Introduction

It has been well established that both nitric oxide (NO) and heme oxygenase (HO; EC. 1.14.99.3)—the latter of which degrades heme into equimolar amounts of biliverdin (BV), carbon monoxide (CO) and free iron (Fe2+)—are important signalling components with multiple biological functions in plants (for review, see Delledonne 2005; Leitner et al. 2009; Shekhawat and Verma 2010). For example, a large number of studies on the roles of NO and HO-1 (an inducible form of HO) have focused on their involvement in abiotic stresses, including drought (García-Mata and Lamattina 2001; Liu et al. 2010), salt stress (Zhao et al. 2007; Xie et al. 2011), cold acclimation (Zhao et al. 2009; Bai et al. 2012), metal toxicity (Noriega et al. 2004; Han et al. 2008; Besson-Bard et al. 2009), and paraquat toxicity (Xu et al. 2012). In addition, these two signalling components have been confirmed to be individually or simultaneously involved in various developmental processes, such as seed germination (Beligni and Lamattina 2000; Liu et al. 2010), stomatal movement (García-Mata and Lamattina 2001; Cao et al. 2007) and root development (Pagnussat et al. 2002, 2003; Correa-Aragunde et al. 2004; Xuan et al. 2008). Further results indicated that NO or HO-1 plays a key role in plant responses to various phytohormones, including gibberellic acid (GA), auxin and abscisic acid (ABA) (Beligni et al. 2002; Cao et al. 2007; Xuan et al. 2008; Wu et al. 2011). In some of these responses, NO production and HO-1 up-regulation as well as CO synthesis occur in parallel, or in short succession of one another, thus suggesting possible crosstalk among phytohormones, NO, HO-1 and CO (Cao et al. 2007; Noriega et al. 2007; Xuan et al. 2008; Liu et al. 2010; Bai et al. 2012).

Normally, a number of NO-regulated physiological processes may be mediated by soluble guanylate cyclase (sGC), or through affecting of specific gene transcription, or interacting with biomolecules (Beligni et al. 2002; Polverari et al. 2003; Parani et al. 2004; Salmi et al. 2007; Shoulars et al. 2008; Wang et al. 2009). NO can induce the activation of sGC by binding to the heme group of the enzyme, thus leading to a transient increase of the cellular messenger cGMP (Stamler 1994). In animal and even plant cells, cGMP is an important component of NO signalling (Polte et al. 2000; Salmi et al. 2007; Pasqualini et al. 2009). In tobacco, for example, it was found that NO treatment could induce a transient increase of cGMP levels, and the GC inhibitor 6-anilino-5,8-quinolinedione (LY83583; LY) was able to suppress NO-induced activation of phenylalanine ammonia lyase (PAL) gene expression (Durner et al. 1998). In Arabidopsis, besides a brassinosteroid receptor (AtBRI1) functioning as a GC in vitro (Kwezi et al. 2007), a recent report showed that the leucine-rich repeat receptor-like kinase receptor AtPepR1 also contained a putative GC domain and had GC activity, thus generating cGMP from GTP (Qi et al. 2010)—although this result has been questioned (Ashton 2011; Berkowitz et al. 2011), GC activity has been determined in Arabidopsis (Ludidi and Gehring 2003; Mulaudzi et al. 2011). cGMP could be detected by mass spectrometry and radio-immune assays, and its content has been demonstrated in different plant species and tissues (Brown and Newton 1992; Penson et al. 1996; Durner et al. 1998). Further results showed that GA-induced cGMP plays an important role in α-amylase synthesis and secretion through regulation of GAMYB transcription in barley aleurone layers (Penson et al. 1996). Use of the fluorescent reporter FlincG as an endogenous cGMP sensor showed that NO and the phytohormones GA, ABA, auxin and jasmonic acid, thought to act through cGMP elevation, rapidly increased cellular cGMP in a physiological relevant context (Isner and Maathuis 2011; Isner et al. 2012).

The cereal aleurone layers provide a convenient model system for studying molecular and signalling mechanisms involved in GA-regulated gene expression, especially α-amylase genes, which are beneficial for seed germination. By contrast, ABA exhibits the opposite effects that block the GA responses in aleurone layers through a mechanism that involves the expression of a distinct set of genes such as PKABA1, an ABA-induced protein kinase gene (Gómez-Cadenas et al. 2001). It has also been accepted that the subunit of a heterotrimeric G protein (Jones et al. 1998), cyclic guanosine monophosphate (cGMP; Penson et al. 1996), Ca2+-dependent and Ca2+-independent events (Gilroy and Jones 1992), reversible protein phosphorylation (Gómez-Cadenas et al. 1999) and reactive oxygen species (ROS), such as hydrogen peroxide (H2O2) (Ishibashi et al. 2012), are involved in α-amylase gene expression through regulation of several cis-elements and transcription factors, including GAMYB (Gubler et al. 1999; Ishibashi et al. 2012). A well-known NO-releasing compound, sodium nitroprusside (SNP), not only delayed GA-induced programmed cell death in barley aleurone layers, but also prolonged the secretion of α-amylase (Beligni et al. 2002). More recently, similar physiological roles of CO and hemin, an inducer of HO-1, were also reported (Wu et al. 2011).

