ž43 kDa. , chymotrypsinogen Ž25 kDa. and ribonuclease A Ž13.7 kDa. were used as standards to construct a standard curve. The void volume was determined using Blue Dextran. BSA and Blue Dextran were chromatographed at each urea concentration and elution time remained constant at up to 5 M urea, suggesting that the permeation properties of the column do not change over this range of urea concentrations. Uversky previously demonstrated that the permeation properties of Superose 12 remain constant at up to 8 M urea w12x. Each sample was chromatographed 2–3 times.
226
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
3. Results and discussion 3.1. Spectral characterization of FTH6 , CDH6 , and FTCDH6 The intrinsic tryptophan fluorescence emission spectra of equimolar concentrations of FTH 6 , CDH 6 and FTCDH 6 are shown in Fig. 1. FTCDH 6 has a total of 4 tryptophans, 3 in the transferase domain ŽW158, W171 and W224. and one in the deaminase domain ŽW462.. Accordingly, the transferase domain exhibits significantly higher fluorescence emission intensity per mole of enzyme than does the deaminase domain. Native FTH 6 and CDH 6 have emission spectra with maximum emission wavelengths of approximately 342 nm ŽFig. 1A. and 326–328 nm ŽFig.
Fig. 1. Intrinsic fluorescence spectra of native and denatured FTH 6 ŽA., CDH 6 ŽB., and FTCDH 6 ŽC.. The emission spectra of 1 m M enzyme in Buffer A " urea. Ž1. Native enzyme, Ž2. enzyme in 5 M urea, Ž3. enzyme in 8 M urea, Ž4. enzyme renatured for 24 hours, and Ž5. equimolar mixture of native FTH 6 and CDH 6 Žpanel C only..
1B., respectively. The maximum emission wavelength for FTCDH 6 , at 338 nm ŽFig. 1C., is intermediate between those of the two domains. Similarly shaped spectra were observed over a range of 10 nM to 2 m M monomer concentration for all three protein species Ždata not shown.. The emission spectrum of an equimolar mixture of FTH 6 and CDH 6 exhibits slightly more maximal fluorescence intensity Ž1.07 times. than octameric FTCDH 6 . Upon denaturation in 8 M urea, all three species undergo a decrease in fluorescence intensity and their l max values are redshifted to between 354 and 357 nm, indicating that all tryptophan residues have become exposed to the solvent. Denaturation is essentially a reversible process since after 24 hours of renaturation all three proteins regain most Žgreater than 75%. of their native fluorescence intensity and activity. As has been observed for other multi-domain proteins, the isolated domains refold more efficiently than the full length enzyme w13,14x. The full length enzyme may be more susceptible to non-specific aggregation because of unpacked surfaces which remain exposed during the slow pairing of domains. Both polar and nonpolar molecules can quench the tryptophan fluorescence of FTH 6 and FTCDH 6 , suggesting that at least one of the tryptophans within the transferase domain is solvent exposed Ž data not shown.. In contrast, the single tryptophan in CDH 6 is not accessible to the polar quenchers Csq or Iy, but can be effectively quenched with acrylamide. This information combined with the l max value observed for CDH 6 suggests that W462 resides within a hydrophobic environment w15x, either buried within the deaminase domain or at a subunit interface. Far-UV CD spectra were obtained for the full length enzyme and both independent domains Ž Fig. 2.. All three species appear to contain significant amounts of a helical structure estimated using the K2D method Ž38% for FTH 6 , 60% for CDH 6 , 43% for FTCDH 6 . w16x. Therefore the circular dichroism signal at 222 nm was used to monitor the urea-induced unfolding of these proteins. The additive spectrum of the two domains and the spectrum of FTCDH 6 are similar at wavelengths above 200 nm. Both exhibit similar minima at 208 and 222 nm, suggesting that the isolated domains retain the helical structure present within the native structure. Because the dynode voltage increases at wavelengths below 200 nm
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
227
Žto approx. 500 volts in the 193–198 region., we are reluctant to interpret differences in the CD signals observed at these shorter wavelengths. 3.2. Urea-induced denaturation of FTH6 , CDH6 and FTCDH6 ; loss of transferase and deaminase actiÕities Since the transferase and deaminase activities exist in separate domains, catalytic activity serves as a domain-specific conformational probe. The isolated transferase domain is inactivated ŽFig. 3A. in a cooperative transition occurring between 1.25 and 2.5 M urea. The isolated deaminase domain becomes inactivated within a similar range of denaturant, between 0.75 and 2.5 M urea. The full length enzyme undergoes loss of transferase and deaminase activities at only slightly higher concentrations of urea than the independent domains, suggesting that both domains are only marginally stabilized when associated within the octamer.
Fig. 3. Percent original signal remaining during urea-induced denaturation of FTH 6 ŽA., CDH 6 ŽB. and FTCDH 6 ŽC.. Final concentrations were 1 m M Žfluorescence and activity. or 10 m M Žcircular dichroism. enzyme in Buffer A plus the appropriate amount of urea. Symbols: Ž`. percent original transferase activity; Žv . percent original deaminase activity; ŽI,B. percent relative fluorescences Ž F y F5 .rŽ F0 y F5 ., where F is the maximal peak height, F5 is the fluorescence in 5 M urea and F0 is the fluorescence in 0 M urea, Žn, '. percent original CD signal at 222 nm, Že,l. position of lmax in nm.
3.3. Changes in intrinsic fluorescence upon denaturation in urea
Fig. 2. Circular dichroic spectra of native FTCDH 6 and the isolated domains, in Buffer A. The molar ellipticity, in degrees cm2 dmoly1 , of: Ž- - -. FTH 6 , Ž PPP . CDH 6 Ž — ., FTCDH 6 Ž-PP-. the additive spectra of FTH 6 and CDH 6 .
