Exploratory Growth in Streptomyces venezuelae Involves a Unique Transcriptional Program, Enhanced Oxidative Stress Response, and Profound Acceleration in Response to Glycerol

ABSTRACT

Exploration is a recently discovered mode of growth and behavior exhibited by some Streptomyces species that is distinct from their classical sporulating life cycle. While much has been uncovered regarding initiating environmental conditions and phenotypic outcomes of exploratory growth, how this process is coordinated at a genetic level remains unclear. We used RNA sequencing to survey global changes in the transcriptional profile of exploring cultures over time in the model organism Streptomyces venezuelae. Transcriptomic analyses revealed widespread changes in gene expression impacting diverse cellular functions. Investigations into differentially expressed regulatory elements revealed specific groups of regulatory factors to be impacted, including the expression of several extracytoplasmic function (ECF) sigma factors, second messenger signaling pathways, and members of the whiB-like (wbl) family of transcription factors. Dramatic changes were observed among primary metabolic pathways, especially among respiration-associated genes and the oxidative stress response; enzyme assays confirmed that exploring cultures exhibit an enhanced oxidative stress response compared with classically growing cultures. Changes in the expression of the glycerol catabolic genes in S. venezuelae led to the discovery that glycerol supplementation of the growth medium promotes a dramatic acceleration of exploration. This effect appears to be unique to glycerol as an alternative carbon source, and this response is broadly conserved across other exploration-competent species.

IMPORTANCE Exploration represents an alternative growth strategy for Streptomyces bacteria and is initiated in response to other microbes or specific environmental conditions. Here, we show that entry into exploration involves comprehensive transcriptional reprogramming, with an emphasis on changes in primary metabolism and regulatory/signaling functions. Intriguingly, a number of transcription factor classes were downregulated upon entry into exploration. In contrast, respiration-associated genes were strongly induced, and this was accompanied by an enhanced oxidative stress response. Notably, our transcriptional analyses suggested that glycerol may play a role in exploration, and we found that glycerol supplementation dramatically enhanced the exploration response in many streptomycetes. This work sheds new light on the regulatory and metabolic cues that influence a fascinating new microbial behavior.

INTRODUCTION

The soil is a highly heterogeneous environment. Essential growth factors like oxygen, carbon, and iron can be present at widely varying levels (1–4), and these nutrient levels can be shaped by biotic (e.g., microbial population density) and abiotic (e.g., water saturation) factors, which are themselves dynamic. Bacterial survival in this complex, fluctuating environment requires both genetic and metabolic flexibility and the ability to effectively compete or cooperate with neighboring microbes.

In the soil, Streptomyces is a genus of ubiquitous Gram-positive bacteria that epitomize both metabolic and developmental adaptability. As saprophytic organisms, Streptomyces species harbor a vast repertoire of catabolic enzymes that allow them to make use of a diverse set of carbon sources, including monomeric carbon sources such as glycerol, mannitol, arabinose, and galactose, through to complex polysaccharides like chitin and cellulose (5). Similarly, the genomes of different Streptomyces species have revealed remarkable metabolic capabilities in the form of abundant biosynthetic gene clusters that direct the production of diverse specialized metabolites, often with little overlap between species (6). These specialized metabolites are often small molecules with potent biological activity and include compounds with the ability to inhibit the growth of bacteria and fungi (e.g., antibiotics) (7, 8), promote iron acquisition (e.g., siderophores) (9), and influence community behaviors (e.g., quorum sensing molecules) (10, 11). This rich library of chemical effectors is proposed to equip the streptomycetes with a powerful repertoire of tools that facilitate either competitive or cooperative interactions with other microbes.

Complementing the robust metabolic capabilities of the streptomycetes is a remarkable developmental flexibility. The classical Streptomyces life cycle begins with spore germination and proceeds through the growth of branching filamentous hyphae, establishing a metabolically active vegetative mycelium. Nutrient depletion or other stressors induce a coordinated switch to reproductive growth, which initiates with the raising of aerial hyphae (12). Further maturation events promote the differentiation of these aerial structures into chains of metabolically dormant spores. These spores are resistant to many abiotic stresses and can be released from their chains and dispersed to new environments (13).

Recently, it was shown that some Streptomyces species can abandon their canonical, classical life cycle when grown on solid surfaces and enter an alternative growth mode called “exploration.” Exploratory growth is phenotypically distinct from classical growth: colonies expand rapidly outward as a vegetative-like mycelium and develop a dense network of wrinkles at the center of the biomass (14). In Streptomyces venezuelae, exploration is initiated in response to specific environmental cues. Stimulation of exploration was originally identified as a competitive response by S. venezuelae to coculture with the yeast Saccharomyces cerevisiae on rich, glucose-containing medium (yeast extract-peptone-dextrose [YPD]) (14). An analogous response can be initiated in the absence of yeast when the same rich medium lacks glucose (yeast extract-peptone [YP]), highlighting a strong repressive effect of glucose on exploration (14).

