In order to redefine the mannitol pathway in the necrotrophic plant pathogen Botrytis cinerea, we used a targeted deletion strategy of genes encoding two proteins of mannitol metabolism, BcMTDH (B. cinerea mannitol dehydrogenase) and BcMPD (B. cinerea mannitol-1-phosphate dehydrogenase). Mobilization of mannitol and quantification of Bcmpd and Bcmtdh gene transcripts during development and osmotic stress confirmed a role for mannitol as a temporary and disposable carbon storage compound. In order to study metabolic fluxes, we followed conversion of labelled hexoses in wild-type and ΔBcmpd and ΔBcmtdh mutant strains by in vivo NMR spectroscopy. Our results revealed that glucose and fructose were metabolized via the BcMPD and BcMTDH pathways respectively. The existence of a novel mannitol phosphorylation pathway was also suggested by the NMR investigations. This last finding definitively challenged the existence of the originally postulated mannitol cycle in favour of two simultaneously expressed pathways. Finally, physiological and biochemical studies conducted on double deletion mutants (ΔBcmpdΔBcmtdh) showed that mannitol was still produced despite a complete alteration of both mannitol biosynthesis pathways. This strongly suggests that one or several additional undescribed pathways could participate in mannitol metabolism in B. cinerea.
Mannitol is one of the most abundant polyols occurring in nature. It is usually the most abundant soluble carbohydrate within the mycelium . In fungi, several physiological functions have been ascribed to D-mannitol. It has been described as being a reservoir of reducing power , a carbon storage compound accumulated in Agaricus bisporus basidiospores  and necessary for the formation of fruit bodies in Stagonospora nodorum . Mannitol also contributes to stress tolerance in fungi. In Aspergillus niger, mannitol appears to be essential for the protection of spores against cell damage under a variety of stress conditions including cold, drought and oxidative stress . In mycorrhizal fungi mannitol plays the role of a carbon translocation compound that enables fungi to assimilate carbohydrates from plant origin [6,7]. Furthermore, mannitol has a role in fungal–plant interactions. During the biotrophic interaction between Uromyces fabae and Vicia faba, levels of mannitol markedly increase in apoplastic fluids of infected leaves and in spores. Mannitol might have a dual function by sequestrating plant hexoses and protecting against ROS (reactive oxygen species) . A role for mannitol in pathogenicity of Alternaria alternata for its host, tobacco, has been demonstrated [9,10]. Mannitol was secreted by A. alternata in response to tobacco extracts while the tobacco plant expressed an endogenous MTDH (mannitol dehydrogenase) during infection [9,11]. The mannitol secreted by A. alternata could also play an antioxidant role, by quenching ROS, whereas the MTDH induced by the plant in response to the fungal colonization would degrade pathogen-produced mannitol, allowing the ROS-mediated plant defence to be effective against the fungus.
The metabolic pathway for mannitol biosynthesis and catabolism (Figure 1) is well described in filamentous fungi and takes place through the mannitol cycle, involving two pathways . The direct reduction of fructose 6-phosphate into mannitol 1-phosphate involves a MPD (mannitol-1-phosphate dehydrogenase; EC 126.96.36.199). Mannitol 1-phosphate is then dephosphorylated into mannitol via a mannitol-1-phosphate phosphatase. This last reaction was described as irreversible, consequently mannitol degradation is traditionally described as occurring through oxidation of mannitol to fructose via a reversible MTDH (EC 188.8.131.52). Both pathways exist in ascomycetes .
Model showing mannitol metabolism and the main carbohydrate conversion pathways in fungi
However, previous reports have challenged the existence of a mannitol cycle. In S. nodorum, MPD is necessary for mannitol catabolism, whereas MTDH is not [4,13]. A mannitol phosphorylation pathway allowing conversion of mannitol into mannitol 1-phosphate might exist. Mannitol pathway explorations conducted in S. nodorum and A. alternata revealed that mannitol synthesis occurred mainly through MPD [13,14]; mpd-deletion mutant strains revealed that the intracellular mannitol concentration decreased by more than 80%, whereas mtdh strains were almost completely phenotypically wild-type. Although mannitol metabolism through mannitol 1-phosphate appears to be the dominant route, the physiological role of the MTDH branch remains unclear.
In a previous report, we showed that mannitol was accumulated during infection of sunflower by the necrotrophic plant pathogen Botrytis cinerea . During pathogenesis, plant hexoses were converted into mannitol, which was the major soluble carbon compound detected in infected tissues. In our previous experiments, we were unable to detect mannitol in fungal growth medium  and hence we suggested this polyol could be used as a translocation compound to sequester plant hexoses. Moreover, the mannitol mainly accumulated during the final steps of fungal development in planta, being stored when conidiophores emerged. Mannitol could be necessary for spore survival or constitute a source of energy for germination. By means of reverse genetics and metabolic investigations, we have explored the physiological role and functioning of the mannitol pathway in B. cinerea.
