Global lessons from antibiotic resistance: metformin-hydrolyzing genes in transposable elements, a new threat for type II diabetic patients?

60 19 Figures: 2 20 Table: 1 21 Content type: Analysis 22 23 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 2

24 Metformin drug, widely used to treat type II diabetic patients, is a major pharmaceutical pollutant 25 of wastewater and rivers. This environmental exposure has driven the evolution of bacteria, such 26 as Aminobacter and Pseudomonas, to degrade metformin via a Ni²⁺ -dependent metformin 27 hydrolase complex (MfmAB). Here we decipher the mechanism of acquisition and horizontal 28 transfer of the mfmAB genes, initially mobilized from Aminobacter chromosomes to Pseudomonas 29 conjugative plasmids via common transposable elements ( IS1182 and IS3/IS6 elements ) in 30 composite transposons , carrying also other genes involved in guanylurea and biguanides 31 degradation ( guuH and bguH). These mobile elements, historically involved in acquisition of 32 antibiotic-resistant genes from the environment before clonal expansion in clinical settings , now 33 threaten to co-select for both metformin-degrading and antibiotic resistance genes in contaminated 34 waters. This represents a global threat for diabetic patients with concurrent infections that should 35 be urgently added in the roadmap of research in the context of One Health. 36 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 3

37 The manufacturing and widespread use of pharmaceutical products during the 20th century 38 have contributed to an increasing diversity of active pharmaceutical ingredients and their 39 metabolites in the environment1. This extensive release of bioactive compounds into ecosystems 40 can disrupt ecological functions and pose a significant threat to human and ecosystems health1–3. 41 Of these compounds, metformin (1,1-dimethylbiguanide), has emerged over the last few decades 42 as a globally dominant pharmaceutical pollutant due to its extensive use and poor biodegradability 43 in wastewater treatment plants (WWTPs) and aquatic environments1,4–6. 44 Originally synthesized in the 1920s, metformin was introduced as a first-line treatment for 45 type II diabetes in France (1959), China (1994), and the USA (1995)7–9. Since then, this drug has 46 become one of the most commonly used pharmaceuticals globally, with over 150 million 47 prescriptions worldwide in 202210. More recently, metformin has been investigated for its potential 48 use in the treatment of cancer, endocrine disorders , and obesity, with expectations for its broader 49 use in the future11. The direct mode of action of metformin remains incompletely characterized11. 50 Evidence now supports the fact that metformin alters the human gut microbiota by mediating its 51 therapeutic effects 12–15. Intravenous administration of metformin to type II diabetes patients, 52 bypassing the human gut, showed limited to no therapeutic efficacy , further emphasizing the 53 important interaction with gut microbiota in therapeutic effects16. After a single oral dose of 0.5 g 54 to 2 g per day, over 80% of metformin is excreted unchanged in urine and feces 17,18. Due to its 55 limited hepatic metabolism and extensive excretion, metformin has become one of the most 56 predominant micropollutants in WWTPs1,4. 57 Over the past decade, microbial metabolism in the activated sludge of WWTPs has been 58 identified as a major driver of metformin biodegradation5. The primary transformation product of 59 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 4 metformin in WWTPs, guanylurea, is known to be environment ally toxic, causing reproductive 60 impairment, reduced larval survival, and neurobehavioral alterations in fish after chronic 61 exposure3. Bioremediation strategies targeting metformin and its metabolites have therefore been 62 proposed to mitigate environmental contamination. Recent independent studies have identified 63 Aminobacter sp. and Pseudomonas sp. strains isolated from activated sludge samples in USA, 64 France, and China in 2021 capable of degrading metformin and utilizing it as their sole source of 65 carbon and nitrogen for growth 19–22. These strains encode a Ni 2+-dependent heterohexameric 66 enzyme composed of two subunits, M fmA and MfmB, which hydrolyses metformin into 67 guanylurea and dimethylamine 23. This enzyme belongs to the ureohydrolase s uperfamily and 68 shares similarity with arginase and agmatinase , highly conserved bacterial enzymes recently 69 identified as metformin targets18–20,22. Agmatinase catalyzes the conversion of agmatine into 70 putrescine, a key precursor in polyamine biosynthesis, essential for bacterial growth , nitrogen 71 metabolism, and stress adaptation18,24. 72 Intriguingly, while mfmAB genes were reported to be chromosomally encoded in 73 Aminobacter spp. strains, they are carried on conjugative plasmids in another Aminobacter spp. 74 and in Pseudomonas spp. strains in the USA, suggesting their potential horizontal genetic transfer 75 (HGT) 19–21. The acquisition of mfmAB via mobile genetic elements (MGEs) (i.e conjugative 76 plasmids) could significantly accelerate the global spread of metformin degradation capabilities 77 with implications for both environmental pollution and potentially impacting human metformin 78 treatment. Previous studies have focused mainly on the origin, role and biological activity of this 79 enzyme but the exact genetic support and movement of such enzymes has not been studied so far. 80 This study aims to elucidate the evolutionary origins of the mfmAB gene cluster and uncover the 81 genetic mechanisms driving its HGT along with other unexpected genes across bacterial species 82 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 5 using advanced computational methods . In the context of One Health, assessing the health risk 83 associated with the spread of such mobile “metformin resistance ” genes, by analogy with the 84 current situation of antibiotic resistance emergence and spread in humans, is critical , given the 85 increasing number of patients worldwide being treated with metformin for type II diabetes. 86

