Báo cáo khoa học: The b-N-acetylglucosaminidases NAG1 and NAG2 are essential for growth of Trichoderma atroviride on chitin doc

Dnag1Dnag2 strains were not able to grow on chitin and chitobi-ose, but the presence of either NAG1 or NAG2 was sufficient to restore growth on chitinous carbon sources in solid media.. a

essential for growth of Trichoderma atroviride on chitin Rube´n Lo´pez-Monde´jar*, Valentina Catalano, Christian P Kubicek and Verena Seidl

Research Area Gene Technology and Applied Biochemistry, Institute of Chemical Engineering, Vienna University of Technology, Austria

Introduction

Chitin is a natural polysaccharide consisting of

b-1,4-linked N-acetylglucosamine (GlcNAc) units, and,

although it is the second most abundant biopolymer,

relatively little is known about its turnover in marine

and soil ecosystems In the sea, chitin is found as the

main compound of the exoskeleton of crustaceans, and

on land it is an essential structural component of insects and the cell walls of filamentous fungi, where it

is covalently linked to other carbohydrates and pro-teins [1,2] Degradation of chitin biomass is achieved

Keywords

chitin degradation; chitinases;

mycoparasitism; N-acetylglucosaminidases;

Trichoderma atroviride

Correspondence

V Seidl, Research Area Gene Technology

and Applied Biochemistry, Institute of

Chemical Engineering, Vienna University of

Technology, Getreidemarkt 9 ⁄ 166-5, 1060

Vienna, Austria

Fax: +43 1 58801 17299

Tel: +43 1 58801 17227

E-mail: vseidl@mail.tuwien.ac.at

Website: http://www.vt.tuwien.ac.at/pg/

Seidl

Present addresses

*Department of Soil Water Conservation

and Organic Waste Management, Centro de

Edafologı´a y Biologı´a Aplicada del Segura

(CEBAS-CSIC), PO Box 164, 30100

Espi-nardo, Murcia, Spain

Department of Tree Science, Entomology

and Plant Pathology ‘G Scaramuzzi’, Plant

Pathology Section, Faculty of Agriculture,

University of Pisa, Via del Borghetto 80,

I-56124 Pisa, Italy

(Received 10 June 2009, revised 26 June

2009, accepted 13 July 2009)

doi:10.1111/j.1742-4658.2009.07211.x

The chitinolytic enzyme machinery of fungi consists of chitinases and b-N-acetylglucosaminidases These enzymes are important during the fungal life cycle for degradation of exogenous chitin, which is the second most abun-dant biopolymer, as well as fungal cell-wall remodelling In addition, involvement of chitinolytic enzymes in the lysis of the host cell wall in mycoparasitic Trichoderma spp has been reported In view of the fact that fungi have on average 15–20 chitinases, but only two b-N-acetylglucosami-nidases, the question arises how important the latter enzymes actually are for various aspects of chitin degradation In this study, the role of two b-N-acetylglucosaminidases, NAG1 and NAG2, was analysed in the myco-parasitic fungus Trichoderma atroviride No b-N-acetylglucosaminidase activity was detected in T atroviride Dnag1Dnag2 strains, suggesting that NAG1 and NAG2 are the only enzymes in T atroviride that possess this activity Dnag1Dnag2 strains were not able to grow on chitin and chitobi-ose, but the presence of either NAG1 or NAG2 was sufficient to restore growth on chitinous carbon sources in solid media Our results demon-strated that T atroviride cannot metabolize chitobiose but only the mono-mer N-acetylglucosamine, and that N-acetylglucosaminidases are therefore essential for the use of chitin as a nutrient source NAG1 is predominantly secreted into the medium, whereas NAG2 mainly remains attached to the cell wall No physiological changes or reduction of the mycoparasitic potential of T atroviride was detected in the double knockout strains, sug-gesting that the use of chitin as carbon source is only of minor importance for these processes

Abbreviations

GH, glycoside hydrolase; GlcNAc, N-acetylglucosamine; NAGase, b-N-acetylglucosaminidase; PDA, potato dextrose agar.

by the concerted action of chitinases and

b-N-acetyl-glucosaminidases (NAGases; EC 3.2.1.52) NAGases

belong to glycoside hydrolase (GH) family 20 in the

CAZy classification (http://www.cazy.org), and, by

def-inition, catalyse the hydrolytic release of terminal,

non-reducing GlcNAc residues, but their highest

sub-strate affinity is for the dimer chitobiose (GlcNAc)2,

which they convert into two GlcNAc monomers [3]

The genomes of ascomycetous filamentous fungi

contain on average 15–20 genes encoding chitinases,

but only two or three genes encoding GH family 20

proteins Potential functions of chitin-degrading

enzymes in fungi include use of exogenous chitin as a

nutrient source and cell-wall remodelling during the

fungal life cycle [4]

Some species of the fungal genus Hypocrea⁄

Tricho-derma, such as T atroviride (teleomorph

Hypo-crea atroviridis), T harzianum, T virens (H virens)

and T asperellum, are mycoparasites, i.e they invade

and destroy fungal cells and feed on the contents of

dead cells Chitinases and NAGases have been

repeat-edly implicated in cell-wall hydrolysis during

myco-parasitic attack (for reviews, see [5,6]) Two NAGases

have been cloned from several Trichoderma spp., and

it was shown that they are active as dimers and that

their gene expression can be induced by chitinous

carbon sources such as GlcNAc,

chito-oligosaccha-rides, colloidal chitin and fungal cell walls [7–13]

