http://www.elsevier.com/locate/bba

Biochimica et Biophysica Ac

Regular paper

Profiling candidate genes involved in wax biosynthesis in

Arabidopsis thaliana by microarray analysis

Patricia Costagliolia,b,*, Jerome Joubesa, Christel Garciaa, Marianne Stef c,

Benoıt Arveilerc, Rene Lessirea, Bertrand Garbaya,b

aLaboratoire de Biogenese Membranaire, CNRS, UMR 5200, Universite Victor Segalen Bordeaux 2,

146 rue leo Saignat, Case 92, 33076 Bordeaux Cedex, FrancebEcole Superieure de Technologie des Biomolecules de Bordeaux, Universite Victor Segalen Bordeaux 2,

146 rue leo Saignat, Case 87, 33076 Bordeaux Cedex, FrancecEA 3669, Genetique Humaine, Developpement et Cancer, Universite Victor Segalen Bordeaux 2, Bordeaux, France

Received 25 November 2004; received in revised form 17 March 2005; accepted 15 April 2005

Available online 10 May 2005

Abstract

Plant epidermal wax forms a hydrophobic layer covering aerial plant organs which constitutes a barrier against uncontrolled water loss

and biotic stresses. Wax biosynthesis requires the coordinated activity of a large number of enzymes for the formation of saturated very-long-

chain fatty acids and their further transformation in several aliphatic compounds. We found in the available database 282 candidate genes that

may play a role in wax synthesis, regulation and transport. To identify the most interesting candidates, we measured the level of expression of

204 genes in the aerial parts of 15-day-old Arabidopsis seedlings by performing microarray experiments. We showed that only 25% of the

putative candidates were expressed to significant levels in our samples, thus significantly reducing the number of genes which will be worth

studying using reverse genetics to demonstrate their involvement in wax accumulation. We identified a beta-keto acyl-CoA synthase gene,

At5g43760, which is co-regulated with the wax gene CER6 in a number of conditions and organs. By contrast, we showed that neither the

fatty acyl-CoA reductase genes nor the wax synthase genes were expressed in 15-day-old leaves and stems, raising questions about the

identity of the enzymes involved in the acyl-reduction pathway that accounts for 20% of the total wax amount.

D 2005 Elsevier B.V. All rights reserved.

Keywords: Microarray; Wax synthesis; Arabidopsis thaliana

1. Introduction

Water is a crucial prerequisite for plant life. During the

evolution process, physiological, anatomical and morpho-

logical adaptation has taken place in plants in order to

maintain a water status suitable for survival even under

adverse environmental conditions. Indeed, specialized

1388-1981/$ - see front matter D 2005 Elsevier B.V. All rights reserved.

doi:10.1016/j.bbalip.2005.04.002

* Corresponding author. Laboratoire de Biogenese Membranaire, CNRS,

UMR 5200, Universite Victor Segalen Bordeaux 2, 146 rue leo Saignat,

33076 Bordeaux Cedex, France. Tel.: +33 5 57 57 16 78; fax: +33 5 57 57

17 11.

E-mail address: Patricia.costaglioli@estbb.u-bordeaux2.fr

(P. Costaglioli).

structures like well-developed cuticles and stomata are

present in fossil specimens of the very earliest terrestrial

plants known [1,2]. Both structures are required to optimize

photosynthetic gas exchange under the constraint of loss of

water to a dry atmosphere [3,4].

Plants living in an aerial environment should develop a

barrier against uncontrolled water loss. Such a barrier

should be efficient, translucent for photosynthetically active

radiation, flexible and self-healing. The plant cuticle, which

controls the movement of water between the outer cell wall

of the epidermis and the atmosphere adjacent to the plant,

combines all these properties: it is a rather thin (0.02–10 Amthick) membrane consisting of a polymer matrix (cutin),

polysaccharides and associated solvent-soluble lipids (cutic-

ta 1734 (2005) 247 – 258

P. Costaglioli et al. / Biochimica et Biophysica Acta 1734 (2005) 247–258248

ular waxes) [5]. Cutin is a three-dimensional polymer of

mostly C16 and C18 hydroxy fatty acids cross-linked by

ester bonds [6–10].

Cuticular wax is a general term for complex mixtures of

very long chain aliphatics lipids, but it also includes

triterpenoids and minor secondary metabolites, such as

sterols and flavonoids. The physical and chemical properties

of cuticular wax determine vital functions for plant. Indeed,

besides playing a major role in protecting the aerial parts of

the plants from uncontrolled water loss [3,4], it protects

plants against ultraviolet radiation and helps minimize

deposits of dust, pollen and air pollutants [11]. In addition,

surface wax is believed to play important roles in plant

defence against bacterial and fungal pathogens [12] and has

been shown to participate in a variety of plant–insect

interactions.

A large number of enzymes is required to carry out the

formation of saturated very-long-chain fatty acids

(VLCFAs) with predominant chain lengths from 20 to 32

carbons [13], and their further transformation in several

aliphatic compounds that constitute the wax layer (Fig. 1).

Two principal wax biosynthetic pathways coexist in the

epidermal cells of plants: an acyl reduction pathway, which

produces primary alcohols and wax esters, and a decarbon-

ylation pathway, leading to the formation of aldehydes,

alkanes, secondary alcohols and ketones [14].

Several wax-deficient mutants have been isolated in

different plant species [11,14–16]. The mutant loci in

Arabidopsis thaliana are termed eceriferum (cer), and 22

Fig. 1. Schematic representation of the wax biosynthetic pathways in Arabido

Arabidopsis stems and leaves. The VLCFAs are synthesized by Elongases, w

components through the acyl reduction pathway (20% of the metabolic flux) an

molecules cross the plasma membrane, most probably through ABC transporters, a

involves LTPs to some degree. In the case of the fatty acyl-CoA reductase (FAR*)

before its further transformation in primary alcohol. This FAR* is therefore distin

independent cer loci have been identified in this plant

model. The corresponding mutants display an abnormal wax

composition and/or a global decrease in wax constituents

[17–20].

In the past few years, five CER genes have been cloned

(CER1, 2, 3, 5 and 6), but the biological function of the

corresponding protein is only known with certainty for

CER5 and CER6, which encode, respectively, an ABC

transporter and a condensing enzyme of the elongase

complex [11,21].

Further to the strategy of using Arabidopsis cer mutants,

some other genes encoding proteins involved in wax

biosynthesis have been cloned and characterized. A

mutation of the GLOSSY 8 (GL8) locus in maize results

in decreased levels of wax components longer than C24. It

has been later demonstrated that the corresponding gene

encodes a reductase involved in VLCFAs synthesis [22,23].

The wax synthase (fatty acyl-coenzyme A: fatty alcohol

acyltransferase), which catalyzes the final step in the

synthesis of linear esters, has been characterized and

partially purified from developing jojoba embryo. This led

to the cloning of the corresponding gene [24]. Very recently,

it has been demonstrated that WIN1/SHINE1, an Arabi-

dopsis thaliana ethylene response factor-type transcription

factor, can up-regulate wax production in leaves and stems

when overexpressed [25,26].

Our current knowledge is thus confined to a very limited

number of genes from a series of different plants, and

several strategies have been developed to discover other

psis. Possible metabolic pathway for wax biosynthesis and transport in

hich are multi sub-units complex. Then, they are converted to different

d the decarbonylation pathway (80% of the flux). The different classes of

nd then travel up to the plant surface. It has been proposed that this transport

of the acyl reduction pathway, the aldehyde is not released from the enzyme

ct from the one involved in the decarbonylation pathway.

P. Costaglioli et al. / Biochimica et Biophysica Acta 1734 (2005) 247–258 249

actors of the wax biosynthesis pathway. One of these

strategies relies on identification in the available database of

candidate genes that may play a role in wax synthesis,

regulation and transport, owing to their predicted function,

their homology to known wax genes or their pattern of

expression. In their recent review, Kunst and Samuels [11]

established a list of 35 genes that encode wax biosynthetic

enzymes. We used their list as a starting point and we

increased it to a total of 147 candidate genes after

examination of publicly available web databases (TIGR,

TAIR, Arabidopsis Membrane Protein Library, The Arabi-

dopsis Lipid Gene Database). Moreover, after the recent

discovery that an ABC transporter plays a role in wax

deposition [21], we added to our list the 135 ABC protein

encoded by the Arabidopsis genome, bringing the total up

to 282 genes.

