Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
2049 Development 124, 2049-2062 (1997) Printed in Great Britain © The Company of Biologists Limited 1997 DEV0091 A leucine-rich repeat containing receptor-like kinase marks somatic plant cells competent to form embryos Ed D. L. Schmidt, Flavia Guzzo, Marcel A. J. Toonen and Sacco C. de Vries* Department of Molecular Biology, Wageningen Agricultural University, Dreijenlaan 3, 6703 HA Wageningen, the Netherlands *Author for correspondence (e-mail: [email protected]) SUMMARY The first somatic single cells of carrot hypocotyl explants having the competence to form embryos in the presence of 2,4-dichlorophenoxyacetic acid (2,4-D) were identified using semi-automatic cell tracking. These competent cells are present as a small subpopulation of enlarged and vacuolated cells derived from cytoplasm-rich and rapidly proliferating non-embryogenic cells that originate from the provascular elements of the hypocotyl. A search for marker genes to monitor the transition of somatic into competent and embryogenic cells in established suspension cell cultures resulted in the identification of a gene transiently expressed in a small subpopulation of the same enlarged single cells that are formed during the initiation of the embryogenic cultures from hypocotyl explants. The predicted amino acid sequence and in vitro kinase assays show that this gene encodes a leucine-rich repeat contain- ing receptor-like kinase protein, designated Somatic Embryogenesis Receptor-like Kinase (SERK). Somatic embryos formed from cells expressing a SERK promoterluciferase reporter gene. During somatic embryogenesis, SERK expression ceased after the globular stage. In plants, SERK mRNA could only be detected transiently in the zygotic embryo up to the early globular stage but not in unpollinated flowers nor in any other plant tissue. These results suggest that somatic cells competent to form embryos and early globular somatic embryos share a highly specific signal transduction chain with the zygotic embryo from shortly after fertilization to the early globular embryo. INTRODUCTION explants used, and it appears that there is quite some variability in the tissue that responds first. A further complicating factor in unraveling the early stages of somatic embryogenesis is the fact that there is almost no evidence to show that the cellular changes observed in particular cells are indeed directly responsible or even necessary for embryo formation. In Daucus carota (carrot), the formation of embryogenic cell cultures usually commences with the incubation of seed-derived seedling hypocotyl explants in auxin-containing medium (De Vries et al., 1988a). As in other species, following an increase in cytoplasmic content, cell division is resumed in provascular cells, but not in cortical or in epidermal cells (Guzzo et al., 1994). Continued cell divisions then result in the formation of masses of small isodiametric cells. These cells can then enlarge and finally detach into the culture medium to contribute to the developing embryogenic cell culture (Guzzo et al., 1994). Such an embryogenic culture contains morphologically and biochemically different cell types, that can be present as single cells or cell clusters (Van Engelen and De Vries, 1992), of which only 1-2% of the cells are actually embryogenic (De Vries et al., 1988a). The term ‘embryogenic’ is defined as the ability to form somatic embryos without further exogeneous application of growth regulators (De Jong et al., 1993). The origin of embryogenic cells, that are usually present as clusters of small cytoplasm-rich cells (Komamine et al., 1990) is not clear and is Somatic or asexual embryogenesis is the process whereby somatic cells develop into plants via characteristic morphological stages. The later stages, in particular, closely resemble zygotic embryo development, and in dicots pass through the globular, heart and torpedo stage. Somatic embryos have been obtained in many different plant species and from a wide variety of starting materials such as microspores, protoplasts, immature embryos, tissue explants and in vitro cultured cells (see for recent reviews Dudits et al., 1995; Mordhorst et al., 1997). The events that take place during the period in which plant cells undergo the transition from somatic to embryogenic cell are poorly understood (reviewed by De Jong et al., 1993). In tissue explants, the first response is often noted to be the rapid replacement of the vacuole with cytoplasm, followed by the first division. In Dactylis glomerata, leaf mesophyll cells respond this way (Trigiano et al., 1989), while in Sorghum bicolor also vascular tissue, close to the wound surface, responds (Wernicke and Brettel 1980). The same responses are observed in Cichorium (Dubois et al., 1991), but in contrast to these studies, in Medicago sativa (Dos Santos et al., 1983) and in Ranunculus sceleratus (Konar et al., 1972), particular types of epidermal cells were noted to be the responsive ones. In all of these examples, in vitro regenerated plantlets were the source of the Key words: Daucus carota L., somatic embryo, embryogenic cell, leucine-rich repeat receptor-like kinase, zygotic embryo, SERK 2050 E. D. L. Schmidt and others thought to involve an auxin dependent transition stage occurring in single cells. Cells in this transition between the somatic and embryogenic cell state are defined as competent cells (Komamine et al., 1990; Toonen et al., 1994), which is an operational definition based on the requirement for exogenous auxin. Recording of the developmental fate of many thousands of individual carrot cells from established embryogenic suspension cultures by cell tracking revealed that competent cells have a highly variable appearance that prevents their identification on the basis of morphology (Toonen et al., 1994). Using this system with cells from an activated carrot hypocotyl explant revealed that a small subset of a particular type of elongated single cells are the ones that first acquire the competence to form embryogenic cells. Apart from the ability to form embryogenic cells, these cells are indistinguishable from the majority of the elongated single cells. It is generally assumed that the formation of plant embryos requires the activation of specific sets of genes (reviewed by Goldberg et al., 1994; Thomas et al., 1993) and many studies have employed differential screening techniques to identify such genes (eg. Wilde et al., 1988; Aleith and Richter, 1990; Wurtele et al., 1993; Heck et al., 1995). The corresponding expression pattern of these genes during zygotic embryogenesis allowed classification into several groups (Goldberg et al., 1989; Sterk and de Vries, 1992). While many of the genes found to be expressed in early somatic embryos appeared to encode genes normally expressed late in zygotic embryogenesis or throughout plant development (reviewed by Zimmerman, 1993), others such as LTP (Sterk et al., 1991) and EMB-1 (Wurtele et al., 1993) have been shown to be expressed at the corresponding, globular, stage in zygotic embryogenesis. Embryo-expressed MADS box-containing regulatory proteins have been identified in both Brassica napus (Heck et al., 1995) and in Arabidopsis thaliana (Rounsley et al., 1995). The Arabidopsis gene, AGL15, is expressed as early as the 8-celled zygotic embryo, and is also present uniformly in torpedo stage embryos, in seedlings and to a lower level in leaves (Rounsley et al., 1995). Several genes have been reported (reviewed by Zimmerman, 1993) that are putative markers for embryogenic cell clusters, but none have been described to date that are reliable markers for the preceding stage of competent cells. There may be several reasons for this, such as the bias in many screening procedures towards more abundantly expressed genes, the low number of competent cells present in embryogenic cultures and the unavailability of rapid and reliable procedures to detect gene expression in single cells. To overcome these difficulties screens were carried out employing a series of carrot cell cultures with widely differing numbers of single competent cells as the starting material. A small number of genes were found for which expression is detectable in single competent cells in embryogenic cell cultures. One of these genes was investigated in more detail and found to encode a receptor-like kinase that appears to mark competent and embryogenic cells. MATERIALS AND METHODS Plant material, cell cultures, hypocotyl explant activation and cell tracking Seedlings of Daucus carota cv. Flakkese were grown for 2-3 weeks in vermiculite, while adult plants of this cultivar were obtained from S&G Seeds (Enkhuizen). Controlled pollination was performed by hand, and complete umbels removed at various days after pollination (DAP). Flower RNA was obtained from three complete umbels for each time point and contained all flower organs including residual pollen grains. Cell cultures were derived from Daucus carota cv. Flakkese and maintained as previously described (De Vries et al., 1988a). Cell suspension culture was carried out at high cell density in B5 medium (Gamborg et al., 1968) supplemented with 2 µM 2,4-D (2,4-dichlorophenoxy-acetic acid; B5-2 medium). Embryo cultures with globular, heart and torpedostage somatic embryos were derived from <30 µm sieved cell cultures cultured at low cell density (100,000 cells / ml) in B5 medium without 2,4-D (B5-0). For hypocotyl explant induction experiments, plantlets were obtained from seed of Daucus carota cv. S. Valery as described previously (Guzzo et al., 1994). The hypocotyls of 1-week old plantlets were divided in segments of 3-5 mm, incubated for various periods of time in B5-2 medium and returned to B5-0 medium. Seven days after explantation and exposure to 2,4-D, the hypocotyl segments were washed in B5-2 medium and subsequently fragmented on a 170 µm sieve and the resulting cells collected to form a fine cell suspension. Immobilization of these cells in B5-0.2 medium was performed in a