PRODUCTION AND CELLULAR LOCALIZATION OF NEUTRAL LONG-CHAIN LIPIDSIN THE HAPTOPHYTE ALGAE ISOCHRYSIS GALBANA AND EMILIANIA HUXLEYI1

Matthew L. Eltgroth, Robin L. Watwood, and Gordon V. Wolfe2

Department of Biological Sciences, California State University Chico, Chico, California 95929-0515, USA

Isochrysis galbana Parke, Emiliania huxleyi(Lohm.) Hay and Mohler, and some related prym-nesiophyte algae produce as neutral lipids a set ofpolyunsaturated long-chain (C37–39) alkenones, al-kenoates, and alkenes (PULCA). These biomarkersare widely used for paleothermometry, but thebiosynthesis and cellular location of these uniquelipids remain largely unknown. By staining with thefluorescent lipophilic dye Nile Red, we found thatI. galbana and E. huxleyi, like many other algae,package their neutral lipid into cytoplasmic vesiclesor lipid bodies. We found that these lipid bodiesincrease in abundance under nutrient limitationand disappear under prolonged darkness andshow that this pattern correlates well with theconcentration of PULCA as measured by TLC. Inaddition, we show that lipid vesicles purified bysucrose density gradient centrifugation consist pre-dominantly of PULCA. We also found significantpools of neutral lipid associated with chloroplasts,and PULCA component profiles in lipid vesiclesand chloroplasts are similar. Examination of cellultrastructure shows conspicuous cytoplasmic andchloroplast lipid bodies, and we suggest thatPULCA may be synthesized in chloroplasts andthen exported to cytoplasmic lipid bodies for sto-rage and eventual metabolism. Our results connectand extend prior observations of lipid bodies andmembrane-unbound PULCA in I. galbana andE. huxleyi, as well as the behavior of PULCA duringnutrient and light stress.

Key index words: alkenoates; alkenones; alkenes; E.huxleyi; haptophytes; I. galbana; lipid bodies; neu-tral lipids; Nile Red; PULCA; TLC; ultrastructure

Abbreviations: ER, endoplasmic reticulum; LB, li-pid body; NR, Nile Red; PULCA, polyunsaturatedlong-chain alkenes, alkenones, and alkenoates;TAG, triacylglycerol

Some prymnesiophyte taxa of the order Isochry-saldes (Isochrysis, Emiliania, Gephyrocapsa, Chrysotila)produce only small amounts of triacylglycerol (TAG)as their neutral lipid (Marlowe et al. 1984a,b). Theyproduce instead a suite of polyunsaturated long-chain

(C37–39) alkenes, alkenones, and alkenoates (Sukenikand Wahnon 1991, Patterson et al. 1994, Fabregaset al. 1998), abbreviated here as PULCA (Fig. 1).These compounds are unlike the cis-polyunsaturatedfatty acids typical of eukaryotic membrane constituentsin that they have two to four unusual trans-alkenebonds occurring at 7-carbon intervals (Fig. 1a),although this has been confirmed only for the C37

alkenones (de Leeuw et al. 1980, Rechka and Maxwell1988a,b) and C37–38 alkenes (Riley et al. 1998). Nosaturated isomers have been reported, but Rontaniet al. (2001, 2004a) detected monounsaturated alke-nones in all taxa. Another set of C31–33 alkenes pro-duced by these taxa has been determined to have cisgeometry (Riley et al. 1998) and are likely biosynthe-tically distinct.

At colder growth temperatures, PULCA are morehighly polyunsaturated (Conte et al. 1995, 1998), andthe proportion of diunsaturated isomers of the C37

methyl alkenones (the ‘‘unsaturation index’’ UK037) was

proposed as a growth temperature proxy (Brassellet al. 1986, Prahl and Wakeham 1987). Because thesecompounds were first discovered in marine coccolith-bearing sediments (de Leeuw et al. 1980, Volkmanet al. 1980a,b), sedimentary core-top UK0

37 has beenwidely adopted by geochemists as a proxy for surfacewater paleotemperature (Muller et al. 1998). However,the calibration of unsaturation index depends onstrain genetics (Conte et al. 1998) and environmentalfactors such as nutrient limitation or light availability(Epstein et al. 1998, 2001, Yamamoto et al. 2000,Versteegh et al. 2001, Prahl et al. 2003), and the utilityof this tool remains limited by our lack of under-standing of the biosynthesis of these compounds.

The cellular location of the PULCA has long been apuzzle. Early reports (Brassell et al. 1986, Prahl et al.1988) assumed they were membrane lipids and sug-gested the increased degree of PULCA unsaturation atcolder growth temperatures played a role in maintain-ing membrane fluidity. One ultrastructural study (vander Wal et al. 1985) hypothesized that an unusual‘‘intermediate’’ layer of the plasma membrane mightbe the location of these compounds. In contrast, Conteand Eglinton (1993) stated that alkenones were notdetected in the membranes of lysed cells, althoughthey provided no data in support of that assertion.However, a recent study by Sawada and Shiraiwa(2004) examined isolated cell membrane fractions ofE. huxleyi and found that alkenones and alkenoates

1Received 11 February 2005. Accepted 29 June 2005.2Author for correspondence: e-mail gwolfe2@csuchico.edu.

1000

J. Phycol. 41, 1000–1009 (2005)r 2005 Phycological Society of AmericaDOI: 10.1111/j.1529-8817.2005.00128.x

associated with internal organelles, such the endoplas-mic reticulum (ER) and coccolith-producing compart-ment, and suggested these PULCA components aremembrane-unbound lipids. Additional PULCA werealso found in the Golgi and plasma membranes and inchloroplast thylakoids. But unlike the ER and cocco-lith-producing compartment, these fractions weredominated by fatty acids typical of membrane lipids.

