(Received for publication, October 28, 1996)
From the Institute of Biological Chemistry, University of Parma,
43100 Parma, Italy, Pharmacia & Upjohn, Nerviano,
20014 Milano, Italy, § Institute of Chemistry, Department of
Biochemistry, University of São Paulo, CP 26077, 05599-970,
São Paulo, Brasil, and the ¶ Department of Biophysics,
Federal University of São Paulo, São Paulo CP
04044-020, Brasil
Angiotensin II AT1A receptor is
coupled to G-protein, and the molecular mechanism of signal
transduction is still unclear. The solution conformation of a synthetic
peptide corresponding to residues 300-320 of the rat AT1A
receptor, located in the C-terminal cytoplasmic tail and indicated by
mutagenesis work to be critical for the G-protein coupling, has been
investigated by circular dichroism (CD), nuclear magnetic resonance
(NMR) and restrained molecular dynamics calculations. The CD data
indicate that, in acidic water, at concentration below 0.8 mM, the peptide exists in a predominantly coil structure
while at higher concentration it can form helical aggregates; addition
of small amounts of trifluoroethanol induces a secondary structure,
mostly due to the presence of helical elements. Using NMR-derived
constraints, an ensemble of conformers for the peptide has been
determined by restrained molecular dynamics calculations. Analysis of
the converged three-dimensional structures indicates that a significant
population of them adopts an amphipathic -helical conformation that,
depending upon experimental conditions, presents a variable extension
in the stretch Leu6-Tyr20. An equilibrium with
nonhelical structured conformers is also observed. We suggest that the
capability of the peptide to modulate its secondary structure as a
function of the medium dielectric constant, as well as its ability to
form helical aggregates by means of intermolecular hydrophobic
interactions, can play a significant role for G-protein activation.
Mutational studies on angiotensin II AT1A1 receptor have indicated that residues or sequences of the cytoplasmic tail and loops, close to the membrane, are critical for G-protein coupling (1-4). As far as the molecular mechanism, although the current working hypothesis is yet very speculative, it is certainly accepted that contacts between the receptor cytoplasmic portions and G-protein must take place and, therefore, that the secondary structure of those sequences is of great importance. The modulation of the coupling due to endocytosis or to a shift, for all the cytosolic components, between some sort of an aggregate state and an "open" one are among the mechanisms proposed (5). It is interesting to note, however, that from a conformational point of view, both mechanisms might require an equilibrium between a helical and an extended conformation.
As far as the receptor C-terminal (CT) portion, the 21-mer near the membrane surface, corresponding to the sequence 300-320 in the rat AT1A receptor, has been predicted to have a high probability to assume an amphipathic helical structure. Moreover, recently, it has been shown that the synthetic peptide representing the fragment 306-320 of the rat AT1A receptor is able to activate purified Gi1, Gi2, and G0 proteins (3).
These observations have prompted us to investigate the solution conformation of the synthetic fragment LFYGFLGKKFKKYFLQLLKYI-NH2, fCT300-320, to verify if indeed it may assume a helical structure and if its secondary structure can be modulated by physical or chemical, physiologically relevant perturbations.
The investigation of the solution conformation of fCT300-320 has been carried out by means of circular dichroism (CD) and two-dimensional 1H-NMR spectroscopy together with restrained molecular dynamics calculations. The peptide was studied in water, as well as in acidic H2O, 2,2,2 trifluoroethanol (TFE) mixtures, at 28 and 5 °C.
The results obtained show that, in acidic water,
fCT300-320 exists in a predominantly coil
structure at concentration lower than 0.8 mM, whereas at
the NMR concentration, it tends to form soluble helical aggregates; in
the presence of TFE, the peptide does not aggregate and its structure
is characterized by a well-defined amphipathic -helix involving the
stretch Leu6-Tyr20, with an increased
flexibility of the helix in the N-terminal segment. Most importantly,
the data clearly demonstrate the existence of an equilibrium between
helical conformations and extended ones: changes in the dielectric
constant of the medium, as well as temperature variations, can modulate
the extent of the helix, and hydrophobic intermolecular interactions
can favor the formation of helical aggregates. All of this evidence
indeed supports the various models and functional evidence requiring a
coil to helix transition of the AT1A receptor CT region
associated to helix-helix interaction as a possible molecular mechanism
for G-protein activation.
