The expression of tobacco knotted1-type class 1 homeobox genes correspond to regions predicted by the cytohistological zonation model

Authors


*For correspondence (fax +81 52 789 5226;
e-mail j45751a@nucc.cc.nagoya-u.ac.jp).

Summary

We have isolated and characterized four tobacco homeobox genes, NTH1, NTH9, NTH20, NTH22 (Nicotiana tabacum homeobox) which belong to the class 1 knotted1-type family of homeobox genes. Comparison of the inferred amino acid sequences of the ELK homeodomains of these genes and previously reported kn1-type class 1 proteins has revealed that the four new tobacco genes belong to distinct subclasses, suggesting that each NTH gene may have distinct functions. Using in situ hybridization and by analysing the distribution of GUS activity in tobacco plants transformed with NTH promoter::GUS constructs, local- ized expression of the three NTH genes was observed in the shoot apical meristem (SAM). In the vegetative SAM, NTH1 and NTH15 showed overlapping expression in the corpus, NTH20 was expressed in the peripheral zone, and NTH9 was predominantly expressed in the rib zone. The expression patterns of the different NTH genes correspond to regions predicted by the cytohistological zonation model, suggesting that each NTH gene specifies the function of the SAM zone with which it is associated.

Introduction

All of the above ground parts of plants develop from an indeterminate group of cells, the shoot apical meristem (SAM). The SAM is formed during embryogenesis, and after seed germination it continuously generates various organs and tissues, such as leaves, stems and flowers, throughout the life of the plant. The SAM is not only the ultimate source of new tissues and organs of the above ground part of plants, but also maintains and replenishes itself (Steeves & Sussex 1989).

Previous morphological studies indicated that various angiosperms share some structural features in the internal organization of the SAM (Steeves & Sussex 1989). The SAM can be divided into three histologically distinct zones, the central zone, the peripheral zone and the rib zone, based on the cell size and the arrangement of the cell divisions (Esau 1977; Lyndon 1990; Steeves & Sussex 1989). The central zone, occupying the centre of the SAM, consists of relatively large cells that result from slower cell divisions in this region, and is thought to act as a source of new undifferentiated cells to replace those that are gradually lost from other zones of the SAM. The peripheral zone is located in the flanking regions of the SAM and consists of smaller, more rapidly dividing cells. Lateral organs such as leaf primordia and axillary buds are generated within this zone. Finally, the rib zone, which is located at the base of the other two zones, gives rise to the pith of the stem.

SAM is also organized into a few fundamental cell layers (Schmidt 1924). The tunica consists of one to three cell layers and covers the underlying corpus. This tunica/corpus structure is a consequence of differences in the orientation of cell division in different parts of the SAM. Cell divisions within the tunica are ordinarily restricted to the anticlinal plane, thus maintaining a layered arrangement of the cells. Within the corpus, the plane of cell division is more variable. Despite extensive knowledge of the complex structural organization of the SAM, very little is known about the molecular mechanisms underlying meristem function in such processes as lateral organ formation.

