Plant Cell, Vol. 10, 1277-1294, August 1998, Copyright © 1998, American Society of Plant Physiologists
Inactivation of a Glycyl-tRNA Synthetase Leads to an Arrest in Plant Embryo Development
Ursula Uwer1,a,
Lothar Willmitzerb, and
Thomas Altmannb
a Institut für Genbiologische Forschung, Ihnestrasse 63, 14195 Berlin, Germany
b Max-Planck-Institut für Molekulare Pflanzenphysiologie, Karl-Liebknecht-Strasse 25, 14476 Golm, Germany
Correspondence to:
Thomas Altmann, altmann{at}mpimp-golm.mpg.de (E-mail), 49-331-9772301 (fax).
 |
ABSTRACT |
|---|
Embryo formation is the first patterning process during vegetative plant growth. Using transposons as insertional mutagens in Arabidopsis, we identified the mutant edd1 that shows embryo-defective development. The insertion mutation is lethal, arresting embryo growth between the globular and heart stages of embryonic development. The mutant phenotype cosegregates with a transposed Dissociation element. Sequences flanking the transposed element were isolated and used to isolate a full-length cDNA clone representing the wild-type EDD1 gene. Complementation of the mutant through Agrobacterium-mediated gene transfer of an EDD1 wild-type copy as well as loss of the transposon concomitant with phenotypic reversion demonstrated that the transposon had caused the mutation. Based on homology to Escherichia coli, the EDD1 gene is predicted to encode a novel glycyl-tRNA synthetase (GlyRS) that has not been identified previously in higher plants. An N-terminal portion of the plant protein is able to direct a marker protein into pea chloroplasts. Thus, the gene identified by the embryo-defective insertion mutation encodes a GlyRS homolog, probably acting within the plastidic compartment.
|
|
INTRODUCTION |
|---|
Pattern formation presents one of the most challenging aspects of plant development. It addresses the organizational aspect of development, namely, the origin of diverse cell types, tissues, and organs at specific locations to form a structurally and functionally meaningful context. In flowering plants, most pattern-forming events occur in the postembryonic sporophyte after seed germination. However, the primary plant architecture, which is fairly uniform among flowering plants, is generated during early embryogenesis (reviewed in Jurgens et al. 1991 
; Jurgens and Mayer 1994 ).
Formation of the embryo body can be divided conceptually into three overlapping phases (West and Harada 1993 ; Jurgens and Mayer 1994 ). During the first morphogenetic phase, an apical–basal axis is established throughout the embryo body, with the root apical meristem at one end and the shoot apical meristem at the other. A radial pattern is created by the formation of three primordial tissue layers, that is, protoderm, procambium, and ground meristem, and bilateral symmetry occurs throughout the development of the cotyledons (Goldberg et al. 1994 ; Meinke 1995 ).
The second phase is characterized by embryo growth and the accumulation of storage reserves (Soueges 1919 ; Muller 1963 ; Goldberg et al. 1989 ). During the final phase, the embryo prepares for desiccation, dehydrates, and enters a period of developmental arrest. The ovule develops into the seed, which is surrounded by an elaborated coat that enables the differentiated embryo to survive harsh environmental conditions until germination and sporophytic development are favorable.
Higher plant embryogenesis was investigated extensively during the past century (Hanstein 1870 ; Vandendries 1909 ; Schnarf 1929 ; Maheshwari 1950 ). Studies using light and electron microscopy have provided detailed descriptions of the morphological and anatomical changes that characterize embryonic development (Schulz and Jensen 1969 ; Tykarska 1979 ; Mansfield and Briarty 1990a , Mansfield and Briarty 1990b , Mansfield and Briarty 1991a , Mansfield and Briarty 1991b ; West and Harada 1993 ; Goldberg et al. 1994 ; Jurgens and Mayer 1994 ).
Cellular differentiation has been studied largely in relation to the biosynthesis and accumulation of storage proteins, lipids, and starch macromolecules, which ultimately serve as nutrients for the postgerminative seedling (Jenner 1982 ; Slack and Browse 1984 ; Shotwell and Larkins 1989 ; Galili et al. 1993 ). The development of somatic embryo systems (De Jong et al. 1993 ; Liu et al. 1993 ; Zimmerman 1993 ) as well as studies on hormone action (Schiavone and Cooke 1987 ; Liu et al. 1993 ; Fischer and Neuhaus 1996 ) have revealed at least some of the signals that direct proper embryo development.
Genetic analysis provides a complementary approach to the study of plant embryo development. It allows genes with essential functions during this critical stage of the plant life cycle to be identified. To date, only a few genes have been isolated that are expressed specifically during early embryogenesis, and very little is known about genes acting on embryo pattern. One way to identify "embryonic" genes is to disrupt normal embryo development by mutation. Embryonic mutations have been identified in a wide range of flowering plants, including Arabidopsis (Muller 1963 ; Jurgens et al. 1991 ; Meinke 1991 ; Franzmann et al. 1995 ), maize (Sheridan and Neuffer 1982 ; Clark and Sheridan 1991 ), barley (Bosnes et al. 1987 ; Felker et al. 1987 ), carrot (Lo Schiavo et al. 1988 ), rice (Nagato et al. 1989 ), and pea (Bhattacharyya et al. 1990 ). Many of these mutations have been induced with chemical mutagens, x-irradiation, or T-DNA from Agrobacterium.
In this report, we describe the isolation and characterization of a recessive embryo-defective mutant of Arabidopsis, edd1 (for embryo-defective development), caused by insertion of a Dissociation (Ds) transposon. Development of mutant embryos diverges from normal embryo formation, because edd1 embryos do not establish normal bilateral symmetry and fail to complete their development. The EDD1 gene was isolated and shown to encode a protein with significant homology to the bacterial glycyl-tRNA synthetase (GlyRS) from Escherichia coli.
|
|
RESULTS |
|---|
Isolation of the edd1 Mutant
A two-component transposon system was used to induce insertional mutants in Arabidopsis (ecotype C24) by using an Accl /DsA transposition line. As described earlier (Altmann et al. 1992 , Altmann et al. 1995 ), this system is based on the maize transposable element Activator (Ac) and consists of an immobilized Ac element, Accl, and a nonautonomous Ds element, DsA. In the presence of Accl, DsA is mobilized. Descendants of the transposition line were screened for novel phenotypes. One mutant line that was identified showed a recessive seed-defective phenotype. This line harbored 25% aborted seeds randomly distributed throughout the siliques. Abnormal seeds are readily distinguishable as white seeds from wild-type green seeds in immature siliques. In mature siliques, affected seeds had a severly shrunken brownish appearance and dried out and died. Germination of mutant seeds was never observed. This type of embryo-defective mutant was named edd1, for embryo-defective development. In wild-type plants, seeds containing defective embryos were observed at much lower frequency (<1%).
