Authors: Mapenda Gaye, Amira Zhor Rim Zinai, Nicholas Armstrong, Gaetan Herbette, Valérie Monnier, Hubert Bassene, Georges Diatta, Masse Sambou, Cheikh Sokhna, Didier Raoult, Florence Fenollar, Oleg Mediannikov
Categories: Research, Fungus combs, Fungus-farming termites, Termitomyces, Antibacterial, Meglutol, Dicrotalic acid
Source: AMB Express
Authors: Mapenda Gaye, Amira Zhor Rim Zinai, Nicholas Armstrong, Gaetan Herbette, Valérie Monnier, Hubert Bassene, Georges Diatta, Masse Sambou, Cheikh Sokhna, Didier Raoult, Florence Fenollar, Oleg Mediannikov
Fungus-farming termites (Macrotermitinae), predominantly found in Africa, are eusocial insects with significant ecological roles. Historically, they have been valued in traditional medicine, human diets, and livestock feed. These termites share a long-standing symbiotic relationship with Termitomyces fungi, which has evolved over millions of years and is critical to their survival and ecological impact. This mutualism promotes a unique monoculture of Termitomyces in the fungus comb while suppressing fungal and bacterial antagonists, likely due to the comb’s structural or chemical properties, sparking interest among researchers. In this study, we conducted an extensive examination of 11 fungus combs associated with five termite species collected in Senegal. Our analysis revealed significant antibacterial properties in the crude extracts of the combs, notably against multidrug-resistant strains. Chemical analyses led to the identification of dicrotalic acid (Meglutol) in the active fractions of two combs from agricultural areas. This compound, likely of plant origin, suggests a link between termite feeding habits and the antimicrobial potential of the combs. Although the exact bioactive compounds responsible for the antimicrobial activity have not yet been fully identified, the presence of various metabolites may explain the maintenance of Termitomyces monocultures and the suppression of pathogens. This also illustrates the complex ecological relationship between Termitomyces and termites, which may work together to produce natural bioactive compounds that suppress pathogens.
The online version contains supplementary material available at 10.1186/s13568-025-01917-2.
Fungus-farming termites (Macrotermitinae, Termitidae: Blattodea) form a monophyletic, basally branching clade and are eusocial insects with a remarkable organizational structure, which, in some cases, includes distinct castes (Aanen et al. 2002; Eggleton 2010; Hölldobler 1990; Hussain et al. 2017). These termites are primarily distributed across sub-Saharan Africa and large regions of Southeast Asia where more than 300 species are grouped into around 12 genera. They are considered the earliest known fungal cultivators in the Old World (Aanen et al. 2002; Aanen and Eggleton 2005; Hasnaoui et al. 2022, 2024; Krishna et al. 2013; Nobre and Aanen 2012).
Beyond their ecological importance, fungus-farming termites are a valuable natural resource, widely used in traditional medicine and as a source of food for both humans and livestock (Costa-Neto 2005; Ouango et al. 2022) However, they are most renowned for their obligate mutualistic relationship with Termitomyces (Agaricales: Lyophyllaceae), a lignocellulolytic fungus that has co-evolved with them (Aanen et al. 2002; Bignell et al. 2011; Schmidt et al. 2022; van de Peppel et al. 2021). This mutualism is initiated with the construction of a specialized fungal structure known as the fungus comb, built within the nest and composed of partially digested plant matter (Arshad and Schnitzer 1987; Poulsen et al. 2014; Rouland-Lefèvre and Bignell 2002).
The fungus comb is organized into two distinct a fresh upper layer and an older, more decomposed lower layer often covered with white fungal nodules (Kusumawardhani et al. 2021), This layered architecture provides a favourable microenvironment for the exclusive cultivation of Termitomyces, the fungal symbiont essential to the colony’s survival (Otani et al. 2019). Termites actively inoculate the plant substrate with fungal spores, shape the comb, and regulate nest conditions such as temperature and humidity to promote fungal development (Korb 2011). In return, Termitomyces degrades complex plant lignin, cellulose, hemicellulose, and pectin into simpler compounds (da Costa et al. 2019; Dashtban et al. 2010; Hyodo et al. 2000; Ohkuma 2003), effectively transforming inedible plant matter into digestible nutrients. As a result, the fungus produces protein-rich nodules, which are consumed by termites at various developmental stages, playing a central role in their nutrition and the sustainability of the colony (Tibuhwa 2012; Wisselink et al. 2020).
This mutualism extends beyond nutrient cycling. Termites also shape the fungal and bacterial microbiota of the comb through selective substrate collection, digestive processing, and the antimicrobial properties of their saliva and gut (Brune and Ohkuma 2011; Moreira et al. 2021; Otani et al. 2019b; Prestwich et al. 1977; Visser et al. 2012). Social behaviours such as waste management, and substrate inoculation also contribute to pathogen suppression and microbial filtering (Aanen et al. 2002; Nobre and Aanen 2012). Moreover, the physicochemical conditions of the nest, including high humidity, stable temperature, and rich organic content act as an ecological filter, further influencing microbial composition.
