Use of palmitoleic acid and its oxidation products for monitoring the degradation of ice algae in Arctic waters and bottom sediments

ice Degradation of palmitoleic acid (C 16:1 x 7 ), the main fatty acid component of sea ice-associated (sympagic) diatoms, was monitored in Arctic sea ice at the beginning of ice melting and in the underlying sinking particles and superﬁcial bottom sediments. In sea ice, degradation of sympagic algae involved biotic oxi- dation induced by 10 S -DOX-like lipoxygenase of unknown salinity-stressed attached bacteria, while photo- and autoxidation were limited. In the water column, strong hydratase and Z / E isomerase activity were observed. Hydration of unsaturated fatty acids seems to be a detoxiﬁcation strategy, which is essential for bacterial survival when associated with free fatty acid-rich environments such as ice algae. In contrast, Z / E isomerisation of palmitoleic acid was attributed to the release of Fe 2+ ions during radical-induced damage of the active site of the bacterial 10 S -DOX-like lipoxygenase and Z / E isomerases. Due to the poor physiological state of their attached bacteria resulting from salinity stress in brine channels or toxicity of free ice algae fatty acids, sympagic algae appeared to be only very weakly biotically degraded within the water column. In bottom sediments, free radicals resulting from 10 S -DOX-like lipoxygenase activity induced a strong autoxidation of the ice algal material. The presence in bottom sediments of a signiﬁcant proportion of oxidation products resulting from 10 S -DOX-like lipoxygenase activity attested to the strong contribution of sea ice-derived OM released during the early stages of ice melt prior to depo- sition in the sediments. However, on the basis of the highest fatty acid photooxidation state observed in these sediments, an additional contribution of highly photooxidized material (ice algal material released at the end of ice melting or open water phytoplankton) seems likely. The degradation of hydroperoxides, resulting from biotic and abiotic degradation of palmitoleic acid, appeared to involve: (i) homolytic cleav- age of the peroxyl group affording the corresponding hydroxy- and oxoacids, (ii) reduction to the corresponding hydroxyacids by peroxygenases, (iii) heterolytic proton-catalysed cleavage and (iv) conversion to allylic 1,4-diols by diol synthases and hydroperoxide isomerases.


Introduction
The thinning and retreat of Arctic sea ice, which is one of the most striking consequences of recent climate change, will likely have a significant impact on Arctic ecosystem functioning in the future (Wassmann et al., 2011). Continued changes in the polar seas and sea ice retreat will also affect marine biogeochemical cycles, with possibly important feedbacks to the climate and marine ecosystems (Arzel et al., 2006).
Reduction of sea ice in the Arctic under the effect of global warming should significantly alter the relative fluxes of phytoplankton and ice algae to the seafloor (Carroll and Carroll, 2003). A climate-change-mediated shift in primary producers would thus impact the structure and function of the sea floor community, which is strongly dependent upon the deposition of organic material from the overlying water column for its energy requirements (McMahon et al., 2006). Moreover, under the effect of global warming the carbon sink potential of ice algae (resulting from their strong aggregation and the stress state of their associated bacteria, should be gradually replaced by the carbon source potential of open water phytoplankton (weakly aggregated and mineralized before the bottom) .
Determining the fate of sympagic (ice-associated) derived particulate organic matter (POM) following its release from sea ice during spring melting is thus an important research objective (Tedesco and Fettweis, 2012), but partitioning processes such as growth of the water column phytoplankton community, grazing, remineralization and export is complicated and currently not well constrained. The strong pulse of particulate OM released into the water column during ice melt provides essential energy at the base of the Arctic food web (Michel et al., 1996;Lizotte, 2001), although some sinks to the seafloor (Riebesell et al., 1991;Fortier et al., 2002;Renaud et al., 2007). Ultimately, the material that is not grazed or re-mineralized during its descent through the water column, can feed the benthos (Boetius et al., 2013) or be stored in sediments.
