Regiospecific Nitrosation of N-terminal-blocked Tryptophan Derivatives by N2O3 at Physiological pH*

N2O3 formed from nitric oxide in the presence of oxygen attacks thiols in proteins to yield S-nitrosothiols, which are believed to play a central role in NO signaling. In the present study we examined theN-nitrosation of N-terminal-blocked (N-blocked) tryptophan derivatives in the presence of N2O3 generating systems, such as preformed nitric oxide and nitric oxide donor compounds in the presence of oxygen at pH 7.4. Under these conditionsN-nitrosation of N-acetyltryptophan and lysine-tryptophan-lysine, respectively, was proven unequivocally by UV-visible spectroscopy as well as 15N NMR spectrometry. Competition experiments performed with the known N2O3 scavenger morpholine demonstrated that the selected tryptophan derivatives were nitrosated by N2O3 with similar rate constants. It is further shown that the addition of ascorbate (vitamin C) induced the release of nitric oxide from N-acetyl-N-nitrosotryptophan as monitored polarographically with a NO electrode. Theoretical considerations strongly suggested that the reactivity of protein-bound tryptophan would be high enough to compete effectively with protein-bound cysteine for N2O3. Our data demonstrate conclusively that N2O3nitrosates the secondary amine function (N indole) at the indole ring of N-blocked tryptophan with high reactivity at physiological pH values.

Nitric oxide ( ⅐ NO) is involved in a great variety of physiological and pathophysiological processes, including the regulation of vascular tonus, immune response, and neurotransmission (1). In mammals the basal level of ⅐ NO (in the aqueous phase) is in the nanomolar range (2). However, the local ⅐ NO concentrations can be increased substantially in areas with high NO synthase activity and/or in hydrophobic regions (3,4), thereby facilitating the formation of dinitrogen trioxide (N 2 O 3 ) (Equations 1 (5) and 2 (6)). the putative formation of S-nitrosothiols is now generally believed to be of high physiological importance in vivo (Ref. 9 and references therein). For instance, S-nitrosothiols can operate as nitric oxide-donating compounds in vitro (Equation 5 (10)), and they are believed to induce ⅐ NO-like biological activities in vivo, probably by releasing ⅐ NO (11,12).
RSNO ϩ ascorbate/Cu 2ϩ 3 RSNO ϩ products ϩ Cu 1ϩ 3 ⅐ NO ϩ RS Ϫ ϩ Cu 2ϩ (Eq. 5) Further, S-nitrosation of thiols is postulated to induce several functional protein modifications in vivo (13)(14)(15)(16)(17)(18). In addition to thiols, secondary amines also react rapidly with N 2 O 3 (7). Consequently, the question arose whether (protein-bound) tryptophan could be nitrosated at the nitrogen atom of the indole ring at physiological pH 7. 4. It has been demonstrated recently that melatonin (N-acetyl-5-methoxytryptamine) is N-nitrosated by NaNO 2 /HCl as well as by ⅐ NO/O 2 at pH 7.4. However, the reported rate constant for the latter reaction is rather low (k(melatonin ϩ ⅐ NO) ϭ 0.5 M Ϫ1 s Ϫ1 ) (19). Keeping the low physiological concentrations of melatonin in mind (e.g. the maximal concentration in human serum is ϳ75 pg/ml (20)), this reaction therefore should not proceed to a significant extent in vivo. A similar reaction has not been established thus far for tryptophan at pH 7.4. However, there are indications that the secondary amine of tryptophan (N indole ) can be nitrosated when the primary amine function (N alanine ) is N-blocked. 1 Consequently, albumin (3,(21)(22)(23) or, more simply, N-acetyltryptophan (24) were N-nitrosated by N 2 O 3 generated in situ by HCl/NaNO 2 (Scheme 1). Nitrosamines in general are well known carcinogens (25). Similarly, Venitt et al. (26) observed mutagenic effects of N-acetyl-N-nitrosotryptophan in bacteria.
