An investigation of the electronic and steric environments of tyrosyl residues in ribonuclease a and Erwinia Carotovoral -asparaginase through fluorescence quenching by caesium, iodide and phosphate ions

Biochimica et Biophysica Acta, 434 (1976) 297-310

© Elsevier ScientificPublishing Company,Amsterdam-- Printed in The Netherlands BBA 37360 AN INVESTIGATION OF THE ELECTRONIC AND STERIC ENVIRONMENTS OF TYROSYL RESIDUES IN R1BONUCLEASE A AND E R W I N I A C A R O T O V O R A L-ASPARAGINASE THROUGH FLUORESCENCE QUENCHING BY CAESIUM, IODIDE AND PHOSPHATE IONS

R. B. HOMER and S. R. ALLSOPP School of Chemical Sciences, University of East Anglia, Norwich, NR4 7TJ ( U.K.)

(Received September 22nd, 1975)

SUMMARY The fluorescence lifetimes and relative quantum yields of several derivatives of tyrosine are reported. The quenching of the fluorescence of these compounds by phosphate, caesium and iodide ions has been investigated; the encounter rate constants, calculated from the quenching parameters and lifetimes, show a clear dependence on the charges borne by quenchers and fluorophores. The ratio of the Stern-Volmer constants of iodide and caesium, ions of similar size, defines an electrostatic parameter sensitive to the charge of the fluorophore which can be evaluated without knowledge of the fluorescence lifetimes. The mean of the encounter rate constants for caesium and iodide ions defines a rate constant which is largely chargeindependent and is used to establish a steric parameter. The two parameters are used to investigate the tyrosine environment in bovine ribonuclease A (EC 3.1.4.23) and Erwinia carotovora L-asparaginase (EC 3.5.1.1). The quantum yield of L-asparaginase (0.12) is very high for a class A protein and may be associated with the absence of disulphide bridges. There was no evidence for more than one type of tyrosine residue from the quenching experiments with either enzyme, an observation which is attributed to efficient energy transfer amongst tyrosine residues. At pH values close to the isoelectric points of the enzymes the electrostatic parameter suggests that the environment of the quenchable tyrosines in L-asparaginase is somewhat more positive than in ribonuclease. In 1 ~ sodium dodecyl sulphate the tyrosine environment of L-asparaginase becomes markedly negative as expected. The steric parameter indicates a lower accessibility of the tyrosine residues in L-asparaginase than in ribonuclease; an illustrative calculation is provided linking the steric parameter with the number of exposed tyrosine residues by taking into account the greater collision frequency of the larger protein molecules and the encounter distance for quenching determined from charge effects on the quenching of the model compounds. The calculation suggests that three tyrosyl residues are accessible in ribonuclease, in good agreement with other studies, but in L-asparaginase the number increases from 0.4 at pH 5.73 to 0.8 at pH 9.16 suggesting a loosening of the enzyme structure at high pH.

298 INTRODUCTION The tryptophanyl fluorescence of proteins has been widely used as a probe ot tryptophan environment [1], it is sensitive to the accessibility of solvent [2] and quenching ions [3], the proximity of ionisable groups [4] and, in some enzymes, the binding of substrates and inhibitors [5]. In contrast, the fluorescence of class A proteins [6], which contain tyrosine but no tryptophan, appears to be rather insensitive to these perturbations, although the number of proteins of this type which have been studied is relatively small [1]. This difference is typified by a comparison of the Lasparaginases of similar physical and enzymic properties [7] from Escherichia coli B and Erwinia carotovora. The tryptophanyl fluorescence of the former enzyme has been shown [5] to be sensitive to pH and substrate and inhibitor binding whereas the tyrosyl fluorescence of the Erwinia carotovora enzyme, which contains no tryptophan, is, as reported in this paper, insensitive to these factors. Lehrer [3] has recently shown that iodide quenching experiments can be used to probe tryptophanyl accessibility in lysozyme and can provide information about the electrostatic environment of the quenched residues. We report an extension of this work to the tyrosyl residues of Erwinia carotovora L-asparaginase, ribonuclease A and model compounds using phosphate, iodide and caesium ions as quenchers. The data obtained from the quenching of the models by I- and Cs + ions enable an electrostatic factor which does not require a knowledge of fluorescence lifetimes, and a steric factor, to be defined, these are used to investigate the environment of quenchable tyrosyl residues in the proteins. EXPERIMENTAL Erwinia carotovora L-asparaginase (kindly provided by Dr. H. E. Wade) and Ribonuclease A (4 times crystallised, British Drug Houses Ltd.) were estimated at 280 nm using Exl~ -----6.1 [7] and 7.2 [8] respectively. L-Tyrosine, N-acetyl-L-tyrosine amide, N-acetyl-L-tyrosine ethyl ester and L-tyrosine ethyl ester were chromatographically homogeneous products from B.D.H. Ltd. N-acetyl-D-tyrosine and 4hydroxy-phenylethylamine were obtained from Sigma and 3-(p-hydroxylphenyl)propionic acid from Koch-Light Ltd. Inorganic chemicals were of "AnalaR" grade from B.D.H. Ltd. Absorption spectra were measured on a Cary 14 spectrophotometer, fluorescence spectra were measured with a Farrand spectrofluorimeter employing 5 nm bandpass slits on both monochromators; the sample compartment was thermostatically kept at 25 ~ 0.1 °C. Potassium iodide solutions contained 10 -4 M sodium thiosulphate to suppress I~- formation [3], their absorbances were checked and if necessary small corrections made to the fluorescence intensity to compensate for the inner filter effect at the exciting wavelength [9]. Quantum yields were determined relative to L-tyrosine using the method of Parker and Rees [9]. Fluorescence lifetimes were determined by the single-photon counting technique on an instrument in Dr. F. Wilkinson's laboratory by Dr. A. C. R. Thornton whose assistance is gratefully acknowledged. The data obtained after deconvolution analysis fitted satisfactorily to a single exponential decay in all cases.

