Fluorescence studies of internal rotation in apohemoglobin α-chains

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 228, No. 2, February 1, pp. 519-524, 1934

Fluorescence

Studies of Internal Rotation in Apohemoglobin

a-Chains

JOSE OTON, DIONIGIO FRANCHI, ROBERT F. STEINER,’ CLARA FRONTICELLI ASUNCION MARTINEZ, AND ENRICO BUCCI’ Department of Chemistry, University of Maryland, Baltimore County, Catonsvih, Maryland 21.%8,and Department of Bio.!ogical Chemistry, School of Medicine, University of Maryland, Baltimore, Maryland Received August 8, 1933, and in revised form September

19, 1983

The molecular dynamics of the apo a-chain of human hemoglobin have been examined using three different fluorescent probes, as well as by circular dichroism. All of these criteria are consistent with a significant loss of organized structure and molecular rigidity for the apo derivative. The apo a-chain thus contrasts with the apo b-chain, which retains considerable rigidity and organized structure.

Cys-93 and Cys-112 positions showed a substantial degree of mobility. The tryptophan groups gave evidence of mobility only at low ionic strength, where there was some indication of the presence of a rotation of restricted amplitude and very short correlation time, possibly reflecting motion which is confined to tryptophan itself, or to tryptophan plus a small number of adjacent amino acids. In the case of the ANS probe, there was no indication of any rotation of short correlation time, suggesting that it is held in a fairly rigid microenvironment. In view of the conformational similarity of the a-chain to the P-chain within the hemoglobin tetramer, it is of interest to compare its properties with those of the p-chain. Any difference in molecular flexibility may have a bearing upon the respective roles of the two in mediating the allosteric properties of hemoglobin. In the present paper, the intrinsic fluorescence of the single tryptophan, the fluorescence of a label attached to the single cysteine sulfhydryl group, and the fluorescence of an ANS complex have been monitored.

The question of the possible existence of internal modes of rotational freedom in native globular proteins has recently attracted considerable attention. Evidence for the presence of such internal rotations has been obtained in particular instances from hydrogen exchange (1, Z), nuclear magnetic resonance (3), oxygen quenching of tryptophan fluorescence (4), and the time decay of fluorescence anisotropy (5, 6). While it is impossible to generalize, it now appears that rotational flexibility may be of wider occurrence in native proteins than was earlier recognized. In a preceding paper, data have been presented upon the fluorescence properties of the apo */3- and apo ‘P-chains of human hemoglobin (3). A fluorescent label (AEDANS)2 was placed specifically at the Cys-93 and Cys-112 positions. The intrinsic fluorescence of the two tryptophan groups Trp-15 and Trp-37, as well as that of a 1,8ANS group bound in the heme pocket, were also examined. The labels placed at the 1 To whom correspondence should be addressed. ’ Abbreviations used: l,&ANS, 1-anilino-S-naphthalene sulfonate; AEDANS, N-iodoacetylaminoethyl5-naphthylamine-1-sulfonate; Apo a-Chains, Hemefree derivative of (Y subunits alkylated with iodoacetamide; CN-heme, Ferric heme combined with cyanide.

EXPERIMENTAL

PROCEDURES

Hemoglobin and the isolated a- and @subunits were prepared as previously described (6-8). Alkylation of the -SH group at a-104 by either iodoacetamide or 519

0003-9361/34 $3.00 Copyright 0 1984by AcademicPress.Inc. All rights of reproductionin nny form reserved.