On the other side, as protein-bound N-nitrosotryptophan was able to both stimulate vasorelaxation and inhibit platelet aggregation (21), there might be a putative physiological significance of N-nitrosotryptophan derivatives. This idea, however, has not been developed further, probably because thiols are now believed to be the main physiological targets for N 2 O 3 in proteins.
In contrast to this opinion, we demonstrate unequivocally here the N 2 O 3 -mediated nitrosation of N-blocked tryptophan (N-acetyltryptophan and the tripeptide lysine-tryptophan-lysine) by N 2 O 3 at physiological pH 7.4. In addition, competition experiments performed with the common N 2 O 3 scavenger morpholine demonstrate that N 2 O 3 reacts quickly with N-terminalblocked tryptophan. Finally, as liberation of ⅐ NO is evident after reaction of N-acetyl-N-nitrosotryptophan with ascorbate, N-nitrosotryptophan is expected to exhibit similar ⅐ NO-donating capabilities as S-nitrosocysteine.
Experimental Conditions-Because nitrosation reactions are sensitive to the presence of metal ions, solutions were exposed to chelating resin (Chelex 100) as described previously (27).
Nitrosation by Preformed ⅐ NO Gas-Nitric oxide ( 14⅐ NO) or 15 Nlabeled nitric oxide ( 15⅐ NO) was prepared daily by the addition of glacial acetic acid to an aqueous solution of K 4 [Fe(CN) 6 ] (0.7 M) containing either NaNO 2 or Na 15 NO 2 (0.72 M) under oxygen-free conditions. The overall stoichiometry is presented by the equation, The reliability of this procedure was confirmed by detecting the release of nitric oxide polarographically with a graphite nitric oxide-sensing electrode (World Precision Instruments, Berlin, Germany).
N-blocked derivatives of tryptophan were nitrosated in oxygen-saturated phosphate/triglycine buffer (250/25 mM, pH 8.0) with the above described preformed ⅐ NO gas by bubbling it gently into the solution through a needle until the pH value had dropped to 7.4.
Nitrosation by NO Donor Compounds-MAMA NONOate and spermine NONOate were prepared as 100-fold stock solutions in 10 mM NaOH at 4°C and used immediately. A stock solution of nitrosocysteine was prepared by S-nitrosation of 100 mM cysteine with equimolar amounts of NaNO 2 in acidic solution (pH 2) at 0°C (24). Nitrosation reactions were performed in 1 ml of potassium phosphate/NaHCO 3 /CO 2 buffer (50 mM/25 mM/5%, pH 7.4, 37°C) in 35-mm dishes.
Nitrosation by NaNO 2 under Acidic Conditions-Substances were exposed to equimolar concentrations of sodium nitrite in H 2 O equilibrated with glacial acetic acid to pH 3.5.
15 N NMR Identification of Nitrosated Products-Immediately after nitrosation, sample aliquots were supplemented with 10% D 2 O and analyzed by 15 N NMR spectrometry. The adducts were identified by 50.67 MHz 15 N NMR spectrometry on a Bruker AVANCE DRX 500 instrument. Chemical shifts (␦) are given in ppm relative to neat nitromethane (␦ ϭ 0) as the external standard.
Decay Kinetic Experiments-The decay kinetics of N-acetyl-N-nitrosotryptophan (100 M) at various temperatures (15-45°C) were determined in air-tight quartz cuvettes by rapid scan monitoring the UVvisible absorption at max ϭ 335 nm (⑀ 335 ϭ 6100 M Ϫ1 cm Ϫ1 ) (24). Between recordings the samples were protected from light. The temperature was maintained at Ϯ 0.1°C.

Formation of N-Nitrosotryptophan in Phosphate
Buffer, pH 7.4 -Evidence exists that under acidic conditions similar to those in the stomach the NaNO 2 -dependent nitrosation of human serum albumin, bovine serum albumin, or the dipeptide glycine-tryptophan yields N-nitrosotryptophan (21,23). The main nitrosating species in this reaction is postulated to be dinitrogen trioxide (N 2 O 3 ) formed from HNO 2 dehydration (7,29). Provided that this is indeed the case, N 2 O 3 formed from ⅐ NO in the presence of oxygen at pH 7.4 should also nitrosate N-terminal-blocked tryptophan derivatives, like N-acetyltryptophan and peptide-associated tryptophan (lysine-tryptophan-lysine).
