Fluorescence quenching studies of bovine growth hormone in several conformational states

154 Biochimica et Biophysica Acta, 955 (1988) 154-163 Elsevier

BBA33153

Fluorescence quenching studies of bovine growth hormone

in several conformat ional states

H.A. Havel, E.W. K a u f f m a n and P.A. Elz inga

Control Research and Development, The Upjohn Company, Kalamazoo, MI (U.S.A.)

(Received 21 March 1988)

Key words: Fluorescence quenching; Tryptophan environment; Protein folding

The use of steady-state fluorescence quenching methods is reported as a probe of the accessibility of the single fluorescent tryptophan residue of bovine growth hormone (bGH, bovine somatotropin, bSt) in four solution-state ¢onformatious. Different bGH conformations were prepared by using previous knowledge of the multi-state nature of the equilibrium unfolding pathway for bGH: alterations in denaturant and protein concentration yielded different bGH conformations (native, monomerie intermediate, associated intermediate and unfolded). Because the intramoleeular fluorescence quenching which occurs in the native st:~te is reduced when the protein unfolds to any of the other conformations, steady.state fluorescence intensity measurements can be used to monitor bGH unfolding as well as the formation of the associated intermediate. These steady-state intensity changes have been confirmed with fluorescence lifetime measurements for the different ¢onformational states of bGH. Fluorescence quenching results were obtained using the quenchers iodide (ionic), aerylamide (polar) and triehloroethanol (non.polar). Analysis of the results for native-state bGH reveals that the tryptophan environment is slightly non-polar (in agreement with the emission maximum of 335 run) and the tryptophan is more exposed to aerylamide than most native-state tryptophan residues which have been studied. The tryptophan is most accessible to all quenchers in the unfolded state, because no sterie restrictions inhibit quencher interaction with the tryptophan residue. The iodide quenching results indieate that the associated intermediate tryptophan is not accessible to iodide, probably due to negative charges inhibiting iodide penetration. The associated intermediate tryptophan is less accessible to all three quenchers than the monomerie intermediate tryptophan, due to tight padfing of molecules in the associated intermediate state.

Introduction

The study of protein structure and dynamics has received considerable attention lately from both the theoretical and an experimental view- point. One way to approach these studies experi-

Abbreviations: bGH, bovine 8rowth hormone, bST, bovine somatotmpin; pGH, porcine growth hormone; TCE, trichloroethanol; Gdn-HCI, quanidine hydrochloride.

Correspondence: H.A. Havel, Control Research and Develop- meat, The Upjohn Company, Kalamazoo, MI 49001, U.S.A.

mentally is to compare the structure and dynamics for different conformations of the same protein, such as the comparison of native and equilibrium folding intermediate structures. Such an approach can be followed by bovine growth hormone (bGH, bovine somatotropin, bSt), as it has been determined previously that the equilibrium folding process for bGH involves at least one monomeric and one self-associated intermediate [1-5].

Bovine growth hormone is a globular protein (191 amino acids, M r -22000) containing about 50~; a-helical structure [2] and a single tryptophan residue. The hormone has been shown to increase

0167-4838/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)

milk production in dairy cattle upon daily admin- istration. The three-dimensional structure of bGH has not been determined, but, recently, Abdel- Meguid et al. [6] have determined the structure of a genetically engineered variant of recombinant porcine growth hormone (pGH). The structure consists of four anti-parallel a-helices arranged in a left-twisted helical bundle; the tryptophan residue is located within one of the a-helical regions near the center of the molecule. Because the bovine and porcine growth hormones have only 18 amino-acid residue differences out of 191 (91% sequence homology) the three-dimensional structure of b o l l is assumed to be very similar to pGH.

The intrinsic fluorescence from tryptophan residues in proteins has often been used as a probe of protein structure and dynamics [7,8]. Recent studies have shown that unique structurel infor- mation can be obtained from studies of th, ~ . fluorescence quenching behavior of proteins upon the addition of extrinsic quenching agents [9,10]. Pro- teins which have only one tryptophan residue, such as bGH, are particularly interesting, because their fluorescence properties should be relatively easy to interpret. We have studied the steady-state quenching of the tryptophan residue in bGH by quenchers with different properties in order to evaluate the utility of fluorescence quenchin 8 methods in solving protein structure problems and to further our understanding of the equilibrium unfolding process for bGH. We also report the changes observed in steady-state fluorescence emission spectra and fluorescence lifetimes as bGH unfolds in Gdn-HCI.

