Aminoimidazole Carboxamide Ribotide Exerts Opposing Effects on Thiamine Synthesis in Salmonella enterica

ABSTRACT

In Salmonella enterica, the thiamine biosynthetic intermediate 5-aminoimidazole ribotide (AIR) can be synthesized de novo independently of the early purine biosynthetic reactions. This secondary route to AIR synthesis is dependent on (i) 5-amino-4-imidazolecarboxamide ribotide (AICAR) accumulation, (ii) a functional phosphoribosylaminoimidazole-succinocarboxamide (SAICAR) synthetase (PurC; EC 6.3.2.6), and (iii) methionine and lysine in the growth medium. Studies presented here show that AICAR is a direct precursor to AIR in vivo. PurC-dependent conversion of AICAR to AIR was recreated in vitro. Physiological studies showed that exogenous nutrients (e.g., methionine and lysine) antagonize the inhibitory effects of AICAR on the ThiC reaction and decreased the cellular thiamine requirement. Finally, genetic results identified multiple loci that impacted the effect of AICAR on thiamine synthesis and implicated cellular aspartate levels in AICAR-dependent AIR synthesis. Together, the data here clarify the mechanism that allows conditional growth of a strain lacking the first five biosynthetic enzymes, and they provide additional insights into the complexity of the metabolic network and its plasticity.

IMPORTANCE In bacteria, the pyrimidine moiety of thiamine is derived from aminoimidazole ribotide (AIR), an intermediate in purine biosynthesis. A previous study described conditions under which AIR synthesis is independent of purine biosynthesis. This work is an extension of that previous study and describes a new synthetic pathway to thiamine that depends on a novel thiamine precursor and a secondary activity of the biosynthetic enzyme PurC. These findings provide mechanistic details of redundancy in the synthesis of a metabolite that is essential for nucleotide and coenzyme biosynthesis. Metabolic modifications that allow the new pathway to function or enhance it are also described.

INTRODUCTION

Thiamine pyrophosphate (TPP) is an essential coenzyme required by all organisms. TPP participates in catalysis required for a number of central metabolic processes, such as the conversion of pyruvate to acetyl coenzyme A (acetyl-CoA), the synthesis of the tricarboxylic acid cycle intermediate succinyl-CoA, and the pentose phosphate pathway. Study of metabolic redundancy in the thiamine biosynthetic pathway has provided a robust model system for identifying the sometimes-subtle interactions that comprise the metabolic network (1, 2). The interactions found to date have demonstrated that multiple biosynthetic pathways can form thiamine biosynthetic intermediates. Additionally, these studies have led to the identification of primary (RidA) and secondary (TrpD), or “moonlighting,” biochemical activities for enzymes not ordinarily involved in thiamine synthesis (3, 4).

The 4-methyl-5-(beta-hydroxyethyl)-thiazole (THZ) monophosphate and 4-amino-5-hydroxymethyl-2-methylpyrimidine (HMP) pyrophosphate moieties are synthesized independently, joined, and finally phosphorylated to form TPP (5). The first five catalytic steps of HMP synthesis are common to the purine biosynthetic pathway and form 5-aminoimidazole ribotide (AIR) (Fig. 1). A mutant strain of Salmonella enterica (purG purH) can bypass the common HMP/purine biosynthetic pathway for thiamine synthesis because an alternate route to forming AIR exists (6). Previous studies showed that this alternate route to AIR is dependent on (i) the accumulation of 5-amino-4-imidazolecarboxamide ribonucleotide (AICAR), (ii) a functional phosphoribosylaminoimidazole-succinocarboxamide (SAICAR) synthetase (PurC; EC 6.3.2.6), and (iii) methionine and lysine in the growth medium.

FIG 1.

The integrated purine, thiamine, and histidine biosynthetic pathways in S. enterica. Relevant biosynthetic pathways are represented schematically. Gene products are indicated next to the reaction they catalyze. Some steps are condensed, and significant metabolites are structurally depicted. Abbreviations: Gln, glutamine; PRPP, phosphoribosyl pyrophosphate; R5P, ribose-5-phosphate; AIR, 5-aminoimidazole ribotide; HMP-P, 4-amino-5-phosphomethyl-2-methylpyrimidine; TPP, thiamine pyrophosphate; N5-CAIR, N5-carboxy-AIR; CAIR, carboxy-AIR; AICAR, 5-amino-4-imidazolecarboxamide ribotide; SAICAR, succino-AICAR; FAICAR, 5-formyl AICAR.

AICAR is a purine biosynthetic intermediate and a by-product of histidine biosynthesis (7, 8). In eukaryotic organisms AICAR is extensively integrated in metabolic networks and has been demonstrated to allosterically and transcriptionally regulate metabolic processes (9, 10). AICAR stimulates AMP-activated protein kinase, which has a prominent role in maintaining energy homeostasis in mammals (11). The accumulation of AICAR is known to affect diverse areas of bacterial metabolism, including adenosine homeostasis (12), gluconeogenesis (13), and symbiosis (14). In S. enterica, AICAR has been shown to indirectly inhibit thiamine synthesis at the ThiC step (15–19) and directly inhibit pantothenate synthesis at the PanC-catalyzed step (18).

The enzyme PurC catalyzes the ligation of carboxy-AIR (CAIR) and aspartate in purine biosynthesis (Fig. 1). AICAR-dependent AIR synthesis in a purG purH mutant was shown to require PurC despite the absence of CAIR and SAICAR, the PurC product (6). These findings suggested that PurC had an alternative activity relevant to AIR formation (6). The structural similarities between AICAR and AIR led us to the straightforward hypothesis that S. enterica could decarbamylate AICAR to AIR. The working model for AICAR-dependent AIR synthesis was that PurC (independently or with other enzymes) converted AICAR to AIR via a CAIR intermediate (6).

In this study, the requirements for AICAR, PurC, and methionine and lysine in an alternative AIR synthesis pathway were investigated. Here, we provide in vivo evidence that AICAR is the precursor to AIR and in vitro evidence that PurC can use AICAR as a substrate to synthesize AIR. Further, the data show (i) an indirect role for methionine and lysine, (ii) the benefit of decreased aspartate levels in AICAR-dependent AIR synthesis, and (iii) that several mutations can independently increase AICAR-dependent thiamine synthesis, emphasizing the inherent plasticity of the metabolic network that allows it to adapt to genetic and environmental perturbations.

MATERIALS AND METHODS

Bacterial strains, media, and chemicals.

