Nutrient-deprived growth of streptomycetes reveals a novel growth phenotype with enhanced antimicrobial activity

The capacity of gram-positive streptomycetes to adapt to various environments is reflected in the sheer size of their genomes and the number of encoded regulatory proteins. However, typical studies of Streptomycetes are conducted using media rich in available nutrients. Here, we demonstrate that S. coelicolor colonies respond to nutrient depletion with a phenotype transition. We refer to this phenotype as “foraging growth” for its submerged and continuous growth on depleted media. The foraging phenotype differ distinctly from colonies on nutrient-rich media, in terms of colony morphology, genomic stability and their secretome. The results demonstrate that adaption to nutrient deprivation through foraging is found throughout the phylogeny, indicating it is highly conserved. Furthermore, foraging S. coelicolor gains the ability to inhibit mold on deprived media and exhibits an altered metabolomic profile. An enhanced competitive activity against gram-negative and gram-positive bacteria was also observed among the tested species. These findings highlight the morphological adaptions of streptomycetes during nutrient deprived growth and demonstrate altered secondary metabolite production and that novel antimicrobial activities can be detected.


Introduction
The gram-positive genus Streptomyces, known for its filamentous morphology and production of bioactive secondary metabolites, plays a crucial role in soil as saprophytes hydrolyzing a wide array of polysaccharides and macromolecules (Chater, Biro, Lee, Palmer, & Schrempf, 2010).Streptomycetes undergo a complex life cycle, beginning with spore germination, progressing through vegetative hyphal growth and branching to form a mycelial network.A differentiation process leads to the development of aerial hyphae and spores, which often coincide with activation of secondary metabolites.The differentiation relies on the bld-genes to control aerial development and the whigenes for spore maturation (McCormick & Flardh, 2012).The life cycle of streptomycetes has extensively been studied on rich media where subsequent nutrient depletion is thought to contribute to mycelial differentiation and secondary metabolite production.
While the importance of optimized growth medium for metabolite production and life cycle progression is widely recognized, few have examined nutrient poor conditions.The abundance of gene regulatory proteins, two-component systems, transporters, and substrate-binding proteins identified in the S. coelicolor genome suggests a significant capacity for adaptation and the utilization of varied nutrient sources (Bentley et al., 2002;Nikolaidis et al., 2023).A recent study on rich media showed that S. coelicolor metabolite production was enhanced because of genomic heterogeneity (Z.Zhang et al., 2020).Additionally, certain streptomycetes exhibit an "exploratory phenotype" when coincubated with yeast or by specific media components alone (S.E. Jones et al., 2017;Shepherdson & Elliot, 2022).As for conventional growth, exploratory growth is closely connected to metabolite production and activates cryptic metabolites (Shepherdson & Elliot, 2022).
Genetic identification of biosynthetic gene clusters (BGCs) has revealed that each Streptomyces spp.
could have the capacity to produce over 20 secondary metabolites (Bentley et al., 2002;Ikeda et al., 2003;Ohnishi et al., 2008;Qin et al., 2017).However, standard cultivation often fails to stimulate the activation of the regulatory pathways responsible for their synthesis (Nett, Ikeda, & Moore, 2009;Ochi & Hosaka, 2013).After exhausting primary substrates, carbon catabolite repression enables the adaptation of growth and metabolism to utilize to less prioritized substrates.Rapidly consumed carbon sources are often found to repress production of secondary metabolites (Romero-Rodriguez et al., 2017).Therefore, altering environmental stimuli to activate silent secondary metabolite production is a promising strategy in the search for new bioactive compounds (X.Zhang, Hindra, & Elliot, 2019).
In this study, we demonstrate that global nutrient depletion in growth media coincides with a morphological transition of S. coelicolor colonies to a phenotype termed "foraging".Our findings reveal that bldA is not essential for foraging growth, although TTA codon-containing gene products were detected in the foraging secretome.Additionally, S. coelicolor exhibited mold-inhibitory activity under nutrient-depleted conditions.Metabolomic analysis showed significant changes in the metabolome, and growth inhibition assays suggested roles for bldA and adpA in antimicrobial activity during nutrient-deprived growth.These results underscore a competitive advantage for foraging colonies on nutrient-deprived media.Furthermore, all strains cultured under these conditions exhibited foraging growth, suggesting this behavior is evolutionarily conserved.This study indicates that nutrient-deprived growth may be a key method to both generate novel bioactive secondary metabolites in addition to providing a more complete understanding of the life cycle.

