Disruption and overexpression of auxin response factor 8 gene of Arabidopsis affect hypocotyl elongation and root growth habit, indicating its possible involvement in auxin homeostasis in light condition


  • Chang-en Tian,

    1. Division of Biological Sciences, Graduate School of Science, Hokkaido University, Sapporo 060-0810, Japan,
    2. Division of Biological Sciences, Graduate School of Environmental Earth Science, Hokkaido University, Sapporo 060-0810, Japan, and
    Search for more papers by this author
    • Present address: Department of Bioengineering, Guangzhou University, Guangzhou 510405, China.

  • Hideki Muto,

    1. Division of Biological Sciences, Graduate School of Science, Hokkaido University, Sapporo 060-0810, Japan,
    2. Division of Biological Sciences, Graduate School of Environmental Earth Science, Hokkaido University, Sapporo 060-0810, Japan, and
    Search for more papers by this author
  • Kanako Higuchi,

    1. Department of Biological Sciences, Graduate School of Science, Tokyo Metropolitan University, Hachioji, Tokyo 192-0397, Japan
    Search for more papers by this author
  • Tomoyuki Matamura,

    1. Division of Biological Sciences, Graduate School of Environmental Earth Science, Hokkaido University, Sapporo 060-0810, Japan, and
    Search for more papers by this author
  • Kiyoshi Tatematsu,

    1. Division of Biological Sciences, Graduate School of Environmental Earth Science, Hokkaido University, Sapporo 060-0810, Japan, and
    Search for more papers by this author
    • Present address: Plant Science Center, RIKEN, Suehirocho 1-7-22, Tsurumi-ku, Yokohama 230-0045, Japan.

  • Tomokazu Koshiba,

    1. Department of Biological Sciences, Graduate School of Science, Tokyo Metropolitan University, Hachioji, Tokyo 192-0397, Japan
    Search for more papers by this author
  • Kotaro T. Yamamoto

    Corresponding author
    1. Division of Biological Sciences, Graduate School of Science, Hokkaido University, Sapporo 060-0810, Japan,
    2. Division of Biological Sciences, Graduate School of Environmental Earth Science, Hokkaido University, Sapporo 060-0810, Japan, and
      (fax +81 11 706 2739; e-mail kty@sci.hokudai.ac.jp).
    Search for more papers by this author

(fax +81 11 706 2739; e-mail kty@sci.hokudai.ac.jp).


Auxin response factor (ARF) family genes play a central role in controlling sensitivity to the plant hormone auxin. We characterized the function of ARF8  in Arabidopsis by investigating a T-DNA insertion line (arf8-1) and overexpression lines (ARF8 OX) of ARF8. arf8-1 showed a long-hypocotyl phenotype in either white, blue, red or far-red light conditions, in contrast to ARF8 OX that displayed short hypocotyls in the light. Stronger and weaker apical dominance, and promotion and inhibition of lateral root formation were observed in arf8-1 and ARF8 OX respectively. Sensitivity to auxin was unaltered in arf8-1 hypocotyls with respect to growth inhibition caused by exogenously applied auxin and growth promotion induced by higher temperatures. ARF8 expression was observed constitutively in shoot and root apexes, and was induced in the light condition in hypocotyls. Free IAA contents were approximately 30% reduced in light-grown hypocotyls of ARF8 OX, but were similar between those of arf8-1 and wild type. Expression of the three GH3 genes was reduced in arf8-1 and increased in ARF8 OX, indicating that they are targets of ARF8 transcriptional control. Because the three GH3 proteins may be involved in the conjugation of IAA as suggested by Staswick et al. (2002), and because two of the three GH3 genes are auxin inducible, ARF8 may control the free IAA level in a negative feedback fashion by regulating GH3 gene expression. ARF family genes seem to control both auxin sensitivity and homeostasis in Arabidopsis.


Since its discovery in the 1920s the plant hormone auxin has played a central role in the growth and development of plants. However, its mechanism of action at the molecular level remains largely unknown. How auxin could affect such a wide range of developmental processes has been especially puzzling. This situation has been drastically changed by the finding of a gene family of auxin response factors (ARFs) in Arabidopsis (Kim et al., 1997; Ulmasov et al., 1997). The founding member of the ARF family was identified as a protein that binds to the auxin-responsive cis-acting element (AuxRE) derived from promoters of an auxin-responsive gene family, GH3 (Ulmasov et al., 1997). Most ARF proteins consist of an N-terminal DNA-binding domain, a middle region and a carboxy terminal domain that is responsible for a protein-to-protein interaction. They activate or repress transcription of downstream genes, depending on the amino acid sequence of their middle region (Hagen and Guilfoyle, 2002).

The Arabidopsis genome contains 23 ARF genes, which appear sufficient to explain a plethora of physiological characteristics of auxin. Currently, however, we only know the physiological functions of three of them from the phenotypes of their loss-of-function mutants: ETTIN (ETT)/ARF3 is necessary for auxin-dependent pattern formation of the gynoecium (Nemhauser et al., 2000; Sessions et al., 1997); MONOPTEROS (MP)/ARF5 functions in vascular development, in both patterning and differentiation (Berleth and Jürgens, 1993; Hardtke and Berleth, 1998; Przemeck et al., 1996); and NON-PHOTOTROPIC HYPOCOTYL 4 (NPH4)/ARF7 operates in differential growth responses of hypocotyls such as phototropism, gravitropism and apical hook maintenance (Harper et al., 2000; Liscum and Briggs, 1995; Watahiki and Yamamoto, 1997). In either case, the loss-of-function mutant exhibits reduced sensitivity to auxin, which results in defects in the specific auxin-dependent phenomenon. Furthermore, the function of NPH4 overlaps partially with that of MP because nph4 mp double mutants exhibit more severe defects in vascular formation than those observed in single mp mutants (Hardtke et al., 2004). The distinct and overlapping functions of the three genes strongly suggest that the diverse array of functions characteristic of auxin is produced by various combinations of 23 ARF genes. Thus, it is imperative to study the function of other uncharacterized ARF genes for a better understanding of auxin-specific actions at the molecular level.

