Carbon Source Dependence and Photostimulation of Conidiation in Hypocrea atroviridis

Hypocrea atroviridis is frequently used as a photomorphogeneticmodel due to its ability to conidiate upon exposure to light.Light is thereby believed to be the primary trigger for sporeformation. In contrast, we show here that conidiation is primarilycarbon source dependent and that illumination plays a catalyticrole; of a total of 95 tested carbon sources, only a small setof carbohydrates, polyols, and sugar acids allowed conidiationin darkness, and on most of them, conidiation was significantlymore strongly expressed in light. In addition, there are alsoa number of carbon sources on which H. atroviridis conidiatesin darkness, but light does not further stimulate the process.Yet on another small set of carbon sources (L-sorbitol, D-fucose,D- and L-arabinose, and erythritol), H. atroviridis shows bettersporulation in darkness than in light. No sporulation was observedon organic acids and amino acids. Mutants with deletions inthe two blue-light receptor proteins BLR-1 and BLR-2 generallyshowed weaker conidiation on a smaller number of carbon sourcesthan did the parental strain, yet they clearly sporulated on15 and 27 of the 95 carbon sources tested, respectively. Ofthe carbon sources supporting sporulation, only 11 supportedthe conidiation of both mutants, suggesting that the BLR-1 andBLR-2 receptors are variously involved in the carbon source-dependentregulation of spore formation. The addition of cyclic AMP, whichhas been reported to lead to conidiation in darkness, both positivelyand negatively affected sporulation and resulted in differenteffects in the parental strain and the two ${Delta}$ blr mutants. Ourdata show that the carbon source is the prime determinant forconidiation and that it influences the organism's regulationof conidiation by means of BLR-1 and BLR-2 and their cross talkwith cyclic AMP.

INTRODUCTION

Top

INTRODUCTION

Hypocrea atroviridis (anamorph Trichoderma atroviride, whichwas previously and is still elsewhere frequently misnamed Trichodermaviride) is able to conidiate upon exposure to light and canthus be used as a photomorphogenetic model organism (2). Incomplete darkness, H. atroviridis has been reported to growinfinitely as a mycelium but, upon nutrient deprivation, initiatesthe development of specialized asexual reproductive structures,conidiophores, and conidia (18). Similarly, a pulse of bluelight (450 to 495 nm) induces sporulation (14).

In Neurospora crassa, the response to blue light is controlledby the zinc finger transcription factors white collar 1 andwhite collar 2 (WC-1 and WC-2, respectively) (19, 21). WC-1and WC-2 interact through PAS domains to form the functionalwhite collar complex (WCC) that binds to the promoters of photoregulatedgenes to rapidly activate transcription in response to light(29). The corresponding orthologues in H. atroviridis, blue-lightregulators 1 and 2 (BLR-1 and BLR-2), have recently been shownto fulfill a similar function (5). Yet the mechanisms by whichthe WCC transduces the signal of light to initiate sporulationin H. atroviridis are not clear yet; following a pulse of bluelight and changes in the plasma membrane potential and in theconcentration of intracellular ATP, the activation of adenylatecyclase (20) and a transient biphasic oscillation in intracellularcyclic AMP (cAMP) levels are the first detectable events (12).Several lines of evidence suggest that this rise in the concentrationof cAMP is important for photoinduced sporulation. The additionof exogenous cAMP promotes sporulation in darkness (1, 22),whereas atropine, a compound known to reduce cAMP levels infungal cells, prevents sporulation even after photoinduction(22). The expression of an antisense version of the pkr1 gene,encoding the regulatory subunit of protein kinase A (PKA), resultedin a nonsporulating phenotype, and the overexpression of thisgene produced colonies that conidiated even in darkness, whereasCasas-Flores et al. showed that overexpression of pkr1 blockedthe induction of early light response genes (6). Interestingly,Casas-Flores et al. also demonstrated that the WCC is necessaryfor conidiation even when conidiation is induced by carbon starvationinstead of illumination but that blue-light-dependent activationof PKA was found to be independent of the WCC. These data suggestcross talk between the WCC and cAMP signaling and between lightand carbon sensing.