Three classes of α-amylase genes have been identified in wheat (Triticum aestivum), but only α-Amy1 and α-Amy2 are active in the aleurone cells of germinating wheat grain, and α-Amy3 is only expressed in developing grain (Huttly et al. 1988). Previous results also showed that the expression of α-Amy2/54, one of members of the α-Amy2 gene family, was enhanced by the application of GA (Huttly et al. 1988; Huttly and Baulcombe 1989). However, molecular mechanisms and corresponding signalling transduction of the expression of this gene remain unknown.

In this paper, our aim was to characterize the roles of NO, cGMP and HO-1 in GA-induced α-Amy2/54 gene expression, with particular emphasis on the temporal signature of HO-1 as a possible signalling component in this event. Further pharmacological and biochemical evidence illustrated that HO-1 is a newly identified component required for NO-induced cGMP-mediated α-Amy2/54 gene expression in GA-treated wheat aleurone layers.

Materials and methods

Chemicals

All chemicals were purchased from Sigma-Aldrich (St Louis, MO, USA) unless stated otherwise. The chemicals used for treatments were GA (50 μM) (catalogue number G7645; Wu et al. 2011); hemin (10 μM; an inducer of HO-1; catalogue number 51280) (Lang et al. 2005; Xuan et al. 2008); zinc protoporphyrin-IX (ZnPPIX, 100 μM; a specific inhibitor of HO-1; catalogue number 282820) (Wu et al. 2011; Bai et al. 2012); sodium nitroprusside (SNP, 100 μM; a well-known NO-releasing compound under light conditions; catalogue number 13451) (Delledonne et al. 1998; Beligni et al. 2002; Bethke et al. 2004; Floryszak-Wieczorek et al. 2006); spermine NONOate (NONOate, 10 μM; another well-known NO-releasing compound; purchased from Cayman chemical company; catalogue number 82150) (Yamasaki and Cohen 2006; Fröhlich and Durner 2011); 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide potassium salt (cPTIO, 200 μM; a scavenger of NO; catalogue number C221) (Delledonne et al. 1998; Pagnussat et al. 2003; Bethke et al. 2004); 6-anilino-5,8-quinolinedione (LY, 50 μM; a GC inhibitor; catalogue number A6563); and 8-bromoguanosine 3′,5′-cyclic monophosphate (8-Br, 10 μM; a cell-permeable cGMP derivative; catalogue number B1381) (Penson et al. 1996; Durner et al. 1998; Polte et al. 2000; Neill et al. 2002; Pagnussat et al. 2003; Hu et al. 2005). Additionally, an old SNP solution (presumely containing equalmolar amounts of ferrocyanide, nitrate and nitrite, two normal products of NO decomposition) (Beligni and Lamattina 2000; Graziano et al. 2002) was obtained as a negative control by maintaining a 100 μM SNP solution for 10 days in the light in an open tube to eliminate NO before its application (Tossi et al. 2009). Because SNP solution is extremely photosensitive (Wang et al. 2002; Floryszak-Wieczorek et al. 2006), half-life for SNP when exposed to light was t ½ = ca.12 h (Floryszak-Wieczorek et al. 2006). So, 10 days is long enough for SNP solution releasing all of the NO. The concentrations of above compounds used in this study were determined in pilot experiments from which significant responses were obtained.

Preparation of CO aqueous solution

CO-saturated aqueous solution was freshly produced (Han et al. 2008; Wu et al. 2011) by bubbling CO gas gently through a glass tube into 50 ml of 5 mM CaCl2 (Kuo et al. 1996) for at least 20 min, a duration long enough to saturate the solution with CO. The saturated stock solution (100 % saturation) was then diluted immediately with 5 mM CaCl2 solution to the saturation of 1 % (v/v).

Preparation of wheat aleurone layers

Aleurone layers were prepared from de-embryonated wheat (Triticum aestivum ‘Yangmai 13’) grains as described previously (Kuo et al. 1996; Beligni et al. 2002; Mrva et al. 2006; Wu et al. 2011). In brief, the embryo and distal end of the grain were removed, and the de-embryonated half-grains were surface-sterilized in 1 % sodium hypochlorite solution for 20 min, and rinsed extensively with sterile water for several times. Then the de-embryonated half-grains were imbibed in sterile water at 25 ± 1 °C for 48 h. Aleurone layers were isolated from the imbibed grain by removing the starchy endosperm under sterile conditions. Afterwards, the isolated aleurone layers were treated as indicated, and corresponding biochemical and molecular analyses were carried out. Layers incubated in 5 mM CaCl2 alone were regarded as the control (Con).

Determination of heme oxygenase and α-amylase activities

Heme oxygenase (HO) activities were analysed using the method described in our previous reports (Han et al. 2008; Wu et al. 2011). The concentration of BV was estimated using a molar absorption coefficient at 650 nm of 6.25 mM−1 cm−1 in 0.1 M HEPES–NaOH buffer (pH 7.2). One unit of activity (U) was calculated by taking the quantity of the enzyme to produce 1 nmol BV per 30 min.

α-Amylase activity in the aleurone incubation medium was measured using the previously reported starch–iodine procedure (Jones and Varner 1967; Beligni et al. 2002; Palma and Kermode 2003). One unit of activity (U) is defined as a change of one absorbance unit at 620 nm min−1.

