Presence of β-Lactamase-producing Enterobacterales and Salmonella Isolates in Marine Mammals

Marine mammals have been described as sentinels of the health of marine ecosystems. Therefore, the aim of this study was to investigate (i) the presence of extended-spectrum β-lactamase (ESBL)- and AmpC-producing Enterobacterales, which comprise several bacterial families important to the healthcare sector, as well as (ii) the presence of Salmonella in these coastal animals. The antimicrobial resistance pheno- and genotypes, as well as biocide susceptibility of Enterobacterales isolated from stranded marine mammals, were determined prior to their rehabilitation. All E. coli isolates (n = 27) were screened for virulence genes via DNA-based microarray, and twelve selected E. coli isolates were analyzed by whole-genome sequencing. Seventy-one percent of the Enterobacterales isolates exhibited a multidrug-resistant (MDR) pheno- and genotype. The gene blaCMY (n = 51) was the predominant β-lactamase gene. In addition, blaTEM-1 (n = 38), blaSHV-33 (n = 8), blaCTX-M-15 (n = 7), blaOXA-1 (n = 7), blaSHV-11 (n = 3), and blaDHA-1 (n = 2) were detected. The most prevalent non-β-lactamase genes were sul2 (n = 38), strA (n = 34), strB (n = 34), and tet(A) (n = 34). Escherichia coli isolates belonging to the pandemic sequence types (STs) ST38, ST167, and ST648 were identified. Among Salmonella isolates (n = 18), S. Havana was the most prevalent serotype. The present study revealed a high prevalence of MDR bacteria and the presence of pandemic high-risk clones, both of which are indicators of anthropogenic antimicrobial pollution, in marine mammals.


Introduction
Antimicrobial resistance (AMR) in bacteria is an emerging public health concern worldwide. Particularly, AMR among Enterobacterales is globally widespread and constitutes a public health threat due to a high prevalence of multidrug-resistant (MDR) strains [1,2]. MDR Gram-negative bacteria may complicate antimicrobial therapy. In addition to resistance classification of bacteria (multidrug resistance (MDR), extensive drug resistance (XDR), and pandrug resistance (PDR)) in the last decade, several novel classification criteria that correlate well with clinical outcomes were introduced (e.g., difficult-to-treat resistance (DTR)). This additional classification criterion will lead to significant benefits in assessing the significance of bacterial resistance in therapeutics [3,4].
Despite the intensively studied occurrence of AMR in bacteria from humans and domestic animals, there is a lack of knowledge about the distribution of AMR among bacteria from marine mammals. The Northeastern Pacific Ocean is a rich habitat that sustains a massive amount of phytoplankton, zooplankton, krill, and fish, all of which support a diverse and abundant community of more than 30 species of marine mammals [10]. Previous observations indicate that the marine environment may act as a reservoir of antimicrobial resistance genes [11][12][13][14]. Different mechanisms can lead to the presence of antimicrobial-resistant bacteria in the oceans: first, anthropogenic antimicrobial-containing runoff, challenging native bacteria to become resistant; second, coastal runoff of already resistant bacteria from terrestrial sources; and third-although less important in this context-antimicrobial production by marine microorganisms [11,15].
Wildlife sampled in habitats impacted by livestock production, solid waste disposal, wastewater treatment, and other anthropogenic influences have been found to result in a higher prevalence of antimicrobial-resistant bacteria compared to less influenced habitats [16]. The global use of antimicrobial agents reaching the oceans via municipal sewage, commercial fish farms, and hospitals [17,18] is likely to increase AMR in bacteria from marine mammals and marine microbial communities [19]. It has been reported that wildlife, in general, might act as an indicator of the burden of resistance within a local environment and could help to identify sources of anthropogenic contamination with antimicrobial resistance determinants [16,[20][21][22]. For these reasons, marine mammals, such as pinnipeds and dolphins residing in coastal waters, have been described as sentinels of the health of coastal marine ecosystems and, consequently, also the health of humans living in or visiting these areas [23][24][25]. Nevertheless, information on AMR in bacteria from marine mammals remains scarce [19,[26][27][28][29][30].
The present study aimed to determine the occurrence and to characterize ESBL-and AmpC β-lactamase-producing Enterobacterales, including non-resistant Salmonella spp., in stranded marine mammals temporarily housed in the Marine Mammal Care Center Los Angeles (MMCCLA), California. Included in this study were California sea lions (Zalophus californianus), northern fur seals (Callorhinus ursinus) and one Guadalupe fur seal (Arctocephalus townsendi) representing Otariid seals, northern elephant seals (Mirounga angustirostris) and harbor seals (Phoca vitulina) representing Phocid seals, and a longbeaked common dolphin (Delphinus capensis) and a pygmy sperm whale (Kogia breviceps) representing Cetaceans.
Seventeen of the 18 Salmonella isolates obtained in this study were susceptible to all tested antimicrobial agents. One Salmonella isolate (ID 42 S.) displayed an AmpC phenotype, but no further resistance to non-β-lactam antibiotics was observed.