Further, enhanced NAGase activities were detected

on non-chitinous carbon sources such as a-glucans

and oligosaccharides containing galactose, and

tran-scriptional upregulation of nag1 and nag2 under the

respective growth conditions was shown [11] In the

same study, basal transcript levels of nag1 and nag2

and corresponding NAGase activities were detected

under non-inducing growth conditions, possibly

sug-gesting a role during fungal cell-wall remodelling

However, while transcriptional regulation of genes

encoding NAGases has been studied in detail,

rela-tively little is known about their functions and

impor-tance in Trichoderma Aspergillus nidulans has only

one NAGase, nagA, which was shown to be strongly

induced during autolysis [14,15] In contrast, in a

T atroviride Dnag1 strain, residual NAGase activity

was found to be as high as 80%, depending on the

substrate [11], which is most likely due to the fact

that T atroviride has two NAGases The

mycopara-sitic abilities of the Dnag1 strain were similar to those

of the wild-type in plate confrontation assays with

plant pathogenic fungi [16], and no phenotypic

changes or alteration of mycoparasitim were detected

in a T asperellum exc2y knockout strain (where

exc2y is equivalent to nag2) [10]

It is obvious that there is still a severe lack of understanding of the physiological relevance of NAGases in fungi NAGase gene expression was found to be upregulated under a variety of growth conditions, but few effects were observed in single knockout strains in Trichoderma Therefore, the question arises as to how important NAGases actu-ally are for various chitinolyitc processes Are they involved in cell-wall remodelling, attack and defence mechanisms (e.g mycoparasitism) and⁄ or are they solely important for chitin sequestration, independent

of the chitinous carbon source? Chitin is the second most abundant biopolymer on earth, but how fungi handle its degradation is still not understood, espe-cially in view of the fact that they have up to 35 chitinases, but only two extracellular NAGases Are NAGases of particular importance for chitin catabo-lism in fungi or are they fully dispensable for this process?

We addressed these questions by creating Dnag1D nag2 strains in T atroviride Here we present data showing that NAG1 and NAG2 are the only extracel-lular NAGases in T atroviride, as Dnag1Dnag2 strains exhibit no residual extracellular NAGase activity, and that their function in cleaving the dimer chitobiose in GlcNAc monomers is essential for the use of chitin

as a nutrient source in T atroviride This is the first time that the ability of a fungus to catabolize chitin has been abolished However, the mycoparasitic potential was not altered in Dnag1Dnag2 strains, sug-gesting that use of chitin as a carbon or nitrogen source is not of major importance for the mycopara-sitic process

Results

Construction of T atroviride Dnag2 and Dnag1D nag2 strains

Two NAGases have been cloned and characterized from several Trichoderma species so far [7–10] In

T atroviride, a Dnag1 strain has been reported previ-ously [16], whereas NAG2 has only been analysed indirectly by comparison of the growth of T atrovi-ride wild-type (WT) and Dnag1 strains on various car-bon sources [11] To elucidate the role of nag2 and the combined roles of nag1 and nag2 in T atroviride, nag2 knockout and nag1 nag2 double knockout strains were generated The Dnag1 strain carries the amdS selection marker, and an hph cassette, confer-ring resistance to hygromycin B, was therefore used for generation of both Dnag2 and Dnag1Dnag2 knock-out strains

The deletion cassette (Fig S1A) was amplified by

PCR from the generated plasmid pVCNAG2 and

transformed into T atroviride WT and Dnag1 strains

(see Experimental procedures) Purified transformants

were screened for deletion of nag2 by PCR (data not

shown) Five Dnag2 strains and four Dnag1Dnag2

strains of the positively identified transformants were

subjected to Southern analysis (Fig S1B), which

confirmed that nag2 had been replaced by the hph

cassette, and showed that only a single copy of the

deletion cassette had been integrated into the genome

Carbon source utilization profiles of T atroviride

Dnag2 and Dnag1Dnag2 strains

The carbon source utilization profiles of all positively

identified Dnag2 and Dnag1Dnag2 knockout strains were

assessed using the Biolog Phenotype MicroArray

sys-tem, which has previously been established for

Tricho-dermaspp [11,17] and allows fast and reliable screening

of growth rates on 95 carbon sources Carbon source

profiling enabled us to analyse the phenotypical

vari-ability of growth among the knockout strains in order

to check for any possible defects unrelated to the nag2

gene knockout due to the transformation procedure,

and also to compare the growth profiles of the

knock-out strains with those of the WT strain Specific growth

rates of the strains were calculated from the increase in

the absorbance at 750 nm between 24 and 42 h – the

time at which active growth occurs on most carbon

sources – and are shown in Fig S2 The inter-strain

var-iability among the five Dnag2 strains and four Dnag1

Dnag2 strains that were studied was extremely low, as

can be seen from the error bars in Fig S2, representing

the standard deviation of the growth rate for the

respec-tive groups of strains

The average carbon source utilization profiles of

the Dnag2 and Dnag1Dnag2 strains were highly

simi-lar to those of the WT (Fig S2), showing that

assimi-lation of the 95 carbon sources assayed was not

altered in the Dnag2 or Dnag1Dnag2 knockout strains

The Dnag1 strain, which has already been

character-ized in detail using the Biolog system [11], also

displayed similar growth rates (data not shown)

