Alpha-arylphorin is a mitogen in the Heliothis virescens midgut cell secretome upon Cry1Ac intoxication

Bacterial toxin

Bacillus thuringiensis var. kurstaki strain HD73 producing Cry1Ac toxin was obtained from the Bacillus Genetic Stock Center (BGSC, Columbus, OH, USA). Bacterial culturing, toxin activation and purification were as described elsewhere (Perera et al., 2009). Purity of activated toxins was assessed by SDS-10% PAGE (Fig. S1), and protein concentration quantified using the Coomassie Plus Protein Assay (Pierce, Waltham, MA, USA) with bovine serum albumin (BSA) as standard. Purified toxin samples were kept at −80 °C until used (less than six months).

Insects and toxin feeding

Eggs from the Cry1Ac-susceptible YDK strain of H. virescens (Gould, 1995) were kindly supplied by Dr. Fred Gould (North Carolina State University). Upon hatching, larvae were reared on artificial diet (Bio-Serv, Flemington, NJ, USA) at 28 °C on an 18L:6D photoperiod.

Establishment of primary midgut cell cultures

All dissections and transfers were performed in the sterile environment of a biosafety cabinet. Preparation and establishment of primary midgut cell cultures were done following protocols described previously (Castagnola, Eda & Jurat-Fuentes, 2011). Briefly, midguts of fourth instar H. virescens larvae were dissected and cleaned of food, peritrophic matrix, and Malphigian tubules using forceps and rinsed in sterile Ringer’s (Barbosa, 1974) containing 0.5% (v/v) gentamicin (Invitrogen, Carlsbad, CA, USA), 0.1% bleach, and 1× antibiotic/antimycotic (Gibco 15240-062; Gibco, Grand Island, NY, USA). Incubation media was prepared by mixing working Grace’s (supplemented Grace’s Insect Medium [Invitrogen] containing 1× antibiotic/antimycotic and 0.1% gentamicin) in sterile Ringer’s solution in a 3:1 ratio (Loeb, Brandt & Birnbaum, 1984). For each primary cell culture preparation, five pools of five to six cleaned midguts were cut in sections and incubated in 2 ml of incubation media for 90 min at room temperature. After incubation, the midgut tissue was homogenized by carefully pipetting up and down and then sieved through 70 µm cell strainers (BD Biosciences, NJ) into a sterile 50 ml conical tube. Tubes were centrifuged (400× g for 5 min at 4 °C) and the supernatants discarded. The pellet in each tube containing mature midgut cells and midgut stem cells was resuspended in 1 ml of working Grace’s media.

Stem cells were separated from mature cells using a density gradient (Loeb & Hakim, 1999). Briefly, samples were overlaid on 3 ml of Ficoll-Paque (GE LifeSciences, Piscataway, NJ, USA) in a 15 ml conical tube, and then centrifuged (600× g for 15 min at 4 °C). After centrifugation, stem cells were collected from the top 0.99 ml, the intermediate 2.75 ml containing tissue debris were discarded, and the mature cells were collected in the pellet. Ficoll-Paque was eliminated from stem and mature cell samples by washing twice with incubation media (600× g for 5 min at 4 °C). Final stem and mature cell pellets were suspended in 0.35 or 1 ml, respectively, of working Grace’s. Stem and mature cell samples that were prepared simultaneously were pooled and the number of cells counted using a hemocytometer (Bright-Line, Horsham, PA). Using this procedure, we reproducibly obtained approximately 8 × 10⁵ stem cells and 1 × 10⁷ mature cells from 30 H. virescens larvae. Cells were counted in a hemocytometer and diluted to 4 × 10⁵ cells/mL (stem cells) or 3.5 × 10⁶ cells/mL (mature cells) with working Grace’s, and kept in a sterile incubator at 26 °C.

