Introduction

Yersinia enterocolitica is a gram-negative, facultative anaerobic, non-sporulating, coccobacillary gamma-proteobacterium belonging to the family of Enterobacteriaceae (Rakin et al. 2015). It is a potential human pathogen with vast serotype diversity. It has been differentiated into 6 biovars viz., 1A, 1B, and 2–5 based on biochemical tests (Virdi et al. 2012; Singh et al. 2014), and greater than 70 serotypes depending on the antigenic variations in the O-antigen of the lipopolysaccharides (Rakin et al. 2015). Y. enterocolitica biovar 1B, a highly virulent biovar, is a food-borne and water-borne pathogen, but blood-transfusion associated septicemia and mortality is also reported. The enteropathogenic Y. enterocolitica invades the intestinal epithelium and multiplies in the lymphatic system (Ahlawat et al. 2020). The highly virulent strains of Y. enterocolitica subsp. enterocolitica posses a virulence plasmid (pYV), heat stable toxin (YstA), high-pathogenicity islands (HPI), chromosomally-encoded type II (T2SS) and type III (T3SS) secretion systems, resistance-gene loci, various metal-uptake operons (Thomson et al. 2006), genomic islands, prophages, putative hemolysins, adhesions, and low redox laccases (Singh et al. 2014, 2016; Ahlawat et al. 2020). Yersinia secretion locus of pYV plasmid encodes a T3SS along with a set of virulence factors, i.e., Yersinia-outer membrane proteins (Yops). The virulence factor T3SS of Yersinia inhibits its phagocytosis in host, while HPI has a role in siderophore yersiniabactin-mediated iron uptake (Rakin et al. 2015).

Conventionally, mouse infection models are used to study the pathology of Y. enterocolitica and other human pathogens. However, due to ethical concerns, limited number of animals can be used. Further, the experimental animals are costly and the maintenance of animal facilities is expensive. Therefore, there is an urgent need for a simple, rapid, easy to handle, inexpensive, and ethically acceptable in vivo model for studying the virulence of enteropathogens. Interestingly, Y. enterocolitica, which is considered as a mammalian pathogen, is also toxic to insects. Yersinia is toxic to Manduca sexta or tobacco hornworm. It also colonizes Caenorhabditis elegans by expanding its intestinal lumen and thus killing the nematode (Ahlawat et al. 2020). Further, Galleria mellonella or wax-moth insect larva is a commonly used alternative infection model for enteropathogens (Alenizi et al. 2016) including C. jejuni (Senior et al. 2011) and Klebsiella pneumoniae (Insua et al. 2013). It is also used to assess the host–pathogen interactions (Hurst et al. 2015). The insect gut provides an alternate and analogous subject to the vertebrate gut, such as the alkaline pH of midgut, apical end of columnar cells in midgut is covered by short hair-like projections called microvilli, and digestive tract of insects is lined with an invertebrate-unique structure called peritrophic membrane (PM) (Zhang and Guo 2011), i.e., widely analogous to the mucous lining of the vertebrate gut (Campbell et al. 2008).

In an earlier work, a label-free quantitative proteomic technique was performed to gain insights into Beauveria bassiana-Bombyx mori interaction. According to this report, the screening of immune molecules against B. bassiana at the proteome level is more accurate and rapid than the screening at the transcriptome level (Lü et al. 2019). Another study based on a label-free proteomic method offers evidence for host defense against B. mori nuclear polyhedrosis virus (BmNPV) infection, thereby suggesting the potential function of insect midgut after oral infection (Zhang et al. 2019). Their short life cycle and no ethical constraints make them an ideal model system for evaluating the virulence of pathogens. Therefore, the current study was performed to develop the larvae of H. armigera (Hübner) (Lepidoptera: Noctuidae) as a rapid-inexpensive in vivo model system to evaluate the effect of an enteropathogen Y. enterocolitica strain 8081 (1B/O:8) on its midgut. Further, the entomopathogenic potential of Y. enterocolitica strain 8081 in the mid gut of H. armigera larvae was evaluated via a label-free proteomic approach, and validated through spectroscopic, biochemical, histopathology and microscopic studies.

Materials and methods

Materials

Guaiacol, isopropyl-β-D-1-thiogalactopyranoside (IPTG), and kanamycin were purchased from Sigma-Aldrich products, Merck KGaA, Darmstadt, Germany Inc., while all other media components were purchased from Hi-Media, Mumbai, India and were of reagent grade or better.

Bacterial strain preparation and H. armigera larval infection

The reference strain, i.e., Y. enterocolitica strain 8081 (1B/O:8) used in the present study is an American Type Culture Collection (ATCC) strain that was generously provided by Prof. J.S. Virdi, University of Delhi, South Campus, New Delhi, India. Y. enterocolitica strain 8081 (1B/O:8) was grown in LB medium at 37 °C, 200 rpm until stationary growth phase (OD600 is ~ 1–5 × 109 colony forming unit (CFU)/mL). The cell pellet obtained by centrifuging cells at 5000 rpm, 10 min, and 4 °C was washed and suspended in normal saline. First-instar H. armigera neonate larvae were obtained from National Bureau of Agricultural Insect Resources (NBAIR) [Bangalore, India] and only healthy-looking larvae were used in the experiment. Two days old (first-instar) larvae of H. armigera were challenged orally with Y. enterocolitica strain 8081 [undiluted (10 µL/mL), 10−1 dilution, and 10−2 dilution] along with the artificial diet, till fourth-instar. The control insect larvae were fed with the artificial diet. The experiments were performed in several replicates with body weight, morbidity, and mortality in insect monitored and recorded. Thereafter, ten-day old (fourth-instar) larvae of both control and Yersinia added diet-fed (YADF) insects were dissected aseptically and the gut was excised for biochemical, histopathology, scanning electron microscopy (SEM), and label-free LC–MS/MS proteomic studies. The insect bioassays were performed at 27 ± 1 °C, RH 65 ± 5%, and 16 h light/8 h dark cycle for eight days [from two days old (first-instar) larvae to ten days old (fourth-instar) larvae]. Proteomic analysis was carried out on ten-day old (fourth-instar) larvae on the basis of morbidity data.

