Bacterial strains and media
A collection of 185 genome-sequenced bacterial isolates, described previously14, was utilized to assemble the synthetic community used in this work. These isolates were obtained from surface-sterilized Brassicaceae roots, primarily Arabidopsis thaliana, grown in two soils from North Carolina, USA35. This isolate collection includes strains V. paradoxus CL14, Arthrobacter CL28, Acinetobacter CL69 and Acinetobacter CL71, which are also used in this work in individual strain contexts. V. paradoxus CL14 ΔHS33, which has a clean deletion of genes with gene ID 2643613677 through 2643613653 was constructed previously14 and used here. Additional strains were obtained from the American Type Culture Collection (ATCC): E. soli LF7 (ATCC BAA-2102), R. pomeroyi (ATCC 700808) and B. japonicum (ATCC 10324). P. phytofirmans PsJN (DSMZ 17436) was obtained from the DSMZ-German Collection of Microorganisms and Cell Cultures. P. putida strain 1290 was generously provided by Johan Leveau (University of California Davis). Pseudomonas strain Root 562 was generously provided by Paul Schulze-Lefert (Max-Planck-Gesellschaft). All bacteria, with exceptions noted below, were routinely grown on LB agar plates (10 g l−1 tryptone, 5 g l−1 yeast extract, 10 g l−1 NaCl, 1.5% (w/v) agar) and in 2xYT liquid medium (16 g l−1 tryptone, 10 g l−1 yeast extract, 5 g l−1 NaCl) at 28 °C. The 175-member (185-member minus 10 Variovorax strains) synthetic community (SC185-10V) was grown on KB medium as was done previously to culture this synthetic community14. B. japonicum (ATCC 10324) was routinely grown on liquid and solidified YM medium (1 g l−1 yeast extract, 10 g l−1 mannitol, 0.5 g l−1 dipotassium phosphate, 0.2 g l−1 magnesium sulfate, 0.1 g l−1 NaCl, 1 g l−1 CaCO4, pH 6.8, solidified with 1.5% agar as necessary) at 28 °C. R. pomeroyi (ATCC 700808) was routinely grown on liquid and solidified LB medium supplemented with 2% sea salt (Millipore Sigma S9883) and solidified with 1.5% (w/v) agar as necessary. M9 base medium was formulated using 1x M9 minimal salts medium (Sigma M6030) supplemented with 2 mM MgSO4, 0.1 mM CaCl2 and 10 µM FeSO4. A carbon source or sources were added to this M9 base medium to support bacterial growth. Unique strains constructed in this study are available upon request.
Bacterial 16S rRNA sequencing
Bacterial colonization of Arabidopsis roots was assessed using a method similar to the previous study14. Roots from 8–10 plants were collected into sterilized 2 ml tubes containing three 4 mm glass beads and root fresh weight in each tube was obtained. Five such samples were collected for each bacterial treatment. The roots were washed three times with sterile distilled water and stored at −80 °C until further processing. The roots were then lyophilised for 48 h using a Labconco freeze-dry system and pulverized using an MPBio tissue homogenizer. DNA was extracted from the root samples and bacterial cell pellets saved from the bacteria for input into the experiment using the DNeasy PowerSoil HTP 96 kit (Qiagen) according to manufacturer instructions. The V3-V4 region of the bacterial 16S rRNA gene was amplified and sequenced as previously described14.
16S amplicon sequence data processing
The 16S sequencing data from synthetic community experiments were processed as previously described14. Briefly, usable read output from MT-Toolbox36 (reads with 100% primer sequences that successfully merged with their pair) were filtered for quality with Sickle37 by not allowing any window with Q score under 20. The resulting sequences were globally aligned to a 16S rRNA gene sequence reference dataset from genome assemblies of the synthetic community members. For strains that do not have an intact 16S rRNA sequence in their assembly, Sanger sequencing was used to obtain the 16S rRNA gene sequence of the strains for inclusion in the reference dataset. The reference dataset also included sequences from Arabidopsis organellar sequences and known bacterial contaminants. Sequence alignment was performed with USEARCH v.7.109038 using the optional usearch_global at a 98% identity threshold. On average, 85% of read sequences matched an expected isolate. The 185 isolates of our 185-member synthetic community could not all be distinguished from one another on the basis of the V3-V4 sequence. They were thus classified into 97 unique sequences encompassing a set of identical (clustered at 100%) V3-V4 sequences coming from a single or multiple isolate strains. An isolate abundance table was created from the sequence mapping results.
