A Functional Analysis of the Purine Salvage Pathway in Acetobacter fabarum

ABSTRACT

Acetobacter species are a major component of the gut microbiome of the fruit fly Drosophila melanogaster, a widely used model organism. While a range of studies have illuminated impacts of Acetobacter on their hosts, less is known about how association with the host impacts bacteria. A previous study identified that a purine salvage locus was commonly found in Acetobacter associated with Drosophila. In this study, we sought to verify the functions of predicted purine salvage genes in Acetobacter fabarum DsW_054 and to test the hypothesis that these bacteria can utilize host metabolites as a sole source of nitrogen. Targeted gene deletion and complementation experiments confirmed that genes encoding xanthine dehydrogenase (xdhB), urate hydroxylase (urhA), and allantoinase (puuE) were required for growth on their respective substrates as the sole source of nitrogen. Utilization of urate by Acetobacter is significant because this substrate is the major nitrogenous waste product of Drosophila, and its accumulation in the excretory system is detrimental to both flies and humans. The potential significance of our findings for host purine homeostasis and health are discussed, as are the implications for interactions among microbiota members, which differ in their capacity to utilize host metabolites for nitrogen.

IMPORTANCE Acetobacter are commonly found in the gut microbiota of fruit flies, including Drosophila melanogaster. We evaluated the function of purine salvage genes in Acetobacter fabarum to test the hypothesis that this bacterium can utilize host metabolites as a source of nitrogen. Our results identify functions for three genes required for growth on urate, a major host waste product. The utilization of this and other Drosophila metabolites by gut bacteria may play a role in their survival in the host environment. Future research into how microbial metabolism impacts host purine homeostasis may lead to therapies because urate accumulation in the excretory system is detrimental to flies and humans.

INTRODUCTION

Acetic acid bacteria (AAB) constitute a diverse group of Alphaproteobacteria found in a wide range of environments and used industrially for vinegar production as well as other processes. AAB of the genus Acetobacter are common in sugar-rich substrates, especially those containing ethanol, e.g., fermenting fruits and vegetables (1, 2). They are also constituents of the gut microbiota of animals that feed on such substrates, including the fruit flies Drosophila melanogaster and Drosophila suzukii (3–6). Studying this group can improve our understanding of host-microbe interactions and potentially enhance the utilization of AAB in biotechnology.

It has been hypothesized that the ecology and evolution of Acetobacter species is shaped by their association with insect hosts (7–9). Drosophila can provide a means of transportation between ephemeral growth substrates and also exerts selective pressures on microbiota via digestion and immunity (10–13). Research on Lactobacillus plantarum, another prominent gut microbe in Drosophila, showed that mutations improving growth in the host’s diet were key to the bacterium’s adaptation to the host environment in a laboratory evolution experiment (14). Other studies found that application of experimental evolution regimes to the host leads to changes in Drosophila microbiome composition, emphasizing the significance of the microbial response to changes in the host environment (15, 16). A growing body of work has developed Drosophila as a successful model for illuminating the impacts of gut bacteria on their animal hosts (17–19), yet our understanding of how this relationship impacts the bacteria is still relatively limited.

A central question in microbiome research is, how do bacteria adapt to association with animal hosts? Put another way, what selective pressures are relevant in the host environment, and how do they shape the microbiome over different timescales? One emerging theme in this line of inquiry is that hosts can select for certain metabolic capabilities in their microbial partners through the provision of particular nutrients or growth substrates (20, 21). These substrates can be consumed as part of the host’s diet or may be molecules produced by the host, e.g., glycans in the mammalian gut (22) or chitin in the light organ of the Hawaiian bobtail squid (23). Provisioning of substrates to the microbiome can shape the composition of the microbial community, as seen in the mammalian gut, or even determine the timing of bacterial colonization and metabolic output, as seen in squid (24). Insight into microbial adaptation to the host environment, including the metabolism of host-provisioned substrates, is essential to understanding how microbiomes function and to the development of microbiome therapies for human disease (25, 26). The Drosophila microbiome model is a novel area for research on this topic.