These data indicate that NAGases are not essential

for normal growth on non-chitinous carbon sources

Dnag2 strains C2332 and A523 and Dnag1Dnag2

strains 713 and 1921 were randomly chosen and used

together with the WT and Dnag1 strain in subsequent

experiments for thorough characterization of their

phenotypes Dnag2-I and Dnag2-II are Dnag2 strains

C2332 and A523, and nag1Dnag2-I and nag1Dnag2-II

are Dnag1Dnag2 strains 713 and 1921, respectively

NAG1 and NAG2 are essential for growth on chitin and chitobiose

Having shown that T atroviride Dnag2 and Dnag1 Dnag2 strains grew normally on non-chitinous carbon sources, we next investigated the role of these enzymes

in growth on chitin Although T atroviride has more than 25 chitinases, chitin is not a good carbon source for T atroviride, even in its pre-treated colloidal form The fungus does not readily use chitin but first forms

a thin mycelium on the surface of the whole agar plate before actually starting to form biomass, which is then strongly linked to sporulation The WT, Dnag1 and Dnag2 strains formed a firm layer of biomass and spores on chitin plates, whereas the Dnag1Dnag2 strains only produced very few spots of sporulating biomass (Fig 1A) On control plates containing potato dextrose agar (PDA) all strains grew and sporulated normally (Fig 1B) These results suggest that the pres-ence of at least one of the two enzymes NAG1 and NAG2 is essential for growth on chitin by hydrolysing the dimer chitobiose (GlcNAc)2, and imply that extra-cellular conversion of the dimer into monomers is nec-essary for assimilation of this carbon source, and that only the monomer can be taken up by the fungus The small amount of biomass that Dnag1Dnag2 strains formed on chitin plates could theoretically result either from the presence of an as yet unidentified third NAGase, or be due to release of GlcNAc mono-mers resulting from the random cleavage of chito-olig-omers by chitinases To test this, we grew the strains

on chitobiose (Fig 1C) Under these conditions, growth of the WT and Dnag1 and Dnag2 strains occurred and was identical, whereas the Dnag1Dnag2 strains did not grow at all, except for a very few extre-mely thin aerial hyphae, which were also found on control medium containing no carbon source; these hyphae therefore most likely result from internal car-bohydrate reserves of the spores These findings prove that NAG1 and NAG2 are together responsible for chitobiose degradation by T atroviride, and that there are no further enzymes in T atroviride that account for this ability

As growth on plates can be misleading, e.g due to varying hyphal thickness, we also quantified the biomass formed on chitin plates The results from biomass quantifications (Fig 2A) reflected the macro-scopic observations from Fig 1B, showing a statisti-cally significant reduction of biomass formation in the Dnag1Dnag2 strains to less than 25% of the WT strain (one-way ANOVA, F(5,6) = 138.44, P < 0.01) Bio-mass formation in the Dnag1 and Dnag2 strains was similar to that in the WT (P > 0.05)

To further prove that NAG1 and NAG2 are the

only enzymes responsible for NAGase activity in T

at-roviride, we assayed their activity in the various

dele-tion strains Fig 2B shows that the activity was indeed

completely absent in the Dnag1Dnag2 strains In the

single knockout strains, a significant reduction of

NAGase levels was detected (one-way ANOVA,

F(5,6) = 71.72, P < 0.01), but the residual activity

was still approximately 60% of that of the WT strain

In addition, the finding that the sum of the NAGase

activities in the Dnag1 and Dnag2 strains totalled more

than 100% of that in the WT suggested that

expres-sion of NAG2 and NAG1, respectively, may be

enhanced in each other’s absence to compensate for

the absence of the other enzyme

NAGase activity is not essential for induction of

chitinases

Brunner et al [16] reported a reduction of chitinase

activities in the Dnag1 strain during growth on

colloi-dal chitin in shake flask cultures, and concluded that

NAG1 may be involved in formation of the inducer

for chitinase gene expression We were therefore

expecting an even more drastic reduction in the double

mutant Consequently, we measured chitinase activities

in the single and double mutants on chitin plates

(Fig 2C) The data confirmed the significant reduction (one-way ANOVA, F(5,6) = 8.20, P < 0.01) of chitin-ase formation in the Dnag1 strain However, this reduction did not occur in the Dnag2 strains, and, most importantly, not in the Dnag1Dnag2 strains either While the reason for the unique behaviour of the Dnag1 strain remains to be elucidated, we neverthe-less conclude that this observation is not connected to the reduction of NAGase activity in the Dnag1 strain because even the complete loss of NAGase activity in the double mutants did not affect chitinase formation

on chitin in these strains (see Discussion for details)

