Serratia marcescens quinoprotein glucose dehydrogenase activity mediates medium acidification and inhibition of prodigiosin production by glucose

Abstract

Serratia marcescens is a model organism for the study of secondary metabolites. The biologically active pigment prodigiosin (2-methyl-3-pentyl-6-methoxyprodiginine), like many other secondary metabolites, is inhibited by growth in glucose-rich medium. Whereas previous studies indicated that this inhibitory effect was pH dependent and did not require cyclic AMP (cAMP), there is no information on the genes involved in mediating this phenomenon. Here we used transposon mutagenesis to identify genes involved in the inhibition of prodigiosin by glucose. Multiple genetic loci involved in quinoprotein glucose dehydrogenase (GDH) activity were found to be required for glucose inhibition of prodigiosin production, including pyrroloquinoline quinone and ubiquinone biosynthetic genes. Upon assessing whether the enzymatic products of GDH activity were involved in the inhibitory effect, we observed that d-glucono-1,5-lactone and d-gluconic acid, but not d-gluconate, were able to inhibit prodigiosin production. These data support a model in which the oxidation of d-glucose by quinoprotein GDH initiates a reduction in pH that inhibits prodigiosin production through transcriptional control of the prodigiosin biosynthetic operon, providing new insight into the genetic pathways that control prodigiosin production. Strains generated in this report may be useful in large-scale production of secondary metabolites.

Fig 1

Growth, prodigiosin production, and culture pH of WT (CMS376) and gdhS mutant (CMS1083) cells grown in LB and LBG. Genotype symbols are consistent for each panel. The gdhS strain is the rig-1 mutant with a wild-type crp gene. Cultures were grown in LB and in LBG (LB supplemented with 110 mM glucose). Culture turbidity (OD₆₀₀ nm) (A), prodigiosin content (A₅₃₄) (B), and culture pH (C) were measured. Time course experiments were performed with 4 independent biological replicates per time point per condition and repeated two times with consistent results. Mean values from one representative time course are shown. Error bars indicate one standard deviation.

Fig 2

Model for the role of quinoprotein glucose dehydrogenases in extracytoplasmic glucose metabolism. d-Glucose is oxidized to d-glucono-1,5-lactone by GdhS in the periplasm, a process requiring both pyrroloquinoline quinone (PQQ) and ubiquinone (1, 32). Ubiquinone is reduced, generating ubiquinol, a molecule that is then converted to ubiquinone by one of several possible ubiquinol oxidases, creating a proton gradient (1, 32). d-Glucono-1,5-lactone is hydrolyzed to d-gluconic acid nonenzymatically and possibly through the action of a theoretical gluconolactonase enzyme. Production of d-glucono-1,5-lactone, its subsequent change to d-gluconic acid, and the establishment of a proton gradient are expected to decrease extracellular pH.

Fig 3

Glucose dehydrogenase is necessary for glucose inhibition of prodigiosin (GIP). Complementation of the rig-10 mutant (Δcrp gdhS) glucose insensitivity phenotype using the wild-type S. marcescens gdhS gene or the E. coli gcd gene expressed from a medium-copy-number plasmid. The vector pMQ132 was used as a negative control. Cultures were incubated at 30°C for 20 h; “− glucose” indicates growth in LB (black bars); “+ glucose” indicates growth in LBG (gray bars). Prodigiosin levels (A₅₃₄), normalized by culture turbidity (OD₆₀₀ nm) (A), and culture pH (B) are shown. The charts show an average of three independent biological replicates per condition and genotype, and error bars indicate one standard deviation. The experiment was performed twice with consistent results. Strain CMS1687 with pMQ132 served as the crp mutant control.

Fig 4

Gluconic acid production, GDH activity, and the effect of d-glucono-1,5-lactone on prodigiosin levels. “+ glucose” indicates growth in LBG; otherwise, strains were grown in LB. (A and B) Gluconic acid (GA) was measured from culture medium to assess GDH activity in different genetic backgrounds and culture conditions over a time course (A) or at 6 h of incubation at 30°C after being subcultured to an A₆₀₀ of 0.1 (B). (C and D) Glucose dehydrogenase activity from crude cellular extracts measured over time. (E) Prodigiosin production measured from d-glucono-1,5-lactone-treated cultures, normalized by culture density and measured at 20 h. (A, C, and E) The charts show the averages of four biological replicates per time point and condition from a representative experiment that was repeated on at least one other occasion with consistent results. (B) The chart depicts the combined data from two experiments performed on different days, using a total of 8 biological replicates per genotype. Error bars indicate 1 standard deviation. WT, CMS376; gdhS, CMS1083; crp, CMS1687; crp gdhS, rig-10; crp pqqE, rig-13; crp ubiB, rig-A21.

Fig 5

Pyrroloquinoline quinone (PQQ) is required for GIP. (A to C) Prodigiosin levels, normalized by culture turbidity, culture pH, and GDH activity of cellular extracts, were determined from cultures grown for 20 h in LBG with (+) or without (−) 3.03 μM PQQ. The crp mutant (Δcrp, CMS1687), crp pqqE mutant (rig-13), and crp gdhS mutant (rig-10) were used for this analysis. Panels A to C show the averages of 8 biological replicates per condition from two separate experiments. (C) Mock, no-protein control. Error bars indicate one standard deviation.

Fig 6

Complementation of the ubiB mutant phenotype. (A and B) Cultures of the Δcrp (CMS1687) and Δcrp ubiB (rig-A21) strains, bearing either an empty vector control (pMQ132) or the wild-type ubiB gene expressed from a medium-copy-number plasmid (pubiB), were grown in LB (− glucose) or LBG (+ glucose) for 20 h. Prodigiosin levels (A₅₃₄), normalized by culture turbidity (OD₆₀₀ nm) (A), and culture pH (B) are shown. The average pH values are shown for the rig-21 mutant. The experiments show the averages of 8 biological replicates per genotype from two experiments performed on different days. The error bars indicate one standard deviation. The asterisk indicates a significant difference between the strain with ubiB, pMQ132, and glucose and the strain with ubiB, pubiB, and glucose, as determined using a Student t test (P < 0.001).

Fig 7

Altered pigA transcription in glucose-rich medium in the WT but not a gdhS mutant. (A and B) β-Galactosidase activity was measured from a chromosomal pigA promoter-lacZ fusion that measured expression of the prodigiosin biosynthetic operon after growth for 20 h. (A) β-Galactosidase activity was measured from WT (CMS376) and gdhS mutant (CMS1083) cultures grown in LBG (+ glucose) and LB (− glucose). Data from eight biological replicates per condition from two separate experiments are shown. (B) Expression of the PpigA-lacZ fusion was inhibited by growth in d-glucono-1,5-lactone at >20 mM in both the WT and the gdhS (rig-1 with restored crp) mutant strains. A representative experiment using 4 biological replicates per condition is shown and was consistent with a repetition of the experiment on a different day. Error bars indicate one standard deviation, and the asterisk indicates that the WT + glucose condition produced a different amount of β-galactosidase activity than all other conditions, as determined by an ANOVA with a Tukey posttest (P < 0.05).