Cell culture

Acanthamoeba, human corneal epithelial (HCE) cells, Fusarium solani, Pseudomonas aeruginosa, and Staphylococcus aureus were prepared as previously described [21]. Briefly, both of the non-pathogenic (ATCC 30011; yeast culture isolate) and pathogenic (ATCC 30868; human corneal isolate) strains of Acanthamoeba castellanii Castellani were purchased from the American Type Culture Collection (ATCC). A keratitis sample isolated from an anonymized patient at the Department of Ophthalmology, Kyung Hee University Hospital (Seoul, Republic of Korea) was sent to our lab to confirm whether the keratitis sample was of Acanthamoebic origin. The clinical isolate shared 98% sequence homology with Acanthamoeba spp. based on the complete 18s rDNA sequence. All three Acanthamoeba strains were axenically cultured in PYG medium at 25°C and after 5 days, the culture media was collected. HCE cells were incubated at 37°C with 5% CO₂ in endothelial cell growth media (KGM BulletKit; Lonza, Portsmouth, NH, USA). P. aeruginosa and S. aureus were cultured in Brain Heart Infusion media at 37°C, while F. solani was cultured in Sabouraud Dextrose (SD) media at 25°C. The culture media of F. solani, P. aeruginosa, and S. aureus were harvested after cultivating for 1 day.

Cloning and peptide antibody production of chorismate mutase

Gene sequence of the A. castellanii Neff chorismate mutase (GenBank accession no: XM_004352878) was used to design the primers for chorismate mutase, which were subsequently used to amplify the coding sequence of A. castellanii Castellanii chorismate mutase by polymerase chain reaction. The primer information is as follows: 5’-ATGCGCTTCCTGCTCGCCTT-3’ (forward), and 5’-TCAATCGGCGTGCTGGACGG-3’ (reverse). The peptide sequence for the immunogen is as follows: NH₂-C-RLQVETLNSEFNAGLPVPVQHAD-COOH. Synthesized peptide and the antibody raised against the peptide were purchased from AbFRONTIER (AbFRONTIER, Seoul, Korea). Briefly, two NZW rabbits were injected intradermally with 1.0 mg of peptide-KLH conjugates in complete Freund’s adjuvant on days 0, and 0.5 mg of peptide in incomplete Freund’s adjuvant on days 28, 42, and 56. The animals were bled on days 35 and 49. After the 3^rd immunization, antisera titer was assessed using an indirect ELISA with the peptide-BSA conjugates as coating antigens until the titer plateaued. After the last immunization, blood samples were acquired by cardiac puncture from each rabbit.

Acanthamoeba specificity of the chorismate mutase antibody by western blot

Acanthamoeba-specificity of the chorismate mutase was confirmed using western blot as previously described [21]. Protein samples from the whole cell lysates and the culture media of Acanthamoeba spp., HCE cells, F. solani, P. aeruginosa, and S. aureus were prepared using a lysis buffer. After determining the concentration of each protein sample, 20 μg of cell lysates and 5 μg of conditioned media were resolved by SDS-PAGE and transferred onto a nitrocellulose membrane. After blocking the membrane with 5% skim milk in TBST for 1 h at RT, membrane was incubated overnight at 4°C with the chorismate mutase polyclonal antibody (1:1000). Membrane was washed with TBST, then incubated with horseradish peroxidase (HRP)-conjugated anti-rabbit antibody (1:5000) (Sigma-Aldrich, St. Louis, MO, USA) at RT for 1 h. Bands were developed using enhanced chemiluminescence (ECL; Thermo Fisher, MA, USA) on an x-ray film in the darkroom.

Immunocytochemistry

HCE cells were seeded at a density of 3 × 10⁵ cells/well on a six-well plate and incubated in KGM BulletKit media at 37°C for 24 h. After incubation, A. castellanii (ATCC 30868) trophozoites were added to HCE cells at a density of 5 × 10⁵ cells/well, and incubated at 37°C for 4 h. After incubation, HCE cells and A. castellanii were fixed with 100% methanol for 5 min at RT and permeabilized using 0.2% Triton X-100 in phosphate-buffered saline (PBST) for 10 min at RT. After washing with PBS, cells were blocked using 1% bovine serum albumin in PBS for 30 min at RT. The cells were incubated overnight at 4°C with chorismate mutase antibody (1:200), washed with PBST, and subsequently incubated with mouse-anti-rabbit IgG-CFL-488 antibody (1:400) (Sigma-Aldrich, St. Louis, MO, USA) for 1 h at RT. Cells were washed with 1X PBS and stained with VECTASHIELD^® Mounting Medium with DAPI (Abcam, Burlingame, CA, USA). Images were acquired using fluorescent microscopy (Leica DMi8, Wetzlar, Germany).

