Rhodnius prolixus uses the peptidoglycan recognition receptor rpPGRP-LC/LA to detect Gram-negative bacteria and activate the IMD pathway.

Insects rely on an innate immune system to recognize and eliminate pathogens. Key components of this system are highly conserved across all invertebrates. To detect pathogens, insects use Pattern recognition receptors (PRRs) that bind to signature motifs on the surface of pathogens called Pathogen Associated Molecular Patterns (PAMPs). In general, insects use peptidoglycan recognition proteins (PGRPs) in the Immune Deficiency (IMD) pathway to detect Gram-negative bacteria, and other PGRPs and Gram-negative binding proteins (GNBPs) in the Toll pathway to detect Gram-positive bacteria and fungi, although there is crosstalk and cooperation between these and other pathways. Once pathogens are recognized, these pathways activate the production of potent antimicrobial peptides (AMPs). Most PRRs in insects have been reported from genome sequencing initiatives but few have been characterized functionally. The initial studies on insect PRRs were done using established dipteran model organisms such as Drosophila melanogaster, but there are differences in the numbers and functional role of PRRs in different insects. Here we describe the genomic repertoire of PGRPs in Rhodnius prolixus, a hemimetabolous hemipteran vector of the parasite Trypanosoma cruzi that causes Chagas disease in humans. Using a de novo transcriptome from the fat body of immune activated insects, we found 5 genes encoding PGRPs. Phylogenetic analysis groups R. prolixus PGRPs with D. melanogaster PGRP-LA, which is involved in the IMD pathway in the respiratory tract. A single R. prolixus PGRP gene encodes isoforms that contain an intracellular region or motif (cryptic RIP Homotypic Interaction Motif-cRHIM) that is involved in the IMD signaling pathway in D. melanogaster. We characterized and silenced this gene using RNAi and show that the PGRPs that contain cRHIMs are involved in the recognition of Gram-negative bacteria, and activation of the IMD pathway in the fat body of R. prolixus, similar to the PGRP-LC of D. melanogaster. This is the first functional characterization of a PGRP containing a cRHIM motif that serves to activate the IMD pathway in a hemimetabolous insect.

Introduction

Insects rely on an innate immune defense system to recognize, activate the appropriate signal transduction pathway, and mount a robust effector immune response to eliminate invading pathogens (Leclerc and Reichhart, 2004, Müller et al., 2008, Li et al., 2018, Vilmos, 1998). The recognition phase relies on pattern recognition receptors (PRRs) that bind to signature molecules from pathogens known as pathogen-associated molecular patterns (PAMPs) (Akira et al., 2006, Janeway, 1989, Lu et al., 2020). This recognition activates appropriate immune signaling pathways that ultimately regulate immune responses including phagocytosis, melanotic encapsulation, and specifically the production of effector antimicrobial peptides (AMPs) (Akira et al., 2006, Janeway, 1989, Lu et al., 2020).

Classical studies on Drosophila melanogaster characterized the Immune Deficiency (IMD) and Toll pathways as master regulators of innate immune responses triggered by fungi and bacteria (Valanne et al., 2011, Kleino and Silverman, 2014). These recognition molecules, signaling pathways, and effector AMPs have been identified in all insects studied to date. Traditionally we have considered that most Gram-negative (Gr-) bacteria stimulate the IMD pathway, while fungi and Gram-positive (Gr+) bacteria stimulate the Toll pathway but there is strong evidence of crosstalk among these and other immune pathways (Nishide et al., 2019, Koyama et al., 2015, Wang et al., 2020). In all invertebrate systems, the detection of microorganisms is achieved by PRRs that activate the IMD and Toll pathways (Lu et al., 2020, Wang et al., 2019, Lin et al., 2020). These PRRs include the Gram-negative binding proteins (GNBPs) involved in the Toll pathway, and the peptidoglycan recognition proteins (PGRPs) involved in the IMD and Toll pathways. Other innate immune pathways, such as the JAK/STAT, RNA interference (RNAi), and autophagy pathways are activated by viruses, bacteria, and parasitoids but the role of PRRs in the activation of these pathways is not as well understood (Bang, 2019, Moy and Cherry, 2013, Blair, 2011).

Insects have multiple genes that encode PGRPs and GNBPs. The repertoire of these PRRs, the microorganisms they detect, the pathways they activate, and AMPs they induce have been well characterized in many holometabolous insects (Wang et al., 2019). These properties, however, are relatively unknown in most hemimetabolous insects. In fact, the existence of a functional IMD pathway itself in several hemimetabolous insect species was questioned when the publication of their genomes reported an absence of many proteins/genes that are considered essential for the function of the IMD pathway in holometabolous insects (International Aphid Genomics Consortium 2010, Benoit et al., 2016, Mesquita et al., 2015, Zumaya-Estrada et al., 2018). Subsequent studies using bioinformatics and RNAi approaches have identified almost all the ‘missing’ members of the IMD pathway in R. prolixus and several other hemimetabolous insects (Nishide et al., 2019, Salcedo-Porras et al., 2019).

PGRPs are proteins that contain a PGRP domain (∼160 amino acids (AAs) in length) homologous to the bacteriophage T7 lysozyme (an N-Acetylmuramyl-L-alanine amidase) that degrades cell wall peptidoglycan (PGN) from bacteria (Dziarski, 2004, Kang et al., 1998). In D. melanogaster, PGRPs are encoded by 13 genes that produce 19 proteins (Royet et al., 2011). These PGRPs can be classified as short (extracellular or intracellular proteins) or long forms (transmembrane, extracellular, or intracellular proteins). From a functional point of view, PGRPs can have amidase activity (catalytic) or not (non-catalytic PRR or regulatory proteins) (Wang et al., 2019, Royet et al., 2011). The former group is involved in the degradation of the bacterial cell wall, while the latter is involved in the binding of bacterial PAMPs, and therefore act as PRRs for the IMD (Kleino and Silverman, 2014) or Toll pathway (Valanne et al., 2011).

