Proteomic characterization and discrimination of Aeromonas species recovered from meat and water samples with a spotlight on the antimicrobial resistance of Aeromonas hydrophila

Abstract Aeromonas is recognized as a human pathogen following ingestion of contaminated food and water. One major problem in Aeromonas identification is that certain species are phenotypically very similar. The antimicrobial resistance is another significant challenge worldwide. We therefore aimed to use mass spectrometry technology for identification and discrimination of Aeromonas species and to screen the antimicrobial resistance of Aeromonas hydrophila (A. hydrophila). A total of 150 chicken meat and water samples were cultured, and then, the isolates were identified biochemically by the Vitek® 2 Compact system. Proteomic identification was performed by MALDI‐TOF MS and confirmed by a microchannel fluidics electrophoresis assay. Principal component analysis (PCA) and single‐peak analysis created by MALDI were also used to discriminate the Aeromonas species. The antimicrobial resistance of the A. hydrophila isolates was determined by Vitek® 2 AST cards. In total, 43 samples were positive for Aeromonas and comprised 22 A. hydrophila, 12 Aeromonas caviae (A. caviae), and 9 Aeromonas sobria (A. sobria) isolates. Thirty‐nine out of 43 (90.69%) Aeromonas isolates were identified by the Vitek® 2 Compact system, whereas 100% of the Aeromonas isolates were correctly identified by MALDI‐TOF MS with a score value ≥2.00. PCA successfully separated A. hydrophila, A. caviae and A. sobria isolates into two groups. Single‐peak analysis revealed four discriminating peaks that separated A. hydrophila from A. caviae and A. sobria isolates. The resistance of A. hydrophila to antibiotics was 95.46% for ampicillin, 50% for cefotaxime, 45.45% for norfloxacin and pefloxacin, 36.36% for ceftazidime and ciprofloxacin, 31.81% for ofloxacin and 27.27% for nalidixic acid and tobramycin. In conclusion, chicken meat and water were tainted with Aeromonas spp., with a high occurrence of A. hydrophila. MALDI‐TOF MS is a powerful technique for characterizing aeromonads at the genus and species levels. Future studies should investigate the resistance of A. hydrophila to various antimicrobial agents.

