APPLIED MICROBIAL AND CELL PHYSIOLOGY

A shift to 50°C provokes death in distinct waysfor glucose- and oleate-grown cells of Yarrowia lipolytica

Thi Minh Ngoc Ta & Lan Cao-Hoang & Cynthia Romero-Guido & Morgane Lourdin &

Hanh Phan-Thi & Sébastien Goudot & Pierre-André Marechal & Yves Waché

Received: 1 July 2011 /Revised: 22 July 2011 /Accepted: 5 August 2011# Springer-Verlag 2011

Abstract Based on the observation that shocks provokedby heat or amphiphilic compounds present some similari-ties, this work aims at studying whether cells grown onoleate (amphiphilic pre-stress) acquire a tolerance to heatshock. In rich media, changing glucose for oleate signifi-cantly enhanced the cell resistance to the shock, however,cells grown on a minimal oleate medium lost their ability togrow on agar with the same kinetic than glucose-growncells (more than 7-log decrease in 18 min compared with 3-log for oleate-grown cells). Despite this difference inkinetics, the sequence of events was similar for oleate-grown cells maintained at 50°C with a (1) loss of ability toform colonies at 27°C, (2) loss of membrane integrity and(3) lysis (observed only for some minimal-oleate-growncells). Glucose-grown cells underwent different changes.Their membranes, which were less fluid, lost their integrity

as well and cells were rapidly inactivated. But, surprisingly,their nuclear DNAwas not stained by propidium iodide andother cationic fluorescent DNA-specific probes but becamestainable by hydrophobic ones. Moreover, they underwent adramatic increase in membrane viscosity. The evolution oflipid bodies during the heat shock depended also on thegrowth medium. In glucose-grown cells, they seemed tocoalesce with the nuclear membrane whereas for oleate-grown cells, they coalesced together forming big dropletswhich could be released in the medium. In some rare casesof oleate-grown cells, lipid bodies were fragmented andoccupied all the cell volume. These results show that heattriggers programmed cell death with uncommon hallmarksfor glucose-grown cells and necrosis for methyl-oleate-grown cells.

Keywords Heat . Membrane integrity. Membrane fluidity.

Fluorescence . Yarrowia lipolytica . Programmed cell death .

Lipid . Oleate . Chromatin changes

Introduction

Understanding the sequence of events resulting in cell deathafter a heat shock is, as all questions concerning the life anddeath of cells, an important concern of modern biology. It isof particular interest in microbiology, field in which theincomplete inactivation of pathogens or other contaminantsresults in health and industrial problems. The resistance ofcells to an increase in temperature depends to a large extenton the physiological state of the cell and on the subsequentphysiological response. In eukaryotic cells, the response toheat stress is relatively conserved from yeast to mammals(Jenkins 2003). This includes, among others, the productionof heat shock proteins, the accumulation of trehalose, a

Ta and Cao-Hoang contributed equally to this work.

T. M. N. Ta : L. Cao-Hoang :C. Romero-Guido :M. Lourdin :H. Phan-Thi : S. Goudot : P.-A. Marechal :Y. Waché (*)Laboratory GPMA, IFR92,Université de Bourgogne and AgroSup Dijon,1, Esplanade Erasme,21000 Dijon, Francee-mail: ywache@u-bourgogne.fr

L. Cao-HoangDepartment of Food Technology,School of Biological and Food Technology,Hanoi University of Science and Technology,1 Dai Co Viet Road,Hanoi, Vietnam

Present Address:T. M. N. TaInstitute of Biotechnology and Environment,Nha Trang University, 02 Nguyễn Đình Chiểu,Nha Trang, Khánh Hòa, Vietnam

Appl Microbiol BiotechnolDOI 10.1007/s00253-011-3537-3

heat-induced transient G0/G1 cell cycle arrest and thesphingosine activation of an ubiquitin protein degradationpathway. A major impact of temperature is the denaturationand aggregation of proteins which accumulate, saturatingthe ubiquitin- and proteasome-degradation systems (Riezman2004). In some cases, heat triggers an apoptotic response.The occurrence of such a reaction is common and welldocumented for multicellular organisms but less defined forunicellular cells (Madeo et al. 2004).

The plasma membrane plays an important role in theresponse to heat shock. Important modifications take placein this structure in response to heat (e.g. synthesis ofsphingolipids (Jenkins 2003) which can include modifica-tion of the level of sterols (Gaspar et al. 2008; Jones et al.2004). In a physical point of view, the plasma membranehas to keep its role of selective barrier at a temperaturewhich can cause a fluidisation of the membrane lipid phase,an increased diffusion of solutes such as protons throughthe lipid bilayer, inactivation of transport proteins and thesubsequent alteration of energetic gradients (Tolner et al.1997; van de Vossenberg et al. 1999). Interestingly, takenglobally, the effect of temperature on membranes presentssome similarities with the effect of lipids. Lipids present inthe cell environment will accumulate in the membranebilayer causing thereby structural and physicochemicalchanges (Weber and de Bont 1996). In particular anddepending on their own structure, they will alter themembrane fluidity or phase transition temperature, thepermeability and, as a result of structural changes, theactivity of enzymes embedded into the bilayer. The effecton cell physiology depends on the structure of theconsidered lipids. Lipids possessing one polar end will beeasily incorporated into the membrane, whereas compoundspossessing more polar groups will be extracted by mem-branes depending on the size and arrangement of the apolarand polar regions (Subczynski and Wisniewska 2000).Moreover, depending on the chain length and location ofthe compound in the bilayer, they will exhibit a verydifferent toxic effect. Short-chain monopolar compoundswill localise in the polar head region and this will changethe balance between the apolar centre and the polar edge ofthe bilayer, causing a destabilisation of the membranewhereas long-chain compounds will change the membranefluidity without perturbing the general arrangement of themembrane (Aguedo et al. 2002, 2003a, b; Weber and de Bont1996). Similarities between heat and chemical shocks havealready been observed concerning the response of Saccha-romyces cerevisiae to ethanol (Piper 1995). However, amajor difference between long-chain lipids and ethanol is thecatabolisation of these substrates. Cells growing in presenceof oleate are submitted to a fluidising stress due to theincorporation of oleate molecules into the membrane butdespite this stress, oleate is a major substrate for yeasts that

