Gene 593 (2016) 143–153

Contents lists available at ScienceDirect

Gene

j ourna l homepage: www.e lsev ie r .com/ locate /gene

Research paper

Functional and transcriptomic analysis of the key unfolded proteinresponse transcription factor HacA in Aspergillus oryzae

Bin Zhou, Jingyi Xie, Xiaokai Liu, Bin Wang ⁎, Li Pan ⁎School of Bioscience and Bioengineering, South China University of Technology, Guangzhou, PR China

Abbreviations: UPR, unfolded protein response; Hdisruption mutant; HacA-CA, a strain that expressed aHacA; WT, wild-type; ER, endoplasmic reticulum; bZIP,ER-associated degradation; CD, Czapek–Dox medium;extract medium; NBRC, NITE Biological Resource Center;polymerase chain reaction; agdA, alpha-glucosidasefluorescent protein; PBS, phosphate-buffered saliphenylindole; RNA-seq, RNA sequencing; AspGD, Aspergthe number of fragments per kilobase of exons per miOntology; FDR, False Discovery Rate; KEGG, Kyoto EncyclbipA, encoding an ER chaperone; pdi, encoding proencoding peptidyl-prolyl cis-trans isomerase; amyB, encquantitative real-time PCR; BP, Biological Processes; CMolecular Functions; glaA, glucoamylase; RESS, Repremechanism; eIF2α, the eukaryotic translation initiation ftranscription factor-4; DTT, dithiothreitol.⁎ Corresponding authors.

E-mail addresses: btbinwang@scut.edu.cn (B. Wang),

http://dx.doi.org/10.1016/j.gene.2016.08.0180378-1119/© 2016 Elsevier B.V. All rights reserved.

a b s t r a c t

a r t i c l e i n f o

Article history:Received 30 April 2016Received in revised form 4 August 2016Accepted 8 August 2016Available online 9 August 2016

HacA is a conserved basic leucine zipper transcription factor that serves as themaster transcriptional regulator inthe unfolded protein response (UPR). To comprehensively evaluate the role of HacA in Aspergillus oryzae, ahomokaryotic hacA disruption mutant (HacA-DE) and a strain that expressed a constitutively active form ofHacA (HacA-CA) were successfully generated, and transcriptome analyses of these mutants were performed.Growth and phenotypic profiles demonstrated that hyphal growth and sporulation were impaired in theHacA-DE and HacA-CA strains that were grown on complete and minimal media, and the growth impairmentwasmore pronounced for the HacA-CA strain. Comparedwith a wild-type (WT) strain, the transcriptome resultsindicated that differentially expressed genes in these mutants mainly fell into four categories: the protein secre-tory pathway, amino acidmetabolism, lipidmetabolism, and carbohydratemetabolism. Furthermore, we identi-fied 80 and 36 genes of the secretory pathway whose expression significantly differed in the HacA-CA strain(compared with theWT and HacA-DE strains) and HacA-DE strain (compared with theWT strain), respectively,which mostly belonged to protein folding/UPR, glycosylation, and vesicle transport processes. Both the HacA-CAand HacA-DE strains exhibited reduced expression of extracellular enzymes, especially amylolytic enzymes,which resulted from the activation of the repression under secretion stress mechanism in response to endoplas-mic reticulum stress. Collectively, our results suggest that the function of HacA is important not only for UPR in-duction, but also for growth and fungal physiology, as it serves to reduce secretion stress in A. oryzae.

© 2016 Elsevier B.V. All rights reserved.

Keywords:Aspergillus oryzaeHacAProtein secretionRepression under secretion stressRNA-seqUnfolded protein response

1. Introduction

In eukaryotic cells, the endoplasmic reticulum (ER) is crucial for theproduction of membrane and secreted proteins, and its production abil-ity is limited by the level of ER-resident chaperones, foldases, and othermodifying enzymes that assist in protein folding. When protein folding

acA-DE, a homokaryotic hacAconstitutively active form of

the basic leucine zipper; ERAD,DPY, dextrose-peptone-yeastORF, open reading frame; PCR,gene;; eGFP, enhanced greenne; DAPI, 4′,6-diamidino-2-illus Genome Database; FPKM,llion mapped reads; GO, Geneopedia of Genes and Genomes;tein disulfide isomerase; ppi,oding alpha-amylase; qRT-PCR,C, Cellular Components; MF,ssion Under Secretion Stressactor-2 alpha; ATF4, activating

btlipan@scut.edu.cn (L. Pan).

requirements exceed the ER's folding capabilities, misfolded proteinscan accumulate and elicit stress to the ER, which will activate theunfolded protein response (UPR) mechanism to decrease ER stress(Feng et al., 2011; Montenegro-Montero et al., 2015; Tanaka et al.,2015). In Saccharomyces cerevisiae andfilamentous fungi, theUPRmain-ly depends on an evolutionarily conserved signaling cascade that is me-diated by the ER-resident transmembrane kinase/endoribonucleaseIRE1 and the basic leucine zipper (bZIP) transcription factor Hac1p/HacA (HacA in filamentous fungi) (Carvalho et al., 2012; Heimel,2015; Montenegro-Montero et al., 2015). In the UPR process, the hac1/hacAmRNAs of yeast and filamentous fungi undergo similar, unconven-tional splicing reactions to produce functional Hac1p/HacA proteins,which are then shuttled into the nucleus where they induce the expres-sion of ER chaperone genes and ER-associated degradation (ERAD)genes to promote the refolding or degradation of unfolded proteins(Moon et al., 2015; Tanaka et al., 2015).

A transcriptome analysis under UPR-inducing conditions in fungisuggested that the target genes of the UPR are predominantly enrichedin functional categories that are associated with the secretory pathway,including ER-resident chaperones, phospholipid metabolism, fatty acidsynthesis, translocation, protein glycosylation, cell wall biosynthesis,vesicular transport, vacuolar protein targeting, and protein degradation

144 B. Zhou et al. / Gene 593 (2016) 143–153

(Travers et al., 2000; Sims et al., 2005; Arvas et al., 2006; Guillemetteet al., 2007; Wang et al., 2010). However, the majority of these studiesinduced the UPR signaling pathway through the use of harsh chemicals(dithiothreitol (DTT) or tunicamycin) (Guillemette et al., 2007;Wang etal., 2010), homologous or heterologous protein expression (Guillemetteet al., 2007; Kwon et al., 2012; Liu et al., 2014), or growth conditions thatinduced secretory hydrolytic enzyme production (Jorgensen et al., 2009;Benz et al., 2014). Additionally, these studies were performed on wild-type (WT) strains; thus, the specific contributions of HacA could not beevaluated. To address this, genome-wide expression profiles have beengenerated for a hacA disruption mutant and a strain that constitutivelyexpresses hacA (Feng et al., 2011; Carvalho et al., 2012; Fan et al.,2015). In filamentous fungi, hacA deletion mutants of Aspergillusfumigatus (Feng et al., 2011),Neurospora crassa (Fan et al., 2015), and As-pergillus niger (Carvalho et al., 2010) have been constructed successfully,and the hacA disruptions resulted in drastic growthdefects inA. niger andN. crassa, and nearly normal growth in A. fumigatus. However, ahomokaryotic hacA disruptionmutant has not been successfully generat-ed in Aspergillus oryzae (Tanaka et al., 2015), and further understandingthe contribution of A. oryzae hacA to ER stress adaptation is still needed.

A. oryzae, an organism that is generally recognized as safe, is a suit-able host for homologous and heterologous protein production(Nevalainen et al., 2005; Ward et al., 2006). To comprehensively evalu-ate the role of HacA as the master regulator of the UPR in the A. oryzaeprotein secretory pathway, we successfully generated a homokaryotichacA disruption mutant (HacA-DE) and a strain that constitutivelyexpressed an activated form of hacA (HacA-CA). Our results demon-strated that hyphal growth and sporulation were impaired in theHacA-DE and HacA-CA strains when they were cultured in completemedium, and minimal media containing glucose, maltose, dextrin, andstarch as carbon sources. Additionally, the growth impairment of theHacA-CA strain was more pronounced than that of the HacA-DE strain.Here, we performed a transcriptome analysis of the HacA-DE andHacA-CA strains to more thoroughly characterize the effect of hacA onthe secretory pathway and whole cell metabolism.

