Global Carbon Utilization Profiles of Wild-Type, Mutant, and Transformant Strains of Hypocrea jecorina

Abstract

The ascomycete Hypocrea jecorina (Trichoderma reesei), an industrial producer of cellulases and hemicellulases, can efficiently degrade plant polysaccharides. However, the catabolic pathways for the resulting monomers and their relationship to enzyme induction are not well known. Here we used the Biolog Phenotype MicroArrays technique to evaluate the growth of H. jecorina on 95 carbon sources. For this purpose, we compared several wild-type isolates, mutants producing different amounts of cellulases, and strains transformed with a heterologous antibiotic resistance marker gene. The wild-type isolates and transformed strains had the highest variation in growth patterns on individual carbon sources. The cellulase mutants were relatively similar to their parental strains. Both in the mutant and in the transformed strains, the most significant changes occurred in utilization of xylitol, erythritol, d-sorbitol, d-ribose, d-galactose, l-arabinose, N-acetyl-d-glucosamine, maltotriose, and β-methyl-glucoside. Increased production of cellulases was negatively correlated with the ability to grow on γ-aminobutyrate, adonitol, and 2-ketogluconate; and positively correlated with that on d-sorbitol and saccharic acid. The reproducibility, relative simplicity, and high resolution (±10% of increase in mycelial density) of the phenotypic microarrays make them a useful tool for the characterization of mutant and transformed strains and for a global analysis of gene function.

The number of fungal species for which genome sequences are available is steadily increasing. The analysis of expressed sequence tags (ESTs) from these fungi regularly identifies genes that do not have orthologs even in closely related species (8, 20, 24). Clearly, genomic and cDNA sequencing is the first step in the identification and validation of gene function. Genome-wide studies of the role of a particular gene usually is made through DNA microarray analyses or proteomics (25). At the same time, phenotypic assays are needed to identify the effects of mutations at the metabolic level and to determine the cellular function of a gene.

The Biolog Phenotype MicroArray (PM) approach (2, 3) is a high throughput system for the identification of carbon sources and other nutrients used for the growth of various microorganisms (2, 3, 15, 27). Biolog PM has general utility for characterizing fungal metabolism (15, 27), but the consistency of the inferred phenotypes in different strains from the same species, mutants, and genetically altered isolates relative to their parental strains often is unknown. We are interested in metabolic engineering of the ascomycete Hypocrea jecorina (Trichoderma reesei) to increase the production of cellulase and other extracellular enzymes. The genome sequence of this fungus is now available (http://gsphere.lanl.gov/trire1/trire1 .home.html), as are cDNA sequences from mycelia grown on glucose or under cellulase-inducing conditions (5, 7, 10, 24). However, knowledge of the genome-wide similarity of physiological profiles in various transformed strains and mutants often is needed to meaningfully interpret the gene expression data.

The objective of the present study was to investigate the global carbon metabolism of H. jecorina per se and to compare it to genetically modified strains obtained after classical mutagenesis or DNA-mediated transformation. We wanted to know (i) if various wild-type isolates differed in growth on individual carbon sources, (ii) whether most of the known H. jecorina mutants used for cellulase research have similar growth rates on individual carbon sources, and (iii) whether DNA-mediated transformation generally alters growth patterns. These data will provide the baseline for PM analysis of global investigations of the physiological effects of knockouts of metabolic and regulatory genes in H. jecorina.

MATERIALS AND METHODS

Strains.

The wild-type and mutant strains of H. jecorina used in the present study, together with their origins and properties, are listed in Table 1.

TABLE 1.