Dnag2 and Dnag1Dnag2 strains have no morphological defects

A hypothesis that was raised previously in several reviews, e.g [18,19], postulated that chitin-degrading enzymes, including NAGases, are involved in cell-wall remodelling during hyphal growth To study the poten-tial involvement of NAG1 and NAG2 in these pro-cesses, a detailed morphological characterization of the

T atroviride Dnag1, Dnag2 and Dnag1Dnag2 strains was carried out It should be noted that no NAGase activity was detected under these growth conditions, using glucose as the carbon source (data not shown) Germination of the strains was followed in liquid

A

B

C

Fig 1 Growth on chitin and chitobiose T atroviride strains (WT, Dnag1, Dnag2 and Dnag1Dnag2) were grown on solid medium (1.5% w ⁄ v agar) containing (A) minimal medium with colloidal chitin, (B) PDA as a control to show normal growth and sporulation of the knockout strains, and (C) minimal medium with chitobiose Dnag2-I and Dnag2-II are Dnag2 strains C2332 and A523, and I and

nag1Dnag2-II are Dnag1Dnag2 strains 713 and 1921, respectively.

cultures, but no differences between the knockout strains and the WT could be detected with respect to the timing and frequency of spore swelling and germi-nation, and the morphology of the germ tubes was also completely normal, indicating that the NAGases NAG1 and NAG2 are not essential for germination in

T atroviride(Fig S3A)

Hyphal morphology was investigated macroscopi-cally and microscopimacroscopi-cally on agar plates using a num-ber of carbon sources including glucose, glycerol, maltotriose, glycogen, glucosamine, GlcNAc and PDA Hyphal extension and colony diameter were measured, but no differences between the knockout strains and the WT were observed on any carbon source, confirm-ing the data from the Biolog analysis (see above) A microscopical analysis of hyphal growth and branching patterns did also not reveal any differences among the analysed strains (Fig S3B), indicating that the NAG-ases NAG1 and NAG2 are not essential for hyphal growth in T atroviride

Further, sporulation rates were measured on various carbon sources by quantification of the numbers of spores formed on agar plates, but again no significant influence of the loss of nag1 and nag2 could be detected The only exception was the carbon source GlcNAc, on which the WT and Dnag1Dnag2 strains produced a similar number of spores, while the Dnag2 strains only produced 11 ± 1% of the number of spores produced by the WT and the number of spores

in the Dnag1 strain was 466 ± 105% The results on all other carbon sources showed no differences between the WT and the single knockout strains, and, most importantly, on none of the investigated carbon sources could any changes in sporulation rates be detected in the Dnag1Dnag2 strains

Comparison of growth and chitinolytic activities

on chitin in liquid and solid media Having determined that NAG1 and NAG2 are essential for growth on chitin in plates and that no residual NAGase activity remained in the double knockout

A

B

C

Fig 2 Biomass and chitinolytic enzyme activities on chitin agar plates (A) Biomass, measured as total protein concentration, (B) NAGase activities and (C) chitinase activities of T atroviride strains (WT, Dnag1, Dnag2 and Dnag1Dnag2) Values for biomass are given per mL of protein extracts under normalized extraction conditions, and enzyme activities were normalized to the biomass and are shown per mg of biomass (total protein) Error bars show SEM val-ues of the measurements D2-I and D2-II are Dnag2 strains C2332 and A523, and DD-I and DD-II are Dnag1Dnag2 strains 713 and

1921, respectively.

strains, we were interested to assess how these results

compare to those of previous studies using submerged

shake flask cultures [12,16], which, however, do not

resemble natural growth conditions for T atroviride

On agar plates, growth on chitin was linked to

sporula-tion (see Fig 1A), and, similarly, visual inspecsporula-tion of

the shake flask cultures showed that the biomass

formed by the WT strain was already green due to

spor-ulation after 48 h, whereas the biomass of the Dnag2

strains was only light green and no sporulation was

observed in the Dnag1 and Dnag1Dnag2 strains To

quantify these observations, samples were taken after

30, 48 and 72 h, and the total protein concentration,

corresponding to biomass formation, was measured for

mycelial and conidial biomass after extraction with

NaOH (Fig 3A) Our results show that growth on

chi-tin in shake flask cultures was similar to that of the WT

in the Dnag2 strains and slightly reduced in the Dnag1

strain, and that almost no growth of the Dnag1Dnag2

strains occurred at all Extracellular NAGase activities

of the Dnag2 strains, normalized to the amount of

bio-mass, were similar to WT levels, whereas those of the

Dnag1 strain were reduced to below 2% and in the Dnag1Dnag2 strains no NAGase activity could be detected at all (Fig 3B) This showed that NAG1 and NAG2 are also the only two enzymes responsible for extracellular NAGase activity in T atroviride in shake flask cultivations Similar results were obtained when cell-wall-bound NAGase activities were also taken into account, except that activities in the Dnag1 strain were approximately 28% and 35% of the WT levels at 48 and 72 h, respectively (Fig 3C) This suggests that NAG2 remains attached to the fungal cell wall although the protein sequence does not contain any membrane-anchoring signals Chitinase activities, normalized to the biomass, paralleled NAGase activities, with a strong reduction of chitinase activities in the Dnag1 and Dnag1Dnag2 strains (Fig 3D) In summary, these results revealed differences in the kinetics of NAG1 and NAG2 formation between submerged and solid-surface cultivations, which also seemed to affect chitinase for-mation in the Dnag1 strain, but confirmed our finding that the presence of either NAG1 or NAG2 is essential for growth on chitin