Despite the selective nature of this environment, the fungus comb hosts diverse bacterial communities, some likely originating from the termite gut via the processed substrate (Mathew et al. 2012; Otani et al. 2016; Yin et al. 2019). The chemical and microbial profile of the comb appears to favour the growth of Termitomyces while inhibiting antagonistic or competitive microorganisms, potentially through antimicrobial compounds produced by the termites, their microbial symbionts, or the comb matrix itself (Nandika et al. 2021, p. 20; Um et al. 2013).
Ethnographic observations in sub-Saharan Africa suggest that fungus combs, termite soldiers, and termite mound soils are traditionally used for their antimicrobial properties (Figueirêdo et al. 2015; Hsieh and Ju 2018; Meyer-Rochow 2017). Recent scientific studies have begun to confirm these practices, reporting that crude extracts of fungus combs from Macrotermitinae termites exhibit antifungal and antibacterial activity (Kusumawardhani et al. 2021; Nandika et al. 2021; Witasari et al. 2022). These bioactivities are likely the result of complex interactions between termite physiology, microbial symbionts, and secondary metabolites from the comb matrix.
Based on previous studies showing the antimicrobial properties of fungus combs and their potential of being a new antimicrobial agents (Nandika et al. 2021, 2023; Witasari et al. 2022), this study aims to investigate whether fungus combs from different termite species exhibit antibacterial activity, and whether this activity varies depending on termite species and mound location. Additionally, the study seeks to isolate and characterize a bioactive compound from Macrotermes bellicosus using chemical and spectrometric methods to explore its activity.
The fungus combs were collected in the Kedougou region of Senegal in July 2021 as shown in Fig. 1; Table 1. The combs and soldier termites were separated in bags and tubes, respectively. All samples were then transferred while maintaining the cold chain and stored at − 20 °C until use.
Fig. 1a Mapping of Senegal. b Collection sites in the Kédougou Dindéfello and Segou. Created by QGIS software 3.38.2 Spatial without Compromise QGIS Web Site
Table 1List of fungal combs and their respective locations in Senegal utilized in the studySample IDTermite species associated with fungus combsSite of collectionCoordinatesDate of collectionSOT01aPseudacanthotermes sp.Dindéfelo, forest12.377978211120945, 12.31517149291925507.07.2021SOT02Pseudacanthotermes sp.Dindéfelo, forest12.377978211120945, 12.31517149291925507.07.2021SOT07 Macrotermes subhyalinus Dindéfelo, forest12.377978211120945, 12.31517149291925507.07.2021SOT08 Macrotermes bellicosus Dindéfelo, forest12.377978211120945, 12.31517149291925507.07.2021SOT10Odontotermes sp.Dindéfelo, forest12.377978211120945, 12.31517149291925507.07.2021SOT12 Ancistrotermes cavithorax Dindéfelo, forest12.377978211120945, 12.31517149291925507.07.2021SOT30 Ancistrotermes cavithorax Ségou, forest12.390661232888291, 12.2858513231289109.07.2021SOT35Odontotermes sp.Ségou, forest12.390661232888291, 12.2858513231289109.07.2021SOT36 Macrotermes subhyalinus Ségou, forest12.390661232888291, 12.2858513231289109.07.2021SOT40 Macrotermes bellicosus Ségou, field12.40135933781269, 12.28325287873392709.07.2021SOT42 Macrotermes bellicosus Ségou, field12.40135933781269, 12.28325287873392709.07.2021
The bioactive compounds from fungus combs were extracted using ethyl acetate as the extraction solvent, and its efficiency was compared to that of other solvents such as methanol and water, previously evaluated in earlier studies. (Fig. 1S) (Caesario et al. 2023; Witasari et al. 2022).
First, 50 g of fungus combs were defrosted, manually crushed, and mixed with 500 mL of ethyl acetate (≥ 99.5% ACS, VWR Chemicals BDH, Rosny-sous-Bois, France). The mixture was incubated for 24 h at room temperature with continuous shaking in a shaker-incubator (MaxQ™ 8000, Thermo Scientific, Illkirch-Graffenstaden, France). After decantation, the organic phase was separated and evaporated to dryness under a nitrogen stream at 40 °C using an evaporator with a fixed heating block (Liebisch, Bielefeld, Germany).
The dried residue was then re-dissolved in 5 mL of HPLC-grade water, vortexed, and centrifuged at 3000 ×g for 5 min to remove insoluble material. The resulting supernatant was subsequently evaporated to dryness, weighed, and stored at 4 °C until further use. The extraction yielded crude extracts that were dark yellow in colour, with a pH of 7 and a concentration of 50 mg/mL.
