INTRODUCTION
Guanylurea (carbamylguanidine) is one of the most widespread water contaminants, originating and accumulating from metformin, a first-line globally important drug (
1,
2). Metformin is widely prescribed as a treatment for type 2 diabetes and has even heightened significance in light of its observed anti-inflammatory, anticancer, and antiaging effects (
3–6). Most recently, metformin has been associated with reduced mortality in COVID-19 patients suffering from preexisting conditions such as diabetes and obesity (
7). Those studies also showed that human drug-metabolizing enzymes do not transform metformin, the drug is largely excreted unchanged, and microbes in wastewater treatment plants and aquatic environments transform it to guanylurea.
Guanylurea has been widely labeled as a recalcitrant “dead-end” product of metformin in numerous wastewater treatment systems (
8,
9). Guanylurea forms, and has been detected in, wastewater treatment plant effluent and coastal waters around the globe (
4,
10–12). Metformin and guanylurea have been reported in surface waters at levels of micrograms per liter, which is extremely high for a pharmaceutical contaminant (
12,
13). Metformin was detected in European surface waters at levels of up to 3 μg/liter, with guanylurea concentrations exceeding those of metformin by an order of magnitude or more (
14–16).
It is also relevant that metformin and guanylurea are charged molecules that bind poorly to granulated activated carbon, the worldwide standard treatment for removing pharmaceuticals, which are largely hydrophobic molecules (
12,
17). In that context, biodegradation is even more crucial for their removal. The levels of these compounds entering municipal wastewater treatment are likely to increase due to metformin’s broadening pharmacological efficacy, its widespread use, projected increases in the rates of obesity and diabetes, and the emerging new uses for the compound (
2,
16). There is also evidence that metformin and guanylurea may impact some aquatic species (
14,
18,
19). In light of these different impacts, guanylurea is considered in many countries to be one of the most prominent emerging water pollutants (
20).
While microbial metabolism of metformin is considered to be the major source of guanylurea, it can also derive from other anthropogenic and natural sources. Guanylurea may also accumulate from the microbial degradation of cyanoguanidine (dicyanodiamide), a common fertilizer additive (
21,
22). Derivatives of guanylurea are utilized in the manufacture of flame retardants as well as propellants for energetic compounds and munitions (
23,
24). Natural products containing the guanylurea moiety have been identified in the red alga
Gymnogongrus flabelliformis. These compounds include the novel amino acid gigartinine and a guanylureido acid named gongrine (
25–27), which are proposed to play a role in nitrogen metabolism (
28) and may contribute to faster spring growth for certain marine plants (
29).
Given the prevalent use of metformin and frequent detection of guanylurea in aquatic environments, research on the microbial degradation of these compounds is essential to develop biotechnological applications for bioremediation and wastewater treatment. Several studies have recently investigated the biodegradation of metformin and guanylurea by microbial communities isolated from activated sludge (
8,
11). In one study not involving metformin, the disappearance of guanylurea was shown to be accompanied by the appearance of guanidine as measured by high-performance liquid chromatography (HPLC) analyses, suggesting guanidine as a potential transformation product (
23). These studies examined the microbial breakdown of guanylurea but did not define metabolic pathways or the genes and enzymes mediating breakdown.
The present study describes the biodegradation of guanylurea by a Pseudomonas mendocina bacterium isolated, by enrichment, from a wastewater treatment plant. This bacterium utilized guanylurea as a nitrogen source for growth. Genome sequencing and bioinformatic analyses led to the identification of genes involved in the biodegradation of ureide and guanidinium compounds. A new member of the isochorismatase-like hydrolase protein family was shown to catalyze the conversion of guanylurea to guanidine, and the enzyme was characterized physically and kinetically. A mineralization pathway to ammonia and carbon dioxide was demonstrated. Insights obtained here may now be used with wastewater metagenome data to predict the intrinsic capacity for guanylurea biodegradation.
RESULTS
Enrichment and isolation of a pure culture growing on guanylurea.
