ND, not determined; \* Intermediate resistance; AMP, ampicillin; MEM, meropenem; CTX, cefotaxime; CTF, ceftiofur; CAZ, Ceftazidime; ETP, ertapenem; IPM, imipenem; NAL, nalidixic acid; CIP, ciprofloxacin; ENR, enrofloxacin; LEV, levofloxacin; STR, streptomycin; KAN, kanamycin; GEN, gentamicin; AMK, amikacin; TET, tetracycline; DOX, doxycycline; FOS, fosfomycin; FFC, florfenicol; TIG, tigecycline; SXT, sulfamethoxazole-trimethoprim (SMZ-TMP).

**Figure 2.** Comparison of the antibiotic resistance rates between the ExPEC and non-ExPEC isolates in this study.

Among the 16 ExPEC isolates resistant to cefotaxime, *bla*CTX-M was found in 15 isolates (10 *bla*CTX-M-9G and 6 *bla*CTX-M-1G), including one isolate WF1-5-13 carrying both *bla*CTX-M-9G and *bla*CTX-M-1G (Table 4). *bla*CTX-M-9G and *bla*CMY-2-type pAmpC-encoding gene were present in isolate AH234 from layer farm B, accounting for the resistances to the thirdgeneration cephalosporins. *fosA3* was found in ten ExPEC isolates resistant to fosfomycin and this gene was only distributed among *bla*CTX-M-positive isolates. For the six colistin and two meropenem resistant ExPEC isolates, the presence of *mcr-1* and *bla*NDM could account for their corresponding resistance, respectively. Notably, the two ExPEC isolates (WF1-5-13 and WF1-5-40) harboring both *mcr-1* and *bla*NDM, carried additional *bla*CTX-M, *fosA3* and *floR* genes, and isolate WF1-5-40 also possessed *rmtB* (Table 4).

#### *3.5. Genetic Relationships of the ExPEC Isolates*

All 22 ExPEC isolates could be successfully analyzed by PFGE and 20 different PFGE profiles were obtained indicating that clonal dissemination of these ExPEC isolates was unlikely (Figure 1). As shown in Figure 1, 12 types were identified by the MLST subtyping method along with one new ST (6-6-804-10-9-1-6 in isolate G-1-5) not previously registered in the *E. coli* MLST database. The most prevalent ST types were ST117 (four) and ST93 (four), followed by ST569 (three), ST1485 (two) and ST2944 (two). Most isolates sharing the same ST types had different PFGE profiles and were from different farms or markets of different cities (Figure 1). For example, the four isolates (AH234, L2-JC-34, WF1-5- 10 and LS-A-3) belonging to ST93 were from four chicken markets/farms in four cities and had different PFGE patterns. Notably, the two ExPEC isolates WF1-5-13 and WF1- 5-21 from the same chicken farm in city Weifang shared identical PFGE pattern and ST type; however, different resistance genotypes and phenotypes were found in these two isolates (Figure 1 and Table 4). Different resistance phenotypes were also found in isolates JS-JC-4 and JS-JC-7, which were from the same market and possessed identical PFGE pattern and ST type (Figure 1 and Table 4).

#### **4. Discussion**

The presence of ExPEC colonizing healthy chickens could be a huge threat to both animal and human health. For China, having the largest consumption of antibiotics in the world, the prevalence and antibiotic resistance of ExPEC among healthy chickens urgently need to be studied. In this study, we investigated the resistance of *E. coli* isolates from healthy chickens of seven layer farms, one white-feather broiler farm and 17 live poultry markets in China, and the ExPEC among these commensal isolates were also characterized.

Both the rates of resistance to tetracycline and doxycycline among isolates in this study were above 80.0%, consistent with that from chickens in China after 2012 [48]. The high resistance rates to these drugs could be attributed to the heavy usage of tetracyclines in poultry, because oxytetracycline, tetracycline, chlortetracycline and doxycycline have been heavily used for decades in animal production including poultry [49]. Tigecycline, a last-resort treatment for human infections caused by MDR Gram-negative bacteria, has never been used in animal husbandry. Luckily, isolate resistant to tigecycline was not found in our *E. coli* isolated during 2015–2017; however, the heavy usage of tetracyclines in animals could increase the prevalence of newly mobile tigecycline-resistance gene *tet*(X4) in *E. coli* and this should be paid more attention [33]. The resistance rates to meropenem (4.9%) and colistin (17.0%), two critically important antimicrobials in human medicine, among the 926 isolates in our study could be well accounted for the presence of *bla*NDM (4.9%) and *mcr-1* (17.0%), respectively (Tables 1 and 2). *qnrS* was the most prevalent PMQR gene in this study, differing from that in humans [50] and animals [51], in which the most prevalent PMQR gene was *oqxAB*. In some countries, *qnrB* was the most prevalent type [52]. The prevalence of CTX-M-type ESBLs (35.6%) in healthy chickens in this study was similar to that (38.5%) in *E. coli* isolates from chickens in China [48], but lower than that in chicken production of India [53]. Notably, except levofloxacin and tigecycline, the resistance rate for each of the 15 antimicrobials tested among *E. coli* from white-feather broilers, was significantly higher than that from brown-egg layers and that from yellow-feather broilers (*p* < 0.05) (Table 1). Such phenomena could be attributed to that consumption of antibiotics in white-feather broilers is the largest among the three types of chickens. The rate of resistance to ampicillin, cefotaxime, colistin, and fosfomycin among *E. coli* from yellowfeather broilers was significantly higher than that from brown-egg layers, respectively (*p* < 0.05). This might because that β-lactams and colistin are often used in early feeding period of the yellow-feather broilers while almost all antimicrobials are forbidden in layer farms during the laying period.

