and C-Terminal Methionine - ACS Publications - American Chemical

Article pubs.acs.org/JPCB

New Insights into the Reaction Paths of 4‑Carboxybenzophenone Triplet with Oligopeptides Containing N- and C‑Terminal Methionine Residues Piotr Filipiak,*,† Krzysztof Bobrowski,‡,§ Gordon L. Hug,†,§ Dariusz Pogocki,∥ Christian Schöneich,⊥ and Bronislaw Marciniak† †

Faculty of Chemistry, Adam Mickiewicz University, 61-614 Poznan, Poland Institute of Nuclear Chemistry and Technology, 03-195 Warsaw, Poland § Radiation Laboratory, University of Notre Dame, Notre Dame, Indiana 46556, United States ∥ Medicinal Chemistry Department, Institute of Biotechnology, University of Rzeszow, 35-959 Rzeszow, Poland ⊥ School of Pharmacy, Department of Pharmaceutical Chemistry, University of Kansas, Lawrence, Kansas 66047, United States ‡

S Supporting Information *

ABSTRACT: The oxidation processes of L-Met-(Pro)n-L-Met peptides that contain two Met residues located on the N and C termini and separated by a defined number (n = 0−4) of proline residues were investigated in aqueous solutions using laser flash photolysis. The use of such peptides allowed for distance control between the sulfur atoms located in the side chains of the Met residues. Interactions between side chains of the Met residues were probed by the observation of transients with σ*-type 2c-3e (S∴S)+, (S∴N)+, and (S∴O)+ bonds as well as of α-(alkylthio)alkyl radicals (αS). This approach enabled the monitoring, in real time, of the efficiency and kinetics of interactions between amino acid chains. Such knowledge is important, inter alia, for long-distance electron-transfer (ET) processes because amino acid side chains can serve as relay stations. The yields of these transients (measured as quantum yields (Φ)) were found to be dependent on the number of Pro residues, however, not dependent in a simple way on the average distance between sulfur atoms in Met residues. A decrease in the yield of the (S∴S)+ species with the number of Pro residues occurred at the expense of an increase in the yields of intramolecular three electron-bonded (S∴O)+/(S∴N)+ radical cations and αS radicals. These observations were rationalized by the fact that the time required for adequate overlap of the bonding orbitals is a key factor effecting the yield of the (S∴S)+ species. The time, however, can be controlled not only by the average distance but also by the specific geometrical and conformational properties of the peptide molecules. centered radical cation (Met>S•+), and (4) an additional reaction involving proton transfer from the protonated amine group to the ketyl radical anion (CB•−) when the peptide contains a protonated unsubstituted amine group. In peptides containing Met residues, the Met>S•+ radical cation can be stabilized through the formation of two-center three electron bonds (2c,3e). This process can occur intermolecularly with the sulfur atom of another peptide molecule (in peptides with a single Met residue). It can also occur intramolecularly with another sulfur atom (in peptides with multiple Met residues) and nitrogen and oxygen atoms located in the peptide bonds, irrespective of the number of Met residues. The efficiency of the formation of intermolecular dimeric (S∴S)+ radical cations and intramolecular (S∴S)+-, (S∴N)+-, and (S∴O)+-bonded

1. INTRODUCTION The photoinduced one electron oxidation of peptides containing Met residues has been thoroughly investigated over more than 20 years.1,2 Water-soluble derivatives of benzophenone (BP), such as 4-carboxybenzophenone (CB), were used as sensitizers. These compounds, upon absorbing UV light, yield an excited singlet state, which undergoes intersystem crossing (ISC) to give the nπ* triplet state (3CB*) on the ketone moiety. The 3CB* excited state can induce oneelectron oxidation of the electron-rich sulfur moiety in the side chain of Met through electron transfer. The global mechanism of the primary processes, following 3CB* formation, involves the formation of a short-lived radical ion pair [CB•−··· Met>S•+], which undergoes further reactions including: (1) back-electron transfer to the original substrates, (2) proton transfer with the formation of the respective ketyl radicals (CBH•) and α-(alkylthio)alkyl radicals, (3) escape of radical ions with the formation of radical anions (CB•−) and the sulfur© XXXX American Chemical Society

Received: February 4, 2017 Revised: April 4, 2017 Published: April 25, 2017 A

DOI: 10.1021/acs.jpcb.7b01119 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B

Chart 1. Structure of Radicals Formed during •OH-Induced and 3CB-Induced Oxidation of Met-(Pro)n-Met Oligopeptides (n = 0−4) at pH 5.75

amide functionalities. Second, because the excited carbonyl triplets can be produced “in vivo”,7,8 they can serve as significant oxidizing agents in biological systems and therefore can lead to enzyme inactivation and structural changes in proteins. The L-Met-(Pro)n-L-Met peptides, investigated in the current work, are oligopeptides that contain two Met residues located on the N and C termini and that are separated by a defined number (n = 0−4) of Pro residues. The use of such peptides allows for distance control between the sulfur atoms located in the side chains of Met residues. Such an approach was successfully applied in studying the distance dependence of (i) the rate constants for the long-range electron transfer between amino acid residues9 and metal centers,10 and (ii) the radiation chemical yields of intramolecular (S∴S)+-bonded radical cations formed via oxidation by •OH radicals.11

species (expressed in quantum yields) depends on structural, geometrical, and conformational properties controlled inter alia by the location, the number, and the optical isomerism of Met residues in the peptide molecule as well as by the character of the peptide chain (open/cyclic) and also by external properties of the system such as the pH of the environment and the concentration of peptides.3−6 The functional groups adjacent to the sulfur center, the primary site of initial oxidation, are clearly key to how the reaction is initiated and how it proceeds. Therefore, effects of neighboring groups on the reactions of the Met>S+• radical cation formed directly by one-electron oxidation of the sulfur atom in the Met residue by carbonyl triplets within particular peptides and protein domains may be of great importance. First, in these environments, Met residues are potentially surrounded by a wide variety of neighboring groups carrying thioether, carboxylate, amine, hydroxyl, and B


Article


filters were used to avoid spurious response from second-order scattering of the monochromator gratings. All experiments were carried out with a gravity-driven flow system and a rectangular quartz optical cell (0.5 × 1 cm). The monitoring-light path length was 0.5 cm. A solution of CB (2 mM) at neutral pH was used as a relative actinometer13 by monitoring its triplet−triplet absorption (ε535 = 6250 M−1 cm−1).14 Typically 5−10 laser shots were averaged for each kinetic trace. All solutions were deoxygenated by bubbling high-purity argon through them. The reservoir of the flow system typically was loaded with 50− 70 mL of solution per transient spectrum. 2.3. Photoexcitation of 4-Carboxybenzophenone (CB) in Aqueous Solution. Owing to their electrophilic character, the nπ* triplet states of aromatic ketones can be used as typical one-electron oxidants.15 For instance, 3CB* was shown to be an efficient one-electron oxidant.14,16−18 4-Carboxybenzophenone (CB) is first excited to the singlet excited state (1CB*), which subsequently undergoes an efficient intersystem crossing to the excited triplet state (3CB*) with a quantum yield of Φ = 1 for 3 CB* formation.16 2.4. Spectral Resolutions of Transient Absorption Spectra. Preliminary to the determination of quantum yields, a set of kinetic traces was collected for a sequence of monitoring wavelengths between 360 and 720 or 760 nm at 10 nm intervals. For each individual kinetic trace acquired, the dataacquisition systems automatically generate 10 kinetic traces on 10 distinct time scales. This redundancy of time scales makes it relatively easy to assemble transient spectra at convenient time delays following the laser pulse. Because there were numerous overlapping optical transients, a spectral-resolution procedure was used.18 It could be used after the transient spectra were collected and assembled for the relevant time windows. In each chosen time window, the resulting spectra were decomposed into the component spectra associated with the various transient species present via a multiple linear regression method based on writing the change in absorbance as

