Journal of Colloid and Interface Science214,319–332 (1999)Article ID jcis.1999.6189, available online at http://www.idealibrary.com on
Chymotrypsin Adsorption on Montmorillonite: Enzymatic Activityand Kinetic FTIR Structural Analysis
M. H. Baron,*,1 M. Revault,* S. Servagent-Noinville,* J. Abadie,† and H. Quiquampoix†
*Laboratoire de Dynamique, Interactions et Re´activite, UPR 1580 CNRS, Universite´ Paris VI, 2 rue Henri Dunant, 94320 Thiais, France;and †UFR de Science du Sol INRA-ENSAM, 2 place Pierre Viala, 34060 Montpellier, France
Received November 4, 1998; accepted February 25, 1999
ohekins at slulth
itivth
m nu-t n ofp
ptivec clayf andt se ofb re-t allyo ene ch asc thee and( yme( ts inf n oft
am pre-t urala latedt ce oft n am Us-i pD-d na lvedv thea dra-t opti-m thev pend-i iarys ed tod thea o thec filesm the@g
Soils have a large solid surface area and high adsorptive capac-ties. To determine if structural and solvation changes induced bydsorption on clays are related to changes in enzyme activity,-chymotrypsin adsorbed on a phyllosilicate with an electroneg-tive surface (montmorillonite) has been studied by transmissionTIR spectroscopy. A comparison of the pH-dependent structuralhanges for the solution and adsorbed states probes the electro-tatic origin of the adsorption. In the pD range 4.5–10, adsorptionnly perturbs some peripheral domains of the protein compared tohe solution. Secondary structure unfolding affects about 15–20eptide units. Parts of these domains become hydrated and othersntail some self-association. However, the inactivation of the cat-lytic activity of the adsorbed enzyme in the 5–7 pD range is dueess to these structural changes than to steric hindrance whenhree essential imino/amino functions, located close to the en-rance of the catalytic cavity (His-40 and -57 residues and Ala-149nd chain residue), are oriented toward the negatively chargedineral surface. When these functions lose their positive charge,
Soil bacteria and fungi involved in the biodegradationrganic matter secrete extracellular enzymes (1, 2). Tnzymes hydrolyze insoluble or adsorbed polymers, maoluble monomers that can reach microorganisms or roote taken up across cell membranes by specific transpor
ems and then metabolized (3). The effect of this extracelnzyme activity is clearly expressed in the few mm aroundoots in the soil, the rhizosphere (4). In particular, the posffect of the symbiosis between fungi and plant roots,
1 To whom correspondence should be addressed. E-mail: baron_mhnrs.fr.
a
319
fsegndys-areee
ycorrhizal association, on the phosphorus and nitrogenrition of the host plant is in a large part due to the secretiohosphatases (5–7) and proteases (8).Soils have a large solid surface area and high adsor
apacities, mainly due to the high surface energy of theraction (9). Enzymes have a high affinity for the claysheir interaction with such surfaces leads to a decreaoth mobility and catalytic activity (10 –18). Two interp
ations are currently invoked to explain the experimentbserved pH shift of the optimal enzyme activity whnzymes are adsorbed on electronegative surfaces sulays: (i) a surface pH effect due to the presence ofnzyme active site in the double diffuse layer (19 –23)ii) a pH-dependent modification of the adsorbed enz13, 16, 24 –29). We present here experimental resulavor of a third mechanism, a pH-dependent orientatiohe enzyme on the surface.
Our aim is to determine fora-chymotrypsin, taken asodel proteolytic enzyme of the rhizosphere, which inter
ation of the interaction is most appropriate and if structnd solvation changes induced by adsorption could be re
o the changes of the enzyme activity. The pH dependenhe activity of this enzyme is modified by adsorption oontmorillonite clay with an electronegative surface (9).
ng FTIR transmission spectroscopy, we have quantifiedependent secondary structures fora-chymotrypsin in solutiond a-chymotrypsin adsorbed on the clay. The time-resoibrational analyses allow specific ionization states ofcidobasic side chains to be related to local folding, or hy
ion, which appear to be determinant parameters for anum enzyme activity. Special attention is devoted to
arious possible orientations of the adsorbed enzyme deng on the location of positively charged sites in the terttructure of the protein. The pD dependence was analyzetermine which of these side chains could be involved indsorption process. The effect of a steric hindrance due tlay will be discussed in relation to the pD-dependent proeasured for the enzymatic activity in solution and forlvt-
8wc er( ge.F L( ofc lesw rgon-s -d sc DClo i-t lowd on ofa e,t lyr s than1 at-i ga same
(%)
320 BARON ET AL.
MATERIALS AND METHODS
hemicals
Dideuterium oxide (D2O) was obtained from the CEFrance), and NaD2PO4 and Na2DPO4 salts for the buffer werbtained by exchange in D2O. A Wyoming montmorilloniteith a size fraction,2 mm, a specific surface area of 800 m2/g,nd an electric charge of 1.253 1026 mol z e2/m2 was usedhea-chymotrypsin was purchased from Sigma (C 4129,
I, from bovine pancreas). The crystalline protein has a33 6 nm3 size (30). It is made of three polypeptide cha
inked by five disulfide bridges. Among the 241 amino aesidues, 14 have carboxylic functions (9 Asp1 5 Glu), 16ave amino/imino functions (13 Lys1 3 Arg), 4, have pheolic functions (4 Tyr), and 2 have imidazole functions (2 Hhe isoelectric point (iep) is 8.6. In the crystalline stateecondary structure is made of parallel and antipar–sheets for;51% of the polypeptide backbone. Only 11%
he backbone has a helical structure (30, 31).The substrate used for the measurement of the cat
ctivity wasN-benzoyl-L-tyrosinep-nitroanilide (BTNA), ob-ained from Sigma (B 6760). The reaction was stoppedddition of 1 mol/L trichloroacetic acid, and the reaction prct, aniline, was measured by its absorbance at 410 nm
wo centrifugations of 10 and 15 min at 40,000g. Citrate (pH,), phosphate (7, pH , 8), and borate (pH. 8) buffers weresed. The ionic strengths of the buffers were estimated
he sodium ion concentration. The final concentrationsere 30 mg/L ofa-chymotrypsin, 77mM BTNA, and 0.7 g/Lf montmorillonite, the final ionic strength was 10 mM, and
emperature was 25°C. The catalytic activity was measureH following three procedures previously described (6, 13
he chymotrypsin in solution, as a control, (ii) the chymotrin in the presence of montmorillonite after 1 h of contactontaining both free and adsorbed chymotrypsin, and (iii)hymotrypsin in the supernatant after centrifugation (15 m
TABpD-Dependent Intensities of the Amide I Components,
0,000g) of the chymotrypsin–montmorillonite suspensieasuring only the free fraction of the chymotrypsin.
