Two Novel Short Peptidoglycan Recognition Proteins (PGRPs) From the Deep Sea Vesicomyidae Clam Archivesica packardana: Identification, Recombinant Expression and Bioactivity

Vesicomyidae clams are common species living in cold seeps, which incorporates symbiotic bacteria into their body maintaining endosymbiosis relationship. As members of pattern recognition receptor (PRR) family, peptidoglycan recognition proteins (PGRPs) recognize pathogen associated molecular patterns and play an important role in innate immunity. In present study, two short PGRPs (ApPGRP-1 and -2) were first identified from Vesicomyidae clam Archivesica packardana. Sequences analysis showed that they have both conserved Zn2+ binding sites (H-H-C) and amidase catalytic sites (H-Y-H-T-C), and phylogenetic tree indicated that they clustered with short PGRPs of other molluscs. PGN assay showed that ApPGRPs could bind Lys-type PGN from Staphylococcus aureus and Dap-type PGN from Bacillus subtilis, and revealed amidase activity with selective zinc ion dependence. rApPGRP-1 and -2 (recombinant ApPGRP-1 and -2) could bind six bacteria with a broad spectrum and had both zinc-dependent and -independent bactericidal activity. ApPGRPs had the complete functions of effectors and partial functions of receptors from PGRPs. Further analyses showed that ApPGRPs from A. packardana might be involved in the endosymbiosis relationship between the host clam and endosymbiotic bacteria as a regulator. The results of these experiments suggested that ApPGRPs were involved in cold seep clams’ immune response. This study provides basic information for further research on the immune mechanisms of deep sea organisms.


INTRODUCTION
The immune system contains innate immunity and adaptive immunity (Janeway, 1989). Innate immune system is the frontline of defense and almost the only defense mechanism for invertebrates to protect the host from invasion by microbes in the surrounding environment (Janeway, 1989). Innate immunity can be triggered by a set of specific receptors termed PRRs. PRRs could recognize the conserved, invariant components on the cell surface of microbes named PAMPs (Yang et al., 2013). PRRs include Toll-like receptors, PGRP, scavenger receptor (SRCR), thioestercontaining proteins (TEP), lipopolysaccharide and beta-1,3glucan binding protein (LGBP), and lectins (Janeway and Medzhitov, 2002;Hoffmann, 2003).
The Bathymodiolus mussel and Vesicomyidae clam are common species in hydrothermal vents and cold seeps (Barros et al., 2015). They have methane-oxidizing (MOX) bacteria or sulfur-oxidizing (SOX) bacteria in their gills as symbionts (Martins et al., 2014). Expression level of BaPGRP was upregulated at 12 h but down-regulated at 24 h in the hydrothermal vents mussel Bathymodiolus azoricus that had been challenged with live V. alginolyticus (Martins et al., 2014). PGRPs from the cold seep mussel B. platifrons had different binding modes to peptidoglycan (PGN) from Gram-negative and Gram-positive microorganisms (Yue et al., 2015). These results suggested that pattern recognition receptors and related molecules are involved in the immune response of deep sea vent/seep organisms, but little information about Vesicomyidae clam is available.
The Vesicomyidae clam A. packardana, previously named Calyptogena packardana, which has been collected from cold seeps (Johnson et al., 2016). In present study, we study the roles of PGRP in immune system of this species. The primary objectives of the present research are: (1) identification and sequence analysis of ApPGRP-1 and -2 molecules; (2) combination and degradation of PGN by rApPGRP-1 and -2; (3) combination and inhibition of microbes by rApPGRP-1 and -2. The roles of PGRPs in the endosymbiosis relationship between the clam and endosymbiotic bacteria have also been discussed.

Sample Collection
Clams were collected from the Malibu Mound (33.902,.735) at a depth of 520 m during a MBARI expedition in 2014. Once clams were brought to the deck, the adductor muscle and gill were immediately removed and stored in RNAlater (Ambion, Austin, TX, United States). Total RNA was extracted and quality was assessed using gel electrophoresis and a spectrophotometer (Nanodrop2000, Thermo Scientific TM NanoDrop TM , United States). Then, the RNA samples were sent out for sequencing at SeqMatic LLC (Fremont, CA, United States) on a HiSeq TM 2000 platform. Over 6 Gbp clean data were obtained for each tissue library.

