Conformational Statistics and Energetics Analysis of Sequential

Conformational Statistics and Energetics Analysis of Sequential Peptides Undergoing Intramolecular Transfer of Excitation Energy ... Conformational st...
0 downloads 0 Views 704KB Size
Download PDF
J. Phys. Chem. 1996, 100, 6835-6844

6835

Conformational Statistics and Energetics Analysis of Sequential Peptides Undergoing Intramolecular Transfer of Excitation Energy B. Pispisa,*,† A. Palleschi,‡ M. Venanzi,† and G. Zanotti§ Dipartimento di Scienze e Tecnologie Chimiche, UniVersita´ di Roma “Tor Vergata”, 00133 Roma, Italy, Dipartimento di Chimica, UniVersita´ di Roma “La Sapienza”, 00185 Roma, Italy, and Centro di Chimica del Farmaco del CNR, c/o UniVersita´ di Roma “La Sapienza”, 00185 Roma, Italy ReceiVed: NoVember 2, 1995; In Final Form: January 3, 1996X

The photophysics of short linear peptides of general formula Boc-Leu-Leu-Lys(P)-(AA)n-Leu-LeuLys(N)-OtBu, where AA ) Ala or Aib (R-aminoisobutyric acid), and P and N are protoporphyrin IX and naphthalene, respectively, covalently bound to -amino groups of lysine side-chains, were investigated in water/methanol 75/25 (v/v) solution by steady-state and time-resolved fluorescence experiments. Quenching of the excited naphthyl chromophore takes place by electronic energy transfer to the porphyrin ground state, and proceeds on a time scale of 3-8 ns. A minor and slower fluorescence lifetime measures the decay of the exciplexes. Quenching efficiencies exhibit a different trend, depending on whether AA ) Ala or Aib, indicating differences in the structural features of the two series of peptides. Consistently, CD spectra suggest that the former compounds populate R-helical structures, while the latter ones possibly attain a 310-helix conformation, in agreement with the proven ability of Aib to form 310-helices in solution. The increased percentage of intramolecular H-bonds in the P(Aib)nN as compared to the corresponding P(Ala)nN peptides, as determined by IR spectra in dilute CD3OD or CDCl3 solution, confirms this conclusion. The fluorescence results were satisfactorily described by a dipole-dipole interaction mechanism, provided the mutual orientations of N and P groups are taken into account, which implies that interconversion among conformational substates of chromophore linkage is slow on the time scale of the transfer process. Conformational statistics analysis shows a rather wide interprobe separation distance distribution for each peptide, owing to the aliphatic portion of the side-chains carrying the chromophores, but theoretical conformational analysis indicates that only a few energetically favored conformers are the major contributors to the energy transfer process.

Introduction A variety of molecules carrying covalently bound porphyrins has been used as model systems for investigating both energy and electron transfer processes,1-5 which represent the natural photochemical pathways for light harvesting.4,6 Typically, the moieties involved in the excitation energy transfer interact by a singlet-singlet mechanism, via a long range dipole-dipole interaction process or via exchange interactions at very short interprobe distances, or by a triplet-triplet mechanism, which is forbidden by the above mechanisms and must occur by an electron-exchange process. In the singlet-singlet transfer pathway both spectral and geometric requirements must be matched. According to Fo¨rster’s theory7 the emission spectrum of the donor must overlap with the absorption spectrum of the acceptor, and the emission and absorption dipoles of the donor and acceptor molecules, respectively, must be at a certain distance and properly oriented. In the Dexter mechanism8 electronic orbital overlap of the donor and acceptor is instead required, which implies a direct contact between the partners. In these studies macromolecular materials, especially synthetic polypeptides, were often employed as carriers to mimic a natural environment,9,10 and the fluorescent probes have also proven to be a valuable tool for obtaining structural and dynamic information on the polymeric matrix in solution. Indeed, † Dipartimento di Scienze e Tecnologie Chimiche, Universita ´ di Roma “Tor Vergata”. ‡ Dipartimento di Chimica, Universita ´ di Roma “La Sapienza”. § Centro di Chimica del Farmaco del CNR, c/o Universita ´ di Roma “La Sapienza”. X Abstract published in AdVance ACS Abstracts, March 15, 1996.

0022-3654/96/20100-6835$12.00/0

Figure 1. Spectral overlap between the normalized donor stimulated fluorescence, FD (full line), and the acceptor extinction coefficient,  (broken line).

determination of interdomain distance distributions and domain motions in globular proteins is usually recovered from sitespecific labeling and fluorescence energy transfer by the timeresolved emission parameters.11,12 We have recently investigated the photophysics of naphthalene (N) and protoporphyrin IX (P) covalently linked to the sidechains of lysine residues in both polypeptides13 and short linear peptides.14 Under conditions where both types of materials attain an ordered conformation, quenching of excited naphthalene takes place by electron transfer from ground-state porphyrin, P f 1N*, in R-helical polypeptides, and by transfer of excitation energy, 1N* f P, in the helical peptides. Figure 1 shows the spectral overlap of the normalized donor fluorescence with the acceptor extinction coefficient. We are currently assessing the issue of the role played by electronic and Franck-Condon factors in this transition, focus© 1996 American Chemical Society

6836 J. Phys. Chem., Vol. 100, No. 16, 1996

Figure 2. Probability distribution of interchromophoric distances in the R-helical P(Ala)nN peptides investigated. Curves a f e correspond to n ) 0 f 4. The helix periodicity is easily recognizable by the trend of the maxima.

