NPH4/ARF7 and ARF19 promote leaf expansion and auxin-induced lateral root formation


(fax +1 919 962 1625; e-mail


Auxin response factors (ARFs) bind auxin response promoter elements and mediate transcriptional responses to auxin. Five of the 22 ARF genes in Arabidopsis thaliana encode ARFs with glutamine-rich middle domains. Four of these can activate transcription and have been ascribed developmental functions. We show that ARF19, the fifth Q-rich ARF, also activates transcription. Mutations in ARF19 have little effect on their own, but in combination with mutations in NPH4/ARF7, encoding the most closely related ARF, they cause several phenotypes including a drastic decrease in lateral and adventitious root formation and a decrease in leaf cell expansion. These results indicate that auxin induces lateral roots and leaf expansion by activating NPH4/ARF7 and ARF19. Auxin induces the ARF19 gene, and NPH4/ARF7 and ARF19 together are required for expression of one of the arf19 mutant alleles, suggesting that a positive feedback loop regulates leaf expansion and/or lateral root induction.


The plant hormone auxin regulates numerous developmental processes, and can affect cell division, cell growth, or cell differentiation depending on the context. For example, auxin can promote or inhibit stem or root cell elongation required for tropic responses, induce cell divisions leading to lateral organ outgrowth (Benkova et al., 2003), and induce differentiation of vasculature (Mattsson et al., 1999; Sieburth, 1999). Plants have diverse auxin responses in part because multiple paralogous genes encode the cellular auxin response machinery (Dharmasiri and Estelle, 2004; Remington et al., 2004; Weijers and Jurgens, 2004), and these paralogs have distinct developmental functions.

Auxin induces many developmental effects by regulating gene expression. Transcription factors called auxin response factors (ARFs) can bind to auxin response elements (AuxREs) in promoters of auxin-inducible genes, and mediate auxin gene induction responses (Ulmasov et al., 1997a, 1999b). Arabidopsis has 22 ARF genes encoding members of this family (Hagen and Guilfoyle, 2002; Liscum and Reed, 2002; Remington et al., 2004). In addition to a conserved N-terminal DNA binding domain, most ARF proteins have a C-terminal dimerization domain that can interact with a corresponding domain in Aux/IAA proteins (Kim et al., 1997; Ulmasov et al., 1999b). Between these conserved N- and C-terminal domains, ARF proteins have a less conserved middle domain that in some cases can activate transcription (Tiwari et al., 2003; Ulmasov et al., 1999a). Five of the Arabidopsis ARF proteins, ARF5/MP, ARF6, ARF7/NPH4, ARF8, and ARF19, have glutamine-rich (Q-rich) middle domains, and ARF5/MP, ARF6, ARF7/NPH4 and ARF8 have been shown to activate transcription in protoplast transient expression assays (Ulmasov et al., 1999a). Other ARF middle domains are not rich in glutamine, and do not activate in these assays (Tiwari et al., 2003; Ulmasov et al., 1999a).

Analyses of plants with mutations in MP/ARF5, ARF6, NPH4/ARF7, and ARF8 genes have revealed the developmental functions of these four Q-rich middle domain ARFs. MP/ARF5 regulates embryo patterning, flower patterning, and vascular differentiation (Aida et al., 2002; Berleth and Jürgens, 1993; Hardtke and Berleth, 1998; Przemeck et al., 1996). NPH4/ARF7 regulates root and hypocotyl tropisms (Hardtke et al., 2004; Harper et al., 2000; Liscum and Briggs, 1996; Stowe-Evans et al., 1998; Watahiki and Yamamoto, 1997), and contributes to embryonic, vascular, and flower patterning in the absence of MP/ARF5 (Hardtke et al., 2004). ARF6 and ARF8 each promote stamen development, and in combination are required for flower bud opening and for both stamen and carpel maturation (P. Nagpal, C.M. Ellis, J.M. Alonso, J.R. Ecker & J.W. Reed, University of North Carolina, Chapel Hill, NC, USA, unpublished results). ARF8 also affects hypocotyl elongation (Tian et al., 2004). Several arf mutants have reduced auxin gene induction responses, consistent with the idea that these ARFs activate gene expression in response to auxin (Hardtke et al., 2004; P. Nagpal, C.M. Ellis, J.M. Alonso, J.R. Ecker & J.W. Reed, University of North Carolina, Chapel Hill, NC, USA, unpublished results; Stowe-Evans et al., 1998). The distinct functions of these ARFs probably arise in part because of distinct expression patterns. However, even when expressed ectopically, NPH4/ARF7 could not rescue an mp mutant (Hardtke et al., 2004), suggesting that MP/ARF5 has biochemical properties that NPH4/ARF7 lacks. These genetic and phenotypic analyses reveal that both the ARF6–ARF8 and MP/ARF5–NPH4/ARF7 pairs have partially overlapping functions. ARF6 and ARF8 form a phylogenetic clade, as do MP/ARF5, NPH4/ARF7 and ARF19 (Remington et al., 2004). Within the latter clade, NPH4/ARF7 and ARF19 are more closely related to each other than either is to MP/ARF5.

Auxin probably activates Q-rich middle domain ARF proteins by regulating turnover of ARF-interacting proteins called Aux/IAAs (Dharmasiri and Estelle, 2004). Arabidopsis has 29 IAA genes encoding Aux/IAA proteins. Most Aux/IAA proteins have four conserved motifs called I–IV. Motif I is a transcriptional repression domain and can repress auxin gene induction responses (Tiwari et al., 2001, 2004). Motif II is recognized by SCFTIR1 and probably other closely related E3 ubiquitin ligases. These ubiquitinate Aux/IAAs and target them for turnover, and auxin stimulates this pathway (Dharmasiri et al., 2003; Gray et al., 2001; Kepinski and Leyser, 2004; Ramos et al., 2001; Tian et al., 2003; Tiwari et al., 2001; Zenser et al., 2001). Motifs III and IV constitute a dimerization domain and can interact with similar motifs in ARF proteins (Kim et al., 1997; Morgan et al., 1999; Ouellet et al., 2001; Ulmasov et al., 1999b). Gain-of-function mutations in motif II of several IAA genes stabilize the corresponding protein and affect developmental responses to auxin (Reed, 2001). In several cases these mutations decrease auxin-induced gene expression (Tatematsu et al., 2004; Tian et al., 2002).

