Vegetables and Fruit as a Reservoir of β-Lactam and Colistin-Resistant Gram-Negative Bacteria: A Review.

Antibacterial resistance is one of the 2019 World Health Organization's top ten threats to public health worldwide. Hence, the emergence of β-lactam and colistin resistance among Gram-negative bacteria has become a serious concern. The reservoirs for such bacteria are increasing not only in hospital settings but in several other sources, including vegetables and fruit. In recent years, fresh produce gained important attention due to its consumption in healthy diets combined with a low energy density. However, since fresh produce is often consumed raw, it may also be a source of foodborne disease and a reservoir for antibiotic resistant Gram-negative bacteria including those producing extended-spectrum β-lactamase, cephalosporinase and carbapenemase enzymes, as well as those harboring the plasmid-mediated colistin resistance (mcr) gene. This review aims to provide an overview of the currently available scientific literature on the presence of extended-spectrum β-lactamases, cephalosporinase, carbapenemase and mcr genes in Gram-negative bacteria in vegetables and fruit with a focus on the possible contamination pathways in fresh produce.

1. Introduction

Fresh produce is considered a good source of minerals, vitamins, phytonutrients and dietary fiber. Accordingly, there is a consensus that a diet rich in vegetables and fruit may decrease the risk of heart diseases and protect against some types of cancer [1]. In 2003, the Food and Agriculture Organization (FAO) of the United Nations and the World Health Organization (WHO) started an initiative worldwide to promote fruit and vegetable intake for health, with a recommended minimum consumption of 400 g of vegetables and fruit per day [2]. Following this recommendation, the intake of fresh produce as ingredients in healthy diets has been increasing and has gained popularity globally [3]. Consequently, the consumption of contaminated fresh produce, such as vegetables and fruit eaten raw, has been associated with an increasing number of outbreaks of foodborne disease [1]. In addition, fresh produce represents a route of human exposure to antibiotic-resistant bacteria and has often served as a reservoir of antibiotic resistance genes, representing a major public health threat [3,4].

In this context, one of the main public health preoccupations worldwide is the emergence of Gram-negative bacteria displaying resistance to oxyimino-cephalosporins (3GCs), carbapenem and colistin [3]. β-lactams, essentially extended-spectrum cephalosporins and carbapenems, are the main therapeutic choices to treat infections caused by resistant Gram-negative bacteria [5,6,7]. However, resistance to these antibiotic drugs has been increasing in recent years mostly through β-lactamase production. Various β-lactamases have been identified worldwide, including penicillinases, extended-spectrum β-lactamases (ESBLs), cephalosporinases (AmpC), and carbapenemases [7]. Given these circumstances, the approved alternative is colistin, but its re-use in clinical practice has led to the appearance of colistin-resistant bacteria, particularly through horizontal transfer (mcr) [8].

The transfer of these multidrug-resistant Gram-negative bacteria to fresh produce may occur during production via animal manure, through the use of contaminated irrigation water, or be linked to humans during the post-harvest stage, as well as during transport, conservation and processing by handlers [9]. The ingestion of antibiotic-resistant bacteria poses a potential public health concern since they are able to colonize the gut and exchange resistance genes with intestinal bacteria during their passage through the intestines which facilitates their further dissemination in the environment [3]. Extended-spectrum β-lactamase, cephalosporinase and carbapenemase producers as well as mcr gene-producing Gram-negative bacteria isolated from fresh vegetables and fruit have been reported in several countries around the world [3,4,10,11].

Thus, the aim of this review is to highlight the current situation of the worldwide dissemination of ESBL, cephalosporinase, carbapenemase and mcr gene-producing Gram-negative bacteria from fresh vegetables and fruit, their genetic characteristics, and possible contamination pathways.

2. β-Lactam Resistance and Gram-Negative Bacteria

β-lactam resistance in Gram-negative bacteria can be attributed to two main mechanisms, these include the acquisition of β-lactamase genes, as well as the modification of the target (penicillin-binding proteins) [12]. β-lactamase enzymes have played an important clinical role and have served as the principal resistance mechanism detected for β-lactam drugs [13,14].

The first enzyme detected presenting β-lactamase activity originated from Bacillus coli in 1940, currently supposed to be the class C, AmpC chromosomal cephalosporinase from Escherichia coli [14]. Given this, various extended-spectrum cephalosporins were introduced in the 1980s, which were stable against penicillinase hydrolysis, such as TEM-1 (TEMoniera) and SHV-1 (sulfhydryl variable). A few years later, Enterobacteriaceae species developed several derivatives of TEM-1, TEM-2 and SHV-1; these variants extended their hydrolysis spectrum to include oxyimino-cephalosporins, hence the term ‘extended-spectrum’ β-lactamases (ESBL) [14,15]. Afterwards, a novel variant of the ESBL family named Cefotaximase-Munchen (CTX-M) was described, which became the predominant ESBL in enterobacterial species worldwide [14], as well as the family of Guyana extended-spectrum β-lactamases (GES) reported as ESBL variants in 2000 [16]. The β-lactamases belonging to Ambler class C, called cephalosporinases, are derived from the ampC gene in the chromosome of various Enterobacteriaceae species [17,18]. In the early 1990s, plasmid-encoded AmpC cephalosporinases were described in species lacking an inducible AmpC enzyme. Afterwards, plasmid-mediated AmpC, such as Dharhan hospital (DHA), cephamycinase (CMY), cefoxitinase (FOX), moxalactamase (MOX) and Ambler Class C (ACC), were reported worldwide [19].

In this worrisome situation, carbapenems were introduced to clinics in the late 1980s and showed significant activity in the treatment of infections caused by AmpC and ESBL-producing Gram-negative bacteria [16,20]. The first carbapenemase reported in Enterobacteriaceae was the Serratia marcescens enzyme (SME-1) in London in 1982. Since then, various carbapenemase enzymes belonging to the Ambler class A β-lactamases have been reported, including imipenemase (IMI-1) and non-metallocarbapanemase class A (NmcA); however, the K. pneumoniae-carbapenemase (KPC) type was the most commonly found [5,16,21]. On the other hand, the first MBL variant was discovered in Bacillus cereus in 1966 and was called the BCII enzyme. Until 1989, only four MBL enzymes had been identified and were all chromosomally encoded, therefore they were deemed clinically negligible. Afterwards, various plasmid-encoding class B carbapenemases were described, such as Imipenem-resistant Pseudomonas-type carbapenemases (IMP),Verona integron-encoded MBL (VIM),and recently, New Delhi MBL (NDM) [22,23].In class D β-lactamases, several variants with relatively weak carbapenemase activity have also been reported as carbapenemase enzymes, including OXA-48, OXA-58, OXA-24/40 and OXA-23 [22]. In Enterobacteriaceae, class D carbapenemases are mainly represented by the OXA-48-like enzymes [24].

3. Colistin Resistance in Gram-Negative Bacteria

Colistin is a cationic polypeptide antibiotic belonging to the polymyxin family [25]. It was described initially in 1947 in Paenibacillus polymyxa, and it is commonly used in human and veterinary medicines, plant cultivation and animal husbandry [25,26,27]. Although in the 1970s its use was discontinued due to its neuro- and renal toxicity, it was reintroduced in the mid-2000s as a last line therapeutic option for the treatment of extensively drug-resistant (XDR) Gram-negative infections, such as those caused by carbapenem-resistant GNB [26,28].

The initial target site of colistin is lipopolysaccharide (LPS), more exactly lipid A, located in the outer membrane, which plays a major role in cell permeability. The electrostatic interaction between the cationic region of colistin, which is from the diamino-butyric acid (Dab) residues, and the negatively charged phosphate groups of lipid A, replace the magnesium and calcium ions previously united with the phosphate group. This destabilizes the lipid A and increases the permeability of the outer membrane, leading to the entry of colistin by a self-promoted uptake mechanism and eventual bacterial death [26,29]. Another antibacterial mechanism is the inhibition of a crucial respiratory enzyme, the type II NADH-quinone oxidoreductase (NDH-2) in the bacterial cell membrane [29]. The increased use of colistin has led to the emergence of colistin-resistant strains worldwide [25]. Colistin resistance is mainly achieved by modification of LPS, and consequently the reduced or absent affinity for colistin. This mechanism, although universal in Gram-negative bacteria, may differ between species. The lipid A of LPS undergoes changes, essentially due to the addition of positively charged residues such as phosphoethanolamine (PEtn) and/or 4-amino-4-deoxy-L-arabinose (L-Ara4N). These molecules decrease the overall negative charge of LPS, leading to a smaller electrostatic interaction with the positive charges of colistin that inhibits cell lysis [26,30].