In practice, many factors influence the quantity and the quality of the sympagic OM reaching the seafloor. These include the physical characteristics of the water column, including water depth and whether it is well mixed or stratified, the latter being a particular feature of Arctic waters due to the lower salinity caused by ice melt in spring and freshwater runoff. In addition, the fate of sympagic OM is influenced by physiological, morphological and ecological factors such as cell agglomeration and grazing by under-ice and water column heterotrophs (Vancoppenolle et al., 2013). A particular feature of sea ice biota is their ability to produce large quantities of so-called extracellular polymeric substances (EPS), which serve as a cryoprotectant within the ice matrix. Indeed, EPS concentrations in Arctic sea ice are typically at least an order of magnitude higher than for under-ice and open water environments (Krembs and Engel, 2001;Meiners et al., 2003), which promotes aggregation of sympagic algae and thus enhances export efficiency to the seafloor. The contribution of ice algae to the underlying sediments also depends strongly on the remineralization potential of their associated bacteria. Recently, it was observed that during the early stage of ice melting, hypersaline conditions in brine channels strongly impact the bacterial community associated with ice algae . The resulting poor physiological state of these bacteria should likely contribute to the preservation of the algal material prior to deposition on the ocean bottom.
Lipid biomarkers, although only representing a relatively minor fraction of the total OM, can convey important information on the source (terrigenous, marine or bacterial) and degradation state of OM, and their distributions are often more informative than bulk parameters such as total organic carbon (Saliot et al., 2002). Furthermore, the oxidation products of monounsaturated fatty acids appear to be especially useful for differentiating the involvement of photo-and autoxidative degradation processes in suspended and sinking particulate matter and bottom sediment samples (for a review see Rontani, 2012). More recently, analysis of these compounds provided evidence of lipoxygenase-induced degradation in suspended (Galeron et al., 2018) and sinking  particles.
Since the integrity of the strong pulse of particulate organic matter (POM) released during Arctic ice melt remains largely unexamined, we monitored biotic and abiotic degradation processes in ice algae released from melting Arctic sea ice and their transfer to the ocean bottom using pathway-specific degradation tracer products of palmitoleic acid (C 16:1x7 ). Indeed, although present in several bacteria (e.g., de Carvalho and Caramujo, 2014), palmitoleic acid is also the major fatty acid of diatoms (Fahl and Kattner, 1993;Leu et al., 2010), which strongly dominate the biomass in sea ice.

Sampling
Ice cores, sinking particles and superficial bottom sediments were collected during the spring of 2015 at a landfast ice station located near Broughton Island (67°28.766 0 N; 63°47.579 0 W; water column depth: 379 m) north of Davis Strait, Canada (GreenEdge ice camp 2015). Sea ice and sinking particles were sampled during the early stage of ice melt ). An ice core was collected on 4 June 2015 using a 9 cm internal diameter Mark II coring system (Kovacs Enterprises) and sliced in the field. The bottom 1 cm slice, which corresponds to the ice-ocean interface and contains the largest amount of ice algae, was allowed to melt in filtered seawater (FSW; 0.2 mm) using at least 2 parts FSW for 1 part of ice in order to limit osmotic stress and filtered (Whatman GF/F 47 mm). Short-term sediments traps were deployed to collect sinking particles over 48 h at 2 m and 25 m with two mooring lines. Sediment traps were made of polyvinyl chloride (PVC) and had an aperture diameter of 15 cm. The recovery of the samples analyzed in the present study took place on 8-10 June 2015 (at the time of highest flux of sinking material measured during this time series, Lalande, unpublished data). A sub-sample for lipid analysis (corresponding to 4.5% of the total sample) was filtered onto Whatman GF/F 47 mm filter. Superficial bottom sediment samples (ca. 0-1 cm) were also collected using box corers at the same sampling location on board the CCGS Amundsen in October 2015 as part of the GreenEdge project. All the samples were kept frozen (-20°C) until analysis.