2. Material and methods

2.1. Strains and culture conditions

A. oryzae strains used in this study (Table 1) were cultivated in min-imal medium (Czapek–Dox (CD) medium) (Zhou et al., 2015) contain-ing 1 or 2% (w/v) of glucose as a carbon source (or other carbon sourcesas indicated), 0.3% (w/v) NaNO2, 0.1% (w/v) K2HPO4, 0.2% KCl, 0.05%(w/v) MgSO4·7H2O, and 0.001% (w/v) FeSO4·7H2O, pH 5.5; or in com-plete medium (dextrose-peptone-yeast extract (DPY) medium) con-taining 2% (w/v) glucose, 1% (w/v) peptone, 0.5% (w/v) yeast extract,0.1% (w/v) K2HPO4, and 0.05% (w/v) MgSO4·7H2O. When required,10 mM uridine was added.

2.2. Disruption of the hacA gene

The A. oryzae hacA gene (AO090124000074) was deleted in a ΔpyrGstrain (niaD−; ΔpyrG; ku70::ptrA) by replacing its open reading frame

Table 1The A. oryzae strains used in this study.

Strain Strain origin Geno

niaD300a RIB40 niaD−

Δku70 niaD300 niaD−

ΔpyrG Δku70 niaD−

HacA-DE ΔpyrG niaD−

HacA-DE-RE HacA-DE niaD−

HacA-CA HacA-DE-RE niaD−

HacA-CA-eGFP HacA-DE-RE niaD−

a The A. oryzae strain niaD300, a niaDmutant that was derived from the A. oryzae RIB40 stra

(ORF) with a DNA fragment containing pyrG as a selection marker.The ΔpyrG strain construction was performed as described in the Sup-plemental material and methods (Supplementary Figs. S1 and S2). Thecassette used to delete hacA was produced by a fusion polymerasechain reaction (PCR) in three steps: first, the hacA upstream (ATG startcodon) (1.5 kb) and downstream (TAG stop codon) (0.5 kb and1.5 kb) regions and the pyrG gene (1398 bp) containing its promoterand terminator regions were amplified using the primers listed in Sup-plementary Table S1. The pyrG genewas amplified using plasmid ANIp7(Storms et al., 2005) as a template; the hacA upstream and downstreamregionswere amplified using niaD300 genomic DNA as a template. Sec-ond, a first round of fusion PCR was performed using the three frag-ments (1.5 kb hacA upstream and 0.5 kb downstream regions, and thepyrG gene) as template DNAs and hacAUF/AnpyrGR as the outwardprimers. Third, a second round of fusion PCR was performed using thefirst round fusion PCR product and the 1.5 kb hacA downstream frag-ment as template DNAs and hacAUF/hacADR as the outward primers.Then, the DNA fragment obtained from the second round of fusionPCR was inserted into the pMD20 T-Vector (TaKaRa, Shiga, Japan),yielding pΔhacA::pyrG. The DNA fragment obtained after digestingpΔhacA::pyrG with EcoRV was used for A. oryzae transformation, andthe transformation was performed according to the method of Gomiet al. (1987). The hacA-deleted strain was identified by PCR amplifica-tion of chromosomal sequences, as described previously by Wanget al. (2008) and Zhang et al. (2011) (Supplementary Fig. S3). Excisionof the pyrG marker was conducted according to the method ofMaruyama and Kitamoto (2008). The pyrG-excised strainwas identifiedamong 5-fluoroorotic acid-resistant transformants using PCR and phe-notypic screening. Subsequently, one pyrG-excised strain, HacA-DE-RE,was obtained (Supplementary Fig. S4).

2.3. Construction of a constitutively active hacA (hacAi) strain and a hacAi-eGFP strain

To complement the hacA knockout strain with a constitutivelyactivated allele of the hacA gene, a complementation cassette was con-structed in the HacA-DE-RE strain (niaD−; ΔhacA; ΔpyrG; Δku70::ptrA).As a reference, a similar complementation cassette was made using thehacAi-eGFP (enhanced green fluorescent protein) gene. The cassettethat was used to obtain a strain that only expresses the constitutivelyactive hacA gene was produced by fusion PCR in three steps. First, thehacA upstream (ATG) (1.7 kb) and downstream (TAG) (1.4 kb) regions,the alpha-glucosidase gene (agdA) terminator (E12508.1) region, theconstitutively active hacA gene (hacAi, lacking the 20-nucleotide intronof hacA), and the pyrG gene containing its promoter and terminator re-gions (1398 bp) were amplified using the primers listed in Supplemen-tary Table S1. Full-length hacAiwas amplified using plasmid pMDGHRas a template (Zhou et al., 2015); the pyrG gene region was amplifiedusing plasmid ANIp7 as a template; and the hacA upstream, hacA down-stream, and agdA terminator regions were amplified using niaD300genomic DNA as a template. Second, a fusion PCR was conductedusing the four fragments (hacAi, agdA terminator, pyrG, and hacA down-stream) as template DNAs and hacAi-CDS-F/hacAiDR as the outwardprimers (Supplementary Table S1). Then, the DNA fragment obtained

type Reference

NBRC; Δku70::ptrA This study; ΔpyrG; Δku70::ptrA This study; ΔhacA::pyrG; Δku70::ptrA This study; ΔhacA; ΔpyrG; ΔKu70::ptrA This study; ΔhacA::hacA-i + pyrG; Δku70:: ptrA This study; ΔhacA::hacA-i + eGFP + pyrG; Δku70:: ptrA This study

in, was obtained from the NITE Biological Resource Center (NBRC) in Japan.

145B. Zhou et al. / Gene 593 (2016) 143–153

from the fusion PCR was inserted into the pMD20 T-vector, yieldingphacAiP. Third, the hacA upstream (promoter) region was insertedinto the XbaI site of the plasmid phacAiP using the IN-FUSION CloningKit (TaKaRa), which yielded phacAi. The complementation cassettewas sequenced, released from phacAi via EcoRI restriction, and usedfor A. oryzae transformation. Transformants with a targeted integrationof the construct at the hacA locuswere screened byPCR of their chromo-somal sequences (Supplementary Fig. S5), and a quantitative real-timePCR (qRT-PCR) assay was used to assess their hacAi copy number (Sup-plementary Figs. S6 and S7). The qRT-PCR assay is described in detail inthe Supplementary materials and methods. The construction of thehacAi-eGFP strainwas similar to that of the hacAi strain, and it is also de-scribed in detail in the Supplementary materials and methods.

2.4. Fluorescence microscopy

Strains were inoculated into 50 mL of DPY medium and cultivatedfor 40 h at 30 °C. Mycelia were collected and washed with phosphate-buffered saline (PBS, Sigma-Aldrich, St. Louis, MO, USA). Then, themycelia were stained with 0.5 μg/mL 4′,6-diamidino-2-phenylindole(DAPI) (Sigma-Aldrich) for 20 min. The mycelia were washed threetimes in PBS and examined using an LSM710 laser-scanning confocalmicroscope (Carl Zeiss, Oberkochen, Germany), with excitation wave-lengths of 405 and 498 nm and emission wavelengths of 488 and546 nm for DAPI and GFP, respectively. Phase-contrast brightfield andfluorescent images were captured with an AxioCam camera (CarlZeiss) using a 100× oil immersion lens, and they were processedusing ZEN 2012 software.

