Biochimica el Bwphysicc¢Act#, 1076(1991) 1.5 22
1991 El~vicrS¢i¢0¢ePublishersRe+(Biomeali~alDivisiort)0167-483~'/9l/$03.50 ADONIS 0167483891~
15
BBAPRO33774
The characterisation of immobilised lignin peroxidase by flow injection analysis Matthias S. Fawer, Jtirg Stierli, Stephe~ Cliffe and A.rmin Fiechter lnstirule of Bioterhnolo~,', ETH HiJnggerberg Ziirich (Switz~rtandj
(Received28 May 1990) Keywolds: Ligninperoy.idase;Flowinjectionanalysis;immobilizedenzyme lmmohilised lignin peroxidase has been investigated using a flow system in the steady state =h~tdby flow injection analysis (FIA). in the steady state, the extreme sensitivity of the enzyme towards inactivation b), H20 z resulted in a stable response only in the presence of saturating levels of organic substrate and at very low (10 pM) peroxide concentrations. By contrast, the low contact time during FIA led to a stable response to injectimas elf 100/~M H2Oz. At higher peroxide cuncenlrations a reprodma'ble inactivation was observed, allowing a study of factors a/leering both activity and stability, Lignin peroxidase substrates that undergo at least semi-reversible oxidafion/reduetio~ Including high-molecular-weight lignln fractions, could be detected by electreehemical reduction of the oxidation products. With this detaetion system it was possible to demons~atc the role of veratryl alcohol as mediator. This mediated oxidation of I|gnin functioned only when all components were present simultaneously, and was not observed when lignin was
separated from the site of veratt3'! alcohol oxidation. Introduction Lignin peroxidase plays a key role in the breakdown of lignin and has been the subject of an extensive research effort due to its potential application in several biotechaotogical processes [1,2]. The enzyme from Phanerochaete chrysosporium is well characterised [3-7]; multiple isozymes are observed varying in isoelectric point between pH 3.0 and 5.0, but all consisting of a single po!ypeptide chain with polysaecharide content 10-45~, molecular mass 39--43 kDa, and which conrains a single iron protoporphyrin prosthetic group. The native enzyme is oxidised by a peroxide to compound !, which may be denoted as a" Fe(IV)--O P+" species, indicating formation of an organic cation radical within the porphyrin ring system. Upon reaction with substrate, compound 1 undergoes a one electron reduction to compound II, which in turn undergoes a further one electron reduction to the native enzyme [3,8,9]. Whilst two electron reductions are possible, it is suggested that the one electron oxidation of an organic suhstrate to form a cation radical is the primary event; and that the
Abbreviati0n:FIA, flowinjectionanalysis. Correspondence:S. Cliff¢.Instituteof Biotechnology,E"THHOnfitiecberg, CH~,093Zurich,Switzerland.
variety of reactions reported to be eataiysed by the enzyme are due either to chemical reactions of these radicals, or a further one electron reduction [10-12]. Excess hydrogen peroxide leads to formation of an inactive form, compound IlI, and eventually to irreversibIe enzyme inactivation [131. Lignin peroxidase activity is normally assayed by monitoring the production of veratryl aldehyde from vexatryl alcohol and hydrogen peroxide [3]. Veratryl alcohol is also a secondary metabolite of P. chrysosporium, stabilJses the soluble enzyme and has been proposed to mediate the oxidation of native lignin via the initial cation radical product []4-16]. Flow injection analysis (FIA) is an established technique for performing chemical and enzymatic assays [17]. It involves the injection of a small sample into a flowing stream, which is either ¢nixed with a further stream or streams (chemical or soluble enzyme reaction) or passed through a reactor column (immobilised enzyme) before the signal, /n, the form of a peak, is detected in a flow cell The enzyme reactor may be incorporated as an integral part of the flow-cell. FIA also offers an, as yet, underexploited methodology for the systematic characterisation of enzymes. Our interest lay in the possibility of using lignin peroxidase in analysis, an interest prompted by the ability of the enzyme to oxidise polycyclic aromatic hydrocarbons. To this end we have studied the on-
16 zymatic properties of immobilised lignin peroxidase in batch and flow systems, and in particuF.~r by FIA. Materials and Methods All chemicals were of the highest purity available from Fluka, Aldrich or Merck, Phosphate buffels were prepared by mixing their sodium salts. Tartrate, acetate and oxalate buffers were prepared by adjusting the pH of a earboxylic acid solution with 4 M sodium hydroxide, followed by volume adjustment to the final acid concentration as given. Tartrate-phosphate buffers were prepared similarly, the pH being adjusted with solid di-sodium hydrogen phosphate. Hydrogen peroxide solutions were prepared daily and concentrations detcwnined spectrophotometrieally, e24o = 40 M - 1. cm- t 118]. Cyclic voltammetry was performed using a BAS 100 dectroanalyser, with electrodes, a platinum counter, glassy carbon working and an Ag/AgCI/3 M KCI reference, supplied by the