are able to grow on this substrate even at high concentrationwhereas ethanol becomes toxic at far lower concentrations.

The aim of the present study is to investigate whether theimpact of a temperature shift is the same for cells grown onoleate than for cells grown on glucose. The focus willparticularly be on the membrane by monitoring its integrity,fluidity and energetics as well as on the ability of cells toform colonies, on their reducing capacity and on celldamages. The yeast Yarrowia lipolytica will be used as amodel system as it is well adapted to growth on amphiphiliccompounds and the effect of these compounds on itsmembrane has already been investigated (Ta et al. 2010;Romero-Guido et al. 2011).

Material and methods

Strain, media and culture conditions

The strain Y. lipolytica W29 (ATCC20460; CLIB89) wascultured in YPDA medium (yeast extract, 10 gL−1; peptone,20 gL−1; dextrose, 20 gL−1; and agar, 15 gL−1) at 27°C for48 h. After this culture, cells were used to inoculate (A600=0.25 corresponding to 6.5 106 cells mL−1) various liquidmedia: YPD (YPDAwithout agar), YPO (YPD with glucosereplaced by methyl oleate, 5 gL−1 and Tween 80, 0.2 gL−1)or YNBO (yeast nitrogen base, 6.7 gL−1; methyl oleate, 5 gL−1; Tween 80, 0.2 gL−1) and grown at 27°C, 140 rpm inbaffled Erlenmeyer flasks as described earlier (Groguenin etal. 2004; Waché et al. 2001, 2000). Cells were harvested inthe mid-logarithmic growth phase (19 h for YPD, 24 h forYNBO and YPO) and washed twice in phosphate buffersaline (PBS) at pH 5.6 in 25 mM and finally resuspended toA600=0.7 in PBS. Methyl oleate was chosen as the source oflipids as it is the lipid source the most commonly used forlipid metabolism studies and Tween 80 as it is efficient toemulsion methyl oleate and as it does not bring other lipidicmoieties to the medium than oleate moieties.

Heat treatment

Cell suspensions at A600=0.7 in PBS buffer (about 2×107

cells mL−1) were shifted from 27°C to 50°C at 2°C min−1

and maintained at 50°C. The treated cells were harvestedand cooled to room temperature after different times formonitoring of cell number, colony forming ability andfluorescent staining.

Reducing capacity evaluation by methylene blue and abilityto form colonies by the CFU method

The cell suspension was mixed with an equal volume ofmethylene blue and incubated for 5 min at room temper-

Appl Microbiol Biotechnol

ature then visualised and counted using a Malassez cellunder light microscopy (Aguedo et al. 2004). Methyleneblue is assumed to colour all the cells. It can be reduced ormetabolised in metabolic active cells which become thennon-coloured. In the contrary, cells that are coloured in blueare cells having lost their reducing capacity. For the colonyforming unit method (CFU), successive decimal dilutionsof cellular suspensions were plated on YPDA and YNBGA(YNB, 6.7 gL−1; glucose, 20 gL−1; agar, 15 gL−1) andcolonies were counted after 2 days of culture at 27°C.

Fluorescent evaluation of membrane integrity, membraneenergetic, DNA and lipid bodies staining

Cells were exposed to lactone in PBS (20 mM, pH 5.6,A600=2.0) for 60 min then washed twice and resuspendedin the same buffer and adjusted to A600=0.7 for fluorescentstaining. For control of loss of membrane integrity, cellstreated with 3 gL−1 γ-dodecalactone for 90 min aspreviously described were used (Ta et al. 2010). Themembrane integrity marker propidium iodide (PI; Sigma-Aldrich, St. Quentin Fallavier, France) was prepared at200 μg mL−1 in distilled water; a working concentration of2 μg mL−1 was used. This fluorescent probe stains theDNA of the cells that lost their membrane integrity (Deereet al. 1998). In the case of bis-oxonol (BOX), a hydropho-bic fluorophore for membrane potential, marker of the cellmembrane energetic state, the stock solution was preparedin dimethylsulfoxide (DMSO) at 1 mg mL−1; the workingsolution was 5 μg mL−1. Lipid bodies were observed afternile red (NR) staining (Sigma-Aldrich). NR was prepared inacetone solution at a stock concentration of 1 mg mL−1 andused at a concentration of 1 μg mL−1. DNA staining wasevaluated using acridine orange (AO) with stock solution indistilled water at 5 mg mL−1 and used at a concentration of10 μg mL−1 or 4′,6-diamidino-2-phenylindole (DAPI; stocksolution in distilled water at 1 mg mL−1). The observationswere made using a multispectral confocal microscopyEclipse TE 2000E, Nikon, with He/Ar laser multiraiessystem. Excitation wavelengths were 561 nm for PI andNR, 514 for BOX and AO, respectively. The images werevisualised and taken using software EZ-C1 version 3.50.The cell-wall-specific stain calcofluor white (20 mg/L inthe cell suspension from a 1 mg/mL, stock solution indistilled water containing one drop of NaOH) was used insome cases for better visualisation of the cells.