In this work, on the basis of both the obtained results and the current knowledge, we outline a possible general molecular model accounting for the various functions in which the CT tail of the AT1A receptor may be involved.
The 21-amino acid residues peptide
fCT300-320, corresponding to the fragment
300-320 of the cytoplasmic CT domain of the angiotensin II rat
AT1A receptor, was synthesized manually in a 0.2 mM scale under the tert-butyloxycarbonyl
protocol. It was purified by high performance liquid chromatography,
the amino acid composition was analyzed, and the primary sequence was
determined. Fig. 4 shows the sequence with both the numbering for the
peptide residues and the one from the primary structure of rat
AT1A (6).
Circular Dichroism Spectroscopy
CD spectra were recorded at
28 and 5 °C on a Jobin Yvon CD6 spectropolarimeter using a
thermostatic water bath to control the cell temperature. The instrument
was routinely calibrated with an aqueous solution of recrystallized
d-10-camphorsulfonic acid. Ellipticity is reported as mean
residue molar ellipticity, [], (deg cm2
dmol
1). The peptide's concentration was determined
spectrophotometrically from the tyrosine absorbance, using an
274 value of 1412 M
1
cm
1 and amino acid analysis. The concentration-dependence
study, in the range of 0.2-1.2 mM, was performed by
dissolving the peptide in water at pH 4.0 (adjusted using small amounts
of NaOH or HCl solutions). The H2O/TFE cross-titration
experiments were carried out mixing the appropriate aliquots of two 0.2 mM stock solutions, one in water at pH 4.0 and the other in
TFE.
The helical population was estimated by using the mean residue
ellipticity at 222 nm []222 according to the equation
of Chen et al. (7), [
]
= (fH
ik/N)[
]H
, where
[
]
is the observed mean residue ellipticity at
wavelength
, fH is the fraction of helix in
the molecule, i is the number of helical segments,
k is a wavelength-dependent constant (2.57 at
222 nm), N is the total number of residues, and
[
]H
is the maximum mean residue ellipticity for a
helix of infinite length. The expected value of the mean residue
ellipticity for the peptide in a 100% helical conformation was
calculated to be
34666 deg cm2 dmol
1.
The peptide for the NMR sample was
dissolved in 0.7 ml of 70% H2O, 30%
TFE-d3 (v/v), pH 4, to yield a peptide
concentration of 1.6 mM. TFE has been widely used as a
solubilizing and structure-inducing cosolvent for peptides and protein
fragments; in fact, a strong correlation between TFE-enhanced
-helicity and the intrinsic propensity of the sequence to adopt this
structure has been extensively demonstrated (8). In this case though
fCT300-320 is soluble in water, at the NMR
concentration, it tends to form soluble helical aggregates that do not
allow a proper analysis of the NMR spectra. Therefore, an
H2O/TFE mixture has been used to overcome this problem. All
two-dimensional 1H-NMR experiments were recorded with a
pulse field gradient (z axis; 1H, {13C,
15N}) triple resonance probe on a Varian Unity 600 spectrometer, at temperatures of 28 and 5 °C. The proton chemical
shifts were referenced to the H2O signal located at 4.774 ppm at 28 °C and 5.05 ppm at 5 °C, downfield from the external
4,4-dimethyl-4-silapentane-1-sulfonate (0.00 ppm). To assign the
peptide resonance peaks, standard methods were used to perform DQF-COSY
(9, 10), double quantum-two-dimensional (11), clean TOCSY (12), and
NOESY (13, 14) experiments. The TOCSY spectra were acquired using a
Waltz-16 spin-lock sequence at a field strength of 10 kHz and spin-lock
evolution time of 80 ms; NOESY experiments with several mixing times,
(
m), in the range between 150 and 300 ms were carried
out. All two-dimensional experiments were acquired in the
phase-sensitive mode using the method of States et al. (15),
and quadrature detection was achieved using hypercomplex data
acquisition. The spectral width in both dimensions was typically 10 ppm, and 512
1 experiments with 32 transients of 2048 complex points for each free induction decay were recorded. In both
DQF-COSY and TOCSY spectra, water suppression was achieved by low power
continuous wave irradiation during the relaxation delay, while in the
NOESY experiments, the Watergate technique (16) was employed.