Previously, the genes responsible for most of these additions were thought to be due to chromosomal mutations in genes of a two-component regulatory system, such as pmrAB, PhoPQ, and mgrB, which are not transferable [30]. In late 2015, Liu et al. described the mobilized colistin resistance (mcr-1) gene in an E. coli isolate recovered from livestock in China [8,31]. MCR-1 confers resistance by modifying the colistin target through the action of phosphoethanolamine transferase, which ensures the transfer of phosphoethanolamine (PEA) onto the glucosamine saccharide of lipid A, contributing as in chromosomal resistance to reduce the net negative charge of lipid A and consequently, colistin binding [32]. After the discovery of the mcr-1 gene, nine other mcr gene types (mcr-2 to mcr-10) were identified. The second mobile colistin resistance gene, mcr-2, was found initially in E. coli strains isolated from pigs and calves in Belgium [33]. The gene mcr-3 was identified in E. coli from pigs in China [34], and mcr-4 was reported in Salmonella enterica serovar Typhimurium strains isolated from pigs in Italy [35]. In 2017, anovel transposon-associated phosphoethanolamine transferase gene (mcr-5) was described in d-tartrate-fermenting Salmonella enterica subsp. enterica serovar paratyphi B in Germany [36]. In 2018, further variants were described; mcr-6 was identified in Moraxella spp. isolated from pigs in Great Britain [37], while mcr-7 and mcr-8 were described in K. pneumoniae strains isolated from animals (chickens and pigs) in China [38,39]. In 2019, a novel variant mcr-9 was reported in a Salmonella enterica serovar Typhimurium strain isolated from a human in Washington State in 2010 [40] and more recently, Wang et al. reported the detection of an mcr-10 variant in an Enterobacter roggenkampii clinical strain in China [41].

4. Literature Search Strategy and Data Collection

The dissemination of extended-spectrum β-lactamase, cephalosporinase, carbapenemase and MCR-producing Gram-negative bacteria in fresh produce is a major public health threat, since they are a very suitable pathway for the spread of antibiotic-resistant bacteria from farm to fork. Until June 2021, thirty-three molecular studies have revealed the isolation of Gram-negative bacteria producing β-lactamase and mcr genes on fresh vegetables and fruit. They have been used and are accessible through the PubMed database using the following keywords: ‘‘ESBL’’, ‘‘AmpC’’, ‘‘KPC’’, ‘‘VIM’’, ‘‘NDM’’, ‘‘IMP’’, ‘‘OXA-48”, ‘‘mcr’’, ‘‘carbapenem resistance’’, ‘‘fresh vegetables’’, ‘‘vegetables’’ and ‘‘fruit’’.

5. Vegetable and Fruit Isolates with ESBL and Cephalosporinase Genes

A total of nineteen molecular studies reporting the isolation of ESBL-producing Gram-negative bacteria and AmpC genes from vegetables and fruit have been described (Figure 1, Table 1). The first report of ESBL-producing GNB isolates from vegetables and fruit was reported in 2014 in The Netherlands. These bacteria were reported on six vegetable types that are consumed raw (bunched carrots, blanched celery, endive, chicory, iceberg lettuce and radish), and from iceberg lettuce farms. In that study, the blaFONA-5 gene was detected among Serratia fonticola isolates on iceberg lettuce from a farm. In addition, 35 Rahnella aquatilis strains harboring the blaRAHN gene were identified. Of the 35 isolates, 34 strains were producing the blaRAHN-1, and only one R. aquatilis strain carried the blaRAHN-2 gene [42]. After this publication, this level of resistance has been reported in Europe, Africa, Asia and America. Like isolates from humans, animals and the environment, the CTX-M family is the most prevalent type of ESBL-producing Enterobacteriaceae found in vegetables. Similarly, in an Italian study carried out on fresh vegetables, the authors refer to the detection of different ESBL enzymes, including CTX-M-15, CTX-M-1, SHV-12 and RAHN-1 in twenty isolates (the blaCTX-M-15 gene in C. freundii, E. coli and Pantoea agglomerans, the blaCTX-M-1 gene in Enterobacter cloacae, the blaSHV-12 in E. coli and blaRAHN-1 in R. aquatilis). Whereas only four isolates displayed AmpC production, among the four strains obtained, two Hafnia alvei isolates carried a blaACC gene and two E. cloacae harbored a blaDHA-1 gene [43]. A study in The Netherlands investigated the prevalence of third-generation cephalosporin (3GC) resistant Gram-negative bacteria on fresh vegetables. A total of 27 Serratia spp. isolates with an ESBL phenotype harboring a blaFONA variant were obtained, including blaFONA-1 (18.5%), blaFONA-2 (37.0%), blaFONA-3 (7.4%), blaFONA-4 (7.4%), blaFONA-5 (18.5%) and blaFONA-6 (11.1%). The blaSHV-12 gene was detected in one E. coli and two Enterobacter spp. strains; however, one R. aquatilis strain harbored the blaRAHN-1 gene [3]. In Switzerland, two studies reported the detection of blaESBL genes on vegetable samples. In the first study, the authors evaluated the presence of ESBL-producing Enterobacteriaceae in 68 vegetables imported from the Dominican Republic, India, Thailand and Vietnam via the national airport in Zürich, and 101 samples were purchased in the city of Zürich. In total, 60 ESBL producers were retrieved, including blaCTX-M- and blaSHV-producing E. coli (blaCTX-M-15, blaCTX-M-55, blaCTX-M-14, blaCTX-M-65, blaCTX-M-1 and blaSHV-12) and K. pneumoniae strains (blaCTX-M-15, blaCTX-M-14, blaCTX-M-3, blaCTX-M-27, blaCTX-M-63, blaSHV-2, blaSHV-2a, and blaSHV-12). Moreover, blaCTX-M-15 and blaSHV-2 genes were identified in E. cloacae, E. aerogenes and C. sakazakii, respectively [10]. The second study reported the detection of CTX-M group 2, CTX-M-15 and FONA-2 in Kluyvera ascorbata, E. cloacae and S. fonticola isolates from diced tomato, chopped chives and spinach, respectively [44]. A study from Germany described the isolation of seven ESBL-producing E. coli isolates collected by food safety inspectors during 2011–2013 from markets, producers and supermarkets. Of the seven isolates, two strains were positive for blaCTX-M-14 and two other isolates harbored blaCTX-M-15 genes. However, three remaining strains were positive for blaCTX-M-65, blaCTX-M-125 and blaCTX-M-2 genes, respectively [45]. In addition, the blaTEM, blaSHV, blaCTX-M and blaDHA genes were also reported in Romania in different Enterobacteriaceae species (S. marcescens, E. cloacae, E. coli, Klebsiella oxytoca and Proteus vulgaris) [46].

Figure 1

Worldwide distribution of extended-spectrum β-lactamase, cephalosporinase, carbapenemase and mcr-producing GNB on fresh vegetables and fruit.

Table 1

ESBL and cephalosporinase genes reported in Gram-negative bacteria isolates from vegetables and fruit worldwide.