Many people consume chickens daily as a source of animal protein worldwide; hence, hygienic methods of supplying chickens for consumption are critical for public health. Meat can be infected with Aeromonas spp. not only through inadequate processing, cutting and grinding but also by washing carcasses with contaminated water (Ghenghesh, Ahmed, El-Khalek, Al-Gendy, & Klena, 2008;Stratev & Odeyemi, 2016). The poor hygienic conditions associated with the processing of raw meat are considered one of the major causes of Aeromonas spp. contamination of meat products (Encinas, Gonzalez, Garcia-Lopez, & Otero, 1999;Ogu, Madar, Okolo, & Tayubi, 2017;Rajakumar, Ayyasamm, Shanthi, Song, & Lak-shmanaperumalsamy, 2012). Therefore, the genus Aeromonas has been associated with a wide variety of food and waterborne infections worldwide, particularly in less developed countries due to poor personal hygiene and lack of quality water (Odeyemi & Ahmad, 2017). Most Aeromonas spp.
The most important Aeromonas spp. are A. hydrophila, Aeromonas caviae (A. caviae), and Aeromonas veronii biovar sobria (A. veronii bv. sobria). These organisms are pervasive in water and meat (Encinas et al., 1999;Osman, Aly, Kheader, & Mabrok, 2012;Sharma & Kumar, 2011;Trakhna et al., 2009). A. hydrophila represents the most virulent of these species and produces multifactorial virulence factors, including structural features related to adhesion, cell attack, and escape from the phagocytosis process, and certain extracellular factors, such as aerolysin, which leads to lysis and toxicity of the cells (Abrami, Fivaz, Glauser, Parton, & Goot, 1998;Chopra & Houston, 1999;Citterio & Biavasco, 2015). Nevertheless, some species of Aeromonas were isolated formerly from several food products, and the substantial role of foods of animal origin in the distribution of Aeromonas infections is unclear.
Although biochemical methods, 16S rRNA sequencing and housekeeping genes are considered the standard methods for detecting different Aeromonas spp., they are not widely used due to their cost, labor and time requirements (Chen et al., 2014;Morinaga et al., 2013;Soler et al., 2004;Trakhna et al., 2009). In addition, the exactness of these presently available methods is limited, and the precise and rapid identification of Aeromonas at the species level is still problematic (Benagli et al., 2012;Pérez-Sancho et al., 2018).
From this perspective, to increase the rate of Aeromonas identification at the species level, recent studies have confirmed and recommended that matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) as an alternative technique for bacterial identification due to its favorable rapid application (Donohue, Smallwood, Pfaller, Rodgers, & Shoemaker, 2006;Elbehiry, Al-Dubaib, Marzouk, Osman, & Edrees, 2016;Murray, 2010). This technology is an up-to-date approach extensively applied for the identification and discrimination of various microorganisms at the genus and species levels on the basis of MALDI-TOF mass spectra (Elbehiry et al., 2017;Sandrin, Goldstein, & Schumaker, 2013;Vávrová, Balážová, Sedláček, Tvrzová, & Šedo, 2015).
The development of antimicrobial resistance in various types of bacteria is another significant challenge worldwide (Chugh, 2008;Laith & Najiah, 2013;Li & Webster, 2018). Recently, the antibiotic resistance of Aeromonas spp. has increased because resistance was developed not only in clinical isolates but also in strains isolated from different sources of food products (Alcaide, Blasco, & Esteve, resistance among foodborne pathogens has developed, possibly due to the prolonged administration of medications in the livestock used for human consumption (Adebayo, Majolagbe, Ola, & Ogundiran, 2012;Deng et al., 2016). A previous study conducted by Saavedra et al. (2004) illustrated that the widespread use of various groups of beta-lactam antibiotics as a method of prophylaxis and treatment of A. hydrophila in humans and animals is considered one of the main causes of the increasing A. hydrophila resistance to amoxicillin, carbenicillin, and ticarcillin. Furthermore, the existence of resistance genes on mobile elements, such as plasmids, transposons and integrons, assists their rapid spread among microorganisms (Romero, Feijoo, & Navarrete, 2012). Similarly, the data on the incidence of antibiotic resistance in Aeromonas spp., particularly in A. hydrophila recovered from chicken meat and water, are sparse. Based on these previously mentioned data, our study was designed to identify various Aeromonas spp. from chicken meat and water samples using MALDI-TOF MS confirmed by SYBR Green real-time (RT)-PCR and microchannel fluidics electrophoresis assays and to study the antimicrobial resistance of A. hydrophila using Vitek 2 Compact AST cards.

| Sample collection
A total of 150 samples, including chicken meat (n = 75) and water (n = 75), were collected from three different sites (Buraidah, Unaizah, and Albukairyah) in the Al-Qassim region, Saudi Arabia.
The samples were collected five times at nearly monthly intervals (in April, May, June, August, and September 2017). Three hundred grams of each chicken meat sample collected from six randomly selected local retail shops, and supermarkets were placed in a separate sterilized plastic bag for the isolation process. One hundred milliliters of each water sample was collected randomly from private drinking water wells, houses, and retailers. The samples collected from houses and retailers were treated first with a sterile sodium thiosulphate solution (13.2 mg/L) to neutralize chlorine and stop its bactericidal action (Massa, Armuzzi, Tosques, Canganella, & Trovatelli, 1999). All samples were kept under icecold conditions, and bacteriological investigations were carried out within 2 hr of collection. All meat and water samples were processed in the Microbiology Laboratory, College of Public Health and Health Informatics, Qassim University for isolation. Isolates were preserved in Cryobank vials at −80°C until the identification process was carried out.

| Isolation of Aeromonas spp.
Thirty grams of each meat specimen was added to 225 ml of alkaline peptone water (pH 8.4 ± 0.2 at 25°C, Sigma-Aldrich, USA), homogenized in a blender (Stomacher® 400, Thomas Scientific, USA) for 2 minutes and incubated at 30ºC for 18-24 hr. Likewise, 10 ml of each water sample was inoculated in 90 ml of peptone water with 1% NaCl (w/v) at pH 8.6 adjusted with sodium hydroxide and incubated at 30ºC for 18-24 hr. The cultures were streaked onto Aeromonas Isolation Agar (Sigma-Aldrich) containing 5 mg/L ampicillin, which supports the growth of Aeromonas spp. After incubation of all plates at 28ºC for 24-48 hr, the colonies appeared slightly to deep green. Three to five typical colonies were subcultured onto glutamate starch phenol red agar (Sigma-Aldrich), and after incubation, the colonies appeared as yellow colonies surrounded by a yellow zone and were identified primarily as Aeromonas spp. if they were gram-negative, oxidase-positive and glucose fermenting. The Voges-Proskauer reaction; esculin hydrolysis; lysine decarboxylase; and fermentation of arabinose, salicin, and sorbitol were then carried out to differentiate Aeromonas at the species level (A. hydrophila, A. caviae, and A. veronii bv. sobria) according to the method described by Janda, Abbott, and Carnahan (1995).