Strains used in this study are derivatives of S. enterica serovar Typhimurium LT2 and are described in Table 1. Rich medium consisted of Difco nutrient broth (NB; 8 g/liter) and NaCl (5 g/liter). Defined medium consisted of no-carbon E medium, 1 mM MgSO₄ (20–22), and trace minerals (23). Glucose (11 mM) was used as the sole carbon source. Difco BiTek agar (15 g/liter) was added for solid medium. Unless otherwise noted, nutrients were supplemented at the following concentrations: adenine (0.4 mM), methionine (0.3 mM), lysine (0.25 mM), succinate (4 mM), pantothenate (0.1 mM), thiamine (0.1 μM), aspartate (0.3 mM), threonine (0.3 mM). Antibiotics were used at the following concentrations for rich (minimal) medium: tetracycline at 20 (10) μg/ml; kanamycin at 50 (150) μg/ml; chloramphenicol at 20 (5) μg/ml; ampicillin at 100 (15) μg/ml. Isopropyl-β-d-thiogalactoside was used for protein expression for PurC and 5′-methylthioadenosine–S-adenosylhomocysteine nucleosidase (MTAN) purification (1 mM) and for physiology studies (0.1 mM). A stable isotope-labeled form of adenine, [U-¹³C]adenine, was purchased from Moravek Biochemicals, Inc.

TABLE 1.

Bacterial strains used in this study^a

Strain	Genotypeb
DM1	Wild type
DM2	purH355
DM95	thi-885::MudJ
DM6123	purG2324::MudJ STM4068-1 zxx9126::Tn10d(Tc) purH355
DM6124	purG2324::MudJ STM4068-1 zxx9126::Tn10d(Tc)
DM7185	araCBAD thiC1137::Kanr
DM11507	purG1739::Tn10 purH3115
DM11529	purG2324::MudJ purH3115
DM12239	ΔpurH3121::Kanr
DM12240	purG1739::Tn10 ΔpurH3121::Kanr
DM12363	purG1739::Tn10 ΔpurH3121::Kanr aspCp831
DM12369	purG1739::Tn10 ΔpurH3121::Kanr sdh215
DM12372	purG1739::Tn10 ΔpurH3121::Kanr ΔfumAC1
DM12412	purG1739::Tn10 ΔpurH3121::Kanr cyaA3661
DM12729	purG2324::MudJ purH3115 zxx4138::Tn10d(Tc) aspCp831
DM12730	purG2324::MudJ purH3115 zxx4138::Tn10d(Tc)
DM13061	purG2324::MudJ purH3115 mutL431::Tn10d(Tc)
DM13071	purG2324::MudJ purH3115 mtfA::Tn10d(Tc)
DM13072	purG2324::MudJ purH3115 zwf32::Tn10d(Tc)
DM13074	purG2324::MudJ purH3115 gdhA631::Tn10d(Tc)
DM13095	pncB::Tn10d(Tc)
DM13096	pncB::Tn10d(Tc) aspCp831
DM13216	purG1739::Tn10 purH3115 ΔguaA1085::cat ΔguaC1086 purE2154::MudJ
DM13336	purG311 purE3043 ΔSTM4068
DM13394	zxx4138::Tn10d(Tc) aspCp831 pCA24N
DM13395	zxx4138::Tn10d(Tc) aspCp831 pCA24N-aspC
DM13466	purG1739::Tn10 ΔpurH3121::Kanr aspCp831 pCA24N
DM13467	purG1739::Tn10 ΔpurH3121::Kanr aspCp831 pCA24N-aspC
DM14012	ΔpurH3121::Kanr purG311 purE3043 ΔSTM4068
DM14050	ΔpurH3121::Kanr purG311 purE3043 ΔSTM4068 ΔgdhA631::cat
DM14051	ΔpurH3121::Kanr purG311 purE3043 ΔSTM4068 ΔmtfA::cat
DM14052	ΔpurH3121::Kanr purG311 purE3043 ΔSTM4068 zwf21::Tn10d(Tc)
DM14053	ΔpurH3121::Kanr purG311 purE3043 ΔSTM4068 zxx4150::Tn10d(Tc) fumAC1
DM14054	ΔpurH3121::Kanr purG311 purE3043 ΔSTM4068 zxx4150::Tn10d(Tc)
DM14055	ΔpurH3121::Kanr purG311 purE3043 ΔSTM4068 ΔwzxE1::cat cyaA
DM14056	ΔpurH3121::Kanr purG311 purE3043 ΔSTM4068 ΔwzxE1::cat

^aAll strains are S. enterica serovar Typhimurium LT2 variants and were constructed for this study.

^bTn10d(Tc) refers to the transposition-defective mini-Tn10 (Tn10 Δ16 Δ17) construct (45). MudJ refers to the MudJ1734 transposon (46). Plasmid pCA24N-aspC is from the E. coli ASKA library (28), and the empty vector control, pCA24N, is a lab-made derivative.

Growth quantitation.

The cells from overnight cultures (NB medium) were pelleted and resuspended in an equal volume of 0.85% NaCl. A 5-μl aliquot was used to inoculate 195 μl of the desired media. Cultures were grown at 37°C with shaking for up to 36 h in a microplate reader (model EL808; Bio-Tek Instruments). Cell density was measured as the absorbance at 650 nm and reported as the final yield at a given time. Specific growth rates (μ; per hour) were determined as follows: μ = ln[(X/X₀)/T], where X is A₆₅₀ during exponential growth, X₀ is A₆₅₀ at time zero, and T is time (h). Lag times were defined as the period between the inoculation of the medium and a 10% increase in optical density.

Genetic techniques.

Transductional crosses were performed with a high-frequency general transducing mutant of bacteriophage P22 (P22 HT105/1 int-201) (24). Methods for transductional crosses, isolation of transductants from bacteriophage, and identification of bacteriophage-free transductants have been previously described previously (25). Mutant strains were constructed using standard genetic techniques. Insertion deletions were generated using the λ-Red-mediated homologous recombination method (26).

Isolation of suppressor mutations in a purG purH background. (i) Spontaneous mutations.

