Identification of ATP synthase α subunit as a new maternal factor capable of protecting zebrafish embryos from bacterial infection

Abstract

Previous studies have shown that ATP synthase α subunit (ATP5A1) plays multiple roles, but our understanding of its biologic functions remains poor and incomprehensive. Here, we clearly demonstrated that zebrafish ATP5A1 was a newly characterized lipoteichoic acid (LTA)- and LPS-binding protein abundantly stored in the eggs and embryos of zebrafish. Zebrafish ATP5A1 acted not only as a pattern recognition receptor, capable of identifying LTA and LPS, but also as an effector molecule, capable of inhibiting the growth of both gram-positive and -negative bacteria. ATP5A1 could disrupt the bacterial membranes by a combined action of membrane depolarization and permeabilization. We also found that the N-terminal 65 residues were critical for the antibacterial activity of zebrafish ATP5A1. In particular, we showed that microinjection of exogenous recombinant (r)ATP5A1 into early embryos could promote their resistance against pathogenic Aeromonas hydrophila challenge, and this pathogen-resistant activity was markedly reduced by the coinjection of anti-ATP5A1 antibody or by the knockdown with morpholino for atp5a1 but not by the coinjection of anti-actin antibody. Moreover, each egg/embryo contains a sufficient amount of ATP5A1 in vivo to kill A. hydrophila. Furthermore, the N-terminal 65 residues 1–65 of ATP5A1 α subunit (rA_1–65) with in vitro antibacterial activity also promoted the resistance of embryos against A. hydrophila, but the N-terminal 69 residues 66–134 (rA_66–134) or C-terminal residues 135–551 (rA_135–551) of ATP5A1 α subunit without in vitro antibacterial activity did not. Finally, we showed that the antibacterial activity of the N-terminal 65 residues of ATP5A1 α subunit was conserved throughout animal evolution. Collectively, these results indicate that ATP5A1 is a novel maternal immunocompetent factor that can protect the early embryos of zebrafish from bacterial infection. This work also provides a new viewpoint for understanding the biologic roles of ATP5A1, which is ubiquitously present in animals.—Ni, S., Zhou, Y., Chen, Y., Du, X., Zhang, S. Identification of ATP synthase α subunit as a new maternal factor capable of protecting zebrafish embryos from bacterial infection.

ATP synthase (EC3.6.3.14), also known as complex V, is a ubiquitous membrane enzyme that is found in bacteria, plants, and animals. In most organisms, the primary function of the enzyme is ATP generation from ADP and inorganic phosphate. The ATP synthase complex consists of 2 structural domains, F₁ and F₀, connected by 2 stalks. The catalytic domain F₁ executes the ATP synthesis and hydrolysis reactions, and the membrane-embedded domain F₀ mediates the proton transport (1–3). F₁ comprises 5 subunits with the stoichiometry α₃β₃γδε, whereas F₀ comprises ab₂c_10–15 subunits, in which c subunit number varies according to species (4). The ATP synthase α subunit (ATP5A1) was first identified in yeasts (5) and has now been documented in various species including shrimp, sea urchin, ascidian, fish, and mammals such as mouse, rat, and humans (6–11). In humans, the ATP5A1 is encoded by atp5a1 gene (1), which is mapped onto the chromosome 18q12-q21 (12). Human atp5a1 gene contains 12 exons that can form at least 2 splice variants that differ in the 3′ UTR (6).

Numerous studies have addressed the functions of the ATP5A1. Several lines of evidence have shown that the ATP5A1 plays some other important roles in the eukaryotic cell in addition to its involvement in the generation of ATP (13–15). For example, ATP5A1 was associated with the pathogenesis of glioblastoma (16), the progression of renal cell carcinoma (17), and the neurodegenerative process of Alzheimer’s disease (18, 19). ATP5A1 appeared to also be involved in the aging process because both decreased expression and carbonylation of the ATP5A1 were observed with age (7, 20). Moreover, the defects in atp5a1 gene were shown to be either lethal (21, 22) or cause fatal neonatal mitochondrial encephalopathy (23). Despite the enormous progress, we still lack a comprehensive understanding of the biologic functions of ATP5A1.

Lipoteichoic acid (LTA) is a surface-associated adhesion amphiphile from gram-positive bacteria, capable of inducing arthritis, nephritis, uveitis, encephalomyelitis, meningeal inflammation, and periodontal lesions and also triggering cascades resulting in septic shock and multiorgan failure. Thus, LTA can be considered a virulence factor that has an important role in infections and in postinfectious sequelae caused by gram-positive bacteria (24). Recently, we have isolated a protein from zebrafish (Danio rerio) embryos by LTA-conjugated Sepharose CL-4B affinity chromatography and identified it as ATP5A1. This suggests that ATP5A1 may play a role in immune responses against bacterial infection. However, we know nothing about it at present. Therefore, the aims of this study were to explore whether ATP5A1 has any antibacterial activity and, if so, to examine its mode of action using the zebrafish as model.

MATERIALS AND METHODS

Fish and embryos

All the fish used in the experiments were treated in accordance with the guidelines of the Laboratory Animal Administration Law of China (permit SD2007695), approved by the Ethics Committee of the Laboratory Animal Administration of Shandong province. The wild-type zebrafish D. rerio was purchased from a local fish dealer and maintained in the containers with well-aerated tap water at 27 ± 1°C. The fish were fed with live bloodworms and fish flakes (TetraMin; Tetra, Melle, Germany) twice per day. Sexually mature D. rerio were placed in a spawning tank in the late evening at a female:male ratio of 2:1, and the naturally fertilized eggs were collected in the early morning the next day and cultured at 27 ± 1°C until use.

Isolation of LTA-bound proteins and mass spectrometric analysis

The isolation of LTA-bound proteins and mass spectrometric analysis were carried out as previously described by Wang et al. (25), with slight modification. In brief, LTA-conjugated affinity resin was prepared by coupling LTA (from Bacillus subtilis; MilliporeSigma, Burlington, MA, USA) with cyanogen bromide –activated Sepharose CL-4B (GE Healthcare, Waukesha, WI, USA), according to the manufacturer’s instructions. The embryonic extracts were prepared from ∼300 embryos collected at 64–128-cell stages. The LTA-bound proteins obtained were separated by SDS-PAGE and visualized by silver staining according to the methods of Minoda et al. (26). Matrix-assisted laser desorption/ionization–time-of-flight mass spectrometry (MALDI/TOF MS) analysis was performed on a Bruker Ultraflex MALDI/TOF MS mass spectrometer (Bruker, Billerica, MA, USA) and searched against the nonredundant protein sequence database of the National Center for Biotechnology Information (NCBI) using the Mascot search engine (Matrix Science, Boston, MA, USA).

Western blotting

MALDI/TOF MS analysis revealed that one of the proteins was ATP5A1, a highly conserved protein among different species. To examine the distribution of ATP5A1 in the different tissues and at the different developmental stages, the tissues, including the heart, liver, spleen, gut, kidney, muscle, skin, brain, eye, gill, ovary, testis, and tail, were dissected out of zebrafish, and the embryos/larvae, including newly fertilized eggs; 4-, 16-, and 256-cell embryos; high blastulae and gastrulae; 10-somite larvae; 2-d-old larvae; 3-d-old larvae; and 7-d-old larvae, were collected. All the samples were homogenized in PBS (pH 7.4) using a JY92-IIN sonicator (Scientz, Ningbo, China), and the homogenates were centrifuged at 5000 g at 4°C for 10 min. The protein concentration of the supernatants was determined using a Bicinchoninic Acid Protein Assay Kit (ComWin Biotech, Beijing, China). The supernatants were loaded onto and run on a 12% SDS-PAGE gel. Western blotting was performed according to the method of Wang et al. (25). The primary antibody was mouse anti-bovine ATP5A1 antibody (ab110273; Abcam, Cambridge, United Kingdom), which was diluted 1:1000 with PBS (pH 7.4).

Immunohistochemistry

To examine the localization of ATP5A1 in the cells, the embryos at desired stages were fixed in 4% paraformaldehyde in PBS overnight. After washing 4 times with PBS containing 0.5% Triton X-100 (PBST), the embryos were blocked in PBST with 5% goat serum (E510009; Shanghai Sangon Biotech, Shanghai, China), 1% bovine serum albumin (BSA), and 1% DMSO for about 2 h and then incubated with mouse anti-bovine ATP5A1 antibody diluted 1:500 at 4°C overnight. They were then processed as described by Wang et al. (25). For control, the embryos were similarly processed and incubated with rabbit preimmune serum (D601019; Shanghai Sangon Biotech). To stain the nuclei, the embryos were counterstained with 5 μg/ml DAPI (Solarbio, Beijing, China) in PBS for 10 min, washed in PBST for 10 min, and stored at 4°C. The embryos/larvae were observed and photographed under a Leica confocal microscope (Leica, Wetzlar, Germany).