The ARF functions together with a transcriptional repressor, Aux/IAA protein, which shares a carboxy terminal domain with ARF (Kim et al., 1997; Liscum and Reed, 2002; Rouse et al., 1998; Tiwari et al., 2004). Aux/IAA proteins can heterodimerize with ARFs through the carboxy terminal domain, resulting in repression of ARF transcriptional activity. Many of the Aux/IAA genes are primary response genes of auxin (Theologis et al., 1985), whose transcription is promoted by auxin through AuxRE in their promoter sequence. The interaction between ARFs and Aux/IAAs thus constitutes negative feedback regulation of auxin, which probably enables a tight developmental control of auxin responses. This mechanism is best illustrated by the interaction between NPH4/ARF7 and MSG2/IAA19, which is responsible for auxin-mediated tropic responses in hypocotyls (Tatematsu et al., 2004).

In the present study, we carried out a functional characterization of the ARF8 gene, which is closely related to NPH4/ARF7, by studying a T-DNA insertion mutant of Arabidopsis, arf8-1, and overexpression lines of ARF8 (ARF8 OX). We found that arf8-1 and ARF8 OX display long- and short-hypocotyl phenotype, respectively, in the light condition. The present data suggest that the phenotype may result from a change in the level of auxin rather than a change in sensitivity to auxin, and that ARF8 controls the auxin level in a negative feedback manner by transcriptional regulation of the GH3 family genes (Hagen and Guilfoyle, 2002; Nakazawa et al., 2001; Staswick et al., 2002). We present data to support that the ARF family proteins might play a central role in auxin-triggered responses by regulating both the level of auxin and the sensitivity to auxin in plants.


A T-DNA insertion line of ARF8 gene

A T-DNA insertion line of ARF8 was isolated in a Wassilewskija (Ws) background, which we designated arf8-1. A T-DNA was found to be inserted into the third exon of ARF8 in arf8-1 (Figure 1a). An RNA gel blot analysis using a probe corresponding to exons 1–5 showed that arf8-1 produced a reduced amount of two mRNAs which were shorter than the full-length ARF8 mRNA (Figure 1b, arrowhead). We determined the nucleotide sequences of the cDNAs of the transcripts by RT-PCR and 5′-RACE, and found that the shorter mRNA was likely to originate from the authentic ARF8 promoter, and that it was terminated in the T-DNA. This mRNA was predicted to encode the N-terminal 63 amino acid residues of ARF8 protein fused with 11 amino acid residues derived from the T-DNA sequence. The longer mRNA, on the contrary, originated 673 bp inside the T-DNA and appeared to be terminated by the natural ARF8 terminating signal. It was predicted to encode an 87 amino acid residue-long chimeric protein derived from agropine synthesis reductase (MAS1) and agrocinopine synthase, both of which are T-DNA proteins.

Figure 1.

T-DNA insertion line and overexpression lines of ARF8.
(a) ARF8 gene structure and T-DNA insertion sites in arf8-1.
(b) RNA gel blot analysis of ARF8 gene in 3-day-old light-grown seedlings using a part of ARF8 cDNA from exon 1 to 5 as a probe. Two micrograms of poly(A)+ RNA was subjected to electrophoresis in each lane. β-Tubulin is used as a loading control. Numbers indicate the size in nucleotides of ribosomal RNA.
(c) Hypocotyl length of 6-day-old seedlings grown under continuous white light at 23°C. Values are mean ± SEM of three independent experiments, in which 18–22 seedlings were measured. arf8-1; pARF8:ARF8: arf8-1 transformed with pARF8:ARF8, 35S:ARF8: ARF8 overexpression line.
(d) Six-day-old seedlings grown in white light (left) or darkness (right). From left to right: Ws, arf8-1, arf8-1; pARF8:ARF8, and 35S:ARF8. Bars = 5 mm.

arf8-1 seedlings display longer hypocotyls in the light condition, although hypocotyl growth was not affected in the dark (Figure 1c,d). Hypocotyls of the heterozygous arf8-1 seedlings were slightly (approximately 15%), but significantly longer than those of Ws (P = 0.017 in t-test) (Figure 2), and the progeny of heterozygous plants showed a segregation of the long-hypocotyl phenotype of approximately 1:3 (χ2 = 0.08, P > 0.70), indicating that the arf8-1 mutation is a single, semidominant mutation. Transformation of arf8-1 plants with a chimeric gene, which consisted of the ARF8 promoter-containing genomic fragment and a 3′ portion of ARF8 cDNA (pARF8:ARF8), almost restored its long-hypocotyl phenotype. However, the hypocotyl length of the transgenic plants was slightly longer than that of Ws (P = 0.014 in t-test), and was essentially the same as that of the arf8-1 heterozygous seedlings (P = 0.49 in t-test) (Figure 2). As we observed mRNA encoding a truncated ARF8 protein in arf8-1, arf8-1 may not be a perfect loss-of-function mutation. To better understand the genetic basis of arf8-1, we examined the hypocotyl length of seedlings which were obtained by crosses among Ws, arf8-1 and the transgenic plant pARF8:ARF8. These lines contained gene dosages of the wild type and arf8-1 alleles between 0 and 2. The results in Figure 2 show that hypocotyl elongation was promoted almost in proportion to the arf8-1 dosage at the fixed dosage of the wild-type allele, although the increment caused by arf8-1 was much smaller than that caused by the loss of the wild-type allele. Similarly, hypocotyl elongation was more inhibited in proportion to ARF8 dosage when the arf8-1 dosage was fixed at 1 or 2. These results imply that ARF8 inhibits hypocotyl elongation, and that the long-hypocotyl phenotype of arf8-1 mostly results from loss-of-function of the ARF8 gene. However, these results also show that arf8-1 indeed has weak promotive effects on hypocotyl growth. To confirm this, we determined the hypocotyl length of ARF8 OX which harbored ARF8 cDNA driven by the cauliflower mosaic virus (CaMV) 35S promoter (Figure 1b). It exhibited shorter hypocotyls than Ws (Figure 1c,d), which is consistent with the elongation-repressing function of the ARF8 gene in hypocotyls.