In a prior study, we undertook a comparative genomic analysis of Acetobacter species to identify genetic signatures of adaptation to the host environment (27). We compared Acetobacteraceae isolated from laboratory-reared flies, wild-caught flies, as well as food and industrial sources. Among these groups, the most significant genetic differentiation was between isolates from wild and laboratory-reared Drosophila. Lab isolates had all lost flagellar motility genes, and all retained purine salvage genes. Wild fly isolates, in contrast, were mixed in terms of whether they had these loci. The results suggested that persistent association with Drosophila in the lab environment provides selective pressure against motility and for purine utilization (27). However, the functionality of purine salvage genes was not tested, and the ability of Acetobacter to use host metabolites was only suggested.

The possibility of purine utilization by the gut microbiota in Drosophila is noteworthy for several reasons. The major nitrogenous waste product of the fly is a purine, uric acid (in the form of urate), which is excreted through the Malpighian tubules and into the hindgut for elimination in feces (28). Drosophila can express urate oxidase to convert urate to allantoin, but for unknown reasons, urate predominates (29, 30). Several studies have linked either urate or allantoin accumulation to shortened life span in flies, and Drosophila has been employed as an experimental model for hyperuricemia and other diseases affecting the excretory system (29–31). Lang et al. showed that Drosophila accumulates urate crystals in the Malpighian tubules and hindgut when fed a high-purine diet; this effect was exacerbated by knockdown of urate oxidase (29). While the authors implicated insulin signaling and reactive oxygen species as drivers of urate accumulation, they did not investigate any role of microbiota in this process. Yamauchi et al. found increased levels of allantoin correlate with reduced life span in flies but were not caused by changes in urate oxidase expression (30). Instead, they showed that activation of the immune deficiency (IMD) pathway by microbiota member Acetobacter persici drives these changes. They found L. plantarum (but not A. persici) reduced the purine content of the diet, suggesting that microbial purine utilization can also alter purine metabolic homeostasis in flies by limiting intake (30). In summary, purine waste homeostasis impacts Drosophila health and longevity and may be influenced by microbiota at multiple levels. More research on purine utilization by Drosophila microbiome members is needed to clarify the picture.

In this study, we used targeted mutagenesis and complementation to test the function of purine utilization genes in Acetobacter fabarum DsW_054, a strain isolated from wild-caught Drosophila suzukii (27). The purine salvage locus found in this strain is conserved in many Acetobacter species (27) but had yet to be functionally characterized in any AAB. Gene deletion and complementation have not been widely used in Acetobacter, but see references 32 to 35. Acetobacter species associated with Drosophila have only been modified with transposon mutagenesis (36, 37) and overexpression (38), but not via allelic exchange. In this study, we employed an allelic exchange strategy to make unmarked, in-frame deletions in A. fabarum. We confirmed the function of three genes, which encode important enzymatic steps in purine utilization by this bacterium.

RESULTS

Growth of Acetobacter with hypoxanthine as a sole source of nitrogen.

Many bacteria are capable of breaking down purine nucleotides as a source of nitrogen (39, 40). Our prior research found a putative purine salvage locus present in the genomes of some, but not all, Acetobacter species isolated from Drosophila (27) (Fig. 1). To correlate the presence of this locus with the ability to break down purines, we tested the capacity of Acetobacter isolates to grow on hypoxanthine as a sole source of nitrogen in M9-lactate medium. Media with NH₄Cl or allantoin as nitrogen sources were also tested. We found 14 of 17 strains were capable of growing in minimal medium with inorganic nitrogen (NH₄Cl) (Table 1). Of these, 9 could also grow on hypoxanthine as a sole source of nitrogen. We saw close, but not complete, concordance between the presence of the purine salvage locus and the ability to grow on hypoxanthine: Acetobacter malorum DsW_057 was able to grow on hypoxanthine despite lacking the locus, and Acetobacter okinawensis DsW_060 was not able to grow despite having it (Table 1). Aside from these inconsistencies, having the locus was a good predictor of growth on hypoxanthine based on Fisher’s exact test (P < 0.01). A subset of strains with the purine salvage locus was also able to grow on allantoin, an intermediate in the pathway (Fig. 1B) and the ultimate by-product of Drosophila purine metabolism.

FIG 1.