Gene Identification and Expression Vector Construction
Homologs of the peptidoglycan recognition protein (PGRP) gene were found through searching the A. packardana transcriptome (data not yet published) using TBLASTN 1 with previously published PGRP genes as a query. Two new PGRPs (ApPGRP-1 and -2) were identified.

Sequence Analyses
Sequence comparison was conducted in the BLAST program 1 . We calculated theoretical isoelectric point and molecular weight in the ProtParam program 2 . SignalIP4.1 3 was used to predict the signal peptide, and the SMART program (Letunic and Bork, 2018) 4 was used to predict functional domains. Multiple protein sequences were aligned by using Clustal W program (Larkin et al., 2007) (version1.83 5 ). A neighbor-joining (NJ) tree was established based on the deduced amino acid sequences with 1,000 bootstrap replicates by using MEGA v5.0 software (Tamura et al., 2011) 6 . The predicted tertiary structures were constructed in the SWISS-MODEL program 7 and checked in Deepview/Swiss-Pdb Viewer 4.0 8 (Guex and Peitsch, 1997).

Expression, Purification and Western Blotting of Recombinant Proteins
Verified transformants were cultured in LB medium (yeast extract, 5 g, tryptone, 10 g, sodium chloride, 10 g of 1 L) with 100 µg/ml ampicillin, 20 µg/ml chloramphenicol, 0.5 mg/ml Larabinose. The culture temperature was set at 37 • C with shaking at 200 rpm. Two hours later, tetracycline was added into the LB medium to a final concentration of 2 ng/ml. L-arabinose and tetracycline are inducers of chaperone proteins dnaK-dnaJ-grpE and groES-groEL, respectively. Chloramphenicol is the resistance marker of plasmid pG-KJE8. When optical density at 600 nm (OD 600 ) reached 0.5, the medium was cooled to 15 • C and incubated for more than 30 min. After that, Isopropyl-hd-thiogalactoside (IPTG) was added and the final concentration was 0.1 mM. The medium was cultured for another 24 h at 15 • C. Then, bacteria were precipitated at 8,000 g for 5 min at 4 • C. The recombinant proteins pCOLD II-ApPGRP-1 and -2 were present in the supernatant after sonication. Proteins were purified with Ni-NTA-Sefinose Column (Sangon Biotech, Shanghai, China), and eluted with 300 mM imidazole under non-denaturing conditions. The obtained proteins were dialysed, concentrated, and then stored at −80 • C before use. The protein samples (before induction, after induction, the supernatant and precipitate after sonication) were separated in 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE). Western blotting procedures were set as follows: proteins were transferred from gels to PVDF membranes (Immobilon-membrane, Millipore, MA, United States) by Trans-Blot SD Semi-Dry Electrophoretic Transfer Cell (DYCP-40C, Beijingliuyi, Beijing, China). After that, the primary anti-6 × His antibody (diluted 1: 5000 with 5% skim milk; ab18184, abcam, Cambridge, United Kingdom) was incubated overnight with the membranes. The membrane was then washed with TBST (50 mM Tris-HCl, 50 mM NaCl, 0.05% Tween20, pH 7.2) three times. After that, secondary antibody (diluted 1: 10000; ab6789, abcam, Cambridge, United Kingdom) was incubated with the PVDF membranes for 2 h, followed by three washes with TBST. The membranes were incubated for 5 min with Pierce ECL Western Blotting Substrate (Thermo Scientific, MA, United States), and the Chemiluminescence imaging system (ChemStudio, Analytikjena, Jena, Germany) was used to detect chemical signals.

