Chemiluminescence Associated with Singlet Oxygen Reactions with Amino Acids, Peptides and Proteins^†

Materials. Rose Bengal (RB); 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox); 2-phenyl-1,2-benzisoselenazol-3[2H]-one (Ebselen); Xylenol Orange; reduced glutathione; l-tryptophan (98%); l-tyrosine (Tyr) (>99%); l-cysteine (>99%); l-methionine (Met) (>99%); l-histidine (His) (>99%); Ala-Trp (>99%); Trp-Ala (>99%); Ala-Tyr (>99%); Tyr-Ala (>99%); Gly-His (>99%); His-Gly (> 99%); ascorbic acid; lipid-free albumin from human serum (HSA); lipid-free albumin from bovine serum (BSA); phosphatase alkaline type I-S from bovine intestinal mucous (AP); lysozyme from chicken egg white (LYZ); ribonuclease A, type III-A from bovine pancreas (RA); insulin from bovine pancreas (INS); catalase from bovine liver; and Poly-l-Trp were purchased from Sigma (USA). Ala-Trp-Ala (94%) and Ala-Tyr-Ala (>99%) were from Bachem Feinchemikalien AG (Switzerland). Solutions were prepared employing twice-distilled water pretreated with Chelex-100 (Sigma), 1 g L⁻¹ for 12 h.

Singlet oxygen reactions. Singlet oxygen was generated in the irradiation of RB in air-saturated solutions. To this purpose, light from a halogen lamp was filtered through a 400 nm cut-off glass filter. The solution containing the RB (5 μM) and the target molecule (amino acids, peptides or proteins) in 10 mM phosphate buffer, pH 7.0, were kept at fixed temperature (4°C or 25°C) during the irradiation (5 min.). Peroxide formation was evaluated by a FOX methodology based on the formation of a Xylenol Orange/Fe(III) complex following the peroxide promoted oxidation of Fe(II) in a 1 mM phosphate buffer solution. The formation of the complex was quantified by the absorption of the sample at 560 nm (21). Calibrations carried out in the presence of the buffer were indistinguishable from those obtained in distilled water.

Measurement of CL from pre-irradiated samples. CL intensities were measured in a Beckman LS 6500 scintillation counter working in the out-of-coincidence mode, using the narrow tritium iso-set module. The photomultiplier had a nearly constant response up to 650 nm. The equipment was calibrated employing the hemin-hydrogen peroxide-luminol reaction as a standard (9), rendering an efficiency of 2.3 × 10²¹ counts/Einstein.

After irradiation of solutions containing RB and the substrates, 3 mL of the sample was transferred to a polyethylene vial and introduced into the scintillation counter. The elapsed time from irradiation to the start of the CL measurement was about 30 s. When the photolysis was carried out at 4°C, the CL was measured in solutions kept at near 0°C by their introduction into previously frozen vials containing 0.5 g of ice (buffer solution). Each one of these vials was placed in a larger glass vial to minimize its heating rate during the recording time.

The spectral distribution of the CL was assessed from the changes in intensity observed when different filtering solutions were introduced between the polyethylene vials containing the solution and the external glass vials. Three filters were employed: a concentrated solution of RB (cut-off 580 nm), and solutions of potassium dichromate (0.05 M) in either 6 M HCl (cut-off 540 nm) or 3 M NaOH (cut-off 490 nm).

In order to establish the mechanism of the CL, antioxidants (700 μM Trolox or 100 μM ascorbic acid), a peroxidase-like compound (120 μM Ebselen), Fe(II) (5 mM FeSO₄ ) or a metal chelator (0.1 mM dietylentriaminepentaactic acid [DTPA]) were added after the irradiation. In order to evaluate the role of singlet oxygen in the formation of the CL precursors and in the CL process, azide (1 or 5 mM) was added prior to or after the RB irradiation. To further assess the role of singlet oxygen in the production of the CL precursors, irradiations were also performed in deuterated phosphate buffer, pD 7.0, or in enriched deuterium oxide solutions. In some experiments, photolyzed solutions were diluted in heavy water (10:1) in order to assess the role of singlet oxygen in the after-photolysis emission.