Purine utilization in Acetobacter fabarum DsW_054. (A) Purine utilization locus with mutations described in this study diagrammed on the top and genes labeled on the bottom. Deletion mutations are depicted as black bars on gene arrows, with nucleotide changes summarized above. Numbers indicate base pair positions relative to the start of the reading frame, showing the removal of bases in mutants; new stop codons added via allelic exchange are highlighted in red. (B) Predicted gene functions are summarized except for guaD (guanine deaminase), a LysR-type regulator (unlabeled black arrow) and predicted permease (white arrow). (B, Left) a summary of the purine salvage pathway is shown; (B, Middle) enzyme(s) responsible for each step; (B, Right) genes predicted to encode the enzymes. Gene names were assigned based on protein sequence similarity to characterized homologs and are listed next to NCBI locus tags.

TABLE 1.

Growth of Acetobacter strains in M9-lactate media with different nitrogen sources^a

Strain	Data for:
	Salvage locus	NH4Cl	Hypoxanthine	Allantoin
Acetobacter sp. strain DmW_043	0	0	0	0
Acetobacter persici DmL_053	0	0	0	0
Acetobacter sp. DsW_059	0	0	0	0
Acetobacter okinawensis DsW_060	1	1	0	0
Acetobacter orientalis DsW_061	0	1	0	0
Acetobacter nitrogenifigens DsW_063	0	1	0	0
Acetobacter orientalis DmW_045	0	1	0	0
Acetobacter cibinongensis DmW_047	0	1	0	0
Acetobacter pomorum DmCS_004	1	1	1	0
Acetobacter malorum DmCS_006	1	1	1	0
Acetobacter fabarum DsW_054	1	1	1	0
Acetobacter malorum DsW_057	0	1	1	0
Acetobacter tropicalis DmCS_005	1	1	1	1
Acetobacter tropicalis DmW_042	1	1	1	1
Acetobacter indonesiensis DmW_046	1	1	1	1
Acetobacter tropicalis DmL_050	1	1	1	1
Acetobacter indonesiensis DmL_051	1	1	1	1

^aPresence (1) or absence (0) of the purine salvage locus was determined by genomic analyses in reference 27. Growth in M9-lactate media with different nitrogen sources was assessed visually after 48 h with any discernible increase in turbidity scored as 1 and no change scored as 0. Rows progress from strains that grew on the least number of nitrogen sources to those that grew on all of them.

Genetic investigation of purine salvage in A. fabarum.

We adopted a genetic approach to test the function of genes in the purine salvage locus of A. fabarum DsW_054, which encodes several enzymes required for the recovery of nitrogen from purines (Fig. 1). These include homologs of xdhABC encoding xanthine dehydrogenase, which converts hypoxanthine and xanthine to urate in the first committed step of the pathway. An allelic exchange strategy was used to introduce an unmarked deletion within the xdhB reading frame to see if it was required for growth on hypoxanthine (Fig. 1A; see Materials and Methods for details). Wild-type A. fabarum (WT) and mutant strains were grown in M9-lactate media with hypoxanthine or urate as sources of nitrogen. Inactivation of xanthine dehydrogenase in ΔxdhB blocked growth on hypoxanthine but not urate (Fig. 2A), which is consistent with our conception of the pathway. Reintroducing the xanthine dehydrogenase genes (xdhABC) on a plasmid complemented the mutant, restoring growth on hypoxanthine.

FIG 2.

Genetic analysis of purine utilization. Bar graphs summarize cell density measurements at 600 nm of cultures in M9-lactate media. Data are means ± standard deviation; n = 9. Nitrogen sources used include hypoxanthine (HX), urate (UA), and ammonium chloride (NH₄). (A) Comparison of WT to the xanthine dehydrogenase mutant ΔxdhB with either empty pCM62 (p62) or pCM62 with genes xdhABC (p62_XDH). (B) Growth phenotypes of two Tn5 insertion mutants in the permease gene on three different nitrogen sources, compared to WT.

Upstream of xdhABC, there is a predicted permease gene (Fig. 1A). We tested the function of this gene via two independent Tn5 insertion mutants from an arrayed, sequenced transposon mutant library (37). These mutations significantly reduced the ability of A. fabarum to grow on hypoxanthine as a sole source of nitrogen (Fig. 2B). Permease mutations did not impact growth on urate, however (Fig. 2B), suggesting that this transporter acts on hypoxanthine but not urate.

urhA is required for urate oxidation.

