OH-Radical Oxidation of Lung Surfactant Protein B on Aqueous Surfaces

Air pollutants generate reactive oxygen species on lung surfaces. Here we report how hydroxyl radicals (·OH) injected on the surface of water react with SP-B1–25, a 25-residue polypeptide surrogate of human lung surfactant protein B. Our experiments consist of intersecting microjets of aqueous SP-B1–25 solutions with O3/O2/H2O/N2(g) gas streams that are photolyzed into ·OH(g) in situ by 266 nm laser nanosecond pulses. Surface-sensitive mass spectrometry enables us to monitor the prompt (<10 μs) and simultaneous formation of primary On-containing products/intermediates (n≤5) triggered by the reaction of ·OH with interfacial SP-B1–25. We found that O-atoms from both O3 and ·OH are incorporated into the reactive cysteine Cys8 and Cys11 and tryptophan Trp9 components of the hydrophobic N-terminus of SP-B1–25 that lies at the topmost layers of the air–liquid interface. Remarkably, these processes are initiated by ·OH additions rather than by H-atom abstractions from S–H, C–H, or N–H groups. By increasing the hydrophilicity of the N-terminus region of SP-B1–25, these transformations will impair its role as a surfactant.


INTRODUCTION
Human lung alveoli are covered by a 0.1-0.5 µm thick epithelium lining uid (ELF) that has a large surface area ∼885,000 cm 2 (vs. the ∼4,500 cm 2 of the airway). 1) ELF contains surfactant proteins B and C (∼10 wt% of total ELF surfactant material), [1][2][3][4][5][6] in addition to water-soluble antioxidants such as glutathione ([GSH] ∼100-500 µM), ascorbic acid ([Asc] ∼100 µM), 1,7) uric acid, and liposoluble α-tocopherol. [7][8][9][10][11] SP-B functions reduce the energy required to expand lungs during aspiration and prevent their collapse upon exhalation. 12) A recent simulation study revealed that SP-B induces the formation of bilayer reservoirs by monolayer folding, helps attaching the disconnected bilayer aggregates to a monolayer at the air-water interface, and facilitates lipid transfer between these structures. 13) SP-B de ciency is known to cause severe lung dysfunction and in ammation. 14) Human infants genetically lacking SP-B cannot survive. 15) It is apparent that the oxidative degradation of SP-B's surface activity should induce acute syndromes. 2,[16][17][18][19] Recent studies suggested that the inhalation of particulate matter (PM) produces ·OH in ELF. [20][21][22] In vivo studies have shown that ·OH produced in ELF exposed to O 3 (g) enhances bronchoalveolar permeability. 23) Given the abundance of antioxidants in ELF (see above), SP-B oxidative damage can be triggered only by ·OH produced or deposited close to SP-B. 24) Because inhaled gas-phase oxidants are very reactive and the SP-B surfactant is present at the ELF surface, the relevant reactive events are expected to take place at the air-water interface rather than in the bulk liquid. A previous study found that the heterogeneous reaction of O 3 (g) with SP-B 1-25 (a model 25-residue polypeptide of human SP-B) 25) in water-methanol droplets yields several products that incorporate O-atoms in a few seconds. 3) e products from the droplets exposed to O 3 (g) or the solution treated O 3 -bubbling were found to be quite di erent, 3) implying that the mechanism at the gas-liquid interface is di erent from that in bulk solution. 7) e authors also generated reactive intermediates in bulk water with Fenton's reagent (Fe 2+ +H 2 O 2 ) to test whether the chemistry actually involved ·OH radicals. However, we recently found that the Fenton reaction at the air-water interface produces oxo-ferryl species (Fe IV =O) in high yields, which behave as O-atom donors rather than as H-abstracting species. [26][27][28][29][30] In fact, a recent study suggested that Fenton's reaction produces <10% ·OH in bulk water at physiological pH 6-7. 28) Since the production of Fe IV =O is signi cantly enhanced in Fenton's reaction at the air-water interface, 26) establishing the mechanism of the oxidation of SP-B 1-25 initiated by ·OH at the air-water interface requires experiments involving a direct, bona de source of ·OH in situ.
