Conserved and essential transcription factors forcellulase gene expression in ascomycete fungiSamuel T. Coradetti, James P. Craig, Yi Xiong, Teresa Shock1, Chaoguang Tian2, and N. Louise Glass3

Department of Plant and Microbial Biology, University of California, Berkeley, CA 94720

Edited by Jay C. Dunlap, Dartmouth Medical School, Hanover, NH, and approved March 23, 2012 (received for review January 15, 2012)

Rational engineering of filamentous fungi for improved cellulaseproduction is hampered by our incomplete knowledge of transcrip-tional regulatory networks. We therefore used the model filamen-tous fungus Neurospora crassa to search for uncharacterizedtranscription factors associated with cellulose deconstruction. Ascreen of a N. crassa transcription factor deletion collection identi-fied two uncharacterized zinc binuclear cluster transcription factors(clr-1 and clr-2) that were required for growth and enzymatic ac-tivity on cellulose, but were not required for growth or hemicellu-lase activity on xylan. Transcriptional profiling with next-generation sequencing methods refined our understanding of theN. crassa transcriptional response to cellulose and demonstratedthat clr-1 and clr-2 were required for the bulk of that response,including induction of all major cellulase and some major hemicel-lulase genes. Functional CLR-1 was necessary for expression of clr-2and efficient cellobiose utilization. Phylogenetic analyses showedthat CLR-1 and CLR-2 are conserved in the genomes of most fila-mentous ascomycete fungi capable of degrading cellulose. In As-pergillus nidulans, a strain carrying a deletion of the clr-2 homolog(clrB) failed to induce cellulase gene expression and lacked cellulo-lytic activity on Avicel. Further manipulation of this control systemin industrial production strains may significantly improve yields ofcellulases for cellulosic biofuel production.

Hydrolysis of cellulose to glucose is a major bottleneck incellulosic biofuel production processes (1). In nature, fungal

cellulases act within a complex repertoire of enzymes to syner-gistically deconstruct plant cell wall polysaccharides (1, 2) (Fig.S1) and industrial production of fungal cellulases is expected toplay a major role in next-generation biofuel production (1). Acombination of process engineering and random mutagenesis offungal species has greatly improved hydrolytic enzyme productionfrom industrial strains (3). However, rational engineering ofregulatory networks will likely be necessary to further increaseproductivity and specificity of enzyme secretion.Fungal cellulase gene expression and secretion are tightly con-

trolled at the transcriptional level (4–6). The most extensive workon cellulase gene transcription in filamentous fungi has been donein Aspergillus niger and Hypocrea jecorina (Trichoderma reesei) (3,7). Several small metabolites have been identified as solubleinducers of expression of cellulase genes in H. jecorina, includinglactose (8), cellobiose (9), and sophorose (10), and xylose inducesexpression of some cellulase genes in A. niger (5). These inducersdrive expression of overlapping but distinct assemblages of cellu-lase, hemicellulase, and β-glucosidase genes.In Aspergilli and H. jecorina, expression of cellulase and hemi-

cellulase genes requires the conserved fungal transcription factorXlnR/XYR1 (5, 7, 11). Although xlnR/xyr1 homologs are importantfor hemicellulase induction in Fusarium species and in Neurosporacrassa, they do not play a significant role in cellulase gene expres-sion (12–14). Other genus-specific transcriptional activators (suchas ACE2 in H. jecorina) appear to have some specificity for par-ticular inducers, but deletion strains for these transcription factorsstill induce significant cellulase secretion (15). These findingssuggest that a complex network of transcriptional activators is re-quired for cellulase and hemicellulase gene induction and secretionin filamentous fungi.

N. crassa colonizes freshly burned plant material (16, 17). N.crassa shows robust growth on cellulosic material, which is en-abled by a diverse hydrolytic secretome (6, 18). An initial study onN. crassa’s utilization of cellulose yielded promising candidategenes that were subsequently used in engineering strategies forimproved cellulosic biofuel production in Saccharomyces cer-evisiae (19) and contributed insights into a role for polysaccharidemonooxygenases in cellulose deconstruction (20, 21).Transcriptional studies of N. crassa identified an uncharac-

terized predicted transcription factor that was induced on cel-lulosic substrates (6). That discovery prompted a broader searchfor transcription factors necessary for degradation and utilizationof cellulose by assessing a near-full genome deletion strain setin N. crassa (22). Using these tools, we identified two tran-scription factors (clr-1 and clr-2) that are required for degra-dation of cellulose and explored the effects of clr-1 and clr-2deletion mutants on the transcriptional response to cellulose.Homologs of clr-1 and clr-2 are present in the genomes of a widevariety of filamentous ascomycete species capable of degradingplant cell-wall material. We show that in the distantly relatedfungus, Aspergillus nidulans, a homolog of clr-2 (clrB) was alsorequired for cellulolytic activity, but a homolog of clr-1 (clrA)modulated cellulase gene expression. These data indicate thathomologs of CLR-1 and CLR-2 play an important role in plantcell-wall degradation in filamentous ascomycete fungi.

