Photoproducts of phytoplanktonic sterols: a potential source
of hydroperoxides in marine sediments?
Jean-FrancËois Rontani *, Daphne Marchand
Laboratoire d'OceÂanographie et de BiogeÂochimie (UMR 6535), Centre d'OceÂanologie de Marseille (OSU), Campus de Luminy, case 901, 13288 Marseille, France
Received 23 June 1999; accepted 29 November 1999 (Returned to author for revision 31 August 1999)
Abstract
A detailed study of the lipid composition of Recent sediments of Carteau Bay (Gulf of Fos, Mediterranean Sea) has made possible the detection of signi®cant amounts of 5-stenol photoproducts of phytoplanktonic origin. Photo-degradation of 5-stenols in senescent phytoplanktonic cells seems to play a role in the degradation of these
com-pounds in the marine environment. These reactions lead to the production of4-6a/b- and6-5a-hydroperoxysterols mainly in esteri®ed and bound forms, which appeared to be relatively well preserved in the sediments. This surprising stability could be attributed to: (i) the weak reducing properties of sul®des towards these hydroperoxides or (ii) the protection of these compounds in intact phytoplanktonic debris. Destruction of hydroperoxides and allylic rearrange-ment of6-5a-hydroperoxysterols (to the corresponding
5-7a/b- derivatives) takes place at the bottom of the core
analyzed. The detection of high amounts of 5,6-epoxy-24-ethylcholestan-3b-ol (52 ng/g dry sediment at 3.5 cm depth), (resulting probably from the oxidation of 24-ethylcholest-5-en-3b-ol by hydroperoxides in the absence of molecular oxygen) strongly suggests that hydroperoxysterols may play a role in the degradation of organic matter in anoxic sediments. Due to their greater stability in sediments,4-6a/b-hydroperoxysterols will be more reliable in situ markers
of type II photodegradation processes (i.e. those involving singlet oxygen) than6-5a-hydroperoxysterols. Dehydration of the reduced sterol photoproducts described in the present work might constitute a potential source of steratrienes, which are often present in the sediments.#2000 Elsevier Science Ltd. All rights reserved.
Keywords:Type II photodegradation processes; Sterols; Senescent phytoplankton; Recent sediments; Tracers; Hydroperoxides
1. Introduction
Steroidal alcohols (sterols) are comparatively stable in sediments and hence they have a long geological record (Gagosian et al., 1982). Moreover, they possess struc-tural features, such as positions of double bonds, nuclear methylation and patterns of side-chain alkylation, which can be restricted to a few groups of organisms (Volk-man, 1986). Consequently, sterols in seawater act as good tracers of biogenic material and food chains
(Steudler et al., 1977) and in Recent sediments they constitute excellent biomarkers for tracing diagenetic transformations (Mackenzie et al., 1982).
Though sterols are generally considered to be more stable than the majority of the organic compounds pro-duced by phytoplankton (Gagosian et al., 1982), only a small part of the sterols produced in the euphotic zone (less than 1% in the equatorial Atlantic Ocean; Gago-sian et al., 1982) reach the sediment. This disappearance is generally considered to be the result of biodegrada-tion (Johannes and Satomi, 1966), bacterial degradabiodegrada-tion (Iturriaga, 1979) and coprophagy (Smayda, 1969).
We have recently demonstrated that 5-stenols (the
predominant biogenic sterols in most environments) can
0146-6380/00/$ - see front matter#2000 Elsevier Science Ltd. All rights reserved. P I I : S 0 1 4 6 - 6 3 8 0 ( 9 9 ) 0 0 1 5 6 - 4
www.elsevier.nl/locate/orggeochem
* Corresponding author. Tel.: 4-9182-9623; fax: +33-4-9182-6548.
also be quickly photodegraded in senescent cells of phytoplankton (Rontani et al., 1997,1998). This photo-degradation, which involves type II (i.e. involving sing-let oxygen) photoprocesses (Rontani et al., 1997), results in the formation of 5a- and 6a/b-hydroperoxides. The surprising stability of these hydroperoxides observed in dead phytoplanktonic cells (no degradation products detected after 21 days of incubation at 20C) strongly
suggests that in the marine environment, measurable amounts of these photoproducts can survive the long transport times to the ocean ¯oor.
In the present work, we provide evidence of the pre-sence of such photoproducts in Recent sediments of the Gulf of Fos (Mediterranean Sea) in order to determine: (i) whether the photodegradation processes participate eciently in the degradation of 5-stenols in the
euphotic zone of the oceans, and (ii) whether these photoproducts are suciently stable to play the role of tracers of photodegradation processes.
2. Experimental
2.1. Sediment sampling
The top layer (20 cm) of the sediment was collected with a manual corer under 9 m of water. The sealed cores were maintained in isotherm bags (containing dry ice) during their transportation to Marseilles, where they were stored atÿ20C until analysis. Station 35 (as
numbered in the SEDIFOS program) in Carteau Bay (Gulf of Fos, Mediterranean Sea) was chosen as the site for this study. Its suitability for this work was based on: (i) minor variance in particle size distribution with depth (Bonin et al., 1999) suggesting minimal variation of sedimentary conditions over the pro®le studied, (ii) relatively high concentrations of chlor-ophyll phytyl sidechain photodegradation products (ratio photoproducts/unchanged chlorophyll phytyl chain=0.3; Rontani et al., 1996) and (iii) very weak irradiance at the water-sediment interface owing to the resuspension of sediments (Barranguet, 1994). The oxic layer of the sediment was 4 mm deep (Bonin et al., 1999) and the sedimentation rate of this zone is approximately 0.5±1 cm yearÿ1(Grenz et al., 1990). The sediment
con-tains a macrofaunal assemblage characteristic of muddy sand. Polychaetes dominated the benthic macrofauna (70%) and crustaceans were the second most dominant group (25%). More than 80% of the organisms were located in the upper 4 cm of sediment (Gilbert et al., 1998).