Characterization of Genotypic Resistance
Within the group of β-lactamase genes, bla CMY (n = 51) was the most prevalent and was detected solely (n = 13) or together with bla CTX-M , bla OXA , bla SHV , and/or bla TEM . All sequenced bla CMY genes were identified as bla CMY-2 , except one E. coli isolate (ID 171) harboring bla CMY-4 . Sequencing of bla CMY was not carried out in intrinsically AmpCproducing Enterobacterales, such as Citrobacter spp. and Enterobacter spp. The gene bla TEM was detected in 38 isolates-solely (n = 1) or together with bla CMY , bla CTX-M , bla OXA, and/or bla SHV . All bla TEM genes were identified as bla TEM-1 . bla SHV genes were detected in 13 different isolates. The broad-spectrum β-lactamase genes bla SHV-11 (n = 3) and bla SHV-33 (n = 8) were detected in 11 of the 12 K. pneumoniae isolates, together with bla TEM and bla CMY (n = 8) or together with bla TEM , bla CMY, and bla OXA (n = 3). The ESBL gene bla SHV-12 (n = 2) was harbored by the two L. amnigena isolates (IDs 130, 132), which also carried the gene bla DHA-1 . The genes bla CTX-M-15 and bla OXA-1 were observed in seven isolates, respectively (Tables 1 and 2).  The gene bla CMY-2 was detected in the Salmonella isolate 42 S. In two of the E. coli isolates (IDs 3 and 202) displaying an AmpC phenotype, no β-lactamase gene was detected, but mutations in the promoter and attenuator region of the chromosomal ampC gene were found at positions -18 (G→A), -1 (C→T) and +58 (C→T) in E. coli isolate 3 and at position -32 (T→A) in E. coli isolate 202. Further, mutations were found at position -18 (G→A), -1 (C→T), and +58 (C→T) of the ampC gene of E. coli isolates 8/2 and 21, at positions -18 (G→A) and -1 (C→T) in E. coli 192/2, at position +58 (C→T) in E. coli 17, and at position -28 (G→A) of the ampC gene of E. coli isolates 197/1 and 224a.
The analysis of the quinolone resistance-determining regions (QRDR) of gyrA and parC revealed mutations that resulted in amino acid exchanges at positions 83 (Ser→Leu) and 87 (Asp→Asn) in GyrA of five ciprofloxacin-resistant isolates and at position 80 (Ser→Ile) in ParC of six ciprofloxacin-resistant isolates.
Even though phenotypic resistance to tetracycline was present, none of the tested tet genes were detected in C. freundii complex isolate 224b.
Antimicrobial resistance phenotypes and genes are summarized in Tables 1 and 2.

Molecular Typing Methods
The most common E. coli phylogenetic group was A (n = 8). Four E. coli isolates each were assigned to the groups B1, B2, D, and F, while three isolates represented phylogroup E (Tables 1 and 3).
The multidrug transporter gene mdfA, which confers resistance to a diverse group of biocides and antimicrobial agents [33], was detected in all 12 isolates analyzed by WGS.
The determination of plasmids via the PlasmidFinder software revealed that the predominant plasmid was IncFIB (n = 7). In the E. coli isolate 209, the bla TEM-1B gene was carried by an IncI1-Iα plasmid (Supplementary Tables S2 and S3).
The mlplasmids analyses predicted that the majority of the antimicrobial resistance genes might be plasmid-encoded, whereas most virulence genes appeared to be chromosomally-encoded (Supplementary Tables S3 and S4
In all Salmonella isolates (n = 18) the transcriptional regulator gene hilA and Salmonella enterotoxin gene stn were detected. The virulence gene tviA was not found.