Ten independent cultures of DM12240 (purG1739::Tn10 ΔpurH3121::Kan^r) were grown overnight in NB. Cells were harvested by centrifugation and resuspended in 0.85% NaCl. Approximately 1 × 10⁸ cells from each cell suspension were spread onto solid glucose medium with adenine containing (i) no additives, (ii) methionine, (iii) lysine, (iv) succinate, or (v) pantothenate and incubated at 37°C (22). Colonies arose on the methionine-, succinate-, and pantothenate-containing media, and a maximum of three colonies per plate were streaked for isolation on nonselective rich medium and characterized further. A pool containing >80,000 random Tn10d(Tc) insertions was used to map spontaneous suppressor mutations. Tn10d(Tc) insertions found to be linked to suppressor mutations were used to reconstruct relevant strains.

(ii) Insertions.

A P22 lysate grown in a pool of cells containing >80,000 random Tn10d(Tc) insertions was used to transduce DM11529 (purG2324::MudJ purH3115) to tetracycline resistance (Tc^r) on solid NB medium. The Tc^r transductants were screened for a Thi⁺ phenotype on glucose medium containing adenine, methionine, and tetracycline. Tc^r Thi⁺ transductants were reconstructed by transduction.

Derivatization and purification of the pyrimidine moiety of thiamine.

A 500-ml culture of DM13216 (purG purH guaA guaC purE) was grown to an A₆₅₀ of approximately 0.8 in defined medium. The medium contained glucose (16.5 mM), guanine (0.2 mM), methionine, lysine, and either unlabeled adenine (150 μM) or a 1:1 mixture of unlabeled adenine and [U-¹³C]adenine (75 μM:75 μM). Cells were pelleted, and the pyrimidine moiety of thiamine was derivatized and isolated as previously described (27). A crude extraction of TTP was achieved by boiling the cells in 0.1 N HCl for 20 min. Following a pH adjustment with NaOH, the cell debris was cleared by centrifugation. A mixture of ethanol and ethanethiol was added, and the solution was boiled for 2.5 h, during which time TPP was cleaved by the ethanethiol nucleophile, displacing thiazole diphosphate and forming 2-methyl-4-amino-5-[(ethyl-thio)methyl]pyrimidine (EtMP). EtMP was then purified by a series of methylene chloride extractions. A stream of nitrogen was used to volatilize the methylene chloride, and the EtMP residue was resuspended in ethyl acetate. Samples were submitted for positive time of flight [(+) TOF] mass spectral analysis at the University of Wisconsin Biotechnology Center.

Purification of SAICAR synthetase (PurC; EC 6.3.2.6).

PurC purification was previously described (6). PurC with an N-terminal hexahistidine tag was expressed from JW2461-2 cells (containing pCA24N-purC) and purified using nickel ion affinity chromatography. Buffer used for purification was 50 mM Tris, pH 7.4, with 100 mM NaCl, 5% glycerol.

Purification of MTAN (EC 3.2.2.9).

JW0155-1 (carrying pCA24N-pfs) (28) was used to inoculate 50 ml of SB medium supplemented with 30 μg/ml chloramphenicol and incubated for ∼18 h at 37°C with shaking. A 40-ml aliquot was used to inoculate 3 liters of the same medium. When the A₆₅₀ reached 0.8, MTAN expression was induced with 0.1 mM isopropyl-β-d-thiogalactoside, cells were incubated at 28°C for an additional 10 h, the mixture was pelleted by centrifugation (7,000 × g for 10 min at 4°C), and the cell pellet was stored. The cell pellet was washed (2 times) and resuspended in binding buffer (50 mM KPO₄ [pH 7.5], 100 mM KCl, 5% glycerol). DNase and lysozyme were added at 0.01 mg/ml, cells were broken with a French press (three passes at 1,500 lb/in²), and cell debris was pelleted by centrifugation (18,000 × g for 30 min at 4°C). A 0.22-μm filter was used to remove particulates from the cell lysate, which was then loaded onto a Ni-nitrilotriacetic acid Superflow resin (Qiagen). Two column volumes (CV) of binding buffer, followed by 15 CV of wash buffer (50 mM KPO₄ [pH 7.5], 300 mM KCl, 5% glycerol, 10 mM imidazole) were passed over the column until the A₂₈₀ of the eluent was zero. MTAN was eluted with an imidazole gradient (10 to 500 mM) over 20 CV. The MTAN concentration was determined with a bicinchoninic acid protein assay kit (Thermo Scientific).

PurC assays.

Previously described activity assays for forward (29) and reverse (30) PurC catalysis were modified by substituting AICAR for the PurC substrate and eliminating any coupling components. The final reaction volumes were 50 μl and contained the following: 50 mM HEPES (pH 7.8), 20 mM KCl, 6 mM MgCl₂, 2.5 mM ATP, 2.5 mM AICAR, 100 μM PurC, and 4 μM TdPurE (the AIR carboxylase from Treponema denticola) for forward catalysis. For conditions that favored reverse catalysis, the final reaction volumes were 50 μl and contained the following: 50 mM Tris, 20 mM KCl, 6 mM MgCl₂, 2.5 mM Na₃AsO₄, 2.5 mM ADP, 2.5 mM AICAR, 100 μM PurC, and 4 μM TdPurE. TdPurE was a gift from L. D. Palmer (31). TdPurE, which can interconvert AIR and CAIR, was added in case AIR synthesis proceeded through a CAIR intermediate. Variations of the assay mixture included 5 mM aspartate and ranged from pH 5 to 10. Reaction mixtures were incubated at 37°C for 16 h. Stock solutions of AICAR were prepared using an extinction coefficient (ε₂₆₈) of 12,348 M⁻¹ cm⁻¹ (29). Stock solutions of ADP and ATP were prepared using an extinction coefficient (ε₂₅₉) of 15,400 M⁻¹ cm⁻¹ (32).

ThiC-coupled bioassays (EC 4.1.99.17).