Fig 3 Biomass and chitinolytic enzyme activities upon growth on chitin in shake flask cultures Values for biomass are given per mL of trea-ted culture extract Enzyme activities were normalized to the biomass and are shown per mg of biomass (total protein) (A) Biomass, mea-sured as total protein concentration, (B) extracellular NAGase activities, (C) total (extracellular and cell-wall-bound) NAGase activities, and (D) extracellular chitinase activities Mean values from one representative experiment are shown Filled diamonds, WT strain; filled triangles, Dnag1 strain; grey circles, Dnag1 strain A523 (D2-II); grey diamonds, Dnag2 strain C2332 (D2-I); open circles, Dnag1Dnag2 strain 713 (DD-I); crosses, Dnag1Dnag2 strain 1921(DD-II).

Autolysis is not altered in Dnag1Dnag2 strains

As T atroviride Dnag1Dnag2 strains could not grow on

chitin, we reasoned that T atroviride would also no

longer be able to recycle GlcNAc from its own cell

wall, and its ability to autolyse would be altered

Autolysis was studied by growing T atroviride strains

in submerged cultivations with glucose as the carbon

source and measuring the decrease of biomass after

entering the stationary phase (i.e when glucose had

been depleted) This phase was observed after 35 h of

cultivation for all strains However, up to 90 h of

culti-vation, no differences in the decline of biomass due to

autolysis and the corresponding concentration of

extracellular proteins were detected between the strains

(data not shown) NAGase activities under these

con-ditions were consistent with the results for the

respec-tive strains shown during submerged growth on chitin,

i.e no NAGase activity was observed in the double

knockout strains and NAG1 activitiy seemed to be

predominantly extracellular, whereas the majority of

NAG2 was apparently attached to the cell walls (data

not shown) These results indicate that recycling of

GlcNAc via NAG1 and NAG2 during autolysis is not

of major importance for T atroviride

Mycoparasitism is not affected by the lack of

NAGase activity

To analyse whether the inability to use chitin would

affect the mycoparasitic activity of T atroviride, plate

confrontation assays with two plant pathogenic fungi,

the basidiomycete Rhizoctonia solani and the

ascomy-cete Botrytis cinerea, were performed The experiment

was carried out on both PDA plates and on plates

with minimal medium and nutrient limitations (various

nitrogen and glucose concentrations, see Experimental

procedures for details) However, no effect was

observed under any of these conditions (data not

shown); all tested knockout strains were as efficient as

the WT in parasitizing both host fungi This result

shows that the inability to use chitin as a carbon

source during mycoparasitism does not affect the

antagonistic potential of T atroviride This finding

does not eliminate the possibility that chitin in the cell

walls of the host fungi is hydrolysed by chitinases, but

indicates that it does not need to be metabolized

Discussion

In this study, we assessed the function of fungal GH

family 20 NAGases in fungal chitin catabolism

Rela-tively little is known about chitin turnover by fungi, and

it is especially difficult to determine the importance of chitin degradation for the mycoparastic process The number of chitinolytic enzymes in mycoparasitic Trichoderma spp is much higher than in other fungal genera [20], but the number (two) and amino acid sequences of NAGases are much more conserved when compared to other fungal genomes Until now, apart from their transcriptional regulation, nothing was known about the roles and physiological relevance of NAGases in fungi We therefore generated Dnag1Dnag2 double knockout strains in T atroviride in order to study the role of NAGases in chitin degradation

No extracellular NAGase activity was detected in Dnag1Dnag2 strains under any of the tested growth con-ditions, which indicates that NAG1 and NAG2 are indeed the only extracellular NAGases in T atroviride under the tested conditions Our data show that the presence of either of these enzymes is essential and sufficient for degradation of chitobiose and growth on chitin Analysis of the T atroviride genome database (http://genome.jgi-psf.org/Triat1/Triat1.home.html) re-vealed that, in addition to NAG1 (protein ID 136120) and NAG2 (protein ID 41039), the genome contains a third ORF encoding a GH family 20 protein (ID 33962); however, this is highly dissimilar to NAGases [NAG1 and NAG2 have a sequence similarity of 70% positives (e 0.0), but compared to the third GH 20 protein the similarity is only 37% positives (e-04 to e-08)], strongly suggesting a different substrate spectrum for this enzyme Interestingly, while most fungi possess three NAGases, the A nidulans genome also contains only two GH family 20 proteins, one of which is highly simi-lar to NAGases and the other to T atroviride protein