IMD pathway PGRPs bind to DAP-type PGN from Gr- bacteria and Bacillus sp., while Toll pathway PGRPs bind to Lys-type PGN from Gr+ bacteria. The amidase activity of PGRPs depends on the binding of Zn2+ ions and requires a 5 amino acid motif in the PGN-binding grove that is conserved with the bacteriophage T7 lysozyme. PGRPs that function as PRRs (dmPGRP-LC, dmPGRP-LE, dmPGRP-LA), lack one or several of these AA residues (Wang et al., 2019, Royet et al., 2011, Reiser et al., 2004).

The Toll pathway PGRPs in D. melanogaster (dmPGRP-SA and dmPGRP-SD) are secreted extracellular proteins involved in the recognition of Gr+ bacteria (Bischoff et al., 2004, Wang et al., 2006, Leone et al., 2008). The IMD pathway in D. melanogaster has PGRPs with different functions; some are involved in activating the pathway while others downregulate its activity. These IMD pathway PGRPs bind to the DAP-type PGN, and may be intracellular (dmPGRP-LE) (Chevée et al., 2019, Gottar et al., 2002) or membrane-bound (dmPGRP-LC and dmPGRP-LA) (Gottar et al., 2002, Gendrin et al., 2013). Other PGRPs block the detection of PGN by binding to dmPGRP-LC and preventing IMD pathway activation (dmPGRP-LF) (Maillet et al., 2008) (see [8] for a review and graphical description of these PGRPs). In other insects some PGRPs can detect both Gr- and Gr+ bacteria; in Tribolium castaneum the transmembrane PGRP-LA is involved in the activation of the IMD pathway by Gr- and Gr+ bacteria, and in the stinkbug Plautia stali the transmembrane PGRP-L2 also recognizes both types of bacteria and activates both pathways (Nishide et al., 2019, Koyama et al., 2015).

A signature characteristic of PGRPs involved in the activation of the IMD pathway (i.e. dmPGRP-LC and dmPGRP-LE) is a short N-terminus motif called a cryptic RIP Homotypic Interaction Motif (cRHIM). cRHIMs were described in insects due to similarities with the human kinase RIP1 and RIP3 (Kaneko et al., 2006, Kleino et al., 2017). In humans, these proteins induce necrosis upon viral infection by forming homotypic interactions of their RHIM (Li et al., 2012). RIP1 and RIP3 create amyloid fibrils (fibrous protein aggregates comprised of many of these proteins), that are hypothesized to function as signaling cores in the control of innate immune pathways (Kagan et al., 2014). In D. melanogaster, cRHIMs are present in dmPGRP-LC, dmPGRP-LE, dmPGRP-LA, dmIMD, dmPIRK (a negative regulator of IMD pathway signaling), and in dmRelish (the Nf-kb transcription factor of the IMD pathway). This cRHIM is required for PGRPs to activate the IMD pathway (Kaneko et al., 2006, Kleino et al., 2017). In the current model, Gr- bacteria are detected by dmPGRP-LC or dmPGRP-LE. These proteins form homodimers, and their cRHIMs seed the formation of amyloid fibrils that interact with the intracellular adaptor protein dmIMD (Kaneko et al., 2006, Kleino et al., 2017, Kleino and Silverman, 2019).

RHIMs are normally around 25 AAs in length and their sequence is loosely conserved; there is great variability in the RHIM AA sequence even among proteins from the same organism. This motif has a core of 4 AAs; in most RHIMs the core sequence is VxxG or IxxG, but only the fourth residue is completely conserved in all RHIMs (Kleino et al., 2017, Kajava et al., 2014). Most RHIM residues are hydrophobic and are predicted to form a 3 fold beta-sheet (Kleino et al., 2017, Kajava et al., 2014) that allows the formation of amyloid fibrils (Kajava et al., 2014).

In this paper, we describe the PGRP repertoire of Rhodnius prolixus, a hemipteran bloodfeeding insect. This insect has served as a model for insect physiological studies for over 80 years (Wigglesworth, 1933), and is a major vector of the parasite Trypanosoma cruzi that causes ∼12.000 human deaths annually in the Americas (WHO et al., 2020). We characterized one of these genes, rpPGRP-LC/LA, using bioinformatics and silencing approaches. This gene encodes PGRP proteins with transmembrane domains and N-terminal cRHIMs, which are characteristics of other insect PGRPs involved in the recognition of Gr- bacteria. We used RNAi to demonstrate that silencing the expression of rpPGRP-LC/LA effectively inhibits the recognition of Gr- bacteria and disrupts the expression of AMPs regulated by the IMD pathway in R. prolixus. To our knowledge, this is the most complete characterization of PGRPs in triatomines, which opens new avenues of research to investigate the recognition of pathogens in hemimetabolous insects, and potentially to disrupt parasite transmission by triatomines.

Materials and methods

Insect rearing

Two R. prolixus colonies were used. Samples obtained for the transcriptome assembly were obtained from a colony maintained in the insectary of the Institute of Medical Biochemistry at the Federal University of Rio de Janeiro (UFRJ). These insects were fed on live rabbits at three-week intervals and maintained at 28 °C and 80–90% relative humidity under a photoperiod of 12 h of light and 12 h of dark. Recently moulted (1–2 days) fifth instar nymphs were used in the experiments and were maintained under the standard rearing conditions described. Animal care and experimental protocols were conducted following the guidelines of the institutional care and use committee (UFRJ Committee for Evaluation of Animal Use for Research), which are based on the National Institutes of Health Guide for the Care and Use of Laboratory Animals (ISBN0-309-05377-3). The protocols were approved by the Committee for Evaluation of Animal Use for Research (CAUAP) from the UFRJ, under registry number 115/13. Dedicated animal care facility technicians carried out all aspects related to rabbit husbandry under strict guidelines to ensure careful and consistent handling of the animals.

