Nutrient Availability Shifts the Biosynthetic Potential of Soil-Derived Microbial Communities

Secondary metabolites produced by microorganisms are the main source of antimicrobials and other pharmaceutical drugs. Soil microbes have been the primary discovery source for these secondary metabolites, often producing complex organic compounds with specific biological activities. Research suggests that secondary metabolism broadly shapes microbial ecological interactions, but little is known about the factors that shape the abundance, distribution, and diversity of biosynthetic gene clusters in the context of microbial communities. In this study, we investigate the role of nutrient availability on the abundance of biosynthetic gene clusters in soil-derived microbial consortia. Soil microbial consortia enriched in high sugar medium (150 mg/L of glucose and 200 mg/L of trehalose) had more biosynthetic gene clusters and higher inhibitory activity than those enriched in low sugar medium (15 mg/L of glucose + 20 mg/L of trehalose). Our results demonstrate that experimental microbial communities are a promising tool to study the ecology of specialized metabolites.


Introduction
The chemical products of microbial secondary metabolism (also called natural products) modulate interactions within and between species and are thus a major means through which the microbial world communicates [1]. Secondary metabolites have had an enormous impact on modern medicine: they are the main source of antimicrobials used to treat infections, they are used as therapeutics for cancer and other important human diseases, and they are used as immunosuppressants that enable life-saving transplantation surgeries [2]. Soil microbes have been the primary discovery source for these secondary metabolites, often producing complex organic compounds with specific biological activities [3,4]. The enzymes that assemble microbial natural products are encoded by genes located in biosynthetic gene clusters (BGCs). While it is widely hypothesized that secondary metabolism broadly shapes microbial ecological interactions, little is known about the factors that influence the abundance, distribution, and diversity of biosynthetic gene clusters in the context of microbial communities [5].
Studies that compare the biosynthetic potential of microbial communities sourced from different soils have unclear extrapolative value and little is known of how environmental factors can contribute to enrichment for secondary metabolism on finer scales [4]. Comparisons between United States soil communities from New Hampshire and Arizona suggest that the arid desert soils of Arizona may harbor more antagonistic, inhibitory compounds than the forest soils of New Hampshire [6]. They observed a diversity of Nonribosomal peptide synthetases (NRPSs) and polyketide synthases (PKSs) domains in arid soils when compared with forest soils. One hypothesis is that this is due to the harsh, nutrientpoor conditions of the soil that may lead to increased pressures on nutrient acquisition and/or other means of competition. BGCs and their producing organisms are found in almost every known microbial niche and covary with some environmental factors [7]. For example, even within soils sourced from the same rhizospheres, biosynthetic capacity shifts depending on soil depth [3]. However, these studies are often limited to correlative descriptions subject to sampling biases and systematic assessments with sufficient experimental control remain lacking [4].
Many bacteria dedicate large portions of their genomes to BGCs, sometimes in excess of 25% of all genetic material [8]. BGCs are often maintained vertically over evolutionary 1 3 64 Page 2 of 7 timescales [9,10], implying that they are adaptively important in their natural settings [9,11]. Further, the producers of secondary metabolites often do not express their BGCs under typical laboratory conditions. A growing body of evidence suggests many biotic and abiotic cues can elicit the production of different secondary metabolites [4,12]. Recently, Hurley et al. [13] observed that the taxonomy and inhibitory profile of the bacteria isolated from four United States soil samples used within the Tiny Earth project is strongly affected by the selective media used. Potato dextrose agar (PDA) enriched for strains that inhibited Acinetobacter baylyi and Pseudomonas putida, while tryptic soy agar (TSA) enriched for Erwinia carotovora-inhibiting strains. Understanding the link between secondary metabolisms and nutrient availability has fundamental implications across microbial ecology, including the ecology of antagonism, community maintenance, invasion, niche construction, and niche defense. The aim of this study was to evaluate if carbon source availability can affect the biosynthetic potential of enriched microbial communities.  [14]. The inoculated flasks were incubated at 30 ℃ and 225 rpm. Optical density at 600 nM was measured every 24 h to monitor growth. Microbial consortia grown in the LS medium were named PCA1-3 and microbial consortia grown on HS medium were named PCB1-3. Once the stationary phase had been reached, the cultures were centrifuged and the cell pellets were frozen at − 80 ℃. Supernatants were filter sterilized using 0.22 μM Durapore® membranes (Millipore) and kept refrigerated to be used with certain microbes in a 96-well growth assay to determine antimicrobial potential.