CHART 1

ing on the different conformational rigidity, and hence nuclear mobility, between the ordered backbone chains of the polypeptides and short linear peptides. The peptides examined, whose sequence is presented in Chart 1, show rather wide distribution probabilities of center-to-center distance between N and P molecules, owing to the conformational mobility of the aliphatic portion of the chromophore linkages. This is illustrated in Figure 2 for P(Ala)nN, according to conformational statistics analysis.14a,b Nevertheless, the idea was that only a few, highly populated, low-energy conformational states are actually involved in the transfer process, because the structural requirements for an efficient energy transfer are believed to be rather severe so as to optimize the separation distance and mutual orientation of the probes. To verify this hypothesis, we performed a theoretical conformational analysis on the aforementioned peptides. We report here the results of this investigation, together with new data on the photophysics of the peptides in water/methanol 75/25 (v/v) solution. For comparison, we also summarize the most relevant data of the conformational statistics analysis, as well as some of the spectroscopic and fluorescence results previously reported.14 Computational data confirm the foregoing idea, emphasizing the care that must be taken when using conformational statistics data alone in interpreting excited-state phenomena. They shed further light on the relationship between the observed effects and structural features of the compounds examined. Experimental Section Materials. The series of sequential peptides of Chart 1 are formed by two Leu-Leu-Lys triads (Leu ) L-leucine; Lys ) L-lysine) linked together through a spacer of n ) 0-4 L-alanine (Ala) residues or n ) 1-2 R-aminoisobutyric acid (Aib)

Pispisa et al. residues. These compounds are denoted as P(Ala)nN or P(Aib)nN, respectively, while the corresponding blank samples as (Ala)nN and (Ala)nP or (Aib)nN and (Aib)nP, depending on the bound chromophore. Owing to the presence of two carboxylic groups in P, both monomeric and cross-linked dimeric protoporphyryl peptides were synthesized, but only the monomeric species will be reported here. Amino acid coupling was carried out either by the mixed anhydride method (MA) or by the dicyclohexylcarbodiimidehydroxybenzotriazole method (DCCI-HOBt). tert-Butyloxycarbonyl (Boc) and tert-butyl ester (OtBu) were used to protect the amino and carboxyl terminus, respectively. The identity and purity of the intermediates were checked by thin layer chromatography on silica gel plates (Merck), mass spectrometry, and NMR spectroscopy. Extensive chromatographic purification of the final materials led to optical purities higher than 90%, according to NMR measurements. All peptide syntheses were extensively described elsewhere.14b,c Methods. Steady-state fluorescence measurements were carried out on a SPEX Fluoromax spectrofluorometer, operating in SPC mode (λex ) 280, λem ) 340 nm), as already reported.14 In all cases the solutions were previously bubbled for about 20 min with ultrapure nitrogen. Lifetime experiments were carried out at the European Laboratory of Non-Linear Spectroscopy (Florence) by using a dye laser (Rhodamine 6G) as the excitation source, pumped by a ML-NdYAG laser, and SPC detection. The dye laser output (≈20 µJ at 580 nm) was then doubled by a 1 cm KDP crystal to obtain the 290 nm radiation used in these experiments. The decay curves were fitted by a nonlinear least-squares analysis to exponential functions by an iterative deconvolution method. Transient absorption experiments were carried out by the flash-photolysis setup already reported.14b Absorption spectra were recorded on a Jasco 7850 apparatus, and circular dichroism (CD) measurements were carried out on a Jasco J-600 instrument with appropriate quartz cells. Molar ellipticities are given on a per residue basis, i.e., [Θ] ) (Θ/ (naCl)), where Θ is the ellipticity (mdeg), C the peptide concentration (mM), na the number of amino acids in the peptide, and l the optical path length (cm). IR spectra were carried out on a Perkin-Elmer 983 IR spectrophotometer, using CaF2 cells. 1H-NMR spectra were recorded on a Bruker AM 400 instrument operating at 400.135 MHz. Other apparatuses were already reported.13,14 Results and Discussion Steady-State and Time-Resolved Fluorescence Data. A typical example of steady-state fluorescence spectra in water/ methanol 75/25 (v/v) solution (λex ) 280 nm) is illustrated in Figure 3. A substantial quenching of N singlet emission by the bound protoporphyrin molecule can be noted, while the amount of exciplex is rather small very likely because of the polarity of the solvent, since the exciplex emission markedly decreases on going from methanol to the mixed solvent.14 The fluorescence quantum yield of the oligopeptides in the mixed solvent (ΦPN) and the efficiency of the quenching process, EN, as given by [1 - (ΦPN/ΦN)], are reported in Table 1, ΦN being the quantum yield of the peptides carrying the N group only.15 While the quenching efficiency, EN, in methanol solution parallels with that of the fluorescence rise in P, EP, thus indicating that the energy lost in the deactivation of the excited naphthyl chromophore is nearly completely transferred to the porphyryl group, the quenching of N* in water/methanol 75/25 shows that the fluorescence quantum yield of P is negligibly

Analysis of Sequential Peptides

J. Phys. Chem., Vol. 100, No. 16, 1996 6837 TABLE 2: Relative Quantum Yields of Porphyrin Emission as a Function of Added Watera % H2O (by vol) sample

0

25

50

75

P(Ala)2N (Ala)2P

1 1

0.604 0.702

0.181 0.401

0.057 0.028

a λ ) 308 nm. The emission range explored was 600-720 nm; ex sample concentration 1.4 × 10-6 M. Quantum yield uncertainty within 10%.

Figure 3. Fluorescence spectrum of P(Aib)2N and P(Ala)2N (a and b) and of the corresponding blank samples (Aib)2N and (Ala)2N (c and d) in water/methanol 75/25. Insert: wavelength region where exciplex emission occurs.

TABLE 1: Energy Transfer Efficiencies in Water/Methanol 75/25 (v/v) and Methanol, from Steady-State Measurements H2O/CH3OH a

sample

EN

P(AA)0N P(Ala)1N P(Ala)2N P(Ala)3N P(Ala)4N P(Aib)1N P(Aib)2N

0.76 0.58 0.84 0.68 0.62 0.90 0.91

CH3OH

Ep

b

a

EN

Epb

0.61 0.59 0.82 0.77 0.55 0.98 0.81

0.58 0.53 0.78 0.66 0.43 0.95 0.81

a λex ) 280, λem ) 340 nm; EN ) 1 - (ΦPN/ΦN). ΦPN is the quantum yield of N* in P(AA)nN and ΦN that in the blanks, (AA)nN (ΦN ≈ 0.22). b λex ) 280, λem ) 630 nm; EP ) [(IPN/IP) - 1](PCP/NCN), where IP ) PCPΦP and IPN ) PCPΦP + NCNΦPEP are the fluorescence intensities of porphyrin in the blanks [(AA)nP] and in the P(AA)nN peptides, respectively, while CP and CN are the concentration of the acceptor and donor chromophores.