manufacturer. The working electrode was polished (alumina), washed in a sonic bath for 30 s and then cycled ten times through the voltage range, and in the electrolyte used for the subsequent voltammogram. Lignin peroxidase was a kind gift from Dr. Multi Leisola of Cultor, Finland (lot No. 884320). The ~,attially purified extracellular fraction was sterile fihered on arrival and stored at 40C. lso~electric focussing showed a similar isozyme profde Io that reported by Ginmoff et at. [5]. Lignin peroxide concentration was determined to be 0.1 raM, ~409ffi !65 ram - I . cm -I [191 and activity, 131 international units (IO). ntl -t. One IO corresponds to the quantity of enzyme producing 1 /tmol-w.in-i veratrylaldehyde [3] under the following conditions: in 1 mi 0.1 M sodium tartrate (pH 3.0), 2 mM veraffyl alcohol (final concentrations and volume) at 250C. The reaction was started by addition of 2 mM H202 (0,1 ml) and the initial increase in absurbance al 310 am was measured with a Shiraadzu UV-240 spectrophotometer (s~t0 = 9300 M - ' - em- ~). Immobilisation procedures: For each immobilisation. 2 ml of the lignin pero~idase solution was exchanged into the appropriate buffer by passage through Sephadex G-25 (PD-10 column, Pharmacia). All coupling buffers contained 2 raM veratryl alcohol, Three methods were investigated: (A) Glutaraldehyde-activated glass beads - 500 rag dry beads (Daltosil 500, 0.1-0.2 nun diameter~ Serva) previously derivitised with 3-aminopropyltfiethoxysilane using the aqueous procedure, and activated as described by Wectall [20] - were incubated with 2 ml enzyme solution in 50 mM sodium phosphate (pH 7.0), 24 h at 4=C. (B) Epoxy activated acrylic beads - 2 g beads, Euperglt C (Ri~hm) - were incubated with 2 mi enzyme
solution in 50 mM sodium phosphate (pH 6.5), 24 h at 4°C and 24 h at room temperature. (C) To hydrazide derivitised agarose - the support was prepared by addition of 3 ml 107o anhydrous hydrazinv in isopropanol to 25 ml suspension of N-hydroxysuccinimide activated agarose-Affigel 15 (Bio-Rad) in isopropanol at 4°C. The derivitised beads were washed and stored at 4°C. Lignin peroxidase with oxidised sugar residues was prepared by reaction with 50 mM sodium periodate in 0.1 M sodium acetate (pH 4_5) for 24 h at 4°C in the dark, followed by a second passage ovel a PD-10 column. 1.5 mi bead suspension was incubated with 2.0 ml enzyme solmion in 50 ruM se~4ium acetate (pH 4.5), for 24 h at 4°C. Coupling yields were calculated from the fall in absorbance at 409 nm and by the fall in soluble enzyme activity. After coupling, the beads were washed with wztcr, 0.5 M NaCI, water again and finally with 10 mM tartrate phosphate (pH 4.0) containing 2 mM veratryl alcohol and stored at 4°C. The activity of the preparations was measured in a batch assay: a known q1~tntity of beads was added to a PD-10 colunm (with the Sephadex previously removed) containing 10 ml 0.1 M sodium tartmte (pH 3.0), 2.2 mM veratryl alcohol. 1 ml of 2.2 mM H202 was added and the columns rotated at 30 rpm. After 5 rain, an aliquot of the snpcmatant was collected and the absorbanec at 310 am compared to that of a control. The effectiveness factor ~ of the immobifised preparations was calculated, where ~ is the ratio of the observed activity ( I U / r ~ beads), to the calculated activity (IU soluble enzyme bound/rag beads). For use in the flow system, beads were packed into silicone or polyvtnylchlofide tubing, and retained behind porous polyethylene frits (30 t~m Porex). For kinetic and stability experiments small columns, 1 mm in diameter and 10-30 ram long were used which gave conversions c -407o. To investigate the reductive response to additional substrates, larger columns of diameter 2,0 mm, length 10-30 nun were used, which gave conversions of 40-955. The FIA unit was constructed in this institute. It consists of two, four channel pumps (Ismat~c SA, G[attbrag& CH), an injection and a six-way valve (Raheodyne); each valve driven by an electric actualor; and an automated sampler (a converted U|Irorae fraction collector from LKB). The mechanical parts were controRed by a home made micro-controller system (lute[ 8052 AH Basic) which provided the interface to the user via a key-board and display, and also took over data acquisition and processing. Software was written on a personal computer (Macintosh SE) and saved in a non-volatile memory in the microprocessor. Via R,e key-board it was possible to set the number of samples, injections per sample, lime between injections and integration stop/ start. Two detectors were used in series; an electrochemical detector (Metrohm 641), and a spectrupholometer
17
.-~-., IE
. . . . 2~ . . . . .
t
Fig, t. FIA schemafor LigninIWsoxidaseexperiments.The solid fines indicatethe normalset up. Scheme2a and b w~¢ used to instigate the possibility of mediation by veratryl alcohol at a disla~e. Pl - sampleloadingpump:. P'2.- main pump; SV - six way valveto select runningbulfer RB; B2 - secondbuffer; IV ~ i~j~tion valve: AS - aatomaledsampler: IE - immobilisedenzyme:,ECD - electrochemicalde~ector;S - slz~trophotomele~and W - waste.
