Structural and Functional Analysis of PGRP-LC Indicates Exclusive Dap-Type PGN Binding in Bumblebees

Peptidoglycan recognition proteins (PGRPs) play an important role in the defense against invading microbes via the recognition of the immunogenic substance peptidoglycan (PGN). Bees possess fewer PGRPs than Drosophila melanogaster and Anopheles gambiae but retain two important immune pathways, the Toll pathway and the Imd pathway, which can be triggered by the recognition of Dap-type PGN by PGRP-LCx with the assistance of PGRP-LCa in Drosophila. There are three isoforms of PGRP-LC including PGRP-LCx, PGRP-LCa and PGRP-LCy in Drosophila. Our previous study showed that a single PGRP-LC exists in bumblebees. In this present study, we prove that the bumblebee Bombus lantschouensis PGRP-LC (Bl-PGRP-LC) can respond to an infection with Gram-negative bacterium Escherichia coli through binding to the Dap-type PGNs directly, and that E. coli infection induces the quick and strong upregulation of PGRP-LC, abaecin and defensin. Moreover, the Bl-PGRP-LC exhibits a very strong affinity for the Dap-type PGN, much stronger than the affinity exhibited by the PGRP-LC from the more eusocial honeybee Apis mellifera (Am-PGRP-LC). In addition, mutagenesis experiments showed that the residue His390 is the anchor residue for the binding to the Dap-type PGN and forms a hydrogen bond with MurNAc rather than meso-Dap, which interacts with the anchor residue Arg413 of PGRP-LCx in Drosophila. Therefore, bumblebee PGRP-LC possesses exclusive characteristics for the immune response among insect PGRPs.


Introduction
Innate immunity is an important line of defense that protects insects from pathogen invasion. It can be activated by interactions between germline-encoded pattern recognition receptors (PRRs) and pathogen-associated molecular patterns (PAMPs) [1][2][3][4]. As an important type of PRR, peptidoglycan recognition proteins (PGRPs) were first discovered in the hemolymph of silkworms and participate in an antimicrobial host defense mechanism [5]. PGRPs have been shown to be conserved from mammals to insects, except in lower metazoans (nematodes) and plants that do not have PGRPs [6].
Nevertheless, insects possess more diverse PGRPs than mammals; for instance, Drosophila melanogaster has 13 PGRP genes and Anopheles gambiae possesses seven PGRP genes [6]. Based on their length, insect PGRPs can be divided into two classes: short PGRPs (PGRP-S) and long PGRPs (PGRP-L). Each PGRP contains a single peptidoglycan (PGN)-binding domain. These PGRPs recognize the bacterial or fungal PGN and then trigger two important immune pathways, the Toll

Response of Bl-PGPR-LC to Infection with Gram-Negative and Gram-Positive Bacteria
The present study showed that both the Gram-positive bacterium Staphylococcus aureus and the Gram-negative bacterium Escherichia coli can trigger the upregulation of PGRP-LC and the antimicrobial peptides abaecin, defensin and hymenoptaecin ( Figure 1). However, over time, the degree of response differs between S. aureus and E. coli. After E. coli infection, the PGRP-LC gene showed no change within 6 h, while hymenoptaecin was significantly upregulated ( Figure 1A). After 12 h, the PGRP-LC, abaecin and defensin genes were strongly upregulated ( Figure 1A). However, after 24 h, an obvious decrease in the degree of upregulation of the PGRP-LC, abaecin and defensin genes was observed compared to that at 12 h of infection; in contrast, hymenoptaecin was upregulated, which was similar to the result at 6 h of infection ( Figure 1A). In contrast to the E. coli infection, the S. aureus infection only caused the upregulation of PGRP-LC and defensin after 12 h and none of the four genes were upregulated within 6 h ( Figure 1B). After 24 h, all four genes were upregulated significantly, and the upregulation was obviously higher than that at 12 h of infection ( Figure 1B).

