A single rye plant possesses a root surface of about 640 square metres (Dittmer 1937). During limitation of immobile nutrients, such as iron (Fe) and phosphate (P), plants increase their root surface by the development of extra root hairs to achieve an improved mobilization and uptake of the respective nutrient. In this work, root hair development adapted to phosphate and iron deficiency has been investigated. In this process, root hair formation is linked to iron and phosphate homeostasis.

1.1  Iron homeostasis in plants

As a transition metal, iron is able to change its oxidation state, making it an important component of biological redox systems such as photosynthesis or the respiratory chain. Although iron is the fourth most abundant element in the lithosphere (http://www.matpack.de/Info/Nuclear/Elements/lithosphere.html), under aerobic conditions it is poorly available for plants, because Fe3+ is highly insoluble due to precipitation of iron hydroxide or iron phosphate. In anaerobic environments (e.g. in waterlogged soils) the concentration of the soluble Fe2+ increases. When Fe2+ is taken up in excess, it is toxic, because it reduces oxygen into superoxide radicals that generate hydrogen peroxide. In the Fenton reaction, hydroxyl radicals are produced. The hydroxyl radicals cause lipid peroxidation and other, non-selective, oxidative damage (Marschner 1995). Thus, in living organisms iron is chelated and its homeostasis underlies a strict regulation (Schmidt 1999, 2003, Hell & Stephan 2003).

1.1.1  Acquisition and uptake of iron from the soil

A visible iron deficiency symptom is an interveinal chlorosis in young leaves, as several steps of chlorophyll synthesis depend on iron (Marschner 1995). For improved iron mobilization and uptake, a variety of physiological and morphological adaptation reactions have evolved. Morphological changes of the root system inducible by a low Fe availability are an enhanced lateral root development and an increased number of root hairs (Moog et al. 1995, Pinton et al. 1998, Landsberg 1996, Schmidt 1999). Also root tip swellings were observed (Landsberg 1996). In legume or Proteaceen species, proteoid roots, which are dense clusters of short laterals with a high root hair density, develop in response to a low Fe supply (White & Robson 1989, Arahou & Diem 1997). According to López-Bucio et al. (2003), the main processes affecting the overall root system architecture are the primary root meristem activity leading to changes in the primary root length, lateral root formation, which increases the exploratory capacity of the root system, and the development of additional root hairs, which increases the surface area of the root. The Fe-deficiency-responsive formation of rhizodermal transfer cells, which have labyrinth-like cell wall inversions, in a variety of species increases the apoplast-symplastic surface area if the root (Kramer et al. 1980, Schmidt 1999).


Iron uptake is accompanied by an acidification of the rhizosphere by members of the H+-ATPase gene family. The lowering of the pH increases iron hydroxide solubility and proton extrusion energizes the membrane for uptake (Marschner 1995, Schmidt 1999, Palmgren 2001). Furthermore, phenolics like piscidic acid and alfafuran, organic acids, mainly citrate and malate, and riboflavin are secreted during Fe starvation. These substances serve as iron chelators, antimicrobial agents, or provide electrons for Fe3+ reduction and contribute to iron acquisition (Marschner 1995, Schmidt 1999, Abadía et al. 2002, Ae et al. 1990, Masaoka et al. 1993, Landsberg 1986, Susín et al. 1993).

In dicots, Fe depletion induces a reduction-based iron uptake system in the root consisting of a ferric-chelate reductase and a high-affinity iron transporter, referred to as strategy I (Römheld 1987). Strategy II is confined to grasses and is characterized by enhanced release of phytosiderophores, which derive from the precursor nicotianamine (NA) and form stable complexes with Fe3+. These complexes are taken up by a specific transporter identified as the oligopeptide transport protein ZmYS1 (YELLOW STRIPE1; Römheld & Marschner 1986, Curie et al. 2001). Strategy I and II are not mutually exclusive since the transporter and the reductase are present also in rice roots (Schmidt 2006).

The gene responsible for the reductase activity in strategy I plants is FRO2 (FERRIC-CHELATE REDUCTASE/OXIXASE2) belonging to the superfamily of flavocytochromes in Arabidopsis and its ortholog FRO1 in pea (Robinson et al. 1999, Waters et al. 2002). FRO1 mRNA is localized in the rhizodermis, cortex, and vascular cylinder of the root and in the mesophyll and parenchyma of the leaf (Waters et al. 2002). In Arabidopsis, eight FRO genes exist that are expressed differently in all plant organs and are predicted to function on different subcellular membranes (Mukherjee et al. 2006).


The main transporter of strategy I plants is the ZIP metal transporter IRT1 (IRON REGULATED TRANSPORTER1; Eide et al. 1996, Eckhardt et al. 2001, Cohen et al. 2004). IRT1 and FRO2 are induced under Fe deficiency and are coregulated (Connolly et al. 2003, Vert et al. 2003). IRT1 is repressed at the protein level by Fe (Connoly et al. 2002). It is localized in the plasmalemma of the rhizodermis, cortex, and anthers (Vert et al. 2002).