Insects for the knockdown (KD) assays were selected from a colony of R. prolixus (Strain CDC, NR-44077) that is maintained in the Animal Care facility at Simon Fraser University, Burnaby BC, Canada. This colony was obtained originally from. Dr. Ellen Dotson and Alice Sutcliffe from the Centers for Disease Control and Prevention through distribution by BEI Resources, NIAID, NIH. Insects were fed with defibrinated rabbit blood (Hemostat Laboratories, USA) using artificial membrane feeders at three-week intervals and maintained at 28 °C and 80–90% relative humidity under a photoperiod of 12 h of light and 12 h of dark. Recently moulted (3–4 days) fifth instar nymphs were used in the experiments and were maintained under the standard rearing conditions described.

PGRP homolog search

A combination of strategies was used to identify the complete sequence of PGRPs in R. prolixus. To identify full length cDNA sequences, we assembled a de novo transcriptome (see below) and analyzed it along with other published R. prolixus transcriptomes (Ribeiro et al., 2014, Latorre-Estivalis et al., 2017). To identify PGRP coding transcripts we used the nucleotide and amino acid sequences of the three annotated PGRPs in the R. prolixus genome RproC3 (Mesquita et al., 2015, Giraldo-Calderón et al., 2015) (RPRC014061, RPRC012777, and RPRC007262) as queries in BLAST searches. We also scanned these transcriptomes using HMMER3’s hmmscan tool (Finn et al., 2011) with the HMM profile of the N-acetylmuramoyl-L-alanine amidase PF01510 domain from Pfam that is present in all PGRPs (El-Gebali et al., 2019). Putative transmembrane domains were identified using TMHMM v. 2.0 online server (Sonnhammer et al., 1998). The individual N-terminal amino acid sequences from these putative PGRPs were examined for the presence of cRHIMs by creating multiple sequence alignments with PGRP and IMD proteins from holometabolous insects and the human RIP1 and RIP3 proteins. The prediction of secondary structures in putative cRHIMs was done with SOPMA (Geourjon and Deléage, 1995) and multiple sequence alignments were built using the MUSCLE algorithm and edited manually using MEGA X software (Kumar et al., 2018).

Transcriptome assembly

We created a de novo transcriptome assembly from fat body tissues of R. prolixus that had been challenged with Gr- (Enterobacter cloacae) or Gr+ (Staphylococcus aureus) bacteria. Library construction and Illumina sequencing were done at the Biomedical Research Centre sequencing core at the University of British Columbia. RNA-seq libraries were constructed using the NEBnext Ultra ii Stranded mRNA (New England Biolabs, Canada). Sequencing was performed on the Illumina NextSeq 500 with Paired-End 80bp × 80bp reads. Sequence data were demultiplexed using Illumina's bcl2fastq2. The process yielded 560,758,510 total raw reads (∼31 million reads per sample). Sample quality processing was done using established protocols (Anon, 2020a). Sequence quality correction and adaptor trimming were performed with rCorrector (Song and Florea, 2015) and TrimGalore (https://www.bioinformatics.babraham.ac.uk/projects/trim_galore/). Ribosomal RNA contamination was removed by mapping the reads to the ribosomal RNA SSUParc and LSUParc from SILVA (Quast et al., 2013) using bowtie2 (Langmead and Salzberg, 2012), and only paired reads were further used in the assembly. Reads were analyzed with FastQC (Anon, 2020b) and MultiQC (Ewels et al., 2016) software throughout the quality processing. All the remaining Paired-End reads were used as inputs for the de novo assembly. Quantification of transcripts was done with trinity (Grabherr et al., 2011), bowtie2 (Langmead and Salzberg, 2012), RSEM (Li and Dewey, 2011), and edgeR (Robinson et al., 2010), and reported as Transcripts Per Million (TPM).

Evaluation of predicted PGRP transcripts

We verified the existence of PGRPs predicted in the de novo transcriptome by mapping the 8h PBS control RNA-seq samples onto the R. prolixus genome RproC3 (Mesquita et al., 2015, Giraldo-Calderón et al., 2015) using the program STAR (Dobin et al., 2013). Genomic scaffolds with PGRPs were processed and graphed to display read coverage and evidence of splice junctions with the program ggsashimi (Garrido-Martín et al., 2018). We compared PGRPs predicted from the de novo transcriptome with the genomic mapping to identify putative chimeric transcripts. We used PCR and subsequent Sanger sequencing to confirm or reject the presence of chimeric sequences, and used only confirmed sequences for subsequent analyses.

PCR amplification of PGRPs

Primers were designed for use in PCR amplification and sequencing of PGRP regions containing cRHIMs, and to confirm the expression of these transcripts in cDNAs synthesized from the RNA of fat body tissues from immune challenged insects. All PCR reactions used the ABM 2x PCR Taq Mastermix (ABM, Canada). Sequences were deposited in GenBank under accession numbers MW115914 to MW115929. The primer sequences and PCR conditions are listed in Supplementary File 1.

Phylogenetic analysis

Multiple protein sequence alignments were made for rpPGRPs. Full length representative PGRP sequences from multiple insects were selected from different genome assembly projects. Functionally characterized PGRPs were used to classify the different clades into PGRPs with or without amidase activity. The tunicate Oikopleura dioica PGRP protein sequence (Uniprot ID: Q5QFD0w) was used as the outgroup. Protein alignments were made using MUSCLE with default parameters in the program MEGA X (Kumar et al., 2018). Maximum Likelihood analyses were done using IQ-TREE v. 2.0 (Minh et al., 2020). Best-fit models for each gene alignment were selected based on the Bayesian Information Criterion (BIC). Branch support was assessed by 1000 ultrafast bootstrap replicates.

