Salmonella infection, generally referred to as salmonellosis, is a gastrointestinal disorder caused by consuming Salmonella-contaminated foods [1]. Salmonella infection is typically linked to animal products such as poultry, egg, beef, and pork; however, other foods including fruits, vegetables, dried foods, infant formula, and pet foods may also be contaminated [2]. Based on the White–Kaufmann scheme, Salmonella is divided into two species, Salmonella bongori and Salmonella enterica, and S. enterica is further divided into six subspecies, i.e., S. enterica subsp. arizonae, S. enterica subsp. diarizonae, S. enterica subsp. enterica, S. enterica subsp. haoutenae, S. enterica subsp. indica, and S. enterica subsp. salamae [3, 4]. Three surface structure determinants, including lipopolysaccharide (LPS), flagella, and capsular polysaccharide, differentiate them into approximately 2,659 serotypes [5, 6]. Among these, approximately 1,547 serotypes emanate from S. enterica subsp. enterica [6]. Salmonella serotypes are primarily transmitted via animal products, such as meats and eggs, can infect both humans and animals, and are generally known as nontyphoidal Salmonella (NTS) serotypes [7]. Common NTS infection symptoms include diarrhea, fever, abdominal cramps, and vomiting, and most healthy people recover within a week without hospitalization and medical treatment [8]. However, these NTS serotypes can be invasive (defined as invasive nontyphoidal Salmonella, iNTS) and cause fatal bacteremia and systemic infection, resulting in life-threatening complications [9].

According to the culture-confirmed human infection survey conducted by the Laboratory-based Enteric Disease Surveillance (LEDS) system, the 20 most prevalent serotypes in decreasing order were Salmonella Enteritidis, Salmonella Newport, Salmonella Typhimurium, Salmonella Javiana, Salmonella I 4,[5],12:i:-, Salmonella Infantis, Salmonella Muenchen, Salmonella Montevideo, Salmonella Braenderup, Salmonella Thompson, Salmonella Saintpaul, Salmonella Heidelberg, Salmonella Oranienburg, Salmonella Mississippi, Salmonella Typhi, Salmonella Bareilly, Salmonella Berta, Salmonella Agona, Salmonella Paratyphi B var., and Salmonella Anatum, which were associated with 69.2% incidence of culture-confirmed salmonellosis in the United States of America (U.S.) in 2016 [10]. Interestingly, 360 other serotypes were responsible for 16.5% salmonellosis incidence [10]. These 20 prevalent serotypes have been closely associated with human infection worldwide for the last decade, although the pecking orders varied with time and location [10]. We aimed to develop an on-site rapid diagnostic method to identify the dominant serotypes usually transmitted between humans and animal-based foods. The devised method targets 18 of the top 20 serotypes reported by LEDS system and excluded S. Typhi and S. Paratyphi because they do not colonize or inhabit animals as reservoir, except higher primates [5, 7].

From an epidemiological perspective, the surveillance of environmental serotypes is crucial for monitoring the emergence of pathogenic serotypes and preventing their transmission through the food chain [11, 12]. A dot-blot technique is a straightforward and rapid diagnostics where a small amount of bacterial lysate is directly dotted on a membrane and subsequently probed with a specific antibody to detect the target antigen [13]. Dot-blot assays are substantially more sensitive than enzyme-linked immunosorbent assay (ELISA) because nitrocellulose and polyvinylidene difluoride (PVDF) membranes have better protein- and LPS-binding capacities than the polystyrene plates used in ELISA [14]. Since this technique does not require highly skilled personnel or the expensive equipment needed for traditional immunology-based diagnosis, dot-blot immunoassays are affordable in most laboratories [15]. To exploit dot-blot immunoassay for a simple and reliable diagnosis for prevalent Salmonella serotypes, we determined a combination of antibodies to identify multiple dominant NTS serotypes and improved the method so that, excluding pre-cultivation time, the assay could be completed in 3 h.

