Stress inducible proteinase inhibitor diversity in Capsicum annuum

Background Wound-inducible Pin-II Proteinase inhibitors (PIs) are one of the important plant serine PIs which have been studied extensively for their structural and functional diversity and relevance in plant defense against insect pests. To explore the functional specialization of an array of Capsicum annuum (L.) proteinase inhibitor (CanPIs) genes, we studied their expression, processing and tissue-specific distribution under steady-state and induced conditions. Inductions were performed by subjecting C. annuum leaves to various treatments, namely aphid infestation or mechanical wounding followed by treatment with either oral secretion (OS) of Helicoverpa armigera or water. Results The elicitation treatments regulated the accumulation of CanPIs corresponding to 4-, 3-, and 2-inhibitory repeat domains (IRDs). Fourty seven different CanPI genes composed of 28 unique IRDs were identified in total along with those reported earlier. The CanPI gene pool either from uninduced or induced leaves was dominated by 3-IRD PIs and trypsin inhibitory domains. Also a major contribution by 4-IRD CanPI genes possessing trypsin and chymotrypsin inhibitor domains was specifically revealed in wounded leaves treated with OS. Wounding displayed the highest number of unique CanPIs while wounding with OS treatment resulted in the high accumulation of specifically CanPI-4, -7 and −10. Characterization of the PI protein activity through two dimensional gel electrophoresis revealed tissue and induction specific patterns. Consistent with transcript abundance, wound plus OS or water treated C. annuum leaves exhibited significantly higher PI activity and isoform diversity contributed by 3- and 4-IRD CanPIs. CanPI accumulation and activity was weakly elicited by aphid infestation yet resulted in the higher expression of CanPI-26, -41 and −43. Conclusions Plants can differentially perceive various kinds of insect attacks and respond appropriately through activating plant defenses including regulation of PIs at transcriptional and post-translational levels. Based on the differentially elicited CanPI accumulation patterns, it is intriguing to speculate that generating sequence diversity in the form of multi-IRD PIs is a part of elaborative plant defense strategy to obtain a diverse pool of functional units to confine insect attack.


Background
Plants have evolved elaborate defense strategies composed of constitutive and inducible responses in order to cope with herbivore challenges. The induced defenses commence only when herbivore-derived signals are perceived by the plants. A wide array of studies has reported the induction of direct and indirect plant defenses in response to herbivory and other biotic stresses [1][2][3][4]. Insect damage, mechanical wounding and/or elicitors in insect oral secretions (OS), such as fatty acid amino acid conjugates, volicitin, inceptins, caeliferins, and glucose oxidase, stimulate the local and systemic release of signaling intermediates like systemin and/or jasmonic acid; these then amplify the defense cascade throughout the plant [5][6][7]. Though the major consequence of herbivory is wounding, plants' responses to insect feeding are more complex due to the elicitors present in insect OS [8]. Defense responses entail the regulated activation of plant defense genes and the suppression of growth-related genes [8,9]. As a result, defensive metabolites and/or proteins accumulate in plants within the local tissues damaged by herbivores as well as systemically in undamaged tissues.
The accumulation of trypsin and chymotrypsin-like proteinase inhibitors (PIs) throughout the aerial tissues of tomato and potato plants was demonstrated to be a direct consequence of insect-mediated damage or mechanical wounding [10]. Thus, serine PIs represent one of the best examples of locally and systemically induced responses in Solanaceous plants [11][12][13][14][15][16]. The constitutive expression of PIs, which has been reported to occur in storage organs and the reproductive tissues of plants, may fulfill anti-insecticidal as well as other endogenous functions in planta [4,13,[16][17][18].
Most Solanaceae members contain the multi-gene family encoding Pin-II type PIs [4,16,19], which possesses considerable sequence diversity resulting from variations in tandem sequence repeats, domain duplications and circularly permuted domain organizations [20]. A distinct feature of these PIs is the presence of tandem repeats of a 50-amino-acid polypeptide called inhibitory repeat domain (IRD), which can vary from 1 to 8 with interconnecting linker peptides. Each IRD contains 8 conserved cysteines (Cys) along with a reactive site for targeting a serine protease. Gene duplication events have resulted in the evolution of the multi-domain Pin-II family of PIs with structurally and functionally divergent IRDs [21]. Horn et al. [22] isolated a set of IRDs resulting from the differential proteolysis at the linker peptide separating the subunits of a 7-domain precursor from methyl-jasmonate-elicited N. attenuata leaves. The sequence variability in the multi-gene family of Pin-II PI proteins, their regulated expression and their posttranslational processing are together responsible for generating a PI cocktail effective in defense and/or endogenous function [4,16,23].
