Identification of genes expressed in maize root cortical cells during lysigenous aerenchyma formation using laser microdissection and microarray analyses

Authors

  • Imene Rajhi,

    1. Graduate School of Agriculture and Life Sciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo, Tokyo 113–8657, Japan
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Takaki Yamauchi,

    1. Graduate School of Agriculture and Life Sciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo, Tokyo 113–8657, Japan
    Search for more papers by this author
    • These authors contributed equally to this work.

  • Hirokazu Takahashi,

    1. Graduate School of Agriculture and Life Sciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo, Tokyo 113–8657, Japan
    Search for more papers by this author
  • Shunsaku Nishiuchi,

    1. Graduate School of Bioagricultural Sciences, Nagoya University, Furo-cho, Chikusa, Nagoya 464–8601, Japan
    Search for more papers by this author
  • Katsuhiro Shiono,

    1. Graduate School of Agriculture and Life Sciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo, Tokyo 113–8657, Japan
    Search for more papers by this author
  • Ryosuke Watanabe,

    1. Graduate School of Agriculture and Life Sciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo, Tokyo 113–8657, Japan
    Search for more papers by this author
  • Ahmed Mliki,

    1. Biotechnology Center, Borj Cedria Science and Technology Park, Route Touristique Borj Cédria-Soliman, B.P. 901, 2050 Hammam-Lif, Tunisia
    Search for more papers by this author
  • Yoshiaki Nagamura,

    1. National Institute of Agrobiological Sciences, 2-1-2 Kannondai, Tsukuba, Ibaraki 305–8602, Japan
    Search for more papers by this author
  • Nobuhiro Tsutsumi,

    1. Graduate School of Agriculture and Life Sciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo, Tokyo 113–8657, Japan
    Search for more papers by this author
  • Naoko K. Nishizawa,

    1. Graduate School of Agriculture and Life Sciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo, Tokyo 113–8657, Japan
    2. Research Institute for Bioresources and Biotechnology, Ishikawa Prefectural University, 1–308 Nonoichimachi, Ishikawa 921–8836, Japan
    Search for more papers by this author
  • Mikio Nakazono

    1. Graduate School of Agriculture and Life Sciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo, Tokyo 113–8657, Japan
    2. Graduate School of Bioagricultural Sciences, Nagoya University, Furo-cho, Chikusa, Nagoya 464–8601, Japan
    Search for more papers by this author

Author for correspondence:
Mikio Nakazono
Tel: +81 52 789 4017
Email: nakazono@agr.nagoya-u.ac.jp

Summary

  • To adapt to waterlogging in soil, some gramineous plants, such as maize (Zea mays), form lysigenous aerenchyma in the root cortex. Ethylene, which is accumulated during waterlogging, promotes aerenchyma formation. However, the molecular mechanism of aerenchyma formation is not understood.
  • The aim of this study was to identify aerenchyma formation-associated genes expressed in maize roots as a basis for understanding the molecular mechanism of aerenchyma formation. Maize plants were grown under waterlogged conditions, with or without pretreatment with an ethylene perception inhibitor 1-methylcyclopropene (1-MCP), or under aerobic conditions. Cortical cells were isolated by laser microdissection and their mRNA levels were examined with a microarray.
  • The microarray analysis revealed 575 genes in the cortical cells, whose expression was either up-regulated or down-regulated under waterlogged conditions and whose induction or repression was suppressed by pretreatment with 1-MCP.
  • The differentially expressed genes included genes related to the generation or scavenging of reactive oxygen species, Ca2+ signaling, and cell wall loosening and degradation. The results of this study should lead to a better understanding of the mechanism of root lysigenous aerenchyma formation.

Introduction

The aerenchyma is a specialized tissue consisting of longitudinal gas spaces, which enables the internal movement of gases (e.g. O2, CO2, ethylene and methane) in plant roots, petioles and stems (Armstrong, 1979; Colmer, 2003). The internal transport of oxygen via the aerenchyma from shoots to roots is especially important for survival under waterlogged conditions. In general, aerenchyma can be classified into two main types: schizogenous aerenchyma and lysigenous aerenchyma (Jackson & Armstrong, 1999; Seago et al., 2005). Schizogenous aerenchyma is formed by the creation of gas spaces between cells as a result of highly regulated cell separation and differential cell expansion, without cell death taking place. Lysigenous aerenchyma is formed by the creation of gas spaces as a result of death and the subsequent lysis of some cells (e.g. root cortical cells), and is observed in many crops, such as barley, maize, rice and wheat (Jackson & Armstrong, 1999; Evans, 2003).

Many wetland plant species (e.g. rice and Juncus effusus) constitutively form lysigenous aerenchyma in roots under well-drained soil conditions, and its formation is enhanced on soil waterlogging. On the other hand, lysigenous aerenchyma in nonwetland plants, including maize, is not normally formed under well-drained soil conditions, but is induced by waterlogging, hypoxia, mechanical impedance and even under aerobic conditions by nutrient deficiency (Drew et al., 1979; He et al., 1992, 1996a). Because aerenchyma formation can be induced in maize roots by external stimuli, maize has often been used as a model plant for understanding the mechanism of aerenchyma formation.

Ethylene has been implicated in lysigenous aerenchyma formation in maize and rice (Drew et al., 1979; Jackson & Armstrong, 1999; Shiono et al., 2008). In maize roots, ethylene biosynthesis is stimulated by enhancing the activities of two ethylene biosynthetic enzymes (1-aminocyclopropene-1-carboxylic acid (ACC) synthase and ACC oxidase) under hypoxic conditions (He et al., 1996a). Indeed, hypoxic treatment increases the production of ethylene in maize roots by several fold within 3 h (Geisler-Lee et al., 2010). The treatment of maize roots with inhibitors of ethylene action (e.g. silver ions) or ethylene biosynthesis (e.g. aminoethoxyvinylglycine (AVG), aminooxyacetic acid (AOA) and cobalt chloride) effectively blocks aerenchyma formation under hypoxic conditions (Drew et al., 1981; Konings, 1982; Jackson et al., 1985). Moreover, aerenchyma can be induced by treatment with ethylene, even under aerobic conditions (Jackson et al., 1985). These observations indicate that ethylene works as a trigger for inducible aerenchyma formation in maize roots. Ethylene-responsive aerenchyma formation is affected by chemical inhibitors or stimulators of programmed cell death and other signaling pathways (He et al., 1996b). These analyses suggest that heterotrimeric G-protein-, phospholipase C (PLC)-, inositol 1,4,5-trisphosphate (IP3)- and calcium-dependent signaling pathways are involved in the process of lysigenous aerenchyma formation in maize roots (He et al., 1996b; Drew et al., 2000).

In the late stage of lysigenous aerenchyma formation, the cell wall is degraded enzymatically. Initially, the location of esterified pectin and de-esterified pectin in the cell wall of the maize cortex is changed during cell death (Gunawardena et al., 2001), and subsequently the cell wall is degraded by the combined action of pectolytic, xylanolytic and cellulosolytic enzymes (Jackson & Armstrong, 1999). Indeed, the activities of cellulase (CEL), xylanase and pectinase, all of which are involved in the loosening or degradation of the cell wall, are enhanced in maize roots under waterlogged conditions (Jackson & Armstrong, 1999). On the other hand, the expression of genes encoding expansin, which promotes cell wall extensibility by the breaking of hydrogen bonds between hemicellulose and cellulose, is induced by ethylene (Rose et al., 2000). A gene encoding xyloglucan endo-transglycosylase (XET) is up-regulated in maize roots after 12 h of flooding, and induction is inhibited by treatment with an ethylene biosynthesis inhibitor (Saab & Sachs, 1996).

