Hormonally controlled expression of the Arabidopsis MAX4 shoot branching regulatory gene


(fax +44 1904 32 8682; e-mail hmol1@york.ac.uk).


The Arabidopsis MORE AXILLARY BRANCHING 4 (MAX4) gene is required for the production of a long-range, graft-transmissible signal that inhibits shoot branching. Buds of max4 mutant plants are resistant to the inhibitory effects of apically applied auxin, indicating that MAX4 is required for auxin-mediated bud inhibition. The RAMOSUS 1 (RMS1) and DECREASED APICAL DOMINANCE 1 (DAD1) genes of pea and petunia, respectively, are orthologous to MAX4 and function in a similar way. Here we show that, despite the similarities between these three genes, there are significant differences in the regulation of their expression. RMS1 is known to be upregulated by auxin in the shoot, suggesting a straightforward link between the RMS1-dependent branch-inhibiting signal and auxin, whereas we find that MAX4 is only upregulated by auxin in the root and hypocotyl, and this is not required for the inhibition of shoot branching. Furthermore, both RMS1 and DAD1 are subject to feedback regulation, for which there is no evidence for MAX4. Instead, overexpression studies and reciprocal grafting experiments demonstrate that the most functionally significant point of interaction between auxin and MAX4 is post-transcriptional and indeed post-synthesis of the MAX4-dependent graft-transmissible signal.


In many plant species, the presence of an actively growing primary shoot apex exerts a strong inhibitory effect on the outgrowth of the axillary buds below it, a phenomenon termed apical dominance (reviewed in Cline, 1994). Since the 1930s, when it was first proposed by Thimann and Skoog (1934), a large body of evidence has accumulated to support a central role for apically derived auxin in mediating apical dominance. In Arabidopsis, analysis of the effects of the loss of AUXIN RESISTANT1 (AXR1) gene function have been central in establishing the role of auxin. Mutations in AXR1 impair auxin signalling by reducing the efficiency of auxin-induced destabilization of the Aux/IAA transcriptional repressor proteins (Gray et al., 2001), thereby dampening auxin-induced transcriptional upregulation. These auxin signalling defects correlate with increased shoot branching and bud outgrowth resistant to the inhibitory effects of apically applied auxin (Lincoln et al., 1990; Stirnberg et al., 1999).

The mechanisms by which auxin inhibits bud growth are still unclear. Auxin must act indirectly because radiolabelled auxin moves down the stem in the polar transport stream, but does not accumulate in inhibited buds (Booker et al., 2003; Everat-Bourbouloux and Bonnemain, 1980; Hall and Hillman, 1975). Furthermore, direct measurement suggests that auxin levels actually increase in buds as they begin to grow out (Gocal et al., 1991; Hillman et al., 1977). In Arabidopsis, grafting studies and expression of the wild-type AXR1 gene from tissue-specific promoters in an axr1 mutant background have identified the xylem-associated tissues at the node as the site of auxin action in the inhibition of bud growth (Booker et al., 2003). These tissues include the main conduit for auxin transport down the stem (Galweiler et al., 1998). Hence, it seems likely that auxin is perceived in the stem as it is transported down the plant and this signal is somehow relayed into the bud.

The search for factors that could act in the transmission of the inhibitory message into the buds identified additional mutants with relatively unpleiotropic increases in axillary bud outgrowth. Among those identified were the more axillary branching (max) mutants of Arabidopsis, the ramosus (rms) mutants of pea and the decreased apical dominance (dad) mutants of petunia (Beveridge, 2000; Booker et al., 2004, 2005; Morris et al., 2001; Napoli, 1996; Snowden et al., 2005; Sorefan et al., 2003; Stirnberg et al., 2002). Reciprocal grafting demonstrated that for many of these mutants, wild-type root stocks were able to restore a wild-type branching habit to mutant shoots (Beveridge et al., 2000; Booker et al., 2004, 2005; Morris et al., 2001; Napoli, 1996; Sorefan et al., 2003; Turnbull et al., 2002). This suggests the existence of an upwardly mobile signal that inhibits bud outgrowth. Because the reciprocal grafts with wild-type shoots grafted to mutant roots also have wild-type shoot branching, this signal must be produced in both shoots and roots.

In Arabidopsis, mutations at four MAX loci have been described (Booker et al., 2004, 2005; Sorefan et al., 2003; Stirnberg et al., 2002). Double mutant and reciprocal grafting analyses demonstrate that all these genes act in a single pathway. MAX1, MAX3 and MAX4 act in the production of the upwardly mobile signal, with MAX3 and MAX4 acting upstream of MAX1. MAX2 most likely acts in the perception and/or transduction of the signal (Booker et al., 2005). The molecular identities of the MAX genes are consistent with this model. MAX3 and MAX4 are divergent members of the carotenoid cleavage dioxygenase (CCD) family of enzymes (Booker et al., 2004; Sorefan et al., 2003). MAX3 (CCD7) and MAX4 (CCD8) can cleave a range of carotenoid substrates when expressed in Escherichia coli (Booker et al., 2004; Schwartz et al., 2004), although the in planta substrates for these enzymes remain to be determined. MAX1 encodes a cytochrome p450 family member (Booker et al., 2005). MAX2 encodes an F-box protein, suggesting MAX2-mediated targeted protein degradation operates in the MAX pathway to control bud outgrowth (Stirnberg et al., 2002).

