Function of type–2 Arabidopsis hemoglobin in the auxin-mediated formation of embryogenic cells during morphogenesis


For correspondence (e-mail


Suppression of Arabidopsis GLB2, a type–2 nonsymbiotic hemoglobin, enhances somatic embryogenesis by increasing auxin production. In the glb2 knock-out line (GLB2/), polarization of PIN1 proteins and auxin maxima occurred at the base of the cotyledons of the zygotic explants, which are the sites of embryogenic tissue formation. These changes were also accompanied by a transcriptional upregulation of WUSCHEL (WUS) and SOMATIC EMBRYOGENESIS RECEPTOR KINASE (SERK1), which are markers of embryogenic competence. The increased auxin levels in the GLB2/ line were ascribed to the induction of several key enzymes of the tryptophan and IAA biosynthetic pathways, including ANTHRANILATE SYNTHASE (α subunit; ASA1), CYTOCHROME P79B2 (CYP79B2) and AMIDASE1 (AMI1). The effects of GLB2 suppression on somatic embryogenesis and IAA synthesis are mediated by increasing levels of nitric oxide (NO) within the embryogenic cells, which repress the expression of the transcription factor MYC2, a well-characterized repressor of the auxin biosynthetic pathway. A model is proposed in which the suppression of GLB2 reduces the degree of NO scavenging by oxyhemoglobin, thereby increasing the cellular NO concentration. The increased levels of NO repress the expression of MYC2, relieving the inhibition of IAA synthesis and increasing cellular IAA, which is the inductive signal promoting embryogenic competence. Besides providing a model for the induction phase of embryogenesis in vitro, these studies propose previously undescribed functions for plant hemoglobins.


Hemoglobins are important heme-containing proteins that are ubiquitous in nature. Initially described in animals, where they perform a number of functions associated with their ability to bind oxygen and other small gaseous ligands, hemoglobins have been characterized in a variety of organisms, including bacteria, fungi and plants. The first hemoglobin isolated from plants was leghemoglobin, commonly found in nodules of nitrogen-fixing species, which facilitates the movement of oxygen in the nodules and prevents the inactivation of nitrogenase (reviewed in Smagghe et al., 2009). The presence of a novel hemoglobin in a non-symbiotic plant species, Trema tormentosa (Bogusz et al., 1988), suggested the existence of plant hemoglobins that had additional functions beyond their involvement in symbiotic relationships. Although the major role assigned to non-symbiotic hemoglobins is to scavenge nitric oxide (NO), especially in events related to hypoxia, recent studies suggest an apparent wider function of hemoglobins, especially during developmental processes ranging from seed storage product accumulation to in vivo and in vitro shoot formation (reviewed in Hill, 2012).

Three distinct classes of plant hemoglobins have been identified based on their oxygen binding ability, phylogenetic characteristics and expression patterns (Hunt et al., 2001). Although class–3 hemoglobins show structural similarities with the truncated bacterial globins (Watts et al., 2001), class–1 and -2 hemoglobins are characterized by a three-on-three α–helical loop surrounding the hemoglobin moiety, which is a distinct feature of vertebrate globins (Hargrove et al., 2000). Differences between class–1 and -2 hemoglobins are apparent in their oxygen binding affinities, with an approximate Michaelis constant, Km, of 2 nm for members of class 1 and 150 nm for members of class 2 (Hoy and Hargrove, 2008; Dordas, 2009). Because of the very high affinity for oxygen, Hill (1998) discounted any involvement of class–1 hemoglobins in oxygen carrying and sensing functions. The Arabidopsis genome contains three hemoglobin genes: GLB1, GLB2 and GLB3, each belonging to their respective class and with unique functions. Like other class–1 hemoglobins (Dordas et al., 2003, 2004), GLB1 scavenges NO produced under severe hypoxia, thus fulfilling a protective role during stress conditions (Perazzolli et al., 2004). GLB1 is induced rapidly in Arabidopsis roots exposed to reduced oxygen levels, and the ectopic upregulation of GLB1 reduces cellular NO, increasing plant survival to hypoxic conditions and contributing to the maintenance of cellular energy status (Hunt et al., 2002). These effects are also shared by class–1 hemoglobins from other species, as demonstrated in Medicago sativa (alfalfa), where the overexpression of a class–1 Hordeum vulgare (barley) hemoglobin improves the growth of hypoxic roots by reducing aerenchyma formation and enhancing growth and energy charge through NO scavenging mechanisms (Dordas et al., 2003). In the same study it was also shown that the suppression of barley hemoglobin reduced the ability of the plants to adapt to hypoxia and resulted in a heavy accumulation of NO.

