Chlorohydrin formation from unsaturated fatty acids reacted with hypochlorous acid

Chlorohydrin formation from unsaturated fatty acids reacted with hypochlorous acid

ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS Vol. 296, No. 2, August 1, pp. 547-555, 1992 Chlorohydrin Formation from Unsaturated Reacted with Hypochl...

932KB Sizes 0 Downloads 16 Views

ARCHIVES

OF BIOCHEMISTRY

AND

BIOPHYSICS

Vol. 296, No. 2, August 1, pp. 547-555, 1992

Chlorohydrin Formation from Unsaturated Reacted with Hypochlorous Acid Christine C. Winterbourn,*‘l

Fatty Acids

Jeroen J. M. van den Berg, Esther Roitman, and Frans A. Kuypers

Children’s Hospital Oakland Research Institute, Oakland, California 94806; and *Department Christchurch School of Medicine, Christchurch Hospital, Christchurch, New Zealand

Received January

28, 1992, and in revised form March 16, 1992

Stimulated neutrophils produce hypochlorous acid (HOCl) via the myeloperoxidase-catalyzed reaction of hydrogen peroxide with chloride. The reactions of HOC1 with oleic, linoleic, and arachidonic acids both as free fatty acids or bound in phosphatidylcholine have been studied. The products were identified by gas chromatography-mass spectrometry of the methylated and trimethylsilylated derivatives. Oleic acid was converted to the two 9,10-chlorohydrin isomers in near stoichiometric yield. Linoleic acid, at low HOCl:fatty acid ratios, yielded predominantly a mixture of the four possible monochlorohydrin isomers. Bischlorohydrins were also formed, in increasing amounts at higher HOC1 concentrations. Arachidonic acid gave a complex mixture of mono- and bischlorohydrins, the relative proportions depending on the amount of HOC1 added. Linoleic acid appears to be slightly more reactive than oleic acid with HOCl. Reactions of oleic and linoleic acids with myeloperoxidase, hydrogen peroxide, and chloride gave chlorohydrin products identical to those with HOCl. Lipid chlorohydrins have received little attention as products of reactions of neutrophil oxidants. They are more polar than the parent fatty acids, and if formed in cell membranes could cause disruption to membrane structure. Since cellular targets for HOC1 appear to be membrane constituents, chlorohydrin formation from unsaturated lipids could be significant in neutrophil-mediated cytotoxicity. 0 1992

Academic

of Pathology,

Press.

Inc.

Stimulated neutrophils produce hypochlorous (HOCl) via the myeloperoxidase (MP0)2-catalyzed

acid re-

r To whom correspondence should be addressed. * Abbreviations used: BSTFA, bis(trimethylsilyl)trifluoroacetamide; GC, gas chromatography; MPO, myeloperoxidase; MS, mass spectrometry; PBS, phosphate-buffered saline (50 mM phosphate, 110 mM NaCl, pH 7.4); PC, phosphatidylcholine; TLC, thin-layer chromatography; Me&X, trimethylsilyl; l&O, palmitic acid, l&O, stearic acid; l&l, oleic acid, 182, linoleic acid, 20~2, icosadienoic acid; 20:3, icosatrienoic acid; 20:4, arachidonic acid.

0003-9861/92 $5.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form

action of H202 with Cl-. HOC1 is a highly reactive oxidant and is thought to play an important role in both microbial killing and inflammatory tissue injury by neutrophils (1). The MPO/H202/C1system is toxic to both bacteria and mammalian cells. The mechanism of toxicity is not well understood, although membrane sites are thought to be prime targets (2,3). HOC1 reacts rapidly with many biological molecules. These include thiols, thioethers, and amines which form chloramines (4,5). Surprisingly little attention has been given to its reactions with membrane lipids. Neutrophils can cause lipid peroxidation (6-8). The reaction requires added iron and is not myeloperoxidasedependent. In fact myeloperoxidase-derived HOC1 inhibits peroxidation, and peroxidation is decreased in liposomes pretreated with HOC1 (8). These results suggest that the myeloperoxidase system could cause lipid modification other than peroxidation and have led to the present investigation. In this paper we show that HOC1 and MPO/H202/C1react with oleic (l&l), linoleic (18: 2), and arachidonic (20:4) acids, causing a loss in unsaturation and an increase in polarity. The chemistry of addition of HOC1 to double bonds to give chlorohydrins is well understood (9). Using gas chromatography-mass spectrometry (GC-MS), we have characterized the major products as isomeric chlorohydrin derivatives of the fatty acids. EXPERIMENTAL

