Correspondence: Johan H.J. Leveau, Netherlands Institute of Ecology (NIOO-KNAW), Boterhoeksestraat 48, 6666 GA Heteren, The Netherlands. Tel.: +31 26 479 1306; fax: +31 26 472 3227; e-mail email@example.com
The isolation and annotation of an 8994-bp DNA fragment from Pseudomonas putida 1290, which conferred upon P. putida KT2440 the ability to utilize the plant hormone indole 3-acetic acid (IAA) as a sole source of carbon and energy, is described. This iac locus (for indole 3-acetic acid catabolism) was identified through analysis of a plasposon mutant of P. putida 1290 that was no longer able to grow on IAA or indole 3-acetaldehyde and was unable to protect radish roots from stunting by exogenously added IAA. The iac locus consisted of 10 genes with coding similarity to enzymes acting on indole or amidated aromatics and to proteins with regulatory or unknown function. Highly similar iac gene clusters were identified in the genomes of 22 bacterial species. Five of these, i.e. P. putida GB-1, Marinomonas sp. MWYL1, Burkholderia sp. 383, Sphingomonas wittichii RW1 and Rhodococcus sp. RHA1, were tested to confirm that bacteria with IAA-degrading ability have representatives in the Alpha-, Beta- and Gammaproteobacteria and in the Actinobacteria. In P. putida 1290, cat and pca genes were found to be essential to IAA-degradation, suggesting that IAA is channeled via catechol into the β-ketoadipate pathway. Also contributing to the IAA degrading phenotype were genes involved in tricarboxylate cycling, gluconeogenesis, and carbon/nitrogen sensing.
The existence of this IAA-degrading phenotype can be explained in two, not mutually exclusive, ways. On the one hand, IAA is just one of many compounds that bacteria have adapted to as sources of carbon and/or nitrogen. Complete mineralization of IAA and growth on IAA have been reported for representatives of the genera Pseudomonas (Gieg et al., 1996; Leveau & Lindow, 2005), Arthrobacter (Mino, 1970), Alcaligenes (Claus & Kutzner, 1983) and Bradyrhizobium (Jarabo-Lorenzo et al., 1998). For these bacteria, plants and many plant-associated microorganisms (Costacurta & Vanderleyden, 1995; Patten & Glick, 1996; Spaepen et al., 2007) represent natural sources of IAA. On the other hand, IAA is not just any compound: as a growth hormone, it has an essential biological function in plants, governing many physiological processes (Teale et al., 2006). Thus, mineralization of IAA, or transformation into a biologically inactive form, offers bacteria the potential for manipulation of IAA-related plant activities. An example from the fungal world is Marasmius perniciosus, which causes witches' broom disease on cacao, a condition that is characterized by an imbalance in IAA levels resulting from IAA-oxidizing enzymes produced by the fungus (Krupasagar & Sequeira, 1969). A functionally homologous phenomenon is the ability of Pseudomonas savastanoi to convert IAA to the biologically inactive IAA–lysine (Kuo & Kosuge, 1969). This bacterium is also one of many that produce IAA and can thus sabotage plant physiology by adding to the endogenous IAA pool in plants (Costacurta & Vanderleyden, 1995; Patten & Glick, 1996; Spaepen et al., 2007).