CL following the reaction of free amino acids with singlet oxygen

The CL profiles of amino acid solutions after irradiation (500 μM, 5 min irradiation) are shown in Fig. 1. These data show that only pre-irradiated solutions containing Tyr and Trp, and to a minor extent those containing His, produce a CL significantly larger than that of the control sample (RB irradiated alone). The CL profile can be well fitted to biexponential decays. These decays are dominated by the fast component, whose lifetimes and amplitudes are included in Table 1. These values show that the CL precursors generated in the three systems have lifetimes shorter than 5 min.

CL following the reaction of peptides with singlet oxygen

In order to evaluate the effect of the peptide linkage upon the formation of CL precursors, the emission from several peptides containing either Trp, His or Tyr moieties were measured. The values obtained are included in Table 1. The data show that the presence of the peptide linkage produces only minor changes in the CL of Tyr or His groups, although the decay seems to be slower in the Tyr tripeptide than in the corresponding dipeptides.

In contrast, in Trp-bearing peptides, there were noticeable differences among the tested compounds. An interesting feature of these data is that the emission intensity was strongly dependent upon the peptidic bonds of the compound. In particular, it was observed that at 25°C the initial CL followed the order

In order to see if these large differences were due to a fast decay during the time elapsed between photolysis and CL measurements, we carried out photolysis and CL measurements at lower temperatures. As expected, slower decays were observed, allowing a more precise estimation of the “initial” CL. The values obtained under these conditions showed a much smaller difference between Ala-Trp and Ala-Trp-Ala. This indicates that the large difference observed between these compounds when the measurements were carried out at 25°C was mostly due to a faster decay of the CL precursors in the tripeptide. On the other hand, a difference of ca three orders of magnitude remained in the emissions of Ala-Trp and Trp-Ala, reflecting true differences in the efficiency of CL precursor formation.

Time profiles of the CL arising from pre-irradiated samples containing Ala-Trp and Trp-Ala are shown in Fig. 2. The data showed a noticeable difference between the behavior of the two peptides, a difference that was emphasized by the data collected in Table 1. In fact, while the behavior of Trp-Ala was similar to that of free Trp, the Ala-Trp dipeptide presented a CL that was nearly three orders of magnitude higher than that of the free peptide. The same difference was observed when the photolysis and the CL measurements were carried out at a lower temperature, a condition that should minimize the decomposition of the intermediates in the time elapsed between the photolysis and the CL intensity measurement. The tripeptide, particularly at the lower temperature, had an emission of the same order of magnitude as the Ala-Trp dipeptide, and much higher than that of the Trp-Ala dipeptide. These results indicate that formation of the peptide linkage at the amino group of the Trp moiety leads to a noticeable increase in emission intensity. This difference is not due to differences in the extent of oxidation, as the consumption of the compounds during the irradiation period, as well as the amount of peroxides, is similar (Table 2).

CL following the reaction of proteins with singlet oxygen

The CL emitted after irradiation of RA and INS, proteins lacking Trp moieties, was considerably smaller than that of the Trp-bearing proteins (Table 3). However, it must be noted that Trp-lacking proteins, and in particular RA, emit a CL that is more than an order of magnitude higher than that of Tyr-containing small peptides.

When proteins with Trp groups (HSA, BSA, AP and LYZ) were photolyzed in the presence of RB, their CL showed characteristics similar to those previously described for Ala-Trp and Ala-Trp-Ala peptides. Typical examples are shown in Fig. 3. A common feature of these data is that the decay was more complex than that of the small peptides and could only be fitted to multiexponential decays. We have thus characterized the process by the initial intensity, the integrated intensity (from 0 to 20 min), and the time required for the CL intensity to decay to half of the initial value. These data are summarized in Table 3. It can be concluded that the after-photolysis emission of the protein is even more intense than that of the peptides. Furthermore, these data also show that, in spite of the high CL intensities of all Trp-bearing proteins, there are noticeable differences in the efficiency of the process. In fact, emission from pre-irradiated HSA and BSA was considerably more intense than that of the other proteins.