Leakage of 260-nm absorbing compounds during heattreatment

Cells grown for 19 h in YPD medium were recovered bycentrifugation and washed with PBS buffer until theabsorbance at 260 nm of the supernatant was inferior to

0.1 (Sá-Correia et al. 1989). Cell suspensions at A600=0.7in PBS buffer were submitted to the heat treatment orincubated at 27°C (control). Samples collected at differenttimes were centrifuged (13,000 rpm, 5 min, 4°C), thesupernatant was transferred to a 1-ml cuvette and theabsorbance at 260 nm was determined.

Fluidity evaluation

The fluidity of lipid phases was evaluated through thefluorescent anisotropy using 1,6-diphenyl-1,3,5-hexatriene(DPH) as probe for the hydrophobic core of the membranesand its trimethyl ammonium derivative (TMA-DPH), whichhas a positively charged amino group anchoring thecompound at the membrane surface in contact with thewater. This latter probe could not be used for YNBO cellsdue to interactions between the probe and the cell envelope.The probes were prepared in dimethylformamide at a stockconcentration of 1 mM and conserved at −20°C. The cellsuspension was prepared as described above but resuspendedto a lower concentration (A600=0.6). After addition of 2 μl ofthe fluorescent probe at 1 mM into 3 mL cell suspension, thesample was placed in a thermostated chamber, protected fromlight, at 27°C and stirred. The measurements were made witha Jobin Yvon-Horiba model Fluorolog 3 spectrofluorometer.The excitation wavelength was set to 360 nm and theemission to 450 nm. The measured fluorescence intensitieswere corrected for background fluorescence and lightscattering from the unlabelled samples. Fluorescence anisot-ropy (r) was calculated as follows:

r ¼ IVV � GIVHð Þ= IVV þ GIVHð Þ;G ¼ IHV=IHH

where IVVand IVH are the fluorescence intensities determinedat vertical and horizontal orientations of the emissionpolarizer when the excitation polarizer is set in the verticalposition. This applies similarly for IHV and IHH with thehorizontal excitation polarizer. G is a correction factor forbackground fluorescence and light scattering (Aguedo et al.2003a; Ouvry et al. 2002)

Membrane fluidity was also evaluated by measuringfluorescence generalised polarisation of 2-dimethylamino-6-lauroyl-naphthalene (laurdan; Sigma). Laurdan fluores-cence emission spectra were recorded in the range from 400to 550 nm, using both 350- and 390-nm excitation wave-lengths, whereas the fluorescence excitation spectra wereobtained in the range from 320 to 420 nm, using both 440-and 490-nm emission wavelengths. Generalised polarisa-tion (GP) was calculated by the following formula:

GP ¼ I440 � I490ð Þ= I440 þ I490ð Þwhere I440, I490 are the intensities measured, at eachexcitation wavelength (from 320 to 420 nm), on the

Appl Microbiol Biotechnol

fluorescence excitation spectra obtained by fixed emissionwavelength of 440 and 490 nm, respectively. The mem-brane fluidity was evaluated as the GP value obtained at350 nm.

Results

Loss of ability to grow on agar plates affects primarily cellsgrown on glucose

After the heat shift to 50°C, the cells grown on YPD, YPO orYNBO were counted and observed on a Malassez cell tomonitor the evolution from the initial number (2×107

cells mL−1). The number of cells grown on glucose (YPD)remained constant for at least 3 h (results not shown). It canbe noted that cells undergoing the temperature shift had beenharvested in log growing phase and many cells wereobserved with their daughter cell attached to their surfaceeven after a 3-h incubation at 50°C, showing no shift to theG0 cell cycle stage. For cells grown on oleate, cell counts didnot significantly change during the 3 h at 50°C for cellsgrown on the rich YPO whereas the number of cells grownon the minimal YNBO medium decreased as early as duringthe temperature shift (13% reduction compared with theinitial number of cells when the temperature reached 50°C)and continued to decrease afterwards (about 20% reductionafter 3 h at 50°C). For oleate-grown cells also, associationsbetween mother and daughter cells were still observable after3 h (but in a lesser extent than for glucose-grown cells).

The ability of cells to form colonies on agar media is aclassical way to evaluate the physiological state of micro-organisms as cells in a state of quiescence, death or too muchdamaged cannot grow. To evaluate the effect of the heat shockon this property, the number of cells able to form coloniesduring the heat treatment (CFU) was monitored on rich(YPDA) (Fig. 1a) and minimal (YNBGA; data not shown)glucose agar media. Results were similar regardless of therecovery medium. Cells grown on glucose or minimal oleatemedium behaved similarly with about 30–40% loss of abilityto form colonies when reaching 50°C and no more detectablecolony after 18 min at this temperature. YPO cells were lessperturbed by the heat shock and the decline in the number ofcolony forming unit was of 3-log after 18 min and no morecolony was observed after 120 min.