TIR Spectroscopy
For protein in H2O the contribution of thedOH of waterirectly bonded to the protein could be significant in the Amrange. Moreover, this hydration can be modified by ads
ion (32, 33). Thus for adsorbed hydrophilic proteins a quitative analysis of Amide I absorption profiles exclusivelyerms of backbone structure and solvation is ambiguous.2Oas used instead of H2O (the dOD mode is in the 1200- t250-cm21 range). This raises the question of possible speecondary structures in D2O (34). Indeed, D2O could reducinetics, but, to date, CD (35), as NMR analyses performemall proteins, have not displayed structural difference (3D2O buffers (0.055 mol/L) were initially prepared at pDith Na2DPO4 salts (pD5 pH 1 0.4) (37). DCl or NaODoncentrated solutions in D2O were added to the initial buffpD 8) to obtain six different pD values in the 4–12.5 ranor the solutions,a-chymotrypsin was dissolved at 5 mg/m2 3 1024 mol/L). The suspensions contained 5 mg/mLhymotrypsin and 10 mg/mL of montmorillonite. All sampere prepared at room temperature (20–22°C) and in an aaturated glovebag to avoid any H2O contamination. The adition of a-chymotrypsin ora-chymotrypsin–clay mixtureould modify the pD of the aqueous buffers adjusted withr NaOD. The pD was always stabilized 2 h after these add
ions. Especially for initial pD between 9.5 and 11.5, a secrease to ca. 8.5 was always measured after additi-chymotrypsin ora-chymotrypsin and clay. In this pD rang
o reach a structural equilibrium,a-chymotrypsin apparenteleases more protons than it captures. The difference, les028 mol/L proton, means that the ratio of slowly deproton
ng acidic functions (COOH,F-OH) on slowly protonatinmino/imino basic sites is greater than one. When the
nitial buffers were used, the resulting pD varied by less
1pressed as Percentage of the Overall Amide Intensity,ideuterium Oxide Solution
lm1 utiw mW pet d ina lidps eros was r ths r 1m edm
35c ins hanb spes mHh thH wa ighs So d ot ecov numb(
p cur-v ity oft threem ipalc eo riva-t lightw suf-fi 4).T f twoA dt erall1 2).T Glusa1 oT lns tionsi fre-q spe-c re.T Mar-q ndh ss-i to fita etersw ctra.I som wasi fitsf stedt gavep forc 1638c uen-
(%)
321CHYMOTRYPSIN ADSORPTION ON MONTMORILLONITE
.2 unit. The average pD listed in Tables 1 and 2 resultsets of two, three, or even four, reproducible experimehese pD values slightly differ from solutions to suspensihus, comparisons between solution and adsorbed statese performed for pD that could differ by 0.2 unit. For 4,D , 9, the FTIR spectra of the supernatant for thea-chymo-
rypsin–clay suspensions showed that more than 95% orotein was adsorbed on the clay. For pD 11.8, half ofrotein remained in the supernatant.FTIR transmission spectra were recorded on a Perkin-E
720 spectrometer equipped with a DTGS detector. Resolas set at 4 cm21, using boxcar. A Balston air dryer frohatman (UK) strongly reduced the water vapor in the s
rometer during the measurements. Solutions were inserteCaF2 cell with a 50-mm spacer. Suspensions of the so
hases were inserted between two CaF2 plates with a 25-mmpacer, allowing easy cleaning. For each sample numpectra were collected from five min to 3 h. Time zeroetup when the buffer was added onto the solid protein oolid protein–clay mixtures. Inside the spectrometer, aftein, the temperature was stable and the air dryer reachaximum efficiency.Spectral differences were computed in the 1800- to 1
m21 wavenumber range, on one hand, between proteolutions and the corresponding buffers and, on the otheretween protein–clay mixtures and corresponding clay suions. Adjustments were made using a computed spectruOD (difference spectrum between pure D2O and partiallyydrogenated D2O spectra) to reduce as much as possibleOD contribution in the 1470-cm21 spectral range. The narrond residual H2O vapor bands were canceled by a slpectral smoothing before second-derivative treatments.nd-derivative and curvature calculations were performe
he difference absorption spectra and compared to self-dolution procedures to establish for all spectra the sameer of principal components in the 1800- to 1500-cm21 range38). The curvature analysis corresponds to a homemade
TABpD-Dependent Intensities of the Amide I Components,
uted analysis, based on a second derivative, to amplifyatures in a spectral profile and also to reduce the intenshe lateral wings for the detected bands. Indeed theethods gave for all spectra a similar number of princ
omponents at similar wavenumbers (61 cm21, or less). Somf the initial difference spectra and also some second-de
ive or curvature spectra are reproduced in Fig. 1. The savenumber shifts of the identified components were notcient to be valuably introduced in our spectral fitting (3hus the same sets of nine Amide I (Tables 1 and 2) and omide II components (1549 and 1532 cm21) were introduce
o compute the decomposition of all spectra. To fit the ov800- to 1500-cm21 range, other bands were included (Fig.hey correspond to identified side chains (39, 40). Asp andide chains give absorptions at ca. 1715 cm21 for nCOCOOH andt 1584 (Asp) and 1567 cm21 (Glu) for naCOO2, 1605 and518 cm21 are assigned to Tyr, and 1592 cm21 is assigned trp/His vibrations. The COND2 functions of the Asn and Gide chains (23 residues) should have significant contribun the Amide I range (40). However, the correspondinguencies are not predictable enough to valuably includeific Asn and Gln contributions in our curve-fitting proceduhe same least-squares iterative program (Levenberg–uardt), with a fixed empirical half-bandwidth at half-baeight (10 cm21) and also a fixed empirical Lorentzian/Gau
an profile (0.25/0.75), has been applied (32, 33, 38, 41)ll difference spectra. These constant band shape paramere empirically chosen to give the best fit for all the spe
ntensities were the only allowed variations (Fig. 2). Withany constraints, although the number of components
mportant, computed intensities resulting from successiveor one spectrum were always identical. We have also tehat decomposition of smoothed or unsmoothed spectraractically the same results, with only 0.4% variationomponents with the closest wavenumbers (1644 andm21) and even less for the others. The constancy in freqies and band shapes for any spectral decomposition allo
2pressed as Percentage of the Overall Amide Intensity,in Buffered Dideuterium Oxide Suspension
e ge oft ro-g a)r ther ithin1 Dido ties( oree ithc ying,t ples.S se of2 De
oteinb ms onalm ideI iesa hiftsw tionw er,os ide Ic mideI s areg s int ithpm llc ctralr s, ca.1 dedp The1 Thee licals n ofa eter-m em -c le,2 ofC isa stal-l vedi le-s H-b forb l anda as-
-t riuo ctrs
322 BARON ET AL.
omparative quantitative analysis of intensity changes foromponent from one spectrum to another.Protein flexibility is related to the rate of the peptide NH/
xchange in D2O. This exchange is evaluated by the measent of the decrease of Amide II intensities (residual CO
FIG. 1. Absorbance spectra (1500–1750 cm21) for chymotrypsin in soluion or adsorbed on montmorillonite in suspension in buffered dideutexide for three pD ranges (A); second-derivative and “curvature spepectra for pD 4.5 (B). Solution (—); adsorbed state (- - -).
