The maize (Zea mays L.) RTCS gene encodes a LOB domain protein that is a key regulator of embryonic seminal and post-embryonic shoot-borne root initiation

Authors


(F. Hochholdinger: fax +49 7071 29 5042; e-mail frank.hochholdinger@zmbp.uni-tuebingen.de or G. Taramino: fax +1 302 695 2726; e-mail graziana.taramino@cgr.dupont.com).

Summary

Maize has a complex root system composed of different root types formed during different stages of development. The rtcs (rootless concerning crown and seminal roots) mutant is impaired in the initiation of the embryonic seminal roots and the post-embryonic shoot-borne root system. The primary root of the mutant shows a reduced gravitropic response, while its elongation, lateral root density and reaction to exogenously applied auxin is not affected. We report here the map-based cloning of the RTCS gene which encodes a 25.5 kDa LOB domain protein located on chromosome 1S. The RTCS gene has been duplicated during evolution. The RTCS-LIKE (RTCL) gene displays 72% sequence identity on the protein level. Both genes are preferentially expressed in roots. Expression of RTCS in coleoptilar nodes is confined to emerging shoot-borne root primordia. Sequence analyses of the RTCS and RTCL upstream genomic regions identified auxin response elements. Reverse transcriptase-PCR revealed that both genes are auxin induced. Microsynteny analyses between maize and rice genomes revealed co-linearity of 14 genes in the RTCS region. We conclude from our data that RTCS and RTCL are auxin-responsive genes involved in the early events that lead to the initiation and maintenance of seminal and shoot-borne root primordia formation.

Introduction

An elaborate root stock architecture is required for efficient water and nutrient uptake of the developing plant. Maize (Zea mays L.) displays a complex root stock architecture composed of different root types formed during different phases of development (Feldman, 1994; Hochholdinger et al., 2004a,b). During embryogenesis a primary root is laid down at the basal pole of the embryo while a variable number of seminal roots are formed at the scutellar node. These roots are only relevant during the early stages of seedling development in maize (Hochholdinger et al., 2004a,b). At later stages of development an extensive post-embryonic shoot-borne root system forms the major backbone of the adult root stock. Shoot-borne roots are formed in whorls at consecutive nodes of the shoot and are endogenously initiated opposite to collateral vascular bundles (Martin and Harris, 1976). Nodal roots that are formed at underground nodes are designated crown roots while roots initiated in aboveground nodes are designated brace roots (Hetz et al., 1996). The maize root stock develops about 70 shoot-borne roots during its life cycle which are organized in approximately six whorls of densely packed underground crown roots and two to three whorls of aboveground brace roots (Hoppe et al., 1986). Crown roots which form the major part of the adult root stock are the basis for lodging resistance of the plants and are responsible for the major part of the water uptake via their lateral roots (McCully and Canny, 1988).

Thus far, two mutants of maize affected in shoot-borne root formation have been isolated. The monogenic recessive rt1 (rootless1) mutant (Jenkins, 1930) is lacking all shoot-borne roots at the higher nodes while there is only a slight difference in the number of crown roots at the first two nodes. In contrast, the rtcs (rootless concerning crown and seminal roots) mutant (Hetz et al., 1996) was identified by its complete lack of embryonically formed seminal roots and post-embryonically formed shoot-borne roots. Histological analyses revealed that the mutation in the RTCS gene blocks the initiation of the affected root types (Hetz et al., 1996). Thus, the primary root and its lateral roots remain the only root type in the rtcs mutant. Despite the lack of post-embryonic roots, mutant plants can be grown to maturity if they are protected from dehydration and lodging with external supports (Hetz et al., 1996).

Recently, a novel plant-specific gene family designated lateral organ boundaries (LOB; Iwakawa et al., 2002; Shuai et al., 2002) has been defined. Proteins of this family do not have a recognizable functional motif but contain a conserved LOB domain. Mutations in LOB domain genes helped to define their functions in leaf venation in Arabidopsis (Iwakawa et al., 2002), in patterning of stem cells in axiallary meristems of maize (Bortiri et al., 2006a) and in adventitious root formation in rice (Oryza sativa; Inukai et al., 2005; Liu et al., 2005).

In this paper, we report the positional cloning and characterization of the maize RTCS gene which encodes for a LOB domain protein involved in seminal and post-embryonic shoot-borne root formation in maize.

Results

The rtcs primary root is not affected in elongation and lateral root formation

Previous works on the rtcs mutant (Hetz et al., 1996; Hochholdinger et al., 2000; Sauer et al., 2006) focused exclusively on its developmental defects in seminal and shoot-borne root formation (Figure 1). Therefore, we subjected the single remaining primary root to detailed physiological analyses. First, we compared the length and number of lateral roots between wild-type and rtcs seedlings 10 days after germination. We did not detect any significant differences between the two genotypes for either measurement. Wild-type primary roots had an average length of 35.5 ± 0.7 cm while rtcs primary roots had an average length of 36.9 ± 1.1 cm. Similarly, 10-day primary roots of wild-type seedlings contained an average number of 222 ± 10 lateral roots, whereas the total number of lateral roots on the rtcs primary root was 203 ± 10. Thus, no difference was detectable between wild-type and rtcs seedlings concerning primary root elongation and lateral root density.

Figure 1.

