Dong 2015

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

Mol Genet Genomics

DOI 10.1007/s00438-015-1123-6

ORIGINAL ARTICLE

GDSL esterase/lipase genes in Brassica rapa L.: genome‑wide


identification and expression analysis
Xiangshu Dong1,4 · Hankuil Yi1 · Ching‑Tack Han2 · Ill‑Sup Nou3 · Yoonkang Hur1

Received: 8 June 2015 / Accepted: 18 September 2015


© Springer-Verlag Berlin Heidelberg 2015

Abstract GDSL esterase/lipase proteins (GELPs), a very expression profiling revealed that BrGELPs were expressed
large subfamily of lipolytic enzymes, have been identified in various tissues, including floral organs, implying that
in microbes and many plants, but only a few have been BrGELPs play diverse roles in various tissues and during
characterized with respect to their roles in growth, develop- development. Ten percent of BrGELPs were specifically
ment, and stress responses. In Brassica crops, as in many expressed in fertile buds, rather than male-sterile buds,
other species, genome-wide systematic analysis and func- implying their involvement in pollen development. Analy-
tional studies of these genes are still lacking. As a first step ses of EXL6 (extracellular lipase 6) expression and its co-
to study their function in B. rapa ssp. pekinensis (Chinese expressed genes in both B. rapa and Arabidopsis, as well
cabbage), we comprehensively identified all GELP genes as knockdown of this gene in Arabidopsis, revealed that
in the genome. We found a total of 121 Brassica rapa this gene plays an important role in pollen development in
GDSL esterase/lipase protein genes (BrGELPs), forming both species. The data described in this study will facilitate
three clades in the phylogenetic analysis (two major and future investigations of other BrGELP functions.
one minor), with an asymmetrical chromosomal distribu-
tion. Most BrGELPs possess four strictly conserved resi- Keywords BrGELP subfamily · SGNH hydrolase ·
dues (Ser-Gly-Asn-His) in four separate conserved regions, Extracellular lipase 6 · Brassica rapa · Pollen development
along with short conserved and clade-specific blocks, sug-
gesting functional diversification of these proteins. Detailed
Introduction

Communicated by S. Hohmann. GDSL esterases/lipase proteins (GELPs) usually have a


Ser-His-Asp active site, with the serine-containing motif
Electronic supplementary material The online version of this located closer to the N terminus (Brick et al. 1995; Upton
article (doi:10.1007/s00438-015-1123-6) contains supplementary
material, which is available to authorized users.
and Buckley 1995). GELPs are an important subfam-
ily of lipolytic enzymes with various biochemical activi-
* Yoonkang Hur ties (Akoh et al. 2004; Leščić Ašler et al. 2010). They are
ykhur@cnu.ac.kr also referred to as SGNH hydrolases in reference to the
1 conserved residues Ser-Gly-Asn-His present in four con-
Department of Biological Sciences, College of Biological
Science and Biotechnology, Chungnam National University, served blocks (I, II, III, and V) (Brick et al. 1995; Upton
Taejon, Republic of Korea and Buckley 1995; Mølgaard et al. 2000; Akoh et al. 2004).
2
Department of Life Science, Sogang University, The GELP genes can be found throughout all kingdoms
Seoul 100‑611, Republic of Korea of life. The characteristics and features of GELP genes
3
Department of Horticulture, Sunchon National University, have been analyzed in plants at the genomic level in Arabi-
Suncheon, Jeonnam, Republic of Korea dopsis, Sorghum bicolour, Populus tricholarpa, Oryza
4
School of Agriculture, Yunnan University, Kunming, Yunnan, sativa, maize and others (Ling 2008; Volokita et al. 2011;
China Chepyshko et al. 2012). Among them, several GELP genes