The MAX4, RMS1 and DAD1 genes have all been shown to be orthologous, suggesting that aspects of shoot branching control are conserved in all three species (Snowden et al., 2005; Sorefan et al., 2003). In addition, mutations in a rice MAX2 homologue have recently been shown to result in an increased branching phenotype (Ishikawa et al., 2005), indicating an ancient origin for this pathway, which has been conserved in both monocot and dicot species.

All of these branching mutants so far tested exhibit auxin resistant bud outgrowth (Beveridge et al., 2000; Sorefan et al., 2003), suggesting that the upwardly mobile signal is involved in auxin-mediated bud inhibition. In pea, Reverse transcription-Polymerase Chain Reaction (RT-PCR) analysis has shown that the expression of RMS1 in nodal tissue is upregulated by auxin and downregulated by removal of the primary apex: an in vivo source of auxin (Foo et al., 2005; Sorefan et al., 2003). This immediately suggests a mechanism by which auxin could act through increasing the synthesis of the RMS-dependent signal to inhibit branching. In addition, mutations in rms3, rms4 and rms5, as well as mis-sense mutation in rms1, result in increased RMS1 expression, suggesting that either increased branching or lack of RMS-dependent signalling results in feedback upregulation of RMS1 expression. The petunia DAD1 gene also shows this feedback regulation in shoots, although to a lesser extent (Snowden et al., 2005). This suggests that RMS1 and DAD1 expression may represent an important point of control for shoot branching in pea and petunia respectively, although the functional significance of these transcriptional changes has not been firmly established. In contrast, expression of the Arabidopsis orthologue, MAX4, is upregulated by auxin in roots, but not in nodal tissue (Sorefan et al., 2003). This suggests that transcriptional control of RMS1 and MAX4 play different roles in the regulation of shoot branching and that the interaction of the RMS pathway with auxin differs significantly from the interaction between the MAX pathway and auxin. Here, we describe experiments aimed at testing these hypotheses further, including characterization of MAX4 promoter::GUS (M4p::GUS) expression in detail, and analysis of the effects of overexpression of MAX4.


M4p::GUS expression

The use of the M4p::GUS construct to characterize tissue-specific MAX4 expression was first reported by Sorefan et al. (2003). Three independent M4p::GUS transgenic lines were examined and found to show identical GUS expression patterns. Strong expression was observed in the root tips of 5-day-old seedlings and weak expression in nodal tissue in the mature stem. In addition, upregulation of MAX4 expression was found to occur in the root elongation zone following 24 h auxin treatment. However, in contrast to the situation for the orthologous RMS1 gene in pea, no evidence was found for auxin-induced upregulation of MAX4 expression in nodal tissue.

The low levels of MAX4 expression make it difficult to detect MAX4 transcript by Northern blot analysis. Semiquantitative RT-PCR, using RNA extracted from 7-day-old seedlings, confirmed that levels of M4p::GUS expression are similar to those of the endogenous MAX4 gene (Figure 1a), indicating that the M4p::GUS plants are a useful tool for characterizing MAX4 expression. Histochemical GUS staining revealed that M4p::GUS expression was first evident in root tissue soon after germination. Faint staining was seen in root tip cells of 1-day-old seedlings, which became stronger after 2 days growth (Figure 1b,c). A shorter period of GUS staining showed that M4p::GUS expression was restricted to the columella root cap (Figure 1d). This staining was also seen in lateral roots (Figure 1e). Expression was not seen in early lateral root primordia, but became evident as the lateral root tip emerged from the primary root. Root tip expression was maintained throughout the life cycle of the plant and was present even after 10 weeks growth, when outgrowth of lateral shoot branches had ceased and senescence of aerial tissue had begun (data not shown). The only other site of strong and consistent M4p::GUS expression was in the stylar and stigmatic tissues of mature flowers (Figure 1f,g) and in siliques (Figure 1h).

Figure 1.

(a) Semiquantitative RT-PCR comparing levels of endogenous MAX4 transcript with GUS transcript expressed from the MAX4 promoter (M4p::GUS) in RNA extracted from whole 7-day-old M4p::GUS transgenic seedlings.
(b–h) Histochemical staining for GUS activity in M4p::GUS transgenic plants showing expression in germination seedlings (b, c); primary root tip of 5-day-old seedling (d); lateral root tip of 18-day-old seedling (e); immature siliques (f, g) and mature siliques (h).

