Activation tagging of the two closely linked genes LEP and VAS independently affects vascular cell number

Authors

  • Eric Van Der Graaff,

    Corresponding author
    1. Institute of Plant Biology, University of Zürich, Zollikerstrasse 107, CH-8008 Zürich, Switzerland, and
    Search for more papers by this author
    • Current address: University of Cologne, Botanical Institute II, Gyrhofstrasse 15, 50931 Cologne, Germany.

  • Paul J.J. Hooykaas,

    1. Clusius Laboratory, Institute of Molecular Plant Sciences, Leiden University, Wassenaarseweg 64, 2333 AL Leiden, the Netherlands
    Search for more papers by this author
  • Beat Keller

    1. Institute of Plant Biology, University of Zürich, Zollikerstrasse 107, CH-8008 Zürich, Switzerland, and
    Search for more papers by this author

* For correspondence (fax +49 221 470 5039; e-mail E.Graaff@uni-koeln.de).

Summary

The complex dominant Arabidopsis thaliana mutant lettuce (let) shows the conversion of the leaf petiole into a leaf blade caused by an ectopic leaf blade formation. This is the result of the activation tagging of the LEAFY PETIOLE (LEP) gene encoding an AP2/EREBP-like transcription factor. Here, we report that in addition to this leafy petiole phenotype, the size of the vascular bundles is increased in all aerial organs in let as a result of an increase in the number of xylem, phloem (pro)cambial and pericycle cells. This vascular phenotype is caused by activation tagging of the two genes VASCULAR TISSUE SIZE (VAS) and LEP. These genes are closely linked and arranged in tandem. Activation tagging of LEP only caused a specific increase in the number of xylem cells. This increased xylem cell number, together with the ectopic leaf blade formation, indicates that LEP functions as a cell division-promoting factor. The activation tagging of VAS only resulted in a specific increase in phloem (pro)cambial and pericycle cells. We conclude that activation tagging of LEP and VAS results in additive phenotypes. Insertional mutants for LEP and VAS display wild-type vascular development, indicating the relevance of activation tagging for functional analysis of novel genes involved in plant development.

Introduction

Plant vascular tissue consists mainly of xylem, responsible for the transport of water and dissolved nutrients and phloem, responsible for transporting photosynthetic products. Vascular tissue is also important for the transport of signal transducing molecules and confers mechanical strength to the plant (Fukuda, 1997). During the growth of plants, xylem and phloem cells differentiate from new cells that are continuously produced from the meristematic (pro)cambial cells by the process of cell division (Aloni, 1987). After the elongation of organs ceases, the differentiation of the primary phloem and xylem is completed and secondary thickening is initiated (Shininger, 1979). A panel of the meristematic tissue between the primary phloem and xylem does not differentiate and prolongs its mitotic activity, thereby forming the cambium. In stems, the meristematic activity can spread laterally from the vascular bundles into the interfascicular parenchyma creating a closed sheet of meristematic tissue (Dengler, 2001; Shininger, 1979). The continuous cell division activity in the cambium serves for its self-maintenance and the derivatives on its flanks differentiate into secondary phloem and xylem (Dengler, 2001; Savidge, 2001). In leaves, the cell division activity of the procambium cells are limited to the early stages of primordium development (Aloni, 2001) and, like in species lacking secondary vascular growth, all procambial cells are consumed during vascular differentiation (Dengler, 2001).

The study of the development of vascular tissue in plants and in the Zinnia tracheary element differentiation system (reviewed in Aloni, 2001; Kuriyama and Fukuda, 2001) showed that many plant hormones are involved in vascular tissue formation. Auxin seems to be one of the major controlling signals for vascular tissue differentiation (Aloni, 2001) and the pre-patterning of vascular strands (reviewed in Berleth and Mattson, 2000; Dengler, 2001), while brassinosteroids (BR) play an important role in tracheary element differentiation (Yamamoto et al., 2001). In Arabidopsis mutants defective in BR synthesis, the number of xylem cells is decreased, whereas that of phloem can be either normal or increased compared to wild type (Choe et al., 1999; Szekeres et al., 1996). In addition, the interfascicular parenchyma cells can be reduced in number or lacking (Choe et al., 1999). This suggests that BR can initiate vascular tissue specification and/or regulate the balance between xylem and phloem cell number in the vascular bundles (Kuriyama and Fukuda, 2001).

Despite the importance of vascular tissue for plant development, our understanding of the processes in vascular tissue formation is limited and few genes have been characterised that influence the cell number in vascular tissue without compromising its proper functioning. Recently, genes have been identified, which are expressed specifically at distinct phases of vascular tissue differentiation or in a subset of vascular tissue cell types (Hertzberg et al., 2001; Kuriyama and Fukuda, 2001; Zhao et al., 2000). The subsequent functional analysis of such genes and mutants displaying an altered vascular tissue formation will lead to a better understanding of the molecular mechanisms underlying vascular tissue development.