Developmental Profiles of edd1 Embryos
We compared the phenotypes of wild-type and mutant embryos from siliques of edd1 plants at various stages of embryo development. As shown in Figure 1A, a wild-type globular-stage embryo contains a spherical embryo proper attached to a linear suspensor of seven to nine cells, including the hyphophysis at the base of the embryo proper. Due to local cell divisions within the apical embryo region, a triangular embryo proper is formed (Figure 1B). At this stage of development, the embryo proper displays three primary tissue layers—protoderm, ground meristem, and procambium. Cotyledon primordia arise at the flanks of the upper tier, producing a heart-shaped embryo (Figure 1C). At the same time, the hypophysis divides to form a region of cells from which the root apical system will develop. During the torpedo stage (Figure 1D), the cotyledons and hypocotyl elongate, and the suspensor begins to degenerate. Chlorophyll accumulation becomes visible throughout the embryo and the embryo sac. The mature embryo proper consists of two cotyledons adjacent to the hypocotyl, and the embryo occupies nearly the whole seed (data not shown).
To understand the role of the EDD1 gene, we tried to elucidate the gene's primary site and time of action during embryo development. edd1 embryos appeared normal through the early-globular embryo stage, and we could not distinguish between mutant and wild-type embryos. By contrast, edd1 embryos were readily identified within siliques that contained wild-type embryos at heart stage. At this time of silique development, edd1 embryos are retarded in their growth and do not display the bilateral symmetry of wild-type embryos (Figure 1E).
Microscopic analysis of isolated edd1 embryos from siliques that contained wild-type embryos at the torpedo stage revealed that enlargement of mutant embryos occurred but never reached wild-type levels and was not accompanied by normal cotyledon or axis differentiation. Individual mutants displayed variable phenotypes. Most edd1 embryos failed to develop a pair of cotyledons but showed an enlarged apical region of the embryo proper due to abnormal cell divisions within the embryo (Figure 1F). In rare cases, formation of one or two small apical bulges could be observed (Figure 1G and Figure 1H). Chlorophyll accumulation within mutant embryos or the surrounding embryo sac never became visible. The radial arrangement of tissues seemed to be unaffected: mutant embryos were bordered by a normally formed protoderm, and central vascular strands could be distinguished. Abnormal development was also detected within the suspensor region of mutant embryos. This region was enlarged in contrast to the simple linear cell file of wild-type embryos (Figure 1I, indicated by an arrowhead).
To determine the growth and differentiation potential of mutant embryos cultured in vitro, we removed edd1 embryos from immature seeds at various developmental stages and transferred them to hormone-free synthetic nutrient medium. Under these conditions, mutant embryos enlarged and formed small plantlets (Figure 1J). None of the cultured mutants was green, but some produced purple sectors. Despite a high frequency of regeneration, cultured plantlets failed to continue their growth and turned into calli. Fertile plants were not obtained. From this observation, we conclude that either the culture conditions were inappropriate for complete rescue of mutant embryos or the EDD1 gene product is essential for normal vegetative growth.
Isolation of the EDD1 Gene
In plants segregating for the edd1 phenotype, the presence of a novel DsA insertion not detectable in the original Accl/DsA line could be monitored using DNA gel blot analysis (data not shown). The transposed DsA element was separated from the Accl element by crossing the edd1 line to the untransformed wild type, and progeny plants were selected that lacked the Accl element (Figure 2A). Descendants of plants that carried the stabilized DsA element (DsA*) were tested for linkage between the transposed element and the edd1 mutation (Figure 2B). Among the 110 plants tested (76 edd1/+ plants and 34 wild-type +/+ plants), no recombination between the mutant locus and the inserted DsA element was observed. This strict linkage between the DsA insertion and the mutant locus strongly indicated that the DsA insertion caused the embryo-defective phenotype.
View larger version (56K):
[in this window]
[in a new window]
|
Figure 2.
Genomic DNA Gel Blot Analysis.
Total DNA from edd1 segregants displaying either the mutant (e) or wild-type (wt) phenotype was digested with EcoRI-HindIII and hybridized with a 32P-labeled 3' HindIII-BamHI fragment of Ac. Arrows indicate the positions of the corresponding 3' fragments of Accl and the transposed DsA* at the new locus.
(A) The edd1 mutant was crossed to the untransformed wild type, and F1 plants were tested for separation of Accl and DsA. edd1 segregants marked with an asterisk (e*) only contained the transposed DsA* element; the Accl as well as the DsA element in the original locus were no longer present in these plants.
(B) To demonstrate linkage between the novel DsA* locus and the mutant phenotype, we analyzed progeny plants of the edd1 mutant with the stabilized DsA* element for recombination events. Plants showing the mutant phenotype (e) all contained the DsA* element, whereas plants displaying wild-type characteristics (wt) did not.
|
|
To study the molecular and biochemical basis of the edd1 phenotype, we used the transposon sequence to screen a genomic library made from DNA isolated from heterozygous edd1/+ plants to isolate DNA adjacent to both termini of the DsA element in the mutant. A 3.5-kb EcoRI-SalI chromosomal fragment was subcloned; it contained the complete DsA element and 0.5 and 1.0 kb of flanking genomic sequences, respectively. The sequence of the flanking regions was determined from both directions, and the cloned fragments were used to screen wild-type Arabidopsis cDNA libraries, resulting in the isolation of several cDNA clones. The complete nucleotide sequence of the longest cDNA clone (p7/9) was determined. It comprises 3438 bp and has an open reading frame (ORF) extending from nucleotide 32 to nucleotide 3235, with the stop codon TGA and a putative polyadenylation signal, AATAAA (Joshi 1987 ), downstream of the translational stop. The nucleotide sequence was submitted to the EMBL database (accession number AJ003069).
Expression of the EDD1 gene was examined using RNA blot analysis of poly(A)-enriched RNA isolated from various Arabidopsis tissues. Isolated mRNA from leaves, buds, flowers, stems, and roots was hybridized using the complete Arabidopsis EDD1 cDNA as a probe. EDD1 transcripts are of low abundance (Figure 3A), and no tissue-specific expression of the examined gene could be monitored compared with the ACTIN1 gene control (Figure 3B; Nairn et al. 1988 ). Interestingly, a second transcript of ~0.9 kb that cross-hybridized with the EDD1 probe could be detected. This transcript was detected only when using the full-length cDNA clone as a hybridization probe. When we used the Ds flanking regions or a 5' cDNA fragment as a probe, this mRNA was not detected.
View larger version (46K):
[in this window]
[in a new window]
|
Figure 3.
Analysis of EDD1 Expression in Different Organs of Arabidopsis Wild-Type Plants.