M4p::GUS expression was also observed more weakly and less consistently in several other tissues. In agreement with previous findings (Sorefan et al., 2003), M4p::GUS expression was observed in nodal tissue surrounding buds (Figure 2a). Staining was never seen in the buds themselves, or the young outgrowing branches. In the majority of nodes (59/78), no visible staining was observed. In the remainder, weak staining was present both above and below the bud and occasionally in the petioles of cauline leaves. GUS staining was observed in nodal tissue supporting buds of different sizes, at different positions along the primary inflorescence and in primary inflorescences of different heights. No correlation was found between any of these factors and the presence of GUS staining (data not shown). Thus, it is unclear why MAX4 is expressed at detectable levels in some stems but not in others. Expression of M4p::GUS was also occasionally seen in hypocotyl tissue and petioles of rosette leaves (data not shown). No other sites of expression were observed in wild-type (WT) plants at any stage of development.

Figure 2.

The effect of auxin on MORE AXILLARY BRANCHING 4 (MAX4) expression.
Histochemical staining for GUS activity in MAX4 promoter::GUS (M4p::GUS) transgenic plants.
(a) Untreated wild-type nodal tissue.
(b, c) Nodal tissue of wild-type plants following 72 h application of 1 μm NAA to the cut apical surface.
(d) Nodal tissue of untreated Yucca auxin-overproducing plants.
(e) RT-PCR of MAX4 mRNA amplified from the stems of nodal segments with or without 72 h application of 1 μm NAA (auxin) to the cut apical surface.
(f–h) Primary root tip treated with 1 μm NAA for 6 h (f), 8 h (g) and 16 h (h). For untreated controls see Figure 3(a).
(i, j) Hypocotyls of wild-type (i) and Yucca (j) seedlings.

Auxin and MAX4 expression in nodal tissue

RT-PCR analysis of RMS1 expression in pea found a strong upregulation in nodal tissue in response to apically supplied auxin (Foo et al., 2005; Sorefan et al., 2003). In contrast, no evidence for upregulation of MAX4 expression in nodal tissue was found using either RT-PCR or M4p::GUS analysis (Sorefan et al., 2003). To investigate further whether MAX4 was transcriptionally regulated by auxin at the node, we examined whether application of exogenous auxin or elevated levels of endogenous auxin could result in increased M4p::GUS expression. Nodes of M4p::GUS plants carrying buds of <1.5 mm in length, which still have the capability to respond to auxin (Chatfield et al., 2000), were excised and placed between blocks of agar with 1 μm Naphthalene acetic acid (NAA) or a control treatment of ethanol added to the apical block. Staining in nodal tissue remained variable, as shown by representative nodes in Figure 2(a–c). GUS staining was seen both in untreated nodes (Figure 2a) and in nodes treated with 1 μm NAA for up to 72 h (Figure 2b), with little evidence for upregulation of M4p::GUS expression in NAA treated nodes. In each case, the majority of nodes had no visible staining (Figure 2c); control (24/26 unstained) and NAA (17/26 unstained).

To test whether M4p::GUS expression in these assays reflects the expression of the endogenous MAX4 gene, RT-PCR experiments were performed using RNA extracted from the stems of excised nodal segments treated in the same way as for the GUS experiments with, or without 72 h exposure to 1 μm apically applied NAA. There was no difference in the accumulation of MAX4 message in auxin-treated versus untreated samples (Figure 2e).

To determine whether increased levels of endogenous auxin could induce MAX4 expression, M4p::GUS was crossed into the yucca background. yucca plants overproduce auxin (Zhao et al., 2001). In this background, the majority of nodes (6/11) again exhibited no GUS staining (Figure 2d). Staining in yucca nodes was of similar intensity to that seen in WT. These data do show a general trend towards more GUS-positive nodes in high-auxin stems; however, in every case, the majority of stems showed no detectable GUS activity and the RT-PCR experiment did not detect any evidence for auxin-induced expression of MAX4 in stem tissue.

Auxin and MAX4 expression in roots and hypocotyls

In contrast to the situation in nodal tissue, application of exogenous 1 μm NAA for 24 h reliably upregulates M4p::GUS expression in the elongation zone of the root (Sorefan et al., 2003). To determine the time course of MAX4 induction, 1 μm NAA was added to seedlings for different lengths of time. Figure 2(f) shows that 6 h treatment with exogenous auxin did not increase M4p::GUS expression above that seen in untreated seedlings. However, 8 h incubation was sufficient to induce expression in the root elongation zone (Figure 2g). This expression was maintained and perhaps increased, following 16 h incubation (Figure 2h; for untreated roots see Figure 3a). A similar induction in hypocotyl tissue was also observed (data not shown). In the yucca background, expression of M4p::GUS was evident in untreated hypocotyls (Figure 2j), whereas untreated WT plants only rarely showed staining in this tissue (Figure 2i). No expression was evident in the elongation zone of the root of untreated [M4p::GUS]yucca plants, but such expression was induced following treatment with exogenous auxin (data not shown).

Figure 3.

Auxin-induced expression of M4p::GUS is AXR1-dependent.
Histochemical staining for GUS activity in M4p::GUS transgenic plants.
(a, c, e) Wild-type seedlings and (b, d, f) axr1-12 seedlings.
(a, b) Untreated roots.
(c, d) Roots treated with 1 μm NAA for 24 h.
(e, f) Hypocotyls treated with 1 μm NAA for 24 h.
For untreated wild-type hypocotyls see Figure 2(i).