An important developmental function of auxin is to induce lateral roots. Lateral root primordia can initiate in the early differentiation zone of the root meristem (Dubrovsky et al., 2000), but only grow out in response to endogenous auxin delivered from the shoot (Bhalerao et al., 2002; Casimiro et al., 2003; Reed et al., 1998b). Exogenous auxin can induce additional lateral roots to form and grow out, and can also promote adventitious root formation on stems or other tissues. Several gain-of-function iaa mutations affect production of lateral or adventitious roots (Fukaki et al., 2002; Knee and Hangarter, 1996; Leyser et al., 1996; Rogg et al., 2001; Tatematsu et al., 2004; Tian and Reed, 1999), but it is not known which ARFs they target.

In this work we have studied the properties of ARF19, the last Arabidopsis glutamine-rich middle domain ARF to be characterized. We find that ARF19 can activate gene expression just as MP/ARF5, ARF6, NPH4/ARF7, and ARF8 can. Mutations in the ARF19 gene cause no dramatic phenotype on their own, but in combination with mutations in ARF7/NPH4, they decrease auxin-induced gene activation. In addition, arf19 nph4 double mutants have small leaves and make very few lateral or adventitious roots.


ARF19 can activate transcription

To determine whether ARF19 can activate auxin response genes containing TGTCTC AuxREs, we carried out transfection assays with Arabidopsis leaf mesophyll protoplasts. We improved upon previous transfection assays with ARF effector genes (Tiwari et al., 2003; Ulmasov et al., 1999a) by transfecting effector genes into nph4-1 mutant protoplasts containing single copy, stably integrated DR5:GUS or 2XD0:GUS reporter genes. Mesophyll cells that lack NPH4/ARF7 have little auxin-responsive gene expression unless transfected with an ARF activator (Tiwari et al., 2005; see below), and auxin regulates the integrated reporter genes much more tightly than it does transfected reporter genes.

Figure 1 shows that both integrated DR5:GUS and 2XD0:GUS reporter genes were expressed at much lower level in nph4-1 mesophyll protoplasts than in wild-type protoplasts. Transfection with either 35S:NPH4/ARF7 or 35S:ARF19 effector genes restored auxin-responsive gene expression of both reporter genes, while transfection with 35S:ARF1, a transcriptional repressor, did not. RT-PCR assays revealed that NPH4/ARF7 and ARF19 effector genes, but not an ARF1 effector gene, also increased auxin-responsive expression of endogenous IAA1 (At4g14560), IAA19 (At3g15540), and HAT2 (At5g47370) genes (data not shown). These results indicate that ARF19 can activate transcription of auxin-responsive promoters in the presence of auxin, similarly to the other four Q-rich ARFs in Arabidopsis.

Figure 1.

GUS activity of single copy, integrated auxin-responsive reporter genes in wild-type and nph4-1 mutant protoplasts transfected with 35S:NPH4/ARF7 and 35S:ARF19 effector genes.
(a) DR5:GUS reporter gene.
(b) 2XD0:GUS reporter gene. Standard errors are indicated.

ARF19 and NPH4/ARF7 are expressed in multiple tissues

To determine the cellular locations of NPH4/ARF7 and ARF19 expression, we examined X-gluc staining in plants carrying PNPH4/ARF7::GUS or PARF19::GUS promoter:GUS fusions (Figure 2). PNPH4/ARF7:GUS was expressed in petioles and blades of cotyledons and leaves, and in roots (Figure 2a–d). Thick sections of X-gluc-stained PNPH4/ARF7:GUS leaves revealed expression in all cell layers including vasculature, mesophyll and epidermis (data not shown), consistent with published in situ hybridization data (Hardtke et al., 2004). The root expression was strongest in the vasculature and in more mature parts of the root, and also appeared in root meristems (Figure 2i–l). PARF19::GUS was expressed in vasculature of cotyledons and leaves, stems and roots, with strongest staining in the young portions of roots including the root meristem (Figure 2e–h,m–o). In roots, PARF19::GUS staining was much stronger than PNPH4/ARF7::GUS staining. Both GUS fusion genes were also expressed in flower pedicels and in sepals, stamen filaments, and anthers, and strongly in the floral organ abscission zone below the silique (data not shown). In flowers, PARF19::GUS expression was more restricted to vascular tissue than was PNPH4/ARF7::GUS expression. It is possible that sequences in addition to those upstream of the start codon regulate transcription of these genes, or that post-transcriptional regulation is important. However, phenotypes we have observed in leaves and roots of mutant seedlings (see below) indicate that these expression patterns are relevant to the functions of NPH4/ARF7 and ARF19.

Figure 2.

Expression of PNPH4/ARF7:GUS and PARF19:GUS fusions in transgenic plants. (a–d, i–l) PNPH4/ARF7:GUS seedlings. (e–h, m–o) PARF19:GUS seedlings. (a, b, e, f) 4-day-old dark-grown seedlings; (c, g) 4-day-old light-grown seedlings; (d, h–o) 7-day-old light-grown seedlings. (l, o) primary root tips; (i–k, m, n) lateral roots. Staining in PARF19:GUS roots was extremely strong, and therefore the seedlings shown in n and o were stained for 2 h instead of 16 h.

Isolation of arf19 mutations

To determine the developmental functions of ARF19, we isolated mutations in the ARF19 gene from the Salk Institute sequence-tagged T-DNA insertion collection (Alonso et al., 2003). The arf19-3 insertion (SALK021481) is 12 bp upstream of the start codon (Figure 3a), and caused production of a slightly larger transcript than in wild-type plants (about 4.5 kb, Figure 3b). The arf19-3 transcript was also more abundant than the wild-type transcript. The arf19-4 insertion (SALK009879) interrupted the fourth intron and caused production of a transcript of about 10 kb (Figure 3b). As the probe used to detect these transcripts was from the 3′ portion of the gene downstream of the T-DNA insertion sites, the larger transcripts in the two mutants may have resulted from fusion of the 3′ portion of the ARF19 gene to T-DNA sequences. In an nph4 mutant background, where arf19 phenotypes were apparent (see below), arf19-3 and arf19-4 alleles were each recessive, and they each caused similar phenotypes when homozygous, suggesting that both are loss-of-function mutations. The severity of the effects of the two mutations was similar, suggesting that they may each be null mutations. In addition, preliminary results with an anti-ARF19 antibody suggest that ARF19 protein is present at a decreased level in the arf19-3 mutant and is absent in the arf19-4 mutant (T. J. Guilfoyle, unpublished data). The transcripts seen in the mutants may be poorly translated (arf19-3) or translated in an incorrect reading frame (arf19-4).