In this paper we report a mass-spectrometric study speci cally designed to address these issues, in which aqueous SP-B   (a 25-residue polypeptide NH 2 -1 FPIPLPYCWLCRALIKRIQAMIPK 25 G-COOH) reacts with gas-phase ·OH at the air-water interface. SP-B 1-25 has been used as a representative model peptide that mimics SP-B functions, because of the similarity of their chemical and physical properties. 4,6) Our experiments were conducted in a novel setup in which aqueous SP-B 1-25 microjets were intersected with gas-phase ·OH streams generated in close proximity to the surface via pulsed laser photolysis of O 3 (g)/H 2 O(g)/O 2 (g)/N 2 (g) mixtures, [31][32][33][34][35][36] while continuously monitoring the composition of the interfacial layers of the microjets via pneumatic ionization mass spectrometry. [37][38][39] EXPERIMENTAL SECTION e present experimental setup is essentially the same as the one reported elsewhere. 31) e charged product species generated on the surface of SP-B 1-25 (aq) microjets during τ ∼10 µs contact times (τ is the lifetime of the microjets before being pneumatically nebulized) with O 3 (g) or ·OH(g) streams are monitored in-situ by online mass spectrometry (Fig. S1, Agilent 6130 Quadrupole LC/MS Electrospray System). 31 . ese conditions ensure that the detected species correspond to the truly initial stages of reaction. [34][35][36] e all-12 C isotopomer of SP-B 1-25 ≡Phe 1 -Pro 2 -Ile 3 -Pro 4 -Leu 5 -Pro 6 -Ty r 7 -Cys 8 -Tr p 9 -Leu 10 -Cys 11 -Arg 12 -A la 13 -Leu 14 -Ile 15 -Lys 16 -Arg 17 -Ile 18 -Gln 19 -Ala 20 -Met 21 -Ile 22 -Pro 23 -Lys 24 -Gly 25 , has a molecular mass of 2928 Da. Samples are injected at 100 µL min −1 into the spraying chamber of the mass spectrometer through a grounded stainless steel needle (100 µm bore) coaxial with a sheath issuing nebulizer N 2 (g) at a high velocity (v g ≈160 m/s). 38) e surface speci city of our experiments has been demonstrated previously. 26,37,38,40,41) It should be emphasized that the products we observe are formed when gaseous reactants collide with the intact aqueous jets as they emerge from the nozzle, i.e., before jets are broken up into sub-micron charged droplets by the nebulizer gas. 26) Since 266 nm pulses ash every 100 ms, and the microjets breaks up within ∼10 µs a er being ejected from the nozzle, it is safe to assume that the observed phenomena always take place on the surface of fresh solutions. 31) We veri ed that in the absence of O 3 (g) neither reactant signals are a ected nor new product signals appear upon 266 nm pulsed irradiation, thereby excluding the photolysis SP-B 1-25 (aq) as the source of detected products (see Fig. S2). In our experiments, ⋅OH(g) will stick to the surface of water in nearly every collision, 31,[33][34][35][42][43][44][45][46] and react with SP-B 1-25 (R1), or dimerize into H 2 O 2 (R2) at the air-water interface.
(R2) e preference of ·OH for interfacial layers relative to the bulk liquid is supported both by theoretical and experimen-tal studies. 31,32,34,[43][44][45][46] is e ect should enhance the probability of ·OH reactions with hydrophobic residues at the air-water interface before it di uses into bulk phase. Figure 1A shows a positive ion mass spectrum of 43 µM SP-B 1-25 (aq) microjets exposed to O 2 (g)/H 2 O(g)/N 2 (g), and to O 3 (g)/O 2 (g)/H 2 O(g)/N 2 (g) with the 266 nm laser on and o . e three major peaks at m/z=586.7, 733.0 and 977.0 correspond to multiply protonated [SP-B 1-25 +m H] m+ for m=5, 4 and 3, respectively. It is apparent that 4+ is the most abundant species under present conditions. We found that the relative abundances of the most protonated species increase at lower SP-B 1-25 (aq) concentrations (Fig. S3), a phenomenon that we tentatively ascribe to decreased repulsion among sparser polycationic chains in the air-water interfacial layers sampled by our technique. Herea er, our analysis will largely focus on the evolution of the m/z + =733.0 [SP-B . We did not detect n≥6 products in our experiments. Importantly, the application of 266 nm laser pulses enhances all product signals (Fig. 1B, red trace), thereby suggesting that the detectable products of SP-B 1-25 oxidation by O 3 and ⋅OH might be identical.

RESULTS AND DISCUSSION
is nding is in accordance with previous reports on the ozonation of amino acids, 48) and the ⋅OHinitiated oxidation of proteins. 49) Figure 2 shows [SP-B 1-25 +4 H+n O] 4+ mass spectral signal intensities from SP-B 1-25 (aq) microjets exposed to H] 3+ depletion is signi cantly larger than for the other protonated species suggests structural changes that depend on m. [50][51][52] Species of smaller m values behave as if they increase the exposure of the more hydrophobic (surfaceactive) amino acid residues to O 3 and ·OH.