Inhibition Assays
To determine if the metabolites produced in the six cultures at the stationary phase had antimicrobial properties, we measured the inhibitory activity of the spent media against a panel of selected bacteria. The bacterial target panel selected was compromised of bacterial species Pseudomonas fluorescens (ATCC13525), Klebsiella pneumoniae (ATCC23357), Bacillus subtilis (ATCC6051), and Salmonella typhimurium LT2 (Nickerson-Arizona State). Target bacteria were grown on a 96-well plate containing 10 μL sterile Thermo Scientific Iso-Sensitest broth + 90 μL microbial community supernatant in quadruplicate for each treatment. Growth was measured using optical density at 600 nM via a BioTek plate reader and Gen5 software. Readings were taken at time points 0, 24, and 48 h after inoculation and plates were grown at 37 ℃ and 26 ℃ depending on which bacterial species was being tested (P. fluorescens and B. subtilis at 26 ℃, K. pneumoniae and S. typhimurium at 37 ℃). Controls were frown with 90 μL of sterile M63 minimal medium with carbon source + 10 μL sterile Thermo Scientific Iso-Sensitest. The optical density data measured from the control were compared to the data from each respective microbial community supernatant to determine if the community supernatants contained metabolites with inhibitory activity against the pathogen panel.

Metagenomic DNA Extraction, Sequencing, and Analysis
DNA extraction and purification were performed for all six samples using the QIAamp BiOstic Bacteremia DNA Kit and protocol (QIAGEN). The purified sample DNA was sequenced at the Microbial Genome Sequencing Center (MiGS) core facility in Pittsburgh on the Illumina NextSeq platform. Sequence reads were filtered and trimmed using the default settings of fastp (Chen et al., 2018). Filtered reads were taxonomically classified using the Kaiju software using the NCBI BLAST nr + euk database [15]. A co-occurrence network was built using the SparCC [16] program implemented in the MicrobiomeAnalyst platform [17] based on the Spearman correlation between genus distributions across the datasets. Metagenomes were assembled using SPAdes 3.13.0 [18]. Biosynthetic gene clusters (BGCs) were annotated with antiSMASH 5.0 using the default settings [19]. Contigs with BGCs were taxonomically classified using a Last Common Ancestor (LCA) approach implemented in the Contig Annotation Tool (CAT) [20]. The short reads of the metagenome datasets used in this study were deposited in the NCBI Short Read Archive (SRA) accession numbers from SRR15633242-SRR15633247.