Figure 3.

Insertion mutations in ARF19.
(a) Diagram of the ARF19 gene indicating positions of T-DNA insertion mutations and start and stop codons. Black bars, exons encoding the open reading frame; white bars, introns; gray bar, region cloned in promoter:GUS fusion construct used in Figure 2.
(b, c) RNA blot hybridization of ARF19 in nph4-1, arf19-3, arf19-4 and double-mutant seedlings treated without or with 1.0 μm IAA for 2 h. Hybridization with ACTIN or rRNA gene probes served as loading controls.

We observed no obvious developmental phenotype in arf19-3 or arf19-4 single-mutant plants, suggesting that ARF19 may function redundantly with another ARF protein. The closest paralog of ARF19 in Arabidopsis is ARF7/NPH4 (Remington et al., 2004), and we therefore constructed double mutants between nph4-1 (Harper et al., 2000; Liscum and Briggs, 1996) and each arf19 mutation. We also constructed an nph4-6 arf19-4 double mutant with a T-DNA insertion (nph4-6, SALK040394) in NPH4/ARF7. The nph4-1 arf19-3, nph4-1 arf19-4, and nph4-6 arf19-4 double mutants each had the same spectrum of phenotypes. That the two independently isolated arf19 alleles and the two nph4 alleles used all had similar phenotypic effects in these double mutant combinations indicates that mutations in both loci caused the phenotypes we describe.

NPH4/ARF7 and ARF19 activate auxin-responsive genes

We used the single and double mutants to determine whether ARF19 contributes to auxin-activated gene expression. We examined auxin-induced gene expression in wild-type, nph4-1, arf19-4, and nph4-1 arf19-4 seedlings. As shown in Figure 4(a), auxin induced IAA1, IAA19, and GH3-4 (At1g59500) genes in wild-type and mutant plants. These genes were induced to similar degree in nph4-1 and arf19-4 plants as in wild-type plants. In the nph4-1 arf19-4 double mutant, the genes were induced less than in wild-type seedlings. These data indicate that ARF19 and NPH4/ARF7 both contribute to auxin induction of a common set of genes.

Figure 4.

Effects of nph4 and arf19 mutations on auxin-induced gene expression.
(a) RNA blot hybridization of indicated genes in nph4-1, arf19-4 and nph4-1 arf19-4 seedlings treated without or with 1 μm IAA for 2 h. Numbers beneath pairs of blots indicate auxin-responsive gene signal levels normalized to 18S rRNA, relative to signal in auxin-treated wild-type seedlings.
(b) Expression of transfected auxin-responsive reporter genes DR5:GUS and 2XD0:GUS in Arabidopsis leaf mesophyll protoplasts. Protoplasts were isolated from wild-type (left), nph4-1 mutant (middle) and arf19-3 mutant (right) leaves, transfected with auxin-responsive reporter genes and incubated with or without 1 μm 1-NAA. GUS activities were measured 20–22 h after transfections. Standard errors are indicated.

To test whether NPH4/ARF7 and ARF19 regulate expression in mesophyll cells, we used RT-PCR assays to test auxin induction of several genes in wild-type, nph4-1, and arf19-3 mutant mesophyll protoplasts. Auxin induced IAA1, IAA19, and HAT2 expression in wild-type protoplasts, and this induction was unaffected in arf19-3 protoplasts. However, auxin induced these genes much less in nph4-1 and nph4-1 arf19-4 protoplasts than in wild-type protoplasts (data not shown). These results suggest that NPH4/ARF7 regulates expression of multiple auxin-responsive genes in leaf mesophyll cells, but that ARF19 is less important in these cells. The arf19-3 and arf19-4 mutations may have had no effect in these assays because NPH4/ARF7 is expressed in leaf mesophyll cells whereas ARF19 is more strongly expressed in leaf veins but not in mesophyll cells (Figure 2). In contrast, nph4-1 may have had little effect in the whole-seedling experiments (Figure 4a) because these included tissues where both NPH4/ARF7 and ARF19 are expressed. The nph4-1 mutation did not affect all auxin-regulated genes in mesophyll protoplasts. Auxin induced expression of GH3-1 (At2g14960), GH3-2 (At4g37390), and GH3-3 (At2g23170) to a similar strong degree in wild-type, nph4-1 and arf19-3 protoplasts (data not shown).

Promoters of the genes tested in the mutants all have TGTCTC sequences that may function as AuxREs. Although these putative AuxREs have not been functionally defined, it is likely that ARF19 and NPH4/ARF7 act through these elements. To test whether nph4 or arf19 mutants were defective in gene expression directed by functionally defined TGTCTC AuxREs, we tested the DR5:GUS (Ulmasov et al., 1995, 1997b) and 2XD0:GUS (Liu et al., 1994; Murfett et al., 2001; Ulmasov et al., 1995) auxin-responsive reporter genes using transfection assays with wild-type and mutant mesophyll protoplasts. Figure 4(b) shows that auxin-induced DR5:GUS and 2XD0:GUS expression were reduced by more than 80% in nph4-1 protoplasts compared with wild-type protoplasts. In contrast, auxin-induced reporter gene expression was reduced by <50% in arf19-3 protoplasts.

To identify other cells in which NPH4/ARF7 and ARF19 affect auxin-induced gene expression, we monitored expression of a PSHY2/IAA3::GUS reporter gene (Tian et al., 2002) in nph4, arf19, and nph4 arf19 backgrounds. In wild-type plants, PSHY2/IAA3::GUS was expressed in hypocotyls of dark-grown seedlings, and in hypocotyls, petioles, cotyledons, and expanding leaves in light-grown seedlings (Tian et al., 2002) (Figure 5a). We did not previously detect expression in roots. However, using a modified staining protocol (see Experimental procedures), we observed staining in the center of the root just behind the meristem in a fraction of wild-type plants. This staining was stronger and was observed more consistently after growth of the seedlings on auxin (Figure 5b). In cotyledons, auxin did not change the expression pattern of this reporter fusion (Figure 5a). The auxin transport inhibitor NPA had little effect on expression in cotyledons but eliminated expression in roots, suggesting that the root expression may depend on transported auxin. The arf19-4 mutation by itself had no effect on these expression patterns, in either roots or cotyledons. In contrast, the nph4-1 mutation reduced PSHY2/IAA3:GUS expression in cotyledons but had no effect in roots. This reduction occurred in mesophyll cells of the cotyledon, but not in the petioles. In the nph4-1 arf19-4 double mutant, cotyledon staining was eliminated and root staining was decreased but not eliminated. These results indicate that NPH4/ARF7 is required for expression of this auxin-regulated gene in mesophyll cells, and that NPH4/ARF7 and ARF19 together promote expression of this gene in petioles and root cells. In root cells other factors apparently also contribute to auxin-induced expression.