Bacterial Strains and Growth Conditions

Salmonella strains used in this study are listed in Tables S1 and S2. Other bacterial strains used include Escherichia coli DH5α (ATCC 53868), Vibrio cholerae ATCC 14033, Staphylococcus aureus ATCC 29213, and Bacillus cereus ATCC 14579. Gram-negative bacteria, i.e., Salmonella, E. coli, and V. cholera were cultivated in Luria–Bertani (LB) broth (Becton, Dickinson and Company, USA.) and gram-positive bacteria, i.e., S. aureus and B. cereus were cultivated in brain–heart infusion (BHI) broth (Becton, Dickinson and Company). All bacterial strains were cultivated at 37°C.

Antibodies

Ten different Salmonella antibodies were selected among commercially available antibodies and are listed in Table 1. The antibodies were produced in mouse or rabbit using diverse immunogens, including LPS (whole or partial), flagella, and whole cell preparation. Ab 3, Ab 5, Ab 6, and Ab 7 were rabbit polyclonal, whereas Ab 1, Ab 2, Ab 4, Ab 8, Ab 9, and Ab 10 were mouse monoclonal (Table 1). They were purchased from Abcam (USA.) and Invitrogen (Waltham, USA). The secondary antibodies used were goat anti-mouse IgG–HRP conjugate (#1706516; Bio-Rad Laboratories, Inc., USA), goat anti-rabbit IgG–HRP conjugate (#31460; Invitrogen), goat anti-mouse IgG–AP conjugate (A3562; Sigma-Aldrich, USA), and goat anti-rabbit IgG–alkaline phosphatase (A3687; Sigma-Aldrich).

Table 1 . Ten primary antibodies used in this study..

ID	Product	Immunogen	Host	Isotype
Ab 1	Salmonella Typhimurium LPS antibody [1E6] (ab8274)	S. Typhimurium LPS	Mouse (mAb)	IgG1
Ab 2	Salmonella LPS antibody [se-01] (ab243104)	S. Typhimurium LPS	Mouse (mAb)	IgG1
Ab 3	Salmonella polyclonal antibody (PA1-7244)	Salmonella O & H antigens	Rabbit (pAb)	IgG
Ab 4	Salmonella monoclonal antibody [F68C] (MA1-7443)	Salmonella sp. preparation	Mouse (mAb)	IgG1
Ab 5	Salmonella antibody (ab35156)	Mixture of S. Enteritidis, S. Typhimurium and S. Heidelberg	Rabbit (pAb)	IgG
Ab 6	Salmonella antibody (ab252742)	Mixture of S. Enteritidis, S. Typhimurium and S. Heidelberg	Rabbit (pAb)	IgG
Ab 7	Salmonella polyclonal antibody (PA1-20811)	Mixture of S. Enteritidis, S. Typhimurium and S. Heidelberg	Rabbit (pAb)	IgG
Ab 8	Salmonella LPS monoclonal antibody [F62C] (MA1-7447)	Salmonella sp. common core	Mouse (mAb)	IgG2b
Ab 9	Salmonella LPS monoclonal antibody [D46J] (MA5-18257)	Salmonella sp. common core	Mouse (mAb)	IgG2a
Ab 10	Salmonella LPS monoclonal antibody [A99H] (MA5-18258)	Salmonella sp. common core	Mouse (mAb)	IgG2a

