Strains, media, and culture conditions

An overview of all strains, plasmids, and cosmids used can be found in Table A in S1 Text. Streptomyces venezuelae NRRL B-65442 was grown on solid maltose-yeast extract-malt extract (MYM) medium [53] at 30°C for 2 days for spore stock generation. Spore stocks for Streptomyces coelicolor M145, Streptomyces avermitilis K139 [54], WAC303 [55], and WAC288 [55] were similarly prepared with incubation times of up to 4 days. Bacillus subtilis (PY79) BDR2662 i::PrpsJ-mCherry was kindly provided by Prof. D.Z. Rudner and cultured in lysogeny broth (LB) unless stated otherwise. Escherichia coli DH5α, Pseudomonas aeruginosa PAO1, and Rhodococcus jostii RHA1 were cultured in LB unless stated otherwise.

Growth on whole grasshoppers was performed by adding 10 mL of autoclaved tap water to a single autoclaved grasshopper (Fresh Feeders—Exotic Nutrition) placed in a petri dish. Spores were added at a density of 10⁶ spores per mL, and petri dishes were incubated statically at 30 °C.

For liquid cultures, bacteria were typically inoculated at a density of 10⁶ spores/mL unless stated otherwise. The minimal chitin medium contained, per liter: 5 g of chitin flakes derived from shrimp (CAS: 1398-611-4), 0.6 g of MgSO₄·7H₂O, 2 g of (NH₄)₂SO₄ (when specified), 150 mL of NaH₂PO₄/K₂HPO₄ buffer (0.1 M, pH 6.8), 1 mL of trace elements solution [53], and 850 mL of tap water. The minimal exoskeleton medium contained 5 g of crushed grasshopper exoskeleton (Fresh Feeders - Exotic Nutrition), which replaced the chitin in the minimal chitin medium. Glucose medium consisted of 5 g of d-glucose and always included 2 g of (NH₄)₂SO₄ as a nitrogen source, replacing the chitin in the minimal chitin medium. Cultures were generally grown in 5 mL volumes within 25 mL glass tubes and incubated at 225 RPM and 30°C. The raw data for all measurements described herein are contained in S2 Data.

Microscopy

Phase-contrast images were taken with a Zeiss Axio Lab 5 upright microscope, equipped with a Zeiss Axiocam 208 color camera. Morphological studies on S. venezuelae grown on grasshoppers by Scanning Electron Microscopy were performed using a prisma environmental scanning electron microscope by Thermo Scientific. Pieces of grasshopper with bacterial biomass were cut and fixed with 1.5% glutaraldehyde (1 h). Subsequently, samples were dehydrated (70% acetone for 15 min, 80% acetone for 15 min, 90% acetone for 15 min, and 100% acetone for 15 min), and critical point dried. Hereafter, the samples were coated with gold using a sputter coater before imaging.

Resazurin assays

Resazurin assays were used to measure metabolic activity. Spores were inoculated in either chitin or glucose medium containing resazurin (CAS:62758-13-8) (15 μg/mL). One milliliter was transferred to a 48-well-plate and covered with a Breathe-Easy sealing membrane. The plate was then incubated at 30 °C with low orbital shaking in a Synergy plate reader for 12 h. Emission at 590 nm was measured every 10 min (excitation: 550 nm). The curves represent the mean of three independent 1 mL cultures for each time point.

RNA sequencing

S. venezuelae was grown in triplicate in 20 mL minimal glucose medium and 20 mL minimal chitin medium, both supplemented with (NH₄)2(SO₄) for 24 h. 5 mL of each culture was pelleted, frozen in liquid nitrogen, and kept at −80 °C before RNA extraction. RNA was extracted from the cell pellets using the Qiagen RNeasy Micro Kit (Qiagen, no. 74004) according to the manufacturer’s protocol, with a minor modification: 0.5 mL of 0.5 mm glass beads were added to samples in lysis buffer followed by bead beating (six times 1 min beating versus one minute on ice). RNA integrity was verified by gel electrophoresis before being sent for sequencing at Genome Quebec. There, total RNA was quantified and its integrity was assessed using a LabChip GXII (PerkinElmer) instrument. rRNA was depleted from 250 ng of total RNA using QIAseq FastSelect (-5S/16S/23S Kit 96rxns). cDNA synthesis was achieved with the NEBNext RNA First Strand Synthesis and NEBNext Ultra Directional RNA Second Strand Synthesis Modules (New England BioLabs). Library preparation was then completed using the NEBNext Ultra II DNA Library Prep Kit for Illumina (New England BioLabs). Adapters and PCR primers were purchased from New England BioLabs. Libraries were quantified by Kapa Illumina GA with Revised Primers-SYBR Fast Universal kit (Kapa Biosystems). Average DNA fragment sizes were determined using a LabChip GXII (PerkinElmer) instrument. The libraries were normalized, pooled, and then denatured in 0.05 N NaOH and neutralized using hybridization buffer (HT1). The pool was loaded at 200 pM on an Illumina NovaSeq S4 lane using the NovaSeq Xp Workflow as per the manufacturer’s recommendations. The run was performed for 2 × 100 cycles (paired-end mode). A phiX library (Illumina) was used as a sequencing control and mixed with libraries at 1% level. Base calling was performed with RTA v3. Program bcl2fastq2 v2.20 was then used to demultiplex samples and generate fastq reads.

