SEARCH

SEARCH BY CITATION

Keywords:

  • ubiquitin;
  • HECT-ubiquitin ligase;
  • trichome;
  • endoreplication;
  • gibberellin;
  • Arabidopsis

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary Material
  9. References
  10. Supporting Information

Attachment of one or more ubiquitins (Ubs) to various intracellular proteins has a number of roles in plants including the selective removal of regulatory proteins by the 26S proteasome. The final step in this modification is performed by ubiquitin-protein ligases (E3s) that promote Ub transfer to appropriate targets. One important family of E3s is defined by the presence of a HECT domain, an active site first found at the C-terminus of the human E3 (E6-AP). Using a consensus HECT domain as the query, we identified a family of seven HECT-containing ubiquitin-protein ligases (UPL1–UPL7) in Arabidopsis thaliana that can be grouped into four subfamilies. The UPL3 and UPL4 subfamily encodes approximately 200-kDa proteins with four Armadillo repeats similar to those in the nuclear pore protein importin-α, suggesting that these E3s identify their targets through binding to nuclear localization sequences. Although T-DNA disruptions of the UPL3 locus do not affect overall growth and development of Arabidopsis, the mutants show aberrant trichome morphology. Instead of developing three branches, many upl3 trichomes contain five or more branches. The upl3 trichomes also often undergo an additional round of endoreplication resulting in enlarged nuclei with ploidy levels of up to 64C. upl3 plants are hypersensitive to gibberellic acid-3 (GA3), consistent with the role of gibberellins in trichome development. The phenotype of upl3 mutants is similar to that of kaktus, a previously described set of trichome mutants with supernumerary branches. Genetic analyses confirmed that upl3 mutants and kaktus-2 are allelic with kaktus-2 plants harboring a splice-site mutation within the UPL3-transcribed region. Collectively, the data indicate that the ubiquitination of one or more activator proteins by UPL3 is necessary to repress excess branching and endoreplication of Arabidopsis trichomes.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary Material
  9. References
  10. Supporting Information

Plants use a repertoire of mechanisms to modulate the levels of critical regulatory proteins that control development and physiology and promote survival in adverse environments. One such mechanism is the selective degradation of proteins via the ubiquitin (Ub)/26S proteasome proteolytic pathway. In this pathway, proteins committed for degradation are first modified with a chain of multiple Ubs. Ubiquitinated proteins are then recognized and degraded by the 26S proteasome, a multicatalytic protease complex that degrades the target but releases the Ub moieties intact (Hershko and Ciechanover, 1998; Pickart, 2001). Remarkably, recent analysis of the Arabidopsis thaliana genome showed that over 1300 loci (or 5% of the proteome) encode components of the Ub/26S proteasome pathway, making it one of the most elaborate regulatory mechanisms in plants (Vierstra, 2003). This importance is also supported by emerging genetic studies that have connected individual Ub/26S proteasome elements to nearly all aspects of a plant's life cycle, including cell division, embryogenesis, photomorphogenesis, hormone responses, disease resistance, and senescence, to name a few (Hellmann and Estelle, 2002; Vierstra, 2003).

Substrates of the Ub/26S proteasome pathway include abnormal proteins and a multitude of naturally short-lived normal proteins, which, in Arabidopsis, could include thousands of different polypeptides (Vierstra, 2003). These targets enter the pathway via an ATP-dependent conjugation cascade, involving the sequential action of E1s, E2s, and E3s (Ub-activating, -conjugating, and -ligating enzymes, respectively). In the final step, an E3 recruits both the target protein and an Ub-E2 intermediate and then stimulates ligation of the C-terminal glycyl carboxyl group of the Ub moiety to free lysyl ε-amino groups in the target (Hershko and Ciechanover, 1998; Pickart, 2001). Through reiterative rounds of conjugation, a poly-Ub chain is ultimately assembled, using one of the several lysines in Ub to form subsequent isopeptide bonds. Because they are responsible for selecting the myriad of appropriate targets, it is not surprising that the families of E3s are the most diverse components in the pathway. The Arabidopsis genome alone encodes at least 1100 distinct E3 components with most of these involved in target binding (Gagne et al., 2002; Kosarev et al., 2002; Vierstra, 2003).

One important class of Ub-protein ligases is the HECT E3s. These single polypeptides were first identified by the presence of a conserved 350-amino acid domain called the HECT domain, based on its homology to the C-terminus of human E6-Associated Protein (E6-AP), a ligase that assists in the degradation of p53 (Huibregtse et al., 1995). They are unique among E3 types in that they directly participate in Ub transfer, using a conserved cysteine in the HECT domain to form a Ub-E3 thiol-ester intermediate that serves as the proximal Ub donor (Scheffner et al., 1995). Comparative X-ray crystallography of the HECT domains from E6-AP and another human HECT E3, WWP1/AIP5, revealed a lobed structure; the 100-residue C-terminal lobe contains the active-site cysteine and the 250-residue N-terminal lobe contains the binding pocket for the E2-Ub intermediate (Huang et al., 1999; Verdecia et al., 2003). A hinge between the lobes is predicted to allow the active site to flex toward the E2 and presumably then toward the target to promote Ub transfer. Upstream of the HECT domain are often one or more interaction motifs (e.g. Armadillo, C2, WW (Trp-Trp), Poly(A)-Binding Protein domain (Poly(A), Filamen-type immunoglobulin domain (IG_FLMN), Regulator of Chromosome Condensation (RCC1), Ub-associated (UBA), Ub-interacting motif (UIM), Ub-like (UBL), and Ankyrin (Bates and Vierstra, 1999; Harvey and Kumar, 1999; Letunic et al., 2002; Mitsui et al., 1999; Wang et al., 1999)). Some likely participate in target recognition, whereas others may be important for localization, regulation, and/or Ub-binding.

Most eukaryotes contain multiple HECT-E3s (Schwarz et al., 1998). For example, the yeast (Saccharomyces cerevisae) genome encodes 5 HECT E3s, Schizosaccharomyces pombe 7, Drosophila at least 13, and the human genome potentially encodes over 50 (Schwarz et al., 1998). Some are able to ubiquitinate more than one target. For example, human E6-AP not only targets p53 (Huibregtse et al., 1995), but it also helps degrade the DNA repair protein HHR23A (Kumar et al., 1999), and is affected in the hereditary neuronal disease Angelman Syndrome (Matsuura et al., 1997; Nawaz et al., 1999). Yeast RSP5 targets the uracil, GAP1, and FUR4 permeases (Hein et al., 1995), the ZRT1 zinc transporter (Gitan and Eide, 2000), and the large subunit of RNA Polymerase II for ubiquitination (Huibregtse et al., 1997) through interactions between its WW domains and a proline-rich PPxY motif in the targets (Harty et al., 2000).

Searches of the various plant genomes indicated that plants also express multiple HECT E3s. Previously, we described UPL1 and UPL2 as two highly similar 405-kDa HECT E3s from Arabidopsis and demonstrated their ligase activity in vitro (Bates and Vierstra, 1999). Here, we searched the near-complete Arabidopsis genome sequence for HECT domain-containing proteins and discovered five new HECT E3s named UPL3–UPL7 that can be grouped by structure into three subfamilies (UPL3/4, UPL5, and UPL6/7). The presence of a variety of domains upstream of the HECT domain suggests that individual members of the UPL1–UPL7 family have distinct sets of targets and functions.

To help define the range of functions under HECT E3 control in Arabidopsis, we assembled a library of upl1–upl7 mutants that were generated using Agrobacterium tumefaciens T-DNA as an insertional mutagen (Sessions et al., 2002). From analysis of a collection disrupting UPL3 in particular, we show here that this Ub ligase is involved in trichome development. Trichomes are specialized epidermal cells that project from the leaf surface and provide a range of benefits, which include acting as a physical barrier to insect attack, evaporation and excess light, and as a site to synthesize and release volatile protectants (Szymanski, 2000). In Arabidopsis, they are formed early during leaf and stem development from the epidermal pavement through sequential steps of trichome initiation, which include three rounds of genome endoreplication, local outgrowth of the cell beyond the plane of the leaf blade, and emergence of the first branch. A final round of endoreplication and formation of a secondary branch generates the characteristic triradiate trichome architecture (Hulskamp et al., 1994; Szymanski, 2000; Szymanski et al., 2000).

