|Home | About | Journals | Submit | Contact Us | Français|
Both the neurotransmitter serotonin and the unnatural amino acid 5-hydroxytryptophan (5HT), contain the 5-hydroxyindole chromophore. The photochemistry of 5HT is being investigated in relation to the multiphoton excitation of this chromophore to produce a characteristic photoproduct with green fluorescence (‘hyperluminescence’). Laser flash photolysis (308 nm) of 5HT in aqueous solution at neutral pH produces both the neutral 5-indoloxyl radical (λmax 400-420 nm) and another transient absorption with λmax 480 nm and lifetime of 2 μs in deaerated solutions. Based on quenching by oxygen and β-carotene, the species at 480 nm is identified as the triplet excited state of 5HT. In acidic solution a new oxygen-insensitive intermediate with λmax 460 is assigned to the radical cation of 5HT. Time-resolved measurements of luminescence at 1270 nm have shown that the triplet state of 5HT is able to react with oxygen to form singlet excited oxygen (1O2*) with a quantum yield of ~0.1. However, 5HT has also been found to be an effective quencher of singlet oxygen with a second order rate constant of 1.3 × 108 dm3 mol-1 s-1. The results are discussed in the light of recent observations on the multiphoton-excited photochemistry of serotonin.
The 5-hydroxyindole chromophore of the unnatural amino acid, 5-hydroxytryptophan (5HT), and the neurotransmitter, serotonin (5-hydroxytryptamine), confers at least two significant spectroscopic differences compared with the indole chromphore of tryptophan. Firstly, the absorption spectrum of 5-hydroxyindoles is red-shifted compared with that of tryptophan allowing selective excitation of 5HT in the presence of tryptophan [1-4]. The ultraviolet fluorescence of 5HT has a higher quantum yield than tryptophan, but shows less solvent sensitivity [2,3,5] and is similar to that from serotonin . These properties have allowed 5HT to be used as a probe of protein structure and dynamics [7-9]. Imaging of serotonin fluorescence within mast cells through selective 3-photon excitation has also been reported . Secondly, both 5HT and serotonin are induced to produce an unusual and characteristic green, ca. 500 nm, luminescence (‘hyperluminescence’) on multiphoton excitation using sub-picosecond near infrared laser pulses [11-15]. Shear et al  have shown that overall this is a 6-photon process, involving an initial 4-photon photochemical reaction leading to a transient product with a lifetime of hundreds of microseconds followed by two-photon excitation of the intermediate to produce the green fluorescence. Similar fluorescence may also be produced by two colour (308 and 430 nm) excitation of 5HT with nanosecond pulsed lasers . More recent studies have indicated that the fluorescing intermediate may be produced by a 3-photon near infrared excitation plus a further single photon excitation delayed by up to tens of nanoseconds . This implies the potential role of an intermediate state with a lifetime longer than that of the excited singlet. Despite these investigations the identity of the photochemical product responsible for hyperluminescence has not yet been established, although the effects of several scavengers of radicals and excited states have been reported [14-17]. Capillary electrophoresis experiments with serotonin indicate the unknown intermediate is similar in size and charge to serotonin itself .
The interest in spectroscopic properties of 5HT suggests that the photochemistry of 5-hydroxyindoles deserves more detailed investigation. Our previous laser flash photolysis investigations at 308 nm  indicated that photoionisation to produce the hydrated electron and the 5-indoloxyl radical was one of the main photochemical pathways, confirming previous work using 248 nm laser photolysis . The triplet state of 5HT has been observed by optically detected magnetic resonance and phosphorescence at low temperatures [19-21]. At 135 K the phosphorescence lifetime is long (τ = 4.8 s), but is reduced to 29 μs at 274 K . In contrast, room temperature phosphorescence of tryptophan in thoroughly deaerated solution has a significantly longer lifetime (1.4 ms at 274 K) and is also more long-lived than the transient absorption (τ ~ 14 - 20 μs) assigned to the triplet state observed by laser flash photolysis (22,23). This apparent discrepancy has been rationalised in terms of a radical recombination mechanism by Fletcher et. al. . The triplet states of phenols have also been detected [25-27], but their absorption spectra at λ > 300 nm have low extinction coefficients and overlap spectra of the radical that is also produced by photolysis.