The equilibrium unfolding process for bGH [1-5] can be represented as follows:

N ~ I ~ U

In

where N represents the native state; I the monomeric intermediate; I n the associated intermediate (n = 3 to 5), and U the unfolded state. The two intermediates are highly populated at 3.8 M Gdn- HCI or 8.5 M urea. The unfolding mechanism results in the observation of different denaturation mid-points for different probes of bGH unfolding [3].

155

Experimental

Materials Recombinant DNA-derived bGH was obtained

by expression in Escherichia coli carrying a temperature-sensitive runaway plasmid into which had been inserted the bGH gene sequence and a tryptophan promoter system [11]. The purification of bGH was done using the procedure described by Evans and Knuth [12]. Pituitary-derived bGH was obtained from A.F. Padow (UCLA Medical Center). Recombinant DNA-derived and pituitary-derived bGH were used interchangeably in these experiments. Ultrapure Gdn-HCI was obtained from Sehwarz/Mann. Quenchers were obtained from the following sources: potassium iodide from Fisher, acrylarnide from Sigma, and triehloroethanol from Kodak. All other reagents were analytical grade.

Methods Sample preparation. All solutions contained 50

mM ammonium bicarbonate buffer (pH 8.5) and were made by mixing appropriate ratios of bGH dissolved in ammonium bicarbonate buffer, buffer alone, and an 8.0 M Gdn-HCI solution containing buffer. To obtain samples of bGH in different conformational states, variations in Gdn-HCI and bGH concentration were used: native (0.05 mg/,',3 = 2.3 #M bGH, 0.0 M Gdn-HCI), mono- merle intermediate (0.05 mg/ml = 2.3 #M bGH, 3.8 M Gdn-HCl), associated intermediate (0.25 mg/ml= 11.4 pM bGH, 3.8 M Gdn-HCI) and unfolded (0.05 mg/ml= 2.3 #M bGH, 6.0 M Gdn-HCI). Separate samples were prepared for each quencher concentration.

bGH concentration. The bGH concentration was determined by measuring solution absorbance at 278 nm using an absorption coefficient of 15 271 M- l - e r a -1 [11.

Ultraviolet absorbance. Ultraviolet absorbance spectra were measured at ambient temperature (about 23 ° C) with an IBM 9420 spectrophotome- ter using 1-cm-pathlength cells.

Steady-state fluorescence. Fluorescence intensity measurements were made at 25.0"C with an SLM 8000C spectrofluorometer in the photon-counting mode using 1 cm x 4 mm cells. Excitation of the tryptophan residue alone was assured by using an

156

excitation wavelength of 295 or 300 nm; an emission wavelength of 350 nm was used for all quenching studies. Both excitation and emission bandpasses were either 4 or 8 nm. Corrections were made for the wavelength-dependent response of the excitation monochromator by using a quantum counter (3 mg/ml Rhodamine B in eth- ylene glycol) and in the emission monochromator by using vendor-supplied correction factors. Blank (solvent) spectra were subtracted from all sample spectra.

The raw fluorescence intensity values for samples containing quencher were corrected for the background fluorescence of the solution by subtraction: from measurements of the fluorescence intensity for the hi,test and lowest quencher concentration solutions (without pr~t.ein) and assuming linear behavior, a background was calculated for each solution and subtracted from the raw fluorescence intensity value. A correction for the inner filter effect was also done using the following equation [7]:

Fcorr - Fobs antilog((Aex + Aem)/2 ) O)

where Fcorr is the corrected fluorescence intensity; Fobs is the observed (uncorrected) fluorescence intensity; Aex is the solution absorbance at the excitation wavelength; and Aem is the solution absorbance at the emission wavelength. The inner-filter correction was done after background subtraction.

Time.resolved fluorescence. Time-correlated single-photon counting techniques were used to obtain fluorescence fifetimes [13]. The excitation pulse train was generated with a picosecond laser system from Spectra-Physics. A frequency-doubled mode-locked Nd:YAG laser pumped a cav- ity-dumped dye laser which was further frequency-doubled to obtain excitation pulses with a repe- tition frequency of 4 MHz at 295 nm. The temperature of the sample was maintained at 20.0°C using a Neslab recirculating bath. The sample chamber and most electronic components were obtained from PRA International. The vertically polarized laser beam was routed through a Glan- Thompson polarizer and emission was detected at 54.7°C from vertical (the magic angle) through a Polaroid polarizer. The emission wavelength of

380 nm was selected with a Jobin-Yvon H-10 monochromator using a 4 nm bandpass in combi- nation with a Schott KV-370 non-fluorescing filter which passes wavelengths longer than 370 nm. The filter was used to eliminate contributions from scattered light that might pass through the monochromator. A Hamamatsu proximity-type microchannel plate photomultiplier tube (MCP- PMT) was used as detector. The emitted photon count rate was adjusted to about 3 kHz with a Newport variable attenuator in the excitation beam; this was done to eliminate pulse pile-up and changes in the impulse response of the MCP- PMT due to count rate. The time-to-amplitude converter was used in the 'reverse' configuration with start pulses from fluorescent photons detected by the MCP-PMT and stop pulses from laser pulses detected by a Spectra-Physics photo- diode. Decay curves were collected until at least 20 000 counts were accumulated in the peak chan- nel.

