A. brasilense contains a functional ethylene-binding protein, AzoEtr1

In plants, ethylene perception is mediated by a family of receptors that contain a conserved N-terminal domain formed by three transmembrane α-helices and seven conserved amino acids known to be required for binding to AtETR1 from A. thaliana [31,52]. Using the amino acid sequence of this domain from AtETR1, we previously determined that many bacterial species contain genes that encode proteins predicted to contain a functional ethylene-binding domain [27,32]. The genome of A. brasilense Sp7 contains one such gene (locus tag OH82_RS30505) that we call Azoetr1. This gene is predicted to encode a protein 525 aa long with an ethylene-binding domain at the N-terminus comprised of three transmembrane α-helices followed by a PAS (Per-Arnt-Sim), and histidine kinase domain at the C-terminus (Figs 1A, S1A). We observed the accumulation of AzoEtr1-YFP at one cell pole when tagged at the C-terminus (Fig 1B). In dividing cells, focal accumulation of AzoEtr1-YFP was observed at poles distal to cell division suggesting that AzoEtr1 is localized at the polar flagellum site. There is a predicted response regulator encoding gene (Azorr^etr1) located just downstream of Azoetr1 in the genome (locus tag OH82_RS30510) that is predicted to encode a protein that is 135 amino acids long and contains no other recognizable output motif (S1B, S2 Figs). Upstream of the Azoetr1 gene in the genome and in the same orientation are genes predicted to encode a peptide chain release factor 3, an aspartate-semialdehyde dehydrogenase, a maltodextrin or glycogen phosphorylase, and a hypothetical protein (S2 Fig). There is a small overlap in nucleotides between the start codon of Azorr^etr1 and the stop codon of Azoetr1 suggesting that the two genes are co-transcribed, perhaps as part of an operon (S3 Fig). This gene orientation suggests that AzoEtr1 is a receptor for a two-component signaling system.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. AzoEtr1 binds ethylene.

A) Predicted domain structure of the AzoEtr1 protein based on sequence homology. Numbers indicate the amino acid range predicted to form each domain. Black rectangles denote predicted transmembrane α-helices that form the ethylene binding domain. Arrows denote the relative location of point mutants constructed in this study. The positions of the gene disruptions using insertion of either a tetracycline (AzoEtr1::Tet^R) or a gentamycin (AzoEtr1::Gm^R) resistance gene are shown. B) Fluorescence confocal microscopy was used to image AzoEtr1-YFP constitutively expressed in A. brasilense as described in the materials and methods. Arrows mark focal accumulation of AzoEtr1-YFP at the poles distal to cell division. Scale bar = 1 µm. C) Alignment of the AzoEtr1 predicted ethylene binding domain with this domain from the receptor characterized in Synechocystis (SynEtr1) and the five receptors from A. thaliana (AtETR1, AtERS1, AtETR2, AtEIN4, AtERS2). The seven amino acids that correspond to the residues required for binding of ethylene to AtETR1 as determined with alanine-scanning mutagenesis [31] are marked with circles. Grey highlights identical amino acids. The approximate locations of the three transmembrane helices are marked based on Wang et al (2006) [31]. D) Ethylene-binding activity to equal amounts of yeast expressing the binding domain of AzoEtr1 fused to GST (AzoEtr1(1-130)GST) or empty vector. Data were normalized to binding activity in AzoEtr1(1-130)GST-expressing yeast. *** p value ≤ 0.001 compared to empty vector as determined by Student’s t-test. E) Ethylene-binding activity to equal amounts of A. brasilense comparing wild-type (Sp7), AzoEtr1:Tet^R, and AzoEtr1::Gm^R. Data were normalized to ethylene-binding activity in wild-type. Different letters denote statistical differences as determined using ANOVA with a p value ≤ 0.05. D,E- Saturable ethylene binding was calculated by subtracting the amount of ethylene bound in the presence of excess ¹²C-ethylene (non-specific binding) from the amount of ethylene bound in the absence of ¹²C-ethylene (total binding) as described in the materials and methods. Data shown represents the average ± SD.

https://doi.org/10.1371/journal.pgen.1011587.g001

Alignment of the N-terminal region of AzoEtr1 with the five ethylene receptors from A. thaliana and the one receptor from Synechocystis showed that AzoEtr1 has 23 amino acids that are identical to the other characterized ethylene receptors (Fig 1C). This includes the seven amino acids necessary for the binding of ethylene to AtETR1 [31]. Other species of Azosprillum also contain putative ethylene receptors with these seven conserved amino acids (S4 Fig). We tested whether AzoEtr1 could directly bind ethylene by expressing the coding sequence for the first 130 amino acids of AzoEtr1 fused to glutathione S-transferase (AzoEtr1[1-130]GST), in Pichia pastoris. Using radioligand binding assays with ¹⁴C₂H₄ we found that cells expressing AzoEtr1[1-130]GST contained saturable ethylene-binding sites (Fig 1D). Control experiments with P. pastoris expressing the empty vector (pPICZ) confirmed that this yeast has no detectable saturable ethylene binding activity as shown previously [32,53]. These results demonstrate that the N-terminal portion of AzoEtr1 is capable of directly binding ethylene.

Based on these results, we examined the binding of ethylene to A. brasilense Sp7 and found that these cells had saturable ethylene binding sites (Fig 1E). Disruption of AzoEtr1 with either a tetracycline resistance insert (AzoEtr1::Tet^R) downstream of helix 3 or a gentamicin resistance insert (AzoEtr1::Gm^R) between helices 1 and 2 (Fig 1A) resulted in cells with little or no saturable ethylene binding (Fig 1E). Thus, A. brasilense Sp7 cells bind ethylene and this binding is mediated by AzoEtr1.

