A constitutive serine protease inhibitor suppresses herbivore performance in tea (Camellia sinensis)

Abstract Protease inhibitors promote herbivore resistance in diverse plant species. Although many inducible protease inhibitors have been identified, there are limited reports available on the biological relevance and molecular basis of constitutive protease inhibitors in herbivore resistance. Here, we identified a serine protease inhibitor, CsSERPIN1, from the tea plant (Camellia sinensis). Expression of CsSERPIN1 was not strongly affected by the assessed biotic and abiotic stresses. In vitro and in vivo experiments showed that CsSERPIN1 strongly inhibited the activities of digestive protease activities of trypsin and chymotrypsin. Transient or heterologous expression of CsSERPIN1 significantly reduced herbivory by two destructive herbivores, the tea geometrid and fall armyworm, in tea and Arabidopsis plants, respectively. The expression of CsSERPIN1 in Arabidopsis did not negatively influence the growth of the plants under the measured parameters. Our findings suggest that CsSERPIN1 can inactivate gut digestive proteases and suppress the growth and development of herbivores, making it a promising candidate for pest prevention in agriculture.


Introduction
Insect herbivores of tea plants are estimated to cause an 11%-55% loss in yield, which is worth US$500 million to US$1 billion worldwide [1].Genetically modifying plants to improve their resistance to insect pests is an environment-friendly strategy for pest control.One possible approach is the use of defense-related metabolites, such as protease inhibitors [2,45].Plant protease inhibitors can inhibit digestive proteases of insect herbivores and, consequently, reduce the uptake of necessary amino acids for herbivore growth and development [2,3].
Serine protease inhibitor (SERPIN) is one of the largest superfamilies of protease inhibitors in plants [4].Most SERPINs inhibit serine proteases, but some also inhibit cysteine proteases [5].In addition to their involvement in physiological processes, SERPINs play critical roles in defense against herbivores by suppressing the enzymatic activity in the insect alimentary canal, which inhibits the absorption of vegetable proteins [6].In past decades, a number of transgenic plants overexpressing SERPINs have been generated to test the increase in plant resistance against insect herbivores.For instance, AtSerpin1 inhibits cysteine and serine protease activities of several leaf herbivores in Arabidopsis (Arabidopsis thaliana).Transgenic Arabidopsis lines overexpressing AtSerpin1 dramatically reduce the larval growth of cotton leafworm (Spodoptera littoralis) [7].Expression of a barley SERPIN in tomato shows increased resistance to tomato leaf miner (Tuta absoluta) [8].Heterologous expression of two fused protease inhibitors, a maize SERPIN and a potato carboxypeptidase inhibitor, in rice enhances defense against the striped stem borer (Chilo suppressalis) [9].While plenty of SERPINs obtained from crop plants have been demonstrated to confer resistance to a wide variety of herbivores [7,[9][10][11][12], it remains unclear whether and how SERPINs inf luence herbivore resistance in tea plants.
Generally, inducible protease inhibitors function as part of plant active defense responses, as they are elicited upon herbivory-related signal perception [6].They are expressed and accumulated in specific tissues that are being or going to be attacked [8,13].This tight spatial-temporal regulation not only allows plants to respond specifically and efficiently, but also controls plant defensive investment in a cost-saving way [14,46].However, this transcriptional and posttranscriptional control of inducible protease inhibitors is more susceptible to environmental factors, and it can even provide an opportunity for adapted herbivores to avoid effective defenses, leading to an unstable resistance phenotype.For instance, the spider mite (Tetranychus evansi) manipulates the expression and activity of inducible protease inhibitors, thereby suppressing the defense responses of tomato plants [15].Thus, we argue that constitutively expressed protease inhibitors may be crucial or even more reliable in guaranteeing effective defenses against herbivores.Although many inducible protease inhibitors have been identified, there are limited reports available on the biological relevance and molecular basis of constitutive protease inhibitors in herbivore resistance.
Here, we identified a novel constitutively expressed serine protease inhibitor, CsSERPIN1 from tea plants.We characterized For the detailed statistical information, refer to Tables S1 and S2 (see online supplementary material).(f-i) Correlations between the expression of CsSERPIN1, CsSERPIN2, CsSERPIN3, and CsSERPIN4 genes, and larval weight gain.The Pearson's product-moment r, R 2 , and P values of correlations are indicated.its expression patterns under both abiotic and biotic stresses, and determined its inhibitory activities both in vitro and in vivo.We clarified how CsSERPIN1 protects plants against two destructive pests, tea geometrid (Ectropis obliqua) and fall armyworm (Spodoptera frugiperda) by using a reverse genetic approach.Our results suggest that CsSERPIN1 has great potential for exploitation in pest control in agriculture.