Enzyme-linked immunosorbent assay (ELISA)

IgG antibody levels specific to A. castellanii were determined by ELISA. A checkerboard titration of each antigen and serum were designed for the optimal antigen concentration (100 to 0.0001 μg/μl) and serum dilutions (1:5 to 1:5000). Cell lysates and the conditioned media of Acanthamoeba were used as coating antigens. Briefly, in 96-well microplates, each well was coated with the antigen dissolved in carbonate coating buffer (0.1 M sodium carbonate, pH 9.5) at 4°C overnight. Plates were washed three times with PBS containing 0.05% Tween 20 (PBST) and blocked with 0.2% gelatin for 1 h at 37°C. All antibodies were diluted in PBS. Diluted samples were incubated at 37°C for 1 h and plates were washed. HRP-conjugated goat anti-rabbit IgG (Cusabio Co. Ltd, Wuhan, China) was added at 1:1000 dilution and incubated for 1 h at 37°C. O-phenylenediamine (OPD; Zymed, San Francisco, CA) was dissolved in citrate-phosphate buffer (pH 5.0) with 0.03% H₂O₂ and subsequently used as substrate. The optical density values at 450 nm were read using an EZ Read 400 microplate reader (Biochrom Ltd., Cambridge, UK). Sera of naïve rabbits were used as negative control.

Identification of chorismate mutase in A. castellanii

Secretory proteins isolated from both non-pathogenic (Fig 1A) and pathogenic (Fig 1B) A. castellanii were resolved by SDS-PAGE and compared for analysis. As illustrated in Fig 1, several protein bands that were specifically present only in the pathogenic Acanthamoeba strain were observed. One such protein, as indicated by the arrow in Fig 1B, was selected for identification. Mass spectrometry and NCBI database search results revealed that the selected protein showed the highest similarity with the chorismate mutase of A. castellanii Neff. In this study, the full-length open reading frame of chorismate mutase from pathogenic A. castellanii Castellani was identified by polymerase chain reaction (GenBank accession number MN630520). The gene consists of 585 bp and encodes 194 amino acids with a calculated mass 21.34 kDa. Protein homology search results revealed that this protein showed 96.9% similarity with chorismate mutase 1 (putative) of A. castellanii Neff and 71.6% similarity with chorismate mutase 2 (subfamily protein) of A. castellanii Neff (Fig 2 and Table 1). Peptide analysis was performed using the SignalP 4.1 server database, which predicted the presence of the signal peptide at amino acid positions 1 to 19.

Fig 1

Secretory proteins of A. castellanii.

Secretory proteins between non-pathogenic Acanthamoeba (ATCC 30011) (A) and pathogenic Acanthamoeba (ATCC 30868) (B) were compared by SDS-PAGE. M; protein size marker, lane 1; conditioned media of Acanthamoeba. An arrow indicates a specific protein band of pathogenic Acanthamoeba origin.

Fig 2

Alignment of the chorismate mutase amino acid sequences of Acanthamoeba spp..

Aligned amino acid sequences of A. castellanii Neff chorismate mutase 1 (ELR23397), A. castellanii Neff chorismate mutase 2 (ELR15898), and A. castellanii Castellani chorismate mutase (MN630520) were compared. ClustalX multiple sequence alignment was used to produce the alignment, and the degree of conservation is represented by different shading. AcN; Acanthamoeba castellanii Neff, AcC; Acanthamoeba castellanii Castellani, CM; chorismate mutase.

Table 1

Identities of chorismate mutase amino acids as determined by protein BLAST.

Name	Accession No.	Length (aa)	Identities
AcC-CM	MN630520	194	194 (100%)
AcN-CM1	ELR23397	195	189 (96.9%)
AcN-CM2	ELR15898	197	144 (71.6%)
Fo-CM	EWY95915	223	34 (13.7%)
Fo-CM2	EWY95914	266	37 (13.3%)
Pa-CM	OHQ59631	185	36 (16.0%)
Pa-CM2	KJJ16793	365	37 (9.3%)
Sa-CM	OHS92057	363	30 (6.7%)

AcC; Acanthamoeba castellanii Castellani, CM; chorismate mutase, AcN; Acanthamoeba castellanii Neff, Fo; Fusarium oxysporum, Pa; Pseudomonas aeruginosa, Sa; Staphylococcus aureus.

Production of chorismate mutase antibody

To assess the applicability of the chorismate mutase protein for AK diagnosis, we produced a polyclonal peptide antibody against the chorismate mutase of A. castellanii. Based on the hydrophobicity and antigenicity analysis profiles, two highly immunogenic epitopes were identified at amino acid position 127–134 and 172–194. Amino acid sequences of chorismate mutase from Acanthamoeba was compared with that of other causes of keratitis such as Fusarium spp., Pseudomonas aeruginosa, and Staphylococcus aureus. Protein homology search results revealed that chorismate mutase of Acanthamoeba showed 13.7%, 13.3%, 16.0%, 9.3%, and 6.7% similarity with chorismate mutase 1 or 2 of F. oxysporum, P. aeruginosa, and S. aureus, each respectively (Fig 3 and Table 1). Based on the amino acid homology search, amino acids at positions 172–194 were selected for peptide antibody production (boxed area in Fig 3).