13
Mol Genet Genomics

in Arabidopsis have been cloned and characterized (Oh stratified for 3 days at 4 °C, and then placed in the growth
et al. 2005; Zhang et al. 2006; Agee et al. 2010). Despite room. The plants were then kept under a lid of transpar-
the fact that they play important roles in development, mor- ent polythene for 5–6 days to maintain high humidity and
phogenesis, lipid metabolism, abiotic stress responses, and support simultaneous germination. For in vitro culture,
pathogen defense (Hong et al. 2008; Kim et al. 2008; Kwon seeds were sterilized with 30 % bleach with 0.1 % Triton
et al. 2009; Lee et al. 2009; Updegraff et al. 2009; Chepy- X-100 (Sigma, USA), stratified for 3 days at 4 °C, and
shko et al. 2012; Tan et al. 2014), only a small fraction of sown in square dishes (100 mm × 100 mm × 20 mm)
plant GELPs have been characterized, and to date no sys- containing half-strength MS media (Duchefa Bioche-
tematic study of Brassicaceae GELPs has been reported. mie, The Netherlands) supplemented with 1 % sucrose
The pollen coat protects pollen grains from excess desic- and 0.8 % PhytoAgar. For B. rapa, seeds were sown in
cation after dehiscence, and also plays a key role in medi- 55 mm × 55 mm × 90 mm pots containing potting soil,
ating early contact between pollen grains and the stigma, and 3-week-old seedlings were transferred into a 4 °C cold
thus contributing to pollen adhesion to the stigma and room for vernalization to induce flowering. After 60 days
facilitating hydration (Edlund et al. 2004; Updegraff et al. of vernalization, the plants were returned to the 22 °C
2009). The pollen coat consists of a lipid-rich compound growth room.
that is synthesized by tapetal cells during pollen mitotic
divisions (Quilichini et al. 2014). The oleosin domain pro- Identification of GELP genes in B. rapa
tein GRP17, the first pollen coat protein to be isolated, is
required for rapid hydration of pollen grains after landing The full genome sequence of B. rapa was downloaded
on the stigma (Mayfield and Preuss 2000). Besides oleo- from BRAD (http://brassicadb.org/brad/; V1.5) (Wang
sin domain-containing proteins, pollen coat proteins con- et al. 2011), and 3537 novel genes recently identified by
sist primarily of GDSL lipases/extracellular lipases (EXLs) deep RNA-Seq were also added for analysis (Devisetty
(Mayfield et al. 2001). Arabidopsis has six EXLs (EXL et al. 2014). Protein sequences of these genes were used as
1–6), which are clustered in the genome and all belong queries to search against the profile Hidden Markov Model
to GELPs subfamily; two of them, EXL4 and EXL6, are (HMM) (version 3.1b2 with Pfam HMM library Pfam
enriched in the pollen coat (Mayfield et al. 2001). EXL4 is 28.0) (Finn et al. 2011). In total, 153 protein sequences
required for efficient hydration of Arabidopsis pollen, and with PF00657 (E value below 0.1) were obtained. Among
loss-of-function mutation in this gene causes a delay in pol- them, only 124 proteins contain the catalytic triad serine,
len hydration (Updegraff et al. 2009). The function and role aspartate, and histidine in the GDSL and WDXXH motifs
of EXL6 have not been well characterized, although the found at the N- and C- termini. The genes without cata-
gene is known to be a target of a key transcription factor lytic triad were considered as the pseudoenzymes (Table
involved in Arabidopsis pollen development, ABORTED S1). Subsequently, three genes Bra016151, Bra017133, and
MICROSPORES (AMS) (Xu et al. 2014). Bra028825 were excluded from our analyses, because these
In this study, we systematically analyzed the phyloge- genes also have other protein domain(s) besides PF00657
netic relationships and expression patterns of Brassica rapa (Table S1). As in B. rapa, the GELP genes in Arabidop-
GELPs. In addition, we found that knockdown of EXL6 sis were identified using protein sequences from TAIR 10
in Arabidopsis led to a male-sterile phenotype. Further- (https://www.arabidopsis.org/). A total of 112 genes encod-
more, co-expression analysis suggested that EXL6 plays ing GELPs were identified from A. thaliana after HMM
a role during pollen exine formation in both B. rapa and search. Only 98 genes were found to contain the catalytic
Arabidopsis. triad (Table S2).

Sequence alignment and phylogenetic analysis


Materials and methods
Sequence alignment and phylogenetic analysis were car-
Plant materials and growth condition ried out based on previously described criteria (Volokita
et al. 2011). Full protein sequences were aligned using
Arabidopsis thaliana (L.) Heynh var. Columbia (Col- ClustalW, implemented in MEGA6 (Tamura et al. 2013),
0) was used for transformation. Both Chinese cabbage with GONNET as the protein weight matrix, a gap-open-
(Brassica rapa ssp. pekinensis, inbred line Chiifu) and ing penalty of 10, and a gap-extension penalty of 0.1.
Arabidopsis plants were grown at 22 ± 0.5 °C with a Phylogenetic trees were constructed using two methods,
light intensity of 140 ± 2 μmol/m2/s on a long-day cycle neighbor-joining (NJ) and maximum likelihood (ML).
(16 h L/8 h D). For Arabidopsis, seeds were sown in In the NJ method, the phylogenetic tree was constructed
60 mm × 60 mm × 55 mm pots containing potting soil, in MEGA6 using the Poisson model with the following

13
Mol Genet Genomics

options: homogeneous pattern, complete deletion, and Table 1. Semi-RT-PCR products were separated on 1.5 %
1000-replicate bootstrap. For ML method, the initial tree agarose gels and stained with ethidium bromide.
were generated by implementing NJ/BioNJ method, and
tree improvement was achieved by nearest-neighbor-inter- Pollen viability
change (NNI) and branch support analysis using 1000 rep-
licates bootstrap with Poisson model. For pollen viability and pollen development progression,
flowers collected from WT and EXL6 antisense- or sense-
Generation of EXL6 knockdown Arabidopsis transgenic Arabidopsis plants were fixed in Carnoy’s fixa-
tive (6:3:1 ethanol:chloroform:acetic acid) for 2 h. After
The full-length coding region of AtEXL6 was cloned from fixation, anthers were dissected and stained with a solution
first-strand cDNA using the AtEXL6-F primers (Table 1). of Malachite green, acid fuchsin, and Orange G for about
The amplified fragments were inserted into T&A cloning 12 h as previously described (Peterson et al. 2010).
vectors (RBC T&A Cloning Kit, Real Biotech Corporation,
Taiwan). After confirming the EXL6 sequence in the T&A Identification of genes co‑expressed with EXL6
vector by sequencing, the fragment was further cloned into
pCambina 3300-35S binary vectors in the sense and anti- AtEXL6 was used as a bait gene for genome-wide co-
sense directions, and transformed into plants to knockdown expression analysis using Expression Angler to identify
AtEXL6 expression. Arabidopsis thaliana Col-0 was trans- putative genes with functions similar to that of AtEXL6
formed with Agrobacterium tumefaciens GV3101 carrying (Toufighi et al. 2005). A cutoff threshold of 0.85 was used
the binary plasmids by the floral dip method (Peterson et al. for the Pearson Correlation Coefficient. Expression pattern
2010). The transformants were selected on plates contain- analysis was performed using the Arabidopsis eFP browser
ing 25 mg/ml glufosinate in MS medium (Sigma) and con- (http://bar.utoronto.ca/efp/cgi-bin/efpWeb.cgi) (Winter
firmed by genomic DNA PCR. et al. 2007). To isolate genes co-expressed with BrEXL6,
previously published microarray results were used as
Reverse transcription PCR and qRT‑PCR source data (Dong et al. 2013a, b). Three B. rapa probes,
Brapa_ESTC003525, Brapa_ESTC000535, and Brapa_
Various tissue organs, sampled from B. rapa and Arabi- ESTC010981, were used to identify the ESTs co-expressed
dopsis, were stored at −70 °C until use. RNA was iso- with BrEXL6. Identified ESTs were used to obtain the B.
lated using the TRIZOL® Reagent (Invitrogen, USA). One rapa gene identification number based on the BlastN result
microgram of total RNA from each sample was used in in the BRAD database (http://brassicadb.org/brad/ with cut-
reverse transcription. The first-strand cDNA was synthe- off below 1e−5). GO enrichment analysis was performed
sized using the ReverTra Ace-α kit (Toyobo, Osaka, Japan). using agriGO (http://bioinfo.cau.edu.cn/agriGO/index.php)
The concentration of synthesized cDNA was determined, (Du et al. 2010).
and the cDNA was diluted to 12.5 ng/μl for PCR. Semi-
quantitative RT-PCR (semi-RT-PCR) was generally per- Conserved motifs search
formed with a 5 min denaturation at 94 °C, followed by 25
or 30 cycles of 94 °C for 30 s, 54 °C for 30 s, and 72 °C for To identify conserved motifs in BrGELPs, the full-length
60 s. All primer sequences used in this study are listed in protein sequences of BrGELP genes were submitted to

