Arts & Architecture

A novel pathway to enzyme deactivation: The cutinase model

Description
A novel pathway to enzyme deactivation: The cutinase model
Published
of 7
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  A Novel Pathway to EnzymeDeactivation: The Cutinase Model R. P. Baptista, 1 L. Y. Chen, 1 A. Paixa˜o, 2 J. M. S. Cabral, 1 E. P. Melo 1,3 1 Centro de Engenharia Biolo´gica e Quı ´mica, Instituto Superior Te´cnico, Av.Rovisco Pais, Lisboa, Portugal; telephone: 351-218419068; fax: 351-218419062; e-mail: emelo@ualg.pt  2  Instituto Superior de Engenharia de Lisboa, Instituto Polite´cnico de Lisboa, Portugal  3  Universidade do Algarve, FERN, Campus de Gambelas, Faro, Portugal  Received 13 September 2002; accepted 3 December 2002 DOI: 10.1002/bit.10641 Abstract:  Cutinase in aqueous solution at pH 4.5 deacti-vates following a parallel pathway. At 53°C, 88% of thecutinase molecules are in the unfolded conformation,which can aggregate with a reaction order of 3 if theprotein concentration is high (  12 µ M  ). The aggregatesshow a sixfold increase in size as determined by dynamiclight scattering. This aggregation process is the firstphase observed during a deactivation experiment; how-ever, after significant cutinase depletion and maturationof the aggregates, a first-order step starts to dominateand a second phase independent of the protein concen-tration is observed. Kinetic partitioning between aggre-gation and first-order irreversible changes of the un-folded conformation can occur during enzyme deactiva-tion when the equilibrium between the native and theunfolded conformation is shifted and kept toward theunfolded conformation.  © 2003 Wiley Periodicals, Inc.  Bio- technol Bioeng   82:  851–857, 2003. Keywords:  enzyme deactivation; cutinase stability; par-allel pathway; protein aggregation INTRODUCTION An improved knowledge of enzyme deactivation pathwayswould significantly enhance the yield of some processes andprovide valuable physical insight into the structure–functionrelationships of enzymes. Enzyme deactivation (also calledlong-term stability) is usually studied by activity measure-ments. The enzyme is typically incubated at a fixed tem-perature and the time-dependent loss of activity is measuredafter cooling the sample. The kinetics of irreversiblechanges, such as aggregation, misfolding, and adverse co-valent reactions (Volkin and Klibanov, 1989), are thus mea-sured and information on the irreversible pathways can begained based on the nature of the deactivation profile. Veryoften irreversible deactivation is thought to be an isolatedprocess and not as proceeding from a thermodynamic equi-librium between the native and the reversibly unfolded con-formations of the enzyme (N ↔ U). Formally, this is notcorrect, but in some cases an apparent value for the rateconstant is the only consequence (Baptista et al., 2000; Zaleand Klibanov, 1983). The first-order characteristic of thedeactivation in the classical two-step pathway correlatingthe reversible and the irreversible steps (N ↔ U → I) is notaffected.Non-first-order deactivation profiles are less common,except perhaps for systems such as immobilized enzymes,and usually require the assumption of more complex path-ways to fit the data. Indeed, a series-type pathway has al-ready been used to broadly classify non-first-order enzymedeactivations (Sadana, 1991). This mechanism is conceptu-ally simple, involving three possible enzyme states, all of which may be active (  E  →  E  1 →  E  2 ), and is rather flexiblebecause it is able to categorize most, if not all, deactivationprofiles presented in the literature. In contrast, parallel-typepathways have been used rather infrequently to describe thecomplex kinetics of enzyme deactivations, mainly for tworeasons. First, if each parallel step is a single transition to anirreversible deactivated form of the enzyme, then experi-mental kinetics are of the first order and thus indistinguish-able from a single-step pathway. Second, non-first-orderkinetics can be considered a parallel pathway only if het-erogeneity exists in the enzyme sample (Sadana, 1991). Theactive enzyme state exhibits heterogeneity either in the spe-cific activity or in the activation energy for deactivationleading to parallel steps. Macroheterogeneity (the heteroge-neity expressed by isoenzymes or by structurally similarenzymes that are easily separable by adequate purification)is not evoked for studies using electrophoretically homoge-neous samples. Microheterogeneity (refering to the propertyof certain highly purified proteins for showing molecularheterogeneity when screened by high-resolution immuno-chemical, biochemical, or biophysical techniques) seems tohold for very specific cases. Intramolecular