87 Identification of bacterial genomes containing mfmAB and homologous sequences 88 To investigate the distribution and evolutionary origins of mfmAB, the mfmA (1,073 bp) 89 and mfmB (1,046 bp) sequences from Aminobacter sp NyZ550 (NCBI accessions WAX94658.1, 90 WAX94657.1) were used as queries for BLASTn and BLASTp searches against the NCBI non-91 redundant database in April 2025 . To identify potential functional homologs across genera, 92 searches were conducted with a threshold of > 90% identity and > 90% coverage . For each 93 bacterial strain carrying mfmAB homologs, relevant metadata were collected, including the 94 geographical location, date , and source of isolation and previously reported metformin-95 hydrolyzing capacity. In addition, whole-genome phylogenetic analysis of all Aminobacter spp. in 96 the NCBI database was performed using TYGS25. 97 Comparative genomic analysis of metformin-hydrolyzing strains 98 To elucidate the genetic context and potential transfer mechanisms of mfmAB across 99 bacterial species, only genomes of strains with confirmed metformin hydrolase activity were 100 selected for comparative genomic analysis. Plasmid classification and r eplication origins were 101 determined using MOB-Typer (v.3.0.3) and Ori-Finder, which identify relaxase types and predict 102 plasmid mobility26,27. Sequence homology comparisons were performed using the Blastn NCBI 103 tool. 104 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 6 To investigate structural variations and the potential role of MGEs in the horizontal transfer 105 of mfmAB across bacterial species, i nsertion sequences and other MGEs associated with mfmAB 106 were identified using ISfinder and mobileOG-db28. Genome alignment and synteny analysis were 107 conducted using Easyfig (v.2.2.2) to visualize the genomic context and compare the organization 108 of mfmAB-harboring regions across different bacterial genomes29. 109 Illustrations in this manuscript were initially created using BioRender 110 (https://www.biorender.com/) to conceptualize the figures under a full publication license . All 111 figures were subsequently redrawn manually using Microsoft PowerPoint. 112

113 Genomic identification and evolutionary origin of mfmAB in metformin-hydrolyzing 114 strains 115 BLASTn analysis of mfmAB sequences against the NCBI database identified 15 bacterial 116 genomes carrying mfmAB homologs with ≥ 99% coverage and 90%-99% identity, including three 117 Pseudmonas and 12 Aminobacter strains (Table 1). Each genome harbored a single copy of the 118 mfmAB homolog within a highly conserved eight-gene cluster spanning approximately 8.2 kb. This 119 cluster encodes a HupE/UreJ family protein, two arginase and agmatinase family proteins (MfmB 120 and MfmA), two hydrogenase nickel incorporation -associated proteins (hypA and hypB), a 121 TetR/AcrR family transcriptional regulator, a cystosine permease (codB) , and a second XRE 122 family transcriptional regulator (Fig. 1a). Notably, 11 of the 12 Aminobacter strains carried the 123 mfmAB homolog on their chromosome, suggesting that the active metformin-hydrolyzing mfmAB 124 likely originated from Aminobacter spp. 125 A total of six strains, including three Aminobacter spp. (NyZ550, MD1, and MET-1) and 126 three Pseudomonas sp. strains (MET -2, KHPS1, and KHPS2) , were previously reported to 127 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 7 hydrolyze metformin through enzymatic assays19–22 (Table 1). These isolates were recovered from 128 WWTPs in France in 2012, in China and the USA in 202119–22. Interestingly, a genome-based 129 phylogenomic analysis indicated that these metformin-hydrolyzing Aminobacter strains, isolated 130 from three different continents, belong to three different lineages (Supplementary Fig. 1). 131 Furthermore, localization analysis of mfmAB revealed that in two Aminobacter strains 132 (NyZ550 from China and MD1 from France), mfmAB was chromosomally encoded, whereas in 133 four strains from the USA ( Aminobacter MET-1 and three Pseudomonas strains), mfmAB was 134 located on plasmids, indicating evidence of the HGT of mfmAB across bacterial species , 135 particularly in the USA (Table 1). 136 Transposition of mfmAB from the chromosome to a plasmid of Aminobacter sp. 137 The origin of replication repC of Aminobacter sp. plasmid pMET-1 containing mfmAB has 138 82% identity and 100% coverage with a plasmid from Rhizobiaceae bacterium strain isolated in 139 an activated sludge from Hong Kong (GCA_023953835.1). Comparative genomic alignment 140 between pMET -1 and the chromosomes of Aminobacter sp. MD1 and NyZ550, revealed a 141 homologous region of approximately 56.86 kb containing mfmAB cluster (Fig. 1a). Notably, this 142 region on pMET -1 was flanked by a complete IS1182- family insertion sequence ( IS). Further 143 examination revealed the presence of two inverted repeats left (IRL) and right (IRR) associated 144 with IS1182, along with an additional IRL structure located at the end of the homologous region 145 (Fig. 1a). Importantly, a 5 bp-directed repeat sequence indicative of a target site duplication (TSD) 146 was observed surrounding this inserted region, providing strong evidence of a recent transposition 147 of mfmAB from the chromosome to plasmid pMET-1 of Aminobacter sp. mediated by IS1182 (Fig. 148 1a, b ). Together, these findings support that the transposition of the mfmAB is rather a recent 149 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 8 evolutionary event that is compatible with the recent selective pressure with metformin by human 150 prescription in patients with type II diabetes. 151 Horizontal transfer of an mfmAB-carrying Aminobacter plasmid followed by 152 transposition into Pseudomonas plasmids 153 All four metformin-hydrolyzing strains from Minnesota, USA, harbored an mfmAB cluster 154 on plasmids encoding a complete set of conjugation-associated genes. These included a type IV 155 secretion system (T4SS) with virB4 (pMET-1, pMET -2, pKHPS2) and tra cluster genes 156 (pKHPS1), as well as an origin of transfer (oriT), relaxase, and transfer coupling protein (Fig. 1a). 