(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .



Evolutionary innovation through transcription factor rewiring in microbes is shaped by levels of transcription factor activity, expression, and existing connectivity [1]

['Matthew J. Shepherd', 'Milner Centre For Evolution', 'Department Of Life Sciences', 'University Of Bath', 'Bath', 'United Kingdom', 'Division Of Evolution', 'Genomic Sciences', 'School Of Biological Sciences', 'University Of Manchester']

Date: 2023-11

The survival of a population during environmental shifts depends on whether the rate of phenotypic adaptation keeps up with the rate of changing conditions. A common way to achieve this is via change to gene regulatory network (GRN) connections—known as rewiring—that facilitate novel interactions and innovation of transcription factors. To understand the success of rapidly adapting organisms, we therefore need to determine the rules that create and constrain opportunities for GRN rewiring. Here, using an experimental microbial model system with the soil bacterium Pseudomonas fluorescens, we reveal a hierarchy among transcription factors that are rewired to rescue lost function, with alternative rewiring pathways only unmasked after the preferred pathway is eliminated. We identify 3 key properties—high activation, high expression, and preexisting low-level affinity for novel target genes—that facilitate transcription factor innovation. Ease of acquiring these properties is constrained by preexisting GRN architecture, which was overcome in our experimental system by both targeted and global network alterations. This work reveals the key properties that determine transcription factor evolvability, and as such, the evolution of GRNs.

Data Availability: All raw data for this study is available on the Open Science Framework (OSF), and can be accessed at https://osf.io/pcdhx/ . Data is also available in supplementary data files where cited. RNAseq files are available on NCBI GEO accession GSE228016.

(A) Flagellar motility rescue experiment route map outlines typical progression of motility evolution. Pathway diagrams display key components of the primary (B, Taylor and colleagues [ 37 ] ) and alternative (C, this study ) rewiring pathways for rescue of flagellar gene expression. Genes are coloured in white, protein components in green. Mutational targets and their effects are shown in red, and whether a mutation occurs in the first or second evolutionary step indicated. Figure created with BioRender.com .

While there exists some empirical evidence of the importance of transcription factor rewiring in driving evolutionary innovation [ 25 , 34 ] and generation of novel transcription factors through gene duplication [ 35 ], in each case, it is unknown why the transcription factor in question rewired as opposed to any of the other regulators within their protein family. For example, structurally and functionally similar phage repressor proteins λ cI and P22 c2 differ in their ability to regulate expression from non-cognate sites with λ cI being more evolvable, a property that correlates with higher robustness to mutation [ 36 ]. What role—if any—GRN structure plays in these processes and whether the rewired transcription factors were unique in their ability to rewire within each study system is also undetermined. To address these questions, we set out to investigate the evolution of transcription factor rewiring using rescue of flagellar motility via rewiring of NtrC documented by Taylor and colleagues [ 37 ]. In this model system, Pseudomonas fluorescens SBW25 is engineered to be non-motile via deletion of the master regulator for flagellar synthesis (fleQ) and abolishment of biosurfactant production. Placing these mutants in soft agar plates results in strong selection for rescue of motility—bacteria will grow, exhaust available nutrients, and starve, unless they acquire a mutation that restores motility, allowing them access to uncolonised areas of the agar plate [ 37 , 38 ]. Under strong selection for motility, the bacteria rapidly and reliably evolve new regulatory network wiring to rescue flagellar motility (a schematic of these experiments is laid out in Fig 1A ) [ 37 ] with the same transcription factor (ntrC) repurposed to rescue flagellar-driven motility each time, to the exclusion of all other homologous transcription factors within the protein family [ 39 ]. However, we do not know the factors that determine this evolutionary preference. To test this, we constructed a double fleQ ntrC knockout and placed this mutant under strong selection for swimming motility. This forces evolutionary utilisation of an alternative transcription factor in order to rescue motility and is an established method for unveiling hidden evolutionary pathways [ 40 ]. We reveal a hierarchy among transcription factors that are rewired to rescue lost flagellar function, with alternative rewiring pathways only revealed after the preferred co-opted transcription factor, NtrC, is eliminated. Identification of an additional transcription factor capable of rewiring within the same GRN background allows us to investigate rules governing when and where a transcription factor rewires within its GRN and is important for understanding how these regulatory systems innovate during adaptation to environmental challenges.