disulfide inter-change, limited enzymatic nicking, or de-amidation of a fewasparagine and glutamine residues before or during purifi-cation are examples of microheterogeneity.This work proposes a parallel pathway for enzyme deac-tivation based on two competing steps with different ordersof reaction. This pathway was shown to describe cutinasedeactivation in aqueous solution at pH 4.5. Cutinase is an Correspondence to:  E. MeloContract grant sponsor: Fundac¸a˜o para a Cieˆncia e Tecnologia (FCT)Contract grant numbers: SFRH/BD/2821/2000; PRAXIS/BIO/14314/98 © 2003 Wiley Periodicals, Inc.  enzyme displaying lipolytic activity (E.C.3.1.1.3) with amolecular weight of around 22 kDa and an isoelectric pointof 7.6 (Koops et al., 1999). Evidence supporting this path-way has come from analyses of deactivation profiles, in-cluding the dependence on enzyme concentration, and fromprotein size determinations by dynamic light scattering. MATERIALS AND METHODSChemicals and Cutinase Production Cutinase from  Fusarium solani  was cloned and expressed in  Escherichia coli  WK-6 strain, a kind gift of Corvas Inter-national (Ghent, Belgium). The recombinant enzyme wasproduced and purified to a lyophilized powder of >95%purity (w/w) using an osmotic shock to disrupt cells, fol-lowed by an acid precipitation and two anionic exchangerchromatographic steps (DEAE-Sephacel and Q-Sepharose)(Carvalho et al., 1999). Cutinase solutions at pH 4.5 wereprepared in 0.1  M   acetate buffer unless otherwise stated andprotein concentration was measured spectrophotometrically(  296nm  2293  M  −1 cm −1 ). D(+)-Trehalose dihydrate from Saccharomyces cerevisiae  was purchased from Sigma Co.,and salts were of analytical grade. Irreversible Deactivation Experiments Solutions of cutinase at pH 4.5 were heated from 25°C to53°C and kept at this temperature while samples were takenperiodically. Before assaying for lipase activity, eachsample was rapidly cooled and maintained at room tempera-ture. The hydrolytic activity was then determined titrimetri-cally by the pH-stat method (Brocklehurst, 1995) with atitrator (Model 702-SM, Metrohm Titrino), at pH 8.5 androom temperature. The reaction mixture consisted of 30 mLof 3 m  M   Tris-HCl buffer with 3% (w/v) arabic gum asemulsifying agent and 140 m  M   of tributyrin as substrate.The reaction was initiated with the addition of cutinase so-lution to achieve an enzyme concentration of 150 n  M  , andthe amount of NaOH was then recorded every 10 s.The apparent rate constants of deactivation were obtainedfrom a nonlinear fit of the plot  a t   /  a T    25°C performed withO RIGIN  software according to a double exponential equa-tion: a t     a T  = 25°C  =   a t     a T  = 25°C  t  =   +  A 1st e − t k  1st +  A 2nd  e − t k  2nd where  a  is activity,  A  is amplitude, and  k   the rate constant.When more than one deactivation phase is present, apparentrate constants are usually calculated from nonlinear fits,because no choice has to be made with regard to the end of one phase and the beginning of the next. Residuals wereused to confirm the double exponential character of the data.The heating time was taken into account for deactivation;that is, the zero time of the experiment was at 25°C. At25°C, all cutinase molecules are in the native state. Duringthe heating period, some cutinase molecules deactivate ir-reversibly (the amount depends on cutinase concentration)and can be accounted for in  A 1st  if 25°C is taken as the timezero of the experiment. Thus,  A 1st  and  A 2nd  could be relatedto the fraction of cutinase molecules that proceeded to eachirreversible state if a parallel pathway was considered. Thermal Unfolding Thermal unfolding of cutinase at pH 4.5 was monitored byfluorescence emission using a spectrofluorimeter (ModelMPF-3, Perkin-Elmer) with 90° geometry. Cutinasesamples were heated at an approximately constant heatingrate (1.5°C/min) using an auxiliary water bath and the tem-perature of the protein solution was measured to an accu-racy of ±0.1°C using a thermometer (Model 51K/J, Fluka)immersed in the cell while recording the fluorescence in-tensity. The single tryptophan residue (Trp69) of cutinasewas excited selectively at 296 nm and emission was mea-sured at 335 nm. Upon cutinase unfolding, the tryptophanresidue was pulled from the quenching effect of the disul-fide bond between Cys31 and Cys109 and the quantumyield increased (Melo et al., 2001). Slit widths of 2.5 and 5nm were used for excitation and emission light, respec-tively. A narrow excitation slit was necessary to preventcleavage of the disulfide bridge noted earlier (Neves-Petersen et al., 2002).Thermal unfolding was studied based on a two-statemodel, which assumes that only the cutinase native andunfolded