AzoEtr1 is a functional ethylene receptor that affects biofilm formation

We wished to determine whether AzoEtr1 is a functional ethylene receptor that initiates changes in cell physiology, behavior, or both. Ethylene has previously been shown to increase biofilm formation in the cyanobacterium Synechocystis [32] leading us to examine this trait in A. brasilense. To do this we treated A. brasilense with 100 ppb ethylene for three days under biofilm conditions. This level of ethylene is commonly used in plant research and has been shown to affect the physiology of two cyanobacteria, Synechocystis and Geitlerinema sp. PCC 7105, with no obvious adverse effects [32,37,40]. Ethylene reduced biofilm formation by Sp7 and both the AzoEtr1::Gm^R and AzoEtr1::Tet^R disruptants reduced biofilm formation in either the absence or presence of 100 ppb ethylene (Figs 2A, S5A). The application of ethylene did not measurably alter total cell growth under these conditions (Fig 2B). By contrast, disruption of AzoEtr1 led to increased growth (Figs 2B, S5B). Thus, the reduction in biofilm formation is not a consequence of reduced growth. Disruptant cells also showed alterations in other traits associated with stress such as corrugation of the colony surface, increased sensitivity to H₂0₂, and higher accumulation of carotenoids (S6A-D Fig). The application of ethylene to Sp7 cells did not alter the colony surface or H₂0₂ sensitivity. Disruption of AzoEtr1 did not affect aerotaxis (S6E Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Ethylene reduces biofilm formation and A. thaliana root colonization.

A) Biofilm formation after 3 days in the presence or absence of 100 ppb ethylene was assayed using crystal violet staining. The extent of biofilm formation was compared between wild-type (Sp7) and AzoEtr1::Gm^R, and AzoEtr1::Gm^R transformed with Azoetr1-YFP or two point mutants (Azoetr^D35A-YFP, Azoetr1^C75A-YFP). All three were co-transformed with a plasmid containing Azorr^etr1 using a separate plasmid. B) Growth was evaluated by measuring the OD₆₀₀ of the planktonic culture in each assay well in samples treated as in A. C) Proteins from membranes of AzoEtr1::Gm^R and AzoEtr1::Gm^R transformed with either Azoetr1-YFP, Azoetr^D35A-YFP, or Azoetr1^C75A-YFP were analyzed with immunoblots probed with anti-GST antibodies to confirm the expression of the receptors. D) Biofilm formation was allowed to occur for 3 or 5 days in the presence or absence of 100 ppb ethylene compared to 3 days in the absence of ethylene followed by 2 days with ethylene. A, B, D) Data is the average ± SEM. Different letters denote statistical differences as determined using ANOVA with a p value ≤ 0.05. E) The ability of A. brasilense cells to form colony forming units on ethylene-insensitive ein2-5 A. thaliana roots was evaluated 24 hours post-inoculation in the absence or presence of 100 ppb ethylene. Data is the average ± SEM. ** p value ≤ 0.01 compared to the no ethylene control as determined by Student’s t-test. F) Representative fluorescent micrographs of YFP expressing A. brasilense on the surface of ein2-5 roots in the absence or presence of 100 ppb ethylene. Scale bars 10 µm. G) Light microscopic images of A. brasilense cells grown under biofilm conditions for 1 day showing cell attachment. H) SEM images of A. brasilense cells grown under biofilm forming conditions for 3 days. G, H) Wild-type cells (Sp7) treated with ethylene-free air or 100 ppb ethylene and AzoEtr1::Gm^R cells treated with ethylene-free air are shown.

https://doi.org/10.1371/journal.pgen.1011587.g002

To confirm that disruption of AzoEtr1 causes a reduction in biofilm formation, we transformed AzoEtr1::Gm^R cells with Azoetr1-YFP, as well as Azoetr^D35A-YFP, or Azoetr1^C75A-YFP mutant genes predicted to not bind ethylene. In AtETR1, C65 is needed to coordinate the copper cofactor required for ethylene binding and aligns with C75 in AzoEtr1 [29]. Computational modeling predicts that the D25 residue in AtETR1 (aligns with D35 in AzoEtr1) is important for positioning H69 (H79 in AzoEtr1) in helix 2 to coordinate the copper cofactor; D25 is also predicted to form a polar bond with a K91 in helix 3 (K98 in AzoEtr1) which is involved in receptor output [54]. Mutation of the C65 or D25 residue eliminates ethylene-binding activity in AtETR1 and mutation of the comparable C residue in SynEtr1 also eliminates binding activity [31,32,52]. AzoEtr1 has characteristics of a two-component receptor which typically signal via a response regulator protein. The presence of a predicted response regulator encoding gene (Azorr^etr1) in the genome next to Azoetr1 suggests that the output of this signaling system is via AzoRR^etr1. Because we predicted that the AzoEtr1::Gm^R disruption would disrupt Azorr^etr1, all three of these lines were co-transformed with parental Azorr^etr1 using a separate plasmid as detailed in the materials and methods. Transformation with Azoetr1-YFP rescued biofilm formation levels in the absence of ethylene and the application of 100 ppb ethylene reduced this similar to what was observed in Sp7 (Fig 2A). Transformation of AzoEtr1::Gm^R cells with either Azoetr^D35A-YFP or Azoetr1^C75AYFP mutant genes failed to rescue biofilm formation. Transformation with Azoetr1-YFP reduced cell growth to wild-type levels, whereas, the mutant receptor transgenes had a smaller effect on cell growth (Fig 2B). It is possible that the ectopic expression of Azorr^etr1 on a separate plasmid from AzoEtr1 could be reducing the relative expression of one or both genes to affect these results. Expression of the transgenic expressed AzoEtr1 receptor proteins was confirmed with immunoblots (Fig 2C) and the relative transcript abundance of Azorr^etr1 was similar between the complementation lines (S7 Fig). This supports the conclusion that the mutant receptors are non-functional, although, it does not rule out other confounding factors from using two plasmids for ectopic protein expression. This pattern of rescue is similar to what is observed in comparable mutants in the Synechocystis ethylene receptor, SynEtr1, resulting in a non-functional protein [32]. In contrast, this is distinct from results in plants where comparable mutations result in a protein that is functional and signals, but fails to bind ethylene and turn off [52,55]. The transformation of AzoEtr1::Gm^R cells with empty vector plasmids did not rescue the biofilm phenotype and had variable effects on total cell growth (S8 Fig).