Phenotype transition from conventional S. coelicolor colony morphology correlate with nutrient depletion.
In a phenotypic screening of different media formulation, we observed a unique growth morphology of S. coelicolor after 60 days on 0.5x diluted TSA medium.The colonies had developed an unusual thinner, veil-like, outer growth zone, distinct from the conventional S. coelicolor morphology (Fig. 1A).This morphology, termed "foraging growth", was further investigated using time-lapse imaging over 75 days and revealed a morphological transition around day 50 (Video S1).
This transition appears linked to nutrient availability, as it occurred earlier on smaller media volumes and later on undiluted TSA.Foraging growth was evident already after 21 days when grown in ⌀ 4 cm dishes with 0.5xTSA.Using these dishes, we monitored nutrient depletion by inoculating plates with spores and sampled the agar every three days for 27 days and analyzed nutrient content using mass spectrometry.Samples were taken 0.5 cm away from the colony, to ensure that the samples would not be contaminated with hyphae, and the very edge of the dish.Altogether, 51 metabolites were annotated, including sugars, amino acids, and fatty acids, most of which depleted over time (Fig. 1B).
Nutrient levels proximal to the colony mirrored those at the dish-edge (Figure 1 -figure supplement 1A).Although the levels of e.g., glucose and fructose were depleted at foraging onset, others e.g., levels of glutamine and lactic acid initially increased before decreasing.Some, e.g., leucine and phenylalanine, remained stable until day 12, before decreasing (Fig. 1C, Figure 1 -figure supplement 1B).The levels of some, e.g., sucrose, stayed constant throughout, and only five metabolites increased more than one-fold (Fig. 1C).However, supplementing these metabolites in 0.5xTSA and MS agar did not induce foraging growth.These findings demonstrate that the morphological transition coincides with media nutrient depletion.

Nutrient depletion promotes foraging growth.
Modified Streptomyces-minimal medium only contain two carbon sources: agar and 55 mM glucose (Kieser et al., 2000).Observing that foraging growth commenced upon carbon source depletion, we excluded glucose, making agar the sole carbon source in the modified medium, hereafter referred to as foraging agar media (FAM).Colonies on FAM displayed similar characteristics to those on 0.5xTSA post-transition, exhibiting a thin, veil-like submerged expansion eventually covering the entire plate (Fig. 2A).Over time, lineages developed within the colony, some of which became pigmented (Fig. 2B).The amount of aerial hyphae and pigment production at the inoculation site was influenced by the presence of metabolizable nutrients, such as glycerol, in the spore storage buffer (Fig. 2C), underscoring the effect of nutrients on the growth characteristics of the foraging phenotype.FAM supplemented with glucose or fructose effectively inhibited foraging growth, leading to conventional raised colonies (Fig. 2D).Notably, sucrose supplementation of FAM did not prevent foraging growth, indicating the inability of S. coelicolor to metabolize sucrose.Time-lapse imaging captured the progressive expansion and sector formation of a foraging colony on FAM (Fig. 2E, Supplementary Video S2).Introduction of glucose or tryptone to a foraging colony induced raised colony growth and strong pigment production, typical for growth in nutrient-rich conditions (Fig. 2F, 2G, Supplementary Video S3 and S4).Scanning Electron Microscopy (SEM) imaging of FAM revealed predominantly submerged hyphal growth, with limited hyphae at the surface (Fig. 2H).Aerial hyphae were found both at the site of inoculation and uncommonly at other patches throughout the colony.In contrast, growth on nutrient-rich media such as 0.5x TSA revealed that vegetative hyphae were primarily piled up on the surface, with only a few extending further into the agar.
Sector formation was observed during foraging growth (Fig. 1, 2A, 2B, Video S1 and S2).To investigate these sectors for potential genomic instability as reported for S. coelicolor grown on nutrient rich media (Z.Zhang et al., 2020), a lineage from the initial inoculation, a subsequent sector 'A', and its subsequent sectors 'I' and 'IV', were sequenced using the Illumnia platform.Analysis using breseq software (Deatherage & Barrick, 2014) showed no significant genomic instability compared to growth on nutrient-rich media.Only three mutations were identified: two mutations in sector A, in the chromosomal replication initiator protein DnaA (SCO3879, E520K) and in a hypothetical protein (SCO1511, Q8CK26).Sector IV had an additional mutation in a putative regulator (SCO5582, Q9ZBP7) (Figure 2 -figure supplement 1).These mutations did not visibly affect colony morphology on rich media.Altogether, these results demonstrate that S. coelicolor behaves distinctly during nutrient deprived growth.