Figure 5.

Expression of PSHY2/IAA3::GUS fusion in mutant backgrounds. 6-day-old seedlings of the indicated genotypes carrying the PSHY2/IAA3::GUS fusion were grown for 4 days on plates containing either no addition, 100 nm IAA, 100 nm NAA, or 10 μm NPA, and stained with X-Gluc.
(a) Cotyledons.
(b) Root tips.

Auxin induces ARF19 expression

Previous microarray experiments indicated that auxin can induce ARF19 expression (Tian et al., 2002). RNA blot hybridizations of RNA from whole seedlings confirmed this to be the case (Figure 3b,c). We did not observe alterations in the pattern or intensity of X-gluc staining of plants with the PARF19::GUS construct after auxin treatment (data not shown), suggesting that auxin increases ARF19 expression only in cells already expressing baseline levels of ARF19. We may have failed to observe increased intensity of X-gluc staining in PARF19::GUS plants because of the strong staining observed in the absence of auxin and the stability of the β-glucuronidase reporter protein. Auxin did not induce the arf19-3 transcript, and induced the arf19-4 transcript in some experiments but not others (Figure 3b,c). The larger arf19-4 transcript disappeared entirely in the nph4-1 arf19-4 double mutant (Figure 3b,c). These results indicate that NPH4/ARF7 is required for expression of arf19-4, and suggest that, if arf19-4 transcription uses regulatory elements in common with wild-type ARF19 transcription, then NPH4/ARF7 may also promote ARF19 expression. In that case, as the nph4-1 mutation by itself did not eliminate expression or auxin-inducibility of wild-type ARF19, the results suggest that NPH4/ARF7 and ARF19 together may activate ARF19 in a positive feedback loop. Future work will test this hypothesis more rigorously.

arf19 mutations enhance nph4 hypocotyl and leaf phenotypes

The mutants had developmental phenotypes in hypocotyls, leaves and roots. arf19-4, nph4-1, and arf19-4 nph4-1 seedlings had similar hypocotyl lengths and primary root lengths as wild-type seedlings (Table 1). Dark-grown nph4 mutant seedlings grow less upright than do wild-type seedlings, reflecting a defect in hypocotyl gravitropism (Harper et al., 2000). Whereas arf19 single-mutant seedlings grew upright similarly to wild-type seedlings, nph4 arf19 double mutants had a more extreme deviation from vertical growth than did the nph4-1 single mutant (Table 1). Thus, arf19 mutations enhance the effect of nph4-1 on hypocotyl gravitropism. Root gravitropism was also affected in the double mutants more severely than in the single mutants (Weijers et al., 2005).

Table 1.  Morphology of wild-type, nph4, arf19, and double-mutant plants
MeasurementWild typenph4-1arf19-3arf19-4nph4-1 arf19-3nph4-1 arf19-4
  1. Values are mean ± SD (n).

  2. aSignificantly (P < 0.05) different from wild-type population, based on t-test.

  3. bEpidermal cells were counted in imaged fields of 3130 μm2.

Root length, 7 days, mm20.5 ± 3.1 (15)23.2 ± 4.6 (13)20.4 ± 4.4 (31)20.0 ± 3.5 (28)20.0 ± 4.2 (12)19.8 ± 4.4 (10)
Hypocotyl length, 6 days dark, mm10.8 ± 1.1 (20)11.2 ± 1.5 (20)9.0 ± 1.2 (40)a9.6 ± 1.1 (40)a13.0 ± 3.1 (20)a10.4 ± 3.0 (20)
Mean hypocotyl angle from vertical,°5.7 ± 4.1 (69)18.8 ± 17 (62)a8.9 ± 11.2 (48)a8.0 ± 7.9 (89)a32.9 ± 27.3 (54)a93.8 ± 68 (30)a
Adventitious roots from hypocotyl2.9 ± 0.8 (23)1.4 ± 0.9 (21)a2.7 ± 1.1 (15)2.5 ± 1.0 (27)0 ± 0.2 (21)a0 ± 0 (22)a
Lateral roots, 12 days9.9 ± 3.5 (22)7.6 ± 5.8 (10) 5.3 ± 2.1 (15)a 0.1 ± 0.3 (41)a
Rosette diameter, 33 days40.1 ± 3.6 (12)32.8 ± 5.3 (11)a 39.7 ± 3.2 (11) 28.1 ± 3.4 (12)a
Leaf 6 (42 days)
 Petiole length, cm9.5 ± 2.3 (10)8.9 ± 2.1 (10) 10.9 ± 1.3 (10) 9.9 ± 1.6 (10)
 Leaf blade length, cm14.4 ± 3.4 (10)13.4 ± 2.5 (10) 15.6 ± 2.5 (10) 10.7 ± 1.7 (10)a
 Leaf blade width, cm10.3 ± 1.8 (10)9.5 ± 1.3 (10) 10.7 ± 1.1 (10) 7.9 ± 1.0 (10)a
 Leaf blade area, cm2116100 131 66
 Cells per unit area, leaf baseb83 ± 15 (4)104 ± 17 (5) 75 ± 10 (4) 124 ± 29 (5)
 Average cell area, leaf base, μm23830 42 25
 Cells per unit area, leaf middleb80 ± 13 (4)84 ± 10 (5) 87 ± 10 (5) 118 ± 9 (5)a
 Average cell area, leaf middle, μm23937 36 27
 Cells per unit area, leaf tipb86 ± 9 (4)87 ± 10 (5) 77 ± 8 (5) 124 ± 20 (5)a
 Average cell area, leaf tip, μm23636 41 25

Harper et al. (2000) found that the ethylene precursor ACC could suppress the agravitropic growth of nph4 hypocotyls, and hypothesized that ethylene could activate another factor to promote gravitropism in dark-grown seedlings. The redundant action of NPH4/ARF7 and ARF19 suggested that ARF19 might be the proposed factor. However, we found that ACC could suppress the agravitropic growth of nph4-1 arf19-3 and nph4-1 arf19-4 double mutants (data not shown), indicating that the hypothesized ethylene-regulated factor was not ARF19.