Figure 2.

Dosage effects of ARF8 and arf8-1 on elongation of hypocotyl.
Values are mean ± SEM of three independent experiments, in which hypocotyl length was measured for 18–22 seedlings grown for 6 days under continuous white light at 23°C. arf8-1; pARF8:ARF8: arf8-1 transformed with pARF8:ARF8.

Morphological phenotype of arf8-1

arf8-1 seedlings showed longer hypocotyls only in the light condition (Figures 1d and 3). Fluence-rate response curve analyses showed that arf8-1 hypocotyl length increased in a dose-dependent manner when expressed relative to the length of Ws hypocotyls (Figure 3a). arf8-1 hypocotyls were longer than those of Ws when grown in either white, far-red, red or blue light (Figure 3b), indicating that ARF8 functions under the control of phytochromes and cryptochromes. The growth habit of roots was also affected by arf8-1 and ARF8 OX: roots of arf8-1 as well as Ws (Rutherford and Masson, 1996) grew in a slanting manner when grown on vertically oriented plates (Figure 4a, left and middle). However, ARF8 OX roots grew almost downward (Figure 4a, right). Furthermore, formation of lateral roots was slightly but significantly promoted in arf8-1 than in Ws at an earlier stage of seedling development, and was clearly suppressed in ARF8 OX (Figure 4b). In the mature stage, arf8-1 and ARF8 OX displayed stronger and weaker apical dominance than Ws (Figure 4c). arf8-1 also showed a little lower fecundity, and the rosette leaves of ARF8 OX were wrinkled in contrast to the rather flat surface of Ws leaves. Except for these differences, the morphologies of arf8-1 and ARF8 OX were similar to those of the wild type.

Figure 3.

Fluence-rate response curves of hypocotyl elongation of arf8-1.
(a) Effects of white light condition.
(b) Effects of monochromatic far-red (top), red (middle) and blue (bottom) light conditions. Ws (open circle) and arf8-1 (closed circle) seedlings were grown for 6 days under continuous light conditions at 23°C. Values are mean ± SEM of three independent experiments, in which 18–22 seedlings were measured.

Figure 4.

Morphology of arf8-1 and the ARF8 overexpression line grown in the white light condition.
(a) Eight-day-old light-grown seedlings grown on vertically oriented agar plates. From left to right, Ws, arf8-1 and ARF8 OX.
(b) Formation of lateral roots of Ws (open circle), arf8-1 (closed circle) and ARF8 OX (triangle) observed after germination. Data represent mean ± SEM of five independent experiments, in which nine to 11 seedlings were observed.
(c) Number of inflorescence stems of 2-month-old plants. Data represent mean ± SD of five to 10 plants.

ARF8 is expressed mainly in hypocotyls of light-treated seedlings

Tissue specificity of ARF8 gene expression was examined using an ARF8 promoter:GUS reporter fusion gene containing the 2.1 kb-long 5′ upstream region of the ARF8 gene (Figure 5b,c). GUS staining was observed in shoot and root meristems irrespective of the light condition. However, ARF8 was strongly expressed in hypocotyls only in the light condition. The light-induced expression was seen in either blue, red or far-red light (Figure 5d). The results of semi-quantitative RT-PCR showed that ARF8 mRNA level of hypocotyls was nearly 100 times higher in the light than in the dark (Figure 5a). These results are consistent with the abnormal growth of hypocotyls observed only in the light condition. We also examined the change in ARF8 gene expression when seedlings were brought into the light condition from the dark (Figure 5e). GUS staining was detected only in the upper portions of hypocotyls of light-treated seedlings, suggesting that ARF8 was expressed only in the cells that started to elongate after the transfer to the light.

Figure 5.

Tissue-specific and light-induced expression of ARF8 gene.
(a) Semi-quantitative RT-PCR of ARF8  with total RNA isolated from hypocotyls of 6-day-old dark- or light-grown seedlings.
(b) GUS staining of wild-type seedlings transgenic for the pARF8:GUS fusion. Seedlings were grown in the dark (left) or under white-light condition (right) for 3 days at 23°C.
(c) Close-up view of upper portions of seedlings in (b).
(d) Effects of monochromatic light irradiations on ARF8 expression. From left to right: seedlings were grown under white (12 W m−2), blue (200 μmol m−2sec−1), red (72 μmol m−2 sec−1) or far-red (400 μmol m−2 sec−1) light condition for 3 days at 23°C.
(e) Effects of white-light treatments on ARF8 expression. From left to right: seedlings grown in the dark for 72, 66, 48 and 24 h, then treated with white-light for 0, 6, 24 and 48 h respectively. Seedlings were grown hydroponically at 23°C.
Bars = 2 mm in (b), (d) and (e), and 0.2 mm in (c).

