Modified apolipoprotein pattern after irradiation of human high-density lipoproteins by ultraviolet B

Modified apolipoprotein pattern after irradiation of human high-density lipoproteins by ultraviolet B

167 Bioch#nica et Biophyxica Acts, 11-'8 (191)2) 167-173 ,i~ 1992 Elsevier Science Publishers B.v. Aii righi..-., . . . . . ,.d ~;1~;5-2761~/92/$[15...

704KB Sizes 0 Downloads 50 Views

167

Bioch#nica et Biophyxica Acts, 11-'8 (191)2) 167-173 ,i~ 1992 Elsevier Science Publishers B.v. Aii righi..-., . . . . . ,.d ~;1~;5-2761~/92/$[15.t1~1

BBALIP 541)10

Modified apolipoprotein pattern after irradiation of human high-density lipoproteins by ultraviolet B Suzanne Salmon ", Ren6 Santus ~', Jean Claude Mazi~re ~'", Michel Aubailly " and Josiane Haigle ~'

'* l~]tlSt'tll~l NallOlltll d 7hstom' Naturelh,. Lahoratoire tie Ph,-~ico.('hmde tlt. l: 4~h~tnaliml Bioiogkllw. Pari.v (France~ and ~' Laboratoire th' Biochimic. Factdt? tfl" Mt:dccinc 5ai,t .-lntoim; Paris (France)

(Received 27 Fcbrt,a%' 1992) (Revised manus~:,tpl received 10 June 1992)

Key words: Pholoperoxidalion: I-IDL: Apolipoprotcin A-I; Apolipoprolein A-It: SDS-PAGE; Immunoblol: Monoclonal antibt~dy

The ultraviolet B-induced destruction of tryptophan residues and lipid pcroxidation of high-density lipoproteins is accompanied by the immediate and marked structural modific:qion of the apolipoproteins, as revealed by SDS-oolyacrylamide gel elcctrophoresis and immunoblot with specific mont~clonat antibodies. Formation of several polymers of apolipoprotein A-I, apolipoprotein A-I! or both apolipoprotcins occurred, although apolipoprotcin A-il did not contain any Trio residue. These results suggest that initial photochemical damage can be transferred via intramacromolecular proce~,~es to other sites within the same apolipo0rotein and by intermaeromolecular reactions from apolipoprotcin A-I to other apolilaoproteins, in both cases, lipid peroxidation enhances the propagation of the initial photochemical damage. The physiological significance of this work is discussed with respect to the low-light doses required for the alterations of the high-density lipoproteins.

Introduction

In previous studies [1,2], we have demonstrated that the Trp residues of serum lipoproteins are susceptible to ultraviolet (UV) B-induced photolysis. Their photooxidation triggers lipid peroxidation and leads to the consumption of vitamin E and carotenoids, the major classes of antioxidants found in lipoproteins [ 1,2]. We have emphasized the possible role of the H D L as important photochemical targets of the interstitial fluid feeding the epidermal cells [2]. The major HDL fraction in human serum is constituted by the so-called H D L 3 composed of two main proteins: the apoli0oprotein A-I (apo A-I) and the apolipoprotein A-il (apo A-ll). The other apolipoproteins of the particle represent less than 10% of the total protein content. Apolipoproteins A-I and A-ll are irregularly distributed. About 75% of H D L a contain two molecules of apo A-I

Correspondence to: R. Santus, Museum National d'Histoire Naturelle, Laboratoire de Phy,fico-Chimie de l'Adaptation Biofogique.

