A homeodomain-leucine zipper I transcription factor,MeHDZ14,regulates internode elongation and leaf rolling in cassava (Manihot esculenta Crantz)

Xiaoling Yu,Xin Guo,Pingjuan Zhao,Shuxia Lia,,d,Liangping Zou,Wnbin Li,Ziyin Xu,Ming Png,Mngbin Ruana,,d,

a National Key Laboratory for Tropical Crop Breeding,Sanya 572025,Hainan,China

b Key Laboratory of Biology and Genetic Resources of Tropical Crops,Institute of Tropical Bioscience and Biotechnology,Chinese Academy of Tropical Agricultural Sciences,Haikou 571101,Hainan,China

c Key Laboratory for Biology and Genetic Resources of Tropical Crops of Hainan Province,Hainan Institute for Tropical Agricultural Resources,Haikou 571101,Hainan,China

d Sanya Research Institute,Chinese Academy of Tropical Agricultural Sciences,Sanya 572025,Hainan,China

e College of Tropical Crops,Hainan University,Haikou 570228,Hainan,China

Keywords: HD-Zip transcription factor Drought Internode elongation Leaf rolling Cassava

ABSTRACT Drought stress impairs plant growth and other physiological functions.MeHDZ14,a homeodomainleucine zipper I transcription factor,is strongly induced by drought stress in various cassava cultivars.However,the role of MeHDZ14 in cassava growth regulation has remained unclear.Here we report that MeHDZ14 affected plant height,such that a dwarf phenotype and altered internode elongation were observed in transgenic cassava lines. MeHDZ14 was found to negatively regulate the biosynthesis of lignin.Its overexpression resulted in abaxially rolled leaves.The morphogenesis of leaf epidermal cells was inhibited by overexpression of MeHDZ14,with decreased auxin and gibberellin and increased cytokinin contents. MeHDZ14 was found to regulate many drought-responsive genes,including genes involved in cell wall synthesis and expansion.MeHDZ14 bound to the promoter of caffeic acid 3-Omethyltransferase 1(MeCOMT1),acting as a transcriptional repressor of genes involved in cell wall development.MeHDZ14 appears to act as a negative regulator of internode elongation and epidermal cell morphogenesis during cassava leaf development.

1.Introduction

Homeodomain-leucine zipper(HD-Zip)transcription factors are a subset of homeobox genes that are unique to the plant kingdom[1] and contain homeodomains with tightly linked leucine-zipper motifs.The HD-Zip transcription-factor family has been classified into four subfamilies(I-IV)by their molecular structures[1].They have been identified and characterized in the model plant speciesArabidopsisand rice[1].However,none has been functionally characterized in cassava (Manihot esculentaCrantz),a tropical staple food crop.

HD-Zip I transcription factors function in plant responses to abiotic stress and phytohormone signaling.TaHDZipI-5improved drought tolerance in wheat [2].Overexpression ofZmHDZ10conferred both drought and salt resistance in transgenic rice andArabidopsis[3].Overexpression ofMdHB7-likeincreased salinity tolerance in transgenic apple [4].Overexpression ofCaHDZ12improved tolerance to osmotic stress and increased sensitivity to abscisic acid(ABA)in transgenic tobacco and chickpea[5].AtHB12was strongly induced by osmotic stress and ABA treatments [6]and acted as a positive regulator of phosphatase 2C (PP2C),a key component of the ABA signaling pathway inArabidopsis[7].

Abiotic stress-responsible HD-Zip I transcription factors contribute to plant development.During the early development of inflorescence stems,AtHB12inhibits the expression ofgibberellin 20 oxidase1 (GA20OX1),leading to inhibition of stem elongation[8].Leaf growth is controlled by tight coordination of cell division and cell expansion.Endoreduplication,a common process in higher plants during which cells increase their size by increasing their ploidy level,is considered [9] one of the key determinants of cell expansion.AtHB12 acts as a positive regulator of leaf growth by promoting cell expansion and endoreduplication [10].The mechanism by which AtHB12 promotes both cell expansion and endoreduplication is reduction of S6 kinase 1 activity.The reduction of S6 kinase 1 activity reduced cell size and ploidy level as well as down-regulated expression of cell cycle-regulatory proteins in leaves [11].

Various strategies have been developed by cassava cultivars to adapt to drought,such as early growth quiescence and growth reduction [12].Recently [13-17],several drought-responsive cassava genes have been functionally identified and characterized.However,the molecular mechanisms by which cassava cultivars respond to drought remain unclear.In a previous study,we identified a HD-Zip I transcription factor,MeHDZ14,that responds to drought in several cassava cultivars [18].But its role in regulating cell size and growth is unknown.