Western Blotting

Immunoblotting was conducted as described previously [16]. In brief, bacterial cells at the stationary growth phase were centrifuged at 12,000 ×g for 5 min, resuspended in 1× Laemmli sample buffer (Bio-Rad Laboratories, Inc.), and boiled for 5 min. The bacterial lysates were analyzed using 10% sodium dodecyl sulfate polyacrylamide electrophoresis (SDS-PAGE) gels and the protein fragments were transferred to PVDF membrane (Bio-Rad Laboratories, Inc.). Transfer buffer contained 25 mM Tris base, 0.2 M glycine, and 20% methanol (pH 8.5). The membrane was blocked with 5% skim milk solution (Tris Buffered Saline; TBS, 0.1% Tween-20, and 5% w/v skim milk) for 30 min. TBS contained 20 mM Tris-HCl and 150 mM NaCl. The membranes were then treated with primary antibodies (Table 1) at a dilution ratio of 1:20,000 and horseradish peroxidase (HRP)-conjugated secondary antibodies at a dilution ratio of 1:6,000 for 1 h, respectively. Antibody dilution buffer was TBS containing 0.1%Tween-20 and 5% skim milk and wash buffer (TBS-T) was TBS containing 0.1% Tween-20. All chemical reagents were purchased from Sigma-Aldrich, unless otherwise specified. As a size marker, Precision Plus Protein Dual Color Standards (#1610374; Bio-Rad Laboratories, Inc.) was used. Immunoblotting signals were developed using ECL Western Blotting Detection Reagents kit (GE Healthcare, USA.) and visualized using the ChemiDoc MP System (Bio-Rad Laboratories, Inc.).

Dot-Blot Assay

The experimental flowchart is shown in Fig. 1A. Bacterial cells were cultivated in LB broth until an OD₆₀₀ of 3.0 (approximately 3 × 10⁹ CFU/ml) was reached and 500 μl cell suspension was centrifuged at 12,000 ×g for 5 min. The cells were resuspended in 100 μl lysis buffer (63 mM Tris-HCl, 0.1% β-mercaptoethanol, 2% SDS, and 10%glycerol) and boiled for 5 min. The cell lysate was treated with 10 μl DNase (Turbo DNA-free kit, #AM1907; Invitrogen) and RNase (#EN0531; Thermo Fisher Scientific, USA) at 37°C for 20 min and subsequently with 5 μl protease K (20 mg/ml, #AM2544; Invitrogen) at 59°C for 20 min. Then, 5 μl cell lysate solution was spotted on a nitrocellulose membrane (Bio-Rad Laboratories, Inc.) and dried for 20 min (Fig. 1B). The nitrocellulose membrane was blocked with 5% skim milk solution for 20 min and treated with primary antibodies (20 min for each antibody). Ab 8 and Ab 9 were diluted in 5% skim milk at a ratio of 1:100, whereas Ab 3 was diluted in 5% skim milk at a ratio of 1:100,000. When Ab 3 was applied together with Ab 8 or Ab 9, the membrane was first treated with Ab 8 or Ab9 (monoclonal antibody) before adding Ab 3 (polyclonal antibody). Antibody was diluted in TBS-T containing 5% skim milk. The membrane was washed using TBS-T buffer twice and treated with alkaline phosphatase-conjugated secondary antibodies diluted in 5% skim milk at a ratio of 1:2000 for 20 min. The membrane was washed using TBS-T buffer twice and treated with 20 mL substrate solution [0.4% nitrotetrazolium blue chloride (NBT, #N6876; Sigma-Aldrich) and 0.2% 5-bromo-4-chloro-3-indolyl phosphate disodium salt (BCIP, #B6149; Sigma-Aldrich) in a working buffer (0.1 M NaCl, 50 mM MgCl₂, 0.1 M Tris-HCl)] for 20 min. The membrane was dried after purple spots developed.

Figure 1. Scheme of the dot-blot immunoassay. (A) Flowchart of the designed method with the estimated time. (B) A diagram showing the designed locations of tested serotypes on a nitrocellulose membrane. White and purple circles are indicative of negative and positive interaction, respectively, with tested antibodies. Cell lysates of 18 Salmonella serotypes (A1– D3) used in optimizing the assay were spotted onto nitrocellulose membranes. Lysis buffer solution (D4) and Escherichia coli cell lysate (D5) were spotted as controls.