Interestingly, in all of these studies, little attention hasbeen paid to the possible regulatory effect of carbon metabolism,which can, however, be expected from the fact that carbon cataboliterepression is known to influence sporulation in bacteria andfungi (9, 26). Chovanec et al. (7) previously reported thatphotoconidiation of H. atroviridis is differently influencedby a small number of carbon sources. The objectives of the presentstudy were to examine these findings on a broader scale andto investigate the roles of the blue-light receptor proteinsBLR-1 and BLR-2 and signaling by cAMP in this process.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Strains.
H. atroviridis strain IMI 206040 (wild type; initially depositedas Trichoderma harzianum IMI 206040 but reidentified by meansof sequence analysis of internal transcribed spacers 1 and 2of the rRNA gene cluster, NCBI GenBank accession no. AF278795)and the two ${Delta}$ blr-1 and ${Delta}$ blr-2 mutants derived from it (5) wereused in this study. All cultures were maintained on plates containing3% potato dextrose agar (PDA) and subcultured monthly.

PM analysis.
The Biolog Phenotype MicroArray (PM) technique (Biolog Inc.,Hayward, CA), as optimized for Trichoderma previously (8), wasused to investigate the effect of constant darkness and lighton carbon assimilation patterns. To prepare the inoculum, thestrains were precultivated on 3% PDA plates for 5 days. Conidiationof the H. atroviridis ${Delta}$ blr-1 and ${Delta}$ blr-2 mutant strains was stimulatedby mechanical injury of the mycelium on PDA with a sterile scalpel(5). Inocula were prepared by rolling a sterile, wetted cottonswab over sporulating areas of the plates. The conidia werethen suspended in sterile Biolog FF inoculating fluid (0.25%Phytagel, 0.03% Tween 40), gently mixed, and adjusted to a transmissionof 70% at 590 nm. Ninety microliters of this conidial suspensionwas then dispensed into each of the wells of the PM plates,and the plates were incubated either with alternating illumination(12 h of 1,800-lx white light) and darkness or with constantdarkness at 28°C. The optical density at 750 nm (OD₇₅₀)(mycelial growth) and OD₄₉₀ (mitochondrial activity) were measuredafter 12, 18, 24, 36, 42, 48, 60, 66, 72, 96, and 168 h usinga microplate reader (Biolog Inc., Hayward, CA). In order toavoid contact of the plates incubated in darkness with light,each data point was assayed with a single plate. Each assaywas repeated three to six times in a series of independent experiments.Due to a generally decreased growth rate of blr mutants, theOD₇₅₀ at 66 h was chosen as a reference time point for the growthassays, since it allowed a comparison of biomass formation onall carbon sources when the majority of growth curves were inthe linear phase.

To study the effects of the addition of dibutyryl-cAMP, stocksolutions in sterile, distilled water were prepared separatelyand added to the corresponding inoculation fluids at a concentrationof 1 mM.

To quantify conidiation, we developed a nominative scale whichdifferentiates between the complete absence of conidiophoresand conidia (0), the formation of immature pustules withoutmature conidia (1), and four levels of conidiation intensity(2 to 5) from weak diffuse sporulation (2) up to the developmentof a thick conidial mat covering the whole well (5). The levelswere distinguished by visual inspection of the plate imagesat a high resolution and cross-checked by plate examinationunder a light microscope. An example is given in the inset ofFig. 1.