INSERUM U 312, 43 rue Cuvier. 75231 Paris ct~dex 05, France. Abbreviations: A ~ . a0oli0oprotein; CTAB. cetyltrimethylammonium bromide; DOX, desferrioxamine; HDL3, high-density lipoprotein (fraction 3): PAGE. oolyacwlamide gel electrophoresis; TBARS, thiobarbituric acid reactive substances.

and two molecules of aim A-il, while the rest is devoid of apo A-ll [3]. The amino acid sequence of apolipoproteins is well known. Apoli0oprotein A-I is a single polypeptide chain of 243 residues (molecular mass: 28 kDa). It contains 4 Trp and it has no eysteine nor cystine [4], The apolipoprotein A-I! is a dimer of a 77 amino acid polypeptide connected by a disulphide bond (molecular mass: 17.4 kDa). Its primary structure is characterized by the absence of arginine, tryptophan, histidine and free cysteine [5]. The 4 T r p / m o l of the apo A-I are the only amino acids that can absorb wavelengths longer than 295 nm, corresponding to the onset of the solar UV spectrum on earth. The a m A-I concentration in the interstitial fluid is about 15/.tM [6] in the 50 /.tm-thiek layer of the epidermis susceptible to the UVB radiations [7]. Since there is a rapid equilibration of plasma HDL with H D L of the extravascular compartment [8], it can therefore be considered that lipoproteins cannot be neglected as mediators in the overall effects of UVB light on human skin through the specific or unspecific interactions of photochemically-modified H D L with human skin ceils. It is well known that free radical formation in proteins can induce their cleavage or cross-linking [9]. A consequence of apolipoprotein alteration may be a disfunction of the recognition of the lipoproteins by their respective receptors and therefore, of the

168 metabolism of cholesterol [10]. It has been shown that aggregated chemically-modified HDL are no longer recognized by the HDL receptors [11] and cannot export cholesterol from the cell lipoprotein. Moreover, auto- or copper-oxidized and malonaldialdehyde-modified HDL containing aggregated apolipoproteins have a reduced cholesterol-accepting capacity [ 12,13]. To help understanding possible modifications of biological activity of HDL towards human skin cells under mild UVB exposure, we present here results on the apolipoprotein pattern obtained after HDL photooxidation with UVB light doses well below the minimal erythemal dose corresponding to a mild sunburn (e.g., the so-called MED in dermatology) [I4]. Materials and Methods

High-density lipoproteins (HDL) were prepared, in presence of 0.5 mM EDTA, by sequential ultracentrifugation of serum obtained from normal volunteers and taken as the 1.125-1.21 fraction (HDL3). Before experiments, EDTA was removed by dialysis against 4 I of 10 mM phosphate buffer (pH 7.2) with 0.15 M NaCi (PBS) during 1 day. All solutions were prepared in triply distilled water. Irradiation at 290 nm _
tyrosine or histidine content, was determined by colorimetric assays [17,18]. The apolipoprotein pattern of HDL 3 was studied according to Laemmli [19] on I mm-thick, SDS-containing 12% or 7.5% polyacrylamide gel slabs in a Mini Protean II system. Material. chemicals and prestained molecular weight standards were pt!reha~ed f:om BioRad (94200 iwy.sur-Seine, France). Samples prepared without reduction by 2-mercaptoethanol contained 5-10 # g of proteins of irradiated or unirradiated HDL. The electrophoresis was carried out at 201) V until the Bromophenol blue marker reached the bottom of the gel. The proteins were either stained with Coomassie blue R-250, or electrically transferred to nitrocellulose sheets during 30 rain at 50 V [20]. Nitrocellulose transfers were saturated overnight at 8°C in 5% low fat milk in PBS and then incubated with monoclonal antibodies against apo A-I or apo A-II. Bound immunogiobulins were revealed with peroxidase-labelled antimouse-lgG, and the peroxidase activity was detected by reaction with the 4-ehloro-l-naphthol (0.05%)-hydrogen peroxide (0.01%) substrate. Anti-apo A-II monoclonal antibodies and peroxidase-labelled immunoglobulins were purchased from Biosis (60200 Compi~:gne, France). Monoclonal antibodies against apo A-I were a gift from Dr. M. Ayrault-Jarrier [21]. Results

The irradiation of aerated HDL3 leads to immediate Trp destruction and to the formation of lipid peroxide decomposition products (TBARS) as exemplified in Fig. 1 for various experimental cond:tions, it can be seen that in aerated solutions, addition of 1%

A 40

A

~5

,

1.6

,

30 2O

0

i O.8

2

4

6

8

10

Irradiation time (rain.)

i

0

2

4

6

O

10

Irradiation time (mln.)