The shift in temperature provokes a loss of membraneintegrity and structural changes

To get further inside the mechanisms explaining thedifference observed at the cellular level depending on thecarbon source, the behaviour of probes reporting the stateof the membrane was investigated. To evaluate whether the

loss of ability to grow was due to membrane damages, themembrane integrity probe propidium iodide was used. Thisnucleic acid marker compound possessing two positivecharges becomes membrane permeant only if membranesare damaged and have lost their integrity (Cao-Hoang et al.2008). The number of cells stained with PI is shown onFig. 1b. For glucose-grown cells, only 5% of the cells werestained by PI after 3-h incubation at 50°C compared to100% for cells grown on the two different lipid media. Themembrane integrity loss began after 18 min for lipidminimal medium and after 80 min for YPO-grown cells.For both cultures, this period corresponds to the time of

50°C 100

10

U (

%)

1

CF

U

0.127°C

0.01

0

initial 0 17 min 120 min

(%)

cells

ar

ked

Ma

120

100

80

60

40

20

0

0 30 60 90 120 150 180

Holding time at 50°C (min)

m0.2

A26

0nm

A

0.1

015050 1000

Time (min)

a

b

c

Fig. 1 Cell ability to form colonies (a) and membrane integrity asevaluated through PI marked cells (%) evolution during the shift andmaintained at 50°C (b). The temperature profile is given in (a) fromthe initial time (27°C, before the temperature shift) to the time at 50°C.For YPD cells, as the staining with PI was not positive, membraneintegrity was checked by monitoring the leakage of cellular content ofYPD cells heated at 50°C or kept at 27°C as a control (c). White squares,YPD; black squares, YPO; black up-pointing triangles, YNBO; andblack circles, YPD kept at 27°C. Data presented have been calculatedfrom at least six independent repetitions

Appl Microbiol Biotechnol

complete loss of ability to form colonies. To explain whyglucose-grown cells were not stained by PI despite theirlack of ability to form colonies, the loss of membraneintegrity was monitored by another method based onleakage of cellular content. The significant increase in theabsorbance at 260 nm for heated cells compared with non-heated control cells showed that despite the absence of PIstaining, the heat treatment provoked a loss in membraneintegrity (Fig. 1c). Cells were then carefully observed inconfocal microscopy after staining with PI (Fig. 2) and, ascontrols for DNA staining, with other DNA intercalatingagent such as DAPI, acridine orange (not shown) and thehydrophobic Hoechst 33342. Beside the 5% cells that werenormally stained by PI, the other cells remained unstainedeither by PI, DAPI or acridine orange (stainings with DAPIand acridine orange, which are similar to PI, are not shown)but were clearly stained by Hoechst 33342 in the nucleusand mitochondria (Fig. 2). The fact that the more polarstains did not stained cells suggests that nuclear DNA hadpartially lost its stainability. To check whether nuclear DNAhad become more hydrophobic, a further staining wascarried out with the apolar lipophilic nile red and with theanionic lipophilic bis-oxonol (Fig. 2). For YPD-grown cellsmaintained at 50°C, both compounds stained membranesand lipid bodies in the cell but a difference appeared in thestaining by these two probes as BOX coloured more the cellnucleus (Fig. 2). This nuclear staining was not observed foroleate-grown cells but in these conditions this could beexplained by the presence of lipid bodies. Indeed, the stainhad a high affinity for these organelles which could result inthe absence of staining of the other components of the cellsuch as DNA and mitochondria (results not shown).

Another relatively non-permanent probe was used to betterdefine the impact of the heat stress on the cells: methylene

blue. This probe is classically used to discriminate betweennon-coloured cells actively reducing methylene blue and cellscoloured in blue by the probe. Surprisingly, although thisprobe can stain Y. lipolytica cells after various stresses(Aguedo et al. 2004; Ta et al. 2010), the amount of cellsstained after the heat shock was very low: about 10% forYPD cells after 3 h and less than 1% for oleate-grown cells(results not shown). It can be noted that for dividing cells,the proportion of daughter cells stained by methylene blue orpropidium iodide was far higher than average while themother cells were not stained (results not shown).

The energetic state of the cell membranes was alsoinvestigated with the membrane potential probe bis-oxonol.This compound stained cells from the beginning of the timeat 50°C regardless of the culture medium. Actually, forunshocked cells only the part out of the polarised plasmamembrane (hydrophobic part of the cell wall) was coloured(Fig. 2D) and after the temperature shift, all the lipid-richparts of the YPD-grown cells were marked such asmitochondria or various membranes suggesting that cellshad lost both the plasma membrane gradient but also themitochondrial energy. For oleate-grown cells, mitochondriawere not observable although such organelles are numerousin these actively respiring cells (Fig. 2) but the absence ofcolouration of mitochondria in cells grown on this mediummight come from the presence of hydrophobic lipid bodiesfor which the probes possessed a higher affinity than formitochondria (Fig. 5B, C).

Membrane fluidity probes reveal different membranestructures depending on the substrate

As the kinetics of loss of membrane integrity depended onthe culture medium, we focussed on the membrane fluidity.