Mutant analysis and transgene studies showed that genes belonging to the HD-ZIP family of transcription factors are important for vascular tissue formation. The expression of the Arabidopsis ATHB-8 gene (Baima et al., 1995) was shown to be associated with provascular/procambial tissue. Loss-of-function mutants for the Arabidopsis IFL1 gene lacked the formation of xylem fibres and vessel elements in inflorescence stems (Zhong and He, 1999), while the overexpression of ATHB-8 and IFL1 (Baima et al., 2001; Zhong and He, 2001) caused an increased xylem and interfascicular fibre formation. Interestingly, the activation tagging of the patatin-like gene STURDY resulted in an increased xylem cell number and interfascicular fibre formation (Huang et al., 2001) similar to the phenotype conferred by overexpression of ATHB-8 and IFL1. The increase in xylem cell number caused by overexpression of ATHB8, ILF1 and STURDY took place without a significant effect on phloem cell number and therefore does not occur at the expense of the cell number in another vascular cell type like observed for several BR mutants (Choe et al., 1999; Szekeres et al., 1996). This suggests that the ATHB8, ILF1 and STURDY genes act on the regulation of cell division activity in vascular tissue development. Increased expression of another type of HD-ZIP gene, ATHB-2, resulted in a decreased secondary vascular tissue growth, whereas reduced expression levels promoted secondary growth (Steindler et al., 1999). Therefore, ATHB-2 also acts on the regulation of cell division activity in vascular tissue development, but in contrast to ATHB8, ILF1 and STURDY as a negative regulator. The rice OSHOX1 gene (Scarpella et al., 2000) is expressed specifically in provascular/procambial tissue like the Arabidopsis ATHB-8 gene (Baima et al., 1995). However, overexpression of OSHOX1 led to premature vascular differentiation without an alteration in vascular cell number (Scarpella et al., 2000). This indicates that not all the HD-ZIP genes involved in vascular tissue formation act on the control of cell division activity.

Previously, we reported that activation tagging of the AP2/EREBP-like transcription factor LEAFY PETIOLE (LEP) causes the leafy petiole phenotype displayed by the dominant lettuce (let) mutant (van der Graaff et al., 2000). Here, we report that the size of the vascular bundles is increased in all aerial organs in let, as a result of an increase in the number of xylem, phloems (pro)cambial and pericycle cells. This vascular phenotype is caused by activation tagging of the two genes VASCULAR TISSUE SIZE (VAS) and LEP. These genes are closely linked and arranged in tandem. The activation tagging of VAS and LEP independently affects vascular tissue development, with VAS specifically affecting phloem (pro)cambial and pericycle cell number and LEP specifically affecting that of xylem.

Results

The lettuce mutant exhibits an altered vascular tissue formation

The dominant lettuce (let) mutant was initially isolated because of its conspicuous leafy petiole phenotype (van der Graaff et al., 2000). Histological analysis of let now revealed that the mutant displays an increased size of the vascular bundles in the leaf midrib, young hypocotyls, bracts, inflorescence stem, cotyledons and floral organs due to an increased vascular cell number (Figure 1a–n; Table 1; data not shown). The leaf venation pattern is unaffected in the mutant leaves (data not shown). The increased cell number is specific for the vasculature and the cell number in other tissues is not altered. However, due to the increased vascular bundle size, the cell shape and size in the cell layers surrounding the vasculature can be affected in let. This is most prominent for the parenchyma cells surrounding the main vascular bundle in the mutant leaves (Figure 1c–f). The vascular phenotype of let is specific for aerial organs, since vascular tissue formation in roots is similar to wild type (Figure 1o,p), even for much older roots exhibiting secondary thickening (data not shown). We conclude from these results that the increase in vascular tissue size in let is caused by an increase in xylem, phloem (pro)cambial and pericycle cell number.

Figure 1.

Vascular tissue size is affected in aerial organs of let.

Transverse sections of different organs of wild type (a,c,e,g,i,k,m,o) and let (b,d,f,h,j,l,n,p).

(a,b) Leaf 6 taken from 21-day-old plants was sectioned at 25% (c,d) and 75% (e,f) of the leaf length measured from the leaf base.

(g,h) One-week-old hypocotyls.

(i,j) Bracts.

(k,i) First node of inflorescence stem 2 weeks after the onset of bolting.

(m,n) Cotyledons.

(o,p) Top 5 mm from 1-week-old roots.

(q) Leaf 6 from 35SDE-LEP-VAS transgenic (21-day-old) exhibiting a strong leafy petiole phenotype, compare panel (f) for position of section and vascular tissue size in let. Scale bars: 50 µ m.

Table 1.  Vascular cell number is increased in the leaves of let and 35SDE-LEP-VAS transgenics
LineLeaf blade widtha
nbXylemcPhloemcnXylemPhloem
  • a Petioles and leaf blades were analysed for the largest leaves (leaf 5 and 6) from 21-day-old soil-grown plants. Petioles were sectioned at 25% and leaf blades at 75% (see Figure 1a,b for position).

  • b

    Number of plants analysed.

  • c

    Number of cells counted over the width of the main vascular bundle at the height of the xylem and phloem.

  • d

    Standard deviation is given in parenthesis.

  • e

    Strength of the leafy petiole phenotype displayed by the 35SDE-LEP-VAS transgenics.