 Phenotype of 10-day-old wild-type (a) and rtcs (b) seedlings. Close up of wild-type (c) and rtcs (d) coleoptilar nodes illustrate that the mutant does not form shoot-borne roots.
CR, crown root; SR, seminal root; PR, primary root; LR, lateral root. Size bars: (a), (b) 2 cm; (c), (d) 2.5 mm.

The rtcs primary root displays a delayed gravitropic response but no effect upon exogenous auxin application

We investigated the gravitropic response of wild-type and rtcs primary roots (Figure 2). While the gravitropic curvature in the first 2 h after induction was similar in both genotypes, rtcs primary roots showed a reduced gravitropic bending compared to the corresponding wild-type primary roots between 2.5 and 12 h after gravitropic stimulation. Since the gravitropic response is an auxin-mediated process (Perrin et al., 2005), we also analyzed the reaction of primary root length and the number of lateral roots on different types of exogenously applied natural [indole-3-acetic acid (IAA)] and synthetic [2,4-dichlorophenoxy acetic acid (2,4-D), α-naphthalene acetic acid (αNAA)] auxins. Exogenous auxin application between 10−8 and 10−6 m had a similar inhibitory effect on maize primary root elongation in the wild-type as well as in the rtcs mutant. Similarly, high concentrations (10−6 m) of all three exogenously applied auxins led to a similar reduction in the number of lateral roots in wild-type and rtcs primary roots (data not shown). This indicates that only subtle auxin-related defects concerning gravitropism are detectable in the primary root of the rtcs mutant.

Figure 2.

 Gravitropic response of the wild-type (black square) and the mutant rtcs (white square) primary roots.
Seedlings were grown in paper rolls for 3 days and subsequently turned by 90° (time 0). The curvatures were measured after the indicated time points. Each square represent the mean of 13 measurements. The mutant rtcs shows a significantly reduced gravitropic response between 2.5 and 12 h after induction. The asterisks indicate a significant differences in the t-test (two sided; unequal variance of the two populations) at a level α of 3%; error bars indicate the standard error of the mean.

RTCS gene cloning

The RTCS gene was identified via a map-based cloning strategy. The mapping population consisted of approximately 2000 plants of a segregating F2 population generated through intercrosses of the original mutant (background DK105) into the inbred line B73. Three hundred and sixty-seven plants in that mapping population were scored as rtcs mutants based on their lodging phenotype after 40 days in the field. Previous mapping information placed the rtcs locus on the short arm of chromosome 1 between the markers BNLG 1014 and BNLG 1429 (Hetz et al., 1996; Krebs et al., 1999). Our mapping population comprising 367 rtcs mutants yielded 34 recombinants with marker BNLG 1014 and 63 recombinants with marker BNLG 1429. In order to obtain additional genetic markers near the RTCS locus, simple sequence repeat (SSR) markers were retrieved in the interval between BNLG 1014 and BNLG 1429 from the maize genome database (http://www.maizegdb.org) and deployed for genotyping the F2rtcs plants. In particular, marker UMC 1685 yielded seven recombinants while marker UMC 1160 revealed 11 recombinants, indicating a distance of 0.95 and 1.5 cM from the rtcs locus, respectively (Figure 3a). Markers UMC1685 and UMC1160 were physically positioned by hybridization onto a single maize contig (data not shown). The physical distance between the two markers comprised approximately 10 overlapping bacterial artificial chromosomes (BACs). By generating a cleaved amplified polymorphic sequence (CAPS) marker (CapsB74.m9) from the end sequence of BAC clone b74.m9, the genomic region comprising the rtcs locus was confined to two overlapping BAC clones, b35.m15 and b74.m9. CapsB74.m9 was mapped very close to RTCS, at a distance of 0.13 cM (separated by one recombination breakpoint). Based on this information, we sequenced the BAC clone b74.m9. The 170 kb of BAC b74.m9 was searched for the presence of open reading frames (ORFs) and four genes were identified and considered as RTCS candidates (Appendix S1 highlighted in grey). By sequencing wild-type and mutant alleles of these putative genes, we identified a 5-bp insertion in the mutant allele of one of the genes carrying a LOB domain (Iwakawa et al., 2002; Shuai et al., 2002). The insertion, which causes a frame-shift and introduces a premature stop codon, is 227 bp downstream of the putative ATG start codon (Figure 3b). We therefore concluded that this gene represents the RTCS locus.

Figure 3.

Map based cloning of the RTCS gene. (a) Fine mapping of the RTCS gene between the markers UMC 1685 and capB74.m9.
(b) Gene and protein structure of the RTCS gene and the RTCS protein and the mutation sites of the three alleles rtcs-1 to rtcs-3.
(c) Expression of RTCS mRNA in different rtcs alleles. Reverse transcriptase-PCR was conducted using RTCS-specific primers on RNA extracted from 10-day-old coleoptilar nodes. While rtcs-1 and rtcs-2 generate a rtcs transcript, no rtcs transcript can be detected in allele rtcs-3. Histone 2A was used as a control.
(d) Dendrogram of the closest maize and rice relatives of RTCS and the RTCS flanking gene, RTCN. The phylogenetic analysis was performed as described in Zimmermann and Werr (2005).