RNA gel blot analysis of EDD1 expression is shown. Approximately 3 µg of poly(A)-enriched RNA from leaf, bud, open flower, stem, and root was loaded per lane. The blot was hybridized successively with the Arabidopsis EDD1 cDNA and the Arabidopsis ACTIN1 gene (Nairn et al. 1988 ). The arrows at right indicate the detected transcripts. (A) RNA gel blot hybridized with the EDD1 cDNA. In addition to the EDD1 transcripts, a second transcript of ~0.9 kb that cross-hybridizes with the EDD1 probe was detected. (B) Same blot as shown in (A) hybridized with the Arabidopsis ACTIN1 gene.
|
|
The EDD1 Gene Encodes GlyRS
The predicted EDD1-encoded protein is 1068 amino acids long, with an apparent molecular mass of ~117 kD. Homology searches using the BLAST service of the National Center for Biotechnology Information (Bethseda, MD) revealed that the Arabidopsis protein shares no significant similarity to any known eukaryotic protein. As shown in Figure 4A, significant homology was found between the EDD1-encoded protein and bacterial GlyRS from E. coli (Keng et al. 1982 ; Webster et al. 1983 ). GlyRS belongs to the family of aminoacyl-tRNA synthetases (AARSs), an essential but structurally diverse class of enzymes that catalyze the addition of amino acids to specific tRNA molecules. GlyRS specifically ligates the amino acid glycine to its cognate tRNA species. The E. coli enzyme is a tetramer of two subunits, 
2ß2 (Keng et al. 1982 ). A single mRNA encodes both subunits of the enzyme with the termination site for the ORF in frame with an initiation site for the ß ORF located nine nucleotides downstream of the translational stop codon. In contrast, only a single ORF was found for the plant homolog described in this study, where both subunits (named domain a and b for the EDD1-encoded protein) are translated into a single polypeptide chain.
View larger version (19K):
[in this window]
[in a new window]
|
Figure 4.
Molecular Analysis of the edd1 Locus.
(A) A comparison of two segments of the Arabidopsis EDD1 protein with the corresponding protein sequence from the E. coli GlyRS is shown. Asterisks indicate identical amino acids in both sequences. The dash indicates a stop codon within the bacterial sequence. The homology between EDD1 and the E. coli protein is most prominent in the N-terminal protein half (see upper alignment), ranging from 60 to 70% identical amino acids. Homology is lower in the C-terminal half, which corresponds to the bacterial ß subunit (see lower alignment). (B) Insertion site of the DsA element. The structure of the DsA-disrupted gene is shown in comparison with the edd1 cDNA. Black boxes represent exons, dashed lines show their position within the cDNA, and introns are indicated by white regions between exons. The DsA element inserted into an exon region that encodes parts of the N-terminal protein half with homology to the bacterial subunit. (C) Integration of DsA took place close to an exon/intron border region, as indicated by arrows. The 8-bp duplication is depicted in boxes. The duplicated sequence contains parts of the 3' consensus splice sequence of the neighboring intron region.
|
|
Sequence comparison between the full-length cDNA clone and the subcloned genomic fragment enabled us to discriminate coding and noncoding regions within the subcloned region and to determine the exact integration site of the transposed DsA element. As shown in Figure 4B, the DsA element integrated into an exon region of domain a. Upon integration of the DsA element, a duplication of 8 bp of plant DNA (GGTAT TGA) at the target locus took place. The 8-bp duplication is a characteristic typical of transposable elements belonging to the Ac/Ds family (Pohlmann et al. 1984 ). The target site itself contains parts of the 3' consensus splice site of the neighboring intron region (Figure 4C). The insertion site indicates that edd1 is a loss-of-function mutation and that this allele does not produce a functional protein.
For the isolation of the EDD1 promoter, we subcloned and analyzed a 4-kb BglII genomic fragment that cross-hybridized with a 5'-terminal probe of the isolated plant cDNA. This fragment contained 1294 bp of the promoter region and was shown to be extremely AT rich (Figure 5A). A putative TATA box could be located 107 or 135 nucleotides upstream of the cDNA start site. Homology searches comparing the Arabidopsis EDD1 promoter region with eukaryotic promoter databases revealed short stretches of homology to other plant promoters, such as the -phaseolin promoter from bean (Anthony et al. 1990 ) and the legumin legA promoter from pea (Lycett et al. 1984 ; see also Figure 5B). Both genes are predominantly expressed during embryogenesis in the cotyledons of these plants.
The EDD1-Encoded GlyRS Is Targeted to Chloroplasts
The similarity of the EDD1 protein to the bacterial enzymes extends almost throughout the whole ORF (except from the 61 amino acids at the extreme N terminus of the plant protein) and is most prominent in the N-terminal half of the protein, ranging from 60 to 70% identical amino acids. The C-terminal part of the protein shows a slightly lower degree of similarity (50% identical amino acids). The N-terminal end of the protein (amino acids 1 to 61) is characterized by a positive net charge with a pI of 12.48, a high percentage of serine (18.3%) and arginine (13.3%) residues, and a complete lack of the amino acids glutamate, aspartate, and tyrosine. These characteristics are indicative of a potential function of this N-terminal region as a transit peptide for import into plastids (von Heijne et al. 1989 ; von Heijne and Nishikawa 1991 ).
We were interested to determine whether the EDD1 protein is post-translationally imported into plastids and tested the import capacity of the in vitro–synthesized EDD1 protein into isolated pea chloroplasts. Coupled in vitro transcription and translation of the full-length EDD1 cDNA yielded only low levels of the 110-kD EDD1 protein, and no chloroplast import studies could be performed with the complete protein. To circumvent the problem of an insufficient amount of the EDD1 protein, we fused parts of the cDNA encoding the putative transit peptide to the uidA gene, encoding the 68-kD ß-glucuronidase (GUS) protein of E. coli (Jefferson 1987 ).
Translation of the chimeric gene yielded a single fusion protein of ~75 kD (Figure 6, lane 1). In the presence of intact pea chloroplasts under the appropriate import conditions, 70 to 80% of the added chimeric preprotein was found to be processed into a protein of ~65 to 70 kD (Figure 6, lane 2)—a size to be expected if the putative transit peptide were cleaved. The addition of the protease thermolysin to the reaction mixture digested the preprotein in the supernatant, whereas the mature protein imported into the chloroplasts was protected and remained intact (Figure 6, lane 3). The addition of a detergent destroyed the integrity of the plastid membrane. The truncated protein was no longer protected by the chloroplast membrane and was digested by the protease (Figure 6, lane 4). We conclude from these data that the N-terminal part of the EDD1 protein comprising the amino acids 1 to 61 is capable of mediating transport of the GUS protein into pea chloroplasts in vitro and that the isolated plant gene most probably encodes a plastidic GlyRS isoform.
View larger version (19K):
[in this window]
[in a new window]
|
Figure 6.
Plastidic Targeting of the Chimeric GUS Protein.
