Distinct properties of proteases and nucleases in the gut, salivary gland and saliva of southern green stink bug, Nezara viridula

Stink bugs negatively impact numerous plant species of agricultural and horticultural importance. While efforts to develop effective control measures are underway, the unique digestive physiology of these pests presents a significant hurdle for either protein- or nucleotide-based management options. Here we report the comparative biochemical and proteomic characterization of proteases and nucleases from the gut, salivary gland and saliva of the southern green stink bug, Nezara viridula. The pH optimum for protease activity was acidic (5 to 6) in the gut with the primary proteases being cysteine proteases, and alkaline (8 to 9) in the saliva and salivary gland with the primary proteases being serine proteases. The serine proteases in saliva differ biochemically from trypsin and chymotrypsin, and the cathepsins in the gut and saliva showed distinct properties in inhibitor assays. Nuclease activity (DNase, RNase, dsRNase) was concentrated in the salivary gland and saliva with negligible activity in the gut. The most abundant proteins of the gut (530) and salivary gland (631) identified by proteomic analysis included four gut proteases along with eight proteases and one nuclease from the salivary gland. Understanding of N. viridula digestive physiology will facilitate the design of new strategies for management of this significant pest.


Results
Salivary gland and gut morphology. N. viridula adults have two salivary glands with two major lobes, the principal salivary gland (PSG) and the accessary salivary gland (ASG), along with a salivary duct connected at the junction of the PSG and ASG (Fig. 1). The salivary glands were flanked by and attached to the first section of the midgut. The gut is divided into four sections (M1 to M4) as described by Tada et al. 13 (Fig. 1). Section M4 has a crypt-like structure attached to the hindgut, which harbors symbiotic microorganisms and does not function in digestion 14 . Total proteolytic activity and pH optima. The total protease activity and specific protease activity for N. viridula gut, salivary gland and saliva using azocasein as substrate are shown in Fig. 2. While the total protease activity was highest in the gut, the specific activity was lowest in the gut compared to saliva and salivary gland (Fig. 2). The high specific protease activity in saliva reflects the relatively low protein content. The optimum pH for protease activity in the salivary gland and saliva was 8 to 9 and in the gut was 5 to 6 ( Fig. 2).
Inhibitor assays. The profile of protease inhibition differed significantly between the gut and salivary gland or saliva (Fig. 3). The cysteine protease inhibitor E-64 inhibited more than 90% of gut protease activity but did not affect protease activity from saliva or salivary gland. However, protease activity in the salivary gland and saliva was significantly inhibited (~80% inhibition) by a class-specific serine protease inhibitor PMSF. The trypsin inhibitor TLCK inhibited 20%, 80% and 70% of protease activity from the gut, salivary gland and saliva respectively. The chymotrypsin inhibitor TPCK inhibited around 40% of protease activity from saliva or salivary gland, but only 10% activity from gut (Fig. 3). These results indicate that cysteine proteases predominate in the gut, with trace amounts of trypsin-like and chymotrypsin-like activity. In contrast, serine proteases predominate in the saliva and salivary gland, along with significant metalloprotease activity.
Specific substrate and class specific inhibitor assays. To delineate the specific types of protease activity present, different chromogenic substrates were used with a combination of inhibitors which are specific to a selective protease. All five enzyme types were detected in the gut (Fig. 4). Salivary gland and saliva had significantly higher chymotrypsin-like, trypsin-like and cathepsin activities compared to the gut (P ≤ 0.05; Student's t-test), and no cysteine protease activity. In contrast, cysteine protease activities were the most abundant in the gut compared to saliva and salivary gland. PMSF significantly inhibited protease activity from salivary gland and saliva, but TLCK and TPCK did not significantly inhibit trypsin and chymotrypsin activity (Fig. 4). These results indicate the presence of serine proteases with different specificity (other than trypsin and chymotrypsin) in salivary gland and saliva. As expected, E-64 significantly inhibited gut cysteine protease activity, and EDTA inhibited aminopeptidase activity from gut, saliva and salivary gland. While E-64 inhibited ~50% gut cathepsin, the same activity was not inhibited by E-64 in salivary gland and saliva (Fig. 4). This result indicates that cathepsins in the gut are cysteine proteases (sensitive to E64) while cathepsins in saliva belong to a different cathepsin class. Taken together, high cysteine protease activity was found in the N. viridula gut, while high serine protease, cathepsin and aminopeptidase activities were found in salivary gland and saliva.
