Indole-3-Acetic Acid Biosynthesis Is Deficient in Gluconacetobacter diazotrophicus Strains with Mutations in Cytochrome c Biogenesis Genes

Bacterial strains, vectors, and growth conditions.The bacterial strains and plasmids used are presented in Table 1. G. diazotrophicus wild-type PAl5 (ATCC 49037) and ccm mutants were maintained in either DYGS or LGIP medium (40). Escherichia coli strains were grown at 37°C in Luria-Bertani broth. Antibiotics for G. diazotrophicus were added at the following concentrations (in micrograms per milliliter): tetracycline, 100; kanamycin, 200; streptomycin, 700. For the isolation of membranes, cells were grown aerobically at 30°C in 3 liters of LGIP medium supplemented with 1.0 mM (NH₄)₂SO₄.

Mutagenesis.Transposon mutants were generated by conjugation of strain PAl5 with E. coli S17-1 carrying the suicide plasmid pSUP1021, which contains a Tn5 transposon that confers kanamycin resistance. Conjugations were performed on DYGS (pH 6) plates and incubated for 36 h at 30°C. After conjugation the cells were resuspended in LGI medium, diluted, and plated onto DYGS medium with kanamycin. Transconjugants were picked and stored as libraries in 96-well microtiter plates. The site-directed insertion mutants were generated using the Ω fragment (streptomycin) as described previously (41).

Tn5 screening.Tn5-containing transconjugants were grown in minimal LGI medium supplemented with tryptophan (100 μg/ml) as an IAA precursor in 96-well microtiter plates, and after 4 days of incubation the amount of IAA produced by each transconjugant was determined using Salkowasky reagent (18). IAA mutant candidates were further characterized by thin-layer chromatography (TLC) and high-performance liquid chromatography (HPLC) analysis.

HPLC analysis.Extraction and HPLC of G. diazotrophicus cultures for IAA were performed according to the methods of Costacurta et al. (11) with slight modifications. Bacterial cultures (20 ml) were made cell free by centrifugation at 3,000 × g and filtration, and supernatants were extracted three times with ethyl acetate after adjusting the pH to 2.8. Five- to 15-μl aliquots of the filtered extracts were injected into an Alltech, type Econosphere C185U column (250 by 4.6 mm) equipped with a differential UV detector absorbing at 280 nm. The isocratic solvent used for reverse-phase chromatography was acetonitrile-glacial acetic acid (1%) in water (10:90). The flow rate was adjusted to 1 ml/min. Peak retention times were compared with those of chemically synthesized IAA standards and quantified by comparison of peak areas.

Analytical methods.Preparation of membrane proteins, sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), staining for heme, spectral analysis of cytochromes, and determination of dehydrogenase activities and respiratory activities were performed as described by Flores-Encarnacion et al. (16) with a few modifications. Briefly, cells in 3 liters of liquid culture were obtained after 36 h of growth with shaking at 250 rpm. Cells were pelleted by centrifugation and then washed twice with TCM buffer (50 mM Tris-HCl [pH 7.4] containing 5 mM CaCl₂ and 5 mM MgCl₂). Membranes were isolated, and protein concentrations were measured by a modification of the Lowry method. For spectral analysis, membranes (8 mg) were resuspended in 50% (vol/vol) glycerol and analyzed in an SLM-Aminco DW 2000 spectrophotometer. Samples were reduced with a few grains of sodium dithionite in the presence or absence of KCN, and difference spectra at 77K were recorded. The c-type cytochrome spectra were measured after membranes were precipitated with 0.01 N HCl in acetone as described by Goodhew et al. (22). Heme peroxidase staining was performed to detect c-type cytochrome bands. Membranes (3 mg) were washed with trichloroacetic acid (10% [vol/vol]) and separated by SDS-PAGE. SDS treatment removed noncovalently bound hemes, so that only c-type cytochrome was revealed after heme peroxidase staining. Protein blotting and detection were performed with the Amersham (Piscataway, N.J.) enhanced chemiluminescence Western blotting detection reagents.

Oxidase activities were determined as described by Flores-Encarnacion et al. (16) with various substrates at the following final concentrations: 3 mM NADH, 10 mM glucose, 10 mM ethanol, 10 mM acetaldehyde, 10 mM ascorbate plus 2 mM N,N,N′,N′-tetramethyl-p-phenylenediamine (TMPD) or 10 mM ascorbate plus 1.5 mM 2,3,5,6-tetrachloro-1,4-benzenediol (THQ). The oxidase activities were measured polarographically with a Clark oxygen electrode in 2 ml of 50 mM potassium phosphate buffer (pH 7.4 or 6.0) at 30°C. Dehydrogenase activities were measured spectrophotometrically with TMPD as the alternative electron acceptor, as described by Flores-Encarnacion et al. (16).

