Functional Characterization of Terpene Synthases from Masson Pine (Pinus massoniana) under Feeding of Monochamus alternatus Adults

Abstract

Conifers have evolved sophisticated terpenoid defenses for protection against herbivores and pathogens. Pinus massoniana Lamb. is the most widely distributed pioneer afforestation and resin tree species in China, but is seriously harmed by pine wilt disease. Monochamus alternatus is the main vector of pine wilt disease in China. Monoterpenes, sesquiterpenes and diterpenes, the main secondary defensive compounds of P. massoniana, are catalyzed by different terpene synthases (TPSs), which participate in the important defense pathways against external biotic and abiotic stresses. Here, we aimed to identify the terpene synthases (TPSs) in P. massoniana, responding to the feeding of M. alternatus, and to characterize the functions and products of the mono-TPSs. We identified six differentially expressed TPS genes in the P. massoniana fed upon by M. alternatus, including four mono-TPS and two sesqui-TPS genes. The functions of the four mono-TPSs were verified by analysis of the main product and by-products of these mono-TPSs. (+)-α-Pinene, (−)-α-pinene, and limonene were the major products of TPS (+)-α-pinene, TPS (−)-α-pinene, and TPS limonene, respectively, but TPS (−)-β-pinene only catalyzed a trace amount of (−)-β-pinene in the products. Our findings shed light on the potential relationships between the structure of terpene synthases and their corresponding products.

1. Introduction

Conifers rely on constitutive and induced defense systems, primarily based on terpenoids, against various biotic and abiotic stresses [1,2,3]. The coniferous terpenoid defense system consists of monoterpenoids, sesquiterpenoids, diterpenoids, and the corresponding resin ducts and cells [4,5]. Conifers can actively change the composition of terpenoids to defend against insect herbivores, affecting insect behavior and performance or attracting natural enemies [6,7,8]. For example, a methyl jasmonate-stimulated feeding treatment affected the release of terpene volatiles from five different conifers [9]. Methyl jasmonate treatment changes the chemical composition of Pinus sylvestris and influences the meal properties of Hylobius abietis [10]. Many herbivorous insects have evolved to use host-produced terpenoids, particularly volatile monoterpenoids and sesquiterpenoids, as critical chemical signals for host recognition and location. For instance, Dendroctonus frontalis and D. terebrans have olfactory and behavioral responses to volatiles (α-pinene, β-pinene, myrcene, limonene, and 4-allylanisole) released from the resins excreted by their hosts [11]. Furthermore, Xyleborus glabratus is attracted to sesquiterpenes, such as α-copaene, α-cubebene, α-humulene, and calamenene, emitted from the cambium of their lauraceous hosts [12].

The components and composition of constitutive terpenoids vary under different biotic and abiotic stresses [13], which is directly related to the distinct gene expression levels of terpene synthases (TPSs) under different conditions [14,15]. In Picea abies and Pinus massoniana (Pm), terpenoid biosynthesis pathways are activated under feeding stress or methyl jasmonate induction. This activation leads to upregulated gene expression of mono-TPS PaJF67 and Pm TPSs, which finally enhances the biosynthesis of some terpenoids, such as α-pinene, β-pinene, and (+)-3-Carene [16,17,18]. Identification and characterization of TPS genes for genetic improvement can improve the genetic improvement of plants and enhance resistance against invaders. For example, transgenic soybean with overexpression of GmAFS showed significantly improved resistance to soybean cyst nematode [19]; expression of the 7-epizingiberene synthase and Z-Z-farnesyl-diphosphate synthases in the glandular trichomes of cultivated tomato resulted in the resistance of tomatoes to several herbivore pests [20].

In plants, TPSs catalyze the formation of many monoterpenes, sesquiterpenes, and diterpenes with different structures, using precursors [21]. Monoterpenes, sesquiterpenes, and diterpenes are synthesized from geranyl diphosphate (GPP), farnesyl diphosphate (FPP), and geranylgeranyl diphosphate, respectively. They are, in turn, synthesized from different numbers of C5s by isoprenyl transferase. Terpenoids usually exist as enantiomers but often with one enantiomer as the dominant product [22]. Terpene synthases can produce one or more products [23]. The production of sesquiterpenes is usually lower than that of monoterpenes and diterpenes but has the highest structural diversity in coniferous trees. Due to difficulty in the structural identification of sesquiterpenes and the diversity of sesquiterpene synthase products, the physiology and biochemistry of sesquiterpene synthases have barely been studied in coniferous trees. Phylogenetic analysis based on protein sequences has divided gymnosperm and some angiosperm TPSs into seven subfamilies: TPS-a, TPS-b, TPS-c, TPS-e/f, TPS-g, and TPS-h [24,25,26,27,28]. The TPS-d subfamily is entirely composed of TPSs from gymnosperms that are involved in primary and secondary metabolism [21]. Some TPS genes in P. abies, P.sitchensis, P. glauca, Abies grandis, Pseudotsuga menziesii, P. taeda, P. contorta, and P. banksiana have been identified, cloned, and functionally verified [29,30,31,32,33,34].