Treatment
Samples (filters or sediments) were reduced with excess NaBH 4 or NaBD 4 after addition of MeOH (25 ml; 30 min) to reduce labile hydroperoxides to alcohols which are more amenable to analysis using gas chromatography-mass spectrometry. Water (25 ml) and KOH (2.8 g) were then added and the resulting mixture saponified by refluxing (2h). After cooling, the mixture was acidified (HCl, 2 N) to pH 1 and extracted with dichloromethane (DCM; 3 Â 20 ml). The combined DCM extracts were dried over anhydrous Na 2 -SO 4 , filtered and concentrated via rotary evaporation at 40°C to give total lipid extracts (TLEs). TLEs were derivatized by dissolving them in 300 ml of pyridine/bis(trimethylsilyl)trifluoroacetamide (BSTFA; Supelco; 2:1, v:v) and silylated (50°C, 1 h). After evaporation to dryness under a stream of N 2 , the derivatized residue was dissolved in ethyl acetate/BSTFA (to avoid desilylation) and analyzed using gas chromatography-electron ionization quadrupole time of flight mass spectrometry (GC-QTOFMS).
A different treatment was employed to estimate the relative proportions of hydroperoxides and their ketonic and alcoholic degradation products. Subsamples were extracted three times with chloroform-MeOH-H 2 O (1:2:0.8, v:v:v) using ultrasonication. The supernatant was separated by centrifugation at 3500g for 9 min. To initiate phase separation, purified H 2 O was added to the combined extracts. The upper aqueous phase was extracted three times with DCM and the combined organic phase and DCM extracts were filtered and the solvent removed via rotary evaporation. The residue obtained after extraction was dissolved in 4 ml of DCM and separated in two equal subsamples. After evaporation of the solvent, degradation products were obtained for the first subsample after acetylation, which induced conversion of hydroperoxides to the corresponding ketones (Mihara and Tateba, 1986), and saponification. The second subsample was reduced with NaBD 4 and saponified. Comparison of the amounts of alcohols present after acetylation and NaBD 4 reduction made it possible to estimate the amount of hydroperoxides and alcohols present in the samples, while deuterium labeling after NaBD 4 reduction allowed us to estimate the proportion of ketones present in the samples (Marchand and Rontani, 2003).
Acetylation was carried out in 300 ll of a mixture of pyridine and acetic anhydride (2:1, v:v), which was allowed to react at 50°C overnight and then evaporated to dryness under a stream of N 2 .

GC-QTOFMS
Oxidation products of palmitoleic acid were identified and quantified using an Agilent 7890B/7200A GC-QTOF System (Agilent Technologies, Parc Technopolis -ZA Courtaboeuf, Les Ulis, France). A cross-linked 5% phenyl-methylpolysiloxane (Macherey Nagel; Optima 5-MS Accent) column (30 m Â 0.25 mm, 0.25 lm film thickness) was employed. Analysis was performed with an injector operating in pulsed splitless mode at 280°C and the oven temperature programmed from 70°C to 130°C at 20°C/min, then to 250°C at 5°C/min and then to 300°C at 3°C/min. The carrier gas (He) was maintained at 0.69 Â 10 5 Pa until the end of the temperature program. Instrument temperatures were 300°C for the transfer line and 230°C for the ion source. Accurate mass spectra were obtained across the range m/z 50-700 at 4 GHz. The QTOFMS instrument provided a typical resolution ranging from 8009 to 12,252 from m/z 68.9955 to 501.9706. Perfluorotributylamine (PFTBA) was utilized for daily MS calibration. Compounds were identified by comparison of their TOF mass spectra, accurate masses and retention times with those of standards, either purchased or synthesized in the laboratory (see following Section). Quantification of each compound involved extraction of specific accurate fragment ions, peak integration and determination of individual response factors using external standards.

