A Review of SHV Extended-Spectrum β-Lactamases: Neglected Yet Ubiquitous

β-lactamases are the primary cause of resistance to β-lactams among members of the family Enterobacteriaceae. SHV enzymes have emerged in Enterobacteriaceae causing infections in health care in the last decades of the Twentieth century, and they are now observed in isolates in different epidemiological settings both in human, animal and the environment. Likely originated from a chromosomal penicillinase of Klebsiella pneumoniae, SHV β-lactamases currently encompass a large number of allelic variants including extended-spectrum β-lactamases (ESBL), non-ESBL and several not classified variants. SHV enzymes have evolved from a narrow- to an extended-spectrum of hydrolyzing activity, including monobactams and carbapenems, as a result of amino acid changes that altered the configuration around the active site of the β -lactamases. SHV-ESBLs are usually encoded by self-transmissible plasmids that frequently carry resistance genes to other drug classes and have become widespread throughout the world in several Enterobacteriaceae, emphasizing their clinical significance.


INTRODUCTION
Thanks to their ability to inhibit cell wall biosynthesis, β-lactams remained the first-line defense against bacterial infections for over 20 years, before resistant bacteria appeared in clinical practice.
Resistance to this class of drugs can be the result of antibiotic target site alteration, prevention of antibiotic access by altered permeability or forced efflux, or antibiotic degradation (Wilke et al., 2005). The latter, represents the primary resistance mechanism in Gram-negative bacteria producing β-lactamase enzymes able to covalently bind the carbonyl moiety of the β-lactam ring and hydrolyze its amide bond (Fisher et al., 2005). Naturally occurring chromosomally located βlactamases are quite common in Gram-negative bacteria; likely evolved from penicillin-binding proteins, when produced in small quantity they do not significantly contribute to antibiotic resistance. It was the appearance of the first plasmid-mediated β-lactamase TEM-1 (Datta and Kontomichalou, 1965) to designate the beginning of an unstoppable phenomenon in the 1960s. Ever since, the introduction of new natural or synthetic drugs to replace old ones in an attempt to limit the insurgence of antibiotic resistant bacteria triggered a chain reaction providing bacteria with a constant selective pressure driving the expansion of different resistance mechanisms (Medeiros, 1997).
In recent years β-lactamases have extensively diversified in response to the clinical use of new generations of β-lactams (penicillin, cephalosporins, carbapenems, and monobactams) leading to the need of classification schemes. Based on primary structure (Ambler, 1980), enzymatic properties and biochemical attributes (Bush et al., 1995), and the increasingly available amino acid sequences (Bush and Jacoby, 2010) four major classes (A, B, C, D) can be acknowledged. Serine β-lactamases belonging to class A are the most abundant (Philippon et al., 2016), with more than 500 enzymes, including the most clinically significant extended spectrum β-lactamases (ESBL) variants, i.e., CTX-M-, TEM-, and SHV-type enzymes (Bush and Fisher, 2011).
Although, SHV enzymes did not undergo the explosive dissemination observed for CTX-M-type variants (Canton et al., 2012), in recent years they have been found in several Enterobacteriaceae outside of the typical clinical hosts Klebsiella pneumoniae and Escherichia coli, with a rising allele variability (http://www.lahey.org/studies), and in different environmental niches. Many admirable works describing the biochemistry, the genetics and the evolution of SHV β-lactamases have appeared over the last years. The aim of this review is to provide the readers with an updated overview on SHV β-lactamases, their amino acid variants and spectrum of activity, and to describe the occurrence of plasmid-associated SHV enzymes in Enterobacteriaceae and their epidemiological significance.

ORIGIN AND DIVERSITY OF THE SHV FAMILY
The first bla SHV-1 gene was identified in the 1970s in E. coli (Pitton, 1972). The encoded enzyme SHV−1 (sulfhydryl reagent variable) proved its activity against penicillins and first generation cephalosporins (Matthew et al., 1979) and was confirmed part of the conjugative plasmid p453 (Barthélémy et al., 1988; Table 1). The most likely ancestor of the plasmid-mediated SHV−1 is a chromosomal species-specific penicillinase detected in fecal K. pneumoniae isolates from neonates (Haeggman et al., 1997). The enzyme showed a typical antibiogram with penicillin rather than cephalosporin resistance and a marked inhibition by clavulanic acid. How bla SHV-1 moved from the chromosome to the plasmid does not have a conclusive explanation since the proposed association with a transposable element (Nugent and Hedges, 1979) has not been confirmed.
