Network of nutrient-sensing pathways and a conservedkinase cascade integrate osmolarity and carbonsensing in Neurospora crassaLori B. Hubermana,b, Samuel T. Coradettia,b, and N. Louise Glassa,b,c,1

aPlant and Microbial Biology Department, University of California, Berkeley, CA 94720; bEnergy Biosciences Institute, University of California, Berkeley,CA 94720; and cEnvironmental Genomics and Systems Biology Division, Lawrence Berkeley National Laboratory, Berkeley, CA 94720

Edited by Jay C. Dunlap, Geisel School of Medicine at Dartmouth, Hanover, NH, and approved September 1, 2017 (received for review May 10, 2017)

Identifying nutrients available in the environment and utilizing themin the most efficient manner is a challenge common to all organisms.The model filamentous fungus Neurospora crassa is capable of uti-lizing a variety of carbohydrates, from simple sugars to the complexcarbohydrates found in plant cell walls. The zinc binuclear clustertranscription factor CLR-1 is necessary for utilization of cellulose, amajor, recalcitrant component of the plant cell wall; however, ex-pression of clr-1 in the absence of an inducer is not sufficient toinduce cellulase gene expression. We performed a screen for un-identified actors in the cellulose-response pathway and identifieda gene encoding a hypothetical protein (clr-3) that is required forrepression of CLR-1 activity in the absence of an inducer. Using clr-3mutants, we implicated the hyperosmotic-response pathway in thetunable regulation of glycosyl hydrolase production in response tochanges in osmolarity. The role of the hyperosmotic-response path-way in nutrient sensing may indicate that cells use osmolarity as aproxy for the presence of free sugar in their environment. Thesesignaling pathways form a nutrient-sensing network that allowsN. crassa cells to tightly regulate gene expression in response toenvironmental conditions.

MAP kinase cascade | carbon metabolism | hyperosmotic response |carbon catabolite repression | plant biomass degradation

Accurately sensing and responding to nutrients is a challengecommon to all organisms. Complex signaling networks have

evolved to efficiently deploy resources required to harvest andutilize nutrients with the least energy expended by the cell. Inhumans, inaccurate nutrient sensing can result in type II diabetesand obesity (1). Mutations in nutrient-sensing pathways also play arole in cancer progression, since rapid growth of tumors causesphysiological changes that result in abnormal nutrient require-ments and utilization (2). In fungi, inaccurate nutrient sensing andsignaling can result in slow growth or an inability to appropriatelyutilize nutrients in the environment (3).Saprophytic filamentous fungi are capable of consuming a

wide variety of carbohydrates from simple sugars to the complexcarbohydrates found in plant cell walls. Utilization of thesecomplex carbohydrates requires the cell to activate expression ofgenes encoding secreted enzymes that degrade insoluble carbo-hydrates into sugars that can be subsequently imported into thecell (3). If these enzymes are not produced, the cell cannot utilizethese complex carbon sources (4–6). However, production ofsuch enzymes when preferred carbon sources are present resultsin a competitive disadvantage (7). Thus, filamentous fungi haveevolved a complex nutrient-sensing network that queries thestate of the environment to activate expression of these enzymesonly when complex carbohydrates are present and preferredcarbon sources are absent (8–10). In this study, we use the cel-lulolytic response of the filamentous fungus Neurospora crassa asa model to investigate the interplay of various signaling pathwaysthat regulate the cellular response to preferred and nonpreferredcarbon sources.

Cellulose, the major component of the plant cell wall, is apolymer of β-(1–4)–linked glucose units that is highly recalcitrantto degradation. In fungi that can utilize cellulose, lignocellulolyticgene expression is repressed when preferred carbon sources arepresent through a process known as “carbon catabolite repression”(11). There are several transcription factors involved in carboncatabolite repression (3). The best studied is the zinc fingertranscription factor cre-1 (NCU08807), the N. crassa ortholog ofSaccharomyces cerevisiae MIG1 (12, 13). CRE-1 represses theexpression of cellulase genes in response to a range of simplesugars, including glucose, and products of cellulose degradation,such as the disaccharide cellobiose (14).When cellulose is present in the absence of preferred carbon

sources, the induction of cellulolytic genes in N. crassa is de-pendent on two zinc binuclear cluster transcription factors: clr-1(NCU07705) and clr-2 (NCU08042) (Fig. S1A) (4). In the ab-sence of an inducer, clr-1 is expressed but unable to activate theexpression of cellulase genes. When an inducer, such as a deg-radation product of cellulose, is present, CLR-1 activates theexpression of a small number of genes, including several β-glu-cosidases and the transcription factor clr-2 (4). CLR-2 is re-sponsible for the majority of cellulase gene expression (15).Deletion of either clr-1 or clr-2 abolishes the cellulolytic responseand eliminates the ability of N. crassa cells to utilize cellulose as acarbon source (4). Constitutive expression of clr-2 results in theactivation of cellulase gene expression even in the absence of aninducer (15). However, expression of clr-1 in the absence of an

Significance

Microbes have evolved complex signaling networks to identifyand prioritize utilization of available energy sources. For manyfungi, such as Neurospora crassa, this entails distinguishing be-tween an array of carbon sources, including insoluble carbohy-drates in plant cell walls. Here, we identified a repressor of thecellulose-response pathway in N. crassa. Using this derepressedmutant, we implicated the conserved hyperosmotic-responseMAP kinase pathway in regulating the response of N. crassa toinsoluble carbohydrates. We hypothesize that fungal speciesthat degrade plant biomass use osmolarity as a proxy for solublesugar in the environment to regulate their nutritional responses,enabling tailored production of lignocellulases. This findingcould help in battling fungal plant diseases and in the pro-duction of second-generation biofuels.

Author contributions: L.B.H., S.T.C., and N.L.G. designed research; L.B.H. and S.T.C. per-formed research; L.B.H. and S.T.C. analyzed data; and L.B.H. and N.L.G. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Data deposition: RNA-seq data used in this study were deposited in the Gene ExpressionOmnibus (GEO) database, https://www.ncbi.nlm.nih.gov/geo (accession no. GSE95681).1To whom correspondence should be addressed. Email: Lglass@berkeley.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1707713114/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1707713114 PNAS | Published online September 25, 2017 | E8665–E8674

GEN

ETICS

PNASPL

US

Dow

nloa

ded

by g

uest

on

Oct

ober

28,

202

0

inducer is not sufficient to elicit the full cellulolytic response,leading us to hypothesize that there could be additional, un-identified genes involved (4).To test this hypothesis, we screened for mutants that activated

the cellulolytic response in the absence of an inducer. From thisscreen, we identified a repressor of CLR-1 activity, which wenamed “clr-3.” Deletion of clr-3 resulted in CLR-1–dependentcellulase gene expression in the absence of an inducer. Furtheranalysis of mutants identified in the screen led to the discoverythat the hyperosmotic response (OS) MAP kinase pathway isinvolved in the cellulolytic response in N. crassa.Our data support a model in which the regulation of cellulase

gene expression is the result of the integration of a range ofenvironmental inputs, which are transmitted via nutrient-sensingpathways. One of these is the OS pathway, which we hypothesizeuses osmolarity to sense the presence of free, soluble sugar, asopposed to insoluble, crystalline carbon sources found in plantcell walls. Together with genes, such as cre-1, involved in carboncatabolite repression, clr-3, working through the transcriptionfactor clr-1, and the OS pathway form a nutrient-sensing networkthat transmits information about available nutrients to tightlyregulate the expression of genes necessary to consume non-preferred carbon sources.

ResultsCLR-3 Acts Through CLR-1 to Repress the Cellulolytic Response. Activa-tion of cellulase gene expression in N. crassa requires CLR-1 andCLR-2 (Fig. S1A) (4). Expression of clr-2 is directly regulated byCLR-1, and constitutive expression of clr-2 is sufficient to activatethe cellulolytic response (15, 16). However, transcription of clr-1does not result in the activation of cellulase gene expression in theabsence of an inducer (4).To identify additional regulators of the cellulolytic response,

we placed a hygromycin-resistance marker (HygR) under the reg-ulation of the pmo-4 (NCU01050) promoter, which is bound byboth CLR-1 and CLR-2 and is activated in response to cellulose(Fig. S1A) (16). Wild-type cells containing the Ppmo-4-HygR con-struct grew when exposed to medium containing the inducercarboxymethyl cellulose (CMC) and hygromycin. However,when these cells were exposed to a noninducing carbon source,such as xylan, growth was inhibited when hygromycin was added(Fig. S1B). We mutagenized the Ppmo-4-HygR strain using thechemical mutagen N-methyl-N-nitro-N-nitrosoguanidine (NTG)and screened for mutants able to grow on medium containingxylan and hygromycin (Fig. S1B).We first characterized two mutants, M13 and M201, and con-

firmed that these two mutants expressed cellulases even in theabsence of an inducer. To mitigate possible effects caused by dif-ferential growth speed in media containing different carbonsources, we grew M13 and M201 cells in medium containing su-crose as the sole carbon source and then washed and transferredthe mycelial mass to medium containing an inducing carbon source(cellulose), a noninducing, nonrepressing carbon source (xylan), orstarvation conditions [Vogel’s salts (17) lacking a carbon source].Secreted cellulase activity was measured 72 h posttransfer. Toidentify the causative mutations in M13 and M201 that resulted incellulase activity in the absence of an inducer, we used bulkedsegregant analysis followed by Illumina sequencing.Analysis of sequencing data of the bulk progeny pools showed

that both M13 and M201 contained nonconservative missensepoint mutations in conserved regions in a hypothetical protein,NCU05846 (Fig. S2A). These were the only mutations in codingregions present exclusively in progeny that exhibited inappro-priate cellulase secretion. Both mutants exhibited a modestamount of secreted cellulase activity in the absence of an inducerbut wild-type levels of secreted cellulase activity in the presenceof cellulose (Fig. 1A). A strain carrying a full deletion of NCU05846exhibited an identical cellulase-secretion phenotype, which was