Within GRNs, a key mechanism of innovation is transcription factor rewiring, in which transcription factors can gain or lose regulatory connections to target genes creating new network architectures and opportunity for phenotypic innovation [ 18 ]. Rewiring events can have dramatic effects on the transcriptome [ 19 ] that can drive phenotypic diversification. However, the majority of past studies on transcription factor rewiring involve retrospective experimental dissection of networks that have already diverged [ 20 – 25 ], which allows inference of past rewiring events but does not address the evolutionary factors driving the rewiring process. For rewiring to occur, a transcription factor must first have the potential interaction with non-cognate regulatory targets, allowing regulation of a new gene [ 26 ], as is the case for innovation in other proteins [ 27 , 28 ]. The potential for non-cognate interactions comes from built-in homology between paralogous components of different regulatory networks—a consequence of gene duplication and divergence [ 29 ]. Non-cognate interactions can become meaningful and drive adaptation if favoured by natural selection [ 30 , 31 ]. However, non-cognate interactions cannot be commonplace within GRNs as this will likely result in dysregulation of gene expression and fitness costs for an organism in the environment to which it is adapted [ 32 , 33 ].

During conditions of environmental upheaval and niche transition events, rapid phenotypic adaptation is essential for evolutionary success [ 1 – 3 ]. To understand patterns of diversification in novel environments, we need to understand why some evolutionary transitions occur more rapidly than others, and what allows some organisms to succeed where others fail. Key to this is understanding the evolution of gene regulatory networks (GRNs), control circuits common throughout the domains of life that determine the magnitude and timing of gene expression [ 4 ] in response to environmental and internal signals [ 5 , 6 ]. GRNs are frequent sites of adaptive mutation driving phenotypic evolution [ 7 – 9 ] and often underscore adaptation to—and survival in—new and changeable environments [ 10 – 13 ]. Alterations to GRNs can also enhance drug resistance and stress responses in the face of environmental challenges [ 14 – 17 ]. Determining how these networks evolve and identifying rules governing when and how a regulatory circuit adapts will allow us to better understand their role in determining the evolutionary success of an organism.