Preparation of midgut cell secretomes

Purified mature midgut cells in working Grace’s were seeded (3.5 × 10⁶ cells) in individual wells of a 12-well culture-treated plate (Corning, Corning, PA, USA). The Cry1Ac toxin concentration (1 µg/ml) used as treatment to induce the midgut secretome was chosen based on its ability to induce low cytotoxicity (percent mortality 15.84 ± 0.29 compared to 8.09 ± 0.99 in buffer treatment) as measured by trypan blue staining of primary H. virescens mature cell cultures treated for 18 h at 26 °C. The purified Cry1Ac toxin (1 µg/ml) or the corresponding volume of control buffer (20 mM TRIS/HCl, 0.3 M NaCl pH 8.0), were added to the cultures and incubated for 18 h at 26 °C. After incubation, media supernatant containing the proteins secreted by the midgut cells was collected by centrifugation (2,000× g for 20 min at 4 °C). Two independent biological samples were used with each tested in triplicate. All resulting collected secretomes (six) were pooled to prepare each secretome category (buffer or Cry1Ac), and were used either for MS/MS identification or for bioactivity assays as described below. Proteomes were stored at −80 °C until used.

Secretome samples for 1DGel LC/MS/MS analysis were concentrated to 50–100 µl and the media exchanged to 20 mM TRIS/HCl pH 8.0 buffer using centrifugal filter devices (3-kDa MWCO; Millipore, MA), following manufacturer’s instructions. Proteins in concentrated samples were quantified using the Qubit fluorometer (Invitrogen, Carlsbad, CA, USA) and then diluted to 2 mg/ml in 20 mM Tris/HCl pH 8.0 buffer. Samples were shipped to NextGen Sciences (Ann Arbor, MI, USA) for proteomic analysis.

Stem cell bioactivity assays

Primary H. virescens midgut stem cell cultures were stained with the viability stain calcein AM (Invitrogen) following manufacturer’s instructions and then counted and gated using flow cytometry. Purified stem cells were seeded into individual wells of a 12-well plate at a concentration of 4 × 10⁴ cells/ml and a final volume of 500 µl. Treatments of Cry1Ac or buffer-induced secretomes were applied to the wells (final well volume was 1 ml). The time zero control was measured by adding working Grace’s media. After 30, 60 and 180 min of incubation time, proliferation and differentiation of cells were measured by staining for viability with calcein AM (Invitrogen, Carlsbad, CA, USA) and analyzed for side scatter and green (calcein AM) fluorescence by counting 20,000 non-gated events in an LSRII flow cytometer (BD Bioscience, San Jose, CA, USA), as previously described (Castagnola, Eda & Jurat-Fuentes, 2011). At least two wells for the same treatment were measured per experiment, and experiments were replicated thrice.

Proteomic analysis of primary midgut cell culture secretome

Proteins in secretome samples (10 µg) were separated by 1D SDS-10% PAGE using the NuPAGE Bis-Tris mini gel system (Invitrogen, Carlsbad, CA, USA) following manufacturer’s instructions, and then each of the sample lanes was sliced in five cross-sections that were subjected to in-gel digestion in a ProGest workstation (Genomic Solutions, Ann Arbor, MI, USA). Briefly, samples were reduced with 10 mM DTT at 60 °C, and then allowed to cool to room temperature before being alkylated with 100 mM iodoacetamide. Tryptic digestion was done at 37 °C for 4 h, and reactions were stopped by addition of formic acid (0.1% final concentration). Analysis of peptides generated by the tryptic digestion through liquid chromatography coupled to tandem mass spectrometry (LC/MS/MS) was performed at NextGen Sciences (Ann Arbor, MI, USA) using a ThermoFisher LTQ Orbitrap XL mass spectrometer. Tandem mass spectra were analyzed at MS Bioworks (Ann Arbor, MI, USA) using Mascot (Matrix Science, London, UK) and queried against a custom H. virescens transcriptome database (Perera et al., 2015) translated in the 6 possible frames. Search parameters included a fragment ion mass tolerance of 0.50 Da and a parent ion tolerance of 10.0 PPM. Iodoacetamide derivative of cysteine was specified as a fixed modification, while S-carbamoylmethylcysteine cyclization (N-terminus), deamidation of asparagine and glutamine, and oxidation of methionine and acetylation of the N-terminus were specified as variable modifications.