Table 1  List of primers used in this study


Primer name Primer sequence Size (bp) Usage
Forward (5′–3′) Reverse (5′–3′)

AtExl6real AAACCATTCAGAAAAATTTTGGAC CTAAGACTGAATCACCAAAAGCAA 152 Real-time PCR


AtExl4real AAACATAAGAGAGTCTAAACACTGTGG GAATCTCCAAAAGCCAAAAGC 159 Real-time PCR
BrEXL6 CGGCGACTAGTCGAACACTGAG ACAACCTGCAGTTTTCGCTTCTG 520 Semi-RT-PCR
AtEXL6 GGACTTATTTTCGATATGTTTAGGGG TCTTCCCGTCGGAAATTTATAGTCA 222 Semi-RT-PCR
AtEXL6-F GGCGGACTTATTTTCGATATGTTTAG TTAGCTAGGCAAGGCCTTTCGCTAT 1047 FCDS, vector constructs
GGG CTG
35S CAATCCACTTGCTTTGAAGAC CCATCATTGCGATAAAGGAAAG 135 gDNA PCR for transgenic
plant confirmation
At Actin7 AGTGGTCGTACAACCGGTATTGT GAGGATAGCATGTGGAACTGAGAA 96 Internal control

13
Mol Genet Genomics

Fig. 1  Genomic distribution of the BrGELP genes and pseudoen- zymes (red color) present on each chromosome. Bra035295 (pseu-
zymes (red color) on Brassica rapa chromosomes. Chromosome doenzyme), Bra038972, Bra040203, and Bra040640, which are cur-
number is indicated at the top of each chromosome. Numbers in rently mapped to scaffolds, are not shown. The blue color indicate the
parentheses indicate the number of BrGELP genes and pseudoen- centromere positions (color figure online)

MEME (Bailey et al. 2009). For a single motif, follow- BrGELP genes were present on all 10 B. rapa chromo-
ing criteria were used: number of repetitions, 0 or 1 per somes, with a non-uniform distribution: except for three
sequence; maximum number of motifs, 100; optimum BrGELPs (Bra038972, Bra040203, and Bra040640), all
motif width, 6–12. Only motifs with E value ≤1e−10 were other BrGELP genes could be mapped on chromosomes
selected for further analysis. A01–A10 (Fig. 1). Four chromosomes contained more than
57 % of all BrGELPs in the genome: chromosomes A07
(17.4 %), A02 (14.0 %), A08 (14.0 %), and A09 (12.4 %)
Results (Supplementary Table 1; Fig. 1). The pericentromere region
of chromosome A07 and the short arm of chromosome
The B. rapa GDSL esterase/lipase genes A08 contained chromosomal regions with especially high
BrGELP density (Fig. 1).
To identify the members of BrGELP gene subfamily, the Gene structure analysis revealed that BrGELP genes
entire B. rapa genome sequence (V1.5) and 3537 novel consisted of five exons on average. Eighty-six BrGELP
transcripts recently identified by deep sequencing were genes (71.1 %) had five exons interrupted by four introns
downloaded from the BRAD database (Wang et al. 2011) (Supplementary Table 1, Supplementary Fig. 1), as revealed
and UC Davis Phytonetworks Genome Browser (http:// by a global analysis of gene structure in the B. rapa
phytonetworks.ucdavis.edu/) (Devisetty et al. 2014), genome (5.03 exons per gene) (Wang et al. 2011); similar
respectively. After HMM search, 121 BrGELP genes and results were obtained from the analysis of OsGELP genes
29 BrGELP pseudoenzymes (missing one or two residues in rice (Chepyshko et al. 2012). The rest of the genes had
of the catalytic triad) were isolated. Three genes contain- 1 exon (1 gene, Bra009496), 2 exons (2 genes), 3 exons (9
ing other protein domain(s) in addition to PF00657 were genes), 4 exons (14 genes), 6 exons (6 genes), or 10 exons
not used in our BrGELP candidate genes analysis (Supple- (3 genes) exons.
mentary Table 1).The locus ID, genome location, coding
sequence (CDS) length, and protein length of the resulting Phylogenetic analysis, motif and orthologs identification
genes are listed in Supplementary Table 1. The CDS sizes of the BrGELP genes
of the BrGELP genes varied from 966 bp (Bra039701) to
2673 bp (Bra025794), with the average sequence length To obtain insight into the evolutionary relationship of the
of 1142 bp. The genomic sequences of the BrGELP genes BrGELP genes, we carried out a phylogenetic analysis
ranged from 1104 to 11,289 bp (Supplementary Table 1). of the BrGELP and A. thaliana GELP (AtGELP) genes.