B C G H

E FD

HA

Fig. 2 Staining of YPD-grown cells with various fluorescentprobes. A Propidium iodide, control with cells after treatment withγ-dodecalactone for membrane permeabilisation, B PI after 4 h at50°C, C Hoechst 33342 after 4 h at 50°C, D bis-oxonol before

treatment, E BOX after 2 h at 50°C, F Nile red after 2 h at 50°C,untreated cells grown on YPD (G) or YPO (H) observed on TEM.Bar, 1 μm. Solid arrows, nucleus; dashed arrows, mitochondria

Appl Microbiol Biotechnol

To investigate this parameter, we used the apolar DPH andthe monopolar TMA-DPH (Fig. 3). DPH has an affinity forall lipid phases of the cells and diffuses easily throughmembranes. When localised in a membrane, its fluores-cence anisotropy accounts for the apolar part in the centreof the lipid bilayer. TMA-DPH is a derivative of DPHwhich possesses a cationic group anchoring the compoundat the polar head end of the membrane bilayer anddecreasing its diffusivity across membranes. In a mem-brane, its fluorescence anisotropy will account for the

fluidity of the membrane less in depth than DPH (Kaiserand London 1998; Pebay-Peyroula et al. 1994) and the factthat it is anchored on the polar head layer decreases theglobal movement of the compound. For both probes, theratio of the fluorescence anisotropy for lipid-grown cellswas lower than for glucose-grown cells, showing that theprobes were located in a more fluidic phase. Despite thedifference of initial value, the effect of temperature on thevariation of fluorescence anisotropy was equivalent for cellsgrown on the various media (Fig. 3A, B). Interestingly, thiseffect was very low for the polar TMA-DPH suggestingthat the ordered part of the membrane at the polar end wasmuch less sensitive to temperature than the less ordered partinside the hydrophobic core of the membrane (Fig. 3A, B).From this result, showing a low modification of themembrane surface in response to the heat shift, it seemedinteresting to evaluate fluidity with another probe anchoredto the membrane surface. Laurdan, a monopolar proberesponding, in addition to the fluidity of the environment,to its polarity was chosen (Fig. 3C). Surprisingly, theresponse of the probe was comparable to that of DPH witha marked phase variation for glucose-grown cells and tothat of TMA-DPH with only a slight change withtemperature with YNBO cells. Notably, the generalisedpolarisation was similar at the growth temperature regard-less of the culture medium (GP around 0.35 at 27°C).

Holding cells at 50°C triggers a rigidification of the DPHenvironment for glucose-grown cells

It is also particularly interesting to note that the DPH-fluorescence anisotropy of cells grown on YPD continued toincrease when the cells were maintained at 50°C (Fig. 4),whereas it was stable for lipid-grown cells. This shows adynamic modification of the environment of the probe in stableconditions of temperature suggesting either an increase in themembrane rigidity or a migration of the probe to a highly rigid

0.32

0.3

0.28

0.26

0.24Ani

sotr

opy

(r)

Ani

sotr

opy

(r)

0.22A

B

0.20 10 20 30 40 50

Temperature (°C)

0.25

0.20

0.15

0.10

0.05

00 10 20 30 40 50

Temperature (°C)

0.45

0.4

0.35

0.3

0.25

C0.2

Gen

eral

ised

pol

aris

atio

n (G

P)

15 20 25 30 35 40Temperature (°C)

Fig. 3 Effect of temperature on membrane fluidity as evaluated usingfluorescent probes: a TMA-DPH; b DPH; c laurdan. YPD (whitesquares); YPO (black squares); and YNBO (black up-pointingtriangles). The data presented are representative of the results obtainedin at least six independent repetitions

600.16

500.14

400.12

300.1

200.08

Ani

sotr

opy

(r)

100.06

00.040 500 1000 1500 2000

Time (s)

Fig. 4 Membrane fluidity evaluation during temperature shift andholding at 50°C using DPH: YPD (diamonds), YPO (black circles)and temperature (grey line). The data presented are representative ofthe results obtained in at least six independent repetitions

Appl Microbiol Biotechnol

phase. For DPH, confocal laser scanning microscopy coupledwith spectral analysis showed clearly that the probe was in theplasma membrane after incorporation (results not shown).

Heat shock stimulates the coalescence of lipidbodies or, in some cases, their fragmentation

Y. lipolytica is an oleaginous yeast which stores lipids inlipid bodies inside the cell. The observation of the changesin the structure of these droplets during the heat shockcould inform us on the physiological state of the cells.During log-phase, there are normally about two or three

small lipid bodies inside the cells grown on glucose, threeor four medium-sized lipid bodies in cells grown onminimal oleate medium and more than four big dropletsfor cells grown on YPO (Fig. 5). The effect of thetemperature shift on the lipid bodies’ structure was almostnot detectable for cells grown on glucose but a modificationof the localisation occurred and, after the shift, theseparticles were seen in the neighbourhood of the nuclearmembrane (Fig. 5). For oleate-grown cells, the modificationof the structure of the lipid bodies was notable. For cellsgrown on both oleate-media, the number of lipid bodiesinside the cells decreased after 1 to 3 h but the size of the

A

a b c d e

B

a b c d e

C

a c d

b c d e

b e

ddd50°C

D

a27°C 0 1h 2h 3h

Holding time at 50°C

Fig. 5 Evolution of lipid bodies during temperature shift and holdingat 50°C. AYPD, B YNBO and C, D YPO. Small letters refer to timeholding at 50°C: a before the shift, b arrived at 50°C, c 1 h, d 2 h and

e 3 h (see scheme at the bottom right of the figure). Solid arrows, LBs;dashed arrows, nuclear membrane. To facilitate observation, cellswere marked with calcofluor in some cases

Appl Microbiol Biotechnol

remaining droplets grew after that time showing that thelipid particles coalesced inside the cells (Fig. 5B, C).Moreover, for YNBO-grown cells, cell lysis resulted in theliberation of lipid bodies (and of the numerous lipiddroplets adsorbed on the cell surface after growth in thismedium) in the environment (Fig. 5B). We observed alsofor some YPO-grown cells a drastic fragmentation of thelipid bodies resulting in some cases in the covering of allthe cell area by lipid aggregates (Fig. 5D).