More than half of these variants (n = 99) has not been classified yet due to absence of biochemical characterization. Figure 1 illustrates a phylogenetic analysis of 149 out of the 189 SHV β-lactamase variants whose amino acid sequences were available online (http://www.lahey.org/studies), as of July 2016. Unlike other β-lactamase families (D'andrea et al., 2013;Evans and Amyes, 2014), there is no clear clustering of the different subgroups, as also mirrored by gene based analysis (Supplementary Figure S1). Among the majority of unclassified variants, subgroup 2b and the few 2br variants are scattered all over the tree. Subgroup 2be showed clustering of most of the ESBL variants (including SHV-2a, , together with few non-classified enzymes . It has been proposed that SHV β-lactamases descended from an unidentified ancestor holding an extended spectrum phenotype (2be) and that subgroup 2b derived from it (Hall and Barlow, 2004). Our analysis showed that several of SHV ESBL variants were scattered along the tree with short branch lengths with neighboring 2b or unknown variants within the SHV phylogeny (i.e., SHV−40, SHV−11, and SHV−35; Figure 1), supporting the hypothesis that they evolved from multiple variants, probably within the antibiotic era. Among the non-ESBL variants, bla  represents one of the most successful and, together with bla SHV-1 , the likely source of evolution for the existing SHV ESBL variants. bla SHV-11 was first identified as plasmid-encoded in clinical K. pneumoniae from Switzerland (Nüesch-Inderbinen et al., 1997) and ever since has been isolated worldwide.
Although, nearly displaced, together with TEM, by CTX-M enzymes over the years (Canton et al., 2012), 46 ESBL bla SHV genes have been described so far ( Table 1). The first report of SHV-mediated resistance to third-generation cephalosporins was in 1983 with the isolation and characterization of bla SHV-2 , encoded by plasmid pBP60 in a German clinical isolate of Klebsiella ozaenae and showing only a few nucleotide mismatches with bla SHV-1 (Kliebe et al., 1985). In a few years four other ESBL variants were identified as plasmid-encoded in clinical K. pneumoniae, showing variable gene homologies with the bla SHV-1 and bla SHV-2 sequences (50-90%): bla SHV-2a encoded by conjugative plasmid pZMP1 (Podbielski et al., 1991); bla SHV-3 on pUD18 (Nicolas et al., 1989); bla  , widely disseminated from France as a result of a single K. pneumoniae clone diffusion (Arlet et al., 1990(Arlet et al., , 1994; and bla SHV-5 able to hydrolyze broad-spectrum cephalosporins and monobactams (Gutmann et al., 1989). Of these first variants, the most epidemiologically successful were bla SHV-2a and bla SHV-5 , which will be further discussed, together with bla SHV-2 and bla SHV-12 , in a dedicated paragraph (Section Expansion toward New Ecological Niches). Interestingly, bla SHV-3 and bla SHV-4 have been only sporadically detected since their first description. bla SHV-3 seems to be geographically restricted to the USA where it was detected in E. coli of animal origin, associated with other antibiotic resistance genes such as bla CTX-M-15 , bla CTX-M-24 , bla CMY-2 , and/or bla TEM-1 (Shaheen et al., 2011). bla SHV-4 was identified also in Enterobacter aerogenes and Citrobacter diversus in different countries (Arpin et al., 1996;El Harrif-Heraud et al., 1997;Pitout et al., 1998). *Isolation or first description. **Non ESBL genes bla FIGURE 1 | Maximum likelihood amino acid tree of 149 SHV-type β-lactamases. Variants whose sequence has not been released in GenBank as of July 2016, that show partial sequence or are identical to others (http://www.lahey.org/studies/) were not included in the analysis. SHV-180 and SHV-181 share the same sequence as well as SHV-121 and SHV-136. The tree was implemented in Mega version 6.06 (Tamura et al., 2013). Solid circles represent: red, extended-spectrum β-lactamases (2be; n = 46); green, broad-spectrum β-lactamases (2br, n = 5); and blue, penicillinases (2b, n = 30). Unclassified alleles are reported in black (n = 68).
Most of SHV ESBLs (25 out of 46) are unique cases, with only one report so far. Seventeen variants are exclusively found in clinical K. pneumoniae: bla SHV-6 , bla SHV-13 , bla SHV-16 , bla SHV-18 , bla SHV-23 , bla SHV-45 , bla SHV-64, bla SHV-66, bla SHV-86, bla SHV-90, bla SHV-91, bla SHV-98, bla SHV-99, bla SHV-100, bla SHV−104, bla SHV-105 , and bla SHV-134 . These variants have been described worldwide (Brazil, Portugal, Algeria, USA, Tunisia, Netherlands, France, South Africa, Colombia, and China) and are mostly associated to plasmids ( Table 1). Some of these variants are sporadically accompanied by other antibiotic resistance genes like in the case of: (i) bla SHV-45 encoded by an IncA/C plasmid together with bla CTX-M-2 and bla SHV-27 (Dropa et al., 2015); (ii) bla SHV-134 encoded by an IncFII plasmid accompanied by a second plasmid carrying bla VIM−1 (Sánchez-Romero et al., 2012); (iii) and bla SHV-105 , conferring reduced susceptibility to ceftazidime, ceftriaxone, and aztreonam together with bla SHV-1 , and bla SHV-5 (Jones et al., 2009). One of the oldest variants, bla SHV-6 , was only described in France in 1991 in a K. pneumoniae clinical case (Arlet et al., 1991). It might be speculated that the 180 kb plasmid encoding bla SHV-6 and conferring decreased susceptibility to ceftazidime and aztreonam was not stable or it reduced bacterial strain fitness preventing a successful dissemination. Four variants have been described only in clinical E. coli: (i) bla SHV-15 , described together with bla CMY-2 in a strain imported from India into the United Kingdom (http://www.lahey.org/ studies/); (ii) bla SHV-24 , identified in Japan on a transferable 150 Kb plasmid conferring high-level resistance to ceftazidime but not cefotaxime and cefazolin (Kurokawa et al., 2000); emergence of SHV-24 might have been driven by the extensive use of ceftazidime in Japan, enabling bacterial survival in high concentrations of this drug; (iii) bla SHV-102 , recovered in a Spanish hospital and hydrolyzing cefotaxime and ceftazidime ; (iv) and bla SHV-129 , detected in an abscess specimen from a patient hospitalized in Italy in 2008 (Lascols et al., 2012).
