Reactive oxygen species induced structural alterations of substance P

Substance P (SP1-11) was exposed to a continuous flux of superoxide ( O 2 - ) or hydroxyl radicals (.OH) in a hypoxanthine (HX)/xanthine oxidase (86 mU) system in the presence of 1 mM deferoxamine and 40 mM D-mannitol or 50 μM FeCI3. 6H2O and 50 μM EDTA, respectively. O 2 - caused fragmentation between the Phe7 and Phe8, whereas .OH induced cleavage also between the Phe8 and Gly9. Reactive oxygen species H2O2 and HCIO did not cause fragmentation, but modification of the amino acid side chains and/or aggregation with altered hydrophobicity in reverse phase high performance liquid chromatography compared to native SP1-11. Furthermore, exposure of SP1-11 to phorbol myristate acetate preactivated neutrophils resuited in products similar to those observed upon exposure to superoxide or hydroxyl radicals in a cell-free HX/xanthine oxidase system. This study suggests that, in contrast to rigid proteins, fragmentation is relatively easily induced in a small peptide like SP1-11, perhaps due to strain on the peptide and t-carbon bonds caused by the movable, random coil configuration acquired by SP1-11 in an aqueous solution. Oxidative modification might modulate paracrine actions of SP1-11 at site of inflammation.


Introduction
Substance P (SPl-11) is an undecapeptide, which consists of Arg-Pro-Lys-Pro-Gln-Gln-Phe-Phe-Gly-Leu-Met-NH2 .1 SP1-11 has been localized to small unmyelinated, slowly conducting C-type polymodal nociceptors. 2 In addition to its possible role in pain transmission in the dorsal horn of the spinal cord, 3 it has been implicated in neurogenic inflammation. 4 Local release by axon reflex as a result of antedromic conduction causes wheal and flare, 4 but various paracrine effects such as synthesis and secretion of type I matrix metalloproteinase (MMP-1 or' fibroblast-type' interstitial collagenase) and prostaglandin E 2 (PGE2) by synoviocytes, and of interleukin-lfl and tumour necrosis factor-o by monocytes, have also been described. 6 In addition to its role in vasoregulation and cytokine and proteinase secretion, substance P can activate polymorphonuclear neutrophils (PMN) and monocyte/macrophages to produce superoxide (O-) via the NADPH oxidase pathway, v'8 In many inflammatory diseases hypoxia-reperfusion syndrome is another and possibly more important source of oxygen derived free radicals (ODFR). 9 During ischaemia xanthine dehydrogenase is proteolytically cleaved to xanthine oxidase by an enzyme activated by Ca 2+ efflux from mitochondria. At the same (C) 1992 Rapid Communications of Oxford ktd time, chemical energy from adenosine triphosphate (ATP) is utilized and hypoxanthine (HX) is produced. During reperfusion xanthine oxidase will catalyse the conversion of HX to xanthine and further to uric acid (Equation 1, see below), with bimolecular oxygen acting as an electron acceptor in both reactions. It seems, therefore, that released at a site of inflammation, will be exposed not only to degradative enzymes but also to various reactive oxygen species. This prompted us to study the possible effect of such compounds on substance P in vitro.

Materials and Methods
Preparation of synthetic substance P: Synthetic substance P was first purchased from Sigma Chemical Company (St Louis, MO, USA) or Cambridge Research Biochemicals (Cambridge, UK). Because relatively large amounts were needed, substance P was also synthesized according to the solid-phase method with Applied Biosystems (Foster City, CA, USA) 430A peptide synthesizer using tBoc chemistry and p-methyl-BHA-resin.
Mediators of Inflammation. Vol 1992 355 Production of oxygen-derived free radicals" Synthetic soluble substance P (1 mg/ml) in RPMI-1640 (Gibco) was exposed to Oor hydroxyl radical ('OH) produced in a hypoxanthine/xanthine oxidase (EC 1.2.3. 22) system in the presence of 1 mM deferoxamine (Desferal(R), Ciba Pharmaceutical Co) and 40 mM t)-mannitol or 50/.tM FeC13"6HiO (Mallinckrodt, Paris, KY, USA) and 50 #M ethylenediaminetetraacetic acid (EDTA) (Sigma), respectively, as described in detail elsewhere. 1'11 HX was added in excess (up to a 10mM final concentration). Deferoxamine chelates iron by occupying all of its four coordination sites and iron is thus altered to a catalytically inactive form, 12 whereas mannitol is an effective hydroxyl radical scavenger. 13 In the presence of these reagents, alterations observed are caused by O-, although H202 is also formed in a so-called one-electron auto-oxidation or dismutation (Equation 2). Iron was added as iron(III) ions, so that the reaction would start immediately after addition of xanthine oxidase" in this case, iron(III) ions are first reduced to iron(II) ions (Equation 3). Iron(I1) ions can then act as an electron donor and catalyse the Fenton reaction, in which H202 is converted to hydroxyl ion and hydroxyl radical (Equation 4). The combination of Equations 3 and 4 is the so-called Haber-Weiss reaction (Equation 5). When "OH was produced, EDTA was added in addition to iron(III) ions to increase the effective concentration of iron, which is difficult to dissolve in water based buffers; iron complexed with EDTA is redox reactive. All reactions were started by adding xanthine oxidase, performed under constant stirring at 22C and stopped by addition of 10/lg/ml superoxide dismutase (EC 1.15.1.1) and catalase (EC 1.11.1.6). In addition to ODFR, SP>ll was also exposed to reactive oxygen species H202 (1 and 0.1 mM) or HClO (added as calcium salt, 1 and 0.1 mM) The HC10 reaction was stopped by L-methionine as scavenger, 4 which at the end of the incubation was added to 10mM final concentration.