Discussion

The impact of growth in the presence of amphiphilic orhydrophobic compounds on the cell response to stresses hasnot been much investigated. Although long-chain monop-olar compounds are usually not toxic, they are extracted bymembranes and perturb their structure and homeoviscosity(Weber and de Bont 1996). Our results highlight this pointas the fluorescence anisotropy of DPH and TMA-DPH wassignificantly lower for cells grown on oleate. In correlationwith this observation, the staining by PI of heat-shockedoleate-grown cells was much higher, suggesting that thehigh fluidity could decrease membrane integrity. However,after additional tests for glucose-grown cells, it could beestablished that these cells had also lost their membraneintegrity. We investigated then whether beside the globalfluidity, a relationship could be established between thisloss of integrity and the membrane structure. The fluidity ofoleate-grown cells reflects probably the presence of oleateembedded in the membrane which modifies the structure ofthe bilayer. As methyl oleate fluidises both the hydrophobiccore and the polar ordered part of the membrane, it is likelythat the carboxylate-end anchors the compound in thephospholipid polar head part while the acyl-chain isembedded inside the apolar part. However, although thefluidity of the polar part is enhanced, the structure is notcompletely changed if we consider the rather low effect ofthe temperature shift on the fluidity of this region for bothglucose- and oleate-grown cells. We can suppose that themethyl oleate orientation, between the polar group and thecis double bond of the molecule, is more or less normal tothe membrane surface, perturbing thus fluidity but not theglobal order of the membrane polar edge. To investigatemore in depth this effect, we used laurdan as a polarlipophilic probe reflecting membrane fluidity near themembrane surface but also, through a dependence offluorescence to polarity, the interface state. Interestingly,this probe exhibited a rather different behaviour dependingon the culture medium. A phase transition was notable withglucose- but not with oleate-grown cells. It can be assumedthat the membrane interface is differently organised depend-ing on the culture medium especially at the polar interface

but it is also possible that, as a result from this differentstructure, laurdan locates more or less in depth. In that case,laurdan would be more in depth in a region undergoing anotable phase transition with an interface composed almostexclusively of phospholipids whereas, in the presence ofoleate, it would be more in the surface of the membrane inthe region less sensitive to the heat shift. These structuraldifferences at the edge of membranes might result indifferences in the membrane permeabilisation upon treat-ments. We have suggested that this modification could becaused by the incorporation of oleate into the bilayer but itis also possibly due to the cell adaptation to oleateinvolving a change in the membrane composition. Forinstance, we observed that, beside modification of cellenvelope properties (Aguedo et al. 2003c), cells grown onoleate possess more ergosterol than those grown on glucose(Ta et al. 2010).

Despite some similarities in the kinetics of permeabilisa-tion of the membrane and of loss of ability to form coloniesof YPD and YNBO cells, fundamental differences wereobserved. For instance, for YPD cells maintained at 50°C,DNA was poorly stained by propidium iodide (5% of thecells were stained). Stainability by PI is not only used toevaluate cell death but also apoptosis. Apoptotic cellsundergo normally hyperchromicity, however, there aredifferent stages as early apoptotic cells are only slightlyfluorescent and the hyperchromicity appears only at thefinal stages (Bedner et al. 1999). In our study, no PIhyperchromicity was observed even after long periods andit is rather unlikely that cells could stay at the early stage ofprogrammed cell death (PCD) for so long. It seems morelikely that this low staining observed for three DNAintercalating probes could be explained by changes in thestructure of DNA. Differential staining depending on thechromatin condensation has already been observed forhuman cells and the authors postulated that these differ-ences could be due to differences in the in situ accessibilityof DNA to PI (Bedner et al. 1997). Interestingly the nucleusof heat-shocked YPD cells that was not stainable by thecationic PI became stainable by the anionic and hydropho-bic bis-oxonol and, to a lesser extent, by the unchargedhydrophobic nile red. This suggests that the modificationsof the chromatin structure perturbing DNA stainability arerelated to hydrophobicity and electrostatic changes. Anoth-er specificity of glucose-grown cells was the dramaticincrease in membrane fluidity when exposed to 50°C. Sucha phenomenon has already been reported for apoptotichuman cells (Artwohl et al. 2008) and could be related tothe production of membrane degrading-reactive oxygenspecies as proposed by (Feng and Zhang 2004).