The physiological role of PULCA appears to resem-ble the role of other neutral lipids, which often serve asenergy reserves. Pond and Harris (1996) showed thatin Emiliania huxleyi, PULCA are present in all growthphases but cellular pools increase in stationary phase,as is typical for storage lipids (Henderson and Sargent1989, Hodgson et al. 1991). Epstein et al. (1998) andPrahl et al. (2003) observed increased PULCA accu-mulation under N or P limitation, up to 10%–20% ofcell C in stationary phase, although this varied con-siderably among strains. Epstein et al. (2001) andPrahl et al. (2003) also showed that under energy-depleted growth conditions (prolonged darkness),PULCA pools decrease due to catabolism. Other eu-karyotes that store TAGs often compartmentalize theseinto lipid vesicles or lipid bodies (LBs) (Murphy 2001).These droplets can be visualized with the fluorescentstain Nile Red (NR) (Cooksey et al. 1987, Brzezinskiet al. 1993, Dempster and Sommerfeld 1998, Kimuraet al. 2004), which selectively stains neutral lipids.Oleaginous fungi (Kamisaka et al. 1999) and algae

(Schneider and Roessler 1994) accumulate very highamounts of neutral lipid and have been studied for oilproduction and as aquaculture food stocks. Amonghaptophytes, there is little known about LBs, but Liuand Lin (2001) observed lipid droplet formation andaccumulation in Isochrysis galbana, especially as cellsentered the stationary phase.

In this study, we present data on the cellular loca-tion of the long-chain neutral lipids in the haptophytesI. galbana and E. huxleyi. We identified LBs within thesealgae, showed their behavior is consistent with a role inenergy storage, and determined the composition ofpurified LBs to be PULCA.

MATERIALS AND METHODS

Cultures. Emiliania huxleyi and I. galbana cultures (Table 1)were obtained from the Provasoli-Guillard National Centerfor Culture of Marine Phytoplankton (CCMP) or were a giftfrom Suzanne Strom, Western Washington University. Emi-liania huxleyi CCMP 1742 is a synonym for strain 55a, thecalibration standard for alkenone paleothermometry (Prahland Wakeham 1987). Emiliania huxleyi CCMP 1516 is acalcifying strain chosen by the DOE, U.S. Dept. of Energyfor genome sequencing (B. Read and T. Wahlund, personalcommunication). Cultures were grown in 100–200 mL vo-lumes of f/2 or f/20 media under 80 mmol photons �m–2 � s–1

cool-white illumination at 161 C in a 16:8-h light:dark cycle.In some experiments, we grew cells with either nitrate orphosphate reduced to 10% f/2 levels to specifically induce Nor P limitation. For cells incubated in complete darkness, thetemperature and other conditions remained the same. Cellnumbers were determined by hemocytometer or by CoulterCounter (Coulter Electronics, Ltd., Luton, U.K.) using a70-mm orifice.

NR staining, microscopy, and spectrofluorometry. Cells fixedwith glutaraldehyde (final concentration 0.5%) were stained in1-mL aliquots with 6mL of an NR stock solution (500mg �mL�1

in acetone) and counterstained with one drop of 15mg �mL�1

DAPI. We viewed wet mounts with an Olympus BX51 micro-scope (Melville, NY, USA) and a Pixera Penguin 600EScharged coupled device camera (Los Gatos, CA, USA) forbright-field, Nomarski DIC, or fluorescence using blue excita-tion (455 nm). A Schott glass bandpass output filter centered at495 nm (Oriel optics no. 51720, Stratford, CT, USA) was addedto mask chl emission. Spectrofluorimetry was performed on3 mL stained fixed cells with an Ocean Optics (Dunedin, FL,USA) USB2000 spectrofluorometer driven by a Sutter Instru-ments (Novato, CA, USA) Lambda-LS Xe light source with492 � 10 nm band pass filter (Edmund Optics #NT46–042,Barrington, NJ, USA).

Cell fractionations. Cells were pelleted in a clinical centri-fuge, and the pellets were repeatedly subjected to osmoticlysis with a protease inhibitor cocktail containing 1 mM

FIG. 1. PULCA structures. (a) C35-alkene-based PULCA ske-leton. Diunsaturated isomers at C14 and C21 are shown, with sitesof additional trans-desaturation at C7 and C28 denoted as inter-rupted double bonds. Longer C36 and C37 analogs occur withadditional carbons at the R group. (b–d) Terminal R groupsinclude C37 methyl/C38 ethyl alkenones (b), C36 methyl or ethylalkenoates (c), and C37 alkenes (d).

TABLE 1. CCMP algal cultures used in this study.

CCMP strain Synonyms Origin Latitude Growth (1 C) Calcification

Emiliania huxleyi1742 VAN556, 55a, NEPCC55a N.E. Pacific 50 N 8–25 No1516 CCMP2090 S. Pacific 2 S 14–22 Under low nutrients370 (451B, F451) Norway 59 N 11–16 No374 (89E, CCMP1949) Gulf of Maine 42 N 11–22 NoIsochrysis galbana1323 ISO, Strain ‘‘I’’, NEPCC2 Irish Sea 54 N 3–28 No

HAPTOPHYTE NEUTRAL LIPIDS 1001

EDTA, 1 mM PMSF, 1 mM benzamidine, and 10 mg �mL�1

leupeptin and pepstatin A (Sigma-Aldrich, St. Louis, MO,USA). Cell debris and unlysed cells were removed by low-speed centrifugation (3000 rpm, 5 min), and the remainingcell-free homogenate was collected. Two to 3 mL of the cell-free homogenate was layered over a discontinuous sucrosegradient consisting of 5 mL 60%, 5 mL 45%, 5 mL 30%, and5 mL 15% sucrose. The gradients were ultracentrifuged at30,000 g for 16 h at 41 C. Buoyant LB fractions were collectedfrom the surface by pipette. Chloroplast fractions werecollected from the 45%–60% sucrose interface and wereresupended in small volume of water before lipid extraction.

Lipid extractions. Cells or cell fractions were collected byfiltration onto GF/F (Whatman, Florham Park, NJ, USA)filters, by centrifugation, or by addition of aqueous samplesdirectly to the extraction solvents. Lipids were extractedusing a modified Bligh and Dyer procedure (Bligh andDyer 1959). Briefly, filters or pellets were immediately placedin 10:5:4 CHCl3/MeOH/H2O and vortexed and then incu-bated for 2–24 h. Chloroform and water were added to a finalconcentration of 10:10:9 CHCl3/MeOH/H2O to produce twophases, which were separated by centrifugation in a clinicalcentrifuge. The organic layer was removed, and the remain-ing aqueous layer was reextracted with chloroform. Thecollected extracts were evaporated to dryness under N2 at501 C. Samples were resuspended in 2:1 CHCl3/MeOH,capped under N2, and stored at � 201 C until analysis.