The motivation to study the oxidation mechanism of the LMet-(Pro)n-L-Met peptides induced photochemically by the CB excited triplet state is related to our recent studies of these peptides where the distance between sulfur atoms turned out to be a key factor in the formation of intramolecular (S∴S)+bonded radical cations in very acidic environment in which all of these peptides exist in the cationic form.11 There is also a key difference in the primary products when oxidation occurs by a strong one-electron oxidant such as 3CB* compared with oxidation by •OH radicals (Chart 1).11 The short-lived radical ion-pair [CB•−···>S+•] is formed, following 3CB* quenching, and this radical ion-pair decays by several reaction channels (including formation of the Met>S+• radical cation) which are controlled by the pH of the reaction environment.1 On the contrary, in the case of •OH-induced oxidation, the initially formed >S•−OH radical can decay using protons from water to form a monomeric sulfur-centered radical cation Met>S+•. However, the reaction pathways involving the Met>S+• radical cation should be the same, irrespective of the character of the oxidant and the pH of the reaction environment. Therefore, this approach enables monitoring, in real time, of the efficiency and kinetics of interactions between Met side chains at pH values where the peptides exist in the zwitterionic form. In this work, we investigated the oxidation processes starting from the transients formed by photochemically induced triplet states of 4-carboxybenzophenone (3CB*) at pH 5.7 to unravel the mechanisms of oxidation of Met residues in peptides with the changing distance between them. These oligopeptides can also serve as primary models for protein molecules containing multiple Met residues (e.g., calmodulin, prion proteins), where, for example, higher order structures have thioether functional groups located at various distances.

2. EXPERIMENTAL METHODS 2.1. Materials. The peptides Met-Met, Met-Pro1-Met, MetPro2-Met, Met-Pro3-Met, and Met-Pro4-Met were synthesized by the Biochemical Resource and Service Laboratory of The University of Kansas. They were purified to >95% purity and characterized by mass spectrometry. 4-Carboxybenzophenone (CB) was from Aldrich. The other chemicals were obtained as follows: Perchloric acid (HClO4) was purchased from Aldrich Chemical (Milwaukee, WI), and reagent grade NaOH was obtained from J. T. Baker (Center Valley, PA). The deionized water for time-resolved experiments (18 MΩ resistance) was purified in a reverse osmosis/deionization water system from Serv-A-Pure Co. The pH was adjusted by the addition of either NaOH or HClO 4 . Solutions were subsequently purged for at least 30 min per 500 mL of sample with high-purity argon. 2.2. Laser Flash Photolysis. The nanosecond laser flash photolysis system at the Notre Dame Radiation Laboratory was used for the time-resolved experiments. The data acquisition system has been previously described in detail.12 Laser excitation at 337.1 nm from a Laser Photonics PRA/Model UV-24 nitrogen laser (operated at about 4−6 mJ, pulse width ∼8 ns) was at a right angle with respect to the monitoring light beam. The detection system consisted of a pulsed xenon lamp (1 kW) as the monitoring light source and a SPEX 270 M monochromator coupled to a Hamamatsu R955 photomultiplier (PMT). The signal from the photomultiplier was processed by a LeCroy 7200 digital storage oscilloscope (OSC 7242 B) and a PC-AT compatible computer (PC). Cut-off

ΔA(λi) =

∑ εj(λi)aj j

(1)

where εj is the extinction coefficient of the jth species and the regression parameters, aj, are equal to the concentration of the jth species times the optical path length of the monitoring light. The sum in eq 1 is over all species present. For any particular time delay of an experiment, the regression analysis included equations such as eq 1 for each λi under consideration. The results of this procedure could then be used to produce concentration profiles of the transients with time. The reference spectra of these transients have been obtained in previous work.19 The molar absorption coefficients of the relevant transients, which will be further identified below, are the transients from CB (its triplet-state 3CB* ε540 = 6250 M−1 cm−1, radical anion CB•− ε650 = 7600 M−1 cm−1, and ketyl radical CBH• ε570 = 5200 M−1 cm−1),14,20 the α-(alkylthio)alkyl radicals, (αS), λmax = 290 nm and ε290 = 3000 M−1 cm−1),21 the dimeric (S∴S)+-bonded radical cation ((S∴S)+, λmax = 490 nm and ε490 = 5530 M−1 cm−1),22 and the (S∴O)+-bonded radical cation ((S∴O)+, λmax = 390 nm and ε390 = 3000 M−1 cm−1).19 Although various (S∴N)+ species were identified, they were incorporated as rescaling (using λmax = 390 nm and ε390 = 4520 M−1 cm−1) certain (S∴O)+ contributions because the spectral shapes of (S∴N)+ and (S∴O)+ are very similar.22 C


Article

The Journal of Physical Chemistry B Spectral resolutions from laser flash photolysis required taking into account the intramolecular (S∴S)+ radical cation. In experiments, the concentration of the Met-(Pro)n-Met substrates were kept low enough so that the formation of intermolecular (S∴S)+ was unlikely on the time scales monitored. To obtain a good-quality reference spectrum for intramolecular (S∴S)+, a nominal (S∴S)+ spectrum was taken from a report on the OH-oxidation of Met-Met at low concentration in an aqueous solution.22 This spectrum still had the possibility of having αS contributions in it. This αS contribution was estimated to be 20% for an analogous reaction in methionine.23 So to get the “uncontaminated” intramolecular (S∴S)+ spectrum, 20% of the αS spectrum21 was subtracted from the empirical (S∴S)+ spectrum.22 After the transient spectra were assembled, each of these spectra was resolved into component spectra associated with the various transient species deemed to be present. These spectral resolutions were made by fitting the reference spectra to the observed transient spectra via eq 1. Using this spectralresolution technique, concentrations of the transients were determined at any desired time delay following the laser pulse and are displayed as concentration profiles. 2.5. Determination of Quantum Yields. The resulting concentration profiles for the 3CB* quenching experiments at pH 5.75 were extrapolated back to the end of the laser pulse to make estimates of the initial quantum yields. Relative actinometry was used with separate cells of CB at a concentration such that the optical densities of the solutions at 337 nm were matched in the CB actinometry cell and in the quenching-solution cell. This gave the concentration of the triplet state from which the quantum yields could be calculated as a ratio of concentrations. This is possible because the triplet quantum yield of CB is normally taken to be one.16