Monochamus alternatus is a wood-boring insect, widely distributed from East to South Asia, including China, Japan, Korea, Laos, Taiwan, Vietnam, and some other parts of the world, including Canada, Denmark, and the United States [35]. The larvae of this beetle feed and pupate in the phloem and xylem of host plants, the newly emerged adults mate after pre-mating feeding, and the female adults spawn in the ovipositing incisions [36] (Figure 1). M. alternatus can spread pine wood nematodes (Bursaphelenchus xylophilus), which are the causal agent of pine wilt disease, during adult feeding [37,38]. P. massoniana is an important tree species widely distributed in southern China because of its high tolerance of acidic soils and high oleoresin yield, with a distribution area of about 20 million square kilometers in China. However, P. massoniana is the preferred host of M. alternatus and is seriously damaged by pine wilt disease. Since the disease was first discovered in China in 1982, it has killed billions of pine trees, causing huge economic losses, and is the main threat to the healthy pine forests widely distributed in China [39].

In our previous study, the terpene-based defense of P. massoniana was induced by M. alternatus adult feeding. Transcriptome sequencing and paraffin sections showed that most TPSs of P. massoniana showed increased expression, which was followed by the formation of traumatic resin ducts and release of terpenoids [18,40]. Among the released volatile terpenoids, monoterpenoids and sesquiterpenoids had the highest contents, including α-pinene, β-phellandrene, β-pinene, β-myrcene, camphene, caryophyllene, and longifolene. In addition, many diterpenes are released. We also proved that methyl jasmonate induction significantly promoted the release of volatile terpenoids from P. massoniana [41]. At the same time, the pre-induction of methyl jasmonate significantly reduced the feeding area of the M. alternatus adult [41]. These findings indicate that the terpenoids are important secondary metabolites of P. massoniana in response to M. alternatus adult feeding, and TPSs are involved in the biosynthesis of terpenoids. Terpenoids play a major role in M. alternatus host and mate localization, as well as in chemical communication among the pine wood nematode, M. alternatus, and P. massoniana [42,43,44]. Some TPSs in P. massoniana have been identified, such as Pm GGPPS2, which affects the synthesis of diterpenoids [45], as well as Pm TPS4 and Pm TPS21, which regulate the synthesis of α-pinene and longifolene and participate in the defense against pine wood nematodes [46]. Nevertheless, the links between the structures of TPSs and their products in P. massoniana remain unclear.

In this study, we first modelled the structures of six TPSs in P. massoniana that had been found to be significantly induced by M. alternatus adult feeding, including four mono-TPS genes, TPS (+)-α-pinene, TPS (−)-α-pinene, TPS (−)-β-pinene, and TPS limonene, and two sesqui-TPS genes, TPS selinene and TPS longifolene, after cloning and analyzing their sequences. Prokaryotic expression was then performed for monoterpene synthesis to determine their final products. The results from this study will help reveal the potential correlations between the structures of TPSs and their corresponding terpene products, which will provide a theoretical basis for the transgenic breeding of M. alternatus resistant Masson pines to control pine wilt disease.

2. Materials and Methods

2.1. Plant and Insect Materials

A two-year-old P. massoniana seedling from Pingxiang City, Jiangxi Province (26°76′ N, 114°28′ E) in China was planted individually in a greenhouse (26 ± 0.5 °C, relative humidity = 70 ± 5%, 16:8 h light/dark photoperiod) with nutrient medium (humus/turf soil/perlite, 3:1:1 by volume) and weekly deep watering [40]. The third-instar larvae of M. alternatus were collected from P. massoniana host trees in Quanjiao County, Anhui Province, China (41°31′ N, 117°74′ E). Subsequently, they were reared individually on an artificial diet at a stationary temperature of 26 ± 0.5 °C, a relative humidity of 60 ± 5%, and a 16:8 h light/dark photoperiod, as described by Chen Rui-Xu et al. [40]. Newly emerged adults were collected and individually kept in 250-mL glass flasks, provisioned with fresh P. massoniana twigs as food in climate-controlled rooms (26 ± 0.5 °C, relative humidity = 60 ± 5%, 16:8 h light/dark photoperiod). Adult beetles (mixture of male and female), visually estimated to be of similar physiological conditions, were chosen, and starved for 24 h before use. Finally, the plant materials were authenticated by Prof. Kongshu Ji from Nanjing Forestry University, China. The voucher specimen (NF1001839) was deposited at Dendrological Herbarium, Nanjing Forestry University (Institution code from Chinese Virtual Herbarium).

2.2. Sampling

The feeding treatment was administered according to Chen Rui-Xu et al. [18]. Healthy P. massoniana seedlings under similar growth conditions were selected and divided into treatment and control groups. The starved adults of M. alternatus were caged individually on a seedling stem using a wire mesh (25 cm × 25 cm, hole size = 4 mm²) surrounding the stem of P. massoniana seedlings. After feeding for 24 h, the wire mesh and insects were removed (N = 3). Subsequently, the seedlings of the treatment and control groups were sampled. The stems of P. massoniana, 5 cm above and below the feeding center of M. alternatus, were cut into small sections (1–2 cm long), transferred to ribonuclease-free centrifuge tubes, and quickly frozen in liquid nitrogen. The stems were stored at −80 °C.