Several different PI proteins and genes with 1-to 4-IRDs have been identified and characterized from C. annuum (CanPIs) tissues [16,[24][25][26][27][28][29]. There was substantial variability in the induced expression of CanPIs upon aphid infestation, virus infection, chewing by insects and mechanical wounding. The abundance of transcripts did not always result in higher CanPI proteins, though they were well correlated in lepidopteran-infested C. annuum leaves. Furthermore, these studies indicated that many CanPIs are expressed simultaneously, but the significance of such PI expression diversity in C. annuum remains unclear.
In order to examine the potential functional specificities of the various isoforms of CanPI in C. annuum, we addressed the following questions: (i) Does elicitation increase PI isoform diversity? (ii) How specialized is the induction response to a particular treatment? Following experimental inductions of C. annuum leaves, we investigated the diversity in CanPI transcript and protein profiles. Sequencing revealed 24 novel CanPI transcripts, increasing the total known to 47. Selective analysis of PI activity in proteomes using 1D and 2D electrophoresis followed by mass spectrometry revealed local and systemic responses in PI activity.

Differential regulation of CanPIs upon induction
The amplification of cDNA derived from uninduced and induced leaves with CanPI gene-specific oligonucleotides yielded transcripts of 789, 614, 445 and 267 bp, representing 4-, 3-, 2-and 1-IRD CanPIs, respectively. Semi-quantitative analysis revealed differential CanPI expression in uninduced and induced leaves ( Figure 1A). In comparison to CanPIs with 2-IRDs, those with 4-and 3-IRDs showed higher abundance in wounded leaves treated with water or OS ( Figure 1A). In aphid-infested leaves, while the overall expression of CanPIs was low, 3-IRD transcripts were prominent. The amplified transcripts were cloned and 60 representative clones from each treatment (25 in case of uninduced) were sequenced to confirm the identities of the CanPIs. This analysis detected novel 4-, 3-and 2-IRD subtypes based on variations in amino acid composition. Based on this analysis, in addition to the 23 CanPIs previously reported [16,26,27], 24 novel CanPIs were identified including seven 2-IRD, fourteen 3-IRD, and three 4-IRD CanPIs (Table 1). 1-IRD CanPIs identified in uninduced leaves were not detected under any of the induction treatments. Details regarding the treatmentspecific representation of CanPIs and their IRD composition are summarized in Table 1. The frequency of occurrence of individual CanPIs out of the 60 clones sequenced per treatment was analyzed and is referred to as abundance in leaf under a particular treatment (Table 1).
Variation in the abundance of CanPIs was apparent ( Figure 1B), with 3-IRD CanPIs being highest (from 40 to 60%) either in uninduced or induced leaves. The abundance of 4-IRD CanPIs was increased in leaves subjected to wounding and treated with OS (38%), compared to aphid-infested (12%) leaves or in wounded leaves treated with water (17%). The proportion of 2-IRD CanPIs ranged from the lowest (10%) in wounded leaves treated with OS to the highest (20%) in aphid infested and wounded leaves treated with water. The differential expression of the various subtypes of CanPIs (with respect to their IRD composition; see Table 1) resulted in an induction-specific CanPI profile. CanPI-4, -7, -10, -24,-41 and −43 showed the highest representation in leaves across all induction treatments (Table 1). In wounded leaves treated with OS, CanPI-4 and −7 showed the highest frequency, whereas CanPI-41 and −43 were represented most highly in aphid-infested leaves. Wounded leaves treated with water showed the highest number of expressed CanPIs (25) as well as a wide representation of several unique CanPIs (10 in number) in low frequencies (Table 1, Figure 1C). Responses to wounding with OS appear more specialized as suggested by the expression of a few CanPIs (16) in high frequencies (Table 1, Figure 1C). Both the treatments share more CanPIs (3-and 4-IRD type) possibly due to the standardization of the amount of wounding between these two treatments. Aphid-infested leaves also accumulated transcripts of 7 unique CanPIs representing a diverse array of PIs. Only transcripts specific for CanPI-3, -5, -7, -8 and −10 could be analyzed from these tissues. CanPI-7 (4-IRD) showed constitutive expression in uninduced as well as induced leaves, while CanPI-3 and −5 (3-IRD PIs) showed low accumulation levels in wounded leaves treated with OS ( Figure 1D). CanPI-8 and −10 (4-IRD PIs) were differentially expressed in wounded leaves treated with water or OS, and completely absent in aphid-infested leaf tissues ( Figure 1D).