Exhaustive genetic analyses have identified over 25 gene products required for correct trichome differentiation in Arabidopsis. These include GIBBERELLIC ACID-1 (GA1), involved in GA biosynthesis, SPINDLY (SPY), an O-linked N-acetyl glucosamine transferase required for repressing GA responsiveness (Olszewski et al., 2002; Perazza et al., 1999; Tseng et al., 2002), several presumed transcription factors (e.g. TRANSPARENT TESTA-GLABRA (TTG1), GLABROUS1 and 3 (GL1 and GL3; Szymanski et al., 2000), and components that regulate the microtubule cytoskeletal network (e.g. ZWICHEL (ZWI), FRAGIL FIBER-2 (FRA2), and ANGUSTAFOLIA (AN); Ilgenfritz et al., 2003). In the upl3 mutants, leaf trichomes initiate normally, but then acquire extra branches beyond the normal set of three and often undergo an additional round of endoreplication, indicating that the latter steps in trichome development are compromised. These findings suggest that the UPL3-directed ubiquitination and subsequent degradation of one or more positive regulatory proteins are essential to limit excess trichome cell growth and DNA replication.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary Material
  9. References
  10. Supporting Information

The UPL1–UPL7 gene family in Arabidopsis

In both yeast and animals, the family of HECT E3s comprise an important class of ubiquitination enzymes (Pickart, 2001). To define the complexity of HECT E3s in plants, we searched the Arabidopsis DNA databases for related proteins using the HECT domain from Arabidopsis UPL1 and human E6-AP as queries (Bates and Vierstra, 1999; Huibregtse et al., 1995). This search identified a number of cDNA and genomic sequences that were condensed to seven unique loci designated UPL1–UPL7, that included UPL1 and UPL2, which were described previously (Figure 1a; Bates and Vierstra, 1999). All the encoded proteins contained an approximately 350-amino acid HECT domain at the C-terminal end that included a highly conserved region surrounding the positionally invariant cysteine required to catalyze Ub transfer (Figure 1b). The UPL proteins could be grouped into four subfamilies (UPL1/2, UPL3/4, UPL5, and UPL6/7) based on intron/exon positions of the corresponding genes, protein sequence and length, and the presence of additional protein motifs upstream of the HECT domain as predicted by the smart and pfam databases (Figure 1a, data not shown). All seven UPL genes are expressed and thus are likely to be functional. Expressed sequence tags (ESTs) were found for UPL1–UPL6 (http://www.arabidopsis.org). Subsequently, we were able to detect the transcripts of all seven by RT-PCR of total seedling RNA (Bates and Vierstra, 1999; Figure 3a, data not shown).

image

Figure 1. Structure and organization of the HECT E3 protein family (UPLs) in Arabidopsis.

(a) Protein organization of the UPLs. Position of the signature HECT domain and other motifs potentially important for UPL function are shown. UBA, Ub-associated; UIM, Ub-interacting motif, UBL, Ub-like. The amino acid sequence length and percent sequence similarity of related members are indicated on the right.

(b) Alignment of the HECT domain surrounding the active-site cysteine of the UPL1–UPL7 family with human E6-AP. Identical and similar amino acids are shown in black and grey boxes, respectively. The active-site cysteine and the end of the protein sequences are identified by the arrowhead and asterisks, respectively.

(c) Alignment of the four Armadillo repeats of UPL3 and UPL4 with the four Armadillo repeats of an Arabidopsis importin-α (accession number At1g09270). The asterisks in the consensus sequence identify amino acid positions often conserved among Armadillo repeats from diverse eukaryotic proteins (Szurek et al., 2001).

Download figure to PowerPoint

image

Figure 3. Identification of T-DNA mutants affecting the expression of UPL3.

(a) Genomic organization of UPL3 and the positions of the T-DNA insertions. Exons and introns are indicated in boxes and lines, respectively. The coding regions for the HECT domain and the Armadillo repeats are shown in the gray boxes. The C locates the codon for the active-site cysteine in the HECT domain. The putative untranslated regions are indicated by the white boxes. A comparison of the UPL3 sequence surrounding the 3' splice site of the sixth intron in wild-type Ler and the kak-2 mutant is shown (intron-6, lowercase; exon-7, uppercase). The G to A transition in the kak-2 mutant at nucleotide 2048 is in reverse type and the 3' splice-site consensuses are in gray.

(b) RT-PCR analysis of the upl3-1 and upl3-2 mutants. Total RNA was isolated from young wild-type (WT) and mutant seedlings and then reverse transcribed using primers #1 or #2 specific for the UPL3 gene, the positions of which are shown in panel (a). The RT products for UPL3 were then subjected to PCR using the #1 + #3 (RT#1) or #2 + #3 (RT#2) primer combinations. RT-PCR of PAE2 using the same RNA preparations was included as a control. A similar set of PCR reactions using cDNA or genomic DNA as the template is included in each panel for comparison.

(c) RT-PCR analysis of WT (Ler) and kak-2 total mRNA. Primers #4 and #5 flank introns 3 and 4 upstream of the kak-2 splice-site mutation. Primers #6 and #7 flank introns 6 and 7 and span the 3' splice site mutation (nucleotide 2408) in intron 6 (see (a)).

Download figure to PowerPoint

As described previously by Bates and Vierstra (1999), UPL1/2 are approximately 405-kDa proteins containing a UBA domain that is common among a variety of proteins involved in Ub metabolism, and thus may be important for Ub binding (Hofmann and Bucher, 1996). We also discovered a potential UIM downstream of the UBA domain (Figure 1a). This hydrophobic patch, first found in the 26S proteasome subunit RPN10 as a binding site for poly-Ub chains (Fu et al., 1998b) and subsequently detected in a variety of other proteins that interact with Ub (Hofmann and Falquet, 2001), provides UPL1/2 with two possible sites for Ub association.

For UPL3, the intron/exon boundaries and the full coding region were determined by combining a partial cDNA (AV529735.1), encompassing approximately 70% of the coding region and a 450-bp 3' untranslated region (UTR) ending in a poly(A) tail, with a RT-PCR product generated for the 5' end of the coding region. Comparison of this full-length cDNA sequence with the chromosomal sequence revealed 16 introns interrupting an 1888-amino acid (or 203 kDa) coding region (Figures 1a and 3a). Although the UPL3 exons could be detected in the Arabidopsis genomic DNA database (http://www.arabidopsis.org), the current annotation (1/10/2003) has the exons divided between two predicted open reading frames (At4g38610 and At4g38600), instead of merged into one complete gene. Twelve of the 16 introns were predicted correctly. Of the remaining four introns, two were mispredicted at both ends (eighth and fourteenth), one was mispredicted at the 3' end (ninth), and a new intron was identified where the At4g38610 and At4g38600 annotations were joined. In addition, we discovered that the second and third introns of At4g38600 are mispredicted and actually comprise part of the tenth exon.

Using our re-annotated UPL3 sequence, the likely gene and protein organizations of UPL4 were deduced. UPL4 encodes a protein with 54% amino acid sequence similarity to UPL3 with the main difference being the absence of a 225-residue region 650 amino acids from the C-terminus of UPL4. Both UPL3 and UPL4 are predicted by smart to contain a cluster of four Armadillo repeats (170–392) that could represent an important protein–protein interaction site (Figure 1a,c). These repeats and the approximately 50 amino acids upstream are most similar to the Armadillo-containing region of Arabidopsis importin-α (Merkle, 2001; Szurek et al., 2001). Importin-α (also known as karyopherin-α) and its yeast and animal relatives form part of the nuclear pore apparatus that helps shuttle proteins into and out of the nucleus (Merkle, 2001; Szurek et al., 2001). The Armadillo-repeat region of importin-α, in particular, is thought to serve as the receptor for nuclear localization sequences (NLS) (Merkle, 2001; Szurek et al., 2001), suggesting that the related domain in UPL3/4 could help identify nuclear target(s) via their NLS.

The gene organization of UPL5 was deduced from six partial cDNAs (AU528677.1, AU228803, AU237716, AV523091, AV528677, and AV564048 (http://www.arabidopsis.org)), intron/exon prediction programs, and the expected structure of the HECT domain. A stop codon immediately precedes the expected translation initiation codon in the 5' cDNA (AU528677.1), supporting the conclusion that the UPL5 locus synthesizes only an 873-amino acid protein. The 195-residue N-terminal region contains three stretches similar to the UBL domain, a motif with weak homology to Ub (Figure 1a). UBL domains, which are also typically located at the N-terminal end, are found among a diverse group of proteins including RAD23, Parkin, and DSK2, and presumably help these proteins associate with other components of the Ub system through interactions with Ub-binding domains such as UIM and UBA (Hofmann and Bucher, 1996; Hofmann and Falquet, 2001). Downstream of the UBL repeats is a region (residues 272–296) homologous to the C-type lectin-binding domain (100% match to the Prosite consensus). This region could promote interactions of UPL5 with sugars, suggesting that glycosylated proteins are potential substrates.