We tested whether ethylene disperses biofilms by first allowing biofilm formation for three days in ethylene-free air and then applying 100 ppb ethylene for 2 days. This was compared to biofilm formation after 3 or 5 days in the absence of ethylene (Fig 2D). Our results demonstrate that ethylene did not disperse biofilms that had already formed because the amount of biofilm formed was statistically indistinguishable between 3 days with no ethylene and 3 days with no ethylene followed by application of ethylene. However, additional biofilm formation was blocked since more biofilm formed after 5 days with no ethylene compared to 3 days with no ethylene followed by ethylene for 2 days. These results demonstrate that AzoEtr1 is a functional ethylene receptor that affects bacterial traits involved in stress resistance and host colonization, the first to be described in a proteobacterium.

AzoEtr1 is a functional ethylene receptor that affects root colonization

Ethylene inhibits biofilm formation resulting in reduced cell aggregation. There are many sources of ethylene in soil including plants and bacteria [20]. To determine whether A. brasilense Sp7 uses ethylene as a signal to communicate with other A. brasilense cells, we measured ethylene production by the bacteria. To do this we grew liquid cultures in sealed flasks overnight, normalized the OD₆₀₀ to 1.0, and incubated the cells with known precursors of ethylene biosynthesis for two hours. We sampled the headspace using a laser-based detector. No ethylene was detected above background levels even in the presence of precursors known to increase ethylene production in various bacteria species and plants [56] (S9 Fig). Thus, A. brasilense does not produce ethylene under the conditions tested.

Ethylene could act as a cross-kingdom signal from plants to A. brasilense as EPS production and biofilm formation are important for colonization of plant roots by A. brasilense [57,58]. This led us to question whether ethylene and AzoEtr1 affect root colonization. To evaluate the influence of ethylene on the ability of bacterial cells to colonize roots, we used ethylene insensitive 2-5 (ein2-5) A. thaliana and never ripe (nr) tomato which do not respond to ethylene [59,60], so that we were only affecting bacteria cells with the application of ethylene. Application of 100 ppb ethylene reduced the ability of A. brasilense to colonize the roots of both ein2-5 and nr one day post-inoculation (Figs 2E,F, S10A). Disruption of AzoEtr1 also reduced tomato root surface colonization and fluorescence microscopy showed a loss of cell aggregates on tomato root surfaces (S10B-D Fig).

To investigate the nature of the surface attachment that was reduced by the application of ethylene, we allowed cells to attach to poly-L-lysine-coated slides [61] so that we could visualize the cells under these conditions (Fig 2G). This showed that untreated wild-type cells attached on the slide surface as a dense biomass after 24 h. In contrast, the application of 100 ppb ethylene or disruption of AzoEtr1 resulted in little biomass buildup on the slide and only individual cells or small clusters were visible. To examine for the presence of exopolysaccharide (EPS)-dependent cell aggregation, scanning electron microscopy was performed on samples grown under biofilm conditions (Fig 2H). These images show that wild-type cells in ethylene-free air form large aggregates of cells that contain extracellular fibrils. Both the ethylene-treated and mutant strains did not show large aggregations, and where small aggregates of cells existed, there was little extracellular material visible. Together, these results suggest that ethylene inhibits the EPS production and the aggregation of cells, thereby preventing attachment to surfaces and accumulation of cell biomass.

A. brasilense responds to low levels of ethylene within hours of application

Plants have differential responses over a wide range of ethylene concentrations from over 100 ppm down to 0.2 ppb where transient responses occur within minutes of application [62,63]. We previously showed that the cyanobacteria Synechocystis and Geitlerinema have physiological responses to the application of ethylene at concentrations as low as 8 ppb (the lowest dose examined), and changes in the transcript levels of specific genes in Synechocystis are altered by as low as 1 ppb ethylene [40]. Therefore, we were interested in the sensitivity of A. brasilense to ethylene. To examine this, we conducted biofilm assays over a range of ethylene concentrations from 8 to 700 ppb. This revealed that ethylene down to 8 ppb reduced biofilm formation to a similar magnitude as all higher concentrations used (Fig 3A). Thus, the threshold for this change was below 8 ppb, similar to what we reported for the effects of ethylene on the physiology of cyanobacteria [40].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. A. brasilense responds to low levels of ethylene.

A) Cells were exposed to different levels of ethylene for 24 hours under biofilm conditions and crystal violet used to measure biofilm formation. Different letters denote statistical differences as determined using ANOVA with a p value ≤ 0.05. B) Cells were exposed to 100 ppb ethylene for different times and real-time RT-qPCR used to measure the transcript levels of two genes. Data were normalized to housekeeping genes as described in the materials and methods and then to the levels in the ethylene-free controls at that time point. The average ± SEM shown (* p value ≤ 0.05 compared to the no ethylene control at that time point as determined by Students t-test). C) The gene transcript abundance of these two genes was examined with real-time RT-qPCR as a function of ethylene dose 6 hours after application of the indicated dose of ethylene. Data were normalized to housekeeping genes as noted in the materials and methods and then to the no ethylene control. The average ± SEM shown. * p value ≤ 0.05 compared to no ethylene control as determined by ANOVA.