thin layer of phytagel (Toonen et al., 1994). After 1 week of further culture, 2,4-D was removed by washing the plates with B5-0 medium. This allowed embryos to develop beyond the globular stage. Development of the immobilized cells was recorded using a modified procedure described by Toonen et al. (1994). The main change involved a new MicroScan program for automatic 3-axis movement to scan all cells in the phytagel (Toonen and de Vries, 1997). Nucleic acid isolation and analysis RNA was isolated from cultured cells and plant tissues as described by De Vries et al. (1988b). Poly(A)+ RNA was obtained by purification by oligo(dT) cellulose (Biolabs). For RNA gel blot analysis samples of 10 µg total RNA were electrophoresed on formamide gel, and transferred to Nytran-plus membranes (Schleicher & Schuell). For spot-dot northern analysis, 5 µg of total RNA was denatured and spotted onto nytran-plus filters using a hybridot manifold (BRL). Hybridization of RNA blots took place at 42°C in hybridization buffer containing 50% formamide, 6× SSC, 5× Denhardt, 0.5% SDS and 0.1 mg/ml salmon sperm DNA. Genomic DNA was isolated according to the method of Sterk et al. (1991). Samples of 10 µg genomic DNA were digested with different restriction enzymes and separated on agarose gel, and transferred to Nytran-plus membranes. Hybridization of DNA blots was performed as previously described (Sterk et al., 1991). Following hybridization, filters were washed under stringent conditions (3× 20 minutes in 0.1% SSC, 1% SDS, at 65°C). Filters were exposed to Kodak XOmat AR film. Nucleotide sequence analysis was performed on an ABI 373A automated DNA sequencer (Applied Biosystems). The sequences reported here have been deposited in the GenBank database, accession number U93048. SERK promoter-luciferase expression A 2200 bp HindIII/DraI genomic DNA fragment, with the DraI site 42 bp upstream of the predicted translational start codon of SERK, was cloned into the binary vector pMT500 (Toonen et al., 1996b) containing the firefly luciferase reporter gene (Millar et al., 1992). The resulting construct was transformed into carrot cells as described for the AtLTP1 promoter-luciferase constructs (Toonen et al., 1996b). A subpopulation of <50 µm diameter cells from a primary transformed suspension culture was embedded in phytagel with B5-0 medium at a concentration of 100,000 cells/ml. In single suspension cells, luciferase activity was not detectable with the CCD camera system described (Toonen and de Vries, 1997), even after a one hour exposure. It was only possible to measure luciferase activity in single cells by using a CCD camera (Photometrics) without a microscope, directly over the cells. Since luciferase is quickly deactivated in the presence of luciferin, measurements were made immediately after the addition of luciferin (Promega) to a final concentration of 20 µM in order to obtain the signal from the accumulated luciferase. The spatial resolution in the digitized images of the CCD camera is restricted to a pixel size of 30 by 30 µm. To dis- Somatic embryo receptor-like kinase 2051 tinguish luciferase activity from e.g. cosmic radiation, only signals observed in three consecutive 1-hour exposures were retained and positioned over the microscopic images of the same area of the culture dish. Development of the cells was recorded for 13 days as described by Toonen and de Vries (1997). A detailed analysis of the SERK promoter and of the luciferase-expressing cells will be presented elsewhere. Screening procedures Two independent cDNA libraries were constructed with equal amounts of poly(A)+ RNA from total established cell cultures grown for 6 days in B5-2 medium, sieved <125 µm cell cultures grown for 6 days in B50 medium, and sieved <30 µm cell cultures grown for 6 days in B5-0 medium. cDNA synthesis and cloning into the Uni-ZAPTM XR vector was performed according to the manufacturers protocol (Stratagene). Differential screening of the cDNA libraries was performed essentially as described by Scott et al. (1991). RNA was isolated from either three embryogenic or three non-embryogenic cell cultures, which were grown for 7 days in B5-2 after sieving through a 30 µm mesh. First strand cDNA synthesis was performed on 4 µg total RNA using AMV reverse transcriptase (RT; Gibco BRL). [32P]dATP labeled probes were prepared using random prime labeling on first strand cDNA. Pooled probes from high embryogenic or non-embryogenic cell populations were hybridized to two pairs of nitrocellulose filters, each containing 1000 plaques from one cDNA library. After washing for 3× 20 minutes in 0.1% SSC, 1% SDS at 65°C, hybridization was visualised by autoradiography for 2 days on Kodak XOmat film. Plaques that only showed signal with the embryogenic transcript probe were purified by two further rounds of screening. In order to identify cDNA clones which are expressed at low levels in the <30 µm sieved cell population, cold plaque screening was performed as described by Hodge et al. (1992). Plaques from the differential screening that did not show any signal after 7 days of autoradiography were purified by two further rounds of screening. The resulting clones were used as probes for characterization of the expression pattern of the corresponding genes. Differential display RT-PCR Differential display (dd) of mRNA was performed essentially as described by Liang and Pardee (1992). cDNA synthesis took place by annealing 1 µg of total RNA in 10 µl buffer containing 200 mM KCl, 10 mM Tris-HCl (pH 8.3), and 1 mM EDTA with 100 ng of one of the following anchor primers: (5′-TTTTTTTTTTTGC-3′), (5′TTTTTTTTTTTCTG-3′), (5′-TTTTTTTTTTTCA-3′). Annealing took place by heating the mix for 3 minutes at 83°C followed by incubation for 30 minutes at 42°C. Annealing was followed by the addition of 15 µl prewarmed cDNA buffer containing 16 mM MgCl2, 24 mM Tris-HCI (pH 8.3), 8 mM DTT, 0.4 mM dNTP, and 4 Units AMV reverse transcriptase (Gibco BRL). cDNA synthesis took place at 42°C for 90 minutes. First strand cDNA was phenol/chlorophorm extracted and precipitated with ethanol using glycogen as a carrier. The PCR reaction was performed in a reaction volume of 20 µl containing 10% of the synthesized cDNA, 100 ng of anchor primer, 20 ng of one of the following 10-mer primers: (5′-GGGATCTAAG-3′), (5′TCAGCACAGG-3′), (5′ GACATCGTCC-3′), (5′-CCCTACTGGT-3′), (5′-ACACGTGGTC-3′), (5′-GGTGACTGTC-3′), 2 µM dNTP, 0.5 Unit Taq enzyme in PCR buffer (10 mM Tris-HCl (pH 9.0), 1.5 mM MgCl2, 50 mM KCl, 0.01% gelatin and 0.1% Triton X-100) and 6 nM [α-32P]dATP (Amersham). PCR parameters were 94°C for 30 seconds, 40°C for 1 minute, and 72°C for 30 seconds for 40 cycles using a Cetus 9600 (Perkin-Elmer). Amplified and labeled cDNAs were separated on a 6% denaturing DNA sequencing gel. Gels were dried without fixation and bands were visualized by 16 hours of autoradiography using Kodak XOmat film. Bands containing differentially expressed cDNA fragments of 150-450 nucleotides were cut out of the gel and DNA was extracted from the gel slices by electroelution onto DE-81 paper (Whatmann). Reamplification of the cDNA fragments was performed using the same PCR cycling parameters as described above but with reaction buffer containing 2.5 µM of both the 10-mer and the anchor oligo and 100 µM dNTP. Reamplified PCR products were cloned into pBluescript vector II SK (Stratagene). RT-PCR 2 µg of total RNA from adult plant tissue or cell cultures was annealed at 42°C with 50 ng oligo (5′-TCTTGGACCAGATAATTC-3′) in 10 µl annealing buffer (250 mM KCl, 10 mM Tris-HCl pH 8.3, 1 mM EDTA). After 30 minutes annealing, 10 units AMV-reverse transcriptase were added in a volume of 15 µl cDNA buffer (24 mM TrisHCl pH 8.3, 16 mM MgCl2, 8 mM DTT, 0.4 mM dNTP). The reverse transcription reaction took place for 90 minutes at 42°C. PCR amplification of SERK cDNA was carried out with two specific oligos for the SERK kinase domain, (5′-CTCTGATGACTTTCCAGTC-3′) and (5′-AATGGCATTTGCATGG-3′). Amplification was carried out with 30 cycles of 30 seconds at 94°C, annealing at 54°C for 30 seconds and extension at 72°C for 1 minute, followed by a final extension for 10 minutes at 72°C. SERK PCR products were then separated by agarose gel electrophoresis, blotted and hybridized with a radiolabeled kinase domain of SERK cDNA. Following hybridization, filters were washed under stringent conditions (3× 20 minutes in 0.1% SSC, 1% SDS, at 65°C). The resulting hybridizing band of 680 bp was in agreement with the expected size of the PCR product. Independent RT-PCR experiments were performed twice with similar results. Hybridization with an 18S ribosomal probe of a northern blot loaded with similar amounts of RNA as used for the RT-PCR reactions confirmed the integrity and amounts of RNA used in the experiment. Whole-mount in situ hybridization Whole-mount in situ hybridizations were performed essentially as previously described (Engler et al., 1994). Cell cultures and somatic embryos were immobilized on poly-L-lysine coated slides during fixation to improve handling. Whole-mount in situ hybridization on explants took place by embedding hypocotyls from 7-day old plantlets in 3% Seaplaque agarose (Duchefa) and processing them in Eppendorf tubes. Transverse as well as longitudinal sections were made with a vibrotome (Biorad Microcut). Sections of 50-170 µm thick were incubated in B5-2 medium for a minimum of 3 days to induce formation of embryo-forming cells. Optimal induction was achieved with longitudinal hypocotyl sections with a thickness of at least 90 µm. To obtain proliferating, non-embryogenic cell cultures, hypocotyl sections were exposed to 2,4-D for only 1 day, and subsequently transferred to B5-0 medium (Guzzo et al., 1994). Whole-mount in situ hybridization on developing seeds was performed by removing the chalazal end of the seeds to allow easier probe penetration. After hybridization, the enveloping layers of integuments and endosperm were carefully removed to expose the developing embryos. All samples were fixed for 60 minutes in