Concurrent with morphological development, exploring colonies reshape the physicochemical properties of their surrounding environment through the emission of trimethylamine, a small basic volatile compound (14). The airborne diffusion of trimethylamine away from the colony promotes a steady rise in the local pH, which in turn confers a competitive advantage to exploring Streptomyces species. It promotes the exploration of nearby S. venezuelae cultures growing on rich, glucose-containing medium (typically an exploration-repressive condition) and simultaneously lowers the bioavailability of iron by forming poorly soluble hydroxides (15). This self-imposed iron limitation both strengthens the response of exploring S. venezuelae cultures and inhibits the growth of other microbes in the vicinity.

While the environmental triggers of exploration initiation and propagation are becoming clearer, there is much that remains to be understood about this unusual microbial behavior and the factors that influence it. Exploration in Streptomyces appears to be largely independent of regulators that control classical development (14). This raises an interesting possibility that exploration is encoded as an independent genetic program, with both defined regulatory circuits and distinct metabolic pathways that may be activated or downregulated over time. To better understand the pathways that drive Streptomyces exploration, we examined gene expression changes over time in S. venezuelae exploring cultures using RNA sequencing (RNA-seq). We observed the global reprogramming of transcription during the maturation of an exploring colony, with a notable upregulation of genes associated with primary metabolism, respiration, and the oxidative stress response. A number of regulatory elements belonging to different functional categories also exhibited changes in expression over time, including known regulators of classical development, extracytoplasmic function (ECF) sigma factors, and second messenger systems. We further discovered that glycerol supplementation of the growth medium radically enhanced the exploration response of S. venezuelae and altered the metabolic output of these cells. This phenomenon was unique to glycerol, as a similar response could not be induced by supplementation with other alternative carbon sources. Conversely, the response to glycerol was not unique to S. venezuelae, as multiple exploration-competent wild Streptomyces isolates exhibited analogous enhanced exploratory growth in the presence of glycerol.

RESULTS

Gene expression is globally reprogrammed during exploration on YP medium.

We previously observed that S. venezuelae exploration occurs independently of many of the key regulators of classical development (i.e., the bld and whi genes), suggesting that a different suite of signaling pathways may function to coordinate the switch into exploration. To gain insight into the genes that are activated and repressed during S. venezuelae exploration, we performed RNA sequencing on transcripts isolated from S. venezuelae colonies growing on exploration-promoting medium (YP) over time. Representative time points were selected based on colony morphology to capture each phase of the exploration process: early (colony establishment), mid (establishment of a “core” region and start of exploration), and late (robust exploration) (Fig. 1A). Analysis of gene expression data between the early and late time points revealed over 1,000 genes that were differentially expressed and distributed across the chromosome (Fig. 1B). Reducing the complexity within the data to Clusters of Orthologous Groups (COGs), we saw a high frequency of differential expression for genes participating in primary metabolism (COG terms carbohydrate transport and metabolism, amino acid transport and metabolism, and inorganic ion transport and metabolism) (Fig. 1C). When identifying target genes for follow-up validation studies, we first focused on operons that showed consistent trends in expression over time and genes that participated in similar pathways.

FIG 1.

RNA sequencing of cultures exploring on YP over time. (A) Representative colonies chosen as time points for RNA isolation and sequencing. Early, mid, and late cultures were grown for 2, 5, and 9 days, respectively. (B) Differential gene expression over time in YP exploring cultures. The fold changes in the expression of each gene between the early and late time points were plotted, and these were arranged by their position on the S. venezuelae chromosome. Data points that are statistically significant are shown in red (P < 0.01). (C) Genes that showed statistically significant (P < 0.01) differential expression between early and late time points were assigned Clusters of Orthologous Groups (COG) identifiers, and their distributions were plotted. A key for the corresponding COG terms is presented to the right of the graph.

Among the most statistically significant differentially expressed genes were those involved in inorganic nitrogen (highest transcript levels early) and sulfur (highest transcript levels late) metabolism (Fig. 2A). Therefore, we tested the exploration response when the concentrations of these nutrients were altered under exploration-promoting conditions. We supplemented YP with nitrate, sulfite, and sulfate at a range of concentrations (10 μM to 1 mM) and found that in all instances, the effect on S. venezuelae exploration was negligible (Fig. 2B). From this, we concluded that nitrogen and sulfur acquisition were likely important but that sufficient concentrations of these nutrients could be obtained from the growth medium such that additional supplementation had no effect.