1.1.2 Iron release into the xylem and long-distance transport through the xylem

After iron has entered the rhizodermis, the main chelator of Fe2+ in the symplast is the ubiquitous aminoacid nicotianamine (NA, Hell & Stephan 2003, Stephan et al. 1996, von Wirén et al. 1999). During Fe starvation, nicotianamine synthase is strongly upregulated in the root (Wintz et al. 2003). Radial transport of Fe2+-NA from the rhizodermis to the xylem vessels takes place on a symplasmic route (Stephan et al. 1996). Immunolabeling showed that NA is deposited mainly in the vacuoles of root stelar cells (Pich et al. 1997). In the iron-overaccumulating pea mutants brz (bronze) and dgl (degenerated leaflets), vacuolar NA is increased suggesting a function of NA also in detoxification of excess iron (Pich et al. 2001).

Xylem-loading occurs mainly in the xylem parenchyma; it is regulated independently from the uptake into the rhizodermis and cortex (Marschner 1995). Long-distance transport via the xylem occurs as an Fe3+-citrate complex (Tiffin 1966, Pich et al. 1994, López-Milán et al. 2000), meaning a reoxidation of the iron when released into the xylem. So far, no ferroxidase has been described in higher plants, but Herbik et al. (2002) showed that the multicopper ferroxidase-like protein FLP participates in high-affinity iron uptake of Clamydomonas, whose iron uptake mechanism involves a reductive step as for strategy I plants (Eckhardt & Buckhout 1998).


Efflux of citrate into the apoplast for Fe3+ chelation is catalyzed by FRD3 (FERRIC REDUCTASE DEFECTIVE3), a member of the multidrug and protein efflux family (MATE), which is able to transport citrate (Rogers & Guerinot 2002, Durrett et al. 2006). Mutation of FRD3 leads to ferric iron precipitates in the root vascular cylinder and leaf apoplast and to a decreased iron content within the leaf cells (Delhaize 1996, Green & Rogers 2004). FRD3 functions root-specific and is expressed in the pericycle and the vascular cylinder (Green & Rogers 2004).

The plasma membrane Fe2+-NA transporter AtYSL2 (YELLOW STRIPE-LIKE2) is expressed in the endodermis, pericycle, and most strongly in the xylem parenchyma of the root and shoot (DiDonato et al. 2004, Schaaf et al. 2005). The function of YSL2 is not entirely clear. Because xylem parenchyma cells play a key role in ion secretion into the xylem and in the reabsorption of ions from the xylem sap along the pathway to the shoot (Marschner 1995), the localization of AtYSL2 indicates a function in one of these processes. AtYSL2 is repressed by Fe deficiency (DiDonato et al. 2004, Schaaf et al. 2005). The oligopeptide transporter OPT3, which shares sequence similarities with YSL2, is increased in roots of Fe-starved plants. OPT3 transports iron, independent of the presence of NA (Wintz et al. 2003).

The natural resistance-associated macrophage protein NRAMP1 transports Fe3+ and is, together with NRAMP3 and NRAMP4, induced in the roots of Fe-deficient Arabidopsis. NRAMP2 is upregulated under iron sufficient conditions (Curie et al. 2000, Thomine et al. 2000). NRAMP3 and NRAMP4 are located in the tonoplast of the vascular tissues of roots and shoots suggesting an involvement in long-distance transport (Thomine et al. 2003). In the nramp3nramp4 double mutant, iron is not mobilized from vacuolar globoids, indicating a function of NRAMPs in the export of iron from vacuolar metal pools under iron starvation (Lanquar et al. 2005).


The vacuolar iron transporter VIT1 is involved in sequestration of iron into vacuoles. VIT1 is expressed throughout the plant independent of the iron availability but most strongly in seeds. In wildtype embryos, iron is concentrated in the provascular strands of the cotyledons, hypocotyls, and radicle, which is abolished in vit1 (Kim et al. 2006).

1.1.3 Iron uptake by leaf cells, iron storage, and phloem-transport into sink organs

After the iron has reached the leaf apoplast, uptake by leaf cells depends on the reduction of iron, most likely catalyzed by members of the FRO family (Mukherje et al. 2006). Also photoreduction has been observed (Brown et al. 1979). Chelation of Fe2+ by NA within the leaf cells is supported by the fact that the NA-deficient tomato mutant chloronerva reveals iron precipitates along the veins and a permanent interveinal chlorosis rescued by exogenous NA application (Scholz 1989). The chlorosis indicates a lack of mobile iron within the intercostal mesophyll. Cloning of chloronerva revealed that the mutated gene encodes nicotianamine synthase (Ling et al. 1999).