Statistical analysis of data

The data of insect weight was calculated as mean values [± standard deviation]. The statistical analysis was performed by one-way Analysis of Variance (ANOVA) using Microcal Origin 6.0 (Microcal Origin Software, Northampton, Massachusetts). The mean values obtained from each sample were compared by using an unpaired t-test. A probability value < 0.05 was regarded as statistically significant.

Proteomic analysis of YADF insect gut

Protein preparation, digestion and nano-LC–MS/MS

The proteomic analysis of the excised gut of the YADF insects was performed as discussed in our previous reports (Ahlawat et al. 2020). The excised gut of the ten days old (fourth-instar) larvae were briefly sonicated in extraction buffer. After the breakdown of tissue the protein was precipitated using 8:1 (Acetone:TCA). The pellets obtained were washed by using dithiothreitol-acetone. Thereafter, the protein was solubilized and quantified using nanodrop, reduced, alkylated, and diluted by using 25 mM ABC (pH ~ 7.0). The digestion was done by adding trypsin in 1:50 ratio and trypsinization was stopped by adding FA. Further, the samples were dried and reconstituted in FA. The digested peptides were analyzed using a nano LC–MS/MS. The peptides were loaded onto a 2 µm × 50 cm PepMap RSLC C18 pre-column and then eluted at 300 nL/min flow-rate by using a gradient of solution A and B for 123 min. The full-scan MS spectra (m/z 350–2000) were obtained on a Q-Exactive Orbitrap in a positive-ion mode (Jain et al. 2019).

Data analysis and bioinformatics studies

Raw mass spectrometry data files acquired from the mass spectrometer were handled using proteome discoverer software (ver. 2.2, Thermo Fisher Scientific, USA). The acquired MS files were searched against the database of protein sequence related to Helicoverpa and Yersinia available in UniProt (http://www.UniProt.org/). SEQUEST-HT results were filtered with the Percolator-based scoring to enhance the accuracy and sensitivity of the peptide identification. The proteins and peptides were deduced from spectrum identification results by using peptide spectrum matches (PSMs). Proteins and peptides were validated at a false discovery rate (FDR) of 0.1% assessed by using decoy-hit distribution. The proteins were selected by q-value (≤ 0.05) by using in-built statistical packages of proteome discoverer (ver. 2.2, Thermo Fisher Scientific, USA). The results were filtered by Xcorr (< 1.5), PSMs (≥ 5) for proteins and ΔCn (< 0.01) for peptides with a high confidence (Jain et al. 2019).

Protein quantification was done using the total MS/MS spectrum count of the identified proteins. Additional criteria were employed to enhance the confidence. Normalization of identified PSMs among LC–MS/MS runs was done by dividing individual PSMs of proteins by total PSMs and average of % PSM count was used for evaluating the fold-changes for treatment conditions, as described earlier (Ahlawat et al. 2020). For contrasting relative intensities of proteins between control and YADF insects groups, samples were assessed by using cumulative-confident-normalized PSMs value. Functional assignment and annotation by gene ontology (GO) terms (www.geneontology.org/), InterPro terms (InterProScan, EBI: https://www.ebi.ac.uk/), enzyme classification (EC) codes, and metabolic pathways (Kyoto Encyclopaedia of Gene and Genomes, KEGG) were determined by OmicsBox (https://www.biobam.com/omicsbox/) and Blast2GO (https://www.blast2go.com/). The heat map cluster was made using Multi Experiment Viewer (MeV) software suite (ver. 4.9; http://mev.tm4.org/) and protein–protein interaction networks (PPIs) were analyzed via Search Tool for the Retrieval of Interacting Genes/Proteins (STRING) software (ver. 11.0; https://string-db.org/). The mass spectrometry proteomics data have been deposited to the ProteomeXchange consortium via the proteomics identification database (PRIDE; https://www.ebi.ac.uk/pride/) (Perez-Riverol et al. 2019) partner repository with the dataset identifier PXD021304.

Effect of Y. enterocolitica infection on H. armigera larvae gut

Biochemical and spectroscopic studies

To study biochemical changes the gut homogenate was prepared from the excised ten days old (fourth-instar) larval gut. Catalase (CAT), an antioxidant enzyme, catalyzes the breakdown of hydrogen peroxide (H2O2) to less toxic products (Rajendran et al. 2014). Catalase assay was done according to the procedure described earlier (Cuéllar-Cruz et al. 2009). It measures the decomposition of H2O2 by enzyme catalase. One unit of catalase decomposes 1 µM of H2O2 per minute at 25 °C and pH 7 under the specified conditions. The amount of H2O2 decomposed was calculated and the results were expressed as nmol of H2O2 decomposed/min/mg protein. Lipid peroxidation (LPO), defined as deterioration of membrane lipids by free radicals that result in the formation of stable oxidation products (Kurutas 2016), was assayed via the modified method of Wills (1966). The stable oxidation products of LPO are commonly used as the biomarkers of oxidative/nitrosative stress or damage and the results were expressed as nmol malondialdehyde (MDA)/mg protein. Glutathione (GSH) was estimated as described by Kumar et al. (2008). The results were expressed as µg glutathione/mg protein. GSH is an intracellular antioxidant that can accommodate the loss of a single electron due to the presence of sulfur containing thiol group (Kurutas 2016). Thereafter, Lowry’s method was used to determine the protein concentrations, using bovine serum albumin (BSA) as a standard. Further, the ten days old (fourth-instar) larval gut samples were freeze-dried by using the lyophilizer (Allied Frost, India) and the change in structure was examined by performing the fourier-transform infra-red (FTIR) spectroscopy on a FTIR-5300 spectrophotometer (Alpha FTIR, Bruker, Germany) to record data in 500–4000 cm−1 scanning range.