Scaffold (version Scaffold_4.8.2; Proteome Software Inc., Portland, OR, USA) was used to validate MS/MS based peptide and protein identifications. Peptide identifications were accepted if they could be established at greater than 95% probability as specified by the Peptide Prophet algorithm (Keller et al., 2002). Protein identifications were accepted if they could be established at greater than 99.0% probability and contained at least 2 identified peptides. Protein probabilities were assigned by the Protein Prophet algorithm. Proteins that contained similar peptides and could not be differentiated based on MS/MS analysis alone were grouped to satisfy the principles of parsimony. Weighed spectral counts calculated for each protein using an internal weighting function in Scaffold that apportions shared peptides among proteins according to peptide exclusivity, were used for quantitative comparisons. Given the low number of replicates used (multiple independent secretome replicates pooled into a single sample for MS analysis for each category), we used the Fisher’s exact test, as described elsewhere (Zhang et al., 2006), to compare the relative abundance between the Buffer and Cry1Ac sample categories (Zhang et al., 2006). Significance was established at p < 0.005 to increase the probability of accurate detection of differences in protein levels for each of the identified proteins.

Transcriptome profiling

Early fourth instar larvae were exposed to diet supplemented with a sublethal (no mortality after 8 h exposure) dose of Cry1Ac (1 µg/cm²) and observed for feeding. Midguts were dissected from actively feeding larvae after 0, 2 or 8 h and then flash frozen. Total RNA was extracted from midguts using RNAzol® RT (Molecular Research Center, Cincinnati, OH, USA) and enriched for mRNA using PolyATtract® magnetic beads (Promega, Madison, WI, USA). Approximately 500 ng of mRNA was fragmented using NEBNext® Magnesium RNA fragmentation reagents (New England Biolabs, Ipswich, MA, USA), cleaned and concentrated with MinElute Reaction Cleanup kit (Qiagen, Hilden, Germany). Double-stranded cDNA was prepared from fragmented mRNA with random hexamers using SuperScript® Double-Stranded cDNA Synthesis Kit (Invitrogen, Carlsbad, CA, USA). An Illumina TruSeq library preparation kit was used to prepare indexed libraries that were size fractionated to obtain insert lengths of 300 ± 30 bp using the LabChip XT DNA 750 Assay Kit in the LabChip XT instrument (PerkinElmer, Waltham, MA, USA). Size fractionated libraries were quantified by real time quantitative PCR using KAPA Library quantification reagents and standards (KAPA Biosystems, Boston, MA). Individual libraries multiplexed to contain equimolar concentrations were submitted to the USDA-ARS Genomics and Bioinformatics Research Unit (Stoneville, MS, USA) for sequencing (50 cycles) in an Illumina Hiseq200 platform. Sequence reads were pre-processed to remove adapters and to quality trim prior to expression profiling.

Contigs of interest based on the proteomic identification (Table 1) were selected from the transcriptome assembly of H. virescens (Perera et al., 2015) and used as reference sequences in expression profiling with CLC Genomics Workbench (CLC Bio-Qiagen, Aarhus, Denmark). Replicate groups were created using sequence reads from each replicate experiment. Read density for each transcript was normalized for transcript read length using the reads per kilobase of exon model per million mapped reads (RPKM) method (Mortazavi et al., 2008). The genes were then subjected to Baggerley’s test (Baggerly et al., 2003) to obtain weighted fold change of expression levels between the control and experimental replicates. The fold change values were filtered with the false discovery rate (FDR) correction of p-values at 0.0001 (one false discovery in 10,000 comparisons), and then normalized to an internal reference gene (Hv_Contig_26369, β-actin).