bla SHV-46 was only described on a 70 Kb conjugative plasmid also carrying bla TEM-1 and bla KPC−2 in a carbapenem-resistant strain of Klebsiella oxytoca from the urine of a hospitalized patient in New York (USA) in 1998 (Yigit et al., 2003). Finally, bla SHV-34 is an interesting example of extended-spectrum β-lactamase encoded by an epidemic plasmid circulating among Citrobacter koseri, E. coli, and K. pneumoniae in the same US hospital between 1998 and 2000 (Heritage et al., 2003).
Majority of SHV ESBLs have been detected in K. pneumoniae or E. coli (Table 1). bla SHV-30 was the first variant to be detected in an E. cloacae isolate from a blood culture from a solid-organ transplant recipient in the USA in 2003 (Szabó et al., 2005). The gene, previously described in K. pneumoniae and Salmonella (Mulvey et al., 2004;Whichard et al., 2007), was located on a 9.4 Kb plasmid and contributed together with chromosomal ampC, bla SHV-7 , and bla TEM-1 to the antibiotic resistance profile of the E. cloacae isolate, the first of its kind producing two different SHV enzymes. Three other novel ESBL variants have been solely identified as plasmid-encoded in clinical E. cloacae: (i) bla SHV-70 , from a Chinese patient with history of ceftazidime treatment (Ling et al., 2006) and observed in other clinical Chinese settings (Liu et al., 2008); (ii) bla SHV-128 , isolated in Tunisia in 2009, located on an IncFII conjugative plasmid, and conferring resistance to all β-lactams except imipenem (Bourouis et al., 2015); (iii) and bla SHV-183 , for which additional description is not available (http://www.lahey.org/studies/).

SHV EXTENDED-SPECTRUM β-LACTAMASES: CATALYTIC PROPERTIES AND RESISTANCE PHENOTYPE
Extended-spectrum SHV β-lactamases belong to functional group 2be, while very recently they were assigned to subclass A1 of serine β-lactamases, clustering with TEM and CTX-M enzymes among other clinically relevant β-lactamases (Bush, 2013;Philippon et al., 2016). SHV ESBLs consist of two subdomains: an α/β that includes an antiparallel five-stranded β-sheet flanked by α-helices, and an all-α-helical subdomain (Matagne et al., 1998). Similar to TEM β-lactamases (Jelsch et al., 1993), the active site is located within the cleft created by the subdomains and it contains the Ser 70 residue that mediates the nucleophilic attack on the carbonyl group of the β-lactam ring. In the vicinity of this serine residue, several conserved structural and functional amino acid motifs have been identified. These include the Ser 70 XXLys ("SXXK" motif, with X representing variable amino acids), the Ser 130 AspAsn ("SDN" motif), the Glu 166 XXLysAsn ("EXXLN" motif), and the Lys 234 Thr/SerGly ("KTG" motif) (Bush, 2013).
Residue Leu 35 is located further away from the active site of class A β-lactamases and its substitution to Gln (e.g., SHV-2a, SHV-12) has been suggested to have an indirect role in enhancing the extended-spectrum capability of SHV β-lactamases (Nüesch-Inderbinen et al., 1997). In contrast, Gly 238 and Glu 240 amino acids are part of the active site lying near the R1 side chain of the β-lactam (Huletsky et al., 1993). Substitutions in Gly 238 either to Ser (e.g., SHV-2, SHV-2a) or Ala (e.g., SHV-13, SHV-18) displace the β3-strand from the reactive Ser 70 , resulting in a slightly expanded active site. This conformational change improves the binding to and the accommodation of newer cephalosporins with large C7 substituents, thereby expanding the substrate spectrum of these SHV ESBLs to include cefotaxime and to a lesser extent to ceftazidime (Huletsky et al., 1993;Matagne et al., 1998;Nukaga et al., 2003). It has been suggested that Glu 240 substitutions to Arg (SHV-86) or Lys (e.g., SHV-4, SHV-5) cause the ammonium group of the long side-chains of these residues to form an electrostatic bond with the carboxylic acid group on the oxyimino-substituents of ceftazidime and aztreonam (Knox, 1995). This interaction has a dual effect on the hydrolysis of ceftazidime by improving initial binding and facilitating proper positioning within the SHV β-lactamase, whereas the hydrolysis of other β-lactams is less affected (Huletsky et al., 1993). Gly 238 Ser and Glu 240 Lys amino acid substitutions characterize the majority of SHV ESBLs ( Table 2) and mirror those seen in extended-spectrum TEM β-lactamases. Interestingly, a plethora of extended-spectrum SHV and TEM β-lactamases exhibit higher levels of hydrolytic activity against ceftazidime than against cefotaxime (ceftazidimases) ( Table 3). This phenotype was attributed to the Glu 240 Lys substitution, in contrast with most CTX-M β-lactamases lacking this critical substitution and only showing a cefotaximase activity, (Bonnet, 2004).