Data were processed on a Silicon Graphics Personal Iris workstation using the FELIX NMR 2.3 data processing package (BIOSYM Technologies, San Diego, CA). Prior to Fourier transformation, the time domain data were zero-filled in both dimensions to yield 4 K × 2 K matrices and apodized by a 30°-shifted squared sinebell window function to improve resolution; for quantitative measurements of NOESY data, a 90°-shifted squared sinebell window function was applied to avoid distortion of peak intensities. When necessary, a fifth-order polynomial base-line correction algorithm was applied after transformation and phasing.
To obtain distance constraints, cross-peak intensities were estimated
from the 150 ms NOESY spectra recorded at 28 and 5 °C. To relate
these NOE data with interproton distances, a calibration was made using
the distance of 1.8 Å for the well defined geminal -protons, and
the NOE intensities were classified as strong, medium, and weak,
corresponding to upper bound constraints of 3.0, 4.0, and 5.0 Å,
respectively. Lower bounds were taken to be the sum of the van der
Waals radii (1.8 Å) for the interacting protons in all cases. Upper
distance restraints involving nonstereospecific assigned methylenes,
aromatic, and methyl protons were replaced by appropriate pseudo-atoms
(17).
The dihedral angle restraints were derived from the
3JHN-
coupling constants,
estimated in the DQF-COSY spectra, from measurements of separations of
extrema in dispersive and absorptive plots of rows through cross peaks
(18). In these experiments, the digital resolution after zero-filling
along
2 was 0.43 Hz/point; multiple solutions for each
J value were selected.
All calculations were carried out on a Silicon Graphics ONYX computer using the Consistent Valence Force Field (19, 20), implemented in the DISCOVER (BIOSYM Technologies Inc., San Diego, CA) software package, together with the INSIGHT II as graphic interface. The structures of fCT300-320 were computed using the simulated annealing methods (21) in the NMR refine module of the INSIGHT II package. The simulated annealing calculations included some distinct phases. Phase 1 involved randomization of all atomic coordinates followed by minimization of the starting structures using a quadratic potential and very low force constants for each term of the pseudo-energy function, including chiral and NOE constraints. Phase 2 involved simulated annealing with a progressive increment of the force constants up to their full values. Phase 3 involved cooling of the molecule from 1000 to 300 K over 10 ps. At the end of this protocol, the structures were energy minimized with full Consistent Valence Force Field (Morse and Lennard-Jones potentials, coulombic term) by steepest descents and conjugate gradients using several thousand iterations until the maximum derivative was less than 0.001 kcal/Å. Amide bonds were maintained as trans during the calculations. The quality of the final structures was analyzed on the basis of the number of NOE distance violations and of the root mean square deviation values of the backbone.
Fig. 1 shows the CD spectra
for fCT300-320, as a function of concentration,
in water, at pH 4.0 and 28 °C. At concentration higher than 0.8 mM, the occurrence of a transition from an extended
conformation to a helix reach one is indicated by the shift of the
minimum from 198 to 208 nm, together with the net increase of the
negative band at 223-224 nm. On the basis of the EMBL structural
prediction (22, 23)2 and according to the
NMR data (see "NMR Spectroscopy"), the change of the CD spectra as
a function of the peptide concentration can be indicative of the
formation of helical aggregates. In fact, the slight red shift of the
n-* transition would support this hypothesis. The possibility of
formation of coiled coil structures is discarded as the negative band
at 208 nm, due to the parallel-polarized
-
* transition, does not
undergo any shift and the [
]220/[
]208 ratio is relatively constant for concentration above 0.8 mM
(7, 24).
