Requirement of the antimicrobial peptide CRAMP for macrophages to eliminate phagocytosed E. coli through an autophagy pathway

Host-derived antimicrobial peptides play an important role in the defense against extracellular bacterial infections. However, the capacity of antimicrobial peptides derived from macrophages as potential antibacterial effectors against intracellular pathogens remains unknown. In this study, we report that normal (wild type, WT) mouse macrophages increased their expression of the cathelicidin-related antimicrobial peptide (CRAMP) after infection by viable E. coli or stimulation with inactivated E. coli and its product LPS, a process involving activation of NF-κB followed by protease-dependent conversion of CRAMP from an inactive precursor to an active form. The active CRAMP was required by WT macrophages to eliminate phagocytosed E. coli, with participation of autophagy-related proteins ATG5, LC3-II, and LAMP-1 as well as conjugation of the bacteria with p62. The autophagy-mediated elimination of E. coli was impaired in CRAMP−/− macrophages resulting in retention of intracellular bacteria and fragmentation of macrophages. These results indicate CRAMP as a critical component in autophagy-mediated clearance of intracellular E. coli by macrophages.


Introduction
Macrophages comprise an essential part of the innate immune system in response to bacterial infections (Rosenberger and Finlay, 2003). Because macrophages are highly phagocytic and are readily confronted by pathogenic bacteria, they must be equipped with effective mechanisms either for killing bacteria or controlling their replication to avoid becoming a reservoir of infection. For example, colon macrophages residing in the subepithelial lamina propria (LP) represent the first line defense against invading pathogens hence act as crucial sentinels for the  (Rekha et al., 2015). Bacteria initiate autophagy in macrophages mainly via their pathogen-associated molecular patterns (PAMPs) and damage-associated molecular patterns (DAMPs). Cell surface recognition and cytosolic sensing of invading pathogens by these molecules result in signaling cascades that promote rapid and localized autophagy machinery assembly. For instance, as a cytosolic sensor in macrophages, c-GAS recognizes bacterial DNA to trigger autophagy activation, resulting in ubiquitination of the bacterium or its phagosome by ubiquitin ligases Parkin and Smurf1. Ubiquitin chains subsequently bind to autophagy adaptors, such as p62 and NDP52 that recruit LC3 to deliver bacteria into an autophagosome. In addition, damaged phagosome is also targeted by autophagy via the recognition of host glycan present in the phagosomal lumen through cytosolic lectins of the galectin family. The process is tightly regulated by more than 30 autophagy-related gene products (ATGs). Upon autophagy activation, ATGs, serine/threonine kinase ULK1, and Beclin-1, in association with Atg14 and type III phosphatidylinositol 3-kinase (PI3K) Vps34, promote the formation of a cup-shaped isolation membrane to engulf the cargo to form a double-membrane autophagosome, which then fuses with lysosomes to form an autolysosome in which the engulfed cargo is degraded (Klionsky, 2010). However, the role of autophagy process in macrophage elimination of phagocytosed E. coli is unclear.
LL-37 in human and its mouse orthologue CRAMP are cathelicidin-related antimicrobial peptides, which belong to a family of host-derived antibacterial polypeptides ( In this study, we demonstrate that CRAMP is involved in the elimination of phagocytosed E. coli by mouse macrophages as shown by the retention of phagocytosed E. coli of macrophages deficient in CRAMP. We further provide evidence that CRAMP deficiency results in reduced expression of autophagy-related molecules ATG5, LC3-II, LAMP-1 and p62 and impaired degradation of E. coli conjugated with p62 in macrophages.