FIG 2.

Effects of inorganic nitrogen and sulfur during exploration on YP. (A) Gene expression levels in YP over time (early [2 days], mid [5 days], and late [9 days]) for two regions of the S. venezuelae chromosome with predicted function in inorganic nitrogen (vnz_14960-vnz_14975) (top) and sulfur (vnz_29235-vnz_29270) (bottom) metabolism. The gene organization for each region is depicted above each graph, with vnz gene identification numbers indicated above and gene annotations indicated below each arrow. PFLA, pyruvate-formate lyase-activating enzyme; PFL pyruvate-formate lyase; NT, nitrate/nitrite transporter; NR, nitrite reductase; ST, sulfate transporter; SA, sulfate adenylyltransferase; ASK, adenylyl-sulfate kinase; PAPSR, phosphoadenosine phosphosulfate reductase; SR, sulfite reductase. (B) Effects of nitrogen and sulfur supplementation on exploration. S. venezuelae was spotted onto YP medium supplemented with various concentrations of nitrate, sulfite, and sulfate and imaged after 7 days of growth.

Exploration on YP is associated with strong respiration and oxidative stress responses.

Analysis of our RNA sequencing data further suggested that energy production and respiration pathways were similarly induced over the course of exploration. We saw strong transcriptional upregulation of genes encoding ATP synthase subunits, alongside genes encoding iron-sulfur cluster assembly proteins, which produce essential cofactors responsible for conducting electrons through the supercomplexes of the electron transport chain (Fig. 3A and B). Accordingly, as the respiratory chain is a noted source of intracellular reactive oxygen species (16), we saw a corresponding transcriptional activation of the genes encoding the oxidative stress-responsive catalases and superoxide dismutases (Fig. 3C).

FIG 3.

Upregulation of respiration-associated and oxidative stress response genes during exploration on YP. (A) Expression over time for genes participating in ATP synthesis (vnz_24810-vnz_24845). (B) Expression over time for genes participating in iron-sulfur cluster biogenesis (vnz_07605-vnz_07635). For panels A and B, the gene organization for each region is depicted above each graph, and vnz gene identification numbers are given above and gene annotations are given below each arrow. AS, ATP synthase subunit; MSC, metal-sulfur cluster biosynthetic enzyme; SUF, sulfur assimilation system iron-sulfur cluster assembly protein; CD, cysteine desulfurase; RP, Rieske (2Fe-2S) protein. (C) Gene expression levels in YP-grown cultures over time for predicted catalase and superoxide dismutase (SOD) genes carried by S. venezuelae. (D, top) Catalase enzyme activity was assayed by native PAGE using proteins from cultures harvested at early and late time points from exploring cells grown on YP, together with extracts from classical vegetatively grown cultures. (Bottom) Coomassie-stained SDS-PAGE gels of the same samples served as a loading control for the total protein content. In all panels, early is 2 days, mid is 5 days, and late is 9 days.

To determine whether the increased transcription of these oxidative stress-protective enzyme-encoding genes was correlated with increased enzyme activity, we assessed catalase activity in exploring cultures over time. Native PAGE was used to resolve distinct catalase species from whole-cell lysates harvested from early- and late-stage YP exploring cultures, alongside a vegetative (nonexploring) growth control. We found that at our early time point (before exploration had initiated), activity levels were roughly equivalent to those under classical, vegetative growth conditions (Fig. 3D). However, at the late time point (during robust exploratory growth), there was a pronounced increase in catalase activity, suggesting that these cells had an enhanced capacity to manage reactive oxygen species compared with early and vegetatively grown cultures.

To assess whether increased catalase activity was a prerequisite for robust exploration, we generated strains of S. venezuelae in which the most highly expressed catalase gene in our transcriptional data set (vnz_36165) was either deleted or expressed from a highly active, constitutive promoter. When these strains were grown in parallel with their equivalent wild-type control strain on YP, there were subtle differences in colony architecture compared with that of the wild type; however, the overall rate of exploration was not significantly impacted for either the deletion or the overexpression strain (see Fig. S1 in the supplemental material).

Multiple mechanisms of transcriptional control are differentially expressed during exploration.

To identify regulatory elements that may function in coordinating the exploration response, we extracted differentially expressed genes from the COG analysis (Fig. 1C) that had been assigned to the terms “transcription” or “signal transduction mechanisms” (Table S1). As the RNA sequencing data had suggested that exploratory growth was concomitant with a broad reprogramming of gene expression, we initially prioritized regulators that we expected to function globally in the cell rather than modulating pathway-specific processes. To this end, we identified multiple functional categories of regulators that could fulfill this role, including well-characterized transcription factors that participate in classical development, extracytoplasmic function (ECF) sigma factors, and second messenger-responsive proteins (Table 1; Table S1).