PBS containing 70 mM EGTA, 4% paraformaldehyde, 0.25% glutaraldehyde, 0.1% Tween 20, and 10% DMSO. Samples were then washed, treated with proteinase K for 10 minutes, again washed and fixed a second time. Hybridization solution consisted of PBS containing 0.1% Tween 20, 330 mM NaCl, 50 µg/ml heparin, and 50% deionized formamide. Hybridization took place for 16 hours at 42°C using digoxigenin-labeled sense or antisense riboprobes (Boehringer Mannheim). After washing, the cells were treated with RNase A, and incubated with anti-digoxigenin-alkaline phosphatase conjugate (Boehringer Mannheim) which had been preabsorbed with a plant protein extract. Excess antibody was removed by washing, followed by rinsing in staining buffer (100 mM Tris-HCl pH 9.5, 100 mM NaCl, 5 mM MgCl2, 1 mM levamisole) and the staining reaction was performed for 16 hours in a buffer containing 4-nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indoyl-phosphate. Observations were performed using a Nikon Optiphot microscope equipped with Nomarski optics or brightfield optics. Autophosphorylation assay A 1.4 kb SspI cDNA fragment of the SERK cDNA encoding most of 2052 E. D. L. Schmidt and others the open reading frame was cloned into the pGEX expression vector (Pharmacia). A fusion protein consisting of SERK and the glutathione S-transferase (GST) gene product was isolated and purified as described previously (Horn and Walker, 1994). Purified fusion protein was coupled to glutathione agarose beads (Sigma) and incubated for 20 minutes at 20°C in a volume of 10 µl buffer: 50 mM Hepes (pH 7.6), 10 mM MgCl2, 10 mM MnCl2, 1 mM DTT, 1 µCi [γ-32P] (3,000 Ci/mmol). Excess label was removed by washing the fusion protein/glutatione agarose beads three times for 5 minutes in 50 mM Tris-HCl (pH 7.3), 10 mM MgCl2 at 4°C. Protein was removed from the beads by cooking in SDS-PAGE loading buffer. Equal amounts of protein were separated by SDS-PAGE and protein autophosphorylation was visualized by autoradiography. RESULTS Isolation of cDNA clones that are preferentially expressed in embryogenic cell cultures of carrot In order to increase the chance of obtaining genes expressed in carrot suspension cells competent to form embryos, the number of such cells in a series of established cell cultures was determined. A subpopulation of cells that had passed through a 30 µm nylon sieve was isolated from eight different cultures ranging in age between 2 months and 4 years. In these <30 µm populations, the number of embryos formed from the single cells and small cell clusters was determined and expressed as a percentage of the total number of cells present at the start of embryogenesis. Three sieved <30 µm cultures able to form somatic embryos at a frequency of more than 100 embryos per 10,000 cells were then used as a source for competent cells, and three other cultures that produced less than 1 embryo per 10,000 cells were used as non-embryogenic controls. The assumptions made were that a 100-fold difference in the embryogenic capacity would indicate a similar difference in the number of competent cells, and that this would be sufficient to detect the mRNA of a hypothetical gene only expressed in competent single cells with the screening methods employed. As main selection strategies, cold plaque screening (Hodge et al., 1992) and differential display (dd) RT-PCR (Liang and Pardee, 1992) were used besides conventional differential screening of cDNA libraries. Labeled probes for differential screening were obtained from RNA out of a <30 µm sieved subpopulation of cells from either embryogenic or non-embryogenic cell cultures. Employing these probes in a library screen of approximately 2000 plaques yielded 26 plaques that failed to show any hybridization to either probe. These so-called cold plaques were purified and used for further analysis. From the total number of plaques that did hybridize, about 30 did so only with the probe from embryogenic cells. Differential display reverse transcriptionpolymerase chain reactions (ddRT-PCR) using a combination of one anchor primer and one decamer primer were performed on mRNA isolated from three embryogenic, and three nonembryogenic suspension cultures. About 50 ddRT-PCR fragments were obtained from each reaction. Using combinations of three different anchor and six different decamer primers, a total of approximately 1000 different cDNA fragments was visualized. Six of these PCR fragments were only found in lanes with mRNA from <30 µm populations of cells from embryogenic cultures and with oligo combinations of the anchor primer (5′-TTTTTTTTTTTGC-3′) and the decamer primers (5′-GGGATCTAAG-3′), (5′-ACACGTGGTC-3′), (5′-TCAGCACAGG-3′). Because differential PCR fragments often consist of several unresolved cDNA fragments (Li et al., 1994), cloning proved to be essential prior to undertaking further characterization of the PCR fragments obtained. All clones obtained were subjected to a second screen that consisted of spot-dot northern hybridization performed under conditions of high stringency. This method, which used RNA from entire unsieved embryogenic and non-embryogenic suspension cultures, proved to be a fast and reliable additional selection method. Only one clone (22-28) of the 30 clones obtained after differential screening, proved to be restricted to embryogenic cell cultures while the majority was constitutively expressed. The 26 clones obtained from the cold plaque screening required very long exposure times in the spot-dot northern analysis. Six of these clones failed to show any hybridization signal and 19 proved to be expressed in both embryogenic and non-embryogenic cell cultures (results not shown). One clone (31-50) showed low expression in all embryogenic cultures and in one non-embryogenic culture, but not in the others (Fig. 1). Of the six cloned fragments obtained by ddRT-PCR display, four showed hybridization more or less restricted to transcripts present in embryogenic cultures (6-11, 7-13, 10-25, 11-21; Fig. 1). All clones that passed through the second screening were sequenced. The clones 6-11 and 7-13 were identical to the carrot Lipid Transfer Protein (LTP) gene, previously identified as a marker for embryogenic carrot cell cultures. LTP expression is restricted to embryogenic cell clusters and the protoderm of somatic and zygotic embryos from the early globular stage onwards (Sterk et al., 1991). Therefore, while the LTP gene is not a marker for competent cells, its appearance in the screening confirms the validity of our methods with respect to the cloning of genes expressed early during somatic embryogenesis. cDNA clone 31-50 encodes a leucine-rich repeat containing receptor-like kinase The predicted amino-acid sequence of the cDNA clone 31-50 (Fig. 2A) shows homology with the structural features of plant and animal receptor protein kinases. Because clone 31-50 is Fig. 1. Screening for competent cell markers. 5 µg of total RNA, from non embryogenic cultures with less than 1 embryo per 10,000 cells (−), embryogenic cultures with apprimately 20 embryo per 10,000 cells (+), and high embryogenic cultures with over 100 embryos per 10,000 cells (++), was immobilized and hybridized to labeled DNA probes. The control panel shows hybridization to a constitutively expressed ubiquitin transcript. Exposure of the autoradiographs was for 16 hours, except for clone 31-50 encoding SERK, of which the autoradiograph was exposed for 14 days. Somatic embryo receptor-like kinase 2053 A B -32 ctcatttaattttactttaaaaaataattctatATGAATCGTAACAGTATAAATATATTA M N R N S I N I L 27 28 AATTACATGCAGTTCACTGATGCTTACCTTGACAAATATGGGGTTCTTATGACATTGGAG N Y M Q F T D A Y L D K Y G V L M T L E 87 88 CTTTACAGCAATAACATAAGTGGACCAATTCCTAGTGATCTTGGGAATCTGACAAATTTG L Y S N N I S G P I P S D L G N L T N L * * GTGAGCTTGGACCTATACATGAATAGCTTCTCTGGACCTATACCGGACACATTAGGAAAG V S L D L Y M N S F S G P I P D T L G K 147 SERK P**L**L**L**L*L*NN*LSGPI 5 207 RLK5 P**L**L**L**L*L**N*LSG*I 21 ERECTA P**LG*L**L**L*L**N*L*G*I 19 CTTACAAGGCTAAGATTCTTGCGTCTCAACAACAACAGCCTCTCTGGTCCAATTCCAATG L T R L R F L R L N N N S L S G P I P M * TCACTGACTAATATTACAACTCTTCAAGTCCTGGATTTATCAAACAATCGGCTATCAGGA S L T N I T T L Q V L D L S N N R L S G * CCAGTACCGGATAATGGCTCATTTTCTTTGTTTACACCTATCAGTTTTGCCAATAATTTG P V P D N G S F S L F T P I S F A N N L * AATTTATGTGGACCCGTAACTGGGAGGCCCTGCCCTGGATCTCCCCCATTTTCGCCACCA N L C G P V T G R P C P G S P P F S P P o o CCTCCGTTCATCCCACCATCAACAGTACAGCCTCCAGGACAAAATGGTCCCACTGGAGCT P P F I P P S T V Q P P G Q N G P T G A o o ATTGCTGGGGGAGTAGCTGCTGGTGCTGCTTTACTGTTTGCTGCACCTGCAATGGCATTT I A G G V A A G A A L L F A A P A M A F 267 PRK1 P**L**L**L**L*L**NN**G*I 5 CF-9 PS*L**L**L**LDLSSNNL*G*I 26 TOLL P**LF*H**NL**L*L**N*L**L 15 568 GCATGGTGGCGGAGAAGAAAACCGCGAGAACATTTCTTTGATGTGCCAGCTGAAGAGGAC A W W R R R K P R E H F F D V P A E E D 627 628 CCAGAAGTGCACCTTGGTCAACTGAAGAGGTTTTCTCTGCGAGAATTGCAAGTCGCAACG P E V H L G Q L K R F S L R E L Q V A T 687 688 GATACTTTTAGTACCATACTTGGAAGAGGTGGATTTGGTAAGGTGTATAAGGGACGCCTT D T F S T I L G R G G F G K V Y K G R L 747 748 GCTGATGGCTCACTTGTAGCAGTTAAAAGGCTTAAAGAAGAACGAACACCAGGTGGTGAG A D G S L V A V K R L K E E R T P G G E 807 808 CTGCAGTTTCAAACAGAGGTGGAAATGATTAGCATGGCTGTGCATCGAAATCTTCTGCGT L Q F Q T E V E M I S M A V H R N L L R 867 868 CTACGTGGTTTCTGCATGACACCAACAGAGCGGCTTCTTGTATATCCATACATGGCTAAT L R G F C M T P T E R L L V Y P Y M A N * GGAAGTGTTGCGTCGTGTTTAAGAGAGCGTCAGCCATCAGAACCTCCCCTTGATTGGCCA G S V A S C L R E R Q P S E P P L D W P 927 988 ACTAGGAAGAGGATTGCACTAGGATCTGCTAGGGGGCTTTCTTATTTGCATGACCATTGT T R K R I A L G S A R G L S Y L H D H C 104 1048 GATCCCAAGATTATCCATCGTGATGTAAAAGCTGCAAATATATTATTGGACGAAGAATTT D P K I I H R D V K A A N I L L D E E F 110 1108 GAGGCTGTTGTAGGTGATTTTGGGTTAGCTAGGCTCATGGATTACAAGGATACCCATGTT E A V V G D F G L A R L M D Y K D T H V 116 1168 ACAACTGCTGTAAGGGGTACCTTGGGCTACATAGCTCCCGAGTACCTCTCGACTGGAAAG T T A V R G T L G Y I A P E Y L S T G K 122 1228 TCATCAGAGAAGACCGATGTCTTTGGTTATGGGATTATGCTCTTAGAGCTCATTACTGGA S S E K T D V F G Y G I M L L E L I T G 128 1288 CAGAGAGCTTTTGATCTTGCTCGCCTTGCGAACGATGATGATGTTATGTTGTTGGATTGG Q R A F D L A R L A N D D D V M L L D W 134 1348 GTTAAAAGCCTTTTGAAAGAGAAAAAGTTGGAGATGCTGGTCGATCCTGACCTGCAGAAC V K S L L K E K K L E M L V D P D L Q N 140 1408 AATTACATTGACACAGAAGTTGAGCAGCTTATTCAAGTAGCATTACTCTGTACCCAGGGT N Y I D T E V E Q L I Q V A L L C T Q G 146 1468 TCGCCAATGGAGCGGCCTAAGATGTCAGAGGTAGTCCGAATGCTTGAAGGTGATGGCCTT S P M E R P K M S E V V R M L E G D G L 152 1528 GCAGAAAAGTGGGACGAGTGGCAAAAAGTAGAAGTCATCCATCAAGACGTAGAATTAGCT A E K W D E W Q K V E V I H Q D V E L A 158 1588 CCACATCGAACTTCTGAATGGATCCTAGACTCGACAGATAACTTGCATGCTTTTGAATTA P H R T S E W I L D S T D N L H A F E L 164 1648 TCTGGTCCAAGAtaaacagcatataaaatgtgaatgaaattaatattttttatggttaaa S G P R *** 170 1708 aaaaaaaaaaaaaaa 148 208 268 328 388 448 508 Protein 327 Consensus Sequence 387 447 507 567 C I II 1722 III SERK OsPK10 PRK1 PELLE ATDTFSTILGRGGFGKVYKGRLAD ..SN.CNK..Q....S..L.T.P. LLRASAEV..S.NL.SS..AL.M. DGWSPDNR..Q....D..R.KWK SERK OsPK10 PRK1 PELLE VHRNLLRLRGFCMT H.IH.VK.....TE T.P...P.VAYYYR R.D.I.A.Y.YSIK PTERLLVYPYMANGSVASCLRERQ GPH...A.E......LDKWIFHSK KE.K....D.AS...L..H.H GGKPC...QL.KG..LEAR..AHKAQ VIA VIB SERK R2976 OsPK10 PRK1 PELLE PSEPPLDWPTRKRIALGSARGLSYLHDHCDPK IIHRDVKAANILLDEEF ...A.............N... ...............D. EDDHL...D..FN....T.K..A...QD..S. .V.C.I.PE.V...DN. GNQSR...SS.LK.VK.V.KA.A...NELPSLALP.GHL.SS.V...KYL NPL.A.T.QQ.FS.S..T...IYF..TARGTPL ..G.I.P......QCL SERK RIC2976 OsPK10 PRK1 PELLE EAVVGDFGLARLMDYKDTHV TTAVRGTLGYIAPEY ...F......KP........ ....H..I.H..... I.K.S.....K..TREQS.. F.TL...R..L...W NP.LM.YT.VP.VNLAQ VQHLLVA.K.... QPKI.....V.EGPKSLDA.VEVNK.F..KI.LP..F SERK R2976 OsPK10 PRK1 PELLE LSTGKSSEKTDVFGYGIMLLELITGQRAFDLA ....IL...................K.V.... .TNYAI...S..YS..MV...I.G.RKSY.PS AQQ.RITR....WSL..L...TL..KFPTNYL RNFRQL.TGV..YSF..V...VF..RQVT.RV SERK R2976 OsPK10 PRK1 PELLE KEKKLEMLVDPDLQNNYIDT EVEQLIQVALLCTQGSPMERPKMSEVVRMLE GGNAG.I .......K... ..DS....V........LD....VA.A.I.DE GDLQDIFDAKLKYNDKDG R..TA.K...W.I.DDFYQ..S..K..Q... RDNESAFDKEMNTTKDSQ QIRK.FDIGVA.C.EDLDT.WDLK...QSOQ RQNRM.L.EKHLAAPMGKELD MCMCA.EAG.H..A.D.QD..S.NA.LKRF. SERK R2976 OsPK10 PRK1 PELLE GDGLAEKWDEWQKVEVIHQDVELA PHRTSEWILDSTDNLHAFELSGPR* ..A...R.R....I.IVQ..... GLYKNG.TV...E....V......* .VCEVLQPPVSSQIGYRLYANAFKSSSEEGTSSGMSDYN.DAL.S.VR.....* SLNDKDHGHSNSDQMHDAGV* FTVTD* IV 928 Extracellular LRR Repeats GSLVAVKRLKEERTP GGELQFQTEVEMISMA ..RI...K. .GIGQ .KKE.RS..TI.GSI .QA.V...F. QMNH VAKED.HEHMRRLGRL QLD..I.VMNYRSPNIDQK PV...QSYNELKYLNSI V 987 VII VIII IX X X RLANDDDVMLLDWVKSLL L.RGEG..P...AQTITQ EISEKAHFPSFAFK.LEE A.STGYGTE.AT..DTII PE.ETKKN...Y..QQW XI Fig. 2. Comparison between the SERK amino acid sequence and related sequences. (A) The predicted amino acid sequence of SERK is presented in the single letter code. Asterisks denote potential N-linked glycosylation sites and open circles indicate potential O-glycosylation sites. The putative membrane-spanning region is underlined. The nucleotides are numbered, starting at 1 from the translational start site. Capitals mark the open reading frame. (B) The consensus sequences of SERK LRRs compared with other LRR-containing proteins. Nonconserved amino acids are indicated by an asterisk. The other LRR-containing sequences are from the following proteins: Arabidopsis RLK5 (Walker, 1993), Arabidopsis ERECTA (Torii et al., 1996), Petunia PRKI (Mu et al., 1994), Lycopersicon Cf-9 (Jones et al., 1994) and Drosophila Toll (Hashimoto et al., 1988). (C) Sequence alignment of the predicted kinase domain and the remaining C-terminal part of the carrot SERK protein with protein kinases from Oryza OsPK10 (Zhao et al., 1994), Oryza EST clone R2976 representing a partial cDNA, Petunia PRK1 and Drosophila Pelle (Shelton and Wasserman, 1993). The 11 conserved protein kinase subdomains are indicated I to XI (Hanks et al., 1988). Conserved residues are only indicated in the SERK protein and are represented by a dot in the other sequences. The translation stop sites are indicated by asterisks. 2054 E. D. L. Schmidt and others Fig. 3. The kinase domain of SERK. (A) DNA blot hybridization analysis of carrot genomic DNA hybridized to the HincII-SspI fragment of the SERK cDNA clone encoding the kinase domain. The restriction enzymes indicated were used to digest 10 µg of genomic DNA. (B) Autophosphorylation assays (Mu et al., 1994) were performed using the SERK kinase domain expressed in E. coli. GSTSERK fusion proteins were tested using the 1.4 kb SspI cDNA fragment of the SERK cDNA clone in the reverse orientation (first lane), in the sense orientation in frame with the GST protein resulting in a 70×103 Mr fusion protein (second lane), the GST protein alone (third lane), and an E. coli extract from an untransformed control (fourth lane). Labeled protein was visualized by autoradiography after SDS-PAGE. expressed in embryogenic cell cultures it was renamed Somatic Embryogenesis Receptor Kinase (SERK). The SERK protein contains an N-terminal domain with five leucine-rich repeats (LRRs; Fig. 2B) that is proposed to act as a protein-binding region in LRR receptor kinases (Kobe and Deisenhofer, 1994). Several potential N-glycosylation sites are present in the LRR motifs and one in the intracellular kinase domain (Fig. 2A). Between the extracellular LRR domain of SERK and the membrane-spanning region there is a 32 amino acid region rich in prolines (13), that is unique for the SERK protein. Of particular interest is the sequence SPPPP, which is conserved in extensins, a class of universal plant cell wall proteins (Varner 1 A B and Lin, 1989). The significance of this proline-rich box (Fig. 2A) is not clear, it might act as a hinge region by providing flexibility to the extracellular part of the receptor, or act as a region for interaction with the cell wall. In extensins, usually all prolines in the SPPPP repeat are hydroxylated and are considered to be used as targets for O-linked glycosylation. Although a transmembrane domain is present in the protein, the Nterminal amino acids do not clearly display characteristics of a 3 5 7 12 16 18 1 9 Fig. 4. An example of the embryo formation as observed in a cell tracking experiment. The images of the same area of a culture dish are taken from videotapes used to record development. The number of days of culture after fragmentation of the activated hypocotyl explants is indicated in each frame. The cells were grown in medium with a low amount of 2,4-D for 7 days after fragmentation. After this period, the 2,4-D was removed to allow embryo formation to proceed. Bar: 100 µm. Somatic embryo receptor-like kinase 2055 A a ❉ eg Cc b B Ee d D ❉ ❉ ❉ ❉ ❉ Fig. 5. Examples of cell populations released from activated explants. (A) Based on the size and on the known position of particular cell types within the hypocotyl before fragmentation, 4 morphologically different types of cells were distinguished, and examples of each are shown: black circle: small cytoplasm-rich cell; black square: enlarging cell; white circle: enlarged cell; white square: large cell. The enlarged cell labeled with an asterisk is a competent cell, since it gave rise to an embryo as determined by cell tracking (see Fig. 4). (B-E) Asterisks indicate some of the 24,722 cells recorded that were observed to form embryos in the cell tracking experiment. Bar: 100 µm. signal peptide. The proposed intracellular domain of the protein contains the 11 subdomains characteristic of the catalytic core of protein kinases (Hanks et al., 1988; Fig. 2C). The core sequences HRDVKAAN and GTLGYIAPE in respectively the kinase subdomains VI and VIII suggest a function as a serine / threonine kinase (Hanks et al., 1988). Another interesting feature of the intracellular part of the SERK protein is that the second half of the C-terminal motif resembles an LRR and is also present in two other plant proteins that resemble protein kinases (Fig. 2C). This domain may be involved in mediating the protein-protein interaction necessary for transmission of an intracellular phosphorylation cascade. Hybridization of the SERK cDNA clone to the carrot genome revealed the presence of only a single main hybridizing band, indicating a single SERK gene in the carrot genome (Fig. 3A). No signal was observed after northern blotting of mRNA from embryogenic cell cultures and hybridization with labeled SERK probes (results not shown), reflecting the Table 1. Correspondence between number and cell type of embryo-forming cells and cells expressing the SERK gene Cell tracking In situ hybridization Total no. of cells No. of cells forming embryos (% of total) Total no. of cells No. of cells expressing SERK (% of total) 4976 0 325 0 Enlarging (±16×40 µm) 10764 0 763 0 Enlarged (±35×90 µm) 3511 20 (0.56%) 1593 7 (0.4%) Large (>60×140 µm) 5471 0 850 0 Cell type Small cytoplasm rich (±16×16 µm) 2056 E. D. L. Schmidt and others Fig. 6. SERK gene expression during hypocotyl explant activation. Gene expression is visible as a purple precipitate in individual cells. (A-E) Cell population obtained by mechanical fragmentation of the 2,4 D-treated hypocotyls. Only a few of the enlarged cells show SERK expression (asterisks). (F) Longitudinal vibrotome section of a hypocotyl before 2,4-D treatment shows complete absence of SERK expression in any type of cell. (G-I) Proliferating cell masses that originate from the provascular tissue (longitudinal section). In G a single enlarged cell shows SERK expression while the adjacent rows of morphologically identical cells do not express the gene. In H a single enlarged SERK-expressing cell is in the process of detachment from the surface of the proliferating mass. In I a small cluster of enlarged cells showing SERK expression is present at the surface of proliferating tissue. (J) Proliferating mass of cells developed from the provascular tissue of the hypocotyl 10 days after the beginning of a short (24 hours) 2,4-D treatment followed by culture in basal medium. The lateral root primordia and the enlarged cells detaching from the surface are devoid of SERK mRNA. Bar: 50 µm. low levels of transcript present in