complemented by expressing NCU05846 at the his-3 locus (Fig.1A). N. crassa cells are multinucleate, and it is possible to makeheterokaryons with genetically different nuclei. To assess com-plementation of the cellulase phenotype in M13 and M201, weforced heterokaryons between the mutants and cells that wereeither deleted or wild type for NCU05846 (Fig. S2B). (M13 +ΔNCU05846) and (M201 + ΔNCU05846) heterokaryons failedto complement the inappropriate cellulase-secretion phenotype(Fig. S2C). However, when heterokaryons were made by mixingM13 or M201 with cells wild type at the NCU05846 locus, theexpression of cellulases under noninducing xylan conditions wassignificantly reduced (P < 2 × 10−3) (Fig. S2C). These data in-dicated that mutations in NCU05846 were causative for thecellulase-secretion phenotype; we therefore named NCU05846“clr-3” for “cellulose degradation regulator-3.”The clr-3 promoter is bound by both CLR-1 and CLR-2, and

clr-3 expression is up-regulated in the presence of cellulose (4,16). The CLR-3 protein included a domain of unknown function(DUF1479) with orthologs in sordariomycete species (Fig. S2).The CLR-3 protein family appeared to have undergone dupli-cation and loss, with a variable number of DUF1479-containingparalogs in the genomes of ascomycete and basidiomycete spe-cies. The N. crassa genome also contained a clr-3 paralog,NCU02600, whose expression is up-regulated in the presence ofcellulose (Fig. S2D) (4). To determine whether NCU02600 isinvolved in regulating cellulase gene expression, we measuredthe secreted cellulase activity of cells lacking NCU02600. Nodifferences in the secreted cellulase activity of ΔNCU02600 cellscompared with wild-type cells or any combinatorial effects ofdeleting both clr-3 and NCU02600 were detected (Fig. S2C).Because changes in the secreted cellulase activity of the Δclr-3

mutant could be due to defects in any aspect of cellulase pro-duction, from mRNA abundance to protein secretion, we askedwhether clr-3 was responsible for regulating cellulolytic gene ex-pression. We grew wild-type and Δclr-3 cells in medium containingsucrose as the sole carbon source and subsequently exposed thesecells to either the noninducing carbon source xylan or starvationfor 10 h. qRT-PCR was performed on RNA isolated from thesecultures for two genes whose expression is known to be activatedduring the cellulolytic response: clr-2 and cbh-1 (NCU07340) (4).In wild-type cells, expression of clr-2 is regulated by CLR-1,and expression of the cellulase gene cbh-1 is activated by bothCLR-1 and CLR-2 (4, 16). Under xylan conditions, cells lackingclr-3 showed a 56% increase in clr-2 expression (P = 6 × 10−3) anda 70-fold increase in cbh-1 expression (P = 4 × 10−3) comparedwith wild-type cells (Fig. 1B). The difference in the expressionof clr-2 and cbh-1 in wild-type and Δclr-3 cells during starvation

A Cellulose Xylan Starvation

050

100

Wild

type∆clr

-3

Rel

ativ

e cb

h-1

expr

essi

on

Rel

ativ

e cl

r-2 e

xpre

ssio

n

B Xylan Starvation

024

705030

90

Wild

type∆clr

-3

1,000

150200

2,0003,0004,000

Wild

type

∆clr-3

∆clr-3

clr-3@

his-3

Rel

ativ

e C

MC

ase

activ

ity

00.20.40.60.81.01.2

pmo-4-H

yg

P

RM13

M201

Fig. 1. CLR-3 (NCU05846) negatively regulates the cellulolytic response.(A) Carboxymethyl cellulase (CMCase) activity of enzymes secreted into cul-ture supernatants 72 h posttransfer to the indicated medium by the in-dicated wild-type or mutant cultures. (B) Transcript abundance of clr-2 andcbh-1 relative to act 10 h posttransfer to the indicated medium as measuredby qRT-PCR in wild-type or Δclr-3 (ΔNCU05846) cells. Error bars indicate SDs.Circles indicate the values of individual biological replicates (Dataset S1).

E8666 | www.pnas.org/cgi/doi/10.1073/pnas.1707713114 Huberman et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

28,

202

0

conditions was even more pronounced, with a 16-fold increase in clr-2expression (P = 3 × 10−5) and a 396-fold increase in cbh-1 expression(P = 6 × 10−5) (Fig. 1B). These data indicated that clr-3 regulatescellulase gene expression at the level of transcript abundance.It is possible that CLR-3 acts through the CLR-1/CLR-2

cellulase-induction pathway or, alternatively, it could function inan as-yet unidentified cellulase-regulation pathway. Becausedeletion of clr-3 caused inappropriate expression of clr-2, wehypothesized that CLR-3 repressed the expression of cellulasegenes in a CLR-1–dependent fashion (4, 16). In support of ourhypothesis, cells that lacked both clr-1 and clr-3 were unable tosecrete cellulases even in the presence of cellulose, indicatingthat deletion of clr-3 was unable to suppress the Δclr-1 pheno-type (Fig. 2A). Additionally, constitutive overexpression of clr-1in combination with Δclr-3 resulted in a fourfold increase insecreted cellulase activity during exposure to xylan (P = 3 × 10−6)and a threefold increase in secreted cellulase activity understarvation conditions (P = 10−4) compared with overexpressionof clr-1 in an otherwise wild-type background (Fig. 2A). Thesedata indicated that CLR-3 acts through the CLR-1/CLR-2 path-way to repress the expression of cellulases in the absenceof cellulose.To provide further evidence that CLR-3 acts upstream of

CLR-1 activation, we investigated the effect of deleting clr-3 onthe transcriptome of N. crassa. Previous studies used ChIP se-

quencing (ChIP-seq) and RNA sequencing (RNA-seq) to iden-tify the genes whose promoters are bound by CLR-1 and whoseexpression is activated in a CLR-1–dependent manner (4, 16).Our cellulase activity data (Fig. 2A) indicated that Δclr-1 wasepistatic to Δclr-3, leading us to hypothesize that cells lacking clr-3 will up-regulate the expression of genes whose promoters arebound by and whose expression is regulated by CLR-1. Analysisof RNA-seq data showed that all but one of the genes whosepromoters are bound by CLR-1 were significantly up-regulatedin Δclr-3 cells compared with wild-type cells during starvationconditions (Fig. 2B and Dataset S2). A more comprehensiveanalysis of the transcriptome showed that the expression of72 genes was up-regulated (more than fourfold) in Δclr-3 cellscompared with wild-type cells at both 4 h and 10 h posttransfer tostarvation (clr-3 regulon), and the expression of all but six ofthese genes was also activated during exposure to the crystallinecellulose substrate Avicel (Avicel regulon) (Fig. 2C and DatasetS3) (4, 16). This 66-gene set was highly enriched for genes boundby either CLR-1 or CLR-2 compared with the Avicel regulon asa whole (Fig. 2D and Dataset S2). These data indicated thatCLR-3 acts to repress the action of CLR-1 in the absence ofan inducer.

The Hyperosmotic-Response Pathway Regulates Cellulase GeneExpression. Cellulase activity of Δclr-3 cells in the absence of aninducer was significantly (more than fivefold) less than that ofwild-type cells exposed to cellulose (P < 10−103), indicatingthat other repression or activation mechanisms must be involvedin cellulase induction (Fig. 1A). One of the mutants identified inthe screen, M211, had significantly increased secreted cellulaseactivity under noninducing conditions compared with Δclr-3 cells(P < 3.1 × 10−16) (Fig. S3A). When backcrossed, this mutantsegregated into two groups of progeny with inappropriate cellulaseproduction (Fig. S3 A and B). Progeny in the first group (M211Moderate) were morphologically indistinguishable from both wild-type and Δclr-3 cells and produced cellulases at levels comparableto Δclr-3 cells (Fig. S3 A and B). The second group of progeny(M211 High) had a distinct morphological phenotype that wassimilar to that of the original M211 mutant and exhibited signifi-cantly increased cellulase production compared with Δclr-3 cellsunder noninducing conditions (P < 6.2 × 10−15) (Fig. S3 A and B).We used bulked segregant analysis and Illumina sequencing toidentify a large chromosomal deletion that included clr-3 in boththe M211 Moderate and M211 High progeny (Dataset S4). Fur-ther analysis of the M211 High bulk progeny genomic sequencesshowed a substantial number of additional point mutations, in-cluding a missense mutation in os-1 (NCU02815), os-1R572H, whichis a histidine kinase in the OS pathway (Dataset S4) (18). Histidinekinases are known to transmit information about environmentalstimuli through signal-transduction pathways, and the OS-1R572H

mutation is in the sensor domain of this protein, specifically inthe semistructured connector between the two α-helices of theHAMP domain (present in histidine kinases, adenyl cyclases,methyl-accepting proteins, and phosphatases) (19, 20).The N. crassa OS pathway is homologous to the S. cerevisiae

high-osmolarity glycerol (HOG) pathway and was identified for itsrole in responding to hyperosmotic shock (21, 22). In N. crassa,hyperosmolarity activates the histidine kinase OS-1, which worksthrough a response regulator and a histidine phosphotransferaseto activate the OS MAP kinase cascade [os-4 (NCU03071), os-5(NCU00587), and os-2 (NCU07024)] (18, 23–26). When activatedby hyperosmotic shock, the MAP kinase OS-2 activates the basicleucine zipper domain transcription factor ASL-1 (NCU01345)and other unknown transcription factors, which are responsible foractivating the expression of osmoresponsive genes (Fig. 3A) (27).Because the morphological phenotype of the M211 High

mutant progeny was similar to that of OS pathway mutants, wehypothesized that the OS pathway mutation was responsible for

BA

C D

High

Low

Cellulose Starvation

WT ∆clr

-1∆clr

-2W

T∆clr-3

∆clr-3

(10h)

WT (1

0h)