View larger version (42K):
[in this window]
[in a new window]

FIG. 1. Summed map of conidiation of H. atroviridis wild-type strain IMI 206040 and two blr deletion mutants after 168 h of incubation on 95 carbon sources and water under conditions of darkness and light or darkness with elevated levels of cAMP. The map was composed after two-way joining cluster analysis applied to (i) the carbon sources and (ii) the fungal strains cultivated under different conditions as two groups of variables. The nine-pointed stars, five-pointed stars, and pentagons correspond to carbon sources providing fast (cluster I), good (cluster II), and decreased (cluster III) mycelial growth of all three strains in darkness as estimated after 48 h of incubation. The respective intensities of sporulation (thumbnail images in the inset) are given by a corresponding color as indicated in the color scale. Examples of conidiation grades are estimated according to the nominative scale, where the complete absence of conidiophores and conidia corresponds to 0, the formation of immature pustules without mature conidia corresponds to 1, and true conidiation is measured by one of four levels of intensity, from weak diffuse sporulation (2) up to the development of a thick conidial mat covering the whole well (5). "light" corresponds to an alternating 12-h-light/12-h-darkness cycle; "cAMP" indicates 1 mM of exogenous cAMP.

Statistical data evaluations.
Data from all experiments were combined into a single matrixand analyzed with the STATISTICA 6.1 software package (StatSoft,Inc., Tulsa, OK). All data were subjected to descriptive statisticalevaluations (mean, minimum, maximum, and standard deviationvalues) and checked for outliers.

Cluster analysis (16, 31) was used to detect groups in the dataset. This method was used to simultaneously group both carbonsources and strains/conditions in a two-way joining analysisaccording to their sporulation profile. In most cases, the cluster-joininganalysis was made with Euclidian distance and complete linkageas the amalgamation rule; i.e., distances between clusters weredetermined by the greatest distance between any two objectsin the different clusters. We used a discrete counter plot,which is a graphical representation of two-way joining results,to obtain a carbon source-specific sporulation map based onthe OD₇₅₀ at 168 h. In the merged sporulation map, each datapoint is represented as a color-coded rectangular region. Thecarbon source order remained intact, but the positions of strains/conditionson the combined map were manually modified according to thelogic of the experiment.

When needed, one-way or main-effect analyses of variance wereused to compare the levels of growth of selected strains onindividual carbon sources. Tukey's honestly significant differencetest as implemented in STATISTICA 6.1 was used for post hoccomparisons to detect the contribution of each variable to themain effect of the F test resulting from the analyses of variance.

RESULTS

Top

RESULTS

Carbon source-specific conidiation of H. atroviridis in darkness.
The species-specific carbon source utilization of H. atroviridisbased on its Biolog PM profile has been characterized by Seidlet al. (with strain ATCC 74058 or P1 [25]) and M. A. Friedlet al. (with strain IMI 206040 [submitted for publication]).In these two studies, all carbon sources were divided by joiningcluster analysis into four standard groups based on the correspondingdynamics of the resulting growth curves. Clusters I, II, andIII thereby contained carbon sources which allow fast, moderate,and slow growth, respectively, and cluster IV contained carbonsources on which H. atroviridis grows very poorly or does notgrow at all. In order to check whether the specific setup ofthe Biolog system would influence any of these results, selectedcarbon sources from each group were also tested on agar platesand on standing liquid cultures. Consistent data were obtained(data not shown), thus ruling out any specific interference.

Using the nominative six-step scale developed in this studyto evaluate conidiation intensity on microplates, we first testedthe effect of carbon sources on spore production in darkness.After 168 h of incubation under these conditions, conidiationwas noted on only 48 out of 95 tested carbon sources (the firsttrack in Fig. 1). These included almost all of the carbohydrates,polyols, and sugar acids but, with the exception of N-acetyl-β-D-glucosamine,not the other sugar amines. Conversely, it did not occur onany amino acids, aliphatic acids, or alcohols, with the exceptionof β-hydroxybutyric acid. In order to learn whether thegrowth rates achieved on these carbon sources were determinativefor conidiation, we compared the relative abundances of conidiationwith the growth rates on these carbon sources (Fig. 2). Fromthis, it is evident that the growth rate itself does not correlatewith conidiation. Among the 48 carbon sources which supportedconidiation in darkness, 9 supported fast growth, 10 supportedmoderate growth, and 15 supported only slow growth (clustersI, II, and III, respectively). The remaining 14 carbon sourceswhich supported conidial development provided very poor mycelialgrowth of the fungus (cluster IV). Therefore, conidiation indarkness is strongly carbon source dependent.