Vegetable Type	ESBL/AmpC Gene	Isolation Period	Species	Isolates Number	Country	Other Antibiotic Resistance Genes	Sequence Type	References
Lettuce	bla FONA-5	2011	Serratia fonticola	1	The Netherlands	ND	ND	[42]
	bla RAHN-2		Rahnella aquatilis	1
	bla CTX-M-15	2013–2014	Klebsiella pneumoniae	1	Algeria	aph(3′)-Ia, aadA2, strB, strA, qnrS1, oqxB, oqxA, fosA, mph(A), catA2, sul1, sul2, tet(A), dfrA12	ST219	[48]
	bla DHA-1 bla SHV-101		K. pneumoniae	1		blaOXA-1, aac(6′)Ib-cr, aph(3′)-Ia, aac(6′)Ib-cr, qnrB4, oqxB, oqxA, fosA, mph(A) catB3, ARR-3, sul1	ST882
	blaSHV-28, blaCTX-M-15,		K. pneumoniae	1		blaOXA-1, aac(6′)Ib-cr, aac(3)-Iia, aac(6′)Ib-cr, qnrB66, oqxB, oqxA, fosA, catB3, dfrA14.	ST14
	bla CTX-M-15	2015	Escherichia coli	1	Ecuador	dfrA1, aadA5	ST44	[54]
	bla CTX-M-15			1		None	ST44
	bla CTX-M-14	2017–2018	S. fonticola	1	South Africa	ND	ND	[9]
	bla SHV-154		S. marcescens	1		ND	ND
	bla CTX-M-15	2018	E. coli	1	South Korea	ND	ST2509	[51]
	blaSHV, blaTEM	2019	Proteus vulgaris	1	Romania	ND	ND	[46]
	bla CTX-M-15	ND	K. pneumoniae	1	Brazil	blaOXA-1, blaSHV-110, aac(3)IIa, aac(6′)-Ib-cr, opxAB, drfA14, catA1, tet(A), fosA, opxB	ST198	[56]
Butterhead lettuce	blaFONA-1 (1–6)	2012–2013	S. fonticola	ND	The Netherlands	ND	ND	[3]
Iceberg lettuce	bla RAHN-1	2011	R. aquatilis	ND	The Netherlands	ND		[42]
	blaSHV,blaTEM	2011–2012	K. pneumoniae	2	United States	ND	ND	[52]
	bla CTX-M-1		S. marcescens	1		ND	ND
	blaFONA-1 (1–6)	2012–2013	S. fonticola	ND	The Netherlands	ND	ND	[3]
Tomato	blaCTX-M, blaSHV, blaTEM	2011–2014	E. coli	1	China	ND	ND	[7]
	blaSHV-28, blaCTX-M-15	2013–2014	K. pneumoniae	1	Algeria	aac(3)-Iia, qnrB66, oqxB, oqxA, fosA	ST14	[48]
	blaSHV-28, blaCTX-M-15,			1		blaOXA-1, aac(6′)Ib-cr, aac(3)-Iia, aac(6′)Ib-cr, qnrB66, oqxB, oqxA, fosA, catB3, dfrA14	ST14
	bla CMY-2	2017–2018	Citrobacter freundii	1	South Africa	ND	ND	[9]
	bla CTX-M-14			1		ND	ND
	bla CTX-M-55		E. coli	1		ND	ND
	bla CTX-M-14			1		ND	ND
	bla CTX-M-14			1		blaSHV-1, blaTEM-215	ND
	bla SHV-18		E. asburiae	1		ND	ND
	bla MIR-14			1		bla SHV-26	ND
	bla ACT-29			1		ND	ND
	blaCTX-M-27, blaCTX-M-15		E. cloacae	1		bla SHV-26	ND
	bla MIR-20			1		ND	ND
	blaTEM-3, blaACT-2, blaSHV-18			1		blaTEM-1, blaSHV-11	ND
	blaSHV-18, blaTEM-3		E. cowanii	1		ND	ND
	bla CTX-M-15		K. pneumoniae	1		ND	ND
	bla ACT-10		K. oxytoca	1		ND	ND
	bla CTX-M-55		Proteus mirabilis	1		bla TEM-215	ND
	blaACT-10, blaDHA-18, blaCMY-49		Pseudomonas penneri	1		ND	ND
	bla SHV-18		R. aquatilis	1		bla TEM-215	ND
	bla MIR-16		R. aquatilis	1		ND	ND
	blaSHV-18, blaMIR-16	2017–2018	E. asburiae	1	South Africa	blaTEM-1, blaOXA-1	ND	[9]
Diced tomato	bla CTX-MGroup2	2014	Kluyvera ascorbata	1	Switzerland	ND	ND	[44]
Spinach	bla FONA-2	2014	S. fonticola	1	Switzerland	ND	ND	[44]
	blaCTX-M-group1, blaTEM	2017	E. asburiae	1	South Africa	ND	ND	[47]
	blaCTX-M-group1, blaTEM, blaSHV, blaOXA		E. coli	2		ND	ND
	blaCTX-M-group1, blaTEM, blaSHV, blaOXA		K. pneumoniae	3		ND	ND
	blaCTX-M-group1, blaTEM		E. asburiae	1		ND	ND
	blaCTX-M-group1, blaTEM, blaSHV, blaOXA		E. coli	2		ND	ND
	blaCTX-M-group1, blaTEM, blaSHV, blaOXA		K. pneumoniae	3		ND	ND
	bla CTX-M-group1		R. aquatilis	1		ND	ND
	blaCTX-M-group1, blaTEM, blaSHV, blaOXA		R. aquatilis	1		ND	ND
	blaCTX-M-group1, blaTEM, blaSHV		R. aquatilis	2		ND	ND
	bla CIT		S. fonticola	3		ND	ND
	blaTEM, blaSHV			1		ND	ND
	blaCTX-M-group1, blaTEM, blaSHV, blaOXA			2		ND	ND
	blaCTX-M-group1, blaTEM, blaSHV, blaOXA, blaCIT			1		ND	ND
	blaCTX-M-group1, blaSHV			1		ND	ND
	blaCTX-M-group1, blaTEM, blaSHV, blaCIT			2		ND	ND
	bla CTX-M-27	2017–2018	E. coli	2	South Africa	ND	ND	[9]
	bla MIR-20			1		ND	ND
	blaSHV-18, blaCTX-M-15, blaTEM-3			1		ND	ND
	blaCTX-M-14, blaTEM-3			1		ND	ND
	bla CTX-M-14			1		ND	ND
	bla CTX-M-15			1		ND
	bla CTX-M-55			1		ND	ND
	blaCTX-M-14, blaACT-58			1		ND	ND
	bla CTX-M-14			2		ND	ND
	bla CTX-M-14			2		bla TEM-215	ND
	bla ACT-58		E. asburiae	1		bla TEM-215	ND
	bla CMY-87		E. ludwigii	1		ND	ND
	blaCTX-M-27, blaEC-30		R. aquatilis	1		ND	ND
	bla CTX-M-15			1		bla SHV-11	ND
	blaCTX-M-15, blaSHV-28		S. fonticola	1		ND	ND
	blaCTX-M-14, blaSHV-28			1		ND	ND
	bla MIR-16			1		blaTEM-1, blaOXA-1	ND
	bla CTXM-15			1		bla TEM-215	ND
	blaSHV, blaTEM	2019	S. marcescens	1	Romania	ND	ND	[46]
	bla CTXM		E. cloacae	1		ND	ND
	bla CTXM-15	ND	E. cloacae	1	Brazil	blaOXA-1, blaTEM-1B, blaACT-7, aac(3)-IIa, aac(6′)Ib-cr, ant(3′’)Ia, strA, strB, qnrB, sul2, tet(A), fosA.	ST927	[56]
	bla CTXM-15	ND	E. coli	1		blaTEM-1B, aac(3)IId, aadA5, strA, strB, tet(A)	ST14012
Chopped Spinach	blaCTXM14, blaSHV-142	2017	K. pneumoniae	1	Canada	ND	ST261	[55]
	bla CTXM-27		E. cloacae	1		qnrB2, qnrS1, aac(6′)Ib cr	ND
	bla CTXM-27		E. aerogenes	1		aac(6′) Ib cr	ND
Ceylon spinach	bla CTXM-14	2014	K. pneumoniae	1	Switzerland	ND	ST37	[10]
Water spinach	bla CTXM-15		K. pneumoniae	1		ND	ST16
Cucumber	blaCTX-M, blaSHV	2011–2014	E. coli	2	China	ND	ND	[7]
	bla CTXM-15	2014	E. cloacae	1	Switzerland	ND	ND	[10]
	bla CTXM-15		E. coli	1		ND	ST410
	bla TEM-116	2015–2016	P. mosselii	1	Japan	ND	ND	[49]
	bla MIR-20	2017–2018	E. cloacae	1	South Africa	ND	ND	[9]
	bla SHV-18		R. aquatilis	1		bla OXA-1	ND
	blaCTXM, blaTEM	2019	E. coli	1	Romania	ND	ND	[46]
	blaDHA		E. cloacae	1		ND
Bitter cucumber	bla CTXM-15	2014	E. coli	1	Switzerland	ND	ST131	[10]
Coriander	bla CTXM-55	2011–2014	E. coli	2	China	ND	ST48, ST4680	[7]
	blaCTX-M, blaOXA		Citrobacter freundii	1		ND	ND
	bla CTX-M-55	2018	E. coli	1	Malaysia	blaTEM-1B, aph(3 0)-Ia, aph(300)-Ib, aph(6)-Id, mdf(A), floR, ARR-2, sul2, tet(A), dfrA14	ST155	[50]
	bla CTX-M-65		E. coli	1		aac(3)-IV, aadA5, aph(4)-Ia, oqxA, oqxB, mdf(A), floR, sul1, sul2, tet(A), dfrA17	ST479
Parsley	blaSHV-28, blaCTX-M-15, blaOXA-1	2013–2014	K. pneumoniae	1	Algeria	aac(6′)Ib-cr, aac(3)-Iia, aac(6′)Ib-cr, qnrB66, oqxB, oqxA, fosA, catB3, dfrA14	ST14	[48]
	blaCTX-M-15, blaOXA-1			1		aac(6′)Ib-cr, aac(3)-IIa, strB, strA, aac(6′)Ib-cr,oqxB, oqxA, fosA, catB3, sul2, tet(A), dfrA14	ST45
	bla SHV	2019	K. oxytoca	1	Romania	ND	ND	[46]
Water parsley	bla CTX-M-55	2018	E. coli	1	South Korea	ND	ND	[51]
	blaCTX-M-15, blaTEM-1			1		ND	ST101
	blaCTX-M-14, blaTEM-1			1		ND	ST354
	bla CTX-M-14			1		ND	ST38
Parsley/cilantro	bla CTX-M-15	2015	E. coli	1	Ecuador	None	ST410	[54]
				1		dfrA1, aadA5	ST44
Soy sprouts	bla CTX-M-65	2011–2013	E. coli	1	Germany	floR, aac(6′)-Ib3, sul2, tet(A), fosA3	ST10	[45]
	bla CTX-M-125			1		aph(3′)-II, tet(A), fosA3	ST542
	bla CTX-M-14			1		catA1, floR, aac(6′)Ib-cr, aph(3′)-Ia, aadA5, sul1, sul2, tet(A), dfrA17, fosA3	ST527
	bla CTXM-14	2014	K. pneumoniae	1	Switzerland	ND	ST208	[10]
Sprouts-mixture	bla CTX-M-15	2011–2013	E. coli	1	Germany	blaTEM-1, qnrS1, strA, strB, sul2, tet(A), dfrA14	ST847	[45]
Alfalfa	bla CTX-M-15	2015	E. coli	1	Ecuador	dfrA1, aadA5	ST410	[54]
	bla CTX-M-15			1		None	ST44
	bla CTX-M-15			1		None	ST44
Alfalfa sprouts	bla CTX-M-15	2011–2013	E. coli	1	Germany	blaTEM-1, qnrS1, strA, strB, sul2, tet(A), dfrA14	ST410	[45]
Greenbeans	bla CTX-M-14	2017–2018	E. coli	2	South Africa	ND	ND	[9]
	blaCTX-M-14, blaCMY-2		S. fonticola	1		bla TEM-215	ND
	blaCTX-M-14, blaCMY-161	2017–2018	S. fonticola	1		bla TEM-215	ND
Curry leaves	bla CTXM-15	2014	K. pneumoniae	1	Switzerland	ND	ST307	[10]
	bla CTXM-14		E. coli	1		ND	ST38
	bla CTXM-15		K. pneumoniae	1		ND	ST1742
	bla SHV-12		E. coli	1		ND	ST1656
	bla CTXM-15		K. pneumoniae	4		ND	ST1739, ST1741, ST1881, ST1740
	bla CTXM-1		E. coli	1		ND	ST1555
	bla CTXM-15			1		ND	ST4681, ST152
	bla CTXM-14			1		ND	ST4679
	bla CTXM-55			1		ND	ST10
Mint	blaCTX-M-15, blaSHV-28	2013–2014	K. pneumoniae	1	Algeria	blaOXA-1, aac(6′)Ib-cr, aac(3)-Iia, aac(6′)Ib-cr, qnrB66, oqxB, oqxA, fosA, catB3, dfrA14	ST14	[48]
	blaCTX-M-15, blaSHV-28			1		blaOXA-1, aac(6′)Ib-cr, aac(3)-Iia, aac(6′)Ib-cr, qnrB66, oqxB, oqxA, fosA, catB3, dfrA14	ST14
Chili	bla CTXM-15	2014	E. coli	1	Switzerland	ND	ST405	[10]
Green chili	bla CTXM-15		E. cloacae	1		ND	ND	[10]
	bla CTXM-15		K. pneumoniae	2		ND	ST1740, ST37
	bla CTXM-27			1		ND	ST458
Small chili	bla CTXM-65		E. coli	1		ND	ST167
Chili pepper	blaCTX-M-15, blaSHV-28	2018	K. pneumoniae	1	Malaysia	blaTEM-1B, blaOXA-1, aac(3)-IIa, aac(6 0)-Ib-cr, aph(300)-Ib, aph(6)-Id, aac(6 0)-Ib-cr, oqxA, oqxB, qnrB1, fosA, catB3, sul2, tet(A), dfrA14	ST307	[50]
	blaDHA-1, blaSHV-28		K. pneumoniae	1		oqxA, oqxB, qnrS1, fosA, sul1, tet(A), dfrA1	ST101
Hyacinth bean seeds	bla CTXM-15	2017	E. coli	1	Canada	ND	ST189	[55]
				1		bla TEM-1	ST226
Ginseng	bla TEM-116	2015–2016	Pseudomonas paralactis	1	Japan	ND	ND	[49]
	bla TEM-116			1		ND	ND
	bla TEM-116		P. arsenicoxydans	1		ND	ND
Beets	bla CTX-M-15	2013–2014	K. pneumoniae	1	Algeria	aph(3′)-Ia, aadA2, strB, strA, nrS1, oqxB, oqxA, fosA, mph(A), catA2, sul1, sul2, tet(A), dfrA12	ST219	[48]
Carrot	blaCTX-M-15, blaOXA-1	2013–2014	K. pneumoniae	1		blaTEM-1B, aac(6′)Ib-cr, aac(3)-IIa, strB, strA, ac(6′)Ib-cr, qnrB66, oqxB, oqxA, fosA, catB3, sul2, dfrA14.	ST45	[48]
	bla RAHN-1	2011	R. aquatilis	ND	The Netherlands	ND	ND	[42]
Bunched carrot	blaFONA(1–6)	2012–2013	S. fonticola	ND	The Netherlands	ND	ND	[3]
Arugula	bla RAHN-1	2015–2016	R. aquatilis	4	Italy	ND	ND	[43]
	bla CTX-M-15		C. freundii	4		ND	ND
	bla ACC		Hafnia alvei	2		ND	ND
	bla CTXM-15 bla SHV-106	ND	K. pneumoniae	1	Brazil	blaOXA-1, blaTEM-1B, aac(6)Ib-cr, strA, strB, qnrB1, opxAB, gyrA, parC, tet(A), fosA.	ST2739	[56]
Egg plant	bla CTXM-15	2014	K. pneumoniae	1	Switzerland	ND	ST45	[10]
	bla CTXM-15			1		ND	ST307
	bla CTXM-15		E. cloacae	1		ND	ND
Chinese chive	bla SHV-12	2015–2016	P. parafulva	1	Japan	ND	ND	[49]
Chopped chives	bla CTX-M-15	2014	E. cloacae	1	Switzerland	ND	ND	[44]
Onion	blaFONA-1 (1–6)	2012–2013	S. fonticola	ND	The Netherlands	ND	ND	[3]
	bla TEM-116	2015–2016	Pseudomonas beteli	1	Japan	ND	ND	[49]
Broccoli	bla TEM-116	2015–2016	P. hunanensis	1	Japan	ND	ND	[49]
Cabbage	bla TEM-116	2015–2016	P. hunanensis	1	Japan	ND	ND	[49]
	blaCTX-M, blaSHV	2019	E. cloacae	1	Romania	ND	ND	[46]
	bla CTXM-65	2018	E. coli	1	South Korea	ND	2847	[51]
	bla CTXM-15	ND	E. coli	1	Brazil	blaOXA-1, blaTEM-1B, aac(3)IIa, aac(6′)Ib-cr, aadA5, gyrA, parC, sul1.	ST648	[56]
	bla CTXM-15	ND	E. coli	1		blaOXA-1, aac(6′)Ib-cr, aadA5, aac(6′)Ib-cr, sul1, drfA17, tet(B), mph(A)	ST38
Cut cabbage	bla SHV-12	2015–2016	P. hunanensis	1	Japan	ND	ND	[49]
Bean sprout	bla SHV-12		P. putida	1		ND	ND
Yard long beans	bla CTXM-55	2015–2016	E. cloacae	1	Japan	ND	ND	[49]
	bla SHV-12	2014	Cronobacter sakazakii	1	Switzerland	ND	ST3696	[10]
	bla CTXM-14		E. coli	1		ND	ND
Holy Basil	bla CTXM-15	2014	K. pneumoniae	1	Switzerland	ND	ST36
	bla CTXM-65		E. coli	1		ND	ST58
Okra (marrow)	bla CTXM-14		E. coli	1		ND	ST38
	bla CTXM-15		E. coli	2		ND	ST155, ST443
Okra	bla CTXM-15		K. pneumoniae	2		ND	ST997, ST244
	bla CTXM-15		E. aerogenes	1		ND	ND
	bla CTXM-15		E. cloacae	2		ND	ND
	bla CTXM-15		E. coli	2		ND	ST4682, ST4684
Parwal beans	bla CTXM-15		E. coli	1		ND	ST641
Peppermint	bla CTXM-3		K. pneumoniae	1		ND	ST15
Cha-om (acacia)	bla SHV-12		K. pneumoniae	1		ND	ND
	bla CTXM-55		E. coli	2		ND	ST167, ST393
	bla CTXM-14		E. coli	1		ND	ST58
Garlic chives	bla CTXM-63		K. pneumoniae	1		ND	ST1743
	bla CTXM-55		E. coli	1		ND	ST226
Lemongrass	bla CTXM-14		K. pneumoniae	1		ND	ST1530
Sweet basil	bla SHV-2a		K. pneumoniae	1		ND	ST76
Basil leaves	bla CTXM-65		E. coli	1		ND	ST4683
Celery	bla RAHN-1	2011	R. aquatilis	34	The Netherlands	ND	ND	[42]
	blaSHV-60, blaDHA-1	2013–2014	K. pneumoniae	1	Algeria	blaTEM-1D, aadA1, strB, strA, qnrB4, oqxB, oqxA, fosA, sul1,tet(A), dfrA1	ST236	[48]
Lollo rosso leaves	bla CTX-M-14	2011–2013	E. coli	1	Germany	strA, strB, sul1, dfrA1	ST973	[45]
Lollo rosso and Lollo bionda leaves	bla CTX-M-2		E. coli	1		blaTEM-1, strA, strB, aadA5	ST120
Blanched celery	bla SHV-12	2012–2013	E. coli	1	The Netherlands	ND	ND	[3]
	bla FONA-1		S. fonticola	ND		ND	ND
Radish	bla RAHN-1	2012–2013	R. aquatilis	1	The Netherlands	ND	ND	[3]
	blaFONA(1–6)		S. fonticola	ND		ND	ND
	bla RAHN-1	2011	R. aquatilis	ND		ND	ND	[42]
Chicory	bla RAHN-1	2011	R. aquatilis	ND	The Netherlands	ND	ND	[42]
Endive	bla RAHN-1	2011	R. aquatilis	ND		ND	ND
	blaFONA-1 (1–6)	2012–2013	S. fonticola	ND	The Netherlands	ND	ND	[3]
Iceberg lettuce + arugula	bla SHV-12	2015–2016	E. coli	3	Italy	ND	ND	[43]
	bla CTX-M-15		E. coli	1		ND	ND
Mixed green vegetables	bla CTXM-15	2017	E. cloacae	1	Canada	blaTEM-1, qnrB1, aac(6′) Ib cr	ND	[55]
Sambhar vegetables	blaCTXM15, blaSHV-106		K. pneumoniae	1		ND	ST101
Aster scaber	blaCTX-14, blaTEM-1	2018	E. coli	1	South Korea	ND	ST69	[51]
Perilla leaf	blaCTX-M-27, blaTEM-1			1		ND	ST349
Sweet potato stalk	bla CTX-M-15			1		ND	ST224
Pepper leaf	blaCTX-M-55, blaTEM-1			1		ND	ND
Mapleleaf ainsliaea	bla CTX-M-27			1		ND	ST10
Leafy greens	bla CTX-M	2015–2016	Enterobacterale	1	United States	ND	ND	[53]
	bla CMY		Enterobacterale	6		ND	ND
Frisee salad	blaCTX-M-1, blaDHA-1	2015–2016	E. cloacae	2	Italy	ND	ND	[43]
Frisee salad + carrot	bla CTX-M-15		Pantoea agglomerans	6		ND	ND
Peach	bla CTX-M-15	2013–2014	K. pneumoniae	1	Algeria	aadA2, strB, strA, qnrS1, oqxB, oqxA, fosA, mph(A), catA2, sul1, sul2, dfrA12	ST219	[48]