WT ∆clr

-1∆clr

-2∆clr

-3 (10

h)

WT (1

0h)

CLR-1CLR

-2∆clr

-3W

T

Cellulose Starvation Target of

WT ∆clr

-1∆clr

-2∆clr

-3 (10

h)

WT (1

0h)

CLR-1CLR

-2Avic

el

∆clr-3

WT

Cellulose Starvation Target of

CLR-3

CLR-1 Bound Genes

Avicel Regulonclr-3 Regulon

Cellulose Xylan Starvation

00.20.40.60.81.01.2

Wild

type

∆clr-3

∆clr-1

∆clr-1

∆clr-3

Rel

ativ

e C

MC

ase

activ

ity

ccg-1P-cl

r-1 ∆clr

-3

ccg-1P-cl

r-1

ccg-1P-cl

r-2

ccg-1P-cl

r-2

ccg-1P-cl

r-2

Fig. 2. CLR-3 represses the activity of the transcription factor CLR-1.(A) CMCase activity of enzymes secreted into culture supernatants 72 h post-transfer to the indicated medium by the indicated wild-type or mutant cul-tures. Error bars indicate SDs. Circles indicate the values of individual biologicalreplicates (Dataset S1). (B) Heat map of the expression level of genes whosepromoters are bound by CLR-1 (16). (C) Heat map of the expression level ofgenes whose expression is more than fourfold differentially regulated be-tween wild-type and Δclr-3 cells at both 4 h and 10 h posttransfer to starvationconditions. (D) Heat map of the expression level of genes in the Avicel regulon[as described in Coradetti et al. (4) with the addition of genes whose pro-moters are bound by either clr-1 or clr-2 (16)]. Unless otherwise noted, cellswere harvested for RNA-seq 4 h posttransfer to the indicated medium. Purplebars indicate genes whose promoters are bound by clr-1. Pink bars indicategenes whose promoters are bound by clr-2. Orange bars indicate genes in theAvicel regulon. Green bars indicate genes whose expression is more thanfourfold differentially regulated in Δclr-3 cells compared with wild-type cells atboth 4 h and 10 h posttransfer to starvation.

Huberman et al. PNAS | Published online September 25, 2017 | E8667

GEN

ETICS

PNASPL

US

Dow

nloa

ded

by g

uest

on

Oct

ober

28,

202

0

the increased cellulase activity secreted by the M211 mutants(Fig. S3B). To test this hypothesis, we constructed (M211 + Δclr-3)and (M211 + Δos-1 Δclr-3) heterokaryons and determined theircellulase phenotype (Fig. S3C). As expected, the cellulase-secretion phenotype of the (M211 + Δclr-3) heterokaryon wasindistinguishable from that of Δclr-3 cells (Fig. S3D). However,the secreted cellulase activity of the (M211 + Δos-1 Δclr-3)heterokaryon was sevenfold higher than that of the (M211 +Δclr-3) heterokaryon during starvation conditions (P = 3 × 10−5)(Fig. S3D). These data indicated that the os-1mutation is at leastpartially responsible for the increased cellulase activity of theM211 mutant, although we cannot rule out the possibility thatother mutations in this strain are contributing to its inappropriatecellulase secretion.

To assess the role of the OS pathway in cellulase regulation,we tested strains carrying deletions of members of the OSpathway alone and when these OS mutations were combinedwith Δclr-3. Deletions of members of the OS pathway in anotherwise wild-type background did not have a significant effecton cellulase production (Fig. 3B). However, in a Δclr-3 back-ground, deletions of members of the OS MAP kinase cascadeshowed significantly increased cellulase activity compared withΔclr-3 cells in noninducing conditions, with a twofold increase onxylan and a 10-fold increase under starvation conditions (P <2.1 × 10−15) (Fig. 3B). To confirm that activation of the OS-2MAP kinase was required for proper cellulase regulation, weconstructed a kinase-dead os-2 allele by mutating residues ho-mologous to residues necessary for activation of the S. cerevisiae

A

B

OS-1HyperosmolarityOS-5

MAPKKOS-2MAPK

OS-4MAPKKK

OsmoresponsiveGene Expression

C

D

High

Low

ASL-1

+ unknown TFs

050

100150200250

1,0003,0005,0007,0009,000

11,000

Wild

type∆clr

-3

∆os-1

∆clr-3

∆os-2

∆clr-3

0

2

4

6

40

Wild

type∆clr

-3

∆os-1

∆clr-3

∆os-2

∆clr-3

80120160

Rel

ativ

e cl

r-2 e

xpre

ssio

n

EStarvation

Wild

type

∆clr-3

∆os-1∆os

-2∆os

-1∆clr

-3

∆os-2

∆clr-3

∆cre-1

∆cre-1

∆clr-3

CLR-1

Bound

CLR-2

Bound

CLR-3

RegStarvation

Wild

type

∆clr-3

∆os-1∆os

-2∆os

-1∆clr

-3

∆os-2

∆clr-3

∆cre-1

∆cre-1

∆clr-3

CLR-1

Bound

CLR-2

Bound

CLR-3

Reg

Avicel

Reg

F

Wild

type

∆clr-3∆os

-1

∆os-1

∆clr-3

∆os-4∆os

-5

∆os-4

∆clr-3

∆os-5

∆clr-3

∆os-2

∆clr-3

∆os-1

∆os-4

∆clr-3

∆asl-1

∆asl-1

∆clr-3

Wild typebackground

∆clr-3background

00.20.40.60.81.01.21.4

Rel

ativ

e C

MC

ase

activ

ity

Wild

type

∆os-2

00.20.40.60.81.01.21.4

Rel

ativ

e xy

lana

se a

ctiv

ity

Avicel Regulon clr-3 10h Regulon

∆os-2

CelluloseXylanStarvation

XylanStarvation

200

3

Rel

ativ

e cb

h-1

expr

essi

on

XylanStarvation

222

T171A

Y173A

os-2

∆clr-3

Pccg-1

A828V

-xlr-1

∆os-2

Pccg-1

A828V

-xlr-1

Fig. 3. The OS pathway regulates glycosyl hydrolase gene expression under starvation conditions. (A) The N. crassa OS pathway. The histidine kinase OS-1 isactivated by hyperosmolarity and activates the hyperosmotic MAP kinase cascade, which consists of the MAP kinase kinase kinase OS-4, the MAP kinase kinaseOS-5, and the MAP kinase OS-2. Activated OS-2 activates ASL-1 and other unknown transcription factors, which activate expression of osmoresponsive genes.(B) CMCase activity of enzymes secreted into culture supernatants by the indicated wild-type or mutant cultures 72 h posttransfer to the indicated medium.(C) Transcript abundance of clr-2 and cbh-1 relative to act 10 h posttransfer to the indicated medium as measured by qRT-PCR in the indicated wild-type ormutant cells. (D) Heat map of the expression level of genes in the Avicel regulon. (E) Heat map of the expression level of genes whose expression is more thanfourfold differentially regulated between wild-type and Δclr-3 cells 10 h posttransfer to starvation conditions. Orange bars indicate genes in the Avicelregulon. Purple bars indicate genes bound by CLR-1. Pink bars indicate genes bound by CLR-2. Green bars indicate genes in the clr-3 regulon. (F) Xylanaseactivity of enzymes secreted into culture supernatants 72 h posttransfer to the indicated medium by the indicated wild-type or mutant cultures relative to thexylanase activity secreted by wild-type cells transferred to VMM + xylan. Error bars indicate SDs. Circles indicate the values of individual biological replicates(Dataset S1).

E8668 | www.pnas.org/cgi/doi/10.1073/pnas.1707713114 Huberman et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

28,

202

0

Hog1 MAP kinase (T171 and Y173) to alanine (os-2T171A Y173A)(28, 29). Cells that expressed the kinase-dead os-2T171A Y173A

allele in combination with Δclr-3 showed secreted cellulase ac-tivity at levels comparable to Δos-2 Δclr-3 cells when exposed tocellulose, xylan, or starvation conditions (Fig. 3B).To ensure that this increase in secreted cellulase activity was

due to changes in the transcript abundance of genes involved inthe cellulolytic response rather than a translational or post-translational effect, we measured the expression of clr-2 andcbh-1 in Δos-2 Δclr-3 cells. Deletion of os-2 in combination withΔclr-3 had very little effect on clr-2 and cbh-1 transcript abundancewhen cells were exposed to xylan. However, under starvationconditions a threefold increase in clr-2 expression (P = 3 × 10−6)and a fourfold increase in cbh-1 expression (P = 4 × 10−7) wasobserved, indicating that the OSMAP kinase pathway acted at thetranscript level to regulate the cellulolytic response, particularlyunder starvation conditions (Fig. 3C).Previous studies showed that the histidine kinase OS-1 is at

least partially responsible for activating the OS MAP kinasecascade in response to hyperosmolarity (18, 30). We hypothe-sized that OS-1 would play a similar role in the regulation ofcellulase expression. However, deletion of os-1 in combinationwith clr-3 caused a threefold reduction in secreted cellulase ac-tivity under starvation conditions compared with Δclr-3 cells (P =7.6 × 10−5) (Fig. 3B). Deletion of os-1 had an even more dra-matic effect on cellulolytic gene transcript abundance. Transcriptlevels of clr-2 and cbh-1 were decreased by ninefold and 89-fold,respectively, in Δos-1 Δclr-3 cells under starvation conditionscompared with Δclr-3 cells (P < 6 × 10−5) (Fig. 3C). Additionally,when exposed to xylan, Δos-1 Δclr-3 cells showed a 21-fold de-crease in cbh-1 expression compared with Δclr-3 cells (P = 6 ×10−5) (Fig. 3C).Because the role of OS-1 appeared to oppose that of the OS

