Accumulation of Charantin and Expression of Triterpenoid

Huang , D. W.; Sherman , B. T.; Lempicki , R. A. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources Nat. Pro...
0 downloads 0 Views 4MB Size
Article pubs.acs.org/JAFC

Accumulation of Charantin and Expression of Triterpenoid Biosynthesis Genes in Bitter Melon (Momordica charantia) Do Manh Cuong,† Jin Jeon,† Abubaker M. A. Morgan,† Changsoo Kim,† Jae Kwang Kim,‡ Sook Young Lee,§ and Sang Un Park*,† †

Department of Crop Science, Chungnam National University, 99 Daehak-ro, Yuseong-gu, Daejeon 34134, Korea Division of Life Sciences and Bio-Resource and Environmental Center, Incheon National University, Yeonsu-gu, Incheon 406-772, Korea § Regional Innovation Center for Dental Science & Engineering, Chosun University, 309 Pilmun-daero, Dong-gu, Gwangju 501-759, Korea J. Agric. Food Chem. 2017.65:7240-7249. Downloaded from pubs.acs.org by LA TROBE UNIV on 01/04/19. For personal use only.



S Supporting Information *

ABSTRACT: Charantin, a natural cucurbitane type triterpenoid, has been reported to have beneficial pharmacological functions such as anticancer, antidiabetic, and antibacterial activities. However, accumulation of charantin in bitter melon has been little studied. Here, we performed a transcriptome analysis to identify genes involved in the triterpenoid biosynthesis pathway in bitter melon seedlings. A total of 88,703 transcripts with an average length of 898 bp were identified in bitter melon seedlings. On the basis of a functional annotation, we identified 15 candidate genes encoding enzymes related to triterpenoid biosynthesis and analyzed their expression in different organs of mature plants. Most genes were highly expressed in flowers and/or fruit from the ripening stages. An HPLC analysis confirmed that the accumulation of charantin was highest in fruits from the ripening stage, followed by male flowers. The accumulation patterns of charantin coincide with the expression pattern of McSE and McCAS1, indicating that these genes play important roles in charantin biosynthesis in bitter melon. We also investigated optimum light conditions for enhancing charantin biosynthesis in bitter melon and found that red light was the most effective wavelength. KEYWORDS: bitter melon, triterpenoid biosynthesis pathway, charantin, transcriptome analysis, LED irradiation



INTRODUCTION Advances in next-generation sequencing (NGS) have enabled detailed analysis of genetic networks of any plant species. It has been suggested that Trinity, Trans-ABySS, Velvet-Oases, and SOAP de novo programs are the best tools for analyzing genes at the transcriptomic levels.1 Trinity has been shown to perform well in analyses by the External RNA Control Consortium (ERCC) on human data.2 Recent reports have shown that the Trinity program is one of the best tools for annotating transcriptomes, including for medicinal plants such as Lycoris sprengeri,3 Gerbera hybrida,4 Lycium chinense,5 Raphanus sativus,6 Paulownia fortunei,7 Salvia splendens,8 Rubus coreanus Miquel,9 and Momordica cochinchinensis.10 Shukla et al.11 reported de novo transcriptome sequencing of bitter melon from roots, flower buds, stems, and leaf samples of a gynoecious line (Gy323) and a monoecious line (DRAR1) but did not investigate the triterpenoid biosynthesis pathway of seedlings. Bitter melon (Momordica charantia) is a subtropical creeping plant belonging to the family Cucurbitaceae that is widely used as a medicinal herb. The plant reportedly possesses antioxidant, anti-inflammatory, anti-HIV, and anticancer activities, and has been used to treat diseases such as diabetes, liver disorders, and gastric problems.12−17 Bitter melon contains several bioactive compounds, such as cucurbitane type triterpenoids (kuguacin J, karavilosides, kuguaglycoside C, momordicoside, charantin), triterpene glycosides, phenolic acids, flavonoids, essential oils, saponins, fatty acids (α-eleostearic acid), and proteins (α-momorcharin, RNase MC2, and MAP30).18 The fruit © 2017 American Chemical Society