We estimated 16S rRNA absolute abundance using a plasmid spike-in method39. Synthetic DNA was spiked at known quantities into samples before DNA extraction and the ratio of added to recovered synthetic DNA served as a conversion factor by which the total number of 16S rRNA molecules in a given sample was estimated. We designed a plasmid which included 16S V3-V4 primer binding sequences flanking a randomly generated DNA sequence matching the most frequent length and Guanine + Cytosine (GC) content of amplicons generated using the same primer sequences from wild soil. These sequences were synthesized by Geneart (Invitrogen) and supplied cloned in plasmid pMA-T. The plasmid was transformed into E. coli and isolated using a miniprep spin kit (Qiagen). Specific volumes of this isolated plasmid were then added to individual samples before DNA extraction to spike-in approximately 20% of the predicted 16S copies occurring within the sample. We performed colony-forming units (c.f.u.) counting using similarly treated plant samples (that is, growth on SynCom-inoculated agar plates) to obtain an estimate of the 16S copy number per mg fresh weight of plant roots. We plated serial dilutions of plant root samples ground in MgCl2 on LB to perform c.f.u. counts. The c.f.u. count multiplied by a given sample’s fresh weight were used to calculate sample-specific predicted 16S copy numbers.
Plant growth conditions and root growth inhibition assay
A. thaliana ecotype Col-0 seeds were sterilized in 70% household bleach, 0.2% Tween-20 for 10 min with vigorous agitation and then rinsed 10 times with sterile distilled water. Seeds were stratified at 4 °C in sterile distilled water for 1–2 d. Plants were germinated for 7 d on 0.5x MS agar medium (2.22 g l−1 PhytoTech Labs M-404: Murashige & Skoog modified basal medium with Gamborg vitamins, 0.5 g l−1 MES hydrate, pH adjusted to 5.7, solidified with 1% (w/v) agar) supplemented with 0.5% (w/v) sucrose in vertical 12 ×12 cm square plates under long-day conditions (21 °C/18 °C, 16 h light/8 h dark, day/night cycle). Then 8 to 10 plants were aseptically transferred to 12 ×12 cm plates containing 0.5x MS agar medium without sucrose where the medium surface was spread with the bacterial inoculum. For assays with IAA addition, 100 nM IAA was added to the medium before pouring the plates. The plant root tip location was marked on plates after transfer to record the initial root tip position. The plates containing the plants and bacteria were incubated vertically under short-day conditions (22 °C/18 °C, 9 h light/15 h dark, day/night cycle) for an additional 11 d. Plates were imaged on a document scanner and primary root elongation was determined using imageJ to quantify the change in root tip position from the initial to the final position.
Bacterial inoculation of plants
Individual bacterial strains were grown on agar plates of the media types specified above at 28 °C. Before plant inoculation, a single colony was picked into the appropriate liquid medium and grown at 28 °C to late exponential or early stationary phase. To remove the medium from the bacteria before inoculation, strains were washed three times in sterile 10 mM MgCl2. The optical density at 600 nm (OD600) was measured for each washed strain and normalized to OD600 of 0.01 in 10 mM MgCl2. For plant experiments with mono-association of an individual strain, 100 µl of OD600 = 0.01 washed bacteria was spread on the 12 ×12 cm plate before plant transfer. For experiments in duo-association with Arthrobacter CL28, 100 µl of OD600 = 0.01 washed Arthrobacter CL28 was spread along with 100 µl of OD600 = 0.01 of the second strain.
The 175-member synthetic community (SC185-10V) was prepared as described for the 185-member synthetic community used previously14 by leaving out the 10 isolates from the genus Variovorax. Briefly, 7 d before plant transfer, strains were inoculated individually into 600 µl KB medium in a 96-well plate and grown at 28 °C for 5 d. At 2 d before plant transfer, 20 ul from these 5-day-old cultures were transferred to 380 ul fresh KB medium in a new set of 96-well plates and both sets of plates were returned to the incubator for 2 d. This resulted in two cultures of each strain, one 7 d old and the other 2 d old, which were combined. The OD600 of the strains in each well was measured and the strains were combined while normalizing the OD600 of each strain in the pool. This pool was washed twice with 10 mM MgCl2 and diluted to OD600 = 0.2. For experiments with the SC185-10V SynCom, 100 µl of this OD600 = 0.2 washed pool was spread on 12 ×12 cm plates. For treatments where an additional strain was added to the SC185-10V SynCom, the individual strain was washed as described above, diluted to OD600 = 0.0034 in 10 mM MgCl2, and 100 µl of this dilution was spread on the plates with the SC185-10V SynCom. This addition of the individual strain corresponded to an OD600 three times that of a single strain in the SC185-10V SynCom (0.0034 = (0.2/175) × 3). For the addition of the 10 Variovorax strains to the SC185-10V SynCom experiment, the 10 Variovorax strains were grown individually in 2xYT medium from colonies grown on plates. The OD600 values of the 10 cultures were measured and the 10 strains were pooled while normalizing the OD600 of each strain to the same value. This mixture of the 10 Variovorax strains was then treated as the individual strains for washing and addition of 100 µl of OD600 = 0.0034 to the SC185-10V SynCom on plates.
