J. Phys. Chem. 1995, 99, 11255- 11265
11255
Elementary Processes in Peptides: Electron Mobility and Dissociation in Peptide Cations in the Gas Phase R. Weinkad,” P. Schanen, D. Yang,+S. Soukara, and E. W. Schlag Institut f i r Physikalische und Theoretische Chemie, Technische Universitat Miinchen, 85747 Garching, Lichtenbergstrasse 4, Germany Received: January 24, 1995; In Final Form: April 28, 1995@
Neutral peptides of natural amino acids of the type (X),-Y (n = 1,2,3) are prepared in the gas phase by laser desorption and supersonic cooling. Local ionization is performed by resonant laser excitation in aromatic amino acids (Y) located at the C-terminal end. In a one-color experiment, subsequent UV photofragmentation of the cation is shown to directly reflect the prior charge migration in these large molecules. Peptides are engineered, which show either fragment ions originating from the chromophore or from the opposite N-terminal side (X). The results show that by changing local ionization energies and thus absolute positions of ionic dissociation energies, one has complete control over different paths of chemical reactivity. The length dependence of the process shows, that charge mobility seems to be not the bottleneck for dissociation pathways at high internal energies: charge transfer over more than 10 o-bonds is shown. When we apply a local picture and estimate local ionization potentials, we find, for the peptides used here, that after localized ionization, positive charge is statically localized at the initial prepared site. In a two-color experiment (UV VIS) we observe indications that in the photoexcited tripeptide cation Leu-Leu-Tyr charge transfer can occur at internal energies of about 2.2 eV, an energy at which no dissociation occurs. We interpret the process in terms of direct photoexcitation into a charge transfer (CT) band or by a photoexcitation to a localized state followed by nonradiative relaxation to a CT state. For the charge-transfer process we propose a through bond HOMO electron transfer (hole transfer) as the relevant mechanim. Consequences of our findings for charge migration and fragmentation processes in peptides are discussed.
+
I. Introduction Electron transfer (ET) in proteins is a subject of great topical interest. In this work we produce molecular beams of native peptides, perform a local ionization, and show how the ensuring charge transfer in peptides is directly linked to their reactivity. We investigate charge mobility in the gas phase in native molecular systems namely bare peptide cations composed of natural amino acids, a process relevant to biology and medicine as well as theory. We employ natural peptide chains without attached special donors or acceptors which could have undesirable cooperative side effects on the measurements. In fact, changing local ionizing sites totally changes the reactivity of these peptides. In a two-color experiment at a tripeptide, we find hints for charge transfer occumng independent from dissociation. Polypeptides or large molecules in the gas phase in general open a new world to experimentalists. They consist of subunits such as amino acids, nucleotides, side chains, and aromatic chromophores, which are linked by a-bonds. This heterogeneous composition stimulates one to look at such systems in a picture, where the properties of the supermolecule are described by the interaction of local properties. In addition to undertaking a study of these large systems in the absence of environmental influence, such systems possess a large number of vibrational degrees of freedom, a high density of electronic states and many dissociation channels. Therefore findings and theory observed and developed for small molecules are not necessarily applicable to these systems. This may concem ionization,’ dissociation To whom correspondence should be addressed. Insitute of Chemistry, Academia Sinica. Beijing, China. Abstract published in Advance ACS Absfructs, June 15, 1995. $
’Permanent address: @
0022-3654/95/2099- 11255$09.00/0
dynamics,*charge transfer, and other intramolecular processes. Charge-transfer processes especially play an important role in photosynthesis and in donor-acceptor redox reaction^.