A total of 46 bacterial strains from the WDCM collection (WDCM 874, https://ccinfo.wdcm.org/details?%20regnum=874) were used to evaluate the antimicrobial properties of fungus comb extracts (Table 2).
Table 2Bacterial species list tested against crude comb extractsBacterial strainsWDCM 874 collection n°Culture media Acinetobacter baumannii P0025Columbia AgarAcinetobacter baumannii, AYEP0024Columbia AgarA. baumannii producing NDM carbapenemaseP1886Columbia AgarA. baumannii producing Oxa-23 carbapenemaseP1573Columbia AgarA. baumannii producing Oxa-24 carbapenemaseP1574Columbia Agar Bacillus cereus Q4833Columbia Agar Bordetella bronchiseptica Q5540Columbia Agar Bordetella pertussis Q7922Columbia Agar Campylobacter jejuni Q4401Columbia Agar Corynebacterium diphtheriae Q6123Columbia Agar Clostridioides difficile Q5206Columbia Agar Enterobacter bugandensis P9515Columbia Agar Enterococcus faecalis Q6320Columbia AgarE. faecium resistant to vancomycin (vanA)P3499Columbia AgarE. faecium resistant to vancomycin (vanB)Q2649Columbia AgarEscherichia coli with low-level penicillinaseQ9211Columbia AgarE. coli with high-level penicillinaseQ9210Columbia AgarE. coli producing cephalosporinaseQ9209Columbia AgarE. coli producing extended-spectrum β-LactamaseQ9212Columbia AgarKlebsiella pneumoniae, sensitive to carbapenemQ0220Columbia AgarK. pneumoniae producing OXA-48 carbapenemaseQ1445Columbia Agar Lactobacillus gasseri P5516Columbia Agar Lactobacillus iners P2565Columbia Agar Listeria monocytogenes Q3532Columbia Agar Micrococcus luteus P7806Columbia Agar Mycobacterium abscessus Q7106Middlebrook 7H10 Mycobacterium houstonense Q6899Middlebrook 7H10Mycobacterium tuberculosis resistant to rifampicinQ6975Middlebrook 7H10M. tuberculosis resistant to rifampicinQ6933Middlebrook 7H10M. tuberculosis resistant to rifampicinQ6976Middlebrook 7H10M. tuberculosis resistant to rifampicinQ1644Middlebrook 7H10M. tuberculosis sensitive to rifampicinQ7376Middlebrook 7H10M. tuberculosis sensitive to rifampicinQ7296Middlebrook 7H10 Proteus mirabilis Q5254Columbia Agar Proteus vulgaris Q5838Columbia Agar Pseudomonas aeruginosa Q9213Columbia Agar Salmonella enterica P2209Columbia Agar Serratia marcescens Q2543Columbia AgarS. marcescens producing OXA-48 carbapenemaseQ1873Columbia Agar Staphylococcus aureus Q5098Columbia AgarMethicillin-resistant S. aureus (MRSA)P5549Columbia Agar Stenotrophomonas maltophilia Q5352Columbia Agar Streptococcus pneumoniae P0800Columbia Agar Streptococcus pyogenes Q0656Columbia Agar Vibrio cholerae Q0469Columbia Agar Yersinia enterocolitica Q5864Columbia Agar
Initial screening was performed on Micrococcus luteus (P7806) using the agar well diffusion method (Balouiri et al. 2016; Debalke et al. 2018; Jiménez-Esquilín and Roane 2005) to determine the most active extract.
Subsequently, only one extract (SOT40) was selected for further testing against all bacterial strains due to its initial efficacy against Micrococcus luteus and the large amount of fungus comb collected, which allowed for multiple extractions followed by in-depth analyses to characterise and purify the sample. Although it did not exhibit the highest inhibition, the substantial quantity of material available made it suitable for comprehensive testing.
Bacterial inocula were adjusted to 10^8^ cells/mL using the McFarland 0.5 standard. Fungus-comb extracts were prepared at 50 mg/mL in HPLC-grade water and 50 µL (2.5 mg) of each extract was applied to agar wells. Inhibition zones were measured after incubation, with results compared to normalized values (Tsirinirindravo and Andrianarisoa 2009) using the Graph Pad software version 6.0 at a 5% significance level of and results were recorded separately. Positive controls included antibiotic discs oxacillin (5 µg) and erythromycin (15 µg) (SIRscan, Montpellier, France).
To isolate the bioactive compounds responsible for in vitro antibacterial activity, the SOT40 fraction was purified by High-Performance Liquid Chromatography with Ultraviolet detection (HPLC-UV, Waters, Milford, MA, USA). The separations were carried out using 3 mL of crude extract (50 mg/mL).