299 RESULTS AND DISCUSSION

(a) Quantumyields and lifetimes The c o m p o u n d s listed in T a b l e I were c h o s e n to constitute a series which all c o n t a i n e d the para-substituted p h e n o l c h r o m o p h o r e characteristic o f tyrosine b u t b e a r i n g either positive, n e u t r a l o r negative charges at p H 7. All the c o m p o u n d s h a d e x c i t a t i o n a n d emission m a x i m a at 278 a n d 303 n m respectively. The q u a n t u m yields relative to tyrosine, 0.21 [10], are in g o o d a g r e e m e n t with the available literature d a t a [11, 12] which are given in T a b l e I. The fluorescence lifetime ot tyrosine (3.04 ns) is in r e a s o n a b l e a g r e e m e n t with a p u b l i s h e d value o f 2.6 ns [13] b u t significantly s h o r t e r t h a n the c a l c u l a t e d value o f 7.5 ns [12], 3 - ( 4 - h y d r o x y p h e n y l ) p r o p i o n i c acid a n d 2 - ( 4 - h y d r o x y p h e n y l ) e t h y l a m i n e also have m u c h s h o r t e r lifetimes t h a n those c a l c u l a t e d (7.0 a n d 8.0 ns respectively) f r o m the a b s o r p t i o n spectra a n d q u a n t u m yields [12]. The p o o r a g r e e m e n t between t h e o r y a n d e x p e r i m e n t m a y reflect the i n a d e q u a c y o f the c a l c u l a t i o n for the relatively b r o a d a n d w e a k a b s o r p t i o n b a n d s of these c o m p o u n d s [14]. The n a t u r a l lifetimes (Zo = ~/~) c a l c u l a t e d f r o m the r a d i a t i v e lifetimes, 3, a n d q u a n t u m yields, % v a r y f r o m 12 to 18 ns with a m e a n o f 14.1 -+- 1.9 ns. T h e c o n s t a n c y TABLE I QUANTUM YIELDS AND FLUORESCENCE LIFETIMES OF A SERIES OF DERIVATIVES OF L-TYROSINE, RIBONUCLEASE AND L-ASPARAGINASE AT 25 °C pH

Quantum yield

Fluorescence lifetime r (ns)

Natural radiative fluorescence lifetime To (ns)

L-Tyrosine N-Ac-L-TyrOEt

7.0 7.0

3.04 ! 0.05 0.90 i 0.10

14.5 4. 0.2 18.0 4- 4.0

N-Ac-L-TyrNH2 N-Ac-D-Tyr 3-(4-hydroxyphenyl)-propionicacid

7.0 7.0 7.0

1.32 ± 0.10 3.06 4- 0.05 2.90 4- 0.05

12.0 4- 1.0 14.6 5:0.2 12.1 i 0.2

2-(4-hydroxyphenyl)-ethylamine

7.0

3.05 4. 0.05

13.3 4.4-0.2

L-TyrOEt Ribonuclease

7.0 9.16 5.73 9.16 7.50 5.73 7.16 7.16 7.16

0.21 a 0.05 4- 0.01 (0.06 b) 0.11 ± 0.005 0.21 4- 0.005 0.24 4- 0.005 (0.20 ¢) 0.23 4- 0.005 (0.19 b, 0.22 c) 0.06 i 0.01 0.013 4- 0.002 0.015 4- 0.002 0.11 4- 0.01 0.12 4- 0.01 0.11 4. 0.01 0.13 ± 0.01 0.12 4- 0.01 0.15 4- 0.01

L-Asparaginase 1% Sodium dodecyl sulphate 7 M Urea 4 M Guanidinium chloride

a Teale and Weber [10]. b Cowgill [11]. c Feitelson [12]. a Calculated from ~o9 = z using zo = 14.1 ns. Blumberg et al. [19], pH not stated.

