Minireview Molecular mechanisms underlying glyphosate resistance in bacteria

Glyphosate is a nonselective herbicide that kills weeds and other plants competing with crops. Glyphosate speci ﬁ cally inhibits the 5-enolpyruvyl-shikimate-3-phosphate (EPSP) synthase, thereby depleting the cell of EPSP serving as a precursor for biosynthesis of aromatic amino acids. Glyphosate is toxicologically for animals and humans. it the most-important herbicide in agriculture. its intensive application in agriculture is a serious environmental issue it the A few years after the discovery of the mode of action of glyphosate, it been observed that bacteria evolve glyphosate resistance by acquiring mutations in the EPSP synthase gene, rendering the encoded enzyme less sensitive to the herbicide. The identi ﬁ cation of glyphosate-resistant EPSP synthase variants paved the way for engineering crops tolerating increased amounts of the herbicide. This review intends to summarize the molecular mechanisms underlying glyphosate resistance in bacteria. Bacteria can evolve glyphosate resistance by (i) reducing glyphosate sensitivity or elevating production of the EPSP synthase, by (ii) degrading or (iii) detoxifying glyphosate and by (iv) decreasing the uptake or increasing the export of the herbicide. The variety of glyphosate resistance mechanisms illustrates the adaptability of bacteria to anthropogenic substances due to genomic alterations.


Introduction
Glyphosate (N-(phosphonomethyl(glycine) is a nonselective herbicide that is used in agriculture to kill weeds (Franz, 1979;Duke and Powles, 2008). Glyphosate was first synthesized in 1950 by the swiss chemist Henri Martin, who, however, did not discover its herbicidal effects (Franz et al., 1997;Zimdahl, 2010). Twenty years later, John E. Franz, a chemist working at the agrochemical and agricultural biotechnology corporation Monsanto., observed that glyphosate kills plants (Franz et al., 1997;Zimdahl, 2010). Glyphosate was soon patented for herbicide use and its large-scale production began in the second half of the 1970s (Grossbard and Atkinson, 1985;Duke and Powles, 2008;Benbrook, 2016). Since then, glyphosate has become the most widely used herbicide in global agriculture (Duke and Powles, 2008;Benbrook, 2016;Duke, 2018). Glyphosate is a highly polar and water-soluble compound (Franz, 1979). Therefore, in its pure form, glyphosate has little effectiveness as most plants' hydrophobic cuticle repels it. The herbicide is usually formulated with surface-active substances, so-called surfactants, to facilitate the uptake and thus the effectiveness of glyphosate in killing plants (Knowls, 1998). Once the plant has taken up glyphosate, the herbicide is transported via the phloem to the target in various plant tissues.
Glyphosate specifically inhibits the 5-enolpyruvylshikimate-3-phosphate (EPSP) synthase of the shikimate pathway in archaea, bacteria, Apicomplexa, algae, fungi and plants ( Fig. 1A; Amrhein, 1980, 1984;Amrhein et al., 1983;Comai et al., 1983;Schulz et al., 1984;Kishore and Shah, 1988;Roberts et al., 1998;Roberts et al., 2002;Zhi et al., 2014). The EPSP synthase converts the glycolytic intermediate phosphoenolpyruvate (PEP) and shikimic acid-3-phosphate (S3P) into EPSP,  Fig. 1. The shikimate pathway and the molecular target of glyphosate. A. The shikimate pathway and substances inhibiting its enzymes. 7-Deoxy-seduheptulose (7dSH) and chlorogenic acid were shown to inhibit the dehydroquinate synthase (Brilisauer et al., 2019;Neetu et al., 2020). IMB-T130, curcumin and glyphosate inhibit the dehydroquinate dehydratase (Zhu et al., 2018), the shikimate dehydrogenase  and the enolpyruvyl 3-phosphate (EPSP) synthase  respectively. B. Overly of the EPSP synthase structure models from Streptococcus pneumoniae in complex with shikimate 3-phosphate (S3P) and glyphosate (GS; wheat colour; PDBid: 1RF6; Park et al., 2004)  which serves as an essential precursor for de novo synthesis of the aromatic amino acids, phenylalanine, tyrosine and tryptophan as well as of the vitamins folic acid and menaquinone (Wilson et al., 1998;Herrmann and Weaver, 1999). Moreover, the shikimate pathway provides precursors for secondary plant metabolism (Herrmann and Weaver, 1999). Since the shikimate pathway enzymes are essential for many organisms, the pathway is an attractive target for antibiotics ( Fig. 1A) The inhibition of the EPSP synthase by glyphosate results in the depletion of the cellular