Purification of alpha-arylphorin

Arylphorin levels are highest at the pre-wandering phase of the 5th larval instar in noctuid larvae (Burmester, 2015). Consequently, hemolymph was collected from 60 pharate 5th instar H. virescens larvae by making a small incision at the base of the 1st and/or 2nd proleg and collecting droplets into 15 mL conical tubes containing 5 mg of phenylthiourea to block hemolymph phenoloxidase activity, and maintaining on ice. After collection, hemolymph was frozen at −20 °C until used (no longer than 2 months). Frozen hemolymph was thawed on ice and diluted 5-fold in 20 mM Tris pH 7.9 (buffer A). For fractionation, hemolymph was filtered (0.22 µm) and loaded onto a HiTrap Q HP column (GE Healthcare, Little Chalfont, UK), previously equilibrated with buffer A and connected to an AKTA FPLC system (GE Healthcare, Little Chalfont, UK). Proteins were eluted with a 0–1 M linear gradient of NaCl in 20 mM Tris pH 7.9 (buffer B) at a flow rate of 1 mL/min, collecting 1 mL fractions. To reduce the presence of smaller proteins co-purifying with α-arylphorin, fractions estimated to contain α-arylphorin (based on presence of ∼70 kDa band on electrophoretic observations) were combined and filtered using an Amicon Ultra-15 mL centrifugal unit (Millipore, Billerica, MA, USA) with a MWCO of 50 kDa. After concentration, partially purified α-arylphorin (Fig. S1) was quantified with the Coomassie Plus Protein Assay (Pierce, Waltham, MA, USA) using BSA as the standard, and then aliquoted and maintained at −80 °C until used.

Alpha-arylphorin feeding bioassays

Artificial diet was dispensed into wells of 128-well bioassay trays (both from Bio-Serv, Flemington, NJ, USA) and left to dry in a laminar flow cabinet. A single concentration of α-arylphorin observed to cause midgut hyperplasia in preliminary assays (0.781 µg/mL) was prepared in 20 mM Tris, pH 7.9 and distributed (75 µL) to the dry diet surface of each well and gently swirled to ensure even coating (29.2 ng/cm² final dose per well). Upon drying, a single neonate larva was placed in each well and the wells were sealed with adhesive covers. Larvae were allowed to feed on diet containing α-arylphorin for five days (3rd instar) under standard rearing conditions and then moved to diet that was surface-contaminated with Cry1Ac toxin (0.5 µg/cm²) or toxin buffer, and mortality scored after seven days. Bioassays were conducted with 16 neonate larvae per treatment and replicated three times.

Histological sections

The number of midgut cells in larvae fed for five days on diet supplemented with α-arylphorin or buffer (as described above) was determined by counting the number of nuclei incorporating DAPI (4′,6-diamidino-2-phenylindole) in gut sections. Larvae (4–6 per treatment) were fixed in ice-cold Carnoy’s solution (60% ethanol, 30% chloroform, 10% glacial acetic acid) and stored at 4 °C overnight in biopsy cassettes, and then transferred to freshly prepared 70% ethanol. Larvae were processed using a Tissue-Tek VIP processor (Sakura, Torrance, CA) and embedded in paraplast medium (Sigma–Aldrich, St. Louis, MO, USA). Block sections were obtained by cutting 5 µm slices using a Micron HM355s microtome (Thermo Scientific). For DAPI staining, tissues were mounted on Superfrost Plus Slides (Fisher Scientific, Waltman, MA, USA) following recommended guidelines (Slaoui & Fiette, 2011). Slides prepared for DAPI staining were deparaffinized through two 10 min washes in xylene followed by a rehydration series of ethanol washes (absolute, 95%, 70%) for 5 min each, and finally washed twice for 5 min in distilled water. Tissues were permeabilized by treatment with 10 mM citrate buffer pH 6.0 at 95−99 °C for 20 min and rinsed in two 1-min washes of phosphate buffered saline (PBS, 2 mM KCl,135 mM NaCl, 1.7 mM KH₂PO₄,10 mM Na₂HPO₄, pH 7.4). Blocking was performed in PBS containing 0.1% Tween-20 and 3% BSA for 1 h. After blocking, tissues were mounted in a water-based medium containing DAPI for nuclear detection. The total number of cells in each treatment was estimated by counting the number of DAPI-stained nuclei per 100 µm² in 3 sections from independent larvae for each treatment. Cell size in midguts submitted to experimental or control treatments was not determined.