The first separation was performed on a Spherisorb Phenyl column (10 μm, 4.6 mm x 250 mm, Waters, Milford, MA, USA) featuring a C6-Phenyl phase on non-end capped silica. Fractions showing antimicrobial activity were further separated using a SunFire C18 column (5 μm, 4.6 mm x 250 mm, Waters). Active fractions from this step were subjected to a final purification on an XBridge BEH Amide column (5 μm, 4.6 mm x 250 mm, Waters), which is specifically designed to retain polar compounds. The collected fractions were dried under a nitrogen stream at 40 °C and reconstituted in pure HPLC-grade water for assay. Each run consisted of 300 µL of crude extract.
All separations were performed at a flow rate of 1 mL/min with the columns maintained at 35 °C. The mobile phases consisted of solvent A (water) and solvent B (acetonitrile). A linear gradient was 33–97% B over 20 min for the Spherisorb Phenyl and SunFire C18 columns, and 90–10% B over 20 min for the XBridge BEH Amide column. UV detection was conducted between 200 and 800 nm using a photodiode array detector.
Electrospray Ionization High-Resolution Mass Spectrometry (ESI-HRMS) analyses were performed on a SYNAPT G2 HDMS mass spectrometer (Waters) equipped with an Electrospray Ionization Source. The sample was ionized in positive mode under the following capillary 2.8 kV; orifice 20 V; nebulising gas flow rate (nitrogen): 100 L/h. Similarly, the sample was also ionized in negative mode under the following capillary − 2.27 kV; orifice − 20 V; nebuliser gas flow rate (nitrogen): 100 L/h. Mass spectra (MS) were obtained using a time-of-flight (TOF) analyser. Accurate mass measurements were performed in triplicate with an external calibration. The sample was dissolved in 300 µL of water and then diluted 1/10^4^ in a solution of methanol with 0.1 mM sodium chloride. The solution was infused into the ionization source at a flow rate of 10 µL/min.
NMR analyses were performed with a Bruker Avance II^+^ 600 MHz spectrometer (magnet of 14.1 Tesla) operating under Topspin 3.2 PL6 and equipped with a 5 mm TCI Cryoprobe. The active compound was successively solubilized in 80 µL of D2O 99.90%-d + 0.1% TMPS-d4 (Eurisotop, Gif-sur-Yvette, France, ref. D221B, batchM1751, δ1H = 0.00 ppm, δ13C = 0.00 ppm), in 80µL of CDCl3 99.96%-d (Eurisotop, ref. D029T, δ1H = 7.26 ppm) and in 80 µL of Acetone-d6 99.96%-d (Eurisotop, ref. D038B, δ1H = 2,05 ppm). The solution was placed in a 2 mm OD capillary tube. The NMR spectra were performed using standard sequences from the Bruker library at 300 K. The chemical shifts δ are given in ppm and coupling constants J in Hz. NMR spectra were processed using TopSpin 4.1.4 software.
FTIR analysis was performed on a Bruker Tensor 27 spectrometer equipped with Diamond Platinum ATR. The active compound was solubilized in ~ 20 µL of chloroform, then ~ 10 µL was applied to the crystal. After evaporation of the solvent, FTIR analysis was performed, and the spectrum was processed with OPUS 6.0 software.
The bioactivity of dicrotalic acid was evaluated using both purified (500 µg/mL dry extract) and crude (1000 µg/mL dry matter) extracts of the 11 crude extracts. Ultra-Performance Liquid Chromatography-Mass Spectrometry (UPLC/MS) analysis was performed with an Acquity I-Class chromatography system coupled to a Vion high-resolution Quadrupole-Time-of-Flight (Q-ToF) mass spectrometer (Waters).
Samples were eluted at 0.6 mL/min through a Raptor Polar X ion exchange column (2.7 μm, 30 × 2.1 mm, Restek) maintained at 30 °C. The mobile phases, prepared with ULC/MS-grade solvents and reagents (Biosolve, Dieuze, France), were as solvent A, an aqueous buffer at pH 5 (10% of 200 mM ammonium formate in water and 90% acetonitrile); solvent B, 0.5% formic acid in water. The gradient program was set to maintain 5% solvent B for 1.5 min, ramp to 90% over 1.5 min, and hold for an additional minute.
Chemicals were ionized in negative mode with the following capillary voltage at 1 kV, source temperature at 120 °C, and desolvation temperature at 250 °C. Data were acquired using the MSe survey method with Lock mass correction, employing leucine enkephalin ([M-H]− 554.2620 m/z) as the reference. Dicrotalic acid was detected at 1.5 min ([M-H]− 161.04555 m/z, fragment ion C5H8O2 99.04515 m/z). An internal standard (glyphosate IS) was used and was detected at 3.5 min ([M-H]− 168.00673 m/z).
A calibration curve for dicrotalic acid was generated using pure standards (Cayman Chemicals) diluted in water to prepare six concentrations (10, 50, 100, 200, 500, and 1000 µg/mL). All samples and standards were spiked with glyphosate as an internal standard.
