Ancient Evolution and Recent Evolution Converge for the Biodegradation of Cyanuric Acid and Related Triazines

Abstract

Cyanuric acid was likely present on prebiotic Earth, may have been a component of early genetic materials, and is synthesized industrially today on a scale of more than one hundred million pounds per year in the United States. In light of this, it is not surprising that some bacteria and fungi have a metabolic pathway that sequentially hydrolyzes cyanuric acid and its metabolites to release the nitrogen atoms as ammonia to support growth. The initial reaction that opens the s-triazine ring is catalyzed by the unusual enzyme cyanuric acid hydrolase. This enzyme is in a rare protein family that consists of only cyanuric acid hydrolase (CAH) and barbiturase, with barbiturase participating in pyrimidine catabolism by some actinobacterial species. The X-ray structures of two cyanuric acid hydrolase proteins show that this family has a unique protein fold. Phylogenetic, bioinformatic, enzymological, and genetic studies are consistent with the idea that CAH has an ancient protein fold that was rare in microbial populations but is currently becoming more widespread in microbial populations in the wake of anthropogenic synthesis of cyanuric acid and other s-triazine compounds that are metabolized via a cyanuric acid intermediate. The need for the removal of cyanuric acid from swimming pools and spas, where it is used as a disinfectant stabilizer, can potentially be met using an enzyme filtration system. A stable thermophilic cyanuric acid hydrolase from Moorella thermoacetica is being tested for this purpose.

INTRODUCTION

Rings happen. The spontaneous (abiotic) formation of chemicals, especially aromatic rings, is prominent in nature and complements the formation of chemical compounds by biosynthesis (natural products) and human synthesis (anthropogenic products). The most well-known examples of abiotic ring formation are the alicyclic, benzenoid, and polycyclic aromatic hydrocarbon rings found in petroleum that form in the earth's subsurface at high pressure and temperatures of >200°C (1). In addition to petroleum hydrocarbons, abiotic chemical reactions produce a wide array of oxygen, sulfur, and nitrogen heterocyclic rings.

Heterocyclic rings are known to also form spontaneously (1). At 25°C and 1 atm pressure, isocyanic acid spontaneously cyclizes to form a mixture of cyanuric acid and cyamelide (2). Cyamelide is not very stable, but the partially aromatic cyanuric acid ring is highly resistant to abiotic hydrolysis or other ring opening reactions. Given that isocyanic acid is widespread in the universe, where it is commonly found in gas clouds, meteors, and comets (3), the highly stable cyanuric acid is plausibly an important sink of organic carbon, nitrogen, and oxygen in early planetary systems. Recent studies investigating RNA as an early genetic code found that cyanuric acid and triaminopyrimidine self-assemble in water to create aggregates resembling contemporary nucleic acid base pairs. Triaminopyrimidine forms nucleosides with ribose, and upon heating, the cyanuric acid mixture forms gene-length polymers that have been termed proto-RNA (4, 5). This polymerization suggests that cyanuric acid and related cyclic compounds might have played a role in the initiation of life on earth.

Anthropogenic synthesis of cyanuric acid was first accomplished by Scheele in 1776, although he called it pyruvic acid (6, 7). In 1820, Serrulas intentionally synthesized cyanuric acid by mixing cyanogens in water (6). Since the 1950s, cyanuric acid has become a major industrial chemical due to its easy synthesis from simple precursors, its stability once formed, and its intermediate role in making more elaborate commercial compounds (8).

Cyanuric acid and related s-triazine compounds have found industrial applications in chlorine disinfection processes, in electrical varnishes, the reduction of nitrogen oxides in stationary diesel engine exhaust gases, and as dyes, pharmaceuticals, explosives, polymer intermediates, fire retardants, and herbicides (9–14). With the exception of cyanuric acid, whose origins in nature were previously mentioned, most s-triazine compounds are likely anthropogenic, being synthesized by humans and hence entering the environment only in the past 150 years. The biodegradation of various s-triazines by microorganisms has been studied extensively in the last 20 years (15–35), and many of the biodegradation pathways have been shown to have cyanuric acid as a metabolic intermediate (Fig. 1) (16, 19, 36). Therefore, the occurrence of cyanuric acid on modern-day Earth may be attributable to a combination of abiotic processes, industrial synthesis, and microbial metabolism. This minireview posits that cyanuric acid is an ancient chemical, meaning that its occurrence predates the appearance of humans on Earth, whereas more complex s-triazine compounds synthesized for industrial purposes are relatively new. In this context, the enzyme initiating an attack on cyanuric acid is likely ancient, and more recent evolutionary events provided enzymes that funnel various s-triazine compounds into cyanuric acid as a common intermediate. Channeling new metabolites into an existing pathway is known to be a common feature for de novo pathway evolution (37, 38). The main focus of this minireview is on the cyanuric acid hydrolases, for which the X-ray structure has only recently been solved, and for which biotechnological applications have recently emerged.