Nuclease activities. The DNase, RNase and dsRNase activities in N. viridula gut, salivary gland and saliva were determined. For DNase and RNase activities, the specific activities were highest in saliva and salivary gland, and relatively low in the gut (Fig. 5a). The degradation pattern of DNA visualized on agarose gels supported the relative levels of nuclease activity with complete degradation of DNA by salivary gland and salivary enzymes within 5 min, and relatively little degradation by gut enzymes (Fig. 5b). The specific activity of dsRNase was the highest in saliva at 15.76 U/μ g with only low activity (0.385 U/μ g) in gut and salivary gland samples (Fig. 6). The rate of dsRNA degradation visualized in agarose gels reflected the activity assay data with only minimal dsRNA degradation seen in gut and salivary gland samples at 20 min (Fig. 6). Taken together, these results demonstrate that nucleases (DNase, RNase, dsRNase) are abundant in N. viridula saliva and salivary gland, but present at relatively low levels in the gut.
Proteomic identification of proteases and nucleases. Proteomic analysis to identify proteases and nucleases from N. viridula resulted in identification of 530 and 631 proteins from the gut and salivary gland respectively (Supplementary File 1). Identified enzymes (highlighted in Supplementary File 1) were computationally annotated on the basis of their metabolic function. Approximately 32% of gut and 26% of the salivary gland enzymes were involved in energy metabolism whereas 23% of the enzymes function in protein metabolism (Fig. 7). Significant proportions of the enzymes identified function in detoxification and antioxidation (10% gut and 8% salivary gland), and in carbohydrate metabolism (15%). Proteases comprised 3% and 4.25% of enzymes from the gut and salivary gland respectively (Fig. 7). The higher number of proteases and nucleases detected in the salivary gland compared to the gut supports the higher activity levels for some protease and nuclease groups (e.g. serine proteases and dsRNases) in salivary gland extracts.
Four gut proteases were identified along with eight proteases and one nuclease from salivary gland. The gut proteases were one aminopeptidase, two putative aminopeptidases and one cathepsin-like cysteine proteinase (Table 1). In the salivary gland, aminopeptidases and metalloproteases with cathepsin D, cysteine protease and a nuclease were identified (Table 1). Sequence alignment showed that although the gut and salivary gland proteases belong to similar mechanistic classes, their sequences differ ( Supplementary Fig. S1). An exception was aminopeptidase N-10 which shared 100% sequence identity between the two tissues. Salivary gland proteases, including two metalloproteases, one methionine aminopeptidase and a ubiquitin-specific peptidase formed a separate clade on phylogenetic analysis. A cathepsin L-like cysteine protease from the gut and a "26,29 kDa proteinase-like" cysteine protease from salivary gland shared ~30% sequence identity and formed a separate group. An aspartate protease cathepsin D and a signal peptidase from salivary gland shared very low sequence similarity with other proteases and stand alone in the phylogenetic tree ( Supplementary Fig. S1). Both sequence analysis and activity assays highlight the presence of different proteases and nucleases in the gut and salivary gland.

Discussion
The use by stink bugs of both extra-oral and gut-based digestion of host plant materials using enzymes with distinct pH optima allows for a two-pronged approach for efficient use of nutrient value from plant proteins, and for efficient degradation of toxins or plant defense molecules. The high enzymatic activity for efficient digestion likely reflects the unusual digestive physiology of these insects, with section M4 of the alimentary tract harboring symbiotic bacteria and being blocked to ingested material 14 , requiring complete digestion and absorption of nutrients in sections M1-M3. The enzymatic challenge faced by material ingested by stink bugs poses a significant problem however for development of novel strategies for stink bug management. Insect specific toxins such as those derived from the bacterium Bacillus thuringiensis 15 , and silencing RNAs designed to block translation of insect proteins essential for survival 16,17 , are likely to be inactivated before reaching their target site within the stink bug. Here we undertook biochemical characterization of digestive enzymes in the gut, salivary gland and saliva of N. viridula to define the overall enzymatic composition of these tissues.
The structure of the N. viridula salivary gland was comparable to that of the brown marmorated stink bug, Halyomorpha halys 18 , with two lobes, the accessary and principal salivary glands that produce watery and sheath (or gelling) saliva respectively 19,20 . Consistent with high enzymatic activity in watery saliva of other species 21 , N. viridula saliva had the highest specific protease activity of the three tissues. Based on the alkaline pH optimum, and sensitivity to serine protease inhibitors, the salivary proteases of N. viridula are predominantly of the serine mechanistic class. Interestingly, protease activity in salivary gland and saliva was inhibited by PMSF, a class-specific serine protease inhibitor, while the activity was not inhibited by TLCK and TPCK, specific inhibitors of trypsin and chymotrypsin-like enzymes respectively. N. viridula therefore has serine proteases in saliva that differ from trypsin and chymotrypsin. Similarly, the gut and salivary cathepsins differ as gut cathepsins were inhibited by E-64 (a cysteine proteases inhibitor) but salivary gland and salivary cathepsins were not.