Imaging studies of YADF insect gut

Aseptically excised ten days old (fourth-instar) larval gut samples were stored in 2% paraformaldehyde prior to histopathological analysis. Sections of larval gut tissue (10–15 µm) were sliced using CM3600 Cryomicrotome (Leica Microsystem Ltd., Germany) followed by fixation on a gelatin-coated slide. After fixation, the tissue samples were stained with hematoxylin and eosin (H-E). The gelatin-coated slides were dipped in hematoxylin for 1–5 min and washed with water for 5 min, followed by eosin staining for 25 s and washing. Further, the slides were dipped subsequently in 75%, 95%, and 100% ethanol for 1 min each, followed by xylene treatment for 3 min. Finally, the slides were DPX (dibutylphthalate polystyrene xylene) mounted and observed under the light microscope.

Scanning electron microscopy (SEM) was done commercially at Indian Institute of Technology (IIT), New Delhi, India. For SEM, the gut tissues of fourth-instar larvae were fixed in 2% paraformaldehyde and vacuum-dried. Dried tissues were mounted, coated with palladium, and examined using FEI Quanta 200F integrated with Oxford-EDS system (IE 250 X Max 80), Netherlands at 7 kV.

Results

Proteomic analysis of YADF insect gut

For performing label-free LC–MS/MS proteomic studies, two days old (first-instar) larvae of H. armigera were orally challenged with Y. enterocolitica strain 8081 in the artificial diet till fourth-instar. The insect bioassays were performed at 27 ± 1 °C, RH 65 ± 5%, and 16 h light/8 h dark cycle for eight days [from two days old (first-instar) larvae to ten days old (fourth-instar) larvae]. Thereafter, the ten days old (fourth-instar) larvae were dissected aseptically to obtain the excised gut tissues, which were pooled to perform the proteomic study. The experiments were performed using 22 control and 33 YADF insects larvae (11 each for 10−1, 10−2, and 10−3 dilution). Using label-free proteomic technique, i.e., Orbitrap nano-LC–MS/MS, a total of 1133 (306 exclusively in control insect larvae, 95 exclusively in YADF insects, and 732 shared between both the groups) unique protein accessions having PSM ≥ 1 were identified and quantified with a high confidence in this study (Fig. 1A).

Fig. 1
figure 1

A Distribution of 1133 identified protein accessions with PSMs ≥ 1, B Distribution of 1133 identified proteins accessions with PSMs ≥ 1 based on the match against protein sequences of Helicoverpa and Yersinia available in UniProt, C Distribution of 467 identified protein accessions with PSMs ≥ 5, D Heat map depicting changes in protein expression in H. armigera larvae with and without Yersinia-feeding using protein accessions common to both conditions, E Heat map depicting changes in protein expression in H. armigera larvae with and without Yersinia-feeding using protein accessions present in either group, F Distribution of 467 identified protein accessions with PSMs ≥ 5 based on expression levels

Full size image

Among all the identified proteins, about 2% were of Yersinia (Fig. 1B). The presence of Yersinia proteins in the insect gut tissues might be due to the establishment of Y. enterocolitica strain 8081 into the epithelial cells of the larvae gut, as the excised gut of H. armigera larvae was flushed with phosphate buffered saline (PBS) before performing the preliminary steps of proteomics. The Yersinia-identified proteins included metabolism proteins (13), transport and secretion proteins (2), putative hemolysins (1), secretory systems (2; T3SS:1, T6SS:1), and proteins for sensing, signaling, and regulation (4) (Supplementary Table 1).

GO annotation and KEGG pathway analysis of stringently-filtered insect proteins

On strict filtering of the dataset by using the criterion of PSMs ≥ 5, 467 proteins were stringently identified. Out of 467 proteins, 284 were shared between both groups and 183 (160 in control, 23 in YADF insects) were exclusively determined in a single group (Fig. 1C). The differentially expressed proteins are shown in a heat map format by using normalized PSMs (Fig. 1D, E). The heat map data clearly indicate the diverse intensities (67% up-regulated, 31% down-regulated) of proteins between control and YADF insects (Fig. 1F). Further, identified proteins (467 proteins; PSMs ≥ 5) were classified based on biological process (BP), cellular component (CC), and molecular function (MF) categories in gene ontology (GO) annotations. GO analysis of control insects unique proteins displayed 37.51% BP, 10.77% CC, and 51.68% MF proteins. Comparatively, GO analysis of YADF insects unique proteins had 36.18% BP, 13.7% CC, and 50.1% MF proteins (Fig. 2, Supplementary Table 2). Further, when results of both the groups were compared, extracellular region (CC, GO: 0005576) and transferase activity (MF, GO: 0016740) classes or functions were reported to be unique in control insects. Biological regulation (BP, GO: 0065007), cytoskeleton (CC, GO: 0005856), and extracellular space (CC, GO: 0005615) classes or functions were unique in YADF insects (Fig. 2, Supplementary Table 3). A biological system employs a variety of strategies and mechanisms to ensure their survival under varying conditions. In some conditions, they balance perturbations by supporting the conditions under which their constitutive processes persists. Altogether these mechanisms are linked to the concepts of ‘homeostasis’ and ‘adaptation’, therefore, contribute towards system’s viability against perturbations by modifying its own dynamic behaviour. Conclusively, biological regulation enables an organism to control the outcomes of perturbation by regulating its own constitutive dynamics following remarkable alterations in the internal and external conditions (Bich et al. 2016). Thus, the exclusive presence of biological regulation in YADF insects confirms the encounter of these insects with a perturbation (i.e., an enteropathogen in this case).