juice of bitter melon has been shown to reduce blood sugar levels and has been used in the treatment of diabetes for centuries.19 Charantin, a natural cucurbitane type triterpenoid that is present in the fruit, leaves, and seeds of bitter melon,20−22 has been reported to possess potential hypoglycemic activity.23 Charantin is an equal mixture of β-sitosteryl glucoside (C35H60O6) and 5,25-stigmasteryl glucoside (C35H58O6).24 Like that in other triterpenoid compounds, charantin biosynthesis starts with the formation of acetyl-CoA that is converted into mevalonate, isopentenyl diphosphate (IPP), and squalene by several enzymes: acetoacetyl-CoA thiolase (AACT), HMG-CoA synthase (HMGS), HMG-CoA reductase (HMGR), mevalonate kinase (MK), phosphomevalonate kinase (PMK), mevalonate diphosphate decarboxylase (MVD), farnesyl diphosphate synthase (FPS), and squalene synthase (SQS). In the next step, squalene epoxidase (SE) converts squalene to 2,3-oxidosqualene, which is the ubiquitous precursor for the biosynthesis of sterol and diverse triterpenes in plants.25 Subsequently, 2,3-oxidosqualene is converted to cycloartenol by cycloartenol synthase (CAS) (Figure 1). Recent relevant studies have focused on environmental factors such as light quality, irrigation, and temperature variation that lead to the accumulation of secondary compounds in Received: Revised: Accepted: Published: 7240

April 27, 2017 July 22, 2017 July 24, 2017 July 24, 2017 DOI: 10.1021/acs.jafc.7b01948 J. Agric. Food Chem. 2017, 65, 7240−7249

Article

Journal of Agricultural and Food Chemistry

adjustable light intensity.34,35 Use of LEDs has allowed us to easily analyze the effects of different wavelengths and light intensities on the accumulation of secondary compounds in controlled environments. In this study, we analyzed the transcriptome data of bitter melon and identified several candidate genes that encode enzymes related to triterpenoid biosynthesis pathways. To investigate the spatial distribution of triterpenoid-related genes, we analyzed the expression of these genes and the accumulation of charantin in different organs. In addition, we also investigated whether LED illumination could enhance charantin accumulation in bitter melon. These results will be a valuable resource for further research on bitter melon bioengineering and will provide basic information to increase the yield of medicinal compounds in bitter melon.



MATERIALS AND METHODS

Plant Material. Bitter melon seeds were purchased from Beijing Namo Tech.-Trade Co. Ltd. (Beijing, China). Dehulled seeds were sterilized with 70% (v/v) ethanol for 30 s and 1% (v/v) sodium hypochlorite solution for 10 min. The sterilized seeds were germinated on half Murashige and Skoog (1/2 MS) medium36 with 2.5 mM 2-(Nmorpholino) ethanesulfonic acid (MES), pH 5.7, and 0.8% agar at 25 °C under 16 h light/8 h dark conditions and incubated in a growth chamber for 10 days. For the LED illumination experiment, bitter melon seedlings were transferred in 1/2 MS liquid media and incubated in a growth chamber under white (380 nm), blue (470 nm), or red (660 nm) light wavelengths. The conditions in the LED chamber were temperature, 25 ± 1 °C; humidity, 85 ± 5%; electrical conductivity, 1,200 ± 90 μS/cm; CO2, 1000 ± 100 ppm; light intensity, 160 μmol·m−2.·−1; and photoperiod, 16 h light/8 h dark. Whole seedlings were harvested after 7, 14, or 21 days of illumination. In other experiments, bitter melon plants were grown at an experimental farm at Chungnam National University (Daejeon, Korea). Young leaves, mature leaves, stems, roots, male flowers, female flowers (Figure 2A−F), and four development of stages of fruit (Figure 2G) were collected, immediately frozen in liquid nitrogen, and stored at −80 °C for RNA isolation, or were freeze-dried for high-performance liquid chromatography−ultraviolet (HPLC-UV) analysis. Illumina Sequencing. We used the RNeasy Mini kit (Qiagen, USA) to isolate total RNA from bitter melon seedlings. We removed rRNAs in total RNA using a Ribo-Zero rRNA Removal kit (Epicenter, RZPL11016). Magnetic Oligo (dT) beads were used for the separation and purification of mRNA. Then, mRNA was fragmented to short pieces using divalent cations under elevated temperature, which were used to establish a cDNA library using the TruSeq Stranded Total RNA Sample Prep kit-LT set A and B (Illumina, Rs-122-2301 and Rs-122-2302). The cDNA library was sequenced in 76 bp length