Overall, the available SHV ESBL kinetic parameters show that most of the substitutions lead to more efficient hydrolysis of oxyimino-β-lactams than penicillins, as depicted by the low K cat values for penicillins ( Table 3). While they retain their ability to hydrolyze penicillins, they are not catalytically so efficient compared to SHV-1 (Bush and Singer, 1989) and this is due to the decreased strength of the crucial hydrogen-bonding network needed for penicillin catalysis (turnover). As a consequence, since β-lactam inhibitors (clavulanic acid, sulbactam, and tazobactam) are structurally very similar to penicillin substrates, SHV ESBLs also exhibit increased susceptibility to β-lactam inhibitors compared to SHV-1 ( Table 3) leading to less inhibitor required for inactivation (lower K i and IC 50 s; Tzouvelekis and Bonomo, 1999).

DETECTION
There are at least 46 known SHV-ESBL genes together with more than 150 non-ESBL or unclassified alleles to date (http:// www.lahey.org/studies/). Accurate identification of these variants is essential for surveillance and for epidemiological studies of transmission mode, particularly in clinical setting, where appropriate antimicrobial therapy is critical.
A panel of different phenotypic confirmatory tests is available to determine the presence of extended-spectrum β-lactamases, including SHV-variants: minimum inhibitory concentration (MIC) determination of β-lactam with and without clavulanic acid, double disk synergy test (DDST), inhibitor potentiated disk diffusion test (IPDDT), three-dimensional test (TDT) and commercially available methods (Etest for ESBLs, Vitek ESBL cards, MicroScan panels, and BD Phoenix Automated Microbiology System (Bradford, 2001;Paterson and Bonomo, 2005). Standard microbiological procedures can take up to several days for culture, isolation and characterization and many comparative studies have shown that PCR-based methods have higher sensitivity (Bedenic et al., 2001(Bedenic et al., , 2007Singh et al., 2012), mostly due to variable levels of gene expression. Therefore, PCR and nucleotide sequence analysis (Stürenburg et al., 2003), together with various PCR-based methods, remain the gold standard for extended-spectrum β-lactamase SHV-variants identification.
Chanawong and colleagues developed a PCR-restriction fragment length polymorphism (PCR-RFLP) method to allow the identification of new SHV β-lactamases variants through detection of known mutations that alter recognition sites of restriction endonucleases (Chanawong et al., 2000). PCR-RFLP complements pre-existing PCR-single strand conformational polymorphism (PCR-SSCP) limited by partial gene amplification, thus missing potential mutation sites (M'Zali et al., 1998). PCR-RFLP can also be used in combination with restriction site insertion-PCR (RSI-PCR), a method based on primers mismatches, allowing the unambiguous identification of up to 27 SHV variants by point mutation (Chanawong et al., 2001a). Fluorescently labeled hybridization probes followed by melting curve analysis can also be used to discriminate between ESBL and non-ESBL bla SHV genes (Randegger and Hächler, 2001). This method, termed the SHV melting curve mutation detection method, is also able to categorize SHV ESBL producers into phenotypically relevant subgroups: (i) weak ceftazidime resistance (SHV-6 and SHV-8); (ii) significant resistance to cefotaxime and ceftriaxone and moderate resistance to ceftazidime (SHV-2, SHV-2a, and SHV-3); and (iii), most effective against all expanded-spectrum cephalosporins (SHV-4, SHV-5, SHV-9, and SHV-12). Combined systems can also be developed ad hoc to rapidly screen local epidemiological settings (Chia et al., 2005). A modified SHV melting-curve mutation detection method able to distinguish between prevalent Taiwanese bla SHV genes (SHV-1, SHV-2, SHV-2a, SHV-5, SHV-11, and SHV-12) was combined with a multiplex PCR to identify different β-lactamases genes (bla SHV , bla CTX-M-3 -like, and bla CTX-M-14 ). The design of this method can be easily adapted to other geographic areas where different ESBLs are prevalent. Multiplex real-time PCR assays for the fast detection of extended-spectrum β-lactamase and carbapenemase genes were developed with differential melting curves able to recognize up to 120 different SHV allelic variants (Singh et al., 2016).