2.5. RNA extraction and qRT-PCRs

Equal numbers ofWT,HacA-DE, andHacA-CA conidiawere inoculat-ed into 50 mL of DPY medium and incubated statically at 30 °C. Every2 h, the cultures were gently vortexed for 30 s. After cultivation for40 h, mycelia were collected and 0.5 g (fresh weight) was transferredinto 250-mL shakeflasks containing 50mLof DPYmedium and incubat-ed for an additional 24 h at 30 °C with shaking at 200 rpm. For ER stresstreatment, DTT was added to the culture at a final concentration of20 mmol/L. Cultures without DTT treatment had an equivalent volumeof distilled water added. Mycelia were filtered through filter paperand frozen immediately in liquid nitrogen and stored at −80 °C. TotalRNA extraction and qRT-PCRs were performed as previously described(Ye and Pan, 2008; Zhou et al., 2015).

2.6. Preparation of a cDNA library for RNA sequencing

Total RNA samples and RNA sequencing (RNA-seq) cDNA librarieswere prepared as described by Wang et al. (2010). The cDNA librarywas sequenced on the Illumina sequencing platform (Illumina HiSeq2500, Illumina, San Diego, CA, USA) using the paired-end technologyin a single run (GENE DENOVO, Guangzhou, China), in which 125-bppaired-end reads were obtained. A Perl program was written to selectclean reads by removing low-quality sequences (in which N50% of thebases had a quality of b20 in one sequence), reads with N5% of Nbases (unknown bases), and reads containing adaptor sequences. Theraw sequencing data were deposited in the Sequence Read Archive(http://www.ncbi.nlm.nih.gov/sra/) at the National Center for Biotech-nology Information under the accession number SRP067204.

2.7. RNA-seq data analysis

The resulting reads were aligned to the A. oryzae RIB40 genomethat was retrieved from AspGD (version s01-m08-r01, www.aspergillusgenome.org). The gene expression level was normalized bythe number of fragments per kilobase of exons per million mappedreads (FPKM) (Trapnell et al., 2010). Gene Ontology (GO) terms of A.

oryzae genes were assigned using Cytoscape with the ClueGO plug-in(Bindea et al., 2009). GO terms with false discovery rate (FDR) valuesb0.05 were defined as over-represented. ClueGO analysis parameterswere set in the panel as follows: View Style Setting as Significance,Show only Pathways with p b 0.05, a GO Tree interval from Min Level 3to Max Level 8, and Min 3 genes and 4% genes as a cluster in GO Term/Pathway Selection. The significance of each term or groupwas calculatedwith a two-sided hypergeometric test and p-value correction using theBenjamini-Hochberg method. The corrected gene annotation list has12,357 genes that are associatedwith one ormore GO terms. The report-er algorithmwas used twice to identify GO terms that weremainly asso-ciated with upregulated or downregulated genes. Kyoto Encyclopedia ofGenes and Genomes (KEGG) pathway analyses of differentiallyexpressed genes were performed as described by Kanehisa et al. (2008).

3. Results

3.1. Construction and analysis of the hacA deletion strain and acomplemented strain expressing the activated form of hacA

To obtain an A. oryzae hacA deletion strain (HacA-DE, niaD−;ΔhacA::pyrG; Δku70::ptrA), the hacA ORF was replaced by the pyrG cas-sette. Correct integration of the pyrG cassettewas verified by PCR ampli-fication of the chromosomal sequences (Supplementary Fig. S3).Furthermore, to obtain a complemented strain that expresses a consti-tutively activated form of HacA, we integrated the spliced form ofhacA, which lacks the 20-nucleotide intron, into the hacA locus in theHacA-DE-RE (niaD−;ΔhacA;ΔpyrG;Δku70::ptrA) strain. Transformantswith the correct integration pattern were selected by PCR amplificationof the chromosomal sequences and a qRT-PCR assay of their hacAi copynumber, and an activated form of the hacA gene was successfully inte-grated at the hacA locus in the HacA-CA strain (niaD−; ΔhacA::hacA-i + pyrG; Δku70::ptrA) (Supplementary Figs. S5–S7). Growth of theWT, HacA-DE, and HacA-CA strains on DPY medium and minimalmedia containing glucose, maltose, dextrin, or starch as carbon sourceswas assessed. The radial growth rate (colony size) of both the HacA-DE,and the growth of the HacA-CA strain was impaired more severely thanthat of the HacA-DE strain (Fig. 1a, Supplementary Figs. S8 and S9).Meanwhile, the HacA-DE and HacA-CA mutants exhibited abnormalconidiation (Fig. 1a); compared with the WT strain, the HacA-DE andHacA-CAmutants produced very few spores, and the sporulation defectwas slightly more severe in HacA-DE mutant (Fig. 1b). Additionally, inDPY medium containing DTT, growth of the WT, HacA-DE, and HacA-CA strains was impaired, and the growth of the HacA-CA strain was im-pairedmore severely; in glucosemedium that contained DTT, growth oftheWT, HacA-DE, andHacA-CA strainswas abolished completely. Theseresults indicated that the effects of a constitutively activated UPR differfrom those resulting from the lack of the UPR, and the induction of theUPR that resulted from expressing constitutively active hacA had obvi-ously deleterious effects on vegetative growth and morphology, evenunder acute ER-stress.

TheHacA-DE andHacA-CAmutants were cultured in liquid DPYme-dium at 30 °C, with shaking at 200 rpm, to characterize their UPR sig-nals. However, after 24 h of cultivation, the growth rate of the HacA-CA mutant was extremely poor, which made it difficult to harvestmycelia (Supplementary Fig. S10a). To solve this problem, the HacA-DE and HacA-CA mutants were pre-cultured by inoculating them into50 mL of DPY medium and incubating them statically at 30 °C. After40 h of cultivation, mycelia of the pre-cultured mutants were collected(up to 0.5 g fresh weight) and transferred into 50 mL of fresh DPY me-dium. After an additional 24 h of cultivation at 30 °C with shaking at200 rpm, theHacA-CAmutant showed nearly normal growth comparedto the WT strain (Supplementary Fig. S10b). To determine whether theloss of hacA and the expression of the activated form of hacA were suf-ficient to disrupt and activate UPR signaling in A. oryzae, respectively,the expression levels of UPR-relevant marker genes, including bipA

146 B. Zhou et al. / Gene 593 (2016) 143–153

Fig. 2. Analysis of differentially expressed genes in the WT, HacA-DE and HacA-CA strains. (a) Overview of the number of differentially expressed genes via pairwise comparisons of theHacA-CA andWT, HacA-CA and HacA-DE, and HacA-DE and WT strains. (b) Venn diagrams of the number of overlapping and non-overlapping induced or repressed genes via pairwisecomparisons of the HacA-CA andWT, and HacA-CA and HacA-DE strains.

147B. Zhou et al. / Gene 593 (2016) 143–153

(encoding an ER chaperone), pdi (encoding protein disulfide isomer-ase), ppi (encoding peptidyl-prolyl cis-trans isomerase), amyB(encoding alpha-amylase), and hacA itself, were examined by qRT-PCR. The qRT-PCR analysis showed that the expression levels of theUPR target genes (hacAi, bipA, pdi, and ppi) were significantly higherin the HacA-CAmutant, which is consistent with the results in responseto DTT treatment of theWT strain (Fig. 1c). Additionally, the expressionof UPR target genes (bipA, pdi, and ppi)was downregulated in theHacA-DE mutant, which disrupted UPR signaling (Fig. 1c). The expressionlevels of amyB were largely lower in the HacA-CA, HacA-DE, and WT(DTT treatment) strains (Fig. 1c). However, the constitutive UPR activa-tion in the HacA-CAmutant triggered amuchmore severe downregula-tion of amyB expression compared with that in the HacA-DE and WT(DTT treatment) strains (Fig. 1c). Fluorescence microscopy of theHacA-CA-eGFP strain, inwhichHacAi-eGFP expression is under the con-trol of the hacA promoter, showed a similar pattern of fluorescence, asits hyphae were stained with the nuclear dye DAPI (Fig. 1d), confirming

Fig. 1. Analysis of the HacA-DE and HacA-CA strains. (a) Growth and phenotypic profiles of tminimal agar plates containing 1% of different carbon sources (as indicated) at 30 °C for 5 d. Dwhich induces ER-stress. (b) Differences in the sporulation of the WT, HacA-DE, and HacA-CAand incubated at 30 °C for 5 d. To obtain the number of conidiospores per plate, total conidiothree independent colonies. (c) Expression analysis of UPR genes by qRT-PCR of the hacAi, biwas added to the culture at a final concentration of 20 mM. To the control culture equal amo2 h after applying the stress agent. The glyceraldehyde-3-phosphate dehydrogenase geneanalyzed in triplicate, and values are the means of three replicates ± S.D. (d) Subcellularincubated statically at 30 °C for 40 h. Every 2 h, the culture was gently vortexed for 30 s. Fluowere also stained by DAPI.

that HacA indeed localized to the nucleus. Taken together, the HacA-DEand HacA-CA mutants were constructed successfully to disrupt and ac-tivate, respectively, UPR signaling in A. oryzae, and they were subse-quently used to study the UPR process on a genome-wide scale.