520

OTON

by AEDANS was performed in the following way. The heme was removed from the a subunits following the procedure of Clegg et al (12). The apoderivative was lyophilized and the powder dissolved in 8 M guanidinium-HCl, 0.2 M phosphate buffer at pH 7.8. Oxygen was removed by flushing with nitrogen and dithiothreitol was added to a 50-fold molar excess. After 3 h at 37°C the sample was dialyzed using a Visking tube with an Af, 10,000 cut off against a measured volume of the same solvent so to reduce the excess of dithiothreitol from 50 to 3 M excess. After equilibration, iodoacetamide was added to a 15 M excess over the total amount of -SH groups present in the sample. After 1 h at 37°C the protein was dialyzed against 0.04 M borate at pH 9.0, the precipitate removed by centrifugation, and the supernatant filtered through Sephadex G-25 to eliminate the residual alkylating agents. Amino acid analyzes showed nearly 100% alkylation by iodoacetamide and about 90% substitution with AEDANS. In the rest of this presentation the term apo achains indicates the heme-free derivative of oi subunits, alkylated with iodoacetamide or AEDANS. @Subunits fully alkylated with iodoacetamide were prepared as previously described (13). Static antiotropy measurements. Static anisotropy measurements were made using an Aminco-Bowman spectrofluorometer which was coupled with an Aminco Dasar data acquisition system, as described in a preceding publication (8). The viscosity was systematically varied at constant temperature by the weighed addition of quantities of sucrose or glycerol, in order to obtain a Perrin plot, as described earlier (6, 8). The exciting beam was vertically polarized. The anisotropy, A, is equal to (V - H)/( V + W), where V and H are the vertically and horizontally polarized components, respectively, of fluorescence intensity. Dynamic measurements of anisotrqm/ Measurements of the decay with time of anisotropy were made using an Ortec 9200 nanosecond fluorometer. The details of the procedure have been described elsewhere (6,8). The same least-squares-fitting program as described in Ref. (8) was used in this study to determine correlation times. The exciting beam was unpolarized, so that the anisotropy is equal to (V - H)/(2V + H). Determinations of jhwrescence decay times. Fluorescence decay times were computed using the method of moments and/or nonlinear least squares (9-11). In the case of tryptophan fluorescence, decay times were measured using the TRW instrument (8). Radiationless energy transfer. The efficiency (E) of resonance energy transfer between a fluorescent donor and an acceptor group is given by E = 1 - QDA/QD,

VI

where Qo4 is the quantum yield of the donor in the presence of acceptor and Qo is the quantum yield of

ET AL. donor alone. Rd the donor-acceptor separation responding to 50% efficiency, is given by R, = 9.79 X lOa (QoJK2n-“)1’6

cor-

PI

Here K* is an orientation factor, which is equal to 2/3 for the case of random orientation, which was assumed here; n is the refractive index of the medium (1, 4); and J is the overlap integral, which is given by J=

s

F(X) c(X) X%X

[31 s

F(X) dh

where F is the donor fluorescence intensity, t is the molar extinction coefficient of the acceptor (M-l cm-‘), and X is the wavelength in nanometers. The distance R between donor and acceptor groups in related to E and R, by

[41 In the present case the quenching of tryptophan arising from radiationless energy transfer to bound 1,8-ANS was measured. The emission spectrum of tryptophan in the absence and presence of ANS was determined with a JASCO spectrofluorometer. The absorption spectrum of ANS was determined with a Cary spectrophotometer. The integrations involved in computing Jwere done by trapezoidal intergration using a computer program. Circular dichroism Circular dichroism measurements were made with a JASCO 20 apparatus. RESULTS

conjugate. Figure 1 and AEDANS-104 Table I summarize the results obtained for an AEDANS conjugate with the Cys-104 sulfhydryl of the a-chain. The time decay of fluorescence intensity could be fitted smoothly by a single component of decay time 12.0 ns, using the method of moments (11). Any second component is at least two orders of magnitude smaller in amplitude. There is thus no indication of any heterogeneity of the microenvironment of the label, which would be reflected by fluorescence components of different emission properties. The time decay of fluorescence anisotropy is rapid, with no indication of a more slowly decaying region at longer times (Fig. 1). The essentially exponential decay could be fitted by a single short correlation time close to 0.4 ns. This value is very short in

INTERNAL

ROTATION

IN APOHEMOGLOBIN

00 -2.o 0

LnA n

TIME

(NS)

FIG. 1. Time decay of fluorescence anisotropy for an AEDANS conjugate with the Cys-104 sulfhydryl of the apo a-chain (0.3 mg/ml). The solvent is 40 mM phosphate, pH 5.0, 25°C.