Binding Analysis of rApPGRP-1, -2 to PGN
The assay was done as previous study (Yang et al., 2013) with slight modifications. We first incubated 40 µg of rApPGRP-1 or -2 proteins in 200 µl TBS buffer (50 mM Tris-HCl, 50 mM NaCl, pH 7.2) with L-PGN from Staphylococcus. aureus (Catalog No. 77140, Sigma-Aldrich, MA, United States; 100 µg, 1 mg/ml) and D-PGN from Bacillus subtilis (Sigma-Aldrich, MA, United States; 100 µg, 1 mg/ml). Then, bound and unbound proteins were obtained by centrifugation at 13,000 rpm for 15 min after incubation at 4 • C for 3 h. TBS was used to wash the pellets (bound fraction) three times. After that, 2 × SDS-PAGE loading buffer was used to separate bound proteins from PGN by boiling at 95 • C for 5 min. Samples were analyzed in 10% SDS-PAGE. Western blotting with anti-6 × His antibody was used to detect the target proteins as above.
Binding Analysis of rApPGRP-1, -2 to Microbial Cells The assay was done as described by Feng et al. (2012) with slight modifications. S. aureus, M. luteus, B. subtilis, E. coli, Pichia pastoris, and Saccharomyces cerevisiae colonies were grown in culture medium (LB medium for Gram-positive bacteria, Gramnegative bacteria; YPD medium for fungi). LB medium was the same as above, and YPD medium (1 L) includes: 10 g yeast extract, 20 g peptone, 20 g dextrose. When the OD 600 was close to 0.8 (ca. 1.6 × 10 8 cells/ml), 4 ml of each medium was centrifuged at 5,000 g and the pellets were washed twice with TBS buffer. Microbial cells were re-suspended in 50 µl of TBS buffer and then mixed with 40 µg of rApPGRP-1 or -2 dissolved in 200 µl of TBS buffer. The mixture was incubated for 3 h at 4 • C, and centrifuged at 12,000 g at 4 • C for 15 min. The cell pellets were washed three times with TBS buffer and then suspended in 50 µl of 2 × SDS sample buffer. The samples were heated at 95 • C for 5 min to get the bound protein.
The bound proteins from six kinds of microorganisms were checked in 10% SDS-PAGE and detected using western blotting as above.

Analysis of Amidase Activities
Amidase activity of PGRP could cleave the lactylamide bond between muramic acid and L-alanine, and further cause the dissolution of the PGN (Foster, 2004). The assay was done as described by Mellroth et al. (2003), Yang et al. (2013) with slight modifications. 40 µg L, D-PGN (1 mg/ml) was incubated with 50 µg of rApPGRP-1 or -2 protein in TBS-ZnCl 2 solution (50 mM Tris-HCl, 50 mM NaCl, 100 µM ZnCl 2 , pH 7.2). 40 µg PGN (1 mg/ml) was mixed with TBS buffer which was set as a control. OD 540 was measured per 30 min during a 300 min period by using Varioskan LUX Multimode Microplate Reader (Thermo Fisher Scientific, MA, United States).

Analysis of Antimicrobial Activities
The assay was done as described by  with slight modification. S. aureus, M. luteus, B. subtilis, E. coli, P. pastoris, and S. cerevisiae were cultured to an OD 600 of 0.6 in LB medium at 37 • C or YPD medium at 28 • C (Fungi). 200 µl aliquots of the cultures were pelleted by centrifugation at 5,000 g for 5 min at 4 • C and washed with TBS buffer twice. The pellets were re-suspended in 200 µl TBS buffer. After a 1:50 dilution, 10 µl of the suspensions was mixed with 25 µl 200 µg/ml rApPGRP-1, or 200 µg/ml rApPGRP-2, or TBS (as control), with 10 or 100 µM ZnCl 2 existed. The mixtures were shaken on a Tube Tumbler for 6 h at 26 • C, and then 1 ml LB or YPD medium was added separately. The cultures were shaken at 37 • C or 28 • C overnight before absorbance measurement at 600 nm.

Statistical Analyses
Statistical analysis was performed in SPSS22.0 (IBM Company, NY, United States). Duncan test of one-way ANOVA was used among mean values from different groups. The P-value is 0.05.

Alignment and Phylogenetic Analysis of ApPGRP-1 and ApPGRP-2
Online BLAST analysis showed that amino acid sequences of ApPGRP-1 and -2 are highly homologous with PGRPs from other species. ApPGRP-1 is homologous with PGRPs from H. cumingii (AHK22786.1), H. discus discus (AHB30456.1), and the similarities are 58% and 53%, respectively. In addition, it also has a high homology with an Octopus bimaculoides protein (XP_014778613.1) with 56% identity. ApPGRP-2 has high similarity with PGRPs from C. gigas (XP_011422763.1) and Pinctada fucata (JAS03318.1), and the similarities are 56% and 52%, respectively. ApPGRP-1 and -2 have 44% similarity with each other. Alignment analysis of these two ApPGRPs and other animals' PGRPs showed that the C-portion of the PGRPs is highly FIGURE 1 | Nucleotide and deduced amino acid sequences of ApPGRP-1 (A) and ApPGRP-2 (B). The predicted signal peptide is underlined, and the predicted start and termination codons are circled. The amidase/PGRP domain predicted in SMART program is highlighted in gray, and the black boxes indicate Zn 2+ binding sites.
Phylogenetic analysis of ApPGRP-1, -2 and other species' PGRPs showed that ApPGRP-1 clustered with short PGRPs from abalone H. discus discus and freshwater pearl mussel H. cumingii, and ApPGRP-2 clustered with three short PGRPs from oyster C. gigas, and then these two branches were grouped together (Figure 3).