In this study,we producedMeHDZ14overexpression transgenic cassava lines to investigate its functions in cassava drought response and development.Systematically phenotypic analyses were focused on the regulatory roles ofMeHDZ14in cassava internode elongation and leaf development.The alternation of lignin,gibberellin and auxin biosynthesis pathways were studied in transgenic cassava to declare multiple roles ofMeHDZ14.The down-stream target genes that directly regulated byMeHDZ14were identified to dissect the underlying molecular mechanisms thatMeHDZ14regulating drought responses in cassava.

2.Materials and methods

2.1.Plant materials

Stems of two cassava genotypes(cv.60444 and cv.E1424)were cultured in pots for 90 days under greenhouse conditions(12 h/12 h of light/dark,30 °C/25 °C day/night) at the Institute of Tropical Bioscience and Biotechnology (Haikou,China).Each pot contained 5 kg of well-mixed soil (soil: vermiculite: clay pellets,1:1:1).For production ofMeHDZ14-overexpressing transgenic cassava,friable embryogenic calluses of cassava cv.60444 were transformed usingAgrobacteriumstrain LBA4404 carrying the DNA construct35S:MeHDZ14following Zainuddin et al.[19].The transgenic cassava was identified by Southern blotting as in our previous study [13].

2.2.Drought treatment of cassava seedlings

For drought treatments,water withholding was used to treated 90-d-old seedlings grown under the same greenhouse conditions as two cassava genotypes(cv.60444 and cv.E1424).Continuously watered plants were used as controls (with soil moisture content 35% ± 2%).In each treatment,at least six plants of each genotype were subjected to drought stress for 10 days.Throughout the experiment,soil moisture content in each pot was measured once daily with a moisture sensor.The leaves and stems of droughttreated cassava seedlings were harvested at day 5(at soil moisture content 5.0% ± 1.8%) and day 10 (at soil moisture content 1.5% ±0.5%) during drought stress.Samples harvested from continuously watered seedlings at the same time were used as control.

2.3.Chlorophyll content index and transpiration rate detection

Chlorophyll content index of cassava mature leaves from 180-dold plants (grown in the field at the Wenchang Test Bed of the Institute of Tropical Bioscience and Biotechnology,Hainan,China)was measured with CCM-200 plus chlorophyll meter (OPTISCIENCES,Hudson,NH,USA).Indices from at least 21 mature leaves from wild-type or transgenic cassava lines were recorded.Transpiration rate was measured at 11: 00 and 16: 00 with a Li-6400XT photosynthesis measurement system (LI-COR,Lincoln,NE,USA).Ten mature leaves from the same plants were tested for chlorophyll content.

2.4.Quantitative real-time PCR (qPCR)

Gene expression was quantified by qPCR using gene-specific primers(Table S1).All reactions were performed in triplicate,with SYBR Premix Ex Taq II Kit(Takara,Dalian,Liaoning,China)on a StepOne Real-Time PCR system (ABI,Carlsbad,CA,USA).To quantify the amplified qPCR products,the comparative ΔΔCT method was used.

2.5.Scanning electron microscopy

Apical buds,sixth internodes,and newly expanding and mature leaves were collected from 60-d-old wild-type and transgenic cassava plants grown in pots and were treated following Perera et al.[20].Samples were rinsed with 0.1 mol L-1phosphate buffered saline(PBS,pH 7.4)three times and immediately fixed with 1%tetroxide(OsO4,CAS No.20816-12-0)in 0.1 mol L-1PBS for 2 h at room temperature.The fixed samples were washed with 0.1 mol L-1PBS,followed by dehydration in ethanol and drying with critical point dryer.Samples were attached to metallic stubs using carbon stickers and sputter-coated with gold for 30 s,then observed and imaged using a scanning electron microscope system (Regulus8100,Hitachi,Tokyo,Japan).Cell size in internodes and leaves was estimated with ImageJ 1.53 k (https://imagej.nih.gov/ij).

2.6.Safranin o-fast green staining

Leaf sections were dehydrated by immersion in xylene for 20 min (1),xylene for 20 min (2),100% ethanol for 5 min (1),100%ethanol for 5 min(2),and 75%ethanol for 5 min.The sections were then immersed in safranin O staining solution (#G1031-1,Servicebio,Wuhan,China)for 2 h.They were decolorized by washing with 50%,70%,and 80% ethanol for 15-30 s each and then immersed in plant solid green staining solution (#G1031-2,Servicebio) for 20 s followed by dehydration in 100% ethanol.Finally,the sections were submerged in xylene for 5 min and then sealed with neutral resin and imaged with a microscope system (DS-U3,Nikon,Tokyo,Japan) and photographs were generated by Case-Viewer 2.4 (3DHISTECH,Budapest,Hungary).