13
Mol Genet Genomics

We also analyzed the Arabidopsis genome (TAIR10) as two major and one minor clade, consistent with our results
described above for Chinese cabbage genome. After HMM (Volokita et al. 2011).
search, 112 genes, including 98 GELP genes, 13 GELP Conserved sequences and motifs imply important
pseudoenzymes, and one gene with multi-protein domains, roles in the functions of enzymes. A total of 55 motifs (E
were obtained (Table S2). The unrooted phylogenetic tree value ≤1e−10) were identified from the 121 BrGELP pro-
was constructed for 121 BrGELPs and 98 AtGELPs using teins (Supplementary Table 3). Our analysis of conserved
the NJ method in MEGA6. The sequence alignment is motifs revealed that blocks I, II, III, and V were found in
shown in Supplementary Fig. 2. In summary, the GELP all BrGELPs, and each block was included in different
genes from B. rapa and Arabidopsis could be divided into motifs (2, 4, 8, and 1, respectively) (Fig. 3; Supplementary
three clades (I–III) (Fig. 2); Clade I and II consisted of 13 Table 3). Beside these four conserved blocks, additional 17
and 4 subclades, respectively. The phylogenetic tree gener- well-conserved motifs (with E values below 1e−100) were
ated with ML methods also gave the same overall pattern also detected in most BrGELP genes. The other motifs
except some branch rotation (Supplementary Fig. 3). Pre- were specific to individual subclades in the phylogenetic
vious analysis for GELP genes from land plants revealed tree (Supplementary Table S3). For instance, motifs 18, 21,

Fig. 2  Phylogenetic relationships of BrGELP and AtGELP gene ing options: homogeneous pattern, complete deletion, and 1000-rep-
family. Multiple protein sequence alignment was performed using licate bootstrap. Different clades and branches (subclades) are indi-
ClustalW implemented in MEGA6. The neighbor-joining (NJ) tree cated by the colors of the background and branch lines, respectively
was constructed in MEGA6 using the Poisson model with the follow- (color figure online)

13
Mol Genet Genomics

22, 23, 34, 38, 43, 47, 48, 51, 52, and 53 were only found Fig. 4  Heat map of expression patterns of BrGELP genes (complete▸
in clade I; motifs 25, 26, 27, 29, 30, 31, 32, 33, and 35 were catalytic triad) and pseudoenzymes (incomplete catalytic triad) based
on probe intensity (PI) (extracted from Dong et al. 2013a). S1–S3
specifically detected in clade II (Supplementary Table S3). indicate floral buds from male-sterile B. rapa before the tetrad stage
Because orthologous genes frequently retain similar (S1), at the tetrad stage (S2), and containing aberrant pollen grains
functions (Koonin 2005), identification of orthologous (S3). F1–4 represent fertile B. rapa floral buds before the tetrad stage
GELP genes between Chinese cabbage and Arabidopsis (F1), at the tetrad stage (F2), after the tetrad stage but before contain-
ing mature pollen (F3), and containing mature pollen (F4). Genes in
may allow us to predict the function of specific BrGELP the box are putatively involved in the pollen development
genes. Total 69 orthologous groups were identified using
InParanoid (O’Brien et al. 2005) (Supplementary Table 4).
Expression patterns of BrGELP genes
Among the orthologous gene groups of B. rapa and Arabi-
dopsis, a maximum of three B. rapa orthologs were sug-
The expression pattern of BrGELP genes can provide addi-
gested for each Arabidopsis gene, consistent with the
tional insight into their functions. To determine the expres-
notion that genome triplication occurred in Chinese cab-
sion patterns of BrGELP genes in different tissues, we used
bage after divergence of these two species from their com-
Illumina RNA-Seq transcriptomic data (Cheng et al. 2012).
mon ancestor. For example, the three B. rapa genes GDSL-
Of the 121 genes, 87 genes had significant RPKM (Reads
MOTIF LIPASE 3 (GLIP3: Bra030917, Bra030916, and
Per Kilobase per Million mapped reads) in roots, stems, or
Bra014375) were assigned as orthologs for AT1G53990.
leaves, and were expressed in these tissues in a preferential
For most AtGELP genes, however, only one or two
or specific manner (Supplementary Fig. 4A). We also deter-
BrGELP genes were identified, indicating gene losses dur-
mined the expression patterns of the BrGELP pseudoen-
ing speciation in the Brassica lineage. A previous study
zymes and observed that 20 of the 29 pseudoenzymes had
revealed that extracellular lipases (EXLs) 4, 5, and 6 are
significant RPKM in roots, stems, or leaves (Supplementary
involved in pollen coat development in Arabidopsis (May-
Fig. 4B). For expression patterns in flowers, we used previ-
field et al. 2001). In our analysis of the orthologous rela-
ously published microarray data (Dong et al. 2013a). Among
tionships between Arabidopsis and B. rapa, Bra008229 and
the 121 BrGELPs, 65 genes represented by 107 probes were
Bra103243 were identified as the orthologs of AtEXL4 and
expressed in different stages of floral buds (Fig. 4). Expres-
AtEXL6, respectively, suggesting that they are also involved
sion patterns could be grouped into three types: constitutive,
in pollen development in B. rapa.