levels of aromatic amino acids and thus, in plants death (Gresshoff, 1979). Glyphosate also inhibits the growth of bacteria and other organisms unless the aromatic amino acids are provided by the environment (Gresshoff, 1979;Amrhein et al., 1983;Fischer et al., 1986;Wicke et al., 2019). As the glyphosateinduced deficiency in aromatic amino acids impairs bacterial growth, it is not surprising that the herbicide has a severe effect on the general physiology of the bacteria (Kang et al., 2011;Lu et al., 2013). For instance, a transcriptome analysis of Escherichia coli exposed to glyphosate identified the differential expression of more than 1000 genes, representing about 23% of the genome (Lu et al., 2013). Moreover, sublethal concentrations of glyphosate reduce the susceptibility of enterobacteria to clinically important antibiotics (Kurenbach et al., 2015;Pöppe et al., 2020). Recently, it has been observed that the application of glyphosate increases the prevalence of antibiotic resistance genes in soil microbiomes (Liao et al., 2021). However, this phenomenon seems to be due to an enrichment of antibiotic resistance genes and not caused by a genome-wide glyphosate-induced increase in the mutation rate (Tincher et al., 2017;Liao et al., 2021). Although glyphosate is considered toxicologically safe for animals and humans (Li and Long, 1988), the scientific debate about the toxicity of the weedkiller glyphosate is still ongoing (Arj o et al., 2013;Klingelhöfer et al., 2020). As described above, glyphosate is applied together with surfactants for improving the uptake of the herbicide by the plants. Many studies uncovered that the toxicity of the co-formulants is in fact much higher than that of the herbicide itself (Relyea, 2005;Mesnage et al., 2013Mesnage et al., , 2019Mesnage and Antoniou, 2017;Defarge et al., 2018;Hao et al., 2019).
The existing scientific literature also contains conflicting results about the role of glyphosate in perturbing or changing the microbial activity and composition in the soil. For instance, it has been observed that the microbial activity and the composition of the soil, which had been exposed to glyphosate was significantly stimulated and changed respectively (Haney et al., 2002;Araújo et al., 2003). Other studies demonstrate that the treatment of cultivated soil with glyphosate only slightly altered the composition of the microbial community (Barriuso et al., 2011;Schlatter et al., 2017;Dennis et al., 2018). Moreover, glyphosate did not affect the soil microbial communities associated with crops across diverse farming systems (Kepler et al., 2020). In contrast to this, the effect of glyphosate on the microbiota of animals seems to be unambiguous (Shehata et al., 2013). A recent study revealed that the composition of the honeybee gut microbiota is altered due to the exposure of pure glyphosate and glyphosate present in a commercial herbicide formulation (Motta et al., 2018;Motta and Moran, 2020). Moreover, glyphosate also increases the susceptibility of bees to an infection by opportunistic pathogens like Serratia marcescens (Motta et al., 2018). Thus, the massive global use of glyphosate in agriculture could be the reason for the decline in the populations of bees and other insects.
A few years after the discovery of the mode of action of glyphosate in 1980 , it was observed that bacteria evolve glyphosate resistance by acquiring mutations in the EPSP synthase gene, rendering the encoded enzyme less sensitive to the herbicide (Comai et al., 1983;Barry et al., 1992). However, these observations are not surprising because bacterial populations can reach high cell densities and subpopulations may have spontaneously accumulated beneficial mutations, providing the cells with a selective growth advantage (Gunka et al., 2013). Due to the extensive use of glyphosate in agriculture, also plants evolved mechanisms of glyphosate resistance (Baerson et al., 2002;Sammons and Gaines, 2014). This review intends to summarize the molecular mechanisms underlying glyphosate resistance in bacteria. These organisms can evolve glyphosate resistance by (i) reducing glyphosate sensitivity or elevating production of the EPSP synthase, by (ii) degrading or (iii) detoxifying glyphosate and by (iv) decreasing the uptake or increasing the export of the herbicide. The variety of glyphosate resistance mechanisms illustrates the adaptability of bacteria to anthropogenic substances due to genomic alterations.