From the 282 candidate genes identified, we report in this

study the expression of 204 genes in the aerial part of 15-

day-old Arabidopsis plants: 18 beta-keto acyl-CoA syn-

thases (KCS), 2 beta-ketoacyl-CoA reductases (KCR), 5

trans-2-enoyl-CoA reductases (ECR), 6 fatty acyl-CoA

reductases (FAR), 10 wax synthases (WS), 3 bifunctional

WS/DGAT (Acyl-coA: diacylglycerol acyltransferase), 37

lipid transfer proteins (LTP), 113 ABC transporters and 10

CER-related genes.

2. Materials and methods

2.1. Plant material

Arabidopsis thaliana ecotype Landsberg erecta (Ler-0,

cer1-1, cer3-1 and cer6-2) seeds were obtained from the

Nottingham Arabidopsis Stock Centre (stock number

NW20, N31, N33 and N6242). Seeds were surface-sterilized

and placed on Petri dishes containing germination medium

[27]. The seeds were cold-treated at 4 -C for 2 days to ensure

uniform germination and moved to photoperiodic light.

Arabidopsis plants were grown under long-day conditions

(16-h light/8-h dark) at 22 -C and 80% relative humidity. The

aerial parts of the plants were collected 15 days after sowing,

frozen in liquid nitrogen and stored at �80 -C.Arabidopsis thaliana ecotype Columbia (Col-0) was

used for dark/light experiments and for the tissue distribu-

tion of CER6 and At5g43760 mRNAs. Seeds were sterilized

and plants grown as described above. For dark experiments,

half of the 15-day-old plants were transferred in the dark for

24 h, while the remaining plants were left under normal

conditions. Then, the aerial parts of the plants were

collected, frozen in liquid nitrogen and stored at �80 -C.For the expression study in Arabidopsis organs, plants were

grown under long-day conditions (16-h light/8-h dark) at

22 -C and 80% relative humidity. After 20 days of growth in

Petri dishes the plantlets were transferred to soil in the same

conditions for 30 more days, and then plants were collected

and frozen in liquid nitrogen.

2.2. RNA preparation and fluorescent labeling of the probes

To minimize interplant variability, tissues from a mini-

mum of 20 plants were pooled for each RNA extraction. For

the genotypes cer6-2, cer3-1 and Ler-0, three independent

RNA purifications were performed (only one for cer1-1).

Total RNA was extracted from the aerial parts of plants

using the guanidinium isothiocyanate/cesium chloride pro-

cedure. The amount of RNA was determined by spectro-

photometry at 260 nm and its integrity was assessed by

analyzing the ribosomal RNA bands after gel electro-

phoresis. For each microarray experiment, 20 Ag of RNA

were used as template to synthesize labeled cDNA probe by

using the Agilent Fluorescent Direct Label Kit (Agilent

Technologies) according to the manufacturer’s instructions.

Experiments were performed by using Cy3-dCTP and Cy5-

dCTP (Perkin Elmer/NEN).

2.3. Probe hybridization, slide washing and scanning

Fourteen Agilent Arabidopsis-1 microarrays were used

for the comparison study between the cer mutants and the

wild-type plants. Each array contained 60-mers length

oligonucleotides probes specific for 13704 Arabidopsis

genes and 440 control spots.

200 AL of hybridization solution were used for each

microarray. Hybridization chambers were incubated for 17

h at 60 -C in a hybridization oven (Robbins scientific) with

a rotation setting of 8. At the end of the incubation period,

the slides were washed sequentially at room temperature in

0.5 � SSC and 0.1% SDS for 5 min and 0.06 � SSC for 2

min. Then, they were dried at room temperature by

centrifugation at 400�g. Hybridized microarray slides

were scanned for cyanine3 at 532 nm and cyanine5 at

633 nm with a dual-laser Microarray Scanner (Agilent

technologies) with a 10-Am resolution.

2.4. Data analysis

Spot intensities were quantified with Agilent Feature

Extraction software version A.6.1.1. The settings were as

follows: non-uniformity outlier flagging on, population

outlier flagging on, background subtraction by using the

minimum signal on array and the LOWESS normalization

method. Spots, which were not flagged, were considered for

further analysis. To normalize the results obtained from the

different microarray experiments, we used the average value

obtained with each control sample (Ler-0) [28]. For each

gene, average intensity signals and their standard deviation

were calculated.

2.5. Real-time PCR conditions and analysis

Total RNA was isolated using RNeasy Plant Mini Kit

(Qiagen, USA). Purified RNA was treated with DNase I

using the DNA-free kit (Ambion, Austin, TX) and RNA

Table 1

Primer sequences used in real-time PCR experiment

Locus name Gene name Forward primer Reverse primer PCR size (bp)

At1g02205 CER1 5VAAGGATGGGAAATGCATGAG3V 5VTGATGTGGAAGGAGGAGAGG3V 109

At1g49240 actin 2 5VCCGAGCAGCATGAAGATTAAG3V 5VCATACTCTGCCTTAGAGATCCACA3V 124

At1g67730 GLOSSY 8 5VCTACTTCCTCCGACCATCCA3V 5VCTGGGTTACGAGCAACGAGT3V 158

At1g68530 CER6 5VTCTAGCTCGGTGAAGCTCAAG3V 5VAAGCTCAACGGCGACAATAG3V 108

At3g13920 eIF-4A-1 5VCTGATTTTGACCCGTCGTCT3V 5VAAGACAAACAACAAAGCCGAAT3V 177

At3g55360 TSC13 5VCTGGAGCTTTGGTGCTTACA3V 5VCTGGGGTCCCTCAGATTCTT3V 156

At4g24510 CER2 5VGGCGAAACAAACACGAATG3V 5VTCGGATATCCCAACCATTTC3V 119

At5g02310 CER3 5VAGTCCCCAGTCTGAGCTGAA3V 5VCTGGAGCCATATTTGGGTTG3V 117

At5g60390 EF-1-a. 5VTCGTTATGATCGACTCTCTTATGG3V 5VCCAAAAAGGAGGGAGAGAGAAAG3V 113

P. Costaglioli et al. / Biochimica et Biophysica Acta 1734 (2005) 247–258250

integrity was checked on a 1.5% (w/v) agarose gel. Total

RNA (2 Ag) was reverse transcribed in a 20 AL reaction

using an oligo(dT)18 primer and SuperScripti II reverse

transcriptase (Invitrogen, GmbH), according to the manu-

facturer’s instructions. The cDNA were then diluted 10

times and 2 AL was used as template for real-time PCR

experiment. Gene-specific primers used are listed in Table 1.

To establish a standard curve for quantification, each

PCR product was cloned in pGEM\-T Easy vector

(Promega, USA). Plasmids were quantified by fluorescence

using DNA Quantitation kit (Sigma-Aldrich, France).

Plasmids were serially diluted ranging from 106 to 102

copies/AL and used as template controls in real-time PCR

experiments. All standard samples were assayed in triplicate

wells and experimental samples were assayed in four

replicate wells.

Real-time PCR was performed on a iCycler i (Bio-Rad,

USA). Samples were amplified in a 25-AL reaction

containing 1 � SYBR Green Master Mix (Bio-Rad) and

300 nM of each primers. The thermal profile consisted of 1

cycle at 95 -C for 3 min 30 s followed by 40 cycles at 95 -Cfor 30 s and at 60 -C for 1 min. For each run, data

acquisition and analysis was done using the iCycler i iQ

software (version 3.0a, Bio-Rad).

The relative number of copies of each transcript was

determined by interpolating the Ct values of the unknown

samples to each standard curves. The quantity of actin 2,

EF-1-a and eIF-4A-1 mRNAs in each sample was

determined and used to normalize for differences of total

RNA amount [29].

Table 2

Relative expression of the CER genes in cer mutants and control plants

Gene name Locus name Ratio cer6-2/Ler-0

(n =6)

CER1 At1g02205 0.90T0.24 (1.25)

CER2 At4g24510 0.92T0.18 (0.91)

CER3 At5g02310 0.90T0.42

CER6 At1g68530 0.89T0.25 (1.11)

KCS1 At1g01120 0.97T0.13

WAX2 At5g57800 0.97T0.08

Relative expression of the CER genes was calculated from microarray experime

cer3-1/Ler-0. The ratio given in the column cer1-1/Ler-0 represents the mean v

The ratio calculated using the real-time PCR data are indicated in parenthesis. The

two distinct RNA samples.