In parallel, 24 h past fertilization (hpf) larvae were fixed in 4% polyformaldehyde in PBS and precipitated in a 30% sucrose solution at 4°C overnight. Subsequently, they were embedded in optical cutting temperature compound, and sections were cut at 12-μm thickness at −20°C (Leica). The sections were stained as above, observed, and photographed under a Leica confocal microscope (Leica).

To further verity its subcellular localization, ATP5A1 was also overexpressed in human embryonic kidney (HEK) 293T cell line. The enhanced green fluorescent protein (eGFP) gene was subcloned into the eukaryotic expression vector pcDNA3.1/V5-His A (Thermo Fisher Scientific, Waltham, MA, USA) and the plasmid designated pcDNA3.1/V5/eGFP. Subsequently, the complete coding region of atp5a1 was amplified by PCR using the primer P1 and P2, and the PCR products were digested with HindIII and EcoRI and ligated into the pcDNA3.1/V5/eGFP (which was cut with the same restriction enzymes) upstream to construct the recombinant eukaryotic expression vector, pcDNA3.1/V5/atp5a1/eGFP. HEK 293T cells were seeded in 6-well plates and cultured at 37°C with 5% CO₂ in DMEM containing 10% fetal bovine serum, 10 U/m penicillin, and 100 mg/ml streptomycin. The plasmids pcDNA3.1/V5/eGFP and pcDNA3.1/V5/atp5a1/eGFP were then transfected into HEK 293T cells using Lipofectamine 2000 reagent (Thermo Fisher Scientific) according to the manufacturer’s instructions. At 48 h after transfection, the transfected cells were washed with PBS, fixed with 4% paraformaldehyde, and stained with 1 mg/ml DAPI according to the manufacturer’s instructions. The samples were examined by a Leica confocal microscope (Leica).

Cloning and sequencing of atp5a1

A total of 60 embryos/larvae of zebrafish were collected at about the 128-cell stage and ground in RNAiso Plus (Takara, Kyoto, Japan). Total RNA was isolated from the ground samples according to the manufacturer’s instructions (Omega, Norcross, GA, USA). After digestion with recombinant DNase I (RNase-free) (Takara) to eliminate genomic contamination, the cDNAs were synthesized with a reverse transcription kit (Takara) with an oligo (dT) primer. The reaction was carried out at 37°C for 15 min and inactivated at 85°C for 5 s. The cDNAs were stored at −20°C until use.

A pair of the primers P3 and P4 (Table 1) specific to the ATP5A1 gene was designed using the Primer Premier program, v.5.0 (http://www.premierbiosoft.com/primerdesign/), based on the sequence of zebrafish atp5a1 gene (NM_001077355.1) on NCBI and used for PCR to amplify the gene. The PCR protocol was as follows: an initial denaturation at 94°C for 5 min, followed by 37 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 40 s, and a final extension at 72°C for 10 min. The amplification product was cloned into the pGEM-T vector (Tiangen Biotech, Beijing, China) following the manufacturer’s instructions and transformed into Trans 5α bacteria (TransGen Biotech, Beijing, China). The DNA inserts were sequenced by Shanghai Sangon Biotech to verify for authenticity.

TABLE 1.

Sequences of the primers used in this study

	Primer sequence, 5′–3′
Description	Sense	Antisense
Subcellular localization (P1a and P2 b)	CCCAAGCTTATGCTTTCCGTTCGCGTCGCGGCGG	CCGGAATTCCTCGAAGCTCGAGAGGAAGTTGAGC
Atp5a1 cloning (P3 and P4)	ATGCTTTCCGTTCGCGTCGCGGCGG	CTACTCGAAGCTCGAGAGGAAGTTG
Real-time quantitative PCR (P5 and P6)	GCTGTTGCCTACCGTCAGATGT	TCCACCACCGAAGTTGTCGTTCA
Real-time quantitative PCR (P7 and P8)	GATGCGGAAACTGGAAAGGG	AGGAGGGCAAAGTGGTAAACG
WISH (P9 and P10)	GGCGGAGGTGTCCAGCATT	AGCGTGTTTGCCGTTGTCT
ATP5A1 expression in vitro (P11b and P12 a)	CCGGAATTCATGCTTTCCGTTCGCGTCGCGGCGG	CCCAAGCTTCTACTCGAAGCTCGAGAGGAAGTTG
16s rRNA of A. hydrophila (P13 and P14)	AATACCGCATACGCCCTAC	AACCCAACATCTCACGACAC
A135–551 expression in vitro (P15b and P16a)	CCGGAATTCGCTATTGTGGACGTTCCTGTCGGAG	CCCAAGCTTCTACTCGAAGCTCGAGAGGAAGTTG
A1–65 expression in vitro (P17b and P18a)	CCGGAATTCATGCTTTCCGTTCGCGTCGCGGCGG	CCCAAGCTTAGCTCCAGTGTCGGCTCCAAGAATC
A66–134 expression in vitro (P19b and P20a)	CCGGAATTCGAACTGGAGGAGACCGGCCGCGTGC	CCCAAGCTTCCCTGTTCTCTTGACGATGTCTCCC
MA1–65 expression in vitro (P21c and P22a)	CGCGGATCCATGCTGTCTGTGCGCGTCGCCGCGG	CCCAAGCTTAGACGTGTCAGCTCCCAGAATCCTC
CA1–65 expression in vitro (P23c and P24a)	CGCGGATCCATGATTTCATCAAAGGTTATCGCAG	CCCAAGCTTTGGTGCAGTTCCAAGAATTCTTTGT

Underlined sequences: ^aHindIII site; ^bEcoRI site; ^cBamHI site.

Homology, phylogenetic, and syntenic analyses

The cDNA obtained was analyzed for coding probability with the DNAStar software package v.5.0 (DNAStar, Madison, WI, USA), and the domain and signal peptide were predicted using the Simple Modular Architecture Research Tool (SMART) program (http://smart.embl-heidelberg.de/) and the SignalP 4.1 server (http://www.cbs.dtu.dk/services/SignalP/), respectively. The molecular mass (MW) and isoelectronic point of the mature protein were determined using ProtParam (http://www.expasy.ch/tools/protparam.html). The nuclear localization signal was analyzed by the PSORT II (http://psort.hgc.jp/form2.html). Homology search in the GenBank database (https://www.ncbi.nlm.nih.gov/genbank/) was carried out by Basic Local Alignment Search Tool (BLAST) server (http://www.ncbi.nlm.nih.gov/BLAST/), and the phylogenetic tree constructed by MEGA (v.7.0) using p-distance was based on the neighbor-joining method. The reliability of each node was estimated by bootstrapping with 1000 replications. The numbers shown at each node indicate the bootstrap values (%). The 3-dimensional (3D) structure prediction was performed by Swiss-Model online software at the Expert Protein Analysis System (http://www.expasy.org/) using human ATP5A1 isoform a (Protein Data Bank (http://www.rcsb.org/) code: 2w6e.1.C) as a model. Chromosomal locations of human (Homo sapiens), cow (Bos taurus), mouse (Mus musculus), frog (Xenopus tropicalis), and zebrafish (D. rerio) atp5a1 genes were obtained from the Sequence Viewer (http://www.ncbi.nlm.nih.gov/projects/sviewer) and Ensembl Genome Browser (http://www.ensembl.org).

Real-time quantitative PCR

To detect the expression patterns of atp5a1 in the different tissues and at the different developmental stages, the tissues (kidney, heart, liver, spleen, gut, muscle, gill, eye, ovary, testis, brain, skin, and tail) were dissected out of adult zebrafish, and the embryos/larvae (newly fertilized eggs; 4-, 16-, and 256-cell embryos; high blastulae and gastrulae; 10-somite larvae; 2-d-old larvae; 3-d-old larvae; and 7-d-old larvae) were collected. All the samples were ground in RNAiso Plus (Takara) and stored at −70°C until use. A pair of the primers, P5 and P6 (Table 1), specific to atp5a1 was designed using the Primer Premier program, v.5.0. Real-time quantitative PCR was performed as previously described by Wang et al. (25). The gene β-actin (P7 and P8; Table 1) was chosen as the reference for internal standardization. The expression levels of atp5a1 relative to that of the β-actin gene were calculated by the comparative C_t method (2^−ΔΔCt) (27).