S. coelicolor foraging growth secretome is both Sec and Tat dependent.
S. coelicolor grown on a deprived media such as FAM must utilize the agar as the main carbon source.
Agar degradation was visualized by staining the plates with iodine-based Lugol's solution, which revealed a distinct halo around the colonies indicating substrate degradation (Fig. 3A).Proteins were extracted from the zones outside of foraging colonies and analyzed using mass spectrometry.The resulting protein composition was visualized using proteomaps (Liebermeister et al., 2014).Of the 203 detected proteins, 52 were annotated into six categories: membrane transport, biosynthesis, central carbon metabolism, signal transduction, other enzymes, and folding/sorting/degradation/transcription, as shown in the treemap (Fig. 3B) and the proteins represented in the gene name mosaic map (Fig. 3C).
Signal peptide prediction on all S. coelicolor coding sequences using SignalP 6.0 (Teufel et al., 2022) identified 696 proteins with signal sequences.Prediction of the 203 proteins in the foraging secretome showed: 56 with the Sec translocon, 18 with the Tat translocon and 123 without any signal peptide (Fig. 3D, Supplementary table S1).Proteomaps of these three groups indicated functional overlaps among proteins with Sec, Tat and without signal peptides (Fig. 3E).Comparing our findings with previous studies of S. coelicolor secretomes on rich media (Kim, Chater, Lee, & Hesketh, 2005;Widdick et al., 2006) showed that 24% of the proteins identified in foraging conditions were also detected in nutrient-rich conditions, while 75% were unique to foraging, including the glucose-repressible extracellular agarose DagA (Figure 3 -figure supplement 1A, Supplementary table S2).Comparing the proteins lacking signal peptides from foraging (n=203) and rich conditions (n=401) revealed an overlap of 60 proteins (Supplementary table S4).
PANTHER classification of signal peptide-containing proteins categorized them into four groups (Fig. 3F, Supplementary table S4).Notably, the transporter category included the ABC transporter BldKB, known for its role in carbon catabolite repression and mycelium formation (Chavez et al., 2011;Nodwell & Losick, 1998).The metabolite interconversion category featured 13 hydrolases and one oxidoreductase, while the protein modifying enzyme category included ten proteases and a ubiquitinprotein ligase.Among the unclassified proteins were additional hydrolases and solute binding proteins.The non-signal peptide proteins were classified into 11 categories (Figure 3  In summary, these results uncover the secretome components crucial for foraging growth of S.
coelicolor on depleted media, highlighting the role of both Sec and Tat transport systems in nutrient degradation and uptake of agar components.