RNAi experiments: dsRNA target selection and synthesis

To confirm the role of PGRPs with a cRHIM in immune recognition and AMP regulation, we knocked down the expression of PGRPs that contained a putative cRHIM. We targeted the cRHIM coding sequence from R. prolixus rpPGRP-LC/LA mRNA and performed a selective analysis of the sequences to reduce the potential of off-target silencing during dsRNA injections. The PGRP cDNA sequences were divided into 19 nucleotide fragments and used as queries with BLASTn against the R. prolixus transcript-C3.1 data-set retrieved from Vectorbase (Giraldo-Calderón et al., 2015). Any hits against genes other than PGRPs larger than 17 nucleotides, > 90% identity, and < 3 gaps were mapped to the original cDNA and a region with the least number of potential off-target silencing fragments was selected to amplify a 380 bp fragment (dscRHIM). For a silencing control, the Arabidopsis thaliana AINTEGUMENTA (ANT) gene was subjected to the same analysis and primers were designed to amplify a 516 bp fragment (dsANT). All primers are listed in Supplementary File 1. PCR was used to amplify templates from cDNAs generated from the fat body of insects challenged with Gr− and Gr+ bacteria. The PCR products were ethanol precipitated and sequenced. These confirmed templates then were used to amplify dsRNA templates using the same primers to which we added 5’ T7 polymerase binding sites. These extended PCR products were ethanol precipitated, sequenced, and used to synthesize dsRNA using the TranscriptAid T7 High Yield Transcription Kit (Thermo Fisher Scientific, USA). The dsRNA was ethanol precipitated, visualized on 1% agarose gel, quantified on a Nanodrop 1000 spectrophotometer v. 3.7 (Thermo Fisher Scientific, USA), and resuspended in molecular grade water.

RNAi experiments: dsRNA injection and determination of knockdown (KD) efficiency

Fifth instar insects, 3–4 days after moulting, were injected intrathoracically with 1 μL volumes containing 2.5 μg of dsRNA generated from either rpPGRP-LC/LA or ANT, using a 10 μL Hamilton syringe. Fat body tissues were collected from injected insects 3–6 days after injection. These tissues were used to determine the efficiency and duration of the KD effect using real time quantitative PCR for different rpPGRP-LC/LA isoforms (see below).

Immune challenge

Based on the KD data, insects were immune challenged with bacteria three days after the injection of dsRNA. Gr- bacteria (Escherichia coli ATCC® 11303™) or Gr+ bacteria (Staphylococcus carnosus ATCC® 51365™) were grown in Luria-Bertani (LB) broth overnight at 200 rpm and 37 °C as described previously (Lowenberger et al., 1995). The liquid cultures were centrifuged, and pellets were washed 2 times in PBS (137 mM NaCl; 2,7 mM KCl; 10 mM sodium phosphate at pH7.2). The bacteria were centrifuged once more, and the supernatant discarded. A 000 entomological pin was dipped into the bacterial pellet and inserted into the hemocoel of the insect through the flexible cervical membrane. Each treatment used a minimum of 4 insects, and individual insects were used as replicates.

Tissue isolation, RNA extraction, and cDNA synthesis

The fat body tissues were dissected 8 and 24 h after immune challenge and total RNA was extracted using TRizol reagent (Invitrogen) following the manufacturer's recommendations. The samples were quantified on a Nanodrop 1000 spectrophotometer v. 3.7 (Thermo Fisher Scientific, USA). First strand cDNA synthesis was performed in 20 μL reactions containing 2.0 μg total RNA using an oligo dT primer (MG) with the OneScript cDNA Synthesis Kit (ABM, Canada) and an extension time of 50 minutes. The subsequent cDNA was diluted 1:50 with DEPC water.

Quantitative real-time PCR

Quantitative real-time PCR (qPCR) was used to assess the effect of PGRP KD on the expression of selected AMPs known to have differential expression in insect fat body tissues after infection with bacteria. Primers were designed, or retrieved from the literature, for: α-Tubulin, Prolixicin, Lysozyme-B, Defensin-B, rpPGRP-LC/LAa, and rpPGRP-LC/LAb (Supplementary File 1). All qPCR reactions contained 4.4 μL of cDNA, 300mM of each primer, and 5 μL of PerfeCTa SYBR Green Super Mix (Quanta Biosciences, USA) in a final volume of 10 μL. qPCR was performed on a LightCycler96 thermal cycler (Roche diagnostics, Germany). The qPCR conditions were: 95 °C: 5 min, 40 cycles of 95 °C: 15 s and 60 °C: 40 s, followed by a melt curve analysis to confirm the specificity of the reactions. No-template controls were included with each primer set to verify the absence of exogenous DNA and primer-dimers. Relative differences in transcript levels were calculated using the 2– ΔΔ Ct method (Schmittgen and Livak, 2008, Livak and Schmittgen, 2001) with α-Tubulin as the reference gene. PCR efficiencies (E) for each primer were determined previously using the slope of a linear regression model. Due to sequence similarity, it was not possible to quantify individual rpPGRP-LC/LA isoforms. The primer pairs used to estimate rpPGRP-LC/LAa and -b levels also amplified other rpPGRP-LC/LA isoforms. The silencing values are therefore a conservative measurement as some of the other isoforms are not targeted by dscRHIM RNA but are amplified by the primers we designed (Fig. 1B and Supplementary File 1). The combined expression levels of these other isoforms is a fraction (∼1/6) of rpPGRP-LC/LAa and -b levels and therefore the majority of the silencing reported corresponds to rpPGRP-LC/LAa and -b isoforms (Supplementary File 1).