NHs
ch
-
ntities) as a function of time and is expressed as percentahe overall Amide I intensity (32, 33, 41). For totally hydenateda-chymotrypsin (solid state), Amide II/Amide I (areeaches;40% (spectrum not shown). In solution, most ofandom peripheral polypeptide segments are exchanged w0 min. Thus, the exchange depends on the diffusion of2O
nto internal domains. In all cases, 2 h aftera-chymotrypsinissolution, the NH/ND exchange involved more than;65%f the polypeptide backbone. The Amide II/Amide I intensi%) plotted in Fig. 3 are averaged values from two or mxperiments prepared with the same buffer (cf. earlier). Wareful control of time, temperature, and spectrometer drhere was less than 2% variation between equivalent samincea-chymotrypsin consists of 241 residues, a decrea% in the Amide II/Amide I ratio should involve NH/Nxchange for;12 peptide units.Structural and solvation parameters evaluated for the pr
ackbone rely on the decomposition of the 1700- to 1610-c21
pectral range which mostly corresponds to the vibratiotions of the CO peptide groups (Amide I for CONH, Am
9 for COND) (42–45). Indeed, in our study COND speclways dominated the CONH. Since no major frequency sere observed as a function of time (cf. earlier), no distincas made for Amide I for CONH or COND entities. Moreovur analyses result from comparisons betweena-chymotrypsinamples having a similar extent of exchange. Each Amomponent is expressed as a percentage of the overall Aintensity. The assignments of these Amide I componentiven in Table 1 (32, 33, 39, 41–47). The two weak band
he 1690- to 1680-cm21 spectral range are associated weptide CO in hydrophobic environments (32, 33, 41).a-Chy-otrypsin contains a large amount ofb-sheets; thus a sma
ontribution of these structures is also expected in this speange (see further) (34, 39, 44). The next two component670 and 1660 cm21, are assigned to non-hydrogen-boneptide units in loops in polar environments (46, 47).651-cm21 component is assigned to the helical domains.stimated amount of CO peptide groups involved in the hetructure, ca. 15%, is in agreement with the weak proportiomino acid residues involved in such a structure (12%) dined by X-ray spectroscopy fora-chymotrypsin (30, 31). Thajor band at 1638 cm21 and another at 1630 cm21 are asso
iated withb-sheets. Adding both contributions, for examp5 and 16% fora-chymotrypsin at pD 5.6, the proportionO peptide groups H-bonded inb-sheets reaches 41%. Thislso in good agreement with X-ray spectroscopy for cry
ized chymotrypsin, giving 52% amino acid residues involn double- or triple-strandb-sheets (30, 31). Indeed, in doubtrand b-sheets, only half of the CO peptide units areonded. The relatively wide frequency range accounting-structures could arise from the presence of both parallentiparallel extended strands (30, 31, 44). An alternativeignment is proposed by Casalet al. (48): the 1630-cm21
ma”
c tos ti cob rotm eei ea tein
( f the1 ofb
d inF eforea ntsf ns oft tides rcent-a eachs asf toa ationa f2 threep
a
nte tot sixp eas-i s thefl
mina endc3 stal-l fta 0- to1 .
ins
tho onm ioup
323CHYMOTRYPSIN ADSORPTION ON MONTMORILLONITE
omponent could correspond to externalb-sheets exposedolvent and then hydrated. At ca. 1618 cm21 a weak componen
s assigned to external polypeptide extended strands thate self-associated to other external strands of adjacent polecules (42, 48–51). Accounting for the correlation betw
ncreased absorption in the 1645- to 1635-cm21 spectral rangnd increased rate of NH/ND exchange for several pro
FIG. 2. Computed decompositions of absorbance spectral profiles (uspension (B) in buffered dideuterium oxide at pD 4.5.
FIG. 3. Time dependence of the Amide II intensity (percentage ofverall Amide I intensity) for chymotrypsin in solution (A) or adsorbedontmorillonite in suspension (B) in buffered dideuterium oxide at varDs.
1
uldeinn
s
32, 33, 39, 41), we propose to ascribe the increase o644-cm21 component fora-chymotrypsin to an increaseackbone hydration in random domains.Intensities (%) reported in Tables 1 and 2 and plotte
igs. 3, 9, and 10 correspond to averages (cf. earlier). Bveraging, variations in intensity of the Amide I compone
or equivalent samples were less than 1%. Weak variatiohe Amide I molecular extinction coefficient from one peppecies to another are expected. Thus, the calculated peges cannot precisely indicate the absolute amounts ofpecies. However, any difference of61% or more measuredunction of time, from one pD to another, or from a solutionn adsorbed state is assumed to be significant of solvnd/or structural changes. Sincea-chymotrypsin consists o41 residues, such a change of 1% should involve two oreptide units.
RESULTS AND DISCUSSION
-Chymotrypsin in Solution
NH/ND Exchange from 10 min to 2 h.The time-dependevolution of the Amide II/Amide I intensities (%) relative
he amount of residual CONH units is plotted in Fig. 3A forD values. At any given time, this ratio decreases for incr
ng pD. We conclude that the increase of the pD enlargeexibility of the chymotrypsin backbone.The increase in exchange from pD 4.5 to pD 5.5 at 10
rises from the deprotonation of the three carboxylichains and the external Asp and Glu carboxylic groups (pK 5.3–4.3)displayed in the schematic representation of cry
ized a-chymotrypsin (tosylated), PDB 2CHA (from Birktond Blow (31)). The decrease in absorbance in the 170750-cm21 range and the increase of thensCOO2 bands at ca
21
0–1500 cm21) for chymotrypsin in solution (A) or adsorbed on montmorillonite
e
s
175
584 and 1567 cm (Fig. 1) illustrate this deprotonation. The
s t pD5a fec wlyd uriA ewi fm
cet . Ap n ai th1 ins( aiub ins( eep sidc unwm
arb , thN vef hee mie
p em -194i lowi argedw latea Theu avityw orep
uche thee y
,2 hemt 102A -24c Va9 in thX
,2 d theA threei 0.
,2 um/i 6, and
324 BARON ET AL.
lope of the curve from 10 to 45 min (Fig. 3A) is steeper a.5 than at pD 4.5. During this period, the residualnCOCOOD
bsorption decreases (not shown). This implies that thearboxylic functions still protonated at pD 4.5 are sloeprotonated at pD 5.5. They should correspond to the bsp-102, Asp-194, and Glu-70 side chains (Fig. 4). N
nternal hydrophilic sites (COO2) would favor the diffusion oore water molecules into protein cavities.For pD 7.5 the exchange before 10 min is strongly enhan
hereafter the curve is parallel to that for pD 5.5 (Fig. 3A)D 7.5 all carboxylic functions are deprotonated at 10 mi
ndicated by the absence of any residual absorption in740- to 1700-cm21 range. The two His imidazole side chaHis-40 and His-57), protonated at pD 5.5, should remnprotonated in aqueous medium at pD 7.5 (pK ; 6). Saltridges between His1 and the internal carboxylate side chaAsp-194, Asp-102, and Glu-70) that could be formed betwD 5.5 and 6.5 (Fig. 5) should be disrupted when Hishains become neutral. This apparently enhances the amoater inside the enzyme core, allowing the exchange of;30ore peptide units.At pD 8.6, in agreement with the deprotonation of all c
oxylic groups and imidazole side chains, as for pD 7.5H/ND exchange at 10 min is not increased further. Howe
rom 20 min up to 2 h a significant enhancement of txchange is observed once more. At pD 7.5 the three and terminal groups (Cys-1, Ile-16, and Ala-149) should
FIG. 4. Structural representation of crystallizeda-chymotrypsin (ref PDBCHA). Sticks schematize His-40 and -57 residues. Balls and sticks sc
ize all Asp and Glu carboxylic (carboxylate) residues (only internal Asp-sp-194, and Glu-70 are explicitly numbered), Tyr-146 residue, Asnarboxylic (carboxylate) end chain, and Pro-8 residue. Location of the–Leu-13 carboxylic (carboxylate) end chain segment is not provided-ray 3-dimensional structure (31).
bA
w
ed
d;tse
n
net of
-er,
noe
rotonated (pK ;8.3) (Fig. 6). At pD 8.6 they should bostly unprotonated. Indeed, in the crystalline state, Asp
s close to the Ile-16 amino end chain (Fig. 5). The sncrease of the exchange could depend on a specific enlater solvation if a salt bridge between Asp-194 carboxynd the aminium Ile-16 chain cannot be formed (52).ncompensated negative charge inside the enzymatic could entail greater water diffusion exchanging about 20 meptide units.Similarly, an increase of the pD to 11.6 does not have m
ffect on the NH/ND exchange at 10 min. This agrees withxternal position of all Lys (pK 5 10.5), which should sta
a-,5l-e
FIG. 5. Structural representation of crystallizeda-chymotrypsin (ref PDBCHA). Sticks schematize His-40 and -57 residues, Cys-1, Ile-16, anla-149 amino (aminium) end chain. Balls and sticks represent the
nternal carboxylic (carboxylate) residues: Asp-102, Asp-194, and Glu-7
FIG. 6. Structural representation of crystallizeda-chymotrypsin (ref PDBCHA). Sticks schematize all His, Lys, and Arg (amino/imino or amini
minium) residues and the three amino/aminium end chains: Cys-1, Ile-1la-149.
