ARCHIVES
OF BIOCHEMISTRY
Vol. 276, No. 1, January,
AND
BIOPHYSICS
pp. 236-241, 1990
Distance Measurements Chih-Lueh Department
in Cardiac Troponin Cl
Albert Wang2 and Paul C. Leavis of Muscle Research, Boston Biomedical Research Institute, 20 Staniford Street, Boston, Massachusetts
02114
Received June 19,1989
Intramolecular distance measurements were made in cardiac troponin C (cTnC) by fluorescence energy transfer using Eu3+ or Tb3+ as energy donors and Nd3+ or an organic chromophore as acceptors. The laser-induced luminescence of bound Eu3+ is quenched in EulNdlcTnC with a lifetime of 0.328 ms, compared with 0.43 ms for EuzcTnC. The enhanced decay corresponds to an energy transfer efficiency of 0.25, or a distance of 1.1 nm between the two high affinity sites. We have also labeled cTnC with 4-dimethylaminophenylazophenyl4’-maleimide (DAB-Mal) at the two cysteine residues (Cys-35 and Cys-84). Energy transfer measurements were carried out between Tb3+ bound to the high affinity sites and the labels attached to the domain containing the low affinity site. Upon uv irradiation at pH 6.7, TblcTnCDAB emits tyrosine-sensitized Tb3+ luminescence that decays biexponentially with lifetimes of 1.29 and 0.76 ms. The shorter lifetime is ascribed to energy transfer from Tb3+ to the DAB labels, yielding an average distance of 3.4 nm between the donor and the acceptors. At pH 5.0, however, the luminescence decays exclusively with a single lifetime of 1.31 ms, suggesting that under these conditions all Tb3+ ions are more than 5.2 nm away from the label. Thus cTnC, like skeletal TnC, undergoes a pa-dependent conformational transition which converts an elongated structure at lower pH’s to a rather compact conformation in a more physik? 1990 Academic Press, Inc. ological medium.
The three-dimensional structure of skeletal troponin C (sTnC)3 has been revealed by two groups of X-ray ’ This work was supported by grants from the National Institutes of Health (AR32727 and HL41411 to C.L.A.W., and HL20464 to P.C.L.) and the Muscular Dystrophy Association of America, and by Biomedical Research Grant RR05711. C.L.A.W. is an Established Investigator of the American Heart Association. ’ To whom correspondence should be addressed. ” Abbreviations used: cTnC, cardiac troponin C; DAB-Mal, 4-dimethylaminophenylazophenyl-4’.maleimide; DAB, the 4-dimethylaminophenylazophenyl moiety of DAB-Mal; cTnCuAs, cardiac troponin C with DAB-Ma1 labeled at Cys-35 and Cys-84; EDTA,
crystallographers (1, 2). These studies showed that sTnC is an unusually elongated, dumbbell-shaped molecule. The N-terminal domain that contains the two low affinity Ca*+-binding sites (sites I and II), and the C-terminal domain that contains the two high affinity sites (sites III and IV) are separated by a single a-helical stretch. The entire molecule is about 7.5 nm in length. The stability of this long helix depicted in the crystal structure has been a subject of discussion. It was suggested that the central helix is maintained by salt bridges between the side chains of charged residues (3), although evidence of such intramolecular networks is yet to be established in the more refined X-ray structure. On the other hand small angle X-ray scattering data indicate that the interdomain distance of sTnC in solution appears shorter than that predicted by the crystal structure; this would require some kind of conformational change in the central helix (4). More recently, a number of insertion and/or deletion mutants of TnC in the central helical region obtained through genetic engineering show little structural and functional differences from the wild-type protein (5), suggesting that the long helix, if it exists, must be quite flexible. It is noteworthy that similar conclusions have been reached for calmodulin, a homologous protein of TnC (6, 7). The notion of a flexible central helix in TnC is in particular consistent with several earlier reports that the two halves of TnC may interact with each other (8, 9). Furthermore, since the crystals for diffractional studies on sTnC were obtained at pH 5, one possibility is that the central helical structure is stabilized by acidic medium. We have previously observed using fluor bUL.3nce resonance energy transfer techniques that rabbit sTnC is indeed quite elongated at pH 5, but it appears to be in a more compact configuration at pH around 7 (10). In this report we have extended
ethylenediaminetetraacetic acid; Hepes, 4-(2.hydroxyethyl)-l-piperazineethanesulfonic acid, Pipes, 1,4-piperazinediethanesulfonic acid, sTnC, skeletal troponin C; Euhlgh and Eu,,,, Et?+ ions bound at the high and the low affinity sites, respectively.
236 All
0003.9861/90 $3.00 Copyright 0 1990 by Academic Press, Inc. rights of reproduction in any form reserved.
CONFORMATION
OF CARDIAC
TROPONIN
237
C
this method to cTnC; we found that the pH-dependent structural transition is also taking place in cTnC. MATERIALS
AND METHODS
cTnC was isolated from bovine heart muscle following the procedure of Potter (11). Modification of its two cysteine residues (Cys-35 and Cys-84) was carried out by incubating the protein (2-3 mg/ml) with a five-fold molar excess of DAB-Ma1 (4-dimethylaminophenylazophenyl-4’.maleimide, Molecular Probes, Junction City, OR) at 25°C for 4 h in a solution containing 8 M urea, 2 mM EDTA, 0.1 M KCl, 25 mM Hepes buffer, pH 7.5. Following labeling, the protein was dialyzed against a solution containing 0.1 M KC1 and 25 mM Hepes at pH 7.5. The concentration of cTnC was determined either spectrophot,ometritally using an extinction coefficient (290 nm, 1 cm light path, 1 mg/ ml) of 0.3 or by employing the BCA protein assay reagent (Pierce Chemical Co.). The ratio of label to protein, typically 0.8 to 1.5, was calculated using Q~~)~,,,= 2.48 X 10 Mm ’ cm I. Localization of the DAB-Ma1 binding to the cysteine residues of cTnC was checked by digesting the labeled protein with trypsin (tryp sin:protein = 1:50) for 30 min at room temperature. The reaction was terminated by the addition of a threefold excess of soybean trypsin inhibitor over the trypsin. Peptides were separated by polyacrylamide gel electrophoresis on 15% crosslinked gels in the presence of urea or sodium dodecyl sulfate. Protein bands were stained with Coomassie blue. For further analysis, gel bands were transferred to Immobilon membranes (Millipore Corp.) employing a mini Trans.Blot apparatus (Bio-Rad). Other peptide mixtures were