On the basis of these results, Evans (2003) proposed that selective cell death in the maize root cortex occurs in five stages: (1) perception of hypoxia and initiation of ethylene biosynthesis; (2) perception of an ethylene signal by cells of the mid-cortex; (3) initiation of cell death with loss of ions to the surroundings, plasma membrane invagination and the formation of small vesicles; (4) chromatin condensation, increased activities of cell wall hydrolytic enzymes and the surrounding of organelles by membranes; (5) cell wall degradation, cell lysis and absorption of cell contents and water by the surrounding cells, thereby forming gas spaces (i.e. aerenchyma).

So far, these results have mainly been obtained by morphological, anatomical and pharmacological studies, and thus the molecular mechanism of lysigenous aerenchyma formation remains to be elucidated. To better understand the mechanism of lysigenous aerenchyma formation, it is necessary to identify the genes involved and to determine how they are regulated. In this study, we grew maize under aerobic or waterlogged conditions, with or without pretreatment with an inhibitor of ethylene perception. Because aerenchyma formation occurs specifically in root cortical cells, we used laser microdissection (LM; Nakazono et al., 2003; Nelson et al., 2006) to isolate these cells, and then examined their mRNA levels with a microarray and semi-quantitative reverse transcription-polymerase chain reaction (RT-PCR). As a result, we identified genes that were up-regulated or down-regulated in root cortical cells during aerenchyma formation, and discuss their possible roles.

Materials and Methods

Plant material and growth conditions

Maize (Zea mays L. inbred line B73) caryopses were placed on moist chromatography paper (3MM CHR; Whatman, Maidstone, Kent, UK), rolled up in the paper, placed in a flask half shielded with aluminum foil and incubated in constant light at 28°C as described by Nakazono et al. (2003). Three-day-old aerobically grown seedlings were then subjected to the following two experimental conditions.

Experiment 1: effects of waterlogged conditions on aerenchyma formation in a primary root

After 3 d of growth, the underground part (i.e. roots) of seedlings was submerged in distilled water to create waterlogged conditions. For an aerobic control, the chromatography paper was kept moist, but never submerged.

Experiment 2: effect of ethylene on aerenchyma formation in a primary root under waterlogged conditions

Before waterlogging, 2.5-d-old seedlings were pretreated with 1 ppm of a gaseous ethylene perception inhibitor, 1-methylcyclopropene (1-MCP), for 12 h in a tightly closed container. For a control, the same treatment was used but without 1-MCP.

Anatomical observations

Three-day-old aerobically grown seedlings were further grown for 24 h under waterlogged conditions with or without pretreatment with 1-MCP, or under aerobic conditions. We isolated segments of primary roots at 1.5–2.0 cm from the root–shoot junction for the observation of aerenchyma formation. Transverse sections of primary roots were used to determine the extent of aerenchyma formation (defined as the area of the aerenchyma per area of the whole root on the section). Each section was photographed using a light microscope (ECLIPSE E600; Nikon, Tokyo, Japan) with a CCD camera (DIGITAL SIGHT DS-L1; Nikon). Areas were measured with Image J software (Ver. 1.39u; National Institutes of Health, Bethesda, MD, USA). Three independent experiments were conducted, each using three primary roots.

Laser microdissection (LM)

The basal parts of the primary roots (1.5–2.0 cm from the root–shoot junction) were fixed in 75% ethanol : 25% acetic acid; after dehydration in a graded ethanol series, the tissues were embedded in paraffin and sectioned at a thickness of 16 μm. Serial sections were placed onto PEN membrane glass slides (Molecular Devices, Toronto, ON, Canada) for LM as described by Takahashi et al. (2010). To remove paraffin, slides were immersed in 100% xylene for 5 min, and then in 50% xylene and 50% ethanol for 5 min, and finally in 100% ethanol for 5 min, followed by air drying at room temperature. Cortical cells or stelar cells were collected from the root tissue sections using a Veritas Laser Microdissection System LCC1704 (Molecular Devices).

RNA extraction

Total RNA was extracted from the LM-isolated cortical cells or stelar cells using a PicoPure™ RNA isolation kit (Molecular Devices) according to the manufacturer’s instructions. The extracted total RNA was quantified with a Quant-iT™ RiboGreen RNA reagent and kit (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. The quality of total RNA was assessed using a RNA 6000 Pico kit on an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) as described by Takahashi et al. (2010).

Microarray experiment

Total RNAs (10 ng each) were labeled with a Quick Amp Labeling Kit (Agilent Technologies) according to the manufacturer’s instructions. Aliquots of Cy5-labeled and Cy3-labeled cRNA (750 ng each) were used for hybridization in a 4 × 44k Maize Gene Expression Microarray (Agilent Technologies). The array contains 42, 034 60-mer oligo probes to maize genes. Three biological replicates and a color swap for each replicate were analyzed. The hybridized slides were scanned using a DNA microarray scanner G2505C (Agilent Technologies), and signal intensities were extracted by Feature Extraction software (Version 10.5.1.1; Agilent Technologies). A complete set of microarray data was deposited to the Gene Expression Omnibus (GEO) repository under accession number GSE22943.

Microarray data analysis

For inter-array normalization, a global median normalization was applied across all microarrays to achieve the same median signal intensities for each array, and the false discovery rate (FDR) estimation method was used to obtain P values corrected for multiple testing using R software (http://www.r-project.org/) and the RankProduct package (Breitling et al., 2004). The fold change of each probe between two conditions was calculated using an average of six replicates (three biological replicates and a color swap for each replicate). We identified the genes for which there was more than a 2.0-fold change in expression between the two conditions on average (at least 1.5-fold change in each replicate) and whose FDR P value was < 0.05.

Maize expressed sequence tag (EST) sequences were downloaded from the Dana-Farber Cancer Institute (DFCI) Maize Gene Index (http://compbio.dfci.harvard.edu/tgi/cgi-bin/tgi/gimain.pl?gudb=maize). Maize Gene IDs were identified from the Maizesequence Database (http://www.maizesequence.org/) by BLASTN similarity searches using the maize EST sequences as queries. The maximum E value was set at 0.0001. The top hit rice genes were selected using homology-based searches against the Michigan State University's (MSU's) Rice Genome Annotation Project Database (http://rice.plantbiology.msu.edu/) and the Rice Annotation Project Database (http://rapdb.dna.affrc.go.jp/download/index.html). The maximum E value was set at 0.0001. The putative functions were identified from the MSU Rice Genome Annotation Project Data Download (http://rice.plantbiology.msu.edu/downloads_gad.shtml). The annotations were manually improved using BLASTX searches for sequences matching the maize EST sequences (http://www.ncbi.nlm.nih.gov/genbank/GenbankSearch.html).

For gene ontology (GO) analysis, we merged the same IDs and analyzed the frequency of GO terms of up-regulated and down-regulated genes using GO Slim Assignments (http://rice.plantbiology.msu.edu/downloads_gad.shtml).

Semi-quantitative RT-PCR analysis

Semi-quantitative RT-PCR analysis was performed to confirm the expression pattern of selected genes identified by the microarray analysis. Two replicate samples were used for RNA extraction. First-strand cDNA was synthesized using Superscript III (Invitrogen) from 10 ng of total RNA extracted from root cortical cells or stelar cells as already described. KOD FX (TOYOBO, Tokyo, Japan) was used for subsequent PCR amplification with appropriate primers (Supporting Information Table S1): initial denaturation (94°C for 2 min) and 29–45 cycles of denaturation (94°C for 30 s), annealing (56–62°C for 30 s), extension (68°C for 30 s) and final extension (68°C for 6 min).