(0.85 i 0.15) d (1.9c) (1.9D 2.4 ± 0.4

(95~) (95D 20 4- 4

300 of the natural lifetime of a series of similar fluorophores is often assumed [3] and is to be expected where the substituents, as here, do not perturb the symmetry of the chromophore. L-Asparaginase has excitation and emission maxima at 278 and 303 nm respectively, its quantum yield, 0.12 at pH 7.5, is remarkably large in comparison with that of ribonuclease, determined as 0.015 in accord with the literature values [6,15] and insulin, 0.037 [15]. The quantum yield of L-asparaginase is not as great as that of tyrosine, as has been implied [16], but it is very close to that of N-acetyltyrosine amide which must be a better model of tyrosine in a peptide chain. The high quantum yield may be due to the absence of disulphide bonds which have been shown to quench tyrosine fluorescence in models [17]; peptide bonds which have also been suggested as quenchers of tyrosine fluorescence in proteins [18] would appear to be relatively unimportant quenchers in L-asparaginase. The measured fluorescence lifetime of L-asparaginase, 2.4 ns, leads to a natural lifetime of 20 ns which is rather larger than the mean of the natural lifetimes found for the models. Significantly longer natural lifetimes have been obtained for ribonuclease (95 ns) and insulin (35 ns) [19]. The most probable cause for the discrepancy is that the observed lifetimes refer only to the fluorescent tyrosyl residues whereas the observed quantum yields represent an average over all residues, some of which may be non-fluorescent. Another possibility is that the measured lifetime could be lengthened by energy transfer amongst the tyrosine residues. The quantum yield of L-asparaginase fluorescence is insensitive to pH in the range 3.5-9.2, (data not shown) neither is it affected by the addition of 10 -2 M aspartic acid. These observations are in marked contrast to the behaviour of the tryptophanyl fluorescence of the Eseherichia eoli L-asparaginase [5]. Denaturation of the Erwinia carotovora L-asparaginase by 1 ~-sodium dodecyl sulphate, 7 M urea or 4 M guanidinium chloride produced 10-20 ~ increases in quantum yield whereas the quantum yield of ribonuclease A increases by over 100 ~ on denaturation [20]. These data are in concordance with the high quantum yield of L-asparaginase in suggesting that the tyrosyl residues are subject to little static quenching in the native enzyme. (b) Phosphate quenching

The susceptibility of tyrosine fluorescence to quenching by hydrogen phosphate ions has been proposed as a probe of tyrosine accessibility in proteins [21]. We have examined the effect of HPO 2- on the fluorescence of the tyrosine model compounds and L-asparaginase and found quenching of fluorescence to occur. Collisional quenching is described by the Stern-Volmer [22] equation: (Fo/F) -- 1 = K o [Q]

(1)

where F0 and F are the fluorescence intensities in the absence and presence of quencher concentration [Q] respectively and Ko is the Stern-Volmer constant. Plots of the quenching data according to eqn. 1 showed pronounced curvatures for some of the compounds, Fig. la, suggesting the possibility of ground-state complex formation which has been postulated in analogous systems [21, 23]. We could find no evidence from changes in the absorption spectra or the fluorescence excitation or emission

301 10C

A

!B

"~

200

b

o o

o

150

c o

5£

.

o

'7e

~

o

c o

o

~

d

~

50

1

OC

UU~)

Q4

08

1

1'2

0

04

E Po4 -i

08

Fig. 1. Fluorescence quenching by hydrogen phosphate at pH 7.8, 25 °C. (A) Stern-Volmer plot, eqn. 1, (B) plot according to eqn. 2. The compounds investigatedare indicated by the letters against the line as follows: a, 4-hydroxyphenylethylamine;b, g-tyrosine;c, N-acetyl-D-tyrosine;d, L-tyrosine ethyl ester; e, N-acetyl-c-tyrosineamide; f, N-acetyl-L-tyrosineethyl ester; g, L-asparaginase(pH 7.5). wavelengths in the presence of phosphate to support ground state complex formation, however, the quenching data were analysed by the proposed equation [21, 23], eqn. 2.