Assessment of the Cross-Reactivity of 10 Antibodies to Other Bacterial Pathogens

The cross-reactivity of 10 different Salmonella antibodies (Table 1) was tested using five representative foodborne pathogens, including three gram-negative (S. Typhimurium, E. coli, and V. cholera) and two gram-positive (S. aureus and B. cereus) species (Fig. 2). The bacterial whole cell lysates were run on SDS-PAGE gels and probed with 10 different antibodies individually. All antibodies excluding Ab 8, Ab 9, and Ab 10 exhibited multiple ladder-like bands against Salmonella, while three antibodies of Ab 8, Ab 9, and Ab 10 against the common core oligosaccharide region of Salmonella LPS produced a few low-molecular weight bands with Salmonella cell lysates (Fig. 2). However, all antibodies were not cross-reactive with the other four bacterial species.

Figure 2. Cross-reactivity tests of 10 Salmonella antibodies (Ab 1–10) against five bacterial species in western blotting analyses. ST: Salmonella Typhimurium ATCC 14028; Ec: Escherichia coli ATCC 53868; Vc: Vibrio cholerae ATCC 14033; Sa: Staphylococcus aureus ATCC 29213; Bc: Bacillus cereus ATCC 14579. M: size marker.

Comparison of Antibody Selectivity against 18 Salmonella Serotypes

The detection effectiveness of the antibodies was examined using 18 Salmonella serotypes, i.e., S. Enteritidis, S. Newport, S. Typhimurium, S. Javiana, S. I 4,[5],12:i:-, S. Infantis, S. Muenchen, S. Montevideo, S. Braenderup, S. Thompson, S. Saintpaul, S. Heidelberg, S. Oranienburg, S. Mississippi, S. Bareilly, S. Berta, S. Agona, and S. Anatum, which were responsible for 67.5% Salmonella infection in the U.S. in 2016 [10]. Incidence rates and tested strains of the 18 serotypes are listed in Table S1. The bacterial lysates were analyzed using SDS–PAGE and probed with 10 antibodies (Fig. S1). Ab 8, Ab 9, and Ab 10, which were generated using Salmonella core oligosaccharides as immunogens, displayed fewer bands with low-molecular weights than other antibodies. The selectivity of 10 antibodies against 18 serotypes are shown in Table 2. S. Typhimurium, S. Thompson, S. Saintpaul, and S. Heidelberg serotypes were detected by most tested antibodies. However, S. Newport was not detected by any antibody. Comparing the numbers of serotypes detected by each antibody indicated that Ab 8 had the highest detection rate (13/18) and detected S. Infantis, S. Montevideo, S. Mississippi, and S. Bareilly, which were not detected by the other antibodies (Fig. S1, Table 2). The detection rates of the four polyclonal antibodies were comparable (Ab 3: 8/18; Ab 5, Ab 6, and Ab 7: 7/18) and their selectivity spectra were also similar, overlapping S. Enteritidis, S. Typhimurium, S. Thompson, S. Saintpaul, S. Heidelberg, S. Berta, and S. Agona (Fig. S1, Table 2).

Table 2 . Spectrum of 10 Salmonella antibodies against 18 serotypes in western blotting..