Confirmation of rtcs cloning via independent alleles

Confirmation that this gene actually encodes RTCS was achieved via sequencing of two independent rtcs alleles (rtcs-2 and rtcs-3) which were isolated in F2 segregating populations of Mutator and Ac element stocks, respectively. Both alleles displayed the same phenotype as the original allele rtcs-1, i.e. no seminal or shoot-borne root formation. Crosses of homozygous mutants of rtcs-2 and rtcs-3 with rtcs-1 exclusively yielded mutant progenies, confirming through the lack of complementation that these mutants are allelic with each other. The rtcs-2 allele contained two mutational sites, one insertion in the promoter region (368 bp upstream of the start codon) and a A/G transition at the splicing acceptor site between the intron and exon 2 (Figure 3b). The inserted sequence appeared to be novel with a size of 578 bp (GenBank accession number EF051735) but carried the hallmark of a non-autonomous transposable element with a 18-bp terminal inverted repeat (TIR) sequence, which differs in 1 bp between the 5′ and 3′. The novel transposable element is associated with an 8-bp direct duplication at the site of insertion. Within this element, there is another direct duplication of 26 bp and a 35-bp transposase footprint. The second mutation in rtcs-2 resulted in a shift of the splicing acceptor site to the next available AG dinucleotide in exon 2. The altered splice acceptor site was verified by sequencing a RT-PCR product of rtcs-2. This mutation led to the deletion of 15 bp in the transcript, i.e. five amino acid residues in the RTCS polypeptide. The accumulation of the rtcs transcript was only partially affected in rtcs-2 (Figure 3c). The rtcs-3 mutant contained a canonical Ac transposon insertion in exon 2 of the RTCS gene 33 bp upstream of the stop codon (Figure 3b). While rtcs-1 and rtcs-2 mutants generated RTCS transcript, no transcript was detected in rtcs-3 (Figure 3c). Thus, rtcs-3 appears to be a null allele of the rtcs gene.

RTCS gene structure

A full-length cDNA of 735 bp was retrieved from a lambda cDNA library that was prepared from wild-type coleoptilar nodes (GenBank accession number EF051732). Alignment of the RTCS cDNA and genomic sequences revealed that the RTCS gene (Figure 3b) is composed of two exons, separated by a 96-bp intron, thus resembling the structure of other known members of the LOB domain gene family in Arabidopsis and rice (Inukai et al., 2005; Iwakawa et al., 2002; Liu et al., 2005; Shuai et al., 2002). The RTCS gene encodes a 25.5-kDa protein with 244 amino acid residues, including a 102-amino-acid long LOB domain. The LOB domain contains a conserved C motif, which can possibly form a new zinc finger, typical of the DNA-binding domain of transcription factors, and a leucine-zipper-like domain, possibly involved in dimerization (Figure 3b).

Maize–rice micro-synteny

On the macro-synteny level, chromosome 3 of rice shows co-linearity with maize chromosome 1, which contains the RTCS gene (Gale and Devos, 1998). Since the maize genome has not yet been fully sequenced in this region, we investigated the co-linearity of the region on the micro-synteny level. The public rice BAC OSNJBb0050N02 (AC105734), containing the syntenic region of the RTCS gene, spans a 133-kb long region, including 17 ORFs (OSNJBb0050N02.1-17). We identified the closest maize homologs of all of these 17 ORFs via expressed sequence tags (ESTs). Through PCR with primers designed from these EST sequences, we were able to identify 13 of these genes, including RTCS, in five minimally overlapping maize BACs covering this region (Appendix S1). Interestingly, another LOB domain gene in the region, which we designated RTCS NEIGHBOR (RTCN) (GenBank accession number EF051734), is also found in the corresponding rice region, next to the RTCS ortholog. The RTCN ORF has 44% nucleotide identity to RTCS and the putative RTCN protein shows 72% identity to the RTCS LOB domain region. To explore the possibility that the failed amplification of the four remaining rice genes could be due to the incorrect identification of orthologous EST sequences, we performed dot blot analyses on the five overlapping maize BACs using those four rice genes as hybridization probes. This approach revealed an additional gene (OSNJBb0050N02.17, a phosphate transporter) conserved between rice and maize. However, the remaining three genes failed to be detected. By searching the public rice MPSS database (Nakano et al., 2006) we could identify a low level of expression only for one of the three annotated genes (OSNJBb0050N02.16) that were not identified in the maize BACs. All of the three genes were annotated in rice as hypothetical proteins. Thus, two of the putative genes might be pseudogenes. Interestingly, not only the presence but also the order of the 14 syntenic genes was completely maintained between the two species (Appendix S1).

RTCS is duplicated in the maize genome

The maize genome is a segmental allotetraploid in various regions (Gaut and Doebley, 1998). Since the region on chromosome 1S where the RTCS gene is located shows a duplication on chromosome 9 (Bruggmann et al., 2006), we investigated whether this region harbors a gene similar to RTCS. Indeed, a genomic sequence that was 84% identical to the RTCS gene at the nucleotide level was retrieved via PCR and was designated RTCS-LIKE (RTCL). This gene is located on BAC ZMMBBb0193N08 (AC149475), which physically maps on maize chromosome 9 between marker BNLG128 and UMC1137 (http://www.maizegdb.org). Thus, we concluded that RTCL is the homolo gene of RTCS. The RTCL transcript has a size of 699 bp (GenBank accession number EF051733) and encodes a 232-amino-acid protein of 24.7 kDa. The RTCS and RTCL proteins share an overall identity of 72%, while the LOB domain displays an identity of 88% between these proteins. Based on blast analysis, RTCL does not appear to be present in rice. Hence, while the maize RTCS gene has been duplicated during evolution, no duplication occurred to the rice CRL1 gene.