To investigate whether auxin-induction of MAX4 expression acts via the AXR1 pathway, the M4p::GUS transgene was crossed into the axr1-12 auxin resistant mutant background. In the absence of auxin, expression of M4p::GUS in the axr1-12 background was indistinguishable from that seen in the WT background (Figure 3a,b). When seedlings were supplied with 1 μm NAA for 24 h, upregulation in the elongation zone of axr1-12 seedlings still occurred, but was greatly reduced relative to WT (Figure 3c,d). The auxin-induced upregulation of M4p::GUS expression seen in hypocotyl tissue in the WT background (Figure 3e) was completely abolished in axr1-12 plants (Figure 3f).

Overexpression of MAX4

It is possible that the auxin-induced expression of MAX4 in hypocotyl and root tissue is required for the inhibition of branching in wild-type plants, and the reduced auxin-induced expression of MAX4 in the axr1-12 mutant background could contribute to the bushy phenotype of axr1-12 plants. To test this hypothesis, MAX4 was overexpressed from the Cauliflower Mosaic Virus 35S (CaMV 35S) promoter in the axr1-12 background. The 35S::MAX4 construct was described by Sorefan et al. (2003). A strongly expressing line was selected from six independent lines that were taken to homozygosity and characterized from the original WT Columbia (Col) transformation. This transgene was crossed into the axr1-12 background. [35S::MAX4]axr1-12 plants were found to show no significant alteration in branching from the axr1-12 mutant (P = 0.132;Figure 4a).

Figure 4.

Overexpression of MORE AXILLARY BRANCHING 4 (MAX4).
(a) The number of secondary rosette branches produced by 7-week-old WT and axr1-12 plants (n = 20), with or without overexpression of the MAX4 gene from the CaMV 35S promoter.
(b) Time course of outgrowth for buds (n = 7–24) on isolated nodes of wild-type (squares), max4-1 mutants (triangles) and wild-type plants overexpressing MAX4 from the CaMV 35S promoter (diamonds), with (black symbols) and without (white symbols) apical auxin application.
Error bars represent the standard errors of the means.

We also investigated whether overexpression of MAX4 had any effect on branching in a wild-type background. Preliminary analysis had already indicated that high levels of MAX4 overexpression did not noticeably reduce branching (Sorefan et al., 2003). Quantification confirmed no significant difference in the number of branches produced by WT and [35S::MAX4]WT plants (Figure 4a). To test for branch reduction in a situation where a greater number of branches are formed, max4, WT and [35S::MAX4]WT plants were grown in short days for 28 days before transfer to long photoperiods and decapitation of the primary bolt. The number of rosette branches was scored 10 days later. In these conditions, max4 mutants produced 12.7 ± 0.4 branches (mean ± SE, n = 25), WT produced 6.3 ± 0.2 and [35S::MAX4]WT 6.0 ± 0.2, which is not significantly different from WT.

The response of [35S::MAX4]WT buds to apical auxin was also determined. In the absence of apical auxin, [35S::MAX4]WT buds grew out with similar kinetics to those of WT plants (Figure 4b). In response to apical addition of 1 μm NAA, a slight delay was seen in the outgrowth of [35S::MAX4]WT buds relative to WT. Although the difference is small, it is reproducible and significant at two points of the graph (144 and 168 h).

Post-transcriptional interactions between MAX4 and auxin

Taken together, these data suggest that transcriptional control of MAX4 by auxin does not play a major role in the control of branching in Arabidopsis, either directly or downstream of auxin. We therefore sought to determine whether auxin might regulate the stability of the MAX4 protein. A CaMV 35S::MAX4-myc tagged construct was generated and transformed into max4 plants. This construct was found to restore wild-type branching to the max4 mutant, indicating that the MAX4-myc protein is fully functional. The use of the 35S promoter ensures that any change in the abundance of the protein is not due to transcriptional regulation. To determine the effect of auxin on MAX4-myc accumulation, 7-day-old [35S::MAX4myc]max4 seedlings were treated with 1 μm NAA. Immunoblotting of extracts prepared from these plants showed no increase in the levels of MAX4-myc in response to 2, 4 or 8 h auxin treatment (Figure 5a). Neither was any shift in mobility of the MAX4-myc protein detected, suggesting that auxin does not lead to a mass-shifting covalent modification of MAX4-myc, such as phosphorylation.

Figure 5.

Post-transcriptional interactions between MORE AXILLARY BRANCHING 4 (MAX4) and auxin.
(a) Abundance of MAX4-myc protein expressed from the CaMV 35S promoter in transgenic seedlings treated with 1 μm NAA for 0, 2, 4 or 8 h, detected by Western blotting and anti-myc antibody.
(b) Number of secondary rosette branches produced by chimeric plants generated by hypocotyl grafting (shoot genotype/root genotype).
Error bars represent the standard errors of the means; n = 4–9.