Binding of Recombinant ApPGRP-1 and -2 Proteins to Microbial Cells
Microbe binding assay was done to analyze whether rApPGRP-1 and -2 bound Gram-negative bacteria, Gram-positive bacteria and fungi. Clear bands were detected which suggested that rApPGRP-1 and -2 could bind to six microbes (Figure 7). The band intensities of E.c (lane 4, Figure 7) were weaker than other bands. No bands were observed for the negative control (data not shown). Seeing from these data, the ApPGRPs proteins could bind a wide spectrum of bacteria.

Antimicrobial Activities of rApPGRP-1 and -2
For Gram-positive bacteria, when zinc ions didn't exist, rApPGRP-2 had antibacterial activity against B. subtilis, but no antibacterial activity against S. aureus or M. luteus. In the participation of zinc ions, rApPGRP-1 + 100 µMZn group significantly inhibited S. aureus, which was significant different from the rApPGRP-1 or 100 µMZn groups (P < 0.01), indicating that rApPGRP-1 had strong enough amidase activity to achieve bactericidal activity in the presence of zinc ions. Similarly, rApPGRP-2 + 100 µMZn group showed analogous activity against B. subtilis and rApPGRP-1 + 100 µMZn group showed activity against M. luteus (Figure 9).
For E. coli, S. cerevisiae and P. pastoris, we chose 10 µM Zn 2+ for antibacterial experiments, as 100 µM Zn 2+ alone had a strong antibacterial effect on these microorganisms. For E. coli, rApPGRP-1 and -2 had clear antimicrobial activity in the absence of 10 µM Zn 2+ . In the presence of 10 µM Zn 2+ , 10 µMZn group had a bacteriostatic effect but not significant. As the antibacterial effect of rApPGRP-1 + 10 µMZn group is not more than rApPGRP-1 group and there is no significant difference among these three groups: rApPGRP-1, 10 µMZn and rApPGRP-1 + 10 µMZn group (P > 0.05), suggested that antibacterial ability of rApPGRP-1 may not require zinc ions. A similar situation was also observed in the rApPGRP-2 + 10 µMZn group (Figure 9).

DISCUSSION
In this study, two new short PGRPs (ApPGRP-1 and -2) were identified from A. packardana. Homology analysis indicated that ApPGRPs had relatively high similarity with PGRPs of other organisms, and phylogenetic tree analysis showed that ApPGRPs clustered with most PGRPs from molluscs with high bootstrap values. As mollusc PGRPs are commonly involved in a series of immune responses, rApPGRP-1 and -2 from A. packardana might also play a similar role in regulating diverse immune responses to adapt cold seep habitat.