2.3. RNA Extraction and Gene Sequence Analysis

Pm TPS gene sequences and corresponding gene-specific primers were derived from the previously determined M. alternatus feeding transcriptome [18] (BioProject accession number PRJNA328403). Four monoterpene synthase genes, unigene 26781, unigene 36618, unigene 26791, and unigene 18558, and two sesquiterpene synthase genes, unigene 380 and unigene 12384, were selected for amplification of the coding sequence (CDS). The total RNA was extracted from the sections of P. massoniana using a Miniprep RNA Purification Kit (TIANGEN, Beijing, China) and reverse-transcribed to synthesize complementary DNA (cDNA) using a 5 × All-In-One RT Master Mix (Accurate Biology, China). The SMARTer^® Rapid Amplification of cDNA Ends 5′/3′ Kit (Clontech Laboratories, Inc., Mountain View, CA, USA) was used for 5′ or 3′ rapid amplification of cDNA via polymerase chain reaction to obtain the complete sequences. The amplified products were recovered using the MiniBEST Agarose Gel DNA Extraction Kit Ver.4.0 (Takara Bio Inc., Otsu, Japan). The amplified sequences were ligated to a pCE2 TA/Blunt-Zero vector (Vazyme, Nanjing, China) to identify full-length sequences. The full-length sequences of Pm TPS (+)-α-pinene, Pm TPS (−)-α-pinene, Pm TPS (−)-β-pinene, Pm TPS limonene, Pm TPS selinene, and TPS longifolene were submitted to GenBank (accession numbers MN652001.1, MN652002.1, MN652003.1, MN652005.1, MN652006.1, MN652007.1). The basic local alignment search tool at the National Center for Biotechnology Information was used to identify sequences homologous to Pm TPSs (https://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed on 13 January 2022).

2.4. Bioinformatics Analysis

The protein sequence alignment was conducted using DNAMAN 6.0 software (Lynnon Biosoft, San Ramon, CA, USA. Version number 6.0.3.99). The online platform ExPASY https://web.expasy.org/compute_pi/, accessed on 1 December 2021) was used to calculate isoelectric point, molecular weight, and hydrophobic/hydrophilic prediction. In addition, the online tool TargetP 1.1 (http://www.cbs.dtu.dk/services/TargetP/, accessed on 1 December 2021) was used to predict the amino-terminal (N-terminal) signal peptide sequence. The coding sequences of 85 monoterpene and sesquiterpene synthases from pine trees were downloaded from the NCBI database (Table S1) and aligned using Clustal W software (https://www.genome.jp/tools-bin/clustalw, accessed on 13 December 2021). MEGA 7.0 software (Version 7.0.26) was used to construct the phylogenetic tree as per the maximum likelihood method with 500 bootstrap replicates [47].

SOPMA online software (https://npsa-prabi.ibcp.fr/cgi-bin/npsa_automat.pl?page=npsa_sopma.html, accessed on 26 December 2021) was used to analyze and predict the secondary structures of Pm TPS proteins. In addition, the online software MEME (Multiple EM for Motif Elicitation) (https://meme-suite.org/meme/tools/meme, accessed on 11 January 2023) was used to predict and analyze the motifs of Pm TPS genes. The motif width was set to 2–300, and the number of motifs was set to 20. The amino acid sequences of 91 terpene synthases in the phylogenetic tree were used for motif analysis.

2.5. Homology Modeling and Molecular Docking

Homology models of six Pm TPSs were built using the Swiss-Model online website (https://swissmodel.expasy.org, accessed on 11 January 2022) based on the tertiary structure of α-bisabolene synthase from A. grandis (protein data bank (PDB)no. 3sae.1. A) and taxadiene synthase from Taxus brevifolia (PDB no. 3p5r.1. A). Modeled structures were confirmed to be of high quality, with Ramachandran plot statistics exceeding 90% using PROCHECK [48]. In addition, the substrates GPP and FPP were docked with the model structures of Pm mono-TPSs and sesqui-TPSs, respectively, using SYBYL-X software version 2.0 (Certara, Princeton, NJ, USA).

2.6. Expression and Purification of Recombinant Pm Mono-TPSs

Six Pm TPS-coding sequences were amplified as described above, and a homologous recombination insertion into the pET-32a (+) vector was employed. Homologous recombinant plasmids were verified via sequencing and transferred into BL21 (DE3) competent cells. Luria–Bertani medium (1 L) containing 100 μg/mL Amp was inoculated with five individual colonies and cultured until an optical density at 600 nm (OD600) of at least 0.8 was reached at 37 °C and 220 rpm. Isopropyl-β-D-thiogalactopyranoside was added to a final concentration of 0.5 mM and induced to express at 20 °C and 100 rpm for 12 h. Cultures were collected via centrifugation at 13,523× g for 15 min. The bacterial cells were collected and resuspended in 100 mL resuspension buffer (50 mM PB, 300 mM NaCl, 20 mM imidazole, pH 7.4). Phenylmethylsulfonyl fluoride was added to a final concentration of 1 mM and ultrasonically broken in an ice bath (power 200 W, work 3 s, pause 5 s, for 10 min per 30 mL of resuspended bacterial cells). After centrifugation at 13,523× g for 30 min, the supernatant and precipitate were collected (dissolved in 8 M urea). Subsequently, 20 μL of each was added to 5 μL loading buffer, kept in a 100 °C water bath for 10 min, and then centrifuged for a short time. Sodium dodecyl-sulfate polyacrylamide gel electrophoresis was performed on 7.5% gel (100 V, 120 min). The protein gel was stained with Coomassie brilliant blue and photographed. Recombinant Pm mono-TPS proteins were purified to homogeneity by Ni²⁺-affinity chromatography and then desalted at 4 °C for 16 h in a dialysis bag (14 kDa) with a buffer composed of 25 mM 4-(2-hydroxyethyl)-1-piperazineethanesulphonic acid, 100 mM NaCl, and 1 mM dithiothreitol (DTT). The dialysis target protein was frozen at −80 °C for 4 h and treated in freeze-drying apparatus for more than 24 h until the protein was completely dried. Dried terpene synthase protein was diluted to approximately 5 mg/mL using ddH₂O.