Fig. 2
figure 2

Gene ontology (GO) analysis of differentially expressed protein accessions with PSMs ≥ 5 from normal diet-fed insects (on left) and YADF insects (on right)

Full size image

According to the pathway analysis, 78 assigned EC numbers were identified for the provided unique protein accessions (Supplementary Fig. 1). KEGG pathways (56) were represented by 467 unique protein accessions, including purine metabolism (KO00230: 11 and 4 identified enzymes in control and YADF insects gut, respectively), citrate (TCA cycle) cycle (KO00020: 10 and 6 identified enzymes in control and YADF insects gut, respectively), pyruvate metabolism (KO00240: 9 and 8 identified enzymes in control and YADF insects gut, respectively), glycolysis/gluconeogenesis (KO00010: 9 and 7 identified enzymes in control and YADF insects gut, respectively), starch and sucrose metabolism (KO00500: 8 and 4 identified enzymes in control and YADF insects gut, respectively), and carbon fixation pathways in prokaryotes (KO00720: 7 and 7 identified enzymes in control and YADF insects gut, respectively).

The insect proteins can be differentiated under six categories: proteins involved in (i) melanin synthesis; (ii) stress response and cell redox homeostasis; (iii) cytoskeleton; (iv) iron uptake and storage; (v) anti-pathogenic response; and (vi) metabolism. The proteome comparison between control and YADF insects revealed significantly up-regulated prophenoloxidase 1 (UniProt ID-A0A290U614), prophenoloxidase subunit 2 (UniProt ID-Q2VIY6), serine protease (UniProt ID-O18447), serine protease 5 (UniProt ID-O18436), proteins with serine-type endopeptidase activity (UniProt ID-A0A2W1BP39, UniProt ID-B1NLE3, UniProt ID-O18443, UniProt ID-A0A2W1BGN3, UniProt ID-O18450, UniProt ID-A0A2W1BWI0, and UniProt ID-A0A2W1BVE6), and proteins of serpin family (UniProt ID-A0A2W1BEQ7, UniProt ID-F5B4G8, UniProt ID-A0A2W1BH13, and UniProt ID-A0A290U612); while, significantly down-regulated serine protease inhibitor 6 (UniProt ID-A0A1L5J028) (Table 1).

Table 1 List of up-regulated and down-regulated proteins in YADF insects
Full size table

The proteins related to heat shock response (UniProt ID-C7SIR9, UniProt ID-A0A2H4LI91, UniProt ID-A0A2H4LHM4, UniProt ID-F5BYH9, UniProt ID-A0A2W1BDN4) were significantly up-regulated; whereas, the proteins responsible for cell redox homeostasis (UniProt ID-A0A2W1BQL1, UniProt ID-A0A2W1BSJ9, UniProt ID-A0A2W1BR63, UniProt ID-A0A0A7RB97, UniProt ID-A0A2W1C0K5) and protein folding (UniProt ID-A0A2W1BTU8, UniProt ID-A0A2W1BH57) were significantly down-regulated in YADF insects. The proteins having protective role against oxidative stress, such as catalase (UniProt ID-H9BEW3), glutathione-S-transferase (UniProt ID-A0A0D3M5U4, UniProt ID-A0A2W1BLG9, UniProt ID-A0A291ARU4, UniProt ID-A0A0D3M5T8, UniProt ID-A0A2W1BKA7, and UniProt ID-A0A0D3M5V9), thioredoxin peroxidase (UniProt ID-B2KSE9), and thioredoxin domain containing protein (UniProt ID-A0A2W1C322) were significantly up-regulated in response to the bacterial feeding. Moreover, ferritin (UniProt ID-A0A2W1BW25 and UniProt ID-A0A2W1C098), transferin (UniProt ID-A0A067YAW9), and lipocalin family proteins (UniProt ID-A0A0D3QSH9 and UniProt ID-A0A2W1C146) were significantly up-regulated in YADF insects. The cytoskeleton proteins, including actin (UniProt ID-E2IV58, UniProt ID-A0A221LCD5, UniProt ID-E2IV62) and chitin deacetylase 1 (UniProt ID-D2WPC4) were also found to be up-regulated in the gut of YADF insects. Additionally, proteins with role in cell adhesion (UniProt ID-A0A2W1BP23) and cytoskeleton organization (UniProt ID-A0A2W1BJX1) were significantly down-regulated; while, the protein responsible for actin filament depolymerization (UniProt ID-A0A2W1BIB1) was significantly up-regulated (Table 1). Thus, confirming the affected cytoskeleton of the insect gut in response to Yersinia feeding.

Moreover, digestive enzymes such as carboxyester hydrolase of type-B carboxylesterase/lipase family (14) and of lipase family (2), aminopeptidase of peptidase M1 family (6), α-amylase of glycosyl hydrolase family (5), and alkaline phosphatase of alkaline phosphatase family (3) were observed to be significantly up-regulated in YADF insects. Furthermore, metabolic proteins (30), energy metabolism proteins (23), translation-related proteins (23), lipid binding (7), and 5 proteins with lipid transporter activity were found to be differentially expressed in larvae gut after bacterial infection. Proteins including succinate dehydrogenase [ubiquinone] flavoprotein subunit, mitochondrial (UniProt ID-A0A2W1BZK9), succinate dehydrogenase [ubiquinone] iron-sulfur subunit, mitochondrial (UniProt ID-A0A2W1BX81), complex1_30kDa domain-containing protein with NADH dehydrogenase (ubiquinone) activity (UniProt ID-A0A2W1BUD9), uncharacterized protein with NADH dehydrogenase (ubiquinone) activity (UniProt ID-A0A2W1BXT8), and uncharacterized protein with cytochrome-c oxidase activity (UniProt ID-A0A2W1BJ66) were significantly up-regulated in gut of YADF insects (Table 1).