Pea chloroplasts were incubated with the 35S-labeled chimeric 75-kD GUS protein containing an N-terminal portion of the Arabidopsis GlyRS protein synthesized in a cell-free reticulocyte lysate. Lane 1 contains the in vitro–translated preprotein used in this import assay. Lane 2 contains the protein isolated from chloroplasts recovered after incubation with the radioactively labeled in vitro translation product. Lane 3 contains protein from chloroplasts incubated with the labeled preprotein and post-treated with the protease thermolysin. Lane 4 is as given for lane 3, except that chloroplasts were post-treated with thermolysin and the detergent Triton X-100. Protein molecular mass markers in kilodaltons are given at right. (+), presence; (-), absence; pre, preprotein; m, mature protein.
|
|
Somatic Reversion of the edd1 Mutation
One characteristic of transposon insertion mutants is their instability in the presence of a transposase (TPase) source. Under such conditions, wild-type revertants can occur (Bancroft et al. 1993 ; Chuck et al. 1993 ; Tacke et al. 1995 ). To induce excision of the DsA element from the edd1 locus and thus to generate wild-type sectors, we cross-fertilized the mutant line to a transgenic Arabidopsis line that contained a TPase gene under the control of the cauliflower mosaic virus (CaMV) 35S promoter. This TPase gene fusion was shown earlier to drive high frequencies of somatic and germinal DsA excision in Arabidopsis (Swinburne et al. 1992 ; Long et al. 1993 ). If the edd1 mutation had been caused by the insertion of DsA, instability of the mutant phenotype could occur as the result of DsA transactivation through the introduced TPase. We screened 240 F1 plants (heterozygous for the edd1 mutation) from this cross for somatic reversion events monitored as wild-type side shoots branching from a mutant main stem. Such a situation is expected to occur at low frequency because a fairly large revertant sector that does not cover the main stem must be generated. Furthermore, restoration of gene function most probably requires precise excision, with removal of the duplicated plant DNA sequences due to the position of DsA close to an exon/intron border. Accordingly, only a single variegated plant that showed a revertant sector could be recovered. Seeds from the revertant branch all grew into plants that did not segregate embryo-defective seeds.
Revertant plants were analyzed for the presence of the DsA element by DNA gel blot analysis. As shown in Figure 7, the DsA* present in edd1 was no longer detectable in the revertant progeny (R1 to R10). In these plants, the introduced TPase (Ac) as well as additional fragments (DsA+) potentially representing novel DsA insertion loci could be monitored. To confirm that the edd1 mutation was caused by the DsA insertion, the genomic regions around the termini of DsA* were amplified from genomic DNA of 30 individual progeny plants of the revertant sector and subsequently were analyzed. Polymerase chain reaction (PCR) amplicons that corresponded to the EDD1 gene were subcloned, and 15 of these PCR clones were sequenced. All of the 15 sequenced fragments showed the original wild-type sequence, indicating that DsA excision led to an exact restoration of the EDD1 gene. In addition to the revertant locus, a second PCR product was also amplified. Its sequence showed extensive homologies to the EDD1 gene, indicating the possible presence of a second related gene encoding another GlyRS isoform (data not shown).
View larger version (26K):
[in this window]
[in a new window]
|
Figure 7.
Somatic Reversion and Complementation of the edd1 Mutation.
Revertants were generated by crossing the mutant edd1 line to a transgenic line that contains a TPase gene under the control of the CaMV 35S RNA promoter (Swinburne et al. 1992 ; Long et al. 1993 ). One progeny plant of this cross contained a revertant (wild-type) sector. Seeds from that branch all grew into wild-type plants. Genomic DNA from those revertant plants was digested with EcoRI and HindIII and hybridized with a 32P-labeled 3' HindIII-BamHI fragment of Ac. Revertant plants (R1 to R10) do not contain the DsA* element at the position observed for the mutant line edd1 (e) but segregate for the introduced TPase gene (Ac) and for DsA at new locations (DsA+).
|
|
The chromosomal locations of edd1 and the homologous locus were determined by hybridization to the CIC yeast artificial chromosome (YAC) library (Creusot et al. 1995 ). The map position of the identified YACs was deduced (D. Bouchez, unpublished data), and edd1 was positioned on the lower arm of chromosome 3, close to the markers m409 and csr (ALS gene), at a map position that was ~82 centimorgans (cM), whereas the homologous locus was located on chromosome 5, close to the marker mi125, at a map position that was ~85 cM. Allelism tests were performed with the mutants emb69, emb117, emb149, emb154, and emb172 (kindly provided by D. Meinke, Oklahoma State University, Stillwater), which map to the same chromosomal region on chromosome 3 as does edd1. Our results demonstrate that the EDD1 locus had not been identified earlier by mutation and belongs to a separate complementation group.
Genetic Complementation of the edd1 Mutation
To demonstrate genetic complementation of the mutant edd1 by the isolated gene, we generated transgenic Arabidopsis lines that harbored the plant EDD1 cDNA under the control of the CaMV 35S promoter (Franck et al. 1980 ) or the USP promoter (for unknown seed protein) from fava bean (Baumlein et al. 1991 ; Fiedler et al. 1993 ), respectively. Both promoters differ in their stage and tissue specificity and should drive different expression levels of the cloned cDNA. The two transgenic lines, named 35S–EDD1 and USP–EDD1, were used separately to cross-fertilize heterozygous edd1 plants. In the case of genetic complementation, a reduction of aborted seeds (from 1 of 4 to 1 of 16) within siliques of F1 plants that inherited the edd1 mutation and the transgene was expected.
Introduction of the EDD1 cDNA into the edd1 genetic background under the control of the 35S promoter did not decrease the percentage of defective seeds. Progeny plants from the cross between heterozygous edd1 plants and the 35S–EDD1 transgenic line contained 100% normal seeds, 25% defective seeds, or >25% defective seeds, with seeds being aborted at different developmental stages in the latter case.
In contrast, a reduction in defective seeds was observed in the progeny of two independent lines that resulted from a cross between heteozygous edd1 plants and the USP–EDD1 transgenic line (see Table 1). The fraction of defective seeds varied during silique development. At an early developmental stage, approximately one of 16 of the seeds was defective, and only these aborted seeds contained mutant edd1 embryos (Figure 8B, left). Of the remaining green seeds, approximately three of 16 showed retarded embryo growth. Whereas most of the normal seeds contained bent cotyledon-stage embryos at that time of silique development (Figure 8B, center), such retarded embryos contained torpedo-stage embryos (Figure 8B, right). The segregation ratio is indicative of a situation in which the retarded growth behavior could represent homozygous edd1 embryos complemented by the introduced EDD1 cDNA copy. At a later developmental stage, an increase of defective seeds was observed. The percentage of seeds that contained mutant edd1 embryos still represented approximately one of 16 of the total seed number. In addition to this class of defective seeds, aborted seeds containing torpedo-stage embryos also appeared. They possibly represent homozygous (edd1/edd1) mutants with the transgene present.
View larger version (62K):
[in this window]
[in a new window]
|
Figure 8.
Genetic Complementation of the edd1 Mutation.
edd1 plants were cross-fertilized to the transgenic line USP–EDD1 containing the wild-type EDD1 gene under the control of the USP promoter from fava bean. (A) DNA gel blot analysis of wild-type and complemented plants. Total Arabidopsis DNA was digested with EcoRI and HindIII and hybridized with a 32P-labeled EDD1-specific probe. Several F1 progeny plants of the above-mentioned cross contained the introduced transgene (lanes 1 to 4). wt, wild-type plants; e, edd1 heterozygous plants. (B) Seeds derived from the above-described cross were dissected, and embryo development was studied. Approximately one of 16 of the total seed number contained defective embryos (left). Whereas most of the seeds contained bent cotyledon-stage embryos at that time of silique development (center), approximately three of 16 showed retarded embryo growth, with immature seeds containing torpedo-stage embryos (right).