| Biochemical analysis of Aeromonas using the Vitek 2 Compact system
The Vitek 2 Compact ((bioMérieux. Marcy l'Etoile, France) Gram-Negative Identification (GNI) and antibiotic susceptibility testing (AST) cards were used to identify and determine the antibiotic susceptibilities of Aeromonas spp. according to the manufacturer's recommendations. In brief, 3-4 fresh colonies were suspended in sterilized physiological saline (aqueous 0.45% NaCl, pH 4.5 to 7.0) and thoroughly mixed. The Mcfarland turbidity was adjusted in the range from 0.50 to 0.63 using DensiChekTM (BioMe′rieux, France).

Five milliliters of this suspension was loaded into Vitek 2 ID-GNI and
AST gram-negative (AST-GN04) cards. The Vitek 2 Cassette was finally loaded with cards and suspension tubes into the device. The unknown organisms were compared to the reference strains stored in the Vitek 2 Compact software for proper identification.

| Rapid identification of Aeromonas spp. using MALDI Biotyper
We applied MALDI Biotyper Reference Library for Clinical Applications (MBT-CA) Database version V.3.3.1.2 (Bruker Daltonik, Bremen, Germany) which has been approved by FDA under Section 510(k) as a powerful method for rapid and precise identification and discrimination of Aeromonas spp. The Proteomic identification was performed according to the ethanol/formic acid extraction method designated by Bruker Corporation. Briefly, a fresh colony of overnight culture, incubated at 28°C for 24 hr, was utilized for each isolate and inoculated onto two spots of the target plate, and every colony was then covered with 1 µl of matrix solution (saturated α-cyano-4-hydroxycinnamic acid in 50% acetonitrile and 2.5% trifluoroacetic acid). The microbial spectra were directly produced by applying Compass IVD software, and the identification was directly conducted with a MALDI Biotyper machine.

| Data analysis in MALDI Biotyper
The score value of the unidentified spectrum in the range from zero to three was determined by matching the unknown spectra with the spectra stored in the Bruker library. The accuracy of the strain recognition was detected as designated by the measures of Bruker Daltonik. The device performed the precise detection of species when the log score ranged from 2.3 to 3.0; nevertheless, the species and genus levels were recognized in the range from 2.00 to 2.29 and from 1,700 to 1,999, respectively. Furthermore, a score of 0.00 to 1.69 means that the proof of identity is not reliable. The diverse spectra created by the Microflex LT Compass IVD software were measured in a m/z range from 2,000 to 20,000 Da. To distinguish between Aeromonas spp., mathematical testing of the data sets was generated on the basis of principal component analysis (PCA), and the findings were illustrated in a three-dimensional (3d) score plot created directly by compass software. According to the MBT-CA Database, which contains 47 reference Aeromonas spp. and subspecies, the PCA dendrogram setting was utilized for species grouping.

| DNA extraction
DNA extraction of the field isolates was achieved by QuickGene-810 (AutoGen, Japan) using the QuickGene DNA tissue kit S (DT-S), which was applied according to the manufacturer's recommendations.
Briefly, 3-5 fresh colonies of each sample grown on soybean casein digest agar were transferred into a sterilized microcentrifuge tube containing 180 µl MDT lysis buffer and 20 µl proteinase K, and the lysate was then centrifuged at 8,000 g for 5 min. The supernatant was transferred to a new Eppendorf tube, and 180 µl LDT buffer was added. Two hundred forty microliters of absolute ethanol (Panreac, Barcelona, Spain) was added, and the tube was properly agitated.
The lysate was transferred into the cartridge supplied with the kits and then inserted into the machine. Finally, the concentration and purity of the extracted DNA were determined by the NanoDrop™ 2000 spectrophotometer (Thermo Scientific, MA, USA).