Agar-plate assay of amylase

The agar-plate assay was performed as described previously (Lanahan and Ho 1988; Ikeda et al. 2001). In brief, wheat embryoless half-seeds were placed on 2 % agar plates containing 10 mM sodium acetate and 5 mM CaCl2 at pH 5.0. Different plates were made by adding GA, SNP, NONOate individually or simultaneously to the cooled agar. To detect secreted amylase activity, soluble potato starch (0.2 %) was added to the agar before autoclaving. Agar plates were then developed by incubating the plates in I-KI solution. Half-seeds that synthesized and secreted amylase had transparent halos around them resulting from the digestion of the starch by amylases.

Western blotting analysis for wheat HO-1

Rabbit polyclonal antibody was made against the mature wheat HO-1 with a molecular weight of 26 kDa (Xu et al. 2011). Homogenates obtained for the HO activity assays were also analysed by western blotting. Fifty micrograms of protein from homogenates were subjected to SDS-PAGE using a 12.5 % acrylamide resolving gel (Mini Protean II System, Bio-Rad, Hertz, UK) according to the method described in our previous report (Xuan et al. 2008). Meanwhile, Coomassie brilliant blue staining was used to show that equal amounts of proteins were loaded. Additionally, the films were scanned and analysed using Quantity One v4.4.0 software (Bio-Rad, USA).

NO production determined by using Greiss reagent

NO content was determined using the methods described by Rockel et al. (2002), Zhou et al. (2005) and Liu et al. (2009) with slight modifications. Twenty aleurone layers were ground in a mortar and pestle in 1 ml of 50 mM cool acetic acid buffer (pH 3.6, containing 4 % zinc diacetate). The homogenates were centrifuged at 10,000 g for 15 min at 4 °C. The supernatant was collected. The pellet was washed by 1 ml of extraction buffer and centrifuged as before. The two supernatants were combined and 0.1 g of charcoal was added. After vortex and filtration, the filtrate was leached and collected. The mixture of 1 ml of filtrate and 1 ml of the Greiss reagent was incubated at room temperature for 30 min. Meanwhile, identical filtrate which was pretreated with 200 μM 2-(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide potassium salt (cPTIO), the scavenger of NO, for 20 min, was used as blanks. Absorbance was determined at 540 nm. NO content was calculated by comparison to a standard curve of NaNO2.

Real-time RT-PCR analysis

Real-time quantitative RT-PCR reactions were performed using a Mastercycler® ep realplex real-time PCR system (Eppendorf, http://www.eppendorf.com/) with SYBR® Premix Ex Taq™ (TaKaRa, http://www.takara-bio.com/) according to the manufacturer’s protocols. Using specific primers (Supplementary Table S1), expression levels of α-Amy2/54 (accession number X13580), α-Amy1 (accession number S42213), α-Amy2/8 (accession number X13576), HO-1 (accession number HM014348), GAMYB (accession number AB044084), and PKABA1 (accession number AF519805) are presented as values relative to corresponding control samples at the indicated times, after normalization to actin (accession number AB181991) transcript levels.

Statistical analysis

Where indicated, results are expressed as mean values ±SD of three independent experiments with at least three replicates for each. Statistical analysis was performed using SPSS 16.0 package. For statistical analysis, t test (P < 0.05 or P < 0.01) or Duncan’s multiple test (P < 0.05) was chosen as appropriate.

Results

Both SNP and NONOate increase α-amylase activity in a GA-dependent fashion

Previous results showed that in barley aleurone layers, two NO-releasing compounds (SNAP and SNP) strengthened the GA-stimulated synthesis and secretion of α-amylase (Beligni et al. 2002). To further characterize the sensitivities of wheat embryoless half-seeds and aleurone layers to NO, we first tested the effects of SNP and another NO-releasing compound—spermine NONOate (NONOate)—on amylase and α-amylase activities. Compared to control sample, agar-plate assays (Fig. 1a) showed cleared zones (plaques) on plate containing GA alone that indicated significant secretion of amylase, which was appreciably strengthened by either SNP or NONOate cotreatment. As expected (Floryszak-Wieczorek et al. 2006), we also noticed that SNP only works in the light conditions. Comparatively, the addition of SNP and NONOate alone brought about identical responses, in comparison with control treatment. Subsequent enzyme determination showed that the GA induced the synthesis and secretion of α-amylase, which became more distinct with time, such that α-amylase activity increased from 0 to 11.67 U layer−1 after 48 h of incubation in GA-containing medium (Fig. 1b). Further time-dependent increases in α-amylase activity induced by SNP, and NONOate, only obviously appeared in GA-treated aleurone layers. To attribute a specific role of NO in regulating GA-induced α-amylase, an old SNP solution (containing no NO but ferrocyanide, nitrate and nitrite), was applied as a negative control. Contrary to the inducible effects of SNP, we discovered that old SNP did not exhibit similar significant effects (Supplemental Table S2). These results suggested that both SNP and NONOate-increased α-amylase activity were in a GA-dependent fashion.