Discussion
In this study, we investigated the phenotypic and genotypic characteristics of ESBLand AmpC-producing Enterobacterales, as well as the presence of Salmonella isolates in stranded marine mammals prior to their rehabilitation. The majority of the investigated βlactamase-producing Enterobacterales exhibited an AmpC phenotype, an ESBL phenotype was observed in five E. coli isolates, and both AmpC and ESBL phenotypes were present in two E. coli and two L. amnigena isolates. In the scope of this study, three types of AmpC resistance mechanisms have been identified: plasmid-mediated AmpC (pAmpC) genes, overexpression of chromosomal AmpC in E. coli, and de-repressed AmpC in Citrobacter spp. and Enterobacter spp. AmpC-producers have been reported as particularly common in North America, whereas in Europe, β-lactamases are primarily represented by ESBLs [34]. Among the bla genes investigated in this study, the pAmpC gene bla CMY was detected most frequently. It was in all K. pneumoniae and P. mirabilis isolates, in one Salmonella isolate, and in 18 E. coli isolates identified as bla CMY-2 , which encodes the most prevalent pAmpC enzyme in humans, livestock, and companion animals worldwide [7,9]. Moreover, bla CMY-2 has been detected in different wildlife species [35][36][37][38]. In two E. coli isolates (IDs 206 and 230a), bla CMY-2 is predicted to be encoded chromosomally, which was described in 2015 for the first time in E. coli [39]. In addition, we obtained a high number of isolates exhibiting bla TEM-1 (n = 38) and, less frequently, bla OXA-1 (n = 7). Both bla genes were recently detected in E. coli isolates from harbor seals and grey seals (Halichoerus grypus) in Ireland [40], in which, in contrast to the findings of the present study, bla OXA-1 was more frequently represented than bla TEM-1 . All but one (ID 160) K. pneumoniae isolates carried a bla SHV gene. The broad-spectrum β-lactamase genes bla SHV-11 and bla SHV-33 were detected exclusively in ST466 K. pneumoniae and ST405 K. pneumoniae, respectively. Chaves et al. were the first to detect the novel SHV-type variant SHV-33 in K. pneumoniae, and they hypothesized that the ancestor of SHV-1 β-lactamase originated from the K. pneumoniae chromosomal DNA [41]. Furthermore, the pAmpC gene bla DHA-1 , which was detected in the two L. amnigena isolates (IDs 130 and 132), was previously also obtained in a K. pneumoniae isolate originating from a free-living mouflon in Austria [42].
Besides resistance to β-lactam antibiotics, resistance to other antibiotic classes was common among the investigated Enterobacterales. In total, 71% of the isolates exhibited a MDR phenotype and genotype. The study on E. coli in seals in Ireland determined a high proportion of MDR bacteria (66.6%) in the animals tested [40]. Already in the 1990s, AMR in Enterobacterales appeared widespread among Californian pinnipeds [26]. Wallace et al. [19] observed an increase of AMR in bacteria of stranded pinnipeds of the Northwest Atlantic over a period of 6 years (2004-2010). E. coli especially displayed a considerable increase in resistance to β-lactams, sulfonamides, and aminoglycosides, while Klebsiella spp. exhibited an increase in resistance to aminoglycosides and fluoroquinolones [19].
Biocide susceptibility testing revealed a high proportion of elevated MIC values for chlorhexidine (79%) and isopropanol (31%). Information regarding biocide susceptibility of Enterobacterales isolated from wildlife is scarce. A recent study on K. pneumoniae from Californian pinnipeds investigated environmental persistence, as well as disinfectant susceptibility, of these bacteria [43]. Commonly recommended concentrations of chlorine and hydrogen peroxide turned out to be ineffective in eradication of the biofilm formed by hypermucoviscous K. pneumoniae, whereas ethanol and chloramine-T effectively eradicated K. pneumoniae, irrespective of whether formed in biofilm or not [43].
Among the E. coli isolates, six different phylogenetic groups were determined, with the most common being phylogroup A (n = 8), described as being widespread among human commensal E. coli strains [46]. Contrary to our findings, investigations in Australia revealed the human-associated phylogroups B2 and D as the predominating E. coli phylotypes in pinnipeds [47][48][49].
cgMLST revealed that three E. coli clones (IDs 68, 183, and 230b) belonged to A-O NT :H9-ST167-CH11-0. These isolates were obtained from different individuals over a time-span of several months and exhibited in part highly diverse resistance pheno-and genotypes. This could indicate that these clones were already circulating for an extended period of time within the tested population of marine mammals and in their environment. Therefore, several resistance gene exchanges with other bacteria could have taken place. Related investigations have been reported regarding interspecies transmission of an E. coli ST410 clone between wildlife, humans, companion animals and the environment within several years in Berlin, Germany [50].
E. coli ST167 is considered a pandemic clone of significant public health concern due to its ESBL-producing strains [9]. Rising numbers of ST167 strains carrying the bla NDM carbapenemase gene were being reported and highlighted the threat emanating from this clone [51][52][53]. Guenther et al. were the first to report occurrence of ST167 ESBL-E. coli in wildlife [54]. Moreover, E. coli ST167 was detected in cloacal samples of silver gulls in Australia [55].
The MDR AmpC-producing E. coli isolate 197/1 belonged to ST648, a pandemic high risk clone combining MDR and high-level virulence [56], which has already been isolated from wild birds in Germany [50,57] and zoo animals in Israel [58].
E. coli isolate 209 was identified as ST38, a prevalent human clinical pathogen that has also been reported in wild birds of prey in Mongolia [59], rats in Africa [60], and silver gulls in Australia [61]. Furthermore, ST38 has had an entry in the Enterobase Escherichia/Shigella Database, whose corresponding isolate originated from a sea lion in Ecuador, South America (https://enterobase.warwick.ac.uk/species/index/ecoli, accessed on 21 March 2021) [62]. Both E. coli ST38 and ST648 are known as emerging ExPEC worldwide [63,64].
E. coli ST349, ST963, and ST1431 have also been detected in samples from wild birds [65], and an ST372 E. coli, isolated from a straw-coloured fruit bat in the Republic of Congo, has previously been detected and classified as ExPEC [66]. Additionally, E. coli ST372 has had an entry in the aforementioned database and the respective originated from a common bottlenose dolphin (Tursiops truncatus) in Mexico. There are no reports describing the isolation of E. coli ST484, ST4957, and ST5748 from wildlife to the authors knowledge.
An alarmingly high prevalence of different E. coli virulence genes was found in the isolates investigated. Most of these virulence genes could be put into context with ExPEC and/or UPEC, e.g., cnf1, fimH, fyuA, iroN, iss, iucD, iutA, kpsMII, papA, papC, sitA, and traT [67,68]. E. coli virulence genes associated with ExPEC were also detected in seals in Ireland [40] and in marine sediments from coasts of the Adriatic Sea [69].
K. pneumoniae is a nosocomial pathogen with increasing MDR rates and global dissemination of high risk clones [70]. Although several studies have previously focused on K. pneumoniae in farm and companion animals [71][72][73][74], information about this pathogen in wildlife is still scarce [42,55,[75][76][77][78]. Nevertheless, there are some reports about the isolation of K. pneumoniae from marine mammals along the U.S. Pacific Coast [19,[79][80][81]. Among the K. pneumoniae isolates in the present study, two distinct sequence types were obtained: ST405 (n = 3) and ST466 (n = 9). The three K. pneumoniae isolates (IDs 161, 173, and 175) belonging to ST405 displayed indistinguishable resistance pheno-and genotypes and exhibited 17 different resistance genes each. This might indicate the importance of this clone in the investigated environment. K. pneumoniae ST405 strains isolated from humans have been reported worldwide to harbor ESBL and carbapenemase genes [82][83][84][85]. However, reports about K. pneumoniae ST466 are scarce. One study revealed a context of K. pneumoniae ST466 harboring bla CTX-M-15 with causing neonatal sepsis [86].
Salmonella spp. have been previously isolated from marine mammals in California [87][88][89][90]. Within the scope of the present study, 18 Salmonella isolates belonging to six different serotypes (S. Havana, S. Reading, S. Albany, S. Newport, S. Saintpaul, S. Braenderup) were obtained, and one isolate exhibited AMR. Our findings regarding Salmonella are in general accordance with those of Stoddard et al. [82]. They investigated the prevalence and antimicrobial susceptibility of Salmonella spp. in northern elephant seals on the Californian coast. They further determined a higher prevalence and higher antimicrobial resistance rates of Salmonella in stranded seals than in young pinnipeds that had never entered the seawater on their natal beaches, presumably because stranded animals are more susceptible to infection by pathogens in the marine environment from terrestrial sources [87]. Furthermore, another study on Salmonella in Californian seals revealed that the amount of precipitation within the immediate pre-sampling period correlates positively with the presence of fecal Salmonella spp. [90].
As previously reported, marine mammals living in coastal habitats as predatory species with long lifespans can serve as sentinels of the health of the ocean and its inhabitants [24,91]. The high burden of AMR bacteria among marine mammals reported in this study might be the result of increasing anthropogenic pollution of the marine environment with antimicrobial agents, biocides, and already resistant bacteria. To get more detailed insights into the broad field of possible environmental risk factors and to identify notable sources of contamination, further comparative studies are needed.