To establish known and predicted protein–protein interactions (PPIs) between the cytoskeleton proteins (15), cell-redox proteins (14), and stress-related heat shock proteins or HSPs (6) from H. armigera, an interactome was constructed via STRING11.0 database using B. mori as a reference organism. The constructed interactome had 30 nodes and 53 edges with average node degree of 3.53, average local-clustering coefficient of 0.558, and PPI-enrichment p-value of < 1.0 e−16. Moreover, the interactome consisted of three major clusters: (1) between cell-redox proteins (GSTd3, GSTe3, GSTo1, GSTo2, GSTs2, and LOC733095), (2) between cytoskeleton proteins (Kettin, A1, 100101180, BGIBMGA000613-TA, BGIBMGA003515-TA, BGIBMGA014226-TA, and BGIBMGA001730-TA), and 3) between stress-related HSPs and cell-redox proteins. The cell redox proteins GSTs2 and BGIBMGA002818-TA were found to connect the clusters (1) and (3). Likewise, the cytoskeleton proteins A1 and 100101180 connected the clusters (2) and (3) by interacting with stress-related proteins HSP70 and 14-3-3zeta, respectively. Further, the cytoskeleton protein BGIBMGA000613-TA of cluster (2) was found to interact with the cell redox protein BGIBMGA008023-TA of cluster (3) (Fig. 3). The protein interactions between up-regulated 14-3-3zeta with up-regulated 100101180 (actin filament depolymerization) and down-regulated BGIBMGA008023-TA with down-regulated BGIBMGA000613-TA (motor activity) further strengthened the hypothesis of stress-induced damage of Helicoverpa gut epithelium in response to enteropathogen infection (Fig. 3; Table 2).

Fig. 3
figure 3

Protein interaction network of differentially expressed (of interest) proteins generated with STRING (ver. 11.0). Nodes represent proteins and edges represent protein–protein interactions (PPIs). Empty nodes represent proteins of unknown 3-D structure and filled nodes represent known or predicted 3-D structure. The ‘cyto’, ‘stress’ and ‘redox’ represents cytoskeleton proteins, stress-related HSPs, and cell-redox proteins, respectively

Full size image
Table 2 List of proteins used for protein–protein interactions (PPIs) study
Full size table

Effect of Y. enterocolitica infection on H. armigera larvae gut

The larvae of two days old (first-instar) neonates of H. armigera were orally challenged with virulent strain, i.e., Y. enterocolitica strain 8081 (1B/O:8) in the artificial diet, till fourth-instar. The control insects were fed with the artificial diet. The larvae survival was observed over time after feeding of the bacterial suspension such that no mortality was reported in either group. However, the bacterial infection noticeably affected the feeding, development, and behavior of the infected larvae. Post infection, the larvae become morbid and the feeding was also reduced. The initial average body weight of the larva was 0.22 ± 0.014. The reported body weight of control and YADF insects (fourth-instar) were 375.63 ± 32.59 mg and 253.87 ± 66.78 mg, respectively (Fig. 4A, Supplementary Table 4). The data was statistically analyzed which showed significant reduction in the average body weight of ten days old (fourth-instar) YADF insects (Fig. 4A), where A, B, and C represents 10−1, 10−2, and 10−3 dilution, respectively.

Fig. 4
figure 4

A Reduction in body weight of H. armigera larvae after treatment with different concentration of Y. enterocolitica strain 8081. 8081 (A, B, and C) represents 10–1, 10–2, and 10–3 dilution, respectively and * represents differences between control (artificial diet-fed) insects and YADF insects (unpaired student’s t-test p < 0.05) B Biological FTIR-spectrum showing biomolecular peak assignments for control (in blue) and YADF insects [fourth-instar (in black)] gut samples

Full size image

Biochemical and spectroscopic studies

The gut of ten days old (fourth-instar) larvae were dissected aseptically and homogenized for further analysis. The GSH level was found to slightly decreased (0.41 fold); whereas, the CAT activity increased by 2.52 fold in YADF insects gut. Higher (2.83 fold) MDA level in YADF insects gut showed enhanced LPO in comparison to the control larval gut (Table 3).

Table 3 Catalase (CAT), gluthathione (GSH) and malondialdehyde (MDA) levels in the gut homogenates of control and YADF insects
Full size table

Fourier-transform infra-red (FTIR) spectroscopy provides a fast and label-free tool to study the molecular composition of a sample without perturbing it. It depends on quantitatively measuring the IR active-vibrational modes of the molecular bonds with an electric dipole moment, which change by the atomic displacement. For biological samples, the significant spectral regions quantified are the finger-print region (450 cm−1, 600 cm−1) and amide I/II region (700 cm−1, 1500 cm−1). The higher wavenumber region, i.e., 2550–3500 cm−1 corresponds to stretching vibrations like C–H, S–H, O–H and N–H; whereas, the lower-wavenumber region is associated with bending and C-skeleton finger-print vibrations. Together, these regions provide a finger-print of structure of the specimens (Baker et al. 2014). The FTIR spectroscopic analysis showed a variation with the absence of peaks (in YADF insect) at 3258 cm−1, 2838 cm−1, and 1745 cm−1, indicating N–H stretching, C–H stretching, and C=O stretching, respectively. A shift of peak from 1376 cm−1 (in control) to 1397 cm−1 (in YADF insect) was due to C-H bending (Fig. 4B, Supplementary Table 5). Additional peaks at 998 cm−1 for C–H bending, and 1078 cm−1 for C–C multiple bond stretching was observed in control larval gut. Various additional peaks from 500 to 900 cm−1 were observed in gut of control larvae. Altogether, these variations suggested the change in biological macromolecules including proteins, nucleic acids, lipids, and carbohydrates.