Fig. 1
figure 1

Two well-known NO-releasing compounds SNP and spermine NONOate increase the amount of amylase secreted by GA-treated wheat embryoless half-seeds or aleurone layers. Wheat embryoless half-seeds (a) or aleurone layers (b) were separately placed on starch plates or incubated in a medium containing 5 mM CaCl2 alone (Con) or together with 50 μM GA, 100 μM SNP, 10 μM spermine NONOate (NONOate) individually and simultaneously. Afterwards, amylase agar-plate assay was carried out by staining with I-KI after 48 h of various treatments (a). Meanwhile, α-amylase activities were determined at the indicated time points (b). Data are mean ± SD of three independent experiments with at least three replicates for each. Within the same treatment time points, asterisks indicate significant difference (P < 0.05 and P < 0.01) between GA + SNP or GA + NONOate treatment and GA-treated alone samples according to t test (b)

Full size image

SNP-induced up-regulation of α-Amy2/54 transcription, α-amylase activity, and NO production were sensitive to cPTIO

Previous results confirmed that GA-treated de-embryonated seeds synthesized relatively low pI isozymes (Amy2) of α-amylase, which normally appeared much earlier than other isozymes (Huttly et al. 1988). In our experiments, we compared Amy1, Amy2/8 and Amy2/54 transcripts in GA-treated aleurone layers, and found that α-Amy2/54 was the most highly expressed α-amylase gene induced by GA. For example, after treatment with 50 μM GA for 12 h, α-Amy2/54, Amy2/8, and Amy1 exhibited about 18.2-, 15.4-, and 12.1-fold increases, respectively, compared with the control (5 mM CaCl2, 0 h) sample (Supplemental Fig. S1). In view of the fact that GA-regulated wheat α-Amy2/54 promoter contains two regions essential for its high level expression (Tregear et al. 1995), α-Amy2/54 gene expression was used as a molecular probe to investigate the detailed molecular mechanism after wheat aleurone layers were treated with GA plus NO-releasing compounds.

To confirm the role of NO in the SNP-induced response, we also adopted a pharmacological approach by using the scavenger of NO, cPTIO. Consistent with the previous report (Sarath et al. 2006), the addition of GA in the presence of (especially) or absence of SNP brought about not only the up-regulation of α-Amy2/54 transcripts and α-amylase activity (Fig. 2a), but also enhanced NO production (Fig. 2b). Furthermore, the GA alone- or GA plus SNP-induced responses were markedly reduced by cotreatment with cPTIO. There were significant decreases in α-Amy2/54 transcript and NO production in the sample treated with cPTIO alone, compared with control sample. We also noted that enhanced NO production induced by GA plus SNP treatment apparently preceded the up-regulation of α-Amy2/54 transcript and thereafter α-amylase activity. These data, together with the performances of the negative control of SNP (Supplemental Table S2), highlight the importance of NO in the abovementioned mode of action.

Fig. 2
figure 2

Effects of the NO-releasing compound SNP, NO scavenger cPTIO on α-Amy2/54 expression, α-amylase activity, and NO content. Wheat aleurone layers were incubated in a medium containing 5 mM CaCl2 (Con) alone or together with 50 μM GA, 100 μM SNP, 200 μM cPTIO alone or their combinations. α-Amy2/54 transcription (left) and α-amylase activity (right) were detected after 12 or 24 h of different treatments, respectively (a). Meanwhile, NO contents were detected by using Greiss reagent at the indicated time points (b). Data are mean ± SD of three independent experiments with at least three replicates for each. Within each set of experiments, bars denoted by the different letters are different significantly in comparison with Con treatment at P < 0.05 according to Duncan’s multiple test

Full size image

8-Br-cGMP, a cGMP derivative, mimicked SNP responses

Previous studies both in animals and plants have shown that NO-mediated functions are cGMP-dependent or -independent (Kim et al. 1997; Durner et al. 1998; Shen et al. 1998; Pagnussat et al. 2003). In the following experiments, our results confirmed that the treatment of aleurone layers with the guanylate cyclase inhibitor 6-anilino-5,8-quinolinedione (LY83583; LY) significantly suppressed the inducible effects of GA on α-Amy2/54 transcripts and α-amylase activity (Fig. 3), suggesting the possible occurrence of a GC-like enzyme in this process. Interestingly, 8-Br-cGMP (8-Br), a cGMP derivative known to be biologically effective when applied to plant cells (Durner et al. 1998; Hu et al. 2005; Bai et al. 2012), was able to mimic the strengthening effects of SNP in the up-regulation of α-Amy2/54 transcript and induction of α-amylase activity in GA-treated aleurone layers—all of which could be significantly reversed by cotreatment with LY. These observations might be related to cGMP involvement in GA-induced response in α-amylase gene expression as reported by Penson et al. (1996). Finally, it should be specified that in regard to the above parameters, no significant difference was observed when 8-Br or LY was applied alone, compared to control; further suggesting that the inducible effect of cGMP in α-amylase gene expression might also be GA-dependent, similar to the action of SNP (Fig. 1).