The Protein Prowler predictor [30] revealed ER signal peptides in all CanPIs. The ER signal peptide (SP) sequences of 25 aa at the N-terminal of CanPIs, which showed ten variants, were named SP-1 to SP-10 (Additional file 2: Figure S1).
To investigate the interrelationships and groupings within CanPIs and with out-groups, phylogenetic analysis was carried out. CanPIs formed a distinct cluster from full-length Pin-II PI (4-IRD) from Nicotiana benthamiana used as an out-group. The dendrogram revealed clustering based on the identical component IRDs in CanPIs (Figure 2). Distinct clusters of 4-and 3-IRD PIs were evident with some intermixing for e.g. CanPI-31, -33, -1 and −4. CanPIs showing more aa sequence similarity associated close to each other for e.g.  Table S1.
The deduced amino acid sequences of CanPIs (47 in total) consisted of 28 unique IRDs; of these, 7 had chymotrypsin inhibitory (CI) reactive sites and 21 had trypsin inhibitory (TI) reactive sites. The multiple sequence alignment of IRDs revealed variation mostly within reactive site loops or towards the C-terminal ends (Additional file 3: Figure S2). Five TI IRDs possess 'Lys' while 16 had ' Arg' at the P1 position of the reactive site. CI IRDs had 'Leu' at P1, except IRD 22 which had 'Pro'. IRD 48 (TI) had active site variation with ' Asn' being replaced with ' Asp' ( Figure 3A). Another crucial difference was variation in the number and positioning of cysteine residues per IRD. Four such cysteine variants, namely IRD 9, 11 and 24 were TIs, and IRD 33 was a CI ( Figure 3A).
The relative abundance of IRDs that resulted from various types of inductions was analyzed ( Figure 3B). Representation of IRDs 1, 4, 5, 10, 12, 14, 17, 18, 25 and 37 was high in induced tissues and exhibited treatment specificity. IRDs 1 and 17 were highly represented under all treatments while IRDs 12, 18 and 37, which are TIs, showed high frequency in aphid-infested leaves and wounded leaves treated with water ( Figure 3B). The abundance of IRDs 4 and 5, which are CIs, and IRDs 10, 14 and 25, which are TIs, was distinct in wounded leaves treated with OS ( Figure 3B). Most of the remaining diversity of IRDs is contributed by wounded leaves treated with water or OS ( Figure 3C).
Wounding and insect damage results in quantitative and qualitative changes in CanPI proteins C. annuum leaves were found to have increased PI activity upon induction ( Figure 4A). Significantly higher level   of PI activity was evident in the systemic leaves of wounding treated with water induction and both, the local and systemic leaves of wounding treated with OS induction compared to similar leaves from unwounded control plants ( Figure 4A). The PI activity in aphidinfested leaves was higher than in uninduced leaves; however, it was 1.5-and 3.5-fold less than in wounded leaves treated with water or with OS, respectively. Induced levels of PI activity ranged from 2-fold in leaves wounded and treated with water to 4-fold in leaves wounded and treated with OS. A differential pattern of PI isoform induction was observed in C. annuum leaves in response to various treatments ( Figure 4B). Three prominent CanPI activity bands were detected in leaves that were wounded and treated with OS or with water, while only two PI isoforms could be detected in case of aphid infested and uninduced leaves. This difference indicated induced qualitative diversity in the CanPIs that resulted from these two treatments. Extracts from leaves wounded and treated with water or OS attained an early saturation of H. armigera gut protease (HGP) inhibition (70%) unlike aphid-infested and uninduced tissues, consistent with the quantitative differences amongst the PIs in their activity ( Figure 4C). The HGP of fourth-instar larvae displayed at least seven protease isoforms (HGP-1 to −7), of which only HGP-6 and −7 were able to retain marginal activity in the presence of CanPIs from either uninduced or induced tissues ( Figure 4D).