ResultsEssential Regulators for Cellulose Degradation in N. crassa. Toidentify transcription factors required for cellulose degradation inan unbiased fashion, we screened theN. crassa transcription factordeletion set (22) for mutants with deficient growth on cellulose(Avicel) (SI Materials and Methods). This screen identified ap-proximately two-dozen poorly characterized transcription factorswith a spectrum of effects for growth and activity on Avicel. Al-though strains containing deletions of genes involved in plant cell-wall degradation in H. jecorina and Aspergilli were present in thecollection, including homologs to xlnR (5) (NCU06971), ace1 (23)(NCU09333), and hap2 (24) (NCU03033), in N. crassa thesemutants showed no significant growth defects on Avicel. However,other deletion strains for transcription factors with a known in-fluence on cellulase production were identified, including nit-2(NCU09068, a homolog to the nitrogen regulator areA) (25), pacC

Author contributions: S.T.C., J.P.C., Y.X., and N.L.G. designed research; S.T.C., J.P.C., Y.X.,T.S., and C.T. performed research; S.T.C., J.P.C., Y.X., T.S., C.T., and N.L.G. analyzed data;and S.T.C., J.P.C., and N.L.G. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Data deposition: The profiling data reported in this paper have been deposited in theGene Expression Omnibus (GEO) database, www.ncbi.nlm.nih.gov/geo (accession no.GSE35227) and RNA-Seq data are listed in Dataset S1.1Present Address: Codexis Inc., Redwood City, CA 94063.2Present Address: Tianjin Institute of Industrial Biotechnology, Chinese Academy of Sci-ences, Tianjin 300308, China.

3To whom correspondence should be addressed. E-mail: Lglass@berkeley.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1200785109/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1200785109 PNAS | May 8, 2012 | vol. 109 | no. 19 | 7397–7402

GEN

ETICS

Dow

nloa

ded

by g

uest

on

June

22,

202

0

(NCU00090, pH sensing) (26), and cre-1 (NCU08807, carboncatabolite repression) (27). The strain that exhibited the mostsevere growth defect on Avicel, but not on either sucrose or xylan,contained a deletion of NCU07705, which was previously identi-fied as induced under cellulolytic conditions (6) (Fig. S2). A sec-ond mutant, which contained a deletion of NCU08042, alsoexhibited a particularly severe growth defect on Avicel (Fig. S2).NCU07705 andNCU08042were provisionally named clr-1 and clr-2, respectively, for cellulose degradation regulator 1 and 2.The clr-1 and clr-2 genes encode proteins that belong to the

fungal-specific zinc binuclear cluster superfamily (Fig. S2). Thislarge and diverse family of transcriptional regulators (28) includesmany previously described regulators of alternative carbon me-tabolism, including the S. cerevisiae transcriptional activator Gal4,and the H. jecorina cellulosic regulators XYR1 and ACE2.Members of this family typically have two conserved domains,a zinc binuclear cluster coordinating DNA binding, and a con-served middle homology domain, which is often associated withregulation of transcription factor activity (28). Both CLR-1 andCLR-2 have the canonical domain architecture for zinc binuclearcluster transcription factors, although the middle homology do-main of CLR-2 is truncated (Fig. S2).Deletion strains for clr-1 and clr-2 (Δclr-1 and Δclr-2) grown for

16 h on sucrose and then transferred to Avicel for 24–48 h showedno detectable cellulase activity when assayed for cleavage ofcarboxymethyl cellulose (CMC) (Fig. 1). The Δclr-1 and Δclr-2mutants secreted ∼5% of total protein compared with WT strainsunder cellulolytic conditions (Fig. 1 and Fig. S2) and displayedonly trace levels of xylanase activity. However, when the Δclr-1and Δclr-2 mutants were grown on sucrose for 16 h and thenswitched to xylan, they displayed near WT hemicellulase activityand secreted protein levels (Fig. 1) as well as similar bandingpatterns on SDS/PAGE gels (Fig. S2). The cellulolytic pheno-type segregated with the hygromycin marker used to disrupt

NCU07705 and NCU08042 and was complemented by intro-duction of a WT copy of NCU07705 or NCU08042 at the his-3locus. Neither WT strains nor the clrmutants showed cellulolyticactivity when exposed to hemicellulosic substrates (Fig. 1A).The Δclr-1 and Δclr-2 mutants showed an identical phenotype

when grown on sucrose (WT growth), xylan (WT growth), andcellulose (no growth). However, the Δclr-1 mutant was deficientfor growth on cellobiose, which is the major hydrolytic product ofcellulose degradation, and the Δclr-2mutant grew similarly to WT(Fig. 1D). These data suggest that CLR-1 and CLR-2 may dif-ferentially regulate gene sets associated with plant cell-wall deg-radation, in particular genes encoding β-glucosidase enzymes,which are involved in hydrolysis of cellobiose to glucose (Fig. S1).

Induction of Cellulose Degrading Enzymes in WT N. crassa. To betterunderstand processes by which filamentous fungi sense and re-spond to cellulose in their environment, we used next generationRNA sequencing (RNA-Seq) to profile genome-wide mRNAabundance in WT N. crassa when exposed to Avicel. WT cultureswere grown for 16 h on sucrose and then shifted to Avicel as asole carbon source. RNA samples were taken at 30 min, 1 h, 2 h,and 4 h following the shift, and relative transcript abundanceswere measured.Within 30 min of transfer to Avicel, mRNA abundance for

cellulase genes increased 2- to 10-fold (Fig. 2A). This level oftranscription was maintained until 1 h after the shift, then cel-lulolytic genes were further induced several thousand-fold. Ex-pression patterns for hemicellulase genes exhibited a similar two-stage induction (Fig. S3). A conservative differential expressionanalysis (see Materials and Methods) between Avicel and sucrosecultures at 4 h called nearly 40% of the genome with functionalcategory enrichments (29), dominated by a general decrease in