Embryos are indicated by arrows.
|
|
View this table:
[in this window]
[in a new window]
|
Table 1.
Segregation of Embryo-Defective Seeds in F1 Plants Derived from Crosses between Heterozygous edd1 Plants and the USP–EDD1 Transgenic Line
|
|
Limited DNA gel blot analysis of progeny plants from the cross between the USP–EDD1 transgenic line and edd1 plants was performed using an edd1-specific probe for hybridization (Figure 8A). Arabidopsis wild-type plants were homozygous for the EDD1 allele, whereas edd1 plants (heterozygous plants giving rise to 25% aborted seeds) contained the wild-type as well as the mutant allele. Several progeny plants from the above-mentioned cross contained an additional locus corresponding to the introduced cDNA (Figure 8A, lanes 1 to 4). Plants that were heterozygous for the EDD1 locus (edd1/EDD1) but contained the introduced transgene (numbers 1, 3, and 4) showed a reduced percentage of aborted seeds. We conclude from these data that the introduced USP–EDD1 gene fusion is capable of partially complementing the edd1 mutation. The effect seems to be dependent on the developmental stage of the embryo and probably corresponds to the expression pattern of the USP promoter.
|
|
DISCUSSION |
|---|
In this article, we describe the recessive embryo-defective mutant edd1 of Arabidopsis. This mutant was isolated by using transposon mutagenesis. Genomic plant sequences flanking the transposon in mutant plants have been rescued and used to isolate a full-length cDNA clone representing the wild-type EDD1 gene.
Conclusive evidence that the EDD1 gene had been cloned was obtained by genetic reversion and complementation experiments. In the mutant, a gene encoding a GlyRS has been disrupted upon integration of the transposable element DsA. Reversion of the edd1 phenotype through mobilization of the transposon was shown to occur at low frequency after cross-fertilizing the mutant line to a transgenic Arabidopsis line that contains an active TPase gene fusion. In our experiments, only a single revertant for the DsA-induced edd1 mutant could be rescued, showing a precise excision of the transposon. Precise excision in the Ac/Ds system is rare but has been reported for a revertant of the Ds-induced alcohol dehydrogenase Adh1 mutant (Dennis et al. 1986 ) and the Ds-induced iojap mutant (Han et al. 1992 ). The low reversion frequency indicates that a precise restoration of the original wild-type sequence may be necessary to rebuild protein function. Genetic complementation of the mutant phenotype was achieved by introducing the Arabidopsis EDD1 cDNA under the control of the USP promoter from fava bean into the edd1 genetic background.
Analysis of the EDD1 predicted amino acid sequence revealed a significant homology of the plant protein to GlyRS of E. coli, which belongs to the AARS family (Keng et al. 1982 ; Webster et al. 1983 ). AARSs catalyze the first step in protein synthesis by attaching amino acids to the 3' end of the cognate tRNA in a two-step energy-dependent reaction sequence. At least 20 such enzymes, with one enzyme being specific for each amino acid found in proteins, constitute a minimum set for protein biosynthesis. The sequence homology between the deduced Arabidopsis protein and the protein from E. coli is extremely high, showing up to 70% identical amino acids. A similiar degree of sequence homology between bacterial and eukaryotic synthetase enzymes had not previously been documented because homologies between yeast and bacterial enzymes normally range from 20 to 50% in the portions that can be aligned (Schimmel 1987 ). This high level of similiarity leaves little doubt that the identified gene encodes a plant GlyRS. No significant homology of the plant protein to other so-far characterized eukaryotic GlyRSs from yeast (Kern et al. 1981 ), silkworm (Nada et al. 1993 ), or humans (Ge et al. 1994 ; Shiba et al. 1994 ) has been detected, indicating that the plant enzyme corresponding to the isolated gene shares a common evolutionary history with the enzyme from E. coli, which is distinct from that of eukaryotes.
Results derived from protein uptake experiments show that an N-terminal portion of the Arabidopsis GlyRS is capable of directing a marker protein into pea chloroplasts. We conclude from these data that the plant enzyme described here most probably acts within the plastidic compartment. Nevertheless, additional functions of the GlyRS in other cellular compartments cannot be entirely excluded. However, there is only little evidence for mistargeting or cotargeting of preproteins to different organelles (Hurt et al. 1986 ; Huang et al. 1990 ; Creissen et al. 1995 ). AARSs are essential components of the protein-synthesizing machinery and are found in each cellular compartment, which harbors genetic information and is able to express it. Plant cells contain three such compartments that perform protein synthesis, namely, the nucleoplasm, the mitochondria, and the plastids. Because none of the AARS enzymes has been shown to be encoded by the plant organellar genomes, they probably are nuclear encoded and imported from the cytosol.
Mutation of the edd1 locus does not impair early embryonic growth, and development of mutant embryos proceeds normally until the end of the globular embryo stage. Embryo growth is disturbed during transition to the heart-shaped stage. Limited elongation and enlargement of mutant embryos occur but are not accompanied by normal cotyledon and shoot meristem formation. Mutant embryos display variable phenotypes that are independent of their developmental stage, suggesting that penetrance of the mutation is incomplete. In most cases, cotyledons are completely missing, and only seldom can small apical bulges, possibly representing strongly reduced cotyledons, be detected around the apical region. Microscopic examination of mutant embryos suggested that formation of the radial tissue pattern comprising protoderm, ground meristem, and vascular tissue proceeds normally, indicating that cellular differentiation within the embryo proper occurs in the absence of normal morphogenesis. Growth of the suspensor is deregulated, yielding an enlarged suspensor instead of a single cell file. Abnormal suspensor development has been reported in earlier studies and seems to result directly from aberrant growth of the embryo proper (Akhundova et al. 1978 ; Clark and Sheridan 1991 ; Yeung and Meinke 1993 ; Schwartz et al. 1994 ; Vernon and Meinke 1994 ; Yadegari et al. 1994 ).
From the obtained sequence information, we deduced that the protein encoded by the EDD1 locus probably acts as a plastidic GlyRS. However, its function during plastidic translation does not provide a direct hint to its role during embryogenesis. The most simple explanation is that the EDD1 protein (GlyRS) is one of the components necessary for the biosynthesis of metabolic products that are of general importance for cellular functions. It cannot be excluded that a simple metabolite, such as glucose or a group of substances normally produced by the plastids, becomes limiting during embryogenesis due to the loss of GlyRS function, resulting in an arrest of embryo growth. The observation that mutant embryos can be partially rescued if grown in vitro on nutritive media may suggest that the defect is due in part to such a basic lack of nutrients. However, the medium cannot completely rescue the edd1 mutant. This rather contradicts a pure metabolic deficiency causing the developmental arrest of edd1 embryos and indicates that additional mechanisms are probably affected in the mutant. A model that we propose implies that the edd1 mutation interferes with the exchange of signals between the plastid and the nucleus, which are necessary for normal embryogenesis.