Figure 4. Variations of the relative absorbance (∆A/∆Amax) of the T-T transient band at 470 nm of P(Ala)2N (empty circles) and (Ala)2P (full circles) in methanol by titration with H2O. In both cases there is a complete bleaching of the absorption at 75% added water.

small, so that EP is unmeasurable. This peculiar behavior may be explained by taking into account the following observations: (1) Titration by H2O of the T-T transient absorption at 470 nm of, e.g., P(Ala)2N and (Ala)2P in methanol shows a complete bleaching of the band at 75% (by vol) added water (Figure 4) and (2) titration by H2O of methanol solutions of P(Ala)2N and (Ala)2P leads to a dramatic decrease in the quantum yields of porphyrin emission in the range 600-720 nm (λex ) 308 nm), as shown in Table 2. Both findings indicate that addition of water up to 75% into the methanol solution turns off the usual channels of excited-state relaxation of the

Figure 5. Typical example of the fluorescence decays. Normalized decay profile (λex ) 285, λem ) 340 nm) of P(Aib)2N (a) and (Aib)2N (b) in water/methanol 75/25. The full lines represent the best fit of the data by a two-exponential decay for P(Aib)2N and by a monoexponential decay for the blank. The laser profile is also shown.

porphyrin, i.e., interconversion to the triplet or the singlet state, opening new efficient channels for the dissipation of the energy transferred to P, such as, for instance, vibronic interactions of the bulky porphyrin with the surrounding water molecules. In other words, the intramolecular N* f P energy transfer mechanism of N* quenching observed in methanol is still operative in the mixed solvent medium, but the energy transferred to P is now very quickly consumed with nonradiative pathways. Aggregation effects of P on the observed phenomena should be minor, if any, because the P emission intensity in water/methanol 50/50 (v/v) follows Beer’s law by changing the concentration by 1 order of magnitude (10-6-10-5 M). By inspection of Table 1, it also appears that the quenching efficiency of P(Ala)nN in water/methanol goes through a maximum as n increases and exhibits a different trend as compared to P(Aib)nN. This suggests that the interprobe distance is controlled by factors other than simply the number of AA units in the backbone chain and that the two series of peptides are likely to populate different conformations in solution. We next investigated the fluorescence time decay of excited naphthalene in water/methanol 75/25 (v/v). The results are similar to those obtained in methanol solution.14a,b indicating that addition of water does not significantly alter the relaxation behavior of the systems investigated. A typical example is shown in Figure 5. In all cases, no significant change was observed on varying the sample concentration within 1 order of magnitude (2 × 10-6-2 × 10-5 M), so that interchain effects can be ruled out. The curves were well fitted by a two-component exponential decay, eq 1, while the time decay of the blanks, τ0, was always found to be strictly monoexponential.

I(t) ) R1 exp(-t/τ1) + R2 exp(-t/τ2)

(1)

The shorter decay time, τ1, ranges from 3 to 8 ns, depending on the number of AA residues in the spacer, and can be assigned to the energy transfer process. The longer decay time, τ2,

6838 J. Phys. Chem., Vol. 100, No. 16, 1996

Pispisa et al.

TABLE 3: Fluorescence Lifetimes of N* in H2O/CH3OH 75/25 (v/v) at 25 °Ca sample

τ1 (ns)

R1

τ2 (ns)

R2

χ2

E1b

P(AA)0N P(Ala)1N P(Ala)2N P(Ala)3N P(Ala)4N P(Aib)1N P(Aib)2N

7.8 5.4 4.9 8.3 8.7 3.0 2.5

0.88 0.56 0.95 0.83 0.71 0.97 0.96

45.5 45.0 45.0 45.0 45.0 44.0 45.5

0.12 0.44 0.05 0.17 0.29 0.03 0.04

1.4 1.7 1.4 1.4 1.8 1.2 1.3

0.79 ( 0.04 0.85 ( 0.06 0.86 ( 0.06 0.77 ( 0.05 0.76 ( 0.05 0.92 ( 0.06 0.93 ( 0.06

aλ ex ) 290, λem ) 340 nm; the time decay of N* in the (AA)nN peptides (blanks) is τ0 ) 36.2 ( 1.6 ns. The lifetime uncertainty is better than 5%, while that of the preexponents is around 20%. b Energy transfer efficiency, as given by eq 2.

measures the exciplex decay, being insensitive to both peptide composition and interchromophoric distance. This assignment is supported by the finding that decay measurements at λem ) 470 nm, where the N* emission is definitely minor, show a similar dominant long lifetime component, which is affected, unfortunately, by a rather large uncertainty, owing to the very low fluorescence emission within this wavelength region.

E1 ) 1 - (τ1/τ0)

(2)

Table 3 lists the decay times and the energy transfer efficiencies, eq 2. Inspection of the table indicates that the preexponents of both the short and long components vary with the length of the peptides. Anticipating that all compounds exhibit an ordered, intramolecularly H-bonded structure, where Aib replaces the Ala residues, the preexponent of the short component, R1, exhibits a different trend as the number of AA residue increases. Since this quantity measures the relative population of the species undergoing energy transfer, which in turn depends on the structural features of the matrix, this observation indicates that, if a different chain folding in the two series of peptides occurs, the same number of Aib or Ala intervening residues does not necessarily mean a similar population of donor-acceptor pairs undergoing energy transfer. On the other hand, the preexponent of the long component, R2, is much larger for the hepta- and decapeptide of P(Ala)nN than for the other samples, very likely because the probes in P(Ala)1N and P(Ala)4N are on the same side of the R-helix, one or two turns apart, respectively, favoring a face-to-face orientation, and hence exciplex formation.9b We then addressed the problem of the different structural features in the Ala- and Aib-based peptides by chiroptical and IR spectra. Circular Dichroism and IR Spectra. Earlier spectroscopic results14 have shown that even the shortest member of the series of P(AA)nN in dilute methanol solution exhibits the characteristic features of intramolecularly H-bonded structures.14b,c This was interpreted as due to the presence of the Leu and Ala or Aib residues in the backbone chain, having a high helix-forming potential,16,17 and of the N and P groups in the lysine sidechains, which play a concerted restructuring role on the peptides.14 Typical CD spectra in water/methanol 75/25 (v/v) are shown in Figure 6. The spectral patterns of the Ala- and Aib-based compounds exhibit marked differences. For instance, P(Ala)2N produces an R-helix-like spectrum, with the n,π* and the parallel component of π,π* transitions centered at around 220 and 209 nm, respectively, while P(Aib)2N exhibits only a pronounced dichroic band at 207 nm and a minor and broad band at about 235 nm. Since IR spectral results in dilute CD3OD or CDCl3 solution indicate that P(Aib)2N also attains an intramolecularly

Figure 6. CD spectra in water/methanol 75/25 (v/v) of P(Ala)2N (a), P(Aib)2N (b), (Ala)2N (c), and (Aib)2N (d). The ellipticity is on a per residue basis.