A

C

B

D

0

100

200

WT ∆clr-1 ∆clr-2 WT ∆clr-1 ∆clr-2 Avicel Cultures Xylan Cultures

Prot

ein

(mg/

L) Total Protein

0.0

1.0

2.0

Sucrose Cellobiose Xylan CMC Dry

Wei

ght (

g/L)

WT ∆clr-1 ∆clr-2

0.0

1.0

Rel

. Act

ivity

CMCase Activity 24 hrs 48 hrs

4.0

8.0 Xylanase Activity

0.0

1.0

Rel

. Act

ivity

****** **

****

Growth

Fig. 1. Growth and activity phenotype of clr deletion strains. (A–C) Car-boxymethyl cellulose hydrolysis (CMCase) activity, xylanase activity, and totalprotein in supernatants from WT and clr mutant cultures grown on sucrosefor 16 h and subsequently transferred to Avicel or xylan. (D) Mycelia dryweights from carbon-limited cultures on soluble carbon sources (0.5% wt/volexcept for CMC, which was 2%). *P < 0.05, **P < 0.01, ***P < 0.001.

A

B All Genes (4 hours)

10

100

1000

10000

100000

1h 4h 1h 4h 30m 1h 2h 4h

Sucrose No Carbon Avicel

FPK

M

7340 GH7 9680 GH6 7190 GH6 2240 GH612916 GH618760 GH611050 GH615057 GH7 0762 GH5 0206 CDH

333 Differential Expressed Genes

Log 2 A

vice

l/No

Car

bon

Maximum FPKM

Cellulase Genes

0

5

10

-5

1 10 100 1000 10000 100000

15

-15

Fig. 2. Transcriptional induction by cellulosic substrates in N. crassa asmeasured by RNA-Seq. (A) Message abundance in FPKM for 10 cellulasegenes after transfer of a 16-h sucrose-grown culture to either sucrose, nocarbon source, or Avicel. Error bars show ± 1 SD for conditions with triplicatemeasurements. Conditions without error bars are from single libraries.Genes listed by NCU number (39) and CAZy class (43). (B) Comparison of fullgenome-expression profiles after transfer of a sucrose-grown culture toAvicel vs. no added carbon source. Log2 ratio of Avicel/no carbon FPKM vs.maximum FPKM in either condition. Genes detectable in one condition butnot the other are plotted along the y axis (nonexpressed condition set to 1FPKM). Genes exhibiting differential expression [identified by DESEq (42)and EdgeR (44)] are plotted in red.

7398 | www.pnas.org/cgi/doi/10.1073/pnas.1200785109 Coradetti et al.

Dow

nloa

ded

by g

uest

on

June

22,

202

0

transcription, translation, and mitochondrial biogenesis (Fig. S4).As an alternative reference condition, we shifted cultures tominimal media with no added carbon source. The “no carbon”cultures showed the same initial increase in cellulase gene tran-scription as observed on Avicel, but without the later, moresubstantial induction (Fig. 2B). Only 333 genes were differentiallyexpressed between Avicel and no carbon cultures at 4 h (Fig. 2B).This gene set was enriched for induction of the C-compoundand carbon metabolism, extracellular metabolism, and poly-saccharide-binding functional categories (Fig. S4). Given thesimilarity of transcriptional profiles between Avicel and no car-bon conditions, especially at 1 h posttransfer, we interpret thefirst stage of induction as the relief of carbon catabolite repres-sion (de-repressed phase), followed by a true induction phase inresponse to a cellulosic signal (inductive phase).Of the 333 genes differentially expressed between Avicel and

no carbon conditions, 212 had greater expression on Avicel thanunder either no carbon or sucrose conditions (Fig. S5). We referto this gene set as a conservative “Avicel regulon.” The Avicelregulon included 17 of 21 predicted cellulase and 11 of 19 pre-dicted hemicellulase genes (6), 22 genes encoding enzymes withpredicted activity on polysaccharides, and 56 additional genesencoding proteins predicted to enter the secretory pathway(Dataset S1). Also induced were genes for 13 transporters [in-cluding two recently identified cellodextrin transporters (19)],three components of the endoplasmic reticulum (ER) trans-location complex, two ER disulfide isomerases, two ER chap-erones (HSP70 and calreticulin), seven cytosolic enzymes withpredicted activity on disaccharides, a homolog to Aspergilli andH. jecorina xlnR/xyr1 (5, 30), and both clr-1 and clr-2. Also in-cluded in the Avicel regulon were 92 proteins annotated asconserved hypothetical proteins or with no predicted function, 31of which are predicted to enter the secretory pathway.

Transcriptional Profiles of clr Deletion Mutants. To compare ex-pression of the Avicel regulon with WT, the Δclr-1 and Δclr-2mutants were subjected to transcriptional profiling (Fig. 3). Fourhours after transfer to Avicel media, both clr deletion strains

showed a expression profile similar to WT no carbon samples(Fig. S6) and failed to induce transcription of all cellulases (Fig.3A) and 5 of the 10 most highly expressed hemicellulase genes(Fig. S3). Hierarchical clustering of expression patterns for the212 genes specifically induced on Avicel (Fig. 3B) revealed fourmajor groups of genes with similar expression patterns. A sum-mary of the Avicel regulon and dependence of major categorieson clr-1 and clr-2 is presented in Fig. 3C.Group A (clr-1/2–dependent) consisted of 98 genes that

showed no induction in both Δclr-1 and Δclr-2 mutants. Thisgroup included 16 cellulases, 6 hemicellulases, 15 genes encodingenzymes with activity on polysaccharides, and 14 other enzymespredicted to enter the secretory pathway (Dataset S1). Thegroup also includes three components of the ER translocationcomplex, two disulfide isomerases, as well as two ER chaperones(HSP70 and calreticulin). Genes within group A were the moststrongly induced in the Avicel regulon.Group B (clr-1–dependent) consisted of 28 genes (including