Results

Aerenchyma formation in a maize primary root

To determine a time point for the identification of ethylene-responsive, aerenchyma formation-associated genes by the LM microarray analyses, 3-d-old aerobically grown maize seedlings were kept under waterlogged conditions, with or without pretreatment with 1-MCP, an inhibitor of ethylene perception, for 0, 6, 12, 18 and 24 h, and aerenchyma formation (percentage cross-sectional area) was measured at the basal region of a primary root. Aerenchyma formation started between 18 and 24 h after waterlogging treatment, whereas it was suppressed for at least 24 h after waterlogging treatment when the seedlings were pretreated with 1-MCP (Fig. 1). These results confirm that ethylene works as a trigger for inducible aerenchyma formation under waterlogged conditions. On the other hand, aerenchyma formation was not observed at the basal region of roots of 4-d-old seedlings grown under aerobic conditions (Fig. 1). To perform microarray analysis, we decided to collect root cortical cells at 12 h after the treatment under waterlogged conditions, with or without pretreatment with 1-MCP, or under aerobic conditions.

Figure 1.

Aerenchyma formation of maize (Zea mays) primary roots under waterlogged conditions (closed squares), waterlogged conditions with 1-methylcyclopropene (1-MCP) pretreatment (circles) and aerobic conditions (open squares). (a) The extent of aerenchyma formation (area of the aerenchyma per area of the whole root on the section) was analyzed at each 6 h for a period of 0–24 h after the start of treatment. All values are means (= 9) ± SD. Three roots were subjected to analysis in each of the three experiments. (b) Tissue sections of maize root at 24 h after the start of treatment. Bar, 100 μm.

Microarray analyses combined with LM

Three-day-old aerobically grown maize seedlings were further grown for 12 h under three conditions: under waterlogged conditions, with or without pretreatment with 1-MCP, or under aerobic conditions; then, the basal parts of the primary roots were fixed and tissue sections were prepared for LM. Cortical cells were collected from the tissue sections via LM (Fig. 2). The RNA samples extracted from the LM-isolated cortical cells were labeled with Cy3 or Cy5 dye, and the labeled cDNA from each of three biological replications was hybridized to maize oligo-microarrays. To identify the genes expressed during aerenchyma formation, gene expressions were compared between the waterlogging treatment and the aerobic control (Expt 1) or between the waterlogging treatment without 1-MCP pretreatment and the waterlogging treatment with 1-MCP pretreatment (Expt 2). The resulting data were analyzed as described in the Materials and Methods section. For each experiment, we selected genes whose intensities were > 2.0-fold higher or lower under one condition than under another condition (FDR P value < 0.05). As a result, the signal intensities of 575 genes (c. 1.4%) among the 42,034 gene probes spotted on a microarray slide were significantly different between the two treatments common in Expts 1 and 2. Among them, it was likely that 239 genes (c. 0.6%) were up-regulated and 336 genes (c. 0.8%) were down-regulated under the conditions inducing aerenchyma formation (i.e. waterlogged conditions) (Fig. 3, Tables S2, S3).

Figure 2.

Isolation of cortical cells from paraffin-embedded sections of a maize (Zea mays) primary root using laser microdissection (LM). (a) A root tissue section before LM. (b) A root tissue section after LM. (c) LM-isolated cortical cells. Bars, 100 μm.

Figure 3.

Number of genes up-regulated or down-regulated under waterlogged conditions [without 1-methylcyclopropene (1-MCP) pretreatment]. Genes whose signal intensities were > 2.0-fold higher or lower under one condition than under another condition (FDR P value < 0.05) were considered to be up-regulated or down-regulated, and the genes commonly up-regulated or down-regulated in both experiments were collected. Experiment 1: 12 h waterlogged conditions (WL)/12 h aerobic conditions (Aer). Experiment 2: 12 h waterlogged conditions without 1-MCP pretreatment (−MCP)/12 h waterlogged conditions with 1-MCP pretreatment (+MCP).

Characterization of specific gene clusters based on GO

The up-regulated and down-regulated genes were classified into several categories based on their allocated GO terms using GO Slim Assignments (http://rice.plantbiology.msu.edu/downloads_gad.shtml) (Fig. 4). Approximately 36% of the up-regulated genes and c. 32% of the down-regulated genes were categorized to genes responsive to stress and several stimuli (e.g. abiotic, biotic, endogenous, external and extracellular stimuli). The genes related to cellular process, biosynthetic process and transport were also relatively abundant in both the up-regulated and down-regulated genes. On the other hand, 5.9% of the up-regulated genes, but only 0.002% of the down-regulated genes, were translation-related genes encoding ribosomal proteins, translation initiation factors and translation elongation factors (Fig. 4).

Figure 4.

Gene classification based on gene ontology (GO) for genes commonly up-regulated (open bars) or down-regulated (closed bars) in Zea mays in Expts 1 and 2. The frequency of GO terms was analyzed using GO Slim Assignment. The x-axis and y-axis indicate the names of clusters and the ratio of each cluster, respectively. Only the biological processes were used for GO analysis.

Validation of gene expression patterns by semi-quantitative RT-PCR

Some of the genes shown to be up-regulated or down-regulated by the microarray were also analyzed by semi-quantitative RT-PCR to confirm the change in expression. For this, we selected 13 genes (11 up-regulated and two down-regulated), which included three reactive oxygen species (ROS) generation/scavenging-related genes, three calcium signaling-related genes, three cell wall modification-related genes, one transporter gene and three transcriptional regulation-related genes (Table 1). The semi-quantitative RT-PCR results confirmed the microarray results for each of the genes (Fig. 5).

Table 1.   List of genes confirmed by semi-quantitative reverse transcription-polymerase chain reaction (RT-PCR) analysis
Gene nameMaize ESTa accessionExpt. 1b WL/AerExpt. 2b−MCP/+MCPMaize gene IDscRAP Os IDsdMSU LOC_Os IDseGene annotationf
  1. Aer, aerobic conditions; –MCP, waterlogged conditions without 1-methylcyclopropene (1-MCP) pretreatment; +MCP waterlogged conditions with 1-MCP pretreatment; WL, waterlogged conditions.

  2. aMaize expressed sequence tags (ESTs) in DFCI Maize Gene Index (http://compbio.dfci.harvard.edu/tgi/cgi-bin/tgi/gimain.pl?gudb=maize).

  3. bAverage expression ratio of three biological replicates and a color swap for each replicate.

  4. cMaize Gene IDs in Maizesequence Database (http://www.maizesequence.org/).

  5. dRAP Os IDs in Rice Annotation Project Database (RAP-DB; http://rapdb.dna.affrc.go.jp/).

  6. eMSU’s Loc_Os IDs in Rice Genome Annotation Project Database (http://rice.plantbiology.msu.edu/).

  7. fMSU’s Putative Function in Rice Genome Annotation Project Database. Manually improved annotations from Genbank BLASTX search (http://www.ncbi.nlm.nih.gov/genbank/GenbankSearch.html) are described in parentheses.

Reactive oxygen species (ROS) generation/scavenging
RBOHTC289691117.7210.72GRMZM2G300965Os12g0541300LOC_Os12g35610Respiratory burst oxidase, putative, expressed
MnSODTC3012963.453.16GRMZM2G160629Os05g0323900LOC_Os05g25850Superoxide dismutase, mitochondrial precursor, putative, expressed
MTTC2986720.160.39GRMZM2G164229Os11g0704500LOC_Os11g47809Metallothionein, putative, expressed
Calcium signaling
CBLTC2948446.6911.56GRMZM2G125838Os02g0802400LOC_Os02g55880EF hand family protein, putative, expressed (calcineurin B-like, CBL)
CMLTC2827522.7115.08GRMZM2G467184Os11g0141400LOC_Os11g04560Calmodulin-like protein 1, putative, expressed
CNGCTC2890940.060.04GRMZM2G074317Os03g0758300LOC_Os03g55100Cyclic nucleotide-gated ion channel 2, putative, expressed
Cell wall modification
XETTC2863135.475.20GRMZM2G174855Os07g0529700LOC_Os07g34580Glycosyl hydrolases family 16, putative, expressed (xyloglucan endotransglucosylase, XET)
PGTC2823635.562.55GRMZM2G037431Os01g0636500LOC_Os01g44970Polygalacturonase, putative, expressed
CELTC3145714.3612.59GRMZM2G141911Os02g0123700LOC_Os02g03120Endoglucanase, putative, expressed (cellulase, CEL)
Transporter
H+ATPaseTC30534218.07186.62GRMZM2G450055Os02g0797300LOC_Os02g55400ATPase 8, plasma membrane-type, putative, expressed (plasma membrane H+-ATPase)
Transcriptional regulation
ERFTC28926930.884.33GRMZM2G053503Os01g0313300LOC_Os01g21120AP2 domain-containing protein, expressed (ethylene response factor, ERF)
RAV1LTC3013254.424.91GRMZM2G169654Os01g0693400LOC_Os01g49830B3 DNA-binding domain-containing protein, expressed (AP2/ERF-B3 domain-containing transcription factor RAV1-like)
UVR8LTC3003282.843.60GRMZM2G003565Os04g0435700LOC_Os04g35570Regulator of chromosome condensation domain-containing protein, expressed (UVB-resistance protein UVR8-like)
Figure 5.