Adult arf19 single-mutant plants had rosettes of similar diameter to those of wild-type plants. nph4 mutant plants had slightly smaller rosettes than wild-type plants, and arf19 nph4 double-mutant plants had smaller rosette diameters than either wild-type or single-mutant plants (Table 1; Figure 6). This was especially apparent in plants grown under short-day conditions under bright light. Leaf sizes of the different genotypes correlated with rosette diameter. In a separate experiment, we measured petiole length and leaf blade length and width of leaf number 6 of 42-day-old plants (Table 1). Petiole lengths of nph4-1, arf19-4, and nph4-1 arf19-4 plants were similar to those of wild-type plants. Leaf blades of nph4-1 single-mutant plants were slightly shorter and narrower than those of wild-type plants, although these differences were not statistically significant. Leaf blades of nph4-1 arf19-4 double mutants were significantly shorter and narrower than wild-type or single-mutant leaf blades (Table 1). We compared sizes of epidermal cells of these leaves in leaf surface imprints. Single-mutant leaves had similar numbers of cells per unit area as did wild-type leaves, whereas leaves of the double mutant had more cells per unit area (Table 1). The calculated average epidermal cell size of nph4-1 arf19-4 leaves was 67–69% of average wild-type leaf epidermal cell size for different parts of the leaf, while the average leaf blade area was 56% of wild-type leaf blade area. While these results do not exclude effects on cell division, they indicate that smaller cells largely account for the small leaves in the double mutant and therefore that ARF19 and NPH4/ARF7 promote leaf cell expansion.

Figure 6.

Rosettes of 33-day-old wild-type and mutant plants grown in short days.

NPH4/ARF7 and ARF19 promote lateral and adventitious root formation

Exogenous auxin can inhibit primary root growth (Estelle and Somerville, 1987). nph4 and arf19 single-mutant roots were similarly sensitive to exogenous IAA as were wild-type roots (Figure 7). However, the nph4 arf19 double mutants were much less sensitive, with the dose–response curve shifted by about 50-fold (Figure 7a).

Figure 7.

Effects of nph4 and arf19 mutations on root growth and lateral root formation.
(a) Root elongation over 4 days on medium with the indicated concentration of IAA.
(b) Number of lateral roots formed over 7 days on medium with the indicated concentration of IAA.
(c) Number of lateral roots formed from excised primary roots incubated for 10 days on medium with various concentrations of IAA.
(d) Number of adventitious roots formed from excised hypocotyls incubated for 10 days on medium with various concentrations of IAA. Error bars indicate SD. The legend in panel (b) applies to all four panels. Results very similar to those shown for nph4-1 arf19-4 were obtained for nph4-1 arf19-3 (all four experiments) and nph4-6 arf19-4 (panels a and b) (data not shown). Asterisks indicate values that are unlikely to be the same as wild-type values according to a t-test (P < 0.05).

The most dramatic phenotype of the nph4 arf19 double mutants was an almost complete lack of lateral roots. While nph4 and arf19 single-mutant seedlings each had similar numbers of lateral roots as did wild-type seedlings, nph4 arf19 double-mutant seedlings made no lateral roots (Table 1). Even in the presence of up to 1 μm of exogenous IAA, which stimulates lateral root formation in wild-type plants, double-mutant plants formed very few lateral roots (Figure 7b). Similarly, in an assay for adventitious root formation from the hypocotyl of intact seedlings, double-mutant seedlings made very few adventitious roots (Table 1).

Lateral roots are thought to develop in response to shoot-derived auxin (Reed et al., 1998b), and the decreased number of lateral and adventitious roots in nph4 arf19 double-mutant plants could arise if the leaves produce less auxin. This possibility is reasonable considering that the leaves expand less than do wild-type leaves. To test more directly whether the decreased lateral and adventitious root formation rates were caused by root- and hypocotyl-autonomous defects, we dissected apart roots, hypocotyls, and shoot apices from 3–4-day-old dark-grown seedlings, and tested whether these could produce roots in response to varying levels of exogenous auxin. Seedlings were grown first in darkness to promote hypocotyl elongation, and then grown in white light for a further 10 days before roots were counted. As shown in Figure 7(c), in the presence of exogenous IAA excised roots from nph4 and arf19 single mutants formed fewer lateral roots than did wild-type excised roots. nph4 and arf19 single mutant excised hypocotyls also formed fewer adventitious roots than did wild-type hypocotyls (Figure 7d). nph4 arf19 double mutants had even more extreme decreases in lateral and adventitious root numbers. These results indicate that the decreased lateral and adventitious root production in nph4 arf19 seedlings are automous to the root and hypocotyl.

At the highest IAA concentration tested (10 μm), the nph4 arf19 double mutants did make a small number of lateral and adventitious roots. Moreover, the excised shoot portions of these seedlings containing the shoot apical meristem also produced a few adventitious roots even in the absence of exogenous auxin (data not shown). Therefore, other ARF proteins may be able to stimulate lateral root production in the absence of both NPH4/ARF7 and ARF19.

Limited grafting experiments also suggested that the lateral root phenotypes were root-autonomous (data not shown). Thus, five successful wild-type root:nph4 arf19 shoot grafts produced plants that formed lateral roots, whereas a single successful nph4arf19 root:wild-type shoot graft produced a plant that failed to produce lateral roots. Conversely, the rosettes of wild-type root:nph4 arf19 shoot grafts were smaller than those of wild-type shoot:wild-type root grafted plants, suggesting that the leaf expansion defect of nph4 arf19 plants was also leaf-autonomous.


We have found that ARF19 acts redundantly with NPH4/ARF7 to regulate leaf cell expansion, hypocotyl growth angle, and lateral root formation. Both NPH4/ARF7 and ARF19 activated transcription in protoplast assays, and nph4 arf19 double-mutant plants had decreased gene activation responses to exogenous auxin. They also had decreased endogenous level of arf19-4 transcript and of PSHY2/IAA3:GUS expression in cotyledons. Endogenous auxin promotes leaf expansion, tropisms, and lateral root induction (Bhalerao et al., 2002; Jones et al., 1998; Marchant et al., 2002; Reed et al., 1998b), and our results therefore suggest that the mutant phenotypes arose from decreased gene induction responses to endogenous auxin.