Foraging phenotype is found across the phylogeny of Streptomyces.
To investigate the influence of classical developmental regulators (bld and whi) on foraging growth, we cultured wild-types and developmental-mutants on FAM and found that all tested mutants displayed the foraging phenotype (Figure 4 -figure supplement 1A).Notably, the bldA mutant exhibited foraging growth despite the presence of TTA-codon containing gene products in the foraging secretome.
We then extended the study to various Streptomyces spp., representing different branches of the phylogenetic tree (Table 1), to assess if foraging growth was unique to S. coelicolor.Pairwise average nucleotide identity (ANI) values were utilized to create a heatmap that groups strains according to their genomic similarity, showing that only S. coelicolor and S. lividans shared an ANI >99 (Fig. 4A).All tested strains displayed the foraging growth phenotype on FAM (Fig. 4B), with little to no pigmentation and agar degradation confirmed by Lugol's solution staining (Figure 4 -figure supplement 1B).Similar to S. coelicolor, the foraging growth was inhibited in all strains with the addition of 1% glucose or tryptone to FAM (Figure 4 -figure supplement 1C and 1D).Additionally, from local soil samples grown on FAM, we selected seven colonies exhibiting the foraging morphology.
These were all confirmed to be Streptomyces spp.using MALDI-TOF analysis (Supplementary table S7).
This observation suggests the ubiquity of foraging growth across the Streptomycetes phylogenetic spectrum.Furthermore, the growth behavior of developmental mutants on FAM indicates that these traditional developmental regulators do not control foraging growth.Foraging S. coelicolor displays an altered metabolomic profile.
S. coelicolor colonies grown on FAM were found to inhibit the growth of a contaminating mold (Fig. 5A), identified as Penicillium citrinum through 18S/ITS sequencing.This inhibition was replicated by growing S. coelicolor on FAM for five days before inoculating P. citrinum spores at a distance of 2 cm, resulting in an average 4.3 mm inhibition zone (±0.76 mm, n = 96).Inhibition was not observed when glucose was added to the agar or on any other nutrient-rich media, nor on FAM with pH below 7.
Interestingly, the bldA mutant, is typically unable to produce certain antimicrobial metabolites, inhibited P. citrinum growth on FAM.However, the adpA-mutant strain M851 showed less inhibition compared to the wild-type (Fig. 5B).All tested whi-mutants (-A, -B, -G, -H, and -I) exhibited moldinhibitory activity (Figure 5 -figure supplement 1A).Inhibition of a local Aspergillus spp.isolate was also demonstrated using an overlay assay with a 100x diluted TSA-overlay, since the Aspergillus spp.
did not grow on FAM (Figure 5 -figure supplement 1B).However, no inhibition of Saccharomyces cerevisiae or Candida albicans was detected using the same assay.
To explore the metabolomic profile of S. coelicolor grown on FAM and FAM supplemented with 1% glucose (FAM+G), agar was sampled around the colonies for mass spectrometric analysis.The metabolomic profiles differed between FAM and FAM+G, both in negative (R2(X)=0.75,Q2(X)=0.97) and positive detection modes (R2(X)=0.63,Q2(X)=0.47) (Fig. 5C, 5E).Some features with relatively higher concentrations on FAM-agar (highlighted in green in Fig. 5D, 5F) could be matched with known S. coelicolor metabolites based on monoisotopic molecular weight using the StreptomeDB (Moumbock et al., 2021).Among these annotated features, the signaling molecule γ-butyrolactone was found at higher concentrations in FAM+G.Wailupemycin G was the only annotated metabolite found at considerably higher concentrations on FAM+G (Fig. 5G, Table 2).There were 30 features, with higher levels in the FAM condition, that could not be annotated using their monoisotopic mass and StreptomeDB (Supplementary table S8).Twelve of these features were unique to FAM and not detected in any other tested condition (Figure 5 -figure supplement 1C and 1D).
The mass spectrometry results demonstrate a significant alteration in the metabolomic profile of foraging S. coelicolor.The mold inhibitory assay using mutants suggests that mold inhibition involves adpA and other genes dependent on bldA.Foraging Streptomycetes demonstrate increased antibacterial activity.
To investigate whether mold-inhibition was a common behavior during foraging, all Streptomyces spp.
tested for foraging growth above, were tested for mold inhibition.However, none of the other species inhibited P. citrinum.The altered metabolome of S. coelicolor during foraging, combined with its ability to hinder mold growth, led us to investigate its potential to inhibit bacterial growth.Due to the nutrient-depleted state of FAM, B. subtilis and E. coli were unable to proliferate.Therefore, to test for antibacterial activity, we overlaid FAM with foraging S. coelicolor colonies with diluted LA and then added B. subtilis 168 and E. coli MG1655 on top of the LA layers.The growth was evaluated after incubation using a stereo microscope.This method was applied to all previously studied Streptomyces spp. to assess their antibacterial properties.Most species showed increased inhibitory activity on nutrient-depleted FAM-agar compared to FAM supplemented with glucose (FAM+G) and tryptone (FAM+T) (Fig. 6A).However, S. coelicolor did not exhibit enhanced antibacterial activity on FAM compared to the nutrient-supplemented media.
Analysis of bacterial inhibition by developmental mutants indicated that the bldA-mutant lost its ability to inhibit B. subtilis growth, whereas the adpA mutant acquired the ability to inhibit E. coli (Fig. 6B).In summary, these findings suggest that streptomycetes are more competitive on FAM than on rich agar.The distinct responses of the bldA and adpA mutants further emphasize their roles in the antimicrobial activities of foraging S. coelicolor.

Strains, media, and culture conditions
The strains used in this study are listed in Supplementary Table S9.Streptomyces strains were grown on FAM (1 g (NH4)2SO4, 0.5 g K2PO4x7H2O, 0.1 g FeSO4x7H2O, 10 g agar, in 1L pH 7.5) (Kieser et al., 2000).FAM were supplemented with 1% of various sugars: glucose (FAM+G), sucrose (FAM+S), fructose (FAM+F) or tryptone as an amino acid source (FAM+T) when indicated.To investigate the effect of pH on mold-inhibition the pH of FAM was set using HEPES-buffer and HCl/NaOH.Tryptone soy broth (Thermo Fisher Scientific) was prepared according to instructions and diluted 0.5x when indicated with water and 1.5% agar added before autoclaving.MS-agar (20 g mannitol, 20 g Soy flour, 15 g agar in 1L) agar were used for sporulation and petri dish time-lapse imaging.Effects of accumulated metabolites were conducted by adding 20mg/ml: beta-alanine, meso-erythriol (Merck), Potassium-citramalate monohydrate (Merck), urea (Merck), and pantothenic acid (Merck).All incubations were carried out at 30°C.