Many homeobox genes have now been cloned from various plant species. Comparative analysis of the homeodomains and other conserved motifs of the plant homeobox genes has revealed that the genes fall into several classes (Kerstetter et al. 1994; Lincoln et al. 1994; Vollbrecht et al. 1991). These classes include the kn1-type (HD-kn1), the homeodomain leucine zipper (HD-ZIP: Ruberti et al. 1991; Schena & Davis 1992), the plant homeodomain finger (PHD-finger: Bellmann & Werr 1992; Korfhage et al. 1994; Schindler et al. 1993), the GLABRA2-like homeodomain (HD-GL2: Lu et al. 1996; Rerie et al. 1994), and the BELL1-like homeodomain proteins (HD-BEL1: Quaedvlieg et al. 1995; Reiser et al. 1995). The most extensively characterized gene family is the kn1-type, all members of which share homeodomain sequences similar to that of the maize knotted1 gene (kn1), the first homeobox gene to be identified in plants (Vollbrecht et al. 1991). Based on sequence comparisons, the members of this family fall into two subclasses, class 1 and 2 (Kerstetter et al. 1994). The class 1 genes show higher similarity to kn1 in the homeodomain and are primarily expressed in the SAM and developing stem, whereas the class 2 genes have lower similarity to kn1 in the homeodomain and are expressed in most organs (Kerstetter et al. 1994). Class 1 genes, such as maize kn1 (Sinha et al. 1993; Vollbrecht et al. 1991), tobacco NTH15 (Tamaoki et al. 1997a), and Arabidopsis KNAT1 (Lincoln et al. 1994), show strong expression within the SAM. Ectopic expression of class 1 genes in spontaneous mutants, or in transgenic plants carrying copies of the genes expressed from the cauliflower mosaic virus 35S promoter, causes dramatic alterations in leaf and flower morphology. Furthermore, mutant plants with recessive alleles of the Arabidopsis SHOOT MERISTEMLESS (STM) gene, another member of the class 1 gene family, fail to develop a SAM during embryogenesis (Long et al. 1996). These findings strongly suggest that the class 1 kn1-type genes are responsible for meristem maintenance and/or lateral organ formation from the SAM (Hake et al. 1995).

During organ differentiation in animal embryos, families of homeobox genes with similar structures are often involved in a series of related developmental processes (Gehring 1987; Ingham 1988; McGinnis & Krumlauf 1992). In plants, by analogy, a battery of class 1 kn1-type homeobox genes may cooperate in meristem maintenance and/or lateral organ formation from the SAM. Indeed, it has been reported that the maize class 1 kn1-type homeobox genes comprise a small multi-gene family and that these genes are highly expressed in the shoot apex (Jackson et al. 1994; Kerstetter et al. 1994). With this in mind, we attempted to isolate as many different class 1 homeobox genes from tobacco as possible, and then used in situ hybridization and GUS histochemical staining of transgenic tobacco to compare their expression patterns within the SAM.

Results

Isolation of kn1-type class 1 homeobox genes from tobacco

Previous studies have demonstrated that the kn1-type class 1 homeobox genes encode a well-conserved homeodomain. To isolate new members of the class 1 gene family from tobacco, we designed degenerate oligonucleotide primers corresponding to the conserved amino acid sequences of the homeodomain and performed RT–PCR with total or poly(A)+ RNA from various tobacco organs. Several fragments containing the homeobox sequence were amplified from RNA isolated from shoot buds, flower buds, flowers, roots, stems and young leaves. Sequence analysis revealed that we had isolated nine distinct DNA fragments encoding homeobox sequences, including one corresponding to the previously reported NTH15 (Tamaoki et al. 1997a). From these we chose four fragments for further study because their homeodomain sequences fell into different subgroups when compared with previously reported homeodomains (see below). The four fragments were used as probes to screen for full-length cDNA clones from two tobacco cDNA libraries produced from shoot apices and stems. One clone, designated NTH1 (Nicotiana tabacum homeobox 1), was isolated from the shoot apex cDNA library, while three other cDNA clones, NTH9, NTH20 and NTH22, were isolated from the stem cDNA library.

The four new homeobox genes were completely sequenced. NTH1, NTH9, NTH20 and NTH22 contained open reading frames encoding 326, 323, 357 and 320 amino acids, respectively (Fig. 1). All four predicted proteins shared the three conserved domains: the 64 amino acid homeodomain, the ELK domain located immediately upstream of the homeodomain (Fig. 2a), and the KNOX domain, a recently defined consensus domain located further upstream of the ELK homeodomain. The sequences of all four NTH proteins were well conserved throughout the homeodomain. Most importantly, the amino acid sequence of the third helix, which is thought to be responsible for the recognition of target DNA sequences, was completely conserved in all four NTH proteins. The four invariant amino acids (WPYP), located between helix 1 and helix 2 of all kn1-type homeodomains, were also completely conserved in the NTH proteins.