The 2-D activity profiles of induced leaf samples showed the induction of several novel PI isoforms in the range of isoelectric point (pI) from 4 to 7 and also shift in the pI and/or molecular mass of few isoforms ( Figure 5). Five prominent TI isoforms (TI-1 to TI-5) corresponding to 1-, 2-, 3-and 4-IRD were identified in uninduced leaves, in the pI range of 5 to 7 ( Figure 5). TI-6 (1-IRD) isoform which has major basic shift in pI compared to TI-1, was present in all the three types of induced tissues. TI-3, -4 and −5 were consistently detected in aphid-infested leaves but were absent from wounded leaves treated with water or OS. TI-8 to −13 were present only in wounded leaves treated with water, while TI-14, -15, and −16, corresponding to 4-IRD Can-PIs, were present in wounded leaves treated with OS only. The induced PI activity in wounded leaves treated with water showed a distribution of several TI isoforms with low intensity, whereas the up-regulation of only a few TI isoforms was evident in wounded leaves treated with OS. TI-2 was absent in all induced tissues, consistent with the specificity of some PI isoforms in uninduced leaves. Among the TI isoforms corresponding to 1-, 2-, 3-or 4-IRD CanPIs, more variations were observed for 3-and 4-IRD CanPIs.
Partially purified PIs contained small peptides of about 5.5 to 6.3 kDa equivalent to a single IRD as analyzed on MALDI-TOF-MS ( Figure 6). High molecular mass proteins exhibited very low intensity in the mass spectra, perhaps due to ion suppression effects, and therefore are not considered. Uninduced leaf extracts displayed a single major peak of 5583 Da, whereas aphid-infested leaves had major peaks at 5583 Da and 5616 Da and few low intensity peaks. A peak of 5583 Da was most prominent in wounded leaves treated with water, in addition to several peaks of 5616, 5760 and 5832 Da with high intensity. Extract from leaves wounded and treated with OS displayed 6138 and 5961 Da as the major peaks, with low intensity peaks at 5616, 5832, 6036 and 6301 Da and an absence of the 5583 Da peak. This variation in the molecular masses of peptides is likely a result of the proteolytic processing of the precursor PI proteins to generate multiple functional PI species (equivalent to single IRD from multi-IRD CanPIs) [22,31,32].
The partially purified PI protein from uninduced and induced leaves, displaying varying mass spectral profiles on MALDI-TOF-MS, were separated on Tricine gel, and the proteins were individually excised and processed for in-gel digestion followed by the identification of peptides by MALDI-TOF-MS/MS (Additional file 1: Table S2). The 6 kDa protein in all the extracts showed matches to Dendrogram of CanPIs based on deduced amino acid sequences of full-length genes of 1-to 4-IRDs isolated from C. annuum. Pin-II type PI from N. benthamiana (NbPI, NCBI: ABA42892) is used as an out-group. Details of the CanPIs (1 to 47), composition, number of IRDs, tissue-wise representation and color coding are given in Table 1. Dendrogram was constructed using Lasergene software.
Pin-II proteinase inhibitors from C. annuum. However, due to high homology amongst IRD sequences, the database generated identification to several CanPIs/IRDs (Additional file 1: Table S2).

Tissue specificity of PI accumulation
Flowers, followed by stems and early fruit showed a significantly higher level of PI activity compared to tissue from leaves, roots and the different developmental  B   TI CI CI CI TI TI TI TI TI TI TI TI CI TI TI  stages of fruit (Additional file 4: Figure S3). Flower tissue showed the highest PI activity, while tissue from turning fruit exhibited the least activity with a 7-fold difference. The in-gel PI visualization after resolution on 2-DE for various C. annuum plant parts displayed the qualitative variations in the PI activity across these tissues (Figure 7). The clusters corresponding to 1-, 2-, 3-and 4-IRD CanPIs were prominent in stems, early fruit and flowers. In accordance to the very low TIUs in roots, we could detect only one TI isoform which corresponded to 2-IRD CanPI. The 1-IRD cluster contained several merged TI isoforms and indicated multiple, charge variants due to amino acid sequence differences. Stem tissue in particular showed less diversity in the 3-IRD cluster while exhibiting the highest number of charge variants in the 1-IRD cluster. Stem, early fruit stage and flower tissues revealed the presence of the high molecular mass isoforms of PIs (Figure 7, marked by arrows) that are predicted to be larger than those of 4-IRD CanPIs.

