PHB Biosynthesis Counteracts Redox Stress in Herbaspirillum seropedicae

Abstract

The ability of bacteria to produce polyhydroxyalkanoates such as poly(3-hydroxybutyrate) (PHB) enables provision of a carbon storage molecule that can be mobilized under demanding physiological conditions. However, the precise function of PHB in cellular metabolism has not been clearly defined. In order to determine the impact of PHB production on global physiology, we have characterized the properties of a ΔphaC1 mutant strain of the diazotrophic bacterium Herbaspirillum seropedicae. The absence of PHB in the mutant strain not only perturbs redox balance and increases oxidative stress, but also influences the activity of the redox-sensing Fnr transcription regulators, resulting in significant changes in expression of the cytochrome c-branch of the electron transport chain. The synthesis of PHB is itself dependent on the Fnr1 and Fnr3 proteins resulting in a cyclic dependency that couples synthesis of PHB with redox regulation. Transcriptional profiling of the ΔphaC1 mutant reveals that the loss of PHB synthesis affects the expression of many genes, including approximately 30% of the Fnr regulon.

Keywords: Fnr; bacterial signal transduction; polyhydroxyalkanoates; redox regulation; transcriptomics.

FIGURE 1

The lack of PHB production in H. seropedicae ΔphaC1 influences the levels of c-type cytochromes. (A) Cell suspensions of the wild-type (SmR1), triple fnr mutant (MB231) and phaC1 mutant (ΔphaC1) strains showing the differences in color associated with differential expression of c-type cytochromes. (B) Loading control (stained with Coomassie Brilliant Blue) from the protein extracts used for the ortho-dianisidine staining. (C) Levels of c-type cytochromes in the wild-type (SmR1), triple fnr mutant (MB231), and phaC1 mutant (ΔphaC1) strains were analyzed after separation of protein extracts in Tris-Tricine PAGE followed by o-dianisidine staining for detection of covalently bound heme. The gels in panels (B,C) are representative of four biological replicates. The numbers in panels (B,C) indicate the molecular weight of protein markers in kDa. (D) The expression profile of the fixN-lacZ transcriptional reporter fusion was tested in SmR1, MB231, and ΔphaC1 strains by measuring β-galactosidase activity according to (Miller, 1972; Batista et al., 2013). Results are the mean ± SD from three biological replicates.

FIGURE 2

The Fnr proteins influence PHB production. PHB levels were quantified from cells grown in NFbHP-Malate medium supplemented with 20 mM NH₄Cl, after acid methanolysis and gas chromatography of the wild-type strain (SmR1) and phaC1 mutant (ΔphaC1) when cultivated under low aeration (120 rpm) (A) or under high aeration (350 rpm) (B). Samples were taken either in the mid-log phase or the late-log phase of growth as indicated. The results are representative of two independent experiments performed in biological triplicates. ^∗p < 0.01, and ns—not statistically significant (p > 0.05) according to an unpaired t-test.

FIGURE 3

Oxidative stress is higher in the phaC1 mutant. ROS levels were analyzed using the H₂-DCFDA probe and flow cytometry in the wild-type strain (SmR1) and phaC1 mutant (ΔphaC1) in either the mid-log phase (A) or the late-log phase (B) of growth. The control assay corresponds to measurement of ROS levels in the absence of added compounds. H₂O₂ and N-acetyl-L-cysteine (Nac) indicate respectively, treatment of cells with 5 mM hydrogen peroxide or 5 mM Nac, prior to determination of ROS levels. Data represent the average (±SEM) of two biological replicates analyzed in four technical replicates. ^∗∗p < 0.001, ^∗∗∗p < 0.0001, and ns—not statistically significant (p > 0.05) according to an unpaired t-test.