About 80% of the leaf iron is located in the chloroplasts (Smith 1984). The major iron storage component in the plant is the protein phytoferritin (Briat et al. 1999). Ferritin has ferroxidase activity and consists of a hollow sphere containing an iron phosphate core (Wade et al. 1993, van Wuytswinkel et al. 1995, Harrison & Arosio 1996). Ferritin accumulates mainly in the non-green plastids of roots and shoots and is induced by iron excess (Lobréaux & Briat 1991). The transmembrane protein PIC1 (PERMEASE IN THE INNER ENVELOPE OF CHLOROPLASTS1) is located in the plastids and mediates iron transport. pic1 mutants only grow heterotrophically and are chlorotic and dwarf, indicating that PIC1 may be important for iron transport into the plastids (Duy et al. 2006).


Translocation of iron from the leaf cells to sink organs depends on symplastic transport via plasmodesmata and the phloem, since meristems and newly developing organs do not have complete xylem structures (Hell & Stephan 2003). Evidence for the symplastic iron transport was provided by Zhang et al. (1995), as radioactive iron fed to the xylem did not directly reach the shoot apex but after remobilization from older leaves.

In the phloem sap, iron is predominantly in the Fe3+ state (Maas et al. 1988), meaning a further oxidation of the iron when released into the phloem. As an iron chelator involved in long-distance transport via the phloem, the small iron transport protein ITP belonging to the family of dehydrins has been isolated. ITP specifically binds Fe3+ in vitro and transports iron in vivo (Krüger et al. 2002). The iron in sink organs like beech nuts and young growing tissues is again in the form of Fe2+-NA (Kristensen & Larsen 1974, Stephan et al. 1990, 1994, Pich et al. 1994). Thus, phloem unloading of Fe3+-ITP to Fe2+NA must involve iron reduction. Possible candidates catalyzing this process are members of the FRO gene family (Mukherjee et al. 2006).

The Fe2+-NA transporter AtYSL1 is expressed in the xylem parenchyma of leaves, pollen, and in young silique parts (Le Jean et al. 2005). The seeds of the ysl1 mutant contain less iron and nicotianamine. Thus, YSL1 seems to be crucial for iron loading into sink organs. Hell and Stephan (2003) speculated that NA could have a shuttle function in phloem unloading. In pea seeds, ferritin is the major iron storage component (Marentes & Grusak 1998). Figure 1 summarizes the path of iron through dicotyl plants.


Figure 1 : A model for iron transport through dicotyl plants. The apoplast is marked in grey, the symplast in white. Iron uptake starts in the rhizodermis (Rh). The Fe3+ ions are mobilized by the secretion of phenolics (P), organic acids (OA), and protons (H+) into the rhizosphere. Free or chelated iron diffuses or precipitates within the apoplastic space and is detained by the casparian band. Uptake into the rhizodermal cells takes place after reduction of Fe3+ chelates by FRO2 through IRT1. Inside the root cells, iron is chelated by NA and reaches the cortical (C), endodermal (En), pericycle (Per) and xylem parenchyma cells of the root (Xpr) via the symplastic way through plasmodesmata. NA accumulates in the vacuoles of the endodermis. VIT1 mediates iron sequestration into vacuoles. Xylem loading is mainly accomplished by the xylem parenchyma. The main iron transport species within the xylem is Fe3+ citrate. A prerequisite for xylem transport of iron is the excretion of citrate (Cit) into the apoplast catalyzed by FRD3. For the reoxidation of iron during the transfer from Fe2+-NA towards Fe3+ citrate, no molecular component has been identified so far. It is not clear, whether the iron is oxidized before or after release into the apoplast. FRO3 and YSL2 are induced in the endodermis, pericycle, and the xylem parenchyma cells of root (Xpr) and shoot (Xps), but their function is unknown. From their localization, a role in xylem loading or exchange processes along the pathway to the shoot could be assumed. The vacuolar metal transporter NRAMP also occurs in the shoot xylem parenchyma and may be involved in the remobilization of iron from vacuolar iron pools. After release of Fe3+ citrate into the leaf apoplast, iron uptake by the leaf cells again depends on FRO action. An acidification of the apoplast is also important. The symplastic transport within the leaf cells again occurs as Fe2+-NA. Transport into plastids is mediated by PIC1. The main iron storage component in plastdis is ferritin that has ferroxidase activity. Phloem transport towards sink organs mainly occurs in the Fe3+ state, probably bound to ITP. Phloem parenchyma is not shown; it expresses FRD3 and FRO3. The loading of iron into sink organs is not well understood. A function of Fe2+-NA as the transport form within the cells would comprise a further reduction step catalyzed by FRO gene family members. IRT1, YSL1, and YSL2 are found in sink organs. An important iron storage component within seeds beneath NA is ferritin.