Analysis of the fluorescence intensity decay curves was done using the delta function decon- volution method [14] with 2,5-diphenyloxazole in cyclohexane (~ = 1.20 ns) as the reference fluorophore using our own software written with ASYST (MacMillan Software). The entire experimental decay curve (512 channels) was subjected to analysis for determination of lifetime parameters. Standard statistical tests for goodness-of-fit were used to determine the acceptability of the analysis results: X 2, Durbin-Watson parameter and the runs test [13].

Curve-fitting procedures for quenching results. Linear regression analyses were performed using standard techniques. The fits to the Stern-Volmer equation with static quenching constant (Eqn. 5) were performed using non-linear regression procedures available in ASYST (MacMillan Software).

Results and Discussion

Steady-state fluorescence spectra discriminate between bGH conformations

The steady-state fluorescence emission spectra for the tryptophan residue of bGH in different conformations are distinctive (Fig. 1). The spectrum of the associated intermediate state (I,) is not shown, because it exhibited the same emission

maximum wavelength as the monomefic intermediate (I), although the emission intensity was greater for I , (Table I). About a 2-fold increase in fluorescence emission occurred when the protein was unfolded which was accompanied by a shift in the emission maximum from about 335 nm (N) to 342 nm (I and I , ) to 352 nm (U); some broad- ening of the fluorescence emission spectrum was observed in the transition from the N, I and I , states (FWHM ffi 57 nm) to the unfolded state (FWHM = 64 rim, Table I). The same behavior was observed when bGH was unfolded with urea [4], except the emission maximum for the I and l,, states was closer to that of the native state (Dougherty, J.J., Jr. and Holzman, T~F., unpub- fished data). These results indicate that the native-state tryptophan residue is located in a moderately non-polar region of protein structure (as is the tryptophan in pGH [6]) which is close to an intramolecular quenching group such as a neighboring amino-acid residue side-chain of lysine, histidine, methionine or cystine. Upon initial unfolding from the native state to either of the two intermediate states, the interaction with the quenching group is diminished, but the tryptophan environment is still moderately non-polar. In the unfolded state, the tryptophan is surrounded by polar solvent. If both tyrosine and tryptophan residues were excited by using an excitation wavelength of 280 am (spectrum not shown), the native state spectrum was unchanged from that shown in

157

100

• ..'.."..'"~"t /" ...' '~.~.~ s ." 4 "..

J . " ~ o~.. \

g 7S , ." I.".. • . :

I= / ." "~ "- m / .: ~ ..:

: ' ~ - g / : •~ "x

SO ; ~ "... t~ i : / ~ _ ~'~ ...,...

/ - ~ ~ "'" "~. g ,,o~, "'..:

,"r ",- 25 ""~"'...

.."'

0 ' , , -- I , 30t 325 350 375 400 42!

Wavelength (am)

Fig. 1. Tryptophan emission spectra for native (N, - - - ) , monomedc intermediate (I, - . . . . . ) and unfolded (U, • . . . . . ) bGH at 0.01 m g / m l (0.5/tM) protein concentration. The peak near 390 nm is a spurious signal due to incomplete correction

in the emission monochromator.

Fig. 1, but the unfolded spectrum showed two maxima, one at about 310 nm (typical for tyrosine) and one at 350 nm (typical for tryptophan). Efficient energy transfer from tyrosine to tryptophan in the native state is likely the reason that a single emission band is observed in the native state. In the unfolded state, the energy transfer is less efficient, because of the increased separation of tyrosine and tryptophan residues, and two emission bands are observed, one for tyrosine residues and one for the tryptophan residue. Such

TABLE I

TRYPTOPHAN FLUORESCENCE PARAMETERS FOR BOVINE GROWTH HORMONE IN DIFFERENT CONFORMA-

TIONAL STATES

Conformationa! Steady-state Time-resolved o

state E m i s s i o n R e l a t i v e B a n d w i d t h f] ~'] f2 'r2 f3 1"3 ( ' r ) X2 ?~max (nm) a Quantum Yield b (rim) c (ns) (ns) (ns) (ns)

