ORIGINAL ARTICLE

Lipidomic Analysis of Porcine Olfactory Epithelial Membranesand Cilia

Simona Lobasso • Patrizia Lopalco • Roberto Angelini •

Maristella Baronio • Francesco P. Fanizzi •

Francesco Babudri • Angela Corcelli

Received: 7 April 2010 / Accepted: 6 May 2010 / Published online: 29 May 2010

� AOCS 2010

Abstract The use of the matrix 9-aminoacridine has been

recently introduced in matrix-assisted laser desorption/

ionization time-of-flight (MALDI-TOF) mass spectrometry

analysis of both anionic and cationic phospholipids. In the

present study, we take advantage of this technique to

analyze the lipids of porcine olfactory mucosa and a

membrane fraction enriched in cilia. Thin-layer chroma-

tography (TLC) and 31P-NMR analyses of the lipid extracts

were also performed in parallel. MALDI-TOF-MS allowed

the identification of lipid classes in the total lipid extract

and individual lipids present in the main TLC bands. The

comparison between the composition of the two lipid

extracts showed that: (1) cardiolipin, present in small

amount in the whole olfactory mucosa lipid extract, was

absent in the extract of membranes enriched in olfactory

cilia, (2) phosphatidylethanolamine species were less

abundant in ciliary than in whole epithelial membranes,

(3) sulfoglycosphingolipids were detected in the lipid

extract of ciliary membranes, but not in that of epithelial

membranes. Our results indicate that the lipid pattern of

ciliary membranes is different from that of whole-tissue

membranes and suggest that olfactory receptors require a

specific lipid environment for their functioning.

Keywords Olfactory epithelium � Pig � Lipids �MALDI-TOF-MS � TLC � 31P-NMR

Abbreviations

AC3 Adenylcyclase III

cAMP Cyclic-adenosine-monophosphate

Cer Ceramides

CHOL Cholesterol

CM Ciliary membranes

DMSO Dimethylsulfoxide

DTT Dithiothreitol

Gb5 Globopentaosylceramides

IBMX Isobutylmethylxanthine

MALDI-TOF-MS Matrix-assisted laser desorption/

ionization time-of-flight mass

spectrometry

OR Olfactory receptor

OSN Olfactory sensory neuron

PtdOH Phosphatidic acid

PtdCho Phosphatidylcholine

PtdEtn Phosphatidylethanolamine

p-PtdEtn Plasmenyl-Phosphatidylethanolamine

PtdIns Phosphatidylinositol

PMSF Phenylmethanesulfonyl fluoride

PtdSer Phosphatidylserine

Ptd2Gro Cardiolipin

S-GalCer Sulfoglycosphingolipids

S. Lobasso � R. Angelini � M. Baronio � A. Corcelli (&)

Department of Medical Biochemistry Medical Biology

and Medical Physics, University Aldo Moro, Pl. G. Cesare,

70124 Bari, Italy

e-mail: a.corcelli@biologia.uniba.it

P. Lopalco

Institute for Microelectronics and Microsystems (IMM),

National Research Council (CNR), Lecce, Italy

F. P. Fanizzi

Department of Biological and Environmental Sciences

and Technologies, University of Salento, Lecce, Italy

F. Babudri

Department of Chemistry, University Aldo Moro, Bari, Italy

A. Corcelli

Institute for Chemical-Physical Processes,

National Research Council (IPCF-CNR), Bari, Italy

123

Lipids (2010) 45:593–602

DOI 10.1007/s11745-010-3432-1

CerPCho Sphingomyelin

WM Whole-tissue membranes

Introduction

Inhalation of odors across the surface of the olfactory

epithelium of the animal nose activates the olfactory sig-

naling cascade, which involves the binding of ligands to

receptors localized on primary sensory cells, the olfactory

sensory neurons (OSN). The OSN are bipolar neurons of

the pseudostratified olfactory epithelium, having a thin

sensory axon extending to higher brain regions and a single

dendrite that ends with a knob, from which long fine cilia

protrude, directly projected into the mucous of the olfac-

tory epithelium.

Olfactory cilia are the sites of the sensory transduction

apparatus. The binding of odorants to G-protein-coupled

seven-transmembrane olfactory receptors (OR) activates

the Gaolf subunit of a specific heterotrimeric G-protein

complex, which stimulates the enzyme adenylcyclase III

(AC3) to synthesize the second messenger molecule cyclic-

adenosine-monophosphate (cAMP). cAMP in turn acti-

vates the opening of cyclic-nucleotide-gated ion channels

present on the plasma membrane, generating electrical

signals in the primary sensory axons [1, 2].

Each OSN expresses only one type of OR out of a

repertoire of about 1,000 [3]. Numerous OSN expressing

the same OR are dispersed in the olfactory mucosa, while

their axons converge to form glomeruli in the olfactory

bulb, where a precise distribution or map of odors exists.

It is known that several proteins involved in the sensory

signaling cascade are compartmentalized in specialized

membrane subdomains, called lipid rafts, which are

expected to be spread in the olfactory ciliary membranes

[4, 5].

In the lipidomics era it is surprising to find out that only

a few analyses of the lipids of olfactory mucosa have been

reported in the literature. Old reports described the lipid

composition of plasma membranes isolated from bovine

[6] and rat [7] olfactory mucosa, while no data are avail-

able either on pig olfactory mucosa in toto, or on isolated

olfactory cilia.

It would be helpful to have an overview of the specific

set of lipids in the sensory cilia in order to investigate the

possible functions of lipids in signal transduction, adapta-

tion, xenobiotic metabolism and OSN maturation.

The present study provides the first general character-

ization of the membrane lipids of the neuroepithelial

olfactory mucosa covering pig turbinates and information

on lipids associated with the specialized membrane domain

of pig olfactory cilia.