Secretome bioactivity in primary midgut stem cell cultures

After collecting the proteins secreted by primary mature midgut cell cultures (secretome) in response to Cry1Ac or control buffer treatment, we determined their regenerative properties on primary H. virescens midgut stem cell cultures using a flow cytometry method based on differential calcein staining (Castagnola, Eda & Jurat-Fuentes, 2011). Flow cytometry gates defined by the side scattered light and calcein fluorescence allowed monitoring of changes in the number of stem, mature, and dead cell subpopulations within each sample (Fig. 1). Freshly prepared H. virescens midgut stem cell cultures (time zero in Fig. 1) contained predominantly stem cells (approx. 3.5 × 10⁴ cells or 85%), and a smaller proportion of mature (0.6 × 10⁴ cells or 14%) and dead (0.6 × 10³ or 1%) cells. Treatment with buffer-induced secretome resulted in a non-significant trend (Student’s T-test, p = 0.074) of rapid (30 min.) reduction in the number of stem cells concomitant with a significant increase in the number of dead (Student’s T-test, p < 0.05) and mature cells (Student’s T-test, p < 0.05) (Fig. 1, open symbols). Technical limitations prevented us from determining whether the cell types accounting for the increased dead cell numbers in this treatment were stem, mature or both, but this observation may reflect the low viability in culture of midgut mature cells in the original preparation and additional mature cells originating from stem cell differentiation. After three hours, only a small reduction in the number of dead cells (probably due to cytolysis) was observed in these cultures.

Similarly to treatment with buffer-induced secretome, a significant (Student’s T-test, p < 0.05) initial tendency to increase the number of dead and mature cells and decrease stem cell numbers was detected in cultures treated with Cry1Ac-induced secretome (Fig. 1, black symbols). However, an important observation was that the total number of dead cells after 30 min or 1 h was approximately half in cultures treated with Cry1Ac compared to buffer secretome treatment. After 3 h, the number of stem cells in the Cry1Ac secretome-treated samples had recovered to similar levels observed for the initial conditions, while the number of dead cells had significantly decreased compared to 30 min (Student’s t test, p < 0.05) and the number of mature cells remained constant. The recovery in the number of stem cells between 30 min and 3 h (Student’s T-test, p < 0.05) was suggestive of a regenerative response to Cry1Ac secretome compared to control treatment. This response was not observed when primary midgut stem cell cultures were treated with the same Cry1Ac toxin concentration used to obtain the Cry1Ac-induced secretome (see line estimates in Fig. 1), indicating the presence of mitogens in the Cry1Ac-induced secretome.

Identification of secretome proteins and alterations in response to Cry1Ac intoxication

Proteins in buffer and Cry1Ac-induced secretomes from H. virescens primary mature midgut cell cultures were identified using LC/MS/MS and a custom H. virescens transcriptome database (Perera et al., 2015). A total of 358 proteins were identified in the secretomes (protein False Discovery Ratio [FDR] 0.0%), with 326 and 313 proteins detected in the buffer and Cry1Ac-induced secretomes, respectively. Out of the 358 identified proteins, 281 proteins were common to both secretomes, while 45 proteins were unique to the buffer, and 32 to the Cry1Ac induced secretomes, respectively. Normalized spectral counts for all proteins detected in the Cry1Ac secretome and their corresponding values in buffer secretome are listed in Table S1. The list of identified proteins included proteins involved in physiological functions expected for the gut tissue, such as digestive enzymes, storage and transport proteins, immune and stress-related proteins, and also putative Cry toxin receptors. Intracellular proteins (ribosomal, mitochondrial and cytosolic enzymes, nucleotide-related proteins) were also commonly detected, possibly due to leakage of cell contents after cell death.