these cultures. Detection of the SERK transcript on the original spot-dot northerns was only possible after long exposure times compared with other probes (Fig. 1). Recently a number of LRR receptor-like kinases have been isolated from other plant species (Walker, 1993, 1994). Homology between the SERK protein and receptor-like kinases from Petunia (Mu et al., 1994) and Oryza (Zhao et al., 1994) could be observed (Fig. 2C), as well as with some of the recently identified pathogen resistence gene products (Fig. 2B) like tomato Cf-9 (Jones et al., 1994). BLAST database searches resulted in the identification of the dbEST clone R2976. This partial rice clone shows a 74% identity on the amino acid level with SERK, suggesting that this gene is highly conserved between monocots and dicots. The ability of the SERK protein to autophosphorylate was investigated in vitro, using a previously described autophosphorylation assay (Mu et al., 1994), with a bacterial fusion protein that contained the complete intracellular region of the SERK protein. The bacterially expressed SERK fusion protein indeed proved to be able to autophosphorylate (Fig. 3B), indicating that the SERK protein is able to fulfill a role as a protein kinase. Expression of the SERK gene corresponds to the first appearance of competent cells during hypocotyl activation The formation of competent cells was determined after Fig. 7. SERK gene expression in cell cultures. RT-PCR products obtained from RNA of entire hypocotyls after 10 days of culture are present in lane one (after 2 days of 2,4-D treatment, nonembryogenic culture) and lane two (after 3 days of 2,4-D treatment, embryogenic culture). Lane three shows RT-PCR products obtained from an established non-embryogenic cell culture and lane four an established embryogenic culture. The electrophoresis pattern of the resulting 680 bp PCR fragment was blotted and hybridized with a probe containing the kinase domain of SERK. As a control, a northern blot with similar amounts of RNA as used for the RT-PCR reaction, was hybridized with an 18S ribosomal probe. exposing seed-derived carrot hypocotyl explants to 2,4-D (Guzzo et al., 1994). When carrot hypocotyls are induced with Somatic embryo receptor-like kinase 2057 2,4-D, only the cells of the provascular tissue proliferate. This up the newly initiated embryogenic suspension culture are suggests that the cells derived from this tissue form all the actually competent to form embryogenic cells. The expression different cells, including the embryogenic ones, that are of the SERK gene was determined by whole-mount in situ present in a newly initiated suspension culture. Explant cells hybridization in a population of cells similar to the one used of cortical and epidermal origin only expand, and are quickly for the cell tracking experiment. Examples of the results lost upon subculture. The duration of 2,4-D treatment is obtained are shown in Fig. 6A-E, and the results of 3,531 cells important in the formation of embryogenic cells in activated hybridized with the SERK antisense riboprobe are included in hypocotyls (Guzzo et al., 1994), and has to be at least 3 days, Table 1. Expression of the SERK gene was found to be with an optimal period of around 7 days. In the presence of restricted to only 0.44% of the enlarged cell type. Therefore, 2,4-D, the formation of competent cells and the transition the expression of the SERK gene appears to be closely corretowards embryogenic cells is initiated. After removal of 2,4-D, the formation of somatic embryos from embryogenic cells occurs after 2-3 weeks. The first appearance of single competent cells in the explant was determined experimentally by semi-automatic cell tracking (Toonen et al., 1994) and was performed on large populations of immobilized cells. Hypocotyl explants activated with 2,4-D for 7 days were first washed to remove previously released cells and to ensure that only cells still in the explant were included in the analysis. After mechanical fragmentation of the explants, samples of the resulting population of mainly single cells were immobilized to allow recording of their development by cell tracking. A typical example of the recordings made is shown in Fig. 4. In the immobilized cell populations obtained in this way all the morphologically discernible cell types were present that were also seen in the unfragmented activated hypocotyls. Fig. 5A-E shows the different types of cells, which can be divided into four groups according to mean size and cytoplasmic content (Table 1). Because the same cell types were observed in sections of activated hypocotyl explants (Guzzo et al., 1995), it was possible to predict the position of each type of cell in the explant. Small cytoplasm-rich cells (16×16 µm) are found as the proliferating cells that surround the vascular elements. Enlarging vacuolated cells (16×40 µm) are encountered on the surface of the mass of proliferating cells and these can detach from the surface when fully enlarged (35×90 µm). Large vacuolated cells (more than 60×140 µm) are the non-proliferating remnants of the hypocotyl epidermis and cortical parenchyma. The shape of the enlarging and fully enlarged cells ranged from oval to elongate or triangular. Cell tracking on a total of 24,722 cells released from 7-day activated hypocotyls showed that only 20 single cells formed a somatic embryo. Because of their dependance on Fig. 8. SERK gene expression in established embryogenic cell cultures and in zygotic embryos. Gene expression is visible as a purple precipitate. (A-D) Very continued 2,4-D treatment, the embryo-forming few cells of all morphologically recognizable types of single cells in single cells are still in the competent cell stage. All embryogenic suspension cultures show SERK expression. (E-I) Embryo cell of the embryo-forming single cells belonged to the culture. SERK transcripts are not detectable in large clusters. (E) SERK category of 3,511 enlarged cells (Fig. 5B-E) that transcripts are present in small clusters. (F,G) SERK expression is detected in therefore contained competent cells in a frequency small globular embryos. No signal could be detected using a sense SERK of 0.56%. The cell tracking experiment clearly riboprobe (H) or during later stages of somatic embryogenesis (I). (J) SERK reveals that none of the highly cytoplasmic and transcript is detectable in an approximately 8-celled embryo from a seed rapidly proliferating cells has reached the collected at 14 DAP. Note the absence of any hybridization in the other tissues competent cell state, but that only after elongation of the seed. (K) An early globular embryo at 17 DAP of approximately 100 the first competent cells form. It is also evident that cells. (L) An embryo at the globular-heart transition stage at 17 DAP. Bar: A-I, 50 µm; J-L, 100 µm. only a very limited number of the cells that make 2058 E. D. L. Schmidt and others and the first appearance of competent cells in explants treated for at least 3 days with 2,4-D before being returned to basal medium. Hybridization of northern blots, derived from RNA electrophoresis patterns on formamide gels, never gave any signal after hybridization with SERK cDNA probes, not even after prolonged exposure in a PhosphorImager (results not shown), in line with the very restricted expression pattern of the SERK gene. Fig. 9. SERK gene expression in plant organs. RT-PCR products obtained from flower/seed RNA at 0 to 20 days after pollination (lanes one to five), from leaves (lane six), seedling stems (lane seven) and seedling roots (lane eight). The electrophoresis pattern of the resulting 680 bp PCR fragment was blotted and hybridized with a probe containing the kinase domain of SERK. As a control, a northern blot with similar amounts of RNA as used for the RT-PCR reaction, was hybridized with an 18S ribosomal probe. lated, both qualitatively and quantitatively, with the presence of competent single cells. To obtain insight into the temporal regulation of SERK expression during explant activation, whole mount in situ hybridization was performed on entire intact or handsectioned explants treated for different periods with 2,4-D (Fig. 6F-J). Representative samples were collected from explants that were untreated (Fig. 6F) and treated for 7 days (Fig. 6G, H) or 10 days (Fig. 6I) with 2,4-D. While enlarged cells became present after the first 5 days of culture, the first few SERK-expressing enlarged cells were found after 7 days of culture in the presence of 2,4-D (Fig. 6H). These cells were present at the surface of the mass of proliferating cells originating from the provascular tissue. In the hypocotyls treated for 10 days with 2,4-D, the number of SERK-positive cells had increased to about 3% and included at this stage cells also present in small clusters (Fig. 6I). No SERK transcript was ever detected in small cytoplasm-rich cells or large vacuolated cells. Hypocotyls treated for only 1 day with 2,4-D and subsequently cultured in hormone free medium for 10 days showed proliferating explant cells that gave rise to roots and non embryogenic cell cultures. SERK expression could never be detected in such explants (Fig. 6J). The in situ hybridization results described above were obtained from a relatively small number of explants, so RT-PCR followed by Southern hybridization was performed to obtain more quantitative results. These are shown in Fig. 7 and confirm the close temporal correlation between the expression of the SERK gene The SERK gene is expressed in established embryogenic cell cultures and transiently during somatic and zygotic embryogenesis While the results described so far indicate that competent and embryogenic cell formation is restricted to a particular class of enlarged cells during explant activation, the situation in an established embryogenic cell culture is more complex. Competent single cells in such cultures do not appear to belong to one cell type in particular, but have been shown to originate from all morphologically different cell types. Embryogenic