comparison with the value (‘7 ns) expected for a rigid spherical molecule with the molecular weight of the a-chain and does not greatly exceed the value anticipated for a completely mobile probe (-0.1 ns). The implication is that the flexibility of the polypeptide in the Cys-104 region is such that very little immobilization of the label occurs and there is little or no restriction upon the amplitude of its rapid rotation. It is of interest that this behavior is somewhat reminiscent of that of the AEDANS conjugate with the Cys-112 TABLE

521

a-CHAINS

group of the P-chain, suggesting that the sequencehomology may endow this general portion of both chains with a high degree of flexibility (8). ANS complex with apo a-chain. 1,8-ANS is known to be bound specifically by the heme pockets of hemoglobin and myoglobin and to form a fluorescent complex. The fluorescent complex formed with the a-chain has an emission maximum near 470 nm. The time decay of fluorescence intensity is heterogeneous. A two-component analysis indicated the presence of a primary component of decay time 15 ns, plus a shorter component of decay time after 0.6 ns (Table I). However, these values should be regarded as only apparent, in view of the heterogeneity. A rapid exponential disease in fluorescence anisotropy was observed (Fig. 2). A single-component fit for the shorter range of times (less than 80 ns) yielded a correlation time of 2.0 ns (Table I), which is small in comparison with that expected for the entire molecule. The amplitude of the corresponding rotation is sufficiently large to reduce the anisotropy to a low value (-O.Ol), so that it is difficult to observe any correlation time reflecting slower rotations. It is clear that considerable mobility is being detected by a probe within the heme pocket. This is in contrast to the parallel

I

ROTATIONAL CORRELATION TIMES FROM FLUORESCENCE ANISOTROPY DECAY

AEDANS-104 l&ANS

0.3@ 12.0d 0.3se 12.0” (1 x 10-4) 0.w 9.3d 0.39” 15.3” 2.6”

(-45)

0.4

0.63”

2.0

Note. The solvent is 40 mM phosphate, pH 5.0,25”C. a Amplitude corresponding to a particular decay time of fluorescence intensity. *Decay time of fluorescence intensity (ns). ’ Rotational correlation time from one-component fit (ns). d From one-component fit. e From two-component fit.

0

%E

2.0 (NS)

FIG. 2. Time decay of fluorescence anisotropy for a l,%ANS complex with apo o-chain. The concentration of the apo a-chain is 1.5 mg/ml; that of 1,8-ANS is 0.01 mM. The buffer is 40 mM phosphate, pH 5.0,25”C.

522

OTON

case of the P-chain, for which an ANS probe reflects only the rotation of the entire molecule (8). Trgptophan jluorescence. The single tryptophan (Trp-14) of the apo a-chain has an emission maximum close to 347 nm, which is displaced substantially to longer wavelengths, as compared with the value for the apo b-chain (340 nm), suggesting a more polar environment and hence a greater degree of exposure to solvent for the former. The apo a-chain is also more effectively quenched by acrylamide than the apo P-chain (Fig. 3), in consistency with expectations if a higher extent of accessibility to solvent exists. Figures 4 and 5 show Perrin plots corresponding to tryptophan emission for the apo a-chain. For the conditions studied, all the plots show a high degree of curvature, suggesting that more than one correlation time is present. The apparent correlation times estimated from the limiting asymptotic slopes at high values of T/q are significantly smaller than anticipated for a rigid molecule of this size (Table II), or found for the apo P-chain (8) or apomyoglobin (9), suggesting a considerable mobility of the polypeptide in the region of the tryptophan. From the steep initial slopes at low values of T/v, the correlation

3

0.03

MOLARITY

ET AL.

FIG. 4. Perrin plot for tryptophan fluorescence of apo a-chain (0.4 mg/ml) in 40 mM phosphate, pH 5.3, 25°C. The viscosity (v) is varied by the addition of sucrose (0) or glycerol (0). The polarization (P) was measured as described in the text. The excitation and emission wavelengths were 235 and 345 nm, respectively. The units of T/t, are degrees centipoise-‘.

times computed are very short, suggesting that the motion of a polypeptide segment of limited size is being detected (Table II), possibly the motion of the tryptophan itself. Again, the contrast with the B-chain is considerable; these results indicate a much greater mobility for the a-chain tryptophan.