Figure 1. Concentration profiles of product formation following triplet quenching of CB by Met-Pro-Met (A) and Met-(Pro)4-Met (B), [peptide] = 0.5 mM, [CB] = 2 mM, in Ar-saturated aqueous solutions, pH 5.75, relative actinometry: [3CB*] = 7.3 μM. The legends are inside the Figures. Insets: Evolution of the transient absorption spectra following triplet quenching of CB by Met-Pro-Met (A) and Met(Pro)4-Met (B), [peptides] = 0.5 mM, [CB] = 2 mM, in Ar-saturated aqueous solutions, pH 5.75, relative actinometry: [3CB*] = 7.3 μM. Spectra recorded after different time delays from the laser flash: (■) 270 ns (A and B); (red ●) 900 ns (A), 770 ns (B); (green ▲) 2.25 μs (A), 1.3 μs (B); (blue ▼) 4.5 μs (A), 2.25 μs (B); (cyan ◆) 15 μs (A), 6 μs (B); (pink ◀) 50 μs (A), 30 μs (B); (yellow ◆) 150 μs (A), 120 μs (B).

maximum at 570 nm. This is a strong indication that the CB ketyl radical is present. However, it is broader than the literature spectrum16 for the ketyl radical that further indicates that there are other products present, presumably from MetPro-Met and Met-(Pro)4-Met. The most likely transient species that would be broadening the ketyl radical spectrum at long delay times would have been intramolecular (S∴S)+, intramolecular (S∴N)+, intramolecular (S∴O)+, and the CB•− radical anion (vide infra in the Discussion). An intermolecular (S∴S)+ seems unlikely to appear at short delay times because the concentration of Met-Pro-Met and Met-(Pro)4-Met was only 0.5 mM. Hence, the intermolecular dimeric (S∴S)+ radical cation was not included in the spectral mix for the spectral resolutions. Even though the α-(alkylthio)alkyl radicals (αS) were very likely present following the triplet quenching events, their spectral contributions would be hidden under the ground-state absorption of CB. Therefore, in the spectral region investigated (see insets in Figure 1A,B), which avoids CB ground-state absorption in the near-UV, no αS radical absorption was considered in the spectral resolutions. As will be further rationalized in the Discussion section, both intramolecular (S∴N)+ and intramolecular (S∴O)+ were expected to contribute to the transient spectra, but their respective yields could not be extracted from the resulting transient spectra and could not be deduced a priori from the most plausible mechanism (see below). However, because their spectral shapes are very similar, the spectral shape of intramolecular (S∴N)+ was used in the spectral resolutions of the transient spectra from the triplet-quenching experiments. Such a technique acts as a “place holder” in the spectral resolutions, which allows for the determination of the

3. RESULTS 3.1. 3CB*-Induced Oxidation: Laser Flash Photolysis. The reaction of 3CB* triplets with Met-(Pro)n-Met (n = 0−4) peptides was investigated in Ar-saturated solutions at the concentration of 0.5 mM at pH 5.75. These experimental conditions were chosen for two reasons. First, the lower concentration of the peptides helps to limit any intermolecular (S∴S)+ formation. Second, at pH 5.75 there is a higher population of ω(cis)-conformers compared with that at pH 1,11 which means the distance between terminal Met residues will be different at these two pH values. 3 CB* can be conveniently used to create intramolecularly (S∴S)+-bonded radical cations cleanly (see below) in aqueous solutions at higher pH values where CB exists as an anion in its ground state. The 3CB* used in this work was shown to accept an electron from the sulfur moiety of Met residues, thus yielding a chargetransfer complex [CB•−···>S+•].17 In the initial steps of the photoreaction, this complex can undergo different reaction pathways: back electron transfer (kbt), separation of radical ions (ksep), “in cage” proton transfer from the Met moiety to CB•− (kH), and the proton transfer from the protonated amino group of the peptide to the radical anion CB•− (kNH).4 In the insets to Figure 1A,B, multiple transient spectra are displayed that follow the 337 nm laser flash of an aqueous solution of CB and Met-Pro-Met and of CB and Met-(Pro)4Met, respectively, at pH 5.75. The 270 ns transient spectra are identical to the literature spectrum of the triplet of CB.14 As the time delay increases these spectra evolve into a spectrum with a D


Article


determined as an unknown mixed absorbance contribution, separate concentrations of (S∴N)+ and (S∴O)+ could not be determined and thus are not shown in the concentration profiles in Figure 1A,B. The concentration profiles in Figure 1 were used to compute quantum yields by a relative actinometry method using an optically matched solution of CB as the actinometer. From the actinometry, the initial CB triplet concentration was 7.3 μM. Quantum yields were calculated in this manner for all five Met(Pro)n-Met (n = 0−4) and are listed in Table 1. The quantum yields were determined at their respective maximum concentrations for CB •− and CBH •. The quantum yield of intramolecular (S∴S)+ was calculated at the end of the triplet decay. As described above, (S∴N)+ and (S∴O)+ were accounted for jointly in the spectral resolutions, but their individual concentrations could not be determined. In addition, αS radicals could not be determined because of limitations in the observable spectral region. As will be discussed later, the sum of these three concentrations plus the maximum intramolecular (S∴S)+ should add up to the maximum concentration of CBH•.

concentrations of the other species that are present while still accounting for the absorption of an unspecified ratio of the spectral contributions of intramolecular (S∴N)+ and intramolecular (S∴O)+. Accounting for (S∴N)+ and (S∴O)+ in this manner, spectral resolutions were performed, and examples are given in Figure 2A,B for the time delays of 4.5 and 6 μs, respectively.

4. DISCUSSION 4.1. Mechanism for the 3CB*-Induced Oxidation (pH 5.75). The first step in the 3CB*-induced oxidation of Met(Pro)n-Met (see Scheme 1) is an electron transfer from the sulfur atom in the quencher to the carbonyl group of 3CB*. The result is a radical-ion pair, one member of which is the radical anion of CB and the other is a Met-(Pro)n-Met with a cationic radical site on one of the sulfur atoms. It is reasonable to suspect also that this radical-ion pair will be preferentially formed via oxidation of the N-terminal Met residue because of the favorable approach of negatively charged CB-triplets to the positively charged end of the Met-(Pro)n-Met species that are zwitterions at pH 5.75. That intermediate radical ion-pair can decay via four different pathways: (1) back electron transfer (kbt), (2) charge separation (ksep), (3) proton transfer leading to the formation of αS α-(alkylthio)alkyl radicals, (kH), and (4) proton transfer from an amino-group to the ketyl radical anion (kNH). The latter process would take place only at the Nterminal of the zwitterionic Met-(Pro)n-Met. It is plausible that the monomeric S-centered radical cation sites were not seen in the laser flash experiments because their possible absorption (λmax = 285 nm for [(CH3)2S∴OH2]+)24 would be masked by the ground-state absorption of CB and/or their decay was too rapid. One of the primary products derived from CB was the CB•− with observed quantum yields between 0.04 and 0.08 (see Table 1), depending on the number of Pro residues in the Pro bridges. At this pH (5.75), the CB•− decayed via a fast

Figure 2. Resolution of the spectral components in the transient absorption spectra (A) 4.5 μs after laser pulse following quenching of the CB triplet state by 0.5 mM Met-Pro-Met and (B) 6 μs after laser pulse following quenching of the CB triplet state by 0.5 mM Met(Pro)4-Met; [CB] = 2 mM, in Ar-saturated aqueous solutions, pH 5.75.