To verify the nitrosation of N-blocked tryptophan derivatives by N 2 O 3 at pH 7.4, we selected various N 2 O 3 -generating systems. We employed preformed nitric oxide as well as in situ generation of ⅐ NO from S-nitrosocysteine and spermine NONOate, respectively. In these systems, N 2 O 3 is formed from oxidation of nitric oxide ( ⅐ NO autoxidation) according to Equations 1 and 2 (see the Introduction). Because the formation of N-acetyl-N-nitrosotryptophan from the reaction of N-acetyltryptophan with NaNO 2 at pH 3.5 has been described in detail (24,30), we used this reaction as a reference system. In analogy to the data of Bonnett and Holleyhead (24) we recorded the UV-visible spectrum of N-acetyl-N-nitrosotryptophan with an absorption maximum at 335 nm (Fig. 1A, trace b). Interestingly, an almost identical UV-visible spectrum was observed when N-acetyltryptophan was allowed to react with preformed nitric oxide in the presence of oxygen at physiological pH 7.4 ( Fig. 1A, trace e). This finding indicated very strongly that N 2 O 3 can indeed nitrosate the secondary amine of the indole system in tryptophan. Likewise, the UV-visible spectrum of N-acetyl-N-nitrosotryptophan was also detected during reaction of N-acetyltryptophan with the above mentioned nitric oxide-donating compounds in the presence of air (Fig. 1A, traces c and d).
Because in vivo N-terminal-blocked tryptophan is present primarily in proteins, the N 2 O 3 -mediated nitrosation of the nitrogen atom of the indole ring was also investigated for peptide-bound tryptophan by employing lysine-tryptophan-lysine. In full agreement with the experiments performed with N-acetyltryptophan (see above), all applied N 2 O 3 -generating systems were able to nitrosate peptide-bound tryptophan effectively at the selected pH values (Fig. 1B). In contrast, the nitrosation of L-tryptophan by the N 2 O 3 -generating systems was much less effective (data not shown), which is in line with data from the literature (28,30). Conclusively, only N-terminal-blocked tryptophan but not L-tryptophan is a relevant target for N 2 O 3 at physiological pH.
Detection of N-Nitrosotryptophan by 15 N NMR Spectrometry-Because UV-visible absorption spectra cannot provide unambiguous evidence that the secondary amine function of the indole ring was truly nitrosated, we used 15 N NMR spectrometry as a more reliable analytical tool. In 1986, Dorie et al. (31) recorded the 15  ences in the recording conditions, e.g. different pH values as well as improvements of the 15 N NMR spectrometric techniques during the past years. Nevertheless, nitrosation of the nitrogen atom at the indole ring is proven by this characteristic 15 N NMR spectrum. When N-acetyltryptophan was reacted with authentic 15 NO in the presence of oxygen, only the two new 15 N NMR resonances at 180.1 and 165.6 ppm were detected, consistent with the formation of N-acetyl-N-nitrosotryptophan (Fig. 2B). To the best of our knowledge, this is the first direct proof that a tryptophan derivative can be nitrosated by N 2 O 3 at the nitrogen atom of the indole system at physiological pH.
Analogous to N-acetyltryptophan the tripeptide lysine-tryptophan-lysine was nitrosated at pH 7.4 by preformed 15 NO as proven by the 15 N NMR resonance lines at 181.1 and 166.7 ppm, respectively (Fig. 2C).