Phylogeny of RTCS relatives

A phylogenetic analysis of the closest relatives of the RTCS gene in maize and rice resulted in the dendrogram displayed in Figure 3(d). The maximum-parsimony algorithm and the neighbor-joining approach resulted in similar trees. The maize RTCS and the syntenic rice gene CRL1 (Inukai et al., 2005), which is allelic with the rice ARL1 gene (Liu et al., 2005), were the closest relatives. While 100 of 102 amino acids of the LOB domain are conserved between RTCS and rice CRL1, only 51% of the amino acids outside the LOB domain are identical between RTCS and CRL1. The product of RTCL, the duplicated RTCS homolog, is located in a sub-branch of the RTCS/CRL1 cluster. The ZmCRLL1 gene product, previously believed to be the maize ortholog of the rice CRL1 (Inukai et al., 2005), is grouped in a branch distinct from the RTCS/CRL1 cluster. The LOB domain protein RTCN, whose corresponding gene is located next to the RTCS gene and its rice ortholog OsCRLL1, encoded by the gene next to CRL1, together form an independent clade. In summary, the maize RTCS and rice CRL1 are the most closely related genes and the maize RTCL gene is their closest relative.

The RTCS gene is preferentially expressed in root tissue at very low levels

Massively parallel signature sequencing (MPSS) technology allows for the quantification of 17-bp sequences in populations of 2 × 105 to 2 × 106 cDNAs. These 17-bp signature sequences often correspond to unique cDNAs, thus allowing for the quantification of the abundance of a particular cDNA in a sample representing a particular organ and developmental stage in a defined genetic background (Christensen et al., 2003).

The RTCS signature was present at low levels, between 3 and 13 p.p.m. (parts per million) in libraries related to roots [Figure 4a: primary roots 55 h after germination (VE stage), nodal roots at V5 stage, nodal root tips at V6 stage, whole roots at V6 and V12 stages]. The different development stages of maize (V-stages) are defined in: http://www.extension.iastate.edu/pages/hancock/agriculture/corn/corn_develop/CornPlantStages.html. Interestingly, the RTCS signature was also found, at a level of 4 p.p.m., in a tassel library derived from spikelets harvested at the quartet stage. Expression of RTCS in spikelets might be indicative of the competence of spikelets to form roots under stress conditions as previously observed in rice (Jones and Pope, 1942; Taniguchi and Futsuhara, 1988). The RTCL gene displayed a similar expression pattern in primary roots, whole roots and in the same spikelet library. In addition, the highest level of expression of the RTCL gene was found in the scutellum. Thus, the RTCS and RTCL genes show some overlap in root-specific gene expression. Interestingly, although the RTCS mutant is affected in the embryonic initiation of seminal lateral roots, no expression of the RTCS gene was detected in any of the embryo libraries. This might imply that RTCS is expressed at extremely low levels below 1 p.p.m., or alternatively that it is only expressed at very early or very short phases of embryo development. The LOB domain gene RTCN located next to RTCS on chromosome 1S (Appendix S1) is expressed in most tissues examined including the kernels, silks and roots at different developmental stage. Hence, while RTCS and RTCL display preferential expression in roots and show overlap in their expression patterns RTCN displays different expression patterns indicating functions different from the duplicated RTCS and RTCL genes.

Figure 4.

Expression of the RTCS, RTCL and RTCN genes (a) MPSS analysis of the duplicated genes RTCS and RTCL and the RTCS flanking gene RTCN.
(b) Estimate of relative RTCS versus RTCL ratios in different tissues via sequencing of PCR products generated with oligonucleotide primers deduced from completely conserved sequences between RTCS and RTCL that flanked variable regions between the two genes.
(c) In situ hybridization experiments of 5- and 10-day-old coleoptilar nodes of wild-type plants indicating exclusive RTCS expression in emerging crown root primordia. Boxed parts of the safranin/fastgreen stained cross sections indicate the corresponding section displayed in the sense and antisense in situ hybridization experiments.
(d) Western blot analysis of 5- and 10-day-old wild-type and mutant rtcs coleoptilar node protein extracts with an antibody raised against RTCS peptides.
(e) Auxin induction of the RTCS and RTCL genes between 1 and 3 h. Histone 2A was used as a control.