As a practical matter, when the concentration of the triplet reaches low values at longer delay times, a straightforward spectral resolution places too much emphasis on the triplet, even after the triplet has clearly decayed. To get more realistic concentrations of the other transients in this time range, the beginning triplet concentrations are extrapolated to longer delays by fitting the triplet concentrations at short time delays to a single exponential fit. Then, triplet concentrations can be extrapolated to longer times. These values at longer times were then fixed in the appropriate time windows, while the other species were subjected to the multilinear regression of the spectral resolution. Spectral resolutions were done for time windows up to 150 μs after the laser pulse. The resulting concentrations can be calculated from the spectral resolution for each species contributing to the final spectral mix at each wavelength. These concentrations are plotted in Figure 1A,B as a function of time. Because the absorbance of (S∴N)+ and (S∴O)+ were

Table 1. Quantum Yields of Transients Formed after 3CB*-Induced Oxidation of Met-(Pro)n-Met peptides (n = 0−4) Φ from laser flash photolysisa peptide

(S∴S)+ (end of triplet)

CBH•(max.)

CB•−(max.)

Met-Met Met-Pro1-Met Met-Pro2-Met Met-Pro3-Met Met-Pro4-Met

0.14 0.20 0.19 0.10 0.10

0.44 0.45 0.40 0.47 0.41

0.08 0.08 0.08 0.06 0.04

(S∴N)

+

+ (S∴O)

+

+ αSb

0.30 0.25 0.21 0.37 0.31

Uncertainties were estimated to be ±15%. bΦ (αS + (S∴N)+ + (S∴O)+) was calculated by subtraction of the quantum yield values for (S∴S)+ from CBH (max)

a

E


Article


Scheme 1. Mechanism of the 3CB*-Induced Oxidation of the C- and N-Terminal Met Residue in Met-(Pro)n-Met at pH 5.75

previous discussion connected to the OH-induced mechanism11 addresses this issue. 4.2. Quantitative Description of the 3CB*-Induced Oxidation. The sum of the yields of radicals derived from the peptides will be equal to the sum of the yields of radicals derived from CB. As discussed above, not all of the radicals from the peptides can be monitored in the laser flash photolysis experiments with optical detection; however, all of the radicals from CB can be seen in these experiments. Furthermore, from the time dependence of the concentration profiles in Figure 1AB, it can be seen that the CB•− radicals, formed in the separation of radical ion pairs, disappear rapidly at pH 5.75. Such behavior has been observed previously at pH well below the pKa = 8.216 of CBH•, indicating that the CB•− radical ions are being rapidly protonated forming CBH•.25 On the contrary, the concentration profiles of CBH• reach their maxima between 4 and 6 μs. On the basis of these considerations, the yield of CBH• can be taken as the full yield of radicals either from the peptides or from the CB. This was the basis for subtracting the observed concentrations of intramolecular (S∴S)+ species from the maximum concentrations of CBH• to give the sum of the yields of the other peptide radicals, that is, intramolecularly (S∴N)+-bonded five-membered species, the intramolecularly (S∴O)+-bonded six-membered species, and the αS α(alkylthio)alkyl radicals. Further quantitative information on individual quantum yields for these three species is not accessible directly from the data obtained from laser flash photolysis experiments with optical detection. However, using the underlying rate constants from a recent pulse radiolysis study,11 some estimates can be made for these other quantum yields, for example, that for (S∴O)+. The key to

protonation reaction that can be seen in the concentration profiles in Figure 1A,B. Ketyl radicals (CBH•) (total quantum yields between 0.40 to 0.47) can be formed by protonation of CB•− either in the reaction complex (in the kNH or kH reactions) or in the bulk, following the escape of CB•− from the reaction cage. In summary, at pH 5.75, CB•− produced CBH• using protons from water as well as from the protonated amino group or from the carbons alpha to the sulfur radical cationic site. Complementary with CB•− formation, the monomeric S-centered radical cationic site on Met-(Pro)nMet decayed by four different pathways: (1) forming intramolecularly (S∴S)+-bonded multimembered ring radical cations (e.g., for Met-(Pro)1-Met: Φmax = 0.20 at 3.75 μs and for Met-(Pro)4-Met: Φmax = 0.10 at 6.0 μs (end of triplet); see Figure 1A,B), (2) forming intramolecularly (S∴N)+-bonded 5membered species, (3) forming intramolecularly (S∴O)+bonded six-membered species, and (4) forming αS α(alkylthio)alkyl radicals. The formation of intramolecularly five-membered (S∴N)+bonded species at the N-terminal Met came from the kNH reaction (see Scheme 1S in the Supporting Information (SI)).4 Another type of intramolecular (S∴N)+-bonded fivemembered species can be also considered in Scheme 1. The formation of this species would involve the interaction of the lone pairs of the nitrogen atom from the peptide bond and the unpaired electron of the sulfur radical cation at the C-terminal Met. Although this is potentially possible, no such species was observed during 3CB*-induced oxidation of Pro-Met.4 Therefore, the presence of this species was not further considered. As far as the formation of the S∴O-bond species is concerned, our F


Article


k + kαS ⎞ 1 1 ⎛ 1 = (k SO + kαS)τdiff ⎜1 + SO ⎟+ ΦSS + Φsep ⎝ Φsep kact ⎠

this line of analysis is that the quantum yields of (S∴S)+ formation were directly measured, and, in addition, a mathematical expression can be written for them that incorporates the previously determined fundamental rates11 through the branching ratio shown graphically in Scheme 1. This approach is pursued in Subsection 4.2.1. The total quantum yield for the sum of the ketyl radical (CBH•) and its radical dianion (CB•−) can be estimated from the maximum of the quantum yield for CBH• because CB•− is rapidly protonated at pH 5.75 being more that 2 pH units below the pKa of CBH•. From Table 1 the maximum for each of the peptides studied was nearly the same and equal to ∼0.43. That leaves a quantum yield of ∼0.57 for back electron transfer (kbt). Within the radical-ion pair, three additional processes occur. These three processes involve charge separation by the diffusion apart of the radical ions (ksep) and two proton-transfer processes involving the methylene and methyl groups adjacent to the sulfur atoms (kH) and the protonated amine group at the N-terminal (kNH), leading to transient photoproducts (Scheme 1 and Scheme 1S in the SI). At the first glance, none of these three processes coming from the decay of the radical-ion pair should be S−S distance-dependent. Thus we tentatively assume that the efficiency of each of these three reaction channels is independent of the length of the proline chains in the peptides studied. 4.2.1. Utilizing the Approach of Szabo et al. to Incorporate Intramolecular Diffusive Motions into the Reaction Dynamics. A mathematical expression for the quantum yield of (S∴S)+ formation can be written that incorporates the previously determined fundamental rates11 through the branching ratio of >S+• into various decay channels shown graphically in Scheme 1. On the basis of Scheme 1, the quantum yield of (S∴S)+ formation can be written as the product of (1) the probability for radical-ion pair formation times, (2) the probability for radical-ion pair separation times, and (3) the branching ratio for the decay of the monomeric sulfur radical cation into the (S∴S)+ channel. Combining the first two factors as the quantum yield for radical-ion-pair separation,Φsep, the quantum yield for (S∴S)+ is given by eq 2 ΦSS + = Φsep

k SS + k SS + + k SO + kαS

(6)

The key equation that was used for the average-reaction or diffusion time,27 τdiff = 1/kdiff (or average first contact time), for diffusing together of our S-to-S distances is given by eq 7 1 kdiff

= τdiff = D−1

∫a

L

−1 dx {peq (x)(

∫x

L

dy peq (y))2 }

(7)

where the diffusion constant (D) was assumed, initially, to be 5 × 10−6 cm2 s−1 (a typical intermolecular D for a small protein), independent of distance. The boundary conditions27 for the diffusion processes were (i) reflecting at the maximum extension of the S−S distance, L, and (ii) absorbing when the S−S distance was equal to the contact distance, a, which was taken to be 3.6 Å.28 (We are taking this to be the sum of the van der Waals radii of two sulfur atoms.) The peq(rS−S) in eq 7 are the equilibrium distributions (statistical weights) of the intramolecular S−S distances in a Met-Pron-Met peptide. These peq(rS−S) distributions were computed from trajectories following all-atom Langevin dynamics (LD). 11 For a justification of using the statistical weights, peq(rS−S), from LD trajectories at pH 1 for the peq(rS−S) distributions at pH 5.75, see Section 4.4. Plotting 1 versus τdiff should give a straight line. Such a ΦSS +

plot is shown in Figure 3.