In Africa, the first recorded ESBL and/or cephalosporinase-positive GNB was observed in 2019 in South Africa. In this report, 545 vegetable samples including spinach, cucumbers, tomatoes, green beans and lettuce, were collected from street-trading greengrocers, mobile trolley vendors, formal retailers and vendors at two farmers markets from September 2017 to May 2018. ESBL genes were detected in 39 strains, while AmpC production was observed in 20 strains belonging to 10 genera of Enterobacteriaceae including S. fonticola, Serratia marcescens, E. coli, E. cloacae, Enterobacter asburiae, Enterobacter cowanii, Enterobacter ludwigii, R. aquatilis, K. pneumoniae, Klebsiella oxytoca, Citrobacter freundii, Proteus mirabilis and Proteus penneri. Different blaCTX-M genes were obtained, including blaCTX-M-14 (n = 15), blaCTX-M-15 (n = 6), blaCTX-M-27 (n = 4) and blaCTX-M-55 (n = 3). In addition, the blaTEM-3 gene (n = 3), as well as blaSHV genes encoding blaSHV-18 (n = 6), blaSHV-28 (n = 1), and blaSHV-154 (n = 1) were detected. Three isolates carried more than one ESBL gene; two strains (E. cowanii and E. coli) harbored the blaTEM-3 gene in association with blaSHV-18 and blaCTX-M-14 genes, respectively, while one E. coli isolate carried blaCTX-M-14, blaSHV-18 and blaTEM-3 genes. AmpC genetic determinants were observed in 18 of 58 (31%) isolates, 17 strains carried only one pAmpC gene, including blaMIR-20 (n = 4), blaMIR-16 (n = 3), and blaACT-58 (n = 2), and one isolate each harbored blaMIR-14, blaCMY-2, blaACT-2, blaACT-10, blaACT-29, blaEC-30, blaCMY-161 or blaCMY-87, respectively. As well, one P. penneri isolate harbored three AmpC genetic determinants (blaDHA-18, blaCMY-49 and blaACT-10). Among these 17 isolates, five strains (Enterobacter spp. (n = 2), R. aquatilis (n = 1), E. coli (n = 1) and S. fonticola (n = 1)) also carried ESBL genes [9]. Another report from South Africa described the detection of twenty enterobacterial isolates, identified as E. asburiae, E. coli, K. pneumoniae, R. aquatliis and S. fonticola, harboring different ESBL and AmpC genes, including blaCTX-M-group1, blaTEM, blaSHV, blaOXA and blaCIT genes [47]. In Algeria, Mesbah Zekar et al. reported the identification of multi-drug resistant K. pneumoniae isolates in fresh fruit and vegetables purchased in Bejaia city. In this study, eleven K. pneumoniae isolates harbored multiple ESBL genes, and blaCTX-M-15, blaOXA-1, blaSHV-101 and blaSHV-28 were described. In addition, two K. pneumoniae strains coharbored blaDHA-1 with ESBL genes [48].