DNA manipulations, sequencing, and analysis.The DNA region flanking the Tn5 integration was sequenced using the primer 5′-CGTTCAGGACGCTAC-3′, complementary to bases 17 to 34 within the Tn5 inverted repeat. Sequencing was performed with an ABI automated sequencer (model 373A) by using a PRISM Ready Reaction Dye Deoxy terminator cycle sequencing kit (Perkin-Elmer, Boston, Mass.; DNA sequencing facility, University of Arizona). DNA sequences were identified by using the BLAST server of the National Center for Biotechnology Information accessed over the Internet. Computer-assisted sequence analysis was performed using programs of the University of Wisconsin package version 10.0 Genetics Computer Group software.

Production of IAA by wild-type strain PAl5.In an attempt to elucidate the IAA biosynthetic pathway(s) in G. diazotrophicus, IAA and possible intermediates were analyzed by HPLC and TLC. The HPLC elution profiles of acidic ethyl acetate extracts contained two minor peaks and three major peaks: two early-eluting unknown substances at 9 and 11 min, ILA, IAA, and indole-3-carboxylic acid (ICA) (Fig. 1). The sharp peak in the beginning of the chromatogram is an internal standard. TLC of ethyl acetate extracts from G. diazotrophicus cultures showed IAA, ILA, and TOL, but no ICA could be detected by this method (data not shown). ILA and TOL are intermediates of the IpyA pathway. ICA is a decarboxylated, relatively inactive form of IAA. Therefore, the results indicated that G. diazotrophicus produces IAA through the IpyA pathway. Quantification of the indole derivatives was achieved by measuring the peak area of the HPLC chromatogram, and yields of 5 to 7 μg of IAA/ml were found in the cell-free supernatants of stationary cultures of G. diazotrophicus PAl5.

• Open in new tab
• Download powerpoint

FIG. 1.

Reversed-phase HPLC chromatogram of the culture fluid of G. diazotrophicus wild-type PAl5 (a) and IAA⁻ mutant strain MAd22 (b) after 4 days of growth. The cultures were grown in minimal medium (LGI) supplemented with 100 μg of Trp/ml. AU, absorption unit.

Isolation and genetic analysis of a mutant deficient in IAA biosynthesis.Various approaches to clone an indole-3-pyruvic acid decarboxylase gene (ipdC gene), a key enzyme in the IpyA pathway, were not successful. In order to isolate genes encoding enzymes involved in IAA biosynthesis, a different genetic approach was adopted using the transposon Tn5. Following Tn5 mutagenesis, 2,500 exconjugants were analyzed for IAA production as described in Materials and Methods. Three mutant candidates with decreased IAA and one IAA-overproducing mutant candidate were further characterized by HPLC and TLC (data not shown). Among IAA⁻ mutant candidates, MAd10 was further characterized, because MAd10 excreted only ∼6% of the amount of IAA produced by the parent strain in liquid culture and it is neither a Trp auxotroph nor a Nif⁻ strain (data not shown). The Tn5-linked region of MAd10 was cloned and sequenced; it encoded gene products with a high similarity to several antibiotic synthetase enzymes from Bacillus subtilis, Streptomyces sp., and Pseudomonas sp. (27, 29, 39). This result suggested that the IAA⁻ phenotype of MAd10 was not linked to the Tn5 insertion site. When the region containing the Tn5 and flanking DNA from MAd10 was inserted into the chromosome of wild-type strain PAl5, the resulting strain was IAA⁺. This also indicated that the Tn5 insertion in the antibiotic synthetase region in MAd10 was not linked to the IAA⁻ phenotype.

Complementation of the mutant MAd10.To determine where the mutation causing the IAA⁻ phenotype was located on the chromosome, a pLAFR3 cosmid library of G. diazotrophicus was introduced into MAd10 (46). Two cosmids complementing the IAA⁻ phenotype of MAd10 were isolated after screening 2,000 tetracycline-resistant transconjugants using the colorimetric method described earlier. These cosmids had overlapping restriction sites, which indicated that they probably contained the same gene or genes complementing decreased IAA production in MAd10. The region from one of the cosmids, pSL10, was subcloned to determine the smallest fragment that complemented the IAA⁻ phenotype. A 5.6-kb HindIII/EcoRI fragment (pSL12) successfully complemented the IAA⁻ phenotype, restoring wild-type IAA levels (Table 2).

Sequence analysis.To determine the nucleotide sequence, pSL12 was subcloned to smaller fragments and each was sequenced. Examination of the sequence revealed five contiguous open reading frames (ORFs) (Fig. 2), the deduced products of which had high similarity to proteins involved in cytochrome c biogenesis (ccm genes [cytochrome c maturation genes]) in E. coli (37 to 48% identity) or cyc genes in Bradyrhizobium japonicum (38 to 52% identity). The ccm operon sequence is reported under GenBank accession number AY456185 . Products of the ccmABCDEFGH genes in E. coli are required for the formation of holocytochrome c, the products being subunits of an ABC transporter (CcmABC), a heme chaperone (CcmE), a putative cytochrome c heme lyase (CcmEF), a postulated redox system (CcmGH), and a membrane protein that stabilizes the complex (CcmD) (reviewed in references 26, 48, and 49). In B. japonicum these genes are named cyc and are arranged somewhat differently than the ccm genes in E. coli and now in G. diazotrophicus (data not shown).