N 335 0.4 57 0.17 0.11 0.47 1.24 0.36 4.95 2.38 1.59 I 342 1.0 58 0.03 0.05 0.51 1.25 0.46 5.21 3.03 2,,42 In 342 0.7 56 0.03 0.10 0.36 1.33 0.61 4.99 3.53 2.98 U 352 0.8 64 0.05 0.49 0.39 1.74 0.56 4.11 3.00 1.47

Unfolded proteins e > 350 . . . . 0.39 1.7 0.61 4.2 3.2 -

" Estimated uncertainty :!: 2 rim. b Integrated emission intensity from 300 to 450 rim, estimated uncertainty :i:0.1. c Full width at half maximum, estimated uncertainty + 2 nm. d Results from fit to Eqn. 2: f~ = a i l r J (Za i~ ) ; (~') -- Z f ~ . e Average results for 10 proteins at red edge of emission spectrum [17].

158

a phenomenon has been observed previously by Brand and Cagan for monellin [15] and by Khan et al. [16] for thermolysin.

The equilibrium denaturation curve for monomeric bGH monitored with tryptophan fluorescence at 350 is shown in Fig. 2; the denaturation curve is similar to that reported previously using 280 um excitation and 340 nm emission to monitor bGH unfolding [3]. No significant differences were observed between denaturation curves using a single emission wavelength and those generated by integrating the emission intensity from 300 to 450 rim. The fluorescence intensity results shown in Fig. 2 were not corrected for the, slight dependence of tryptophan emission intensity on Gdn-HCI concentration (see caption to Fig. 2), but this does not affect the conclusions drawn. The monomeric intermediate exhibited the largest fluorescence intensity, about 2.5-times the native state emission and displayed slightly more fluorescence emission than the unfolded state (Ta- ble I). This latter result indicates that the monomeric intermediate is more shielded from both intramolecular quenching and quenching from bulk solution constituents than is the unfolded state. The formation of the associated inter-

I= | 7s

e ¢,) t : ¢D

| o s0 n-

2S !

o.o 1 .o £0 3:0 ,:o s'.o e:o Gdn HCI Concentration (M)

Fig. 2. Equilibrium denaturation curve for 0.01 mg/ml (0.5 FM) bGH as monitored by tryptophan fluorescence emission at 350 nm. No correction was made for the slight effect of Gdn-HCI concentration on tryptophan emission: a 1 /zM NATA solution exhibited a linear increase in fluorescence

emission with a shallow slope of about 5~ per M Gdn-HCI.

100

80

60

0.0 012 0 4 O.S 0 8 1.0

bGH Concentration (mg/ml)

Fig. 3. Effect of bGH concentration on tryptophan fluorescence emission at 350 nm in 3.8 M Gdn-HCI. The emission intensities have been corrected for bGH concentration and therefore represent molar tryptophan emission at each protein

concentration.

mediate can be monitored with tryptophan emission by increasing bGH concentration at 3.8 M Gdn-HCI (Fig. 3, [4]); the results in Fig. 3 were obtained by monitoring a single emission wavelength (350 nm), but, since the emission spectra for the I and I , states exhibited the same width and maximum wavelength (Table I), the results would not be significantly different if the integrated intensity from 300 to 450 nm were reported. The associated intermediate has about 70~ of the emission intensity of the monomeric intermediate on a molar basis (Table I), perhaps due to intermolecular quenclfing from neighboring molecules in the associated state, but its emission maximum is identical to that shown in Fig. 1 for the monomeric intermediate.

Fluorescence decay curves indicate complex tryptophan photophysics

Measurements of tryptophan fluorescence life- tim~ allow further discrimination of the struct- ural differences between bGH conformations; lifetime measurements are also necessary for the calculation of bhnolecular quenching rate constants (see below). Intensity decays have been obtained for the native, monomeric intermediate, associated intermediate and unfolded states of

bGH. The decay curves were analyzed as a sum of exponentials:

I ( t ) =~.a , exp(- t/~) (2) i

where ~,~ is the fluorescence lifetime of component i with amplitude ai. The best fits for all bGH conformations were obtained assuming three components (Table I). Previous workers have used flash lamp sources and conventional photomultiplier tube detectors to show that native state lifetimes of tryptophan in a large number of proteins are sensitive to differences in protein structure [8,17], but all unfolded proteins exhibit two tryptophan lifetimes of about 1.7 and 4.2 ns [17]. Our results were obtained with a picosecond laser as the excitation source and a MCP-PMT for detector. This system yields an impulse response which more closely resembles a &function than a flash lamp-PMT system (FWHM of about 0.5 ns for a laser-MCP-PMT system vs. 8 ns for a flash lamp- PMT system), thus we have more accurately ex- tracted short lifetime contributions to the fluorescence decay curves. As a result we have observed three components for the unfolded state of bGH. Two of the components in the unfolded state are virtually identical to those observed by Grinvald and Steinberg [17] for other proteins in the unfolded state, and the third component con- tributes only about 5~ to the observed fluorescence and exhibits a short lifetime (0.5 ns).