2.7.Measurement of phenylpropanoid and endogenous phytohormone contents

Levels of phenylpropanoid compounds and endogenous phytohormones,including salicylic acid (SA),indole-3-acetic acid (IAA),trans-zeatin,cis-zeatin,isopentenyladenine (iP),isopentenyladenosine (iPR),gibberellin acid 1 (GA1),gibberellin acid 1(GA3),gibberellin acid 4 (GA4),and gibberellin acid 7 (GA7),were measured by extraction and separation using liquid chromatography electrospray ionization tandem mass spectrometry (LC-ESIMS/MS)following Ross et al.[21].Approximately 100 mg of mixed leaf sample was extracted with 1.5 mL of methanol formic acid solution (methanol: formic acid: water 7.8: 0.2: 2).Each sample consisted of homogenized mature leaf tissue (from third to fifth leaves)from three plants of each line and treated as one biological replicate.Three replicates were used.

2.8.Phytohormone treatments with transgenic cassava in vitro plantlets

For plant hormone treatments,apical shoots～1 cm long were cut from 30-d-old transgenic cassava and wild-typein vitroplantlets and cultured on cassava basic medium containing GA3(10 μmol L-1and 25 μmol L-1) or IAA (10 μmol L-1and 25 μmol L-1).Plantlets were cultured at 26 °C under 12 h light/12 h dark in a versatile environmental test chamber for 14 d.For GA3 treatments,internode length (of first and second internodes) and root numbers of at least six plants from each line were measured.For IAA treatments,root numbers of at least six plants from each line were measured.

2.9.Lignin content measurement

Total cell wall was isolated from 200 mg of stem tissues from wild-type and transgenic cassava following Kong et al.[22].Total lignin from～2 mg of isolated cell wall was dissolved in 100 μL acetyl bromide solution(25%)and diluted to 2 mL with acetic acid.After centrifugation,the supernatant was measured with a microplate reader in a measurement system at 280 nm.Cell-wall content of lignin (μg mg-1) was calculated as

where ABS is absorption value,Coeff is absorption coefficient,optical distance is 0.539 cm,volume is 2 mL,and weight is weight of isolated cell wall.

2.10.RNA-seq and transcriptome analysis

Three mature leaves(the third,fourth,and fifth leaves from the apical bud,7-15 days after their emergence) from 60-d-old wildtype and transgenic pot-grown cassava plants were collected for RNA sequencing.RNA-seq was performed by the Beijing Genomics Institute(Shenzhen,Guangdong,China)and three biological replicates were used for each sample.Reads were aligned to the cassava genome (https://phytozome-next.jgi.doe.gov/Manihot esculenta,v8.1)with HISAT 2.1.0[23].Using the DEGseq method[24],differentially expressed genes (DEGs) were defined using the following criteria: |log2(fold change)| ＞ 1 and adjustedP-value (Qvalue) ≤0.001.Gene Ontology (GO) enrichment analysis was performed according to the definition of DEGs in the Goseq R software package [25].GO terms with |log2(fold change)| ＞ 0 andQvalue ≤0.001 were considered enriched.

2.11.Yeast one-hybrid assay

A DNA fragment containing three tandem repeat HD-Zipbinding elements (5′-ATTTAATTAAATTA -3′) was synthesized and inserted into thepAbAivector.The resulting 3×HDZBelement: pAbAiwas transformed into yeast strain Y1HGold to generate bait-specific reporter strains.The full-length cDNA sequence ofMeHDZ14was cloned into thepGADT7vector.MeHDZ14:pGADT7was transformed into bait-specific reporter strains.The binding between MeHDZ14 and the HDZB-element was determined by bacterial plaque formation of transformed yeasts on SD/-Leu/AbA(200 ng mL-1) medium.

2.12.Electrophoretic mobility shift assay (EMSA)

The full-length cDNA sequence ofMeHDZ14was cloned into thePATX-SUMOexpression vector and transformed into the RosettaEscherichia colistrain.Bacterial cells were harvested and lysed tin buffer (PBS pH 7.5+10% glycerol).The recombinant protein was purified by affinity vs.His-Tag on Ni resin (Roche,Basel,Switzerland).Probes were designed and labeled with biotin with a probe labeling kit (#GS008,Beyotime,Shanghai,China).Then 2 μg of the recombinant protein MeHDZ14 was mixed with theMeCOMT1-P1-biotin,MeCOMT1-P1,andMeCOMT1-P1m probes respectively as designed at 25 °C for 20 min.Western blotting with streptavidin-HRP conjugate was used to detect DNA shift bands using a chemiluminescent EMSA kit (#GS009,Beyotime).Images were acquired with an ImageQuant LAS4000mini chemiluminescence imaging system (GE Healthcare Life Science,Boston,MA,USA).