Functional tests using isolated cells or organs indicate that organ-autonomous action of NPH4/ARF7 and ARF19 promotes lateral root production. Detached roots and hypocotyls of nph4 arf19 double-mutant plants produced very few lateral roots even in the presence of exogenous auxin, and a wild-type shoot did not restore lateral root production when grafted to a double-mutant rootstock. The effects of the mutations on lateral root formation are likely to be cell-autonomous because lateral roots arise from pericycle cells, and NPH4/ARF7 and ARF19 promoter:GUS fusions were expressed in root vasculature. Both were also expressed in root meristems, and double mutants were also resistant to inhibition of root growth by auxin. Such organ- (and probably cell-) autonomous function is consistent with the function of these proteins as transcription factors. nph4 and arf19 single mutations had no effects on these root phenotypes, suggesting that NPH4/ARF7 and ARF19 have largely redundant functions in roots under the conditions we tested.

The actions of NPH4/ARF7 and ARF19 in leaves and hypocotyls are more distinguishable. PNPH4/ARF7:GUS was expressed strongly in vasculature of hypocotyls, and weakly in cortical and epidermal cells, whereas PARF19::GUS was expressed mainly in vasculature of hypocotyls. Consistent with these patterns, nph4-1 had a larger effect on hypocotyl growth angle than did arf19 mutations, and nph4 arf19 double mutants had even more randomized growth direction, indicating that NPH4/ARF7 and ARF19 act partially redundantly in hypocotyls, with NPH4/ARF7 playing a slightly larger role. Mesophyll cells expressed PNPH4/ARF7:: GUS but not PARF19::GUS, whereas petioles and vascular cells expressed both fusions. Consistent with these patterns, nph4 mutations decreased auxin-responsive gene expression in mesophyll protoplasts and PSHY2/IAA3:GUS expression in the mesophyll of cotyledons, but arf19 mutations had little effect on expression in mesophyll. However, nph4 arf19 double mutants also lacked PSHY2/IAA3:GUS expression in cotyledon petioles, and had smaller leaves than did single mutants, indicating that both ARFs promote leaf cell expansion. The decreased cell expansion in nph4 arf19 double-mutant leaf blades compared with wild-type or nph4 leaf blades suggests that ARF19 may have some non-cell-autonomous effects in leaves, because we observed little PARF19::GUS expression outside vasculature in leaf blades. Alternatively, the promoter fusions may indicate only a subset of the normal domains of expression of these genes or the corresponding proteins.

As discussed above, Aux/IAA proteins probably regulate development by repressing gene activation by Q-rich ARFs. Consistent with the model that Aux/IAA proteins regulate NPH4/ARF7 and ARF19, plants with gain-of-function mutations in motif II of any of several different IAA genes have phenotypes that overlap those of nph4 arf19 double mutants. For example, shy2/iaa3, slr1/iaa14, msg2/iaa19, and iaa28 mutations each reduced or eliminated lateral root formation (Fukaki et al., 2002; Rogg et al., 2001; Tatematsu et al., 2004; Tian and Reed, 1999). Based on promoter:GUS fusions, these IAA genes were each expressed in stele and/or pericycle cells, consistent with these phenotypes. SHY2/IAA3 was expressed principally near the root meristem (Figure 5b), whereas SLR1/IAA14 and IAA28 had stronger expression in more mature zones of the root (Fukaki et al., 2002; Rogg et al., 2001), and MSG2/IAA19 had little baseline expression in the root but was strongly induced in roots by exogenous auxin (Tatematsu et al., 2004). The differences among these expression patterns in the root pericyle or stele suggest that different Aux/IAA proteins might regulate lateral root formation in different zones of the root or at different stages, providing finer control of lateral root formation. Analogously, several gain-of-function iaa mutations affect root growth inhibition by auxin, leaf expansion, and tropisms (Reed, 2001), and different Aux/IAA proteins may regulate NPH4/ARF7 and ARF19 in root meristems, hypocotyls, or leaves at distinct developmental moments. Although these conclusions based on gain-of-function alleles should be viewed cautiously (Reed, 2001), some of these Aux/IAA proteins probably indeed regulate NPH4/ARF7 and ARF19 in wild-type plants.

Light conditions affect the degree of leaf expansion and the direction of hypocotyl growth, and concentrations of multiple nutrients affect lateral root formation and outgrowth (Lopez-Bucio et al., 2003; Malamy and Ryan, 2001; Vandenbussche et al., 2003). Environmental stimuli might therefore regulate NPH4/ARF7 and ARF19 activity. Analyses of nph4 single-mutant phenotypes have in fact indicated that NPH4/ARF7 regulates hypocotyl phototropism (Harper et al., 2000; Stowe-Evans et al., 1998, 2001). If light or nutrient levels affect Aux/IAA protein stability or activity, then a subset of Aux/IAA proteins that can regulate lateral root production, phototropism, or leaf expansion may mediate inputs from these environmental signals. Alternatively, non-Q-rich ARFs may interact with NPH4/ARF7 and ARF19, either by dimerizing with them directly or by competing for promoter binding. For example, ethylene promotes, and light inhibits, ARF2 turnover in seedlings (Li et al., 2004), suggesting that light and ethylene might regulate NPH4/ARF7 and/or ARF19 through effects on ARF2.

Even when entire roots are exposed to auxin, only a subset of pericycle cells divide to form lateral roots. This observation implies that stochastic and/or feedback mechanisms allow local groups of cells to cooperate to form a lateral root. Such self-organization presumably requires local differences in auxin concentrations that are set up by polar auxin transport (Benkova et al., 2003). In addition, intracellular positive and negative feedback loops may regulate NPH4/ARF7 and ARF19 activity in cells destined to form a lateral root. Auxin induces SHY2/IAA3, SLR1/IAA14, MSG2/IAA19, and many other IAA genes, and this induction may dampen lateral root formation response at auxin concentrations below some threshold. Conversely, auxin induces ARF19, and this positive feedback may promote lateral root formation by pericycle cells that have surpassed some threshold of auxin concentration, or in which the auxin signal persists for longer than some minimal duration. Moreover, gain-of-function axr3 mutations increase auxin-induced gene induction and adventitious root numbers (Leyser et al., 1996), suggesting that AXR3/IAA17 and other Aux/IAA proteins may dimerize with Aux/IAAs that act negatively, thereby constituting an additional potential positive feedback mechanism. Once a threshold of auxin concentration and/or ARF activation is reached, then self-reinforcement of the induced state through positive feedback may ensure that lateral roots do not abort at an incipient stage. The ARF19 promoter has several possible AuxREs that might mediate this positive feedback, including two centered at 804 and 824 bp upstream of the ATG.