separated by HPLC, employing a Synpak C-4 column and a O-90% CH&N gradient in 0.1% trifluoroacetic acid. The column was monitored at 230 and 460 nm to detect the labeled peptides. Isolated peptides were identified either from their amino acid compositions, determined on a Beckman system 7300 amino acid analyzer, or by sequence analysis, performed on a Model 477A protein sequencer (Applied Biosystems). Lanthanide ions (in hexahydrated chloride forms, 99.9% purity) were purchased from Alpha Ventron (Danvers, MA). Metal luminescence decay measurements were carried out as follows: The Et?’ or Tb”-containing samples were excited either indirectly with a medium intensity ultraviolet flash lamp (a Corning 7-54 cutoff filter was used to excite only the protein aromatic residues) or directly with a pulsed (10 Hz) nitrogen-pumped tunable dye laser (Molectron DL 14). Rhodamine 590 and coumarine 460 were used as the laser dyes for Eu” excitation and coumarine 500 was used for Tb” excitation. A photomultiplier equipped with a gating circuit was used as the detector, which was gated off for 0.1 ms to prevent overloading with the scattered light or prompt luminescence. The excitation spectra of proteinbound Eu“’ were obtained by manually scanning the excitation wavelength through either the 578580 nm region (for the ‘F,,-“I),, transition) or the 464-470 nm region (for the ‘F,,-“Dz transition) in 0.04.nm intervals and monitoring the intense emission at 615 nm with the use of a cutoff filter (R-62) in front of the photomultiplier. For the lifetime measurements decay data were signal averaged in 256 channels; the digital data points were fed to a PDP/ll-04 microcomputer for further data analysis. Decay parameters extracted from these data were used to calculate energy transfer efficiencies, which in turn yielded distance information. Other experimental details have been described previously (12). The theory of resonance fluorescence energy transfer has been wellestablished (13). The efficiency of energy transfer, E, is given by
where rda and rd are the luminescence lifetimes of the donor (Eu”’ or Tb:“) in the presence and absence of the acceptor (Nd:‘+ or DAB), respectively; r is the actual donor-acceptor separation, and R,, is the critical distance of 50% energy transfer. R,, is further defined as
570.5
579.0
579.5
1 580.0
Wavelength(nm) FIG. 1. The ‘Fo-“Do excitation spectra of Eu,cTnC. From the bottom, n = 0.3, 0.6, 1.0, 2.0, and 3.0. Protein concentrations were 25-30 pM in 0.1 M KC1 and 25 mM Pipes, pH 6.7. The emission intensity represents the total photon counts accumulated over 128 sweeps at each excitation wavelength collected at 0.4 ms following the laser pulse. R,, = 8.79
X
10 Lin--4$K2J(cm’i)
where n is the refractive index; 4 the donor quantum yield; K’ the orientation factor; and J the donor-acceptor spectral overlap integral. The values of these parameters are n = 1.4 (14); @ = 0.49 for Tb”’ (12) and around 0.2 for Eu”’ (15); K’ = 4; and J is calculated by numerical integration from the emission spectra of the donor and the absorption spectra of the acceptor. In this work published values of R,, for the Eu”‘-Nd”’ pair (0.9 nm, see (16)) and for the Tb:‘+-DAB pair (3.2 nm, see (17)) were used. It should be noted that R,, is not very sensitive to the value of each individual parameter; for example, change in J by a factor of 2 results only in a 12% change in Ro. The largest uncertainty usually arises from the orientation factor (K~); but since we are using lanthanide ions as the energy donor and acceptor, whose degenerate energy levels render these probes essentially isotropic, the value of f for K’ is well justified and the uncertainty is greatly minimized. The overall uncertainty of the data in this work is estimated to be less than 10%.
RESULTS AND DISCUSSION Bovine cTnC binds only three Ca2+ per protein molecule owing to the fact that its first Ca2+-binding site is defective (18). In what follows, we use luminescent lanthanide ions (Eu”+ and Tb3+) to substitute for Ca2+ at the Ca2+-binding sites. Et?+, for example, when bound to cTnC, emits enhanced metal luminescence upon irradiation with a laser light source. Figure 1 shows the exci-
238
WANG AND LEAVIS
.z g g! C
160
-
80
-
8 ii sf .-t
1
E 3 464.5
465.0
465.5
Wavelength
466.0
466.5
467.0
466.5
467.0
(nm)
60 I
464.5
-
465.0
465.5
Wavelength FIG.
2.
The ‘Fo-‘D2 excitation
466.0
(nm)
spectra of Eu,cTnC.
length compared to the spectrum of the n = 1 state (Fig. 2B). Assuming that there are only three metal-binding sites in cTnC, we therefore assign the 5-3 difference spectrum as that of the Eu3+ bound at the low affinity site (Eu,,,). In the following experiments Eu3+ ions bound at the high affinity sites of cTnC (Euhigh) are excited at 464.90 nm where the interference from Eulow is minimal. Specific excitation of Eulow has not been possible because of the extensive overlap between its spectrum and that of Euhigh; we therefore chose to excite at 466.43 nm where Eulow and Euhigh have about the same luminescence intensity. The luminescence of Euhigh (for Eu,cTnC with n = 1 to 3 when excited at 464.90 nm) decays with a single lifetime of 0.43 ms, which is comparable with that of calmodulin (15, 17). When Eu,cTnC was excited at 466.43 nm, the metal luminescence decayed nonexponentially. From the decay curve we have obtained two lifetimes, one is around 0.44 ms and the other is shorter (0.23 ms, see Table I). Since at this wavelength Eu3+ ions bound at both classes of sites are being excited (see above), we assign the component with the longer lifetime as Euhigh and the short-lived component as Eulow. The fact that Eulow has a more quenched emission suggests that site II of cTnC has a more open structure so that an extra quencher (such as water molecule) can be accommodated. This is consistent with its lower affinity for the metal ions. In the presence of Nd3’, which exhibits an absorption spectrum that overlaps with the emission spectrum of Eu”‘, and therefore acts as an energy acceptor for Eu3’ emission, the luminescence decay of Euhigh becomes non-
(A) From the
bottom, n = 1.0, 2.0, 3.0, and 5.0. Other information is the same as in Fig. 1. (B) The excitation spectra corresponding to Eu”+ bound at the two classes of sites of cTnC. The trace in closed circles (0) is the difference spectrum obtained by subtracting the spectrum of n = 3 from that of n = 5; the trace in open circles (0) is the spectrum of n = 1.