Lipid analysis. Total lipids were separated by TLC using adouble development system (Olsen and Henderson 1989) onsilica HPTLC plates (60 mm particle size, Camag Scientific,Wilmington, NC, USA). The development of polar lipids wasperformed with 25:25:25:10:9 Me acetate/isopropanol/CHCl3/MeOH/KCl (0.9% aqueous solution). For nonpolarlipids, a solvent containing a 80:20:2 mixture of hexane/diethyl ether/glacial acetic acid was used. Plates were scannedat 600 dpi on a flatbed scanner to detect pigments. Lipidswere detected by spraying with 3% (w/v) cupric acetate in 8%(v/v) H3PO4 and charring for 10 min at 1601 C. Standardsincluded phosphatidylcholine, phosphatidylglycerol, digalac-tosyl diacylglyceride, cholesterol, a TAG mixture, and a fattyacid methyl ester (FAME) mixture (Sigma-Aldrich). Detectionlimit was about 5 mg for individual lipid components. Plateswere scanned again, and TIFF images were analyzed fordensitometry with Kodak 1D image analysis software (NewHaven, CT, USA). For densitometry, calibrations standardcurves (5–25 mg) included phosphatidylcholine (phospholi-pids), digalactosyl diacylglyceride (glycolipids), and TAGs(PULCA). GC analysis of PULCA fractions was performedas described in Prahl et al. (2003).

TEM. Cells pelleted for TEM were fixed in0.5� Karnovsky’s fixative (pH 7.4) for 30 min and rinsedtwice for 5 min each in 0.1 M Sorensen’s phosphate buffer.For standard staining, cells were postfixed in 1% OsO4 for30 min and rinsed in 0.1 M Sorensen’s phosphate buffer. Forosmium thiocarbohydrazide-osmium staining (Willinghamand Rutherford 1984, Guyton and Klemp 1988), cells werepostfixed with a saturated aqueous solution of thiosemicar-bazide, followed by a 5-min buffer rinse, a 20-min postfix in1% OsO4, and a 5-min buffer rinse. All samples weredehydrated in a series of 70, 90, and three 100% acetonesolutions, each for 10 min. Pellets were infiltrated with a50:50 mixture of Epon/acetone overnight and then em-bedded in 100% Epon resin and polymerized for 48 h at651 C. Sections were cut on an LKB ultramicrotome andcollected on 300 mesh Cu grids. Sections were stained in 2%uranyl acetate for 5 min, rinsed three times for 15 s each indistilled water, then stained in Fahmy lead for 5 min, followedby a rinse for 15 s in 20 mM NaOH, then three 15-s rinses indistilled water. Sections were viewed and photographed on

an Hitachi H-300 transmission electron microscope(Schaumberg, IL, USA), and negatives were digitized with aMicrotek scanner (Carson, CA, USA) at 3600 dpi.

RESULTS

LBs observed by epifluorescence and LM. Preliminarytests with NR showed optimal staining at 4 mg �mL� 1,with emission at 580–625 nm (Fig. 2). Under NRstaining, Isochrysis and Emiliania cells in late expo-nential phase showed red-staining cell membraneswith large yellow-staining lipid vesicles (Fig. 3a).When a Schott glass bandpass filter was added tothe emission path to block red fluorescence, someyellow staining was also visible in chloroplasts andsometimes around the cell membrane (Fig. 3b). Lipidbodies varied in number and size among cells, butmost cells had multiple LBs, typically located next tothe chloroplasts or at the periphery of cells. No-marski DIC images of cells without NR stainingshowed LBs visible as refractive inclusions (Fig. 3c),and with bright-field illumination, LBs often showeda blue coloration (Fig. 3d).

LB and PULCA behavior during nutrient and lightstress. We next examined what happens to thesebodies during times of lipid anabolism or catabolism.Emiliania huxleyi 1742 cells that reached stationaryphase due to N or P limitation showed a qualitativeincrease in LBs under NR staining (Fig. 4a, day 10).The TLC showed major increases of PULCA pools instationary phase (Fig. 4b, days 4–10), and neutrallipid accounted for nearly all the increased lipid overthis time (Fig. 4c). In contrast, when E. huxleyi cellswere placed into the dark for 72 h, NR-stained cellsshowed nearly complete loss of NR signal at 580–625 nm, with only partial loss of chl fluorescence(Fig. 5). The TLC analysis showed that in the dark,PULCA stores were markedly reduced, althoughthere were changes in polar lipids as well (Fig. 6a).

2000

1500

1000

500

0550 600 650 700 750

Emission λ (nm)

Flo

ures

cenc

e (a

rbit

rary

)

16

4,8

2

0

FIG. 2. Fluorescence emission spectra of Pyrenomonas salinastained with 0, 2, 4, 8, or 16mg �mL� 1 NR. Excitation at 492 nm.Chlorophyll peak is at 675 nm, whereas NR emission fromneutral lipid (TAG) at 580 nm (arrow) increases up to 4mg �mL�1.

M. L. ELTGROTH ET AL.1002

Isochrysis cells in prolonged darkness exhibited asmall decrease in PULCA, with relatively little changein polar lipids (Fig. 6b). However, even after 7 days indarkness, we did not observe large NR fluorescencereductions in I. galbana. Rather, we observed thatneutral lipid seemed to move from chloroplasts tocytoplasmic LBs (Fig. 6, c and d). This pattern wasalso characteristic of lighted but nutrient-limitedIsochrysis (Fig. 3a), although total PULCA pools in-creased in the light.

PULCA in subcellular fractionations. After sucrosedensity ultracentrifugation, NR staining of the buoy-ant fraction of stationary phase cell lysates revealednumerous LBs (Fig. 7a), as did the chloroplast frac-tion (not shown). The TLC analysis of the buoyantfraction revealed that LBs are comprised primarily ofneutral alkenones and alkenes, with some phospha-tidylcholine and a sterol or carotenoid (Fig. 7b, lane2). The latter might account for the pigmentation ofLBs under bright-field microscopy (Fig. 3d). TheTLC also revealed a significant amount of neutral

lipids in the chloroplast fraction (Fig. 7b, lane 3).Isochrysis galbana and E. huxleyi cultures producedsimilar results, but E. huxleyi lipids associated withLB were more difficult to recover after centrifuga-tion, even when cells had notable lipid vesicles (notshown). The GC analysis of cell fractions showedsimilar profiles of PULCA components (Fig. 8). Thispattern was observed for several E. huxleyi strains aswell as I. galbana, although component profiles werestrain- or species-specific (Table 2).