Relatively little is known about the function of class–2 hemoglobins. GLB2 overexpression increases plant survival in a similar fashion to GLB1 overexpression (Hebelstrup et al., 2006), and it scavenges cellular NO (Hebelstrup and Jensen, 2008; Hebelstrup et al., 2012); however, the induction of GLB2 by low temperature (Trevaskis et al., 1997) or cytokinin (Hunt et al., 2001), conditions that do not affect GLB1 expression, suggests distinct functions for the two genes. Expression pattern analyses revealed the preferential expression of GLB2 in developing organs, such as immature seeds and fruits, young leaves and somatic embryos (Hendriks et al., 1998; Hunt et al., 2002; Wang et al., 2003), suggesting the gene might be required in tissues with high energy demand. Vigeolas et al. (2011) observed an elevated energy state followed by increased oil accumulation in Arabidopsis seeds overexpressing GLB2, which would support this view. No information is currently available on the function of GLB3.

Emerging evidence indicates that non-symbiotic hemoglobins are also expressed under normoxic conditions, thus suggesting a potential role in developmental processes (Hill, 2012). Profound phenotypic abnormalities during post-embryonic growth were induced by the altered expression of both GLB1 and GLB2 (Hebelstrup et al., 2006). Whereas increasing levels of GLB1 and GLB2 affect meristem function by accelerating the transition of vegetative meristems into inflorescence meristems, the silencing of GLB1 delays the flowering transition and causes the formation of aereal rosettes at lateral meristems (Hebelstrup and Jensen, 2008). The effects of hemoglobin on meristem function were further analysed during shoot organogenesis, which was encouraged by the ectopic expression of GLB1 and GLB2, possibly through the induction of genes participating in cytokinin perception and signalling (Wang et al., 2011). Taken together these preliminary observations suggest potential roles of hemoglobins during morphogenesis.

The execution of morphogenic events is best exemplified during somatic embryogenesis: i.e. in the formation of embryos from cultured somatic cells. This process is influenced by changes in the environment and culture conditions, and is generally favored by mild hypoxia (Thorpe and Stasolla, 2001), which mimics the low oxygen environment accompanying zygotic embryo development (Rolletschek et al., 2003). Besides producing a large number of embryos, somatic embryogenesis is a good ‘proof of concept’ model to examine the early induction phases associated with the somatic–embryonic transition (Mordhorst et al., 2002; Su et al., 2009). In Arabidopsis, somatic embryogenesis is a two-step process initiated by culturing early cotyledonary zygotic embryos on an induction medium containing the auxin 2,4–dichlorophenoxyacetic acid (2,4–D), which is the inductive signal required for the formation of embryogenic cells (Raghavan, 2004). Removal of this growth regulator evokes the development of somatic embryos (Bassuner et al., 2007). In an effort to assess the role played by hemoglobins during Arabidopsis somatic embryogenesis, we induced somatic embryos in lines with altered hemoglobin levels. These included lines ectopically expressing the three hemoglobin genes (lines 35S::GLB1, 35S::GLB2 and 35S::GLB3), as well as two homozygous lines in which the genes were either knocked out (lines GLB2−/− and GLB3−/−) or downregulated via RNAi (line GLB1-RNAi). Transcriptional and structural analyses were also conducted to elucidate the differential response of the lines in culture. Our results reveal a role of GLB2 in mediating somatic embryogenesis, and identify this gene as a potential modulator of auxin synthesis and transport.


Somatic embryogenesis is affected by altered expression of hemoglobin genes

Somatic embryogenesis in Arabidopsis is a two-step process requiring an induction (I) phase (14 days) in which the embryogenic tissue is induced from the adaxial sides of the cotyledons in the presence of 2,4–D, followed by a developmental (D) phase (9 days) that culminates with the formation of the somatic embryos (Figure 1a). Lines with altered levels of GLB1, GLB2 and GLB3, fully characterized in previous studies (Hebelstrup et al., 2006; Hebelstrup and Jensen, 2008; Wang et al., 2011), were used to induce somatic embryogenesis. Whereas the ectopic expression of the three hemoglobin genes (lines 35S::GLB1, 35S::GLB2 and 35S::GLB3) and the knock-out of GLB3 (line GLB3−/−) did not have any pronounced effect on the number of embryos produced in the culture, the suppression of GLB1 (line GLB1-RNAi) inhibited embryo formation (Figure 1b). This was in contrast to the GLB2 knock-out line (GLB2−/−), which showed a threefold increase in the production of somatic embryos (Figure 1b,c). This line was used for further experiments in order to examine the mechanisms underlining the beneficial role of GLB2 suppression on somatic embryogenesis.