Based on the molecular criteria of Johnson et al. [6], 2.4% (22/926) of the healthy chicken fecal *E. coli* isolates were qualified as ExPEC in this study. The ExPEC could asymptomatically colonize the gut of a fraction of healthy animal population and survive

in extra-intestinal environments, causing diseases in animals and humans through the food chain [54]. The threat to human health posed by the healthy chicken ExPEC isolates in this study could been further proved by that healthy poultry ExPEC were capable to adhere or invade human intestinal epithelial [14]. In this study, virulence markers *iutA* and *KpsM* II were the two most prevalent genes among the ExPEC isolates from healthy chickens, consistent with the finding about MDR *E. coli* in healthy chickens in Brazil [14]. Besides isolation methods, geographic locations and management practices, different classification criteria for ExPEC has been the main factor contributing to differences in frequency of ExPEC between studies. We will focus the studies using the same PCR-based screening method for ExPEC as that in our study. The detection rate of ExPEC in our samples was 2.4%, similar to that (4.7%, 5/108) among chicken egg *E. coli* isolates (*p* > 0.05), but lower than that (21.5%, 130/606) in chicken meat isolates reported in the USA (*p* < 0.05) [55]. Notably, the prevalence of ExPEC isolates (2.4%, 22/926) in our study was also significantly lower than that (13.2%, 40/304) from farmed healthy chickens in Quebec, Canada [20]. This might be explained by that boiled DNA extracts from total cultures of samples were initially screened for all possible ExPEC strains in the previous study, contributing to a high recovery rate of ExPEC. In this study, the detection rate (7.7%, 6/78) of ExPEC among white-feather broilers was significantly higher than that (1.6%, 6/371) from brownegg layers and that (2.1%, 10/477) from yellow-feather broilers in live poultry markets (*p* < 0.05). The phenomenon could be attributed to the selective pressure on the dissemination of ExPEC posed by antimicrobials frequently used in white-feather broilers. This was proved by that highly similar PFGE patterns were found in the two ExPEC isolates WF1-5-13 and WF1-5-21 from the same Farm (Figure 1).

In this study, the most two prevalent ST types of the 22 ExPEC isolates were ST117 and ST93 (Figure 1). Since ST117 and ST93 types of ExPEC had been found to be associated with meningitis of humans in Brazil in 1999 (http://enterobase.warwick.ac.uk/species/ecoli/ search\_strains?query=st\_search (accessed on 7 October 2020)), both ST types of *E. coli* have caused sepsis among humans [56,57] and disseminated among humans around the world including the European countries and China (http://enterobase.warwick.ac.uk/species/ ecoli/search\_strains?query=st\_search (accessed on 7 October 2020)). Besides ST2944, all other STs in this study have been also found in human isolates (http://enterobase.warwick.ac.uk/species/ecoli/search\_strains?query=st\_search (accessed on 7 October 2020)). Serogroups of the ExPEC isolates from animals were rarely studied, although the APEC isolates from diseased poultry were often serotyped. In this study, O78 was the most predominant serogroup among ExPEC isolates from healthy chickens, followed by O26 and O86. This was slightly different from that of a previous study about APEC from Korea in which O78 was the most prevalent serogroup followed by O2 and O53 [15], both of which were not identified in our study. Notably, O78 was also a common serogroup among human ExPEC isolates from neonatal meningitis in Europe [58]. The distribution rate of each serogroup in this study was also different from that of APEC obtained between 2005 and 2008 in Guangdong, China [59]. The presence of serogroup O86 in two ExPEC from healthy chickens in this study are of interest since this serogroup was only identified in human ExPEC strains in Brazil rather than in strains from poultry in a previous report [57]. In the present study, one ExPEC isolate from healthy chicken belonged to O18 serogroup, which was frequently found in APEC from diseased avian in the United States and also the main serogroup of ExPEC isolates causing newborn meningitis in the Europe [58,60,61]. O83 and O45, which were prevalent in neonates with *E. coli* meningitis from the Netherlands [61] and France [62], respectively, were also found in our ExPEC isolates. All these results show that the ExPEC isolates from healthy chickens in this study might transmit to humans and the prevalence of ExPEC isolates in healthy animals should be monitored in the future.

Notably, all the ExPEC isolates in this study were MDR isolates. Although the use of antibiotics in animals does not select ExPEC strains intrinsically [25], it will favor the dissemination of ExPEC with MDR phenotypes among healthy animals, contributing to the emergence of MDR ExPEC in human infections [22]. Almost all ExPEC isolates in this study were resistant to ampicillin (100%), sulfamethoxazole-trimethoprim (100%), tetracycline (95.5%), and doxycycline (90.9%), which were all used extensively in animal husbandry in China, further favoring the dissemination of ExPEC among animals and humans by co-selection. In recent years, ExPEC isolates producing ESBLs or AmpC in human infections have been increasing [63], and such pathogens have been also found in healthy poultry in Brazil recently [14]. In our study, CTX-M-type ESBLs were found in 15 of the 22 ExPEC isolates from healthy chickens. The increase in ESBLs or AmpC among ExPEC from poultry will inevitably reduce the therapeutic options of ExPEC infections in humans, because cephalosporins are important to human medicine. Fosfomycin has been widely recommended for treating uncomplicated urinary tract infection especially caused by ESBLs-producing or fluoroquinolone-resistant ExPEC isolates [64]. However, ten ExPEC isolates in this study harbored the fosfomycin resistance gene *fosA3* and all ten isolates also carried *bla*CTX-M, among which six were resistant to fluoroquinolones (Table 4). Such pathogens in healthy chickens will pose a great threat to human health because they will compromise the efficacies of fosfomycin, fluoroquinolones and cephalosporins. For ExPEC from different markets/farms, three isolates (YJ-JC-8, WF1-5-21 and LS-A-7) carrying virulence genes *papA* and *iutA* belonged to serogroup O78 and they were all ST117 type, however, different PFGE and resistance profiles were present in the three isolates (Figure 1 and Table 4). This indicates that the ExPEC isolates have undergone a complex evolutionary process resulting in genetically diverse isolates although they share identical ST type and serogroup at the beginning. Even for ExPEC isolates with identical PFGE pattern and ST type from the same farm, such as WF1-5-13 and WF1-5-21, different resistance genotypes and phenotypes were also observed, further proving the complex evolutionary process within ExPEC.

