Nutrient-dependent influence of red, green, blue, and UV-A light on the production of destruxin A, B, and E in Metarhizium brunneum bioreactor batch cultures

Destruxins are cyclic hexadepsipeptides with a broad bioactivity spectrum, including insecticidal, phytotoxic, antiviral, antimicrobial, and anticancer properties, making them highly valuable metabolites in agriculture, medicine, and biotechnology. Due to their critical role in insect-fungal and plant-fungal interactions, understanding the environmental factors triggering their production is of high interest. This study examined whether different light wavelengths affect the production of the destruxins A, B, and E in Metarhizium brunneum BIPESCO 5, and whether any light effect was independent of the nutritional state of the culture. To achieve this, M. brunneum was grown in bioreactor batch cultures using two different media (i.e., nutrient-rich Sabouraud-4 %-glucose (S4G) medium versus nitrogen-limited (N-lim) minimal medium). Five illumination conditions were applied to each tested medium, comprising either constant darkness, or red light (660 ± 25 nm), green light (528 ± 33 nm), or blue light (455 ± 20 nm) at an intensity of 1.4 mW cm -2 as well as UV-A (340/375 + 11 nm) irradiation at an intensity of 1.1 mW cm -2. Destruxin production proved to be affected by the illumination conditions, with distinct light responses depending on the medium composition. Overall, cultures grown in the N-lim minimal medium exhibited significantly higher destruxin A and B concentrations compared to those in the nutrient-rich S4G medium. Shorter wavelengths (UV-A, blue, and green light) enhanced destruxin A and B production in the S4G medium but reduced their levels in the N-lim medium, whereby the destruxin A:B ratios remained constant within each nutritional condition. This study adds to the growing evidence of complex, nutrient-dependent light responses in fungal secondary metabolism, suggesting that destruxin biosynthesis is a finely tuned adaptive mechanism, possibly as a tool for niche exploitation, rather than a simple on/off metabolic reaction.

Fungi of the genus Metarhizium are ubiquitously present (Brancini et al., 2022) and occupy a remarkable bandwidth of ecological niches, encompassing saprotrophs and rhizosphere competent organisms, (growth promoting) endophytes in plants, and insect pathogens (St Leger and Wang, 2020; Brancini et al., 2022). Indeed, some Metarhizium species are among the best studied entomopathogenic fungi (Stone and Bidochka, 2020), making them keystone species for use as biological pest control agents (Brancini et al., 2022).

One reason for the fascinating ecological versatility of the genus Metarhizium is the highly flexible metabolism of these fungi, which allows rapid shifts between various ways of life (Lovett and St Leger, 2015; St Leger and Wang, 2020; Stone and Bidochka, 2020) and to endure demanding environmental conditions such as nutritive, temperature, and oxidative stress, desiccation, or UV-A and UV-B exposure (Rangel et al., 2011; Rangel et al., 2015; Chandra Teja and Rahman, 2016; Paixão et al., 2017; Brancini et al., 2021; Corval et al., 2024; Liu et al., 2024). In this respect, it is of little surprise that this metabolic plasticity is also reflected in the comparable broad chemodiversity within the genus, which is able to produce a plethora of secondary metabolites such as destruxins, serinocyclins, metachelins, fungerins, aurovertins, metacridamides, helvolic acids, metarhizins, ovalicins, cytochalasins, or swainsonine (Yadav et al., 2019).

Among the predominant secondary metabolites with highly interesting bioactivity spectra are the so-called destruxins: cyclic hexadepsipeptides with up to date over forty known analogs (Barelli et al., 2022), which differ due to structural variations either in the R-group of the amino acid residues, the hydroxy acid, or the N-methylation (Pedras et al., 2002; Taibon et al., 2014). Quantity and the spectrum of the excreted destruxins also appear to be closely linked to insect specificity and plant host preference, highlighting the importance of these metabolites in insect-fungal and plant-fungal interactions (Xu et al., 2016; Barelli et al., 2022). Aside from their insecticidal and phytotoxic properties, destruxins have garnered increasing interest in medical research due to their broad bioactivity spectrum, which comprises amongst others antiviral, antimicrobial, anticancer, antiangiogenic, antiproliverative and immunosuppressive traits (Chen et al., 1997; Pedras et al., 2002; Liu and Tzeng, 2012; Dornetshuber-Fleiss et al., 2013).

The precise trigger factors leading to destruxin production in Metarhizium species are still somewhat unclear and at times even contradictory. Apart from species and strain specific differences (Amiri-Besheli et al., 2000; Barelli et al., 2022), a number of general key factors have been identified, which are typical trigger factors for secondary metabolite production in fungi: medium composition and cultivation form (Liu et al., 2000; Wang et al., 2004; Barelli et al., 2022), pH value (Wang et al., 2004; Hu et al., 2006; Liu et al., 2007), oxygen availability (Liu et al., 2007; Dong et al., 2016), temperature (Hu et al., 2006; Borisade et al., 2016), oxidative stress and culture age (Wang et al., 2005), the presence of other organisms (Barelli et al., 2022) or external stressors such as the mycotoxin zearalenone (Nowak et al., 2024). Although light is known to trigger the production of various secondary metabolites in fungi (Yu et al., 2023a), it is still unclear if this also holds true for destruxin biosynthesis. However, there is a growing body of evidence that Metarhizium species are highly susceptible to changes in the environmental light conditions, which significantly affect the growth, morphology, virulence, physiology, or stress tolerance of these fungi (Oliveira et al., 2018; Brancini et al., 2019; Dias et al., 2021b; Brancini et al., 2022; Vrabl et al., 2023; Corval et al., 2024).