Estimation of autoxidative, photooxidative and 10S-DOX degradation
The role played by autoxidation in the degradation of palmitoleic acid was estimated by considering the proportion of its specific Z-oxidation products (Frankel, 1998) and the water temperature according to the approach previously described by Marchand and Rontani (2001). After subtraction of the amounts of oxidation products of autoxidative origin, it remained to determine the relative parts played by photooxidative and enzymatic processes in the degradation. Taking into account the production of equal amounts of 9-E (8) and 10-E (9) oxidation products during the photooxidation of the D 9 monounsaturated fatty acids (Frankel, 1998) and their specific allylic rearrangement to 11-E (5) and 8-E (6) isomers, respectively (Porter et al., 1995), the contribution of photooxidative degradation was estimated to be 2 Â (9-E + 11-E). Concerning 10S-DOX degradation, this was obtained from the difference between (10-E + 8-E) and (9-E + 11-E) oxidation products, to which was added the amount of 7,10dihydroxyhexadec-8(E)-enoic acid (21) formed (Galeron et al., 2018).
Isomeric allylic dihydroxyhexadecenoic acids could be also detected (Table 1, Fig. 3). Threo 7,10-dihydroxyhexadec-8(E)enoic (21), threo and erythro 8,11-dihydroxyhexadec-9(E)-enoic (22) and erythro 8,11-dihydroxyhexadec-9(Z)-enoic acids (23) could be formally identified by comparison of their accurate mass spectra ( Fig. 4a and c) and retention times with those of reference compounds. The compound eluting just after threo 7,10dihydroxyhexadec-8(E)-enoic acids (21) (Fig. 3), and exhibiting exactly the same mass spectrum, was attributed to the erythro diastereoisomer. Indeed, it was observed previously that threo and erythro diastereoisomers of TMS derivatives of dihydroxyhexadecanoic acids eluted on an apolar column in the same order (Hansel and Evershed, 2009). Erythro 9,12-dihydroxyhexadec-10 (E)-enoic acids (24) were tentatively identified on the basis of the accurate mass fragmentations of their trimethylsilyl derivatives (Fig. 4b) and the abundance of the m/z 147 peak in their mass spectra, which is similar to those of the other E isomers (Fig. 4a and c). Indeed, this fragment ion, which arises from interaction between two TMS groups, is always more intense in the EI mass spectrum of the Z isomer compared to its E counterpart (Zaikin and Halket, 2009). The percentage of these isomeric dihydroxyhexadecenoic acids increased from 0.1% in sea ice to 6% in superficial bottom sediments.
A significant increase of the proportion of hexadec-9(E)-enoic (palmitelaidic) acid relative to palmitoleic acid was apparent in the deeper sediment trap and in superficial bottom sediments (Table 1).

Degradation of palmitoleic acid in sea ice (0-1 cm)
Type II photosensitized oxidation (i.e. involving singlet oxygen) of palmitoleic acid produces equal proportions of isomeric 9-and 10-hydroperoxides (2 and 3) with an allylic E double bond (Frankel, 1998), which can subsequently undergo highly stereoselective radical allylic rearrangement to 11-E and 8-E hydroperoxides (1 and 4), respectively (Fig. 5;Porter et al., 1995). In contrast, autoxidation (i.e. free radical induced oxidation) affords a mixture of 9-E (2), 10-E (3), 11-E (1), 11-Z (5), 8-E (4), and 8-Z (6) hydroperoxides, also exhibiting equal proportions of the major 9-E (2) and 10-E (3) isomers (Fig. 5;Frankel, 1998). The strong predominance of the 10-E hydroxyacid (9) observed in sea ice after NaBH 4 reduction ( Fig. 1) is thus very surprising. Similar profiles of palmitoleic acid oxidation products were observed previously in other sea ice samples (Amiraux et al., submitted for publication) and in sinking particles underlying melting sea ice , also in the Canadian Arctic, and attributed to the involvement of a specific bacterial enzymatic process under the hypersaline conditions found in brine channels. Indeed, a 10S-DOX-like lipoxygenase able to convert palmitoleic acid to 10(S)hydroperoxyhexadec-8(E)-enoic acid (3) was previously isolated from Pseudomonas aeruginosa 42A2 (Guerrero et al., 1997;Busquets et al., 2004). Palmitoleic acid is a major constituent of numerous bacteria (Oliver and Colwell, 1973;Viso and Marty, 1993), although it is also the predominant fatty acid in sea ice diatoms (Fahl and Kattner, 1993), which make up the major part of the biomass in sea ice. The presence of high proportions of 10hydroperoxyhexadec-8(E)-enoic acid (3) in sea ice was thus attributed to the degradation of ice algae by some of their attached bacteria , consistent with the autonomous action of prokaryotic lipoxygenases on the membrane phospholipids of their eukaryotic cell hosts (Vance et al., 2004;Garreta et al., 2013).