Activated sludge from the Saint Paul, MN, municipal wastewater treatment plant was used as a source of bacteria for enrichment on guanylurea as the sole nitrogen source. After the first five transfers, guanylurea utilization was indicated by a significant increase in optical density in liquid culture and appearance of individual colonies on agar plates containing LB or guanylurea minimal medium. Subsequent transfers and plating of individual colonies led to the isolation of a pure culture with guanylurea-degrading capability as demonstrated by HPLC. No guanylurea was detected in the spent growth medium of the cultures after 24 h of incubation. Control media which were not inoculated did not exhibit significant guanylurea disappearance (<5%). The 16S rRNA sequence analysis identified the bacterium as a Pseudomonas mendocina, and it was designated here as strain GU.
Pseudomonas mendocina strain GU was able to utilize guanidine, agmatine, and urea as sole sources of nitrogen for growth, using citrate and acetate as carbon sources (see Table S1 in the supplemental material). The doubling time with guanylurea was 44 min. Doubling times with other nitrogen-rich compounds analogous to guanylurea ranged from 66 to 92 min: agmatine (66 min), guanidine (81 min), and urea (92 min). No growth was observed when biuret, metformin, or cyanoguanidine was used as the sole nitrogen source.
Genome sequencing and analysis of P. mendocina strain GU.
The strain was subjected to Illumina sequencing with 36-fold coverage. The
P. mendocina strain GU genome showed a high degree of gene sequence relatedness (see Table S2) and synteny (see Fig. S1) with
P. mendocina ymp, which was isolated from Yucca Mountain hazardous waste repository (
30). Properties of the two genomes are compared in
Table 1. Overall, the two genomes shared 98% average nucleotide sequence identity, and 77.6% of the predicted proteins in the genome of
P. mendocina strain GU shared >95% sequence identity to proteins present in the genome of strain ymp (Table S1). The major difference was the presence of an additional 665 coding sequences in strain GU (12.4% of the genome). Taken together, only 10% of the shared proteins were less than 95% identical.
Identification and purification of a guanylurea-degrading enzyme.
Guanylurea degradation genes or enzymes had not previously been identified. In light of this, the genome sequence of
P. mendocina strain GU was mined for the presence of genes that might encode enzyme classes that act on the guanidine or urea functional groups that, in combination, compose guanylurea. Genes encoding enzymes annotated to be active with the guanidinium compounds arginine and agmatine were identified but, based on the flanking genes, were thought not to be relevant to guanylurea. The organism has genes encoding an active urease, but urease was previously tested for activity with guanylurea and shown to be inactive (
31). A gene encoding a protein annotated from BLAST comparisons as “cysteine hydrolase” and “nicotinamidase” was analyzed in more detail. The translated amino acid sequence showed 48% sequence identity to a biuret hydrolase (BiuH) from
Rhizobium leguminosarum that had been characterized functionally (
32). We noted that biuret is structurally analogous to guanylurea, having only an oxygen replacing one of the guanylurea guanidinium nitrogen atoms (
Fig. 1A and
B). In this context, a synthetic gene encoding the biuret hydrolase homolog (NCBI accession no.
MBF8163004.1) was obtained and expressed in
Escherichia coli, and the His-tagged enzyme was purified in one step to homogeneity via a Ni affinity column. Incubation of the purified enzyme with guanylurea produced 1 mol of ammonia per mol of guanylurea, identifying this enzyme to be capable of degrading guanylurea. No redox cofactors or oxygen was required, indicating that the enzyme was a guanylurea hydrolase.
Characterization of the guanylurea hydrolase enzyme.
