Strigolactone acts downstream of auxin to regulate bud outgrowth in pea and Arabidopsis.

During the last century, two key hypotheses have been proposed to explain apical dominance in plants: auxin promotes the production of a second messenger that moves up into buds to repress their outgrowth, and auxin saturation in the stem inhibits auxin transport from buds, thereby inhibiting bud outgrowth. The recent discovery of strigolactone as the novel shoot-branching inhibitor allowed us to test its mode of action in relation to these hypotheses. We found that exogenously applied strigolactone inhibited bud outgrowth in pea (Pisum sativum) even when auxin was depleted after decapitation. We also found that strigolactone application reduced branching in Arabidopsis (Arabidopsis thaliana) auxin response mutants, suggesting that auxin may act through strigolactones to facilitate apical dominance. Moreover, strigolactone application to tiny buds of mutant or decapitated pea plants rapidly stopped outgrowth, in contrast to applying N-1-naphthylphthalamic acid (NPA), an auxin transport inhibitor, which significantly slowed growth only after several days. Whereas strigolactone or NPA applied to growing buds reduced bud length, only NPA blocked auxin transport in the bud. Wild-type and strigolactone biosynthesis mutant pea and Arabidopsis shoots were capable of instantly transporting additional amounts of auxin in excess of endogenous levels, contrary to predictions of auxin transport models. These data suggest that strigolactone does not act primarily by affecting auxin transport from buds. Rather, the primary repressor of bud outgrowth appears to be the auxin-dependent production of strigolactones.

Classical decapitation and replacement experiments by Thimann andSkoog (1933, 1934) suggested that the plant hormone auxin originating from the shoot tip acted as a repressive signal for axillary bud outgrowth at nodes below the shoot tip. However, unraveling the action of apically derived auxin has been challenging, mainly because auxin was found to move strictly downward in the vascular cambium of the stem (the polar auxin transport stream) and apparently cannot change direction to move upward to enter axillary buds and branches (Snow, 1937;Hall and Hillman, 1975;Morris, 1977;Morris and Thomas, 1978;Bangerth, 1989;Prasad et al., 1993;Booker et al., 2003). In addition, while apically derived auxin moved downward through live cells (Morris and Thomas, 1978), the inhibiting influence was able to be transmitted upward through dead tissue (Snow, 1929). Thus, this "secondary inhibiting influence" of auxin was proposed to act via an inhibiting substance that moved up into buds through the transpiration stream (Snow, 1929(Snow, , 1937. However, despite attempts at identifying the second messenger of auxin action, such a substance was never found (Bangerth, 1989).
Seminal work on auxin canalization by Sachs (1968Sachs ( , 1969 led to the idea that auxin saturation in the transport stream of the main stem could block auxin transport from lateral sources. It was found that the direction of the vascular connections of new buds was influenced by the presence or absence of the apex or subtending leaf (Sachs, 1968). This meant either that auxin depletion in an established vascular stream (e.g. after decapitation by removal of the apex) attracted the formation of new vasculature or that auxin levels in an established vascular stream repelled the formation of new vasculature (i.e. the new vasculature would join other vasculature where there is less auxin). Sachs (1970) proposed that the transport stream of an intact stem, full of auxin, repelled the development of vasculature from buds and thus blocked their outgrowth. Decapitation depleted the apically derived stream of auxin and thus released the buds to grow.
The work of Sachs (1968Sachs ( , 1969 seemingly bypassed the need for a second messenger, as auxin could influence lateral auxin transport from a distance, such as at the vascular junctions in the main stem. This theory initially relied on the idea that a bud needed to form a vascular connection before it could grow. However, large dormant buds that are not actively growing were found to have functional vascular connections (Ali and Fletcher, 1970;Peterson and Fletcher, 1973). Indeed, dormant buds are highly developed and would presumably require vascular connections for early development and growth before they enter a stage of dormancy. Additionally, buds could become active and later reenter a stage of dormancy (Stafstrom and Sussex, 1992;Shimizu and Mori, 1998). In contrast to these findings, Sorokin and Thimann (1964) reported that xylem strands from an axillary bud did not connect with those from the main stem until after release from apical dominance. These vascular strands strengthened with time after release from apical dominance (Sorokin and Thimann, 1964). However, Marr and Blaser (1967) showed that the strengthening of vascular connections of induced axillary buds occurred after, not before, visible bud outgrowth. Sachs (1981) later refined his idea to suggest that it was the increase in bud vasculature that allowed buds to grow.
The studies of Sachs (1968Sachs ( , 1969 focused on vascular development; however, Bangerth (1989) showed that auxin (in this case, indole-3-acetic acid [IAA]) export out of a branch was correlated with its ability to repress the outgrowth of other buds or branches. Thus, it was proposed that auxin export may be a requirement of bud outgrowth, and auxin transport from an earlier developed branch might simply inhibit auxin transport from later developed buds. This competition was suggested to occur at the junctions where the auxin transport streams meet (Bangerth, 1989).
Cytokinins are a class of phytohormone that promotes cell division, and application of cytokinin to buds can induce outgrowth (Sachs and Thimann, 1967). While likely a trigger for bud release, cytokinin may also promote auxin production and basipetal auxin transport out of growing buds, which consequently represses the production of cytokinin lower in the stem and limits its availability for other buds (Bangerth et al., 2000;Tanaka et al., 2006;Shimizu-Sato et al., 2009).