Figure 1. Triterpenoid biosynthetic pathway in plants. Red indicates the genes which were monitored in this study. AACT, acetoacetyl-CoA thiolase; HMGS, HMG-CoA synthase; HMGR, HMG-CoA reductase; MK, mevalonate kinase; PMK, phosphomevalonate kinase; MVD, mevalonate diphosphate decarboxylase; IDI, isopentenyl diphosphate isomerase; FPS, farnesyl diphosphate synthase; SQS, squalene synthase; SE, squalene epoxidase; β-AS, β-amyrin synthase; CPQ, cucurbitadienol synthase; CAS, cycloartenol synthase; DMAPP, dimethylallyl diphosphate; IPP, isoprenyl diphosphate.

plants. In particular, irradiation by different wavelengths of light with light-emitting diodes (LEDs) have been reported in a range of plant species such as lettuce,26,27 buckwheat,28,29 lemon balm,30 pea seedling,31 chinese cabbage,32 and rapeseed33 with promising results, although varying among the different plant species. However, it was not reported in bitter melon. LEDs were used as sources of artificial lighting in plant production systems because they have many advantages including small size, long operating lifetime, wavelength specificity, low heat emission, high energy-conversion efficiency, and

Figure 2. Phenotype of bitter melon. Male flowers (A), female flowers (B), stems (C), young leaves (D), mature leaves (E), roots (F), and four developmental stages fruits (G). The scale bars represent 5 cm. 7241

DOI: 10.1021/acs.jafc.7b01948 J. Agric. Food Chem. 2017, 65, 7240−7249

Article

Journal of Agricultural and Food Chemistry

(http://www.uniprot.org) with GO annotation, Arabidopsis protein database from the Arabidopsis Information Resource (TAIR10, https://www.arabidopsis.org), Brassica rapa protein database from BRAD (http://brassicadb.org/brad), and Protein database from COG (http://www.ncbi.nlm.nih.gov/COG) with COG functional category annotation. For sequences that could not be predicted by BLAST, we used Augustus (http://bioinf.uni-greifswald.de/augustus) with the Arabidopsis thaliana model to predict “CDS” and orientation. GO annotation for each unigene was transferred from BLASTX hits in the UniProt protein database. After obtaining GO annotations for all transcripts, WeGo (http://wego.genomics.org.cn) was utilized to determine gene function before finally performing GO classification at the category 2 level.41 To quantify bitter melon unigene expression, we aligned preprocessed high-quality reads of bitter melon transcript sequences and calculated the expression values with the aligned read counts for each unigene. Bowtie242 software was used to align the high-quality reads on the unigene sequences, and eXpress (http://bio. math.berkeley.edu/eXpress)43 was used to quantify the differentially expressed genes (DEGs). Finally, DAVID software (https://david. ncifcrf.gov) was used to functionally classify the DEGs.44 Identification of Candidate Genes Related to Triterpenoid Biosynthetic Pathways. To identify triterpene biosynthesis genes in the bitter melon transcriptome database, functional genes related to triterpene were selected by evaluating the annotated file in the GenBank database using BLAST. The genes that showed maximum identity and similarity were selected for further transcriptional study. RNA Isolation and cDNA Synthesis. Total RNA was isolated from the different plant organs (young leaves, mature leaves, stems, roots, male flowers, female flowers, and four development of stages fruit) of bitter melon using Easy BLUE Total RNA Kit (iNtRON, Seongnam, Korea), and the quality was determined by running a 1.2% agarose gel. ReverTra Ace kit (Toyobo Co., Ltd., Japan) was used for the synthesis of cDNA. The prepared cDNA was then diluted 20-fold prior to performing quantitative real-time PCR. Quantitative Real-time PCR Analysis. Primers for the 15 genes involved in the triterpene pathways were designed using the Primer3 Web site (http://bioinfo.ut.ee/primer3-0.4.0/primer3/),45 and their sequences are provided in Table S1. For real time-PCR, the reaction mixture (20 μL) contained 2 μL of primers (10 ppM), 5 μL of diluted cDNA (20-fold dilution), and a PCR mixture of 2× SYBR Green (10 μL) and water (3 μL). The PCR conditions were 3 min at 95 °C; 41 cycles of 15 s at 95 °C, 15 s at 56 °C, and 15 s at 72 °C.