It was important to examine the kinetics and substrate specificity of the enzyme described above to determine if this enzyme was competent to support significant guanylurea degradation and if that was the major native activity. The purified enzyme was active over a pH range of 5.5 to 8.5, with highest specific activity at pH 8.0. At the pH optimum, the specific activity of purified guanylurea hydrolase with saturating guanylurea was 13 μmol per min per mg (
Table 2). The
kcat/
Km was higher (3-fold) than for the average enzyme with its cognate substrate, as compiled from a study of thousands of enzymes by Davidi and coworkers (
33). The enzyme showed minimal activity on biuret, with 0.06% the specific activity measured on guanylurea. Moreover, no activity (<0.01 μmol per min per mg) was detected with acetylurea, dicyandiamide, phenylurea, nitroguanidine, and 2-imino-thiobiuret (
Fig. 1C). These data indicate that the major biological function of the enzyme studied here is to release ammonia from guanylurea, allowing
P. mendocina strain GU to grow on guanylurea.
Elucidating the guanylurea hydrolase reaction product.
In addition to ammonia, guanidine was stably produced in incubations with guanylurea hydrolase and guanylurea. This could arise from hydrolytic cleavage of the terminal urea C-N bond or the subterminal N-C bond (
Fig. 1B). The former cleavage reaction would produce ammonia and carboxyguanidine, which readily decarboxylates, forming stable guanidine. The subterminal bond cleavage would produce guanidine directly, along with carbamate. Carbamate at neutral pH in water has a half-life of 70 ms (
34), rapidly converting to carbon dioxide and ammonia. So, both reactions produce ammonia and guanidine rapidly from unstable intermediates. However, it is now possible to “observe” carboxyguanidine because an enzyme that rapidly converts it to allophanate and ammonia was recently discovered (
35). In that previous study, guanidine carboxylase produced highly unstable carboxyguanidine that was rapidly hydrolyzed into a more stable compound, allophanate, via carboxyguanidine deiminase (CgdAB). Allophanate hydrolase was used to convert allophanate to ammonia and carbon dioxide. Here, an analogous experiment was performed by adding the enzymes CgdAB and allophanate hydrolase to GuuH reactions. If carboxyguanidine was formed, the additional enzymes would release an additional three equivalents of ammonia. If guanidine and carbamate were formed, no additional ammonia would be released. None of the enzymes have activity with guanidine.
In two separate experiments, carried out as described in the Materials and Methods, no increase in ammonia was observed in kinetic (short-term incubations) or stoichiometric (long-term incubations) experiments, indicating that guanidine is produced directly. In these experiments, an up to 40-fold excess of coupling enzymes were added, which would have released additional ammonia from carboxyguanidine, if that had been produced by GuuH.
Gene regions of the P. mendocina strain GU genome related to guanylurea mineralization.
With the observation that guanidine and carbamate are the likely products of the GuuH enzyme, and carbamate that decays spontaneously liberates one ammonia molecule, further ammonia release would require enzymes implicated in guanidine degradation. Guanidine metabolism has only recently been elucidated, and special identifying features of the genes were reported (
35). That information was used to aid in bioinformatic analysis of predicted protein-encoding regions in the
P. mendocina strain GU genome (
Fig. 2).
First, the guanylurea hydrolase gene (
guuH), identified in this study, did not appear to be contiguous to genes encoding enzymes involved in related metabolism that we could discern (
Fig. 2A). The only identifiable surrounding genes were ABC transporters of unknown function. A gene region likely to be involved in guanidine metabolism was localized ∼400 genes distant (
Fig. 2B). The large gene clearly encodes an enzyme including the biotin-binding domain and other regions diagnostic for carboxylases (
36). The carboxylase is further identified as a guanidine-metabolizing enzyme by a specific active region demarcated in a previous study (
35). Guanidine carboxylases contain an aspartic acid residue in the active site at a position typically occupied by asparagine in fungal and bacterial urea carboxylases (
35), and the protein here contained that diagnostic aspartate. Even stronger evidence is provided by cooccurrence of the
cgdAB genes next to our annotated guanidine carboxylase (
Fig. 2B). The
cgdAB genes are present in ∼7,000 bacterial genomes, and 96% of the time, they colocalize with a guanidine carboxylase (
35). The CgdA and CgdB proteins were shown previously to form a complex and to transform the product of guanidine carboxylase, carboxyguanidine, to allophanate. Lastly, guanidine metabolism gene expression was shown to be regulated by guanidine riboswitches (
37), and these are found directly upstream of the enzyme-encoding genes just described (
Fig. 2B).