3.2. Summary of the transcriptomes of the HacA-DE, HacA-CA, and WTstrains

We performed a transcriptome analysis of the WT, HacA-DE, andHacA-CA strains based on RNA-seq data, and obtained a total of 14.77,18.58, and 17.74million readswith an average length of 125 bp, respec-tively; 13.68 (92.67%), 17.32 (93.22%) and 16.18 (91.22%) million readswere uniquely mapped to the genome, with a tolerance of a 2-bp mis-match, which represented 37, 47, and 43-fold coverage of the A. oryzaeRIB40 genome, respectively (Supplementary Table S2).

We further analyzed the RNA-seq data with TopHat and Cufflinks(Trapnell et al., 2012) to conduct a genome analysis of differential

he WT, HacA-DE, and HacA-CA strains. The strains were grown on DPY agar medium orithiothreitol (20 mM) is a strong reducing agent that prevents disulfide bond formation,strains. The WT, HacA-DE, and HacA-CA strains were inoculated onto DPY medium platesspores for each colony were harvested and counted. Conidial yield data are the means ofpA, pdi, ppi, and amyB genes. ×, without hacA gene. For UPR induction of WT strain, DTTunts of water was added. Samples from the stressed and unstressed cultures were taken(gapdh) was used as an endogenous control to normalize the data. Each sample waslocalization of hacAi-eGFP. The strain was inoculated into 50 mL of DPY medium andrescence was evaluated using a Zeiss LSM710 laser-scanning confocal microscope. Nuclei

148 B. Zhou et al. / Gene 593 (2016) 143–153

gene expression among the HacA-DE, HacA-CA, and WT strains. Com-pared with the WT strain, the expression of 775 genes in the HacA-CAmutant was upregulated significantly (FDR b 0.05), while the expres-sion of 1991 genes was downregulated; the expression of 1374 genesin the HacA-DE mutant was upregulated significantly (FDR b 0.05),while the expression of 972 genes was downregulated. Comparedwith the HacA-DE mutant, the expression of 481 genes in the HacA-CA mutant was upregulated significantly (FDR b 0.05), while the ex-pression of 2386 genes was downregulated (Fig. 2a and SupplementaryTable S3). Additionally, as shown in the Venn diagram, 296 upregulatedgenes overlapped between the HacA-CA andWT strains, and the HacA-CA and HacA-DE strains, while 1147 downregulated genes overlapped(Fig. 2b and Supplementary Table S4).

3.3. Impacts of the constitutively activated form of HacA on the tran-scriptome profile of A. oryzae

To determine the function of the constitutively activated form ofHacA in the HacA-CA mutant, overrepresented GO terms among the

Fig. 3. KEGG pathways analysis of differentially induced or repressed genes in the HacA-CA strathe HacA-CA mutant strain compared with the WT strain.

differentially expressed genes were identified using the WT strain as acontrol. For this analysis, the upregulated genes were enriched in 246GO terms, which included 158 terms of ‘biological processes’ (BP), 50terms of ‘cellular components’ (CC), and 38 terms of ‘molecular func-tions’ (MF). Additionally, the downregulated genes were enriched in343 GO terms that included 229 BP terms, 39 CC terms, and 75 MFterms (Supplementary Table S5 p1). GO and KEGG metabolic pathwayanalyses indicated that the upregulated genes were specifically over-represented in fourmajor functional categories: i) the protein secretorypathway included protein processing in the ER, protein export, and N-glycan biosynthesis; ii) amino acid metabolism consisted of 18 aminoacids, including lysine, phenylalanine, tyrosine, and tryptophan; iii)translation linked to aminoacyl-tRNA biosynthesis; and iv) metabolismof cofactors and vitamins included pantothenate and CoA biosynthesis,riboflavin metabolism, vitamin B6 metabolism, and one folate carbonpool. These results suggested that constitutively expressing the activat-ed form of HacA upregulated the expression of genes that encode pro-teins that are involved in protein secretion, amino acid biosynthesis,and aminoacyl-tRNA biosynthesis, which improve protein production.

in. Representation of the main significantly induced and repressed pathways (p b 0.05) in

149B. Zhou et al. / Gene 593 (2016) 143–153

Conversely, the downregulated genes in the HacA-CA mutant weremainly enriched in fivemain functional categories: i) carbohydrate me-tabolism, including starch, sucrose, fructose, mannose, and galactosemetabolism, pentose and glucuronate interconversions, the citric acidcycle, glycolysis/gluconeogenesis, and pyruvate, glyoxylate, anddicarboxylate metabolism; ii) amino acid metabolism linked to the me-tabolism of eight amino acids, including tyrosine, glutathione, and va-line; iii) lipid metabolism comprising fatty acid, alpha-linolenic acid,glycerolipid, and arachidonic acid metabolism, synthesis and degrada-tion of ketone bodies, and sphingolipid and linoleic acid metabolism;iv) glycan degradation and glycosphingolipid biosynthesis; and v) per-oxisome transport and catabolism (Fig. 3; Supplementary Tables S5 p1and S6 p1).We assumed that the expression of genes involved in carbo-hydratemetabolism and energy productionwas depressed in the HacA-CAmutant to slowhyphal growth. Furthermore, to gain insights into theimpact of the constitutive activation of HacA on A. oryzae, theWT strainand the HacA-DEmutant were used as reference strains to detect genesthatwere differentially expressed in theHacA-CAmutant. GO andKEGGanalyses indicated that 296 upregulated genes were specificallyenriched in protein processing in the ER, N-glycan biosynthesis, andprotein export (Supplementary Tables S5 p2 and S6 p2). Additionally,1147 over-represented, downregulated genes that overlapped betweenthe HacA-CA and WT strains, and the HacA-CA and HacA-DE strainswere classified primarily into four main categories: carbohydrate,amino acid, and lipid metabolism, and transport and catabolism (Sup-plementary Tables S5 p2 and S6 p2). Upregulated genes that overlapped

Fig. 4. KEGG pathways analysis of differentially induced or repressed genes in the HacA-DE strathe HacA-DE mutant strain compared with the WT strain.

between the HacA-CA and WT strains, and the HacA-CA and HacA-DEstrains might be used to obtain a detailed mechanistic picture of theprotein secretion response of the A. oryzae secretory machinery, whilethe downregulated genes that were linked to central metabolicpathways might reflect the growth defect that was observed in theHacA-CA mutant.