Validation of expression for genes selected from the microarray analysis with semi-quantitative reverse transcription-polymerase chain reaction (RT-PCR). Semi-quantitative RT-PCR analysis of the selected genes was performed with appropriate primers (Supporting Information Table S1). The alphanumeric symbols in parentheses indicate the Maize Gene IDs of the Maizesequence Database. The ubiquitin gene (UBQ) was used as a control. Aer, aerobic conditions; –MCP, waterlogged conditions without 1-methylcyclopropene (1-MCP) pretreatment; +MCP, waterlogged conditions with 1-MCP pretreatment; ROS, reactive oxygen species generation/scavenging; TP, transporter; WL, waterlogged conditions.

Tissue-specific gene expression analysis

To examine whether the ethylene-mediated waterlogging-responsive expression of the selected genes was associated with aerenchyma formation, we used LM to collect sections of cortical cells (aerenchyma-forming tissue) and stelar cells (nonaerenchyma-forming tissue) from cross-sections of primary roots that had been exposed to waterlogged or aerobic conditions for 12 h (Fig. 6a), and performed semi-quantitative RT-PCR (Fig. 6b). Under waterlogged conditions, all of the 13 selected genes (Table 1) were up-regulated or down-regulated in cortical cells. Eight (RBOH, MnSOD, CBL, CML, CNGC, XET, ERF and UVR8L) were also up-regulated or down-regulated in stelar cells, but the difference in the mRNA levels between waterlogged and aerobic conditions was greater in cortical cells than in stelar cells (Fig. 6b).

Figure 6.

Tissue-specific expression analysis of genes selected from the microarray analysis. (a) Isolation of stelar cells (stele) and cortical cells (cortex) from paraffin-embedded tissue sections of a maize (Zea mays) primary root using laser microdissection. Bars, 100 μm. (b) Semi-quantitative reverse transcription-polymerase chain reaction (RT-PCR) analysis of the selected genes. The alphanumeric symbols in parentheses indicate the Maize Gene IDs of the Maizesequence Database. The ubiquitin gene (UBQ) was used as a control. Aer, aerobic conditions; Co, cortex; ROS, reactive oxygen species generation/scavenging; St, stele; TP, transporter; WL, waterlogged conditions.

ROS generation/scavenging-related genes

The production of ROS has been implicated in diverse physiological processes, including programmed cell death, in plants (Overmyer et al., 2003). One of the major sources of ROS in plants is a reaction mediated by NADPH oxidase, which is responsible for the conversion of O2 to superoxide anion (O2), thereby leading to the production of hydrogen peroxide (H2O2) (Overmyer et al., 2003). Here, we found several ethylene-mediated waterlogging-responsive genes related to ROS generation or ROS scavenging (Table 2). The up-regulated genes include those encoding respiratory burst oxidase homolog (RBOH), glutathione S-transferase and manganese superoxide dismutase, and the down-regulated genes include those encoding RBOH and metallothionein (MT). The RBOH gene products are involved in ROS generation, and the other gene products are involved in ROS scavenging. Among these genes, RBOH (GRMZM2G300965) showed a 117-fold higher expression level under waterlogged conditions than under aerobic conditions, and the induction was partially suppressed by treatment with 1-MCP (Fig. 5, Table 2). As shown in Fig. 6(b), up-regulation of RBOH expression was observed in both cortical cells and stelar cells, but the mRNA levels appeared to be slightly higher in cortical cells than in stelar cells under waterlogged conditions. On the other hand, the MnSOD gene (GRMZM2G160629) was up-regulated preferentially in cortical cells under waterlogged conditions. Interestingly, the MT gene (GRMZM2G164229) was constitutively expressed in both cortical cells and stelar cells under aerobic conditions, but the MT mRNA levels were decreased specifically in cortical cells under waterlogged conditions (Fig. 6b).

Table 2.   List of genes related to reactive oxygen species (ROS) generation/scavenging, calcium signaling and cell wall modification, whose expression was up-regulated or down-regulated in maize root cortex during aerenchyma formation
Maize ESTa accessionExpt. 1b WL/AerExpt. 2b−MCP/+MCPMaize gene IDscRAP Os IDsdMSU LOC_Os IDseGene annotationf
  1. Aer, aerobic conditions; –MCP, waterlogged conditions without 1-methylcyclopropene (1-MCP) pretreatment; +MCP, waterlogged conditions with 1-MCP pretreatment; WL, waterlogged conditions.

  2. aMaize ESTs in DFCI Maize Gene Index (http://compbio.dfci.harvard.edu/tgi/cgi-bin/tgi/gimain.pl?gudb=maize).

  3. bAverage expression ratio of three biological replicates and a color swap for each replicate.

  4. cMaize Gene IDs in Maizesequence Database (http://www.maizesequence.org/).

  5. dRAP Os IDs in Rice Annotation Project Database (RAP-DB; http://rapdb.dna.affrc.go.jp/).

  6. eMSU’s Loc_Os IDs in Rice Genome Annotation Project Database (http://rice.plantbiology.msu.edu/).

  7. fMSU’s Putative Function in Rice Genome Annotation Project Database. Manually improved annotations from Genbank BLASTX search (http://www.ncbi.nlm.nih.gov/genbank/GenbankSearch.html) are described in parentheses.

  8. gAC234190.1 indicates contig ID.