CsSERPIN1 may be involved in tea plant defenses against herbivores
To determine the potential roles of SERPINs in herbivore resistance, we first quantified the larval growth of a destructive pest, tea geometrid, on 33 tea accessions.We observed variation in herbivore performance among different tea accessions (Fig. 1a).We then searched the annotated SERPIN genes in the tea genome, and identified only four genes with an intact serpin domain (PF00079).These four SERPINs belong to the MEROPS inhibitor family I4, clan ID.Their serpin domain consists of nine α-helices and three β-sheets, along with a semi-conserved reactive center loop that contains the active site recognized by the target protease [6].The full-length sequences of the four identified SERPINs (referred to as CsSERPIN1, CsSERPIN2, CsSERPIN3, and CsSERPIN4) were cloned and validated by reverse transcription PCR and Sanger sequencing.Our phylogenetic analysis of these four SERPINs, along with 24 herbivore-related SERPINs from 14 plant species, revealed that CsSERPIN1 and CsSERPIN2 were homologous to AtSerpin1 in Arabidopsis (Fig. S1, see online supplementary material).We further quantified the transcriptional levels of these four SERPIN genes in the second and third fully expanded leaves of 33 tea accessions.These specific leaf positions were chosen as they were utilized for the above herbivore resistance assessment.We found that the expression level of CsSERPIN1, but not the remaining three SERPIN genes, was negatively correlated with herbivore growth (Fig. 1b-i; Tables S1 and S2, see online supplementary material).These data suggest that the expression of CsSERPIN1 is associated with the resistance to tea geometrid.

CsSERPIN1 is a constitutively expressed serine protease inhibitor
As many defense-related SERPINs are induced in vegetative tissues by insect herbivores and defense-related signaling [16][17][18], we first investigated whether CsSERPIN1 expression is inf luenced by herbivory using QRT-PCR.CsSERPIN1 transcript levels were not affected by herbivore attack or herbivore-related signaling molecules, including jasmonic acid (JA), salicylic acid (SA), ethylene, and abscisic acid (ABA) (Fig. 2a-f).We then challenged tea plants with a regularly encountered pathogen, Colletotrichum gloeosporioides, but found this pathogen infection did not inf luence CsSERPIN1 expression either (Fig. 2a and g).To further test whether abiotic stresses could alter the expression of CsSERPIN1, we treated tea plants with cold (4 • C) or heat (38 • C) stresses and examined CsSERPIN1 transcription levels.However, these abiotic stresses did not significantly change the expression of CsSERPIN1 (Fig. 2a and h-i).To rule out the possibility that the minimally affected expression of CsSERPIN1 was due to experimental false negatives, we measured the expression levels of eight genes: CsLOX7 [19], CsJAZ2 [20], CsUGT87E7 [21], CsACS1 [22], CsSnRK2.1 [23], CsCBF1 [24], CsHSP90 [25], and CsGSTU19 [26], which are known to exhibit a robust response to different treatments.We observed strong induction of these positive control genes under each treatment in our experiments (Fig. 2b-i).
To further consolidate the expression pattern of CsSERPIN1, we conducted a transcriptomic analysis of tea plants after herbivore attack using RNA sequencing.We found that CsSERPIN1, along with CsSERPIN3 and CsSERPIN4, were not affected at 3 h and 24 h after herbivore attack, while CsSERPIN2 was significantly induced at 24 h (Fig. 2j; Data S1, see online supplementary material).CsSERPIN1 exhibited significantly higher expression levels compared to the other 10 detected SERPINs, including CsSERPIN2, CsSERPIN3, and CsSERPIN4 (Data S1, see online supplementary material).We also observed strong induction of the positive control gene CsLOX7 at 3 h and 24 h after herbivory, which was consistent with the results above (Data S1, see online supplementary material).Therefore, CsSERPIN1 is abundantly expressed in tea plants, and its expression pattern under our assessed biotic and abiotic stresses is highly reliable and not a result of false-negative outcomes.
To exclude the effects of host genotype on CsSERPIN1 expression, we quantified its transcript levels in 33 tea accessions after herbivore attack.Consistently, CsSERPIN1 was not affected in each of the above accessions (Fig. S2, see online supplementary material).Thus, CsSERPIN1 might be a constitutively expressed SERPIN, and it is possible that its transcriptional regulation is minimally affected by biotic and abiotic stresses.

