Identification and analysis of CYP450 genes from transcriptome of Lonicera japonica and expression analysis of chlorogenic acid biosynthesis related CYP450s

Identification of CYP450 genes in L. japonica

The transcriptome data of L. japonica generated from different sequencing platforms including 454 GS-FLX, Illumina HiSeq2000, and Illumina GA II was downloaded from the NCBI SRA database with accession numbers SRR290309, SRR342027, SRR576924, SRR576925 and SRR766791. Four datasets were assembled and annotated. To identify putative CYP450 genes, both Hidden Markov Model (HMM) method and BLAST method were used. For HMM method, P450.hmm file which represents the Hidden Markov Model of the cytochrome P450 family was initially downloaded from Pfam (http://pfam.xfam.org/), and then, HMMER3 software (Eddy, 2011) was used to search P450.hmm against L. japonica deduced amino acid database. For BLAST method, 19,047 full length plant CYP450 sequences were retrieved from UniProt (http://www.uniprot.org/). These sequences were used as queries to tblastn against L. japonica transcriptome assembly with an E-value cutoff of 1e−5. After filtering out the repeated results, the coding sequences of the resultant subjects were retrieved. Finally, results from the two methods were integrated and corrected manually. The identification methods were conducted for the four datasets of L. japonica and the results were also integrated and corrected. The corrected L. japonica CYP450s were further submitted to NCBI Conserved Domain Search (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi) to predict the conserved domain. Sequences with complete cytochrome P450 domain were selected for further analysis.

Classification and characterization of L. japonica CYP450 genes

L. japonica CYP450s were classified into different families and subfamilies according to the sequence similarity using sequences from Cytochrome P450 Homepage as reference sequences. If the amino acid sequences of L. japonica CYP450s showed >40%, >55%, or >95% sequence similarity with reference sequences, they were classified into the same family, subfamily, or allelic variant, respectively (Nelson, 2009).

The deduced amino acid sequences of L. japonica CYP450s were subjected to Multiple Expectation Maximization for Motif Elicitation (MEME, http://meme-suite.org/) (Bailey et al., 2009) analysis for identification of conserved motifs. Sequences of the four conserved CYP450 motifs including heme-binding region, PERF motif, K-helix region and I-helix region were extracted and then subjected to WEBLOGO (http://weblogo.berkeley.edu/) (Crooks et al., 2004) to create the sequence logos.

Phylogenetic analysis of predicted CYP450 genes

A total of 63 representative sequences from plant CYP450 families were selected for phylogenetic analysis with 151 L. japonica CYP450 sequences. Specifically, CYP450 sequences whose functions had already been identified were preferentially selected. Multiple sequence alignment was performed using MUSCLE 3.6 software (Edgar, 2004). The result of alignment was imported to MEGA4 (Tamura et al., 2007) and phylogenetic analysis was performed. The phylogenetic tree was constructed using the Neighbor-Joining algorithm with the Poisson model and pairwise deletion. Bootstrap testing with 1,000 replications was used to test the phylogenetic tree. The Newick format file of bootstrap consensus tree was exported and then modified using EvolView (http://www.evolgenius.info/evolview/) (Zhang et al., 2012).

Gene ontology and KEGG pathway analysis

Blast2GO (http://www.blast2go.com/) (Conesa et al., 2005) was used to perform Gene ontology (GO) annotation of L. japonica CYP450s. These predicted genes were functionally categorized according to three different criterions including cellular component, molecular function and biological process. The GO terms of all L. japonica CYP450s were extracted and subjected to Web Gene Ontology Annotation Plot (WEGO, http://wego.genomics.org.cn/cgi-bin/wego/index.pl) (Ye et al., 2006) to plot GO annotation results. KEGG annotation that maps the L. japonica CYP450s to possible KEGG pathway for biological interpretation of systemic functions was also conducted using Blast2GO.