The next step in purine salvage is the oxidation of urate to 5-hydroxyisourate (HIU) (Fig. 1B) (41). Our prior study referred to gene HK20_RS08215 of A. fabarum as the urate oxidase responsible for this step based on RAST annotation (27). However, BLASTp and Conserved Domain Database searches (NCBI) identified HK20_RS08215 as homologous to the allantoinase PuuE (42), an enzyme that functions later in the pathway (Fig. 1B). BLASTp searches of A. fabarum and other Acetobacter genomes did not return matches to any characterized urate oxidases despite the presence of genes HIUase (pucM) and 2-oxo-4-hydroxy-4-carboxy-5-ureidoimidazoline decarboxylase (or OHCUdc; pucL) for the subsequent conversion of HIU to allantoin. Others have shown that several nonorthologous enzymes can oxidize urate (40, 43, 44). Of these enzymes, the urate hydroxylase HpyO of Xanthamonas shares 59% amino acid similarity to a predicted oxidoreductase encoded just upstream of puuE in A. fabarum as well as other Acetobacter species (Fig. 1A). We hypothesized that this oxidoreductase gene encodes the enzyme responsible for urate oxidation in the purine salvage pathway. To test this, we constructed an internal deletion in this reading frame, introducing a stop codon as the third codon. The resulting mutant was unable to grow on hypoxanthine or urate as sole sources of nitrogen, but reintroducing the gene on a plasmid restored these abilities (Fig. 3A). Given this evidence, we named the gene urhA for urate hydroxylase of Acetobacter.

FIG 3.

UrhA functions in urate degradation. Bar graphs summarize cell density measurements at 600 nm of cultures in M9-lactate media. Data are means ± standard deviation; n = 9. Nitrogen sources used include hypoxanthine (HX), urate (UA), and allantoin (AL). (A) Comparison of WT to the urate hydroxylase mutant ΔurhA with empty pCM62 (p62) or pCM62 bearing urhA (p62_urhA). (B) Kinetic assay of growth and urate utilization in M9-lactate with 2 mM urate as the source of nitrogen. Solid lines show cell density, and dashed lines show urate concentration in cultures of the indicated strains. Mean ± standard deviation is shown; n = 9. (C) Heterologous expression of genes from A. fabarum in A. orientalis DsW_061; p62_puuE encodes predicted allantoinase, and p62_urhAOp encodes four genes from the locus (depicted above and described in Fig. 1).

To provide additional evidence that UrhA degrades urate, we conducted a kinetic assay correlating the growth of A. fabarum strains in minimal media with the depletion of urate. Increases in the cell density of the WT and complemented ΔurhA mutant were accompanied by concomitant decreases in urate concentration (Fig. 3B). The ΔurhA pCM62 control showed no change in cell density or urate concentration.

We also tested whether heterologous expression of urhA and other genes in the locus could confer the ability to metabolize urate on other bacteria. Acetobacter orientalis DsW_061 is unable to grow on hypoxanthine or allantoin (Table 1) and lacks the first five enzymes in the purine salvage pathway (listed in Fig. 1B). Interestingly, its genome does encode an allantoate amidohydrolase, ureidoglycine transaminase, and urease. This suggests that acquisition of just one gene by DsW_061, allantoinase puuE, could enable growth on allantoin, and acquisition of genes urhA, pucM, pucL, and puuE could enable growth on urate. We tested this by introducing two different plasmids to DsW_061, one with all four genes and one with only puuE. Consistent with our prediction, both constructs enabled DsW_061 to grow on allantoin, but only the construct that included urhA enabled growth on urate (Fig. 3C). This supports the conclusion that urhA encodes a urate hydroxylase and is consistent with the predicted functions of the other genes in the locus.

puuE is required for growth on purines and allantoin.

The fact that introduction of puuE to A. orientalis DsW_061 enabled growth on allantoin provides good evidence that this A. fabarum gene encodes a functional allantoinase. To further test this, we disrupted the gene with a small internal deletion and introduced stop codons into the reading frame (Fig. 1A). The ΔpuuE mutant was unable to grow on either hypoxanthine or urate, but growth was restored on both substrates via complementation with the puuE gene (Fig. 4A). These results are consistent with our conception of the purine salvage pathway and the requirement of allantoinase for nitrogen recovery (Fig. 1B). Interestingly, we did not observe growth of WT A. fabarum when allantoin was provided as the sole source of nitrogen despite the fact that allantoin is an intermediate in the salvage pathway (Table 1). We hypothesized that A. fabarum is unable to transport allantoin. To test this, we cloned an allantoin permease gene from Escherichia coli, allP (45), and introduced it to A. fabarum on a plasmid. This construct allowed the WT strain to grow on allantoin, but not the ΔpuuE mutant, consistent with puuE encoding allantoinase (Fig. 4B). Addition of pallP to the ΔurhA strain enabled growth on allantoin, but with delayed kinetics compared to the WT (Fig. 4B). These data are consistent with the prediction that urhA is responsible for urate degradation, which is not required for growth on allantoin. The longer lag phase exhibited by this strain could result from the ΔurhA mutation impacting the regulation or expression of puuE downstream (Fig. 1A).