cells, that do not require exogenous auxin treatment, are thought to be present only in the form of clusters of at least 3-4 cells and not as single cells (Toonen et al., 1994). SERK expression was found in all morphologically discernible single cell types that were present in an embryogenic cell culture (Fig. 8A-D) at a frequency between 0.1 and 0.5% (results not shown) depending on the cell type. In non-embryogenic cultures, SERK-expressing cells were never encountered. Unfortunately, the non-embryogenic culture in which the original spot-dot northern showed SERK expression was lost and could not be included in this analysis. We expect that in this particular cell culture the competent cell stage is initiated but not completed. As was observed in the activated explants, SERK expression was not restricted to single cells, but also occurred in small clusters of 2 to 8 cells (Fig. 8E). Since clusters of this size are known to consist of embryogenic cells, these data show that SERK expression is not restricted to competent single cells, but may persist in small clusters of embryogenic cells. SERK expression was also followed during the course of somatic embryo development. While, in small globular somatic embryos of up to about 100 cells, there is a high level of expression (Fig. 8F,G), no SERK expression was encountered during the mid to late globular, heart (Fig. 8I) and torpedo-stages of somatic embryogenesis. Whole-mount in situ hybridization on partially dissected carrot seeds showed that the SERK gene was only expressed in early embryos up to the globular stage (Fig. 8J,K). No expression luciferase day 1 day 1 day 2 day 3 day 6 day 9 day 13 Fig. 10. Luciferase expression under control of the SERK promoter. Luciferace activity of immobilised cells was recorded at day 1 with a CCD camera (left image). The single pixel in the left image measures 30 by 30 µm. Video cell-tracking of the cells was performed for a period of 13 days (light microscopic images). The light microscope images were sized to match the CCD image. Somatic embryo receptor-like kinase 2059 was seen in seedlings, roots, stems, leaves, developing and mature flower organs, pollen grains and stigmas before and after fertilization (results not shown). Tissues in the developing seed such as seed coat, integuments, all embryo sac constituents before fertilization, as well as the endosperm at all stages of development investigated, did not show any SERK expression. Later stages of carrot zygotic embryos (Fig. 8L) were also completely devoid of SERK mRNA. These results were confirmed using RT-PCR (Fig. 9) and indicate that no SERK mRNA accumulates in any of the adult plant organs nor in flowers prior to pollination. The first occasion when SERK expression can be detected is in flowers at 3 days after pollination (DAP), at which stage fertilization has taken place and endosperm development has commenced. SERK mRNA remains present in flowers up to 20 DAP, corresponding to the early globular stage of the zygotic embryo (Lackie and Yeung, 1996). Therefore, the SERK mRNA as detected by RT-PCR in flowers at 3 and 7 DAP is likely to come from SERK gene expression in the zygotes, because in carrot the zygote remains undivided up to 1 week after pollination (Lackie and Yeung, 1996). Attempts to perform wholemount in situ hybridisation on seeds containing only zygotes have so far been unsuccessful. SERK promoter-luciferase expression during somatic embryogenesis To determine directly whether SERK-expressing cells indeed develop into somatic embryos, transformed carrot suspension cultures containing a SERK promoter-luciferase construct were analysed for luciferase expression in cell cultures sieved through a 50 µm mesh to enrich for single cells and small cell clusters. Development of the immobilized cells after recording the luciferase images was determined using automated cell tracking (Toonen et al., 1994, 1997). The origin of nine torpedo stage somatic embryos was determined this way. Of these, three developed from a single cell that showed luciferase activity at day 1, four developed from cell clusters consisting of 2-6 luciferase-expressing cells while two embryos developed from single cells that failed to show a detectable level of luciferase activity at day 1. The somatic embryo shown in Fig. 10 originated from a luciferase-expressing two-celled cluster. These results demonstrate that most somatic embryos develop from single cells and small cell clusters expressing SERK at day 1. Somatic embryos beyond the globular stage did not show luciferase expression, confirming the transient SERK gene expression pattern (results not shown). DISCUSSION The SERK gene as a marker for the competent cell stage in somatic embryogenesis In plants, embryo formation in the absence of the fusion of two gametes is a widespread phenomenon. It occurs naturally in certain species in the developing ovule, as exemplified by apomictic embryogenesis (Koltunow, 1993), or on the surface of leaves as in Malaxis (Taylor, 1967). More commonly, embryogenesis can be induced experimentally with a wide variety of tissue explants as the starting material after treatment of explants with synthetic growth regulators. It is generally accepted that the genetic and physiological consti- tution of the donor plant as well as the age and type of the explant are important parameters in succesfully inducing somatic embryogenesis. However, knowledge about the first events that take place during the transition of somatic cells into embryogenic cells is largely lacking. There appear to be two main reasons for this. The first is that cellular changes observed in somatic explant cells that have responded to the inducing treatment in general have not been proved to be essential for the formation of embryogenic cells. The second is that no specific markers have been described so far that reliably predict which explant cells will become embryogenic. The process of cellular reactivation and the subsequent formation of embryogenic cells in carrot explants has been described by Guzzo et al. (1994, 1995). That work showed that a particular elongated cell type appeared in culture, derived from small rapidly proliferating cytoplasmic cells that themselves derived from reactivated provascular cells. It was further shown cytologically, that some of the elongated cells underwent an asymmetrical division. After continued culture, small clusters of dividing cytoplasmic cells appeared that resembled the proembryogenic masses seen in established embryogenic suspension cultures (Guzzo et al., 1994, De Vries et al., 1988a). The fact that only a limited number of cells actually undergoes the transition of somatic into embryogenic cell is postulated to be the result of the presence of different sets of auxin receptors (Filippini et al., 1992). One of the goals of the work presented here was to determine which cells of a carrot hypocotyl explant have completed the transition of a somatic cell into an embryogenic cell. The results revealed that the first cells to become competent belong to a type of enlarged cells that are detaching from the surface of the mass of proliferating cells, confirming the previous predictions (Guzzo et al., 1994). All other cells, including the majority of the enlarged cells present, were completely unresponsive. This result is in contrast to the generally accepted idea that a population of small rapidly dividing meristematic cells are the ones that are competent to become embryogenic. Instead, the results presented here, together with the cytological observations (Guzzo et al., 1994), demonstrate that competent cells arise first as elongating cells, when still attached to the explant. This is a situation that is strikingly similar to the rapid increase in cell volume observed in plant egg cells after fertilization (Mansfield and Briarty, 1991). Unlike the zygote, it was so far not possible to predict which elongated cell on the surface of the explant will become competent. This raises the question whether the observed correspondence between elongation and competent cell formation is causal or merely reflects a particular state common to most cells present on the surface of the explant. To answer this question, it was essential to achieve another goal of this work: to obtain markers that are able to distinguish precisely between competent and non-competent cells. The expression of the SERK gene described here was found to be very tightly correlated with the ability of cells of the correct morphology to attain the competent cell state. In line with the much less defined morphology of competent single cells in established suspension cultures (Toonen et al., 1994), no clear cell type specificity was apparent in cells expressing the SERK gene in those cultures. The difference between these findings and the 2060 E. D. L. Schmidt and others results obtained employing activated explants is at present not explained. In comparison with other markers that have been suggested to distinguish individual embryo-forming cells, such as callose (Dubois et al., 1991) and the monoclonal antibody JIM8 (Pennell et al., 1992), the SERK gene appears to be quite specific under culture conditions. While the presence of the JIM8 epitope was restricted to embryogenic cell cultures, cell tracking of cells labeled with this antibody failed to show a correlation with the ability of these cells to develop into somatic embryos (Toonen et al., 1996a). As shown in this work, such a correlation was established for the SERK gene by cell tracking of cells expressing luciferase under the control of the SERK promoter. The availability of a vital marker for competent cells offers the possibility of following, with great precision, the events that take place during formation of such cells. It could for instance be a useful tool to help determine the frequency of pseudomeiotic segregation events proposed to accompany embryogenic cell formation (Giorgetti et al., 1995). The identification of the SERK gene and other markers for competent cells