(Shimadzu 120-02 single beam equipped with a 100 izl flow through quartz cure!to from Hellma). The electrochemical flow cell was of the wall jet type (Metrohm 656), which incorporates an Ag/AgCI/3 M KCI referonce electrode and a gold counter electrode. Oxidation products were quantified by reduction at a glassy carbon declrode. The output from the two detectors w ~ fed over a two channel chart recorder to the micr~ processor, end for each peak a print out of height and area was obtained. A scheme for the FIA experiments is shown in Fig. 1, with the solid lines indicating the sebup for most experiments. Scheme 2a and b indicate the inclusion of a second hufferstream carrying both veratryl alcohol and a third subs!rate which was used to inv~tigate the possibility of veratryl alcohol mediated oxidation at a distance. The volume of the injection loop was determined to be 100 Ill. Unless otherwise stated, the operating conditions were: ambient temperature (22 ° C), flow rate 1.0 ml- rain 1, running buffer 50 mM tartrate phosphate (pl-I 3.0). Veratryl alcohol was added at the same concentration (routinely 2 raM) to both running buffer and samples. For steady-state experiments, hydrogen peroxide was added to the running buffer, whilst for FIA, hydrogen peroxide (routindy 1130pM) and any additional substrates were injected together. In order to exclude arlifacts dee to loss of enzyme activity, all experiments were conducted in a cyclic fashion, e.g.. from pH 2.0 t h r o u ~ to pH 5.0 and back through to pH 2.0. This was particularly in,per!ant when conducting experiments ~vit,~ varying veratryl alcohol concentrations. Results lmmobilisation yield
.
With the three methodologies A, B and C, coupling .yields were 95, 68 and 53~ with ~ = 0.02, 0.02 and 0.05,
respectively. From the subsequent kinetic analysis it became clear that the low observed activity was due primarily to diffusional constraints and not a toss of active enzyme. Due to their favourable performance in flow systems, subsequent experiments were carried oat with the enzyme immobilised on controlled pore glass by methodology A. Further optimisation was not attempted. Whilst the $1utaraldehyde method has been used successfully for the immobifisation of a figoin pero~dase from Trametes Versicolor [21], the results of Olsson el al. [22] with horse radish peroxidase, indicate that immobilisation chemistry and enzyme preparation may affect markedly the value of ~ obtained. FIA - delector response
The spectroscopic peak height and area was highly reproducible: relative standard deviation = 1.25 (n = 50, peak area, injection of 100 pM H202, 2 mM veratryl alcohol). In general four injections per sample were recorded and values averaged. Under conditions of in. stabiIRy (low [veratryl alcohol] or high [H202]), the number of injections was reduced to two. Unlike peak height, the spectroscopic peak area is not affected by dispersion, but the latter is inversely proportional to flow rate, To calculate the conversion X ( X ~ nmol veratryl aldehyde formed/nmol H202 injected), a calibration curve of peak a~ea x flow rate against nmol veratryl aldehyde injected was constructed. "[he electrochemical response was similarly reproducible, but subject to gradual electrode fouling, which became problematical in the presence of subs!rates) such as phenol and lignia. The response was not calibrated, and wa~ evaluated as peak height in hA. Control injections were always made, e.f~, with and without HzO 2. In general the control signal was insignificant, except in the ~mse of oxidation of a third snbstrate, where the oxidation of veratryl alcohol alone gave a background value. C~ntrol injections were also made without an enzyme column to exclude non-er~zymatic reactions. Initial activity
During initial FIA experiments, a continuous increase in the yield of veratryl aldehyde was observed over a 2-3 h period. A burst of concentrated peroxide, e.g., 3-4 injections of 2 ram H~O2 stabilised the re. sponse. This burst was repeated before each set of experiments, typically once daily. When the immobilised enzyme was exposed to other subs!rates, and in particular high molecolnr lignin fractions, this burst was more frequently ~.xiuired. It has been suggested that bound phenolics are responsible for the reduction of compound 1 in the absence of an added subs!rate [23]. We suspect that the activated enzyme or the initial calion radical products react with compounds bound either to the support (the same effect was also observed
18
pH optimum of e:eraJryl alcohol oxidation
lime (~n)
FiB. 2. lnactiv=don of immobillsed lig,oia pcrordda.s¢ by H20z, Campari~n l~twecm steady slate (A-O and FIA (Q. ID, "h~e.~ relative spectTophotonaetdc~¢spo~sc b ptotted againsl tiros, Cooditia~as: 50 mM tartratC phoSlgaatebuffer (pH 3.0)+2 raM vc~atryl alcohol. A - l0/tM H202; B 50 pM H:O2; C - 100 #M HzOz. Dgri~l~ F|A {with injeCaons ¢~¢ry 3 mini am _ 100 ~.M ltzOz; • - 1 mM HzOz,
with Eupergit) or to the enzyme itself. The burst oxidises these substances. The burst did not leave the enzyme in an activated state as no production of vcratryl aldehyde was seen upon suhzeqtnent flow Of the alcohol in the absence of hydrogen peroxide.
Enzyme stability With sequential batch assays we observed a steady loss in activity. This prompted a systematic investigation, using the flow system, into the stability of the immobilised enzyme at various veratryl alcohol and HzO z concentrations, during FIA and in the steady state. Increasing veratryl alcohol, up to saturating levels of I0 raM, gave a corresponding increase in stability. The fall in activity was dependent upon the veratryl alcohol/HzOz ratio and the exposure time. Thus during FIA, a constant response over 50 injections was observed when this ra|io was >~20. In the steady state, a stable response, less than 5,% fail over 1 h was observed when this ratio was ~ 200 (Fig. 2). This limits steadystate experiments to very low peroxide levels, and thus a signal at the limits of detection. With FIA a wider, if still very limited range of experimental conditions is available.. The constant FIA response at 100 txM H202, 2 ram verattyl alcohol enabled further studies. In ad. dition it was possible to investigate factors affecting inactivation with injections at higher H20: coneentra, tions,
For veratryi alcohol oxidation, the soluble enzyme has an optimum close to pH L0, with inactivation below this pH value, and a decrease to almost zero activity at pH 5.0 [20]. We observed a rely similar profile for the immobilized enzyme in sodium tartrate, phosphate and tartrate/phosphate buffers. The stability of the immobilized preparation (as measured by noting the fail in activity upon injection of 1 mM HzO2) was identical at pH 3.0 and 4.0. At pH 2.0 a reduced stability was observed and below this pH value the enzyme was rapidly inactivated by H~O~.