Imaging studies of YADF insects gut

The scanning electron micrographs of the surface of control and YADF insects midgut epithelium appeared with densely packed microvilli. However, cell loosening and widening was observed in YADF insects midgut epithelium (Fig. 5A, B). This is in concurrence with down-regulated proteins with role in cell adhesion and cytoskeleton organization. Histopathological observations do not show any significant difference between control and YADF insects midgut epithelium besides slight erosion or cell loosening (Fig. 5C, D).

Fig. 5
figure 5

Scanning electron microscopy (SEM, X5000) of midgut of H. armigera larvae: A control B treated with Y. enterocolitica strain 8081.Hematoxylin–eosin (H-E) staining (X40) of midgut of H. armigera larvae: C control D treated with Y. enterocolitica strain 8081

Full size image

Discussion and conclusion

Upon oral challenge with Y. enterocolitica strain 8081, no mortality but significant reduction in average body weight was observed in H. armigera larvae. This was supported by an earlier report in G. mellonella-Y. enterocolitica strain 8081 infection model (Alenizi et al. 2016). As reported earlier, 57 colony forming unit (CFU) of Y. enterocolitica strain 8081 killed almost all larvae of G. mellonella at 15 °C, but 40 CFU killed only 4% larvae at 30 °C; thus, showing the temperature-dependent toxicity of this pathogen in G. mellonella infection model (Springer et al. 2018). Similarly, another study on developing G. mellonella larvae as an in vivo model system for evaluating the virulence of Shigella showed time-dependent larval death accompanied by melanization (Barnoy et al. 2017). According to a previous work, the oral feeding of Y. enterocolitica (WA-314 pYVO8 +) to 6–8 weeks old female C57BL/6 mice caused a significant reduction in mean weight, two days post infection (Hering et al. 2016).

On performing label-free proteomics, putative hemolysin and secretory systems were emerged as the main Yersinia-identified proteins along with proteins for sensing, signaling and regulation. Previous study involving genomic comparison approach identified 329 genes that were present in highly-insecticidal strains (Yersinia intermedia and Yersinia frederiksenii) and absent in weakly-insecticidal strain (Yersinia enterocolitica strain W22703). Similar to the proteins identified in the present study, the products of previously-identified genes belong to the following categories: lipoproteins and other membrane proteins, metabolism, resistance toward toxic substances, sensing, signaling and regulation, transport and secretion, putative adhesins, toxins, hemolysins and secretory systems, stress response, iron uptake and storage (Springer et al. 2018). The gram-negative entomopathogenic bacteria have developed various strategies including specialized secretion systems to interact with the insects to kill them. The secretion systems transport several proteins from bacterial cytosol to surface, into the environment or directly into the host cells, where they have role in stimulating the virulence (McQuade and Stock 2018). From an earlier work, Y. frederiksenii and Y. intermedia were shown to carry a T6SS, which along with other functions, adds to the virulence (Springer et al. 2018). Additionally, the pathogenicity of Y. enterocolitica depends on T3SS. The secretion system facilitates the injection of virulence factors or Yops into the cytosol of the host cells. Yops modulates the cytoskeleton, thereby inhibiting the phagocytosis and inducing the cell death (Rakin et al. 2015). Inside insect hosts, Y. enterocolitica are confronted by changing and hostile surrounding conditions, which they have to sense and adapt to for their persistence. Interestingly, bacteria have evolved the two-component systems to sense and to react rapidly to the changing surrounding conditions. They are involved in metabolite utilization, regulation of symbiosis or virulence, and in adaptation to many stress factors. Additionally, in an earlier work, five virulence-associated (vag) genes including vagH, a methyltransferase were reported to be linked to virulence in mice as shown by infection with Yersinia. Other methyltransferases, vagB and vagE were found to be not essential but contained within operons involved in virulence (Garbom et al. 2004). Another report suggested that the ABC transporter genes irp6 and irp7 are required for mouse virulence in Y. enterocolitica O:8 (Brem et al. 2001). Further, nudA gene encoding a Nudix hydrolase is reported to be a virulence factor and important for resisting stress in Legionella pneumophila (Edelstein et al. 2005). The product of YPTB0333 gene encoding a transcriptional regulator from LysR family is found to control genes encoding Yfe iron-uptake system and polymyxin B resistance. Such that, using a mouse model of oral infection, it was shown that YPTB0333 is required for the colonization of Peyer's patches and mesenteric lymph nodes by Y. pseudotuberculosis (Arafah et al. 2009). Further, penicillin-binding protein 1B [Escherichia coli O44:H18 (strain 042/EAEC)], p-hydroxybenzoic acid efflux pump subunit AaeB [E. coli O26:H11 (strain 11368/EHEC)], and UDP-4-amino-4-deoxy-L-arabinose–oxoglutarate aminotransferase [E. coli O26:H11 (strain 11368/EHEC)] was found to be associated with antibiotic resistance (Manrique et al. 2011). Thus, the proteomic occurrence of the above mentioned pathogenic proteins in YADF insects gut supports the induced infection.