Fig. 3
figure 3

8-Br-cGMP, a cGMP derivative, mimics the effects of the NO-releasing compound SNP on the induction of α-Amy2/54 expression and α-amylase activity. Wheat aleurone layers were incubated in a medium containing 5 mM CaCl2 (Con) alone or together with 50 μM GA, 100 μM SNP, 10 μM 8-Br-cGMP (8-Br), or 50 μM LY83583 (LY) alone or their combinations. α-Amy2/54 transcription (left) and α-amylase activity (right) were detected after 12 or 24 h of different treatments, respectively. Data are mean ± SD of three independent experiments with at least three replicates for each. Within each set of experiments, bars denoted by the different letters are different significantly in comparison with Con treatment at P < 0.05 according to Duncan’s multiple test

Full size image

The SNP- and 8-Br-induced responses were sensitive to zinc protoporphyrin IX, an inhibitor of HO-1, and reversed by CO

As reported previously, cGMP is a crucial mediator in the endothelial regulation of HO-1 expression; and endothelial protection afforded by the NO and cGMP systems is causally related to the induction of HO-1 (Polte et al. 2000). To test the possible role of HO-1 in NO and cGMP signalling, leading to the induction of α-amylase gene expression, we added the potent HO-1 inhibitor zinc protoporphyrin IX (ZnPPIX; Wu et al. 2011), and one of the by-products of HO-1 [carbon monoxide (CO) aqueous solution] to GA-treated aleurone layers. ZnPPIX clearly prevented the induction action of GA with or without SNP or 8-Br treatments on α-Amy2/54 transcripts and α-amylase activity, respectively (Fig. 4). However, when exogenous 1 % saturation of CO aqueous solution was added together with ZnPPIX in the presence of GA plus 8-Br or SNP, the α-Amy2/54 gene expression and α-amylase activity inhibited by ZnPPIX treatment returned close to levels displayed in aleurone layers treated with GA plus 8-Br or SNP samples. However, the application of ZnPPIX alone did not produce any different effects as compared to control. These data, together with the identification of HO-1 as a NO and cGMP target molecule in animals (Polte et al. 2000), highlight that at least in wheat aleurone layers, there is a functional link among NO, cGMP, and HO-1.

Fig. 4
figure 4

The SNP- and 8-Br-cGMP-induced α-Amy2/54 gene expression and increases of α-amylase activity are sensitive to ZnPPIX, but reversed by CO. Wheat aleurone layers were incubated in a medium containing 5 mM CaCl2 (Con) alone or together with 50 μM GA, 10 μM 8-Br-cGMP (8-Br), 100 μM SNP, 100 μM ZnPPIX, or 1 % CO-saturated aqueous solution alone or their combinations. α-Amy2/54 transcription (left) and α-amylase activity (right) were detected after 12 or 24 h of different treatments, respectively. Data are mean ± SD of three independent experiments with at least three replicates for each. Within each set of experiments, bars denoted by the different letters are different significantly in comparison with Con treatment at P < 0.05 according to Duncan’s multiple test

Full size image

HO-1 and CO exhibit similar induction of α-amylase activity and α-Amy2/54 transcription via a GA-dependent fashion

To further confirm the role of HO-1 in the GA-induced α-amylase activity, the HO-1 inducer hemin was used. Time course experiments (Fig. 5a) illustrated that, compared with GA-treatment alone, exposure of wheat aleurone layers to 10 μM hemin as well as 1 % saturation of CO aqueous solution for 48 h caused a progressive increase in α-amylase activity only in GA-containing medium. Subsequent molecular evidence of α-Amy2/54 gene expression (Fig. 5b) showed good agreement with these results. However, BR and Fe2+ did not have the similar effects like CO in the induction of α-amylase activity (Wu et al. 2011).

Fig. 5
figure 5

Changes of α-amylase activity, α-Amy2/54 and HO-1 transcripts, HO activity and HO-1 protein level in response to hemin, CO, and ZnPPIX. Wheat aleurone layers were incubated in a medium containing 5 mM CaCl2 (Con) alone or together with 50 μM GA, 10 μM hemin, 1 % CO-saturated aqueous solution, or 100 μM ZnPPIX alone or their combinations. α-Amylase activity (a) was determined at the indicated time after various treatments, α-Amy2/54 transcription was also determined after 12 h of incubation (b). Meanwhile, HO-1 transcript and HO activity (c) were also determined after 8 h of different treatments. HO-1 protein expression was analyzed by western blotting (d), and Coomassie Brilliant Blue-stained gels are present to show that equal amounts of proteins were loaded (e). The number below the band (d) illustrates the relative abundance of the corresponding HO-1 protein compared with that of the Con sample. Data are mean ± SD of three independent experiments with at least three replicates for each. Within the same treatment time points, asterisks indicate significant difference (P < 0.05 and P < 0.01) between GA + Hemin or GA + CO treatment and GA-treated alone samples according to t test (a). Within each set of experiments, bars denoted by the different letters are different significantly in comparison with Con treatment at P < 0.05 according to Duncan’s multiple test (b, c)

Full size image

Subsequent results showed that GA treatment up-regulated HO-1 transcript, and induced HO activity as well as higher HO-1 protein level after 8 h of treatment. The effects of GA were obviously strengthened by cotreated with hemin (Fig. 5c–e). However, the application of ZnPPIX differentially reduced the GA- and GA plus hemin-induced HO activity, but did not alter the HO-1 transcripts and HO-1 protein contents. Similar responses were also observed when ZnPPIX was applied alone. These data indicate that HO-1 and CO might be required for hemin-induced α-Amy2/54 gene expression; and this response was regulated through a GA-mediated process.