We chose to analyze our decay curves using discrete lifetime components although we are aware that this model has been questioned lately [18-22] with the proposal made that a better description would be a continuous distribution of lifetime components. The distribution could be interpreted as arising from the effects of restricted internal rotation of the tryptophan residue when it is covalently bound to the polypeptide chain. Dif- ferent rotational conformers of the tryptophan could have different distances from an intramolecular quenching group and hence exhibit different fluorescence lifetimes. Such a model has been proposed by Tanaka and Mataga [23] for horse apocytochrome c and Tanaka et al. [24] for erabutoxin B, for example. Our primary purpose in these experiments, however, is to determine an

159

average lifetime in order to calculate bimolecular quenching rate constants and it is unl/kely that the deterrrfination of average lifethne depends on the specific model used to fit the observed decay Curves.

Fluorescence quenching results reveal tryptophan accessibility

Quenching of the intrinsic fluorescence of proteins by the addition of extrinsic agents is governed by the ability of the quencher to get into close contact with a fluorophore [7]. The access can be restricted by either steric or chemical barriers to quencher penetration. The steric barriers are not rigid (they fluctuate due to conformational oscilla- tions of the protein backbone on the nanosecond timescale) and act to inh/bit the penetration of large quenching agents through the protein matrix. The chemical barriers are due to the chemical environment in which the fluorophore is located: quenching agents will only penetrate those regions of protein structure which do not present unfavorable energy barriers to penetration, i.e., regions of protein structure which are chenucaUy compatible with the quencher molecule; an example would be the preference of non-polar quenchers for non-polar regions of protein structure.

Robbins et al. [25] have recently shown the utility of using three different quenchers to probe the accessibility of the tryptophan residues in terminal deoxynueleotidyl transferase; each of the three quenchers has a quenching efficiency (~,) close to 1.0 [9] and differs in molecular size and chemical nature, and hence the protein presents different steric and chemical barriers for each quencher° We have followed the same approach in our studies of the single tryptophan in bGH and have used the ionic quencher, iodide [26], the non-ionic polar quencher, acrylamide [27], and the non-polar quencher, trichloroethanol [28]. The re- suits were analyzed using the Stern-Volmer equation for dynamic quenching [7]:

Fo/F= 1 + Ksv [Q] = 1 + kqr0 [Q! (3)

where F and F 0 are the fluorescence intensities with and without quencher; Ksv is the Stern- Volmer quenching constant; [Q] is the quencher concentration, kq is the bimolecular quenching

160

TABLE II

PARAMETERS FOR QUENCHING OF bGH TRYPTOPHAN FLUORESCENCE BY IODIDE, ACRYLAMIDE AND TRI- CHLOROETHANOL a

Conform. Iodide Acrylamide Trichloroethanol

state Ksv kq KSV kq Ksv kq V (M- ]) (M -1) (×10 9 IV[- 1. s -1 ) (M -1) (X10 9 M-I . s -1 ) (M -1 ) (X10 9 M-I . s -1 )

N 1.9±0.1 b 0,80+0.06 2.7±0.2 1,1 4-0.1 1.8±0.5 0.75±0.21 - I 2.0±0.4 b 0,7 +0.1 4.4±0.1 1,4±0.04 3.2±0.6 c 1.0 -I-0.2 5.3±0.3 c I , 0.7+0.1 0.19+0.02 2.7±0.2 0,8±0.04 2.6±0.8 c 0.75±0.23 6.6±0.5 c U 2.2±0.1 0.72:t:0.05 7.2+0.2 2,4+0.06 7.4±0.2 2.5 ±0.1 -

a All Ksv results from linear regression fit to Eqn, 3 ± S.D., unless noted otherwise. b Linear regression for KI quenching of N and I performed only on results at [KI] < 0.08 M. c Ksv and V for TCE quenching of I and In from non-linear regression fit to Eqn. 4 + S,D.

rate constant; and ~'o is the unperturbed fluorescence lifetime. We used the average lifetimes (~ ~ )) contained in ~rable I for ~'0. It is implicit with this approach to analyzing the quenching results that the addition of quencher does not affect the structure of the protein or aftect the equilibria between different conformational states of bGH.