3. Results and discussion

3.1. Analysis of wax genes expression in eceriferum mutants

To reveal new Arabidopsis genes that might be involved

in the biosynthesis of cuticular waxes, we performed a

comparative analysis between normal Arabidopsis plants

and cer mutants. We used the following criteria to choose

the cer mutant candidates: (i) the mutated gene was

identified and cloned; (ii) the phenotype of the mutant

was clearly established; (iii) biochemical studies showed a

dramatic decrease in total wax loads and/or an abnormal

wax composition. Based on these criteria, we selected the

cer6-2, cer1-1 and cer3-1 mutants. As shown by numerous

studies, these three mutants had a striking defect in the

production of waxes [17,18,30,31].

These mutant plants were grown, RNA samples from the

aerial part of 15-day-old plants were purified and used for the

microarray experiments. We chose this developmental stage

because it has been demonstrated that the difference in wax

load per leaf area between cer mutants and the wild type was

maximum at this stage [32]. Thereafter, 14 independent

microarray experiments were performed: six with cer6-2/

Ler-0, six with cer 3/Ler-0 and two with cer1/Ler-0.

The experimental data obtained from these microarray

experiments for six CER genes are shown in Table 2. These

data clearly indicated that unexpectedly, none of the well-

characterized CER genes displayed differential expression

of at least twofold or more between the cer mutants and the

wild-type plants. Additional real-time PCR experiments

Ratio cer3-1/Ler-0

(n =6)

Ratio cer1-1/Ler-0

(n =2)

1.51T0.21 (1.54) 1.51 (1.66)

0.92T0.45 (0.83) 0.89 (0.60)

0.86T0.61 (0.71) 1.1 (1.25)

0.88T0.28 (0.67) 1.05 (0.71)

0.92T0.08 Flagged spots

0.93T0.11 1.09

nts. Data are meanTS.D. values of six experiments for cer6-2/Ler-0 and

alue of two experiments.

y represent the mean value of two independent experiments performed with

P. Costaglioli et al. / Biochimica et Biophysica Acta 1734 (2005) 247–258 251

confirmed these results (Table 2). Moreover, none of the 204

candidate genes printed on the microarrays passed the

twofold expression cutoff which was selected. Altogether,

these experimental data showed that the mRNA steady-state

levels measured for the CER genes in the cer6-2, cer1 and

cer3 mutants were close to the normal values, although a

dramatic defect in wax synthesis takes place. We thus

concluded that the identification of genes involved in wax

biosynthesis throughout the comparison of gene expression

between normal and cer mutants was a fruitless approach.

We therefore proceeded to search another experimental

model to screen our candidate genes.

3.2. Expression of wax genes in the aerial parts of wild type

15-day-old Arabidopsis

Wax deposition starts as soon as epidermal cells are

exposed to air [5], and lasts for several weeks. In

Arabidopsis, the total epicuticular wax load per leaf area

is similar at 15 and 25 days after germination, while the leaf

area increases significantly during the same time period,

suggesting that wax biosynthesis is very active at these

stages [32]. In older plants, wax biosynthesis seemed to be

less active: the wax load decreased to 60% of the maximum

in Arabidopsis leaves 40 days after germination [32]. These

observations suggest that the expression of the genes

involved in wax biosynthesis should be elevated in the

aerial part of the plant during the first 4 weeks after

germination. Unfortunately, to our knowledge, only one

study was devoted to the expression of a wax biosynthetic

gene, CER6, during Arabidopsis development. The corre-

sponding mRNA was detected in 1-day-old seedlings,

reached maximal steady-state levels in 8-day-old seedlings,

and was detected throughout development in Arabidopsis

leaves [33].

To improve our view of the situation, we measured the

mRNA levels for two CER genes (CER6 and CER3) and

two genes encoding enzymes of the elongases (the KCR

GLOSSY8 and the ECR TSC13) in the aerial parts of

Arabidopsis seedlings at 15 and 28 days after germination.

As a control, we used samples from 15-day-old seedlings

which were placed in the dark for 24 h. Indeed, it has been

previously demonstrated that wax accumulation is modu-

Table 3

Expression of four wax genes in the aerial parts of 15-day-old and 28-day-old Ar

Gene name Locus name mRNA levels (real time-PCR)

15 days 28 days 1

CER6 At1g68530 29340T2877 30319T2292 1

GLOSSY8 At1g67730 5433T470 3482T 254

TSC13 At3g55360 3977T302 4760T176

CER3 At5g02310 3151T213 4217T328

Quantifications of CER6, GLOSSY8, TSC13 and CER3 mRNAs were performed

molecules of each transcript was determined in real-time PCR experiments by inte

the obtained values were normalized with respect to the Actin number of mol

normalization and background subtraction. Data are meanTS.D. values of four ex

lated by light [34] and that CER6 transcription is repressed

in the dark [33]. Our results were generated using real-time

PCR (Table 3). We showed that for each mRNAs, the

steady-state levels were similar at 15 and 28 days after

germination, suggesting that the rate of wax biosynthesis is

elevated and constant during this period. This was con-

firmed by the results obtained from the plants placed in the

dark for 24 h. The mRNA steady-state levels for CER6,

Glossy8 and TSC13 decreased in the absence of light.

Altogether, these results showed that the expression of

wax genes is elevated in the aerial parts of 15-day-old

Arabidopsis seedlings, and we used this criterion to screen

our candidate genes.

3.3. General informations about our microarray data

For the 13704 genes studied (representing half the

genome of Arabidopsis), the average intensity was around

1600, and 2555 genes (18.6%) gave signal intensity �average intensity. It should be noted that the average

intensity of the negative controls (180 blank spots on each

array) was 246T29, and that 6912 genes (50.4% of total)

gave a signal value � 500, indicating that they were not

expressed at significant levels in the samples studied.

Importantly, the relative level of expression measured

with our microarray experiments were quite similar to those

previously obtained by the real-time PCR technique. Data

obtained by the two techniques indicated that CER6 mRNAs

were most abundant, followed by GLOSSY 8, TSC13 and

CER3 mRNAs (Table 3). Thus, at least for this set of genes,

the correlation between the real-time PCR technique and our

microarray data appeared reasonably good.

3.4. Enzymes of the VLCFA biosynthesis pathway

The predominant cuticular wax constituents are all

derived from saturated VLCFAs with predominant chain

length C20–C32 [13,14,35] Therefore, the first step of the

wax biosynthesis pathway consists in the elongation of

C18:0 fatty acid produced in the plastid. Fatty acid

elongation involves four enzymatic reactions, which are

successively catalyzed by a KCS, a KCR, a h-hydroxyacyl-CoA dehydrase (DCH) and an ECR. As of today, we have

abidopsis seedlings

mRNA levels (microarray)

5 days/24 h dark Dark/light ratio 15 days

1229T933 0.382 5190T1131

1669T250 0.307 1971T4552036T68 0.511 1833T427

2781T203 0.882 993T307

using real-time PCR and microarray experiments. The relative number of

rpolating the Ct values of the unknown samples to each standard curve, and

ecules. The microarray results represent the processed intensity after dye

periments (real time-PCR) and nine microarray experiments.

P. Costaglioli et al. / Biochimica et Biophysica Acta 1734 (2005) 247–258252

no information concerning DCH, and therefore we will only

discuss the fates of the KCS, KCR and ECR.

KCS catalyses the condensation of malonyl-CoA with a

long chain acyl-precursor. This is the rate-limiting step of

VLCFA synthesis which determines the acyl chain length of

the VLCFAs produced [36]. Four KCS have been studied

from Arabidopsis. A single condensing enzyme, FAE1,

catalyzes VLCFA synthesis in seeds [36,37], whereas KCS1

[38], FDH [39,40] and CER6 [30,41], have been implicated

in the synthesis of wax components. Millar et al. [41]

showed that there is no significant functional overlap of

CER6 with KCS1 and FDH activities in the stem of

Arabidopsis. Another VLCFA condensing enzyme, CER 60

[30], with high amino acid sequence identity to CER6, did

not appear to make a significant contribution to the

synthesis of stem surface lipids and was expressed at low

levels in the shoots [33].