(2)

which can be transformed into eq 3 k + kαS ⎞ 1 1 ⎛ = ⎜1 + SO ⎟ ΦSS + Φsep ⎝ k SS + ⎠

(3)

Figure 3. Reciprocal quantum yields of (S∴S)+ formation versus average diffusion time computed from the diffusion-controlled rate of S−S closure from the Szabo et al. formula27 using statistical weights, peq(rS−S), for the peptides at pH 1 with D = 5 × 10−6 cm2 s−1.

As in the pulse radiolysis paper,11 kSS+ can be written in terms of the Noyes equation26 1 k SS +

=

1 kdiff

+

1 kact

From the slope and the intercept and the expressions for 1 them in eq 6, Φ (k SO + kαS) = 7.9 × 10 8 s −1 and

(4)

where kdiff is the specific rate constant if the reaction is diffusion controlled and kact is the specific rate constant called “activation-controlled”. Substitution of this Noyes equation into eq 3 gives eq 5 k + kαS k + kαS ⎞ 1 1 ⎛ = + SO ⎜1 + SO ⎟ ΦSS + Φsep ⎝ kdiff kact ⎠

sep

1 Φsep

(1 +

k SO + kαS kact

) = 5.59, respectively. Dividing intercept by

slope one can get intercept/slope =

(5)

equal to 7.1 × 10 Figure 3.

which can be further transformed into eq 6 G

−9

1 1 + k SO + kαS kact

(8)

s from the linear fit to the points shown in DOI: 10.1021/acs.jpcb.7b01119 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B In the pulse radiolysis work, the same systems gave kSO = 7.7 × 105 s−1 based on a typical value for kαS = 1.3 × 106 s−1, which was the key number used to calibrate11 the diffusion times for the relative intramolecular motion of the two sulfur atoms in the Met-Pron-Met peptides at pH 1. As Lapidus et al. found,29 we also found it necessary to calibrate our Szabo et al.27 calculation.11 Lapidus et al. employed temperature to do their calibration.29 We used the typical kαS (1.3 × 106 s−1)30 to help calibrate our kαS, and, as a consequence, this directly also calibrated the intramolecular diffusion constant.11 As previously argued,11 neither kSO nor kαS is expected to vary directly with the number of prolines in the bridges of the Met-Pron-Met peptides. We also do not expect them to vary much with the diffusion constant because deprotonation to form αS does not involve intramolecular diffusion and because the formation of (S∴O)+ involves minimal intramolecular diffusion. So we take kαS and kSO at pH 5.75 to be the same as those determined from the pulse radiolysis work at pH 1.11 The activation process, characterized by kact, also is not expected to differ at pH 5.75 compared with pH 1, where kact was found to be 6.6 × 106 s−1 under the assumption that kact is the same for all five peptides in that study11 at pH 1 and also in the current work at pH 5.75. Putting these values for kαS, kSO, and kact into eq 8 for the expected intercept to slope ratio, the result is 6.35 × 10−7 s. This expected value can now be compared with the value computed from the linear fit to the data in Figure 3 where the intercept-to-slope ratio is 7.1 × 10−9 s. The comparison can then be used to calibrate the intramolecular, diffusive relative motion of the two sulfurs in the peptides at pH 5.75 relative to that at pH 1. These expected and actual intercept-to-slope ratios differ by a factor of 90. This is interpreted to mean that the intramolecular diffusion is slower by a factor of 90 compared with the diffusion constant used in the Langevin dynamics simulation where the diffusion constant was taken to be the same as a typical intermolecular diffusion of small proteins (D = 5 × 10−6 cm2 s−1). This can be seen mathematically in the following manner. Although the expected intercept/slope ratio in eq 8 does not contain the relative diffusion of the sulfurs, on its activated rate, the horizontal axis specifically does contain it. The actual intercept to slope ratio for what is plotted in Figure 3 can be written as eq 9

The idea in the current calibration procedure is to shift the time axis through the diffusion constant to make the intercept to slope ratio at pH 1 (computed solely from the branching ratio using the three rates, kαS, kSO, and kact, determined from pulse radiolysis) to be equal to the intercept to slope ratio at pH 5.75 (computed from Figure 3 where the τdiff was computed from the Szabo et al. formula, eq 7, with a reasonable, but ultimately, arbitrary D). In this case, the intercept-to-slope ratio, computed from eq 8 using the three rates from the pulse radiolysis experiments at pH 1, are taken to be what the ratio should be based on the arguments about kαS, kSO, and kact being the same at pH 1 and pH 5.75. Then, to make the ratio from the plot in Figure 3 be the same using D as the only adjustable parameter, the actual D at pH 5.75 must be 90 times the arbitrary D = 5 × 10−6 cm2 s−1 used to compute the τdiff values in Figure 3. Therefore, we take D equal to (1/90) × 5 × 10−6 cm2 s−1 = 5.6 × 10−8 cm2 s−1 for the intramolecular diffusion constant associated with the relative motion between the two sulfurs in these peptides at pH 5.75. With the establishment of the diffusion constant (5.6 × 10−8 cm2 s−1) for the relative motion of the two sulfurs in Met-PronMet peptides at pH 5.75, the rate constant for the formation of the dimer of the intramolecular (S∴S)+ can be computed from the Noyes equation with kact from the pulse radiolysis analysis and k′diff (or 1/τ′diff) from the Szabo et al. formula, eq 7. The primes on k and τ indicate these are related to the relative intramolecular motion of the two sulfur atoms using the diffusion constant calibrated for the peptides at pH 5.75. Figure 3 is replotted with these recalibrated τ′diff values for the five peptides in Figure 4. The reciprocal (k′diff) of recalibrated average reaction times τ′diff for the five peptides are listed in the third column of Table 2.

( ) 1

⎛ 1 ⎞ Δ ΦSS+ intercept/slope = ⎜ ⎟ / ⎝ ΦSS + ⎠0 Δ(τdiff )

( ) 1

⎛ 1 ⎞ Δ ΦSS+ =⎜ ⎟ / ⎝ ΦSS + ⎠0 D−1Δ(Int ) ⎛ 1 ⎞ ⎛ 1 ⎞ = D−1Δ(Int )⎜ ⎟ ⎟ /Δ⎜ ⎝ ΦSS + ⎠0 ⎝ ΦSS + ⎠

Figure 4. Reciprocal quantum yields of intramolecular (S∴S)+ formation versus average diffusion time computed from the diffusion-controlled rate of S−S closure from the Szabo et al. formula27 using statistical weights, peq(rS−S), for the peptides at pH 1 with D = 5.6 × 10−8 cm2 s−1.