PROCEDURES

Materials. Free fatty acids, phospholipids, and other biochemicals were obtained from the Sigma Chemical Co. (St. Louis, MO). Bis(trimethylsilyl)trifluoroacetamide (BSTFA) plus 1% trimethylchlorosilane was obtained from Pierce (Rockford, IL). Myeloperoxidase was purified as in (10). Sodium hypochlorite was purchased from Fischer. Its concentration was determined by reaction with monochlorodimedon (tm = 19,000 Mm’) (10). All solvents used were of HPLC-grade purity. Reaction of fatty acid micelles with HOCL Micelles of l&l, 182, or 20~4 were prepared by adding the lipid to phosphate-buffered saline (50 mM phosphate, 110 mM NaCl, pH 7.4) (PBS) and sonicating the nitrogenbubbled solution on ice for a few seconds. A known molar amount of

547 Inc. reserved.

548

WINTERBOURN

NaOCl was added to a 2 mu solution of fatty acid (usually 1-2 ml) while mixing continuously on a vortex mixer. After 10-15 min the solution was acidified with HCl to decrease the water solubility of the products and was extracted twice with dichloromethane. The dichloromethane extracts were evaporated under nitrogen and the lipid was dissolved in the appropriate solvent for analysis or derivatization. Reaction ofphosphdipid vesicles with HOCl. A mixture of 1-palmitoyl, 2-oleoyl phosphatidylcholine (l&O, l&l-PC), 1-palmitoyl, 2-linoleoyl phosphatidylcholine (l&O, l&2-PC), and 1-palmitoyl, 2-arachidonoyl phosphatidylcholine (l&O, 20:4-PC) was dried down under nitrogen from a chloroform solution. Subsequently PBS buffer was added to a final phospholipid concentration of 0.75 mM and multilamellar liposomes were formed by vortex mixing. Unilamellar vesicles were formed by extrusion of liposomes through filters with 0.1~pm pores (11). A known amount of NaOCl was added to 1 ml of the vesicle suspension while mixing continuously on a vortex mixer. After incubating at room temperature for 30 min, 100 ~111 mM CaClz was added and the phospholipids were hydrolyzed with 5 IU phospholipase AZ from bee venom in a 30min incubation at 37°C. Subsequently, lipids were extracted with dichloromethane as described above. The level of hydrolysis was tested by TLC on silicagel HR 60 (Merck, Darmstadt, Germany), eluted with chloroform/methanol/0.9% NaCl/acetic acid (100/50/5/16). All phospholipid was found to have been converted to lysophospholipid and free fatty acid. Derivatization and GC-MS analysis of the reaction products were performed as described below. Thin-layer chromatography. Modification of free fatty esters was monitored by TLC on silica gel plates eluted ether (bp 60-80”C)/ether/acetic acid (30:70:1). Spots by spraying with 40% HzSOl and heating. Iodine vapor tory for staining chlorohydrins.

acids or methyl with petroleum were visualized was unsatisfac-

Reaction of fatty acids with MPO. To solutions of l&l or l&2 in PBS, additions of 20 pM HxO, were made at lo-min intervals. MPO (13 nM) was added at the start and at every second addition of HzOz. At alternate additions of HzOz, 2 pM ascorbate was added to reactivate any reversibly inactivated MPO (12). It was necessary to maintain a low H202 concentration to minimize HxO*-dependent inactivation of the enzyme and competition by HxOx for the HOCl. Nevertheless, addition of further MPO during the reaction was necessary because the enzyme underwent HOCl-dependent inactivation. After IO-20 additions of HzOx, the lipid was acidified and extracted as for HOCl-treated samples. A second procedure was also used, in which lipid solutions containing MPO were infused with HzOz at 10 pM/min and ascorbate at 0.5 pM/ min. The H202 concentration, which was monitored with a peroxide electrode, was maintained below 10 pM by making further additions of MPO (13 nM). If either of these procedures was not followed, HOC1 yields were low and there was only a small amount of lipid modification.

ET AL. HP 5970 A MSD mass spectrometer (Hewlett-Packard, Palo Alto, CA) equipped with a 15-m DB-1 fused silica column (id = 0.25 mm, film thickness 0.25 wrn, J&W Scientific, Folsom, CA). The sample (1~1) was splitless injected with helium as the carrier gas and the quadrupole mass spectrometer was operated in electron impact mode at 70 eV. The column was kept at 50°C for 3 min and subsequently heated at a rate of 27’C/ min to 18O’C followed by an increase of 5”C/min to 310°C.

RESULTS

Loss of Unsaturatiun Treatment of either l&l micelles (not shown) or egg phosphatidylcholine liposomes (Fig. 1) with NaOCl resulted in a near stoichiometric loss of double bonds. There was an accompanying decrease in turbidity of the micelle suspension measured as AA 7oo(not shown), suggesting an increase in water solubility of the products. Rates of reaction were followed by adding monochlorodimedon at intervals after the NaOCl. With 0.5 mM (0.15 mg/ml) 18: 1, the NaOCl disappeared with a half-life of about 40 s and was all consumed within 3 min. A similar reaction rate was observed with egg PC. Rates were approximately twice as fast at pH 5.8 and 30 times slower at pH 9.5 than they were at pH 7.4. Since HOC1 has a pK, of 7.5 these results suggest that HOC1 rather than OCl- is the reactive species. Throughout the manuscript, therefore, although NaOCl was added to the lipid, the reactions have been attributed to HOCl. Separation of Products by TLC TLC of dichloromethane extracts showed that the reaction of l&l or 18~2with HOC1 gave products with increased polarity. With l&l (R+l55), there was one major product band with R,Q.40. 18:2 (Rfl.55) gave groups of products at R,Q.2-0.4 and R,O.O5-0.1.After methylation,