In insect, proteomics suggested increased melanization, i.e., insect innate immune response to infection and damage to host, which can be defined as the synthesis and deposition of melanin to encapsulate the pathogens at the injury site, accompanied by coagulation and opsonization. It is similar to the formation of abscess in mammalian infection (Tsai et al. 2016). Melanin synthesis is catalyzed by enzyme phenoloxidase (PO) that is produced as an inactive prophenoloxidase (PPO). Insect PPO is a crucial innate immunity protein due to its involvement in humoral and cellular defense. During a microbial infection, PPO cascade is activated by serine proteases that cleave PPO to PO. The activated PO transforms phenolic substrates to quinones, which polymerize to form melanin around wounds and invading pathogens. PO also produces reactive oxygen species (ROS), cytotoxic quinines, and reactive nitrogen that cause tissue toxicity; and therefore, its activation is tightly regulated by protease inhibitors. In a previous proteomic study on B. bassiana-infected silkworm, the expression levels of serine protease (NP_001036826) was up-regulated and of serine protease inhibitors, i.e., chymotrypsin inhibitor CI-8A (AAK52495), chymotrypsin inhibitor fb (AAK52496), fungal protease inhibitor F-like (XP_004924401), and serine protease inhibitor 7 precursor (NP_001139701) were down-regulated after B. bassiana infection. Thus, suggesting their role in melanization and defense against the fungal pathogens (Lü et al. 2019). Another study on baculovirus-infected H. armigera, reported that the infection induced serpin-5 and serpin-9 inhibits clip-domain serine proteases (cSPs), which resulted in reduced PO activity, to suppress the host melanization for the pathogen survival (Yuan et al. 2017).

The proteomics study further confirmed the affected metabolism of the insect gut in response to Yersinia feeding. Previous proteomic work on Wolbachia strain wStr-infected Aedes albopictus C/wStr1 cells, reported elevated energy metabolism proteins in infected cells. It may supplement the amino acids as well as the precursors for lipid and nucleic acid biosynthesis, which cannot be synthesized by the bacteria. Further, up-regulated superoxide dismutases (SODs), thioredoxins (Trxs), and peroxidases (PODs) indicated a response to oxidative stress in the infected cells. Collectively, altered metabolic signaling, protein degradation pathways and autophagy; with elevated TCA cycle and amino acid metabolism proteins suggests that the bacterial replication within the host-derived vacuoles escapes xenophagy, and sequesters the host-derived amino acids (Baldridge et al. 2017). Another proteomic study, to evaluate the alterations in C. elegans during Pseudomonas aeruginosa PAO1 infection, suggested the changed expression of TCA cycle proteins and oxidative phosphorylation machinery. Thereby, inducing the oxidative stress impaired protein homoeostasis mechanisms (Mir and Balamurugan 2019). Moreover, in the present study, ferritin, transferrin, and proteins with protective role against oxidative stress and heat shock response were significantly up-regulated; whereas, the proteins responsible for cell redox homeostasis were significantly down-regulated in response to the bacterial feeding. In this regard, a previous study on silkworm showed high expression of B. mori thioredoxin peroxidase (BmTPx) enzyme in response to baculovirus infection, suggesting the protective role of BmTPx against oxidative stress (Lee et al. 2005). Further, insect thioredoxin (Trx) is a thiol-dependent redox system, which regulates the responses generated by oxidative stress, with a prominent role in maintaining the cell redox homeostasis (Nappi and Christensen 2005). Ferritin and transferrin have important roles in iron storage and transport, respectively. Ferritins store excess iron and play a crucial role in iron homeostasis and management of the oxidative stress response in the invertebrates. Crustacean ferritins are implicated in innate immunity against the pathogens and resistance to various stresses (Tang et al. 2019). As known previously, specific innate immune responses limit iron availability to microbes during infection by decreasing the plasma iron concentrations within hours of infection. This response has been documented in mammals, including humans and other vertebrates (Ganz 2018). In a recent work, novel red fluorescent protein (RFP) was purified from the midgut of silkworm, whose N-terminal sequencing predicted chbp gene of lipocalin gene family with anti-pathogenic activities (Manjunatha et al. 2018). In corroboration to proteomics data, the biochemical assays depicting higher MDA level in YADF insects gut suggested enhanced LPO in comparison to the control larval gut. This causes loss of membrane integrity. Further, enhanced LPO and altered levels of antioxidant enzymes results in oxidative stress; thus, producing deleterious effects on the growth and development of H. armigera (Ahlawat et al. 2020). As reported previously, haemocytes produce free radicals to battle infections. Increased production of free radicals were observed between control group and Streptococcus pneumoniae (strain TIGR4 and D39)-infected larvae (Cools et al. 2019). Another study on Tenebrio molitor larvae infected with Heterorhabditis beicherriana suggested increased activities of superoxide dismutase (SOD), peroxidase (POD), catalase (CAT), and tyrosinase (TYR). Other detoxifying enzymes, like glutathione S-transferase (GST), carboxylesterase (CarE), and acetylcholinesterase (AChE) were also increased at lower infection rates suggesting that the host anti-oxidative response and detoxifying enzymes played an important role in the defensive reaction to infection (Li et al. 2016). However, ROS are also generated during respiration by respiratory complexes of mitochondrial electron-transport chain. Released ROS induces oxidative stress that caused increased expression of the stress response and cell redox homeostasis proteins. Further, increased activity of antioxidant enzymes was reported in entomopathogenic nematodes-infected G. mellonella larvae that suggests increased oxidative stress and anti-oxidative responses in response to the infection (Wu and Liu 2012).