Materials

The nasal cornets of pig were kindly provided immediately

after sacrifice by the slaughterhouse in Ruvo di Puglia (Bari,

Italy). The Cyclic AMP [3H] assay system was purchased

from Amersham (Freiburg, Germany). Odorants (Sigma–

Aldrich, St. Louis, MO) were usually prepared as stock

solutions in ethanol or DMSO. All organic solvents used

were commercially distilled and of the highest available

purity (Sigma–Aldrich). Plates for TLC (Silica gel 60A,

10 9 20 cm, 0.25 mm thick layer), obtained from Merck

(Darmstadt, Germany), were washed twice with chloroform/

methanol (1:1, by vol) and activated at 120 �C before use.

The following lipid standards were obtained from Avanti

Polar Lipids, Inc. (Alabaster, AL): 1,2-dimyristoyl-sn-gly-

cero-3-phosphate (sodium salt) (14:0 PtdOH), 1,2-dimyri-

stoyl-sn-glycero-3-phospho-L-serine (sodium salt) (14:0

PtdSer), 1,2-dipalmitoleoyl-sn-glycero-3-phosphoethanola-

mine (16:1 PtdEtn), 1,2-dioleoyl-sn-glycero-3-phospho-

(10-myo-inositol) (ammonium salt) (18:1 PtdIns), 10,30-bis

[1,2-dimyristoyl-sn-glycero-3-phospho]-sn-glycerol (sodium

salt) (14:0 Ptd2Gro), 1,2-dimyristoleoyl-sn-glycero-3-phos-

phocholine (14:1 (D9-Cis) PtdCho), 1,2-distearoyl-sn-gly-

cero-3-phosphocholine (18:0 PtdCho), 1,2-di-O-phytanyl-

sn-glycero-3-phosphocholine (4ME 16:0 diether PtdCho),

N-palmitoyl-D-erythro-sphingosylphosphorylcholine 16:0

CerPCho (d18:1/16:0). The matrix for MALDI-TOF–MS

(9-aminoacridine hemihydrate) was purchased from Acros

Organics (Morris Plains, MJ).

Methods

Isolation of Membranes Enriched in Olfactory Cilia

Ciliary membranes were detached and isolated from pig

olfactory epithelia as previously described [8, 9]. The pig

olfactory epithelia, situated on both left and right ethmo-

turbinates, were carefully stripped from the underlying

bone and washed in ice-cold Ringer solution (120 mM

NaCl, 5 mM KCl, 1.6 mM K2HPO4, 1.2 mM MgS04,

25 mM NaHCO3, 7.5 mM glucose, pH 7.4) plus 1 mM

PMSF within 10–20 min of slaughter. All operations were

carried at 0–4 �C. The olfactory cilia were detached using a

calcium shock procedure, raising the calcium concentration

in the Ringer solution up to 10 mM. After gently shaking

(20 min at 4 �C), the deciliated epithelia were removed by

centrifugation (7,000g for 5 min). The supernatant was

collected and the pellet incubated again with the solution

containing 10 mM calcium ion for 20 min. After removing

the deciliated epithelia by centrifugation, the supernatants

containing the detached cilia were combined. Membranes

enriched in olfactory cilia were collected by centrifuging at

594 Lipids (2010) 45:593–602

123

27,000g for 15 min, and resuspended in Buffer A (3 mM

MgCl2, 2 mM EDTA, 10 mM Tris/HCl, 1 mM PMSF, pH

7.4) with 5% glycerol. Ciliary membranes were appor-

tioned into small volumes and saved at -80 �C. The pro-

tein concentration was determined by the Bradford method

[10], with bovine serum albumin as standard.

Whole-Tissue Membrane Isolation

Pig olfactory epithelia, carefully dissected as described in

the previous paragraph, were suspended in 5–7 volumes of

hypotonic Ringer solution (Ringer without NaCl) and

homogenized on ice with a Potter homogenizer. Cartilage

fragments and large epithelia pieces were removed by low

speed centrifugation followed by filtration. The filtrate was

then centrifuged twice at 1,500g for 10 min and the pig-

mented pellet removed. The final supernatant was centri-

fuged at 27,000g for 20 min to generate a light yellow

pellet. This pellet was suspended in 5% glycerol-containing

Buffer A and stored at -80 �C.

Adenylcyclase Assay

Adenylcyclase activity was assayed according to a modi-

fied version of the method described previously [11, 12].

All assays were carried out at 37 �C. Briefly 70 ll of cilia

suspension (50–100 lg protein/ml) was mixed with 360 ll

of stimulating buffer containing 200 mM NaCl, 10 mM

EGTA, 50 mM MOPS, 2.5 mM MgCl2, 1 mM DTT,

0.05% sodium cholate, 1 mM ATP, 20 lM GTP, 1 mM

IBMX buffered at pH 7.4, with or without a stimulant

(200 lM odorants or 5 lM forskolin). The incubation of

cilia with the stimulating buffer (15 min at 37 �C) was

stopped by adding 350 ll of ice cold 10% perchloric acid.

Then samples were centrifuged at 2,500g for 5 min at 4 �C;

400 ll of the supernatant was added to 100 ll of 10 mM

EDTA pH 7.0. The samples were then neutralized by

adding 500 ll of a mixture containing 1,1,2-trichlorotri-

fluoroethane and tri-n-octylamine, mixed and then three

phases were obtained by centrifugation; the upper phase

contained the water soluble components and was used to

estimate the amount of cAMP produced during the incu-

bation of cilia with substrates. The cAMP was determined

by radioimmunoassay using an Amersham kit.