C24149.0 (1.1)d20.6 (3.2)86.0 (0.7)12.9 (1.7)
lettuce525.0 (1.4)55.6 (5.7)1015.7 (2.9)36.6 (7.0)
Weake413.0 (0.7)31.3 (1.9)411.0 (1.2)24.3 (3.0)
Intermediatee416.5 (1.1)37.3 (4.4)411.8 (0.8)26.3 (1.9)
Intermediate/stronge420.5 (2.2)48.3 (5.3)416.3 (1.1)34.5 (3.6)
Stronge224.0 (1.0)52.5 (1.5)417.0 (1.6)38.5 (2.3)

The transformation of the 35SDE-LEP-VAS construct, which harbours the activation-tagged LEAFY PETIOLE (LEP) and VASCULAR TISSUE SIZE (VAS) genes (Figure 2a), phenocopied let (van der Graaff et al., 2000), including the mutant leaf vascular tissue phenotype (Figure 1f,q; Table 1). The size of the main vascular bundles in leaves was analysed for four independent 35SDE-LEP-VAS transgenics exhibiting the leafy petiole phenotype at different strengths. This showed a positive correlation between the strength of the leafy petiole phenotype, increased LEP expression levels (van der Graaff et al., 2000), increased vascular cell number (Table 1) and increased VAS expression levels (Figure 3a).

Figure 2.

Schematic representation of the plant DNA flanking the RB of the activator T-DNA insert in let, the T-DNA constructs derived from this flanking plant DNA and the T-DNA insertion sites in the vas and lep knock-out lines.

(a) Part of the plant DNA directly flanking the RB of the T-DNA insert is shown in detail. On this fragment, two genes are located: LEP (41057–41689, At5g13910) and VAS (42266–42794, At5g13900, GenBank AF463514). Numbers indicate the location on the BAC clone MAC12 and asterisks, the T-DNA insertion sites in the lep and vas T-DNA insertional mutants. The small arrows indicate the position of the primers used in this study. 1 and 2, RTL2F and RTL2R; 3 and 4, VAS5′-race and VAS3′-race; 5 and 6, LEP5′-race and LEP3′-race; 7 and 8, WKO-LEP and WKO-VAS. Open boxes: LEP, VAS and GUS coding region. The boxed 3′ indicates part of the LEP 3′ coding region and its 3′ UTR. Black box: CaMV 35S minimal promoter (−90 to 0) with AMV leader sequence. Hatched box: 35S-enhancer region (−393 to −95). Dotted box: 5′ UTR and 3′ UTR regions of LEP and VAS, determined by RACE-PCR. Arrows indicate the different transcripts generated by the T-DNA constructs.

(b,c) T-DNA insertion site in the vas and lep knock-out lines. On the top row, the sequences flanking the T-DNA insertion and on the bottom row, the wild-type sequences are depicted. In bold, the T-DNA and in italic, the filler DNA sequences are shown. Wild-type sequences deleted in the T-DNA insertional mutants are underlined. In brackets, the position relative to the protein translation start site of the wild-type 5′ and 3′ nucleotides flanking the T-DNA insert are given. (b) vas knock-out line in which 3 bp of wild-type sequences were deleted upon T-DNA insertion and 21 bp of filler DNA was inserted. The T-DNA insert is composed of a LB-RB:RB-LB inverted repeat. Dashes were introduced to align the sequences from wild-type and the knock-out line. (c) lep knock-out line in which 52-bp wild-type sequence (425–476) was deleted upon T-DNA insertion.

Figure 3.

VAS expression in wild-type, let and the different transgenic lines used in this study.

Northern blots were probed with the VAS coding sequence amplified from cDNA using the RTL2 primers (upper panels), stripped and probed with the GapC coding region amplified from cDNA using the GapC primers for loading control (lower panels). (a,b,d,e) 20 µg and (c) 15 µg total RNA.

(a) VAS expression in 35SDE-LEP-VAS transgenics (harbouring the activation-tagged LEP and VAS genes), exhibiting either a wild-type leaf development or a weak, intermediate and strong leafy petiole phenotype, respectively.

(b) VAS expression in independent 35SDE-VAS transgenic lines (harbouring the ectopically expressed VAS gene), exhibiting a wild-type vascular development, weak increase (two independent lines) or strong increase (three independent lines) in vascular tissue size.

(c) VAS expression in wild type, let and the 3 DE-VAS transgenics (lines 22-3, 36-4 and 52-4, harbouring the activation-tagged VAS gene) used for the developmental analysis.

(d) VAS expression in wild type at different developmental stages.RNA was isolated from 1-week-old complete seedlings, the aerial part from 1-week-old and 2-week-old seedlings and roots from 2-week-old seedlings. The latter were grown in liquid culture.

(e) VAS expression in different organs from wild type and let. RNA was isolated from cotyledons, expanded leaves, inflorescence stems or flowers from soil-grown plants.

Activation tagging of VAS influences phloem (pro)cambial and pericycle cell number

The possible role of the VAS gene in vascular tissue formation was studied in a vas knock-out line and by the transformation of two different VAS overexpressor constructs to wild type. In the Wisconsin knock-out facility (Sussman et al., 2000), a vas insertional mutant was isolated harbouring an inverted T-DNA repeat insert (LB-RB:RB-LB) 268 bp downstream of the protein translation start site (Figure 2a,b). RT-PCR analysis confirmed complete loss-of-function for VAS, while LEP expression was unaffected. The development of the homozygous vas knock-out line was indistinguishable from Wassilewskija wild type. Furthermore, histological analysis showed similar cell numbers in the vascular tissue of roots and hypocotyls of 7- and 16-day-old plants compared to wild type (data not shown).