Enzyme activators and stabilisers A series of substances was SCreened for their effect upon enzyme activity (by co-injeclion with 100 p.M H202) and upon ear.yme stability (by co-lnjection with 2 mM H20 ~) at concentrations between 10 ttM and 2.1) raM. Little ( ~ 5%) or no effect was observed for the foffowing substances: NaCI, CuSO4, st3,-'liumglyoxylate, potassium ferricyanide, tetra~acetylethylenediamine (a source of acetyl peroxide) and tetramethylethylene, diamine (a catalyst for radical production from peroxides). Ferrous chloride caused rapid inactivation, possibly due to an enzyme catalyzed Fenton reaction and subsequent attack of free hydroxyl radicals upon the enzyme. Sodium oxalate showed a concentration-dependent inhibition that was only partially reversible. MnCI 2 (2 raM) and sodium [ecrocyanide (10 p,M) were slightly activators, + 10%; whilst at higher concentrations inhihition was observed. Reducing agents in general had an inhibitory effect; to test rigorously whether any stabilisation occurred, a series o; experiments was run with the additives present i~ the running buffer. As seen in Fig. 3, fully reversible inhibition was observed only with ~tscorbate. None of the additives had a stabilising effect. The addition of reducing agents during biotech-
C Na~SO~ O Na4Fe (CN)I G ,N~ 0 0 DTT C Ascorbelte C
12' 1t 10, S' S' ?, S S, 4 3 2 I o
~
40
i~
80
10o
Enzyme activity and stability with different peroxides At a concentration of !90 pM, the same activity was
observed with H20 ~ and sodium perhorate: and a similar fall in activity for both peroxides was observed at 2 raM. At 100/~M, no activity was observable with ammonium persulphate or with t.butyl hydrogen peroxide.
Fig. 3 Effect of I~,ducing agents upon ligrdn peroxidase activity and stability. Conditions: 50 raM tartral¢ phosphate butt¢.v {pH 3.0~ + 2 raM veratry, l alcohol, tO i;njcctions of ;2 ]ttM l-i;:Oz'c,'eze made at 3 nnin intecvals in Ire absence of any ~ddi=ivc (C) or in the presence of 100 ~:M reducil'~gagent. A - expectedI~s of ~.ctivitydue to injection
of 2 mM H2Oz.
19 nological processes involving lignin peroxidase has been suggested [24]. The rational would be either to avoid polymefisation of quinoic products and chromophore production or to stabilise the enzyme. Such a stabiliser would have to react with compound lII faster than with compounds 1, il, and H202; a property not apparent amongst the compounds tested.
Organic so/venls The effect of dimelhylformamide, dioxane and acetonitrile upon enzyme activity was investigated in the concentration range 0-20% (v/v). At low concentrations (___3~) all solvents affect the activity with less than 4%. At 10 and 20~ solvenL inhibition was least for die×ann, 7 and 8% respectively followed by dimethylformarnide at 23 and 34~o, and finally by acetonitrile at 36 and 405 respectively. The inhibition was reversible, full recovery of activity being observed aft¢ five injections at each concentration. In general non-water soluble substrates were dissolved in dimethylformamide, due to its high solvating power, and injected at 2~ final solvent concentration. Lignin solubility remained a problem, separation being observed within 1 h at concentrations above 50 mg-1-1. Increasing the organic solvent to 20~ did not improve solubility which was rather pH dependent, being considerably higher above pH 5.0. Enzyme kinetics Soluble enzyme: Two approaches are available to measure the kinetic constants for two substrate reactions [25]. Either one substrate is held in excess ([S~] Kmt ) and the second varied. The process is then repeated vice versa and constants calculated directly from the primal3' kinetic plots. Alternatively both substrates are varied in a systematic way, the kinetic constants being calculated from secondary plots of the data. Since lignin peroxidase is inactivated by high concentrations of hydrogen peroxide, the second alternative was adopted. The initial reaction velocity was measured at series of different veratryl alcohol (0.1-4.0 raM) and hydrogen peroxide concentrations (0.05-0.25 raM). Double reciprocal, Lineweaver-Burk plots gave two series of parallel lines for each substrate; whilst Woolf plots of v against v/[S1] gave two series of lines converging at the x-axis. These results are consistent with a p~ng-pong mechanism of peroxidase reaction. Secondary plots of the y-intercept against y-intercept/IS2] (Fig. 4) were constructed to yield the kinetic constants, K,,= 204 FM (veratryl alcohol) and Kin=44 FM (HzOz). The values calculated here represent an a~'exage for the various isoenaymes present and lie withi,t the range of values reported for the purified isozym~s by other groups [5,6,19]. Immobilised enzyme: Attempts to measure the kinetic constants using the batch assay procedure were unsuc-
0r6
0.4
-
1991 El~vicrS¢i¢0¢ePublishersRe+(Biomeali~alDivisiort)0167-483~'/9l/$03.50 ADONIS 0167483891~
15
BBAPRO33774