As shown in Fig. 3, the fluorescence intensities of FTCDH 6 and the isolated domains are quenched in response to increasing urea. The loss of FTH 6 fluorescence intensity occurs simultaneously with inactivation ŽFig. 3A.. The wavelength of maximum emis-
228
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
sion is red-shifted between 2.5 and 4.5 M urea, suggesting increased solvent exposure of the tryptophan residues. Likewise, CDH 6 undergoes a fluorescence intensity loss which correlates closely with inactivation ŽFig. 3B. , and a large red-shift in lmax occurs between 2.8 and 5 M urea, as the single tryptophan becomes increasingly solvent accessible. When FTCDH 6 is subjected to urea denaturation, the fluorescence intensity loss correlates closely with the loss of the transferase activity ŽFig. 3C.. A gradual increase in lmax occurs between 2.5 and 4 M urea, followed by a steeper increase at 4–5 M urea. Thus both domains undergo conformational changes at low concentrations of urea, as demonstrated by inactivation and quenching of intrinsic tryptophan fluorescence. 3.4. Changes in secondary structure To determine if the conformational change indicated by loss of fluorescence intensity and catalytic function involves protein unfolding, far-UV circular dichroism was used to monitor possible loss of secondary structure ŽFig. 3.. The transferase domain undergoes a biphasic decrease in the a helix CD signal at 222 nm. At 2 M urea FTH 6 is inactive and has undergone a major decrease in fluorescence intensity, however more than 90% of the original a helix signal remains. Between 2 and 3 M urea this domain undergoes significant unfolding, as indicated by a 40% loss in the signal at 222 nm ŽFig. 3A.. Furthermore, the local minimum at 222 nm is lost between 2.5 and 3 M urea Ždata not shown. . A small plateau is observed in the CD profile at 3.5 to 4 M urea, which suggests that a partly unfolded transferase domain may be stable within this concentration range. Further unfolding occurs between 4.5 and 6 M urea. The deaminase domain exhibits a 75% decrease in a-helix signal between 2.5 and 5.5 M urea ŽFig. 3B., after loss of most of the fluorescence and activity has occurred. The midpoint of this transition, between 3.5 and 4 M urea, also correlates with the loss of the local minimum at 222 nm Ždata not shown.. At 5.5 M urea, very little of the original helix signal remains. FTCDH 6 also retains most of the helix signal at 222 nm up to 2 M urea Ž Fig. 3C. ; 35% of the signal is lost between 2 and 3.5 M urea, as is the local
minimum at 222 nm. A second decrease of almost 40% is observed between 5 and 7.5 M urea. Thus the isolated domains and the full length enzyme undergo fluorescence and activity changes at urea concentrations where 90% or more of the helical structure remains intact. This suggests that the initial conformational change is primarily limited to changes in tertiary structure. 3.5. Changes in quaternary structure In some oligomeric proteins inactivation andror loss of tertiary structure occur simultaneously with changes in quaternary structure w17,18x. Therefore we wanted to establish whether the inactivation of FTH 6 , CDH 6 and FTCDH 6 and loss of intrinsic fluorescence observed at 1–2.5 M urea correlated with subunit dissociation. Previously, Findlay and MacKenzie w5x used chemical cross-linking studies to monitor the urea-induced dissociation of porcine liver FTCD. Their studies indicated that the octameric enzyme undergoes coincident loss of both catalytic activities, a major decrease in intrinsic tryptophan fluorescence and dissociation to dimers in 2–3 M urea. The enzyme undergoes further dissociation to monomer at urea concentrations above 4 M. A dimeric, proteolytic-derived transferase active fragment was also observed to dissociate within the 2–3 M urea range, and thus it seemed plausible that the interface formed by this transferase active fragment was the same as the one initially lost within the octamer. Our cross-linking experiments with bis-succinimidyl suberate were consistent with these previous observations Ž data not shown.. Dissociation of native FTCDH 6 to dimers was observed between 1.5 and 2.5 M urea, and further dissociation to monomers between 3.5 and 5 M urea. FTH 6 loses the ability to be cross-linked between 2 and 2.5 M urea, suggesting that dissociation of the transferase dimer occurs within this range. However, conditions for efficient crosslinking of the deaminase dimer could not be obtained even after testing several other cross-linking reagents and conditions. Therefore, proteins were subjected to size exclusion chromatography Ž SEC. to detect volume changes which might signify subunit dissociation Ž an increase in time of elution. or protein unfolding Ža decrease in
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
time of elution.. Fig. 4A shows the chromatographic profiles of FTH 6 in 0 to 5 M urea. At 0 M urea, the transferase domain elutes at 58.2 min which is consistent with a molecular mass of 79 kDa and indicates that it behaves as a dimer. Very little change in either the elution position or profile of this peak is observed at urea concentrations below 2.5 M, suggesting that the inactivation and loss in fluorescence intensity occurring in the 1–2 M range result from a conformational change within the dimer, and not from subunit dissociation. At higher concentrations of urea, the transferase domain peak broadens, suggesting the
229
presence of larger, unfolded species. This change in elution behaviour coincides with unfolding of the domain as detected by far-UV CD and increased solvent exposure of the tryptophan residues. These experiments could not distinguish a specific urea concentration where the transferase dimer dissociates to folded monomer, and it is possible that unfolding and dissociation are coupled to produce an unfolded monomeric species. The denaturation of the deaminase domain was also monitored by SEC ŽFig. 4B.. In the absence of urea, CDH 6 exists as a dimer, eluting with a molecu-
Fig. 4. Size-exclusion chromatography profiles of FTH 6 ŽA., CDH 6 ŽB., and FTCDH 6 ŽC. following 2.5 hours of denaturation at room temperature in increasing concentrations of urea. SEC was performed at 48C using a Superose 6 column equilibrated in Buffer A containing the corresponding concentrations of urea.
230
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
lar mass of 53 kDa, approximately twice its polypeptide size. At 1.5 M urea, the peak becomes increasingly asymmetric with shoulders preceding and following the peak. This mixture may include smaller populations of unfolded monomer andror dimer and folded monomer in addition to the folded dimeric domain. However, the peak corresponding to the dimer remains predominant at 2 M urea, suggesting that the conformational change producing an inactive dimer with decreased intrinsic tryptophan fluorescence also precedes the dissociation of CDH 6 . Native FTCDH 6 behaves predominantly as an octamer with a smaller fraction existing as a smaller oligomer or perhaps an equilibrium mixture of a dimer and a larger species Žestimated molecular mass of 200 kDa. ŽFig. 4C.. Between 2 and 2.5 M urea, dissociation occurs at one type of subunit interface and the elution profile demonstrates an increased population of a smaller species. The loss of one type of subunit interface may be coupled to partial unfolding of the full-length enzyme since a 40% loss of the a helix signal is observed between 2 and 3.5 M urea, and SEC indicates that the population of larger, partly unfolded intermediates is substantial at 3.5 M urea. While cross-linking studies indicate that a second dissociation event occurs between 3.5 and 5 M urea, the elution profiles observed at higher urea concentrations are complicated by the presence of peaks at or near the void volume which represent larger, partially unfolded and possibly aggregated species, which may obscure or prevent the observation of a second dissociation event. 3.6. Domain stability Inactivation and quenching of intrinsic tryptophan fluorescence intensity occur as coincident, cooperative transitions at low urea concentrations, and prior to significant loss of secondary or quaternary structure. This suggests that both types of domain initially undergo a two state change in tertiary structure at low concentrations of urea. This transition can be described as F | F ) , where F represents the native state and F ) the inactive oligomeric species. To further characterize this conformational change, the equilibrium constant, K F ) , and the free energy, DG F ) , were calculated at different urea concentra-
Table 1 Analysis of urea induced denaturation Probe
DG H 2 O F ) mF ) Cm Žkcal moly1 . Žkcal moly1 ŽM urea. My1 .