The data described so far suggest that interaction between auxin and the MAX pathway does not appear to involve significant regulation of MAX4 levels. Consistent with this, previous grafting analysis by Booker et al. (2003) demonstrated that AXR1 acts primarily in the shoot to inhibit bud outgrowth, while in contrast WT roots are able to rescue the branching of a max4 mutant shoot, indicating that MAX4 can act in the root to inhibit bud growth (Sorefan et al., 2003). Thus, it appears that the site of auxin action and the site of synthesis of the MAX4-dependent signal can be separated without affecting shoot branching. To test this idea more directly, reciprocal grafting analysis using the max4 and axr1-12 mutants was undertaken. The results show that self-grafted max4 and axr1-12 controls have the increased branching phenotype characteristic of these mutants (Figure 5b). In contrast, max4 shoots grafted to axr1-12 roots exhibit a WT branching phenotype, similar to wild-type roots grafted to max4-1 mutant shoots (Figure 5b).

These results confirm that the shoot is the main site of action for AXR1 in mediating branch inhibition and that, in plants where MAX4 activity and AXR1 activity never occur in the same cells, branching levels are reduced to those usually observed in self-grafted wild-type plants (Booker et al., 2005). This indicates that auxin interacts with the MAX pathway downstream of the MAX4-catalysed step and downstream of the production of a mobile component, in the synthesis of the MAX-dependent signal.

Feedback regulation and MAX4 expression

In pea, RT-PCR analysis revealed a large increase, of up to three orders of magnitude, in the levels of RMS1 expression in the epicotyls and nodes of rms1, rms3, rms4 and rms5 mutant backgrounds relative to that seen in WT plants (Foo et al., 2005). A qualitatively similar, but quantitatively much smaller result was found in petunia stems, but not in other tissues, where mutations in other DAD genes resulted in up to fourfold increased DAD1 expression (Snowden et al., 2005). This suggests that in rms1, rms3, rms4 and rms5 mutants and to a lesser extent dad mutants, feedback upregulation of RMS1 or DAD1 expression occurs in response to increased shoot branching or, more directly, in response to the absence of RMS or DAD signalling.

To determine whether a similar mechanism of feedback control regulates MAX4 expression in Arabidopsis, M4p::GUS expression was examined in the background of the other max mutants. The results show no differences in M4p::GUS expression in any of the max mutant backgrounds relative to WT in either the root cap columella, or in the elongation zone on addition of auxin (data not shown). In hypocotyl tissue, increased GUS staining relative to WT was observed in the hypocotyl of the max2 mutant, but was unaltered in the other max mutant backgrounds (Figure 6a–e). Furthermore, M4p::GUS expression occasionally occurred at the base of petioles and in the petioles themselves in the max mutant backgrounds (Figure 6h). This staining was observed in only a minority of the [M4p::GUS]max plants examined (12/36), but was never observed in [M4p::GUS]WT plants (0/12). These few tissue-specific differences in MAX4 expression in max mutant backgrounds compared with WT argue against the strong feedback regulation observed in pea. Furthermore, if there is an effect in Arabidopsis, it appears to be less than that observed in petunia. This conclusion is supported by the fact that semiquantitative RT-PCR confirmed similar levels of MAX4 expression in the aerial tissue of 3-week-old max and WT plants (Figure 6f,g).

Figure 6.

MORE AXILLARY BRANCHING 4 (MAX4) expression in max mutant backgrounds.
(a–e) Histochemical staining for GUS activity in the hypocotyls of M4p::GUS transgenic 5-day-old seedlings. (a) WT, (b) max1, (c) max2, (d) max3, (e) max4.
(f) Semiquantitative RT-PCR, showing levels of MAX4 expression in max mutant backgrounds. RNA was extracted from aerial tissue of 3-week-old plants grown under a long photoperiod. A product is visible after 35 cycles in all backgrounds.
(g) Actin control, showing approximately equal loading of all samples.
(h) Histochemical staining for GUS activity in M4p::GUS transgenic max3 rosettes.

Cytokinin and the control of MAX4 expression

Cytokinin is a well-known positive regulator of bud outgrowth and, in pea, cytokinin production and RMS signalling have been shown to interact (Beveridge, 2000; Morris et al., 2001). To investigate whether cytokinin affects MAX4 transcription, the regulation of M4p::GUS expression in response to the synthetic cytokinin, benzyladenine (BA) was characterized in 5-day-old seedlings. The addition of cytokinin alone had no effect on M4p::GUS expression, even following 24 h treatment with 10 μm BA or up to 72 h treatment with 1 μm BA (data not shown). However, incubation with 1 μm BA together with 1 μm NAA resulted in a reduced upregulation of M4p::GUS expression in the elongation zone compared with treatment with NAA alone (Figure 7).

Figure 7.

Effect of cytokinin on M4p::GUS expression.
Histochemical staining for GUS activity M4p::GUS the primary root tip of transgenic plants treated with 1 μm NAA for 24 h, without (a) or with (b) addition of 1 μm benzyladenine (BA) for 24 h.