Analyses were then performed in duplicate under identical conditions. To ensure system integrity, a passivation injection with medronic acid (1.8 mg/mL in 50 methanol) was performed prior to sample analysis.
The antimicrobial activity of pure dicrotalic acid (Cayman Chemical, ref. 33832, Ann Arbor, MI, USA) was also evaluated with M. luteus (P7806). The agar well diffusion method was used to demonstrate the antibacterial activity of this molecule as described above. Four different concentrations of dicrotalic acid in pure HPLC water (mg/mL) were 1, 10, 50, and 100. Then, 50 µL of each stock was added to the agar wells corresponding to 0.05 mg, 0.5 mg, 2.5 mg, and 5 mg, respectively.
The antibacterial activity of 11 crude extracts was tested against M. luteus (P7806) with inhibition zones ranging from 14 to 22 mm. Combs extracts associated to Odontotermes sp. and M. bellicosus species showed the highest efficacy (21–22 mm) (Fig. 2; Table 3).
Fig. 2Bar chart representing the inhibition zones (mm) of 11 fungus comb extracts (50 mg/mL) against Micrococcus luteus
Table 3Inhibition zones of 11 fungus comb extracts (50 mg/mL) against M. luteus, and dicrotalic acid concentrations in bioactive crude fungus comb extracts as determined by HPLC-LC/MSSample IDInhibition zones (mm) with 50 µL (2.5 mg)Concentration of dicrotalic acid by HPLC-LC/MS (µg/mL)SOT01a14–SOT0215–SOT0714–SOT0818–SOT1017–SOT1214–SOT3014–SOT3522–SOT3614–SOT4020225SOT4221175
The concentrated crude extract associated with M. bellicosus (SOT40 mound, 50 mg/mL) demonstrated strong antimicrobial activity against various Gram-positive and Gram-negative bacteria, with the largest inhibition zone for M. luteus (30.6 mm) and the smallest for Escherichia coli producing cephalosporinase (12.8 mm). Antimicrobial activity against M. luteus was maintained even after exposure to extreme pH conditions, followed by neutralization. All results for Gram-negative and Gram-positive bacterial strains are reported in Fig. 3; Table 4.
Fig. 3Bar chart representing the inhibition zones (mm) of the fungus comb extract SOT40 (50 mg/mL) against Gram-negative bacteria, Gram-positive bacteria, and mycobacteria
Table 4Inhibition zones of the fungus comb extract SOT40 (50 mg/mL) against gram-negative and gram-positive bacteria, as well as mycobacteriaBacterial strainsCSUR n°Inhibition zones (mm) Acinetobacter baumannii P002521.7Acinetobacter baumannii, AYEP002416.7A. baumannii producing NDM carbapenemaseP188621.7A. baumannii producing Oxa-23 carbapenemaseP157324.3A. baumannii producing Oxa-24 carbapenemaseP157424.1 Bacillus cereus Q483319.3 Bordetella bronchiseptica Q554024.8 Bordetella pertussis Q792227.8 Campylobacter jejuni Q440117.8 Corynebacterium diphtheriae Q612316.3 Clostridioides difficile Q520620 Enterobacter bugandensis P951516.3 Enterococcus faecalis Q632013E. faecium resistant to vancomycin (vanA)P349919.6E. faecium resistant to vancomycin (vanB)Q264921.9Escherichia coli with low-level penicillinaseQ921120E. coli with high-level penicillinaseQ921018E. coli producing cephalosporinaseQ920912.8E. coli producing extended-spectrum β-LactamaseQ921213.5Klebsiella pneumoniae, sensitive to carbapenemQ022020.6K. pneumoniae producing OXA-48 carbapenemaseQ144521.3 Lactobacillus gasseri P551620.2 Lactobacillus iners P256523.3 Listeria monocytogenes Q353214.6 Micrococcus luteus P780630.6 Mycobacterium abscessus Q710619.4 Mycobacterium houstonense Q689917.4Mycobacterium tuberculosis resistant to rifampicinQ697527.7M. tuberculosis resistant to rifampicinQ693326.6M. tuberculosis resistant to rifampicinQ697630.3M. tuberculosis resistant to rifampicinQ164425.7M. tuberculosis sensitive to rifampicinQ737627M. tuberculosis sensitive to rifampicinQ729626.2 Proteus mirabilis Q5254No inhibition Proteus vulgaris Q5838No inhibition Pseudomonas aeruginosa Q921320 Salmonella enterica P220913 Serratia marcescens Q254325.2S. marcescens producing OXA-48 carbapenemaseQ187325.9 Staphylococcus aureus Q509820.4Methicillin-resistant S. aureus (MRSA)P554920.7 Stenotrophomonas maltophilia Q535219.3 Streptococcus pneumoniae P080023.5 Streptococcus pyogenes Q065624.1 Vibrio cholerae Q046919.1 Yersinia enterocolitica Q586423.3