Construction of vectors with Variovorax CL14 iad gene inserts
Portions of the V. paradoxus CL14 IAA degradation locus were subcloned into broad host range vector pBBR1MCS-232. Primers JMC579 through JMC604 (Supplementary Table 8) were used to amplify 3–5 kb segments of the locus by PCR using Q5 DNA polymerase (New England Biolabs). These primers were designed to amplify sections beginning and ending at gene start codons and with appropriate overlapping sequences for Gibson assembly either into the pBBR1MCS-2 backbone or to the adjacent section to make larger vector inserts, as appropriate. The pBBR1MCS-2 vector backbone was prepared for Gibson assembly by amplifying the vector by PCR using primers JMC577 and JMC578 (Supplementary Table 8) and subsequently treating with DpnI to remove circular vector template. PCR fragments were cleaned up as necessary using the QIAquick PCR purification kit (Qiagen). Appropriate fragments were mixed to construct the vectors by Gibson assembly using HiFi DNA Assembly Mastermix (New England Biolabs) according to manufacturer instructions. Gibson assembly products were transformed into NEB 10beta chemically competent E. coli (New England Biolabs) and selected on LB plates supplemented with 50 µg ml−1 kanamycin. Vectors were miniprepped using either the ZR plasmid miniprep classic kit or Zymo BAC DNA miniprep kit (Zymo Research) and confirmed via restriction mapping with PstI-HF (New England Biolabs) and Sanger sequencing (Genewiz).
To construct vectors that are derivatives of pBBR1::70–66, the Q5 site-directed mutagenesis kit (New England Biolabs) was used for gene deletion. Briefly, vector pBBR1::70–66 was used as a PCR template and portions of this vector were amplified by PCR using primers JMC641 through JMC650 (Supplementary Table 8) and Q5 DNA polymerase (New England Biolabs). PCR products were cleaned up and circularized using KLD Mastermix (New England Biolabs). The product was transformed into NEB 10beta chemically competent E. coli (New England Biolabs) and selected on LB plates supplemented with 50 µg ml−1 kanamycin. Vectors were miniprepped and Sanger sequenced as described above to confirm the construction of the correct vectors.
Conjugation of vectors to V. paradoxus CL14 ΔHS33
Vectors were conjugated into V. paradoxus CL14 ΔHS33 using tri-parental mating. The helper E. coli strain carrying plasmid pRK201340 and donor NEB 10beta E. coli strains containing the pBBR1MCS-2-based vectors with Variovorax IAA degradation locus gene inserts were cultured in LB media containing 50 µg ml−1 kanamycin at 37 °C. V. paradoxus CL14 ΔHS33 was grown in 2xYT medium containing 100 µg ml−1 ampicillin at 28 °C. V. paradoxus CL14 wild type and derivative strains such as ΔHS33 are naturally resistant to ampicillin and this ampicillin selection allows for recovery of only Variovorax from the conjugation reaction. To prepare for conjugation, all bacteria were pelleted by centrifugation at 5,000 × g for 5 min and washed 3 times in 2xYT medium without antibiotics. For each conjugation reaction, equal volumes (100–300 µl) of each of the three washed bacteria: recipient V. paradoxus CL14 ΔHS33, donor NEB 10beta E. coli containing a pBBR1MCS-2-based vector, and helper E. coli pRK2013 were mixed. Control conjugation mixtures of each pair of strains and individual strains alone were performed in parallel to ensure successful selection of exconjugants only from mixtures of all three strains together. Conjugation mixtures were pelleted by centrifugation at 5,000 × g for 5 min, resuspended in 50 µl 2xYT media, transferred to LB media plates without antibiotics and allowed to dry in a laminar flow hood. These conjugation plates were incubated overnight at 28 °C. After 18–24 h, exconjugants were selected by streaking from the pooled conjugation mixtures on the LB plate without antibiotics to LB plates containing 50 µg ml−1 kanamycin and 100 µg ml−1 ampicillin. This selects for only V. paradoxus CL14 ΔHS33 (ampicillin resistant) containing the pBBR1MCS-2-based vector (kanamycin resistant). Individual colonies were picked into and subsequently cultured in 2xYT medium containing 50 µg ml−1 kanamycin and 100 µg ml−1 ampicillin at 28 °C.