Reactive oxygen species (ROS) generation/scavenging
TC289691117.7210.72GRMZM2G300965Os12g0541300LOC_Os12g35610Respiratory burst oxidase, putative, expressed
TC3014605.3920.85GRMZM2G416632Os09g0467200LOC_Os09g29200Glutathione S-transferase, putative, expressed
TC3012963.453.16GRMZM2G160629Os05g0323900LOC_Os05g25850Superoxide dismutase, mitochondrial precursor, putative, expressed
TC3117660.270.20GRMZM2G426953Os01g0835500LOC_Os01g61880Respiratory burst oxidase, putative, expressed
TC2986720.160.39GRMZM2G164229Os11g0704500LOC_Os11g47809Metallothionein, putative, expressed
Calcium signaling
TC2948446.6911.56GRMZM2G125838Os02g0802400LOC_Os02g55880EF hand family protein, putative, expressed (calcineurin B-like, CBL)
TC3135115.752.91GRMZM2G052740Os02g0158100LOC_Os02g06340EH domain-containing protein 1, putative, expressed
TC3071985.573.78GRMZM2G082199Os01g0841700LOC_Os01g62430C2 domain-containing protein, putative, expressed
TC2827522.7115.08GRMZM2G467184Os11g0141400LOC_Os11g04560Calmodulin-like protein 1, putative, expressed
TC2820500.410.28GRMZM2G032852Os03g0128700LOC_Os03g03660CAMK_CAMK_like.17 – CAMK includes calcium/calmodulin-dependent protein kinases, expressed
TC2883820.380.14GRMZM2G180916Os10g0564500LOC_Os10g41490CAMK_CAMK_like.41 – CAMK includes calcium/calmodulin-dependent protein kinases, expressed
TC2989290.360.42GRMZM2G173424Os12g0597000LOC_Os12g40510Calcineurin B, putative, expressed
TC2936250.330.18GRMZM2G412601Os09g0418500LOC_Os09g25100CAMK_KIN1/SNF1/Nim1_like.35 – CAMK includes calcium/calmodulin-dependent protein kinases, expressed
TC2812510.270.23GRMZM2G146720Os11g0134300LOC_Os12g03810CAMK_KIN1/SNF1/Nim1_like.37 – CAMK includes calcium/calmodulin-dependent protein kinases, expressed
TC2825870.270.22GRMZM2G078781Os01g0782800LOC_Os01g57370Cyclic nucleotide-gated ion channel 2, putative, expressed
BM5011480.220.32GRMZM2G096228Os10g0389000LOC_Os10g25010OsCML8 – calmodulin-related calcium sensor protein, expressed
TC2890940.060.04GRMZM2G074317Os03g0758300LOC_Os03g55100Cyclic nucleotide-gated ion channel 2, putative, expressed
Cell wall modification
TC31576010.777.35GRMZM2G043943Os01g0743200LOC_Os01g53990Pectinesterase, putative, expressed
TC2838385.8611.61GRMZM2G131912Os04g0137100LOC_Os04g05050Pectate lyase precursor, putative, expressed
TC2823635.562.55GRMZM2G037431Os01g0636500LOC_Os01g44970Polygalacturonase, putative, expressed
TC2863135.475.20GRMZM2G174855Os07g0529700LOC_Os07g34580Glycosyl hydrolases family 16, putative, expressed (xyloglucan endotransglucosylase, XET)
TC3110075.0723.94GRMZM2G119471Os03g0124900LOC_Os03g03350Polygalacturonase, putative, expressed
TC3145714.3612.59GRMZM2G141911Os02g0123700LOC_Os02g03120Endoglucanase, putative, expressed (cellulase, CEL)
TC3144583.00107.88AC234190.1gOs01g0249100LOC_Os01g14660Expansin precursor, putative, expressed
TC2996142.434.16GRMZM2G048430Os06g0711800LOC_Os06g49760Invertase/pectin methylesterase inhibitor family protein, putative, expressed
TC2803170.420.05GRMZM2G028353Os07g0424400LOC_Os07g24190CESA3 – cellulose synthase, expressed
TC2878320.340.10GRMZM2G424832Os07g0208500LOC_Os07g10770CESA8 – cellulose synthase, expressed
TC3034350.240.24GRMZM2G074792Os01g0766900LOC_Os01g56130CSLC1 – cellulose synthase-like family C, expressed

Calcium signaling-related genes

Many studies have suggested that the cytosolic calcium ion (Ca2+) functions as a second messenger for signaling pathways in response to oxygen deprivation (Subbaiah et al., 1994; Tsuji et al., 2000; Baxter-Burrell et al., 2002). Ca2+ signaling may also be involved in aerenchyma formation in maize roots (He et al., 1996b). In this study, several genes implicated in calcium signaling, whose expression was changed significantly under waterlogged conditions (without 1-MCP pretreatment) in both Expts 1 and 2, were identified (Table 2). They included up-regulated genes encoding calcineurin B-like protein (CBL), Ca2+-binding domain-containing proteins and calmodulin-like protein (CML), and down-regulated genes encoding calcium/calmodulin-dependent protein kinases, CBL, cyclic nucleotide-gated ion channel (CNGC) protein and CML (Table 2). As shown in Fig. 6(b), the up-regulation of CBL (GRMZM2G125838) and CML (GRMZM2G467184) and the down-regulation of CNGC (GRMZM2G074317) were observed in both cortical cells and stelar cells, but the changes in expression were more pronounced in cortical cells than in stelar cells.

Cell wall modification-related genes

The last step of aerenchyma formation involves cell wall loosening and degradation, in which many enzymes, including XETs, expansins, CELs and pectinases, are involved (He et al., 1994; Saab & Sachs, 1996; Jackson & Armstrong, 1999). In this study, several of the up-regulated genes encode these enzymes, including the genes for pectinesterase, pectate lyase, polygalacturonase (PG), XET, CEL, expansin and invertase/pectin methylesterase inhibitor family protein, and several down-regulated genes encode cellulose synthase and cellulose synthase-like C family protein (Table 2). Under waterlogged conditions, two of the three selected cell wall modification-related genes, the genes encoding PG (GRMZM2G037431) and CEL (GRMZM2G141911), were specifically up-regulated in cortical cells, whereas the XET gene (GRMZM2G174855) was up-regulated in both cortical cells and stelar cells (Fig. 6b).

Protein kinase, protein phosphatase and transcriptional regulator genes

Among the waterlogging-sensitive genes, 16 genes were protein kinase genes (four up-regulated and 12 down-regulated) and two genes were protein phosphatase genes (both up-regulated) (Table 3). This is consistent with a previous finding that protein phosphorylation and dephosphorylation are important for the regulation of aerenchyma formation in maize (He et al., 1996b).

Table 3.   List of genes encoding protein kinases and protein phosphatases, whose expression was up-regulated or down-regulated in maize root cortex during aerenchyma formation
Maize ESTa accessionExpt. 1b WL/AerExpt. 2b−MCP/+MCPMaize gene IDscRAP Os IDsdMSU LOC_Os IDseGene annotationf
  1. Aer, aerobic conditions; –MCP, waterlogged conditions without 1-methylcyclopropene (1-MCP) pretreatment; +MCP, waterlogged conditions with 1-MCP pretreatment; WL, waterlogged conditions.

  2. aMaize ESTs in DFCI Maize Gene Index (http://compbio.dfci.harvard.edu/tgi/cgi-bin/tgi/gimain.pl?gudb=maize).

  3. bAverage expression ratio of three biological replicates and a color swap for each replicate.

  4. cMaize Gene IDs in Maizesequence Database (http://www.maizesequence.org/).

  5. dRAP Os IDs in Rice Annotation Project Database (RAP-DB; http://rapdb.dna.affrc.go.jp/).

  6. eMSU’s Loc_Os IDs in Rice Genome Annotation Project Database (http://rice.plantbiology.msu.edu/).

  7. fMSU’s Putative Function in Rice Genome Annotation Project Database. Manually improved annotations from Genbank BLASTX search (http://www.ncbi.nlm.nih.gov/genbank/GenbankSearch.html) are described in parentheses.

  8. gAC217293.3 indicates contig ID.