Figure 1.

Effects of altered hemoglobin expression during somatic embryogenesis. (a) Somatic embryogenesis in Arabidopsis. Zygotic embryos are cultured for 14 days on induction (I) medium containing 2,4–D, which allows the formation of embryogenic tissue from the base of the cotyledons. Growth of somatic embryos is encouraged on development (D) medium devoid of growth regulators.
(b) Somatic embryo number in lines ectopically expressing the three hemoglobin genes (lines 35S::GLB1, 35S::GLB2 and 35S::GLB3), and lines in which the genes were either knocked out (lines GLB2−/− and GLB3−/−) or downregulated via RNAi (line GLB1-RNAi). Values ± SEs are means of at least three biological replicates using 80 explants per replicate. *Statistically significant differences (P < 0.05) from the wild type (WT).
(c) External morphology of somatic embryos produced by WT explants (left panel) and GLB2−/− explants (right panel).

During the embryogenic process transcript levels of GLB2 increased significantly in induction (I) medium containing 2,4–D, before declining during the subsequent developmental (D) period (Figure 2a). Localization of GLB2 was also investigated using a GLB2::GUS reporter line (Heckmann et al., 2006). At the beginning of the culture period [I(0)] no GUS staining was detected in the zygotic embryos (Figure 2b1), whereas after 3 days on induction medium GLB2 was highly induced along the embryonic axis (Figure 2b2). Both the shoot meristem and adaxial base of the cotyledons, which is the site of embryogenic tissue formation (Raghavan, 2004), expressed GLB2 at day 7 on induction medium (Figure 2b3). At the end of the induction period (day 14), GLB2 was mainly localized in the apical part, corresponding with the forming embryogenic tissue (arrows in Figure 2 B4). No GUS expression was detected during later stages of development (Figure 2b5). The activation of GLB2 on the induction medium was only a result of the presence of 2,4–D, as the expression of this gene did not increase in explants germinated on a medium identical to that used for induction but devoid of 2,4–D (Figure 2c,d).

Figure 2.

Expression and localization of GLB2. (a) Relative expression of GLB2 during the different phases of somatic embryogenesis in the wild type (WT). Values ± SEs are means of at least three biological replicates and are normalized to the value of I(0), set to 1. *Statistically significant differences (P < 0.05) from the wild type (WT). I, induction medium; D, development medium. Days in culture are in brackets.
(b) Localization patterns of GLB2 by GUS staining during different phases of induction (1, day 0; 2, day 3; 3, day 7; 4, day 14) and development (5, day 3).
(c) Expression levels of GLB2 during germination of the explants in a medium identical to the induction medium but devoid of 2,4–D.
(d) Localization of GLB2 by GUS staining at day 0 (1) and day 6 (2) in germinating explants in the absence of 2,4–D.

Nitric oxide (NO) involvement during somatic embryo formation

The documented ability of GLB2 to scavenge NO (Hebelstrup and Jensen, 2008) prompted us to investigate whether repression of GLB2 increased NO levels along the adaxial sites of the cotyledons during the formation of embryogenic tissue. From 7 days on induction medium a preferential accumulation of NO was detected in the GLB2−/− line (Figure 3a). In the same line NO accumulation was reduced by 2–phenyl-4,4,5,5-tetramethylimidazoline-1-oxyl-3-oxide (PTIO). The role of NO during somatic embryogenesis was further analysed by manipulating cellular NO through exogenous application of NO scavengers PTIO and 2–(4-carboxyphenyl)-4,4,5,5-tetramethylimidazoline-1-oxyl 3 oxide (cPTIO), and donors sodium nitroprusside (SNP) and S-nitroso-N-acetyl-d,l-penicillamine (SNAP). Embryo production appears susceptible to NO levels, as it was increased in the presence of SNP and SNAP, and was decreased in the presence of PTIO and cPTIO (Figure 3). A 52% decrease in embryo production was also observed in the nitric oxide synthase 1 (nos1) mutant line, which is defective in NO production (Guo et al., 2003). Embryo production from explants suppressing GLB2 was also repressed by PTIO, thus confirming the requirement of NO for the enhancement of somatic embryogenesis observed in the GLB2−/− line (Figure 3c).

Figure 3.