H-bonded structure (see below), one must conclude that the differences in the CD spectra of the two octapeptides reflect different helical conformations in solution. Similar arguments apply to the heptapeptides, P(Ala)1N and P(Aib)1N. Interestingly, the spectral patterns of (Ala)2N and (Aib)2N are such that only those of the former peptide are reminiscent of a random coil, with the lower energy band centered at around 197 nm. This indicates that Ala and Aib residues in P(AA)nN have different helical-forming propensity. As far as the IR spectra are concerned, the most relevant results for P(Ala)nN can be summarized as follows: (1) Appropriate dilution techniques18 minimize self-association, so that intermolecular H-bonding interactions are negligible for all the peptides examined at concentrations below 1 mM. (2) Under these conditions, the band in the amide-A region at around 3430 cm-1 can be assigned to the stretching mode of free peptide N-H and urethane groups while that in the frequency range 3330-3320 cm-1 to intramolecular hydrogen-bonded N-H groups. (3) The (Ala)nN or (Ala)nP peptides in CD3OD or CDCl3 solution at a concentration of about 0.05 mM exhibit a band of low intensity in the frequency region around 3330 cm-1 and a definitely more intense band at around 3430 cm-1. According to the integrated intensity of these bands, the H-bonded N-H groups in these peptides are less than 10%, but they dramatically increase in the compounds carrying both N and P chromophores. (4) Consistently, in the amide-I region (1600-1800 cm-1), corresponding to the CdO stretching vibrations of peptide, urethane, and ester groups,17,19,20 all P(Ala)nN exhibit a band at around 1650-1660 cm-1, corresponding to the CdO stretching of H-bonded peptide moieties.20 Instead, the CdO stretching of free peptide groups, normally observed in the 1670-1700 cm-1 region, was found in the Alabased blanks at around 1672 cm-1, while the CdO stretching of the Ot-Bu group is at about 1730 cm-1. The effect of incorporating the Aib residue(s) into the compounds examined appears to be significant in increasing the amount of intramolecularly H-bonded species. In fact, both P(Aib)nN and (Aib)nN in dilute CDCl3 solution exhibit an intense absorption in the frequency region around 3330 cm-1, indicating that, in contrast to (Ala)nN, the Aib-based blanks also attain an ordered conformation in solution. This finding is suggestive of a better helix-forming propensity of the Aib than Ala residues, in agreement with the foregoing CD data. The same effect is observed in the amide-I region. Where Ab/Af over n ratios are plotted as a function of (n 2) a straight line is obtained for the series of P(Ala)nN. Ab and Af are the integrated intensities of the bands of H-bonded and free N-H groups at around 3325 and 3430 cm-1, respectively, and n is the total number of amide groups, including both the urethane group in the backbone and the two NH-CO groups

Analysis of Sequential Peptides

J. Phys. Chem., Vol. 100, No. 16, 1996 6839 CHART 2: Schematic Representation of Helical Sections in Cylindrical Coordinates of a Hypothetical Chain of poly(L-Ala) in an r-Helix and poly(Aib) in a 310-Helix. Distances in Angstroms

Figure 7. Plot of (Ab/Af)/n against n for P(Ala)nN (full squares), (Ala)nN (empty squares), P(Aib)nN (full triangles), and (Aib)nN (empty triangles). Ab and Af are the integrated intensities of the bands of H-bonded and free N-H groups at around 3325 and 3430 cm-1, respectively, and n denotes the number of amide groups in the peptides, including those in the side-chains. The intercept at (Ab/Af)/n ) 0 corresponds to the number of peptide groups in the backbone chain that are one unit too short for a helical turn formation by an intramolecular hydrogen bond (see text). Concentrations around 0.05 mM in CD3OD or CDCl3.

in the probe linkages. This is shown in Figure 7, where the intercept at (Ab/Af)/n ) 0, which corresponds to the number of peptide groups in the backbone that are one unit too short for a helical turn formation by an intramolecular hydrogen bond, is (n - 2) ) 3. This finding, which is just that expected for an H-bonding scheme of the i f i + 4 (C13) type, strongly suggests that P(Ala)nN attain an R-helical structure in water/methanol 75/25 solution, to an extent depending on the chain length. In this connection, it must be noted that the urethane moiety (Boc) was counted as a “peptide” group,21 its contribution to the folding process being of paramount importance for P(AA)0N, which otherwise would probably not even exhibit the lowest tendency to attain a helical conformation. By contrast, the blanks have an Ab/Af ratio nearly zero, thus confirming the idea that it is the presence of both N and P molecules that exerts a concerted restructuring effect on the peptides. In the case of P(Aib)nN the (Ab/Af) ratio increases about 56% on going from P(Ala)1N to P(Aib)1N and around 26% for the corresponding octapeptides. The former figure is the one theoretically predictable for an R-helical/310-helical transition in a heptapeptide, in agreement with the proven ability of Aib to form 310-helices in solution,17,22 owing to the steric constraints of the gem-dimethyl groups. The latter figure is half the value for a similar transition in an octapeptide, suggesting that P(Aib)2N populates a mixture of the two conformations because it has the critical length for a 310/R backbone transition.22 Furthermore, all (Aib)nN blanks exhibit ordered, intramolecularly H-bonded conformations in solution. Where Ab/Af over n ratios are plotted as a function of (n 2) a straight line is again observed for P(Aib)1N, (Aib)1N, and (Aib)2N, though the intercept at (Ab/Af)/n ) 0 is this time (n 2) ) 2 (Figure 7). This figure is that predictable for a 310helical structure because its H-bonding scheme is of the i f i + 3 (C10) type. By contrast, the (Ab/Af)/n ratios for P(Aib)2N and (Aib)3N fall inbetween the two straight lines. Since the uncertainty in the integrated IR bands is better than 10%, this finding strongly supports the hypothesis that these latter peptides populate a mixture of the two helices because they have the critical length for a 310/R backbone transition.22 The geometric features of these helices are schematically illustrated in Chart 2. Conformational Statistics and Energetics Analysis. Quantitatively, eq 3 is commonly used for evaluating interprobe distances when both distance distribution and orbital overlap

of the chromophores can be ignored. It gives the dependence of the transfer efficiency on the inverse sixth power of the interprobe distance R, R0 representing the distance at which 50% of the excitation energy is transferred.23