The transformation of the VAS overexpressor constructs resulted in a specific effect on vascular cell number, while other aspects of plant development detectable by morphological and histological analysis were not affected (data not shown). For eight out of 19 35SDE-VAS transgenics, an increased vascular cell number was detected. In four lines, only a weak increase in vascular cell number was observed in the hypocotyls, while the remaining four lines showed a strong increase in vascular cell number for both hypocotyls and petioles (data not shown). Northern blot analysis showed a correlation between increased VAS expression levels and increased vascular cell number in the T2 generation of 35SDE-VAS transgenics (Figure 3b). However, the progeny from these transgenics display a wild-type vascular development despite the presence of high VAS expression levels (data not shown). Therefore, the vascular phenotype conferred by ectopic VAS overexpression from the 35SDE-VAS transgene is unstable, probably because of gene silencing specifically in the vasculature.

The activation-tagged DE-VAS construct (Figure 2a) conferred an increased VAS expression level for 37 out of the 46 independent DE-VAS lines analysed. The analysis of vascular tissue development in hypocotyls for 13 DE-VAS lines displaying different strengths of VAS expression revealed a correlation between increased VAS expression and increased vascular cell number. In contrast to the 35SDE-VAS transgenics, both the increased VAS expression and increased vascular cell number proved to be stable in the T3 generation of these activation-tagged DE-VAS transgenics (data not shown).

Vascular tissue development was studied in more detail in wild-type, let and three independent homozygous T3 DE-VAS lines (22-3, 36-4 and 52-4) that display a clear vascular phenotype but different strengths of increased VAS expression (Figures 3c and 4). The cell number in the vascular tissue of hypocotyls was comparable for the three independent DE-VAS lines at all stages analysed, and therefore the data from these lines was combined. The DE-VAS overexpressors showed a significant increase in cell number for the outer cell layers of hypocotyls (phloem (pro)cambium and pericycle) compared to wild type (Figure 4b,c,f). Subsequent analysis of vascular tissue in roots, leaf petioles and inflorescence stems showed that the DE-VAS transgenics displayed an increased phloem (pro)cambial and pericycle cell number in aerial organs (data not shown). The increase in cell number for the outer cell layers of hypocotyls is similar for the DE-VAS overexpressors and let (Figure 4b). However, in contrast to let, the DE-VAS overexpressors did not exhibit a significantly increased xylem cell number in either hypocotyls (Figure 4a) or other aerial organs (data not shown).

Figure 4.

Cell number in the vascular tissue of hypocotyls from wild type, let, DE-VAS and 35SDE-LEP transgenics.

The cell number was counted in vascular tissue from hypocotyls over the axis from phloem to phloem pole (length direction) for 7-, 10-, 12-, 14- and 21-day-old plants (this path is indicated in c–f by arrows). Similar results were obtained when cells were counted in the width direction of the vascular tissue in hypocotyls (data not shown). The significance for the difference in the number of vascular cells in let, 35SDE-LEP and DE-VAS transgenics compared to wild type was tested by one-sided Student's t-test with: *P < 0.05, **P < 0.01 and ***P < 0.001. n = Number of plants analysed per developmental stage.

(a) Xylem cell number.

(b) Outer cell number (phloem (pro)cambium, and pericycle).

(c–f) Representative vascular tissue sections from 21-day-old (c) wild-type (x: 17, o: 21), (d) let (x: 31, o: 29), (e) 35SDE-LEP (x: 22, o: 26) and (f) DE-VAS (x: 25, o: 33) plants, respectively. Between brackets, the cell count is indicated for xylem (x) and the outer cell layers (o), counted along the path indicated by the arrows. The two inner arrows indicate the border between xylem and (pro)cambium based on difference in cell (wall) morphology. The upper and lower arrows indicate the perimeter of the vascular tissue. Scale bars: 50 µm.

VAS shares weak similarity with plant non-specific lipid transfer proteins

The VAS gene (At5g13900; GenBank AF463514) encodes a small putative protein of 151 amino acids (Figure 5) with a molecular weight of 16.5 kDa and a pI of 8.8. Analysis of the VAS protein sequence using the SMART analysis tool (Schultz et al., 1998, 2000) revealed that VAS contains an N-terminal 22-amino acid signal peptide for the secretory pathway, a putative 22-amino acid transmembrane region at the C-terminus and an AAI domain (amino acids 29–109). This AAI domain is found in several protein families, including plant (non-specific) lipid transfer proteins (nsLTPs), (2S) seed storage proteins and the family of trypsin/alpha-amylase inhibitors. According to Southern blot analysis and database searches (data not shown), VAS is a single-copy gene, which shares 100% sequence identity with three expressed sequence tags (EST: AI99385, T46734 and N37786) and a weak similarity with nsLTPs, based on the conserved distribution of eight cysteines characteristic for the AAI domain. The position of the putative transcription start and polyadenylation site of VAS was determined using RACE-PCR (Figure 2a). This showed that VAS contains a small intron of 75 bp, a 38-bp 5′ UTR and a 165-bp 3′ UTR, resulting in a transcript of 659 nucleotides.

Figure 5.

DNA and protein sequence of VAS.

Part of the LEP 3′ UTR (until 41880) and the 5′ UTR (42237–42265) and 3′ UTR (42796–42961) of VAS are shown in bold. The predicted amino acid sequence encoded by VAS is shown below the nucleotide sequence with the eight conserved cysteine residues of the AAI domain in bold and the stop codon indicated by an asterisk. The intron present in the VAS gene is indicated in lower case. The N-terminal signal peptide and C-terminal transmembrane region are underlined. Numbers indicate the location on the BAC clone MAC12.