2.2. Treatment of sediments
All the manipulations were carried out with foil-cov-ered vessels in order to exclude photochemical artifacts.
2.3. Extraction
The wet sediment slices (cut under dim light) were extracted ultrasonically with isopropanol:hexane (4:1, v/v) (Leeuw et al., 1977). Hexane extracts were com-bined and the isopropanol/water phase was ®ltered, con-centrated under vacuum and then extracted three times with chloroform. The chloroform and hexane extracts were combined, dried over anhydrous Na2SO4, ®ltered
and concentrated by rotary evaporation (at 40C).
2.4. Reduction of hydroperoxides
Lipidic extracts of sediments were reduced in metha-nol (25 ml) by excess NaBH4 (10 mg/mg of extract)
using magnetic stirring (15 min at 0C) (Teng et al.,
1973). During this treatment steroidal ketones are also reduced and the possibility of some ester cleavage cannot be excluded.
2.5. Alkaline hydrolysis
Saponi®cation was carried out on both reduced and non-reduced samples (lipid extracts or sediments). After reduction, 25 ml of water and 2.8 g of potassium hydroxide were added and the mixture was directly saponi®ed by re¯uxing for 2 h. In the case of non-reduced samples, an additional 25 ml of methanol was added before saponi®cation. After cooling, the contents of the ¯ask were extracted three times with hexane (sediments were ®ltered through Whatman qualitative ®lters before extraction). The combined hexane extracts were dried over anhydrous Na2SO4, ®ltered and concentrated.
2.6. Hydrogenation
Lipid extracts were hydrogenated overnight under magnetic stirring in methanol with Pd/CaCO3 (10±20
mg/mg of extract) (Aldrich) as a catalyst. After hydro-genation, the catalyst was removed by ®ltration and the ®ltrate was concentrated by rotary evaporation.
2.7. Derivatization
After evaporation of solvents, the residues were taken up in 400 ml of a mixture of pyridine and BSTFA (Supelco) (3:1, v/v) and silylated for 1 h at 50C. After
evaporation to dryness under nitrogen, the residues were taken up in ethyl acetate and analyzed by gas chromato-graphy/electron impact mass spectrometry (GC/EIMS).
2.8. Identi®cation and quanti®cation of sterols and their oxidation products
and quanti®ed (calibration with external standards) by GC/EIMS. For low concentrations or in the case of coelutions, quanti®cation was assessed by selected ion monitoring (SIM) with the diagnostic ions at [Mÿ90]+ for 7-hydroxysterols (Rontani et al., 1997), at [Mÿ143]+ for 6-hydroxysterols (Harvey and Vouros,
1979), at [Mÿ161]+for 3,5,6-triols, at [Mÿ18]+for 5-hydroxystanols, at [Mÿ15]+ for 7-hydroxystanols and
at M+for the other compounds.
GC/EIMS analyses were carried out with a HP 5890 series II plus gas chromatograph connected to a HP 5972 mass spectrometer. The following operative con-ditions were employed: 30 m0.25 mm (i.d.) capillary column coated with HP5 (Hewlett Packard) (®lm thick-ness, 0.25mm); oven temperature programmed from 60 to 130C at 30C minÿ1and then from 130 to 300C at
4C minÿ1; carrier gas (He) pressure maintained at 1.04
bar until the end of the temperature program and then programmed from 1.04 to 1.5 bar at 0.04 bar minÿ1;
injector (on column with a retention gap) temperature, 50C; electron energy, 70 eV; source temperature,
170C; cycle time, 1.5 s.
2.9. Standard compounds
Cholest-5-en-3b-ol, 24-methylcholest-5-en-3b-ol, 24-ethylcholest-5-en-3b-ol and 24-ethylcholesta-5,24(28)E
-dien-3b-ol were purchased from Aldrich and Sigma. 5a -and 6a/b-Hydroperoxides were obtained after photo-sensitized oxidation of the corresponding5-stenols in
pyridine in the presence of haematoporphyrin as sensi-tizer (Nickon and Bagli, 1961). Allylic rearrangement of 5a -hydroperoxides to 7a-hydroperoxides and epimer-ization of the latter to 7b-hydroperoxides was obtained at room temperature in chloroform (Teng, 1990). Sub-sequent reduction of these dierent hydroperoxides in methanol with excess NaBH4aorded the
correspond-ing diols. Hydrogenation of these diols was carried out with Pd/CaCO3 as catalyst. Treatment of 5-stenols
with meta-chloroperoxy-benzoic acid in dry methylene chloride yielded a mixture of 5a,6a-and 5b,6b-epoxides. Heating of these epoxides in the presence of water aorded the corresponding 3b,5a,6b-triols (Holland and Diakow, 1979).