FTH 6
transferase 4.54"0.40 fluorescence 4.52"0.29
2.46"0.22 2.48"0.16
1.85 1.82
CDH 6
deaminase 2.55"0.14 fluorescence 1.91"0.14
1.60"0.08 1.26"0.09
1.59 1.52
FTCDH 6 transferase 6.29"0.41 deaminase 4.31"0.28 fluorescence 4.25"0.23
3.29"0.21 2.93"0.19 2.24"0.11
1.91 1.47 1.90
G H 2 O F ) refers to the free energy of the observed conformational change in tertiary structure in the absence of denaturant and m F ) is the urea dependency of this free energy. G H 2 O F ) and m F ) values were obtained from the intercept and slope of the linear extrapolation model provided by Eq. Ž2.. The data in the transition region of the activity and fluorescence denaturation curves for FTH 6 , CDH 6 and FTCDH 6 shown in Fig. 3 were fit to the linear extrapolation model as described in Section 3.6. Errors represent the standard deviation from that fit. Cm represents the urea concentration at which half of the protein has undergone the conformational change. Cm s DG H 2 O F ) r m F ) .
tions within the transition region according to the following relationship: K F ) s F )rF s f F )r Ž 1 y f F ) . s exp Ž yDG F )rRT . Ž1. where f F ) is the fraction of protein which has undergone the conformational change. Data analysis was performed using the non-linear regression analysis program Enzfitter w w19x. To estimate the change in tertiary stability of the domains, the free energy of this transition in the absence of denaturant was estimated from a linear extrapolation of the values of DG F ) versus the denaturant concentration to 0 M urea, according to Pace w20x.
DG F ) s DG H 2 O F ) q m F ) urea
Ž2.
where the intercept DG H 2 O F ) corresponds to the free energy of this conformational change at 0 M urea and the slope m F ) reflects the cooperativity of the ureainduced transition. Comparison of DG H 2 O F ) values ŽTable 1. and transitional midpoints Ž Cm , Table 1. for both the
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
independent domains and FTCDH 6 demonstrate that the tertiary stabilities of the transferase and the deaminase domains are little improved by octamer formation.
4. Conclusions Whether expressed as an isolated domain or constrained within the octameric structure, the transferase and deaminase domains undergo conformational changes at a similar concentration of urea. These apparent changes in tertiary structure result in inactivation without dissociation of the subunit interfaces or significant loss of overall secondary structure. Inactivation often occurs at lower denaturant concentrations than are required to promote gross unfolding of an enzyme Žreviewed in Tsou, w21x., perhaps because the active site requires relatively more flexibility than the structural core of the enzyme. Since coincident inactivation and loss of intrinsic fluorescence is observed for both types of isolated domains, it is possible that tryptophan residueŽs. are located proximal to both the transferase and the deaminase active sites. While the octameric structure has only a limited effect on the tertiary stability of the domains, it may have a more profound effect on the stability of one type of subunit interface. Both domain dimers appear to undergo concurrent dissociation and unfolding at 2.5–3 M urea. This provides additional, albeit circumstantial, evidence that retention of the subunit interfaces plays an important role in maintaining the folded structure of both the transferase and deaminase domains. Within the FTCD octamer, one type of subunit interface is disrupted at a similar urea concentration as observed for both domain dimers, however the second type of subunit interface appears much more stable within the octamer and is maintained at higher concentrations of denaturant. Findlay and MacKenzie w5x previously suggested that the interface formed by two transferase domains is lost in the first dissociation event, partly because a proteolytically derived transferase-active fragment was observed to undergo dissociation between 2 and 3 M urea. Now that we have determined that the isolated deaminase domain also dissociates within this concentration range, it becomes difficult to differentiate
231
whether, within the octamer, the transferase or the deaminase domain interface is less stable to urea. While the domains show little indication of improved tertiary stability when linked to form the bifunctional enzyme, the covalent association of the domains to form the intact enzyme enhances the stability of one type of subunit interface within the octamer.
Acknowledgements This work was supported by grant MT 4479 from the Medical Research Council of Canada. We wish to thank Dr. J. Turnbull, Department of Chemistry, Concordia University, for access to the circular dichroimeter and secondary structure prediction programs and Dr. J. Turnbull, and Dr. J. Cromlish, Department of Biochemistry, McGill University, for careful reading and useful criticism of the manuscript.
References w1x MacKenzie, R.E. and Baugh, C.M. Ž1980. Biochim. Biophys. Acta 611, 187–195. w2x Paquin, J., Baugh, C.M. and MacKenzie, R.E. Ž1985. J. Biol. Chem. 260, 14925–14931. w3x Beaudet, R. and MacKenzie, R.E. Ž1976. Biochim. Biophys. Acta 453, 151–161. w4x MacKenzie, R.E., Aldridge, M. and Paquin, J. Ž1980. J. Biol. Chem. 255, 9474–9478. w5x Findlay, W.A. and MacKenzie, R.E. Ž1987. Biochemistry 26, 1948–1954. w6x Findlay, W.A. and MacKenzie, R.E. Ž1988. Biochemistry 27, 3404–3408. w7x Murley, L.L., Mejia, N.R. and MacKenzie, R.E. Ž1993. J. Biol. Chem. 268, 22820–22824. w8x Murley, L.L. and MacKenzie, R.E. Ž1995. Biochemistry 34, 10358–10364. w9x Gill, S.C. and Von Hippel, P.H. Ž1989. Anal. Biochem. 182, 319–326. w10x Drury, E.J., Bazar, L.S. and MacKenzie, R.E. Ž1975. Arch. Biochem. Biophys. 169, 662–668. w11x MacKenzie, R.E. Ž1980. Methods Enzymol. 66, 626–630. w12x Uversky, V.N. Ž1993. Biochemistry 32, 13288–13298. w13x Garel, J.-R. Ž1992. in Protein Folding ŽCreighton, T.E. ed.., pp. 405–454, Freeman, New York. w14x Jaenicke, R. Ž1987. Prog. Biophys. Mol. Biol. 49, 117–237. w15x Lakowicz, J.R. Ž1983. Principles of Fluorescence Spectroscopy, Plenum Press, New York. w16x Andrade, M.A., Chacon, P., Merelo, J.J. and Moran, F. Ž1993. Protein Eng. 6, 383–390.