The studies reviewed above have almost exclusively used decapitation and related techniques to induce branching and investigate the role of auxin. Indeed, for this reason, the term apical dominance encouraged a focus on the shoot tip as the source of branching regulation. Importantly, these experiments were valid for investigating the natural phenomenon of the response to decapitation, but the caveat is that auxin may not be the only relevant factor affected by decapitation. Moreover, branching in intact plants may not be regulated by the same processes as those induced by decapitation . Until the isolation and characterization of branching mutants in various species, it was not possible to address these issues. In recent decades, many branching mutants have been isolated that have problems with auxin, cytokinin, or brassinosteroids, for example, and show highly pleiotropic phenotypes typical for these hormones (Lincoln et al., 1990;Azpiroz et al., 1998;Tantikanjana et al., 2001). However, a class of mutants were isolated that displayed a specific increase in bud outgrowth that was not correlated with any known hormonal signal (Beveridge et al., 1996(Beveridge et al., , 1997. Paradoxically, the mutants were found to have generally higher levels of auxin and lower levels of xylem cytokinin, facts difficult to reconcile with ideas about the roles of auxin and cytokinin in regulating bud outgrowth (Beveridge et al., 1997). These mutants were ramosus (rms) in pea (Pisum sativum), decreased apical dominance (dad) in petunia (Petunia hybrida), more axillary growth (max) in Arabidopsis (Arabidopsis thaliana), and particular dwarf (d) mutants in rice (Oryza sativa; Beveridge et al., 1994;Napoli, 1996;Stirnberg et al., 2002;Ishikawa et al., 2005). Grafting studies demonstrated that increased bud outgrowth in some of the mutants was caused by the loss of a long-distance mobile signal (termed SMS; Beveridge, 2006) that moved upward from lower tissues (Beveridge et al., 1994;Napoli, 1996;Foo et al., 2001;Turnbull et al., 2002). Mutant phenotypes were rescued by grafting with wild-type tissue, even in interstock grafts where small pieces of wild-type stem tissue were grafted between mutant rootstock and shoot tissue (Napoli, 1996;Foo et al., 2001). Other mutants were not rescued by grafting but were instead suggested to lack response to SMS (Beveridge et al., 1996;Booker et al., 2005). Grafting studies also showed that outgrowth induced by decapitation in SMS mutant plants cannot be inhibited by IAA applied to the stump unless a wild-type rootstock is present . Thus, SMS production somewhere in the plant is required for IAA to prevent buds growing after decapitation. This means that auxin can move down into the roots and promote the production of SMS and suggests that SMS might in fact be the second messenger for auxin.
Breakthroughs in cloning revealed that rms1, max4, dad1, and d10 SMS synthesis mutant phenotypes were caused by mutations in an orthologous gene, CAROT-ENOID CLEAVAGE DIOXYGENASE8 (CCD8; Sorefan et al., 2003;Snowden et al., 2005;Arite et al., 2007). Likewise, RMS5, MAX3, and D17/HTD1 were found to encode CCD7 (Booker et al., 2004;Johnson et al., 2006;Zou et al., 2006). In contrast, the SMS response mutants, max2, rms4, and d3, were found to be mutated in an orthologous gene encoding an F-box protein (Stirnberg et al., 2002;Ishikawa et al., 2005;Johnson et al., 2006). Recently, the testing of putative carotenoidderived compounds led to the discovery that SMS is a strigolactone or downstream product (Gomez-Roldan et al., 2008;Umehara et al., 2008). Strigolactones are a group of related molecules, thought to be derived from carotenoids (Matusova et al., 2005), that are also involved in promoting arbuscular mycorrhizae symbiosis and parasitic weed seed germination (Cook et al., 1972;Akiyama et al., 2005). It is likely that CCD7 and CCD8 are enzymes that are involved in the production of strigolactones, as the mutants were found to be deficient in strigolactones (Gomez-Roldan et al., 2008; Umehara et al., 2008). Exogenous strigolactone applied to buds or supplied to the roots or the vascular stream was able to rescue the branching phenotype of the ccd7 and ccd8 mutants but not that of the putative SMS response mutants rms4, max2, and d3 (Gomez-Roldan et al., 2008;Umehara et al., 2008).
Prior to the discovery of SMS as a strigolactone, studies with Arabidopsis SMS mutants led to reinterpretation of the theories of Sachs and Bangerth and the establishment of the current auxin transport hypothesis, which proposed that axillary buds compete for limited auxin transport capacity in the main stem (Bennett et al., 2006;. This is based on the premise that in order for an axillary bud to grow it must be able to export auxin and that the main stem of a wild-type plant is saturated with apically derived auxin . A key result in support of the auxin transport hypothesis is that the excessive branching phenotype of Arabidopsis SMS mutant plants could be rescued when grown on medium containing the auxin transport inhibitor N-1-naphthylphthalamic acid (NPA; Bennett et al., 2006). This indicated that auxin transport may be crucial for branching. Bennett et al. (2006) and Lazar and Goodman (2006) also demonstrated that inflorescence stems and rosettes of bolting Arabidopsis SMS mutants had increased expression of genes encoding PIN-FORMED (PIN) polar auxin transport efflux proteins, which correlated with increased protein abundance. An excess of transporters present in the stem at rosette nodes just prior to bud release may encourage auxin to flow from lateral sources and trigger buds to commence growth (Bennett et al., 2006). SMS, therefore, was proposed to act as a regulator of auxin transport by reducing the expression and/or plasma membrane localization of auxin transporters; in the SMS mutants, this inhibition failed to occur (Bennett et al., 2006). It is important to note that because SMS is required for inhibition of decapitation-induced branching by IAA (see above; Beveridge et al., 2000), auxin must regulate SMS production and, according to the auxin transport hypothesis, SMS must then move upward to regulate auxin transport at vascular connections.