TABLE I Summary of Luminescence Decay Parameters for Eu3+ Bound to cTnC in the Presence and Absence of Nd3+ Ions” xex
71
PH
(nm)
(ms)
(A,)
(ms)
Eu,cTnC Eu,cTnC EuHcTnC Eu,NdlcTnC Eu,Nd,cTnC
6.7 6.7 6.7 6.7 6.7
464.90 464.90 464.90 464.90 464.90
0.432 0.436 0.408 0.328 0.315
(100) (100) (100) (92) (99)
0.46 0.53
Eu:,cTnC EqcTnC Eu,Nd,cTnC
6.7 6.7 6.7
466.43 466.43 466.43
0.442 0.441 0.354
(45) (44) (53)
0.237 0.232 0.161
(5’3
Eu,NdlcTnC
5.0*
466.43
0.51
0.26
(62)
Eu,La,cTnC Eu,LazcTnC
6.7 5.0
466.43 466.43
0.55 0.58
(38) (26)
0.27 0.26
(61)
Sample
tation spectra of protein-bound Eu3+ at around 579 nm (for the 7Fo-5Do transition) as cTnC was titrated with EuC13. Although the X,,, shifts slightly toward the longer wavelength (from 578.94 to 579.02 nm) as the low affinity site is being saturated, the resolution is poor between ions at the high affinity sites and those at the low affinity site. In an attempt to improve the separation so that Eu3+ at either class of sites can be specifically excited, the 7Fo-5D2 transition was also monitored using coumarine 460 as the laser dye (15). In contrast to the 7Fo-5Do transition, the 7Fo-5D2 transition of Eu3+ ions results in a rather broad excitation spectrum due to degeneracy at the excited state. The overall shape of the spectra again shows little change during the titration (Fig. 2A), but the difference spectrum of the n = 5 and n = 3 states clearly shows a shift toward the longer wave-
T2
(39)
(AZ)
(8) (1) (55) (47) (74)
” Protein concentrations typically were 15-20 pM in 0.1 KC1 and 25 Pipes, pH 6.7. Samples were excited by laser using coumarine 460 as the laser dye. Data were subjected to one- or two-exponential method-of-moments analyses; see the legend to Fig. 3. * The pH was adjusted by adding 0.1 M HCl. mM
CONFORMATION
OF CARDIAC
1 0
Time
Irnil
FIG. 3. Luminescence decay curves for Eu,cTnC (upper curve) and Eu,Nd,cTnC (lower curve) upon direct excitation. X,, = 464.9 nm. Protein concentrations were 15-20 fiM in 0.1 M KCl, 25 mM Pipes, pH 6.7. Dots represent normalized counts in each channel after base subtraction (1 channel = 0.02 ms). Solid lines are the calculated fitting curves based on the parameters obtained from a two-exponential method-ofmoments analysis. The quality of fit was judged from the magnitude of the mean square residual defined as x”/N = (l/N) X:;“-, (I,,, - I,,)“/ I,,, , where I,,, and I,,, are the calculated and experimental intensities, respectively; N is the total number of points. The values of x’/N were about 1 in all cases.
exponential, the appearance of a short lifetime being indicative of energy transfer (Fig. 3). In a mixture of Eu”+: Nd”+:cTnC = 1:l:l the two lanthanide ions mainly occupy the two high affinity sites and the energy transfer takes place between them. From the measured lifetimes we have calculated the distance between sites III and IV in the C-terminal domain to be 1.1 nm, which agrees well with the crystallographic structure of sTnC. When cTnC is mixed with one Eu3+ and two Nd3+ ions per molecule, the distribution of these lanthanide ions would be such that most of the ions occupy sites III and IV (the high affinity sites) in the C-terminal domain, while the rest bind at site II (the low affinity site) in the N-terminal domain. One would thus expect both intraand interdomain energy transfer to occur. Upon excitation at 464.90 nm Euhigh in Eu,Nd+TnC exhibits emission that decays with the same quenched lifetime as described above. When excited at 466.43 nm, however, the luminescence decay yielded two lifetimes, both being shorter than the unquenched ones (Fig. 4 and Table I).
TROPONIN
239
C
While one of these two lifetimes (0.35 ms) is attributable to Euhigh that undergoes intradomain energy transfer between sites III and IV, the other one (0.16 ms) must result from cTnC molecules containing one Et?+ at either the high or the low affinity site and two Nd3+ acceptors at the remaining sites. This would imply that under the experimental conditions (pH 6.7) the two halves of the cTnC molecule can come close enough for interdomain energy transfer to occur. It should be pointed out that it is extremely difficult to resolve more than two lifetimes from a given decay curve. The observed short lifetime (0.16 ms) may represent a combination of more than one luminescent species, and thus from these lifetime measurements an estimation of the separation distance between the two end-domains has not been possible (see below). To rule out the possibility that the shortened lifetimes might result from protein conformational changes induced by the added second lanthanide ions, we have also examined the luminescence decay of EulLa,cTnC (nonfluorescent La”+ does not absorb energy in the region that Eu”+ emits, and therefore is not an acceptor), in which case the lifetimes were not quenched (Table I). All the foregoing experiments were carried out in a solution buffered at pH 6.7. In order to see the effect of pH on the structure of cTnC, the pH of the solution was lowered by adding 0.1 M HCl. At pH 5.0 the lumines-
0
1
Time
lms]
FIG. 4. Luminescence decay curves for Eu,,cTnC at pH 6.7 (upper curve) and Eu,Nd+TnC at pH 6.7 (lower curve) and Eu,Nd,cTnC at pH 5.0 (middle curve) upon direct excitation. X,, = 466.43 nm. See the legend to Fig. 3 for other information.
240
WANG
AND
B
0
15
30
60
90
Minutes FIG. 5. Localization of labels on cTnC by gel electrophoresis. (A) Urea gel electrophoresis of the time course of tryptic digestion of cTnCnAs (see text for details of digestion) stained with Coomassie blue. (B) Unstained gel of a 30-min digestion showing the single labeled fragment comprising residues 22-39.
cence of Eu,NdzcTnC decayed more slowly, corresponding to less energy transfer or an increase in the distance between the donor ions and the acceptor ions (Fig. 4, Table I). The lifetimes are comparable to the unquenched lifetimes of Eu3+, indicating that some bound Eu3+ ions do not have any acceptor in their vicinity. Thus these data suggest a structural transition of the cTnC molecule between pH 5 and pH 6.7. In contrast to the case of Eu,NdzcTnC, the luminescence of EulLazcTnC does not depend on the pH. We then carried out further interdomain distance mapping in cTnC, taking advantage of the fact that there are two cysteine residues in the N-terminal domain (Cys-35 and Cys-84) and the two high affinity sites located in the C-terminal domain. By attaching a chromophor (such as DAB-mal) to the cysteine residues and placing luminescent ions (Tb3’) at sites III and IV, we hope to observe energy transfer between the two enddomains. To verify that labeling was confined to these sites, tryptic digests of cTnC were run on gel electrophoresis (Fig. 5). The DAB-Ma1 probe whose intense color makes it easily visible on gels was confined to a single peptide band which had an amino acid composition and sequence that identified it as residues 22 to 39. The lack of a second labeled band that would contain Cys-84 is not surprising insofar as trypsin is able to cleave the protein at Arg-83 and Lys-86, leaving the tripeptide CysMet-Lys, which would not be likely to form a band on the gel. Separation of the digest by HPLC (results not shown), however, yielded a major peak corresponding to
LEAVIS
residues 22 to 39 plus a smaller peak containing a mixture of peptides, one of which appeared from amino acid analysis to be Cys-84 to Lys-86. No evidence was found for label on any other peptide derived from the protein. Thus we conclude that the DAB-Ma1 label is confined to the N-terminal half of cTnC. At pH 6.7 the luminescence of TblcTnCDAB upon uv excitation decays primarily (78% of the total emission intensity) with a short lifetime of 0.76 ms while 22% of the total intensity decays with an unquenched lifetime of 1.29 ms (Fig. 6, lower curve). The latter component is most likely due to incomplete labeling of the protein. The short-lived component is ascribed to an energy transfer between Tb3+ and the DAB label. The lifetime yields a distance of 3.4 nm. This distance, which is shorter than one would expect from the crystallographic structure, suggests that under these conditions the Cterminal domain of cTnC is relatively close to the Nterminal domain. When the pH is changed to 5.0, however, the metal luminescence of the same sample decays essentially with a single, unquenched lifetime of 1.31 ms (Fig. 6, upper curve), suggesting that the DAB moiety is now more than 5.2 nm (or 2 X R,J away from the bound Tb3+ (at sites III and IV). Similar results were obtained when the sample was excited directly with a laser light source (Table II).