232
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
w17x Aceto, A., Caccuri, A.M., Sacchetta, P., Bucciarelli, T., Dragani, B., Rosato, N., Federici, G. and Di Ilio, C. Ž1992. Biochem. J. 285, 241–245. w18x Abu-Soud, H.M., Loftus, M. and Stuehr, D.J. Ž1995. Biochemistry 34, 11167–11175.
w19x Leatherbarrow, R.J. Ž1987. Enzfitter w , BiosoftrElsevier, Cambridge, UK. w20x Pace, C.N. Ž1986. Methods Enzymol. 131, 266–280. w21x Tsou, C.-L. Ž1995. Biochim. Biophys. Acta 1253, 151–162.
Monofunctional domains of formiminotransferase-cyclodeaminase retain similar conformational stabilities outside the bifunctional octamer Laura Lea Murley, Robert E. MacKenzie
)
Department of Biochemistry, and Joint Centre for Structural Biology, 3655 Drummond St., Montreal, Que. H3G 1Y6, Canada Received 16 August 1996; revised 7 November 1996; accepted 11 November 1996
Abstract Each identical subunit of octameric formiminotransferase cyclodeaminase consists of a transferase and a deaminase domain connected by a short linker sequence. Both domains can be independently expressed in Escherichia coli as monofunctional dimers and show no indication of associating, suggesting that the linker mediates the only substantial interaction between the transferase and deaminase domains. To better understand the benefits arising from octamer formation, we have used equilibrium unfolding methods to examine the properties of the transferase and deaminase domains independently and within the octamer. Each isolated dimeric domain undergoes an apparent change in tertiary structure at low concentrations of urea Ž- 2 molrl. which results in the concurrent loss of intrinsic fluorescence and catalytic activity. The full length octameric enzyme also undergoes inactivation and a loss of intrinsic fluorescence over this concentration range, without apparent change in secondary or quaternary structure. Between 2 and 2.5 M urea the isolated transferase and deaminase domains dissociate to monomers. However, only one of the subunit interfaces in the octamer is disrupted at this urea concentration and dissociation of the second interface occurs between 3.5 and 5 M urea. While each domain shows similar stability to denaturation within and outside of the octamer, one type of subunit interface achieves increased stability within the full length enzyme. Keywords: Conformational stability; Subunit interface; Domain interaction; Equilibrium unfolding; Octamer; Bifunctional enzyme
1. Introduction
Abbreviations: FTCD, formiminotransferase cyclodeaminase; FTCDH 6 , hexahistidine-tagged formiminotransferase cyclodeaminase; FTH 6 , hexahistidine-tagged formiminotransferase; CDH 6 , hexahistidine-tagged cyclodeaminase; CD, circular dichroism; l max , wavelength of maximum fluorescence emission; SEC, size-exclusion chromatography; BSA, bovine serum albumin; Ni-NTA, nickel-chelated nitrilotriacetic acid matrix; H 4 PteGlu, tetrahydrofolate; DTT, dithiothreitol; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis. ) Corresponding author. Fax: q1 Ž514. 3987384; E-mail: [email protected]
The bifunctional enzyme formiminotransferase cyclodeaminase Ž FTCD. catalyzes two sequential reactions in the histidine degradation pathway. The transferase activity transfers a formimino group from the histidine catabolite form im inoglutam ate to H 4 PteGlu n . The cyclodeaminase catalyzes the deamination of the formimino intermediate producing 5,10-methenylH 4 PteGlu n and NHq 4 . With monoglutamylated folates Ž n s 1., the two activities function independently, however, polyglutamylated intermediate Ž n G 4. can be efficiently channelled from the
0167-4838r97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 6 7 - 4 8 3 8 Ž 9 6 . 0 0 2 0 8 - 7
224
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
transferase to the deaminase active site w1,2x. FTCD preferentially channels pentaglutamates and MacKenzie and coworkers w1,2x have proposed that the polyglutamate tail acts as a noncovalently bound swinging arm which anchors the intermediate to the enzyme while allowing the pteroyl moiety to move between the two active sites. FTCD is composed of 8 identical subunits arranged to form a planar, ring-shaped tetramer of dimers w3,4x. Denaturationrrenaturation w5,6x and deletion analysis studies w7x indicate that the octamer is the smallest functional unit of FTCD to retain both catalytic activities and the ability to channel substrates. The FTCD octamer contains 4 high-affinity polyglutamate binding sites w2x, and Findlay and MacKenzie w5,6x have proposed that the integrity of alternating subunit interfaces within the native octamer is a prerequisite for expression of both catalytic activities. While this suggests that the active sites may be situated at the subunit interfaces, exact numbers of catalytic sites and their relative location within the octamer are not known. We have recently determined that each subunit of FTCD consists of an N-terminal transferase domain and a C-terminal deaminase domain separated by a short linker sequence w8x. The polyglutamate binding site maps to the deaminase domain. These domains can be independently expressed in E. coli and show no indication of interacting with each other once purified. This suggests that the linker mediates the only substantial interaction between the transferase and deaminase domains within each subunit of FTCD. Both domains contain sequences governing subunit association, and exist as monofunctional dimers in solution w8x. It is possible that the octameric structure has evolved from the fusion of two different dimeric monofunctional proteins, such that the subunit interfaces formed by identical domains are necessary for the maintenance of domain structure and function, but the hetero-domain interaction within the subunit is relatively unimportant. If this model were accurate, the two types of domains should behave similarly within and outside of the octamer. The ability of FTCD to channel substrate and the 50% decrease in k cat values observed for each isolated domain w8x argue that the transferase and deaminase domains do not function completely autonomously of each other within the octamer.
Substrate channelling is one of the more obvious benefits of octamer formation. We would like to establish if fusion of the transferase and deaminase domains might result in other advantages, such as increased stability of the domains or subunit interfaces. In this paper we examine the urea-induced denaturation of the transferase and deaminase domains independently and as part of the FTCD octamer.