NPH4/ARF7 and ARF19 might activate any of several candidate target genes to induce lateral root formation. NAC1, encoding a NAC-domain protein and XBAT32, encoding a ring finger protein, each promote lateral root development, and auxin induces both NAC1 and XBAT32 (Nodzon et al., 2004; Xie et al., 2000). Further work will be required to establish whether these two genes are targets, and what other genes may mediate effects on lateral root formation and leaf expansion.

Other workers have also recently characterized phenotypes of arf19, nph4 and nph4 arf19 mutant plants, and obtained results similar to those described here (Okushima et al., 2005). Data are now available on the developmental functions of all five Q-rich ARFs in Arabidopsis. Based on current information, these five account for many of the developmental processes that auxin can induce, including embryo, vascular, and phyllotactic patterning (MP/ARF5 and NPH4/ARF7), tropic growth (NPH4/ARF7), lateral root formation (NPH4/ARF7 and ARF19), leaf expansion (NPH4/ARF7 and ARF19), inhibition of root elongation (NPH4/ARF7 and ARF19), and flowering stem elongation (ARF6 and ARF8). However, these results do not account for all auxin-related responses. For example, auxin still induces lateral roots at a low frequency in nph4 arf19 double mutants, and none of the characterized mutant combinations appears to affect inhibition of shoot branching. Additional redundancy among these five ARFs might be revealed by analyses of higher-order mutants. Another possibility is that the non-Q-rich ARFs might also mediate auxin gene activation responses in particular tissues.

Experimental procedures

Plant material

T-DNA insertions in ARF19 (At1g19220) and NPH4/ARF7 (At5g20730) were identified from the sequenced SALK collection (Alonso et al., 2003), and plants carrying these insertions were identified by polymerase chain reaction using gene-specific and T-DNA left border primers. Based on the insertion endpoint sequence index (Alonso et al., 2003), the arf19-3 insertion (SALK021481) is 12 bp upstream of the start codon, and the arf19-4 insertion (SALK009879) interrupted the fourth intron. The arf19-3 allele was identified using primers ARF19-F1 (5′-GTATCCGATCCCAATCGG-3′) and JMLB (Alonso et al., 2003). The arf19-4 allele was identified using primers ARF19-R1 (5′-GTCTGTTTAGCTTCAAGCC-3′) and JMLB, and the wild-type allele was identified using primers ARF19-F1 and ARF19-R1. The nph4-1 mutation is a DNA deletion and/or rearrangement (Harper et al., 2000; Liscum and Briggs, 1996) and was genotyped using gene-specific primers that span the rearrangement breakpoint (E. Liscum, personal communication). Primers NPH4-5 (5′-TCCTGCTGAGTTTGTGGTTCCTT-3′) and NPH4-6 (5′-GGGGCTTGCTGATTCTGTTTGTTA-3′) span the breakpoint, and therefore amplify a product from the wild-type allele but not the nph4-1 allele, whereas primers NPH4-11 (5′-TGCCTCAGCTCATCGTAAC-3′) and IAA25-3 (5′-GGTGGATTGTGGCCAGCTCAG-3′) amplify a product from both alleles. Co-amplification with both primer pairs allows identification of nph4-1 homozygotes as those giving rise to the NPH4-11/IAA25-3 amplification product but not the NPH4-5/NPH4-6 product. nph4-6 (SALK040394) is an insertion in the 11th (and largest) exon. nph4-6 was identified by PCR amplification with primers JMLB and NPH4-5, and the corresponding wild-type allele was identified with primers NPH4-5 and NPH4-R21 (5′-TCTGCTGGAATATCTGTTGG-3′).

Effector and reporter plasmids

The full-length open reading frame of ARF19 was amplified by RT-PCR using RNA from Arabidopsis suspension culture cells, sequenced to confirm no sequence modifications occurred, and cloned under the control of the 35S double enhancer promoter of cauliflower mosaic virus (CaMV) followed by the translational enhancer from the 5′ leader of tobacco mosaic virus (Tiwari et al., 2003). The 3′ untranslated region of the construct was derived from the nopaline synthetase gene (Tiwari et al., 2003). Full-length ARF1 and NPH4/ARF7 effector constructs have been described previously (Tiwari et al., 2003; Ulmasov et al., 1999a). The DR5:GUS and 2XD0:GUS (in pUC19) reporter gene plasmids have been described (Murfett et al., 2001; Ulmasov et al., 1995, 1997b). Single copy, integrated DR5:GUS and 2XD0:GUS reporter genes were described previously (Murfett et al., 2001; Ulmasov et al., 1997b).

Protoplast isolation

Wild-type and homozygous nph4-1, arf19-3, nph4-1DR:GUS, and nph4-12XD0:GUS Arabidopsis seeds were germinated and grown in pots containing moistened Pro-Mix (Premier Horticulture Inc., Red Hill, PA, USA) at 20°C under continuous light. Leaves from plants that were approximately 4–6 weeks old were used for protoplast isolation. Protoplasts were isolated using a modification of the procedure of Kovtun et al. (2000; and Tiwari et al. (2005). Leaves (approximately 1 g) were cut into 0.5–1 mm strips and transferred to a Petri dish containing 20–25 ml enzyme solution (1% cellulase R10 SERVA Electrophoresis GmbH, Heidelberg, Germany), 0.25% macerozyme R10 (SERVA Electrophoresis GmbH), 0.4 m mannitol, 80 mm CaCl2, 20 mm MES, pH 5.7). After vacuum infiltration for 20 min, the leaf tissue was placed in the dark, gently shaken at 40 rpm on a platform shaker for 90 min, followed by shaking at 80 rpm for an additional 1 min. The protoplasts were filtered (200 μm nylon mesh; Spectrum Laboratories Inc., Rancho Dominguez, CA, USA) and diluted by adding 1/3 volume 200 mm CaCl2. The filtered protoplasts were pelleted at 1000 rpm for 3 min in a Beckman JA7.5 rotor, washed with 25 ml cold W5 solution (154 mm NaCl, 125 mm CaCl2, 5 mm KCl, 5 mm glucose and 1.5 mm MES, pH 5.7), repelleted, and gently resuspended in 25 ml cold W5 solution. The purified protoplasts were incubated on ice for 30 min., and protoplasts were counted using a hemacytometer and a light microscope. The protoplasts were repelleted and resuspended in cold MMg solution (0.4 m mannitol, 15 mm MgCl2, 4 mm MES, pH 5.7) at 3 × 105 protoplasts per ml.