3. Results and discussion
Analyses of the top layer of the sediments revealed the presence of cholest-5-en-3b-ol (cholesterol), 24-methyl-cholesta-5,22E-dien-3b-ol and 24-ethylcholest-5-en-3b-ol (sitosterol) as major 5-stenols. These compounds are
present mainly in ``bound'' (i.e. non-solvent-extractable) form. Smaller amounts of 24-methylcholest-5-en-3b-ol
(campesterol), 24-methylcholesta-5,24(28)-dien-3b-ol and 24-ethylcholesta-5,24(28)E-dien-3b-ol (fucosterol) were also detected. 24-Methylcholesta-5,22E-dien-3b-ol and 24-methylcholesta-5,24(28)-dien-3b-ol, not being commercially available, were identi®ed by comparison of their electron impact mass spectra with mass spectral data described in the literature (Lee et al., 1979; Leeuw et al., 1983).
It was previously determined that the major part of the chlorophyll present in the sediments of station 35 in Carteau Bay originated from diatoms (Barranguet, 1994). The presence of 24-methylcholesta-5,22E-dien-3b -ol and 24-methylch-olesta-5,24(28)-dien-3b-ol con®rms that diatoms are important constituents of the phyto-plankton in that environment (Volkman, 1986; Volk-man et al., 1998). Sitosterol may also have a diatom origin (Volkman, 1986).
In the sediment analyzed, every sterol with a5-double
bond was accompanied by the equivalent 4-3-stenone,
5a(H)-3-stanone and 5a(H)-3b-stanol. The presence of these compounds can be attributed to the well known microbiological conversion of5-stenols as demonstrated
during several incubation experiments with labelled cho-lesterol in recent sedimentary environments (Gaskell and Eglinton, 1975; Mermoud et al., 1984).
GC/EIMS analyses allowed detection of signi®cant amounts of4-3b,6- and
5-3b,7- epimeric unsaturated
diols probably arising from cholesterol and sitosterol in the sediments (Fig. 1A). Mass spectra of (disilylated)5
-3b,7-diols exhibit strong [Mÿ HOSiMe3]+peaks (Fig.
2A), whereas those of (disilylated) 4-3b,6-diols show characteristic [Mÿ 143]+ ions associated with A-ring
loss (Fig. 2B). Photooxidation of5-stenols in senescent phytoplanktonic cells involves type II (i.e. involving singlet oxygen) photoprocesses and gives rise to6-5a -and 4-6a/b- allylic hydroperoxides (Rontani et al.,
1997).6-5a-Hydroperoxides are relatively unstable and
may undergo allylic rearrangement to5-7a
-hydroper-oxides, which in turn epimerize to 5-7b
-hydroper-oxides (Smith, 1981). Such conversions can occur during the photoreaction itself or during product isolation and analysis (Korytowski et al., 1992). We have previously demonstrated that allylic rearrangement does not occur signi®cantly in dead cells of phytoplankton nor during the extraction, reduction and saponi®cation processes employed (Rontani et al., 1997). It occurs quantitatively in the chromatograph during the GC analysis of (dis-ilylated)6-3b,5a-diols when an on column injector is
used, whereas splitless injection results in a complete desilylation (Rontani et al., 1997). To avoid this incon-venient gas chromatographic allylic rearrangement, sterol photoproducts must be hydrogenated before GC/ EIMS analyses (Rontani et al., 1998).
Dierent treatments (Fig. 3) were applied to a slice of super®cial sediments (3±4 cm) in order to determine: (i) whether 5-3b,7-diols arise from allylic rearrangement
of the corresponding6-3b,5a-diols during GC analyses or are actually present in the sediment, and (ii) whether the sediments contain hydroperoxides or alcohols. Comparison of the results obtained in extracts E1, E2
and E3(Table 1) clearly show that most of the4-3b
,6-and5-3b,7-diols are in esteri®ed and bound (i.e. non-extractable) forms. After hydrogenation of double bonds (extracts E4), we detected mainly saturated 3b,5a
-diols (Fig. 1B) with lesser amounts of saturated 3b,6a/b -diols (Table 1). Saturated 3b,5a-diols are silylated only at position 3 by the mixture BSTFA/pyridine and easily lose a water molecule upon electron impact. Conse-quently, their EI mass spectra (Fig. 2C) closely resemble those of4-stenols and comparison of retention times is
needed to identify them unambiguously. Saturated 3b,7a/b-diols are present in extracts E4 at trace level
(Table 1) and represent only 2% of the 5-3b,7-diols content of the corresponding extract E3.It appears that
the most of the5-3,7-diols detected in non-hydrogenated extracts is formed during GC injection and that the allylic
rearrangement of 5-hydroperoxy- and/or 5 -hydro-xysterols is not signi®cant in the ®rst 3 cm of the sediment. The quantities of unsaturated 3b,6- and 3b,7-diols decrease considerably if reduction with NaBH4is
omit-ted before the alkaline hydrolysis (comparison between extracts E3 and E5) (Fig. 1A and C, Table 1). This
decrease is accompanied by increased amounts of the corresponding5-3b-ol-7-ones and3,5-7-ones (Fig. 4, Table 1) in extracts E5. The formation of these ketones,
which are well known products of biradical oxygen autooxidation of sterols (Smith, 1981), can be attributed to: (i) the involvement of autooxidative processes in the water column or in the aerobic zone of the sediments aording5-7a/b-hydroperoxides, (ii) the oxidation of
the corresponding sterols during the saponi®cation pro-cedure employed, or (iii) the degradation of 6-5a
-hydroperoxysterols during alkaline hydrolysis (Balci, 1981) (Fig. 5).