PurC reaction mixtures, 2.5 mM AIR, 22 mM S-adenosylmethionine (SAM), 50 mM HEPES (pH 7.8), and purified MTAN, were each made anoxic by flushing with N₂ for 20 min (in 1.5-ml Eppendorf tubes or glass bottles sealed with rubber stoppers). All reagents were then transferred to an anaerobic chamber outfitted with an automatic airlock (Coy Laboratories, Grass Lake, MI). In the anaerobic chamber, 50 mM HEPES (pH 7.8) was used to make an anoxic solution of 100 mM dithionite. The 4Fe-4S cluster of 51 μM ThiC was reduced by a 10-min incubation with 1.7 mM dithionite. A 15-μl aliquot of each PurC reaction mixture was combined with 35 μl of a master mix (reduced ThiC, SAM, and MTAN) to bring the final reaction volumes to 50 μl. Final concentrations of ThiC assay reagents were as follows: 1.1 mM dithionite, 0.12 mM SAM, 2 μM MTAN, and 32 μM ThiC. Control reaction mixtures contained 35 μl of the ThiC reaction master mix with buffer only or authentic AIR (100 μM, final concentration). Reaction mixtures were incubated for 18 h at room temperature and then removed from the anaerobic chamber. Reaction mixtures were then centrifuged in a microcentrifuge at top speed for 1 min, and 5 μl of the supernatant was pipetted onto a minimal glucose medium plate containing a soft agar overlay inoculated with the HMP auxotroph DM7185 (thiC). The plates were incubated at 37°C for 18 h and assessed for growth. ThiC assay components (ThiC, authentic AIR, and purified SAM) were a gift from L. D. Palmer and were prepared as previously described (19, 31). Previous experiments determined that a bioassay with DM7185 reliably detected 1 pmol of HMP (L. D. Palmer, unpublished data).

Aspartate aminotransferase assays (AspC; EC 2.6.1.57).

Cells were grown in minimal glucose medium (100 ml) to early stationary phase (A₆₅₀, ∼1.3) and stored as pellets at −80°C. Cell pellets were resuspended in 5 ml of 25 mM KPO₄ (pH 7.5), 0.1 mM EDTA, 0.2 mM pyridoxal 5′-phosphate, 0.2 mM dithiothreitol, and 5% glycerol. The resuspended cell pellets were each sonicated twice for 30 s (10-s bursts), and after a 30-min centrifugation (7,000 × g for 10 min at 4°C), the cleared supernatant was transferred to a new tube. The reverse aminotransferase activity was measured by using a previously described assay (33) that coupled oxaloacetate formation to malate dehydrogenase activity and monitored NADH oxidation at 340 nm. Protein concentrations were determined using a bicinchoninic acid assay kit (Thermo Scientific).

AIRs syntheses. (i) CAIRs synthesis.

Carboxy-AIR riboside (CAIRs) was synthesized as previously described (34) by using a 10-fold dilution of all reagents. AICARs (2.5 g) was saponified with 6 N NaOH (10 ml) in a reflux condenser (with soda lime) at 120°C for 4 h. After cooling on ice, the mixture was triturated after the addition of ethanol (20 ml once, 5 ml three times, and 2.5 ml once), discarding the supernatant each time. The substance was dried with a Savant SpeedVac and then triturated after the addition of 5 ml methanol. After 3 h, the supernatant was discarded and the final product was dried with the SpeedVac.

(ii) AIRs synthesis.

AIRs was synthesized as previously described (35) using a 2.5-fold dilution of all reagents and an alternate chromatography technique. CAIRs (0.5 g) was dissolved in 0.25 M NaOAc-AcOH buffer (pH 4.8; 20 ml). Nitrogen was streamed through the solution for 24 h at 27°C. A glass column containing 8 ml of Dowex-50WX8 hydrogen form resin was ionized by flowing 3 CV of 1 M H₂SO₄ over it, followed by water until the pH was >2. After the AIRs solution was loaded onto the column, the column was washed with 25 CV of water. AIRs was eluted with 1 N NaOH. Eluted fractions were collected and analyzed for AIRs with an HPLC system. AIRs-containing fractions were pooled and confirmed for their ability to feed DM6123 and DM6124, as previously described. The pools were then lyophilized, and the powder was resuspended in water, lypophilized again, and stored at −80°C.

HPLC analysis of PurC reaction mixtures and AIRs-containing fractions.

PurC reaction mixtures (10 μl) or putative AIRs-containing fractions (20 μl) were loaded onto a Gemini C₁₈ column (5 μm, 110 Å, 150 by 4.6 mm) from Phenomenex and run at 1 ml/min isocratically (100 mM triethylammonium acetate [pH 7.5]–2% methanol) for 10 min, followed by a 10-min gradient to 75% methanol. Data were extracted from the HPLC system (Shimadzu Scientific Instruments) at 254 nm. A series of 14 solutions (in 50 mM HEPES, pH 7.8) containing 1 μM to 2.5 mM AIR were made using a molar extinction coefficient (ε₂₅₀) of 3,270 M⁻¹ cm⁻¹ (29). These solutions were used to generate a standard curve (concentration of AIR versus the area under the relevant peak) for AIR quantification.

RESULTS

AICAR is a precursor to AIR in vivo.

Genetic experiments with a purG purH mutant showed that AIR could be synthesized in the absence of the first five purine biosynthetic enzymes if AICAR levels were increased (6). A simple model suggested that AICAR was a direct precursor to AIR in vivo; however, it was possible that the effect of AICAR was indirect. Stable isotope labeling studies were used to distinguish between the two possibilities. AICAR is not imported by S. enterica, and so a strategy relying on knowledge of purine salvage and histidine biosynthesis was employed. Uniformly labeled [U-¹³C]adenine was provided to a strain in which all of the AICAR in the cell was derived from exogenously provided adenine via the histidine biosynthetic pathway (Fig. 2A).

FIG 2.

AICAR is the direct precursor to AIR in vivo. (A) Scheme for the [U-¹³C]adenine study, showing AICAR-dependent TPP synthesis and derivatization/purification of the pyrimidine moiety of thiamine. (B and C) Mass spectral data from (+) TOF analysis of the derivatized pyrimidine moiety of thiamine, EtMP, from AICAR-dependent AIR synthesis in DM13216 (purG purH3115 guaA guaC purE). Expected m/z of EtMP in this mode was 184 (+3 if AIR was derived from AICAR). Cells were grown with 0.15 mM unlabeled adenine (B) or a 1:1 mixture of unlabeled and [U-¹³C]adenine, 0.15 mM total (C).

To maximize detection of the label, strain DM13216 (purG purH guaA guaC purE) was constructed (Fig. 1). This combination of mutations ensured that (i) purine synthesis was blocked to prevent canonical AIR synthesis (purG), (ii) AICAR accumulated from histidine biosynthesis (purH), (iii) exchange between cellular pools of adenine- and guanine-containing nucleotides was prevented (guaA, guaC), and (iv) AIR accumulated to increase thiamine synthesis (purE). DM13216 was grown in liquid minimal medium containing unlabeled adenine or a 1:1 mixture of unlabeled and uniformly ¹³C-labeled adenine. In each case, TPP was extracted from the cells and cleaved into THZ pyrophosphate and EtMP. Flux through the histidine biosynthetic pathway using uniformly ¹³C-labeled adenine would result in AICAR with an atomic mass increased by 4. The hypothesized conversion of AICAR to AIR (6) would entail the loss of the [¹³C]carboxamide group and yield a molecule of AIR with three ¹³C atoms (Fig. 2A). Based on the ThiC-catalyzed conversion of AIR to HMP-P, the pyrimidine moiety of thiamine would retain these three ¹³C atoms (36), which would be present in the EtMP derivative. Mass spectrometry was used to analyze the purified pyrimidine moiety (in the form of EtMP) from each growth condition.