Results and Discussion
Soil microbial communities were enriched in minimal medium containing either 15 mg/L of glucose + 20 mg/L of trehalose (LS) or 150 mg/L of glucose and 200 mg/L of trehalose (HS). Supernatants of HS communities inhibited the growth of B. subtilis (t = 5.413, df = 3, P = 0.012), K. pneumoniae (t = 3.846, df = 3, P = 0.031), P. fluorescens (t = 9.565, df = 3, P = 0.002) and S. typhimurium (t = 5.249, df = 3, P = 0.13). LS supernatants inhibited the growth of K. pneumoniae (t = 4.65, df = 3, P = 0.019) and S. typhimurium (t = 4.801, df = 3, P = 0.017) (Fig. 1). Some studies have found that high concentrations of nutrients (e.g., glucose) can inhibit the production of secondary metabolites in some bacteria [21], while others find that different carbon sources can influence the production of secondary metabolites in taxon-specific ways [22]. To our knowledge, our study is the first to evaluate the effect of glucose and trehalose concentrations on the inhibitory activity of undefined microbial consortia. While this approach does not enable the identification of the major producers of inhibitory molecules, it can shed light on how nutrient availability influences interspecies interactions. The number of metagenomic reads after quality filtering and trimming ranged from 9 to 10 million per sample. Metagenomic assemblies ranged from 51 to 390 Mb (Supplementary Table 1). Taxonomic classification of reads indicated that the HS microbial communities were dominated by fungi, while LS microbial communities were   (Fig. 2a). Previous studies have evaluated the effect of nutrient availability on microbial community diversity, abundance, and composition [23]. For example, experiments with soils amended with glucose have found high concentrations (> 8 mg C/g) favored the growth of fungi over bacteria, potentially attributed to differences between the kingdoms in optimal osmotic potential [24].
Functional annotation of the contigs in Clusters of Orthologous Genes (COGs) categories revealed that genes in the "Chromatin structure and dynamics" (t = 8.214, df = 4, P = 0.007) and "Secondary metabolites biosynthesis, transport, and catabolism" (t = 5.915, df = 4, P = 0.018) categories were overrepresented in the HS microbial communities (Fig. 3). "Signal transduction mechanisms" genes (t = 4.156, df = 4, P = 0.021) were overrepresented in the LS microbial communities (Fig. 3). Most biosynthetic gene clusters (BGCs) were overrepresented in the HS microbial communities. Type I Polyketide Synthase (T1PKS) and Nonribosomal Peptide Synthetase (NPRS) biosynthetic gene clusters were more abundant in the HS communities, while BGCs encoding bacteriocins and siderophores were more abundant in the LS communities (Table 1). Correlation analyses of microbial genera in the communities revealed many negative correlations between bacteria belonging to the Enterobacteriaceae family (e.g., Salmonella) and the two most abundant fungal genera (Trichoderma and Fusarium; both belonging to the Sordariomycetes class) (Fig. 4a). This is consistent with the results obtained in inhibition assays (see below). Most BGCs in the HS microbial communities were classified to the Sordariomycetes taxonomic class, while most of the BGCs in the LS microbial communities were classified to the class Actinobacteria (Fig. 4b). Many members of these taxonomic classes are prolific producers of natural products [25,26] indicating that our approach enriches for different taxa of antibiotic-producing microorganisms.
Microbial communities with more BGCs showed more inhibitory activity toward the bacterial pathogens used in this study. Traditionally, inhibition assays and other screening assays for antimicrobial activity are performed with axenic cultures and the use of mixed cultures for antibiotic discovery is still relatively new [27,28]. Co-culture is an effective approach to promote the activation of silent biosynthetic gene clusters and can facilitate the discovery of new natural products [29,30]. In this study, we demonstrate that enriched microbial communities derived from environmental, complex communities (e.g., soil and feces) could be screened for the ability to produce novel antibacterial compounds.

Conclusion
In conclusion, we show that laboratory-enriched microbial communities are a promising tool to study the ecology of secondary metabolites. High and low sugar conditions can enrich for microbial communities with different inhibitory phenotypes and different biosynthetic gene contents. This suggests that nutrient availability in nature may be linked to the diversity and distribution of secondary metabolites within microbial communities, including BGCs that produce molecules involved in interspecies antagonism. Future studies involving chemical libraries, metatranscriptomic, and metabolomic approaches will contribute to our understanding of the environmental and nutritional conditions that are favorable for the production and/or selection of novel secondary metabolites. Funding This research did not receive any specific grant from funding agencies in the public, commercial, or nonprofit sectors. Support for M.G.C. was provided by grant 2020-67012-31772 (Accession  PCA1  PCA2  PCA3  PCB1  PCB2  PCB3   NRPS  15  13  12  58  55  91  NRPS-like  5  5  6  23  32  22  Terpene  21  16  8  16  13  17  T1PKS  1  1  1  32  15  27  Bacteriocin  7  9  6  0  0  0  Siderophore  3  6  3  0  2  0  NRPS|T1PKS  0  1  0  4  3  3  Other  12  14  15  6