Stimulated of CRAMP production in macrophages by E. coli products
To obtain evidence for the importance of CRAMP for macrophages to eliminate phagocytosed E. coli, we generated macrophages from bone marrow (BM) cells of control CRAMP +/+ mice (LysMCre -CRAMP F/F ). After infection with E. coli isolated from the feces of naïve mice, the production of CRAMP by macrophages progressively increased and reached the maximal level by 20 h (Fig. 1A, B). Inactivated E. coli also stimulated control macrophages to produce CRAMP as confirmed by Western blotting (Fig. 1C).
We further observed revealed that stimulation of control macrophages by inactivated E. coli induced rapid phosphorylation of IB, shown by an increase in total IB due to de novo synthesis ( Fig. 1F) (Karin, 1999). In contrast to control macrophages, there was a significantly diminished phosphorylation of IκB-α in CRAMP -/macrophages from Myeloid CRAMP -/mice ( Fig. 1F). The CRAMP production by control macrophages in response to inactivated E. coli was attenuated by a selective IκB-α inhibitor BAY117082 (Fig. 1G). Thus, activation of NF-κB is critical for macrophages to produce CRAMP in response to stimulation by E. coli and its product LPS.

Requirement of CRAMP for macrophages to eliminate phagocytosed E. coli
To examine the role of CRAMP in macrophage elimination of phagocytosed E. coli, a mouse RAW 264.7 cell line, used as an in vitro model, was co-cultured with inactivated E. coli for 20 h. RAW 264.7 cells expressed high level of CRAMP with few endocytosed inactivated E. coli ( Fig.   2A, Upper panel). Preincubation of RAW 264.7 cells with BAY117082 reduced the production of CRAMP with increased phagocytosed E. coli within the cells ( Fig. 2A, Lower panel). The 7 bactericidal activity of CRAMP was also shown by a synthetic peptide that directly killed E. coli in vitro (Fig. 2B).
CRAMP is normally stored in lysosomes of macrophages as an inactive precursor, which is converted to an active form through cleavage by proteases (Shinnar et al., 2003;Zanetti, 2004) such as intracellular elastase-like serine protease (Rosenberger et al., 2004). We found that Elastatinal, an elastase inhibitor, attenuated the capacity of macrophages to eliminate phagocytosed E. coli (Fig. 2C). Therefore, CRAMP production and conversion are critical for macrophages to eliminate both phagocytosed and extracellular E. coli.

(K, I moved the source of CARMP + macrophages to preceding paragraphs)
As shown in These results indicate that CRAMP was required for macrophages to timely eliminate phagocytosed E. coli.

Involvement of autophagy pathway in CRAMP-mediated elimination of phagocytosed E. coli by macrophages
We then tested whether lysosomal hydrolases in macrophages are required for autophagic elimination of inactivated E. coli. Treatment of CRAMP +/+ control macrophages or RAW264.7 8 cells with E64d, an inhibitor of cathepsins B and L, or pepstatin A, an inhibitor of cathepsin D, that suppress autolysosomal digestion, protected E. coli from autophagic elimination by the cells (Fig. 4A-C). Thus, lysosomal proteases are important for autophagic degradation of inactivated E. coli by macrophages, indicating the dependence on autophagy.
We further found that there was a reduced expression of the autophagy-related protein ATG5, which is involved in the extension of the phagophoric membrane in autophagic vesicles Participation of CRAMP in the autophagy pathway in macrophages for E. coli elimination was further demonstrated by reduced fluorescence intensity of LC3B + and LAMP-1 + (Lysosomal associated membrane protein 1) and increased fluorescence intensity of p62 + in CRAMP -/macrophages as compared to CRAMP +/+ counterparts after stimulation with inactivated E. coli for 12 h (Fig. 5C-E). The bacteria showed reduced colocalization with LAMP-1 (Fig. 5D), but increased colocalization with p62 ( Fig. 5E) in CRAMP -/macrophages, indicating that CRAMP deficiency impaired degradation of bacteria conjugated with p62, resulting in retention of E. coli in the cells.