Fig. 3  Motif logo of four conserved blocks found in BrGELP proteins: I (a), II (b), III (c), and V (d). Conserved residues Ser-Gly-Asn-His in
blocks are indicated with red triangles (color figure online)

13
Mol Genet Genomics

13
Mol Genet Genomics

stage-specific during flower development, and fertile bud- overrepresented among genes co-expressed with AtEXL6
specific. BrEXL4 (Bra008229, containing complete cata- and BrEXL6 (Fig. 5c, d). Taken together, our analyses sug-
lytic triad, and Bra014533, a pseudoenzymes without gested that EXL6 may be required for proper pollen devel-
Serine residue) and BrEXL6 (Bra103243 for three ESTs), opment, especially pollen exine formation, in both B. rapa
the predicted orthologs of AtEXL4 and AtEXL6, respec- and Arabidopsis.
tively, were specifically expressed in fertile floral buds of To confirm the expression patterns, we validated
B. rapa. In addition, 14 BrGELPs genes (27 probes) were BrEXL6 and AtEXL6 expression by RT-PCR (Fig. 6).
also expressed in a fertile bud-specific manner. Among AtEXL6 was specifically expressed before floral stage
these, expression patterns for two orthologous BrGELPs for 12, and BrEXL6 was highly expressed in fertile bud 3.
each Arabidopsis gene were supported by multiple probes: The RT-PCR results were consistent with our expression
Bra016475 (Brapa_ESTC001010, Brapa_ESTC010869, pattern analysis (Fig. 5a, b), which revealed specific or
Brapa_ESTC047743) and Bra025794 (Brapa_ESTC000842, high expression in floral buds at the later stage. To eluci-
Brapa_ESTC003502, Brapa_ESTC011093) for AT1G20132; date gene function, we generated knockdown mutants of
and Bra017361 (Brapa_ESTC009337, Brapa_ESTC019325, AtEXL6 by antisense expression and sense co-suppression
Brapa_ESTC030427) and Bra026529 (Brapa_ESTC009284, driven by the CaMV35S promoter (Fig. 7a). In total, four
Brapa_ESTC025896, Brapa_ESTC026051) for AT2G03980. independent transgenic plants (out of 26 plants obtained)
These results suggest that a large number of BrGELP genes, exhibited down-regulation of AtEXL6 and shared similar
including BrEXL6, might also be involved in pollen develop- defects in pollen grains (Fig. 7, Supplementary Fig. 5A).
ment in B. rapa. However, these transgenic plants exhibited normal veg-
etative growth relative to wild-type plants (Supplementary
Conserved function of GELPs in pollen development Fig. B). To explore whether the mutant phenotypes in the
in B. rapa and Arabidopsis transgenic plants correlated with AtEXL6 expression lev-
els, we performed semi-RT-PCR. As shown in Fig. 7b,
To investigate the function of Bra103243 (BrEXL6), one AtEXL6 expression was almost completely suppressed in
of the GELPs specifically expressed in fertile floral buds the sense-16 line, which exhibited severe pollen defective
but not in male-sterile flower buds, we carried out co- phenotypes. Compared to wild-type and sense-16 plants,
expression analysis using previously published microarray lines antisense-3 and antisense-9 (with moderate reduc-
data (Dong et al. 2013a, b). ESTs that were co-expressed tion in AtEXL6 expression) exhibited intermediate levels
with three probes for BrEXL6, Brapa_ESTC003525, of pollen development (Fig. 7b, Supplementary Fig. 5A).
Brapa_ESTC000535 and Brapa_ESTC010981, were con- These results indicate that sufficient expression of AtEXL6
sidered to be co-expressed ESTs. The B. rapa gene ID is required for normal pollen development in Arabidopsis.
was obtained for these ESTs using the BlastN service in To narrow down the stage at which EXL6 functions,
BRAD. To identify genes co-expressed with AtEXL6, we we monitored the pollen development process in WT and
used Expression Angler (http://bar.utoronto.ca/ntools/cgi- AtEXL6 transgenic plants, and observed significant differ-
bin/ntools_expression_angler.cgi) (Toufighi et al. 2005). ences after the tetrad stage. At this stage, cytoplasm deg-
When the threshold value for Pearson’s correlation coef- radation was observed in microspore pollen of transgenic
ficient (PCC) was set at 0.85, 36 B. rapa genes (70 ESTs) plants, whereas no similar phenotype was observed in WT
and 25 Arabidopsis genes were identified as co-expressed plants (Fig. 7c). Based on the results obtained from expres-
with BrEXL6 and AtEXL6, respectively (Supplementary sion analysis and transgenic knockdown of AtEXL6, we
Tables 5, 6). The expression patterns of these genes/ESTs conclude that AtEXL6 is involved in pollen development
were reconstructed using previously published micro- after the tetrad stage in Arabidopsis, and BrEXL6 may have
array data and the Arabidopsis eFP browser (http://bar. a similar function in Chinese cabbage.
utoronto.ca/efp/cgi-bin/efpWeb.cgi) (Winter et al. 2007;
Dong et al. 2013a). BrEXL6 and its co-expressed ESTs
were specifically expressed in fertile floral buds, with Discussion
the highest expression in floral buds, at and after the tet-
rad stage, but before containing mature pollen (Fig. 5a). Identification and analysis of BrGELPs
In Arabidopsis, the highest expression levels of AtEXL6
and its co-expressed genes were observed at floral stage Recently, genome-wide analyses have been performed
10–12, representing the microspore pollen development on several gene superfamilies in Brassica species; these
stage (Table 5B). GO enrichment analysis of the co- include B. rapa phytocyanins (BrPCs) (Li et al. 2013),
expressed genes using agriGo (Du et al. 2010) revealed AP2/ERF in B. rapa (Song et al. 2013) and B. olera-
that genes involved in pollen exine formation were highly cea (Thamilarasan et al. 2015), B. rapa non-specific