Discussion
C. annuum produces an array of PI genes exhibiting regulated expression under induced conditions. CanPIs are strongly elicited by wounding and upon Lepidopteran insect attack owing to the high cellular damage and plant's perception of herbivore-specific elicitors as compared to aphid infestation. Based on observed CanPI accumulation upon elicitation, it is intriguing to speculate that generating sequence diversity in the form of multi-IRD PIs is part of a plant's defense strategy to obtain a pool of diverse functional units to confine insect attack. Among Solanaceae, different species of Nicotiana display PI genes containing 2 and 4 to 8 IRDs [19]. In addition, 2-IRD PIs from tomato (Solanum lycopersicum) and S. nigrum have been well characterized [4,33]. Simultaneously expressed PIs with varying IRD composition have been reported in N. alata stigma (4-and 6-IRDs; [34]), in N. glutinosa infected with TMV (6-and 8-IRD; [35]) and in N. attenuata in response to herbivory (7-IRD; [22]). In addition to the previously reported 23 PI genes from C. annuum pericarp, developing fruit and stem [16,26,27], 24 new CanPIs were isolated and characterized from the induced leaves in the present study. Among these 47 CanPI genes, 9 contained 4-IRDs, 20 contained 3-IRDs, 15 contained 2-IRDs and 3 contained a single IRD thus contributing to the diversity. Consistent with the previous studies, 3-IRD PIs were highly abundant in C. annuum leaves (Table 1, Figure 2). We observed a strong upregulation of 3-and 4-IRD CanPI transcripts in induced leaves when compared to uninduced leaves ( Figure 1A, Table 1). Reciprocating patterns obtained at the PI activity level corroborate the induction specific regulation of CanPIs. Other reports on C. annuum have also demonstrated an increase in CanPI expression in local and systemic leaf tissues upon elicitation by aphids, viruses, insect feeding and mechanical wounding [16,25,26]. The stronger induction of PI activity in distant leaves than in local leaves challenged with the same treatment was interesting and suggested strong distant signalling. Despite the difficulties of directly comparing the aphid infestation treatment with other treatments carried out in the present study, an interesting pattern of accumulated CanPIs is evident at transcript and PI activity level. CanPI transcript abundance in aphid infested leaves was much lower than that observed in wounded leaves treated with water or OS ( Figure 1A). Aphids have been found to elicit defense related genes including PIs but the responses are low as compared to that elicited by chewing insect attack [36,37]. However, studies have also found aphid induced transcriptional signatures of salicylic acid signaling and no increase in PI transcripts [38]. In the present study, transcripts of CanPI-41 and −43 are highly accumulated under aphid infestation while CanPI-8 and −11 remain either un-elicited or suppressed by aphid damage; the latter being up-regulated by wounding or W+OS treatment. The suppression of selected wound induced responses, rather than the lack of cellular damage, might also be responsible for the low responses to aphid attack [36]. Aphid infestation induced CanPI activity though much less as compared to wounded leaves treated with water or OS ( Figure 4A). A unique aphid induced CanPI signature was evident in the 2D activity profiles ( Figure 5).