2.7. TPS Enzyme Assays and Gas Chromatography–Flame Ionization Detection/Gas Chromatography–Mass Spectrometry (GC–FID/GC–MS) Analysis

The TPS enzyme activities were tested according to a previously published procedure (Martin et al. 2002), with modifications. Before the assay, the frozen proteins were placed at 37 °C until thawing. The diluted proteins were pre-equilibrated with assay buffers (25 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (pH 7.4), 100 mM KC1, 10 mM MgCl₂, 5% (v/v) glycerol, and 5 mM DTT). Enzyme activity was assessed using 1 mL of the pre-equilibrated proteins with the addition of 25 μL GPP (50 mM, with 50% v/v methanol, perform nitrogen purge before use). All enzyme assays were overlaid with 800 µL of hexane (Aladdin Biological Technology Co., Ltd., Shanghai, China) and 400 ng/mL isobutyl benzene (Aladdin Biological Technology Co., Ltd.) to collect the released volatiles, then incubated at 37 °C, 80 rpm for 1 h, 3 h, and 5 h. The upper layer hexane (200 µL) was taken into a sample vial for GC–FID/GC–MS analysis. Chromatographic grade n-hexane, (+)-α-pinene, (−)-α-pinene, (−)-β-pinene, R-(+)-limonene, linalool, and myrcene standard were purchased from Aladdin Biological Technology Co., Ltd.

The Pm mono-TPS products were analyzed on a GC–FID fitted with an HP-5 column (30 m × 0.25 mm × 0.25 μm, Agilent Technologies Inc., Santa Clara, CA, USA). The injection volume was 1 μL of the hexane extract. The oven temperature was programmed from an initial temperature of 50 °C for 3 min to 250 °C at increments of 10 °C min⁻¹. The GC–MS analysis was performed using the same program as that for GC–FID. The MS conditions were obtained as follows: an ion source, 230 °C; electron energy, 70 eV; and a scan range, 50–500 mass units; helium was used as the carrier gas, 1.0 mL/min. Molecule identification was performed using the National Institute of Standards and Technology 14 library. The results were confirmed by comparing the retention times of target monoterpenes with authentic standards, including (+)-α-pinene, (−)-α-pinene, (−)-β-pinene, R-(+)-limonene, linalool, and myrcene (Aladdin Biological Technology Co., Ltd.). If available, identification was confirmed by the Kovats index or authentic standards [49]. The C₇-C₄₀ saturated alkanes were purchased from Sigma-Aldrich (St. Louis, MO, USA).

3. Results

3.1. Sequence, Phylogenetic, and Motif Analysis of Six Pm TPSs

To characterize the terpene synthase genes involved in the M. alternatus feeding induced defense in P. massoniana, six genes, Pm TPS (+)-α-pinene, Pm TPS (−)-α-pinene, Pm TPS (−)-β-pinene, Pm TPS limonene, Pm TPS selinene, and Pm TPS longifolene, which were differentially expressed after M. alternatus inoculation, were identified in the P. massoniana transcriptome [38]. The BLAST alignment showed hits with known TPS genes based on the presence of conserved sequence characteristics shared by all TPSs. The six predicted proteins encoded by the above Pm TPSs ranged between 580–635 amino acids in size, with predicted pIs of 5.71, 5.53, 5.74, 5.44, 5.6, and 5.24, respectively. All Pm mono-TPSs had signal peptide sequences at the N-terminus, ranging from 36–48 aa in length, and localized to chloroplasts (Table S2). In contrast, Pm sesqui-TPSs did not contain signal peptide sequences (Table S2). Hydrophobic/hydrophilic analyses showed that all six Pm TPSs were hydrophilic (Figure S1).

As shown in Figure 2a, the mono-TPSs Pm TPS (+)-α-pinene, Pm TPS (−)-α-pinene, Pm TPS (−)-β-pinene, Pm TPS limonene and the sesqui-TPS Pm TPS selinene contain the arginine-tryptophan motif RRX₈W, which is conserved among most mono-TPSs and some sesqui-TPSs, near the N-terminus. The RRX₈W motif was missing in Pm TPS longifolene. The highly conserved aspartate-rich motif DDXXD and RXR motif were present in these six Pm TPSs. However, another conserved motif, DTE, was not found in P. massoniana. This is crucial for chelating divalent cations, typically Mg²⁺, in the carboxyl-terminal (C-terminal) domain [50]. Phylogenetic analysis revealed that four monoterpene and two sesquiterpene synthases of P. massoniana were clustered into two subfamilies (TPS-d1 and TPS-d2) with monoterpene and sesquiterpene synthases from other pine plants, except for P. Abies TPS E (E-α-farnesene), P. engelmannii × P. glauca TPS E (E-α-farnesene synthase/E-β-ocimene), and P. teada TPS (α-farnesense), which formed into one clade as monoterpene synthases, in line with previous studies [25,29,51,52] (Figure 2b). Among them, Pm TPS (+)-α-pinene grouped with α-pinene of P. kesiya. Furthermore, these were grouped with TPS α-pinene of P. contorta and P. banksiana. The Pm TPS limonene was grouped with P. sitchensis TPS (−)-limonene and P. abies TPS (−)-limonene. The Pm TPS (−)-β-pinene clustered with TPS (−)-β-pinene and TPS α-terpineol from several Pinus spp. Two sesqui-TPSs, Pm TPS longifolene and P. sylvestris TPS longifolene, were grouped, whereas Pm TPS selinene was grouped with P. sylvestris TPS caryophyllene/humulene (cary/humu), and slightly distant from the A. grandies δ-selinene synthase, annotated as monofunctional but heterologous.