clr-2) that were not induced above no carbon levels in the Δclr-1mutant and had intermediate expression levels in the Δclr-2strain. This group of genes was the least well-supported clusterand many could be included in group A. This group is dominatedby 16 conserved hypothetical proteins, but also includes sixenzymes with predicted function on polysaccharide and oligo-saccharides (including a putative β-glucosidase), a xlnR/xyr1 ho-molog (NCU06971), and a major exoglucanase, NCU07190.Group C (clr-modulated) consisted of 43 genes, the expression

levels of which were modulated by clr-1 and/or clr-2, but whichwere at least partially induced over no carbon levels in bothdeletion mutants (Dataset S1). This group includes 18 secretedenzymes, including three lipases, three monoxygenases, fiveenzymes with predicted activity on polysaccharides or dis-accharides, and four transporters, including NCU08114, a high-affinity cellodextrin transporter (19).Group D (clr-independent) consisted of 42 genes with WT in-

duction in both mutants. This group included four highly expressedhemicellulase genes and several genes with predicted functionrelated to xylose or other pentose utilization (Dataset S1).

A C

DBGroup A

clr-1/2 dependentGroup B

clr-1 depGroup Cmodulated

Group Dindependent

WT No Carbon

∆clr-2 Avicel

WT Avicel 1.0 0.7 0.3 0.0-0.3-0.7-1.0

∆clr-1 Avicel

0 10 20 30 40 50 60

Cellulases Hemicellulases

Polysaccharide Active Protein Processing

Others Hypotheticals Transporters

Transcription Factors Carbon Metabolism

Others Hypotheticals

Sec

rete

d N

on S

ecre

ted

Number of Genes

clr-1/2 dependent clr modulated clr-1 dependent clr independent

10

100

1000

10000

100000

WT No Carbon

∆clr-1 Avicel

∆clr-2 Avicel

WT Avicel

FPK

M

7340 GH6 9680 GH6 7190 GH6 2240 GH61 2916 GH61 8760 GH61 5057 GH7 0762 GH5 0206 CDH 1050 GH61

0

500

1000

1500

0 1 2 3 4

FPK

M

Hours After Media Shift

clr-1 in WT (No Carbon) clr-2 in WT (No Carbon)clr-1 in WT (Avicel)clr-2 in WT (Avicel)clr-1 in ∆clr-2 (Avicel) clr-2 in ∆clr-1 (Avicel)

Fig. 3. The clr regulon. (A) FPKM of 10 cellulase genes in WT and clr deletion strains when exposed to Avicel for 4 h. Error bars show ± 1 SD for conditionswith triplicate measurements. Conditions without error bars are from single clr mutant libraries. (B) Expression patterns for the 212 genes in the Avicelregulon in WT and the clr mutants. Group A includes 98 genes that were clr-1/2–dependent, Group B includes 28 genes that were clr-1–dependent, Group Ccontained 43 genes that were clr-modulated, and group D included 42 genes that were clr-independent. (C) Distribution of genes in functional categories (29)among groups identified in B. (D) Time course of expression of clr-1 and clr-2 in WT and clr mutant strains after transfer to Avicel from sucrose media.

Coradetti et al. PNAS | May 8, 2012 | vol. 109 | no. 19 | 7399

GEN

ETICS

Dow

nloa

ded

by g

uest

on

June

22,

202

0

We used a selection of promoter regions of the most stronglyinduced cellulase genes to search for conserved motifs that couldfunction as binding sites for CLR-1/CLR-2 or other factors incomplex with them. We identified several putative motifs witha variety of established algorithms (Fig. S7). However, whenconsidered with the genome at large, occurrence of these motifsin a promoter region for a gene had no correlation with relativeexpression of that gene on Avicel.

Transcriptional Regulation of clr-2 by clr-1. When WT N. crassa wastransferred to Avicel, clr-1 and clr-2 initially showed an expres-sion level similar to no carbon controls (de-repressed phase)(Fig. 3D). However, although clr-1 was more highly expressed inthe de-repressed phase, clr-2 showed significantly higher ex-pression levels in the inductive phase. Importantly, in the Δclr-1mutant, expression of clr-2 was abolished (Fig. 3D). Expressionlevels of clr-1 were unaffected in the Δclr-2 mutant. These datasuggest that clr-2 is under transcriptional regulation by CLR-1and that the relative abundance of CLR-1 under de-repressedconditions is insufficient for induction of clr-2 expression. Thus,some signaling to CLR-1 through posttranslational modificationor alteration in protein interactions is likely required for its ac-tivation (see Fig. 5).

