Boron and calcium induce major changes in gene expression during legume nodule organogenesis. Does boron have a role in signalling?


  • Dedication: The authors wish to dedicate this investigation to Dr Adam Kondorosi who has been a beacon to us all in both science and life.

The requirement of trace amounts of boron (B) for plant growth was reported as early as 1923 (Warington, 1923) and over several decades this micronutrient has been related to most biochemical, physiological or anatomical plant processes. Cell wall stability through borate cross-linking of apiosyl residues of the pectin polysaccharide rhamnogalacturonan II (RGII) is the most convincingly demonstrated role of B (O’Neill et al., 2004). Moreover, there is increasing evidence that B is involved in the control of expression of some genes that are important for plant metabolism (Beato et al., 2010) or cell differentiation and organogenesis (Lowatt, 1985; Reguera et al., 2009). Furthermore, the ability of boric acid to inhibit the growth of cell lines of prostate and breast cancer (Meacham et al., 2007) or abnormal embryonic development in animals under low-B conditions (Rowe & Eckhert, 1999) points to a function of the micronutrient that is related to signalling mechanisms and to the control of gene expression.

Further progress on this research topic will depend on new methodologies with greater analytical capability and on the use of biological models that are highly dependent on the micronutrient. Such a model is available in relation to the symbiotic interaction between legumes and nitrogen (N2)-fixing rhizobia, which triggers the organogenesis of a new root organ, the nodule, following a highly regulated process (Foucher & Kondorosi, 2000; Oldroyd & Downie, 2008). The high sensitivity of nodule development to B deficiency is in line with there being a higher B content in nodules than in other plant tissues (Redondo-Nieto et al., 2003). Symptoms of B deficiency appear in nodules earlier than in other plant organs that can develop using B storage in seed (Bolaños et al., 1994). This supports the utility of nodule organogenesis as a model for B-nutrition studies. Indeed, 3 wk after inoculation with rhizobia, B-deficient low functional nodules led to plants suffering from N deficiency. The reduction of the shoot and root FW of those plants (which was c. 40% lower than B-sufficient controls) was attributed mainly to a lack of N rather than to an effect of B deficiency on root and shoot growth (Bolaños et al., 1994).

Typical symptoms of B deficiency appear in the structure of symbiotic nodules with irregular walls as a result of an abnormal assembly of components, such as pectins and proline- and hydroxyproline-rich glycoproteins (Bonilla et al., 1997a,b). In addition, B deficiency affects each stage of the symbiosis (summarized in Table 1): early plant bacteria recognition (Redondo-Nieto et al., 2001), the infection process (Bolaños et al., 1996; Reguera et al., 2010a), symbiosome development (Bolaños et al., 2001; Redondo-Nieto et al., 2007) and cell differentiation during nodule organogenesis (Reguera et al., 2009). Interestingly, several of those B deficiency symptoms, mainly early preinfection and infection events or nodule organogenesis, but not symbiosome development, could be partially prevented by addition of Ca2+ (Redondo-Nieto et al., 2003).

Table 1.   Main effects of boron (B) deficiency (–B) and the supplement of Ca2+ (−B + 2Ca2+) on the development of the legume–rhizobia symbiotic interaction
Symbiotic event−B−B + 2Ca2+ (Redondo-Nieto et al., 2003)
 nod gene activity (Redondo-Nieto et al., 2001)InhibitedRestored
 Root colonization (Redondo-Nieto et al., 2001)DecreasedRestored
Rhizobial infection
 Infection thread development (Redondo-Nieto et al., 2001Reguera et al., 2010b,c)Prematurely abortedPartially restored
 Rhizobial endocytosis (Bolaños et al., 1996)DecreasedPartially restored
 Bacteroid proliferation (Bolaños et al., 1996)DecreasedRestored
Symbiosome development
 Symbiosomal PsNLEC 1 glycoprotein (Bolaños et al., 2001)Abnormally glycosylatedNot restored
 Peribacteroid membrane RGII-glycoproteins (Redondo-Nieto et al., 2007)AbsentNot studied
Nodule organogenesis
 Anaphase-promoting complex (ccs52 gene) (Reguera et al., 2009)Down-regulatedThis study
 Hydroxyproline/proline-rich proteins (Bonilla et al., 1997a)Not covalently boundNot studied
 Cell wall pectins (Bonilla et al., 1997b)DisorganizedPartially restored

To test the interesting hypothesis that B (and the B–Ca2+ relationship) is affecting the overall control of gene expression during nodule development, we grew the legume model Medicago truncatula inoculated with Sinorhizobium meliloti 1021 in B-sufficient or B-deficient media supplemented with different Ca2+ concentrations (see details in Redondo-Nieto et al., 2003).