Fungal strain and growth conditions
B. cinerea B05.10 was maintained at 21 °C on rich medium as described previously . B. cinerea developmental stages were obtained from mycelia grown for 2–6 days (mycelium and sporulating mycelium) on cellophane sheets deposited on to 2% (w/v) glucose-rich solid medium. Conidia were harvested from 12-day-old mycelium grown on 2% (w/v) glucose-rich solid medium. Germinating conidia were obtained by cultivating 12-days-old spores (approx. 107 cells per ml) in 2% (w/v) glucose-rich liquid medium with shaking at 110 rev./min. For osmotic stress experiments, mycelia were grown for 2 days on cellophane on solid Gamborg medium (pH 5.0) supplemented with 2% (w/v) glucose . Mycelia were then transferred on to fresh solid medium supplemented with 2% (w/v) glucose and 1 M NaCl for 30 min to 24 h. For osmotic stress response analysis by in vitro NMR spectroscopy, mycelia were first transferred on to plates containing radiolabelled fructose [1% (w/v) fructose containing 10% (w/w) 13C-fructose] for 4 h and then transferred for 1 and 4 h on to fresh solid medium supplemented with 2% (w/v) glucose and 1 M NaCl. Mycelia, conidia and germinating conidia were harvested, frozen in liquid nitrogen and then stored at −80 °C. In vivo NMR analyses were performed in a perfusion medium containing 0.1% KNO3, 0.005% KCl, 0.01% MgSO4, 0.01% CaSO4 and 13C-labelled glucose or fructose as indicated. B. cinerea transformant strains were selected on rich medium supplemented with 70 μg/ml hygromycin (InvivoGen) or with 100 μg/ml nourseothricin (Werner BioAgents).
In vitro germination and conidiation assays
Conidia were harvested in 0.5% Tween 80 from 12-day-old mycelium grown on solid rich medium. For in vitro germination assays, 500 conidia of each strain were spread on solid rich medium in 140-mm-diameter plates. Young germinations were counted after 4 days of incubation. For conidiation assays, 1000 conidia were spread on solid rich medium in 24-well plates. After 12 days of growth, conidia produced in each well were harvested and counted.
Phytopathogenicity assays were performed using sunflower cotyledons as hosts as described previously . Cotyledons from 1-week-old germlings were infected at the end of a dark period by depositing a 5-mm mycelium disk near the tip of the leaves. Necrosis was detectable at 24 h.p.i. (hours post infection) by the appearance of a brown colour surrounding the starting point of infection. At 48 h.p.i., half of the cotyledons were macerated and necrosed. The whole cotyledon was infected at 72 h.p.i. Conidiation began at 96 h.p.i. and was achieved at 120 h.p.i.
Plasmid construction and transformation of B. cinerea
Transformation of B. cinerea B05.10 was performed using the hph gene as a selectable marker. To disrupt the BcMPD (B. cinerea MPD)- and BcMTDH (B. cinerea MTDH)-encoding genes, Bcmpd and Bcmtdh constructs containing the hygromycin-resistance cassette (OliC promoter–hph gene coding sequence–tub1 terminator) flanked by 5′- and 3′-Bcmpd and Bcmtdh genomic DNA fragments were constructed (Supplementary Figure S1 available at http://www.BiochemJ.org/bj/427/bj4270323add.htm). The hygromycin cassette was released from pLOB1 (provided by Professor P. Tudzynski, Institut für Botanik, Universität Münster, Münster, Germany) after EcoRI and EcoRV digestion and then cloned in pBKSII (Stratagene). Flanking regions (5′-Bcmpd, 3′-Bcmpd, 5′-Bcmtdh and 3′-Bcmtdh fragments) were amplified using the P1/P2, P3/P4, P5/P6 and P7/P8 primers pairs respectively (Supplementary Table S1 available at http://www.BiochemJ.org/bj/427/bj4270323add.htm and Supplementary Figure S1) and were cloned on each side of the hygromycin cassette (Supplementary Figure S1) to form pTD5 and pTD6. To obtain the pTD10 plasmid, Bcmtdh 5′- and 3′-flanking regions were amplified using P9/P10 and P11/P12 respectively (Supplementary Figure S1 and Supplementary Table S1). Amplified fragments were cloned in pCB04 on each side of the nourseothricin cassette, which contained the Streptomyces nat1 (N-acetyltransferase; conferring nourseothricin resistance) gene flanked by the A. nidulans OliC promoter and the B. cinerea tub1 terminator . The preparation of protoplasts and transformation were adapted from procedures described previously . The fungal cell wall was digested using 50 mg/ml glucanex (Novozymes) for 2 h at 26 °C. Wild-type strains were transformed with 10 μg of NotI-linearized vector pTD5 or pTD6 to obtain ΔBcmtdh and ΔBcmpd mutants respectively. The development of a double mutant was achieved by transforming the ΔBcmpd strain generated above with 10 μg of ScaI-linearized vector pTD10. PCR analyses were performed to ensure that replacement of the gene of interest by the selection cassette had occurred. Genomic DNA from ΔBcmpd transformants was amplified using P13/P14 and P15/P16 primers (Supplementary Table S1), to check the 5′- and 3′-insertions respectively. Genomic DNA from ΔBcmtdh transformants was amplified using P17/P14 primers and P18/P19 primers to verify 5′- and 3′-insertions respectively (Supplementary Table S1). Bcmtdh gene replacement by the nourseothricin cassette was verified using primers P20/P21 and P22/P19 to check 5′- and 3′-insertions respectively (Supplementary Table S1). Southern blot analyses were performed to ensure the single insertions. DNA (5 μg), digested with EcoRI for single mutants or with HindIII for double mutants, was separated on a 0.8% agarose gel and transferred on to a nylon membrane. A 32P-labelled probe (the hph coding sequence for single mutants or the 5′-Bcmtdh flanking region for the double mutants) was obtained from plasmids pLOB1 and pTD5 using the Megaprime™ DNA Labeling system (Amersham Biosciences). Hybridization was carried out as described previously .