Wild-type and mutant strains of Hypocrea jecorina used in this study

Strain	Nature	Origin and/or propertiesa
QM 6a (ATCC 13631 = CCM F-560 = DSM 768 = IMI 192654 = VTT D-74083)	ex type	Solomon Islands
QM 9123	Mutant of QM 6a	Produces 2× cellulases of QM 6a
QM 9414	Mutant of QM 9123	Produces 5× cellulases as QM 6a
QM 9978	Mutant of QM 9123	Produces no cellulases
TU-6	pyr4 mutant of QM 9414	Uridine auxotroph
G.J.S 85-236	Wild type	Sulawesi, Indonesia, teleomorph
TUB F-1034	Wild type	Taiwan, anamorph
TUB F-430	Wild type	Sri Lanka, anamorph
TUB F-733	Wild type	Parana, Brazil, anamorph
CBS 836.91	Wild type	Para, Brazil, teleomorph
TUB F-1066	Wild type	Northern Argentina, anamorph

^aThe term “teleomorphs” refers to cultures derived from ascospores of H. jecorina, whereas “anamorphs” were directly isolated from soil.

Transformation and genetic analysis of obtained isolates.

H. jecorina strains QM 9414 and QM 9978 (11) were transformed with plasmid pRLMex30, which carries a bacterial hygromycin B phosphotransferase gene (hph) under the control of the H. jecorina pki1 promoter and cbh2 termination signals (17). Transformant isolates were selected on plates containing 50 μg of hygromycin (Calbiochem, San Diego, CA)/ml. DNA was extracted by using standard protocols (1) and subjected to Southern analysis (1) to confirm plasmid integration and to determine the number of inserted copies.

For Southern analyses, DNA was digested to completion with either PstI or EcoRI (each has a single restriction site in the hph gene) and hybridized to a 1.0-kb NsiI/XbaI fragment of hph that does not include any of the H. jecorina homologous, i.e., promoter or terminator, sequences. The number of copies was calculated from the number of bands resulting from different loci of integration and from signal intensities resulting from concatemerization of the cassette. The intensity of the hph fragment bands was quantified densitometrically with the GS-800 Calibrated Densitometer (Bio-Rad, Fremont, CA) after different exposure times. Controls with strains that lacked hph were included.

Biolog Phenotype MicroArray technique.

Global carbon assimilation profiles were evaluated by using Biolog FF MicroPlate (Biolog, Inc., Hayward, CA). The FF MicroPlate test panel contains 95 wells, each with a different carbon-containing compound(s), and one well with water.

H. jecorina strains were grown on 2% (wt/vol) malt extract agar under ambient laboratory conditions with diffuse daylight at 25°C. The inoculum was prepared after conidial maturation (2 to 3 days) by rolling a sterile, wetted cotton swab over conidiating areas. Conidia were suspended in 16 ml of sterile phytagel (0.25% Phytagel, 0.03% Tween 40) in disposable borosilicate test tubes (20 by 150 mm). The spore suspension was mixed in a vortex mixer for 5 s and adjusted to an A₅₉₀ of 75% ± 2%. Then, 90 μl of the conidial suspension was dispensed into each test well. Microplates were incubated in the dark at 30°C, and the A₇₅₀ was used to measure mycelial growth at 12, 18, 24, 36, 42, 48, 66, and 72 h. Each strain was analyzed in at least three independent experiments using separately prepared inocula.

For analysis of the uridine auxotrophic strain H. jecorina TU-6 (11), media also were supplemented with 10 mM uridine.

Statistical analysis.

Data from all experiments were combined in a single matrix and analyzed with the STATISTICA 6.1 (StatSoft, Inc., Tulsa, OK) software package. All data were subjected to descriptive statistical evaluations (mean, minimum, maximum, and standard deviation values) and checked for outliers. Conditions resulting in outliers were reevaluated, and the outlier value was replaced in the analysis if the repeat assay was concordant with the results from the other assays.

Cluster analysis (13, 29) was used to detect groups in the data set. This method was used to group carbon sources utilized by a particular strain, to identify strains with similar utilization profiles of particular carbon source(s), and to simultaneously group both carbon sources and strains in a two-way joining analysis. In most cases, the cluster-joining analysis was made with Euclidian distance and complete linkage as the amalgamation rule, i.e., distances between clusters were determined by the greatest distance between any two objects in the different clusters.