PGN Binding Specificity and Amidase Activities of PGRPs
L-PGN or D-PGN can be specifically and preferentially recognized by PGRPs. In Drosophila, L-PGN from Gram-positive bacteria could trigger the toll signal pathway by PGRP-SA or Frontiers in Physiology | www.frontiersin.org PGRP-SD. Dap-type PGNs from Bacillus and Gram-negative bacteria could stimulate the IMD pathway (Michel et al., 2001;Hoffmann, 2003;Kaneko et al., 2004). In molluscs, rCfPGRPS1 from Chlamys farreri displays affinity to L-PGN from S. aureus (Yang et al., 2010). CgPGRP-S1S from pacific oyster displays specific binding activity to D-PGN, but not to L-PGN (Iizuka et al., 2014). In this study, PGN binding assays revealed that ApPGRPs have PGN-binding activity toward D, L-PGN which was also found in HcPGRPS1 from H. cumingi (Yang et al., 2013). It could be seen that PGRPs from different organisms have their specific PGN binding spectrums.
We also measured amidase activities of the PGRPs. In rPGRP-S from amphioxus, with the participation of Zn 2+ , higher hydrolyzing activity was detected when using Lys-PGN as substrate compared to Dap-type PGN (Yao et al., 2012). In the present study, rApPGRP-1 and -2 showed some but not significant degradation activity for L-PGN in the presence of Zn 2+ . However, rApPGRP-1 also showed somehow amidase activity toward D-PGN in the absence of Zn 2+ . These results are consistent with previous study . We therefore conclude that the amidase activity of these ApPGRPs is Zn 2+selective dependent. Microbial Binding Specificity and Antibacterial Activity rApPGRP-1 and -2 bound six bacteria with a broad spectrum. Recombinant ApPGRP-1 exhibited antibacterial activity that inhibited the growth of S. aureus and M. luteus in the presence of Zn 2+. This is consistent with the results from Drosophila (Mellroth and Steiner, 2006), C. farreri (Yang et al., 2010). On the other hand, rApPGRP-1 and-2 had obvious antimicrobial activity for E. coli and P. pastoris in the absence of Zn 2+ . This antimicrobial activity might similar to BmPGRP-S5 from Bombyx mori, which has obvious antibacterial activity toward Gram-positive bacteria M. luteus and S. aureus, Gramnegative bacteria S. marcescens, and E. coli . Similarly, amphioxus rPGRP-S can also suppress the growth of P. pastoris in the absence of Zn 2+ (Yao et al., 2012). In our study, rApPGRP-1 and -2 displayed amidase activity and bactericidal activity in Zn 2+ -dependent or -independent manner. These results indicated that the bactericidal effect of ApPGRPs might be a direct effect of amidase-mediated lysis of PGN as there are conserved sites (H-Y-H-T-C) for amidase activity in ApPGRP-1, -2.

Functions of ApPGRPs
Peptidoglycan recognition proteins mainly have three functions: recognition receptor, regulator, and effector (Figure 10; Chen and Lv, 2014). ApPGRPs bound PGN and six kinds of microorganisms, which indicate they have partial functions of PGRPs as receptors. After that, IMD, PO, or Toll pathway would be activated, and hosts produce diverse immune responses (Michel et al., 2001;Bischoff et al., 2004;Park et al., 2007;Pal and Wu, 2009). However, the roles of ApPGRPs in these pathways are still not well studied. From the result that ApPGRPs have amidase activity to degrade PGN and could kill microorganisms, we speculate that they can function as effector of PGRPs. To explore the possible roles of ApPGRPs as a regulator, further in vivo study is needed.

Trade Off Between Immune Response and Symbiotic Relationship Maintaining
The clam A. packardana in the study is from cold seep and has sulfur-oxidizing in its gills as a symbiont (Johnson et al., 2016). ApPGRPs can bind microbes, cleave PGN and display antibacterial activity, which indicated that PGRPs from A. packardana might play vital roles in regulating diverse immune responses. Organisms that incorporate symbiotic bacteria into their bodies must maintain a stable symbiotic relationship, therefore as a regulator, PGRPs might play important roles in this process. Given that the large amount symbionts are existed in cold seep clam gills (Johnson et al., 2016), ApPGRPs might function as a regulator (enhance or inhibit) in this tissue. To maintain a long-term symbiotic relationship, weevils from the genus Sitophilus express a PGRP which can decrease the biological activity of PGN from symbiotic bacteria (maybe), and therefore avoid PGN from stimulating the host to generate higher immune response (Anselme et al., 2006;Moya et al., 2008). In deep sea organisms, PGRPs' roles as regulator are not very clear. We just found that PGRPs from the Atlantic vent mussel B. azoricus might participate in the immune response at early point as they establish symbiotic relationship with bacteria (Martins et al., 2014). PGRPs from cold seep mussel B. platifrons might have different roles in different tissues, and BpPGRPs might recognize symbiont bacteria in the gill and function in the immune response in the visceral mass (Sun et al., 2017).

CONCLUSION
In conclusion, two short PGRPs from the cold seep clam A. packardana were identified and preliminarily analyzed. Functional assays showed that ApPGRPs have biological function and participate in the immune response. By comparison with PGRPs from other invertebrates, we hypothesize that ApPGRP might also be involved in the endosymbiosis relationship between the host and endosymbiotic bacteria as a regulator. Taken together, our study provides some basic information for further study on the immune and symbiotic mechanisms of vent/seep molluscs.

ETHICS STATEMENT
All applicable international, national, and/or institutional guidelines for the care and use of animals were followed.

AUTHOR CONTRIBUTIONS
XK and HZ conceived and designed the experiments, analyzed the data, and wrote the paper. XK, HL, YL, and HZ performed the experiments. All authors reviewed the manuscript.