RNA isolation and transcript quantification
Biological materials were frozen in liquid nitrogen and kept at −80 °C. Samples were crushed in liquid nitrogen and the total RNA was extracted by phenol/chloroform separation and lithium chloride precipitation . For qRT-PCR (quantitative real-time PCR) experiments, 20 μg of total RNA from each sample was treated with DNAseI (Ambion), to ensure the absence of genomic DNA, and PCR was performed using the DNaseI-treated total RNA as the template and Taq DNA polymerase (MP Biomedicals). The quality of total RNA was verified using the Agilent 2100 Bioanalyzer, Agilent RNA 6000 Nano reagents and RNA Chips. Total DNaseI-treated RNA (5 μg) was treated with Thermoscript reverse transcriptase (Invitrogen) as described by the manufacturer. qRT-PCR experiments were performed with the ABI PRISM 7900HT Sequence Detection System (Applied Biosystems) using the Power SYBR® Green PCR Master Mix (Applied Biosystems) according to the manufacturer's instructions. Relative quantification was based on the 2ΔCT method using Bcact1 and BctubA (actin and tubulin) as calibrator reference genes. As gene expression profiles were similar using both controls, only the results obtained using BcactA transcripts are shown. The amplification reaction was as follows: 95 °C for 10 min; 50 cycles of 95 °C for 15 s and 60 °C for 1 min; 95 °C for 15 s; 60 °C for 15 s; and 95 °C for 15 s. Three independent replicates, prepared from independent biological samples, were analysed. Primers used for qRT-PCR are shown in Supplementary Table S1.
PCA (perchloric acid) extracts were prepared from 5–10 g of B. cinerea mycelia according to the method described in . Values are given in mg per g of fresh weight of fungal material. In vitro (tissue extracts) and in vivo (perfused tissues) spectra were recorded on a Bruker NMR spectrometer (AMX 400, wide bore) equipped either with a 10-mm or a 25-mm multinuclear-probe tuned at 161.9 or 100.6 MHz for 31P- and 13C-analyses. The 2H-resonance of 2H2O was used as lock signal. For in vitro measurements, 13C-NMR and 31P-NMR acquisitions were performed as described in . Assignments and quantifications were made after running a series of spectra of the extracts, together with known amounts of authentic compounds at different pHs, for the extracts submitted to the 31P-NMR analysis. In particular we wished to be certain that the peak of resonance appearing at 4.77 p.p.m. in the samples incubated with mannitol corresponded to mannitol 1-phosphate. In vivo measurements were performed using 48 h.p.i. sunflower cotyledons as described in .
Polyol and sugar detection by TLC
Mannitol was extracted from conidia, germinating conidia, mycelia or conidiating mycelia frozen in liquid nitrogen, which were crushed and then lyophilized. In order to obtain comparable results, 100 mg of desiccated fungal extracts were suspended in 900 μl of water and boiled for 10 min for each sample. Supernatants were analysed by TLC. A sample of 1 μl was deposited on SIL G-25 TLC plates (Macherey–Nagel) and separated in an acetonitril/ethylacetate/propanol/water mixture (17:4:4:3, by vol.). Sugars and polyols were revealed using 0.5% KMnO4 and 1 M NaOH, and identified by using standard sugars or polyols (1 μl of 0.5% standard solution).
Preparation of cell extracts and enzyme assays
Cell extracts were prepared as described in . Lyophilized mycelia were ground, resuspended in extraction buffer (50 mM phosphate buffer, pH 7.0, containing 0.5 mM EDTA and 5 mM 2-mercaptoethanol), then centrifuged for 30 min at 20000 g at 4 °C. Supernatants were collected and desalted on a Zeba Desalt Spin column (ThermoScientific) before enzyme activity measurements. Enzyme assays were performed as described in  using 50 μg of protein in a final volume of 1 ml and by monitoring absorbance changes of NADP+/NADPH at 340 nm. For MTDH activity, the reduction of fructose to mannitol was assayed in a reaction mixture of 10 mM Hepes, pH 7.0, containing 0.6 M fructose and 0.2 mM NADPH. The oxidation of mannitol to fructose was conducted in a reaction mixture of 10 mM Hepes, pH 9.0, containing 0.4 M mannitol and 2 mM NADP. For MPD activity, the reduction of fructose 6-phosphate was assayed in a reaction mixture of 10 mM Hepes, pH 7.0, containing 5 mM fructose 6-phosphate and 0.3 mM NADH. For the reverse reaction, the oxidation of mannitol 1-phosphate was measured using 10 mM Hepes, pH 9.0, containing 5 mM mannitol 1-phosphate and 0.5 mM NAD. Enzyme activity was expressed as nmoles of cofactor (reduced or oxidized NADPH)/min per μg of protein.