Because the control of bud outgrowth involves inputs and regulatory loops from multiple signals, efforts to unravel the hormone interactions have been challenging . In particular, the exact role of auxin content and transport in the regulation of bud outgrowth has been questioned Dun et al., 2006;Ferguson and Beveridge, 2009). Indeed, big and bud1 mutants in Arabidopsis have reduced auxin transport and enhanced branching (Gil et al., 2001;Dai et al., 2006), in contrast to Arabidopsis SMS mutants, which have increased auxin transport and enhanced branching (Bennett et al., 2006). Additionally, depletion of auxin content is not always sufficient to induce bud outgrowth Ferguson and Beveridge, 2009). The recent identification of SMS provides the crucial missing link to assess branching control in plants. We report here on experiments designed to pin down the importance of auxin in bud outgrowth regulation. We show that branching induced by auxin depletion in the main stem following decapitation of pea plants was completely blocked by strigolactone application and that strigolactone reduced branching in Arabidopsis auxin response mutant plants. In addition, the application of an auxin transport inhibitor to pea buds slowed their outgrowth, but only after several days, whereas the response to strigolactone application was rapid. Taken together with auxin transport experiments, our results suggest that auxin transport from buds is not the initial trigger of bud release, although it may be crucial for ongoing bud outgrowth, as we have suggested previously Ferguson and Beveridge, 2009). Rather, it seems likely that auxin promotes strigolactone biosynthesis in the main stem, implying that strigolactone acts as the classical second messenger in apical dominance.

Strigolactone Completely Represses Bud Outgrowth after Decapitation
In addition to causing bud outgrowth, decapitation has been shown to deplete IAA levels in the main stem (Foo et al., 2005;Morris et al., 2005), which in turn may reduce strigolactone levels (Foo et al., 2005;Johnson et al., 2006). If auxin in the stem inhibits bud outgrowth via strigolactones, then strigolactone application to wild-type buds should prevent their outgrowth regardless of decapitation. Indeed, repeated application of GR24 (a synthetic strigolactone; Akiyama et al., 2005) directly to the uppermost axillary bud of wildtype pea plants following decapitation completely inhibited bud outgrowth (Fig. 1, A and B). Untreated buds at lower nodes grew out as normal (data not shown; note branches in Fig. 1A). The fact that GR24 can block bud outgrowth after decapitation implies that apically derived auxin is not required for strigolactone action.

NPA Inhibits Sustained But Not Early Bud Outgrowth
NPA rescues the branching phenotype of SMS mutants of Arabidopsis, which suggests that strigolac-tones may act similarly to auxin transport inhibitors (Bennett et al., 2006). To test whether blocking auxin transport out of a bud inhibits bud outgrowth in a similar way to strigolactone treatment, we applied inhibitive quantities of either NPA or GR24 to small buds of SMS mutant (ccd8) and decapitated wild-type pea plants and compared outgrowth over time (Fig. 1, B and C). Buds were treated at a time when the ccd8 buds had not grown in comparison with the wild-type buds (e.g. at day 0, wild-type buds at node 2 were 0.97 6 0.07 mm and ccd8 buds were 0.81 6 0.03 mm in length). GR24 treatment completely inhibited bud growth in ccd8 and decapitated wild-type plants from the outset, while unexpectedly, comparable NPA-treated buds grew normally for the initial days before growth was suppressed compared with control ccd8 and decapitated wild-type plants. It is possible that bud swelling (as opposed to actual bud outgrowth) caused by an auxin buildup in NPA-treated buds could have contributed to the initial increase in size of NPA-treated ccd8 and decapitated wild-type buds. However, this is unlikely, as no increase in bud size was observed in comparable buds of intact wildtype plants after NPA treatment (Fig. 1D).
Cytokinin application can trigger bud outgrowth (Sachs and Thimann, 1967) in a similar way to decapitation and SMS deficiency ( Fig. 1, B-D). Therefore, we used a cytokinin to induce bud outgrowth in order to further corroborate the decapitation and SMS mutant results. A bioactive cytokinin, 6-benzylaminopurine (BA), was applied with or without NPA to buds of intact wild-type plants. Again, prior to slowing, buds treated with BA and NPA grew initially like BA-onlytreated plants (Fig. 1D). These data imply that auxin transport out of a bud is not required for their initial growth but instead may be important for sustained bud growth.