A search in the Arabidopsis Lipid Gene Database [42]

revealed the existence of 16 supplementary KCS-related

genes bringing-up the total number to 21. To get insights

into the different families that belong to this class of

enzyme, we aligned the corresponding sequences with

CLUSTAL W (1.83) [43] and trees were constructed using

TreeView [44] (Fig. 2). From this tree, four main families of

KCS were identified, each corresponding to one of the four

KCS previously described in Arabidopsis. The smallest

family comprised two members, CER6 and CER60. As

expected from previous studies, we measured high levels of

Fig. 2. Expression of the members of the KCS family in 15-day-old shoots. The d

softwares, based on the amino acid sequences of the 21 KCS-related genes identifi

intensities TS.D. of six to nine different experiments. ND: not determined (correspo

genes whose mRNA expression levels are high in the samples studied.

CER6 expression in the aerial parts of the seedlings.

However, the fact that CER60 was expressed to substantial

levels, representing more than half of CER6 levels, was

quite surprising knowing that it has been reported that this

gene was expressed at a lower level than CER6 [33]. This

may indicate that CER60 plays a particular role in wax

deposition at certain stages of development that require

higher levels of wax production, as previously suggested

[33]. From the seven sequences belonging to the FDH

family, three were not expressed in our samples. A search in

the Arabidopsis Lipid Gene Database indicated that no EST

exists for At5g49070 and At1g71160, suggesting that these

two genes may represent in fact pseudogenes, or genes that

are expressed in unusual physiological conditions. Con-

cerning At3g52160, one EST has been identified in a flower

cDNA library, but owing to its lack of expression in our

samples, it seems unlikely that this gene plays a role in stem

or leaf wax biosynthesis. As expected on the basis of

current knowledge, FDH was expressed to substantial

amounts, its expression level being similar to that of

CER6. Three other genes, closely related to each other,

At5g04530, At1g07720 and At2g28630 were expressed at

moderate levels. Their possible implication in wax biosyn-

thesis should be evaluated.

Six genes closely related to the seed-specific condensing

enzyme FAE1 were identified in the Arabidopsis genome.

One of them, At4g34250 was not printed on the micro-

arrays used, but data from the Arabidopsis Lipid Gene

endrogram has been generated using CLUSTALW (V 1.83) and TreeView

ed in the Arabidopsis genome. The levels of expression are means of signal

nding target absent from the microarray used). The black boxes indicate the

P. Costaglioli et al. / Biochimica et Biophysica Acta 1734 (2005) 247–258 253

Database showed that five EST out of seven originated from

seed cDNA libraries, suggesting that this KCS may play a

role in seed oil production. Such a role has been

demonstrated for FAE1 [36,37], and as expected, we did

not measure any significant level of transcripts for this gene

in our samples. Thus, three genes belonging to this family

appeared to be expressed to moderate levels in the aerial

part of 15-day-old Arabidopsis seedlings: At1g19440,

At2g16280 and At2g15090. These three genes may catalyse

a condensing reaction in leaves and stems similar to that

achieved by FAE1 in the Arabidopsis seeds.

The last family, composed of six members, corresponds

to the KCS1 gene characterized by Todd et al. [38]. From

the four sequences present on our array, two were expressed

to moderate levels, At1g01120 (KCS1) and At2g26640, and

one, At5g43760, to levels comparable to those measured for

CER6. The expression of the At5g43760 gene was further

studied using the real-time PCR technique. First, we

compared the level of expression of the corresponding

RNA in 15-day-old plants which were grown under normal

light conditions or placed for 24 h in the dark. The amount

of At5g43760 mRNA was significantly reduced in the

absence of light. We measured an At5g43760 mRNA dark/

light ratio of 0.557 which is comparable to the one measured

for TSC13 (Table 3). We also measured the expression of

the corresponding transcripts in different Arabidopsis

organs and we compared the tissue distribution with that

of the CER6 mRNA (Fig. 3). Our results showed that these

Fig. 3. Tissue distribution of CER6 and At5g43760 mRNAs. The relative

levels of each mRNA were determined by real-time PCR in roots, stems,

cauline leaves, rosette leaves, flowers and siliques. The relative number of

copies of each transcript was determined by interpolating the Ct values of

the unknown samples to each standard curves. The quantity of actin 2, EF-

1-a and eIF-4A-1 mRNAs in each sample was determined and used to

normalize for differences of total RNA amount. Results are presented as the

percentage of the maximal level measured. The data represent the means

TS.D. of four replicates from two independent experiments.

two mRNA species were co-regulated, being highly

expressed in cauline leaves, at a slightly lower level in

rosette leaves, stems, flowers and siliques, and not

significantly expressed in roots. Thus, the abundance of

At5g43760 transcripts in the aerial part of the plants, its co-

expression with CER6 in different organs and in the light/

dark experiment suggested that this KCS may play a role in

wax deposition. Moreover, it has been recently demonstra-

ted that the gene product of At5g43760 is able to catalyze

the formation of VLCFAs when expressed in yeast [45]. The

putative role played by this gene during wax biosynthesis

will be evaluated by using one of the four SALK mutants

which are available.

Finally, from the 18 KCS studied, twelve were expressed

in the aerial parts of the plants, and among them, three

presented quite high levels of mRNA expression: CER6,

FDH and At5g43760.

The second step in the VLCFA synthesis requires the

reduction of the h-ketoacyl-CoA to form a 3-hydroxyacyl-

CoA, reaction catalyzed by a KCR. The recent identification

of a putative KCR, YBR159w, in the genome of S.

cerevisiae [46] led to a search for orthologs in the

Arabidopsis genome, and two candidates were identified:

At1g67730 and At1g24470. It has been demonstrated that

At1g67730 functionally complements the ybr159D mutant

[46], and that mutation in the maize gene ortholog to

At1g67730 led to a wax-less phenotype referred to as

glossy8 [22,23]. No known biological role has been

assigned to the At1g24470 gene. The mRNA expression

levels were measured for these two genes in our samples,

and we obtained values of 1971T455 (n =9) for At1g67730

and 427T217 (n =8) for At1g24470. These results showed

that At1g67730 was expressed to significant levels, as

expected for a gene encoding a component of the elongase

complex. On the other hand, due to its low level of

expression, the probability that the second KCR candidate

identified in the Arabidopsis genome plays an important

role in synthesis of very-long chain precursors of the wax

pathway is very low.

The final reaction of VLCFA synthesis is catalyzed by a

trans-2,3 Enoyl-CoA reductase. The first gene coding for

this class of enzyme, TSC13, has been identified in yeast

[47]. It has been shown that TSC13 is essential for yeast

viability, and that it catalyses a step in the fatty acid

elongation cycle for acyl-CoA substrates of all chain

lengths. A similarity search using the TSC13 protein

sequence led to the identification of an ortholog in the

Arabidopsis genome, At3g55360, which was described to

be similar to mammalian steroid 5-alpha-reductase. Further

searches in the TAIR database allow the identification of

five genes with the same similarity description: At1g72590,

At2g16530, At2g38050, At3g55360 and At5g16010. Align-

ment of these five sequences showed that the first three were

closely related, suggesting that they may catalyze similar

reactions, but they were not expressed in our samples. On

the contrary, the At3g55360 and At5g16010 relative

Table 4

Genes of the acyl-reduction pathway

Locus NAME Expression level TS.D.

FAR At3g11980 131T55

At3g44540 318T106

At3g44550 294T143At3g44560 299T110

At3g56700 nd

At4g33790 nd

At5g22420 121T73At5g22500 321T134

Wax synthase At1g34490 289T193

At1g34500 181T113

At1g34520 328T239At3g51970 257T108

At5g51420 221T62

At5g55320 88T41At5g55330 121T57

At5g55340 nd

At5g55350 147T67

At5g55360 120T45At5g55370 nd

At5g55380 601T392

For each gene, average intensity signals and their standard deviation were

calculated from the nine measurements, when at least six replicate values

were not flagged.

nd: not determined = corresponding target absent from the microarray.

P. Costaglioli et al. / Biochimica et Biophysica Acta 1734 (2005) 247–258254

expression values were 1833T427 (n =8) and 9636T1507(n =9), respectively. As stated above, the former is an

ortholog of TSC13 which can functionally complement a

yeast tsc13 mutant [48]. The level of expression measured

for At5g16010 was significantly higher than those measured

for At3g55360, suggesting that the corresponding protein

could play an important role at this developmental stage.

After their synthesis by the elongase complex, the

VLCFAs enter either the acyl reduction pathway, where

they are transformed into primary alcohols and wax esters,

or the decarbonylation pathway, to form the aldehydes,

alkanes, secondary alcohols and ketones.