The characterisation of immobilised lignin peroxidase by flow injection analysis Matthias S. Fawer, Jtirg Stierli, Stephe~ Cliffe and A.rmin Fiechter lnstirule of Bioterhnolo~,', ETH HiJnggerberg Ziirich (Switz~rtandj
(Received28 May 1990) Keywolds: Ligninperoy.idase;Flowinjectionanalysis;immobilizedenzyme lmmohilised lignin peroxidase has been investigated using a flow system in the steady state =h~tdby flow injection analysis (FIA). in the steady state, the extreme sensitivity of the enzyme towards inactivation b), H20 z resulted in a stable response only in the presence of saturating levels of organic substrate and at very low (10 pM) peroxide concentrations. By contrast, the low contact time during FIA led to a stable response to injectimas elf 100/~M H2Oz. At higher peroxide cuncenlrations a reprodma'ble inactivation was observed, allowing a study of factors a/leering both activity and stability, Lignin peroxidase substrates that undergo at least semi-reversible oxidafion/reduetio~ Including high-molecular-weight lignln fractions, could be detected by electreehemical reduction of the oxidation products. With this detaetion system it was possible to demons~atc the role of veratryl alcohol as mediator. This mediated oxidation of I|gnin functioned only when all components were present simultaneously, and was not observed when lignin was
separated from the site of veratt3'! alcohol oxidation. Introduction Lignin peroxidase plays a key role in the breakdown of lignin and has been the subject of an extensive research effort due to its potential application in several biotechaotogical processes [1,2]. The enzyme from Phanerochaete chrysosporium is well characterised [3-7]; multiple isozymes are observed varying in isoelectric point between pH 3.0 and 5.0, but all consisting of a single po!ypeptide chain with polysaecharide content 10-45~, molecular mass 39--43 kDa, and which conrains a single iron protoporphyrin prosthetic group. The native enzyme is oxidised by a peroxide to compound !, which may be denoted as a" Fe(IV)--O P+" species, indicating formation of an organic cation radical within the porphyrin ring system. Upon reaction with substrate, compound 1 undergoes a one electron reduction to compound II, which in turn undergoes a further one electron reduction to the native enzyme [3,8,9]. Whilst two electron reductions are possible, it is suggested that the one electron oxidation of an organic suhstrate to form a cation radical is the primary event; and that the
Abbreviati0n:FIA, flowinjectionanalysis. Correspondence:S. Cliff¢.Instituteof Biotechnology,E"THHOnfitiecberg, CH~,093Zurich,Switzerland.
variety of reactions reported to be eataiysed by the enzyme are due either to chemical reactions of these radicals, or a further one electron reduction [10-12]. Excess hydrogen peroxide leads to formation of an inactive form, compound IlI, and eventually to irreversibIe enzyme inactivation [131. Lignin peroxidase activity is normally assayed by monitoring the production of veratryl aldehyde from vexatryl alcohol and hydrogen peroxide [3]. Veratryl alcohol is also a secondary metabolite of P. chrysosporium, stabilJses the soluble enzyme and has been proposed to mediate the oxidation of native lignin via the initial cation radical product []4-16]. Flow injection analysis (FIA) is an established technique for performing chemical and enzymatic assays [17]. It involves the injection of a small sample into a flowing stream, which is either ¢nixed with a further stream or streams (chemical or soluble enzyme reaction) or passed through a reactor column (immobilised enzyme) before the signal, /n, the form of a peak, is detected in a flow cell The enzyme reactor may be incorporated as an integral part of the flow-cell. FIA also offers an, as yet, underexploited methodology for the systematic characterisation of enzymes. Our interest lay in the possibility of using lignin peroxidase in analysis, an interest prompted by the ability of the enzyme to oxidise polycyclic aromatic hydrocarbons. To this end we have studied the on-
16 zymatic properties of immobilised lignin peroxidase in batch and flow systems, and in particuF.~r by FIA. Materials and Methods All chemicals were of the highest purity available from Fluka, Aldrich or Merck, Phosphate buffels were prepared by mixing their sodium salts. Tartrate, acetate and oxalate buffers were prepared by adjusting the pH of a earboxylic acid solution with 4 M sodium hydroxide, followed by volume adjustment to the final acid concentration as given. Tartrate-phosphate buffers were prepared similarly, the pH being adjusted with solid di-sodium hydrogen phosphate. Hydrogen peroxide solutions were prepared daily and concentrations detcwnined spectrophotometrieally, e24o = 40 M - 1. cm- t 118]. Cyclic voltammetry was performed using a BAS 100 dectroanalyser, with electrodes, a platinum counter, glassy carbon working and an Ag/AgCI/3 M KCI reference, supplied by the manufacturer. The working electrode was polished (alumina), washed in a sonic bath for 30 s and then cycled ten times through