Auxin sensitivity is not changed in arf8-1

Because ARF8 is closely related to ARF7/NPH4 which controls auxin sensitivity in hypocotyls (Stowe-Evans et al., 1998; Watahiki and Yamamoto, 1997), we determined auxin sensitivity in arf8-1 and ARF8 OX. We found that the growth of hypocotyls of both lines was inhibited by exogenous IAA in the dark to a similar extent to Ws (Figure 6). We also checked whether hypocotyl growth is affected by temperature in the light condition in arf8-1 as higher temperatures increase the level of auxin in Arabidopsis (Gray et al., 1998). Ws hypocotyls grew 1.42 ± 0.21-fold (n = 3) longer at 29°C than at 23°C at a light fluence rate of 16 W m−2, and a similar increase (1.38 ± 0.15-fold, n = 3) was observed in arf8-1. These results indicate that auxin sensitivity of hypocotyls is not influenced by disruption and overexpression of the ARF8 gene.

Figure 6.

Effects of exogenous IAA on elongation of arf8-1 (left) and ARF8 OX (right) hypocotyls.
After induction of germination, seedlings of Ws (open circle) and arf8-1 or ARF8 OX (closed circle) were grown in the presence of various concentrations of IAA for 5 days in the dark at 23°C. Data represent mean ± SD of about 20 seedlings.

Next, we checked gravitropism and phototropism of hypocotyls of arf8-1 and ARF8 OX as defects in tropistic responses are often associated with altered sensitivity to auxin. As shown in Figure 7(a,b), the gravitropic response of arf8-1 hypocotyls was not significantly different from that of the wild type (P = 0.33 at 24 h in t-test), and the phototropic response was reduced by only 20% from the wild-type level (P = 0.02 at 24 h in t-test). Furthermore, ARF8 OX hypocotyls exhibited essentially the same responses as those of Ws in both gravitropism and phototropism (P = 0.08 at 24 h in t-test in gravitropism) (Figure 7c,d).

Figure 7.

Tropic responses of arf8-1 (a and b) and ARF8 OX (c and d) hypocotyls.
Gravitropic (a and c) and secondary positive phototropic responses (b and d) were observed in 3-day-old etiolated hypocotyls of Ws (open circle) and arf8-1 or ARF8 OX (closed circle). Each data represent mean ± SEM of four to six (a and b) or three (c and d) independent experiments, where about 17 seedlings were measured.

Free auxin contents were decreased in the ARF8 overexpression line

As no significant changes were detected in auxin sensitivity in arf8-1, we next determined the content of free IAA in hypocotyls of 6-day-old light-grown arf8-1 and ARF8 OX with gas chromatography single-ion-monitoring mass spectrometry. As shown in Table 1, the level of free IAA was significantly lower in hypocotyls of ARF8 OX than of Ws (P = 0.007 in t-test), but the level was similar between Ws and arf8-1.

Table 1.  Free IAA levels in hypocotyls of 6-day-old light-grown seedlings
GenotypeIAA contents (ng g−1 fresh weight)
  1. Values are mean ± SEM of four independent experiments, in which three measurements of 50–100 hypocotyls were carried out.

Ws5.02 ± 0.39
arf8-15.41 ± 0.20
35S:ARF83.96 ± 0.36

GH3 family genes may be targets of ARF8 transcription factor

The short-hypocotyl phenotype of ARF8 OX resembled the phenotype of the overexpression lines of DFL1 (Nakazawa et al., 2001), a member of the GH3 primary auxin-responsive gene family (Hagen and Guilfoyle, 2002; Hagen et al., 1984). The ArabidopsisGH3 gene family has 19 members divided into three subgroups (groups I, II and III) according to their sequence similarity (Staswick et al., 2002). More importantly, Staswick et al. (2002) have recently shown that the GH3 proteins have an enzymatic activity in vitro for adenylation of IAA, salicylic acid and/or jasmonic acid, and that the group II proteins including DFL1, AtGH3a (Tanaka et al., 2002) and YDK1 (Takase et al., 2004) react with IAA. They have further proposed that the group II proteins may be enzymes involved in the conjugation of IAA with amino acids such as Lys. These recent findings prompted us to examine whether the GH3 genes were targets of ARF8 by measuring mRNA levels of the eight group II genes in arf8-1 and ARF8 OX. RNA gel blot analyses did not show any signals for the four group II genes, YDK1, At1g59500, At2g14960 and At2g47750 (data not shown). However, mRNA levels of DFL1, AtGH3a and At1g28130 decreased by approximately 25–50% in arf8-1, and increased by approximately 50% in ARF8 OX compared with those of Ws; the level of At2g23170 was, however, not influenced by the ARF8 gene (Figure 8). We also examined the mRNA level of the group I genes, JAR1/FIN219 (Hsieh et al., 2000; Staswick et al., 1992, 2002) and DFL2 (Takase et al., 2003). JAR1 reacts with jasmonic acid in vitro (Staswick et al., 2002). Interestingly, the mutations in ARF8 did not affect the expression of either gene. We further measured the mRNA level of SUR2 which encodes cytochrome P450 CYP83B1 involved in auxin homeostasis (Barlier et al., 2000), and found that SUR2 expression was not dependent on ARF8. These results indicate that at least three group II genes of the GH3 family including DFL1 and AtGH3a may be targets of the ARF8 transcription factor.

Figure 8.

RNA gel blot analysis of GH3 family genes and SUR2.
One microgram of poly(A)+ RNA isolated from 3-day-old light-grown seedlings was subjected to electrophoresis in each lane. β-Tubulin was used as a loading control.