(a) Localization of NO at the base of the cotyledons (sites of embryogenic tissue formation) of wild-type (WT) explants (left panel), GLB2−/− explants (middle panel) and GLB2−/− explants treated with cPTIO (right panel) and cultured for 7 days on induction medium.
(b) Somatic embryo formation in WT explants cultured in the presence of NO scavenging compounds (PTIO or cPTIO) or NO donor compounds (SNP or SNAP). Values ± SEs are means of at least three biological replicates using 80 explants per replicate. *Statistically significant differences (P < 0.05) from the wild type (WT + H2O). As a control, explants were cultured with solvents used to dissolve the NO-scavenging or -releasing compounds.
(c) Effects of PTIO supplementations on somatic embryo production of the GLB2−/− line.

GLB2 affects auxin biosynthesis during the induction phase of somatic embryogenesis

The beneficial effect of suppressed GLB2 on embryo production was further examined by investigating the synthesis and transport of natural auxin, IAA, which acts as the inductive signal for the formation of embryogenic tissue (Raghavan, 2004). In Arabidopsis, the biosynthesis of IAA mainly occurs through the tryptophan pathway (reviewed by Zhao, 2010), which is negatively regulated by MYC2, a basic helix-loop-helix (bHLH) domain-containing transcription factor (Dombrecht et al., 2007). During the second half of the induction phase the expression of MYC2 was significantly reduced in the GLB2−/− line (Figure 4a). The accumulation of MYC2 transcripts was also dependent on NO levels. The inclusion of the NO donor SNP repressed MYC2 at the end of the induction medium, whereas an upregulation of MYC2 occurred in tissue cultured with the NO scavenger PTIO (Figure 4b). Of note, somatic embryo formation was favored in the myc2 mutant line, whereas it was repressed in the line constitutively expressing MYC2 (35S:MYC2), and also in lines overexpressing MYC2 in a GLB2 mutant background (GLB2−/− 35S:MYC2) (Figure 4c).

Figure 4.

(a) Expression level of MYC2 during the induction (I) phase of somatic embryogenesis in the wild type (WT) and in the GLB2−/− line. Values ± SEs are means of at least three biological replicates, and are normalized to the value of WT at I (day 0), set to 1. *Statistically significant differences (P < 0.05) from the WT at the same day in culture.
(b) Effects of the NO donor SNP or NO scavenger PTIO on MYC2 expression during the induction phase of the WT line. Values ± SEs are means of at least three biological replicates and are normalized to the value of untreated control (C) tissue at I (day 7), set to 1. *Statistically significant differences (P < 0.05) from C at the same day in culture.
(c) Somatic embryo number in the WT line, the myc2 mutant line, the 35S MYC2 line and the GLB2−/− line overexpressing MYC2. Values ± SEs are means of at least three biological replicates using 80 explants per replicate. *Statistically significant differences (P < 0.05) from the WT.

Collectively, these results suggest that the enhancement of embryo formation observed in the GLB2−/− line might be the result of reduced MYC2 expression, in a response mediated by NO.

Transcriptional analysis of the tryptophan-dependent auxin biosynthetic pathway revealed precise changes in the expression of key enzymes evoked by the repression of GLB2. Two early enzymes of tryptophan synthesis, ASA1 (anthranilate synthase α–subunit) and PAI3 (anthranilate isomerase), are induced in the GLB2−/− line at days 14 and 7, respectively, on induction medium (Figure 5). A similar upregulation at the beginning of the culture medium [I(0)] was observed for TSA1 (tryptophan synthase α–subunit). Several enzymes participating in the later steps of tryptophan and IAA synthesis were also upregulated in the GLB2−/− line. These included TSB1 (tryptophan synthase β–subunit), YUC4 (yucca4), CYP79B2 (cytochrome P450 CYP79B2) and AMI1 (amidase1) (Figure 5). The requirement for AMI1, the last IAA biosynthetic enzyme converting indole-3-acetamide to IAA (Pollmann et al., 2003) in the GLB2-mediated embryogenic response, was further confirmed by the reduction in embryo number observed in the GLB2−/− ami1 double mutant lines (Table 1). Many of the genes upregulated in GLB2−/− tissue were responsive to NO levels, as they were induced by the NO donor SNP and repressed by the NO scavenger PTIO (Figure 6). The transcriptional activation of the IAA biosynthetic pathway in the GLB2−/− line was accompanied by increased endogenous levels of IAA (Figure 7a) and a reduced requirement of exogenous auxin to sustain elevated somatic embryo production (Figure 7b). Immunocalization analyses revealed elevated levels of IAA on the adaxial side of the cotyledons (the site of embryogenic tissue formation) of the GLB2−/− explants cultured on induction medium (Figure 7c).