Fig 3

Alignment of the chorismate mutase amino acid sequences from different organisms.

Aligned amino acid sequences of F. oxysprum chorismate mutase (EWY95915), F. oxysprum chorismate mutase 2 (EWY95914), S. aureus chorismate mutase (OHS92057), P. aeruginosa chorismate mutase (OHQ59631), P. aeruginosa chorismate mutase 2 (KJJ16793), and A. castellanii Castellani chorismate mutase (MN630520) were compared. The boxed sequence corresponds to the antigen peptides used to raise the anti-chorismate mutase polyclonal antibody. Fo; Fusarium oxysporum, Sa; Staphylococcus aureus, Pa; Pseudomonas aeruginosa.

Confirming the specificity of a chorismate mutase peptide antibody

The specificity of the chorismate mutase antibody was demonstrated by western blot analysis using cell lysates and conditioned media of HCE cells, non-pathogenic Acanthamoeba, pathogenic Acanthamoeba, Acanthamoeba spp. isolated from a clinical sample, F. solani, P. aeruginosa, and S. aureus (Fig 4). Western blot results revealed that the chorismate mutase antibody did not react with HCE cell lysate and conditioned media (lane 1 of Fig 4A and 4B) but reacted with Acanthamoeba spp. cell lysate and conditioned media (lanes 2 to 4 of Fig 4A and 4B). The estimated sizes of the chorismate mutase protein and the secretory chorismate mutase in conditioned media were 21.34 kDa and 19.34 kDa, respectively. While both of these were detected from the conditioned media of Acanthamoeba spp., their presence was not confirmed in HCE cells, F. solani, P. aeruginosa, and S. aureus lysates (Fig 4B). Several non-specific bands were detected in the P. aeruginosa cell lysates, S. aureus cell lysate and its conditioned media (lane 7 of Fig 4A and 4B).

Fig 4

Specific antibody response of chorismate mutase.

The differential diagnosis of Acanthamoeba keratitis was confirmed by western blot analysis of cell lysates (A) and conditioned media (B) using the polyclonal peptide antibody raised against chorismate mutase of A. castellanii. Lane 1: HCE cells, lane 2: non-pathogenic Acanthamoeba, lane 3: pathogenic Acanthamoeba, lane 4: clinical isolate of Acanthamoeba, lane 5: F. solani, lane 6: P. aeruginosa, and lane 7: S. aureus.

To further validate the chorismate mutase antibody specificity, immunocytochemistry was performed using HCE cells were cultured with A. castelannii (Fig 5). As expected, DAPI staining revealed the cellular nucleic acid contents of HCE cells (Fig 5B), whereas HCE cells remained unstained when incubated with chorismate mutase antibody (Fig 5C). On the contrary, strong interaction between the chorismate mutase antibody and Acanthamoebae trophozoites was observed (Fig 5C). Faint DAPI staining was observed from Acanthamoeba, which could be attributed to its smaller size in comparison to HCE cells. A. castellanii cysts were also detected with chorismate mutase antibody (Fig 6A), however, F. solani, P. aeruginosa, and S. aureus were not detected with that antibody (Fig 6B–6D). Overall, the chorismate mutase antibody demonstrated its potential for selective and differential detection of Acanthamoeba via immunocytochemistry.

Fig 5

Immunocytochemistry using the A. castellanii chorismate mutase antibody.

HCE cells cultured with A. castellanii trophozoites for 4 h were visualized under a fluorescent microscope. Bright-field (A), DAPI-stained (B), indirect immunocytochemical labeling (C), and merged images (D) were acquired at 400x magnification.

Fig 6

Confirming the species-specificity of the A. castellanii chorismate mutase antibody using immunocytochemistry.

A. castellanii cysts (A), F. solani (B), P. aeruginosa (C), and S. aureus (D) were visualized under a fluorescence microscope. Bright-field, DAPI-stained, and indirect immunocytochemical labeling images were acquired at 400x magnification.

ELISA titer of anti-chorismate mutase antibody

As seen in Fig 7, serological analysis of immune sera indicated that a rabbit injected with A. castellanii chorismate mutase possessed Acanthamoeba-specific IgG. High IgG antibody titers were observed from both A. castellanii cell lysates (1/5000) and its conditioned media (1/5000) (Fig 7A and 7B). In addition, Acanthamoeba-specific IgG was demonstrated to be highly sensitive, with a detection limit as low as 0.1 ng for both Acanthamoeba cell lysate and the conditioned media (Fig 7C and 7D).

Fig 7

A. castellanii chorismate mutase-specific IgG response.

The chorismate mutase polyclonal antibody was titrated using A. castellanii cell lysates (A) and conditioned media (B). Sensitivity of the antibody was also determined using cell lysates (C) and conditioned media (D). positive; immunized serum, negative; non-immuned serum.