Protoplast transfection and GUS assays

Protoplasts were transfected by a modified polyethylene glycol (PEG) method described by Kovtun et al. (2000). Approximately 6 × 104 protoplasts in 0.2 ml MMg solution at 20°C were mixed with 10 μg supercoiled effector plasmid DNA for protoplasts with integrated reporter genes or with 10 μg effector plasmid DNA plus 10 μg reporter plasmid DNA for protoplasts without integrated reporter genes. A 35S:CAT effector gene (Tiwari et al., 2003; Ulmasov et al., 1999a) was added to equalize amounts of transfected DNA in assays that did not contain an ARF effector gene. An equal volume of 40% (w/v) PEG3350 (Sigma-Aldrich, St Louis, MO, USA) prepared with 0.1 m Ca(NO3)2 and 0.4 m mannitol solution (pH 10) was added to the mixture, and the transfection cocktail was incubated at room temperature for 20 min. W5 solution (0.8 ml) was added slowly without mixing, and the protoplast mixture was incubated for another 10 min. Following incubation, the solution was mixed, and protoplasts were pelleted by centrifugation at 1000 rpm for 3 min. The transfected protoplasts were gently resuspended in 1 ml WI solution (0.5 m mannitol, 20 mm KCl, 4 mm MES, pH 5.7) with or without 1 μm 1-NAA and were incubated at room temperature in the dark for 20–22 h. GUS activity was assayed as described by Liu et al. (1994), with the exception that the protoplasts were lysed in 100 μl of luciferase cell culture lysis reagent (Promega, Madison, WI, USA), and 10 μl of lysate was added to 100 μl of 1 mm 4-methylumbelliferyl β-d-glucuronide solution.

Promoter:GUS fusion constructions

NPH4/ARF7 and ARF19 promoter:GUS constructs were cloned just upstream of the GUS start codon in a modified pPZP211 vector that contained GUS-nos and some upstream restriction sites derived from pEBGUS (Hagen et al., 1991). For ARF7/NPH4, a fragment containing 2248 bp upstream of the ARF7/NPH4 start codon was amplified using primers 5′-CCACTTTCCACTGCAGACACCATACGGAC-3′ (PstI site underlined, includes nucleotides 56700–56672 of BAC AF296832) and 5′-CCTTCAACAGGATCCGGAGAAACTCCATTTG-3′ (BamHI site underlined, includes nucleotides 54440–54470 of BAC AF296832), and was cloned into a BamHI site upstream of GUS. The BamHI site was created in the sequence 30 bp downstream of the NPH4/ARF7 start codon, so this construct contains both the NPH4/ARF7 ATG and the GUS ATG. For ARF19, a fragment containing 2379 bp upstream of the start codon was amplified using primers 5′-CACCTCTCGAGCATGCTGCACCCAC-3′ (XhoI site underlined, includes nucleotides 24572–24548 of BAC AC069143) and 5′-CTAGCTCGAGCCATTTGATGGAGCTTTCATGG-3′ (XhoI site underlined, includes nucleotides 22194–22215 of BAC AC069143), and was cloned into a SalI site upstream of the GUS gene, eliminating the restriction site. In this case the GUS ATG was used as the start codon.

Gene expression in plants

RNA hybridizations were performed as described previously (Tian et al., 2002) using probes generated by PCR from genomic DNA templates. Hybridization bands in the blots in Figure 4 were quantitated using ImageQuant software (Amersham Biosciences, Piscataway, NJ, USA). For detecting PSHY2/IAA3:GUS expression, seedlings were immersed in ice-cold 90% acetone for 15–20 min. We previously omitted this acetone fixation step (Tian et al., 2002). The seedlings were then washed three times with distilled water and then immersed in 50 mm phosphate buffer (pH 7.0) for 30 min. The buffer was then replaced by 50 mm phosphate buffer pH 7.0 containing 2 mm X-Gluc, vacuum infiltrated for 20 min and incubated at 37°C for 24 h or until blue color was observed.

Phenotypic analyses

Conditions for growth of plants on agar plates and in soil were as described previously (Reed et al., 1998a). For rosette diameters, the mean of three measurements taken at roughly 60° angles from each other were taken for each plant. Data for rosette diameters and leaf sizes shown in Table 1 was from plants grown under 8 h days at approximately 180 μmol m−2 sec−1 light. To determine leaf cell numbers, epidermal imprints were taken in superglue and the images measured using NIH Image as described (Elumalai et al., 2002). Hypocotyl growth angles were measured using NIH Image to analyze images of dark-grown seedlings grown on vertically oriented plates. Auxin root growth resistance assays were as described (Estelle and Somerville, 1987; Tian and Reed, 1999). Induction of adventitious root formation in intact plants was as described (Tian and Reed, 1999). For the dissected seedling experiments, seedlings were grown on vertical plates for 3–4 days in darkness, and were then cut using microsurgical scissors at the base and the top of the hypocotyl. The dissected root, hypocotyl and shoot segments were then incubated on plates containing the indicated concentrations of IAA for a further 10 days in light (100 μmol m−2 sec−1) before being examined for lateral or adventitious roots under a dissecting microscope. Grafting experiments were performed as described (Turnbull et al., 2002).


We thank Christine Ellis and Qing Tian for help with RNA blot hybridizations, Eric Tomko for making promoter:GUS fusion constructs, Paul Reeves and Sara Ploense for help preparing figures, and Mannie Liscum for providing primers for the nph4-1 diagnostic assay. This work was supported by NIH grant R01-GM52456 to J. W. R. and National Science Foundation Grant MCB 00800096 to T. J. G. and G. H.