https://doi.org/10.1371/journal.pgen.1011587.g003

Biofilm assays were conducted over a long period and flow-through chambers were used to deliver ethylene so that O₂ and CO₂ were kept constant. However, a limitation of this method is the difficulty in reliably delivering ethylene doses below approximately 8 ppb. We therefore conducted shorter-duration experiments examining changes in the transcripts of two selected genes using much lower ethylene doses where ethylene was simply injected into sealed chambers for several hours. For this we chose OH82_RS30850 (annotated as pyruvate dehydrogenase complex E1 subunit beta) and OH82_RS14895 (annotated as an amino acid permease) because our RNA sequencing (RNA-seq) showed that both of these genes are upregulated by the application of ethylene (see next section). We first determined how quickly gene transcript changes could be measured using real-time RT-qPCR. In Synechocystis, ethylene rapidly affects the transcript levels of genes with changes occurring within 30 min [32]. Here we observed no change in transcript levels before four hours after addition of 100 ppb ethylene and OH2_RS30850 did not show a response until six hours after ethylene (Fig 3B). Therefore, we conducted dose-response measurements of these genes 6 h after application of ethylene treatment (Fig 3C). The transcript abundance of OH82_RS30850 increased with ethylene doses as low as 0.1 ppb. A similar trend was observed for OH82_RS14895; however, a statistically significant increase was not observed until 1 ppb. These ethylene levels are below the threshold for most plant responses.

Ethylene causes wide-spread changes in the metabolism of A. brasilense

To further explore the effects of ethylene on bacterial physiology, we used RNA-seq to examine transcriptomic changes in cells grown in liquid culture with NaNO₃ as a nitrogen source for 4 h in the presence or absence of 100 ppb ethylene. Principal component analysis (PCA) showed that the ethylene-treated samples clustered separately from the controls (S11A Fig). However, one of the control samples did not cluster with the other two controls; therefore we performed a new PCA analysis without this sample (S11B Fig). Further analyses were conducted without this control sample. Using an adjusted p value (Benjamini-Hochberg Procedure) less than 0.05, we observed that 395 of the 6274 genes responded to ethylene with a log₂ change ≥ |0.5|. Of these, 51 were downregulated and 344 were upregulated (Fig 4A and S1 Data). Gene Ontology analysis showed that many of the annotated genes are predicted to be involved in metabolism including energy production and the metabolism or transport of amino acids, carbohydrates, lipids, secondary metabolites, coenzymes, and inorganic ions (Fig 4B). Of these 395 transcripts, 36 had a log₂ change ≥ |1.0| with 31 upregulated and five downregulated (Fig 4C and Table 1). When examining the top 36 genes in the context of the entire differentially expressed gene list, many grouped with other genes in the genome that were similarly regulated by ethylene with an adjusted p value ≤ 0.05, but with smaller amplitude changes in transcript abundance (S1 Table). These included 15 genes annotated to be involved in transport, 11 potentially involved in redox reactions, and several components of the pyruvate dehydrogenase complex. There were also several other genes that grouped together on the genome and showed significant upregulation, but did not include the top 36 genes (S2 Table). In these groups six genes are annotated as ABC transporters, eight that contain coenzyme A, and six hydratases/dehydratases. Thus, genes in these groups are likely to be co-transcribed and may represent operons. The levels of Azoetr1 and Azorr^etr1 transcripts were not significantly altered by ethylene treatment in the RNA-seq dataset. qPCR of these genes revealed only a small transient change in Azoetr1 and no change in Azorr^etr1 caused by the application of ethylene (S12 Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Global gene transcript changes caused by ethylene.

A. brasilense cells were grown in liquid culture as detailed in the materials and methods for 4 h in the presence or absence of 100 ppb ethylene. Samples were prepared as detailed in the materials and methods and RNA-seq analysis and bioinformatics analyses was carried out by GeneWiz (South Plainfield, NJ, USA). A) Heatmap showing relative abundance of each gene in individual control (no ethylene) and ethylene-treated samples. Included are gene transcripts that have a log₂ change ≥ |0.5| with an adjusted p-value ≤ 0.05. B) Pie charts showing the percent of genes downregulated (top) or upregulated (bottom) by ethylene in various Gene Ontology categories using the Kyoto Encyclopedia of Genes (KEGG). C) Volcano plot showing global transcriptional changes caused by ethylene treatment. Each data point represents a gene. Genes with an adjusted p-value ≤ 0.05 and a log2 fold ≥ 1 are indicated by red dots. Genes with an adjusted p-value ≤ 0.05 and as log2 fold change ≤ -1 are indicated by blue dots.

https://doi.org/10.1371/journal.pgen.1011587.g004

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Top 36 Gene Transcripts Affected by Ethylene^a.

https://doi.org/10.1371/journal.pgen.1011587.t001

RNA-seq analyses predicted that ethylene treatment elicited major metabolic changes in the bacteria. Therefore, we performed untargeted metabolomics to assess the patterns of metabolites in cells treated with ethylene versus untreated cells under the same conditions as used for the RNA-seq experiment. Cells treated with ethylene had significantly altered metabolite profiles between treated and untreated cells as determined by partial least square discriminant analysis (Fig 5A). Of the 123 metabolites detected, 69 were significantly altered (p ≤ 0.1) between the two conditions with 60 enriched and nine depleted (S3 Table). Many metabolites in carbohydrate and carbon metabolism such as 3-phosphoglycerate, glyceraldehyde 3-phosphate, fructose 1,6-bisphosphate, glucose phosphate, phosphoenolpyruvate, pyruvate, and UDP-glucose were increased by ethylene treatment. A key regulator of nitrogen metabolism, 2-oxoglutarate, was also increased [64,65]. By contrast, the second messenger cyclic di-GMP, which is involved in the regulation of biofilm formation and metabolism [66,67], was reduced by ethylene. Pathway analysis showed that nucleotides and analogs were the most highly altered pathways, followed by amino acid metabolism (Fig 5B). These observations suggest that nitrogen metabolism is a major target of ethylene signaling in this bacterium.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Ethylene alters the levels of many metabolites.