TABLE 1.

Global regulators that are differentially expressed over the course of exploration

Product type and gene	Log2 fold change (early/late)	P value
Regulators of classical development
bldM (vnz_22005)	1.12	2.76 × 10−3
bldN (vnz_15655)	2.50	1.01 × 10−4
whiH (vnz_27205)	2.70	1.03 × 10−11
WhiB-like proteins
whiB (vnz_13645)	1.37	1.29 × 10−6
wblA (vnz_16495)	3.56	2.35 × 10−36
wblE (vnz_24255)	1.20	1.68 × 10−4
whiB-like (vnz_00590)	−2.49	1.15 × 10−4
ECF sigma factors
sigE (vnz_15840)	3.30	7.46 × 10−37
sigQ (vnz_22610)	3.02	1.82 × 10−13
Second messenger-responsive proteins
eshA (vnz_35035)	5.82	1.14 × 10−33

As the majority of these regulators (except vnz_00590) were more highly expressed early in the exploration process, with transcript levels dropping as exploration progressed, we wondered whether any of these systems were important for relaying signals needed to initiate exploration within the cell. In the set of regulators that overlapped those coordinating classical development, we identified bldM, bldN, and whiH. BldN is an alternative sigma factor that directs the expression of multiple development-specific genes, including bldM, which encodes a transcription factor that functions both alone and in conjunction with WhiI to control the transcription of genes needed for reproductive growth (17, 18). We know from previous work that mutants defective in bldN and, to a lesser extent, bldM are capable of exploring but do so at a reduced rate (14). WhiH acts at a later stage in classical development and is important for spore formation (12). The deletion of its associated gene has previously been shown to have a negligible impact on the progression of exploration (14).

The WhiB-like (Wbl) proteins comprise a subcategory of classical development regulators. This family of typically short proteins (<150 amino acids) is specific to the Actinobacteria, and their associated genes are often found in multiple copies (e.g., 11 paralogs are encoded within the S. venezuelae chromosome). While the functions of many of these proteins remain unknown, three members (WhiB, WhiD, and WblA) have been shown in different Streptomyces species to act as transcription factors that coordinate distinct stages of reproductive growth (19). A defining feature shared by Wbl proteins is the conserved presence of a [4Fe-4S] cluster that is necessary for protein function and is sensitive to decomposition by both O₂ and NO (19). Although we noted strong upregulation of protein-coding genes functioning in iron-sulfur cluster biosynthesis during exploration (Fig. 3), deleting the founding member of the Wbl family, whiB, still allowed robust exploration. Notably, known targets of WhiB (including ftsZ and ftsW) were not found to be differentially expressed over the course of exploration.

The third set of global regulators that we identified in our differential expression analyses was the extracytoplasmic sigma factors sigE and sigQ. As their name implies, these sigma factors are commonly activated upon sensing an environmental stimulus. Often, ECF sigma factors are regulated posttranslationally by anti-sigma factors. Curiously, SigE and SigQ are two exceptions; instead, they are under the transcriptional control of the genome-adjacent two-component systems cse(A)BC and afsQ1-afsQ2, respectively (20–23). The SigE regulon has recently been defined in Streptomyces coelicolor by chromatin immunoprecipitation; many of its target genes contribute to cell envelope function and integrity (24). Much less is known about SigQ; it is broadly conserved among Streptomyces species, and mutational studies in S. coelicolor have implicated it in both development and antibiotic production (22, 25). Given that less was known about SigQ, and given its conserved nature, we were interested in probing its role in exploration. We created both sigQ deletion and overexpression strains and compared their growth to that of a wild-type strain on YP medium. Robust exploration was observed for all strains (Fig. 4A). Subtle differences were noted for the sigQ overexpression strain relative to the wild type, including a reduced rate of expansion coupled with more wrinkling and apparent biomass formation in the core region of the colony.

FIG 4.

Contribution of differentially expressed global regulators to exploration. Mutant strains of the global regulators sigQ (A) and eshA (B) were spotted onto YP medium, alongside wild-type S. venezuelae, and assessed for exploratory growth. Images were taken after 9 days of growth.