TC28123023.496.29GRMZM2G147051Os05g0436100LOC_Os05g36050Serine/threonine protein kinase, putative, expressed
CO45869018.066.11GRMZM2G092604Os09g0293500LOC_Os09g12240Serine/threonine protein kinase
BRI1-like 1 precursor, putative, expressed
TC2992856.3910.13GRMZM2G091338Os01g0789200LOC_Os01g57940Tyrosine protein kinase domain-containing protein, putative, expressed
TC2825714.803.18GRMZM2G113668Os03g0857600LOC_Os03g64050Receptor protein kinase, putative, expressed
TC2936633.123.12GRMZM2G126765Os06g0208700LOC_Os06g10650Tyrosine phosphatase family protein, putative, expressed
TC2939602.522.40GRMZM2G015610Os04g0449450LOC_Os04g37660Protein phosphatase protein, putative, expressed
TC2820500.410.28GRMZM2G032852Os03g0128700LOC_Os03g03660CAMK_CAMK_like.17 – CAMK includes calcium/calmodulin-dependent protein kinases, expressed
TC3063610.390.38GRMZM2G074262Os01g0864700LOC_Os01g64490Protein kinase domain-containing protein, expressed
TC2883820.380.14GRMZM2G180916Os10g0564500LOC_Os10g41490CAMK_CAMK_like.41 – CAMK includes calcium/calmodulin-dependent protein kinases, expressed
TC2936250.330.18GRMZM2G412601Os09g0418500LOC_Os09g25100CAMK_KIN1/SNF1/Nim1_like.35 – CAMK includes calcium/calmodulin-dependent protein kinases, expressed
TC3088020.310.18GRMZM2G119521Os10g0432001LOC_Os10g29620Tyrosine protein kinase domain-containing protein, putative, expressed
TC3152300.290.40GRMZM2G126858Os04g0616700LOC_Os04g52640SHR5-receptor-like kinase, putative, expressed
TC2812510.270.23GRMZM2G146720Os11g0134300LOC_Os12g03810CAMK_KIN1/SNF1/Nim1_like.37 – CAMK includes calcium/calmodulin-dependent protein kinases, expressed
TC3145810.250.20GRMZM2G092776Os03g0113000LOC_Os03g02190Protein kinase domain-containing protein, expressed
TC3088360.230.12GRMZM2G132591Os10g0174800LOC_Os10g09620OsWAK108 – OsWAK receptor-like protein kinase, expressed
TC3126320.220.12GRMZM2G135359Os03g0294800LOC_Os03g18370MRH1, putative, expressed
TC3058400.150.17AC217293.3gOs12g0615100LOC_Os10g07556Wall-associated receptor kinase-like 22 precursor, putative, expressed
TC3026130.150.20GRMZM2G158252Os10g0362300LOC_Os10g21810Histidine kinase, putative, expressed

Of the 34 genes encoding putative transcriptional regulators, including transcription factors of > 20 different families, 13 genes were up-regulated and 21 genes were down-regulated (Table 4). Among these genes, the gene encoding an AP2 domain-containing protein (GRMZM2G053503), which is similar to ethylene response factor (ERF), showed a strong (c. 30-fold) increase in expression under waterlogged conditions, and the induction was partially suppressed by pretreatment with 1-MCP (Fig. 5, Table 4). The expression of a gene [designated as RAV1-like (RAV1L) in this study; GRMZM2G169654] encoding a protein containing a B3 DNA-binding domain, which is homologous to an AP2/ERF domain and B3 domain containing transcription factor RAV1, was also induced under waterlogged conditions, and seemed to be controlled by ethylene (Fig. 5, Table 4). Another up-regulated gene is a regulator of chromosome condensation domain-containing protein (GRMZM2G003565). It is a UVR8-like gene (UVR8L), which is homologous to the Arabidopsis UVR8 gene (Table 4). Three other up-regulated genes are related to histone modification. These include a jmjC domain (jumonj-C-domain)-containing protein (GRMZM2G417089) and two histone acetyltransferases (GRMZM2G069886 and GRMZM2G106673) (Table 4). Under waterlogged conditions, the ERF (GRMZM2G053503) and UVR8L (GRMZM2G003565) genes were mainly up-regulated in cortical cells and the RAV1L (GRMZM2G169654) gene was up-regulated only in cortical cells (Fig. 6b).

Table 4.   List of genes encoding transcriptional regulation-related proteins, whose expression was up-regulated or down-regulated in maize root cortex during aerenchyma formation
Maize ESTa accessionExpt. 1b WL/AerExpt. 2b−MCP/+MCPMaize gene IDscRAP Os IDsdMSU LOC_Os IDseGene annotationf
  1. Aer, aerobic conditions; –MCP, waterlogged conditions without 1-methylcyclopropene (1-MCP) pretreatment; +MCP, waterlogged conditions with 1-MCP pretreatment; WL, waterlogged conditions.

  2. aMaize ESTs in DFCI Maize Gene Index (http://compbio.dfci.harvard.edu/tgi/cgi-bin/tgi/gimain.pl?gudb=maize).

  3. bAverage expression ratio of three biological replicates and a color swap for each replicate.

  4. cMaize Gene IDs in Maizesequence Database (http://www.maizesequence.org/).

  5. dRAP Os IDs in Rice Annotation Project Database (RAP-DB; http://rapdb.dna.affrc.go.jp/).

  6. eMSU’s Loc_Os IDs in Rice Genome Annotation Project Database (http://rice.plantbiology.msu.edu/).

  7. fMSU’s Putative Function in Rice Genome Annotation Project Database. Manually improved annotations from Genbank BLASTX search (http://www.ncbi.nlm.nih.gov/genbank/GenbankSearch.html) are described in parentheses.

  8. gAC232238.2 indicates contig ID.

TC28926930.884.33GRMZM2G053503Os01g0313300LOC_Os01g21120AP2 domain-containing protein, expressed (ethylene response factor, ERF)
TC28828925.115.39GRMZM2G417089Os03g0430400LOC_Os03g31594jmjC domain-containing protein, expressed
TC2922027.8922.50GRMZM2G068973Os01g0816100LOC_Os01g60020NAC domain transcription factor, putative, expressed
CD9665496.419.14GRMZM2G069886Os01g0246100LOC_Os01g14370Histone acetyltransferase HAC5, putative, expressed
TC2930584.804.82GRMZM2G106673Os09g0347800LOC_Os09g17850Acetyltransferase type B catalytic subunit, putative, expressed
TC2930564.6910.36GRMZM2G106673Os06g0194400LOC_Os06g09420B3 DNA-binding domain-containing protein, expressed
TC3013254.424.91GRMZM2G169654Os01g0693400LOC_Os01g49830B3 DNA-binding domain-containing protein, expressed (AP2/ERF-B3 domain-containing transcription factor RAV1-like)
TC3044084.3923.27GRMZM2G148074Os10g0561800LOC_Os10g41230Homeobox-associated leucine zipper, putative, expressed
TC3029704.2444.59GRMZM2G108865Os03g0764900LOC_Os03g55610Dof zinc finger domain-containing protein, putative, expressed
TC3008034.2110.77GRMZM2G172621Os07g0679500LOC_Os07g48180bZIP transcription factor domain-containing protein, expressed
TC2857124.164.05GRMZM2G125522Os05g0449900LOC_Os05g37730MYB family transcription factor, putative, expressed
TC3032063.7416.46GRMZM2G460472Os11g0158500LOC_Os11g06010Helix–loop–helix DNA-binding protein, putative, expressed
TC3003282.843.60GRMZM2G003565Os04g0435700LOC_Os04g35570Regulator of chromosome condensation domain-containing protein, expressed (UVB-resistance protein UVR8-like)
TC2831260.410.42GRMZM2G093895Os12g0173125LOC_Os12g07480TCP family transcription factor, putative, expressed
TC2884020.410.29GRMZM2G142768Os02g0817600LOC_Os02g57250OsIAA10 – auxin-responsive Aux/IAA gene family member, expressed
TC2999460.400.39GRMZM2G154641Os01g0848400LOC_Os01g62920Homeodomain protein, putative, expressed
TC2811700.370.36GRMZM2G171365Os03g0122600LOC_Os03g03100OsMADS50 – MADS-box family gene with MIKCc type-box, expressed
TC3127510.360.33GRMZM2G104551Os09g0532900LOC_Os09g36250MYB family transcription factor, putative, expressed
TC3135130.350.02AC232238.2gOs01g0859500LOC_Os01g64020Transcription factor, putative, expressed
TC2959290.340.27GRMZM2G037630Os03g0411100LOC_Os03g29760Nuclear transcription factor Y subunit, putative, expressed
TC2832990.320.30GRMZM2G328481Os06g0211200LOC_Os06g10880bZIP transcription factor, putative, expressed
TC2804810.320.16GRMZM2G485184Os12g0510900LOC_Os12g32620OsWLIM1 – LIM domain protein, putative actin-binding protein and transcription factor, expressed
TC3016780.310.26GRMZM2G104390Os01g0343300LOC_Os01g24070GATA zinc finger domain-containing protein, expressed
TC2922950.300.25GRMZM2G180328Os01g0884300LOC_Os01g66120No apical meristem protein, putative, expressed
TC3066500.270.29GRMZM2G144196Os08g0242800LOC_Os08g14450RNA polymerase sigma factor, putative, expressed
TC2898040.270.44GRMZM2G356439Os04g0678400LOC_Os04g58190Dof zinc finger domain-containing protein, putative, expressed
TC3155610.260.33GRMZM2G320549Os01g0922800LOC_Os01g69850OsMADS65 – MADS-box family gene with MIKC* type-box, expressed
TC3029570.170.24GRMZM2G031094Os11g0152700LOC_Os12g05680Transcription factor, putative, expressed
TC2906960.160.18GRMZM2G452178Os07g0129700LOC_Os07g03770Homeobox domain-containing protein, expressed
TC3017870.160.25GRMZM2G003944Os11g0175700LOC_Os11g07460TCP family transcription factor, putative, expressed
TC3096650.160.14GRMZM2G472671Os08g0543700LOC_Os08g43070Helix–loop–helix DNA-binding domain-containing protein
TC3139750.130.11GRMZM2G009406Os08g0490100LOC_Os09g29960Dof zinc finger domain-containing protein, putative, expressed
TC2858320.130.18GRMZM2G002128Os05g0449900LOC_Os05g37730MYB family transcription factor, putative, expressed
TC3021850.090.05GRMZM2G172657Os07g0586900LOC_Os07g39820SHR, putative, expressed