^,^ Natural electron-transfer (ET) complexes such as large molecular systems can have a size of several thousands of daltons. For such descriptions of large supermolecules a localized picture will be used: Molecular subunits are acting as electron donor (D) or electron acceptor (A) and are linked by a initially nonconducting molecular spacer (Sp) or bridge (B). The bridge, which is promoting ET, is a flexible or a rigid chain, which in biological systems often consists of proteins or polypeptides. Electron transfer is started by local photoexcitation of the donor or acceptor, thus providing a driving energy for charge separation. In natural systems or special designed supermoleculess ET can take place over distances of more than 10 8, and is then called long-range ET. Many authors have contributed to the understanding of CT mechanisms. “Through bond”, “through space”, and HOMO and LUM03 electron-uansfer mechanisms have been investigated theoretically and experimentally. The most prevalent mechanism is the through bond mechanism, which has been described theoretically by Marcus6 and others. In. solution experiments many different parameters have been varied to test the validity of theories. Well defined rigid bridges of different length have been used to observe the length dependence of CT ratess The solvent dependence has been measured in order to investigate influences of the environment on CT p r o c e s s e ~ . ~ ~ ~ The effect of the solvent on the ET can be strong and complex: For example, the solvent can serve as heat bath, as vibrational background continuum, and as a reorganizing or shielding medium for separated charges. Investigations in the gas phase are useful in attempting to understand environmental influences but so far have beefl limited 0 1995 American Chemical Society
11256 J. Phys. Chem., Vol. 99, No. 28, 1995 to small thermal evaporizable molecular systems. Theoretical considerations by Bixon et aL9 and Jortner et al.'O suggest that in neutral isolated molecules the ET should have a limited range for charge separation. In such cases this is due to energetic reasons since in the gas phase, the separated charges are not stabilized. Nevertheless, in the gas phase, D-B-A systems show charge separation, but in most of these cases the molecule collapses under strong Coulomb attraction between opposite charges resulting in an intramolecular exciplex formation.' ' - I 3 Thus, in most cases charges are not separated over long distances. Even direct excitation into an emissive charge transfer state was observed in a flexible linked D-A compound.I4 However, so far there is only one case where evidence for a long-range charge transfer in a neutral molecule in the gas phase has been observed.15-" In weakly bound clusters in the gas phase an intermolecular exciplex formation was observed by Piuzzi et a1.,'8-20a process that could be seen as an example of "through space" ET transfer. The case of charge transfer in cations and anions is different from the charge separation for neutrals: there is one net charge, and for its transfer no Coulomb force between positive and negative charges has to be overcome. As a consequence, in cations there should be no intrinsic restriction for distances of charge migration.l o In cations differences in local ionization energies can be a driving force for charge migration. Therefore, in contrast to neutral molecules, ET in ions can occur even in the ground state as Levy et al. showed.?' They combined site selective resonant MPI (laser 1) with time-delayed resonant photodissociation for probing the charge site (probe laser 2). For bichromophoric molecules of the type D-B-A they found charge transfer to occur directly after ionization and to be faster than their experimental time resolution of 2 ns. They proposed a hole (HOMO electron) transfer in the cation ground state as the relevant mechanism:2' In a local picture of the molecules the positive charge is optimizing its energy by intramolecular hopping to the chromophore which has the lowest ionization potential. Similar observations have been made for clusters (see below). In our work we use native peptides as samples consisting of one or several nonaromatic and only one aromatic amino acid. Measured and estimated ionization potential show that in our experiments electron transfer does not occur directly after ionization but needs further photon absorption in the cation. As for electronic states in cations, couplings between electronic states (localized and nonlocalized) can be strong. As observed for the special case of homo dimer clusters, coupling of electronic states can occur in causing large splittings2* and charge resonance bands.23 Similar couplings can be expected for molecules such as peptides consisting of same repeptitive constituents, here amino acids. As for hetero dimers, optical excitation to a CT band has been found in the NzOf-Ar clu~ter.'~Here the CT band is close in energy to the ionization energy of the attached Ar atom leading to Ar' as dissociation product. CT bands due to neighbored chromophores can be as well expected for large molecular cations consisisting of mostly isolated parts. For the amino acid glycine, for example, the first two photoelectron bands can be correlated directly to an ionization of lone-pair electrons of the nitrogen (onset at 8.8 eV (IP)25)and the oxygen in the carbonyl groupe (onset at 9.9 eV25). They can be, in first order, seen as local bands where charge is localized. In this way in glycine the second band can be seen as a charge transfer band. In contrast to neutrals such CT bands do not require any substantial shift in geometry due to the missing or small Coulomb interaction of the single