Loss of unsaturation. The effect of HOC1 on the degree of unsaturation of 181 micelles and egg PC liposomes was determined by measuring the change in iodine number (13) upon adding known amounts of NaOCl to the lipid (2-5 mg in PBS). The rate of disappearance of NaOCl was measured by determining the amount remaining at short intervals after the start of the reaction by adding monochlorodimedon and measuring AAm,. Derivatization. Samples were analyzed either as [email protected] ester/MesSi ether derivatives or as methyl ester/MqSi ether derivatives. For conversion of fatty acids to methyl esters prior to MesSi derivatization, samples were heated in 6 M HCl in methanol at 70°C under nitrogen for 1 h. An equal volume of water was added and the methyl esters were extracted into dichloromethane. For MeeSi derivatization, samples (either free fatty acids or methyl esters) were dissolved in 100 ~1 acetonitrile. BSTFA (50 ~1) containing 1% trimethylchlorosilane was added and the mixture was heated at 70°C for 1 h. The samples were then used for GC-MS analysis without further extraction. Gas chromatography-mass spectrometry. GC-MS analyses were carried out using a HP 5790 A Series gas chromatograph coupled to a

0.0

2.0

pmoles

4.0

6.0

8.0

HOC1 added

FIG. 1. Loss of unsaturation of egg phosphatidylcholine upon treatment with NaOCl. Phosphatidylcholine (4 mg) was reacted with NaOCl and the degree of unsaturation analyzed using the iodine number assay. Each point represents the mean of three assays, with variation falling within the symbol size.

FATTY

ACID

549

CHLOROHYDRINS

the TLC patterns were similar except that the bands all had slightly higher Rf values due to the lower polarity of the methyl esters. Thus, TLC provides a simple method of monitoring the outcome of the reaction before proceeding to derivatization and GC-MS. Identification

of Products

El

Using GC-MS

Figure 2 shows the GC-MS total ion chromatograms after reaction of l&l (Fig. 2A), 18:2 (Fig. 2B), and 20:4 (Fig. 2C) with HOC1 and subsequent conversion of the lipid extracts to Me$i ester/Me& ether derivatives (i.e., without prior methylation of the fatty acids). The l&l reaction mixture shows the starting material eluting at 14.7 min and essentially only one product peak at 20.2 min. In the 18:2 reaction mixture, the unreacted fatty acid eluted at 14.5 min with two clusters of product peaks at around 19.8 and 23.9 min. Unreacted 20:4 eluted at 16.9 min, with product clusters at 22 and 26 min. Products of the Reaction of Oleic Acid with HOC1

14

16

18

20

22

24

200'

"5)

150,

1

The mass spectra and proposed structures and fragc mentation of the products obtained from reaction of 18: i7j iz 100, 1 with HOC1 are given in Fig. 3. When analyzed as MeaSi ester/Me$li ether derivatives (Fig. 3A), the obtained mass 5spectrum indicated the presence of an isomeric mixture 6 50. of (9,lO)chlorohydrin derivatives of stearic acid (18:0), compounds I and II, as would be expected after addition of HOC1 to the double bond in 18:l. The molecular ion with expected m/z 478 was not observed. Instead, the 0. fragment ion after loss of a methyl group at m/z 463 ((M14 18 16 20 22 24 CH,)+) was found. Other characteristic fragment ions were found at m/z 317, 263, 215, and 365, corresponding to fragmentation next to the -0TMS groups in compounds I and II as shown in Fig. 3A. The fragment ions at m/z 463, 365, and 263 show the characteristic isotope distribution pattern indicative of the presence of Cl (35C1:37C1 isotope ratio approximately 3:l). This identification and monochlomhydrins 202 bischlorohydrins fragmentation pattern was confirmed by analyzing the products as methyl ester/Me$Si ether derivatives. The total ion chromatogram (not shown) gave one significant product peak with a retention time of 19.0 min. The mass spectrum of this peak (Fig. 3B) does not show a molecular 50 ion at m/z 420 but there were low intensity fragment ions corresponding to (M-CH3)” and (M-OCH3)+ at m/z 405 and 389, respectively. Other characteristic fragment ions 0 were m/z 259, 263, and 307, corresponding to fragmen16 20 18 22 24 26 tation next to the -0TMS groups in compounds III and RETENTION TIME (min) IV as shown in Fig. 3B. The presence of Cl in fragments m/z 405,389,307, and 263 was indicated by the Cl isotope FIG. 2. GC-MS total ion chromatograms, after reaction with NaOCl distribution pattern as explained above. Thus, the product of (A) MeeSi-derivatized l&l (NaOCl:fatty acid ratio approx 1:2); (B) 18~2 (NaOChfatty acid ratio approx 1:l); and (C) structures derived from the methyl ester and the MesSi Me&-derivatized Me,Si-derivatized 20:4 (NaOChfatty acid ratio approx 2:l). The idenester fragmentation data together clearly establish an tification of the products as isomeric mixtures of monochlorohydrins isomeric mixture of l&O (9,10)-chlorohydrins as the and bischlorohydrins was achieved on the basis of characteristic fragment products of the reaction of HOC1 with l&l. ions as described in Figs. 3-5 and the text.