Results

Evolutionary rescue of flagellar motility can occur in the absence of ntrC through de novo mutation to an alternative two-component system When previously challenged to rescue flagellar motility in the absence of FleQ, rewiring almost exclusively occurred through the NtrC transcription factor [37,38]. Perhaps we only see NtrC co-opted to rescue FleQ function because it is the only transcription factor capable of doing this? It is known that within families of homologous transcription factors, there is variation in the ability to bind non-cognate sites [36], so it is possible that NtrC is unique in its ability to rewire. FleQ and NtrC are part of a family of 22 structurally related transcription factors called RpoN-dependent enhancer binding proteins (RpoN-EBPs), many of which are predicted to be more structurally similar to FleQ than NtrC through 3D structural modelling [39]. To identify if any other RpoN-EBPs were capable of rewiring, the double knockout non-motile P. fluorescens (ΔfleQΔntrC) was challenged to rescue flagellar motility in 0.25% agar lysogeny broth (LB) plates. Motile zones reemerged in a two-step manner (a slow-swimming variant, then a faster-swimming variant) as typical of previous studies [37]. Motile isolates were sampled and whole genome resequenced at each step. Motility-granting mutations were identified in the gene PFLU1131 for all first-step motile isolates (n = 15), and in 13/15 cases, this was the only mutant gene (S1 File). PFLU1131 is unstudied outside of our own study system [40,41] and encodes a putative sensor kinase, a protein which—in response to a signal—will modulate the activity of a cognate transcription factor through phosphorylation or dephosphorylation. The cognate transcription factor is typically grouped in the same operon as the kinase, and the PFLU1131 gene is situated in an operon between 2 other genes, PFLU1130 and PFLU1132, which encode a hypothetical GNAT-acetyltransferase and a putative RpoN-dependent transcriptional regulator, respectively. PFLU1132 is a known FleQ-homolog [39] and together with PFLU1131 forms a putative two-component system (a kinase and regulator pair, a common regulatory system in bacteria [42]). The most frequent mutation in PFLU1131 was an identical in-frame 15-bp deletion (Fig 2A) in the histidine-kinase phospho-acceptor domain (73%). This mutation results in loss of 5 amino acids (368-GEVAM-372) in the protein product (henceforth referred to as PFLU1131-del15). Other mutations included a highly similar 15-bp deletion resulting in loss of 5 amino acids a few amino acids downstream (369-EVAMG-373), as well as a single nucleotide polymorphism (SNP) resulting in A375V. All of these mutations (86% of the first-step mutations) cluster to the same 26 bp of the 1,770-bp PFLU1131 ORF and result in amino acid changes in a site directly adjacent to the catalytically active H-box [43] (amino acids 376–382; Fig 2A), suggesting a significant effect on the catalytic function of the putative kinase. We additionally constructed a PFLU1132 knockout in the ΔfleQΔntrC background to identify any further RpoN-EBPs capable of rewiring. ΔfleQΔntrCΔPFLU1132 failed to rescue flagellar motility within the 6-week assay cutoff (S2 File; n = 192), suggesting evolutionary rewiring of other transcription factors may not be easily achieved under the conditions we tested. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. In a ΔfleQΔntrC background, motility rescue initially occurs through of mutations in PFLU1131, which have a net up-regulatory effect on the transcriptome. (A) Diagram of PFLU1131 protein, indicated positions of first-step motility-rescuing mutations. Predicted protein domains (pfam) are indicated in lilac; amino acid positions are indicated as numbers below each domain. Mutations are indicated in blue boxes, with the number of replicate lines gaining each mutation indicated in brackets. (B) FleQ-homolog encoding gene PFLU1132 is essential for rescued flagellar motility in the PFLU1131-del15 mutant strain and depends on the presence of the kinase mutation. Scale bars (white) = 12.6 mm. Transcription factor gene PFLU1132 is deleted and reintroduced as a single-copy chromosomal insertion expressed from an L-rhamnose–inducible promoter system (rhaSR-PrhaBAD). The same complementation lacking the PFLU1131-del15 mutation as well as an empty expression system transposon were included as further controls. Photographs of motility after 1-day incubation in 0.25% agar LB plates supplemented with or without 0.15% L-rhamnose for induction of transcription factor expression. In all cases, the genetic background for these mutants and constructs is a ΔfleQΔntrC background. (C) Volcano plots indicating impact of PFLU1131-del15 mutation on the transcriptome relative to the ΔfleQ ancestor. Red points indicate significantly differentially expressed genes. Triangles indicate flagellar genes; squares indicate PFLU1131/2 and adjacent genes; and circles indicate all other genes. Vertical and horizontal dashed lines indicate significance cutoff values on both the x and y axes. Data underlying this figure can be found in S6 File. https://doi.org/10.1371/journal.pbio.3002348.g002 Together, these results suggest that in the absence of FleQ and NtrC, flagellar motility can be rescued via rewiring of an alternative transcription factor in the same protein family, PFLU1132. To confirm that observed motility phenotypes were dependent on the PFLU1131/2 two-component system, the PFLU1132 gene was deleted and complemented with and without the presence of the most common first-step kinase mutation (PFLU1131-del15). Knockout of PFLU1132 abolished flagellar motility, and complementation restored motility only in the presence of the kinase mutation (Fig 2B). This experiment was repeated for ntrBC and produced the same result (S1 Fig). The PFLU1131-del15 mutation also grants flagellar motility when ntrC is present (i.e., in a ΔfleQ background), so does not depend on the absence of ntrC to function (S2 Fig)—ruling out the possibility that the presence of NtrC suppresses the alternative pathway. Transcriptomic analysis of the PFLU1131-del15 mutant by RNA sequencing indicates an alteration to the activity of the PFLU1131/2 two-component system that results in a net up-regulatory effect on the transcriptome (Fig 2C). These regulatory changes include up-regulation of the flagellar genes (S3 Fig) and the PFLU1130/1/2 operon. In sum, in the “primary rewiring route,” repeatable mutations in the NtrBC (Fig 1B) two-component system were found to rescue flagellar motility. In the absence of both FleQ and NtrC, we identified the “alternative rewiring route,” via PFLU1131/2 (Fig 1C), which was also capable of rescuing flagellar motility.