FIG 4.

Functional analysis of the predicted allantoinase puuE. (A) Bar graph summarizing cell density measurements of cultures in M9-lactate media. Data are means ± standard deviation; n = 9. Nitrogen sources used include hypoxanthine (HX) and urate (UA). WT is compared to the allantoinase mutant ΔpuuE with either empty pCM66 (p66) or pCM66 bearing puuE (p66_puuE). (B) Cell density measurements of liquid cultures over time, grown in M9-lactate containing allantoin as the sole source of nitrogen. Strains without plasmids (top of the legend) were compared to those with pCM62 bearing allP, the allantoin permease gene from E. coli (bottom of the legend). Strains without plasmids did not exhibit growth, and the corresponding curves are obscured at the bottom of the graph. Data are means ± standard deviation; n = 9.

In summary, genetic analyses of the purine salvage locus, coupled with growth experiments in minimal media, confirmed the predicted functions of several key genes in this pathway. They indicate that xdhB, urhA, and puuE are required for growth on purines as a sole source of nitrogen. Additionally, our results suggest A. fabarum has transporters for hypoxanthine and urate, but not allantoin.

DISCUSSION

This study demonstrates that Acetobacter can utilize a purine salvage pathway to grow on urate as a sole source of nitrogen. The results support the hypothesis that these bacteria can consume purine metabolites produced by Drosophila. Our genetic approach confirmed the function of three genes in the purine salvage locus, conserved in many Acetobacter isolates. Here, we discuss the significance of bacterial purine utilization to host health and to interactions among members of the microbiome.

Urate is the predominant nitrogenous waste excreted by Drosophila. Its accumulation can significantly shorten life span (29), and so can that of its breakdown product, allantoin (30). As noted in the introduction, Drosophila can convert urate to allantoin with its own urate oxidase, but prior studies have found they produce mainly urate (29, 30). Yamauchi et al. showed that microbiota can modulate levels of allantoin in Drosophila by either utilizing purines in the diet (thereby reducing host consumption of purines) or by influencing host production of allantoin (30). Our results suggest a third way: gut bacteria could reduce levels of urate and/or allantoin by consuming these waste products.

While it seems clear that gut microbiota impact purine homeostasis in Drosophila, this relationship is likely complex, with factors such as diet, host age, and microbiome composition all playing a role in determining the outcome. A number of studies have reported diet-dependent changes in urate concentrations in Drosophila (46, 47), but they did not control for the composition of the gut microbiota, which itself can be modified by diet (48). For Yamauchi et al., A. persici drove an age-dependent increase in allantoin via activation of the IMD pathway (30). However, A. persici is not able to utilize purines (Table 1) (30), leaving open the question as to whether an Acetobacter species that could utilize purines would have the same impact on the host. Future experiments using defined diets and gnotobiotic animals are needed to tease apart the relative contributions of host consumption, host metabolism, and bacterial metabolism on purine homeostasis in Drosophila. These advances will maximize the utility of Drosophila as a biomedical research model for renal disease (31).

We found that Acetobacter species varied in their ability to utilize purines and allantoin, with some only able to use the former and not the latter. This is significant because the host itself can convert urate to allantoin. This conversion could, theoretically, benefit certain species over others in the gut microbiome if they are dependent on these host-provisioned substrates for nitrogen. Our data show that A. fabarum is unable to use allantoin due to the absence of a transporter (Fig. 4), suggesting that this could be the deciding factor for whether Acetobacter species with the purine salvage locus can utilize allantoin. Is it possible that some species specialize in urate utilization and others in allantoin? Such a scenario could provide a mechanism to partition the niche and promote Acetobacter diversity in the microbiome. An analogous situation has been well established for Bacteroides species in the human gut, whose diverse polysaccharide utilization profiles promote coexistence and even cooperation among closely related species (22, 49). Both empirical and modeling data have shown significant and complex metabolic interactions between members of the Drosophila microbiome, which can significantly impact host traits and fitness (50–53). Future research will delve into the impacts of host-provisioned urate and allantoin on interactions among Drosophila microbiome species.