RNA sequencing reads were aligned to the S.venezuelae genome assembly (GCA_001886595.1_ASM188659v1) using the STAR aligner [56]. First, we indexed the genome assembly using “STAR genomeGenerate” with “—genomeSAindexNbases 8” to support bacterial genomes [56]. We then used “STAR alignReads” to align paired end reads with “--alignIntronMin 20 --alignIntronMax 19 --outFilterMultimapNmax 20” parameters which prevents STAR from incorporating introns and gap junctions into alignments, thereby making STAR compatible with bacterial genomes [56]. Quality control of raw aligned reads were summarized using fastqc [57]. We the then filtered PCR duplicates with “Picard MarkDuplicates” using VALIDATION_STRINGENCY=LENIENT [58] and filtered reads with a quality score <30 with “samtools view” [59]. To view these reads on the Interactive Genome Viewer (IGV) [60], we counted the reads per kilobase per million (PRKM) normalized alignments and converted the alignment “bam” file to a “bigwig” file using DeepTools “bamCoverage” [61]. These filtered reads were then assigned to genes in the “GCA_001886595.1_ASM188659v1_genomic” genome annotation using featureCounts “-s 1 -Q 30” to allow for high-quality strand-specific alignments [62]. Filtered and counted RNA-seq reads were then subjected to data normalization. First, we filtered rRNA genes, which account for a high proportion of sequenced reads and can impact downstream analysis. Then, we RPKM normalized the remaining counts using the edgeR R package [63]. Genes with 5 or fewer reads in at least 2 of the 3 biological replicates were excluded. Gene-expression between samples was measured using Pearson’s correlation and plotted into a heatmap with the Pheatmap R package (https://cran.r-project.org/web/packages/pheatmap/pheatmap.pdf) (Fig A in S1 Text). Then, we performed Principal Component Analysis (PCA) between samples and plotted the first two PCs with ggplot2 (Fig B in S1 Text). Differential expression analysis between glucose and chitin conditions were performed with edgeR [63], with gas the “control” condition. Differentially expressed genes (DEGs) were then plotted using a Manhattan plot using ggplot2. Raw sequencing data is included in S1 Data.

Chitinase annotation and architecture

We initially identified six proteins as putative GH18 family chitinases (vnz_02735; vnz_05055; vnz_23100; vnz_24855; vnz26400; vnz35060) and one GH19 family chitinase (vnz_16685) by using the Streptomyces Annotation Server (https://strepdb.streptomyces.org.uk). We used the catalytic domains of these putative chitinases as queries for BLAST searches against the S. venezuelae genome and this led to the identification of three additional putative GH18 family chitinases (vnz_12680; vnz_23445; vnz_12765). To predict their enzyme architecture, we utilized UniProt’s integrated databases (InterPro; PROSITE; Pfam; SUPfam) and identified putative binding domains, catalytic domains, and putative transmembrane regions (Table B in S1 Text), which were then structurally examined using the alpha fold plug-in.

Phylogeny

We conducted phylogenetic analyses using Molecular Evolutionary Genetics Analysis 11 (MEGA11) software [64]. We performed an alignment of the amino acid sequences for the catalytic domains of the nine GH18 family chitinases present in the S. venezuelae genome, and these same chitinases together with 15 previously identified chitinases from the GH18 family from various bacteria [65,66], utilizing the MUSCLE plug-in with its default settings. We applied the maximum likelihood algorithm, coupled with the Jones-Taylor-Thornton substitution model, to deduce the unrooted phylogenetic trees. To evaluate the confidence in the branching patterns of the resulting trees, we performed a bootstrap analysis with 500 replicates. The chitinases from S. venezuelae were categorized into subfamilies based on their phylogenetic relationships with the previously characterized chitinases.

Chitinase conservation

We conducted a protein BLAST (BLASTP) search using full-length protein sequences of all the identified S. venezuelae chitinases, recombinase (vnz_26845), and vnz_04420 (gene in chloramphenicol biosynthetic gene cluster) to assess conservation across bacterial taxa, with a particular focus on the Streptomycetaceae family. The BLAST search was performed against the NCBI protein database using stringent thresholds: query coverage of 80%–100%, sequence identity of 50%–100%, and an E-value of ≤1E−05. To ensure comprehensive retrieval of homologous sequences, the Maximum Target Sequences parameter was set to 5,000.

The total number of hits and the proportion of hits from the Streptomycetaceae family for each query sequence are summarized in Table E in S1 Text.

Construction of S. venezuelae mutants

Gene deletions were generated using ReDirect technology [67]. In short, coding sequences on a cosmid vector carrying large fragments (30–40 kb) of S. venezuelae genomic DNA were replaced by an oriT-containing apramycin resistance cassette (PCR amplified from pIJ773) with primers as indicated in Table C in S1 Text. The mutant cosmids were introduced into the methylation-deficient Escherichia coli strain ET12567 carrying the conjugation helper plasmid pUZ8002, followed by conjugation into S. venezuelae. Exconjugants were screened for double-crossover events, and gene deletions were confirmed by PCR using various combinations of primers positioned to confirm the presence of the apramycin resistance cassette and absence of the wild-type gene (Table C in S1 Text).

Mutant strains were complemented by expressing the wild-type gene in trans, after cloning them into the integrating vector pMS82 [68] behind the constitutive ermE* promoter to ensure high levels of transcription [69] using restriction enzymes NdeI (Promega) and XhoI (Promega) and T4 ligase (New England BioLabs). Primers are listed in Table C in S1 Text.