Table 1. Percentage decrease of Arabidopsis somatic embryos produced by the ami1 mutant lines (SALK_069970 and CS875054) and the GLB2−/− ami1 double mutant lines, compared with the GLB2−/− line (set at 100%)
Line% decrease
  1. Values ± SEs are the mean of three biological replicates.

ami1 (SALK_069970)72 ± 6
ami1 (CS875054)62 ± 6
GLB2−/− ami1 (SALK_069970)58 ± 12
GLB2−/− ami1 (CS875054)69 ± 9
Figure 5.

Expression levels of genes participating in tryptophan and IAA biosynthesis during the 14–day induction (I) period of somatic embryogenesis in the wild type (WT) and in the GLB2−/− line. Values ± SEs are means of at least three biological replicates and are normalized to the value of the WT at I (day 0), set to 1. *Statistically significant differences (P < 0.05) from the WT at the same day in culture. Abbreviations: ASA1, anthranilate synthase (α–subunit); PAI3, anthranilate isomerase; IGPS, indole-3-glycerolphosphate synthase; TSA1, tryptophan synthase (α subunit); TSB1, tryptophan synthase (β–subunit); TAA1, tryptophan amino transferase; CYP79B2, cytochrome P79B2; YUC4, YUCCA4; AMI1, amidase1.

Figure 6.

Effects of applications of the NO donor SNP and/or NO scavenger PTIO on the expression levels of selected genes (see Figure 5 for gene names) involved in tryptophan and IAA synthesis during the induction (I) phase of wild-type (WT) somatic embryogenesis. Values ± SEs are means of at least three biological replicates, and are normalized to the value of untreated control tissue (C) at I (day 7), set to 1. *Statistically significant differences (P < 0.05) from C at the same day in culture.

Figure 7.

(a) Endogenous IAA levels in the wild-type (WT) and GLB2−/− explants cultured for 14 days on induction medium. Values ± SEs are means of at least three biological replicates. *Statistically significant differences (P < 0.05) from the WT.
(b) Somatic embryogenesis in the WT and GLB2−/− lines, supplemented with different levels of 2,4–D in the induction medium.
(c) Immunolocalization of IAA at the base of the cotyledons of WT (1) or GLB2−/− (2) explants after 7 days on induction medium. Primary antibody was omitted from control sections (3).

GLB2 affects auxin transport during the induction phase of somatic embryogenesis

Expressions of the auxin-efflux transporters PIN1 and PIN2 were measured on different days during the induction phase of somatic embryogenesis. Compared with wild-type (WT) tissue, the repression of GLB2 increased the expression of PIN1 from day 7 on the induction medium, reaching a maximum value at day 14 (Figure 8a). This was in contrast to the expression of PIN2, which was not altered in the GLB2−/− line (Figure 8b). We also analyzed the distribution of both PIN1 and PIN2 proteins along the cotyledons of the explants, from which embryogenic tissue generates during exposure to 2,4–D. The PIN1 signal was apparent on the adaxial side of the cotyledons of WT embryos from early days on induction medium (Figure 8c1). Visible signs of polar localization were only detected by the end of the induction period (day 14; Figure 8c2). The localization of PIN1 in the WT suggested the establishment of an auxin flow directed towards the apical region of the embryogenic tissue (Figure 8c3). In the GLB2−/− line, polarization of PIN1, which is indicative of polar transport of auxin from the shoot meristem along the adaxial side of the cotyledons, was observed at day 7 on induction medium (Figure 8c4). Furthermore, polar PIN1 signal in the same line occurred in distinct domains of the embryogenic tissue (Figure 8c5), which are possible sites of somatic embryo formation. A longitudinal view of these domains suggests a PIN1 distribution favoring the acropetal movement of auxin (Figure 8c6). Heavy accumulation of PIN1 was detected in emerging GLB2−/− somatic pro-embryos (Figure 8c7).

Figure 8.

Expression levels of PIN1 (a) and PIN2 (b) in wild-type (WT) and GLB2−/− explants cultured on induction (I) medium. Values ± SEs are means of at least three biological replicates and are normalized to the value of WT at I (day 0), set to 1. *Statistically significant differences (P < 0.05) from the WT at the same day in culture.
(c) Immunolocalization of PIN1 in WT (1–3) and GLB2−/− explants (4–7) at day 3 (1), day 7 (4) and day 14 (2, 3 and 5–7) of induction.
(d) Immunolocalization of PIN2 in WT explants (1, day 3; 2, day 7; 3, day 14) and GLB2−/− explants (4, day 3; 5, day 7; 6, day 14) of induction. Protein polarization (arrowhead): predicts auxin flow.