A. brasilense cells were treated with 100 ppb ethylene or ethylene-free air for 24 h. At this time, untargeted metabolomics analysis was conducted as described in the Materials and Methods. A) Partial least squares discriminant analysis of data. B) Pathway analysis using the Kyoto Encyclopedia of Genes (KEGG) showing pathways of the 69 differentially accumulated metabolites.

https://doi.org/10.1371/journal.pgen.1011587.g005

Ethylene mediates changes in the balance between carbon and nitrogen metabolism

Our transcriptomic data indicated that carbon metabolism was upregulated and nitrogen metabolism was downregulated by ethylene. Non-targeted metabolomics also indicated that ethylene treatment increased the abundance of metabolites related to glycolysis and gluconeogenesis. These observations suggest that the perception of ethylene can alter the balance between carbon and nitrogen metabolisms in A. brasilense.

Three genes with regulatory roles in nitrogen metabolism were downregulated by ethylene in the RNA-seq datasets: glnA (OH82_RS05785) encoding a glutamine synthetase, glnB (OH82_RS05780) encoding a PII regulatory protein, and amtB (OH82_RS13415) encoding an ammonium transporter. All three gene products have been shown to regulate GlnAB and play important roles in the regulation of nitrogen metabolism in A. brasilense [68–73] as well as many other bacteria. GlnB is a PII protein implicated in regulating bacterial metabolism in response to changes in the cellular carbon to nitrogen ratio [68,70,72,73] and is a major regulator of nitrogen metabolism including free-living nitrogen fixation. A. brasilense also possesses a second PII protein, GlnZ, that plays a distinct role from that of GlnB in the regulation of nitrogen fixation [72,70,74–75]. The glnZ gene was not altered by ethylene in our RNA-seq data.

To validate the changes in the abundance of glnB and amtB as a result of ethylene treatment, we used real-time RT-qPCR to compare the expression of glnB and amtB under nitrogen fixing conditions and included glnZ as a negative control. Consistent with the RNA-seq data, we found that the application of 100 ppb ethylene reduced the expression of both glnB and amtB but had no effect on glnZ (Fig 6A). The promoter region of Azoetr1 contains a putative binding site for the sigma-54 factor RpoN (S3 Fig). RpoN controls many nitrogen-related behaviors in A. brasilense including nitrogenase expression [76]. Given the regulatory role of GlnB in the induction of nitrogen fixation and the role of RpoN in controlling nitrogen fixation gene expression, we tested the effect of ethylene treatment on the induction of nifH (OH82_RS22550) that encodes a subunit of the nitrogenase enzyme [77–79]. As expected, the transcript levels of nifH were reduced in the presence of 100 ppb ethylene (Fig 6A). Azoetr1 expression was similar when grown in the presence of either NaNO₃ or NH₄Cl as the nitrogen sources (Fig 6B). Consistent with ethylene signaling being linked to nitrogen metabolism, Azoetr1 expression was increased either in the absence of a nitrogen source (nitrogen fixation conditions) or in the presence of 2-oxoglutarate which mimics the high carbon to nitrogen ratio signal that regulates GlnB activity. Together these findings are consistent with ethylene mediating changes in the balance between carbon and nitrogen metabolism, most likely by affecting the abundance of glnB transcript levels, thus altering the ability of cells to respond to shifts in the carbon-to-nitrogen ratio.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Ethylene alters the metabolic balance and stimulates poly-hydroxybutyrate (PHB) accumulation.

A) Cells were grown with no nitrogen source in the presence and absence of 100 ppb ethylene for 24 hours at which time samples were processed and real-time RT-qPCR used to measure the transcript levels of the indicated genes. Data were normalized to expression of housekeeping genes and then to expression in the absence of ethylene as described in the materials and methods. Data is the average ± SEM. * p value < 0.05; ** p value < 0.01 compared to the no ethylene control as determined by Student’s t-test. B) Cells were grown overnight with 10 mM of the indicated nitrogen source or no added nitrogen with no shaking. At this time RNA was extracted and real-time RT-qPCR used to measure the transcript level of Azoetr1. Data was normalized to levels in the presence of NaNO₃ and housekeeping genes as described in the materials and methods. Data is the average ± SEM. Statistical difference from NaNO₃ was determined with ANOVA. n.s. not significant; * p value < 0.05; ** p value < 0.01. C, D) Cells were grown in MMAB in the presence and absence of 100 ppb ethylene for 24 hours at which time the cells were stained with Nile red. C) Representative DIC merged with fluorescent images of cells in control cells treated with 100 ppb ethylene. Scale bar = 1 µM. D) Stack graph showing the number of PHB granules accumulated per cell in control (n = 389) and ethylene-treated cells (n = 251).

https://doi.org/10.1371/journal.pgen.1011587.g006

An imbalance between nitrogen and carbon metabolism regulates the accumulation of poly-hydroxybutyrate (PHB) which is also regulated by GlnB in A. brasilense [80]. PHB serves as a carbon reserve for many bacteria and its accumulation in A. brasilense promotes stress endurance [81]. Using cells stained with the lipophilic dye Nile Red and microscopy, we showed that ethylene treatment increased the number of cells that contained PHB granules, including a statistically significant increase (Mann–Whitney U test p < 0.001) in the median number of PHB granules per cell from 1 to 3 under these conditions (Fig 6C, D and S2 Data). These results are consistent with the RNA-seq data showing a decrease in the expression of a gene encoding a poly-hydroxybutyrate depolymerase, phaZ, (OH82_RS11420), which is involved in the breakdown of PHB [82].