Evolution of glyphosate resistance by decreasing the glyphosate sensitivity or elevating the production of the EPSP synthase Ten years after the herbicidal effects of glyphosate were discovered, its molecular target was identified . As described above, the glyphosate-dependent inhibition of the EPSP synthase prevents the formation of the aromatic amino acids, phenylalanine, tryptophan and tyrosine (Fig. 1A). Biochemical and structural analyses of the EPSP synthase from E. coli revealed that the binding of the two substrates PEP and S3P cause a large conformational change in the enzyme resulting in a catalytically competent active site (Anton et al., 1983;Schönbrunn et al., 2001;Priestman et al., 2005;Pollegioni et al., 2011). The crystal structures of the enzyme in complex with S3P and glyphosate suggests that the herbicide acts as a transition state analogue ( Fig. 1B; Schönbrunn et al., 2001;Park et al., 2004;Pollegioni et al., 2011). In fact, glyphosate replaces PEP in the active site of the EPSP synthase, thereby preventing the synthesis of EPSP, which is the precursor of de novo synthesis of the aromatic amino acids.
To enable glyphosate application on fields for killing unwanted plants competing with crops, it was necessary to develop crops containing a glyphosate-insensitive EPSP synthase. The first EPSP synthase variant carrying a single amino acid substitution (Pro101Ser) and showing a reduced sensitivity towards glyphosate was identified in the enteric bacterium Salmonella typhimurium. For this purpose, S. typhimurium was grown in the presence of toxic glyphosate levels (0.35 g L À1 -2 g L À1 ) and subjected to chemical mutagenesis to facilitate the evolution of an aroA * gene encoding an EPSP synthase, less sensitive to glyphosate (Comai et al., 1983;Stalker et al., 1985). The aroA* was used to generate the first genetically modified tobacco plant (Nicotiana tabacum), tolerating increased levels of glyphosate . Even though the genetically modified plant was still sensitive to glyphosate, the work by Comai and colleagues (.,1985), paved the way for the development of further EPSP synthase variants with reduced glyphosate sensitivity (della-Cioppa et al., 1987;Padgette et al., 1995;Liu et al., 2015;Liang et al., 2017;Fartyal et al., 2018;Liu and Cao, 2018). In the following years, insensitive EPSP synthase variants from various organisms have been obtained by directed evolution and site-directed mutagenesis as well as by cultivating the organism of choice in the presence of toxic glyphosate levels (Padgette et al., 1991;Eschenburg et al., 2002;Healy-Fried et al., 2007;Kahrizi et al., 2007;Funke et al., 2009;Cao et al., 2012;Firdous et al., 2018;Wicke et al., 2019). A glyphosate-insensitive EPSP synthase variant was also identified in the environmental Agrobacterium sp. strain CP4 that was isolated from the waste feed of a glyphosate production factory (Barry et al., 1992). The EPSP synthase variant from the Agrobacterium sp. CP4 strain was used to generate the Roundup Ready® soybean, which tolerates high amounts of glyphosate (Barry et al., 1992). Thus, glyphosate resistance by target modification in bacteria also evolves in nature. Structural analysis of the Agrobacterium sp. CP4 EPSP synthase revealed that an alanine residue at position 100 (Ala100) does not allow glyphosate to bind to the active site (Funke et al., 2006). Usually, in the EPSP synthase of plants, there is a glycine residue at this position, which makes the active site accessible for glyphosate (Pollegioni et al., 2011). The same amino acid exchange in the EPSP synthase of Klebsiella pneumoniae and E. coli (Gly96Ala) also increases glyphosate resistance of the bacteria ( Fig. 1C; Sost and Amrhein, 1990;Eschenburg et al., 2002). Like in bacteria, the modification of the glyphosate target by the acquisition of mutations is also a very widespread resistance mechanism in plants ( Baerson et al., 2002;Heap, 2014;Sammons and Gaines, 2014;Zhang et al., 2017;Wicke et al., 2019;McElroy, 2020).