Whole-mount in situ hybridization

The fertilized eggs; 2-cell embryos; high blastulae and gastrulae; 10-somite larvae; and 1-, 2-, and 3-d-old larvae were collected. A fragment of atp5a1 was PCR amplified using the primer pair P9 and P10 (Table 1) and subcloned into vector pGEM-T (Tiangen Biotech). The constructed vector was digested by the NcoI enzyme (Takara), and the atp5a1-specific antisense probe labeled with digoxigenin-linked nucleotides was synthesized by Sp6 RNA polymerase (Thermo Fisher Scientific). Whole-mount in situ hybridization (WISH) was performed as described by Thisse and Thisse (28). The embryos/larvae were observed and photographed under the Nikon SMZ1000 stereomicroscope (Nikon, Tokyo, Japan).

Expression and purification of recombinant ATP5A1

The cDNA encoding complete ATP5A1 was amplified by PCR using the sense primer P11 (EcoRI site is underlined) and the antisense primer P12 (HindIII site is underlined) (Table 1) as described by Wang et al. (25). The PCR product was digested with EcoRI (Takara) and HindIII (Takara) and subcloned into the pET28a expression vector (Merck GmbH, Darmstadt, Germany) digested previously with the same restriction enzymes. The insert was verified by DNA sequencing, and the expression construct was designated pET28a/atp5a1.

The cells of Escherichia coli BL21 (DE3; TransGen Biotech) were transformed with the plasmid pET28a/atp5a1 and then cultured in Luria broth containing 50 μg/ml kanamycin. The cultures were diluted 1:100 with LB and further incubated at 37°C for 3 h. The expression of the recombinant protein was induced by addition of isopropyl β-d-thiogalactoside (IPTG) to the cultures at a final concentration of 0.1 mM. After overnight induction at 37°C, the bacterial cells were collected, and the recombinant (r)ATP5A1 was purified as previously described by Du et al. (29). The purified protein was refolded by dialysis, analyzed by SDS-PAGE on a 12% gel, and stained with Coomassie Brilliant Blue R-250. In parallel, the purified rATP5A1 was also characterized by Western blotting. Protein concentration was determined by the bicinchoninic acid method using BSA as a standard.

To express the thioredoxin (TRX)-histidine (His)-tag peptide for control, E. coli BL21 cells were also transformed by plasmid pET32a. The induction and purification of recombinant TRX-His-tag peptide were performed as previously described.

Assay for affinity of rATP5A1 to bacteria, LTA, LPS, and lipid A

Western blotting was used to detect the binding of rATP5A1 to the gram-positive bacteria Micrococcus luteus [American Type Culture Collection (ATCC) 49732], B. subtilis (ATCC 6633), and Staphylococcus aureus (ATCC 25923) and the gram-negative bacteria E. coli (ATCC 25922), Vibrio anguillarum (ATCC 43308), and Aeromonas hydrophila (ATCC 35654) by the method of Wang et al. (25). The binding of rATP5A1 to the gram-positive and gram-negative bacteria was also assayed by the method of Li et al. (30), with slight modification. Briefly, the gram-positive bacteria M. luteus, B. subtilis, and S. aureus were mixed with either FITC-labeled rATP5A1 or FITC-labeled rATP5A1, which had been preincubated with LTA, and the gram-negative bacteria E. coli, V. anguillarum, and A. hydrophila were mixed with either FITC-labeled rATP5A1 or FITC-labeled rATP5A1, which had been preincubated with LPS. After incubation at 28°C for 2 h, the mixtures were observed and photographed under an Olympus BX51 fluorescence microscope (Olympus, Tokyo, Japan). The FITC-labeled TRX-His-tag peptide was used as control. To examine the affinity of rATP5A1 to LTA and LPS, ELISA was used as previously described by Wang et al. (25). Both rATP5A1 and BSA (used as control) were biotinylated with biotinamidohexanoic acid N-hydroxysuccinimide ester (NHS-LC-biotin; Heowns, Tianjin, China).

LPS consists of lipid A and polysaccharide. Thus, we also tested whether rATP5A1 binds to lipid A. A stock solution of lipid A purchased from MilliporeSigma was prepared by dissolving in DMSO, giving a concentration of 1 mg/ml. Aliquots of 50 μl of 80 μg/ml lipid A in 10 mM PBS (pH 7.4) were applied to each well of a 96-well microplate, air dried in 25°C, and processed as previously described.

To test whether rATP5A1 has any lectin activity, aliquots of 50 μl of 80 μg/ml LPS in 10 mM PBS (pH 7.4) were applied to each well of a 96-well microplate and air dried at 25°C overnight. The plate was incubated at 60°C for 30 min to further fix LPS, and the wells were each blocked with 200 μl of 1 mg/ml BSA in 10 mM PBS (pH 7.4) at 37°C for 2 h. Meanwhile, aliquots of 50 μl of 25 μg/ml rATP5A1 or BSA (control) in 10 mM PBS (pH 7.4) containing different concentrations (0, 2.5, 5, 10, and 20 mg/ml) of various sugars d-glucose, d-galactose, d-mannose, N-acetyl-d-galactosamine (GlcNAc), and N-acetyl-d-mannosamine were preincubated at room temperature for 1 h and then added into each well. After incubation at room temperature for 3 h, the wells were processed and measured as previously described by Gao et al. (31).

Assay for antibacterial activity of rATP5A1 in vitro

The gram-positive bacteria M. luteus, B. subtilis, and S. aureus and the gram-negative bacteria E. coli and V. anguillarum were cultured in LB medium at 37°C to midlogarithmic phase and collected by centrifugation at 6000 g at room temperature for 10 min. The gram-negative bacterium A. hydrophila was also cultured in tryptic soy broth medium at 28°C to midlogarithmic phase and collected by centrifugation. The bacterial pellets were washed in 20 mM Tris-HCl buffer (pH 7.4) 3 times, resuspended, and adjusted to a density of 1 × 10⁵ cells/ml. The antibacterial activity of rATP5A1 was assayed as by the method of Du et al. (29). In addition, transmission electron microscopy (TEM) was used to observe the effects of rATP5A1 on the fin structure of the bacteria tested. Briefly, aliquots of 500 μl of the bacterial suspensions containing 5 × 10⁷ cells/ml were mixed with 500 μl of rATP5A1 (80 μg/ml). In parallel, aliquots of 500 μl of the bacterial suspensions were mixed with 500 μl PBS as control. The mixtures were incubated at 25°C for 1 h, fixed in 2.5% glutaraldehyde in 100 mM PBS, and then dropped onto 400-mesh carbon-coated grids and allowed to stand at room temperature for 3 min for negative staining. Excess fluid was removed by touching the edge of filter paper. The grids were then put into 2% phosphotungstic acid for 3 min and dried by filter paper. Observation was performed under a Jeol JSM-840 transmission electron microscope (Jeol, Tokyo, Japan).

Assay for membrane depolarization activity of rATP5A1

The membrane depolarization activity of rATP5A1 was assayed with 3, 3′-dipropylthiacarbocyanine iodide (DiSC₃-5, a potential-dependent distributional fluorescent dye; MilliporeSigma). The gram-positive bacteria M. luteus, B. subtilis, and S. aureus and the gram-negative bacteria E. coli, V. anguillarum, and A. hydrophila were cultured as above. The bacterial cells in the midlogarithmic phase were harvested by centrifugation at 6000 g for 10 min, washed in 5 mM HEPES buffer (pH 7.4) containing 20 mM glucose, and resuspended in 5 mM HEPES buffer containing 20 mM glucose and 100 mM KCl, giving a concentration of 1 × 10⁷ cells/ml. Aliquots of 100 μl of the bacterial suspensions were each applied to a well of a 96-well plate. A stock solution of DiSC₃-5 was added to the bacterial suspensions, yielding a final concentration of 0.5 μM, and subjected to incubation at room temperature for 30 min to get a steady baseline of fluorescence intensity. The bacterial suspensions were then mixed with 100 μl of different concentrations of rATP5A1, giving the desired concentrations of 20 and 40 μg/ml, respectively. HEPES buffer containing 20 mM glucose was used as control. Changes in fluorescence intensity were continuously recorded for 30 min with a GENios plus spectrofluorometer (Tecan, Männedorf, Switzerland) at an excitation wavelength of 622 nm and an emission wavelength of 670 nm.

Assay for membrane permeabilization activity of rATP5A1

Next, flow cytometry was used to measure the effect of rATP5A1 on the bacterial membranes of the gram-positive bacteria M. luteus, B. subtilis, and S. aureus and the gram-negative bacteria E. coli, V. anguillarum, and A. hydrophila. The culture and collection of the bacteria were as described above. After washing 3 times with PBS, the bacterial pellets were suspended in PBS and adjusted to a density of 1 × 10⁶ cells/ml, and then rATP5A1 was added to the bacterial suspensions, yielding final concentrations of 10, 20, and 40 μg/ml. For control, the bacteria were mixed with PBS alone. The mixtures were incubated at 28°C for 2 h and fixed with 10 μM propidium iodide (PI; a DNA-intercalating agent and a fluorescent molecule; MilliporeSigma) solution under dark conditions at 4°C for 15 min. The bacterial cells stained by PI were examined using a FC500MPL flow cytometer (Beckman Coulter, Brea, CA, USA). Data were analyzed using WinMDI, v.2.9, software (Scripps Research Institute, San Diego, CA, USA).