Bacterial strains and growth conditions

Wild-type Azospirillum brasilense Sp7 and AzoEtr1 mutants in the same background were used throughout this study. Bacterial strains and plasmids used in this study are listed in Table 2. Cells were grown at 28°C with shaking in either tryptone-yeast (TY) or Minimal Media for A. brasilense (MMAB) supplemented with 10 mM NaNO₃ unless otherwise indicated. For cultures grown in the absence of nitrogen or 2-oxoglutarate, an overnight culture grown in TY was resuspended in the designated media and adjusted to an OD₆₀₀ of 1.0. These cultures were grown overnight without shaking to allow for biological nitrogen fixation. Growth in TY was supplemented with appropriate antibiotic selections (ampicillin 200 µg ml^-1 for all A. brasilense, 10 µg ml^-1 tetracycline for Tet^R insertions, and 50 µg ml^-1 gentamycin for Gm^R insertions).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. List of strains and plasmids used in this study.

https://doi.org/10.1371/journal.pgen.1011587.t002

Localization of AzoEtr1

Gateway cloning, according to the manufacturer’s recommendations, was used to create C-terminal YFP fusions of AzoEtr1 using the pRH005 destination vector. Gene specific primers were prepared and Sp7 genomic DNA was used as a template. PCR products were then subjected to BP and LR cloning steps to move the PCR product first into pDONR2.1, an expression vector before being moved into pRH005, the destination vector, yielding the YFP fusion constructs under constitutive expression. These constructs were conjugated to wild-type A. brasilense Sp7 cells. Cells expressing AzoEtr1-YFP were grown and image stacks were captured using a Leica

SP8 confocal microscope. Images were analyzed by the maximum intensity projection of 25 stacks. Images were processed using Lightning deconvolution software.

Disruption of Azoetr1 and complementation in A. brasilense

Azoetr1 was disrupted with either a tetracycline resistance insert (AzoEtr1::Tet^R) downstream of helix 3 or a gentamicin resistance insert (AzoEtr1::Gm^R) between transmembrane helices 1 and 2. Insertional mutagenesis was performed as previously described [111]. Gateway cloning according to manufacturer recommendations was used to create the C-terminal YFP fusion of AzoEtr1 using the pRH005 destination vector using the 500 bp upstream of Azoetr1 as the promoter. Gene specific primers were made and Sp7 genomic DNA was used as a template. PCR products then underwent BP and LR cloning steps to move the PCR product first into pDONR2.1, an expression vector before being moved into pRH005, the destination vector, yielding the YFP fusion constructs. To create AzoEtr1^D35A and AzoEtr1^C75A mutants containing the D35A and C75A point mutations respectively, Azoetr1 was placed into the pUC19 plasmid and site-directed mutagenesis was performed using the primers listed in S4 Table. After mutagenesis, the same primers used for wild-type Azoetr1 gateway cloning were again used to move AzoEtr1^D35A and AzoEtr1^C75A into the pRH005 vector for expression. DH5α E. coli were transformed with constructs confirmed by sequencing and mated into AzoEtr1::Gm^R cells using the triparental conjugation protocol for A. brasilense [111]. Complementation with Azorr^etr1 was performed by amplification with gene specific primers encoding restriction cut sites. After digestion of both the template DNA and pBBR1-MCS3 plasmid, Azorr^etr1 was ligated into the pBBR1-MCS3 plasmid putting Azorr^Etr1 expression under the lac promoter. pBBR1-MCS3 containing Azorr^etr1was mated into AzoEtr1::Gm^R pRH005-Etr1, AzoEtr1::Gm^R pRH005-Etr1^D35A, and AzoEtr1::Gm^R pRH005-Etr1^C75A, using a triparental conjugation protocol. No induction of Azorr^Etr1 with IPTG was necessary because leaky expression was enough to give comparable levels of Azorr^etr1 as wild-type A. brasilense (S7 Fig).

Cloning of Azoetr1 and expression of AzoEtr1 in P. pastoris

When we started this project, A. brasilense Sp7 had not yet been sequenced. Therefore, we cloned AzoEtr1 from A. brasilense Sp245 (now known as A. baldaniorum) [112] and PCR amplified a truncated version (the first 390 bp) of the gene corresponding to the coding region for the putative ethylene-binding domain (first 130 amino acids). Amplification was carried out by PCR with a 5’ EcoR1 restriction site and a 3’XhoI restriction site using 5’-TGAATTCATGTTCGGTGGCGTGGAAGCCTTC-3’ and 5’-GTCTCGAGGTCGGCGAGTTGCGTGGC-3’ primers. Sequencing A. brasilense Sp7 showed that the amino acid sequences of A. brasilense Sp7 and A. baldaniorum are identical for the first 130 amino acids used for ethylene-binding assays (S4 Fig). The PCR fragment was subcloned into a pGEM-T vector and DH5α E. coli cells transformed with this plasmid. The plasmids were isolated and AzoEtr1 excised, and subsequently ligated into the P. pastoris expression vector, pPIKZA containing a GST tag at the 3’ end of the gene, as previously described [32]. P. pastoris was transformed with these plasmids as previously described by [53]. AzoEtr1[1-130]GST was expressed in yeast cells using the protocol described in the Invitrogen P. pastoris manual. Following at 48-hour induction with methanol, cells were harvested, washed once with water, frozen in liquid nitrogen, and stored at -80°C until use.

Whole-cell ethylene-binding assays

¹⁴C₂H₄ was obtained from ViTrax Radiochemicals (Placentia, CA, USA) and trapped in mercury perchlorate as previously described [52,113]. For both bacteria and yeast, cells were thawed and whole-cell ethylene-binding assays were carried out as previously described [29,52,113]. For yeast cells, 1 g fresh weight per sample was used and for bacteria cells, 3 g fresh weight per sample were used. The cells were exposed to 400 ppb ¹⁴C₂H₄ in the presence or absence of excess ¹²C₂H₄. Saturable ethylene binding was calculated by subtracting the amount of ethylene bound in the presence of excess ¹²C₂H₄ (non-specific binding) from the amount of ethylene bound in the absence of ¹²C₂H₄ (total binding). All experiments were repeated at least three times. Yeast data was normalized to saturable binding in cells expressing AzoEtr1, and bacterial cells to saturable binding in A. brasilense Sp7.