Several pathways have been suggested for the bacterial mineralization or transformation of IAA (Fig. 1). Tsubokura et al. (1961) described a bacterium isolated from air that attacks IAA by splitting the indole ring to produce 2-formaminobenzoylacetic acid. Bradyrhizobium japonicum metabolizes IAA via dioxindole-3-acetic acid, dioxindole, isatin, isatinic acid and anthranilic acid (Jensen et al., 1995). This pathway overlaps partially with the mineralization of IAA by an Alcaligenes sp. via isatin, anthranilic acid and gentisate (Claus & Kutzner, 1983). Decarboxylation of IAA to skatole has been reported for several bacteria (Yokoyama et al., 1977; Deslandes et al., 2001), including a Pseudomonas sp. from soil, that converts skatole to catechol (Proctor, 1958). The latter is a confirmed intermediate in the mineralization of IAA by Pseudomonas putida 1290 (Leveau & Lindow, 2005), Pseudomonas sp. LD2 (Gieg et al., 1996) and an Arthrobacter species (Mino, 1970). For the two Pseudomonas species, it was shown that catechol is ortho-cleaved to cis,cis-muconic acid and channeled into the β-ketoadipate pathway (Harwood & Parales, 1996). Other types of bacterial biotransformation of IAA are (1) transformation to indole-3-methanol by Rhizobium phaseoli (Ernstsen et al., 1987) and (2) conjugation, as exemplified by the production of IAA-lysine by Pseudomonas savastanoi (Kuo & Kosuge, 1969).
Surprisingly, almost nothing is known about the genetics that underlie the IAA degradative pathways described above. Except for the iaaL gene from P. savastanoi, encoding the IAA-lysine synthase (Glass & Kosuge, 1986), no other genes have been linked to IAA-catabolic enzyme activities. This lack of knowledge has seriously hampered studies of the biology, ecology, and evolution of the bacterial IAA degradative phenotype. Without DNA sequences, no tools exist to assess and quantify the occurrence, diversity, and activity of IAA-degradative genes in natural bacterial populations. Furthermore, the unavailability of IAA degradative genes has held back exploitation of such genes in the study of IAA functioning in plants, which actually owes much of its progress to bacterial genes encoding the synthesis of IAA (Hedden & Phillips, 2000).
In the current study, the authors' aim was to be the first to identify and describe bacterial genes involved in the biotransformation and mineralization of IAA. For this, Pseudomonas putida 1290 was chosen as a model strain (Leveau & Lindow, 2005). This bacterium was shown previously to use IAA efficiently as the sole source of carbon, nitrogen, and energy. Here, it is described how a combination of plasposon mutagenesis and functional complementation was used to identify a gene locus in P. putida 1290 that carries genes encoding IAA degradative activity.
Materials and methods
Bacterial strains and growth conditions
The bacterial strains used in this study are listed in Table 1. Pseudomonas strains were grown at 30 °C on King's B (KB) medium (King et al., 1954) or on M9 minimal medium (Sambrook et al., 1989) supplemented with one of the following carbon sources at a final concentration of 5 mM: IAA (Sigma-Aldrich, Steinheim, Germany), indole 3-acetaldehyde-sodium bisulfite addition compound, (Sigma-Aldrich), indole 3-acetamide (Sigma-Aldrich), benzoate (Sigma-Aldrich) or glucose. Escherichia coli strains were grown at 37 °C on Luria–Bertani (LB) medium. Where applicable, antibiotics were used at the following final concentrations: chloramphenicol, 12.5 μg mL−1 (Cm12.5); kanamycin, 50 μg mL−1 (Km50); and rifampicin, 40 μg mL−1 (Rif40).