The final category of regulators identified as being differentially expressed was the second messenger-binding proteins. EshA is a conserved cAMP-binding protein that has been implicated in different processes depending on the species. In S. coelicolor, eshA deletion results in the reduced production of the antibiotic actinorhodin, while development is unaffected (26). Conversely, in Streptomyces griseus, eshA deletion abolishes aerial mycelium formation and reduces the production of the antibiotic streptomycin (27). eshA has been shown to be controlled by BldM and cotranscribed with two 2-methylisoborneol synthase-encoding genes (mibA and mibB) (13). Consistent with these observations, both mibA and mibB were also downregulated during exploration (Table S1). As eshA exhibited one of the largest fold changes in expression over time during exploration and has previously been tied to developmental transitions, we wondered if EshA may also have a regulatory role in exploring cultures. We found that an eshA mutant was indistinguishable from the wild type with respect to its exploration capabilities (Fig. 4B).

Glycerol profoundly affects exploration.

We had previously established that the carbon source (glucose) had a strong effect on exploration. When expanding our analysis of transcriptional regulators to include those that had more pathway-specific functions, we identified significant differential expression for the regulator of glycerol catabolism, gylR, with similar observed trends for its downstream regulon members. The gyl region consisted of an operon of three genes (gylFKD) that encoded a permease, a glycerol kinase, and a glycerol-3-phosphate dehydrogenase, respectively (Fig. 5A); the expression levels of gylR and the gylFKD operon were highest at the earliest time point, and expression rapidly tapered off as exploration proceeded (Fig. 5B). We wondered how cells would respond if their growth medium was supplemented with glycerol and if this might delay the onset of exploration. To test this, we compared the growth of wild-type S. venezuelae on YP and YP supplemented with glycerol (YPG). Unexpectedly, the addition of glycerol dramatically accelerated S. venezuelae exploration (Fig. 5C). Compared to their YP-grown counterparts, YPG explorers grew rapidly (Fig. 5D; Video S1); colonies initially spotted at the center of a standard 10-cm petri plate had expanded to cover the entire plate within 7 to 9 days of growth; such a colony size was not typically observed for YP explorers, even with extended incubation times. YPG-grown colonies also developed a distinct colony morphology compared to that of YP-grown cultures, having a much more intricate network of wrinkles extending throughout the colony. Growth on YPG further yielded significantly more condensation within the plate lids than with YP-grown cultures, suggesting enhanced respiration rates (Video S1). Beyond the differences in exploration growth rates and colony architectures, we also found that YPG-grown colonies began secreting a bright orange pigment into the underlying medium after approximately 5 days of growth (Fig. 5E).

FIG 5.

Effect of glycerol on exploration. (A) Organization of the glycerol utilization operon in S. venezuelae. The three genes of the gylFKD operon are transcribed from a single promoter. (B) Transcript levels in YP at early (2 days), mid (5 days), and late (9 days) time points for the glycerol catabolic gene cluster. (C) Images of representative S. venezuelae colonies spotted onto YP and YPG 2 to 9 days after inoculation. (D) Growth curves in solid medium comparing the rates of surface area expansion for S. venezuelae grown on YP and YPG. The average from six replicates was plotted under each condition. Error bars represent 1 standard deviation. (E) Representative images of a YPG-grown colony from the top and bottom sides of the plate showing the secretion of an orange pigment into the underlying agar. Images were taken after 7 days of growth. (F) Growth of wild-type S. venezuelae, a glycerol operon deletion mutant (ΔgylR-D), a complemented strain (Δgyl/+gyl) (middle right) in which the wild-type cosmid carrying the gylR-D operon was integrated into the chromosome by homologous recombination, and a complementation vector control strain (Δgyl Δgyl) (far right) in which a mutant cosmid with the gylR-D operon deleted was integrated into the chromosome by homologous recombination spotted onto YP and YPG. Representative images were taken after 7 days of growth.

To confirm that these changes in S. venezuelae exploration were mediated by glycerol uptake and metabolism, we generated a mutant strain in which gylR and the gylFKD operon were deleted. When inoculated onto YPG plates, the mutant grew as if it were “blind” to glycerol, phenotypically resembling the wild-type strain grown on YP (Fig. 5F). The dramatic YPG exploration phenotype could be restored to the mutant by reintroducing wild-type copies of these genes in trans (Fig. 5F).

Given the remarkable changes in exploration that stemmed from growth on glycerol, we were curious if the effects were specific to glycerol or if a similar effect could be conferred by supplementing the medium with any alternative carbon source. To test this, we grew wild-type S. venezuelae on YP (no-glucose-containing) plates supplemented with a variety of carbon sources (Fig. S2). Of the eight carbon sources tested, none enhanced exploration in the way that glycerol did.