(Fo/FI)- 1 [Q]

- kq -? K, + kq K, [O]

(2)

Here Ka is the ground state association constant and kq the collisional quenching constant; values of these constants obtained from the plots of Fig. lb are given in Table II. The data for both tyramine and tyrosine ethyl ester gave unsatisfactory plots with small negative slopes although the original Stern-Volmer plots were linear. No trend in the Ko or kq values with the molecular charge can be discerned, however, when the bimolecular rate constants for collisional quenching, k3, are calculated using the fluorescence lifetime k3 : KQM it is clear from Table II that the rate constants increase as the positive charge on the fluorophore increases. L-Asparaginase fluorescence is weakly quenched by phosphate ions with k3 an order of magnitude smaller than ka for the model compounds. There was no evidence that any fraction of the tyrosine residues remained unquenched at high phosphate concentrations (see below (d)). The low value of k3 could arise either through an electrostatic effect, if there was a high negative charge around the quenched residues amplifying the effect found with the model compounds, or through steric inaccessibility of the tyrosyl residues to collision with the quenching ion.

(c) Quenching by iodide and caesium ions, model compounds A separation of the effects of electrostatically and sterically induced inaccessi-

302 TABLE II FLUORESCENCE QUENCHING OF DERIVATIVES OF L-TYROSINE AT pH 7.8 AND OF L-ASPARAGINASEBY PHOSPHATE AT 25 °C Fluophore

Overall charge Constants from eqn. 1a:

Constantsfrom eqn. 2~:

KQ(M-') k3/109(M-''s-') k~iM-') Ka(M-1) L-Tyrosine Zero N-Ac-L-TyrOEt Zero N-Ac-L-TyrNHz Zero N-Ac-D-Tyr Negative L-TyrOEt Positive 2-(4-Hydroxyphenyl)ethylaminePositive L-Asparaginase (pH 7.5) Zero

9,3b 3.4 4.3b 7.8b 6.7 15.0 0.53

3.1 4.0 3.3 2.6 7.7 5.00 0.22

6.8 1.4 3.0 4.4

0.46 0.66 0.33 0.27

a Errors in these constants are estimated at 10~. b These values were obtained from curved Stern-Volmer plots by extrapolation to low quencher concentration. ¢ Errors in these constants are estimated at 20 %

bility to quenchers is possible in principle as the following consideration of the collisional process shows. The diffusion controlled rate constant k o for a collisional process between uncharged spheres is given by eqn. 3 [24] kD : 2 R T (2 + rF/ro + ro/rv)/3000

(3)

where r v and r o are the radii of the fluorophore and quencher respectively and ~ is the viscosity. Debye [25] has developed this for the case where the cellisions are between ions; the electrostatic effect on the rate constant may be expressed by eqn. 4. kD(io.s) ----ko6/(exp 6 -- 1)

(4)

where 6 ZFgQeZ/,~kTa, here z is the charge on the ions and a is their distance of closest approach, e is the electronic charge, e the dielectric constant of the solvent and k the Boltzmann constant. Quenching ions of opposite charge but similar radii and hence similar diffusion coefficients should have similar collisional rate constants (k3 + and k3- for positive and negative ions respectively) with a neutral fluorophore but divergent values when the fluorophore is charged. The ratio k 3 - / k 3 + is readily obtainable from the Stern-Volmer constants for the ions KQ- and KQ+ as the lifetime of the fluorophore, which in many cases is unavailable, cancels (eqn. 5). This is an advantage over methods employing only a single quenching ion [3] where knowledge of the lifetime is essential and steric factors are not readily separated from electrostatic ones. =

E = K o - / K o + ~---k3-/k3 + = exp zFe2/e k T a

(5)

The experimentally determinable ratio K o - I K o + defines an electrostatic parameter E which depends only on the charge of the fluorophore provided that the quenching ions are of similar size. Iodide and caesium ions provide an almost ideal pair of ions

303 in this respect, their crystal radii are I - , 2.16/~, Cs +, 1.69 A, with a radius of 2.30 A for the hydrated caesium ion [26]. Their atomic numbers I = 53, Cs = 55 are very close so that if fluorescence quenching depends on enhanced intersystems crossing due to the heavy atom effect collisions of both should be similarly effective [27]. The data obtained for the quenching of the tyrosine model compounds by potassium iodide and caesium chloride were analysed by the Stern-Volmer eqn. 1. In all cases linear plots were obtained; a selection of these is shown in Fig. 2 and the K o and derived k3 values are tabulated in Table III. All the data were obtained at pH 7.5, 25 °C and I ~- 0.8 adjusted with potassium chloride which did not quench.

4C

.c2

'

4C ~ c a

2£ f d

OO

d.2

d4

d6

d8

Fig. 2. Stern-Volmer plots for quenching by (A) iodide and (B) caesium at pH 7.5, 1 = 0.8 (KCI), 25 °C. The compounds are identified by the letters in the legend to Fig. 1 with the addition of: 3-(4-hydroxyphenyl)propionic acid, h.

Using 2.2 and 4.0 A for the radii of the quenching ion and the phenolic fluorophore respectively the encounter rate constant for the ions with the neutral fluorophores was calculated from eqn. 3 to be k o ~ 8.0" 109 M-1. s-1 which agrees with the experimental k3 values, Table lII, within factors of 2 for I - and 7 for Cs +. In all cases the observed value is less than that calculated and the difference between the two ions suggests that the collisions of Cs + with the fluorophore are not as efficient at quenching as those of I - and may indicate that other mechanisms of quenching apart from enhanced intersystems crossing, such as electron or proton transfer may contribute to the quenching by iodide [28].