FIG 1.

Sources of cyanuric acid. (A) Abiotic production from isocyanic acid that likely occurred on early Earth. (B) In the modern world, the metabolism of diverse anthropogenic s-triazine compounds funnels into cyanuric acid as a common intermediate.

RECENT EVOLUTION OF ATRAZINE METABOLISM

There is ample multifaceted evidence that the initial reactions in the metabolism of the s-triazine herbicide atrazine, which funnels into cyanuric acid, evolved and spread in recent evolutionary time. Prior to 1993, atrazine was considered to be poorly biodegraded in the field (39), and now, it is found to be highly biodegradable (23). Second, the AtzA enzyme that initiates atrazine metabolism via hydrolytic dechlorination is 98% identical to melamine deaminase, and these enzymes are readily interconvertible via two point mutations (15, 40). Third, the atzA gene has been identified on six continents in highly divergent microorganisms, and the sequences are >99% identical (28). Fourth, the atzA gene is almost invariably encoded by a plasmid-carried gene, and the gene is shown to be flanked by identical IS1071 sequences in highly divergent organisms (24, 27, 29). Last, the atzABC genes, which encode enzymes metabolizing atrazine to cyanuric acid, are not organized into an operon and are thought to be constitutively expressed in Pseudomonas sp. strain ADP (29, 41). There is uniform agreement in the scientific literature that the enzymes initiating the metabolism of atrazine have recently evolved and spread globally due to the selective pressure of atrazine, which serves as a source of nitrogen for those organisms that can biodegrade it (16, 17, 23–27, 30–35, 38, 42).

CYANURIC ACID HYDROLASE GENES AND ENZYMES

In contrast to the metabolism of atrazine discussed above, the data are consistent with bacteria having evolved the enzymes to liberate ammonia from cyanuric acid long before industrial s-triazine production started 150 years ago (12, 21, 36, 43–51). Sequence divergence, broad phylogenetic distribution, and the fixation of the cyanuric acid degradation genes on chromosomes and, in some cases, within operons, all point to an ancient origin of the gene function (29, 52–54).

To date, there has been only one enzyme family identified with the catalytic potential to open the cyanuric acid ring, which is known as the cyanuric acid hydrolase (CAH)/barbiturase family. CAH and barbiturase enzymes act on structurally analogous 6-membered ring substrates; yet each enzyme is specific and does not react with the other's substrate (Fig. 2) (55–57). CAH produces biuret via a hydrolytic ring opening, followed by decarboxylation. Barbiturase catalyzes the second step in the oxidative pyrimidine degradation pathway. While most organisms utilize a reductive pyrimidine degradation pathway, some actinobacteria use barbiturase in an oxidative catabolic pathway (53, 55, 56). Unlike barbiturase, cyanuric acid hydrolases are more broadly distributed throughout bacteria and fungi (53, 58), but they are still rare overall, appearing in <0.3% of the >51,000 completed microbial genomes surveyed here. To date, cyanuric acid hydrolase activity has experimentally been demonstrated in strains of Rhizobium, Methylobacterium, Pseudomonas, Acidovorax, Actinobacteria, Moorella, and the fungus Sarocladium, but it is not known to be present in archaea, plants, or animals (48, 53, 58–62).

FIG 2.

Reactions catalyzed by cyanuric acid hydrolase and biuret hydrolase (A) and barbiturase (B).