First-step mutations in the alternative rewiring pathway rewire another FleQ homologous transcription factor In the primary rewiring pathway, NtrC (a homolog of FleQ) was recruited to recover lost FleQ function. This occurred as a two-step process: an initial motility-granting mutation to the gene ntrB encoding NtrC’s cognate kinase and a secondary motility-enhancing mutation to the helix-turn-helix (HTH) DNA binding domain of the NtrC transcription factor. Similarly, in the ΔfleQΔntrC background strain, first-step mutations restored slow motility, showing that alternative rewiring routes were possible, but not utilised in the presence of ntrC. In 13/15 first-step mutants, only a single mutation to PFLU1131 was identified, suggesting that only this mutation is required for flagellar motility to be rescued. To confirm this, the PFLU1131-del15 mutation was introduced into the ancestor (ΔfleQΔntrC). The resulting engineered strain was motile (S2 Fig). As no mutational change to the DNA binding domain of PFLU1132 is needed, this suggests a non-specific mechanism through which this transcription factor induces the flagellar genes. To investigate this, our transcriptomic data were used to assess the impact of the PFLU1131-del15 mutation on the expression of a list of genes [44] controlled by all 22 homologous transcription factors (part of the RpoN-EBP family) present in P. fluorescens SBW25 [45]. Regulatory changes predictably include up-regulation of the flagellar genes (S3 Fig) and the PFLU1130/1/2 operon but also include up-regulation of many other genes. In particular, we saw significant up-regulation for 54% of all RpoN-EBP controlled genes in the alternative rewiring pathway (Fig 3). For comparison, 70% of RpoN-EBP controlled genes were up-regulated in the primary rewiring pathway (via NtrC) [37]. Flagellar genes account for 12% of all RpoN-EBP controlled genes, and while the native regulatory targets of PFLU1132 are unknown, it is unlikely that it natively regulates 54% of these. Similarly, the first-step motile mutant (ΔfleQ ntrB-T97P) reported in Taylor and colleagues [37] leads to up-regulation of genes known to be involved in nitrogen assimilation (as expected as mutations are located in the Ntr pathway); however, it also results in up-regulation of many genes with no known regulatory link to NtrC. The large percentages of the RpoN regulon induced by these specific kinase mutations, in both the primary and alternative rewiring pathways, suggest first-step mutations confer a state of modified regulatory activity to PFLU1132 resulting in more frequent non-cognate interactions across many RpoN-EBP controlled genes including the flagellar genes. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. First-step mutations result in up-regulation of flagellar and other RpoN-EBP controlled genes. Heatmap of Log2Fold changes in gene expression for RpoN-EBP controlled genes after first-step motility mutations, relative to ΔfleQ ancestor. ΔfleQΔntrC background and PFLU1131-del15 mutant (a first-step mutant, also ΔfleQΔntrC background) differentially expressed genes were determined by RNA sequencing (this work). ΔfleQ first-step (NtrB-T97P) differentially expressed genes were identified by RNA microarray and originally reported by Taylor and colleagues [37]. List of RpoN-EBP controlled genes was obtained from Jones and colleagues [44]. Groups of genes by function are indicated by the coloured bars above the plot: red: flagellum; blue: nitrogen; teal: other/unknown. Data underlying this figure can be found in S7 File. https://doi.org/10.1371/journal.pbio.3002348.g003