Figure 1.

Alignment of five NTH homeodomain proteins from tobacco.

The homeodomain is indicated by a bold underline and the ELK region is represented with a bold dashed underline. The thinner underline indicates the KNOX domain.

Figure 2.

Comparison of the ELK homeodomains between the NTH proteins and other KN1-type class 1 homeodomain proteins.

(a) Alignment of the ELK homeodomains of the NTH proteins and other class 1 proteins. Residues conserved among all of the sequences are shaded. The ELK domain is boxed.

(b) Phylogenetic tree of the class 1 plant homeodomains based on the alignment in (a). The tree was calculated using the UPGMA program (Unweighted Pair Group Method with Arithmetic mean).

To examine the structural relationship between the NTH genes and previously reported kn1-type class 1 genes, we analysed the ELK homeodomains using the UPGMA program (Unweighted Pair Group Method with Arithmetic mean, Fig. 2b). The four NTH genes fell into different subgroups in the resulting phylogenetic tree. For example, NTH15 was identified as a member of the group containing Arabidopsis STM (Long et al. 1996), tomato TKn2 (Harven et al. 1996) and soybean SBH1 (Ma et al. 1994), whereas NTH20 was grouped with other homeobox genes such as apple KNAP1 and KNAP2 (Watillon et al. 1997), tomato TKn1 (Harven et al. 1996) and Arabidopsis KNAT1 (Lincoln et al. 1994). NTH22 was most similar to POTH1 from potato and was less similar to Arabidopsis KNAT2 (Lincoln et al. 1994). NTH1 and NTH9 did not cluster with any other genes.

Expression patterns of the NTH genes

To characterize the expression patterns of the NTH genes, RNA gel-blot analysis was performed (Fig. 3). The four NTH genes were expressed at different levels in various organs. The NTH1 transcript was most abundant in shoot buds, but was also detected at a lower level in flower buds, mature flowers, roots and stems. NTH9 was mainly expressed in shoot buds and flower buds, and was much less abundant in other organs. NTH20 was mainly expressed in stems, with lower expression in shoot buds and roots. Similarly, the highest expression of NTH22 was observed in stems, with lower expression in flower buds and flowers, and very low levels in other tissues. These results show that each of the four NTH genes has a unique expression pattern.

Figure 3.

RNA gel-blot analysis of the NTH genes in different organs of tobacco.

RNA was isolated from shoot buds (SB), flower buds (FB), flowers (FL), roots (RT), vegetative stems (ST), unexpanded leaves (YL) and expanded leaves (OL). Ten micrograms of total RNA were loaded per lane. The blots were probed with gene-specific probes which excluded the region encoding the homeodomain to avoid cross-hybridization. To show the relative quantity of RNA in each lane, the ubiquitin probe was used.

All four NTH genes were expressed in both vegetative and reproductive meristems, even though these meristems give rise to different types of lateral organs. To investigate the detailed expression pattern of the NTH genes within both meristems, we performed in situ hybridization using antisense probes specific for each NTH gene. Expression of NTH1, NTH9 and NTH20 was observed in median longitudinal sections of the vegetative SAM of mature plants (Fig. 4), whereas no clear signal was observed in sections probed for NTH22 (data not shown). Control sense-strand probes did not give clear hybridization signals, except for some faint signals in immature vascular strands of stems and leaves. Similar hybridization was also observed with all of the antisense probes, and was therefore interpreted as non-specific hybridization (data not shown). NTH1 mRNA, indicated by purple staining, was observed in the corpus region of the SAM, whereas little or no expression was observed in the outer, tunica cell layers (Fig. 4b). NTH1 expression appeared to be down-regulated at the position of the leaf primordia. We confirmed the pattern of NTH1 expression within the shoot apical region by GUS staining of tobacco plants transformed with a chimeric construct containing the NTH1 promoter fused to the E. coli uidA gene (see Experimental procedures). GUS activity, indicated by blue staining, was specifically localized in the corpus region, with no expression in the position of leaf primordium initiation (arrowhead in Fig. 4c). This result confirms that NTH1 is preferentially expressed in the corpus region but is down-regulated in the region of leaf primordium initiation. This localization of NTH1 expression is essentially the same as that described for NTH15 mRNA (Fig. 4a and Tamaoki et al. 1997a).

