Introduction

Elongating plant axial organs, such as hypocotyls, epicotyls, coleoptiles, tendrils, and roots, exhibit a revolving movement and their tips draw a circular or elliptic orbit. This growth movement was first described in the nineteenth century, and Darwin and Darwin (1880) termed it circumnutation. It has long attracted many plant researchers, who have taken multiple approaches in their studies using various plant species including oat, rice, pea, French bean, Arabidopsis thaliana, and sunflower (Stolarz 2009). Since Israelsson and Johnsson (1967) proposed that gravity induces oscillatory movements in plant organs, the relation between gravity and circumnutation has been extensively investigated using clinostat or space flight experiments (Brown 1993; Brown and Chapman 1984; Chapman and Brown 1979). It was recently verified genetically that gravity sensing is essential for circumnutation. Inflorescence stems of the scr mutant of A. thaliana, which cannot sense gravity due to the defect of stem endodermal cells containing amyloplasts (Fukaki et al. 1998), did not exhibit circumnutation (Kitazawa et al. 2005). This abnormal phenotype was restored by introducing the morning glory SCR into the mutant concomitantly with the recovery of endodermis tissue. However, the mechanism and significance of circumnutation are not yet completely understood.

Auxin, one of the major phytohormones involved in the regulation of plant growth and development, is known to promote cell elongation through two pathways (Badescu and Napier 2006; Velasquez et al. 2016). The faster pathway is by activation of the plasma membrane H+-ATPase, which leads to acid growth (Hager 2003; Takahashi et al. 2012), and the slower one is through the regulation of gene expressions (Chapman and Estelle 2009). Auxin is mainly produced in the shoot apex and young leaves (Ljung et al. 2002) from where it flows down to the basal regions via the phloem and/or by cell-to-cell transport mediated by carrier proteins (Michniewicz et al. 2007). In both gravitropism and phototropism, after seedlings sense gravity or light stimulus, auxin is asymmetrically distributed in the organs, leading to the differential growth of epidermal cells in the opposing sides (Liscum et al. 2014; Morita 2010). Auxin is also supposed to play important roles in circumnutation. Britz and Galston (1982a) reported that α-naphthylphthalamic acid (NPA), an established inhibitor for polar auxin transport (Murphy et al. 2002), inhibited the nutation of pea epicotyls. NPA also blocked circumnutation of morning glory shoots (Hatakeda et al. 2003) and of pea roots (Kim et al. 2016). Britz and Galston (1982b) further demonstrated that de-capped pea epicotyls did not exhibit nutation, but resumed nutation after the exogenous application of indole-3-acetic acid (IAA), an endogenous auxin. Finally, inflorescence stems of the auxin-resistant axr-2 mutant of A. thaliana exhibited little circumnutation (Hatakeda et al. 2003). However, the distribution pattern of auxin as well as its mode of action associated with circumnutation has not yet been clarified.

Plant hormones are known to affect the direction of elongation of axial organs by altering the orientation of cortical microtubules in the epidermis tissue (Fischer and Schopfer 1997; Shibaoka 1994). Cortical microtubules regulate the orientation of cellulose microfibrils deposited onto the innermost surface of the cell wall, and determine the direction of plant cell elongation (Lloyd 2011). Given that the epidermis tissue limits the growth pattern of axial organs (Kutschera and Niklas 2007), circumnutation may also be driven by the specific mode of epidermal cell elongation. Cortical microtubules can play some role in circumnutation by regulating the mode of epidermal cell elongation. On the other hand, accumulating knowledge demonstrates that the orientation of cortical microtubules can be altered in response to the mechanical stress generated in multi-cellular tissues (Landrein and Hamant 2013). It has even been proposed that mechanical stress can drive circumnutation (Baskin 2007). In the epidermal cells of adzuki bean [Vigna angularis (Willd.) Ohwi & H.Ohashi] epicotyls, cortical microtubules exhibit dynamic reorientation in a cyclic manner (Mayumi and Shibaoka 1996; Takesue and Shibaoka 1998). Moreover, auxin plays a primary role in driving the cycle, whereas its relation to circumnutation has not been investigated.

In this study, by combining careful handling of intact plants and liquid chromatography–mass spectrometry (LC–MS), we succeeded in quantifying endogenous hormones in adzuki bean epicotyls exhibiting circumnutation. We found, for the first time, that IAA and gibberellin A1 (GA1) were asymmetrically distributed in the bending zone of circumnutation. The epicotyl region where the asymmetric distributions of hormones were detected exhibited the highest responsiveness to these hormones in terms of elongation enhancement. Thus, we propose that the differential growth of epidermal cells induced by the asymmetrically distributed hormones drives circumnutation.

Materials and methods

Plant materials

Adzuki bean (V. angularis cv. Erimowase) seeds stored at 4 °C were transferred to a growth chamber (LH-60FL12-DT; Nippon Medical and Chemical Instruments Co., Ltd., Osaka, Japan) for 1 h to adapt to 27 ± 1 °C. Thereafter, the seeds were sown in 200 ml pots filled with vermiculite and warm water. The pots were covered with plastic wraps to maintain high humidity and exposed to continuous white light (30 µmol m−2 s−1, FL10D; NEC Co., Tokyo, Japan) from above. Three days later, the plastic wraps were removed so as not to interfere with epicotyl growth.