Fig. 1

The five PGRP genes and isoforms in Rhodnius prolixus. A. Representation of the 5 genes and the 13 predicted PGRP protein isoforms in the fat body de novo transcriptome. rpPGRP-LC/LA encodes 2 isoforms with a N-terminal intracellular cRHIM (Black rectangle) and encodes PGRP domains with 2 different AA sequences, each represented with a unique color. The other 4 PGRP genes lack a predicted cRHIM. Proteins with transmembrane domains are represented as crossing a bilipid bilayer. Isoform names for PGRP isoforms are below each protein. B. Expression levels for the 13 PGRP isoforms in the fat body of R. prolixus from the de novo transcriptome of immune activated insects. Expression values in Transcripts Per Million (TPM) are shown for the 8h post PBS injection control samples. These values align vertically with the rpPGPRs shown in panel 1A. Transcript levels from the same sample are linked with a unique-type line. Similar expression levels and trends are seen in all samples and treatments (Supplementary File 1). C. The N-terminus sequences of PGRP and IMD proteins that contain a cRHIM and are involved in IMD pathway signaling are shown. Residue conservation is denoted in blue, and darker colors represent higher levels of conservation. The red box denotes the cRHIM. D. The regions of putative secondary structures of selected proteins that contain RHIMs (as predicted by SOPMA (Geourjon and Deléage, 1995)) are shown. Beta-turns and extended strands are shown in dark purple, alpha helices are shown in light yellow. The underscore sequence in Hs-RIP1 and Hs-RIP3 denote the sequences that form the 3-fold beta-sheets involved in amyloid formation (Kajava et al., 2014). Dm: Drosophila melanogaster, Ag: Anopheles gambiae, Am: Apis mellifera, As: Anopheles stephensi, Gm: Glossina morsitans, Nl: Nilaparvata lugens, Tm: Tenebrio molitor, Aa: Aedes aegypti, rp: Rhodnius prolixus, Tc: Tribolium castaneum, Hs: Homo sapiens, B: Bombyx mori.

Statistical analysis

All treatments were tested for normality using the Shapiro-Wilk normality test and compared using the unpaired Student's T–test or Mann Whitney U Test. Calculations and graphs were made using R 3.0.14 and JMP 13.1. P-values lower than 0.05 were considered significantly different. Relative transcript levels were expressed as means with whiskers representing ± SEM.

Data availability

The data used in the analyses can be found in Supplementary File 1, and include: PGRP and IMD sequences and accession numbers, a phylogenetic tree Newick file, the TPM for individual rpPGRP-LC/LA isoforms, and the qPCR Ct values and fold change calculations.

Results

PGRP repertoire in Rhodnius prolixus

Five genes encoding proteins with putative PGRP domains were identified by BLAST and HMMER searches (Fig. 1A, Supplementary File 1, and Supplementary File 2). These sequences align with incomplete annotations of the genes RPRC007262, RPRC007261, RPRC014061, RPRC014065, RPRC012777, and the scaffold KQ34073. A total of 13 PGRP isoforms are encoded by these genes: Four are encoded by rpPGRP-LC/LA, one by rpPGRP-SC/SA, two by rpPGRP-1, four by rpPGRP-2, and two by rpPGRP-3 (Fig. 1A and Supplementary File 2).

Although most of the PGRPs have long N-terminal sequences, only two isoforms of rpPGRP-LC/LA contained a putative cRHIM (Fig. 1A, 1C, and 1D); the remaining 2 isoforms of this gene use an alternative exon that encodes a different intracellular sequence. Alignments of RHIM sequences from human and insect proteins with the N-terminal regions of the R. prolixus PGRPs show conservation of specific AAs in most of the R. prolixus isoforms (Fig. 1D). However, only the AA sequences in rpPGRP-LC/LAa and -b have a region predicted to contain a 3-fold beta-sheet (Fig. 1D), and we propose that this is a functional cRHIM. The other R. prolixus PGRP sequences are rich in polar AAs, which are not common in beta-sheets, and likely are not functional cRHIMs (Fig. 1D).

The average expression levels of PGRP isoforms from all the fat body de novo transcriptome samples were highest for rpPGRP-LC/LAa (47.0 TPM, CI95 43.2 TMP – 50.9 TPM) and -LC/LAb (33.5 TPM, CI95 29.6 TPM – 37.4 TPM), both of which contain putative cRHIMs (Fig. 1B and Supplementary File 1).

The next highest expression levels occur in rpPGRP-LC/LA isoforms that lack a cRHIM: rpPGRP-LC/LAa’ (14.3 TPM, CI95 10.2 TMP – 17.9 TPM) and -b’ (8.8 TPM, CI95 4.9 TPM – 12.7 TPM). Similar expression values were seen in rpPGRP-SC/SA, a gene similar in sequence to rpPGRP-LC/LA, and which encodes proteins with a TM domain and two PGRP domains, but which lacks an intracellular region (14.3 TPM, CI 10.5 TPM – 17.3 TPM) (Fig. 1B and Supplementary File 1).

The remaining isoforms have very low expression levels, less than 5 TPM (Fig. 1B and Supplementary File 1). rpPGRP-LC/LAa and -LC/LAa’, have identical extracellular sequences but variable intracellular sequences. rpPGRP-LC/LAb and -b’ also share the extracellular sequence but these are different from the rpPGRP-LC/LAa and -a’ sequences. The intracellular region of rpPGRP-LC/LAa’ and -b’ is identical in sequence (Supplementary File 1).

To evaluate the evolutionary relationships ofR. prolixus PGRPs, we built Maximum Likelihood (ML) phylogenetic trees with PGRP protein sequences from diverse insect orders including other triatomine species (Fig. 2A). The phylogenetic tree contains 2 main branches, one clade contains PGRPs with known amidase activity while the other contains PGRPs without amidase activity that usually are involved in signaling within the IMD pathway (Wang et al., 2019). All the PGRPs from R. prolixus, and other triatomines, cluster in the latter clade. The closest PGRP with known function to triatomine PGRPs is D. melanogaster PGRP-LA. The triatomine clade is divided into 4 well-supported subclades (Fig. 2B). Other triatomine species have 3 or 4 PGRP encoding genes, each located within one of the 4 subclades (Fig. 2B).

