Comparative transcriptional profiling of orange fruit in response to the biocontrol yeast Kloeckera apiculata and its active compounds
Global changes in citrus gene expression profiles
To obtain an overall picture of the gene regulation, biocontrol yeast, the ether extract
and active compound were used to treat citrus and two independent microarray analyses
were performed for each treatment. To reduce experimental variation, two sets of six
fruit exocarps were harvested from each treated and untreated (water control) fruit.
After the removal of low-quality and internal reference probes, a total of 20,083
transcripts were reliably detected in the microarray analysis. Microarray analysis
gene changes in citrus exocarp revealed as many as 801, 339 and 608 differentially
expressed genes (DEGs) that showed a significant (P 0.05) change in expression (?1.3-fold) after 24 h of incubation with K. apiculata, the extract and PEA (Fig. 1). We further analysed these genes in subsequent experiments.
Fig. 1. Number of differentially expressed genes (DEGs) in citrus after statistical analysis.
Venn diagram shows the number of up-regulated and down-regulated genes that are expressed
in common or in special between K. apiculata, 2-phenylethanol (PEA) and the extract treatment
All DGEs were aligned against the Arabidopsis database by using the Citrus HarvEST software, and detailed descriptions of the sequences
are shown in Additional file 1. GO categories were assigned to the 1052 DEGs using the Blast2GO program (http://www.blast2go.org). The DEGs were categorized into 22 groups based on their biological processes, as
shown in Fig. 2. Response to stimulus (267), metabolic process (495), cellular process (501), pigmentation
(208), biological regulation (225), multicellular organismal process (142) and developmental
process (135) were the major categories. Categories based on the cellular component
revealed that the responsive genes were mainly related to cell (564), cell part (564),
organelle (440) and organelle part (159). With regard to molecular function, the DEGs
were classified as catalytic activity (385), binding (470), transcription regulator
activity (56), transporter activity (53), molecular transducer activity (40), electron
carrier activity (36), enzyme regulator activity (19), antioxidant activity (11),
structural molecule activity (9), translation regulator activity (5) and nutrient
reservoir activity (2). Of these, the plastid and intracellular organelle were the
major sub-cellular organelles involved in the citrus response.
Fig. 2. Functional categorization of global pattern of gene expression in citrus in response
to different treatment based on GO annotation
The responsive genes were further assessed using KEGG pathway analysis (http://www.genome.jp/kegg/) (Additional file 1). The most represented pathways are phenylpropanoid biosynthesis (26), limonene and
pinene degradation (18), ABC transporters (14), proteasome (3), lysosome (5), oxidative
phosphorylation (9), flavonoid biosynthesis (20), the regulation of autophagy (3),
calcium signalling pathway (13), apoptosis (18), fatty acid metabolism (7), MAPK signalling
pathway (25), phenylalanine, tyrosine and tryptophan biosynthesis (7), citrate cycle
(TCA cycle) (5), flavone and flavonol biosynthesis (14), starch and sucrose metabolism
(13), arachidonic acid metabolism (4), phenylalanine metabolism (7), ascorbate and
aldarate metabolism (6) and carotenoid biosynthesis (5). Most of these pathways were
consistent with biological processes that were already identified by GO analysis.
Some of these pathways were related to the defence response based on previous knowledge,
such as phenylpropanoid biosynthesis and the calcium signalling pathway 26], 31], 32], 36].
Change pattern of gene expression in citrus in response to K. apiculata
Of the 801 DEGs in orange exocarp tissue treatment with K. apiculata, 56 % of the annotated genes were down-regulated and 44 % were up-regulated. Furthermore,
the microarray data for the probes of the significant dataset were mapped to Arabidopsis using the MapManBin software (http://ppdb.tc.cornell.edu/dbsearch/searchacc.aspx). The data obtained from this analysis are presented in Fig. 3 and Additional file 2. Major and minor CHO metabolism (10), glycolysis (3), fermentation (1), TCA (3),
mitochondrial electron transport/ATP synthesis (3), lipid metabolism (17), amino acid
metabolism (12), redox (6), nucleotide metabolism (9), DNA (10) and proteins (82)
associated with the cell (11) showed more down-regulated than up-regulated genes,
while more up-regulated genes were found in PS (5), N-metabolism (2), hormone metabolism
(17) and secondary metabolism (14) associated with the cell wall (9) in response to
K. apiculata treatment. A number of new genes that are potentially related to defence responses
were identified in this study. Based on microarray and previous data 26], 31], 32], 36], hormone, reactive oxygen species (ROS), lipid, secondary metabolite, cell wall,
stress, phenyalanine metabolism related genes were selected for further analysis.
Figure 4 and Additional file 3 summarize the changes in these defence-related genes.