d oi oft ccu( esg der
tabf thrw ctioa ptids fav ai
-p 0, 46 ouA rys soa an1
inpv tern wae itht eraH slod occ Fig3
mi( ed
b ainu lightd i-r .T andI
minttao . Ina thisc Tyrr 1%vm l ands
d at2 n ofp s iss thep orh ndedc withp self-a as-s dent.T of thea reaseo twoe af-f sidec
f the
,2 rest
,2 -155r
325CHYMOTRYPSIN ADSORPTION ON MONTMORILLONITE
eprotonated rather than protonated at lower pD without mfying the internal hydration (Fig. 6). A fast deprotonationhe two external Tyr side chains (Fig. 7) should also opK 5 10.1). From 20 min up to 2 h the exchange becomreater than at pD 8.6. This is associated with the slowotonation of the two internal Tyr residues (Fig. 7).
At pD 12.4 a weak supplementary exchange is detecrom the first measurements. A partial deprotonation ofelatively external imino Arg-154 chain (pK 5 12.5) couldeaken a bridge with the adjacent Asp-72 carboxylate funnd separate two close and relatively hydrophobic petrands (Fig. 8). Such a structural transition would haveored an immediate exchange of the adjacent peptide dom
Secondary structures and hydration.The spectral decomositions were performed for spectra recorded at 10, 20, 30, and 120 min at 6 pD. The evolution of the intensity of fmide I components (%) representative ofb-sheet secondatructures, direct backbone hydration, and protein self-astion, as a function of time, are only illustrated for pD 4.51.6 in Figs. 9A and 9B, respectively.For pD 4.5 the intensity of all Amide I components rema
ractically constant between 10 min and 2 h (less than 1%ariation), meaning that the structural equilibrium with exal carboxylate and internal carboxylic side chains (Fig. 4)stablished within 10 min. For pD 5.5 some variations w
ime were noted (not shown). While the relatively periphis residues are assumed to be rapidly protonated, aeprotonation of the internal carboxylic functions shouldur, as inferred from the study of the NH/ND exchange (A).For pD 7.5, because equilibrium was attained within 30
not shown), internal Asp and Glu side chains were assum
FIG. 7. Structural representation of crystallizeda-chymotrypsin (ref PDBCHA). Sticks schematize His-40 and -57 residues; balls and sticks rep
he four Tyr residues.
d-
r
p-
lee
ne-
ns.
5,r
ci-d
s
-s
lw-.
nto
e rapidly deprotonated and His imidazole groups to remnprotonated. For pD 8.6, between 10 min and 1 h, a secrease ofb-structure (1638 cm21) was compensated by dect hydration of;3–4 peptide units (1644 cm21, not shown)his could reflect slow structural changes at the Asp-194
le-16 levels (Fig. 5).For pD 11.6, some instability was also observed from 10
o 1 h (Fig. 9B). A small amount ofb-structure (1638 cm21) isransformed into another kind ofb-folding (1630 cm21) andlso into directly hydrated peptide units (1644 cm21). Theverall slow structural variation involves 5–7 peptide unitsgreement with the NH/ND exchange results (Fig. 3A),ould be related to slow deprotonation of the internalesidues (Fig. 7). As before, there was no more thanariation between 1 and 2 h. Thus, at any studied pD,a-chy-otrypsin in solution was assumed to reach a structura
olvation equilibrium after 2 h.Only the intensities of the Amide I components obtaineh are displayed in Table 1. They are plotted as a functio
D in Fig. 10. The evolution of the secondary structurehown in part A. The evolution of the solvation state ofeptide backbone, with “free” (polar and hydrophobic)ydrogen-bonded carbonyls, is shown in part B. The H-boarbonyls are bonded either with water molecules oreptide –NH– from adjacent protein molecules (proteinssociation). The intensities of the Amide I componentsigned to free internal carbonyls are almost pD indepenhese domains are not influenced by the ionization statescidobasic side chains of the protein. The very weak decf the a-helix content when pD increases means that thexternal helical regions in chymotrypsin are not strongly
ected by electronic charges carried by the acidobasichains (Figs. 4 and 6).From pD 4.5 to pD 5.6 and 7.5, the slight decrease o
ent
FIG. 8. Structural representation of crystallizeda-chymotrypsin (ref PDBCHA). Sticks schematize His-40 and -57, Arg-154, Phe-71 and Leuesidues; balls and sticks represent Asp-72 residue.
c ectp enf -n –1p het omp
g ef ngem astsw ons( om-i te,o anA
nt at1 able1 t(( nlyi s, ap ign-mb ydra-t Atp saltb ailedt
d andb thet d allL hesec Tyrr truc-t otont osei daryt tc ews ree-mi sedh –30i
-f e-m soci-a weenw Thisi ns ah ngea nectt . Al vorp 8.6,t nt,i teina the
s e ins .
326 BARON ET AL.
omponent at 1644 cm21 means a weak reduction of the direptide hydration. In the meantime, both Amide I compon
or b-structures, at 1638 and 1630 cm21, increase simultaeously (Table 1, Fig. 10A). Such changes involve 8eptide units in all. The local folding is apparently all set w
he deprotonation of internal carboxylic side chains is c2
FIG. 9. Time-dependent Amide I components (%) fora-chymotrypsin inolution, pD 4.5 (A) and pD 11.6 (B), or adsorbed on montmorillonituspension, pD 4.5 (C) and pD 11.8 (D), in buffered dideuterium oxide
leted. An internal electrostatic interaction between the COOc
ts
2n-
roup of Asp-194 and the NH31 group of Ile-16 could hav
avored this constraint (Fig. 5). The larger NH/ND exchaeasured from pD 4.5 to pD 5.5 to pD 7.5 (Fig. 3A) contrith this result. The deprotonation of the carboxylic functi
17 functions, including the 3 end chains) (Fig. 4), all becng highly hydrophilic, could, as for the NH/ND exchange ravercome the effect of the local folding depending onsp-194. . . Ile-16 salt bridge.On the contrary, at pD 8.6, the increase of the compone
644 cm21 demonstrates an increase of direct hydration (Tand Fig. 10B). This is connected with a decrease inb-shee
1638 cm21 in Table 1). Indeed, the other kind ofb-structure1630 cm21) slightly increases (Table 1). The changes onvolve 8–12 peptide units in all (Figs. 10A and 10B). Thuarallel strand becoming antiparallel is unlikely. The assent of the 1630-cm21 band to hydratedb-domains (48) is inetter agreement with the simultaneous direct backbone h
ion (Fig. 10A) and the faster NH/ND exchange (Fig. 3A).D 8.6, in comparison to pD 7.5, the disruption of theridge between Ile-16 and Asp-194 (Fig. 5) could have ent
hese structural and solvation changes.From pD 8.6 to pD 11.6 the contribution ofb-structures
ecreases strongly (52) and hydration increases in random-domains (Table 1 and Figs. 10A and 10B). At pD 11.6,
wo His imidazole groups, the three amine end chains, anys side chains should stay unprotonated (Fig. 6). Thanges and a fast deprotonation of the two externalesidues (171 and 146) (Fig. 7) are assumed to involve sural changes in relatively peripheral regions. Then slow prransfers involving the internal Tyr-228 (Fig. 7) and the clnternal Arg-230 residues could lead to the slower seconransition between 10 min and 1 h (Fig. 9B). This differenharge distribution from that at pD 8.6 gives rise to ntructural and solvation features (Figs. 10A and 10B). In agent with a previous Raman study (52), theb-sheet unfolding
nvolves 10–15 peptide units. It is followed by an increaydration of ca. 10% of the backbone, meaning around 20
nternal peptide units.Compared to pD 11.6, only a very smallb-domain is un
olded at pD 12.4 (;3 peptide units). Internal hydration rains practically constant, but some more protein self-astion is observed. Self-association minimizes contacts betater and unfolded hydrophobic domains of the protein.