Considering (i) the diverse light conditions Metarhizium species encounter in nature (Brancini et al., 2022), (ii) the influence of light on the virulence in these organisms, and (iii) the role of destruxins in insect-fungal and plant-fungal interactions (Xu et al., 2016; Barelli et al., 2022), we hypothesized that the biosynthesis of these metabolites is responsive towards wavelength-specific changes in the illumination conditions. If light indeed affects destruxin biosynthesis, we furthermore hypothesized that its effect might have a modulatory role rather than being a simple trigger factor (e.g., the effects of different wavelengths are strongly dependent on nutritional conditions). This second hypothesis was based on two main observations: (i) In fungi there is increasing evidence for a tight interplay between nutritional conditions and the extent or nature of the light response (Tisch and Schmoll, 2010; Yu et al., 2023a). (ii) Previous studies, including our own work (Kato et al., 2003; Atoui et al., 2010; Vrabl et al., 2022), indicated that the nutritional state can be decisive in determining if and how light triggers the production of secondary metabolites.

In this study, we aimed to test these two aforementioned hypotheses: First, we aimed to investigate whether destruxin A, B, and E production in the biotechnologically relevant production strain Metarhizium brunneum BIPESCO 5 is affected by changes in the illumination conditions (i.e., darkness versus red, green, blue, and UV-A exposure) using standardized bioreactor batch cultures. Secondly, we aimed to test whether wavelength-specific effects are independent of the nutritional state of the culture. The latter is highly relevant to potential biotechnological exploitation and to our understanding of the physiological and ecological roles of these valuable metabolites.

Chemicals

Unless stated otherwise, all chemicals were purchased from Merck (Darmstadt, Germany). Glucose and agar were obtained from Roth (Karlsruhe, Germany). Prior to this study, standards for destruxin A, B, and E analysis (NMR purity >98%) were isolated in-house from Metarhizium brunneum strain BIPESCO 5 (CBS 123710) as described earlier (Seger et al., 2006).

Organism and precultures

Metarhizium brunneum strain BIPESCO 5 (CBS 123710) was obtained from the BIPESCO Team Innsbruck (Institute of Microbiology, University of Innsbruck, Austria) and used throughout the study.

For conidial production, the fungus was cultivated on a Saboraud-4 %-glucose (S4G) agar plate and incubated for 12 days at 25 °C under uncontrolled light conditions and a relative air humidity of approximately 60%. The conidia were then washed off the plate with 10 mL of 0.1% Tween 80. To remove any remaining mycelia from the conidial suspension, the solution was filtered through a pipette stuffed with cotton wool (Newmeyer, 1990). The conidia were counted and adjusted to 2 × 10 7 mL -1. Aliquots á 1 mL were stored in cryovials at −80 °C.

The precultures were cultivated in a 500 mL Erlenmeyer flask containing 100 mL Sabouraud-4 %-glucose (S4G) medium (i.e., Sabouraud-2 %-dextrose broth and 2% D (+) Glucose). For inoculation, 1 mL of the conidial suspension was thawed and then transferred under sterile conditions into the medium. Subsequently, the culture was grown in darkness on a rotary shaker at 110 rpm for 72 h at 25 °C and a relative air humidity of approximately 60%.

Bioreactor batch experiments

Media

For the bioreactor experiments, the fungus was grown in two different media: As a complex medium, S4G broth was used, which was additionally supplemented with 1 g L −1 antifoam agent (Clerol FBA 5075, Cognis, Germany). To achieve defined nutrient-limited conditions, a defined ammonium-limited minimal medium (N-lim) was applied as previously described (Vrabl et al., 2019) with the following composition (mM): glucose × 1 H 2 O (400), (NH 4) 2 SO 4 (5), NH 4 Cl (10), KH 2 PO 4 (5.8), MgSO 4 × 7 H 2 O (1.6). In addition, the medium was supplemented with 1 g L −1 antifoam (Clerol FBA 5075, Cognis, Germany). Glucose and salts were autoclaved separately and, after cooling to room temperature, combined aseptically. Afterwards, a trace element solution was sterilized by filtration through a 0.22 μm cellulose acetate filter and added aseptically (10 mL L −1) to the combined medium. The trace element solution consisted of (g L -1): Fe(II)SO 4 × 7 H 2 O (1.0), Mn(II)SO 4 × 1H 2 O (0.46), ZnCl 2 (0.4), Cu(II)SO 4 × 5 H 2 O (0.1), and CaCl 2 × 2 H 2 O (0.6). To avoid precipitation, the trace elements were filled up to 1 L with distilled water, and the pH was adjusted to pH 3 with 5 M HCl.