Singlet oxygen (1Δg) may be produced by photosensitization reactions involving energy transfer from both singlet and triplet states of organic molecules in solution, although the latter are more commonly important because of their relatively longer lifetimes [27-29]. Singlet oxygen is a reactive species that is quenched by both chemical and physical mechanisms by biomolecules, including indoles and phenols [30-33]. A relatively specific physical quencher of singlet oxygen is 1,4-diazabicyclo[2,2,2]octane (DABCO) . Recently Gostkowski et al  have found that multiphoton-induced luminescence from serotonin is enhanced on addition of DABCO, suggesting that in its absence singlet oxygen might be produced and destroy serotonin within the typically femtolitre volume of multiphoton excitation.
We have therefore undertaken studies to identify transient intermediates formed by laser flash photolysis at 308 nm of 5HT under a range of conditions. Formation and quenching of singlet oxygen by 5HT was also investigated using time-resolved singlet oxygen luminescence measurements.
All chemicals were obtained from Sigma-Aldrich, were of the highest purity available and used as received. Laser flash photolysis at 308 nm employed a Lumonics PM846-I excimer laser producing ca 10 ns pulses with intensities of ca 10 mJ/pulse. The kinetic single beam spectrometer comprised a Xe lamp (Applied Photophysics), grating spectrometer (Bentham Instruments TM300) and Hamamatsu photomultiplier tube (type R928). Data was collected using a digital oscilloscope (Tektronix TDS 3012). Time-resolved singlet oxygen luminescence was detected using a liquid nitrogen-cooled germanium photodiode (North Coast EO-817P) in conjunction with a 1270 nm interference filter. The experiments involving photosensitised formation of singlet oxygen with Rose Bengal used the 532 nm frequency-doubled output (ca. 5 mJ/pulse, ca 7 ns) of a Continuum Powerlite 8000 Nd:YAG laser.
Transient spectra from laser flash photolysis of 5HT in aqueous solution buffered to pH 7.0 are shown in Figure 1. As previously reported for deaerated (argon-saturated) solutions , excitation at 308 nm leads to photoionisation as demonstrated by formation of the hydrated electron with λmax ca. 720 nm . Figure 1A shows the transient absorption spectra at 200 ns and 8 μs after the laser pulse in oxygenated solutions of 5HT (0.2 mmol dm-3) at pH 7.0. The main feature of these spectra is the absorption maximum at 400-420 nm which has previously been identified at the neutral 5-indoloxyl radical (16,35-37) formed by electron and proton loss from the 5-hydroxyindole (HO-Ind) part of 5HT (reaction ). The 5-indoloxyl radical appears not to react with oxygen and is comparatively stable with a lifetime of hundreds of microseconds under the conditions of our experiment and has previously been shown to decay through a second order process .
In the spectrum measured at 200 ns after the laser pulse there is also evidence for the hydrated electron (λmax ~ 720 nm) and an additional weak absorption at about 460-500 nm. The rapid reaction of e-aq with O2 (k2 = 1.9 × 1010 dm3 mol-1 s-1 would lead to a lifetime of about 50 ns in oxygen-saturated solution. The inset to Figure 1A (trace (a)) shows the transient decay at 480 nm; the estimated lifetime is 200 ns which indicates that at this wavelength another species other than the hydrated electron is present. The small increase in absorbance on the scale of ten microseconds illustrated in this transient was also routinely observed and has been ascribed to a minor reaction involving the conversion of a low yield of nitrogen- to the oxygen-centred 5-indoloxyl radical after the initial photolysis .
The hydrated electron reacts rapidly with N2O (9.7 × 109 dm3 mol-1 s-1) and the higher solubility of N2O (ca 22 mmol dm-3) in aqueous solution compared with oxygen (ca. 1 mmol dm-3) should ensure that the hydrated electron in N2O-saturated solution is too short lived (τ ~ 5 ns) to be observed in our experiment under such conditions. Figure 1B shows that on laser flash photolysis of N2O-saturated solutions of 5HT the spectrum measured 200 ns after the laser pulse is similar to that in oxygenated solutions but with a slightly more prominent shoulder at about 480 nm. Trace b) in the inset Figure 1A shows that at 480 nm this component has a lifetime of 2.2 μs. Subtraction of the remaining spectrum at 6 μs after the laser pulse from that at 200 ns is also shown in Figure 1B and shows a peak at 480 nm. We ascribe this transient absorption to the excited triplet state of 5HT (35HT). Measurements of the decay of this species in solutions saturated with varying ratios of N2O and O2 indicated that it reacted with oxygen with a second order rate constant of (7 ± 1) × 109 dm3 mol-1 s-1.