MAP kinase pathway with respect to cellulase regulation understarvation conditions, in contrast to its role in OS MAP kinasepathway activation under hyperosmotic conditions (23–25, 30), wehypothesized that the epistatic relationship between OS-1 and theOS MAP kinase cascade might also differ in these two conditions.To test this, we deleted os-1 and os-4 in combination with clr-3.The pattern of secreted cellulase activity of Δos-1 Δos-4 Δclr-3cells was indistinguishable from that of Δos-4 Δclr-3 cells, sug-gesting that, as during the response to hyperosmotic shock, thehistidine kinase OS-1 was upstream of the MAP kinase kinasekinase OS-4 during the response to starvation (Fig. 3B).In N. crassa, the basic leucine zipper transcription factor ASL-

1 regulates the expression of some OS-2–dependent genes; how-ever, deletion of asl-1 does not cause sensitivity to hyperosmoticshock, and not all OS-2–dependent genes are regulated by ASL-1,implying that the OS pathway also works through other, unknowntranscription factors to regulate gene expression (27). To test thepossibility that regulation of cellulase gene expression by the OSpathway was mediated by ASL-1, we measured the secreted cel-lulase activity of cells lacking asl-1. As with deletion of members ofthe OS pathway, deletion of asl-1 in an otherwise wild-type cellhad no effect on secreted cellulase activity under any of the con-ditions we tested (Fig. 3B). However, unlike deletion of membersof the OS pathway, cells lacking asl-1 and clr-3 showed a level ofsecreted cellulase activity that was indistinguishable from that ofΔclr-3 cells, indicating that the OS pathway does not act throughasl-1 to regulate cellulases (Fig. 3B).There are two possible mechanisms for the regulation of cel-

lulase gene expression by the OS pathway. One possibility is thatclr-3 and the OS pathway work in the same pathway to derepressCLR-1 activity. Alternatively, the OS pathway could play a rolemore similar to that of carbon catabolite repression and regulatecellulases through an independent pathway. To distinguish be-tween these two possibilities, we used RNA-seq to investigate therole of the OS pathway in responding to starvation, the nutrient

condition in which the OS pathway appeared to have the greatestrole in regulating cellulase expression. The OS pathway had thelargest phenotypic effect on cellulase production in cells lackingclr-3, so we first investigated the transcriptomes of Δos-1 Δclr-3and Δos-2 Δclr-3 cells. Consistent with secreted enzyme activity,of the 215 genes in the Avicel regulon, 104 were significantly(more than twofold) down-regulated in Δos-1 Δclr-3 cells, and69 were significantly (more than twofold) up-regulated in Δos-2Δclr-3 cells compared with Δclr-3 cells (Fig. 3D and Dataset S2).In contrast, only four genes in the Avicel regulon were signifi-cantly (more than twofold) up-regulated in Δos-1 Δclr-3 cells,and only 11 genes were significantly down-regulated in Δos-2Δclr-3 cells compared with Δclr-3 cells (Fig. 3D and Dataset S2).These data indicated that the OS pathway has a global effect oncellulase gene expression under starvation conditions.We hypothesized that genes regulated by os-1 and os-2 that

were not in the Avicel regulon might provide a clue as to whetherthe OS pathway regulated cellulase expression in the same path-way as clr-3 or through a separate pathway. Corroborating thesecreted cellulase activity, deletion of either os-1 or os-2 alone didnot have a significant effect on the expression of the vast majorityof genes in the Avicel regulon (Fig. 3D and Dataset S2). However,to our surprise, 55 genes that were significantly (more than four-fold) up-regulated inΔos-2 cells compared with wild-type cells (os-2 starvation regulon) were also differentially expressed in Δclr-3cells compared with wild-type cells after exposure to starvation for10 h (clr-3 10 h regulon). If clr-3 and os-2 were working throughthe same pathway, the expression level of these genes should besimilar in single clr-3 and os-2 mutants and in Δos-2 Δclr-3 cells.However, if clr-3 and os-2 were working through separate path-ways, it was likely that deletion of both clr-3 and os-2 would resultin an additive effect. Analysis of the transcriptomes of Δos-2 Δclr-3cells showed that nearly all the genes in both the os-2 starvationregulon and the clr-3 10 h regulon were significantly (more thantwofold) up-regulated compared with deletions of either os-2 orclr-3 alone, indicating that os-2 and clr-3 act in separate pathways(Fig. 3E and Dataset S3).One gene of particular interest that was up-regulated in both

Δclr-3 and Δos-2 cells compared with wild-type cells after 10 hexposure to starvation was the invertase gene (NCU04265),which is necessary for the consumption of sucrose, with a fivefoldincrease in transcript abundance in Δos-2 cells compared withwild-type cells during starvation (Dataset S3) (31). This up-regulation may indicate that os-2 has a general role in repres-sing genes involved in the consumption of nonpreferred carbonsources when conditions are not favorable to producing theseproteins. If the OS pathway is involved in the general regulationof genes necessary for the utilization of nonpreferred carbonsources as opposed to a specific role in cellulase regulation,deletion of os-2 should significantly affect the production ofother enzymes, such as xylanases. To test this hypothesis, wedeleted os-2 in cells that expressed a constitutively active versionof the transcription factor xlr-1 (NCU06971) (Pccg-1-xlr-1

A828V),which is responsible for activating the expression of xylanasegenes in the presence of xylan (5, 16). We observed a greaterthan twofold increase in xylanase secretion during starvation con-ditions in Pccg-1-xlr-1

A828V Δos-2 cells compared with Pccg-1-xlr-1A828V

cells (P = 1.4 × 10−11), indicating that the effect of the OS pathway-mediated response to starvation is not limited to cellulase generegulation (Fig. 3F).

The OS Pathway Exhibits Tunable Regulation of Cellulase Expressionin Response to Changes in Osmolarity. The OS pathway mediatesresponses to hyperosmolarity, oxidative stress, and heat stress byregulating the expression of an overlapping set of genes (22, 24,32). It is possible that the OS pathway mediates a general stressresponse to a variety of stresses, including the nutrient stressexperienced by cells exposed to starvation and poor carbon

Huberman et al. PNAS | Published online September 25, 2017 | E8669

GEN

ETICS

PNASPL

US

Dow

nloa

ded

by g

uest

on

Oct

ober

28,

202

0

sources, whose physiological effect on the cell is mitigated by theregulation of a core set of stress-response genes. An alternativehypothesis is that the OS pathway-mediated response to starva-tion is a hypoosmotic response to the low osmolarity that resultsfrom a lack of soluble sugar in the environment.To distinguish between these two hypotheses, we first analyzed

the expression of gcy-1 (NCU04923), a gene involved in theproduction of glycerol in response to hyperosmotic shock, usingqRT-PCR under hyperosmotic and starvation conditions (32).Corroborating previous results, we saw a ninefold OS-2–de-pendent increase in gcy-1 transcript abundance after a 2-h ex-posure to hyperosmotic shock caused by the addition of 1 MNaCl compared with osmonormal conditions; this increased ex-pression was sustained, albeit at a reduced level, for up to 10 h(P < 10−4) (Fig. 4A). However, there was no increase in gcy-1transcript abundance after exposure to starvation compared withosmonormal conditions with sucrose as the sole carbon source(Fig. 4A). To confirm that gcy-1 expression was dependent on theosmolarity of the environment even in the absence of a carbonsource, we increased the solute concentration in medium lackinga carbon source from 16 mM to 1,288 mM. The transcriptabundance of gcy-1 in wild-type cells was dependent on osmo-

larity and increased to levels comparable to those observed inmedium containing sucrose + 1 M NaCl (P = 2 × 10−4, df = 1,F = 33, one-way ANOVA) (Fig. S4A). These data indicated thatthe response to starvation does not activate processes, such asglycerol production, that are instrumental in the OS pathway-dependent response to hyperosmotic shock.Since the OS pathway was not activating a general stress re-

sponse in response to starvation, we hypothesized that the OSpathway regulates cellulase expression in response to changes inosmolarity. We tested this hypothesis in two complementary ways:by observing the effect of low osmolarity on cells whose celluloseresponse is derepressed and by increasing the osmolarity of theenvironment of cells with a strongly induced cellulose response tolevels at which gcy-1 expression was significantly activated.Assuming complete dissociation of all salts, we calculated the

osmolarity of the solutes in Vogel’s salts, the N. crassa growthmedium (17), to be 161 mM. If cellulase production is regulatedin response to osmolarity, reducing the concentration of Vogel’ssalts should result in increased cellulase activity. Since we wantedto distinguish the effects resulting from changing osmolarity oncellulase expression from those due to direct activation of thecellulose response by external carbon sources, we measured thesecreted cellulase activity of Δclr-3, Δos-1 Δclr-3, and Δos-2Δclr-3 cells in medium lacking a carbon source. We reduced theconcentration of Vogel’s salts so that the resulting osmolaritywas 16 mM and then increased the osmolarity with NaCl. As weincreased the osmolarity of the medium from 16 mM to 161 mM,we saw a significant decrease in the secreted cellulase activity ofΔclr-3 cells (P = 6 × 10−6, df = 1, F = 31, one-way ANOVA) (Fig.4B). This result was significantly different from the pattern ofsecreted cellulase activity of either Δos-1 Δclr-3 or Δos-2 Δclr-3cells (P < 0.01, df = 4, F > 3.7, two-way ANOVA) (Fig. 4B). Tocontrol for potential effects caused by the addition of NaCl, wealso changed the osmolarity of the medium by altering theconcentration of Vogel’s salts and observed a similar effect onΔclr-3 cells (P = 0.005, df = 1, F = 17, one-way ANOVA) (Fig.S4B). The cellulase activity of Δos-1 Δclr-3 cells was consistentlylow and that of Δos-2 Δclr-3 cells was consistently high across thevarious Vogel’s concentrations (P < 0.02, df = 2, F > 2.4, two-way ANOVA) (Fig. S4B). To confirm that this effect was due toregulation of cellulase gene expression by the OS pathway inΔclr-3 cells, we measured expression of clr-2 and cbh-1 as weincreased the Vogel’s concentration from 16 mM to 161 mM andsaw a significant decrease in the transcript abundance of bothgenes (P < 8 × 10−5, df = 1, F > 68, one-way ANOVA) (Fig.S4C). This effect was not observed in Δos-2 Δclr-3 cells (P <0.004, df = 2, F > 9, two-way ANOVA) (Fig. S4C). These dataindicated that cellulase expression in cells with a derepressedcellulose response is regulated by the osmolarity of the envi-ronment in an OS pathway-dependent fashion.If osmolarity is serving as a proxy for the amount of soluble