Further steps in the cyanuric acid degradation pathway involve a biuret hydrolase from either the amidase signature family, like AtzE/TrzE (29), or from the cysteine hydrolase superfamily (63), with each generating allophanate as a product. Allophanate hydrolases, such as AtzF/TrzF, are also amidase signature family proteins and catalyze the conversion of allophanate to 2 moles each of ammonia and carbon dioxide (29, 64). In sequenced genomes, the genes encoding biuret hydrolase and allophanate hydrolase are sometimes found in close proximity to the CAH. In Pseudomonas sp. ADP, the three genes are organized into a tightly controlled cyanuric acid degradation operon that responds to nitrogen limitation and cyanuric acid concentration (29, 52), a sophisticated regulatory structure that contrasts sharply with the lack of regulation observed with the atzABC genes in the same microorganism. The AtzDEF proteins were also shown to form a large multienzyme complex with the potential to funnel metabolites to more efficiently degrade cyanuric acid and produce ammonia for metabolic growth (65). This orchestration at the gene and protein levels implies multiple levels of coordination of function for these biodegradation activities.

EARLY EVOLUTION AND RECENT GENE RESURGENCE

A recent update of the CAH/barbiturase family conducted here identified 169 different sequences (see Table S1 in the supplemental material), compared to 41 sequences analyzed previously (53). Moreover, the current analysis, unlike the previous study, examined the correlation of sequence phylogeny and taxonomy. The current sequences come from a wide variety of organisms that include bacteria, fungi, and eukaryotic algae. The functional diversity of these organisms is immense, with organisms initially isolated for photosynthetic potential (66); the ability to produce vinegar by fermentation (67); the ability to oxidize ferrous iron and inorganic sulfur compounds at low temperatures (68); symbiotic nodulation and nitrogen fixation (69); pathogenesis (70, 71); and bioremediation in the agricultural, mining, explosives, and petroleum industries (30, 68, 72, 73).

Neighbor-joining and maximum likelihood family protein trees were produced with these new sequences (Fig. 3), similarly to the tree used in our 2012 paper (53). The trees produced by both PHYLIP and MEGA 6.06 were in agreement with the results presented here (74, 75). The previous paper grouped the 41 sequences available at the time into clades based upon sequence identity and was used primarily to identify diverse proteins for further characterization. In this paper, 128 sequences were added, and the increased sampling allowed the identification of taxonomic conservation within clades based upon the organisms containing the CAH proteins. A detailed taxonomic analysis of the organisms containing each protein sequence (see Fig. S1 and Table S2 in the supplemental material) was superimposed visually onto the tree by the addition of colors indicating specific taxonomic classifications at the phylum and class levels to uncover additional insights into the relationship of these sequences with the organisms from which they originate. When looking at how protein functionalities have evolved and are found in diverse organisms, an analysis of this type is extremely informative, elucidating horizontal gene transfers, gene duplications, and functional divergence events. For instance, horizontal gene transfer was shown to occur during PcpB evolution (76) and urea aminolyase/urea carboxylase distribution (77). Paralogs involving gene duplication and dissemination were also shown in this manner for the small multiresistance protein family, which was linked to larger multidrug transporters (78). The beneficial and widespread use of this technique has even promoted the development of PhyloView, a specific software capable of applying taxonomic information to protein trees (79). We chose to manually conduct a similar analysis due to the limited number of proteins involved and the database requirements of the program.

FIG 3.

Maximum likelihood phylogenetic tree for the CAH/barbiturase family matched with the phylogeny of the organisms containing them (indicated in color) and instances in which that matching does not hold (dotted oval). Moving from left to right across the top of the figure, the clades consist of organisms with the following taxonomy: fungi (brown); Actinobacteria (light blue), which has a single betaproteobacterium (gray); Micromonas (green yellow); Firmicutes (salmon); and Alphaproteobacteria (proceeding left to right in the clade) dominated by Rhizobium (purple), some Agrobacterium spp. (purple), a subset of photosynthetic Bradyrhizobium spp. (cyan), the Azorhizobium protein that has a solved crystal structure (purple), the one cloned Methylobacterium sp. with known CAH activity (red), and a subclade with all uncharacterized Methylobacterium spp. (red) adjacent to two gammaproteobacteria (tan). Barbiturases and related proteins (dark blue) are located on the tree's upper right. The three clades that fail to show taxonomic conservation are displayed along the bottom of the protein tree. The first consists of proteins from organisms shown to biodegrade industrial s-triazine compounds (dark green), a second set of bradyrhizobia (cyan), and a branch that contains Firmicutes (salmon), including M. thermoacetica ATCC 39073 CAH, a gammaproteobacterium (tan), and a cyanobacterium (sea green). The second divergent taxonomy clade on the lower right side of the figure has subclades of the gammaproteobacteria (tan), Firmicutes (salmon), alphaproteobacteria (purple), betaproteobacteria (gray), and a third set of bradyrhizobia (cyan). The third clade, containing nine members, is highly divergent from any sequence of known function and has members from the Cyanobacteria, Actinobacteria, Firmicutes, Alphaproteobacteria, and Betaproteobacteria. The asterisks indicate an experimentally confirmed function. X indicated structures available with A. caulinodans ORS571 (top center, PDB ID 4NQ3) and Pseudomonas sp. ADP (lower left, PDB IDs 4BVQ, 4BVR, 4BVS, and 4BVT).