Similarly, we were interested to know whether the glycerol effects on exploration were specific to S. venezuelae or if this effect was also observed for other streptomycetes. As the exploration response is not universally conserved among streptomycetes (or may be triggered by as-yet-undiscovered conditions for some species), we assembled a panel of 21 wild Streptomyces isolates that showed robust exploration on YP. For these strains, glycerol supplementation consistently resulted in colonies with increased wrinkling (21/21) and pigmentation (17/21) (Fig. 6). Glycerol supplementation was also frequently associated with accelerated exploration; over 60% (13/21) of these strains yielded colonies that were larger on YPG than on YP (e.g., WAC 5485) (Fig. 6). A full summary of phenotypes for the tested strains can be found in Table S2 in the supplemental material. In all, these data suggest that glycerol can dramatically alter exploration dynamics, colony architecture, and metabolism in diverse streptomycetes.

FIG 6.

Effects of glycerol on exploration in other Streptomyces species. Different wild Streptomyces isolates from the Wright Actinomycete Collection (WAC) were spotted alongside S. venezuelae on YP and YPG media. Representative colonies of select strains were photographed following 7 days of growth.

DISCUSSION

Our investigation into the transcriptional profile of exploring S. venezuelae revealed that there was a global remodeling of gene expression over time in YP-grown cells. Just as changes were not localized to specific regions of the chromosome, the functions of differentially expressed genes were similarly diverse. We saw notable up- and downregulation of genes participating in organic and inorganic nutrient acquisition, primary metabolism, and cellular defense, alongside other cellular functions and pathways. Known regulators that participate in classical development had, at most, subtle effects on exploratory growth, implying that additional strategies must contribute to the control of this behavior. We were unable to implicate a specific regulatory element in coordinating exploration, indicating either that there are exploration-specific regulators that remain to be discovered or that initiating and promoting exploration require changes in regulator activity, which would not necessarily be captured in our transcriptional analyses. There remain many regulatory genes whose expression changed significantly over the course of an exploration growth cycle, however, and it will be interesting to test the effects of some of these less well-characterized regulators on exploration, both alone and in combination with others.

We found that exploration was associated with an enhanced oxidative stress response, involving the upregulation of multiple superoxide dismutase- and catalase-encoding genes, and significantly increased catalase enzyme activity. Manipulating the levels of the most highly expressed catalase had minimal effects on exploration, suggesting considerable functional redundancy shared among these related enzymes. Notably, one of the regulatory genes whose expression decreased as exploration proceeded was wblA. The WblA group of regulators has been shown to negatively regulate the oxidative stress response in S. coelicolor (52), Corynebacterium glutamicum (53), and Mycobacterium tuberculosis (54). It is conceivable that WblA has a similar role in S. venezuelae, and its downregulation during exploration is needed to mount an effective oxidative stress response.

Our work here, alongside previous investigations into exploration (14, 15), suggests that increased respiration is a hallmark of Streptomyces exploration. We observed significant upregulation of ATP synthesis and iron-sulfur cluster biogenesis in our transcriptional analyses here, and we have previously demonstrated that a functional copy of the alternative cytochrome bd oxidase is necessary to drive S. venezuelae exploration when grown next to yeast on YPD (14). A key biochemical property that distinguishes the cytochrome bd oxidase from the primary cytochrome aa₃ oxidase is a higher affinity for molecular oxygen (28–30). If respiration in exploring cultures is proceeding at a higher rate than classical vegetative growth, we may expect that local microenvironments of the colony could quickly become hypoxic, necessitating the need for stronger oxygen capture strategies. This may also present a self-imposed environment conducive to the function of the oxygen-sensitive Wbl proteins. The intricate surface wrinkling patterns that we observe exclusively in exploring colonies may help cells respond to oxygen depletion by maximizing the cell surface area for gas exchange. Indeed, increased colony wrinkling in response to oxygen depletion or impaired respiration appears to be a common response in microbial biofilm communities, including those of Pseudomonas aeruginosa (31–33), Bacillus subtilis (34), Candida albicans (35), and Aspergillus fumigatus (36).

An unexpected outcome of probing the transcriptional program of explorer cells was the fascinating response of these colonies to glycerol supplementation. In the presence of glycerol, rates of exploratory growth readily outpaced those of traditional YP-grown explorers, and colonies developed more complex wrinkling morphologies. Curiously, this enhancement appeared to be specific to glycerol; none of the other tested alternative carbon sources had the same exploration-promoting effect. This suggested that the role of glycerol during this growth mode may extend beyond simply acting as an additional carbon substrate for biomass and energy production.