Fig. 2, A and B, shows the CD
spectra of 0.2 mM fCT300-320 at
various H2O/TFE percentages (v/v), pH 4, at 28 and
5 °C, respectively. At both temperatures, upon addition of TFE as
low as 10%, a significant change of the CD spectrum is observed. The
shift of the negative band from 198 to 208 nm, together with the
appearance of an additional negative band at 222 nm and of a positive
band at 196 nm, suggest that the peptide is switching from an unordered
state to a helix-rich one. The absence of any isodichroic point
indicates the existence of an equilibrium between multiple
conformations.
As observed in the concentration dependence study, both at 28 and
5 °C, when increasing the TFE concentration, the behavior of the
parallel-polarized -
* transition at 208 nm does not show any red
shift and any modification of the
[
]220/[
]208 ratio is detected, thus
confirming that the TFE-induced helical structures do not tend to
aggregate in a coiled coil fashion.
The analysis of the spectra using the equation of Chen et
al. (7) indicates that, at 28 °C, the -helix content is
50% at 30% TFE, and here it stabilizes all the way through 100%
TFE (Fig. 2A, inset). When the temperature is
lowered at 5 °C, a plateau is reached at 40% TFE with a helix
content of about 60% (Fig. 2B, inset).
Initially,
one- and two-dimensional 1H-NMR experiments of
fCT300-320 were recorded in water in the
concentration range of 1.5-3.0 mM, varying the pH between
3.0 and 6.0 (data not shown). Under these conditions, the proton
resonances line-width were larger than expected for a peptide of this
size, in agreement with the presence of intermolecular aggregation as
it was suggested by the CD data. Consequently, according to the CD
results, the peptide was dissolved in 70% H2O, 30%
TFE-d3 (v/v). In this condition, the NMR
resonance lines presented the expected high resolution shape, allowing
a proper analysis of the spectra. Following to the sequential
assignment procedure (17), TOCSY spectra were used to identify the
amino acids spin systems, and NOESY spectra were used both to obtain
interresidue connectivities and to distinguish between equivalent spin
systems. Assignments of the amino acids side-chain resonances were
confirmed by DQF-COSY and double quantum-two-dimensional spectra;
aromatic amino acids (Phe, Tyr) were identified using their
-aromatic NOESY connectivities. The interpretation of the two-dimensional 1H-NMR spectra presenting a unique set of
resonances suggests that the peptide is either in a monomeric state or
in a symmetric dimeric one. All protons have been identified, and the
chemical shifts are listed in Table I for 28 and
5 °C, together with the NH-
CH coupling constants
(3JHN-
).
|
Several NMR parameters
provide information on the amount of secondary structure of peptides in
solution. A fast and simple method, complementary to the analysis of
the NOE connectivities, is based on the observed deviations of CH
proton chemical shifts with respect to the random coil values (25). The
-protons secondary shifts for fCT300-320 are
reported in Fig. 3, and it is evident that, except for
Phe2, Leu6, and Ile21 at 28 °C
and only for Phe2 at 5 °C, all the
-protons
resonances move upfield with a slightly higher deviation at 5 °C,
thus suggesting the existence of a helical region extending between
Lys8 and Tyr20 at both temperatures. The more
pronounced deviation of the secondary
CH chemical shifts observed in
the C-terminal part of the peptide indicates that the helical structure
in that portion is more stable as compared with the N-terminal
part.
The summary of the interresidue NOEs at 28 °C is reported in Fig.
4A, showing a good correlation with the CH
secondary shifts. In fact, medium and strong connectivities are found
between the adjacent amide protons from Phe10 to
Ile21 as can be seen also in the NH-NH region of the 150 ms
NOESY spectrum displayed in Fig. 5A, indeed
supporting the existence of a significant population of conformers
containing a well defined
-helix in that part of the peptide.