We used a discrete counter plot, which is a graphical representation of two-way joining results, to obtain a carbon utilization map based on A₇₅₀ at 48 h. The algorithm for this analysis as implemented in STATISTICA 6.1 is limited to 50 cases, so we analyzed the data as four independent matrixes, with the contents of each matrix corresponding to the main groups of carbon sources. To combine the outcomes in a single plot, each of the four data sets was normalized by including values for maximum and minimum carbon source utilization (QM 6a mycelial production on γ-aminobutyrate and glucuronamide, respectively). In the merged carbon source utilization map, each datum point is represented as a color-coded rectangular region. The carbon source order remained intact, but the position of strains on the combined map was manually modified according to the nature of strains that most closely related strains (e.g., parental strain and mutants derived from it) were located by the side of each other. One-way or main-effect analyses of variance (ANOVAs) were used to compare the growth of selected strains on individual carbon sources. Tukey's honest significant difference (Tukey HSD) test as implemented in STATISTICA 6.1 was used for post hoc comparisons to detect the contribution of each variable to the main effect of the F test resulting from the ANOVA. The summed data matrixes also were evaluated following factor analysis and multidimensional scaling to detect additional relationships between variables.

RESULTS

Growth measurements by PM and their reproducibility.

We grew H. jecorina QM 9414 on d-glucose, d-galactose, d-lactose, glycerol, l-sorbose, d-xylose, xylitol, and l-arabitol in submerged culture in shake flasks, on agar plates, and compared the resulting growth curves with those obtained by PM. Good concordance was found in all cases between growth in the different types of cultures. This correlation held even for l-sorbose, a carbon source that may alter hyphal morphology (9) and was initially expected to be an exception to the correlation between biomass and A₇₅₀. Thus, the increase in absorbance values in PMs provides a reliable means to measure mycelial growth of H. jecorina.

The reproducibility of the PM results was tested for all 11 strains (Table 1) individually. No significant differences were detected in the main-effect ANOVA between independent experiments. The total number of measurements obtained for three parallel plates for each strain, i.e., 95 carbon sources and water, at eight different time points, were subjected to cluster analyses. The A₇₅₀ value for each time point (averaged over all carbon sources) usually grouped together in the three parallel experiments, with only very short linkage distances between values from different plates. The detected three larger clusters correspond to (i) spore germination (12, 18, and 24 h), (ii) growth (36, 42, and 48 h), and (iii) transition to idiophase and/or sporulation (66 and 72 h). In the latter case, the statistical distances between the time points were equal to the differences between the plates, thus providing an additional indicator of either sporulation or the cessation of growth on the third day of incubation. For most of the subsequent analyses we used only turbidity values between 12 and 48 h for the characterization of mycelial growth and to compare strains under investigation. These results also indicate that the data are highly reproducible.

Carbon source utilization profile of the H. jecorina ex-type strain QM 6a.

Growth of this strain on 95 carbon sources and water varied (Fig. 1). Cluster I contained the carbon sources that enabled the fastest growth and included several monosaccharides, oligosaccharides, the polyols (d-arabinitol and erythritol), and γ-aminobutyric acid. Cluster II contained the carbon sources that enabled good growth, but the turbidity increases were almost linear and reached only about half of the biomass densities of the compounds in Cluster I (Fig. 2). Cluster II also included the primary amino acids (l-alanine, l-aspartic acid, and l-glutamic acid) and some carbohydrates and polyols. Growth on Tween 80 (polyoxyethylensorbitan monooleate) and d-sorbitol was similar, suggesting that the d-sorbitol moiety in Tween 80 was used preferentially over the oleic acid component. The third cluster (III) contained many rare sugars, sugar acids, organic acids, and amino acids, all of which enabled only slow growth. Utilization of these carbon sources was still incomplete at 48 h (Fig. 2). The water control case grouped within these carbon sources, with the growth observed presumably resulting from the catabolism of the phytagel spore carrier (Phytagel is a bacterial heteropolysaccharide composed of glucuronic acid, rhamnose, and glucose) and/or the utilization of the nutrient reserves within the spores. It is thus doubtful whether the detected slow mycelial development on the other carbon sources of cluster IV is due to utilization of them or due to the same processes as growth on water. Two of the members of cluster IV—d-lactic acid methyl ester and glucuronamide—supported significantly less growth than that observed in the water control (ANOVA, post hoc Tukey HSD test, P = 0.012 and P = 0.004, respectively).