Reactivity of N-blocked Tryptophan Derivatives Toward N 2 O 3 -For a discussion of the putative significance of tryptophan nitrosation in vivo it is necessary to determine the rate constants of the reaction of N-acetyltryptophan or its deriva-tives with N 2 O 3 at pH 7.5. As mentioned above, Blanchard et al. (19) reported that melatonin reacts with ⅐ NO in the presence of oxygen at pH 7.4 with a second-order rate constant of only 0.5 M Ϫ1 s Ϫ1 . However, these experiments were performed in the presence of 200 mM Hepes buffer, which is known to react with reactive nitrogen/oxygen species to yield both H 2 O 2 (33, 34) and a ⅐ NO-donating compound (35). Therefore we evaluated the rate constants of N-nitrosation of N-blocked tryptophan derivatives with N 2 O 3 by a competition method employing morpholine as the N 2 O 3 scavenger (2,36,37). The reaction of 2 mM N-acetyltryptophan with 0.5 mM MAMA NONOate yielded about 360 M N-acetyl-N-nitrosotryptophan. Morpholine competitively inhibited the MAMA NONOate-induced nitrosation of N-acetyltryptophan in a nonlinear manner with an IC 50 value of 13.4 mM (Fig. 3). The protonated form of morpholine, that is morpholinium, with a pK a value of 8.23 at 37°C (38), is not expected to react with N 2 O 3 (39). Thus, the fraction of unprotonated morpholine is about 15.7% at pH 7.5. Taking this fraction into account, the true IC 50 value is given by 13.4 mM ϫ 0.157 ϭ 2.1 mM. Conclusively, as 2 mM N-acetyltryptophan was used (see above), N 2 O 3 reacts at 37°C with virtually identical rate constant with both N-acetyltryptophan and morpholine. To verify the presumption that only the unprotonated fraction of the secondary amine effectively reacts with N 2 O 3 , a control experiment was performed with piperazine as a competitive N 2 O 3 scavenger. At pH 7.5 about 99% of piperazine (pK a ϭ 5.55 (7)) exists in the unprotonated form; thus, as expected, piperazine was more effective than morpholine in inhibiting the MAMA NONOate-induced nitrosation of N-acetyltryptophan (Fig. 3). From inspection of Fig. 3 it can be deduced that MAMA NONOate-induced nitrosation of N-acetyltryptophan was halfmaximally inhibited at a piperazine concentration of 2.7 mM. The observation that the corrected IC 50 value of morpholine is lower than the experimental IC 50 value of piperazine (Table I) is in agreement with the report that N 2 O 3 reacts faster with morpholine than with piperazine (7). Not unexpectedly, the reactivity of other N-terminal-blocked tryptophan derivatives toward N 2 O 3 was found to be similar (Table I). To compare the rate constant of these nitrosation reactions with other nitrosation reactions reported in the literature, one experiment with N-acetyltryptophan was performed at 25°C and a rate constant of 4.4 ϫ 10 7 M Ϫ1 s Ϫ1 was obtained. Thus, N-blocked tryptophan reacts rather fast with N 2 O 3 (see the Discussion).
Decomposition of N-Acetyl-N-nitrosotryptophan-In contrast to common, stable nitrosamines, protein-bound N-nitrosotryptophan has been reported to undergo slow decay at pH 2 (23). To quantify this capability at pH 7.4, we analyzed the decomposition of N-acetyl-N-nitrosotryptophan by monitoring its absorption at 335 nm. At 37°C N-acetyl-N-nitrosotryptophan decayed in a first-order manner (Fig. 4A). Similarly, first-order decay kinetics were also observed at other temperatures (15-45°C, data not shown). From the excellent Arrhenius plot of the rate data (Fig. 4B) an activation barrier of E a ϭ 13.2 Ϯ 0.1 kcal mol Ϫ1 and an A-factor of 1.7 Ϯ 0.3 ϫ 10 5 s Ϫ1 were extracted. This A-factor appears to be extremely low for a simple first-order (homolysis) decomposition. From the Arrhenius parameters the half-life of N-acetyl-N-nitrosotryptophan at physiological conditions (T ϭ 37°C, pH 7.4) can be calculated to 140 min. Thus, the N-nitrosamines of N-blocked tryptophan derivatives are rather long-lived intermediates.