Most of the clades in the family's protein tree (Fig. 3, top and left) are conserved within specific taxonomic classes, consistent with vertical descent of the genes, with a few random insertions that are interpreted as intermittent horizontal gene transfers. The barbiturases and related proteins are all within the Actinobacteria and are clustered in the tree's upper right. Most of the remaining sequence clades are CAHs, with each clade being represented by at least one experimentally confirmed CAH. The upper part of the tree represents groups of fungi; Actinobacteria; Micromonas; Firmicutes; and Alphaproteobacteria, composed of Rhizobium, a subset of photosynthetic Bradyrhizobium, including Bradyrhizobium sp. strain BTAi1 and Bradyrhizobium sp. strain ORS278, and Methylobacterium.

There are three clades that failed to show this taxonomic conservation, which are displayed along the bottom of the protein tree (Fig. 3). The first contains a subclade of proteins from organisms isolated for their ability to biodegrade industrial s-triazine compounds, including Pseudomonas (gammaproteobacteria), Arthrobacter (actinomycetes), Acidovorax (betaproteobacteria), and Aminobacter (alphaproteobacteria), and whose genes are, in some cases, known to be carried on self-transmissible plasmids. The genes for the proteins were largely identified in soil bacteria and likely spread via horizontal gene transfer. Other subclades of this group are a second set of bradyrhizobia, including Bradyrhizobium diazoefficiens USDA 110, and a more distant branch of this clade that contains Firmicutes, including Moorella thermoacetica ATCC 39073 CAH, which has been well characterized.

The remaining two clades with divergent taxonomies have unknown function. One contains a third set of Bradyrhizobium spp., including B. elkanii, and the other contains nine proteins, five of which have paralogs in their genomes, scattered throughout the protein tree. Two of those five proteins have been cloned and were shown not to have either CAH or barbiturase activity, being consistent with a novel function(s) yet to be discovered.

Three different clades were listed as containing Bradyrhizobium species (Fig. 3). These three clades seem to be divided among the three major superclades of the Bradyrhizobium: the B. japonicum-B. diazoefficiens, B. elkanii, and photosynthetic Aeschynomene symbionts (80). When analyzing Bradyrhizobium taxonomy and species-level identification, the B. elkanii split has been considered an early event (81), followed by the coevolution of plant and bacteria to produce the Nod-independent Aeschynomene-nodulating bradyrhizobia. The Bradyrhizobium species-level identification might mirror the evolution of the CAH family members, with the B. elkanii enzymes being more divergent, while the photosynthetic Aeschynomene bradyrhizobia cluster with other Alphaproteobacteria, and B. japonicum-B. diazoefficiens clusters with the proteins from the s-triazine-degrading bacteria. This observation, the limited number of family members in the entire set of sequenced space, the phylogenetic distribution, the unique protein scaffold (discussed more below), general protein instability, and the fixation of the cyanuric acid degradation genes on chromosomes, plasmids, and, in some cases, within operons, all point to a protein class that was old and dying out; this class, however, received a recent resurgence due to selective pressure imposed by the large-scale industrial production and environmental distribution of s-triazines.