304 TABLE III FLUORESCENCE QUENCHING OF DERIVATIVES OF L-TYROSINE BY CAESIUM AND IODIDE AT pH 7.5, 25 °C, I = 0.8 (KC1) Fluorophore

Overall charge

Caesium chloride a Potassium iodide E ~ ro

k~/lO ~

~o

ko'~/109

(M-1 .s-l)~

kdlO ~

(M-l) (M-I.s -1) (M-t) (M-l.s -1) L-Tyrosine N-Ac-L-TyrOEt N-Ac-L-TyrNH2 N-Ac-D-Tyr 3-(4-hydroxyphenyl)propionic acid 2-(4-hydroxyphenyl)ethylamine L-Tyr-OEt

Zero Zero Zero Negative

3.6 1.6 1.7 4.7

1.2 1.8 1.3 1.5

14.4 4.3 6.7 9.9

4.7 4.7 5.0 3.2

4.0 2.7 3.9 2.1

3.0 3.3 3.2 2.4

Negative 6.5

2.2

8.5

2.9

1.3

2.6

Positive 3.1 Positive 0.67

1.0 0.79

20.9 12.1

6.8 14.2

6.7 18.0

3.9 7.5

a Errors in these constants are estimated at 10% b E = k3-/k3

+.

c kD'M= (k3- + k3+)12. The distance of closest approach of the quencher to the fluorophore, a, can be calculated from eqn. 4 by using the data for the neutral and charged fluorophores in Table III. The value of a obtained is 9.5 :~ 1.0/k which is significantly larger than the sum of the radii of the fluorophore and quencher (6.2 A), this suggests that the quenching cross-section is greater than the radii imply, although it may arise through the blanketing effect of high ionic strength on electrostatic interactions. Eqn. 4 is strictly applicable to low ionic strengths only, however, it appears to give reasonable results at the higher ionic strength used here. High ionic strengths are necessary to make accurate measurements of the Stern-Volmer constants when z is short. The charge on the fluorophore is clearly reflected in the k 3 values as was observed for phosphate. In general the results for Cs + do not show as good discrimination as the iodide or phosphate results but the k 3 values for the quenching of the two positively charged fluorophores by iodide or phosphate are in poor agreement. The electrostatic parameter E (eqn. 5) amplifies the effect of charge on k3 found for the individual ions and falls into the reasonably distinct ranges of ~< 2.1 for negative, 2.7-4.0 for neutral and >~ 6.7 for positive fluorophores (Table III) and thus provides a parameter which can be used for determining the charge around quenchable fluorophores in proteins. An expression which excludes the effect of charge and which can subsequently be used to define a steric parameter may be intuitively defined as the average of kaand k3 +, kD' = (k3- q- k3+)/2. It can be demonstrated by substitution in eqn. 4 that this is an excellent approximation to the calculated encounter rate constant for uncharged species provided I ZFZo I ~< 1 with a ~ 6/k. The experimental results bear this out, in Table III where ko' is given it may be seen to fall in the range 2.4--3.9. l09 M - l - s -1 with a mean of 3.1. l09 M - l ' s -~ if the anomalously high value for Ltyrosine ethyl ester is excluded. Due to the greater effectiveness of iodide as a quencher there is a residual correlation of ko' with charge, this could be eliminated by increasing all the k3 + values for caesium by a factor k 3 - / k a + determined for the neutral fluoro-

305 phores, which would correct for the lower efficiency of caesium, but we have not done this as it does not materially affect the results in the later sections. The rate constant kD' can be used to define a steric parameter S = k~,p/kDM, the ratio of the charge independent collisional rate constant of the protein to that of the model compounds, this has the value of unity for complete accessibility and should be sensitive to the accessibility of the tyrosyl residues in the protein but largely insensitive to their electrostatic environment.

(d) Quenching by iodide and caesium ions, proteins The fluorescence of L-asparaginase and ribonuclease was quenched by iodide and caesium ions and the data obtained fitted eqn. 1, Figs. 3 and 4. The Stern-Volmer

e°

Q4

~"

%

02

c

o

d

O.C~ oo

o ~ d2

d4

~

O1

d~

d.8

Fig. 3. Stern-Volmer plots for the quenching of L-asparaginase fluorescence by iodide and caesium at I = 0.8 (KCI), 25 °C. The quenchers and conditions are denoted by the letters as follows: (a) Cs +, 1 ~ sodium dodecyl sulphate, pH 7.2; (b) I - , pH 9.16; (c) I - , 1 ~ sodium dodecyl sulphate, pH 7.2; (d) I - , pH 5.73; (e) Cs +, pH 9.16; (f) Cs +, pH 5.73.