Isolation and identification of local isolates
The inhibited mold was isolated from the FAM dish and recultivated on TSA.The agar plate was sent to Eurofins for 18S/ITS sequencing.For isolation of local streptomycetes, soil samples from potatoes and plant soil were dried at room temperature before being deposited on FAM.The plates were incubated at 37°C until foraging colonies were observed after approx.5-10 days.Colonies with characteristic foraging phenotypes were re-streaked on TSA and characterized using MALDI-TOF (Bruker).The local Aspergillus spp. was isolated from a soil sample and assessed morphologically.

Colony imaging
For all imaging of small colonies spores were inoculated on agar media and visualized using a Nikon SMZ1500 dissection microscope and a Nikon Digital sight DS-Fi1 camera, with either surface illumination using LEDs or oblique sub-stage illumination.Whole petri dishes were imagined using a Nikon D7000 DSLR.Agar time-lapses were recorded using a Raspberry Pi 3 model B with camera module v2 mounted in a 10x lens (Olympus).The device was placed in a 30°C dark room with an LED lamp programmed to illuminate the plate only during image capture, one frame per hour.ImageJ distribution Fiji (Schindelin et al., 2012) was used to process single frames into videos and Adobe Photoshop CC software for still images.

Electron microscopy
Colonies of S. coelicolor on agar were cut, fixed overnight at 4°C in 2.5% glutaraldehyde/0.1Msodium cacodylate buffer, dehydrated in series of ethanol, critical point dried and coated with a 5 nm Au/Pd before being analyzed with a Merlin Field Emission SEM (Zeiss) using SE-HE2 and in-lens detectors.
Images were processed using Adobe Photoshop CC software.

Inhibition assays
For mold inhibition, Streptomyces spores were inoculated on FAM and incubated at 30°C for 5 days.
Spores from P. citrinum were inoculated 2 cm from the foraging colony.The plates were incubated for 14 days and observed daily for inhibition zones and imaged as described above.For bacterial inhibition, plates with Streptomyces spp.were incubated for 20 days before a thin layer of 100x diluted TSA (1.5% agar) was placed on top of agar and incubated overnight at 4°C.Overnight cultures of E. coli and B. subtilis were diluted and 2x10 7 CFU in 100 µl was added to the agar overlays and incubated overnight at 37°C.Inhibition zones were assessed and imaged using a Nikon SMZ1500 dissection microscope and a Nikon Digital sight DS-Fi1 camera.

Foraging secretome protein analysis
The agar outside of foraging S. coelicolor on FAM incubated for 20 days was extracted.The agar was crushed and incubated overnight with water in a 50 ml Falcon tube.Proteins in the supernatant were precipitated by addition of TCA and incubated on ice of 4 minutes before centrifugation at 20,000g for 5 min.The pellet was washed twice with 200 l ice-cold acetone.The pellet was dried, dissolved in protein sample buffer and loaded on an SDS-PAGE.Protein dense regions were cut and sent for mass spectroscopy identification to Linköping University, where general in-gel digestion was performed as described by Shevchenko et al. (Shevchenko, Tomas, Havlis, Olsen, & Mann, 2006).