Western blot analysis
BcMPD and BcMTDH were produced as His6-tagged fusion proteins in Escherichia coli M15 cells using the QiaExpress® System (Qiagen). cDNAs were amplified by PCR with forward primers containing NcoI restriction sites and reverse primers containing BglII sites (Supplementary Table S1). PCR products were cloned in to pQE60 (Qiagen). Optimal protein expression was achieved using 2 mM IPTG (isopropyl β-D-thiogalactoside) after 4 h of induction at 37 °C and 1 mM IPTG after 20 h of induction at 23 °C for BcMPD and BcMTDH respectively. BcMPD was detected in inclusion bodies, whereas BcMTDH remained in the soluble fraction. Purification of tagged proteins was performed on Ni-NTA (Ni2+-nitrilotriacetate) columns (Qiagen), under denaturing and native conditions for BcMPD and BcMTDH respectively, according to manufacturer's instructions. Anti-BcMPD and anti-BcMTDH sera were obtained by immunization of rabbits with the purified proteins (Covalab) according to protocols approved by the Comité d'Ethique de l'Expérimentation Animale de l'Université de Bourgogne, France. Detection of BcMPD and BcMTDH in the fungal strains was performed by Western blotting. Total cell extract proteins were quantified in order to load 75 μg of each extract in each lane. Proteins were then separated by SDS/PAGE (10% gels) and blotted on to a nitrocellulose membrane. Nitrocellulose membranes were probed with rabbit polyclonal anti-BcMPD (at a dilution of 1:2500) and anti-BcMTDH (at a dilution of 1:2500) antisera. Bands were visualized with ECL (enhanced chemiluminescence) using HRP (horseradish peroxidase)-conjugated goat anti-(rabbit IgG) antibody (at a dilution of 1:40000). The detection was performed as described in the manufacturer's instructions for the SuperSignal® West Pico ECL Western detection kit (Pierce Biotechnology).
RESULTS AND DISCUSSION
Construction of Bcmpd and Bcmtdh mutant strains
To understand the role of mannitol in B. cinerea (Figure 1), we searched for MTDH and MPDH coding sequences in order to construct single and double deletion mutant strains; both Bcmtdh and Bcmpd genes were identified in the genome sequence of B. cinerea (http://www.broad.mit.edu/annotation/genome/botrytis_cinerea/) and a low-stringency Southern blot indicated that only one copy of each gene was present in the B. cinerea genome (results not shown). Each sequence (of length 1314 bp and 862 bp for Bcmpd and Bcmtdh respectively) had one intron. Sequence analyses revealed that BcMPD and BcMTDH had the features of mannitol-1-phosphate 5-dehydrogenases and mannitol-2-dehydrogenases and that the proteins would be 43 and 28 kDa respectively. BcMTDH shares 71% and 78% identity with S. nodorum MDH1 (the homologous MTDH in S. nodorum) and A. alternata MTDH respectively; BcMPD shares 62% and 59% identity with S. nodorum MPD1 and A. alternata MPDH, respectively. BcMPD belongs to the LDR (long-chain dehydrogenases/reductases) superfamily, whereas BcMTDH falls into the SDR (short-chain dehydrogenase/reductase) superfamily [23,24].
To make a Bcmpd deletion construct, a region including 567 bp of the promoter, the ORF (open reading frame) and 328 bp of the sequence 3′ of the gene was replaced by a hygromycin-resistance cassette. In the case of the Bcmtdh deletion vector, 60 bp of the promoter and 423 bp of the coding sequence were replaced by the hygromycin-resistance cassette (Supplementary Figures S1A and S1B). Protoplasts of the B. cinerea B05.10 strain were transformed with the linearized vectors and hygromycin-resistant transformants were tested by PCR (results not shown) and Southern blot analyses (Supplementary Figure S1D). In both cases, two transformants, ΔBcmpd19 and ΔBcmpd24 for Bcmpd, and ΔBcmtdh11 and ΔBcmtdh16, for Bcmtdh, were found to have undergone gene replacement as expected (Supplementary Figure S1D). Gene deletions were confirmed further through transcriptional and Western blot analysis and revealed an absence of transcripts and proteins in the mutants when compared with the wild-type strain (Figures 2A and 2B). Interestingly, when compared with the wild-type strain, Bcmtdh expression was clearly increased in the ΔBcmpd mutant, whereas the Bcmpd transcription level was not affected in the ΔBcmtdh mutant. Western blot experiments showed that BcMPD and BcMTDH proteins seemed to be more abundant in both deletion mutants (Figure 2B). Enzymatic assays (Figure 2C) confirmed that Bcmpd deletion abolished MPD activity. However, in the Bcmtdh mutant, MTDH activity was strongly reduced but not completely eliminated and a low level of MTDH activity remained (Figure 2C). All experiments presented in this study were conducted on both of the independent transformants Bcmpd19 and Bcmpd24 for the BcMPD experiments and Bcmtdh11 and Bcmtdh16 for BcMTDH experiments; in both cases, results were clearly similar for all experiments. Consequently, the results shown in the present paper are for the Bcmpd19 and/or Bcmtdh16 deletion mutant strains.