Recording and analysis of epicotyl circumnutation and elongation

Vinyl tape 1 mm in width was carefully pasted on each 5-day-old greened epicotyl every 3 mm from the tip (Fig. 1a). Thereafter, epicotyls were put in a clear, colorless acrylic box maintained at 27 °C with an electric panel heater (MP-916; TRiO Co., Osaka, Japan) under continuous white light (30 µmol m−2 s−1). Epicotyls were photographed every 10 min in the plane parallel (x view) or perpendicular (y view) to the first leaves (Fig. 1a, c), by programmable cameras (Ltl-5210B; Little Acorn Outdoors, Green Bay, WI, USA). Circumnutation and epicotyl elongation were analyzed from sequentially recorded images using the Tracker Video Analysis and Modeling Tool (http://www.cabrillo.edu/~dbrown/tracker/).

Fig. 1
figure 1

Greened adzuki bean epicotyls used for growth movement analyses. Photographs of 5-day-old greened epicotyls were taken in the plane parallel (a; x view) or perpendicular (c; y view) to the first leaves. Four pieces of tapes 1 mm in width were pasted onto each epicotyl every 3 mm from the tip to trace the elongation pattern. Regions between the tip and the top piece, between each piece, and between the bottom piece and the base, respectively, were defined sequentially as regions I–V. b Scheme to determine the angle of curvature of circumnutation and the bending zone on the epicotyl. The angle of curvature of circumnutation was defined as the angle (θ) between the longitudinal center line of the basal straight region of the epicotyl (solid line) and that of the apical straight region (broken line). The bending zone was defined as the region between two positions where the longitudinal center line of the epicotyl diverges from that of the basal straight region (open circle) and where from that of the apical straight region (open diamond). Bar 1 cm

Full size image

The angles between the apical region and the basal region of the epicotyl were independently measured in the x and y views (Fig. 1b) and processed with fast Fourier transform to obtain the periods and amplitudes. Since the number of samples processed with fast Fourier transform is restricted to 2x, we determined to process every 25 samples, at an interval of 10 min, for 320 min (5.3 h) in total, to keep balance between the time-resolution and the precision of the analysis. The integrated periods (P) and amplitudes (A) were obtained from the periods in the x view (Px) and the y view (Py), and the amplitudes in the x view (Ax) and the y view (Ay) by the following equations, respectively.

$$P=\frac{{{P_x}+{P_y}}}{2}$$
$$A=\sqrt {A_{x}^{2}+A_{y}^{2}}$$

Scanning electron microscopy

Sections 1 cm in length were taken from 6-day-old greened epicotyls and buried in silicone resin (Extrude Wash; Kerr, Romulus, MI, USA). After the resin had dried up, the sections were removed and the silicone molds were filled with epoxy resin (DEV-TUBE S-208; ITW, Inc., Glenview, IL, USA) to make replicas. Prepared replicas were coated with gold and observed with a scanning electron microscope (SU6600; Hitachi, Ltd., Tokyo, Japan).

Spray treatments of intact epicotyls

After the 5-day-old greened adzuki bean epicotyls had been observed for 24 h to confirm that they exhibited normal circumnutation, they were gently sprayed with test solutions containing auxin transport inhibitors or hormones from four directions at right angles to each other. Tween 20 was added to the test solutions at 0.1% to reduce the surface tension of the solutions and facilitate the efficient spread of the applied solution drops onto the surface of the epicotyls.

Quantification of endogenous phytohormones

Sections 0–1, 1–2, and 2–3 cm below the tip of the 6-day-old greened epicotyls were carefully separated along the longitudinal axis into either convex/concave or pre-convex/pre-concave pairs with respect to the plane of curvature of circumnutation using a razor blade (Fig. 4a). They were then frozen in liquid nitrogen. The contents of five different plant hormones [GA1; IAA; abscisic acid (ABA); salicylic acid (SA); jasmonic acid (JA)] were simultaneously determined by LC–MS, essentially according to Tsukahara et al. (2015). In brief, approximately 0.1 g (fresh weight) of dissected tissues was ground in liquid nitrogen, followed by extraction in 80% acetonitrile containing 1% acetic acid at 4 °C for 1 h. Internal standards of plant hormones were added to the extraction buffer at this step. The composition of internal standards are described elsewhere (Tsukahara et al. 2015). The hormones were fractionated into three parts by a solid-phase extraction comprised of a reverse-phase (Oasis HLB, Waters Corporation, Milford, Massachusetts, USA), cation-exchange (Oasis MCX, Waters Co.), and anion-exchange cartridges (Oasis WAX, Waters Co.), and each fraction was analyzed by triple-quadrupole LC–MS (Agilent 6410, Agilent Technologies, Santa Clara, California, USA) according to Tsukahara et al. (2015).

Effects of phytohormones on the elongation of epicotyl sections

Sections 0–1, 1–2, and 2–3 cm below the tip were taken from the 6-day-old greened epicotyls and were floated in the control solution (10 mM KCl, 2% sucrose, and 2 mM PIPES, pH 7.0) for 1 h to release the endogenous hormones. The sections were then treated with different hormone combinations dissolved in the control solution. The length of each section was recorded using a stereomicroscope (Stemi DV4; Carl Zeiss, AG, Oberkochen, Germany) and a digital camera (GR DIGITAL; Ricoh, Co., Ltd., Tokyo, Japan), and analyzed using ImageJ software (https://imagej.nih.gov/ij/).