LB ultrastructure. The transmission electron mi-crographs of stationary phase E. huxleyi cells showedlarge dark-staining lipid bodies in the cytoplasm(Fig. 9). As with light micrographs (Fig. 3), TEMshowed lipid bodies typically associated with chlor-oplasts, usually appressed to the chloroplast mem-brane (arrows, Fig. 10, a and b). Chloroplastscontained smaller dark-staining bodies adjacent tothylakoid membranes (arrows, Fig. 10, c and d).Osmium thiocarbohydrazide-osmium staining pro-duced very dark LBs, whereas standard osmium

FIG. 3. Light micrographs ofLBs. (a, b) NR-stained epifluor-escence. (a) Isochrysis galbanawithout emission bandpass filter.Neutral lipid appears yellow, po-lar lipids and chl, red. (c) Emilia-nia huxleyi 1742 with emissionfilter to remove red fluores-cence. (c) Nomarski DIC imagesof E. huxleyi 1516. Ragged cocco-sphere is visible on cell exterior.Spherical lipid vesicles (arrows)are seen associated with chloro-plasts. (d) Bright-field image ofE. huxleyi 1742. Note blue colora-tion of LBs (arrows). Cell dia-meters 3–5mm.

HAPTOPHYTE NEUTRAL LIPIDS 1003

staining gave more variable results. We saw noevidence of a membrane bounding the bodies. Iso-chrysis galbana cells showed similar features (notshown).

DISCUSSION

Here, we provide evidence that the PULCA-produ-cing taxa I. galbana and E. huxleyi have conspicuouscytosolic LBs easily visible by LM and EM that consistprimarily of NR-staining PULCA neutral lipid. NRlabeling showed that LBs increase at stationary phase(P or N limitation) and decrease in the dark, exactly thebehavior observed for bulk PULCA (Prahl et al. 2003)and consistent with an energy storage role, as suggestedby several workers (Pond and Harris 1996, Epsteinet al. 1998, Prahl et al. 2003). Rontani et al. (2004b)recently showed that Chrysotila lamellosa can degradealkenones via epoxidation, adding support to this hy-pothesis. Cell fractionations show that LBs consist pri-marily of PULCA, whereas chloroplasts contain asignificant pool as well. This idea is supported byelectron micrographs that show conspicuous osmophilic

LBs in both the cytoplasm and plastids. Although wecannot rule out a pool of PULCA in cell membranes,our results suggest most of this lipid resides in LBs.

LB ultrastructure. Most eukaryotic cells produceLBs consisting of a hydrophobic core of neutral lipids(typically containing TAGs but also sterol esters orwax esters) surrounded by a phospholipid-proteincoat (Zweytick et al. 2000, Murphy 2001). Some alsocontain sterols or carotenoids (Zweytick et al. 2000).Our fractionations suggest that I. galbana andE. huxleyi have LBs consisting mostly of neutralPULCA but with some sterol or phospholipids,although there is no evidence of a membrane. Lightmicrographs and lipid separations also suggest somepigment associated with the LBs. In other algae,carotenoids can co-occur with neutral lipids in struc-tures such as chromoplasts (Murphy 2001).

Lipid bodies have not been well studied for hapto-phytes, although they are common in other algae(Cooksey et al. 1987, Dempster and Sommerfeld1998). This may be due to the assumption that chry-solaminarin, a b1 ! 3 glucan found in chloroplast

FIG. 4. Emiliania huxleyi 1742 LBand PULCA production under nu-trient depletion over a 10-day incu-bation. (a) Examples of NR-stainedcells in exponential phase (day 0,left) or late stationary phase (day10, right). (b) TLC of total lipidextracts on days 0, 4, 5, 7, and 10.Lipid standards consist of (bottom totop) phosphatidylcholine, phospha-tidylglycerol, digalactosyl diacylgly-ceride, cholesterol, TAG, and fattyacid methyl ester. (c) Cell numbers(circles) and major lipid pools (bars),as quantified by TLC densitometry.Stationary phase on days 4–10 isdenoted by gray shading.

M. L. ELTGROTH ET AL.1004

pyrenoids, is the main storage metabolite in coccolitho-phorids (Pienaar 1994). Authors previously describingthe ultrastructure of E. huxleyi have failed to identifyneutral lipid droplets, although their images showosmophilic bodies that resemble ours strikingly, parti-cularly for cells in nutrient depletion (Wilbur andWatabe 1963 [fig. 19], Klaveness 1972 [fig. 2]). A

more recent review of cell ultrastructure (Pienaar1994) noted that lipid droplets are often locatedadjacent to the plastid, as we observed, and becomemore abundant as the cells age, but no specific studieswere cited. Liu and Lin (2001) provided the firstultrastructural description of cytosolic LBs in I. gal-bana. They documented increasing LB production instationary phase and suggested these are synthesizedin the chloroplast before being exported to the cyto-plasm. Curiously, they did not connect the large LBs toPULCA, speculating instead that those neutral lipidswere part of cellular membranes.

PULCA location and biosynthesis. Neutral lipidsynthesis is typically associated with membrane-bound enzymes, and possible sites of biosynthesisinclude thylakoids, chloroplast membranes, or theER. Several indirect lines of evidence suggest thatPULCA biosynthesis is associated with the chloro-plasts. Our TEM preparations usually showed LBsadjacent to chloroplast membranes and often showedincipient LBs in the chloroplasts. Our fractionationexperiments demonstrated that chloroplasts alwayscontained a large amount of PULCA, even in sta-tionary phase cultures with abundant cytosolic LBs.The increased production of PULCA under nutrientlimitation (Bell and Pond 1996, Pond and Harris1996, Epstein et al. 1998, Prahl et al. 2003, our

3000

2000

1000

0550 600 650 700 750

Emission λ (nm)

Flo

ures

cenc

e (a

rbit

rary

)

FIG. 5. Spectrofluorometric analysis of NR-stained Emilianiahuxleyi 1742 cells in light (upper curve) and after 72-h darkincubation (lower curve), showing near total loss of NR peak at625 nm.