Worryingly, besides *mcr-1* and *bla*NDM, the two ExPEC isolates WF1-5-13 and WF1-5-40 carried additional *bla*CTX-M, *fosA3* and *floR* genes, with isolate WF1-5-40 also harboring *rmtB* (Table 4). The presence of such ExPEC isolates co-harboring *bla*NDM and *mcr-1* in healthy chickens in this study will threaten the health of consumers because such pathogens will not only compromise the efficacies of cephalosporins, fosfomycin and aminoglycosides, but also threaten the usage of carbapenems and colistin, two critically important antimicrobials used for serious infections caused by MDR ExPEC [65]. *mcr* has been also found in two ExPEC isolates from diseased poultry in Brazil [18] and two ExPEC isolates from healthy ducks in China [66]. All four ExPEC isolates carrying *mcr* from animals in the two previous reports were susceptible to carbapenems, although NDM-producing ExPEC isolates susceptible to colistin have been reported in humans [67]. To the best of our knowledge, this is the first report about *mcr-1*-positive ExPEC isolates harboring *bla*NDM from healthy chickens.

#### **5. Conclusions**

In conclusion, we observed that the resistances in *E. coli* from white-feather broilers were more serious than those from layer farms and those from live-poultry markets in China, respectively. This study also reported that 2.4% of these *E. coli* isolates from healthy chickens were qualified as ExPEC using a molecular detection method. The most predominant serogroup of these ExPEC isolates was O78, followed by O26 and O86, and almost all serogroups identified in our study were frequently reported in human ExPEC isolates in many countries, suggesting that ExPEC isolates from healthy poultry could be a source of potentially virulent ExPEC causing multiple diseases in humans. Notably, all the ExPEC isolates in this study possessed MDR phenotypes and most showed resistances to cephalosporins and fosfomycin, which made co-selection of these ExPEC possible when corresponding drugs were used. More worryingly, six ExPEC isolates in this study carried *mcr-1*, including two harboring both *bla*NDM and *mcr-1*, which could compromise both the efficacies of carbapenems and colistin. The presence of MDR ExPEC isolates in healthy chickens, especially those carrying *mcr-1* and/or *bla*NDM, is alarming and will

pose a serious health threat to consumers. Interventions need to be taken to reduce these pathogens in the chicken intestine and prevent clinical ExPEC infections in humans by reducing transmission via poultry products. Further studies are required for monitoring the prevalence of MDR ExPEC in healthy chickens in China and other countries. To our knowledge, this is the first report of *mcr-1*-positive ExPEC isolates harboring *bla*NDM from healthy chickens.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/ani11041112/s1, Table S1: Origins of the samples collected in this study.

**Author Contributions:** B.-T.L. designed the study, secured the funds, and supervised the whole work. M.Z. collected the samples and carried out the laboratory work. P.-P.M., W.-S.L. and X.L. also contributed in the laboratory work. B.-T.L., M.Z., P.-P.M. and X.-Y.L. wrote the original draft of the manuscript. B.-T.L., M.Z. and Y.-Z.L. did the data analysis and interpretation. B.-T.L. and M.Z. critically reviewed and edited the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was supported by grants from the Scientific and Technological Projects of Qingdao (19-6-1-94-nsh and 21-1-4-ny-11-nsh), the Shandong Major Science and Technology Innovation Projects (2019JZZY010719), Major Agricultural Application Technology Innovation Project of Shandong Province (SD2019XM004) and the Scientific and Technological Projects of Qingdao (21-1-4-ny-10-nsh).

**Institutional Review Board Statement:** All animal experiments were carried out in accordance with guidelines issued by the Qingdao Agricultural University Animal Care and Use Committee (approval number, QDAUA-2015-033).

**Data Availability Statement:** No new data were created or analyzed in this study. Data sharing is not applicable to this article.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article S. aureus* **Biofilm Protein Expression Linked to Antimicrobial Resistance: A Proteomic Study**

**Cristian Piras 1,†, Pierluigi Aldo Di Ciccio 2,†, Alessio Soggiu 3, Viviana Greco 4,5, Bruno Tilocca 1, Nicola Costanzo 1, Carlotta Ceniti 1, Andrea Urbani 4,5, Luigi Bonizzi 3, Adriana Ianieri <sup>6</sup> and Paola Roncada 1,\***


**Simple Summary:** Biofilm formation represents one of the most effective forms of bacterial persistence in surfaces where nutrients are available or in the tissues of living hosts as humans or animals. Such persistence is due to the high rate of antimicrobial resistance of this shell conformation. It often represents a burden when the pathogen colonizes niches from where it is not removable such as food facilities, farm facilities or parts of living organisms. In this study, we investigated biofilm formation mechanisms and enhanced antimicrobial resistance of 6 different *S. aureus* strains. The detected mechanisms were primarily related to the control of catabolites, the production of proteins with moonlighting activities and the detoxification of compounds with antimicrobial activities (i.e., alcohol). Glycolysis and aerobic metabolisms were found to be less active in the biofilm conformation. Consequently, less H2O2 production from aerobic metabolism was translated into a measurable under-representation of catalase protein.