1.1.4 Regulation of iron homeostasis

An important mediator regulating the strategy I and the morphological adaptations to Fe deficiency is the bHLH protein FER, also designated as FRU (FER-LIKE REGULATOR OF IRON UPTAKE) or FIT (FE-DEFICIENCY INDUCED TRANSCRIPTION FACTOR, Jacoby et al. 2004, Colangelo & Gueriot 2004). FER was initially cloned from tomato (Ling et al. 2002). The fer mutant is chlorotic and fails to induce ferric-chelate reductase activity or FRO2 and IRT1 transcription (Brown et al. 1971, Ling et al. 2002). Homozygous fru or fit knock-outs of Arabidopsis reveal the same phenotype and have reduced iron content (Jacoby et al. 2004, Colangelo & Guerinot 2004). The expression of 72 out of 179 iron-regulated genes depend on FIT (Colangelo & Guerinot 2004). Grafting experiments revealed that FER acts root-specific (Brown et al. 1971). In the late meristematic region of tomato roots, FER is located in the nucleus of the rhizodermis and outer cortical cell layers. In the mature root hair zone FER transcripts were found in the vascular cylinder between the xylem and the phloem poles (Brumbarova & Bauer 2005, Ling et al. 2002). In Arabidopsis, FRU shows a similar expression pattern (Jacoby et al. 2004). This implies that in addition to controlling the uptake from the soil, FER/FRU/FIT also might regulate xylem-loading or long-distance transport within the root. FER/FRU/FIT shows highest expression in response to iron deficiency at the transcriptional and posttranscriptional level and is down-regulated at generous iron supply (Jacoby et al. 2004, Colangelo & Guerinot 2004, Brumbarova & Bauer 2005). Since FER/FRU/FIT is controlled by iron, the iron status is sensed upstream of FER action (Brumbarova & Bauer 2005). FIT overexpressers had no obvious phenotype, which led Colangelo & Guerinot (2004) to conclude the protein levels are not altered in these lines or an additional factor is necessary for regulation of the target genes.

Auxin and ethylene are assumed to be involved in the iron-deficiency-responsive induction of transfer cells and extra root hairs, because application of either hormone mimics the morphological Fe-deficiency phenotype, and Fe starvation increases the synthesis of auxin and ethylene (Landsberg 1984, 1996, Schmidt & Bartels 1996, Schmidt et al. 2000, Römheld & Marschner 1986, Romera et al. 1999). Pharmacological inhibition of ethylene synthesis or action represses the Fe stress syndrome (Romera & Alcántara 1994). Auxin transport inhibitors or ethylene inhibitors suppress root hair development of iron sufficient or deficient Arabidopsis. Auxin- or ethylene-insensitive mutants were not able to produce root hairs in the presence or absence of Fe (Schmidt & Schikora 2001). Because auxin can enhance ethylene production, it is supposed the impact of auxin on the morphological Fe deficiency responses could be mediated through ethylene (Yu & Yang 1979, Kim et al. 1992, Romera et al. 2006).

1.1.5 Linking of iron homeostasis to the homeostasis of other nutrients


In Fe-starved Arabidopsis, Thimm et al. (2001) showed a transcriptional induction of several primary metabolism enzymes. αamylase and glycolysis enzymes are upregulated in the shoot, whereas citrate cycle, respiration, and fermentation are induced in the root, indicating that triosephosphates are mobilized beyond photosynthesis yield in the shoot and are sent towards the root to supply the need for energy and reduction equivalents. Anaplerotic sequences are increased in the root fitting with the excretion of organic acids (Thimm et al. 2001).

Several components involved in iron homeostasis also bind other transition metals as listed in Table 1. Therefore, iron homeostasis is linked with the homeostasis of other transition metals.

Table 1 : Affinity of several components involved in iron homeostasis to divalent metals other than iron.





Fe, Cu, Mn

Norvell et al. 1993, Welch et al. 1993


Fe, Zn, Mn, Co, Cd

Rogers et al. 2000


Fe, Cu, Zn, Mn

Thomine et al. 2000, 2003


Fe, Zn, Mn, Cd

Stephan et al. 1996


Fe, Cu, Zn, Mn, Ni, Cd

Schaaf et al. 2004



Fe, Cu

Fe, Mn

Duy et al. 2006

Kim et al. 2006


Phosphate strongly chelates iron, in the soil as well as in the plant (Marschner 1995). To that effect, phosphate starvation increases iron accumulation involving ferritin expression, which marks iron excess. High phosphate supply induces IRT1 transcription normally induced by iron deficiency. This interaction leads to a mutual impact of iron and phosphate homeostasis (Hirsch et al. 2006).

1.2 Phosphate homeostasis in plants

Phosphate is required in the range of 0.3-0.5% of the plant dry matter for optimal growth. In cells, it occurs as phosphate ester or an anhydride bond, which are stable, but can also easy be hydrolysed. These properties make phosphate be an important structural element in nucleic acids and phospholipids and play a central role in energy metabolism and regulatory phosphorylation. The high phosphate demand of plant cells stands in contrast to the low P availability of the most soils. Although abundant, phosphate is highly immobile because of strong interactions with Fe, Ca, Al, and organic compounds (Marschner 1995, Raghothama 1999). Therefore, plants have evolved various strategies for P acquisition from the soil and internal remobilization of phosphate (Raghothama 1999, Raghothama & Karthikeyan 2005, Poirier & Bucher 2002, Franco-Zorrilla et al. 2004, Hammond et al. 2004, Ticconi & Abel 2004).