The motif types and permutations of P. massoniana (Figure 2c) and its homologous TPSs (Table S3) were analyzed using the online software MEME (Version 5.5.5). The number and species of the 20 predicted motifs ranged from 11 to 50 aa in length. All Pm mono-TPSs contained 20 motifs, except for Pm TPS (−)-β-pinene. Thirteen motifs (motifs 1, 2, 3, 4, 5, 7, 8, 9, 10, 11, 12, 14, and 17) were found in all Pm TPSs. The conserved domains DDXXD, RRX₈W, and RXR were found in motifs 1, 11, and 3, respectively. Motifs 15, 18, 19, and 20 were found only in the Pm mono-TPSs but not in the Pm sesqui-TPSs. Most of the Pm mono-TPSs contained all 20 motif structures in the order 18-11-12-20-7-13-8-19-9-5-17-3-15-1-6-4-2-14-10-16. Motifs 1, 5, 7, 8, 11, and 12 appeared in all sequences. None of the Pm sesqui-TPSs contained motifs 18 or 20. The motifs of TPS D1 and TPS D2 had their characteristics. In particular, the clade of four farnesene and three bisabolene synthases had a long, non-conservative region between motifs 12 and 7, which was not found at the corresponding sites in other clades (Figure S2).

3.2. Prediction of the Secondary Structure of Pm TPSs

The secondary structure of Pm TPS amino acid sequences was predicted using SOPMA software (Figure 3a). The main secondary structures of the six Pm TPSs were α-helices, accounting for 63.27%, 63.91%, 63.04%, 63.62%, 68.95%, and 68.28%, respectively (Table S4). The α-helix content of sesqui-TPSs was slightly higher than that of mono-TPSs. The random coil contents were 26.66%, 26.22%, 29.25%, 30.56%, 29.45%, 24.01%, and 23.79%, respectively. The content of β-sheet/corner (2.92%–4.14%) and extension chain (3.18%–5.52%) was the lowest. The β-sheet/turn and extended strands are interspersed in the α-helix structure, and random coils connect all the structures. Owing to the presence of a signal peptide, the α-helix content in the first 100 amino acids of Pm mono-TPSs was less than that of Pm sesqui-TPSs. All Pm mono-TPSs had a higher content of extended strands than Pm sesqui-TPSs, except for Pm TPS (−)-β-pinene.

SMART software was used to analyze the functional domains of Pm TPSs (Figure 3b). Two Pfam domains, terpene-synth and Terpene-synth-C, have been characterized. The Terpene-synth domain of Pm mono-TPSs starts from the N-terminal 73–75 aa region and ends at the 264–270 aa region, while the terpene-synth-C domain starts from the N-terminal 302–312 aa region and ends at the 492–577 aa region. In Pm sesqui-TPSs, the terpene-synth domain starts from 18/27 and ends at 211, while the terpene-synth-C domain starts from 247/257 and ends at 512/523. In addition, there is a low-complexity region in the N-terminus of Pm TPS (+)-α-pinene, as shown in purple. The sequences were SIVPSMSMSSTTSVS and SMSMSSTTSVS, both serine-rich.

3.3. Homology Modeling and Molecular Docking of Pm TPSs

The predicted Pm TPS structures, as shown in Figure 4, revealed that Pm TPSs comprised two structural domains, the N- and C-terminal, with each domain consisting of several α-helices, β-sheets/turns, and random coils/loops. The C-terminal domain forms 9–17 α-helices from helix A to helix K. The structures of helix C, helix D2, and helix H1 in Pm TPS (+)-α-pinene, Pm TPS (−)-β-pinene, and Pm TPS limonene had 2–3 fewer amino acid residues than those in Pm TPS (−)-α-pinene, Pm TPS selinene, and Pm TPS longifolene. Helix J is presented as a complete long helix (Pm TPS (+)-α-pinene, Pm TPS (−)-β-pinene, and Pm TPS limonene) or equally divided into two short helices (Pm TPS (−)-α-pinene, Pm TPS selinene, and Pm TPS longifolene), according to the template. Other α-helices were conserved in all Pm TPSs.

The results of Ramachandran plot evaluation (Figure S3) revealed that the structural score of Pm TPSs obtained via this modeling were of a high-quality model: 90.7%–94.8% of residues were found in the most favored region, 4.1%–8.2% of residues were found in additionally allowed regions, and 0.4%–1.5% of residues were found in generously allowed and disallowed regions. This justified the stability without any steric hindrance and reliance on the predicted model with acceptable qualities. All TPSs formed active pockets at the C-terminus, where GPP and FPP molecules form hydrogen bonds with different amino acid residues. The docking experiment showed that GPP and FPP could bind to monoterpene and sesquiterpene synthases, respectively. The confidence scores of the six terpene synthases were 4, 4, 4, 4, 3, and 2 for Pm TPS (+)-α-pinene, Pm TPS (−)-α-pinene, Pm TPS (−)-β-pinene, Pm TPS limonene, Pm TPS selinene, and Pm TPS longifolene, respectively. The total scores were 7.23, 7.24, 6.31, 7.68, 7.32, and 5.98 for Pm TPS (+)-α-pinene, Pm TPS (−)-α-pinene, Pm TPS (−)-β-pinene, Pm TPS limonene, Pm TPS selinene, and Pm TPS longifolene, respectively.