FIG. 1.

(Right side) Utilization of carbon sources by H. jecorina QM 6a (solid line) and the two early cellulase mutants QM 9123 (triangles) and QM 9414 (black circles). The order of the carbon sources is the rank of the growth on 95 carbon sources and water, based on the A₇₅₀ value at 48 h for strain QM 6a. Standard deviations are given by error bars. (Left side) Joining cluster analysis applied to carbon sources based on their profiles at 12, 18, 24, 36, 42, and 48 h. Numbers on this side are the same as those on the right side and indicate the compound.

FIG. 2.

Growth curves of carbon sources belonging to clusters I to IV in Fig. 1. A single line corresponds to one carbon source.

Carbon source utilization profiles of cellulase-overproducing mutants.

Compared to the parental strain QM 6a, the first improved cellulase producer mutant QM 9123 grew slower on most carbon sources of clusters I and II. Wild-type growth generally was restored for most carbon sources in the second generation mutant QM 9414 (Fig. 1). Growth on three carbon sources—adonitol, 2-ketogluconate, and γ-aminobutyric acid—was significantly slower for both mutants and was negatively correlated with cellulase formation (Fig. 3). Growth by the cellulase-overproducing strains on d-sorbitol and d-saccharic acid was significantly higher than for the wild-type strains. The increase in turbidity of strain QM 9123 on γ-aminobutyric acid at 72 h (Fig. 3B) was due to the onset of sporulation, which increased A₇₅₀ in a manner not proportional to the biomass increase. This effect also was observed for strain QM 6a but not for strain QM 9414.

FIG. 3.

Growth of H. jecorina QM 6a (•), QM 9123 (□), and QM 9414 (○) on carbon sources, for which statistically significant differences among them were detected. (A) Adonitol; (B) 2-keto-d-gluconic acid; (C) γ-aminobutyric acid; (D) salicin; (E) sorbitol; (F) saccharic acid. Standard deviations are given by vertical bars. Values with different letters are significantly different (ANOVA, post hoc Tukey HSD test, P < 0.05).

Phenotypic diversity of H. jecorina wild-type isolates.

There was more variation in carbon source utilization among six recently isolated wild-type strains and the ex-type strain, QM 6a, than there was between QM 6a and the cellulase-overproducing mutants derived from it (Fig. 4). Growth patterns on carbon sources from Cluster I were the most variable, with both QM 6a and the cellulase-overproducing mutants having intermediate growth rates. QM 6a grew faster on most of the carbon sources in clusters II, III, and IV, whereas all of the wild-type isolates grew significantly better than QM 6a on N-acetyl-d-glucosamine and on fumaric, quinic, succinic, and bromosuccinic acids (ANOVA, post hoc HSD test, P < 0.001). The six recently isolated wild-type strains had considerable phenetic concordance of growth on carbon sources from clusters III and IV. The phenetic differences between these six isolates was significantly smaller than the differences between QM 6a and any of them (ANOVA, post hoc HSD test, P = 0.03) but greater than the differences between the various mutants of QM 6a. Thus, the ex-type strain of H. jecorina, QM 6a, is not phenetically representative of the recently isolated wild-type strains.

FIG. 4.