**Abstract:** Antimicrobial resistance (AMR) represents one of the most critical challenges that humanity will face in the following years. In this context, a "One Health" approach with an integrated multidisciplinary effort involving humans, animals and their surrounding environment is needed to tackle the spread of AMR. One of the most common ways for bacteria to live is to adhere to surfaces and form biofilms. *Staphylococcus aureus* (*S. aureus*) can form biofilm on most surfaces and in a wide heterogeneity of environmental conditions. The biofilm guarantees the survival of the *S. aureus* in harsh environmental conditions and represents an issue for the food industry and animal production. The identification and characterization of biofilm-related proteins may provide interesting insights into biofilm formation mechanisms in *S. aureus*. In this regard, the aims of this study were: (i) to use proteomics to compare proteomes of *S. aureus* growing in planktonic and biofilm forms in order to investigate the common features of biofilm formation properties of different strains; (ii) to identify specific biofilm mechanisms that may be involved in AMR. The proteomic analysis showed 14 differentially expressed proteins among biofilm and planktonic forms of *S. aureus*. Moreover, three proteins, such as alcohol dehydrogenase, ATP-dependent 6-phosphofructokinase, and fructose-bisphosphate aldolase, were only differentially expressed in strains classified as high biofilm producers. Differentially regulated catabolites metabolisms and the switch to lower oxygenrelated metabolisms were related to the sessile conformation analyzed.

**Citation:** Piras, C.; Di Ciccio, P.A.; Soggiu, A.; Greco, V.; Tilocca, B.; Costanzo, N.; Ceniti, C.; Urbani, A.; Bonizzi, L.; Ianieri, A.; et al. *S. aureus* Biofilm Protein Expression Linked to Antimicrobial Resistance: A Proteomic Study. *Animals* **2021**, *11*, 966. https://doi.org/10.3390/ ani11040966

Academic Editor: Amit Vikram

Received: 3 March 2021 Accepted: 24 March 2021 Published: 31 March 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

**Keywords:** *Staphylococcus aureus*; planktonic cells; biofilm; proteomics; food safety; antimicrobial resistance

#### **1. Introduction**

Humanity is already facing a challenge related to antimicrobial resistance (AMR). Such a burden will become worse due to the massive use of antimicrobials such as alcohol-based products for hands and workplaces sanitization necessary to mitigate the transmission of coronavirus disease 2019 (COVID-19). These key precautions may create an ecological pressure on microorganisms and contribute to the emergence of AMR in microbial populations that can colonize human body and the environment.

The use of biocides in the environment (such as farms and food industries) has already created a phenomenon known as AMR cross-resistance [1–3]. Biofilm formation contributes to enhance AMR resistance by physical and biochemical means [4]. A biofilm is defined as "a microbially derived sessile community characterized by cells that are irreversibly attached to a substrate or interface or to each other, are embedded in an autoproduced matrix of extracellular polymeric substances (which is composed of protein, DNA and polysaccharide) and exhibit an altered phenotype with respect to growth rate and gene transcription" [5]. It is well known that bacteria growing as biofilms might be physiologically distinct from the same bacteria growing as free-swimming planktonic cells [6,7].

Briefly, biofilms allow bacteria to better resist harsh environmental conditions [8]. Such a conformation can be found everywhere where there is a source of nutrients such as in the food-processing environment or zootechnical industry (food-processing equipment, milk collection and storage facilities) [9]. Biofilms-enhanced resistance to disinfectants/antimicrobials/antibiotics represents a threat for food industries and farms [10]. The biofilm, in fact, protects the bacteria from detaching by cleaning agents and from being killed by disinfectants [11]. However, biofilm protection mechanisms appear to be different from those responsible for resistance to conventional antibiotics [12]. First, the extracellular polymeric substances (EPS) matrix delays or prevents antimicrobial action, either by limiting disinfectants diffusion or by chemical interaction/inactivation with proteins and extracellular polysaccharides [13]. Other factors can play a role in this feature, such as the bacterial growth rate, the heterogeneity within the biofilm, the general stress response, quorum sensing mechanisms, the induction of a certain biofilm phenotype and the over-expression of efflux pumps [14]. In addition, biofilm activities include the upregulation of virulence factors and secretion of extracellular polymers [15]. Horizontal gene transfer plays an important role in AMR. The small intra-cellular distance typical of biofilms facilitates the spread of resistance genes and generates the presence of extracellular DNA in the biofilm matrix [16].

Among bacteria, *Staphylococcus aureus* (*S. aureus*) is able to form biofilm on most surfaces and under almost all the environmental conditions found in food industries [17]. It is a commensal and opportunistic pathogen and under certain conditions, may cause a wide range of infectious diseases such as skin infections, bacteremia, endocarditis, pneumonia and food poisoning. *S. aureus* biofilm mode of growth is regulated by complex genetic factors and can produce at least two different types of biofilm: ica operon-dependent (i.e., promoted by the ica operon) and ica operon-independent [17]. A study carried out by Resch et al. (2005) identified more than 160 genes that were significantly over-expressed during biofilm growth conditions. Those genes encoded for binding factors, polysaccharide intracellular adhesion (PIA) and peptidoglycan modeling factors [7]. Additionally, many proteins have been implicated as important components in cellular adhesion and biofilm matrix development [18]. These include surface-associated proteins (protein A), fibrinogenbinding proteins (FnBPA and FnBPB), biofilm-associated protein (Bap) and clumping factor B (ClfB).