Visualization of cortical microtubules

Cortical microtubules adjacent to the outer tangential wall of epidermal cells of epicotyls were visualized with indirect immunofluorescence microscopy according to the methods described in Sakiyama and Shibaoka (1990) after minor modifications. A mouse monoclonal antibody against α-tubulin (Amersham Japan, Tokyo, Japan) and Alexa488-conjugated goat anti-mouse IgG (Molecular probes, California, USA) were used as the primary and secondary antibodies, respectively. Specimens were observed with a light microscope (BX50; Olympus, Co., Tokyo, Japan) and fluorescence images were taken using a charge coupled device camera (Coolpix; Nikon, Co., Tokyo, Japan). The angle of cortical microtubules to the longitudinal axis of the cells was measured using ImageJ software (Fig. 7a) and separated into four classes: longitudinal (0–30°), oblique (30–60°), transverse (60–90°), and random (no dominant alignment of microtubules).

Results

Circumnutation and elongation pattern of greened adzuki bean epicotyls

First, the movement of 5-day-old adzuki bean epicotyls grown under continuous white light (Fig. 1, Movie S1) was recorded every 10 min by time-lapse imaging for 48 h (Fig. 2a). The angle between the apical region and the basal region of the epicotyl continued to oscillate during the observation in both x and y views (see Fig. 1 for the definition). We processed these results with fast Fourier transform (Fig. S1a–i). The periods of detected oscillations were maintained more or less constant at around 60 min during the observation (Fig. 2b), while the amplitudes increased from 10° to 25° in the x view and to 35° in the y view in the first 15 h, and then gradually decreased thereafter (Fig. 2c). The trajectory projections of the epicotyl tip on a horizontal plane were reconstructed from two-view measurements of the tip orientation angle (Fig. S1a′–i′). The tip of the greened epicotyl moved in a slightly irregular manner from 0 to 10 h, drawing circles with unfixed orbits. Finally, the tip turned to move in a circular orbit. All of the observed epicotyls eventually exhibited a clockwise movement in the top view, while 40% of the epicotyls exhibited a counterclockwise movement during the early irregular movement. The trajectory of the tip was an ellipse rather than a perfect circle (Fig. 2d), and approximately 20% of the epicotyls exhibited a pendulum-like movement (Fig. 2e). Indeed, the amplitude of oscillation in the y view was significantly larger than that in the x view (Fig. 2c). The contour of adzuki bean epicotyl is like a misshapen hexagon with two vertexes on the X-axis, the center line of the epicotyl cross section parallel to the x view, and two concave edges parallel to the X-axis (Fig. S2). These structural characteristics might provide a preference in the oscillation direction to the epicotyls.

Fig. 2
figure 2

Circumnutation of greened adzuki bean epicotyls under white light. a Changes in the angle of circumnutation in a representative 5-day-old epicotyl during a 48 h observation period. The solid line shows θ in x view and the broken line shows θ in y view. b, c Parameters of circumnutation detected with fast Fourier transform. The mean period (b) and amplitude (c) were determined every 5.3 h in the x view (open circles) and in the y view (closed squares), respectively (Px, Py, Ax, and Ay in “Materials and Methods”). The vertical bar on each point shows SE (n = 5). d, e Horizontal projections of movements of epicotyl tips at 10 min intervals. The tip exhibited elliptical (d) or pendulum-like movements (e). In each panel, an open and a closed triangle indicate the starting and final positions of the epicotyl tip, respectively. f The lengths of greened epicotyls under white light were measured for 48 h. See Fig. 1 for definition of the regions I–V. Open circles, open diamonds, and gray triangles show the positions of the basal end of the bending zone, of the apical end of the bending zone, and where the elongation of a defined region was completed, respectively. The vertical bar on each point shows SE (n = 5)

Full size image

To clarify which epicotyl region was responsible for the curvature of circumnutation, we analyzed the elongation pattern of epicotyls (Fig. 2f). The mean length of the 5-day-old greened epicotyls was approximately 6.5 cm. Over the next 48 h, epicotyls under white light elongated at a constant rate of approximately 0.16 cm h−1. Elongating epicotyls could be separated into two regions: the more apical region, in which epicotyls continued to elongate (elongating region), and the more basal region, in which epicotyls had elongated completely (completed region). The elongation of epicotyls completed from the most basal region, and in regions V, IV, and III (see Fig. 1a for the definition), sequentially, at around 27, 36, and 42 h of observation (gray triangles in Fig. 2f). By tracing the bending zone of circumnutation on each epicotyl (see Fig. 1b for definition), we found that the bending zone was always located in the region 1–3 cm below the tip (regions intercepted by the open diamond and open circle in Fig. 2f). Markedly, its basal end (open circles in Fig. 2f) was always almost identical to the apical end of the completed region (gray triangles in Fig. 2f), and accordingly, moved up toward the tip along the longitudinal axis of the epicotyl. We also found that the epicotyls did not exhibit any twisting movement since the orientation of the marking tape did not change during the observation. Consistent with this, the epidermal cell files run almost parallel to the longitudinal axis of the epicotyl (Fig. S3).