mplies that the extra unfolding at pD 12.4 rather concerydrophobic region. As proposed from the NH/ND exchanalysis, the partial Arg-154 unprotonation could discon
wo external neighboring strands in the protein (Fig. 8)arger flexibility of these hydrophobic domains would farotein self-association in solution. In the pD range 4.5–
he 1618-cm21 Amide I component is weak and constandicating that at the concentration used (5 mg/mL), proggregation is very weak, if any. Indeed, in this pD rangeomponent at 1618 cm21 could be assigned to a vibration
al
c sidc sl thei1 na oct
tionas thee .1e ent Gs lyp
ofa exc
fferedd rentiateds
327CHYMOTRYPSIN ADSORPTION ON MONTMORILLONITE
ontribution of the Tyr-OH side chains (40). When thesehains are deprotonated (Tyr-O2), the vibrational mode iowered to 1604 cm21 (53). Thus in the present case,ncrease of the absorbance at 1618 cm21 from pD 11.6 to pD2.4 when all Tyr side chains should be deprotonated isssigned to tyrosinate side chains but to protein self-ass
ion.
Correlation between NH/ND exchange, structure, solvand enzymatic activity. The pH dependence ofa-chymotryp-in catalytic activity is presented in Fig. 11. Below pH 5nzyme has no activity. Our results for pD 4.5 (pH 4mphasize that such inactivity is not related to protein d
uration but to the protonation state of the internal Asp andide chains. Carboxylate forms are essential in the catarocess (54–56).From pH 5 to 7 (pD 5.4 to 7.4), the hydrolytic activity
-chymotrypsin increases progressively. The NH/ND
FIG. 10. pD-dependent Amide I components (%) fora-chymotrypsin inideuterium oxide (2 h): secondary structures (A, C) (forb-sheets the contribuolvated domains (B, D).
solution (A, B) or adsorbed on montmorillonite in suspension (C, D) in bution of hydrated and unhydrated domains (Tables 1 and 2) are added); diffe
hange, and thus water diffusion, was also increased. Accos
e
otia-
,
)a-lutic
- FIG. 11. pH-dependent hydrolysis of BTNA witha-chymotrypsin inolution or witha-chymotrypsin adsorbed on montmorillonite in suspens
rd- ion.
i op tivW ine itA eew 8.3tmt rupt ug-g
Oua s au gea othm tont tem olaf ativc et tha
a
-d idI ril-l
ds p t2 per owe oc
ngi filea witr ed.t , ba oe ec tont thea ecf sev ouh
frot wep ans
u rptiond
teine clayn aterc waterr sorp-t
steri thana om-p st eed,f tant( ex-c struc-t
t erea d 120m ide Ic 9D,r
iredt ons( ains,t 49a f thep clay( ND
(a kss 16 andA andA
328 BARON ET AL.
ng to the pH dependence of the activity, only the fractionrotein with unprotonated imidazole is assumed to be achen the maximum of the activity is reached, the Ile-16 am
nd chain is apparently still protonated. Its interaction wsp-194 appears to draw a specific favorable folding. Indhen the structural contraction is released at pD 8.7 (pH
he catalytic activity decreases. The catalytic activity ofa-chy-otrypsin is known to depend on pH-dependent His1 proton
ransfers (54–56). A local unfolding depending on the dision of an Asp-194. . . Ile-16 salt bridge has also been sested (57).At pH above 10.5 the enzyme has no more activity.
nalysis in the 11.2–12 pH range (pD 11.6–12.4) confirmnfolding ofb-secondary structures (52). Most of the chanre assumed to be in relatively peripheral domains, butore internal changes could be connected with the depro
ion of the two internal Tyr residues (Fig. 7). Too many waolecules inside the catalytic cavity could interfere with p
unctional groups of the substrate (56). Indeed, a negharge on the Tyr-146 end chain, not far from the entranche catalytic cavity (Fig. 7), might have also canceledffinity of the enzyme for its substrate.
-Chymotrypsin Adsorbed on Montmorillonite
NH/ND exchange from 10 min to 2 h.The time depenence of the amount of residual CONH peptide units (Am
I/Amide I (%)) for a-chymotrypsin adsorbed on montmoonite in the presence of D2O is shown in Fig. 3B for 5 pD.
At pD 4.5, the exchange is reduced at 10 min compareolution, whereas, in contrast, it is favored from 30 min uh. Adsorption is assumed initially to prevent relatively
ipheral domains of the protein from contact with water. Hver, the greater exchange later (2 h) proves that water mules finally penetrate further inside the protein.For pD 5.9, in comparison to pD 4.5, the NH/ND excha
nitially increases (10 min). Then the time-dependent prore parallel. Initially adsorption reduces the exchangeespect to solution (pD 5.5); then the exchange is enhanche first stage, adsorption protects weakly buried domainsfter 1 h, internal domains finally encounter more water mcules than in solution. For the solutions, the enhancedhange from pD 4.5 to 5.5 was associated with the deproion of the internal carboxylic side chains (Fig. 4). Fordsorbed state, the intensity of thenCO absorption of tharboxylic groups (1700–1750 cm21) as a function of pDollowed that in solution (Figs. 1 and 2). Based on this obation, the time dependence of the NH/ND exchange shave been similar for the two states.Once again, at pD 7.7 adsorption prevents some NH
he exchange for the first hour. However, in contrast to loD, after 2 h, the exchange is similar for the adsorbedolution states. Such similar exchange for the buried pe
ptie
fe.ohd,),
-
rnsera-
rre
ofe
e
too--le-
eshInutl-x-a-
r-ld
mrd
de
nits emphasizes that specific features induced by adsoo not involve buried domains of the protein.At pD 8.7, at 10 min the exchange for the adsorbed pro
xceeds that for the protein in solution, meaning that theo longer protects external domains of the protein from wontact. The similar exchange measured after 2 h, wheneaches internal domains, confirms that, as at pD 7.7, adion involved only external domains of the protein.
For pD 11.8, the NH/ND exchange was also initially fan the presence of montmorillonite, but to a lesser extentt pD 8.7. The difference rapidly disappeared (30 min). Cared to the solution (pD 11.6), at 2 h adsorption even seem
o prevent exchange in buried or hydrophobic regions. Indor this pD we have found half of the protein in the supernacf. Experimental). We assume that protein moleculeshanging on the electronegative surface are, on average,urally influenced by this surface.
Secondary structures and hydration (2 h).For a-chymo-rypsin adsorbed on montmorillonite, decompositions wlso performed for spectra recorded 10, 20, 30, 45, 60, anin at five pD values. The time dependencies of the Am
omponents for pD 4.5 and 11.8 are plotted in Figs. 9C andespectively, as examples.