sugar in the environment, we would expect to see a similar effect inwild-type cells under inductive conditions as osmolarity is in-creased with NaCl. As predicted, we observed a significant de-crease in the secreted cellulase activity of wild-type cells exposed tocellulose as we increased the salt concentration, culminating ina complete abolition of cellulase activity at 1 M NaCl (P = 8.0 ×10−28, df = 1, F = 668, one-way ANOVA) (Fig. 4C). To control forpossible effects caused by high concentrations of NaCl, we alsotested the secreted cellulase activity of cells lacking the threemajor extracellular β-glucosidases (ΔΔΔβ-glucosidase) (NCU00130,NCU04952, and NCU08755), which allowed the induction of cel-lulase gene expression when these cells were exposed to cellobiose(14). Consistent with results of cultures with added NaCl, the se-creted cellulase activity decreased in ΔΔΔβ-glucosidase cells as theconcentration of the nonionic osmolyte cellobiose increased, an ef-fect not due solely to carbon catabolite repression (P < 2.4 × 10−9,df = 1, F > 64, one-way ANOVA) (Fig. S4D).

2h 2h10h 10hWild type ∆os-2

81216

40

20

Rel

ativ

e gc

y-1

expr

essi

on SucroseSucrose+1 M NaClStarvation

A B∆os-2∆clr-3∆clr-3 ∆os-1

∆clr-3

00.20.40.60.81.01.21.41.61.8

0 40 80 120 160Rel

ativ

e C

MC

ase

activ

ity

[Solute] (mM)

0 20 40 60 80 1000

0.20.40.60.81.01.21.41.6

Rel

ativ

e C

MC

ase

activ

ity

Hours post-inoculation

∆os-2Wildtype ∆os-1

0 0.2 0.4 0.6 0.8 1.00

0.20.40.60.81.01.2

Rel

ativ

e C

MC

ase

activ

ity

[NaCl] (M)

Wildtype ∆os-1

C D

Fig. 4. The OS pathway regulates cellulase gene expression in response tochanges in osmolarity. (A) Transcript abundance of gcy-1 relative to act at 2 hand 10 h posttransfer to the indicated medium asmeasured by qRT-PCR in wild-type or Δos-2 cells. Transcript abundance of gcy-1 is shown relative to thetranscript abundance of gcy-1 in wild-type cells exposed to sucrose 2 h post-transfer. (B) CMCase activity of enzymes secreted into culture supernatants bythe indicated mutant cells 72 h posttransfer to 0.1× Vogel’s salts lacking acarbon source with varying concentrations of NaCl added as needed to reachthe indicated solute concentration. Solute concentration indicates the totalconcentration of dissolved solutes (16mM from Vogel’s salts and any additionalsolutes from NaCl). CMCase activity is shown relative to that present in theculture supernatant of Δclr-3 cells exposed to 16mM solute. (C) CMCase activityof enzymes secreted into culture supernatants 72 h posttransfer to VMM +cellulose with the indicated concentration of NaCl by wild-type and Δos-1cultures. (D) CMCase activity of enzymes secreted into culture supernatants atthe indicated time after the inoculation of equal concentrations of conidia intoVMM + cellulose. CMCase activity is shown relative to that present in the cul-ture supernatant of wild-type cells 96 h postinoculation into VMM + cellulose.Error bars indicate SDs. Circles indicate the values of individual biological rep-licates (Dataset S1). Horizontal spread of individual data points surrounding anaverage is shown to aid in viewing the y values of similar replicates. It does notindicate any differences in the x values of these data points.

E8670 | www.pnas.org/cgi/doi/10.1073/pnas.1707713114 Huberman et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

28,

202

0

Because cells lacking os-2 are extremely sensitive to increasesin osmolarity, it was not possible to measure cellulase activity ofΔos-2 cells when exposed to high concentrations of NaCl (22).However, deletion of os-1 caused a significant insensitivity toNaCl concentration with respect to secreted cellulase activity,resulting in significantly higher secreted cellulase activity athigher concentrations of NaCl compared with wild-type cells(P = 1.5 × 10−14, df = 4, F = 28, two-way ANOVA) (Fig. 4C). Asimilar dependence on OS-1 was seen in ΔΔΔβ-glucosidase Δos-1 cells exposed to increasing concentrations of cellobiose (P =0.045, df = 3, F = 2.8, two-way ANOVA) (Fig. S4D). These dataindicated that cells regulate cellulase expression in response toosmolarity in an OS pathway-dependent fashion.If cells are using osmolarity to regulate the expression of plant

cell wall-degrading enzymes, it is possible that the effects of theOS pathway would be particularly noticeable after direct in-oculation into a medium where the only carbon source is crys-talline and thus does not affect the osmolarity of the medium.We tested this hypothesis by directly inoculating wild-type, Δos-1, and Δos-2 cells into medium containing cellulose as the solecarbon source and measuring secreted cellulase activity. De-letion of os-1 resulted in significantly decreased cellulase activitycompared with wild-type cultures (P = 0.03, df = 4, F = 3, two-way ANOVA), while deletion of os-2 resulted in significantlyincreased cellulase activity compared with wild-type cultures(P = 10−5, df = 4, F = 8, two-way ANOVA) (Fig. 4D). Theseresults indicated that the OS pathway plays a role in regulatingglycosyl hydrolase expression immediately after colonization of acrystalline carbon source, such as cellulose.

The OS Pathway-Mediated Regulation of Glycosyl Hydrolases Does NotAct Through the Canonical Carbon Catabolite-Repression Pathway.One of the genes in the os-2 starvation regulon was cre-1, whichencodes a transcription factor that is a regulator of carbon ca-tabolite repression (Fig. 5A and Dataset S3) (13). Deletion of cre-1increases activation of the cellulolytic response when preferredcarbon sources, such as sugars released during the breakdown ofcellulose, are present (11). Because the expression of cre-1appeared to be os-2 dependent, and CRE-1 represses cellulasegene expression, we asked whether the OS pathway workedthrough CRE-1 to repress the cellulolytic response.To test whether CRE-1 is downstream of the OS pathway

during cellulase regulation, we deleted cre-1 in combination withclr-3 and investigated whether the pattern of cellulase expressionmatched that of cells lacking clr-3 in combination with a memberof the OS pathway. When exposed to xylan, cells lacking bothcre-1 and clr-3 exhibited a twofold increase in cellulase activitycompared with cells lacking clr-3 (P = 1.7 × 10−19) (Fig. 5B).Corroborating the cellulase activity data, we observed a 13-foldincrease in clr-2 transcript abundance and a 28-fold increase incbh-1 transcript abundance in Δcre-1 Δclr-3 cells compared withΔclr-3 cells (P < 4.8 × 10−5) (Fig. 5C). However, during starva-tion, secreted cellulase activity of Δcre-1 Δclr-3 cells was notsignificantly different from that of Δclr-3 cells (Fig. 5B). Indeed,the transcript abundances of clr-2 and cbh-1 actually decreasedby fourfold and eightfold, respectively, in Δcre-1 Δclr-3 cells ascompared with Δclr-3 cells (P < 10−3) (Fig. 5C).The differences in cellulase expression in cells lacking clr-3 in

combination with cre-1 as opposed to members of the OSpathway led us to wonder whether cre-1 and the OS genes wereworking through separate pathways to regulate cellulase geneexpression. If true, we predicted that a strain carrying cre-1, os-5,and clr-3 mutations would show an additive effect for cellulaseactivity. As hypothesized, a Δcre-1 Δos-5 Δclr-3 strain showed agreater than 65% increase in secreted cellulase activity whenexposed to xylan compared with either Δcre-1 Δclr-3 cells or Δos-5 Δclr-3 cells (P < 5.7 × 10−7) (Fig. 5B). These data supported

the hypothesis that CRE-1 and the OS pathway function inseparate pathways to regulate cellulase gene expression.To investigate this hypothesis further, we performed a global

analysis of gene expression in Δcre-1 Δclr-3 compared with Δclr-3cells and found that nearly 80% of the genes in both the clr-3 andAvicel regulons were significantly (more than twofold) down-regulated in Δcre-1 Δclr-3 cells compared with Δclr-3 cells un-der starvation conditions (Fig. 3D and Dataset S2). Although thedown-regulation of genes in the Avicel regulon in Δcre-1 Δclr-3cells was reminiscent of cells lacking both os-1 and clr-3, furtheranalysis of the RNA-seq data showed that, unlike in the tran-scriptome of Δos-1 Δclr-3 cells, the expression of 49 genes in theclr-3 10 h regulon, but not in the Avicel regulon, was increased byat least twofold in Δcre-1 Δclr-3 cells compared with Δclr-3 cells(Fig. 3E and Dataset S3). This pattern of gene expression wassubstantially different from that in cells lacking clr-3 and eitheros-1 or os-2, suggesting that cre-1 is not downstream of the OSpathway in the starvation response.Our data indicated that the OS pathway regulates cellulase

expression due to changes in solute concentration, while thetranscription factor cre-1 regulates cellulase expression in responseto the presence of preferred carbon sources (11). To confirm thatthe OS pathway does not regulate cellulase expression in responseto preferred carbon sources, we investigated the response ofN. crassa to the nonhydrolysable glucose analog 2-deoxyglucose.When wild-type cells are exposed to medium containing 2-deox-yglucose in combination with a nonpreferred carbon source,such as cellulose, cells are unable to grow because the presence of2-deoxyglucose represses the cellulolytic response (33). We in-oculated cells carrying deletions of cre-1, clr-3, os-1, and os-2 alone