Effect of auxin transport inhibitors on the circumnutation of adzuki bean epicotyls

Auxin transport inhibitors or de-capping of organs substantially disturbed circumnutation in several plant species (Britz and Galston 1982a, b; Hatakeda et al. 2003; Kim et al. 2016). To know whether auxin also plays a role in the circumnutation of adzuki bean epicotyls, we examined the effect of auxin transport inhibitors. After we confirmed that the epicotyls exhibited normal circumnutation (after the amplitude reached the peak value at around 15–20 h of observation with a constant period of 60 min), those epicotyls were sprayed with test solutions at 24 h (Fig. 2c). While a mock treatment with 0.1% Tween 20 alone did not affect the period of circumnutation, 0.1% Tween 20 plus 20 µM 2,3,5-triiodebenzoic acid (TIBA) persistently increased the period from 60 to 140 min in 18 h treatment (Fig. 3a). The movement of the epicotyl tip was slightly disordered in 8 h of the mock treatment, but it recovered to an elliptical trajectory in 12 h of the mock treatment (Fig. S4e′–i′). In contrast, after treatment with TIBA, the epicotyl tip no longer drew a circular trajectory (Fig. S5e′–i′).

Fig. 3
figure 3

Effect of 2,3,5-triiodebenzoic acid (TIBA) on the period and amplitude of circumnutation of greened adzuki bean epicotyls. Greened epicotyls exhibiting circumnutation were treated either with 0.1% Tween 20 alone (mock) or together with 20 µM TIBA (TIBA) at 24 h of observation (indicated by arrows). The period (a) and amplitude (b) of circumnutation were determined every 5.3 h by integrating the values calculated in the x view and the y view (see “Materials and methods”). The asterisks and daggers indicate significant differences between “mock” and “TIBA” and between “no treat” and “mock” or “TIBA,” respectively, by a Tukey’s test (P < 0.05). The vertical bar on each point shows SE (n = 5)

Full size image

The mock treatment decreased the amplitude of circumnutation at a descending rate of 4.3 ± 0.6° h−1 in the first 2 h of treatment (Fig. 3b, mock), which was significantly higher than that of non-treated epicotyls (0.1 ± 1.0° h−1) during the same observation period (P < 0.05; Tukey’s test). The effect of the mock treatment was transient, and the descending rate of amplitude over the next 2–12 h was 0.5 ± 0.4° h−1, indicating recovery of circumnutation. The TIBA treatment also decreased the amplitude of circumnutation at a descending rate of 4.8 ± 0.9° h−1 in the first 2 h of treatment (Fig. 3b, TIBA), similar to that of the mock-treated epicotyls (P > 0.05; Tukey’s test) and significantly higher than that of the non-treated epicotyls during the same observation period (P < 0.05; Tukey’s test). However, the descending rate of amplitude over the next 2–12 h of TIBA treatment was 1.6 ± 0.2° h−1, significantly higher than that of the mock-treated epicotyls (P < 0.05; Student’s t test). The amplitudes never recovered and maintained a level lower than 10° until the end of the observation period. Importantly, the TIBA treatment never affected the elongation rate of epicotyls (Fig. S6), indicating that it specifically impaired circumnutation. Another auxin transport inhibitor 9-hydroxyfluorene-9-carboxylic acid (HFCA) exhibited essentially the same effects on circumnutation as those of TIBA (Fig. S7). The amplitudes decreased to 5° in 4 h of the HFCA treatment, and never recovered until the end of the observation period. The periods finally increased to 90 min. These results strongly suggest the involvement of auxin in the maintenance of circumnutation in adzuki bean epicotyls, as demonstrated in other plant species.

Distribution of phytohormones in adzuki bean epicotyls exhibiting circumnutation

In gravitropism and phototropism, the asymmetric distribution of auxin in bending axial organs is thought to be the direct cause of tropic curvature, in which a higher auxin level brings about an enhanced elongation of one side compared with the other side (Haga et al. 2005; Haga and Iino 2006; Iino 1991; Li et al. 1991; Parker and Briggs 1990; Went and Thimann 1937). However, as far as we know, no reports have examined the distribution pattern of auxin associated with circumnutation. Taking advantage of its large size, we investigated the distribution of endogenous hormones in adzuki bean epicotyls. Since the plane of curvature rotates along the longitudinal axis of the epicotyl during circumnutation, we attempted to compare hormone levels between not only the convex and concave sides at a certain time point (convex/concave) but also what would be the convex and concave sides at the next time point (pre-convex/pre-concave). After epicotyl sections 0–1, 1–2, and 2–3 cm below the tip were separated along the longitudinal axis into two halves with respect to the plane of curvature of circumnutation (Fig. 4a), namely, convex/concave or pre-convex/pre-concave pairs, the abundances of five different phytohormones in the sections were quantified by LC–MS (Fig. 4b–f′; Table S1). In the convex/concave pairs, we could not detect any significant difference in IAA abundance at any region along the longitudinal axis of the epicotyl (Fig. 4b). On the other hand, in the pre-convex/pre-concave pairs, IAA was distributed significantly more in the pre-convex half than in the pre-concave half of sections 1–2 cm below the tip (Fig. 4b′). Furthermore, GA1, which promotes the elongation of adzuki bean epicotyl sections in concert with auxin (Shibaoka 1972), and ABA, which inhibits the elongation of dwarf pea epicotyls (Sakiyama and Shibaoka 1990; Sakiyama-Sogo and Shibaoka 1993), were also distributed significantly more in the pre-convex half of sections 1–2 cm below the tip (Fig. 4c, c′, d, d′). These distribution patterns were specific to hormone species, and may not result from the experimental procedures, since JA and SA did not exhibit such characteristic distribution patterns (Fig. 4e, e′, f, f′). Furthermore, when we statistically analyzed the IAA content in epicotyl half sections 0–1 cm below the tip, we could not detect any significant differences among any parts of epicotyl (Table S2). Therefore, differences in the contents of hormones between the convex/concave pairs and the pre-convex/pre-concave pairs may be in a margin of error.