Considering the concerns for food safety associated with *S. aureus* biofilms and the high cost of managing this issue in the food industry, a better knowledge of the mechanisms involved in *S. aureus* biofilm growth mode is essential. To date, several studies have focused on pathogenicity and only a few have addressed differences in protein expression of *S. aureus* due to biofilm formation [19,20]. The identification and characterization of proteins linked with biofilm could provide interesting insights on the mechanism and/or process of biofilm formation in *S. aureus*.

According to this premise, the aims of this study were: (i) to compare proteomes of *S. aureus* growing in planktonic and biofilm forms, in order to investigate the common features of biofilm formation properties of six different strains; and (ii) to identify possible biofilm mechanisms that may be involved in AMR. The employment of 6 different strains will help with the comprehension of biofilm formation mechanisms more representative of the *S. aureus* species rather than be focused on mechanisms typical of a single strain.

#### **2. Materials and Methods**

#### *2.1. Bacterial Strains*

A total of six biofilm-forming *S. aureus* strains were analyzed in this study. In details, three *S. aureus* reference strains (ATCC 35556, ATCC 12600, ATCC 29213) and three foodrelated isolates (wild-types) were used in the experiment. The food related-strains were isolated from food (n.1) and food handlers (n.2), respectively.

Stock cultures were stored at −80 ◦C. All strains were incubated for 24 h at 37 ◦C in tryptone soy broth (TSB, Oxoid S.p.A., Milan, Italy) before each experiment. All these strains have been grown both in the planktonic and in the sessile form (biofilm cultures) and analyzed through 2D electrophoresis coupled with matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS).

The sessile (biofilm cultures) and planktonic cells were prepared as follows:


The ability of *S. aureus* isolates to produce biofilms was determined according to the protocol described by Di Ciccio et al., 2015 [21]. In all cases, all experiments were repeated in triplicate. Briefly, polystyrene tissue culture plates (6 wells—961 mm2) were used as substratum for biofilm formation at 37 ◦C. Cultures of *S. aureus* were prepared, from overnight tryptone soy agar (TSA, Oxoid S.p.A., Milan, Italy) growth, in TSB by incubating at selected temperature: 37 ◦C. Cultures were then washed three times with sterile phosphate-buffered saline (PBS; pH 7.3) (Sigma-Aldrich S.r.l., Milan, Italy) and diluted with fresh TSB to reach a concentration of about 10<sup>8</sup> colony-forming units (CFU) mL−<sup>1</sup> by reading the optical density (OD) level at 550 nm (UV Mini-1240—Shimadzu, Long Beach, CA, USA). Three milliliters (ml) of the standardized inoculum were then added to polystyrene tissue culture plates (well—35 mm diameter). Samples were then incubated at 37 ◦C. After 24 h incubation, non-adherent cells were removed by washing each well three times with sterile PBS. After adding sterile PBS (3 mL), biofilm in wells was dislodged mechanically by scraping vigorously using a sterile cell-scraper. Finally, the cells were harvested by centrifugation (4000 rpm, 10 min., 4 ◦C, Beckman, J2-MC, centrifuge). The resulting pellets, washed and resuspended in sterile PBS, were centrifuged again (4000 rpm, 10 min., 4 ◦C). The cells were washed several times and pelleted by five centrifugations. Finally, the supernatant was removed and the pellet from the biofilm cultures grown was stored at −80 ◦C until use for proteomic studies (the pellets from the biofilm cultures had a weight of 50 mg).


*S. aureus* reference strains (ATCC 35556, ATCC 12600, ATCC 29213) and food-related isolates *S. aureus* (281, 402, 184) were used. An overnight culture was created by inoculating a colony of *S. aureus* into 5 mL of TSB for 24 h at 37 ◦C. After incubation, the *S. aureus* culture was centrifuged for 10 min at 4000 rpm, 4 ◦C. The supernatant was then replaced with sterile PBS, and pellet was resuspended by thoroughly mixing with pipette. The cells were washed several times and pelleted by five centrifugations (4000 rpm, 10 min, 4 ◦C). Finally, the supernatant was removed and the pellet from the overnight cultures grown was stored at −80 ◦C until use for proteomic studies (the pellets from the planktonic cultures had weights: 50 mg).

#### *2.2. Proteomic Analysis*


We diluted 50 milligrams of cellular pellet of the different *S. aureus* strains in 700 μL of 2DE buffer containing 7 M urea, 2 M thiourea, 4% CHAPS, 1% DTT, and protease inhibitors (GE-Healthcare) according to manufacturer instructions.

To ensure the complete disruption of the collected bacterial cells, the samples were processed with 6 cycles of 1-min bead beating interspersed by a cycle of centrifuge. For this purpose, into the sample was added the same amount (1:1 *v/w*) of 0.1 mm zyrcouniumsilica beads (300 μg beads added to 300 μL of buffer + the volume of the pellet). The bead beating cycle was conducted at 4000 rpm for 1 min with the purpose to avoid overheating. Then, the samples were centrifuged at 12,000× *g* for 5 min at 4 ◦C in order to chill and disperse the foam. This operation was repeated 6 times. After the 6th cycle, samples were centrifuged for 20 min and the supernatant was stored in another tube for subsequent proteomics analysis.