the voltage range, and in the electrolyte used for the subsequent voltammogram. Lignin peroxidase was a kind gift from Dr. Multi Leisola of Cultor, Finland (lot No. 884320). The ~,attially purified extracellular fraction was sterile fihered on arrival and stored at 40C. lso~electric focussing showed a similar isozyme profde Io that reported by Ginmoff et at. [5]. Lignin peroxide concentration was determined to be 0.1 raM, ~409ffi !65 ram - I . cm -I [191 and activity, 131 international units (IO). ntl -t. One IO corresponds to the quantity of enzyme producing 1 /tmol-w.in-i veratrylaldehyde [3] under the following conditions: in 1 mi 0.1 M sodium tartrate (pH 3.0), 2 mM veraffyl alcohol (final concentrations and volume) at 250C. The reaction was started by addition of 2 mM H202 (0,1 ml) and the initial increase in absurbance al 310 am was measured with a Shiraadzu UV-240 spectrophotometer (s~t0 = 9300 M - ' - em- ~). Immobilisation procedures: For each immobilisation. 2 ml of the lignin pero~idase solution was exchanged into the appropriate buffer by passage through Sephadex G-25 (PD-10 column, Pharmacia). All coupling buffers contained 2 raM veratryl alcohol, Three methods were investigated: (A) Glutaraldehyde-activated glass beads - 500 rag dry beads (Daltosil 500, 0.1-0.2 nun diameter~ Serva) previously derivitised with 3-aminopropyltfiethoxysilane using the aqueous procedure, and activated as described by Wectall [20] - were incubated with 2 ml enzyme solution in 50 mM sodium phosphate (pH 7.0), 24 h at 4=C. (B) Epoxy activated acrylic beads - 2 g beads, Euperglt C (Ri~hm) - were incubated with 2 mi enzyme
solution in 50 mM sodium phosphate (pH 6.5), 24 h at 4°C and 24 h at room temperature. (C) To hydrazide derivitised agarose - the support was prepared by addition of 3 ml 107o anhydrous hydrazinv in isopropanol to 25 ml suspension of N-hydroxysuccinimide activated agarose-Affigel 15 (Bio-Rad) in isopropanol at 4°C. The derivitised beads were washed and stored at 4°C. Lignin peroxidase with oxidised sugar residues was prepared by reaction with 50 mM sodium periodate in 0.1 M sodium acetate (pH 4_5) for 24 h at 4°C in the dark, followed by a second passage ovel a PD-10 column. 1.5 mi bead suspension was incubated with 2.0 ml enzyme solmion in 50 ruM se~4ium acetate (pH 4.5), for 24 h at 4°C. Coupling yields were calculated from the fall in absorbance at 409 nm and by the fall in soluble enzyme activity. After coupling, the beads were washed with wztcr, 0.5 M NaCI, water again and finally with 10 mM tartrate phosphate (pH 4.0) containing 2 mM veratryl alcohol and stored at 4°C. The activity of the preparations was measured in a batch assay: a known q1~tntity of beads was added to a PD-10 colunm (with the Sephadex previously removed) containing 10 ml 0.1 M sodium tartmte (pH 3.0), 2.2 mM veratryl alcohol. 1 ml of 2.2 mM H202 was added and the columns rotated at 30 rpm. After 5 rain, an aliquot of the snpcmatant was collected and the absorbanec at 310 am compared to that of a control. The effectiveness factor ~ of the immobifised preparations was calculated, where ~ is the ratio of the observed activity ( I U / r ~ beads), to the calculated activity (IU soluble enzyme bound/rag beads). For use in the flow system, beads were packed into silicone or polyvtnylchlofide tubing, and retained behind porous polyethylene frits (30 t~m Porex). For kinetic and stability experiments small columns, 1 mm in diameter and 10-30 ram long were used which gave conversions c -407o. To investigate the reductive response to additional substrates, larger columns of diameter 2,0 mm, length 10-30 nun were used, which gave conversions of 40-955. The FIA unit was constructed in this institute. It consists of two, four channel pumps (Ismat~c SA, G[attbrag& CH), an injection and a six-way valve (Raheodyne); each valve driven by an electric actualor; and an automated sampler (a converted U|Irorae fraction collector from LKB). The mechanical parts were controRed by a home made micro-controller system (lute[ 8052 AH Basic) which provided the interface to the user via a key-board and display, and also took over data acquisition and processing. Software was written on a personal computer (Macintosh SE) and saved in a non-volatile memory in the microprocessor. Via R,e key-board it was possible to set the number of samples, injections per sample, lime between injections and integration stop/ start. Two detectors were used in series; an electrochemical detector (Metrohm 641), and a spectrupholometer
17
.-~-., IE
. . . . 2~ . . . . .
t
Fig, t. FIA schemafor LigninIWsoxidaseexperiments.The solid fines indicatethe normalset up. Scheme2a and b w~¢ used to instigate the possibility of mediation by veratryl alcohol at a disla~e. Pl - sampleloadingpump:. P'2.- main pump; SV - six way valveto select runningbulfer RB; B2 - secondbuffer; IV ~ i~j~tion valve: AS - aatomaledsampler: IE - immobilisedenzyme:,ECD - electrochemicalde~ector;S - slz~trophotomele~and W - waste.
