Construction of recombinant antibody expression gene
We genetically synthesized codon-optimized VH and VL genes of MEDI3902 and introduced each DNA into the pcDNA3.1(-) vector, which is a widely used vector for protein or antibody expression in mammalian cells. Prior to the VH or VL sequence, we added the Kozak sequence followed by the IL-2 signal sequence to increase the expression yield. We prepared each plasmid and used them for transient co-transfections of HEK293F cells (Fig. 1A).
Expression and purification of recombinant antibody
We expressed rAb through a suspension culture of HEK293F cells, which has been widely used for large-scale production of proteins, including antibodies. We injected two plasmids, pcDNA3.1::anti-P. aeruginosa H chain and pcDNA3.1::anti-P. aeruginosa L chain, with PEI and cultured the cells. The supernatant was collected and purified using protein A (PA) affinity beads, and the buffer was changed to PBS using ultrafiltration. Next, we performed SDS-PAGE analysis with or without reduction to resolve the structure of the generated rAb (Fig. 1B). After denaturing the sample using DTT and heating, two bands were observed, which were corresponded to the H chain and L chain, whose amino acid-based calculated size was 51.9 kDa and 25.7 kDa, respectively. In the case of the native sample without denaturation, no band around 51.9 kDa or 25.7 kDa was observed, indicating that excess H and L chains were not present. Only the expected full-sized Ab was present in the sample, and the rAb was successfully expressed with correct folding. It is worth noting that almost no extra bands were observed in the sample, indicating that the purification was very high only after PA affinity purification without additional size-exclusion chromatography, which can cause the loss of Ab. We confirmed that 3.36 mg of purified rAb was obtained per 150 mL of culture.
Antigen-binding efficiency of recombinant antibody
We confirmed the antigen-binding efficiency of rAb against three P. aeruginosa strains: P. aeruginosa UCBPP-PA14, ATCC 27853, and ATCC BAA-2108. P. aeruginosa UCBPP-PA14 is a susceptible strain, isolated from a human burn patient [31]. ATCC 27853 is a susceptible strain, isolated from a hospital blood specimen [32]. BAA-2108 is a multidrug-resistant P. aeruginosa strain that was isolated from a cystic fibrosis patient during a clinical test for evaluating the efficacy of aerosolized tobramycin [33]. We cultured the three strains and diluted them. We plated the cells on a 96-well plate and performed indirect ELISA using rAb as the primary antibody. At that time, we investigated the antigen-binding efficiency of both rAb and two commercial anti-P. aeruginosa Abs, monoclonal Ab (mAb) and polyclonal Ab (pAb). We used HRP-conjugated goat anti-human IgG-Fc antibody, HRP-conjugated goat anti-mouse IgG2a antibody, and HRP-conjugated goat anti-rabbit IgG antibody as a secondary antibody for rAb, mAb, and pAb, respectively (Fig. 2A). As a result, the signals of total nine ELISA platforms with different antibodies and antigens increased in an antigen-concentration-dependent-manner, indicating the antigen-binding efficiency of each antibody against each strain (Fig. 2B and Table 2). When we compared the responses of three antibodies against P. aeruginosa UCBPP-PA14, two commercial Abs showed a broader detection range than rAb, whereas the EC50 values were similar. When we used ATCC 27853 as an antigen, pAb showed a very low LOD value (3.5 CFU), indicating the usefulness of this Ab for sensitive detection of ATCC 27853. In the case of ATCC BAA-2108, the LOD values of mAb and rAb were similar and the LOD value of pAb was higher than that of the other two Abs. Overall, among the three Abs, which showed antigen-binding efficiency, commercial pAb showed a higher response than mAb and rAb. Nonetheless, the sensitivities of these Abs against each antigen, except pAb against ATCC 27853, were not sufficient to detect clinical level; 1-104 CFU/mL 31.
Table 1
Sequences of variable domains in recombinant antibody.
|
VH
|
VL
|
Nucleotide
|
gaggtgcagctggtggaatctggcggcggacttgttcaacctggcggctctctgagactgag ctgtgccgcttccggcttcacctttagcagctacgccatggactgggtccgacaggctcctggc aaaggccttgaatgggtgtccgccatcaccatgtctggcatcaccgcctactacaccgacgac gtgaagggcagattcaccatcagccgggacaacagcaagaacaccctgtacctgcagatgaacagc
|
gccatccagatgacacagagccccagcagcctgtctgcctctgtgggagacagagtgaccatca cctgtagagccagccagggcatcagaaacgacctcggctggtatcagcagaagcctggcaagg cccctaagctgctgatctacagcgccagcacactgcagagcggagtgcctagcagattttctggc agcggctccggcaccgatttcaccctgaccatatctagcctgcagcctgaggacttcgccacc
|
Amino
acid
|
EVQLVESGGGLVQPGGSLRLSCAASGFTFSSYAMDWVRQAPGKGLEWVSAITM SGITAYYTDDVKGRFTISRDNSKNTLYLQMNSLRAEDTAVYYCAKEEFLPGTHYF YGMDVWGQGTTVTV
|
AIQMTQSPSSLSASVGDRVTITCRASQGIRNDLGWYQQKPGKAPKLLIYSAS TLQSGVPSRFSGSGSGTDFTLTISSLQPEDFATYYCLQDYNYPWTFGQGTKVEI
|
Table 2
EC50 and LOD values of Abs that were determined from the titration curves of indirect ELISA.