Biogenesis of plastids is tightly coupled with temporal and spatial stages of plant development. Thus, a key component of plastid development is the coordination of gene expression between the nuclear and plastidic genomes that are separated by several membranes and are vastly different in complexity, structure, gene organization, and regulatory mechanisms. Several lines of evidence indicate that the developmental status of the plastid is determined by the same developmental program that controls the cell in which it resides (Taylor 1989 ). Genetic information in the nucleus therefore ultimately regulates the expression of the plastid genome as well as the expression of nuclear genes encoding plastid proteins. However, little is known about the molecular nature of the genes that regulate chloroplast development or the way this regulation is achieved. Only a few mutations in nuclear genes that regulate chloroplast development have been subjected to molecular analysis. These mutants are affected in a variety of chloroplast functions, including protein targeting to the thylakoid (hcf106 of maize; Martienssen et al. 1989 ; Voelker and Barkan 1995 ), maintenance of ribosomes within the plastid (iojap of maize; Walbot and Coe 1979 ; Han et al. 1992 ), chloroplast division (arc mutants of Arabidopsis; Pyke and Leech 1992 , Pyke and Leech 1994 ; Robertson et al. 1995 ), and communication between chloroplast and mitochondrion (chm of Arabidopsis; Martinez-Zapater et al. 1992 ).
Furthermore, the developmental stage of the chloroplast itself appears to regulate the expression of nuclear genes coding for chloroplast-destined proteins (Harpster et al. 1984 ; Mayfield and Taylor 1984 ; Taylor 1989 ). For example, in photooxidative mutants of maize or in a variety of plants in which chloroplast development is arrested with an inhibitor, light-regulated nuclear genes such as CAB (encoding the chlorophyll a/b binding proteins of photosystem II) or RBCS (encoding the small subunit of ribulose 1,5-bisphosphate carboxylase) are not expressed. Thus, it has been hypothesized that signals from chloroplasts regulate nuclear gene expression (Oelmuller 1989 ; Taylor 1989 ). A signal from the plastid, known as the plastid factor, has long been postulated to provide the nuclear genome with information about the developmental status of the plastid in each cell (Bradbeer et al. 1979 ; reviewed in Taylor 1989 ). The plastid factor is proposed to coordinate gene expression in the nucleus with the needs of the chloroplast.
A block of protein synthesis within the plastids resulting from a lack of GlyRS activity is expected to interfere with this complex signal exchange program. The dependence of the signal generation on plastid mRNA translation could be based either on a malfunction within the plastid sensed by the signal generating machinery or on the block of biosynthetic pathways that involve plastid-encoded activities. The latter may affect the synthesis of the hypothesized plastid signal itself.
Plastid differentiation in plant embryos has been investigated by Schulz and Jensen 1969 . They demonstrated that plastids within the ground meristem and protoderm tend to form internal lamellae, which is reminiscent of developing thylakoids, early during embryogenesis. These morphological alterations coincide with the appearance of transcripts for two chloroplast proteins: immature Capsella seeds already express the nuclear-encoded RBCS gene as well as the plastid-encoded PSBA gene (QB protein of photosystem II). Transcripts are uniformly distributed throughout the torpedo-stage embryo, with the exception of procambial cells (Fiebig et al. 1990 ). The expression of the plastidic AARS enzymes and their tRNA substrates is linked to plastid differentiation and is partly under developmental and light-regulated control. Exposure of dark-grown Euglena cells to light results in a marked increase in the level of plastidic AARS and parallels chloroplast biogenesis (Reger et al. 1970 ; Parthier et al. 1972 ; Reinbothe and Parthier 1990 ). In cotton, AARSs show a four- to 16-fold increase of expression in cotyledons during germination (Brantner and Dure 1975 ). In contrast to the Euglena system, this increase does not require an exposure to light in cotton but appears to be a preprogrammed aspect of the germination process.
The effect of the edd1 mutation first becomes visible during the transition phase. Mutant embryos do not display normal cotyledon initials but develop an abnormal embryo structure without chlorophyll accumulation. This suggests that edd1 acts at an early stage of chloroplast development, before chlorophyll biosynthesis is activated. Because the edd1-encoded GlyRS is a plastid-localized protein and chloroplast function is obviously impaired in the mutant, we hypothesize that the developing chloroplast creates a signal required for cotyledon morphogenesis and that edd1 plants are defective in the production or transmission of that signal. A different mutation that most probably also affects protein synthesis in plastids and leads to an arrest of embryogenesis has been described by Tsugeki et al. 1996 . In this mutant, a transposon insertion in the Arabidopsis SSR16 gene, which is highly homologous to the gene encoding the S16 ribosomal protein from Neurospora crassa, causes an embryo-defective lethal mutation.
Coupling between plastidic and cellular differentiation has been demonstrated through mutations in several loci that are involved in both chloroplast development and palisade parenchyme morphogenesis, for example, in bean (Zaumeyer 1938 , Zaumeyer 1942 ; Wade 1941 ) and Arabidopsis (Kirk and Tilney-Basset 1978 ; Wetzel et al. 1994 ). Furthermore, mutation of DCL (for defective chloroplasts and leaves) or DAG (for differentiation and greening) loci of Antirrhinum blocks the development of chloroplasts and interferes with palisade cell formation (Chatterjee et al. 1996 ; Keddie et al. 1996 ).
The DAG as well as the DCL loci encode proteins with unknown function. These proteins are targeted to the plastids. In the affected tissue, the palisade layer is missing and replaced with smaller spherical-shaped cells similiar to those observed in spongy mesophyll. Plastids from the mutant tissue are small, contain very few or no thylakoids, and resemble proplastids. Mesophyll cells of the Arabidopsis chloroplast division mutant arc6 contain two to six enlarged chloroplasts instead of ~100 normal-sized chloroplasts (Robertson et al. 1995 ). Although the internal structures of the enlarged arc6 chloroplasts are not radically different from normal granal and thylakoid structures, the shape of mesophyll cells is different in arc6 leaves, leading to an altered leaf morphology in mutant plants. The arc6 mutation probably affects proplastid division.
The effect of dag, dcl, and arc6 on leaf morphology is different from that of many other genes affecting chloroplast development, such as olive of Antirrhinum (Hudson et al. 1993 ), hcf106 of maize (Miles 1994 ), and albina and xantha of barley (Henningsen et al. 1993 ). These mutants have partially developed but defective chloroplasts without having any morphogenetic effects on palisade development. This could imply that only signals generated early in the development of chloroplasts are necessary for stimulating the final divisions and expansions of cells of the palisade mesophyll.
In our model (see Figure 9), EDD1 is involved in the production of a very early signal that is required to initiate the first step in the differentiation of proplastids, which in turn triggers cotyledon morphogenesis. DCL, DAG, and ARC6 are required at a later time for the development of the mesophyll tissue. In contrast, OLIVE, HCF106, ALBINA, and XANTHA act late during chloroplast development without affecting cell and tissue morphogenesis.