Metabolomic analysis
Metabolic profiling of media nutrients by GC-MS was performed at the Swedish Metabolomics Center in Umeå, Sweden.For the nutrient analysis of agar, six replicates of agar plugs were collected from the edge of the dish and 0.5 cm away from the S. coelicolor colonies on 0.5xTSA medium.This sampling was performed on six different ⌀ 4 cm plates.Three control samples were taken at every time point from three sterile 0.5xTSA dishes.Sample preparation was performed according to A et al. (A et al., 2005).The samples were stored at -80 °C until analysis.Small aliquots of the remaining supernatants were pooled and used to create quality control (QC) samples.The samples were analyzed in batches according to a randomized run order on GC-MS.Derivatization and GCMS analysis were performed as described previously A et al. (A et al., 2005).Non-processed MS-files from the metabolic analysis were exported from the ChromaTOF software in NetCDF format to MATLAB-R2020a (Mathworks, Natick, MA, USA), where all data pre-treatment procedures (base-line correction, chromatogram alignment, data compression and Multivariate Curve Resolution) were performed.The extracted mass spectra were identified by comparisons of their retention index and mass spectra with libraries of retention time indices and mass spectra (Schauer et al., 2005).Mass spectra and retention index comparison was performed using NIST MS 2.2 software.Annotation of mass spectra was based on reverse and forward searches in the library.Masses and ratio between masses indicative of a derivatized metabolite were especially notified.The mass spectrum with the highest probability indicative of a metabolite and the retention index between the sample and library for the suggested metabolite was ± 5 the deconvoluted "peak" was annotated as an identification of a metabolite.The annotated nutrients were plotted using GraphPad Prism (version 9.0.2) to show trends over time and ridge plots were generated using R version 4.2 and packages ggplot2 3.4.4and ggridges 0.5.4.
For untargeted metabolite analysis, agar stubs were punched out from just outside of S. coelicolor colonies on FAM and FAM+G.For each condition, twelve samples were isolated from plates inoculated with S. coelicolor, in addition to two control samples each from sterile plates.The weight of the samples were adjusted to 100 mg before the addition of 1 mL extraction buffer (90/10 v/v methanol:water).Internal standards (13C9-Phenylalanine, 13C3-Caffeine was, D4-Cholic acid, D8-Arachidonic Acid, 13C9-Caffeic Acid) were added to each sample.The sample was shaken at 30 Hz for 3 min in a mixer mill, and proteins were precipitated at +4 °C for 2h on ice.The sample was centrifuged at +4 °C, 14 000 rpm, for 10 min.The supernatant was transferred to a microvial and solvents evaporated.Before analysis, the sample was re-suspended in 10 + 10 µL methanol and water.The set of samples were analyzed in positive mode and negative mode.The chromatographic separation was performed on an Agilent 1290 Infinity UHPLC-system (Agilent Technologies, Waldbronn, Germany).

DNA sequencing of foraging sectors
Two levels of sectors formed during foraging on FAM were grown in TSB-media and 600 mg of pellet was sent to MicrobesNG for Illumnia sequencing.Quality controlled and trimmed reads with 30x coverage were obtained and analyzed using breseq software (Deatherage & Barrick, 2014) with default settings, using the S. coelicolor A(3)2 reference genome (Genbank: GCA_000203835.1, (Bentley et al., 2002)) updated using wild-type sequence data to accommodate for genomic variation in lab strain.The reads of descendent sectors were mapped against this updated reference genome.
Illumnia read data used in this study have been deposited to SciLifeLab Data Respository with the dataset identifier DOI: https://doi.org/10.17044/scilifelab.c.7171122.

Statistical analysis
In this study, certain measurements in the untargeted metabolomics analysis fell below the detection limits, resulting in missing values.To ensure an accurate representation of our findings and maintain the integrity of the statistical analysis, these missing values were omitted from the dataset.
Consequently, all values presented in the violin plots depict observable data trends and variations above the detection threshold.All statistical analyses were conducted using GraphPad Prism (version 9.0.2).Details on the specific type of statistical analysis and the criteria for significance testing are provided in the legends of the respective figures.