Docking simulation results are shown in Figure 5. The side chains of aspartic acid (Asp) 484 and tyrosine (Tyr) 555 formed three hydrogen bonds and one hydrogen bond with the diphosphate group of GPP, respectively, in the substrate-binding pocket in Pm TPS (+)-α-pinene. The amino acids Asp 484 and Tyr 555 are distributed in helices H2 and J, respectively. For Pm TPS (−)-α-pinene, four residues in the binding pocket form hydrogen bonds with GPP: the amino acid glutamine (Glu) 410 in the middle of helix F, the amino acid Glu 489 in the H2-α1 loop, the amino acid serine (Ser) 432 in the G1-G2 loop, and the amino acid arginine (Arg) 473 in the middle of helix H2. The amino acid Arg 473, Glu 410, Glu 489, and Ser 432 formed three, two, two, and one hydrogen bonds with the diphosphate group of GPP, respectively. For Pm TPS (−)-β-pinene, only hydrogen bonds were formed between the side chain of the amino acid threonine (Thr) 559, which is distributed in helix K, and the diphosphate group of GPP. For Pm TPS limonene, five residues in the binding pocket formed hydrogen bonds with GPP: cysteine (Cys) 343 and Asn 346 in helix D, the amino acid Tyr 319 in helix C, the amino acid Tyr 425 in helix F, and the amino acid Ser 451 near the end of helix G1. Each of the above amino acids formed one hydrogen bond with the diphosphate group of GPP.

The side chain of two closely related amino acids, Asp 321 and Asp 325 in helix D, formed a hydrogen bond with the diphosphate group of GPP in the substrate-binding pocket of Pm TPS selinene. For Pm TPS longifolene, four residues in the binding pocket formed hydrogen bonds with FPP: Asp 331 in helix D, the amino acid Ser 433 near the end of helix G1, and Arg 473 and lysine (Lys) 484 in helix H2. The amino acid Arg 473, Asp 331, Ser 433, and Lys 484 formed three, two, one, and one hydrogen bonds with the diphosphate group of FPP, respectively.

3.4. Heterologous Expression and Functional Verification of Pm Mono-TPSs

All cDNAs of Pm TPSs cloned into the pET32 (+) vector with thioredoxin and histidine (HIS) tags were successfully expressed in E. coli DH5α competent cells, and recombinant proteins were then purified with His 60 Ni Superflow Resin and Gravity Columns (Figure S4). Following an in vitro enzyme activity assay, Pm TPS (+)-α-pinene was confirmed to encode a mono-TPS enzyme that catalyzes the formation of mostly (+)-α-pinene, β-pinene, linalool, and a few minor R-(+)-limonene, as detected by GC–FID and GC–MS analysis (Figure 6, Figures S5 and S6, Table S5). Pm TPS (−)-α-pinene catalyzed the formation of mostly (−)-α-pinene, as well as myrcene, R-(+)-limonene, and linalool. Pm TPS β-pinene catalyzed the formation of mostly β-pinene and linalool, as well as a few (+)-α-pinene and R-(+)-limonene. Pm TPS limonene catalyzed the formation of mostly R-(+)-limonene and a few (−)-α-pinene and linalool. The GC–MS analysis of the reaction products catalyzed by Pm TPS (+)-α-pinene with GPP as a substrate showed that the reaction product concentration of (−)-α-pinene was 120 μg/mL after 1 h of reaction, whereas that of Pm TPS limonene, Pm TPS (+)-α-pinene, and Pm TPS β-pinene were 6, 1, and 1.36 μg/mL, respectively (Table S5). With the extension of the incubation time, the main product of Pm mono-TPSs varied and increased during 1–3 h of the in vitro reaction, except for Pm TPS (−)-β-pinene. Within 3–5 h of incubation, the concentrations of the main products of Pm TPS (−)-α-pinene decreased from 125.6 to 114 μg/mL, whereas the concentrations of Pm TPS (+)-α-pinene and Pm TPS limonene increased from 2.43 and 6.67 μg/mL to 2.6 and 8.57 μg/mL. The by-products of Pm TPS (+)-α-pinene, linalool, and (−)-β-pinene were detected after 1 h and 3 h of incubation, respectively, and gradually increased to 2.333 and 0.393 μg/mL after 5 h of incubation. Limonene was detected after 5 h of incubation. Linalool and (−)-α-pinene, the by-products of Pm TPS (−)-β-pinene, were detected after 1 h of incubation, but the content of linalool exceeded that of the main product β-pinene after 3 h of incubation. The content of (−)-α-pinene gradually increased to 0.67 μg/mL after 5 h of incubation but was lower than that of the main product (−)-β-pinene at 1.03 μg/mL. Limonene was detected after incubation for 5 h. The levels of limonene and (+) -α-pinene, by-products of Pm TPS (−)-α-pinene, accounted for approximately 5% of the main product and were detected after 1 h of incubation. The levels of (−)-α-pinene and linalool, two by-products of Pm TPS limonene, were lower than those of Pm TPS (−)-α-pinene by-products, and were detected after 1 and 3 h of incubation, respectively.