Two-dimensional (2D) electrophoresis was run in all samples: 100 micrograms of protein were loaded on a 7 cm strip through active rehydration performed overnight at 50 V in a buffer containing 7 M urea, 2 M thiourea, 2% CHAPS, 0.5% ampholytes 3–10 Amersham, and 26 mM DTT. For isoelectric focusing (IEF), the following protocol was applied: 100 V/1 h linear, 250 V/2 h linear, 4000 V/5 h linear, 4000 V step/50,000 total volt-hours (VhT), using a protean IEF platform.

Once the final amount of VhT was reached, immobilized pH gradient (IPG) strips were frozen up to the next step or directly equilibrated in two steps of 15 min under gentle stirring. The first step of equilibration was performed in buffer (6 M UREA, 2% SDS, 0.05 M Tris-HCl pH 8.8, 20% glycerol) supplemented with 1% DTT *w/v* and the second step was performed in a buffer with the addition of 2.5% *w/v* iodoacetamide. The IPG strips were put in a 12% home-made acrylamide gel and IEF run under constant amperage of 15 mA per gel, until the bromophenol blue (BFB) reached the front. The gels were then eliminated from the plates, washed three times with double-distilled water and spotted overnight (ON) with Coomassie Brilliant Blue.

Using an Imagescanner III (GE Healthcare) the gels were digitalized. The image analysis was performed using SameSpots software (Version 4.5, Nonlinear Dynamics U.K.). All imported images were checked for quality (saturation, ending) and spots, with a *p*-value lower than 0.05, were manually excised for subsequent mass spectrometry (MS) analysis and protein identification. If the MALDI MS/MS identification was obtained with a MASCOT score higher than 40, the protein was analyzed via GO for the comprehension of its function/role.


Protein identification was performed according to previous studies [22,23].

Briefly, after different steps of dehydration, reduction and alkylation, the excided single spots were digested with a solution of 0.01 μg/μL of porcine trypsin (Promega, Madison, WI, USA) at 37 ◦C o.n., and peptides were concentrated using C18 ZipTip (Millipore, Bedford, MA, USA). they were then co-crystallized with a solution of αciano-4-hydroxycinnamic acid and spotted on a Ground Steel plate (Bruker-Daltonics, Bremen, Germany).

The MS analysis was performed on a Ultraflex III MALDI-TOF/TOF spectrometer (Bruker-Daltonics) in positive reflectron mode.

External calibration was performed using the standard peptide mixture calibration (m/z: 1046.5418, 1296.6848, 1347.7354, 1619.8223, 2093.0862, 2465.1983, 3147.4710; Bruker-Daltonics).

FlexAnalysis 3.3 software (Bruker-Daltonics) was used for the selection of the monoisotopic peptide masses of each mass spectra. After an internal calibration on autolysis peaks of porcine trypsin (m/z: 842.509 and 2211.104) and exclusion of contaminant ions (known matrix and human keratin peaks), the created peak lists were analyzed by MASCOT v.2.4.1 algorithm (www.matrixscience.com, accessed on 23 March 2021) searching against SwissProt 2021\_database restricted to Firmicutes and *Staphylococcus aureus* (11,196 sequences) taxonomy.

The parameters used for database search are the following: carbamidomethylation of cysteines and oxidation on methionine as fixed and variable modifications respectively; one missed cleavage site allowed for trypsin; 70 ppm as maximal tolerance.

Mascot protein scores greater than 50 were considered significant (*p* < 0.05) for protein identification assignment.

To confirm the identifications, MS/MS spectra were also acquired by switching the instrument in LIFT mode with 4–8 × <sup>10</sup><sup>3</sup> laser shots. For the fragmentation, precursor ions were manually selected, and the precursor mass window was automatically set. Spectra baseline subtraction, smoothing (Savitsky–Golay) and centroiding were operated using Flex-Analysis 3.3 software.

The parameters used for the database search are the following: carbamidomethylation of cysteines and oxidation on methionine as fixed and variable modifications respectively; one missed cleavage; 50 ppm and 0. 5 Da as mass tolerance for precursor ions and for fragments respectively. The taxonomy was restricted to *Staphylococcus aureus* (10,227 sequences).

The confidence interval for protein identification was set to 95% (*p* < 0.05) and only peptides with an individual ion score above the identity threshold were considered correctly identified.

#### **3. Results**

The proteomic analysis was performed in order to discover the mechanisms of biofilm formation common to all analyzed *S. aureus* strains. Six different strains with different biofilm formation indexes were analyzed in the planktonic form and the biofilm form. For each strain, biofilm formation, expressed as BPI, was calculated as follows: "BPI = [ODmean biofilm surface (mm2) <sup>−</sup>1] × 1000". All isolates were defined in different categories (weak, moderate or strong producers) on the basis of their BPIs values (Table 1).


**Table 1.** Biofilm formation index (BPI) of *S. aureus* strains on polystyrene at 37◦ included in this study.

The analyzed strains included: *S. aureus* ATCC 35556, already described as a strong biofilm producer [24,25]; *S. aureus* ATCC 12600, classified as moderate biofilm producer [21]; three food isolates of *S. aureus* classified as strong (281), moderate (402) and weak biofilm producer (184); *S. aureus* ATCC 29213 measured as weak biofilm producer. BPI on polystyrene at 37 ◦C was used as the measure for all the experimental procedures in this work. All the strains with BPI below 0.300 were considered weak biofilm producers. In these cases, the biofilm layer was phenotypically barely visible and not stable in its structure. Such a phenotype was confirmed by the extremely low BPI below 300. For

this reason, four strains (A, B, C and D) were considered as part of the moderate/high biofilm-producing group, while the remaining two (E and F) showed a phenotype closer to the low/non-forming biofilm group.