Conservation of CLR Protein Sequences and Function in FilamentousAscomycete Fungi. A phylogenetic analysis of CLR-1 and CLR-2protein sequences showed broad conservation across filamen-tous ascomycete fungi. Maximum-likelihood phylogenetic treesof CLR-1 and CLR-2 homologs largely recapitulate previouslypublished fungal trees (31) (Fig. S8). One notable exception tothis congruence of species and gene trees was the CLR-1 ho-molog from H. jecorina. Homologs for both proteins can be foundin most filamentous ascomycete species; exceptions are the an-imal fungal pathogens Coccidioides immitus and Coccidioidesposadasii, and the closely related saprophyte Uncinocarpus reesii,which have lost their clr-1 homolog. Additionally, a clr-2 ho-molog was not identified for Grosmannia clavigera, which causesblue-stain disease on pines.To assess whether clr-1 and clr-2 homologs function to regu-

late genes involved in plant cell-wall deconstruction in otherfilamentous ascomycete species, we created deletion strains inthe distantly related fungus A. nidulans in AN5808 (clrA) andAN3369 (clrB). Similar to N. crassa Δclr-1 and Δclr-2 mutants,the A. nidulans ΔclrA and ΔclrB deletion strains were deficientfor cellulase and xylanase activity, as well as total protein se-cretion when pregrown glucose cultures were transferred toAvicel (Fig. 4). Enzyme activity was abolished in the ΔclrB mu-tant, but the ΔclrA mutant showed ∼50% of WT activity (Fig. 4 Aand B). Both deletion mutants were deficient for growth oncellobiose, although ΔclrB was more strongly affected (Fig. 4C).Consistent with enzyme data, the induction pattern of majorcellulase genes in the ΔclrB mutant was several thousand-foldless than WT (Fig. 4D), but in the ΔclrA mutant, the averageinduction was two- to fourfold less. On a per-gene basis, thisdecrease was not statistically significant (P < 0.05) for three offour tested cellulases (P = 0.049, 0.052, 0.105, and 0.121 forAN1273, AN7230, AN0494, and AN5175, respectively), butconsidering all of the genes together, the null hypothesis thatΔclrA has WT levels of cellulase gene expression was not sup-ported. From these data we conclude that clrA has a less im-portant role in cellulase induction in A. nidulans compared withclr-1 in N. crassa. However, the function of CLR-2/ClrB as anessential activator for cellulase gene expression and activity isconserved between N. crassa and A. nidulans, two of the mostwidely divergent species of filamentous ascomycete fungi.

DiscussionBy using genomic resources for N. crassa, we identified twoconserved transcription factors that are important regulators ofgenes encoding enzymes that are required for deconstruction ofcellulose. Despite the complete loss of induced cellulase geneexpression and cellulase secretion in Δclr-1 and Δclr-2 strains,growth and hemicellulase secretion were normal on xylan. Theseresults are in contrast to the function of XlnR/XYR1 in Aspergilliand H. jecorina, where this transcription factor is necessary forregulation of some cellulase and hemicellulase genes (3, 30).However, a XlnR/XYR1 homolog is not necessary for cellulasegene expression or activity in N. crassa or Fusarium species (12–14). In these species, XlnR/XYR1 homologs regulate the ex-pression of hemicellulase genes and are required for utilizationof hemicellulose. These data suggest that some fungal specieshave evolved independent mechanisms for expression of cellu-lase and hemicellulases genes in response to distinct inducersfrom cellulose or hemicellulose. CLR-1 and CLR-2 are appar-ently components of such a specific pathway for cellulose sensingin N. crassa.Cellobiose and its transglycosylation products have been

considered the primary inducing molecules for fungal cellulases(9, 32). In this study, the Δclr-1mutant grew poorly on cellobiose,suggesting that CLR-1 is a crucial element in N. crassa’s cello-biose sensing mechanism. Under this model (Fig. 5), CLR-1 isactivated for induction of its target genes in the presence ofcellobiose or its products. Activated CLR-1 promotes expressionof several genes necessary for efficient import and utilization ofcellobiose, as well as that of clr-2. CLR-2 then directly inducescellulase and select hemicellulase gene expression, possibly ina heterocomplex with activated CLR-1.Homologs of clr-1 and clr-2 are present in genomes of fila-

mentous ascomycete fungi, particularly those in association withplants, suggesting conserved regulatory mechanisms associatedwith genes within the Avicel regulon. In support of this hy-pothesis, deletion mutants for the A. nidulans clr-1 and clr-2homologs (clrA and clrB, respectively) revealed both conserved

A

B

***

0

1

WT ∆clrA ∆clrB WT ∆clrA ∆clrB Rel

. Act

ivity

on A

vice

l 48 hrs 72 hrs 120 hrs

0

100

200

WT ∆clrA ∆clrB Pro

tein

(mg/

L)on

Avi

cel

1

10

100

1000

10000

WT ∆clrA ∆clrB

Fold

Cha

nge AN1273 GH6

AN7230 CDH AN0494 GH7 AN5176 GH7

0

2

4

WT ∆clrA ∆clrB

Dry

Wei

ght

(g/L

)

Glucose Cellobiose

CMCase Xylanase

Total Protein

Induction

GrowthC

D

*

*

*** *

* **

* * **

*** **** * **

** ** ***

Fig. 4. Phenotype of A. nidulans clr deletion strains. (A) Enzyme activity ofculture supernatants from ΔclrA and ΔclrB mutants grown on glucose andthen shifted to Avicel medium. (B) Total protein in supernatants of culturesgrown on Avicel. (C) Mycelial dry weights from WT and clr mutants fromcultures on glucose and cellobiose (0.5% wt/vol). (D) Induction of selectedcellulase genes in WT and the clr mutants following an 8-h shift to Avicel, byquantitative RT-PCR. Statistical significance by one tailed, unequal variancet-test. *P < 0.05, **P < 0.01, ***P < 0.001.

7400 | www.pnas.org/cgi/doi/10.1073/pnas.1200785109 Coradetti et al.