Laser flash photolysis of 5HT in deaerated methanol and dimethylsulfoxide (DMSO) gave essentially similar transient spectra as in the aqueous solutions described above. The main difference was that compared with aqueous solution, in methanol 35HT was shorter-lived (τ = 0.56 μs), whereas in DMSO it had a longer lifetime (4.0 μs). Further evidence for the identification of the 480 nm band as 35HT was obtained through the use of β-carotene which has a very low yield of intersystem crossing and which is therefore not directly excited to the carotene triplet, but which can act as a triplet energy acceptor (38,39). Addition of low concentrations of β-carotene to solutions of 5HT in DMSO resulted in an increased rate of decay of the 35HT absorption at 480 nm, and the appearance of a shoulder in the transient absorption spectrum at 520 nm characteristic of the β-carotene triplet-triplet absorption (data not shown).
On addition of > 1 mol dm-3 of HCl to solutions of 5HT, the transient absorption spectrum changes to exhibit a single peak at 460 nm. The lifetime of this species is similar in both oxygen and nitrous oxide saturated solutions and decayed exponentially with τ = 1.15 μs in 80% HCl. Transient spectra measured over a range of HCl concentrations are shown in Figure 2. The species with an absorption maximum at 460 in very acidic solutions is tentatively assigned to the radical cation of 5HT which is the initial photolysis product from photoionisation and which deprotonates to the neutral indoloxyl radical at higher pH values (see Reaction ). A plot of absorbance for this species versus the Hammett acidity function, H0, indicates a pKa of −0.7 ± 0.1. This is somewhat higher than pKa ≈ -2 for simple phenols  and compares with the pKa of the 4-aminophenol radical cation (pKa 2.2 ). The pKa value emphasises the tendency towards the iminosemiquinone nature of the 5-indoloxyl radical.
The observation of the triplet excited state of 5HT suggests the possibility of triplet energy transfer to ground state triplet oxygen to yield the 1Δg excited state of oxygen:-
The formation of excited state singlet oxygen was detected by time resolved emission at 1270 nm from singlet oxygen, employing a fast germanium photodiode, after pulsed 308 nm laser excitation of solutions of 5HT in air-saturated D2O/H2O 80/20 v/v containing 40 mmol dm-3 phosphate (equimolar Na2HPO4and NaH2PO4)). A range of solutions were prepared with absorbance values at 308 nm (A308) up to ~ 0.3. A transient luminescent signal at 1270 nm was recorded from these solutions, the intensity of which increased with increasing concentration and absorbance of the solution at 308 nm. The lifetime of the observed signal decreased from 16.2 μs at 30 μmol dm-3 5HT to 11.8 μs at 90 μmol dm-3 5HT. Extrapolation to zero 5HT concentration gave a lifetime of 18 μs. This is consistent with the lifetime of singlet oxygen in D2O/H2O (80/20 v/v) taking data from the compilation of Wilkinson et al (31), and confirms that the observed luminescence arises from singlet oxygen. Furthermore, the luminescence was observed to be quenched by the addition of azide ion. The decrease in singlet oxygen lifetime with increasing 5HT concentrations indicates that 5HT is also a singlet oxygen quencher (see below). The quantum yield for singlet oxygen formation from 5HT was obtained by comparison of the emission intensities from 5HT compared with those obtained from solutions of Rose Bengal under similar conditions, for which the quantum yield is reported as 0.76 (29). In order to account for the observed fast formation of singlet oxygen in the air-saturated solutions, emission intensities were determined by integration of the intensity-time decay profiles and corrected for the observed quenching of the singlet oxygen luminescence lifetime by both Rose Bengal and 5HT. Figure 3 shows plots of the intensity of singlet oxygen emission from solutions of Rose Bengal or 5HT in D2O/H2O (80/20 v/v) having a similar range of absorbance values at 308 nm. The relative slopes give a quantum yield of 0.12 ± 0.01 for formation of singlet oxygen from photoexcitation of 5HT at 308 nm. This also indicates a lower limit for the quantum yield for the triplet in 5HT in aqueous solution. This compares with a quantum yield of 0.27 for fluorescence from the excited singlet state of 5HT (5).