is is consistent with experiments showing that more hydrophobic long-chain species react more extensively with ·OH(g) than the less hydrophobic shorter-chain ones. 31,32,34,35) is is also consistent with a recent report showing that antioxidant activity of free amino acids in Fenton's reaction increases with  hydrophobicity. 53) Figure 3 shows reactant decay and product enhancements upon irradiation in similar experiments as functions of 266 nm pulse energy. In Fig. 3, laser energies (x-axis) at 1, 5, 10, 20, 30, and 40 mJ pulse −1 correspond to [·OH(g)] 0 ≈2, 9, 18, 33, 46, and 57 ppmv, respectively. e actual [·OH] colliding to the microjets are expected to be smaller than these estimated values. 34,35) It is apparent that the full extent of irradiation e ects is already reached at the lowest pulse energies. Further increases in pulse energy only achieve the partial degradation of the n=1 and 3 [SP-B 1-25 +4 H+n O] 4+ products. [SP-B 1-25 +4 H+3 O] 4+ in particular disappears above ∼20 mJ pulse, possibly via photo-degradation or secondary reactions. e e ect of laser energy on product formation (Fig. 3)  Our results are consistent with the rapid incorporation of up to 5 O-atoms to SP-B 1-25 by both O 3 and ·OH at the air-water interface. Since (i) the reported rate constants for the reactions of cysteine (k Cys+ozone >7.0×10 6 M −1 s −1 ) and tryptophan (k Trp+ozone =7.0×10 6 M −1 s −1 ) with O 3 in bulk water are >10 2 times larger than for the rest of the amino acids (except for methionine), 48) (ii) the corresponding reactions with ·OH (k Trp+ · OH =1.3×10 10 M −1 s −1 ), (k Cys+ · OH =3.5×10 10 M −1 s −1 ) are di usionally controlled, (iii) both cysteines Cys 8 and Cys 11 , and tryptophan Trp 9 , are embedded in the hydrophobic section of SP-B 1-25 that resides at the topmost layers of air-water interface, 3,54) we infer that these are the three amino acid residues attacked by both O 3 and ·OH. More speci cally, we envision that in the initial stages of the oxidation process O 3 will transfer one O-atom to the sul ydryl R-S-H group of Cys to produce the corresponding sulfenic acid R-S-OH, 55) whereas O 3 also adds to the pyrrole ring of Trp to produce the corresponding primary POZ and secondary SOZ Trp-O 3 ozonides (Scheme S1). We also consider that the S-containing methionine Met 21 residue, 56) by being situated in the hydrophilic C-terminal side buried in water (as shown by MD simulations by Goddard and coworkers), 3) is less likely to be oxidized by both O 3 and ·OH than the Cys residues closer to the air-water interface. e fact that both O 3 and ·OH preferably reside at the air-water interface rather than in bulk [44][45][46] makes it less likely that they will di use towards the Met 21 residue located in bulk.
We also tested whether the mechanism of Trp ozonation at the air-water interface follows the same course as in bulk water in our experimental setup by exposing free L-tryptophan (aq) to O 3 (g)/O 2 (g)/H 2 O(g)/N 2 (g) mixtures in the absence/presence of 266 nm pulses (Fig.  S5). We found that the interfacial ozonation of anionic Ltryptophan (m/z − =203) produces species that incorporate one to three O-atoms, which are detected at m/z − =219 (+1O), 235 (+2O) and 251 (+3O), along with a signal at m/z − =207. ese species match the masses of the reported products of Trp ozonation 3,57) and hydroxylation 58,59) in bulk water. See Scheme S1 for assignments. We note, however, that the formation of kynurenine (Kyn) from the oxidation of Trp 9 in SP-B 1-25 , which would have led to a peak at m/z + =734=[2928+48 (+3O)−44 (−CO 2 )+4 (+4H)]/4, is absent from our mass spectra (Fig. 1B). us, from the fact that we observe the incorporation of up to 5 O-atoms into SP-B 1-25 , we conclude that Trp 9 accepts 3 O-atoms, and Cys 8 and Cys 11 one O-atom each during the ozonation and hydroxylation of SP-B 1-25 at the air-water interface under present conditions. Note that H-abstraction from Trp by ·OH in the presence of O 2 would have led to the formation of a peroxyl radical, m/z − =203−1 (−H)+32 (+O 2 )=234, and possibly to an alcohol m/z − =234−16+1=219, and carbonyl m/z − =234−16−1=217, from the disproportionation of the peroxyl radical, as in the case of alkyl and aromatic carboxylic acids. 31,32,34,35) eir conspicuous absence proves that under present conditions all oxidation processes, both those initiated by ozonation and hydroxylation, are initiated by O-atom transfers or ·OH additions to the S-center of the Cys and the pyrrole ring of Trp, rather than by H-atom abstraction from the myriad C-H bonds available in SP-B 1-25 . e exceptional reactivity of S-atoms in biomolecules for ⋅OH-addition is consistent with our GSH+·OH and GSSG+·OH studies at the air-water interface, 36) suggesting that interface-speci c phenomena are general and stem from the peculiar nature of interfacial water as a reaction medium. 26,[60][61][62] We also note that these highly selective oxidations imply "molecular recognition" processes, possibly mediated by water networks. 63,64) us, our results suggest that these hitherto unknown interface-speci c radical recognition processes may play central roles in lung surface chemistry. e occurrence of up to 5 O-atom transfers to a single SP-B 1-25 unit implies that oxidants are always in excess over reactive centers at the interface.