ARF8 may be involved in regulation of auxin level

Here we show that arf8-1, a T-DNA insertion line of ARF8, has long hypocotyls, and that an overexpression line of ARF8, ARF8 OX, has short hypocotyls. In either case, the aberrant phenotype is observed only in the light condition. Growth of Arabidopsis hypocotyls has been intensively studied as a model system of elongation growth as their epidermal and cortical cells grow simply by cell elongation without cell division (Gendreau et al., 1997). Hypocotyl growth is influenced by both environmental and hormonal factors (Collett et al., 2000). The former includes light (Fankhauser and Chory, 1997), temperature (Gray et al., 1998) and mechanical stimulation (Johnson et al., 1998), while the latter includes auxin as well as other plant hormones such as gibberellin (Richards et al., 2001), brassinosteroid (Fankhauser and Chory, 1997) and ethylene (Smalle et al., 1997). In the case of auxin, changes in both auxin level and sensitivity are known to affect hypocotyl growth, especially in the light condition. Free IAA contents are increased by overexpression of either the bacterial Trp monooxygenase gene (iaaM; Romano et al., 1995), ArabidopsisYUCCA gene which encodes flavin monooxygenase (Zhao et al., 2001), or the cytochrome P450 CYP79B2 gene (Zhao et al., 2002). The latter two genes encode enzymes in Trp-dependent auxin biosynthesis in Arabidopsis. IAA contents are also increased by mutations in SUR1/ALF1 (Boerjan et al., 1995; Celenza et al., 1995) or SUR2/RNT1 genes (Bak and Feyereisen, 2001; Barlier et al., 2000; Delarue et al., 1998), or by incubation at higher temperatures (Gray et al., 1998). In any case, Arabidopsis exhibits the long-hypocotyl phenotype in the light condition. On the other hand, IAA level is reduced by overexpression of iaaL, whose product conjugates IAA to Lys (Jensen et al., 1998), or by disruption of both CYP79B2 and CYP79B3 genes (Zhao et al., 2002). Hypocotyls of these lines are shorter than those of the wild type. Growth inhibition of hypocotyls by treatment with phytotropin 1-naphthylpthalamic acid, which is observed only in the light condition, might also be caused by reduction in IAA contents in hypocotyls due to a decrease in the flow of auxin from the shoot apex (Jensen et al., 1998). Furthermore, short hypocotyls in the light condition can also be caused by changes in the sensitivity to auxin by mutations in auxin signaling components such as AXR1 (Leyser et al., 1993) and AXR3/IAA17 (Leyser, 2002; Rouse et al., 1998), which decrease and increase the sensitivity to auxin respectively (Collett et al., 2000).

In this context, the long-hypocotyl phenotype of arf8-1 might result from a decrease in auxin sensitivity as a related transcription factor ARF7/NPH4 controls auxin sensitivity in tropic responses (Stowe-Evans et al., 1998; Watahiki and Yamamoto, 1997). Therefore, we examined auxin sensitivity in arf8-1 and ARF8 OX by measuring growth inhibition of hypocotyls by exogenously applied auxin in the dark. No significant changes were detected in seedlings of arf8-1 and ARF8 OX. We also checked promotion of hypocotyl growth by higher temperatures (Gray et al., 1998) in arf8-1, and found that the mutant responded to temperature increase in a similar fashion to wild type. Although it is not simple to interpret the effects of temperature on hypocotyl growth because of complex regulation of IAA synthesis by temperature (Zhao et al., 2002), this result may suggest unaltered sensitivity of arf8-1 hypocotyls to auxin. Furthermore, tropic responses of hypocotyls were not so much affected in arf8-1 and ARF8 OX, although they are severely affected in arf7/nph4 (Liscum and Briggs, 1995; Watahiki and Yamamoto, 1997). We next determined the free IAA contents of hypocotyls and found that the auxin content in ARF8 OX hypocotyls is lower than those of the wild type, but that it is similar between arf8-1 and the wild type. In fact, all the other phenotypes of ARF8 OX examined here are consistent with lower auxin contents: lower apical dominance and suppression of lateral root formation. These characteristics are known to be induced by reduced auxin levels (Boerjan et al., 1995; Celenza et al., 1995; Delarue et al., 1998).

The short-hypocotyl phenotype observed in ARF8 OX is also associated with overexpression of DFL1, DFL2, and YDK1, GH3 family genes of Arabidopsis (Nakazawa et al., 2001; Takase et al., 2003, 2004). All the auxin-inducible GH3 genes except for JAR1/FIN219 are classified as group II genes whose products can adenylate IAA in vitro (Staswick et al., 2002). The mRNA level of three group II genes –DFL1, AtGH3a and At1g28130 – decreased in arf8-1 and increased in ARF8 OX, indicating that they are expressed through ARF8. DFL1 (Nakazawa et al., 2001) and AtGH3a (Tanaka et al., 2002) are auxin-inducible, while gene expression of At1g28130 is not regulated by auxin (our unpublished data; Hagen and Guilfoyle, 2002).

Taking the findings of Staswick et al. (2002) into account, the above results may explain how overexpression of ARF8 causes short hypocotyls and suppression of lateral root formation. If adenylation of IAA by the three above-mentioned GH3 proteins leads to conjugation of IAA with amino acids as suggested by Staswick et al. (2002), this may cause a decrease in IAA contents in ARF8 OX hypocotyls and roots, leading to auxin-depleted phenotypes such as short hypocotyls and suppression of lateral root formation. Although IAA levels were reduced in ARF8 OX hypocotyls, we did not detect significant differences between those of wild-type and arf8-1 hypocotyls. According to the above-mentioned assumption, a decreased expression level of the GH3 genes observed in arf8-1 should lead to an increase in the IAA level, which may result in promotion of hypocotyl elongation. IAA levels may have increased in arf8-1 hypocotyls, but the increase may have been too small to detect. The levels of free IAA in the light condition are increased by about fourfold in the case of overexpression of iaaM (Romano et al., 1995), about 1.6-fold for YUCCA overexpression (Zhao et al., 2001), about 1.3-fold for CYP79B2 overexpression (Zhao et al., 2002), and about 10-fold in the case of loss-of-function mutations in SUR2/RNT1 (Delarue et al., 1998). These increases in free IAA result in two to fivefold increases in hypocotyl elongation. In most cases hypocotyl growth is inhibited even in the dark, and apical dominance is strengthened. The degree of aberration in the phenotypes including increased hypocotyl growth was smaller in arf8-1 than in the above studies, suggesting that the change in the IAA contents would be difficult to detect in arf8-1.