Effect of mannitol pathway alteration on Bcmpd and Bcmtdh transcript levels, proteins and enzymatic activities
Conversion of hexoses by BcMPD and BcMTDH pathways
To investigate the respective role of the BcMPD and BcMTDH branches in mannitol metabolism, hexose assimilation and conversion through the mannitol pathway was analysed in a plant infection context by in vivo NMR spectroscopy. Transfer rates of 13C1-glucose or 13C2-fructose by ΔBcmpd or ΔBcmtdh strains were followed in real time using sunflower cotyledons infected by B. cinerea (Table 1) and compared with the results obtained previously with the wild-type strain . 13C1-Glucose was converted by the ΔBcmpd and ΔBcmtdh mutant strains into 13C1-trehalose, glycogen and 13C1/6-mannitol as shown previously . Although transfer of 13C1 from glucose to mannitol was strongly impaired in ΔBcmpd, the amounts of labelled trehalose and glycogen were increased. In ΔBcmtdh, the labelling transfer rate of 13C1 from glucose into 13C1/6 mannitol was similar to the wild-type strain, whereas the amounts of labelled trehalose and glycogen were significantly lowered. 13C2-fructose was converted by all the strains into a single compound, mannitol. However, ΔBcmtdh transferred 13C2 from fructose to mannitol with a lower efficiency compared with the wild-type strain (a 30% decrease). In vivo NMR data showed trehalose (and glycogen) labelling during growth in the presence of glucose, but not in the presence of fructose. These results clearly demonstrate that trehalose accumulation is a consequence of the MPD pathway deletion and that it is preferentially used during glucose assimilation by B. cinerea. Trehalose (and glycogen) synthesis is directly linked to glucose 6-phosphate (Figure 1). Mannitol biosynthesis is also connected to this metabolite via fructose 6-phosphate, used as a substrate by BcMPD. Consequently, mannitol synthesis from glucose by BcMPD could deplete the pool of glucose 6-phosphate. ΔBcmpd mutants, which are deficient in mannitol biosynthesis, could produce trehalose and glycogen to prevent glucose 6-phosphate accumulation.
|Labelling transfer rate (μmoles per g of fresh weight per h)|
|3 mM Glucose||3 mM Fructose|
|Allocation of labelled carbon||WT||ΔBcmpd||ΔBcmtdh||WT||ΔBcmpd||ΔBcmtdh|
|Labelling transfer rate (μmoles per g of fresh weight per h)|
|3 mM Glucose||3 mM Fructose|
|Allocation of labelled carbon||WT||ΔBcmpd||ΔBcmtdh||WT||ΔBcmpd||ΔBcmtdh|
Mannitol is mobilized during in vitro development of B. cinerea
Metabolic profiling revealed that mannitol is found in large amounts in developing B. cinerea mycelium . To try to determine the role of mannitol in B. cinerea development, we analysed the sporulation and germination rate. No obvious differences were observed between the wild-type and single mutant strains (Supplementary Figure S2 available at http://www.BiochemJ.org/bj/427/bj4270323add.htm). The lack of effect of mannitol gene deletion on sporulation and/or spore germination prompted us to analyse sugar and polyol content by TLC after growth in the presence of glucose (Figure 3). For this purpose, fungal extracts were analysed during distinctive phases of development: growing mycelium (2-day-old cultures); conidiation (6-day-old cultures); mature conidia (collected from 12-day-old cultures); and germination of conidia (performed for 2, 4 and 6 h).
Sugar content and Bcmpd and Bcmtdh transcript levels during in vitro development
Mannitol, and to a lesser extent trehalose, were accumulated during sporulation and in mature conidia in the wild-type strain. These compounds rapidly degraded during germination and were almost undetectable 2 h after germination initiation (Figure 3A). The ΔBcmtdh strain exhibited a similar profile to wild-type, whereas the sugar profile revealed by TLC was modified in ΔBcmpd. Mannitol content was markedly reduced in mycelium, whereas stored trehalose drastically increased in mature conidia and mycelium (2-day- and 6-day-old). Surprisingly, TLC profiles revealed an almost unaltered mannitol content in mutant strain conidia.
Bcmpd and Bcmtdh gene expression was also analysed during the development phases in the wild-type strain. The results presented in Figure 3(B) show that both genes exhibited similar expression patterns during the in vitro development of the wild-type strain. Despite the dissimilar expression levels, their transcription was particularly high during sporulation (6-day-old cultures) and in mature conidia. In contrast, during germination, transcript and protein levels drastically decreased (Figures 3B and 3C). It should be noticed that we detected higher levels of BcMTDH protein, particularly at 2 h after germination initiation. This could be explained by the major role played by this protein in mannitol degradation during germination. Nevertheless, both mannitol metabolism branches could participate in mannitol accumulation in conidiating mycelium and spores.