Figure 4.

In situ localization of NTH mRNAs and histochemical analysis of GUS activity in the vegetative meristem of transgenic tobacco plants carrying NTH GUS constructs.

Median longitudinal sections through vegetative shoot apices of mature tobacco plants (approximately 60 days old) were used for the in situ hybridization and GUS staining analyses. (a), (b), (d) and (f): in situ hybridization with antisense RNA probes against NTH15, NTH1, NTH9, and NTH20, respectively. (c), (e) and (g): GUS staining of transgenic tobacco carrying NTH1::GUS, NTH9::GUS and NTH20::GUS, respectively. Note that NTH1 and NTH15 are mainly expressed in the corpus (a, b, c), NTH9 is expressed in the rib zone (d,e), and NTH20 is expressed in the peripheral zone (f, g). Arrowheads indicate the leaf initiation site. lp, leaf primordium; rm, rib meristem; cz, central zone; pz, peripheral zone. All bars indicate 50 μm.

Expression of NTH9 was localized in the lower parts of the SAM, known as the rib zone, whereas little or no expression was observed in the upper region of the SAM or in the procambium. Occasionally, faint staining was seen in young leaves and leaf primordia (Fig. 4d). We have interpreted this staining as non-specific since similar staining was also observed with the sense-strand control probe. Indeed, GUS staining of tobacco plants transformed with an NTH9 promoter::GUS chimeric gene was concentrated in the rib zone (Fig. 4e).

In the case of NTH20, expression was localized to the peripheral zone of the SAM, while fainter staining was seen in the central zone and leaf primordia (Fig. 4f). The suppression of NTH20 in the central zone and leaf primordia was confirmed by the serial longitudial and cross-sections. A similar expression pattern was also seen in tobacco plants transformed with an NTH20 promoter::GUS chimeric gene (Fig. 4g). GUS staining clearly shows down-regulation of NTH20 expression in the central zone, along with strong expression in axillary buds that was not seen with in situ hybridization (Fig. 4f). Such expression in the peripheral zone was also seen in the case of the Arabidopsis KNAT1 gene (Lincoln et al. 1994). The similar expression patterns of NTH20 and KNAT1 are consistent with the high similarity of their primary structures (Fig. 2).

As reported previously, NTH15 was also strongly expressed in both early and late floral meristems (Fig. 5a and Tamaoki et al. 1997a). The expression patterns of the four new NTH genes in the reproductive SAM were further examined by in situ hybridization. In the inflorescence meristems, strong NTH1 expression was observed in the meristem centre and in the region where the floral meristems would emerge (Fig. 5b, arrowhead). NTH9 mRNA was also observed in a similar region to NTH1 (Fig. 5c). In contrast, expression of NTH20 and NTH22 was not detected in the inflorescence meristem, but was seen in the basal regions of the meristem, in a zone corresponding to the border regions of the secondary inflorescence meristems. No expression of the NTH genes was observed in floral meristems (data not shown).

Figure 5.

In situ localization of NTH mRNAs in the infloresence meristem.

Longitudinal sections through inflorescence meristems were used for in situ localization of NTH mRNAs. Expression of NTH15 (a), NTH1 (b) and NTH9 (c) is evident in the centre of inflorescence meristem and the floral meristem initiation sites (indicated with arrowheads). NTH20 (d) and NTH22 (e) are not expressed in the meristem, but rather in regions basal and peripheral to the meristem (indicated with arrowheads). No staining could be detected in the meristem region hybridized with a sense NTH20 probe (f). All bars indicate 50 μm.