Colony-forming unit (CFU) assays for Streptomyces

Spores were inoculated in 5 mL of minimal chitin medium or MYM liquid medium in 25 mL glass tubes to minimize shear stress and facilitate chitin attachment. Cultures were incubated at 30 °C with shaking at 225 rpm. After three days of incubation in MYM medium and six days in chitin medium, cultures were vortexed three times for 10 seconds each, and 10-fold serial dilutions were prepared. Two microlitre spots were inoculated to MYM agar and incubated overnight. CFU counts were determined from the highest dilution factor that supported growth. Bar graphs represent the averages of three independent cultures.

Liquid chromatography–mass spectrometry (LC–MS)

To detect chitobiose in culture supernatants, spent medium from 4-day cultures of wild-type S. venezuelae and the dasBC deletion strain were filtered and dried. The dried spent medium was resuspended in half the original volume in LC–MS grade water. One-hundred microlitres was injected into an Acquity UPLC BEH C18 column (1.7 µm, 2.1 × 50 mm) in parallel with blank controls. The separation gradient was at 0.2 mL/min, 95% A for 1 min, increasing linearly to 60% A over 22 min, to 5% A at 24 min and held at 5% for 2 min: A—water + 0.1% formic acid and B—acetonitrile + 0.1% formic acid. Samples were analyzed by electrospray ionization with MS/MS on selected masses (447.16; 425.18; 244.08) on a QTOF (Waters Xevo G2S-QTOF).

S. venezuelae—B. subtilis co-cultivation

B. subtilis BDR2662 sacA::PrpsJ-mCherry was kindly provided by Prof. D. Rudner. Bacteria were cultured overnight at 30 °C with shaking at 225 rpm. The overnight culture was then diluted 1:50 into MYM or minimal chitin medium with (NH₄)2(SO₄). Similarly, 10⁶ spores/mL of either wild-type S. venezuelae or the dasBC deletion mutant were added to generate a co-culture. One milliliter of each condition was transferred to a 48-well plate and covered with a Breathe-Easy sealing membrane. The plate was incubated at 30 °C with low orbital shaking in a Synergy plate reader for 16–48 h, measuring fluorescence emission at 620 nm every 10 min (excitation: 550 nm). The curves represent the mean of three 1 mL cultures for each time point. To determine the CFUs of B. subtilis in co-cultures, a 1:1000 dilution of an overnight B. subtilis culture was added to 1 mL of MYM or chitin medium containing (NH₄)₂(SO₄) with 10⁶ spores/mL of either wild-type S. venezuelae or the dasBC deletion mutant in a 48-well plate. Cultures were incubated as static cultures for 3 days, followed by preparation of 10-fold serial dilutions. A 5 μL sample from each dilution was spotted onto LB agar and incubated overnight. CFU counts derived by determining the highest dilution factor that supported growth. Bar graphs represent the averages of three independent cultures.

Assessment of S. venezuelae spent medium toxicity on B. subtilis

B. subtilis BDR2662 sacA::PrpsJ-mCherry (PY79) was cultured overnight at 37 °C with shaking at 225 rpm. The overnight culture was then diluted 1:1000 into LB. To every 450 microliters of LB and B. subtilis cells, 50 μL of 10× concentrated S. venezuelae spent medium was added (see below). For each culture, 0.5 mL were transferred to a 48-well plate and covered with a Breathe-Easy sealing membrane. The plate was incubated at 37 °C with high double-orbital shaking in a Synergy plate reader for 16–48 h, measuring fluorescence emission at 590 nm every 10 min (excitation: 550 nm). The curves represent the mean of three 0.5 mL cultures for each time point. Concentrated spent media samples were prepared from 3 mL of spent medium of S. venezuelae wild-type and dasBC deletion strains grown on chitin for 16 h, which were filtered and evaporated. The resulting residue was dissolved in 300 μL milliQ water. Spent media concentrates were prepared in triplicate.

Streptomyces can grow using chitin as sole source of carbon and nitrogen

We carried out two experiments to determine whether Streptomyces venezuelae could grow using raw chitin as carbon and nitrogen source. First, we applied spores to sterilized grasshopper carcasses. We observed vigorous growth, with the insect being covered by a dense hyphal mat within 2–3 days (Fig 1A). We imaged this growth using electron microscopy and observed a dense mycelium with individual hyphae appearing to grow along the length of the exoskeleton (Fig C in S1 Text). We measured the grasshopper masses before and after 6 weeks of incubation with S. venezuelae spores, normalizing to an uninoculated control insect. The combined mass of the remaining material and the Streptomyces hyphal mat showed a 27% decrease relative to the insect prior to incubation (Fig 1B), which could be due to consumption and CO₂ release during respiration by S. venezuelae.

Download:

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

Fig 1. Streptomyces venezuelae growth on insect (-exoskeleton) and chitin.

A) S. venezuelae overgrowing a grasshopper, visualized by an extensive hyphal mat covering the exoskeleton of the insect. B) Growth of S. venezuelae (S. vnz) lead to a 27% loss of grasshopper biomass after an incubation period of 6 weeks. Bars represent the averages of eight biological replicates and the error bars show standard deviations. C) S. venezuelae hyphal mass coating a piece of insect exoskeleton D) S. venezuelae growing on a piece of chitin. E) Growth on chitin supports sporulation by S. venezuelae, illustrated by the green spore pellet (bottom panel—arrow) and the observation of spores by microscopy (top panel). F) Growth of S. venezuelae (S. vnz) on chitin results in biomass loss of over 50% after an incubation period of six weeks. Bars represent the averages of five biological replicates and the error bars show standard deviations. G) The metabolic marker resazurin reaches its peak in fluorescence earlier for cultures grown on chitin compared with glucose. Insert: bar graph showing CFUs per milliliter after 6 days of growth on either chitin medium or glucose medium. Scalebars: C and D: 50 microns. E: 5 microns. The data underlying this figure can be found in S2 Data.