In Asia, Usui et al. analyzed 130 samples of fresh vegetables collected from seven supermarkets in Japan, 10 out of the 130 samples contained ESBL-producing Pseudomonas spp. including; P. hunanensis, P. putida, P. parafulva, P. beteli, P. mosselii, P. paralactis and P. arsenicoxydans. These isolates harbored the blaSHV-12 or blaTEM-116 ESBL gene [49]. In China, a nationwide survey investigated the prevalence of ESBL-producing Enterobacteriaceae from retail food, where four isolates were obtained. Three were identified as E. coli and one as C. freundii isolated from retail vegetables, including tomatoes, cucumber and coriander. The C. freundii isolate carried blaCTX-M and blaOXA genes, while two E. coli isolates harbored blaCTX-M and blaSHV genes and one other E. coli strain carried blaCTX-M, blaSHV and blaTEM genes [7]. In Malaysia, ESBL or AmpC genes were detected in two E. coli (blaCTX-M-55 and blaCTX-M-65) and two K. pneumoniae isolates (blaCTX-M-15, blaSHV-28 and blaDHA-1) from coriander and chili pepper respectively [50]. In addition, different CTX-M variants were described from E. coli isolates in South Korea including CTX-M-14, CTX-M-15, CTX-M-27, CTX-M-55 and the CTX-M-65 variant [51].