157 These genetic features suggest that these plasmids are mobilizable and capable of conjugative 158 transfer. However, comparative plasmid analysis revealed that pMET-1 in Aminobacter sp. strain 159 is distinct from the plasmids in the three Pseudomonas sp. strains (Fig. 1a). The alignment 160 sequence of pMET-1 with those of Pseudomonas sp. strains, identified a conserved 8.2 kb region 161 containing mfmAB (Fig. 1a). At the distal end of this region, various complete ISs were observed, 162 including an IS3 element in pMET -2 and pKHPS2, and an IS6 element in pKHPS1 (Fig. 1a). 163 Importantly, each Pseudomonas sp. strain harbored only one plasmid, suggesting that the mfmAB 164 cluster in three Pseudomonas sp. strains was not acquired via direct conjugation and stable 165 maintenance of an Aminobacter conjugative plasmid such as pMET-1. Instead, the transfer likely 166 occurred through an initial conjugative transfer of the mfmAB-carrying plasmid from Aminobacter 167 sp. to the Pseudomonas sp. strains, followed by the IS-mediated transposition of the mfmAB cluster 168 into resident Pseudomonas plasmids, with a subsequent loss of the original donor plasmid (Fig. 169 1b). 170 Interestingly, in three Pseudomonas plasmids, this cluster of 8.2 kb mfmAB-containing 171 region was found within a larger 23 kb composite transposon flanked by IS3000 elements, 172 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 9 belonging to the Tn3-family transposase that contains a cluster of genes encoding polyamine ABC 173 transporter substrate-binding proteins and a partial spermidine/putrescine ABC transporter potF 174 (Fig. 1a). A BLASTn search for these transporter genes revealed high sequence identity (88%) and 175 coverage (98%) with homologous regions in the chromosome of Pseudomonas sp. strains, 176 suggesting that this cluster was initially chromosomally encoded before being transposed onto 177 plasmids (Fig. 1b) . The truncation of potF adjacent to the mfmAB cluster indicates that the 178 transposition of mfmAB cluster, mediated by either IS3 or IS6 elements, occurred after the initial 179 acquisition of Tn3-family transposon containing these transporter genes (Fig. 1b). 180 Plasmid-mediated metformin/biguanide/guanylurea degrading in Pseudomonas sp. 181 strains 182 Comparative sequence analysis of the three Pseudomonas plasmids showed that pMET-2 183 of Ectopseudomonas mendocina and pKHPS2 of Pseudomonas hydrolytica revealed a high 184 similarity (99.99% identity and 92% coverage), strongly supporting the occurrence of conjugative 185 transfer of mfmAB-carrying plasmid between these two Pseudomonas sp. strains. In contrast, 186 pKHPS1 exhibited significant genetic divergence but shared two distinct homologous regions with 187 pMET-2 and pKHPS2 (Fig. 1a). In addition to Tn3-like transposon containing mfmAB, these 188 plasmids harbored a second 8 kb-conserved region, that is flanked by IS66, encoding guanylurea 189 hydrolase (guuH) and biguanide aminohydrolase (bguH), enzymes known to degrade guanylurea 190 and biguanide, respectively20,30 (Fig. 1a and Fig. 2). This cluster was rather transposed by the IS66 191 element, known to be a transporter -related insertion sequence between different plasmids of 192 Pseudomonas sp. strains31 (Fig. 1a, b). 193

194 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 10 The Red Queen runs on metformin: plasmid evolution and bacterial adaptation to 195 anthropogenic pollutants 196 As one of the most widely used medications, metformin has become one of the most 197 frequently detected pharmaceuticals in wastewater, with concentrations reaching up to 16 000 ng/L 198 in the USA1,4,5,7,20. This persistent environmental presence is believed to impose selective pressure 199 on microbial communities, facilitating HGT and the evolution of metformin -degrading 200 capabilities, such as those conferred by the mfmAB gene cluster32 (Fig. 1a, b and Fig. 2). The recent 201 and almost concomitant discovery of environmental bacteria capable of hydrolyzing metformin on 202 three different continents (China, Europe, and the America) is intriguing and we hypothesized that 203 this was not a coincidence but rather a widespread phenomenon on a global scale due to the recent 204 epidemic of diabetic patients and the massive pollution of water environments by this drug (Fig 205 2). Previous studies hypothesized a chromosomal origin for mfmAB gene cluster in Aminobacter 206 in the environment, none comprehensively characterized its genetic context or clarify the precise 207 mechanism/scenario of its mobilization from the chromosome of Aminobacter to various plasmids 208 using advanced computational method s20,23,33. Here we confirm the Aminobacter chromosomal 209 origin of these genes and we show that these genes evolved convergently in three independent 210 lineages (Asia, Europe, and North America), supporting Darwinian principles of adaptation (Fig. 211 1a, b and Supplementary Fig. 1). A precise comparative analysis of the available genomes allowed 212 us to reconstruct the scenario of mobilization of this cluster in the bacterial species analyzed (Fig. 213 1b). The cluster was firstly mobilized from the chromosome to a conjugative plasmid in 214 Aminobacter by transposition via a common IS1182. Conjugation of this plasmid in to 215 Pseudomonas resulted in this transposon jumping to another Pseudomonas conjugative plasmid, 216 on at least two independent occasion s via an IS3 or an IS6 element , into larger composite 217 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 11 transposons that already contain an additional cluster of four genes associated with putrescine 218 uptake34,35 resulting in the truncation of adjacent genes (Fig. 1a, b). Finally, a second 8 kb-cluster 219 encoding for guanylurea hydrolase ( guuH) and biguanide aminohydrolase ( bguH), enzymes 220 known to degrade guanylurea and biguanide, was also transposed in these plasmids using an IS66 221 element20,30 (Fig. 1a, b). This work highlights the very recent and dynamic process of evolution of 222 these plasmids in the context of environmental pollution by metformin and the constant and 223 formidable adaptation of bacteria to survive in rapidly changing environments polluted by 224 anthropogenic activities. This evolutionary trajectory aligns with the Red Queen hypothesis, which 225 suggests that organisms must continuously evolve not just to gain advantage, but simply to survive 226 under environmental pressures. Although environmental pollution by metformin has persisted for 227 decades, m etformin hydrolase was discovered only recently in 20211,4–6. This suggests that 228 bacteria initially endured metformin exposure without this enzyme, likely relying on alternative 229 resistance mechanisms. The transposable elements characterized in our study may represent a key 230 evolutionary adaptation, enabling bacterial survival under prolonged metformin selection pressure. 