For KI quenching (Fig. 4, Table II), the native and monomeric intermediate states of bGH exhibited quenching curves which were linear at low KI concentrations (less than 0.08 M) but show a plateau as the KI concentration is increased. This type of behavior for the native state may be due to charge effects on bGH structure, since the ionic strength was not controlled in these experiments or could indicate two populations of fluoro- phores [7], one whose tryptophan residue is accessible and one where it is not. For the monomeric intermediate (in 3.7 M Gdn-HCI), it is not likely that the curvature in the quenching results is due to charge effects, because the ionic strength of this solution was high. However, in order to minimize these possible charge effects, the quenching rate constants for the native and monomeric intermediate states reported in Table II were obtained from a fit to Eqn. 3 using initial slopes from the results shown in Fig. 4. The kq values for tryptophan in the native and monomeric intermediate states of bGH are indistinguishable from the kq for the unfolded state (0.72.10 9 M - 1 . s - 1 ) and therefore may represent the quenching rate constant for the small fraction of molecules which are in the unfolded state in these solutions.

The tryptophan fluorescence in the associated intermediate state of bGH was difficult to quench with iodide (kq - 0.2.10 9 M - 1 - s - l ) , indicating that the tryptophan was substantially protected from iodide penetration due to steric or environment effects. Steric effects could limit the access of iodide, because the ionic radius for iodide, although generally considered to be small enough to quench tryptophan in most proteins, is too large (2.20 A) for it to get close enough to quench the fluorescence of the tryptophan residue.

1,8

1.6 !

ik 1.4

u. /1

T

1.0 ~ -

t u a 1 u a 0.0 0.05 0.10 0. 5 0.20 0.25 0.30

KI Concentration (Ni)

Fig. 4. Stern-Vo~ner plot of bGH fluorescence quenching using iodide as quencher. Lines drawn are from linear regression fit to Eqn. 3. Unfolded-state results are mean values from four separate determinations; error bars indicated on unfolded line are ± S.D. For other states, the results are representative data points from two separate determinations; no error bars are shown for these states. Key for bGH states: o, native (N); 4, monomeric intermediate (I); D, associated intermediate (I,);

and ®, unfolded (U).

Environment effects, such as non-polar regions or negative charges, could also limit penetration of iodide ions and would seem to be a more likely cause for low fluorescence quenching than steric effects. Previous studies of bGH quenching by KI [29] in the native state provided results sin'filar to those observed here. However, more quenching was observed for bGH in the unfolded state in the previous report, yielding a Stern-Volmer quenching rate constant of 4.1 M -1 which is about 2-times the value reported here (Table II). The cause of this difference may be differences in protein purity and/or buffer (pH 7.5 tds-HCl with 5.0 M Gdn-HC1 vs. the pH 8.5 ammonium bicarbonate with 6.0 M Gdn-HCI used here).

For acrylamide quenching of tryptophan fluorescence, all the conformations of bGH exhibited linear dependence on quencher concentration (raw data not shown, quenching rate constants are contained in Table II). The results for each conformation can be explained by different steric barriers to acrylamide access ili ghe different conformations. The difference in quenching rate constants between the native and monomeric intermediate states indicates that the monomeric intermediate state has fewer steric restrictions to acrylamide access than the native state, i.e., the monomeric intermediate is more expanded than the native state as has been demonstrated with size-exclusion chromatography [4] and may be depicted as a loosening of the packing in the four a-helical bundle [6]. The associated intermediate state was least accessible to acrylaraide of all the conformational states, because close contact between molecules in the associated state limits acrylamide penetration. Quenching of the unfolded state was most easily done when compared to the other bGH conformations, because there are obviously fewer steric restrictions to acrylamide access in the unfolded conformation.

Since acrylamide is often used to probe tryptophan accessibility in proteins, it is possible to compare the quenching rate constants determined here with those determined for other single tryptophan proteins and for tryptophan in solution. The results in Table III show that the tryptophan residue in all four conformat/.onal states of bGH (even the associated intermediate s~ate) was more accessible than the single tryptophan in four other

161

TABLE III

COMPARISON OF ACRYLAMIDE QUENCHING RATE CONSTANTS FOR SINGLE TRYPTOPHAN PROTEINS

Protein kq( X 109 M - !. s - l) Ref.

Azurin 0.04 30 Cod parvalbumin 0.11 31 RNAase T 1 0.3 30

0.2 32 0.2 33

Human serum albumin 0.5 30 bGH-I,,. 0.8 this work Nuclease 0.9 30 bGH-N 1.1 this work bGH-I 1.4 this work Monellin 1.9 9

2.0 30 bGH-U 2.4 this work Glucagon 3.7 30 ACTH 4.2 30 Tryptophan 5.4 34 lndole 7.1 35

proteins, notably azurin which has a kq value about 25-times smaller than the kq values observed here. The native-state tryptophan in bGH was more exposed to acrylamide quenching than were the tryptophan residues in five of the s/x native state proteins for which acrylamide quenching rate constants have been reported. Only monellin [9] has a native-state tryptophan that was more exposed to acrylamide than the tryptophan residue in native state bGH. However, the tryptophan residue was more protected from acrylamide quenching in all the bGH conformational states (even the unfolded state) when compared to the small peptides, glucagon and ACTH, or free tryptophan in solution.