Discussion
The primary habitat of streptomycetes, the soil, is a dynamic environment with variable environmental factors.Laboratory cultivation on a narrow range of rich media may not fully represent adaptations occurring in the diverse, competitive soil environment.This study reports a novel phenotype-transition of S. coelicolor colonies upon nutrient depletion.
Metabolomics analysis showed that the phenotype transition on 0.5xTSA coincided with nutrient depletion, leading to the "foraging phenotype".Media composition is known to be crucial for life-cycle progression, with nutrient depletion often referred to as the trigger of differentiation and sporulation, though the process is acknowledged to be complex (Barka et al., 2016;Bush, Tschowri, Schlimpert, Flardh, & Buttner, 2015;Flardh & Buttner, 2009;McCormick & Flardh, 2012).Instead of a sporulation, S. coelicolor responded to nutrient depletion with a phenotype transition.Four distinct trends in metabolite dynamics were distinguishable: depletion, accumulation, initial accumulation followed by depletion, and unchanged levels.The trend of initial accumulation followed by depletion is likely due to the breakdown of macromolecules in the TSA-media.It is established that fluctuations of metabolite levels during growth are a result of the breakdown and uptake of external sources and subsequent release into the surroundings (Douglas, 2020).Steady accumulation may occur from byproduct export or independently of metabolite overflow (Pinu et al., 2018).Supplementation of the accumulated metabolites to media did not trigger foraging growth, suggesting they do not act as signals for phenotype transition.However, two accumulated metabolites, β-alanine and pantothenic acid, are crucial in various biochemical pathways.β-alanine, essential for pantothenic acid synthesisa precursor of Coenzyme A -enhances respiration in alkaline soils (Datta et al., 2017).Their release might influence the microbial landscape in the natural habitat of S. coelicolor.
Nutrient levels proximal to the colony and at the edge of the dish closely followed each other, explaining the persistence of the foraging phenotype even as it expands into unexplored areas.
In contrast to rich media, where low molecular weight organic compounds are abundant (Kieser et al., 2000;Shirling & Gottlieb, 1966), soil environments typically have transient availability of such nutrients (D.L. Jones & Murphy, 2007;Rousk & Jones, 2010).The amount of complex carbohydrates and plant debris of fertile soils are considered high, but there is also a flux of available nutrients throughout the year that the soil inhabitants need to accommodate for (Randewig, Marshall, Näsholm, & Jämtgård, 2019).FAM, our minimal defined agar medium without glucose, supports spore germination and foraging growth, but growth reverted to conventional phenotype upon the addition of nutrients.This indicates that the complex polysaccharides in agar are not preferred over simpler organic compounds.Nitrogen and phosphate also play a crucial role in cycle progression and metabolite production.Our FAM media, with higher nitrogen levels than typically found in soil, may influence these processes (Rousk & Jones, 2010).Nitrogen-concentrations higher than 1 mM allow synthesis of glutamate independent of ATP, but low levels require the activity of glutamate synthetase (Reuther & Wohlleben, 2007).Glutamate synthetase II activity its mRNA levels correlate with mycelial differentiation on solid media demonstrating its importance in mycelial development (Fink, Weissschuh, Reuther, Wohlleben, & Engels, 2002).High levels of nitrogen in rich media cause repression of secondary metabolite production (Hodgson, 2000).Further, high levels of inorganic phosphate have been shown to inhibit the production of several secondary metabolites (Bibb, 2005;Hobbs et al., 1992;Martin, 2004).Accordingly, we detected an increase of some metabolites during FAM-growth, but we also detected a decreased production of several metabolites.Microdiffusion experiments in soils have revealed that there is a significant flux of organic nitrogen (Inselsbacher & Nasholm, 2012;Inselsbacher, Öhlund, Jämtgård, Huss-Danell, & Näsholm, 2011;Leitner et al., 2017).
A nitrogen source in the form of organic ammonium, as e.g., asparagine prevents foraging growth on FAM-plates.It will therefore be of future interest to evaluate the effect of altered nitrogen and phosphate levels as well as organic versus inorganic sources, on foraging growth and metabolome.
The foraging secretome contained 74 proteins with predicted signal peptides.Comparing the foraging secretome with rich growth secretome (Widdick et al., 2006) showed that 75% of foraging secretome was not found during rich growth.This underscores the adaptation that takes place upon nutrient depletion that results in the foraging phenotype.Despite the presence of bldA-dependent proteins in the secretome, a bldA-mutant still displayed foraging growth, suggesting bldA-regulation is not essential for this phenotype.The developmental protein BldKB was identified in the secretome, suggesting a role in foraging.Additionally, the foraging behavior was observed in various Streptomyces species, indicating it may be a common feature across the genus (Labeda et al., 2017;Labeda et al., 2012).Multi locus sequence analysis suggests that Kitasatospora is a distinct genus, and suggests that foraging is not limited to Streptomycetes, since foraging growth was also observed for K. setae.
The study demonstrates that foraging growth conditions affect secondary metabolism.The gained ability of S. coelicolor to inhibit mold-growth together with the overall enhanced potential of other Streptomyces spp. to inhibit gram-positive and gram-negative bacteria suggests that there are also significant alterations of their metabolomes.This shows potential for using FAM growth in screening for active secondary metabolites not produced on rich media.However, the low nutrient levels used for the overlay assays could also influence the target bacteria rendering them more susceptible to the antimicrobial effects of foraging colonies.It is well established that growth conditions, e.g., choice of carbon source, affect secondary metabolism of streptomycetes (Sanchez et al., 2010;van Keulen & Dyson, 2014) and other strategies are also being employed to activate silent BGC (Ochi & Hosaka, 2013;van der Meij, Worsley, Hutchings, & van Wezel, 2017).A recent discovery shows that secondary metabolite production is significantly increased as a result from genetic instability on nutrient rich MSagar (Z.Zhang et al., 2020).We could however not find any support for major mutational deletions during foraging growth when investigating the sectors formed during foraging growth.This further emphasizes that the survival strategies of S. coelicolor differ between depleted and rich growth conditions.
The mechanism by which foraging growth is regulated remains to be elucidated.Identifying the regulatory components would result in understanding the required nutrient deprived conditions.
Importantly, knowledge about its regulation could also be utilized as a tool for manipulating the production of the secondary metabolites.In summary, this study demonstrates that nutrient-deprived growth leads to a distinct growth phenotype and altered metabolome, with enhanced antimicrobial properties.This novel growth behavior and its metabolic implications offer new insights into environmental adaptation and competitiveness of streptomycetes under nutrient deprivation with potential implications in anti-microbial drug discovery.