E/(1 - E) ) (R0/R)6

(3)

The transfer efficiency can be measured by fluorescence intensity or fluorescence lifetimes, while R0 is evaluated by eq 4,7,24 where n is the refractive index, ΦN the donor fluorescence quantum yield in the absence of transfer, and J the overlap integral, as given by eq 5, where A(λ) (M-1 cm-1) is the extinction coefficient of the acceptor, FD(λ) is the corrected fluorescence emission spectrum of the donor, and the spectral distribution of donor fluorescence is normalized to unity.

R0 ) (9.79 × 103)[(2/3ΦNJ)/n4]1/6

(4)

∫0∞FD(λ) A(λ)λ4 dλ J) ∫0∞FD(λ) dλ

(5)

The spectral overlap integral in water/methanol 75/25 (v/v) is 3.62 × 10-14 M-1 cm3 and hence R0 is 32.9 Å. The factor 2/3 in eq 4 is the isotropic average value of the orientation parameter κ2.9b,12,25,26 When the donor and acceptor molecules do not rotate fast enough to randomize their orientations during the donor lifetime, κ2 is, instead, given by24

κ2 ) cos2 Θ(3 cos2 γ + 1)

(6)

The geometric parameters for eq 6 are illustrated in Figure 8, where R is the distance between the center of mass of N* and P, γ the angle that the transition dipole moment of N* makes with the line joining the center of mass of N* and P, E the vector representing the electric field at the center of mass of P

6840 J. Phys. Chem., Vol. 100, No. 16, 1996

Figure 8. Angles defining the orientation between the transition dipole moments of the donor and acceptor molecules N* and P and the electric field vector E due to N* acting at the center of mass of the P molecule. R is the line joining the centers of mass of N* and P.

Pispisa et al. CHART 3: Structural Geometry of the Peptides Investigated, Showing the Rotational Angles of the Chromophore Linkages (χ) and Those of the Backbone Chain Fixed in the r-Helix for the Ala-Based and in the 310-Helix for the Aib-Based Peptides. The intramolecular Center-to-Center Distance between N and P for a Given Conformation m Is Denoted as Rm ) |Rm|. Hydrogen Atoms Are Omitted for Clarity

Figure 9. Energy transfer efficiencies 〈E〉s, calculated according to eqs 9 and 10, where κm2 is given by eq 6 for each conformation m of the Ala-based (full squares) and Aib-based (full circles) peptides. The experimental transfer efficiencies from lifetime measurements in water/ methanol 75/25, E1 (eq 2), are also reported (empty symbols).

by the transition dipole moment of N*, and Θ the angle between E and the transition dipole moment of P. Since the magnitude of E is the vectorial sum of its radial and angular components, i.e., it is proportional to [(2 cos γ)2 + sin2 γ]1/2 ) (3 cos2 γ + 1)1/2,24a the component of E along the transition dipole moment of P, E′P, is proportional to cos Θ(3 cos2 γ + 1)1/2. As a result, the value of κ2 that gives the E′P2 dependence on the orientations of both N* and P is given by eq (6). Where an average transfer efficiency, 〈E〉s, comprising the parameter κ2 that takes into account the angular relationship between the chromophores, is evaluated in place of eq 3, the results of Figure 9 are obtained. This because the interprobe distance distribution cannot be ignored and the orientation of the chromophores is not randomized during the donor lifetime. Calculations were based on the rotational isomeric state model for the probe linkages,27 making use of the R-helical (φ ) -57°, ψ ) -47°, ω ) 180°), or 310-helical coordinates (φ ) -60°, ψ ) -30°, ω ) 180°) for the backbone chain of the Ala- or Aib-based peptides, respectively, as schematically illustrated in Chart 3. The good agreement between 〈E〉s and E1, as given by eq 2, indicates that biased chromophore orientations contribute to the measured distances.14b This prompted us to perform a conformational energetics analysis to gather information on the geometric and steric constraints that control the structural features, and hence both the interchromophoric distance and probe orientation of the energetically most favored peptides. Conformational energy calculations were carried out using the set of parameters already reported28 and focusing on the stereochemistry of the side-chains carrying the fluorophores, with the backbone chain conformation fixed in an R-helix for the Ala-based peptides or in a 310-helix for the Aib-based ones. The total potential energy was evaluated in terms of nonbonding, electrostatic, and hydrogen bond interactions, with a refinement process that minimizes the conformational energy in the multidimensional space of all torsional angles. Threefold torsional potential functions, with barriers V0k ) 1.2 and

2.8 kcal/mol, were adopted for rotation around the N-C and C-C bonds. Standard bond angles and bond lengths were used for both the backbone and side-chains.29 Nonbonded interatomic interactions were evaluated by using a 6-12 LJ potential function,28,30 and Coulombic interactions were assessed by assigning partial atomic charges for each atom in the peptides and a distance-dependent dielectric constant.28a In summary, the following equation was used to evaluate the total energy for the mth conformer:

Um,tot ) COUL + NB + HB + TOR

(7)

where COUL, NB, and HB are the sum of all pairwise electrostatic, nonbonded, and H-bond interactions, respectively, and TOR is given by s

TOR ) ∑(V0k/2)(1 + cos 3χk)

(8)

k)1

s being the number of torsional rotations around the dihedral angle χk (Chart 3).

Analysis of Sequential Peptides

J. Phys. Chem., Vol. 100, No. 16, 1996 6841

We were thus able to evaluate the average transfer efficiency 〈Et〉, according to eqs 9 and 10:

〈Et〉 ) ∑Et,mPm

(9)

m

Et,m )

1 2 Rm 1+ 3κm2 R0

[ ( )] 6

(10)

where Et,m is the transfer efficiency in conformer m, and the sum is taken over all conformations of both donor and acceptor chromophore linkages. Rm and κm2 are the interprobe distance and orientation parameter, respectively, in the same conformer, and Pm the corresponding probability, as given by eq 11, where k is the Boltzmann constant, and U′m is the difference between the total energy of the conformer, Um,tot, and the lowest energy among all conformers of the given peptide, Umin, as given by eq (12).