VAS is expressed in vascular tissue

VAS expression in wild-type plants was detected in roots and young seedlings using Northern blot analysis (Figure 3d,e). In addition, RT-PCR analysis was performed to study organ specificity of VAS expression. This showed that VAS is expressed in all wild-type organs (data not shown). VAS expression is significantly increased in aerial organs of let (Figure 3e).

Attempts to analyse VAS expression using in situ RNA hybridisation proved to be without success, even for wild-type roots in which VAS expression is highest. Therefore, uidA fusion constructs were used to examine tissue specificity conferred by the VAS promoter. Transformation of the pVAS-GUS construct (Figure 2a) showed that the VAS promoter confers expression to the central stele of roots and the lower part of the hypocotyl (Figure 6 a,b). Histological analysis showed that GUS activity is detectable in endodermal cells and very weak in vascular tissue, while GUS staining is absent from cortex and epidermis (data not shown).

Figure 6.

Expression pattern of VAS and LEP.

(a–i) uidA analysis of VAS promoter activity with (a,b) pVAS-GUS and (c–i) DEpVAS-GUS transgenics.

(j–p) RNA in situ hybridisation of wild-type LEP expression. (a) 3-day-old seedling, (b) Detail from 5-day-old root showing GUS activity specific for the central stele, (c) 7-day-old seedling, (d) Detail from 5-day-old seedling showing GUS activity in emerging leaf, (e,f) Young leaves, (g) cotyledon from 10-day-old seedling, (h) flower, (i) bract, (j–l) Longitudinal sections through hypocotyl from 7-day-old seedling hybridised with LEP sense probe (j) and antisense probe (k,l), respectively, (m,n) transverse sections through hypocotyl from 10-day-old seedling hybridised with LEP sense (m) and antisense probe (n), respectively and (o,p) transverse sections through the leaf blade from a young leaf hybridised with LEP sense (o) and antisense probes(p), respectively. Arrows indicate in situ signals in (n) and (p). Scale bars: (a–i) 1 mm, except (b,d) 0.2 mm and (j–o) 50 µm.

In order to analyse the altered VAS expression caused by activation tagging of the VAS gene, the doubled enhancer region (DE) of the CaMV 35S promoter used as activator tag was cloned in front of the VAS promoter region in the pVAS-GUS construct. Thus, this promoter driving uidA expression in the resulting DEpVAS-GUS construct (Figure 2a) is identical to the one used for overexpression of VAS in the DE-VAS construct that conferred a stable increased vascular cell number. Analysis of more than 100 primary transformants and the progeny of 54 transgenics showed that transformation of the DEpVAS-GUS construct resulted in two classes of transgenics. The minority of the transgenics display weak, patchy expression in vascular tissue and cells throughout all plant organs (data not shown). In the majority of transgenics, this activation-tagged uidA fusion construct confers expression specifically to the vascular tissue of all organs analysed (Figure 6c–i). Histological analysis of such plants showed that GUS staining is restricted to vascular tissue including the endodermis (data not shown). These results show that the activation-tagged VAS promoter activity as conferred by this DEpVAS-GUS construct correlates with the vascular phenotype displayed by the DE-VAS transgenics.

Activation tagging of LEP influences xylem cell number

In let, the cell number for xylem, phloem (pro)cambium, and pericycle is increased, whereas activation tagging of VAS (DE-VAS transgenics) results in a specific increase in the cell number for the outer cell layers only (Figure 4). Therefore, vascular tissue development was studied in a lep insertional mutant and LEP overexpressors (35SDE-LEP transgenics) to understand this difference in vascular phenotype. In the Wisconsin knock-out facility (Sussman et al., 2000), a lep insertional mutant was isolated harbouring a T-DNA insert 424 bp downstream of the protein translation start site (Figure 2a,c). RT-PCR analysis confirmed complete loss-of-function for LEP, while VAS expression was unaffected. The homozygous lep knock-out line exhibited a wild-type development. Histological analysis showed similar cell numbers in the vascular tissue of roots and hypocotyls of 7- and 16-day-old plants compared to Wassilewskija wild-type plants (data not shown).

The vascular cell number was analysed for seven independent 35SDE-LEP transgenics displaying different strengths of increased LEP expression and the leafy petiole phenotype. This analysis showed that those lines exhibiting both a clear leafy petiole phenotype and an increased LEP expression also display an increased vascular cell number. Three of such 35SDE-LEP transgenics were analysed in more detail. Despite clear differences in the levels of LEP expression between them, the number of cells in the vascular tissue was comparable for all stages analysed (data not shown) and the data from the three lines was combined. The 35SDE-LEP transgenics displayed a specific increase in xylem cell number in hypocotyls (Figure 4a,c,e) and other aerial organs (data not shown). This increase in xylem cell number for hypocotyls is similar for the 35SDE-LEP overexpressors and let (Figure 4a). However, in contrast to the DE-VAS transgenics and let, the cell number in the outer cell layers is not significantly affected (Figure 4b).

LEP is expressed in xylem

Analysis of LEP expression with in situ RNA hybridisation using Paraplast embedded tissue detected wild-type LEP expression in vascular tissue of young leaves, petioles and hypocotyls (Figure 6) in addition to previously reported expression in leaf primordia and developing leaf blades (van der Graaff et al., 2000). Using sense probes, no significant signals were obtained (Figure 6j,m,o). The staining in the vascular tissue was always observed for cells associated with tracheary elements and could not be observed in cells located outside the xylem and therefore appears to be xylem specific (Figure 6k,l,n,p). In let and 35SDE-LEP transgenics, a similar xylem-specific expression pattern was observed (data not shown), indicating that the activation tagging of LEP caused cell type-specific upregulated LEP expression.