Discussion

To better understand the molecular mechanism of aerenchyma formation in maize root cortical cells, we screened for genes whose expression changed in response to ethylene under waterlogged conditions, and found 239 up-regulated genes and 336 down-regulated genes. Unsurprisingly, many of the genes (c. 36% of the up-regulated genes and c. 32% of the down-regulated genes) are known to be responsive to stress or other stimuli (Fig. 4). However, it is not clear why many translation-related genes (5.9%) are included in the up-regulated genes, but not in the down-regulated genes. It is known that translation of many normal cellular mRNAs is extremely limited in maize roots under anoxia, whereas mRNAs for anaerobic proteins (related to anaerobic metabolism, such as glycolysis and fermentation) are selectively translated (Sachs et al., 1980; Bailey-Serres, 1999). The selective translation under oxygen deprivation is important for energy conservation and facilitates the transition to anaerobic metabolism (Branco-Price et al., 2008). Thus, to understand the roles of the differentially expressed genes (including the up-regulated translation-related genes), it is necessary to examine whether their mRNAs are effectively translated in root cortical cells under waterlogged conditions.

RBOH, a plant homolog of gp91phox in mammalian NADPH oxidase, has an important role in ROS-mediated signaling, such as the defense response, programmed cell death and development in plants (Torres et al., 2002; Foreman et al., 2003; Takeda et al., 2008; Yoshioka et al., 2009). Indeed, the Rop (RHO-like small G-protein of plants)-dependent H2O2 production mediated by NADPH oxidase, the activity of which is stimulated by Ca2+, contributes to the induction of expression of ADH and RopGAP4 genes in Arabidopsis under oxygen deprivation (Baxter-Burrell et al., 2002). In rice, ethylene-induced, H2O2-mediated epidermal cell death, which precedes the emergence of adventitious roots, is regulated by NADPH oxidase (Steffens & Sauter, 2005, 2009). Here, we found that one RBOH gene (GRMZM2G300965) was up-regulated and another RBOH gene [ZmRBOHA (Lin et al., 2009); GRMZM2G426953] was down-regulated during aerenchyma formation in maize roots, implying that the roles of the two RBOH proteins may be different. The maize up-regulated RBOH is homologous to rice OsRBOHH (Wong et al., 2007), Arabidopsis AtRBOHB (Torres et al., 1998) and potato StRBOHB (Yoshioka et al., 2001) (data not shown). In potato, treatment of tubers with hyphal wall components (HWCs) from Phytophthora infestans causes a rapid and transient oxidative burst (i.e. H2O2 accumulation; phase I), followed by a massive oxidative burst (phase II) (Yoshioka et al., 2001). It is likely that StRBOHA contributes to phase I of the oxidative burst, and that other RBOHs (StRBOHB, StRBOHC and StRBOHD) contribute to phase II (Yamamizo et al., 2006). Both oxidative bursts are inhibited by the serine/threonine protein kinase inhibitor K252a or the extracellular Ca2+ chelator ethyleneglycol-bis(β-aminoethylether)-N,N′-tetraacetic acid (EGTA) (Kobayashi et al., 2007). These observations raise the possibility that, in maize, waterlogging-induced up-regulation of RBOH is involved in H2O2 production, and the H2O2 induces cell death (i.e. aerenchyma formation) in root cortical cells. On the other hand, we found that the expression of the gene encoding MT (GRMZM2G164229), which works as an ROS scavenger (Wong et al., 2004), was repressed under waterlogged conditions, and that the repression seemed to be ethylene dependent (Fig. 5, Table 2). Interestingly, the rice Metallothionein2b (MT2b) gene is down-regulated in response to ethylene and H2O2 in epidermal cells, thereby amplifying the accumulation of H2O2 produced by NADPH oxidase (i.e. RBOH), to induce cell death (Steffens & Sauter, 2009). Similarly, the cortical cell-specific down-regulation of the maize MT gene (Fig. 6b) may contribute to higher accumulation of the RBOH-produced H2O2, which induces cell death in the cortical cells for lysigenous aerenchyma formation. In stele, the MT gene is constitutively expressed even under waterlogged conditions (Fig. 6b), which might reduce the amount of H2O2 produced by RBOH. Indeed, it has been demonstrated recently that the down-regulation of MT2b or application of H2O2 promotes aerenchyma formation in internodes of rice stems (Steffens et al., 2010). On the basis of these results, it is possible that similar mechanisms may regulate epidermal cell death in rice, aerenchyma formation in rice internodes and aerenchyma formation in maize root cortex. Further functional analyses of the up-regulated RBOH gene and the down-regulated MT gene in maize are necessary to examine this possibility.

Genes related to Ca2+ signaling, including CBL, CML and calcium/calmodulin-dependent protein kinase, were also identified as the up-regulated or down-regulated genes in response to waterlogging treatment (without 1-MCP pretreatment) (Table 2). It has been proposed that, under oxygen deprivation, Ca2+ is released from the apoplast and from mitochondria into the cytoplasm, and the elevated cytosolic Ca2+ provokes subsequent activation of kinases and phosphatases, resulting in the activation of the expression of genes responsible for aerenchyma formation (Subbaiah & Sachs, 2003). Treatments with thapsigargin and caffeine, which increase intracellular Ca2+ levels, stimulated CEL activity and aerenchyma formation under aerobic conditions, whereas both EGTA (a Ca2+ chelator) and ruthenium red (an inhibitor of Ca2+ fluxes from organelles) prevented the increase in CEL activity and aerenchyma formation by decreasing cytosolic Ca2+ in maize roots, even under anaerobic conditions (He et al., 1996b). It seems that Ca2+ can bind directly to Ca2+-binding EF-hand motifs in the N-terminal region of RBOH (i.e. NADPH oxidase) and stimulate its activity (Keller et al., 1998; Sagi & Fluhr, 2001; Oda et al., 2010). On the other hand, it has been reported that calcium-dependent protein kinase activates RBOH by phosphorylation of its N-terminal region (Kobayashi et al., 2007). On the basis of these results, an interaction between Ca2+ signaling and RBOH-mediated H2O2 production might be important for programmed cell death in root cortical cells.