FIG. 6. LB and PULCA mod-ification during prolonged dark-ness for Emiliania huxleyi 1742 (a)and Isochrysis galbana (b–d). (a)TLC of E. huxleyi 1742 total lipidextracts before (lane 1) and after(lane 2) 3 days of darkness. (b)TLC of I. galbana total lipid ex-tracts before (lane 1; day 0) andafter 3 days (lanes 2–3) or 7 days(lanes 4–5) in light or dark. HC,hydrocarbons (alkenes); EK,ethyl ketones; MK, methyl ke-tones. TLC lipid standards consistof (bottom to top) phosphati-dylcholine, phosphatidylglycerol,digalactosyl diacylglyceride, cho-lesterol, TAG, and fatty acidmethyl ester. (c, d) NR stainingof I. galbana cells before (c) andafter (d) dark incubation.

HAPTOPHYTE NEUTRAL LIPIDS 1005

observations) also suggests these neutral lipids mayfunction in chloroplasts as a sink for excess reducingpower, as hypothesized by Yammoto et al. (2000).

One possible interpretation is that chloroplastssynthesize PULCA, which are then packaged intoLBs and exported to the cytoplasm, as suggested byLiu and Lin (2001) for I. galbana. Our observations ofIsochrysis under both light and nutrient limitationsupport this hypothesis, as both growth-limiting statescaused a net movement of NR-staining neutral lipidfrom chloroplasts to conspicuous LBs, although totalneutral lipid increased in the light but decreased in thedark. Vascular plant chloroplasts often contain lipid

bodies (plastoglobuli) adjacent to thylakoids. Althoughplastoglobuli are generally not thought to be able tomove across chloroplast membranes, Guiamet et al.(1999) reported plastoglobuli were exported fromsenescing soybean chloroplasts to cytosolic sites ofcatabolism, and other reports suggest lipid bodiesmay be translocated across membranes, even fromcell to cell (Murphy 2001). Our micrographs clearlyshow that even in stationary phase cells, chloroplaststructure was well preserved, with conspicuous thyla-koid membranes, and no indication of plastid degra-dation. This suggests active production of neutrallipids by chloroplasts, rather than conversion orscavenging during senescence.

However, another possibility is that PUCLA aresynthesized in the ER, which then packages theminto LBs via budding or translocation of protein-targeted lipids, as for nonphotosynthetic organisms(Murphy and Vance 1999, Zweytick et al. 2000, Mur-phy 2001). Sawada and Shiraiwa (2004) found alke-nones and alkenoates in E. huxleyi ER fractions,denoted by the enzyme NADPH:cyt c reductase, andsuggested that the alkenones were present as mem-brane-unbound micelles. Our evidence supports theirhypothesis that PULCA are not primarily membranelipids but strongly suggests they are contained in lipiddroplets rather than micelles. Sawada and Shiraiwa(2004) also found only small PULCA pools in thechloroplast thylakoids, in contrast to our results.They used Percollt rather than our sucrose gradients,and we suspect that different fractionation techniquesled to our different results. In particular, our ‘‘chlor-oplast’’ fraction, denoted by pigment, probably in-cluded other internal membranes such as ER andGolgi as well. Jeffrey and Anderson (2000) showedE. huxleyi ER membranes are continuous with the outermembranes of the chloroplast, so we cannot distin-guish PULCA produced in the ER from that producedin chloroplasts, and delivered via the ER into cytoplas-mic lipid vesicles. Sawada and Shiraiwa (2004) alsosuggested the coccolith-producing vesicle, denoted byuronic acid, was a major pool of alkenones andalkenoates. However, in our cultures PULCA biosynth-esis does not appear linked to calcification.

We found that PULCA components in LBs andchloroplasts appear to be similarly distributed. Sawadaand Shiraiwa (2004) also found similar results differentE. huxleyi cellular pools. However, in our experiments,UK0

37 predicted a growth temperature far lower than the161 C culture conditions for E. huxleyi 1742. Althoughthis might have been an artifact of separating cellfractions by overnight ultracentrifugation at 41 C,whole cells not subjected to centrifugation also hadthese unusual signatures. We have observed beforethat under N- or P-limited stationary phase, cellsaccumulate predominantly triunsaturated isomers, asreported for E. huxleyi (Epstein et al. 1998, Prahl et al.2003) and I. galbana (Versteegh et al. 2001). In parti-cular, the development of these lipids as a tool forpaleotemperature determination requires that we un-

FIG. 7. PULCA distributions in Isochrysis galbana cell fractio-nation experiments. (a) NR-stained buoyant fraction followingcell lysis and density centrifugation. Bright fluorescing spots areneutral lipid vesicles; faint NR crystals are also visible. (b) TLCanalysis of lipids from I. galbana. Lane 1: standard mixture(phosphatidylcholine, phosphatidylglycerol, cholesterol, TAG/fatty acid methyl ester poorly separated); lane 2: buoyant (LB)fraction; lane 3: chloroplasts; lane 4: whole cell lipids. PULCAcomponents: MK, methyl ketones (alkenones); EK, ethyl ke-tones; HC, hydrocarbons.

60%

50%

40%

30%

20%

10%

0%

frac

tion

of

tota

l

A36

:2m

A36

:3m

K37

:2m

K37

:3m

K37

:4m

K38

:2e

K38

:3e+

A36

:2e

K38

:2m

K38

:3m

K39

:2e

K39

:3e

PULCA component

E. huxleyi 1742LBsChlpsWhole Cells

FIG. 8. PULCA components of cell fractions for Emilianiahuxleyi 1742, shown as a fraction of total PULCA for buoyantlipid bodies, chloroplasts (Chlps), and whole cells. A36:2m,diunsaturated C36 methyl alkenoate; K37:2m, C37 methyl ketone(alkenone); K38:2m, C38 ethyl ketone (alkenone), etc.

M. L. ELTGROTH ET AL.1006

derstand their biosynthesis and desaturation by factorsother than growth temperature. Why these trans-un-saturated nonmembrane lipids should be desaturatedat colder temperature, if they do not contribute tomembrane fluidity, is an unanswered question. Cur-rently, little is known about the genetics of lipidbiosynthesis in these algae, but the sequencing of theE. huxleyi genome, now in progress (B. Read and T.Wahlund, personal communication) should greatlyhelp to identify genes involved with neutral lipidbiosynthesis and transport.