Fig. I. Destruction (in ~;) of Trp residues (A) and corresponding TBARS production (B) in UVB irradiated HDL~ soluti'~.s in absence {closed symbt)ls) and Jr. presence ol I'~ SDS {open symbols). Data are the average of a least three determinations. The absorbed light dose was 0.4 J rain - i corresponding to an incident dose of 6.7 J rain -I, e, air: A, N20: I1, Ar: ©, a i r + S D S : zx, N20+SDS: t:],

Ar + SDS.

169 SDS before the irradiation does not greatly change the Trp destruction (Fig, !A) but abolishes thc TBARS production (Fig. IB). These data thus confirm and complement our previous observations [2], The formation of primary radical species resulting from Trp excitation could not only trigger the radic~t chain reactions of the lipid peroxidation [2] but may also be responsible for changes in the apolipoprotein structure or conformation. Possible protein structure alterations were therefore studied by SDS-PAGE experiments carried out on irradiated and control sampies. Fig. 2 shows the pattern of the main apo!ipo proteins of irradiated HDL. Immunoblotting using monoclonal anti-apo A-I or anti-apo A-ll antibod,~es demonstrates that the apt) A-I and apo A-II antigenic domains are present in several high molecular mass bands formed during the irradiation (Figs. 3 and 4). The high molecular mass bands appear within thc first minutes of irradiation of the air-saturated solutions, as illustrated by Fig. 3 showing the bands obtained after 3 min of irradiation, when about 85% of the Trp residues are still intact and little TBARS are formed. As will be discussed later, one minute of irradiation under our experimental conditions (see ReL 1 and references cited therein) is equivalent to an absorbed energy of 0.13 J c~a -'~. The proportion of high molecular mass bands increases with the irradiation time and after 20 min only high molecular mass polymers, ( > 100 kDa) are observed (Fig. 4B). it must be also mentioned that unirradiated control samples are not modified after bubbling with air during the 20 rain of the experiment (Fig. 4A). Since the radical chain reactions of the lipid peroxidation are propagated by redox metal ions [22], the k Da 205 116 80

46.s

ilb

A-I A-H

Time

0

I 0

20

(rain.)

Fig. 2. SDS-polyacrylamide gel ¢lectropho~esis pattern of aerated HDL 3 solutions irradiated during 0, 10 and 20 rain with the incident energy given in Fi~. I. The flat gel electrophoresis is stained with Coomassie Blue R250. Each lane contains 10 ~tg of proteins.

4 -II

A-I

k Da

k Da

....

• <' I

106

~

t 49.5

'~

~

205 116 46.5

I 32.5 27.5

Time (rain.) 0 3 6

d 10

0

3

6

10

Fig. 3. lmmunoblotsshowingthe modificationso[ apo A-1 (left) and apo A-II {right) during UVB irradiation. Aerated HDL3 solutions were irradiated during O, 3. 6 and I0 rain with the incident energy given in Fig. 1, See Materials and Methods for eleetrophoresis and immunotqot6ag co,ditions, Each lane contains 5 pg o[ proteins.

possible intervention of metal ion traces present in the HDL solutions was investigated [2]. When desferrioxamine, a strong ferric iron-complexing agent was added before irradiation, the Trp destruction rate was decreased (about 28% vs. 35% after 10 rain of irradiation) and the TBARS formation was markedly inhibited (about 0.5 vs. 1.5 nmol)(data not shown), which confirms previous results [2]. The formation of high molecular mass polymer was also slowed down by desferrioxamine (Fig. 4C). Table ! shows the molecular masses calculated from Coomassie Blue-colored gel electrophoresis or immunoblots. It should be noted that cross-linking occurs, producing dimers and trhners of apo A-I of calculated molecular masses 55 + 4 kDa and 82 + 4 kDa or apo A-I! bands of calculated molecular masses 36 + 1 kDa and 49 + 3 kDa. Aggregates of higher molecular mass are also formed. They are recognized by both antibodies, They could represent tetrainers of apo A-I (4 Apo A-i = 112 kDa) or associations between apo A-! and apo A-II. On the contrary, addition of 1% SDS before the irradiation inhibits their formation (Fig. 5) despite marked Trp destruction (Fig. 1A). Hydrated electrons (e,L,) are produced by the monophotonie photoionization of Trp by 300 nm radiations [23]. The possible involvement of the strongly reducing e:,a in the observed apolipoprotein changes was investigated by saturating the irradiated solutions with Ar. In At-saturated solutions, only one apo A-i polymer (-- 80 kDa) is formed whereas the apo A-II is not modified (Fig. 6) and the high molecular mass hand~ are absent. The N,_O is a good scavenger of the hydrated electrons which under N20 saturation are transformed into the strongly oxidizing hydroxyl radicals. Under NzO saturation (Fig. 6), polymers of apo