In plants, complexes of Ca2+ sensors (CBLs) and their targets [CBL-interacting protein kinases (CIPKs)] form a complex network of Ca2+ signaling, and are responsible for environmental adaptation processes (Luan et al., 2009; Weinl & Kudla, 2009), implying that the up-regulated CBL (GRMZM2G125838) and down-regulated CBL (GRMZM2G173424) might be involved in adaptation (e.g. aerenchyma formation) to waterlogged conditions. The genes (GRMZM2G074317 and GRMZM2G078781) encoding proteins similar to cyclic nucleotide-gated ion channel AtCNGC2 and AtCNGC4, respectively, were included in the down-regulated genes. AtCNGC2 is involved in influxes of Ca2+ and K+ in a cyclic nucleotide-dependent fashion (Leng et al., 1999). It is noteworthy that mutations of AtCNGC2 and AtCNGC4 genes [designated defense, no death 1 (dnd1) and dnd2, respectively] cause a phenotype that shows reduced ability to produce the hypersensitive response (HR) in response to avirulent Pseudomonas syringae pv. glycinea (Clough et al., 2000; Jurkowski et al., 2004).

Several genes related to cell wall loosening and degradation were up-regulated under waterlogged conditions, and it is likely that their induction was controlled by ethylene (Table 2). We found that a gene (GRMZM2G174855) encoding XET, a cell wall loosening enzyme, was up-regulated in both cortical cells and stelar cells in response to waterlogging (Figs 5, 6b). Previously, Saab & Sachs (1996) reported that XET mRNA was strongly accumulated in maize seedlings under flooding. Treatment with an ethylene biosynthesis inhibitor, AOA, under flooded conditions prevented the development of aerenchyma in maize roots and totally suppressed the accumulation of XET mRNA, suggesting that ethylene-responsive XET induction is involved in aerenchyma formation through cell wall loosening and degradation (Saab & Sachs, 1996). The XET gene identified in this study is not the same as the XET gene reported by Saab & Sachs (1996), suggesting that at least two ethylene-responsive XET genes are strongly expressed in maize roots under waterlogged conditions. The up-regulation of other genes related to cell wall loosening or degradation (e.g. pectinesterase, pectate lyase, PG and CEL) may also contribute to the activation of hydrolytic enzymes, including CEL, xylanase and pectinase, in maize roots under waterlogged conditions (Jackson & Armstrong, 1999). Indeed, the expression of genes encoding PG (GRMZM2G037431) and CEL (GRMZM2G141911) was up-regulated specifically in cortical cells under waterlogged conditions (Fig. 6b). On the other hand, we found that the genes for cellulose synthase (GRMZM2G028353 and GRMZM2G424832) and cellulose synthase-like C family protein (GRMZM2G074792) were down-regulated, suggesting that this down-regulation promotes cell wall degradation via repression of cellulose synthesis. We also identified a gene (GRMZM2G450055) encoding plasma membrane H+-ATPase as a cortical cell-specific up-regulated gene (Figs 5, 6b, Table 1). It is proposed that the extrusion of intracellular H+ into the cell wall by plasma membrane H+-ATPase results in a decrease in apoplastic pH, which induces cell wall loosening, possibly mediated by low-pH-activated expansins and XETs (Frias et al., 1996; Shieh & Cosgrove, 1998). Thus, the expression of the XET gene (GRMZM2G174855) is up-regulated in both cortical cells and stelar cells, but the activity of XET protein might be enhanced preferentially in cortical cells because the gene encoding plasma membrane H+-ATPase shows cortical cell-specific induction of expression under waterlogged conditions. In this way, the up-regulated H+-ATPase gene might be involved in cell wall loosening in cortical cells during cell death.

Under waterlogged conditions, the ERF gene (GRMZM2G053503) was preferentially up-regulated in cortical cells and the RAV1L gene (GRMZM2G169654) was specifically up-regulated in cortical cells (Figs 5, 6b, Table 4). The up-regulation of these genes was suppressed by 1-MCP pretreatment (Fig. 5, Table 4). Recently, Licausi et al. (2010) have identified two Arabidopsis hypoxia-inducible ERF genes, HRE1 and HRE2, which belong to group VII of the ERF family in Arabidopsis (Nakano et al., 2006), and have proposed that HRE1 and HRE2 play a partially redundant role in the tolerance of plants to anaerobic stress by enhancing anaerobic gene expression and ethanol fermentation. Group VII of the ERF family also contains Arabidopsis RAP2.2 (Hinz et al., 2010), rice SUB1A (Fukao et al., 2006; Xu et al., 2006; Fukao & Bailey-Serres, 2008), and rice SNORKEL1 and SNORKEL2 (Hattori et al., 2009), all of which play important and distinct roles in survival under hypoxia or submergence. Interestingly, the maize up-regulated ERF is highly homologous to the Arabidopsis HRE2 protein, suggesting that the maize ERF gene, like the Arabidopsis HRE2 gene, is involved in the adaptation of plants to waterlogged conditions. However, to date, it is unclear whether transcriptional regulation by this ERF affects aerenchyma formation in maize roots, and thus further functional analysis of the ERF gene is necessary. On the other hand, it has been reported that the expression of the Arabidopsis RAV1 gene is induced by treatment with ACC (a precursor of ethylene biosynthesis) and that the RAV1 protein positively controls leaf senescence, which is a developmentally programmed cell death process (Woo et al., 2010). Similarly, in maize, the cortical cell-specific RAV1L protein may be positively involved in programmed cell death (i.e. in aerenchyma formation) in root cortical cells under waterlogged conditions.

Three genes related to histone modification were induced under waterlogged conditions (without 1-MCP pretreatment) in both Expts 1 and 2 (Table 4). One of the histone modification-related genes (GRMZM2G417089) encodes a jmjC domain-containing protein. Recently, some jmjC domain-containing proteins have been shown to be histone demethylases (Mosammaparast & Shi, 2010). We have reported previously that submergence and re-aeration of rice cause dynamic and reversible changes of the histone methylation and acetylation states for the genes involved in anaerobiosis (Tsuji et al., 2006). Similarly, it is possible that dynamic histone modifications occur in chromatin at particular genes in the maize cortex in response to ethylene under waterlogged conditions, and that the three maize up-regulated gene products contribute to the changes in histone methylation and acetylation.

In conclusion, in this study, we found that genes related to many types of molecular function (e.g. ROS generation or scavenging, Ca2+ signaling and cell wall modification) were up-regulated or down-regulated in root cortical cells under waterlogged conditions, and their expression was likely to be regulated by ethylene. We are also currently conducting microarray analysis for the identification of the inducible aerenchyma formation-associated genes of rice. By comparison of these microarray data and the identification of the genes up-regulated or down-regulated in common in maize and rice during aerenchyma formation, good candidate genes for functional analyses may be selected. The data should provide a basis for an understanding of the molecular mechanism of inducible lysigenous aerenchyma formation in plants.

Acknowledgements

We thank Drs T. D. Colmer, W. Armstrong, J. Armstrong, A. I. Malik, M. B. Jackson, J. Bailey-Serres, P. Perata, M. Sauter, B. Steffens, R. D. Hill, P. S. Schnable, H. Yoshioka, K. Takeda, A. Oyanagi, K. Kawaguchi, F. Abe, S. Arimura, K. Ohtsu and T. Abiko for stimulating discussions. We thank Dr Y. Mano and F. Omori for amplifying the seeds of maize (inbred line B73) and stimulating discussions. We thank H. Kamakura for help with the LM experiment and R. Motoyama for help with the microarray analysis. This work was supported in part by a grant from the Bio-oriented Technology Research Advancement Institution (Promotion of Basic Research Activities for Innovative Biosciences), a grant from the Ministry of Agriculture, Forestry and Fisheries of Japan, and grants-in-aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Ancillary