Evolutionary rescue of motility through alternative rewiring pathway is significantly constrained If alternative rewiring routes are available to natural selection—why have we not seen them utilised in the presence of the primary rewiring route? In the alternative rewiring pathway, we measured the time taken for swimming mutants to evolve and the strength of the evolved phenotype in comparison to the primary rewiring route. Time to emergence (the length of time taken for motility to evolve on the soft agar plate) was recorded for all replicate experimental lines of the ΔfleQΔntrC ancestor, as well as ΔfleQ as a comparison. Strikingly, only 9.4% of replicate experimental lines (n = 160) of ΔfleQΔntrC evolved within the 6-week assay cutoff compared to 100% for the ΔfleQ background (n = 22). Initial first-step motile ΔfleQΔntrC strains also evolved within an average of 18.7 days from the assay start (Fig 4A), significantly longer than ΔfleQ took to rescue motility (4.2 days, P = 0.002, Dunn test). After first-step PFLU1131 mutations, second-step mutants evolved rapidly in ΔfleQΔntrC lines within an average of 2.2 days (Fig 4A). This was faster than the second-step mutants in the ΔfleQ background, which took 3.3 days to emerge (P = 0.0017, Dunn test). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. Motility rescue through the alternative pathway is significantly slower and provides a poorer motility phenotype that displays lower pleiotropic costs. (A) Time to emergence of motile zone (days) for first- and second-step mutants in the ΔfleQΔntrC and ΔfleQ backgrounds. Percentages beneath each plot indicate proportion of replicate evolving that step within the 6-week assay cutoff. Significant differences in mean time to emergence are indicated as follows: P < 0.0001 = ****, P < 0.001 = ***, Ns. = p > 0.05 (Dunn test). (B) Race assay as measure of motility fitness. Distance moved over 24 hours in 0.25% agar LB plates measured relative to the ΔfleQ ntrB-T97P mutant. P value indicated is generated from a two-sample Wilcox test. (C) Fitness in LB broth measured as area under the 24-hour growth curve (AUGC) relative to ΔfleQ ntrB-T97P. P values above plots generated from Dunn tests. AUGC values and growth curve plots are provided in S4 Fig. In parts B and C, PFLU1131-del15 is present in a ΔfleQΔntrC genetic background. For all boxplots: box represents first to third quartile range; middle line represents median value; whiskers range from quartiles to maxima and minima. Data underlying parts A, B, and C can be found in S8, S9, and S10 Files, respectively. https://doi.org/10.1371/journal.pbio.3002348.g004 The greatly increased time and low frequency for rescue motility in an ΔfleQΔntrC background may be reflective of a small pool of accessible PFLU1131 mutations that can trigger rewiring of PFLU1132, perhaps due to a small mutational target size. One replicate line gained a de novo mutation in the DNA mismatch-repair gene mutS (S1 File), resulting in a frameshift and probable loss of function [46]. This strain, along with strains derived from it, possessed large numbers of additional SNPs (>80) including an SNP in PFLU1131. Loss of function to mutS is known to enhance mutability in Pseudomonads [47], and its occurrence may increase access to motility-rescuing mutations in PFLU1131. From a phenotypic perspective, the first-step mutations in the PFLU1131 pathway in a ΔfleQΔntrC strain provide a far poorer motility phenotype than the analogous first-step mutations in a ΔfleQ background (Fig 4B). In a race assay, PFLU1131-del15 mutants swam 0.31 mm for every 1 mm swam by the most common first-step mutant in the primary rewiring route (P = 0.000356, Wilcox test). One possible explanation for this poor motility is a lack of significant up-regulation for the flagellar filament subunit FliC (S3 Fig) in the PFLU1131-del15 mutant. The PFLU1131-del15 mutant additionally does not confer a significant defect to growth in shaking LB broth compared to its ancestral strain (Fig 4C; average relative area under the growth curve (rAUGC) of 1.55 and 1.48, respectively; P = 0.1132, Dunn test) and grew significantly better than the most common first-step mutant in the primary rewiring route (rAUGC of 1; P = 0.0081, Dunn test). The poor motility phenotype and the lack of significant fitness cost associated with PFLU1131-del15 likely indicates that the PFLU1131/2 system is a weaker activator of the flagellar motility that incurs less severe pleiotropy compared to NtrBC. These findings indicate that the alternative pathway is the far poorer option for rescuing flagellar motility and flagellar gene expression compared to the primary NtrBC pathway—motility rescue in the ΔfleQΔntrC background took significantly longer to occur, was significantly less frequent, and provides a significantly poorer flagellar motility phenotype. This solution also conferred a lower pleiotropic fitness cost, which in of itself may provide benefit but likely reflects to low flagellar expression and weaker rewiring. However, it is not clear why this should be the case—both NtrC and PFLU1132 are FleQ-homologs and are gaining mutations to their cognate kinases that facilitate rewiring. This suggests that there are other factors that constrain the evolutionary innovation of PFLU1132.