Discussion
In this study, we elucidated previously uncharacterized macrophage effector mechanisms for elimination of phagocytosed E. coli. Viable E. coli infection or inactivated E. coli stimulation of mouse macrophages increase intracellular production and extracellular release of CRAMP by activation of NF-κB to trigger autophagy-dependent degradation of the bacteria (as summarized in Fig. 6). Interestingly, although both LPS and the chemotactic peptide fMLF are the products of E. coli, only LPS is able to up-regulate CRAMP expression in macrophages, indicating that Our current study showed that CRAMP deficiency is associated with reduced expression of autophagy-related proteins ATG5, LC3-II, LAMP-1 in macrophages after phagocytosis of E.
coli. However, the changes in p62 are different. p62 (A170 or SQSTM1) is an accessory autophagy-targeting molecule that directs cytosolic proteins to autophagosomes, a process critical for the elimination of intracellular bacteria. p62 delivers specific cytosolic components, including ribosomal protein S30 (rpS30) and additional ubiquitinated proteins, to autophagic organelles to be processed into bactericidal products. In the absence of p62, the cells are unable to generate neo-antibacterial factors, resulting in non-functional autophagy despite maturation, thereby failing to effectively eliminate intracellular bacteria (Ponpuak et al., 2010). The degradation of p62 is a widely used marker to monitor autophagic activity because p62 binds to  CO2 overnight. The non-adherent cells were collected, centrifuged and re-cultured in tissue culture dishes (1×10 6 cells/ml) with addition of DMEM with 50 ng/ml M-CSF for 3 days. The medium was replaced on day 7 and fully differentiated macrophages were harvested. CRAMP +/+ macrophages were generated from BM of control (LysMCre -CRAMP F/F ) mice (referred to as control cells) and CRAMP -/macrophages were generated from BM of Myeloid cell-specific

Isolation of fecal E. coli
The fecal E. coli isolated from naïve mice was aerobically cultured on Violet red bile lactose (VRBL) agar for 24 h. Single bacterial colonies were identified as E. coli O22H8 by complete genome sequencing. E. coli identified was cultured in LB agar at 37 o C, 180 RMP for 24 h, then determined for concentrations at OD600nm = 0.4 corresponding to ~2 × 10 8 colony forming unit 13 (CFU)/ml). E. coli suspension was aliquoted in 1 ml volumes and stored at -80 o C for future use. When necessary, live or heat-inactivated E. coli was labeled with FITC (Isomer I, Sigma) following the manufacturer recommended procedures.
After culture at 37 o C in 5% CO2, cell supernatant was harvested at different time points to measure CRAMP concentration with ELISA using a Mouse CRAMP ELISA Kit (MyBioSource, CA).

In vitro killing of E. coli by CRAMP
E. coli was diluted at 5 × 10 4 in 100 µl/well on 96-well plates followed by culture with various concentrations (0.01 -100 µg/ml) of synthetic murine CRAMP (Hycult Biotech) at 37°C for 2 h.
The bacterial suspension was then serially diluted with PBS and plated on nutrient agar plates at 37°C for 24 h. The number of E. coli treated with CRAMP was quantitated and expressed as the percentage of the number of untreated bacteria as a control.

Western immunoblotting
BM-derived CRAMP +/+ control and CRAMP -/macrophages or CT26 mouse colon epithelial cell line (ATCC) grown in 60-mm dishes to sub-confluency were cultured for 3 h in FCS-free media.
After treatment with inactivated E. coli, the cells were lysed with 1× SDS sample buffer (62.5 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, and 50 mM dithiothreitol), then sonicated for 15 s and heated at 100 o C for 5 min. Cell lysate was centrifuged at 12,000 rpm (4 o C) for 5 min, and protein concentrations of the supernatants were measured by DC Protein Assay (Bio-Rad).
The lysates with titrated proteins were electrophoresed on 10% SDS-PAGE precast gels (Invitrogen) then transferred onto ImmunoBlot polyvinylidene membranes (Bio-Rad), which were blocked with 5% nonfat milk. Phosphorylated IκB-α were detected using phospho-specific Abs according to the manufacturer's instructions. After incubation of the membranes with a horseradish peroxidase-conjugated secondary Ab, protein bands were detected with a Super Signal Chemiluminescent Substrate (Pierce) and the images were quantitated using a G-BOX GeneSnap system (SYNGENE). For detection of total IκB-α, β-actin, ATG-5, LC3B, p62 and CRAMP, the membranes were stripped with Restore Western Blot Stripping Buffer (Pierce) followed by incubation with specific Abs.

Statistics
All experiments were performed at least three times with three replicate samples. Statistical analysis was performed using GraphPad Prism by two-tailed Student's t test or 1-way ANOVA with Kruskal-Wallis Test. Data with error bars represent mean ± SEM and P values less than 0.05 (P < 0.05) were considered statistically significant.