On the American continent, different Enterobacteriaceae isolates harbored blaSHV, blaTEM and blaCTX-M-1 as well as blaCTX-M and blaCMY genes and were detected from iceberg lettuce and leafy greens, respectively in the United States [52,53]. Moreover, seven E. coli isolates carrying the blaCTX-M-15 gene were reported from leaf lettuce, alfalfa and parsley/cilantro in Ecuador [54], while blaCTX-M-14, blaCTX-M-15, blaCTX-M-27, blaSHV-106 and blaSHV-142-positive Enterobacteriaceae were reported in Canada from imported vegetable samples [55], and the blaCTX-M-15 gene in Brazil [56].

6. Vegetables and Fruit Isolates with Carbapenemase Genes

Only eight reports have revealed the isolation of Gram-negative bacteria producing carbapenemase genes from vegetables and fruit (Figure 1, Table 2). The first study describing carbapenemase-producing Gram-negative bacteria from fresh vegetable samples was published in 2015. Samples were purchased from different retail shops specializing in Asian food from three different cities in Switzerland, imported from Vietnam, Thailand, and India. In this study, only one Klebsiella variicola strain carrying the blaOXA-181 gene was isolated from a coriander sample from Thailand/Vietnam, and the obtained isolate coharbored a quinolone resistant gene (qnrS1). These data suggest that the international production of imported fresh vegetables constitutes a possible reservoir for the spread of carbapenemase-producing Gram-negative bacteria, especially Enterobacteriaceae [57].

Table 2

Carbapenemases and mcr genes reported in Gram-negative bacteria isolates from vegetables and fruit worldwide.

Vegetables Type	Carbapenemase/mcr Gene	Isolation Period	Species	Isolates Number	Country	Other Antibiotic Resistance Genes	Sequence Type	Plasmid Type	Reference
Coriander	bla OXA-181	2015	Klebsiella variicola	1	Switzerland	qnrS1	ND	IncX3	[57]
Lettuce	blaKPC-2 and blaNDM-1	2015	Escherichia coli	1	China	blaDHA-1, fosA3, floR aacA4, tet(D), sul1, armA, mph(E), msr(E), erm(B), strA, strB	ST877	IncA/C (blaNDM-1), Untypable (blaKPC-2)	[60]
	bla OXA-48	2016	K. pneumoniae	1	Algeria	bla TEM-1	ST391	ND	[11]
	bla NDM-5	2017	E. coli	1	China	blaCTX−M−1G, fosA3, floR	ST4762	X3	[61]
				1		blaCTX−M−1G, fosA3, floR, oqxAB.	ST4762	X3
	bla NDM-1		C. freundii	1		fosA3, floR, qnrB	ND	ND
Parsley	bla NDM-1	2015	K. pneumoniae	1	Japan	blaSHV-28, blaSHV-1, blaTEM-1A, blaCTX-M-15, blaCTX-M-14b, blaTEM-1A, blaOXA-9, fosA, oqxAB, tet(D), aac(69)-Ib, aadA1, aph(39)-VI, aph(6)-Id, aph(39)-VIb, aph(39’)-Ib, aac(69)-Ib-cr, qnrS1	ST15	ND	[59]
	bla OXA-48	2016	K. pneumoniae	1	Algeria	None	ND	ND	[11]
	bla OXA-48	2019	E. cloacae	1	Romania	ND	ND	ND	[46]
	bla KPC		K. oxytoca	1		ND	ND	ND
Baby leaf mix	bla NDM-1	2015	K. pneumoniae	1	Japan	blaCTX-M-15, blaOXA-9, blaTEM-1A, blaSHV-28, blaCTX-M-14b, fosA, oqxAB, aac(69)-Ib, aadA1, aph(39)-VI, aac(69)-Ib-cr, qnrS1, aph(6)-Id, aph(39)-VIb, aph(39’)-Ib.	ST15	ND	[59]
	blaOXA-66, blaOXA-72		A. baumannii	1		blaADC-25, blaOXA-66, blaOXA-72, sul2, tet(B), aac(3)-Ia aac(69)-Ip, aph(39’)-Ib, aph(6)-Id	ST2	GR2 (blaOXA-72)
Cucumber	bla KPC-2	2017	K. pneumoniae	1	China	qnrB, oqxAB	ST23	F35:A-:B1	[61]
				1		blaCTX−M−1G, qnrB, oqxAB	ST23	F35:A-:B1
	bla NDM-5		E. coli	1		fosA3, floR, qnrB	UT	ND
				1		blaCTX−M−1G, fosA3, floR,.	ST4762	ND
	blaNDM-5 and blaKPC-2		E. coli	1		None	ND	ND
Curly endive	bla NDM-5	2017	E. coli	1	China	blaCTX−M−1G, fosA3, floR, rmtB	ST167	ND	[61]
	bla NDM-1		C. freundii	1		fosA3, floR, qnrB	ND	ND
				1		fosA3, floR, oqxAB, qnrB, rmtB	ND	ND
				1		fosA3, floR, qnrB	ND	X3
Tomato	bla OXA-48	2016	K. pneumoniae	1	Algeria	None	ST1877	ND	[11]
	bla NDM-1	2017	C. freundii	1	China	blaCTX−M−1G, fosA3, floR, qnrB	ND	X3	[61]
Leaf rape	bla NDM-5	2017–2018	E. coli	1	China	mcr-1, fosA3	ST156	X3	[4]
Spinach	bla NDM-9			1		mcr-1, fosA3	ST2847	Untypable
	bla KPC	2019	Morganella morganii	1	Romania	ND	ND	ND	[46]
Vegetables (ND)	bla OXA-72	ND	Acinetobacter calcoaceticus	2	Lebanon	ND	ND	ND	[58]
Lettuce	mcr-1	2013	E. coli	1	Portugal	blaTEM-1, aadA1y, aph(4)-Ia, estX-12, floR, sat2, strA, strB, sul2, tetA	ST1716	ND	[62]
		2013–2014	E. coli	1		blaTEM-1, aac(3)-Iv, aadA1, aph(4)-Ia, aph(6)-Id, mdf(A)-type, tetA, sul2, floR	ST1716	ND	[63]
		2016	E. coli	1	China	blaCTX-M-14, floR, fosA3, oqxAB	ST795	IncHI2	[64]
				1		blaCTX-M-55, floR	ST2505	ncI2
		2015	E. coli	1		blaCTX-M-55, rmtB, floR, fosA3.	ST156	IncI2
		2016	E. coli	1		floR	ST48	IncX4
		2015	Raoultella ornithinolytica	2		blaCTX-M-14, floR, fosA3, oqxAB	NA	IncHI2
	mcr-1	2018	E. coli	1	South Korea	blaTEM-1 and blaCTX-M-55	ST10	ND	[65]
	mcr-1	2017–2018	E. coli	1	China	ND	ST10	X4	[66]
				1		ND	ST2705	HI2
Tomato	mcr-1	2016	E. coli	1	China	blaCTX-M-14, floR, fosA3, oqxAB	ST69	IncHI2	[64]
		2015	E. coli	2		floR	ST206	chromosome
	mcr-1	2017–2018	E. coli	1	China	ND	ST713	X4	[66]
				1	China	ND	UT	I2
Leaf rape	mcr-1	2017–2018	K. pneumoniae	1	China	blaNDM-5, fosA3	ST156	X4	[4]
			E. coli	1	China	ND	ST744	X4	[66]
Green Pepper	mcr-1	2017–2018	E. cloacae	1	China	ND	ND	ND	[66]
	mcr-1		E. coli	1		ND	ST5873	X4
Spinach	mcr-1	2017–2018	E. coli	1	China	blaNDM-9, fosA3	ST2847	I2	[4]
	mcr-1			1		ND	ST2253	I2	[66]
Cha-om	mcr-1	2014	E. coli	1	Switzerland	bla CTX-M-55	ST167	ND	[67]
Basil leaves	mcr-1	2014	E. coli	1		bla CTX-M-65	ST4683	ND
Cucumber	mcr-1	2017–2018	E. coli	2	China	ND	ST744	X4	[66]
				2		ND	ST1115	I2
Carrot	mcr-1	2017–2018	E. coli	1		ND	ST5539	X4
			E. coli	1		ND	ST13	I2
Curly endive	mcr-1	2017–2018	E. coli	1		ND	ST13	X4
Pak choi	mcr-1	2017–2018	E. coli	1		ND	ST648	I2
Apple	mcr-1	2016	E. coli	1	China	aadA2, aadA1,floR, cmlA1, sul2, sul3, tetA, tetM, dfrA12,mdfA	ST189	IncFIA	[68]
Orange	mcr-1		K. pneumoniae	1		blaSHV-110, qnrS1, oqxA, oqxB, fosA6, sul1, tetA, dfrA1	ST442	IncHI1