Inoculation of the bioreactor

Prior to the inoculation of the bioreactor, the mycelium of the preculture was separated from the medium under aseptic conditions using a sterile nylon mesh (pore size 60 µm). The resulting mycelial mass was washed with a two-fold volume of sterile deionized water and subsequently transferred under aseptic conditions into the bioreactor. For the experiments using S4G as the medium, the bioreactor medium was inoculated with a single preculture. Due to the significantly lower growth rates in the N-limited minimal medium, the bioreactors were inoculated with three precultures.

Cultivation conditions

All bioreactor batch experiments were performed in MiniFors 2 bioreactors (Infors HT, Bottmingen, Switzerland) with a working volume of 4.5 L. The culture was stirred at 800 rpm with two Rushton impellers (one at the bottom of the stirrer shaft and the second one 12 cm above with aligned blades). In addition, the bioreactor vessel was equipped with baffles. Throughout the cultivation, the temperature was kept constant at 25 °C, the aeration at 0.6 vvm (volume of gas per volume of medium per minute), and the pH with 5 M NaOH at pH 5.6 for experiments with the S4G medium (i.e., initial pH of the native medium) and at pH 5 for N-lim experiments. Carbon dioxide evolution and oxygen consumption were measured with an exit gas analyzer (Infors HT, Bottmingen, Switzerland).

Irradiation conditions

The irradiation of the bioreactors was performed using a slightly modified setup as previously described (Vrabl et al., 2022), which was expanded by an LED module fitted with Osram Oslon SSL 150 LEDs for green light and a UV module fitted with Seoul Viosys UV-A LEDs. For each tested nutrient medium, five illumination conditions were applied, comprising either constant darkness, or red light (660 ± 25 nm), green light (528 ± 33 nm), or blue light (455 ± 20 nm) at an intensity of 1.4 mW cm −2 as well as UV-A (340/375 ± 11 nm) irradiation at an intensity of 1.1 mW cm −2 (see also Supplementary Figure S1; Supplementary Table S1). The LED-panels were placed in front of the bioreactor vessels (Figure 1) and subsequently covered with a light-tight fabric to avoid light contamination. This fabric was also used for cultivations in complete darkness.

Light measurements

Irradiance was determined using a PM100D console and an S120VC photodiode power sensor, which had a measurement uncertainty of ±5% (280–439 nm) (Thorlabs, Newton, NJ, United States). Spectral power distributions were acquired using a MAYA 2000 Pro spectrometer (Ocean Insights, Orlando, Florida, United States), which was equipped with an HC-1 diffraction grating and a 5 µm entrance slit, providing a spectral resolution of 0.66 nm full width at half maximum (FWHM). The spectrometer was also fitted with a QP600-1-SR-BX fiber optic (Ocean Insights, Orlando, Florida, United States) and a CC-3-UV-S cosine corrector (Ocean Insights, Orlando, Florida, United States). Peak wavelength measurements were performed using a BTS256-EF general lighting spectral luminance meter (Gigahertz-Optik, Munich, Germany) with a measurement uncertainty of ±1 nm. Photon flux density was determined from measured irradiance and peak wavelength as described earlier (Walter and Schöbel, 2023).

Sampling

Each batch experiment (Figure 2) was conducted over a period of 90 h, whereby for each tested condition, at least two independent bioreactor batch cultivations were performed with triplicate samples at each sampling time point. The initial sample was collected before inoculation, while subsequent samples were taken immediately after inoculation (0 h) and at 18, 20, 22, 24, 42, 46, 66, 70, and 90 h. Prior to collecting the actual sample, approximately 15 mL were withdrawn from the bioreactor with a syringe to clear the dead volume of the sampling pipe. In the next step, depending on the biomass concentration in the culture broth, a total of approximately 30–40 mL was collected from the bioreactor, which was subsequently used to determine the mycelial dry mass, residual nutrient concentrations, and destruxin concentrations (see 2.4 Analytics).

Analytics

Dry mass

For gravimetric dry mass (DM) estimation, a defined volume of culture broth (either 5 or 10 mL, depending on the biomass concentration of the culture broth) was suction-filtered through a cellulose acetate filter (pore size 0.45 µm; Sartorius, Germany). The resulting filtrate was quickly transferred into Eppendorf tubes, immediately frozen at −40 °C (FRYKAtherm KB 18-40, FRYKA-Kältetechnik GmbH, Esslingen, Germany), and subsequently transferred to −20 °C until later analysis. The filter cake was dried to constant weight at 105 °C for 24 h in a Memmert UM 400 incubator (Memmert, Schwabach, Germany). After drying, the filters were allowed to cool at room temperature in a vacuum desiccator. The final dry mass was then determined gravimetrically by subtracting the initial filter weight from the post-drying weight.