The architecture of exploring Streptomyces colonies is reminiscent of those of many bacterial biofilms. Notably, glycerol has been implicated in inducing or enhancing the establishment of bacterial biofilms in multiple systems. In the model organism B. subtilis, cultures supplemented with a combination of glycerol and manganese exhibit robust biofilm formation, with the histidine kinase KinD (for which no obvious homolog exists in S. venezuelae) being responsible for sensing and transducing this environmental signal (37). In Listeria monocytogenes, glycerol specifically induces biofilm formation at the air-liquid interface of aerobic broth cultures (38). Similar effects are seen in the pathogen P. aeruginosa, where growth on glycerol supports increased biofilm formation, in part through the resulting overproduction of Pel polysaccharide (39). In the yeast C. albicans, glycerol biosynthesis genes are upregulated during biofilm growth relative to planktonically growing cells. Disruption of glycerol synthesis in this fungus leads to broad transcriptional changes, including reduced capacities for biofilm formation and adherence, highlighting the role of this metabolite in the genetic regulation of biofilm development (40). Similarly, in the soil bacterium Janthinobacterium lividum, glycerol induces two phenotypic responses: increased biofilm formation (as evidenced by increased extracellular polysaccharide production) and overproduction of the purple pigment violacein (41). The production of an unknown, but highly conserved, pigment was also observed for the majority of glycerol-grown exploring Streptomyces cultures.

Given the strong response of exploring cultures to glycerol, it is worth considering how this observation is potentially reflective of the natural environment. The soil houses a particularly rich community of diverse microbes and is prone to large fluctuations in nutrient availability. In response to elevated glucose concentrations, yeasts like S. cerevisiae can produce and secrete glycerol as a by-product of fermentation (42, 43). This parallels what we know about exploration in S. venezuelae, where coculture with S. cerevisiae on YPD and subsequent glucose depletion from the growth medium allow a robust exploration response. It is conceivable that the accelerated rate of exploration in response to exogenous glycerol supplementation may be recapitulating an ecologically relevant strategy when encountering yeast under nutrient-rich growth conditions.

In all, we have shown that exploration requires a significant respiratory investment and employs regulatory cascades that have some overlap but are largely distinct from those of classical development. How glycerol changes the exploration process and what metabolic changes accompany exploration under all conditions are questions of interest for future investigation.

MATERIALS AND METHODS

Strains, plasmids, media, and culture conditions.

Strains, primers, and plasmids used in this study are listed in Tables S3 to S5 in the supplemental material. S. venezuelae NRRL B-65442 was grown in liquid MYM (1% malt extract, 0.4% yeast extract, 0.4% maltose) for cultivation overnight and on solid MYM (2% agar) for spore stock generation and vegetative (nonexploratory) growth controls. Spore stocks for wild isolates were similarly prepared. For exploration experiments, 10 μL of a culture of S. venezuelae (MYM [10 mL]) grown overnight was spotted onto solid YP medium (1% yeast extract, 2% peptone, 2% agar) additionally supplemented with 2% carbon source (e.g., dextrose or glycerol) and/or additional nutrients (sodium nitrate, sodium bisulfite, and sodium sulfate [10 μM, 100 μM, or 1 mM]) where appropriate. In exploration experiments where the growth of multiple strains was being compared, cultures grown overnight were diluted with MYM to a normalized optical density at 600 nm (OD₆₀₀). For experiments assessing the growth of wild Streptomyces isolates, due to the frequent nondispersed growth of different strains in liquid MYM, 5 μL of spore stock solutions was directly spotted onto the relevant solid medium. All Streptomyces cultures were grown at 30°C.

Construction of S. venezuelae mutants.

Gene deletions were generated using ReDirect technology (44). Coding sequences on a cosmid vector carrying large fragments (30 to 40 kb) of S. venezuelae genomic DNA were replaced by an oriT-containing apramycin (ΔgylRFKD [equivalent to Δvnz_06115-06130 (deletion of vnz_06115 through vnz_06130)], Δvnz_36165 [catalase-encoding gene], ΔeshA [Δvnz_35035], and ΔsigQ [Δvnz_22610]) or hygromycin (Δbla or ΔgylRFKD, for the complementation control) (see below) resistance cassette. In creating the eshA deletion strain, the coding sequence and the flanking 2-kb up- and downstream sequences were PCR amplified and cloned into the pCR2.1-TOPO vector between the HindIII and SpeI restriction sites (as this region lacked an appropriate cosmid for deletion), before the eshA gene was targeted for replacement with the apramycin resistance cassette, as described above. The mutant cosmids/plasmids were introduced into the nonmethylating Escherichia coli strain ET12567/pUZ8002, followed by conjugation into S. venezuelae. The resulting exconjugants were screened for double-crossover events, and gene deletions were verified by PCR using combinations of primers located upstream of, downstream of, and internal to the deleted regions (Table S5).