Although a possible involvement of group II GH3 proteins in the conjugation of IAA is solely based on their enzymatic activities in vitro (Staswick et al., 2002), our results suggest the presence of feedback regulation of the IAA level through transcriptional control of GH3 expression by ARF8, as most of the group II GH3 genes are auxin-inducible. In contrast to ARF3, 5 and 7 that regulate the sensitivity to auxin in morphogenesis and tropic responses (Mattson et al., 2003; Nemhauser et al., 2000; Stowe-Evans et al., 1998; Watahiki and Yamamoto, 1997), ARF8 appears to be specifically involved in the regulation of auxin level. It is worth mentioning that another mechanism of negative feedback regulation of auxin level has been reported. SUR2, which encodes CYP83B1 (Barlier et al., 2000), negatively regulates auxin production in Arabidopsis by catalyzing the first committed step in indole glucosinolate biosynthesis which decreases flux from Trp to IAA. As SUR2 is also auxin-inducible, having AuxRE in its promoter (Barlier et al., 2000), it may also participate in the negative feedback regulation of auxin level (Bak and Feyereisen, 2001; Barlier et al., 2000). We show that SUR2 expression is not influenced by ARF8 mutations, indicating the presence of multiple independent feedback mechanisms for auxin homeostasis.

ARF8 gene expression is affected by light

Although ARF8 is expressed constitutively in shoot and root apical regions, ARF8 gene expression is strongly induced in hypocotyls in the light condition. However, ARF8 expression is not induced by light in portions of the hypocotyl that have already elongated in the dark; ARF8 seems to be light-inducible only in hypocotyl cells that start to elongate in the light condition. The light-inducible expression of ARF8 is consistent with our observation that elongation defects of arf8-1 hypocotyls are observed only in the light. The gene expression profile of ARF family genes has not been examined comprehensively. The best studied genes among them are ARF7/NPH4 and ARF5/MP. ARF7 expression is observed ubiquitously, and is generally stronger in the small, non-vacuolated cells of young organs; the expression domain of ARF5 is more restricted to the central regions of various organs (Hardtke et al., 2004). Thus, ARF8 is the first ARF gene whose expression is regulated by environmental conditions. With gene expression affected by light, ARF8 may integrate auxin and light signals to control plant development.

Recently, red1 mutant, allelic to the above-mentioned sur2, has been reported to display long-hypocotyl phenotype specifically in continuous red-light condition and not in continuous far-red-light condition (Hoecker et al., 2004). SUR2/RED1 expresses specifically in red-light condition compared with far-red-light condition. This phenotype is a sharp contrast to that of ARF8: arf8-1 hypocotyls elongate in either far-red, red or blue light, and ARF8 expression is also induced in either light condition. Interaction between auxin and light signals seems to occur redundantly in either a specific or non-specific manner with respect to light quality.

The arf8-1 allele may have a small promotive effect on hypocotyl elongation

Although arf8-1 has a T-DNA insertion in its third exon, it contains smaller amounts of two ARF8-derived transcripts, suggesting that it may not be a null allele. In fact, F1 seedlings from a cross between arf8-1 and the wild type display slightly longer hypocotyls than the wild type. Furthermore, introduction of the wild-type ARF8 gene almost, but not completely, restores the long-hypocotyl phenotype of arf8-1. We examined the hypocotyl length of seedlings obtained from various crosses among the wild type, arf8-1 and arf8-1 with the ARF8 transgene. The results show that hypocotyl length is usually increased when the copy number of arf8-1 increases, indicating that arf8-1 has a slight promotive effect on hypocotyl elongation. Therefore, it is supposed that seedlings of the null arf8 allele would exhibit longer hypocotyls than the wild type, but that they must be a little shorter than those of arf8-1.

One of the two transcripts found in arf8-1 encodes a protein containing a 63-amino acid-long N-terminal portion of the ARF8 protein. Truncation of the N-terminal 26 amino acid residues disrupts DNA-binding activity of ARF1 (Ulmasov et al., 1999b); the N-terminal portion also corresponds to the N-terminal 26 amino acids in ARF8. Therefore, the arf8-1 truncated protein may participate in DNA binding of ARF8. The elongation-promoting function of arf8-1 may be due to arf8-1 truncated protein interfering with DNA binding of other ARFs whose functions overlap the function of ARF8. arf8-1 also produces an mRNA that encodes a small fusion protein between agropine synthesis reductase and agrocinopine synthase derived from the inserted T-DNA. Thus, we cannot exclude the possibility that the chimeric protein may have a weak function to promote hypocotyl elongation.