Table 1. Summary of Annotations of Bitter Melon Transcripts

all transcripts transcripts BLASTed against transcripts BLASTed against transcripts BLASTed against PROT transcripts BLASTed against transcripts BLASTed against transcripts BLASTed against transcripts BLASTed against transcripts CDS predicted all annotated transcripts

NR NT SWISSBRAD TAIR COG GO

number of annotated transcripts

ratio (%)

88,703 43,635 36,768 34,214

100 49.19 41.45 38.57

41,726 41,031 14,579 34,214 40,383 51,745

47.04 46.26 16.44 38.57 45.53 58.34

paired-end (PE) reads in an Illumina NextSeq500 sequencer (Illumina Inc., San Diego, CA, USA) to produce 68,073,862 raw sequencing reads. Illumina Sequencing was performed at the Life is Art of Science (LAS) company in Seoul, Korea (http://www.lascience.co.kr/). Read Filtration and Assembly. To assess the quality and the adapter sequences of the raw read data, we used FastQC (http:// www.bioinformatics.babraham.ac.uk/projects/fastqc) with the default options. Before assembly, we used the adapter trimming software CutAdapt (https://pypi.python.org/pypi/cutadapt) to remove the adapter sequences and low-quality bases from the raw reads.37 The high-quality reads were then assembled using the Trinity package (http://trinityrnaseq.github.io)38,39 and evaluated by Translate S/W (http://trinityrnaseq.github.io). To reduce redundant operations in gene annotation or following downstream analyses, we used CD-HitEST (http://weizhongli-lab.org/cd-hit)40 to cluster the bitter melon transcriptome contigs based on sequence similarity. The longest contig in each cluster was chosen as a unigene for the corresponding bitter melon gene. Unigene Function Annotation. To identify the bitter melon transcripts and estimate their biological functions, we performed a homology search of the unigene sequences, with a significance threshold E-value ≤10−5, in various public protein and nucleotide databases: BLASTX, BLASTN, NCBI nr protein and nt nucleotide database (ftp://ftp.ncbi.nlm.nih.gov), SwissProt protein database

Figure 3. Classification of NR annotation results of bitter melon (Momordica charantia) transcripts. 7242

DOI: 10.1021/acs.jafc.7b01948 J. Agric. Food Chem. 2017, 65, 7240−7249

Article

Journal of Agricultural and Food Chemistry Table 2. Comparison of Triterpenoid Biosynthetic Genes of Bitter Melon with the Most Orthologous Genesa genes of M. charantia