SNP-induced HO-1 gene expression and HO activity were sensitive to cPTIO and LY

To assess if HO-1 is associated with the SNP response leading to α-Amy2/54 gene expression, we studied the SNP-induced expression of this enzyme in detail. In our further experiments, 100 μM SNP together with GA significantly up-regulated HO activity and HO-1 protein level, in comparison with the individual GA or SNP treatment (Fig. 6a, b). To further clarify the roles of NO and cGMP in the induction of HO-1 gene expression, the NO scavenger cPTIO, the GC inhibitor LY, and the potent inhibitor of HO-1 ZnPPIX were used together with GA plus SNP. Remarkably, both cPTIO and LY (in particular) significantly prevented the enhancement of HO-1 gene expression and corresponding HO activity. In similar conditions, and consistent with previous results (Fig. 5c–e) in which the effect of ZnPPIX was investigated in HO-1 gene expression, there was a significant decrease of HO activity but not in HO-1 transcripts. cPTIO applied alone triggered a weak decrease in HO activity rather than altering HO-1 mRNA. Additionally, no significant effects occurred when exogenous superoxide dismutase (SOD, 50 U/ml) was added in medium containing GA plus SNP (Fig. 6d, e), thus ruling out the interfering effects conferred by ROS in the HO-1 gene expression (Yannarelli et al. 2006).

Fig. 6
figure 6

Effects of SNP, cPTIO, LY, ZnPPIX and superoxide dismutase (SOD) on HO-1 transcript, HO-1 protein level and HO activity in GA-treated wheat aleurone layers. Wheat aleurone layers were incubated in a medium containing 5 mM CaCl2 (Con) alone or together with 50 μM GA, 100 μM SNP, 200 μM cPTIO, 50 μM LY, 100 μM ZnPPIX, or 50 U/ml SOD alone or their combinations. HO activity (a, e), HO-1 protein content (b), and its transcripts (d) were determined after 8 h of treatments. The number below the band (b) indicates the relative abundance of the corresponding HO-1 protein compared with that of Con sample. Coomassie Brilliant Blue-stained gels are present to show that equal amounts of proteins were loaded (c). Data are mean ± SD of three independent experiments with at least three replicates for each. Bars denoted by the different letters are different significantly in comparison with Con treatment at P < 0.05 according to Duncan’s multiple test

Full size image

Induction of HO-1 in response to cGMP

The reduction effect of ZnPPIX on 8-Br-induced α-Amy2/54 gene expression in GA-treated wheat aleurone layers (Fig. 4) suggests that HO-1 might be a positive trigger. To confirm this possibility, we studied the effect of 8-Br and LY on the gene expression of HO-1. Aleurone layers exposed to 50 μM GA together with 8-Br or LY for 8 h showed marked increases or decreases in HO-1 gene expression, HO activity and HO-1 protein level, compared with GA treated alone (Fig. 7). Contrary to these results, combination with LY in 8-Br-treated aleurone layers resulted in dramatically reduced effects. In addition, 8-Br-treated alone sample exhibited an increase of HO activity, which was consistent with the changes of HO-1 gene and protein level. Negative responses were observed in the sample treated by LY alone.

Fig. 7
figure 7

Effects of LY83583 and 8-Br-cGMP on HO-1 transcript, HO activity, and HO-1 protein level in GA-treated wheat aleurone layers. Wheat aleurone layers were incubated in a medium containing 5 mM CaCl2 (Con) alone or together with 50 μM GA, 10 μM 8-Br-cGMP (8-Br), 50 μM LY83583 (LY) alone or their combinations. After incubation for 8 h, HO-1 transcript and HO activity (a) and HO-1 protein level (b) were determined. The number below the band (b) indicates the relative abundance of the corresponding HO-1 protein compared with that of Con sample. Coomassie Brilliant Blue-stained gels are present to show that equal amounts of proteins were loaded (c). Data are mean ± SD of three independent experiments with three replicates for each. Within each set of experiments, bars denoted by the different letters are different significantly in comparison with Con treatment at P < 0.05 according to Duncan’s multiple test

Full size image

Expression profiles of GAMYB and PKABA1

GA-regulated Myb transcription factor (GAMYB) is a component of GA signalling in cereal aleurone cells. Molecular analysis showed that GA up-regulates GAMYB transcription in wheat plants (Zhang et al. 2007). The ABA-induced protein kinase PKABA1 is a component of the signal transduction pathway leading to suppressed GA-induced gene expression in cereal grains (Holappa and Walker-Simmons 1995; Gómez-Cadenas et al. 1999; Johnson et al. 2008). To identify the target genes of NO- and cGMP-mediated GA signalling in the synthesis of α-amylase gene expression, we analyzed the inter-relationships among NO, cGMP, HO-1, CO, GAMYB, and PKABA in wheat aleurone layers. As expected, the expression of GAMYB in aleurone layers treated with GA at 4 and 8 h after treatment were significantly higher than in controls (Fig. 8a). There were contrasting changes in the expression profiles of PKABA1 transcripts (Fig. 8b). Similar to the responses of CO, up-regulation of GAMYB and down-regulation of PKABA1 by GA were drastically amplified in aleurone layers cotreated by GA and SNP or 8-Br (in particular). However, although there was a slight increase of GAMYB gene expression in wheat aleurone layers treated with GA plus cPTIO, LY or ZnPPIX, respectively, the significant induction of PKABA1 was found to be modulated after 8 h of various treatments. Especially, the GA plus ZnPPIX rapidly induced the PKABA1 gene as early as 4 h post-treatment. These changes were consistent with previous effects of these treatments in α-Amy2/54 gene expression and secretion of α-amylase (Figs. 2, 3, 4, 5), suggesting that under GA conditions, GAMYB and PKABA1 might be the target genes of NO and cGMP signalling which consequently promoted the synthesis of α-Amy2/54 mRNA.