Dow

nloa

ded

by g

uest

on

June

22,

202

0

and divergent aspects. Induction of cellulase genes required clrBbut not clrA. However, ΔclrA was deficient for cellobiose utili-zation. Thus, ClrA may have a conserved role in filamentousascomycete species in a cellobiose-sensing pathway. Similar to N.crassa clr-2, in A. nidulans, clrB was absolutely required for cel-lulolytic activity. However, in Aspergilli XlnR also regulates ex-pression of some cellulase genes (5). These data suggest diversityin the signaling pathways that activate cellulases through XlnRand ClrB in filamentous fungi. Further comparisons of the CLR-1/CLR-2 and XlnR regulons in diverse filamentous fungi willdefine conserved and divergent features associated with regula-tion of genes involved in plant cell-wall deconstruction.We identified a 212-gene set that specifically responded to

crystalline cellulose (Avicel regulon) with a majority of this geneset requiring clr-1, clr-2, or both to reach WT-induced expressionlevels. This gene set likely comprises the essential core structuraland enzymatic components for efficient degradation of cellulose.The gene set not only includes all highly expressed cellulasegenes, but many hemicellulase genes, indicating that compli-mentary enzymatic activities are required to effectively degradecomplex substrates (33). In addition, dozens of secreted proteinswith no predicted function or only a broad homology-based an-notation showed clr-dependent induction. Homology searcheswith N. crassa genes within the Avicel regulon against the A.nidulans and H. jecorina genomes showed that many of thesegenes are conserved (Dataset S1). This collection of proteinsprovides a rich source for identification of new enzymes withsynergistic activities on complex substrates, as recently dem-onstrated (20, 21). Interestingly, included in this gene set weresome well-characterized proteins with functions in secretion, sug-gesting that modification of the N. crassa secretory pathway isneeded to accommodate the dramatic increase in protein secre-tion that accompanies cellulosic metabolism.Promoter analyses did not reveal a statistically significantly

enriched binding motif in clr-dependent genes. Full genomeanalyses of DNA binding interactions by ChIP-Seq of both CLR-1and CLR-2 in context with other transcriptional activators, such asthe XlnR homolog in N. crassa (xlr-1) (14), will elucidate the mo-lecular mechanism of induction and regulation of genes involvedin plant cell-wall deconstruction. Further characterization andmanipulation of the components of the cellulase regulatorynetwork will enable rational engineering of more diverse andproductive industrial strains for cellulolytic enzyme production.

Materials and MethodsStrains. The WT reference strain and background for all N. crassa mutantstrains was FGSC 2489 (22). Deletion strains for clr-1 and clr-2 (FGSC 11029

and FGSC 15835, respectively) were obtained from the Fungal Genetics StockCenter (http://www.fgsc.net/) (34). The A. nidulans reference strain was FGSC4A. Gene deletions in A. nidulans were carried out by transforming FGSCA1149 (pyrG89; pyroA4; nkuA::argB) by established protocols (35). Trans-formants were crossed to LO1496 (fwA1, pyrG89, nicA2, pabaA1, from BerlR. Oakley Department of Molecular Biosciences, University of Kansas, Law-rence, KS) to remove nkuA::argB and pyroA4.

Culture Conditions. N. crassa cultures were grown on Vogel’s minimal medium(VMM) (36). Carbon sources were 2% wt/vol unless otherwise noted. Conidiawere inoculated into 100 mL liquid media at 106 conidia/mL and grown at 25 °C in constant light and shaking (200 rpm). A. nidulans cultures were grown onminimal medium (MM) (37). Carbon sources were 1% wt/vol unless otherwisenoted. Conidia were inoculated into 100mL liquid media at 4 × 106 conidia/mLand grown at 37 °C in constant light and shaking (200 rpm).

Media Shift Experiments. N. crassa cultures were grown 16 h on VMM,centrifuged at 2,390 × g for 10 min, and washed with VMM (36) withouta carbon source. Mycelia were resuspended in 100 mL VMM with 2% (wt/vol)carbon source [sucrose, cellulose (Avicel PH-101; Sigma-Aldrich) or hemi-cellulose (beechwood xylan, Sigma-Aldrich)] or with no carbon source added.Total RNA was extracted as described in ref. 6. For enzyme activity assays,culture supernatants were sampled at 24 and 48 h, centrifuged at 2,390 × gtwice to remove mycelia, and stored at 4 °C for analysis within 24 h.

A. nidulans cultures were grown 16–17 h on MM-glucose. A 15 mL samplewas taken at time 0. The remaining culture was filtered through miracloth,washed, and transferred to 100 mL MM containing 1% Avicel. RNA wasextracted as above and mRNA abundance was compared between the 8 h andtime 0 samples by quantitative RT-PCR (38). Fold-induction was calculated as theratio of the mRNA level normalized to act A at 8 h vs. act A at time 0. For en-zyme activity assays, culture supernatants were sampled at 48–120 h, centri-fuged at 2,390 × g twice to remove mycelia and stored at 4 °C for analysis.

RNA Sequencing and Transcript Abundance. Libraries were prepared withstandard protocols from Illumina and sequenced on the Illumina GenomeAnalyzer IIx and HiSEq 2000 platforms (SI Materials Methods). Sequencedlibraries were mapped against predicted transcripts from the N. crassaOR74A genome (39) (v10) with Tophat v1.2.0 (40). Transcript abundance(FPKM, fragments per kilobase unique exon sequence per megabase of li-brary mapped) was estimated with Cufflinks v0.9.3 (41) using upper quartilenormalization and mapping against reference isoforms. Profiling data areavailable in Dataset S1 and at the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/; accession no. GSE35227).