CsSERPIN1 is an active serine protease inhibitor in vitro
To investigate whether CsSERPIN1 has the inhibitory activity, we expressed it in Escherichia coli, and purified the protein using an affinity method (Fig. 3a).Next, we tested the inhibitory effects of purified CsSERPIN1 on the activities of two digestive proteases, chymotrypsin and trypsin, in vitro (Fig. 3b).Remarkably, trypsin activity was highly susceptible to CsSERPIN1, with inhibition ranging from 20-87% depending on the amount introduced.CsSER-PIN1 also caused more than 20% inhibition of chymotrypsin activity.These results indicate that CsSERPIN1 is an active protease inhibitor in vitro and that chymotrypsin is much less susceptible than trypsin to CsSERPIN1.

CsSERPIN1 inhibits larval performance and protease activity in vivo
To further evaluate the potential of CsSERPIN1 to inhibit larval digestive proteases, we introduced the purified CsSERPIN1 protein into an artificial diet.The tea geometrid larvae fed on the diets containing CsSERPIN1 gained much less weight and produced smaller pupae than those fed on control (Con) or maltose-binding protein (MBP)-containing diets (Fig. 3c and d).To investigate the underlying physiology, we dissected larvae at the end of the feeding assay and analysed the protease activity in their gut.We found the larvae fed on the diet incorporating CsSERPIN1 had much lower chymotrypsin and trypsin activities than those fed on the control diet (Fig. 3e).These findings suggest that CsSERPIN1 could inactivate larval digestive proteases in vivo, thereby suppressing larval performance.

Overexpression of CsSERPIN1 increases tea resistance to tea geometrid larvae
To investigate the role of CsSERPIN1 in plant defense against herbivores, we transiently overexpressed CsSERPIN1 in tea plants with Agrobacterium tumefaciens-mediated transformation.QRT-PCR analysis showed that the overexpression of CsSERPIN1 persisted for at least 5 days post infiltration (Fig. 4a).Furthermore, the larvae fed on CsSERPIN1 overexpression plants grew more slowly than larvae fed on control or empty vector (EV) transformed plants (Fig. 4b).CsSERPIN1, thus, enhances tea resistance to herbivores.

Heterologous expression of CsSERPIN1 increases herbivore resistance in Arabidopsis
As stable transformation methods are currently unavailable for tea plants, we heteroexpressed CsSERPIN1 in Arabidopsis to further assess its suppressive effects on herbivore growth.We verified the expression levels of CsSERPIN1 in two homozygous lines (L2 and L5) using QRT-PCR and immunoblot assays (Fig. 5a).These two transgenic lines showed indistinguishable growth and morphology from wild-type (WT) plants, as indicated by the measured parameters (Fig. 5b).Lepidopteran larvae, fall armyworms, were allowed to fed on the CsSERPIN1-heteroexpressed and WT plants, and the increase in larval weight was determined.The weight gain was significantly lower on both day 5 and day 7 for fall armyworm larvae that fed on the CsSERPIN1-heteroexpressed lines compared to those on WT plants (Fig. 5c).Additionally, we found that larvae feeding on CsSERPIN1-heteroexpressed plants had much lower gut trypsin and chymotrypsin activities than those feeding on WT plants (Fig. 5d).Taken together, these results suggest that the inhibitory effects of CsSERPIN1 on herbivores extend beyond tea plants.

CsSERPIN1 contributes to herbivore resistance variation of tea plants
We hypothesized that the growth variation of tea geometrid on the 33 accessions might be due to the altered gut protease activities mediated by CsSERPIN1.To test this, a group of tea geometrid larvae were allowed to feed on the three most resistant (ZX10, ZX19, and ZX48) and three most susceptible (ZX13, ZX16, and ZX18) accessions identified above, respectively, and the activities of trypsin and chymotrypsin in larval gut were quantified.The larvae feeding on resistant accessions consumed much fewer leaves compared to those feeding on susceptible accessions (Fig. 6a), indicating that our resistance phenotype assessment was sufficiently reliable.Consistent with our hypothesis, the gut trypsin and chymotrypsin activities were much higher after feeding on susceptible accessions compared to those on resistant accessions (Fig. 6b).Together, CsSERPIN1 is implicated in the herbivore resistance variation of tea plants.