Fig. 2

Maximum likelihood phylogenetic tree of PGRPs from selected insects. A. Insect PGRPs separate into a clade of PGRPs with amidase activity (bottom clade highlighted in brown) and a clade of PGRPs with no predicted amidase activity (top clade highlighted in purple) (Wang et al., 2019). All triatomine PGRPs are contained within the latter clade and are shown in blue. Some clades containing multiple species are collapsed for display purposes. B. A detailed tree of the triatomine PGRP clade shows four main groups. Trees are represented with transformed branches to display evolutionary relationships more clearly. A Newick file with the complete tree and branch lengths can be found in Supplementary File 1. Clade support is shown as percentage values of 1000 ultrafast bootstrap replicates.

The R. prolixus PGRP genes encode 7 different PGRP domains; rpPGRP-LC/LA encodes 2, rpPGRP-SC/SA encodes another 2, and each of the other PGRP genes encodes a unique PGRP domain. An analysis of the PGRP domains showed that they lack essential residues required for amidase activity (Fig. 3).

Fig. 3

The PGRP domain sequences from selected insects. PGRP sequences with amidase activity are shown in the top section and PGRPs without amidase activity are shown in the middle section. The 5 amino acid residues required for amidase activity are highlighted in red and their positions are indicated with an arrow. Rhodnius prolixus PGRPs are shown in the lower section and lack several of these required residues. Dm: Drosophila melanogaster, Ag: Anopheles gambiae, Am: Apis mellifera, As: Anopheles stephensi, Gm: Glossina morsitans, Nl: Nilaparvata lugens, Tm: Tenebrio molitor, Aa: Aedes aegypti, Rp: Rhodnius prolixus.

RNAi indicates that rpPGRP-LC/LA is involved in the immune response against Gram- bacteria

The expression of AMP transcripts in the fat body tissues of R. prolixus that had been injected with dsRNA was measured 8 and 24 h after immune challenge. AMP expression levels were compared between insects that had been injected previously with dsANT (as a control) or dscRHIM (complementary to the sequence in rpPGRP-LC/LAa and -b transcripts coding for a putative cRHIM). cRHIM KD reduced the expression of rpPGRP-LC/LAa by 93%, and rpPGRP-LC/LAb by 79% compared with dsANT injected insects (Fig. 4A). These isoforms share their extracellular sequence with rpPGRP-LC/LAa’ and rpPGRP-LC/LAb’, respectively, therefore these silencing levels are likely very conservative.

Fig. 4

Effects of knocking-down rpPGRP-LC/LAa and b on the expression of fat body AMPs in 5th instar Rhodnius prolixus challenged with Gram-negative (Gr-) or Gram-positive (Gr+) bacteria. Insects were injected with 2.5 μg of dsRNA complementary to the cRHIM sequence in rpPGRP-LC/LAa, and -b, or ANT, a plant gene that serves as a dsRNA injection control. Silencing efficiency for rpPGRP-LC/LAa and rpPGRP-LC/LAb was measured 3 days post dsRNA injection (A). At this point, the insects were injected with Gr- bacteria (Escherichia coli ATCC® 11303™) (B and D) or Gr+ bacteria (Staphylococcus carnosus ATCC ® 51365™) (C and E). The relative expression of selected AMPs was evaluated using the ΔΔCT method (Schmittgen and Livak, 2008, Livak and Schmittgen, 2001) and data are presented as fold differences between dscRHIM and dsANT silenced insects at 8 (B and C) and 24 hours (D and E) post immune challenge (hpi). Bars represent the mean transcript levels ± 95% CI. Means were compared using the unpaired Student's T–test. P values are displayed above each compared group.

In cRHIM KD insects infected with Gr- bacteria the levels of Defensin-B, Lysozyme-B, and Prolixicin decreased to 28% (CI95 16–33%), 50% (CI95 38–65%), and 23% (CI95 -12–36%) respectively of their corresponding controls at 8 hours post bacterial inoculation (Fig. 4B). A stronger reduction was seen 24 h post bacterial inoculation; the levels of Defensin-B, Lysozyme-B, and Prolixicin decreased to 13% (CI95 0–22%), 17% (CI95 12–26%), and 17% (CI95 -1–36%), respectively of their corresponding controls (Fig. 4D).

In contrast, 8 h after treatment with Gr+ bacteria, the levels of AMP expression did not differ significantly between insects that had received dscRHIM or dsANT (Fig. 4C). A reduction in AMP levels was seen 24 h after Gr+ bacteria treatment, Defensin-B, Lysozyme-B, and Prolixicin levels decreased to 13% (CI95 -130–37%), 27% (CI95 -46–44%), and 32% (CI95 -189–101%) although the reduction was not statistically different from the AMP levels in the dsANT control (P-value > 0.05) (Fig. 4).

Discussion

The recognition of pathogens as nonself is paramount in order to activate all components of the innate immune response repertoire. This is especially relevant for insects such as R. prolixus that live in environments filled with microorganisms. This species also has served as the most important hemimetabolous insect model to study basic insect physiology, but in the wild also serves as a major vector of T. cruzi, the causal agent of human Chagas disease. Understanding how triatomines such as R. prolixus recognize microbial pathogens may open new avenues of research into how and why they do not also eliminate T. cruzi. Usually, triatomines exposed to sympatric T. cruzi strains (i.e. that exist in the same geographic area as the vector) do not eliminate the parasites, but they can eliminate, or reduce the numbers of parasites from allopatric T. cruzi strains (Dworak et al., 2017, Brenière et al., 2016). These results likely depend on specific recognition mechanisms unrelated to the PGRPs discussed here.