was facilitated by the availability of a range of suspension cultures differing strongly in the number of competent single cells present. This strategy avoided the potential problem that a non embryogenic cell line would be employed in which competent cell formation had occurred, but in which a later stage in embryogenic cell formation would have been inhibited, thus reducing the chance of finding genes expressed during the early competent cell states. In addition, the availability of in situ hybridization methods that allowed the visualization of gene expression in single suspension cells while preserving cell morphology, together with cloning methods aimed at avoiding selection towards abundantly expressed genes, explains why similar experiments carried out previously only yielded genes expressed in later stages, such as proembryogenic masses (Choi and Sung, 1984; Wilde et al., 1988; Aleith and Richter, 1990). The SERK gene is transiently expressed in embryogenesis The analysis of the expression pattern of the SERK gene during embryogenic cell formation and during somatic embryogenesis revealed that SERK expression continues during proembryogenic mass formation and also during somatic embryogenesis up to about the 100-celled globular stage. Because after this stage expression in somatic embryos is completely abolished, the SERK gene shows a transient expression pattern, one that was also found in zygotic embryogenesis, perhaps as early as the zygote. This transient expression pattern, following the classification system for embryo-expressed genes would place the SERK gene in class 2, comprising very few genes exclusively expressed during early embryogenesis (Sterk and De Vries, 1992). So far, none of the genes identified using embryo mutational analysis, such as Bio-1 (Schneider et al., 1989), Prolifera (Springer et al., 1995), EMB30/Gnom (Shevell et al., 1994), Fusca-1 (Castle and Meinke, 1994), Knolle (Lukowitz et al., 1996) and STM (Long et al., 1996) exhibit an expression pattern that is restricted to the embryo. Thus, the expression pattern of the SERK gene points to a function in a signal transduction cascade only required for the first seven or so cell divisions of the plant embryo. While the nature of the signal, its transduction and its importance are not clear yet, it is clear that this cascade is reproduced with great fidelity in somatic embryogenesis. The possible function of the SERK gene in early plant embryogenesis The predicted SERK protein sequence resembles a leucinerich repeat (LRR) receptor kinase protein, a class of plant proteins that was originally described by Chang et al. (1992). Some members of this class of plant receptor-like kinases are known regulators of developmental processes, like the Arabidopsis Clavata protein (Meyerowitz, 1995) and the Arabidopsis Erecta protein (Torii et al., 1996). Others, like the Petunia PRK1 protein seem to be involved in signal transduction during pollen development or pollination (Mu et al., 1994). Expression of this particular Petunia gene is restricted to pollen and pollen tubes prior to fertilization in contrast to the carrot SERK gene of which no expression can be detected in pollen. PRK1 and SERK therefore regulate non-overlapping processes, separated from each other by the process of fertilization. Other plant receptor-like proteins with LRRs are involved in pathogen resistence, presumably by the specific binding to elicitors (reviewed by Dangl, 1995). The specificity of protein-protein interactions mediated by LRR-containing proteins are most probably due to the composition of the nonconsensus residues within the LRRs (Kobe and Deisenhofer, 1994). The presence of the perfect consensus sequence SPPPP found in extensins and some types of arabinoalactan proteins or AGPs suggests an interaction of the extracellular part of the SERK protein with components of the cell wall. The SPPPP domain is considered to be a target sequence for arabinosylation onto hydroxylated prolines (reviewed by Carpita and Gibeaut, 1993). All prolines in the SPPPP consensus are normally hydroxylated. Whether a glycosylated SPPPP sequence is mediating a possible anchoring to specific regions of the cell wall remains to be determined. If so, it would be an elegant mechanism to prevent free movement of receptor molecules and yet prevent inflexibility preventing dimerization, which would be difficult to reconcile with covalent attachment to a cell wall polymer. While the average primary cell wall has a thickness of approximately 50 nm (Pruitt et al., 1993) and the maximum size of the entire extracellular domain can only be about 15 nm when present as an α helix, the extracellular ligand binding domain is likely to be completely embedded within the cell wall. The most likely type of ligand for SERK will therefore consist of a cell wall-diffusible peptide. Peptides effective in inducing plant responses, such as systemin (McGurl et al., 1992) and ENOD40 (Van de Sande et al., 1996) have been described. Thus, it appears that while LRR containing receptor-like protein kinases play several roles in plant development, intercellular peptides are now being uncovered that are likely signal molecules that can activate developmental processes mediated through such receptors. Clues about the function of SERK might be found in the homology between SERK and two proteins in Drosophila that are required for the establishment of the dorsoventral polarity in the embryo. The kinase domain of SERK shows homology with the Drosophila Pelle protein, a serine/threonine kinase Somatic embryo receptor-like kinase 2061 involved in formation of the dorsoventral axis during embryogenesis (Shelton and Wasserman, 1993). The Drosophila Pelle protein itself is activated by the Toll transmembrane receptor (Hashimoto et al., 1988; Govind and Steward, 1991), of which the ligand-binding domain, as in SERK, consists of LRRs. In the plant embryo sac and in the activated explant a situation may exist whereby an unknown inducer is present uniformly, but embryo formation awaits the presence of the SERK protein. Such a model may fit with the restricted expression pattern found for the SERK gene both in vivo and in vitro. It is also in line with the hypothesis that in plants inductive interactions mediated by diffusable signal molecules are an important regulatory mechanism (reviewed by Schmidt et al., 1994). Direct evidence for the existence of cell inductive processes in plants was recently presented by Van Den Berg et al. (1995) for the Arabidopsis root. While most of the elements concerning the origin and targets of processes of cell to cell communication in early plant embryogenesis are lacking, the SERK gene described here may represent a significant part of a mechanism that is essential for the formation of plant cells destined to become embryos. We would like to thank Wim Reidt and Conny Eijkelboom for practical assistance. Marijke Hartog and Tony van Kampen are acknowledged for nucleotide sequence analysis, Dorus Gadella for help with luciferase measurement and Olaf Sonneveld (S&G Seeds, Enkhuizen, the Netherlands) for the carrot plant material. E. D. L. S. and M. A. J. T. are supported by the Technology Foundation (STW). F. G. is supported by EMBO, the Italian government and the EC program PTP-Biotech. The rice clone R2976 was made available by the Rice Genome Research Program (RGP). REFERENCES Aleith, F. and Richter, G. (1990) Gene expression during induction of somatic embryogenesis in carrot cell suspensions. Planta 183, 17-24. Chang, C., Schaller, G. E., Patterson, S. E., Kwok, S. F., Meyerowitz, E. M. and Bleecker, A. B. (1992). The TMK1 gene from Arabidopsis codes for a protein with structural and biochemical characteristics of a receptor protein kinase. Cell 4, 1263-1271. Choi, J. H. and Chung, Z. R. (1984). Two-dimensional gele analysis of carrot somatic embryonic proteins. Plant Mol. Biol. Rep. 2, 19-25. Castle, L. A. and Meine, D. W. (1994). A FUSCA gene of Arabidopsis encodes a novel protein essential for plant development. Plant Cell 6, 25-41. Carpita, N. C. and Gibeaut, D. M. (1993). Structural models of primary cell walls in flowering plants-consistency of molecular structure with the physical properties of the walls during growth. Plant J. 3, 1-30. Dangl, J. L. (1995). Pièce de Resistance: novel classes of plant disease resistance genes. Cell 80, 363-366. De Jong, A. J., Schmidt, E. D. L. and De Vries, S. C. (1993). Early events in higher plant embryogenesis. Plant. Mol. Biol. 5, 367-377. De Vries, S. C., Booij, H., Meyerink, P., Huisman, G., Wilde, H. D., Thomas, T. L. and Van Kammen, A. (1988a). Acquisition of embryogenic potential in carrot cell-suspension cultures. Planta 176, 196-204. De Vries, S. C., Hoge, H., and Bisseling, T. (1988b). Isolation of total and polysomal RNA from plant tissues. In Plant Molecular Biology Manual (ed. S. B. Gelvin, R. A. Schilperoort and D. P. S. Verma), pp. 1-13. Kluwer Academic Publishers, Dordrecht. Dos Santos, A. V. P., Cutter, E. G., Davey, M. R. (1983). Origin and development of somatic embryos in Medicago sativa L. (Alfalfa). Protoplasma 117, 107-115. Dubois, T., Guedira, M., Dubois, D. and Vasseur, J. (1991). Direct somatic embryogenesis in leaves of Cichorum – A histological and SEM study of early stages. Protoplasma 162, 120-127. Dudits, D., Gyorgyey, J., Bogre, L. and Bako, L. (1995). Molecular Biology of somatic embryogenesis. In In vitro Embryogenesis in Plants. (ed. T. A. Thorpe), pp. 267-308. Kluwer Academic Publishers, Dordrecht. Engler, J. A., Van Montagu M. and Engler G. (1994). Hybridization in situ of whole mount messenger RNA in plants. Plant Mol. Biol. Rep. 12, 321-331. Filippini, F., Terzi, M., Cozzani, F., Vallone, D. and Lo Schiavo, F. (1992). Modulation of auxin-binding proteins in cell-suspension: II. Isolation and initial characterization of carrot cell variants impaired in somatic embryogenesis. Theor. Applied Genetics 84, 430-434. Gamborg, O. L., Miller, R. A. and Ojima, K. (1968). Plant cell cultures: I. Nutrient requirements of suspension cultures of soybean root cells. Exp. Cell. Res. 50, 151-158. Giorgetti, L., Vergara, M. R., Evangelista, M., LoSchiavo, F., Terzi, M. and Ronchi, V. N. (1995). On the occurrence of somatic meiosis in embryogenic carrot cell cultures. Mol. Gen. Genet. 