(9)

and is equal to 7.1 × 10−9 s. In eq 9, Int stands for the result of the multiple integrations over peq(rS−S) in eq 7. From eq 8, the intercept-to-slope was 6.35 × 10−7 s computed above. We use this difference in the two values of the intercept-to-slope ratio to calibrate the relative diffusion of the intramolecular S−S motion. Similar to the pulse radiolysis analysis, we do not change the peq(rS−S) distribution derived from the Langevin dynamics, we change D to 5.6 × 10−8 cm2 s−1 to make the intercept-to-slope ratios the same from the two different ways of writing and computing it.

Having the recalibrated diffusion constant for the relative motion of the two sulfurs with a peptide at pH 5.75, the rate, kSS+, for the formation of intramolecular (S∴S)+ can be computed from a Noyes equation26 1 1 1 = + k SS + k′diff kact (10) H


Article

The Journal of Physical Chemistry B Table 2. Rates and Quantum Yields Computed from the Underlying Dynamics Met-Met Met-Pro-Met Met-Pro2-Met Met-Pro3-Met Met-Pro4-Met

τ′diff (ns)

k′diff (s−1)

kSS+ (s−1)

ΦSS+a

ΦSOb

ΦαSc

Φremandd

25 37 13 480 530

× × × × ×

× × × × ×

0.17 0.17 0.18 0.10 0.10

0.023 0.025 0.022 0.049 0.051

0.040 0.042 0.037 0.083 0.086

0.24 0.18 0.14 0.24 0.17

4.0 2.7 8.0 2.1 1.9

7

10 107 107 106 106

5.7 5.37 6.17 1.67 1.57

6

10 106 106 106 106

a

Computed from eq 2. bComputed from eq 12. cComputed from eq 13. dSum of quantum yields of radicals formed in processes associated with kNH and kH.

using k′diff computed from the Szabo et al. equation with D = 5.6 × 10−8 cm2 s−1 and the kact from pH 1 and the analysis of the pulse radiolysis data. The resulting values of kSS+ for the five peptides studied are listed in the fourth column of Table 2. 4.2.2. Quantum Yields. The plot of 1 versus τ′diff in

processes (last column of Table 2), characterized by the rates kNH and kH shown in Scheme 1, could be estimated by subtracting the theoretically computed quantum yields in Table 2, columns 6 plus 7, from the experimentally measured sum of the quantum yields of (S∴O)+, αS, and (S∴N)+ in the last column of Table 1. Fifth, from the third column in Table 2, it can be seen that the three shortest peptides have k′diff that are larger than kact = 6.6 × 106 s−1 for intramolecular (S∴S)+ formation, whereas the two longest peptides have k′diff smaller than kact. Thus intramolecular (S∴S)+ formation is activationlimited in the three shortest peptides and is diffusion limited in the two longest peptides. Similar kinetics were found in the dynamics of intramolecular (S∴S)+ formation at pH 1 in the pulse radiolysis experiments.11 4.3. Sulfur−Sulfur Distance Dependence in the Formation of Intramolecular (S∴S)+ Radical Cations. Some of the >S+• radicals resulting only from the charge separation channel (ksep) do form intramolecularly (S∴S)+bonded radical cations, and this reaction pathway is S−S distance-dependent. On the contrary, the other two secondary reaction pathways, involving the >S+• radicals, which lead to either αS radicals or (S∴O)+-bonded radical cations, are not directly S−S distance-dependent. Their indirect S−S distance dependence is again due to the fact that these reactions are competitive with the formation of the intramolecularly (S∴S)+bonded radical cations.11 Moreover, αS radicals are also formed via proton-transfer process (kH) (Scheme 1). The ΦSS+ values obtained from laser flash photolysis were also plotted against the number of Pro residues (Figure 5). Except for Met-Met, there was a qualitative correlation between the ΦSS+ listed in Table 1 and the number of proline residues. In particular, as the number of prolines increased, the

ΦSS +

Figure 4 should still follow the modified eq 6 as k + kαS ⎞ 1 1 ⎛ 1 = (k SO + kαS)τ′diff ⎜1 + SO ⎟+ ΦSS + Φsep ⎝ Φsep kact ⎠ (11)

From the slope of the linear fit in Figure 4 and the kSO and kαS from the pH 1 analysis of the pulse radiolysis data,11 Φsep is calculated to be 0.24. Φsep can also be calculated from the intercept of the linear fit in Figure 4 using the same two rates, kSO and kαS, in addition to kact from the pH 1 analysis of the pulse radiolysis data.11 This result coming from the intercept for the value of Φsep is the same as that calculated from the slope. With the values of Φsep and the five values for kSS+ available, along with the values of kSO and kαS, it is possible to compute a theoretical ΦSS+ from these fundamental rates and the separation probability using eq 2. It is also possible to extend the quantitative characterization of the mechanism slightly forward by computing ΦSO k SO k SS + + k SO + kαS

(12)

kαS k SS + + k SO + kαS

(13)

ΦSO = Φsep

and ΦαS ΦαS = Φsep

coming through the monomeric sulfur radical and also the sum of the quantum yields of two of the other processes coming out of the radical ion pair, associated with the rates kNH and kH; see Scheme 1 and Scheme 1S in the SI. The results of these quantum yield calculations are listed in Table 2. From these newly computed parameters using rates carried over from the analysis of the pulse radiolysis data at pH 1 and the rates of kSS+, several new perspectives can be seen. First, Φsep, assumed to be constant for the five peptides is computed to be 0.24 from both the slope and intercept of the linear fit to the 1 versus τ′diff in Figure 4 based on eq 11. Second, using ΦSS +

this Φsep, τ′diff, and the fundamental rates from the pH 1 analysis, the resulting ΦSS+ in the fifth column of Table 2 is in reasonable agreement with the experimental ΦSS+ in the second column of Table 1. Third, with the underlying rates, ΦSO could be estimated, along with ΦαS coming through the ion pair state; neither of these could be measured experimentally with laser flash photolysis. Fourth, the sum of the quantum yields of the

Figure 5. Dependence of ΦSS+ (■) on the number (n) of proline residues in the case of 3CB-induced oxidation of Met-(Pro)n-Met peptides at pH 5.75. Values for quantum yields are ±15%. I