Interestingly, a recent study with the Gram-positive soil bacterium Bacillus subtilis revealed that the EPSP synthase does not permit any changes that increase the resistance of the enzyme to glyphosate (Wicke et al., 2019). On the one hand, any change in the amino acid sequence of the B. subtilis EPSP synthase could reduce the enzyme activity too much to permit the survival of the bacteria. On the other hand, the limited evolvability of the B. subtilis EPSP synthase could be due to the essentiality of the enzyme (Jordan et al., 2002;Koo et al., 2017;Wicke et al., 2019). Alternatively, the enzyme could fulfil additional cellular functions, which could be disturbed through any change in the amino acid sequence. It will be very interesting to elucidate the underlying reason for the essentiality of the aroE encoding EPSP synthase gene in B. subtilis.
Beside direct modification of the EPSP synthase gene, bacteria can also evolve glyphosate resistance by elevating the production of the EPSP synthase. The elevated production of the EPSP synthase can be achieved in two ways: (i) by overexpressing the coding gene or (ii), by gene amplification (Fig. 1D; Chekan et al., 2016). Indeed, E. coli mutants with enhanced expression of the aroA EPSP synthase gene due to promoter-up mutations can easily be isolated by incubating the bacteria with sublethal amounts of glyphosate Stalker et al., 1985;Wicke et al., 2019). Moreover, the glyphosate sensitivity of bacteria can be simply reduced by overexpression the aroA gene using a multicopy plasmid (Rogers et al., 1983). The elevated cellular levels of the EPSP synthase reduce the toxicity of the herbicide by titrating the herbicide away, allowing a subfraction consisting of a non-inhibited enzyme to synthesize sufficient EPSP for amino acid biosynthesis (Fig. 1D). The enhanced cellular levels of the EPSP synthase due to selective gene amplification increase the resistance to glyphosate in the same way (Fig. 1D). This mechanism of glyphosate resistance can be observed in bacteria (Wicke et al., 2019) and even seems to be the dominant resistance mechanism in plants ( Gaines et al., 2010Gaines et al., , 2011Jugulam et al., 2014;Dillon et al., 2017).
To conclude, the evolution of glyphosate resistance by overexpressing a glyphosate-sensitive EPSP synthase is common in bacteria and plants. The development of glyphosate-resistant organisms due to the intensive application the herbicide in agriculture is a serious environmental issue (Schütte et al., 2017). For instance, the increased use of glyphosate may negatively affect the biodiversity (Motta et al., 2018). The current status of the glyphosate-resistant weeds, which are emerging worldwide, can be tracked at the International Herbicide-Resistant Weed Database (http://www.weedscience.org).
Glyphosate resistance due to degradation of the herbicide Glyphosate belongs to the large group of phosphonic acids or amino phosphonates (Studnik et al., 2015). As other phosphonates, glyphosate is resistant to chemical hydrolysis, thermal decomposition and photolysis due to a stable C P bond (Kononova and Nesmeyanova, 2002). Given the fact that large quantities of glyphosate have been used in agriculture, it is not surprising that microorganisms exist that can break down the herbicide (Borggaard and Gimsing, 2008; for excellent reviews see Singh and Walker, 2006;Hove-Jensen et al., 2014). Indeed, several Gram-negative and Gram-positive bacteria have been described to degrade glyphosate and/or use the herbicide as a source of phosphorous (Shinabarger and Braymer, 1986;Pipke et al., 1987a;1987b;Fitzgibbon and Braymer, 1988;Pipke and Amrhein, 1988a;1988b;Liu et al., 1991;Dick and Quinn, 1995;Penaloza-Vazquez et al., 1995;Castro Jr et al., 2007;Sviridov et al., 2011Sviridov et al., , 2012Sviridov et al., , 2015Kryuchkova et al., 2014;Yu et al., 2015). More recently, it has been demonstrated that glyphosate degradation seems to occur also in fungi and plants (see below; Duke, 2011;Rojano-Delgado et al., 2012;Vemanna et al., 2017;Pan et al., 2019). However, it is likely that the degradation pathways evolved not specifically to decompose glyphosate, but to protect the cells from naturally occurring phosphonates and to access the phosphorous present in these phosphonates or to exploit them as a phosphorous source during phosphate starvation (Metcalf and van der Donk, 2009). However, the pathways for breaking down glyphosate can also be considered as a mechanism conferring resistance to the herbicide.