Fig. 8
figure 8

Expression profiles of GAMYB and PKABA1. Wheat aleurone layers were incubated in a medium containing 5 mM CaCl2 (Con) alone or together with 50 μM ABA, 50 μM GA in the absence or presence of 100 μM SNP, 10 μM 8-Br-cGMP (8-Br), 1 % CO-saturated aqueous solution, 200 μM cPTIO, 50 μM LY83583 (LY), or 100 μM ZnPPIX. After incubation for 4 or 8 h, GAMYB (a) and PKABA1 (b) transcripts were determined. The expression levels of the genes are presented as values relative to those of Con samples. Data are mean ± SD of three independent experiments with three replicates for each. Within each set of experiments, bars denoted by the different letters are different significantly in comparison with Con treatment at P < 0.05 according to Duncan’s multiple test

Full size image

Discussion

It has been well established that endogenous NO and cGMP serve as an important signalling system that modulates plant growth and development (for review, see Delledonne 2005; Leitner et al. 2009). In this report, we discover that HO-1, at least partially in our experimental conditions, plays a vital role in NO and cGMP signalling leading to α-Amy2/54 gene expression in GA-treated wheat aleurone layers.

The first line of evidences supporting this conclusion was that induction of α-amylase activity triggered by GA was amplified in a time-dependent fashion after cotreatment with exogenous NO in the form of the NO-releasing compounds, SNP or NONOate, for a 48-h period, reaching a maximum within 24–48 h of treatment (Fig. 1b). However, the addition of SNP or NONOate alone resulted in identical responses compared to controls. These results were confirmed by the amylase agar-plate assay (Fig. 1a), and the SNP responses were also supported by the significant up-regulation of the most abundant transcripts of α-Amy2/54 (Fig. 2a and Supplementary Fig. S1). By contrast, the above phenomenon was not observed in old SNP solution-treated samples (a negative control) (Supplementary Table S2; Tossi et al. 2009). Moreover, the inducible responses triggered by GA with or without SNP were sensitive to cPTIO, a scavenger of NO, which was confirmed to significantly inhibit NO content in aleurone layers (Fig. 2b). Therefore, these results clearly indicate that maintenance of NO overproduction in GA-treated aleurone layers (Supplementary Fig. S2) was critical for α-Amy2/54 gene expression and thereafter enhanced α-amylase activity; and that alterations of endogenous NO contents is a function fitting well with GA-induced amylase synthesis. Although we cannot exclude the possibility that the chemicals used in the present testing might not be specific for NO, these results collectively illustrated that reestablishment of NO homeostasis in GA-treated aleurone layers was an essential event in triggering α-Amy2/54 gene expression.

Subsequent work combined with our previous results (Wu et al. 2011) showed that increased HO-1 gene expression is one of the earliest responses involved in the signalling pathways triggered by GA in wheat aleurone layers. The time points of maximal induction of HO-1 transcripts and its protein level as well as HO activity induced by GA were preceded by corresponding changes of α-amylase activity in wheat aleurone layers (Figs. 1b, 5), although this inducible response was relatively weak compared to the action of ABA (Wu et al. 2011). These findings are consistent with previous reports showing that auxin, cytokine and ABA could induce HO-1 gene expression in plants (Cao et al. 2007; Xuan et al. 2008; Huang et al. 2011). However, the basic mechanisms related to these aspects are poorly understood and need further research.

Ample animal research has shown that HO, CO and NO not only modulate one another’s function within the cell, but also require each other to impart their effects in most situations (Dulak and Józkowicz 2003). In the present study, remarkably, up-regulation of HO-1 gene expression was also amplified when plants were cotreated by GA and SNP or the HO-1 inducer hemin, both of which were blocked by the addition of ZnPPIX, the specific inhibitor of HO-1 (Figs. 5c–e, 6). Similarly, the SNP- and hemin-triggered up-regulation of HO-1 was still observed in soybean seedlings (Noriega et al. 2007; Santa-Cruz et al. 2010) and wheat aleurone layers (Wu et al. 2011). We then examined whether HO-1 induced by GA or SNP, individually or simultaneously, affected α-Amy2/54 gene expression and thereafter α-amylase activity in aleurone layers. It was shown that exogenous hemin or CO, one of the by-products of HO-1, enhanced early GA-induced α-Amy2/54 gene expression and α-amylase activity in aleurone layers but had no apparent effect in the absence of GA (Fig. 5a, b). These results further confirmed that increases of α-amylase activity induced by HO-1 were GA-dependent. This parallels the situation encountered in the action of exogenous NO-releasing compounds, in which both SNP and NONOate were shown to increase α-amylase activity only in the presence of GA (Fig. 1).