A

!zl

550

WINTERBOURN

When the reaction was carried out at slightly basic pH (a-7.8), a second (minor) reaction product peak was found in the total ion chromatogram of the MesSi ester/Me$Si ether derivatives with a retention time of 17.1 min. Its mass spectrum showed no evidence of the presence of Cl. On the basis of the fragmentation data (not shown) this product could be identified as the 9,10-epoxide of l&O (M+ = 370). Under slightly alkaline conditions, the epoxide could form through elimination of HCl from the chlorohydrin (9). It was not present in the methyl ester/ MesSi ether samples because of hydrolysis back to the chlorohydrin by CH30H/HC1 during derivatization (14). This product was not found in significant amounts at pH 7.4.

Products of the Reaction of Linoleic Acid with HOC1 The product clusters observed in the total ion chromatogram of the MeaSi ester/MesSi ether derivatives, after reaction of 182 with HOC1 (Fig. 2B), each consisted of (at least) three partially separated peaks. The proportions of both groups of peaks relative to l&2 increased with increasing amounts of HOC1 added, as did the proportion of the second cluster (retention time 23.9 min) relative to the first cluster (19.8 min) (Table I). Absolute quantification of the relative product yields is not possible without knowing the response factors for each component. However, by comparing the peak areas with the 18:2:

2 e

loo-

ET AL. TABLE

I

Relative Peak Areas of the Major Components in Total Ion Chromatograms of MesSi Ester/MegSi Ether Derivatives after the Reaction of Linoleic Acid with Different Amounts of NaOCl Peak area

Approx 18:2:NaOClratio

l&2

19.8-min cluster

23.9-min cluster

51 2:l 12

0.49 0.39 0.08

0.38 0.40 0.40

0.13 0.21 0.52

NaOCl ratios it would appear that responses are higher for the products than for the unmodified fatty acid. GCMS of the equivalent methyl ester/Me,Si ether derivatives also gave two product clusters (not shown) with retention times of 19.0 and 22.8 min. The mass spectra of the first cluster of product peaks (19.7-19.9 min) in Fig. 2B showed some common fragment ions and others that were unique to one peak. Two characteristic spectra are shown in Figs. 4A and 4B. The fragmentation patterns are consistent with an isomeric mixture of the (9,10)- and (12,13)-monochlorohydrins of 18: 1 (compounds V-VIII in Fig. 4). No molecular ion was observed, but each peak in the cluster showed an (M-

317

A

60-

s g

6.

p

40-

5x

161 I 100

365

215 !a3 L

I 200

I 300

400

(M-CH3)+ 463

-I 500

m/z

m

CH3 -cCH2P--CH

4’ OTMS I”“i”H?&bO,CHs

215 60 263’

I J(

4osd

215 i 00

m

CH3 -(CH2)7 ~~-KH~~CH3 CH

m/z

FIG. 3. Mass spectra, fragmentation patterns, and identification of the reaction products of 18:l with HOCl, analyzed as (A) Me,Si esters/ MesSi ethers or (B) methyl esters/Me$Si ethers. Addition of HOC1 to the double bond in l&l is shown to lead to the formation of two isomeric 18:O chlorohydrins (Me$Si esters I and II: &f+ = 478, (M-CH3)+ = 463; methyl esters III and IV: iVf+ = 420, (M-CHJ+ = 405).

FATTY

% 100 .z so E s 6o f 40 .s =2 20

-A

551

CHLOROHYDRINS

ClI~3-7

355

El

CHs -!CHzk-CH=

CH-

CHz-CH ’

I’ CH

(CHz)7-

COOTMS

129 t

317

7

0

ACID

213 200

100

t’

451(5x)

213

403 L

26, 300

251d

WCHS)+

I

400

1

TM50

, 600

500

rjil

CH3 -(CHzh-CH=

ml2

CH-

CH?.+:H

Cl :H--(CHPD-

COOTMS

L---e5 1’3IOO- B a .= so t =2 6O2p 40- 129 s 2

200, 100

lylll

317

491 (5x)

(M.CH3)+

173 I ,

200

300

359

449

I

I 400

m

500

COOTMS

CHJ

COOTMS

I 600

FIG. 4. Mass spectra and reaction products from the reaction of 18~2 with HOC1 present in the first product cluster (RT 19.9 min, Fig. 2B), analyzed as fatty acid MesSi esters. Mass spectrum A is a characteristic spectrum early in the cluster (19.85 min), whereas spectrum B represents a slightly later-eluting product mixture (19.95 min). The products were identified as l&l monochlorohydrin isomers (M’ = 476, (M-CHs)+ = 461). Of the possible products V-VIII, products V and VI were predominant (see text).