231 Since metformin inhibits agmatinase , which is a key enzyme to produce putrescine, 232 bacteria were able to counteract this inhibition by expressing a putrescine uptake transporter , 233 enabling direct environmental acquisition of this nutrient 18,34 (Fig. 2). The transposition of the 234 mfmAB cluster into this initial transposon and truncation of this transporter was not deleterious for 235 the bacteria because the hydrolysis of metformin restores the biosynthesis of putrescine by the 236 agmatinase. However, metformin hydrolase produces guanylurea that is thought to be a toxic 237 compound for bacteria and detoxification of this compound could take place due to the guanylurea 238 hydrolase guuH present in the second transposable element that makes it possible to further use 239 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 12 guanylurea as a source of carbon and nitrogen20 (Fig. 2). Once again, transposition proves critical 240 for bacterial survival and adaptation to shifting nutrient conditions. 241 Natural occurrence of metformin analogs and global lessons from antibiotic 242 resistance dissemination 243 Metformin is an alkaloid biguanide with two methyl substituents, forming 244 dimethylbiguanide. It is derived from galegine, a natural analog of metformin isolated from the 245 traditional herb Galega officinalis7,8. The synthesis of metformin simply relies on the ability of 246 arginine to transfer its amidine group to a precursor of galegine by a transamidination reaction 36. 247 Interestingly, a vast selection of guanidine derivatives can be found in nature and may explain the 248 ancient origin of the “metformin hydrolase” cluster8,36. Indeed, secondary metabolites bearing a 249 guanidine moiety have been isolated from several bacterial species such as Streptomyces spp., 250 Amycolatopsis orientalis , and Xenorhabdus strains36. Non -ribosomal peptides, many of which 251 with agmatine or a modified arginine residue have been found in Cyanobacteria sp., such as 252 Microcystis sp36. Guanidine derivates have also been isolated from the marine sponge Crambe 253 crambe and plants such as Solanum cernuum and marine nudibranch Actinocyclus papillatus 254 (Mollusca)36. An imidazole -bis-guanidine alkaloid has been isolated from zoanthid Epizoanthus 255 illoricatus36. The structural similarity, leading to prolonged bacterial exposure to guanidine 256 analogs in natural environments, may have driven the evolutionary transfer of metformin -257 hydrolyzing enzymes. Furthermore, the widespread presence of these analogs could explain the 258 independent emergence of an arginase –agmatinase-related gene pair identified in Aminobacter 259 strains from three different continents, likely as a result of convergent evolution 19–21. The 260 widespread distribution of agmatinase family enzymes across diverse bacterial phyla suggests an 261 ancient evolutionary connection between bacterial metabolism s and guanidine derivatives 23. 262 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 13 Notably, the metformin metabolism may have originated from ancestral pathways linked to 263 agmatinase activity18,37. The agmatinase homologs might have played a potential role as a reservoir 264 for novel metformin-metabolizing enzymes. The evolutionary scenario of mfmAB in metformin-265 hydrolyzing strains can be conceptualized through a framework that mirrors the emergence and 266 dissemination of antimicrobial resistance genes 38–42 (Fig. 2). The recently described plasmid-267 mediated mobile colistin -resistant (mcr-1) gene that threatens the clinical efficacy of colistin in 268 patients infected by multidrug -resistant bacteria is the best example of the evolution of an 269 environmental ancestral gene to a specialized enzyme mobilized via transposition in clinical 270 bacteria within a very efficient and rapid turnaround time38,43. Since its first report in Escherichia 271 coli isolates from pigs in China, the mcr-1 gene has been isolated worldwide in humans and animal 272 reservoirs, and in environmental sources such as wastewater, soil, and rivers38,44,45. The widespread 273 use of colistin in agriculture has facilitated the rapid and uncontrolled dissemination of plasmid -274 mediated mcr-1 genes among bacteria 43. Global lessons from this example have shown that the 275 ancestor of this gene has evolved from different bacterial species in the aquatic environment as a 276 convergent evolution according to Darwin’s theory before being mobilized via transposition events 277 in conjugative plasmids in animals. Colistin selective pressure in animal husbandry then amplified 278 the spread in animals that later transferred to clinically relevant bacteria , and that is now 279 pandemic46. A similar concern is emerging with antifungal agents, as the extensive agricultural 280 use of azole -based fungicides is now a major driver of resistance in Aspergillus species and 281 probably the emergence of Candida auris 47. Here we show that various IS elements and 282 transposition events from the chromosome of an environmental progenitor to conjugative plasmids 283 of metformin hydrolases are exactly the same as those recently described for the antibiotic 284 resistance global crisis 42,48. Indeed, IS1182, IS3 and IS6 are the most frequent mechanism s 285 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 14 involved in the transposition of current pandemic antibiotic-resistant genes (vancomycin resistance 286 operon vanA, mcr-1, extended -spectrum beta -lactamases, carbapenemases, aminoglycosides 287 modifying enzymes, etc.) frequently organized into plenty multiple different multidrug-resistant 288 composite transposon s within various conjugative plasmids 41,42,48–50. The high mobility and 289 rapidity with which these MGEs reorganize themselves according to the selection pressures 290 exerted in the ecosystems in which bacteria evolve, in particular anthropogenic environments 291 polluted simultaneously by antibiotics and metformin and its metabolites, pose a major risk to 292 humanity of transfer of this transposon of enzymatic metformin “resistance” genes onto multidrug-293 resistant transposons and/or conjugative plasmids in clinical bacteria (Fig. 1c). 