Bacterial Isolates
Samples were collected from April 2019 to December 2020 upon admission of the stranded marine mammals to the Marine Mammal Care Center Los Angeles (MMCCLA) as part of diagnostic health evaluation. In all but four cases, samples were collected prior to any antimicrobial treatment. Overall, 145 rectal swabs, 130 oral swabs, four wound swabs, two swabs of a blow hole, and two swabs of the nictitating membrane were sampled from 148 individual animals, including 122 California sea lions (Zalophus californianus), 15 northern elephant seals (Mirounga angustirostris), six harbor seals (Phoca vitulina), two northern fur seals (Callorhinus ursinus), one Guadalupe fur seal (Arctocephalus townsendi), one long-beaked common dolphin (Delphinus capensis), and one pygmy sperm whale (Kogia breviceps).
Each sample was preincubated at 37 • C overnight in buffered peptone water (BPW) (Merck, Germany). For selective isolation of β-lactamase-producing Enterobacterales, 200 µL of the incubated sample were cultured at 37 • C overnight in BPW supplemented with cefotaxime (1 mg/L) (BPWCTX) and then cultivated at 37 • C overnight on MacConkey agar (Oxoid, Basingstoke, United Kingdom) supplemented with cefotaxime (1 mg/L) (Mac-CTX). In parallel, another 200 µL of the preincubated sample were cultured at 42 • C overnight in Rappaport-Vassiliadis-broth (RV) (Oxoid, Basingstoke, UK) and in Selenitbroth (Merck, Darmstadt, Germany) and then cultivated at 37 • C on BD™ XLD Agar (Xylose-Lysine-Desoxycholate Agar) (BD, Heidelberg, Germany) for selective isolation of Salmonella sp. After incubation on MacCTX, one colony representing a distinct colony morphotype was regrown on BD MacConkey II Agar (MC) at 37 • C overnight. After incubation on XLD, the respective colonies of Salmonella isolates showed the typical colony appearance of Salmonella and were selected for further characterization. Colonies selected from MacCTX and XLD were identified to the species level by matrix-assisted laser desorption/ionization-time-of-flight (MALDI-TOF) mass spectrometry (Bruker Daltonik, Heidelberg, Germany). Isolates, which grew on MacCTX and were confirmed as belonging to Enterobacterales, were examined for ESBL production by combination disk tests using cefotaxime and ceftazidime with and without clavulanic acid (Becton Dickinson, Heidelberg, Germany) according to the Clinical and Laboratory Standards Institute (CLSI) standards [92]. Furthermore, cefoxitin (30 µg) (BD, Heidelberg, Germany) was added to this test to detect AmpC phenotypes.