Fig. 3. Dierent treatments applied to Carteau Bay sediments.
Table 1
Concentrations (ng/g dry sedimenta) of sitosterol oxidation products detected in the extracts obtained after dierent treatments (as described in Fig. 3) of Carteau Bay sediments (slice 3±4 cm)
Compound E1 E2 E3 E4 E5
24-Ethylcholest-5-en-3b,7a/b-diols 5 20 31 trb 7
24-Ethylcholest-4-en-3b,6a/b-diols tr 12 18 tr tr
24-Ethylcholesta-3b,7a/b-diols ndc nd nd tr nd
24-Ethylcholesta-3b,5a-diol tr tr 7 35 tr
24-Ethylcholesta-3b,6a/b-diols tr tr tr 15 tr
24-Ethyl-5,6-epoxycholestan-3b-ols (b+a) tr 52 tr tr 21
24-Ethylcholesta-3b,5a,6b-triol tr 7 tr tr 7
24-Ethyl-3b-hydroxycholest-5-en-7-one nd nd nd nd 9 (ndd)
24-Ethylcholesta-3,5-dien-7-one nd nd nd nd 17 (25d)
a Accuracy estimated to be 5 ng/g dry sediment. b tr=traces (amounts <5 ng/g dry sediment). c nd=not detected.
The ®rst hypothesis can be discarded. In fact, if5 -3b-ol-7-ones and 3,5-7-ones or their parent
5-7a/b
-hydroperoxides were initially present in the sediment analyzed, we would recover relatively large amounts of saturated 3b,7-diols and 7-stanols after reduction and hydrogenation. This is not the case, since the quantities
of saturated 3b,7-diols detected in extracts E4represent
only a small proportion (4%) of the amounts of5-3b
-ol-7-ones present in extracts E5and we failed to detect
7-stanols. Metal ions present in the sediments can pro-mote autooxidation of sterols during the hot saponi®-cation procedure employed. This hypothesis can be
Fig. 4. Electron impact mass spectra of (A) 3b -hydroxy-24-ethylcholest-5-en-7-one (silylated) and (B) 24-ethylcholesta-3,5-dien-7-one.
Fig. 6. Electron impact mass spectra of (A) 5,6-epoxy-24-ethylcholestan-3b-ol (silylated) and (B) 24-ethylcholesta-3b,5a,6b-triol (disilylated at positions 3 and 6).
discarded since the production of 5-3b-ol-7-ones was also observed after an attempt of cold saponi®cation (overnight) carried out under nitrogen (Table 1). The lack of 3,5-7-ones observed after this treatment
demonstrates that the dehydration of 5-3b-ol-7-ones
during saponi®cation requires heating. The greater part of 5-3b-ol-7-ones and
3,5-7-ones results probably
from the degradation of the corresponding 6-5a -hydroperoxides during the hot saponi®cation treatment (Fig. 5). These dierent results seem to indicate thatthe degradation of sterol hydroperoxides operates slowly in the sediment. Though sul®des produced by sulfate redu-cing bacteria in the anoxic zone of the sediments are known to easily reduce hydroperoxides (Mihara and Tateba, 1986), we observed that these species are not very ecient reducers of 5a-hydroperoxysterols (Ron-tani 1999, (unpubl. data). This result could explain the surprising preservation of these hydroperoxides observed in the core analyzed.
Grimalt et al. (1991) previously detected dierent cholesta-3,5-dien-7-ones in sediments of the Santa Olalla Lagoon. These authors considered that the formation of these compounds is mediated via hydroperoxides and
attributed their production to water column oxidation processes. Our results support these conclusions and sug-gest that the dienones detected in the surface sediments of this area probably resulted from the degradation during the treatment (Soxhlet extraction + alkaline hydrolysis) of the corresponding 6-5a-hydroperoxysterols pro-duced in the euphotic zone of the water column.
The presence of hydroperoxides in the sediment is supported by the detection of relatively high amounts of isomeric 5,6-epoxy-24-ethylcholestan-3b-ols [with a large proportion of the b isomer (ratio b/a=8)] in extracts E2 (Table 1, Fig. 6A). It is well known that
attack of sterols by the ®rst-formed sterol hydroper-oxides yields both isomeric 5,6-ephydroper-oxides (5a,6a-epoxides and 5b,6b-epoxides in the ratio 1:8 to 1:11; Smith and Kulig, 1975). The formation of these epoxides was pre-viously observed after heating of aqueous dispersions of cholesterol under aerobic conditions (Chicoye et al., 1968; Smith and Kulig, 1975) or under nitrogen in the presence of 5a- and 7a/b-hydroperoxides (Smith and Kulig, 1975). It is important to note that under aerobic conditions 7-hydroperoxides constitute major sterol degradation pro-ducts (ratio 7-hydroperoxides/5,6-epoxides=18) (Smith
Fig. 7. Depth pro®les of cholesterol and its photoproducts (quanti®ed in extracts E3) in sediment core sections of Carteau Bay. (* The proportions of5-3b,7a/b-diols and
and Kulig, 1975). As was demonstrated above after hydrogenation, 7-derivatives are present in very low amounts in the slice of sediment analyzed (ratio satu-rated 3b,7-diols/5,6-epoxides=0.02). Consequently, we attributed the production of these epoxides to oxidation of 5-stenols by their own 5- and 6-hydroperoxides under anaerobic conditions. It is generally considered that in sediments, spontaneous oxidation of Fe(II) and other metal ions at the surface of oxide minerals generates a superoxide radical, which decomposes to H2O2.