Of the 24 Gram-negative bacterial strains tested, three Enterobacteriaceae strains (E. coli producing cephalosporinase, E. coli producing extended-spectrum β-lactamase (ESBL), and Salmonella enterica) were the least susceptible, with inhibition zones ranging from 12.8 to 13.5 mm. Other Gram-negative bacteria showed higher sensitivities (16.7–27.8 mm), including Acinetobacter baumannii,* A. baumannii* AYE, A. baumannii producing OXA-23 carbapenemase, A. baumannii producing Oxa-24 carbapenemase, (A) baumannii producing New Delhi Metallo-β-Lactamase (NDM) carbapenemase, Bordetella bronchiseptica,* (B) pertussis*,* Campylobacter jejuni*,* Enterobacter Bugandensis*,* E. coli* with low level penicillinase (LLP), E. coli with high level penicillinase (HLP), K. pneumoniae sensitive to carbapenem, K. pneumoniae producing Oxa-48 carbapenemase, Pseudomonas aeruginosa,* Serratia marcescens*,* S. marcescens* producing Oxa-48 carbapenemase, Stenotrophomonas maltophilia,* Vibrio cholerae*, and Yersinia enterocolitica. Among these, B. pertussis (Q7922) showed the largest inhibition zone (27.8 mm). However, no activity was observed against Proteus mirabilis and Proteus vulgaris after 48 h of incubation. (Fig. 2S; Table 4)
All tested Gram-positive bacterial strains were highly susceptible to the crude extract. Among them, Enterococcus faecalis and Listeria monocytogenes showed relatively lower sensitivity (13–14.6 mm), compared to other Gram-positive strains (16.3–30.6 mm), including Bacillus cereus, Corynebacterium diphtheriae, Clostridioides difficile, E. faecalis, vancomycin-resistant Enterococcus faecium (vanA and vanB), Lactobacillus gasseri, Lactobacillus iners, Mycobacterium abscessus, Mycobacterium tuberculosis, Mycobacterium houstonense, methicillin-resistant Staphylococcus aureus (MRSA), Streptococcus pyogenes, and Streptococcus pneumoniae (Fig. 2S; Table 4).
During the initial purification on a Phenyl Spherisorb column, the chromatogram revealed several prominent peaks. Of the six fractions tested for antimicrobial activity against M. luteus (Fig. 4a), only one showed antibacterial activity (Fig. 3S-1a). This fraction, likely containing a polar compound with low retention on the reverse-phase column, was further purified.
Fig. 4UV-Visible Absorbance chromatogram of the pooled fractions using the C6-phenyl Spherisorb. a SunFire C18 and XBridge BEH. b Amide Columns. c Identified fractions are numbered according to the colors and the red line indicates the antimicrobial activity
In the second purification step, the chromatogram displayed five major peaks (Fig. 4b), each of which was tested for antibacterial activity, only one fraction showed activity (Fig. 3S-1b). The increased retention of the active compound on this column compared to the C6 Phenyl column suggested a lack of aromatic groups in its structure. The final separation produced a cleaner chromatogram with well-resolved peaks (Fig. 4c). Of the four fractions obtained, one demonstrated antibacterial activity against M. luteus (Fig. 3S-1c).
The purified antibacterial compound exhibited a UV spectrum with a maximum absorption at 206 nm (Fig. 5), likely associated with carbonyl functional groups. After purifying the bioactive compound through three successive HPLC-UV steps, additional analyses, including electrospray ionization mass spectrometry (ESI-MS) and nuclear magnetic resonance (NMR), were performed to determine its chemical structure in greater detail. Retention time alignment between the extract and standard showed slight variations between runs and was therefore not used as the sole criterion for compound identification.
Fig. 5UV-Visible absorption chromatograms of the purified SOT40 fraction and dicrotalic acid (meglutol) obtained using SunFire C18 and XBridge BEH Amide column
Dicrotalic acid was analysed from approximately 7 mg of the purified and bioactive fraction obtained from the SOT40 extract. The molecular formula C6H10O5, was established from (ESI-HRMS) data showing a sodium adduct peak at m/z 185.0424 [M + Na]^+^ (calculated for C6H10O5Na^+^, 185.0420) and a pseudo-molecular ion peak at m/z 161.0459 [M-H]^−^ (calculated for C6H9O5^−^, 161.0455) (Fig. 6a, b), suggesting the occurrence of two degrees of unsaturation (Fig. 4S a and b, Table 1S and 2S). The IR spectrum showed the presence of a broad band with signals in 1739 and 1711 cm^− 1^, which seems to correspond to the open form of dicrotalic acid. The broad signal could correspond to intermolecular hydrogen bonding with the acid and alcohol functions present. The ^13^C NMR spectra showed only four signals due to an axial symmetry corresponding to two equivalent quaternary carbon of acid function C-1/C-5 (178.0 ppm), another quaternary carbons with a hydroxyl function C-3 (72.9 ppm), two equivalent methylene carbons C-2/C-4 (48.1 ppm), and a methyl carbon C-6 (29.3 ppm).