Assay for antibacterial activity of ATP5A1 in embryos/larvae

As rATP5A1 showed antibacterial activity in vitro, we wondered whether ATP5A1 plays any role in the immune defense in early embryos. The 64–128-cell-stage embryos of zebrafish collected were washed 3 times with sterilized distilled water, homogenized, and centrifuged at 5000 g at 4°C for 10 min. The embryo extracts were then used to examine the antibacterial activity against the gram-positive bacteria M. luteus, B. subtilis, and S. aureus and the gram-negative bacteria E. coli, V. anguillarum, and A. hydrophila as previously described by Wang et al. (32). To precipitate native ATP5A1 in the embryo extracts, 1 ml of the extracts was mixed with 0.5 μg of mouse anti-ATP5A1 antibody (ATP5A1Ab). For control, 1 ml of the embryo extracts was mixed with 0.5 μg of anti–β-actin antibody (AcAb) (CWO263A; Cwbio, Beijing, China). We then assayed whether ATP5A1 has any ability to protect developing embryos in vivo. Fifty 8-cell-stage embryos were microinjected into the yolk sac with 6 nl of sterilized PBS (blank control), mouse ATP5A1Ab solution (0.33 ng), AcAb solution (0.33 ng), purified rATP5A1 solution (0.6 ng), or BSA solution (0.6 ng) and challenged 1 h later by injection of 6 nl of live A. hydrophila (pathogenic to zebrafish) suspension containing about 500 bacterial cells. The mortality rate was recorded, and the cumulative mortality rate was calculated at 24 h after the bacterial injection. To confirm the specificity of the antimicrobial activity of rATP5A1 in vivo, mouse ATP5A1Ab was injected together with rATP5A1 into the embryos, which were then challenged by injection of live A. hydrophila. For control, the embryos were coinjected with AcAb and rATP5A1 and treated similarly.

To verify the killing of A. hydrophila by developing embryos, 8-cell-stage embryos were microinjected with live A. hydrophila as above. Five embryos were collected each time at 0, 12, and 24 h after the bacterial injection. The normal embryos were similarly collected as control. The total DNAs were isolated from each embryo and used for PCR analyses as previously described by Wang et al. (32). PCR was carried out to amplify a specific region of the A. hydrophila 16s rRNA gene using the primer pair P13 and P14 (Table 1).

Assay for antibacterial activity of atp5a1-morpholino-knockdown embryos/larvae

The sequence-specific antisense oligonucleotides of 25 nt, CGCGACGCGAACGGAAAGCATTTTT [morpholino for atp5a1 (atp5a1-MO)] and CCTCTTACCTCAGTTACAATTTATA [control morpholino (control-MO)] were designed and synthesized by Gene Tools (Philomath, OR, USA). One hundred 1–2-cell-stage embryos were microinjected with 12 ng/6 nl of atp5a1-MO, and half of the embryos were challenged 24 h later by injection of 6 nl of live A. hydrophila suspension (∼500 cells). For control, the embryos were injected with the control-MO and treated similarly. The efficacy of atp5a1-MO was confirmed by the method of Du et al. (29) through Western blotting. All the embryos were cultured at 27°C, and the mortality was evaluated 24 h later after the challenge with A. hydrophila.

Titration of ATP5A1 content in eggs/embryos

Exactly 120 embryos collected at 0, 12, and 24 hpf were washed 3 times with sterilized 0.9% saline, homogenized, and centrifuged at 5000 g at 4°C for 5 min. The supernatants were pooled and used to measure the content of ATP5A1 by ELISA as previously described by Wang et al. (25).

Assay for structure-activity relationship

To determine the structure-activity relationship, the cDNA regions encoding the N-terminal 65 residues 1–65 of D. rerio ATP5A1 (A_1–65; with the C-terminal 486 residues depleted), the N-terminal 69 residues 66–134 (A_66–134; which consists of the ATP_synt_ab_N domain per se), and the C-terminal residues 135–551 (A_135–551; with the N-terminal 134 residues depleted) were amplified by PCR using the pairs of primers P15 and P16, P17 and P18, and P19 and P20 (Table 1). The expression vector plasmids pET28a/a1-65, pET28a/a66-134, and pET28a/a135-551 were constructed and transformed into E. coli BL21 cells. The expression and purification of the recombinant proteins were performed as above, and the purified proteins were verified by Western blotting using anti-His-Tag mouse monoclonal antibody. The antibacterial activity of rA_1–65, rA_66–134, and rA_135–551 against the gram-positive and gram-negative bacteria as well as their affinity to LTA and LPS were both assayed as above. In addition, rA_1–65, rA_66–134, and rA_135–551 were also microinjected into zebrafish embryos, followed by challenge of live A. hydrophila, to test their antibacterial activity in vivo.

Assay for antibacterial activity of ATP5A1 homologs in mouse and ascidian

Next, we tested whether the antibacterial activity of ATP5A1 was conserved during evolution. As the antibacterial activity of rA_1–65 was comparable to that of rATP5A1 in zebrafish, we thus amplified the cDNA regions encoding the N-terminal 65 residues of mouse (M. musculus) ATP5A1 (GenBank accession no. NP_031531.1; MA_1–65) and the N-terminal 65 residues of ascidian (Ciona intestinalis) ATP5A1 (GenBank accession no. NP_001027729.1; CA_1–65) by PCR using the pairs of primers P21 and P22, P23 and P24. The expression vector plasmids pET28a/ma1-65 and pET28a/ca1-65 were constructed and transformed into E. coli BL21 cells. The recombinant polypeptides rMA_1–65 and rCA_1–65 were prepared, and their antibacterial activities as well as affinity to LTA and LPS were similarly assayed as previously described.

Statistical analyses

All the assays were performed in triplicate (technical replicates), and each experiment was repeated 3 times (biologic replicates). Statistical analyses were performed using Prism 5 (GraphPad Software, La Jolla, CA, USA). The significance of difference was determined by 2-way ANOVA. The difference at P < 0.05 was considered significant. All of the data were expressed as means ± sd.

RESULTS

ATP5A1 is an LTA-binding protein abundant in eggs/embryos

The proteins eluted from the LTA-conjugated Sepharose CL-4B column were resolved by SDS-PAGE, and 6 main bands (1–6) were identified (Fig. 1A). MALDI/TOF MS analysis revealed that the bands 1–6 were tubulin β chain (unpublished results), angiopoietin 4 (unpublished results), receptor-interacting protein 1 (unpublished results), ATP synthase subunit α (Fig. 1B), human antigen R (ELAV)-like RNA binding protein 1 (unpublished results), and kinetochore protein Nuf2 (unpublished results), including 445, 496, 661, 364, 91, and 454 aa, respectively. Western blotting showed that ATP5A1 was primarily present in the heart, liver, muscle, skin, ovary, and testis, with relatively higher levels in the ovary (Fig. 1C). This suggests that ATP5A1 was abundantly distributed in the ovary. Accordingly, ATP5A1 was also detected in the fertilized eggs and embryos/larvae up to 7 d (Fig. 1D). These indicate that ATP5A1 was an LTA-binding protein stored mainly in the eggs/embryos of zebrafish.

Figure 1.

Identification of ATP5A1 as an LTA-binding protein. A) SDS-PAGE of the proteins isolated from the embryo extracts of zebrafish on LTA-conjugated Sepharose CL-4B affinity resin. Lane M, marker; lane 1, embryo extracts; lane 2, effluent fractions after Tris-HCl washing; lane 3, effluent fractions when all proteins were washed clean; lane 4, effluent fractions containing the adsorbed proteins 1–6. B) The protein sequences from MALDI/TOF MS analysis of A, band 4. The matched peptides are shown in red. C) Western blotting (a) and quantifications (b) of ATP5A1 in different tissues including heart (H), liver (L), spleen (Sp), gut (Gu), kidney (K), muscle (Mu), skin (Sk), brain (Br), eye (E), gill (Gi), ovary (O), testis (Te), and tail (Ta). D) Western blotting (a) and quantifications (b) of ATP5A1 at the different developmental stages, including zygote (0 h), 4-cell stage (about 1 h), 16-cell stage (∼1.5h), 256-cell stage (∼2.5 h), high blastula stage (∼3.5 h), 50% epiboly stage (∼5.5 h), 10-somite stage (∼15 h), 2-d postfertilization (dpf), 3-dpf, and 7-dpf. Data are presented as means ± sd.