I+
4
s *.2 3 Q)
1
2
Time
3
lmsl
FIG. 6. Luminescence decay curves for Tb,cTnt&s at pH 6.7 (lower curve) and Tb,cTnC& at pH 5.0 (upper curve) upon indirect excitation. The sample was excited with a uv flash lamp (1 channel = 0.04 ms). See Materials and Methods and the legend to Fig. 3 for other information.
CONFORMATION TABLE
OF CARDIAC
II
Summary of Luminescence Decay Parameters for Tb3’ Bound to Unlabeled and DAB-Labeled cTnC”
72
71 Sample Tb,cTnC
PH
Excitation
(ms)
C-4,)
6.7
“V “V laser’ laser “V U” laser laser
1.32 1.42
(100) (100) (100) (100)
5.0h Tb,cTnC Tb,cTnC,>m Tb,cTnCh
6.7 5.0 6.7 5.0 6.7 5.0
1.28
1.36 0.76 1.31 0.66 0.68
(ms)
(AZ)
(78)
1.29
(98) (100)
3.77
(22) (2)
(34)
1.32
(66)
’ Protein concentrations were the same as those in Table I. ’ pH was adjusted by adding 0.1 M HCl or 0.1 M NaOH. ’ For direct excitations a solution of coumarine 500 was used as the laser dye; X,, = 489.5 nm.
Since the luminescence of Tb3+ bound to unlabeled cTnC exhibits at both pH 6.7 and 5.0 a single lifetime of 1.3 ms (Table II), the observed change in the decay patterns indicates a pH-dependent structural transition in the cTnC molecule: at pH 5.0 cTnC exits in an elongated conformation as described by the X-ray diffraction studies; at neutral pH’s, however, cTnC assumes a more compact conformation so that the two end-domains are only 3.4 nm apart. This conclusion is consistent with the observations made with the Eu-Nd energy transfer (see above). Such a pH-dependent conformational change is very similar to what has been observed for sTnC in solution (10). The data presented in this paper, along with similar data for skeletal TnC (10) and calmodulin (19), call into question the relationship between the structures of these proteins determined from crystals grown at acidic pH and their solution structures at neutral pH values. In all cases the proteins assume a more compact structure than would be expected from the crystal data that brings their high and low affinity domains closer together. This suggests that the rigidity of the central helix implicit in the crystal structures is at least partially lost. The recent studies of Persechini and Kretsinger (7) in which the two domains of calmodulin are crosslinked by a bifunctional reagent spanning, at most, 1.9 nm provides direct corroborating evidence for a nonhelical bend between the two domains. If this is indeed the case, observations of interdomain communication in isolated skeletal TnC (8, 9) in which metal binding to one domain affects the metal binding kinetics and extrinsic fluorescence of a probe in the other are problematic insofar as it is difficult to imagine how a flexible central tether can act as a
TROPONIN
241
C
transducer of conformational change within the molecule. On the other hand, the possibility that hydrophobic pockets in the two domains may be brought into proximity by the helical bend such that they might jointly form a binding site for a target protein, as has been recently postulated for the calmodulin-myosin light chain kinase complex (20), opens new avenues for investigations of calmodulin and TnC complexes. ACKNOWLEDGMENTS The authors thank Miss Qian Zhan and Miss Elizabeth Gowell for their technical assistance, and also Dr. Renne Lu and Miss Anna Wong for the peptide sequence analysis.
REFERENCES 1. Herzberg, O., and James, M. N. G. (1985) Nature (London) 653-659. 2. Sundaralingam, M., Bergstrom, R., Strasburg, G., Rao, S. T., chowdhury, P., Greaser, M., and Wang, B. C. (1985) Science 945-948. M., Drendel, W., and Greaser, M. (1985) 3. Sundaralingam, Natl.
Acad.
4. Heidorn,
Sci. USA
82,
313, Roy227, Proc.
7944-7947.
D. B., and Trewhella,
J. (1988) Biochemistry
27,
909-
915.
5. Xu, G.-Q., and Hitchcock-DeGregori, 263, 13,962-13,969. 6. Putkey, J. A., Ono, T., VanBerkum, (1988)
J. Biol.
Chem.
263,
S. E. (1988) J. Biol. Chem. M. F. A., and Means, A. R.
11,242-11,249.
7. Persechini, A., and Kretsinger, R. H. (1988) J. Biol. Chem. 12,175-12,178. 8. Wang, C.-L. A., Leavis, P. C., and Gergely, J. (1983) J. Biol. 258,9175-9177. 9. Grabarek,
Z., Leavis, P. C., and Gergely, J. (1986) J. Biol.
263, Chem. Chem.
261,608-613. 10. Wang, C.-L. A., Zhan, Q., Tao, T., and Gergely, J. (1987) J. Biol. Chem.
262,9636-9640.
11. Potter, J. D. (1982) in Methods in Enzymology (Frederiksen, D. W., and Cunningham, L. W., Eds.), Vol. 85, pp. 241-263, Academic Press, San Diego.
12. Wang, C.-L. A., Tao, T., and Gergely, J. (1982) J. Biol. Chem. 257,8372-8375. 13. Stryer, L. (1978) Annu. Reu. Biochem. 47,819-846. 14. Fairclough, R. H., and Cantor, C. R. (1978) in Methods in Enzymology (Hirs, C. H. W., and Timasheff, S. N., Eds.), Vol. 48, pp. 347-379, Academic Press, San Diego. 15. Mulqueen, P., Tingey, J. M., and Horrocks, W. Dew., Jr. (1985) Biochemistry
24,6639-6645.
16. Horrocks,
W. Dew., Jr., Rhee, M.-J., Snyder, A. P., and Sudnick, D. R. (1980) J. Amer. Chem. Sot. 102,3650-3652. 17. Wang, C.-L. A. (1986)J. Biol. Chem. 261, 11,106-11,109. 18. Leavis, P. C., and Kraft, E. L. (1978) Arch. Biochem. Biophys.
186,411-415. 19. Wang, C.-L. A. (1989) Biochemistry 28,4816-4820. 20. Persechini, A., and Kretsinger, R. H. (1988) J. Cardiouasc. macol. 12, Suppl. 5, Sl-S12.