Another gene region (genes 300 to 302) in the
P. mendocina strain GU genome encoded an allophanate hydrolase, an adjacent urea carboxylase, and a regulatory protein, respectively (
Fig. 2C). The carboxylase protein encoded by gene 301 contained an asparagine instead of aspartate at its active site, a feature consistent with a preference for urea as the substrate undergoing carboxylation (
35). Urea carboxylation produces allophanate, and allophanate hydrolase can degrade that to ammonia and carbon dioxide. In the presence of guanylurea, the five genes highlighted in red in
Fig. 2 encode four proteins (CgdA and CgdB function as one protein) that constitute a complete pathway for guanylurea that can liberate all four nitrogen atoms as ammonia (
Fig. 2D).
In addition to the genomics and bioinformatic inferences described above, there is direct experimental data consistent with the pathway shown in
Fig. 2D. First,
P. mendocina strain GU grows readily on guanidine as a sole nitrogen source. Only one metabolic pathway for guanidine is currently known and that proceeds as shown in
Fig. 2, via guanidine carboxylase, CgdAB, and allophanate hydrolase (
35), for which genes are present in
P. mendocina strain GU.
To further test the pathway shown in
Fig. 2D experimentally, we carried out ammonia stoichiometry experiments. If all of the relevant genes shown in
Fig. 2 are expressed, it is predicted that 4 equivalents of ammonia would be produced for each guanylurea molecule consumed. To test this hypothesis, parallel cultures of
P. mendocina strain GU were grown with limiting amounts of nitrogen, and then cell densities were determined when growth ceased. As shown in
Fig. 3, the growth on 1 mol of guanylurea, containing four nitrogen atom equivalents, was the same as with 4 mol of ammonium ion. For example, separate growth cultures containing 1.92 mM ammonium chloride or 0.48 mM guanylurea each grew to an optical density at 600 nm (OD
600) of 0.8. Both contained the same number of nitrogen atoms, equivalent to 1.92 mM ammonia. These data are consistent with guanylurea being completely mineralized to release all the nitrogen atoms contained within the compound. One can draw out a chemically plausible pathway in which guanylurea undergoes a direct deamination reaction to produce biuret, but biuret was negative as a growth substrate and metabolism producing biuret would only yield one nitrogen equivalent, not four as was observed.
Guanylurea hydrolase is a new member of the isochorismatase hydrolase-like protein family.
A query of the Pfam database with the GuuH amino acid sequence gave a match to the isochorismatase hydrolase-like (IHL) protein family (PF00857.20) with an E value of 7.1e
−42. GuuH (GH in
Fig. 4) consists of 231 amino acids and matched extensively over amino acids 26 to 217 with biuret hydrolase (BH) and triuret hydrolase (TH) (
Fig. 4). IHL proteins are sometimes denoted as cysteine hydrolases, and GuuH contains a cysteine, C171, that aligns with a cysteine in BH and TH and found in a highly conserved region (
Fig. 4A). X-ray structures have been determined for biuret (
38) and triuret hydrolases (
39), and those studies have revealed the cysteine aligning with C171 in GuuH acts as a nucleophile to catalyze initial bond cleavage. The resulting acyl enzyme intermediate further undergoes hydrolysis to complete the reaction. Those structural and mechanistic studies have also revealed that the biuret and triuret hydrolase active sites both contain a D-K-C catalytic triad, of which all the residues reside in GuuH in alignment and in highly conserved regions of the protein overall (
Fig. 4A).