(Sawyer, 1991). Although attack of H2O2on5-stenols
can constitute an additional source of 5,6-epoxides (Smith et al., 1978), the absence of signi®cant amounts of 7-hydroperoxides does not support the involvement of such a pathway. We also detected 24-ethylcholesta-3b,5a,6b-triol (Table 1; Fig. 6B), arising probably from hydration of the corresponding 5,6-epoxides (Smith, 1981). We failed to detect signi®cant amounts of 5,6-epoxycholestan-3b-ol and cholesta-3b,5a,6b-triol in the dierent extracts analyzed.
Signi®cant amounts of cholesta-3b,5a-diols and traces of cholesta-3b,6a/b-diols were also detected in extracts E3 (Table 1). We attributed the production of these
compounds to an anaerobic microbial hydrogenation of the corresponding 5- and 6-hydroxysterols similar to that responsible for the conversion of5-stenols to
sta-nols (Gagosian et al., 1980; Leeuw and Baas, 1986). It can be noted that the quantities of cholesta-3b,5a-diols decreases if the reduction with NaBH4is omitted before
alkaline hydrolysis (comparison between extracts E3and
Fig. 8. Depth pro®les of sitosterol and its photoproducts (quanti®ed in extracts E3) in sediment core sections of Carteau Bay. (* The proportions of5-3b,7a/b-diols and
6-3b,5a-diols were determined after hydrogenation). Table 2
Apparent ®rst-order degradation rate constants and correlation coecients for sterolsaand their photodegradation productsb
Compound k(yearÿ1) r2
24-Ethylcholest-5-en-3b-ol 0.20 0.76 24-Ethylcholest-6-en-3b,5a-diol 0.20 0.81 24-Ethylcholest-4-en-3b,6a-diol 0.26 0.82 24-Ethylcholest-4-en-3b,6b-diol 0.27 0.86
a The values were not calculated for cholesterol and its pho-toproducts owing to the presence of a large contribution of this sterol, probably of macrofaunal origin, in the ®rst centimetres of sediment, which biased considerably the comparison of the degradation rate constants.
E5; Table 1). Some hydrogenation seems to take place
before reduction of the 5-hydroperoxy- group aording cholestan-3b-ol-5a-hydroperoxides. Alkaline hydrolysis of such compounds must result in the formation of ter-tiary alkoxyl radicals, which can either abstract a hydrogen atom (from other molecules) to give the cor-responding diols or undergob-fragmentation processes leading to A- or B- ring cleavage.
Cholesterol and sitosterol photoproducts were quan-ti®ed in extracts E3of dierent core sections covering 18
cm of sediment. Figs. 7 and 8 depict the concentration pro®les with depth of sterols and their photoproducts in the core sections investigated. Concentrations of 6
-3b,5a- and4-3b,6a/b-diols quickly decrease with depth
(Figs. 7 and 8). The ®rst-order degradation rates of 24-ethylcholest-6-en-3b,5a-diol and 24-ethylcholest-4-en-3b,6a/b-diols (calculated for the ®rst 10 cm of sediment) are relatively close to that of their parent sterol (Table 2). Consequently, information relative to the photo-degradation state of phytoplankton carried by the ratio 5- or 6-hydroxysterols/intact sterols must not be strongly altered during early diagenesis. However, hydrogenation experiments showed that at the bottom of the core6
-3b-ol-5a-hydroperoxides and/or 6-3b,5a-diols decay via allylic rearrangement to their5-3b,7a/b-derivatives
(Figs. 7 and 8). This can be attributed to: (i) a progressive destruction of the structural integrity of cell membranes with increasing core depth or (ii) the well known e-cient microbial degradation of unsaturated fatty acids in sediments (Cranwell, 1976). The absence of allylic rear-rangement of6-5a-hydroperoxysterols in phytodetritus was in fact previously attributed to the high9
unsatu-rated fatty acid content of phytoplanktonic cell mem-branes (Rontani et al., 1997). Korytowski et al. (1992) have previously observed that 5a-hydroperoxides are more stable in membranes containing unsaturated phospholipids than in those containing saturated phos-pholipids. This stability was attributed either to hydro-gen bonding between the unsaturated fatty acyl chain of phospholipids and 5a-hydroperoxides which could hin-der the allylic rearrangement (Nakano et al., 1980), or to dierences of polarity in the carbon 7±10 zone of the fatty acyl chain (where sterols tend to localise in phos-pholipid/sterol bilayers; MacIntosh, 1978).
7-Hydroxysterols cannot constitute speci®c tracers of photooxidative processes, since they may also be formed by autoxidation (Van Lier and Smith, 1970; Smith, 1981) and bacterial degradation (Krischenoski and Kieslich, 1993; Mahato and Garai, 1997) of 5-stenols. Conse-quently, due to their relative stability in sediments, 4 -6/-hydroperoxy- and 4-6/-hydroxysterols [which are produced in higher proportion in cell membranes than in homogeneous solutions (Korytowski et al., 1992)] may be considered to be better tracers of photooxidative pro-cesses than 6-5-hydroperoxy- and
6-5 -hydro-xysterols.