RTCS displays a higher expression than RTCL in the coleoptilar node and primary root

The rtcs mutation completely inhibits the initiation of shoot-borne and seminal roots. However, while the mutation also causes a reduced gravitropic response, primary roots display normal elongation, lateral root formation and response towards application of exogenous auxin. We therefore examined the possibility of differential expression of two closely related genes, RTCS and RTCL, in the primary root and coleoptilar node, which might correlate with the root-type-specific phenotype of the mutant. For this purpose, we isolated poly(A) RNA from 10-day-old coleoptilar nodes and 5- and 10-day-old primary roots and performed RT-PCR analysis using a primer set designed in the sequence conserved between RTCS and RTCL, hence amplifying the cDNA from both genes at the same time (for primer sequences, see Appendix S2). Differential expression was assessed by sequencing at least 420 clones of each PCR fragment. The RTCS and RTCL cDNA molecules were distinguished by sequence polymorphisms present between the two genes. In all tested tissues, the RTCS gene displayed a significantly higher expression than the RTCL gene (Figure 4b). This result was in accordance with other semi-quantitative RT-PCR experiments we performed with gene-specific primer sets, which showed at least a 10-fold difference in expression levels of two genes (data not shown). Despite the predominant RTCS expression, there were differences in RTCL expression level in the examined tissues. While in 10-day-old coleoptilar nodes only 6% of the analyzed clones were RTCL, in 10- and 5-day-old primary roots 9% and 12% of the analyzed clones were of RTCL origin, respectively. Hence, while RTCS expression in all analyzed tissues was significantly higher than RTCL expression, RTCL displayed higher expression levels in primary roots than in coleoptilar nodes.

RTCS is expressed in emerging crown root primordia

Since the major defect in rtcs is the complete absence of shoot-borne roots, RTCS gene expression was investigated via in situ hybridization experiments in coleoptilar nodes during the early stages of crown root initiation (Figure 4c). Analyses were performed on coleoptilar nodes of wild-type seedlings 5 and 10 days after germination. In rare instances, crown roots were already initiated in 5-day-old coleoptilar nodes (Figure 4c). In 10-day-old coleoptilar nodes, crown root meristems were always present in the wild type. Expression of RTCS coincides with the newly formed root primordia.

The RTCS protein is expressed in 10-day-old wild-type coleoptilar nodes

In order to compare the expression of the RTCS protein in wild-type and rtcs coleoptilar nodes, we performed Western blot analyses with protein extracted from 5- and 10-day-old coleoptilar nodes and a specific antibody raised against a short region of the RTCS protein. Each protein extract was prepared from two independent biological replicates and loaded on two adjacent lanes (Figure 4d). In protein extracts of wild-type coleoptilar nodes, a 25.5 kDa band was detected at 10 days. At 5 days, the protein was not detected, which is in accordance with the observation that crown roots are typically not initiated at this early stage of development. Hence, accumulation of RTCS protein coincides with the development of shoot-borne root primordia.

The RTCS and RTCL promoters contain auxin response elements and the genes are auxin inducible

Promotor analysis of the RTCS and RTCL genes via the PLACE database (Higo et al., 1999) revealed putative auxin response element (AuxRE) motifs in both genes; six elements in the 3-kb RTCS promoter region and 13 elements in the 3-kb RTCL promoter region. We therefore tested the auxin inducibility of the two genes. In line with the previous expression studies (Figure 4b) RTCS displays a much higher expression before induction than RTCL. The expression of both genes in 10-day-old primary roots was induced by the application of 5 μm of the synthetic auxin αNAA after 1 h of treatment and remained induced after 3 h (Figure 4e). Auxin inducibility and the presence of AuxRE in the promoters of RTCS and RTCL indicate that these genes might belong to the family of early auxin response genes (Guilfoyle et al., 1998).

Discussion

In recent years, a number of mutants affected in various aspects of maize root formation have been identified (reviewed in: Feix et al., 2002; Hochholdinger et al., 2004a,b, 2005). The phenotypes of these mutants indicated that the formation of the maize root system is controlled by development-specific regulatory networks that partially overlap during the formation of the different root types (Hochholdinger et al., 2004a,b). So far, only one gene of these mutants (rth1) has been cloned (Wen et al., 2005). The RTH1 gene encodes a sec3 homolog that is involved in root hair elongation in maize (Wen et al., 2005). Here we report the cloning and characterization of RTCS, which is involved in the post-embryonic initiation of crown and brace roots from consecutive shoot-nodes and the embryonic initiation of the seminal roots at the scutellar node (Hetz et al., 1996). Besides rtcs, the rum1 mutant (Woll et al., 2005) is the only other known maize mutant that is affected in both embryonic and post-embryonic root formation (Woll et al., 2005). However, while both mutants are affected in seminal root initiation, rum1 is defective in the initiation of the post-embryonic lateral roots, and rtcs is defective in the initiation of the post-embryonic shoot-borne roots.

The recent availability of densely populated genetic maps of maize (http://www.maizegdb.org), the complete sequencing of the rice genome (Goff et al., 2002) and the synteny between the maize and rice genomes (Gale and Devos, 1998) have made it increasingly feasible to clone maize genes via a map-based cloning approach (Bortiri et al., 2006b). We took advantage of these new technical innovations and cloned the RTCS gene via a map-based approach. Interestingly, our micro-synteny analysis of the RTCS genomic region demonstrated not only the conservation of 14 genes that are positioned in this region but also that the order of these genes is completely maintained between these two grass genomes. The RTCS gene is located on chromosome 1S and encodes a 25.5-kDa protein with a single LOB domain. The LOB domain has recently been identified among several plant proteins (Iwakawa et al., 2002; Shuai et al., 2002). So far, gene products containing LOB domains have been implicated in various aspects of plant development including leaf venation in Arabidopsis (Iwakawa et al., 2002), patterning of axillary meristems in maize (Bortiri et al., 2006a), and adventitious root formation in rice (Inukai et al., 2005; Liu et al., 2005). Since RTCS maps in the syntenic region of the two allelic rice genes CRL1 (Inukai et al., 2005) and ARL1 (Liu et al., 2005), RTCS is most probably the maize ortholog of this gene. The proteins encoded by the rice gene CRL1 and maize RTCS display 98% sequence identity within the LOB domain but only 51% identity outside the LOB domain. Since both genes are affected in shoot-borne root formation, this might indicate that the sequence of the LOB domain is of greater importance for the developmental function of LOB domain proteins than their variable parts.