Why PULCA? Evolutionary and ecological aspects..What might be the fitness benefit of producing lipid

bodies containing such unusual neutral lipids? Ron-tani et al. (1997) found the trans-polyunsaturatedPULCA are notably resistant to photodegradationand are more stable to singlet oxygen than carote-noids, and Mouzdahir et al. (2001) showed the trans-polyunsaturated PULCA C37–38 alkenes are morephotostable than cis-polyunsaturated C31–33 alkenes,suggesting the double-bond geometry is a key factor.However, Mouzdahir et al. (2001) further noted thatthe alkenones’ extreme resistance to photodegrada-tion was higher than expected from stereogeometryalone and speculated they were membrane-unboundlipids. Our observations of the packaging of PULCAinto lipid bodies support that idea. Emiliania huxleyi in

TABLE 2. PULCA components (as a percentage of total PULCA) in different cell fractions.

Fraction

Emiliania huxleyi 1742 E. huxleyi 1516 Isochrysis galbana

WC LB Chlp WC LB Chlp WC LB Chlp

A36:2 Me 1 1 1 2 1 2 1 1 1A36:3 Me 2 2 2 1 1 1 2 2 2K37:2 Me 9 5 7 6 5 6 5 3 4K37:3 Me 49 53 52 39 42 40 54 57 58K37:4 Me 4 6 4 0 0 0 3 4 4K38:2 Et 7 5 6 9 8 8 8 5 5K38:3 Et þ A36:2 Et 20 22 21 29 30 30 24 25 24K38:2 Me 2 1 2 1 1 1 0 0 0K38:3 Me 3 1 2 8 8 8 1 1 1K39:2 Et 1 0 0 1 1 1 0 0 0K39:3 Et 2 3 2 3 3 3 1 2 2

WC, whole cells; LB, lipid bodies; Chlp, chloroplasts. A36:2 Me, C36:2 methyl alkene; K38:2 Et, C38:2 ethyl ketone (alkenone), etc.

FIG. 9. Ultrastructure of Emiliania huxleyi 1516 cells. (a, d)Osmium thiocarbohydrazide-osmium stained sections. (b, c)Standard staining. LB, lipid body; N, nucleus; M, mitochon-drion; G, Golgi; Chlp, chloroplast. Magnification, 7000�.

FIG. 10. Emiliania huxleyi 1516 ultrastructure (detail). (a, b)Osmium thiocarbohydrazide-osmium stained sections. Arrowsshow lipid bodies adjacent to chloroplast membrane. (c, d)Standard staining. Arrows show plastid LBs on thylakoid mem-branes. LB, lipid body; Chlp, chloroplast.

HAPTOPHYTE NEUTRAL LIPIDS 1007

particular thrives under high-light low-nutrient con-ditions (Nanninga and Tyrrell 1996) and shows littlephotoinhibition at high irradiances (Paasche 2002,and references therein). Therefore, PULCA might beevolutionarily favored over TAGs because they are amore photostable form of energy storage.

Finally, we suggest these unusual lipids may alsohave a trophic benefit for their producers. We ob-served during grazing of E. huxleyi by the phagotrophicdinoflagellate Oxyrrhis marina that PULCA componentswere degraded slowly, if at all, by O. marina after near-total consumption of prey cells (not shown). Only slightchanges in component profiles were recovered after 24h from predator food vacuoles. Experiments with theciliate Helicostomella showed similar results, thoughgrazing rates were low (S. Strom, personal commu-nication). We speculate that the unusual geometry ofthese molecules may limit their degradation by gra-zers, making PULCA in effect ‘‘algal Olestra.’’ Emilianiais a cosmopolitan taxon that forms notable bloomsworldwide, and high-light, low-nutrient blooms maybe trophically unsuitable due to accumulation ofPULCA. Although Isochrysis is a widely used aquacul-ture food stock, noted for its highly beneficialpolyunsaturated fatty acids content (Sukenik and Wah-non 1991), neutral lipid production in haptophytes ishighly strain specific (Conte et al. 1998), and thetrophic quality of wild strains likely reflects the balancebetween polyunsaturated cis-membrane and trans-neutral lipids. A long-standing issue for the use ofthese compounds in paleothermometry is how UK0

37 ismodified by passage through the food web andsedimentary deposition (Prahl et al. 1993, Teeceet al. 1994, 1998). We suggest that grazer modification,like sedimentary diagenesis (Prahl et al. 1993, Teeceet al. 1994, 1998), may have minimal impact onUK0

37 signatures, and future work should focus ongenetic and environmental factors affecting bio-synthesis.

We thank Fred Prahl and Margaret Sparrow for GC analysis ofPULCA fraction components, and Suzanne Strom and KelleyBright for grazing experiments. John Mahoney providedadvice on cell fractionations, and Richard Demaree helpedwith TEM preparations. Maureen Conte shared unpublisheddata, and Morris Kates provided advice on lipid separations.Two anonymous reviewers provided helpful comments on themanuscript. This work was supported by NSF grants OCE-9986306, OCE-0350359, and DUE-0126618.

Bell, M. V. & Pond, D. 1996. Lipid composition during growth ofmotile and coccolith forms of Emiliania huxleyi. Phytochemistry41:465–71.

Bligh, E. G. & Dyer, W. J. 1959. A rapid method of total lipidextraction and purification. Can. J. Biochem. Physiol. 37:911–7.

Brassell, S. C., Eglinton, G., Marlowe, I. T., Pflaumann, U. &Sarnthein, M. 1986. Molecular stratigraphy: a new tool forclimatic assessment. Nature 320:129–33.

Brzezinski, M. A., Reed, D. C. & Amsler, C. D. 1993. Neutral lipidsas major storage products in zoospores of the giant kelpMacrocystis pyrifera (Phaeophyceae). J. Phycol. 29:16–23.

Conte, M. H. & Eglinton, G. 1993. Alkenone and alkenoatedistributions within the euphotic zone of the eastern North

Atlantic: correlation with production temperature. Deep-SeaRes. 40:1935–61.

Conte, M. H., Thompson, A., Eglinton, G. & Green, J. C. 1995.Lipid biomarker diversity in the coccolithophorid Emilianiahuxleyi (Prymnesiophyceae) and the related Gephyrocapsa ocea-nica. J. Phycol. 31:272–82.