Although the experiments reported above show that photoexcitation of 5HT leads to singlet oxygen formation, it is well known that indoles and phenols are also highly efficient quenchers of singlet oxygen (31-33 and references therein). Experiments were therefore undertaken in which singlet oxygen was generated in oxygen-saturated D2O (at pD ~7.8) solutions by pulsed laser photolysis of Rose Bengal at 532 nm, where 5HT does not absorb. In all cases the singlet oxygen decays were strictly first order. The singlet oxygen decay rate (kobs) increased with the addition of increasing concentration of 5HT according to:-
where k0 is the solvent and temperature dependent decay rate in the absence of 5HT, and where k2 is the second order rate constant for quenching of singlet oxygen by 5HT. The value of k at 25 °C was determined to be (1.26 ± 0.04) × 108 dm3 mol-1 s-1. Compared with values for singlet oxygen quenching measured in D2O/acetonitrile mixtures (33) this is about three times the singlet oxygen-quenching rate constant for tryptophan (4.1 × 107 dm3 mol-1 s-1), and is over one hundred times faster than the rate constant for quenching of singlet oxygen by N-acetyl tyrosine ethyl ester at ~ pD 7 (8.5 × 105 dm3 mol-1 s-1). The comparatively low one-electron reduction potential at pH 7 (E0’) of the 5-indoloxyl radical (640 mV (18)) correlates with the more rapid quenching of singlet oxygen by 5HT compared with tyrosine (E0’ 930 mV (42)) and tryptophan (E0’ 1020 mV (42)).
The first order rate constant for decay of singlet oxygen in solutions containing a range of 5HT concentrations were measured between 15 °C and 75 °C as shown in Figure 4. Activation parameters were obtained from the Arrhenius plot of natural logarithm of the second order rate constant versus reciprocal temperature. The activation energy for the second order rate constant with 5HT was 15.0 ± 0.6 kJ mol-1. The corresponding enthalpy of activation (ΔH‡) and entropy of activation (ΔS‡) were 12.5 ± 0.6 kJ mol-1 and −47.8 ± 1.8 J K-1 mol-1 respectively.
Both the effect of quencher oxidation potential and the thermodynamic parameters are consistent with the proposed mechanism for singlet oxygen quenching by phenols and amines involving reversible formation of an exciplex with considerable charge transfer character (43,44). An overall mechanism is shown in Scheme 1, emphasising the apparent paradox of 5HT being involved in both formation and quenching of singlet oxygen. The efficiency of the sensitisation of singlet oxygen by a triplet state organic molecule can be influenced by several factors (30,45). These include the triplet state energy and oxidation potential (Eox) of the sensitiser that may favour a charge transfer process (charge transfer is negligible if the Eox of the sensitiser is greater then 1.8 V vs SCE, (30)). Other factors also play a role in determining the quenching mechanism where both energy and charge transfer are thermodynamically favourable. These include the nature of the triplet excited state, i.e. n-π* or π-π*, and structural factors within the collision complex formed between oxygen and the organic molecule for example as determined for amines (46). Of particular importance is how this can influence the ability of the nascent singlet oxygen to escape from the collision complex and, as in this case, avoid the back reaction of the nascent singlet oxygen being quenched by 5HT. The mechanism of this is likely to be through a charge transfer process indicated in scheme 1, although we have found no evidence for the final separation into radical species even in the case of a quencher phenol with low oxidation potential (33). The fact that we observe a shortening of the singlet oxygen lifetime with increasing 5HT concentration demonstrates that separation of the collision complex occurs and diffusion of singlet oxygen and 5HT occurs following triplet sensitisation.
The results have demonstrated that in addition to the neutral 5-indoloxyl radical, laser flash photolysis at 308 nm of 5HT in neutral aqueous solution generates the triplet state of 5HT with a T-T absorption at 480 nm and with a lifetime of 2.2 μs. In very acidic solutions, the radical cation is also observed. Observation of time-resolved luminescence at 1270 nm shows that energy transfer from the triplet state of 5HT generates singlet molecular oxygen with a quantum yield of ~0.1. However 5HT is also a quencher of singlet oxygen, with a rate constant higher than for either tryptophan or phenol. The direct observation of the 5HT triplet state in fluid solution at room temperature will allow a more detailed investigation of the involvement of this state as an intermediate in the generation of the species responsible for ‘hyperluminescence’. The possibility might be considered that the either the neutral radical, or the triplet state might be the states induced by three photon absorption and further excited by subsequent one-photon process as described by Gostkowski et al (17). However the lifetime of at least tens of nanoseconds determined for this intermediate excludes it being the radical cation, for which a sub-nanosecond lifetime in neutral solution is estimated on the basis of its pKa. The observation of sensitisation of singlet oxygen formation from photo-excited 5HT also accounts for the increase in hyperluminescence intensity in the presence of the singlet oxygen quencher DABCO.
We are grateful to the Biotechnology and Biological Science Research Council for a grant in support of this work.