2.d

"

°

u-°l~ 1C b

O.O

02

04 06 [Quencher] (M)

08

Fig. 4. Stern-Volmer plots for the quenching of ribonuclease fluorescence by (a) iodide and (b) caesium at I = 0.8 (KCI), 25 °C, pH 5.73 (Q) and pH 9.16 (O).

306 TABLE IV FLUORESCENCE QUENCHING OF L-ASPARAGINASE AND RIBONUCLEASE BY CAE. SIUM AND IODIDE AT 25 °C, I ~ 0.8 (KCI) Enzyme

pH Caesium chloride" KQ k3+/109 ( M - ' ) ( M - I ' s -')

L-Asparaginase

5.73 <0.03 <0.013 7.50 0.06 0.025 9.16 0.10 0.042

in 1~sodium dodecylsulphate 7.16 Ribonuclease 5.73 9.16

0.51 0.21 0.68 0.35 0.85 0.45

Potassium iodide'

Eb

kD,p S° (M-l"s-1) c (=

k o ' p / k D , M)

KQ k3_/109 (M-')(M-1"s -') 0.16 0.067 0.24 0.10 0.34 0.14 0.27 0.11 3,1 1.6 2.4 1.3

~>5.3 <0.04 4.0 0.063 3.4 0.092 0.5 4.6 2.8

0.16 0.98 0.87

<0.013 0.020 0.030 0.052 0.32 0.28

a Errors in these constants are estimated at 10~. b E -- k3_/k3+.

c ko,p = (k3_ + k3+)/2. with kD'M = 3.1.109 M-l.s -l,

d S ~ ko,p/kt,,M

constants given in Table IV show that in general the enzymes are less susceptible to quenching than the model compounds. The fact that the data for the enzymes gave linear plots in the Stern-Volmer equation implies that all their tyrosyl residues can be quenched at a sufficiently high quencher concentration and with the same SternVolmer constant. Lehrer [3] has derived a modified Stern-Volmer equation, eqn. 6, Fo/(Fo

--

F)

--

1 1 [Q]. fa Ko ~ fa

(6)

which enables the fraction of accessible fluorophores f~ to be obtained; the other symbols are the same as those in eqn. 1, Application of eqn. 6 to the quenching data yields a value of./a = 1 for both L-asparaginase and ribonuclease (Fig. 5) again suggesting that all the tyrosine residues are accessible to quenching. In contrast several tryptophan containing proteins have yielded lower values off~, despite yielding linear Stern-Volmer plots, suggesting only partial accessibility of the tryptophan residues [29] Our data for the iodide quenching of ribonuclease are not in agreement with Burstein's [30] who obtained a non-linear Stern-Volmer plot which he interpreted in terms of differing accessibilities of tyrosyl residues. A possible source of the discrepancy is the partial absorption of the exciting radiation by formation of iodine in the concentrated potassium iodide solutions, we found that the addition of sodium thiosulphate as suggested by Lehrer [3] effectively prevented this problem. The failure to delineate more than one class of tyrosyl residue could be taken to suggest that all the 48 tyrosyl residues in L-asparaginase and the 6 in ribonuclease are equally accessible to collision with the quenching ions. This is very unlikely. Titration studies have shown that only about 1 0 ~ of the tyrosyl residues of Lasparaginase ionise normally [16] (Homer, R. B. and Allsopp, S. R., in preparation) and there is much evidence that only 2 or 3 of the ribonuclease tyrosyl residues are accessible [31 ]. The apparent equivalence of the tyrosyl residues in the quenching ex-

307

o

20 16( i

i

u...°

u. °

120 u. °

~.o

~-°

80

4.0

21o

41o 1

Nf Fig. 5. Fluorescence quenching by iodide of L-asparaginase at pH 9.16 (©, right hand ordinate) and ribonuclease at pH 5.73 (0, left hand ordinate) both at I = 0.8 (KCI), 25 °C. The data are plotted according to eqn. 6, the intercept at 1/[KI] = 0 is unity for both plots.

periments could result from energy transfer amongst the tyrosyl residues which is known to occur from fluorescence depolarisation studies of ribonuclease [32]. Efficient energy transfer to a residue accessible to collisional quenching would made the detection of unquenched residues impossible in these experiments. Energy transfer may also occur amongst tryptophanyl residues [1, 32] but as the position of the tryptophan emission (unlike that of tyrosine) and hence the overlap integrals for energy transfer depend on environment, total transfer amonst the whole population is less probable and selective quenching of lysozyme fluorescence has been achieved by energy transfer to salicylamide [33].