Conte, M. H., Thompson, A., Lesley, D. & Harris, R. P. 1998.Genetic and physiological influences on the alkenone/alkeno-ate versus growth temperature relationship in Emilianiahuxleyi & Gephyrocapsa oceanica. Geochim. Cosmochim. Acta 62:51–68.

Cooksey, K. E., Guckert, J. B., Williams, S. W. & Callis, P. R. 1987.Fluorometric determination of the neutral lipid content ofmicroalgal cells using Nile Red. J. Microbiol. Methods 6:333–45.

de Leeuw, J. W., van der Meer, F. W., Rijpstra, W. I. C. & Schenck,P. A. 1980. Unusual long chain ketones of algal origin. InDouglas, A. G. & Maxwell, J. R. [Eds.] Advances in OrganicGeochemistry 1979. Pergamon, Oxford, pp. 211–7.

Dempster, T. A. & Sommerfeld, M. R. 1998. Effects of environ-mental conditions on growth and lipid accumulation inNitzschia communis (Bacillariophyceae). J. Phycol. 34:712–21.

Epstein, B. L., D’Hondt, S. & Hargraves, P. E. 2001. The possiblemetabolic role of C37 alkenones in Emiliania huxleyi. Org.Geochem. 32:867–75.

Epstein, B. L., D’Hondt, S., Quinn, J. G., Zhang, J. & Hargraves, P.E. 1998. The effect of dissolved nutrient concentrations onalkenone-based temperature estimates. Paleoceanography13:122–6.

Fabregas, J., Cid, A., Torres, E., Sukenik, A. & Herrero, C. 1998.Effects of nitrogen source and growth phase on proximatebiochemical composition, lipid classes and fatty acid profile ofthe marine microalga Isochrysis galbana. Aquaculture 166:105–16.

Guiamet, J. J., Pichersky, E. & Nooden, L. D. 1999. Mass exodusfrom senescing soybean chloroplasts. Plant Cell Physiol.40:986–92.

Guyton, J. R. & Klemp, K. E. 1988. Ultrastructural discriminationof lipid droplets and vesicles in atherosclerosis: value ofosmium-thiocarbohydrazide-osmium and tannic acid-para-phenylenediamine techniques. J. Histochem. Cytochem.36:1319–28.

Henderson, R. J. & Sargent, J. R. 1989. Lipid composition andbiosynthesis in ageing cultures of the marine cryptomonad,Chroomonas salina. Phytochemistry 28:1355–61.

Hodgson, P. A., Henderson, R. J., Sargent, J. R. & Leftley, J. W.1991. Patterns of variation in the lipid class and fatty acidcomposition of Nannochloropsis oculata (Eustigmatophyceae)during batch culture. I. The growth cycle. J. Appl. Phycol.3:169–81.

Jeffrey, S. W. & Anderson, J. M. 2000. Emiliania huxleyi (Hapto-phyta) holds promising insights for photosynthesis. J. Phycol.36:449–52.

Kamisaka, Y., Noda, N., Sakai, T. & Kawasaki, K. 1999. Lipidbodies and lipid body formation in an oleaginous fungus,Mortierella ramanniana var. angulispora. Biochim. Biophys. Acta1438:185–98.

Kimura, K., Yamaoka, M. & Kamisaka, Y. 2004. Rapid estimationof lipids in oleaginous fungi and yeasts using Nile redfluorescence. J. Microbiol. Methods 56:331–8.

Klaveness, D. 1972. Coccolithus huxleyi (Lohm.) Kamptn. II. Theflagellate cell, aberrant cell types, vegetative propagation andlife cycles. Br. Phycol. J. 7:309–18.

Liu, C.-P. & Lin, L.-P. 2001. Ultrastructural study and lipid forma-tion of Isochrysis sp. CCMP1324. Bot. Bull. Acad. Sinica 42:207–14.

Marlowe, I. T., Brassell, S. C., Eglinton, G. & Green, J. C. 1984a.Long chain unsaturated ketones and esters in living algae andmarine sediments. Org. Geochem. 6:135–41.

Marlowe, I. T., Green, J. C., Neal, A. C., Brassell, S. C., Eglinton, G.& Course, P. A. 1984b. Long chain (n-C37-C39) alkenones inthe prymnesiophyceae. Distribution of alkenones and otherlipids and their taxonomic significance. Br. Phycol. J. 19:203–16.

M. L. ELTGROTH ET AL.1008

Mouzdahir, A., Grossi, V., Bakkas, S. & Rontani, J.-F. 2001. Visiblelight-dependent degradation of long-chain alkenes in killedcells of Emiliania huxleyi and Nannochloropsis salina. Phytochem-istry 56:677–84.

Muller, P. J., Kirst, G., Ruhland, G., von Storch, I. & Rosell-Mele, A.1998. Calibration of the alkenone paleotemperature indexUk037 based on core-tops from the eastern South Atlantic andthe global ocean (601N–601S). Geochim. Cosmochim. Acta62:1757–72.

Murphy, D. J. 2001. The biogenesis and functions of lipid bodies inanimals, plants and microorganisms. Prog. Lipid Res. 40:325–438.

Murphy, D. J. & Vance, J. 1999. Mechanisms of lipid-body forma-tion. Trends Biochem. Sci. 24:109–15.

Nanninga, H. J. & Tyrrell, T. 1996. Importance of light for theformation of algal blooms by Emiliania huxleyi. Mar. Ecol. Prog.Ser. 136:195–203.

Olsen, R. E. & Henderson, R. J. 1989. The rapid analysis of neutraland polar marine lipids using double-development HPTLCand scanning densitometry. J. Exp. Mar. Biol. Ecol. 129:189–97.

Paasche, E. 2002. A review of the coccolithophorid Emiliania huxleyi(Prymnesiophyceae), with particular reference to growth,coccolith formation, and clacification-photosynthesis interac-tions. Phycologia 40:503–29.

Patterson, G. W., Tsitsa-Tsardis, E., Wikfors, G. H., Gladu, P. K.,Chitwood, D. J. & Harrison, D. 1994. Sterols and alkenones ofIsochrysis. Phytochemistry 35:1233–6.

Pienaar, R. N. 1994. Ultrastructure and calcification of coccolitho-phores. In Winter, A. & Siesser, W. G. [Eds.] Coccolithophores.Cambridge University Press, Cambridge, pp. 13–37.