Interestingly, there are a number of phenotypic differences between the rice crl1 and maize rtcs mutants. While both mutants display defects in shoot-borne root initiation, the maize phenotype is not only restricted to this post-embryonic root type but can already be detected during embryogenesis by the lack of seminal root primordia (Hetz et al., 1996). This difference might be due to the fact that rice does not produce seminal roots, presumably because CRL1 is repressed in seminal root primordia. On the other hand, the rice mutant has altered crown root and lateral root formation (Inukai et al., 2005), while the rtcs mutant has no defect in lateral root formation on the primary root, the only root produced. A possible explanation for this difference could be the fact that the RTCS gene was duplicated during the evolution of the maize genome. This duplication created the RTCS-LIKE (RTCL) gene on chromosome 9 with overlapping functionality to RTCS. In contrast, rice has only a single-copy gene, CRL1. Our analysis of RTCS and RTCL expression in coleoptilar node and primary roots revealed the higher expression of RTCS over RTCL in all analyzed tissues. However, the expression of RTCL in developing primary roots was 1.5 to 2 times higher than in the coleoptilar node. This might imply that RTCS is relevant for the initiation of shoot-borne root formation and that low levels of RTCL in the coleoptilar node cannot compensate for the lack of RTCS transcript in the rtcs mutant. On the other hand, RTCS and RTCL might have a cooperative, partly redundant function in primary root development, which might be indicated by the higher level of RTCL expression in the primary root. This could explain why only the auxin-regulated gravitropic response of the rtcs mutant was impaired in the primary root while other auxin-regulated responses such as primary root length and lateral root density were not affected.

As previously demonstrated, the rtcs mutation specifically blocks shoot-borne root initiation in the coleoptilar node. However, the competence for cell division is still maintained in the coleoptilar node, as shown via the expression of the competence marker cdc2 (Hochholdinger and Feix, 1998a) and the initiation of lateral shoots (tillers) in the rtcs coleoptilar node when introgressed into inbred lines like tiller-rich gaspe flint (Hochholdinger and Feix, 1998b).

Both RTCS and RTCL genes are auxin inducible and contain auxin response elements in their promoter. This might indicate that RTCS functions as an auxin-responsive factor that is involved in auxin-mediated initiation of shoot-borne root primordia. Recent comparative proteome analyses of wild-type and mutant rtcs coleoptilar nodes revealed that several components involved in auxin signal transduction were differentially expressed between these genotypes (Sauer et al., 2006). The cloning of the RTCS gene now allows the interpretation of these data in the context of auxin signaling and shoot-borne root formation. Auxin-binding protein 1 (ABP1) is believed to be a component of the auxin signaling network by acting as a high-affinity auxin receptor (Chen, 2001). Auxin-binding protein 1 was highly expressed in wild-type coleoptilar nodes 5 days after germination and was subsequently down-regulated to low expression levels between 5 and 10 days after germination, thus indicating a feedback inhibition of this gene during early coleoptilar node development. In rtcs, the feedback inhibition of this putative auxin receptor, which might act in the earliest stages of auxin signal transduction, was not detected, but instead the gene and protein maintained its high expression level (Sauer et al., 2006). On the other hand, high auxin concentrations can stimulate cell division via a second, unidentified auxin pathway, possibly mediated by a low-affinity auxin receptor (Chen, 2001). This second pathway is most likely coupled with a heterotrimeric G-protein and it appears that ABP1 and G-protein pathways must crosstalk to bring about normal development (Chen, 2001). This model is supported by the observation that a subunit of a heterotrimeric G-protein was down-regulated in 10-day-old wild-type coleoptilar nodes (Sauer et al., 2006). Further, it has been demonstrated that calmodulin is involved in auxin signal transduction by binding to the SAUR class of early auxin-responsive genes (Yang and Poovaiah, 2000). Since calmodulin is down-regulated in the rtcs mutant (Sauer et al., 2006), the RTCS gene appears to be an early auxin-responsive gene, directly or indirectly interacting with this signal transduction component.

In summary, the RTCS gene plays a central role in the auxin-mediated initiation of seminal and shoot-borne roots. This is suggested by its auxin inducibility, its AuxRE sequence promoter elements and proteomics data that indicate differential expression of various components of the auxin signal transduction machinery. We propose that RTCS represents an auxin response factor that inhibits the upstream ABP1 and a trimeric G-protein while stimulating downstream calcium-dependent signaling via the up-regulation of calmodulin. Future experiments that modulate RTCS expression and identify additional interaction partners might help optimize cereal root systems to improve yield performance.

