Elicitation of Antimicrobial Active Compounds by Streptomyces-Fungus Co-Cultures

Abstract

The bacteria of the genus Streptomyces and Basidiomycete fungi harbor many biosynthetic gene clusters (BGCs) that are at the origin of many bioactive molecules with medical or industrial interests. Nevertheless, most BGCs do not express in standard lab growth conditions, preventing the full metabolic potential of these organisms from being exploited. Because it generates biotic cues encountered during natural growth conditions, co-culture is a means to elicit such cryptic compounds. In this study, we explored 72 different Streptomyces-fungus interaction zones (SFIZs) generated during the co-culture of eight Streptomyces and nine fungi. Two SFIZs were selected because they showed an elicitation of anti-bacterial activity compared to mono-cultures. The study of these SFIZs showed that co-culture had a strong impact on the metabolic expression of each partner and enabled the expression of specific compounds. These results show that mimicking the biotic interactions present in this ecological niche is a promising avenue of research to explore the metabolic capacities of Streptomyces and fungi.

Keywords: Streptomyces; antimicrobial agents; biosynthetic gene cluster; co-culture; fungus; specialized metabolism.

Figure 1

Streptomyces-fungus co-culture setup and screening. (A) Illustration of a Streptomyces-fungus co-culture after 14 days of growth. (B) Screening of bioactive molecules elicited during the co-culture. The compound extracts of mono-cultures of fungi and Streptomyces (rows and columns named “control”) in GA medium were compared with the extracts resulting from their co-culture in a bioassay experiment against the growth of B. subtilis ATCC6633. The values in the table indicate the percentage of inhibition of B. subtilis ATCC6633 after a 24 h growth period in comparison with a control grown in absence of the extract. The two co-cultures between S. commune 6601-A with S1D4-11 and S1D4-23 (highlighted in blue) were selected as they presented a significant (t-test p value < 0.05) impact on the growth of B. subtilis ATCC6633 in comparison with the controls.

Figure 2

Antimicrobial activities of S1D4-11-FIZ and S1D4-23-FIZ against different Bacilli. Inhibition was quantified as a percentage of growth inhibition in comparison with a control without extract. The growth was measured by spectrometry at OD 600 nm after 24 h of growth. The different Bacilli strains are (A) Bacillus subtilis ATCC6633, (B) Bacillus sp. RB2-2 N12, (C) Bacillus sp. RB2-2 N10, and (D) Bacillus sp. RB2-1 N16. Statistical difference was assessed with a t-test. *** = p value < 0.005 for both comparisons between fungus and Streptomyces-fungus interaction zone (SFIZ) and Streptomyces and SFIZ.

Figure 3

Gas chromatography-mass spectrometry (GC-MS) spectrum comparison. (A) Partial least square-discriminant analysis (PLS-DA)comparison of GC-MS spectra of mono- and co-cultures. (B) Heat-map of the first 750 discriminant features (26%) revealed by PLS-DA. The scale indicates the relative abundance of features calculated by centered-reduced of initial intensity.

Figure 4

Distribution of common and specific features between SFIZs and controls. LC-MS metabolite profiles were recorded and analyzed with GNPS. The Venn diagram compares specific and common features between each SFIZ and its controls as well as SFIZ features between the two experiments. The number of SFIZ specific features for each experiment is indicated in the dashed square. +m: positive mode; −m: negative mode; ft.: features.

Figure 5

Molecular networking of ions produced during mono- and co-cultures. Data were obtained byLC-MS/MS in both positive and negative modes and were analyzed through GNPS. Each spectrum is shown as a node and related metabolites are linked by edges. Node color represents fractions by a chromatographic separation (C18 reverse phase) prior MS/MS detection. For both SFIZ extracts, the activity against the indicator strain was only found in two fractions (N° 22 and 23). To identify potential candidates induced during the studied interactions, we selected compounds found in fractions 22 and 23 and in SFIZs but not in the controls (Figure 6). As control extracts were solubilized in dimethyl sulfoxid (DMSO) (see previous paragraph), we restricted this analysis to SFIZ extract molecules found in both DMSO and methanol, recognizing that some methanol-soluble and DMSO-insoluble candidates could be omitted.

Figure 6

Identification of specific compounds in bioactive fractions. LC-MS spectra of fractions N° 22 and N° 23 (named S1D4-11-FIZ F22 and S1D4-23-FIZ F23, respectively) with anti-Bacillus activity were compared with single culture controls and with total extracts of SFIZ in DMSO, which also has similar activity. Compounds with potential activities (highlighted in red) are those at the intersection of the total SFIZ extract and in the fraction considered. The number of features found in each condition is indicated. Features were analyzed and compared with DEREPLICATOR+ in order to build the Venn diagram.