	Antibody (Ab No.)
	1	2	3	4	5	6	7	8	9	10	3+8	3+9	8+9
S. Enteritidis	−3	+	+ +	−	+ +	+ +	+ +	+	+ +	+	+ +	+ +	+ +
S. Newport	−	−	−	−	−	−	−	−	−	−	−	−	−
S. Typhimurium	+ + +	+ + +	+ + +	+ + +	+ + +	+ + +	+ + +	+	+ +	+ +	+ + +	+ + +	+ +
S. Javiana	−	−	−	−	−	−	−	+	+	+ + +	+	+	+
S. I 4,[5],12:i:-	−	−	−	−	−	−	−	+	−	+ + +	+	−	+
S. Infantis	−	−	−	−	−	−	−	+ +	−	−	+ +	−	+ +
S. Muenchen	−	−	−	−	−	−	−	−	+	−	−	+	+
S. Montevideo	−	−	−	−	−	−	−	+ +	−	−	+ +	−	+ +
S. Braenderup	−	−	−	−	−	−	−	+ +	−	−	+ +	+	+ +
S. Thompson	+ + +	+ + +	+ + +	+ +	+ + +	+ + +	+ + +	−	+	+	+ + +	+ + +	+
S. Saintpaul	+ + +	+ + +	+ + +	+ + +	+ + +	+ + +	+ + +	−	+ +	-	+ + +	+ + +	+ +
S. Heidelberg	+ + +	+ + +	+ + +	+ + +	+ + +	+ + +	+ + +	−	+ +	+	+ + +	+ + +	+ +
S. Oranienburg	−	−	−	−	−	−	−	+ +	−	+	+ +	−	+ +
S. Mississippi	−	−	−	−	−	−	−	+	−	−	+	−	+
S. Bareilly	−	−	−	−	−	−	−	+ +	−	−	+ +	−	+ +
S. Berta	−	−	+ + +	−	+ + +	+ + +	+ + +	+	+ +	+	+ + +	+ + +	+ +
S. Agona	−	+ + +	+ + +	−	+ + +	+ + +	+ + +	+	+ +	−	+ + +	+ + +	+ +
S. Anatum	−	−	−	−	−	−	−	+ +	+	+	+ +	+	+ +
Detection rate1	4/18	6/18	7/18	5/18	7/18	7/18	7/18	13/18	10/18	9/18	16/18	11/18	17/18
Incidence (%)2	14.8	32.4	35.3	16.9	33.2	33.2	33.2	49.8	42.2	43.3	54.8	44.3	57.4

¹Detection rate: Number of detected serotypes out of tested 18 serotypes..

²Incidence (%): Percentage of salmonellosis incidence associated with the detectable serotypes. Incidence was retrieved from the culture-confirmed human infection survey by LEDS. https://www.cdc.gov/nationalsurveillance/pdfs/2016-Salmonella-report-508.pdf..

³Symbols of +/-: Detected/not detected. More symbols indicate stronger detection signals..

Owing to the selectivity and signal intensity (responsible for the sensitivity in dot-blot assays), Ab 3, Ab 8, and Ab 9 were selected for further dot-blot immunoassays. Ab 3 showed the highest detection rate among the four polyclonal antibodies and exhibited strong binding signals in the western blotting analysis. Ab 8 and Ab 9, despite having moderate signal intensities, demonstrated broad serotype spectra, covering 13 and 10 serotypes, respectively. The selectivity of three combinations (Ab 3+8, Ab 3+9, and Ab 8+9) was predicted based on the western blotting analyses (Table 2). Ab 3+8 and Ab 8+9 were estimated to detect 16 and 17 serotypes, respectively, which are responsible for 54.8% and 57.4% salmonellosis incidence, respectively, in the U.S. in 2016 [10]. Ab 3+9 was expected to detect 11 serotypes, which was inferior to the other two combinations.

Application of a Combination of Two Antibodies to Dot-Blot Immunoassays

Dot-blot assays were preformed using Ab 3, Ab 8, and Ab 9 individually (Fig. 3A) and in combination (Fig. 3B). The cell lysates of 18 Salmonella serotypes were dotted on nitrocellulose membranes, probed with primary and alkaline phosphatase-conjugated secondary antibodies, and treated with BCIP/NBT substrate solution as described in Materials and Methods (Fig. 1). Detection limit of three antibodies was estimated at approximately 10⁶ CFU in dot-blot assays using S. Typhimurium ATCC 19585 (Fig. S2). With regard to the differential antibody binding affinities between serotypes, Salmonella cells of approximately 7.5 × 10⁷ CFU were used for each tested serotype. The serotype selectivity of each antibody was compared between western blotting and dot-blot analyses in Table 3. Most serotypes detected in western blotting experiments were also detected in dot-blot assays with a few exceptions. Ab 3 failed to detect S. Braenderup and S. Thompson in dot-blot assays but detected S. I 4,[5],12:i:-and S. Anatum. Ab 8 did not detect S. Typhimurium and S. Mississippi in dot-blot analyses but detected S. Thompson. Ab 9 did not detect S. Javiana, S. Muenchen, and S. Thompson but detected S. I 4,[5],12:i:-.