CAH ENZYMOLOGY

Eight CAH family enzymes have been cloned, purified, and characterized, including proteins from Pseudomonas sp. ADP, M. thermoacetica ATCC 39073, B. diazoefficiens (formerly B. japonicum) USDA 110, Rhizobium leguminosarum bv. viciae 3841, Methylobacterium sp. strain 4-46, Sarocladium sp. strain CA, and two paralogs from Azorhizobium caulinodans ORS 571 (53, 58). CAH from additional s-triazine-degrading bacteria have been observed, without complete characterization of the CAH protein (42, 62). Most consist of 340 to 400 amino acids per subunit and have subunit molecular masses of 36 to 43 kDa. Size-exclusion chromatography confirmed dimer and tetramer states of the CAH proteins from Pseudomonas sp. ADP and A. caulinodans ORS 571, respectively (82, 83). From steady-state kinetic analysis, it has been determined that k_cat values of the seven proteins range from 3 to 73 s⁻¹, and K_m values range between 23 and 400 μM (53, 58). The percent identities of pairwise sequence comparisons between experimentally confirmed CAH proteins are as low as 32%, indicative of extensive divergence within the CAH proteins (53, 58). Yet, for most CAH enzymes, the k_cat/K_m values are between 10⁵ and 10⁶ M⁻¹ s⁻¹. The reasonable k_cat/K_m values and the fact that other compounds tested were shown not to be reactive are consistent with the idea that cyanuric acid is the native substrate for these enzymes. In comparison, the barbiturases and barbiturase-like proteins share >67% identity in pairwise sequence comparisons. The one purified and characterized barbiturase has a k_cat/K_m of 1.6 × 10³ M⁻¹ s⁻¹, but despite the modest activity, the genomic and physiological contexts strongly support the idea that barbituric acid is the native substrate for these proteins (53, 55, 56).

CAH proteins were shown not to need a metal for activity (53). Moreover, they are generally unstable for storage and manipulation, which are undesirable properties for applying these enzymes for bioremediation. However, a CAH from the thermophile M. thermoacetica ATCC 39073 was identified and has superior temperature stability in both ambient and elevated temperatures, thereby showing more potential for bioremediation applications (59). Family analysis failed to link CAH proteins with other proteins that were more extensively characterized (53), resulting in limited mechanistic insights until X-ray crystallographic studies provided details of the active structure (82–84).

CAH STRUCTURES

The crystal structures of CAH from A. caulinodans ORS571 (PDB identification [ID] 4NQ3) and Pseudomonas sp. ADP (PDB IDs 4BVQ, 4BVR, 4BVS, and 4BVT) revealed that the enzyme consists of three structurally homologous domains that form a β-barrel-like structure with external α-helices (Fig. 4) (82, 83). The 3-fold symmetry of the protein mirrors the trifold symmetry of the cyanuric acid substrate and resulted in a trimeric active site with serine, lysines, arginines, and glycines symmetrically contributed by each domain. Although the three sections of the protein share only 13 to 17% sequence identity, the reflection of active-site residues in each third of the protein suggests that gene duplication occurred in the distant past, followed by gene fusion and sequence divergence to create the modern three-domain barrel structure (82, 83). This active-site structure positions three serines as potential nucleophiles, each with an activating lysine residue to form three Ser-Lys dyads (Fig. 4). Bioinformatic and mutagenic studies with the Azorhizobium protein focused attention on S226 as the potential serine nucleophile, activated by K156, with sole sequence conservation across the entire family and a 20- to 40-fold-greater decrease in activity compared to that of the other two serine mutants.

FIG 4.

A. caulinodans ORS571 CAH crystal structure (PDB ID 4NQ3). (A) Trimeric domains are each colored differently (blue, residues 1 to 99; magenta, residues 104 to 245; cyan, residues 246 to 355). The trimeric active site is also shown with a Ser-Lys dyad (red and brown, respectively) from each of the domains. Bound barbituric acid is shown colored by atom type, with carbon shown in gray, oxygen in red, nitrogen in blue, and hydrogen in white. (B) Enlargement of the Ser-Lys dyads (red and brown, respectively). (C) Overlay of each of the trimeric domains from CAH and AroH chorismate mutase (light green, PDB ID 1XHO).

The closest related structure to the CAH fold is the AroH-type chorismate mutase (CM) trimer (PDB ID 1XHO; Dali Z-score, 7.8) (83). CM catalyzes the Claisen rearrangement of chorismate to prephenate, using a pericyclic mechanism and an active site that stabilizes a polar chair-like transition state, allowing for the million-fold rate enhancement provided by the enzyme over chemical conversion (85, 86). Each of the CM monomers is equivalent to a domain within CAH. CM, however, contains its active site on the exterior face of the barrel at subunit interfaces, rather than in the interior of the barrel, like CAH. The positional difference in the active site and the absence of sequence linkages make it impossible to determine whether these two proteins are related via divergent or convergent evolution. Like CAH, AroH has limited prevalence in nature, where it is mainly found in Gram-positive bacteria of the Bacillus/Clostridia group and a few Gram-negative bacteria (87). Unlike CAH however, CM catalyzes a critical activity, catalyzing the first committed step in the biosynthesis of aromatic amino acids. The AroQ-type chorismate mutase, which consists of a four-helix bundle, is much more prevalent, and it has been hypothesized that the AroQ CM is in fact displacing the AroH CM (88).