Pond, D. W. & Harris, R. P. 1996. The lipid composition of thecoccolithophore Emiliania huxleyi and its possible ecophysiolo-gical significance. J. Mar. Biol. Assoc. UK 76:579–94.

Prahl, F. G., Collier, R. B., Dymond, J., Lyle, M. & Sparrow, M. A.1993. A biomarker perspective on prymnesiophyte productiv-ity in the northeast Pacific Ocean. Deep-Sea Res. 40:2061–76.

Prahl, F. G., Muelhausen, L. A. & Zahnle, D. L. 1988. Furtherevaluation of long-chain alkenones as indicators of paleocea-nographic conditions. Geochim. Cosmochim. Acta 52:2303–10.

Prahl, F. G. & Wakeham, S. G. 1987. Calibration of unsaturationpatterns in long-chain ketone compositions for paleotempera-ture assessment. Nature 330:367–9.

Prahl, F. G., Wolfe, G. V. & Sparrow, M. A. 2003. Physiologicalimpacts on alkenone paleothermometry. Paleoceanogaphy18:1025–31.

Rechka, J. A. & Maxwell, J. R. 1988a. Characterisation of alkenonetemperature indicators in sediments and organisms. Org.Geochem. 13:727–34.

Rechka, J. A. & Maxwell, J. R. 1988. Unusual long chain ketones ofalgal origin. Tetrahedron Lett. 29:2599–600.

Riley, G., Teece, M. A., Peakman, T. M., Raven, A. M., Greene, K.J., Clarke, T. P., Murray, M., Leftley, J. W., Campbell, C. N.,Harris, R. P., Parkes, R. J. & Maxwell, J. R. 1998. Long-chainalkenes of the haptophytes Isochrysis galbana and Emilianiahuxleyi. Lipids 33:617–25.

Rontani, J.-F., Beker, B. & Volkman, J. K. 2004a. Long-chainalkenones and related compounds in the benthic hapto-phyte Chrysotila lamellosa Anand HAP 17. Phytochemistry 65:117–26.

Rontani, J.-F., Beker, B. & Volkman, J. K. 2004b. Regiospecificoxygenation of alkenones in the benthic haptophyte Chrysotilalamellosa Anand HAP 17. Phytochemistry 65:3269–78.

Rontani, J.-F., Cuny, P., Grossi, V. & Beker, B. 1997. Stability oflong-chain alkenones in senescing cells of Emiliania huxleyi:effect of photochemical and aerobic microbial degradation onthe alkenone unsaturation ratio (UK0

37). Org. Geochem. 26:503–9.Rontani, J.-F., Marchand, D. & Volkman, J. K. 2001. NaBH4

reduction of alkenones to the corresponding alkenols: a usefultool for their characterisation in natural samples. Org. Geochem.32:1329–41.

Sawada, K. & Shiraiwa, Y. 2004. Alkenone and alkenoic acidcompositions of the membrane fractions of Emiliania huxleyi.Phytochemistry 65:1299–307.

Schneider, J. C. & Roessler, P. 1994. Radiolabeling studies of lipidsand fatty acids in Nannochloropsis (Eustigmatophyceae), anoleaginous marine alga. J. Phycol. 30:594–8.

Sukenik, A. & Wahnon, R. 1991. Biochemical quality of marineunicellular algae with special emphasis on lipid composition. I.Isochrysis galbana. Aquaculture 97:61–72.

Teece, M. A., Getliff, J. M., Leftley, J. W., Parkes, R. J. & Maxwell, J.R. 1998. Microbial degradation of the marine prymnesiophyteEmiliania huxleyi under oxic and anoxic conditions as a modelfor early diagenesis: long chain alkadienes, alkenones andalkyl alkenoates. Org. Geochem. 29:863–80.

Teece, M. A., Maxwell, J. R., Getliff, J. M., Parkes, R. J., Briggs, D.E. G. & Leftley, J. W. 1994. Laboratory degradation of lipids ofthe marine prymnesiophyte Emiliania huxleyi and significancefor sediment studies. In Eglinton, G. & Kay, R. L. F. [Eds.]Biomolecular Palaeontology: Lyell Meeting Volume. NERC Spec.Publ. 94/1, Natural Environment Research Council, Swindon,UK, pp 5–8.

van der Wal, P., Leunissen-Bijvelt, J. J. M. & Verkleij, A. J. 1985.Ultrastructure of the membranous layers enveloping the cell ofthe coccolithophorid Emiliania huxleyi. J. Ultrastruct. Res. 91:24–9.

Versteegh, G. J. M., Riegman, R., de Leeuw, J. W. & Jansen,J. H. F. 2001. UK0

37 values for Isochrysis galbana as a function ofculture temperature, light intensity and nutrient concentra-tions. Org. Geochem. 32:785–94.

Volkman, J. K., Eglinton, G., Corner, E. D. S. & Forsberg, T. E. V.1980a. Long-chain alkenes and alkenones in the marinecoccolithophorid Emiliania huxleyi. Phytochemistry 19:2619–22.

Volkman, J. K., Eglinton, G., Corner, E. D. S. & Sargent, J. R.1980b. Novel unsaturated straight-chain methyl and ethylketones in marine sediments and a coccolithophore Emilianiahuxleyi. In Douglas, A. G. & Maxwell, J. R. [Eds.]. Advances inOrganic Geochemistry 1979. Pergamon, Oxford, pp. 219–27.

Wilbur, K. M. & Watabe, N. 1963. Experimental studies oncalcification in molluscs and the alga Coccolithus huxleyi. Ann.N. Y. Acad. Sci. 109:82–112.

Willingham, M. C. & Rutherford, A. V. 1984. The use of osmium-thiocarbohydrazide-osmium (OTO) and ferricyanide osmium-related methods to enhance membrane contrast and preserva-tion in cultured cells. J. Histochem. Cytochem. 32:455–60.

Yamamoto, M., Shiraiwa, Y. & Inouye, I. 2000. Physiologicalresponses of lipids in Emiliania huxleyi and Gephyrocapsa oceanica(Haptophyceae) to growth status and their implications foralkenone paleothermometry. Org. Geochem. 31:799–811.

Zweytick, D., Athenstaedt, K. & Daum, G. 2000. Intracellular lipidparticles of eukaryotic cells. Biochim. Biophys. Acta 1469:101–20.

HAPTOPHYTE NEUTRAL LIPIDS 1009