Bombus PGRP-LC Strongly Bound to Dap-Type PGN and Exhibited Unique Anchor Residues in Comparison with Drosophila PGRP-LCx
The bumblebee Bl-PGRP-LC is upregulated after infection with S. aureus and E. coli, so the bumblebee Bl-PGRP-LC protein should have the ability to bind to the Lys-type PGN and the Daptype PGN. However, the extracellular region of Bl-PGRP-LC is able to bind to the Dap-type PGN rather than the Lys-type PGN and shows a high affinity for the Dap-type PGN ( Figure 2A). As a control, the known Dm-PGRP-LCx exhibited no binding to the Lys-type PGN or the Dap-type PGN ( Figure 2B). Interestingly, the honeybee (A. mellifera) Am-PGRP-LC exhibited a weak affinity for the Dap-type PGN and no binding to the Lys-type PGN, but the short PGRP-LC had no ability to bind to the Dap-type PGN or the Lys-type PGN ( Figure 2C). To eliminate the influence of the Glutathione Stransferase (GST) tag on the binding of PGNs, the GST protein was purified and the binding test was performed on the PGNs. The GST protein did not bind to the Dap-type PGN or the Lys-type PGN ( Figure 2D). Based on the contact residues of Drosophila PGRP-LCx for the binding of TCT from the PGRP-LCx/PGRP-LCa/TCT complex, we mutated the 14 corresponding residues of Bl-PGRP-LC, namely, His 282 , Thr 283 , His 305 , Arg 309 , Ser 312 , Tyr 316 , Arg 330 , His 338 , Phe 340 , Asn 343 , His 390 , Ser 394 , Arg 395 and Leu 397 . All these residues were mutated to Ala. None of these mutations showed any effect on the recognition of Lys-type PGN by Bl-PGRP-LC ( Figure 3). In addition, the His 282 , Thr 283 , His 305 , Arg 309 , Ser 312 , Tyr 316 , Arg 330 , His 338 , Phe 340 , Asn 343 , Ser 394 , Arg 395 and Leu 397 mutations did not alter the binding to Dap-type PGN ( Figure 3A). Remarkably, the mutation of the His 390 of bumblebee Bl-PGRP-LC completely abolished binding to the Dap-type PGN ( Figure 3B). Therefore, we hypothesize that the residue His 390 is the main anchor residue of Bl-PGRP-LC for the recognition of the Dap-type PGN.

Bombus PGRP-LC Strongly Bound to Dap-Type PGN and Exhibited Unique Anchor Residues in Comparison with Drosophila PGRP-LCx
The bumblebee Bl-PGRP-LC is upregulated after infection with S. aureus and E. coli, so the bumblebee Bl-PGRP-LC protein should have the ability to bind to the Lys-type PGN and the Dap-type PGN. However, the extracellular region of Bl-PGRP-LC is able to bind to the Dap-type PGN rather than the Lys-type PGN and shows a high affinity for the Dap-type PGN ( Figure 2A). As a control, the known Dm-PGRP-LCx exhibited no binding to the Lys-type PGN or the Dap-type PGN ( Figure 2B). Interestingly, the honeybee (A. mellifera) Am-PGRP-LC exhibited a weak affinity for the Dap-type PGN and no binding to the Lys-type PGN, but the short PGRP-LC had no ability to bind to the Dap-type PGN or the Lys-type PGN ( Figure 2C). To eliminate the influence of the Glutathione S-transferase (GST) tag on the binding of PGNs, the GST protein was purified and the binding test was performed on the PGNs. The GST protein did not bind to the Dap-type PGN or the Lys-type PGN ( Figure 2D). Based on the contact residues of Drosophila PGRP-LCx for the binding of TCT from the PGRP-LCx/PGRP-LCa/TCT complex, we mutated the 14 corresponding residues of Bl-PGRP-LC, namely, His 282 , Thr 283 , His 305 , Arg 309 , Ser 312 , Tyr 316 , Arg 330 , His 338 , Phe 340 , Asn 343 , His 390 , Ser 394 , Arg 395 and Leu 397 . All these residues were mutated to Ala. None of these mutations showed any effect on the recognition of Lys-type PGN by Bl-PGRP-LC ( Figure 3). In addition, the His 282 , Thr 283 , His 305 , Arg 309 , Ser 312 , Tyr 316 , Arg 330 , His 338 , Phe 340 , Asn 343 , Ser 394 , Arg 395 and Leu 397 mutations did not alter the binding to Dap-type PGN ( Figure 3A). Remarkably, the mutation of the His 390 of bumblebee Bl-PGRP-LC completely abolished binding to the Dap-type PGN ( Figure 3B). Therefore, we hypothesize that the residue His 390 is the main anchor residue of Bl-PGRP-LC for the recognition of the Dap-type PGN.