Lipid Extraction

Total lipids were extracted from membranes using the

Bligh and Dyer method [13] in the presence of 0.01% DTT

as antioxidant. 6 ml of methanol/chloroform (2:1, by vol)

was added to a 1.6 ml membrane suspension (about 6 mg

proteins). The mixture was gently shaken for 15 min and

then centrifuged to collect the supernatant. The residue

pellet was re-extracted by adding 7.6 ml of methanol/

chloroform/water (2:1:0.8, by vol). The mixture was again

shaken for 15 min and centrifuged. Then 4 ml each of

chloroform and 0.2 M KCl were added to the combined

supernatant extracts to obtain a two-phase system, chlo-

roform and methanol/water (1:0.9, by vol). After complete

phase separation, the lipid-containing chloroform phase

was brought to dryness under argon; dried lipids were

weighed, suspended in a small chloroform volume and

saved at -20 �C.

Thin-Layer Chromatography

Total lipid extracts were analyzed by TLC on silica gel

plates and lipids were eluted with Solvent A, chloroform/

methanol/acetic acid/water (85:15:10:3.5, by vol). For

2D-analysis, the plates were developed using Solvent B,

chloroform/methanol/ammonium hydroxide (65:25:5, by

vol), in the first dimension, and Solvent A in the second

dimension. The solvent used for separation of neutral lipids

on 1D-TLC was Solvent C, hexane/ethyl ether/acetic acid

(70:30:1, by vol). Individual phospholipids were identified

by reference to authentic lipid standards (Sigma–Aldrich).

Lipids were detected by spraying plates with 5% sulfuric

acid, followed by charring at 120 �C. Additional confir-

mation of the identity of lipids was obtained using (a)

molybdenum blue reagent (Sigma–Aldrich), specific for

phospholipids, (b) 0.5% a-naphthol in methanol/water (1:1,

by vol), specific for glycolipids, (c) ninhydrine 0.25% in

acetone/lutidine (9:1, by vol), for free amino groups [14].

To analyze in detail the various lipid components of the

extracts, bands present on 1D-TLC, developed in Solvent

A, were scraped and lipids extracted from silica, as pre-

viously described [14]; then lipid bands were analyzed by

mass spectrometry.

MALDI-TOF Mass Spectrometry

Lipid analysis was performed as previously described [15].

Briefly total lipid extracts (10 mg/ml; dissolved in chlo-

roform/methanol (1:1, by vol)) were diluted from 20 to

200 ll with isopropanol/acetonitrile (60:40, by vol). After

mixing 10 ll of diluted sample with 10 ll of 9-amino-

acridine (10 mg/ml; dissolved in isopropanol/acetonitrile

(60:40, by vol)), 0.3 ll of the mixture was spotted on the

instrument plate. The same procedure was followed to

analyze the lipid standards (1 mg/ml). MS analysis was

performed on a Bruker Microflex spectrometer (Bruker

Daltonics, Bremen, Germany). Mass spectra were acquired

in the positive and negative mode by averaging 600 con-

secutive laser shots (50 shots per subspectra). Synthetic

lipid standards (Avanti Polar Lipids) were used as external

standards for calibration.

Lipids (2010) 45:593–602 595

123

NMR Spectroscopy

31P-NMR analysis of phospholipids present in the total

lipid extract was performed by following the previously

described method [16, 17]. The method is based on the use

of a methanol reagent containing D2O and a dissolved

EDTA salt, prepared as follows. The cesium salt of EDTA

was prepared by titrating a 0.2 M suspension of EDTA free

acid with CsOH to a pH of 6.0, at which point free EDTA

was in solution; EDTA salt solutions were evaporated to

dryness on a freeze-dry apparatus, dissolved in a minimum

volume of D2O to exchange labile 1H for 2D, dried a

second time and dissolved in D2O to a concentration of

0.2 M. The final methanol reagent was prepared by dis-

solving 1 ml of D2O-EDTA solution in 4 ml of methanol.

The use of D2O is solely in order to provide a deuterium

reference signal for magnetic resonance field-frequency

stabilization; it is not essential for signal narrowing.

Between 1 and 5 mg of phospholipid standards and 12 mg

of total lipid extract were dissolved in 0.8 ml of deuterated

chloroform. To this solution 0.4 ml of methanol reagent

(containing Cs/EDTA) was added, and the mixture stirred

gently. Two liquid phases were obtained, a larger chloro-

form phase and a smaller water phase. By using a Pasteur

pipette, the sample was placed in an NMR test tube, where

it separated within 1 min. The sample tube turbine was

adjusted so that only the chloroform phase was detected by

the NMR spectrometer’s receiver coil. Magnetic field sta-

bilization was obtained through the deuterium resonance of

deuterated chloroform. Unless otherwise specified, samples

were analyzed with proton broad-band decoupling to

eliminate 1H-31P multiplets. Under these conditions each

spectral resonance corresponded to single phosphorus. 31P

chemical shifts were relative to 85% H3PO4 as an external

standard. Samples were analyzed using a Bruker DRX500

Avance instrument (Bruker Daltonics, Bremen, Germany).

Results

Whole-tissue membranes (WM) of neuroepithelial cells

were isolated from pig olfactory epithelium and total lipids

were extracted to be analyzed; in addition to the WM, a

membrane fraction enriched in olfactory cilia (CM) was

isolated by following the so-called calcium-shock method

[8, 9], briefly described in Fig. 1.

Ciliary membrane enrichment was estimated by assay of

odor-stimulated AC3 activity (i.e. an activity marker for

olfactory cilia); Table 1 reports the specific AC3 activity

both in membranes enriched in olfactory cilia and in

the whole-tissue membranes, in the absence and in the

presence of stimulants. Basal, forskolin-stimulated and

odor-stimulated AC3 activity was observed in both the

membrane preparations, but it is evident that forskolin- and

odor-stimulated activities were higher in the membrane

fraction enriched in cilia.

Total lipids were extracted from the two different

membrane preparations isolated from epithelial cells of pig

nose; the lipid/protein ratio was about 1.2 in both mem-

brane fractions. Figure 2 shows the two-dimensional TLC

analyses of the total lipid extracts of CM (a) and WM (b).