In WT tissue no PIN2 polarization was detected during the first days in culture (Figure 8d1), and occurred only after day 7 on induction medium (Figure 8d2). Like PIN1, the distribution of PIN2 was suggestive of an acropetal movement of auxin towards the apical regions of the emerging embryogenic tissue. Embryogenic masses exhibited a strong PIN2 signal (Figure 8d3). No apparent differences in PIN2 localization patterns were observed between the WT (Figure 8d1–3) and the GLB2−/− line (Figure 8d4–6).


Plant cells exhibit remarkable plasticity, allowing them to de-differentiate and embark on new developmental pathways leading to the production of tissues, organs or new plants. This characteristic is best exemplified in culture, where somatic cells can be induced to produce embryos through a process referred to as somatic embryogenesis. In Arabidopsis, somatic embryo production requires two distinct steps: an induction phase needed for the acquisition of embryogenic competence, and a developmental phase in which embryogenic cells form embryos (Bassuner et al., 2007). Whereas the developmental phase occurs in the absence of growth regulators, the induction phase requires auxin, which acts as the inductive signal promoting the de-differentiation of somatic cells (Raghavan, 2004). PIN1-mediated auxin maxima in clusters of cells at the base of the cotyledons of the zygotic explants have been described as the early and necessary events for the formation of embryogenic tissue, and ultimately for the production of somatic embryos (Su et al., 2009). Suppression of the Arabidopsis non-symbiotic hemoglobin, GLB2, increases the production of somatic embryos (Figure 1b), elevates auxin levels (Figure 7a) and decreases the concentration of 2,4–D in the induction medium required to achieve significant somatic embryo production (Figure 7b). In the glb2 knock-out line (GLB2−/−), auxin maxima and early polarization of PIN1 were detected at the base of the cotyledons, where embryogenic cells originate during induction (Figure 7c). Explants of this line also had increased levels of AtWUS and AtSERK1 (Figure S1). These two genes, the expression levels of which are elevated by auxin (Hecht et al., 2001; Su et al., 2009), are often used as markers of embryogenic competence, and their ectopic expression triggers the vegetative-to-embryogenic transition (reviewed by Karami et al., 2009). The effect of GLB2 suppression on auxin levels was also confirmed by transcriptional studies showing the activation of key enzymes of tryptophan and auxin biosynthesis during the induction period of GLB2−/− explants (Figure 5). Key enzymes induced by the suppression of GLB2 are ASA1, regulating the first committed step of tryptophan biosynthesis (Bender and Fink, 1998), and CYP79B2 and AMI1, with the former being responsible for diverting tryptophan into IAA production (Zhao et al., 2002), and the latter participating in the last step of IAA synthesis (Mano et al., 2010). Overexpression of either CYP79B2 or AMI1 is sufficient to elevate IAA levels (Zhao et al., 2002; Mano et al., 2010), and the abolishment of AMI1 function in the GLB2−/− line compromises embryo formation (Table 1). Collectively, these results demonstrate that somatic embryogenesis in Arabidopsis is enhanced by the suppression of GLB2 as a result of the increased production of inductive signal IAA at the presumptive sites of embryogenic tissue and somatic embryo formation.

During embryonic and post-embryonic growth, auxin synthesis is influenced by a multitude of factors, including morphogenic and developmental events (reviewed by Zhao, 2010). The possible role of hemoglobins, including GLB2, in affecting auxin is limited, as the molecules are known only to bind small gaseous ligands and perform a few oxidative reactions (reviewed by Smagghe et al., 2009). A potential point of functional intersection of auxin and hemoglobins is NO (reviewed by Hill, 2012), which is scavenged by GLB2 (Hebelstrup and Jensen, 2008; Hebelstrup et al., 2012).

Fluctuation of endogenous NO levels has pronounced implications on growth and development. Emerging evidence has demonstrated the role played by NO during plant defense mechanisms (Delledonne et al., 1998), abiotic stress responses (Mackerness et al., 2001; Gould et al., 2003) and morphogenic responses mediated by plant growth regulators (Leshem and Haramaty, 1996; Beligni and Lamattina, 2000; Seregelyes et al., 2003). Specifically, NO and auxin have been linked in a variety of processes ranging from root hair development, lateral and adventitious root formation, root gravitrophic bending and root responses to iron deficiency (reviewed by Hill, 2012). Hemoglobins, which are known to oxygenate NO, forming nitrate, have also been implicated in affecting some of these responses. The present study demonstrates that GLB2 (Figure 2b), NO (Figure 3a) and auxin (Figure 7c) appear in the same region of the cotyledons where embryogenic cells develop. Furthermore, the NO scavengers PTIO and cPTIO suppress embryo formation in both WT (Figure 3b) and GBL2−/− lines (Figure 3c), whereas the NO donors SNP and SNAP, added to the induction medium, substantially enhance embryo formation in WT tissue (Figure 2b). From this it can be concluded that changes in NO availability during the induction period influence the development of embryogenic tissue, as also reported in alfalfa cultured cells (Otvos et al., 2005), and that the suppression of GLB2 increases the level of NO in regions of the cotyledon where embryogenic tissue and somatic embryos develop. Furthermore, the regulation of several tryptophan and IAA biosynthetic genes by the application of SNP or PTIO (Figure 6) supports the notion that the increase in IAA level observed in the GLB2−/− line is mediated by NO.