Extraction and quantification of CGA

The L. japonica used for this study was maintained at the Germplasm Nursery in Institute of Botany, Jiangsu Province and Chinese Academy of Sciences, Nanjing, Jiangsu Province. Flower buds and flowers samples for CGA and RNA extraction were collected at five stages: young alabastrum (YA, ≤1.5 cm), green alabastrum (GA, 2.0–3.0 cm), while alabastrum (WA, 3.2–4.4 cm), silvery flower (SF, about 5 cm), and golden flower (GF, about 5 cm). The extraction and quantification of CGA were conducted as described in Chinese Pharmacopoeia with minor modifications (Chinese Pharmacopoeia Commission, 2010). Briefly, dried buds or flowers were separately comminuted with a miler, and 0.2 g of each solid sample (40 mesh) was extracted with 25 mL of 50% aqueous methanol by ultrasonication (250 W, 35 kHz) for 30 min. After cooling to room temperature, the extracts were replenished to earlier weights with 50% aqueous methanol. Then, 5 mL of the extracts were diluted to 25 mL with 50% aqueous methanol and filtered with 0.45 µm Millipore filter membranes. An Agilent 1200LC series HPLC system was used to analyze the CGA levels. Separations were performed on an Agilent TC-C18 reserved-phase column (5 µm, 250 mm × 4.6 mm) at 25 °C. The mobile phase was composed of acetonitrile-0.4% H₃PO₄ (13:87). The flow rate was 1 mL/min and fractions were monitored at 327 nm. Components were identified by comparison of the retention times of the eluting peaks to those of commercial standards under the same conditions.

RNA extraction and qRT-PCR

Total RNA from five samples was extracted using RNAiso Plus (Takara, Tokyo, Otsu, Shiga, Japan) according to the manufacturer’s instructions. RNA quality and concentration were measured using a ND-1000 UV spectrophotometer (Nanodrop Technologies, Wilmington, DE, USA). First-strand cDNA was synthesized using 3 µg of total RNA with M-MLV reverse transcriptase (Promega, USA) in a 25 µl reaction system. For quantitative real-time reverse transcriptional PCR (qRT-PCR), each reaction was prepared according to the manufacturer’s instructions using SYBR® Premix Ex TaqTM II (Takara) and 2 µl of diluted cDNA as a template. The qRT-PCR reactions were conducted on the qTOWER2.2 Real Time PCR Systems (Analytik, Jena, Germany). The L. japonica actin gene was used as a control to normalize the relative expression levels of target genes. Gene-specific primers used for qRT-PCR were listed on Table S1. All results were representative of three independent experiments.

Identification and classification of CYP450 genes in L. japonica

Cytochrome P450 is one of the most massive gene superfamilies that is comprised of a number of families and subfamilies. In the present study, by integrating the results from different datasets and manual correction, we in total identified 151 putative CYP450s with complete cytochrome P450 domain in L. japonica. Among them, nine CYP450s had been previously reported and the other 142 CYP450s were identified here for the first time in L. japonica. Based on sequence similarity, we classified the 151 CYP450 genes from L. japonica into 10 clans consisting of 45 families and 76 subfamilies (Table 1). Among them, the CYP71 clan, which represents the whole set of A-type CYP450 genes, contains 71 genes belonging to 19 families (CYP71, CYP73, CYP75–CYP84, CYP89, CYP92, CYP93, CYP98, CYP701, CYP706, and CYP736). The non-A type CYP450 genes of L. japonica contains the remaining 80 genes, which belongs to nine CYP clans (CYP51, 72, 74, 85, 86, 97, 710, 711, and 727) and 26 families (CYP51, CYP72, CYP714, CYP715, CYP721, CYP734, CYP749, CYP74, CYP85, CYP87, CYP88, CYP90, CYP707, CYP716, CYP722, CYP724, CYP728, CYP729, CYP86, CYP94, CYP96, CYP704, CYP97, CYP710, CYP711, and CYP727). The largest CYP family of L. japonica is CYP71 and CYP72, containing 17 and 18 members, respectively.