Fig. 3. MapManBin analyses of the common or in special up-regulated and down-regulated genes
in citrus between K. apiculata, PEA and the extract treatment
Fig. 4. Cluster analysis of the expression profiles of resistance-related differentially expressed
genes in citrus by MeV (http://www.tm4.org/mev.html). Each column represents a sample, and each row represents a single gene. The diagram
was generated using log
2
-transformed ratio values, and colours indicate relative signal intensities. Genes
down-regulated in the treatment compared to control are depicted in green, and up-regulated
genes are depicted in red
The first noticeable pathway is the hormone metabolism pathway. In total, 17 differentially
expressed genes are involved in hormone metabolism, including the ethylene (ET)–signalling pathway of eight ethylene response factor (ERF) genes (Cit.18086.1.S1_at, Cit.22763.1.S1_s_at, Cit.2675.1.S1_s_at, Cit.2677.1.S1_at,
Cit.17142.1.S1_s_at, Cit.18673.1.S1_at, Cit.20640.1.S1_at, Cit.16845.1.S1_at); jasmonic
acid (JA)-signalling pathway of one hydroperoxide lyase (HPL) (Cit.10444.1.S1_at) and five allene oxide synthase (AOS) genes (Cit.905.1.S1_at, Cit.6011.1.S1_at, Cit.23585.1.S1_at, Cit.31140.1.S1_at, Cit.996.1.S1_s_at)
and five abscisic acid (ABA)-signalling pathway genes (Cit.13424.1.S1_at, Cit.5225.1.S1_at,
Cit.10675.1.S1_at, Cit.13166.1.S1_at, Cit.34429.1.S1_s_at) (Fig. 4 and Additional file 3). Most of these genes were highly expressed in yeast-treated orange tissue, such
as ERF (Cit.2677.1.S1_at), which was up-regulated 4.0-fold according to the microarray data.
In addition, two polyamine (polyamine oxidase), three auxin-responsive and five gibberellic
acid (GA; gibberellin receptor, gibberellin oxidase) genes were down-regulated.
Reactive oxygen species (ROS) accumulation has been well studied for biocontrol yeast-induced
defence responses in fruits 15]–18]. The second group of metabolic pathways is involved in the redox and antioxidation
pathway. In total, five genes involved in antioxidant biosynthesis were down-regulated,
including monodehydroascorbate reductase (Cit.3320.1.S1_s_at, Cit.3318.1.S1_at), superoxide
dismutase (SOD, Cit.5267.1.S1_at), catalase (CAT, Cit.8351.1.S1_s_at) and peroxidase
(POD, Cit.8515.1.S1_s_at) (Fig. 4 and Additional file 3). For example, SOD (Cit.5267.1.S1_at) was down-regulated 5.5-fold according to the
microarray data, and the qRT-PCR data were consistent with these results, demonstrating
that the level of SOD was 2.2-times lower in response to K. apiculata-treatment than in CK. Moreover, Cytochrome P450 plays an important role in the redox
pathway and has been well characterized 43]–45]. Over nine different cytochrome P450 genes were detected in our microarray data,
such as monooxygenase/p-coumarate 3-hydroxylase, monooxygenase 83B1 and ent-kaurenoate
oxidase.
The third group of metabolic pathways consists of signalling pathway and pathogenesis-related
(PR) proteins. The signalling pathway genes included 13 genes for calcium and 25 genes
for MAPK signalling, most of which were up-regulated in response to K. apiculata application. The other genes encoding for chitinase (Cit.15242.1.S1_at, 1.8-fold)
and ?-1,3-glucanase (Cit.10558.1.S1_s_at, 1.4-fold) were stimulated by K. apiculata application (Fig. 4 and Additional file 3). In addition, five different disease resistance protein genes (TIR-NBS-LRR class)
were also found in our study.
The fourth group of significant K. apiculata-responsive genes included secondary metabolic processes, such as the phenylpropanoid
pathway, limonene and pinene degradation, flavone and flavonol biosynthesis and carotenoid
biosynthesis (Fig. 3). These genes were induced following K. apiculata application. A significant increase in the expression of the genes encoding for chalcone-flavanone
isomerase (Cit.17011.1.S1_s_at), cinnamoyl-CoA reductase (Cit.13313.1.S1_s_at), violaxanthin
de-epoxidase (Cit.30844.1.S1_s_at) and shikimate 5-dehydrogenase (Cit.25466.1.S1_at)
was observed. These genes are mainly involved in to lignin and flavanol biosynthesis.
Moreover, genes involved in the biosynthesis of the other secondary metabolites, such
as carotenoid and terpenes, were down-regulated, including p-coumarate 3-hydroxylase (Cit.30567.1.S1_at), 3-chloroallyl aldehyde dehydrogenase
(Cit.30574.1.S1_s_at), ent-kaurenoate oxidase (Cit.13587.1.S1_at) and carotenoid isomerase
(Cit.29769.1.S1_s_at).