Construction of V. paradoxus CL14 gene deletions
Unmarked gene deletions in V. paradoxus CL14 were constructed as described previously14 using the suicide vector backbone pMo130 originally developed for gene knockouts in Burkholderia spp.41. Primers JMC203 and JMC204 (Supplementary Table 8) were used to amplify the pMO130 vector backbone by PCR. This product was subsequently treated with DpnI (New England Biolabs) to digest circular template DNA. Primers JMC605 through JMC612 and JMC671 through JMC677 (Supplementary Table 8) were used to amplify flanking regions for the gene deletion targets from V. paradoxus CL14 genomic DNA. All PCR was performed using Q5 DNA polymerase (New England Biolabs) and products were cleaned up, as appropriate, with the QIAquick PCR purification kit (Qiagen). These PCR products were assembled into suicide vectors using HiFi Gibson Assembly Mastermix (New England Biolabs), transformed into chemically competent NEB 5alpha E. coli (New England Biolabs), and selected on LB plates with 50 µg ml−1 kanamycin. Vectors were miniprepped using the ZR plasmid miniprep classic kit (Zymo Research) and confirmed by Sanger sequencing (Genewiz). Confirmed vectors were transformed into the chemically competent bi-parental mating strain E. coli WM3064. Transformants were selected at 37 °C on LB media supplemented with 50 µg ml−1 kanamycin and 0.3 mM diaminopimelic acid (DAP), and single colonies picked into LB medium also with 50 µg ml−1 kanamycin and 0.3 mM DAP.
Bi-parental mating was performed by growing E. coli WM3064 containing the appropriate suicide vector as described above, and V. paradoxus CL14 was grown in 2xYT medium containing 100 µg ml−1 ampicillin at 28 °C. Both E. coli and Variovorax were washed separately three times using 2xYT medium, then mixed in a 1:1 ratio and pelleted. All centrifugation steps were performed at 5,000 × g for 5 min. The pelleted conjugation mixtures were resuspended in 1/10 the volume of 2xYT, plated on LB agar with 0.3 mM DAP and grown at 28 °C overnight. Exconjugants from these plates were streaked out and grown on LB agar with 100 µg ml−1 ampicillin, 50 µg ml−1 kanamycin, and no DAP at 28 °C. These strains were purified by streaking and growing on plates of the same medium once more. These strains with suicide vector integration were then grown once in liquid LB containing 100 µg ml−1 ampicillin and 1 mM isopropyl 1-thio-d-galactopyranoside (IPTG) at 28 °C and then streaked on plates containing media with 10 g l−1 tryptone, 5 g l−1 yeast extract, 100 g l−1 sucrose, 1.5% agar, 100 µg ml−1 ampicillin and 1 mM IPTG. Colonies from these plates were picked and grown in the same liquid media. These strains were then assessed for gene deletion by PCR using primers JMC657 through JMC660 and JMC697 through JMC699 (Supplementary Table 8). The Quick-DNA miniprep kit (Zymo Research) was used to isolate all genomic DNA for PCR screening. To purify the knockout strains, they were streaked and grown out three times on LB plates containing 100 µg ml−1 ampicillin before a final PCR confirmation. To check the purity of the final strains, PCR was performed with one primer outside the deletion region and one inside the deleted gene to ensure no product is produced for the knockout strain. The sequences for the primers used for this PCR reaction (JMC691, JMC717, JMC718, JMC693 and JMC694) can be found in Supplementary Table 8.
Measurement of bacterial growth and IAA degradation
Individual strains were grown in 5 ml cultures in various media types supplemented with IAA at 28 °C and 250 r.p.m. To screen the V. paradoxus CL14 ΔHS33 pBBR1 vector complemented mutants, 2xYT medium supplemented with 0.1 mg ml−1 IAA was used. For comparison of other V. paradoxus CL14 mutants, M9 medium with 15 mM succinate and 0.1 mg ml−1 (0.57 mM) IAA was used. For comparison of IAA-degrading strains from diverse genera, M9 medium with 0.1% (w/v) casamino acids (Bacto) and 0.1 mg ml−1 IAA was used. For R. pomeroyi, 2% (w/v) sea salts were added to this M9 medium with casamino acids and IAA. The pBBR1 vector library in E. coli NEB 10beta was screened in LB medium supplemented with 0.04 mg ml−1 IAA and grown at 37 °C and 250 r.p.m. For all media types, IAA was solubilized in 100% ethanol at 20 mg ml−1 and diluted to 0.1 mg ml−1 in the media, resulting in 0.5% (v/v) ethanol in the media.