(e) The accessibility and environment of the tyrosyl residues in L-asparaginase and ribonuclease The electrostatic parameter E increases for both enzymes as the pH is lowered, Table IV, this is analogous to the change in E observed for an increase in positive charge on the model compounds and for the proteins probably reflects the overall increase in positive charge with decreasing pH. The isoelectric point for L-asparaginase at high ionic strength is 7.65 [34] consequently at pH 7.5 the value of E should reflect the local charge around the quenched tyrosyl residues, the value of 4.0 is at the top of the range found for the neutral models. The larger Stern-Volmer constant found with the dinegative HPO]- anion than with the iodide ion for the quenching of Lasparaginase fluorescence (Tables II and IV) could be explained by the electrostatic effect of a mildly positive environment about the quenched tyrosine residues which would be compatible with the value of E reported. Ribonuclease has an isoelectric point of 9.5 and at the higher pH studied, 9.16, E = 2.81 which is significantly lower than for L-asparaginase and near the bottom of the range for neutral models. This result does not support the suggestion [30] that there is a positively charged lysyl residue near the quenched tyrosine. The most dramatic effect of charge exhibited by

308 the electrostatic parameter was with L-asparaginase in 1 ~o sodium dodecyl sulphate where caesium was a more effective quencher than iodide (Fig. 3) yielding E = 0.5, a value much lower than any obtained with the model compounds and compatible with the model of sodium dodecyl sulphate-denatured proteins proposed by Reynolds and Tanford [35] in which protein molecules are sheathed in the anionic detergent. The steric parameter S, Table IV, is very much smaller for L-asparaginase than for ribonuclease suggesting that the tyrosyl residues in the latter are much more accessible; the pH dependence of S for L-asparaginase is much greater than that for ribonuclease suggesting that the accessibility of the tyrosyl residues of L-asparaginases increases with pH; there is a further increase in accessibility in the presence of sodium dodecyl sulphate. The value of S depends on both the relative collision frequency of the quenchers with the proteins and models and the quenching efficiency of the collisions. The collisional rate constants can be calculated from eqn. 3 and will be larger for the proteins because of their greater radius, as the calculated ratios kDp/kuM ~ 4.2 for L-asparaginase (radius 36 A) and 2.4 for ribonuclease (radius 18 A) demonstrate. Although the proteins undergo more collisions per unit time with quenchers than free tyrosine does, the efficiency of these collisions in quenching tyrosine fluorescence will be much lower for proteins where tyrosine only occupies a very small part of the surface layer. In a simple model S would be proportional to the number of accessible tyrosyl residues NT with the proportionality constant equal to the product of the ratio of the collisional rate constants calculated above, and the probability of a collision with a protein hitting a tyrosyl residue. The latter can be approximated by the ratio of the effective volume of a tyrosyl residue to the volume of the surface shell within which quenching can occur. The charge dependence of k3 led to a distance of closest approach of 9.5 ~: 1.0 A of which 7.3 A_can be apportioned to the tyrosyl residue if the quenchers have a radius of 2.2 A. Treating the tyrosyl residues and protein molecules as spheres the fraction of the volume of the outer spherical shell, thickness 7.3 A, occupied by a tyrosyl residue is 0.042 in ribonuclease and 0.0085 in L-asparaginase. Hence for ribonuclease S = 2.4.0.042 Nx and for L-asparaginase S ~ 4.2. 0.0085 Nx. Although these calculations must be subject to considerable uncertainty the experimental value of S for ribonuclease suggests that 3 tyrosyl residues are in the surface layer which is in excellent agreement with other work [20, 31]. For L-asparaginase, substitution of the experimental value of S gives ~ 0.4--0.8 accessible tyrosyl residues, increasing with pH, compared with about 4 by titration [16]. The agreement here is fair especially if the increase in accessibility with pH, which may reflect a general loosening of the tertiary or quaternary structure, continues to higher pH where the titration result was obtained. CONCLUSIONS We have not been able to substantiate a report [30] of differential quenching of tyrosyl residues in proteins and concluded that energy transfer between the residues is important. Although the tyrosine fluorescence of L-asparaginase is insensitive to pH and substrate it could be quenched by inorganic ions. The quenching by phosphate left an ambituity as to whether the observed reduction in the bimolecular collisional