So far, several glyphosate degradation pathways have been identified in microorganisms. For the details of the underlying biochemistry of the pathways, we would like to refer to the excellent review by Howe and coworkers (Hove-Jensen et al., 2014). The so-called C P-lyase pathway is widespread among bacteria. It is the only pathway that enables the bacteria to release phosphorous from glyphosate ( Fig. 2; Villarreal -Chiu et al., 2012;Hove-Jensen et al., 2014;Manav et al., 2018). Upon phosphate starvation, glyphosate is taken up via ATP binding cassette (ABC) transport systems in proteobacteria and cyanobacteria (Hove-Jensen et al., 2014). In the first step of the C P-lyase pathway, glyphosate is converted to 5-phosphoribosyl-1-diphosphate (PRPP) and N-methylglycine (sarcosine; Zeleznick et al., 1963;Shames et al., 1987;Murata et al., 1988;Zhang and van der Donk, 2012;Seweryn et al., 2015). N-Methylglycine may be converted to glycine via the transfer of the methyl group to tetrahydrofolic acid in C 1 metabolism (Hassan-Abdallah et al., 2005). Alternatively, N-methylglycine may be converted to H 2 O 2 , formaldehyde and glycine (Meskys et al., 2001). Independent of the fate of Nmethylglycine in metabolism, in both cases, the inorganic phosphate derived from glyphosate may be released from PRPP (Fig. 2).
The C P-lyase is a multi-enzyme complex that has been investigated for more than four decades in E coli and other bacteria (Kamat et al., 2011;Hove-Jensen et al., 2014). The enzyme complex consists of five subunits, which are involved in the transport, metabolism of the intermediates and products and in the preparation of the final cleavage via a radical reaction and post-catalytic metabolism (Kamat et al., 2011). The multi-enzyme complex is encoded by the phnCDEFGHIJKLMNOP operon whose expression is regulated by the phosphate regulon (Villarreal-Chiu et al., 2012). The C P-lyase cleaves a wide range of phosphonates including glyphosate. Interestingly, the final cleavage of the C P bond can lead to methane release (Kamat et al., 2013). Another pathway that is responsible for glyphosate degradation begins with the O 2 -dependent oxidation reaction, which can be catalysed by a glyphosate oxidoreductase (GOX) and glycine oxidase ( Fig. 2; Pollegioni et al., 2011). The oxidation of glyphosate yields in aminomethylphosphonate (AMPA) and glyoxylic acid. AMPA can be further converted by the enzymes of the C P-lyase pathway and phosphate may be released from PRPP (Fig. 2). The glyoxylate can be converted to CO 2 in the glyoxylic acid cycle. The pathway of glyphosate degradation via AMPA is less abundant in nature (Hove-Jensen et al., 2014)). However, it has been suggested that the prolonged exposure of environmental bacteria to glyphosate has led to alterations in one or more genes for improving their capacity for glyphosate cleavage (Hove-Jensen et al., 2014). Indeed, the bacterium Achromobacter sp. strain LBAA that was isolated from a glyphosate waste stream treatment facility, uses a GOX that confers glyphosate resistance to tobacco plants in a modified form (Barry and Kishore, 1995).
In recent years, other enzymes like phosphonoacetaldehyde hydrolases, phosphonoacetate hydrolases and phosphonopyruvate hydrolases that degrade phosphonates have been discovered (Huang et al., 2005;Agarwal et al., 2011;Villarreal-Chiu et al., 2012). These enzymes are also known collectively as 'phosphonatases', which often have a high substrate specificity. Another class of enzymes promotes the 2-oxoglutarate-dependent cleavage of phosphonates. This pathway seems to be closely associated with the phn operon encoding the C P-lyase activity (McSorley et al., 2012).
It has also been observed that glyphosate degradation can occur by other means. For instance, the igrA (increased glyphosate-resistant gene A) gene product of the Pseudomonas sp. strain PG2982 increases glyphosate resistance of the bacterium (Fitzgibbon andBraymer, 1988, 1990). A recent study uncovered that IgrA belongs to a superfamily of NADPH + H +dependent aldo-keto-reductases (AKR), which are present in a variety of organisms including fungi and plants (Jez and Penning, 2001;Vemanna et al., 2017;Pan et al., 2019). IgrA (AKR1) of the Pseudomonas sp. strain PG2982 as well as other AKR homologues bind glyphosate and convert it to AMPA and glyoxylate (Pan et al., 2019). Recently, it has been demonstrated that overproduction of IgrA also increases glyphosate resistance in transgenic rice (Fartyal et al., 2018). Moreover, enzymes that do not play an obvious role in breaking down glyphosate can be modified via directed evolution in such a way that they degrade the herbicide.