Discussion
Inducible protease inhibitors have been shown to inf luence plant defenses against herbivores, but the role of constitutive protease inhibitors in herbivore resistance is still unclear.Our study helps fill these gaps by characterizing a novel constitutively expressed serine protease inhibitor, CsSERPIN1, which inactivates digestive proteases in vitro and in vivo, and consequently, suppresses the growth and development of herbivores in both Arabidopsis and tea plants.
While some herbivore-related protease inhibitors are induced by herbivore attack or herbivory-related signaling [16][17][18], our results show that the transcriptional levels CsSERPIN1 were minimally affected by biotic and abiotic stresses.Herbivore attack, pathogen infection, cold, heat, as well as defense-related signaling molecule treatments, did not alter the expression of CsSER-PIN1.The expression pattern of CsCERPIN1 was confirmed by the transcriptomic analysis after herbivory using RNA sequencing.Furthermore, we found that this unaffected pattern was independent of tea genotypes, suggesting that CsSERPIN1 might be a constitutive gene in tea plants.However, we cannot rule out the possibility that other untested factors may inf luence the expression of CsSERPIN1.
The effects of SERPINs on insect digestive protease activity have been extensively studied [8,9,16,27].Similarly, our in vitro study shows that CsSERPIN1 can inhibit trypsin and chymotrypsin proteases.Feeding assays with artificial diet and heteroexpression plants further confirmed the inhibitory ability of CsSERPIN1 in the gut of tea geometrids and fall armyworms.Given that serine proteases have been found in several insect orders including Lepidoptera, Diptera, Orthoptera, Hymenoptera, and Coleoptera [2,28], it is quite possible that CsSERPIN1 can suppress the growth of various herbivores.However, more research is necessary to confirm this possibility.Besides serine proteases, some SERPINs could inhibit the cysteine proteases.For instance, AtSerpin1 can significantly suppress the activities of cathepsin B and L protease in aphids.Feeding assays with artificial diet indicate that AtSer-pin1 is toxic to aphid nymphs with 50% mortality [7].Thus, it would be interesting to test whether CsSERPIN1 has the similar activity and effects on herbivores that mainly rely on cysteine proteases for digestion in future studies.
In plants, serine proteolysis is a key cellular process that is carried out by a diverse group of serine proteases [29].SERPINs are essential for the regulation of endogenous serine proteases, enabling precise control of proteolytic processes that ultimately ref lect on proper plant growth and development [30].Intriguingly, we did not observe any differences in plant growth mediated by CsSERPIN1.CsSERPIN1-heteroexpressed Arabidopsis plants showed similar root length and biomass to WT plants.The lack of clear phenotypic differences suggests that CsSERPIN1 either has only marginal effects on plant growth or that its inf luence on plant growth is not ref lected in the assessed parameters.Similarly, silencing SERPIN SPI2a, SPI2b, or SPI2c also does not inf luence the growth and development of Arabidopsis [16].Nevertheless, more detailed experiments should be conducted to assess whether CsSERPIN1 impacts plant growth and development.
In our study, CsSERPIN1 reduced the performance of two lepidopteran species.Feeding assays with artificial diet revealed that CsSERPIN1 interfered with the growth and pupation of tea geometrid.Transient overexpression of CsSERPIN1 in tea plants further confirmed its suppressive effects on larval growth.Notably, we were unable to include the key experiment to prove the suppressive effects of CsSERPIN1, where loss of function of CsSERPIN1 would result in increased larval weight gain, as the stable transformation technology is currently unavailable in tea plants.Instead, we attempted to transiently knock down CsSERPIN1 with antisense oligodeoxynucleotides, but were unsuccessful, probably because the methods can vary depending on the properties of target genes.Nevertheless, further investigation of knock out/down of CsSERPIN1 and its impact on herbivores will be necessary in future studies.Additionally, heterologous expression of CsSERPIN1 in Arabidopsis demonstrated its notable reduction ability in the weight gain of fall armyworms.These results are consistent with previous reports that have highlighted the importance of SERPINs in reducing the performance of lepidopteran herbivores, either when engineered into transgenic plants or introduced into artificial diets [7,8,16].While CsSERPIN1 has been observed to significantly suppress herbivores growth, it is important to note that plants have a variety of defensive compounds and multiple mechanisms to defend against herbivores.CsSERPIN1 only plays a role in one aspect of this complex defense system, rather than being the sole determinant of it.Because the resistant tea accessions could have multiple pathways to resist herbivores, including independent mechanisms involving CsSERPIN1, a detailed dissection of the defense mechanisms in the resistant tea accessions may offer valuable insights into the specific role of CsSERPIN1 within the intricate defense system.This understanding could provide a valuable source of durable resistance for developing new tea cultivars.
Emerging studies show that lepidopteran herbivores can adjust their digestive proteolytic processes to counteract the inhibitory activities of plant protease inhibitors [31][32][33].To overcome this herbivore adaptation, researchers suggest that it may be essential to integrate more than one inhibitor [11,12,34].Therefore, in an agricultural context, while CsSERPIN1 could reduce herbivore attack, it will be wise to combine it with other protease inhibitors or protease inhibitor families, such as inducible protease inhibitors, to achieve irreversible effects on target herbivores.
Taken together, although many plant protease inhibitors have been identified and proven to play an important role in plant defense, to the best of our knowledge, CsSERPIN1 is the first described constitutive SERPIN involved in herbivore resistance in tea plants.As CsSERPIN1 is not inf luenced by environmental factors and does not appear to negatively affect plant growth, it has significant potential for use in herbivore control.