Fig. 6a Positive electrospray ionization mass spectrum (MS) of the SOT40 sample. b negative electrospray ionization mass spectrum (MS) of the SOT40 sample
The ^1^H NMR spectra showed only three signals corresponding to two signals of the diastereotopic proton of methylene protons H-2/H-4 as doublets (J 14.8) at 2.70 ppm and 2.73 ppm due to the presence of a C-3 asymmetric carbon atom, and one methyl as a singlet at 1.42 ppm. The HMBC spectrum showed cross-peak correlation between the proton H-6 of the methyl at 1.42 ppm with the carbon C-3 at 72.9 ppm and C-2/C4 at 48.1 ppm, and the proton H-2/H-4 of the methylene at 2.70 ppm and 2.73 ppm with the carbon C-1/C5 at 178.0 ppm, C-3 at 72.9 ppm, C-2/C4 at 48.1 ppm, and C-6 at 29.3 ppm. Due to the presence of exchangeable protons of the hydroxyl and acid functions, these protons were not observable. We attempted to perform the analysis in acetone and chloroform but without success. All results are reported in (Fig. 5S and Table 3S).
Nevertheless, dicrotalic anhydride, an alternative structure corresponding to the NMR results, is spontaneously hydrolysed to dicrotalic acid in presence of water. Thus, the final active chemical in aqueous solutions is therefore dicrotalic acid (meglutol) (Figs. 6S and 7). NMR analysis quantified the purity of the analysed extract at 90% of dicrotalic acid.
Fig. 7Chemical structure of dicrotalic acid
This technique was used to confirm the presence and purity of dicrotalic acid in the SOT40 crude and HPLC-UV purified extracts, and to investigate its presence in other fungus combs. First, we confirmed its presence in a purified SOT40 fungus comb extract with a purity of about 70% (335 µg / 500 dry matter). Among the 11 crude fungus comb extracts (1000 µg/mL dry matter), dicrotalic acid was detected in samples SOT40 and SOT42 from Ségou, with a concentration of 225 µg/mL and 175 µg/mL, respectively (Table 3). This screen confirms the presence of dicrotalic acid in crude and purified extracts associated to M. bellicosus from Ségou, Senegal. However, this compound was not detected in the samples from Dindéfelo, Senegal.
Dicrotalic acid demonstrated antimicrobial activity against the M. luteus strain (P7806), producing a zone of inhibition zone of 9 mm at a minimum concentration of 10 mg/mL, equivalent to 0.5 mg (50 µL). In comparison, dicrotalic acid standards at concentrations of 50 mg/mL and 100 mg/mL produced inhibition zones of 24 mm and 33 mm, respectively.
In this study, we investigated the in vitro antibacterial properties of fungus comb extracts collected in Senegal. These complex structures and sometimes termite mound soil have long been used in traditional medicine for treating various diseases (Luoba et al. 2004; Njiru et al. 2011; van Huis 2017). We also identified a major compound in two extracts, which may be primarily responsible for the observed antibacterial activity. All samples were collected in a single region, so any influence of wider geographic or seasonal variation could not be assessed, a point we recognize as a limitation of the present work.
Our initial screening of ethyl acetate crude extracts from 11 fungus combs associated with five termite species revealed antibacterial activity against M. luteus in all samples, though activity levels varied. Extracts from M. bellicosus and Odontotermes sp. appeared to be the most potent, showing larger zones of inhibition, particularly SOT35 (Odontotermes sp., 22 mm), followed by SOT42 and SOT40 (M. bellicosus). In contrast, combs from Pseudacanthotermes sp., M. subhyalinus, and A. cavithorax exhibited weaker activity, with inhibition zones around 14 mm. These results are consistent with previous findings reporting the antimicrobial activity of soldier extracts from M. bellicosus, including activity against clinical strains such as Staphylococcus aureus (Hammoud Mahdi et al. 2020). However, further studies are needed to better understand the mechanisms and ecological roles related to termite species and geographical variation. Because inhibition assays were performed only once per strain, except for M. luteus, which we replicated for method validation, the numerical values reported here should be interpreted with caution and regarded as exploratory.
The effectiveness of the ethyl acetate extracts is likely due to the solvent’s capacity to extract essential antibacterial metabolites, such as polyketides, macrolides, and alkaloids, given its relatively non-polar nature. This choice of solvent is supported by previous studies, which reported that ethyl acetate extracts yield significantly larger inhibition zones than those obtained using hexane, water, or methanol (Lajoie et al. 2022; Nandika et al. 2021, 2023). Additionally, the subsequent water solubilization step ensures the capture of more polar compounds that are not soluble in the organic solvent, thus enhancing the overall extraction of bioactive molecules.