The RMS1 gene of pea, the MAX4 gene of Arabidopsis and the DAD1 gene of petunia are orthologous (Snowden et al., 2005; Sorefan et al., 2003). They are all required for the production of a long-range, graft-transmissible upwardly mobile signal that inhibits branching. They each function in pathways in which mutations in each of the other pathway genes result in similar phenotypes: rms3, rms4 and rms5 in pea; max1, max2 and max3 in Arabidopsis; and dad2 and dad3 in petunia (Beveridge, 2000; Booker et al., 2004, 2005; Morris et al., 2001; Napoli, 1996; Snowden et al., 2005; Sorefan et al., 2003; Stirnberg et al., 2002; Turnbull et al., 2002). A MAX2 orthologue has also recently been described in rice (Ishikawa et al., 2005). Despite the wide-spread conservation of this shoot branching control system, comparison of the data we describe here on the expression of MAX4 with published data on the expression of RMS1 and DAD1 demonstrates major differences in the way in which these pathways are regulated.

Of particular interest is the relationship between the RMS and MAX genes and auxin. The axillary buds of rms and max mutants show resistance to the inhibitory effects of apical auxin, indicating that the MAX/RMS pathways are required for full auxin-mediated inhibition of bud outgrowth (Beveridge et al., 2000; Sorefan et al., 2003). In pea, RMS1 expression was found to be dramatically upregulated in nodal tissue in response to apical auxin addition and downregulated by decapitation, which removes an endogenous auxin source (Foo et al., 2005). Although the functional significance of this upregulation has not been established, it immediately suggests a mechanism for auxin-mediated bud inhibition, via transcriptional upregulation of RMS1 and consequent increased production of the RMS-dependent branch-inhibiting signal. In contrast, our data show little change in MAX4 expression in nodal tissue in response to auxin.

Auxin treatment does result in AXR1-mediated upregulation of MAX4 expression in the elongation zone of the root and in hypocotyl tissue. Consistent with this, expression was observed in untreated hypocotyls of Yucca mutants, which overproduce auxin (Zhao et al., 2001). In contrast, exogenous-auxin-independent expression was not observed in the elongation zone of Yucca roots. It is possible that the levels of ectopic auxin accumulation in the root elongation zone are not high enough to induce MAX4 transcription.

To test the functional significance of the auxin-induced expression of MAX4, we carried out MAX4 overexpression experiments and grafting studies with the axr1 mutant. The results demonstrate that auxin-regulated MAX4 expression in the root and hypocotyl does not drive, nor is it required for the inhibition of branching. These data suggest that, in contrast to its pea orthologue, the transcriptional control of MAX4 through auxin signalling is not a major point of interaction between these pathways in controlling bud outgrowth.

Because auxin-induced MAX4 transcription is apparently not required for bud growth inhibition, we investigated the possibility that auxin might promote MAX4 activity at the protein level. The abundance of myc-tagged MAX4 protein was found to be unaltered by incubation with auxin and no mass change, which could be indicative of a post-translational modification, was observed. Although it is still possible that auxin could act to increase activity of the MAX4 protein in another way, the analysis of the effects of reciprocal grafting between max4 and axr1 mutants indicates that auxin signalling and MAX4 can be active in completely separate tissues and still result in wild-type branching. Thus, the major point of interaction between auxin and the MAX pathway is after the MAX4-catalysed reaction and after the production of a mobile compound. This could involve auxin regulating the nodal transcription or function of genes downstream of MAX4 in the MAX pathway. Alternatively, auxin may act post-synthesis of the MAX-dependent compound. This second hypothesis seems more likely because the major site for auxin action has previously been shown to be the shoot (Booker et al., 2003), but expression in the root only is sufficient to suppress branching for each of the three known genes involved in the synthesis of the MAX-dependent compound (Booker et al., 2004; Turnbull et al., 2002), AXR1 expression in xylem-associated cells in the axr1-12 background was found to be sufficient to restore to wild-type the highly branched phenotype of the axr1-12 mutant (Booker et al., 2003). As the MAX-dependent compound is acropetally but not basipetally mobile, it may be transported in the xylem (Turnbull et al., 2002). Thus, there is a likely juxtaposition of the upwardly mobile compound, the critical site for auxin signalling and the main conduit for auxin transport, suggesting that the main point of interaction between these pathways could be in this tissue.

The effect of MAX-pathway feedback on MAX4 expression

In addition to upregulation by auxin, RMS1 expression is also altered in the backgrounds of other rms mutants. Four of the rms mutants exhibit large increases in RMS1 expression (Foo et al., 2005). The same is true, but to a lesser extent in petunia, where upregulation of the orthologous DAD1 gene is observed in some tissues in the dad mutant backgrounds (Snowden et al., 2005). These data provide evidence of feedback upregulation, in which reduced RMS or DAD pathway function leads to upregulated RMS1 or DAD1 transcription. In pea, the feedback regulation of RMS1 transcription is dependent on RMS2 gene function (Foo et al., 2005). The rms2 mutant has a phenotype similar to the other rms mutants and current models for the regulation of branching in pea suggest that the RMS2 gene is required to activate optimal transcription of the other RMS genes, because RMS1 is expressed at below-wild-type levels in rms2 mutants. In contrast to the situation in pea, we found little evidence for significant upregulation of MAX4 expression in max mutant backgrounds. This suggests that the RMS2-dependent feedback regulation system operating in pea does not occur in Arabidopsis, at least at the level of MAX4 transcription.