Table 3 . Spectrum of three Salmonella antibodies against 18 serotypes in dot-blot assay..

	Antibody (Ab No.)
	3		8		9		3+8	3+9	8+9
	WB1	DB2	WB	DB	WB	DB	DB	DB	DB
S. Enteritidis	+ +5	+ +	+	+	+ +	+	+ +	+ +	+
S. Newport	−	−	−	−	−	−	−	−	−
S. Typhimurium	+ + +	+ +	+	−	+ +	+	+ +	+ +	+
S. Javiana	−	−	+	+	+	−	+	−	+
S. I 4,[5],12:i:-	−	+ +	+	+	−	+	+ +	+ +	+
S. Infantis	−	−	+ +	+ +	−	−	+ +	−	+
S. Muenchen	−	−	−	−	+	−	−	−	-
S. Montevideo	−	−	+ +	+ +	-	−	+ +	−	+
S. Braenderup	+	−	+ +	+	−	−	+	−	+
S. Thompson	+ + +	−	−	+	+	−	+	−	+
S. Saintpaul	+ + +	+ +	−	−	+ +	+	+ +	+ +	+
S. Heidelberg	+ + +	+ +	−	−	+ +	+	+ +	+ +	+
S. Oranienburg	−	−	+ +	+	−	−	+ +	−	+
S. Mississippi	−	−	+	−	−	−	−	−	−
S. Bareilly	−	−	+ +	+ +	−	−	+ +	−	+
S. Berta	+ + +	+ +	+	+	+ +	+	+ +	+ +	+
S. Agona	+ + +	+ +	+	+	+ +	+	+ +	+ +	+
S. Anatum	−	+	+ +	+	+	+	+ +	+	+
Detection rate3	8/18	8/18	13/18	12/18	10/18	8/18	15/18	8/18	15/18
Incidence (%)4	35.3	36.8	49.8	40.6	42.2	36.8	53.7	36.8	53.7

¹WB: western blotting analysis..

²DB: Dot-blot immunoassay..

³Detection rate: Number of detected serotypes out of tested 18 serotypes..

⁴Incidence (%): Percentage of salmonellosis incidence associated with the detectable serotypes. Incidence was retrieved from the culture-confirmed human infection survey by LEDS. https://www.cdc.gov/nationalsurveillance/pdfs/2016-Salmonella-report-508.pdf..

⁵Symbols of +/-: Detected/not detected. More symbols indicate stronger detection signals..

Figure 3. Dot-blot assays of three antibodies against 18 Salmonella serotypes. Cell lysates of 18 Salmonella serotypes were spotted as described in Fig. 1B and probed with Ab 3, Ab 8, and Ab 9, (A) individually or (B) in combination. As negative controls, lysis buffer solution (D4) and Escherichia coli cell lysate (D5) were tested in parallel.

Dot-blot assays were conducted using three combinations of two antibodies (Fig. 3B). Considering the weak binding signals of monoclonal antibodies, higher Ab 8 or Ab 9 concentrations were added before adding Ab 3 (Fig. 1). The optimized dot-blot assay using Ab 3+8 or Ab 8+9 identified 15 of 18 serotypes (Fig. 3B). The diagnosed 15 serotypes were overlapped between two combinations and were estimated to be associated with 53.7% salmonellosis cases in the U.S. in 2016 (Table 3). However, the detection sensitivity of Ab 8+9 was inferior to that of Ab 3+8, likely because of the weak signal of monoclonal antibodies.

The designed dot-blot immunoassay using Ab 3+8 was applied to diagnose 18 Salmonella strains belonging to seven dominant serotypes (Fig. 4). Fifteen strains categorized into S. Enteritidis (five strains), S. Typhimurium (three strains), S. Montevideo (two strains), S. Infantis (one strain), and S. Bareilly (four strains) were successfully detected within 3 h. Noticeably, Ab 3+8 produced faint dot-blot signals against S. Newport (two strains) and S. Muenchen (one strain), which were not detectable in western blotting.