When cells were grown with unlabeled adenine, a single peak (m/z 184) with the expected mass-to-charge ratio for EtMP was detected (Fig. 2B). In contrast, when a 1:1 mixture of unlabeled and [U-¹³C]adenine was provided in the growth medium, the m/z 184 peak was joined by an equally prominent peak, m/z 187, the anticipated mass-to-charge ratio for EtMP derived from ¹³C-labeled AICAR (Fig. 2C). These data confirmed that three ¹³C atoms were incorporated into the pyrimidine moiety of approximately half of the cellular TPP and supported the conclusion that AICAR was the in vivo precursor to AIR in the purG purH mutant.

PurC converts AICAR to CAIR.

The confirmation that AIR was derived from AICAR in vivo, combined with the requirement for PurC, led to a model in which PurC had a direct role in converting AICAR to AIR. PurC was purified and incubated with AICAR under reaction conditions conducive to the forward and reverse reactions of PurC at various pHs. To account for the possibility that AIR synthesis proceeds through a CAIR intermediate (6), the AIR carboxylase from Treponema denticola (TdPurE) was included in an otherwise-identical set of reaction mixtures. This enzyme strongly favors the conversion of CAIR to AIR (37) and would thus increase the amount of product detected. When analyzed via HPLC, reaction mixtures containing PurC, AICAR, ADP, Na₃AsO₄, and TdPurE (pH 9) yielded a product that coeluted with an authentic AIR standard (Fig. 3A). Control reactions showed that product formation required PurC and AICAR.

FIG 3.

PurC converts AICAR to CAIR in vitro. (A) A product peak was observed (A₂₅₄) when PurC, AICAR, ADP, Na₃AsO₄, and TdPurE (pH 9) reaction mixtures were incubated for 16 h at 37°C (black line). The product was not detected in identical reaction mixtures lacking AICAR (dashed line) or PurC (dotted line). The product coeluted with an authentic standard of AIR (gray line). (B) Soft agar inoculated with DM7185 (thiC) was overlaid on a minimal glucose medium plate and used for a bioassay of in vitro reaction products. In vitro PurC reactions were coupled to ThiC activity, and 5 μl of each coupled reaction mixture was pipetted on the agar (positions 1 to 4). After an 18-h incubation at 21°C, turbid zones of growth indicated that HMP-P was formed in PurC reaction mixtures containing 2.5 mM AICAR, 2.5 mM ADP/Na₃ASO₄, 100 μM PurC, and 4 μM TdPurE (position 1). A lack of growth indicated HMP-P was not formed when reaction mixtures lacked PurC (position 2), AICAR (3), or ADP/Na₃AsO₄ (4). An uncoupled ThiC reaction (100 μM authentic AIR as the substrate) was diluted 1/100 (position 5) was used as a positive control for the ThiC reaction, and a reaction mixture lacking AIR (position 6) was used as a negative control.

Formation of AIR in reaction mixtures containing PurC, AICAR, ADP, Na₃AsO₄, and TdPurE (pH 9) was verified in a coupled assay with ThiC, which converts AIR to the HMP moiety of thiamine. HMP was then qualitatively assessed by bioassay, where the reaction mix was used to support growth of a thiC mutant (Fig. 3B). Previous work determined that this bioassay is able to detect 1 pmol HMP (Palmer, unpublished). Control reaction mixtures lacking PurC, AICAR, or ADP-Na₃AsO₄ failed to generate detectable HMP. The reaction mix spotted in position 5 of Fig. 3 contained authentic AIR in place of the PurC-mediated reaction mix in the ThiC assay. This positive result confirmed both the activity of the ThiC protein and the specificity of the thiC mutant for HMP.

Together, these data showed that PurC has a direct role in the formation of AICAR-dependent AIR synthesis and supported the hypothesis that PurC converts AICAR to CAIR via a reaction analogous to the reverse reaction with PurC. Importantly, the in vitro requirement for AIR carboxylase (TdPurE) does not directly correlate with an in vivo role for the AIR carboxylase-mediated conversion of CAIR to AIR, as it has been shown that a purG purH purE mutant strain maintains AICAR-dependent AIR synthesis (6). A standard curve of the authentic AIR standard (1.5 to 25 μM) showed that the PurC-mediated reaction generated 11.2 μM AIR over the 16-h incubation period. The weak product formation in the implemented assay suggested other conditions, or additional components, are required for optimal activity. Numerous components of the assay mixture were altered in an effort to increase the formation of AIR, including substrate concentrations and pHs. No significant changes in AIR formation were observed.

Dissecting the opposing roles of AICAR on thiamine synthesis.

purH mutants require thiamine, because accumulated AICAR indirectly compromises the ThiC reaction (15–17, 19, 38), at least in part by causing a decrease in CoA levels (18). purG purH mutants require AICAR for PurC-dependent thiamine synthesis but also face the AICAR-dependent inhibition of ThiC detected in a purH mutant (6). Thiamine-independent growth of purG purH requires the addition of (i) methionine and lysine, (ii) succinate and pantothenate, or (iii) methionine and pantothenate (Fig. 4). The combination of methionine and lysine, or succinate alone, lowers the cellular thiamine requirement (39). Methionine and pantothenate have been shown to individually and additively restore thiamine synthesis (i.e., ThiC activity) in a purH mutant (18, 40). Pantothenate restores ThiC activity by bypassing a constraint at PanC and increasing CoA levels, while methionine restores ThiC activity by an unknown mechanism (18). Therefore, the nutrients that allow growth of purG purH mutants in the absence of thiamine must either (i) reduce the cellular thiamine requirement, (ii) decrease the negative impact of AICAR on ThiC, (iii) increase AICAR-dependent thiamine synthesis, or (iv) act via a combination of these actions. AICAR-dependent AIR synthesis in vitro did not depend on any of these nutrients, suggesting that these metabolites modulated the metabolic network to allow AICAR-dependent AIR synthesis to satisfy the cellular thiamine requirement.