https://doi.org/10.1371/journal.pbio.3003292.g001

The grasshopper exoskeleton is composed primarily of chitin in addition to proteins and other carbohydrates [70], which are all potential sources of carbon and nitrogen to Streptomyces. To investigate whether the exoskeleton alone would support bacterial growth we dried the grasshoppers extensively and dissected away the intra-abdominal material to liberate exoskeleton. We then crushed the dissected exoskeleton, sterilized the resulting flaky powder, and added this to sterile phosphate buffer (pH 6.8) and trace elements, resulting in an exoskeleton medium. Again, S. venezuelae was able to grow efficiently, illustrated by the formation of aggregates consisting of exoskeletal fragments and tightly-associated hyphae within 2 days (Fig 1C). Together, these results show that Streptomyces can grow using dead insect exoskeleton as sole nutrient, supporting their status as saprophytes.

Second, we created a defined medium consisting of commercially sourced, practical grade α-chitin flakes from shrimp, suspended in phosphate buffer and trace elements. Colloidal chitin as well as additional carbon sources, such as agar, amino acids, or yeast extract, were deliberately excluded [40,41,44–49,71]. Again, we found that S. venezuelae could grow efficiently on this defined medium (Fig 1D) and completed its life cycle with the production of 10⁸ spores per mL after 5–6 days (Fig 1E and inset 1G). Moreover, we found that in inoculated static chitin cultures, over 50% of the weight had escaped the system after 6 weeks of incubation, again likely due to CO₂ release during respiration (Fig 1F). Note that these measurements include the Streptomyces biomass, suggesting that a greater proportion of chitin had been metabolized.

To determine if this trait is specific to S. venezuelae, we tested four other Streptomyces species for growth in this restrictive medium: the evolutionarily diverged Streptomyces coelicolor and Streptomyces avermitilis, alongside the environmental isolates WAC288, and WAC303 from the Wright Actinomycetes collection [55]. All four species generated substantial biomass after 5 days (Fig D in S1 Text). In contrast, E. coli, Pseudomonas aeruginosa, Bacillus subtilis, and Rhodococcus jostii, all of which are common soil inhabitants, were unable to grow on this medium (Fig E in S1 Text). This indicates that the growth observed for the diverse Streptomyces species is not a universal property of bacteria, including soil dwellers.

Optical density was not a feasible means of quantifying S. venezuelae growth under these conditions due both to the filamentous nature of Streptomyces cells, and the presence of light scattering chitin flakes. We therefore used resazurin, which is taken up by metabolically active cells and reduced by cellular respiratory dehydrogenases to resorufin, which fluoresces at 590 nm [72]. In this way, we compared the metabolic activity of S. venezuelae in chitin medium with cultures using glucose and ammonium as carbon and nitrogen sources (see materials and methods). We found that the cells reached their peak reducing potential in the chitin medium 25% faster than in the glucose medium (Fig 1G). Furthermore, growth on the defined glucose medium did not support efficient spore formation. This is supported by CFU measurements after six days of growth: chitin-grown cultures yield approximately 10⁸ spores per milliliter, whereas glucose-grown cultures reach only 10⁶ to 10⁷ CFUs per milliliter (Fig 1G, inset).

Collectively, these data show that S. venezuelae readily grows and completes its life cycle using raw chitin as primary source of carbon and nitrogen. These observations suggest that Streptomyces have not only incredibly specialized metabolic capabilities, but have also evolved a remarkable primary metabolism, that enables them to take advantage of this abundant but insoluble nutrient source.

Growth on chitin results in a distinct gene expression profile

To determine how growth on chitin influences gene expression, we carried out an RNA sequencing experiment, comparing growth in liquid minimal chitin medium to growth in liquid minimal glucose medium. In this instance, we added ammonium to both growth conditions to focus primarily on the use of chitin (versus glucose) as a source of carbon. Cultures were harvested after 24 h of growth to allow for the accumulation of sufficient amount of biomass for both growth conditions. RNA was extracted, followed by the preparation of cDNA libraries, which were then sequenced.

The transcript profiles for growth on glucose and chitin were significantly different, with 2,681 out of 7,462 genes exhibiting changes of 2-fold or greater (FDR ≤ 0.05) (Fig 2A and S1 Data). Out of those, 1,460 genes had significantly higher transcript levels during growth on chitin.

Download:

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

Fig 2. Transcriptional response of Streptomyces venezuelae during growth on chitin.

A) Manhattan plot of S. venezuelae on glucose vs. chitin after 24 h of growth. Each dot represents a gene. Red dots are significantly differentially expressed with an FDR cutoff of 0.05. Genes with a positive Log2 fold change are upregulated during growth on chitin. Orange dots represent the 40 genes that are most significantly upregulated (lowest False Discovery Rate) during growth on chitin (Table D in S1 Text). Genes that were further investigated in this study are labeled. B) The 40 genes that are most significantly upregulated during growth on chitin (orange dots) have been assigned a Cluster of Orthologues Genes (COG) group. C) Phylogeny, domain architecture, and expression of S. venezuelae chitinases. The genome encodes one GH19 and nine GH18 chitinases with various chitin-binding domains. Subfamily assignments are shown. The first number indicates log₂ fold change (chitin vs. glucose); the second shows log₂ counts per million. The data underlying this Figure can be found in S1 and S2 Data. Please open S3 Data to view the tree-file.