Structural Insight into PGRP-LC from Different Bee Species
The overall structure of Bl-PGRP-LC retains the classical PGRP structure and overlaps with Drosophila PGRP-LCx. The residues His 365 , Thr 366 , His 388 , Asp 395 , Tyr 399 , Arg 413 , Gly 424 , Asn 426 , His 473 , Ser 477 and Glu 480 of Drosophila Dm-PGRP-LCx respond to the binding to the Dap-type PGN. In comparison, the Bl-PGRP-LC/TCT complex structure shows that the residues Arg 309 , Phe 340 , His 390 (His 473 in Dm-PGRP-LCx) and Ser 394 (Ser 477 in Dm-PGRP-LCx) interact with TCT and stabilize the binding of TCT ( Figure 4A). We found that His 390 is the anchor amino acid for Bombus Bl-PGRP-LC, which is different from the anchor amino acid Arg 413 in Drosophila Dm-PGPR-LCx ( Figure 3B). Moreover, the residue His 390 forms a hydrogen bond with the MurNAc of TCT rather than the meso-Dap of TCT, which interacts with the corresponding anchor residue Arg 413 in Dm-PGRP-LCx. Therefore, we hypothesize that the residues that stabilize the interaction between Bl-PGRP-LC and the Dap-type PGN are not conserved in Drosophila Dm-PGRP-LCx, especially the anchor residue, and that the binding to the Dap-type PGN might be different. Moreover, the extracellular regions of PGRP-LCs of different bee species, including Apis cerana, A. mellifera, A. dorsata, A. florea, Bombus lantschouensis, Dufourea novaeangliae, Melipona quadrifasciata, Osmia bicornis and Megachile rotundata, share a high sequence identity. The anchor amino acid His 390 and the other contact residues Arg 309 , Phe 340 and Ser 394 are completely conserved among the different bee species ( Figure 4B). Hence, we hypothesize that the structural characteristics of PGRP-LC for Dap-type PGN binding are extremely similar among bees.