2.13.Dual luciferase (LUC/REN) reporter assays

Reporter plasmids were constructed by cloning the promoters of candidate genes separately intopGreenII0800-lucto formpromoter: pGreenII0800-lucconstructs.The35S: MeHDZ14and35S:GFPconstructs were used as effectors in the analysis.Assays were performed following Hellens et al.[26].LUC/REN relative value was measured by using the Dual-Luciferase Reporter Assay System(#RG028,Beyotime).

2.14.Statistical analyses

One-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons test was performed using GraphPad Prism version 9.00 for Windows (GraphPad Software,La Jolla,CA,USA,https://www.graphpad.com).

3.Results

3.1.Overexpression of MeHDZ14 caused dwarfing in transgenic cassava

Drought reduced the growth of cassava cultivars cv.60444 and cv.E1424 (Fig.1A,B).Drought stress significantly reduced the growth of cassava cv.60444,while only slightly reducing that of cassava cv.E1424 (Fig.1C).Drought drastically reduced internode lengths in cv.60444,while not affecting the length of internode in E1424 (Fig.1D).We previously identified a HD-Zip I transcription factor [18],MeHDZ14,which is homologous to AtHB12 (Fig.S1),is involved in drought stress responses in several cassava cultivars.Expression ofMeHDZ14in stems and leaves was strongly induced by drought in cv.60444 but only slightly induced in cv.E1424(Fig.1E,F).TheMeHDZ14promoter showed high transcriptional activity in both stems and leaves (Fig.S2).The presence ofMeHDZ14was associated with reduction of growth in cassava cultivars under drought stress.

Fig.1. MeHDZ14 overexpression affected growth of transgenic cassava seedlings.Drought stress reduced growth of cassava cv.60444(A)and cv.E1424(B).(C)Drought stress affected plant height of both cultivars.(D) Drought stress reduced relative internode length of cv.60444 and cv.E1424.Expression analysis of MeHDZ14 in stems (E) and leaves(F)of two cassava cultivars under drought stress(n=3).DS-5d,drought stress for 5 days;DS-10d,drought stress for 10 days.(G)Seedlings of wild-type and transgenic cassava were grown in pots under normal conditions.WT,wild type(cv.60444);OE#5 and OE#7 represent MeHDZ14-overexpressing transgenic cassava lines#5 and line#7,respectively.12 DAP,12 d after planting in pot;60 DAP,60 days after planting in pot.Scale bars,5 cm.(H)Leaf weight,(I)plant height,and(J)crown diameter of wild-type and transgenic cassava (n=9).(K) qPCR analysis of MeHDZ14 in transgenic cassava.Error bars represent ± SD and different letters indicate differences at P ＜0.05 (ANOVA,Tukey’s multiple comparisons test).

Eight independent transgenic lines were identified by Southern blotting (Fig.S3A),of which two (OE#5 and OE#7) were selected for further study.No significant phenotypic difference between the wild-type and transgenic cassava was observed inin vitroplantlets,even those treated with PEG and ABA(Fig.S3B).A dwarf phenotype was observed only in OE#5(Fig.1G).Leaf weight,plant height,and crown diameter were decreased in OE#5 in comparison with both the wild type and OE#7 (Fig.1H-J).Expression ofMeHDZ14in OE#5 was higher than that in OE#7 (Fig.1K).

3.2.MeHDZ14 reduces the elongation of internode and lignin biosynthesis in transgenic cassava

To test whetherMeHDZ14influenced internode elongation,the length of 12 internodes (from the fifth to the 16th) in wild-type and transgenic cassava seedlings was measured (Fig.2A).Overexpression ofMeHDZ14reduced internode length in OE#5 but did not affect that in OE#7(Fig.2B).When the morphology of the epidermal cells of a selected internode (the ninth) was examined by scanning electron microscopy (Fig.2C),the shape of epidermal cells was substantially altered by overexpression ofMeHDZ14,particularly in OE#5(Fig.2D-F).The length and area of internode epidermal cells were also reduced (Fig.2G,H).Thus,the reduced length of cells contributed to the shorter internodes of OE#5.Given thatMeHDZ14overexpression did not affect the number of nodes in transgenic cassava plants (Fig.2I),the shorter stem is likely to have been caused by shorter internodes in OE#5.Overexpression ofMeHDZ14also reduced the thickness of epidermal cells in OE#5 (Fig.2J-L).