ATP5A1 is distributed in cytoplasm of embryonic cells

Immunohistochemical examination revealed that ATP5A1 was localized in the blastoderm of fertilized eggs; 4-, 16- and 256-cell embryos; blastulae (Fig. 2Aa–d); and the epidermis and neural tube of the head of 24-h larvae (Fig. 2Bb, c). As shown by the high resolution of frozen sections in Fig. 2B, ATP5A1 was apparently localized in the cytoplasm. This was further confirmed by the overexpression of ATP5A1 in HEK 293T cells (Fig. 2C). These suggest that ATP5A1 was a protein distributed in the cytoplasm of embryonic cells.

Figure 2.

Distribution of zebrafish ATP5A1 in the cells of different developmental stage embryos and subcellular localization of ATP5A1 in HEK 293T cells. A) Immunohistochemical localization of ATP5A1 in the different developmental stages: 2-cell stage (about 0.75 h) (a), 8-cell stage (∼1.25 h) (b), 16-cell stage (∼1.5 h) (c), 256-cell stage (∼2.5 h) (d), 50% epiboly stage (∼5.5 h) (e), 10-somite stage (∼15 h) (f), 24-h postfertilization (24-hpf) (g), and 16-cell embryo (h), which was incubated with rabbit preimmune serum as a negative control. Nucleus was visualized by DAPI staining (blue). Scale bars, 100 μm. B) High resolution of A. a, b) High resolution of Ae (a) and Ag (b). Scale bars, 25 μm. c) High resolution of a frozen section view shown in Ag. Scale bars, 10 μm. C) Subcellular localization of ATP5A1 in HEK 293T cells. The HEK 293T cells were transiently transfected with pcDNA3.1/V5/eGFP or pcDNA3.1/V5/atp5a1/eGFP. After 48 h, the cells were imaged by fluorescence microscopy. The nucleus was stained by DAPI. Scale bar, 25 μm.

Structure, phylogenetics, and synteny of atp5a1

Zebrafish ATP5A1 cDNA obtained was 1876 bp long and contained an open reading frame of 1656 bp, a 5′-UTR of 51 bp, and a 3′-UTR of 169 bp (Supplemental Fig. S1). The open reading frame coded for a protein of 551 aa with a calculated molecular mass of ∼59.74 kDa and an isoelectronic point of 8.87. Zebrafish ATP5A1 included ATP-synt_ab_N domain (at residues 68–134), F1_ATPase_alpha domain (at residues 136–417), and ATP-synt_ab_C domain (at residues 426–530; Fig. 3A). No nuclear localization signal was identified in zebrafish ATP5A1 (Supplemental Table S1), consistent with its localization in the cytoplasm of embryonic cells. Sequence alignment showed that zebrafish ATP5A1 was 91.3, 89.5, 92.0, 92.9, 92.4, 85.8, 77.1, and 84.4% identical to mammalian, avian, reptile, amphibian, fish, amphioxus, ascidian, and sea anemone ATP5A1 proteins (Supplemental Fig. S2), respectively. The structure prediction using Swiss-Model online software revealed that the 3D structure of zebrafish ATP5A1 was closely similar to that of human ATP5A1, both consisting of 21 α-helices and 24 β-sheets (Fig. 3B).

Figure 3.

Structure, phylogenetics, and synteny of zebrafish ATP5A1. A) Domain structure of ATP5A1 predicted by the SMART program. B) 3D structures of the ATP5A1 generated by Swiss-Model online software using human ATP5A1 isoform a (Protein Data Bank code: 2w6e.1.C) as the model. C) Phylogenetic tree constructed by MEGA, v.7.0, using the neighbor-joining method. The accession numbers of ATP5A1 proteins used are as follows: H. sapiens ATP5A1 isoform a (NP_001001937.1), H. sapiens ATP5A1 isoform b (NP_001244263.1), H. sapiens ATP5A1 isoform c (NP_00124264.1), Pongo abelii ATP5A1 (NP_0011266.846.1), M. musculus ATP5A1 (NP_031531.1), Rattus norvegicus ATP5A1 (NP_075581.1), Sus scrofa ATP5A1 (NP_001172071.1), B. taurus ATP5A1 (NP_777109.1), Corvus cornix cornix ATP5A1 (XP_019147514.2), Gallus gallus ATP5A1 (NP_989617.2), Pogona vitticeps ATP5A1 (XP_020662197.1), Alligator mississippiensis ATP5A1 (KYO28859.1), X. tropicalis ATP5A1 (NP_001025610.1), Salmo salar ATP5A1 (NP_001133132.1), Oncorhynchus mykiss ATP5A1 (XP_021457645.1), Astyanax mexicanus ATP5A1 (XP_007256648.2), Boleophthalmus pectinirostris ATP5A1 (XP_020787965.1), Sparus aurata ATP5A1 (AGV76807.1), Fundulus heteroclitus ATP5A1 (XP_012723449.1), D. rerio ATP5A1 (NP_001070823.1), Monopterus albus ATP5A1 (XP_020449614.1), Exaiptasia pallida ATP5A1 (KXJ16138.1), Branchiostoma belcheri ATP5A1 (XP_019644160.1), and C. intestinalis ATP5A1 (NP_001027729.1). D) Syntenic map of the genomic segment where atp5a1 resides in the chromosomes of zebrafish and other vertebrates including humans. Boxes represent genes, and the direction of box indicates the direction of gene.

The phylogenetic tree constructed using the sequences of ATP5A1 proteins available showed that zebrafish ATP5A1 was clubbed together with other teleost ATP5A1 proteins, forming an independent clade branched from mammalian, avian, and amphibian counterparts (Fig. 3C). This well reflected the established phylogeny of chosen organisms. Syntenic analysis showed that zebrafish atp5a1 was mapped onto chromosome 21 and closely linked to haus1, resembling that of human, cow, mouse, and frog atp5a1 genes (Fig. 3D). This suggests that the linkage of atp5a1 with haus1 was conserved throughout vertebrate evolution.

Tissue- and stage-specific expression of atp5a1

Real-time quantitative PCR showed that atp5a1 was highly expressed in the eye, heart, muscle, liver, ovary, testis, and brain and at a lower level in the gill, gut, kidney, and tail (Fig. 4A). As shown in Fig. 4B, atp5a1 mRNA was abundant in the zygotes, which decreased continuously when development went on. However, some atp5a1 mRNA still remained in 7-d larvae. These indicate that the maternal atp5a1 mRNA level was relatively higher in the fertilized eggs, which gradually decreased with development. Interestingly, the comparison between mRNA level (Fig. 4) and protein level (Fig. 1) revealed that there was some difference between atp5a1 mRNA and its protein content in the respective tissues. This might be because of the existence of translational control in addition to transcriptional control of this factor.

Figure 4.

Expression patterns of zebrafish atp5a1 in the different tissues and at the different developmental stages. A) Expression profiles of zebrafish atp5a1 in the different tissues including spleen (Sp), gill (Gi), eye (E), heart (H), skin (Sk), muscle (Mu), kidney (K), liver (L), gut (Gu), ovary (O), testis (Te), tail (Ta), and brain (Br). B) Expression profiles of zebrafish atp5a1 at the different developmental stages including zygote (0 h), 4-cell stage (about 1 h), 16-cell stage (∼1.5h), 256-cell stage (∼2.5 h), high blastula stage (∼3.5 h), 50% epiboly stage (∼5.5 h), 10-somite stage (∼15 h), 2-d postfertilization (dpf), 3-dpf, and 7-dpf. β-Actin was chosen as the internal control for normalization. Relative expression data were calculated by the method of 2^−∆∆Ct. The vertical bars represent means ± sd. (n = 3). The data are from 3 independent experiments performed in triplicate. C) Expression of atp5a1 during early development detected by WISH. Stages of embryonic development: newly fertilized egg (0 h) (a); 2-cell-stage embryo (∼0.75 h) (b); high blastula–stage embryo (∼3.5 h) (c); 50% epiboly-stage embryo (∼5.5 h) (d); 10-somite larvae (∼15 h) (e); 1-d-old larvae (f); 2-d-old larvae (g); and 3-d-old larvae (h).