description

orthologous genes

McAACT (407 aa) acetoacetyl-coenzyme A Cucumis melo (CoA) thiolase (XP_008462041.1) Siraitia grosvenorii (AEM42969.1) Cucumis sativus (XP_004144550.1) McHMGS (443 aa) HMG-CoA synthase Siraitia grosvenorii (AEM42970.1) Cucumis sativus (XP_004147967.1) Cucumis melo (XP_008448952.1) McHMGR1 (592 aa) (HMG-CoA) reductase Cucumis melo (XP_008448157.1) Cucumis sativus (XP_004139979.1) Malus domestica (XP_008348952.1) McHMGR2 (584 aa) (HMG-CoA) reductase Cucumis melo (XP_008451185.1) Cucumis sativus (XP_004144569.1) Malus domestica (XP_008348952.1) McMK (386 aa) mevalonate kinase Cucumis sativus (XP_004139185.1) Cucumis melo (XP_008454825.1) Cucumis melo (XP_008454844.1) McPMK (507 aa) mevalonate-5-phosphate Momordica charantia kinase (AKO82483.1) Cucumis melo (XP_008466028.1) Cucumis sativus (XP_004136210.1) McMVD (418 aa) mevalonate diphosphate Siraitia grosvenorii decarboxylase (AEM42974.1) Cucumis melo (XP_008446768.1) Cucumis sativus (XP_004150845.1) McIDI (234 aa) isopentenyl diphosphate Siraitia grosvenorii isomerase (AEM42977.1) Cucumis melo (XP_008443691.1)

a

genes of M. charantia

identity (%) 96

McFPS1 (342 aa)

96 95 96

McFPS2 (342 aa)

94 94 87

McSQS (417 aa)

87 80 90

McSE (430 aa)

90 80 87

McCPQ (636 aa)

87 84 99

McCAS1 (391 aa)

92 91 94

McCAS2 (319 aa)

94 93 97

description

orthologous genes

Cucumis sativus (XP_004139102.1) farnesyl diphosphate Cucumis melo synthase (XP_008463030.1) Cucumis sativus (XP_004149409.1) Siraitia grosvenorii (AIS86112.1) farnesyl diphosphate Cucumis melo synthase (XP_008463060.1) Cucumis sativus (XP_004138366.1) Siraitia grosvenorii (AEM42979.1) squalene synthase Siraitia grosvenorii (AEM42980.1) Cucumis sativus (XP_011649259.1) Cucumis melo (XP_008461009.1) squalene epoxidase Cucumis sativus (XP_004136919.1) (squalene monooxygenase) Cucumis melo (XP_008455523.1) Gossypium raimondii (XP_012449747.1) cucurbitadienol synthase Siraitia grosvenorii (AEM42982.1) Cucumis melo (XP_008459938.1) Cucumis sativus (NP_001292630.1) cycloartenol synthase Siraitia grosvenorii (AEM42981.1) Cucumis melo (XP_008462186.1) Cucumis sativus (XP_004141754.1) cycloartenol synthase Siraitia grosvenorii (AEM42981.1) Cucumis sativus (XP_004141754.1) Cucumis melo (XP_008462186.1)

identity (%) 95 90 90 96 96 94 94 94 93 93 80 88 74 94 91 90 93 90 89 94 92 91

96

aa, amino acid. phase of MeOH/water (98:2 v/v) was used; 20 μL was injected for each sample. Identification and quantification of charantin was carried out by comparing the retention times and the peak areas with reference to a charantin external standard with 92.2% purity (ChromaDex Inc., 10005 Muirlands Blvd., Suite G, First Floor, Irvine, USA) or by the direct addition of the standard into the sample. Quantification and analysis were each performed in triplicate. Statistical Analysis. For HPLC statistical analysis, data were analyzed by the statistical analysis software (SAS version 9.2, SAS Institute Inc., USA). All data are given as the average mean and standard deviation of triplicate experiments. The experimental data were subjected to analysis of variance (ANOVA), and significant differences among the means were determined using Duncan’s multiple-range test.