To determine whether auxin and NO increase in response to hemoglobin suppression as part of the same mechanism influencing the development of embryogenic cells, we have examined the expression of the bHLH transcription factor, MYC2, which negatively regulates auxin expression (Dombrecht et al., 2007). MYC2 transcript levels were significantly lower in the GLB2−/− line during the induction period (Figure 4a), and although repression of MYC2 has a promoting effect on somatic embryogenesis, overexpression of MYC2 reduces embryo number (Figure 4c). Thus, the suppression of GLB2, without externally altering NO availability, decreases MYC2. These same conditions increased the expression of auxin biosynthesis genes (Figure 5), and resulted in the increased detection of NO in regions of the cotyledon where embryogenic cells develop (Figure 3a). The addition of the NO donor SNP to the induction medium also decreased MYC2 expression (Figure 4b) and enhanced the embryogenic process (Figure 3b). Conversely, the introduction of the NO scavenger PTIO elevated MYC2 expression (Figure 4b) and reduced the number of embryos formed in culture (Figure 3b). The data fit a model (Figure 9) in which the suppression of GLB2 reduces the degree of NO scavenging, thereby increasing the cellular NO concentration. The increased NO levels repress the synthesis of MYC2, relieving the inhibition of IAA synthesis and improving the embryogenic competence of the cells.

Figure 9.

Suggested model summarizing the interaction among GLB2, NO and IAA during the induction phase of somatic embryogenesis in Arabidopsis.

This work suggests that class–2 non-symbiotic hemoglobins play a role in regulating the synthesis and transport of auxins by altering the level of the signal molecule, NO, in specific cells.

Experimental procedures

Plant material and treatments

Arabidopsis (ecotype Col–0) seeds in which the hemoglobin genes were overexpressed (lines 35S::GLB1, 35S::GLB2 and 35S::GLB3), downregulated by RNAi (line GLB1-RNAi) or knocked out (GLB2−/−) were generated and characterized as previously described (Hebelstrup et al., 2006; Hebelstrup and Jensen, 2008). The knock-out line GLB3−/− was obtained from the SALK collection of T–DNA insertional mutants (Alonso et al., 2003), and was characterized in previous work (Wang et al., 2011). The GLB2::GUS reporter line was generated by Heckmann et al. (2006). The myc2 mutant line and the MYC2 overexpressing line (35S:MYC2, kindly provided by Prof. Kazan) were characterized previously (Dombrecht et al., 2007). GLB2−/− ami1 double mutant lines were generated by crossing, using two different AMI1 T–DNA insertion lines (SALK_069970 and CS822368) in which the amidase 1 gene (At1G08980) was disrupted (Figure S2).

For the constitutive expression of MYC2 in the GLB2−/− line, full-length cDNA of Arabidopsis MYC2 (AT1G32640) was inserted in the Gateway entry clone vector pDONR_221 (Invitrogen, according to the instruction manual, and then transferred into the pEarley 100 final destination vector carrying the 35S promoter (Earley et al., 2006). The construct was introduced into Agrobacterium tumefaciens strain GV3101, which was used to spray transform the GLB2−/− line according to the method described by Elhiti et al. (2010). Transformed plants were screened by quantitative RT-PCR (Figure S3).

Arabidopsis somatic embryogenesis was induced by culturing bent-cotyledon zygotic embryos on induction medium containing 2,4–D for 14 days, followed by a transfer onto a hormone-free development medium in which somatic embryos developed (Bassuner et al., 2007). The NO scavengers cPTIO and PTIO were dissolved in water and applied at a concentration of 100 μm. The NO donors SNP and SNAP were dissolved in 30% ethanol and DMSO, respectively, and applied at a concentrations of 10 μm. Applications were performed by dispensing 10 μl of solution directly on the explants every other day in culture.