Motility enhancing second-step mutations suggest that alterations to global gene regulatory network can facilitate rewiring in the alternative pathway Following the first-step mutations in PFLU1131 that unlock flagellar motility, our motile isolates develop second-step mutations that act to boost motility speed (S5A Fig), often accompanied by significant pleiotropic fitness costs for growth in LB broth (S5B Fig). These second-step mutations offer clues to the nature of constraining factors limiting innovation through rewiring of PFLU1132, as they represent evolutionary solutions to the poor motility provided to the first-step mutations. Whole genome resequencing identified a diverse set of motility enhancing second-step mutations occurring at both a local (PFLU1131/2 locus) and a global regulatory scale (Fig 5A; full details of all mutations provided in S1 File, second step n = 18 (several first step lineages generates multiple second-step zones)). In contrast to the frequent second-step mutations observed in the HTH DNA-binding domain of NtrC [37], no analogous mutations were observed in the same domain of PFLU1132. Mutations that did occur in the PFLU1132 gene were parallel and identified in only 2 lines, impacting the receiver domain of the transcription factor, which may further modify activity of the PFLU1132 regulator by altering the interactions with its kinase. One line gained a secondary PFLU1131 mutation, an SNP in the HATPase domain along with the first-step SNP A547D already present (Fig 2B), which may further boost kinase interactions. A mutation to the promoter region of the PFLU1131/2 operon was also observed. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 5. Diverse second-step mutations indicate multiple global regulatory strategies for enhancing rewiring through PFLU1132. (A) Diagram of mutational distribution across the SBW25 chromosome for first- and second-step motility mutants. Each ring represents the chromosome, with black dots indicating non-synonymous mutations. The size of each dot is proportional to the number of mutations occurring in that locus in independent replicates. Mutational target names are coloured by functional category: red: PFLU1131/2; purple: PFLU1583/4; blue: metabolic; green: global regulatory; grey: other. (B) Principal component analysis of transcriptomic data for ancestral, first-step, and second-step motility mutants. (C and D) Volcano plots of differentially expressed genes for second-step mutations relative to ΔfleQ ancestor (mutants are themselves ΔfleQΔntrC backgrounds). Red points are significantly differentially expressed. Triangles indicate flagellar genes; squares indicate PFLU1131/2 and adjacent genes; and circles indicate all other genes. Vertical and horizontal dashed lines indicate significance cutoff values on both the x and y axes. (E) Log2Fold change in gene expression for RpoN-EBP controlled genes after PFLU1583 Δ48–74 mutation—relative to the first-step PFLU1131-del15 mutant (mutants are themselves ΔfleQΔntrC backgrounds). RpoN-EBP controlled genes grouped by function are indicated by the coloured bars above the plot: red: flagellum; blue: nitrogen; teal: other/unknown. Data underlying parts B, C, and D can be found in S6 File, and part E in S11 File. https://doi.org/10.1371/journal.pbio.3002348.g005 Other second-step mutations can be grouped into 3 other broad categories. (i) The first are mutations in the operon PFLU1583/4, accounting for 24% of second-step mutations (Fig 5A, purple). This gene pair encodes a putative anti-sigma factor PP2C-like phosphatase and a putative STAS family anti-sigma factor antagonist (PFLU1583 and PFLU1584, respectively). Mutations in PFLU1583 were generally loss of function suggesting loss of repression on an unknown sigma factor. This may act on RpoN, the partner sigma factor of PFLU1132, in regulating gene expression; however, this cannot be tested easily through rpoN knockout, and P. fluorescens SBW25 encodes approximately 31 other putative sigma factors [45], which may instead be the mechanistic targets of this mutation. Transcriptomic analysis of a PFLU1583 mutant (PFLU1583Δ48–74) identified sigma factor RpoE as up-regulated; however, rpoE knockout did not negate the motility enhancing effect of the mutation (S6A Fig). To test that PFLU1583/4 are not acting to enhance motility through another RpoN-EBP, we engineered PFLU1583Δ48–74 in a ΔfleQΔntrC background in the absence of a PFLU1131 mutation. This strain was immotile, indicating that these anti-sigma factor mutations depend on the function of the PFLU1131/2 system to confer a swimming phenotype (S6B Fig). (ii) The second category of mutations impact core gene expression components that will affect global GRN function (Fig 5A, green). These included mutations to rho, rpsK, and rpoC that encode core gene expression machinery and will impact expression of most genes in the genome. (iii) The final category of motility enhancing mutations fit a metabolic theme (Fig 5A, blue). These mutations occur in genes involved in central carbon metabolism, which will likely have a significant impact on global gene expression through alterations to central carbon flux, likely incurring pleiotropic fitness trade-offs [48], and may act on RpoN indirectly—for example, via core carbon metabolic regulator CbrB (an RpoN-EBP)—or signal to the PFLU1131 sensor kinase via internal metabolic flux. The diversity and nature of these second-step mutations suggest global regulatory strategies to facilitate flagellar gene expression through the alternative rewiring pathway. To understand their regulatory impact, transcriptome analysis was performed on a pair of representative mutants: the most common second-step mutant (anti-sigma factor mutant, PFLU1583 Δ48–74), which represented a mutation with global regulatory effects, and the PFLU1131/2 promoter mutation as a representative of mutation with predicted “local” regulatory effects. Together, these mutations represent the 2 loci that constitute 56% of identified second-step mutations. Principal component analysis indicates that the transcriptomic profile of the anti-sigma factor mutant differs significantly from the profile of the PFLU1131/2 promoter mutation. Both mutations resulted in similar variation across PC1 relative to the first-step PFLU1131-del15 mutant, but opposite variation in PC2 (Fig 5B). Both the PFLU1131/2 promoter and anti-sigma factor mutations result in net up-regulatory effects on the transcriptome reflected by positive skewed volcano plots (Fig 5C and 5D) albeit with differing expression patterns. Both mutations have the effect of further up-regulating RpoN-EBP controlled genes with 22% and 60% being up-regulated for the anti-sigma factor mutant and the PFLU1131/2 promoter mutant, respectively (Fig 5E). In general, for the alternative rewiring pathway, genomic and transcriptomic analysis of second-step mutations indicate enhancement of rewiring through the PFLU1132 transcription factor, albeit utilising differing (local and global) mechanisms. These results highlight that diverse mutations with both targeted and global regulatory effects can help facilitate transcription factor rewiring.