170

A-II

A-I

A

B

C

A

B

C

10 2 0

].02(}

I,,:. Da

k Da 106 -IIi so -ii 49.5 - - ~ ~:::. . . .

..,':~

2"/

.~..

(

i8.5 - i Time (min.)

.

.

.

.

0

10

if::

~ , ~ .

10 20

20

10 20

.

.

.

16 20

.

0

.

.

Fig. 4. lmmunot:lols of aerated HDI..~ solutions after 0. I0 and 20 rain of irradiation with the incident energy given in Fig. I (experimental conditions: see Fig. 3). (A) Unirradiated solutions bubbled with air during 0, 10 and 20 min (B) Irradiated solutions (C) Same as (B) but solutions contained 50 ~ M desferrioxamine (unirradiated controls with desferrioxamine are identical to those given in (A) and are not shown).

A-I and ape A-II of molecular mass < 100 kDa are produced, They slowly form possibly because of some oxygen contamination revealed by the production of a few TBARS (Fig. 1B), Experiments were also per-

restores the formation of high molecular mass polymers ( > 100 kDa). It was observed that their rate of formation was faster than in the aerated solutions. Thus, the polymer content after 10 min of irradiation

f o r m e d u n d e r saturation with the N z O - O z gas mixture. T h e p r e s e n c e o f O z in t h e gas m i x t u r e m u s t a l l o w f o r the propagation of the peroxldatlve chain reactions after the reaction of the hydroxyl radicals. As already o b s e r v e d [2], t h e T r p p h o t o d e s t r u c t i o n r a t e w a s accele r a t e d ( a b o u t 4 5 % vs. 3 5 % a f t e r 10 m i n o f i r r a d i a t i o n ) w h e r e a s T B A R S f o r m a t i o n w a s n o t a b l y i n h i b i t e d (0,7

vs. 1.5 nmol) (data not shown). The presence of oxygen

A-I k Da

le6 - - ~ 80 "-"~'~

TABLE l

Molecular mass of bands produced by UVB irradiation of high.densiO' lipoproteins Molecular masses were calculated from SDS-PAGE (12 or 7,5% acrylamide) or immunoblo:s obtained in a Mini Protean II system using prestained molecular mass standards. Because of the short migration length in Miniprotcan 11 gel slabs, given values are only estimates. The number ,.ff experiments is shown in parentheses, Calculated molecular mass (kDa) Colored NativeapoA-ll Native ape A-I band 3 band4 band 5 band0 band 7 band8 band 9

t4_+ 4(7) 27+ I (7) -

52+ 2(5) 68 78:t: 6(3) 1155:24(4) 207+ 12 (3)

Ape A-I blot 28_+ I (10) ~

55:1:4 (6) 67 82_+ 4 (7; 93+ 3 (5) 120_+17 (3) 19(I

Ape A-Ii blot 16±1 (9) 36+ 1 (5) 49+3(7) 63 + 8 (4) 90 123+3(3) 210

a2.s -!i:

lS.$ -Time (rain.)