Total RNA was isolated from nodules derived from four replicate experiments with three different B and Ca2+ treatments, and following the procedure described in Mergaert et al. (2003), and retrotranscribed in the presence of [α-33P]-deoxycytidine triphosphate to obtain 33P-labelled cDNA. Transcriptome analysis was performed by hybridization of the cDNA on a nylon membrane printed with probes from a cDNA library prepared from M. truncatula nodule mRNA (Györgyey et al., 2000; Mergaert et al., 2003). The probe set on the arrays included genes coding for nodulins (nodule-specific genes), cell cycle proteins, signal transduction proteins, transcription factors, components and regulators of the cytoskeleton, proteins involved in cell wall synthesis, a group representing diverse proteins and a set of clones with no homology to previously characterized proteins. In total, 192 unique genes, repeated three times, were represented on the arrays (see more details and the full data in the Supporting Information, Table S1).

Overall, a general shift was detected in the gene expression of B-deficient nodules, affecting > 70% of genes analysed. The majority were up-regulated genes, although some genes crucial for nodule development and function (discussed later) were down-regulated under B deficiency (Fig. 1). Among the affected gene categories (Table 2), it is noticeable that transcripts coding enzymes important for cell wall synthesis (i.e. cellulose synthase, chitinases) or for structural proline-rich and glycine-rich wall proteins were accumulated. This was likely a secondary response to the unstable cell wall structure under B deficiency, resulting in a demand of cell wall components that increases the expression of genes encoding the wall-building machinery. Short-term alteration of other cell wall-related genes in B-deficient Arabidopsis roots has also been reported (Camacho-Cristóbal et al., 2008), and therefore a primary effect of B deficiency on the expression of wall-modifying and wall-synthesis enzyme genes is also possible.

Figure 1.

Pie charts showing the effect of boron (B) deficiency and the addition of a supplement of Ca2+ on gene expression in nodules of Medicago truncatula. Under B deficiency and normal Ca2+ concentration (−B + Ca), 64% of genes analysed were up-regulated and 12% were down-regulated in response to B deficiency. The addition of Ca2+ (−B + 2Ca) fully (to 90–100%) or partially (to at least 50%) restored the effects of B deficiency on the expression of 77% of genes affected. The supply of extra calcium had none (up-regulated section) or little effect (< 50%, included in low prevention section) on the expression of the 23% genes down- or up-regulated in response to B-deficiency genes (see details and full data in Supporting Information Table S1).

Table 2.   Important examples of genes grouped by functional categories whose expression was affected by boron (B) deficiency (−B) during the development of Medicago truncatula symbiotic root nodules, indicating whether supplement of double concentration of Ca2+ (−B+2Ca2+) could prevent the effects of B deprivation
Category/gene−B−B + 2Ca2+
  1. MAPK, mitogen-activated protein kinase; MRP-like ABC transporter, multidrug resistance associated protein-like ATP-binding cassette transporter; NCR, nodule-specific cysteine-rich nucleotide; Ser/Thr RLK, Serine/Threonine Receptor-Like Kinase.