At pD 4.5 (Fig. 9C) and 5.9 (not shown), the time requo attain structural equilibrium is longer than for the solutiFig. 9A). Several external protonated Lys and Arg side chhe two peripheral His1 side chains, and the flexible Ala-1minium end chain, which are oriented on the same side orotein, should be rapidly adsorbed by the electronegativeFigs. 6 and 12). From 10 to 40 min, given the slower NH/xchange for the adsorbed state (Fig. 3B), the amount of
FIG. 12. Schematic representation of an orientation fora-chymotrypsinref PDB, 2CHA) adsorbed on montmorillonite below pH 7 via His1, Lys1,nd Arg1 side chains and Ala1-149 aminium end chain. Balls and sticchematize His residues; sticks represent Lys and Arg residues and Ile-la-149 aminium end chains. Only His and Arg residues and Ile-16la-149 aminium end chains are explicitly numbered.
water
w n ir tha thp nalr itht B)A
thea ). At ateb ainp oteb thip get
n ib Fig9 Tys tio( 40m sine
igs1 lvat oup . Ai lesc utio( pede ad
tlyr Thl .H en oe sec ter nfA Thp np egt tera ths leta ptiu tras
get C)F
s .1 orp-t theH thec tatici niume -t
2 hi l ofs d Fig.1 la-t erh 7.5( 0C).W ed),t Theo thei llow-i da
rt localb sim-i andF theN tates( endab
thea ationa hisc itivec obaln lec-tr hains( ithn otheroc g itsd ge inn
omep d in-v (33,4
tion,a -z ed
329CHYMOTRYPSIN ADSORPTION ON MONTMORILLONITE
hich reaches internal hydrophilic domains of the proteieduced by the steric hindrance of the clay. Thus, fordsorbed state, a slow conformational modification ofrotein, from 30 min to 1 h, finally favors hydration in interandom domains andb-domains (Fig. 9C), corroborated whe larger NH/ND exchange (from 40 min to 2 h, Fig. 3dsorption also entails some protein self-association.For pD 7.7, as for the solution, the conformation of
dsorbed protein was stabilized at ca. 30 min (not shownhis pD the His imidazole groups should remain unprotonut the external Ala-149 and internal Ile-16 amino end chrotonated (Fig. 6). At pD 8.7, some fluctuations are netween 30 and 40 min (not shown). As for the solution inD range, these fluctuations are related to proton exchan
he internal Ile-16 aminium/amine end chain.For pD 11.8, the slow enhancement of the hydratio
-structures and in random domains from 10 to 30 min (D) is related to a slow deprotonation of the two internalide chains, as in solution (Fig. 7). In comparison to soluFig. 9B), the reduction of the slow internal hydration (fromin to 2 h) confirms an influence of the clay on chymotrypven in this pD range (Fig. 9D).The Amide I intensities given in Table 2 and plotted in F
0C and 10D provide information on the structure and soion of chymotrypsin adsorbed on montmorillonite at variDs when structural equilibrium has been reached (2 h)
ncrease of the pD from 4.5 to 11.8 obviously induceshange for the adsorbed protein than for the protein in solFig. 10). This is good agreement with the weaker pD deence of the NH/ND exchange after 2 h (Fig. 3). As for thexchanges, the larger effects of the clay on chymotrypsinetected at pD below 7.7 (Tables 1 and 2 and Fig. 10).At pD 4.5, helical domains (Fig. 8) are only very sligh
educed (Fig. 10C) compared to the solution (Fig. 10A).arger unfolding involvesb-sheets of;10–15 peptide unitsydrated b-sheets (1630 cm21) become dominant over thonhydrated ones (1638 cm21) and direct hydration is alsnhanced (1644 cm21) (Tables 1 and 2 and Fig. 10). Thehanges apparently take place in two steps: a first staeached within 10 min and a second step lasts to 1 h. Theeatures cannot arise from a different COOD/COO2 ratio forsp or Glu side chains in the adsorbed state (cf. earlier).rimary changes are associated with a rapid orientatioositively charged external functions toward the electron
ive surface (Figs. 6 and 12). An optimization of these inctions with the mineral surface over 40 min would explainlower secondary structural transition. This is in compgreement with the progressive unshielding of some penits, which are easily exchanged once all the structuralitions induced by adsorption are established.For pD 5.9, the unfolding resulting from adsorption is lar
han at pD 4.5 (;20–25 peptide groups) (Figs. 10A and 10or the solutions, increasing pD from 4.5 to 5.6 enta
ileb
see
.
tdsdsat
n.rn
,
.-
snsn
n-
re
e
isew
eofa--eeden-
r.d
pecific dehydration and weakb-folding (Table 1 and Figs0A and 10B). This folding is apparently restricted by ads
ion (Table 2 and Figs. 10C and 10D). The orientation ofis1 side chains and Ala-149 aminium end chain towardlay surface could hinder cooperative internal electrosnteractions between imidazoles, carboxylates, and amind groups (Fig. 5), favoring theb-folding observed in solu
ion.At pD 7.7, although water diffusion inside the protein at
s similar to what happens in solution (Fig. 3), the leveecondary structure remains weaker (Tables 1 and 2 an0C). This confirms that most of the unfolding involves re
ively peripheral regions (;10–15 peptide units). On the othand, the local folding observed in solution at pD 5.9 andFig. 10A) is now set up for the adsorbed state (Fig. 1
hen His imidazole groups are unprotonated (uncharghey are no longer oriented toward the clay surface.rientation of the adsorbed protein could have enabled
midazole side chains to adopt structural arrangements ang the formation of the Asp-194. . . Ile-16 electrostatic bons in solution (Fig. 5).At pD 8.7, the level of unhydratedb-sheet remains lower fo
he adsorbed state (Tables 1 and 2), but, as in solution, the-unfolding is observed. Thus, direct hydration becomes
lar in the solution and adsorbed states (Tables 1 and 2igs. 10B and 10D). These features occur just whenH/ND exchange profiles are the closest for the two s
Fig. 3). In both cases, the deprotonation of the Ile-16minium chain should entail the same Asp-194. . . Ile-16 saltridge disruption.For pD 11.8, the small differences between solution and
dsorbed state arise from slight reductions of direct hydrnd hydratedb-structures (Tables 1 and 2 and Fig. 10). Tonfirms that besides a greatly reduced number of posharges around the globular protein and in spite of its glegative charge, chymotrypsin is still influenced by the e
ronegative clay. This could be via the external Arg1-230esidue located on the opposite side of many Lys side cFig. 12). Compared to pD 8.6, at pD 11.8, the protein wumerous deprotonated Lys side chains should adopt anrientation on the clay. Specific interactions via Arg1 sidehains could specifically shield an internal domain, reducinirect hydration, on one hand, and the rate of the exchanearby domains, on the other hand (see earlier).All along the 4.5–11.8 pD range adsorption entails s
Correlation between NH/ND exchange, structure, solvand enzymatic activity. Adsorption completely inhibits enyme activity below pH 7. The catalytic activity is reduc
etween pH 7 and 8.5 but recovers above pH 9 (Fig. 11).