4. Discussion

In this study, we identified six TPS genes in P. massoniana, including four mono-TPS and two sesqui-TPS genes. We believe that these genes are involved in the response of P. massoniana to M. alternatus invasion. In vitro enzyme activity assay also verified that Pm mono-TPSs can catalyze substrate GPP to produce volatile terpenes. In our previous study [18,38,39], we proved that exogenous MeJA treating and M. alternatus adult feeding could induce a resin-based defense of P. massoniana, including secondary resin duct formation, induced resin terpenoid accumulation, and gene transcriptional changes related to terpenoid synthesis in metabolic pathways (Figure 7). TPSs catalyze complex carbocation cascade reactions on the prenyl diphosphate substrate, resulting in cyclic or linear terpene backbones [53], which determines the formation and composition of oleoresin terpenes. Therefore, studying the relationship between TPS structure, action site and production is helpful to elucidate the mechanism behind the synthesis of resin with complex composition, as well as the understanding, development and application of terpene synthase in vitro terpene biosynthesis.

The total length of Pm TPSs was similar to that reported in Pinaceae [54]. Each Pm mono-TPSs contained a peptide sequence of approximately 50 amino acid residues, similar to SaTPS1 identified from Santalum album [55]. Studies have shown that a lack of the N-terminal peptide FaNES1 leads to changes in subcellular localization, resulting in different products [56]. The Pm sesqui-TPSs had no peptide sequence, suggesting that it may function in the cytoplasm, whereas the sesqui-TPSs from Solanum habrochaites (santalene/bergamotene synthases) are located in chloroplasts, indicating that sesquiterpenes can also be synthesized in plastids [57]. Among the three characteristically conserved domains of TPS, RxR and DDxxD were found in all the sequences. The RRx8W domain was found not only in monoterpenes but also in Pm TPS selinene. Monoterpene synthases lacking the RRx8W motif are responsible for the production of acyclic mono-TPSs, a characteristic of the TPS-g subfamily [58]. Previous studies on the functional verification of menthol limonene synthase have shown that the tandem arginine sequence RR in RRx8W might catalyze the transformation of GPP to an intermediate product [59], suggesting that Pm TPS selinene can bind the substrate GPP. Additionally, in A. grandis, RRx8W was also found in the N-terminus of δ-selinene synthase, and its heterologous expression could produce 52 sesquiterpenes [32]. The conserved domain NSE/DTE was not found in Pm TPSs nor the reported TPS sequences of P. taeda, P. contorta, P. banksiana, and P. nigra subsp. laricio [22,34,60], suggesting that the conserved domain NSE/DTE may not be conserved in Pinus.

In the phylogenetic tree, we found that the amino acid sequences of TPSs in congeners were more similar to those of TPSs with the same main products in species from different genera (Figure 2b). For example, the distance between Pm TPS seli (Pm TPS selinene) and A. grandis TPS d-seli, both of which produce selinene, was greater than the distance between Pm TPS seli and Pinus sylvestris TPS cary/humu, which are congeners but have different main products. This feature is also found in other species. The similarity between caryophyllene synthase from Oryza sativa, Arabidopsis thaliana, Artemisia annua, and Cucumis sativus is less than 40%. In contrast, caryophyllene synthase from Oryza sativa has 40%–42% similarity with the synthases that produce other sesquiterpenes from Zea mays [61]. No special motifs were found in the sesquiterpene sequences used to construct the phylogenetic tree. None of the Pm sesqui-TPSs contained motifs 18 or 20. Picea abies TPS E, E-alpha-far, Pinus taeda TPS alpha-far, and Picea engelmannii × Picea glauca TPS (E, E)-alpha-far/(E)-beta-oci in TPS D1 contained an additional motif 15 compared to the other sesqui-TPSs. The absence of these motifs suggests a potential difference in the evolution of mono- and sesqui-TPSs. Furthermore, in the phylogenetic tree, motif 19 was absent in all sesqui-TPS sequences of Pinus, including P. massoniana. We hypothesize that this may provide evidence for the functional differentiation of TPSs occurring before species differentiation.

Previous study has shown that the catalytic function of TPSs is affected by kinetics rather than thermodynamics, meaning that the types of intermediates initially formed in the reaction can affect the formation of subsequent products [62]. The formation of multiple centromeres in the active center can generate a variety of substrates, leading to the formation of by-products in fewer reaction steps [63]. Molecular docking simulations showed that the substrate-binding pockets of sesqui-TPSs were larger than those of mono-TPSs. The substrate of sesqui-TPS is FPP (C15), which contains three double bonds and is more complex than GPP (C10), which contains two double bonds [64]. Because of the increase in the chain length of the substrate, the number of sesquiterpene intermediates is much greater than that of monoterpene intermediates, which enriches the diversity of product structures [65]. Larger pockets may facilitate the entry of substrates and the formation of products. Secondary metabolic enzymes tend to produce mechanistic elasticity during evolution, giving rise to new enzymes with a wider range of substrates and a variety of new products [66].