Fig.2. MeHDZ14 overexpression affected elongation of internode in transgenic cassava.(A)Diagram indicates internode position of 60-DAP cassava seedling.WT,wild type(cv.60444);OE#5 and OE#7 represent MeHDZ14 overexpressing transgenic cassava lines#5 and#7,respectively.(B)Internode length of wild-type and transgenic cassava.12 internodes of each plant were measured(n=5).(C)Internodes of wild-type and transgenic cassava.Photographs were acquired by scanning electron microscope.Scale bars,1 mm.Internode epidermal cells of wild-type cassava(D),transgenic cassava OE#5(E),and OE#7(F).The added green line marks outline of several epidermal cells.Scale bars,50 μm in (D-F).(G) Length of internode epidermal cells of wild-type and transgenic cassava.(H) Area of internode epidermal cells of wild-type and transgenic cassava.(I)Node numbers of wild-type and transgenic 60-DAP seedlings.An internode transection of wild-type (J),transgenic cassava OE#5 (K),and OE#7 (L).Asterisks indicate epidermal cells.Scale bars,50 μm in (J-L).Error bars represent ± SD and different letters indicate differences at P ＜0.05 (ANOVA,Tukey’s multiple comparisons test).

Lignin functions in the structure and repair mechanism of the plant cell wall[27].Pathways of lignin biosynthesis in plants have been reviewed [28-30].Given that overexpression ofMeHDZ14causes a shape alteration in internode epidermal cells,it may be speculated thatMeHDZ14is involved in the biosynthesis of lignin in cassava.The contents of key components of the phenylpropanoid pathway in wild-type and transgenic cassava were measured (Fig.S4).The contents of lignin and phenylalanine were reduced in transgenic cassava,in particular in OE#5 (Fig.S4).The content of SA was also greatly reduced in transgenic cassava(Fig.S4).These findings suggest thatMeHDZ14impairs lignin biosynthesis by repressing phenylalanine synthesis in cassava.

3.3.MeHDZ14 overexpression caused rolled leaves in transgenic cassava

Leaf rolling is an important agronomic trait in crop breeding[31].HD-Zip transcription factors are essential in the control of leaf rolling in rice [32].MeHDZ14overexpression resulted in major abaxially rolled leaves in transgenic cassava lines (Fig.3A).Leaf rolling was observed only in mature and not in newly expanding leaves (Fig.3A,B).The leaf-rolling index (LRI) indicated that leaf rolling occurred with the development of leaves in both OE#5 and OE#7 (Fig.3C).But overexpression ofMeHDZ14reduced leaf blade length only in OE#5(Fig.3D,E).Thus,MeHDZ14induced leaf rolling in transgenic cassava plants during leaf development.The HD-Zip transcription factor caused leaf rolling and reduced leaf transpiration rate in rice [33].MeHDZ14overexpression greatly reduced the chlorophyll content index of cassava leaves (Fig.3F)but not leaf transpiration rate (Fig.3G).

Fig.3. MeHDZ14 overexpression enhanced leaf rolling in transgenic cassava.(A)Overexpression of MeHDZ14 caused adaxially rolled leaves in transgenic cassava.WT,wild type(cv.60444);OE#5 and OE#7 represent MeHDZ14-overexpressing transgenic cassava lines#5 and#7,respectively.N,newly expanding leaves;M,mature leaves.Scale bars,5 cm.(B)Transections of newly expanding and mature leaves from wild-type and transgenic cassava.Ad,adaxial epidermis;ab,abaxial epidermis.Scale bars,1 mm.(C)Leaf rolling index of newly expanding and mature leaves from wild-type or transgenic cassava(n=9).(D)Mature leaves of wild-type and transgenic cassava.Scale bar,5 cm.(E)Blade length of mature leaves of wild-type and transgenic cassava.(F)Chlorophyll content index of mature leaves from wild-type and transgenic cassava.(G)Transpiration rate of mature leaves from wild-type and transgenic cassava(n=10).Error bars represent±SD,different letters indicate differences at P ＜0.05(ANOVA,Tukey’s multiple comparisons test).

3.4.MeHDZ14 regulates morphogenesis of leaf epidermis in transgenic cassava

Scanning electron microscopy showed no difference in cell morphogenesis of leaf primordia between wild-type and transgenic cassava(Fig.4A,B).Severe cell shrinkage was observed in the epidermis of both newly expanding and mature leaves from OE#5 transgenic cassava,but only in mature leaf epidermis of OE#7(Fig.4C).But guard-cell morphology was not affected byMeHDZ14overexpression (Fig.4D).In transverse sections of mature leaves,the shape of epidermal cells was altered in transgenic cassava,especially in OE#5 (Fig.4E).No change was seen in the shape of epidermal cells of newly expanding leaves from either transgenic line (Fig.S5).

Fig.4. MeHDZ14 overexpression inhibited cell expansion during leaf development in transgenic cassava.(A)P1 leaf primordium,(B)P2 leaf primordium,(C)newly expanding leaf,and(D)mature leaf of wild-type and transgenic cassava.Photographs were acquired by scanning electron microscope.Scale bars,200 μm in(A and B),and 40 μm in(C and D).(E)Transections of mature leaves from wild-type and transgenic cassava.WT,wild type(cv.60444);OE#5 and OE#7 represent MeHDZ14-overexpressing transgenic cassava lines#5 and line#7,respectively.Ad,adaxial epidermis;ab,abaxial epidermis.Scale bars,50 μm.Red asterisks indicate the cells that shape was changed significantly compared to wild-type.