Sequence alignment and domain prediction

Sequence alignment was performed using Clustal Omega software (https://www.ebi.ac.uk/jdispatcher/msa/clustalo). Genomic predictions of Azoetr1 and Azorr^etr1 were performed using the Basic Local Alignment Search Tool (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Protein domain predictions were done using Simple Modular Architecture Research Tool (http://smart.embl-heidelberg.de/) [114,115].

Crystal violet biofilm staining

Biofilm formation was assayed by crystal violet staining [84] in 12-well PVLC plates with the lid pierced above each well with a 27-gauge needle and covered with micropore surgical tape to ensure equal gas flow between the wells. The plates were placed in airtight chambers and treated with the specified concentration of ethylene for 3 days before crystal violet staining. After 3-days, liquid was removed from the wells and stained with 0.1% crystal violet for 20 minutes. Crystal violet was then removed and each sample washed with distilled water three times. After washing, the crystal violet was solubilized in 95% ethanol and the absorbance at 600nm (A₆₀₀) was measured. Growth of cells in separate samples was determined by measuring the OD₆₀₀ of the planktonic culture in each well. Data represent the analysis of at least three biological replicates for each condition which were each analyzed with three technical replicates

Abiotic surface attachment on slide assays

A. brasilense cultures were grown in TY media overnight with appropriate antibiotics and adjusted to an OD₆₀₀ of 1.0 in 1mM NaNO₃ supplemented MMAB used in previous biofilm experiments. Microscope slides were coated with poly-L-lysine to increase cell adhesion [61] prior to the application of 10 µL samples. These were placed into 500 mL mason jars without cover slips and sealed with lids fitted with rubber septa to inject ethylene. Humidity was maintained in the chambers with wet paper. Samples were treated with either 100 ppb ethylene or ethylene-free air and incubated at 28°C for 24 h. After this time, the slides were removed and imaged on a Zeiss Axio Observer Z1 at 100x.

Scanning electron microscopy

A. brasilense was grown in liquid MMAB media with antibiotics and incubated with shaking at 28°C until the culture reached OD₆₀₀ of 0.4. The samples were centrifuged and washed with Chemotaxis (Che) buffer (10 mM phosphate buffer plus 0.1 mM EDTA) three times, resuspended in biofilm media, and placed in Petri dishes. Silicon chip specimen supports (TED PELLA, Inc.) were sterilized with 95% (v/v) ethanol, coated with poly-L-lysine to increase cell adhesion [61], placed in Petri dishes with cells. The samples were placed in an airtight container and sealed with lids fitted with rubber septa to inject ethylene. Samples were treated with 100 ppb ethylene or ethylene-free air and incubated at 28°C without shaking for 72 h to allow formation of biofilm. The samples were prepared for scanning electron microscopy and visualized using a Hitachi TM3030 scanning electron microscope.

Root colonization assays

Root colonization assays were performed on tomato and A. thaliana with a modification of a prior study [116]. Briefly, tomato seeds were surface sterilized with 50% (v/v) bleach for 30 min to remove and kill microbes on the seed coat, washed three times with distilled H₂O, plated on 0.8% (w/v) agar plates containing ¼ strength Murashige and Skoog salts [117] and grown for 3d in a 16 h:8 h light:dark cycle at 28°C. Seedlings were then transferred to 50 mL chambers (2 seedlings per chamber) containing Fahraeus media with 0.6% (w/v) agar and inoculated with A. brasilense. Four chambers were used for each assay. Colonization assays were performed in the presence or absence of 100 ppb ethylene for 24 hours. The roots were then removed and gently rinsed with distilled water followed by root tissue homogenization using a bead homogenizer. The resulting mixture was serially diluted to 10^-8 and plated to count the CFUs. The colonization index was calculated as previously described [116], and normalized to either wild-type or ethylene-free conditions.

For A. thaliana colonization, seeds were surface sterilized with 70% (v/v) ethanol and 10% (v/v) bleach plus 0.1% (v/v) triton X-100. They were then stratified for 3 days and plated on 0.8% (w/v) agar plates containing ¼ strength Murashige and Skoog salts [117] and grown for 4d in light at 22°C. The seedlings were dipped into concentrated bacteria placed on fresh agar plates. Colonization was allowed to occur for 24 h in the presence or absence of 100 ppb ethylene. Whole seedlings were weighed and homogenized before serial dilution and plating for CFU counting. CFUs per gram of plant fresh weight were calculated.

Fluorescent microscopy of root colonization

In some cases, root colonization was performed as described above using YFP-expressing A. brasilense. These were generated by conjugation with the gateway destination vector pRH005 to express KanR promoter-YFP fusion. For tomato colonization, assays were carried out in microscopy slide-in-chambers containing one tomato plant as previously described [118]. These chambers were maintained at 28°C for 24 h before imaging using a Leica SP8 confocal microscope. The colonization of A. thaliana roots was performed as described above. After 24 h, the plants were removed from the wells and placed onto slides with 5 µL of water. Images were taken on a Zeiss Axio Observer Z1 microscope.

Ethylene measurements

To measure ethylene production by A. brasilense, liquid cultures were grown to an OD₆₀₀ of 1.0 and sealed for 2 h before sampling the headspace using an ETD-300 photoacoustic ethylene detector (Sensor Sense B.V., Nijmegen, The Netherlands). Some samples were supplemented with 1 µM 1-aminocyclopropane-1-carboxylate, 1 mM methionine, or 1 mM 2-(methylthio)ethanol as these are known precursors for ethylene biosynthesis in plants and bacteria [56]. The experiments were replicated twice.