FIGURE 4

The phaC1 mutant is more sensitive to methyl viologen. To determine the sensitivity of H. seropedicae strains to the superoxide generator methyl viologen, the wild-type (A) and the phaC1 mutant (B) strains were cultured in liquid media to an OD_600nm of 0.5, when different concentrations of superoxide were added. In panels (A,B), the closed squares indicate the growth of the strains without methyl viologen addition, while the circles, diamonds and open squares, indicate the growth profile upon addition of 0.5, 1.0 and 2.0 mM of methyl viologen, respectively. In panel (C), the percentage of growth rate decay between the wild-type (black lines—squares) and phaC1 mutant (gray lines—circles) is shown. Typical growth rates (100%) were 0.46 ± 0.003 /h for the wild-type and 0.43 ± 0.002 /h for the phaC1 mutant in the absence of methyl viologen. The data is representative of three independent biological replicates. ^∗p < 0.0001 according to an unpaired t-test.

FIGURE 5

The H. seropedicae ΔphaC1 strain produces more superoxide than the wild-type (SmR1). Cells were grown in NFbHP-Malate media supplemented with 20 mM of NH₄Cl until the late-log phase (OD_600nm = 1.0), treated with the superoxide specific probe DHE and then visualized by fluorescence microscopy as described in Section “Materials and Methods.” In panel (A), one representative image for each strain is shown. In panel (B), the relative fluorescence quantification is shown. The quantitative analysis was made as described in Section “Materials and Methods.” The results are given as the mean ± standard deviation from quantifications performed for four different images from each strain (two for each biological replicate). AU indicates arbitrary units of normalized fluorescence intensity. ^∗p < 0.0001 according to a two-sample t-test.

FIGURE 6

Activity of the pfixN-lacZ reporter fusion upon addition of methyl viologen (MV) or H₂O₂. (A) Representative growth profile of H. seropedicae SmR1 upon addition of MV. Samples were collected at the indicated times for measuring the pfixN-lacZ expression in panel (C). (B) Representative growth profile of H. seropedicae SmR1 upon addition of H₂O₂. Samples were collected at the indicated times for measuring the pfixN-lacz activity in panel (D). (C,D) Activity of the Fnr-dependent transcriptional fusion pfixN-lacZ was assayed upon addition of MV or H₂O₂ respectively. The arrows, in panels (A,B), indicate the addition of different amounts of MV or H₂O₂ as indicated. The color codes for symbols and bars in the graphs are as follows: white, growth under high aeration or 350 rpm (350); gray, growth under low aeration or 120 rpm (120); black, growth upon switch from 350 to 120 rpm (switch); yellow, switch from 350 to 120 rpm and addition of 0.2 mM of MV (s-0.2 MV); blue, switch from 350 to 120 rpm and addition of 0.5 mM of MV (s-0.5 MV); orange, switch from 350 to 120 rpm and addition of 1.0 mM of MV (s-1.0 MV); green, switch from 350 to 120 rpm and addition of 0.2 mM of H₂O₂ (s-0.2 H₂O₂); red, switch from 350 to 120 rpm and addition of 0.5 mM of H₂O₂ (s-0.5 H₂O₂); dark blue, switch from 350 to 120 rpm and addition of 1.0 mM of H₂O₂ (s-1.0 H₂O₂).

FIGURE 7

Levels of adenine nucleotides in the wild-type SmR1 and ΔphaC1 strains. The metabolites were analyzed by LC-MS as described in Section “Materials and Methods” and the peak heights were compared. Box plots represent the data of three biological replicates analyzed in three technical replicates. ^∗p < 0.00001 according to a two-sample t-test.

FIGURE 8

Levels of malate and acetate detected in the supernatant of cultures from the wild-type SmR1 and ΔphaC1 strains. Both metabolites were quantified using NMR spectroscopy as described in Section “Materials and Methods.” In panels (A,B), the time course of malate consumption during different stages of growth are shown in the wild-type (SmR1) and ΔphaC1 strains, respectively. In panel (C), the relative levels of acetate detected in the supernatants of the wild-type (red dots) and ΔphaC1 (blue dots) are shown.