Plants and insects
To investigate the involvement of CsSERPIN1 in herbivore resistance, we used tea (Camellia sinensis) and Arabidopsis (A. thaliana) plants in this study.All 33 tea accessions were 3-year-old plants, grown at Shengzhou Experimental Station in Zhejiang Province, China (120 • 48 46 E, 29 • 44 45 N).These tea accessions, except for the 'Longjing 43' (LJ) accession, are our in-house breeding materials, which were carefully selected from various local tea populations across China.Some of them are still undergoing the breeding process and do not have assigned accession numbers yet.The detailed information on the three most resistant and susceptible tea accessions studied is listed in Table S3 (see online supplementary material).The 3-year-old tea cultivar 'LJ' was used for CsSER-PIN1 expression analysis and transient overexpression assays.Healthy tea plants of similar size were selected for experiments.
We generated CsSERPIN1-heteroexpressed Arabidopsis plants (L2 and L5) using an established protocol described in the section 'Heterologous expression of CsSERPIN1 in Arabidopsis'.Arabidopsis seeds were germinated on MS medium and grown for 10 days in MS medium before being transferred individually into round pots with commercial potting soil in a climate chamber (23 • C, 50% relative humidity, 12 h/12 h photoperiod).Thirty-day-old wildtype plants and heteroexpressed lines were used for phenotype assessment and herbivore resistance assays.

Herbivore resistance evaluation of tea accessions
To evaluate the herbivore resistance of different tea accessions, we used larval growth and leaf consumption area as parameters.The second and third fully expanded leaves with short stems of each accession were freshly detached and placed into a hydrated f loral foam for moisturizing.Leaves, together with the f loral foam, were then put into a 10 cm square petri dish.Groups of seven three-day-old tea geometrid caterpillars with similar length were selected and pre-starved for 6 h before being introduced into the petri dish.The leaves were replaced once a day to ensure that caterpillars have sufficient fresh leaves.Larval weight was measured seven days after the start of assay (n = 51-56).To quantify the consumed leaf aera, the remaining leaf after larvae feeding were scanned, and the lost area was assessed (n = 8).

CsSERPIN1 cloning and characterization
The cDNA sequence of CsSERPIN1 was amplified using the primers CsSERPIN1-F and CsSERPIN1-R, which were designed based on the CsSERPIN1 sequence (NCBI accession no.LOC114291199) and are listed in Table S4 (see online supplementary material).The amplified product was cloned into a pEASY-blunt zero vector (TransGen Biotech, Beijing, China) and subsequently sequenced.Structural domain analysis was performed with SMART (Simple Modular Architecture Research Tool, http://smart.embl-heidelberg.de)and PFAM (http://pfam.xfam.org/)databases.

Phylogenetic analysis
Phylogenetic analysis was performed using the MEGA 11.0 program [35].The protein sequences were aligned using the ClustalW method in MEGA 11.0 with the following parameters: pairwise alignment gap opening penalty of 10, gap extension penalty of 0.1, multiple alignment gap opening penalty of 10, gap extension penalty of 0.2, the use negative matrix was turned off, and the delay divergence cutoff (percentage) was set at 30.The resulting alignment was then used to construct an unrooted tree using the neighbor-joining method with bootstrap analysis (n = 1000, amino acid, Poisson model, rate and patterns: uniform rates, data subset to use: pairwise deletion, traditional tree without modification for graphics).Statistical tests were performed, and the bootstrap values for the branches are shown in the generated tree.