13
Mol Genet Genomics

Fig. 5  Expression pattern and GO enrichment analysis of EXL6s and and purple lines indicate expression patterns of three BrEXL6 ESTs,
their co-expressed genes. a Expression patterns of BrEXL6 and its Brapa_ESTC010981, Brapa_ESTC000535, and Brapa_ESTC003525.
co-expressed genes in sterile and fertile B. rapa floral buds, based on b Expression pattern of AtEXL6 and its co-expressed genes in various
previously published microarray data (Dong et al. 2013a). S1–3 indi- tissues of Arabidopsis, performed using the Arabidopsis eFP Browser
cate the floral buds from male-sterile B. rapa, before the tetrad stage (http://bar.utoronto.ca/efp/cgi-bin/efpWeb.cgi). FS stands for “flower
(S1), after the tetrad stage (S2), and containing aberrant pollen grains stage”. Red line indicates the expression pattern of AtEXL6. c, d GO
(S3). F1–4 represent fertile B. rapa floral buds before the tetrad stage enrichment analysis of genes co-expressed with BrEXL6 and AtEXL6,
(F1), at the tetrad stage (F2), after the tetrad stage but before contain- respectively (color figure online)
ing mature pollen (F3), and containing mature pollen (F4). Red, blue

lipid transfer proteins (BrnsLtp) (Li et al. 2014), B. rapa


expansin (BrEXP) (Krishnamurthy et al. 2014), and the
B. rapa MADS-box genes (Duan et al. 2015). Previously,
the analysis of GELP subfamily in plants has been reported
for Arabidopsis, rice and other land plants (Ling 2008;
Volokita et al. 2011; Chepyshko et al. 2012; Jiang et al.
2012). However, a detailed analysis of the GELP subfam-
ily in Brassica has not been reported to date. In this study,
we carried out genome-scale identification and expression
analysis of B. rapa GELPs, and characterized the function
of AtEXL6, a possible Arabidopsis ortholog of BrEXL6.
As in Arabidopsis, the 121 BrGELP genes identified here
Fig. 6  Expression levels of EXL6. a Analysis of BrEXL6 expression
in various tissues of B. rapa. Shoots were sampled from 3-week-old were asymmetrically distributed on chromosomes 1–10,
Chiifu seedlings. F1–4 represent fertile B. rapa floral buds before the with several densely populated regions on certain chromo-
tetrad stage (F1), at the tetrad stage (F2), after the tetrad stage but somes (Fig. 1) (Ling 2008). These genes can be phyloge-
before containing mature pollen (F3), and containing mature pollen netically grouped into two major and one minor subfamily
(F4). b Analysis of AtEXL6 expression in various tissues of wild-type
(Col-0) Arabidopsis using semi-RT-PCR. FS1–12 flower stage 1 to (Fig. 2), as in Arabidopsis and other species (Ling 2008;
stage 12, FS13–16 flower stage 13 to stage 16 Chepyshko et al. 2012; Volokita et al. 2011), consistent

13
Mol Genet Genomics

Fig. 7  Analysis of AtEXL6-knockdown transgenic plants. a Sche- b Expression levels of AtEXL6 in floral buds of WT and transgenic
matic representation of AtEXL6 gene structure and DNA regions used plants. c Comparisons of pollen development among WT and T1
for the antisense and sense constructs. White boxes indicate untrans- AtEXL6 transgenic plants. ‘Antisense’ and ‘sense’ indicate transgenic
lated regions (UTRs), gray boxes represent exons, lines connecting lines derived from antisense and sense constructs, respectively. The
boxes represent introns, and dashed arrows indicate the orientation of green arrows indicate the defects. The numbers after FS indicates flo-
antisense and sense fragments in the transgenic constructs. AtEXL6- ral development stages. Bar 20 μm (color figure online)
F and -R indicate the positions of primers used for semi-RT-PCR.

with a high degree of similarity in the genome structures and GLIP2) are in Clade II (subclade II-b) (Kwon et al.
of Arabidopsis and Brassica species (Krishnamurthy et al. 2009; Lee et al. 2009), and hot pepper pathogen- and stress-
2015; Thamilarasan et al. 2014). Among the GELPs, the responsive genes (CaGLIP1 and CaGL1) are in Clade I
well-characterized genes belong to Clade I and II (Fig. 2). (subclade I-m) (Hong et al.2008; Kim et al. 2008). BnLIP2,
For example, Arabidopsis bacterial resistance genes (GLIP1 involved in rapeseed germination, morphogenesis, and

13
Mol Genet Genomics

flowering, is in Clade II (subclade II-d) (Ling et al. 2006). Functional analysis of EXL6
By contrast, the pollen fertility-related genes EXL4–6
(Updegraff et al. 2009; Xu et al. 2014) are grouped in Clade EXL6 and its co-expressed genes in Arabidopsis and
I (subclade I-b), along with many other genes specifically B. rapa, which are enriched in pollen exine formation and
expressed in male-fertile floral buds (Fig. 2). lipase genes, are transcriptionally activated during the early
Compared with other fully sequenced plant genomes stage of pollen development and significantly downregu-
so far (Ling 2008; Volokita et al. 2011; Chepyshko et al. lated at the final stage of pollen maturation (Figs. 5, 6; Sup-
2012), the Chinese cabbage genome contains among the plementary Tables 5, 6) (Dong et al. 2013a). To elucidate
highest number of BrGELPs (121). The high number of the function of AtEXL6, for which no T-DNA knockout
GELP subfamily members in this vegetable crop could mutant lines were available in Arabidopsis, we generated
be related to functional diversification in order to use a knockdown mutants by over-expression of EXL6 sense and
wide range of substrates (Volokita et al. 2011), as well as antisense transcripts (Fig. 7, Supplementary Fig. 5). We
the genome triplication event in this lineage (Wang et al. found that several AtEXL6 knockdown transgenic plants
2011). The presence of conserved motifs in most of the 121 exhibited defective pollen grain phenotypes that were well
BrGELPs (Fig. 3; Supplementary Table 3), in particular the correlated with the reduction in AtEXL6 expression level.
four strictly conserved residues Ser-Gly-Asn-His in blocks, A detailed analysis of pollen development revealed that
I, II, III, and V (Akoh et al. 2004), strongly suggests that cytoplasmic degradation occurred after the tetrad stage in
most of the identified genes encode functional GELP the AtEXL6 knockdown lines (Fig. 7c). In summary, we
enzymes, and that these motifs play specific roles in diver- demonstrated that AtEXL6 plays an essential role in Arabi-
sity of substrate selection or multifunctionality. Functional dopsis pollen development; based on phylogenetic analysis
information about unknown genes can be deduced from the and expression profiling, we propose that BrEXL6 is an
study of orthologs with known functions (Volokita et al. ortholog of AtEXL6.
2011), although the presence of several copies of B. rapa
orthologs corresponding to each Arabidopsis gene may Acknowledgments This research was supported by research grants
from Golden Seed Project (Center for Horticultural Seed Develop-
imply functional divergence, redundancy, or degeneration ment, nos. 213003-04-3-CG100 and 213003-04-3-SB230), Ministry
(Fig. 2). The phylogenetic grouping suggests that a subset of Agriculture, Food and Rural Affairs (MAFRA), Ministry of Oceans
of BrGELP genes are orthologous to the Arabidopsis pol- and Fisheries (MOF), and Rural Development Administration (RDA).
len fertility-related genes EXL4–6.
Compliance with ethical standards