As shown in Fig. 4C, pan-expression of atp5a1 was observed in the embryos before segmentation stages (Fig. 4Ca–e), and then atp5a1 mRNA was detected primarily in the head region of 1-, 2-, and 3-d-old larvae (Fig. 4Cf–h), which have the heart, brain, and eyes as the main organ at these stages. These suggest that ATP5A1 was somehow involved in early development, especially in the formation of the heart, brain, and eyes.

rATP5A1 binds to gram-positive and gram-negative bacteria, LTA, LPS, and lipid A

ATP5A1 was expressed in E. coli BL21 (DE3) and purified by affinity chromatography on a Ni-nitrilotriacetic acid resin column. The purified recombinant protein, rATP5A1, yielded a single band of ∼63.56 kDa on SDS-PAGE after Coomassie Blue staining, corresponding to the expected size (Fig. 5A). Western blotting showed that the purified rATP5A1 reacted with the mouse anti-His-tag monoclonal antibody, indicating that rATP5A1 was correctly expressed.

Figure 5.

SDS-PAGE and Western blotting of rATP5A1 and its binding to gram-positive and gram-negative bacteria, LTA, LPS, and lipid A. A) SDS-PAGE. Lane M, marker; lane 1, negative control (empty vector pET28a); lane 2 and 3, total cellular extracts from E. coli BL21 containing expression vector pET28a/atp5a1 before induction (2) or induced with IPTG (3); lane 4, purified rATP5A1; 5, Western blotting of rATP5A1. B, C) Western blotting about interaction of rATP5A1 with gram-positive (B) and gram-negative (C) bacteria. Lane M, marker; lane 1, purified rATP5A1; lane 2, 4, and 6, M. luteus, B. subtilis, and S. aureus (B) or E. coli, V. anguillarum, and A. hydrophila (C) incubated in the presence of rATP5A1; lane 3, 5, and 7, M. luteus, B. subtilis, and S. aureus (B) or E. coli, V. anguillarum, and A. hydrophila (C) incubated in the absence of rATP5A1. D) No affinity of TRX-His-tag peptide with the gram-positive and gram-negative bacteria. Lane M, marker; lane 1, purified TRX-His-tag; lane 2, 3, 4, 5, 6, and 7, M. luteus, B. subtilis, S. aureus, E. coli, V. anguillarum, and A. hydrophila incubated in the presence of TRX-His-tag peptide. E, F). Binding of FITC-labeled rATP5A1 to gram-positive (Ea, d, g) and gram-negative (Fa, d, g) bacteria; binding of FITC-labeled rATP5A1 preincubated with LTA to gram-positive bacteria (Eb, e, h) or preincubated with LPS to gram-negative bacteria (Fb, e, h); no binding of FITC-labeled TRX-His-tag peptide to gram-positive (Ec, f, i) and gram-negative (Fc, f, i) bacteria. Scale bars, 20 μm. G, H, I) Interaction of rATP5A1 with LTA (G), LPS (H), and lipid A (I) revealed by ELISA. J) Effects of various sugars on the interaction of rATP5A1 with LPS. Each point in the graph represents the mean ± sd (n = 3). The data are from 3 independent experiments performed in triplicate. The bars represent means ± sd. GalNAc, N-acetylgalactosamine; GlcNAc, N-acetylglucosamine; OD, optical density.

Figure 5B, C show the interaction of rATP5A1 with the gram-positive bacteria M. luteus, B. subtilis, and S. aureus and the gram-negative bacteria E. coli, V. anguillarum, and A. hydrophila, revealed by Western blotting. It was apparent that rATP5A1 bound to all the bacteria tested, whereas the TRX-His-tag peptide used as control showed little binding to the same bacteria examined (Fig. 5D). In agreement with above observations, FITC-labeled rATP5A1 also bound to all the gram-positive and gram-negative bacteria tested (Fig. 5Ea, d, g, Fa, d, g), but the FITC-labeled TRX-His-tag peptide did not (Fig. 5Ec, f, i, Fc, f, i). These indicate that zebrafish rATP5A1 was able to interact with both the gram-positive and gram-negative bacteria. This not only supports that ATP5A1 was an LTA-binding protein but also suggests that it was an LPS-binding protein as well.

ELISA was also used to test the interaction of rATP5A1 with LTA and LPS. It showed that rATP5A1 bound to both LTA and LPS in a dose-dependent manner, whereas BSA used as negative control did not (Fig. 5G, H). These confirmed without doubt that zebrafish ATP5A1 was indeed an LTA- and LPS-binding protein. In addition, we also demonstrated that the FITC-labeled rATP5A1, which was preincubated with LTA or LPS, displayed reduced binding to the gram-positive or gram-negative bacteria, respectively (Fig. 5Eb, e, h, Fb, e, h). These suggest that rATP5A1 interacted with the gram-positive and gram-negative bacteria via LTA and LPS, respectively.

We also found that rATP5A1 had an affinity to lipid A (a core component of LPS) comparable with that of LPS (Fig. 5I). Besides, the binding of rATP5A1 to LPS was not inhibited by any of the sugars tested, even at a concentration as high as 20 mg/ml (Fig. 5J). These indicate that rATP5A1 interacted with LPS via lipid A and had little lectin activity.

Antibacterial activity of rATP5A1

As zebrafish ATP5A1 showed affinity to LTA and LPS, we thus wondered if it had any antibacterial activity. As shown in Fig. 6A, the growth of all the gram-positive and gram-negative bacteria tested were inhibited by rATP5A1 in a dose-dependent manner, indicating that in addition to binding to LTA and LPS, rATP5A1 showed a wide spectrum of antibacterial activity. At the concentration of 40 μg/ml, the antibacterial activity of rATP5A1 against the gram-positive and gram-negative bacteria was nearly comparable to that of 2 μg/ml kana. TEM examination revealed that rATP5A1 caused a direct damage to the cells of bacteria (Fig. 6B) (i.e., the cell surfaces were dissolved, the membranes were disrupted, and the cytoplasm became less dense). These indicate that rATP5A1 was a bactericidal agent capable of inhibiting the growth of both the gram-positive and gram-negative bacteria.

Figure 6.

Antibacterial activity of rATP5A1 and its effects on bacterial structures. A) Antibacterial activities of rATP5A1 against gram-positive bacteria M. luteus, B. subtilis, and S. aureus and gram-negative bacteria E. coli, V. anguillarum, A. hydrophila. Each point in the graph represents the mean ± sd (n = 3). The data are from 3 independent experiments performed in triplicate. The bars represent means ± sd. B) TEM. M. luteus (a), B. subtilis (b), S. aureus (c), E. coli (d), V. anguillarum (e), and A. hydrophila (f) incubated with PBS. M. luteus (g), B. subtilis (h), S. aureus (i), E. coli (j), V. anguillarum (k), and A. hydrophila (l) incubated with rATP5A1. OD, optical density.

rATP5A1 causes membrane depolarization and permeabilization

Because rATP5A1 had antibacterial activity, we tentatively tested its mode of action. The membrane depolarization activity of rATP5A1 was assayed using DiSC₃-5. As shown in Fig. 7A, the fluorescence intensities of all the bacterial cells (M. luteus, B. subtilis, S. aureus, E. coli, V. anguillarum, and A. hydrophila) treated with rATP5A1 increased significantly compared with control. This indicates that rATP5A1 caused depolarization of the bacterial plasma membrane. Flow cytometry revealed that few of the nontreated bacterial cells (M. luteus, B. subtilis, S. aureus, E. coli, V. anguillarum, and A. hydrophila) showed a PI fluorescent signal (Fig. 7B and Supplemental Fig. S3), indicating that they had intact and viable cell membranes. By contrast, a significant (P < 0.001) proportion of the cells treated with rATP5A1 displayed a fluorescent signal, and the number of cells with a fluorescent signal increased with increasing doses of rATP5A1, indicating that their cell membranes were no longer intact and became permeable to PI. These suggest that rATP5A1 was able to disrupt the bacterial membranes by a combined action of membrane depolarization and permeabilization.

Figure 7.

rATP5A1 was able to disrupt the bacterial membranes by a membranolytic mechanism including a combined action of membrane depolarization and membrane permeabilization. A) The rATP5A1 caused depolarization of the bacterial plasma membrane. B) The effects of rATP5A1 on the membrane integrity of M. luteus, B. subtilis, S. aureus, E. coli, V. anguillarum, and A. hydrophila, cells analyzed by flow cytometry. All data are expressed as means ± sd; n = 3. The data are from 3 independent experiments performed in triplicate. The bars represent means ± sd. **P < 0.01, ***P < 0.001.

ATP5A1 is involved in antibacterial defense of early embryos

The protein concentration of the extracts prepared from 64- and 128-cell embryos was ∼5 mg/ml. The embryonic extracts showed conspicuous antibacterial activity against the gram-positive bacteria M. luteus, B. subtilis, and S. aureus and the gram-negative bacteria E. coli, V. anguillarum, and A. hydrophila. Interestingly, this inhibition of bacterial growth activity was significantly reduced by the preincubation of the extracts with ATP5A1Ab, but not by the preincubation with AcAb (Fig. 8A). These suggest that ATP5A1 was the molecule directly associated with the antibacterial activity of the embryo extracts.