Spectral distribution of the CL arising from pre-oxidized Trp-containing peptides and proteins

The spectral distribution of the CL arising from preoxidized peptides and proteins was assessed employing cut-off solutions. The results obtained are summarized in Table 4. These data indicate that the observed CL is not due to singlet oxygen dimol emission and that it can be attributed to the emission from carbonyl group containing products derived from the singlet oxygen-mediated oxidation of the parent molecules. We cannot discard a priori a mechanism involving the dark formation of singlet oxygen (from endoperoxides formed during the photolysis) and its reaction with the remaining peptides and/or their products. However, the observed CL was not reduced by the addition of 5 mM sodium azide, a singlet oxygen quencher via a physical mechanism, after the photolysis of Ala-Trp. On the other hand, its addition prior to photolysis almost totally abolished the after-photolysis emission. This would indicate that the CL precursors are produced in singlet oxygen-mediated reactions taking place during the photolysis and not in Type I photooxidation processes. This conclusion is further supported by the increase in after-photolysis CL observed when the irradiation was performed in (partially) deuterated water (Table 5). Furthermore, the production of chemiluminescent precursors is associated with the presence of oxygen as no CL was observed when the pre-irradiation was carried out in nitrogen-purged solutions (data not shown).

Effect of additives upon the CL following the reaction of Trp containg peptides and proteins

In order to obtain some insight into the characteristics of the products that lead to the post-irradiation CL and the mechanism of the process, several additives were added after irradiation, prior to the CL measurements. Typical results are shown in Fig. 4, and main conclusions are summarized in Table 6. The total suppression of the CL elicited by Fe(II) or Ebselen addition would support the proposal that the emission arises from the cleavage of peroxides. In this regard, the CL in this system behaves as that observed in other protein oxidations (8) and in red cells or brain homogenate oxidations (22–24). However, a peculiar feature of the present data is the lack of effect of free radical scavengers (Trolox or ascorbic acid). This contrast with results obtained in other systems where, at least partially, the CL emitted from oxidized products is enhanced by free radical processes (7–9).

The effect of catalase is less straightforward. Its addition slightly increases the rate of CL decay, without any effect on the initial intensity. These results could be attributed to a slow suicidal decomposition of hydroperoxide intermediates (see following scheme) by catalase (25).

Estimation of the CL quantum yield

A comparison of the integrated CL with the amount of modified Trp groups allows a rough estimation of the CL quantum yield, defined as the number of photons emitted by the reacted Trp moiety. In the evaluation of the integrated CL it has been assumed that:

With these assumptions, the quantum yield for Ala-Trp is 2.5 × 10⁻⁸, showing, that in spite of the high intensities emitted, the CL pathway corresponds to a very minor fraction of the whole reaction between Trp groups and singlet oxygen. Regarding the multiplicity of the emitting species, it is interesting to note that the CL intensity is very similar in air-saturated or oxygen-saturated solutions (data not shown). This would indicate that it corresponds to fluorescence from a singlet state and/or phosphorescence from a very short-lived triplet.

Mechanism of the CL process

The very low CL quantum yields makes difficult to establish the mechanism of the process. The CL can arise from a main process for which a very low proportion of the products are able to emit, and/or for a very minor process with high luminescence yields. The proposed reaction scheme (Scheme 1) is based on the first premise and only considers the main processes. This reaction scheme is based on previous studies regarding the intermediates formed in singlet oxygen-mediated oxidation of Trp groups (11,18,20,26,27). From the present results, it is proposed that, if the NH₂ group is not blocked (as in Trp and Trp-Ala), most of the reacting groups follow a dark pathway through the intermediate C. The process

takes place in several steps and, in the absence of free radicals, would not provide a CL pathway (7). On the other hand, in the highly CL compounds (Ala-Trp, Ala-Trp-Ala or the proteins) reaction via intermediate C is blocked. The major pathway involves intermediate E, leading to emission through a process such as

For Ala-Trp, an additional reaction pathway through intermediate F could contribute to the observed CL. A dark decomposition of their peroxide precursors (E and F) by Fe(II) or Ebselen would explain the quenching of the CL elicited by their addition.