0 air

3

10

0

3 10 N20

3

10

ARGON

Fig. 5. Immunoblots showing the ape A-I of HDL 3 solutions irradiated during O, 3 and 10 rain with the incident energy given in Fig. t in the presence of 1% SDS, Experimental conditions are the same as in Fig. 3. Solutions were bubbled with air, N:O and At as indicated on the figure,

171

A.II

A-I

plete antioxidant Lleaching [2]. Self-aggregation of the apolipoproteins is a consequence of light absorption by Trp residues since unirradiated samples do not form aggregatc~. Directs of apo A-! or apo A-il and higher polymers of apo A-I or of both apoliproteins are formed, in relation to our observations, it has been shown that polymerisafion of apo A-I occurring thR)ugh non-covalent dimer formation can induce tetramcr-c or octameric association with increased a-helicity [24]. It should bc noticed that apo A-II also produces dimers although this protein, unlike apo A-I, does not contain Trp residues and cannot be direcHy altered by the UVB irradiation. The formation of high molecular mass polymeric material from HDL undergoing oxidation is not exceptional. Thus, Zawadski et al. recently described the formation of high molecular mass apo A-l-containing particles during the Cu-'+-mediated oxidation of dialyzed plasma [25] as observed here under irradiation of isolated HDL. The addition of 50 /.tM desferrioxamine inhibits TBARS formation by about 70%, while it diminishes the tryptophan photodegradation by only 9_5%. At the same time, the apolipoprotein modification measured by SDS-PAGE occurs to a lesser extent. It is therefore probable that Fenton-type reactions catalyzed by trace amounts of redox metal ions bound to the lipoproteins are involved in their alteration. Regarding the N , O - O ,

k Da 106 8 0 OINID

49.$~ 32.s~

27.5~ ,.,:- ~

~.8.5 ....

O

Time !min.)

0

3

10

2O

0

3

10

ARGON

k Da

i?~

$ 0 lib

106 80BII

49.5

-

32,5 32.$~

2"I.5~

27.$

ill iiii

18..~

IS.5

Time (rain.)

effect, an increase in Trp photolysis (about 10%) leads to a higher rate of aggregation with less TBARS formation most probably because of site specific hydroxyl

0

3

N20

10

20

3

I0

N20

10

N20-02

Fig. 6 Immunoblots showing the modifications of apt) A-I (tort) and

alto A-il (risht) of HDL~ solutions irradiated during I), 3, I(I and 211 min with the incident energy given in Fig. I. Experimental conditions are the same as in Fig. 3. Solutions were bubbled with Ar (top) or N_,O(bottom).

of the N20-O2-saturated solutions was equivalent to that obtained after 20 min of irradiation of the aerated solutions (Fig. 6). Discussion

The main conclusion that can be drawn from this study on the effect of UVB radiations on lipoproteins is that circulating HDL apolipoproteins are s t r ~ g ! y altered by these radiations. Thus, alteration of the apoliltmprotein pattern, which requires oxygen, already occurs in the first minutes of irradiation before corn-

radicals formation [2] which react in the vicinity of their formation site (i.e., photoionized Trp residues). It is likely that under these conditions the apolipoproteins are the primary targets of the hydroxyl radicals. An important result is the absence of polymers originating from apo A-I in aerated samples irradiated in the presence of 1% SDS. The parallel inhibition of TBARS formation demonstrates the key role played by lipid peroxidation in the processes leading to protein degradation. As to the mechanism of the SDS effect, it is likely that the SDS micelles destroy the HDL structure. In particular, they separate apolipoproteins from the lipidic core whose lipids are shared by the many SDS micelles according to a Poisson distribution. This statistical distribution abolishes the lipid 0eroxidation triggerod by the Trp photolysis. The oxygen dependence of apo A-If degradation is also clearly evidenced by results obtained with deaerated solutions in which the damage on the apo A-! is not propagated to the apo A-II. In related studies, Kochevar described the appearance of high molecular mass proteins after UV irradiation of erythrocyte membranes. She concluded that Trp residues were responsible for these alterations and that they were more likely at the origin of the biological effects of UV radiation on membranes than oxidized phosphotit'ids [26]. The present study strongly