Utilization of 95 carbon sources by various wild isolates of H. jecorina. The order of carbon sources corresponds to that in Fig. 1 (H. jecorina QM 6a). The strains used were G.J.S 85-236 (⧫), TUB F-1034 (▪), TUB F-430 (▴), TUB F-733 (⋄), CBS 836.91 (□), and TUB F-1066 (▵). Standard deviations are given by bars. The light gray background corresponds to the variability seen in the early cellulase mutants, taken from Fig. 1.

Growth patterns of QM 9978, a cellulase-negative mutant.

The phenotype profile of QM 9978 mutant did not differ significantly from that of its parent, QM 6a (Fig. 5). Most of the changes again occurred in the utilization of carbon sources in cluster II, where QM 9978 had decreased ability to grow on d-galactose and N-acetyl-d-glucosamine, whereas its relative ability to utilize salicin and l-alanine increased. The observations were confirmed at 48 h by ANOVA (post hoc HSD test; P < 0.05 for all listed compounds). Thus, the loss of the ability to induce cellulase formation did not alter the global carbon utilization phenotype profile of QM 9978.

FIG. 5.

Summed map of global carbon utilization profiles of the ex-type QM 6a strain of H. jecorina, cellulase-negative mutant strain QM 9978; uridine auxotrophic (pyr4) mutant TU-6; early cellulase mutant QM 9414; and two triplets of hygromycin B-resistant transformants (the number of integrated copies is given in roman numerals). The map was composed after several two-way joining cluster analyses applied to carbon sources (i) and fungal strains (ii) as two groups of variables. For the pedigrees and relationships of the strains, see Table 1. Due to the low variability in carbon sources from clusters III and IV, only carbon sources from clusters I and II are shown; the respective growth (A₇₅₀ after 48 h) is given by a corresponding color as indicated in the color scale.

Growth patterns of TU-6 (pyr4).

H. jecorina mutant TU-6 carries a mutation at pyr4, which encodes orotidine-5-phosphate-decarboxylase-encoding (11, 12). This strain frequently is used as an auxotrophic recipient strain for DNA-mediated transformation. This mutant—when supplemented with 10 mM uridine—has a slightly different growth pattern from that of QM 9414, primarily due to its improved utilization of erythritol, β-methyl-d-glucoside, d-galactose, d-ribose, glycerol, and xylitol (Fig. 5). Utilization of other carbon sources by TU-6 was not different from that of its parent strain QM 9414.

Carbon utilization patterns in transformants.

A site-directed integration/transformation system is not available for H. jecorina. Instead, DNA integration usually occurs at ectopic sites, and incidental epistatic effects associated with the transformation process or the ectopic integration event could remain undetected. Triplets of transformant strains derived from either QM 9414 and QM 9978 were evaluated (Fig. 5). When strains with different numbers of DNA insertions were compared, the number of heterologous gene copies was not correlated with alterations in carbon utilization pattern. Even single-copy integration events could result in significant changes in the carbon utilization profile. Relative to the parental strain, all three QM 9414 transformant mutants grew poorly, e.g., on maltotriose, dextrin, erythritol, d-galactose, lactose and arbutin (48 h, ANOVA, post hoc Tukey HSD test, P < 0.01). The three QM 9978 transformant strains had carbon utilization patterns that differed both from the parental strain and from each other. After 48 h of incubation, transformant CPK 1028 utilized xylitol, l-arabinose, d-ribose, and erythritol (ANOVA, post hoc Tukey HSD test, P < 0.05) and d-galactose (ANOVA, post hoc Tukey HSD test, P < 0.0001) at significantly increased rates but maltose at a significantly decreased rate (ANOVA, post hoc Tukey HSD test, P < 0.001). The two other transformants of QM 9978 (CPK 1027 and CPK 1029) did not use any of carbon sources more effectively but were impaired in their ability to utilize xylitol and glycerol (Fig. 5). Thus, transformation to hygromycin B resistance can alter carbon utilization profiles of the resulting transformants in various different and sometimes contradictory ways.