294 Potential clinical impact of metformin -hydrolyzing enzymes and future therapeutic 295 strategies 296 To date, there is no evidence of the presence of this transposon in the human or animal 297 digestive microbiota, which means that transfer from the environment has not yet taken place or 298 has not yet been reported because it has not been specifically searched for. However , clinical 299 reports of metformin treatment failure in patients with type II diabetes along with the expected 300 increase in diabetic patients in the years to come raises concerns about the possible HGT of mfmAB 301 from environmental bacteria to human -associated microbes potentially compromising the 302 therapeutic efficacy of metformin38,40,51 (Fig. 1c). Interestingly, it has been recently discovered 303 that specific enzymes encoded by the human gut and oral microbiome can inactivate another 304 antidiabetic drug, acarbose, through phosphorylation, suggesting a specific form of microbiome -305 mediated resistance that may impact both microbial competition and the drug’s clinical 306 effectiveness52. Global lessons from antibiotic resistance may be helpful to start designing clinical 307 studies to look for potential synergistic effect of metformin. Notably, sulfonylurea agents such as 308 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 15 gliclazide, which also exhibit antimicrobial properties, have demonstrated therapeutic benefit s 309 when used in combination with metformin in patients with suboptimal glycemic control compared 310 to metformin alone53. Similarly, we believe that it is also timely to encourage research looking for 311 metforminase inhibitors similarly to beta -lactamase inhibitors. Metformin hydrolase is a Ni2+ 312 dependent enzyme that belongs to a large superfamily of metalloproteins with diverse biological 313 functions distributed in all domains of life including archaea, bacteria, CPR, and humans54. In the 314 context of antibiotic resistance , the development of metallobetalactamase inhibitors remains 315 challenging, however novel boronate -based inhibitors such as taniborbactam was shown to be 316 effective against such enzymes 55. Collaborative efforts with clinical microbiologist s and 317 endocrinologists should be undertaken to promote the development of boronate -based inhibitors 318 and/or other metalloenzyme inhibitors in this context. Similarly, the inactivation of such 319 metalloenzymes by metal chelators that remove nickel ions or metal -based drugs that replace 320 nickel with another metal are other avenues of research that should be considered in the future. 321 In conclusion, our study shows that currently the mobilization of the transposon carrying 322 mfmAB into plasmids of Aminobacter and Pseudomonas has been reported only in environmental 323 water in Minnesota, USA, suggesting that it would be opportune to develop global strategies and 324 take decisions requiring evidence-based risk assessments. The potential co-selection and formation 325 of composite transposon s carrying all metformin-hydrolyzing, antibiotic and heavy metals 326 resistance genes, that could later spread to clinical settings should urgently be monitored and 327 actively surveyed around the world (Fig. 2) 56,57. Lessons should be learned from the emergence 328 and spread of antibiotic resistance in the context of One Health to limit such scenario. Finally, 329 many other pharmaceutical agents widely used in humans for different diseases are now known to 330 be major pollutants and to be degraded by bacterial communities in the environment ( such as 331 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 16 paracetamol, ibuprofen, carbamazepine )58–60 (Fig. 2). Therefore, we strongly believe that a new 332 research field, that we propose to call “PharmacoMicrobioResistance ”, (i.e. microbial resistance 333 to non-antibiotic human-targeted drugs) should be added, to encompass pharmaceutical agents 334 other than antibiotics in the current research proposals in the context of One Health. 335 Declarations 336 Competing Interests: The authors declare that they have no competing interests. 337 Ethical approval: Not required. 338 Funding 339 This work was supported by the French Government under the “Investissements d’avenir” 340 (Investments for the Future) programme managed by the Agence Nationale de la Recherche (ANR, 341 fr: National Agency for Research), (reference: Méditerranée Infection 10-IAHU-03). 342

We are very grateful to the IHU Méditerranée Infection, Marseille, France 343 for financial support. 344 345 346 347 348 349 350 351 352 353 354 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 17