Phar-
OF BIOCHEMISTRY
Vol. 276, No. 1, January,
AND
BIOPHYSICS
pp. 236-241, 1990
Distance Measurements Chih-Lueh Department
in Cardiac Troponin Cl
Albert Wang2 and Paul C. Leavis of Muscle Research, Boston Biomedical Research Institute, 20 Staniford Street, Boston, Massachusetts
02114
Received June 19,1989
Intramolecular distance measurements were made in cardiac troponin C (cTnC) by fluorescence energy transfer using Eu3+ or Tb3+ as energy donors and Nd3+ or an organic chromophore as acceptors. The laser-induced luminescence of bound Eu3+ is quenched in EulNdlcTnC with a lifetime of 0.328 ms, compared with 0.43 ms for EuzcTnC. The enhanced decay corresponds to an energy transfer efficiency of 0.25, or a distance of 1.1 nm between the two high affinity sites. We have also labeled cTnC with 4-dimethylaminophenylazophenyl4’-maleimide (DAB-Mal) at the two cysteine residues (Cys-35 and Cys-84). Energy transfer measurements were carried out between Tb3+ bound to the high affinity sites and the labels attached to the domain containing the low affinity site. Upon uv irradiation at pH 6.7, TblcTnCDAB emits tyrosine-sensitized Tb3+ luminescence that decays biexponentially with lifetimes of 1.29 and 0.76 ms. The shorter lifetime is ascribed to energy transfer from Tb3+ to the DAB labels, yielding an average distance of 3.4 nm between the donor and the acceptors. At pH 5.0, however, the luminescence decays exclusively with a single lifetime of 1.31 ms, suggesting that under these conditions all Tb3+ ions are more than 5.2 nm away from the label. Thus cTnC, like skeletal TnC, undergoes a pa-dependent conformational transition which converts an elongated structure at lower pH’s to a rather compact conformation in a more physik? 1990 Academic Press, Inc. ological medium.
The three-dimensional structure of skeletal troponin C (sTnC)3 has been revealed by two groups of X-ray ’ This work was supported by grants from the National Institutes of Health (AR32727 and HL41411 to C.L.A.W., and HL20464 to P.C.L.) and the Muscular Dystrophy Association of America, and by Biomedical Research Grant RR05711. C.L.A.W. is an Established Investigator of the American Heart Association. ’ To whom correspondence should be addressed. ” Abbreviations used: cTnC, cardiac troponin C; DAB-Mal, 4-dimethylaminophenylazophenyl-4’.maleimide; DAB, the 4-dimethylaminophenylazophenyl moiety of DAB-Mal; cTnCuAs, cardiac troponin C with DAB-Ma1 labeled at Cys-35 and Cys-84; EDTA,
crystallographers (1, 2). These studies showed that sTnC is an unusually elongated, dumbbell-shaped molecule. The N-terminal domain that contains the two low affinity Ca*+-binding sites (sites I and II), and the C-terminal domain that contains the two high affinity sites (sites III and IV) are separated by a single a-helical stretch. The entire molecule is about 7.5 nm in length. The stability of this long helix depicted in the crystal structure has been a subject of discussion. It was suggested that the central helix is maintained by salt bridges between the side chains of charged residues (3), although evidence of such intramolecular networks is yet to be established in the more refined X-ray structure. On the other hand small angle X-ray scattering data indicate that the interdomain distance of sTnC in solution appears shorter than that predicted by the crystal structure; this would require some kind of conformational change in the central helix (4). More recently, a number of insertion and/or deletion mutants of TnC in the central helical region obtained through genetic engineering show little structural and functional differences from the wild-type protein (5), suggesting that the long helix, if it exists, must be quite flexible. It is noteworthy that similar conclusions have been reached for calmodulin, a homologous protein of TnC (6, 7). The notion of a flexible central helix in TnC is in particular consistent with several earlier reports that the two halves of TnC may interact with each other (8, 9). Furthermore, since the crystals for diffractional studies on sTnC were obtained at pH 5, one possibility is that the central helical structure is stabilized by acidic medium. We have previously observed using fluor bUL.3nce resonance energy transfer techniques that rabbit sTnC is indeed quite elongated at pH 5, but it appears to be in a more compact configuration at pH around 7 (10). In this report we have extended
ethylenediaminetetraacetic acid; Hepes, 4-(2.hydroxyethyl)-l-piperazineethanesulfonic acid, Pipes, 1,4-piperazinediethanesulfonic acid, sTnC, skeletal troponin C; Euhlgh and Eu,,,, Et?+ ions bound at the high and the low affinity sites, respectively.
236 All
0003.9861/90 $3.00 Copyright 0 1990 by Academic Press, Inc. rights of reproduction in any form reserved.
CONFORMATION
OF CARDIAC
TROPONIN
237
C
this method to cTnC; we found that the pH-dependent structural transition is also taking place in cTnC. MATERIALS
AND METHODS
cTnC was isolated from bovine heart muscle following the procedure of Potter (11). Modification of its two cysteine residues (Cys-35 and Cys-84) was carried out by incubating the protein (2-3 mg/ml) with a five-fold molar excess of DAB-Ma1 (4-dimethylaminophenylazophenyl-4’.maleimide, Molecular Probes, Junction City, OR) at 25°C for 4 h in a solution containing 8 M urea, 2 mM EDTA, 0.1 M KCl, 25 mM Hepes buffer, pH 7.5. Following labeling, the protein was dialyzed against a solution containing 0.1 M KC1 and 25 mM Hepes at pH 7.5. The concentration of cTnC was determined either spectrophot,ometritally using an extinction coefficient (290 nm, 1 cm light path, 1 mg/ ml) of 0.3 or by employing the BCA protein assay reagent (Pierce Chemical Co.). The ratio of label to protein, typically 0.8 to 1.5, was calculated using Q~~)~,,,= 2.48 X 10 Mm ’ cm I. Localization of the DAB-Ma1 binding to the cysteine residues of cTnC was checked by digesting the labeled protein with trypsin (tryp sin:protein = 1:50) for 30 min at room temperature. The reaction was terminated by the addition of a threefold excess of soybean trypsin inhibitor over the trypsin. Peptides were separated by polyacrylamide gel electrophoresis on 15% crosslinked gels in the presence of urea or sodium dodecyl sulfate. Protein bands were stained with Coomassie blue. For further analysis, gel bands were transferred to Immobilon membranes (Millipore Corp.) employing a mini Trans.Blot apparatus (Bio-Rad). Other peptide mixtures were