Lipid band identification was performed by comparison

with authentic standards (St, in Fig. 2c). The two lipid

profiles were similar; five main bands arose from the sep-

aration of polar lipid components during the chromato-

graphic run, while neutral lipid bands could be seen close

to the solvent front. Polar lipid bands corresponded to the

following lipid classes (in Rf order): sphingomyelin

(CerPCho) (as a doublet), phosphatidylinositol (PtdIns),

phosphatidylcholine (PtdCho), phosphatidylserine (Ptd-

Ser), and phosphatidylethanolamine (PtdEtn); cardiolipin

(Ptd2Gro) was a minor lipid component present in the lipid

extract of WM, while it was absent in the lipid extract of

CM. Another difference between the two TLC plates in

Fig. 2a, b was in the intensity of PtdEtn spot, which was

more intense in whole than ciliary membranes. Glyco-

sphingolipids (cerebrosides, gangliosides, sulfatides), rela-

tively abundant components of nervous tissue cells, could

Fig. 1 Calcium-shock scheme for isolating membranes enriched in

olfactory cilia. The pig olfactory epithelia were suspended in ice-cold

Ringer solution containing 10 mM CaCl2 (twice). To separate tissue

fragments, fractions were centrifuged at low speed (7,000g for

5 min). To collect membranes enriched in olfactory cilia, the

combined supernatants were centrifuged at high speed (27,000g for

15 min). The final membrane pellet was resuspended in Buffer A/5%

glycerol and stored at -80 �C

596 Lipids (2010) 45:593–602

123

be not detected by this technique either in whole or in

ciliary membranes. Finally it was difficult to compare the

cholesterol (CHOL) spots in WM and CM, because they

were too close to the solvent front overlapping with other

neutral lipid components, such as fatty acids and diacyl-

glycerols. All together the TLC data in Fig. 2 document

that our preparation of ciliary membranes was devoid of

mitochondrial membranes and is characterized by a lower

PtdEtn content than the whole membranes.

The difference in the PtdEtn content was also evident in

the 1D-TLC lipid profiles (Fig. 3). To further investigate

on the CHOL content of WM and CM, we analyzed their

lipid extracts by 1D-TLC using a solvent for neutral lipids

(as described in ‘‘Methods’’). The results revealed that

cholesterol percentage did not differ between the two lipid

extracts and amounted to about 10% (Fig. 4).

To gain detailed information on the lipid classes present

in pig olfactory mucosal cells, the total lipid extract of

whole-tissue membranes was analyzed by MALDI-TOF-

MS, using the novel matrix 9-aminoacridine allowing a fast

reliable analysis of both zwitterionic and anionic lipid

species [15]. Main lipid bands isolated from preparative

1D-TLC were also analyzed by MALDI-MS. Furthermore

ESI-MS analysis of the lipid extract was performed to

support the identification of lipid classes by MALDI-TOF-

MS (not shown).

Table 1 Adenylcyclase activity in isolated membranes

Basal activity ± SEM Forskolin-stimulated activity ± SEM Odor-stimulated activity ± SEM

Membranes enriched in cilia 197.7 ± 25.3 850.5 ± 39.4 480.1 ± 26.6

Whole-tissue membranes 60.9 ± 18.6 280.3 ± 13.2 150.5 ± 10.2

Cyclase assay was carried out with 5 lM forskolin or with an odor mixture containing 100 lM each of eugenol and citralva, as described in

‘‘Methods’’. The specific activity is reported as pmol cAMP/mg/min. The values are averages of three separate experiments. SEM refers to

standard error of the mean

Fig. 2 2D-TLC of the lipid

extracts of membranes isolated

from pig olfactory epithelium.

Total lipids were extracted from

the membrane preparations

isolated from epithelial cells of

pig nose and from membranes

enriched in olfactory cilia, as

described in ‘‘Methods’’. Eighty

micrograms of the lipid extract

of whole-tissue (WM) and

enriched in olfactory cilia (CM)

membranes were loaded in the

TLC plates (a) and (b),

respectively. The following pair

of solvents was used: Solvent B

chloroform/methanol/

ammonium hydroxide (65:25:5,

by vol), and then Solvent A

chloroform/methanol/acetic

acid/water (85:15:10:3.5, by

vol). Total lipids were detected

by spraying with 5% sulfuric

acid and charring at 120 �C. On

the TLC plate in (c) eight

micrograms of each lipid

standard were loaded. S-GalCer,

galactocerebroside sulfate;

Ptd2Gro, cardiolipin; CHOL,

cholesterol; PtdCho,

phosphatidylcholine; PtdEtn,

phosphatidylethanolamine;

PtdIns, phosphatidylinositol;

PtdSer, phosphatidylserine;

CerPCho, sphingomyelin

Lipids (2010) 45:593–602 597

123

The mass spectrum acquired in the negative mode

(Fig. 5a) showed two main peaks at m/z 885.8 and 788.8,

corresponding to the molecular ions [M - H]-of PtdIns

38:4 and PtdSer 36:2, respectively. The minor peaks at m/z

750.6, 766.7 and 810.7 corresponded to the molecular ions

[M - H]- of plasmenyl-PtdEtn, PtdEtn and PtdSer (all

38:4 species), respectively. Furthermore the peaks at m/z

838.7 and 857.3 were attributed to a PtdSer 40:4 and PtdIns

36:4. Two small peaks were present in the high m/z range;

the peak at m/z 1448.3 was attributed to the molecular ions

[M - H]- of a cardiolipin, having four linoleic acid (18:2)

chains, while the peak at m/z 1475.8 was not assigned. Of

course the apparent amounts of the various lipids in the

MALDI profiles depend on the lipid individual tendency to

ionization together with its abundance in the extracts.

MALDI-TOF-MS analysis also allowed the direct

identification of minor lipid components of olfactory neu-

roepithelial cells, such as sulfoglycosphingolipids (i.e.

sulfatides). A detailed description of the advantages of

MALDI-TOF in the analysis of sulfatides has been recently

reported [18].