355 356 1. Wilkinson, J. L. et al. Pharmaceutical pollution of the world’s rivers. Proceedings of the 357 National Academy of Sciences 119, (2022). 358 2. Wellington, E. M. et al. The role of the natural environment in the emergence of antibiotic 359 resistance in Gram-negative bacteria. Lancet Infect Dis 13, 155–165 (2013). 360 3. Elizalde-Velázquez, G. A., Herrera-Vázquez, S. E., Gómez-Oliván, L. M. & García-361 Medina, S. Health impact assessment after Danio rerio long-term exposure to 362 environmentally relevant concentrations of metformin and guanylurea. Chemosphere 341, 363 140070 (2023). 364 4. Bradley, P. M. et al. Metformin and other pharmaceuticals widespread in wadeable 365 streams of the southeastern united states. Environ Sci Technol Lett 3, 243–249 (2016). 366 5. He, Y., Zhang, Y. & Ju, F. Metformin contamination in global waters: biotic and abiotic 367 transformation, byproduct generation and toxicity, and evaluation as a pharmaceutical 368 indicator. Environ Sci Technol 56, 13528–13545 (2022). 369 6. Briones, R. M. & Sarmah, A. K. Modelling degradation kinetics of metformin and 370 guanylurea in soil microcosms to derive degradation end-points. Environmental Pollution 371 245, 735–745 (2019). 372 7. Bailey, C. J. Metformin: historical overview. Diabetologia 60, 1566–1576 (2017). 373 8. Bailey, C. & Day, C. Metformin: its botanical background. Practical Diabetes 374 International 21, 115–117 (2004). 375 9. Watanabe, C. K. Studies in the metabolism changes induced by administration of 376 guanidine bases. Journal of Biological Chemistry 33, 253–265 (1918). 377 10. Triggle, C. R. et al. Metformin: Is it a drug for all reasons and diseases? Metabolism 133, 378 155223 (2022). 379 11. Di Magno, L., Di Pastena, F., Bordone, R., Coni, S. & Canettieri, G. The mechanism of 380 action of biguanides: new answers to a complex question. Cancers (Basel) 14, 3220 381 (2022). 382 12. Tobar, N. et al. Metformin acts in the gut and induces gut-liver crosstalk. Proceedings of 383 the National Academy of Sciences 120, (2023). 384 13. Wu, H. et al. Metformin alters the gut microbiome of individuals with treatment-naive 385 type 2 diabetes, contributing to the therapeutic effects of the drug. Nat Med 23, 850–858 386 (2017). 387 14. Vich Vila, A. et al. Impact of commonly used drugs on the composition and metabolic 388 function of the gut microbiota. Nat Commun 11, 362 (2020). 389 15. Sun, L. et al. Gut microbiota and intestinal FXR mediate the clinical benefits of 390 metformin. Nat Med 24, 1919–1929 (2018). 391 16. Sum, C. ‐F. et al. The effect of intravenous metformin on glucose metabolism during 392 hyperglycaemia in type 2 diabetes. Diabetic Medicine 9, 61–65 (1992). 393 17. Tucker, G. et al. Metformin kinetics in healthy subjects and in patients with diabetes 394 mellitus. Br J Clin Pharmacol 12, 235–246 (1981). 395 18. Tassoulas, L. J. & Wackett, L. P. Insights into the action of the pharmaceutical metformin: 396 Targeted inhibition of the gut microbial enzyme agmatinase. iScience 27, 108900 (2024). 397 19. Hillmann, K. B. & Niehaus, T. D. Genome sequences of two pseudomonas isolates that 398 can use metformin as the sole nitrogen source. Microbiol Resour Announc 11, (2022). 399 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 18 20. Martinez-Vaz, B. M. et al. Wastewater bacteria remediating the pharmaceutical 400 metformin: Genomes, plasmids and products. Front Bioeng Biotechnol 10, (2022). 401 21. Chaignaud, P. et al. A methylotrophic bacterium growing with the antidiabetic drug 402 metformin as its sole carbon, nitrogen and energy source. Microorganisms 10, 2302 403 (2022). 404 22. Li, T., Xu, Z.-J. & Zhou, N.-Y. Aerobic degradation of the antidiabetic drug metformin by 405 Aminobacter sp. strain NyZ550. Environ Sci Technol 57, 1510–1519 (2023). 406 23. Li, T. et al. Discovery of a Ni2+-dependent heterohexameric metformin hydrolase. Nat 407 Commun 15, 6121 (2024). 408 24. Michael, A. J. Polyamine function in archaea and bacteria. Journal of Biological 409 Chemistry 293, 18693–18701 (2018). 410 25. Meier-Kolthoff, J. P. & Göker, M. TYGS is an automated high-throughput platform for 411 state-of-the-art genome-based taxonomy. Nat Commun 10, 2182 (2019). 412 26. Robertson, J. & Nash, J. H. E. MOB-suite: software tools for clustering, reconstruction 413 and typing of plasmids from draft assemblies. Microb Genom 4, (2018). 414 27. Dong, M.-J., Luo, H. & Gao, F. Ori-Finder 2022: A comprehensive web server for 415 prediction and analysis of bacterial replication origins. genomics proteomics 416 bioinformatics 20, 1207–1213 (2022). 417 28. Siguier, P. ISfinder: the reference centre for bacterial insertion sequences. Nucleic Acids 418 Res 34, D32–D36 (2006). 419 29. Sullivan, M. J., Petty, N. K. & Beatson, S. A. Easyfig: a genome comparison visualizer. 420 Bioinformatics 27, 1009–1010 (2011). 421 30. Wissbroecker, K. B., Zmuda, A. J., Karumanchi, H. & Niehaus, T. D. Biochemical and 422 genomic evidence for converging metabolic routes of metformin and biguanide 423 breakdown in environmental Pseudomonads. Journal of Biological Chemistry 300, 424 107935 (2024). 425 31. Han, C.-G., Shiga, Y., Tobe, T., Sasakawa, C. & Ohtsubo, E. Structural and functional 426 characterization of IS 679 and IS 66-family elements. J Bacteriol 183, 4296–4304 (2001). 427 32. Shi, A., Fan, F. & Broach, J. R. Microbial adaptive evolution. J Ind Microbiol Biotechnol 428 49, (2022). 429 33. Tassoulas, L. J., Rankin, J. A., Elias, M. H. & Wackett, L. P. Dinickel enzyme evolved to 430 metabolize the pharmaceutical metformin and its implications for wastewater and human 431 microbiomes. Proceedings of the National Academy of Sciences 121, (2024). 432 34. Qiao, Z. et al. Structural insights into polyamine spermidine uptake by the ABC 433 transporter PotD-PotABC. Sci Adv 10, (2024). 434 35. Wu, D. et al. Structural basis of substrate binding specificity revealed by the crystal 435 structures of polyamine receptors Spud and Spue from Pseudomonas aeruginosa. J Mol 436 Biol 416, 697–712 (2012). 437 36. Berlinck, R. G. S., Bernardi, D. I., Fill, T., Fernandes, A. A. G. & Jurberg, I. D. The 438 chemistry and biology of guanidine secondary metabolites. Nat Prod Rep 38, 586–667 439 (2021). 440 37. Szymczak-Pajor, I., Drzewoski, J., Kozłowska, M., Krekora, J. & Śliwińska, A. The gut 441 microbiota-related antihyperglycemic effect of metformin. Pharmaceuticals 18, 55 442 (2025). 443 38. Wang, R. et al. The global distribution and spread of the mobilized colistin resistance 444 gene mcr-1. Nat Commun 9, 1179 (2018). 445 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 19 39. Willems, R. J. L. et al. Molecular diversity and evolutionary relationships of Tn 1546 -446 Like elements in Enterococci from humans and animals. Antimicrob Agents Chemother 447 43, 483–491 (1999). 448 40. Liu, Y.-Y. et al. Emergence of plasmid-mediated colistin resistance mechanism mcr-1 in 449 animals and human beings in China: a microbiological and molecular biological study. 