A full-length cDNA is available for UPL6, which was then used along with a partial EST for UPL7 to deduce the full coding sequence of UPL7. Although less related by amino acid sequence alignments (only 27% similar), the UPL6 and UPL7 proteins appear to comprise a subfamily based on the presence of a signature IQ calmodulin-binding site near their N-termini. Intriguing possibilities inferred by this site are that UPL6 and UPL7 (i) ubiquitinate calmodulin directly, (ii) modify calmodulin-interacting proteins using calmodulin as an adaptor, and/or (iii) are regulated by internal Ca2+ through reversible calmodulin binding. In addition, smart predicted two lipid spanning regions, suggesting that UPL6 and UPL7 are anchored to one or more membranes.

Phylogenic analysis of the HECT E3s

To help deduce the functions of the Arabidopsis UPL family, we compared their sequences to all available HECT-containing proteins from the rest of the eukaryotic kingdom in the smart annotation of the GenBank database (as of 2/10/2003; Letunic et al., 2002). Two hundred and four entries were identified, which were then condensed down to 158 representatives that appeared to be unique and of reasonable integrity to allow sequence alignments. (To avoid problems with the phylogenetic analysis, we did not include short sequences containing just the HECT domain, which were prevalent in the various EST populations (e.g. a wheat homolog AAB18822.1 of UPL3/4). All five HECT E3s from the yeast genome were present in this group as were the seven from Arabidopsis and one from rice. An additional 47 and 35 were from humans and mice, respectively, confirming that mammals have greatly expanded the number of HECT E3s compared to yeast and plants (Schwarz et al., 1998).

By clustalx analysis using just the approximately 350 C-terminal amino acids of the HECT domain for the alignment, we generated an unrooted phylogenic tree for the entire collection. As can be seen in Figure 2 and Supplementary Material, a number of clades emerged that often included representatives from several species, suggesting that specific HECT E3 types originated before the evolutionary split of plants, fungi, and animals. When the tree was color-coded for the presence of other motifs upstream of the HECT domain, a strikingly similar clustering of these motifs was also evident (Figure 2). For example, HECT-containing proteins predicted to include C2-WW, IQ-Calmodulin-binding, Poly(A), or Armadillo repeat domains were confined to the separate clades previously defined by the HECT domain alone. Such organization suggests a co-evolution of the HECT domain with other domains potentially important for target specificity, localization, and/or regulation. We note that not all proteins in the various clades were identified as having the signature-upstream motifs. Whether this failure represents the actual absence of such motifs or the possibilities that such motifs were not detected by smart or were missing from the sequences analyzed are not yet known.

image

Figure 2. Phylogenic analysis of the HECT E3 proteins from various animals, fungi and plants.

One hundred and fifty-eight unique HECT-containing proteins were identified with smart (Letunic et al., 2002). Those that contained near-full-length protein sequence were aligned based on the C-terminal 350 residues bearing the HECT domain by clustalx with a bootstrap value of 1000. The tree is color-coded for the presence of additional sequence motifs N-terminal to the HECT domain, the classifications of which are marked on the left. UBA, Ub-associated; C2/WW, Nedd-4 like domain; IG-FLMN, Filamin-type immunoglobulin domain; RCC1, Regulator of Chromosome Condensation; Poly(A), C-terminal domain of Poly(A)-binding protein. For those proteins with two or more additional motifs (e.g. UPL1 and UPL2 which has a UBA and UIM motif), the motif more in common with others in its clade was used as the defining motif. The positions of the Arabidopsis HECT E3s, (UPL1–UPL7) are identified by the arrowheads. The position of the rice HECT E3 (AAK14420.1) similar to UPL6 and UPL7 is located by the asterisk. The bar represents the branch length equivalent to 0.1 amino acid change per residue.

Download figure to PowerPoint

For the Arabidopsis UPL1–UPL7 families, in particular, the individual proteins also clustered with other proteins containing the signature motifs, even though the tree was generated with the HECT domain alone. UPL1/2, UPL3/4, and UPL6/7 subfamilies did not group with each other, but clustered with the UBA-, IQ-calmodulin-binding, and Armadillo-containing HECT-containing proteins from other organisms, respectively (Figure 2). The sole outlier was UPL5. It formed its own clade consistent with the fact that UPL5 was the only representative of the entire collection with a potential C-type lectin-binding domain (Figure 2).

The UPL3/4 subfamily has reasonably close homologs with known functions from other species. Their closest relatives are human TRIP12 and yeast UFD4, Armadillo repeat-containing HECT E3s involved in the ubiquitination of thyroid hormone receptor in the absence of hormone engagement, and the 26S proteasome-associated degradation of proteins bearing Ub moieties at their N-termini, respectively (Lee et al., 1995; Xie and Varshavsky, 2002). The clade containing the UPL6/7 subfamily (IQ-calmodulin-binding domains (Figure 2)) also includes yeast HUL5 and the human KIAA0010. Both the yeast and human proteins associate with the 26S proteasome, suggesting that these HECT E3s spatially connect Ub ligation with target breakdown (Leggett et al., 2002; You and Pickart, 2001).

T-DNA insertional mutants of UPL3

To help define the functions of the Arabidopsis UPL1–UPL7 gene family, we searched the mutation databases for insertional mutants in each of the loci (Krysan et al., 1999; Sessions et al., 2002). Aided by their large genomic footprint, a number of T-DNA insertions were identified for each gene. UPL3, in particular, was interrupted by three T-DNA insertions present in the Syngenta SAIL population. This population was generated with a T-DNA harboring the phosphinothricin acetyltransferase gene, and thus is resistant to the herbicide Basta (Sessions et al., 2002). upl3-3 contained a T-DNA 320-bp upstream of the ATG start codon, whereas upl3-2 and upl3-1 contained a T-DNA that interrupted the tenth exon at the 1236 and 1614 codons and caused small deletions of 18 and 7 bp, respectively (Figure 3a). In all three cases, left borders bracketed the T-DNA. Basta resistance of heterozygous selfed plants segregated in a 3 : 1 ratio, indicating that a single T-DNA integration site was present in each case.

RT-PCR analysis of homozygous mutant plants showed that upl3-1 and upl3-2 mutations, but not the upl3-3 mutation, block expression of the full-length UPL3 transcript (Figure 3b). Here, first-strand cDNAs were synthesized from total seedling RNA using reverse primers flanking the upl3-1 and upl3-2 T-DNA insertion sites (primer #1 or #2; Figure 3b). The RT products were then subjected to PCR using a forward primer (#3) and each of the downstream primers independently (#3 and #1 or #3 and #2). For a control, a similar analysis was performed using primers specific for PAE2 that encodes an α-5 subunit of the 26S proteasome (Fu et al., 1998a). The amplified regions of both PCR reactions spanned several introns, thus allowing us to confirm by size that the products were from the UPL3 and PAE2 mRNAs and not genomic DNA (Figure 3a).

As can be seen in Figure 3b, an RT-PCR product of the appropriate size (minus the eighth, ninth, and/or tenth introns) could be generated from wild-type and upl3-3 mRNAs using either of the RT products (primers #1 or #2) as the template. In contrast, the upstream RT products (#1), but not the downstream RT products (#2), worked for upl3-1 and upl3-2 templates, indicating that the sequence downstream of T-DNA insertion sites was absent from the mutant mRNAs. These truncated mRNAs should, in turn, direct the synthesis of UPL3 polypeptides, missing a substantial portion of the C-terminus including the HECT domain. As the HECT domain is essential for the ubiquitination activity of UPLs (Bates and Vierstra, 1999; Scheffner et al., 1995), the truncated upl3-1 and upl3-2 proteins should be enzymatically inactive even if they accumulate.

Phenotypic analysis of UPL3 mutants

To help identify the process(es) abrogated by the loss of UPL3 activity, we grew homozygous lines for the three upl3 mutants under a variety of conditions and compared their phenotypes to wild-type Col-0. Under optimal growth conditions, the upl3-1 and upl3-2 plants were morphologically indistinguishable from wild type, with no adverse effects seen on germination rate, growth of etiolated plants, development of green seedlings under white light, emergence of rosette leaves, root growth and root hair elongation, days to flowering, fecundity, and senescence. The level of Ub conjugates was also unaffected, indicating that overall, the ubiquitination pathway was not compromised by the mutations (data not shown). However, upon microscopic examination of whole-mount preparations of leaves, a striking alteration of trichome shape was evident. Whereas wild-type Arabidopsis plants (and the upl3-3 mutant) developed triradiate trichomes (Szymanski, 2000; Szymanski et al., 2000), the trichomes on upl3-1 and upl3-2 plants often developed five or more branches with some containing up to seven points (Figure 4a,b). The increased branching phenotype was apparent for trichomes on both the adaxial and abaxial leaf surfaces. Stem trichomes are normally unbranched in Arabidopsis (Szymanski, 2000; Szymanski et al., 2000). In the upl3-1 and upl3-2 backgrounds, a modest increase in the number of two-branched stem trichomes was evident, suggesting that these mutations influence all trichome types.

image

Figure 4. T-DNA disruption of UPL3 alters trichome branching.

(a) A field of trichomes from wild-type (WT) and upl3-2 plants observed by light microscopy.

(b) SEM of representative trichomes from the abaxial surface of WT, upl3-1, and upl3-2 leaves.

(c) Linkage of the T-DNA insertion in upl3-2 to the increase in trichome branches. Leaves from individual 15-day-old plants from a population segregating for the upl3-2 mutation were analyzed by PCR for the presence of the T-DNA in the UPL3 locus and examined microscopically for the number of trichome branches. The average number of branches (±SD) was calculated from the measurement of at least 25 trichomes from each plant. Measurements of 14 plants are shown; the correlation was identical for an additional 16 plants.

Download figure to PowerPoint

Scanning electron micrographs (SEM) of the mutant trichomes showed that they still contained a papillate surface. However, the circumference of the central trunk was often larger and emerged from an enlarged socket of accessory epidermal cells (Figure 4b). Whereas the three branches of wild-type trichomes typically originate close to each other from the trunk and have similar lengths, the supernumerary branches of mutant trichomes often emerged randomly as short spurs from the main branches to create a more reticulated architecture. The spacing of the trichomes, and shape and patterning of pavement epidermal cells and guard cells of the upl3-1 and upl3-2 leaves were indistinguishable to those from wild type (Figure 4a,b), suggesting that the mutations have little, if any, effect on trichome initiation or other aspects of epidermal development.

As can be seen from the analysis of individual plants segregating for the upl3-2 mutation, the trichome phenotype is tightly linked to the T-DNA insertion (Figure 4c). Whereas trichomes containing five or more branches were often seen on homozygous upl3-2 individuals, only rarely were trichomes containing more than three branches seen on wild-type plants. We noticed a slight increase in the percentage of aberrant trichomes in hemizygous upl3-1 and upl3-2 plants. To examine this observation more carefully, we counted the branch number of trichomes from wild-type, hemizygous, and homozygous upl3-1, upl3-2, and upl3-3 plants (Figure 5). Only a small percentage of wild-type and upl3-3 trichomes had more than three branches, whereas most trichomes from homozygous upl3-1 and upl3-2 plants had more than three with the majority having five. The hemizygous upl3-1 and upl3-2 plants displayed an intermediate phenotype; most trichomes still had three branches, but substantially more than wild type had four branches and a few even contained five branches (Figure 5). The effects on hemizygous plants suggest that the mutants display either haploinsufficiency or semidominance. Semidominance could be explained by the T-DNA insertions, prematurely truncating UPL3 transcription. Although missing the HECT domain required for catalysis, the resulting proteins may still contain site(s) for target interaction (e.g. Armadillo repeats) and thus could dominantly interfere with the action of wild-type UPL3.

image

Figure 5. Effect of hemizygous and homozygous upl3 mutations on trichome branching.

The proportion of trichomes were plotted relative to the number of branches from wild-type and hemizygous and homozygous upl3-1 and upl3-2 leaves. At least 200 trichomes were measured for each line.

Download figure to PowerPoint

The structural similarity between the UPL3 and UPL4 proteins, including the presence of a similar Armadillo repeat region upstream of a closely related HECT domain (Figure 1), suggested initially that this E3 pair might be functionally redundant. However, analysis of a T-DNA mutant bearing an insertion within the UPL4 coding region (and thus likely representing a null-activity allele) suggested otherwise. Homozygous upl4-1 plants revealed no obvious abnormal phenotypes when grown under optimal growth conditions and developed normal trichomes (unpublished data).

upl3 trichomes have an increase in nuclear DNA content

One of the early steps in Arabidopsis trichome differentiation is the initiation of three rounds of DNA endoreplication followed later by one additional round, resulting in trichome nuclei with an average DNA content of 32C. Because some, but not all, trichome-branching mutants affect this increase in ploidy, a link between endoreplication and branching has been proposed (Hulskamp et al., 1994; Melaragno et al., 1993). For example, strong alleles of the gl1 and ttg1 mutations, which fail to produce trichomes, also do not display increased endoreplication in non-differentiated pavement cells, while the gl3 mutant fails to form secondary branches and appears to attenuate endoreplication after the third round, resulting in a DNA content of 16C (Hulskamp et al., 1994). In contrast, mutants that form supernumerary branches such as rastifari (rfi), polycomb (pym), triptychon (try), spindly (spy), and kaktus (kak) often undergo additional rounds of endoreplication, resulting in enlarged trichome nuclei with a DNA content of 64C or more (Perazza et al., 1999). Similarly, tetraploid Arabidopsis, in which all cells presumably have a doubled DNA content, shows increased branching (Perazza et al., 1999).

Thus, it was likely that the strong upl3 mutants, like the other supernumerary-branching mutants, rfi, pym, try, spy, and kak, undergo extra round(s) of endoreplication. To examine this possibility, wild-type and homozygous upl3-1 and upl3-2 leaves were fixed and stained with 4′,6-diamidino-2-phenylindole (DAPI), and the DNA content of individual nuclei was determined by quantitative fluorescence imaging in three dimensions. As a ploidy standard, an identical analysis was performed on a population of guard cell nuclei, which should remain at 2C (Szymanski and Marks, 1998). Like those from wild type, the mutant nuclei were typically situated at or near the first branch point. However, an immediately obvious feature of mutant nuclei was their larger relative size, implying that they contain more DNA (Figure 6a). DNA measurements of large populations confirmed this increase in trichome cell ploidy. Whereas wild-type and upl3-3 nuclei typically had a narrow distribution of DNA content focused around 32C, both the homozygous upl3-1 and upl3-2 nuclei had a much broader distribution with some nuclei exceeding 64C (Figure 6b, data not shown). Thus, the nucleus in many mutant trichomes appears to have undergone a partial or complete additional round of endoreplication.

image

Figure 6. Effect of the upl3 mutants on the size and ploidy level of the trichome nucleus.

Leaves were fixed, stained with DAPI and visualized by epifluorescence microscopy.

(a) Fluorescence from representative trichome nuclei from wild-type and homozygous upl3-1 and upl3-2 mutants. Bar represents 10 μm.

(b) Relative fluorescence units (RFU) of individual nuclei as determined by epifluorescence. Ploidy levels (dashed lines) were estimated using a guard cell nuclei standard to determine the DNA content of a diploid nucleus (Szymanski and Marks, 1998). For simplicity, the values for wild type and upl3-3 and values for upl3-1 and upl3-2 nuclei were pooled in (b).

Download figure to PowerPoint

upl3 mutants are hypersensitive to GA

The supernumerary branching phenotype of the constitutive GA-response mutant, spy-5, and the glabrous phenotype of GA biosynthetic mutant, ga1-3, have provided a strong connection between this hormone and trichome differentiation (Perazza et al., 1998). In particular, GA appears to be required for the initiation and subsequent development of trichomes by stimulating the synthesis and/or activity of the transcriptional activation complex containing GL3, TTG1, and GL1 (Payne et al., 2000; Perazza et al., 1998; Szymanski et al., 2000). Given the similarity of the spy-5 and upl3 mutants with respect to trichome development, it was possible that they likewise are altered in their response to GA. To test this possibility, we grew wild-type and mutant seedlings on a GA3-containing medium and examined its effect on a number of responses diagnostic for this hormone (Olszewski et al., 2002). GA3 had no effect on the elongation of both wild-type and mutant roots. Increasing GA3 concentrations increased hypocotyl elongation, with the mutants showing a stronger response, which was indicative of hypersensitivity (Figure 7). Mild hypersensitivity was also seen for auxin (napthyleneacetic acid) and abscisic acid, suggesting that the response to other hormones was affected as well, although weakly. However, not all GA responses were altered; in particular, flowering time under short days and seed germination was similar for the mutants and the wild type (data not shown).

image

Figure 7. Hypersensitivity of the upl3-1 and upl3-2 mutants to GA.

Mutant and wild-type seeds were stratified on GA-free medium and then sown on GA3-containing solid medium. Hypocotyl lengths (±SD) were then measured after 10 days of growth. Left panel, average of at least 10 plants per treatment. Right panels, pictures of two representative hypocotyls from each treatment.

Download figure to PowerPoint

The kak-2 mutation is within the UPL3 locus

As mentioned above, a number of loci have been identified that affect trichome branching and endoreplication, including SPY, TRY, PYM, RFI, and KAK (Perazza et al., 1999). Whereas SPY encodes an O-linked N-acetyl glucosamine transferase (Tseng et al., 2002), the latter four loci remain to be identified. try and rfi mutations also generate twin trichomes emerging from the same socket (Perazza et al., 1999); the absence of this patterning defect in upl3 mutants indicates that they are not allelic to either try or rfi. In contrast, the upl3-1 and upl3-2 phenotype appears indistinguishable to that reported for kak mutants, which display excess branching of leaf trichomes, a modest increase in branched trichomes on the stem, and an extra round of endoreplication, but do not affect trichome spacing and number, or promote the formation of twin trichomes (Perazza et al., 1999). KAK has been mapped to the bottom of chromosome 4 approximately 2.5 cM from the simple sequence length polymorphism marker nga1107 (Perazza et al., 1999). Coincidentally, the UPL3 gene is only 48 kb from nga1107, strongly suggesting that KAK and UPL3 are the same locus.

To examine this possibility, we tested for allelism by crossing homozygous upl3-1 and upl3-2 plants with homozygous kak-2 plants. All the progeny had abnormal trichomes similar to the parents, indicating that UPL3 and KAK are the same locus (data not shown). To define the kak-2 lesion, we sequenced the UPL3 gene from the kak-2 mutant (Perazza et al., 1999) and compared it to the wild-type UPL3 sequence of its Ler parent. In the full coding sequence, we discovered a single G to A transition at position 2408 from the ATG in kak-2 genomic DNA (Figure 3a). Based on its position at the 3' splice junction of the sixth intron (AG to AA), we predicted that the kak-2 allele would disrupt the maturation of the UPL3 mRNA. Sequence analysis of RT-PCR products spanning the kak-2 mutation revealed that instead of using the normal splice site, a cryptic splice site was used that is 14 bp 3' to the kak-2 mutation. This aberrant splicing reaction deleted the 5' end of the seventh exon, which could be observed as a shorter RT-PCR product from the UPL3 mRNA (Figure 3c), and introduced a frame shift with a premature stop codon after seven amino acids. The phenotypic similarity of kak-2 to upl3-1 and upl3-2 suggests that kak-2 likely represents a strong allele.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary Material
  9. References
  10. Supporting Information

HECT E3s represent a large and heterogeneous class of Ub-protein ligases that regulate the levels of a number of critical cell regulators in both yeast and animals (Huibregtse et al., 1995; Pickart, 2001). Similarly, plants contain HECT E3s, with seven (UPL1–UPL7) present in Arabidopsis. Like their yeast and animal counterparts, UPL1–UPL7 are large proteins (96–405 kDa), bearing an assortment of functional motifs preceding a C-terminal consensus HECT domain. The seven proteins from Arabidopsis could be further grouped into four subfamilies based on the presence of one or more signature domains upstream of the HECT domain, including Armadillo repeats, C-type lectin-binding domains, and IQ calmodulin-binding domains that could be important for target specificity. While the substrates of UPL1–UPL7 are not yet known, the presence of these additional domains imply interactions with targets bearing NLSs (UPL3/4), glycosyl groups (UPL5), and calmodulin (UPL6/7). In addition, UBA, UIM, and UBL domains were evident in specific members, suggesting that non-covalent interactions with Ub (UBA and UIM) or with Ub-binding proteins (UBL) are important as well. To date, ubiquitination activity has been confirmed only for UPL1 (Bates and Vierstra, 1999). But, given that the remaining six also contain a HECT domain strongly related to the consensus sequence including the presence of the active-site cysteine, it is quite likely that these proteins are also Ub ligases.

Members of the four Arabidopsis UPL subfamilies were often more similar phylogenetically to HECT E3s from other species than to members of other Arabidopsis subfamilies, suggesting that these subfamilies arose before the split of the animal, fungal, and plant kingdoms. Cluster analysis using the HECT domain alone implied that groups of HECT E3s share important features within this domain that may help promote their catalytic activity. Possibilities include their unique abilities to interact with Ub and/or specific E2s through conserved arrangements of specific amino acids surrounding the active-site cysteine and the E2-binding pocket, respectively (Huang et al., 1999). With respect to the latter possibility, studies with human E6-AP, yeast RSP5, and Arabidopsis UPL1 indicate that HECT E3s specifically associate with a subfamily of E2s that includes yeast UBC4/5, human UBCH5-8, and Arabidopsis UBC8-12 (Bates and Vierstra, 1999; Kumar et al., 1997). Furthermore, within this E2 subfamily, a preferential pairing of individual members with specific HECT E3s has been noted (Kumar et al., 1997). An interesting possibility is that this pairing is reflected in the phylogenetic clustering of the HECT domain, with each clade being defined by the specific E2 isoform in use. We also found that those HECT E3s with similar HECT domains often contain similar accessory domains that could confer analogous target specificity, cellular localization, and/or regulation.

Whereas yeast contains only five HECT E3s, 50 or more may be present in mice and humans, indicating that mammals have greatly expanded the use of this ligase type (Schwarz et al., 1998 and this report). In contrast, no comparable expansion is evident for Arabidopsis. Instead, Arabidopsis appears to have relied more heavily upon two other ligase types for selective ubiquitination, the SCF complex and RING E3s, where recent genome analyses have identified more than 700 and 400 potential isoforms, respectively (Gagne et al., 2002; Kosarev et al., 2002). However, the conservation of specific HECT E3s among a wide array of eukaryotes indicates that they have catalytic activities that cannot be replaced by other E3 types.

Genetic analysis of Arabidopsis UPL3 and our discovery that the DNA lesion in the trichome mutant kak-2 occurs within the UPL3 gene indicate that this HECT E3 participates in the regulation of trichome development. Trichomes emerge from the pavement epidermal cells by a well-defined temporal sequence of events, with genetic studies suggesting that branching and endoreplication are coordinately regulated (Szymanski et al., 2000). Phenotypic analysis of the upl3-1 and upl3-2 mutants (and kak (Hulskamp et al., 1994; Perazza et al., 1999)) implies that they are compromised in the latter stages of trichome development. Normally, the final step(s) involves initiation of the final round of endoreplication to generate a 32C nucleus and creation of the second branch point to form the characteristic triradiate architecture. Instead, the upl3 mutants generate supernumerary branched trichomes that often undergo an additional round of endoreplication resulting in some trichome cells becoming 64C. Consequently, it appears that loss of UPL3 function relieves a critical inhibitory step that normally prevents excess endoreplication and branching toward the end of trichome morphogenesis.

A similar phenotype to upl3-1 and upl3-2 was not evident for a T-DNA mutant affecting UPL4, the closest homolog of UPL3 in Arabidopsis. Under the assumption that UPL3 and UPL4 are expressed at similar levels based on a near equal amount of available ESTs (http://www.arabidopsis.org), we conclude that the function(s) and target(s) of UPL3 are distinct from UPL4 and other members of the HECT E3 family in Arabidopsis. It should be emphasized that UPL3 likely has functions outside of trichome development. The GA hypersensitivity of the upl3 mutants with respect to hypocotyl growth clearly demonstrates a role for UPL3 in this tissue as well. However, the phenotype of upl3-1 and upl3-2 plants outside of trichome morphogenesis are much more subtle, raising the possibility that UPL3 works together with UPL4 in other Arabidopsis cell types.

Based on the predicted activity of UPL3, that of directing the ubiquitination and subsequent removal of one or more short-lived proteins, we speculate that this HECT E3 participates in trichome development by targeting for degradation of one or more activators that promote endoreplication and secondary branching. upl3 mutants would fail to remove these activator(s); their overaccumulation would then continue the branching process and induce extra round(s) of endoreplication. Based on the phenotype alone, possible targets of UPL3 could include proteins genetically epistatic to UPL3, whose loss would generate phenotypes opposite to that of upl3 mutations (i.e. reduced branching and impaired final round of endoreplication) and/or generate the same phenotype as upl3 when overexpressed. Several proteins that approximate these criteria include those encoded by GL1, GL3, ZWI, FRC2, FRC4, and STI (Hulskamp et al., 1994; Ilgenfritz et al., 2003; Luo and Oppenheimer, 1999; Payne et al., 2000).

GL3 is an intriguing candidate based on several criteria. It is a bHLH transcription factor that appears to work as a positive regulator of trichome differentiation in a complex with the WD-40 protein TTG1 and the Myb-transcription factor GL1 (Payne et al., 2000). gl3 plants produce fewer trichomes that are smaller, have less branches, and undergo one less round of endoreplication than the wild type (Hulskamp et al., 1994). The finding that mutations in GL1, GL3, and TTG1 are epistatic to most other branching mutants places this complex at the center of endoreplication and branch regulation (Hulskamp et al., 1994; Payne et al., 2000). The NLS of GL3 could serve as the UPL3-recognition site, using the Armadillo repeats of UPL3 to bind the protein in a similar way that the Armadillo-repeat region of importin-α interacts with NLSs of nuclear proteins during their import (Szurek et al., 2001). Ubiquitination by UPL3 could either remove the whole GL3/TTG1/GL1 complex or just the GL3 subunit. In support of the former, overexpression of GL3 increased trichome number, but did not result in the dramatic supernumerary branching that we have observed for upl3 (Payne et al., 2000), suggesting that multiple proteins may be stabilized in upl3-1 and upl3-2 plants.

Another candidate is ZWI that appears to be epistatic to supernumerary branching mutants and encodes a kinesin motor protein involved in microtubule-driven movement. Overexpression of the ZWI gene has failed to increase the levels of the ZWI protein, suggesting that it is naturally short-lived (A. N. Reddy, unpublished). Other possible targets include members of the FURCA (FRC) family, especially FRC2 and FRC4 (Luo and Oppenheimer, 1999). They appear to act as positive regulators of trichome branching and work downstream in the same pathway as GL3. Loss of each results in two branched trichomes, but it is not yet known if endoreplication is altered. Another reduced branching mutant, sti, shows increased branching upon overexpression, yet endoreplication is not affected in either the mutant or the overexpression lines (Ilgenfritz et al., 2003). Other possible UPL3 targets include cell-cycle checkpoint proteins that promote DNA replication in the absence of cytokinesis. Targeted removal of such proteins may be essential to halt endoreplication. Recent studies have shown that misexpression of the cyclin-dependent kinase inhibitor ICK1/KRP1 in trichomes leads to reduced DNA content, decreased cell size, and reduced trichome branching (Schnittger et al., 2003). Interestingly, this work also revealed additional complexity in the relationship between DNA content and cell size as DNA-dependent and -independent effectors of trichome size were detected (Schnittger et al., 2003).

The upl3-1 and upl3-2 mutants are modestly hypersensitive to GA, consistent with the role of this hormone in promoting trichome initiation through the GL3/TTG1/GL1 complex (Perazza et al., 1998). Perhaps, a positive regulator of trichome development at the distil end of the GA-signaling pathway is stabilized in upl3 mutants. Candidate targets include the positive regulators of the GA response encoded by the Arabidopsis GAMYB-like genes (Gocal et al., 2001). However, given the modest alterations in GA-sensitivity, the regulator is not likely to be a main effector of GA responsiveness.

Whatever its targets, our discovery that UPL3 participates in trichome development demonstrates that ubiquitination and likely degradation of one or more key regulatory proteins play an important role in this morphogenic process. Clearly, the identification of protein(s) that interacts with UPL3 will be instrumental in defining the position of this Ub ligase in the complex trichome regulatory network and ultimately how ubiquitination controls trichome shape. Here, protein interaction studies involving UPL3 and candidate targets may be informative.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary Material
  9. References
  10. Supporting Information

UPL3 DNA sequence analysis

Arabidopsis loci encoding proteins bearing the HECT domain were identified by blast searches of the A. thaliana ecotype Columbia DNA sequence database (http://www.arabidopsis.org) using the HECT domains of UPL1 (AF127564; Bates and Vierstra, 1999) and human E6-AP (Huibregtse et al., 1995) as queries. Full-length coding regions were either deduced from near full-length cDNAs (UPL1–UPL3, UPL5, and UPL6) or predicted by their alignments with the intron/exons of their nearest relatives (UPL4 and UPL7). A near full-length cDNA for UPL3 encompassing the entire predicted coding region was generated by joining the 4289-bp partial cDNA AV529735.1 (KAZUSA, Chiba, Japan) containing the poly(A) tail, the 3' UTR, and the 3' coding sequence with an RT-PCR product that contained the remaining 5' coding sequence. The RT-PCR reactions were performed with 1 μg of total RNA extracted with the Trizol reagent (GIBCO-Invitrogen, Carlsbad, CA, USA) from 15-day-old seedlings. The forward and reverse primers (TTGCCCGGGGGATCCATGGAAACTCGGAGCCGCAA and ACTTGTAGAGCAGGAACCAAGACCTGTGGGTCTTTC) added a BamH1 site prior to the ATG and included a unique Drd1 site at the 5' end of the cDNA. The AV529735.1 cDNA digested with Drd1 and BamH1 was appended to the Drd1/BamH1-digested RT-PCR product. Both strands of the UPL3 cDNA were sequenced using dideoxy nucleotide dye terminator chemistry (Perkin-Elmer, Foster City, CA, USA).

Amino acid sequence alignments were performed using clustalx mac v.1.6b and displayed using mac boxshade v.2.11 (Institute of Animal Health, Pirbright, UK). The annotations for UPL1–UPL7 can be found in the Arabidopsis genome database: UPL1 (At1g55860), UPL2 (At1g70320), UPL3 (At4g38600/38610), UPL4 (At5g02880), UPL5 (At4g12570), UPL6 (At3g17205/17206), and UPL7 (At3g53090). The full-length cDNA sequence for UPL3 was deposited in the GenBank database under accession number AY265959. The UPL3 genomic locus was sequenced in its entirety from the Ler ecotype and kak-2 following PCR amplification of subdomains with ExTaq polymerase (PanVerra, Madison, WI, USA). The RT-PCR primers used to analyze kak-2 and Ler mRNA were #4 (GTGCAAGAAACTACCTTCTGATGCAT); #5 (AAGAATGCTACTAATACCAAGAAGAA); #6, (TGAGATAGTCAACCTAGCGAATGAGC), and #7 (ACTTGTAGAGCAGGAACCAAGACCTGTGGGTCTTTC; see Figure 3a).

Phylogenetic analysis

The smart database and its protein motif search program (http://smart.embl-heidelberg.de) were used to identify 204 distinct eukaryotic loci encoding a HECT domain and to detect other possible protein motifs within the collected sequences. Individual analyses reduced this number to 158 non-redundant proteins. A phylogenetic tree using the C-terminal 350 amino acids encompassing the HECT domain was generated by the neighbor-joining method with a bootstrap value of 1000 (clustalx mac v.1.6b). An expanded version of this tree identifying the species and GenBank accession number for each protein can be found in the Supplementary Material at The Plant Journal web site.

T-DNA insertion mutants of UPL3

A search of the Syngenta/SAIL Arabidopsis T-DNA population (generated with the A. thaliana ecotype Col-0; Sessions et al., 2002) with the UPL3 gene sequence revealed three T-DNA insertion lines at or near the UPL3 locus: 503D6 (upl3-1), 339F5 (upl3-2), and 1278F8 (upl3-3). Each mutant was back-crossed three times to Col-0 wild type and then selfed to generate homozygous mutant plants, as determined by Basta resistance, and by PCR with appropriate combinations of 5' and 3' UPL3 primers and the T-DNA left-border primer (Krysan et al., 1999). Expression of UPL3 was monitored by RT-PCR using the primers #1 (ACTGAGTCCGCCAATTTCACATCATG) and #2 (CAGAAATCGTGCTGCTATTGTCACCA) for the first round of RT followed by 35 cycles of PCR using #3 (GGTAGTCCATGCTGTAGATCAACTTG) and #1 primers. Control RT-PCR reactions were performed with the mRNA for the proteasome α-5 subunit gene PAE2 using the primers (CTGACATTGAGGTTTATCTCAGATCG and CTCAACTCGATAAAATCCATTATCTG) for both the RT and the PCR reactions. The RT reactions were performed using Mu-MLV reverse transcriptase (Promega, Madison, WI, USA) and 1 μg of total RNA from 15-day-old seedlings, and the PCR reactions were performed with ExTaq DNA polymerase (PanVera, Madison, WI, USA).

Plant growth and hormone treatments

Arabidopsis seeds were surface-sterilized, plated on Murashige-Skoog medium (2.15 g l−1), 1% sucrose, 0.7% agar, and stratified at 4°C for 4 days. Seedlings were grown at 22°C under white fluorescent lighting for 16-h light/8-h dark photoperiod. GA3 (Sigma, St Louis, MO, USA) was added to the medium, where indicated. Root and hypocotyl elongation measurements were taken from a minimum of 12 plants after 10 days growth on vertical plates.

Phenotypic analysis of upl3 mutant trichomes

SEM, quantitation of nuclear DNA content, and assessment of trichome branch number distribution were performed on the third or fourth leaves of 22- to 24-day-old seedlings. For SEM, leaves were vacuum-infiltrated with 5% glutaraldehyde in 50 mm sodium cacodylate (pH 7.0) and incubated at 25°C for 4 h with one change of fresh fixative after 2 h. Leaves were then rinsed in 50 mm sodium cacodylate (pH 7.0) at 4°C for 12–14 h prior to dehydration by an ethanol series (Busse and Evert, 1999). Samples were critically point-dried (Samdri 780; Tousimis Corp., Rockville, MD, USA), sputter-coated with 350 Å of gold (Autoconducta Vac IV; SeeVac Inc., Pittsburgh, PA, USA), and visualized with a Hitachi S-570 scanning electron microscope with a lanthanum hexaborate gun set to an accelerating voltage of 5 kV (Hitachi High Technologies, Pleasanton, CA, USA). Digital images were captured on a gatan-digiscan v2.5.6 (Gatan Inc., Pleasanton, CA, USA).

Nuclear DNA quantitation was performed according to Szymanski and Marks (1998). Leaves were fixed in 3 : 1 ethanol:acetic acid and 1 mm MgCl2 for 45 min, cleared in 95% ethanol and 1 mm MgCl2 for 12–14 h, and washed for 1 h in fresh 95% ethanol and 1 mm MgCl2. Cleared leaves were washed three times in 20 mm sodium phosphate (pH 7.0), 150 mm sodium chloride, 1 mm MgCl2 (PBS + MgCl2), and stained in the same buffer plus 1 μg ml−1 DAPI for 15 min. Leaves were washed three times for 15 min each in PBS + MgCl2 and mounted in Vectashield (Vector Labs, Burlingame, CA, USA) under a coverslip. Four leaves from different plants were sampled for each genotype, and twenty trichomes were imaged from one lateral half of each leaf marked by the mid-vein.

DAPI-stained trichome nuclei were visualized using epifluorescent microscopy. Individual trichome images were captured using an Olympus BX51 microscope and Hamamatsu C4742-95 cooled CCD camera. Nuclear images were acquired with a 40× objective lens using metamorph V.4.6 software and collected as a stack of five sections through the nucleus along the Z-axis. Stack files were cropped and then deconvolved using autodeblur V.7.5.3 software (AutoQuant Imaging, Inc., Watervliet, NY, USA) with 20 iterations using the maximization likelihood estimation blind deconvolution method specified by the expectation-maximization algorithm (Holmes et al., 1995). The integrated fluorescent intensity of each nucleus from the maximum projections of the deconvoluted stacks was determined using metamorph V.4.6 software. Similar measurements of 10 guard cell nuclei from each leaf were employed as an internal standard for a diploid nucleus. The values for guard cell nuclei were consistent from leaf to leaf and were used to construct an estimated ploidy scale. Immunoblotting with anti-Ub antibodies were performed as described (Smalle et al., 2003).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary Material
  9. References
  10. Supporting Information

We thank Dr James Busse for assistance with the SEM, Joseph Walker and Jennifer Gagne for technical help, Drs. Steve Goff and Alan Session for access to the Syngenta SAIL mutant collection, the Kazusa DNA Research Institute for the UPL3 partial cDNA AV529735.1, Dr John Larkin for the kak-2 mutant, and Dr Jan Smalle for critically reading the manuscript. For Abby. This work was supported by a grant from the US National Science Foundation Arabidopsis 2010 program to RDV (MCB-0115870).

Supplementary Material

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary Material
  9. References
  10. Supporting Information

Figure S1. An expanded version of Figure 2 including the organism, Accession number, number of amino acids, and domains present for each entry in the phylogenetic tree.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary Material
  9. References
  10. Supporting Information
  • Bates, P.W. and Vierstra, R.D. (1999) UPL1 and 2, two 405-kDa ubiquitin-protein ligases from Arabidopsis thaliana related to the HECT-domain protein family. Plant J. 0, 183195.
  • Busse, J.S. and Evert, R.F. (1999) Pattern of differentiation of the first vascular elements in the embryo and seedling of Arabidopsis thaliana. Int. J. Plant Sci. 160, 113.
  • Fu, H., Doelling, J.H., Arendt, C.S., Hochstrasser, M. and Vierstra, R.D. (1998a) Molecular organization of the 20S proteasome gene family from Arabidopsis thaliana. Genetics, 149, 677692.
  • Fu, H., Sadis, S., Rubin, D.M., Glickman, M., Van Nocker, S., Finley, D. and Vierstra, R.D. (1998b) Multiubiquitin chain binding and protein degradation are mediated by distinct domains within the 26S proteasome subunit Mcb1. J. Biol. Chem. 273, 19701981.
  • Gagne, J.M., Downes, B.P., Shiu, S.H., Durski, A.M. and Vierstra, R.D. (2002) The F-box subunit of the SCF E3 complex is encoded by a diverse superfamily of genes in Arabidopsis. Proc. Natl. Acad. Sci. USA, 99, 1151911524.
  • Gitan, R.S. and Eide, D.J. (2000) Zinc-regulated ubiquitin conjugation signals endocytosis of the yeast ZRT1 zinc transporter. Biochem. J. 346, 329336.
  • Gocal, G.F.W., Sheldon, C.C., Gubler, F. et al. (2001) GAMYB-like genes, flowering, and gibberellin signaling in Arabidopsis. Plant Physiol. 127, 16821693.
  • Harty, R.N., Brown, M.E., Wang, G., Huibregtse, J. and Hayes, F.P. (2000) A PPxY motif within the VP40 protein of Ebola virus interacts physically and functionally with a ubiquitin ligase: implications for filovirus budding. Proc. Natl. Acad. Sci. USA, 97, 1387113876.
  • Harvey, K.F. and Kumar, S. (1999) Nedd4-like proteins: an emerging family of ubiquitin-protein ligases implicated in diverse cellular functions. Trends Cell Biol. 9, 166169.
  • Hein, C., Springael, J.Y., Volland, C., Haguenauer-Tsapis, R. and Andre, B. (1995) NPl1, an essential yeast gene involved in induced degradation of Gap1 and Fur4 permeases, encodes the Rsp5 ubiquitin-protein ligase. Mol. Microbiol. 18, 7787.
  • Hellmann, H. and Estelle, M. (2002) Plant development: regulation by protein degradation. Science, 297, 793797.
  • Hershko, A. and Ciechanover, A. (1998) The ubiquitin system. Annu. Rev. Biochem. 67, 425479.
  • Hofmann, K. and Bucher, P. (1996) The UBA domain: a sequence motif present in multiple enzyme classes of the ubiquitination pathway. Trends Biochem. Sci. 21, 172173.
  • Hofmann, K. and Falquet, L. (2001) A ubiquitin-interacting motif conserved in components of the proteasomal and lysosomal protein degradation systems. Trends Biochem. Sci. 26, 347350.
  • Holmes, T.J., Bhattacharyya, S., Cooper, J.A., Hanzel, D., Krishnamurthi, V., Lin, W., Roysam, B., Szarowski, D.H. and Turner, J.N. (1995) Light microscopic images reconstructed by maximum likelihood deconvolution. In Handbook of Biological Confocal Microscopy (Pawley, J.B., ed.). New York: Plenum Press, pp. 389402.
  • Huang, L., Kinnucan, E., Wang, G., Beaudenon, S., Howley, P.M., Huibregtse, J.M. and Pavletich, N.P. (1999) Structure of an E6AP-UbcH7 complex: insights into ubiquitination by the E2-E3 enzyme cascade. Science, 286, 13211326.
  • Huibregtse, J.M., Scheffner, M., Beaudenon, S. and Howley, P.M. (1995) A family of proteins structurally and functionally related to the E6-AP ubiquitin-protein ligase. Proc. Natl. Acad. Sci. USA, 92, 25632567.
  • Huibregtse, J.M., Yang, J.C. and Beaudenon, S.L. (1997) The large subunit of RNA polymerase II is a substrate of the Rsp5 ubiquitin-protein ligase. Proc. Natl. Acad. Sci. USA, 94, 36563661.
  • Hulskamp, M., Misera, S. and Jurgens, G. (1994) Genetic dissection of trichome cell development in Arabidopsis. Cell, 76, 555566.
  • Ilgenfritz, H., Bouyer, D., Schnittger, A., Mathur, J., Kirik, V., Schwab, B., Chua, N., Jurgens, G. and Hulskamp, M. (2003) Arabidopsis STICHEL gene is a regulator of trichome branch number and encodes a novel protein. Plant Physiol. 131, 643655.
  • Kosarev, P., Mayer, K.F. and Hardtke, C.S. (2002) Evaluation and classification of RING-finger domains encoded by the Arabidopsis genome. Genome Biol. 3, 112.
  • Krysan, P.J., Young, J.C. and Sussman, M.R. (1999) T-DNA as an insertional mutagen in Arabidopsis. Plant Cell, 11, 22832290.
  • Kumar, S., Kao, W.H. and Howley, P.M. (1997) Physical interaction between specific E2 and HECT E3 enzymes determines functional cooperativity. J. Biol. Chem. 272, 1354813554.
  • Kumar, S., Talis, A.L. and Howley, P.M. (1999) Identification of HHR23A as a substrate for E6-associated protein-mediated ubiquitination. J. Biol. Chem. 274, 1878518792.
  • Lee, J.W., Choi, H.S., Gyuris, J., Brent, R. and Moore, D.D. (1995) Two classes of proteins dependent on either the presence or absence of thyroid hormone for interaction with the thyroid hormone receptor. Mol. Endocrinol. 9, 243254.
  • Leggett, D.S., Hanna, J., Borodovsky, A., Crosas, B., Schmidt, M., Baker, R.T., Walz, T., Ploegh, H. and Finley, D. (2002) Multiple associated proteins regulate proteasome structure and function. Mol. Cell, 10, 495507.
  • Letunic, I., Goodstadt, L., Dickens, N.J., Doerks, T., Schultz, J., Mott, R., Ciccarelli, F., Copley, R.R., Ponting, C.P. and Bork, P. (2002) Recent improvements to the SMART domain-based sequence annotation resource. Nucl. Acids Res. 30, 242244.
  • Luo, D. and Oppenheimer, D.G. (1999) Genetic control of trichome branch number in Arabidopsis: the roles of the FURCA loci. Development, 126, 55475557.
  • Matsuura, T., Sutcliffe, J.S., Fang, P., Galjaard, R.J., Jiang, Y.H., Benton, C.S., Rommens, J.M. and Beaudet, A.L. (1997) De novo truncating mutations in E6-AP ubiquitin-protein ligase gene (UBE3A) in Angelman syndrome. Nat. Genet. 15, 7477.
  • Melaragno, J.E., Mehrotra, B. and Coleman, A.W. (1993) Relationship between endopolyploidy and cell size in epidermal tissue of Arabidopsis. Plant Cell, 5, 16611668.
  • Merkle, T. (2001) Nuclear import and export of proteins in plants: a tool for the regulation of signalling. Planta, 213, 499517.
  • Mitsui, K., Nakanishi, M., Ohtsuka, S., Norwood, T.H., Okabayashi, K., Miyamoto, C., Tanaka, K., Yoshimura, A. and Ohtsubo, M. (1999) A novel human gene encoding HECT domain and RCC1-like repeats interacts with cyclins and is potentially regulated by the tumor suppressor proteins. Biochem. Biophys. Res. Commun. 266, 115122.
  • Nawaz, Z., Lonard, D.M., Smith, C.L., Lev-Lehman, E., Tsai, S.Y., Tsai, M.J. and O'Malley, B.W. (1999) The Angelman syndrome-associated protein, E6-AP, is a coactivator for the nuclear hormone receptor superfamily. Mol. Cell Biol. 19, 11821189.
  • Olszewski, N., Sun, T. and Gubler, F. (2002) Gibberellin signaling: biosynthesis, catabolism, and response pathways. Plant Cell, 14, S61–S80.
  • Payne, C.T., Zhang, F. and Lloyd, A.M. (2000) GL3 encodes a bHLH protein that regulates trichome development in Arabidopsis through interaction with GL1 and TTG1. Genetics, 156, 13491362.
  • Perazza, D., Vachon, G. and Herzog, M. (1998) Gibberellins promote trichome formation by up-regulating GLABROUS1 in Arabidopsis. Plant Physiol. 117, 375383.
  • Perazza, D., Herzog, M., Hulskamp, M., Brown, S., Dorne, A.M. and Bonneville, J.M. (1999) Trichome cell growth in Arabidopsis thaliana can be derepressed by mutations in at least five genes. Genetics, 152, 461476.
  • Pickart, C.M. (2001) Mechanisms underlying ubiquitination. Annu. Rev. Biochem. 70, 503533.
  • Scheffner, M., Nuber, U. and Huibregtse, J.M. (1995) Protein ubiquitination involving an E1-E2-E3 enzyme ubiquitin thioester cascade. Nature, 373, 8183.
  • Schnittger, A., Weinl, C., Bouyer, D., Schobinger, U. and Hulskamp, M. (2003) Misexpression of the cyclin-dependent kinase inhibitor ICK1/KRP1 in single-celled Arabidopsis trichomes reduces endoreduplication and cell size and induces cell death. Plant Cell, 15, 303315.
  • Schwarz, S.E., Rosa, J.L. and Scheffner, M. (1998) Characterization of human hect domain family members and their interaction with UbcH5 and UbcH7. J. Biol. Chem. 273, 1214812154.
  • Sessions, A., Burke, E., Presting, G. et al. (2002) A high-throughput Arabidopsis reverse genetics system. Plant Cell, 14, 29852994.
  • Smalle, J., Kurepa, J., Yang, P., Emborg, T.J., Babiychuk, E., Kushnir, S. and Vierstra, R.D. (2003) The pleiotropic role of the 26S proteasome subunit RPN10 in Arabidopsis growth and development supports a substrate-specific function in abscisic acid signaling. Plant Cell, 15, 965980.
  • Szurek, B., Marois, E., Bonas, U. and Van den Ackerveken, G. (2001) Eukaryotic features of the Xanthomonas type III effector AvrBs3: protein domains involved in transcriptional activation and the interaction with nuclear import receptors from pepper. Plant J. 26, 523534.
  • Szymanski, D.B. (2000) The role of actin during Arabidopsis trichome morphogenesis. In Actin: a Dynamic Framework for Multiple Plant Cell Functions (Staiger, C.J., ed.). Dordrecht: Kluwer Academic Publishers, pp. 119.
  • Szymanski, D.B. and Marks, D.M. (1998) GLABROUS1 overexpression and TRIPTYCHON alters the cell cycle and trichome cell fate in Arabidopsis. Plant Cell, 10, 20472062.
  • Szymanski, D.B., Lloyd, A.M. and Marks, M.D. (2000) Progress in the molecular genetic analysis of trichome initiation and morphogenesis in Arabidopsis. Trends Plant Sci. 5, 214219.
  • Tseng, T.S., Thornton, T.M., Gopalraj, M. and Olszewski, N.E. (2002) Ectopic expression of the tetratricopeptide repeat domain of SPINDLY causes defects in gibberellin response. Plant Physiol. 129, 605615.
  • Verdecia, M.A., Joazeiro, C.A., Wells, N.J., Ferrer, J.L., Bowman, M.E., Hunter, T. and Noel, J.P. (2003) Conformational flexibility underlies ubiquitin ligation mediated by the WWP1 HECT domain E3 ligase. Mol. Cell, 11, 249259.
  • Vierstra, R.D. (2003) The ubiquitin/26S proteasome pathway, the complex last chapter in the life of many plant proteins. Trends Plant Sci. 8, 135142.
  • Wang, G., Yang, J. and Huibregtse, J.M. (1999) Functional domains of the Rsp5 ubiquitin-protein ligase. Mol. Cell Biol. 19, 342352.
  • Xie, Y. and Varshavsky, A. (2002) UFD4 lacking the proteasome-binding region catalyses ubiquitination but is impaired in proteolysis. Nat. Cell Biol. 4, 10031007.
  • You, J. and Pickart, C.M. (2001) A HECT domain E3 enzyme assembles novel polyubiquitin chains. J. Biol. Chem. 276, 1987119878.

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. Supplementary Material
  9. References
  10. Supporting Information

Figure S1. An expanded version of Figure 2 including the organism , Accession number, number of amino acids, and domains present for each entry in the phylogenetic tree.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

FilenameFormatSizeDescription
TPJ_1844_sm_FigureS1.eps213KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.