Table 1. Strains, plasmids, and primers used in this study
References or sources
Pseudomonas putida 1290
Able to use IAA, but also benzoate and glucose, as sole source of carbon and energy
Well-characterized soil bacterium; genome sequence available; not able to use IAA as carbon and energy source; obtained from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (Braunschweig, Germany)
pCC1FOS containing genomic fragment of P. putida 1290; confers IAA+ phenotype to P. putida KT2440
pCC1FOS containing genomic fragment of P. putida 1290; confers IAA+ phenotype to P. putida KT2440
pCC1FOS containing genomic fragment of P. putida 1290; confers IAA+ phenotype to P. putida KT2440
pCC1FOS containing genomic fragment of P. putida 1290; confers IAA+ phenotype to P. putida KT2440
pCC1FOS containing genomic fragment of P. putida 1290; contains cat locus
pCC1FOS11 digested with EcoRI and religated
pCC1FOS11Eco digested with MluI and religated
pCC1FOS11Eco digested with XhoI and religated
pCC1FOS11Eco digested with PspXI and religated
pCC1FOS11Eco digested with AscI and religated
AscI/EcoRI digested pCC1FOS11EcoAsc ligated to 1.2-kb AscI/EcoRI fragment of pO4-H06
3.7-kb genomic fragment of P. putida 1290 containing ORF 18 in pCR4Blunt-TOPO
PCR amplification of iacH locus
Construction and screening of a P. putida 1290 plasposon library
Plasposon pTnMod-KmOlacZ (Table 1) was used to create a mutant library of P. putida 1290R. To this end, equal amounts of exponentially growing cells of P. putida 1290R, E. coli DH5α (pTnMod-KmOlacZ), and helper strain E. coli DH5α (pRK2013) were mixed, spotted onto a KB plate, and incubated at 30 °C for 24 h. Cells were then recovered from the plate, resuspended in 10% glycerol, and plated onto M9 medium containing 5 mM glucose, Rif40, and Km50. Of the resulting P. putida 1290R∷KmOlacZ colonies, 4416 were transferred to fresh M9 glucose Rif40 Km50 plates, and then streaked onto M9 plates containing Km50 and either IAA or benzoate. Mutants unable to grow on IAA or benzoate were selected for further study. Their genomic DNA was isolated using an Ultra Clean Soil DNA Kit (Mobio, Carlsbad, CA), digested with PstI (Amersham Life Sciences, Piscataway, NJ), religated with T4 DNA ligase (New England Biolabs, Beverly, MA), and used to electrotransform E. coli EP-Max 10B cells with a Gene Pulser Xcell Microbial System (Bio-Rad, Hercules, CA) according to the manufacturer's instructions. From the resulting transformants, plasmid DNA was isolated using a QIAprep Spin Miniprep Kit (Qiagen Benelux, Venlo, the Netherlands) and used as template DNA in a sequencing reaction (BaseClear, Leiden, the Netherlands) with primers LAC or PP1 (Table 1). Thus, for each mutant the nucleotide sequence of the DNA surrounding the plasposon insertion site was obtained. These sequences, between 228 and 1008 bp in length, were analyzed by lasergene software (DNASTAR, Madison, WI) and used as a query in blast searches at the website of the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov).
Screening of a large-insert DNA library from P. putida 1290 genomic DNA
A CopyControl Fosmid Library Production Kit was used (Epicentre, Madison, WI) to construct a large-insert DNA library of P. putida 1290 genomic DNA as described earlier for Collimonas fungivorans Ter331 (Leveau et al., 2004). This library consisted of 3072 E. coli EPI300 clones, each carrying a random ∼38-kb DNA fragment of the P. putida 1290 genome in vector pCC1FOS (Table 1). A subset of 768 clones, grown overnight in eight microtiterplates containing 150 μL LB Cm12.5 well−1, was screened by PCR for DNA sequences containing the iacH gene (see below) as follows: cells from every row of each microtiterplate were pooled resulting in 64 pools of 12 clones. Pooled cells were pelleted by centrifugation, resuspended in milli-Q water, incubated at 99 °C for 10 min, and centrifuged again. The supernatant was diluted 30 × in milli-Q water and 4.5 μL was used as template DNA in a 15-μL PCR reaction containing 7.5 μL ABsolute QPCR SYBR Green Mix (ABGene, Hamburg, Germany), 2.7 μL of a 2.5 μM iacH locus-specific primer set (Table 1), and 0.3 μL milli-Q water. PCR reactions were set up with a CAS-1200 liquid handling robot (Corbett Research, Sydney, Australia). PCRs were run in a Rotor-Gene 3000 (Corbett Research) using the following settings: 15 min at 95 °C, and then 40 cycles of 15 s at 95 °C, 30 s at 60 °C, and 30 s at 72 °C. Positive pools were identified as those that produced the same fluorescent signal as the positive control (i.e. genomic DNA of P. putida 1290). Positive pools were split into individual clones for a second round of PCR. Individual positive clones were analyzed further by 10 × dilution of o/n cultures in LB Cm12.5 supplemented with 1 × induction solution and incubation for 5 h at 37 °C and 300 r.p.m. Five milliliters of this culture were used in a QIAprep Spin Miniprep Kit protocol (Qiagen). The resulting fosmid DNAs were used for restriction analysis with EcoRI, BamHI and/or HindIII (New England Biolabs, Beverly, MA), and as DNA templates in end-sequencing reactions using primers MR and MF-20 (Table 1).