These data were interpreted to suggest that this type of β-barrel scaffold is very rare in nature and perhaps has been shrinking in prevalence. Most of the cloned CAH and barbiturase proteins are highly unstable with changing temperature and storage conditions (53), and genomics has shown that they are sparsely distributed in bacteria and not found in plants, animals, or archaea. These observations support the idea that the CAH/barbiturases are an exiguous protein family, which may have experienced a resurgence, with recent selection-driven horizontal transfer of the CAH gene upon the introduction of anthropogenic s-triazine compounds.

INDUSTRIAL APPLICATIONS FOR CAH

A major application of industrially produced cyanuric acid is for the disinfection of swimming pools, spas, and other waters for which it is important to maintain chlorine disinfection that might otherwise decline rapidly due to sunlight. Either by stabilizing hypochlorous acid and its conjugate base, hypochlorite (chlorine bleach), directly or by adding trichloroisocyanuric acid, cyanuric acid can be used to maintain chlorine disinfection (9, 89).

When di- or trichloroisocyanuric acid is used directly for disinfection and added repeatedly upon chlorine dissipation, cyanuric acid accumulates to levels that shift the chemical equilibrium, reducing the level of free chlorine in the water and impairing the disinfection process. This necessitates that the water be discarded and new fresh water drawn, which is a stress on water resources in hot dry areas where swimming pools are most prevalent. The removal of the cyanuric acid in situ would allow for a longer life span of the pool water, thereby reducing the time, effort, and natural resources required for water exchange. A natural solution to this problem is the use of CAH to degrade cyanuric acid. With that purpose in mind, CAH has been encapsulated in a porous silica matrix (89) and experimentally tested in a bioreactive pool filter. The M. thermoacetica ATCC 39073 CAH gene was cloned into Escherichia coli and heated to ensure cell nonviability while maintaining CAH functionality. The encapsulated CAH was capable of degrading 10,000 μM cyanuric acid by 70% in 24 h and 100% in 72 h.

An impediment to developing a solution for swimming pool water purification is the susceptibility of the CAH enzyme to inactivation due to oxidization by hypochlorite and hypochlorous acid, which are present in the chlorinated waters (90). Yeom et al. showed that CAH activity declined as a function of increasing chlorine levels in the municipal pool waters that were treated (89). A recent study showed that properly prepared 3-aminopropyltriethoxysilane (APTES)-encapsulated cells could withstand hypochlorite concentrations up to 10 ppm, providing evidence that this could be an effective biohybrid filter medium for treating chlorinated swimming pool and other waters containing cyanuric acid (91). The APTES-encapsulated cells also showed increased permeability, allowing for degradation rates of 29 μmol/min per mg of CAH protein, rivaling purified enzyme rates.

CONCLUSIONS

Cyanuric acid hydrolase enzymes are likely to have an ancient origin, with a structural scaffold that is rare, but the recent influx of anthropogenic s-triazines into the environment is thought to have provided new selective pressure and driven greater dissemination of the enzyme by horizontal gene transfer. The enzyme has a rare and unusual protein fold, as demonstrated by recent X-ray crystallographic data. Cyanuric acid hydrolases open the cyanuric acid ring by promoting the nucleophilic attack of a serine residue on the substrate and subsequent hydrolysis of the enzyme-serine covalent intermediate to yield the ultimate product, biuret. A cyanuric acid hydrolase from the thermophile M. thermoacetica ATCC 39073 is currently being investigated for biotechnological application in the removal of a disinfection by-product from swimming pool waters.

Funding Statement

This work was supported by grants to Lawrence P. Wackett from the Biocatalysis and Synthetic Ecology Initiatives of the BioTechnology Institute at the University of Minnesota, National Science Foundation Partnerships for Innovation grant 1237754, and a grant from the Office of the Vice President for Research MnDRIVE Initiative at the University of Minnesota.