After a long period in which apoptosis was considered asrestricted to higher eukaryotic cells, many examples ofPCD were observed in yeast with induction by several

Appl Microbiol Biotechnol

sources (Büttner et al. 2006; Madeo et al. 1997; Madeo etal. 2004). However, contrasting with mammals, the induc-tion of apoptosis by heat in wild-type yeast has not beenobserved. Although the first mention of an apoptosis-likeprocess in yeast was related to a temperature modification,this concerned a sublethal temperature and a mutant strain(Madeo et al. 1997). From that report, heat shock was notmentioned as a cause of apoptosis-like process in wild-typeyeast. The interest of cellular suicide for a pluricellularorganism is well understood but it seems often less obviousfor unicellular organisms. In the evolution of a population,a character can be selected if it brings an improvement tothe population. PCD can favour the survival of non-damaged cells, increase genetic diversity through meioticrecombination and trigger genetic variation for adaptationof continuously changing conditions (Büttner et al. 2006).In our conditions, the 5–10% of the glucose-grown cellswho did not undergo the PCD process lost their membraneintegrity and therefore did not exhibit significantly longerability to form colonies. Cells grown on oleate survivedlonger and showed no signs of PCD. With the big lipidbodies occupying the cell volume, it could be possible thatthese structures would play a role in the response to the heatshock. For glucose-grown cells, lipid bodies were foundclose to the nuclear membrane after the shift. In oleate-grown cells, no such a movement was observed but lipidbodies coalesced giving a limited number of huge lipidbodies inside the cells. In rare cases, spectacular lipidbodies’ fragmentation was observed suggesting that lipidbodies possessing cells could also undergo PCD althoughhallmarks were different. The lipid metabolism could thusplay a role on the PCD response. The effect of the carbonsource on apoptosis has already been observed in responseto aspirin but the presence of lipids was not investigated(Balzan et al. 2004). For human cells, the inhibiting effectof ceramid protein kinase has been observed (Gómez-Muñoz et al. 2005) and the regulation of ceramid by lipidsis likely.

In conclusion, the presence of oleate in the mediumperturbs the membrane of yeast cells and induces thepresence of lipid bodies. Despite these structural modifica-tions, lipid-grown cells are more resistant to the heat shockthan glucose-grown cells. This study reports that heat shockat lethal temperature can induce a programmed cell deathprocess in yeast but that this process is modified by thepresence of oleate. It also highlights a very unusual DNAchange resulting in the absence of staining by most cationicDNA-specific probes but in the staining of nucleus by theanionic bis-oxonol.

Acknowledgements This work was partially funded by the FrenchAgency for Research (programme Transaronat). A grant was given toTa by the French Ministry for Research and to Phan-Thi by the

Agence Universitaire de la Francophonie. An important part of thework was carried out in the Spectral Imaging Platform of IFR 92. Theauthors are thankful to Christine Bernard-Rojas for technical help.

References

Aguedo M, Beney L, Waché Y, Belin J-M, Gervais P (2002)Interaction of odorous lactones with phospholipids: implicationsin toxicity towards producing yeast cells. Biotechnol Lett24:1975–1979

Aguedo M, Beney L, Waché Y, Belin J-M (2003a) Interaction of anodorant lactone with model phospholipid bilayers and its strongfluidizing action in yeast membrane. Int J Food Microbiol80:211–215

Aguedo M, Beney L, Waché Y, Belin J-M (2003b) Mechanismsunderlying the toxicity of lactone aroma compounds towards theproducing yeast cells. J Appl Microbiol 94:258–265

Aguedo M, Waché Y, Mazoyer V, Sequeira-Le Grand A, Belin J-M(2003c) Increased electron donor and electron acceptor charactersenhance the adhesion between oil droplets and cells of Yarrowialipolytica as evaluated by a new cytometric assay. J Agric FoodChem 51:3007–3011

Aguedo M, Waché Y, Coste F, Husson F, Belin J-M (2004) Impact ofsurfactants on the biotransformation of methyl ricinoleate intogamma-decalactone by Yarrowia lipolytica. J Mol Catal 29:31–36

Artwohl M, Lindenmair A, Sexl V, Maier C, Rainer G, FreudenthalerA, Huttary N, Wolzt M, Nowotny P, Luger A, Baumgartner-Parzer SM (2008) Different mechanisms of saturated versuspolyunsaturated FFA-induced apoptosis in human endothelialcells. J Lipid Res 49:2627–2640

Balzan R, Sapienza K, Galea DR, Vassallo N, Frey H, Bannister WH(2004) Aspirin commits yeast cells to apoptosis depending oncarbon source. Microbiology 150:109–115

Bedner E, Burfeind P, Gorczyca W, Melamed MR, Darzynkiewicz Z(1997) Laser scanning cytometry distinguishes lymphocytes,monocytes, and granulocytes by differences in their chromatinstructure. Cytometry 29:191–196

Bedner E, Li X, Gorczyca W, Melamed MR, Darzynkiewicz Z (1999)Analysis of apoptosis by laser scanning cytometry. Cytometry35:181–195

Büttner S, Eisenberg T, Herker E, Carmona-Gutierrez D, Kroemer G,Madeo F (2006) Why yeast cells can undergo apoptosis: death intimes of peace, love, and war. J Cell Biol 175:521–525

Cao-Hoang L, Marechal PA, Le-Thanh M, Gervais P, Waché Y (2008)Fluorescent probes to evaluate the physiological state and activityof microbial biocatalysts: a guide for prokaryotic and eukaryoticinvestigation. Biotechnol J 3:890–903

Deere D, Shen J, Vesey G, Bell P, Bissinger P, Veal D (1998) Flowcytometry and cell sorting for yeast viability assessment and cellselection. Yeast 14:147–160