Electrotransformation of P. putida strains
Fosmids were introduced into P. putida by electrotransformation as follows: an overnight KB culture was diluted 100 × into fresh KB medium and incubated at 30 °C for 3.5 h, placed on ice for 1 h, and centrifuged for 5 min at 4147 g at 4 °C. Cells were resuspended and washed three times in an equal volume, half volume, and quarter volume, respectively, of ice-cold 10% glycerol. After the final wash, cells were concentrated 50-fold in 10% glycerol. An aliquot of 100 μL was mixed with 8 μL of plasmid or fosmid DNA, transferred to a 0.2-cm cuvette, and electroporated with a Gene Pulser Xcell Microbial System (Bio-Rad) using the settings suggested for P. aeruginosa (i.e. 25 μF, 200 Ω and 2500 V). After addition of 1 mL SOC (Sambrook et al., 1989), cells were incubated for 3 h shaking at 30 °C, and plated on appropriate media.
Root elongation assays and IAA measurements
The ability of wild-type P. putida 1290 and selected plasposon-mutant derivatives to rescue radish roots (variety French Breakfast) from IAA-induced growth inhibition was tested in a root assay as described previously (Leveau & Lindow, 2005). Root lengths were analyzed in statistica (StatSoft, Tulsa, OK) by a standard t-test after passing Levene's test for homogeneity of variance. In supernatants of cultures that were incubated in the presence of IAA, IAA concentrations were measured using Salkowski reagent as described previously (Leveau & Lindow, 2005).
DNA sequencing and annotation
DNA preparations of four fosmids (i.e. pCC1FOS9, −10, −11 and −12) conferring an IAA-degradative (IAA+) phenotype to P. putida KT2440 were pooled and sent to Macrogen (Seoul, Korea), where two libraries were constructed: (1) a small-insert library of 384 clones in plasmid pCR4Blunt-TOPO, and (2) a 96-clone library of transposon-treated fosmids each carrying a random insertion of TN <rOri/kan2>. DNA sequences were obtained from both ends of each cloned insert in the pCR4Blunt-TOPO library or from both regions flanking each TN <rOri/kan2>transposon insertion. DNA sequence reads were assembled using Lasergene's SeqMan module (DNASTAR). FGENESB (SoftBerry, Mount Kisco, NY) and GenDB (Meyer et al., 2003) were used to annotate the consensus DNA of the largest contig.
Identification of an iac locus coding for indole 3-acetic acid catabolism
A total of 4416 plasposon mutants of P. putida 1290 were screened for reduced ability to grow on M9 minimal medium containing IAA as the sole source of carbon and energy. Twelve of such IAA− mutants were identified. Only mutant 26E2 completely failed to produce biomass on M9 IAA plates; all others formed tiny microcolonies. The DNA sequence surrounding the plasposon insertion site in each of the IAA− mutants was determined and analyzed by blastx (Table 2). Eleven mutants carried their plasposon in different ORFs with high coding similarity to genes identified previously in Pseudomonas putida strains. Mutant 26E2 carried the plasposon in an ORF provisionally called iacH, with coding similarity to several amidases from plants and bacteria, but without obvious homologs in any Pseudomonas species, with the exception of P. putida GB-1 (Corstjens et al., 1992; unpublished genome sequence with accession number NZ_AAXR00000000).