Since its first report, carbapenemase genes have been identified in bacteria from different vegetable samples in only five countries across the Asian, African and European continents. Indeed, OXA-72-producing Acinetobacter calcoaceticus strains have been found in two vegetable samples purchased from the same market in Beirut, Lebanon [58]. In Japan, two K. pneumoniae and one Acinetobacter baumannii isolate were collected from vegetable samples in the city of Higashi-Hiroshima. Both K. pneumoniae isolates carried blaNDM-1 with other genes conferring resistance to β-lactams (blaCTX-M-15, blaOXA-9, and blaTEM-1A), aminoglycosides (aac(6’)-Ib, aadA1, and aph(3’)-VI), quinolones (qnrS1) and fluoroquinolones (aac(6’)-Ib-cr). While the obtained A. baumannii isolate coharbored blaOXA-66, blaOXA-72 and genes conferring resistance to sulfonamides (sul2), tetracycline (tet(B)), and streptomycin (strAB) [59]. In China, Wang et al., identified an Escherichia coli strain coproducing blaKPC-2 and blaNDM-1 genes in fresh lettuce from a market in Guangzhou. In addition, this multidrug resistant E. coli strain coharbored fosfomycin resistance genes (fosA3 and floR). This study represents the first report of either blaKPC or blaNDM genes in bacteria obtained from vegetables [60]. Additionally, two other studies from China reported the detection of carbapenemase-positive enterobacterial isolates. The first reported the isolation of twelve carbapenem-resistant isolates obtained from vegetable samples, where the highest detection rate was found in curly endive samples. The authors identified two K. pneumoniae isolates carrying the blaKPC-2 gene, while five of each E. coli and C. freundii strains harbored the blaNDM gene, including four E. coli with the blaNDM-5 gene and five C. freundii with blaNDM-1. Notably, one E. coli strain from a cucumber sample harbored blaNDM-5 and blaKPC-2 genes simultaneously. All C. freundii and E. coli isolates carried fosfomycin resistance genes (fosA3and floR), and all K. pneumoniae and C. freundii isolates harbored the floR gene. However, one strain of E. coli and C. freundii harbored the aminoglycoside resistance gene (rmtB). Quinolone resistance genes including oqxAB and qnrB, were found in four and eight isolates, respectively [61]. The second report signaled the detection of two E. coli isolates carrying blaNDM genes in leaf rape and spinach recovered from two supermarkets in Shandong province; one isolate concomitantly harbored blaNDM-9, mcr-1 and fosA3, while the second isolate carried blaNDM-5, mcr-1 and fosA3 genes [4].

From the African continent only one study from Algeria has been reported. The authors identified three K. pneumoniae isolates harboring the blaOXA-48 gene from lettuce, tomatoes and parsley in Béjaïa city [11]. In Europe and more precisely from Romania, the blaOXA-48 and blaKPC genes were detected in E. cloacae and K. oxytoca isolates from parsley samples [46].

7. Vegetables and Fruit Isolates with the mcr Gene

To date, eight studies have reported mcr-producing Gram-negative bacteria, especially isolates of Enterobacteriaceae species, from fresh produce that mostly originated from China (Figure 1, Table 2). The mcr-1 gene was first reported in 2016 in Switzerland in two out of sixty isolates. The two E. coli isolates carried the mcr-1 gene and coharbored blaCTX-M-55 and blaCTX-M-65 genes, respectively [67]. After this first description in fresh produce, mcr-1-producing GNB isolates on fresh produce were reported in China, where seven E. coli and two Raoultella ornithinolytica isolates were recovered from tomato and lettuce samples between May 2015 and August 2016 in Guangzhou. All the obtained mcr-1-positive strains harbored the florfenicol resistance gene (floR). Of the nine isolates, six strains carried the blaCTX-M gene (four blaCTX-M-14 and two blaCTX-M-55), with five and four strains harboring the fosA3 and oqxAB efflux pump gene, respectively [64]. Moreover, the mcr-1 gene was described in China from a total of 528 fresh vegetable samples, including 18 different types purchased from 53 supermarkets and farmers markets from 23 districts or cities in nine provinces between May 2017 and April 2018. Of the 528 samples analyzed, only 19 samples harbored one or more mcr-positive isolates, and the three highest detection rates were noted in carrots (14.3%), pakchoi (13.3%) and green pepper (7.7%), followed by leaf lettuce (5.6%), leaf rape (4.9%), romaine lettuce (4.3%), tomato (3.5%), spinach (3.2%), cucumber (3.1%), and curly endive (2.4%). In the above study, twenty-four mcr-1-positive isolates were obtained; twenty-three strains were identified as E. coli and one as E. cloacae. Fourteen mcr-1-positive strains coproduced the blaCTX-M gene, nine strains harbored the blaCTX-M-9G gene and three strains carried blaCTX-M-1G. However, the remaining two strains harbored both blaCTX-M-9G and blaCTX-M-1G genes. In addition, eight and two isolates harbored fosA3 and rmtB genes, respectively. Plasmid-mediated resistance to quinolones (PMQR), including oqxAB, qnrS and qnrB genes, were also detected in this study [66]. Additionally, in the same country, two E. coli isolates carrying the mcr-1 gene were isolated from leaf rape and spinach in Shandong province. These isolates coharbored metallo-β-lactamase and fosA3 genes; the first carried blaNDM-5, while the second harbored the blaNDM-9 gene [4]. In 2018, one E. coli isolate carrying the mcr-1 gene recovered from lettuce was reported in South Korea. The obtained isolate coharbored blaTEM-1 and blaCTX-M-55 genes [65]. In Portugal, the mcr-1 gene was reported by Manageiro et al. in 2020, and they documented the presence of this gene in an E. coli strain isolated from a lettuce sample. In silico analysis showed the presence of additional antibiotic resistance genes including blaTEM-1, aac(3)-Iv, aadA1, aph(4)-Ia, aph(6)-Id, mdf(A)-type, tetA, sul2 and floR-type [63]. Another report from Portugal revealed the detection of the mcr-1 gene in an E. coli strain from conventionally produced lettuce. The isolate co-carried blaTEM-1, aph(4)-Ia, floR, sat2, strA, strB, sul2 and tetA genes, while no conventional and organic fruit were positive for the mcr-1 gene [62].

On fruit samples, the mcr-1 gene was detected in E. coli and K. pneumoniae isolates from apple and orange samples recovered in China [68].

8. Contamination Pathways and Genetic Characteristics of β-Lactamases and mcr-Producing Gram-Negative Bacteria

The high diversity of global clones illustrates the extensive spread of ESBL-producing K. pneumoniae and E. coli isolates on vegetables around the world (ST45, ST219, ST15 and ST147 found in K. pneumoniae isolates, and ST410-A, ST44, ST405, ST131 and ST38 in E. coli isolates). In Algeria, sequence type 14, ST45, ST219, ST236, and ST882 have been identified among K. pneumoniae strains recovered on fresh fruit and vegetables carrying ESBL or cephalosporinase genes, including blaCTX-M-15, blaOXA-1, blaSHV-101, blaSHV-28 and blaDHA-1 genes that were mostly (11 of 13) located on the IncFII plasmid, while the IncR plasmid replicon was identified in only one isolate [48]. In Switzerland, twenty-two different sequence types identified in E. coli-positive ESBL have been described on imported vegetables, including ST4684 and ST4683, four of them belonging to the epidemiologically important sequence types ST405 (n = 1), ST131 (n = 2) and ST38 (n = 2) [10]. Similarly, the ST131 E. coli clone is known for its role in the global spread of ESBLs, especially CTX-M-15, and this clone has had an inevitable clinical impact on antibiotic resistance and pathogenicity [69]. In the same study conducted in Switzerland, high clonal diversity was observed among K. pneumoniae strains, with two isolates belonging to the epidemic clones ST15 and ST147 [10]. In Quito, Ecuador, the hyper epidemic clones ST410 and ST44 harboring the blaCTX-M-15 gene have been identified in E. coli isolates from leaf lettuce, alfalfa and parsley/cilantro, and three of them were found on the same integron 1 variable region (dfrA1 and addA5). The five remaining isolates presented blaCTX-M-15 downstream of an insertion sequence element p1 (ISEcp1) [54]. In a survey in Germany, a high diversity of global clones was identified in ESBL-producing E. coli isolates from different vegetable samples. These ESBL determinants were detected on different plasmids as follows: IncHI2 and IncK (blaCTX-M-14, ST527, ST973), IncN (blaCTX-M-65, ST10), IncFIB (blaCTX-M-15, ST410), IncHI2 (blaCTX-M-125, ST542) and IncFIA-FIB (blaCTX-M-2, ST120) [45]. The IncFIB and IncFIC plasmid replicons were found in Pseudomonas hunanensis, and P. putida carried the blaSHV-12 respectively.The plasmid IncK/B was reported in P. paralactis harboring the blaTEM-16 gene [49].