The destruction of the integrity of cell membranes must also favour hydroperoxide degradation. The decrease of 5,6-epoxides and 3b,5a,6b-triol concentra-tions observed at the bottom of the core (Fig. 9) sup-ports this hypothesis; however, this decrease can also be explained by a simple reduction of interactions between
5-stenols and their corresponding hydroperoxides after
destruction of cell membranes.
Photoproducts of cholesterol are present throughout the core in lower proportion than those of sitosterol (Figs. 7 and 8). This can be explained by the fact than (i) zooplankton faecal pellets are considered to be impor-tant sources of cholesterol in marine sediments (Volk-man, 1986) and (ii) photodegradation rates are slower in copepod faecal pellets than in senescent phytoplank-tonic cells (Nelson, 1993). Such an origin of cholesterol and its photoproducts could also explain the lack of 5,6-epoxycholestan-3b-ol in the sediments. Indeed, the modi®cation of the structural integrity of phytoplank-tonic cells during the feeding of zooplankton might result in a decrease in the proximity of hydroperoxides to their parent sterol. The very high concentrations of cholesterol observed in the sublayer (0±2 cm) (Fig. 7) may also be attributed to the macrofauna (Parrish et al.,
1996; Sun et al., 1998), which is mainly located in the upper 4 cm of sediment (Gilbert et al., 1998).
Several authors have mentioned the presence of ster-atrienes with the three double bonds in the AB-ring system in sediments (Gagosian and Farrington, 1978; Wardroper, 1979; Grimalt et al., 1991). The formation of these compounds has been attributed either to a dehydrogenation of the 3,5-steradienes (Leeuw and
Baas, 1986), or to a microbial hydroxylation of5
-ste-nols (Holland and Diakow, 1979) followed by dehydra-tion (Gagosian et al., 1980). Dehydradehydra-tion of the reduced sterol photoproducts described in the present work con-stitutes another potential source of these trienes (Fig 10).
4. Conclusions
Signi®cant amounts of photoproducts resulting from the addition of 1O
2 (Frimer, 1979) to the 5-double
bond of cholesterol and sitosterol have been detected in the sediment core analyzed. These results show that photodegradation of 5-stenols in senescent or dead phytoplanktonic cells plays a role in the degradation of these compounds in the marine environment. Though 24-methylcholesta-5,22E-dien-3b-ol, 24-methylcholesta-5,24(28)-dien-3b-ol and 24-ethylcholesta-5,24(28)E -dien-3b-ol are also present in this sediment, we failed to
detect signi®cant amounts of their photoproducts. This can be attributed to the addition of1O
2to the double bond
localised on the side-chain of these sterols, which likely leads to the production of unidenti®ed photoproducts.
The surprising preservation of hydroperoxysterols observed throughout the core could be attributed to the weak reducing properties of sul®des towards these compounds. These photoproducts, which are present mainly in esteri®ed and bound forms, could also be protected in intact phytoplanktonic membranes of well-silici®ed diatoms. Indeed, it is generally considered that most of the bound materials observed in sediments represent the contents of intact biological debris (Cran-well, 1978; Sun et al., 1993).
6-5a-Hydroperoxysterols (or the corresponding
6
-5a-hydroxysterols) are potential type II photodegradation markers, not only because they are the major products of singlet oxygen attack on the steroidal5-3b- system, but also because biological functionalization of steroids at C-5 is rare. If these compounds are particularly stable in phy-todetritus (Rontani et al., 1997, 1998), they decay slowly in the sediment to their corresponding5-7a/b- deriva-tives, which are not selective markers. Microbial C-6 hydroxylation of some4- steroidal compounds is known
(for reviews see Mahato and Majumdar, 1993; Mahato and Garai, 1997) but results only to the formation ofb
epimers. Consequently,4-6a/b-hydroperoxysterols (or
the corresponding 4-6a/b-hydroxysterols) may be considered as more reliable in situ markers of type II photodegradation processes than6-5a-hydroperoxides
(Korytowski et al., 1992).
High amounts of isomeric 5,6-epoxy-24-ethylchole-stan-3b-ols have been detected in the sediments ana-lyzed. The formation of these compounds was attributed to the oxidation of 24-ethylcholest-5-en-3b-ol by hydroperoxides in the absence of molecular oxygen. These results suggest that hydroperoxysterols can be suciently stable in sediments to play a role in the degradation of organic matter under anoxic conditions. In future studies, we intend to use HPLC/MS±MS ana-lyses in order to con®rm these results and to directly prove the existence of sterol hydroperoxides in the sedi-mentary material.
Steroidal 4-3b,6a/b-diols could be associated with some photoproducts of the chlorophyll phytyl chain (Cuny and Rontani, 1999) and monounsaturated fatty acids (Rontani, 1998) to constitute a ``pool'' of useful indicators of photooxidative alterations of phyto-plankton. This ``pool'' could be used to validate photo-induced oxidative processes in Recent sediments, providing new ways to gain information on current environmental problems related to ozone depletion.
Acknowledgements
Thanks are due to Drs. D. Massias and V. Grossi for the sampling of sediments and to Mr. M. Paul for his careful reading of the English. Drs. R.M.K. Carlson and M.±Y. Sun are sincerely thanked for their very constructive reviews of this paper.
Associate EditorÐJ.K. Volkman
References
Balci, M., 1981. Bicyclic endoperoxides and synthetic applica-tions. Chemical Review 81, 91±108.