RNA extraction and qRT-PCR

For gene expression analysis, RNA was extracted from A. brasilense grown overnight in MMAB in the absence of ethylene or at the designated ethylene concentration. For the gene expression analysis of glnZ, glnB, amtB, and nifH cells were grown overnight in MMAB, washed and resuspended at an OD₆₀₀ of 1.0 in MMAB without a nitrogen source and grown without shaking for 24 hours in the presence or absence of ethylene at the indicated concentration. RNA isolation was performed using the TRIzol:chloroform method with some modifications. Briefly, 2 mL of cells were pelleted before resuspension in 550 µL of TRIzol and dissolution of cell pellets. After cell pellets were dissolved, 250 µL of lysozyme was added, and samples were incubated for 20 min at 37°C. After incubation, the samples were placed on ice for 5 min before the addition of 200 µL chloroform. cDNA synthesis was performed using the Sensifast cDNA synthesis kit. RNA was normalized to 100 ng µL^-1 and 100 ng µL^-1 cDNA was prepared. qPCR was performed using a SYBR-NOROX kit. Primers for quantitative PCR were designed using Integrated DNA Technologies and gene expression levels relative to the previously validated reference genes GlyA and RecA [119] calculated following the 2^-∆∆ct method [120]. The data were then normalized to the expression in the absence of ethylene. The primers used to quantify Azoetr1, Azorr^etr1, OH82-RS30850, and OH82_RS14895 are listed in S5 Table. All other primers used were from prior studies [49,121,122].

RNA-seq

To determine the downstream targets of the ethylene signaling pathway, we conducted an RNA-seq experiment in which cells were grown in liquid culture in MMAB with NaNO₃ for 16 h and then for 4 h in the presence or absence of 100 ppb ethylene. RNA was extracted as previously described. RNA-seq and bioinformatics analyses were performed using GeneWiz (South Plainfield, NJ, USA). The Wald test was used to generate p values. Genes with adjusted p values ≤ 0.05 and absolute log₂ fold changes ≥ 1 were considered differentially expressed. Gene Ontology categories were determined using the Kyoto Encyclopedia of Genes (KEGG). The RNA-seq data discussed in this publication have been deposited in the NCBI GEO Omnibus [123] with GEO Series Accession Number GSE179776 (https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE179776).

Immunoblots

Proteins from membranes isolated from A. brasilense were separated by SDS-PAGE. Wild-type and mutant AzoEtr1 protein expression were analyzed with immunoblots using a 1:1000 dilution of a polyclonal anti-GFP antibody (Abcam). An infrared IRDye800 (LI-COR) secondary antibody was used, and detected using an Odyssey CLx Imager (LI-COR).

Metabolomics

Samples were grown in TY overnight before being washed in Chemotaxis (Che) Buffer, and the OD₆₀₀ was normalized to 1.0. These samples were used to inoculate 10mL of MMAB supplemented with NaNO₃. The samples were grown for 24 h in sealed tubes injected with 100 ppb ethylene or an equivalent volume of ethylene-free air. After 24 h of growth, the cells were pelleted by centrifugation, the supernatant was removed, and the samples were flash-frozen in liquid nitrogen before analysis by LC-MS/MS. Partial least squares differential analysis was performed to determine the separation of the metabolic profiles. The analysis was performed at the Biological and Small Molecule Mass Spectrometry Core Facility in the Department of Chemistry at the University of Tennessee. MetaboAnalyst 5.0 was used for partial least squares discriminant analysis and to conduct pathway analysis [124]. Data represent the analysis of three biological replicates for each condition which were each analyzed with three technical replicates and normalized to tissue fresh weight. Pathway analysis of metabolites was performed using the Kyoto Encyclopedia of Genes (KEGG) (https://www.genome.jp/kegg/).

PHB staining

Cells were grown overnight in TY medium and washed with Che Buffer before adjusting the OD₆₀₀ to 1.0 in Che buffer [118]. These samples were used to inoculate 50 mL MMAB supplemented with 10 mM NaNO₃. The cultures were grown in flow-through chambers and treated with either 100 ppb ethylene or ethylene-free air. Stirring was achieved using a stir bar set to 200 RPM. Growth was allowed to occur for 24 h after which 2 mL of these cultures were collected and washed with Che Buffer before staining with 0.5% (w/v) Nile Red for 30 min., and then imaged using a Zeiss Axio Observer Z1 microscope. Images were analyzed using ImageJ/Fiji (ver. 1.52E) to determine the number of foci per cell. The statistical significance of the differences in the median values was determined using the Mann-Whitney Test.

Oxidative stress assays

Sp7 and AzoEtr1::Tet^R cells were grown on 1.5% (w/v) TY agar for 24 h at which time a 6 mm filter paper disk soaked in 0.3% (v/v) hydrogen peroxide was placed in the center of the agar dish. After 24 h, the zone of inhibition diameter was measured.

Carotenoid extraction

Cultures were grown for three days in TY medium at 28°C at which time cells were pelleted and resuspended in an equal volume of 100% (v/v) methanol. The samples were shaken overnight at room temperature at a speed of 180 RPM. The OD₄₈₅ of the supernatant was determined using an Evolution 60 spectrophotometer (Thermo Scientific).

Colony morphology

The cells were grown overnight in TY medium and adjusted to an OD₆₀₀ of 1.0. 10µL spots were made on TY 1.5% (w/v) agar plates and allowed to dry and grow for three days before imaging with an Olympus SZH10 microscope using a Canon EO5 camera.

Aerotaxis Assay

Aerotaxis assays were performed as previously described with some changes [67]. Cells were grown to an OD₆₀₀ of 0.6 in MMAB and washed in Che buffer before resuspension to an OD₆₀₀ of 0.6. An optically flat capillary tube was then used to draw cells from the cell suspension. This capillary tube was placed in focus under a microscope and a flow nitrogen gas was used for 3 minutes to abolish existing oxygen gradients. The gas flow was then switched to air and band formation was visualized under a light microscope at 4x.

Statistics

ANOVA and Mann-Whitney tests were performed using GraphPad Prism (ver. 10.0.2). Student’s t-test was performed done using Microsoft Excel.