246, 657-662. Goldberg, R. B., Barker, S. J., Perez-Grau, L. (1989). Regulation of gene expression during plant embryogenesis. Cell 56, 149-160. Goldberg, R. B., de Paiva, G. and Yadegari, R. (1994). Plant embryogenesis: zygote to seed. Science 266, 605-614. Govind, S. and Steward, R. (1991). Dorsoventral pattern formation in Drosophila. Trends Genet. 7, 119-125. Guzzo, F., Baldan, B., Mariani, P., LoSchiavo, F. and Terzi, M. (1994). Studies on the origin of totipotent cells in explants of Daucus carota L. J. Exp. Bot 45, 1427-1432. Guzzo, F., Baldan, B., Levi, M., Sparvoli, E., LoSchiavo, F., Terzi, M., and Mariani, P. (1995). Early cellular events during induction of carrot explants with 2,4-D. Protoplasma 185, 28-36. Hanks, S. K., Quinn, A. M. and Hunter, T. (1988). The protein kinase family: conserved features and deduced phylogeny of the catalytic domains. Science 241, 42-52. Hashimoto, C., Hudson, K. L. and Anderson, K. V. (1988). The Toll gene of Drosophila, required for dorsal-ventral embryogenic polarity, appears to encode a transmembrane protein. Cell 52, 269-279. Heck, G. R., Perry, S. E., Nichols, K. W. and Fernandez, D. E. (1995). AGL15, a MADS domain protein expressed in developing embryos. Plant Cell 7, 1271-1282. Hodge, R., Paul, Wyatt, Draper, J. and Scott, R. (1992). Cold-plaque screening: a simple technique for the isolation of low abundance, differentially expressed transcripts of low abundance, differentially expressed transcripts from conventional cDNA libraries. Plant J. 2, 257-260. Horn, M. A. and Walker, J. C. (1994). Biochemical properties of the autophosphorylaton of RLK5, a receptor-like protein kinase from Arabidopsis thaliana. Biochim. Biophys. Acta 1208, 65-74. Jones, D. A., Thomas, C. M., Hammond-Kosack, K. E., Balint-Kurti, P. J. and Jones, J. D. G. (1994). Isolation of the tomato Cf-9 gene for resistance to Cladosporium fulvum by transposon tagging. Science 266, 789-792. Kobe, B. and Deisenhofer, J. (1994). The leucine-rich repeat: a versatile binding motif. Trends Biochem. 19, 415-421. Koltunow, A. M. (1993). Apomixis: embryo sacs and embryos formed without meiosis or fertilization in ovules. Plant Cell 5, 1425-1437. Komamine, A., Matsumoto, M., Tsukahara, M., Fujiwara, A., Kawahara, R., Ito, M., Nomura, K. and Fujimura, T. (1990). Mechanisms of somatic embryogenesis in cell cultures – physiology, biochemistry and molecular biology. In Progress in Plant Cellular and Molecular Biology (ed. H. J. J. Nijkamp, L. H. W. Van Der Plas, A. Van Aartrijk), pp. 307- 313. Kluwer Academic Publishers, Dordrecht. Konar, R. N., Thomas, E. and Street, H. E. (1972). Origin and structure of embryoids arising from epidermal cells of the stem of Ranunculus sceleratus L. J. Cell Sci. 11, 77-93. Lackie, S. and Yeung, E. C. (1996). Zygotic embryo development in Daucus carota. Can. J. Bot. 74, 990-998. Li, F., Barnathan, E. S. and Kariko, K. (1994). Rapid method for screening and cloning cDNAs generated in differential mRNA display: application of Northern blot for affinity capturing of cDNAs. Nucl. Acid Res. 22, 17641765. Liang, P. and Pardee, A. B. (1992). Differential display of eucaryotic messenger RNA by means of the polymerase chain reaction. Science 257, 967-971. Long, J. A., Moan, E. I., Medford, J. I. and Barton, M. K. (1996). A member of the KNOTTED class of homeodomain proteins encoded by the STM gene of Arabidopsis. Nature 379, 66-69. Lukowitz, W., Mayer, U. and Jürgens, G. (1996). Cytokinesis in the Arabidopsis embryo involves the syntaxin-related KNOLLE gene product. Cell 84, 61-71. Mansfield, S. G. and Briarty, L. G. (1991). Early embryogenesis in Arabidopsis thaliana. II. The developing embryo. Can. J. Bot. 69, 461-476. McGurl, B., Pearce, G., Orozco-Cardenas, M. and Ryan, C. A. (1992). 2062 E. D. L. Schmidt and others Structure, expression, and antisense inhibition of the systemin precursor gene. Science 257, 1570-1573. Meyerowitz, E. M. (1995). Pattern formation in Arabidopsis flower development. EDBC 95 congress of the European developmental biology organization. Abstract SI4. Millar, A. J., Short, S. R., Hiratsuka, K., Chua, N. -H., and Kay, S. A. (1992). Firefly luciferase as a reporter of regulated gene expression in higher plants. Plant Mol. Biol. Reporter 10, 324-337. Mordhorst, A. P., Toonen, M. A. J., de Vries, S. C. (1997). Plant Embryogenesis. Critical Reviews in Plant Sciences. Ed. Conger, B. V. (in press) Mu, J-H., Lee H-S. and Kao, T. -H. (1994). Characterization of a pollenexpressed receptor-like kinase gene of Petunia inflata and the activity of its encoded kinase. Plant Cell 6, 709-721. Pennell, R. I., Janniche, L., Scofield, G. N., Booy, H., De Vries, S. C. and Roberts, K. (1992). Identification of a transitional cell state in the developmental pathway to carrot somatic embryogenesis. J. Cell Biol. 119, 1371-1380. Pruitt, R. E., Hulskamp, M., Kopczak, S. D., Ploense, S. E. and Schneitz, K. (1993). Molecular genetics of cell interactions in Arabidopsis. Development 1993 Supplement, 77-84. Rounsley, S. D., Ditta, G. S. and Yanofsky, M. F. (1995). Diverse roles for MADS box genes in Arabidopsis development. Plant Cell 7, 1259-1269. Schmidt, E. D. L., De Jong, A. J. and De Vries, S. C. (1994). Signal molecules involved in plant embryogenesis. Plant Mol. Biol. 26, 1305-1313. Schneider, T., Dinkins, R., Robinson, K., Shellhammer, J. and Meinke, D. W. (1989). An embryo-lethal mutant of Arabidopsis thaliana is a biotin auxotroph. Dev. Biol. 131, 161-167. Scott, R., Dagless, E., Hodge, R., Soufleri, I. and Draper, J. (1991). Patterns of gene exression in developing anthers of Brassica napus. Plant Mol. Biol. 17, 195-207. Shelton, C. A. and Wasserman, S. A. (1993). Pelle encodes a protein kinase required to establish dorsoventral polarity in the Drosophila embryo. Cell 72, 515-525. Shevell, D. E., Leu, W. -M., Gilimor, C. S., Xia, G., Feldmann, K. A. and Chua, N. -H. (1994). EMB30 is essential for normal cell division, cell expansion, and cell adhesion in Arabidopsis and encodes a protein that has similarity to Sec7. Cell 77, 1051-1062. Springer, P. S., Richard, W., McCombie W. M., Sundaresan, V. and Martienssen, R. A. (1995). Gene trap tagging of Prolifera, an essential MCM-2-3-5-like gene in Arabidopsis. Science 268, 877-880. Sterk, P., Booij, H., Schellekens, G. A., Van Kammen, A. and De Vries, S. C. (1991). Cell-specific expression of the carrot EP2 lipid transfer protein gene. Plant Cell 3, 907-921. Sterk, P., and De Vries, S. C. (1992). Molecular markers for plant embryos. In Synseeds: Applications of synthetic seeds to crop improvement (ed. K. Redenbaugh), CRC Press, London. Taylor, R. L. (1967). The foliar embryos of Malaxis paludosa. Can. J. Bot. 45, 1553-1556. Thomas, T. L. (1993). Gene expression during plant embryogenesis and germination: an overview. Plant Cell 5, 1401-1410. Toonen, M. A. J., Hendriks, T., Schmidt, E. D. L., Verhoeven, H. A., Van Kammen, A and De Vries, S. C. (1994). Description of somatic-embryoforming single cells in carrot suspension cultures employing video cell tracking. Planta 194, 565-572. Toonen, M. A. J., Schmidt, E. D. L., Hendriks, T., Verhoeven, H. A., Van Kammen, A. and De Vries, S. C. (1996a). Expression of the JIM8 cell wall epitope in carrot somatic embryogenesis. Planta 200, 167-173. Toonen, M. A. J., Verhees, J. A., Schmidt, E. D. L., Van Kammen, A. and De Vries, S. C. (1996b). AtLTP luciferase expression during carrot somatic embryogenesis. Submitted. Toonen, M. A. J., de Vries, S. C. (1997). Use of video tracking to identify embryogenic cultured cells. Plant Tissue Culture Manual H, (ed. K. Lindsey), pp. 1-45. In press. Torii, K. U., Mitsukawa, N., Oosumi, T., Matsuura, Y., Yokoyama, R., Whittier, R. F. and Komeda, Y. (1996). The Arabidopsis ERECTA gene encodes a putative receptor protein kinase with extracellular leucine-rich repeats. Plant Cell 8, 735-746. Trigiano, R. N., Gray, D. J., Conger, B. V. and McDaniel, K. J. (1989). Origin of direct somatic embryos from cultured leaf segments of Actylis glomerata. Bot. Gaz. 150, 72-77. Van Den Berg, C., Willemsen, V., Hage, W., Weisbeek, P. and Scheres, B. (1995). Cell fate in the Arabidopsis root meristem determined by directional signalling. Nature 378, 62-65. Van de Sande, K., Pawlowski, K., Czaja, I., Wieneke, K., Schell, J., Schmidt, J., Walden, R., Matvienko, M., Wellink, J., Van Kammen, A., Franssen, H. and Bisseling, T. (1996). Modification of phytohormone response by a peptide encoded by ENOD40 of legumes and a nonlegume. Science 273,370-373. Van Engelen, F. A. and De Vries, S. C. (1992). Extracellular proteins in plant embryogenesis. Trends Genet. 8, 66-70. Varner, J. E. and Lin, L. -S. (1989). Plant cell wall architecture. Cell 56, 231239. Walker, J. C. (1993). Receptor-like protein kinase genes of Arabidopsis thaliana. Plant J. 3, 451-456. Walker, J. C. (1994). Structure and function of the receptor-like protein kinases of higher plants. Plant Mol. Biol. 26, 1599-1609. Wernicke, W. and Brettel, R. (1980). Somatic embryogenesis from Sorgum bicolor leaves. Science 287, 138-139. Wilde, H. T., Nelson, W. S., Booij H., De Vries, S. C. and Thomas, T. L. (1988). Gene expression programs in embryogenic and non-embryogenic carrot cultures. Planta 176, 205-213. Wurtele, E. S., Wang, H., Durgerian, S., Nikolau, B. J. and Ulrich, T. H. (1993). Characterization of a gene that is expressed early in somatic embryogenesis of Daucus carota. Plant Physiol. 102, 303-312. Zhao, Y., Feng, X. H., Watson, J. C., Bottino, P. J. and Kung, S. D. (1994). Molecular cloning and biochemical characterization of a receptor-like serine / threonine kinase from rice. Plant Mol. Biol. 26, 791-803. Zimmerman, J. L. (1993). Somatic embryogenesis: a model for early development in higher plants. Plant Cell 5, 1411-1423. (Accepted 7 March 1997)