Bitter melon cyclophilin gene (McCYP) (accession number: HQ171897.1) was used as the positive gene reference for quantifying the expression of the selected genes. The PCR products were analyzed using Bio-Rad CFX Manager 2.0 software. The analysis was repeated three times. High Performance Liquid Chromatography (HPLC) Analysis. The extraction method used for charantin analysis in bitter melon was similar to that described by Kim et al.46 with some modifications. Samples (100 mg) were extracted at room temperature with 2 mL of n-hexane to remove lipids and then dried in an extractor hood. The procedure was repeated twice. Samples were then extracted with 1 mL of MeOH (100%). The extract was vortexed, incubated in a sonicator (Branson Ultrasonic Co., Danbury, CT, USA) for 1 h, and centrifuged (12,000 rpm at 4 °C) for 10 min. The supernatant was filtered through Whatman No. 42 filter paper and allowed to evaporate (Heidolph VV2011, 40 °C). The evaporated extract was resuspended in 1 mL of MeOH (HPLC grade) for HPLC analysis. An Agilent Technologies 1260 series HPLC system (Agilent Technologies, Palo Alto, CA, USA) with a UV detector (204 nm) was used for the analysis. Separation was performed on an Optimapak C-18 column (4.6 mm × 250 mm, 5 μm, 100 Å; RStech, Korea) with a flow rate of 0.8 mL min−1. A mobile



RESULTS AND DISCUSSION RNA Sequencing and De Novo Assembly of Bitter Melon Transcriptome. To obtain an overview of the bitter melon transcriptome, we used an Illumina Nextseq500 system to sequence a cDNA library from total RNA derived from seedlings. After the initial processing, a total of 68,073,862 clean 7243

DOI: 10.1021/acs.jafc.7b01948 J. Agric. Food Chem. 2017, 65, 7240−7249

Article

Journal of Agricultural and Food Chemistry

Figure 4. Expression profiles of triterpenoid biosynthetic genes in bitter melon. Total RNA was extracted from roots, stems, Y-leaves (young leave), M-leaves (mature leaves), F-flowers (female flowers), M-flowers (male flowers), and four fruit stages (stages 1 to 4) of bitter melon and used for qRT-PCR analysis. The relative expression ratio of the transcripts was shown to be calibrated with the expression of McCYP. The height of each bar and the error bars indicate the mean and standard error, respectively, from three independent measurements.

65,540 transcripts for the gynoecious line (Gy323) and 61,490 transcripts for the monoecious line (DRAR1).11 With regard to transcript size, 38,243 of the transcripts (43.11%) were less than 500 bp, 23,001 (25.93%) were 500−1,000 bp, 23,602 (26.61%) of transcripts were 1,000−3,000 bp, and 3,857 (4.35%) were more than 3,000 bp (Figure S1).

pair-end reads were obtained for assembly by Trinity software (Table S2). After selecting the best reads, a total of 126,279 contigs were assembled with an average length of 1,036 bp; the contigs were assembled into 88,703 transcripts with a mean size of 898 nt and N50 of 1,600 nt. This result was higher than that previously reported by Shukla et al., who obtained a total of 7244

DOI: 10.1021/acs.jafc.7b01948 J. Agric. Food Chem. 2017, 65, 7240−7249

Article

Journal of Agricultural and Food Chemistry Functional Annotation and Classification of Bitter Melon. We used BLASTX to annotate 51,745 transcripts against the nonredundant (NR) protein database and the nucleotide (NT) database at the National Center for Biotechnology Information (NCBI) Web site; approximately 58.34% of all transcripts were annotated (Table 1). Of the total 88,703 transcripts, 43,635 transcripts (49.19%) had BLAST hits to known proteins in the NR database, and 36,768 transcripts (41.45%) had BLAST hits to nucleotides in the NT database. In addition, some transcripts were aligned to other public databases, including 34,214 (38.57%) transcripts in the SWISSPROT protein database, 41,726 (47.04%) transcripts in the Brassica database (BRAD), 41,031 (46.26%) transcripts in the Arabidopsis information resource (TAIR) database, 14,579 (16.44%) transcripts in the Clusters of Orthologous Group (COG) database, and 34,214 (38.57%) transcripts in the Gene Ontology (GO) database. The E-value distributions showed that approximately 59.8% of the transcripts had strong similarity (