Some residues are responsible for catalytic elasticity within and around the active sites of TPSs. Changing these residues can regulate the products or catalytic rates [67,68,69]. For example, Xu Jin-Kun et al. [70] reported that Tyr residues (Y573) in S-limonene synthase may be involved in the earlier steps of the reaction, probably by controlling the conformation of the helical linalyl diphosphate intermediate. The replacement of Ile in A. thaliana ent-kaurene synthase with Thr can convert it into a pimaradiene synthase, but the catalytic efficiency is reduced by four times [71]. The replacement of nine amino acids in Nicotiana tabacum can accomplish conversion between the catalytic specificities of 5-epi-aristolochene synthase and premnaspirodiene synthase [68]. Molecular docking between Pm TPSs and the corresponding substrates showed that there were seven kinds of residues in Pm mono-TPSs participating in the catalytic process, Asp, Arg, Tyr, Glu, Trp, Thr, and Cys, which were distributed between helix C and helix K. Asp, Try, Arg, and Lys were found in Pm sesqui-TPSs, distributed between helix D and helix H2, similar to reports from A. grandis [31]. Mn²⁺ is a common catalyst in biological catalytic reactions, often combining with Asp [72], which may be a reason for the elevated catalytic efficiency of some TPSs in the presence of Mn²⁺ [73]. Thus, adding Mn²⁺ may accelerate the production of Pm (+)-α-pinene, Pm selinene, and Pm longifolene. At a pH ≥ 7, Asp and Glu are deprotonated, facilitating the formation of a carbonium ion intermediate. Therefore, when using Pm mono-TPSs in the catalytic reaction, it is critical to maintain a pH above 7 for stable production. The surfaces of the aromatic groups in Tyr and Trp are positively charged, which can stabilize the carbonium ion intermediate [62]. Because Pm sesqui-TPSs have fewer aromatic groups than Pm mono-TPSs, there may be more binding possibilities for the structures of the intermediates and products in Pm sesqui-TPSs than that in Pm mono-TPSs.

During expression vector construction, a purified tag was inserted at the N-terminus to avoid the influence of redundant tag sequences at the C-terminal active site during protein activity tests. The non-removal of peptide sequences in this study did not affect protein expression and in vitro activity tests, as in other studies [22,61,74,75,76,77]. Previous studies have shown that 1–1.5 h is the optimal reaction duration for collecting enzymatic reaction products [22,34]. In this study, most of the main products were produced within 1 h, whereas longer reaction durations only significantly increased the number of by-products. Therefore, under a consistent metal ion concentration and incubation temperature, it is recommended to complete the reaction of TPSs within 1–1.5 h. The products of mono-TPSs mainly possess chirality, with several trace products that have a related stereo-structure [78]. Heterologous expression of monoterpene synthases reported in Pinus is usually named after their main products, such as (+)-α-pinene synthase and (−)-α-pinene synthase from P. taeda, (+)-α-pinene synthases, (−)-α-pinene synthases, and (−)-β-pinene synthase from P. contorta and P. banksiana, as well as (−)-α-pinene synthase and β-pinene synthase from Pinus elliottii [22,34,79]. In this study, the main product profiles of the Pm TPSs were similar to those reported above but at different concentrations. Meanwhile, as observed in the TPSs of P. taeda [22], most Pm TPSs only generated certain enantiomers.

Four Pm mono-TPSs were heterologously expressed. Among them, Pm (−)-α-pinene produced the largest number of products and formed the largest number of hydrogen bonds with GPP. The varied amounts of by-product linalool produced by Pm mono-TPSs suggest that Pm (−)-α-pinene may have a higher catalytic specificity than Pm (+)-α-pinene and Pm (−)-β-pinene. Therefore, the number of active sites and hydrogen bonds between the substrate and enzyme may be positively related to product specificity. Chen Rui-Xu et al. [40] analyzed the volatiles produced by P. massoniana fed on by M. alternatus adults. The ratio of β- pinenes to α-pinenes in the volatiles was approximately 1:100, which is very close to the ratio of these two products in in vitro activity tests. However, the chemical composition of the products in the in vitro activity tests differed from that of the volatiles. This may be due to the catalytic functions of downstream enzymes, such as P450 enzymes, which in plants catalyze the transformation of some TPS products to other compounds [80]. For example, limonene in Mentha canadensis can be converted into (−)-carvone or (−)-isomentadiene by two P450 enzymes [81]. Notably, although the limonene synthase gene was upregulated in P. massoniana fed upon by M. alternatus adults, and limonene was the main heterologous expression product, this compound was not detected in volatiles. The limonene produced may have been converted to pinocarvone by downstream P450 enzymes [40].

5. Conclusions

Terpene synthases catalyze complex multistep reactions that generate hundreds of structurally diverse terpenoids of biological and commercial importance. In a previous study, we showed how defense mechanisms are triggered in P. massoniana through a complex series of changes, including traumatic resin duct formation, which induce terpenoid accumulation, key enzyme activation, and TPS gene expression [18,38]. In this study, we cloned six TPS genes from P. massoniana involved in these protective responses. We then subjected these genes to bioinformatic analysis and in vitro activity tests. Phylogenetic analysis indicated that the evolutionary distance between TPSs of the same species was greater than that between the same products within P. massoniana. Furthermore, we uncovered the structural features that may be responsible for the functional attributes of TPSs through molecular docking, verifying our findings via in vitro activity tests of Pm TPSs. Studying the catalytic mechanism of TPS could aid in understanding the structure–function relationship of TPSs in P. massoniana and provide a theoretical basis for resistant tree breeding through genetic modification.