3.5. Enzymes of the acyl-reduction pathway

The acyl reduction pathway converts VLCFA to very-

long-chain primary alcohols and esters. In many higher

plants, these classes of molecules are found in cuticular

waxes, but some species produce linear esters of long-chain

alcohols and fatty acids as a seed lipid energy reserve.

Indeed, most of our knowledge upon this pathway came

from studies of jojoba seeds.

The first step of the acyl-reduction pathway consists in

the formation of primary alcohols from VLCFA precursors.

This reaction is catalyzed by a fatty acyl-CoA reductase

(FAR) [49]. The first FAR gene ever identified was from

jojoba [50] and this lead to the identification of eight

related-sequences in the Arabidopsis genome [11]. Among

them, Arabidopsis MS2 codes for a tapetum-specific protein

which is essential for pollen formation [51]. Indeed, it has

been proposed that the FAR-like enzymes could be

responsible for the formation of the alcohol moiety of the

wax esters found in the cuticular lipids of Arabidopsis. Our

results did not support this hypothesis. None of the six FAR

genes present on our array was expressed to significant

levels in the aerial part of the 15-day-old plants (Table 4).

Concerning the two other FAR genes, At3g56700 and

At4g33790, no EST associated with At3g56700 was found

in the TAIR website and data available from NASCArrays

indicated that At4g33790 is only expressed in Arabidopsis

roots. In summary, none of the FAR genes identified so far

in the Arabidopsis genome seems to play a significant role

in cuticular wax deposition.

The final reaction of this pathway is catalyzed by a fatty

acyl-CoA: fatty alcohol acyltransferase (E.C.2.3.1.75, wax

synthase, WS). Purification of a WS from embryo of jojoba

led to the cloning of the corresponding cDNA [24]. Since

then, twelve Arabidopsis ORFs with homology to jojoba

WS were identified [11,42], and it has been proposed that

several of this genes are required for cuticular wax

production in Arabidopsis leaves and stems [11]. Our

results did not support this hypothesis. None of the 10

WS genes present on our microarrays was expressed to

significant levels in the aerial parts of 15-day-old seedlings

(Table 3). Moreover, no EST has ever been identified in

Arabidopsis library for any of the WS [42].

Our results showing that neither the FAR genes nor the

WS genes were expressed in 15-day-old Arabidopsis leaves

and stems were quite unexpected. The acyl-reduction

pathway accounts for 20% of the total wax amount [52],

and primary alcohol and ester amounts in 25-day-old plants

of the Ler genotype represent 5% and 2.3% of the total wax

amount in stems, and 3.5% and 19.5% in leaves, respec-

tively [20]. To explain these contradictory results, we can

raise two hypotheses. The first is that the amino acid

sequences of the FAR and WS enzymes involved in primary

alcohol and ester production for wax production are very

different from that of the enzymes involved in the ester

production for seed in jojoba, and thus they have not been

identified yet. Indeed, a novel bifunctional wax ester

synthase/acyl-CoA:diacylglycerol acyltransferase has been

recently characterized in Acinetobacter calcoaceticus [53].

The author identified 11 bifunctional enzyme-related

sequences in the Arabidopsis genome, and we can postulate

that they may play a role in wax biosynthesis. Three of them

were spotted on our microarrays (At3g49210, At3g49190

and At3g49200), but their respective mean signal inten-

sities, 272, 243 and 301, indicated that they were not

expressed to significant levels in our samples. Nevertheless,

data from NASCArrays indicate that two other members of

this family, At5g12420 and At5g37300, were expressed to

low but significant values in a number of different

conditions, and we cannot rule out that they may play a

role in wax ester synthesis.

The second hypothesis is that wax esters and primary

alcohols are synthesized via an alternative pathway. Indeed,

an early biochemical study suggested that primary alcohol

P. Costaglioli et al. / Biochimica et Biophysica Acta 1734 (2005) 247–258 255

production could be a two-step process, which would be

carried out by two separate enzymes, an NADH-dependent

acyl-CoA reductase required for a reduction of VLCFAs to

aldehydes, and an NADPH-dependent aldehyde reductase

necessary for the subsequent reduction of aldehydes to

primary alcohols [54]. This hypothesis was later abandoned

when it was demonstrated that alcohol formation from

VLCFA is carried out by a single fatty acyl-CoA reductase

(FAR) in pea leaves [49] and in jojoba embryo [50], but the

situation in Arabidopsis seedlings has not been evaluated so

far.

3.6. The decarbonylation pathway

In Arabidopsis stems, products of the decarbonylation

pathway account for around 80% of the total wax. This

pathway is initiated by the production of aldehydes from

VLCFA precursors by a membrane-bound fatty acyl-CoA

reductase. These aldehydes are then decarbonylated by an

aldehyde decarbonylase to odd-chain alkanes. The follow-

ing step consists in the hydroxylation of alkanes to

secondary alcohols, and oxidation of secondary alcohols

to ketones.

Surprisingly, although this pathway provides the main

amount of wax in Arabidopsis, little information exists

concerning the identity of the genes that encode these four

different enzymes. CER1 was suggested to encode an

aldehyde decarbonylase in view of the phenotype reported

for the cer1 mutants [18,55]. The obtention of the CER1

sequence (At1g02205) was not helpful in confirming its

biochemical function [56], and the cloning of CER1

orthologs in maize and Kleinia odora raised the possibility

that these genes may encode membrane-bound receptors

[57]. More recently, two teams simultaneously identified a

new Arabidopsis CER1-related gene, designated WAX2 [58]

or YRE [59], presenting 30% amino acid identity with

CER1. Mutation of the corresponding gene resulted to a cer

phenotype, and analysis of the wax composition of the

mutants led to the proposal that WAX2/YRE catalyses the

transformation of acyl-CoA to aldehydes [58,59].

We identified five CER1-related genes in the Arabidopsis

thaliana genome: At1g012190, At1g02205, At2g37700,

At5g28280 and At5g57800. Only two were expressed in our

samples, At1g02205/CER1 (998T401) and At5g57800/

WAX2/YRE (2027T144, n =9), thereby confirming that they

play an important role in wax deposition.

3.7. Other Cer genes (CER2/CER3)

Analysis of the cer2 wax composition showed a deficit in

C28 and C30 components, suggesting that the CER2 protein

plays a role in the elongation of C26 fatty acids [18,55]. The

characterization of CER2 gene provided no clues of its

biochemical function [60,61], but localization of the CER2

protein in nuclear fractions suggested a regulatory role in

cuticular wax accumulation [62]. More recently, the cloning

of an acetyl CoA:deacetylvindoline 4-O-acetyltransferase

from Catharanhtus roseus led to the identification of 13

acyltransferase-related sequences in the Arabidopsis

genome, including CER2 [63]. Interestingly, the 13 acyl-

transferase-related sequences identified contain the Pfam

profile 02458. We identified 41 transferases containing the

Pfam profile 02458 in the TAIR database, and their

sequences were compared (Fig. 4, Supplementary data).

This alignment showed that 6 sequences were more or less

related to the CER2 protein. The mRNA expression levels

of four of them were measured in our microarray experi-

ments. At3g23840 and At2g46110 were not expressed to

above background values in our samples. Surprisingly, the

amount of CER2 mRNA was moderate, suggesting that

CER2 functions as a regulatory protein rather than as a

biosynthetic enzyme. In addition, we found that At4g13840,

also called CER2-like, was expressed to high levels in our

samples. Unfortunately, no other information exists con-

cerning the function of this gene and further analysis of

Arabidopsis mutants will be required to conclude that the

At4g13840 plays any role in wax biosynthesis.

The singularity of CER3 gene (At5g02310) comes from

the fact that no other related-sequences are present in the

Arabidopsis genome. Studies of the biochemical composi-

tion of the cer3 mutant led to the suggestion that the

corresponding protein was involved in fatty acid elongation

[18]. However, sequencing of the CER3 gene revealed the

presence of a putative nuclear localization sequence in the

corresponding protein, and two phosphorylation sites

known to be involved in transport of proteins in nucleus

[15,64]. These observations led to the hypothesis that CER3

may encode a regulatory protein. Knowing that nuclear

regulatory proteins are usually expressed at a low level, our

results were compatible with such a role for CER3. Indeed,

the relative expression level measured for this gene was

993T307 (n =9), which is significant but low when

compared to genes encoding biosynthetic enzymes such as

CER6. This finding was confirmed by our PCR experiment

(Table 3).