2. Materials and methods 2.1. Materials Folic acid and formimino-L-glutamic acid were from Sigma. Urea solutions ŽUltra Pure urea, ICN. were prepared immediately prior to each experiment to minimize the formation of isocyanate. Ni-NTA resin was purchased from Qiagen. The Superose 6 HR 10r30 column was from Pharmacia. 5-formiminoH 4 PteGlu was prepared enzymatically as described previously w2x except that purified recombinant transferase domain w8x was used to catalyze its synthesis. Acidified 5-formiminoH 4 PteGlu was centrifuged in a microfuge for 10 min at 48C to pellet precipitated enzyme, instead of passing the solution through Centriflo cones ŽAmicon.. 2.2. Enzyme purification Recombinant histidine-tagged FTCDH 6 , FTH 6 and CDH 6 were expressed in E. coli and purified as previously described w8x. Enzyme preparations had specific transferase activities of approx. 30 m molrmin per mg for FTCDH 6 , and 22–24 m molrmin per mg FTH 6 , and specific deaminase activities of 18–21 m molrmin per mg for FTCDH 6 and 16–20 m molrmin per mg for CDH 6 when assayed with 0.1 mM 5-formiminoH 4 PteGlu. Single bands were observed upon SDS PAGE. Triton X 100 was replaced with Tween 20 by rebinding the enzyme to Ni-NTA resin and washing and eluting with buffers containing 0.05% Tween 20 in place of Triton X 100. Enzyme was concentrated to approx. 4 mgrml ŽCentriprep, Amicon. and stored at y208C in buffer containing 0.1 M potassium phosphate ŽpH 7.8., 0.05% Tween 20, 40% glycerol and 1 mM DTT.
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
Protein concentrations Ž as monomer. were determined using the calculated extinction coefficients w9x E280 s 37 920 My1 cmy1 for FTCDH 6 , 28 150 My1 cmy1 for FTH 6 and 9770 My1 cmy1 for CDH 6 . 2.3. Unfolding studies Denaturation experiments were performed in Buffer A Ž0.1 M potassium phosphate, pH 7.3, 0.05% Tween 20 and 1 mM DTT. containing the appropriate concentration of urea Ž0 to 8 M. . Initial timecourse experiments indicated that recombinant FTCDH 6 and the isolated domains undergo almost all of their fluorescence and activity loss within the first 100 minutes of incubation at room temperature with a much slower loss thereafter. This is consistent with a previous observation that FTCD purified from pig liver undergoes significant inactivation within the first 2 hours of urea-induced denaturation w5x. In the following experiments, proteins were incubated for 150 minutes at room temperature prior to analysis, a time sufficient to achieve equilibrium, as monitored by intrinsic fluorescence and enzyme activity.
225
priate amount of urea Ž0 to 8 M. , and the emission spectra Žexcitation at 290 nm. were recorded between 320 and 380 nm. The excitation bandpass was set at 3 nm and the emission bandpass at 5 nm. The spectrum of a buffer blank was subtracted from each sample spectrum.
2.6. Circular dichroism The far-UV CD spectra of samples containing 10 m M protein, incubated for 150 minutes at room temperature in Buffer A containing urea Ž0 to 7.5 M., were recorded in a Jasco J-710 model with a cylindrical cuvette of 0.05 cm path length. Scan speed was set at 100 nmrmin and measurements are reported from 5 accumulations. The spectrum of a buffer blank was subtracted from each sample spectrum and molar ellipticity was determined. Dynode voltages at 222 nm remained at or below 300 volts, even at the higher urea concentrations.
2.7. Size exclusion chromatography 2.4. Enzyme actiÕity Proteins Ž1 m M. were incubated for 150 minutes at room temperature. Transferase activity was assayed w10,11x in 0.5 ml incubation volumes for 3 minutes at 308C. The deaminase activity was determined by time drive assay at 308C in assay mix containing 0.1 M potassium phosphate ŽpH 7.3., 0.1 mM 5-formiminoH 4 PteGlu and 35 mM 2-mercaptoethanol, and the formation of the product 5,10methenylH 4 PteGlu was measured at A 355 on a Beckman DU640 spectrophotometer. All activity measurements are the average of assays done in duplicate or triplicate. 2.5. Fluorescence emission The intrinsic fluorescence of FTCDH 6 , FTH 6 and CDH 6 was measured using a Hitachi F3010 fluorescence spectrophotometer equipped with a thermostatically controlled sample holder at 258C. Samples Ž10 nM to 2 m M. were incubated for 150 minutes at room temperature in Buffer A containing the appro-
FTCDH 6 , FTH 6 and CDH 6 Ž1 mgrml. were incubated at room temperature for 150 minutes in Buffer A containing the appropriate concentration of urea Ž0 to 5 M.. 200 m l aliquots were injected onto a Superose 6 HR 10r30 column equilibrated in Buffer A containing the appropriate concentration of urea. Samples were chromatographed at a flow rate of 0.31 mlrmin, at 48C and the absorbance of the eluate was monitored at OD 280 . Ferritin Ž440 kDa., catalase Ž232 kDa., aldolase Ž158 kDa., bovine serum albumin Ž67 kDa., ovalbumin Ž43 kDa. , chymotrypsinogen Ž25 kDa. and ribonuclease A Ž13.7 kDa. were used as standards to construct a standard curve. The void volume was determined using Blue Dextran. BSA and Blue Dextran were chromatographed at each urea concentration and elution time remained constant at up to 5 M urea, suggesting that the permeation properties of the column do not change over this range of urea concentrations. Uversky previously demonstrated that the permeation properties of Superose 12 remain constant at up to 8 M urea w12x. Each sample was chromatographed 2–3 times.
226
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
3. Results and discussion 3.1. Spectral characterization of FTH6 , CDH6 , and FTCDH6 The intrinsic tryptophan fluorescence emission spectra of equimolar concentrations of FTH 6 , CDH 6 and FTCDH 6 are shown in Fig. 1. FTCDH 6 has a total of 4 tryptophans, 3 in the transferase domain ŽW158, W171 and W224. and one in the deaminase domain ŽW462.. Accordingly, the transferase domain exhibits significantly higher fluorescence emission intensity per mole of enzyme than does the deaminase domain. Native FTH 6 and CDH 6 have emission spectra with maximum emission wavelengths of approximately 342 nm ŽFig. 1A. and 326–328 nm ŽFig.