To measure growth, a 200 µl sample was taken from the growing cultures and OD600 was determined on an Infinite M200 Pro plate reader (Tecan). Subsequently, cells were pelleted by centrifugation at 4,200 × g for 15 min and 50 µl of supernatant was transferred to a new 96-well plate and frozen at −80 °C until further analysis. IAA degradation was determined by thawing the plates containing 50 µl aliquots of culture supernatant and combining this with 100 µl of Salkowski reagent (10 mM ferric chloride and 35% perchloric acid)42. This was performed alongside mixing 50 µl of IAA standards with 100 µl of Salkowski reagent in the same 96-well plate format. Colour development was allowed to proceed for 1 h and absorbance was read at 530 nm on the Infinite M200 Pro plate reader (Tecan). The absorbances measured were converted to IAA concentration on the basis of the absorbances measured for the IAA standards.
Liquid Chromatography Dual Mass Spectroscopy (LC–MS/MS) metabolomics on Variovorax IAA degradation
V. paradoxus CL14 was grown in 5 ml cultures of M9 minimal medium supplemented with either 0.1 mg ml−1 IAA, 0.1 mg ml−1 13C6-IAA (with the 6 carbons of the benzene ring of the indole labelled, Cambridge Isotope Laboratories CLM-1896-PK), and/or 15 mM succinate. Cultures and parallel media controls were incubated at 28 °C with shaking at 250 r.p.m. Cultures and media controls were centrifuged (4,200 × g for 15 min at 4 °C) to pellet cells; supernatants were transferred to new tubes and both pellets (intracellular fraction) and supernatants (extracellular fraction) were stored frozen at −80 °C until extraction. All subsequent work was performed over dry ice or in chilled cold blocks. Frozen pellets from the intracellular fraction were thawed for 3 h at 4 °C, then 800 µl of cold LCMS-grade water was added to the pellets with repeated pipetting to break up the pellet until visually homogeneous. Samples were then quickly returned to −80 °C to freeze the suspension. Frozen pellet suspensions and extracellular solutions were lyophilised until dry. The cells from the dried pellet suspensions were lysed and homogenized with a bead mill (BioSpec Mini-Beadbeater-96) using one sterile 3.2 mm steel ball in each tube for 3 rounds of 5 s each with 10 s breaks in between to reduce heat production. Dried extracellular samples were concentrated by resuspension in 100 µl LCMS- grade methanol, vortexed 3 times for 10 s each, water bath sonicated for 20 min, incubated at 4 °C overnight, centrifuged (1,000 × g, 4 °C, 5 min), and the methanol supernatant was dried using a speed vacuum concentrator. On the day of LC–MS/MS analysis, homogenized dry material was suspended in LCMS-grade methanol with internal standard mix (100 µM U-13C/15N-labelled amino acids, SIGMA 767964). Intracellular samples were suspended at 11.1 µl mg−1 of original sample cell pellet wet weight; extracellular samples were suspended at 38.9 µl mg−1 of corresponding cell pellet wet weight from the culture. The solutions were vortexed 3 times for 10 s each, bath sonicated in ice water for 10 min, chilled at −20 °C for 10 min, then centrifuged (10,000 × g, 5 min, 10 °C) to pellet insoluble material. Supernatants containing the methanol extracts were filtered through 0.22 µm PVDF microcentrifuge filtration tubes (10,000 × g, 5 min, 10 °C); filtrates were transferred to glass vials and immediately capped. Filtrates were then analysed by LC–MS/MS using an Agilent 1290 UHPLC system connected to a Thermo Q Exactive Hybrid Quadrupole-Orbitrap mass spectrometer equipped with a heated electrospray ionization (HESI-II) source probe. Extracts were chromatographically separated on a ZORBAX RRHD Eclipse Plus C18, 95 Å, 2.1 × 50 mm, 1.8 µm column (Agilent) for non-polar metabolomics. Separation, ionization, fragmentation and data acquisition parameters are specified in Supplementary Table 7. Briefly, metabolites were separated by gradient elution followed by MS1 and data-dependent (top 2 most abundant MS1 ions not previously fragmented in last 7 s) MS2 collection; targeted data analysis was performed by comparison of sample peaks to a library of analytical standards analysed under the same conditions or by searching the raw data files for predicted m/z values based on structural information of compounds of interest. Three parameters were compared: matching m/z, retention time and fragmentation spectra using Metabolite Atlas (https://github.com/biorack/metatlas)43,44. Identification and standard reference comparison details are provided in Supplementary Table 6. Raw and processed data are available for download at the JGI Joint Genome Portal under ID 1340427. Statistical comparisons were performed using R version 3.6.2, using package agricolae 1.3–5 and stats 3.6.245; boxplots were generated with base R graphics using the boxplot function.