Second-step promoter capture mutation suggests importance of transcription factor expression for rewiring Although second-step mutations were diverse in the alternative rewiring pathway, one promoter capture event in particular provides evidence for the role of increased expression of the rewired transcription factor in strengthening the motility phenotype. To gain capacity to drive expression of non-cognate genes, the PFLU1132 transcription factor will need to first saturate its native regulatory interactions before it can engage in low-affinity interactions with FleQ-controlled genes. High expression of active transcription factor can provide these conditions, elevating the concentration of transcription factor in the cell to permit greater expression of flagellar genes [49,50]. While many of the second-step mutations detailed above may impact PFLU1131/2 expression indirectly (RpoC, Rho), the PFLU1131/2 promoter mutation directly effects expression of the two-component system. This mutation is a 1.59-kbp deletion resulting in total loss of the PFLU1130 gene, and in PFLU1131/2 becoming part of the PFLU1127/8/9 operon (Fig 6A). Upstream of this new combined operon sits a predicted RpoN binding site [44], so this deletion positions PFLU1132 in a new operon under the control of RpoN, which may create a positive feedback loop where this RpoN-dependent regulator drives its own expression—either through rewired regulation or native control of this promoter. This genetic rearrangement significantly up-regulates both PFLU1131 and PFLU1132 with 4.26 and 3.79 Log2Fold increases in expression compared to the non-motile ΔfleQ strain, respectively (Fig 6B, raw values for this and other tested strains provided, S5 File). Up-regulation of this two-component system corresponds with the 60% increase in expression of RpoN-EBP controlled genes discussed above (Fig 5E), indicating an increase in the strength of PFLU1132 rewiring—inducing greater flagellar gene expression (S3 Fig). The 1.59-kbp deletion results in loss of an intergenic region that typically separates the 2 combined operons. This region is predicted to contain a rho-dependent terminator (S3 File), which may typically prevent transcriptional readthrough into PFLU1130/1/2. Such a terminator will restrict PFLU1131/2 expression via readthrough and constrain its ability to achieve high concentrations that may facilitate rewiring. The rho mutation observed in another second-step motile strain may also act to increase readthrough at this site. In comparison to this promoter mutation, the other second-step mutation tested (anti-sigma factor PFLU1583 Δ48–74) does not significantly impact PFLU1131/2 expression (Fig 6B), so this mutation likely influences rewiring of the PFLU1132 transcription factor through a separate mechanism. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 6. Second-step PFLU1131/2 promoter mutation increases PFLU1131/2 expression. (A) PFLU1130/1/2 genetic locus in the SBW25 chromosome, and impact of the 1.59-kbp deletion identified in 1 second-step motile mutant. Location of the RpoN binding site upstream of PFLU1127/8/9 was predicted by Jones and colleagues [44]. The presence of the rho-dependent terminator was determined by use of RhoTermPredict [80]. (B) Log2Fold change in gene expression (relative to ΔfleQ ancestor) heatmap of PFLU1130/1/2 and nearby RpoN-EBP controlled genes, including PFLU1127/8/9 in the ancestral ΔfleQΔntrC strain, first- and second-step motility mutants. Data underlying part B of this figure can be found in S12 File. https://doi.org/10.1371/journal.pbio.3002348.g006 In summary, expression of PFLU1131/2 appears to be a major factor constraining rewiring of this regulatory system and rescue of flagellar motility, and overcoming this constraint facilitates evolutionary innovation through this alternative rewiring route.