CH,)+ ion at m/z 461, indicating a molecular ion of m/z 476. Other characteristic fragment ions are m/z 173,213, 221,261,317, and 365, corresponding to the fragmentation shown in Fig. 4. The abundance of fragment ions m/z 317 and 365 throughout the cluster indicates a preferential addition of HOC1 to the 9,10-double bond of 182 (compounds V and VI). Addition to the 12,13-double bond, represented by compounds VII and VIII, occurred to a lesser extent. The structural identification of the monochlorohydrin reaction products was confirmed by analysis of the methyl ester/Me$i ether derivatives. The spectra (not shown) exhibited a (very low intensity) molecular ion at m/z 418, an (M-CH,)+ ion at 403, and an (M-OCH&+ ion at 387. Other major characteristic fragments were at m/z 221 and 173 (the same fragments as shown in Fig. 4) and at m/z 307,299, and 259. These latter fragments are the methyl ester equivalents of the 365,357, and 317 fragments from compounds V, VI, and VIII. Two characteristic spectra from the second cluster of product peaks (retention time 23.8-24.0 min) in the MesSi ester/Me$i ether total ion chromatogram are shown in Figs. 5A and 5B. The presence of weak (M-CH,)+ ions at m/z 585 indicates a molecular ion of m/z 600, consistent with bischlorohydrin derivatives of 18:0. Characteristic ions were found at m/z. 173, 317, 221, and 365, corresponding to fragmentation of compounds IX-XII as shown in Fig. 5. The isotopic distribution of the latter two fragment ions and that at m/z. 585 indicates the

presence of Cl. The fragment ion at m/z 301 could not be unequivocally assigned, but possibly arises from the fragment ion at m/z 337 (CH,(CH,).&H(OTMS)CHClCH&H(OTMS)‘) after loss of HCl and rearrangement. The methyl ester/Me&Si ether derivatives (retention time 22.7-23.0 min, results not shown) gave a very weak (M-CH&+ ion at m/z 527, indicating a molecular ion of m/z 542, in agreement with the products being 18: 0 bischlorohydrins. Major fragment ions were present at m/z 173 and 221 (see Fig. 5), m/z 301 (possibly 337-HCl, see above), and m/z 259 and 307. The latter are the methyl ester equivalents of the 317 and 365 fragments in Fig. 5. These results confirm that the products in the second cluster are a mixture of positional isomers of 18:0 (9,10)and (12,13)-bischlorohydrins (compounds IX-XII). If the pH was controlled at pH 7.4 or less, the chlorohydrin derivatives were the only significant products. However, if the reaction with NaOCl was carried out at a pH nearer 8, GC separation of the MesSi esters gave another cluster of three peaks eluting after 19.6-19.8 min. The largest fragment in the mass spectra of these peaks was at m/z 515, with no evidence of chlorine. This suggests a molecular mass of 530, which is equivalent to that of MesSi-derivatized dihydroxyoleic acid. This product was not present in the methyl ester samples. It is assumed that it is derived from epoxides that would be formed by HCl elimination, as observed with 181, but absolute identification and the pathway of formation were not pursued further.

552 2,

WINTERBOURN ,

loo-

.=2

80

2 + $!

6o 40-

'i: d3

zo-

A

ET AL.

317 hEI

173

(10X) 129

0, IO0

173

(MG-l3)+

204 217



317

585

I

L,

300

200

CH3 -(CHZh-CH

400

so0

600

CH-(CH2)?-

COOTMS

ml2

lxll loob ‘G

B .

so-

301

5 E

60

.g z

4020-

221

173

365

TMSO 123

204

391 I

f

0-l 100

200

300

400

Cl

(M-CHI)+ (10X) 585

465

337

CHs -KH2k-rH

440

CH-

TMS? c’ CH2 CH-CH-(CH2)7-

COOTMS

I so0

600

m/z

q

,731’

T'.,

FIG. 5. Mass spectra and products from the reaction of 18:2 with HOC1 present in the second product cluster (RT 23.9 min, Fig. 2B), analyzed as fatty acid MesSi esters. Mass spectrum A was recorded early in the cluster (23.85 min), whereas mass spectrum B was obtained more toward the end of the cluster (24.05 min). The products were identified as 180 bischlorohydrin isomers (Af’ = 600, (M-CHJ+ = 585) (see text).