3.4. Impacts of the hacA deletion on the transcriptome profile of A. oryzae

To compare the phenotypes of the HacA-DE mutant and the WTstrain, differentially expressed genes were subjected to GO and KEGGanalyses. The upregulated genes were enriched in 69 GO terms, whichincluded 45 BP, 21 CC, and three MF terms. The down-regulated geneswere enriched in 174 GO terms, which included 113 BP, 22 CC, and 39MF terms (Supplementary Table S5 p3). These results indicated thatthe upregulated genes were over-presented in one major categorythat was linked to the metabolism of 13 amino acids, including glycine,serine, and threonine. Conversely, the downregulated genes in theHacA-DE mutant were enriched in five major categories: i) the proteinsecretory pathway, including genes involved in protein processing inthe ER, and various types of N-glycan biosynthesis; ii) carbohydratemetabolism linked to pyruvate metabolism, the pentose phosphatepathway, starch, sucrose, fructose, and mannose metabolism, andgluconeogenesis; iii) metabolism linked to nine amino acids, includingvaline, leucine, and isoleucine; iv) lipid metabolism, includinggenes involved in glycerolipid, fatty acid, alpha-linolenic acid,

in. Representation of themain significantly induced and repressed pathways (p b 0.05) in

150 B. Zhou et al. / Gene 593 (2016) 143–153

glycerophospholipid and arachidonic acid metabolism, and synthesisand degradation of ketonebodies; and v) transport, includingATP-bind-ing cassette transporters and peroxisomes (Fig. 4; Supplementary Ta-bles S5 p3 and S6 p3). From these results, we reasonably inferred thatamino acid biosynthesis could be increased and that amino acid degra-dation could be decreased tomake up for the lack of protein productionmaterial in the HacA-DE mutant. Additionally, the inhibition of proteinprocessing and N-glycan biosynthesis, as well as the growth reductionresulting from the suppression of carbohydrate metabolism, alleviatedthe ER stress that was caused by the disruption of the UPR in theHacA-DE strain.

3.5. Regulatory models of the protein secretory pathway in the HacA-CAand HacA-DE strains

To construct regulatory models of the protein secretory pathway inthe HacA-CA and HacA-DE strains, we mapped the RNA-seq data forthe HacA-CA mutant, the HacA-DE mutant, and the WT strain to all369 A. oryzae genes encoding secretory pathway components thatwere identified by Liu et al. (2014) (Supplementary Table S7). In theHacA-CA strain, the expression of 80 of the 369 genes in the secretorypathway was significantly changed (FDR b0.05) compared with thatof the WT and HacA-DE strains, among which 78 were upregulatedand two were downregulated (Fig. 5 and Supplementary Table S7 p1).Among the secretory pathway-related genes, the majority of the in-duced genes belonged to ER-related processes, including translocationinto the ER, protein folding/UPR, glycosylation/quality control, proteindegradation, and COPI- and COPII-mediated transport processes.The expression of the majority UPR target genes, such as PDI1, HUT1,KAR2, ERJ5, LHS1, MNS1, MPD1, CPR1 and SSA1, was upregulated, whilethe expression of the important components of the UPR (GCN4(AO090009000459) and IRE1 (AO090005000934) was unchanged. Fur-thermore, comparedwith theWT strain, the expression of 36 of the 369secretory pathway genes significantly differed (FDR b0.05) in theHacA-

Fig. 5. Regulatory model of genes in the A. oryzae protein secretory pathway in three pairwiserepresent differentially expressed genes (FDR b 0.05 and ≥2-fold change), which are designaterepresenting pairwise comparisons ofWT vs. HacA-CA (left part), HacA-DE vs. HacA-CA (middlesignificant changed. N, nucleus; ER, endoplasmic reticulum; E, endosome; V, vacuole; G, Golgi.

DE mutant (Fig. 5 and Supplementary Table S7 p2), among which 11were upregulated and 25 were downregulated. The majority of thedownregulated genes belonged to translocation into the ER, proteinfolding/UPR, glycosylation, and COPII-mediated transport processes(Fig. 5). The expression levels of a number of chaperone-encodinggenes, such as AHA1, STI1, and YDJ1, were also upregulated in theHacA-DE mutant, as this might facilitate protein folding in the absenceof a functional UPR.

3.6. Constitutive activation and deletion of hacA downregulates amylolyticenzyme expression

A. oryzae secretes a large amount of enzymes, include Taka-amylase(amyA-C), glucoamylase (glaA), andα-glucosidase (agdA), which are in-volved in the degradation of starch, and the expression of these amylo-lytic enzymes is mediated by the transcription factor AmyR (Kanemoriet al., 1999). However, under ER stress, the expression of these amylo-lytic enzymes, especially of Taka-amylase, is downregulated, which istermed the repression under secretion stress (RESS) mechanism(Wang et al., 2010). Our RNA-seq data revealed that the expression ofstarch-degrading enzyme-encoding genes amyA (298.3-fold), amyB(263.1-fold), amyC (271.5-fold), glaA (135.1-fold), agdB (42.7-fold),and agdA (19.7-fold)was severely downregulated in theHacA-CA strain(Table 2). In addition, the expression of amyRwas also downregulated(8-fold) in the HacA-CA strain (Table 2). We speculated that HacA acti-vation leads to the inactivation of AmyR, resulting in the downregula-tion of the AmyR regulons, which indicated that the repression ofAmyR regulon-related enzymes and the AmyR regulon under RESS isconsistent with these phenomena in A. niger (Carvalho et al., 2012).However, surprisingly, in the HacA-DE strain, the expression of amyA(4.6-fold), amyB (4.7-fold), amyC (4.8-fold), glaA (4.9-fold), and agdB(42.7-fold), but not agdA (1.6-fold), was downregulated significantly,while amyR was not differentially expressed significantly (1.4-folddownregulation) (Table 2). This might be the reason that the disruption

comparison datasets: WT vs. HacA-CA, HacA-DE vs. HacA-CA, and WT vs. HacA-DE. Boxesd by the names of their homologues in S. cerevisiae. Each box is separated into three parts,part), andWT vs. HacA-DE (right part). Red, upregulated; green, downregulated; gray, no

Table 2Expression values of genes involved in starch metabolism.

Gene ID Gene name Description WT. vs. HacA-CA WT. vs. HacA-DE

Fold change FDR Fold change FDR

AO090003001208 amyR Transcription factor of starch utilization amyR- A.oryzae −8.0 2.60E − 120 −1.4a 7.32E − 09AO090003001591 amyA Alpha-amylase amyA - A.oryzae −298.3 0 −4.6 0AO090120000196 amyB Alpha-amylase amyB - A.oryzae −263.1 0 −4.7 0AO090023000944 amyC Alpha-amylase amyC - A.oryzae −271.5 0 −4.8 0AO090010000746 glaA Glucoamylase glaA - A.oryzae −135.1 0 −4.9 0AO090038000471 agdB Ortholog of alpha-glucosidase agdB - A. nidulans −42.7 1.52E − 201 −42.7 2.88E − 196AO090003001209 agdA Alpha-glucosidase agdA - A.oryzae −19.7 0 −1.6a 1.05E − 39

a Not significantly differentially expressed.

151B. Zhou et al. / Gene 593 (2016) 143–153

of hacA, which resulted in a failure to trigger the UPR under situationsthat demand increased secretory capacity, would be expected to impairsecretion. As a result, the unresolved ER stress triggered the RESS,whichled to reduced amylolytic enzyme secretion and an alteration in theoverall secretory profile, which is similar to results described in A.fumigatus (Richie et al., 2009) and N. crassa (Fan et al., 2015).

4. Discussion

HacA is the master regulator of the UPR, which is a key cellular re-sponse for homeostatic adaptation to ER stress. Here, we examinedthe role of the hacA gene and the UPR control mechanisms of A. oryzaeby constructing a hacA knockout strain (HacA-DE) and a strain that ex-presses constitutively active HacA (HacA-CA). In an earlier study,Tanaka et al. (2015) obtained an A. oryzae hacA disruption mutantonly as a heterokaryon. However, in this study, we successfully generat-ed a homokaryotic hacA disruption mutant derived from a ΔpyrG strain(niaD−; ΔpyrG; Δku70::ptrA), which suggests that hacA is not essentialfor the growth of A. oryzae. We found that hyphal growth and sporula-tion were impaired in both the HacA-DE and HacA-CA strains, even incomplete medium (DPY) (Fig. 1a–b and Supplementary Figs. S8–S9).Similarly, the loss of hac1/hacA resulted in a drastic growth defect andnearly normal growth in A. niger (Carvalho et al., 2010) and A. fumigatus(Feng et al., 2011), respectively, whereas it did not influence vegetativegrowth in yeasts and dimorphic fungi like S. cerevisiae, Candida albicans,Cryptococcus neoformans, and Ustilago maydis under non-stress condi-tions (Heimel, 2015). In A. niger, expression of the intron-less versionof hacA mRNA (hacACA) resulted in a reduced growth rate, and thecells were unable to grow on a medium containing starch as the solecarbon source, but the deletion of hacA in A. niger had a profound effecton fungal growth and morphology, resulting in smaller and more com-pact colonies that barely formed conidia (Carvalho et al., 2012). In ourstudy, compared with the HacA-DE mutant, the growth rate impair-ment was more pronounced for the HacA-CA mutant, both in liquidand on solid media (Fig. 1a, Supplementary Figs. S8–S10). These resultssuggest that no a priori conclusions can be reached regarding the phe-notype of a hacA deletionmutant or that resulting from the overexpres-sion of constitutively active HacA in fungi, even among closely relatedspecies. We assume that HacA must maintain a threshold activitylevel, or deleterious effects, such as an inhibition of cell growth, willoccur.