https://doi.org/10.1371/journal.pbio.3003292.g002

The annotations of the 40 genes with the lowest false discovery rates and the highest comparative transcript levels during growth on chitin are listed in Table D in S1 Text and shown as groups of orthologues genes (COGs) (Fig 2B). We found that the majority of genes could either be attributed to groups C, G, and S, which correspond to “Energy production and conversion” (C), “Carbohydrate metabolism and transport” (G), and “Function unknown” (S), respectively. As we previously observed stronger reducing power in cells grown on chitin (Fig 1G), we examined the genes within the “energy production and conversion” group more closely. We found that this group primarily comprised genes encoding subunits of the NADH-oxidoreductase complex (complex I) (Table D in S1 Text), a component of the respiratory chain in both bacteria and eukaryotes. In S. venezuelae, complex I comprises 14 core subunits (vnz_21055–vnz_21130), which is standard in bacteria. In addition, the S. venezuelae genome encodes 9 subunit homologs (vnz_21265–vnz_21310), which is uncommon for bacteria [73]. Notably, the transcript levels of the homologs were unchanged between chitin and glucose, while those of the 14 core subunits were significantly higher during growth on chitin (Table D in S1 Text). This suggested that metabolic energy production is altered in bacteria grown on chitin.

Classification and functional specialization of S. venezuelae chitinases

We next investigated the genes in group G for “carbohydrate metabolism and transport”, aiming to identify enzymes involved in chitin catabolism. Notably, this group, consisting of 11 genes, was predominantly composed of chitinase-annotated genes, which accounted for 7 of them (Table D in S1 Text). We queried the whole S. venezuelae genome and identified at least 10 putative chitinases in the S. venezuelae genome (see materials and methods). Of these, one had been previously investigated for its antifungal activity [74]; however, aside from that, none of them has been characterized.

We analyzed the protein architecture, phylogeny, conservation, and expression patterns of all putative S. venezuelae chitinases. Most bacterial chitinases have a catalytic domain belonging to glycoside hydrolase family 18 (GH18), which hydrolyzes the β-1,4 linkages between the N-acetylglucosamine (GlcNAc) residues. We found that nine S. venezuelae chitinases belong to the GH18 family, while one belongs to the glycoside hydrolase family 19 (GH19), which is primarily associated with plant chitinases but had previously been reported in Streptomyces too [75]. Eight of these chitinases (vnz_02735; vnz_05055; vnz_12680; vnz_16685; vnz_23445; vnz_24855; vnz_26400; vnz_35060) have signal peptides indicating that they are secreted proteins, while the remaining two chitinases (vnz_12765 and vnz_23100) contain hydrophobic residues at the N-terminus of the protein, suggesting that they are possibly membrane-anchored. The seven chitinase hits from the RNA sequencing data (vnz_05055; vnz_12765; vnz_23100; vnz_23445; vnz_24855; vnz_26400; vnz_35060) are all members of the GH18 family.

GH18 family chitinases can be further divided into subfamilies A, B, and C [65,76], on the basis of their function and evolution: members of GH18 subfamily A are usually processive chitinases that release GlcNAc or chitobiose (a dimer of GlcNAc) from the chitin polymer. In contrast, subfamily B enzymes are primarily endo-acting chitinases that cleave the polymer into oligosaccharides. Finally, subfamily C chitinases have a distinct evolutionary lineage. By constructing an unrooted phylogenetic tree of known bacterial GH18 chitinases from subfamilies A, B and C together with S. venezuelae GH18 chitinases (Fig F in S1 Text), we showed that vnz_05055; vnz_12765; vnz_23445; vnz_24855 belong to group A, vnz_23100; vnz_26400 and vnz_35060 to group B, and vnz_02735 and vnz_12680 to group C (Fig 2C). Individual chitinases were distinguished by their distinct carbohydrate binding domains, which included various families of carbohydrate-binding modules, fibronectin type 3 modules, and, in the case of vnz_02735, a ricin B lectin domain, which is unusual for chitinases. While it is not uncommon for Streptomyces species to have 10 or more chitinases encoded in their genomes [27,52]; the specific biological roles of these enzymes in the bacteria remain largely unexplored. Indeed, protein BLAST analysis demonstrated that the S. venezuelae chitinases are common within the Streptomycetaceae family. All S. venezuelae chitinases yielded between 1,700 and 4,600 hits within the Streptomycetaceae family under stringent conditions (query coverage 80%–100%, sequence identity 50%–100%, and an E-value threshold of ≤1E−05), reflecting their high conservation (Table E in S1 Text and materials and methods). These numbers approach the level of the highly conserved recombinase RecA, which produced 5,600 hits. In contrast, a gene from the chloramphenicol biosynthetic gene cluster, found only in a subset of Streptomycetaceae, resulted in approximately 100 hits, showing limited distribution within the family. These data highlight the widespread presence of multiple chitinases in members of the Streptomyces genus.

Enzymes for chitin degradation and product import

To address the importance of the putative chitinase genes to grow on chitin, we constructed null mutations in genes encoding two members from group A (vnz_24855 and vnz_12765), one group B (vnz_26400), one group C (vnz_02375), and the one GH19 enzyme (vnz_16685), and compared the growth of the mutant strains to their wild-type parent on defined chitin medium and MYM medium, which contains maltose, yeast extract, and malt extract as carbon and nitrogen sources [53].