View larger version (37K):
[in this window]
[in a new window]
|
Figure 9.
Model of EDD1 Gene Action during Embryogenesis.
EDD1 function is required early during chloroplast biogenesis for the expression of plastidic genes. The edd1 mutation affects the synthesis of the chloroplast signal A!. In this model, A! (e.g., a plastid-derived metabolite) serves as the primary signal in a simple signal transduction pathway that activates the expression of nuclear genes involved in cotyledon formation. ARC6, DAG, and DCL act downstream of EDD1. Their activities are required for the development of the thylakoid membrane system and for the synthesis of signal B!, which in turn triggers palisade cell differentiation and controls leaf morphology. OLIVE, HCF106, ALBINA, and XANTHA are also involved in chloroplast development. A mutation in any of these genes leads to defective chloroplasts, but there is no morphogenetic effect. Therefore, we assume that they act downstream of EDD1, ARC6, DAG, and DCL and do not take part in the above-described signal transduction pathway that connects the nucleus and the plastids.
|
|
Molecular studies of RNAs present in developing seeds indicate that 20,000 genes are expressed during embryogenesis in higher plants (Goldberg et al. 1989 ). Although it is reasonable to assume that many of these genes encode basic cellular functions, it was a surprising result that even in the case of mutants classified on the basis of histological analysis as "pattern mutants," the corresponding genes encode rather general functions (Shevell et al. 1994 ; Meinke 1995 ; Busch et al. 1996 ; Lukowitz et al. 1996 ).
The results obtained from the analysis of edd1 as described here represent another example in this respect and emphasize the fact that the function of genes such as EDD1 cannot be dismissed as being merely housekeeping but should be considered essential to the whole process of embryogenesis.
|
|
METHODS |
|---|
Plants and Growth Conditions
The embryo-defective Arabidopsis thaliana mutant edd1 (for embryo-defective development; former designation No. 240) described in this report was initially identified through Activator/Dissociation (Ac/Ds) transposon mutagenesis (Altmann et al. 1995 ). Recessive embryo mutants were maintained as heterozygotes that produced 25% mutant embryos upon self-pollination. All Arabidopsis plants were grown in soil under a 16-hr-day (3000 lux fluorescent light, 20°C) and 8-hr-night (17°C) regime at a relative humidity of 70%. For crosses, flowers of the female parent were emasculated ~2 days before anthesis and cross-pollinated 2 or 3 days later.
Transformation of Arabidopsis and Growth in Aseptic Culture
The Arabidopsis ecotype C24 was used for Agrobacterium tumefaciens–mediated gene transfer. Agrobacterium strain C58C1 with the plasmid pGV2260 (Deblaere et al. 1985 ) was used for transformation. Cotyledon infection, selection, regeneration of transformed plants, and growth of the plants to maturity were performed as described by Schmidt and Willmitzer 1988 .
Nomarski Optics of Cleared Seeds
Seeds were removed from siliques and cleared for 30 min to 24 hr in a lactophenol solution (glycerol–lactate–phenol–H2O, 1:1:2:1 [v/v]) on a microscope slide. Seeds of later developmental stages and defective seeds of edd1 plants required extended clearing periods. Cleared seeds were examined using a Zeiss Axiophot (Carl Zeiss, Inc., Oberkochen, Germany) microscope equipped with Nomarski optics. Photographs were taken using Kodak Ektachrome 320T film.
Tissue Culture of Mutant Embryos
Siliques from heterozygous edd1 plants were surface-sterilized by soaking in 1.5% sodium hypochlorite and 0.1% Tween 80 for 10 min. After several rinses in sterile water, embryos were removed from mutant seeds by using fine forceps and cultured on Gamborg's B5 basal medium (Gamborg et al. 1968 ), supplemented with 10% sucrose and 0.8% agarose. Sucrose concentration was decreased stepwise to 3% during the cultivation procedure.
Construction of Escherichia coli and Agrobacterium Plasmids
DNA manipulations of plasmid DNA were performed according to Sambrook et al. 1989 . Transformation of bacterial cells was performed using the Bio-Rad gene pulser, according to the manufacturer's protocol. To obtain the plasmid pUU1, a 3.5-kb EcoRI-SalI chromosomal fragment was ligated into an EcoRI-SalI predigested pBluescript SK+ vector (Stratagene, La Jolla, CA). For construction of the chimeric uidA gene for the chloroplast import assay, the 5' end of the EDD1 cDNA was amplified by polymerase chain reaction (PCR), using the primers SP5/1N (5'-CCGGCAGCTGGTCACACTCATGGC-3') and SP3/2N (5'-CCGGCAGCTGGAGGCGT TGAA-TCG-3'), corresponding to the 5' and 3' ends of the putative transit peptide, respectively. PCR products of the reaction were cloned into the PCR II vector (Invitrogen, San Diego, CA). The transit sequence was excised using PvuII and cloned upstream of the uidA gene in the vector pBGUS(3) (T. Mozo, unpublished data), which had been predigested with SmaI, to obtain plasmid pSP-BGUS3. pSP-BGUS3 was used for in vitro transcription and translation.
For the construction of the vectors p35S–EDD1 and pUSP–EDD1, which were used for complementation of the edd1 mutant, the EDD1 cDNA was cloned behind the two promoters as follows. For the construction of the vector p35S–EDD1, the EDD1 cDNA was excised from the cDNA clone p7/9 as an Asp718-SmaI fragment, filled in with the Klenow fragment of DNA polymerase I, and ligated in the sense orientation downstream of the 35S RNA promoter in a modified version of the binary vector pBinAR (Hofgen and Willmitzer 1990 ), which had been predigested with XbaI and filled in with the Klenow fragment. For the generation of the plant transformation vector pUSP–EDD1, the cDNA insert was treated in the same way but ligated in the sense orientation downstream of the USP promoter in the plant transformation vector pBinUSP (C. Benning, unpublished results) predigested with XbaI and filled in with the Klenow fragment.
Isolation and Analysis of Plant DNA and RNA
Genomic DNA was isolated as described by Dellaporta et al. 1983 by the addition of a cetyl trimethylammonium bromide precipitation and digested with restriction enzymes. After electrophoretic separation on 1% agarose gels, depurination, and denaturation (in 0.5 M NaOH and 1.5 M NaCl), the DNA was blotted onto BiodyneB membranes (Pall, Portsmouth, UK) by using capillary transfer with 20 x SSC (1 x SSC is 0.15 M NaCl and 0.015 M sodium citrate; Sambrook et al. 1989 ). Hybridization with radioactively labeled DNA fragments (Feinberg and Vogelstein 1983 ) was performed at 65°C in 0.25 M sodium phosphate buffer, pH 7.2, 7% SDS, 1% BSA, and 1 mM EDTA. Total RNA was obtained from frozen plant tissue by using the procedure of Logemann et al. 1987 . Poly(A)+ RNA was isolated from total RNA, according the manufacturer's protocol, using the Oligotex suspension (Qiagen, Hilden, Germany). Approximately 3 µg of poly(A)+ RNA was fractionated in a 1% formaldehyde–agarose gel. Hybridizations and washings were done at high stringency, according to standard procedures, using 32P-radiolabeled probes (Church and Gilbert 1984 ). For the control hybridization, the ACTIN1 gene was radioactively labeled (Nairn et al. 1988 ).