Experimental procedures

Plant materials, growth conditions and measurements of growth

Seeds of Arabidopsis thaliana were imbibed in the dark at 4°C for 2–3 days, surface-sterilized and sown on nutrient agar plates containing half-strength MS salts (Murashige and Skoog, 1962), 1% (w/v) sucrose and 0.9% (w/v) agar (pH 5.8); or in liquid medium containing the above-mentioned nutrient composition without agar. Plants were grown at 23°C under continuous white light (12 W m−2) for 7 days unless stated otherwise. In the case of far-red (740 ± 25 nm), red (660 ± 20 nm) or blue (470 ± 30 nm) light treatments, plates were incubated for 24 h in white light, then were moved into chambers with continuous monochromatic light illumination obtained from light-emitting diodes (Tokyo Rikakikai, Tokyo, Japan) for 6 days. After the seedlings were fixed with 5% formaldehyde and 10% acetic acid, hypocotyl length was measured by using an image scanner (GT-7600U; Epson, Suwa, Japan) and appropriate software (NIH Image or Image-Pro Plus; Media Cybernetics, Silverspring, MO, USA). Light fluence rate was determined by a radiometer (S370 with model 247 sensor; UDT Instruments, Baltimore, MD, USA). arf8-1 was backcrossed twice to wild-type Ws before phenotypic characterization.

Tropic responses

In order to determine gravitropism of hypocotyls, seedlings were grown on vertically oriented plates for 3 days in the dark and then turned 90° to a horizontal position. For second-positive phototropism, 3-day-old etiolated seedlings, grown as above, were irradiated with unilateral blue light at a fluence rate of 0.1 μmol m−2 sec−1 obtained by blue light-emitting diodes. An image of the seedlings was taken with a digital camera (C4040 Zoom; Olympus, Tokyo, Japan) at different times under dim green light. The angle was determined digitally from the image as described above.

Growth inhibition by IAA

After cold treatment and surface-sterilization, seeds in the liquid medium were placed under continuous white light at 23°C for 24 h to induce germination. The medium was exchanged for a medium supplemented with various concentrations of IAA, and seedlings were further grown in darkness for 5 days. Seedlings were then fixed and hypocotyl length was measured as described above.

Screening of T-DNA insertion lines

arf8-1 was isolated from a T-DNA-mutagenized population of Arabidopsis by employing a PCR-based reverse genetic screen (Krysan et al., 1999), using a primer for the T-DNA left border, 5′-CATTTTATAATAACGCTGCGGACATCTAC-3′ (JL202), and primers for 5′ or 3′ ends of ARF8, 5′-CTTGTGTTTAAGGGTTTTATGGATGGTTT-3′ or 5′-GATAAGTCTGATGTGTGTGCACGTTTAAG-3′. The T-DNA insertion site was determined by sequencing PCR products amplified between a primer in T-DNA, 5′-ATAATAACGCTGCGGACATCTACA-3′, and a primer located in the seventh exon of ARF8, 5′-CTACTAGTCGCTTGGCACTGACAA-3′.

Molecular methods

Total RNA was prepared with RNeasy Plant Mini Kit (Qiagen, Hilden Germany). Poly(A)+ RNA was purified from total RNA with GenElute-mRNA Miniprep Kit (Sigma, St. Louis, MO, USA). RT-PCR was performed using Access RT-PCR System (Promega, Madison, WI, USA). Semi-quantitative RT-PCR for ARF8 was carried out using a primer located in the first exon of ARF8, 5′-ATGAAGCTGTCAACATCTGGATTG-3′, and the above-mentioned primer located in the seventh exon. RNA was subjected to RNA gel blot analysis according to Tatematsu et al. (2004). Specific probes for the GH3 genes and SUR2 were prepared by RT-PCR with a pair of specific primers: 5′-CAAAGGCAAAGGGATGTATT-3′ and 5′-GCTCCCAGAATAAGACATAG-3′ for DFL1 (Nakazawa et al., 2001), 5′-TGAAGTTTGCTCCAATTATCGAGC-3′ and 5′-TGATAAAAGTCCAAACAATTCCGC-3′ for AtGH3a (Tanaka et al., 2002), 5′-TCATTGAAGAAATGACTCGGAACC-3′ and 5′-AGACCACATAGCATTTGAGCATAC-3′ for YDK1 (Takase et al., 2004), 5′-GTAGATTACCACCGGAACTTTGCT-3′ and 5′-AGCAAAGTTCCGGTGGTAATCTAC-3′ for At1g59500, and 5′-TCTTGACTCAAGGGTTGTATCTAC-3′ and 5′-AAGTGTTCGTCATTAACATCGACG-3′ for At2g23170, 5′-TAAAGCCTTTGACTTCACGAGAAC-3′ and 5′-CTTATACTGGGAGTTGTTAGTGAG-3′ for At2g14960, 5′-AGTAACTGACAGTACATGCTTTGG-3′ and 5′-TTTAAACGCGCTGTGCCAATCATC-3′ for At2g47750, 5′-GAGAGGACTTTGCTGAAAGTTTGT-3′ and 5′-GCGATATTCTAAAAGTGACGGGTT-3′ for At1g28130, 5′-ATTCACTCAACTTTCTCTCACTCC-3′ and 5′-ATTCACTCAACTTTCTCTCACTCC-3′ for JAR1 (Staswick et al., 2002), 5′-TTTGTCTGAAGACAAAGGGAGTGG-3′ and 5′-TAATGGAGAAGTCTCTCCATCTGC-3′ for DFL2 (Takase et al., 2003), and 5′-GGATCTCTTATTGATTATAGCCGG-3′ and 5′-GTCCAAAAACAAGAGCACTGATAC-3′ for SUR2 (Barlier et al., 2000). The identity of each probe was confirmed by sequencing.