Particularly, 4-IRD PIs were strongly induced in wounded leaves treated with OS ( Figure 1B) highlighting the strong and specific effects of insect elicitors on CanPI regulation. Plant responses to wounding/insect feeding are known to be specifically altered by the plant's perception of herbivore-specific elicitors [8,[39][40][41][42]. High amounts of jasmonic acid and the rapid accumulation of wound-inducible transcripts have been reported in response to insect damage or insect OS when compared to only mechanical wounding [5,40,43]. Wounded leaves treated with water showed the highest number (10) of uniquely expressed CanPIs ( Figure 1C), though in lower frequency as compared to a few CanPIs (−4, -7, -10) and IRDs (4,5,10,14,25) with higher frequency in wounded leaves treated with OS. Specifically high representation of CanPIs with multiple IRDs directed towards enriching the PI blend with both CI and TI activities seems to be an approach adapted by the plant upon Lepidopteran insect attack, helpful in tackling a wide range of insect proteases [31]. Significantly high PI activity ( Figure 4A) and detection of three PI activity bands in wounded leaves treated with water or OS in comparison to two activity bands in uninduced and aphid infested leaves ( Figure 4B) is indicative of quantitative and qualitative variations in the accumulated PI activity. Further characterization by 2D electrophoresis, revealed the presence of multiple charge and/or molecular mass variants observed in wounded leaves treated with water ( Figure 5; TI-7 to −13) and with OS ( Figure 5; TI-14 to TI-16) clearly indicating the induced isoform diversity. The differential isoforms detected mostly corresponded to 3-and 4-IRD PIs and thus correlates with the high CanPI transcript accumulation under these two treatments. The absence of certain TI isoforms in OS treated leaves compared to wounding alone, suggests the suppression of some induced responses, resulting in treatment specific patterns. With respect to HGP inhibition potential, the PI activity in all leaf tissues attained 70% inhibition of HGP ( Figure 4C) and could inhibit almost all the HGP isoforms ( Figure 4D). However, an early saturation of HGP inhibition by proteins from leaves induced by wounding and/or treated with OS, as compared to uninduced leaves, is suggestive of the high quantitative accumulation of PI units in such leaf tissues and higher specific activity against insect gut proteases. Multiple IRDs are known to be generated from precursor N. alata PI proteins in Me-JA-elicited leaves [22,44] and from CanPI precursor proteins by the action of endogenous proteases at the linker regions [16]. An increase in the number and intensity of variant mass peaks equivalent to single IRDs, in wounding with water and with OS protein fractions (Figure 6), suggests the enhanced proteolytic processing of the up-regulated CanPI precursor proteins [22,31,32]. The specific presence of isoform TI-6 and absence of TI-1 ( Figure 5) under all inductions also indicate the accumulation of differential 1-IRD isoforms that are generated as a result of processing of CanPI precursors. Thus, our results substantiate the hypothesis [22] that elicitation leads to over-production of the CanPI precursors and enhanced, differential processing of the precursors by proteases to IRDs, resulting in structurally and functionally diverse processed products. It was also noticed that induction treatment specificity is maintained even at the level at which precursor proteins are processed. Peaks ranging from 5.9 to 6.3 kDa show high intensity in wounding with OS treatment whereas the peaks from 5.5 to 5.8 kDa are prominent in uninduced and other treatments, aphid infested and wounding treated with water. These results affirm that plants can differentially perceive various kinds of biotic stresses and respond appropriately through regulation of PIs at transcriptional and posttranslational levels. Sequence analysis revealed highly homologous CanPIs with an average variance of 4%. The clear absence of partial N-and C-terminal repeats in the CanPI precursors groups them as distinct from N. benthamiana Pin-II PIs (Figure 2). The explicit clustering of C. annuum PIs from all other Solanaceous Pin-II PIs suggests recent evolutionary origins [21]. The diversity in CanPIs can be attributed to individual IRDs, which display a sequence variation ranging from 2 to 20% within the vicinity of the reactive site loops and C-terminal region. Twenty eight unique IRDs, constituting 7 CIs and 21 TIs, follow the H-L type topology, where the sequence repeat is identical to the structural repeat [21]. The induction-specific IRD distribution is predominantly biased towards TIs rather than CIs. It is known that in Lepidopteran insects, trypsin-like proteases are predominant which could be correlated to the relatively high abundance of trypsin specific PIs in plants. Active site variants of TIs 'CPRNC' , 'CPKNC' , 'CPRYC' and 'CPRDC' and two types of CI sites 'CTLNC' and 'CTPNC' were present among all identified 28 IRDs. Interestingly four cysteine variants either, missing one or more conserved cysteine residues, change in position of cysteines or having additional cysteines were identified in the present study ( Figure 3A). Recently, six natural IRD variants with selective losses of cysteine residues have been identified in potato [45]. The loss of cysteine residues is often associated with functional differentiation and suggests positive evolutionary gene selection. However, the nature of mutations and the associated factors responsible for selective losses of cysteines remains unclear. Studies on the significance of such mutations affecting the proteinase binding affinity and structural stability/integrity of IRDs have initiated an active debate on the evolution of disulfide bonds in the Pin-II family ( [45,46], our unpublished results). Particularly in C. annuum, the identification of four such IRDs as a result of various inductions suggests that plants elaborative defenses by expressing modified IRDs that improve their overall activity against target proteases. Among various tissues from C. annuum, flowers revealed the highest accumulation of PI activity (Additional file 4: Figure S3), consistent with a role in protecting the reproductive parts of the plant against pests as reported in tomato [15,44].