FIG 4.

Nutrients that allow AICAR-dependent AIR synthesis to satisfy cellular thiamine requirement in S. enterica. Thiamine-independent growth of a purG purH mutant (DM11507) was quantified in minimal medium containing glucose and adenine. Additional supplements were methionine and lysine (closed circles), pantothenate and succinate (squares), methionine and pantothenate (diamonds), or thiamine (open circles).

Mutations that enhanced thiamine-independent growth of a purG purH mutant were isolated to clarify the opposing roles of AICAR on thiamine biosynthesis. DM12240 (purG ΔpurH3121::Kan^r) cells were spread on minimal medium containing glucose and adenine with (i) no addition, (ii) methionine, (iii) lysine, (iv) pantothenate, or (v) succinate. Colonies arose on the medium containing either succinate or pantothenate at average frequencies of ∼5 × 10⁻⁷ and less frequently on medium with adenine and methionine. Additionally, DM11529 (purG purH) was mutagenized via a random Tn10d(Tc) insertion pool, and resulting mutants were screened for thiamine-independent growth on minimal medium containing glucose, adenine, and methionine. In total, eight independently isolated revertants of purG purH that grew on glucose medium with adenine and methionine were further characterized. The loci and nature of their respective suppressor mutations were determined by sequence analysis. Confirmation that the identified mutations were causative was obtained by the recapitulation of the relevant phenotypes via strain reconstruction (insertions linked to spontaneous mutations or in-frame insertion deletions) (Table 2).

TABLE 2.

Mutations suppress the nutritional requirements of a purG purH mutant^a

Strain	Locus	Gene product	EC no.d	Final yielde (A650) in:
				Met	Thi
DM12363	RBS of aspC	Aspartate aminotransferase	2.6.1.1	0.556	0.710
DM12369	sdhCb	Succinate dehydrogenase cytochrome b556 subunit	1.3.5.1	0.494	0.718
DM12372	fumAC	Fumarate hydratase class I and II	4.2.1.2	0.389	0.677
DM12414	cyaAc	Adenylate cyclase	4.6.1.1	0.553	0.715
DM13061	mutL	DNA mismatch repair protein	NA	0.517	0.577
DM13071	mtfA	Mlc titration factor A	NA	0.520	0.764
DM13072	zwf	Glucose-6-phosphate 1-dehydrogenase	1.1.1.49	0.593	0.637
DM13074	gdhA	Glutamate dehydrogenase	1.4.1.4	0.737	0.740

^aLesions were isolated as suppressor mutations of a purG purH mutant that reduced the nutritional requirement for thiamine-independent growth.

^bThe identity of the spontaneous mutation in DM12369 was inferred based on linkage to the sdh operon and the reconstruction of DM12369 phenotypes when a characterized sdhC insertion was transduced into a purG purH3115 mutant, generating DM12656 (data not shown).

^cThe mutation in cyaA (G1018T) resulted in a protein truncated at residue 339.

^dThe Enzyme Commission number indicate the gene product function. NA, not available (the gene product has not been assigned an EC number).

^eGrowth quantitation in minimal glucose adenine medium containing the indicated addition(s) was quantified based on the average final cell density after 36 h of growth for two independent cultures (error between cultures was ≤10%). Results for purG purH parent strains (DM12240 and DM11529) are not shown, for simplicity. Met, methionine; Thi, thiamine.

The identification of sdh mutations as suppressors raised the possibility that another allele(s) acts to reduce the thiamine requirement of the strain (39). In strains where thiamine synthesis is constrained, an sdh mutation restores thiamine-independent growth by decreasing the thiamine requirement, not by restoring thiamine synthesis (39). The decreased thiamine requirement leads to increased intracellular thiamine pools, which allow unsustainable growth of a thiamine auxotroph in thiamine-free medium. In contrast, mutations that enhanced thiamine synthesis in a purG purH mutant would not be expected to allow growth in the absence of added thiamine in a strain lacking a thiamine biosynthetic gene. In fact, when the other, non-sdh suppressor mutations were transduced into a thiamine auxotroph (DM95), no measurable growth was detected in the absence of thiamine (data not shown). These data confirmed that seven of the mutations (aspCp831 allele, fumAC, cyaA, zwf, gdhA, mtfA, and mutL) did not simply spare thiamine but directly affected thiamine synthesis.

AICAR-mediated inhibition of ThiC is reduced by loss of zwf, gdhA, or mtfA.

Unlike a purG purH strain, a purH mutant strain is not limited for AIR, but instead it cannot synthesize thiamine, solely because of the AICAR-induced block at ThiC. Each of the seven mutations that affected thiamine synthesis was transduced (by linked, or causative, insertion) into a purH mutant (DM12239). Growth of resulting strains was quantified in liquid minimal glucose medium containing adenine as a source of purines. In all but two cases (aspCp831 allele and mutL), the mutations partially restored thiamine-independent growth (Table 3). These data suggested that lesions in fumAC, cyaA, zwf, gdhA, and mtfA either decreased AICAR levels, antagonized the inhibitory effects of AICAR on ThiC, or enhanced ThiC activity independently of AICAR.

TABLE 3.

Suppressor mutations restore thiamine-independent growth of a purH mutant^a

Strain	Relevant genotype	μ (A650) in:
		Ade	Ade Thi
DM12239	purH	NG	0.36 (0.67)
DM14015	purH mutL	NG	0.35 (0.67)
DM14016	purH gdhA	0.16 (0.44)	0.36 (0.62)
DM14017	purH mtfA	0.15 (0.61)	0.22 (0.65)
DM14018	purH zwf	0.19 (0.50)	0.29 (0.61)
DM14019	purH fumAC	0.20 (0.44)	0.36 (0.54)
DM14020	purH	NG	0.36 (0.58)
DM14021	purH aspCp831	NG	0.28 (0.53)
DM14022	purH	NG	0.32 (0.53)
DM14001	purH cyaA	0.12 (0.58)	0.14 (0.61)
DM14002	purH	NG	0.32 (0.60)

^aStrains were grown in minimal glucose medium containing the indicated addition(s). Data shown are the average growth rate (per hour) and the final cell density after 18 h for three independent cultures (errors between cultures were ≤10%). NG, no growth (final cell density was <0.18). Isogenic strains lacking suppressor mutations (DM14020, DM14022, and DM14002) are shown below their corresponding isogenic strain. Ade, adenine; Thi, thiamine.