DISCUSSION

We used Phenotype MicroArrays to assess carbon source utilization profiles of Hypocrea jecorina wild-type and mutant strains. The ex-type strain H. jecorina QM 6a had the most distinctive carbon utilization profile of the seven wild-type strains. We think that these changes could be the result of “domestication” of the organism during its maintenance in the Natick laboratories over the last 60 years (22). Most of the other wild-type strains were brought into laboratory culture in the late 1990s. This “domestication” effect may be due to unintentional artificial selection during various in vitro reinoculation and/or revitalization passages.

As part of a program to increase cellulase production, the ex-type QM 6a strain of H. jecorina has been exposed to various mutagens, including radiation from a linear accelerator (18, 22). The mutants recovered from this mutagenesis had up to fivefold more cellulolytic activity. Even so, the carbon source utilization profiles of these mutants were essentially unchanged. Increased cellulase formation was inversely correlated with growth on adonitol (ribitol), 2-ketogluconate, and γ-aminobutyric acid, and directly correlated with growth on d-sorbitol and saccharic acid. We do not know whether these changes are causally related to cellulase formation. However, γ-aminobutyric acid is a sporulation-specific metabolite in Trichoderma (26), and cellulase expression is triggered during conidiation (14). There also could be a link between d-sorbitol utilization and cellulase formation since d-sorbitol can be converted to l-sorbose by an NADP-dependent ketose reductase (23), and l-sorbose can induce cellulases in H. jecorina (19).

H. jecorina QM 9978, which is derived from QM 6a, cannot produce cellulases (16, 18, 28), although the nature of this alteration is not known (31). This mutant has a carbon utilization profile similar to that of its immediate parent, QM 9123, suggesting that the mutation in this strain has occurred in a gene specifically involved in cellulase expression.

H. jecorina strain TU-6 has a mutation at pyr4. TU-6 has a carbon utilization profile that generally is similar to that of its parental strain, although it grows better on a few carbon sources. These data may be interpreted to mean that growth on these carbon sources is limited in the wild-type parent by the endogenous production of uridine. This effect is comparable to the physiological differences between genetically and nutritionally complemented S. cerevisiae leu2 and ura3 mutants (4, 21) but, to the best of our knowledge, such an effect has not been reported previously for a filamentous fungus.

There is no targeted integration system, e.g., argB in A. nidulans (30), in H. jecorina. Unless selection occurs for a specific locus, e.g., gene disruption, DNA integration usually occurs at ectopic loci. Both the location of the ectopic integration and the number of copies of the foreign DNA integrated may be problematic. The introduction of several copies of a gene controlled by a strong promoter may titrate transcription factors away from other promoters. We used PM to evaluate H. jecorina transformants that differed in the number and location of copies of plasmid pRLM_ex30, which contains the hygromycin resistance marker under the control of the pki1 (pyruvate kinase) promoter (17). Variation between different transformants was due to the site of integration and not the number of copies integrated, since transformants with different numbers of integrated copies of the foreign DNA had similar profiles.

Irrespective of the manner in which mutants were induced, e.g., exposure to UV light or radioactivity or DNA-mediated transformation, the variation in carbon source utilization profile was limited to no more than a few carbon sources. Usually these were compounds in clusters I and II, e.g., the polyols xylitol, erythritol, and d-sorbitol; the aldoses d-ribose, d-galactose and l-arabinose; N-acetyl-d-glucosamine; and the oligosaccharides maltotriose and β-methyl-glucoside that enabled good or better growth. Thus, our data demonstrate that indirect mutagenesis and transformation events alter central carbon metabolism in H. jecorina in similar manners. Desjardins et al. (6) also observed that Gibberella zeae can undergo spontaneous or transformation-induced mutations that significantly altered its virulence.

In summary, PM can give detailed and useful information on metabolic differences between strains of H. jecorina and can be used to physiologically evaluate mutants and gene interactions. The variation seen in transformant strains suggests that PM characterizations should be as much a part of a description of new transformants as are the Southern blots used to identify the location and number of the DNA copies incorporated into the genome.