33962 The complete absence of NAGase activity under all tested growth conditions reported in this study – including enzyme assays of mycelial extracts from agar plates that would also reveal any intracellular NAGase activities – suggests that NAG1 and NAG2 are the only two enzymes that possess this activity in T atroviride In accordance with these findings, it is interesting to note that previous studies showed that nag1 gene expression

is induced by a variety of carbon sources and other stimuli, and is in our experience one of the strongest inducible genes in T atroviride Therefore, NAGases constitute a ‘genomic bottleneck’ for chitin catabolism

in fungi with respect to chitin degradation, as chitin can-not be used as a nutrient source if the essential NAGase activity, dependent on only two enzymes, is absent, despite the presence of approximately 30 chitinases However, the large variability of chitinases, but limited arsenal of NAGases, also suggests that chitinases might have many additional functions (defence mechanisms, enabling accessibility to other substrates, e.g during

mycoparasitism, loosening up of the cell wall during the

fungal life cycle, etc.), for which complete degradation

of chitin into GlcNAc monomers is not important

Of course, we cannot rule out the possibility that

NAGase activities will be detected in subgroups of other

GH families, which could be membrane-bound proteins

and hence be involved in fungal cell-wall remodelling,

but it should be noted that the third GH family 20

pro-tein in T atroviride does not contain a

membrane-anchoring signal GlcNAc recycling has been implicated

in cell-wall formation during the fungal life cycle,

including germination, hyphal growth, fusion and

spor-ulation [18,19,21], but we found no alteration of the

phenotype in Dnag1Dnag2 strains, and these two

enzymes are missing to form GlcNAc from chitobiose

during chitin degradation This shows that neither

NAG1 or NAG2 are involved in these processes in

T atroviride As mentioned above, it is possible that

chitinases, but not NAGases, are sufficient for loosening

of the chitin structure in the cell wall during formation

of hyphal branches and fusions With respect to

sporula-tion, an effect could only be detected on GlcNAc as

car-bon source Although all tested strains showed similar

growth rates on GlcNAc plates, the number of spores

produced was altered in the single knockout strains, but

not in the double knockout strains This suggests the

existence of regulatory mechanisms between GlcNAc

metabolism, NAGases and sporulation during growth

on this carbon source Nevertheless, we conclude from

the finding that there was no difference between the WT

and Dnag1Dnag2 strains on any of the tested carbon

sources with regard to spore formation that NAG1 and

NAG2 are not directly important for this process

We also found no differences during autolysis

between the WT and the Dnag1Dnag2 strains This can

be explained by the fact that the cell walls of hyphal

fragments that undergo autolysis are permeabilized by

chitinases and other hydrolytic enzymes, e.g

glucanas-es The intracellular components, such as

mono-saccharides and proteins, are of higher nutritional value

for the hyphal fragments than the chitinous cell wall and

therefore the recycling of GIcNAc is apparently not of

major nutritional importance during this process It

should be noted that we assayed autolysis with glucose

as carbon source, on which all strains showed the same

growth rate, whereas the previous finding that the Dnag1

strain showed delayed autolysis [16] was based on an

experiment with colloidal chitin as carbon source

How-ever, we do not consider this to be a suitable carbon

source for this experiment, because, as can be seen in

Fig 3A, the Dnag1 strain exhibits slower growth in

sub-merged cultures with chitin than the WT strain does,

and therefore the suspected delay of autolysis reported

by Brunner et al may have been due to the slower growth rate on this carbon source

Despite our findings that extracellular NAGases are not important for any of the studied morphogenetic aspects in T atroviride, our results clearly showed that they are essential for growth on chitin Dnag1Dnag2 strains showed significantly reduced biomass formation upon growth on colloidal chitin, and did not grow at all

on chitobiose This demonstrates that T atroviride Dnag1Dnag2 strains cannot hydrolyse chitobiose, and shows that, even though this fungus has a large array of chitinases, the last step of cleaving the dimer into two monomers is performed by only two enzymes: NAG1 and NAG2 Further, it can be concluded from our results that T atroviride cannot take up the dimer and use it as a carbon source, as has been reported for bacte-ria [22], but depends on extracellular cleavage of chitobi-ose into the monomer GlcNAc, which can then be taken

up by the hyphae and catabolized The small amount of residual biomass that was formed upon growth of the double knockout strain on chitin was probably due to small amounts of GlcNAc that were released by those chitinases, which can randomly cleave chito-oligomers, occasionally leading to the formation of monomers

A comparison between extracellular and total (extra-cellular and cell wall-bound) NAGase activities in liquid medium showed that NAG2 – the only NAGase

in the Dnag1 strain – was predominantly cell-wall bound, whereas large amounts of NAG1 were found

to be released from the cell wall in Dnag2 strains The fact that the remaining NAGase in Dnag1 strains was cell-wall-bound in liquid media possibly limited its access to the substrate, and this limitation of GlcNAc availability in turn resulted in less biomass formation

In Dnag2 strains, on the other hand, the large amounts

of extracellular NAG1 were sufficient to enable a growth rate in submerged cultures similar to that of

WT It is important to consider, however, that sub-merged cultures are not a natural growth medium for