Conclusions
This study suggests that CanPI sequence diversity, tissue specificity and explicit responses to different inductions are part of effective plant defense system. The significance of huge complexity of PIs observed specifically in C. annuum needs to be understood. Recent reports on the endogenous and/or defensive roles of PIs from various Solanaceous species and simultaneous expression of multiple CanPIs constitutively highlight their prospective involvement in many of the plant's complex processes [16,19,23,[47][48][49]. Moreover, up-regulated yet specialized CanPI expression upon wounding and insect infestation provides insights into the evolution of PI based plant defense mechanisms against insects and generates many unanswered questions about their regulation. Essentially, more functional studies need to be performed for specific CanPI genes in order to ascertain their roles under a particular treatment and how this variation accounts for the fitness benefit of the plant under specific biotic stress conditions.

Plant material and induction treatments
C. annuum seeds (cv PhuleJyoti) (diploid) were grown in pots with Soilrite (Mixture of horticulture grade expanded perlite, Irish Peat moss and exfoliated vermiculite in equal ratio; Naik Krushi, Pune, MS, India) and supplemented daily with Hoagland solution. 30-dayold seedlings were transferred to individual pots and grown in a growth chamber maintained at 23°C (±2°C) with a 14 h light photoperiod. Leaves, stems, various stages of fruits (early, mid, turning and late), roots and flowers from mature plants (3 months old) were harvested for screening tissue-specific CanPIs.
All induction experiments were performed on 3month-old plants and a set of 3 plants was taken per treatment for each biological replicate. Leaves were mechanically wounded by rolling a fabric pattern wheel along the length of the lamina (4 to 6 rolls depending on the size of leaf ) and the resulting puncture wounds were immediately treated with water or OS. These were considered local tissue, whereas the non-wounded leaves one node above or below were harvested to measure systemic responses. Local and systemic tissues were collected after 30 h of treatment. OS used was collected from H. armigera larvae and diluted 1:50 (v/v) times in MQ water. Plants were kept in an open garden for passive aphid infestation. Natural infestation by Myzus persicae was observed on C. annuum leaves within a week. The density of aphids was high on the leaf lamina towards the petiole. Leaves with at least 20 nymphs per leaf growing at the same nodes were collected as local tissue whereas non-infested leaves were harvested as systemic tissues. Since the plants were naturally infested by aphids in open conditions, the possible comparisons with this treatment and the other two, which were performed in a highly specific and controlled manner, are limited. Leaves from un-elicited control plants growing at the same nodes were harvested from uninduced plants. All the tissue collections were done at a same time and were flash-frozen in liquid nitrogen and stored at -80°C until further use. Two biological replicates were used for the whole study.
Expression profiling, cloning of Pin-II genes and sequence analysis Total RNA from C. annuum leaf tissues (uninduced and all three inductions-systemic) was isolated using TRI-ZOL (Invitrogen, Carlsbad, CA, USA) followed by DNAse treatment at 37°C. Purified RNA was quantified by spectrophotometry, and 1.5 μg was used for firststrand cDNA synthesis using a reverse transcriptase kit (Promega, Madison, WI, USA). Proof-reading Accuprime Pfx DNA polymerase (Invitrogen) was used to amplify cDNAs from systemic leaf tissues of individual treatments and uninduced leaf in independent PCR reactions using CanPI gene (Genbank accession: AF039398) specific primer pair (CanPin-1F and CanPin-1R; Additional file 1 Table S1A). Amplicons were cloned into pGEMTeasy vector (Promega). 60 cloned fragments from each treatment were sequenced using standard T7 forward and SP6 reverse primers. Sequence editing and analysis was carried out using BioEdit, Clustal-X and Lasergene software. For semi-quantitative analysis of CanPIs within the tissues, independent PCRs were performed with reduced number of cycles (25), Accuprime Pfx DNA polymerase and CanPI gene specific primer pair. Specific primer pairs were designed for individual CanPI genes in order to check the expression of specific CanPIs. However, due to a high degree of similarity/homology within C. annuum PIs, it was possible to design genespecific primer pairs for CanPI-3, -5, -7, -8 and −10 only. The specific pairs of oligonucleotides used for the internal differentiation of individual CanPIs are stated in Additional file 1: Table S1A and B. All PCRs were carried out in technical duplicates.