1.2.1  Acquisition of phosphate from the soil

In more than 80% of all land plants, phosphate acquisition by roots is accomplished by mycorrhiza, a symbiosis formed by plant roots with soil fungi from the order Glomeromycota. The fungus provides the plant with nutrients from the soil, mainly phosphate, in exchange for photosynthetic carbohydrates (Karandashov & Bucher 2004). The fungal mycelium increases the explored soil volume. In non-mycorrhizal plants, extension of P depletion zones is closely related to root hair length, whereas in mycorrhizal plants, P depletion zones by far exceed the root hair cylinder (Marschner 1995). Based on fossil data, mycorrhiza is even thought to be a precondition for colonization of land by the first terrestrial plants (Karandashov & Bucher 2004, Wellmann et al. 2003).


Apart from mycorrhiza, an improved soil exploration is achieved by restructuring root system architecture. Under phosphate starvation, root biomass increases, thereby, in relation to the shoot, which is consistent with the transport carbohydrate and phosphate from the shoot towards the root (Marschner 1995, Jeschke et al. 1996, 1997, Raghothama 1999, Forde & Lorenzo 2001). In P-starved Arabidopsis, primary root growth is reduced caused by a shift from an indeterminate to a determinate developmental program due to loss of meristematic cells (Williamson et al. 2001, Linkohr et al. 2002, Sánchez-Calderón et al. 2005). Primary root growth inhibition is accompanied by an increase in the number and length of laterals (Bates & Lynch 1996, Williamson et al. 2001, Linkohr et al. 2002). Nacry et al. (2005) observed contrasting effects of P starvation on lateral root formation with a strong inhibition of primordia initiation combined with a marked stimulation of their activation. The latter was also described by López-Bucio et al. (2005). In bean, soybean, and pea, adventitious rooting occurs. In addition, the angle of the root growth becomes more horizontal representing a gravitropic response. The production of this shallower root system under P stress is termed 'topsoil foraging', since P availability is normally highest near the soil surface (Bonser et al. 1996, Lynch & Brown 2001). Using a geometric model, Ge et al. (2001) calculated a reduction of inter-root competition within the same plant by topsoil foraging.

In P-deficient rape, spinach, tomato, Arabidopsis, and barley, the root surface is increased by the formation of additional root hairs that are also elongated (Foehse & Jungk 1983, Bates & Lynch 1996, Gahoonia & Nielsen 1997). In addition, root hairs of P-starved plants initiate closer to the meristem (Ma et al. 2003). Phosphate-sensitive root hair elongation is the result of both, an increased growth rate and growth duration (Bates & Lynch 1996). Root hairs are the major site of phosphate uptake (Gahoonia & Nielson 1998). Arabidopsis mutants with short or no root hairs revealed a lower P accumulation under P limitation than wildtype plants (Bates & Lynch 2000). Phosphate efficient Arabidopsis accessions have more and longer root hairs and higher root penetration abilities than inefficient types indicating the importance of explorative root morphology for P acquisition (Narang et al. 2000). Also the development of transfer cells under P shortage has been described (Schikora & Schmidt 2002). A massive increase of the root surface area is achieved by the development of cluster (proteoid) roots in Proteaceae, legumes, and other families (Lamont 2003).

Cluster roots exudate high amounts of citrate, malate, and oxalate, which exchange phosphate from Fe, Al, and Ca salts (Gardner et al. 1983, Grierson 1992). Excretion of organic acids for P acquisition also occurs in species without cluster root formation (Lipton et al. 1987). For refilling the citrate cycle, phosphoenolpyruvate carboxylase is induced (Pilbeam et al. 1993, Johnson et al. 1996). The importance of citrate exudation is demonstrated in Arabidopsis overexpressing citrate synthase that shows an improved growth under P limitation (Koyama et al. 2000). Phosphate-efficient Arabidopsis accessions secret more organic acids than inefficient ones (Narang et al. 2000). A phosphate starvation-inducible MATE is suggested as a potential candidate for organic acid secretion (Vance et al. 2003).


Piscidic acid and alfafuran are secreted for solubilization of P from iron salts (Marschner 1995). Under P limitation, the rhizosphere is acidified, which is attributed to the H+-ATPase activity. The pH decrease can dissolve P from Ca phosphate and is important for phosphate uptake (Moorby et al. 1988, Yan et al. 2002).

To mobilize P from organic components, hydrolytic enzymes are secreted into the rhizosphere. Among these are acid phosphatases with low substrate specifity that are induced by P deficiency (Ueki 1978, Lefebvre et al. 1990, Duff et al. 1991). Acid phosphatase enzyme activity is correlated with synthesis of acid phosphatase mRNA and protein (del Pozo et al. 1999, Miller et al. 2001). In addition, phosphate is remobilized within P-deficient plants, which is achieved by induction of at least one specific intracellular phosphatase in the vacuole (Duff et al. 1991). In Arabidopsis, three purple acid phosphatases are inducible by low P (del Pozo et al. 1999, Li et al. 2002).