Fig. 4
figure 4

Distribution patterns of endogenous hormones in greened adzuki bean epicotyls exhibiting circumnutation. a Scheme demonstrating the four epicotyl parts along the longitudinal axis with respect to the plane of curvature of circumnutation. The convex and concave parts are in the plane of curvature, while the pre-convex and pre-concave parts are perpendicular to the plane. b–f, b′–f′ The contents of indole-3-acetic acid (IAA) (b, b′), gibberellin A1 (GA 1 ) (c, c′), abscisic acid (ABA) (d, d′), jasmonic acid (JA) (e, e′), and salicylic acid (SA) (f, f′) in the epicotyl sections are separated into a convex/concave (b–f) or pre-convex/pre-concave pair (b′–f′). Half sections were taken from 16 seedlings in each measurement, and three independent measurements were conducted. The asterisks indicate significant differences between the pairs by a Student’s t test (P < 0.05). The vertical bar on each point shows SE

Full size image

Effect of phytohormones on the elongation of epicotyl sections

Since three different hormones (IAA, GA1, and ABA) exhibited similar asymmetric distribution patterns in the same epicotyl region (Fig. 4b–d′), we examined their effects on the elongation of epicotyl sections (Fig. 5). First, epicotyl sections 0–1, 1–2, and 2–3 cm below the tip were floated on solutions containing different IAA concentrations for 60 min and their elongations were measured (Fig. 5a). A significant enhancement in the elongation of sections 0–1, 1–2, and 2–3 cm below the tip was detected only at 1 mM, over 10 µM, and over 100 µM IAA, respectively, compared to the sections treated with the control solution (P < 0.05; Tukey’s test). Hence, epicotyl sections 1–2 cm below the tip exhibited the highest responsiveness to IAA. We determined to use 100 µM IAA in the following experiments for two reasons. First, the elongations of sections 1–2 and 2–3 cm below the tip were similar at this concentration. Second, the elongations of sections at 100 µM IAA were smaller than those at 1 mM IAA and this concentration might better detect the effects of gibberellin and ABA.

Fig. 5
figure 5

Effect of indole-3-acetic acid (IAA) and gibberellin A3 (GA3) on the elongation of sections of adzuki bean epicotyls. Epicotyl sections 0–1, 1–2, and 2–3 cm from the tip were treated with different concentrations of IAA (a), or 100 µM IAA plus different concentrations of GA3 (b–d). Elongation of each section was monitored for 60 min at 5 min intervals. See text for statistical analyses. The vertical bar on each point shows SE (n = 5)

Full size image

We then examined the effects of gibberellin A3 (GA3), a substitute for GA1, since GA1 was unavailable in the market. Since GA3 affects the elongation of adzuki bean epicotyl sections only when it is applied together with auxin (Shibaoka 1972), we treated epicotyl sections with 100 µM IAA plus different concentrations of GA3 for 60 min (Fig. 5b–d). GA3 at any concentration did not show any significant effect on the elongation of sections 0–1 cm below the tip at any time point (P > 0.05; Tukey’s test) (Fig. 5b). On the other hand, the elongation of sections 1–2 and 2–3 cm below the tip clearly increased in the presence of GA3 (Fig. 5c, d). Treatment with 100 µM IAA plus 1 mM GA3 for 60 min increased the amount of elongation of epicotyl sections two-fold compared to the treatment with 100 µM IAA alone, and was five-fold larger than the control treatment. Importantly, epicotyl sections 1–2 cm below the tip exhibited a higher responsiveness to GA3 than those 2–3 cm below the tip. When treated with 100 µM IAA plus 100 µM or 1 mM GA3, a significant enhancement in the elongation of sections 1–2 cm below the tip was detected from 15 min and that of sections 2–3 cm below the tip was detected from 20 min, compared to the control treatment (P < 0.05; Tukey’s test). When treated with 100 µM IAA plus 10 µM GA3, a significant enhancement in the elongation of sections 1–2 cm below the tip was detected from 25 min, whereas that of sections 2–3 cm below the tip was detected from 40 min (P < 0.05; Tukey’s test).

We also examined the effects of ABA. Although ABA inhibited the elongation of dwarf pea epicotyls (Sakiyama and Shibaoka 1990; Sakiyama-Sogo and Shibaoka 1993), 100 µM ABA did not affect the elongation of adzuki bean epicotyl sections 1–2 cm below the tip in 60 min of the treatments, regardless of the presence of 100 µM IAA alone or together with 100 µM GA3 (Fig. S8). In summary, our results suggest that the epicotyl region 1–2 cm below the tip has the highest responsiveness to auxin and gibberellin in terms of enhancement of the elongation.