Figure 8.

Antibacterial activity of ATP5A1 in vitro and in vivo. A) Antimicrobial activities of the embryo extract against M. luteus, B. subtilis, S. aureus, E. coli, V. anguillarum, and A. hydrophila. B) The early (8-cell stage) embryos were first microinjected with PBS, BSA, AcAb, ATP5A1Ab, rATP5A1, and rATP5A1 plus AcAb or rATP5A1 plus ATP5A1Ab and then challenged by injection with live A. hydrophila. The development of the embryos was observed, and the cumulative mortality rate was calculated at 24 h after bacterial injection. C) The reduced antibacterial activity of the atp5a1-MO-knockdown embryos. a) Western blotting analysis of the efficacy of atp5a1-MO. β-Actin was used as control. Lane 1 and 2, the levels of ATP5A1 and β-actin in the embryos injected with control-MO (1) or atp5a1-MO (2). b) The 24-h mortality of atp5a1-MO-knockdown and control embryos after the challenge with live A. hydrophila. D) PCR analysis of A. hydrophila 16s rRNA gene. The embryos injected with A. hydrophila only (a), AcAb (b), or ATP5A1Ab (c) followed by injection with A. hydrophila. All data are expressed as means ± sd (n = 3). The data are from 3 independent experiments performed in triplicate. The bars represent means ± sd. *P < 0.05, **P < 0.01. A. h, A. hydrophila; con, control; Ex, embryo extract; M, marker.

To examine whether ATP5A1 played the same role in vivo, 8-cell embryos were each microinjected with ATP5A1Ab to block ATP5A1 action, followed by injection with live A. hydrophila (pathogenic to D. rerio). As shown in Fig. 8B, the majority (∼90%) of the embryos injected with PBS, BSA, AcAb, ATP5A1Ab, rATP5A1, rATP5A1 plus AcAb, or rATP5A1 plus ATP5A1Ab developed normally, with a 24-h cumulative mortality rate of about 10%. Notably, the challenge with live A. hydrophila resulted in a significant increase in the mortality rate of the embryos injected with ATP5A1Ab (resulting in a decrease of ATP5A1), with a 24-h mortality rate of ∼87.7%, whereas the same challenge caused only ∼68.5, 69.2, and 69.3% mortality at 24-h in the embryos injected with either PBS, BSA, or AcAb alone. This indicates that blocking of ATP5A1 action in the embryos was able to reduce their anti–A. hydrophila activity. In comparison, the 24-h mortality rate of the embryos injected with rATP5A1 (resulting in an increase of ATP5A1) was reduced to ∼54.4%, which was markedly lower than that of the embryos injected with AcAb or BSA or PBS alone, suggesting that the increased amount of ATP5A1 contributed to the protection of the embryos against A. hydrophila attack. Moreover, the protecting activity of rATP5A1 against A. hydrophila in the embryos was apparently reduced by the coinjection with ATP5A1Ab, but not by the coinjection with AcAb (24-h mortality rate being ∼85.9% vs. ∼56.7%), suggesting the specificity of ATP5A1 antibacterial activity. In addition, we show that the synthesis of ATP5A1 was reduced in the embryos microinjected with atp5a1-MO (Fig. 8Ca), indicating a successful knockdown of ATP5A1. As shown in Fig. 8Cb, the 24-h mortality of the atp5a1-MO–knockdown embryos increased up to 79%, contrasting with the 63% mortality of the control embryos, after the challenge with live A. hydrophila. This provided additional evidence supporting that ATP5A1 played a protective role in the early embryos. These data together suggest that ATP5A1 was an antibacterial factor capable of protecting zebrafish early embryos from bacterial infection.

To prove the killing of live A. hydrophila by the embryos, PCR was performed to amplify a specific region of A. hydrophila 16S rRNA gene. As shown in Fig. 8D, no band was observed in the control sample (from embryos without injected A. hydrophila), but intense bands were found in the embryos collected soon after the bacterial injection (0 h). The band intensities apparently decreased with time (at 12 and 24 h), suggesting the lysis of the bacterium by the embryos (Fig. 8Da). Similarly, the microinjection of AcAb into the embryos failed to suppress the reduction of the band intensity at 12 and 24 h (i.e., A. hydrophila lysis continued in the embryos) (Fig. 8Db), whereas the injection of ATP5A1Ab into the embryos clearly suppressed the decrease in the band intensity during the initial 12 and 24 h (Fig. 8Dc) (i.e., little A. hydrophila lysis took place in the embryos). These show that ATP5A1 was indeed associated with the killing of live A. hydrophila by the embryos.

We then wondered whether the embryos contained a sufficient endogenous amount of ATP5A1 to inhibit the bacterial growth. ELISA revealed that the contents of ATP5A1 in each of the newly fertilized eggs and 12- and 24-h embryos were about 82.7, 74.2, and 65.1 μg/ml, respectively, and were all markedly higher than 40 μg/ml rATP5A1, a concentration necessary to significantly inhibit the growth of the gram-positive and the gram-negative bacteria in vitro. Thus, the endogenous concentration of ATP5A1 in each egg/embryo was sufficient to kill potential pathogens in vivo, at least at the initial 24 hpf.

N-terminal 1–65 residues are critical for antibacterial activity of ATP5A1

The proteins A_1–65, A_66–134, and A_135–551 expressed in E. coli were purified by affinity chromatography on a Ni-nitrilotriacetic acid resin column. The purified recombinant proteins, rA_1–65, rA_66–134, and rA_135–551, each yielded a single band of 12.08, 11.27, and 49.4 kDa on SDS-PAGE after Coomassie Blue staining, matching the expected sizes, respectively (Fig. 9A, B). Antibacterial assay showed that rA_1–65 still retained an antibacterial activity against all the bacteria tested (Fig. 9C). By contrast, both rA_66–134 and rA_135–551 almost completely lost the antibacterial activity against the bacteria (Supplemental Fig. S4). In addition, rA_1–65 also possessed the affinity to LTA and LPS (Fig. 9D), agreeing with the fact that it retained antimicrobial activity, whereas both rA_66–134 and rA_135–551 lost the affinity to LTA and LPS (Fig. 9E, F), agreeing with the fact that they had little antibacterial activity. This suggests that the antibacterial activity of rA_1–65 (and rATP5A1) was related to its capacity to bind to LTA and LPS. These show that the N-terminal 65 residues of ATP5A1 were the core region indispensable for its antibacterial activity and affinity to LTA and LPS as well.

Figure 9.

Diagram showing zebrafish ATP5A1 truncation, SDS-PAGE, and Western blotting of rA_1–65, rA_66–134, and rA_135–551 and their antibacterial activities and binding to LPS and LTA. A) Diagram showing zebrafish ATP5A1 truncation. B) SDS-PAGE and Western blotting of rA_1–65 (a), rA_66–134 (b), and rA_135–551 (c). Lane M, marker; lane 1, negative control (empty vector pET28a); lane 2, total cellular extracts from E. coli BL21 containing expression vector pET28a/a1-65 (a), pET28a/a66-134 (b), and pET28a/a135-551 (c) before induction; lane 3, total cellular extracts from IPTG-induced E. coli BL21 containing expression vector pET28a/a1-65 (a), pET28a/a66-134 (b), and pET28a/a135-551 (c); lane 4, purified rA_1–65 (a), rA_66–134 (b), and rA_135–551 (c); lane 5, Western blotting of rA_1–65 (a), rA_66–134 (b), and rA_135–551 (c). COOH, free carboxyl end group; H₂N, free amino end group. C) Antibacterial activity of rA_1–65 against M. luteus, B. subtilis, S. aureus, E. coli, V. anguillarum, and A. hydrophila. D–F) Interaction of rA_1–65 (D), rA_66–134 (E), and rA_135–551 (F) with LTA and LPS revealed by ELISA. G) The in vivo bioactivity of rA_1–65, rA_66–134 and rA_135–551. The early (8-cell stage) embryos were first microinjected with PBS, BSA, AcAb, ATP5A1Ab, rA_1–65, rA_66–134, and rA_135–551 and then challenged by injection with live A. hydrophila. The development of the embryos was observed, and the cumulative mortality rate was calculated at 24 h after injection. OD, optical density. All data were expressed as means ± sd (n = 3). The data are from 3 independent experiments performed in triplicate. The bars represent the means ± sd. *P < 0.05, **P < 0.01.