Fig. 1. Intrinsic fluorescence spectra of native and denatured FTH 6 ŽA., CDH 6 ŽB., and FTCDH 6 ŽC.. The emission spectra of 1 m M enzyme in Buffer A " urea. Ž1. Native enzyme, Ž2. enzyme in 5 M urea, Ž3. enzyme in 8 M urea, Ž4. enzyme renatured for 24 hours, and Ž5. equimolar mixture of native FTH 6 and CDH 6 Žpanel C only..
1B., respectively. The maximum emission wavelength for FTCDH 6 , at 338 nm ŽFig. 1C., is intermediate between those of the two domains. Similarly shaped spectra were observed over a range of 10 nM to 2 m M monomer concentration for all three protein species Ždata not shown.. The emission spectrum of an equimolar mixture of FTH 6 and CDH 6 exhibits slightly more maximal fluorescence intensity Ž1.07 times. than octameric FTCDH 6 . Upon denaturation in 8 M urea, all three species undergo a decrease in fluorescence intensity and their l max values are redshifted to between 354 and 357 nm, indicating that all tryptophan residues have become exposed to the solvent. Denaturation is essentially a reversible process since after 24 hours of renaturation all three proteins regain most Žgreater than 75%. of their native fluorescence intensity and activity. As has been observed for other multi-domain proteins, the isolated domains refold more efficiently than the full length enzyme w13,14x. The full length enzyme may be more susceptible to non-specific aggregation because of unpacked surfaces which remain exposed during the slow pairing of domains. Both polar and nonpolar molecules can quench the tryptophan fluorescence of FTH 6 and FTCDH 6 , suggesting that at least one of the tryptophans within the transferase domain is solvent exposed Ž data not shown.. In contrast, the single tryptophan in CDH 6 is not accessible to the polar quenchers Csq or Iy, but can be effectively quenched with acrylamide. This information combined with the l max value observed for CDH 6 suggests that W462 resides within a hydrophobic environment w15x, either buried within the deaminase domain or at a subunit interface. Far-UV CD spectra were obtained for the full length enzyme and both independent domains Ž Fig. 2.. All three species appear to contain significant amounts of a helical structure estimated using the K2D method Ž38% for FTH 6 , 60% for CDH 6 , 43% for FTCDH 6 . w16x. Therefore the circular dichroism signal at 222 nm was used to monitor the urea-induced unfolding of these proteins. The additive spectrum of the two domains and the spectrum of FTCDH 6 are similar at wavelengths above 200 nm. Both exhibit similar minima at 208 and 222 nm, suggesting that the isolated domains retain the helical structure present within the native structure. Because the dynode voltage increases at wavelengths below 200 nm
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
227
Žto approx. 500 volts in the 193–198 region., we are reluctant to interpret differences in the CD signals observed at these shorter wavelengths. 3.2. Urea-induced denaturation of FTH6 , CDH6 and FTCDH6 ; loss of transferase and deaminase actiÕities Since the transferase and deaminase activities exist in separate domains, catalytic activity serves as a domain-specific conformational probe. The isolated transferase domain is inactivated ŽFig. 3A. in a cooperative transition occurring between 1.25 and 2.5 M urea. The isolated deaminase domain becomes inactivated within a similar range of denaturant, between 0.75 and 2.5 M urea. The full length enzyme undergoes loss of transferase and deaminase activities at only slightly higher concentrations of urea than the independent domains, suggesting that both domains are only marginally stabilized when associated within the octamer.
Fig. 3. Percent original signal remaining during urea-induced denaturation of FTH 6 ŽA., CDH 6 ŽB. and FTCDH 6 ŽC.. Final concentrations were 1 m M Žfluorescence and activity. or 10 m M Žcircular dichroism. enzyme in Buffer A plus the appropriate amount of urea. Symbols: Ž`. percent original transferase activity; Žv . percent original deaminase activity; ŽI,B. percent relative fluorescences Ž F y F5 .rŽ F0 y F5 ., where F is the maximal peak height, F5 is the fluorescence in 5 M urea and F0 is the fluorescence in 0 M urea, Žn, '. percent original CD signal at 222 nm, Že,l. position of lmax in nm.
3.3. Changes in intrinsic fluorescence upon denaturation in urea
Fig. 2. Circular dichroic spectra of native FTCDH 6 and the isolated domains, in Buffer A. The molar ellipticity, in degrees cm2 dmoly1 , of: Ž- - -. FTH 6 , Ž PPP . CDH 6 Ž — ., FTCDH 6 Ž-PP-. the additive spectra of FTH 6 and CDH 6 .
As shown in Fig. 3, the fluorescence intensities of FTCDH 6 and the isolated domains are quenched in response to increasing urea. The loss of FTH 6 fluorescence intensity occurs simultaneously with inactivation ŽFig. 3A.. The wavelength of maximum emis-
228
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
sion is red-shifted between 2.5 and 4.5 M urea, suggesting increased solvent exposure of the tryptophan residues. Likewise, CDH 6 undergoes a fluorescence intensity loss which correlates closely with inactivation ŽFig. 3B. , and a large red-shift in lmax occurs between 2.8 and 5 M urea, as the single tryptophan becomes increasingly solvent accessible. When FTCDH 6 is subjected to urea denaturation, the fluorescence intensity loss correlates closely with the loss of the transferase activity ŽFig. 3C.. A gradual increase in lmax occurs between 2.5 and 4 M urea, followed by a steeper increase at 4–5 M urea. Thus both domains undergo conformational changes at low concentrations of urea, as demonstrated by inactivation and quenching of intrinsic tryptophan fluorescence. 3.4. Changes in secondary structure To determine if the conformational change indicated by loss of fluorescence intensity and catalytic function involves protein unfolding, far-UV circular dichroism was used to monitor possible loss of secondary structure ŽFig. 3.. The transferase domain undergoes a biphasic decrease in the a helix CD signal at 222 nm. At 2 M urea FTH 6 is inactive and has undergone a major decrease in fluorescence intensity, however more than 90% of the original a helix signal remains. Between 2 and 3 M urea this domain undergoes significant unfolding, as indicated by a 40% loss in the signal at 222 nm ŽFig. 3A.. Furthermore, the local minimum at 222 nm is lost between 2.5 and 3 M urea Ždata not shown. . A small plateau is observed in the CD profile at 3.5 to 4 M urea, which suggests that a partly unfolded transferase domain may be stable within this concentration range. Further unfolding occurs between 4.5 and 6 M urea. The deaminase domain exhibits a 75% decrease in a-helix signal between 2.5 and 5.5 M urea ŽFig. 3B., after loss of most of the fluorescence and activity has occurred. The midpoint of this transition, between 3.5 and 4 M urea, also correlates with the loss of the local minimum at 222 nm Ždata not shown.. At 5.5 M urea, very little of the original helix signal remains. FTCDH 6 also retains most of the helix signal at 222 nm up to 2 M urea Ž Fig. 3C. ; 35% of the signal is lost between 2 and 3.5 M urea, as is the local