Discussion

The kn1-type class 1 family comprises several subgroups

A comparative study of the kn1-type homeobox genes by Kerstetter et al. (1994) suggested that they can be divided into two classes according to differences in their homeodomains. The ELK homeodomains of the class 1 genes show higher similarity to that of kn1 than do those of the class 2 genes. In this study, we have focused on the class 1 genes since in a variety of plants, these genes are mainly expressed in meristem-rich tissues, such as the vegetative and reproductive meristems. For this reason, the class 1 genes are believed to be involved in meristem development and/or maintenance.

We aligned the ELK homeodomain sequences of the NTH genes and previously reported class 1 genes and performed a phylogenetic analysis based on the alignment (Fig. 2). The sequence similarity within this domain is quite high among both monocot and dicot plants. Indeed, the 15 amino acids comprising the third helix are absolutely conserved in the products of all of the genes. The third helix of the homeodomain has been proposed to be responsible for DNA sequence recognition when homeodomain proteins interact with target DNA sequences (Kornberg 1993). Therefore, the conserved sequence of the third helix of the class 1 homeodomain proteins may indicate that these proteins interact with the same or very similar target DNA sequences. Four amino acids between helix 1 and 2 (WPYP) were also totally conserved in all the class 1 homeodomain proteins. Recently, Bürglin (1997) has reported that three invariant residues between helix 1 and 2 (PYP) are also observed in several homeodomain proteins from other organisms, although not in typical homeodomain proteins such as Antennapedia (Gehring et al. 1994). Based on this unique feature, Bertolino et al. (1995) named this homeobox group the TALE (Three Amino acid Loop Extension) family.

Phylogenetic analysis clearly demonstrates that the class 1 genes can be further divided into small subgroups and that the four NTH genes fall into different subgroups (Fig. 2). For example, NTH20 is clustered with a group comprising KNAP1, KNAP2, TKn1 and KNAT1. NTH15 groups with STM, TKn2 and SBH, and NTH22 is paired with POTH1 to make a small group. Both NTH1 and NTH9 comprise separate groups that do not contain any other genes. While NTH1 and NTH9 are unique in tobacco (Fig. 2b), undiscovered homologues may exist in other plants.

It is noteworthy that the class 1 subgroups that include the NTH genes contain only dicot genes. The monocot genes are clustered into separate groups such as the kn1 group (kn1, OSH1 and Hvknox3), the rs1 group (rs1, knox4 and OSH15), and the lg3 group (lg3, knox5 and knox11). Therefore, the clustering of class 1 genes reflects the distance between monocot and dicot plants. Among the dicot genes, the NTH20 group shows the closest relationship to the monocot kn1 group. Based on this observation, it may be possible that the NTH20 and kn1 groups are orthologous and that the homeobox genes they contain may have homologous functions in dicot and monocot plants. In contrast, we could not find appropriate counterparts of two other well-defined clusters of class 1 genes, STM (dicot) and rs1 (monocot), in monocots and dicots, respectively. We were unsuccessful in an attempt to directly isolate a rice homologue of STM, using PCR with primers based on the consensus sequence of genes in the STM group (Sentoku et al. unpublished data). Furthermore, intensive work by Kerstetter et al. (1994) to isolate maize kn1-related genes also failed to reveal any homologues of the dicot STM group. While these negative results are inconclusive, we infer that monocots may not have a counterpart to the dicot STM group, and that this group of homeodomain proteins may be specially evolved for dicot developmental systems. Alternatively, other as yet unidentified class 1 genes that correspond to the STM and rs1 groups may exist in monocots and dicots, respectively. Further studies should clarify the relationships among the class 1 homeobox genes of monocots and dicots.