Because the half-life of peptide-bound N-nitrosotryptophan was found to be similar to that of N-acetyl-N-nitrosotryptophan (data not shown), one may ask whether it would make any sense for physiological functions to nitrosate tryptophan in vivo. Recently, Harohalli et al. (23) reported that N-nitrosotryptophan very slowly releases ⅐ NO at pH 2, and Blanchard-Fillion et al. (40) observed that nitrosomelatonin spontaneously decays at pH 7.4, thereby releasing nitric oxide at a yield of 71%. In our hands, however, authentic nitrosomelatonin released only negligible amounts of nitric oxide on decomposition in Hepes-free buffer solution (data not shown). Noticeably neither Harohalli et al. (23) nor Blanchard-Fillion et al. (40) directly detected ⅐ NO, e.g. with a NO electrode.
As nitrosothiols release nitric oxide in the presence of vitamin C (10), we compared the potential of ascorbate to induce ⅐ NO release from S-nitrosoglutathione versus N-acetyl-N-nitrosotryptophan (Fig. 5). In the absence of copper ions, the release of ⅐ NO from N-acetyl-N-nitrosotryptophan after the addition of ascorbate was about 5-fold higher than from S-nitrosoglutathione. In contrast, a "simple" N-nitrosamine like N-nitrosomorpholine did not liberate nitric oxide under similar condi-  tions (data not shown). The low ⅐ NO-releasing efficiency of S-nitrosoglutathione is explained by the rigorous depletion of copper ions in the applied buffer solution. In the presence of 10 M Cu 2ϩ , however, S-nitrosoglutathione as well as N-acetyl-Nnitrosotryptophan did release nitric oxide with nearly the same yield in the presence of ascorbate (data not shown). In the absence of vitamin C only negligible amounts of nitric oxide were released from either N-acetyl-N-nitrosotryptophan or S-nitrosoglutathione on decomposition. These results demonstrate unequivocally that N-acetyl-N-nitrosotryptophan has nitric oxide-releasing capabilities similar to those of S-nitrosocysteine.

DISCUSSION
The results presented above clearly show that N 2 O 3 nitrosates the secondary amine function at the indole ring of Nblocked tryptophan with high reactivity at physiological pH values. It is known that indole and indole derivatives are easily attacked by N 2 O 3 under acidic conditions, however, with exclusive C-nitrosation, thus yielding 3-nitroso products (30,41). Such products were not observed for tryptophan because here the C-3 position is blocked by the alanine residue, rendering nitrosation at position N-1 (N indole ) more feasible. Nitrosation at C-2 is possible only for derivatives carrying powerful electron donors at the C-3 position (30). Our data indicate that N-acetyltryptophan, melatonin, and the tripeptide lysine-tryptophan-lysine are nitrosated directly by N 2 O 3 because these nitrosation reactions could be inhibited effectively by the N 2 O 3 scavengers morpholine and piperazine, respectively. In contrast, Turjanski et al. (42) reported that melatonin is nitrosated mainly by a combined attack of the radicals ⅐ NO 2 and ⅐ NO. It should be noted that such a mechanism has never been proven for ordinary amines and that a variety of side products, i.e. N-nitro, C-nitro, and C-nitroso compounds, had to be produced by an operating radical mecha-nism. Noticeably, the 15 N NMR data demonstrated that reaction of 15 N 2 O 3 with N-blocked tryptophan exclusively yields 15 N-nitrosotryptophan. Recently, the reaction between albumin and N 2 O 3 generated from NaNO 2 at low pH has been monitored by UV spectroscopy. On the basis of these measurements it was assumed that tryptophan would be nitrosated also (21,23). However, this conclusion has not been generally accepted (22,43). As there is presently no specific test for N-nitrosotryptophan available and because N-nitrosotryptophan derivatives and S-nitrosocysteine have almost identical UV-visible absorption spectra ( max ϳ335 (24) and ϳ340 nm (30), respectively), the efficiency of nitrosation of both tryptophan and cysteine in proteins is hard to verify experimentally by UV-visible spectroscopy. On the other hand, the effectiveness of these reactions can reasonably be estimated on the basis of the experimental rate constants and the concentrations of both amino acids in proteins. At physiological pH values, tryptophan exists practically exclusively