(1) with 50 #g/ml of phorbol 12-myristate 13-acetate (PMA; Sigma) at 37C for 30min. 5 Even though the PMA concentration we use 5 is rather high, it does not cause cell lysis so that neutrophil viability was > 98% as assessed by trypan blue exclusion. After PMA preactivation, the cells were separated by spinning the neutrophils at 1000 x g for 5 min and resuspended in 1 ml RPMI-1640. 1 mg of substance P was added to each tube and the samples were incubated at + 37C for the time periods indicated and 100 #1 aliquots were taken for analysis. Samples were spun at 1000 x g for 5 min and 50/ll of the supernatent was frozen immediately to be analysed later by HPLC.
High performance liquid chromatography: Separation of synthetic substance P exposed to different reactive oxygen species as described above was done by high performance liquid chromatography (HPLC). Samples were loaded onto a Nova-Pak column (silica, 3.9mm x 15cm, pore diameter 600nm) protected by Guard-Pak TM precolumn module and Bondapak C18 inserts using a WISP TM model 710 B sample processor. After a 2 min loading period a 60 min elution was done with a linear 1 ml/min gradient (0-48%) with 50mM NaH2PO4/H3PO4, pH 3  between the two phenylalanine residues at positions 7 and 8. Exposure to hydroxyl radicals produced similar fragments, but, in addition, two minor oxidation products with retention times 31.17 min and 18.33 min were produced ( Figure 1). One of these had a retention time similar to SP9-11 (18.33 min) and although SP>8 was not available for chromatographic analysis, these two minor hydroxyl radical oxidation products could, by deduction, represent an additional cleavage site between phenylalanine and glycine residues at positions 8 and 9.
If instead of ODFR, SP>I was exposed to reactive oxygen species RiO2 (1 and 0.1 mM) or HC10 (added as calcium salt, 1 and 0.1 mM), fragmentation was not observed (only one reaction product was produced), but rather modification of the amino acid s:,de chains and/or aggregation: H202 and HC10 induced the formation of oxidation products with retention times 33.13min and 32.04 min, respectively. Oxidative modification of synthetic soluble substance P in aqueous solution was timedependent. Fragmentation caused by superoxide and hydroxyl radical seemed to have a somewhat similar temporal profile (Figure 2). Oxidative modification induced by hydrogen peroxide was a relatively slow process (32.04 min product) compared to the almost immediate effect seen when 1   hypochlorous acid was used (33.13 min product) Figure 3). Synthetic SP_ in aqueous solution was rapidly modified by PMA preactivated neutrophils. Peaks with retention times corresponding to those of SPs-11, SP>7 and SP9_ appeared in the supernatant, whereas the peak corresponding to the minor SP>s peak produced by "OH in the HX/xanthine oxidase system was negligible (one representative experiment is shown in Table 1). In addition, peaks eluting at approximately 38.7 min, 40.2 min and 41.3min were also found in the substance P-neutrophil supematant aliquots (not shown).

Discussion
These results obtained using cell free conditions show that SPa_11 can be fragmented by ODFR, both Oand "OH. Oxidative modifications of various free amino acids and proteins have received a lot of attention, 6 but observations on the effect of ODFR on substance P have not been published. The effect of ODFR on various amino acids cannot be extrapolated to proteins or peptides due to their often complex secondary and tertiary structure, which will affect the sites accessible for the initial attack and the subsequent secondary effects such as intramolecular charge transfer reactions. Proteins, perhaps due to the usually relatively rigid structure possibly enforced by intra-and interchain disulphide bonds, are usually not fragmented by ODFR, although such an exposure may make them susceptible or mark them for a subsequent proteolytic attack. 7 According to our finding, however, a small peptide, substance P, was fragmented by both Oand "OH. Substance P in aqueous phase acquires a mobile random coil configuration, 18 with nonregular, nonrepeating dihedral angles of the peptide backbone with or without equilibrium between various random coil conformers. This may strain the peptide or o-carbon bonds and may contribute to their cleavage. Some amino acids are more susceptible to oxidative damage than others. 9 Phenylalanine containing an aromatic benzene ring is easily oxidized. Exposure of phenylalanine to "OH will cause hydroxylation to the p-, ra-and in particular 0-position. Two such radicals can join together to give a dimer, that can lose water to form biphenyl. 2 The close apposition of phenylalanine residues at positions 7 and 8 in a flexible random coil SP1-11 might favour the formation of such biphenyls, which could cause a sudden change, a nick, in the stereoconfiguration. That nonenzymatic, conformational changes can cleave even peptide bonds is suggested by the action of catalytic antibodies. 2 An antibody to vasoactive intestinal peptide (VIP) can cause cleavage of a peptide bond in full-length VIP>2s. Interestingly, and against expectations, the scissile bond was not part of the antigenic determinant involved in the high affinity antigen-antibody binding. Therefore, the relatively strong interactive forces involved in binding of the catalytic antibody to VIP>28 are unlikely to be directly responsible for the peptide bond hydrolysis. Competitive binding studies using various synthetic VIP peptide fragments were used to map the antibody binding epitope to amino acid residues 22-28. This binding was able to cause a cleavage at a distant scissile Glu16--Met 17 bond, perhaps via conformational changes induced by binding of the catalytic antibody.