To gain insights into the impact of constitutively active HacA on theUPR mechanism of A. oryzae, we compared the HacA-CA/WT datasetwith the HacA-CA/HacA-DE dataset. In the HacA-CA/HacA-DE dataset,the HacA-DE strain served as the “blank” background. The tran-scriptome data showed that the constitutive activation of HacA resultedin the induction of genes that are associated with the protein secretorypathway (Supplementary Tables S5 p2 and S6 p2). Our data are consis-tent with previous UPR-related studies in fungal and mammalian cells,in which many secretory functions were upregulated, either directlyor indirectly, by HacA (Travers et al., 2000; Lee et al., 2003; Arvas etal., 2006; Carvalho et al., 2012). HacA is the master regulator of theUPR, and deleting this gene has a great effect on the protein secretory

pathway. In this study, the expression of 19 secretory pathway compo-nent-encoding genes, which are involved in protein folding/UPR, glyco-sylation and COPII-mediated transport processes, was downregulatedin the HacA-DE strain (Fig. 5), in contrast to 78 secretory pathway com-ponent-encoding genes (Fig. 5) that were upregulated in the HacA-CAstrain. This result suggests that in our study the canonical Ire1-HacAiUPR directs a pattern of gene expression that can be broadly dividedinto a basal response and an inducible response, with the basal UPRconstituting 17 of the 78 genes and the inducible UPR representingthe remaining 61; the basal UPR predominates in the absence of ERstress and the inducible UPR dominates under acute ER stress. Inyeast, unfolded proteins directly activate IRE1p by binding to its ERlumen-facing amino-terminal domain. Exon-exon ligation is mediatedby the tRNA ligase Rlg1p/Trl1p, and it generates a processed hac1mRNA (hac1i), which encodes the active transcription factor that or-chestrates the UPR (Heimel, 2015). ER stress triggers the dissociationof Kar2p/BiP from IRE1p, resulting in IRE1p activation, and consequenttransmission of the signal from the ER to the nucleus (Okamura et al.,2000). In earlier studies, ER stress was induced by DTT (Wang et al.,2010) or overexpressing amyB (Liu et al., 2014) in A. oryzae, whichresulted in the upregulation of Kar2/BiP expression, while IRE1(AO090005000934) expression was unchanged. In the present study,compared with the WT strain, Kar2/BiP expression was induced andIRE1 expression was unchanged in the HacA-CA strain, while Kar2/BiPexpression was reduced and IRE1 expression was unchanged inthe HacA-DE strain. Compared with the WT strain, the TRL1(AO090012000873) expression level was higher in the HacA-CA(log2(FC) = 1.40, FDR = 2.36E − 39) and HacA-DE (log2(FC) = 3.20,FDR = 0) strains. In addition, the expression of the protein folding/UPR genes CPR6/CPR7 and YDJ1 and the protein degradation geneYTA12 was upregulated in the HacA-CA and HacA-DE strains, whilethe expression of the vacuolar protein sorting genes AMS1 and VPS70and the cellular export and secretion genes INO1 and NCE102 wasdownregulated. The reason for these differences in expression remainunclear and they require further investigation.

For a more condition-independent view regarding how the A.oryzae transcriptome ensures high-level secretion, we comparedsix transcriptome datasets from three independent studies per-formed on A. oryzae: i) a comparison of the HacA transcriptomethat reflected the permanent activation of the UPR in DPY culturesof the HacA-CA and WT strains, and the HacA-CA and HacA-DEstrains (this study), ii) the α-amylase overexpression transcriptomeobtained from comparisons of the batch fermentations of strainsCF1.1 and A1560, A16 and A1560, and CF32 and A1560(Liu et al.,2014), iii) the UPR transcriptome obtained from an A. oryzae CD culturethat was stressedwith DTT (Wang et al., 2010). This analysis uncovered28 geneswhose transcript levels were commonlymodified under all sixsecretion stress conditions: the expression of 19 geneswas upregulated,and the expression of nine genes was downregulated (SupplementaryTable S8). Thus, the genes from this dataset probably represent genesthat are crucial for coping with stress conditions that target the se-cretion machinery. This set of genes includes ER chaperones andfoldases (KAR2 (AO090003000257), CNE1 (AO090009000313),

152 B. Zhou et al. / Gene 593 (2016) 143–153

ERJ5/SCJ1 (AO090011000874), ERO1 (AO090005000630), HUT1(AO090009000400), JEM1 (AO090020000010), LHS1 (AO090005001222), MPD1 (AO090003000829), and PDI1 (AO090001000733)),genes that are important for the translocation of secretory proteinsinto the ER (SEC62 (AO090026000361) and SEC63 (AO090005001238)), and a gene that is important for protein glycosylation (ALG5(AO090701000739)). Fascinatingly, the expression of two genes thatare involvedwith calcium transport, AO090003000051 encoding a pro-tein with calcium-transporting ATPase activity that localizes to the nu-clear envelope, and AO090701000104 encoding a protein that has apredicted role in the regulation of store-operated calcium entry and isan integral component of the ER, was consistently up-regulated underall six conditions. In eukaryotes, chaperones, such as ClxA/CNE1, needcalcium as a co-factor (Wang et al., 2003); hence, enhanced expressionof ClxA/CNE1 requires higher calcium concentrations in the ER, and cal-cium is of general importance for vesicle fusion and the function of theER andGolgi (Stojilkovic, 2005; Dolman and Tepikin, 2006). Hence, cellshave to mobilize calcium from internal or external stores to ensurehigher fluxes through the secretory pathway.

RESS is another important regulatory mechanism during ER stressconditions. It has been reported that RESS represses the transcriptionof secretory protein-encoding genes under ER stress conditions in fila-mentous fungi and plants, although it is absent from S. cerevisiae(Heimel, 2015). In A. niger (Carvalho et al., 2012), RESS was activatedby introducing the constitutively active form of HacA, whereas in A.fumigatus (Richie et al., 2009) andN. crassa (Fan et al., 2015), disruptionof the UPR by deleting hacA also triggered RESS. Our transcriptome pro-files showed that the expression of genes that are involved in carbohy-drate metabolism was commonly downregulated in the HacA-CA andHacA-DE strains, compared with that in the WT strain, which mightbe causatively linked to the RESS phenomenon. In our study, the activa-tion of the UPR by introducing the constitutively active form of HacA ledto the downregulated expression of genes encoding starch-degradingenzymes, including amyA-C, glaA, agdA, and agdB, but, surprisingly, thedisruption of the UPR in the HacA-DE strain also led to the downregula-tion of amyA-C, glaA, and agdB expression (Table 2). The qRT-PCRanalysis showed that the expression levels of amyB were lower in theHacA-CA and HacA-DE strains, especially the HacA-CA strain (Fig. 1c).This result might explain why the growth of the HacA-CA and HacA-DE strainswere severely affectedwhen theywere grown onmedia con-taining starch as the sole carbon source (Fig. 1a and Supplementary Fig.S8). The transcription of these amylolytic enzymes ismediated by AmyR(Kanemori et al., 1999). In addition, amyR expressionwas downregulat-ed in the HacA-CA strain, whereas amyR was not differentiallyexpressed in the HacA-DE strain. Carvalho et al. (2012) speculatedthat HacA activation could lead to the inactivation of AmyR, therebyexplaining the resulting downregulation of AmyR target genes byRESS. In our study,we also found that the expression of amyR and its tar-get genes was downregulated in the HacA-CA strain, while amyR ex-pression was unchanged and the expression of its targeted genes wasdownregulated in the HacA-DE strain (Table 2). In our previous work(Zhou et al., 2015), we used the amyB promoter to control the expres-sion of the Rhizomucor miehei lipase gene as a model to investigate theRESSmechanism, and we found that an octameric sequence was essen-tial for the downregulated transcription of amyB under secretion stress.These results suggest that both the inactivation of AmyR and theoctameric sequence contributed to the downregulation of expressionof starch-degrading enzyme-encoding genes under ER stress.