One notable difference in the sequences was with a glutamine residue that helps bind substrate in biuret and triuret hydrolases and that was replaced with a glutamate in GuuH, as illustrated in
Fig. 4B. X-ray structures for biuret and triuret hydrolases reveal that the glutamine residue hydrogen bonds to the substrate amide group distal to the one reacting at the active site cysteine (
38,
39). The change to an aligning glutamate, Glu221, in GuuH is logical, chemically, as the negatively charged glutamate would be expected to bind electrostatically to the positively charged guanidinium group, thus aligning the guanylurea substrate in a similar manner as biuret and triuret hydrolases align their substrates. This suggests that the glutamine-to-glutamate change is key for differentiating GuuH enzymes from biuret and triuret hydrolases.
To test this hypothesis, we mutated the glutamine residue in the biuret hydrolase from
Herbaspirillum sp. strain BH-1 (Gln212) that corresponds to Glu221 of GuuH and replaced the residue with a glutamate residue. The resulting enzyme, BiuH Q212E, showed nearly 2 orders of magnitude diminished activity with biuret compared to that of the wild type, and guanylurea hydrolysis became the dominating function of the enzyme, with 40% higher activity with guanylurea than with biuret (see Table S3). The wild-type biuret hydrolase showed no detectable activity with guanylurea (Table S3). This result demonstrated that a single point mutation can convert a biuret hydrolase to an enzyme significantly active with guanylurea. Triuret hydrolases have a larger active site, and the glutamine residue is significantly further away from the catalytic cysteine than in GuuH and biuret hydrolases (
39). These different properties also allow discrimination between the closely related GuuH enzymes, biuret hydrolases, and triuret hydrolases (
Fig. 5).
The defining signatures that identified biuret, triuret, and guanylurea hydrolases (
Fig. 4) were used with sequence similarity network (SSN) analysis to look broadly at the interrelationships of the three homologous enzymes with closely aligning sequences pulled from GenBank. The SSN analysis showed the GuuH enzymes to be more closely related to the biuret hydrolases, consistent with the similar sizes of the substrates and their ready interconversion (
Fig. 5). The observation that there were several separated lobes of GuuH sequences in the SSN may suggest that different sequence clusters of GuuH enzymes arose independently from biuret hydrolases, consistent with our observation of a single point mutation changing activities. Guanylurea may largely derive from metformin and the fertilizer additive cyanoguanidine, both which are only recently in the environment, perhaps suggesting a recent evolutionary origin of at least some GuuHs.
DISCUSSION
There is increasing focus on the occurrence of human pharmaceuticals in surface and wastewaters, and metformin is often at or near the top of the list with respect to the number of source detections and levels (
40). Metformin has a moderate half-life of 5 to 10 days in river simulation studies, with variation related to microbial diversity (
41). However, because of metformin’s extremely high usage, with 70 million prescriptions annually and gram quantities per daily dose, the compound is continuously introduced into surface waters and wastewater treatment plants. For many anthropogenic chemicals in the environment, levels of stable microbial degradation products often exceed those of the parent compound (
42), and this is almost invariably the case with metformin. Guanylurea has long been identified as a relatively persistent and high-level transformation product of metformin (
11,
12). The persistent high levels of guanylurea have led to it being labeled as a “dead-end” product of metformin.
While there have been reports of guanylurea biodegradation (
11,
12,
43), this report identifies specific genes, enzymes, and a metabolic pathway mediating its metabolism. In the present study, we isolated a bacterium from a wastewater treatment plant that degrades guanylurea to meet its nitrogen needs, sequenced the genome, identified and purified a new guanylurea-hydrolyzing enzyme, elucidated a metabolic pathway for guanylurea mineralization, and identified the evolutionary origins and connectivity of the guanylurea hydrolase to homologous enzymes in the IHL protein family, a part of the cysteine hydrolase superfamily.
As of 9 November 2020, in recorded genomes from NCBI and EMBL databases, GuuH is encoded in less than 0.5% of genomes. The potential rare occurrence of GuuH in the environment and the gene regulation in guanylurea metabolism may play a role in the accumulation of guanylurea in certain environments. Guanidine metabolic genes encoding guanidine carboxylase and carboxyguanidine deiminase can be turned on by a guanidine-sensing riboswitch, which is present in
P. mendocina strain GU, but there has been no insight into a suppression mechanism of these genes in the presence of guanidine that could be coordinated by broader nitrogen or carbon metabolism, for example (
37). The entire regulation of guanylurea metabolic genes, which are encoded in three separate regions of the genome, is complex and will be examined in detail in a future study.