states are present (Melo et al., 2001). The fractionof unfolded cutinase (  f  unfolded ) at a given temperature wascalculated according to:  f  unfolded  =   FI T   −  FI  N       FI U   −  FI  N   where FI T   is the fluorescence intensity measured at a giventemperature and FI U   and FI  N   are the values of fluorescenceintensities for the unfolded and native cutinase, respec-tively. FI U   and FI  N   were fitted to linear functions dependingon temperature. Dynamic Light Scattering Dynamic light scattering (DLS) measurements were carriedout using a multiangle apparatus (Brookhaven Corp.)equipped with a He–Ne laser (Model 127, Spectra Physics,with    632.8 nm) with 35-mW power. The diffused lightwas detected by a photomultiplier (placed at a fixed angle of 90°) and analyzed with a 136-channel correlator (ModelBI2030AT). Samples containing 98    M   cutinase in 0.1  M  acetate buffer (pH 4.5) were thermostatted at 298K using awater bath. The solution was filtered through Millipore fil-ters (0.1   m) directly into the optical cell. Diffusion coef-ficients were calculated after fitting the autocorrelationfunction with a double exponential equation (software sup-plied by the manufacturer) because cumulant expansionanalysis is valid only for a single exponential autocorrela-tion function (unimodal distribution). Analysis with the cu-mulant expansion method coincides with the double expo- 852 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 82, NO. 7, JUNE 30, 2003  nential analysis for a unimodal distribution (cutinase sampleat 25°C). C ONTIN  software was used for size distributioninformation (Chu, 1991) to confirm the bimodal distributionobserved for cutinase samples heated to 53°C and cooled to25°C before measurement. The diameter of the cutinasemolecule or its aggregates was calculated using the Stokes–Einstein equation: d   =  k  b T     3   D where  d   is diameter,  k  b  is the Boltzmann constant,  T   istemperature (Kelvin),    is viscosity of water (1.002 × 10 −3 N s m −2 ), and  D  is diffusion coefficient. RESULTS AND DISCUSSIONCutinase Deactivation Profiles Deactivation of cutinase at 53°C and pH 4.5 displays clearlynon-first-order kinetics (Fig. 1A). Two phases were ob-served in a plot of the logarithm of the activity versus time.Experimental data plotted in a semilogarithmic plot couldbe fitted to two straight lines with a transition in between.The first phase was dependent on the cutinase concentra-tion, as expected for an aggregation process, and was re-sponsible for >50% of the loss of activity, depending onprotein concentration. The second phase was independent of the cutinase concentration within experimental error. Theseobservations were confirmed when the deactivation profileswere fitted to a double exponential equation that followedthe data well (fits not shown). These double exponential fitswere carried out for quantification purposes only, with noassumption about the deactivation mechanism.Figure 1B shows the apparent rate constants calculatedfrom the double exponential fits, where  k  app  for the firstphase increased linearly with cutinase concentration accord-ing to a process displaying an order of reaction greater thanone. This linear trend versus concentration was observedpreviously for association reactions such as monomer–dimer transitions (No¨lting, 1999). Determination of the ap-parent order of reaction by plotting the logarithm of theinitial rate versus the logarithm of cutinase concentrationyielded a value of 1.8 (the order of reaction determined withrespect to concentration [Laidler, 1987], inset to Fig. 1B).The amplitudes of the first and second phases, normalizedto the amount of cutinase present initially, were also deter-mined from the double exponential fits just described, andwere plotted versus the cutinase concentration in Figure 1C.Increasing cutinase concentration favors aggregation, caus-ing an increase in the amplitude of the first phase. Thisincrease means that the fraction of cutinase molecules ini-tially present, which became part of the aggregate, had in-creased. Also, the increase in the amplitude of one phaseconcomitant with the decrease in the amplitude of the otherphase, as observed in Figure 1C, indicates a competitiveprocess. As more cutinase molecules deactivated during theaggregation phase, less remained to deactivate during thesecond phase and vice versa. Characterization of Aggregates by DynamicLight Scattering Additional proof of the formation of aggregates was col-lected using dynamic light scattering (Fig. 2). Cutinase at25°C is monomeric even at a concentration of 98    M  . Cu-tinase dimensions are 4.5 × 3.0 × 3.0 nm (Martinez et al.,1992) and, considering the cutinase molecule to be a sphereof equal volume, cutinase diameter was calculated to be 3.4nm, a value close to the value of 3 nm measured by DLS.The C ONTIN  method of analysis (insets in Fig. 2), with re-sults consisting of a distribution function of sizes, was Figure 1.  (A) Deactivation profiles for increasing concentrations of cu-tinase at 53°C and pH 4.5: (  ) 12    M;  (  ) 31    M;  (  ) 50    M;  (  ) 73   M;  and (  ) 98    M  . Solid lines correspond to the fits following a parallelpathway with an aggregation (order of reaction  n  3) and a first-orderstep. Correlation coefficients of the fits were between 0.975 and 0.999. (B)Effect of cutinase concentration on the apparent rate constants, (  ) k  (app)1st ) and (  )  k  (app)2nd , determined by a double exponential fit (seeMaterials and Methods). (Inset) Determination of the apparent reactionorder of the aggregation process by plotting the logarithm of the initial rate(− d  ( a t   /  a T    25°C )/  dt  ) t    0  versus the logarithm of cutinase concentration. (C)Normalized amplitudes of the first (  ) and second phases (  ) versuscutinase concentration determined using a double exponential fit (see Ma-terials and Methods). BAPTISTA ET AL.: CUTINASE AT  P H 4.5 DEACTIVATES FOLLOWING A PARALLEL PATHWAY 853  used to confirm the unimodal or bimodal distribution thatwas determined by a double exponential fit of the autocor-relation function (see Materials and Methods). After 2-minincubation at 53°C, aggregates were detected along withmonomeric cutinase. These aggregates displayed a sixfoldincrease in size when analyzed by a double exponential fit,but should be highly heterogeneous, according to C ONTIN analysis. After 10 min at 53°C, most cutinase was aggre-gated (small amounts of monomeric cutinase were undetect-able by DLS) and the aggregates more defined in size be-cause C ONTIN  revealed a narrow distribution closer to thevalues determined by the double exponential fit. Aggregation of Unfolded Conformationof Cutinase Initially, at 53°C, 88% of cutinase molecules were unfolded(Table I), and therefore one should clarify whether it is thenative or the unfolded conformation that gives rise to theaggregate. The equilibrium between the native and unfoldedconformations was shifted when trehalose was added. In-creasing trehalose concentration shifts the equilibrium to-ward the native conformation (Table I), according to previ-ous observations obtained at pH 9.2 (Baptista et al., 2000),and changes the deactivation profile. When the unfoldedequilibrium was not shifted significantly (for 0.2  M   of tre-halose), deactivation took place even faster in the presenceof trehalose (Fig. 3). For concentrations of 0.5  M   trehalose,and even at 0.8  M  , the first phase became less significantbecause the fraction of unfolded cutinase decreased signifi-cantly. When the unfolded fraction was only 15%, aggre-gation was no longer observed and deactivation obeyedfirst-order kinetics (Fig. 3). This behavior indicates that ag-gregation resulted from the unfolded conformation, becausedecreasing the amount of unfolded molecules while keepingprotein concentration constant made aggregation less likely,perhaps even leading to its suppression. Recent work usinglattice models (Dima and Thirumalai, 2002) and experimen-tal data (Silow et al., 1999) have shown that protein aggre-gation proceeds directly from the unfolded conformation, inaccordance with our observations. Parallel Versus Series-Type Deactivation Pathway The meaning of the amplitude of the activity decay for eachphase in non-first-order kinetics depends on the deactivationpathway being a series or parallel type. For a series type, theamplitude is proportional to the activity of each conforma-tional state on the pathway. It is conceptually difficult toaccept that the non-catalytically active unfolded cutinaseshould give rise to a catalytically active aggregated state,displaying a specific activity that could reach about 40% of the specific activity of the native conformation (for cutinaseconcentrations such as 31    M   or lower the second phasewould start at around 40% of the specific activity obtainedat 25°C). For a parallel-type pathway the amplitude corre-lated with the number of enzyme molecules undergoingeach parallel pathway. The pattern observed for the ampli-tudes of the two phases in Figure 1C, where the increase inthe amplitude of the first phase was concomitant with the Figure 2.  Size distribution measured by DLS and analyzed using adouble exponential fit and C ONTIN  (insets) for cutinase (98    M  ) samples at25°C. The top trace shows the results for the unheated sample and the threefigures underneath show the results for samples incubated during severaldifferent timepoints (2, 10, and 20 min) at 53°C and cooled to 25°C priorto measurements. Table I.  