NPA Inhibits Auxin Transport out of Buds
We have shown that applying NPA directly to axillary buds at the earliest possible stage did not prevent early outgrowth caused by decapitation, ccd8 mutation, or BA application (Fig. 1). These buds are too small to assess their polar auxin transport. Therefore, to confirm that NPA affects auxin transport in buds and to compare its ability with that of GR24, we used larger (13.5 6 0.5 mm) growing buds that enabled [ 3 H]IAA transport measurements and that still exhibited a growth inhibition response to GR24 and NPA treatment (Fig. 2). NPA treatment of these growing buds caused a 57% reduction in growth 3 d after treatment, which was very similar to that of GR24treated buds of this size (Fig. 2C). We showed that while the growth of these older buds was inhibited by GR24 and NPA, only NPA affected the transport of [ 3 H]IAA supplied in the same solution (Fig. 2, A and C). Indeed, whereas NPA almost completely blocked the transport of [ 3 H]IAA in the bud, the profile of [ 3 H]IAA transport in the GR24-treated buds was very similar to that of control buds ( Fig. 2A). Importantly, the inhibition of transport by NPA demonstrated that Figure 1. Effects of GR24 and NPA on initial bud outgrowth in pea. A, The axillary bud at node 5 of decapitated wild-type plants grows out when treated with 0 mM GR24 (left plant) but is inhibited when treated with 2 mM GR24 at daily intervals for 4 d (right plant), while untreated buds grow out as normal at lower nodes of both plants. Photograph was taken 9 d after decapitation. Abbreviations: B, Bud inhibited at node 5; Br, branch growing at node 5; D, decapitated stump. B, Bud length at node 5 of wild-type (Torsdag) plants that were left untreated (intact) or decapitated above node 5 and treated with 0 or 2 mM GR24 or 1 mM NPA at daily intervals for 3 d from 13 d old. Data are means 6 SE (n = 13-14). C, Bud length at node 2 of rms1-1 (ccd8) plants treated at 9 d old with either 0 or 2 mM GR24 or 3.4 mM NPA. Data are means 6 SE (n = 15-16). At day 0, corresponding wild-type buds at node 2 were 0.97 6 0.07 mm in length. D, Bud length at node 5 of wild-type (Torsdag) plants that were treated at 12 d old with 0 or 1 mM NPA, 500 mM BA, or 1 mM NPA and 500 mM BA applied to the bud at node 5 or decapitated above node 5 with 0 mM NPA applied to the bud. Data are means 6 SE (n = 12). [See online article for color version of this figure.] the [ 3 H]IAA measured in the GR24-treated and control buds was transported in the polar auxin transport stream. These results support the premise that auxin transport is important for the growth of these larger buds but provide no evidence that GR24 affects auxin transport.

Auxin Addition to Buds Does Not Trigger Bud Release
Classical canalization experiments by Sachs (1969) demonstrated that increased concentrations of laterally applied IAA, relative to the concentration of IAA applied in the main stem, overcame the "inhibitory effect" of the main stem and allowed the connection of lateral vasculature to the vasculature of the main stem. This was demonstrated by applying IAA to the side of pea epicotyls and observing the ensuing formation of vascular connections. Sachs (1981) hypothesized that this inhibitory effect might also control bud growth. If the auxin transport capacity of the main stem is a limiting factor preventing outgrowth of axillary buds and the same principles are involved, then applying a relatively high dose of auxin to an axillary bud should induce its outgrowth. However, we could not induce any outgrowth in pea buds when we applied a rela-tively high concentration of IAA (239 mM) to the buds (control bud length, 1.17 6 0.09 mm, IAA-treated bud length, 1.18 6 0.06 mm, at 12 d after treatment [n = 15-16]).

Strigolactone Reduces Branching in Auxin Response Increased Branching Mutant Plants
Auxin has been shown to promote the expression of SMS synthesis genes (Sorefan et al., 2003;Bainbridge et al., 2005;Foo et al., 2005;Johnson et al., 2006;Arite et al., 2007). As a result, branching mutants defective in auxin response may have reduced strigolactone biosynthesis, which may be the cause of their increased branching phenotype. If this is the case, and auxin acts to repress bud outgrowth primarily by promoting strigolactone production, then exogenous strigolactone application should rescue the increased branching phenotype of this class of mutants. To test this, we applied GR24 to Arabidopsis axr1 mutant plants, which are defective in auxin response and exhibit an increased branching phenotype (Lincoln et al., 1990). GR24 significantly reduced branching in axr1 ( Fig. 3A; P , 0.0001 by Student's t test). Moreover, the 47% reduction in branch number was similar to that of comparable GR24-treated SMS-deficient plants (Fig. 3A).
To further investigate the relationship between auxin response and shoot branching, axr1 and tir1 afb1 afb2 afb3 quadruple auxin response mutant plants (Dharmasiri et al., 2005b) were grown in plant culture trays and their shoot-branching responses to GR24 supplied to the roots in the medium were measured (Fig. 3B). GR24 again significantly reduced branching in axr1 and also in tir1 afb1 afb2 afb3 mutant plants (P , 0.001 by Student's t test). Branching in the Arabidopsis branched1/teosinte branched1-like1 (brc1/tbl1) mutant, which is mutated in a gene encoding a TCP transcription factor that is thought to function downstream of auxin and SMS perception (Aguilar-Martínez et al., 2008;Finlayson, 2008), was not reduced by GR24 (Fig. 3B).
That GR24 was able to reduce branching in plants with defective response to auxin implies that one function of auxin is upstream of strigolactone in shoot-branching regulation. Since strigolactone can act when auxin response is defective, these results demonstrate that strigolactone action does not require auxin response to function and therefore might act downstream of auxin. These results also support the idea that auxin acts to inhibit bud outgrowth at least in part by promoting strigolactone production.