Discussion

Activation tagging has been successfully employed to isolate novel genes involved in plant development (reviewed in Weigel et al., 2000), even genes for which loss-of-function lines did not show an altered development (Ito and Meyerowitz, 2000; Li et al., 2001). This lack of loss-of-function phenotypes for insertional mutants of the cytochrome P450 CYP78A9 and the serine carboxypeptidase BRS1 is apparently caused by functional redundancy, since several highly similar genes for CYP78A9 and BRS1 are present in the genome of Arabidopsis (Ito and Meyerowitz, 2000; Li et al., 2001). Here, we report that in addition to the previously reported leafy petiole phenotype (van der Graaff et al., 2000), the activation-tagged let mutant exhibits an increased vascular cell number in all aerial tissues. This increased vascular cell number was caused by the activation tagging of the tandemly arranged LEP and VAS genes. Initial analysis of 35SDE-VAS transgenics indicated that VAS overexpression conferred the vascular phenotype displayed by let; hence, this gene was named VASCULAR TISSUE SIZE (VAS). However, more detailed analysis of 35SDE-VAS, DE-VAS and 35SDE-LEP transgenics showed that VAS and LEP independently affect vascular tissue development, with VAS specifically affecting cell number in phloem (pro)cambium, and pericycle and LEP specifically affecting that of xylem.

VAS affects the cell number in phloem (pro)cambium and pericycle

Using uidA fusion constructs, we showed that the VAS promoter confers expression associated with vascular tissue in roots and the lower part of the hypocotyl region, corroborating Northern blot analysis. The VAS protein contains an N-terminal signal peptide for the secretory pathway and an AAI domain. This AAI domain is present in the families of plant non-specific lipid transfer proteins (nsLTPs) (Clark and Bohnert, 1999; Gausing, 1994; Vignols et al., 1997), (2S) seed storage proteins (Chatthai and Misra, 1998; Scarafoni et al., 2001) and trypsin/alpha amylase inhibitors (Gomez-Leyva and Blanco-Labra, 2001; Grosset et al., 1997; Huber et al., 1996). In addition, the AAI domain is present in a group of proteins that supposedly anchor the plasma membrane to the cell wall (Neuteboom et al., 1999). This AAI domain is characterised by the conserved spacing of cysteine residues, responsible for the formation of intramolecular disulphide bridges important for secondary protein structure and function (Baud et al., 1993; Oda et al., 1997; Shin et al., 1995). Both on DNA and amino acid level, VAS shares weak similarity to nsLTPs only. In contrast to nsLTPs, VAS contains a C-terminus that is 40 amino acids longer and in which the last 22 amino acids putatively encode a transmembrane domain. This indicates that VAS could be anchored in the plasma membrane after secretion. Plant nsLTPs use a broad range of lipophilic compounds as substrate (Hollenbach et al., 1997), contain an N-terminal signal peptide, are transported across membranes in vitro and are secreted in vivo (Gausing, 1994). The nsLTP TED4 was shown to function as a protease inhibitor protecting living cells during maturation of tracheary elements, indicating an additional role for nsLTPs as specific protease inhibitors (Endo et al., 2001 and references therein). Recently, in a microarray study, the expression of the poplar orthologue of VAS was specifically increased during the expansion and maturation phase of xylem formation (Hertzberg et al., 2001). This temporal expression pattern would be in agreement with a role as protease inhibitor. However, the activation tagging of VAS did not significantly affect xylem formation and, in contrast, increased phloem (pro)cambial and pericycle cell number. Therefore, it is more likely that VAS functions in the transfer of a lipophilic compound to the cell exterior, which subsequently could act as a signal molecule influencing vascular tissue growth.

LEP affects xylem cell number

Previously, we showed that LEP functions in leaf blade formation and hypothesised that LEP might act as a general cell division-promoting factor, which could be specified through interaction with other tissue-specific factors (van der Graaff et al., 2000). Detailed analysis of vascular tissue development in 35SDE-LEP transgenics showed that the activation tagging of LEP furthermore causes a specific increase in xylem cell number. This role is corroborated by in situ RNA hybridisation analysis, which showed that LEP is expressed in cells associated with tracheary elements. The xylem-specific expression and the preferential increase in the number of xylem cells in vascular tissue indicates that LEP functions in xylem formation in addition to its role in leaf blade formation. In both processes, LEP could act as a positive factor in cell division activity.

The role of LEP and VAS in vascular tissue formation

In let, the complex situation occurred where the insertion of the activator T-DNA element caused an altered expression of two closely linked genes, LEP and VAS. Furthermore, the separate activation tagging of LEP and VAS independently affects a different subset of cells in vascular tissue. These activation-tagged phenotypes indicate that LEP and VAS have reciprocal functions in the development of the vascular tissue. This might explain why LEP and VAS are positioned adjacent to each other. Their close proximity could allow co-ordinated regulation of LEP and VAS expression.