I ini ripe oun .
zya odfi erap -rn that ona rgep
ymac pDE dc tep sa ano letp hem enzm gep yma ntat tionc rst th
p f thec ce( tiono ldr ctedw hisa vityii ion oft 9,a tranceo
ity isl ace,h ed onm ichh tion,w tives s tom tionw s isf iche for thes
pro-t , thea ana back-b pro-t for ap in ana entali nd ont llines ences fast,a de ofs ionc –64)o
ce tionsc ultsf epro-t n oft andh efa For
(c ndr
330 BARON ET AL.
ndeed a major result of our analysis is that the unfoldnduced by adsorption is weak and involves essentially peral domains of the protein (10–15 peptide units). This shot perturb the structure at the level of the enzymatic siteThus, the mechanism responsible for the decrease of en
ctivity on adsorption is probably not a pH-dependent mcation of conformation as previously described for sevroteins: bovine serum albumin (26);Hebeloma cylindrospoum acid phosphatases (6); sweet almond (13) andApergillusiger (16) b-D-glucosidases. In this model, it is assumed
he protein unfolds below its isoelectric point, due to strttractive electrostatic forces between the positively charotein and the negatively charged clay surface.Since adsorption has practically no effect on the enz
ctivity above pH 9, unfolding associated with Lys1 adsorptionannot be responsible for the inhibition observed at lowerlectrostatic interactions between these side chains anlay should not hinder the access of the substrate tonzymatic site (Fig. 13). For pH lower than 6.5 (pD; 7), theresence of the clay close to the two His1 imidazole groupnd to the Ala-149 aminium end chain, located at the entrf the enzymatic cavity, could be the reason for comprotein inactivation (Fig. 12). However, just above pH 7, wost of the His side chains should be deprotonated, theatic activity is far from entirely recovered. In this pH ranroton exchange at the imidazole sites, favorable for enzctivity, could maintain the adsorbed protein with an orie
ion still shielding the catalytic site. Indeed protein orientahanges should be very slow compared to proton transfehe nitrogenous imidazole. Moreover, in this pD range
FIG. 13. Schematic representation of an orientation fora-chymotrypsinref PDB, 2CHA) adsorbed on montmorillonite above pH 8 only via Lys1 sidehains. Balls and sticks schematize His residues; sticks represent Lys aesidues and Ile-16 and Ala-149 amine end chains.
p
gh-ld
mei-l
tgd
e
.thehe
ceeny-
,e-
one
rotonated Ala-149 aminium end chain, at the entrance oatalytic cavity, could still interact with the mineral surfaFigs. 5 and 12A). For increasing pH (pD 7.7), the orientaf the protonated Ile1-16 end group toward Asp-194 couesult from a change of the protein orientation, as expeith a partially deprotonated Ala-149 amino function. Tllows some recovery of enzyme activity. In solution, acti
s effectively optimum when the salt bridge Asp-194. . . Ile-16s set up. For the adsorbed state, the complete deprotonathe Ala-149 amino group, which is only expected at pDppears to be a prerequisite for complete release of the enf the enzymatic cavity from the mineral surface.Such a mechanism, where decrease of catalytic activ
inked to an orientation of the active site toward the surfas already been described for a ribonuclease A adsorbica (58). Ribonuclease A is a “hard” protein (59, 60) whas a relatively large dipole moment of its charge distribuith the positive end pointing toward the region of the acite (61). With time, a translation diffusion process leadolecular reorientation of ribonuclease. A side-on adsorpith the active site facing the mica surface in initial stage
ollowed by a progressive “standing up” of the enzyme, whxposes the active site to the solution and hence accessubstrate.
CONCLUSIONS
In contrast to the complete 3-dimensional structure ofeins that can be established using NMR spectroscopypproach by FTIR spectroscopy is currently limited toverage description of protein secondary structures andone solvation states. However, with FTIR, much larger
eins can be studied and structural parameters obtainedrotein in solution can be compared to results obtaineddsorbed state (32, 33, 41, 62–66). This has a fundam
mportance since protein properties, in many cases, depeheir direct environment. When the 3-dimensional crystatructure of the protein is known, it can be used as a refertate. FTIR analyses then become a high-performance,nd cheap experimental technique to evaluate the amplitutructural deformations resulting from local ionizathanges, mutation (32), or adsorption on electrostatic (62r hydrophobic solid phases (32, 33, 41).In the case of the well-knowna-chymotrypsin proteolyti
nzyme, the present transmission FTIR analysis of soluonfirms that the optimum catalytic activity of ca. pH 8 resrom the convergence of numerous parameters: (i) the donation of the carboxylic side chains, (ii) the deprotonatiohe two His side chains, increasing both protein flexibilityydration, and (iii) a localb-sheet folding resulting from th
ormation of a salt bridge between the Ile1-16 end chainminium group and the Asp-194 side chain carboxylate.H . 10, the complete inhibition of the catalytic activ
Arg
ity
s turu bua int venh isd
rani ei 162 inc tras pDr lvea -i ths e ina t be oua tiow a1 ogn 8.5w aset hica ins ripec rtai ena hedF fieli vityA othi ama ess ateA . Inr est o-t cot –25 rmt
pt.
ic
m-
111111 J.
1 .
11
1
22 .2 .
22
22
2 G.,
2
2 oto-s.ds.),gton,
3
33
3
3 hey,
3 . C.,
3 sc.
33
3
4
331CHYMOTRYPSIN ADSORPTION ON MONTMORILLONITE
hould result not only from peripheral secondary strucnfolding directed by external Lys or Tyr deprotonationslso from internal Tyr deprotonations entailing excessive
ernal hydration in the vicinity of the catalytic center. For eigher pH, partial aggregation of the protein in solutionetected.Our analysis also demonstrates that over the entire pD
nvestigated adsorption ofa-chymotrypsin on montmorillonits mainly determined by electrostatic interactions (6, 13,6, 27, 29). When the protein and the clay are broughtontact, these interactions result in a dynamic structuralition of the protein which can last up to 1 h. In the 4.5–9ange, the unfolding resulting from the adsorption invobout 15–20 peptide units in peripheralb-sheets. Protein flex
bility and hydration are slightly enhanced with respect toolution phase. However, the pH-dependent profile of thctivation of the enzyme in the 5–9 pH range cannoxplained by these changes alone. Most of the inhibition wrise from a steric hindrance of the clay due to its interacith positively charged His (40 and 57) imidazole and Al1-49 end chain aminium that control the initial specific recition of the enzyme for its substrate. At pH higher thanhen the charges at His and Ala-149 are completely rele
he enzyme is adsorbed with a different orientation wllows a recovery of activity, similar to that measuredolution in the same pH range. This infers that neither peral unfolding nor partial self-association ofa-chymotrypsinancels the enzyme activity. These findings have impomplications for the immobilization of active proteases orntioselective proteins on various supports (to be publisurthermore, they are of fundamental interest in research
nvolving the pH dependence of reversible enzymatic actit pH . 11, the enzyme activity is completely inhibited b
n the adsorbed state and in solution. We propose a dynnd nondenaturing exchange between protein moleculolution and those only weakly retained on the clay via isolrg1 side chains to account for their structural similarities
elation to the stability of proteins at solid–liquid interfachese findings puta-chymotrypsin in the class of “hard” preins, such as bovine pancreas ribonuclease (59, 60). Inrast, “soft” proteins, such as bovine serum albumin (269, 60, 63) and some phytases (61), show larger confo
ional modifications on adsorption on solid surfaces.
ACKNOWLEDGMENTS
The authors thank Dr. S. Staunton for critically reading the manuscri
REFERENCES
1. Burns, R. G.,in “Soil Enzymes” (R. G. Burns, Ed.), p. 295. AcademPress, New York, 1978.
2. Theng, B. K. G.,in “Formation and Properties of Clay–Polymer Coplexes” (B. K. G. Theng, Ed.), p. 157. Elsevier, Amsterdam, 1979.
et-
ge
,ton-
s
e-eldn
-,d,
h
h-
nt-).
ds.
icind
,
n-9,a-
3. Burns, R. G.,Soil Biol. Biochem.14, 423 (1982).4. Tarafdar, J. C., and Jungk, A.,Biol. Fertil. Soils3, 199 (1987).5. Tarafdar, J. C., and Marschner, H.,Soil Biol. Biochem.26, 387 (1994).6. Leprince, F., and Quiquampoix, H.,Eur. J. Soil Sci.47, 511 (1996).7. Mousain, D., Matumoto-Pintro, P., and Quiquampoix, H.,Rev. For. Fr.