While the biochemical characteristics of stink bug digestive enzymes have not previously been assessed, the digestive enzymes of two other species of Hemiptera, both predaceous insects in the suborder Heteroptera have been investigated. The salivary and gut proteases of Zelus renardii (Reduviidae) 22,23 showed optimal activity at pH 7 to 8.5, with serine proteases predominant in the salivary gland, and carboxypeptidase and aminopeptidase in Figure 3. Impact of synthetic protease inhibitors on total protease activity in gut, salivary gland and saliva of Nezara viridula. Samples were mixed with inhibitors and pre-incubated for 30 min at 37 °C. Residual protease activity was assayed through spectrophotometric assay using Azocasein as substrate. Protease activity without inhibitor (control) was considered 100%. Statistical comparisons were conducted for each treatment compared to the control (* P ≤ 0.01; * * P ≤ 0.001; * * * P ≤ 0.0001; Student's t-test).
the gut. In contrast to N. viridula, which has aminopeptidase activity in both the gut and saliva, aminopeptidase activity was only detected in the gut of Z. renardii. Cysteine protease activity was not assessed. The pH optima and enzyme distribution in the spined soldier bug, Podisus maculiventris (Pentatomidae) 24 was more comparable to that of N. viridula (Pentatomidae) including differential sensitivity of gut and salivary proteases to different inhibitors.
Similar to the gut environment of Coleoptera and other Hemiptera, and in contrast to that of Lepidoptera 25 , an N. viridula gut pH of 5.5 was recorded with cathepsins as the predominant protease type. These results are consistent with the pH optima and enzymatic composition in the gut of other Hemiptera, such as the predatory bug Podisus maculiventris, Rhodnius prolixus and Euschistus euschistoides 23,24 . Trace amounts of trypsin and chymotrypsin were also found in the gut of P. maculiventris 26 , similar to N. viridula. Knowledge of the proteolytic composition of the N. viridula saliva and gut lumen environments provides for direct suppression of proteases through plant expression of protease inhibitors 27,28 . The application of protease inhibitors to disrupt digestive processes for suppression of lepidopteran pests however resulted in upregulation of other protease types, highlighting plasticity within insect digestive physiology 28 . Similar adjustment may result from exposure of stink bugs to plant-expressed protease inhibitors.
While dsRNase activity has been detected in the gut of insects such as the lepidopteran, Bombyx mori and orthopteran, Schistocerca gregaria [29][30][31] , the release of salivary ribonucleases by Hemiptera into host plant tissues represents a particular problem for application of RNA interference-based approaches for suppression of hemipteran pests. Saliva and salivary gland extract from the tarnished plant bug Lygus lineolaris rapidly degraded dsRNA 32 , dsRNase was released during feeding of the pea aphid, Acyrthosiphon pisum 33 and high dsRNase activity levels were detected in the saliva of N. viridula in the present study. The high level of dsRNase activity in saliva compared to the very low level of activity detected in the salivary gland suggests that dsRNases are produced as zymogens that are activated on release into the saliva. This in turn suggests that the presence of active dsRNases in the salivary gland may be harmful to the stink bug. In addition to dsRNase activity, high levels of specific DNase and RNase activity were detected in the salivary gland and saliva of N. viridula. The DNase activity levels were comparable to that of the purified commercial enzyme (TURBO ™ DNase), suggesting an important role for DNases in the digestive process and / or defense. Nucleases in saliva directly or indirectly protect insects from infection by viruses present in the host plant, be they insect pathogens, or insect vectored viruses of plants 34 . Nucleases therefore are likely to provide defense against pathogens, in addition to mediating digestion of host plant-derived nucleic acids.
The results of functional annotation of enzymes identified by proteomic analysis were consistent with the roles of the gut and saliva in digestion, detoxification, and protection against microbes. Digestive enzymes comprised 23% of identified proteins, consistent with proteomic analysis of H. halys watery saliva in which 29% of identified   proteins were digestive enzymes 18 . The abundance of proteases and nucleases in the gut and salivary gland of N. viridula is similar to enzyme abundance in other insect species 35,36 , and the divergence in sequences between salivary and gut proteases supports the presence of enzymes with distinct properties in these two tissues. The abundance of aminopeptidases in the gut and salivary gland proteomes was in accordance with high aminopeptidase activity in both tissues.  Our results highlight the different enzymatic composition of N. viridula salivary gland and gut for efficient digestion of host plant components, detoxification of proteins used in plant defense responses, and for defense against microbes such as viruses. The distinct profiles of protease and nuclease activities in the two tissues present a unique challenge for the design of control agents that can withstand such a two-pronged enzymatic attack.