3.8. Wax transport

After their synthesis along the secretion pathway, the

wax constituents accumulate in the plasma membrane of the

epidermal cells. Then, they must leave this hydrophobic

environment and move toward the plant surface. It has been

proposed that members of the ABC transporters family may

play a role in this process [11], and this hypothesis has been

recently supported by the demonstration that the cer5

mutant resulted from a point mutation in the At1g51500

gene, which encodes an ABC transporter [21].

In the Arabidopsis genome, 135 sequences have been

identified as putative ABC transporters [65,66]. The mRNA

levels for 113 of 135 Arabidopsis ABC transporter genes

were measured and 60 (51%) did not give above-back-

ground values (Table 5, Supplementary data). The proba-

P. Costaglioli et al. / Biochimica et Biophysica Acta 1734 (2005) 247–258256

bility that one of these genes may play a role in wax

transport is therefore quite low. On the other side of the

scale, only 3 (2.6%) were highly expressed with signal

intensity > 5000, 28 (¨25%) were moderately expressed

(1000<<3000) and 22 were expressed at a low level

(550<<1000). The three highly expressed genes belong to

the distinct subfamilies: Full molecule (At2g26910), Half-

molecule (At5g64940) and soluble proteins (At5g64840).

Unfortunately, there is no published information about these

genes and further analysis will be required before conclud-

ing univocally on their potential involvement in wax

transport. We found that the WBC12 transporter corre-

sponding to the CER5 sequence (At1g51500) is expressed

in our samples at a low level (862T276, n =9), therebyconfirming the findings of Pighin et al. [21] which

suggested that additional transporters should be involved

in wax transport.

After their departure from the plasma membrane, wax

components should cross the cell wall and the cutin matrix

to reach the outermost part of the surface. It is generally

assumed that these highly hydrophobic molecules cannot

move freely in aqueous media, and therefore should bind to

carrier molecules during this process. Although there has

been no direct demonstration of a role for the LTPs in

transferring wax components or cutin across the apoplast,

some experimental evidence suggests their involvement in

cuticular lipid synthesis [11,67–72]. The proteins belonging

to the LTPs family are characterized by the presence of a

consensus motif of eight conserved Cys [73], and 71

putative LTPs were retrieved and classified into eight

groups (Type 1–8) based on the number of amino acids

between the fourth and fifth Cys in the core of the motif

[42].

In this study, the mRNA levels for 37 of the 71

Arabidopsis LTPs genes were measured, and 22 of them

(¨60%) did not give above-background values (Table 6,

Supplementary data). On the other hand, the expression of

15 LTP mRNAs was detected, and five gave quite elevated

values (expression levels > 4000): At2g38540 (LTP1),

At3g51600 (LTP5), At1g27950, At3g08770 (LTP6) and

At4g22490. The high level of expression for LTP1 and

LTP5 in Arabidopsis thaliana leaves has already been

reported [67]. The same study showed that the LTP2 mRNA

quantities were high in leaves during Arabidopsis develop-

ment, but our results did not confirm this observation.

Indeed, our microarray data showed a low level of

expression (717T151) for this gene (At2g38530). The

discrepancy between these results may come from the

different probes used in the experiments. Arondel et al. used

cDNA inserts as probes for their Northern blot experiments

[67], and thus the possibility that their probe hybridized to

closely related mRNAs remains. Indeed, clustering analysis

of the 37 LTPs proteins studied here showed that LTP2

sequence is closely related to the highly expressed LTP1,

demonstrating the risk of cross-hybridization (Fig. 5,

Supplementary data). In contrast, the 60mers-specific

sequences spotted on the microarrays allow a better

discrimination between closely related mRNA sequences.

Another interesting finding is that our data are somewhat

different from those obtained through the study of EST

counts in database. Indeed, by using the frequency of gene

transcripts in unbiased database, it is possible to gather

information regarding relative levels of gene expression

[74]. However, due to differences in the size of libraries, it is

often difficult to compare results from different sources

directly. To circumvent this problem, statistical analysis

such as that developed by Beisson et al. [42] can be used.

We collected the data corresponding to the leaf and seedling

synthetic libraries for the 37 LTPs studied here from their

website (http://www.plantbiology.msu.edu/lipids/genesurvey/

index.htm) (Table 6, Supplementary data). We did not

expect identical results after comparison between data

generated with the aerial parts of 15-day-old plants and

leaf or seedling samples, but we found significant differ-

ences. Indeed, we found no EST counts in the two synthetic

libraries screened for 6 out of the 15 LTP genes that we

found to be expressed in the aerial part of the plants

samples. More surprisingly, the Arabidopsis Lipid Gene

Database indicated eleven EST for LTP3 and seven for

LTP4, whereas neither we nor Arondel et al. [67] measured

any detectable LTP3 and LTP4 mRNA levels in leaf and

aerial part of the plants. This suggests that the extrapolation

of gene expression levels from in silico approaches should

be considered with prudence.

In conclusion, from the 204 candidate genes studied, only

25% were shown to be expressed to significant levels in the

aerial part of Arabidopsis seedlings when wax deposition

takes place, thus significantly reducing the number of genes

which will be worth studying using reverse genetics to

demonstrate their involvement in wax accumulation.

Acknowledgement

The PhD work of Mrs Christel Garcia was supported by a

grant from the Conseil Regional d’Aquitaine. We gratefully

acknowledge funding from the French Ministere de la

Jeunesse, de l’Education Nationale et de la Recherche via

C.N.R.S. and the University Victor Segalen Bordeaux 2.

Appendix A. Supplementary data

Supplementary data associated with this article can be

found, in the online version, at doi:10.1016/j.bbalip.

2005.04.002.

References

[1] D. Edwards, G.D. Abbott, J.A. Raven, Cuticles of early land plants:

a palaeoecophysiological evaluation, in: G. Kerstiens (Ed.), Plant

P. Costaglioli et al. / Biochimica et Biophysica Acta 1734 (2005) 247–258 257

Cuticles: An Integrated Functional Approach, Bios Scientific

Publishers, Oxford, 1996, pp. 1–32.

[2] D. Edwards, H. Kerp, H. Hass, Stomata in early land plants: an

anatomical and ecophysiological approach, J. Exp. Bot. 49 (1998)

255–278.

[3] G. Kerstiens, Cuticular water permeability and its physiological

significance, J. Exp. Bot. 47 (1996) 1813–1832.

[4] M. Riederer, L. Schreiber, Protecting against water loss: analysis of the

barrier properties of plant cuticles, J. Exp. Bot. 52 (2001) 2023–2032.

[5] C.E. Jeffree, Structure and ontogeny of plant cuticles, in: G. Kerstiens

(Ed.), Plant Cuticles: An Integrated Functional Approach, Bios.

Scientific Publishers, Oxford, 1996, pp. 33–82.

[6] P.E. Kolattukudy, Cutin, suberin and waxes, in: P.K. Stumpf (Ed.),

The Biochemistry of Plants, vol. 4, Academic Press, New York,

1980, pp. 571–654.

[7] P.J. Holloway, The chemical constitution of plant cutins, in: D.F.

Cutler, K.L. Alvin, C.E. Price (Eds.), The Plant Cuticle, Academic

Press, London, 1982, pp. 45–85.

[8] P.J. Holloway, J. Wattendorff, Cutinized and suberized cell walls, in:

K.C. Vaughn (Ed.), CRC Handbook of Plant Cytochemistry, vol. 2,

CRC Press, Boca Raton, 1987, pp. 1–35.

[9] T.J. Walton, Waxes, cutin and suberin, in: J.L. Harwood, J. Boyer

(Eds.), Lipids, Membranes and Aspects of Photobiology, Academic

Press, London, 1990, pp. 105–158.

[10] P. Von Wettstein-Knowles, Waxes, cutin and suberin, in: T.S. Moore

(Ed.), Lipid Metabolism in Plants, CRC Press, Boca Raton, 1993,

pp. 127–166.

[11] L. Kunst, A.L. Samuels, Biosynthesis and secretion of plant cuticular

wax, Prog. Lipid Res. 42 (2003) 51–80.

[12] M.A. Jenks, R.J. Joly, P.J. Peters, P.J. Rich, J.D. Axtell, E.N.