Figure 4. Diagnosis of 18 Salmonella strains using dot-blot immunoassay. Eighteen Salmonella strains belonging to seven different serotypes were subjected to the designed dot-blot immunoassay using Ab 3+8. A1: S. Enteritidis ATCC 4931; A2: S. Enteritidis FORC_019; A3: S. Enteritidis FORC_052; A4: S. Enteritidis ATCC 13076; A5: S. Enteritidis MFDS 1004839; B1: S. Typhimurium FORC_030; B2: S. Typhimurium NCCP 12219; B3: S. Typhimurium NCCP 14772; B4: S. Newport NCCP 12235; B5: S. Newport FORC_020; C1: S. Montevideo CCARM 8189; C2: S. Montevideo MFDS 1006814; C3: S. Infantis MFDS 1006818; C4: S. Muenchen KCPB 03; C5: S. Bareilly CCARM 8578; D1: S. Bareilly NCCP 11674; D2: S. Bareilly NCCP 16323; D3: S. Bareilly MFDS 1007637; D4: lysis buffer solution; D5: Escherichia coli ATCC 53868.

Traditional Salmonella detection methods are typically based on biological and physiological differences between Salmonella and the other bacterial species [17, 18]. Therefore, the procedures are labor-intensive and require at least several days [19]. Furthermore, traditional serotyping methods using antiserum agglutination with antigenic determinants require a comprehensive set of antisera and trained experts owing to the convoluted protocol, which hinders their extensive use as a popular diagnostic tool for general laboratories [19, 20]. With the developments in molecular biology, various rapid detection methods have been devised to supplement the drawbacks of conventional Salmonella detection techniques [19, 21]. Polymerase chain reaction (PCR)-based molecular diagnostic techniques are extremely quick and sensitive, but their extensive dissemination is constrained by the high cost of PCR operations [22]. Immunological diagnostic techniques, such as ELISA and immuno-chromatography, enable excellent Salmonella detection with high affinity [23-25]. However, these methods are time-consuming and yield low output. Moreover, the efficacy of antibodies is variable between serotypes and vulnerable to external environmental conditions [26, 27]. Besides, there are several other alternatives to the conventional methods, utilizing mass spectrometry, optical phenotyping, and electrochemical biosensors. Particularly, electrochemical biosensors combined with aptamer technology (such as aptasensors) exhibit excellent rapidity and sensitivity [28] and can be integrated into a portable biosensor device, which merits point-of-care testing [29, 30].

As an alternative immunoassay, dot-blot assays also utilize antibodies to detect target antigens but use nitrocellulose or PVDF membranes instead of polystyrene plates frequently used in ELISA. Antigens such as proteins and LPS bind to nitrocellulose and PVDF membranes more tightly than polystyrene plates [14, 15]. Therefore, dot-blot-based methods have considerably higher sensitivity than ELISA-based methods [14]. Dot-blot assay is a simplified method derived from western blotting. However, dot-blot assays do not require electrophoretic separation of antigen molecules on polyacrylamide gels and the procedures are substantially easier and faster than western blotting. Moreover, dot-blot assays are cost-effective. Besides, spotting bacterial lysates on membranes of randomized sizes enables screening large numbers of specimens for the presence of target antigens.