(Shimadzu 120-02 single beam equipped with a 100 izl flow through quartz cure!to from Hellma). The electrochemical flow cell was of the wall jet type (Metrohm 656), which incorporates an Ag/AgCI/3 M KCI referonce electrode and a gold counter electrode. Oxidation products were quantified by reduction at a glassy carbon declrode. The output from the two detectors w ~ fed over a two channel chart recorder to the micr~ processor, end for each peak a print out of height and area was obtained. A scheme for the FIA experiments is shown in Fig. 1, with the solid lines indicating the sebup for most experiments. Scheme 2a and b indicate the inclusion of a second hufferstream carrying both veratryl alcohol and a third subs!rate which was used to inv~tigate the possibility of veratryl alcohol mediated oxidation at a distance. The volume of the injection loop was determined to be 100 Ill. Unless otherwise stated, the operating conditions were: ambient temperature (22 ° C), flow rate 1.0 ml- rain 1, running buffer 50 mM tartrate phosphate (pl-I 3.0). Veratryl alcohol was added at the same concentration (routinely 2 raM) to both running buffer and samples. For steady-state experiments, hydrogen peroxide was added to the running buffer, whilst for FIA, hydrogen peroxide (routindy 1130pM) and any additional substrates were injected together. In order to exclude arlifacts dee to loss of enzyme activity, all experiments were conducted in a cyclic fashion, e.g.. from pH 2.0 t h r o u ~ to pH 5.0 and back through to pH 2.0. This was particularly in,per!ant when conducting experiments ~vit,~ varying veratryl alcohol concentrations. Results lmmobilisation yield
.
With the three methodologies A, B and C, coupling .yields were 95, 68 and 53~ with ~ = 0.02, 0.02 and 0.05,
respectively. From the subsequent kinetic analysis it became clear that the low observed activity was due primarily to diffusional constraints and not a toss of active enzyme. Due to their favourable performance in flow systems, subsequent experiments were carried oat with the enzyme immobilised on controlled pore glass by methodology A. Further optimisation was not attempted. Whilst the $1utaraldehyde method has been used successfully for the immobifisation of a figoin pero~dase from Trametes Versicolor [21], the results of Olsson el al. [22] with horse radish peroxidase, indicate that immobilisation chemistry and enzyme preparation may affect markedly the value of ~ obtained. FIA - delector response
The spectroscopic peak height and area was highly reproducible: relative standard deviation = 1.25 (n = 50, peak area, injection of 100 pM H202, 2 mM veratryl alcohol). In general four injections per sample were recorded and values averaged. Under conditions of in. stabiIRy (low [veratryl alcohol] or high [H202]), the number of injections was reduced to two. Unlike peak height, the spectroscopic peak area is not affected by dispersion, but the latter is inversely proportional to flow rate, To calculate the conversion X ( X ~ nmol veratryl aldehyde formed/nmol H202 injected), a calibration curve of peak a~ea x flow rate against nmol veratryl aldehyde injected was constructed. "[he electrochemical response was similarly reproducible, but subject to gradual electrode fouling, which became problematical in the presence of subs!rates) such as phenol and lignia. The response was not calibrated, and wa~ evaluated as peak height in hA. Control injections were always made, e.f~, with and without HzO 2. In general the control signal was insignificant, except in the ~mse of oxidation of a third snbstrate, where the oxidation of veratryl alcohol alone gave a background value. C~ntrol injections were also made without an enzyme column to exclude non-er~zymatic reactions. Initial activity
During initial FIA experiments, a continuous increase in the yield of veratryl aldehyde was observed over a 2-3 h period. A burst of concentrated peroxide, e.g., 3-4 injections of 2 ram H~O2 stabilised the re. sponse. This burst was repeated before each set of experiments, typically once daily. When the immobilised enzyme was exposed to other subs!rates, and in particular high molecolnr lignin fractions, this burst was more frequently ~.xiuired. It has been suggested that bound phenolics are responsible for the reduction of compound 1 in the absence of an added subs!rate [23]. We suspect that the activated enzyme or the initial calion radical products react with compounds bound either to the support (the same effect was also observed
18
pH optimum of e:eraJryl alcohol oxidation
lime (~n)
FiB. 2. lnactiv=don of immobillsed lig,oia pcrordda.s¢ by H20z, Campari~n l~twecm steady slate (A-O and FIA (Q. ID, "h~e.~ relative spectTophotonaetdc~¢spo~sc b ptotted againsl tiros, Cooditia~as: 50 mM tartratC phoSlgaatebuffer (pH 3.0)+2 raM vc~atryl alcohol. A - l0/tM H202; B 50 pM H:O2; C - 100 #M HzOz. Dgri~l~ F|A {with injeCaons ¢~¢ry 3 mini am _ 100 ~.M ltzOz; • - 1 mM HzOz,
with Eupergit) or to the enzyme itself. The burst oxidises these substances. The burst did not leave the enzyme in an activated state as no production of vcratryl aldehyde was seen upon suhzeqtnent flow Of the alcohol in the absence of hydrogen peroxide.
Enzyme stability With sequential batch assays we observed a steady loss in activity. This prompted a systematic investigation, using the flow system, into the stability of the immobilised enzyme at various veratryl alcohol and HzO z concentrations, during FIA and in the steady state. Increasing veratryl alcohol, up to saturating levels of I0 raM, gave a corresponding increase in stability. The fall in activity was dependent upon the veratryl alcohol/HzOz ratio and the exposure time. Thus during FIA, a constant response over 50 injections was observed when this ra|io was >~20. In the steady state, a stable response, less than 5,% fail over 1 h was observed when this ratio was ~ 200 (Fig. 2). This limits steadystate experiments to very low peroxide levels, and thus a signal at the limits of detection. With FIA a wider, if still very limited range of experimental conditions is available.. The constant FIA response at 100 txM H202, 2 ram verattyl alcohol enabled further studies. In ad. dition it was possible to investigate factors affecting inactivation with injections at higher H20: coneentra, tions,
For veratryi alcohol oxidation, the soluble enzyme has an optimum close to pH L0, with inactivation below this pH value, and a decrease to almost zero activity at pH 5.0 [20]. We observed a rely similar profile for the immobilized enzyme in sodium tartrate, phosphate and tartrate/phosphate buffers. The stability of the immobilized preparation (as measured by noting the fail in activity upon injection of 1 mM HzO2) was identical at pH 3.0 and 4.0. At pH 2.0 a reduced stability was observed and below this pH value the enzyme was rapidly inactivated by H~O~.