New techniques for ESBL detection are employed alongside PCR-based methods these days. Loop-mediated isothermal amplification (LAMP) was applied to detect SHV-and other ESBL-producing bacteria in meat and proved to be more specific and sensitive than MacConkey agar or cefpodoxime disc methods (Anjum et al., 2013). Commercial DNA microarrays are also proving themselves to be accurate, with sensitivity and specificity values for ESBL detection being high. Up to 53 SHV-variants can be covered on a same array (Leinberger et al., 2010), but on the other hand some alleles may fail to be detected (i.e., SHV-12), as previously reported (Stuart et al., 2012). Because arrays have major limitations to detect novel genes or variants, PCR and sequencing remains essential. Matrixassisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry (MS) is routinely used for bacteria identification and has been recently applied to detect ESBLproducing Enterobacteriaceae from positive blood cultures in clinical practice (Jung et al., 2014;Oviaño et al., 2014). Although, this methodology has yet to be fully validated, preliminary results show 99% sensitivity and 100% specificity, and denote a novel approach to categorize bacteria as ESBL producers.
Pyrosequencing combines standard PCR and sequencing by synthesis to rapidly determine the sequence of a target DNA region; it has been extensively used for the detection of bacterial resistance genes and bacterial community composition (Tang et al., 2016;Tian et al., 2016). This technique has been used to perform mutation analysis of bla SHV to resolve heterogeneous sequences in clinical isolates of K. pneumoniae containing more than one SHV variant (Haanperä et al., 2008). An alternative protocol for pyrosequencing is the singlenucleotide polymorphism (SNP), ideal for the sequencing of mixed templates and determination of SNPs at the position of interest. This protocol has been applied to discriminate between eight bla SHV variants from clinical isolates of E. coli and K. pneumoniae, reporting great reproducibility and ability to discriminate between sequences (Jones et al., 2009). A multiplex pyrosequencing assay coupled with qPCR amplification has also been recently developed to enable rapid and accurate detection of bla SHV and bla TEM -producing Enterobacteriaceae (Deccache et al., 2015). Overall, pyrosequencing can be a useful epidemiological tool for the exact identification of bla SHV as a prerequisite for analyzing the spread of certain SHV variants.
Finally, the advent of whole genome sequencing (WGS) has taken differentiation of bacterial strains and identification of the associated antibiotic resistance gene cargo to another level. Aside from the phylogenetic analysis that WGS provides, the complete resistome of a strain can be unraveled as well as its mobilome, i.e., the mobile genetic elements that are associated with antibiotic resistance diffusion. Only this information can provide us with full understanding of complex genomic structures as observed, for example, in clinical K. pneumoniae genomes carrying (i) nineteen antibiotic resistance genes including bla OXA-1 and bla SHV-28 in the chromosome, bla NDM-1 in a plasmid, and bla OXA-232 in a second plasmid (Kwon et al., 2016); (ii) β-lactamase genes bla KPC−2 , bla SHV-11 , bla TEM−169 , and bla OXA-9 , together with aac(6 ′ -)Ib, aadA2, and aph(3 ′ -)Ia as aminoglycoside resistance encoding genes, mph(A) for macrolides, oqxA and oqxB for quinolone, catA1 for phenicol, sul1 for sulfonamide, and dfrA12 for trimethoprim (Lee et al., 2014); or (iii) six different plasmids, adding up to 0.43 Mbp, coding for six β-lactamases (bla SHV-12 , bla OXA-9 , bla TEM-1 , bla CTX-M-2 , and bla KPC−2 ), together with bla SHV-110 and adhesin-related gene clusters on the chromosome (Perreira Ramos et al., 2014).

EXPANSION TOWARD NEW ECOLOGICAL NICHES
Over the last years the presence of antibiotics as well as antibiotic resistant bacteria has been shown outside the clinical environment, including water, soil and, most notably, food producing animals. When looking at SHV-variants distribution it is evident that in recent years, as for most extended-spectrum β-lactamases (Canton et al., 2012), their presence has been confirmed in virtually all ecological niches (Supplementary Table  S1), making it more challenging to restrain antibiotic resistance diffusion. The most representative cases and variants will be discussed.

Aquatic Environment
In an effort to control the release of antibiotics and antibiotic resistant bacteria in the environment, aquatic environments are being investigated worldwide, whether they be natural, drinking or wastewaters. The latter are particularly worrisome given the high prevalence of bla SHV alleles, as observed in untreated hospital wastewater in Australia (Gündogdu et al., 2013), their possible association with determinants of quinolone and other βlactamase resistance (Calhau et al., 2015;Osinska et al., 2016), and their relatively easy transmission to surface water through waste water treatment plant discharges (Marti et al., 2013). Studies showed that SHV types, together with CTX-M and OXA genes can be significantly decreased by biological treatments such as activated sludge processing and anaerobic digestion, although not all can be effectively eliminated (Yi et al., 2015).