Weinkauf et al. charge. The situation can be more complicated for charge transfer in ions than for neutrals due to the higher density of electronic state^^^,?^ for ions. Large biologically relevant molecules are thermal unstable and thus for examinations in the gas phase have to be evaporated by special laser desorption techniques.?' The two-step laser technique consisting of laser desorption of neutral, intact molecules into a supersonic beam expansion and of subsequent resonant multiphoton ionization and dissociation by W-laser pulses has been proven to be a suitable tool for mass spectrometrical analysis of large nonvolatile neutral molecule^,^^.'^ polypeptide^,?^ and nucleotide^^^ in the gas phase up to a mass 2000 Da without decomposition. In mass spectrometry initial charge localization is rarely observed since most ionization techniques are unspecific as for example in the case of electron impact ionization. Localized charge formation, however, is the essential precondition for observing charge flow in large ionic molecules. This can be done by intracluster or intramolecular localization of ionization via resonant multiphoton ionization of suitable local chromophores, a concept which is rarely used. Dissociative charge transfer, for example, was observed to occur in some hetero dimer clusters directly after localized i ~ n i z a t i o n . ~ 'Localized -~~ ionization in bichromophoric molecules by resonant photoionization, has been applied by Levy and co-workers as mentioned above.?' As for peptides, localized photoabsorption in aromatic side chains of peptides can be deduced from the W spectra of large molecules in liquids: Absorption spectra of large peptides agrees well with absorption of the isolated amino acids. Furthermore localized photoabsorption has been observed in MPI-spectra of some di-34.35and t r i p e p t i d e ~in~ ~the gas phase. This finding corresponds well with the fact that localized o-bonds typically absorb around 210 nm and C=O double bonds have only low absorption cross sections at wavelength around 290 nm (extinction coefficient E = 1.2-1.4 m2/mol). In contrast to this, samples with aromatic chromophores have high absorption cross sections ( E > 30 m2/mol) between 250 and 285 nm. Direct ionization is a very fast process ( s) and can be performed only if the energy of two UV photons (wavelength 250-300 nm) is concentrated on one electron. From resonant and thus localized exciation by the first photon and the long lifetimes of the intermediate states, it follows that ionization is localized at the chromophore. In agreement with this fact site selective ionization and even dissociation in large molecules has been observed by us at large polypeptides in the gas phase by 500 fs laser pulse e ~ c i t a t i o n . ~ ~ Other new developments in mass spectrometry like matrixassisted laser desorptiodionization (MALD138.39),secondary ion mass spectrometry (SIMS'"), r52Cfplasma desorption$' and fast atom-bombardment (FAB42-u) permit investigation of large ions in the gas phase up to a mass of 500 000 Da.38 Ionization here is realized but by adduct formation with H+, Na+, K+, and Ca+ ions, and the position of the adduct is unknown. For electrospray ionization (EI) situation is more complicated by the multiple charged character of the ions ~ b s e r v e d . ~In~contrast .~~ to this by resonant multiphoton ionization ions are formed as singly charged radical cations. Typically ET in neutrals is detected by a large red shift of fluorescence resulting from emission of charge separated states. This ET detection technique, however, cannot be generalized for ions: This is obvious for the ion ground state and easy to understand for electronic excited ionic states because of fast radiationless relaxation processes (usually, in ions electronic
J. Phys. Chem., Vol. 99, No. 28, 1995 11257
Elementary Processes in Peptides 01 Leu - Y
bl Gly-Y
0
I O I 11 NH,-CH-C-NH-CH-C-OH
I '
I1
I Q
CH,I--
I
I
I?-
.I
-
86 Oolton vh. . '
/CH\
CH,
CH,
Tryptophan
Tyraslne
8
Phenylalacline
R=
I CH,
I
H
IPAop,
7 5 eV :286
1 nm
IPAop,
8 0 eV I
281 7 rm
IP
-
Aop,=
8 L eV 266 L nm
Figure 1. Structures of peptides of type Leucine-Y (a) and Glycine-Y (b). Photoexcitation is realized at the residues R. (a) In type Leu-Y dipeptides mass 86 Da can be explained only by electron migration from the N-terminal side to the chromophore residue (R) of Y. (b) In type Gly-Y dipeptides mass 30 Da results from charge migration. The Roepstorff-Fohlman nomenclatures7 for peptide fragment ions is shown. (c) Residues of the aromatic amino acids, where local ionization and charge localization is performed.