Protein extraction and proteinase inhibitor activity assays
Total soluble protein was extracted from 1g of fresh leaf tissue obtained from uninduced and induced (local and systemic for all treatments) C. annuum, using a mixture of water and 5% polyvinylpolypyrrolidone (Sigma, St. Louis, MO, USA). Following protein estimation using Bradford reagent (Bio-Rad Laboratories, Hercules, CA, USA), the trypsin inhibitory activity and H. armigera gut protease (HGP) inhibitory activity was estimated enzymatically in the leaf extracts using the synthetic substrate Benzoyl-DL-Arginyl-p-Nitroanilide (Sigma, St. Louis, MO, USA) as described earlier [28]. Protease inhibitor activity was expressed as trypsin inhibitory units per mg of tissue (TIUs/mg). Various amounts of the leaf protein extracts were titrated against HGP to determine maximum inhibitory activity. All the assays were carried out in technical triplicates with 2 biological replicates and statistical analysis of data was performed using singlefactor ANOVA followed by Tukey's post hoc analysis.
1D and 2D electrophoresis for in-gel identification of proteinase inhibitory activity Equal TIUs of leaf protein extracts were resolved on 12% native-PAGE and further processed in order to visualize trypsin inhibitor isoforms using the previously described gel X-ray film contact print (GXCT) method [50]. To visualize protease activity, protein extracts incubated with HGP were resolved on 8% native-PAGE gel and processed by GXCT [51]. For 2D electrophoresis, acetone-precipitated proteins were re-suspended in rehydration buffer and separated in first dimension (Isoelectric focusing) using 11-cm IPG strips (pH 3-10 NL; Bio-Rad Laboratories, Hercules, CA, USA) as per the manufacturer's protocol. Second-dimension separation was done on 12% SDS-PAGE gel using a Hoefer electrophoresis unit (GE Healthcare Bio-sciences AB, Buckinghamshire, UK) maintained at 24°C and 200 V. Trypsin inhibitor activity was visualized by modified GXCT [52].
All the electrophoresis experiments were carried out in technical triplicates with 2 biological replicates.

Partial purification and proteomic analysis of leaf protein extracts
To identify the proteinase inhibitors from C. annuum leaf tissues, protein extracts were first partially purified in the following manner. Ammonium-sulphate-precipitated proteins (90% saturation) from leaf extracts were resuspended in 50 mM Tris buffer (pH 8.0), and subsequently treated at 65°C for 10 min before being desalted using PD Spin-Trap TM G-25 column (GE Healthcare). Proteins were then separated on DEAE-Sephacel (GE Healthcare) equilibrated with 50 mM Tris buffer (pH 8.0), and flow-through was collected separately from bound proteins that were eluted with a NaCl gradient of 0.25 M to 0.4 M in 50 mM Tris buffer (pH 8.0). Fractions with trypsin inhibitor activity were concentrated and desalted using a PD Spintrap column. The partially purified protein fractions were pooled and qualitatively analyzed by matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF-MS), Voyager-De-STR (Applied Biosystems, Framingham, MA, USA). Mass spectral acquisition was performed with a standard instrumental protocol as described earlier [31]. In brief, 2 μg of protein sample was mixed with 20 μL of freshly prepared sinapinic acid (Sigma-Aldrich) (in 30% acetonitrile [ACN], 0.1% trifluoroacetic acid [TFA]) and spotted on the stainless steel MALDI plate, and spectral profiles were acquired in the range of 1 to 25 kDa. The spectra were analyzed with Data Explorer ™ for regions of interest and processed for advanced base-line correction and noise removal.
To obtain protein sequence data, the partially purified proteins were separated on 16% Tricine SDS-PAGE gel [53] to resolve the low molecular mass proteins, after which protein bands were in-gel digested and analyzed by MALDI-Q-TOF-MS (SYNAPT High Definition Mass spectrometer, Waters Corporation, Milford, MA, USA). Mass spectral acquisition was carried out by MALDI survey method. Protein Lynx Global server version 2.4 software (Waters) was used for data processing and database searches. The MS/MS data was searched against the Pin-II protein database constructed separately using the following parameters: peptide tolerance of 20 ppm, fragment tolerance of 0.05D, one missed cleavage, carbamidomethylation of cysteines and possible oxidation of Methionine.

Additional files
Additional file 1: Table S1. Oligonucleotide primers used for RT-PCR and CanPI internal differentiation.