| Primers used in the study
The isolates were further confirmed to the species level by 16S rRNA, aerolysin (aerA), polar flagella (Fla), and hemolysin (ASA1) genes analysis. A specific 16S rRNA region was carefully chosen for detecting Aeromonas spp. The primer express software, ver. 2.0 (Applied Biosystems, USA) was used to designate the primers, and their specificity was investigated with the BLAST program (Table 1).

| SYBR Green RT-PCR assay and electrophoresis for PCR products
SYBR Green RT-PCR for detection of the A. hydrophila, A. caviae, and A. sobria specific genes was then performed using a 7500 Fast Real-Time PCR System (Applied Biosystems). Briefly, a 20 µl reaction volume containing 10 µl of Maxima SYBR Green qPCR Master Mix (2×), no ROX (Thermo Scientifics), 1 µl forward primer, 1 µl reverse primer, 1 µl target DNA and 7 µl of RNase/DNase free water was used. All reactions were carried out in duplicate. Regular amplification parameters were carried out as follows: 50°C for 2 min, 95°C for 2 min, followed by 40 amplification cycles, each of which comprised 95°C for 15 s and 60°C for 1 min.
Amplification results were expressed by plotting Delta Rn (ΔRn) versus cycle number for detection of the Aeromonas genes. Electrophoresis for PCR products was then carried out using a LabChip GX Touch 24 device (PerkinElmer, USA). DNA 1 K Assay Quick was used for chip and preparation of samples according to the manufacturer's procedures.

| Antimicrobial resistance of A. hydrophila using Vitek 2 Compact AST cards
Vitek 2 Compact AST-GN04 cards (MedexSupply, Passaic, NJ, USA) were used to determine the susceptibility of A. hydrophila to antimicrobial agents. As shown in Table 5, five groups of antibiotics were tested in the present study. All antimicrobial agents were chosen according to these five groups, which can be measured by Vitek 2 system cards (Cockerill et al., 1995). Throughout the evaluation period, A. hydrophila ATCC 35654 was used as a quality control strain and was checked at regular intervals.

| Statistical analysis
The data obtained from our study were imported into the Statistical Package for the Social Sciences (SPSS), and all estimations were carried out using SPSS version 20.0 (SPSS Inc., Chicago, IL, USA).

| Biochemical identification of Aeromonas isolates
The Vitek™ 2   for all Aeromonas spp. Several spectra were distributed within the range from 2,000 to 11,000 m/z (Figure 1), and the higher peaks were determined between 4,000 and 10,000 m/z ( Figure 2).

| Accurate identification of Aeromonas spp. using MALDI-TOF MS
Furthermore, a supplementary mathematical tool called PCA was generated in our study by MALDI Biotyper Compass software to explore the degree of similarity and variation in the protein spectra. Numerous protein spectra of the identified isolates were clarified in three-dimensional (3d) PCA as shown in Figure 3a. Based on the PCA calculation, the impacts of PC1, PC2, and PC3 on the creation of a profile in a percentage plot of the difference elucidated were nearly 45%, 17%, and 9%, respectively ( Figure 3b).
Likewise, we analyzed a single peak for all Aeromonas spp. to explore the distinctive differences in the three Aeromonas spp.

| Antimicrobial resistance of A. hydrophila
Vitek 2 Compact AST-GN04 cards were used to determine the susceptibility of A. hydrophila to antimicrobial agents (