Phylogenetic analysis of predicted CYP450s in L. japonica

Representative members of each plant CYP450 family were selected and used to conduct phylogenetic analysis with 151 CYP450s from L. japonica. The predicted CYP450s were classified into two major branches, A-type (47%) and non-A type (53%) (Fig. 1). There were 10 clans in plants CYP450s. Four clans contained multiple families and were designated by their lowest-numbered family members, CYP71, CYP72, CYP85 and CYP86. The other six clans were designated by their only family, CYP51, CYP74, CYP97, CYP710, CYP711 and CYP727. In L. japonica, all 10 clans were identified. Genes belonging to same clan clustered as one clade. For example, the 72 clan, which comprised 28 CYP450s belonged to six families, were clustered as one clade with the eight representative CYP450s. The CYP71 clan that comprised 71 members belonging to 19 families was the largest clan. Three clans, CYP710, CYP711 and CYP727, had only one member identified for each clan.

Conserved motifs analysis of L. japonica CYP450s

Plant CYP450s shared some typical conserved motifs including heme-binding region, PERF motif, K-helix region and I-helix region, which were important for catalytic activities (Paquette, Jensen & Bak, 2009). The L. japonica CYP450s were divided into A-type and non-A type according to phylogenetic analysis. The deduced amino acid sequences were subjected to MEME to analyze the conserved motifs. The consensus sequences of the heme-binding region, also known as “P450 signature”, were “PFGXGRRXCPG” and “XFXXGXRXCXG” for A-type and non-A type CYP450s, respectively (Fig. 2). The cysteine residues in this motif of two types of CYP450s were universally conserved, which links the heme iron to the apoprotein. The consensus sequences of the PERF motif were also different for two types of CYP450s in L. japonica, which are “PERF” for A-type and “PXRX” for non-A type. The R residues in the PERF motif and E and R residues in the K-helix region were universally conserved, which form a salt bridge that has been proposed to be involved in locking the Cys-pocket in position and assuring the stable association of heme with the protein. The threonine residues in the I-helix region which is involved in oxygen activation was highly conserved in both A-type and non-A type CYP450s. In general, sequences of the typical motifs were conserved in L. japonica CYP450s, and the differences between A-type and non-A type CYP450s in L. japonica were similar with other plants (Chen et al., 2014).

Gene ontology classification of L. japonica CYP450s

Gene ontology (GO) is a classification system for standardized gene functions which classifies genes into three main independent GO categories: cellular component, molecular function and biological process. In this study, GO assignments were conducted to classify the functions of CYP450s from L. japonica using Blast2GO. Results indicated that all 151 CYP450s were mapped to one or more GO terms, of which 145 were assigned to the “cellular component”, 151 to the “molecular function”, and 151 to the “biological process” (Fig. 3). Of these categories, cell, binding, catalytic, and metabolic process were the largest subcategories. Comparison of the GO classification between the A-type and non-A type CYP450s, we found that non-A type CYP450s participated in more molecular functions and biological processes than A-type. For example, GO terms of non-A type CYP450s in molecular function category included demethylase, hydrolase, lyase, and transferase; however, no A-type CYP450s was assigned to these subcategories. In biological process category, non-A type CYP450s participated in more biological processes than A-type, including anatomical structure formation, cellular component organization, developmental process, establishment of localization, growth, localization, multicellular organismal process, and reproduction. The GO annotation provided a valuable clue to investigate the functions of CYP450s in L. japonica.