Additional files
• Figure 1 -figure supplement 1. Nutrient levels during phenotype transition on 0.5xTSA.
• Supplementary table 1. Prediction of signal peptides in the foraging secretome using SignalP 6.0.
• Supplementary table 2. Comparison of the foraging secretome with proteins secreted on rich media (Widdick et al 2006).
• Supplementary table 3. Comparison of cytosolic protein detected during foraging growth and growth on rich media by Widdick et al (2006).
• Supplementary table 4. PANTHER classification of signal peptide containing proteins with regard to protein class.
• Supplementary table 5. PANTHER classification of cytosolic protein with regard to protein class.
• Supplementary table 7. Locally isolates streptomcyetes.where the transition to foraging growth was observed is indicated with the dotted lines on day 21.
Each measurement proximal to the colony and at the edge of the dish were taken in sextuplicate for each time point and normalized against the controls.Source file of nutrient fold change over time in edge and proximal samples used for the plots is available in the Figure 1 -Source data 2. coelicolor grown on FAM and 0.5xTSA for 7 days.The glycerol storage buffer was washed away and replaced with water.A top view of the center of the colonies and at the edge show few hyphae at the surface of the FAM.Hyphae extending through the agar are mainly observed on the FAM-plates, while they predominantly grow on top on 0.5xTSA.

Figure supplement legends
8 µm VanGuard precolumn (Waters Corporation, Milford, MA, USA) held at 40°C.The gradient elution buffers were A (H2O, 0.1 % formic acid) and B (75/25 acetonitrile:2-propanol, 0.1 % formic acid).The compounds were detected with an Agilent 6550 Q-TOF mass spectrometer equipped with a jet stream electrospray ion source.The settings were kept identical between the modes, except the capillary voltage.The data processing was performed using the Recursive Feature Extraction algorithm within Agilent Masshunter Profinder version B.08.00 (Agilent Technologies Inc., Santa Clara, CA, USA).All multivariate statistical investigations (PCA) were performed using the software package SIMCA®-P+ version 15.0.2 (Umetrics, Umeå, Sweden).Annotation of unknown metabolites were performed by comparing monoisotopic molecular weights with S. coelicolor metabolites in the StreptomeDB (Moumbock et al., 2021) using R software (4.2).Metabolomic features were plotted using GraphPad Prism (version 9.0.2).All raw mass spectrometry files used in this study have been deposited to ScieLifeLab Data Repository with the dataset identifier DOI: https://doi.org/10.17044/scilifelab.c.7171122.Average nucleotide identity of streptomycetesFastANI(Jain, Rodriguez, Phillippy, Konstantinidis, & Aluru, 2018) version 1.33 was employed to calculate the Average Nucleotide Identity (ANI) values, assessing the genomic similarity among the 13 species investigated in this study (GenBank assembly IDs: GCA_000010605.1were analyzed using FastANI with its default settings.Pairwise comparisons were conducted across all genomes, and the resulting ANI values were utilized to generate a heatmap using R version 4.2.

Figure 2 -
Figure 2 -figure supplement 1. Sector mutations of foraging colonies.Spores of S. coelicolor were

Figure 4 -
Figure 4 -figure supplement 1.Growth of Streptomyces on FAM.(A) Growth of S. coelicolor wild-

Figure 5 .
Figure 5. Mold inhibitory activity and metabolomic analysis of foraging S. coelicolor.