309 rate constant arose t h r o u g h electrostatic or steric factors. This ambiguity could be removed by the use o f Cs + and I - ions as quenchers t h r o u g h the definition of electrostatic and steric parameters f r o m studies with model compounds. Application o f the parameters to ribonuclease and L-asparaginase showed that the tyrosyl residues in the latter are much less accessible and that those that could be quenched were in a more positively charged environment than those o f the former enzyme. The accessibility of tyrosyl residues in L-asparaginase increased markedly with p H suggesting a loosening o f the structure at high pH. The use of Cs + and I - can be extended to other fluoroph0res. Recent papers have given applications to t r y p t o p h a n fluorescence in proteins [36] and preliminary work on tyrosine and tryptophan-containing proteins has appeared [37]. The principle o f using quenchers with opposite charges (iodide and pyridinium ions) to obtain electrostatic and steric information on the environment of fluorophores has been used successfully on micelles [38, 39] where the effects were even larger than those f o u n d here with sodium dodecyl sulphate-denatured L-asparaginase. Caesium is preferable to pyridinium ion as a positively charged quencher in aqueous solutions as it should have much less propensity to form ground state charge transfer complexes and it is applicable over the whole p H range without becoming deprotonated. ACKNOWLEDGEMENTS We are pleased to acknowledge an extramural contract f r o m the Ministry of Defence which supported this work. REFERENCES 1 Longworth, J. W. (1971) in Excited States of Proteins and Nucleic Acids (Steiner, R. F. and Weinryb, I., eds.), pp. 107-198, Plenum Press, New York 2 Kronman, M. J. and Holmes, L. G. (1971) Photochem. Photobiol., 14, 113-134 3 Lehrer, S. S. (1971) Biochemistry 10, 3254--3263 4 Shinitzky, M. and Goldman, R. (1967) Eur. J. Biochem. 3, 139-144 5 Homer, R. B. (1972) Biochim. Biophys. Acta 278, 395-398 6 Teale, F. W. J. (1960) Biochem. J. 76, 381-388 7 Marlborough, D. I., Miller, D. S. and Cammack, K. A. (1975) Biochim. Biophys. Acta 386, 576-589 8 Bigelow, C. C. and Sonenberg, M. (1962) Biochemistry 1, 197-204 9 Parker, C. A. (1968) Photoluminescence of Solutions, Elsevier, Amsterdam 10 Teale, F. W. J. and Weber, G. (1957) Biochem. J. 65, 476~82 11 Cowgill, R. W. (1967) Biochim. Biophys. Acta 133, 6-18 12 Feitelson, J. (1964) J. Phys. Chem. 68, 391-397 13 Chen, R. F., Vurek, G. G. and Alexander, N. (1967) Science 156, 949-951 14 Strickler, S. J. and Berg, R. A. (1962) J. Chem. Phys. 37, 814-822 15 Cowgill, R. W. (1964) Arch. Biochem. Biophys. 104, 84-92 16 Shifrin, S., Solis, B. G. and Chaiken, I. M. (1973) J. Biol. Chem. 248, 3464-3469 17 Cowgill, R. W. (1970) Biochim. Biophys. Acta 207, 556-559 18 Cowgill, R. W. (1964) Biochem. Biophys. Res. Commun. 16, 332-335 19 Blumberg, W. E., Eisinger, J. and Navon, G. (1968) Biophys. J. 8, A106 20 Cowgill, R. W. (1966) Biochim. Biophys. Acta 120, 196-211 21 Chen, R. F. and Cohen, P. F. (1966) Arch. Biochem. Biophys. 114, 514-522 22 Stern, O. and Volmer, M. (1919) Phys. Z. 20, 183-188 23 Moon, A. Y., Poland, D. C. and Scheraga, H. (1965) J. Phys. Chem. 69, 2960-2966

310 24 Smoluckowski, M. (1918) Z. Phys. Chem. 92, 129-168 25 Debye, P. (1942) Trans. Electrochem. Soc. 82, 265-272 26 Cotton, F. A. and Wilkinson, G. W. (1967) Advanced Inorganic Chemistry, 2nd edn., Wiley, London 27 Kasha, M. (1952) J. Chem. Phys. 20, 71-74 28 Watkins, A. R. (1973) J. Phys. Chem. 77, 1207-1210 29 Holmes, L. G. and Robbins, F. M. (1974) Photochem. Photobiol. 19, 361-366 30 Burshtein, E. A. (1968) Biofizika 13, 718-720 31 Donovan, L. G. (1963) Biochim. Biophys. Acta 78, 474-491 32 Weber, G. (1960) Biochem. J. 75, 345-352 33 Elkana, Y. (1968) J. Phys. Chem. 72, 3654-3661 34 Miller, D. S., Marlborough, D. 1. and Cammack, K. A. (1971) Proc. C.N.R.S. Int. Symp. Lasparaginase, Paris, No. 197, pp. 55-71 35 Reynolds, J. A. and Tanford, C. (1970) J. Biol. Chem. 245, 5161-5165 36 Bukolova-Orlova, T. G., Burstein, E. A. and Yukelson, L. Y. (1974) Biochim. Biophys. Acta 342, 275-280 37 Altekar, W. (1974) FEBS Lett. 49, 208-211 38 Pownall, H. J. and Smith, L. C. (1974) Biochemistry 13, 2590-2593 39 Pownall, H. J. and Smith, L. C. (1974) Biochemistry 13, 2594-2597