View larger version (18K):
[in this window]
[in a new window]

FIG. 2. Relation between conidiation of H. atroviridis wild-type strain IMI 206040 and its biomass density at 72 h of incubation on 95 carbon sources and water. Shaded circles indicate mycelial growth rates in complete darkness, and open squares correspond to growth rates in alternating 12-h-light/12-h-darkness cycles. The conidiation grades are indicated in Fig. 1. The biomass density (OD₇₅₀) for 72 h is given (instead of a growth rate), because at this time all cultures are still in the linear phase of growth and the respective values directly correlate with the growth rates (25).

Photostimulation of conidiation is also carbon source dependent.
In order to investigate whether the well-known photostimulationof conidiation would overrule the above-noted carbon sourceregulation of sporulation or be confined to those 48 carbonsources where conidiation was already present in darkness, wesubjected the microplates to illumination. The data are shownin Fig. 1 (second track). Those carbon sources on which H. atroviridisdid not conidiate in darkness also did not enable conidiationunder illumination. The presence of light enabled sporulationonly on those carbon sources which also empowered conidiationin darkness by increasing the intensity and speed of conidialdevelopment such that the appearance of conidiophores and conidiawere detected earlier during cultivation (data not shown) andmore intensely at 168 h. However, a couple of exceptions werealso noted: in contrast to the situation in darkness, conidiationunder illumination was not observed on D-arabinose or D-gluconicacid.

The regulatory role of the blue-light receptor proteins BLR-1 and BLR-2 in conidiation is carbon source dependent.
Light-induced stimulation of conidiation in H. atroviridis hasbeen demonstrated to be dependent on the function of the blue-lightreceptor proteins BLR-1 and BLR-2 (5). In order to test whethertheir function would also be carbon source dependent, we madeuse of the ${Delta}$ blr-1 and ${Delta}$ blr-2 mutant strains. As shown in Fig.1, conidiation in the dark is indeed significantly impairedin the two mutants on most carbon sources. On average, bothmutants produced only diffuse pustules or single conidiophoresand rarely conidiated to grade 3 or 4 of our scale. However,strong conidiation clearly occurred on sorbitol with the ${Delta}$ blr-1strain and on glucuronic acid with the ${Delta}$ blr-2 strain. In addition,the ${Delta}$ blr-1 and ${Delta}$ blr-2 strains were able to moderately producespores on 15 and 27 carbon sources, respectively. Therefore,the carbon sources supporting the low level of conidiation inthe two mutants were nearly always from the pool of carbon sourcesthat favor the conidiation of the parent strain. A notable exception,however, was ${alpha}$ -methyl-D-glucoside, on which no spores were detectedin the parent strain IMI 206040; however, both mutants formedmature conidia. A comparison of the two mutants showed thatonly 11 carbon sources supported conidiation in both mutants,suggesting that the BLR-1 and BLR-2 receptors were variouslyinvolved in the carbon source-dependent regulation of sporeformation.

The addition of cAMP alters the carbon source dependence of conidiation.
Casas-Flores et al. (5) have described that the addition ofcAMP induces conidiation in H. atroviridis in darkness. Ourstudy confirmed this finding but also showed cAMP to be carbonsource dependent. The addition of cAMP induced conidiation inthe parent and both blr mutant strains on some but not all carbonsources (see the Biolog data in the supplemental material).The most notable effects were (i) the complete reduction ofconidiation on D-xylose (cluster I) and on L-rhamnose and β-hydroxybutyricacid (both from cluster IV) and (ii) the stimulation of sporulationon D-saccharic acid, on which IMI 206040 did not sporulate withoutcAMP addition (Fig. 1).