Colony maintenance, dissection and collection of saliva. N. viridula were reared under controlled conditions (16 h light; 8 h dark, 27-28 °C, 65% humidity) and fed a mixed diet containing green beans, sweet corn, carrots and raw peanuts. Adults were chilled for one hour over ice to render them immobile and then dissected. For enzyme assays, the midgut and salivary glands were isolated under a dissection microscope in a plastic Petri dish filled with phosphate-buffered saline (PBS: 0.1 M pH 7.0). Isolated tissues were snap frozen in liquid nitrogen and kept at − 80 °C until use. Saliva from N. viridula adults was collected as described previously 13 . Briefly, adult insects were chilled on ice for five min, then placed in a Petri dish, ventral side up and observed under a dissecting microscope. As the insects warm to room temperature, watery saliva is released from the tip of the beak. This saliva was collected into filter pipet tips and expelled into a 1.5 ml tube containing 10 μ L of phosphate-buffered saline (0.1 M pH 7.0). Samples were stored at − 80 °C until use.
Preparation of gut and salivary gland extracts. Whole guts and salivary glands were ground in liquid nitrogen to a fine powder using a mortar and pestle. The tissue powder was homogenized in phosphate-buffered saline (0.1 M pH 7.0) at a ratio of 1:6 (w/v) and the homogenate was centrifuged twice at 10,000 g at 4 °C for 20 min. The supernatant was collected and used as gut or salivary gland extract. Protein concentrations in the extracts were determined by the Bradford method 37 using BSA as standard.
Total proteolytic activity assay. Total proteolytic activity in gut, salivary gland and saliva (5 μ g of protein from each sample) was determined using azocasein as substrate 38 with 1% (w/v) solution of the substrate in assay buffer (0.1 M Tris-HCl pH 7.0). Gut, salivary gland extracts and saliva (10 μ L) were added to 200 μ L of substrate and incubated overnight at 37 °C. The reaction was terminated by adding 300 μ L of 5% trichloroacetic acid (TCA). Precipitated proteins were centrifuged (10,000 g, 10 min) and 0.5 mL of the supernatant was added to 0.5 mL of 1 M sodium hydroxide. The absorbance was measured at 450 nm (Nanodrop 2000 spectrophotometer; Thermo Fisher Scientific, Waltham, MA, USA) and an increase of 0.01 OD per min defined as one protease activity unit.
Optimum pH of proteolytic activity. The pH optima of N. viridula proteolytic activities in different tissues were determined by adding extracts (10 μ L each) to buffers (100 μ L) with pH ranging from 4 to 10 (pH 4 to 10 buffers; sodium-acetate pH 4, 5 phosphate pH 6, 7 Tris-HCl pH 8 and glycine-NaOH pH 9, 10) added to 100 μ L of azocasein prepared in distilled water. The assays were carried out as described above.
Protease inhibition assays. The synthetic inhibitors PMSF (5 mM), TPCK (100 μ M), TLCK (100 μ M), E-64 (10 μ M) and EDTA (10 μ M), were mixed with gut extract, salivary gland extract and saliva samples (5 μ g protein per sample in a volume of 10 μ L) and incubated for 30 min at 37 °C. The protease extract-inhibitor mixture was added to azocasein solution (200 μ L, prepared in phosphate buffer 0.1 M pH 6.0 for gut and Tris-HCl buffer 0.1 M pH 8.5 for salivary gland and saliva). Reactions were incubated at 37 °C and terminated by addition of TCA (300 μ L, 5%). Protease activity was determined as described above.
Specific substrate and class-specific inhibitor assays. The specific substrates and inhibitor combinations used were 1) BApNA and TLCK or PMSF, 2) SAAPFpNA and TPCK or PMSF, 3) LpNA and EDTA, 4) pGPLpNA and E-64, and 5) AApNA and E-64. Substrates were solubilized in DMSO prior to dilution in buffer (phosphate buffer pH 6.0 for gut, Tris-HCl pH 8.5 for salivary gland and saliva). Synthetic inhibitors listed above were mixed with the gut and salivary gland extract, and saliva samples (10 μ L each) and incubated for 30 min at 37 °C. Reaction mixtures (20 μ L) were added to 100 μ L of each substrate (10 mM) and incubated overnight at 37 °C. Reactions were terminated by adding acetic acid (100 μ L, 30%). The absorbance was measured at 410 nm and protease activity defined as production of 1 μ mol of p-nitroaniline/mg of protein/min at 37 °C.