172 suggests that the implication of lipid peroxidation cannot be excluded. Propagation of the radical chain reactions of lipid peroxidation is not necessartly the only mechanism which accounts for the intra- and inter-apolipoprotein cross-linking. There are other mechanisms which, in addition to lipid peroxidation, may contribute to the light-~induced polymer formation by propagating the primary photochemical damage. First, the photoionization of Trp residues leads to the formation of the neutral tryptophanyl radical (see Ref. 23 for key referenccs). This radical can oxidize intact Tyr residues either in proteins [27] or in macromolecular assemblies Iikc the Fd phage through long-range interactions [28]. Two Tyr residues located at positions 100 and 115 in apo A-I are close to Trp 108 in the same helix [29]. The native apo A-ll dimer contains 8 Tyr [5]. It is possible that interactions a n d / o r spatial relaxation of A-I and A-II apolipoproteins in the native HDL are such that some of the 8 Tyr residues are close enough to be oxidized by a tryptophanyl radical. Secondly, during Trp photolysis, disulphide bonds, histidine and other basic amino acids (Lys and Arg) may be destroyed by a rcductivc reaction due to an intramolecular electron transfer from excited Trp to the peptide backbone [3(}]. The primary and secondary structures of HDL apolipoprotcins are well known. Most of the apolipoproteins have highly homologous 22 residue-peptide segments with high amphiphilic a-helix content [31] interrupted by proline or glycine residues which introduce a fl-turn in paired helices [32]. The amino acid sequence of apo A-I shows that Trp residues are located at position 8 in a random coil region and at positions 50, 72 and 1(}8 in highly tlelical rcgions. The fact that His residues are absent in apo A-If and located in apo A-I between residues 135-199, far from "Irp residues, may explain the absence of any His degradation during the irradiation (data not shown). According to the proposed model, the Trp residue at position 108 in a paired helix, has two adjacent basic Lys. During irradiation, modification of Lys occurs in the HDL to a lesser extent than in low-density lipoprotcins (unpublished observations) perhaps because only this single Trp has Lys residues in its vicinity (data not shown}. In addition, Brasseur ct al. demonstrated that fluorescent probes tinkcd to the phospholipid acyl chains arc close to the Trp residue located at position 108 [33]. This site. adjacent to the lipidic chains, may be a determinant of the UV-induced oxidizing process. It should bc noted that the UVB-induced modification of HDL arc important because the latter may lead to the impairment of their biological function. Indeed, ~.hc secondury structu,c of the apolipoprot,:h,~, partier,lady the conservation of the a-helix structure in the HDL particle, may be important for their binding to the high-affinity ceil membrane receptors [34]. Whereas

the role of oxidized low-density lipoproteins in the formation of the atheromatous plaque is well documented [35], the HDL is supposed to protect the easily oxidizable LDL from autooxydation [36-38]. On the contrary, very little is known on the metabolism of oxidized high-density lipoprotein. Azizova et al. reported a decrease in the cholesterol-accepting capacity of auto-oxidized HDL [12]. In agreement with their data, we demonstrated a decrease in the cholesterol efflux from cells by malonaldialdehyde-modified or copper-oxidized HDL [13]. In UV-exposed skin, the modified HDL from the interstitial fluid may impair reverse cholesterol transport from skin cells and cell membrane protection, if modified HDL can be exported to plasma, one may speculate that they may be less protective against low-density lipoprotein oxida.'.ion. This brings us to an imlaortant question arising from this study. Are the absorbed UVB light doses used under our experimental conditions (0.13 J cm -~) of physiological significance? The minimal erythema doses (MED) for fair-skinned people are about 40 mJ cm -2 at 300 am, 50 mJ em -2 at 304 nm and 1000 mJ cm -2 at 313 nm [14]. At these wavelengths corresponding to the onset of Trp residues in the HDL of the intertitial fluid bathing the epidermis, the transmittance of the stratum corneum is about 45% of the incident light at 313 nm and 32% at 304 nm [14,39]. We have pointed out that HDL are, with albumin, one of the major chromophores of the intertistial fluid. Since all the UVB light is absorbed in the 50/xm of the epidermal layer [39], it can be estimated that the absorbed light dose corresponding to ! MED around 310 nm is about 50 J cm -~ in the epidermal layer. Accordingly, there is more than enough light to induce marked alterations of the apolipoprotcins at doses corresponding to 1 MED. Furthermore, as emphasized in a previous study on photooxidized LDL [!], it is important to note that lipoproteins are an interesting model for the study of the propagation and the amplification of a photooxidative stress by lipid pcroxidation, via post-irradiation dark reactions. Indeed, this work demonstrates that the initial photodamage occurring at the level of chromophoric Trp residues can be transferred to other contiguous proteins only if these proteins can interact with oxygen and peroxidized lipids. The potential pl,ysiological significance of this work also implies that transition metal ions can be bound to lipoproteins to propagate the radical chain reaction of the lipid peroxidation [2]. Under normal physiological conditions, these ions arc sequestered in transport and storage proteins. However, it has been demonstrated that UVB irradiation of skin leads to the release of non-home iron [40]. Photoreduction of ferric ions stored in ferritin [41] a n d / o r metal ion containing protein degradation after