Fig. 3 1D-TLC lipid profiles of whole-tissue (WM) and enriched in

olfactory cilia (CM) membranes. Total lipids (60 micrograms) were

detected by charring. Solvent A was used

Fig. 4 Cholesterol content of whole-tissue (WM) and enriched in

olfactory cilia (CM) membranes. Ten micrograms of each lipid

extract was loaded on the TLC plate. The following solvent was used:

Solvent C, hexane/ethyl ether/acetic acid (70:30:1, by vol). Only the

CHOL bands are shown (note that with Solvent C polar lipids do not

separate and are all together at the sample deposition line). The

quantitative analysis of CHOL content was performed by video-

densitometry (ImageJ software), using standard CHOL loaded on the

same plate to construct the calibration curve

Fig. 5 MALDI-TOF mass spectrum profiles of the lipid extract of pig

epithelial membranes in the negative (a) and positive (b) mode. The

insets enlarge the two regions m/z 900–1,000 and 1,400–1,500 and

show the comparison of the lipid profile of membranes isolated from

the whole epithelium (lower line) with that of membranes enriched in

olfactory cilia (upper line). PtdOh, phosphatidic acid; p-PtdEtn,

plasmenyl-phosphatidylethanolamine; S-GalCer, galactosylceramide

sulfated; Gb5, globopentaosylceramides

598 Lipids (2010) 45:593–602

123

The peak at m/z 906.6 in the inset of the mass spectrum

of Fig. 5a can be attributed to a sulfoglycosphingolipid,

precisely to the sulfated galactosylceramide (S-GalCer),

consisting of a C24:0 hydroxy-fatty acid plus the sphingoid

4-sphingenine (d18:1).

The mass spectrum acquired in the positive mode

(Fig. 5b) was dominated by PtdCho species: the peaks at

m/z 734.3, 760.4, 786.6 and 810.5 corresponded to the

molecular ions [M ? H]? of PtdCho 32:0, 34:1, 36:2 and

38:4, respectively; the minor peak at m/z 703.4 can be

attributed to the molecular ion [M ? H]? of the sphingo-

lipid CerPCho 16:0.

Furthermore the peaks at higher m/z suggested the pres-

ence of other complex glycosphingolipids, such as

gangliosides. The peaks between m/z 1,400 and 1,600

were attributed to globopentaosylceramides (Gb5), which

give strong [M ? Na]? ions, corresponding to glyco-

sphingolipids all containing a C18-sphingosine base and

various fatty acids. The oligosaccharide chain of these

glycosphingolipids consists of two N-acetylgalactosamine,

two galactose and one glucose residue. In particular the

major peaks at m/z 1,564.9 and 1,542.2 appear to correspond

to the [M ? Na]? and [M ? H]? ions of species containing

saturated fatty acids with 24 carbon atoms respectively,

while that at m/z 1,514.9 to the [M ? H]? ion of a specie

having C22:0. The minor peaks at m/z 1,487.5 and 1,592.7

can be attributed to the [M ? H]? ion of Gb5 with C20:0

and to the [M ? Na]? ion of Gb5 with C26:0, respectively.

Galactocerebrosides, giving signals in the range m/z

800–850, were not well distinct from the background.

The lipid extract obtained from the membrane fraction

enriched in cilia was also analyzed by MALDI-TOF-MS;

the mass spectra obtained (negative and positive mode)

were similar to those of whole-tissue membranes. Some

differences were found in the spectrum acquired in the

negative mode, in the intervals of m/z values 900–1,000

and 1,200–1,600. These regions of the MALDI-TOF mass

spectrum are reported in the insets of Fig. 5a, in order to

allow direct comparison of MALDI-TOF-MS lipid profiles

of the two different lipid extracts. It can be seen that in the

mass spectrum of CM lipid extract: (a) cardiolipin was

absent, in agreement with previous TLC data (see Fig. 2a);

(b) additional peaks were present between m/z 930–960

suggesting that the ciliary membranes contained other

sulfoglycosphingolipids, besides the S-GalCer at m/z 906

(see inset in Fig. 5a).

This last finding arises from the particularly high sensi-

tivity of MALDI-MS in the detection of sulfoglyco-

sphingolipids in the tissue lipid extracts; the cluster centered

at m/z 934 indicates, on qualitative basis, the presence of an

enrichment of long-chain sulfoglycosphingolipids (S-Gal-

Cer 26:0) in the specialized membrane of cilia.

Tables 2 and 3 report the main phospholipid and gly-

colipid classes, resulting from the MALDI-TOF-MS anal-

ysis of the total lipid extracts and individual TLC bands.

Like lipids of nervous tissue, polar lipids of the olfactory

mucosa contain mainly arachidonic (20:4) and stearic

(18:0) fatty acids.

Finally, in order to check for quantitative differences in

lipid composition between the whole-tissue and ciliary

membranes, the total lipid extracts of both the membrane

preparations were analyzed by 31P-NMR spectroscopy. In

order to assign NMR peaks to the lipid components in the

extracts, authentic standard phospholipids were analyzed

under the same experimental conditions and their chemical

shifts are reported in Table 4.