The only exception to this observation is in the max2 hypocotyl, where we consistently observe GUS expression above wild-type levels. It is important to note that max2, uniquely among the max mutants, has early seedling phenotypes suggestive of defects in light signalling, including a long hypocotyl (Stirnberg et al., 2002). It is therefore possible that the upregulation of MAX4 in max2 hypocotyls reflects this hypocotyl elongation phenotype, rather than feedback regulation.

Control of MAX4 expression by cytokinin

Another factor known to control branching is cytokinin, which exerts a strong and direct stimulatory effect on the outgrowth of buds. Basally supplied cytokinin can overcome the inhibitory effects of apical auxin (Chatfield et al., 2000). Measurements of cytokinin levels in the rms mutants of pea suggest that the RMS pathway also exerts feedback control of cytokinin levels, again via the action of the RMS2 gene (Beveridge, 2000), because the rms1, rms3, rms4 and rms5 mutants all have decreased levels of root xylem sap cytokinin, whilst the rms2 mutant has elevated root xylem sap cytokinin (Beveridge, 2000; Morris et al., 2001).

It is not known what effect the max mutants have on root xylem sap cytokinin, but we have shown that the MAX pathway and cytokinin interact in the root in that the auxin-stimulated upregulation of MAX4 expression in the elongation zone is attenuated by simultaneous application of cytokinin. If the same cytokinin interactions occur in the MAX and RMS pathways, it would suggest the existence of a balanced regulatory loop between the root and shoot in which increased cytokinin production in the root would result in decreased MAX4/RMS1 expression. The combination of more cytokinin and less MAX/RMS-dependent compound production by the root would result in increased shoot branching, which would result in feedback inhibition on root cytokinin production and increased MAX4/RMS1 expression. In addition to RMS2, shoot-derived auxin could form a component of this loop, because more active branches would increase auxin supply to the roots, where, as a result, MAX4/RMS1 transcription would be upregulated (Figure 3) and cytokinin synthesis would be downregulated (Nordström et al., 2004).

The effects of overexpression of MAX4

Thus far, none of the transcriptional changes described above has been shown to be important in achieving a wild-type shoot branching habit. In pea, this is due to the difficulty in manipulating gene expression levels experimentally. In Arabidopsis, overexpression of MAX4 from the CaMV 35S promoter provides an easy method for constitutive expression of MAX4. The lack of major phenotypic effects from such overexpression are consistent with the idea that post-transcriptional regulation is more important in controlling MAX-pathway function. Alternatively, overexpression of other components of the pathway in addition to MAX4 may be needed to increase levels of the inhibitory compound and further reduce bud outgrowth. For example, it has been shown that MAX3 and MAX4 can act sequentially in cleavage of the same carotenoid substrate (Schwartz et al., 2004). It would not, therefore, be surprising if overexpression of MAX3 and MAX4 together was required in order to increase MAX pathway flux and cause increased inhibition of bud outgrowth. It is also possible that more significant phenotypic effects may occur in as yet untested environmental conditions. This is perhaps hinted at by the increased auxin-induced bud inhibition observed in the 35S-MAX4 plants. Alternatively, the interacting hormonal network that regulates branching may compensate for overexpression of MAX4 such that the branch-inhibiting effects of increased MAX4 action are balanced by post-transcriptional feedback regulation and/or branch promoting factors such as increased cytokinin production.


The four Arabidopsis MAX genes are all members of multigene families, but in each case their closest family relatives are not in the plant kingdom, indicating that ancestral genes from each family were recruited to regulate shoot branching in plants. That this event was relatively early in the evolution of flowering plants is demonstrated by the conservation of MAX gene function across monocots and dicots. Here, we have shown that the regulation of MAX4 expression by auxin and by MAX-pathway feedback is not conserved with its pea orthologues. This suggests the interesting possibility that diversity in shoot architectural forms might be partly achieved through variation in MAX-orthologous-pathway regulation between species. Studies of regulation of family members introduced transgenically into heterologous species will help to determine whether this variation includes cis- and/or trans-acting factors.

Experimental procedures

Sterile plant growth conditions

Arabidopsis seedlings were grown under sterile conditions on vertically oriented petri dishes or, for older plants, in 1-L Weck jars containing Arabidopsis thaliana salts (ATS) agar medium (Wilson et al., 1990). Seeds were surface sterilized and cold treated at 4°C for 2 to 4 days. Seedlings were grown at 21°C under a 16-h-light/8-h-dark regime and a light intensity of 100–120 μmol m−2 sec−2. These plants were used to examine GUS expression and for the excised nodes used in split plate assays (see below). Plants expressing the 35S::MAX4-myc construct were grown in liquid ATS at 22°C under a 16-h photoperiod for 7 days.

Generation of transgenic plants

The MAX4 promoter::GUS (M4p::GUS) construct and transgenic lines were as described by Sorefan et al. (2003). WT T3 plants expressing this construct were crossed into the Yucca, axr1-12 and max mutant backgrounds. F3 homozygotes were determined by selection for homozygosity for the appropriate mutant phenotype and resistance to kanamycin.