Gene expression analysis
We used QRT-PCR to examine the transcriptional levels of genes.In tea plants, the relative gene expression levels were determined using a 2 − Ct method.The house-keeping gene CsGAPDH of tea plants was employed as an internal standard to normalize cDNA concentrations.The protocols for this method has been described in detail in our previous study [19].As CsSERPIN1 gene is absent in wild-type Arabidopsis, we employed a standard curve method rather than 2 − Ct method to quantify the expression levels of CsSERPIN1 in both wild-type and CsSERPIN1-heteroexpressed Arabidopsis plants.This method does not depend on the CsSERPIN1 expression of wild-type plants as calibrator samples.Therefore, the relative expression of CsSERPIN1 in the heteroexpressed lines can be accurately calculated even in the absence of CsSERPIN1 in wild-type Arabidopsis.For a comprehensive understanding of the standard curve method, please refer to Wong and Medrano [36].Specifically, we used serial dilutions of a specific cDNA standard to generate a standard curve, and then applied the cycle threshold (Ct) values of the samples to the curve to calculate the relative expression levels of CsSERPIN1.The CsSER-PIN1 relative expression level in Arabidopsis is the transcript level of CsSERPIN1 relative to the Arabidopsis house-keeping gene glyceraldehyde-3-phosphate dehydrogenase (AtGAPDH).QRT-PCR primers for the house-keeping and tested genes, including CsSER-PIN1, CsSERPIN2 (LOC114256809), CsSERPIN3 (LOC114271199), and CsSERPIN4 (LOC114271172) are all listed in Table S4 (see online supplementary material).

Plant treatments
For herbivory treatment, two third-instar tea geometrid larvae were allowed to feed on the second and third fully expanded leaves, which were consistent with the tissue used for the herbivore resistance evaluation of tea plants.The leaf and the larvae were covered with a small mesh pocket (9 cm × 11 cm) to prevent larvae from escaping.Leaves covered with empty bags were used as control.The whole leaves were harvested at specified time intervals after herbivore attack (n = 5-6).
For the treatment of defense signaling molecules, JA, ABA, SA, or 1-aminocyclopropane-1-carboxylic acid (ACC, for ethylene treatment) (Solarbio, Beijing, China) was first dissolved in ethanol of a small volume, and further diluted in the sterilized Milli-Q water to get working solutions with a concentration of 200 μM for JA, ABA, 1 mM for SA, or 100 μM for ACC.The prepared working solutions were then thoroughly sprayed onto the leaf surface.The control plants were treated similarly with Milli-Q water and ethanol indicated above.The second leaves were harvested at specified time intervals after treatments (n = 5).
For pathogen infection, the spore suspension of Colletotrichum camelliae was inoculated onto tea leaves, which were mechanically pieced as described preciously [37].Sterile water was inoculated as a control.The infected leaves were harvested at specified time intervals after treatments (n = 5).
For cold and heat treatment, tea plants were treated at 4 • C (cold stress) or 38 • C (heat stress).The control plants were treated at 22 • C. The apical second and third young leaves were harvested at specified time intervals after treatments (n = 5).

CsSERPIN1 expression and purification in E. coli
The ORF of CsSERPIN1 was inserted into a pMAL-c2X vector, in frame with a maltose binding protein (MBP) tag, and then transformed into E. coli strain BL21(DE3).After incubation in ampicillincontaining LB medium for 20 h at 37 • C, the culture cells were diluted 1:50 and grown until OD600 was 0.7.IPTG was added to a final concentration of 1 mM, and the cells were further incubated at 16 • C and 140 rpm overnight for the protein expression.The fused proteins were purified by MPB-binding resin (New England BioLabs, Ipswich, Massachusetts, USA) following the manufacturer instructions.Concentrations of purified protein were quantified using the Bradford method (Bradford Protein Assay Kit, Solarbio, Beijing, China).The purified proteins were further confirmed by immunoblot blot using an anti-MBP monoclonal antibody (Genscript, Nanjing, China) and a horseradish peroxidaseconjugated secondary antibody (Solarbio, Beijing, China).

Larval gut extraction
To extract the larval gut, fifth-instar larvae were cold-anesthetized and dissected.The midguts and contents from ten larvae were pooled together, weighed in tubes, and subsequently homogenized in a volume of 0.15 M sodium chloride equivalent to the gut weight.Homogenates were centrifugated, and supernatants were collected.The total protein content in supernatants was determined using a Bradford assay (Bradford Protein Assay Kit, Solarbio, China).