T atroviride, and whether the NAGases were cell-wall-bound or extracellular did not influence growth under more natural conditions on agar plates, and therefore all single knockout strains reached WT growth levels Another interesting finding was that the sum of the NAGase activities measured in single knockout strains exceeded that of the WT This indicates that NAG1 and NAG2 can compensate for each other Such findings are reminiscent of similar data for knockouts of the cello-biohydrolases CBHI and CBHII in Trichoderma reesei, for which a Dcbh1 strain showed increased cbh2 tran-script levels in comparison to the parental strain [23] Chitinase formation on chitin plates was reduced in the Dnag1 strain but not in the Dnag2 and Dnag1Dnag2

strains The decrease in chitinase activities in the Dnag1

strain therefore cannot be directly related to the absence

of NAG1, because it was not observed in the double

knockout strains, of which the Dnag1 strain is the

pro-genitor A more likely explanation is that this effect is

caused by NAG2, possibly also due to its increased

expression in the Dnag1 strain in comparison to the WT

Elevated NAG2 levels could affect the concentration of

the chitinase inducer formed from chitin, e.g by

hydro-lysis or by transglycosylation This will be an interesting

topic for further studies In shake flask cultures, growth

on chitin is generally slow and inefficient As can be seen

by the delayed onset of chitinase formation, measured

activities at 30 h were extremely low for all strains

However, although the WT and single knockout strains

showed an increase in biomass at later time points,

bio-mass in the double knockout strains stayed constant or

even decreased slightly This suggests that due the

lim-ited contact time of chitinases with the substrate in

shake flask cultivations, and possibly also the altered

expression profile of various chitinases, chitinases do

not release enough GlcNAc from the random cleavage

of chito-oligomers to enable residual biomass

forma-tion, as hypothesized for growth on chitin agar plates

Therefore, we conclude that the small amount of

myce-lial biomass that is formed in the first 30 h of the

cultiva-tion – probably from the 0.05% w⁄ v of peptone that is

added to liquid cultures to ensure efficient and

homoge-nous germination – is most likely dead at later time

points, which explains why no chitinase activities were

found in the double knockout strains in the shake flask

experiment The sensitivity of the enzymatic

measure-ment was not the limiting factor, because biomass in the

other strains was also relatively low (Fig 3A), the

attenuance of the enzymatic assays were in a good

sensi-tivity range of the method Chitin is an insoluble carbon

source, and to avoid effects due to limited substrate

accessibility, we conclude from the comparison of solid

and liquid cultures that growth of T atroviride on chitin

should be preferably carried out in solid substrate or

stationary cultivations

The role of chitinolytic enzymes in the mycoparasitic

process has received a lot of attention and has been

the subject of several studies in Trichoderma spp [5]

Our findings imply that NAGases are fully dispensable

for this process These results do not rule out the

pos-sibility that chitinases are important for attack of the

host, but clearly show that the use of chitinous cell

walls from the host as a carbon source is not relevant

for the antagonistic potential of T atroviride Our

find-ings suggest that in soil or on decaying wood – the

two natural habitats of T atroviride – the

mycoparasit-ic lifestyle probably involves successful competition for

nutrients and living space with other fungi rather than sequestration of chitin as a nutrient source

Analysis of chitin metabolism in fungi is compli-cated due to the large number of enzymes that are involved In this study, we elucidated the final extra-cellular steps of this process, and found that, in

T atroviride, NAG1 and NAG2 are the only enzymes responsible for the final step in chitin degradation The availability of these mutants will enable us to perform further studies on the use of chitinous carbon sources and chitinase expression in T atroviride, which will be the next steps towards understanding this versatile aspect of fungal metabolism

Experimental procedures

Strains

T atroviride P1 (ATCC 74058), referred to as wild-type (WT), and the amdS+ nag1 disruption strain T atroviride P1ND1 [16] were maintained on potato dextrose agar (Lab

M Limited, Bury, UK), and stock cultures were kept at )80 C Escherichia coli strain JM109 (Promega, Madison,

WI, USA) was used for plasmid propagation

Fungal cultivation conditions

The growth of fungal transformants on 95 carbon sources was assessed using Biolog phenotype microarrays (Biolog, Hayward, CA, USA) according to the protocol recently developed for Trichoderma spp [17] For growth assays on agar plates, minimal medium [24] with 1.5% w⁄ v agar and supplemented with 1% w⁄ v of the various carbon sources was used The carbon sources included mono- and disaccha-rides, which were purchased from Sigma (St Louis, MO, USA), and colloidal chitin, which was prepared as described previously [25] Agar plates were incubated at 25C under a

12 h light⁄ dark diurnal cycle Chitobiose growth assays were performed in six- well plates containing 650 lL minimal medium + 1.5% agar and 0.5% carbon source per well due

to the high costs of the substrate Plate confrontation assays with Rhizoctonia solani and Botrytis cinerea were performed

as described previously [26] on PDA and also on agar plates with minimal medium salt composition and (a) glucose limitation (0.2% w⁄ v), (b) nitrogen limitation (0.14 gÆL)1 ammonium sulfate), or (c) glucose and nitrogen limitation Shake flask cultivations were prepared in minimal medium containing 0.05% peptone to ensure efficient germination and 1% w⁄ v of either glucose or colloidal chitin Cultures were grown in rotary incubators (Multitron 2 shaking incu-bator, Infors, Bottmingen, Switzerland) at 28C and