The study of innate immune responses in triatomines, or in hemimetabolous insects in general, has lagged considerably behind studies in holometabolous insects such as D. melanogaster. In fact, we still lack a complete understanding of many aspects of the composition and regulation of their innate immune system. In the publication on the genome of R. prolixus (Mesquita et al., 2015), and also in the genomes of other hemimetabolous insects (International Aphid Genomics Consortium 2010, Benoit et al., 2016), the IMD pathway was described as being either absent or nonfunctional because of the absence of many canonical proteins present in all holometabolous insects. This was probably due to the use of automated genome annotation programs, significant sequence divergence, and the presence of many short introns (Panfilio et al., 2017). Subsequently, other members of the IMD pathway were identified (Salcedo-Porras and Lowenberger, 2019) and the pathway was confirmed to be functional in responding to microbial insult, especially Gr- bacteria using RNAi (Salcedo-Porras et al., 2019). Now that the canonical IMD pathway is considered present and functional in R. prolixus, the question of how pathogens are recognized to initiate the overall immune responses within the IMD pathway was addressed.

Using a de novo fat body transcriptome, phylogenetic analyses, and genomic comparisons we identified five genes that encode thirteen PGRP proteins in R. prolixus. One of these genes, rpPGRP-LC/LA, encodes proteins with intracellular cRHIMs that are also present in other insect PGRPs that activate the IMD pathway. Knocking down transcripts that have this cRHIM reduced the expression of AMPs that are induced by Gr- bacteria and regulated by the IMD pathway (Salcedo-Porras et al., 2019, Vieira et al., 2020), suggesting that rpPGRP-LC/LA is a Gr- bacteria PRR involved in the IMD pathway of triatomines.

The original studies with D. melanogaster showed that while some PGRPs function as amidases that directly control bacterial infections by digesting PGN, others act as PRRs of cell wall PGNs and activate the IMD pathway. Most insects have both types of PGRPs (Nishide et al., 2019, Wang et al., 2019, Zhou et al., 2019). We analyzed multiple triatomine genomes and transcriptomes, and identified triatomine PGRPs that are all contained within well-supported subclades, but we did not identify any PGRPs with a predicted amidase activity. This might be a unique characteristic of triatomines as other hemipterans have both amidase and PRR PGRPs (Nishide et al., 2019, Zhou et al., 2019), but in contrast, the Pea aphid has no PGRPs (Elsik, 2010). rpPGRP-LC/LA resembles two PGRPs found in D. melanogaster. From an evolutionary perspective, and based on our phylogenetic data, rpPGRP-LC/LA is closely related to D. melanogaster PGRP-LA, as is another PGRP found in another hemimetabolous insect; P. stali PGRP-L2 (Nishide et al., 2019). From a functional perspective, our KD assays on AMP expression in the R. prolixus fat body indicate that rpPGRP-LC/LA is a functional analog of dmPGRP-LC. In D. melanogaster dmPGRP-LA acts in the respiratory tract as the main PRR for the IMD pathway (Gendrin et al., 2013), but in the fat body, this function is carried out by dmPGRP-LC (Werner et al., 2000, Choe et al., 2005). As we knocked down all rpPGRL-LC/LA isoforms, and only evaluated one tissue, it remains to be seen if there are tissue-specific patterns of expression and functions for rpPGRP-LC/LA in R. prolixus.

A consistent characteristic of all holometabolous insect PRR PGRPs from the IMD pathway is an N-terminal cRHIM (Wang et al., 2019, Kleino et al., 2017, Kagan et al., 2014, Kleino and Silverman, 2019). We could only detect this motif in rpPGRP-LC/LA. This gene can generate PGRPs with 2 different intracellular regions via alternative splicing of 2 exons, but only one of these exons encodes a cRHIM, producing the isoforms rpPGRP-LC/LAa and -b. This motif is present in various IMD pathway molecules of D. melanogaster and is believed to help form supramolecular organizing centres (SMOC) (Kagan et al., 2014, Kleino and Silverman, 2019). In humans, SMOCs aid the signaling of necroptosis (Kajava et al., 2014). In D. melanogaster they are proposed to be involved in the IMD pathway signaling (Kleino and Silverman, 2019). The presence of this motif in rpPGRP-LC/LAa, and -b supports the idea that SMOCs are highly prevalent and rapidly amplify innate immune recognition events (Kagan et al., 2014). Normally the cRHIM would form amyloids with the IMD adaptor protein, but this protein, has not yet been identified or characterized in R. prolixus. Ongoing studies have identified candidate IMD proteins from the de novo transcriptome (Salcedo and Lowenberger, unpublished) and these will be characterized for amyloid formation.

Drosophila melanogaster PGRP-LC, -LE, and -LA are the main IMD pathway PRRs that are highly expressed in the fat body, gut, and trachea, respectively (Gendrin et al., 2013, Kaneko et al., 2006, Leader et al., 2018). The high expression of rpPGRP-LC/LAa and -b (both contain a cRHIM) in the fat body of R. prolixus suggest that these proteins are the main PRRs in this tissue. Due to their low expression levels and their lack of a predicted cRHIM, it is unlikely, that PGRPs other than rpPGRP-LC/LA are involved in the IMD pathway signaling activation in triatomine fat body tissues. Although a careful dissection of the fat body tissue for the transcriptome construction was done, small amounts of other tissues such as trachea and hemocytes may have contaminated our samples. Therefore, we cannot rule out the idea that rpPGRP-LC/LAa and –b might also function as the principal PRRs in these tissues.