Expression characteristics of BrGELPs Conflict of interest Dong X declares that he has no conflict of inter-
est. Yi H declares that he has no conflict of interest. Han C-T declares
Because only a few plant GELPs have been functionally that he has no conflict of interest. Nou I-S declares that he has no con-
flict of interest. Hur Y declares that he has no conflict of interest.
characterized so far, we analyzed BrGELP expression
patterns to obtain additional clues about their functions
(Fig. 4, Supplementary Fig. 4). Many BrGELPs are prefer-
entially or specifically expressed in roots, stems, or leaves
References
(Supplementary Fig. 4), implying that they play specific Agee AE, Surpin M, Sohn EJ, Girke T, Rosado A, Kram BW et al
roles required in these tissue types. (2010) MODIFIED VACUOLE PHENOTYPE1 is an Arabidop-
Three distinct types of expression patterns (constitu- sis myrosinase-associated protein involved in endomembrane
tive, stage specific during floral bud development, and protein trafficking. Plant Physiol 152:120–132
Akoh CC, Lee GC, Liaw YC, Huang TH, Shaw JF (2004) GDSL fam-
male-fertile bud-specific) were observed in the floral bud, ily of serine esterases/lipases. Prog Lipid Res 43:534–552
which includes both sporophytic and gametophytic tissues Bailey TL, Boden M, Buske FA, Frith M, Grant CE, Clementi L et al
(Fig. 4). The identification of many BrGELP genes, includ- (2009) MEME SUITE: tools for motif discovery and searching.
ing the putative orthologs of AtEXL4 and AtEXL6s, which Nucleic Acids Res 37:W202–W208
Brick DJ, Brumlik MJ, Buckley JT, Cao JX, Davies PC, Misra S et al
in contrast to their male-sterile counterparts are specifi- (1995) A new family of lipolytic plant enzymes with members in
cally expressed in fertile floral buds, raises the possibility rice, arabidopsis and maize. FEBS Lett 377:475–480
that these BrGELPs are involved in pollen development. In Cheng F, Wu J, Fang L, Sun S, Liu B, Lin K et al (2012) Biased gene
Arabidopsis, AtEXL6 is a target of AMS, a master regulator fractionation and dominant gene expression among the subge-
nomes of Brassica rapa. PLoS One 7:e36442
that coordinates pollen wall development and sporopollenin Chepyshko H, Lai CP, Huang LM, Liu JH, Shaw JF (2012) Multi-
biosynthesis (Xu et al. 2014). The pollen coat protein EXL4 functionality and diversity of GDSL esterase/lipase gene family
affects lipid composition of the pollen coat and is important in rice (Oryza sativa L. japonica) genome: new insights from
for efficient pollination (Updegraff et al. 2009). bioinformatics analysis. BMC Genom 13:309