173

cell photokilling, may contribute to effective ferrous ion production [40].

Acknowledgements The authors arc plcascd to acknowledge stimulating discussions on UVB actinometry with Dr. M. Bazin. W e w i s h t o t h a n k M i s s M. A u c I a i r f o r h e r skillful h e l p

in preparing the Iipoproteins, Dr. M. AyrauIt-Jarricr (INSERM U 32, Hopital Henri-Mondor 94010 Cr&eil. France) for her hclpful gift of monoclonal antibodies, and CIBA-GEIGY laboratories (925(~) Rueii-Malmatson, France) for their generous gift of Desfcral. J.C. Mazihre wishes to thank "La Liguc Nationalc Franqaisc Contre

le C a n c e r , C o m i t 6 d e Paris" f o r p a r t i a l s u p p o r t .

References I Salmon, S., Mazi6re, J.C., Santus. R., Morlff:re, P. and Bouchemal, N. (1990) Photochem. Photohiol. 52. 541-545. 2 Salmon, S., Mazi6re, J.(',, Santus. R. and Morli~:re. P, (iqgl) Bioehim. Biophys. Acta 11)86. I-6. 3 Atmeh, R.F., Shepherd, J. and Packard. ('J. (1983) Biochim. Biophys. Aeta 751, 175-188. 4 Law. S.W. and Brewer, |I.B. Jr, (19841 Proc. Natl, Acad. Sci. USA 81.66-71). 5 Knot[. T.J., Priestley, LM.. Urdca. M. and Scott, J, (19$4) Biochem. Biophys. Res, Commun. 12t}, 734-74tl. 6 Vessby, B., Gustafsson. S., Chapman, J.M., l tellsing, K. and Lithell, tl. (1987)J. Lipid Res. 28, t~29-641, 7 Tyrelk R.M. and Pidoux, M. (!t~87} Cancer Rex. 47. 1825-1829. /4 Shrimp. C.tl., Dory, L. and Roheim, P.S, (1957) J. Lipid Res, 28, 225-237. 9 Hunt, J.V. and Dean. R.T. (1q89) Biochcm. Biophyx. Rex. Common. 162, 11)76-1084. HI Oram. J,F., Brinton, E.A. and Biermam E . L (ltJ83). J. Clin. invest. 72, 1611-1621. I 1 Brinton, E.A., Oram. J.F., Chen, C.H., AIbers, J.J. and Bierman, E.L. (19861 J. Biol. Chem, 261,495-503. 12 Azizova, O.A. and P:masenko, O.M.. Vol'Ntwa. I".V., Vladimirov, Y.A. {1989) Free Radio. Biol, Med. 7. 251-257. 13 Salmou, S., Mazi~:re, C,, Auclair, M., Themn. L., Santus, R, and Mazihre, J,C. (1'9~,~2)Bioehim. Biophys. Acta 1125, 230-235. 14 Parrish. J.A., Jaenicke. K.F. :rod Anderson, R.R. (It~821 Phot0chem. Pholobiol, 3h, 187-191. 15 Aus:, S.D. (Di86) in tlandhtn~k of methods fl)r o ~ g e n radical research (Greenwald. R.A.. ed.) 2(}3-2117. CRC Press. Boca Raton,