KEGG pathway analysis of L. japonica CYP450s

In order to further understand the biological functions of CYP450s in L. japonica, pathway-based analysis was performed. Given that a CYP450 could be assigned to one or more KEGG pathways as well as GO terms, 47 (31.1%) CYP450s were totally assigned to 25 KEGG pathways (Fig. 4). The 25 pathways could be mainly grouped into six classes, including lipid metabolism, amino acid metabolism, metabolism of cofactors and vitamins, metabolism of terpenoids and polyketides, biosynthesis of other secondary metabolites, and xenobiotics biodegradation and metabolism. In the class of ‘biosynthesis of other secondary metabolites’, after removing duplicate hits, ten CYP450s (CYP73A-m13469, CYP73A-m177245, CYP73A-m8810, CYP75B-m13120, CYP76A-m155830, CYP78A-m152788, CYP93B-m79556, CYP98A-m184946, CYP98A-m43608 and CYP736A-m18282) were found to be involved in the biosynthesis of phenolic compounds including phenylpropanoid, stilbenoid, diarylheptanoid and gingerol, flavonoid, flavone and flavonol, and isoflavonoid. All ten CYP450s belonged to CYP71 clan. In the class of ‘metabolism of terpenoids and polyketides’, nine CYP450s (CYP72A-m132911, CYP72A-m20456, CYP72A-m206268, CYP72D-m62754, CYP72D-m75640, CYP714A-m189781, CYP714E-m17561, CYP714E-m205273 and CYP734A-m842) were found to be involved in ‘monoterpenoid biosynthesis’, all of which belonged to CYP72 clan. Three CYP450s (CYP701A-m150262, CYP701A-m27329 and CYP728B-m166264) were found to be involved in ‘diterpenoid biosynthesis’, among them, two belonged to CYP71 clan and one belonged to CYP85 clan. Five CYP450s (CYP707A-m213600, CYP707A-m35702, CYP707A-m47109, CYP707A-m212742 and CYP728B-m166264) were found to be involved in ‘carotenoid biosynthesis’, all of which belonged to CYP85 clan.

CYP450s involved in CGA biosynthesis

CGA is the most major active ingredient in L. japonica and the biosynthetic pathway of CGA has been investigated in many plants. In CGA biosynthetic pathway, C4H and C3H are the two CYP450-encoded enzymes that participate in the two steps of hydroxylation. In the present study, three C4H and two C3H genes were identified and cloned from L. japonica. Among them, two LjC4Hs and one LjC3H have been previously reported. The newly identified C4H and C3H were designated as ‘LjC4H3’ (GenBank accession number: KX845341) and ‘LjC3H2’ (GenBank accession number: KX845342), respectively. The C4Hs belonged to CYP73A subfamily and C3Hs belonged to CYP98A subfamily. Phylogenetic analysis indicated that two clades were clustered for C4Hs and C3Hs from L. japonica and other plants (Fig. 5).

Because CGA was mainly accumulated in flower bud of L. japonica, buds and flowers in different developmental stages were selected to explore the relationship of C4H and C3H expressions and CGA contents. HPLC analysis was used to measure CGA concentrations in different developmental stages of buds and flowers. As shown in Fig. 6, the percentage of CGA contents decreased during the flower development. Nevertheless, with the increase of bud or flower weights, the total CGA contents increased from young alabastrum (YA) to while alabastrum (WA) stage and reached peak at the WA stage. After flowering, the total CGA contents decreased quickly during flower development. Furthermore, qRT-PCR was conducted to analyze the transcriptional levels of CGA biosynthetic genes in the different developmental stages of buds and flowers, including the five CYP450s identified in this study. The two LjC3Hs exhibited oppositing expression patterns, the transcriptional levels of LjC3H1 increased but that of LjC3H2 decreased during the flower development (Fig. 7). The expression patterns of three LjC4Hs were quite different and the relative expression levels of LjC4H3 was obviously higher than those of the other two (Fig. 7). Interestingly, the expression patterns of LjPAL1, LjC4H3, LjC3H1 and LjHQT were quite similar, which exhibited a trend of decreasing first and then increasing. Considering the gene expressions with CGA contents, only LjC3H2 exhibited a similar pattern with CGA concentrations.