220 rpm, and kept in constant light to enable sporulation, which is linked to growth on this carbon source on agar plates (this study) and was also observed previously to occur

in shake flask cultures (V Seidl, unpublished results) All

experiments were performed at least in duplicate

Determination of fungal growth, biomass

production and sporulation

The increase in colony diameter on agar plates was measured

daily To measure biomass from agar plates containing

col-loidal chitin as carbon source, agar pieces of equal size were

cut out, ground to a fine powder in liquid nitrogen and

suspended in 1 mL of buffer (100 mm Tris, pH 8.0, 1 mm

EDTA) The suspension was kept on ice and sonicated five

times for 10 s each, and centrifuged for 10 min at 13 000 g,

4C The supernatant was subsequently used to measure

total protein concentration, corresponding to the biomass,

and also NAGase and chitinase enzyme activities (see below)

The protein content was determined using the Bradford

pro-tein assay (Bio-Rad, Hercules, CA, USA) with BSA as the

standard Sporulation rates on agar plates were determined

by quantitatively harvesting spores from an agar plate using

a 0.9% NaCl + 0.05% Tween solution and counting the

spores using a haemocytometer

For submerged cultures containing soluble carbon

sources, mycelial dry weight was recorded by withdrawing

40 mL aliquots from the cultures, suction filtration through

a glass wool filter, followed by extensive washing with tap

water, and drying at 80C to constant weight For

sub-merged cultures containing colloidal chitin, the biomass

was determined by taking 1 mL samples and lysing them

by addition of 0.2 mL 0.1 m NaOH for 3 h at 30C The

samples were then centrifuged for 10 min at 13 000 g and

the supernatant was used to measure the total protein

con-centration, corresponding to the biomass, by the Bradford

protein assay using BSA as the standard All extractions

and measurements were performed at least in duplicate

Enzyme assays

Samples from shake flask cultures were centrifuged for

10 min at 13 000 g and 4C, and the supernatants were used

for extracellular enzyme activity measurements Total

(extra-cellular and cell-wall-bound) enzyme activities from shake

flask cultures were measured using samples containing

myc-elia and from agar plates using protein extracts as described

above NAGase activities were measured using a

modifica-tion of the method described by Yagi et al [27], which is

based on the release of p-nitrophenol from the respective aryl

chitosides Samples of between 5 and 100 lL were added to

0.5 mL of a solution containing 50 mm potassium phosphate

buffer, pH 6.7, and 300 lgÆmL)14-nitrophenyl

N-acetyl-b-d-glucosaminide, and the volume made up to 600 lL with

buffer Enzyme assays were incubated at 30C with gentle

agitation, reactions were terminated after 15 min by addition

of 0.4 mL 0.4 m Na2CO3, and absorbance was measured at

405 nm Control measurements of enzyme activities were

performed by omitting the substrate from the phosphate buf-fer Chitinase activities were determined using the same method but with 4-nitrophenyl b-d-N,N¢,N¢¢-triacetylchitotri-ose as substrate Enzymatic activities were calculated based

on the release of 4-nitrophenol using a molar extinction coefficient of 18.5 mmol)1Æcm)1

Statistical evaluation

Statistical analysis of the results, as specified in the various sections, was performed using graphpad instat software version 8.0 (San Diego, CA, USA)

Microscopic analysis

For microscopic analysis, an inverted T300 microscope (Ni-kon, Tokyo, Japan), equipped with differential interference contrast optics, was used, and images were captured using

a DXM1200F digital camera (Nikon) and digitally pro-cessed using photoshop CS3 (Adobe, San Jose, CA, USA) Germination was observed by placing 50 lL samples on large cover slips, and hyphae were imaged directly on agar pieces that were cut out from plates using the inverted agar method described previously [28]

Plasmid construction

The UniProt accession number of T atroviride NAG1 is P87258 The T atroviride nag2 gene was identified in the

T atroviride genome database (http://genome.jgi-psf.org/ Triat1/Triat1.home.html) using a previously cloned frag-ment of nag2 [11], GenBank⁄ EMBL ⁄ DDBJ accession num-ber DQ364461 (UniProt Q0ZLH7), for a BLAST search The query yielded a single specific hit (protein ID 41039) For the nag2 deletion vector, 1.5 kb of the up- and down-stream non-coding regions of T atroviride nag2 were ampli-fied from T atroviride P1 genomic DNA using primer pairs

A⁄ B and C ⁄ D, respectively (Table 1), using the GoTaq system (Promega), with 200 nm of each primer in the PCR reactions and reaction conditions according to the manu-facturer’s instructions The hph cassette from pRLMEX30 [29] was cut out using XhoI⁄ HindIII and ligated into an XhoI⁄ HindIII-digested pBluescript SK(+) vector (Strata-gene, La Jolla, CA), resulting in vector pBS31 The PCR fragment of the nag2 upstream region was ligated into the pGEM-T Easy vector (Promega), cut out again using the NotI restriction sites, and ligated into NotI-digested pBS31 The resulting plasmid was digested with ApaI, and the amplified nag2 downstream region was digested correspond-ingly and ligated, resulting in the nag2 knockout vector pVCNAG2 The correct orientation of the fragments was checked using several control restriction digests The 5.8 kb nag2 deletion cassette was amplified using primers E and F (Table 1) using the Long Template Expand PCR System