Further hydrolytic enzymes induced upon P starvation are extracellular RNases (Nürnberger et al. 1990, Löffler et al. 1992). The RNase genes RNS1 and RNS2 are upregulated in P-deficient Arabidopsis, but also during senescence. From putative signal sequences, excretion into the extracellular matrix and rhizosphere for P mobilization from the soil and recycling within the plant is presumed (Taylor et al. 1993, Bariola et al. 1994). An extracellular cyclic nucleotide phosphodiesterase together with the other nucleolytic enzymes enables the complete utilization of P from nucleotides (Abel et al. 2000). Arabidopsis grown with DNA as the only P source showed normal biomass production, indicating an excretion of DNase (Chen et al. 2000).

1.2.2 Phosphate transport


Phosphate uptake is driven by a proton gradient generated through activity of an H+-ATPase and takes place as proton symport (Ullrich-Eberius et al. 1984, Sakano 1990, Muchhal et al. 1997). In plants, phosphate transporters are grouped into three families, Pht1, Pht2, and Pht3 (Rausch & Bucher 2002). They belong to major facilitator superfamily, which occurs ubiquitously and transports small solutes along chemiosmotic gradients (Raghothama 1999, Pao et al. 1998). High-affinity phosphate transporters (Pht1) have been identified in many species (Muchhal et al. 1996, Smith et al. 1997, Kai et al. 1997, Leggewie et al. 1997). The Pht1 family in Arabidopsis consist of nine members that are differentially expressed in root and shoot tissues (Mudge et al. 2002). Pht1;1-4 are induced by low P in the rhizodermis, preferentially in root hairs, and in the cortex. Pht1;1 and Pht1;4 play a significant role in P acquisition since the pht1;1pht1;4 double mutants show a 75% reduction in P uptake capacity even under sufficient P supply (Shin et al. 2004). The Pht1 high-affinity transporters take up phosphate from the apoplast into the symplast during acquisition from low-P soil as well as during long-distance transport, uptake into leaf cells, and remobilization processes within the plant in the course of the life-cycle and under P limitation (Mudge et al. 2002). Intracellular trafficking of Pht1 towards the plasma membrane is regulated by PHF1 (PHOSPHATE TRASPORTER TRAFFIC FACILITATOR1), which is localized in the endoplasmatic reticulum and related to SEC12 proteins of the early secretory pathway. PHF1 enables the endoplasmatic reticulum exit of Pht1 and is induced by P starvation (González et al. 2005).

Pht2;1 of Arabidopsis is a low-affinity phosphate transporter. It is mainly expressed in green tissues irrespective of the P status of the shoot. Its ortholog in potato is also transcribed in roots and tubers (Daram et al. 1999, Rausch et al. 2004). Pht2 was detected in the inner chloroplast membrane (Ferro et al. 2002, Rausch et al. 2004). It is expressed in the leaf cells, most strongly in the vasculature and pericycle of root and shoot, but also in flowers, siliques, and seeds, and it is induced by light. Thus, Pht2 is under the control of developmental signals, light, and sink strength (Rausch et al. 2004). Members of the Pht3 family are deduced to be mitochondrial P transporters (Rausch & Bucher 2002, Kiiskinen et al. 1997, Takabatake et al. 1999).

1.2.3 Xylem loading of phosphate and distribution within the plant 

The concentration of inorganic phosphate in the cytoplasm of root cells is generally low. After uptake into rhizodermis and cortex, P is quickly incorporated into organic molecules. When the P availability is high, the phosphate content of the vacuoles increases. Release into the xylem occurs again as free phosphate (Marschner 1995).


An Arabidopsis mutant defective in xylem loading is pho1. It reveals reduced phosphate content in the shoot along with increased anthocyanin production while root P concentration is slightly elevated. P uptake into hypocotyls of derooted pho1 plants is not impaired (Poirier et al. 1991). PHO1 is a transmembrane protein that shows highest sequence homology with the Rcm1 mammalian cell surface receptor for retroviruses and with the Syg1 protein from yeast involved in the mating pheromone signal transduction pathway (Hamburger et al. 2002). Syg1 interacts with the heterotrimeric G protein complex (Spain et al. 1995). PHO1 is expressed in the mature vascular system of the root and lower hypocotyl, where it is most pronounced in pericycle and xylem parenchyma cells (Hamburger et al. 2002). Evidence for P transport by PHO1 is lacking. Thus, PHO1 may be is a subunit of a multisubunit transporter or it may influence P export into the xylem indirectly (Hamburger et al. 2002).

The principle transport form of P in the xylem is inorganic phosphate, and the P concentration in the xylem sap is constant, independent of the P supply (Bieleski 1973, Mimura et al. 1996 Jeschke et al. 1997). The cytoplasmic P content of the leaf cells remains unaffected by the P availability while vacuolar P is exhausted under P deficiency, which coincides with the apoplastic P concentration (Mimura et al. 1996).

The main pathway for P distribution is the phloem. Phosphate moves quickly from the xylem to the phloem and is very mobile within the plant so that an individual P atom can make several circuits through the organism (Bieleski 1973).