Residual nutrients

Photometric quantification of residual glucose, ammonium, and phosphate in the culture filtrates from the triplicate samples at each sampling time point was conducted as previously described (Vrabl et al., 2012): Glucose was measured using the enzymatic GOPOD kit (K-GLUC; Megazyme International Ireland Ltd., Wicklow, Ireland). Ammonium was determined with a modified Berthelot (indophenol) assay (Rhine et al., 1998), and phosphate was measured using a molybdate-based colorimetric method (Illmer, 1996).

Destruxins

Culture filtrate processing with solid phase extraction and subsequent HPLC–DAD-MS/MS analysis was performed following a previously established protocol (Taibon and Strasser, 2016). The destruxins A, B, and E as, the most abundant species in the samples, were quantified. Calibration was performed using destruxins A, B, and E, respectively, at concentrations ranging from 1000 μg mL −1 to 1 μg mL −1 (r 2 ≥ 0.99) using pure substances. The content of destruxins from the samples was quantified via the area under the curve. Data analysis was carried out using Quant analysis (Bruker Daltonics, Bremen, Germany) as software for quantification.

Delimitation of growth phases

The end of the exponential growth phase was estimated as previously described (Vrabl et al., 2019) using a combination of nutritional parameters (i.e., depletion of the first limiting nutrient) and characteristic shifts in available online parameters such as respiration and/or pH (examples see Supplementary Figure S2).

Specific growth rates

The specific growth rate (µ) during the exponential growth phase was derived from the slope of the carbon dioxide evolution curve.

Graphics and statistical analysis

The experimental workflow was created via Biorender (BioRender, Toronto, Ontario, Canada). Photomaterials were arranged in Microsoft Powerpoint (Microsoft Corporation, Redmond, Washington, United States). All further figures were done with OriginPro 2025 (OriginLab Corporation, Northampton, Massachusetts, United States).

Time resolution single bioreactor batch cultivations

In this study, a total of 21 independent bioreactor cultivations were performed. For each bioreactor batch experiment, triplicate samples were taken at defined sampling time points in course of the cultivation and were presented as mean ± standard deviation (SD). Mean and SD were calculated in Microsoft Excel (Microsoft Corporation, Redmond, Washington, United States). Each experimental condition (i.e., nutritional condition and light regime) was performed in independent duplicate bioreactor batch cultivations, except for the green light treatment in the N-lim medium, which was performed in triplicate. The time course of each single bioreactor batch cultivation can be found in the Supplementary Figures S4–S7.

Comparison of endpoint data

Destruxin production was assessed by comparing the endpoints of all bioreactor batch cultivations after 90 h of cultivation. All statistical analyses were performed separately for each cultivation medium, with light regime treated as the experimental factor. Statistical analyses were performed using OriginPro 2025 (OriginLab Corporation, Northampton, Massachusetts, United States). The data were tested for normality using the Shapiro-Wilk test. The homogeneity of variance was assessed via Levene’s Test for equality of variances. In case the data met the assumptions of normality and homoscedasticity, the differences among light treatments were analyzed using one-way analysis of variance (ANOVA). When significant main effects were detected, pairwise comparisons were performed using the Sidak post hoc test. If the assumption of normality was violated, differences among light treatments were analyzed using the Kruskal–Wallis test. To assess relationships between destruxin A and destruxin B concentrations, Pearson correlation analyses were performed for each medium individually as well as for the combined dataset. Differences were considered statistically significant at p < 0.05.

Overall growth and destruxin production dynamics

The effect of various illumination conditions on the destruxin A, B, and E production of M. brunneum BIPESCO 5 was explored in two experimental series that differed significantly in nutritional conditions. In the first series of experiments, the fungus was grown in the full medium Saboraud-4 %-glucose (S4G), while in the second set of experiments, M. brunneum was cultivated in a defined nitrogen-limited (N-lim) minimal medium. Independent of the illumination conditions, the growth of M. brunneum followed the typical growth curve of filamentous fungi under carbon excess conditions (Supplementary Figure S4): a lag phase was followed by an exponential growth phase which was terminated once the first essential nutrient was exhausted (i.e., phosphate for S4G-medium, ammonium for the nitrogen-limited minimal medium; Supplementary Figure S2). Afterward, a subsequent prolonged secondary growth phase occurred before the cultures entered the stationary growth phase. The overall culture pigmentation increased in all cultures after the exponential growth phase had ended (Supplementary Figures S8, S9).

Despite a higher inoculum density, M. brunneum cultures grown in the N-lim minimal medium showed a significantly slower growth (specific growth rates from µ = 0.041 h -1 to 0.069 h −1) compared to cultures grown in the S4G-medium (specific growth rates ranging from µ = 0.11 h −1 to 0.16 h −1). Consequently, the biomass concentration of the S4G-grown cultures was approximately twice as high after 90 h of cultivation time as in cultures cultivated in the minimal medium (see Supplementary Figure S3).

