Filling in the Gaps in Metformin Biodegradation: a New Enzyme and a Metabolic Pathway for Guanylurea

Abstract

The widely prescribed pharmaceutical metformin and its main metabolite, guanylurea, are currently two of the most common contaminants in surface and wastewater. Guanylurea often accumulates and is poorly, if at all, biodegraded in wastewater treatment plants. This study describes Pseudomonas mendocina strain GU, isolated from a municipal wastewater treatment plant, using guanylurea as its sole nitrogen source. The genome was sequenced with 36-fold coverage and mined to identify guanylurea degradation genes. The gene encoding the enzyme initiating guanylurea metabolism was expressed, and the enzyme was purified and characterized. Guanylurea hydrolase, a newly described enzyme, was shown to transform guanylurea to one equivalent (each) of ammonia and guanidine. Guanidine also supports growth as a sole nitrogen source. Cell yields from growth on limiting concentrations of guanylurea revealed that metabolism releases all four nitrogen atoms. Genes encoding complete metabolic transformation were identified bioinformatically, defining the pathway as follows: guanylurea to guanidine to carboxyguanidine to allophanate to ammonia and carbon dioxide. The first enzyme, guanylurea hydrolase, is a member of the isochorismatase-like hydrolase protein family, which includes biuret hydrolase and triuret hydrolase. Although homologs, the three enzymes show distinct substrate specificities. Pairwise sequence comparisons and the use of sequence similarity networks allowed fine structure discrimination between the three homologous enzymes and provided insights into the evolutionary origins of guanylurea hydrolase.IMPORTANCE Metformin is a pharmaceutical most prescribed for type 2 diabetes and is now being examined for potential benefits to COVID-19 patients. People taking the drug pass it largely unchanged, and it subsequently enters wastewater treatment plants. Metformin has been known to be metabolized to guanylurea. The levels of guanylurea often exceed that of metformin, leading to the former being considered a "dead-end" metabolite. Metformin and guanylurea are water pollutants of emerging concern, as they persist to reach nontarget aquatic life and humans, the latter if it remains in treated water. The present study has identified a Pseudomonas mendocina strain that completely degrades guanylurea. The genome was sequenced, and the genes involved in guanylurea metabolism were identified in three widely separated genomic regions. This knowledge advances the idea that guanylurea is not a dead-end product and will allow for bioinformatic identification of the relevant genes in wastewater treatment plant microbiomes and other environments subjected to metagenomic sequencing.

Keywords: CgdAB; Pseudomonas mendocina; biodegradation; enzyme; genome; guanidine; guanylurea hydrolase; metformin.

FIG 1

(A to C) Compounds and reaction in the present study.

FIG 2

Gene regions identified in P. mendocina strain GU and biodegradative pathway. (A) Genes 3481 to 3487; (B) genes 3879 to 3882; (C) genes 300 to 302. (D) Metabolic pathway via enzyme reactions and NH₃ stoichiometry.

FIG 3

Nitrogen growth comparison by setting molar amounts of nitrogen in ammonium chloride equivalent to guanylurea. Growth was conducted as described in Materials and Methods.

FIG 4

Sequence alignment and structural comparisons of guanylurea hydrolase (GH) with biuret hydrolase (BH) and triuret hydrolase (TH). (A) Sequence alignment was made using NCBI COBALT alignment tool with the BH (PLY61274.1) and TH (PLY61272.1) sequences from Herbaspirillum BH-1. (B and C) Key amino acid positions differentiating guanylurea (E211) and biuret (Q215) hydrolases, derived from Protein Data Bank structure 6AZQ. Homology model of GuuH made with BiuH as a template (48% sequence identity). The C170S variant of BiuH is catalytically dead, which allows for cocrystals with biuret.

FIG 5

Sequence similarity network (SSN) of 3,700 sequences of IHL proteins determined or predicted to degrade triuret, biuret, or guanylurea. There are >1,900 sequences ranging in identity from 45% to 100%. SSN built using EFI-EST with a cutoff of E−65 and visualized with Cytoscape.