Phylogenomic analysis
To guide the delineation of the IAA degradation operons across the bacterial tree of life, we constructed two Hidden Markov Model (HMM) profiles of the genes iacC and iacD by subsetting all homologous genes from the previously validated operons (Extended Data Fig. 4). In parallel, we downloaded the assembly files for all available complete genomes deposited in the NCBI RefSeq 202 repository46. For the 220,000 assembly files downloaded, we performed open reading frame (ORF) prediction using prodigal. We then used the two HMM profiles described above to query the predicted ORFs. Utilizing ad hoc scripts, we constructed a table of HMM hits along the genomes scanned and subset genomic loci where both iacC and iacD genes appeared adjacent to one another. The logic of using the iacC and iacD genes as anchor genes for our search is that the adjacent physical location of both iacC and iacD homologues is a conserved feature across all previously experimentally validated IAA-degrading operons (Extended Data Fig. 4). Next, for each region containing the adjacent iacC and iacD homologue genes, we extracted the gene neighbourhood adjacent to the anchor hit by extracting the amino acid sequence of ORFs +10 kb and −10kb with respect to the anchor hit. Using hmmscan from the Hmmer v3.1.b2 suite47, we performed HMM profiling in all ORFs extracted via our neighbourhood delineation against the COG database version 2003. Finally, we used the COG profiles across the neighbourhoods to create a matrix describing the prevalence of COGs across the regions (candidate regions) with the adjacent iacC and iacD homologue genes.
For each genome containing at least one candidate region, we performed taxonomic classification using the GTDB database48. Due to the size of our estimated genomic matrix and to reduce potential biases due to over-representation of certain lineages within RefSeq, we performed principal coordinate analysis (PCoA) using a reduced matrix containing one representative candidate region per species. Species labelling was obtained from the GTDB taxonomic classification described above. PCoA was performed using the oh.pco function from the ohchibi package49, taking as input a binary version of the gene matrix described above. We classified candidate reads into the two types of IAA-degrading operon (iac-like and iad-like), utilizing a majority count-based approach using marker COGs conserved between the previously experimentally validated IAA-degrading operons (Extended Data Fig. 4). Specifically, for each potential operon, we determined the prevalence of COGs that a priori (Extended Data Fig. 4) showed differential prevalence across the two degrading operons (for example, iacA, iacB and iacI are markers of the iac operon, while iorB/iadB and iotA/iadA are exclusive markers of the iad-like operon). Hybrid gene clusters were defined as operons that exhibited the hallmark COGs of both operons.
In parallel, we performed phylogenetic inference over all the genomes belonging to genera with at least one representative strain harbouring any of the two types of IAA-degrading operon. This phylogenetic tree was constructed using a super-matrix-based approach as previously described35. Finally, for each genus with at least one assembly harbouring a positive IAA-degrading operon, we estimated the prevalence of the trait across the genus by dividing the total number of isolates with detectable IAA degradation locus by the total number of isolates belonging to that genus in the dataset. In addition, to see the phylogenetic evenness of the distribution of the IAA degradation trait across each genus, we calculated the phylogenetic ratio by calculating the ratio between the average phylogenetic distance (computed via the cophenetic.phylo function from the ape R package50) of isolates with a detectable IAA degradation locus and the total average phylogenetic distance of all isolates within that genus. We constructed the MarR phylogeny using the MarR sequences from candidate regions with 100% markers of one of the two types of IAA-degrading operon. Amino acid sequences of the MarR homologues were aligned using MAFFT51 and phylogenetic inference was performed using FastTree 252.
RNA-seq on Variovorax strains
V. paradoxus CL14 was grown in 5 ml cultures of M9 minimal medium supplemented with 15 mM succinate and 0.5% (v/v) ethanol alone or containing IAA. IAA was at a final concentration of 0.1 mg ml−1 in the medium to which it was added. Cultures were prepared at a starting OD600 of 0.02 and incubated at 28 °C, shaking at 250 r.p.m. Cells from all samples were collected for RNA-seq at 18 h to ensure IAA was still present in the cultures of strains that degraded IAA most rapidly. Cells were pelleted by centrifuging the culture at 4,200 × g for 15 min and removing the supernatant. Cell pellets were frozen at −80 °C before RNA extraction. To extract RNA, cells were lysed in TRIzol reagent (Invitrogen) according to manufacturer instructions for lysis and phase separation. After these steps, RNA was purified from the aqueous phase using the RNeasy mini kit (Qiagen) including the optional on-column DNase digestion with RNase-free DNase set (Qiagen). Total RNA was quantified using the Qubit 2.0 fluorometer (Invitrogen) and RNA-seq libraries were prepared using the Universal Prokaryotic RNA-Seq Prokaryotic AnyDeplete kit (Tecan) according to manufacturer instructions. The resulting libraries were pooled and sequenced on the Illumina HiSeq4000 to generate 50 bp single-end reads.