CsSERPIN1 inhibitory activity
The ability of CsSERPIN1 to inhibit the activities of chymotrypsin and trypsin was determined using previously described methods [27,38,39].Brief ly, the trypsin activity was analysed using α-N-benzoyl-DL -arginine-p-nitroanilide hydrochloride (BApNA) as a substrate with a concentration of 1.5 mM in 0.8% (v/v) DMSO.The enzyme (4.7 × 10 −5 M) and purified CsSERPIN1 protein (3-18 μg) were thermo-equilibrated at 37 • C for 10 min before addition of substrate to start the reaction.The buffer used was 50 mM Tris-HCl, pH 8.0.After 37 • C incubating for 20 min, 30% acetic acid (v/v) was added to stop the reaction.The protease activities were measured by the absorbance at 410 nm.For gut protease activities, the larval gut protein extracts (10 μg) and the substrate were incubated in buffer for 20 min.Appropriate blanks were uses in all the assays.Chymotrypsin protease was quantified similarly as described above, except N-succinyl-alanine-alanineproline-phenylalanine-p-nitroanilide (SAAPFpNA) was used as a substrate.

Transient overexpression of CsSERPIN1 in tea plants
The ORF of CsSERPIN1 was cloned into the pBWA(V)HS-GFP vector and the resulting pBWA(V)HS-CsSERPIN1-GFP plasmid was used to transform A. tumefaciens.Third leaves of 3-year-old tea plants were infiltrated with the transformed Agrobacterium using the A. tumefaciens-mediated infiltration method [40].The empty vector pBWA(V)HS-GFP and infiltration buffer were infiltrated as controls.The entire leaves were harvested for determination of CsSERPIN1 overexpression using QRT-PCR (n = 5).

Heterologous expression of CsSERPIN1 in Arabidopsis
CsSERPIN1 was cloned into a pBWA(V)HS vector carrying a FLAG epitope tag to yield an heteroexpression construct, pBWA(V)HS-CsSERPIN1-f lag.The recombinant vector was then transferred into A. thaliana Columbia-0 via Agrobacterium-mediated transformation, as described previously [41].Two T 2 homozygous lines (L2 and L5) were selected for subsequent experiments.The expression levels of CsSERPIN1 were confirmed by QRT-PCR and immunoblot assays.An anti-FLAG antibody (Sigma-Aldrich, Darmstadt, Germany) and a horseradish peroxidase-conjugated secondary antibody (Solarbio, Beijing, China) were used for the immunoblot assays.

Effect of purified CsSERPIN1 on tea geometrid
Freshly prepared artificial diet was cut into small pieces (10 × 10 × 10 mm) and separated into petri dishes [42].Twenty microliters of 800 μg ml −1 purified CsSERPIN1-MBP proteins were added to each diet, resulting in a concentration of 16 μg ml −1 in the diet.The amount of CsSERPIN1-MBP protein added was based on its inhibitory effects, which was measured in in vitro assays.As shown in Fig. 3b, 12 to 18 μg of CsSERPIN1-MBP protein in 1 ml reaction buffer (12-18 μg ml −1 ) resulted in strong inhibition of trypsin and chymotrypsin activities, with more than 70% and 20% inhibition, respectively.Therefore, 16 μg of CsSERPIN1-MBP protein was introduced into 1 cm 3 artificial diet (16 μg ml −1 ) to test its inhibitory effects on larva performance.Groups of ten three-day-old tea geometrid caterpillars of similar size were prestarved for 6 h and introduced into the petri dishes.The diets with proteins were replaced every half-day to ensure constant protein activity and sufficient fresh food source for the caterpillars.Diets with MBP or distilled water were used as controls.Larval weight was quantified 5 d and 8 d after the start of the assay, and the pupal length and weight were measured after larval pupation (n = 24-30).

Effect of overexpressed CsSERPIN1 on tea geometrid
Two three-day old tea geometrid caterpillars of similar size were introduced to feed on the control, empty vector, or CsSERPIN1trasnformed tea plants.The larvae were confined to the leaves of each plant with a small mesh pocket (9 cm × 11 cm) to prevent escape.The leaves were replaced once half of them were consumed.Larval weight was quantified 5 d after the start of the assay (n = 26).

Effect of heteroexpressed CsSERPIN1 on fall armyworm
A second-instar fall armyworm larva was pre-weighted and then allowed to feed on the wild type and two CsSERPIN1heteroexpressed lines.Each plant and larva were individually confined in a transparent plastic cup with a lid (10 cm in height and 8 cm diameter) to prevent larvae from escaping.The plants were replaced once half of leaves were consumed.Larval weight was quantified 5 and 7 d after the start of the assay (n = 25).