For instance, the B. subtilis glycine oxidase ThiO, which converts a variety of substrates (Nishiya and Imanaka, 1998;Job et al., 2002;Molla et al., 2003;Pedotti et al., 2009a;2009b), was also subjected to directed evolution for generating enzyme variants that are useful to obtain crops with enhanced glyphosate resistance (Pedotti et al., 2009a;2009b). The glycine oxidase is essential for thiamine biosynthesis in B. subtilis (Settembre et al., 2003). A variant of the glycine oxidase carrying three amino acid exchanges showed a 210-fold enhanced catalytic efficiency towards glyphosate (Pedotti et al., 2009a;2009b). The expression of a plantoptimized variant of the B. subtilis glycine oxidase indeed conferred glyphosate resistance to alfalfa (Medicago sp.; Nicolia et al., 2014). To conclude, native promiscuous enzymes of bacteria, fungi and plants can be employed to reduce the toxicity of glyphosate (Duke, 2011;Rojano-Delgado et al., 2012;Vemanna et al., 2017;Pan et al., 2019).

Glyphosate detoxification by covalent modification
Glyphosate resistance might also occur by other means (Pollegioni et al., 2011). It has been observed that Streptomyces and other species producing glufosinate (phosphinotricin), which inhibits the glutamine synthetase (Bayer et al., 1972;Fraser and Ridley, 1984;VanDrisse et al., 2016), protect themselves by converting the harmful substance to a non-inhibitory acetylated form ( Fig. 3A; Whermann et al., 1996). The detoxification of harmful substances by covalent modification is a very common strategy among antibiotic producers (i.e., Streptomyces species). Similar to the N-acetylated glufosinate, it has been demonstrated that also the N-acetylated form of glyphosate does not inhibit the EPSP synthase ( Fig. 3B; Castle et al., 2004). The idea that N-acetylation of glyphosate provides an alternative strategy to engineer glyphosate-resistant crops stimulated the search for a glyphosate-N-acetyltransferase (GAT) with high catalytic efficiency. For this purpose, several hundred Bacillus isolates were screened for their ability to generate N-acetylglyphosate (Castle et al., 2004). The screen revealed that the acetyltransferase from Bacillus licheniformis showed the highest GAT activity. After identifying the gat gene, the GAT was subjected to directed evolution to improve the catalytic efficiency of the enzyme, yielding a variety of variants (Castle et al., 2004;Siehl et al., 2005Siehl et al., , 2007. One of the GAT variants was structurally analysed to get insights into the reaction mechanism of the enzyme (Siehl et al., 2007;Pollegioni et al., 2011). Recently, glyphosate acetylation was reported to occur in Achromobacter sp. Kg 16 (Shushkova et al., 2016). Even though the natural substrate of the GAT homologues is unknown, the work performed on the enzyme illustrates the power of directed evolution to obtain biocatalysts having the desired properties. Since gat homologues are found in many Bacillus species, the encoded enzymes certainly pIay important roles in detoxifying harmful substances produced by the own cell or other bacteria. The GAT encoded by the gat gene, which was subjected to directed evolution, did not show evidence of allergenicity or toxicity when orally applied to mice (Delaney et al., 2008). The GAT could therefore replace glyphosate-insensitive EPSP synthase variants or both, the GAT and insensitive EPSP synthase variants, could be used to enhance the glyphosate resistance of crops (Castle et al., 2004;Guo et al., 2015;Stokstad, 2004). In the past years, the GAT has indeed been successfully introduced into tobacco and cotton to enhance their resistance to glyphosate (Liu et al., 2015;Liang et al., 2017). Furthermore, the gat gene has been used as a selection marker to genetically engineer bacteria (Norris et al., 2009).