The PERK-mediated phosphorylation of the eukaryotic translationinitiation factor-2 alpha (eIF2α) is an adaptive response in higher eu-karyotes to lower protein translation and protein import into the ER.In contrast to most mRNAs, translation of the mRNA encoding theGcn4p homologue ATF4 is induced when eIF2α is phosphorylated(Heimel, 2015). eIF2α is also required for the translation of selectivemRNAs, such as that encoding activating transcription factor-4 (ATF4)(Vattem and Wek, 2004). ATF4 is involved in the regulation of UPR

genes that are involved in ERAD, metabolism, and apoptosis (Fels andKoumenis, 2006). Gcn4p and CpcA are the ATF4 homologues of S.cerevisiae and filamentous fungi, respectively. Both S. cerevisiae and A.niger lack an obvious PERK homologue. Gcn2p phosphorylates eIF2α,leading to a global reduction of protein synthesis and the stimulationof Gcn4p translation, which has been shown to control amino acidbiosynthesis (Carvalho et al., 2012). In our transcriptome analysis,a gcn2 homologue (AO090701000211) and a cpcA homologue(AO090009000459) were not differentially expressed in the HacA-CAand HacA-DE strains, compared with theWT strain. However, our tran-scriptome profiles showed that the major categories of significantly af-fected geneswere related to amino acidmetabolism in these two strains(Figs. 3 and 4; Supplementary Tables S5 and S6). Hence, we concludethat the amino acid biosynthesis and amino acid metabolism in theHacA-CA and HacA-DE strains might not be activated by CpcA. Thismight be due to the slower growth of these strains at the time of har-vest, or by feedback from ER stress.

Our transcriptome profiles also showed that the expression of genesthat are involved in glycerolipid, fatty acid, alpha-linolenic acid, and ar-achidonic acid metabolism, as well as synthesis and degradation of ke-tone bodies, was commonly downregulated in both the HacA-CA andHacA-DE strains, suggesting that they have potential defects in mem-brane homeostasis and nutrient levels.

5. Conclusions

Here, we successfully generated a homokaryotic hacA disruptionmutant (HacA-DE) and a strain that constitutively expresses an activat-ed form of HacA (HacA-CA). Growth and phenotypic profiles demon-strated that hyphal growth and sporulation were impaired in theHacA-DE and HacA-CA strains, and the growth rate impairment wasmore pronounced for the HacA-CA strain. The combination of a geneti-cally defined, constitutively activated HacA mutant and a hacA disrup-tion mutant have provided a solid basis for a genome-wide expressionanalysis to study the response of A. oryzae toward ER stress. The resultsindicate that the differentially expressed genes in these strains aremainly involved in the protein secretory pathway, amino acid metabo-lism, lipid metabolism, and carbohydrate metabolism. Additionally, wecompared the transcriptome of the HacA-CA strain with four other rel-evant transcriptomes of A. oryzae. Overall, 28 genes were found to haveeither elevated (19 genes) or lowered (nine genes) transcript levelsunder all conditions that were examined, thus defining the core set ofgenes that are important for ensuring high protein traffic through thesecretory pathway. Furthermore, the constitutive expression of activat-ed HacA and the disruption hacA both had a negative effect on the ex-pression and, consequently, the production of extracellular enzymes.We suggest that the function of HacA is important not only for UPR in-duction, but also for growth and fungal physiology, as it serves to reducesecretion stress in A. oryzae. Therefore, an increased understanding ofthe physiological role of the UPR during fungal development will helpto reveal how cells cope with a highly active UPR without inducing un-wanted side effects.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.gene.2016.08.018.

Acknowledgements

We wish to acknowledge the financial support by the State 863Project (grant no2014AA021304), the Science andTechnology PlanningProject of Guangdong Province (grant nos. 2013B010404007,2013B090800003, and 2016A050503016), the Fundamental ResearchFunds for the Central Universities (grant nos. 2015ZP032 and2015ZZ040), and the Science and Technology Planning Project ofGuangzhou City (grant no. 201510010191).

153B. Zhou et al. / Gene 593 (2016) 143–153

References

Arvas, M., Pakula, T., Lanthaler, K., Saloheimo, M., Valkonen, M., Suortti, T., Robson, G.,Penttila, M., 2006. Common features and interesting differences in transcriptional re-sponses to secretion stress in the fungi Trichoderma reesei and Saccharomycescerevisiae. BMC Genomics 7, 32.

Benz, J.P., Chau, B.H., Zheng, D., Bauer, S., Glass, N.L., Somerville, C.R., 2014. A comparativesystems analysis of polysaccharide-elicited responses in Neurospora crassa revealscarbon source-specific cellular adaptations. Mol. Microbiol. 91, 275–299.

Bindea, G., Mlecnik, B., Hackl, H., Charoentong, P., Tosolini, M., Kirilovsky, A., Fridman,W.H., Pages, F., Trajanoski, Z., Galon, J., 2009. ClueGO: a Cytoscape plug-in to decipherfunctionally grouped gene ontology and pathway annotation networks. Bioinformat-ics 25, 1091–1093.

Carvalho, N.D., Arentshorst, M., Jin Kwon, M., Meyer, V., Ram, A.F., 2010. Expanding theku70 toolbox for filamentous fungi: establishment of complementation vectors and re-cipient strains for advanced gene analyses. Appl. Microbiol. Biotechnol. 87, 1463–1473.

Carvalho, N.D.S.P., Jorgensen, T.R., Arentshorst, M., Nitsche, B.M., van den Hondel,C.A.M.J.J., Archer, D.B., Ram, A.F.J., 2012. Genome-wide expression analysis upon con-stitutive activation of the HacA bZIP transcription factor in Aspergillus niger reveals acoordinated cellular response to counteract ER stress. BMC Genomics 13, 350.

Dolman, N.J., Tepikin, A.V., 2006. Calcium gradients and the Golgi. Cell Calcium 40,505–512.

Fan, F.Y., Ma, G.L., Li, J.G., Liu, Q., Benz, J.P., Tian, C.G., Ma, Y.H., 2015. Genome-wide anal-ysis of the endoplasmic reticulum stress response during lignocellulase production inNeurospora crassa. Biotechnol. Biofuels 8, 66.

Fels, D.R., Koumenis, C., 2006. The PERK/eIF2alpha/ATF4module of the UPR in hypoxia re-sistance and tumor growth. Cancer Biol. Ther. 5, 723–728.

Feng, X., Krishnan, K., Richie, D.L., Aimanianda, V., Hartl, L., Grahl, N., Powers-Fletcher,M.V., Zhang, M., Fuller, K.K., Nierman, W.C., Lu, L.J., Latge, J.P., Woollett, L., Newman,S.L., Cramer Jr., R.A., Rhodes, J.C., Askew, D.S., 2011. HacA-independent functions ofthe ER stress sensor IreA synergize with the canonical UPR to influence virulencetraits in Aspergillus fumigatus. PLoS Pathog. 7, e1002330.

Gomi, K., Iimura, Y., Hara, S., 1987. Integrative transformation of Aspergillus oryzae with aplasmid containing the Aspergillus nidulans argB gene. Agric. Biol. Chem. 51, 2549–2555.