CRE-1

PreferredCarbon Source

Acellulases

xylanases

C

B Cellulose Xylan Starvation

Rel

ativ

e C

MC

ase

activ

ity

00.20.40.60.81.01.21.4

Wild

type

∆clr-3

∆cre-1

∆cre-1

∆clr-3

∆os-5

∆clr-3

∆cre-1

∆os-5

∆clr-3

∆os-1

∆clr-3

Wildtype ∆cre-1

∆os-1 ∆os-2

∆clr-3 ∆clr-3∆cre-1

∆clr-3∆os-1

∆clr-3∆os-2

Cellulose+2-deoxyglucose

Wildtype ∆cre-1

∆os-1 ∆os-2

∆clr-3 ∆clr-3∆cre-1

∆clr-3∆os-1

∆clr-3∆os-2

CelluloseD

Xylan Starvation

020406080

100120140160180

Wild

type∆clr

-3

∆cre-1

∆clr-3

∆os-1

∆clr-3

∆os-2

∆clr-3R

elat

ive

clr-2

exp

resi

on

02,0004,0006,0008,000

10,00012,000

Rel

ativ

e cb

h-1

expr

essi

on

Wild

type∆clr

-3

∆cre-1

∆clr-3

∆os-1

∆clr-3

∆os-2

∆clr-3

Fig. 5. The OS pathway does not act through the transcription factor CRE-1.(A) Preferred carbon sources activate the transcription factor CRE-1, which re-presses the expression of genes involved in the utilization of nonpreferredcarbon sources. (B) CMCase activity of enzymes secreted by the indicated wild-type or mutant cultures into culture supernatants 72 h posttransfer to the in-dicated medium. (C) Transcript abundance of clr-2 and cbh-1 relative to act 10 hposttransfer to the indicated medium as measured by qRT-PCR in the indicatedwild-type or mutant cells. Error bars indicate SDs. Circles indicate the valuesof individual biological replicates (Dataset S1). (D) Pictures of the indicatedwild-type and mutant cells inoculated into 3 mL of liquid VMM + cellulose orVMM + cellulose + 0.2% 2-deoxyglucose in round-bottomed 24-well plates andallowed to grow for 8 d in constant light with constant shaking. Orange cellmasses are live N. crassa. White circles at the bottoms of wells are unconsumedcellulose. This is a representative image of three biological replicates.

Huberman et al. PNAS | Published online September 25, 2017 | E8671

GEN

ETICS

PNASPL

US

Dow

nloa

ded

by g

uest

on

Oct

ober

28,

202

0

and in combination in medium containing either cellulose orcellulose with 2-deoxyglucose. All of these cells were able to growin medium containing cellulose; however, the additional presenceof 2-deoxyglucose prohibited the growth of cells lacking os-1, os-2,and/or clr-3 but not cre-1, suggesting that the OS pathway does notregulate cellulase gene expression in response to the chemicalsignature of preferred carbon sources (Fig. 5D).

DiscussionRegulation of gene expression in response to nutrient availabilityrequires cells to integrate signals from a variety of signalingpathways that query the state of the extracellular environment.The wide variety of carbohydrates utilized by saprophytic fila-mentous fungi makes these organisms an excellent model systemto study nutrient-sensing networks. Using a forward geneticscreen, we identified members of signaling pathways involved inboth the direct activation of cellulase gene expression in re-sponse to cellulose and in general regulation of genes involved inthe utilization of nonpreferred carbon sources. Our study revealedthat CLR-3, a cellulase repressor, regulates the activity of thetranscription factor CLR-1 in response to cellulose and that theOS pathway is involved in the regulation of genes required forthe consumption of nonpreferred carbon sources. These path-ways work together with other highly conserved pathways, suchas the carbon catabolite-repression pathway, to form a nutrient-sensing network that tightly regulates the expression of genesrequired for plant biomass utilization (Fig. 6).

clr-3 Encodes a Repressor of CLR-1 Activity. Induction of cellulasegene expression is accomplished by the transcription factors CLR-1 and CLR-2. Our data indicate that in the absence of celluloseCLR-3 represses the activity of CLR-1. When cellulose is present,the repressive activity of CLR-3 is relieved, and CLR-1 activatesthe expression of a small number of cellulases and other genesnecessary for the utilization of cellulose, including the transcrip-tion factor clr-2 (Fig. 6). CLR-2 is responsible for activating theexpression of the majority of cellulase genes (4, 15). Our data

indicate that CLR-3 acts to repress the activity of the transcriptionfactor CLR-1 in the absence of an inducer. Although homologs ofclr-1 occur throughout ascomycete species, the role of CLR-1 asthe regulator of the cellulolytic response does not appear to bebroadly conserved (3). Thus, we hypothesize that functionalorthologs of clr-3 may exist only in species in which CLR-1 regu-lates the cellulolytic response, consistent with the conservation ofCLR-3 among sordariomycete species (Fig. S2D).CLR-3 is a DUF1479-containing protein. Although proteins

containing this domain exist in both prokaryotes and eukaryotes,their function is unknown. A recent study in Candida albicansshowed that a gene encoding a DUF1479-containing protein, Gig2,is up-regulated in response to the sugar N-acetylglucosamine;Gig2 may bind the related sugar N-glycolylneuraminate or N-ace-tylneuraminate (34). If Gig2 and CLR-3 function similarly, thisobservation may indicate a role for CLR-3 in binding a sugar mol-ecule, such as cellobiose, or a derivative, thus, inactivating the re-pressive function of CLR-3. Identifying potential metabolites boundto CLR-3 may help clarify the soluble signaling molecule re-sponsible for directly activating the cellulolytic response in N. crassa.

The OS Pathway Regulates the Expression of Genes Involved in theUtilization of Nonpreferred Carbon Sources in Response to Changes inOsmolarity. The N. crassa OS pathway was initially identified forits role in responding to hyperosmotic shock but also has a moregeneral role in cellular stress, including responses to heat stress,oxidative stress, circadian rhythm, and simultaneous exposure toheat stress and starvation (21, 24, 32, 35, 36).We identified a role for the OS pathway in the general regu-

lation of genes involved in the utilization of nonpreferred carbonsources in response to changes in osmolarity. Deletion of membersof the OS MAP kinase cascade results in increased expression ofplant cell wall-degrading enzymes under noninducing conditions incells with derepressed (hemi)cellulolytic responses. In Trichodermareesei, cells lacking the orthologous os-2 MAP kinase (tmk3) showdecreased cellulase production after direct inoculation into liq-uid culture. However, the Δtmk3 mutant shows severe pleiotro-pic growth defects, and Δtmk3 cells grown on solid medium donot show the same reduction in cellulase production (37). TheT. reesei hypercellulolytic mutant RUT-C30 shows defects inosmotic homeostasis (38). Our work suggests that this osmoticdefect may be partially responsible for the increased cellulaseproduction of RUT-C30.Several additional examples of a role for OS MAP kinase

pathways in nutrient sensing exist. In mammalian neuronal cells,the p38 MAP kinase (orthologous to os-2) has been shown toregulate O-linked β-N-acetylglucosamine protein modifications,particularly in response to glucose deprivation (39). In S. cerevisiaeand C. albicans, the p38 MAP kinase ortholog Hog1 functionswith the nutrient-sensing Kss1 MAP kinase to regulate morpho-logical responses to the nonpreferred carbon source, galactose(40). However, unlike the N. crassa OS pathway-mediated re-sponse to starvation, the S. cerevisiae Hog1-mediated response tothe nonpreferred sugar galactose activates the expression of genesrequired for the response to hyperosmotic shock, including genesin the glycerol-production pathway (40).To our surprise, the upstream histidine kinase OS-1 has an

opposing role in the regulation of genes involved in the con-sumption of nonpreferred carbon sources in N. crassa. Understarvation conditions, cells lacking os-1 and clr-3 show decreasedcellulase gene expression compared with Δclr-3 cells. The role ofOS-1 as a repressor of the OS MAP kinase cascade is consistentwith the role of a homologous protein in S. cerevisiae, Sln1, ahistidine kinase partially responsible for regulating the activation ofHog1 (41, 42). However, previous studies in N. crassa indicate thatOS-1 acts as an activator of the OS MAP kinase cascade duringexposure to hyperosmotic shock (18, 30). Additionally, OS-1 hasbeen hypothesized to act as an osmosensor, perhaps via the HAMP

CLR-1 CLR-1

cellulose

CLR-3

PreferredCarbon Source

CRE-1

Starvation

OsmoresponsiveGene

ExpressionASL-1

+unknown TFs

Starvation(Hypoosmolarity)

OS-1

Free Sugar(Hyperosmolarity)

TF?

OS-5OS-2 OS-4

cellulases

clr-2

Fig. 6. Model: Cellulolytic genes are regulated by carbon source and os-molarity. CLR-3 represses the activity of CLR-1 in the absence of cellulose,and the action of CLR-3 is, in turn, repressed by the presence of cellulose. TheOS pathway regulates the lignocellulolytic response by sensing the changesin osmolarity that occur due to the presence or absence of sugars. Thepresence of substantial free sugar or hyperosmotic conditions activates theOS MAP kinase pathway, activating ASL-1 and other, unknown transcriptionfactors (TF) to activate osmoresponsive gene expression and repress glycosylhydrolase production. Starvation or hypoosmotic conditions cause OS-1 torepress the OS MAP kinase pathway, allowing increased glycosyl hydrolaseexpression in conditions in which free sugar is in short supply. Carbon ca-tabolite repression acts through the transcription factor CRE-1 to repress theexpression of cellulases in the presence of preferred carbon sources andactivate the expression of cellulases in the absence of a carbon source.