Intact SMS Mutant Shoots Do Not Show Enhanced Auxin Transport Capacity
Experiments by Sachs (1968Sachs ( , 1969 led to the idea that auxin saturation in the stem vasculature may act to directly block bud outgrowth. This hypothesis was recently expanded to explain the increased branching phenotype of Arabidopsis SMS mutants. These mutants are reported to have increased expression and localization of auxin transport carriers and increased auxin transported in segments of their inflorescences (Bennett et al., 2006). As a result, it was suggested that the increased branching phenotypes of these mutants were due to their auxin transport streams not being saturated (Bennett et al., 2006;. We tested this auxin transport hypothesis by applying increasing concentrations of unlabeled IAA, spiked with fixed amounts of [ 3 H]IAA, to the shoot apex of young intact pea plants or the inflorescence apex of young intact Arabidopsis plants. As shown previously for pea, this treatment led to a wave of IAA that moved down the stem at the speed of polar auxin transport (1 cm h 21 ; Fig. 4, A-E; Morris et al., 2005;Ferguson and Beveridge, 2009). NPA applied in a ring around the stem of pea plants prevented the movement of this wave, indicating that the IAA applied moved in the polar auxin transport stream (Fig. 5). Following the same application method, IAA transported in Arabidopsis inflorescences also moved in a distinct wave traveling at approximately 1 cm h 21 (Fig.  4, F and G).
Having established this auxin transport method for inflorescences of intact Arabidopsis plants, comparisons of auxin transport were made between the wild type and SMS synthesis branching mutants of both species (ccd8 of pea and ccd7 of Arabidopsis) for different auxin concentrations. For each species, the lowest concentration was chosen to only marginally increase the auxin supply to the stem. However, the fact that we were able to get any exogenous auxin into the polar auxin transport stream of wild-type plants indicated that it was not saturated and therefore not functioning at full capacity. When 0.14 to 0.6 mM (50-210 ng) IAA was applied, there was no difference in IAA transport between wild-type and ccd8 mutant pea plants, and the profiles were similar to those at the weakest concentration of applied IAA (0.023 mM; Fig.  4, A-C). Application of these concentrations provided an additional 25% to 200% of the total endogenous IAA content of about 0.41 ng across the whole segment Morris et al., 2005). At 2.9 mM (1,007 ng) applied IAA, ccd8 mutants showed a clear peak of auxin transport, similar to that observed using lower IAA concentrations. While the wild type showed a peak front at this same position, it had greater amounts of IAA than the ccd8 mutant closer to the shoot tip. This enhancement behind the peak front, which presumably also occurred to some extent in the mutant, may have been due somewhat to IAA diffusion. However, the total amount of IAA taken up and transported at this concentration was also greater in the wild type than in the ccd8 mutant (Fig. 4J), which is not easily explained by diffusion. Similarly, when 14 mM (5,000 ng) IAA was applied, the ccd8 mutant was not able to transport more IAA than the wild type and the peak front still occurred at the same position as for other concentrations (Fig. 4, A-E and J). In this case, the amount of IAA taken up and transported in the peak front was about seven to eight times that of the endogenous IAA across the whole segment. That the position of the peak front was not different across a wide range of IAA concentrations suggests that the majority of this IAA was likely in the polar auxin transport stream. These data suggest that pea stems are not normally saturated and instead have the capacity to rapidly respond to, and transport, additional quantities of IAA. Moreover, the pea ccd8 mutant does not appear to have enhanced auxin transport properties.
Using a range of lower IAA concentrations suited to the small Arabidopsis inflorescences, similar findings were detected for Arabidopsis as for pea in that wildtype and ccd7 inflorescence shoots could load and transport exogenously applied IAA. Moreover, when 23 to 250 mM (4-44 ng) IAA was applied, little difference in IAA transport was observed between ccd7 mutant and wild-type shoots (Fig. 4, G-I). At the lowest IAA concentration (2.3 mM), we repeated previously reported findings with isolated segments that SMS mutant inflorescences can load and transport more IAA than the wild type (Fig. 4, F and I). However, the findings at higher concentrations of applied IAA clearly demonstrate that the lack of branching exhibited by wild-type pea and Arabidopsis plants is not directly related to saturation of their auxin transport streams, as both species can take up and transport additional IAA (Fig. 4).
Interestingly, for both species, the total amount of IAA transported relative to the amount applied steadily decreased with increasing concentration (Fig. 4, I and J). For example, when 0.023 mM IAA was applied to wild-type pea apices, 12% was transported, compared with only 1.4% transported when 14 mM was applied. Comparable percentage reductions were observed here using intact Arabidopsis inflorescences, and similar findings have been reported previously using pea cuttings (Baadsmand and Andersen, 1984). This implies that the main mechanism for loading and/or transporting IAA is not diffusion, because the higher concentrations should not have reduced the percentage uptake. Indeed, the lack of a clear change in the rate of transport of the peak front at different concentrations of applied IAA implies that diffusion is not an important variable for at least this part of the auxin wave (for review, see Kramer, 2008).
Independent experiments showed that IAA application to the stem of intact plants had no effect on the growth of buds below. Whereas the IAA (3 mg g 21 ) applied in a lanolin ring to an expanding internode of ccd8 pea plants was absorbed and stimulated elongation of the internode (data not shown), the length of the bud below was unaffected by the auxin treatment. The mean bud lengths were 5.2 6 2.1 cm and 6.2 6 1.4 cm in control and auxin-treated plants, respectively (17 d after treatment [n = 9]; plants in the Parvus background had six leaves expanded at the time of treatment). This supports the idea that additional auxin transported in the main stem does not prevent bud outgrowth in the absence of strigolactone production in mutant plants.

DISCUSSION
The current auxin transport hypothesis promotes the idea that SMS acts upstream of auxin by regulating PIN-dependent auxin transport in the stem (Bennett et al., 2006). In contrast, strigolactone completely inhibited decapitation-induced bud outgrowth in pea (Fig. 1, A and B), supporting the idea that strigolactone actually functions downstream of auxin in the main stem in its regulation of bud outgrowth in decapitated plants (Fig. 6). This nicely fits the classical second messenger theory of apical dominance (Snow, 1929(Snow, , 1937 and is consistent with our previous models of shoot branching (Beveridge, 2000;Ferguson and Beveridge, 2009). Indeed, this apical dominance experiment showed that any changes in auxin level, movement, or signaling caused by removal of the main auxin supply in the shoot in no way prevented the inhibition of bud outgrowth by strigolactone application to buds. Any action of strigolactone itself in buds, therefore, seems to be independent of auxin content in the stem.