To further assess the impact of the fumAC, cyaA, zwf, gdhA, and mtfA mutations on ThiC activity, each was transduced into strain DM14012 (purG purH purE STM4068). In this strain, AIR is derived from in vivo phosphorylation of exogenously provided AIR riboside (AIRs), and AICAR that accumulates is derived only from histidine biosynthesis. These features remove the involvement of the purine pathway and indicate that thiamine synthesis is dependent only on ThiC activity. The relevant strains were inoculated onto a soft agar overlay on solid glucose minimal medium with adenine as previously described (17, 41). A titration of AIRs showed that strains containing mutations in zwf, gdhA, or mtfA required 10-fold less AIRs to support growth than the parental strain DM14012 (Fig. 5). In contrast, neither the fumAC mutation nor cyaA mutation impacted the amount of AIRs required for growth of DM14012 (data not shown). Importantly, none of the seven mutations decreased the AIRs requirement in the strain DM13336, which carries a wild-type purH and does not accumulate AICAR (data not shown). These data suggested that the zwf, gdhA, and mtfA alleles antagonized the impact of AICAR and restored ThiC activity in a purH mutant. The roles of fumAC and cyaA remained unclear based on this test, but they could require a component of the network disrupted in the DM14012 strain for their effect (e.g., PurEK). The suppressor mutations in the aspCp831 allele and mutL had no effect on the thiamine requirement of a purH mutant and were candidates for specifically affecting AICAR-dependent AIR synthesis. The mutator phenotype of a mutL mutant strain complicated interpretation of its effect, and this locus was not considered further.

FIG 5.

Lesions in zwf, mtfA, and gdhA restore ThiC activity. (A) DM14012 (purG purE purH STM4068) was overlaid (in soft agar) onto minimal glucose medium containing 0.4 mM adenine as the purine source. (B to D) A purG purE purH STM4068 mutant containing mutations in zwf (DM14052) (B), mtfA (DM14051) (C), or gdhA (DM14050) (D) were overlaid onto identical medium. Additional supplementation was added by spotting 3 μl of dilutions of the AIRs or 0.5 pmol thiamine in 0.5 μl to the top of the soft agar at different positions. The concentration of the AIRs stock solution was determined to be ∼300 mM.

Decreased aspartate increases AICAR-dependent AIR synthesis in vivo.

The suppressor mutation in DM12363 was a transition (A to G) 8 bases upstream of the aspC translation start site, in the putative Shine-Dalgarno box (ACCTCGTC). The relevant allele was designated aspCp831 and was hypothesized to affect the translation of the aspC gene product. The aspC gene encodes aspartate aminotransferase (AspC; EC 2.6.1.57), which synthesizes aspartate from oxaloacetate and glutamate.

Strain DM12363 (purG purH aspCp831) was evaluated for growth on media with the relevant nutrients in the absence of thiamine. The growth data in Table 4 show that the aspCp831 allele conferred thiamine-independent growth of a purG purH mutant on minimal glucose adenine medium containing methionine, pantothenate, or succinate. Weak growth with methionine/lysine compared to methionine alone suggested that the suppressor mutation resulted in lysine sensitivity in S. enterica.

TABLE 4.

Thiamine-independent growth of the purG purH aspCp mutant^a

Nutrient addition(s)	Final yield (A650)	Growth rate (h−1)	Lag time (h)
Met	0.472	0.2902	10.0
Met-Lys	0.426	0.1225	15.0
Succ	0.444	0.2419	4.5
Pant	0.422	0.1749	21.0
Thi	0.515	0.3058	5.0

^aGrowth of strain DM12363 (purG purH aspCp) was quantified in minimal medium containing glucose, adenine, and the indicated nutrient addition(s). Data shown are the averages from three independent cultures. Final yield, final cell density after 36 h. The lag time was defined as the period between inoculation of the medium and a 10% increase in optical density. The standard error between cultures was ≤10%. Met, methionine; Lys, lysine; Succ, succinate; Pant, pantothenate; Thi, thiamine.

When the aspCp831 allele was transduced into a wild-type background, it caused a noticeable growth defect in minimal glucose medium and sensitivity to exogenous lysine (Fig. 6A). Results from three experiments supported the conclusion that the relevant consequence of the aspCp831 mutant allele was the decreased wild-type AspC in the cell. In the first experiment, pCA24N-aspC, expressing a wild-type copy of aspC, was transformed into wild-type and purG purH genetic backgrounds that carried the aspCp831 allele. Quantitation of growth in the resulting strains (DM13394/5 and DM13468/9) showed that pCA24N-aspC reversed the effects of the aspCp831 allele, i.e., restored growth in the wild-type background and eliminated thiamine-independent growth in the purG purH background (data not shown). A second experiment determined that the addition of aspartate to the medium similarly reversed the phenotypes caused by the aspCp831 allele in both the wild-type and purG purH backgrounds (Fig. 6). Finally, the aspartate aminotransferase activities of DM13095 (wild type) and DM13096 (aspCp831) were measured in cell extracts. As anticipated, the activity in the aspCp831 allele mutant (0.12 ± 0.00 μmol/min/mg of protein) was lower than that of the wild type (0.46 ± 0.06 μmol/min/mg of protein). Taken together, these data supported the conclusion that the aspCp831 allele decreased translation of AspC, resulting in a lower AspC activity and cellular aspartate concentration. Together, these data indicated that decreased aspartate in vivo, directly or indirectly, increases AICAR-dependent AIR synthesis. Although aspartate has numerous fates in the cell, it is interesting that it is a substrate of the canonical PurC reaction (Fig. 1).

FIG 6.

Growth phenotypes caused by the aspCp831 allele. (A) Growth of DM13096 (aspCp831) was evaluated in minimal glucose medium containing no addition (solid circles), lysine (solid squares), lysine and aspartate (triangles), or aspartate (diamonds). Growth of the isogenic wild-type strain, DM13095, is shown for comparison with no addition (open circles) and lysine (open squares). (B) Growth of DM12363 (purG purH aspCp831) was evaluated in minimal glucose medium containing adenine and methionine. Additional supplements were none (closed circles), aspartate (closed triangles), methionine and lysine (open circles), methionine, lysine, and aspartate (open triangles), and aspartate and thiamine (squares).