separated by HPLC, employing a Synpak C-4 column and a O-90% CH&N gradient in 0.1% trifluoroacetic acid. The column was monitored at 230 and 460 nm to detect the labeled peptides. Isolated peptides were identified either from their amino acid compositions, determined on a Beckman system 7300 amino acid analyzer, or by sequence analysis, performed on a Model 477A protein sequencer (Applied Biosystems). Lanthanide ions (in hexahydrated chloride forms, 99.9% purity) were purchased from Alpha Ventron (Danvers, MA). Metal luminescence decay measurements were carried out as follows: The Et?’ or Tb”-containing samples were excited either indirectly with a medium intensity ultraviolet flash lamp (a Corning 7-54 cutoff filter was used to excite only the protein aromatic residues) or directly with a pulsed (10 Hz) nitrogen-pumped tunable dye laser (Molectron DL 14). Rhodamine 590 and coumarine 460 were used as the laser dyes for Eu” excitation and coumarine 500 was used for Tb” excitation. A photomultiplier equipped with a gating circuit was used as the detector, which was gated off for 0.1 ms to prevent overloading with the scattered light or prompt luminescence. The excitation spectra of proteinbound Eu“’ were obtained by manually scanning the excitation wavelength through either the 578580 nm region (for the ‘F,,-“I),, transition) or the 464-470 nm region (for the ‘F,,-“Dz transition) in 0.04.nm intervals and monitoring the intense emission at 615 nm with the use of a cutoff filter (R-62) in front of the photomultiplier. For the lifetime measurements decay data were signal averaged in 256 channels; the digital data points were fed to a PDP/ll-04 microcomputer for further data analysis. Decay parameters extracted from these data were used to calculate energy transfer efficiencies, which in turn yielded distance information. Other experimental details have been described previously (12). The theory of resonance fluorescence energy transfer has been wellestablished (13). The efficiency of energy transfer, E, is given by
where rda and rd are the luminescence lifetimes of the donor (Eu”’ or Tb:“) in the presence and absence of the acceptor (Nd:‘+ or DAB), respectively; r is the actual donor-acceptor separation, and R,, is the critical distance of 50% energy transfer. R,, is further defined as
570.5
579.0
579.5
1 580.0
Wavelength(nm) FIG. 1. The ‘Fo-“Do excitation spectra of Eu,cTnC. From the bottom, n = 0.3, 0.6, 1.0, 2.0, and 3.0. Protein concentrations were 25-30 pM in 0.1 M KC1 and 25 mM Pipes, pH 6.7. The emission intensity represents the total photon counts accumulated over 128 sweeps at each excitation wavelength collected at 0.4 ms following the laser pulse. R,, = 8.79
X
10 Lin--4$K2J(cm’i)
where n is the refractive index; 4 the donor quantum yield; K’ the orientation factor; and J the donor-acceptor spectral overlap integral. The values of these parameters are n = 1.4 (14); @ = 0.49 for Tb”’ (12) and around 0.2 for Eu”’ (15); K’ = 4; and J is calculated by numerical integration from the emission spectra of the donor and the absorption spectra of the acceptor. In this work published values of R,, for the Eu”‘-Nd”’ pair (0.9 nm, see (16)) and for the Tb:‘+-DAB pair (3.2 nm, see (17)) were used. It should be noted that R,, is not very sensitive to the value of each individual parameter; for example, change in J by a factor of 2 results only in a 12% change in Ro. The largest uncertainty usually arises from the orientation factor (K~); but since we are using lanthanide ions as the energy donor and acceptor, whose degenerate energy levels render these probes essentially isotropic, the value of f for K’ is well justified and the uncertainty is greatly minimized. The overall uncertainty of the data in this work is estimated to be less than 10%.
RESULTS AND DISCUSSION Bovine cTnC binds only three Ca2+ per protein molecule owing to the fact that its first Ca2+-binding site is defective (18). In what follows, we use luminescent lanthanide ions (Eu”+ and Tb3+) to substitute for Ca2+ at the Ca2+-binding sites. Et?+, for example, when bound to cTnC, emits enhanced metal luminescence upon irradiation with a laser light source. Figure 1 shows the exci-
238
WANG AND LEAVIS
.z g g! C
160
-
80
-
8 ii sf .-t
1
E 3 464.5
465.0
465.5
Wavelength
466.0
466.5
467.0
466.5
467.0
(nm)
60 I
464.5
-
465.0
465.5
Wavelength FIG.
2.
The ‘Fo-‘D2 excitation
466.0
(nm)
spectra of Eu,cTnC.
length compared to the spectrum of the n = 1 state (Fig. 2B). Assuming that there are only three metal-binding sites in cTnC, we therefore assign the 5-3 difference spectrum as that of the Eu3+ bound at the low affinity site (Eu,,,). In the following experiments Eu3+ ions bound at the high affinity sites of cTnC (Euhigh) are excited at 464.90 nm where the interference from Eulow is minimal. Specific excitation of Eulow has not been possible because of the extensive overlap between its spectrum and that of Euhigh; we therefore chose to excite at 466.43 nm where Eulow and Euhigh have about the same luminescence intensity. The luminescence of Euhigh (for Eu,cTnC with n = 1 to 3 when excited at 464.90 nm) decays with a single lifetime of 0.43 ms, which is comparable with that of calmodulin (15, 17). When Eu,cTnC was excited at 466.43 nm, the metal luminescence decayed nonexponentially. From the decay curve we have obtained two lifetimes, one is around 0.44 ms and the other is shorter (0.23 ms, see Table I). Since at this wavelength Eu3+ ions bound at both classes of sites are being excited (see above), we assign the component with the longer lifetime as Euhigh and the short-lived component as Eulow. The fact that Eulow has a more quenched emission suggests that site II of cTnC has a more open structure so that an extra quencher (such as water molecule) can be accommodated. This is consistent with its lower affinity for the metal ions. In the presence of Nd3’, which exhibits an absorption spectrum that overlaps with the emission spectrum of Eu”‘, and therefore acts as an energy acceptor for Eu3’ emission, the luminescence decay of Euhigh becomes non-
(A) From the
bottom, n = 1.0, 2.0, 3.0, and 5.0. Other information is the same as in Fig. 1. (B) The excitation spectra corresponding to Eu”+ bound at the two classes of sites of cTnC. The trace in closed circles (0) is the difference spectrum obtained by subtracting the spectrum of n = 3 from that of n = 5; the trace in open circles (0) is the spectrum of n = 1.