The CaMV35S::MAX4 construct and transgenic lines were generated as described by Sorefan et al. (2003). The line expressing MAX4 at the highest levels was crossed into the max4 background and found to rescue the mutant phenotype. This line was also crossed into the axr1-12 background. F3 plants homozygous for the axr1-12 mutation and the 35S::MAX4 construct were selected on the basis of their hygromycin resistance and the axr1-12 mutant phenotype.

The myc tagged MAX4 construct was generated by PCR amplifying the coding sequence of MAX4 cDNA using the forward primer (5′-CTGGTACCATGGCTTCTTTGATCACAACC-3′), incorporating a KpnI site 5′ to the ATG, and the reverse primer (5′-AATCTAGATCAAAGATCTTCTTCAGAAATAAGCTTTTGTTCATCTTTGGGGATCCAGCA-3′) including sequence encoding the myc epitope tag, and a 3’ XbaI site. The PCR product was ligated into the pCR2.1-Topo vector (Invitrogen, Paisley, UK) cut with KpnI and XbaI, and was ligated into a modified version of the pCambia 1300 vector (Cambia, Canberra, Australia).

The construct was transformed into wild-type and max4-1 plants via Agrobacterium tumefaciens using the floral dip method of Clough and Bent (1998). Homozygous T3 lines were obtained by self-fertilization, followed by selection on hygromycin and basta. Four independent homozygous lines were obtained and their branching phenotype characterized. A line showing the complete rescue of the mutant branching phenotype was chosen for analysis of the MAX4 protein.

Hormone treatments

Seedlings were removed from ATS plates and placed in Eppendorf tubes containing liquid ATS with the appropriate hormone treatments or an equal volume of ethanol. The Eppendorf tubes were returned to the 21°C growth conditions until transfer of the seedlings into GUS stain. Nodes were excised from plants and placed in split plates as described below.

GUS staining

Histochemical analysis of the GUS reporter enzyme was performed by incubating tissue in GUS staining buffer containing 0.3 mg ml−1 X-Gluc (5-bromo-4-chloro-3-indolyl-β-glucuronide), 50 mm potassium phosphate (pH 7), 0.1 mm potassium ferrocyanide, 0.1 mm potassium ferricyanide, 10 mm EDTA, 0.1% (volume in volume, v/v) Triton X-100. Samples were incubated for 1 to 16 h in reaction buffer and destained in 70% (v/v) ethanol before observation. For observation of root tip staining, tissue was cleared by incubation in 0.24 m HCl in 20% (v/v) methanol, followed by incubation in 7% (v/v) sodium hydroxide in 60% (v/v) ethanol (Malamy and Benfey, 1997), and rehydrated in a graded ethanol series.

Split plate assay

The hormone responses of buds on isolated nodal stem segments were assessed as described in Chatfield et al. (2000).

Decapitation experiment

Arabidopsis plants were grown in 4-cm square compartments (P40; Cookson Plantpak, Maldon, UK) containing F2 compost treated with Intercept 70WG (both Levington Horticulture, Ipswich, UK). After 3 days cold treatment at 4°C, trays were transferred to a growth cabinet set to 21°C, with a light intensity of 200 μmol m−2 sec−1 and 8-h photoperiods. After 28 days under these conditions, plants were transferred to 16-h photoperiods to induce flowering (Greb et al., 2003). Primary bolts were decapitated once they reached 10 to 15 cm in length to encourage the outgrowth of secondary rosette branches. Rosette branches were counted 10 days after decapitation.

Grafting analysis

Grafting was performed on seedlings as described by Turnbull et al. (2002). The max4, axr1-12 and WT (Col) seedlings were used. Grafted seedlings were transferred to compost in 4-cm square compartments as described above. Plants were grown for 5 weeks under long day conditions before characterization of the shoot branching phenotype and verification of graft integrity.

Semiquantitative RT-PCR

For the quantification of MAX4 expression relative to M4p::GUS expression, M4p::GUS seedlings were grown for 7 days in liquid culture. For analysis of MAX4 expression in the max mutant backgrounds, plants were grown on soil under a long photoperiod for 3 weeks before harvesting of aerial tissue. RNA was extracted using Trireagent (Sigma, St Louis, MO, USA), treated with RQ1 DNase (Sigma) and reverse transcribed using Superscript II (Invitrogen) according to manufacturers’ instructions. PCR was carried out using a standard cycle with actin, GUS, MAX4 or specific primers, spanning the first intron of the gene. Five microlitres was taken from the reaction after 20 cycles and after every five subsequent cycles. Samples were run on 1% agarose gels to visualize the PCR products.

Analysis of MAX-myc protein levels

Following 7 days growth of MAX4-myc transgenic seedlings in liquid ATS, 1 μm NAA was added to the media for 2, 4 or 8 h. Tissue was frozen and protein extracts were prepared, processed and analysed by Western blot as described in Kepinski and Leyser (2004).


We thank Stephen Day for critical reading of the manuscript, the York horticultural staff for plant care and Harry Klee for helpful discussions. This work was funded by the Biotechnology and Biological Sciences Research Council of the UK.