Given its high abundance in spores, the idea of mannitol acting as a carbohydrate reserve, as suggested by several reports [25,26], can be considered. The previous results on sporulation and germination rates in fungal pathogens where mannitol metabolism has been affected by gene disruption strategies are controversial. In S. nodorum both mpd and mpd/mtdh mutants affected the production of conidia [4,13], whereas in A. alternata an effect was only seen in the mpd/mtdh double mutant . As in S. nodorum and A. alternata, B. cinerea accumulates mannitol in sporulating mycelium [4,10] but, despite a lower mannitol content, in vitro or in planta conidiation was not affected in the mutants (Supplementary Figure S2). A reduced mannitol concentration could allow sporulation. Moreover, trehalose could be stored or degraded instead of mannitol.
Mannitol is involved in the osmotic stress response in B. cinerea
To further dissect the mannitol pathway, it was necessary to identify physiological conditions that could induce synthesis or degradation of the compound in B. cinerea. An osmotic shock is known to induce remobilization of the intracellular carbon pool in fungi . Most reports have so far excluded a role for mannitol in enabling fungi to cope with osmotic stress. Sclerotinia sclerotiorum, Saccharomyces cerevisiae and A. nidulans store glycerol in response to a hyperosmotic stress [28–30], whereas Magnaporthe grisea, Cladosporium fulvum and S. nodorum accumulate arabitol [31–33]. In B. cinerea, the osmotic stress response is controlled through glycerol accumulation . However, mannitol, which is directly connected to central carbon metabolism could also be implicated in the osmotic stress response. To address this question, we investigated the polyol content of wild-type B. cinerea and the ΔBcmpd and ΔBcmtdh mutant strains during growth under hyperosmotic conditions (Figure 4A).
Intracellular sugar content and Bcmpd and Bcmtdh transcript levels in wild-type, ΔBcmpd and ΔBcmtdh strains during an osmotic stress response
In the wild-type strain, mannitol was transiently degraded during the stress response. Its concentration decreased between 1 and 4 h of stress and then increased from 9 to 24 h. In parallel, glycerol accumulated as mannitol was degraded (Figure 4A). Mannitol degradation during the osmotic stress response could suggest that it has a contribution to osmo-adaptation. Analysis of Bcmpd and Bcmtdh gene expression by qRT-PCR revealed that Bcmpd was transiently down-regulated, whereas Bcmtdh transcription increased as the mannitol pool decreased (Figure 4B). After 1 h of stress, Bcmtdh gene expression reached its maximal level with a 4.5-fold expression increase. BcMPD and BcMTDH protein profiles were analysed by Western blot (Figure 4C). During osmotic stress, BcMPD was degraded from 0 to 1 h, whereas BcMTDH was accumulated. Hence, BcMTDH could be preferentially dedicated to mannitol degradation during osmotic stress response.
In the ΔBcmtdh mutant, the TLC profile was similar to that of the wild-type strain (Figure 4A). ΔBcmpd showed a constant and low mannitol content, whereas trehalose, produced by this mutant strain, was degraded concomitantly, resulting in glycerol accumulation. Whereas in the wild-type and in the ΔBcmtdh mutant the mannitol pool was partially restored from 9 to 24 h of stress, the trehalose pool was not restored in the ΔBcmpd mutant. Expression profiles of Bcmpd or Bcmtdh genes in the wild-type and both mutant strains were similar (Figure 4B). However, Western blot experiments showed that in the ΔBcmtdh strain BcMPD was not degraded during osmotic stress (from 0 to 4 h). Therefore whereas transcriptional control of the Bcmpd gene was the same, post-translational control differed in the wild-type and ΔBcmtdh mutant. The mannitol content decreased during the osmotic stress response in ΔBcmtdh. This could be due to the low MTDH activity that is still detected in ΔBcmtdh cell extracts. On the other hand, BcMPD could participate in mannitol degradation. Consequently, this could suggest, contrary to the proposed cycle , that the mannitol 1-phosphate dephosphorylation reaction could be reversible to allow mannitol degradation through the MPD pathway (Figure 1).
Our results show that mannitol degradation in wild-type and Bcmtdh strains paralleled glycerol accumulation during the osmotic stress response, suggesting that mannitol could constitute a carbon store for B. cinerea, used to quickly synthesize glycerol for osmoprotection. For this purpose, the mannitol pool of the B. cinerea wild-type strain was labelled using 13C2-fructose (fructose is almost solely converted into mannitol in the wild-type strain). An osmotic shock (of 1 M NaCl), in the presence of unlabelled glucose, was applied to mycelium and allocation of the labelled carbon during the osmotic response was then followed by in vitro NMR spectroscopy (Supplementary Figure S3 available at http://www.BiochemJ.org/bj/427/bj4270323add.htm). Before osmotic stress, labelled carbon was found exclusively in mannitol molecules (at C2 and C5), whereas between 1 and 4 h of stress, mannitol was degraded and 50% of the labelled carbon originating from mannitol was allocated for glycerol synthesis and detected in 13C2-glycerol. This result demonstrates that mannitol is not directly implicated in the osmotic stress response, but could contribute indirectly to osmoprotection via carbon relocation to glycerol molecules. This direct connection between mannitol and glycerol in the response to stress underlines the central role of mannitol in B. cinerea carbon metabolism. In wild-type and single mutants, mannitol and trehalose disappearance correlated to glycerol accumulation and, in order to participate in this transient response, trehalose could also be mobilized into glycerol (Figure 4).