49, 67 (1997).8. Plassard, C., Chalot, M., Botton, B., and Martin, F.,Rev. For. Fr.49, 82
(1986).0. McLaren, A. D.,J. Phys. Chem.58, 129 (1954).1. Armstrong, D. E., and Chesters, G.,Soil Sci.98, 39 (1964).2. Albert, J. T., and Harter, R. D.,Soil Sci.115,130 (1973).3. Quiquampoix, H.,Biochimie69, 753 (1987).4. Quiquampoix, H.,Biochimie69, 765 (1987).5. Fusi, P., Ristori, G. G., Calamai, L., and Stotzky, G.,Soil Sci. Soc. Am.
21, 911 (1989).6. Quiquampoix, H., Chassin, P., and Ratcliffe, R. G.,Prog. Colloid Polym
Sci.79, 59 (1989).7. Gianfreda, L., and Bollag, J. M.,Soil Sci. Soc. Am. J.58, 1672 (1994).8. Violante, A., De Cristofaro, A., Rao, M. A., and Gianfreda, L.,Clay
Miner. 30, 325 (1995).9. McLaren, A. D., and Estermann, E. F.,Arch. Biochem. Biophys.68, 157
(1957).0. Durand, G.,C. R. Acad. Sci.259,3397 (1964).1. Goldstein, L., Levin, Y., and Katchalski, E.,Biochemistry3, 1913 (1964)2. Aliev, V. S., Gusev, V. S., and Zvyagintsev, D. G.,Vestn. Mosk. Univ
Biol. Pochvovied.31, 67 (1976).3. Douzou, P., and Petsko, G. A.,Adv. Protein Chem.36, 423 (1984).4. Sandwick, R. K., and Schray, K. J.,J. Colloid Interface Sci.115, 130
(1987).5. Sandwick, R. K., and Schray, K. J.,J. Colloid Interface Sci.121,1 (1988).6. Quiquampoix, H., and Ratcliffe, R. G.,J. Colloid Interface Sci.148,343
(1992).7. Quiquampoix, H., Staunton, S., Baron, M. H., and Ratcliffe, R.
Pintro, P. T., Ratcliffe, R. G., and Staunton, S,in “Proteins at InterfaceII: Fundamentals and Applications” (T. H. Horbett and J. L. Brash, EACS Symposium Series No. 602, p. 321. Am. Chem. Soc., WashinDC, 1995.
0. Mattews, B. W., Sigler, P. B., Henderson, R., and Blow, D. M.,Nature214,652 (1967).
1. Birktoft, J. J., and Blow, D. M., PDB, 2CHA (1975).2. De Collongue, B., Sebille, B., and Baron, M. H.,Biospectroscopy2, 101
(1996).3. Boulkanz, L., Balcar, N., and Baron, M. H.,Appl. Spectrosc.49, 1737
(1995).4. Caughley, B. W., Dong, A., Bhat, K. S., Ernst, D. S. F., and Caug
W. S.,Biochemistry30, 7672 (1991).5. Dong, A., Matsuura, J., Allison, S. D., Chrismann, E., Manning, M
and Carpenter, J. F.,Biochemistry35, 1450 (1996).6. Nigles, M., and O’Donoghue, S. I.,Prog. Nucl. Magn. Reson. Spectro
32, 107 (1998).7. Glasoe, P. K., and Long, F. A.,J. Phys. Chem.64, 188 (1960).8. Baron, D.,in “Logiciels pour la chimie,” p. 282. Socie´te Francaise de
Chimie, Paris, 1991.9. Wantyghem, J., Baron, M. H., Picquart, M., and Lavialle, F.,Biochemistry
29, 6608 (1990).0. Venyaminov, S. Y., and Kalnin, N. N.,Biopolymers30, 1243 (1990).
4
4
444
4 .4 .4
4 .5 .5 .,
5 .5 , an
5 ica
555
5 .5666
6 nstas-echt/
6, F.,
66
332 BARON ET AL.
1. Boulkanz, L., Vidal-Madjar, C., Balcar, N., and Baron, M. H.,J. ColloidInterface Sci.188,58 (1997).
2. Arrondo, J. L. R., Muga, A., Castresana, J., and Goni, F. M.,Prog.Biophys. Mol. Biol.59, 23 (1993).
3. Byler, D. M., and Susi, H.,J. Ind. Microbiol.3, 73 (1988).4. Byler, D. M., and Susi, H.,Biopolymers25, 469 (1986).5. Surewicz, W. K., and Mantsch, H. H.,Biochim. Biophys. Acta952, 115
(1988).6. Baron, M. H., de Loze, C., and Fillaux, F.,Biopolymers11, 2063 (1972)7. de Loze, C., Baron, M. H., and Fillaux, F.,J. Chim. Phys.75,632 (1978)8. Casal, H. L., Ko¨hler, U., and Mantsch, H. H.,Biochim. Biophys. Acta957,
11 (1988).9. Jackson, M., and Mantcsh, H. H.,Biochim. Biophys. Acta1118,139 (1988)0. Jackson, M., Haris, P. I., and Chapman, D.,Biochemistry30,9681 (1991)1. Wetzel, R., Becker, M., Bahlke, J., Billwitz, H., Bo¨hem, S., Ebert, B
Haman, H., Krumbiegel, J., and Lassman, G.,Eur. J. Biochem.104,469(1980).
2. Heremans, L., and Heremans, H.,Biochim. Biophys. Acta999,192 (1989)3. Baron, M. H., de Loze, C., Mejean, T., Coulange, M. J., Turpin, P. Y.
Chinsky, L.,J. Chim. Phys.80, 729 (1983).4. Rawn, J. D., “Biochemistry” (N. Paterson, Ed.). Carolina Biolog
Supply Co., Burlington, 1989.
d
l
5. Blow, D.,Acc. Chem. Res.9, 145 (1976).6. Henderson, R.,J. Mol. Biol. 54, 341 (1970).7. Sigler, B. P., Blow, D. M., Mattews, B. W., and Henderson, R.,J. Mol.
Biol. 35, 143 (1968).8. Lee, C. S., and Belfort, G.,Proc. Natl. Acad. Sci. U.S.A.86,8392 (1989)9. Norde, W., and Lyklema, J.,J. Colloid Interface Sci.66, 257 (1978).0. Haynes, C. A., and Norde, W.,Colloids Surf. B2, 517 (1994).1. Yoon, B. J., and Lenhoff, A. M.,J. Phys. Chem.96, 3130 (1992).2. Pantazaki, A., Baron, M. H., Revault, M., and Vidal-Majdar, C.,J. Colloid
Interface Sci.207,324 (1998).3. Baron, M. H., and Quiquampoix, H.,in “Fifth International Conference o
the Spectroscopy of Biological Molecules” (T. Theophanides, J. Anasopoulo, and N. Fotopoulos, Eds), p. 109. Kluwer Academic, DordrNorwell, MA, 1993.
4. Matumoto-Pintro, P. T., and Quiquampoix, H.,in “Microbiologie Indus-trielle et Environnement” (B. Baleux, M. Demazeaud, C. DiviesGendre, and R. Moletta, Eds.), p. 195. Socie´teFrancaise de MicrobiologieParis, 1997.
5. Pistorius, A. M. A.,Spectrosc. Eur.7(4), 8 (1995).6. Heise, H. M., Marbach, R., Koschinsky, T., and Gries, F. A.,Appl.