RNA-seq data analysis
The V. paradoxus CL14 RNA-seq sequence data were analysed as described previously14. Briefly, the raw reads were mapped to the V. paradoxus CL14 genome (fasta file available at https://github.com/isaisg/variovoraxRGI/blob/master/rawdata/2643221508.fna) using bowtie253 with the ‘very sensitive’ flag. Hits to each individual coding sequence were counted and annotated using the function featureCounts from the R package Rsubread54, inputting the V. paradoxus CL14 gff file (available at https://github.com/isaisg/variovoraxRGI/blob/master/rawdata/2643221508.gff) and using the default parameters with the flag allowMultiOverlap = FALSE. Finally, DESeq255 was used to estimate Differentially Expressed Genes (DEGs) between treatments, with the corresponding fold-change estimates and False Discovery Rate (FDR) adjusted P values. For visualization purposes, we performed z-score standardization of each gene across samples and we visualized this standardized expression values utilizing a heat map constructed using ggplot256. These data can be found in Supplementary Table 9.
MarR protein expression and purification
The coding sequence for each gene can be found in Supplementary Table 10. MarR expression plasmids were synthesized as codon-optimized genes for E. coli expression by BioBasic in the pLIC-His N-term vector (pMCSG7) and transformed into E. coli BL21 (DE3) Gold cells for expression. Cells were grown in the presence of ampicillin in LB medium with shaking at 225 r.p.m. at 37 °C to an OD600 of 0.5, at which point the temperature was reduced to 18 °C. At an OD600 of 0.8, protein expression was induced by the addition of 0.1 mM IPTG and incubation continued overnight. Cells were collected by centrifugation at 4,500 × g for 20 min at 4 °C in a Sorvall (model RC-3B) swinging bucket centrifuge. Cell pellets were resuspended in buffer A (20 mM potassium phosphate, pH 7.4, 50 mM imidazole, 500 mM NaCl), DNase, lysozyme and a Roche Complete EDTA-free protease inhibitor tablet. Resuspended cells were sonicated and clarified via centrifugation at 17,000 × g for 60 min in a Sorvall (model RC-5B) swinging bucket centrifuge. The lysate was applied to a nickel-nitrilotriacetic acid HP column (GE Healthcare) on an Aktaxpress Fast Performance Liquid Chromatography (FPLC) system (Amersham Bioscience) and washed with buffer A. Protein was eluted with buffer B (20 mM potassium phosphate, pH 7.4, 500 mM imidazole, 500 mM NaCl). Fractions containing the protein of interest were combined and passed over a HiLoadTM 16/60 SuperdexTM 200 gel filtration column. Proteins were eluted in S200 buffer (20 mM HEPES, pH 7.4, 300 mM NaCl). Fractions were combined and concentrated for long-term storage at −80 °C.
MarR mutant proteins were created by site-directed mutagenesis using primers from Integrated DNA Technologies. The mutant plasmids were sequenced to confirm the mutations. The mutants were produced and purified using E. coli BL21 (DE3) Gold as described above.