The resistance to glyphosate can also be enhanced by other enzymes that covalently modify the herbicide. For instance, the hygromycin phosphotranferases Hph and GlpA from E. coli and Burkholderia pseudomallei respectively, can phosphorylate glyphosate, thereby conferring glyphosate resistance ( Fig. 3C; Rao et al., 1983;Penaloza-Vazquez et al., 1995). To conclude, various enzymes have been described that confer glyphosate resistance by covalently modifying the herbicide. Since environmental bacteria often come into contact with glyphosate, certainly other enzymes will be discovered that have gained the ability to detoxify the herbicide.
Glyphosate resistance due to decreased uptake and increased export Although glyphosate has been used in agriculture for almost 50 years, it has long been unknown how the herbicide enters a living cell. Recently, it has been observed that the difference in the resistance to glyphosate in different strains of baker's yeast Saccharomyces cerevisiae is due to genetic variations in the genes encoding the transporters Dip5 and Pdr5 ( Fig. 4; Rong-Mullins et al., 2017). Dip5 is a transporter with low substrate specificity that transports glutamate, aspartate, glutamine, asparagine, serine, alanine and glycine (Regenberg et al., 1998(Regenberg et al., , 1999. Dip5 belongs to the amino acid-polyamine-organocation (APC) superfamily of transport proteins (Saier, 2000a). Pdr5 is an ABC efflux transporter that is involved in the detoxification of the yeast cell during exponential growth phase by exporting a variety of substrates (Rogers et al., 2001;Mamnun et al., 2004;Golin et al., 2007). The inactivation of the Dip5 gene increased glyphosate resistance. By contrast, the deletion of the Pdr5 encoding gene decreased glyphosate resistance. These observations suggest that Dip5 and Pdr5 are indeed involved in the uptake and export respectively, of glyphosate in yeast (Rogers et al., 2001). The study by Rong-Mullins and colleagues (.,2017) provides initial evidence for the involvement of amino acid transporters in translocating glyphosate across the membranes in a unicellular eukaryotic organism. It has also been reported that amino acid transporters could be involved in the transport of glyphosate across mammalian epithelial tissues (Xu et al., 2016). Recently, it has been demonstrated that the enhanced export of glyphosate via fungal and bacterial transport proteins reduces the cellular toxicity of the herbicide. For instance, the overproduction of the uncharacterized membrane proteins MFS40 and YhhS from Aspergillus oryzae and E. coli respectively, share similarity with the major facilitator secondary transporter superfamily, enhances glyphosate resistance of E. coli ( Fig. 4; Staub et al., 2012;Tao et al., 2017). However, the precise functions of the transporters remain to be determined. Experimental evidence suggests that YhhS is involved in the efflux of pentose sugars (Koita and Rao, 2012).
Altered glyphosate transport has also been recognized as a mechanism of herbicide resistance in plants (Ge et al., 2010;Chekan et al., 2016). It has been observed that in glyphosate-resistant horseweed Conyza canadensis, the herbicide was rapidly translocated from the cytoplasm to the vacuoles (Ge et al., 2010). The molecular basis for the transport of glyphosate has been attributed to ABC transporters and a tonoplast intrinsic protein (TIP; Peng et al., 2010;Yuan et al., 2010;Nol et al., 2012). It has been suggested that TIPs regulate water transport across the membrane of the vacuole (Chrispeels and Maurel, 1994;Maurel, 1997). Currently, it is hypothesized that the TIP facilitates water influx into the cell to reduce the cellular concentration of the herbicide and that the ABC transporters translocate the herbicide into the vacuole in glyphosate-resistant horseweed (Nol et al., 2012). Thus, like in yeast, glyphosate resistance can also be mediated by altered transport of the herbicide in plants. However, the molecular details how the transport proteins mediate glyphosate resistance in plants remain to be elucidated.