Neither lep nor vas knock-out plants display a vascular phenotype distinct from wild type. Given the fact that LEP is a member of a large gene family, it can be expected that genetic redundancy or degeneracy caused by overlapping expression patterns and function from AP2/EREBP family members can compensate for the loss of LEP activity. However, VAS is a unique gene and therefore a loss-of-function phenotype could be expected. The lack of an altered phenotype for vas knock-out plants indicates that VAS is not essential for plant development or that parallel pathways regulating vascular cell number exist. Given the importance of vascular tissue for plant development, the latter is not unlikely. Because of the lack of loss-of-function phenotypes for LEP and VAS, the possible role of LEP and VAS in plant development can only be deduced from their wild-type expression pattern and overexpression analysis.

The increased vascular cell number caused by the activation tagging of LEP and VAS did not occur at the expense of the cell number in other vascular cell types, like observed for several BR mutants (Choe et al., 1999; Szekeres et al., 1996) in which increased phloem cell number was accompanied by a decrease in that of xylem. Therefore, the LEP and VAS activation-tagged phenotypes are most likely caused by a stimulation of the formation of new cells by cell division rather than a stimulation of vascular differentiation or a shift in the balance of the number of (pro)cambial cells available for phloem and xylem differentiation.

The HD-ZIP transcription factor ATHB8 is the earliest available marker for provascular/procambial cells (Baima et al., 1995) and its expression can be observed in the stele bordering the initial cells of the root apical meristem. The expression of VAS is absent from the root tips of primary and lateral roots (Figure 6a,b), suggesting that VAS is expressed at a later vascular developmental stage than ATHB8. LEP expression is only observed in xylem parenchyma cells and appears to be absent from (pro)cambial cells. This suggests that both VAS and LEP do not directly influence the number of procambial cells. The cell number for all vascular cell types correlated with the expression levels of the HD-ZIP transcription factor ATHB-2 (Steindler et al., 1999) and therefore ATHB-2 appears to regulate overall (pro)cambial activity. In contrast, only specific vascular cell types are affected by the activation tagging of LEP or VAS. Thus, it is most likely that LEP and VAS either asymmetrically stimulate (pro)cambial activity or stimulate cell division activity of the derivatives from the (pro)cambium. The pericycle is ontogenetically part of the vascular cylinder (Scheres et al., 1995) and, consequently, the pericycle cell number would also be increased by the activation tagging of VAS, since it lies on the phloem side of the (pro)cambium. Because of the difficulty to morphologically distinguish the (pro)cambial cells from their parenchymatic derivatives, it is not certain whether the activation tagging of VAS actually influences the number of initial cells in the (pro)cambium.

The xylem phenotype resulting from activation tagging of LEP resembles the overexpression of ATHB8, ILF1 and STURDY, giving rise to an increased xylem cell number (Baima et al., 2001; Huang et al., 2001; Zhong and Ye, 2001) without a significant effect on that of phloem. However, in contrast to the increased interfascicular fibre formation displayed by the overexpression of ATHB8, ILF1 and STURDY, no obvious effect on interfascicular fibre formation was observed for 35DE-LEP transgenics and let. This suggests that LEP functions in a pathway that is different from ATHB8, ILF1 and STURDY or that LEP functions downstream of ATHB8 and IFL1 and regulates a subset of the pathway(s) regulated by ATHB8 and IFL1.

Since LEP displays tissue and cell type-specific enhanced expression due to the activation tagging and the expression of LEP correlates with the overexpression phenotype, a conclusive deduction of its role in plant development could be made from the LEP activation-tagged phenotype. Unfortunately, the exact cell specificity of both wild-type VAS and its activation-tagged expression could not be determined because of too low expression levels. Therefore, the VAS activation-tagged phenotype might be neomorphic. However, the VAS transcript is present in wild-type aerial tissues and the activation-tagged DEpVAS-GUS construct conferred vascular specific activity in aerial organs, comparable to its activity in wild-type roots. This suggests that VAS is also involved in vascular tissue formation in aerial organs. More detailed analysis using VAS protein localisation studies will be required to prove its proposed role in vascular tissue development.

Experimental procedures

Plant materials and growth conditions

Seeds of the transgenic lines and wild type (all Arabidopsis thaliana ecotype C24, except for the Wisconsin knock-out lines that were in the Wassilewskija background) were grown as described before (van der Graaff et al., 2000) unless otherwise stated. The 35SDE-LEP-VAS, 35SDE-LEP and 35SDE-VAS constructs (Figure 2a) were transformed to Arabidopsis using a root transformation protocol (Vergunst et al., 1998) and the DE-VAS, pVAS-GUS and DEpVAS-GUS constructs (Figure 2a) were transformed using the floral dip method (Clough and Bent, 1988). The constructs 35SDE-LEP-VAS and 35SDE-LEP were described before (van der Graaff et al., 2000). The 35SDE-VAS construct was made by cloning the CaMV 35SDE activator tag in the AccI restriction site at position 42174 (numbers refer to the position on the BAC clone MAC12), 63 bp upstream of the VAS coding region. DE-VAS was obtained by cloning the CaMV 35S promoter doubled enhancer region (DE) of the activator tag upstream of the genomic region from the EcoRI restriction site at position 41564 to the HindIII site at position 43200. Fusing the region from 41564 (EcoRI) to 42249 (fragment generated using the Promega Erase-a-Base kit) transcriptionally to the uidA reporter gene of pGPTV-BAR (Becker et al., 1992) resulted in the pVAS-GUS construct. DEpVAS-GUS was made by fusing the doubled enhancer (DE) in front of the VAS promoter region of the pVAS-GUS construct. In the DE-VAS, pVAS-GUS and DEpVAS-GUS constructs, the VAS promoter region was extended with part of the 3′ region of the LEP gene located directly upstream of this VAS promoter to ensure that the full-length VAS promoter was present.