Table 2 MALDI-TOF signals of phospholipids of pig olfactory epithelium

Class of phospholipid Ion Total fatty acid carbon n:n of double bonds

(32:0) (32:1) (34:1) (34:2) (36:0) (36:2) (36:4) (38:4) (40:4) (16:0) (18:0) (22:0) (24:0) (36:4)

(36:4)

Ptd2Gro [M - H]- – – – – – – – – – – – – – 1,448

PtdOH [M - H]- – – – – 701 – – 723 – – – – – –

PtdEtn [M - H]- – – – – – 742 – 766 – – – – – –

p-PtdEtn [M - H]- – – – – – – – 750 – – – – – –

PtdCho [M ? H]? 734 – 760 – – 786 – 810 – – – – – –

PtdSer [M ? H]? – 734 – – – – – – – – – – – –

PtdSer [M - H]- – – – 760 788 – – 810 838 – – – – –

PtdIns [M - H]- – – – – – – 857 885 – – – – – –

CerPCho [M - H]? – – – – – – – – – 703 731 787 815 –

Data reported are obtained from the analysis of total lipid extracts and bands isolated from preparative TLC. The m/z values are reported. The

numbers (x:y) denote the total length and number of double bonds of both acyl chains, respectively, except for PtdEtn plasmalogen species

(denoted with p-PtdEtn), in which the acyl chain at the sn-1 position is replaced with an alkenyl. For CerPCho species, the numbers in brackets

correspond to the length and number of double bonds of the acyl chain, attached to the sphingosine base. For Ptd2Gro species, the numbers refer

to the two pairs of acyl chains

Lipids (2010) 45:593–602 599

123

By comparing the two NMR spectra illustrated in Fig. 6

it can be seen that PtdCho, PtdIns, PtdSer, PtdEtn and

CerPCho species were present in both the lipid extracts of

WM (a) and CM (b); PtdSer and CerPCho peaks (at about

0.9 ppm) were very close to each other and not well

resolved in the case of spectrum in panel a, because WM

possibly contain more PtdSer and CerPCho species

than CM.

The sums of the peak areas in the two spectra were

13.87 and 15.39 in CM and WM, respectively; this indi-

cates that the total lipid phosphorus per mg of total lipid

extract in CM was slightly lower than in WM and repre-

sents an indirect indication of the presence of an higher

proportion of glycosphingolipids in cilia, compared to

whole membranes. The last main peak at 1.05 ppm was

assigned to a plasmenyl-PtdEtn lipid, in agreement with a

previous literature report [19]; also the small, but clearly

visible peak at 0.18 ppm can be attributed to plasmalogen

PtdCho species, as well [19]. The Ptd2Gro peak in the

NMR profile of whole membranes was absent because its

relative proportion in the lipid extract was below the sen-

sitivity limit of NMR analysis.

The differences in the area peaks of the two NMR

spectra indicate that the proportions of various lipid classes

in the two lipid extracts were different. The area ratios

PtdEtn species/(PtdSer ? CerPCho) in CM and WM were

1.23 and 1.62 respectively, indicating that the proportion of

PtdEtn species (as the sum of diacyl- and acyl-alkyl forms)

was lower in CM extract in agreement with results from

TLC analyses (Figs. 2, 3).

Table 3 MALDI-TOF signals of glycolipids of pig olfactory epithelium

Class of

glycolipid

Ion m/z No. of

sulfated

residue

Chain length (carbon n:n of double bonds) Oligosaccharides (no. of residues)

Sphingoid Acyl Glc Gal GalNAc

S-GalCer 24:0 [M - H]- 906.6 1 18:1 24:0h – 1 –

S-GalCer 26:0 [M - H]- 934.5 1 18:1 26:0h – 1 –

Gb5 26:0 [M ? Na]? 1,592.7 – 18:0 26:0 1 2 2

Gb5 24:0 [M ? Na]? 1,564.9 – 18:0 24:0 1 2 2

Gb5 24:0 [M ? H]? 1,542.2 – 18:0 24:0 1 2 2

Gb5 22:0 [M ? H]? 1,514.9 – 18:0 22:0 1 2 2

Gb5 20:0 [M ? H]? 1,487.5 – 18:0 20:0 1 2 2

S-GalCer, galactosylceramide sulfate; Gb5, globopentaosylceramide; Sphingoid, amino-alcohol chain; Acyl, fatty acid; h, hydroxylated fatty

acid; Glc, glucose; Gal, galactose; GalNAc N-acetylglucosamine

Table 4 31P-chemical shifts of standard phospholipids

Lipid ppm

PtdCho 0.115

PtdIns 0.621

CerPCho 0.911

PtdSer 0.894

PtdEtn 1.066

Ptd2Gro 1.148

Chemical shifts are referenced to 85% H3PO4 as an external standard.

The samples were prepared in the Cs/EDTA analytical reagent, as

described in ‘‘Methods’’

Fig. 6 31P-NMR spectra of the lipid extract of whole-tissue (a) and

enriched in olfactory cilia (b) membranes. The areas of peaks are

reported below the x-axis

600 Lipids (2010) 45:593–602

123

Discussion

Lipid research can offer important elements to complete

our understanding of the genesis of sensory perception

pathologies. For example anosmia, i.e. the inability to

perceive odors, can arise from loss of olfactory cilia or

impairment in the olfactory signaling cascade [20]. The

complete comprehension of the mechanisms of ciliogenesis

and assembly of membrane signaling molecules requires an

integration of lipid and protein studies.

The present paper describes the results of a study on

cellular lipids of porcine olfactory mucosa, previously

selected to investigate the response of olfactory cilia to

explosives, in the frame of an investigation to understand

the molecular basis of the ability of an animal nose to

detect buried landmines [21]. In the pig nose the olfactory

mucosa is well distinct from the respiratory mucosa, while

in other mammals technical difficulties are often involved

in obtaining olfactory tissue from ethmoids without con-

comitant excision of non-olfactory tissue [7].

As the olfactory mucosa consists of three cell types

(olfactory sensory neurons, supporting and basal cells),

results of whole membrane lipid analyses refer to the

average lipid composition of membranes arising from the

different cell types in the epithelium.

A combination of a number of different analytical

techniques was used to analyze in detail the lipids of

olfactory neuroepithelium together with the novel MALDI-

TOF-MS approach based on the use of the versatile matrix

9-aminoacridine [15].