Complementation of the ΔgylRFKD mutation was accomplished by the introduction of a copy of the S. venezuelae cosmid Sv-3-E04 (containing wild-type copies of these genes plus the associated upstream and downstream sequences) into the mutant strain. To facilitate conjugation and selection for cosmid integration into the S. venezuelae genome, the ampicillin resistance gene (bla) on the vector backbone was replaced with an oriT-containing hygromycin resistance cassette. To control for the additional up- and downstream sequences included on the cosmid, an analogous construct was generated by replacing the wild-type gylRFKD coding region with an oriT-containing hygromycin resistance cassette. Both constructs were introduced into the gylRFKD mutant strain for phenotypic comparison with the wild-type and mutant strains.

To generate the catalase overexpression construct, the coding sequence of vnz_36165 was amplified from S. venezuelae genomic DNA with primers incorporating EcoRI and BamHI restriction enzyme recognition sites, which were subsequently used to clone the fragment directly downstream of the strong constitutive ermE* promoter sequence in pMC500 (Text S1). The promoter-gene fragment was then excised with KpnI and HindIII and subcloned into the equivalently digested integrating plasmid pMS82 (Table S4).

To generate the sigQ overexpression construct, a 487-bp fragment containing the ermE* promoter was digested out from pIJ12251 using PvuII and EcoRV and subcloned into EcoRV-linearized pMS82. The coding sequence of sigQ (vnz_22610) was subsequently amplified from S. venezuelae genomic DNA with primers incorporating NdeI and XhoI restriction enzyme recognition sites (Table S5). The resulting amplicon was then cloned into the pMS82-ermE*p construct following digestion by NdeI and XhoI (Table S4).

Time-lapse videos of exploring colonies.

Cells were inoculated onto exploration medium, after which the plates were placed in a 30°C incubator on an Epson Perfection V800 photo scanner that had been programmed to acquire one image every hour. The time course images were then compiled in a video format as sequential single frames.

Growth curves in solid medium.

Photos of growing colonies were taken at set time intervals. All image analyses were performed using ImageJ. The scale was established for each image by setting the diameter of the petri dish to 10 cm. The perimeter of the colony was traced, and the area of the captured region was determined.

RNA isolation, library preparation, and cDNA sequencing.

RNA was isolated, as described previously (45), from two independent replicates of S. venezuelae grown on solid exploration medium for the specified incubation times (2, 5, and 9 days of growth on YP). For all samples, rRNA was depleted using a Ribo-Zero rRNA depletion kit. cDNA and Illumina library preparation were performed using a NEBNext Ultra directional library kit, followed by sequencing using un-paired-end 80-bp reads on the HiSeq Illumina platform. All bioinformatic analyses were carried out using packages available through the free open-access platform Galaxy (https://usegalaxy.org). Reads were aligned to the S. venezuelae genome using Bowtie2 (46) and then sorted, indexed, and converted to BAM format using SAMtools (47). Transcript level normalization and analyses of differential transcript levels were conducted using DESeq2 (48). COG analysis was performed using the open-source online tool EggNOG-mapper (49).

Native PAGE assay for catalase enzyme activity.

In-gel assays for catalase activity were conducted as described previously by Weydert and Cullen (50). Briefly, S. venezuelae cells were harvested from solid medium at time points and conditions of interest before being resuspended in 1 mL of lysis buffer (100 mM NaCl, 5% glycerol, 10 mM Tris [pH 8], and one cOmplete Mini EDTA-free protease inhibitor pellet [Roche] per 15 mL). Cell lysates were prepared by sonication (6 cycles of 15 s of 40% duty followed by at least 15 s on ice) (Branson Cell Disruptor 350 sonifier) followed by centrifugation at >15,000 × g for 5 min at 4°C. The protein concentration of the lysate supernatants was determined by the Bradford assay (51). For each sample, 40 μg of total protein was loaded and resolved on an 8% nondenaturing acrylamide gel (19:1 acrylamide/bisacrylamide ratio) under constant voltage at 125 V for 2.5 h. After separation, the recovered gels were rinsed three times with MilliQ water and then soaked in 0.003% hydrogen peroxide for 10 min with mild shaking. Following equilibration, gels were rinsed three times with MilliQ water and then stained by the simultaneous addition of 10 mL of 2% ferric chloride and 10 mL of 2% potassium ferricyanide with gentle agitation. As soon as bands of clearing could be distinguished from the background staining of the gel, the excess stain solution was poured off, and the gel was thoroughly washed with MilliQ water to stop the reaction. As an additional loading control to compare protein concentrations between samples, 40 μg of protein was run under standard SDS-PAGE conditions on a 10% denaturing acrylamide gel (19:1 acrylamide/bisacrylamide ratio) under constant voltage at 125 V for 1 h. Following separation, proteins were stained with Coomassie brilliant blue G-250.

Data availability.

RNA-seq data were submitted to the NCBI GEO repository and assigned the accession number GSE186259.