states are closer in energy than in neutrals) or fragmentation at this energy level. Therefore, in this work it is proposed to detect charge migration either by direct observation of characteristic positive fragment ions (one-color experiment) or by resonant multiphoton dissociation of the cation (two-color experiment). As is well known, fragment ion peak patterns in mass spectra of molecular ions directly yield structural information on the molecules in question. A long list of rules exist, which help to interpret fragmentation of relatively small molecular ions. The empirical rule of Stevenson,47 for example, states that in a fragmentation process of a cation the positive charge prefers to reside on the fragment with the lowest ionization potential; This rule is a statement of experience from unselective ionization techniques and relatively small molecules. The extension of this rule to large and very large molecules as well as to cases of localized ionization is not obvious. It should be noted that Stevenson's rule refers to the charge distribution in the resultant fragments. In our experiments here, we have localized ionization and hence are concerned with charge migration in the intact molecule prior to fragmentation, a different though related issue. We wish to demonstrate local ionization, photodissociation, and mass spectrometry to be a new method for monitoring electron migration alon the peptide chain over more than 10 o-bonds (RDA > 10 We here interpret reactivity, of dipeptides and other polypeptides of the type (X),-Y (all L form) containing only one aromatic amino acid at the C-terminal peptide side in terms of electron mobility. We always chose samples where nonaromatic amino acids are used as bridge or N-terminal end and where aromatic amino acids are situated at the C-terminal end of the peptide. Figure 1 shows, as an example, the molecular structures of dipeptides of type Leu-Y (a) and Gly-Y (b). The free amino group is the N-terminus and the free carbonyl group is the C-terminus. Siteselective positive charge location in the cation is performed by resonant two-photon ionization in the aromatic side chains. These chromophoric residues (R) of tryptophan, tyrosine, and phenylalanine its ionization potentials2s.26and resonant wavel e n g t h ~ " . ~are ~ presented in Figure IC. MPI spectra are avail-
1).
able for the peptides Gly-Trp and G l ~ - G l y - T r p . As ~ ~ for the peptides with unknown MPI spectra, laser wavelengths were tuned to the neutral resonances of the chromophores shown in Figure IC. Neither the amino acids glycine (Gly) nor leucine (Leu) absorb at 250-280 nm and have been used as N-terminal donor or bridge. By energetical considerations we can show that electron transfer in the cationic ground state of peptides is not the relevant mechanism in contrast to previous CT experiments with molecules2' or c l ~ s t e r s ~ (see l - ~ ~below). For the observation of the existence of charge migration in peptide cations we detect characteristic photofragment ions which comes from the N-terminal end, opposite to the site of initial charge location. For photofragmentation the same UV laser is used as for ionization (one-color experiment). This technique is well suited for detection of the existence of CT in peptides and large molecules and shows how strong reactivity and charge transfer are correlated. To find out if charge transfer can occur independently of dissociation we performed a two-color experiment. Ionization is performed by a UV laser and ion dissociation by a visible (VIS) laser. For observation of charge flow we apply a charge and site sensitive multiphoton excitation by VIS light. This excitation scheme is similar to the resonant multiphoton photodissociation spectroscopy, a technique which was shown to be well suited for spectroscopy of short lived excited states in ion^.^^,^' We consider a model describing the peptides by local ionization and thus local absolute dissociation energies. Consequences of our findings for charge migration and fragmenation of peptides are discussed. MPI mass spectra of small peptides have been published p r e v i ~ u s l y > ~but - ~here ~ we wish to suggest that resonant MPI fragment mass spectra of peptides are a signature of charge mobility.