In all cultures, destruxin A, B, and E production started after the exponential growth phase was terminated and increased most of the time steadily until the end of the cultivation at t = 90 h (Supplementary Figures S5–S7).

Effect of the medium composition on the destruxin production

To evaluate the effect of medium composition on destruxin production in M. brunneum, we compared endpoint data from all bioreactor batch cultivations at t = 90 h. While the destruxin production occurred under all tested conditions (Figures 3–5), there was a clear medium dependence with respect to destruxin quantity, which was even more pronounced when the data were normalized to biomass (mg g -1 DM).

Destruxin A

Considering all tested illumination conditions, the destruxin A concentrations were roughly twice as high in the N-lim minimal medium (up to 42.9 mg L −1, darkness) compared to cultures grown in the nutrient-rich S4G medium (up to 19.1 mg L −1, green light) (Figures 3A, 4A). This trend was even more distinct in the biomass normalized data (Figures 3B, 4B, 5A), where up to five times more destruxin A was formed per gram dry mass in N-lim cultures (i.e., up to 8.6 mg g −1 DM, darkness) than in S4G grown cultures (i.e., up to 1.7 mg g −1 DM, green light).

Destruxin B

Similar to destruxin A, the destruxin B concentrations of all experiments taken together were approximately twice as high in cultures grown in the N-lim minimal medium (up to 22.7 mg L −1, darkness) than in the S4G medium (up to 8.6 mg L −1, blue light) (Figures 3C, 4C). Again, this trend was more pronounced in the biomass normalized data (Figures 3D, 4D, 5B), where almost up to seven times more destruxin B per gram dry mass was present in the N-lim cultures (i.e., 4.7 mg g −1 DM, darkness) than in S4G grown cultures (i.e., up to 0.7 mg g −1 DM, green light).

Destruxin E

In contrast to destruxin A and destruxin B, strong fluctuations in the destruxin E content could be observed in the parallels (Figures 3E, 4E). Interestingly, the absolute destruxin E levels were higher in cultivations with the complex medium than in the minimal medium. However, this difference was absent once the data were normalized to biomass (Figures 3F, 4F, 5C).

Destruxin ratios

Independent of the illumination conditions, the destruxin A to B ratios remained more or less stable within one cultivation medium, i.e., in the S4G medium, the ratio was roughly around 70:30, while the N-lim medium resulted in a ratio of roughly 60:40 (Figure 6). Due to the large variations of the destruxin E concentrations, which are possibly due to the degradation to destruxin E-diol (Seger et al., 2004), the ratios between all measured destruxins (A:B:E) were not taken into account.

Effect of light on destruxin production

To examine how light influences destruxin production in M. brunneum, endpoint data from all bioreactor batches at t = 90 h were analyzed separately for cultures grown in S4G medium and under nitrogen-limited conditions, with light quality treated as the experimental factor. Intriguingly, we found that destruxin production was light-responsive, but the effect was reversed depending on the nutritional conditions of the culture.

Destruxin A

In cultures grown in S4G medium, the light regime had a significant effect on destruxin A production (one-way ANOVA, F(4,5) = 30.35, p = 0.001). Post-hoc comparisons using the Sidak test showed that cultures exposed to short-wavelength light (UV-A, blue, and green) produced significantly more destruxin A than cultures grown under red light or in darkness (p < 0.05; Figure 3A). No significant differences were detected among UV-A, blue, and green light treatments, nor between red light and darkness (Sidak test, p > 0.05). When destruxin A concentrations were normalized to fungal biomass, the light regime still had a significant effect on destruxin production (one-way ANOVA, F(4,5) = 10.79, p = 0.011). Post-hoc comparisons revealed that only cultures incubated under green light produced significantly more destruxin A per gram dry mass than cultures grown in darkness (Sidak test, p < 0.05; Figure 3B).

Under ammonium limitation, the light regime significantly affected destruxin A production (one-way ANOVA, F(4,6) = 7.99, p = 0.014; Figure 4A). Post-hoc comparisons using the Sidak test showed that significantly more destruxin A was formed in the dark than under the influence of UV-A irradiation (p < 0.05). Although cultures exposed to red light exhibited higher destruxin A concentrations than cultures exposed to shorter wavelength light (UV-A, blue light, and green light), these differences were not statistically significant (Sidak test p > 0.05). Once the data were normalized to fungal biomass, the effect of light regime remainded significant (one-way ANOVA, F(4,6) = 16.63, p = 0.002, Figure 4B). Cultures grown in darkness produced significantly more destruxin A per gram dry mass than those subjected to short wavelength light (UV-A, blue light, and green light) (Sidak test, p < 0.05). Also, red light-exposed cultures excreted significantly more destruxin A per gram dry mass than those exposed to shorter wavelength light (Sidak test, p < 0.05).