minimum at 222 nm. A second decrease of almost 40% is observed between 5 and 7.5 M urea. Thus the isolated domains and the full length enzyme undergo fluorescence and activity changes at urea concentrations where 90% or more of the helical structure remains intact. This suggests that the initial conformational change is primarily limited to changes in tertiary structure. 3.5. Changes in quaternary structure In some oligomeric proteins inactivation andror loss of tertiary structure occur simultaneously with changes in quaternary structure w17,18x. Therefore we wanted to establish whether the inactivation of FTH 6 , CDH 6 and FTCDH 6 and loss of intrinsic fluorescence observed at 1–2.5 M urea correlated with subunit dissociation. Previously, Findlay and MacKenzie w5x used chemical cross-linking studies to monitor the urea-induced dissociation of porcine liver FTCD. Their studies indicated that the octameric enzyme undergoes coincident loss of both catalytic activities, a major decrease in intrinsic tryptophan fluorescence and dissociation to dimers in 2–3 M urea. The enzyme undergoes further dissociation to monomer at urea concentrations above 4 M. A dimeric, proteolytic-derived transferase active fragment was also observed to dissociate within the 2–3 M urea range, and thus it seemed plausible that the interface formed by this transferase active fragment was the same as the one initially lost within the octamer. Our cross-linking experiments with bis-succinimidyl suberate were consistent with these previous observations Ž data not shown.. Dissociation of native FTCDH 6 to dimers was observed between 1.5 and 2.5 M urea, and further dissociation to monomers between 3.5 and 5 M urea. FTH 6 loses the ability to be cross-linked between 2 and 2.5 M urea, suggesting that dissociation of the transferase dimer occurs within this range. However, conditions for efficient crosslinking of the deaminase dimer could not be obtained even after testing several other cross-linking reagents and conditions. Therefore, proteins were subjected to size exclusion chromatography Ž SEC. to detect volume changes which might signify subunit dissociation Ž an increase in time of elution. or protein unfolding Ža decrease in
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
time of elution.. Fig. 4A shows the chromatographic profiles of FTH 6 in 0 to 5 M urea. At 0 M urea, the transferase domain elutes at 58.2 min which is consistent with a molecular mass of 79 kDa and indicates that it behaves as a dimer. Very little change in either the elution position or profile of this peak is observed at urea concentrations below 2.5 M, suggesting that the inactivation and loss in fluorescence intensity occurring in the 1–2 M range result from a conformational change within the dimer, and not from subunit dissociation. At higher concentrations of urea, the transferase domain peak broadens, suggesting the
229
presence of larger, unfolded species. This change in elution behaviour coincides with unfolding of the domain as detected by far-UV CD and increased solvent exposure of the tryptophan residues. These experiments could not distinguish a specific urea concentration where the transferase dimer dissociates to folded monomer, and it is possible that unfolding and dissociation are coupled to produce an unfolded monomeric species. The denaturation of the deaminase domain was also monitored by SEC ŽFig. 4B.. In the absence of urea, CDH 6 exists as a dimer, eluting with a molecu-
Fig. 4. Size-exclusion chromatography profiles of FTH 6 ŽA., CDH 6 ŽB., and FTCDH 6 ŽC. following 2.5 hours of denaturation at room temperature in increasing concentrations of urea. SEC was performed at 48C using a Superose 6 column equilibrated in Buffer A containing the corresponding concentrations of urea.
230
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
lar mass of 53 kDa, approximately twice its polypeptide size. At 1.5 M urea, the peak becomes increasingly asymmetric with shoulders preceding and following the peak. This mixture may include smaller populations of unfolded monomer andror dimer and folded monomer in addition to the folded dimeric domain. However, the peak corresponding to the dimer remains predominant at 2 M urea, suggesting that the conformational change producing an inactive dimer with decreased intrinsic tryptophan fluorescence also precedes the dissociation of CDH 6 . Native FTCDH 6 behaves predominantly as an octamer with a smaller fraction existing as a smaller oligomer or perhaps an equilibrium mixture of a dimer and a larger species Žestimated molecular mass of 200 kDa. ŽFig. 4C.. Between 2 and 2.5 M urea, dissociation occurs at one type of subunit interface and the elution profile demonstrates an increased population of a smaller species. The loss of one type of subunit interface may be coupled to partial unfolding of the full-length enzyme since a 40% loss of the a helix signal is observed between 2 and 3.5 M urea, and SEC indicates that the population of larger, partly unfolded intermediates is substantial at 3.5 M urea. While cross-linking studies indicate that a second dissociation event occurs between 3.5 and 5 M urea, the elution profiles observed at higher urea concentrations are complicated by the presence of peaks at or near the void volume which represent larger, partially unfolded and possibly aggregated species, which may obscure or prevent the observation of a second dissociation event. 3.6. Domain stability Inactivation and quenching of intrinsic tryptophan fluorescence intensity occur as coincident, cooperative transitions at low urea concentrations, and prior to significant loss of secondary or quaternary structure. This suggests that both types of domain initially undergo a two state change in tertiary structure at low concentrations of urea. This transition can be described as F | F ) , where F represents the native state and F ) the inactive oligomeric species. To further characterize this conformational change, the equilibrium constant, K F ) , and the free energy, DG F ) , were calculated at different urea concentra-
Table 1 Analysis of urea induced denaturation Probe
DG H 2 O F ) mF ) Cm Žkcal moly1 . Žkcal moly1 ŽM urea. My1 .