The relatively low proportions of 9E-and 11E-hydroxyacids (Fig. 1a) is consistent with the low efficiency of Type II photooxidation processes on ice algae, attributed previously to the good physiological state of algae at the time of sampling . Autoxidation also appeared to be limited in the sea ice sample as shown by the low amounts of specific 8Z-and 11Zhydroxyacids (Fig. 1a). Table 1 Concentration of palmitoleic acid and its degradation products in sea ice, sinking particle and surficial bottom sediment samples collected in Davis Strait.

J
The degradation of all these labile isomeric hydroperoxyacids arising from photooxidative, autoxidative and enzymatic processes in sea ice involves: (i) homolytic cleavage of the peroxyl group affording the corresponding hydroxy-and oxoacids (Frimer, 1979), (ii) reduction to the corresponding hydroxyacids by peroxygenases, which catalyze an intermolecular transfer of oxygen from hydroperoxides to the double bonds of fatty acids yielding alcohols and epoxides (Blée and Schubert, 1995;Blée, 1998), (iii) heterolytic proton-catalysed cleavage leading to the formation of two carbonyl fragments (Hock cleavage) initiated by migration of groups to positive oxygen (Frimer, 1979), and (iv) isomerization to allylic 1,4-diols by hydroperoxide isomerases (Fig. 6). It is important to note that this last process affords erythro isomers (Jernerén et al., 2010), which are different to the threo isomers resulting from diol synthase activity.

Degradation of palmitoleic acid in sinking particles
The dominance of Nitzschia frigida (1-20%) and Fragilariopsis oceanica (2-24%) (well-known diatom species in sea ice; Poulin et al., 2011;Ren et al., 2014) in the sediment trap material (Lalande, unpublished data) confirms a sea ice origin for the majority of sinking particles. The high contribution of ice diatoms to this material is also supported by the strong dominance of palmitoleic acid in TLEs (Fahl and Kattner, 1993). Ice algae appeared to be very weakly degraded within the water column, with degradation percentages of palmitoleic acid of only 4.8%, 5.4%, and 4.3% in ice, upper and deeper sediment traps, respectively. This relative preservation is likely attributed to the poor physiological state of bacterial communities associated with ice algae resulting from salinity stress in brine channels .
Stereospecific allylic rearrangement of 10-hydroperoxyhexadec-8(E)-enoic acid (3) to 8-hydroperoxyhexadec-9(E)-enoic acid (4), which appeared to be very limited in sea ice (Fig. 1a), acted very intensively in sinking particles (Fig. 1b). The extent of this rearrangement increased with time (Porter et al., 1995), with higher proportions of the 8-E isomer observed in sinking particles confirming the aging of the material collected in the sediment traps.