There is a lack of key residues that indicate a role in amidase activity in all triatomine PGRPs. In other insects, there are PGRPs without amidase activity that do not activate the IMD pathway (Nishide et al., 2019, Wang et al., 2019, Zhou et al., 2019, Tavignot et al., 2017, Basbous et al., 2011, Song et al., 2018). For instance, D. melanogaster dmPGRP-LF is a negative regulator of the IMD pathway that blocks the homodimerization of dmPGRP-LC (Tavignot et al., 2017, Basbous et al., 2011). This protein has a very short intracellular region (23 AA) and 2 PGRP extracellular domains. rpPGRP-SC/SA also has 2 PGRP domains and lacks a cRHIM. Similarly, rpPGRP-LC/LAa’ and -b’ lack a cRHIM but only have 1 PGRP domain. These 3 proteins might act as negative regulators of the IMD pathway in a similar way to dmPGRP-LF. The longer (>50 AAs in length) N-terminal intracellular region in rpPGRP-LC/LAa’ and -b’ might also interfere with the formation of amyloids, further blocking IMD pathway activation. rpPGRP-1, -2, and -3a also lack a cRHIM and have variable intracellular region lengths, but we cannot state that these proteins modulate IMD pathway activation. In D. melanogaster, secreted PGRPs usually function as amidases (Wang et al., 2019), except for dmPGRP-LE that is an IMD pathway PRR. In R. prolixus, only rpPGRP-3b lacks a transmembrane domain, and its PGRP domain is predicted to lack amidase activity. The low expression of rpPGRP-3b and the absence of a cRHIM suggest that this protein is not likely a PRR for the IMD pathway in the fat body.

Many insects have at least one PGRP with amidase activity, usually named PGRP-LB, that strongly downregulates the activation of the IMD pathway (Wang et al., 2019, Zaidman-Rémy et al., 2018, Dawadi et al., 2018). Infections with a few bacteria can be controlled by baseline immune system components. To prevent the activation of the IMD pathway, PGRP-LB hydrolyses PGN, and avoids the unnecessary metabolic costs associated with a systemic immune response (Wang et al., 2019, Zaidman-Rémy et al., 2018, Dawadi et al., 2018, Zou et al., 2007, Paredes et al., 2011). In other insects, PGRP-LB also regulates the activation of immune responses towards beneficial endosymbionts (Zaidman-Rémy et al., 2018). The absence of any gene resembling PGRP-LB in triatomines is therefore unexpected as R. prolixus relies heavily on microbial symbionts in the intestinal tract. Although we predict that no PGRP in R. prolixus will digest PGNs this should be corroborated experimentally.

The IMD pathway in R. prolixus activates the transcription of Def-A, Def-B, Lys-B, and Prolixicin (Salcedo-Porras et al., 2019, Vieira et al., 2020). Our KD of rpPGRP-LC/LAa and -b downregulated the expression of these AMPs when the KD insects were challenged with Gr- bacteria, indicating the role of this PGRP as a PRR in the IMD pathway. There were no significant differences in AMP expression 8 hours after the KD insects were challenged with Gr+ bacteria. A trend towards a reduced expression of all three AMPs was seen 24 hpi after infection with Gr+ bacteria, but these differences were not statistically different from the controls. The expression of AMP transcripts in triatomines and other insects is a fast inducible response with a peak in AMP transcripts 4 – 8h post bacterial exposure (Nishide et al., 2019, Jo et al., 2019, Sanda et al., 2019, Erler et al., 2011, Lopez et al., 2003, Ursic-Bedoya et al., 2011, Ursic-Bedoya et al., 2008). The late downregulation in our Gr+ bacteria infections may imply that the induction of these AMPs is not directly related to the Gr+ bacterial challenge and that it is a result of interactions with the IMD pathway and crosstalk among multiple immune pathways. Cross talk between the Toll and IMD pathways is known in P. stali 24 h post bacterial stimulation and this effect is dependent on PGRP-L2 and the PRR lysin motif proteins (Nishide et al., 2019). Similarly, the Toll and IMD pathways can synergistically control the production of AMPs in other insects, although individual AMPs normally are regulated principally by only one of these pathways (Tanji et al., 2007, Tinoco-Nunes et al., 2016). In triatomines, the pharmacological inhibition of the Toll pathway downregulates the expression of Defensins and Prolixicin in the midgut, supporting a potential cross talk between the IMD and toll pathway in these insects (Vieira et al., 2018).

Conclusions

Our data serve to confirm that the IMD pathway is functional in R. prolixus, and likely in other triatomines, and that it regulates the expression of potent AMPs that regulate microbial infections. In other insects, the PAMP associated with Gr- bacteria that is detected by PGRPs is the DAP-type PGN, and we expect that rpPGRP-LC/LAa and -b will each have a similar binding affinity.

The IMD pathway in R. prolixus may now be considered to be the full, predicted, and expected canonical immune pathway described in holometabolous insects such as D. melanogaster. Determining how triatomines and other hemimetabolous insects use PGRPs and other PRRs will allow us to better understand the evolution of pathogen recognition mechanisms in insects, and how PRRs interact with other components of innate immune pathways.

Funding

This work was supported in part by an NSERC Discovery Grant [RGPIN261940 and 261940] to CL; by a MITACS Globalink Research Award [IT08650] to NSP; and by SFU President's Research Grant and Graduate Fellowships to NSP. The funders had no role in study design, data collection, and analysis, decision to publish, or preparation of the manuscript.

CRediT authorship contribution statement

Nicolas Salcedo-Porras: Conceptualization, Methodology, Visualization, Formal analysis, Investigation, Writing - original draft, Writing - review & editing. Shireen Noor: Investigation, Writing - review & editing. Charley Cai: Investigation. Pedro L. Oliveira: Conceptualization, Methodology, Visualization, Writing - original draft, Writing - review & editing. Carl Lowenberger: Conceptualization, Methodology, Writing - original draft, Writing - review & editing, Supervision.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.