Discussion
The bumblebee has fewer genes for innate immunity, which is the unique line of defense against invading pathogens, than the dipteran insects D. melanogaster and A. gambiae. These few immune genes are involved in the recognition, signaling and effector functions, such as genes encoding PGRPs, which are important PRRs and can recognize PGNs to activate important immune pathways, the Toll and Imd pathways [2,6,31,32,34]. In contrast to the corresponding Drosophila protein, our recent study proved that bumblebee PGRP-SA preferentially binds to the Dap-type PGN rather than the Lys-PGN, and that the residue that responds to the binding to the Dap-type PGN is conserved in Apidae [35]. Our subsequent study proved that the bumblebee B. lantschouensis PGRP-LC exists as a single isoform, notably different from the three isoforms observed in Drosophila [37]. In this study, B. lantschouensis PGRP-LC was shown to be able to respond to the Gram-negative bacterium E. coli, and Bl-PGRP-LC was able to bind to the DAP-type PGN directly. E. coli infection triggers a quicker and stronger upregulation of genes, including PGRP-LC and the antimicrobial peptide genes abaecin and defensin, than an infection with S. aureus. PGRP-LC expression showed no change at 6 h, and the reaction of PGRP-LC peaked at 12 h after E. coli infection. Although the upregulation level of PGRP-LC was high at 24 h after S. aureus infection, Bl-PGRP-LC could not bind to the Lys-type PGN, so this upregulation might be mediated by another unknown mechanism rather than by the recognition of the Lys-type PGN by PGRP-LC. Therefore, we hypothesize that the cell wall of Gram-negative bacteria first needs to be degraded to be recognized by PGRP-LC after infection in bumblebees. These results also confirm that PGRP-LC recognizes Dap-type PGNs that constitute the bacterial cell wall [6,9]. Moreover, bumblebee PGRP-LC preferentially responded to Gram-negative bacteria through the recognition of DAP-type PGNs, where PGRP-LC expression in infected bumblebees showed a parabolic trend between 6 h and 24 h and reached a maximum at 12 h.
Moreover, bumblebee Bl-PGRP-LC could recognize Dap-type PGNs with a very high affinity, and this result was consistent with the observation that bumblebee PGRP-LC exhibited an immune response to Gram-negative bacteria. However, Drosophila PGRP-LCx has been reported to recognize Dap-type PGNs with the cooperation of PGRP-LCa, and the Drosophila PGRP-LCa/PGRP-LCx/TCT complex determines the contact residues contributing to the interactions between the TCT and PGRP-LCx, especially the anchor residue Arg 413 , which interacts with meso-Dap [17]. Nevertheless, the mutation of the 12 corresponding residues did not affect the binding to the Dap-type PGN, except for the His 390 Ala mutation, which completely abolished the binding to the Dap-type PGN in Bl-PGRP-LC (Figure 3). The corresponding residue His 473 in Drosophila PGRP-LCx only participated in the interactions; hence, we proved that Bl-PGRP-LC adopts the distinct anchor residue His 390 to form hydrogen bonds with MurNAc rather than meso-Dap to stabilize the interaction with the Dap-type PGN. The Bl-PGRP-LC/TCT complex showed that the help of residues Arg 309 , Phe 340 and Ser 394 is required for stabilization. For now, we confirm that bumblebee PGRP-LC exhibits specific binding characteristics for the Dap-type PGN, and that these contact residues are conserved among different bee species. The results indicate that although bumblebees possess relatively few PGRPs, these limited PGRPs have evolved a precise function and structure for the recognition of diverse pathogens.
The honeybee A. mellifera is a more eusocial insect than the bumblebee B. lantschouensis in Apidae [38]. A. mellifera possesses a special form of immunity, namely, "social immunity"; however, B. lantschouensis inhabits much smaller colonies than A. mellifera, with fewer social divisions, which might greatly diminish the defense provided by "social immunity" [39,40]. Therefore, the higher affinity of Bl-PGRP-LC for the Dap-type PGN than that of Am-PGRP-LC might reflect a stronger individual antibacterial immunity in bumblebees than in honeybees. Moreover, bumblebee PGRP-SA also showed a higher affinity for the Dap-type PGN than that of honeybee PGRP-SA in a previous study [35]. However, more studies on the mechanism of activation of the Toll and Imd pathways by PGRPs are needed, and would help explain the specific immune characteristics of bees and the interesting differences in the immune responses of different bee species.
In conclusion, bumblebee PGRP-LC could respond to an infection by Gram-negative bacteria and act as an effective receptor for binding to the Dap-type PGN. Moreover, the recognition of the Dap-type PGN is controlled by the specific anchor residue His 390 in the immune response to bacterial infection. In addition, bumblebee Bl-PGRP-LC exhibits a stronger binding to the Dap-type PGN than honeybee Am-PGRP-LC.

Bacterial Infection and Real-Time Quantitative PCR Verification
Worker bees at 10 days of age with individual weights between 0.19 and 0.21 g were selected from colonies of the Asian bumblebee B. lantschouensis and then reared in separate hives in the same surroundings and fed pollen and sugars to remove unhealthy bees over the next three days. In total, 30 healthy workers (15 workers per group) were starved for 12 h in preparation for feeding with the bacteria Staphylococcus aureus and Escherichia coli. Finally, 40 µL of bacterial culture (OD value = 1.0) mixed with sugar water at 1:1 was fed to the bees, and the time of feeding for each bee was recorded. Fifteen healthy workers were used as a control, and were fed with 40 µL sugar water. After 6 h, 12 h or 24 h, workers were directly frozen in liquid nitrogen and stored at −80 • C until RNA extraction. There were 5 workers per timepoint in each treatment and control group. The total RNA of each worker from each treatment and timepoint was extracted by using TRIzol (Thermo Fisher Scientific, Waltham, MA, USA), and cDNA was synthesized with 5× All-In-One RT MasterMix with the AccuRT Genomic DNA Removal reactions (Applied Biological Materials, British Columbia, Canada). All the samples for relative quantitative real-time PCR (qRT-PCR) samples were run on an ABI 7500 system (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA) with Bestar SybrGreen qPCR Mastermix (DBI ® Bioscience, Ludwigshafen, Germany). Each reaction was performed in a total volume of 20 µL that contained 5 µL of first-strand cDNA as a template and processed with a preincubation program of 5 min at 95 • C, followed by 40 cycles of 15 s at 95 • C and 45 s at 60 • C and a final step for 10 min at 72 • C. Gene-specific primers were designed with Primer Premier 5 (PREMIER Biosoft, San Francisco, CA, USA) ( Table 1). The bumblebee α-actin gene was used as an internal reference for gene expression. Gene expression levels were calculated using the 2 −∆∆Ct method [41]. The credibility of 2 −∆∆Ct values was evaluated by the p-value from t-test statistics. The mean threshold cycle values for each gene were obtained from three independent PCRs of the different individuals.