Ligand binding studies by isothermal titration calorimetry (ITC)
All ITC measurements were performed at 25 °C using an Auto-ITC200 microcalorimeter (MicroCal/GE Healthcare). The buffer employed was 20 mM HEPES, pH 7.4, 50 mM NaCl and 0.5% dimethly sulfoxide (DMSO) for protein/ligand binding and 20 mM HEPES, pH 7.4 and 300 mM NaCl for DNA/protein binding experiments. For ligand binding experiments, the calorimetry cell (volume 200 ml) was loaded with MarR wild-type, mutant or homologue protein at a concentration of 50 μM. The syringe was loaded with a ligand concentration of 0.5 or 2 mM. For DNA binding experiments, wild-type MarR_73 did not bind any of the DNA oligos examined; however, we hypothesized that this arose from the ability of this native receptor to remain bound to ligands retained from its recombinant expression in E. coli. Thus, we employed the MarR_73 S28A protein with reduced ligand binding capacity. Here, the calorimetry cell was loaded with duplex oligo at a concentration of 25 μM and the syringe was loaded with MarR S28A mutant protein, which was necessary to prevent ligand binding during expression and purification, at a concentration of 0.5 mM. A typical injection protocol included a single 0.2 μl first injection followed by 20 1.5 μl injections of the syringe sample into the calorimetry cell. The spacing between injections was kept at 180 s and the reference power at 8 μcal s−1. The data were analysed using Origin for ITC version 7.0 software supplied by the manufacturer and fit well to a one-site binding model. Two independent ITC measurements were performed for each condition. A non-integer N value (for example, 0.73 in Fig. 2a) indicates that some protein monomers may not be in an active conformation, and thus do not bind ligand. Additionally, small measurement errors in assessing the protein or ligand concentrations may also contribute to non-integer N values in ITC. To confirm that 300 mM NaCl did not negatively impact DNA binding, MarR_73 S28A was examined by ITC in 150 mM NaCl. In this condition, the KD for the 22 bp duplex was 0.428 ± 0.002 μM (N = 1.75 ± 0.014), while the KD for 24 bp duplex was 0.151 ± 0.025 μM (N = 2.51 ± 0.26).
Protein crystallography
V. paradoxus MarR_73 was crystallized using the sitting drop vapour diffusion method at 20 °C in conditions outlined in Supplementary Table 4. Crystallization drops were set up using the Oryx4 protein crystallization robot (Douglas Instruments) and contained 0.15 μl protein and 0.15 μl well solution. For all V. paradoxus MarR_73 wild-type conditions, ligands were added at 10-fold molar excess before crystallization trials and crystals appeared within 2–5 d. V. paradoxus MarR_73 with the S28A and R46A mutations was crystallized in similar conditions as the wild-type protein. Similarly, P. putida MarR_iacR, B. japonicum MarR_Bj1, A. baumannii MarR_Ab and E. soli MarR_Es were crystallized using vapour diffusion methods in sitting drop trays at 20 °C and crystals appeared within 3–5 d. All crystallization conditions are outlined in Supplementary Table 4. Crystal specimens were cryoprotected with the well solution supplemented with glycerol to 20% (v/v) (Supplementary Table 4). X-ray diffraction data were collected at the Advanced Photon Source beamline 23-ID-D (Supplementary Table 3). Diffraction images were reduced using either XDS or Denzo and scaled with either Aimless or Scalepack57,58,59. The V. paradoxus MarR_73 structure in complex with IAA was determined by molecular replacement using the structure of 3CDH as a search model in Phaser60. All subsequent structures of V. paradoxus MarR_73 were determined using the V. paradoxus MarR_73 IAA complex structure (PDB: 7KFO) as a search model. The P. putida MarR_iacR and B. japonicum MarR_Bj1 structures were determined by molecular replacement using the structure of 3CJN as the search model. P. putida MarR_iacR (PDB: 7KUA) was subsequently used as the search model for molecular replacement to solve A. baumannii MarR_Ab and E. soli MarR_Es. A nickel ion was placed in the model of MarR_Ab. The following ions or molecules were examined and refined in this location in the MarR_Ab structure: water, Na, Mg, K, Ca, Mn, Fe, Co, Ni, Cu, Zn and Ba. Water, Na, Mg, K, Ca and Ba were deemed unacceptable in this site due to poor difference density. Of the remaining ions considered, there were no sources of Mn, Fe, Co, Cu or Zn in the protein expression media, protein purification buffers, protein storage buffer, crystallization condition or cryoprotectant solutions. Thus, we concluded that the ion present in this structure is Ni due to the use of a nickel-affinity column during the protein’s purification. It is unclear why this ion remained bound to MarR_Ab even after the subsequent size exclusion chromatography purification step, or why such an ion is only observed in this structure of the proteins examined. All structures were refined with either Phenix.refine or Refmac using iterative model building in Coot to the final parameters outlined in Supplementary Table 361,62. MarR_73 is a dimer with one protein monomer in the asymmetric unit and the dimer generated by crystallographic symmetry. PDB accession codes and associated crystallographic data are reported in Supplementary Table 3.
Statistics and reproducibility
No statistical method was used to predetermine sample size, but our sample sizes are similar to those reported in previous publications14,63,64. No data were excluded from the analyses. The experiments were randomized. The investigators were not blinded to allocation during experiments and outcome assessment. Where not stated, data distribution was assumed to be normal, but this was not formally tested.
Reporting summary
Further information on research design is available in the Nature Research Reporting Summary linked to this article.
Source: Ecology - nature.com