1~ Petcr~tm, G.t,. (I't77) ~nal. Biochcm. ~3, 34h-356. 17 ! lahecb, A.F.S.A. 119u~) Anal. Biochem. 14, 32~-33h. IS Sok~Iovsky, M. and ValIse, B.L (1906) Biochcmistpi 5. 357435~I. !q [.ac,,,mli. U K { IO"/(}j Nature 227. 6,ql)-fiSS. 2t) Towbin. t1.. Slaehlin, T. and (7ordon, J. (197tD Proc. Naii. ,\cad. Sci. USA 78. 43511-4354. 21 Peril, E.. &yrault-Jarrier. M,, Pasti._'r, D., Robin, tt.. Pohmovski. J., Aragon, I., 1[eGaud, E. and Pau, B. 119871 Bmehim. Biophys. Acta qlq. 2S7-2qf~. 22 Aust. S.D.. Morchousc. L.A. and Thomas. C.E, ~ltlS51 Free Radio. Biol. Med. 1.3-25. 23 Bazin, M., Patlerson. L.K, and Santus, R. (1983} J. Phys. Chem. ,'¢,7, W,q I'!9,. 24 Donov~.n, J.M., Bcnedek. G.B, and Carev, M.C. (1987) BiochcmistR.." 2ft. SI Ih-SI25, 25 Zawadski, Z., Milne, R.W. and Marcel. Y.L (Iqt/i) J, Lipid Res. 32, 243-25o. 26 Kochevar. I.E. (lq~l) Photoehem. Photobiol. 52, 795-8110. 27 Priitz, W.A.. Siebert, F., Butler, L, Land, E.J., Menez, A, and Monlenay-Garcstier, T. 119,'12) Bk~him. Biophys, Acta 7115, 139149. 28 Ghiron. CA., Santus. R., Bazin, M.. Butler, .!,, Land. E.,I. and Swalk}w. AJ. (t986) BiLvchim. Biophys. Acta Sfq, 3¢G-36& 29 Marcel, Y.L, Provost, R.. Koa, tt,. Raffai, E.. Dac, V.N., Fruchart, J.('. ~:nd Rassart, E. (1991) J. Biol. Chem. 266, 31~q3653. 30 Pigau[I, C and Gerard, D. (1989) Pholochcm. Photobiol, 50, 23-2g. 3t Li. W.II,, Tanimura. M.. LuG. C.C., Dalta. S. and Chan. L. (ItlHS} J. lipid Res. 2% 245-271. 32 Nakagaw;.l, S.II., Lau, tl,S.tt., Kfizdy, F.J. and Kaiser, E.T, (limB5) J. Am. ('hem. So¢. IO7, 7l}B7-7092. 33 Brasxeur. R., De Meuner, J.. Vanh~, B.. Gt~)rmaligh. E,. Ruyxxchacrl, J,M. and Rosscncu, M, (19~01 Biochim, Biophys. Acla I043. 245-252. 34 llerzyk. E.. {_)wen. J.S. and Chapman. D. Ilt~Sg) Biochim. Biophys. Acla ~62. 131-142. 35 Sleinberg, D., Parthasarathy, S., Carew, T.E,, Kht~), J,C. and Wilz'um, J.L, (It}89) N. Engl. J. Mud, 320, 9t5-924. 3fi Ohta. T., Takata. K,, Horiuchi, S., Morim~, Y, and Matsuda, I. (It~St)) FEBS Lelt. 257, 435-438. 37 Parthasarathy, S., Barnett, J. and Fong. L.G. {Iq~}l|) Biochim. Biophys, Acta lIMA. 275-283, 38 Mackncss. tl.l.. ArroL S. and DurringEon, P.N. { l~}l) FEBS Lelt. 28t~, 152-154. 39 Magnus, I.A. (1976) in Dermatological Photobiology: clinit:;d and experimental aspects, pp. 11-21, Blackwell Publishers. 40 Bisaetl, D.L., Chatterjec, R. and H:mnon, D.P. (1~}91) Photochcm. Photobiol, 54, 215-223, 41 Aubailly, M.. Sanlus, R. and Salmon, S. (It~)l) Pholochem. Photobiot. 54, 76t1-773.