Barranguet, C., 1994. RoÃle des microphytes benthiques dans l'oxygeÂnation de l'interface eau-seÂdiment de deux systeÁmes conchylicoles MeÂditerraneÂens: Golfe de Fos, Etang de Thau. Ph.D. thesis, University of Aix-Marseille II, Marseille. Bonin, P., Omnes, P., Chalamet, A., 1999. Simultaneous
occurrence of denitri®cation and nitrate ammoni®cation in sediments of the French Mediterranean Coast. Hydro-biologia 389, 169±182.
Chicoye, E., Powrie, W.D., Fennema, O., 1968. Isolation and characterisation of cholesterol-5b,6b-oxide from an aerated aqueous dispersion of cholesterol. Lipids 3, 335±339. Cranwell, P.A., 1976. Decomposition of aquatic biota and
sediment formation: lipid components of 2 blue-green algal species and of detritus resulting from microbial attack. Freshwater Biology 6, 481±488.
Cranwell, P.A., 1978. Extractable and bound lipid components in a freshwater sediment. Geochimica et Cosmochimica Acta 42, 1523±1532.
Cuny, P., Rontani, J.±F, 1999. On the widespread occurrence of 3-methylidene-7,11,15-trimethylhexadecan-1,2-diol in the marine environment: a speci®c isoprenoid marker of chlor-ophyll photodegradation. Marine Chemistry 65, 155±165. Frimer, A.A., 1979. The reaction of singlet oxygen with ole®ns:
the question of mechanism. Chemical Review 79, 359±387. Gagosian, R.B., Farrington, J.W., 1978. Sterenes in surface
sediments from the Southwest African shelf and slope. Geo-chimica et CosmoGeo-chimica Acta 42, 1091±1101.
Gagosian, R.B., Smith, S.O., Lee, C., Farrington, J.W., Frew, N.M., 1980. Steroid transformations in recent marine sedi-ments. In: Douglas, A.G., Maxwell, J.R. (Eds), Advances in Organic Geochemistry 1979 Pergamon, pp. 407±419. Gagosian, R.B., Smith, S.O., Nigrelli, G.E., 1982. vertical
transport of steroid alcohols and ketones measured in a sediment trap experiment in the equatorial Atlantic Ocean. Geochimica et Cosmochimica Acta 46, 1163±1172.
Gaskell, S.J., Eglinton, G., 1975. Rapid hydrogenation of sterols in a contemporary lacustrine sediment. Nature 254, 209±211. Gilbert, F., Stora, G., Bonin, P., 1998. In¯uence of
bioturba-tion on denitri®cabioturba-tion activity in Mediterranean coastal sediments: an in situ experimental approach. Marine Ecology Progress Series 163, 99±107.
Grenz, C., Hermin, M.-N., Baudinet, D., Daumas, R., 1990.In situbiochemical and bacterial variation of sediments enri-ched with mussel biodeposits. Hydrobiologia 207, 153±160. Grimalt, J.O., Yruela, I., Saiz-Jimenez, C., Toja, J., de Leeuw,
J.W., Albaiges, J., 1991. Sedimentary lipid biogeochemistry of an hypereutrophic alkaline lagoon. Geochimica et Cos-mochimica Acta 55, 2555±2577.
Holland, H.L., Diakow, P.R.P., 1979. Microbial hydroxylation of steroids. 5. Metabolism of androst-5-en-3,17-dione and related compounds by Rhizopus arrhizus ATCC 11145. Canadian Journal of Chemistry 57, 436±440.
Harvey, D.J., Vouros, P., 1979. In¯uence of the 6-trimethylsilyl group on the fragmentation of the trimethylsilyl derivatives of some 6-hydroxy- and 3,6-dihydroxy- steroids and related compounds. Biomedical Mass Spectrometry 6, 135±143. Iturriaga, R., 1979. Bacterial activity related to sedimentary
particulate matter. Marine Biology 55, 157±169.
Johannes, R.E., Satomi, M., 1966. Composition and nutritive value of faecal pellets of a marine crustacean. Limnology and Oceanography 11, 191±197.
Korytowski, W., Bachowski, G.J., Girotti, A.W., 1992. Photo-peroxidation of cholesterol in homogeneous solution, iso-lated membranes, and cells: comparison of the 5-aand 6a -hydroperoxides as indicators of singlet oxygen intermediacy. Photochemistry and Photobiology 56, 1±8.
Krischenoski, D., Kieslich, K., 1993. Two novel microbial conversion products of progesterone derivatives. Steroids 58, 278±281.
Lee, C., Farrington, J.W., Gagosian, R.B., 1979. Sterol geo-chemistry of sediments from the western North Atlantic Ocean and adjacent coastal areas. Geochimica et Cosmochi-mica Acta 43, 35±46.
Leeuw, J.W.de, Simoneit, B.R.T., Boon, J.J., Rijpstra, W.I.C., de Lange, F., van der Leeden, J.C.W. et al., 1975. Phytol derived compounds in the geosphere. In: Campos, R., Goni, J. (Eds.), Advances in Organic Geochemistry 1975. Enadimsa, Madrid, pp. 61±79.
Leeuw, J.W. de, Rijpstra, W.I.C., Schenck, P.A., Volkman, J.K., 1983. Free, esteri®ed and residual bound sterols in Black Sea Unit I sediments. Geochimica et Cosmochimica Acta 47, 455±465.