It is important to note the possibility that association with bacteria that metabolize urate could allow Drosophila to benefit from nitrogen waste recycling (NWR), where microbiota use inaccessible nitrogen waste to produce amino acids that are accessible to the host. NWR is common in symbioses between microbes and insects, including some ants (54) and xylem-feeding insects with intracellular symbionts (55). In Drosophila, the possibility of NWR would be most relevant to larvae developing in nutrient-poor substrates where reingestion of gut microbes can provide a significant contribution to protein nutrition (56, 57).

We provide several lines of genetic evidence that urhA encodes an enzyme that oxidizes urate. The sequence similarity of UrhA to HpyO of Xanthamonas suggests it may have a similar reaction mechanism; HpyO requires FAD as a cofactor and NAD(P)H as a cosubstrate (44). A Phyre2 search (58) with the UrhA sequence matched its structure to that of a cyclohexanone monooxygenase with 100% confidence. This enzyme, from Rhodococcus, has one FAD and two NADP cofactors, and it catalyzes the addition of oxygen to cyclic ketones; its biochemistry is consistent with the transfer of an oxygen atom from dioxygen to the substrate and the formation of water as by-product (59). Further research into the biochemistry of UrhA and other members of this family is needed to test if a similar mechanism is used by urate hydroxylases. Enzymes that break down urate are of interest as possible therapeutics for gout and other conditions linked to hyperuricemia in humans (60). The fact that urate hydroxylases in the HpyO family do not produce hydrogen peroxide as a by-product is a potential benefit in their development for this purpose.

In summary, the purine salvage pathway allows Acetobacter to use urate, or other purine metabolites, as a sole source of nitrogen. Future studies will examine whether this capacity contributes to the fitness of the bacteria in the host environment. The impact of microbial urate metabolism on host physiology and microbiota interactions also represent exciting areas for further investigation. Finally, the allelic exchange tools implemented in this study enhance the utility of Acetobacter as a model for studying host-microbe interactions.

MATERIALS AND METHODS

Bacteria strains and growth conditions.

Bacterial strains used are listed in Table 2; all were cultured at 30°C with shaking for liquid cultures (250 rpm). E. coli was grown in LB (10 g L⁻¹ tryptone, 5 g L⁻¹ yeast extract, and 5 g L⁻¹ NaCl) with antibiotics at the following concentrations, when necessary: 150 μg mL ⁻¹ ampicillin, 50 μg m L⁻¹ kanamycin, and 10 μg mL⁻¹ tetracycline. S. cerevisiae was grown on YPD-lite (8 g L⁻¹ yeast extract, 8 g L⁻¹ peptone, 8 g L⁻¹ dextrose, and 16 g L⁻¹ agar) or complete supplemental mixture without uracil (20 g L⁻¹ dextrose, 20 g L⁻¹ agar, 6.7 g L⁻¹ yeast nitrogen base, and 2 g L⁻¹ complete supplement mixture [CSM]-uracil). A. fabarum was cultured in YPD-lite with antibiotics at the following concentrations, when necessary: 50 μg mL⁻¹ kanamycin and 20 μg mL⁻¹ tetracycline.

TABLE 2.