 MRP-like ABC transporterUp-regulatedRestored
 14-3-3 like proteinUp-regulatedPartially restored
 14-3-3 like proteinUp-regulatedPartially restored
Signal transduction
 GDP dissociation inhibitorUp-regulatedRestored
 Small GTP-binding proteinUp-regulatedRestored
 Small GTP-binding proteinUp-regulatedRestored
 Small GTP-binding protein (RAB1)Up-regulatedRestored
 AMP activated protein kinase betaUp-regulatedRestored
 GTPase-activiting proteinUp-regulatedRestored
 Calmodulin-binding proteinUp-regulatedRestored
 Ser/Thr RLKUp-regulatedRestored
 Ser/Thr RLKUp-regulatedUp-regulated
 Ser/Thr RLKUp-regulatedUp-regulated
 Transcription facto TFIIIUp-regulatedRestored
 Poly(ADP-ribose) polymeraseUp-regulatedRestored
 Myb-like DNA binding proteinUp-regulatedRestored
 WRKY3 DNA binding proteinUp-regulatedUp-regulated
Cell wall
 Cellulose synthaseUp-regulatedRestored
 Putative ripening-related proteinUp-regulatedRestored
 Acidic chitinaseUp-regulatedRestored
 Glycine-rich protein 2Up-regulatedUp-regulated
 Prolin-rich proteinUp-regulatedUp-regulated
 Prolin-rich protein (PRP1)Up-regulatedUp-regulated
 GRP5-like proteinUp-regulatedUp-regulated
 Glycine-rich protein 1Up-regulatedUp-regulated
 Glycine-rich protein 3Up-regulatedUp-regulated
 Leghaemoglobin 1Down-regulatedRestored
 Leghaemoglobin 3Down-regulatedRestored
 Leghaemoglobin 4Down-regulatedRestored
 Leghaemoglobin 5Down-regulatedRestored
 MtN27 nodulinDown-regulatedRestored
 MtN3 nodulin-like proteinDown-regulatedRestored
 MtN22 nodulinUp-regulatedUp-regulated
 Early N-75 nodulinUp-regulatedUp-regulated
 NCR125Up-regulatedPartially restored
 8 NodulinUp-regulatedUp-regulated
 Nodulin 25 precursorUp-regulatedUp-regulated
DNA metabolism
 Histone H1Up-regulatedPartially restored
 Histone H2AUp-regulatedUp-regulated
 Phospholipid glutathione peroxidaseUp-regulatedRestored
 Putative disease resistance proteinUp-regulatedRestored
 Hypersensitive-induced response proteinUp-regulatedRestored
 Cadmium-induced proteinUp-regulatedRestored
 Superoxide dismutaseUp-regulatedRestored
 Ascorbate peroxidaseUp-regulatedRestored
 Wound-induced proteinUp-regulatedRestored
 Putative myosin heavy chainUp-regulatedRestored
 Tubulin beta-1-chainUp-regulatedRestored
Cell cycle
 cdc2 v1Down-regulatedRestored
 cdc2Ms E (ADN7605)Down-regulatedRestored
 cdc2Ms D (ADN7604)Down-regulatedRestored
 cdc2Ms F (ADN7606)Down-regulatedRestored
 cdc2 v2Down-regulatedRestored

Genes typically involved in nodulation were either up-regulated (including several early and late nodulins) or down-regulated, for example, leg-haemoglobin genes previously reported in B-deficient Pisum sativum nodules (Reguera et al., 2010c). One important effect of B deficiency is an aberrant nodule organogenesis in which abnormal cell proliferation, accompanied by a lack of differentiation, was attributed to the down-regulation of the ccs52a gene (Reguera et al., 2009), which is involved in cell cycle arrest and transition of mitotic to endoreduplication cycles (Cebolla et al., 1999; Vinardell et al., 2003). In our study, not only ccs52a but also other genes involved in cell cycle regulation, including some cdc genes, were down-regulated, although expression of cyclin genes was not significantly affected by B deficiency (Table S1). Moreover, other transcripts involved in cell cycle regulation were accumulated. This indicates a perturbation in the cell cycle regulation during low-B nodule organogenesis. Other important categories of genes up-regulated by B deficiency included most of those coding for proteins involved in signal transduction and genes expressed in response to several types of stresses, which may be expected under B deficiency, as has been reported in low-B cultures of Nicotiana tabacum BY-2 cells (Kobayashi et al., 2004).

Interestingly, the effect of the lack of the micronutrient on the expression of most genes (> 75% of the up-regulated or down-regulated ones) was, at least partially, prevented by adding Ca2+ during plant growth and nodulation (Fig. 1, Tables 2, S1; −B + 2Ca treatment). Some of these results also support the suggestion by Koshiba et al. (2010) that Ca2+ influx plays a role in B-deprivation stress signalling. Accordingly, what is reported here agrees with those previous reports where application of a higher dose of Ca2+ to low-B N. tabacum BY-2 cells also affected the B-responsive gene expression. The M. truncatula genes with expression restored following Ca2+ application included most genes involved in cell cycle regulation, signal transduction and stress-responsive genes, enzymes involved in cell wall synthesis, several leghaemoglobins and nodule-specific cysteine-rich nucleotides (NCRs). However, although some NCRs govern terminal bacteroid differentiation (Van de Velde et al., 2010), the expression of other NCRs and nodulins or structural cell wall proteins were not restored to normal levels by the addition of Ca2+. This correlates with our previous report that Ca2+ can prevent abnormal nodule tissue differentiation under B deficiency but not the alteration of cell wall structure or aberrant bacteriod differentiation (Redondo-Nieto et al., 2003).