Nuclease activity assays. DNase, RNase and dsRNase activities were determined on the basis of the release of acid soluble hydrolysis products from DNA, RNA or dsRNA 39 . Calf thymus DNA or yeast RNA was used as substrate, dissolved in 20 mM Tris-HCl, pH 8.0 containing 25 mM NaCl, 10 mM MgCl2, and 5 mM CaCl2. Gut, salivary gland extracts and saliva (5 μ g protein per sample in a volume of 10 μ L) were added to 200 μ L DNA or RNA (100 μ g/mL) and the reaction was carried out at 37 °C for 30 min. TURBO ™ DNase (5 units) (Life Technologies, Carlsbad, CA, USA) and RNase A (20 μ g) (Thermo Scientific, Waltham, MA, USA) were used as positive controls. The reaction was stopped by addition of 300 μ L of 10% TCA containing 20 mM sodium pyrophosphate. For dsRNase activity assays, dsRNAs were prepared using the Ambion ® MEGAscript ® RNAi Kit (Life Technologies, Carlsbad, CA, USA) following the manufacturer's protocol. A non-relevant dsRNA (Brugia malayi cathepsin F gene sequence) was used as substrate. Samples (10 μ L) were added to 20 μ L substrate (0.1 μ g/μ L in buffer as described above) and the reaction was carried out at 37 °C for 30 min. The reaction was stopped by the addition of 30 μ L of 10% TCA containing 20 mM sodium pyrophosphate. The nucleic acids from all reactions Scientific RepoRts | 6:27587 | DOI: 10.1038/srep27587 were precipitated on ice for 1 h and centrifuged (10,000 g, 10 min). Absorbance of the supernatant was measured and one unit of enzyme activity defined as an increase in A 260 of 0.01 per min.
Degradation of nucleic acids. Calf thymus DNA (1 μ g/μ L) was incubated with gut, salivary gland extract and saliva (2 μ g protein per sample in a volume of 5 μ L) at 37 °C for 5, 10, or 30 min. Samples were run and visualized on agarose gels (2%). N. viridula nuclease activity was assayed using GFP dsRNA and non-relevant dsRNA (Brugia malayi cathepsin F, L1 and L2). dsRNAs (2 μ L each) were incubated with tissue samples (5 μ L each) for 5, 10, or 20 min and degradation assessed following agarose gel electrophoresis. Nucleic acids in the gel were visualized using ethidium bromide.
Proteomics. Gut (midgut and hindgut) and salivary gland (ASG, PSG and salivary duct) tissues were isolated from adult N. viridula, snap frozen, and stored at − 80 °C. Tissues were ground in liquid nitrogen and mixed with 1mL Trizol in a vial containing glass beads. The mixture was homogenized using FastPrep ® -24 homogenizer (MP24, MP Biomedicals, Solon, OH, USA) and centrifuged (10,000 g/20min/4 °C). Proteins from the supernatant were isolated using the Invitrogen Protein Isolation Kit (Invitrogen, Carlsbad, CA, USA). Supernatant and pellet samples were mixed with Agilent Isoelectric focusing (IEF) OFFGEL buffer, separated by SDS-PAGE (12% gel) and visualized by silver staining (Supplementary Fig. S2). Protein bands were excised from the gel and in gel digested with trypsin according to published procedures 40 42 with Scaffold delta-mass correction. Protein identifications were accepted if they could be established at > 99.0% probability and contained at least 3 identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm 43 . Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. Proteins with enzymatic activity were functionally annotated with GO terms from gene_association.goa_uniprot (downloaded Nov 21, 2015) 44 . For phylogenetic analyses, protease sequences representing the peptides identified in gut and salivary gland were extracted from the database (BMSB translated transcriptome_ 20150105 and TrEMBL_LarvalBugs_20141201 database) that were used for proteomic identification. The sequence analysis was performed using EMBL-EBI Multiple Sequence Alignment Tool, ClustalW2 (http://www.ebi.ac.uk/Tools/msa/clustalw2/) and phylogenetic tree was generated.

Statistical Analysis.
All experiments with the exception of proteomic analysis were performed with two biological replicates, each with three technical replicates and with each sample assayed twice. The Student t-test was used to assess statistical significance for enzyme activity assays.