Preparation of Proteins
The wild-type B. lantschouensis PGRP-LC protein (Bl-PGRP-LC) was expressed and purified as described in a recent report from our laboratory [37]. Drosophila PGRP-LCx (Dm-PGRP-LCx) was synthesized by GENEWIZ (Suzhou, China) based on the sequence deposited in GenBank (accession number: 2f2l), A. mellifera PGRP-LC and all Bl-PGRP-LC mutants were obtained via PCR with primers designed by Primer Premier 5, and the three segments were ligated to the PGEX-6p-1 vector (previously stored in our laboratory). Then, these plasmids were transformed into the E. coli strain BL21 (DE3) for expression and purification following the protocols for the wild-type PGRP-LC. Expression from these plasmids was induced with 0.5 mM IPTG. The bacteria were harvested by centrifugation at 8000× g for 5 min and were then resuspended in cold phosphate-buffered saline (PBS). After sonication, the samples were centrifuged at 11,000 ×g, and the supernatants were filtered and loaded onto a GST-tag column or a His-tag column. Finally, the collected proteins were further purified by chromatography on a Superdex 200 16/60 column (GE Healthcare, Boston, MA, USA).

Lys-Type and Dap-Type PGN-Binding Assays
This assay was performed as previously described with minor modifications: the experiment was conducted at 4 • C by incubating 30 µg of purified wild-type Bl-PGRP-LC or mutant Bl-PGRP-LC with 300 µg of insoluble PGN in 300 µL of binding buffer containing 20 mM Tris-HCl (pH 8.0) and 50 mM NaCl on a shaking platform for 1.5 h with Dm-PGRP-LCx as a control [17]. Lys-type and Dap-type PGNs from Micrococcus luteus and Bacillus subtilis were provided by Sigma-Aldrich (Merck KGaA, Darmstadt, Germany). After centrifuging the incubation mixture at 16,000× g for 5 min, the bound protein was retained in the insoluble PGN pellet and the remainder of the protein was left in the supernatant. Then, 30 µL of the supernatant was used for SDS-PAGE analysis and the PGN-free Bl-PGRP-LC was visualized by staining with Coomassie blue.

Amino Acid Sequence Alignment and Phylogenetic and Structural Analyses
Alignment of the amino acid sequences of the PGRP-LC proteins from different bee species was performed by using the Clustal Omega server (https://www.ebi.ac.uk/Tools/msa/clustalo/) (Conway Institute UCD, Dublin, Ireland) and Jalview Desktop (http://www.jalview.org/) (University of Dundee, Dundee, UK). The structure of Bl-PGRP-LC was modeled by SWISS-MODEL (https://swissmodel. expasy.org/) (Swiss Institute of Bioinformatics Biozentrum, University of Basel Klingelbergstrasse, Basel, Switzerland) and evaluated based on the GMEQ and QMEAN values [42][43][44][45] with the template (PDB ID: 2f2l), and then, the docking of the Bl-PGRP-SA/TCT complexes was completed by the AutoDock4.2 program (The Scripps Research Institute, San Diego, CA, USA) [46]. The structural analyses were completed by the PyMOL 1.8.6 (DeLano Scientific LLC) (Schrödinger, New York, NY, USA) and CCP4 programs (Research Complex at Harwell (RCaH) STFC Rutherford Appleton Laboratory Harwell Science and Innovation Campus Didcot Oxon, Oxford, UK). The funders had no role in study design, data collection and analysis, the decision to publish, or the preparation of the manuscript.