The null mutants of subfamily A chitinases vnz_24855 (which had the highest differential transcripts abundance on chitin amongst all chitinases) and vnz_12765 had impaired growth in chitin medium as measured by CFUs assays after 6 days (Fig 3A). As a result of efficient sporulation, the wild type produced an average of 10⁸ CFU/mL in chitin culture at this time point, whereas the vnz_24855 and vnz_12765 null mutants yielded approximately 10⁶ CFU/mL, as sporulation in these strains had only just begun. It is important to note that the mutants are capable of reaching higher CFU levels, but require additional time to do so. Both mutant phenotypes could be complemented by introducing a functional copy of their respective genes into their associated null mutant in trans (Fig 3A). Importantly, the growth of both mutants was similar to wild-type in MYM liquid cultures (Fig H in S1 Text), supporting a specific role for these two chitinases during growth on chitin. These data, together with the RNA sequencing data, suggests that the subfamily A chitinases in S. venezuelae are important for growth on chitin flakes when provided as only carbon and nitrogen source.

Download:

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

Fig 3. Colony-forming units (CFUs) of wild-type and mutant strains grown on chitin.

A) Bar graph: Wild-type (WT) Streptomyces venezuelae strains typically produce an average of 10⁸ CFUs per mL in chitin cultures, whereas the vnz_24855 knock-out (KO), vnz_12765 KO and vnz_26400 KO reach around 10⁶ CFUs per mL within the same time frame. Complemented strains (CP) for the vnz_24855 and 12765 chitinases reach comparable CFU’s as wild-type cultures. KO’s of vnz_02735 and 16685 maintain similar amounts of spores compared to WT. Bars represent the averages of three biological replicates and their respective error bars show standard deviations. B) After 6 days of growth on chitin, wild-type S. venezuelae bacteria formed large pellets around the chitin flakes and are sporulating (green pellet indicated by the arrow). In contrast, only few hyphae are visible for the dasBC KO. Inserts: pellets from 0.5 mL culture samples. Scale bar: 50 micrometer. The data underlying this figure can be found in S2 Data.

https://doi.org/10.1371/journal.pbio.3003292.g003

We then assessed the role of group B chitinases in supporting growth on chitin. A null mutation in vnz_26400, the group B chitinase with the highest differential transcript levels in the RNA sequencing data, had similar growth defects as observed for the group A mutants. The vnz_26400 deletion mutant yielded ~ 10⁶ CFUs per mL compared to 10⁸ for the wild-type when grown on chitin (Fig 3A), but grew like wild-type in MYM cultures (Fig H in S1 Text). The null mutant strains for those chitinases that showed less abundant transcript levels (subfamily C: vnz_02735 and GH19: vnz_16685) did not exhibit the same CFU drops when grown on chitin (Fig 3A).

In order to grow on the liberated oligosaccharides resulting from the action of extracellular chitinases, S. venezuelae must have a mechanism to import these metabolites. We noted that transcripts from the dasABCD operon—encoding the ABC transporter for chitobiose uptake in S. coelicolor [51] - were some of the most highly differentially expressed genes when comparing growth on chitin and glucose medium (Table D in S1 Text). To test their importance to growth on chitin, we deleted the dasB and dasC genes, which encode the transmembrane subunits of the transporter, and characterized the growth of the resulting strain as we did for the chitinase mutants.

Loss of dasB and dasC led to a severe growth delay, resulting in no detectable spores after 6 days of growth, indicated by the absence of green spore pigment in the cell pellet. (Fig 3B—bottom left insets). As expected, the growth and development of the dasBC mutant was equivalent to wild-type in MYM (Fig H in S1 Text). Microscopic examination of chitin flakes in the cultures (Fig 3B) revealed some residual growth for the dasBC deletion mutant. We presume that this residual growth stemmed from the uptake of other products of chitin degradation.

These data demonstrate that S. venezuelae possesses an elaborate chitinolytic system where members of the group A and group B chitinases play important roles in generating metabolizable material from insoluble chitin. It further shows that the DasABC complex is important for transporting the products of this catabolic activity into cells for further metabolism.

The Das transporter prevents resource-sharing

The profound effect of the dasBC mutation on the S. venezuelae life cycle supported a critical role for this transporter in chitin catabolism. We were intrigued, however, by the ability of the mutant to grow on chitin, albeit at greatly reduced levels, as this suggested that chitinases may still be secreted and functioning to breakdown chitin. If the Das importer is indeed the dominant importer of chitin breakdown products, we imagined that these products might accumulate in culture medium in the absence of uptake capacity. We therefore conducted LC-MS analyses of spent culture supernatant from wild-type and dasBC cultures (see materials and methods). After four days of growth, we identified a peak with m/z = 447.16 from the dasBC culture that was not present in our wild-type samples (Fig 4A). This agrees with the m/z value of the sodium (Na) adduct of chitobiose. Importantly, the spectrum showed the co-occurrence of m/z = 425.18 (Fig 4B), corresponding to the m/z value for the M + H ion of chitobiose. Further MS–MS analysis of the 447.16 ion showed fragments consistent with the sodium adduct of the GlcNAc monomer (244.08), as well as the same ion without water (226.07) (Fig 4C); this fragmentation pattern supports the accumulation of chitobiose in the dasBC culture. By measuring the area under the curve (AUC), we determined that chitobiose accumulation was at least 20 times greater in the spent medium of the dasBC null mutant (Fig 4C—inset) than the wild-type. Thus, even though growth on chitin is greatly compromised, the chitinolytic system in the S. venezuelae dasBC deletion mutants is clearly active and is releasing significant amounts of chitobiose.