Urban waters are also exposed to relatively high population densities and therefore are often unprotected from biological contaminants, with people playing a crucial role in antibiotic resistance dissemination in the environment. Unusual finding of SHV-producing Stenotrophomonas maltophilia in a swimming recreational Serbian lake and its transient presence during summer months can be considered as a proof of its anthropogenic origin, given its nature of emerging nosocomial pathogen (Novovic et al., 2015). Similar conclusions can be drawn for SHV-producing K. pneumoniae and E. cloacae isolated from a Bangladeshi lake, which receives waste water from surrounding residents, commercial buildings and clinics in Dhaka city (Haque et al., 2014), as well as for artificial water reservoirs in Poland (Wolny-Koladka and Lenart-Boron, 2016), or urban surface waters in Malaysia (Tissera and Lee, 2013). In recent surveillance studies of different rivers and lakes in Switzerland, bla SHV-12 -producing Enterobacteriaceae were isolated only in 4% of the cases (Zurfluh et al., 2013), although this variant is predominant in clinical Swiss isolates (Nüesch-Inderbinen et al., 1997). bla SHV-12 was also detected in Enterobacteriaceae from seawater, together with tet(A) and sul2 in Portugal (Alves et al., 2014), and plasmid-encoded together with bla TEM-1 and/or bla CTX-M-1 in Croatia (Maravic et al., 2015). Finally, data on ESBL-producing Enterobacteriaceae isolated from drinking water is also increasing, reporting SHV alleles in rural water reservoirs in China (Zhang et al., 2015), or drinking water sources for First Nation communities in Canada (Fernando et al., 2016).

Food Producing Animals
Food producing animals have become subject of increasing interest after several studies demonstrated that resistant strains of animal origin can be associated to human infections, possibly through the food chain (Hasman et al., 2005). Majority of SHV variants in this reservoir belong to bla SHV-2 , bla SHV-2a , bla SHV-5 , and bla SHV-12 (Supplementary Table S1) owing to their successful association with conjugative plasmids (see Section Plasmid epidemiology of bla SHV-2 , bla SHV-2a , bla SHV-5 , and bla SHV-12 ).
Surveillance activities in healthy animals worldwide are generating a tremendous amount of data on ESBL distribution. Most SHV β-lactamase producers are E. coli from swine and broiler fecal samples as observed in China (Tian et al., 2012); in Spain, with bla SHV-2 associated with bla CTX-M-9 and bla SHV-12 with bla CTX-M-1 , in pigs and broilers respectively (Blanc et al., 2006); in layers, cattle, and broilers but not in swine in Japan (Hiki et al., 2013;Kameyama et al., 2013); and in the Netherlands, where healthy broilers carried bla SHV-2 in combination with bla TEM-1 or bla TEM−135 . Other Enterobacteriaceae like K. pneumoniae and Citrobacter freundii were positive for bla SHV-2 or bla SHV-12 from poultry and swine, respectively (Machado et al., 2008).
Finding ESBL producers in food producing animals is also mirrored by positive food samples worldwide, mostly retail chicken meat, as reported in Tunisia, with E. coli carrying bla SHV-5 isolated from different butcheries, supermarkets, and local markets (Jouini et al., 2013), or Salmonella enterica carrying bla SHV-12 in Japan (Noda et al., 2015). The crosscontamination between food producing animals and retail meat has been internationally demonstrated due to the detection of plasmid-borne SHV variants, such as bla SHV-2 and bla SHV-2a from Canadian chicken meat and abattoir chicken cecum (Pouget et al., 2013) or bla SHV-2 and bla TEM-1 in Japan (Hiroi et al., 2011), presenting the potential for horizontal transfer between Enterobacteriaceae as a high public health concern.

Wildlife, Companion Animals, and Vegetables
ESBL diffusion has been studied extensively in Enterobacteriaceae from humans and livestock, whereas information on antibiotic resistance in the environment is still limited. Yet, the dissemination success of bla SHV-12 is confirmed by its introduction into the wildlife, notably in birds, as reported in Spain (Alcalá et al., 2015), the Netherlands (Veldman et al., 2013), Poland (Literak et al., 2010), and the Czech Republic (Dolejská et al., 2009). This success is likely associated to predominant avian clones and to efficient plasmids (Table 4,  Figure 3) of the IncN incompatibility group, described to be more frequent in pathogenic than in commensal avian and human E. coli strains (Johnson et al., 2007). bla  was also detected in E. coli from several birds of prey in Portugal, alone or in associations with bla TEM−1b (Pinto et al., 2010).
Lastly, SHV variants have been detected in imported vegetables in Switzerland together with bla SHV-12 for the first time in the opportunistic foodborne pathogen Cronobacter sakazakii whose potential to cause bacteremia and meningitis is an actual concern (Zurfluh et al., 2015). Similar results were observed in vegetables collected in South Korea , salads in the Netherlands (Reuland et al., 2014), and Spain (Egea et al., 2011), displaying a new route of introduction for ESBLs and pathogenic Enterobacteriaceae.