Article

The Journal of Physical Chemistry B measured ΦSS+ values decreased (Figure 5). However, it should be noted that the largest change in ΦSS+ occurred when the number of proline residues was changed from 2 and 3. Interestingly, the dependence of the ΦSS+ values on the number of Pro residues was weaker compared with the analogous dependence of G((S∴S)+) (radiation chemical results) on proline number.11 Furthermore, as the number of Pro residues increased, the sums of the Φ values for (S∴N)+, (S∴O)+bonded species and α-(alkylthio)alkyl radicals showed a very poor correlation with proline number. A possible explanation of the latter behavior might be the following: It should be noted that part of the αS radicals and the (S∴N)+-bonded radical cations are formed in proton-transfer processes that are not directly competitive with the formation of the intramolecularly (S∴S)+-bonded radical cations. 4.4. Langevin Dynamics and Choice and Meaning of Diffusion Constant. The five peptides in this study are zwitterions at pH 5.75. Langevin dynamics simulations similar to those done on the monocations in the pulse radiolysis study11 at pH 1 were initially done on the zwitterions for the current work. However, because these Langevin dynamics trajectories were done in a continuum solvent with no explicit water molecules, the peptides had a strong tendency to artificially form salt bridges between the charged N- and Cterminals of the zwitterions. LD simulations on these peptides at pH 111 largely avoided salt bridges because the C-terminal was protonated and uncharged. To avoid salt bridges, it was assumed that the diffusive motions associated with the intramolecular relative S−S distances were not affected by whether the acid group was protonated or not. To this end, the initial average reaction times from the Langevin dynamics at pH 1 were used also for the current pH 5.75 studies. Both in the pulse radiolysis study at pH 1 and in the current laser flash photolysis study at pH 5.75, the average reaction time, τdiff, in the Szabo et al. equation (eq 7) was calculated taking the diffusion constant equal to 5 × 10−6 cm2 s−1, which is a typical diffusion constant for small proteins. However, even in the pH 1 studies, it was necessary to calibrate the diffusion constant based on a typical rate (1.3 × 106 s−1) for deprotonation of monomeric sulfur radical cations of thioethers.30 That diffusion constant for the relative intramolecular diffusive motion of the two sulfurs turned out to be (1/283) × 5 × 10−6 cm2 s−1 = 1.8 × 10−8 cm2 s−1, where 283 was the calibration factor found for τdiff at pH 1. For comparison, the calibrated diffusion constant at pH 5.75 is (1/90) × 5 × 10−6 cm2 s−1 = 5.6 × 10−8 cm2 s−1, as computed above. Both calibrations are ultimately based on the choice of 1.3 × 106 s−1 for deprotonation of monomeric sulfur radical cations of thioethers. Although the factor of three difference between the two calibrated diffusion constants at pH 1 and pH 5.75 may be simply a consequence of the assumptions made above, there may be a physical basis for this difference. One of the assumptions of this work is that the diffusive intramolecular motion of the relative position of S−S in the unoxidized peptides was the same as the intramolecular diffusion of an unoxidized sulfur and an oxidized one. The diffusion of two unoxidized sulfurs in our Langevin dynamical calculation leading to the distribution of peq(rS−S) in eq 7 would be expected to be little affected by the positive charge on the Nterminal amino group at pH 1 and pH 5.75 or by the negative charge on the C-terminal group at pH 5.75.

However, at pH 1 the oxidized sulfur on either the N- or Cterminal methionine would be repelled by the positive charge at the N-terminal of the peptide chain. This would lead to a slower relative motion for the C-terminal oxidized sulfur moving toward the unoxidized N-terminal sulfur. The positive charge in the amino group would lead to a faster relative motion for an N-terminal oxidized sulfur, which would be repelled by the positively charged protonated amino group at pH 1. Because the oxidizing agent in the pulse radiolysis experiment at pH 1 is the uncharged hydroxyl radical, the probability of which of the two sulfurs would be oxidized should be about the same. Thus the speeding-up and slowingdown depending on which sulfur was oxidized should largely cancel out. In the laser flash photolysis experiments at pH 5.75, there are two significant differences in regard to the electrostatic situation with that in the pulse radiolysis experiments at pH 1. First, the peptides are zwitterions, and, second, the oxidizing agent is a negative ion. Thus it would be expected that the oxidizing agent would be more likely to attack the N-terminal methionine. (The difference in energetics (ΔG) for electron-transfer reactions from the N-terminal and C-terminal sulfur atoms is expected to be small (see the SI).) The resulting oxidized sulfur would be simultaneously repelled by its nearby positively charged, protonated amino group and attracted by the negative charge at the deprotonated acid group on the C-terminal methionine. So the N-terminal oxidized sulfur at pH 5.75 would face a stronger intramolecular electrostatic field than the field at pH 1, and there would be inadequate compensation by C-terminal-oxidized sulfurs going slowly in the other direction because of the decreased probability that these C-terminal oxidations would be taking place, or at least they would be appropriately less weighted in the average reaction times.

5. CONCLUSIONS The quantum yields (ΦSS+) of intramolecular three electronbonded (S∴S)+ species formed in Met-(Pro)n-Met peptides in the zwitterionic form do not depend in a simple way on the average distance between sulfur atoms in Met residues. They depend weakly on the number of proline residues and follow exactly the same trend found for the radiation chemical yields of the (S∴S)+ species formed in the same peptides present in the cationic form via oxidation by •OH radicals.11 In particular, as the number of prolines increased, the measured ΦSS+ values decreased and the largest change in ΦSS+ occurred when the number of proline residues was changed from two and three. These features were also quantitatively reproduced by Langevin dynamics (LD) and a simple statistical mechanical theory of Szabo, Schulten, and Schulten (SSS) applied previously for peptides in the cationic form. This analysis showed that the formation of a contact between terminal Met residues in the peptides with zero to two Pro residues was controlled by the activated formation of (S∴S)+, whereas that in the peptides with three and four Pro residues was controlled by the relative diffusion of the sulfur radical cation and unoxidized sulfur atom. The diffusion-limited rates were calculated, however, using an apparent chain diffusion coefficient (D) that is higher by a factor ∼3 than the apparent chain diffusion coefficient of peptides in the cationic form. This higher value of D and consequently smaller values of the average diffusion times (τ′diff) for LFP results may arise from differences in the J