Table 2. blastx analysis of DNA sequences flanking plasposon insertion sites in IAA− mutants
Using iacH-specific primers (Table 1), four iacH-carrying fosmid clones were identified by PCR in a large-insert library of P. putida 1290 genomic DNA. Restriction analysis of these fosmids pCC1FOS9, −10, −11 and −12 revealed that their inserts overlapped over a region of c. 16 kb (not shown). Pseudomonas putida KT2440 (Nelson et al., 2002), which is not able to utilize IAA as a carbon end energy source (Leveau & Lindow, 2005), was transformed with fosmid pCC1FOS9, −10, −11 or −12. In each case, transformants were obtained with the ability to grow on M9 IAA agar plates. In liquid M9 IAA medium, growth of P. putida KT2440 carrying pCC1FOS9, −10, −11 or −12 occurred at a rate μmax of 0.51 h−1 and was characterized by concomitant and complete removal of IAA from the medium (Fig. 2). The same IAA+ phenotype was observed after transformation of P. putida KT2440 with fosmid pCC1FOS11Eco, a truncated derivative of pCC1FOS11 carrying the fragment that is shared between pCC1FOS9, −10, −11 and −12 (Fig. 3). No IAA+ transformants were obtained after transformation of KT2440 with only water or with negative control fosmid pCC1FOS15 (Table 1).
Shotgun DNA sequencing of pooled pCC1FOS9, −10, −11 and −12 revealed 39 ORFs on the largest assembled contig. ORFs 9 through 27 could be mapped to the DNA insert on pCC1FOS11Eco (Fig. 3). Transposon insertion analysis of pCC1FOS9-12 revealed that inactivation of ORFs 19, 21, 23, 24, 25 and 26 did not affect the IAA+ phenotype, whereas insertions in ORFs 9, 11, 12, 13, 14, 17 and 18 abolished or reduced the ability to grow on IAA (Fig. 3). This was independently confirmed by testing deletion derivatives of pCC1FOS11Eco for their ability to confer an IAA+ phenotype to P. putida KT2440 (Fig. 3). As only pCC1FOS11EcoAscPLUS scored positive, ORFs 19 through 27 seem dispensable for growth on IAA. From this, it was concluded that ORFs 9 through 18, collectively referred to as the iac locus, encode a pathway for indole 3-acetic acid catabolism. The DNA sequence of pCC1FOS11EcoAscPLUS carrying the 8994-bp iac locus has been submitted to the DDBJ/EMBL/Genbank database (accession number EU 360594).
Within the iac locus, ORF18 (iacI) seemed absolutely required for growth utilization of IAA by P. putida 1290: the IAA− phenotype of mutant 26E2 (Fig. 4a), carrying a plasposon insertion in the iacH gene (ORF17), was fully restored after complementation with pCC1FOS11EcoASCPLUS but not with pCC1FOS11EcoAsc (Fig. 3). Given the location and orientation of iacI relative to iacH, the IAA− phenotype of mutant 26E2 could be explained as a polar, negative effect on iacI expression as a result of the plasposon insertion in iacH.
It was found that mutant 26E2 was also affected in its ability to use indole-3-acetaldehyde as a carbon and energy source (Fig. 4b). Indole 3-acetamide could not be used as a growth substrate by either wild-type P. putida 1290 or mutant 26E2 (Fig. 4c). Radish root elongation assays revealed that of all the IAA− mutants, only 26E2 was significantly (P=0.00031) affected in its ability to abolish the deleterious effect of IAA on root elongation (Table 3). The simplest explanation for this is to assume that iacH and/or iacI are involved in the early steps of the IAA degradation pathway. Mutants blocked further downstream in the degradation pathway (e.g. 39D2, 40D5, 52F8 and 64F5) would be expected to retain activity towards IAA and thus the ability to rescue roots from IAA toxicity (Table 3).