Previous studies have shown that auxin positively regulates SMS synthesis gene expression (Sorefan et al., 2003;Bainbridge et al., 2005;Foo et al., 2005;Johnson et al., 2006;Arite et al., 2007), also suggesting that strigolactones could function downstream of auxin to inhibit bud outgrowth. Auxin regulation of the expression of the SMS synthesis gene CCD8 was demonstrated to be AXR1 dependent in Arabidopsis (Bainbridge et al., 2005). Direct application of GR24 to the buds of the Arabidopsis auxin response mutant axr1 reduced its increased branching phenotype by a similar magnitude as SMS-deficient plants (Fig. 3). Growth of axr1 and tir1 afb1 afb2 afb3 auxin response mutant plants in medium containing GR24 also led to an inhibition of shoot bud outgrowth, whereas the tbl1/brc1 mutant, which is thought to act downstream of strigolactone response (Aguilar-Martínez et al., 2008;Finlayson, 2008), showed no response to strigolactone treatment, as expected (Fig. 3B). This implies that the auxin response mutants may branch, at least in part, due to endogenous strigolactone depletion and that auxin response is not necessarily required for strigolactone to inhibit shoot branching. The reduced branching in GR24-treated tir1 afb1 afb2 afb3 quadruple mutant plants suggests that AXR1 regulation of strigolactone biosynthesis might be mediated by the SCF TIR1/AFB ubiquitin ligase complex, which functions as an auxin receptor and targets proteins for ubiquitination (Dharmasiri et al., 2005a(Dharmasiri et al., , 2005bKepinski and Leyser, 2005). Therefore, we propose that one role for auxin in mediating apical dominance is through auxin inducing the expression of strigolactone biosynthesis genes (Fig. 6). This could occur in vascular cambial  . Pathway for auxin and strigolactone action in regulating axillary bud growth. Arrows represent promotion, while flatended lines represent inhibition. Auxin promotes strigolactone biosynthesis gene expression. The MAX2/RMS4 F-box protein is required for strigolactone inhibition of bud release. Auxin transport out of an axillary bud, which can be inhibited by NPA, is then required for a bud that has been released to proceed to sustained bud growth. cells, through which auxin is transported down the stem (Morris and Thomas, 1978). AXR1 and strigolactone biosynthesis genes are indeed known to be expressed in those cells (Booker et al., , 2005Sorefan et al., 2003). Whereas AXR1 may have additional targets, TIR1 is thought to more directly affect auxin responses. Once technologies are available, strigolactone content should be measured in axr1 and tir1 afb1 afb2 afb3 shoots.
It is unlikely that the increased branching phenotype of axr1 and possibly tir1 afb1 afb2 afb3 quadruple mutant plants is due entirely to strigolactone depletion. It is likely that auxin response mutants have increased cytokinin content due to reduced auxin regulation of cytokinin biosynthesis (Nordströ m et al., 2004;Bennett et al., 2006). This may contribute to the additive branching phenotype of axr1 and SMS double mutant plants and, possibly in combination with impaired feedback signaling , to the poor response of axr1 shoots to grafting with wildtype rootstocks. Auxin signaling might also be involved in other aspects of bud outgrowth regulation, particularly the continued elongation of the growing bud once it has been released to grow. Nevertheless, whatever the action of auxin and strigolactone, strigolactone inhibition of branching in auxin response mutants suggests that strigolactone function is at least partly downstream of auxin and auxin response, consistent with the second messenger role for strigolactone. Sachs (1969) demonstrated that a high concentration of IAA in the stem repelled the formation of new vascular connections from lateral sources. It was proposed that auxin is at a saturation point in the stem and, therefore, that there would be no more room in the transport stream for lateral auxin to enter (Sachs, 1981). However, when a greater concentration of IAA was applied to a lateral point, the repulsion by the main stem could be overcome (Sachs, 1969). Indeed, consistent with canalization theory, auxin was shown to promote its own efflux by enhancing PIN activity (Paciorek et al., 2005). However, direct application of IAA did not induce bud outgrowth in wild-type pea. This supports an earlier experiment where IAA applied to pea buds after decapitation actually blocked outgrowth (Thimann, 1937). Note that the amount we chose for this experiment is equivalent to a mid-range application, in order to provide an amount significantly higher than endogenous levels (Fig. 4J).
Based on differences between PIN protein abundance in the wild type and SMS-deficient mutants, Bennett et al. (2006) suggested that reduced PIN abundance in the main stem leads to saturation of the auxin transport stream and prevents bud outgrowth in wild-type plants, whereas the main stem of SMS mutants is not saturated by auxin. Although increased amounts of IAA were transported in intact ccd7 mutant shoots of Arabidopsis compared with the wild type at very low concentrations of applied IAA, this did not occur in ccd8 mutants of pea compared with the wild type and was not observed for the highconcentration applications in either species (Fig. 4). Moreover, we found that in pea and Arabidopsis, wild-type and SMS mutant plants could instantly transport additional exogenous IAA in the main stem and SMS mutants could not take up and transport excess IAA above the level of the wild type when high IAA concentrations were applied (Fig. 4). Together with observing the appropriate rate for polar auxin transport (1 cm h 21 ) at different concentrations, we used localized NPA treatment below the shoot tip of pea to show that the applied auxin was indeed moving in the polar auxin transport stream (Fig. 5). Altogether, this demonstrates that the auxin transport stream in the main stem of wild-type pea and Arabidopsis plants is not functioning at saturation and that SMS synthesis mutants do not have increased capacity to transport higher levels of IAA. Therefore, it is unlikely that overcoming saturation of the main stem auxin transport stream is involved in triggering bud outgrowth.