Several families of transcription factors, including the WRKY, R2R3-MYB, bHLH (basic
helix-loop-helix) and WD40 genes, showed significant transcriptional changes in response
to K. apiculata application, as revealed by the microarray data.
Comparative analysis of gene expression in citrus between different treatment
To further analyze the response of the orange exocarp tissue to the extract and PEA,
a total of 339 and 608 genes were identified. The expression profiles in response
to the extract and PEA were similar to that found for K. apiculata; 57.4 % of the 803 DEGs in response to K. apiculata were also altered in the extract/PEA treatments (Fig. 1). The common up-regulated and down-regulated genes in citrus between K. apiculata-PEA and special in K. apiculata were listed in Additional file 4. The distribution of the genes among the GO and KEGG functional categories indicated
that a large number of defence-related genes were also included in these DEGs. Orange
exocarp tissue responded similarly to K. apiculata, the extract and PEA (Figs. 3 and 4).
Verification of microarray data by qRT-PCR analyses
To confirm that the DEGs identified by the microarray gene expression were indeed
differentially expressed, 20 genes were selected based on their biological significance
for confirmation in a biologically independent experiment using qRT-PCR, including
SOD-, ET-, JA- and ABA-related genes, which were detected in the microarray data and
bioinformatic analyses. The relative transcript abundance patterns for the stress
application were compared using the transcriptome data. The results of the qRT-PCR
experiments revealed that most of the genes showed the same expression pattern as
the microarray data (Fig. 5), such as Cit.2677.1.S1_at, which was 3.2-times higher in response to K. apiculata-treatment than in CK in qRT-PCR data, and 4.0-times higher for the microarray data.
Fig. 5. Verification of the microarray results by qRT-PCR. Black bar: qRT-PCR results for
the genes. Grey bar: microarray data for the genes. Each qRT-PCR reaction was carried
out in triplicate for three repeats. Columns and bars represent the means and standard
error (n = 3) respectively
Pathogenesis-related (PR) proteins activity
Of the PR proteins, chitinase and ?-1,3-glucanase are two of the most fully characterized
enzymes that are capable of hydrolyzing the polymers of fungal cell walls 27] Furthermore, the accumulation of chitinase and ?-1,3-glucanase is important in retarding
fungal growth and decreasing the spoilage of fruits caused by fungal pathogens. The
level of chitinase and ?-1,3-glucanase was observed to be significantly higher after
inoculation of fruit with yeast, the extract and PEA compared with the control (Fig. 6ab), which is consistent with the microarray data. In addition, some PR-proteins could
act in cell wall reinforcement by catalyzing lignification, such as PR-9 46]. As illustrated in Fig. 7, the application of K. apiculata and the extract to discs of orange peel increased the lignin content in fruit peel
relative to the water control peel discs 60 h after application.
Fig. 6. PR protein activity and polyamine and H
2
O
2
content between K. apiculata, PEA and the extract treatment. a Chitinase activity; (b): ?-1,3-glucanase activity; (c): H
2
O
2
content; (d–f): polyamines content. The results in all the histograms are expressed as means ±
standard errors. Mean values for different treatments at each time point are labelled
with different letters to indicate significant differences at the level P 0.05 according to Duncan’s multiple range test
Fig. 7. Lignin content between K. apiculata, PEA and the extract treatment. The results in all the histograms are expressed as
means ± standard errors. Mean values for different treatments at each time point are
labelled with different letters to indicate significant differences at the level P 0.05 according to Duncan’s multiple range test
Hydrogen peroxide (H
2
O
2
) level in orange tissue
Reactive oxygen species (ROS) burst has been shown to regulate the yeast response
processes 15]. Our study showed that K. apiculata treatment resulted in a high level of intracellular H
2
O
2
when applied to oranges (Fig. 6c); this level decreased dramatically 12 h after the application of yeast to citrus
fruit, although the statistic analysis showed that there were significant differences
(P 0.05) between K. apiculata-treatment and control at the point of 24 h, 36 h and 48 h. The extract and PEA did
not enhance the level of H
2
O
2
.
Polyamine level in orange tissue
Polyamines, mainly diamine putrescine (Put), triamine spermidine (Spd) and tetraamine
spermine (Spm), act as an important source of H
2
O
2
production and have been suggested to be involved in the response to pathogen attack
or responsible for enhanced disease resistance in higher plants 47]. In yeast-treated citrus, the level of Put, Spd and Spm were observed to be lower
than normal control, especially at 24 h (Fig. 6). In the extract and PEA-treated citrus, there was not a rule can be followed for
Spm, Put and Spd (Fig. 6d, e, f).