Statistical analyses of weighed spectral counts identified 4 proteins with significant different abundance (Fisher’s Exact test with Benjamini–Hochberg correction; p < 0.005) when comparing buffer to the Cry1Ac-induced secretome (Table 1). Differentially present proteins included digestive enzymes (aminopeptidase and glucosidase), a protein involved in processes of cell growth and proliferation (elongation factor 1 gamma), and a storage protein (arylphorin subunit alpha). Both the Cry1A receptor aminopeptidase and the elongation factor 1 gamma were unique to the buffer secretome, while only α-arylphorin was unique to the Cry1Ac secretome. An uncharacterized glucosidase was more abundant in the Cry1Ac secretome (Table 1).

Transcriptome profiling in larvae exposed to Cry1Ac for 2 and 8 h was used to confirm the levels of expression of the contigs matching to the proteins identified as differentially present in the buffer- and Cry1Ac-induced secretomes. We chose these time points as an early (2 h) and late (8 h) times in the Cry1Ac intoxication process based on previous reports (Forcada et al., 1996; Spies & Spence, 1985). When fold difference in expression levels after 2 h of exposure to Cry1Ac was estimated using the RKPM method (Mortazavi et al., 2008), we observed reduced expression of contigs 997 (aminopeptidase), 6027 (elongation factor), and 2924 (glucosidase), while expression of α-arylphorin was increased more than 9-fold (Fig. 2), in agreement with the proteomic findings. In contrast, after 8 h of exposure to Cry1Ac, only contig 997 had slightly increased expression, while all other three contigs had reduced expression (Fig. 2). When expression of arylphorin contigs detected in the H. virescens transcriptome was examined, we observed that a group of related arylphorin contigs was up-regulated in response to Cry1Ac, while a second group did not respond or had reduced expression levels (Fig. S2). Contig 2879 (α-arylphorin) detected in the secretome analysis and that was also unique to the Cry1Ac secretome (not included in Table 1 since p < 0.016), also displayed increased expression after exposure of larvae to Cry1Ac (Fig. S2).

Interestingly, when comparing levels of expression of arylphorin contigs in the Hv transcriptome (Perera et al., 2015) after exposure to Cry1Ac (Fig. S2), we found differences in expression among arylphorin contigs that were distantly related. For instance, in contrast to contigs matching to arylphorin subunit alpha being up-regulated, contig 19307 matching to arylphorin subunit beta of H. armigera (accession XP_021195808.1; E-value 2.00E−90) was down-regulated after exposure to Cry1Ac. Contig 35193 matching with 74% sequence identity to arylphorin from H. armigera (accession AEO51737.1; E-value 5.00E−40) did not display drastic changes in expression after larvae were exposed to Cry1Ac.

Effect of α-arylphorin on H. virescens susceptibility to Cry1Ac toxin

Given that α-arylphorin was the only protein unique to the Cry1Ac-induced secretome among the differentially present proteins between the buffer and Cry1Ac-induced secretomes, we concentrated our efforts on the putative role of this protein in the midgut response to Cry1Ac intoxication. To test the effect of α-arylphorin during the response to Cry1Ac intoxication, we purified arylphorin from H. virescens hemolymph and fed it to neonates for five days before exposing them to Cry1Ac toxin. Mass spectrometry and electrophoretic analyses of the purified α-arylphorin sample supported that it was >87% pure (Fig. S1). Larvae feeding on this purified α-arylphorin sample for 5 days developed midgut hyperplasia when compared to controls (Fig. 3A). This hyperplasia resulted from a significant increase in the number of cells per unit of midgut surface (Fig. 3B). Exposure of larvae fed buffer without α-arylphorin to Cry1Ac toxin (0.5 µg/cm²) for seven days resulted in 34 ± 5% mortality. In contrast, a significant reduction (Student’s t-test; p < 0.05) in mortality to 15 ± 3% was detected in larvae that had been pre-exposed to α-arylphorin (29.2 ng/cm²) before exposure to Cry1Ac (Fig. 3C). No significant difference was detected between mortality in control larvae (3% mortality) and larvae exposed to α-arylphorin for five days (Fig. 3D) (Mann–Whitney Rank Sum Test used due to data failing normality test; p < 0.05).