Neuropeptide peptide bonds are cleaved by various well described exo-and endopeptidases. Actually, ODFR usually lead to formation of a-carbon centred radicals and, in the presence of oxygen, peroxyl radicals, which decompose to fragment the polypeptide chain at the 0-carbon rather than at peptide bond. 4 Accordingly, the cleavage sites observed in the present study, i.e. between the phenylalanine residues at positions 7 and 8 and, with "OH also between the phenylalanine and glycine at positions 8 and 9, may be located at the 0-carbon rather than the peptide bond. Obviously, the hypothesis about the fragments of substance P generated by oxygen radical exposure has to be validated in future studies with appropriate nuclear magnetic resonance and mass spectroscopy studies allowing the identification of the structure of the generated metabolites.
In contrast to Oand "OH, H202 is, by definition, not a radical: it does not contain unpaired electrons i.e. electrons occupying an atomic or molecular orbital by itself. It is not particularly reactive either, compared to ODFR. On the other hand, H202, like water, as a small, uncharged polar molecule, has a very high Reactiue oxygen species and sbstance P permeability coefficient compared to e.g. charged O-. It can therefore, pass through lipid biomembranes. It can also, according to Equation 4 (see above), lead to the formation of the highly reactive hydroxyl radical in the presence of transition metal ions like iron and copper. In the present study, when H202 was added to an aqueous solution, it did not cause fragmentation of SP_I, but instead a slowly progressive structural modification, which was reflected in altered hydrophobicity of the reaction product in reverse phase HPLC. In contrast, HC10 caused a rapid modification of SP>, which again was reflected in altered hydrophobicity. It was noteworthy that the HC10-induced product had a different retention time than the H202-induced oxidation product.
Although not studied in structural detail, it is likely that HC10 caused conversion of the methionine residue at position 11 to a corresponding methionine sulphoxide (MetSO) or perhaps even to a methionine sulphone (MetSO2). 22 Also this reaction pathway would seem to be of potential relevance in vivo in inflammation, because under such circumstances HC10 is formed from H202 by myeloperoxidase in the presence of chloride ion.
Chloride ion is one of the most common ions in the extracellular tissue fluid and myeloperoxidase is stored in and, upon activation, released from the primary or azurophil granules of the PMN. 23 PMA is a direct activator of the protein kinase C and will, by phosphorylation to serine and threonine residues of some as yet mostly unknown target proteins, activate the cell membrane NADPH oxidase to produce O-. This will spontaneously dismutate to H202 at a rate of 2 x 10 -5 Ms -. In the presence of trace amounts of iron, always present in 'average' reaction mixtures without added metal, usually at about 1/IM as assessed by atomic absorption analysis, 2 hydroxyl radicals also are likely to be formed. The interpretation of PMA preactivated neutrophil experiments is complicated by cellular uptake and by simultaneous activation of neutrophil-mediated exocytosis as was suggested by the presence of peaks other than those produced in the cell free HX/xanthine oxidase system. Neutrophils contain in their primary granules cathepsin G, which, according to the bond specificity is able to cleave substance P. In addition, neutrophils contain an integral membrane protein known as neutral endopeptidase EC 3.4.24.11, also known as the common acute lymphoblastic leukaemia associated antigen CALLA and as enkephalinase, which is also able to cleave substance p.24,25 Therefore, even if untreated neutrophils would cause substance P degradation or superoxide scavenger superoxide dismutase would not be able to inhibit substance P degradation, it is not possible to exclude the role of oxidative degradation because there are alternative and complementary degradative pathways. However, these results suggest that substance P exposed to PMA preactivated neutrophils is modified as if exposed to superoxide or hydroxyl radicals in a cell free HX/xanthine oxidase system in vitro.
It therefore seems that various ODFR and reactive oxygen species can cause oxidant specific and time-dependent modification of synthetic SP1-11 in aqueous phase by inducing fragmentation, modification of side chains and aggregation. According to the present findings, it seems likely that oxidative modification also has to be taken into consideration when the paracrine actions of SP1-11 in inflammatory diseases and processes are regarded: the distance from the axon terminal to the potential site of action may at the paracrine site of action be more than 10 000 longer than in the more concealed synaptic spaces.