0.04

ACRYLAMIDE

FIG. 3. A comparison of the quenching by acrylamide of the apo a- (0) and apo @chains (m). The ordinate is the ratio of unquenched (lo) to quenched (1) intensities. The concentration of protein is 0.1 mg/ml. The excitation and emission wavelengths are 230 and 340 nm, respectively. The buffer is 40 mre phosphate, pH 5.3, 25’C.

FIG. 5. Perrin plot for tryptophan fluorescence of apo a-chain (0.4 mg/ml) in 20 rnrd phosphate, pH 6.7, 25°C. The other conditions are the same as for Fig. 4.

INTERNAL TABLE

ROTATION

IN APOHEMOGLOBIN

II

ROTATIONAL CORRELATION TIME FROM STATIC POLARIZATIONS OF TRYPTOPHAN FLUORESCENCE Solvent

7

ma (ns)

9’ (ns)

pH 5.3, 25°C

3.1

0.6

6.7

pH 6.7, 25°C

3.3

0.6

5.1

40 rnM

phosphate,

20 mM phosphate,

a From initial slope at low T/T values. *From asymptotic slope at high T/n values.

Radiationless energy transfer between tryptophan and 1,8-ANS. The progressive addition of l,%ANS to apo a-chain (0.3 mg/ ml in 40 InM phosphate, pH 5.0) results in a progressive reduction in the intensity of tryptophan fluorescence, without a change in shape of the emission band. An extrapolation of relative intensity at 345 nm (exciting at 390 nm) versus l/[ANS] to 11 [ANSI = 0 yields a value of 0.40 for the limiting relative intensity when the molecule is saturated with l,&ANS and a value of 0.60 for E. Application of Eq. [4] leads to a value for R/R, of 0.93 and a value of 20 + 3 A for R (assuming a value of 2/3 for K2). This is significantly increased from the crystallographic distance, 16.5 A, between Trp-14 and the heme group. Circular dichroism. The circular dichroism spectrum of the apo a-chain shows greatly reduced negative ellipticities in the peptide region (205-225 nm), as compared with the values for the intact a-chain (Fig. 6). This is suggestive of a substantial reduction in a-helical content for the isolated apo a-chain, whose conformation in intact hemoglobin is similar to that of the pchain. It should be stressed that the CD spectrum shown in Fig. 6 for the apo a-chains is similar to that reported by Yip et al. (14) for nonalkylated apo (Y subunits. The apo a-chains recombined stoichiometrically with CN-heme in a similar way to what previously observed for alkylated P-subunits (13); addition of CN-heme produced practically complete recovery of the CD spectrum of native subunits in the far-uv region of the spectrum, but not in the Soret

523

a-CHAINS

region. In this region the maximum height of the positive peak of the reconstituted material at 421 nm was about 2/3 that of the native a subunits. The apo a-chains coupled with AEDANS showed an ellipticity at 222 nm lower than that of the apo a-chains (by about 20%). Also, these chains did not recombine with CN-heme. This suggests a further change in the conformation of the system produced by the fluorophore. This was probably due to the presence of a negative charge in the substituent. DISCUSSION

The most remarkable feature of the results summarized above is the high degree of mobility of fluorescent probes located at three different positions within the apo a-chain of human hemoglobin. The contrast with the apo *b-chain, which shows indication of considerable internal rigidity (8), is considerable. The most dramatic difference arises for a l,&ANS probe located within the heme pocket. In the case of the apo AP-chain such a probe shows essentially no rotational mobility, apart from that associated with the rotation of the molecule as a whole, whereas in the case of the apo a-chain the amplitude of the rotation of short correlation time is sufficient to mask that of longer correlation time reflecting the overall motion.