Products of the Reaction of Arachidonic Acid with HOC1 From the results with l&l and 182, a multitude of products would be expected from the reaction of HOC1 with a fatty acid containing four double bonds. The investigation was restricted, therefore, to HOCl:20:4 ratios of <2:1 which gave only two clusters of multiple product peaks. Identification of the first cluster in the MesSi ester/ MeaSi ether chromatograph (Fig. 2C) as various icosatrienoic acid (20~3)monochlorohydrin isomers (molecular mass 500) is based on the presence of a fragment ion at m/z 485 corresponding to (M-CH,)+. Similarly, the presence of a m/z 609 fragment in the mass spectra of the second cluster is indicative of icosadienoic acid (2O:Z)bischlorohydrin isomers (molecular mass 624). Both these fragments showed a Cl isomer distribution. Since our main aim was to demonstrate formation of chlorohydrin species rather than determine the isomeric distribution, full analyses of fragmentation patterns were not performed. However, fragment ions expected for cleavage adjacent to OTMS groups resulting from HOC1 addition across each double bond of 20:4 could be identified (Table II). The relatively greater abundance of m/z 309 and 261 could indicate preferential attack at the 5,6-position, but further analysis is required to confirm this. GC-MS of the methyl ester/Me$Si ethers supported the identification. As well as unreacted 20:4 (retention time 15.3 min) there were two clusters of peaks at 20.420.8 min and 24.9-25.7 min. There was too much fragmentation to observe molecular ions at M+ 442 for the monochlorohydrins or M+ 566 for the bischlorohydrins,

or the equivalent (M-CH,)+ fragments, but major fragment ions equivalent to those in the MeaSi ester spectra were identifiable (Table II). Relative Reactivity of 18:l and 18:2 An equimolar mixture of l&l and 182 was reacted with either 0.1 or 0.2 mol of NaOCl per mole total fatty acid. GC-MS analysis of the MeBSi ester/Me$i ether derivatives identified peaks corresponding to the monochlorohydrin of 18:l and the mono- and bischlorohydrins of 18: 2. Increasing the NaOCl concentration resulted in the consumption of more of the starting material and an increase in the amount of the bis- relative to the monoderivative of 182 (Table III). Although an accurate assessment of relative reactivities cannot be made without knowing the response factors for the different components, the data suggest that HOC1 shows some preference for 182 over 18:l. However, it is clear that there is not a high degree of selectivity for di- or monounsaturated fatty acids. Reaction of HOC1 with Phospholipid Vesicles Phospholipid vesicles, composed of an equimolar mixture of 16:0,18:1-PC, l&0,18:2-PC, and l&0,20:4-PC, were reacted at a NaOCl:total phospholipid ratio of approx 1:2). GC-MS analysis of the products as the MesSi ester/ MesSi ether derivatives (not shown) revealed a mixture of the chlorohydrins described above for each individual fatty acid. In the case of l&2 and 20:4 mainly monoch-

FATTY

553

ACID CHLOROHYDRINS TABLE II

Assignment

of Major Fragments

Me,Si ester/Me&

in the Mass Spectra of the Products of the Reaction of 20~4 with HOC1 Me ester/MesSi ether derivatives

ether derivatives Identity

4 309 261 221 173 389 349 301 313

Identity

ml2

CH(OTMS)CHCl(CH,),CO(OTMS)+ CH(OTMS)(CH,),CO(OTMS)+ CHs(CHx)&HClCH(OTMS)+ CH3(CH2),CH(OTMS)+ CH(OTMS)CHCl(CH,CH=CH),(CH,),CO(OTMS)+ CH(OTMS)(CH,CH=CH),(CHr)&O(OTMS)+ CH(OTMS)CH2CH=CH(CH2),CO(OTMS)+ Possibly 349-HCl

251 203 221 173 331 291

CH(OTMS)CHCl(CH,),CO(OCH,)+ CH(OTMS)(CH,),CO(OCH,)+ CH&H,),CHClCH(OTMS)+ CH,(CH,),CH(OTMS)+ CH(OTMS)CHCl(CH,CH=CH),(CH,)&O(OCH3)+ CH(OTMS)(CH,CH=CH),(CH,),CO(OCH,)+

255

Possibly 291-HCl

Note. The reaction conditions were the same as those for Fig. 3C. The fragments were present in one or more of the peaks in both product clusters.

lorohydrin products were detected. These results indicate that phospholipid fatty acyl groups as well as free fatty acids form chlorohydrins with HOCl. They also confirm that the reaction is not highly selective for polyunsaturated compared with monounsaturated fatty acids.