A more comprehensive study about the proteins mediating glyphosate transport in bacteria was recently performed with B. subtilis (Wicke et al., 2019). Previously, it has been reported that the addition of glutamate, glutamine, proline and arginine decreases the glyphosate sensitivity of B. subtilis (Fischer et al., 1986). The amino acids probably inhibit the glyphosate uptake, thereby preventing an increase of the cellular concentration to toxic levels. In the same study, it was mentioned that the growth rate of B. subtilis in medium containing toxic amounts of glyphosate (2.5 mM), rapidly increases after prolonged incubation of the bacterial cultures, which could be due to the emergence of glyphosate-resistant B. subtilis mutants (Fischer et al., 1986). Indeed, the study by Wicke and co-workers (Wicke et al., 2019) revealed that B. subtilis rapidly evolves glyphosate resistance by acquiring loss-of-function mutations in the gltT gene enoding the high-affinity glutamate/asparate symporter GltT ( Fig. 4; Zaprasis et al., 2015). Moreover, further adaptation of a gltT mutant led to the inactivation of the gltP gene encoding the low-affinity glutamate transporter GltP (Fig. 4;Tolner et al., 1995;Wicke et al., 2019). Additional genetic as well as metabolome analyses confirmed that GltT is the major glyphosate transporter in B. subtilis (Wicke et al., 2019). Furthermore, the glutamate transporter GltP of E. coli, which shares 52% overall sequence identity with GltT from B. subtilis, was also shown to mediate glyphosate transport. Both GltP and GltT belong to the dicarboxylate/ amino acid:cation (Na + or H + ) symporter family (Slotboom et al., 1999;Saier, 2000b). It is interesting to note that the suppressor screen by Wicke and coworkers (Wicke et al., 2019) did not led to the inactivation of the aimA gene that was recently shown to encode the major glutamate transporter AimA in B. subtilis (Krüger et al., 2021). It is tempting to speculate that AimA has a higher specificity for transporting glutamate as compared with GltT and GltP. Indeed, GltT also mediates transport of the herbicide glufosinate, which inhibits the glutamine synthetase ( Fig. 4; Bayer et al., 1972;Fraser and Ridley, 1984;Wicke et al., 2019). To conclude, in prokaryotic as well as in eukaryotic systems glyphosate and structurally similar substances may enter the cell and organelles via amino acid transporters (Cherepenko and Karpenko, 1999;Wicke et al., 2019). In fact, it is very common among bacteria to inactivate transporter genes upon exposure to toxic amino acid analogues (Zaprasis et al., 2014;Commichau et al., 2015).

Conclusions and outlook
The variety of glyphosate resistance mechanisms that have been identified and characterized illustrate the ability of bacteria to adapt to anthropogenic substances due to genomic alterations. The characterization of glyphosate-resistant bacteria and the understanding of the molecular mechanism underlying glyphosate resistance paved the way for genetically engineering plants Fig. 4. Glyphosate resistance mediated by reduced uptake and export. GltT and GltP, Bacillus subtilis glutamate transporters; YhhS and MFS40, major facilitator secondary transporters from Escherichia coli and Aspergillus oryzae respectively; Pdr5 and Dip5 are amino acid and ABC efflux transporters respectively, from Saccharomyces cerevisiae. Glufosinate and glyphosate inhibit the glutamine synthetase (GlnA) and the EPSP synthase AroE respectively.
[Color figure can be viewed at wileyonlinelibrary.com] tolerating increased levels of the herbicide. These studies certainly contributed to the fact that glyphosate has become the most widely used herbicide in global agriculture. However, intensive application of glyphosate in agriculture is a serious environmental issue because it may negatively affect the biodiversity in the soil and in the insect gut microbiota. In addition to glyphosate, other substances have been identified and characterized in recent years that selectively inhibit the shikimate pathway ( Fig. 1A; Han et al., 2006;Zhu et al., 2018;Brilisauer et al., 2019;Neetu et al., 2020). A very promising substance is the deoxy sugar 7-deoxy-sedoheptulose (7dSH) that is produced by the cyanobacterium Synechococcus elongatus and inhibits the dehydroquinate synthase ( Fig. 1A; Brilisauer et al., 2019). It could be shown that 7dSH inhibits the growth of Arapidopsis thaliana seedlings more strongly that glyphosate when applied at the same substrate concentration (Brilisauer et al., 2019). However, even if 7dSH could be a potential substitute for glyphosate, the deoxy sugar has not yet been classified with regard to its effects on soil microbiota and other organisms. Moreover, in the future it will be interesting to evaluate whether a glyphosate-resistant B. subtilis gltT mutant lacking the GltT glyphosate transporter is suitable to identify glyphosate uptake systems from plants (Wicke et al., 2019). To address this issue, plant-derived cDNA libraries could be used to screen for B. subtilis gltT transformants that are unable to grow with glyphosate due to restored uptake of the herbicide. Comai, L., Sen, L.C., and Stalker, D.M. (1983)