11. Experimental Setup The experimental setup is similar to that of our former experiment^.*^-^^,^*.^^ In the first vacuum chamber the nonvolatile dipeptides and polypeptides were laser desorbed as neutrals by CO? laser pulses (40 &/pulse), introduced into a supersonic beam produced by expansion of Ar with 3 bar of back-pressure through a pulsed nozzle (diameter 400 pm). Introduction of the laser desorbed molecules into the Ar beam enhances cooling and enables sample molecules to be transported efficiently into the ionization chamber. Due to the relatively large distance between desorption area and the nozzle (about 1.5 mm) the cooling effect is expected to be much less efficient for laser-desorbed molecules than for volatile molecules expanded as a premixture. To fulfill the condition of collisionless flight of the ions in the mass spectrometer,the merged beam was skimmed and introduced into the differentially pumped ion source of a reflectron time-of-flight instrument (lo-' mbar). One-Color Experiment. For resonant laser ionization (two W photons) and laser dissociation (third W photon) we used an excimer laser pumped dye laser (Lambda Physik, FL 3001, 7-10 ns pulse width) which was frequency doubled to a W output energy of 400-600 pJ. The laser was focused in the ion source by using a cylindrical lens of 110 mm focal length. To increase or decrease laser intensity, the size of the laser spot was changed by setting the lens at well-defined positions out of focus. Ion counts have been corrected for different sample volumes adressed by the laser beam. Ions were mass analyzed by a reflectron time-of-flight mass spectrometer using multichannel plates as detector and a transient recorder for digitizing the signal.
Weinkauf et al.
11258 J. Phys. Chem., Vol. 99, No. 28, 1995
b)
Oissociation
C T C l Ion
I
uv -u
Neutral
X Y Y X Figure 2. Photoexcitation schemes for local resonant two-photon ionization and cation photoexcitation (X = Gly,Leu; Y = Trp,Tyr, Phe). (a) One-color excitation scheme: dissociation of peptide cations is performed by one-photon excitation (same laser as for ionization).Characteristic fragment ions show prior charge migration. (b) Two-color excitation scheme: in peptide cations resonant two-photon excitation by visible light is charge sensitive at the aromatic chromophore (Y).If charge transfer occurs after the first photon absorption, neither absorption of a second photon nor dissociation is possible. TABLE 1: Ionization Energies of Free Amino Acids IP first band ref“ 8.8 eV N, lone pair 25 glycine leucine 8.5 eVb N, lone pair 25 tryptophan 1.5 eV aromatic ring 26 tyrosine 8.0 eV aromatic ring 25 8.4 eV aromatic ring 25 phenylalanine Estimated onset of He photoelectron spectra. Reduction due to additional CH, groups taken into account. Two-Color Experiment. As in the one-color experiment the local ionization is done by resonant two-photon ionization with the frequency-doubled UV output of an excimer laser pumped dye laser (pulse energy 100 pJ, pulse width 5 ns, wavelength range 260-280 nm). By carefully adjusting the W-laser intensity it is possible to achieve ionization without any fragmentation called “soft ionization”. This indicates that neutral fragmentation and subsequent ionization of fragments does not contribute to our mass spectra and that no further W photon is absorbed in the cation. Photodissociation is then exclusively performed by a second laser (no time delay to laser 1; pulse energy 1.5 d, pulse width 7 ns, wavelength 520-560 nm).
111. Results 111.1. One-Color Experiments. The process of electron transfer in peptide cations can be uniquely detected by the observation of characteristic photofragment ions, the occurrence of which can be explained only by charge migration. The Roepstorff-Fohlman nomenclatures7of fragmentation of peptides is included in Figure lb. The A fragment ions, which would indicate electron flow from the N-terminal side to the ionized positive chromophore, are marked with For peptides containing N-terminal leucyl (Leu, Figure la) a fragment ion mass indicating charge migration would be the mass 86 Da. Analogeously for N-terminal glycyl (Gly) containing peptides this would be the mass 30 Da (see Figure lb). For this experiment we have chosen a one-color excitation scheme as shown in Figure 2a) for a dipeptide X-Y. Local ionization is performed at the aromatic amino acids by resonant two-photon ionization (W photons). W excitation (same color, same laser) of the cation results in dissociation which can be preceded by charge transfer. Absorption of the cation occurs into a broad band of the peptide (see below). Because of static positive charge location
+.