FTH 6
transferase 4.54"0.40 fluorescence 4.52"0.29
2.46"0.22 2.48"0.16
1.85 1.82
CDH 6
deaminase 2.55"0.14 fluorescence 1.91"0.14
1.60"0.08 1.26"0.09
1.59 1.52
FTCDH 6 transferase 6.29"0.41 deaminase 4.31"0.28 fluorescence 4.25"0.23
3.29"0.21 2.93"0.19 2.24"0.11
1.91 1.47 1.90
G H 2 O F ) refers to the free energy of the observed conformational change in tertiary structure in the absence of denaturant and m F ) is the urea dependency of this free energy. G H 2 O F ) and m F ) values were obtained from the intercept and slope of the linear extrapolation model provided by Eq. Ž2.. The data in the transition region of the activity and fluorescence denaturation curves for FTH 6 , CDH 6 and FTCDH 6 shown in Fig. 3 were fit to the linear extrapolation model as described in Section 3.6. Errors represent the standard deviation from that fit. Cm represents the urea concentration at which half of the protein has undergone the conformational change. Cm s DG H 2 O F ) r m F ) .
tions within the transition region according to the following relationship: K F ) s F )rF s f F )r Ž 1 y f F ) . s exp Ž yDG F )rRT . Ž1. where f F ) is the fraction of protein which has undergone the conformational change. Data analysis was performed using the non-linear regression analysis program Enzfitter w w19x. To estimate the change in tertiary stability of the domains, the free energy of this transition in the absence of denaturant was estimated from a linear extrapolation of the values of DG F ) versus the denaturant concentration to 0 M urea, according to Pace w20x.
DG F ) s DG H 2 O F ) q m F ) urea
Ž2.
where the intercept DG H 2 O F ) corresponds to the free energy of this conformational change at 0 M urea and the slope m F ) reflects the cooperativity of the ureainduced transition. Comparison of DG H 2 O F ) values ŽTable 1. and transitional midpoints Ž Cm , Table 1. for both the
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
independent domains and FTCDH 6 demonstrate that the tertiary stabilities of the transferase and the deaminase domains are little improved by octamer formation.
4. Conclusions Whether expressed as an isolated domain or constrained within the octameric structure, the transferase and deaminase domains undergo conformational changes at a similar concentration of urea. These apparent changes in tertiary structure result in inactivation without dissociation of the subunit interfaces or significant loss of overall secondary structure. Inactivation often occurs at lower denaturant concentrations than are required to promote gross unfolding of an enzyme Žreviewed in Tsou, w21x., perhaps because the active site requires relatively more flexibility than the structural core of the enzyme. Since coincident inactivation and loss of intrinsic fluorescence is observed for both types of isolated domains, it is possible that tryptophan residueŽs. are located proximal to both the transferase and the deaminase active sites. While the octameric structure has only a limited effect on the tertiary stability of the domains, it may have a more profound effect on the stability of one type of subunit interface. Both domain dimers appear to undergo concurrent dissociation and unfolding at 2.5–3 M urea. This provides additional, albeit circumstantial, evidence that retention of the subunit interfaces plays an important role in maintaining the folded structure of both the transferase and deaminase domains. Within the FTCD octamer, one type of subunit interface is disrupted at a similar urea concentration as observed for both domain dimers, however the second type of subunit interface appears much more stable within the octamer and is maintained at higher concentrations of denaturant. Findlay and MacKenzie w5x previously suggested that the interface formed by two transferase domains is lost in the first dissociation event, partly because a proteolytically derived transferase-active fragment was observed to undergo dissociation between 2 and 3 M urea. Now that we have determined that the isolated deaminase domain also dissociates within this concentration range, it becomes difficult to differentiate
231
whether, within the octamer, the transferase or the deaminase domain interface is less stable to urea. While the domains show little indication of improved tertiary stability when linked to form the bifunctional enzyme, the covalent association of the domains to form the intact enzyme enhances the stability of one type of subunit interface within the octamer.
Acknowledgements This work was supported by grant MT 4479 from the Medical Research Council of Canada. We wish to thank Dr. J. Turnbull, Department of Chemistry, Concordia University, for access to the circular dichroimeter and secondary structure prediction programs and Dr. J. Turnbull, and Dr. J. Cromlish, Department of Biochemistry, McGill University, for careful reading and useful criticism of the manuscript.
References w1x MacKenzie, R.E. and Baugh, C.M. Ž1980. Biochim. Biophys. Acta 611, 187–195. w2x Paquin, J., Baugh, C.M. and MacKenzie, R.E. Ž1985. J. Biol. Chem. 260, 14925–14931. w3x Beaudet, R. and MacKenzie, R.E. Ž1976. Biochim. Biophys. Acta 453, 151–161. w4x MacKenzie, R.E., Aldridge, M. and Paquin, J. Ž1980. J. Biol. Chem. 255, 9474–9478. w5x Findlay, W.A. and MacKenzie, R.E. Ž1987. Biochemistry 26, 1948–1954. w6x Findlay, W.A. and MacKenzie, R.E. Ž1988. Biochemistry 27, 3404–3408. w7x Murley, L.L., Mejia, N.R. and MacKenzie, R.E. Ž1993. J. Biol. Chem. 268, 22820–22824. w8x Murley, L.L. and MacKenzie, R.E. Ž1995. Biochemistry 34, 10358–10364. w9x Gill, S.C. and Von Hippel, P.H. Ž1989. Anal. Biochem. 182, 319–326. w10x Drury, E.J., Bazar, L.S. and MacKenzie, R.E. Ž1975. Arch. Biochem. Biophys. 169, 662–668. w11x MacKenzie, R.E. Ž1980. Methods Enzymol. 66, 626–630. w12x Uversky, V.N. Ž1993. Biochemistry 32, 13288–13298. w13x Garel, J.-R. Ž1992. in Protein Folding ŽCreighton, T.E. ed.., pp. 405–454, Freeman, New York. w14x Jaenicke, R. Ž1987. Prog. Biophys. Mol. Biol. 49, 117–237. w15x Lakowicz, J.R. Ž1983. Principles of Fluorescence Spectroscopy, Plenum Press, New York. w16x Andrade, M.A., Chacon, P., Merelo, J.J. and Moran, F. Ž1993. Protein Eng. 6, 383–390.
232
L.L. Murley, R.E. MacKenzier Biochimica et Biophysica Acta 1338 (1997) 223–232
w17x Aceto, A., Caccuri, A.M., Sacchetta, P., Bucciarelli, T., Dragani, B., Rosato, N., Federici, G. and Di Ilio, C. Ž1992. Biochem. J. 285, 241–245. w18x Abu-Soud, H.M., Loftus, M. and Stuehr, D.J. Ž1995. Biochemistry 34, 11167–11175.
w19x Leatherbarrow, R.J. Ž1987. Enzfitter w , BiosoftrElsevier, Cambridge, UK. w20x Pace, C.N. Ž1986. Methods Enzymol. 131, 266–280. w21x Tsou, C.-L. Ž1995. Biochim. Biophys. Acta 1253, 151–162.