Strains and plasmids used in this study

Strain or plasmid	Relevant characteristics	Source or reference
Strains
Escherichia coli S17-λpir	Conjugation donor	G. O’Toole
Escherichia coli K-12 MG165	Wild type	J. Peters
Saccharomyces cerevisiae URA	Uracil auxotroph	G. O’Toole
Acetobacter fabarum DsW_054	From wild D. suzukii	27
A. fabarum ΔurhA	Urate hydroxylase mutant	This study
A. fabarum ΔpuuE	Allantoinase mutant	This study
A. fabarum ΔxdhB	xdhB gene mutant	This study
A. fabarum 12-H9	Tn5::HK20_RS08230	37
A. fabarum 48-B2	Tn5::HK20_RS08230	37
Acetobacter pomorum DmCS_004	From lab D. melanogaster	64
Acetobacter tropicalis DmCS_005	From lab D. melanogaster	64
Acetobacter malorum DmCS_006	From lab D. melanogaster	64
Acetobacter tropicalis DmW_042	From wild D. melanogaster	27
Acetobacter sp. strain DmW_043	From wild D. melanogaster	27
Acetobacter orientalis DmW_045	From wild D. melanogaster	27
Acetobacter indonesiensis DmW_046	From wild D. melanogaster	27
Acetobacter cibinongensis DmW_047	From wild D. melanogaster	27
Acetobacter tropicalis DmL_050	From lab D. melanogaster	27
Acetobacter indonesiensis DmL_051	From lab D. melanogaster	27
Acetobacter persici DmL_053	From lab D. melanogaster	27
Acetobacter malorum DsW_057	From wild D. suzukii	27
Acetobacter sp. strain DsW_059	From wild D. suzukii	27
Acetobacter okinawensis DsW_060	From wild D. suzukii	27
Acetobacter orientalis DsW_061	From wild D. suzukii	27
Acetobacter nitrogenifigens DsW_063	From wild D. suzukii	27
Lactobacillus brevis DmCS_003	From lab D. melanogaster	64
Plasmids
pMQ460	Kanr, I-SceI	63
pMQ460-xdhBKO	xdhB deletion construct	This study
pMQ460-OxiKO	Uricase deletion construct	This study
pMQ460-PuuEKO	Allantoinase deletion construct	This study
pMQ337	AraC PBAD SecI	63
pCM62	Tetr, replicates in A. fabarum	65
pCM66	Kanr, replicates in A. fabarum	65
pDN5	Tetr, AraC PBAD SceI	This study
p62_urhA	pCM62 plus uricase gene	This study
p66_puuE	pCM66 plus allantoinase gene	This study
p62_puuE	pCM62 plus allantoinase gene	This study
p62_XDH	pCM62 plus xdhABC	This study
p62_urhAOp	pCM62 plus four-gene operon	This study
p62_allP	pCM62 plus allantoin permease	This study

Chemically defined medium experiments.

Media were assembled from sterile components, including autoclaved 5× M9 salts base, 1 M MgSO₄, 1 M CaCl₂, deionized water, and filter-sterilized 20% lactic acid. Final concentrations were 0.78 g L⁻¹ Na₂HPO₄, 0.3 g L⁻¹ KH₂PO₄, 0.05 g L⁻¹ NaCl, 0.1 g L⁻¹ NH₄Cl, 4 mM MgSO₄, 0.2 mM CaCl₂, and 1% lactic acid. Alternative nitrogen sources were added to replace NH₄Cl at a concentration of 0.125%. They were allantoin (Sigma-Aldrich; catalog no. 05670), sodium hypoxanthine (MP Biomedicals; catalog no. 0210545105), and sodium urate (Sigma-Aldrich; catalog no. U2875). Cultures in 2 mL of defined medium were inoculated with 5 μL of an overnight culture of Acetobacter and then grown with shaking (250 rpm) at 30°C for 24 h (or longer as indicated). Cell density was observed visually or quantified by absorbance at 600 nm. The kinetic assay was performed with M9-lactate containing 2 mM urate. To determine urate concentration in the culture supernatant, cells were pelleted at 16,000 × g for 2 min, and absorbance of the supernatant measured at 290 nm using a Nanodrop spectrophotometer (Thermo). Absorbance values were converted to concentration using a standard curve.

Plasmid and strain construction.