E8672 | www.pnas.org/cgi/doi/10.1073/pnas.1707713114 Huberman et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

28,

202

0

domains found in group III hybrid histidine kinases (26, 43). Thediffering roles for OS-1 in regulating cellulase expression in re-sponse to hyperosmotic shock versus starvation led us to hypoth-esize that the OS pathway could be responding to the changes inosmolarity caused by the presence or absence of soluble sugars toregulate the expression of genes involved in the consumption ofnonpreferred carbon sources (Fig. 6).Although we identified a role for the OS pathway in regulating

the expression of cellulase genes in response to changes in os-molarity, there are still several unanswered questions about howthis is accomplished. ASL-1 is the only transcription factor knownto be downstream of the OS pathway in N. crassa. However, ourdata taken together with the observation that Δasl-1 cells are notsensitive to hyperosmolarity (27) indicate that other transcriptionfactor(s) are likely involved in regulating the transcriptional re-sponse to changes in osmolarity. It will be the role of future studiesto identify which transcription factor(s) or other downstreamtargets are activated by the OS pathway, whether any other genesare involved in mediating the phosphorylation of the OS pathway,and how that affects the regulation of genes involved in the con-sumption of nonpreferred carbon sources.

A Network of Nutrient-Sensing Pathways Regulates Cellulase GeneExpression. Our data support a model in which signals transmittedthrough multiple signaling pathways are integrated by the cell toregulate the expression of the (hemi)cellulolytic response. Genesencoding enzymes involved in plant biomass deconstruction mustbe both released from repression by transcription factors regulatedby the presence of preferred carbon sources and activated bytranscription factors that are regulated specifically by componentsof plant biomass, such as cellulose (Fig. 6).We hypothesize that the OS pathway uses the osmolarity of the

environment as a proxy for the concentration of available solublesugar to regulate the expression of genes required for the utiliza-tion of insoluble or otherwise nonpreferred carbon sources (Fig. 6).When the histidine kinase OS-1 senses an increase in the osmo-larity of the environment, it activates the OS MAP kinase cascade,which in turn recruits ASL-1 and other, unknown transcriptionfactors to activate osmoresponsive gene expression to increase theinternal osmolarity of the cell so that it matches the external os-molarity more closely (18, 32, 44). Our data indicate that this OSpathway-mediated response also recruits an as-yet unknown com-ponent(s) to repress the expression of genes, such as cellulases, thatare involved in the consumption of nonpreferred carbon sources.We hypothesize that, during starvation or initial inoculation ontoan insoluble carbon source when no free sugar is available, OS-1 represses the activity of the OS MAP kinase cascade. This, inturn, releases the transcriptional inhibition of cellulase genes.The OS pathway plays an important role in cellulase regulation.

However, other pathways, such as the carbon catabolite-repressionpathway, which signals through the transcription factor cre-1, alsoplay a critical role in regulating cellulase expression (11). Unlikethe OS pathway, the carbon catabolite-repression pathway dis-tinguishes between the types of carbohydrates available to regulategene expression (3). This may explain why cre-1 had a significantrole in repressing cellulase gene expression when cells were ex-posed to the plant cell wall component xylan, while deleting eitheros-1 or os-2 in combination with clr-3 had very little effect oncellulase gene expression when xylan is the sole carbon source.Our data indicate that the expression of cellulase genes in

N. crassa is regulated by at least three distinct pathways: CLR-mediated cellulase induction, the canonical carbon catabolite-repression pathway, which includes CRE-1, and the OS pathway.Further studies into the mechanisms by which these three pathwaysinteract with each other and with other nutrient-sensing pathways toregulate cellulase gene expression could provide clues about howother eukaryotic nutrient-sensing networks interact. This approachcould also lead to the identification of novel targets for battling

infection by pathogenic fungi, since accurate nutrient sensingis required for the progression of mammalian disease causedby Aspergillus fumigatus infections (45). Additionally, studieson nutrient sensing and cellulase production in saprophyticfungi have implications for increasing the production of plantcell wall-degrading enzymes, which can be harvested as part ofa pipeline for the generation of second-generation biofuels andother high-value products.

Materials and MethodsN. crassa Strains and Culturing. The strains used in this study are listed in TableS1. All strains were derived from the wild-type reference strain Fungal Ge-netics Stock Center (FGSC) 2489 using standard genetic techniques and wereconfirmed by PCR and DNA sequencing (46, 47). The only exceptions are theΔasl-1 strains. Details of their construction and the construction of hetero-karyons can be found in SI Materials and Methods.

N. crassa cultures were grown on Vogel’s minimal medium (VMM) withcarbon sources added at 2% (wt/vol) unless otherwise noted (17). Specifics ofthe carbon sources used can be found in SI Materials and Methods. Unlessotherwise noted, cells were grown from freezer stocks on VMM + sucrose +1.5% agar (Fisher Scientific) slants for 2 d at 30 °C in the dark and 4 d at25 °C in constant light before inoculation into the indicated medium at1 × 106 conidia/mL for all experiments.

Forward Genetic Screen. Ppmo-4-HygR conidia were harvested, washed, and

resuspended in 25 mL PBS with NTG (TCI America) for 2 h before plating onVMM (without calcium) + 0.5% xylan + 0.11 M Tris (pH 8) + 262 μg/mLhygromycin B (Invitrogen) + 1.5% agar. Details of the mutagenesis procedurecan be found in SI Materials and Methods. Hygromycin-resistant colonies weretransferred to slants containing VMM + 0.5% xylan + 200 μg/mL hygromycin B.To ensure that hygromycin resistance was due to inappropriate activation ofthe cellulolytic response, the cellulase activity of mutants grown on VMM +xylan was tested. Mutants that showed cellulase activity in the absence of aninducer were subjected to bulked segregant analysis.

DNA Sequencing and Bulked Segregant Analysis. Mutants that showed cellulaseactivity in the absence of an inducer were backcrossed to the parentalPpmo-4-Hyg

R strain, and ascospore progeny were divided into mutant-like andparental-like pools. Genomic DNA isolation and bulked segregant analysiswere modified from Heller et al. (48). For details see SI Materials and Methods.

Enzyme-Activity and Gene-Expression Assays. Enzyme-activity assays and as-says for gene expression were modified from Coradetti et al. (4). For detailssee SI Materials and Methods.

RNA Sequencing and Transcript Abundance. Libraries were prepared from totalRNA using standard Illumina protocols. RNA was sequenced on either anIllumina HiSeq. 2000 or 4000 system and analyzed using HiSat2 v2.0.5 (49) andCufflinks v2.2.1 (50). Details can be found in SI Materials and Methods.

Statistical Significance Tests. For RNA-seq data, experiments were done inbiological triplicate, and statistical significance was determined usingCufflinks v2.2.1 (50). For all qRT-PCR and secreted enzyme activity exper-iments at least three biological replicates were done. Biological replicateswere independent cultures inoculated on either the same or independentdays. The exact numbers for all qRT-PCR and secreted enzyme activityexperiments are shown in Dataset S1.

Statistical significance was determined using a two-tailed homoscedastic(equal variance) Student’s t test unless otherwise noted. Values of bars andlines in bar and line graphs, respectively are the mean of the biologicalreplicates, and error bars in all figures are SDs. Circles on bar and line graphsindicate the values of individual replicates.

Unless otherwise noted, carboxymethyl cellulase (CMCase) activity isshown relative to the CMCase activity present in the culture supernatant ofwild-type cells 72 h posttransfer to VMM + cellulose, and clr-2 and cbh-1expression is shown relative to the expression of these genes in wild-typecells exposed to VMM + xylan.

Data Availability. RNA-seq data used in this study were deposited in the GeneExpression Omnibus at the National Center for Biotechnology Informationand are accessible through GEO series accession no. GSE95681. The numericalvalues used to generate all qRT-PCR and secreted-enzyme activity graphs areshown in Dataset S1.

Huberman et al. PNAS | Published online September 25, 2017 | E8673

GEN

ETICS

PNASPL

US

Dow

nloa

ded

by g

uest

on

Oct

ober

28,

202

0

ACKNOWLEDGMENTS. We thank the Fungal Genetics Stock Center forstrains. This work was supported by an Energy Biosciences Institute Grant (toN.L.G.), NIH National Research Service Award (NRSA) Individual PostdoctoralFellowship 5 F32 GM113372 (to L.B.H.), and NIH NRSA Trainee Grant 2 T32

GM7127-36 A1 (to S.T.C.). N.L.G. was partially supported by the LaboratoryDirected Research and Development Program of Lawrence Berkeley Na-tional Laboratory under US Department of Energy Contract DE-AC02-05CH11231.

1. Lindsley JE, Rutter J (2004) Nutrient sensing and metabolic decisions. Comp BiochemPhysiol B Biochem Mol Biol 139:543–559.

2. Shaw RJ (2006) Glucose metabolism and cancer. Curr Opin Cell Biol 18:598–608.3. Huberman LB, Liu J, Qin L, Glass NL (2016) Regulation of the lignocellulolytic response

in filamentous fungi. Fungal Biol Rev 30:101–111.4. Coradetti ST, et al. (2012) Conserved and essential transcription factors for cellulase

gene expression in ascomycete fungi. Proc Natl Acad Sci USA 109:7397–7402.5. Sun J, Tian C, Diamond S, Glass NL (2012) Deciphering transcriptional regulatory

mechanisms associated with hemicellulose degradation in Neurospora crassa.Eukaryot Cell 11:482–493.

6. van Peij NN, Visser J, de Graaff LH (1998) Isolation and analysis of xlnR, encoding atranscriptional activator co-ordinating xylanolytic expression in Aspergillus niger. MolMicrobiol 27:131–142.

7. New AM, et al. (2014) Different levels of catabolite repression optimize growth instable and variable environments. PLoS Biol 12:e1001764.

8. Orejas M, MacCabe AP, Pérez González JA, Kumar S, Ramón D (1999) Carbon ca-tabolite repression of the Aspergillus nidulans xlnA gene. Mol Microbiol 31:177–184.