PLASMID EPIDEMIOLOGY OF bla SHV-2 , bla SHV-2A , bla SHV-5 , AND bla SHV-12 The role of plasmids in the successful spread of β-lactamase genes has been extensively described (Carattoli, 2009(Carattoli, , 2013 and, among the SHV family, it finds its best examples in bla SHV-2 , bla SHV-2a , bla SHV-5 , and bla SHV-12 . Combination of these alleles with different dissemination machineries has brought the enzymes to reach diverse niches worldwide (Figure 2). Plasmids belonging to seven replicon types (A/C, F, HI2, I1, L/M, N, X3) have been shown to drive the epidemiology of these four predominant SHV ESBLs, although their distribution varies on the plasmid families (Table 4). Other rep families that have been only incidentally associated with extended-spectrum SHV β-lactamases include the ColE, K, P, and R ( Table 4).
Remarkably, bla SHV-12 on IncI1 plasmids belonging to pST26 have been identified among E. coli isolates of human and animal origin (Accogli et al., 2013;Jones-Dias et al., 2015), indicating the potential transmission of these bla SHV-12 -encoding vehicles from human to animals and/or vice versa.

Miscellaneous Plasmids
bla SHV-12 has been incidentally found on: (i) a ColE plasmid from S. enterica serotype Typhimurium DT104b in Spain (Herrera-Leon et al., 2011); (ii) an IncK plasmid from K. pneumoniae in the United Kingdom (Timofte et al., 2014); (iii) an IncP plasmid from E. cloacae in Taiwan (Chen C. M. et al., 2015); and (iv) a plasmid assigned to the R replicon type from K. pneumoniae in Portugal (Rodrigues et al., 2014). E. coli and K. pneumoniae encoding bla SHV-2 on IncK plasmids were recovered from animal and human sources in the Netherlands and in Uruguay, respectively García-Fulgueiras et al., 2011). IncP plasmids encoding bla SHV-2a from animals in Canada and bla SHV-5 from human in Uruguay have also been reported (García-Fulgueiras et al., 2011;Pouget et al., 2013). Finally, a number of reports highlight the presence of bla SHV-12 on mostly conjugative non-typeable plasmids, according to the PCR-based replicon-typing scheme (Carattoli et al., 2005a). These plasmids of human origin, varying between 50 and 140 Kb in size, were mostly detected among E. coli from the United Kingdom (Doumith et al., 2012) and K. pneumoniae from Tunisia (Elhani et al., 2010) and United Arab Emirates (Sonnevend et al., 2013), underscoring that their dissemination is wider than we know.

OUTSIDE OF THE ENTEROBACTERIACEAE AND A FEW PECULIAR SHV ESBLS
SHV β-lactamases have virtually invaded all human, environmental and animal sceneries, mostly associated to Enterobacteriaceae. In recent years, the first reports of alternative bacterial hosts have been described, notably in Aeromonads, ubiquitous in aquatic habitats and occasionally able to cause human infections. bla SHV-12 was detected, in association with bla FOX-2 and bla CTX-M-15 , on the chromosome of the foodborne pathogens A. caviae and Aeromonas hydrophila from wild-growing mussels from Croatia (Maravić et al., 2013). The first identification of plasmid-encoded SHV-12, together with VIM-1, occurred in clinical A. caviae accountable for a newborn bloodstream infection (Antonelli et al., 2016). The coproduction of these enzymes highlights the potential risks for public health and the role of Aeromonads as reservoirs and dissemination tools of resistance determinants in both environmental and clinical settings. Occasionally, SHV ESBL-producing Pseudomonas aeruginosa can be detected in nosocomial settings and can pose a serious threat as healthcare-associated infection in many regions of the world. bla SHV-2a was first identified on the chromosome of a 1995 clinical P. aeruginosa strain, with high sequence homology to plasmid pMPA2a from K. pneumoniae indicating a likely enterobacterial gene origin (Naas et al., 1999). Subsequent studies demonstrated the insertion of bla SHV alleles into P. aeruginosa chromosome: bla SHV in China (Chen Z. et al., 2015) and Iran (Shahcheraghi et al., 2009); bla SHV-5 (Poirel et al., 2004) and bla SHV-12 (Neonakis et al., 2003) in Greece; and bla SHV-2a in Tunisia (Mansour et al., 2009) and France (Hocquet et al., 2010;Jeannot et al., 2013). The role of IS26 in the mobilization of bla SHV-12 was demonstrated by the chromosomal insertion of an IS26 composite transposon (>24 kb) thanks to the co-mobilization of antibiotic resistance aac(6 ′ )-Ib, which confers amikacin resistance, likely occurred during the clinical course of a burn infection, immediately after amikacin administration (Uemura et al., 2010). bla SHV-5 , bla SHV-11 , bla SHV-12 were also detected in different combinations, together with bla TEM-1b , on various plasmids in Thailand (Chanawong et al., 2001b).
Finally, one of the most effective associations outside of the Enterobacteriaceae is with Acinetobacter baumannii, contributing to the worrisome spread of ESBL-producing strains especially in clinical outbreaks (Blackwell et al., 2016). Plasmid transfer from nosocomial SHV-encoding Enterobacteriaceae seems to be responsible for this phenomenon, as observed for SHV-12 in the Netherlands (Naiemi et al., 2005), or SHV-5 in the USA, a country where this variant is the most prevalent ESBL gene in Enterobacteriaceae (Naas et al., 2007).