Expression of the NTH genes occurs in SAM regions defined by the zonation model

Previous morphological studies, based on cytological features and cell division patterns, have suggested that the SAM is organized into several different regions, such as tunica and corpus, or central, peripheral and rib zones (Steeves & Sussex 1989). We have shown here that the expression of three of the NTH genes reveals a similar zonation of the SAM during vegetative development (summarized in Fig. 6). The expression of NTH1 was confined to the corpus region, excluding the tunica layers, and this pattern was essentially the same as that seen for another NTH gene, NTH15 (Fig. 4a and Tamaoki et al. 1997a). In contrast, NTH9 and NTH20 were expressed in the rib and peripheral zones, respectively.

Figure 6.

Schematic representation of the expression of the NTH genes in the vegetative meristem.

The expression of NTH1 and NTH15 is localized to the corpus as indicated by vertical stripes. NTH9 and NTH20 are predominantly expressed in the rib zone (horizontal stripes) and peripheral zone (diagonal stripes), respectively.

According to the tunica/corpus concept, the corpus is an internal group of cells with an irregular cell division pattern and is thought to provide cells for the interior portion of organs and stems (Schmidt 1924). Furthermore, according to the zonal feature, the peripheral zone consists of small cells with a higher frequency of cell division and the rib zone consists of longitudinally arranged files of cells (Steeves & Sussex 1989). The observations that lateral organs are formed in the peripheral zone and the pith of the stem originates in the rib zone indicate that these zones represent regions in which undifferentiated cells derived from the central zone develop toward a more specified state in which they can be incorporated into organs (Esau 1977; Lyndon 1990; Steeves & Sussex 1989). Based on these observations, we believe that the SAM is organized into specialized regions, each with a distinct function, such as maintainance of undifferentiated cells, determination of cell fate, or formation of organs or tissues.

In animals, it has been shown that multiple homeobox genes are expressed in a characteristic pattern to determine cell fate and to form specific organs or tissues during the process of embryogenesis (Gehring 1987; Ingham 1988; McGinnis & Krumlauf 1992). In Drosophila, for example, segment-specific expression of particular homeobox genes, known as homeotic selector genes, controls a series of other genes to bring about segment-specific organ formation. In other words, the identity of each segment of the Drosophila embryo is specified by regionalized expression of the homeobox genes. By analogy to the mechanism of specification in Drosophila, NTH1 and NTH15 may help specify corpus identity and NTH9 and NTH20 may act as zonal regulators as described above, specifying the identity of the rib and peripheral zones, respectively.

A study supporting such functional regionalization of the meristem at the molecular level has been reported by Fleming et al. (1993), who analysed the expression patterns of several meristem-specific genes within the SAM. For example, a gene encoding arginine decarboxylase (ADC) was predominantly expressed in the corpus, whereas ribosomal protein L38 genes (rpl38) showed specific expression in the peripheral zone. These observations indicate that regionalization of the SAM occurs not only at the histological level, but also at the functional level. Because homeobox genes are believed to encode transcriptional regulators, genes such as those studied by Fleming et al. are good candidates for downstream genes whose expression is regulated by the products of homeobox genes. By regulating the expression of downstream genes, the localized expression of homeobox genes may confer region-specific functions.

In contrast to the variety of expression patterns seen in the vegetative shoot meristem, NTH expression in the inflorescence meristem displayed only two different patterns. Three NTH genes, NTH1, NTH9 and NTH15, were mainly expressed in the centre of the inflorescence meristem, whereas NTH20 and NTH22 expression was not detected in the reproductive SAM itself, but only in the border region between two inflorescence meristems. Such a simplified expression pattern in the inflorescence meristem is also observed for knox homeobox genes in maize (Jackson et al. 1994). Four maize homeobox genes, kn1, rs1, knox3 and knox8, are expressed uniformly throughout the corpus of the inflorescence meristem, even though these genes show distinct expression patterns in the vegetative meristem. Thus, in the vegetative meristem, these homeobox genes are probably involved in different events with distinct biological roles as mentioned above, but in the inflorescence meristem they may cooperate in common developmental event(s). In other words, these homeobox genes may have different biological function(s) in different meristems. Disappearance of NTH20 expression in the inflorescence meristem and localization of its expression to the peripheral region of the vegetative meristem resembles the expression pattern of Arabidopsis KNAT1 in vegetative and reproductive shoots (Lincoln et al. 1994). The similarity of their expression patterns suggests that these two genes may have similar function(s) in Arabidopsis and tobacco, and indeed they also fall into the same small sequence homology group. Down-regulation of NTH20 in the inflorescence meristem indicates that the gene does not function in reproductive meristems, whereas the gene is actively expressed and may actively function in the vegetative meristem. The change in expression of NTH20 and KNAT1 may be one of the essential events for the change from vegetative to reproductive stages.