Guillemette, T., van Peij, N.N.M.E., Goosen, T., Lanthaler, K., Robson, G.D., van den Hondel,C.A.M.J.J., Stam, H., Archer, D.B., 2007. Genomic analysis of the secretion stress re-sponse in the enzyme-producing cell factory Aspergillus niger. BMC Genomics 8, 158.

Heimel, K., 2015. Unfolded protein response in filamentous fungi-implications in biotech-nology. Appl. Microbiol. Biotechnol. 99, 121–132.

Jorgensen, T.R., Goosen, T., den Hondel, C.A.M.J.J., Ram, A.F.J., Iversen, J.J.L., 2009.Transcriptomic comparison of Aspergillus niger growing on two different sugars re-veals coordinated regulation of the secretory pathway. BMC Genomics 10, 44.

Kanehisa, M., Araki, M., Goto, S., Hattori, M., Hirakawa, M., Itoh, M., Katayama, T.,Kawashima, S., Okuda, S., Tokimatsu, T., Yamanishi, Y., 2008. KEGG for linking ge-nomes to life and the environment. Nucleic Acids Res. 36, D480–D484.

Kanemori, Y., Gomi, K., Kitamoto, K., Kumagai, C., Tamura, G., 1999. Insertion analysis ofputative functional elements in the promoter region of the Aspergillus oryzae Taka-amylase A gene (amyB) using a heterologous Aspergillus nidulans amdS-lacZ fusiongene system. Biosci. Biotechnol. Biochem. 63, 180–183.

Kwon, M.J., Jorgensen, T.R., Nitsche, B.M., Arentshorst, M., Park, J., Ram, A.F., Meyer, V.,2012. The transcriptomic fingerprint of glucoamylase over-expression in Aspergillusniger. BMC Genomics 13, 701.

Lee, A.H., Iwakoshi, N.N., Glimcher, L.H., 2003. XBP-1 regulates a subset of endoplasmic re-ticulum resident chaperone genes in the unfolded protein response. Mol. Cell. Biol.23, 7448–7459.

Liu, L., Feizi, A., Osterlund, T., Hjort, C., Nielsen, J., 2014. Genome-scale analysis of the high-efficient protein secretion system of Aspergillus oryzae. BMC Syst. Biol. 8, 73.

Maruyama, J., Kitamoto, K., 2008. Multiple gene disruptions by marker recycling withhighly efficient gene-targeting background (ΔligD) in Aspergillus oryzae. Biotechnol.Lett. 30, 1811–1817.

Montenegro-Montero, A., Goity, A., Larrondo, L.F., 2015. The bZIP transcription factorHAC-1 is involved in the unfolded protein response and is necessary for growth oncellulose in Neurospora crassa. PLoS One 10, e0131415.

Moon, H.Y., Cheon, S.A., Kim, H., Agaphonov, M.O., Kwon, O., Oh, D.B., Kim, J.Y., Kang, H.A.,2015. Hansenula polymorpha Hac1p is critical to protein N-glycosylation activitymodulation, as revealed by functional and transcriptomic analyses. Appl. Environ.Microbiol. 81, 6982–6993.

Nevalainen, K.M., Te'o, V.S., Bergquist, P.L., 2005. Heterologous protein expression in fila-mentous fungi. Trends Biotechnol. 23, 468–474.

Okamura, K., Kimata, Y., Higashio, H., Tsuru, A., Kohno, K., 2000. Dissociation of Kar2p/BiPfrom an ER sensory molecule, Ire1p, triggers the unfolded protein response in yeast.Biochem. Biophys. Res. Commun. 279, 445–450.

Richie, D.L., Hartl, L., Aimanianda, V., Winters, M.S., Fuller, K.K., Miley, M.D., White, S.,McCarthy, J.W., Latge, J.P., Feldmesser, M., Rhodes, J.C., Askew, D.S., 2009. A role forthe unfolded protein response (UPR) in virulence and antifungal susceptibility in As-pergillus fumigatus. PLoS Pathog. 5, e1000258.

Sims, A.H., Gent, M.E., Lanthaler, K., Dunn-Coleman, N.S., Oliver, S.G., Robson, G.D., 2005.Transcriptome analysis of recombinant protein secretion by Aspergillus nidulans andthe unfolded-protein response in vivo. Appl. Environ. Microbiol. 71, 2737–2747.

Stojilkovic, S.S., 2005. Ca2+-regulated exocytosis and SNARE function. Trends Endocrinol.Metab. 16, 81–83.

Storms, R., Zheng, Y., Li, H.S., Sillaots, S., Martinez-Perez, A., Tsang, A., 2005. Plasmid vec-tors for protein production, gene expression andmolecular manipulations in Aspergil-lus niger. Plasmid 53, 191–204.

Tanaka, M., Shintani, T., Gomi, K., 2015. Unfolded protein response is required for Asper-gillus oryzae growth under conditions inducing secretory hydrolytic enzyme produc-tion. Fungal Genet. Biol. 85, 1–6.

Trapnell, C., Williams, B.A., Pertea, G., Mortazavi, A., Kwan, G., van Baren, M.J., Salzberg,S.L., Wold, B.J., Pachter, L., 2010. Transcript assembly and quantification by RNA-Seqreveals unannotated transcripts and isoform switching during cell differentiation.Nat. Biotechnol. 28, 511–515.

Trapnell, C., Roberts, A., Goff, L., Pertea, G., Kim, D., Kelley, D.R., Pimentel, H., Salzberg, S.L.,Rinn, J.L., Pachter, L., 2012. Differential gene and transcript expression analysis ofRNA-seq experiments with TopHat and cufflinks. Nat. Protoc. 7, 562–578.

Travers, K.J., Patil, C.K., Wodicka, L., Lockhart, D.J., Weissman, J.S., Walter, P., 2000. Func-tional and genomic analyses reveal an essential coordination between the unfoldedprotein response and ER-associated degradation. Cell 101, 249–258.

Vattem, K.M., Wek, R.C., 2004. Reinitiation involving upstream ORFs regulates ATF4mRNA translation in mammalian cells. Proc. Natl. Acad. Sci. U. S. A. 101,11269–11274.

Wang, H., Entwistle, J., Morlon, E., Archer, D.B., Peberdy, J.F., Ward, M., Jeenes, D.J., 2003.Isolation and characterisation of a calnexin homologue, clxA, from Aspergillus niger.Mol. Gen. Genomics. 268, 684–691.

Wang, Y., Xue, W., Sims, A.H., Zhao, C., Wang, A., Tang, G., Qin, J., Wang, H., 2008. Isolationof four pepsin-like protease genes from Aspergillus niger and analysis of the effect ofdisruptions on heterologous laccase expression. Fungal Genet. Biol. 45, 17–27.

Wang, B., Guo, G., Wang, C., Lin, Y., Wang, X., Zhao, M., Guo, Y., He, M., Zhang, Y., Pan, L.,2010. Survey of the transcriptome of Aspergillus oryzae via massively parallel mRNAsequencing. Nucleic Acids Res. 38, 5075–5087.

Ward, O.P., Qin,W.M., Dhanjoon, J., Ye, J., Singh, A., 2006. Physiology and biotechnology ofAspergillus. Adv. Appl. Microbiol. 58, 1–75.

Ye, L., Pan, L., 2008. A comparison of the unfolded protein response in solid-statewith submerged cultures of Aspergillus oryzae. Biosci. Biotechnol. Biochem. 72,2998–3001.

Zhang, J., Mao, Z., Xue,W., Li, Y., Tang, G., Wang, A., Zhang, Y.,Wang, H., 2011. Ku80 gene isrelated to non-homologous end-joining and genome stability in Aspergillus niger.Curr. Microbiol. 62, 1342–1346.

Zhou, B., Wang, C., Wang, B., Li, X., Xiao, J., Pan, L., 2015. Identification of functional cis-el-ements required for repression of the Taka-amylase A gene under secretion stress inAspergillus oryzae. Biotechnol. Lett. 37, 333–341.