-U
at the aromatic chromophore after localized ionization (see Table 1) W excitation has to be localized at the aromatic chromophore as well (neutral Gly, Leu does not absorb this wavelength). Reactivity and thus the possibility to observe charge transfer and charge mobility can be steered by variation of local dissociation thresholds. This is accomplished by variation of amino acids in the electron donor or electron acceptor position. Here ionization potentials of the amino acid constituents of the peptides becomes relevant. In Figure 3 the correlation between local ionization potentials and the energetical position of the dissociation threshold is demonstrated for N-terminal (a) and C-terminal (b) amino acids. Ionization potentials of the amino acids are given in Table 1 and are included in Figure 3. Because for some of the amino acids there are no experimental data available, estimated values are included which still can contain large errors. Within each class of amino acids, either N-terminal (Nterminal fragmentation) or C-terminal (chromophore fragmentation), the bond which has to be broken is the same, thus we can assume the dissociation energy to be similar within one class: 0.5 eV for the N-terminal fragmentation, because here the product ion H?NCH-R+ is strongly stabilized (low IP);1.5 eV for the C-terminal side, which mostly represents the C-C bond cleavage energy. As a result of that, the absolute position of the fragmentation threshold of the ion (absolute here means referred to the neutral molecule) in good approximation moves with the ionization potential of the amino acid and the order of dissociation thresholds can thus be given. Isomerization of N-terminal amino acids (scrambling of CH3 groups) does not change the situation significantly. In the case of chromophore fragmentation isomerization to the tropylium cation (for Phe) or related species for Tyr and Trp as well should not change the dissociation threshold substantially. In a rough approximation one can say that for the main part it is the ionization potential of the amino acids which influences the energetic position of the dissociation thresholds. We now transfer this assumption to peptides assembled of amino acids by assuming local ionization potentials and local dissociation energies. Thus we can shift local dissociation energies by changing amino acids in the peptide. For the following section it is only necessary to keep the order of ionization potentials and thus dissociation thresholds in mind: For N-terminal amino
Elementary Processes in Peptides
J. Phys. Chem., Vol. 99, No. 28, 1995 11259
a i N - terminal amino acids ( N -terminal fragmentation\ dissociation threshold
/\
0.5 e\l
------/ 0.5eV
.8 eV 3n
///
///
///
‘/ / Jeutral
Leu bl
G ~ Y
C - terminal amino acids Ichromophore
fragmentation I
dissociation thresholds
/I
TrP
-
------b -- -
TY r
Phe
Figure 3. Correlation of local ionization potentials of the amino acids with the absolute energetic position of the local dissociation threshold in peptide cations: (a) N-terminal amino acids: (b) C-terminal amino acids. Note that this ionic dissociation channels are inherently correlated with the positive charge to be in the relevant local part of the peptide.
acids (donor) I P G > ~ ~IPL~“, and for C-terminal amino acids (acceptor) IPPhe > IPT,, > IPT,. 111.1. Results: One-Color Experiments. In the one-color experiments laser intensity was usually 5 x IO6 W/cm2. Neither laser wavelength nor intensities substantially influence peak heights for the relevant fragment ions. As shown p r e v i ~ u s l y , * ~ - ?laser ~ . ~desorption ~.~~ with subsequent resonant two photon ionization can be performed without producing notable fragmentation, which clearly demonstrates that ionization of neutral fragments does not contribute to our mass spectra. Dipeptides. Laser-desorbed dipeptides have been investigated by Walter et al.52and by Li et al.56at high laser intensities of 1 x lo8 W/cm2 and (1-5) x IO7 W/cm2 leading to complicated fragment pattem due to multiphoton absorption in the cation. We have reinvestigated the dipeptides of the type Leu-Y and Gly-Y (Y = Trp, Tyr, Phe) by resonant ionization mass spectrometryat low laser intensities of (2-5) x lo7W/cm2 (one photon absorption in the cation) in order to be just above dissociation threshold and enhance effects of ET. In our experiments either Leu or Gly serves as an electron donor at the N-terminal end of the peptide and either Trp, Tyr, or Phe is used as the site of localized ionization at the C-terminal side of the peptide (acceptor). In Figure 4 MPI fragment mass spectra of the dipeptides Leu-Y (a-c) and Gly-Y (d-f) are shown (Y = Trp, Tyr, Phe). The local ionization potential and hence local absolute fragmentation energy at the C-terminal amino acids increases from a to c and analogeous from d to f. Note that the ionization potential of Leu (Figure 4a-c) is lower than the ionization potential of Gly (Figure 4d-f). Leu-Y-Dipeptides. In the MPI fragment mass spectra of the N-terminal leucyl-containing peptides Leu-Y (see Figure 4a-