P. aeruginosa strain
|
Antibody
|
EC50 (CFU)
|
LOD (CFU)
|
UCBPP-PA14
|
pAb
|
3.08 ± 0.31 x 106
|
2.92 x 105
|
UCBPP-PA14
|
mAb
|
7.46 ± 3.27 x 105
|
4.69 x 105
|
UCBPP-PA14
|
rAb
|
n.d.
|
n.d.
|
ATCC 27853
|
pAb
|
5.03 ± 0.98 x 103
|
3.50 x 101
|
ATCC 27853
|
mAb
|
2.77 ± 0.37 x 105
|
4.01 x 104
|
ATCC 27853
|
rAb
|
1.01 ± 0.28 x 105
|
2.43 x 104
|
ATCC BAA-2108
|
pAb
|
1.30 ± 0.08 x 104
|
4.42 x 104
|
ATCC BAA-2108
|
mAb
|
2.15 ± 0.27 x 105
|
1.37 x 104
|
ATCC BAA-2108
|
rAb
|
2.91 ± 0.30 x 105
|
7.62 x 104
|
Sandwich ELISA for sensitive detection of P. aeruginosa
In the case of indirect ELISA, the directions of antigen molecules for attachment to the plate differ because of the random seeding of the antigen molecules to the plate. Thus, the epitope of the antigen can be attached to the plate. In sandwich ELISA, two antibodies, capturing antibody and detecting antibody, bind specifically to each epitope. Therefore, the binding capability between antigen and antibody in sandwich ELISA is higher than that in indirect ELISA, resulting in higher sensitivity. Moreover, the selectivity of sandwich ELISA is usually higher than that of indirect ELISA because two antibodies are used to “sandwich” the antigen, which has an advantage when complex samples are used, because only the antigen is specifically immobilized to the capturing antibody rather than the entire complexed sample to the plate. Thus, sandwich ELISA can be more versatile when used for detecting pathogens in complicated in vivo samples, such as food and blood. In addition, P. aeruginosa can form a biofilm when attached to the surface of a 96-well plate for indirect ELISA 32. Therefore, sandwich ELISA is more suitable than indirect ELISA for forming planktonic cells.
Based on these points, we performed indirect ELISA as well as sandwich ELISA. The most important step in sandwich ELISA is selecting the best pair for capturing and detecting Abs. Therefore, we first screened pairs from the combinations of pAb, mAb, and rAb. Each Ab was seeded onto a plate, and the wells were blocked. Afterwards, we added 30 to 107 CFU of ATCC 27853, UCBPP-PA14, ATCC BAA-2108 or PBS and washed the wells. Next, we added a capturing Ab, followed by an HRP-conjugated Ab, which binds to each capturing Ab (Fig. 3A). All pairs showed higher signals in the presence of antigen than in the absence of antigen (Fig. 3B-3D). Among them, four pairs, mAb–pAb, pAb–mAb, rAb-pAb, and rAb-mAb (in the order of capturing Ab–detecting Ab), showed a relatively higher signal to background ratio (S/B) than the other two pairs, mAb-rAb and pAb-rAb. Although the responses of mAb-rAb and pAb-rAb can be improved by optimizing the ELISA conditions and/or by using more appropriate HRP-conjugated Ab that has higher secondary antibody-binding affinity, we moved to the next step by using these four pairs because these selected pairs showed high enough responses with S/B against three pathogens.
We seeded various concentrations of antigen on the plate and performed a sandwich ELISA. Antigen-concentration-dependent responses were observed with the use of all antibody pairs against each antigen, and EC50 and LOD values were calculated (Table 3). When we used UCBPP-PA14 as an antigen, the EC50 values between pairs were similar (approximately 105 CFU), but the pAb-mAb pair showed the lowest LOD value of 8.5 CFU, which was significantly lower than others (103-104 CFU) (Fig. 4). In the case of ATCC 27853, the mAb-pAb pair showed the lowest EC50 (4.10 ± 0.39 x 103 CFU) and LOD (2.04 x 102 CFU) (Fig. 5). For ATCC BAA-2108, the rAb-pAb pair showed a lower EC50 (1.39 ± 0.22 x 103 CFU) and LOD (23 CFU) than the others (Fig. 6). It is worth noting that all LOD values using an rAb for detecting three strains of P. aeruginosa were below 103 CFU, indicating its usefulness for the sensitive detection of each pathogen.
Table 3
EC50 and LOD values of Abs that were determined from the titration curves of sandwich ELISA.
Antigen
|
Capturing antibody
|
Detecting antibody
|
EC50 (CFU)
|
LOD (CFU)
|
UCBPP-PA14
|
mAb
|
pAb
|
1.43 ± 0.44 x 105
|
1.39 x 104
|
UCBPP-PA14
|
pAb
|
mAb
|
4.62 ± 2.80 x 105
|
8.50 x 101
|
UCBPP-PA14
|
rAb
|
pAb
|
4.10 ± 1.68 x 105
|
3.34 x 103
|
UCBPP-PA14
|
rAb
|
mAb
|
2.00 ± 0.06 x 105
|
7.46 x 103
|
ATCC 27853
|
mAb
|
pAb
|
4.10 ± 0.39 x 103
|
2.04 x 102
|
ATCC 27853
|
pAb
|
mAb
|
4.56 ± 1.11 x 104
|
5.78 x 102
|
ATCC 27853
|
rAb
|
pAb
|
7.01 ± 0.79 x 103
|
7.85 x 102
|
ATCC 27853
|
rAb
|
mAb
|
8.21 ± 0.95 x 104
|
6.30 x 103
|
ATCC BAA-2108
|
mAb
|
pAb
|
9.19 ± 0.56 x 103
|
3.79 x 102
|
ATCC BAA-2108
|
pAb
|
mAb
|
1.20 ± 0.04 x 104
|
7.70 x 101
|
ATCC BAA-2108
|
rAb
|
pAb
|
1.39 ± 0.22 x 103
|
2.30 x 101
|
ATCC BAA-2108
|
rAb
|
mAb
|
1.63 ± 0.17 x 104
|
1.84 x 103
|