Reaction of 18:l and 18:2 with Myeloperoxidase

Solutions of l&l and l&2 were treated with serial ador continuous infusion of Hz02, in the presence of Cl- and MPO. With both fatty acids, products with increased polarity were seen on TLC. No reaction was seen if the MPO was omitted. GC-MS identified the products as the same chlorohydrin isomers as those seen with HOC1 treatment. With l&2, only very small amounts of bischlorohydrins were detected. Even allowing for the likely greater response from the chlorohydrin derivatives, the relative peak areas (Table IV) suggest reasonably efficient conversion of H202 to chlorohydrins. ditions

TABLE III Relative Amounts of Products Formed in the Reaction of an Equimolar Mixture of l&l and 18:2 with NaOCl Relative peak area mol NaOCl added/ mol total fatty acid Component

0.1

0.2

18:l monochlorohydrins l&2 monochlorohydrins 19~2bischlorohydrins Unreacted 18~1plus l&2

1.0 1.74 0.91

1.0 1.17 1.08

4.5

2.8

Note. GC-MS of the MesSi ester/MesSi ether derivatives of the total reaction mixture was carried out and peak areas in the total ion chromatograms were determined.

DISCUSSION

This study has shown that HOCl, a major neutrophil oxidant, reacts with unsaturated fatty acids to give chlorohydrins as the major products. With 18:1, mass spectral evidence shows that both isomers of the 9,10-chlorohydrin were formed. With 18:2 and 20~4, isomeric mixtures of monochlorohydrins were predominant at low HOChfatty acid ratios, with the proportions of bischlorohydrins increasing with increasing amounts of HOC1 added. All the double bonds in the polyunsaturated fatty acids were susceptible, resulting in a mixture of isomers. However, the GC-MS fragmentation suggests some preference for the 9,10-position of 18~2and possibly the 5,6-position of 20: 4. There appeared to be only a slight preferential reactivity of l&2 over 18:l. Unsaturated fatty acyl groups in PC gave the same chlorohydrin products as the free fatty acids. The reactions of 18:l and l&2 with the myeloperoxidase/HsOz/C1l system gave the same products as those obtained with HOCl. The areas of the product peaks after GC-MS suggest that much of the HzOz gave rise to chlorohydrins. To achieve this efficiency it was necessary to ensure complete conversion of HzOz to HOC1 by the MPO and to minimize competing reactions of HOC1 with H202 and MPO and inactivation of the enzyme (12, 15, 16). This required low H202 concentrations and several additions of MPO. A single addition of MPO and Hz02 at the beginning of the experiment gave very little reaction. Thus, efficient chlorination of unsaturated lipids by the MPO system can occur but requires stringent conditions. Previous work on oxidative modification of lipids by neutrophils has concentrated mainly on peroxidation (68, 17-19). Little attention has been given to other reactions of MPO or HOCl. Hubbard and co-workers (20-22) investigated the peroxidase-catalyzed iodination of arachidonic acid by HzOz and II. Using several peroxidases

554

WINTERBOURN TABLE

ET AL. IV

Conversion of Oleic and Linoleic Acids to Chlorohydrins by Reaction with Myeloperoxidase, Hydrogen Peroxide, and Chloride Relative Fatty acid

HzO, addition

l&l

Continuous infusion Serial additions Serial additions Serial additions

l&2

mol H,O,/mol

fatty acid”

Unmodified

0.14 0.21 0.16 0.26

fatty acid

0.88 0.86 0.76 0.64

peak areab Monochlorohydrin 0.12 0.14 0.24 0.36

’ Total amount of HzOz added per mole of 181 or 182, in the presence of MPO and 0.14 M NaCl. * From GC-MS total ion chromatograms of MeaSi-derivatized samples.

including MPO, they showed by GC-MS that iodohydrin and iodolactone isomers are formed. Our study has demonstrated comparable reactions of Cl- and MPO to form chlorohydrins. We did not see chlorolactones with l&l or l&2, presumably because the positions of the hydroxyl groups would not favor for lactonization (21). We have shown that chlorohydrins are formed when reagent HOC1 or the MPO system reacts with unsaturated fatty acyl groups. Whether they are formed when the lipid is exposed to stimulated neutrophils has not yet been established. However, there is indirect evidence for such reactions. Sepe and Clark (23,24) observed an MPO-dependent increase in permeability of unsaturated liposomes, in the absence of any peroxidation, and suggested chlorohydrin formation as a possible mechanism. Also, inhibition of neutrophil-mediated peroxidation of liposomes by MPO-derived HOC1 (8) could be explained by HOC1 addition to the unsaturated lipids. Whether membrane lipids form chlorohydrins when cells are exposed to HOC1 or neutrophils will depend on how well they can compete with other constituents for HOCl. With both l&l and egg phosphatidylcholine, we measured stoichiometric loss of double bonds on treatment with HOCl. However, the reaction of HOC1 with unsaturated fatty acids is slower than, for example, its reactions with thiol and amino groups (4,5). If chlorohydrins are formed within membranes, they could have pronounced effects. The increase in polarity, apparent both on chromatography and as increased water solubility of HOCl-treated micelles, is likely to be disruptive to the physical organization of the membrane and its function as a permeability barrier. The possibility of metabolism, e.g., by phospholipases, also remains to be explored. Another potentially significant reaction of MPO-derived HOC1 with unsaturated lipids is the formation of chlorohydrin derivatives of eicosanoids. There are several reports of MPO-dependent inactivation of leukotrienes and prostaglandins (25-27) that may be explained by this mechanism. If these reactions do occur physiologically, then the interesting question arises as to whether the