Proteomics analysis was carried out to compare the sessile versus the planktonic phenotype; however, a separated analysis was performed, including only the moderated to strong biofilm producers. The differentially represented proteins were chosen according to the Progenesis SameSpots provided analysis of variance (ANOVA) *p*-value. The topmost significant ones were chosen to be analyzed via MALDI-TOF MS/MS peptide mass fingerprinting (PMF) and peptide fragment fingerprinting (PFF) if necessary. Of the chosen spots, only the ones successfully identified with a MASCOT score higher than 40 were considered for subsequent Gene Ontology (GO), metabolism and pathway analysis.

As reported in Table 2, 14 proteins were differentially expressed among *S. aureus* planktonic and sessile groups. Of these, 11 were differentially expressed when considering all the strains together with a *p*-value ≤ 0.05 (column: regulation in planktonic vs. sessile). Alcohol dehydrogenase, ATP-dependent 6-phosphofructokinase and Fructose-bisphosphate aldolase differential expression were significant for the medium/high biofilm-forming sub-group (high biofilm producers' column). This classification was done according to the observation of the datasets that clearly showed how the representation trend of some of the differentially expressed proteins was clearly not following the same path in the weak biofilm forming strains. As previously mentioned, this was the case for alcohol dehydrogenase, ATP-dependent 6-phosphofructokinase and fructose-bisphosphate aldolase.

If considering the entirety of the differentially regulated proteins, five were found to be over-represented in the sessile versus planktonic group, and 9 proteins were found to be under-represented. This low number of detected proteins might be due to the high heterogeneity of the different strain analyzed. Three of the five over-represented proteins were involved in carbon metabolism or in stress response. Interestingly, alcohol dehydrogenase and 30 s ribosomal proteins are involved in antimicrobials resistance mechanisms, i.e., detoxification.

On the other hand, under-represented proteins such as 2,3-bisphosphoglyceratedependent phosphoglycerate mutase, alkyl hydroperoxide reductase subunit C, ATPdependent 6-phosphofructokinase, catalase etc. were mostly involved in energy and oxygen-related metabolism.

All data are shown in Table 2 and the image of the differentially represented proteins is shown in Figure 1a. For each protein, it is represented the related figure from the image analysis software. Table 2 indicates the *p*-values obtained from the built-in ANOVA test of the Progenesis SameSpots software. For each protein it is provided with the UNIPROT name and accession number (first two columns of the table); the SameSpots coding number, which represents the code provided by the image analysis software; the Mascot score identification obtained by the combined MALDI peptide mass fingerprint together with the peptide fragment fingerprint for the MS/MS identification; the number of matched peptides and the mascot score; and the ANOVA *p*-value obtained by comparing the planktonic and sessile form of all strains and just moderate/high biofilm producers (last column, the values of normalized volume for each spot are provided in Supplementary Materials, Table S1).

**Table 2.** List of differentially represented proteins among the six different strains analyzed under planktonic and biofilm conditions. As in the last two columns, the analysis was performed, including all the analyzed strains and, subsequently, excluding the low biofilm producers (last column). OS= organism name. Every significant *p*-value (lower than 0.05) is printed in bold.


Figure 1a provides a graphic representation of the Coomassie Brilliant Blue stained entire proteins as detected by the image analysis software. The top four rows show high and moderate biofilm producers' spots, while the two rows at the bottom indicate the low biofilm producers.

Figure 1b shows the graphic representation of the most relevant differentially regulated proteins and metabolisms among the two analyzed *S. aureus* phenotypes. Biological functions were manually checked after each GO search and subsequently reported in the scheme in Figure 1b.

**Figure 1.** (**a**) Graphic representation of the differentially expressed proteins mostly relevant to the regulation of the described mechanisms/pathways. (**b**) Representation of the differentially regulated proteins and the related modulated mechanisms.

#### **4. Discussion**

Biofilms growth is the preferred strategy for the expansion and survival of many clinically and environmentally relevant microorganisms [5]. *S. aureus* is one well-known biofilm-forming pathogen capable of colonizing medical devices [26], food contact surfaces [21] and farm industry facilities [27]. In the biofilm form, *S. aureus* can successfully cope against strong stress conditions [28] and persist on the surfaces of food-processing plants [4], leading to recurrent contamination of both fresh and processed foods worldwide [29–32]. From this perspective, biofilm formation represents a severe threat because of its difficult removal linked to the extremely high tolerance to antimicrobials and antibiotics. Improving knowledge about its formation mechanisms and pathways is mandatory to better design possible and practical intervention strategies. Studies performed on single strains (strain-specific mechanisms) documented the over-representation of fibrinogen-binding protein and the accumulation-associated protein (Aap) in *S. aureus* cells growing embedded in the biofilm matrix in comparison to those growing in the planktonic form [33,34]. Also, increased production of staphylococcal accessory regulator A (SarA) was shown in biofilm formation [20].

All these and many other studies extensively describe the physiology of *S. aureus* biofilm formation that is specific to the strain analyzed. However, it is not considered that diverse *S. aureus* strains may have different mechanisms and pathways of biofilm formation.

In the current study, we employed a comparative proteomic approach to understand better the process of biofilm formation and the possible mechanisms involved in the enhancement of antimicrobial resistance. To achieve this result, we performed a differential proteomics analysis of planktonic versus sessile *S. aureus* isolates and ATCC strains. Six different strains with a wide range of biofilm formation indexes were employed in order to maximize the possibility to detect general mechanisms more representative of the *S. aureus* specie.

The whole comparison allowed the discovery of 14 proteins differentially regulated between the planktonic and sessile group and, three of those (alcohol dehydrogenase, ATP-dependent 6-phosphofructokinase and fructose-bisphosphate aldolase) were specific to the high biofilm-producing strains.