To clone a shorter mRNA of arf8-1, RT-PCR was carried out with the primer in the first exon of ARF8 and the (dT)18 reverse primer. 5′-RACE was conducted to determine the 5′ end of a longer mRNA of arf8-1 by the use of 5′-Full RACE Core Set (Takara, Ohtsu, Japan). First-strand synthesis was primed using a 5′ phosphorylated T-DNA-specific primer, 5′-(P)-TTGCATTGAGACCGATGTTCGTTC-3′. After ligation with T4 RNA ligase, nested PCR was performed: primers for the first PCR were 5′-GGTTTTTTCTTGTGGCCGTCTTTG-3′ and 5′-CATATCTTCCACACGTGAAAATGC-3′, and those for the second PCR were 5′-TTTTCAAATCAGTGCGCAAGACG-3′ and 5′-CAACGAGAAGGAAATTGTCGTGAA-3′. The RT-PCR or PCR products were cloned into pT7blue T-vector (Novagen, San Diego, CA, USA).

Transgenic lines

For molecular complementation of arf8-1 plants, a 3.8-kb genomic fragment which spanned from 2.1-kb 5′ upstream region to the NheI site in the ninth exon of ARF8 gene, was fused with a fragment of ARF8 cDNA from the NheI site to its 3′ end. The fusion sequence, designated as pARF8:ARF8, was subcloned into pBI-H1 (Kimura et al., 1993). The 3.8-kb sequence was amplified from Columbia genomic DNA by PCR with a 5′ promoter primer, 5′-ATATGTCGACTGAGGCTTGATAGGCAATTCGAAG-3′ (SalI site is underlined), and 5′-GCATGCGAAAGCGCATCCCAACTG-3′ (the NheI site is 63 bp upstream of this primer). The construct was confirmed by sequencing. The pARF8:GUS reporter plasmid was constructed by placing a 2.1-kb 5′ upstream region of ARF8 gene ahead of GUS cDNA in a binary vector, pBI101 (Clontech, Palo Alto, CA, USA). This promoter sequence was amplified from Columbia genome DNA by PCR with the 5′ promoter primer and 5′-GCGCTCTAGAGTCTAATTCAACTTCAAGAAACCA-3′ (XbaI site is underlined). These constructs were confirmed by sequencing. The pBI vectors were introduced into Agrobacterium tumefaciens strain pGV2260 by electroporation, which was then used to transform Arabidopsis by flower dip method (Clough and Bent, 1998). For ARF8 overexpressors, CaMV 35S promoter:ARF8 containing HA epitope at the 5′ end of the ARF8 ORF, which was a generous gift of Dr T. J. Guilfoyle (Ulmasov et al., 1999a), was subcloned into pBI-H1. Of the 13 transgenic lines harboring 35S promoter:ARF8, three lines overexpressed ARF8 judged by RNA gel blot analysis.

GUS staining

GUS expression was examined by incubating seedlings in 100 mm sodium phosphate, pH 7.0, containing 1 mm 5-bromo-4-chloro-3-indolyl-β-d-glucuronide, 5 mm K3Fe(CN)6, 5 mm K4Fe(CN)6, 1 mm EDTA and 0.1% Tween-20, at 25°C for 24 h as described by Tatematsu et al. (2004).

Quantification of endogenous IAA

For gas chromatography single-ion-monitoring mass spectrometry (GC-SIM-MS) analyses of free IAA, 50–100 hypocotyls of 6-day-old light-grown Arabidopsis seedlings were ground in liquid nitrogen using a mortar and pestle. After addition of 1.8 ng of [13C6]IAA (Cambridge Isotope Lab, Andover, MA, USA) as an internal standard (2 ng 100 mg−1 fresh weight), the sample was extracted in 80% acetone and 0.1 mg ml−1 2,6-di-tert-butyl-4-methylphenol (BHT) twice. Following centrifugation, the supernatant was brought to a water phase, which was then partitioned twice against ether containing 0.01 mg ml−1 BHT at pH 2.0. The combined ether phase was concentrated, and the organic phase was discarded after adjusting the pH to 10.0 with 2% NaHCO3. The aqueous phase was partitioned against ether as described above after adjusting the pH to 2.0. The ether phase was dried, and the residual IAA dissolved in methanol was then purified by HPLC using a Nucleosil N(CH3)2 column (Senshu, Tokyo, Japan) and a mobile phase of methanol with 0.03% acetic acid. The purified IAA fraction, dried and trimethylsilylated with N-methyl-N-trimethylsilyltrifluoroacetamide (MSTFA), was subjected to splitless injections into a GC-SIM-MS system (QP5050A; Shimadzu, Kyoto, Japan) equipped with a capillary column (DB-1, 0.25 mm i.d. ×30 m, film thickness 0.25 μm; J&W Scientific, Folson, CA, USA). A linear temperature gradient was applied from 80 to 280°C with an increase of 20°C min−1. The injection temperature of the GC was 250°C, the ion source temperature of the MS was 250°C and a helium flow of 1.2 ml min−1 was applied. The ionization potential was 70 eV, and the scan time was 0.2 sec. The percentages of IAA molecules labeled with 13C were calculated from the relative intensities of m/z 202–208 and 319–325 ions after subtraction of background.


We thank Dr Kenzo Nakamura for pBI-H1, and Dr T. J. Guilfoyle for CaMV 35S promoter:ARF8 construct. We also acknowledge the Arabidopsis Biological Resource Center for providing us GH3 cDNAs. Screening for ARF8 disruption lines was conducted in part by Biotechnology Center, University of Wisconsin. This work was carried out in part in the Laboratory of Genetic Research, Center for Advanced Science and Technology, Hokkaido University. C.-e.T. was supported by a Postdoctoral Fellowship of Japan Society for the Promotion of Science, and this work was supported in part by Grants-in-Aid for Scientific Research in Priority Areas from the Ministry of Education, Culture, Sports, Science and Technology to T.K. and K.T.Y.