Experimental procedures

Identification of novel rtcs alleles

We identified two mutants similar to the rtcs-1 reference mutant phenotype in our regular screening experiments of transposon-tagged F2 populations of Mu-active × B73 and Ac × tester crosses, respectively. The rtcs-2 mutant was identified in the BC1F2 progeny developed by crossing B73 with BT94 8D-10 (Mu-active line, which was originally provided by Dr Mike Freeling, University of California, Berkeley) in 2004. The rtcs-3 mutant was identified in 2005 in the selfed progenies of the cross W23 P-w*YCO63 × W22 r1.sc.m3.Ds. The rtcs-2 and rtcs-3 mutants were crossed with the reference rtcs-1 mutant allele and proved allelism.

Phenotypic and histological analysis of RTCS

Seeds of wild-type and rtcs were germinated in paper rolls (Anchor Paper, http://www.anchorpaper.com) as previously described (Hetz et al., 1996) in a plant growth chamber at 60% humidity and 28°C, under a 16-h light, 8-h dark regime. Determination of the lateral root density was performed by digitizing primary root pictures via a scanner (HP Scanjet 7400C, Hewlett-Packard Company, http://www.hp.com/) and subsequently analyzing the scans with Image Pro Express software (Media Cybernetics Inc., http://www.mediacy.com/). The gravitropic curvature of the primary root was measured in seedlings grown in paper rolls at 28°C in the dark as previously described (Woll et al., 2005). Histological analyses and investigations of the impact of the application of exogenous auxin on root development were performed as described in Woll et al. (2005).

Map-based cloning and validation of the RTCS gene

The RTCS gene was initially mapped based on the information provided by Hetz et al. (1996) and Krebs et al. (1999) with SSR and CAPS markers, using 367 homozygous mutant rtcs plants derived from crosses between the inbred lines DK105 and B73. The RTCS gene was finally fine-mapped to BAC b74.m9 covering a 170-kb region of the maize Mo17 genome. The BAC b74.m9 was shotgun sequenced according to Tarchini et al. (2000) and homology searches against public and private (DuPont) databases were used to identify candidate genes in the region. A full-length cDNA clone was obtained by screening a lambda cDNA library generated from wild-type coleoptilar nodes with the ZAP-cDNA cloning kit (Stratagene, http://www.stratagene.com/) according to the manufacturer’s instructions. Total RNA was isolated via an mRNA purification kit (Amersham Pharmacia, http://www5.amershambiosciences.com/) from 14-day-old B73 coleoptilar nodes. Southern blot analyses comparing rtcs-2 and wild-type genomic DNA revealed a 7-kb EcoRI polymorphic band tightly co-segregating with the mutant phenotype in a population of 64 segregating plants. The rtcs-2 mutant allele was cloned by preparing a size-fractionated library in the lambda Zap II cloning vector (Stratagene). A restriction map of the 7.0-kb EcoRI cloned fragment was prepared and the 7.0-kb clone was further subcloned into smaller pieces in the pBlueScript vector (Stratagene) and sequenced. The rtcs-3 mutant allele was amplified using a gene-specific primer from the RTCS gene and an Ac-specific primer (for primer sequences see Appendix S2). The insertion site was identified by cloning the amplified fragment in a TA vector (Invitrogen, http://www.invitrogen.com/) and subsequent sequencing. Similarity searches were performed with the blastn algorithm (Altschul et al., 1997). Sequences were aligned with the DNAstar program (http://www.dnastar.com/). The RTCL gene was identified via PCR using cDNA generated from RNA of 10-day-old primary roots of the inbred line B73 and two primers that were deduced from the coding sequence of the maize RTCS gene (for primer sequences, see Appendix S2).

Micro-colinearity

A set of 10 overlapping clones surrounding the BAC clone b74.m9, containing the RTCS gene, were retrieved based on the contig information of the DuPont database. tblastn analyses were performed on the DuPont EST database to find the best corn homologous sequences of the 17 ORFs annotated in the rice BAC OSNJBb0050N02 (GenBank accession numbers: DN203700, EB641902, DV507769, EC897544, DV025880, DN224414, CK827367, DV169288, CO533354, EE026732, DV528420, AC148166, DR819256, BM500338). Subsequently, primers were designed to PCR amplify the 17 genes using the 10 overlapping maize BACs as a DNA template (for primer sequences see Appendix S2). The order of the maize genes was determined based on the presence/absence of the PCR amplification on the different overlapping BACs.

For the four rice genes whose orthologous corn genes did not show any PCR product, dot blot hybridizations were performed using standard procedures (Sambrook et al., 1989). Fifty nanograms of DNA from a subset of five minimally overlapping maize BAC clones (Figure 3b) was spotted on a Hybond N+ membrane (Amersham) and hybridized with probes deduced from the four sequences: OSNJBb0050N02.05, OSNJBb0050N02.08, OSNJBb0050N02.16 and OSNJBb0050N02.17.