Lastly, proteomics confirmed the affected cytoskeleton of the insect gut in response to Yersinia infection, as suggested by the histopathological and SEM analysis. Earlier report, also suggested depolymerization of actin stress fibres in eukaryotes after T3SS-injected YopE, an essential virulence determinant of Yersinia pseudotuberculosis (Von Pawel-Rammingen et al. 2000). Another study, reported reduced peritrophic membrane permeability (due to down-regulated midgut-specific chitin deacetylase-like protein) as a possible mechanism used by H. armigera to reduce the susceptibility to baculovirus (Jakubowska et al. 2010). Further, variations in IR peaks suggested change in biological macromolecules, which might be mediated by the generated ROS that impairs cell activities, such as gene expression and membrane function. In a previous work, damaged DNA in haemocytes of G. mellonella larvae was observed after Actinobacillus pleuropneumoniae infection. A. pleuropneumoniae cytotoxic factors cause damage to haemocyte DNA, thereby, killing and reducing the number of these cells in the larva. This effect is one of the mechanisms used by microbial cells to evade the host immune system (Pereira et al. 2015). Another study on Y. enterocolitica (WA-314 pYVO8 +) infected mice model system showed that the pathogen invade the colon mucosa by altering the tight junction (TJ) protein expression and architecture (Hering et al. 2016). Most of digestive and absorptive functions occur in the midgut of insect, such that any damage in midgut affects its growth and development, which eventually reduces the body weight (Marzban et al. 2013).

Finally, PPIs among proteins was studied via STRING11.0 software that revealed three interaction clusters generated: one between cell-redox proteins, another between cytoskeleton proteins, and third between stress-related HSPs and cell-redox proteins. Recent study based on interactome analyses (via STRING10.5) of C. elegans (host)—Salmonella enterica serovar Typhi (interacting pathogen) proteins demonstrated HSP-90 as a central player which coordinated other identified proteins and had role during pathogenic defense. Further, it was revealed to be up-regulated during (24 and 48 h) infection. This suggested that HSP-90 is required for regulatory events during infection in the host (Balasubramaniam et al. 2019). Another recent label-free proteomic work on LPS-537 treated sea urchin Paracentrotus lividus showed that LPS-treatment can modulate several processes like cytoskeleton re-organization, stress by induction of HSPs, and energetic homeostasis. The interactomes were also constructed by the early workers. It showed four clusters, mainly composed of cytoskeleton & cytoskeleton-related factors, RAS superfamily GTPase members, heat shock proteins (HSPs), and histones. The results confirmed that the cytoskeleton re-organization is a last process that is restored after the treatment (Inguglia et al. 2020).

Hence, pathogenicity is an outcome of both host and microbial factors—a concept described as the “damage-response framework”. Insect consists of cellular and humoral innate immune responses, where cellular component is composed of hemocytes and humoral defense includes soluble immune factors of hemolymph. The most significant immune response in insects is melanin synthesis, which occurs in response to the microbial invasion (Krachler et al. 2021). The key component of Yersinia–host cell interaction is its linkage with neutrophils during tissue infection. Neutrophils as crucial cells of innate immune system, it sense and kill pathogens by releasing signaling molecules and via degranulation, formation of neutrophil extracellular traps (NETs), generation of ROS, and phagocytosis, respectively. After infection with Yersinia, neutrophils migrate rapidly (with a delay in comparison to the recruitment by a yop mutant) to the Peyer’s patches and eventually encase the bacterial colonies such that they become the most adjacent cell-type to Yersinia. The delay demonstrates that a very early Yop injection into tissue cells modulates cytokine and chemokine release and retards the neutrophil recruitment. For instance, the replication of Y. pestis in draining lymph nodes induces very little host cytokine response that is inhibited by T3SS along with the other components of virulence plasmid. But later, neutrophils with other phagocytes are recruited to infected lymph nodes and IL-17 is produced. Similarly, immune responses to Y. enterocolitica and Y. pseudotuberculosis in Peyer's patches are marked by neutrophil influx and TH17/TH1 response (Davis 2018). In return, neutrophils are successively injected with Yops by Yersinia during infection as Yersinia has an ability to combat neutrophil invasion within the tissues. The T3SS effector proteins manipulate the host defense system to allow pathogen survival by inhibiting phagocytosis and inflammatory pathways that withstand T3SS innate immune recognition (Schubert et al. 2020).

Further, the existing evidences also suggest the ability of Yersinia to withstand ROS, as its growth is not influenced by the presence or absence of host-derived ROS (Davis 2018). Thus, pathogenicity of Yersinia in vertebrates has been shown to be affected by both bacteria-driven processes and host response-driven damage (Krachler et al. 2021). The above accounted host (mammals including humans) response to Yersinia is also observed in present study in H. armigera host. Yersinia infection in both hosts (mammals and H. armigera) induces the expression of proteins related to innate immune response, oxidative stress, and cytoskeleton. Further, similar Yersinia pathogenic proteins such as T3SS and others were also observed in both hosts (mammals and H. armigera) (Table 4). Thus, based on the theory of “damage-response framework” it can be concluded that Yersinia undoubtedly establishes the infection condition inside the insect midgut that truly mimics mammals gut interactions.

Table 4 Pathogenic and host-response proteins common in both Yersinia infected mammals and H. armigera model (present study)
Full size table

In conclusion, the experimental findings suggest that H. armigera larva can be considered as a potential in vivo model for studying the enteropathogenic infection by Y. enterocolitica strain 8081. Label-free proteomic analysis via an Orbitrap nano-LC–MS/MS helped to know the interaction of the enteropathogen with the insect host. In present study, the active infection by Y. enterocolitica strain 8081 in this model system led to its survival and growth, which is dependent upon the production of intact virulence factors like T3SS, T6SS, and hemolysins. In response, H. armigera responded to bacterial infection by significant induction of melanization, iron uptake proteins, and anti-pathogenic response proteins of the lipocalin family to limit the infection. In brief, Y. enterocolitica strain 8081 infection induces innate immune response in the insect gut that synthesizes melanin around the invading pathogen. In a cascading effect, ROS is generated that caused damage in YADF insects gut (Fig. 6). Thus, H. armigera can be used as an alternate, rapid, efficient, ethically acceptable, and cost-effective insect model system to study the pathogenesis of enteropathogens, including Y. enterocolitica.

Fig. 6
figure 6

Summary of Y. enterocolitica infection induced protein alterations in H. armigera model system

Full size image