The results shown here indicate that the lipid composition

of olfactory mucosa in pig is similar to that of bovine and rat

[6, 7]. Phospholipids account for about 85–90% (by weight)

of the total lipid of mucosa, with zwitterionic lipids (PtdEtn,

PtdCho and CerPCho) being more abundant (about 70%)

than anionic species (PtdSer, PtdIns, Ptd2Gro, PtdOH). Most

phospholipids show a polyunsaturated fatty acid content

with the arachidonic acid (20:4) residue predominating. The

fatty acids in sphingomyelin, however, are totally saturated

and include C16:0, C18:0, C22:0, and C24:0 chains.

Although separation of acyl-acyl from alkyl-acyl forms of

phospholipids cannot be easily achieved by TLC, NMR anal-

yses reveal that peak areas of alkyl-acyl PtdEtn almost equal

the diacyl-PtdEtn species, as previously reported for mam-

malian brain tissue [22]. MALDI-TOF-MS analyses indicate

that the main PtdEtn species contain C18 and C20 chains.

As regards glycosphingolipids, sulfoglycosphingolipids

having C24:0 hydroxy-fatty acid and the sphingoid

4-sphingenine (d18:1), and pentaosylceramides containing

a C18-sphingosine base and various fatty acids, were only

detected by MALDI-TOF analysis.

Here information on the lipid composition of a mem-

brane fraction enriched in olfactory cilia is also reported. In

the past biochemical exploration of the olfactory cilia was

mostly based on a membrane preparation protocol estab-

lished by Chen and Lancet [9]. Thus, the examination of

the cilia preparations yields many of the molecular details,

which support the current concept for olfactory signal

transduction. In the present study we isolated a membrane

fraction enriched in ciliary membranes of about the same

quality as the analogous preparation obtained by other

authors from rat tissue [23]. The lack of cardiolipin in the

lipid extract of cilia represents a good internal index of the

absence of mitochondrial and possibly other intracellular

membranes in our preparation. However, we cannot

exclude the presence of microvillar fragments of support

cells in our ciliary membrane preparation.

In summary, the present analysis of the lipid extracts of

whole-tissue membranes and enriched ciliary membranes

shows that: (1) PtdEtn species are less abundant in ciliary

membranes than in total epithelial membranes (TLC and

NMR findings), and (2) long-chain sulfoglycosyl-

sphingolipids (S-GalCer 26:0) are enriched in ciliary

membranes compared to the crude olfactory membranes.

The roles of sulfoglycosylsphingolipids in metabolism and

functions of nervous tissue have been recently reviewed

[24]. In principle, given the importance of lipids in signal-

ing, any difference in the lipid bilayer composition might be

considered a factor that can affect the membrane trans-

duction properties. A decrease of membrane PtdEtn content

was previously described in brain pathologies [22]. In

olfactory cilia a decrease of PtdEtn level implies an increase

in other membrane lipids, such as CerPCho, PtdSer and

PtdIns, having a well known role in the assembly of lipid

rafts (CerPCho, PtdSer) and signal transduction (PtdIns).

On the other hand, PtdEtn is preferentially located in the

inner membrane leaflet, where it is associated with acidic

phospholipids such as PtdIns [25]. Being cone-shaped

phospholipid, PtdEtn can influence the determination of

membrane curvature; it has been reported that intracellular

tubular membranes of Golgi apparatus have a minor

amount of PtdEtn compared to plasma membrane [25]. The

possibility of a correlation between the low PtdEtn content

and the high curvature in cross section of tubules of

olfactory cilia remains to be investigated.

Although the presence of cholesterol-rich rafts in the

olfactory cilia has been indirectly suggested by some bio-

chemical studies [4, 5], here no cholesterol enrichment was

found in the lipid extract of the ciliary membranes. How-

ever it should be considered that the analysis of the CM

lipid extract gives information on the average lipid com-

position of the membrane lining the long (up to 100 lm)

tubular structure of cilia, which presumably consist of

different biochemical and functional domains.

In particular as it has been shown that cholesterol is

absent at the necklace (or base) of the ciliary structure [26,

Lipids (2010) 45:593–602 601

123

27], cholesterol in the CM lipid extract might represent the

average of cholesterol in the proximal and distal tubular

membranes. In conclusion, we cannot exclude the presence

of cholesterol-rich domains (i.e., rafts) in ciliary mem-

branes on the basis of the lack of cholesterol enrichment in

the lipid extract of ciliary membranes.

Another element of complexity in the interpretation of

lipid data arises from the recent finding that the chemo-

sensory apparatus of cilia is also present at the level of the

emergency cone of the sensory olfactory axon [28].

Here besides reporting the first lipidomic data on

olfactory mucosa in toto, we show for the first time the

presence of different lipid domains in the membranes of

neuroepithelial olfactory cells. Our study complements

recent proteomic studies, in which many ciliary proteins,

that mediate chemo-electrical transduction, amplification

and adaptation of the primary sensory signal, have been

identified [29–32].

Acknowledgments We thank F. Naso of the University of Bari for

the use of the MALDI-TOF instrument. This work was supported by

the Italian Ministry of Defence (Contract n. 685/18.12.2003), by

Regione Puglia (Grant code 15, Sens&MicroLab) and by Fondazione

Cassa di Risparmio di Puglia, Bari, Italy.