Isolation of the EDD1 Gene
Genomic DNA was isolated from soil-grown edd1 plants, partially digested with Sau3A, and ligated with the EMBL3 
vector predigested with BamHI (Stratagene). The ligated DNA was packaged in vitro using GigaPack Gold II (Stratagene). Bacterial strains K803 or SRB (Stratagene) infected with the recombinant phages were plated on NZCYM media (Sambrook et al. 1989 ). Approximately 250,000 phages were screened under stringent hybridization conditions, according to Sambrook et al. 1989 . The 1.2-kb HindIII-BamHI fragment of pKU-3 (Baker et al. 1987 ), representing the 3' Ac fragment, was used as hybridization probe. Probes were radiolabeled using the random primed labeling kit (Boehringer Mannheim, Germany). A 3.5-kb EcoRI-SalI chromosomal fragment of one positive clone that hybridized to the 3' Ac probe was subcloned into pBluescript SK+ (Stratagene; clone pUU1) and used for sequence analysis and for subsequent hybridizations.
For PCR detection of the transposon footprint, DNA was extracted from leaves of revertants, as described by Edwards et al. 1991 . The PCR reaction was performed with synthetic primers, TrapS3 (5'-CAGT TGTCT TCTACAAACCG-3') and TrapS5 (5'-GACACACTC-AAT T TCAAGTC-3') (TIB MOLBIOL, Berlin, Germany), corresponding to the 5' and 3' flanking regions of the DsA. Amplified fragments were separated on an agarose gel, cut out, purified using the QIAquick gel extraction kit (Qiagen), and ligated directly into the cloning vector PCR II for sequence analysis.
EDD1 cDNA Cloning
A 3.5-kb EcoRI-SalI fragment from pUU1 was used as a probe to screen an Arabidospis C24 ZAP cDNA library. This library was prepared from bud, flower, and silique mRNA by using the ZAP II cDNA synthesis kit (Stratagene). Plaques (5 x 105 plaque-forming units) were screened under stringent conditions. Positive cDNAs were isolated and subcloned into pBluescript SK+ by in vivo excision, according to the manufacturer's protocol. The longest cDNA clone (clone p7/9), comprising 3448 bp, was used for sequencing. The genomic fragments described earlier as well as the full-length cDNA clone were sequenced on both strands by using synthetic oligonucleotides (TIB MOLBIOL). Sequence reactions were performed using 35S-dATP and the T7 sequencing kit (Pharmacia, Bromma, Sweden), according to the manufacturer's protocol. Sequence analyses were performed using the GCG sequence analysis software package (Genetics Computer Group Inc., Madison, WI) and the BLAST Network service of the National Center of Biotechnology Information (Bethesda, MD). Comparison between the EDD1 protein from Arabidopsis and the homologous protein from E. coli was done with the Pile-up and Pretty box programs (Genetics Computer Group Inc.).
Mapping the edd1 Mutation
To determine the genetic position of the edd1 locus on the Arabidopsis genome, we screened the CIC yeast artificial chromosome (YAC) library (Creusot et al. 1995 ) by using the complete edd1 cDNA clone as a hybridization probe. Several YAC clones falling into two classes cross-hybridized with the probe. The clones CIC3B10, CIC3D2, CIC11D12, and CIC11G6 showed strong hybridization signal, whereas the YAC clones CIC3A4, CIC6E4, CIC5F5, CIC9D5, CIC10B9, and CIC12F8 cross-hybridized only weakly. All of these YAC clones were previously mapped to two different regions of the Arabidopsis genome. The clones CIC3B10, CIC3D2, CIC11D12, and CIC11G6 are positioned on chromosome 3, close to the markers m409 and csr (ALS gene), at a map position that is ~82 centimorgans (cM). All of the other clones are anchored on chromosome 5, close to the marker mi125, at a map position that is ~85 cM (Schmidt et al. 1997 ; Camilleri et al. 1998 ). To characterize both loci, we amplified isolated YAC DNA from the above-mentioned clones by using the primers utilized earlier to distinguish the EDD1 gene and the additional homolog. PCR amplification of the clones CIC3B10, CIC3D2, CIC11D12, and CIC11G6 yielded a 360-bp fragment representing the EDD1 locus, whereas PCR amplification of CIC3A4, CIC10B9, CIC6E4, CIC5F5, CIC12F8, and CIC9D5 produced a 380-bp band representing the locus homologous to EDD1 mentioned earlier. From these data, we conclude that edd1 is positioned on the lower arm of chromosome 3.
In Vitro Transcription and Translation
In vitro transcription and translation were performed according to the supplier's instruction by using either the (TNT)-coupled reticulocyte lysate system or the TNT-coupled wheat germ extract system (Promega, Madison, WI) in the presence of 35S-methionine (Amersham, Braunschweig, Germany).
Import into Isolated Pea Chloroplasts
Intact pea chloroplasts were isolated from pea seedling (7 to 10 days old), as described by Bartlett et al. 1982 . Import reactions were performed in the light for 30 min at 25°C in a total volume of 240 µL containing 20 µL of the in vitro translation product, chloroplasts corresponding to 80 µg of chlorophyll, 12 µL of 1 M Hepes/KOH, pH 8.0, 32 µL of 2 M sorbitol, 18 µL of 250 mM methionine, and 2.4 µL of 0.1 M Mg-ATP. Chloroplasts were pelleted and resuspended in 600 µL of 50 mM Hepes and 0.33 M sorbitol, pH 8.0. One-third of the chloroplast fraction was treated with 2.5 µL of thermolysin (10 mg/mL) and 10 µL of 0.1 M CaCl2 for 20 min on ice, whereas the remaining chloroplasts were incubated on ice for the same period of time. Protease treatment was stopped by the addition of 10 µL of 0.5 M EDTA. Chloroplasts of a second fraction that were preincubated on ice were disrupted by adjusting the reaction mixture to 1% Triton X-100. Unbroken chloroplasts were reisolated through a 45% Percoll cushion by centrifugation at 4500g for 8 min, washed in 50 mM Hepes and 0.33 M sorbitol, pH 8.0, and resuspended in 1 x SDS sample buffer. The proteins were analyzed by electrophoresis (Laemmli 1970 ) followed by fluorography with Amplify (Amersham).
|
|
FOOTNOTES |
|---|
1 Current address: Max-Planck-Institut für Molekulare Pflanzenphysiologie, Karl-Liebknecht-Strasse 25, 14476 Golm, Germany. 
|
|
ACKNOWLEDGMENTS |
|---|
We thank James Lloyd for critically reading the manuscript and Carola Recknagel for technical assistance.
Received December 29, 1997; accepted June 16, 1998.
|
|
REFERENCES |
|---|
ABSTRACT
INTRODUCTION