Virulence Genes
Detection and analysis of virulence genes was performed using primer and probe sets derived from microarray-based methods, as described previously [119,120], and used in custom made microarrays from INTER-ARRAY (INTER-ARRAY by fzmb GmbH, Bad Langensalza, Germany) according to manufacturers instructions. The list of virulenceassociated genes is available at INTER-ARRAY website (https://www.inter-array.com/ #microarrays, accessed on 30 April 2021) upon request.

Characterization of Salmonella Isolates
The antigenic formula of all Salmonella isolates was determined by the National Reference Centre for Salmonella Austria using standard agglutination methods (ISO/TR 6579-3, 2014), and the serotype name was assigned according to the White-Kauffmann-Le Minor scheme [135].

Conclusions
In conclusion, the present study contributes to the growing evidence that β-lactamaseproducing and multidrug-resistant Enterobacterales, as well as Salmonella, are currently part of the microbiome of wild animals. The presence of pandemic clones in samples originating from marine mammals demonstrates the complexity in the dissemination of antimicrobial drug resistance and highlights the public health threats. Furthermore, the resistance and virulence genes frequently encoded on mobile genetic elements can easily be transferred horizontally and also between different species in such a habitat and ecosystem, emphasizing the need of a One Health approach to tackling the global AMR crisis. Institutional Review Board Statement: Ethical review and approval were waived for this study, because all clinical samples were obtained within the scope of routine diagnostic health assessments and only subsequently evaluated within the scope of this study and, therefore, not subject to reporting obligations of the Ethics and Animal Welfare Commission of the University of Veterinary Medicine in Vienna.

Informed Consent Statement: Not applicable.
Data Availability Statement: All data is contained within the article or supplementary material.