3.5.MeHDZ14 affects phytohormone sensitivity and contents in transgenic cassava

Phytohormones are essential growth regulators for cell morphogenesis in plants[34].Application of 10 μmol L-1of gibberellin(GA3)induced much longer internodes in wild-type than in transgenic cassava(Fig.5A,B),whereas application of 25 μmol L-1GA3 increased internode length only in transgenic cassava (Fig.5A,B).GA3 application inhibited root development in both wild-type and transgenic cassava both (Fig.5A,C).But application of IAA induced more roots in transgenic cassava than in the wild type(Fig.5A,D).

Fig.5. MeHDZ14 overexpression affected phytohormones sensitivity in transgenic cassava.(A)In vitro plantlets of wild-type and MeHDZ14-OE transgenic cassava were grown on cassava basic medium (control) or on medium containing GA3 or IAA.Scale bars,10 mm.(B) Length of two internodes (first and second) of wild-type and MeHDZ14-OE transgenic cassava plantlets under GA3 treatment.(C)Root numbers of wild-type and MeHDZ14-OE transgenic cassava plantlets under GA3 treatment.(D)Root numbers of wild-type and MeHDZ14-OE transgenic cassava plantlets under IAA treatment.Error bars represent ± SD (n= 5),different letters indicate differences at P ＜0.05 (ANOVA,Tukey’s multiple comparisons test).

The content of IAA was greatly decreased in leaves of transgenic cassava,especially in those of OE#5(Fig.S6A).In contrast,the contests of the cytokinintrans-zeatin,cis-zeatin,iP,and iPR were increased in leaves of transgenic cassava,especially in OE#5,relative to wild-type cassava (Fig.S6B-E).Contents of the gibberellins GA1,GA3,GA4,and GA7 were significantly decreased in leaves of OE#5 (Fig.S6F-I).

3.6.MeHDZ14 regulates drought-responsive and cell-wall-related genes in cassava

In leaves ofMeHDZ14-overexpressing cassava,6495 genes were up-regulated and 6138 down-regulated (Fig.S7A).In leaves of wild-type cassava,drought stress up-regulated 6143 genes and down-regulated 6434 genes (Fig.S7B).MeHDZ14regulated more than half of drought-responsive genes (Fig.S7C).GO enrichment analysis revealed that many GO terms associated withMeHDZ14regulation were identical to those associated with drought response (Fig.S7D,E).MostMeHDZ14drought-responsive genes were enriched in the GO terms cell,cell part,intracellular,and intracellular parts (Fig.S7D,E).

Phytohormones are the initial trigger of morphogenetic changes in leaf epidermal pavement cells [34,35].Phenotypic analyses(Figs.5,S6) suggested thatMeHDZ14may affect cell morphogenesis of leaf epidermis by modulating hormone biosynthesis.Expression of several genes involved in GA and IAA biosynthesis was down-regulated byMeHDZ14overexpression,as indicated by RNA-seq and qPCR analysis(Fig.S8A,B).Yeast-one-hybrid analysis suggested that MeHDZ14 binds to promoters of genes involved in auxin biosynthesis,including tryptophan aminotransferase related(MeTAR2.1)and indole-3-pyruvate monooxygenase(MeYUCCA2.1) (Fig.S8C).

The expression of many genes involved in cell-wall development was significantly down-regulated byMeHDZ14overexpression.We chose six such genes (MeLIP1,MeEXPA1,MeCOMT1,MeRD22,MeEXPA10,andMeDWF4) to determine whetherMeHDZ14directly regulates their transcription (Table S2).All six were down-regulated byMeHDZ14overexpression (Fig.6A-F).The yeast-one-hybrid assay showed that MeHDZ14 bound to the HD-Zip-binding element (HDZB-element) in the promoters of the six genes (Fig.6G).In an EMSA,MeHDZ14 bound to the HDZBelement in the promoter ofMeCOMT1(Fig.6H).The LUC/REN dual luciferase report assay indicated that transient expression ofMeHDZ14reduced the activity of these six promoters (Fig.6I).