WAVELENGTH

hm)

FIG. 6. Circular dichroism for apo a-chain (-) (0.1 mg/ml) and the cyanomet a-chain (- - -) (0.1 mg/ ml). The solvent is 0.5 M borate, pH 9.0, 2-mm-pathlength cuvettes at 25°C.

524

OTON

ET AL.

A similar contrast exists in the case of sitive to small energetical changes as, for the single tryptophan of the a-chain, as example, those produced by the different compared with the two tryptophans of the hydrophobic and hydrophilic characteris*P-chain. In the latter case, there is only tics of a few surface amino acid side chains. For this reason it cannot be excluded that indication of detectable mobility of limited amplitude at the lowest ionic strength ex- the missing five residues are critical. It is amined and of no mobility at higher elec- relevant to note that the stability of the trolyte levels (8). The tryptophan of the heme-free hemoglobin derivatives is cona-chain appears, in contrast, to have ex- siderably increased in the CY+ dimers of tensive rotational freedom. The circular heme-free hemoglobin (15). dichroism results indicate a major loss in a-helical content. ACKNOWLEDGMENTS It is difficult to avoid the conclusion that the a-chain of human hemoglobin posThis work was supported in part by NIH Grants sesses extensive molecular flexibility and HL13164 and AM30322. Computer time and facilities were supported in part by the Computer Network of that this flexibility is present to a significantly greater extent than in the case of the University of Maryland. the apo *P-chains. It is also of interest that the separation REFERENCES between Trp and l,&ANS group located in ENGLANDER, S. W., AND MAUEL, C. (1972) J. Biol. the heme pocket (9) is slightly increased Chem 247, 238’7. over the crystallographic value. The imENGLANDER, S. W., AND ROLFE, A. (1973) J. Biol. plication is that the conformation of the Chem 249, 4852. apo a-chain in solution is not entirely VISSCHER, R. B., AND GURD, F. R. N. (1975) J. Biol. equivalent to that of the intact a-chain Chems 250, 2238. within the hemoglobin tetramer. LAKOWICZ, J. R., AND WEBER, G. (1973) BiodwmIn view of the similar sequences and ali&y 12,417l. MUNRO, I., PECHT, I., AND STRYER, L. (1979). Proc. most identical tertiary structure of the (YNatl. Acad Sci. USA 76,56. and P-subunits, it is difficult to find an exBUCCI, E., FRONTICELLI, C., FLANIGAN, K., PERLplanation for their differences. The cu-subMAN, J., AND STEINER, R. F. (1979) Biopolymers units lack the five residues which constitute 18, 1261. the D helix in the p-subunits. This is the 7. BUCCI, E., and FRONTICELLI, C. (1965). J. Bid Chem only notable difference between the two 240, 551. kind of chains. Those residues, when pres8. OTON, J., BUCCI, E., STEINER, R. F., FRONTICELLI, ent in the P-subunits, do not seem to conC., FRANCHI, D., MONTEMARANO, J. X., AND tribute much to the structure of the proMARTINEZ, A. (1981) J. BioL Chem. 256, 7248. tein, since they are all situated externally 9. STRYER, L. (1965) J. Mol. Biol. 13, 433. on the surface of the molecule and do not 10. GRINVALD, A., and STEINBERG, I. Z. (1974) And Biochem. 59, 583. appear to establish interactions with the heme. It is possible that some residues 11. ISENBERG, I., DYSON, R. D., AND HANSON, R. (1973) Biophys J. 13,109O. critical for the stability of the apo deriv12. CLEGG, J. B., NAUGHTON, M. A., WEATHERHALL, ative have been substituted in the a-subD. J. (1966) J. Mol. Biol 19, 91. units. We have at present no experimental 13. FRONTICELLI-BUCCI, C., AND Buccx, E. (1975) Bioapproach which may identify such resichemistry 14.4451. dues. It should be stressed that the relative 14. WAKS, M., YIP, Y. K., AND BEYCHOCK, S. (1973) J. instability of the heme-free (Y and p isoBiol Chem. 246, 6462. lated subunits suggest that they possess a 15. KOWALCZYK, J., AND BUCCI, E. (1933) Biochemistry, in press. marginal stability, which make them sen-