Independent triplicate cultures were sampled and analyzed for the WTstrain on Avicel, sucrose, and no carbon at 4 h after the media shift. Rawcounts for reads mapping to unique exons were tallied with HTSEq (42) andused as inputs for both the EdgeR and DESeq software packages. Genes witha multiple-hypothesis–adjusted P value below 0.05 from both models weretaken as differentially expressed between conditions.

FPKM of genes in the Avicel regulon were hierarchically clustered with theCluster 3.0 software suite (http://bonsai.hgc.jp/∼mdehoon/software/cluster/software.htm). Before clustering, FPKMs were log-transformed, normalizedacross strains/conditions, and centered on the geometric mean across strains/conditions on a per-gene basis. The average linkage method was used forcluster generation, with uncentered correlation as the similarity metric.

Enzyme Activity Assays. Relative enzymatic activity of culture supernatantswas assayed with Remazol brilliant Blue R-conjugated CMC and xylan(birchwood) kits from Megazyme. Total protein was determined with theBradford assay (BioRad).

Statistical Significance Tests.Unless otherwise noted, all statistical significancetests were done with a one-tailed homoscedastic (equal variance) t-test. *P <0.05, **P < 0.01, ***P < 0.001.

ACKNOWLEDGMENTS. The authors thank Spencer Diamond for his work onthe initial screens of putative Neurospora crassa transcription factor mutantson cellulosic substrates. This work was supported by a grant from the EnergyBiosciences Institute and a Miller Institute Professorship (to N.L.G.) and Na-tional Institutes of Health National Research Service Award Trainee Grant 2T32 GM 7127-36 A1 (to S.T.C.). We are pleased to acknowledge use ofmaterials generated by P01 GM068087 Functional analysis of a modelfilamentous fungus.

CLR-1

CarbonCataboliteRepression

CLR-1

clr-2

β-glucosidases

cellulases

transporters

Cellobiose

clr-1

CLR-2Glucose

Derepression

Cellulose

Fig. 5. Model of cellulase induction mediated by CLR-1 and CLR-2 in N.crassa. When de-repressed, genes encoding cellulases, β-glucosidases, cello-biose transporters, and CLR-1 are expressed at low levels. CLR-1 is activatedin the presence of cellobiose, promoting expression of cellodextrin trans-porters and β-glucosidase genes. Activated CLR-1 is necessary for increasedexpression levels of clr-2 and may also regulate its own expression. CLR-2(possibly in concert with CLR-1) drives expression of the Avicel regulon, re-leasing more cellobiose in a positive feedback loop.

Coradetti et al. PNAS | May 8, 2012 | vol. 109 | no. 19 | 7401

GEN

ETICS

Dow

nloa

ded

by g

uest

on

June

22,

202

0

1. Himmel ME, et al. (2007) Biomass recalcitrance: Engineering plants and enzymes forbiofuels production. Science 315:804–807.

2. Jordan DB, et al. (2012) Plant cell walls to ethanol. Biochem J 442:241–252.3. Kubicek CP, Mikus M, Schuster A, Schmoll M, Seiboth B (2009) Metabolic engineering

strategies for the improvement of cellulase production by Hypocrea jecorina. Bio-technol Biofuels 2:19.

4. Ilmén M, Saloheimo A, Onnela ML, Penttilä ME (1997) Regulation of cellulase geneexpression in the filamentous fungus Trichoderma reesei. Appl Environ Microbiol 63:1298–1306.

5. Gielkens MM, Dekkers E, Visser J, de Graaff LH (1999) Two cellobiohydrolase-encod-ing genes from Aspergillus niger require D-xylose and the xylanolytic transcriptionalactivator XlnR for their expression. Appl Environ Microbiol 65:4340–4345.

6. Tian C, et al. (2009) Systems analysis of plant cell wall degradation by the model fil-amentous fungus Neurospora crassa. Proc Natl Acad Sci USA 106:22157–22162.

7. Stricker AR, Mach RL, de Graaff LH (2008) Regulation of transcription of cellulases-and hemicellulases-encoding genes in Aspergillus niger and Hypocrea jecorina (Tri-choderma reesei ). Appl Microbiol Biotechnol 78:211–220.

8. Mandels M, Reese ET (1957) Induction of cellulase in Trichoderma viride as influencedby carbon sources and metals. J Bacteriol 73:269–278.

9. Mandels M, Reese ET (1960) Induction of cellulase in fungi by cellobiose. J Bacteriol79:816–826.

10. Mandels M, Parrish FW, Reese ET (1962) Sophorose as an inducer of cellulase in Tri-choderma viride. J Bacteriol 83:400–408.

11. van Peij NN, Gielkens MM, de Vries RP, Visser J, de Graaff LH (1998) The transcrip-tional activator XlnR regulates both xylanolytic and endoglucanase gene expressionin Aspergillus niger. Appl Environ Microbiol 64:3615–3619.

12. Brunner K, Lichtenauer AM, Kratochwill K, Delic M, Mach RL (2007) Xyr1 regulatesxylanase but not cellulase formation in the head blight fungus Fusarium graminea-rum. Curr Genet 52:213–220.

13. Calero-Nieto F, Di Pietro A, Roncero MI, Hera C (2007) Role of the transcriptionalactivator xlnR of Fusarium oxysporum in regulation of xylanase genes and virulence.Mol Plant Microbe Interact 20:977–985.

14. Sun J, Tian C, Diamond S, Glass NL (2012) Deciphering transcriptional regulatorymechanisms associated with hemicellulose degradation in Neurospora crassa. Eu-karyot Cell 11:482–493.

15. Aro N, Saloheimo A, Ilmén M, Penttilä M (2001) ACEII, a novel transcriptional acti-vator involved in regulation of cellulase and xylanase genes of Trichoderma reesei. JBiol Chem 276:24309–24314.