In situ hybridization experiments

For in situ hybridization experiments, maize coleoptilar nodes were fixed in 4% paraformaldehyde for 12 h at 4°C. After dehydration in a graded ethanol series and subsequent clearing in a graded Histoclear series (National Diagnostics, https://http://www.nationaldiagnostics.com/) the specimens were embedded in paraffin and 10–12 μm sections were prepared using a Leica RM2145 microtome (Leica, http://www.leica-microsystems.com/). Samples were deparaffinized using Histoclear and dehydrated in an ethanol series. Hybridization was performed with 98-bp digoxigen in-labelled RNA sense and antisenseprobes of the 3′-untranslated region (UTR) of RTCS (positions 784–882 bp of the full-length cDNA) amplified via PCR and cloned into the pGEM-T vector (Promega, http://www.promega.com/) in both orientations using the T7 primer for in vitro transcription according to the manufacturer (Roche; http://www.roche.de). After 2 h of pre-hybridization in 1 × SPB buffer [50% formamide, 1.25 ×  salts (10 ×  salts: 3 m NaCl, 0.1 M TRIS-HCl, pH 6.8, 0.1 M sodium phosphate, pH 6.8, 50 mm EDTA; TRIS = 2-amino-2-(hydroxymethyl)-1,3-propanediol), 12.5% dextransulfate, 1.25 mg ml−1 tRNA, 0.625 mg ml−1 polyA+, 1.25 ×  Denhardt’s solution], hybridization took place at 45°C for about 16 h. Slides were washed afterwards (5 ×  SSC/ 50% formamide) and treated with RNaseA (10 μg ml−1) at 37°C for 30 min and washed several times with NTE (0.5 m NaCl, 10 mm TRIS-HCl pH 8.0, 0.5 mm EDTA). The immunological detection was conducted by incubating the specimens with nitro-blue tetrazolium chloride (NBT)/5-bromo-4-chloro-3′-indolyphosphate p-toluidine salt (BCIP) according to the manufacturer (Roche). Samples were analyzed under an Axioplan2 microscope (Zeiss, http://www.zeiss.com/) and results were documented with a ProgRes C14 camera (Jenoptik, http://www.jenoptik.com/).

RT-PCR

Reverse trancriptase-PCR was performed with cDNA that was synthesized with Superscript III (Invitrogen) reverse transcriptase from 1 μg DNase treated total RNA. Polymerase chain reaction was performed in a Perkin-Elmer 9700 thermocycler using the GC-2 Advantage kit (BD Biosciences, http://www.bdbiosciences.com/) and a PCR program of 94°C for 3 min, followed by 27 cycles of 94°C/30 sec, 58°C/30 sec, 68°C/1 min, and a final step of 68°C for 3 min.

For the auxin induction experiment, B73 seedlings were germinated in paper rolls for 10 days in a phytochamber as described above, then transferred to a 5 μmαNAA solution (αNAA prepared from 0.1 mαNAA in 80% ethanol). Control plants were grown in distilled water. Ribonucleic acid was isolated from primary roots after 0, 1, 2 and 3 h of αNAA exposure and from control plants at the same time points. The PCR reactions were performed with RTCS- and RTCL-specific oligonucleotide primers. Histone 2A (GenBank AAB04687) was used as a constitutively expressed gene (for primer sequences see Appendix S2).

Western blot

Coleoptilar nodes of 5- and 10-day-old wild-type and rtcs seedlings germinated in paper rolls in a phytochamber under the conditions described above were excised with a razor blade, immediately frozen in liquid nitrogen and subsequently pulverized with a mortar and pestle. The pulverized tissue was resuspended in a solution containing 8 m urea, 50 mm TRIS pH 8.0 and 20 mmβ-mercaptoethanol. Thirty micrograms of protein per lane was separated on a 12% SDS-PAGE gel. Proteins were transferred to nitrocellulose membranes via the semi-dry blotting technique for 1 h at 100 mA per gel (PerfectBlue ‘Semi-Dry’ Elektroblotter, PeqLab, http://www.peqlab.de/). Membranes were blocked with TBS-T (137 mm NaCl, 192 mm glycine, pH 7.6, 0.5% Tween-20) and 3% non-fat dried milk (Sucofin, Zeven) for 1 h. Polyclonal RTCS-specific antibodies were generated against the peptide HAFEQAGADDDDGRQG which is located in the variable part of the RTCS protein outside the LOB domain. Immunoreaction with this RTCS-specific primary antibody (1:10 000) was performed for 2 h at room temperature (RT; 25°C). The membrane was subsequently washed three times for 10 min in TBS-T. The secondary antibody (1:2500 dilution of anti-rabbit antibody conjugated with horseradish peroxidase) was incubated for 1 h at RT. Subsequently, the membrane was washed four times for 10 min in TBS-T. Chemiluminescent detection of signals was performed with the SuperSignal West Femto kit (Pierce, http://www.piercenet.com).

Novel materials described in this publication may be available for non-commercial research purposes upon acceptance and signing of a material transfer agreement. In some cases such materials may contain or be derived from materials obtained from a third party. In such cases, distribution of material will be subject to the requisite permission from any third-party owners, licensors or controllers of all or parts of the material. Obtaining any permissions will be the sole responsibility of the requestor. Plant germplasm and transgenic material will not be made available except at the discretion of the owner and then only in accordance with all applicable governmental regulations.

Acknowledgements

We thank Angela Dressel and Caroline Marcon (both University of Tuebingen) for excellent technical assistance and management of the experimental field, Drs Victor Llaca and Stefan Deschamps (both DuPont, Wilmington, DE) for sequencing BAC clones, and Heidi Ridnour (DuPont) for producing antisera. Root research in F.H.’s laboratory is supported by the Deutsche Forschungsgemeinschaft (DFG; award HO2249/4), the framework program ‘heterosis in plants’ (award HO2249/6), the DFG Sonderforschungsbereich 446 ‘Mechanisms of cell behavior in eukaryotes’ and a research grant by DuPont.

Ancillary