Experimental procedures

Plant growth conditions

Tobacco seeds (Nicotiana tabacum cv. Samsun NN) were sterilized in 5% sodium hypochlorite for 5 min and germinated on germination medium (Murashige and Skoog salts with 1% sucrose and 0.5% gelangum) under continuous light at 25°C. The seedlings were transplanted to soil and grown at 25°C in a 16 h light/8 h dark cycle.

Isolation of NTH genes

To isolate novel knotted-type class 1 genes from tobacco, total RNA was extracted from shoot buds, flower buds, flowers, roots, stems, young leaves and expanded leaves. RT–PCR was performed with these RNAs using one of two oligonucleotides, GCGGATCCAA (A,G) CTTCC (A,G,C,T) AA (A,G) GA (A,G) GC or GCGGATCCAA (A,G) CTTCC (A,G,C,T) AA (A,G) GA (C,T) GC, as the 5′ primer, and AAGCTTTG (A,G) TT (T,A,G) AT (A,G) AACCA (A,G) TT (A,G) TT as the 3′ primer. These primers were designed based on the conserved region of the kn1-type class 1 homeodomain. The amplified fragments were sequenced to confirm that they encoded sequences similar to the class 1 kn1-type homeodomain region. By this method, we isolated nine different partial fragments encoding homeodomains. These fragments were used as probes for isolation of full-length cDNA clones.

For construction of cDNA libraries, total RNA was extracted from shoot apices or stems of mature tobacco plants. Poly(A)+ RNA was isolated by passage through an oligo(dT)-cellulose column (type III, Becton Dickinson Labware). Double-stranded cDNA was synthesized from each poly(A)+ RNA preparation and cloned into the EcoRI site of λgt11 (Stratagene). Screening was performed in 50% formamide, 6× SSC, 5× Denhardt's solution, 0.5% SDS, and 0.1 mg ml–1 salmon sperm DNA at 42°C for 14 h.

Nucleotide sequences were determined by the dideoxynucleotide chain-termination method using an automated sequencing system (ABI 373A). The cDNA clones were completely sequenced in both strands. The cDNA and inferred amino acid sequences were analysed using GENETYX computer software (Software Kaihatsu Co., Japan).

RNA gel-blot analysis

Total RNA was isolated from a variety of tobacco tissues including shoot buds, flower buds, mature flowers, roots, stems, young leaves (approximately 2–3 cm long) and fully expanded leaves. Ten micrograms of each RNA preparation were resolved by electrophoresis, transferred to Hybond N membrane (Amersham), and probed using the 5′ regions of the various cDNAs. The probes excluded the ELK homeodomain to avoid cross-hybridization due to the high conservation of this region. Hybridization was performed at 65°C in a solution containing 10% dextran sulphate, 6× SSC, 5× Denhardt's solution, 0.5% SDS and 0.1 mg ml–1 salmon sperm DNA. Filters were washed with 2× SSC, 0.1% SDS at room temperature and then further washed in 0.2× SSC, 0.1% SDS at 65°C.