Pm )

exp(-U′m/kT) exp(-U′m/kT) ∑ m

(11)

Therefore, the uncertainty in the absolute value of the total energy, Um,tot, arising from the empirical terms of eq 7 is overcome by the fact that eq 11 is based on the energy difference of conformers, Umin playing the role of a reference energy.

U′m ) Um,tot - Umin

(12)

Briefly, we searched for all relative minima in the multidimensional conformational space of the peptides in two steps: first, by minimizing the energy in terms of eq 7, and then by allowing the minimum energy structure to undergo small changes in the internal rotation angles to match the experimental transfer efficiency, provided the energy did not vary more than 1%. This because even the deepest energy minimum structures do not necessarily exhibit the best probe orientation for an efficient energy transfer process, as experimentally determined. The agreement between experimental and calculated transfer efficiencies could be obtained at very low expense of energy for two reasons. The first is a topological reason in the sense that the variations of the internal rotation angles of the sidechains carrying the chromophores are distributed over a number of bonds, so that, on the whole, the structure behaves as though the overall perturbation was very small. The second is a geometric reason, i.e., if the angle γ between the transition moment vectors is close to 90° (eq 6 and Figure 8), very small perturbations in the geometry of the chromophores produce a remarkable change in the κm2 factor and thus in the transfer efficiency, eq 10. Since the probes in the low-energy conformations are usually close to each other, with a parallel or almost parallel arrangement, their mutual orientation could be easily adjusted to match the experimental energy transfer efficiency at a definitely small energy cost.31 Collisions with solvent molecules would supply the required energy in solution. By contrast, as the conformer energy increases with respect to the deepest energy minimum, the chromophores tend to lie far apart. Optimization between calculated and experimental transfer efficiencies now requires a relatively large variation in the internal rotation angles, which would be accompanied by a significant change in the conformer energy. All these features are summarized in Table 4, where the molecular parameters of the most populated conformers for each

TABLE 4: Molecular Parameters of the Most Stable Conformers of the Ordered Linear Peptidesa peptide

Um,totb

Rmc

Et,md

Pme

P(Ala)0N

-3.84 -2.86 -1.18

4.35 5.34 10.85

0.780 0.784 0.997

83.0 15.9 0.9

P(Ala)1N

-5.33 -4.56 -3.87

8.44 10.15 11.32

0.851 0.850 0.850

71.0 19.5 6.1

P(Ala)2N

-11.02 -10.28 -9.92

7.43 7.53 7.12

0.863 0.857 0.864

69.2 19.9 10.8

P(Ala)3N

-8.18 -6.76

12.56 11.45

0.770 0.767

90.8 8.2

P(Ala)4N

-7.05 -4.62 -4.28

6.16 10.06 5.84

0.753 0.759 0.999

97.5 1.6 0.9

P(Aib)1N

-5.85 -3.64

6.35 6.17

0.921 0.999

95.6 2.3

P(Aib)2N

-12.03 -9.43 -9.26 -8.74

10.00 9.58 10.64 9.99

0.927 0.943 0.919 0.923

97.3 1.1 0.8 0.3

〈Et〉f

E1g

0.783

0.78

0.855

0.85

0.862

0.86

0.772

0.77

0.755

0.76

0.923

0.92

0.927

0.93

a

Calculations were performed with the backbone chain of P(Ala)nN in an R-helix and that of P(Aib)nN in a 310-helix. b From eq 7, in kcal/ mol. c P-N center-to-center distance, in Å. d From eq 10, with the orientation factor κm2 as given by eq 6. e From eq 11. f From eq 9.g From eq 2.

peptide are listed, and in Figures 10 and 11, where the molecular model of the most stable structures of some peptides are illustrated. The results indicate that the deepest energy minimum structure for each peptide is populated more than 70% and attains the experimental transfer efficiency by varying the internal rotation angles by less than 1° at very low expenses of energy. By inspection of Table 4, it also appears that among the energetically most favored structures of P(AA)nN, only the hexapeptide exhibits a center-to-center distance (Rm) shorter than 5 Å, which may be considered as a separation distance at which a Dexter-type energy transfer mechanism could apply8. In this case, the E/(1 - E) ratio is exponentially dependent on the interchromophoric distance,25,32-34 i.e.:

E/(1 - E) ) (2π/kDp)J′V2

(13)

where kD is the emission rate constant of the donor and J′ the overlap integral, as obtained by spectroscopic measurements, given by

J′ ) ∫0 FD(λ) A(λ) dλ ∞

(14)

with the normalization conditions

∫0∞FD(λ) dλ ) ∫0∞A(λ) dλ ) 1 and

V2 ) K exp(-2R/a)

(15)

where V is the electronic matrix coupling element, a the average

6842 J. Phys. Chem., Vol. 100, No. 16, 1996

Pispisa et al.

Figure 10. Molecular models of the most stable structures of P(AA)0N (a), P(Ala)1N (b), and P(Aib)1N (c), with the backbone chain in an R-helix (a and b) and in a 310-helix (c), viewed along the helical axis. The structures on the left side are the deepest energy minimum conformations (eq 7), while those on the right side are the conformations after optimization between calculated and experimental energy transfer efficiency.

Bohr radius, and K a quantity with dimensions of energy, corresponding to the electronic matrix coupling at orbital contact. Very recently Scholes and Ghiggino34 have shown, however, that for a naphthalene dimer at an intermolecular distance as short as 4.25 Å the dipole-dipole term in the multipole expansion for the electronic coupling between dipole-allowed

electronic transitions accounts for more than 90% of the interaction energy. This because within the separation distance of 4-5 Å, the higher order terms of the multipole expansion (dipole-octopole and octopole-octopole) counterbalance the short range component of the electronic coupling, thus leaving the limiting dipole-dipole term as the major contributor to the interaction energy.34 Although these calculations refer to a

Analysis of Sequential Peptides

J. Phys. Chem., Vol. 100, No. 16, 1996 6843 of the two conformations because it has the critical length for a 310/R backbone transition.22 Concluding Remarks

Figure 11. Molecular model of P(Ala)2N in the deepest energy minimum, viewed perpendicularly to the helical axis. Note the almost parallel orientation of the backbone chain with P and N groups, leading to a very stable structure.