Figure 1 .
Figure 1.Expression of CsSERPIN1 correlates with herbivore growth in tea plants.(a) Larval weight gain of tea geometrid feeding on different tea accessions for 7 days (± SD, n = 51-56).(b-e) CsSERPIN1, CsSERPIN2, CsSERPIN3, and CsSERPIN4 expression levels in different tea accessions (± SD, n = 6).For the detailed statistical information, refer to TablesS1 and S2(see online supplementary material).(f-i) Correlations between the expression of CsSERPIN1, CsSERPIN2, CsSERPIN3, and CsSERPIN4 genes, and larval weight gain.The Pearson's product-moment r, R 2 , and P values of correlations are indicated.

Figure 2 .
Figure 2. Expression pattern of CsSERPIN1 under different treatments.(a) Heatmaps of CsSERPIN1 expression under herbivore attack, jasmonic acid (JA), salicylic acid (SA), ethylene, abscisic acid (ABA), pathogen infection, cold and heat stress treatments.The color intensity indicates the fold changes of the CsSERPIN1 expression in treated plants relative to control plants (n = 3-6).(b-i) The detailed expression patterns of CsSERPIN1 (+ SE, n = 3-6) under various treatments, including herbivore attack (b), treatment with jasmonic acid (JA, c), salicylic acid (SA, d), ethylene (e), abscisic acid (ABA, f), pathogen (g), cold (4 • C, h), or heat (38 • C, i) stress, at different time points.Inserts are included to show the expression levels of marker genes for each treatment as positive controls.Asterisks indicate significant differences between treatments at the same time points (two-way ANOVA followed by pairwise comparisons through FDR-corrected LSMeans; * P < 0.05; * * P < 0.01).(j) Heatmaps of CsSERPINs expression under herbivore attack for 3 h and 24 h.The color intensity indicates the fold changes of CsSERPINs expression in treated plants relative to control plants using RNA sequencing data (n = 5).For detailed information, refer to Data S1 (see online supplementary material).

Figure 3 .
Figure 3. CsSERPIN1 suppresses larval performance and protease activities in vitro and in vivo.(a) Immunoblot analysis of purified CsSERPIN1 protein.The anti-maltose-binding protein (MBP) monoclonal antibody was used to detect the CsSERPIN1-MBP fusion protein.(b) The inhibitory activity of CsSERPIN1 against trypsin and chymotrypsin enzymes in vitro (+ SE, n = 4).(c, d) Larval weight gain (c), pupal weight and length (d) of tea geometrid fed on artificial diet containing purified CsSERPIN1 or MBP, or control (Con) diet (± SD, n = 24-30).(e) Gut trypsin and chymotrypsin activities of tea geometrid fed on artificial diet containing purified CsSERPIN1 or MBP, or control diet (+ SE, n = 4).Statistical significance for each comparison is indicated by different letters (one-way ANOVA, P < 0.05).

Figure 5 .
Figure 5. Heterologous expression of CsSERPIN1 inhibits larval growth and gut protease activities in Arabidopsis.(a) Expression levels (top panel, + SE, n = 5) and protein accumulation (bottom panel) of CsSERPIN1 in wild-type (WT) Arabidopsis plants and CsSERPIN1-heteroexpressed lines (L2 and L5).The immunoblotting was performed using a f lag antibody to detect CsSERPIN1, or an actin antibody to detect actin as a loading control.(b) Root length and shoot biomass of WT Arabidopsis and CsSERPIN1-heteroexpressed lines (left two panels, + SE, n = 12).Growth phenotypes of 12-and 30-day-old WT and CsSERPIN1-heteroexpressed plants are shown (right two panels).(c) Larval weight gain of fall armyworms fed on WT and CsSERPIN1-heteroexpressed plants (± SD, n = 25).(d) Gut trypsin and chymotrypsin activities of fall armyworms fed on WT and CsSERPIN1-heteroexpressed plants (+ SE, n = 4).Statistical significance for each comparison is indicated by different letters (one-way ANOVA, P < 0.05).

Figure 6 .
Figure 6.CsSERPIN1 contributes to herbivore resistance variation of tea plants.(a) Leaf area consumed by tea geometrid larvae feeding on three resistant (ZX10, ZX19, and ZX48) and three susceptible (ZX13, ZX16, and ZX18) tea accessions (+ SE, n = 8).(b) Gut trypsin and chymotrypsin activities in tea geometrid larvae feeding on the three resistant and three susceptible tea accessions, respectively (+ SE, n = 4).Statistical significance for each comparison is indicated by different letters or asterisks (one-way ANOVA, P < 0.05).