13
Mol Genet Genomics

Devisetty UK, Covington MF, Tat AV, Lekkala S, Maloof JN (2014) Mayfield JA, Preuss D (2000) Preuss, Rapid initiation of Arabidopsis
Polymorphism identification and improved genome annotation of pollination requires the oleosin-domain protein GRP17. Nat Cell
Brassica rapa through deep RNA sequencing. G3 4:2065–2078 Biol 2:128–130
Dong X, Feng H, Xu M, Lee J, Kim YK, Lim YP et al (2013a) Com- Mayfield JA, Fiebig A, Johnstone SE, Preuss D (2001) Gene fami-
prehensive analysis of genic male sterility-related genes in Bras- lies from the Arabidopsis thaliana pollen coat proteome. Science
sica rapa using a newly developed Br 300K oligomeric chip. 292:2482–2485
PLoS One 8:e72178 Mølgaard A, Kauppinen S, Larsen S (2000) Rhamnogalacturonan
Dong X, Kim WK, Lim YP, Kim YK, Hur Y (2013b) Ogura-CMS in acetylesterase elucidates the structure and function of a new fam-
Chinese cabbage (Brassica rapa ssp. pekinensis) causes delayed ily of hydrolases. Structure 8:373–383
expression of many nuclear genes. Plant Sci 199–200:7–17 O’Brien KP, Remm M, Sonnhammer EL (2005) Inparanoid: a com-
Du Z, Zhou X, Ling Y, Zhang Z, Su Z (2010) agriGO: a GO analy- prehensive database of eukaryotic orthologs. Nucleic Acids Res
sis toolkit for the agricultural community. Nucleic Acids Res 33:D476–D480
38:W64–W70 Oh IS, Park AR, Bae MS, Kwon SJ, Kim YS, Lee JE, Kang NY, Lee
Duan W, Song X, Liu T, Huang Z, Ren J, Hou X et al (2015) Genome- S, Cheong H, Park OK (2005) Secretome analysis reveals an
wide analysis of the MADS-box gene family in Brassica rapa Arabidopsis lipase involved in defense against Alternaria bras-
(Chinese cabbage). Mol Genet Genomics 290:239–255 sicicola. Plant Cell 17:2832–2847
Edlund AF, Swanson R, Preuss D (2004) Pollen and stigma struc- Peterson R, Slovin JP, Chen C (2010) A simplified method for dif-
ture and function: the role of diversity in pollination. Plant Cell ferential staining of aborted and non-aborted pollen grains. Int J
16:S84–S97 Plant Biol 1:e13
Finn RD, Clements J, Eddy SR (2011) HMMER web server: inter- Quilichini TD, Douglas CJ, Samuels AL (2014) New views of tape-
active sequence similarity searching. Nucleic Acids Res tum ultrastructure and pollen exine development in Arabidopsis
39:W29–W37 thaliana. Ann Bot 114:1189–1201
Hong JK, Choi HW, Hwang IS, Kim DS, Kim NH, Choi DS et al Song X, Li Y, Hou X (2013) Genome-wide analysis of the AP2/ERF
(2008) Function of a novel GDSL-type pepper lipase gene, transcription factor superfamily in Chinese cabbage (Brassica
CaGLIP1, in disease susceptibility and abiotic stress tolerance. rapa ssp. pekinensis). BMC Genom 14:573
Planta 227:539–558 Tamura K, Stecher G, Peterson D, Filipski A, Kumar S (2013)
Jiang Y, Chen R, Dong J, Xu Z, Gao X (2012) Analysis of GDSL MEGA6: molecular evolutionary genetics analysis version 6.0.
lipase (GLIP) family genes in rice (Oryza sativa). Plant Omics Mol Biol Evol 30:2725–2729
5(4):351–358 Tan X, Yan S, Tan R, Zhang Z, Wang Z, Chen J (2014) Characteriza-
Kim KJ, Lim JH, Kim MJ, Kim T, Chung HM, Paek KH (2008) tion and expression of a GDSL-like lipase gene from Brassica
GDSL-lipase1(CaGL1) contributes to wound stress resistance by napus in Nicotiana benthamiana. Protein J 33:18–23
modulation of CaPR-4 expression in hot pepper. Biochem Bio- Thamilarasan SK, Park JI, Jung HJ, Nou IS (2014) Genome-wide
phys Res Commun 374:693–698 analysis of the distribution of AP2/ERF transcription factors
Koonin EV (2005) Orthologs, paralogs, and evolutionary genomics. reveals duplication and CBFs genes elucidate their potential
Annu Rev Genet 39:309–338 function in Brassica oleracea. BMC Genom 15:422
Krishnamurthy P, Hong JK, Kim JA, Jeong MJ, Lee YH, Lee SI Toufighi K, Brady SM, Austin R, Ly E, Provart NJ (2005) The Botany
(2015) Genome-wide analysis of the expansin gene superfam- Array Resource: e-Northerns, Expression Angling, and promoter
ily reveals Brassica rapa-specific evolutionary dynamics upon analyses. Plant J 43:153–163
whole genome triplication. Mol Genet Genomics 290:521–530 Updegraff EP, Zhao F, Preuss D (2009) The extracellular lipase EXL4
Kwon SJ, Jin HC, Lee S, Nam MH, Chung JH, Kwon SI et al (2009) is required for efficient hydration of Arabidopsis pollen. Sex
GDSL lipase-like 1 regulates systemic resistance associated with Plant Reprod 22:197–204
ethylene signaling in Arabidopsis. Plant J 58:235–245 Upton C, Buckley JT (1995) A new family of lipolytic enzymes?
Lee DS, Kim BK, Kwon SJ, Jin HC, Park OK (2009) Arabidopsis Trends Biochem Sci 20:178–179
GDSL lipase 2 plays a role in pathogen defense via negative Volokita M, Rosilio-Brami T, Rivkin N, Zik M (2011) Combining
regulation of auxin signaling. Biochem Biophys Res Commun comparative sequence and genomic data to ascertain phyloge-
379:1038–1042 netic relationships and explore the evolution of the large GDSL-
Leščić Ašler I, Ivić N, Kovačić F, Schell S, Knorr J, Krauss U et al lipase family in land-plants. Mol Biol Evol 28:551–565
(2010) Probing enzyme promiscuity of SGNH hydrolases. Wang X, Wang H, Wang J, Sun R, Wu J, Liu S et al (2011) The
Chembiochem 11:2158–2167 genome of the mesopolyploid crop species Brassica rapa. Nat
Li J, Gao G, Zhang T, Wu X (2013) The putative phytocyanin genes Genet 43:1035–1039
in Chinese cabbage (Brassica rapa L.): genome-wide identifica- Winter D, Vinegar B, Nahal H, Ammar R, Wilson GV, Provart NJ
tion, classification and expression analysis. Mol Genet Genomics (2007) An “Electronic Fluorescent Pictograph” browser for
288:1–20 exploring and analyzing large-scale biological data sets. PLoS
Li J, Gao G, Xu K, Chen B, Yan G, Li F et al (2014) Genome-wide One 2:e718
survey and expression analysis of the putative non-specific lipid Xu J, Ding Z, Vizcay-Barrenan G, Shi J, Liang W, Yuan Z et al (2014)
transfer proteins in Brassica rapa L. PLoS One 9:e84556 ABORTED MICROSPORES acts as a mater regulator of pollen
Ling H (2008) Sequence analysis of GDSL lipase gene family in wall formation in Arabidopsis. Plant Cell 26:1544–1556
Arabidopsis thaliana. Pak J Biol Sci 11:763–767 Zhang Z, Ober JA, Kliebenstein DJ (2006) The gene controlling the
Ling H, Zhao J, Zuo K, Qiu C, Yao H, Qin J et al (2006) Isolation and quantitative trait locus EPITHIOSPECIFIER MODIFIER1 alters
expression analysis of a GDSL-like lipase gene from Brassica glucosinolate hydrolysis and insect resistance in Arabidopsis.
napus L. J Biochem Mol Biol 39:297–303 Plant Cell 18:1524–1536

13

You might also like