Article


zophenone Triplet State in Aqueous Solution. Competition Between Intramolecular Two-Centered Three-Electron Bonded (SS)+ and (SN)+ Formation. Photochem. Photobiol. 2000, 72, 1−9. (7) Cilento, G.; Adam, W. Electronic Excitation in Dark Biological Processes. In Chemical and Biological Generation of Excited States; Adam, W., Ed.; Academic Press: New York, 1982; pp 277−307. (8) Cilento, G. Generation of Electronically Excited Triplet Species in Biochemical Systems. Pure Appl. Chem. 1984, 56, 1179−1190. (9) Bobrowski, K.; Poznański, J.; Holcman, J.; Wierzchowski, K. L. Long range electron transfer between proline-bridged aromatic amino acids. In Photochemistry and radiation chemistry. Complementary methods for the study of electron transfer; Wishart, J. F., Nocera, D. G., Eds.; Advanced Chemistry Series, vol. 254; American Chemical Society: Washington, DC, 1998; pp 131−143. (10) Isied, S. S.; Ogawa, M. Y.; Wishart, J. F. Peptide-Mediated Intramolecular Electron Transfer: Long-Range Distance Dependence. Chem. Rev. 1992, 92, 381−394. (11) Filipiak, P.; Bobrowski, K.; Hug, G. L.; Pogocki, D.; Schöneich, C.; Marciniak, B. Formation of a Three-Electron Sulfur-Sulfur Bond as a Probe for Interaction between Side Chains of Methionine Residues. J. Phys. Chem. B 2016, 120, 9732−9744. (12) Thomas, M. D.; Hug, G. L. A Computer-Controlled Nanosecond Laser System. Comput. Chem. (Oxford, U. K.) 1998, 22, 491−8. (13) Carmichael, I.; Hug, G. L. Triplet-Triplet Absorption Spectra of Organic Molecules in Condensed Phases. J. Phys. Chem. Ref. Data 1986, 15, 1−250. (14) Hurley, J. K.; Linschitz, H.; Treinin, A. Interaction of Halide and Pseudohalide Ions with Triplet Benzophenone-4-Carboxylate: Kinetics and Radical Yields. J. Phys. Chem. 1988, 92, 5151−5159. (15) Scaiano, J. C. Intermolecular Photoreductions of Ketones. J. Photochem. 1973, 2, 81−118. (16) Inbar, S.; Linschitz, H.; Cohen, S. G. Quenching, Radical Formation, and Disproportionation in the Photoreduction of 4Carboxybenzophenone by 4-Carboxybenzhydrol, Hydrazine, and Hydrazinium Ion. J. Am. Chem. Soc. 1981, 103, 7323−7328. (17) Bobrowski, K.; Marciniak, B.; Hug, G. L. 4-Carboxybenzophenone-Sensitized Photooxidation of Sulfur-Containing Amino Acids. Nanosecond Laser Flash Photolysis and Pulse Radiolysis Studies. J. Am. Chem. Soc. 1992, 114, 10279−10288. (18) Marciniak, B.; Bobrowski, K.; Hug, G. L. Quenching of Triplet States of Aromatic Ketones by Sulfur-Containing Amino Acids in Solution. Evidence for Electron Transfer. J. Phys. Chem. 1993, 97, 11937−11943. (19) Bobrowski, K.; Hug, G. L.; Pogocki, D.; Marciniak, B.; Schöneich, C. Sulfur Radical Cation-Peptide Bond Complex in the One-Electron Oxidation of S-Methylglutathione. J. Am. Chem. Soc. 2007, 129, 9236−45. (20) Hurley, J. K.; Sinai, N.; Linschitz, H. Actinometry in Monochromatic Flash Photolysis: The Extinction Coefficient of Triplet Benzophenone and Quantum Yield of Triplet Zinc Tetraphenyl Porphyrin. Photochem. Photobiol. 1983, 38, 9−14. (21) Hiller, K.-O.; Asmus, K.-D. Tl2+ and Ag2+ Metal-Ion-Induced Oxidation of Methionine in Aqueous Solution. A Pulse Radiolysis Study. Int. J. Radiat. Biol. Relat. Stud. Phys., Chem. Med. 1981, 40, 597− 604. (22) Bobrowski, K.; Holcman, J. Formation and Stability of Intramolecular Three-Electron S∴N, S∴S, and S∴O Bonds in OneElectron-Oxidized Simple Methionine Peptides. Pulse Radiolysis Study. J. Phys. Chem. 1989, 93, 6381−6387. (23) Hiller, K.-O.; Masloch, B.; Göbl, M.; Asmus, K.-D. Mechanism of the OH-Radical Induced Oxidation of Methionine in Aqueous Solution. J. Am. Chem. Soc. 1981, 103, 2734−2743. (24) Chaudhri, S. A.; Göbl, M.; Freyholdt, T.; Asmus, K.-D. A Method to Generate and Study (CH3)2S+• Radical Cations. Reduction of DMSO by H• Atoms in Aqueous HClO4 Solutions. J. Am. Chem. Soc. 1984, 106, 5988−5992.

electrostatic situation present in cationic and in zwitterionic forms of these peptides after oxidation. A decrease in the yield of the (S∴S)+ species does not show a clear correlation with the sums of the Φ values for (S∴N)+/(S∴O)+ species and α-(alkylthio)alkyl radicals (αS). This behavior can be rationalized by the fact that part of (S∴N)+ and αS radicals is formed in proton-transfer processes, which are not directly competitive with the formation of (S∴S)+-bonded radical cations.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcb.7b01119. Scheme 1S. Mechanism of the 3CB*-induced oxidation of the N-terminal Met residue (kNH reaction channel) in the Met-(Pro)n-Met at pH 5.75 (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*Tel. +48 61 829-1738. Fax. +48 61 829-1555. E-mail: piotrf@ amu.edu.pl. ORCID

Piotr Filipiak: 0000-0001-9141-2163 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the National Science Centre in Poland under Grant UMO-2013/11/B/ST4/00811 (P.F.) and by the U.S. Department of Energy, Office of Basic Energy Science under Award Number DE-FC02-04ER15533 (K.B.). P.F. and K.B. acknowledge the NDRL accelerator staff and Professor Ian Carmichael for his hospitality during their stays. D.P. acknowledges the Interdisciplinary Centre for Computational Modeling in the University of Rzeszow for the possibility of performing computations (the computational grant: G-017). This is document number NDRL-5166 from the Notre Dame Radiation Laboratory.

■

REFERENCES

(1) Bobrowski, K.; Houée-Levin, C.; Marciniak, B. Stabilization and Reactions of Sulfur Radical Cations: Relevance to One-Electron Oxidation of Methionine in Peptides and Proteins. Chimia 2008, 62, 728−734 and references therein.. (2) Ignasiak, M. T.; Marciniak, B.; Houée-Levin, C. A Long Story of Sensitized One-Electron Photooxidation of Methionine. Isr. J. Chem. 2014, 54, 248−253 and references therein.. (3) Ignasiak, M. T.; Pedzinski, T.; Rusconi, F.; Filipiak, P.; Bobrowski, K.; Houée-Levin, C.; Marciniak, B. Photosensitized Oxidation of Methionine-Containing Dipeptides. From the Transients to the Final Products. J. Phys. Chem. B 2014, 118, 8549−8558. (4) Hug, G. L.; Bobrowski, K.; Kozubek, H.; Marciniak, B. Photooxidation of Methionine Derivatives by the 4-Carboxybenzophenone Triplet State in Aqueous Solution. Intramolecular Proton Transfer Involving the Amino Group. Photochem. Photobiol. 1998, 68, 785−796. (5) Marciniak, B.; Hug, G. L.; Bobrowski, K.; Kozubek, H. Mechanism of 4-Carboxybenzophenone-Sensitized Photooxidation of Methionine-Containing Dipeptides and Tripeptides in Aqueous Solution. J. Phys. Chem. 1995, 99, 13560−13568. (6) Hug, G. L.; Bobrowski, K.; Kozubek, H.; Marciniak, B. PhotoOxidation of Methionine-Containing Peptides by the 4-CarboxybenK


Article

The Journal of Physical Chemistry B (25) Bobrowski, K.; Marciniak, B. The Kinetics of the Acid-Base Equilibrium of 4-Carboxybenzophenone Ketyl Radical. A Pulse Radiolysis Study. Radiat. Phys. Chem. 1994, 43, 361−364. (26) Noyes, R. M. Effects of Diffusion Rates on Chemical Kinetics. Progress in Reaction Kinetics and Mechanism 1961, 1, 129−160. (27) Szabo, A.; Schulten, K.; Schulten, Z. First Passage Time Approach to Diffusion Controlled Reactions. J. Chem. Phys. 1980, 72, 4350−4357. (28) Bondi, A. van der Waals Volumes and Radii. J. Phys. Chem. 1964, 68, 441−451. (29) Lapidus, L. J.; Steinbach, P. J.; Eaton, W. A.; Szabo, A.; Hofrichter, J. Effects of Chain Stiffness on the Dynamics of Loop Formation in Polypeptides. Appendix: Testing a 1-Dimensional Diffusion Model for Peptide Dynamics. J. Phys. Chem. B 2002, 106, 11628−11640. (30) Mönig, J.; Goslich, R.; Asmus, K.-D. Thermodynamics of S∴S 2σ/1σ* Three-Electron Bonds and Deprotonation Kinetics of Thioether Radical Cations in Aqueous Solution. Ber. Bunsenges. Phys. Chem. 1986, 90, 115−121.

L


and C-Terminal Methionine - ACS Publications - American Chemical

Recommend Documents