Table 3. Ability of wild-type or mutant strains of Pseudomonas putida 1290 to abolish the effect of 1 mM exogenous IAA on radish root elongation
Mean (± SD) radish root length (mm)
86 ± 21
33 ± 22
89 ± 15
57 ± 17
73 ± 19
56 ± 17
77 ± 17
60 ± 24
68 ± 24
78 ± 22
64 ± 25
71 ± 17
78 ± 21
72 ± 23
58 ± 26
52 ± 20
80 ± 21
78 ± 17
Functional gene prediction of the iac locus and occurrence of iac gene clusters in other bacteria
The iac locus appeared to consist of three putative transcriptional units (Fig. 3): (1) an operon of ORFs 9–15 (provisionally named iacA, iacB, iacC, iacD, iacE, iacF and iacG), (2) an operon of ORFs 17 and 18 (iacH and iacI), and (3) a single gene (ORF16 or iacR) located upstream and transcribed divergently from ORF17. Table 4 lists the results of a blastp analysis for each of the iac gene products. Among these, four genes, i.e. iacA, iacC, iacD and iacH, with substantial coding similarity to enzymes that have been shown to act on aromatic N-heterocyclic compounds, were identified (Table 4). For two genes in the iac locus, no function could be predicted based on sequence similarity. blastp analysis of the iacB and iacI gene products revealed 21 and 22 homologs for each, respectively, in Genbank. Annotated as (conserved) hypothetical proteins, these 43 homologs were encoded on 21 finished or draft bacterial genome sequences (Fig. 5): three Gammaproteobacteria, five Alphaproteobacteria, ten Burkholderia species, and three species belonging to the class Actinobacteria, order Actinomycetales, suborder Corynebacterinae. On the near-complete genome sequence of uncultivated Burkholderia sp. SAR-1 (Venter et al. 2004), homologs of iacB and iacI were also identified (Fig. 5).
Table 4. ORF prediction for the iac locus on pCC1FOS11EcoAscPLUS
Conserved hypothetical protein; P. putida GB-1; 90%; 9e-84
Strikingly, in all of the 22 genomes mentioned above, homologs of iacB and/or iacI were locally clustered with genes showing high coding similarity to at least five of the other genes from the iac locus in P. putida 1290 (Fig. 5). Synteny of iac genes was differentially conserved among these genomes. Most noticeable was the fully preserved gene order in P. putida GB-1, compared to P. putida strain 1290. Overall, the gene syntenies found in Actinobacteria and Alpha- and Gammaproteobacteria appeared more similar to each other than to those in the Burkholderia strains.
Based on the discovery of iac-like genes in other bacteria, several strains listed in Fig. 5 were tested for their ability to grow on IAA. Indeed, P. putida GB-1, Burkholderia sp. 383, Sphingomonas wittichii RW1, and Rhodococcus sp. RHA1 all produced biomass on M9 minimal medium containing 5 mM IAA as the sole source of carbon and energy. The same observation was made for Marinomonas sp. MWYL1 (M. Wexler, pers. commun.). These results confirmed that IAA utilizers have representatives in the Alpha-, Beta- and Gammaproteobacteria as well as among the high G+C Gram-positive bacteria.
Involvement of cat and pca genes in the IAA degradation pathway
Based on sequence data from other plasposon mutants of P. putida 1290 (Table 2), it was assumed that the IAA degradation pathway features catechol as a central metabolite. In three mutants (30E1, 39B11 and 64F5), the plasposon was inserted into a catR-like gene, which, in other pseudomonads, is required for the catBCA-encoded conversion of catechol via cis,cis-muconic acid to 3-oxoadipate enol-lactone (Harwood & Parales, 1996). Failure to express the catBCA operon in mutants 30E1, 39B11, and 64F5 would explain the brown coloring of IAA plates by these strains (not shown) as the accumulation of catechol. Another mutant, 52F8, carried a plasposon in the catC gene, which probably blocked the expression of downstream, located catA for catechol 1,2-dioxygenase, also explaining its brown color on IAA plates (not shown) as an accumulation of catechol. The apparent involvement of cat genes during IAA degradation is in agreement with the previous observation that cell extracts of IAA-grown P. putida 1290 cells showed induced levels of catechol 1,2-dioxygenase activity (Leveau & Lindow, 2005).