in the reactive nonprotonated form. Hence, the rate constant of N 2 O 3 with protein-bound tryptophan can be assumed to be k ϭ 4.4 ϫ 10 7 M Ϫ1 s Ϫ1 (see "Results"). In contrast, thiols are only marginally deprotonated at pH 7.4. From the average pK a values of ϳ8.2 of thiolates in proteins (22), it can be deduced that only 13.5% of the cysteine residues are viable targets for N 2 O 3 . Thus, as the rate constants for reaction of thiolates (RS Ϫ ) with N 2 O 3 is about k ϭ 2 ϫ 10 8 M Ϫ1 s Ϫ1 (2), protein-bound cysteine should react at pH 7.4 with N 2 O 3 with a rate constant of k (cysteine ϩ N 2 O 3 ) ϭ 2 ϫ 10 8 M Ϫ1 s Ϫ1 ϫ 0.135 ϭ 2.7 ϫ 10 7 M Ϫ1 s Ϫ1 . Thus, the rate constant of the reaction of protein-bound tryptophan with N 2 O 3 is, somewhat unexpectedly, estimated to be about 63% higher than the rate constant of protein-bound cysteine with N 2 O 3 . To verify that this conclusion can be extended to the reaction rate, the amounts of both tryptophan and cysteine residues in proteins were verified by searching the RCSB Protein Data Bank (44) for those proteins that are believed to be S-nitrosated (Table II). From the data in Table II it can be deduced that in the selected proteins the total amount of cysteine is higher than the total amount of tryptophan. With regard to the differences in the rate constants (see above), one can now estimate that N 2 O 3 reacts primarily (40 -90%) with tryptophan. Thus, we hypothesized that protein-bound tryptophan should be preferentially nitrosated under physiological conditions. To verify this prediction, we are currently developing highly sensitive protocols for the detection of both N-nitrosotryptophan and S-nitrosothiols in proteins.
In this paper we have demonstrated additionally that Nnitrosotryptophan has the capability of releasing nitric oxide in the presence of ascorbate. This implies that N-nitrosotryptophan might operate as a nitric oxide carrier, a property that to date has been attributed more or less exclusively to S-nitrosothiols (7,11,45,46). However, as the chemistry of N-nitrosotryptophan derivatives in physiological environment is not very well developed, it is as yet too early to consider N-nitrosotryptophan derivatives as "harmless" compounds in biological sys-  tems. It might be assumed that potentially harmful reactive nitrogen species like peroxynitrite (40) or N 2 O 3 are deactivated by their reaction with tryptophan residues and that this may them give some antioxidative functions. However, the observation of Venitt et al. (26) that N-acetyl-N-nitrosotryptophan at 5-15 mM induces mutagenicity in bacteria is, in our view, easily explained by its capability of releasing nitric oxide, which is known to induce such effects at unphysiologically high concentrations (47). In order not to be misunderstood, at present harmful effects of N-nitrosotryptophan derivatives cannot be ruled out. It should be remembered that harmful reactions are also known for S-nitrosothiols. For example, it has been reported that S-nitrosothiols react with thiols to yield nitroxyl (48), which generates hydroxyl radicals and/or peroxynitrite in the absence and presence of oxygen, respectively (49). Because N-nitrosotryptophan derivatives are rather longlived and yet not "indefinitely" stable compounds (the hydrolysis of N-acetyl-N-nitrosotryptophan is expected to yield the harmless products tryptophan and nitrite (28)), one may speculate that this could further decrease the putative mutagenic potential of N-nitrosotryptophan. On the other hand, the observed half-life (t1 ⁄2 Ͼ 2 h) is so long that N-nitrosotryptophan derivatives (N indole NO) may participate in physiological processes, e.g. in the transport and release of nitric oxide. In fact, Zhang et al. (21) observed that peptide-bound N-nitrosotryptophan induces vasorelaxation of rabbit aortic rings, a function that is typical for freely diffusing nitric oxide (45). In conclusion, we hypothesize that a putative physiological potential of N-nitrosotryptophan should be more important than its pathophysiological one. This feature remains to be clarified in the near future. In any case, we have identified, in addition to cysteine, a second major target for N 2 O 3 in proteins. This fact will strongly influence the general understanding of protein nitrosation.