Trichloroethanol (TCE) was the only quencher of bGH fluorescence emission to show a positive deviation from linearity as shown by the results for the monomeric and associated intermediate states (Fig. 5, Table II). This may be an indication that both static and dynamic quenching occurred in these conformational states as observed previously with TCE [28]. Precipitation of native-state bGH was observed at TCE concentrations greater than 0.1 M which may be evidence for static quenching of native state bGH fluorescence as well. When static quenching occurs, it is ap-

162

10.0

8.0

6.0

L 4.0

2.0

0.0 . o o o.~s o.~o o~s o~o o~=s o~o

TCE Concentration (M)

Fig. 5. Stern-Volmer plot of bo l l fluorescence quenching using tricldorocthanol as quencher. Straight lines are from linear regression fit to Eqn. 3; curved lines are from non-linear regression fit to Stern-Voimer equation with static quenching constant (Eqn. 4): V== 5.3 M -1 and Ksv = 3.2 M -! for the monomeric intermediate (1) state and V = 6.6 M - 1 and Ksv = 2.6 M -1 for the associated intermediate (In) state. Results are from a single determination. Key for bGH states: o, native (N); z, monomeric intermediate (1); El, associated intermediate

(in); and O, unfolded (U).

propriate to analyze the results with another mod- ification of the Stern-Volmer equation [9,27]:

Fo/F= ( l + Ksv [Q]) e vlQ! (4)

where I / i s the static quenching constant repre- senting an active volume element surrounding the excited fluorophore. When the results from TCE quenching of tryptophan in the monomeric and associated intermediate were analyzed with Eqn. 4, 'good fits to the observed data were obtained (Fig. 5). The determined values for V in the monomeric intermediate (5.3 M-1) and the associated intermediate (6.6 M -1) were significantly greater than those usually determined for TCE [9] as a static quencher of protein fluorescence (2.0 M- l ) and may indicate that TCE accumulated in the vicinity of the tryptophan residue in these conformational states of bGH. This was observed previously for human serum albumin where V = 25 M -1 at pH 5.5 and V= 5.0 M -1 at pH 2.5 [28]. However, the interpretation of V values is difficult, because the determination of V values is usually less reproducible than the determination o f Ksv va lues [9].

The quenching of bGH fluorescence in the native state by TCE was similar to that obselved for acrylamide and greater than that observed for iodide (Table II). This probably reflects the preference of tryptophan for non-polar regions and is consistent with the X-ray structure of pGH [6]. For TCE quenching, the native, monomeric intermediate and associated intermediate states were indistinguishable. However, the unfolded state was clearly more accessible to TCE than the other three conformational states. The fact that the associated intermediate conformation was among the least accessible bGH conformations could be interpreted as an indication of a polar tryptophan environment in the associated intermediate state; this interpretation is in agreement with second-derivative ultraviolet absorbance results for the tryptophan residue in the associated intermediate state [4]. The emission maximum at 345 nm for the associated intermediate, however, indicates that the tryptophan environment in the a~soc;~ed intermediate is less polar than in the unfolded state while the second-derivative ultraviolet absorbance results indicate that the associated intermediate tryptophan is more polar that the unfolded state tryptophan.

The lack of significant fluorescence quenching by iodide (Table II) and low fluorescence quenching by acrylamide and TCE in the associated intermediate state of bGH are likely due to both steric restrictions (intermolecular packing of molecules in the associated state) and chemical barriers (a polar, negatively charged environment for the tryptophan in the associated state). It has been argued previously [4] from circular dichroism spectroscopy of the tryptophan residue that the tryptophan residue in the associated state is held in a rigid environment near the interface between associating molecules. The results reported here are consistent with that interpretation.

Acknowledgements

B.S. Hanna of Kalamazoo College acquired the results in Figs. 1-3 while completing a Senior Individu~,dized Project at The Upjohn Company. We are grateful for helpful discussions with D.N. Brems, J.J. Dougherty and T.J. Thamann.

References

1 Burger, H.G., Edelhoch, H. and Condliffe, P.G. (1966) J. Biol. Chem. 241, 449-457.

2 Holladay, L.A., Hammonds, R.G., Jr. and Puett, D. (1974) Biochemistry 13, 1653-1661.

3 Brems, D.N., Plaisted, S.M., Havel, H.A., Kauffman, E.W., Stodola, J.D., Eaton, L.C. and White, R.D. (1985) Bio- chemistry 24, 7662-7668.