A significant increase of the proportion of 10hydroxyhexadecanoic acid (14) was observed in sinking particles (Fig. 1, Tables 1 and 3) suggesting an enhancement in the activity of hydratases within the water column. Interestingly, it was previously demonstrated that myosin cross-reactive antigen (MCRA) proteins acting as a hydratase on 9(Z) double bonds of fatty acids, play a role in adherence of bacteria (Volkov et al., 2010). Indeed, free unsaturated fatty acids (and notably palmitoleic acid, Wille and Kydonieus, 2003) are toxic for many bacteria due to their deleterious effect on bacterial cellular membranes (Greenway and Dyke, 1979). They also inhibit enoyl-ACP reductase and thus disrupt bacterial fatty acid synthesis (Zheng et al., 2005). Hydration of unsaturated fatty acids may thus represent a detoxification mechanism in bacteria harboring MCRA enzymes. Such detoxification could be essential for bacterial colonization and survival on free fatty acid-rich environments such as ice algae (Falk-Petersen et al., 1998). However, it may be noted that such enhancement of hydratase activity was not previously observed in sinking particles collected in spring 2012 at a landfast ice station in Resolute Passage (Nunavut, Canada) (Rontani et al., 2016), probably due to the higher aggregation state of the ice algae , resulting from an enhanced EPS content. In this case, the presence of high amounts of EPS around ice algae cells could have limited the toxicity of unsaturated fatty acids toward bacteria. The intense Z/E isomerisation of palmitoleic acid observed in the deeper sediment trap samples (Table 1) may result from the involvement of thiyl radicals. Indeed, functionalised aliphatic thiols (glutathione, methionine-containing proteins), which are present in living organisms in considerable amounts (Ferreri et al., 2005), are extraordinarily efficient antioxidants that protect the cells against consequences of damage induced by free radicals and hydroperoxides (Wlodek, 2002). However, this role as repairing agents is counterbalanced by the formation of thiyl radicals, which are efficient catalysts for Z/E isomerisation of lipids in biological membranes (Ferreri et al., 2005). Due to the presence of significant amounts of hydroperoxyacids in sinking particles (Table 1), an induction of Z/E isomerisation of palmitoleic acid by thiyl radicals resulting from the reaction of thiols with these compounds is therefore possible. However, the involvement of such an isomerisation process is not supported by the very low E/Z ratios of unsaturated fatty acids measured previously in highly photooxidized sinking particles (with very high hydroperoxide contents) collected in spring 2012 at a landfast ice station in Resolute Passage (Nunavut, Canada) (Rontani et al., 2016).
It may be noted that Z/E isomerisation of unsaturated fatty acids can also be catalyzed non-enzymatically by Fe 2+ ions (Holtwick et al., 1997;Li and Han, 2008  during radical-induced damage of the active site of the bacterial 10S-DOX-like lipoxygenase (Fuch and Spiteller, 2014) and Z/E isomerase (a cytochrome C-type protein containing an ion Fe 2+ , Holtwick et al., 1997) could thus be at the origin of the Z/E isomerisation of palmitoleic acid in diatoms (Fig. 5).

Degradation of palmitoleic acid in superficial bottom sediments from north of Davis Strait
The dominance of 10-hydroxyhexadec-8(E)-enoic acid (9) (arising from 10S-DOX-like lipoxygenase activity) among palmitoleic acid oxidation products observed in the sediments investigated (Fig. 1c) was also previously detected in several surficial sediments collected in different zones of Arctic (Amiraux, 2017). It may be attributed to a strong contribution of ice-derived OM released to Arctic sediments during the first stage of ice melting and containing bacteria stressed by salinity. It is well known that the generation of radicals in the course of the lipoxygenase catalytic cycle can act as a catalyst in autoxidation processes (Fuch and Spiteller, 2014). Moreover, when lipoxygenase activity becomes very high, increasing amounts of free radicals may damage the active site of these enzymes and release Fe 2+ ions (Sato et al., 1992;Fuch and Spiteller, 2014), which may very efficiently catalyze the reduction of hydroperoxides to alkoxyl radicals (Schaich, 2005) and thereby induce free radical oxidation chains. The increase of autoxidation observed in surface sediments (Table 3) is thus likely initiated by the 10S-DOX-like lipoxygenase activity present in sinking ice algae .