The results presented here have utility in at least two distinct areas. The first is in environmental studies, specifically, where the fates of metformin, cyanoguanidine (
44), and guanylurea are being examined. The results here present a roadmap for identifying specific bacteria, genes, and metabolic enzymes that would be present in a particular environment of interest. For example, wastewater metagenomic data (
45) may now be culled to look for the presence of these genetic capacities. The second outgrowth of the data would be in medical studies. Metformin prescription rates are expected to rise given its efficacy for treating type 2 diabetes, obesity, and now, perhaps, COVID-19 (
7). Metformin is not known to be metabolized by human drug-metabolizing enzymes, but it is known to impact the bacteria in human intestines (
46). Indeed, several studies indicate that the human gut microbiome plays a significant role in modulating metformin’s therapeutic effects, but the mechanisms of that modulation are currently obscure (
47). In light of the overlapping environmental and medical implications, the metabolism of metformin, which largely appears to occur via guanylurea, has heightened importance.
It is intriguing to consider that new GuuH enzymes may be arising in recent evolutionary time to handle new anthropogenic chemical inputs from the expanding usage of metformin. Biuret and triuret hydrolases appear to be ancient enzymes that have diverged with the taxonomy of the bacteria that harbor them (
39,
48). Triuret hydrolases have diverged significantly from biuret hydrolases, a phenomenon clearly indicated by the visual separation on the sequence similarity network shown in
Fig. 5. In contrast, the enzymes identified by sequence signatures as GuuH are interspersed and emerging from several regions of the biuret hydrolase enzyme cluster, as shown in
Fig. 5. One explanation for this would be that GuuH activities have arisen multiple times from biuret hydrolase precursors. Given that a single mutation can impart significant guanylurea hydrolase activity upon a biuret hydrolase, it is plausible that recent evolution is occurring in different global waters as metformin input increases.
There are other precedents for recent evolution of biodegradative enzymes in response to new anthropogenic environmental inputs. For example,
trans-3-chloroacrylate dehalogenase from
Pseudomonas cichorii 170 that degrades the nematicide 1,3-dichloropropene was thought to have arisen in recent evolutionary times from 4-oxalocrotonate tautomerase (
49,
50). A similar observation has been made with anthropogenic
s-triazine compounds, such as the herbicide atrazine, in which atrazine chlorohydrolases arose independently from divergent members of the amidohydrolase protein superfamily (
51,
52). Moreover, the fate of anthropogenic chemicals and their metabolites in the environment has been observed to change in recent times as the result of microbial enzyme evolution (
53). Given that GuuH enzymes can readily arise from a biuret hydrolase via simple mutation(s), and the increasing environmental prevalence of compounds giving rise to guanylurea, we expect that guanylurea will increasingly lose its designation as a “dead-end” metabolite.
MATERIALS AND METHODS
Enrichment cultures and isolation of guanylurea-degrading bacteria.
The bacterial
Pseudomonas mendocina strain GU was isolated from a sample of activated sludge collected at the metropolitan wastewater treatment plant in Saint Paul, MN. Isolation was achieved by enrichment culture with citrate-acetate medium and 1 g of sludge per 50 ml of minimal medium as the inoculum. The minimal medium contained the following per liter of deionized water: 5.45 g K
2HPO
4, 0.2 g MgSO
4·7H
2O, 0.1 g NaCl, 1.5 g 20 mM sodium acetate, and 20 mM sodium citrate (
54). Guanylurea (1 mM) was then added as the sole nitrogen source. Enrichments and isolates were grown at 30°C in a shaking incubator at 200 to 225 rpm. Cultures were transferred in 10-fold dilutions into fresh medium every 7 days. Individual isolates were obtained by plating 10-fold serial dilutions of the enrichments on selective guanylurea plates, transferred to Luria-Bertani (LB) plates, and then isolated by streaking on LB and minimal medium plus guanylurea plates until pure.