Five of the IAA− mutants (39D2, 40D5, 44D7, 49F3 and 53D11) also lacked the ability to grow on benzoate as the sole source of carbon and energy. This suggests a merger of the degradation pathways of IAA and benzoate at 3-oxoadipate enol-lactone, which is converted to acetyl-CoA and succinyl-CoA by the products of genes pcaD, pcaIJ and pcaF. In several Pseudomonas species, expression of these genes is positively regulated by the pcaR gene product. In mutant 39D2, a homolog of this gene was knocked out. The IAA− phenotype of another mutant, 40D5A, may be explained as a polar effect of a plasposon in pcaB on the downstream expression of pcaD.
Identification of other genes essential for IAA degradation
Products of the pca pathway, i.e. acetyl-CoA and succinyl-CoA, are typically fed into the tricarboxylic acid (TCA) cycle. It was hypothesized that malate : quinone oxidoreductase, encoded by the mqo-1 gene that is knocked out in mutants 49F3 and 53D11, is required for efficient TCA cycling, as has been suggested for other bacteria (Molenaar et al., 2000). Furthermore, it was assumed that the pgk gene for phosphoglycerate kinase, which is disrupted in mutant 44D7, is necessary for gluconeogenesis during growth on IAA, based on the reported phenotypes of pgk mutants of P. putida strain A.3.12 (Aparicio et al., 1971). Analysis of IAA− mutants 45A5 and 47H11 revealed a gene with high coding similarity to the sensor/histidine kinase component (Mascher et al., 2006) of the two-component regulatory systems MhaSR in P. putida U (Arias-Barrau et al., 2005) and CbrAB in P. aeruginosa PAO (Nishijyo et al., 2001). Mutants of P. aeruginosa PAO1 that were inactivated in cbrAB were unable to grow on several N-substrates including arginine, ornithine, histidine, spermidine, putrescine and agmatine. It was suggested that CbrAB controls the expression of several catabolic pathways in response to changing intracellular C : N ratios (Nishijyo et al., 2001). Thus, their homolog in P. putida 1290, provisionally designated CbrA, is a protein possibly involved in IAA sensing.
Given the fact that for two of the iac genes (i.e. iacB and iacI) no function could be predicted from sequence homology, and that at least one of them (i.e. iacI) is essential for IAA degradation, any prediction of the IAA-degradative pathway as it operates in P. putida 1290 remains speculative and needs further biochemical investigation. In preliminary experiments, it was observed that colonies of E. coli carrying the iacA gene from P. putida 1290 turned a blue color on LB plates. This presumed production of indigo from indole has been linked to several genes related to iacA (Hart et al., 1990; Kim & Oriel, 1995; Solaiman & Somkuti, 1996; Drewlo et al., 2001; Alemayehu et al., 2004; Choi et al., 2004; Lim et al., 2005), and targets IacA as a possible candidate for the initial attack on IAA. The contribution of iacH to the mineralization of IAA by P. putida 1290 remains enigmatic. The predicted IacH product carries the hallmarks of an EC 3.5 amidase/amidohydrolase (Chebrou et al., 1996). However, together with its closest homologs in Fig. 5, it clearly clusters away from those amidases for which a function has been demonstrated. Among the latter, amino acid sequence is not necessarily a reliable predictor for substrate specificity, which can be altered by a single amino acid substitution (Okada et al., 1983). Furthermore, many of these amidases have dual specificities, cleaving amide as well as ester and nitrile bonds (Patricelli & Cravatt, 2000; Pollmann et al., 2003; Cilia et al., 2005). Future studies will reveal the substrate(s) for IacH and ascertain whether these substrates and their products feature in the IAA mineralization pathway. It is unlikely that iacH is involved in the conversion of indole-3-acetamide (IAM) to IAA: if it were, it would be difficult to explain why P. putida 1290 cannot grow at the expense of IAM (Fig. 4c).