16. Turner BC, Perkins DD, Fairfield A (2001) Neurospora from natural populations: Aglobal study. Fungal Genet Biol 32:67–92.

17. Davis RH, Perkins DD (2002) Timeline: Neurospora: A model of model microbes. NatRev Genet 3:397–403.

18. Phillips CM, Iavarone AT, Marletta MA (2011) Quantitative proteomic approach forcellulose degradation by Neurospora crassa. J Proteome Res 10:4177–4185.

19. Galazka JM, et al. (2010) Cellodextrin transport in yeast for improved biofuel pro-duction. Science 330:84–86.

20. Phillips CM, Beeson WT, Cate JH, Marletta MA (2011) Cellobiose dehydrogenase anda copper-dependent polysaccharide monooxygenase potentiate cellulose degrada-tion by Neurospora crassa. ACS Chem Biol 6:1399–1406.

21. Beeson WT, Phillips CM, Cate JH, Marletta MA (2012) Oxidative cleavage of celluloseby fungal copper-dependent polysaccharide monooxygenases. J Am Chem Soc 134:890–892.

22. Colot HV, et al. (2006) A high-throughput gene knockout procedure for Neurosporareveals functions for multiple transcription factors. Proc Natl Acad Sci USA 103:10352–10357.

23. Aro N, Ilmén M, Saloheimo A, Penttilä M (2003) ACEI of Trichoderma reesei is a re-pressor of cellulase and xylanase expression. Appl Environ Microbiol 69:56–65.

24. Zeilinger S, Mach RL, Kubicek CP (1998) Two adjacent protein binding motifs in thecbh2 (cellobiohydrolase II-encoding) promoter of the fungus Hypocrea jecorina(Trichoderma reesei ) cooperate in the induction by cellulose. J Biol Chem 273:34463–34471.

25. Lockington RA, Rodbourn L, Barnett S, Carter CJ, Kelly JM (2002) Regulation by car-bon and nitrogen sources of a family of cellulases in Aspergillus nidulans. FungalGenet Biol 37:190–196.

26. Tilburn J, et al. (1995) The Aspergillus PacC zinc finger transcription factor mediatesregulation of both acid- and alkaline-expressed genes by ambient pH. EMBO J 14:779–790.

27. Sun J, Glass NL (2011) Identification of the CRE-1 cellulolytic regulon in Neurosporacrassa. PLoS ONE 6:e25654.

28. MacPherson S, Larochelle M, Turcotte B (2006) A fungal family of transcriptionalregulators: the zinc cluster proteins. Microbiol Mol Biol Rev 70:583–604.

29. Ruepp A, et al. (2004) The FunCat, a functional annotation scheme for systematicclassification of proteins from whole genomes. Nucleic Acids Res 32:5539–5545.

30. Stricker AR, Grosstessner-Hain K, Würleitner E, Mach RL (2006) Xyr1 (xylanase regu-lator 1) regulates both the hydrolytic enzyme system and D-xylose metabolism inHypocrea jecorina. Eukaryot Cell 5:2128–2137.

31. McLaughlin DJ, Hibbett DS, Lutzoni F, Spatafora JW, Vilgalys R (2009) The search forthe fungal tree of life. Trends Microbiol 17:488–497.

32. Znameroski EA, et al. (2012) Induction of lignocellulose degrading enzymes in Neu-rospora crassa by cellodextrins. Proc Natl Acad Sci USA 109:6012–6017.

33. Benko Z, Siika-aho M, Viikari L, Reczey K (2008) Evaluation of the role of xylogluca-nase in the enzymatic hydrolysis of lignocellulosic substrates. Enzyme Microb Technol43(2):109–114.

34. McCluskey K (2003) The Fungal Genetics Stock Center: From molds to molecules. AdvAppl Microbiol 52:245–262.

35. Szewczyk E, et al. (2006) Fusion PCR and gene targeting in Aspergillus nidulans. NatProtoc 1:3111–3120.

36. Vogel HJ (1956) A convenient growth medium for Neurospora. Microbiol Genet Bull13:42–46.

37. Vishniac W, Santer M (1957) The thiobacilli. Bacteriol Rev 21:195–213.38. Livak KJ, Schmittgen TD (2001) Analysis of relative gene expression data using real-

time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25:402–408.39. Galagan JE, et al. (2003) The genome sequence of the filamentous fungus Neurospora

crassa. Nature 422:859–868.40. Langmead B, Trapnell C, Pop M, Salzberg SL (2009) Ultrafast and memory-efficient

alignment of short DNA sequences to the human genome. Genome Biol 10:R25.41. Roberts A, Trapnell C, Donaghey J, Rinn JL, Pachter L (2011) Improving RNA-Seq ex-

pression estimates by correcting for fragment bias. Genome Biol 12:R22.42. Anders S, Huber W (2010) Differential expression analysis for sequence count data.

Genome Biol 11:R106.43. Cantarel BL, et al. (2009) The Carbohydrate-Active EnZymes database (CAZy): An

expert resource for Glycogenomics. Nucleic Acids Res 37(Database issue):D233–D238.44. Robinson MD, McCarthy DJ, Smyth GK (2010) edgeR: A bioconductor package for

differential expression analysis of digital gene expression data. Bioinformatics 26(1):139–140.

7402 | www.pnas.org/cgi/doi/10.1073/pnas.1200785109 Coradetti et al.

Dow

nloa

ded

by g

uest

on

June

22,

202

0