9. Orejas M, MacCabe AP, Pérez-González JA, Kumar S, Ramón D (2001) The wide-domain carbon catabolite repressor CreA indirectly controls expression of theAspergillus nidulans xlnB gene, encoding the acidic endo-beta-(1,4)-xylanase X(24).J Bacteriol 183:1517–1523.

10. Piñaga F, Fernández-Espinar MT, Vallés S, Ramón D (1994) Xylanase production inAspergillus nidulans: Induction and carbon catabolite repression. FEMS Microbiol Lett115:319–323.

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

12. Nehlin JO, Ronne H (1990) Yeast MIG1 repressor is related to the mammalian earlygrowth response and Wilms’ tumour finger proteins. EMBO J 9:2891–2898.

13. de la Serna I, Ng D, Tyler BM (1999) Carbon regulation of ribosomal genes inNeurospora crassa occurs by a mechanism which does not require Cre-1, the homologueof the Aspergillus carbon catabolite repressor, CreA. Fungal Genet Biol 26:253–269.

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

15. Coradetti ST, Xiong Y, Glass NL (2013) Analysis of a conserved cellulase transcriptionalregulator reveals inducer-independent production of cellulolytic enzymes in Neu-rospora crassa. Microbiologyopen 2:595–609.

16. Craig JP, Coradetti ST, Starr TL, Glass NL (2015) Direct target network of the Neu-rospora crassa plant cell wall deconstruction regulators CLR-1, CLR-2, and XLR-1.MBio6:e01452-e15.

17. Vogel H (1956) A convenient growth medium for Neurospora (medium N). MicrobialGenetics Bulletin 13:42–43.

18. Ochiai N, et al. (2001) Characterization of mutations in the two-component histidinekinase gene that confer fludioxonil resistance and osmotic sensitivity in the os-1mutants of Neurospora crassa. Pest Manag Sci 57:437–442.

19. Stewart V, Chen LL (2010) The S helix mediates signal transmission as a HAMP domaincoiled-coil extension in the NarX nitrate sensor from Escherichia coli K-12. J Bacteriol192:734–745.

20. Hérivaux A, et al. (2016) Major sensing proteins in pathogenic fungi: The hybridhistidine kinase family. PLoS Pathog 12:e1005683.

21. Mays LL (1969) Isolation, characterization, and genetic analysis of osmotic mutants ofNeurospora crassa. Genetics 63:781–794.

22. Zhang Y, Lamm R, Pillonel C, Lam S, Xu JR (2002) Osmoregulation and fungicide re-sistance: The Neurospora crassa os-2 gene encodes a HOG1mitogen-activated proteinkinase homologue. Appl Environ Microbiol 68:532–538.

23. Jones CA, Greer-Phillips SE, Borkovich KA (2007) The response regulator RRG-1 func-tions upstream of a mitogen-activated protein kinase pathway impacting asexualdevelopment, female fertility, osmotic stress, and fungicide resistance in Neurosporacrassa. Mol Biol Cell 18:2123–2136.

24. Banno S, et al. (2007) Roles of putative His-to-Asp signaling modules HPT-1 and RRG-2, on viability and sensitivity to osmotic and oxidative stresses in Neurospora crassa.Curr Genet 51:197–208.

25. Fujimura M, et al. (2003) Putative homologs of SSK22 MAPKK kinase and PBS2 MAPKkinase of Saccharomyces cerevisiae encoded by os-4 and os-5 genes for osmoticsensitivity and fungicide resistance in Neurospora crassa. Biosci Biotechnol Biochem67:186–191.

26. Schumacher MM, Enderlin CS, Selitrennikoff CP (1997) The osmotic-1 locus of Neu-rospora crassa encodes a putative histidine kinase similar to osmosensors of bacteriaand yeast. Curr Microbiol 34:340–347.

27. Yamashita K, et al. (2008) ATF-1 transcription factor regulates the expression of ccg-1and cat-1 genes in response to fludioxonil under OS-2 MAP kinase in Neurosporacrassa. Fungal Genet Biol 45:1562–1569.

28. Westfall PJ, Thorner J (2006) Analysis of mitogen-activated protein kinase signalingspecificity in response to hyperosmotic stress: Use of an analog-sensitive HOG1 allele.Eukaryot Cell 5:1215–1228.

29. Han J, Lee JD, Bibbs L, Ulevitch RJ (1994) A MAP kinase targeted by endotoxin andhyperosmolarity in mammalian cells. Science 265:808–811.

30. Yoshimi A, Kojima K, Takano Y, Tanaka C (2005) Group III histidine kinase is a positiveregulator of Hog1-type mitogen-activated protein kinase in filamentous fungi.Eukaryot Cell 4:1820–1828.

31. Meachum ZD, Jr, Colvin HJ, Jr, Braymer HD (1971) Chemical and physical studies ofNeurospora crassa invertase. Molecular weight, amino acid and carbohydrate com-position, and quaternary structure. Biochemistry 10:326–332.

32. Noguchi R, et al. (2007) Identification of OS-2 MAP kinase-dependent genes inducedin response to osmotic stress, antifungal agent fludioxonil, and heat shock in Neu-rospora crassa. Fungal Genet Biol 44:208–218.

33. Xiong Y, Sun J, Glass NL (2014) VIB1, a link between glucose signaling and carboncatabolite repression, is essential for plant cell wall degradation by Neurospora crassa.PLoS Genet 10:e1004500.

34. Ghosh S, et al. (2014) N-acetylglucosamine (GlcNAc)-inducible gene GIG2 is a novelcomponent of GlcNAc metabolism in Candida albicans. Eukaryot Cell 13:66–76.

35. Lamb TM, Goldsmith CS, Bennett L, Finch KE, Bell-Pedersen D (2011) Direct tran-scriptional control of a p38 MAPK pathway by the circadian clock in Neurosporacrassa. PLoS One 6:e27149.

36. Plesofsky N, Higgins L, Markowski T, Brambl R (2016) Glucose starvation alters heatshock response, leading to death of wild type cells and survival of MAP kinase sig-naling mutant. PLoS One 11:e0165980.

37. Wang M, et al. (2013) A mitogen-activated protein kinase Tmk3 participates in highosmolarity resistance, cell wall integrity maintenance and cellulase production reg-ulation in Trichoderma reesei. PLoS One 8:e72189.

38. Seidl V, et al. (2008) The Hypocrea jecorina (Trichoderma reesei) hypercellulolyticmutant RUT C30 lacks a 85 kb (29 gene-encoding) region of the wild-type genome.BMC Genomics 9:327.

39. Cheung WD, Hart GW (2008) AMP-activated protein kinase and p38 MAPK activateO-GlcNAcylation of neuronal proteins during glucose deprivation. J Biol Chem 283:13009–13020.

40. Adhikari H, Cullen PJ (2014) Metabolic respiration induces AMPK- and Ire1p-dependent activation of the p38-Type HOGMAPK pathway. PLoS Genet 10:e1004734.

41. Posas F, et al. (1996) Yeast HOG1 MAP kinase cascade is regulated by a multistepphosphorelay mechanism in the SLN1-YPD1-SSK1 “two-component” osmosensor. Cell86:865–875.

42. Maeda T, Wurgler-Murphy SM, Saito H (1994) A two-component system that regu-lates an osmosensing MAP kinase cascade in yeast. Nature 369:242–245.

43. Meena N, Kaur H, Mondal AK (2010) Interactions among HAMP domain repeats act asan osmosensing molecular switch in group III hybrid histidine kinases from fungi.J Biol Chem 285:12121–12132.

44. Lamb TM, Finch KE, Bell-Pedersen D (2012) The Neurospora crassa OS MAPK pathway-activated transcription factor ASL-1 contributes to circadian rhythms in pathway re-sponsive clock-controlled genes. Fungal Genet Biol 49:180–188.

45. Beattie SR, et al. (2017) Filamentous fungal carbon catabolite repression supportsmetabolic plasticity and stress responses essential for disease progression. PLoSPathog 13:e1006340.

46. 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.

47. McCluskey K, Wiest A, Plamann M (2010) The Fungal Genetics Stock Center: A re-pository for 50 years of fungal genetics research. J Biosci 35:119–126.

48. Heller J, et al. (2016) Characterization of greenbeard genes involved in long-distancekind discrimination in a microbial eukaryote. PLoS Biol 14:e1002431.

49. Kim D, Langmead B, Salzberg SL (2015) HISAT: A fast spliced aligner with low memoryrequirements. Nat Methods 12:357–360.

50. Trapnell C, et al. (2012) Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. Nat Protoc 7:562–578.

51. Srivastava M, et al. (2012) An inhibitor of nonhomologous end-joining abrogatesdouble-strand break repair and impedes cancer progression. Cell 151:1474–1487.

52. Galagan JE, et al. (2003) The genome sequence of the filamentous fungus Neurosporacrassa. Nature 422:859–868.

53. Langmead B, Salzberg SL (2012) Fast gapped-read alignment with bowtie 2. NatMethods 9:357–359.

54. McKenna A, et al. (2010) The genome analysis toolkit: A MapReduce framework foranalyzing next-generation DNA sequencing data. Genome Res 20:1297–1303.

55. Katoh K, Standley DM (2013) MAFFT multiple sequence alignment software version 7:Improvements in performance and usability. Mol Biol Evol 30:772–780.

56. Jones DT, Taylor WR, Thornton JM (1992) The rapid generation of mutation datamatrices from protein sequences. Comput Appl Biosci 8:275–282.

57. Kumar S, Stecher G, Tamura K (2016) MEGA7: Molecular evolutionary geneticsanalysis version 7.0 for bigger datasets. Mol Biol Evol 33:1870–1874.

58. Xiong Y, et al. (2014) The proteome and phosphoproteome of Neurospora crassa inresponse to cellulose, sucrose and carbon starvation. Fungal Genet Biol 72:21–33.

E8674 | www.pnas.org/cgi/doi/10.1073/pnas.1707713114 Huberman et al.

Dow

nloa

ded

by g

uest

on

Oct

ober

28,

202

0