On the other hand, the values of the
3JHN- coupling constants, mostly
around 6-7 Hz (Table I), together with the relatively weak intensities
for the (i, i + 3) NOEs and the coexistence with the intense C
H-NH (i, i + 1) NOEs from the
middle to the N terminus of the peptide suggest the presence of
conformational fluctuations in that stretch.
The data collected at 5 °C show all the cross-peaks found at
28 °C, together with some new NOEs involving residues in the N-terminal region, Figs. 4 and 5B. The presence of extensive
medium range CH-NH (i, i + 3) and C
H-C
H
(i, i + 3) NOEs in the stretch between
Phe5 to Tyr20 suggests that almost the entire
peptide exists in an
-helical conformation. A large number of NH-NH
(i, i + 1) NOEs from Tyr3 to
Ile21, together with the majority of the
3JHN-
in the range of 4.7-6.0
Hz, strongly support this possibility.
Qualitative analysis of NOE patterns only
provides a crude description of the secondary structure of
fCT300-320. The possible three-dimensional
structures of the peptide were calculated following the molecular
dynamics protocol described under "Materials and Methods," using a
final set of 187 NOE-derived distance constraints at 28 °C and 220 distance constraints at 5 °C (Table II) together with
18 dihedral angles restraints at both temperatures.
|
A total of 100 structures were calculated for each temperature, of which 92 were energy minimized using the program DISCOVER; 8 structures were rejected because of the presence of D-amino acids, topological mirror images, and misfold. After minimization, 18 structures were selected, at both temperatures, based on the low residual distance violations (less than 0.4 Å) and low dihedral angles violations (less than 5°) and were used to represent the solution three-dimensional structure of fCT300-320. Table II shows the statistics for the selected structures as well as their energies.
Fig. 6, A and B, displays the 18 final Molecular Dynamics structures of
fCT300-320, for both temperatures, superimposed
for the minimum backbone deviation between residues
Lys12-Lys19 at 28 °C and residues
Leu6-Lys19 at 5 °C. As shown in Fig.
6A, at 28 °C, an -helical folding is well defined in
the C-terminal portion of the peptide. Indeed, at variance with the
observed NOE pattern, Fig. 4A, on the basis of the analysis
of the main chain hydrogen bonds and of the dihedral angles
and
, carried out using the program DSSP (26), a stable
-helix
results to extend only between residues Phe14 and
Lys19. The wide conformational spread observed in the
N-terminal portion of the peptide can be interpreted as indicative of
the presence of a structurally flexible
-helix.
Fig. 6B shows that, at 5 °C, a well defined -helical
conformation extends over the peptide stretch
Leu6-Tyr20, confirming that the temperature
exerts a stabilizing effect on the helical structure. The region
containing residues 1-6 is still poorly defined due to the limited
number of constraints available. In addition, it should be noted that,
in this part of the peptide, there are two Gly residues known to
promote an increase in
and
angular freedom and flexibility.
The hydrogen bonds observed for the converged structures of the peptide at both temperatures are listed in Table II and are in agreement with the other data.
The bottom of Fig. 6, A and B, presents a
schematic representation of the least-energy structure of the peptide
where the amphipathic nature of the -helix at both temperatures is
evident.
The results presented in this study allow delineating the
preferential solution conformations of the synthetic linear
peptide fCT300-320, corresponding to
the fragment 300-320 of the CT cytoplasmic tail of the rat
AT1A receptor and indicated to be essential for the
interaction with the G-protein -subunit. The obtained structures can
be related to the active form in the AT1A receptor as it
has been shown that the peptide is able to activate G-protein GTP binding (3).
d3 shows a rather stable helical conformation spanning from the central part of the peptide toward the C-terminal end as it is shown in Fig. 6. In acidic water, below 0.8 mM, the peptide is in a predominantly extended conformation, whereas at higher concentration, it tends to form helical aggregates (Fig. 1).
The molecular mechanism by which the AT1A receptor
interacts with G-protein is yet to be clearly understood. However, all models proposed so far imply the need of a contact between some cytosolic fragments of the receptor and portions of the G-protein. To
achieve this goal, the conformational properties of these fragments clearly play a crucial role and, in particular, the existence of a
helix to coil transition seems to be the most probable molecular mechanism that may permit the association and subsequent dissociation between the AT1A receptor and the G-protein
-subunit.