450 Lancet Infect Dis 16, 161–168 (2016). 451 41. Acman, M. et al. Role of mobile genetic elements in the global dissemination of the 452 carbapenem resistance gene blaNDM. Nat Commun 13, 1131 (2022). 453 42. Partridge, S. R., Kwong, S. M., Firth, N. & Jensen, S. O. Mobile genetic elements 454 associated with antimicrobial resistance. Clin Microbiol Rev 31, (2018). 455 43. Olaitan, A. O., Chabou, S., Okdah, L., Morand, S. & Rolain, J.-M. Dissemination of the 456 mcr-1 colistin resistance gene. Lancet Infect Dis 16, 147 (2016). 457 44. Kieffer, N., Nordmann, P. & Poirel, L. Moraxella species as potential sources of mcr-like 458 polymyxin resistance determinants. Antimicrob Agents Chemother 61, (2017). 459 45. Snesrud, E. et al. A model for transposition of the colistin resistance gene mcr-1 by IS 460 Apl1. Antimicrob Agents Chemother 60, 6973–6976 (2016). 461 46. Khedher, M. Ben et al. Massive analysis of 64,628 bacterial genomes to decipher water 462 reservoir and origin of mobile colistin resistance genes: is there another role for these 463 enzymes? Sci Rep 10, 5970 (2020). 464 47. High time to tackle drug-resistant fungal infections. Nature 640, 569–569 (2025). 465 48. Razavi, M., Kristiansson, E., Flach, C.-F. & Larsson, D. G. J. The association between 466 insertion sequences and antibiotic resistance genes. mSphere 5, (2020). 467 49. Vo, T., Pontarotti, P., Rolain, J.-M. & Merhej, V. Mechanisms of acquisition of the vanA 468 operon among vancomycin-resistant Staphylococcus aureus genomes: The tip of the 469 iceberg? Int J Antimicrob Agents 63, 107154 (2024). 470 50. Varani, A., He, S., Siguier, P., Ross, K. & Chandler, M. The IS6 family, a clinically 471 important group of insertion sequences including IS26. Mob DNA 12, 11 (2021). 472 51. Weiss, T. et al. Secondary metformin monotherapy failure in individuals with type 2 473 diabetes mellitus. BMJ Open Diabetes Res Care 9, e002127 (2021). 474 52. Balaich, J. et al. The human microbiome encodes resistance to the antidiabetic drug 475 acarbose. Nature 600, 110–115 (2021). 476 53. Chandrappa, S., Shetty, S., Narayana, A. & Rajarathnna, K. A comparative study of 477 efficacy and safety of glimepiride versus gliclazide as an add on to metformin in Type 2 478 diabetes mellitus at a tertiary care hospital. Natl J Physiol Pharm Pharmacol 1 (2024) 479 doi:10.5455/njppp.2024.14.03142202430032024. 480 54. Diene, S. M. et al. Origin, diversity, and multiple roles of enzymes with metallo-β-481 Lactamase fold from different organisms. Cells 12, 1752 (2023). 482 55. Mojica, M. F., Rossi, M.-A., Vila, A. J. & Bonomo, R. A. The urgent need for metallo-β-483 lactamase inhibitors: an unattended global threat. Lancet Infect Dis 22, e28–e34 (2022). 484 56. Gillieatt, B. F. & Coleman, N. V. Unravelling the mechanisms of antibiotic and heavy 485 metal resistance co-selection in environmental bacteria. FEMS Microbiol Rev 48, (2024). 486 57. Murray, L. M. et al. Co-selection for antibiotic resistance by environmental contaminants. 487 npj Antimicrobials and Resistance 2, 9 (2024). 488 58. Rastogi, A. & Tiwari, M. K. Microbial degradation of pharmaceuticals. in 183–210 489 (2022). doi:10.1007/978-981-16-8367-1_9. 490 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 20 59. Ha, H., Mahanty, B., Yoon, S. & Kim, C.-G. Degradation of the long-resistant 491 pharmaceutical compounds carbamazepine and diatrizoate using mixed microbial culture. 492 Journal of Environmental Science and Health, Part A 51, 467–471 (2016). 493 60. Fetyan, N. A. H. et al. Bacterial degradation of ibuprofen: insights into metabolites, 494 enzymes, and environmental fate biodegradation of ibuprofen by Achromobacter species. 495 Microbiol Res (Pavia) 15, 2298–2315 (2024). 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 21 Table 517 Table 1. mfmAB homologs identified by BLASTn analysis, detailing genomic localization, 518 strain origins, and validation of metformin-hydrolyzing activity. 519 Figures 520 Fig. 1. Genomic reconstruction and proposed evolutionary scenario for the acquisition and 521 mobilization of the mfmAB gene cluster. a, Comparative genomic alignment of mfmAB-522 containing regions using Easyfig. Coding sequences (CDSs) are shown as arrows: mfmAB cluster 523 (red), homologs (black), insertion sequences (green), putrescine transporter genes (violet), 524 biguanide aminohydrolase (brown), guanylurea hydrolase (blue), and type IV secretion system 525 (T4SS) genes (dark blue). Gray shading (81–100% BLASTn identity) indicates conserved regions. 526 b, Hypothetical model of mfmAB horizontal gene transfer (HGT) and evolution in Aminobacter 527 and Pseudomonas spp. Chromosomes (black lines) and plasmids (circles) are illustrated with CDSs 528 represented as thick lines and color -coded as in panel a. Dashed arrows and circles indicate 529 potential intermediate steps, while solid arrows point post-HGT strains. c, Proposed dissemination 530 of mfmAB to animal and human microbiota via HGT, driven by co-selection with antibiotic/heavy 531 metal resistance genes (yellow). Gene transfer is suggested to occur via conjugation and 532 transposition. Created in BioRender. Bittar, F. (2025) https://BioRender.com/utwl1xw. 533 Fig. 2. Wastewater environmental microcosm: the PharmacoMicrobioResistance paradigm. 534 Degradation pathways of metformin and biguanide by environmental bacteria. Metformin -535 hydrolyzing isolates encode MfmAB enzymes (red), the putrescine transporter PotF (violet), 536 biguanide aminohydrolase BguH (brown), and guanylurea hydrolase GuuH (blue), type IV 537 secretion system (T4SS) genes (dark blue), all represented as thick lines. These genes, associated 538 with various transposable elements (green), are located on plasmids (circles). Additional enzymes 539 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 22 for complete catabolism of metformin such as guanidine carboxylase (GC), carboxyguanidine 540 deiminase (CgdAB), and allophanate hydrolase (AtzF) which are shown as thick orange lines and 541 encoded on the chromosome (black lines). Transporters for guanylurea SugE (grey), metformin 542 (light orange) and biguanides (light blue) are represented as rectangles. Created in BioRender. 