We also tested the antibacterial role of rA_1–65 in vivo. When 8-cell embryos were each microinjected with rA_1–65, rA_66–134, or rA_135–551, about 90% of the embryos developed normally. The challenge with live A. hydrophila caused a significant increase in the mortality rate of the embryos injected with PBS, BSA, rA_66–134, rA_135–551, or AcAb, with a 24-h cumulative mortality rate of ∼65.3, 63.3, 62.0, 67.3, and 62.7%, respectively, whereas the same challenge induced only ∼42% cumulative mortality at 24 h in the embryos injected with rA_1–65. Moreover, the embryo-protecting role of rA_1–65 was counteracted by the coinjection with ATP5A1Ab, but not by coinjection with AcAb (Fig. 9G). It was thus clear that rA_1–65, with antimicrobial activity in vitro, also showed antimicrobial activity in the embryos, but rA_66–134 and rA_135–551, with no antibacterial activity in vitro, did not.

Conservation of antibacterial activity of N-terminal 65 residues

As ATP5A1 was highly conserved during evolution, we thus tested whether the N-terminal 65 residues of ATP5A1 from other animals other than zebrafish had antibacterial activity. The peptides rMA_1–65 and rCA_1–65 representing the N-terminal 65 residues of mouse and ascidian ATP5A1 were expressed in E. coli (Fig. 10A) and used to examine the antibacterial activity and affinity to LTA and LPS. As shown in Fig. 10B, C, the growth of all the bacteria tested was inhibited by rMA_1–65 and rCA_1–65 in a dose-dependent manner. In addition, rMA_1–65 and rCA_1–65 also possessed the affinity to LTA and LPS (Fig. 10D). It was clear that the N-terminal 1–65 residues of both ascidian and mouse ATP5A1 proteins, like that of zebrafish ATP5A1, exhibited a wide spectrum of antibacterial activity against both gram-positive and gram-negative bacteria. This suggests that the antibacterial activity of the N-terminal 65 residues of ATP5A1 (and ATP5A1 itself, by extrapolation) was conserved during animal evolution.

Figure 10.

SDS-PAGE and Western blotting of rMA_1–65 and rCA_1–65 and their antibacterial activities and binding to LTA and LPS. A) SDS-PAGE and Western blotting of recombinant protein rMA_1–65 (a) and rCA_1–65 (b). Lane M, marker; lane 1, negative control (empty vector pET28a); lane 2, total cellular extracts from E. coli BL21 containing expression vector pET28a/ma1-65 (a) and pET28a/ca1-65 (b) before induction; lane 3, total cellular extracts from IPTG-induced E. coli BL21 containing expression vector pET28a/ma1-65 (a) and pET28a/ca1-65 (b); lane 4, purified rMA_1–65 (a) and rCA_1–65 (b); lane 5, Western blotting of rMA_1–65 (a) and rCA_1–65 (b). B, C) Antibacterial activity of rMA_1–65 (B) and rCA_1–65 (C) against M. luteus, B. subtilis, S. aureus, E. coli, V. anguillarum, and A. hydrophila. D) Binding of rMA_1–65 and rCA_1–65 to LTA and LPS. OD, optical density. All data are expressed as means ± sd (n = 3). The data are from 3 independent experiments performed in triplicate. The bars represent means ± sd.

DISCUSSION

Previous studies have shown that ATP5A1 plays multiple roles, including involvement in the generation of ATP, pathogenesis/progression of carcinoma, and the aging process. However, our understanding of the biologic functions of ATP5A1 remains rather poor and incomprehensive. A new finding of our study is that ATP5A1 is not only a pattern recognition receptor, capable of identifying LTA and LPS, but also an effector molecule, capable of inhibiting the growth of both gram-positive and gram-negative bacteria. Potential modes of actions of antibacterial proteins include binding to or inserting into bacterial membrane, which have fatal depolarization of the normally polarized membrane, formation of physical pores, scrambling of the usual distribution of lipids between the leaflets of the bilayer, and damage to critical intracellular targets. We show that ATP5A1 functions by a membranolytic mechanism including interaction with bacterial membrane via LTA and LPS (especially lipid A), membrane depolarization, and membrane permeabilization. These indicate that ATP5A1 may play essential roles in fish nonspecific immune defenses by its ability to recognize potential pathogens via interacting with LTA and LPS and its ability to kill them by a combined action of membranolytic mechanisms including membrane depolarization and membrane permeabilization, thereby preventing or limiting bacterial infection.

Another important finding of our study is the identification of ATP5A1 as a newly characterized maternal immune-relevant factor functioning in the eggs and early embryos of zebrafish. First, ATP5A1 is an LTA- and LPS-binding protein abundantly stored in the eggs and embryos. Second, the antibacterial activity of the embryonic extracts is clearly correlated with ATP5A1 level because it is markedly abolished by the preincubation of the extracts with ATP5A1Ab, but not by the preincubation with AcAb . Importantly, an injection of exogenous rATP5A1 into early embryos apparently promotes their resistance against pathogenic A. hydrophila challenge, and this pathogen-resistant activity of the embryos is significantly reduced by the coinjection of ATP5A1Ab (but not by the coinjection of AcAb) or by the knockdown with atp5a1-MO. Finally, the level of ATP5A1 in the egg/embryo is significantly higher than the concentration (40 μg/ml) necessary to kill potential pathogens. Collectively, these suggest that ATP5A1 can protect the early embryos in vivo from the attacks by pathogens like A. hydrophila. However, it needs to be pointed out that ATP5A1 may function only when the pathogens have penetrated the embryos because it is localized in the cytoplasm.

Fishes are the basal vertebrates that have an innate and adaptive system comparable with that of mammals. Currently, little is known about the role of ATP5A1 in the development of mammalian species, though it has been shown that a defect of atp5a1 in mice resulted in death in utero (21). Given the conservation of the antibacterial activity of the N-terminal 65 residues of ATP5A1 throughout the chordate evolution, we thus believe that mammalian ATP5A1 plays a similar role in early development. However, this demands further study in the future.

Proteins are built as chains of amino acids, which are closely correlated with their functions. It is notable that the N-terminal 65 residues are the critical sequence contributing to the antibacterial activity of ATP5A1. It is also notable that the antibacterial activity against the gram-positive and gram-negative bacteria and the affinity to LTA and LPS coexist in both rA_1–65 and rATP5A1, but neither the antibacterial activity nor the affinity to LTA and LPS are present in rA_66–134 and rA_135–551, suggesting a correlation between antibacterial activity and LTA- and LPS-binding activity. Probably, this correlation forms the basis for the functional sites of rATP5A1 and rA_1–65 simultaneously involved in multiple activities, including interacting with the bacterial signature molecules LTA and LPS and destabilizing/disrupting the bacterial cell membranes.

In summary, this study demonstrates for the first time that ATP5A1 is a maternal immune factor functioning in the eggs and early embryos of zebrafish and is capable of not only recognizing LTA and LPS but also killing the gram-positive and gram-negative bacteria via interacting with the bacteria and disrupting their plasma membranes. Our study also shows that the N-terminal 65 residues are critical for the antibacterial activity of ATP5A1 proteins in ascidian, zebrafish, and mouse, manifesting the high conservation of the antibacterial activity of ATP5A1 during animal evolution. This work also provides a new viewpoint for understanding the biologic roles of ATP5A1, which is ubiquitously present in animals.

Glossary

3D

3-dimensional

A_{1–65/66–134/135–551}

residues 1–65/66–134/135–551 of Danio rerio ATP5A1

AcAb

anti–β-actin antibody

ATCC

American Type Culture Collection

ATP5A1

ATP synthase α subunit

ATP5A1Ab

anti-ATP5A1 antibody

atp5a1-MO

morpholino for atp5a1

BSA

bovine serum albumin

CA

Ciona intestinalis ATP5A1

control-MO

control morpholino

DiSC₃-5

3,3′-dipropylthiacarbocyanine iodide

eGFP

enhanced green fluorescent protein

HEK

human embryonic kidney

His

histidine

hpf

hours past fertilization

IPTG

isopropyl β-d-thiogalactoside

LTA

lipoteichoic acid

MA

Mus musculus ATP5A1

MALDI/TOF MS

matrix-assisted laser desorption/ionization–time-of-flight mass spectrometry

PBST

PBS containing 0.5% Triton X-100

PI

propidium iodide

TEM

transmission electron microscopy

TRX

thioredoxin

WISH

whole-mount in situ hybridization

AUTHOR CONTRIBUTIONS

S. Ni and S. Zhang designed the experiments and wrote the paper; S. Ni, Y. Zhou, Y. Chen, and X. Du performed the experiments; S. Ni and S. Zhang analyzed the data; and all authors approved the final version of the manuscript.

Supplementary Material

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.