Fig.6. MeHDZ14 acted as transcriptional repressor of plant genes involved in cell morphogenesis.(A-F)MeHDZ14 negatively regulated the expression of six genes involved in cell-wall development in transgenic cassava.(G) HD-ZIP binding elements in the promoters of six genes involved in cell-wall development and yeast one-hybrid assay of MeHDZ14 and HDZB-element.HDZB-element,HD-ZIP binding DNA-element.(H) MeHDZ14 binding to the HDZB-element of the MeCOMT1 promoter in the electrophoretic mobility shift assay.(I) Relative luciferase (LUC/REN) activity of the promoters in tobacco leaves with transient expression of MeHDZ14.Error bars represent ± SD (n= 3),different letters indicate differences at P ＜0.05 (ANOVA,Tukey’s multiple comparisons test).

4.Discussion

Cassava genotypes adapt to drought by using strategies such as early growth quiescence or reduction of growth under drought stress [12].In a previous study,we evaluated drought tolerance of 134 cassava genotypes,of which two cassava cultivars,cv.60444 and cv.E1424,were chosen and used for drought treatment in the present study [36].They can also be used as model cassava genotypes to analyze the regulation of drought-induced early growth quiescence.HD-Zip transcription factors are involved in response to abiotic stress and act in hormonal regulation during plant growth [8,10].Comparison of gene expression between the two model cassava cultivars cv.60444 and cv.E1424 suggests thatMeHDZ14plays a role in the regulation of drought-induced early growth quiescence in cassava (Fig.1A-F).

MeHDZ14 is homology to ATHB12 (Fig.S1),which negatively regulates the elongation of inflorescence stems inArabidopsis[8].Thus,MeHDZ14may play roles similar to those ofATHB12in regulating the growth of stems and thereby of plant height.Elongation of internode is essential for growth of stems.Our result(Fig.2)suggests that length reduction of internodes can be partly accounted for by decreased length of stem epidermal cells inMeHDZ14-OE transgenic cassava.It can be inferred thatMeHDZ14negatively regulated stem cell elongation and thus reduced the length of internodes in transgenic cassava.The reduction of internodes can be partly explained by a reduction of GA3 content in transgenic cassava (Figs.5,S6).

HD-Zip transcription factors act as positive regulators of lignin biosynthesis in rice [37,38].In contrast,our findings indicate thatMeHDZ14negatively regulates lignin biosynthesis (Fig.S4).A tomato HD-Zip transcription factor,SlHB8,negatively regulates lignin biosynthesis and influences stem development [39].These studies suggest that HD-Zip transcription factors play contrasting roles in the regulation of lignin biosynthesis.The finding that the phenylpropanoid pathway was inhibited by overexpression ofMeHDZ14can be partially explained by the reduction of phenylalanine in transgenic cassava (Fig.S4).This finding suggests thatMeHDZ14negatively regulates phenylalanine biosynthesis in cassava.Our finding that the level of SA was reduced in transgenic cassava (Fig.S4) is in agreement with the reported [40] impairment of plant growth by SA.Several HD-Zip transcription factors have been reported to be involved in plant’s immunity via the SA signaling pathway [41,42].qPCR analysis indicated that SA negatively regulates the expression ofMeHDZ14in leaves of cassava cv.60444 (Fig.S9).Thus,it can be speculated thatMeHDZ14regulates the plant immune system by triggering the SA signaling pathway.

Although genes including HD-Zip transcription factors that control leaf rolling have been identified in rice and other plant species[32,43,44],none have been mapped or cloned in cassava plants to date.Moderate leaf rolling minimizes shading between leaves,leading to increased photosynthetic efficiency [31].In our study,MeHDZ14overexpression induced leaf rolling while leaving leaf length unchanged in transgenic cassava OE#7(Fig.3C-E).Leaf rolling is the primary visible symptom and one of the survival mechanisms used by plants against drought stress,reducing transpiration rate and canopy temperature in rice and maize[45,46].Transgenic cassava showed a lower chlorophyll content index and normal transpiration rate,(Fig.3F,G)indicating that leaf rolling increased photosynthetic efficiency.However,it did not affect the drought tolerance of transgenic cassava in a closed environment (Fig.S3B).Heat stress alters the plant’s membrane lipid composition,altering leaf orientation by reducing cell size and leading to either leaf rolling or increased water absorption from the soil [47].Under severe drought and heat stress conditions,greater leaf rolling may be associated with higher chances of recovery when moisture stress is relieved [48,49].Future work should investigate whether leaf rolling affects tolerance to drought and heat stress in transgenic cassava.

The leaf epidermis is a biomechanical shield that affects the size and shape of the organ [50].Phenotypic analysis (Figs.4,S5) indicated thatMeHDZ14regulates the morphogenesis of epidermal pavement cells,resulting in abaxially rolled leaves in transgenic cassava.Molecular-genetic and physiological analyses have investigated hormone signaling in epidermal differentiation [35].Plant hormones such as auxin and cytokinin,are essential signaling hormones for the morphogenesis of pavement cells in plants [35,50].HD-Zip transcription factors regulate organ development and tissue formation via auxin and cytokinin signaling in plants [51-53].In our study,MeHDZ14inhibited IAA biosynthesis possibly by directly regulating genes such asTARandYUCCAinvolved in IAA biosynthesis (Fig.S8),while promoting cytokinin biosynthesis in transgenic cassava(Fig.S6).This finding suggests thatMeHDZ14may regulate the morphogenesis of pavement cells by modulating crosstalk between auxin and cytokinin.