Download:

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

Fig 4. Chitobiose accumulation in the dasBC deletion mutant enhances Bacillus subtilis growth in co-culture.

A) Extracted ion chromatogram for chitobiose (N,N″-diacetylglucosamine dimer—structure inserted; M + Na: 447.16). Shown are representative peaks for S. venezuelae wild-type (WT) and the deletion mutant (ΔdasBC). B) Chitobiose spectrum (MS¹). Besides the M + Na adduct and its isotopes, the M + H ion (425.18) and its isotopes are detected too. C) MS/MS spectrum of ion 447.16. Amongst the fragments detected are the ions 244.08 and 226.07, which represent the Na adduct of GlcNAc with and without H₂O, respectively. Insert: Area under the curve (AUC) of the 226.07 ion in spent media of wild-type vs. mutant bacteria. D) B. subtilis growth curves. B. subtilis expressing the red fluorophore mCherry does not exhibit growth on chitin medium, as indicated by the absence of fluorescence. When co-cultured with the Streptomyces venezuelae dasBC deletion mutant (dasBC KO), there is a marked increase in fluorescence output compared to B. subtilis grown alone on chitin medium or when co-cultured with wild-type S. venezuelae (WT). Datapoints are the averages of three biological replicates and the error bars show standard deviations. E) Colony-forming units (CFU’s) per milliliter of culture for B. subtilis. The number of B. subtilis cells is higher in a co-culture with the S. venezuelae dasBC deletion mutant compared to B. subtilis grown alone or in co-culture with wild-type S. venezuelae. Bars represent the averages of three biological replicates and their respective error bars show standard deviations. The data underlying this figure can be found in S2 Data.

https://doi.org/10.1371/journal.pbio.3003292.g004

We then asked whether B. subtilis, which cannot grow in chitin medium (Fig E in S1 Text), could grow when co-cultured with S. venezuelae, indicating that the streptomycete shares the resources it liberates. To assess this, we added equal numbers of B. subtilis cells and S. venezuelae spores to chitin medium. We followed B. subtilis growth by measuring the fluorescence of an mCherry transgene (OD₆₀₀ measurements would have been confounded by the presence of S. venezuelae and insoluble chitin flakes). When co-cultured with wild-type S. venezuelae, little fluorescence was observed, suggesting that B. subtilis had little apparent growth. In contrast, co-culture with the dasBC deletion mutant supported B. subtilis growth, as indicated by an increase in fluorescent signal over time (Fig 4D) and a substantial rise in colony-forming units (Fig 4E). These results suggest that in the absence of a functional DasABC transport complex, chitin breakdown products accumulate extracellularly, making them accessible to other microbes such as B. subtilis. B. subtilis grew normally during co-culture with both the wild-type S. venezuelae and the dasBC mutant in MYM medium, showing that co-cultures with S. venezuelae are not inherently detrimental to B. subtilis (Fig G in S1 Text). Pseudomonas aeruginosa and Rhodococcus jostii were unable to grow in defined chitin medium even when co-cultured with wild type or dasBC mutant of S. venezuelae. This implies that, in addition to being unable to use chitin as sole source of carbon, nitrogen and energy, they were not able to use any of the products released during chitin catabolism by S. venezuelae.

We did, however, want to address the possibility that the poor B. subtilis growth during co-culture with wild-type S. venezuelae might be due to antibiotic production by wild-type Streptomyces but not the dasBC mutant. We therefore grew cultures of the wild-type and dasBC mutant on chitin, prepared concentrated spent media extracts, and assessed the growth of B. subtilis when exposed to these. Neither the wild-type extract nor the dasBC mutant extract inhibited the growth of B. subtilis (Fig I in S1 Text). This observation is consistent with the RNA sequencing data, which indicates that the biosynthetic gene clusters responsible for chloramphenicol production (vnz_04400 - vnz_04475) and other secondary metabolites, such as venemycin, watasemycin, and thiazostatin (vnz_02195 - vnz_02535), are not transcriptionally activated during the first 24 hours of growth on chitin (S1 Data) [77].

We conclude that the Das transporter serves both to import critical products of extracellular chitin degradation products and, simultaneously, to prevent the sharing of this material, with other microbes. This suggests a novel view of an important ABC transporter’s role in nature that we suggest is critical for the evolutionary success of this lifestyle.

Chitobiose is a critical bioavailable chitin breakdown product

Growth on chitin in S. venezuelae is facilitated by a chitinolytic system consisting of up to 10 chitinases and the DasABC importer for chitobiose. Some chitinase mutations did not confer a growth defect on chitin. Most likely these chitinases have redundant roles with other enzymes or, conceivably, serve different functions, like antifungal activity [79–81]. The chitinase deletion mutants that exhibited a phenotype consistently showed delayed sporulation, indicating that the loss of a single chitinase gene can have tangible effects. Even these effects appear to be mitigated over time under laboratory conditions, likely due to functional redundancy within the system. Importantly, we showed that the DasABC importer is a crucial part of this metabolic pathway. This suggests that chitobiose is a critical, bioavailable chitin catabolite. Modest growth was, however, possible in the absence of a functional Das ABC transporter, suggesting that at least one other breakdown product can be imported through a currently unknown importer. A reasonable candidate is GlcNAc, which is likely to be released during the chitin degradation process. It has been further established that in S. coelicolor, GlcNac is imported via a specialized phosphotransferase system [82]. The gene cluster encoding these genes is also conserved in streptomycetes, including S. venezuelae (vnz_13025 – vnz_13035). Transcript levels of these genes are notably higher when S. venezuelae is grown on chitin compared to glucose (S1 Data).