Growth studies.
Growth studies were conducted in triplicates in citrate-acetate medium containing 1 mM each of their respective nitrogen source: guanylurea, biuret, guanidine, urea, or metformin. Nitrogen-free citrate-acetate medium served as a negative control. Citrate-acetate medium containing 6 mM ammonium chloride (NH4Cl) served as a positive control. Cell growth was monitored spectrophotometrically at 600 nm initially at 12- to 24-h intervals. Cultures with short lag phases were monitored every 4 to 6 h.
DNA extraction and PCR testing for taxonomic identification.
Genomic DNA was purified utilizing a Qiagen DNeasy blood and tissue kit (Qiagen, Valencia, CA) and Promega Wizard genomic DNA purification kit (Promega, Madison, WI). PCR amplification for taxonomic identification of bacterial strains was performed with universal 16S primers, 530F (5′-GTGCCAGCMGCCGCGG-3′) and 1492R (5′-GGTTACCTTGTTACGACTT-3′), using Q5 high-fidelity polymerase (New England BioLabs, Ipswich, MA). Amplification of the 1.4-kb DNA 16S rRNA fragment was achieved using the following conditions: 98°C for 2 min, and then 35 cycles consisting of 95°C for 1 min, 60°C for 30 s, and 72°C for 1 min; a final extension of the PCR product was performed at 72°C for 5 min. All PCRs were carried out using a concentration of 0.5 μM for each primer. Amplification products were purified using the QIAquick DNA extraction kit and sequenced (Functional Biosciences, WI, USA) to determine the identity of the microbial isolate.
HPLC analysis.
HPLC testing was conducted with a Waters IC-Pak anion HC column, 4.6 by 150 mm. Isocratic mobile phase consisted of a 5 μM KPO4 buffer at pH 8.0, and the flow rate was 0.5 ml/min. Eluents were monitored at 194 nm. Standards were in 125 mM KPO4 buffer, pH 8.0. Spent medium was analyzed by HPLC after cells were grown in citrate-acetate medium containing 1 mM guanylurea for 24 h. Cells were centrifuged at 3,000 × g for 15 min, and supernatant was then collected and filter sterilized through a 0.2-μm filter. Controls included sterilized N-free medium and 1 mM guanylurea solution. Enzyme was prepared at a concentration of 1 μg/ml in 125 mM KPO4 buffer, pH 8.0, with 1 mM guanylurea. Samples were incubated either overnight at room temperature or for 30 min and boiled for 2 min to inactivate enzyme.
Genome sequencing and analysis.
Total genomic DNA from microbial isolates was sequenced using a Roche GS 454 FLX system and standard LR 70 chemistry. Illumina Nextera XT library preparation and sequencing (on a MiSeq with V3 chemistry and 300-bp paired-end reads) services were provided by the University of Minnesota Biomedical Genomics Center (BMGC, Saint Paul, MN, USA). Adaptors and low-quality bases were trimmed from raw reads with trimmomatic v 0.36.
De novo assembly was performed using SPAdes v 3.13.0 (
55). Initial genome annotation was performed with prokka v 1.12 (
56). Sequence similarity networks were made using the enzyme function initiative-enzyme similarity tool (EFI-EST) (University of Illinois at Urbana-Champaign), retrieving 10,000 sequences related to GuuH (
57). The network was visualized in Cytoscape, and the clusters of guanylurea hydrolase, biuret hydrolase, and triuret hydrolase were identified as presented in
Fig. 5 (
58).
Protein expression and chromatography.
The putative GuuH gene (NCBI accession no.
MBF8163004.1) was codon optimized for
E. coli and synthesized by Integrated DNA Technologies (Coralville, IA). The gene was cloned into a pET28b+ vector with Gibson assembly master mix (New England BioLabs) using NdeI and HindIII restriction sites with an N-terminal 6×His tag and transformed into BL21-DE3
E. coli cells (New England BioLabs). Site-directed mutants of the biuret hydrolase from
Herbaspirillum sp. BH-1 (NCBI accession no.