Our results show that the conformation of fCT300-320 is sensitive both to the polarity of the medium and to the temperature in a way that the peptide can be forced to shift from an extended configuration to a helical one whose extension can be modulated by the strength of the external agents.
In the AT1A receptor, the 300-320 stretch is expected to be at the membrane interface, and the amphipathic property of the membrane could be an ideal factor to modulate its conformation. It has been suggested that AT1A receptor internalization might have conformational rather than functional requirements (27). In this respect, a variable interaction between the membrane and the AT1A CT tail could be favored by the conformational characteristics of that fragment.
Theoretical predictions of fCT300-320 structure, verified by our experimental results, have provided evidence for the amphipathic nature of the helix and concentration dependence studies have indicated the tendency to form helical aggregates. In addition, there is evidence, reported by Conchon et al. (27), indicating that the binding of the agonist to the receptor induces a lateral shift of the complex, and various authors are stressing the importance of the presence, in the receptor structure, of stretches that may utilize the charged surface of an amphipathic helix to achieve G-protein activation (28-30). All combined, it becomes quite tempting to speculate on a possible general molecular model for the many events triggered by ligand binding in the extracellular portion of the AT1A receptor.
The propagation of the agonist-induced conformational changes through
the transmembrane segments results in a lateral shift of the receptor
that might bring its cytoplasmic portions closer to the G-protein. This
event would favor the interaction of the polar part of the receptor CT
tail amphipathic helix with the G-protein -subunit forming one of
the anchoring points between the two proteins, the other being with the
third intracellular loop (3). Concomitantly, affinity reasons would
bring the hydrophobic part of the helix, still exposed to the medium,
to interact with the nearby membrane surface, and this interaction
might be responsible for the receptor internalization. Interestingly,
one of the internalization motifs proposed for many G-protein coupled
receptors is NPXXY (31) that, though in this case there are
yet some doubts about its function, is present in the AT1A
receptor, and it is located in position 298-302. The presence of
Pro299 at the membrane interface is quite relevant since
Pro residues are often observed at the N terminus of
-helices (32).
In addition, similarly to what has been suggested for the
internalization motif of an integral membrane protein of the
trans-Golgi network (33), Pro299, being part of
the internalization motif, could define the beginning of the helical
stretch assuring, at the same time, a certain degree of flexibility
that may be required for a better interaction with the membrane.
Termination of the signal has been associated to a desensitization
mechanism that can be induced by phosphorylation of the Ser and Thr
residues of the receptor within the CT region (34). Indeed this event,
changing the polarity of the surrounding, could induce unfolding of
that fragment and decouple the receptor from the G-protein
-subunit.
In conclusion, there are no doubts that distinct opinions regarding the role of the AT1A CT tail are still present. In fact, while some authors are primarily stressing its role in the receptor internalization process induced by agonist binding (35, 36), others indicate a specific involvement in G-proteins interaction (1, 3, 4).
Based on our results, we believe that the receptor CT tail may be involved in all of these processes and that, due to the complexity of the events associated to the formation of the ternary complex, including the agonist, the receptor and the heterotrimeric G-protein, the experimental conditions used so far by the various researchers do not allow observing all its functions simultaneously, thus biasing the definition of a correct molecular model.
As suggested by Yeagle et al. (37) in relation to rhodopsin,
we believe that, for the AT1A receptor as well, studying
all the soluble peptides of the cytoplasmic portions of which the biological activity can be verified will allow delineating a profile of
the receptor surface that is involved in the interaction with the
G-protein -subunit.
Studies are now underway to better define the flexibility of the CT region as well as to describe the structural properties of the other cytosolic portions of the AT1A receptor involved in G-protein coupling and signal transduction.
The Interfaculty Center for Measurements (CIM) is kindly acknowledged for the use of the NMR and computing facility.