In situ hybridization

In situ hybridization was performed using digoxigenin-labelled sense or antisense RNA produced from the NTH cDNA coding regions. The probes excluded the highly conserved ELK homeodomain and the poly(A) region. Hybridization and immunological detection of the hybridized probes were performed according to the method of Kouchi & Hata (1993).

Screening of genomic libraries

Nuclear genomic DNA was isolated from leaves of 60-day-old tobacco plants. DNA purified by the CTAB method was partially digested with Sau3AI and subjected to sucrose density gradient centrifugation to enrich for fragments of approximately 20 kbp. The fragments were cloned into the BamHI site of λEMBL3 (Stratagene). The following were used as probes to screen for the promoter regions of the various NTH genes: a 320 bp BglII fragment of NTH1, a 410 bp HindIII fragment of NTH9, a 570 bp EcoRI fragment of NTH20 and a 590 bp StuI fragment of NTH22. Screening was performed under the same conditions as used for the cDNA screening.

Construction of NTH promoter: GUS chimeric genes and histochemical analysis of GUS activity

The 5′ sequences of the various NTH genes, containing the 5′ flanking plus 5′ non-coding regions were cloned into pBluescriptII. These promoter sequences were amplified by PCR using an M13 primer that annealed within the vector just outside the multiple cloning site, and 3′ primers specific for each of the NTH genes that anneal just upstream of the translation initiation site. Suitable linker sequences were included at the 5′ end of the 3′ primers to facilitate subsequent cloning: NTH1, BamHI linker; NTH9;SalI linker; NTH20;BamHI linker; NTH22;SalI linker. The PCR products were cloned into pCRII (Invitrogen) and sequenced to confirm that no nucleotide substitutions had occurred during PCR. The NTH1 promoter sequence (approximately 3.0 kb) was inserted into HindIII–BamHI-cut pBI101 (Clontech Laboratories, Inc.) which carries the structural gene for β-glucuronidase (GUS) and the terminator sequence of the nopaline synthase (NOS) gene. Similarly, the NTH9 (3.0 kb) and NTH20 (4.0 kb) promoters were cloned into HindIII–SalI-and HindIII–BamHI-cut pBI101, respectively. The NTH22 promoter sequence (2.8 kb) in pCRII was re-cloned into the EcoRI site of pBluescriptII, then introduced into SalI–BamHI-cut pBI101. Construction of NTH15::GUS has already been described in a previous paper (Tamaoki et al. 1997b).

Histochemical assays for GUS activity were performed essentially as described by Matsuoka & Sanada (1991). Seedlings and shoot apices were embedded in 5% agar, and the agar blocks sectioned (50–70 μm thick) with a microslicer. Sections were placed in 50 mM sodium phosphate buffer, pH 7.0, 20% methanol, and 1 mM 5-bromo-4-chloro-3-indolyl-β-d-glucuronide, then vacuum-infiltrated for 20 min and incubated at 37°C until blue colour appeared. An ethanol wash was performed to stop the reaction and to remove chlorophyll.

Transformation and regeneration of tobacco

The NTH promoter::GUS constructs were introduced into Agrobacterium tumefaciens LBA4404 by electroporation. Agrobacterium-mediated transformation of Nicotiana tabacum, cv Samsun NN was performed with leaf discs as previously reported (Matsuoka & Sanada 1991). Transgenic plants were selected on medium containing 100 mg l–1 kanamycin.

Acknowledgements

This research was supported by a Grant-in-Aid for Scientific Research on Priority Areas (The Molecular Bases of Flexible Organ Plans in Plants) from the Ministry of Education, Science and Culture (Japan) and Special Coordinating Funds for Promoting Science and Technology from the Science and Technology Agency (Japan) to M.M., and by Research Fellowships of the Japan Society for the Promotion of Science for Young Scientists to A.N.

Footnotes

  1. DDBJ/EMBL/GenBank nucleotide sequence database accession numbers AB025573 (NTH1), AB025713 (NTH9), AB004785 (NTH15), AB025714 (NTH20) and AB025715 (NTH22).

Ancillary