chlorohydrin derivatives have biological activity or can modulate the inflammatory response. Thus, further work needs to be carried out to determine whether chlorohydrins are formed in more complex physiological situations and, if so, whether they are involved in microbicidal activity or inflammatory tissue injury by neutrophils. ACKNOWLEDGMENTS We are grateful to Dr. Bertram H. Lubin for his support of this work and Dr. Cedric Shackleton for his expert adviceon massspectrometric analysis. We also thank Ms. Maggie Yee for her expert technical assistance. The present study was supported by the Medical Research Council of New Zealand, the R. G. Bell Trust, and National Institute of Health Grants HL 27059, HL 21061, HL 20985, DK 32094, and DK 34400.

REFERENCES 1. Klebanoff, S. J. (1988) Inflammation: Basic Principles and Clinical Correlates (Gallin, J. I., Goldstein, I. M., and Snyderman, R., Eds.), pp. 391-443, Raven Press, New York. 2. Rosen, H., Orman, J., Rakita, R. M., Michel, B. R., and VanDevanter, D. R. (1990) Proc. N&l. Acad. Sci. USA 87, 10,048-10,052. 3. Schraufstatter, I. U., Browne, K., Harris, A., Hyslop, P. A., Jackson, J. H., Quehenberger, O., and Cochrane, C. G. (1990) J. Clin. Inuest. 85,554-562. 4. Albrich, J. M., McCarthy, C. A., and Hurst, J. K. (1981) Proc. N&l. Acad. Sci. USA 78, 210-214. 5. Winterbourn, C. C. (1985) Biochim. Biophys. Actu 840, 204-210. 6. Stossel, T. P., Mason, R. J., and Smith, A. L. (1974) J. Clin. Invest. 64,638645. 7. Carlin, G., and Arfors, K. E. (1985) J. Free Radicals Biol. Med. 1, 437-442. 8. Winterbourn, C. C., Monteiro, H. P., and Galilee, C. F. (1990) Btichim. Biophys. Acta 1056, 179-185. 9. Marsh, J. (1977) Advanced Organic Chemistry: Reactions, Mechanisms and Structure, McGraw-Hill Kogakusha, Tokyo. 10. Kettle, A. J., and Winterbourn, C. C. (1988) Biochem. J. X2,529536. 11. Mayer, L. D., Hope, M. J., and Cullis, P. R. (1986) B&him. Btiphys. Acta 858, 223-230. 12. Bolscher, B. G. J. M., Zoutberg, G. R., Cuparus, R. A., and Wever, R. (1984) Biochim. Biophys. Acta 784, 189-191. 13. Kates, M. (1972) Techniques of Lipidology: Isolation, Analysis and Identification of Lipids, North-Holland, Amsterdam.

FATTY

ACID

CHLOROHYDRINS

14. Taguchi, V. Y. (1990) in Gas Chromatography: Biochemical, Biomedical and Clinical Applications (Clement, R. E., Ed.), pp. 129179, Wiley, New York. 15. Held, A. M., Halko,

D. J., and Hurst, J. K. (1978) J. Am. Chem.

Sot. 100,5732-5740. 16. Vissers, M. C. M., and Winterbourn,

C. C. (1987) B&rem.

J. 245,

277-280. 17. Thomas, M. J., Shirley, L. R. (1986) Biochemistry

P. S., Hedrick, 25, 8042-8048.

C. C., and DeChatelet,

18. Carlin, G., and Djursater,

R. (1988) Free Radical Res. Commun. 4,

251-257. 19. Claster, S., Chiu, D. T., Quintanilha, 64, 1079-1084.

A., and Lubin, B. (1984) Blood

555

20. Turk, J., Henderson, W. R., Klebanoff, S. J., and Hubbard, W. C. (1983) Biochim. Biophys. Acta 751,189-200. 21. Boeynaems, J. M., and Hubbard, W. C. (1980) J. Biol. Chem. 255, 9001-9004. 22. Boeynaems, J. M., Reagan, D., and Hubbard, W. C. (1981) Lipids

16,246-249. 23. Sepe, S. M., and Clark, R. A. (1985) J. Zmmunol. 134, 1888-1895. 24. Sepe, S. M., and Clark, R. A. (1985) J. Zmmunol. 134,1896-1901. 25. Henderson, W. R., Jorg, A., andKlebanoff, S. J. (1982) J. Zmmunol.

128,2609-2614. 26. Goetzl, E. J. (1982) Biochem. Biophys. Res. Commun. 106, 270275. 27. Paredes, J. M., and Weiss, S. J. (1982) J. Biol. Chem. 257, 27382740.