543 Bittar, F. (2025) https://BioRender.com/oaz6y67. 544 Supplementary data 545 Supplementary Fig. 1. Phylogenetic tree based on whole -genome sequences of Aminobacter 546 spp. available in the NCBI database. Strains known to hydrolyze metformin are highlighted in 547 bold red, while non-metformin-hydrolyzing strains are shown in bold black. Other strains, shown 548 in regular font, have not been previously tested for metformin degradation20,21,33. 549 .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 23 Table 1. mfmAB homologs identified by BLASTn analysis, detailing genomic localization, 550 strain origins, and validation of metformin-hydrolyzing activity. 551 a incomplete genome. 552 WWTPs, wastewater treatment plants. ND, not determined33.553 BLASTn alignment results of the mfmAB against NCBI Refseq (E-value <1e-10) Strain origins and isolation timeline Functional characterization of MfmAB activity Organism Percent Identity Query coverage Genome assembly [Accession number hit] Genomic context (Size in bp) Source of isolation Country (region/state)- Year of isolation Metformin degradation References Aminobacter sp. NyZ550 100.00% 100% GCA_025628925 [CP114190] Chromosome (4,947,010) WWTPs China- 2021 Confirmed 22 Aminobacter niigataensis MD1 99.91% 100% GCA_946995915 [NZ_OX341517] Chromosome (5,500,801) WWTPs France (Strasbourg)- 2012 Confirmed 21 Aminobacter sp. MET-1 99.86% 100% GCA_026521175a [JAMSHL010000190] Plasmid pMET-1 (160,141) WWTPs USA (Minnesota)- 2021 Confirmed 20 Ectopseudomonas mendocina MET-2 99.63% 100% GCA_026503285 [CP098607] Plasmid pMET-2 (90,221) WWTPs USA (Minnesota)- 2021 Confirmed 20 Ectopseudomonas hydrolytica KHPS2 99.49% 100% GCA_024205225 [CP100554] Plasmid pKHPS2 (83,915) WWTPs USA (Minnesota)- 2021 Confirmed 20 Pseudomonas sp. KHPS1 99.58% 99% GCA_024205205 [CP100552] Plasmid pKHPS1 (79,836) WWTPs USA (Minnesota)- 2021 Confirmed 20 Aminobacter sp. MSH1 92.41% 100% GCF_003071665 [CP026265] Chromosome (5,301,880) Soil Denmark- 2007 Negative 33 Aminobacter niigataensis DSM 7050 92.32% 100% GCF_014200015a [NZ_JACHOT010000005] Chromosome contig (383,642) Soil Japan (Niigata)- 1992 Negative 33 Aminobacter sp. DSM 101952 Root100 91.90% 100% GCF_001424795a [NZ_LMCL01000001] Chromosome contig (720,165) Root Germany- 2013 Negative 33 Aminobacter aganoensis DSM 7051 91.71% 100% GCF_014206975a [NZ_JACHOU010000005] Chromosome contig (301,578) Soil Japan (Niigata)- 1992 Negative 33 Aminobacter sp. P9b 92.35% 100% GCF_045345995 [CP150082] Chromosome (4,937,309) Ground water Germany- 2016 ND - Aminobacter sp. SR38 91.56% 100% GCF_014843375 [CP062112] Chromosome (5,667,809) Soil France- 2000 ND - Aminobacter carboxidus DSM 1086 90.44% 100% GCF_014863355a [NZ_JACZEP010000010] Chromosome contig (155,224) Soil Russia- 1977 ND - Aminobacter anthyllidis D-10A 90.24% 100% GCF_029872135a [NZ_JARXNE010000001] Chromosome contig (562,047) Sediment China- 2022 ND - Aminobacter anthyllidis LMG 26462 90.11% 100% GCF_018555685a [NZ_JAFLWW010000004] Chromosome contig (670,596) Root France- 2011 ND - .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 24 Fig. 1. Genomic reconstruction and proposed evolutionary scenario for the acquisition and mobilization of the mfmAB gene cluster. a, Comparative genomic alignment of mfmAB- containing regions using Easyfig. Coding sequences (CDSs) are shown as arrows: mfmAB cluster (red), homologs (black), insertion sequences (green), putrescine transporter genes (violet), biguanide aminohydrolase (brown), guanylurea hydrolase (blue), and type IV secretion system (T4SS) genes (dark blue). Gray shading (81–100% BLASTn identity) indicates conserved regions. b, Hypothetical model of mfmAB horizontal gene transfer (HGT) and evolution in Aminobacter and Pseudomonas spp. Chromosomes (black lines) and plasmids (circles) are illustrated with CDSs represented as thick lines and color-coded as in panel a. Dashed arrows and circles indicate potential intermediate steps, while solid arrows point post-HGT strains. c, Proposed dissemination of mfmAB to animal and human microbiota via HGT, driven by co-selection with antibiotic/heavy .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 25 metal resistance genes (yellow). Gene transfer is suggested to occur via conjugation and transposition. Created in BioRender. Bittar, F. (2025) https://BioRender.com/utwl1xw. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 26 Fig. 2. Wastewater environmental microcosm: the PharmacoMicrobioResistance paradigm. Degradation pathways of m etformin and biguanide by environmental bacteria. Metformin - hydrolyzing isolates encode MfmAB enzymes (red), the putrescine transporter PotF (violet), biguanide aminohydrolase BguH (brown), and guanylurea hydrolase GuuH (blue), type IV secretion system (T4SS) genes (dark blue), all represented as thick lines. These genes, associated with various transposable elements (green), are located on plasmids (circles). Additional enzymes for complete catabolism of metformin such as guanidine carboxylase (GC), carboxyguanidine deiminase (CgdAB), and allophanate hydrolase (AtzF) which are shown as thick orange lines and encoded on the chromosome (black lines). Transporters for g uanylurea SugE (grey), metformin .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 27 (light orange) and biguanide s (light blue) are represented as rectangles. Created in BioRender. Bittar, F. (2025) https://BioRender.com/oaz6y67. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint 28 Supplementary data Supplementary Fig. 1. Phylogenetic tree based on whole -genome sequences of Aminobacter spp. available in the NCBI database. Strains known to hydrolyze metformin are highlighted in bold red, while non-metformin-hydrolyzing strains are shown in bold black. Other strains, shown in regular font, have not been previously tested for metformin degradation20,21,33. .CC-BY 4.0 International licenseavailable under a was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint (whichthis version posted April 28, 2025. ; https://doi.org/10.1101/2025.04.28.651035doi: bioRxiv preprint