Overall, the results show that B nutrition affects the expression of a majority of genes during nodule organogenesis. To our knowledge, there is no other nutrient deficiency, including Ca2+, that results in such an alteration of gene expression. It has been proposed that the disturbed pectic network triggers low B-stress signalling, leading to reactive oxygen species (ROS) production and cell death in cultured tobacco BY-2 cells (Koshiba et al., 2009). However, as mentioned earlier, symptoms of B deficiency appeared in nodules earlier than in shoots or roots (Bolaños et al., 1994), and B deficiency apparently did not elicit ROS production and oxidative damage in nodules (Reguera et al., 2010c). Moreover, short-term studies performed by our group on B-deficient nodule development have shown that expression of some of the genes related to the symbiotic process, in particular those induced early after inoculation with rhizobia, were also affected (Redondo-Nieto et al., 2001; Reguera et al., 2009, 2010b). Furthermore, B deficiency also negatively regulates the physiology and development of living forms without cell walls, including animals (Rowe & Eckhert, 1999), and therefore it is unlikely that altered cell wall stability in low-B nodules is the main or only cause of this loss of gene regulation. Why then is there such a far-reaching effect? The key role of B on animal physiology is consistent with the fact that low B induced the mitogen-activated protein kinase (MAPK) pathway in cultured animal cells and that mammalian cell lines mutated in B transporters stop developing and proliferating (Park et al., 2004, 2005). The finding that B in bacteria is a signalling molecule (Chen et al., 2002) and that borate compounds can interact with regulatory proteins (see several examples in Golbach & Wimmer, 2007) admits the exciting possibility that B also exerts its action through interaction with transcription factors, hence explaining such a wide alteration of gene expression, as was hypothesized by González-Fontes et al. (2008). Supporting this hypothesis, Kasajima et al. (2010) have suggested that WRKY6 is a transcription factor involved in response to B deficiency in Arabidopsis.

Other possibilities come from the fact that several components of plant cell signalling and signal transduction pathways during organogenesis processes are known to be affected by B deficiency. Several studies on B effects in plant and animal development and metabolism point to a role for B in extracellular matrices and/or membrane functions (Brown et al., 2002). There is a well-known relationship linking B, phytohormones and development (Wang et al., 2006; Martín-Rejano et al., 2011), albeit this seems to be insufficient to explain the high B requirement for nodule organogenesis previously reported (Redondo-Nieto et al., 2003). Boron has been described as a modulator of host plant–rhizobia interactions by interacting with legume arabinogalactan proteins important for nodule infection and development (Reguera et al., 2010a). Moreover, several reports described abnormal glycosylation of glycoproteins apparently involved in cell-to-cell signalling during symbiosome development in nodules (Bolaños et al., 2001; Redondo-Nieto et al., 2007), and a similar role for B during animal organogenesis has been hypothesized (Redondo-Nieto et al., 2008). In a preliminary study (Bolaños et al., 2011), we have detected altered overall protein N glycosylation as a very early response to B deprivation during legume nodule and Arabidopsis root or shoot apical meristem development. This consisted of abnormal accumulation of high-mannose-type N-glycans, typical for the endoplasmic reticulum. As described earlier, B deficiency led to uncontrolled cell proliferation and aberrant cell differentiation, and interestingly, accumulation of high-mannose N-glycans is characteristic of some breast cancer lines (de Leoz et al., 2011). Therefore, the role of B in glycosylation mechanisms important for organogenesis has to be explored in more detail.

Moreover, using this M. truncatula nodule organogenesis model, an approach to identify molecules involved in cell signalling, other than glycans or transcription factors, affected by B and/or able to interact with borate could be designed. An exciting report by Ricardo et al. (2004) proposed that borate minerals could play a crucial role in an early ‘RNA world’ of life on Earth by stabilizing cyclic ribose; therefore, the structural and functional stability of microRNAs by borate is a possibility that should be tested. Also, inositides or adenylates, which are potential targets of B (Ralston & Hunt, 2001; Bolaños et al., 2004), could be part of signalling pathways affected by B deficiency. The fact that Ca2+ prevents altered expression of most genes affected by B deficiency suggests an effect of B on Ca2+-mediated signalling pathways. In this regard, it has been demonstrated that boric acid can prevent proliferation of some tumour cells by affecting the release of Ca2+ by cyclic ADP ribose (cADPr), and that this effect is based on its binding capacity to NAD+, the substrate of ADP ribosyl cyclase (Barranco et al., 2008). Therefore, in addition to exploring a primary interaction of B with transcription factors or with signalling and regulatory molecules such as glycans or microRNAs, it will be interesting to determine whether B deficiency affects Ca2+ influx (as suggested by Koshiba et al., 2010) and/or also interferes with Ca2+-releasing mechanisms that are dependent on adenylates or inositides. Studies like this will shed new light on the role of B in the regulation of gene expression and on the effect of the B–Ca2+ relationship on gene expression.


This work was supported by the Ministerio de Ciencia e Innovación BIO2008-05736-CO2-01 and by the MICROAMBIENTECM Program from the Comunidad de Madrid.