Bcmpd deletion reveals a mannitol phosphorylation pathway
Radial growth experiments revealed that ΔBcmtdh mutants were able to grow as well as the wild-type strain on mannitol when used as the sole carbon source (Supplementary Figure S4 available at http://www.BiochemJ.org/bj/427/bj4270323add.htm). Moreover, mannitol degradation occurred in ΔBcmtdh during the osmotic stress response and germination of spores. This degradation might be due to the remaining MTDH activity detected in ΔBcmtdh and could indicate the existence of another pathway to degrade mannitol, probably via mannitol 1-phosphate. To confirm this hypothesis, wild-type, ΔBcmpd and ΔBcmtdh were grown for 2 days in the presence of 2% (w/v) mannitol. Metabolic spectra of mycelial PCA extracts were then analysed by in vitro31P-NMR spectroscopy; 31P-NMR spectra for ΔBcmpd are shown in Figure 5. Mannitol 1-phosphate peaks were detected at 4.77 p.p.m. for each strain only after growth on mannitol. MPD is the only known enzyme in fungi able to produce mannitol 1-phosphate from fructose 6-phosphate . Consequently, biosynthesis of mannitol 1-phosphate in the ΔBcmpd strain requires a pathway which could phosphorylate mannitol into mannitol 1-phosphate. Moreover, Bcmpd and Bcmtdh gene expression analysis, performed after growth of wild-type strain on synthetic medium containing 2% (w/v) mannitol as the sole carbon source, confirmed this hypothesis. We found that Bcmpd expression levels were 35-fold higher than those of Bcmtdh (Supplementary Figure S5 available at http://www.BiochemJ.org/bj/427/bj4270323add.htm). This suggests that mannitol catabolism could occur through phosphorylation of mannitol and subsequent conversion in fructose 6phosphate via the MPD pathway.
Bcmpd deletion reveals a novel mannitol phosphorylation pathway in B. cinerea
Mannitol metabolism in fungi has been proposed to occur as a cycle . In this cycle, mannitol is synthesized through the MPD pathway and degraded via the MTDH pathway. However, several reports do not support the metabolism of mannitol operating as a cycle [4,14]. The ability of the ΔBcmpd strains to use mannitol as sole carbon source clearly indicates that mannitol can be metabolized through other routes, most probably through mannitol 1-phosphate. The remaining objection was the absence of mannitol kinase activity in fungi [2,25]. In the present work, detection of mannitol 1-phosphate when ΔBcmpd mutants were grown on mannitol reveals for the first time the occurrence of a mannitol phosphorylation pathway in fungi. As a matter of fact, this phosphorylation of mannitol might be specific to fungi (or at least to B. cinerea), as it has not been detected in any vascular plant cells or tissues that were pre-incubated in the presence of mannitol (R. Bligny and E. Gout, unpublished work). Accumulation of mannitol 1-phosphate could result from a one-step reaction involving a mannitol kinase and ATP, or from a series of several enzymatic reactions that remain to be discovered.
Deletion of BcMPD and BcMTDH pathways did not abolish mannitol metabolism in B. cinerea and revealed a new metabolic route
Development and pathogenicity were not affected in ΔBcmpd and ΔBcmtdh mutant strains (Supplementary Figure S2). In vitro and in planta conidiation and in vitro spore germination tests also showed that these mutant strains were not affected. Therefore it was necessary to completely disrupt mannitol synthisis by creating a ΔBcmpdΔBcmtdh double mutant in order to assign a physiological role for mannitol.
To construct such a double deletion strain, the ΔBcmpd mutant strain was transformed with a Bcmtdh-disruption cassette containing a nourseothricin-resistance marker (Supplementary Figures S1A–C). To check that deletion and homologous recombination has occurred, nourseothricin-resistant strains were screened by PCR (results not shown) and by Southern blotting (Supplementary Figure S1D). Two strains, db32.3 and db35.2, were selected as double mutants. Transcriptional analyses and Western blotting confirmed deletion of both genes in the double mutants (Figures 2A and 2B). However, as for the ΔBcmtdh single mutant, MTDH activity was still detected in the double-mutant strains, with a reduction of 70% compared with the wild-type strain (Figure 2C).