In contrast, these sediments exhibit a higher proportion of photooxidation products of palmitoleic acid than the particles settling during the early stages of ice melting (Table 3), suggesting the contribution of an additional source of highly photooxidized sinking material. Photooxidative damage is strongly dependent on the physiological state of phototrophic cells (Nelson, 1993;Merzlyak and Hendry, 1994). At the end of the ice melt process, downward percolation of melt water strongly reduces the overall salinity in brine channels, affecting the physiological state of ice algae and increasing their susceptibility towards photooxidation . A weak contribution of highly aggregated ice algae settling during the end of ice melting, where palmitoleic photooxidation percentage was very high (mean value at 30 m: 60%; Rontani et al., 2016) could thus explain the relatively high photooxidation state of palmitoleic acid observed in these sediments. It was previously observed that strands of the under-ice diatom Melosira arctica, which can cover up to 40-80% of the underside of undisturbed ice floes (Syvertsen, 1991;Gutt, 1995) and contain a high proportion of palmitoleic acid (Falk-Petersen et al., 1998), are not used as food in the pelagic zone and sink rapidly to the seafloor (Boetius et al., 2013). A contribution of strongly photooxidized M. arctica could thus also explain the high photooxidation state of sedimentary palmitoleic acid. However, the high chlorophyll content measured in these algal deposits (Boetius et al., 2013) (attesting to their very weak photooxidation state) allows us to discard a significant contribution of this algal material to the superficial bottom sediments analyzed. Despite its lower aggregation and thus sedimentation rate, a weak contribution of open water phytoplankton, which is strongly affected by photooxidation processes in summer in the Arctic , cannot be totally excluded.
The relatively similar palmitelaidic acid/palmitoleic acid ratios observed in superficial bottom sediments and in the deeper sediment trap dominated by sympagic algae (Table 1) confirms the strong contribution of sea ice-derived OM to bottom sediments, a conclusion consistent with that derived following a previous analysis of vaccenic acid isomerisation in these samples .

Conclusions
In the present work, we detected an intense 10S-DOX-like lipoxygenase activity in sea ice, attributed to the involvement of a specific bacterial enzymatic process under the hypersaline conditions found in brine channels during the first stages of ice melting. This enzymatic activity, well-known to produce radicals, is probably at the origin of the strong autoxidation of ice algal material observed in the underlying sediments.
In the water column, the degradation of ice algal material appeared to be very limited, probably due to the presence of bacteria stressed by salinity in the overlying sea ice or by the high free fatty acid content of ice algae. The degradation of hydroperoxides (arising from 10S-DOX-like lipoxygenase oxidation, photo-and autoxidation in ice) involves peroxygenases, diol synthases, hydroperoxide isomerases and abiotic processes (homolytic and heterolytic cleavages, allylic rearrangement). The increasing hydratase activity (a well-known detoxification process in bacteria) observed in sinking particles, which attests to the toxicity of free unsaturated FA of ice algae toward their attached bacteria, constitutes another explanation of the relative preservation of ice algae within the water column. Z/E isomerization of palmitoleic acid was observed in the deeper trap and in surface sediments and was attributed to Fe 2+ ions released during the degradation of active sites of bacterial lipoxygenases and Z/E isomerases. These ions could then enhance autoxidation of ice algal material in the underlying sediments.
Comparison of the profiles of palmitoleic acid degradation products in superficial bottom sediments and in sinking particles allowed us to propose ice algae settling during the early stages of ice melting as important contributors to Arctic sediments. A weaker contribution of ice algal material, strongly photooxidized at the end of sea ice melting, also seems likely. Due to their strong aggregation and low remineralizing potential, ice algae seem to contribute more significantly than open water phytoplankton to the export of carbon to Arctic sediments.