Destruxin B

Similar to destruxin A, the production of destruxin B in the S4G medium tended to be higher in short-wavelength light (UV-A, blue light, green light) exposed cultures than in red light or in darkness, although these differences were not statistically significant (Kruskal–Wallis test, χ 2 (4) = 7.31, p = 0.12, Figure 3C). When destruxin B concentrations were normalized to fungal biomass, the effect of light regime became significant (one-way ANOVA, F(4,5) = 44.18, p < 0.001, Figure 3D). Post-hoc comparisons using the Sidak test revealed that destruxin B production increased with shorter wavelengths, meaning that UV-A, blue-light and green-light irradiated cultures had significantly more destruxin B in the culture filtrate compared to cultures exposed to red light or darkness (p < 0.05).

In the N-lim minimal medium, the effect of the light regime on destruxin B production was reversed compared to S4G-grown cultures of M. brunneum (one-way ANOVA, F(4,6) = 9.61, p = 0.009; Figure 4C). Post-hoc comparisons using the Sidak test showed that cultures grown in darkness or exposed to red light produced significantly more destruxin B than cultures exposed to UV-A (p < 0.05) Following data normalization to fungal biomass, the effects of light remained significant but less pronounced (one-way ANOVA, F(4,6) = 9.92, p = 0.008; Figure 4D). Cultures grown in darkness produced significantly higher amounts of destruxin B per gram dry mass compared to those that were exposed to green light or to UV-A light (Sidak test, p < 0.05).

Destruxin E

No significant effect of the light regime on destruxin E production was detected in either cultivation medium (one-way ANOVA, p > 0.05) (Figures 3E,F, 4E,F).

Destruxin ratios and correlations

Although exposure to light altered the concentration of destruxin A and destruxin B, the ratios between the two metabolites remained unaffected for both media (Figure 6). Due to the strong fluctuations in the destruxin E concentrations, the ratios of all analyzed destruxins (A:B:E) were not calculated. Despite stable destruxin A:B ratios, the destruxin A and B concentrations were strongly and positively correlated across wavelength-specific light treatments (Figure 7). This correlation was significant in both media (S4G (Pearson correlation, r = 0.80, p < 0.05, n = 10); N-lim (Pearson correlation, r = 0.94, p < 0.05, n = 11)). When data from both media were combined, the correlation remained highly significant (Pearson correlation, r = 0.97, p < 0.05, n = 21).

Destruxins are secondary metabolites of high agricultural, medical, and biotechnological interest. So far, numerous key factors have been identified which trigger or impede destruxin synthesis in Metarhizium species. Yet, it is still unclear whether or not light affects their production, as is the case for many other fungal secondary metabolites (Yu et al., 2023a). In Metarhizium species, light has been identified to be both, a regulatory signal and a stress factor, which trigger a multitude of cellular responses (Brancini et al., 2019; Brancini et al., 2022). For instance, exposing germinating conidia or growing mycelium to light strongly influenced the subsequent tolerance towards external stressors such as heat, osmotic pressure, or UV-B irradiation, and affected the virulence against host insects (Brancini et al., 2022). Interestingly, the observed light responses were not only dependent on the applied wavelength or the duration of the exposure, but were cross-linked to other factors such as nutritional conditions (Rangel et al., 2015; Brancini et al., 2016; Brancini et al., 2021; Dias et al., 2021a; Dias et al., 2021b). The latter is not restricted to Metarhizium species per se: Several studies highlighted that light can also have a modulatory effect on fungal metabolism and can trigger a different cellular response (e. g. secondary metabolite production) in a fungus depending on its nutritional state (i.e., nutrient-dependent light response; see Section 4.2).

In this study, we investigated for the first time whether the production of destruxin A, B, and E in Metarhizium brunneum BIPESCO 5 is affected by illumination conditions and whether the effects of light differ between two tested nutritional conditions. Of particular interest was a potential interaction between light and nutritional status, which would suggest a light-dependent modulation of metabolic fluxes in Metarhizium species, as has been observed in other fungi.

Meta perspective: overall growth and medium-dependent destruxin production

For the purpose of the study, the fungus was cultivated in two nutritionally distinct media: either in a Saboraud-4 %-glucose medium (S4G), a nutrient rich medium, which is frequently used for organisms of this genus, or in a defined nitrogen-limited (N-lim) minimal medium. The latter was chosen because nitrogen limitation is one known factor which triggers destruxin synthesis in this genus.

Overall, the growth curves of M. brunneum BIPESCO 5 exhibited in both tested media substantially prolonged deceleration phases after the first essential nutrient was depleted and before the culture entered the stationary phase, which is common for filamentous fungi grown under carbon excess conditions (Vrabl et al., 2019). For the S4G medium, it can be deducted from a combination of online and offline-parameters that the cultures were most probably facing phosphate limited conditions (Supplementary Figure S2): phosphate was depleted very early on during the experiment, and the corresponding respiration curve was typical for fungi under phosphate limited conditions (Vrabl et al., 2019). The calculated growth rates were within the reported range of growth rates in literature for Metarhizium sp., i.e., 0.02–0.16 h -1 (Dorta et al., 1996; Tamerler et al., 1998; van Breukelen et al., 2011), whereby the N-lim minimal medium supported lower specific growth rates and final biomass concentrations than the cultures grown in the complex S4G medium.