TABLE I Summary of Luminescence Decay Parameters for Eu3+ Bound to cTnC in the Presence and Absence of Nd3+ Ions” xex
71
PH
(nm)
(ms)
(A,)
(ms)
Eu,cTnC Eu,cTnC EuHcTnC Eu,NdlcTnC Eu,Nd,cTnC
6.7 6.7 6.7 6.7 6.7
464.90 464.90 464.90 464.90 464.90
0.432 0.436 0.408 0.328 0.315
(100) (100) (100) (92) (99)
0.46 0.53
Eu:,cTnC EqcTnC Eu,Nd,cTnC
6.7 6.7 6.7
466.43 466.43 466.43
0.442 0.441 0.354
(45) (44) (53)
0.237 0.232 0.161
(5’3
Eu,NdlcTnC
5.0*
466.43
0.51
0.26
(62)
Eu,La,cTnC Eu,LazcTnC
6.7 5.0
466.43 466.43
0.55 0.58
(38) (26)
0.27 0.26
(61)
Sample
tation spectra of protein-bound Eu3+ at around 579 nm (for the 7Fo-5Do transition) as cTnC was titrated with EuC13. Although the X,,, shifts slightly toward the longer wavelength (from 578.94 to 579.02 nm) as the low affinity site is being saturated, the resolution is poor between ions at the high affinity sites and those at the low affinity site. In an attempt to improve the separation so that Eu3+ at either class of sites can be specifically excited, the 7Fo-5D2 transition was also monitored using coumarine 460 as the laser dye (15). In contrast to the 7Fo-5Do transition, the 7Fo-5D2 transition of Eu3+ ions results in a rather broad excitation spectrum due to degeneracy at the excited state. The overall shape of the spectra again shows little change during the titration (Fig. 2A), but the difference spectrum of the n = 5 and n = 3 states clearly shows a shift toward the longer wave-
T2
(39)
(AZ)
(8) (1) (55) (47) (74)
” Protein concentrations typically were 15-20 pM in 0.1 KC1 and 25 Pipes, pH 6.7. Samples were excited by laser using coumarine 460 as the laser dye. Data were subjected to one- or two-exponential method-of-moments analyses; see the legend to Fig. 3. * The pH was adjusted by adding 0.1 M HCl. mM
CONFORMATION
OF CARDIAC
1 0
Time
Irnil
FIG. 3. Luminescence decay curves for Eu,cTnC (upper curve) and Eu,Nd,cTnC (lower curve) upon direct excitation. X,, = 464.9 nm. Protein concentrations were 15-20 fiM in 0.1 M KCl, 25 mM Pipes, pH 6.7. Dots represent normalized counts in each channel after base subtraction (1 channel = 0.02 ms). Solid lines are the calculated fitting curves based on the parameters obtained from a two-exponential method-ofmoments analysis. The quality of fit was judged from the magnitude of the mean square residual defined as x”/N = (l/N) X:;“-, (I,,, - I,,)“/ I,,, , where I,,, and I,,, are the calculated and experimental intensities, respectively; N is the total number of points. The values of x’/N were about 1 in all cases.
exponential, the appearance of a short lifetime being indicative of energy transfer (Fig. 3). In a mixture of Eu”+: Nd”+:cTnC = 1:l:l the two lanthanide ions mainly occupy the two high affinity sites and the energy transfer takes place between them. From the measured lifetimes we have calculated the distance between sites III and IV in the C-terminal domain to be 1.1 nm, which agrees well with the crystallographic structure of sTnC. When cTnC is mixed with one Eu3+ and two Nd3+ ions per molecule, the distribution of these lanthanide ions would be such that most of the ions occupy sites III and IV (the high affinity sites) in the C-terminal domain, while the rest bind at site II (the low affinity site) in the N-terminal domain. One would thus expect both intraand interdomain energy transfer to occur. Upon excitation at 464.90 nm Euhigh in Eu,Nd+TnC exhibits emission that decays with the same quenched lifetime as described above. When excited at 466.43 nm, however, the luminescence decay yielded two lifetimes, both being shorter than the unquenched ones (Fig. 4 and Table I).
TROPONIN
239
C
While one of these two lifetimes (0.35 ms) is attributable to Euhigh that undergoes intradomain energy transfer between sites III and IV, the other one (0.16 ms) must result from cTnC molecules containing one Et?+ at either the high or the low affinity site and two Nd3+ acceptors at the remaining sites. This would imply that under the experimental conditions (pH 6.7) the two halves of the cTnC molecule can come close enough for interdomain energy transfer to occur. It should be pointed out that it is extremely difficult to resolve more than two lifetimes from a given decay curve. The observed short lifetime (0.16 ms) may represent a combination of more than one luminescent species, and thus from these lifetime measurements an estimation of the separation distance between the two end-domains has not been possible (see below). To rule out the possibility that the shortened lifetimes might result from protein conformational changes induced by the added second lanthanide ions, we have also examined the luminescence decay of EulLa,cTnC (nonfluorescent La”+ does not absorb energy in the region that Eu”+ emits, and therefore is not an acceptor), in which case the lifetimes were not quenched (Table I). All the foregoing experiments were carried out in a solution buffered at pH 6.7. In order to see the effect of pH on the structure of cTnC, the pH of the solution was lowered by adding 0.1 M HCl. At pH 5.0 the lumines-
0
1
Time
lms]
FIG. 4. Luminescence decay curves for Eu,,cTnC at pH 6.7 (upper curve) and Eu,Nd+TnC at pH 6.7 (lower curve) and Eu,Nd,cTnC at pH 5.0 (middle curve) upon direct excitation. X,, = 466.43 nm. See the legend to Fig. 3 for other information.
240
WANG
AND
B
0
15
30
60
90
Minutes FIG. 5. Localization of labels on cTnC by gel electrophoresis. (A) Urea gel electrophoresis of the time course of tryptic digestion of cTnCnAs (see text for details of digestion) stained with Coomassie blue. (B) Unstained gel of a 30-min digestion showing the single labeled fragment comprising residues 22-39.
cence of Eu,NdzcTnC decayed more slowly, corresponding to less energy transfer or an increase in the distance between the donor ions and the acceptor ions (Fig. 4, Table I). The lifetimes are comparable to the unquenched lifetimes of Eu3+, indicating that some bound Eu3+ ions do not have any acceptor in their vicinity. Thus these data suggest a structural transition of the cTnC molecule between pH 5 and pH 6.7. In contrast to the case of Eu,NdzcTnC, the luminescence of EulLazcTnC does not depend on the pH. We then carried out further interdomain distance mapping in cTnC, taking advantage of the fact that there are two cysteine residues in the N-terminal domain (Cys-35 and Cys-84) and the two high affinity sites located in the C-terminal domain. By attaching a chromophor (such as DAB-mal) to the cysteine residues and placing luminescent ions (Tb3’) at sites III and IV, we hope to observe energy transfer between the two enddomains. To verify that labeling was confined to these sites, tryptic digests of cTnC were run on gel electrophoresis (Fig. 5). The DAB-Ma1 probe whose intense color makes it easily visible on gels was confined to a single peptide band which had an amino acid composition and sequence that identified it as residues 22 to 39. The lack of a second labeled band that would contain Cys-84 is not surprising insofar as trypsin is able to cleave the protein at Arg-83 and Lys-86, leaving the tripeptide CysMet-Lys, which would not be likely to form a band on the gel. Separation of the digest by HPLC (results not shown), however, yielded a major peak corresponding to
LEAVIS
residues 22 to 39 plus a smaller peak containing a mixture of peptides, one of which appeared from amino acid analysis to be Cys-84 to Lys-86. No evidence was found for label on any other peptide derived from the protein. Thus we conclude that the DAB-Ma1 label is confined to the N-terminal half of cTnC. At pH 6.7 the luminescence of TblcTnCDAB upon uv excitation decays primarily (78% of the total emission intensity) with a short lifetime of 0.76 ms while 22% of the total intensity decays with an unquenched lifetime of 1.29 ms (Fig. 6, lower curve). The latter component is most likely due to incomplete labeling of the protein. The short-lived component is ascribed to an energy transfer between Tb3+ and the DAB label. The lifetime yields a distance of 3.4 nm. This distance, which is shorter than one would expect from the crystallographic structure, suggests that under these conditions the Cterminal domain of cTnC is relatively close to the Nterminal domain. When the pH is changed to 5.0, however, the metal luminescence of the same sample decays essentially with a single, unquenched lifetime of 1.31 ms (Fig. 6, upper curve), suggesting that the DAB moiety is now more than 5.2 nm (or 2 X R,J away from the bound Tb3+ (at sites III and IV). Similar results were obtained when the sample was excited directly with a laser light source (Table II).