To perform a broader antibacterial screening, extract SOT40 was selected due to the larger quantity of material available, which allowed for repeated extractions and analyses. While it did not exhibit the strongest inhibition, its volume made it suitable for extensive testing. This extract displayed broad-spectrum antibacterial activity against nearly all 46 bacterial strains tested, except for Proteus mirabilis and P. vulgaris.
Our findings align with previous studies reporting antimicrobial properties of fungus comb extracts from termite species such as Macrotermes gilvus which inhibit wood-staining fungi (Nandika et al. 2021). In their study, Nandika et al. demonstrated that fungus comb extracts exhibited significant antifungal activity against Aspergillus foetidus, with inhibition rates varying depending on the extraction solvent used and achieving complete fungal growth inhibition with certain extracts. Beyond their role in controlling wood-staining fungi, these findings suggest broader applications for biological control strategies. Given the growing need for environmentally friendly alternatives to synthetic fungicides in agriculture, fungus comb-derived compounds could represent a valuable natural resource for crop protection.
Furthermore, our study provides new evidence that fungus combs possess a chemically diverse array of bioactive compounds originating from plants, termites, and associated microbial symbionts. This diversity likely contributes to maintaining the disease-free monocultures of Termitomyces cultivated within termite mounds (Agarwal et al. 2024; Otani et al. 2019a).
Notably, this is the first study to report the antibacterial activity of these extracts against highly pathogenic and multidrug-resistant bacterial strains. Such findings hold significant promise not only for sustainable agriculture, where these extracts could serve as biocontrol agents to reduce postharvest losses, but also for potential medical applications. Previous studies have already shown their efficacy against postharvest pathogens like Aspergillus flavus and A. niger (Witasari et al. 2022), and the remarkable resilience of Termitomyces monocultures offers a compelling model for future crop protection strategies. In addition, the isolation of metabolite-producing actinobacteria from fungus combs presents promising opportunities for the discovery of new antibiotics.
To identify the compounds responsible for these antibacterial effects, we used HPLC-UV, NMR, FTIR, ESI-MS, and LC-MS analyses. These techniques enabled the purification of an active fraction with a purity of 70–90%. The combination of NMR/FTIR and ESI-HRMS data led to the identification of the major compound in SOT40 as dicrotalic acid, also known as meglutol, a product of hydrolysis of the alkaloid dicrotaline found in Crotalaria dura and C. globifera (Adams and Duuren 1953). It is worth noting that minor variations in retention time were observed between chromatographic runs conducted on different dates (Fig. 5), which limits the use of retention time alone as conclusive evidence for compound identification in this study. Therefore, compound identification was based on spectroscopic and mass spectrometric data.
Although meglutol is known as a plant and human metabolite with antilipemic properties, including lowering cholesterol, triglycerides, and β-lipoproteins, and inhibiting hydroxymethylglutaryl-CoA reductase (Beg and Siddiqi 1967; da Rosa et al. 2020; Di Padova et al. 1984; Lupien et al. 1973, p. 19; Medina-Franco et al. 2005), it had never been reported in association with termite fungus combs. While used medically for these purposes, its potential antibiotic effects have not been previously explored. Our results show that dicrotalic acid (meglutol) exhibits antibacterial activity with a minimum effective concentration of 0.5 mg/mL against M. luteus.
LC-MS screening of all 11 crude extracts revealed the presence of meglutol in two extracts from fungus combs of M. bellicosus collected in agricultural fields in Ségou, but not in samples from Dindéfelo. Interestingly, despite the absence of meglutol in the Dindéfelo samples, these still showed comparable antibacterial activity. This suggests the presence of other, yet unidentified, bioactive compounds contributing to the antimicrobial effect.
Previous studies have demonstrated synergistic interactions between multiple compounds from Termitomyces, the fungus comb matrix, or associated bacteria (Benndorf et al. 2018; Chen et al. 2023; Mahamat et al. 2018; Sitati et al. 2021; Visser et al. 2012; Yin et al. 2019). It is likely that the antimicrobial activity in fungus comb extracts results from complex interactions between multiple active constituents.
This study highlights the potent antibacterial effects of crude fungus comb extracts and identifies dicrotalic acid (meglutol) as one of the major active compounds, although it remains unclear whether it originates from Termitomyces, another fungus, or the termites’ own metabolism. The most plausible explanation is that meglutol derives from plant material incorporated into the nest, which opens new avenues for phytochemical studies of local vegetation around termite mounds.
Overall, additional bioactive compounds contributing to the antibacterial activity remain to be identified. The observed activity likely results from a combination of multiple compounds. Further investigations using comprehensive chemical characterizations and functional analyses are needed to uncover and elucidate these compounds. Our findings contribute to the chemical ecology of fungus combs and support their potential as a source of novel antimicrobial agents. Future research should focus on detailed chemical profiling and bioassays of additional fungus comb samples to discover new bioactive molecules.
Below is the link to the electronic supplementary material.
Supplementary Material 1