couple of naphthalene molecules, they are equally applicable to the case of interacting nonidentical chromophores,34 such as the N-P pair. As a result, a Dexter-type singlet transfer mechanism would only be a minor contributor to the overall energy transfer process. This also agrees with the results reported by Speiser et al.,33 showing that only when the R0 value, eq 4, is small and comparable to the interchromophoric distance is the dipole-dipole interaction weak as compared to the exchange interaction. In other words, unless R0 < 10 Å, energy will be transferred mainly by the Fo¨rster mechanism.33b Accordingly, calculations suggest to us that a Dexter-type transfer pathway is unlikely even in the case of the hexapeptide.35 Figure 10 illustrates a typical example of the most stable structures before and after energy transfer optimization. Each couple of conformations looks very similar, the bulky chromophore molecules always exhibiting a face-to-face orientation. Nonbonded and π interactions stabilize this geometry, which is expected to be further stabilized in the water/methanol mixture by hydrophobic interactions between P and N groups. Interestingly enough, this may be the reason why P(Ala)2N exhibits the highest transfer efficiency despite the fact that, on the basis of the R-helix periodicity, the CR atoms of the probe linkages are farther away than in the other peptides, being five residues apart. Owing to both the length and flexibility of the aliphatic portion of the probe linkages, the deepest energy minimum conformation, which heavily contributes to the observed transfer process, sees the two chromophores lying close each other, as shown in Figure 11. Taken together, all these features emphasize the care necessary in interpreting excited-state phenomena when conformational statistics data alone are available. Finally, where calculations for P(AA)0N, P(Aib)1N, and P(Aib)2N were carried out by fixing the backbone chain either in an R-helix or in a 310-helix, owing to the uncertainty in the structural features of these peptides, a good agreement between calculated and experimental transfer efficiency could be obtained only when the backbone of P(AA)0N was in the R-helix and that of P(Aib)1N in the 310-helix. This agrees with IR and chiroptical results. By contrast, in the case of P(Aib)2N both R- and 310-helical backbone conformations give good results, which emphasizes the sensitivity of the empirical energy functions used. In fact, the foregoing IR spectral data suggest that the Aib-based octapeptide very likely populates a mixture

Three major conclusions are drawn from the present study. First, the quenching efficiency from fluorescence intensity measurements in water/methanol 75/25 (v/v) follows a different trend as the number of AA units in the spacer of the backbone chain varies, depending on whether AA ) Ala or Aib. This is a clear indication of different structural features in solution of the two series of P(Ala)nN and P(Aib)nN peptides, a hypothesis fully supported by CD and IR spectral results. Second, conformational energy calculations show that of the 1 118 124 conformers of the probe linkages (1836 for the P linkage and 609 for the N linkage), as evaluated by conformational statistics by using the rotational isomeric state model, and discarding both the g+g- and g-g+ conformations for their very low probability and those in which the chromophore is less than 5 Å from the helical axis, for steric reasons, only a few, low-energy conformations for each peptide are the major contributors to the energy transfer process. They are characterized by a parallel, often face-to-face, orientation of the probes. These findings emphasize both the care that must be taken when using conformational statistics data for interpreting excited-state phenomena and the usefulness of theoretical conformational calculations in providing a valuable insight into the relationship between molecular structure and energy transfer. Third, orientation effects between the N and P groups had to be taken into account for a correct interpretation of fluorescence decay data, implying that the relaxation time of the interconversion among conformational substates of the chromophores linkages is definitely longer than 10 ns. Acknowledgment. This work was supported in part by the National Research Council (C.N.R.) and in part by a Commission of the European Communities under contract No. Ge1*CT920046. We thank Professor R. Righini (Universita´ di Firenze) for helpful discussions. Supporting Information Available: Two-page text outlining the conformational statistics analysis, and two other pages with two tables. The first contains the rotational isomeric states, relative energies, and statistical weights of the donor and acceptor linkages and the second one the values of the 17 dihedral angles of both probe linkages in the deepest energy minimum structure for each peptide, together with the same angles after optimization to transfer efficiency (4 pages). Ordering information is given on any current masthead page. References and Notes (1) Gust, D.; Moore, T. A.; Moore, A. L.; Devadoss, C.; Liddell, P. A.; Hermant, R.; Nieman, R. A.; Demanche, L. J.; De Graziano, J. M.; Gouni, I. J. Am. Chem. Soc. 1992, 114, 3590. Gust, D.; Moore, T. A.; Moore, A. L. Acc. Chem. Res. 1993, 26, 198. Gust, D.; Moore, T. A.; Moore, A. L.; Mcpherson, A. N.; Lopez, A.; De Graziano, J. M.; Gouni, I.; Bittermann, E.; Seely, G. R.; Gao, F.; Nieman, R. A.; Ma, X. C.; Demanche, L. J.; Hung, S.-C.; Luttrull, D. K.; Lee, S.-J.; Kerrigan, P. K. J. Am. Chem. Soc. 1993, 115, 11141. (2) Prathapan, S.; Johnson, T. E.; Lindsey, J. S. J. Am. Chem. Soc. 1993, 115, 7519. Wagner, R. W.; Lindsey, J. S. J. Am. Chem. Soc. 1994, 116, 9759. (3) Osuka, A.; Yamada, H.; Maruyama, K.; Mataga, N.; Asahi, T.; Ohkouchi, M.; Okada, T.; Yamazaki, I.; Nishgimura, Y. J. Am. Chem. Soc. 1993, 115, 439. (4) Balzani, V.; Scandola, F. Supramolecular Photochemistry; Ellis Horwood: New York, 1991; p 110. (5) Wasielewski, M. R. Chem. ReV. 1992, 92, 435. Johnson, D. G.; Niemczyk, M. P.; Minsek, D. W.; Wiederrecht, G. P.; Svec, W. A.; Gains, G. L., III; Wasielewski, M. R. J. Am. Chem. Soc. 1993, 115, 5692.