The dissemination of carbapenemase-producing E. coli is polyclonal, where multiple STs have been reported. The sequence type 877 was reported in an E. coli strain coproducing blaKPC-2 and blaNDM-1 genes isolated from fresh lettuce. These genes were located on 64 and 118 kb plasmids, designated as plasmids pHNTS79-KPC and pHNTS79-NDM, respectively [60]. Furthermore, ST1877 and ST391 have been detected in OXA-48-positive K. pneumoniae isolates from lettuce, tomatoes and parsley in Algeria, where the ST391 clone is considered as an emergent carbapenemase-producing lineage of clinical importance [11]. The blaOXA-181 detected in a Klebsiella variicola isolate from a coriander sample was mediated by the IncX3-type plasmid of 51-kb [57]. Indeed, in Japan the epidemic clones ST15 and ST2 were reported among K. pneumoniae that carried the blaNDM-1 gene and the A. baumannii strain coharbored blaOXA-66, and blaOXA-72 genes, respectively [59]. ST15 is a relatively common NDM-positive K. pneumoniae lineage, and it has been found in various countries across different continents, almost all of which were isolated from humans [70]. IncX3 plasmids carrying the blaNDM gene have been identified in E. coli and C. freundii strains isolated from cucumbers, and the identified IncX3 plasmid was identical or highly similar (99%) to the IncX3 plasmids identified from patients in other countries. In addition, similar F35:A-:B1 plasmids were described in two blaKPC-2-producing-K. pneumoniae isolates belonging to ST23 obtained from different cities. Two E. coli isolates carrying the blaNDM-5 gene isolated from cucumber and romaine lettuce samples in different cities in China shared an identical PFGE pattern and sequence type (ST4762); however, one E. coli strain belonged to ST167 [61].

Among E. coli isolates, the mcr-1 gene was found in multiple STs from different countries, ST10 in South Korea, and ST167 and ST4683 in Switzerland [65,67]. ST156 and ST2847 have been identified in China from leaf rape and spinach samples. In the latter, mcr-1 genes were located on the~60-kb IncI2 plasmid or the ~33-kb IncX4 plasmid; even as blaNDM-5 was on the ~46-kb IncX3 plasmid while blaNDM-9 was on the ~120-kb untypeable plasmid. The detected plasmids were highly similar to those from patients and animals described in different countries [4]. In a Chinese study, six sequence types, including ST795, ST2505, ST69, ST156, ST48 and ST206, were described in seven E. coli isolates recovered from lettuce and tomato samples. For the four E. coli isolates, mcr-1 genes were located on IncHI2, IncI2 or IncX4 plasmids, while for the two Raoultella ornithinolytica isolates, the mcr-1 gene was located on the IncHI2 plasmid [64]. Moreover, in China, sixteen STs along with a new ST type were identified, while the most prevalent STs were ST744 and ST224. In this study, different plasmid replicons were detected, including IncX4, IncI2 and IncHI2; where IncX4 was the most detected and shared highly similar RFLP profiles, although they were from different cities and fresh vegetables. The mcr-1 gene was located on the IncX4 plasmid of ∼33 kb in size [66].

In addition to the above results, the study found that patients and animals shared identical or highly similar plasmids with vegetables [4,61]. Various other studies reported that vegetables may become contaminated with multidrug resistant bacteria from soil, manure fertilization, irrigation water or through direct contamination by humans [2]. In this context, the major way in which antibiotic resistance enters the soil is through the use of animal manure [71]. In Australia, Zhang et al. explored the impact of cattle and poultry manure application on the resistome in lettuce and soil microbiomes, including the rhizosphere, root endosphere, leaf endosphere and phyllosphere. In addition, they identified potential transmission routes of antibiotic resistance genes in the soil–plant system. The authors reported that poultry manure application increased antibiotic resistance genes in the rhizosphere, root endophyte and phyllosphere, while cattle manure use increased the abundance of antibiotic resistance genes only in the root endophyte, suggesting that poultry manure may have a stronger impact on lettuceresistomes. Moreover, the authors also identified an overlap of antibiotic resistance gene (ARG) profiles between lettuce tissues and soil, which indicates that plant and environmental resistomes are interconnected, and confirmed the transmission of antibiotic resistance genes from manured soil to vegetables. Two main transmission pathways were reported: an internal pathway through plant tissues and an external pathway via aerosol from the atmosphere to the plant surface. Thus, in the external pathway, sixty-nine ARGs were shared between poultry manure-amended soil and the phyllosphere of lettuce, while in the internal pathway 47 genes were common between rhizosphere soil and the root endophyte [72]. This finding was consistent with previous studies which reported that the phyllosphere resistome was significantly more abundant and diverse than the endophytic resistome in organic vegetables [73,74]. Another report from China described the impact of the long-term use of inorganic (chemical) and organic (manure) fertilizers on antibiotic resistance genes in greenhouse soils growing vegetables. The results showed that both inorganic and organic use increased the abundance and the diversity of soil ARGs, with a difference in the dominant ARG types. ARG abundance and diversity were both higher in organic fertilizer [75]. These data confirmed those of previous reports indicating that fertilizer application, especially organic fertilizer such as animal manure, raised ARG abundance and diversity in soil compared to soil without fertilization [75,76,77].

Several studies have reported food-borne human outbreaks linked to the consumption of fresh vegetables and fruit irrigated with wastewater and indicated that the type of irrigation practice plays a vital role in the contamination of farm produce [78]. In this context, Araújo et al. characterized the presence of E. coli isolates on vegetables and in irrigation water sampled from 16 household farms in Portugal. In this later study, different commonly acquired genes such as blaTEM, tetA and tetB and plasmids (IncFIC, IncFrep and IncFIB) were detected in isolates in both water and vegetable samples. In addition, rep-PCR typing results detected the same STs and identical clones in vegetables and water, suggesting cross-contamination. These results suggest that irrigation groundwater is a reservoir of antibiotic resistant E. coli and may enter the food chain via vegetable consumption [79]. Makkaew et al. evaluated the contamination of lettuce by E. coli, grown under four diverse methods of wastewater irrigation: open spray, open drip, spray under plastic sheet cover, and drip under plastic sheet cover. E. coli contamination was reported in all lettuce samples in both open and covered spray beds in all types of spray beds. An equal level of microbial quality of spray bed lettuce and submersed lettuce irrigated with wastewater containing 1299.7 E. coli MPN/100 mL was detected, and this result was similar in both laboratory and experimental investigations [80]. In Ghana, Antwi-Agyei et al. reported that irrigation with partially treated and untreated wastewater is a key risk factor for the observed contamination of 80% of produce samples, with a median concentration ranging from 0.64 to 3.84 log E. coli/g produce, while ready-to-eat salad was the most contaminated with 4.23 log E. coli/g [81].

9. Conclusions

This review provides a reference for an enhanced understanding of the global risk of fresh vegetables and fruit in the transmission of multidrug resistant Gram-negative bacteria and emphasizes the necessity of paying close attention to these products as a future public health issue. Given that fresh produce is often consumed raw, this allows the transfer of these antibiotic resistance genes to human gut bacteria. It is now even more important that more investigations should be performed in order to survey the emergence and transmission of these genes to humans from farm to fork. In addition, suitable measures, including the improvement of water quality and agricultural practices, need to be considered to ensure consumer safety worldwide.