Oligonucleotide primers for PCR

The following oligonucleotide primers (Figure 2a) were used: RTL2F (42337), 5′-GAT AAA TGG GCA GAG CGT-3′; RTL2R (42733), 5′-CAA TAG GAC TGA GAA AAG GT-3′; VAS5′-race (42565), 5′-ATG TTC TCC ACA TCG AGC AGG CAA C-3′; VAS3′-race (42328), 5′-GGG ATG CGG ATA AAT GGG CAG AGC-3′; LEP5′-race (41607), 5′-GAC GAG TAG TCG TCA CCG GTC CAG-3′; LEP3′-race (41294), 5′-TGC CTC CTT CCT CAT CCG TCA CC-3′; WKO-LEP (41003), 5′-TCG GAC ATT TAT TGA TCT GTG TAT GCA TA-3′; WKO-VAS (42827), 5′-CAA AGC TGC TAT AGA CAA TGA GAG ATT CA-3′; GapCF, 5′-AGC TCG TCG CTG TCA ACG-3′; GapCR, 5′-GAC AGC CTT GGC AGC TCC T-3′. The numbers between brackets indicate the position of the primers on the BAC clone MAC12.

Analysis of gene expression

Total RNA isolation, Northern blotting, RT-PCR and in situ RNA hybridisation was performed as described before (van der Graaff et al., 2000). Expression of VAS was detected using the RTL2 primers amplifying a 322-bp fragment from cDNA and a 396-bp fragment from genomic DNA. RT-PCR was performed on cDNA samples prepared with an oligo dT15 primer for 20, 25, 30 and 35 cycles. After gel electrophoresis, VAS expression was detected using Southern blotting for the samples amplified at 20 or 25 cycles.

Isolation and molecular characterisation of T-DNA insertional mutants for LEP and VAS

The primers WKO-LEP and WKO-VAS were used to screen for T-DNA insertional mutants of LEP and VAS in the Wisconsin knock-out facility (Sussman et al., 2000). Siblings of the lep and vas knock-out lines were genotyped by PCR using either the LEP-race primers or the VAS-race primers in combination with the T-DNA border primers of the Wisconsin knock-out facility. The same primers were used to isolate the T-DNA border/plant DNA junctions of the T-DNA insertions in the lep and vas lines. The resulting PCR fragments were sequenced to determine exact T-DNA integration sites in lep and vas. The expression of LEP and VAS was analysed using cDNA prepared from homozygous lep and vas lines using the RTL1 (van der Graaff et al., 2000) and the RTL2 primers, respectively.

RACE-PCR

The SMART RACE cDNA amplification kit (Clontech) was used for 5′ and 3′ RACE-PCR with the VAS5′-race and VAS3′-race primers for analysis of the VAS transcription start and polyadenylation site and confirmation of the intron size and position, respectively. The resulting PCR products were cloned in the pGEM-T-easy vector (Promega, Switzerland) and at least 10 independent clones were sequenced.

Vascular tissue analysis

The plants used for the analysis of root and hypocotyl vascular tissue were grown in tissue culture under continuous light. For initial analysis of vascular cell number, the seeds from hemizygous 35SDE-VAS and DE-VAS transgenics were grown on antibiotic selective medium to identify the transgenic siblings. For the detailed analyses of vascular tissue (Figure 4), seeds from the homozygous DE-VAS transgenics were grown without antibiotic selection. The seeds from the hemizygous lettuce mutant and 35SDE-LEP transgenics were grown without antibiotic selection and transgenics siblings were identified because of their leafy petiole phenotype. Plant material was fixed in 2% glutaraldehyde in 0.1 m sodium cacodylate buffer (pH 7.2) for 8 h at room temperature, dehydrated through an ethanol series and embedded in epon. Sections (1 µm) were stained with toluidine blue and mounted in epon. The number of cells was counted in transverse sections as follows. For hypocotyls, the number of cells was counted over the axis from phloem to phloem pole (length direction) and over the perpendicular axis (width direction). Phloem, (pro)cambial, and pericycle cells were classified as outer cell layers. In petioles and leaves, the total width of the main vascular bundle was represented by the number of cells counted at the height of the xylem (dorsal region) and those counted at the height of the phloem (ventral region). Significance for difference in the number of cells compared to wild type was tested by one-sided Student's t-test.

GUS staining

Plants were stained for 8 or 24 h in GUS staining buffer (Jefferson et al., 1987) and either cleared in 70% ethanol or fixed in 2% glutaraldehyde in 0.1 m sodium cacodylate buffer (pH 7.2) for 8 h at room temperature. The fixed plants were dehydrated through an ethanol series and embedded in LR white for histological analysis of GUS activity (2–3-µm sections).

Acknowledgements

We would like to thank Gerda Lamers, Amke Den Dulk-Ras and Tonny Regensburg-Tuink for assistance with the anatomical analysis, Thorsten Schnurbusch for valuable assistance on statistical analysis, Jean-Jacques Pittet for assistance with the digital processing of the figures and Dr Christoph Ringli and Dr Bert van der Zaal for critical reading of the manuscript. E.v.d.G was supported by grants from the Leiden University (Stichting BVS) and the Swiss National Science Foundation (31-51055.97).

VAS GenBank accession number: AF463514.

Ancillary