References

1. Firestein S (2001) How the olfactory system makes sense of

scents. Nature 413:211–217

2. Munger SD, Leinders-Zufall T, Zufall Z (2009) Subsystem

organization of the mammalian sense of smell. Annu Rev Physiol

71:115–140

3. Buck L, Axel R (1991) A novel multigene family may encode

odorant receptors: a molecular basis for odor recognition. Cell

65:175–187

4. Schreiber S, Fleischer J, Breer H, Boekhoff I (2000) A possible

role for caveolin as a signaling organizer in olfactory sensory

membranes. J Biol Chem 275:24115–24123

5. Brady JD, Rich TC, Le X, Stafford K, Fowler CJ, Lynch L,

Karpen JW, Brown LR, Martens JR (2004) Functional role of

lipid raft microdomains in cyclic nucleotide-gated channel acti-

vation. Mol Pharmacol 65:503–511

6. Koyama N, Sawada K, Kurihra K (1971) Isolation and some

properties of plasma membranes from bovine olfactory epithe-

lium. Biochim Biophys Acta 241:42–48

7. Russel Y, Evans P, Dodd GH (1989) Characterization of the total

lipid and fatty acid composition of rat olfactory mucosa. J Lipid

Res 30:877–883

8. Anholt RR, Aebi U, Snyder H (1986) A partially purified prepa-

ration of isolated chemosensory cilia from the olfactory epithe-

lium of the Bullfrog, Rana catesbeiana. J Neurosci 6:1962–1969

9. Chen Z, Pace U, Heldman J, Shapira A, Lancet D (1986) Isolated

frog olfactory cilia: a preparation of dendritic membranes from

chemosensory neurons. J Neurosci 6:2146–2154

10. Bradford MM (1976) A rapid and sensitive method for the

quantitation of microgram quantities of protein utilizing the

principle of protein-dye binding. Anal Biochem 72:248–254

11. Schandar M, Laugwitz K, Boekhoff I, Kroner C, Gudermann T,

Schultz G, Breer H (1998) Odorants selectively activate distinct G

protein subtypes in olfactory cilia. J Biol Chem 273:16669–16677

12. Boekhoff I, Tareilus E, Strotmann J, Breer H (1990) Rapid

activation of alternative second messenger pathways in olfactory

cilia from rats by different odorants. EMBO J 9:2453–2458

13. Bligh EG, Dyer WJ (1959) A rapid method of total lipid

extraction and purification. Can J Biochem Physiol 37:911–917

14. Kates M (1986) Techniques of lipidology laboratory techniques in

biochemistry and molecular biology, vol 3. Elsevier, Amsterdam

15. Sun G, Yang K, Zhao Z, Guan S, Han X, Gross RW (2008)

Matrix-assisted laser desorption/ionization time-of-flight mass

spectrometric analysis of cellular glycerophospholipids enabled

by multiplexed solvent dependent analyte-matrix interactions.

Anal Chem 80:7576–7585

16. Meneses P, Glonek T (1988) High resolution 31P NMR of

extracted phospholipids. J Lipid Res 29:679–689

17. Corcelli A, Lattanzio VMT, Mascolo G, Papadia P, Fanizzi FP

(2002) Lipid-protein stoichiometries in a crystalline biological

membrane: NMR quantitative analysis of the lipid extract of the

purple membrane. J Lipid Res 43:132–140

18. Cheng H, Sun G, Yang K, Gross RW, Han X (2010) Selective

desorption/ionization of sulphatides by MALDI-MS facilitated using

9-aminoacridine as matrix. J Lipid Res. doi:101194/jlrD004077

19. Pearce JM, Komoroski RA, Mrak RE (2009) Phospholipid

composition of postmortem schizophrenic brain by 31P NMR

spectroscopy. Magn Reson Med 61:28–34

20. Jenkins PM, McEwen DP, Martens JR (2009) Olfactory cilia:

linking sensory cilia function and human disease. Chem Senses

34:451–464

21. Corcelli A, Lobasso S, Lopalco P, Dibattista M, Araneda R,

Peterlin Z, Firestein S (2010) Detection of explosives by olfac-

tory sensory neurons. J Hazard Mater 175:1096–1100

22. Kakela R, Somerharju P, Tyynela J (2003) Analysis of

phospholipid molecular species in brains from patients with

infantile and juvenile neuronal-ceroid lipofuscinosis using liquid

chromatography-electrospray ionization mass spectrometry.

J Neurochem 84:1051–1065

23. Washburn KB, Turner TJ, Talamo BR (2002) Comparison of a

mechanical agitation and calcium shock methods for preparation

of a membrane fraction enriched in olfactory cilia. Chem Senses

27:635–642

24. Eckhardt M (2008) The role and metabolism of sulfatides in the

nervous system. Mol Neurobiol 37:93–103

25. Van Meer G, Voelker DR, Feigenson GW (2008) Membrane

lipids: where they are and how they behave. Nat Rev Mol Cell

Biol 9:112–124

26. Cuevas P, Gutierrez Diaz JA (1985) Absence of filipin-sterol

complexes from the ciliary necklace of ependymal cells. Anat

Embryol (Berl) 172:97–99

27. Satir P, Christensen ST (2007) Overview of structure and func-

tion of mammalian cilia. Annu Rev Physiol 69:377–400

28. Maritan M, Monaco G, Zamparo I, Zaccolo M, Pozzan T,

Lodovichi C (2009) Odorant receptors at the growth cone are

coupled to localized cAMP and Ca2? increases. Proc Natl Acad

Sci USA 106:3537–3542

29. Klimmeck D, Mayer U, Ungerer N, Warnken U, Schnolzer M,

Frings S, Mohrlen F (2008) Calcium-signaling networks in

olfactory receptor neurons. Neuroscience 151:901–912

30. Mayer U, Ungerer N, Klimmeck D, Warnken U, Schnolzer M, Frings

S, Mohrlen F (2008) Proteomic analysis of a membrane preparation

from rat olfactory sensory cilia. Chem Senses 33:145–162

31. Mayer U, Kuller A, Daiber FC, Neudorf I, Warnken U, Schnolzer

M, Frings S, Mohrlen F (2009) The proteome of rat olfactory

sensory cilia. Proteomics 9:322–334

32. Saavedra MV, Smalla KH, Thomas U, Sandoval S, Olavarria K,

Castillo K, Delgado MG, Delgado R, Gundelfinger ED, Bacigalupo

J, Wyneken U (2008) Scaffolding proteins in highly purified rat

olfactory cilia membranes. Chem Senses 19:1123–1126

602 Lipids (2010) 45:593–602

123