| D ISCUSS I ON
The accurate identification of various pathogens is an essential step of diagnosis, and the time-to-result obtained is very significant to start the selected treatment as soon as possible. Rapid and accurate analytical tools are necessary for monitoring the food and water safety and screening of any undesirable pathogens, which may cause noteworthy health hazards upon consumption. Aeromonas spp. are known to cause various infections in humans. Because its developing significance as an emerging pathogen isolated from food and water, it is imperative to combat this bacterium (Praveen, Debnath, Shekhar, Dalai, & Ganguly, 2016).
As culture-and biochemical-based identification of different microorganisms are difficult and time-consuming, MALDI-TOF MS was significantly used here for the early identification and discrimination of various pathogens from environmental samples by introducing a simple, rapid, precise, and low-cost identification method compared to other methods (Elbehiry et al., 2016;Singhal, Kumar, Kanaujia, & Virdi, 2015;van Belkum, Welker, Pincus, Charrier, & Girard, 2017).
Recently, MALDI-TOF MS has been revealed to be an important method for the rapid identification of bacterial threats that might contaminate drinking water and food products (Singhal et al., 2015).
In our study, the identification rates for Aeromonas isolates were 21/22 (95.45%) for A. hydrophila, 12/12 (100%) for A. caviae and 9/9 (100%) for A. sobria. These findings prove that the mass spectral data generated by MALDI Biotyper Compass 2.0 Software for all isolates were satisfactory to differentiate between the genus Aeromonas at the species and strain levels. The higher level of precise identification compared to that of the former studies might be due to the updated Compass 2.0 database utilized in our study (Lo et al., 2015;Seng et al., 2009). Similar results were obtained by Donohue et al. (2007), who used the m/z signature of the recognized Aeromonas reference isolates to allocate species of unidentified environmental isolates.
They reported that MALDI-TOF MS quickly and precisely categorized fourteen species and four subspecies of Aeromonas, including A. hydrophila, A. caviae, A. jandaei, and A. veronii bv. sobria, which were the most clinically significant species of the genus Aeromonas.
A previous study of Aeromonas isolates was also conducted by Donohue et al. (2006), who indicated that the signals created by Therefore, proteomic identification has been illustrated to be a powerful tool for species identification. Another MS study was evaluated by Lamy, Kodjo, Laurent, and CoIBVH Group (2011) for proteomic identification of aeromonads. They found that the genus-level precision was detected at 100% compared with rpoB gene sequencing, which makes this system one of the most reliable and rapid techniques for the identification of various microorganisms.  (Bizzini et al.., 2011;Carrasco et al., 2016;Croxatto, Prod´hom, G., & Greub, G., 2012;Lau et al., 2014). Similar results were obtained by Benagli et al. (2012), who tested 741 clinical and environmental Aeromonas isolates using MALDI-TOF MS and found that 93% of these strains were positively identified with a score value ≥2.00. In  (Persson, Al-Shuweli, Yapici, Jensen, & Olsen, 2015;Trakhna et al., 2009).
The distribution of drug resistance among A. hydrophila was also evaluated in our study, as previous surveys showed the development of this pathogen as one of the major opportunistic human pathogens (Laith & Najiah, 2013;Rey et al., 2009 According to Alcaide et al. (2010), the resistance of Aeromonas spp. to various antibiotics has been augmented because emerging resistance has been established not only in clinical isolates but also in Aeromonas spp. recovered from water and food. Another study described by Adebayo et al. (2012) revealed that the frequency of bacterial resistance to different antibiotics in food and food products has increased throughout the last few years, potentially due to their extensive use in livestock raised for human feeding.

Current research illustrates a high frequency of possibly virulent
A. hydrophila among Aeromonas spp. in chicken meat and water samples. Through this study, we confirmed that MALDI-TOF MS used for the identification of Aeromonas isolates is a powerful, cost-effective, and accurate method and was able to distinguish A. hydrophila, A. caviae, and A. sobria strains based on PCA and single-peak analysis. RT-PCR and microchannel fluidics electrophoresis assays can F I G U R E 5 Higher peaks intensity (3,667, 4,351, 7,335 and 9,635 Da) were detected in A. sobria (green), whereas they were missed in A. hydrophila (red) and A. caviae (blue). Moreover, a higher peak intensity (7,347 Da) was detected in A. caviae (blue) while it was missed in A. hydrophila (red) and A. sobria (green) be used as a confirmatory diagnostic method for MALDI-TOF MS.
Future studies will address the application of this method for the direct identification and differentiation of Aeromonas spp. in food or water samples. Moreover, our findings demonstrate that the A. hydrophila strains in the sample had developed antibiotic resistance.
Consequently, the progression of resistance may be predictable; accordingly, the number of effective antibiotics is declining. Because A. hydrophila may threaten human health, the transmission of resistance may have bad impacts for humans.

ACK N OWLED G EM ENTS
This research was made possible with the support of Qassim University, Kingdom of Saudi Arabia.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

AUTH O R S CO NTR I B UTI O N
AE, EM, and EEA designed the study. AE, EM, EEA, AA, and MH performed all the experiments, and AE, MA and EM analyzed the data and wrote the manuscript. All authors read and approved the final manuscript.

E TH I C S S TATEM ENT
Permission has been received from the owners of shops and retailers.

DATA ACCE SS I B I LIT Y
All data generated or analyzed during this study are included in this published article.