• Open in new tab
• Download powerpoint

FIG. 2.

Physical and restriction maps of G. diazotrophicus ccm operon and ccm mutants. (Top) The 5.6-kb G. diazotrophicus genomic fragment harboring the ccm operon, with restriction sites indicated. B, BamHI; Bg, BglII; E, EcoRI; H, HindIII; K, KpnI; M, MscI; Sp, SphI; St, StuI. (Bottom) A series of ccm mutants in which ccm genes were disrupted by deletion or insertion of a streptomycin (Ω) cassette.

Characterization of the ccm mutants.To confirm that one or more of the ccm genes are responsible for the IAA⁻ phenotype, several mutant strains were generated by inserting Ω (Sm^r) cassette and/or by deleting a portion of the genes to yield the strains MAd20, MAd21, and MAd22 (Fig. 2). Mutagenesis results indicated that a mutation in any of the ccmCEF genes results in the IAA⁻ phenotype.

IAA production by both ccm mutants and the wild type was investigated by HPLC analysis. The ccm mutants produced a barely detectable amount of IAA (4 to ∼6% of the wild type) after 4 days of growth. The detectable IpyA pathway intermediate, ILA, was also reduced significantly compared with wild type (Fig. 1; Table 2). The mobilization of a plasmid (pSL12) containing the functional ccm operon restored the ability to produce IAA in the three new mutants (Table 2). This indicated that a defect in cytochrome c production is responsible for the IAA mutant phenotype.

The cytochrome c content of G. diazotrophicus cells was characterized with membranes obtained from cells as described by Flores-Encarnacion et al. (16). Noncovalently bound hemes were removed by washing membranes with 0.01 N HCl in acetone to determine cytochrome c content without the spectral interference of b-type cytochromes. The dithionite-reduced minus persulfate-oxidized spectra (77K) of acidic acetone-treated membranes from wild type and ccm mutants were determined (Fig. 3B). The spectra of wild-type membranes revealed the characteristic peaks at 417, 520, and 549.5 nm originating from the c-type cytochromes, while these peaks were absent or significantly reduced in the ccm mutants (Fig. 3B). Different spectra produced by cyanide treatment of the reduced preparation revealed the presence of an a-type cytochrome in both ccm mutants and the wild type, which was accompanied by a large enhancement of the signals at 589 nm. Additionally, b-type cytochromes were retained in ccm mutant membranes, as suggested by shoulders at 430, 529, and 560 nm in reduced plus CN⁻ minus oxidized spectra (Fig. 3A).

• Open in new tab
• Download powerpoint

FIG. 3.

Difference spectra at 77K for c-type cytochromes associated with membranes of G. diazotrophicus PAl5 and its ccm mutants (MAd10, MAd20, MAd21, and MAd22). (A) Reduced plus KCN minus oxidized spectra. Samples were reduced with dithionite in the presence of 1.0 mM KCN and oxidized with air. (B) Dithionite-reduced minus persulfate-oxidized spectra of acidic acetone-treated membranes. An 800-μg aliquot of protein was used in each sample.

To analyze the ccm mutant phenotypes in more detail, membrane fractions of aerobically grown wild-type and mutant cells were analyzed on SDS-PAGE for the presence of heme-associated peroxidase activity (heme staining). In the wild-type cells, four c-type cytochromes with apparent molecular masses of 67, 56, 52, and 45 kDa were detected (Fig. 4). No membrane-bound c-type cytochromes were detected in membranes prepared from the ccm mutants.

• Open in new tab
• Download powerpoint

FIG. 4.

Membrane-bound c-type cytochromes of G. diazotrophicus PAl5 and its ccm mutants detected by heme staining after SDS-PAGE. Each lane contains membrane protein samples. Lane a, PAl5 (200 μg); lanes b to e, ccm mutants MAd10, MAd20, MAd21, and MAd22 (1 mg each); lane f, horse heart cytochrome c (10 μg).

Since in many organisms cytochrome c proteins are involved in respiration, oxidase, and dehydrogenase activities, these activities of the wild-type and mutant strains were measured polarographically with a Clark oxygen electrode and spectrophotometrically with TMPD as an electron acceptor. In decreasing order, glucose, NADH, acetaldehyde, ascorbate plus THQ, ethanol, ascorbate plus TMPD, and lactate were the best substrates for the oxidase activities associated with G. diazotrophicus membranes (Table 3). In the case of ccm mutants, no ethanol, acetaldehyde, and lactate oxidase activities were detected; however, similar glucose and NADH oxidase activities were retained. The dehydrogenase activities of ccm mutants were significantly decreased when acetaldehyde, ethanol, or lactate was used as substrate, but the activities with glucose or NADH were similar to those of the wild type (Table 4).