Mahato, S.B., Majumdar, I., 1993. Current trends in microbial steroid biotransformation. Phytochemistry 34, 883±898. Mahato, S.B., Garai, S., 1997. Advances in microbial steroid
biotransformation. Steroids 62, 332±345.
MacIntosh, T.J., 1978. The eect of cholesterol on the structure of phosphatidylcholine bilayers. Biochimica et Biophysica Acta 513, 43±58.
Mackenzie, A.S., Lamb, N.A., Maxwell, J.R., 1982. Steroid hydrocarbons and the thermal history of sediments. Nature 295, 223±226.
Mermoud, F., WuÈnsche, L., Clerc, O., GuÈlacËar, F.O., Buchs, A., 1984. Steroidal ketones in the early diagenetic transfor-mations of5-sterols in dierent types of sediments. Organic Geochemistry 6, 25±29.
Mihara, S., Tateba, H., 1986. Photosensitized oxygenation reactions of phytol and its derivatives. Journal of Organic Chemistry 51, 1142±1144.
Nakano, M., Sugioka, K., Nakamura, T., Oki, T., 1980. Inter-action between an organic hydroperoxide and an unsatu-rated phospholipid and alpha-tocopherol in model membranes. Biochimica et Biophysica Acta 619, 274±286. Nelson, J.R., 1993. Rates and possible mechanism of
light-dependent degradation of pigments in detritus derived from phytoplankton. Journal of Marine Research 51, 155±179. Nickon, A., Bagli, J.F., 1961. Reactivity and geochemistry in
allylic systems. I. Stereochemistry of photosensitized oxyge-nation of monoole®ns. Journal of the American Chemical Society 83, 1498±1508.
Parrish, C.C., Yang, Z., Lau, A., Thompson, R.J., 1996. Lipid composition ofYoldia hyperborea(Protobranchia),Nephthys ciliata (Nephthyidae) and Artacama proboscidea (Ter-ebellidae) living at sub-zero temperatures. Comparative Bio-chemistry and Physiology 114B, 59±67.
Rontani, J.±F, 1998. Photodegradation of unsaturated fatty acids in senescent cells of phytoplankton: photoproduct structural identi®cation and mechanistic aspects. Journal of Photochemistry and Photobiology A114, 37±44.
Rontani, J.-F., Raphel, D., Cuny, P., 1996. Early diagenesis of the intact and photooxidized chlorophyll phytyl chain in a Recent temperate sediment. Organic Geochemistry 24, 825±832.
Rontani, J.-F., Cuny, P., Aubert, C., 1997. Rates and mechan-ism of light-dependent degradation of sterols in senescing cells of phytoplankton. Journal of Photochemistry and Pho-tobiology A111, 139±144.
Rontani, J.-F., Cuny, P., Grossi, V., 1998. Identi®cation of a pool of lipid photoproducts in senescent phytoplanktonic cells. Organic Geochemistry 29, 1215±1225.
Sawyer, D.T., 1991. Oxygen Chemistry. Oxford University Press, Oxford, New York.
Smayda, T.J., 1969. Some measurements of the sinking rates of faecal pellets. Limnology and Oceanography 14, 621±625. Smith, L.L., 1981. The Autoxidation of Cholesterol. Plenum
Press, New York pp.119±132.
Smith, L.L., Kulig, M.J., 1975. Sterol metabolism XXXIV. On the derivation of carcinogenic sterols from cholesterol. Can-cer Biochemistry and Biophysics 1, 79±84.
Smith, L.L., Kulig, M.J., Miller, D., Ansari, G.A.S., 1978. Oxidation of cholesterol by dioxygen species. Journal of the American Chemical Society 100, 6206±6211.
Steudler, P.A., Schmitz, F.J., Ciereszko, L.S., 1977. Chemistry of coelenterates. Sterol composition of some predator±prey pairs on coral reefs. Comparative Biochemistry and Physiol-ogy 56B, 385±392.
Sun, M.-Y., Wakeham, S.G., 1998. A study of oxic/anoxic eects on degradation of sterols at the simulated sediment± water interface of coastal sediments. Organic Geochemistry 28, 773±784.
Sun, M.-Y., Lee, C., Aller, R.C., 1993. Laboratory studies of oxic and anoxic degradation of chlorophyll-a in Long Island Sound sediments. Geochimica et Cosmochimica Acta 57, 147±157.
Teng, J.I., 1990. Oxysterol separation by HPLC in combination with thin layer chromatography. Chromatogram 11, 8±10. Teng, J.I., Kulig, M.J., Smith, L.L., Kan, G., van Lier, J.E.,
1973. Sterol metabolism. XX. Cholesterol 7-hydroperoxide. Journal of Organic Chemistry 38, 119±123.
Van Lier, J.E., Smith, L.L., 1970. Sterol metabolism. VII. Autoxidation of cholesterol via hydroperoxides inter-mediates. Journal of Organic Chemistry 35, 2627±2632. Volkman, J.K., 1986. A review of sterol markers for marine
and terrigeneous organic matter. Organic Geochemistry 9, 83±100.
Volkman, J.K., Barrett, S.M., Blackburn, S.I., Mansour, M.P., Sikes, E.L., Gelin, F., 1998. Microbial biomarkers: a review of recent research developments. Organic Geochemistry 29, 1163±1179.