Radial growth experiments revealed that the double mutants were able to grow as well as the wild-type strain on mannitol as sole carbon source (Supplementary Figure S4). Furthermore, germination, sporulation and infection processes were not altered in ΔBcmpdΔBcmtdh compared with wild-type (Supplementary Figure S2). TLC profiles of intracellular sugars and polyols performed during development of ΔBcmpdΔBcmtdh revealed that mannitol was still produced in double mutants (Figure 6). Whereas the mannitol store was severely lowered in mycelium, it was not affected in conidia (Figure 6). Quantification of intracellular metabolites by in vitro NMR spectroscopy in 2-day-old ΔBcmpdΔBcmtdh mycelium revealed that mannitol was synthesized from both glucose and fructose (Table 2). When compared with the wild-type strain, mannitol content in ΔBcmpdΔBcmtdh mycelium after growth on glucose or fructose was decreased by 70% and 31% respectively. The low MTDH activity still detected in double mutant strains could explain why these mutants are still able to produce mannitol, principally from fructose. Moreover, this compound was still degraded during double mutant spore germination (Figure 6). This may suggest that the remaining MTDH activity is able compensate for the mannitol production in double-mutant conidia, but not in mycelium, and that it is implicated in mannitol degradation during germination. These last findings strongly suggest that mannitol could be synthesized and degraded through other metabolic routes.
|2% (w/v) Glucose||2% (w/v) Fructose|
|2% (w/v) Glucose||2% (w/v) Fructose|
Sugar and polyol content during in vitro development of ΔBcmpdΔBcmtdh
Our results are in agreement with those reported in the case of the A. alternata mtdh/mpd double mutant that was still able to grow on mannitol as the sole carbon source . In that case, the authors suggested that the fungus contained other enzymes that allowed utilization of mannitol as a substrate. The ability of the B. cinerea double mutant to metabolize mannitol, together with the presence of an MTDH activity, suggest that additional unrelated gene(s) encoding an MTDH activity should be present in B. cinerea genome. A new MTDH showing no similarity with any known fungal MTDH has been described in Tuber borchii ; the phylogenetic analysis showed this MTDH to be the first example of a fungal MTDH belonging to the MDR (medium-chain dehydrogenases/reductase) superfamily. Consequently, the T. borchii enzyme identified a new group of proteins forming a distinct subfamily of polyol dehydrogenase . A BLAST search (http://www.broadinstitute.org/annotation/genome/botrytis_cinerea/Home.html) for possible homologous proteins to the novel T. borchii MTDH  was therefore performed. Surprisingly, our BLAST results clearly pointed to a B. cinerea sequence (BC1G_15343) annotated as an alcohol dehydrogenase; this sequence shared 84% identity with the T. borchii MTDH and only 11% identity with BcMTDH. In T. borchii, the MTDH produces mannitol from fructose. Such an enzyme, with a high preference for fructose and mannitol is a good candidate to perform mannitol synthesis from fructose and degrade mannitol during development in the double-mutant B. cinerea strains. However, further experiments are necessary to clearly implicate this new gene in B. cinerea mannitol metabolism.
In conclusion, a potential physiological role has been assigned to the MTDH pathway, in that it is likely to be dedicated to a favourable conversion of fructose into mannitol and, conversely, to mannitol degradation during osmotic stress response. On the other hand, we have shown that the MPD pathway could also be implicated in mannitol catabolism, which would abrogate the existence of a mannitol cycle. Finally, analyses of double mutants revealed the existence of a new mannitol pathway to parallel the MTDH pathway functions. During plant infection, mannitol metabolism could help pathogens to efficiently sequester plant hexoses. Gene expression studies revealed a regulated developmental control for mannitol pathway. The interconnection of mannitol metabolism with the central carbohydrate pathway suggests that these processes could regulate carbon metabolic fluxes. Mannitol then constitutes an easily mobilizable carbon store, used for growth and dissemination, but also to cope with stresses and to maintain hyphal turgor pressure.
Botrytis cinerea mannitol-1-phosphate dehydrogenase
Botrytis cinerea mannitol dehydrogenase
hours post infection
perchloric acid, qRT-PCR, quantitative real-time PCR
reactive oxygen species
Pascale Cotton, Thierry Dulermo and Richard Bligny designed the research. Thierry Dulermo, Richard Bligny and Elisabeth Gout performed the NMR experiments. Thierry Dulermo, Christine Rascle and Pascale Cotton performed the molecular and biochemical experiments. Geneviève Billon-Grand performed the Western blotting experiments. Thierry Dulermo, Pascale Cotton, Richard Bligny, Elisabeth Gout and Christine Rascle analysed the data. Richard Bligny and Christine Rascle contributed to a critical review of the maniscript. Pascale Cotton and Thierry Dulermo wrote the manuscript.
We thank Anne-Marie Boisson (Physiologie Cellulaire & Végétale, Unité Mixte de Recherche 5168, Institut de Recherche en Technologies et Sciences pour le Vivant, Commissariat à l'Energie Atomique, Grenoble, France) for technical assistance and M. Wésolowski-Louvel (Génétique Moléculaire des Levures, Unité Mixte de Recherche 5240, Université Lyon1, Villeurbanne, France) for critical reading of the manuscript prior to submission.
Thierry Dulermo was supported by a doctoral scholarship from the Région Rhône-Alpes (Cluster 9), France.