Ribosomal proteins are involved in biofilm regulation/formation and enhanced antimicrobial resistance [35,36]. Interestingly, changes at the ribosomal protein isoforms can shape the response to antibiotics by modifying the affinity of tetracyclines, chloramphenicol, macrolides (e.g., erythromycin) and aminoglycosides (e.g., streptomycin) for the transcription machinery. Hence, a switch in the composition of ribosomal subunits could be involved in biofilm formation and the different susceptibility to antimicrobial molecules [37].

Fructose bisphosphate aldolase and catabolite control protein A (ccpA) are overrepresented in the biofilm conformation versus the sessile condition. The first is an essential enzyme of the glycolytic pathway with virulence functions shaped according to its cellular localization (i.e., moonlighting properties) [38]. As a moonlight protein, it is often expressed in the bacterial surface [39] where it has been linked to virulence in several bacterial pathogens, such as *Francisella* [40], by directly affecting cell migration through its interference with the actin polymerization process.

Similarly, fructose bisphosphate aldolase expression is induced in oxygen depletion conditions, and it has also been associated with transcriptional regulator functions [39]. Catabolite control protein A (ccpA) was found to be massively over-represented in both high and low biofilm producers growing in the sessile conditions. This might be explained by the requirements of the typical multi-layered packed structure of the biofilm, which needs a tight control of nutrients availability, catabolites and secondary metabolites (e.g., ethanol, reactive oxygen species (ROS) etc.). Indeed, nutrients depletion or catabolites accu-

mulation would exert toxic/detrimental effects on the bacterial community itself. In Gram+ bacteria, ccpA expression regulates the synthesis of capsular polysaccharides, toxigenic exoproteins and promotes biofilm formation [25]. Similarly, *S. epidermidis* biofilm formation is positively regulated by ccpA and causes tricarboxylic acid (TCA) cycle repression [41]. This demonstrates that the management of carbon and energy flow by regulating the enzymes involved in glycolytic/fermentative metabolism [42] represents an essential element for the proper formation of biofilm. Accordingly, previous evidence reported that environmental acidification or other phenomena associated with rapid metabolism of carbohydrates occurring in bacteria growing in sessile conditions are regulated by ccpA throughout the modulation of pfka and gpma expression [42,43]. Moreover, the structural organization of the biofilm is likely to result in the accumulation of toxic secondary metabolites such as ethanol from fermentation processes. This may explain the detected increased expression of alcohol dehydrogenase (adh) in the sessile growing strains. The oxygen depletion in the biofilm's inner layers may cause a metabolic shift towards the mixed alcoholic fermentation with increased ethanol concentration that needs to be promptly detoxified by the induction of adh [44–46]. The hypothesis of the metabolic shift towards fermentation and ethanol production is also supported by the under-expression of PfkA and gpmA, which are active in pyruvate production. By limiting the production of pyruvate, sessile cells control the pathways towards any possible fermentative process. Thus, the reduced abundance of PfkA and gpmA in the sessile bacteria might represent the effect of a negative feedback modulation of the fermentative process to protect the bacteria from the toxicity of their secondary metabolites. Analogously, the reduced abundance of catalase, the enzyme active in ROS detoxification, may be a consequence of the reduced oxygen availability in the bacterial samples growing in biofilm form [47]. Such a reduction in the hydroperoxide detoxification power is confirmed by the down-regulation of 3 different catalase isoforms and of alkyl hydroperoxide reductase subunit C (Q2FJN4). This may help to explain the high power of low doses of hydrogen peroxide to dissolve the biofilm conformation [48].

#### **5. Conclusions**

The comparative top-down proteomics (2D-electrophoresis–MALDI TOF) approach used here identified some possible biofilm formation mechanisms of *S. aureus* strains with a wide range of biofilm formation indexes. Biofilm is one of the essential strategies for bacterial virulence and persistence over a wide variety of surfaces and unfavourable conditions and it facilitates survival and resistance in the presence of antimicrobial compounds [49]. Comparison of high- and low-biofilm forming strains in sessile and planktonic form highlighted common mechanisms as the catabolite control and the modulation of the detoxification machinery aimed at avoiding self-inhibition/toxicity (i.e., ethanol detoxification). Glycolysis and aerobic metabolisms seem to be down-regulated in favor of possible fermentation pathways that might be responsible for ethanol production and, possibly, for the induction of alcohol dehydrogenase production.

This study is characterized by using a top-down proteomics approach that led the differential quantification of intact proteofoms. On the other hand, this approach limits the detection of differentially represented, less-abundant proteins. Complementing these data with shotgun proteomics and metabolomics is desired to support the observed evidence and to discover potential biomolecular targets to contrast and/or attenuate this phenomenon.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/ani11040966/s1, Table S1: Raw normalized volume.

**Author Contributions:** Conceptualization, C.P., P.A.D.C., A.S., P.R.; Data curation, C.P., P.A.D.C., A.S., V.G.; Formal analysis, C.P., P.A.D.C., A.S., V.G.; Funding acquisition, A.I., L.B., A.U., P.R.; Investigation, C.P., P.A.D.C., A.S., V.G., B.T.; Methodology, C.P., P.A.D.C., A.S., V.G., P.R., Project administration, A.I., P.R.; Resources, A.I., P.R., L.B., A.U.; Writing—original draft, C.P., P.A.D.C., A.S., P.R.; Writing—review and editing, C.P., P.A.D.C., A.S., V.G., B.T., N.C., C.C., A.I., P.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


#### *Article*