In contrast, dot-blot immunoassays possess a critical drawback owing to antibody specificity. Commercial antibodies against Salmonella generally recognize bacterial cell-surface epitopes present in the outer membrane protein, flagella, and LPS. The challenge is that even within the same species, structural components on the surface may vary between bacterial strains [31, 32]. Contrariwise, different bacterial species may have some similarity in their membranous structure. LPS molecules are complex glycolipids containing a hydrophobic moiety (lipid A), a core oligosaccharide, and a long-chain repeat-unit polysaccharide (O-antigen). Salmonella lipid A is a well-conserved archetypal structure comprising a bisphosphorylated glucosamine disaccharide carrying six acyl chains [33]. Although the core oligosaccharide architecture is also relatively conserved within a species, the O-antigen consists of variable oligosaccharide repeats, where the sugars and structural arrangements are different [34]. Owing to its hyper-variability in terms of length and sugars of repeat units, the O-antigen moiety provides various epitopes, determining the serological specificity between bacterial strains [35]. According to the Kauffman–White serological classification system, Salmonella has 46 O-serogroups [36]. Owing to the variable specificity and selectivity of antibodies between Salmonella serotypes, a single antibody may not detect every Salmonella serotype [37, 38]. Therefore, conventional dot-blot assays have been exploited to identify a specific Salmonella serotype such as S. Enteritidis. However, most monoclonal antibodies developed against S. Enteritidis had the potential to display cross-reactions with other Salmonella serotypes [15, 39, 40]. In the same context, the conventional Salmonella antibodies generated using LPS and/or flagella moieties as immunogens have the potential to interact with bacterial species with similar membranous structure. Especially, E. coli has many structural similarities to Salmonella spp. in LPS, flagella, and outer membrane proteins [41, 42]. Therefore, the reactivity of Salmonella antibodies should be interpreted carefully.

The dot-blot immunoassay developed in this study combined two different Salmonella antibodies with broad cross-reactivity against diverse Salmonella serotypes to diagnose multiple serotypes with high incidence rates in humans. The combination of Ab 3 and Ab 8 identified 15 predominant Salmonella serotypes (Fig. 3B; Table 3), which were responsible for 53.7% culture-confirmed salmonellosis cases in the U.S. in 2016 [10]. Notably, approximately 70% incidence of culture-confirmed salmonellosis was attributable to only 20 serotypes and the rest was ascribed to 360 serotypes (16.5%) and unknown agents (14.3%) [10]. The prevalent Salmonella serotypes were generally similar with minor changes for decades [10]. As a feasibility test, the developed dot-blot assay was applied to seven serotypes of the dominant 20 serotypes (Fig. 4): S. Enteritidis, S. Typhimurium, S. Newport, S. Infantis, S. Muenchen, S. Montevideo, and S. Bareilly. Three serotypes of S. Enteritidis, S. Typhimurium, and S. Newport are typically recognized as the top 3 serotypes with the highest incidence rates [10]. Salmonellosis attributable to the other 4 serotypes is growing recently, although the case numbers are much lower than those by the top 3 serotypes [43-46]. Fifteen strains belonging to S. Enteritidis, S. Typhimurium, S. Montevideo, S. Infantis, and S. Bareilly were successfully detected, whereas 3 strains of S. Newport and S. Muenchen produced weak dot-blot signals. Interestingly, Ab 3 and Ab 8 failed to detect S. Newport and S. Muenchen in western blotting (Fig. S1, Table 2). In comparison with PVDF membranes, nitrocellulose membranes used in the dot-blot assay might exert better binding affinity with various antigenic materials including oligosaccharides, lipids, and proteins, as observed previously [47, 48]. Varied attempts, such as pretreating membranes to increase hydrophobic interactions, adding more antibodies, and using more sensitive reporter enzymes, could be carried out to improve the detection sensitivity.

The drawback of conventional Salmonella dot-blot immunoassays, which is attributable to the cross-reactivity of most commercial antibodies across diverse Salmonella serotypes, was exploited as a counterplot to diagnose multiple serotypes using two antibodies in this study. Combination of polyclonal and monoclonal antibodies broadened the detection spectrum, covering 15 serotypes, and intensified the dot-blot signals. Combination with different and more antibodies can extend the diagnostic spectrum. Considering the simplicity and rapidity, the devised dot-blot immunoassay can be employed for an on-site rapid diagnosis of harmful Salmonella serotypes in the food industry.

This work was supported by a grant (2021M3A9I4026029) from the Bio & Medical Technology Development Program of the National Research Foundation of Korea (NRF), funded by the Ministry of Science & ICT, and a grant (2021N100) of the Commercializations Promotion Agency for R&D Outcomes (COMPA), funded by the Korea government (MSIT).

The authors have no financial conflicts of interest to declare.