NO, IAA, and PIN localization and measurements

Nitric oxide was assayed histochemically with 4,5,-diaminofluorescein diacetate (DAF–2 DA), which has been reported to be effective in localizing NO in various Arabidopsis tissues (Hebelstrup et al., 2008). Explants were incubated in a buffer [10 mm 2–(N–morpholino)ethanesulfonic acid (MES)-Tris, pH 5.6, 0.1 mm CaCl2 and 10 mm KCl] containing 10 mm DAF–2 DA for 2 h, washed twice in the same buffer and sectioned using a vibratome. Sections (50 mm) were used to visualize NO using an Olympus FV500 multilaser confocal microscope with fluorescein iso-thiocyanate filter (excitation 495 nm; emission 515 nm) and fluorescence was captured using fluoview 4.3.

Immunolocalization of endogenous IAA was carried out as described by Thomas et al. (2002), with some minor modifications. Plant material was first pre-fixed in freshly prepared 4% aqueous 1–ethyl-3-(3-dimethyl-aminopropyl)-carbodiimide hydrochloride at 4°C for 2 h, and then post-fixed in FAA (10% formalin, 5% acetic acid and 50% ethanol) overnight at 4°C. The fixed tissue was dehydrated in an ethanol series, embedded in paraplast, sectioned (10 μm) and deparaffinised in xylene. The sections were incubated in blocking solution [full-strength phosphate-buffered saline (PBS) solution, pH 7, 0.1% Tween 20, 1.5% glycine and 5% bovine serum albumin (BSA)] at room temperature (22°C) for 1 h. A 150–μl portion of monoclonal primary IAA antibodies (1 mg ml−1; Sigma-Aldrich,, diluted 1:200 in 10 mm PBS containing 0.8% BSA, was applied to the sections and incubated in a high-humidity chamber for 4 h at room temperature. The slides were washed first in 10 mm PBS containing 0.88 g L−1 NaCl, 0.1% Tween 20 and 0.8% BSA for 5 min, and then in 10 mm PBS with 0.8% BSA for 5 min, in order to remove any excess Tween 20. The slides were incubated in 200 μl of secondary antibodies [anti-mouse IgG alkaline phosphatase conjugate (1 mg ml−1); Promega,] overnight in a high-humidity chamber, washed twice in full-strength PBS containing 0.88 g L−1 NaCl, 0.1% Tween 20 and 0.8% BSA for 10 min, and then incubated in water for 15 min to remove the excess of secondary antibodies. Samples were stained using 250 μl of Western blue (Promega) for 40 min.

For IAA measurements, explants cultured for 14 days on induction medium were freeze-dried and IAA analysis was performed at the National Research Council of Canada – Plant Biotechnology Institute, Saskatoon, Canada ( by high-performance liquid chromatography electrospray tandem mass spectrometry (HPLC-ES-MS/MS) using deuterated internal standards, as described previously (Chiwocha et al., 2003). IAA measurements earlier (before 14 days) in the culture were impossible to perform because of the minute size of the tissue (zygotic embryo explants).

Immunostaining of PIN proteins was performed on wax-embedded (9:1, PEG400 distearate:1–hexadecanol 99%; Sigma-Aldrich) explants, exactly as described by Carraro et al. (2006). Briefly, the tissue was fixed in a 4% paraformaldehyde, full-strength PBS solution for 1 h, under vacuum conditions at room temperature. Embedded tissue was sectioned (10 μm) and incubated overnight with 1:200 anti-PIN1 or anti-PIN2 (Santa Cruz Biotechnology Inc., in full-strength PBS, containing 1% BSA. The sections were then washed twice for 10 min in full-strength PBS and stained with the secondary antibodies conjugated with Alexa488 (Molecular Probes, now Invitrogen, The slides were washed twice for 10 min with full-strength PBS and visualized with fluorescence microscopy.

Gene expression analysis by quantitative RT-PCR

Extraction of RNA, synthesis of cDNA and analysis of gene expression by quantitative RT-PCR were performed exactly as described in Elhiti et al. (2010). All primers used are listed in Table S1. The relative level of gene expression was analyzed with the 2−∆∆Ct method described by Livak and Schmittgen (2001) using actin (AY139999) as a reference.

Statistical analysis

Unless specified, all experiments were performed using at least three biological replicates, and Tukey's post-hoc test for multiple variance was used to compare differences among samples (Zar, 1999).


This work was supported by NSERC Discovery Grants to R.D.H. and C.S., and by The Danish Research Council for Independent Research | Technology and Production Sciences, to K.H.H. The authors thank the Manitoba Institute of Cell Biology Confocal Imaging facility for technical assistance with image capture.