From Fig. 5, it is concluded that the distribution of iac genes is remarkably discontinuous within the bacterial tree. While the finished genome sequences of 81 Alphaproteobacteria, 144 Gammaproteobacteria, and 47 Actinobacteria are available to date (http://www.genomesonline.org), clusters of highly homologous iac genes were found in only three (DSM12444, RW1, HaA2), two (ATCC 17978, MWYL1), and three (R, IFM 10152, RHA1) of those genomes, respectively. As for the 49 finished Betaproteobacterial genomes, iac homologs were present only in six Burkholderia species (AMMD, HI2424, 383, E264, LB400, ATCC 17616) and notably absent in 10 other Burkholderia species (four Burkholderia mallei, four Burkholderia pseudomallei, Burkholderia vietnamiensis G4, and Burkholderia cenocepacia AU1054).
Having linked the P. putida 1290 phenotype of IAA catabolism to the iac gene locus, a significant contribution has been made towards assigning a putative biological function to several iac gene homologs in the public databases for which no function had yet been described. Future research efforts will be targeted towards establishing the functionality of these iac homologs and determining the significance of the observed differences in iac gene synteny in the different bacterial species (Fig. 5). Also, the link between IAA and iac allowed identification of several of the iac-bearing bacteria as previously unrecognized IAA degraders. Interestingly, these strains originated from environments known to be natural sources of IAA, such as the plant rhizosphere (Marinomonas sp. MWYL1) and soil (Burkholderia sp. 383, Rhodococcus sp. RHA1), but also from less obvious environments such as river water (S. wittichii RW1). Two of the bacterial species that were identified as carrying iac genes are clinical isolates, i.e. N. farcinica IFM 10152 and Acinetobacter baumannii ATCC 17978. Strains of A. baumannii are commonly isolated from hospitalized patients, where they are known to be able to infect, among other things, the urinary tract. Interestingly, urine is also a known source of IAA (Kogl et al., 1934), which allows one to start speculating on the role of iac genes in the infection process of A. baumannii isolates.
Intriguingly, B. japonicum USDA 110, the only confirmed IAA degrader (Jarabo-Lorenzo et al., 1998) for which a genome is available (Kaneko et al., 2002), conspicuously lacks iac genes, except for a highly homologous iacA homolog. The gene product of the latter, NrgC (accession number AAG 61032), is mentioned in several studies as being regulated by nitrogen fixation regulator NifA (Nienaber et al., 2000) and expressed in symbiosomes (Hoa et al., 2004). However, this gene has not been linked previously to IAA, which, in the context of the authors' findings, seems to be an exciting avenue for further study, given (1) the stimulatory role of IAA in nodule formation by B. japonicum (Boiero et al., 2007) and (2) the location of nrgC on a nod/nif symbiosis island (Gottfert et al., 2001) with close proximity to genes encoding a nodulation protein (accession number AAG61031) and a nodulation formation efficiency protein (accession number AAG61037). The apparent absence of additional iac homologs from the B. japonicum USDA 110 genome furthermore suggests that this strain possesses genes other than iac homologs for IAA degradation.
The authors thank Étienne Yergeau for his help with the statistical analysis of the radish root elongation data, Dr Robert van der Geize for sending Rhodococcus sp. RHA1, Dr Peter Vandamme for Burkholderia sp. strain 383, Dr Jan Tommassen for P. putida GB-1 and Dr Margaret Wexler for testing Marinomonas MWYL1 for growth on IAA. The authors are also indebted to Dr M. Wexler, Dr Andy Johnston, and Dr Hanspeter Kohler for their comments on drafts of the manuscript. This is publication 4231 of the Netherlands Institute of Ecology (NIOO-KNAW). This research was supported by The Netherlands Organization for Scientific Research (NWO) Innovational Research Incentive Grant VENI 863.03.005.