6844 J. Phys. Chem., Vol. 100, No. 16, 1996 (6) Photochemical ConVersion and Storage of Solar Energy; Connolly, J. S., Ed.; Academic Press: New York, 1981. (7) Fo¨rster, T. Ann. Phys. (Leipzig) 1948, 2, 55; Discuss. Faraday Sec. 1959, 27, 7. (8) Dexter, D. L. J. Chem. Phys. 1953, 21, 836. (9) (a) Sisido, M.; Egusa, S.; Imanishi, Y. J. Am. Chem. Soc. 1983, 105, 1041. (b) Inai, Y.; Sisido, M.; Imanishi, Y. J. Phys. Chem. 1990, 94, 6237. (10) McLendon, G. Acc. Chem. Res. 1988, 21, 160 and references therein. (11) Haran, G.; Haas, E.; Szpikowska, B. K.; Mas, M. T. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 11764. (12) McWherter, C. A.; Haas, E.; Leed, A. R.; Scheraga, H. A. Biochemistry 1986, 25, 1951. (13) (a) Pispisa, B.; Venanzi, M.; D’Alagni, M. Biopolymers 1994, 34, 435. (b) Pispisa, B.; Venanzi, M.; Palleschi, A. Faraday Trans. 1994, 90, 1857. (14) (a) Pispisa, B.; Venanzi, M.; Palleschi, A.; Zanotti, G. J. Mol. Liq. 1994, 61, 167. (b) Pispisa, B.; Venanzi, M.; Palleschi, A.; Zanotti, G. Macromolecules 1994, 27, 7800. (c) Pispisa, B.; Venanzi, M.; Palleschi, A.; Zanotti, G. Biopolymers 1995, 36, 497. (15) Eaton, D. F. Pure Appl. Chem. 1988, 60, 1107. (16) Padmanabhan, S.; Marqusee, S.; Ridgeway, T.; Laue, T. M.; Baldwin, R. L. Nature 1990, 344, 268. (17) Benedetti, E.; Di Blasio, B.; Pavone, V.; Pedone, C.; Santini, A.; Crisma, M.; Toniolo, C. Molecular Conformation and Biological Interactions; Balaram, P., Rameseshan, S., Eds.; Indian Academy of Sciences: Bangalore, 1991; p 497 and references therein. (18) Mizushima, S.; Shimanouchi, T.; Tsuboi, M.; Souda, R. J. Am. Chem. Soc. 1952, 74, 270. (19) Aubry, A.; Protas, J.; Boussard, G.; Marraud, M.; Nee´l, J. Biopolymers 1978, 17, 1693. (20) Pulla Rao, Ch.; Nagaraj, R.; Rao, C. N. R.; Balaram, P. Biochemistry 1980, 19, 425. (21) Benedetti, E.; Pedone, C.; Toniolo, C.; Nemethy, G.; Pottle, M. S.; Scheraga, H. A. Int. J. Pept. Protein Res. 1980, 16, 156. (22) Basu, G.; Bagchi, K.; Kuki, A. Biopolymers 1991, 31, 1763. (23) Cheung, H. C. Topics in Fluorescence Spectroscopy; Lakowicz, J. R., Ed.; Plenum Press: New York, 1991; Vol. 2, Chapter 3. (24) (a) Steinberg, I. Z. J. Chem. Phys. 1968, 48, 2411. (b) Grinvald, A.; Haas, E.; Steinberg, I. Z. Proc. Natl. Acad. Sci. U.S.A. 1972, 69, 2273.

Pispisa et al. (25) Valeur, B.; Mugnier, J.; Pouget, J.; Bourson, J.; Santi, F. J. Phys. Chem. 1989, 93, 6073. (26) Bedna´r, B.; Morawetz, H.; Shafer, J. Macromolecules 1985, 18, 1940. (27) Flory, P. J. Statistical Mechanics of Chain Molecules; Interscience: New York, 1969. Mattice, W. L.; Suter, U. W. Conformational Theory of Large Molecules; Wiley: New York, 1994. (28) (a) Pispisa, B.; Palleschi, A. Macromolecules 1986, 19, 904. Pispisa, B.; Palleschi, A.; Paradossi, G. J. Phys. Chem. 1987, 91, 1546. Pispisa, B.; Paradossi, G.; Palleschi, A.; Desideri, A. J. Phys. Chem. 1988, 92, 3422. (b) Chiessi, E.; Palleschi, A.; Paradossi, G.; Venanzi, M.; Pispisa, B. J. Chem. Res. (S) 1991, 248. Chiessi, E.; Branca, M.; Palleschi, A.; Pispisa, B. Inorg. Chem. 1995, 34, 2600. (29) Wunderlich, H.; Mootz, D. Acta Crystallogr., Sect. B 1971, 27, 1684. Andersen, A. M. Acta Chem. Scand., Ser. B 1975, 29, 239. (30) Nemethy, G.; Pottle, M. S.; Scheraga, H. A. J. Phys. Chem. 1983, 87, 1883. (31) Anticipating that the deepest energy minimum structure for each peptide is characterized by a face-to-face arrangement of the probes (Figures 10 and 11), one is tempted to say that an orientation of the transition moment vectors close to 90° brings about the energy transfer process, while a paired orientation of the transition moments favors exciplex formation. (32) Haggquist, G. W.; Katayama, H.; Tsuchida, A.; Ito, S.; Yamamoto, M. J. Phys. Chem. 1993, 97, 9270. (33) (a) Speiser, S.; Katriel, J. Chem. Phys. Lett. 1983, 102, 88. (b) Levym, S.-T.; Rubin, M. B.; Speiser, S. J. Am. Chem. Soc. 1992, 114, 10747. Speiser, J. L.; Wang, B.; Harriman, A. J. Am. Chem. Soc. 1995, 117, 704. (34) Scholes, G. D.; Ghiggino, K. P. J. Phys. Chem. 1994, 98, 4580. (35) Referring to P(AA)0N, which exhibits the shortest interprobe centerto-center separation distance (〈R〉 ) 4.6 Å), we used eq 7 to evaluate the electronic matrix coupling at orbital contact, K. The energy transfer efficiency was E ) 0.78, from lifetime measurements, the overlap integral for the Dexter-type singlet mechanism (J′) was calculated to be 3.61 × 10-3 from spectral results, kD ) τD-1 was 2.76 × 107 s-1, and the average Bohr radius employed was a ) 0.625 Å.34 Then, K ) 1.12 × 10-11 erg, which differs by some 8 orders of magnitude from the value reported for electronic energy transfer via exchange interaction in bichromophoric molecules,33a

JP953258Y