Arabidopsis and pea have the ability to transport excess IAA, which, for example, may act as a buffer during circadian fluctuations in auxin levels (Covington and Harmer, 2007) or when new primordia or flower buds produce and transport new auxin into the polar auxin transport stream (Benková et al., 2003). Moreover, although a small difference in IAA transport was observed between wild-type and ccd7 Arabidopsis inflorescences when low concentrations of IAA were applied (Fig. 4, F and I), it is unlikely that this is a primary cause for branching in plants, because such a difference is not always observed in pea (Figs. 4, A-E and J, and 5;Beveridge et al., 2000). Multiple repeat experiments found no difference between wild-type and SMS mutant pea plants in the transport of low concentrations of IAA (data not shown). One would expect the primary mechanism of strigolactone function to be conserved across plant species, especially due to high conservation of the biosynthetic pathway and its regulation by auxin ( Fig. 6; Sorefan et al., 2003;Bainbridge et al., 2005;Foo et al., 2005;Zou et al., 2006;Arite et al., 2007;Gomez-Roldan et al., 2008;Umehara et al., 2008). Interestingly, bud vascular traces were found to be repelled from leaf vasculature in the Arabidopsis ccd8 mutant background . This may indicate that the production of auxin in young leaves is increased in the SMS mutants, but it does not distinguish whether this is a cause, or feedback consequence, of the branching phenotype. It is likely that higher endogenous auxin levels, caused by feedback regulation of auxin content in a failed attempt to synthesize more SMS to inhibit branching, has led to these vasculature differences and an increase in PIN abundance in ccd mutants of Arabidopsis .
If auxin transport out of a bud is critical for triggering bud release, then an auxin transport inhibitor like NPA should completely block bud outgrowth from the earliest stage. When we applied NPA to buds of an SMS synthesis mutant of pea, ccd8, we could not phenocopy the early bud repression shown by GR24 (Fig. 1C). While GR24 application prevented bud outgrowth in pea ccd8 mutants, NPA application allowed early bud growth to occur, only inhibiting the sustained bud growth that occurred in the control ccd8 plants (Fig. 1C). This was also the case for buds of decapitated wild-type pea plants (Fig. 1B). In addition, NPA was unable to inhibit early bud outgrowth induced by the cytokinin, BA (Fig. 1D). In contrast, NPA seemed to slow bud growth only after several days (Fig. 1, B-D). The early outgrowth seen in NPA-treated buds (Fig. 1, B-D) was not due to swelling induced by auxin accumulation, because intact wild-type NPAtreated buds did not grow at all (Fig. 1D). In addition, the later inhibitory effect on continued bud growth was not due to NPA affecting strigolactone via the loss of auxin in the stem, as ccd8 mutants are in any case unable to produce strigolactone and decapitated plants have had their major auxin supply removed. The different outgrowth responses of buds treated with either NPA or GR24 are also not due to GR24 being more efficient at inhibiting auxin transport, as GR24 treated together with [ 3 H]IAA had no effect on the transport of [ 3 H]IAA in a growing bud, yet NPA did (Fig. 2). At this later developmental stage, both treatments had a similar effect on bud outgrowth, reducing it by about half (Fig. 2). By assessing [ 3 H]IAA transport from the main shoot tip, we have observed that NPA remains effective at blocking [ 3 H]IAA transport for at least 7 d (data not shown), indicating that the initial NPA treatment applied to tiny buds could potentially have remained active after several days. Apart from other possible side effects of NPA, these data suggest that normal auxin transport may only be required for ongoing bud growth rather than being the initial trigger of bud release ( Fig. 6; Morris et al., 2005;Dun et al., 2006) and that strigolactone and NPA act quite differently. This also provides an explanation for strigolactone acting downstream of auxin, even though NPA reduces branch lengths ( Fig. 1; Bennett et al., 2006).
Early stages of bud outgrowth stimulated by reduced strigolactone signaling would lead to enhanced primordium development and growth of the bud, increasing auxin supply and export to stems. Whatever the precise action of strigolactone in preventing the initial bud release, an outcome is that auxin transport from inhibited buds will be reduced compared with growing buds. Once growing, buds synthesize auxin (Gocal et al., 1991) and its export may enhance vascular connections and nutrient flow to further stimulate the growing bud. Consequently, NPA suppresses continued bud growth (Bennett et al., 2006) but does not suppress the earliest bud growth. An outcome of the failure to inhibit bud release under strigolactone deficiency in SMS mutants or decapitated plants that have greatly suppressed CCD7 and CCD8 gene expression is that stem auxin level can increase along with enhanced auxin transport. This hypothesis is consistent with previous findings that cytokinin triggered bud release while auxin promoted the subsequent elongation of buds (Sachs and Thimann, 1967). What remains to be seen is how strigolactones might interact with cytokinin to regulate shoot branching, especially since auxin and strigolactone deficiency is not always sufficient to promote bud outgrowth unless cytokinin biosynthesis genes are activated (Ferguson and Beveridge, 2009).