Fraction of cutinase unfolded at 53°C measured by fluorescenceemission during a thermal unfolding scan.[Trehalose] (  M  )  f  unfolded  at 53°C0.0 0.880.2 0.860.5 0.500.8 0.281.0 0.15 854 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 82, NO. 7, JUNE 30, 2003  decrease in the amplitude of the second phase, indicates aparallel-type deactivation pathway, as illustrated by the fol-lowing scheme: N   →←   K  U k  1st U agg k  2nd I where  N  and  U  are the native and the reversible unfoldedconformations, respectively,  U agg  is the aggregate formedfrom the unfolded state, and  I  is the irreversible state formedfollowing a first-order step.Indeed, deactivation data in Figures 1A and 3 were fittedafter solving the differential equation describing this path-way (see Appendix): − da    dt   =  k  1st  a n +  k  2nd  a where  a  is activity,  t   is time,  k  1st  and  k  2nd  are the intrinsicrate constants of the two phases, and  n  is the order of theaggregation phase. Apparently, the analytical solution of this equation for  n    2 should be able to fit the data,because the order of reaction was calculated to be 1.8. How-ever, this was not the case, as with  n  2 the data cannot befitted (see Fig. 4), substantiating the large error associatedwith the determination of the order of reaction with respectto concentration. The solution for  n  3 fits rather well theexperimental data (as shown in Figs. 1A and 3), and theintrinsic rate constants for this pathway are presented inTables II and III. The values of   k  2nd  are scattered (explana-tion for this scatter is given in the Appendix), but they arecomparable to the value measured in the presence of 1  M  trehalose when the aggregation phase was absent. This isfurther evidence against a series-type mechanism because itis not expected that both the aggregated and the unfoldedstate would deactivate with the same rate constant. Proteinaggregation mechanisms, namely in amyloid diseases, canbe described by a seeded-nucleated model where a nucleuscomposed of at least a dimer acts as a template for proteinassembly (Dima and Thirumalai, 2002). Protein aggregationcan also be thought of as a step-growth “polymerization”reaction (Allcock and Lampe, 1990), occurring by consecu-tive aggregation to a dimer, trimer, tetramer, and so on,reaching large aggregates. Based on this mechanism, theorder of reaction for the aggregation mechanism does notequal the number of protein molecules in the mature aggre-gate, as measured by DLS. Aggregation must occur in thepresence of a “seed,” which seems to be a trimer for cuti-nase, because data can only be fitted for  n  3.Parallel pathways can explain two phases on deactivationprofiles, but they have been used rather infrequently, if atall, because it must be assumed that the enzyme state ex-hibits heterogeneity either in the specific activity or in theactivation energy for deactivation. The concept of micro-heterogeneity, where an enzyme may consist of a largenumber of closely related species, would allow parallelpathways, but it seems to hold only for specific cases, suchas intramolecular disulfide interchange in serum albumin orenzymatic nicking prior to purification (Sadana, 1991). Ob-viously, (micro)heterogeneity is intrinsic to the ensemble of unfolded conformations, and parallel pathways have beendescribed for folding (Guo and Thirumalai, 1995). Despitethe ensemble of unfolded conformations, cutinase deactiva-tion at pH 9.2 displays first-order kinetics (Baptista et al.,2000). (Micro)heterogeneity of the ensemble of unfolded Table II.  Intrinsic rate constants for cutinase deactivation at 53°C andpH 4.5 assuming a parallel pathway with different orders of reaction.[Cutinase](   M  ) k  1st  × 10 5 [min −1 (m  M   /min) −( n −1) ]  k  2nd  × 10 4 (min −1 )12 3.28 0.00731 0.42 6.550 0.87 0.01973 0.95 63.998 1.00 1.1 Figure 3.  Deactivation profiles for cutinase (31    M  ) at 53°C and pH 4.5in the presence of increasing concentrations of trehalose: (  ) 0.0  M;  (  )0.2  M;  (  ) 0.5  M;  (  ) 0.8  M;  and (  ) 1.0  M  . Solid lines are the fitsfollowing a parallel pathway with an aggregation (order of reaction  n  3)and a first-order step, except for 1  M   trehalose, where a single first-orderdeactivation step fits the data. Correlation coefficients of the fits werebetween 0.982 and 0.999. Figure 4.  Simulation of Eq. (1) for  n  2 (thin solid line) and  n  3(thick solid line) and comparison to experimental data obtained with 50   M   cutinase. For  n  2, the biphasic nature of deactivation is too smoothand is unable to fit the data. BAPTISTA ET AL.: CUTINASE AT  P H 4.5 DEACTIVATES FOLLOWING A PARALLEL PATHWAY 855
Search
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks