SEARCH

SEARCH BY CITATION

Keywords:

  • abscisic acid;
  • viviparous1;
  • abscisic acid insensitive 3;
  • Arabidopsis;
  • maize

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The maize Vp1 gene and abi3 gene of Arabidopsis are believed to be orthologs based on similarities of the mutant phenotypes and amino acid sequence conservation. Here we show that expression of VP1 driven by the 35S promoter can partially complement abi3–6, a deletion mutant allele of abi3. The visible phenotype of seed produced from VP1 expression in the abi3 mutant background is nearly indistinguishable from wild type. VP1 fully restores abscisic acid (ABA) sensitivity of abi3 during seed germination and suppresses the early flowering phenotype of abi3. The temporal regulation of C1-β-glucronidase (GUS) and chlorophyl a/b binding protein (cab3)-GUS reporter genes in developing seeds of 35S-VP1 lines were similar to wild type. On the other hand, two qualitative differences are observed between the 35S-VP1 line and wild type. The levels of CRC and C1-GUS expression are markedly lower in the seeds of 35S-VP1 lines than in wild type suggesting incomplete complementation of gene activation functions. Similar to ectopic expression of ABI3 (Parcy et al., 1994), ectopic expression of VP1 in vegetative tissue enhances ABA inhibition of root growth. In addition, 35S-VP1 confers strong ABA inducible expression of the normally seed-specific cruciferin C (CRC) gene in leaves. In contrast, ectopic ABA induction of C1-GUS is restricted to a localized region of the root elongation zone. The ABA-dependent C1-GUS expression expanded to a broader area in the root tissues treated with exogenous application of auxin. Interestingly, auxin-induced lateral root formation is completely suppressed by ABA in 35S-VP1 plants but not in wild type. These results indicate VP1 mediates a novel interaction between ABA and auxin signaling that results in developmental arrest and altered patterns of gene expression.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The maturation and germination phases of seed development are normally separated by a quiescent, desiccation tolerant phase. Maize viviparous (vp) mutants cause precocious seed germination and loss of desiccation tolerance, suggesting that the integration of these three developmental phases is disrupted. Therefore, vp mutants are instructive of the mechanisms of development from seed maturation to germination. Most of the vp mutants affect either synthesis or perception of abscisic acid (ABA) (Neill et al., 1986; Robichaud and Sussex, 1986; Tan et al., 1997). The viviparous 1 (vp1) mutant, which has an ABA insensitive phenotype (Robichaud and Sussex, 1986), encodes a complex transcription factor (McCarty et al., 1991; McCarty, 1995).

VP1 shares significant sequence similarity with the ABSCISIC ACID INSENSITIVE 3 (ABI3) factor of Arabidopsis (Giraudat et al., 1992). In addition to conferring ABA insensitivity, abi3 mutations disrupt normal seed development and produce green embryos during the seed maturation program (Finkelstein and Somerville, 1990; Giraudat et al., 1992; Nambara et al., 1992). Chlorophyll accumulation is normally observed during seed germination but not during seed maturation, suggesting abi3 and vp1 disrupt the timing of the seed maturation program similarly. Thus, the product of vp1 in maize and abi3 in Arabidopsis are likely to be key regulators in the integrated seed maturation and germination developmental program (McCarty, 1995). In recent years, a series of papers have revealed that ABI3 has functions not only in seed development but also in vegetative growth (Kurup et al., 2000; Robinson and Hill, 1999; Rohde et al., 2000; Rohde et al., 1999). The fact that overexpression of ABI3 causes several ectopic phenotypes (Parcy and Giraudat, 1997; Parcy et al., 1994; Tamminen et al., 2001) indicates a capability of ABI3 to interact unknown components on vegetative cells.

There are four prominent conserved domains in VP1 and ABI3, designated A1, B1, B2 and B3 (Giraudat et al., 1992). B3 is a DNA binding domain that is found in several other transcription factor families in plants (Suzuki et al., 1997; Ulmasov et al., 1997). Functional analyses have defined three discrete functions of VP1. First, VP1 is a co-activator of various seed specific ABA-regulated genes (McCarty et al., 1991; Vasil et al., 1995). Co-activation is strongly dependent on ABA signaling (Carson et al., 1997; Vasil et al., 1995) and is mediated principally by G-box related cis-elements. Co-activation of G-box coupled promoters does not require B3 and is mediated by physical interactions with other ABA-regulated G-box binding factors (Carson et al., 1997; Hobo et al., 1999). Second, VP1 is a direct activator of the maize C1 gene (Hattori et al., 1992; Kao et al., 1996; Suzuki et al., 1997). Activation of C1 is mediated by binding of B3 to the Sph element in the C1 promoter (Hattori et al., 1992; Kao et al., 1996; Suzuki et al., 1997). Third, VP1 is a repressor of germination specific gene expression in the aleurone of the maize endosperm (Hoecker et al., 1995, 1999). Like co-activation of ABA regulated genes, the repressor function of VP1 does not require B3, suggesting that repression is also mediated primarily by protein–protein interactions with other transcription factors (Hoecker et al., 1999).

Despite similarities of the four domains among VP1/ABI3 proteins, other regions of the proteins are highly divergent among plant species, especially between monocot and eudicot (Bobb et al., 1995; Chandler and Bartels, 1997; Giraudat et al., 1992; Hattori et al., 1994; McCarty et al., 1991). Therefore, it is unknown whether these genes are functional orthologs. Furthermore, the interaction of VP1 with ABA signaling has not yet been addressed well in planta although it has been analyzed extensively by transient assays. The ectopic expression of ABI3 revealed that it is able to interact with ABA signaling pathways in vegetative tissues (Parcy et al., 1994). However, expression of ABI3 driven by the 35S promoter in Arabidopsis appears to cause co-suppression in seeds (Parcy et al., 1994), making interpretation of seed phenotypes difficult in 35S-ABI3 transgenic Arabidopsis.

Here we introduced maize VP1 into abi3–6, a deletion mutant of abi3 (Nambara et al., 1994) in Arabidopsis to address whether VP1 of maize and ABI3 of Arabidopsis are functionally conserved. Furthermore, ectopic expression of VP1 revealed a novel interaction between auxin and ABA in the roots.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Analysis of the 35S-VP1 transgene in the abi3–6 mutant background

In order to determine whether maize VP1 could complement the abi3–6 mutation, a series of transgenic Arabidopsis lines carrying a cauliflower mosaic virus (CaMV) 35S-VP1 gene were constructed in a wild type background then crossed to abi3–6 mutant plants. Three independent lines of 35S-VP1 were generated by using the root culture method (Huang and Ma, 1992). We chose one of these lines, G4, for further analysis after confirming by Southern blot that the line had a single copy of the transgene (data not shown). We used the abi3–6 allele (Nambara et al., 1994) to cross with the G4 line because this allele is null for ABI3 protein due to a large internal deletion (Nambara et al., 1994; Nambara E., pers. comm.).

In order to fix the transgene in a homozygous abi3–6 mutant background, F3 seeds from two independent plants were plated on kanamycin (Km) plates and brown seeds were selected from kanamycin resistant (Kmr) plants showing approximately 3 : 1 segregation of the seed color (Table1) and retested for Kmr in the F4. The selected F3 lines were expected to be hemizygous for the transgene and homozygous for abi3–6. Uniformly Kmr lines were selected in the F4 and homozygosity for the abi3–6 deletion allele was confirmed by Southern blot hybridization using an ABI3 specific probe (data not shown). As a further genetic test, the presumptive 35S-VP1, abi3–6 plants were test-crossed with the abi3–6 mutant line and evaluated in the F1 and F2. As expected, the cross yielded all brown seeds in F1 and 3 : 1 segregation in the F2. From these experiments we concluded that a single copy of the 35S-VP1 transgene is sufficient to complement the green seed and desiccation intolerant seed phenotypes of the abi3–6 mutation (Figure 1).

Table 1.  Segregation of F3 seeds of 35S-VP1 transgenic Arabidopsis
  F3 selfcross
browngreenχ2
G4No. 1 No. 2 223 279 85 92 p=0.29 p=0.93
image

Figure 1.  The seeds of wt, abi3–6, and 35S-Vp1 plants.

Mature sliques were collected from Columbia wild type (a), abi3–6 (b), and 35S-Vp1 transgenic plant, G4 (c). The white bar represents 1 mm as a size marker.

Download figure to PowerPoint

To determine whether 35S-VP1 transformants could be selected directly by rescue of the desiccation tolerance seed phenotype of abi3–6, we transformed abi3–6 plants with the 35S-Vp1 transgene using the floral dip method (Clough and Bent, 1998). Seed from treated siliques were allowed to fully desiccate. Of 28 of 30 independent brown seed events produced Kmr plants that segregated approximately 3 : 1 brown : green seed in the T2. Two of these lines, V5 and V26, haboring a single copy and two copies of the transgene, respectively, as confirmed by Southern blot hybridization, were further characterized.

ABA sensitivity of the 35S-VP1 transgenic Arabidopsis on seed germination

Seed of severe abi3 mutations such as abi3–6 are completely insensitive to ABA inhibition of seed germination (Giraudat et al., 1992; Nambara et al., 1994). We determined whether 35S-VP1 restored ABA sensitivity during germination (Figure 2). As expected, germination of Columbia wild type seed was inhibited on to ABA containing germination media, whereas the abi3–6 seeds showed no sensitivity at ABA concentrations up to 50 µm. The three independent transgenes, however, fully restored the ABA sensitivity in the abi3–6 mutant background. All three transgenes conferred ABA dose response curves that were very similar to wild type suggesting that quantitative variation in VP1 expression had little influence on ABA sensitivity.

image

Figure 2.  ABA sensitivity of the seeds of Col (wild type), abi3–6, and three 35S-VP1 transgenic plants (G4, V5, V26) during seed germination.

The dry seeds from wild type and 35S-VP1 plants and the immature seeds from abi3–6, respectively, were collected and imbibed at 4°C for 4 days. The seeds were then germinated on the plates without ABA or with ABA at various concentration for 4 days under continuous light. Fifty seeds were plated on at each concentration and the radicle emergence was used as the maker of germinated seed.

Download figure to PowerPoint

Expression of VP1/ABI3 regulated genes during seed development

To evaluate complementation of ABI3 regulated gene expression in the developing seed, we analyzed the expression of VP1/ABI3 regulated genes in G4 transgenic line. First we conducted Northern analysis for a seed specific ABA regulated gene, cruciferin C (CRC), during seed development. Expression of the CRC gene normally requires ABI3 (Figure 3a, Nambara et al., 1995; Parcy et al., 1997), however, the amount of CRC transcript was not affected by VP1 expression during seed development even though VP1 transcript was detected throughout seed development (Figure 3). Second, we generated C1-β-glucronidase (GUS) transgenic Arabidopsis. The C1-GUS reporter gene is expressed during the late seed maturation stage in Arabidopsis, similar to the expression pattern of the C1 gene in maize (Figure 4a, McCarty et al., 1989). The C1-GUS transgene was then introduced into the abi3–6 and 35S-VP1/abi3–6 by crossing. The expression of C1-GUS during seed maturation is completely eliminated in the abi3–6 mutant. Expression of C1-GUS is partially restored in the presence of 35S-VP1, yeilding approximately 20% of wild type GUS activity at seed maturity. These results indicated a quantitative difference between 35S-VP1 and endogenous ABI3 expressed from the ABI3 gene.

image

Figure 3.  Expression of CRC in the developing seeds of Col (wild type), abi3–6, and 35S-VP1 transgenic plants G4.

The sliques of each plants were collected at 7, 9, 11, 13, 15 days after pollination, respectively, and total RNA was isolated from each sample.

(a) Six microgram of total RNA was loaded on agarose gel and Northern hybridization was conducted with CRC probe, VP1 probe and ABI3 probe after the samples were transferred onto nylon membrane. The membrane was exposed to X-ray film for 6 h for CRC probe and VP1 probe and for 4 days for ABI3 probe, respectively.

(b) The activity for each band of the membrane was measured by Phosphorimager and shown as relative acitivity with respect to Col 7DAP.

Download figure to PowerPoint

image

Figure 4.  Expression of C1-GUS and cab3-GUS in developing seeds. (a) C1-GUS expression during seed development. GUS activity was measured in extracts from developing seeds or matured seeds of C1-GUS, C1-GUS/abi3–6, and C1-GUS/abi3–6/G4 at indicated days after pollination. GUS activity was expressed as 4-methylumbelliferone (MU) production per h per seed.

(b) Cab3-GUS expression during seed development. GUS activity was measured in the extract from developing seeds or matured seeds of cab3-GUS, cab3-GUS/abi3–6, and cab3-GUS/abi3–6/G4 at indicated days after pollination. GUS activity was expressed as 4-methylumbelliferone (MU) production per h per seed.

Download figure to PowerPoint

Nambara et al. (1995) has reported that expression of the chlorophyl a/b binding protein 3 (cab3) is repressed during seed development and the repression is blocked in the abi3–6 mutant. To test whether VP1 complemented the repression function of ABI3, we evaluated the effect of 35S-VP1 on expression of the cab3-GUS transgene in the abi3–6 background. As shown in Figure 4(b), expression of cab3-GUS is repressed during seed development in wild type Arabidopsis seed, and the repression is suppressed by abi3–6, which is consistent with the previous report (Nambara et al., 1995). The repression of cab3-GUS is restored in the developing seeds of 35S-Vp1/abi3–6 seed, indicating that VP1 functions as a repressor in Arabidopsis. Although 35S-VP1 mRNA is present through seed development, cab3-GUS repression is first detected at 7 days after pollination, 2 days prior to the initial induction of C1-GUS in wild type and 35S-VP1/abi3–6 seeds at about 9 days. This suggests that the timing of repression and gene activation mediated by VP1 is differentially regulated in embryos. The wild type ABI3 gene is also transcriptionally active very early in embryo development (Parcy et al., 1994) indicating that post-transcriptional regulation of VP1 and ABI3 may be important in embryo development.

35S-VP1 suppresses the early flowering phenotype of abi3

Kurup et al. (2000) has reported that abi3 causes early flowering under short day and long day conditions. We measured flowering time of 35S-VP1 transgenic lines to see whether VP1 complements the flowering phenotype of the abi3 under long day conditions. Consistent with the results of Kurup et al. (2000) the abi3 mutant flowered 10 days earlier than wild type Col under our long day conditions (Table 2). The flowering time of V5 and V26 was comparable with that of col, indicating that 35S-VP1 complements the early flowering phenotype of abi3. The strongest expressor 35S-VP1 line, G4, prolonged flowering time relative to Col wild type, indicating that VP1 can alter the timing of the transition from vegetative development to reproductive development.

Table 2.  Flowering of col, abi3 and 35S-VP1 under long day condition (n=17–26)
 colabi335S-VP1
G4V5V26
Days of flowering44.9 ± 2.234.3 ± 2.657.9 ±3.942.6 ± 2.742.3 ±2.9
Number of leaves at flowering15.6 ± 1.113.1 ± 1.427.4 ± 3.116.2 ± 2.115.3 ± 1.9

35S-VP1 and 35S-ABI3 induce similar ectopic ABA responses in Arabidopsis

Ectopic expression of ABI3 has been previously reported to enhance ABA inhibition of root growth (Parcy et al., 1994). As shown in Figure 5, sub-micromolar concentrations of ABA promote growth of both wild type and transgenic roots, whereas micromolar concentrations of ABA inhibit growth. ABA sensitivity of abi3–6 roots was slightly less than wild type. In contrast, the 35S-Vp1/abi3–6 lines showed strongly enhanced sensitivity to exogenous ABA. This result is in qualitative agreement with the ectopic root phenotype conferred by ABI3 (Parcy et al., 1994).

image

Figure 5.  ABA sensitivity of root growth.

The 4-days-old seedlings from Col, abi3–6, and the three 35S-VP1 lines were transferred onto plates containing no ABA or with various concentrations of ABA. The root lengths were scored after 3 days in continuous light. The values are shown as a percentage of the wild type seedlings grown on the plate without ABA. Error bars indicate the standard error of the mean of 15–20 seedlings.

Download figure to PowerPoint

In vegetative tissues expression of 35S-ABI3 causes ectopic ABA induction of several seed-specific genes including CRC and Em (Parcy et al., 1994). To test whether VP1 has a similar effect, Northern analysis was performed for expression of CRC (Figure 6). Exogenous application of ABA (5–50 µm) caused strong induction of CRC in leaves of all three 35S-Vp1 lines whereas no CRC transcripts were detected in ABA treated leaves of wild type and abi3–6 (Figure 6a,b). The expression of the VP1 transgene was nearly unaffected by ABA treatment (Figure 6a). These results indicate that VP1 and ABI3 function similarly in vegetative tissues. Moreover, the transcripts for CRC are detected at relatively low levels of exogenous ABA, suggesting that the ectopic expression of CRC is induced by ABA in a physiologically relevant concentration range.

image

Figure 6.  Ectopic induction of CRC in leaves by exogenous ABA application in 35S-VP1 plants.

The 2-wk-old seedlings from Col, abi3–6, and the three 35S-VP1 lines were transferred onto plates containing no ABA or various concentrations of ABA as indicated. After the seedlings were grown for 48 h, total RNA was isolated from the leave tissue.

(a) Six microgram of total RNA was loaded on agarose gel and northern hybridization was conducted with the CRC probe and VP1 probe after the samples were transferred onto nylon membrane.

(b) The activity for each band of the membrane was measured by Phosphorimager and shown as the relative activity with respect to the G4 5 µm treatment.

Download figure to PowerPoint

35S-VP1 confers a novel interaction of ABA and auxin signaling in roots

We have previously shown that VP1 transactivation of C1-GUS in maize cells is restricted to embryo and endosperm tissues, whereas co-activation of the G-box coupled Em gene shows broad tissue specificity (Kao et al., 1996). Ectopic induction of C1-GUS was not detected in leaves of 35S-VP1 lines either in the presence or absence of exogenous ABA (data not shown). Interestingly, however, significant VP1 dependent induction of C1-GUS acitivity was observed in roots of ABA treated seedlings (Figure 7a). Figure 7(b) shows that the GUS stain was restricted to the distal region of the root elongation zone. The results indicate that activation of C1-GUS by VP1 is conditioned by ABA signaling and another unknown development signal specific to the root elongation zone of Arabidopsis plants.

image

Figure 7.  Ectopic induction of C1-GUS in roots of 35S-VP1 plants by exogenouly applied ABA.

(a) The 2-wk-old seedlings of C1-GUS and C1-GUS/abi3–6/G4 grown under continuous light were transferred onto the plates without ABA or with 50 µm ABA combined with without IAA or with 50 µm IAA and incubated continuously under dark for 48 h. GUS activity was measured in root extracts from each seedling and expressed as pmolMU per h per mg protein. Error bars indicate the standard error of the mean of seedlings for each treatment.

(b) For histochemical assay of GUS, the C1-GUS and C1-GUS/abi3–6/G4 plants were grown and incubated as described in (a) except that the seedlings were incubated with 5 µm ABA and/or 5 µm IAA as well. The seedlings were incubated in GUS staining buffer. The seedlings treated with 50 µm ABA and/or 50 µm IAA were incubated for 24 h and otherwise the seedlings were incubated for 48 h. The white bar in the control photo represents 1 mm.

Download figure to PowerPoint

Because auxin has been known to localize at root elongation zone (Costa and Dolan, 2000; Doerner, 2000), we have tested whether auxin affected the ABA-dependent activation of C1-GUS in 35S-VP1 plants. The GUS activity in the roots of the ABA treated seedlings was 2-fold enhanced in the presence of exogenous auxin (Figure 7a). In addition to GUS expression in the elongation zone, 50 µm auxin-treated roots showed GUS staining along the root axis, especially in the vascular tissue (Figure 7b). Interestingly, associated with the enhanced ABA dependent expression of C1-GUS by auxin, the auxin-induced lateral root formation was suppressed by the presence of ABA in 35S-VP1/C1-GUS but not in C1-GUS plants. The ABA inhibition of the auxin-induced lateral root formation and the enhanced ABA dependent expression of C1-GUS by auxin in 35S-VP1 plants were observed even at physiologically relevant concentration (see 5 µm ABA and 5 µm indole-3-acetic acid (IAA) treatment). These results suggest that auxin potentiates VP1 mediated ABA response in roots. Moreover, in the presence of ectopic VP1 expression ABA becomes an auxin antagonist.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We have shown that maize VP1 complements critical aspects of the abi3 mutant of Arabidopsis. The shape and color of the seed, desiccation tolerance, ABA sensitivity of seed germination, flowering time and repression of cab3 are restored by the expression of VP1 in the abi3 mutant. Furthermore, over-expression of VP1 causes a variety of ectopic ABA responses similar to those associated with over-expression of ABI3 (Parcy et al., 1994). On the other hand, expression of the CRC and C1-GUS markers was not fully restored by 35S-VP1 in the abi3–6 mutant during seed development. These results indicate that 35S-VP1 complements a subset of ABI3 functions in the Arabidopsis. VP1 function in Arabidopsis is evidently strongly conditioned by interactions with other tissue specific factors. Over-expression of VP1 also revealed a novel interaction between auxin and ABA mediated by VP1 in Arabidopsis root.

In the seed, failure to complement activation of CRC, and partial restoration of C1-GUS regulation by 35S-VP1 suggests a functional difference between ABI3 and VP1 in effecting gene activation. While we can not rule out the possibility that quantitative differences in VP1 expression relative to endogenous ABI3 are responsible for reduced activation of CRC and C1-GUS, the repression of cab3-GUS and ABA sensitivity of germination are fully restored in 35S-VP1/abi3–6 seeds indicating that any quantitative differences do not affect these functions.

The qualitative difference in complementation of CRC and C1-GUS activation and cab3-GUS repression indicates that these functions are distinct. In maize, distinct domains of VP1 are required for Em and C1 activation and α-amylase repression, respectively (Carson et al., 1997; Hoecker et al., 1995). A critical difference is that C1 activation requires B3, the conserved DNA binding domain of ABI3/VP1 proteins, whereas co-activation and repression functions are mediated by the more divergent N-terminal domain. Because repression of cab3 expression was fully complemented by 35S-VP1, the repressor domain of VP1 appears functionally well conserved in Arabidopsis despite being less conserved at the sequence level. Because the endogenous Arabidopsis ABI3 gene evidently can regulate maize C1-GUS effectively, ABI3 and any associated transcription factors properly recognize the cis regulatory elements in C1 promoter indicating that ABI3 binding to the Sph element is conserved in Arabidopsis. Hence, we do not attribute weak activation of C1-GUS by 35S-VP1 to DNA binding specificities of endogenous transcription factors that must bind to the C1 promoter in Arabidopsis. Rather it is more likely that the deficiency lies in VP1’s ability to interact with factors that complex with or regulate ABI3. This view is reinforced by the fact that an unrelated endogenous ABI3 regulated gene, CRC, is completely non-responsive to 35S-VP1 in seeds. This qualitative molecular phenotype indicates a specific deficiency in gene activation functions of VP1 in the developing seed that extends to native genes of Arabidopsis. It has not been determined whether CRC activation by ABI3 is mediated by the co-activator domain, direct DNA binding via B3, or a combination of both functions. Thus, it is unclear whether deficiency in a single function of VP1 relative to ABI3 can account for the deficiencies in C1-GUS and CRC regulation. Unlike C1-GUS activation, CRC expression in seeds is partially independent of ABI3 (Figure 3; Parcy et al., 1994). Expression for several ABA regulated genes including CRC is only moderately reduced in ABA deficient mutant backgrounds (Pang et al., 1988; Parcy et al., 1994), indicating a significant component of expression is independent of ABA signaling.

VP1 also complements ABI3 function in vegetative development. Arabidopsisabi3 causes early flowering under short day and long day conditions (Kurup et al., 2000). 35S-VP1 complements the early flowering phenotype of abi3. Strong overexpressor significantly prolonged flowering time relative to wild type, indicating that VP1 has a capacity to regulate the transition from vegetative development to reproductive development.

The ectopic effects of 35S-VP1 expression reveal a complex dependence on other developmental factors. In contrast to the situation in the developing seed, 35S-VP1 confers very strong ectopic ABA activation of CRC in vegetative tissues. Hence, the deficiency in VP1 regulation of CRC is apparently limited to the seed. This is consistent with the notion that VP1 lacks the ability to interact effectively with one or more seed specific partners of ABI3. We can not fully rule out the alternative possibility that endogenous ABA concentrations normally present in developing seeds are insufficient to cause induction in the presence of VP1. We note that induction of CRC mRNA in leaves is detected at exogenous ABA concentrations that are comparable (1–5 µm) with in vivo ABA levels present in siliques of Arabidopsis (Koornneef et al., 1984; Nambara et al., 2000).

In contrast to CRC, ectopic activation of C1-GUS is restricted to a narrow band in the distal region of the root elongation zone implying that C1-GUS regulation is limited by a different set of developmental signals. Because this pattern of ectopic induction does not resemble the specificity of the 35S promoter in roots (Benfey et al., 1990), we attribute this pattern to a developmental interaction with VP1 function rather than an effect due to regulation of the transgene. Overall, the tissue specificity of C1-GUS in Arabidopsis is consistent with the regulation of the C1 promoter in maize (Kao et al., 1996), where C1 expression is restricted to seed tissues (Cone et al., 1986). In transient expression assays, over-expression of VP1 is insufficient to transactivate C1-GUS in maize leaf cells (Kao et al., 1996).

Overall, these results are consistent with a model in which VP1 is able to interact with some, but not all proteins that interact with ABI3. Extensive genetic analyses in Arabidopsis have uncovered several transcription factors that may interact either directly or indirectly with ABI3 (Finkelstein, 1994; Parcy et al., 1997). These include LEC1 encoding a protein homologous to HAP3 (Lotan et al., 1998), FUS3 encoding a B3 protein (Luerssen et al., 1998), ABI4 encoding a protein with AP2 domain (Finkelstein et al., 1998) and ABI5 encoding a b-Zip protein (Finkelstein and Lynch, 2000). In particular, ABI5 is related to rice TRAB1, which interacts specifically with rice VP1 (Hobo et al., 1999) and DPBF-1, which binds to the embryo specific/ABA response element in Dc3 promoter of sunflower (Kim et al., 1997). Because these genes are expressed in seeds preferentially or specifically, these could be factors that contribute to the disruption of CRC expression in 35S-VP1 transgenic seed. In particular, LEC1, FUS3 and ABI3 were shown to co-ordinate CRC expression at seed development (Parcy et al., 1997). Moreover, CRC expression is not reduced by abi3–1, which has a point mutation in a conserved B3 that affects DNA binding affinity (Parcy et al., 1997; unpublished data MS and DRM), suggesting that the N-terminal domain of VP1 but not the B3 DNA binding domain is critical for regulation of CRC expression.

Overall, our results indicate that only a subset of ABI3 regulated gene expression is essential for the acquisition of desiccation tolerance and developmental arrest. The relative importance of gene repression and gene activation functions of VP1 and ABI3 in this respect is still unclear. The repressor activity of VP1 and endogenous ABI3 is manifest early in embryo development, several days prior to induction of CRC and C1-GUS in wild type and 35S-VP1 embryos (Figure 4). At least a subset of VP1/ABI3 activated gene expression is non-essential for maturation, whereas, in both maize (Carson et al., 1997; Hoecker et al., 1995) and Arabidopsis studies, integrity of the repressor function of VP1 is correlated with an arrested, desiccation tolerant embryo phenotype suggesting that gene repression may be essential for embryo survival. Importantly, 35S-VP1 does not cause co-suppression of the wild type ABI3 gene (data not shown), which can complicate interpretation of transgenic seed phenotypes (Parcy et al., 1994). In any case, the 35S-VP1 transgene and mutant derivatives that selectively remove functional domains provide a tool that can be used to more clearly delineate the subset of maturation specific gene regulation that is essential for embryo survival.

The pattern of ectopic induction of C1-GUS is correlated with a novel interaction of ABA and auxin conferred by VP1. In wild type roots, auxin treatment induces prolific lateral root initiation along the root axis and swelling of the root tip at the high concentration. This well-characterized auxin response of roots is not normally repressed by ABA. By contrast, ABA fully inhibits auxin induced lateral initiation of the 35S-VP1 transgenic lines. The ABA inhibition of the auxin stimulated lateral root initiation is associated with enhanced expression of the C1-GUS reporter along the root axis, indicating that the ABA response is in turn enhanced by auxin in the presence of VP1. These results indicate that VP1 is capable of mediating a specific interaction between auxin and ABA signaling networks.

The inhibition of lateral root initiation offers an intriguing model for developmental arrest that may provide an insight into the normal function of VP1/ABI3 in seed development. It remains unknown whether a VP1/ABI3 mediated interaction between auxin and ABA signaling also occurs during normal seed development. However, auxin is maintained at high levels during mid and late seed development, presumably to promote endoreduplication and expansion of endosperm cells in maize kernels (Jensen and Bandurski, 1994). The levels of ABA also increase late in seed development (Hole et al., 1989; Koornneef et al., 1984). Thus, VP1/ABI3 might cause developmental arrest by interacting with ABA and auxin signaling during seed maturation. While the two groups independently have shown components mediating interactions of ABA and ethylene (Ghassemian et al., 2000; Beaudoin et al., 2000), no transcription factors mediating both ABA and auxin signaling have been found so far. The auxin resistant mutant in Arabidopsis, axr2 (Nagpal et al., 2000; Wilson et al., 1990), might be a candidate for mediating an ABA and auxin interactions because the mutant is also resistant to ABA and ethylene (Wilson et al., 1990).

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Arabidopsis transformation

The root transformation was performed as previously described (Huang and Ma, 1992). Transformation by the floral dip method was performed as described (Clough and Bent, 1998). Transformants were confirmed by selection on Kanamycin plates (see Results).

Plasmid construction

Restriction fragments containing full-length CaMV 35S promoter and Vp1 coding sequences were isolated from the 35S-VP1 plasmid (McCarty et al., 1991) and introduced separately into the pBI101 vector (Clonetech, Inc., Palo Alto, CA, USA). The structure of the resulting CaMV 35S-VP1 gene construct was confirmed by restriction digestion and DNA sequencing.

Generation of cab3-GUS and C1-GUS transgenic Arabidopsis

Cab3-GUS transgenic Arabidopsis seeds were a generous gift of Dr Joanne Chory of the Salk Institute (La Jolla, CA, USA). C1-GUS transgenic Arabidopsis plants were generated by C.-Y. Kao (manuscript in preparation). To place cab3-GUS and C1-GUS in the abi3–6 mutant background, transgenic plants were crossed with abi3–6 homozygotes. F2 lines homozygous for the transgene and the abi3–6 mutation were derived by simultaneous selection for Kanamycin resistance and the green seed phenotype. The resulting cab3-GUS/abi3–6 or C1-GUS/abi3–6 transgenic lines were crossed with 35S-VP1/abi3–6 (line G4) to generate cab3-GUS/abi3–6/35S-VP1 and C1-GUS/abi3–6/35S-VP1 plants. The expected genotypes were confirmed by scoring segregation for brown and green seeds and by Southern blots probed with VP1 and ABI3 specific probes.

Northern analysis

Total RNA was prepared as previously described (Verwoerd et al., 1989). Northern analysis was performed using standard methods with high stringency washes.

Growth condition for flowering

The seeds of col, abi3, and three 35S-VP1 lines were sterilized and then planted onto soil. The seeds were germinated and grown until the flower buds were initiated from the plants under long day condition (16 h-light/8 h-dark) in our growth room.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We thank Dr Joanne Chory for providing cab3-GUS transgenic Arabidopsis seeds and Dr Eiji Nambara for abi3–6 and CRC gene. This work was supported by National Science Foundation (Awards 9724000 and 0080175) to D.R.M. and by the Florida Agricultural Experiment Station (Journal series R-08365).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References