Gibberellin promotes internode elongation in plants [54,55],and accumulation of gibberellin in leaves is essential for internode elongation[56].MeHDZ14negatively regulated gibberellin biosynthesis in leaves of transgenic cassava,mainly in OE#5 transgenic cassava (Fig.S6).This finding can partly explain the repression of internode elongation that was displayed in OE#5 rather than in OE#7 transgenic cassava (Fig.2).TheMeHDZ14homologATHB12,affected gibberellin biosynthesis inArabidopsisstems by directly regulating the expression of agibberellin 20-oxidase(GA20OX)gene[8].Future studies should investigate whetherMeHDZ14directly regulatesGA20OXexpression in cassava.

HD-Zip transcription factors are involved in plant adaptation to environmental conditions such as drought stress [52,57].These members regulate a group of genes involved in cell expansion,division,and differentiation inArabidopsis[2,10,37,57-59].Indeed,transcriptome analysis indicated thatMeHDZ14regulates numerous drought-responsive genes described by GO terms such as cell and cell part (Fig.S7).The expression of several genes involved in cell-wall development,includingGDSL lipase(MeLIP1) andexpansin(MeEXPA1andMeEXPA10) genes,was down-regulated inMeHDZ14transgenic cassava(Table S2;Fig.6A-F).Further analysis suggested thatMeHDZ14may regulate these cell-wall-related genes by directly binding to a HDZB-element within their promoters (Fig.6G-I).InArabidopsis,HD-Zip transcription factors bind directly to the L1 boxcis-element in theLIP1andEXP8promoters,thereby regulating cell expansion via GA signaling in the epidermis[10,58].ATHB12may affect cell expansion in the brassinosteroid pathway by regulating the expression level ofDWF4inArabidopsis[10].The finding that overexpression ofMeHDZ14reduced the expression level ofMeEXPA1andMeDWF4in transgenic cassava(Fig.6F) suggests thatMeHDZ14may affect cell expansion in both GA and brassinosteroid pathways.

Repressing lignin biosynthesis caused drought sensitivity in cassava [16].Caffeic acid 3-O-methyltransferase (COMT) is essential for the biosynthesis of lignin in plants [60].Although several HD-Zip transcription factors have been identified as regulators of lignin biosynthesis in plants,their role in the direct regulation of the expression level ofCOMTis still unknown.In the present study,the expression of aCOMTgene,MeCOMT1,was down-regulated byMeHDZ14overexpression in cassava (Table S2;Fig.6C).MeHDZ14 directly bound to the HD-ZIP binding element in the promoter ofMeCOMT1and reduced the transcriptional activity of theMeCOMT1promoter (Fig.6G-I).These results suggest thatMeHDZ14might regulate the expression ofMeCOMT1,thereby affecting lignin biosynthesis.

In summary,we have identified the functional role of a transcription factor,MeHDZ14,that is involved in the regulation of internode elongation and leaf rolling in transgenic cassava(Fig.7).This study advances our understanding of a poorly studied transcription factor,MeHDZ14,and could assist the development of new cassava cultivars.

Fig.7.Proposed role of MeHDZ14 in modulating drought response and leaf development in cassava.

CRediT authorship contribution statement

Xiaoling Yu:Resources,Data curation,Investigation,Writing -original draft.Xin Guo:Investigation,Data curation,Visualization,Formal analysis.Pingjuan Zhao:Resources,Investigation,Data curation.Shuxia Li:Investigation,Data curation.Liangping Zou:Methodology,Data curation.Wenbin Li:Investigation,Validation.Ziyin Xu:Methodology,Data curation.Ming Peng:Supervision.Mengbin Ruan:Conceptualization,Writing -review &editing,Funding acquisition,Project administration.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This work was supported by the China Agriculture Research System (CARS11-HNCX),the Major Science and Technology Plan of Hainan Province(ZDKJ2021012),the Central Public-interest Scientific Institution Basal Research Fund for Chinese Academy of Tropical Agricultural Sciences (1630052022008),the National Key Research and Development Program of China (2018YFD1000501),the National Natural Science Foundation of China (31501378),and the Hainan Yazhou Bay Seed Lab (B21HJ0303).

Appendix A.Supplementary data

Supplementary data for this article can be found online at https://doi.org/10.1016/j.cj.2023.03.001.