DISCUSSION

AICAR-dependent AIR synthesis in a purG purH mutant required (i) the accumulation of AICAR from the histidine biosynthetic pathway, (ii) the post-AIR purine biosynthetic enzyme PurC, and (iii) methionine and lysine in the medium (6). The work described here investigated the role of each of the above requirements in AIR synthesis and the complex effects of AICAR, as it simultaneously restores and inhibits thiamine synthesis. The data showed that (i) AICAR serves as the in vivo precursor to AIR, (ii) PurC can directly participate in the conversion of AICAR to AIR, (iii) methionine and lysine are not directly involved in AIR synthesis, and (iv) lowering the cellular aspartate concentration increases AICAR-dependent AIR formation. Additional data showed that modulation of the metabolic network allowed AICAR-dependent AIR synthesis to satisfy the thiamine requirement of the cell, suggesting that this route of thiamine synthesis could potentially be used by some organisms.

Genetic data were unable to reveal whether the need for AICAR accumulation to generate AIR reflects a direct or indirect role of this metabolite (6). When a 1:1 ratio of ¹²C-containing adenine and ¹³C-containing adenine was provided to the appropriate mutant strain (purG purH guaA guaC purE), a similar ratio of EtMP to EtMP plus 3 atomic mass units (amu) was observed. The route of adenine through the salvage pathway to ATP and further through histidine biosynthesis would result in a 1:1 ratio of AICAR to AICAR plus 4 amu (Fig. 2A). The conversion of AICAR to AIR would necessarily eliminate one ¹³C atom (via the loss of the carboxamide carbon), leaving three ¹³C atoms that would be incorporated into the pyrimidine moiety of thiamine via ThiC (36). The data in Fig. 2C show the ratio of m/z 187 and 184 species that would be expected if AICAR were converted to AIR. Together, these results showed that AICAR could serve as a direct precursor to AIR in vivo.

Genetic data showed a requirement for PurC in AICAR-dependent AIR synthesis, but these data did not distinguish between a direct or indirect role for this enzyme in AIR formation. In the simplest scenario, PurC catalyzed the conversion of AICAR to CAIR or AIR. This scenario was attractive, since AICAR has structural similarities to CAIR, a canonical PurC substrate. The AICAR-dependent formation of AIR was reconstituted in vitro and required PurC and an AIR carboxylase. The in vitro requirement for AIR carboxylase suggested that PurC converts AICAR to CAIR. As an AIR carboxylase is not required in vivo, thiamine is increased in the absence of an AIR carboxylase (6), and CAIR is known to be unstable and spontaneously decarboxylate to AIR (42–44), we hypothesize that in vivo the conversion of CAIR to AIR can occur spontaneously or via AIR carboxylase (Fig. 7). The qualitative detection of AIR with a coupled assay using ThiC confirmed that PurC is directly involved in the AICAR-dependent AIR formation detected in vivo. Further biochemical analysis of the PurC-catalyzed CAIR synthesis would be needed to uncover optimal components and conditions for the activity. Such studies would further clarify the mechanism of AICAR-dependent AIR synthesis in vivo.

FIG 7.

Opposing effects of AICAR on thiamine biosynthesis in S. enterica. The model for AICAR-dependent AIR synthesis and points of metabolic cross talk are represented schematically. Biochemical reactions are represented by the solid black arrows, and the corresponding gene products are indicated next to the reaction they catalyze. Points of metabolic cross talk are indicated by gray arrows (contribution) or lines (inhibition) from the relevant metabolite to the metabolic process affected. The hypothesized effects of suppressor mutations are indicated with dashed black arrows or lines next to the relevant gene name.

Unlike the above components, a number of nutrients (e.g., methionine and lysine) allowed thiamine synthesis in vivo but did not directly participate in AIR synthesis in vitro. Rather, these conditions and a number of suppressor mutations modulated the metabolic network to allow the thiamine requirement of the cell to be satisfied by AICAR (Fig. 7). Based on our understanding of this pathway, there are three potential mechanisms that would allow AICAR-dependent AIR synthesis to satisfy the thiamine requirement of a strain. First, the cellular thiamine requirement could be lowered (e.g., by sparing succinyl-CoA) (39). Second, inhibition of ThiC activity by AICAR could be ameliorated (17, 18); finally, AICAR-dependent AIR synthesis could be stimulated.

Six of the eight mutations that restored thiamine synthesis to the purG purH strain also restored growth to a purH mutant strain. This result suggested that these mutations (sdh, zwf, gdhA, mtfA, fumAC, and cyaA) impacted the effect of AICAR on the ThiC reaction. It is interesting that all of these mutations impacted some aspect of the major central carbon metabolic pathways in the cell (e.g., tricarboxylic acid cycle, pentose phosphate pathway). The remaining two mutations (aspCp831 and mutL) appeared to specifically increase AICAR-dependent AIR synthesis. While mutL was not pursued due to its mutator phenotype, the aspCp831 mutation altered the translational level of AspC and effectively decreased aspartate levels. As aspartate is a substrate for the canonical forward reaction of PurC, decreased aspartate levels are expected to shift the equilibrium of the forward and reverse PurC reactions toward the reverse reaction. The aspCp831 mutation, therefore, may support the in vitro data that suggested that AICAR is converted to CAIR via a reaction analogous to the reverse reaction catalyzed by PurC. Importantly, the AspC reaction is the first reaction in a cascade of amino acid biosynthetic pathways, and lowered AspC activity would be expected to have rippling effects on various pathways that may enhance AICAR-dependent AIR synthesis by a distinct or additional mechanism. Significantly, the above conclusions are based on the interpretation of potentially indirect effects of a mutation and thus do not identify specific biochemically testable models.

The ability of S. enterica to satisfy its thiamine requirement using AICAR-dependent AIR synthesis requires AICAR accumulation and metabolic adjustments to balance the opposing effects of AICAR. Here, we have shown that mutations, exogenous nutrients, or both can achieve this balance (Fig. 7). This suggests that while AICAR-dependent AIR synthesis may not significantly contribute to thiamine synthesis in the wild-type strain with normal purine biosynthetic flux, the metabolic network of S. enterica is amenable to changes that can activate AICAR-dependent AIR synthesis. The understanding of the metabolic network of this system has been expanded with genetic and biochemical data from many studies. The understanding of the complexity of this network continues to expand, and continued genetic analyses are not likely to answer remaining questions in a timely manner. The investigation and ultimate understanding of the perturbations that impact the system, and their metabolic and nutritional consequences, will benefit from the implementation of a systems approach that can result in testable predictions. The integration of these new approaches while maintaining the power of traditional biochemical genetics analysis will contribute to a broader understanding of metabolic network robustness and plasticity.