Among all ESBL SHV β-lactamases, few enzymes deserve special consideration because of their unique enzymatic features. SHV-38 is a unique allelic variant of the SHV family to have an expanded-spectrum to carbapenems. It was first described in K. pneumoniae from France (Poirel et al., 2003) and it holds a point mutation (Ala 146 Val) compared to the chromosome-encoded SHV-1. Among all 46 available SHV ESBL variants, only SHV-38 possess the Ala 146 Val substitution (Table 2), likely inducing subtle structural conformational changes favoring imipenem but not meropenem hydrolysis (Walther-Rasmussen and Høiby, 2007).
SHV-129 is a novel clinically acquired variant identified in 2012 from an Italian E. coli isolate ( Table 1; Lascols et al., 2012) and it represents an interesting example of enzyme evolution due to antibiotic pressure. Alongside two well-known amino acid substitutions (Gly 238 Ser, Glu 240 Lys), SHV-129 contains new substitutions, Arg 275 Leu and Asn 146 Asp). The latter was recently demonstrated to be the first global suppressor substitution identified in the SHV β-lactamase family (Winkler and Bonomo, 2016), likely helping in protein stabilization and functionality, as well as in the ability of the enzyme to acquire additional substitutions. It is also proposed that due to the increasing clinical use of cefepime, SHV-129 might have evolved from SHV-2 or SHV-5 in an alternative conformation to expand its spectrum to hydrolyze cefepime, as mirrored by the kinetic parameters of the three enzymes (Table 3).
Finally, SHV-2 can be located on both chromosome and self-transmissible plasmids ( Table 4; Supplementary Table S1). Association of bla SHV-2 with RCS47, a P1-like bacteriophage that infects and lysogenizes E. coli and several other enteric bacteria, was recently reported (Billard-Pomares et al., 2014). bla SHV-2 is flanked by two IS26 elements that likely drove the insertion in the phage backbone. The P1-like prophages were found with high prevalence in natural E. coli of both animal and human origin, including ESBL-producing isolates. This kind of association was already reported for other β-lactamases (bla TEM , bla CTX-M , and mecA) from river and urban sewage water (von Wintersdorff et al., 2016), suggesting that bacteriophages might play a wider role in favoring horizontal transfer of antibiotic resistance determinants than initially thought (Muniesa et al., 2013).

CONCLUDING REMARKS
Tzouvelekis and Bonomo suggested than "it will not be surprising if (SHV) enzymes will continue to expand their substrate spectrum as long as the current antibiotics, or novel ones derived from the basic β-lactam structure, are used" (Tzouvelekis and Bonomo, 1999). In the last two decades we observed the appearance of multiple SHV-type variants, with few ones significantly expanding their substrate. One exception is represented by SHV-38, the only known SHV allelic variant able to hydrolyze carbapenems (Poirel et al., 2003), a feature that has not been associated with any TEM or CTX-M enzyme. In this image resides the fate of SHV extended β-lactamases, unable to undergo the dominant propagation observed, for instance, for the CTX-M family but yet contributing to β-lactam resistance in a not negligible way.
The persistence of SHV enzymes in the bacterial community might also be secured by co-selection with emerging resistance genes. Association of bla SHV-12 with IncX3 plasmids carrying carbapenemase genes bla KPC−2 and bla NDM has been observed in recent years (Table 4) and it seems to be a phenomenon occurring in clinical carbapenem-resistant Enterobacteriaceae worldwide (Kassis-Chikhani et al., 2013;Sonnevend et al., 2013;Partridge et al., 2015;Huang et al., 2016). As highlighted in this review, the association of successful variants bla SHV-2 , bla SHV-2a , bla SHV-5 , and bla SHV-12 with different families of conjugative plasmids (IncA/C, IncF, IncHI2) might also underlie the colonization of virtually all ecological niches encompassing food producing animals, aquatic environment, wildlife, companion animals, and vegetables. Plasmid mediated transfer from nosocomial Enterobacteriaceae enabled SHV dispersion toward alternative bacterial hosts such as the emerging nosocomial pathogens of aquatic origin S. maltophilia and A. caviae, or contributed to the worrisome spread of ESBLproducing strains of A. baumannii and P. aeruginosa. Most interestingly, the ubiquitous presence of SHV ESBL genes and plasmids is suggestive for transmission in human, animals, and the environment, most likely through the food chain, highlighting the potential risks for public health and endorsing a one health research approach.
Overall, SHV ESBL enzymes have kept a stable role in antibiotic resistance over the years. Allele diversification is still occurring, the latest variant being identified in E. cloacae in 2014 (bla SHV-183 ), and effective associations with new genetic platforms are taking place helping expansion toward novel bacterial hosts and reservoirs.

AUTHOR CONTRIBUTIONS
The paper was written by AL and DC, and reviewed by DM. All authors discussed, read, contributed to and approved the final manuscript.