The DasABC importer limits chitobiose availability to competitors during chitin catabolism

The DasABC transporter is a known transporter of chitobiose in S. coelicolor [51]. BLAST searches confirm its presence in the vast majority of sequenced Streptomyces genomes, as well as in Streptacidiphila and Kitasatospora species. The S. coelicolor operon is transcriptionally controlled by the repressor DasR, a pleiotropic regulator of metabolism and antibiotic production [83]. Importantly, DasR relieves repression upon GlcNAc binding. This induces transcription of the dasABCD operon, where dasA encodes the extracytoplasmic chitobiose-binding protein, while dasB and dasC encode permeases, and dasD encodes an intracellular β-N-acetylglucosaminidase that hydrolyzes the transported chitobiose into GlcNAc subunits.

Importantly, we found that DasABC prevented the accumulation of chitobiose into the environment during extracellular chitin catabolism. These findings highlight a key aspect of Streptomyces saprophytic growth: efficient uptake of chitin-derived oligomers limits their availability to other microbes. B. subtilis was only able to proliferate on minimal chitin medium when co-cultured with the dasBC deletion mutant, suggesting that the accumulation of extracellular chitobiose in the absence of a functional importer supported its growth. Importantly, while B. subtilis encodes chitinase-like enzymes, there is no direct evidence from the available literature that it can import chitobiose or that it encodes a specific chitobiose transporter [84,85]. We therefore ruled out an alternate explanation whereby chitobiose import triggered the biosynthesis of antibiotic compounds exclusively in wild-type S. venezuelae, that prevented B. subtilis growth. We would note, however, that we have not ruled out the possibility that other compounds released by the dasBC mutant during growth on chitin (including GlcNAc oligomers) contribute to the growth of B. subtilis.

It is interesting that secondary metabolites did not appear to play a role in the competition for chitin, at least under the conditions we have employed. Work by other investigators [86,87] has demonstrated that some interspecies competition involves the explicit induction of specialized metabolism in streptomycetes. Indeed, B. subtilis is highly sensitive to chloramphenicol, one of the specialized metabolites in the S. venezuelae armamentarium. In this case, however, the competitive edge was attributed to the activity of an efficient ABC transporter. The density of microbial life in the soil [28,88] and the existence of antibiotic resistance mechanisms in many microbes [89] suggests that robust nutrient sequestration could drive Streptomyces’ success in their soil environment. Also, the physical structure of soil and its organic particles limits diffusion [90,91]. Therefore, transporter-mediated privatization could reinforce spatial heterogeneity in nutrient availability and promote the formation of microcolonies enriched in saprophytic species like Streptomyces [91,92].

Chitin catabolism on Earth

We show that Streptomyces are robust chitin metabolisers, joining the ranks of other established chitin-degrading microbes, like many organisms of the genus Vibrio and Serratia marcescens. S. marcescens has four experimentally validated chitinolytic enzymes: ChiC, an endo-acting non-processive chitinase, as well as ChiA and ChiB, which are processive chitinases [93]. A fourth enzyme is a lytic polysaccharide monooxygenase that introduces oxidative chain breaks [94]. Our in silico analysis suggests that the S. venezuelae chitinolytic system encompasses all of these chitin-degrading functions and that the genes responsible for these functionalities are induced upon exposure to chitin (Fig 2C, S1 Data, and Table D in S1 Text). It would be interesting to see whether S. marcescens can also utilize insoluble raw chitin as a sole source of carbon and nitrogen.

The Vibrionaceae offer the only other well-understood example of chitin metabolism, in this case exclusively found in marine ecosystems [95,96]. Similar to Streptomyces, these bacteria can utilize chitin as their sole carbon and nitrogen source [97]. The Vibrio cholerae genome also encodes 10 secreted chitinases, and similar to our findings for S. venezuelae, some of these enzymes play a significant role in supporting growth on chitin, while others have no considerable effect [98]. In Vibrio, secreted chitinases degrade chitin extracellularly into oligosaccharides of varying lengths. The resulting oligosaccharides are imported through a dedicated outer membrane channel called chitoporin. Once in the periplasm, the oligosaccharides are degraded into GlcNAc and chitobiose, which are then transported into the cytoplasm where they enter primary metabolism [97]. Importantly, Vibrio forms dense biofilms that prevent the diffusion of these breakdown metabolites away from the Vibrio cells [99]. Noteworthy, we observed chitin-attached pelleted growth, which can be considered a biofilm in the otherwise fragmented S. venezuelae. A further difference; however, is that the Vibrio system also includes chitin-responsive chemotaxis proteins that attract cells to chitin-containing substrates [97,100]; whether an analogous system operates in Streptomyces remains to be seen. Given the high metabolic value and global abundance of chitin, it would be expected that multiple organisms have devised distinct but effective strategies to it use as a nutrient.

As primary agents in the formation and decomposition of organic matter, soil microorganisms are central to maintaining the global carbon balance [88]. This work demonstrates Streptomyces’ remarkable capacity for chitin decomposition, providing definitive evidence of this ability and identifying key players involved in this mode of growth. Members of the Streptomyces genus appear to achieve this without broadly sharing the products of saprophytic digestion, a trait that has likely contributed to their evolutionary success over the past ~380 million years [101]. Therefore, alongside Vibrio and Serratia, Streptomyces species have arguably played key roles in the cycling of the earth’s nitrogen and carbon, thereby impacting everything from soil nutrient content to climate.