Effect of exogenously applied auxin and gibberellin on the circumnutation of adzuki bean epicotyls

To verify whether the asymmetric distributions of auxin and gibberellin are critical for the maintenance of circumnutation, we applied these hormones exogenously to intact adzuki bean epicotyls by a spray treatment, which should disturb the distribution pattern of endogenous hormones. As expected, the period of circumnutation of the 6-day-old greened epicotyls increased from 60 to 120 min in an 8 h IAA treatment (Fig. 6a). This effect was transient, and the period recovered to 60 min in an 18 h IAA treatment. The movement of the epicotyl tip was disordered in a 2 h IAA treatment, but was also restored to the elliptical trajectory in an 18 h IAA treatment (Fig. S9e′–i′). The descending rates of amplitude in the first 2 h and over the next 2–12 h of IAA treatment were 6.3 ± 1.0° and 0.9 ± 0.4° h−1 (Fig. 6b), not significantly different from the rate in mock-treated epicotyls over a respective period of observation (P > 0.05; Student’s t test). The treatment with 100 µM IAA did not affect the elongation rate of epicotyls (Fig. S6). On the other hand, we could not detect any remarkable effect of exogenously applied GA3 at 100 µM on circumnutation (Fig. 6c, d). Neither the period (Fig. 6c) nor the amplitude (Fig. 6d) was significantly different from those of mock-treated epicotyls at any time point (P > 0.05; Tukey’s test). The elongation rate of epicotyls was also never affected by the GA3 treatment (Figs. 6c, S6, S10).

Fig. 6
figure 6

Effect of indole-3-acetic acid (IAA) and gibberellin A3 (GA3) on the period and amplitude of circumnutation of greened adzuki bean epicotyls. Greened epicotyls exhibiting circumnutation were treated with 0.1% Tween 20 alone (mock) or together with 100 µM IAA (IAA) (a, b) or 100 µM GA3 (GA 3 ) (c, d) at 24 h of observation (indicated by arrows). The period (a, c) and amplitude (b, d) of circumnutation were determined every 5.3 h by integrating the values calculated in the x view and the y view (see “Materials and methods”). The asterisks and daggers indicate significant differences between “mock” and “IAA” or “GA 3 ”, and between “no treat” and “mock”, “IAA”, or “GA 3 ”, respectively, by a Tukey’s test (P < 0.05). The vertical bar on each point shows SE (n = 5)

Full size image

Orientation of the cortical microtubules in adzuki bean epicotyls exhibiting circumnutation

Auxin and gibberellin are involved in determining the mode of organ elongation through the regulation of cortical microtubule orientation (Shibaoka 1994). To ascertain the possibility that cortical microtubules participate in the regulation of differential growth of epicotyls exhibiting circumnutation, we examined the orientation of cortical microtubules in epidermal cells in the four parts of the bending zone (Figs. 4a, 7a). Epidermal cells only in the concave part exhibited a higher percentage of longitudinal cortical microtubules, whereas cells in the other three parts exhibited a more or less similar frequency of longitudinal, oblique, and transverse microtubules (Fig. 7b). Since any sign of reorientation of cortical microtubules in the pre-convex and/or pre-concave parts was not detected, which precedes the differential growth between the convex and concave parts, we have tentatively concluded that the reorientation of cortical microtubules could not be the cause for asymmetric changes in the mode of elongation of epicotyls associated with circumnutation. The predominance of longitudinal cortical microtubules in the concave part may be a result of mechanical stress produced by the epicotyl curvature. No disturbance in the period of circumnutation after the treatment of intact epicotyls with microtubule-disrupting or -stabilizing reagents (Fig. S11) also supports our conclusion.

Fig. 7
figure 7

Orientation of cortical microtubules in epidermal cells in the bending zone of greened adzuki bean epicotyls exhibiting circumnutation. a Cortical microtubules adjacent to the outer tangential wall of epidermal cells in the bending zone of greened epicotyls were visualized by indirect immunofluorescent microscopy. The orientation of microtubules was separated into four classes: L longitudinal (0–30° to the longitudinal axis of cells), O oblique (30–60°), T transverse (60–90°), and R random (no dominant alignment of microtubules). b The percentages of cells exhibiting each orientation of microtubules are shown separately in the four parts of the bending zone. The vertical bar on each column shows SE. We conducted three independent experiments, in which each 100–200 cells were observed

Full size image

Discussion

Circumnutation and asymmetric distributions of auxin and gibberellin

In the case of gravitropism and phototropism, the asymmetric distribution of auxin in the opposing sides of axial organs causes differential growth (Liscum et al. 2014; Morita 2010). Although auxin has long been assumed to play an important role in circumnutation (Britz and Galston 1982b; Hatakeda et al. 2003; Kim et al. 2016), how auxin is distributed in organs exhibiting circumnutation and how it regulates circumnutation have not yet been investigated. In this study, after specifying that the bending zone of circumnutation of adzuki bean epicotyls was in the region 1–3 cm below the tip (Fig. 2f), we further revealed that IAA together with GA1 was asymmetrically distributed there. The pre-convex half of epicotyls 1–2 cm below the tip contained significantly larger amounts of these hormones than the pre-concave half (Fig. 4b, b′). As far as we know, this is the first report on the asymmetric distribution of hormones associated with circumnutation. Furthermore, the epicotyl sections of this region exhibited the highest responsiveness to exogenously applied IAA and GA3. Compared to epicotyl sections 0–1 and 2–3 cm below the tip, epicotyl sections 1–2 cm below the tip responded to lower concentrations of IAA and GA3 (Fig. 5b–d), and besides, when treated with 100 µM IAA plus 10 µM GA3, a significant enhancement in the elongation of epicotyl sections 1–2 cm below the tip was detected earlier than that of sections from other regions (Fig. 5c, d).