Enzyme activators and stabilisers A series of substances was SCreened for their effect upon enzyme activity (by co-injeclion with 100 p.M H202) and upon ear.yme stability (by co-lnjection with 2 mM H20 ~) at concentrations between 10 ttM and 2.1) raM. Little ( ~ 5%) or no effect was observed for the foffowing substances: NaCI, CuSO4, st3,-'liumglyoxylate, potassium ferricyanide, tetra~acetylethylenediamine (a source of acetyl peroxide) and tetramethylethylene, diamine (a catalyst for radical production from peroxides). Ferrous chloride caused rapid inactivation, possibly due to an enzyme catalyzed Fenton reaction and subsequent attack of free hydroxyl radicals upon the enzyme. Sodium oxalate showed a concentration-dependent inhibition that was only partially reversible. MnCI 2 (2 raM) and sodium [ecrocyanide (10 p,M) were slightly activators, + 10%; whilst at higher concentrations inhihition was observed. Reducing agents in general had an inhibitory effect; to test rigorously whether any stabilisation occurred, a series o; experiments was run with the additives present i~ the running buffer. As seen in Fig. 3, fully reversible inhibition was observed only with ~tscorbate. None of the additives had a stabilising effect. The addition of reducing agents during biotech-
C Na~SO~ O Na4Fe (CN)I G ,N~ 0 0 DTT C Ascorbelte C
12' 1t 10, S' S' ?, S S, 4 3 2 I o
~
40
i~
80
10o
Enzyme activity and stability with different peroxides At a concentration of !90 pM, the same activity was
observed with H20 ~ and sodium perhorate: and a similar fall in activity for both peroxides was observed at 2 raM. At 100/~M, no activity was observable with ammonium persulphate or with t.butyl hydrogen peroxide.
Fig. 3 Effect of I~,ducing agents upon ligrdn peroxidase activity and stability. Conditions: 50 raM tartral¢ phosphate butt¢.v {pH 3.0~ + 2 raM veratry, l alcohol, tO i;njcctions of ;2 ]ttM l-i;:Oz'c,'eze made at 3 nnin intecvals in Ire absence of any ~ddi=ivc (C) or in the presence of 100 ~:M reducil'~gagent. A - expectedI~s of ~.ctivitydue to injection
of 2 mM H2Oz.
19 nological processes involving lignin peroxidase has been suggested [24]. The rational would be either to avoid polymefisation of quinoic products and chromophore production or to stabilise the enzyme. Such a stabiliser would have to react with compound lII faster than with compounds 1, il, and H202; a property not apparent amongst the compounds tested.
Organic so/venls The effect of dimelhylformamide, dioxane and acetonitrile upon enzyme activity was investigated in the concentration range 0-20% (v/v). At low concentrations (___3~) all solvents affect the activity with less than 4%. At 10 and 20~ solvenL inhibition was least for die×ann, 7 and 8% respectively followed by dimethylformarnide at 23 and 34~o, and finally by acetonitrile at 36 and 405 respectively. The inhibition was reversible, full recovery of activity being observed aft¢ five injections at each concentration. In general non-water soluble substrates were dissolved in dimethylformamide, due to its high solvating power, and injected at 2~ final solvent concentration. Lignin solubility remained a problem, separation being observed within 1 h at concentrations above 50 mg-1-1. Increasing the organic solvent to 20~ did not improve solubility which was rather pH dependent, being considerably higher above pH 5.0. Enzyme kinetics Soluble enzyme: Two approaches are available to measure the kinetic constants for two substrate reactions [25]. Either one substrate is held in excess ([S~] Kmt ) and the second varied. The process is then repeated vice versa and constants calculated directly from the primal3' kinetic plots. Alternatively both substrates are varied in a systematic way, the kinetic constants being calculated from secondary plots of the data. Since lignin peroxidase is inactivated by high concentrations of hydrogen peroxide, the second alternative was adopted. The initial reaction velocity was measured at series of different veratryl alcohol (0.1-4.0 raM) and hydrogen peroxide concentrations (0.05-0.25 raM). Double reciprocal, Lineweaver-Burk plots gave two series of parallel lines for each substrate; whilst Woolf plots of v against v/[S1] gave two series of lines converging at the x-axis. These results are consistent with a p~ng-pong mechanism of peroxidase reaction. Secondary plots of the y-intercept against y-intercept/IS2] (Fig. 4) were constructed to yield the kinetic constants, K,,= 204 FM (veratryl alcohol) and Kin=44 FM (HzOz). The values calculated here represent an a~'exage for the various isoenaymes present and lie withi,t the range of values reported for the purified isozym~s by other groups [5,6,19]. Immobilised enzyme: Attempts to measure the kinetic constants using the batch assay procedure were unsuc-
0r6
0.4
-