From a meta-perspective, the production of destruxin A, B, and E started under all tested conditions after the onset of the first depleted nutrient, as is frequently observed for secondary metabolites (Tudzynski, 2014). Both, the absolute and biomass-normalized amounts of destruxin A and B in the two tested media, along with the observed changes in their ratios under different nutritional conditions, correspond with previously reported data from Metarhizium sp. (Liu et al., 2000; Seger et al., 2004; Wang et al., 2005; Hu et al., 2006; Wang et al., 2004 #162; Golo et al., 2014; Dong et al., 2016)). Interestingly, the destruxin A and B concentrations in the N-limited minimal medium were significantly higher, both in absolute and biomass-normalized terms, than in the nutrient-rich S4G medium. In principle, this observation underlines previous findings from the literature that nutritional conditions, including the type and availability of the nitrogen source, have a considerable influence on the production of these metabolites (Liu et al., 2000; Wang et al., 2004; Barelli et al., 2022). Moreover, it is well-documented that a reduced availability of nitrogen can lead to elevated levels of secondary metabolites in fungi (Tudzynski, 2014). However, beyond that, the exact cause of the observed increased levels in the N-limited minimal medium in our study is unclear as in the genus Metarhizium the regulatory role of nitrogen in secondary metabolism, and destruxin biosynthesis in particular, is still incompletely understood. This is partly due to the fact that reports on whether nitrogen limitation increases destruxin production in this genus vary considerably (or are at times even contradictory). While differences in species or strains, cultivation conditions, growth rates, and analytical approaches may explain some variability in the data, recent studies have highlighted that nitrogen regulation in fungi is far from simple. Besides complex regulatory mechanisms and metabolic reallocations, which are affected by nitrogen availability or the nature of the nitrogen source, nitrogen regulation is further influenced by a tight crosstalk with other environmental signals (Tudzynski, 2014; Yu et al., 2023b). Consequently, many aspects, such as a potential regulation of the destruxin biosynthetic gene cluster via the nitrogen-responsive transcription factor AreA (Tudzynski, 2014), have yet to be explored.

Interplay between light wavelength and nutrient availability on destruxin A and B production

While nutrient availability has already been recognized as a key factor influencing destruxin production in Metarhizium species, our findings identify light wavelength as an additional, yet unexplored, environmental variable. Unfortunately, due to the strong fluctuations in the destruxin E concentrations, which were most probably caused by its hydrolisation to destruxin E-diol (Seger et al., 2004), we were unable to determine any light-dependency for this metabolite.

The light-induced effects in destruxin A and B production were strongly influenced by the nutritional context: (i) Cultures grown in the N-limited minimal medium responded more to changes in illumination conditions than those grown in the nutrient-rich S4G medium (with phosphate as the likely first limiting nutrient). (ii) Shorter wavelengths (UV-A, blue light, and green light) caused opposing effects on the destruxin A and B production depending on nutritional conditions. In the N-limited minimal medium, the exposure to shorter wavelengths reduced both the absolute destruxin concentrations (mg L -1) and the biomass-normalized yields (mg g -1 DM). The opposite was the case in the nutrient-rich S4G medium, where the same wavelengths enhanced the production of these metabolites compared to the cultures grown in darkness. As discussed in more detail later, these results are intriguing, as they suggest that light acts not as a simple on/off switch for destruxin biosynthesis in M. brunneum, but rather modulates metabolic allocation. In this respect, the destruxin A:B ratios are of special interest. Although light strongly affected the absolute concentrations of destruxin A and B, the destruxin A:B ratio remained constant within each nutritional condition (Figure 6), resulting in a strong positive correlation between the two metabolites (Figure 7). This positive correlation between destruxin A and B production was first described for M. anisopliae by Wang and coworkers and led to the hypothesis that the genes encoding these two compounds are clustered within the genome (Wang et al., 2004), a hypothesis which was later confirmed (Wang et al., 2012). To our knowledge, a direct light regulation of this gene cluster has not yet been reported. However, these medium-specific but light-independent destruxin A:B ratios suggest that the metabolic pathway might be governed by the nutritional status to some degree, while light modulates the total flux through this pathway.

Our data add to the increasing body of evidence that nutrient-dependent light responses may be a fundamental, albeit insufficiently understood, aspect of fungal physiology. First observations linking metabolic responses to light with the nutritional state of fungal cultures already date back decades (Carlile, 1965). Since then, nutrient-dependent light responses have been reported to affect a wide range of fungal physiology, such as growth (Friedl et al., 2008b), culture (micro)morphology (Pawlik et al., 2020; Vrabl et al., 2022), sporulation (Friedl et al., 2008a), enzyme production Stappler (Janusz et al., 2016; Stappler et al., 2017; Dias et al., 2020), or sensitivity towards stressors (Dias et al., 2020). From the perspective of the present study, it is of particular interest that the production of several other secondary metabolites was also reported to be a nuanced metabolic response to available nutrients and ambient light conditions. For instance, in Penicillium ochrochloron, the production of the antibiotic polyketide xanthoepocin was strongly influenced by an interplay between nutritional and light conditions (Vrabl et al., 2022). Similarly, in Aspergillus nidulans, Kato and coworkers showed that the influence of light on the biosynthesis of the mycotoxin sterigmatocytosin had opposing results depending on the medium composition: the type of the medium (i.e., complex nutrient rich versus minimal media) and availability of carbon and nitrogen sources were decisive (Kato et al., 2003) whether or not sterigmatocytosin was preferably produced under light exposure or in complete darkness. Furthermore, in this fungus, the light effect could be reversed when the carbon source concentration was altered (Atoui et al., 2010). In both studies, the authors concluded that the effect of light on the biosynthesis of sterigmatocytosin is a complex crosstalk between light and multiple other environmental cues. This is also in line with recent findings in Trichoderma reesei, which revealed that nutritional signaling and light responses are closely linked (Tisch et al., 2014; Stappler et al., 2017). Although the understanding of this phenomenon remains incomplete, recent studies have started to gradually uncover a complex interplay between the light sensitivity of many metabolic pathways and environmental signals (Tisch and Schmoll, 2010; Tisch et al., 2014; Schmoll et al., 2022; Yu et al., 2023a). From a meta-perspective, such a crosstalk between light and nutrient availability (or environmental cues in general) has significant implications for fungal physiology, ecology, and, in particular, our understanding of the regulation of secondary metabolism. For example, apparent contradictions in experimental results for the same organism may not simply arise from minuscule differences in experimental settings, such as nutritional (Atoui et al., 2010) or illumination conditions (Brancini et al., 2022), but rather reflect a multilayered adaptive response to changing environments.

In this respect, the results of this study and the available literature suggest that destruxin biosynthesis in M. brunneum BIPESCO 5 might also be a flexible, finely tuned response to environmental signals, rather than the consequence of a straightforward, on/off metabolic reaction. In principle, such a physiological plasticity aligns with the ecological versatility of Metarhizium species as insect pathogens, (rhizosphere-competent) saprotrophs, and plant symbionts (St Leger and Wang, 2020). These various ways of life demand not only a high metabolic flexibility in response to vastly different and rapidly changing nutritional states, but also to permanently altered illumination conditions. The latter can range from absolute darkness in the soil to green-light-dominated phyllospheres, to sun-exposed surfaces, which are subjected to the diurnal cycle of the sun (i.e., daily changes in irradiance intensity and spectrum). Furthermore, considering the large bioactivity spectrum of destruxins and their apparent nutrient-dependent light response, the question arises whether these metabolites may serve as multifunctional tools to exploit diverse ecological niches. While the role of destruxins in insect pathogenicity is the focus of multiple studies (e.g., (Xu et al., 2016; Barelli et al., 2022), their potential contributions to endophytic plant colonization, defense against herbivores, plant growth promotion, or competition with other microorganisms occupying the same ecological niche remain less clear. Exploring the role of destruxins in these processes is far from being of mere academic interest. Besides providing a deeper understanding of the ecological role of the fungi that produce these compounds, it may also contribute to the design of more effective biocontrol agents or biotechnological approaches for high-yield production of these valuable metabolites. From a biotechnological perspective, carefully co-optimizing nutrient and irradiation regimes during process design could facilitate downstream processing, increase product yields, and shift the destruxin profile toward a preferred analog. However, further mechanistic and scale-up studies will be required to substantiate this potential. Rigorous experimental standardization, including cultivation conditions (Egli, 2015) and light conditions (Schmoll et al., 2022), in terms of both wavelength and intensity, will be pivotal to future work aimed at elucidating the metabolic and molecular mechanisms underlying destruxin production in Metarhizium species.

Conclusions

In this study, we could identify light wavelengths as an additional environmental variable that modulates destruxin A and B biosynthesis in Metarhizium brunneum BIPESCO 5. The effect of light was strongly dependent on the nutritional status of the culture: Shorter wavelengths (UV-A, blue, and green) enhanced destruxin A and B production in the nutrient-rich S4G medium but reduced it in the N-limited minimal medium. Despite wavelength-driven changes in destruxin A and B concentrations, the destruxin A:B ratio remained medium-specific and stable. These observations suggest that the metabolic pathways for destruxin biosynthesis may be activated by nutrient status, while light modulates the overall flux through these pathways. Taken together, our results hint at a complex, nutrient-dependent light response in destruxin biosynthesis, suggesting a finely tuned, interdependent adaptive regulatory network that might facilitate more fitness for niche exploitation.

These findings also align with and extend prior evidence that interactions among light, metabolic pathways, and environmental cues are a fundamental yet underexplored form of metabolic regulation in fungi. Such crosstalk has implications for our understanding of the physiology and ecological functions of these organisms and, consequently, for biotechnological optimization. Instead of treating light as a simple on/off parameter, co-optimizing wavelength and nutrient regime could be a powerful tool for redirecting metabolic fluxes and increasing product yields.