Increasing transcription factor expression in motile mutants results in faster motility phenotypes To test the importance of transcription factor gene expression for rewiring, we made use of the complementation strains presented in Figs 2C and S1. These experiments produced strains where the primary (ntrC) and alternative (PFLU1132) rewired transcription factor genes are deleted from their native loci and reintroduced on an L-rhamnose titratable promoter system. This was done in the presence of their respective kinase mutations that conferred first-step slow spreading motility in each rewiring pathway. Concentration of L-rhamnose added to the media positively correlated with distance moved (Fig 7A) for both the primary rewiring pathway (i.e., NtrC expression system; ρ = 0.984, Spearman test) and the alternative rewiring pathway (i.e., PFLU1132 expression system; ρ = 0.865, Spearman test). We confirmed increasing L-rhamnose concentration increased expression from the L-rhamnose titratable system (RhaSR-PrhaBAD) in our bacterial strain backgrounds by testing LacZ activity with increasing L-rhamnose concentration for lacZ under control of the L-rhamnose titratable promoter construct in the ΔfleQ and ΔfleQΔntrC backgrounds. In both cases, LacZ activity positively correlated with L-rhamnose concentration (ρ = 0.981 and ρ = 0.955, respectively, Spearman tests), indicating that increasing L-rhamnose concentration results in increasing expression of the gene introduced in the rhaSR-PrhaBAD expression system (S7 Fig). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 7. Impact of RpoN-EBP gene expression on motility speed. (A) Mean motility speed (mm relative to ΔfleQ ntrB-T97P) plotted against increasing L-Rhamnose concentration (%w/v). Mean of 6 biological replicates for each point. First-step motility mutants ΔfleQ ntrB-T97P and ΔfleQΔntrC PFLU1131-del15 had their respective RpoN-EBPs knockout out and reintroduced on a miniTn7 transposon with the RpoN-EBP under control of the PrhaBAD and rhaSR rhamnose-inducible expression system. In full, these were ΔfleQ ntrB-T97P ΔntrC miniTn7[rhaSR-PrhaBAD-stRBS-ntrC] (rha-ntrC) and ΔfleQΔntrC PFLU1131-del15 ΔPFLU1132 miniTn7[rhaSR-PrhaBAD-stRBS -PFLU1132] (rha-1132). (B) Motility speed (mm relative to ΔfleQ ntrB-T97P) of common pathway ntrB-T97P first-step mutant, PFLU1131-del15 first-step and PFLU1131/2 promoter second-step mutants. Whiskers represent standard deviation above and below the mean value. Data underlying parts A and B of this figure can be found in S13 and S14 Files, respectively. https://doi.org/10.1371/journal.pbio.3002348.g007 This allows us to conclude that increasing expression of ntrC and PFLU1132 in the presence of their first-step kinase mutations granted stronger motility phenotypes, facilitating their rewiring. The expression level of a transcription factor can therefore significantly impact its propensity for rewiring, and more highly expressed transcription factor may be expected to engage in non-canonical regulatory interactions more readily [51].

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3002348

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/

You are viewing proxied material from magical.fish. The copyright of proxied material belongs to its original authors. Any comments or complaints in relation to proxied material should be directed to the original authors of the content concerned. Please see the disclaimer for more details.