Ashworth, Chemically induced cuticle mutation affecting epidermal

conductance to water vapor and disease susceptibility in sorghum

bicolor (L.) moench, Plant Physiol. 105 (1994) 1239–1245.

[13] C. Cassagne, R. Lessire, J.J. Bessoule, P. Moreau, A. Creach, F.

Schneider, B. Sturbois, Biosynthesis of very-long-chain fatty acids in

higher plants, Prog. Lipid Res. 33 (1994) 55–69.

[14] P. Von Wettstein-Knowles, Biosynthesis and genetics of waxes, in:

R.J. Hamilton (Ed.), Waxes: Chemistry, Molecular Biology and

Functions, Oily Press, Dundee, 1995, pp. 91–129.

[15] B. Lemieux, Molecular genetics of epicuticular wax biosynthesis,

Trends Plant Sci. 1 (1996) 312–318.

[16] C. Mariani, M. Wolters-Arts, Complex waxes, Plant Cell 12 (2000)

1795–1798.

[17] M. Koornneef, C.J. Hanhart, F. Thiel, A genetic and phenotypic

description of eceriferum (cer) mutants in Arabidopsis thaliana,

J. Hered. 80 (1989) 118–122.

[18] A. Hannoufa, J. McNevin, B. Lemieux, Epicuticular wax of

eceriferum mutants of Arabidopsis thaliana, Phytochemistry 33

(1993) 851–855.

[19] M.A. Jenks, H.A. Tuttle, S.D. Eigenbrode, K.A. Feldmann, Leaf

epicuticular waxes of the eceriferum mutants in Arabidopsis, Plant

Physiol. 108 (1995) 369–377.

[20] A.M. Rashotte, M.A. Jenks, K.A. Feldmann, Cuticular waxes on

eceriferum mutants of Arabidopsis thaliana, Phytochemistry 57

(2001) 115–123.

[21] J.A. Pighin, H. Zheng, L.J. Balakshin, I.P. Goodman, T.L. Western, R.

Jetter, L. Kunst, A.L. Samuels, Plant cuticular lipid export requires an

ABC transporter, Science 306 (2004) 702–704.

[22] X. Xu, C.R. Dietrich, M. Delledonne, Y. Xia, T.J. Wen, D.S.

Robertson, B.J. Nikolau, P.S. Schnable, Sequence analysis of the

cloned glossy8 gene of maize suggests that it may code for a beta-

ketoacyl reductase required for the biosynthesis of cuticular waxes,

Plant Physiol. 115 (1997) 501–510.

[23] X. Xu, C.R. Dietrich, R. Lessire, B.J. Nikolau, P.S. Schnable, The

endoplasmic reticulum-associated maize GL8 protein is a component

of the acyl-coenzyme A elongase involved in the production of

cuticular waxes, Plant Physiol. 128 (2002) 924–934.

[24] K.D. Lardizabal, J.G. Metz, T. Sakamoto, W.C. Hutton, M.R. Pollard,

M.W. Lassner, Purification of a Jojoba embryo wax synthase, cloning

of its cDNA, and production of high levels of wax in seeds of

transgenic Arabidopsis, Plant Physiol. 122 (2000) 645–655.

[25] P. Broun, P. Poindexter, E. Osborne, C.Z. Jiang, J.L. Riechmann,

WIN1, a transcriptional activator of epidermal wax accumulation in

Arabidopsis, Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 4706–4711.

[26] A. Aharoni, S. Dixit, R. Jetter, E. Thoenes, G. van Arkel, A. Pereira,

The SHINE clade of AP2 domain transcription factors activates wax

biosynthesis, alters cuticle properties, and confers drought tolerance

when overexpressed in Arabidopsis, Plant Cell 16 (2004) 2463–2480.

[27] D. Valvekens, M. Van Montagu, M. Van Lijsebettens, Agrobacterium

tumefaciens-mediated transformation of Arabidopsis thaliana root

explants by using kanamycin selection, Proc. Natl. Acad. Sci. U. S. A.

85 (1988) 5536–5540.

[28] Y.H. Yang, S. Dudoit, P. Luu, D.M. Lin, V. Peng, J. Ngai, T.P. Speed,

Normalization for cDNA microarray data: a robust composite method

addressing single and multiple slide systematic variation, Nucleic

Acids Res. 30 (2002) e15.

[29] J. Vandesompele, K. De Preter, F. Pattyn, B. Poppe, N. Van Roy, A.

De Paepe, F. Speleman, Accurate normalization of real-time quanti-

tative RT-PCR data by geometric averaging of multiple internal

control genes, Genome Biol. 3 (2002) 1–11.

[30] A. Fiebig, J.A. Mayfield, N.L. Miley, S. Chau, R.L. Fischer, D. Preuss,

Alterations in CER6, a gene identical to CUT1, differentially affect

long-chain lipid content on the surface of pollen and stems, Plant Cell

12 (2000) 2001–2008.

[31] D. Preuss, B. Lemieux, G. Yen, R.W. Davis, A conditional sterile

mutation eliminates surface components from Arabidopsis pollen

and disrupts cell signaling during fertilization, Genes Dev. 7: (1993)

974–985.

[32] M.A. Jenks, H.A. Tuttle, K.A. Feldmann, Changes in epicuticular

waxes on wildtype and eceriferum mutants in Arabidopsis during

development, Phytochemistry 42 (1996) 29–34.

[33] T.S. Hooker, A.A. Millar, L. Kunst, Significance of the expression of

the CER6 condensing enzyme for cuticular wax production in

Arabidopsis, Plant Physiol. 129 (2002) 1568–1580.

[34] P. Von Wettstein-Knowles, P. Avato, J.D. Mikkelsen, Light promotes

synthesis of the very long chain fatty acyl chains in maize wax, in: D.

Mazliak, P. Benveniste, C. Costes, R. Douce (Eds.), Biogenesis and

Function of Plant Lipids, Elsevier/North Holland Biomedical Press,

NY, 1980, pp. 271–274.

[35] D. Post-Beittenmiller, Biochemistry and molecular biology of wax

production in plants, Annu. Rev. Plant Physiol. Plant Mol. Biol. 47

(1996) 405–430.

[36] A.A. Millar, L. Kunst, Very-long-chain fatty acid biosynthesis is

controlled through the expression and specificity of the condensing

enzyme, Plant J. 12 (1997) 121–131.

[37] D.W. James, E. Lim, J. Keller, I. Plooy, E. Ralston, H.K. Dooner,

Directed tagging of the Arabidopsis FATTY ACID ELONGATION1

(FAE1) gene with the maize transposon activator, Plant Cell 7 (1995)

309–319.

[38] J. Todd, D. Post-Beittenmiller, J.G. Jaworski, KCS1 encodes a fatty

acid elongase 3-ketoacyl-CoA synthase affecting wax biosynthesis in

Arabidopsis thaliana, Plant J. 17 (1999) 119–130.

[39] A. Yephremov, E. Wisman, P. Huijser, C. Huijser, K. Wellesen, H.

Saedler, Characterization of the FIDDLEHEAD gene of Arabidopsis

reveals a link between adhesion response and cell differentiation in the

epidermis, Plant Cell 11 (1999) 2187–2201.

[40] R.E. Pruitt, J.P. Vielle-Calzada, S.E. Ploense, U. Grossniklaus, S.J.

Lolle, FIDDLEHEAD, a gene required to suppress epidermal cell

interactions in Arabidopsis, encodes a putative lipid biosynthetic

enzyme, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 1311–1316.

[41] A.A. Millar, S. Clemens, S. Zachgo, E.M. Giblin, D.C. Taylor, L.

Kunst, CUT1, an Arabidopsis gene required for cuticular wax

biosynthesis and pollen fertility, encodes a very-long-chain fatty acid

condensing enzyme, Plant Cell 11 (1999) 825–838.

P. Costaglioli et al. / Biochimica et Biophysica Acta 1734 (2005) 247–258258

[42] F. Beisson, A.J. Koo, S. Ruuska, J. Schwender, M. Pollard, J.J.

Thelen, T. Paddock, J.J. Salas, L. Savage, A. Milcamps, V.B. Mhaske,

Y. Cho, J.B. Ohlrogge, Arabidopsis genes involved in acyl lipid

metabolism. A 2003 census of the candidates, a study of the

distribution of expressed sequence tags in organs, and a web-based

database, Plant Physiol. 132 (2003) 681–697.