CONCLUSION
Our findings support previous ideas that auxin induces the production of a second messenger to regulate bud outgrowth. We propose that strigolactone could act as a second messenger for auxin action and that this messenger directly represses bud outgrowth (Fig. 6). Auxin export from buds, however, seems to be critical for ongoing bud growth, rather than as the initial trigger (Fig. 6). In this case, it seems that auxin has two actions: one is the involvement of auxin levels in regulating bud release, while the other is via auxin transport being necessary for sustained bud growth.

Plant Material, Growth Conditions, and Treatments
For all garden pea (Pisum sativum) experiments, plant growth conditions were as described by Ferguson and Beveridge (2009), except that plants were sometimes grown in 9:1 composted fine slash:medium river sand potting mix (Bassett Barks). Nodes were numbered acropetally from the first scale leaf, and lengths of lateral buds and branches were recorded using digital calipers. For Arabidopsis (Arabidopsis thaliana) experiments, unless otherwise stated, plants were grown as reported by Gomez-Roldan et al. (2008).
For direct applications in Arabidopsis, solutions were given in 0.1% Tween 20. For Arabidopsis root treatments, plants were germinated and grown in plant culture trays (Phytatray II; Sigma) containing 0 or 5.8 mM GR24 in standard Arabidopsis Murashige and Skoog growth medium at 24°C in a growth chamber (Conviron) with fluorescent lighting and 18-h daylength. The number of rosette branches longer than 5 mm was counted when the plants were 63 d old.

IAA Transport in the Main Stem and Overloading of the Polar Auxin Transport Stream
IAA transport in the main stem and the overloading of the polar auxin transport stream were analyzed using methods similar to those outlined by Beveridge et al. (2000) and Morris et al. (2005). For pea experiments, various concentrations of IAA, each containing 34 kBq [ 3 H]IAA (American Radiolabeled Chemicals; specific activity, 20 Ci mmol 21 ), were dissolved in 50% ethanol. Total IAA concentrations of the solutions were 0.023 mM (8.04 ng), 0.14 mM (49.74 ng), 0.6 mM (209.04 ng), 2.9 mM (1,007.04 ng), and 14 mM (4,998.04 ng). Two microliters of these solutions was applied to the apical bud of 19-d-old wild-type (cv Parvus [L77]) or rms1-1 (ccd8) mutant (WL5237) plants having seven leaves fully expanded. The radiolabel was taken up and transported over a 4-h period. Following this, the internode tissue beginning directly below the apical region was harvested from individual plants and divided into 3-mm equal-length sections.
The same method was used to analyze IAA transport in the main inflorescence stem of Arabidopsis wild-type and max3-11 (ccd7) plants (Co-lumbia ecotype), except that the total IAA concentrations of the solutions were 2.3 mM (0.4 ng; including 1.85 kBq [ 3 H]IAA mL 21 ), 23 mM (4 ng), and 250 mM (44 ng; including 18.5 kBq [ 3 H]IAA mL 21 ), only 1 mL of the solution was applied to apices of 4-week-old, newly bolted plants, and the radiolabel was transported over a 2.5-h period.
Radioactivity was extracted directly from the segments of individual samples in 2 mL of Ultima Gold liquid scintillant (Perkin-Elmer Life and Analytical Sciences) gently shaken overnight as outlined by Morris et al. (2005). Radioactivity was analyzed using a Packard Tricarb 1600 TR Liquid Scintillation Analyzer (Packard Instruments) and recorded as dpm. For pea, dpm was converted to ng of IAA based on 1 dpm being equivalent to 3.95 fg of [ 3 H]IAA. For Arabidopsis, dpm was converted to ng of IAA based on the dpm readings from samples of known ng of [ 3 H]IAA applied to each plant. Total IAA transported was calculated from segments 0.6 to 4.5 cm (Arabidopsis) and 0.6 to 9.0 cm (pea) from the shoot apex based on the assumption that the same percentages of [ 3 H]IAA and IAA were transported.

IAA Transport in NPA-Treated Plants
IAA transport in the main stem of NPA-treated plants was measured as in the overloading experiment, except that a ring of lanolin containing either 0 or 1 mg g 21 NPA with 4 mL g 21 DMSO and 100 mL g 21 ethanol was applied around the uppermost expanding internode, 10 mm below the oldest unexpanded leaf, of 17-d-old wild-type (Parvus [L77]) and rms1-1 (ccd8) mutant (WL5237) pea plants. Two hours after treatments were applied in lanolin, 17 kBq [ 3 H]IAA was applied to the apical bud and the resulting transport was measured as described, except using 5-mm stem segments.
IAA Transport out of Axillary Bud Apices rms1-1 (WL5237) plants were grown for 9 d before axillary buds were removed from nodes 1 and 2 to encourage the growth of the axillary buds at upper nodes. When the plants were 20 d old, plants with a growing bud at node 4 that was 10 to 20 mm in length were selected. These buds had basal internodes of 8 to 17 mm in length. Two microliters of solution containing 14.8 kBq (3.5 ng) of [ 3 H]IAA, 1% acetone, 0.15% DMSO, 50% ethanol, and either 0 or 10 mM GR24 or 1 mM NPA was applied to the bud inside the stipules of the first two nodes of the unexpanded leaves. Plants were also treated, but without [ 3 H]IAA, for bud outgrowth measurements. [ 3 H]IAA transport was measured as described above, except the treatments were left for 1 h, after which the bud internode was harvested into 1.57-mm segments.