1.2.4 Metabolic adaptations to P shortage


In addition to mobilization of P by intracellular nucleases and phosphatases, phosphate is released from phosphoenolpyruvate (Poirier & Bucher 2002, Hammond et al. 2004). Starch is synthesized in the leaves (Foyer & Spencer 1986, Fredeen et al. 1989, Ciereszko et al. 2001). The conversion of triose phosphate into starch leads to a release of P within the chloroplast (Poirier & Bucher 2002). Membrane phospholipids are replaced with non-phosphorus galactolipids and sulfolipids (Poirier & Bucher 2002, Hammond et al. 2004). During P starvation, the portion of P translocated from the shoots to the roots is increased (Drew & Saker 1984, Jeschke et al. 1996, 1997).

If the intracellular P concentration is low, ATP synthesis through photosynthetic and respiratory electron flow is limited. Plants acclimatize with structural re-assembly of the photosynthetic apparatus and the use of alternative oxidase pathway. Also several key enzymes of the Calvin Cycle are inhibited. In addition, metabolic adaptation responses are induced that bypass ATP- and P-dependent enzymes of the glycolysis by using pyrophosphate or forbearing phosphorylation steps (Poirier & Bucher 2002).

Phosphate starvation leads to an increased anthocyanin synthesis. As anthocyanins absorb UV light, they could protect nucleic acids from UV radiation during P-limited photosynthesis (Poirier & Bucher 2002, Hammond et al. 2004).

1.2.5 Integration of local and systemic signals by the PHR1/PHO2/At4 pathway


The physiological and morphological phosphate deficiency responses in plants are thought to be differentially controlled by a systemic signaling mechanism communicating the P status of the leaf cells to the root and by a local signaling mechanism perceiving the P availability in the soil. Split root experiments and replenishment studies showed that root hair elongation is induced by local P deprivation regardless of P status of the plant (Bates & Lynch 1996). Primary root growth slows down when the root tip grows from a Pdepleted environment into a Prich segment and shows a further growth reduction when again reaching a P-depleted zone. This indicates an interaction of local and systemic signals in regulating primary root meristem activity. Lateral root density was not affected by heterogeneously supplied phosphate, whereas lateral root elongation was greatly reduced outside a P-rich patch (Linkohr et al. 2002). Transcription of the high-affinity transporters LePT1 and LePT2 in the roots is systemically repressed in split-root experiments (Liu et al. 1998). If Arabidopsis plants are resupplied after P limitation, transcript levels of high-affinity transporters and other P starvation-inducible genes decline rapidly indicating a local response (Müller et al. 2004).

The gene Mt4 from Medicago is strongly induced in the roots following phosphate starvation and is downregulated by phosphate fertilization or infection with mycorrhiza (Burleigh & Harrison 1997). As Mt4-like genes, TPSI1 (TOMATO PHOSPHATE STARVATION-INDUCED1), AtIPS1 (INDUCED BY PHOSPHATE STARVATION1), At4, and OsPI1 (PHOSPHATE LIMITATION-INDUCIBLE1) have been identified (Liu et al. 1997, Burleigh & Harrison 1999, Martín et al. 2000, Wasaki et al. 2003). In split-root plants, Mt4 is systemically repressed.

In a mutant screen for plants with impaired AtIPS1 reporter gene activity, the MYB-CC protein PHR1 (PHOSPHATE STARVATION RESPONSE1) has been identified. The phr1 mutant does not accumulate anthocyanin under P depletion and reveals reduced expression of At4, acid phosphatase (AtACP5), ribonuclease (RNS1), and phosphate transporter (AtPt1). Root hairs are not altered in the mutant (Rubio et al. 2001). PHR1 is homologous to PSR1 (PHOSPHORUS STARVATION RESPONSE1) from Chlamydomonas; both are localized in the nucleus. PSR1 level is strongly increased upon P starvation, whereas PHR1 is only moderately induced under low P conditions (Rubio et al. 2001, Wykoff et al. 1999). PHR1 binds as a dimer to an imperfect palindromic sequence in the promoter regions of AtIPS1, AtIPS3, At4, Mt4, TPSI1, AtACP5, PAP1, RNS1, and AtPT1 (Rubio et al. 2001).


The pho2 mutant overaccumulates phosphate in the leaves, while root P concentration stays normal (Delhaize & Randall 1995). The P accumulation in the shoot is a result of an increased uptake rate, increased translocation from the root to the shoot, and an impaired remobilization from the shoot to the root. Grafting experiments revealed that pho2 acts root-specific in causing the high P content of the shoot. The increased uptake rate of pho2 disappeared when the shoot was removed. Thus, PHO2 might be involved in long-distance signaling (Dong et al. 1998, Bari et al. 2006). PHO2 encodes a ubiquitin-conjugating E2 enzyme (Chiou et al. 2006, Aung et al. 2006, Bari et al. 2006).