PLY61274.1) and
guuH were made using the Q5 site-directed mutagenesis kit from New England BioLabs. The
guuH gene was expressed by growing cells in lysogeny broth (LB) medium with 50 μg/ml kanamycin at 37°C and 200 rpm to an OD
600 of 0.6 in a shake flask. The culture was cooled to 14°C and induced with 1 mM isopropyl β-
d-1-thiogalactopyranoside (IPTG) and, with the same agitation, incubated for 20 h. Induced cells were harvested at 4,000 ×
g and resuspended in lysis buffer (20 mM phosphate, 200 mM NaCl, pH 7.4). Cells were lysed with 3 passes in a French press at 10,000 lb/in
2. The cells were centrifuged for 90 min at 19,000 ×
g, and the supernatant was passed through 0.2-μm filter prior to loading into a GE Life Sciences AKTA fast liquid protein chromatography (FPLC) system for injection onto a GE Life Sciences HisTrap HP 5-ml column. After washing to limit nonspecific protein binding, GuuH was eluted with a linear gradient from 0.1 M to 0.25 M imidazole in lysis buffer over 10 column volumes with flow rate set at 2 ml/min, and fractions were collected. To determine the oligomeric state of GuuH, gel filtration was performed using a GE Healthcare HiLoad 16/600 Superdex 200-pg column. The column was equilibrated with 50 mM Tris-200 mM sodium chloride at pH 8, and the sample was injected onto the column and washed with 1 column volume at 1 ml/min flow rate, with GuuH eluted as a homotetramer at ∼96 kDa.
Enzyme assays.
Enzyme activity was measured by detection of ammonia release using the Berthelot reaction as previously described (
59). Experiments were performed at room temperature in 125 mM sodium phosphate dibasic buffer at pH 8. Specific activity at 25°C was determined at the pH optimum for the enzyme using the appropriate buffer (acetate, pH 5.5 to 6.5; phosphate, pH 7.0 to 8.0; borate, pH 8.5 to 10.5) with 1 mM guanylurea.
CgdAB from
Pseudomonas syringae, a heterodimeric enzyme (NCBI accession no.
WP_005764729.1 and
WP_005764727.1, respectively), and allophanate hydrolase from
Pseudomonas sp. strain ADP (NCBI accession no.
WP_011117193.1) were employed in coupled enzyme assays to determine if carboxyguanidine is formed during GuuH hydrolysis of guanylurea (
35). The CgdAB protein was a gift from Martin St. Maurice of Marquette University, and it was shown to be active in a coupled assay with guanidine carboxylase with guanidine as the substrate. Enzyme kinetic assays were performed in a 125 mM sodium phosphate dibasic buffer solution, pH 8, with 1 mM guanylurea as the substrate. To initiate the reaction, GuuH was added to a final assay concentration of 0.27 μg/ml. CgdAB was added to a final assay concentration of 15.6 μg/ml and AtzF protein to a final concentration of 7.38 μg/ml. Ammonia release was measured utilizing the Berthelot assay (
59). The assay was measured in a Beckman Coulter DU 640 UV-visible (UV-Vis) spectrophotometer.
A coupled enzyme assay to determine stoichiometries of the guanylurea and ammonium liberated utilized GuuH, AtzF, and CgdAB proteins. The reactions were performed overnight in a solution containing 50 μM guanylurea in 125 mM sodium phosphate, pH 8. Enzyme concentrations were 60 μg/ml GuuH, 15.6 μg/ml CgdAB, and 7.38 μg/ml AtzF. Reaction mixtures were incubated at 20°C overnight, and stoichiometries were determined by ammonia release utilizing the Berthelot reaction as described previously.
Data availability.
Genome sequences for
Pseudomonas mendocina strain GU are available in GenBank at BioProject accession
PRJNA675777 or BioSample accession
SAMN16722328.