[43] J.D. Thompson, D.G. Higgins, T.J. Gibson, CLUSTALW: improving

the sensitivity of progressive multiple sequence alignment through

sequence weighting, position-specific gap penalties and weight matrix

choice, Nucleic Acids Res. 22 (1994) 4673–4680.

[44] R.D. Page, TREEVIEW: an application to display phylogenetic

trees on personal computers, Comput. Appl. Biosci. 12 (1996)

357–358.

[45] S. Trenkamp, W. Martin, K. Tietjen, Specific and differential

inhibition of very-long-chain fatty acid elongases from Arabidopsis

thaliana by different herbicides, Proc. Natl. Acad. Sci. U. S. A. 101

(2004) 11903–11908.

[46] F. Beaudoin, K. Gable, O. Sayanova, T. Dunn, J.A. Napier, A

Saccharomyces cerevisiae gene required for heterologous fatty acid

elongase activity encodes a microsomal beta-keto-reductase, J. Biol.

Chem. 277 (2002) 11481–11488.

[47] S.D. Kohlwein, S. Eder, C.S. Oh, C.E. Martin, K. Gable, D. Bacikova,

T. Dunn, Tsc13p is required for fatty acid elongation and localizes to a

novel structure at the nuclear-vacuolar interface in Saccharomyces

cerevisiae, Mol. Cell. Biol. 21 (2001) 109–125.

[48] K. Gable, S. Garton, J.A. Napier, T.M. Dunn, Functional characte-

rization of the Arabidopsis thaliana orthologue of Tsc13p, the enoyl

reductase of the yeast microsomal fatty acid elongating system, J. Exp.

Bot. 55 (2004) 543–545.

[49] J. Vioque, P.E. Kolattukudy, Resolution and purification of an

aldehyde-generating and an alcohol-generating fatty acyl-CoA reduc-

tase from pea leaves (Pisum sativum L.), Arch. Biochem. Biophys.

340 (1997) 64–72.

[50] J.G. Metz, M.R. Pollard, L. Anderson, T.R. Hayes, M.W. Lassner,

Purification of a jojoba embryo fatty acyl-Coenzyme A reductase and

expression of its cDNA in high erucic acid rapeseed, Plant Physiol.

122 (2000) 635–644.

[51] M.G.M. Aarts, R. Hodge, K. Kalantidis, D. Florack, Z.A. Wilson, B.J.

Mulligan, W.J. Stiekema, The Arabidopsis MALE STERILITY 2

protein shares similarity with reductases in elongation/condensation

complexes, Plant J. 12 (1997) 615–623.

[52] D. Post-Beittenmiller, The cloned Eceriferum genes of Arabidopsis

and the corresponding Glossy genes in maize, Plant Physiol. Biochem.

36 (1998) 157–166.

[53] R. Kalscheuer, A. Steinbuchel, A novel bifunctional wax ester

synthase/acyl-CoA: diacylglycerol acyltransferase mediates wax ester

and triacylglycerol biosynthesis in Acinetobacter calcoaceticus ADP1,

J. Biol. Chem. 278 (2003) 8075–8082.

[54] P.E. Kolattukudy, Enzymatic synthesis of fatty alcohols in Brassica

oleracea, Arch. Biochem. Biophys. 142 (1971) 701–709.

[55] J.P. McNevin, W. Woodward, A. Hannoufa, K.A. Feldmann, B.

Lemieux, Isolation and characterization of eceriferum (cer) mutants

induced by T-DNA insertions in Arabidopsis thaliana, Genome 36

(1993) 610–618.

[56] M.G.M. Aarts, C.J. Keijzer, W.J. Stiekema, A. Pereira, Molecular

characterization of the CER1 gene of Arabidopsis involved in

epicuticular wax biosynthesis and pollen fertility, Plant Cell 7 (1995)

2115–2127.

[57] J.D. Hansen, J. Pyee, Y. Xia, T.-J. Wen, D.S. Robertson, P.E.

Kolattukudy, B.J. Nikolau, P.S. Schnable, The glossy1 locus of maize

and an epidermis-specific cDNA from Kleinia odora define a class of

receptor-like proteins required for the normal accumulation of

cuticular waxes, Plant Physiol. 113 (1997) 1091–1100.

[58] X. Chen, S.M. Goodwin, V.L. Boroff, X. Liu, M.A. Jenks, Cloning

and characterization of the WAX2 gene of Arabidopsis involved in

cuticle membrane and wax production, Plant Cell 15 (2003)

1170–1185.

[59] T. Kurata, C. Kawabata-Awai, E. Sakuradani, S. Shimizu, K.

Okada, T. Wada, The YORE-YORE gene regulates multiple aspects

of epidermal cell differentiation in Arabidopsis, Plant J. 36 (2003)

55–66.

[60] Y. Xia, B.J. Nikolau, P.S. Schnable, Cloning and characterization of

CER2, an Arabidopsis gene that affects cuticular wax accumulation,

Plant Cell 8 (1996) 1291–1304.

[61] V. Negruk, P. Yang, M. Subramanian, J.P. McNevin, B. Lemieux,

Molecular cloning and characterization of the CER2 gene of

Arabidopsis thaliana, Plant J. 9 (1996) 137–145.

[62] Y. Xia, B.J. Nikolau, P.S. Schnable, Developmental and hormonal

regulation of the Arabidopsis CER2 gene that codes for a nuclear-

localized protein required for the normal accumulation of cuticular

waxes, Plant Physiol. 115 (1997) 925–937.

[63] B. St-Pierre, P. Laflamme, A.M. Alarco, V. De Luca, The terminal O-

acetyltransferase involved in vindoline biosynthesis defines a new

class of proteins responsible for coenzyme A-dependent acyl transfer,

Plant J. 14 (1998) 703–713.

[64] A. Hannoufa, V. Negruk, G. Eisner, B. Lemieux, The CER3 gene of

Arabidopsis thaliana is expressed in leaves, stems, roots, flowers and

apical meristems, Plant J. 10 (1996) 459–467.

[65] P.A. Rea, Z.-S. Li, Y.-P. Lu, Y.M. Drozdowicz, E. Martinoia, From

vacuolar GS-X pumps to multispecific ABC transporters, Annu. Rev.

Plant Physiol. Plant Mol. Biol. 49 (1998) 727–760.

[66] R. Sanchez-Fernandez, T.G. Emyr Davies, J.O.D. Coleman, P.A. Rea,

The Arabidopsis thaliana ABC protein superfamily, a complete

inventory, J. Biol. Chem. 276 (2001) 30231–30244.

[67] V. Arondel, C. Vergnolle, C. Cantrel, J.C. Kader, Lipid transfer

proteins are encoded by a small multigene family in Arabidopsis

thaliana, Plant Sci. 157 (2000) 1–12.

[68] D.P. Ma, H. Tan, Y. Si, R.G. Creech, J.N. Jenkins, Differential

expression of a lipid transfer protein gene in cotton fiber, Biochim.

Biophys. Acta 1257 (1995) 81–84.

[69] J. Pyee, P.E. Kolattukudy, The gene for the major cuticular wax-

associated protein and three homologous genes from broccoli

(Brassica oleracea) and their expression patterns, Plant J. 7 (1995)

49–59.

[70] M.B. Trevino, M.A. OConnell, Three drought-responsive members of

the nonspecific lipid-transfer protein gene family in Lycopersicon

pennellii show different developmental patterns of expression, Plant

Physiol. 116 (1998) 1461–1468.

[71] A.K. Sohal, J.A. Pallas, G.I. Jenkins, The promoter of a Brassica

napus lipid transfer protein gene is active in a range of tissues and

stimulated by light and viral infection in transgenic Arabidopsis, Plant

Mol. Biol. 41 (1999) 75–87.

[72] B. Hollenbach, L. Schreiber, W. Hartung, K.J. Dietz, Cadmium leads

to stimulated expression of the lipid transfer protein genes in barley:

implications for the involvement of lipid transfer proteins in wax

assembly, Planta 203 (1997) 9–19.

[73] J.C. Kader, Lipid-transfer proteins in plants, Annu. Rev. Plant Phisiol.

Plant. Mol. Biol. 47 (1996) 627–654.

[74] K. Okubo, N. Hori, R. Matoba, T. Niiyama, A. Fukushima, Y. Kojima,

K. Matsubara, Large scale cDNA sequencing for analysis of

quantitative and qualitative aspects of gene expression, Nat. Genet. 2

(1992) 173–179.