I+
4
s *.2 3 Q)
1
2
Time
3
lmsl
FIG. 6. Luminescence decay curves for Tb,cTnt&s at pH 6.7 (lower curve) and Tb,cTnC& at pH 5.0 (upper curve) upon indirect excitation. The sample was excited with a uv flash lamp (1 channel = 0.04 ms). See Materials and Methods and the legend to Fig. 3 for other information.
CONFORMATION TABLE
OF CARDIAC
II
Summary of Luminescence Decay Parameters for Tb3’ Bound to Unlabeled and DAB-Labeled cTnC”
72
71 Sample Tb,cTnC
PH
Excitation
(ms)
C-4,)
6.7
“V “V laser’ laser “V U” laser laser
1.32 1.42
(100) (100) (100) (100)
5.0h Tb,cTnC Tb,cTnC,>m Tb,cTnCh
6.7 5.0 6.7 5.0 6.7 5.0
1.28
1.36 0.76 1.31 0.66 0.68
(ms)
(AZ)
(78)
1.29
(98) (100)
3.77
(22) (2)
(34)
1.32
(66)
’ Protein concentrations were the same as those in Table I. ’ pH was adjusted by adding 0.1 M HCl or 0.1 M NaOH. ’ For direct excitations a solution of coumarine 500 was used as the laser dye; X,, = 489.5 nm.
Since the luminescence of Tb3+ bound to unlabeled cTnC exhibits at both pH 6.7 and 5.0 a single lifetime of 1.3 ms (Table II), the observed change in the decay patterns indicates a pH-dependent structural transition in the cTnC molecule: at pH 5.0 cTnC exits in an elongated conformation as described by the X-ray diffraction studies; at neutral pH’s, however, cTnC assumes a more compact conformation so that the two end-domains are only 3.4 nm apart. This conclusion is consistent with the observations made with the Eu-Nd energy transfer (see above). Such a pH-dependent conformational change is very similar to what has been observed for sTnC in solution (10). The data presented in this paper, along with similar data for skeletal TnC (10) and calmodulin (19), call into question the relationship between the structures of these proteins determined from crystals grown at acidic pH and their solution structures at neutral pH values. In all cases the proteins assume a more compact structure than would be expected from the crystal data that brings their high and low affinity domains closer together. This suggests that the rigidity of the central helix implicit in the crystal structures is at least partially lost. The recent studies of Persechini and Kretsinger (7) in which the two domains of calmodulin are crosslinked by a bifunctional reagent spanning, at most, 1.9 nm provides direct corroborating evidence for a nonhelical bend between the two domains. If this is indeed the case, observations of interdomain communication in isolated skeletal TnC (8, 9) in which metal binding to one domain affects the metal binding kinetics and extrinsic fluorescence of a probe in the other are problematic insofar as it is difficult to imagine how a flexible central tether can act as a
TROPONIN
241
C
transducer of conformational change within the molecule. On the other hand, the possibility that hydrophobic pockets in the two domains may be brought into proximity by the helical bend such that they might jointly form a binding site for a target protein, as has been recently postulated for the calmodulin-myosin light chain kinase complex (20), opens new avenues for investigations of calmodulin and TnC complexes. ACKNOWLEDGMENTS The authors thank Miss Qian Zhan and Miss Elizabeth Gowell for their technical assistance, and also Dr. Renne Lu and Miss Anna Wong for the peptide sequence analysis.
REFERENCES 1. Herzberg, O., and James, M. N. G. (1985) Nature (London) 653-659. 2. Sundaralingam, M., Bergstrom, R., Strasburg, G., Rao, S. T., chowdhury, P., Greaser, M., and Wang, B. C. (1985) Science 945-948. M., Drendel, W., and Greaser, M. (1985) 3. Sundaralingam, Natl.
Acad.
4. Heidorn,
Sci. USA
82,
313, Roy227, Proc.
7944-7947.
D. B., and Trewhella,
J. (1988) Biochemistry
27,
909-
915.
5. Xu, G.-Q., and Hitchcock-DeGregori, 263, 13,962-13,969. 6. Putkey, J. A., Ono, T., VanBerkum, (1988)
J. Biol.
Chem.
263,
S. E. (1988) J. Biol. Chem. M. F. A., and Means, A. R.
11,242-11,249.
7. Persechini, A., and Kretsinger, R. H. (1988) J. Biol. Chem. 12,175-12,178. 8. Wang, C.-L. A., Leavis, P. C., and Gergely, J. (1983) J. Biol. 258,9175-9177. 9. Grabarek,
Z., Leavis, P. C., and Gergely, J. (1986) J. Biol.
263, Chem. Chem.
261,608-613. 10. Wang, C.-L. A., Zhan, Q., Tao, T., and Gergely, J. (1987) J. Biol. Chem.
262,9636-9640.
11. Potter, J. D. (1982) in Methods in Enzymology (Frederiksen, D. W., and Cunningham, L. W., Eds.), Vol. 85, pp. 241-263, Academic Press, San Diego.
12. Wang, C.-L. A., Tao, T., and Gergely, J. (1982) J. Biol. Chem. 257,8372-8375. 13. Stryer, L. (1978) Annu. Reu. Biochem. 47,819-846. 14. Fairclough, R. H., and Cantor, C. R. (1978) in Methods in Enzymology (Hirs, C. H. W., and Timasheff, S. N., Eds.), Vol. 48, pp. 347-379, Academic Press, San Diego. 15. Mulqueen, P., Tingey, J. M., and Horrocks, W. Dew., Jr. (1985) Biochemistry
24,6639-6645.
16. Horrocks,
W. Dew., Jr., Rhee, M.-J., Snyder, A. P., and Sudnick, D. R. (1980) J. Amer. Chem. Sot. 102,3650-3652. 17. Wang, C.-L. A. (1986)J. Biol. Chem. 261, 11,106-11,109. 18. Leavis, P. C., and Kraft, E. L. (1978) Arch. Biochem. Biophys.
186,411-415. 19. Wang, C.-L. A. (1989) Biochemistry 28,4816-4820. 20. Persechini, A., and Kretsinger, R. H. (1988) J. Cardiouasc. macol. 12, Suppl. 5, Sl-S12.
Phar-