4 Havel, H.A., Kauffman, E.W., Plaisted, S.M. and Brems, D.N. (1986) Biochemistry 25, 6533-6538.

5 Brems, D.N., Plaisted, S.M., Kauffman, E.W. and Havel, H.A. (1986) Biochemistry 25, 6539-6543.

6 Abdel-Meguid, S.S., Shieh, H.-S., Smith, W.W., Dayringer, H.E., Violand, B.N. and Bentle, L.A. (1987) Proc. Natl. Acad. Sci. USA 84, 6434-6437.

7 Lakowicz, J.R. (1983) Principles of Fluorescence Spec- troscopy, Plenum Press, New York.

8 Beechem, J.M. and Brand, L. (1985) Annu. gev. Biochem. 54, 43-71.

9 Eftink, M.R. and Ghiron, C.A. (1981) Anal Biochem. 114, 199-227.

10 Ho, C.N., Patonay, G. and Warner, I.M. (1986) Trends Anal. Chem. 5, 37-43.

11 Olson, ER., Olsen, M.K., Kaytes, P.S., Patel, H.P., Rockenbach, S.K., Watson, E.B. and Tomich, C-S.C. (1987)

.. A.S.M. Abstr., 151. 12 Evans, T.W. and Knuth, M.W. (January 15, 1987) World

Patent WO 8700204. 13 O'Connor, D.V. and Phillips, D. (1984) Time-Correlated

Single Photon Counting, Academic Press, New York. 14 Zuker, M., Szabo, A.G., Bramall, L., Krajcarski, D.T. and

Selinger, B. (1985) Rev. Sci. Instrum. 56, 14-22. 15 Brand, J.G. and Cagan, R.H. (1977) Biochim. Biophys.

Acta 493, 178-187. 16 Khan, S.M., Darnall, D.W. and Birnbaum, E.g. (1980)

Biochim. Biophys. Acta 624, 1-12.

163

17 Grinvald, A. and Steinberg, I.Z. (1976) Biochim° Biophys. Acta 427, 663-678.

18 James, D.R. and Ware, W.R. (1985) Chem. Phys. Lett. 120, 455-459.

19 Albery, W.J., Bartlett, P.N., Wilde, C.P. and Darwent, J.R. (1985) J. Am. Chem. Soc. 107, 1854-1858.

20 Alcala, J.R., Gratton, E. and Prendergast, F.G. (1987) Biophys. J. 51, 597-604.

21 Alcala, J.R., Gratton, E. and Prendergast, F.G. (1987) Biophys. J. 51, 925-936.

22 Lakowicz, J.R., Cherek, H., Gryczynski, 1., Joshi, N. and Johnson, M.L. (1987) Biophys. Chem. 28, 35-50.

23 Tanaka, F. and Mataga, N. (1987) Biophys. J. 51, 487-495. 24 Tanaka, F., Kaneda, N., Mataga, N., Tamai, N., Tamazaki,

I. and Hayashi, K. (1987) J. Phys. chem. 91, 6344-6346. 25 Robbins, D.J., Deibel, M.R., Jr. and Barkley, M.D. (1985)

Biochemistry 24, 7250-7257. 26 Lehrer, S.S. (1971) Biochemistry 10, 3254-3263. 27 Eftink, M.R. and Ohiron, C.A. (1976) J. Phys. Chem. 80,

486-493. 28 Eftink, M.R., Zajicek, .LL. and Ohiron, C.A. (1977) Bio-

chim. Biophys. Acta 491, 473-481. 29 Maddaiah, V.T. and Collipp, P.J. (1972) FEBS Lett. 23,

208-210. 30 Eftink, M.R. and Ohiron, C.A. (1976) Biochemistry 15,

672-680. 31 Eftink, M.R. and Hagaman, K.A. (1985) Biophys. Chem.

22, 173-180. 32 James, D.R., Demmer" D.R., Steer, R.P. and Verrall, R.E.

(1985) Biochemistry 24, 5517-5526. 33 Chen, L. X.-Q., Longworth, J.W. and Fleming, G.R. (1987)

Biophys. J. 51, 865-873. 34 Eftink, M.R. and Ghiron, C.A. (1981) Photochem. Photo-

biol. 33, 749-752. 35 Lakowicz, J.R., Johnson, M.L., Gryczynsld, I., Joshi, N.

and Laczko, G. (1987) J. Phys. Chem. 91, 3277-3285.

Fluorescence quenching studies of bovine growth hormone in several conformational states

Documents

Transcript of Fluorescence quenching studies of bovine growth hormone in several conformational states