The occurrence of cGMP in plants has been unambiguously demonstrated. Accordingly, several roles of cGMP, including acting as a second messenger of defence gene induction (Durner et al. 1998) and auxin-induced adventitious rooting (Pagnussat et al. 2003), have been confirmed. For example, two exogenous NO-generating systems were able to induce dramatic and transient increases in endogenous cGMP levels in tobacco leaves and suspension cells, which is responsible for the induction of PAL gene expression (Durner et al. 1998). These results clearly suggested that one of the important downstream components of NO signalling in mammals, cGMP, might also be functional in plants. As observed for GA-treated wheat aleurone layers, several experimental arguments suggest that the HO-1 is, at least partly, involved in cGMP-induced α-Amy2/54 gene expression and α-amylase activity.

First, we observed that 8-Br, a cGMP derivative, was able to mimic the amplifying effects of one of the NO-releasing compounds (SNP) on the induction of α-Amy2/54 gene expression, which was also confirmed by the increased α-amylase activity in wheat aleurone layers upon GA treatment (Fig. 3). This finding is consistent with previous reports in barley aleurone layers, showing that cGMP is required for GA-induced GAMYB and α-amylase mRNA accumulation (Penson et al. 1996).

Second, the above cGMP and NO (in particular) responses were respectively suppressed by ZnPPIX (an inhibitor of HO-1) and LY (an inhibitor of NO-inducible GC). These effects were significantly reverted when the CO aqueous solution was also added (Figs. 3, 4).

Third, it is noteworthy that 8-Br promoted the induction of HO-1 mRNA in the presence of GA, which was paralleled with changes of its protein level and HO activity; contrasting phenomena were observed when LY was added together with GA (Fig. 7), all of which were consistent with the up- and down-regulation of α-Amy2/54 gene expression (Fig. 3). This result parallels the situation encountered in animals, in which HO-1 is a cGMP-sensitive endothelial gene and conclusively establishes a causal relationship between HO-1 induction and endothelial protection by the NO and cGMP systems (Polte et al. 2000). Therefore, based on these findings, the induction of HO-1 is clearly responsible for the induction of α-Amy2/54 gene expression afforded by the NO and cGMP systems in GA-treated wheat aleurone layers.

To further confirm the above deduction, we then examined whether NO induced by GA altered the HO-1 gene expression in aleurone layers. As expected (Noriega et al. 2007), the HO-1 transcripts, its protein level and HO activity in aleurone layers during GA treatment were markedly induced after exogenous NO treatment in wheat aleurone layers, all of which were sensitive to the addition of cPTIO and particularly of LY (Fig. 6). In fact, GA increased the endogenous NO levels before the onset of transcription of α-Amy2/54 gene and thereafter α-amylase activity (Fig. 2 and Supplementary Fig. S2; Wu et al. 2011). These results clearly confirm that NO produced by GA was related to the later induction of α-amylase gene expression. Although Arabidopsis aleurone layers responded to NO, GA, and ABA, previous pharmacological evidence showed that NO is upstream of GA in a signalling pathway leading to vacuolation of protein storage vacuoles (Bethke et al. 2007). Above discrepancy may reflect the complexity of GA signalling.

Subsequently, to assess the target molecules of NO and cGMP in HO-1-mediated GA signalling in aleurone layers, we focused on GAMYB and PKABA1, both of which are master-regulating molecules in α-amylase gene expression (Ishibashi et al. 2012). Further molecular evidence showed that, in addition to the inhibition of α-Amy2/54 gene expression and α-amylase activity (Figs. 2, 3, 4), the scavenger of NO cPTIO, the GC inhibitor LY, and the potent inhibitor of HO-1 ZnPPIX completely reversed the induction or inhibition effects of GA on GAMYB or PKABA1 transcription (Fig. 8). Contrasting responses were observed after the application of SNP, 8-Br or CO together with GA, suggesting that NO- and cGMP-mediated HO-1 up-regulation promoted the induction of α-Amy2/54 gene expression by the modulation of GAMYB and PKABA1 mRNA in the presence of GA.

Combined with previous reports (Penson et al. 1996; Wu et al. 2011), the events that followed treatment of wheat aleurone layers with GA are summarized (Fig. 9). The first and rapid response was an overproduction of NO within 1 h of treatment, which might be accompanied by increased cGMP levels. Their concomitant presence explains the induction of HO-1 gene expression at later time periods (4–12 h), thus resulting in the up-regulation of α-Amy2/54 gene expression and thereafter increased α-amylase activity. It is noteworthy that, above events in aleurone layers with the addition of NO and cGMP releasing systems plus GA exhibit a similar but stronger pattern.

Fig. 9
figure 9

Time scale of events following treatment with 50 μM GA. The intensity of the shading of the bars corresponds to the levels of NO, cGMP (Penson et al. 1996), HO activity (also see Wu et al. 2011), and α-amylase activity. The rectangle represents corresponding time courses

Full size image

To our knowledge, the present study provides the first mechanistic description of putative functions of NO and cGMP in wheat aleurone layers exposed to GA. Both NO-releasing compounds and cGMP derivatives displayed amplifying effects on the induction of α-Amy2/54 gene expression and thereafter increased α-amylase activity triggered by GA. We suggest that these effects may be due to the up-regulation of HO-1, which was recently confirmed to act as a vital candidate that can integrate environmental stimuli and phytohormones into various plant developmental processes (Shekhawat and Verma 2010).