The reason why asymmetric hormone distributions were detected in the pre-convex/pre-concave pairs and not in the convex/concave pairs, as in gravitropism or phototropism, could be explained by the latent period in hormone responses. We found that, in epicotyl sections 1–2 cm below the tip, the latent period between the application of IAA plus GA3 and the detection of a significant enhancement in elongation was 15 min (Fig. 5c). This is consistent with previous reports, which demonstrated that auxin stimulated elongation growth with a latent period of 10–15 min in various plant species (Badescu and Napier 2006; Evans 1974). For example, Barkley and Evans (1970) precisely monitored the elongation of stem sections of pea and cucumber using a custom-made chamber filled with the growth medium, and showed that 100 µM IAA increased the elongation rate from 10 min after the treatment began. Takahashi et al. (2012) measured the elongation of A. thaliana hypocotyl sections put on agar plates containing IAA, and showed that 10 µM IAA increased the elongation rate from around 10 min. Therefore, we can assume that the mode of elongation of the bending zone of adzuki bean epicotyls is determined by the hormonal status 15 min before, and that circumnutation with a 60 min period is maintained by the differential growth reflecting the hormonal status 15 min before, which is shifting sequentially in the circumferential direction.

As reported in pea epicotyls (Britz and Galston 1982a), morning glory shoots (Hatakeda et al. 2003), and pea roots (Kim et al. 2016), auxin-transport inhibitors impaired the circumnutation of adzuki bean epicotyls (Figs. 3, S7). Furthermore, exogenously applied IAA, which was expected to disturb the distribution pattern of endogenous IAA, transiently increased the period of circumnutation (Fig. 6a). These results strongly suggest that the distribution pattern of endogenous IAA is critical to maintain circumnutation, supporting our hypothesis described above. On the other hand, exogenously applied GA3 did not affect the circumnutation of adzuki bean epicotyls (Fig. 6c, d). We raise two possible explanations for the results. One is that the exogenous GA3 could not fully permeate into the epicotyl, and the other is that the endogenous level of gibberellin was already saturated in the epicotyl. The latter is suggested from the results that different concentrations of GA3 applied with IAA produced similar levels of enhancement in the elongation of epicotyl sections 1–2 cm below the tip (Fig. 5c). Although the endogenous concentration of GA1 estimated from the LC–MS measurements is much lower than 10 µM, the lowest concentration of GA3 exogenously applied (Figs. 4c, 5c), we suppose that the endogenous gibberellin might be localized only in the limited part, where the local concentration of gibberellin is expected to be higher than 10 µM. In any case, the significance of asymmetric distribution of gibberellin in the regulation of circumnutation needs to be further investigated.

Possible mechanisms for the asymmetric distribution pattern of hormones

The next important subject is how the circumnutation-specific asymmetric distributions of auxin and gibberellin are maintained. In the case of auxin, auxin carriers involved in polar auxin transport may play important roles. For example, PIN3, the major auxin efflux carrier functioning in lateral auxin transport, is responsible for the asymmetric distribution of auxin in organs exhibiting gravitropism and phototropism in A. thaliana (Friml et al. 2002). Rakusová et al. (2011) reported the details of PIN3 relocalization during gravitropism in hypocotyls. In vertically placed hypocotyls, the signals from PIN3-GFP were detected at the plasma membrane in both the outer and inner sides of endodermis cells. After hypocotyls were placed horizontally for gravistimulation, the GFP signals at the outer sides of the endodermis cells in the upper half of the hypocotyl gradually disappeared. Since DR5rev::GFP fluorescence was detected in the lower half of the hypocotyl after gravistimulation, this relocalization of PIN3 must be responsible for the asymmetric distribution of auxin. Similarly, Ding et al. (2011) reported PIN3 relocalization during phototropism. When hypocotyls were illuminated unilaterally with white light, the signals from PIN3-GFP at the outer sides of endodermis cells in the illuminated half of the hypocotyl gradually disappeared. Since DR5rev::GFP fluorescence was detected in the shaded half of the hypocotyl after unilateral illumination, this relocalization of PIN3 must be responsible for the asymmetric distribution of auxin.

Such polar relocalization of putative auxin carriers could also explain the asymmetric distribution of auxin in adzuki bean epicotyls. In epicotyls exhibiting circumnutation, gravity and/or light from above would stimulate the displacement of auxin carriers from the outer sides of endodermis cells in the convex part, and auxin would start to be asymmetrically distributed to the concave part. Since auxin takes some time to move across the epicotyl and circumnutation continues, auxin would be asymmetrically distributed more in the pre-convex part than in the pre-concave part. To ascertain this possibility, once the auxin carriers that participate in circumnutation are identified, their localization patterns associated with circumnutation should be urgently investigated.