c) the fragment ion mass 86 Da is prominent or even present solely. As Figure 4a shows, the fragment ion with mass 86 Da can be explained only by a positive charged N-terminal fragment (A fragment). Although the positive net charge in the peptide has been initially created in the chromophoric side chain of the C-terminal amino acid, the positive charge now is found to reside on the N-terminal fragment. For the formation of this N-terminal fragment ion the positive charge has to move over five bonds. The dominance of this process in Leu-Tyr and LeuPhe is shown by the complete absence of other fragment peaks (Figure 4b,c). Only for Leu-Trp the masses 130 Da (fragment from the aromatic side group, see insert in Figure 4a) and 187 Da (Z-fragment) indicate that both fragment channels are comparable in energy and that there is some positive charge stabilization at the aromatic side chain. The series a to c in Figure 4 shows that in accordance with the local ionization and thus absolute local dissociation thresholds of the chromophoric amino acids the fragment channel resulting in ET can be further enhanced by increasing the ionization energy of the electron acceptor. Gly-Y-Dipeptides. The situation changes dramatically for the dipeptides of the type Gly-Y. Note that the ionization potential of Gly is higher as that of Leu. In Gly-Trp (see Figure 4d) only very small intensities for the N-terminal fragment ion with mass 30 Da (see Figure 4b) are observed. Again the fragment ion with the mass 130 Da results from bond cleavage at the side chain indicating now strong stabilization of the positive charge at the indole ring. By comparing Leu-Trp (Figure 4a) and Gly-Trp (Figure 4d), the tendency of the positive charge to stabilize at the side chain of Trp in Gly-Trp has to be correlated with the higher ionization and thus the local dissociation threshold energy of Gly in comparison with Leu. For Gly-Tyr and Gly-Phe (see Figure 4e,f) the intensity of the N-terminal fragment ion with mass 30 daltons is strong. For Gly-Tyr the presence of chromophore fragment ions (mass 75 and 107 Da) as well as Z-fragments (164 Da) indicate that charge stabilization at the chromphore and the N-terminal fragment are comparable. For the dipeptide Gly-Phe the chromophore fragment ion with masses 91 and 104 Da as well as Z-fragments (147 Da) are very small. Obviously by variation of local ionization potentials we can switch fragment ion formation and hence charge position from the chromophoric side chain to the N-terminal end of the dipeptides. The observation of N-terminal fragment ions requires charge flow. For some cases the selectivity of the fragmentation process, which includes ET is surprisingly high. This selectivity in the dissociation is missing in the MPI mass spectra of Walter et al.52 and Li et because of their rich and complicated fragmentation at higher photon levels of the cation. Nevertheless the fragment ions characteristic for ET can be found in their mass spectra. Tripeptides. Analogous to our study of the dipeptides, the tripeptides Leu-Leu-Y and Gly-Gly-Y (Y = Trp, Tyr, Phe) have been investigated by resonant MPI mass spectrometry. Comparing the results on tripeptides with those on dipeptides should bring some affirmation of the observed charge transfer effect. The tripeptides Gly-Gly-Phe and Gly-Gly-Trp have been investigated before by Li et al.56by multiphoton ionization and dissociation at high laser intensities of (1- 5 ) x lo7 W/cm2. Some of our mass spectra of the tripeptides measured at low laser intensity ((2-5) x lo6 W/cm*) are shown in Figure 5. MPI mass spectra of dipeptides Leu-Tyr and Gly-Trp (a, c) and tripeptides Leu-Leu-Tyr and Gly-Gly-Trp (b, d) are compared. The laser wavelength has been set to agree with the absorption of the chromophore and has been the same for di- and
Weinkauf et al.
11260 J. Phys. Chem., Vol. 99, No. 28, 1995
1
130
L l y -Trp
M+ 261
M-le,
- f)
M' 218
C)
Leu - Phe
-
M' I222
I 30
86 10L 1L7 91 I
7