us, the initial attack on any one of them is followed by additional, successive O-atom transfers until all such centers reach their limiting degree of oxidation during contact times. However, since we have shown that GSH and free cysteine can add up to 3 O-atoms under similar conditions, 36,55) it appears that rst-generation sulfenic acids S(O)-H are not further oxidized in SP-B 1-25. is is because they may rapidly form S-S bridges or, due to conformational changes, they become buried in the hydrophilic segment of the polypeptide. 65) In this context, we note that the fate of sulfenic acid could be determined by the microenvironment. 65,66) For example, a S-OH can persist for several hours in human serum albumin, in which 34 of 35 cysteine-resides can create disul debonds thereby minimizing the number of available reduced S-atoms. 67) Another study revealed that the hypervalent S-atoms of sulfenic acids can form covalent complexes with the N-atoms of neighboring histidine residues, a phenomenon that prevents the over-oxidation of cysteine sulfenic acid. 68) erefore, it is conceivable that neighboring Trp 9 's indole-N-atoms may stabilize the Cys 8 -OH and Cys 11 -OH in a similar way. An additional e ect is that since sul nic S(O 2 )-H (pK a ≈2) and sulfonic acids S(O 2 )-OH (pK a <1) are much more acidic than S(O)-H (pK a ≈7.6), 55) they would deprotonate at pH ∼6 thereby decreasing the net charge of [SP-B 1-25 +m H+O] m+ from m (m−1) and shi ing the mass signals. However, the fact that we did not observe such behavior (Fig. S3) led us to exclude the oxidation of Cyssulfenic acid residues. e mechanism of SP-B 1-25 oxidations shown here is generally consistent with previous studies. 3,6,47) However, the previous report using the Fenton reaction for oxidizing SP-B 1-25 by ·OH in bulk water showed O n -products up to n=10, which is in contrast with our observations (n=1-5). 3) We ascribe this to the following reasons: 1) di erences between bulk vs. interfacial mechanism, 2) the very di erent time scales between the two systems, and 3) the Fenton reaction may largely produce oxo-ferryls rather than ·OH. First, as observed in GSH/GSSH+·OH experiments, 36) the di erence indeed comes from where the ·OH-reaction occurs, that is, in water vs. at the air-water interface. We recently reported that ·OH-reaction of amphiphilic species are remarkably di erent in bulk water vs. at the air-water interface. 32,34,35) Second, in the previous report the measurement was performed >12 h a er the Fenton reaction starts, 3) while in the present study the time scale is <10 µs. ird, as mentioned above, the Fenton reaction may produce oxoferryl species in >90% yields in bulk water at pH 6-7, 28) and even more so at the air-water interface. 26) In the present work, ·OH is generated by well-established gas-phase photo-reactions. 31) Hydroxylation, by increasing the hydrophilicity of the N-terminus, is generally expected to degrade the tensioactive properties of SP-B. 6,47,69) e formation of intra/intermolecular S-S bonding via sulfenic R-S-OH groups, that likely occurs in a longer time scale, could also disturb SP-B function as surfactant. 5,6,13) Since inhalation of PM and O 3 evidently induces reactive oxygen species (ROS) in ELF of our lungs, 20) the present results are directly linked to adverse health e ects of air pollutants impairing the role of SP-B as the surfactant.

CONCLUSION
We report a mass spectrometric study on how aqueous SP-B 1-25 is oxidized by O 3 and ·OH at the air-water interface. By using a novel method that combines online pneumatic ionization mass spectrometry with pulse laser ash photolysis, 31) we were able to detect the intermediates/ products generated in the initial stages of the oxidation of SP-B 1-25 by ·OH on water surfaces. Our results suggest that two Cys-residues and a Trp-residue in hydrophobic N-terminal side are the major targets for both O 3 and ·OH, rather than H-atom abstraction from the multiple C-H/N-H bonds available in SP-B 1-25 . We infer that this remarkable interface-speci c radical recognition process is what determines the observed chemistry. ese chemical transformations increase the polarity of the SP-B 1-25 hydrophobic section, promote the formation of disul de links therein and, therefore, are deemed to impair its role as the surfactant that prevents lung collapse upon expiration.

SUPPORTING INFORMATION AVAILABLE
Additional data and experimental details. is material is available free of charge via the Internet.