On the other hand, how the gibberellin distribution is maintained is more uncertain since the mechanism of gibberellin transport has not yet been revealed (Yamaguchi 2008). Although the possibility that unknown gibberellin carriers are involved in the maintenance of gibberellin distribution still remains, another possibility should also be considered. Since GA1 synthesis is promoted through auxin signaling (Ross et al. 2000; Yamaguchi 2008), a local synthesis of gibberellin might contribute to maintain the distribution pattern. In pea epicotyls, IAA increases the level of GA1 through upregulating the expression of PsGA3ox1, which encodes the enzyme that converts gibberellin A20 (GA20) to GA1 (O’neill and Ross 2002). Therefore, if the IAA level exhibits a periodic change, the gibberellin level may also change periodically.

Driving mechanisms for circumnutation

In adzuki bean epicotyls, the basal end of the bending zone was almost always identical to the apical end of the completed region, and as epicotyls elongated, it moved up along the longitudinal axis of the epicotyl (Fig. 2f). These results indicate that epidermal cells constructing the bending zone are regularly turning over with their elongation growth. In the apical region of the bending zone, the elongation of young epidermal cells is sequentially accelerated in a circumferential direction by the asymmetrically distributed hormones, resulting in the initiation of differential growth between the opposing sides of the epicotyl. After being involved in the maximal differential growth in the middle region, older epidermal cells sequentially complete the elongation in the basal region of the bending zone, resulting in the termination of epicotyl differential growth. This driving mechanism of circumnutation seems to be different from that proposed in the French bean. In the twining shoots of French beans, whose epidermal cell file is almost vertical, as in adzuki bean epicotyls, partially reversible length changes in epidermal cells were associated with the revolving movement (Caré et al. 1998). This reversible cell elongation is possibly caused by alternating changes in turgor pressure, ionic composition, and water permeability (Badot et al. 1990; Comparot et al. 2000; Millet et al. 1988).

On the other hand, it has also been proposed that skewed cell files drive circumnutation. When the stalk of a female flower of aquatic eelgrass (Vallisneria asiatica) elongated toward the water surface, it rotated around the longitudinal axis (Kosuge et al. 2013). After the bud reached the water surface, the stalk ceased rotation, and thereafter, exhibited both circumnutation and rotation. When the female flower exhibited circumnutation, the peduncle was strongly twisted and the epidermal cell files were strongly skewed. With experiments using an artificial model, Kosuge et al. (2013) proposed that the helical growth of the peduncle drives circumnutation. These studies indicate that different mechanisms drive circumnutation in different organs depending on their characteristic cellular organizations and modes of cell elongation.

Characteristics of circumnutation

We showed that the period of circumnutation of adzuki bean epicotyls was maintained at approximately 60 min (Fig. 2b). Although circumnutation occurs in various plant species, the period of circumnutation differs from species to species. In etiolated rice coleoptiles, the period of circumnutation slightly increased from 160 to 200 min with the elongation of coleoptiles (Yoshihara and Iino 2005). In A. thaliana, hypocotyls grown under white light exhibited two patterns of circumnutation with a shorter period of 30 min (SPN) and a longer period of 90 min (LPN). The period lengths of both SPN and LPN tended to increase during a 3-day observation (Schuster and Engelmann 1997). One of the possible regulators for the periodicity of circumnutation is a circadian rhythm. The period of circumnutation in A. thaliana inflorescence stems under continuous white light fluctuated in a circadian manner, and was disturbed in the toc1 and elf3 mutants, which are deficient in circadian clock functions (Niinuma et al. 2005). However, as described above, adzuki bean epicotyls did not exhibit such fluctuation in the period of circumnutation (Fig. 2b). This difference is attributable to growth conditions. Niinuma et al. (2005) cultivated A. thaliana seedlings under light and dark cycles, but we raised adzuki been seedlings under continuous white light (see “Materials and Methods”). Someya et al. (2006) raised A. thaliana under continuous light and showed that the period of circumnutation of inflorescence stems did not fluctuate under continuous light conditions. Growth conditions might affect circadian clock responsiveness and/or hormone status in plant organs and cause differences in circumnutation characteristics even in the same plant species.

The direction of movement is also an important parameter of circumnutation. We showed that all adzuki bean epicotyls eventually exhibited a clockwise circumnutation (Fig. 2d, e, Fig. S1). On the other hand, in the case of rice coleoptiles, the ratio of counter-clockwise circumnutation to clockwise was approximately 3:1 (Yoshihara and Iino 2005). In A. thaliana hypocotyls, the direction of SPN was usually counter-clockwise, while that of LPN was usually clockwise (Schuster and Engelmann 1997). French bean twining shoots always exhibited counter-clockwise movements (Millet et al. 1984). Therefore, the mechanisms that regulate the period of circumnutation and determine the direction of movement are tremendously diverse. Smyth (2016) pointed out that the handedness of some kinds of helical growth is variable, and the direction might reflect an early developmental asymmetry in the shoot apical meristem. If the distribution pattern of auxin in the bending zone of circumnutation reflects that in the shoot apical meristem, it might explain how the direction of circumnutation is determined. This should be another important subject to be investigated in future. Although each plant species or organ harbors a unique system, it eventually makes similar periodic and circular movements. This fact suggests that circumnutation has some conserved significance in plants that we have not yet been able to disclose.