Feng Z, Zhang J-t (2004) Protective effect of melatonin on β-amyloid-induced apoptosis in rat astroglioma c6 cells and its mechanism.Free Radic Biol Med 37:1790–1801

Gaspar ML, Jesch SA, Viswanatha R, Antosh AL, Brown WJ,Kohlwein SD, Henry SA (2008) A block in endoplasmicreticulum-to-Golgi trafficking inhibits phospholipid synthesisand induces neutral lipid accumulation. J Biol Chem283:25735–25751

Gómez-Muñoz A, Kong JY, Parhar K, Wang SW, Gangoiti P, GonzálezM, Eivemark S, Salh B, Duronio V, Steinbrecher UP (2005)Ceramide-1-phosphate promotes cell survival through activationof the phosphatidylinositol 3-kinase/protein kinase B pathway.FEBS Lett 579:3744–3750

Appl Microbiol Biotechnol

Groguenin A, Waché Y, Escamilla Garcia E, Aguedo M, Husson F,LeDall M-T, Nicaud J-M, Belin J-M (2004) Genetic engi-neering of the β-oxidation pathway in the yeast Yarrowialipolytica to increase the production of aroma compounds. JMol Catal 28:75–79

Jenkins GM (2003) The emerging role for sphingolipids in theeukaryotic heat shock response. Cell Mol Life Sci 60:701–710

Jones DL, Petty J, Hoyle DC, Hayes A, Oliver SG, Riba-Garcia I,Gaskell SJ, Stateva L (2004) Genome-wide analysis of the effectsof heat shock on a Saccharomyces cerevisiae mutant with aconstitutively activated cAMP-dependent pathway. Comp FunctGenomics 5:419–431

Kaiser RD, London E (1998) Location of diphenylhexatriene (DPH)and its derivatives within membranes: comparison of differentfluorescence quenching analyses of membrane depth. Biochem-istry 37:8180–8190

Madeo F, Fröhlich E, Fröhlich KU (1997) A yeast mutant showingdiagnostic markers of early and late apoptosis. J Cell Biol139:729–734

Madeo F, Herker E, Wissing S, Jungwirth H, Eisenberg T, Fröhlich KU(2004) Apoptosis in yeast. Curr Opin Microbiol 7:655–660

Ouvry A, Waché Y, Tourdot-Maréchal R, Diviès C, Cachon R (2002)Effects of oxidoreduction potential combined with acetic acid,NaCl and temperature on the growth, acidification, and mem-brane properties of Lactobacillus plantarum. FEMS MicrobiolLett 214:257–261

Pebay-Peyroula E, Dufourc EJ, Szabo AG (1994) Location of diphenyl-hexatriene and trimethylammonium-diphenyl-hexatriene in dipal-mitoylphosphatidylcholine bilayers by neutron diffraction. BiophysChem 53:45–56

Piper PW (1995) The heat shock and ethanol stress responses of yeastexhibit extensive similarity and functional overlap. FEMSMicrobiol Lett 134:121–127

Riezman H (2004) Why do cells require heat shock proteins to surviveheat stress? Cell Cycle 3:61–63

Romero-Guido C, Belo I, Ta TMN, Cao-Hoang L, Alchihab M,Gomes N, Thonart P, Teixeira JA, Destain J, Waché Y (2011)Biochemistry of lactone formation in yeast and fungi and itsutilisation for the production of flavour and fragrance com-pounds. Appl Microbiol Biotechnol 89:535–547

Sá-Correia I, Salgueiro SP, Viegas CA, Novais JM (1989) Leakageinduced by ethanol, octanoic and decanoic acids in Saccharomy-ces cerevisiae. Yeast 5:S123–S127

Subczynski WK, Wisniewska A (2000) Physical properties of lipidbilayer membranes: relevance to membrane biological functions.Acta Biochim Pol 47:613–625

Ta TMN, Cao-Hoang L, Phan-Thi H, Tran HD, Souffou N, Gresti J,Marechal PA, Cavin JF, Waché Y (2010) New insights into theeffect of medium chain length lactones on yeast membranes.Importance of the culture medium. Appl Microbiol Biotechnol87:1089–1099

Tolner B, Poolman B, Konings WN (1997) Adaptation of micro-organisms and their transport systems to high temperatures.Comp Biochem Physiol A Physiol 118:423–428

van de Vossenberg JL, Driessen AJ, da Costa MS, Konings WN(1999) Homeostasis of the membrane proton permeability inBacillus subtilis grown at different temperatures. BiochimBiophys Acta 1419:97–104

Waché Y, Laroche C, Bergmark K, Møller-Andersen C, Aguedo M, LeDall MT, Wang H, Nicaud J-M, Belin J-M (2000) Involvement ofacyl coenzyme A oxidase isozymes in biotransformation ofmethyl ricinoleate into γ-decalactone by Yarrowia lipolytica.Appl Environ Microbiol 66:1233–1236

Waché Y, Aguedo M, Choquet A, Gatfield IL, Nicaud J-M, Belin J-M(2001) Role of β-oxidation enzymes in γ-decalactone productionby the yeast Yarrowia lipolytica. Appl Environ Microbiol67:5700–5704

Weber FJ, de Bont JA (1996) Adaptation mechanisms of micro-organisms to the toxic effects of organic solvents on membranes.Biochim Biophys Acta 1286:225–245

Appl Microbiol Biotechnol