Allelic exchange constructs targeting xdhB, urhA, and puuE were built in pMQ460 using recombination cloning in yeast (61–63). Briefly, two DNA fragments of about 1 kb were amplified by PCR, one upstream and one downstream of the chromosomal site to be modified, and then cloned adjacent to each other in pMQ460. The primers used included additional nonannealing bases to facilitate homologous recombination with the vector and each other (Table 3). For xdhB, the construct excluded 828 bp of the reading frame. For urhA, the construct excluded 77 bp of the reading frame and added a stop codon in the third codon position. In the case of puuE, the construct excluded 35 bp of the reading frame and added two stop codons in their place. The allelic exchange constructs were introduced into A. fabarum via conjugation with E. coli S17-λpir, and single-crossover integrants selected for on 50 μg mL⁻¹ kanamycin and 0.2% acetic acid plates. Integrations were confirmed by PCR by using one primer annealing to chromosomal sequence outside the construct and one on the pMQ460 backbone. The pMQ460 plasmid includes a site for the I-SceI nuclease (63). To select for cells that have eliminated pMQ460 via two recombination events, the nuclease was introduced on a separate plasmid, causing a double-stranded break. We constructed pDN5 for this purpose, PCR amplifying araC-P_BAD-I-SceI from pMQ337 and cloning it into pCM62, a plasmid that can replicate in Acetobacter (38). Strains with a confirmed integration of a pMQ460 construct were cultured overnight without selection, and then pDN5 was introduced via conjugation and selection on 20 μg mL⁻¹ tetracycline, 0.2% acetic acid, and 0.2% arabinose. Loss of pMQ460 and successful allelic exchange in the resulting strains were confirmed by kanamycin sensitivity, PCR, and sequencing. A single overnight passage without tetracycline was sufficient to cure the mutant strains of pDN5.

TABLE 3.

Oligonucleotide primers used in this study^a

Primer name	Sequence
urhAKO up Fwd	ttgcatgcctgcaggtcgactctagaggatccCTTCCAGCACCATGGATTTGATCATCTG
urhAKO up Rev	ccgttgtgctgatggttattCCATTTATCAGTCCACATTATGCAGATCGG
urhAKO dwn Fwd	actgataaatggaataaCCATCAGCACAACGGACGGGATG
urhAKO dwn Rev	acagctatgaccatgattacgaattcgagctcgCTGTGCCAATGATCAGGAAGTCGAAC
urhAKO verify	GCAGGAGGTCAGGAGCAGTTCC
urhA express fwd BamHI	aaaaggatccGTGGACTGATAAATGGGTATGGAACAGC
urhA express rev EcoRI	taagaattcGTTGATCACAAACTGCACGGCTATACG
Operon express rev EcoRI	attacgaattcgagctcgCCACACTGGAGCGGTTAAGCACC
puuEKO up Fwd	agcttgcatgcctgcaggtcgactctagaggatccGTGGTTCTGGCGACCGGCATAG
puuEKO up Rev	tgcacggctatacgcctattaGCCATAGCCGATCATGTCACGCG
puuEKO dwn Fwd	atgatcggctatggctaatagGCGTATAGCCGTGCAGTTTGTGATC
puuEKO dwn Rev EcoRI	aaacagctatgaccatgattacgaattcgagctcgGCATCGGGCATGGCGTTAACCTG
puuEKO verify BamHI	taaaggatccCAGTGCATCTACAACTTTACGTTCTCC
xdhBKO up Fwd	caagcttgcatgcctgcaggtcgactctagaggatcCGTATGGTCTGGGCACGCTG
xdhBKO up Rev	gcgtggacccgatacCAGGCTGGACAGATGGTGGAG
xdhBKO dwn Fwd	catctgtccagcctgGTATCGGGTCCACGCAGGTC
xdhBKO dwn Rev	cagctatgaccatgattacgaattcgagctcggtaccCTGCTCTATGTTCAGCACTGCG
xdhBKO verify Fwd	GAGCATGCTGGCAACAGCACAGC
xdhBKO verify Rev	GCTTCCAGCGGCACAACCCTG
XDH exp Fwd XbaI	ttttctagaCGACTTCCTGATCATTGGCACAG
XDH exp Rev BamHI	taaggatcCAACGTAACTGCTGCACACTTG
araC-SceI Fwd EcoRI	atatgaattcCAGTCACGACGTTGTAAAACGACG
araC-SceI Rev PstI	atatctgcagGCTACTCCGTCAAGCCGTCAATTG
AllP Fwd XbaI	ttttctagaGAGGGATTTATGGAACATCAGAGAAAAC
AllP Rev BamHI	ttatggtaccCTCAATGTTCAATATCGGGATTAATTAACC

^aAnnealing bases are capitalized. Restriction sites used in the study are bolded.

Complementation constructs (and pDN5) were constructed via PCR, restriction digest, and ligation into pCM62 or pCM66. Restriction sites utilized are indicated in the primer names in Table 3. The construct p62_UrhA_Operon was constructed in a similar fashion using primers UrhA express fwd BamHI and Operon express rev EcoRI. Allantoin permease, allP (syn. ybbW), was cloned from E. coli K-12 MG1655 (45).