The addition of cAMP to the cultivation medium of both ${Delta}$ blr mutantsalso essentially changed the carbon source dependence of sporulation,and these changes were different from those seen with the parentstrain IMI 206040. Thus, while the addition of cAMP reducedthe conidiation of the last strain on D-xylose, the ${Delta}$ blr-1 mutantformed mature conidia on this compound and the ${Delta}$ blr-2 mutantdid not sporulate on D-xylose either with or without the additionof cAMP.

DISCUSSION

Top

DISCUSSION

Conidiation serves a dual function in fungi. It produces reproductivestructures for dispersal and allows survival under the adverseconditions of environmental stress. Obviously, these two functionsoverlap because dispersal from a given habitat is mostly associatedwith one or another type of stress (UV light, starvation, alteredhumidity, and other stresses). Consequently, it would not besurprising for conidiation to be triggered by a variety of suchconditions, of which, so far, only macronutrient limitationand light have received major attention. The results of thisstudy add to this list the type of carbon source as a dominantparameter. Previous studies investigated photoconidiation oncommon fungal solid media, which contain mostly glucose, sucrose,or maltose as the carbon source. Our data on these carbon sourcesessentially confirm the results of others, i.e., the stimulationof conidiation by light (2). However, contrary results wereobtained with other carbon sources (D-fucose, erythritol, L-arabinose,and D-arabinose), and light was essentially without effect on,e.g., rhamnose, glucosamine, and xylitol. Also, strong sporulationwas obtained on a number of carbon sources in the absence oflight. It is important to note in this context that we haveobserved conidiation under every condition tested (darkness,light, the addition of cAMP, the loss of function of BLR-1/BLR-2)only on carbohydrates and their derivatives (polyols and sugaracids) and, in contrast to Chovanec et al. (7), never on aminoacids, aliphatic acids, or aliphatic alcohols. A plausible explanationfor these findings would be that different carbon sources enabledifferent growth rates and that conidiation would thereforebe observed with those sources on which the fungus has, at thetime of analysis, already reached the point of starvation. Alternatively,the fungus may grow on some carbon sources at a very low ratefor maintenance metabolism and therefore may start to sporulate.If this were the case, however, we should have seen a correlationbetween conidiation and either high or low growth rates. However,although a weak relationship between growth rate and conidiationwas seen, it was not sufficient to use this relationship asthe entire explanation, thus indicating that it is the natureof the carbon source that determines whether or not the conidiationof the fungus can be triggered by light, thus making it a parameterof prime importance. Unfortunately, the lack of knowledge aboutthe biochemistry and regulation of catabolism of many of thecarbon sources employed in this study prevents a more detaileddiscussion of the individual results obtained. Yet we wouldnevertheless like to emphasize that it is well known that lightprovokes severe changes in carbon metabolism, such as proteinphosphorylation and low glucose uptake in Aspergillus orneatus(17) and the inhibition of glucose consumption in Physarum polycephalum(24). Recent work, for which global expression analysis wasused for H. atroviridis (23) and Hypocrea jecorina (24a), revealedthat the expression of genes involved in the transport, signaling,and catabolism of monosaccharides and related components, andin mitochondrial energy production, is stimulated by light.We therefore speculate that light stimulates the expressionof genes involved in some metabolic networks but that it doesnot affect (or even inhibit [24a]) that of others. The differentresponses to light on different carbon sources would thereforebe due to the relative levels of dependence of their catabolismon these light-affected metabolic networks.

A similar explanation may be used to explain the different effectsof the addition of cAMP. Data available in the literature suggestthat cAMP is involved in blue-light-induced conidiation in Trichoderma(12, 13, 20), and the addition of exogenous dibutyryl-cAMP toTrichoderma cultures growing on rich medium even triggers conidiationin darkness (1, 6). Again, we confirmed these data with themajor carbon sources used in the literature but obtained divergentand even contrary results with others. This could be relatedto the fact that, unlike what is expected from stimulation bycAMP, conidiation in H. atroviridis and other fungi is regulatedby a pathway which is negatively controlled by PKA (4, 6). Casas-Floreset al. (6) suggested that H. atroviridis uses separate signalingpathways for stimulation by external and intracellular cAMP,i.e., the classical G protein/PKA pathway for the former andan alternate pathway that uses a membrane receptor for exogenouscAMP like the one which has been reported for N. crassa (3).We could confirm the presence of such a receptor gene in theH. atroviridis genome sequence (C. P. Kubicek, unpublished data).Thus, if this receptor is not expressed on one or more of thesecarbon sources, the fungus will not be able to respond to thestimulus. As for the involvement of the G protein/PKA pathway,it is clear that this pathway leads to strongly different internalcAMP pools, depending on whether or not the respective carbonsource involves a G protein-coupled receptor or transporter.

While the ${Delta}$ blr-1 and ${Delta}$ blr-2 mutant strains exhibited stronglydecreased conidiation, our results contribute two novel findings:first, the mutations do not completely block sporulation butonly strongly decrease it and, second, they still show (light-insensitive)sporulation on some carbon sources (notably glucuronic acidand 2-ketogluconic acid). Casas-Flores et al. (6) hypothesizedthat the BLR proteins are not only involved in photosensingbut may also perceive or transduce the signal originating fromthe lack of glucose (e.g., by redox and oxygen sensing via theirPAS domains). This explanation may be further adapted to explainthe results of the present study, thereby using recent findingson the mechanism of apoptosis for comparison. Sporulation andapoptosis share a number of common physiological events, suchas mycelium vacuolization and protein degradation (10, 10a,11). Thrane et al. (30) recently emphasized the role of a caspase-likeactivity, a central marker for apoptosis (32), in Aspergillusnidulans during sporulation. Apoptosis is influenced by thecellular redox potential, i.e., the ratio of oxidized to reducedglutathione (15), which coincides with the necessity of NADPH(27) and glutathione (28) for conidiation. Although we are notaware of any work comparing the glutathione and NADPH poolsand their turnover in fungi on different carbon sources, theirstoichiometric involvements in the catabolism of, e.g., hexosesand pentoses are different. It can therefore be expected thatthe intracellular pools of glutathione and of NADPH/NADP andNADH/NAD will also differ. We therefore extend the model ofCasas-Flores et al. (6), i.e., that the BLR proteins act asredox and oxygen sensors, by hypothesizing that the observeddifferences in levels of conidiation on different carbon sourcesare due to different redox potentials in the cell during growthon them.

Apart from these theoretical considerations, our findings mayalso offer some practical hints. First, strains of H. atroviridisare used as biocontrol agents, and for this purpose, a formulationof conidia is used. Our study may help to design optimal conditionsfor the conidiation of these strains. Second, many fungal mutantstrains have no conidiating phenotype or only poorly conidiate.Our data suggest that this phenotype may not be universal, buta careful investigation may identify carbon sources on whichthe respective strain can still form conidia.

ACKNOWLEDGMENTS

We thank A. Herrera-Estrella, Irapuato, Mexico, for providingus with the ${Delta}$ blr-1 and ${Delta}$ blr-2 mutant strains.

This work was supported by a grant from the Austrian ScienceFoundation (FWF-P17325) to C.P.K.

FOOTNOTES

* Corresponding author. Mailing address: Research Area of Gene Technology and Applied Biochemistry, Institute of Chemical Engineering, Vienna University of Technology, Getreidemarkt 9-1665, A-1060 Vienna, Austria. Phone: 43 1 58801 17202. Fax: 43 1 58801 17299. E-mail: druzhini{at}mail.zserv.tuwien.ac.at

Published ahead of print on 2 November 2007.

Supplemental material for this article may be found at http://aem.asm.org/.

REFERENCES

Top