The miRNA399 is strongly induced by low P in roots and most prominently in shoots. The primary transcripts of miRNA399 decrease rapidly after re-addition of P. This response is highly specific to P. The target of miRNA399 is the 5' UTR of an ubiquitin-conjugating E2 enzyme (UBC24), which is an allel of PHO2 (Sunkar & Zhu 2004, Fujii et al. 2005, Chiou et al. 2006, Bari et al. 2006). E2 proteins are part of the ubiquitin-proteasome pathway tagging proteins for degradation (Ciechanover 1998, Hellmann & Estelle 2002). Transgenic plants that overexpress UBC24 without the 5' UTR show less induction of AtPT1 and less inhibition of primary root growth. Lateral root density and expression of At4 and RNS1 were not altered (Fujii et al. 2005).

Overexpression of miRNA399 leads to downregulation of UBC24 /PHO2, which resembles the pho2/ubc24 phenotype (Fujii et al. 2005, Chiou et al. 2006, Aung et al. 2006).


From these results it is hypothesized that the abundance of PHO2 in high P conditions leads to a direct or indirect downregulation of high-affinity uptake systems preventing overload. Upon P starvation, PHO2 is posttranscriptionally downregulated by miRNA399 whereby repression of P uptake is alleviated. In addition, a component controlling the internal P concentration of shoot cells or the translocation from the shoot to the root seems to be modified by PHO2 (Dong et al. 1998, Chiou et al. 2006, Bari et al. 2006). PHO2 orthologs with miRNA399 binding sites were found in many angiosperms, but not in Physcomitrella patens or C. reinhardtii, indicating the system is conserved in vascular plants and may have originated during the evolution of higher plants (Bari et al. 2006).

Expression of 21 P starvation-inducible genes affected in pho2 is also altered in phr1. The high primary transcript level of miRNA399 in P-deficient plants is impaired in phr1, and the PHR1 cis-element is found upstream of the miRNA399s, placing PHR1 superior to miRNA399 and its posttranscriptional modification of PHO2 (Bari et al. 2006).

Also At4AtIPS, At4.1, and At4.2 possess miRNA399 binding sites. Thus, in addition to the strong transcriptional induction of the At4-like genes in response to P deprivation, also a posttranscriptional control by miRNA399 may occur. At4 is expressed in the endodermis and vascular tissues of P-starved roots and after prolonged P depletion also in the cortex and epidermis (Shin et al. 2006). AtIPS1 expression, in contrast, is found throughout the plant (Martín et al. 2000). Loss of at4 function impacts the allocation of P from the shoots to the roots leading to an increased P content in the shoot, but in contrast to pho2, re-allocation occurs only under P starvation and the uptake rate into the root is not altered in at4 (Shin et al. 2006).


Phosphate-deficient siz1 plants show an enhanced primary root growth inhibition, extensive lateral root and root hair development, increased root/shoot ratio, and higher anthocyanin accumulation, even though intracellular P levels are similar to the wildtype. Transcript abundance of Pht1;4 and acid phosphatase is higher in siz1; AtIPS1 and AtRNS1 are lower. AtSIZ1 is a small ubiquitin-like modifier (SUMO) E3 ligase. From its sequence an involvement in chromatin organisation and ligase activity are predicted, which is consistent with SIZ1 localization in the nucleus. A sumoylation target of SIZ1 is PHR1. Thus, SIZ1 is an important positive and negative upstream regulator of P starvation responses. SIZ1 mRNA is more abundant in roots than in shoots and is only slightly responsive to P deprivation (Miura et al. 2005).

1.2.6 Further regulators of Pstarvation responses

The phospholipase Dζ2 is induced in roots and shoots upon P starvation. The pldζ2 mutant shows reduced phospholipid hydrolysis under limiting P conditions. Although the root meristem is disorganized in pldζ2, leading to a decreased primary root length and root hairs developing closer to the apex, PLDζ2 is suggested to have a metabolic rather than a regulative function (Cruz-Ramírez et al. 2006).

The bHLH protein OsPTF1 is induced in Pdeficient roots and constitutively expressed in shoots. Overexpressing lines show higher root length and surface area resulting in increased P uptake rates in P-limited conditions. One-hundred-fifty-eight genes related to metabolism, nutrient transport, and regulation are significantly changed in the overexpressers, of which almost all have Ebox elements in their promoters as typical binding sites for bHLH transcription factors (Yi et al. 2005).


In pdr2 (P i deficiency response2), primary root growth inhibition and lateral root development is enhanced under low P conditions. The primary root and newly formed laterals subsequently lose meristem activity followed by the development of secondary and tertiary laterals, respectively, together leading to a small stunted root system. Anthocyanin and starch accumulation and P starvation-inducible gene expression are enhanced in P-deficient pdr2. Phosphate content is higher in pdr2 grown without P and lower when supplied with DNA as a phophate source. Meristem function can be restored by Phi indicating pdr2 disrupts sensing of low external P availability (Ticconi et al. 2004). The pdr2 phenotype is reminiscent of the experiment of Torrey (1950), in which removal of the primary root tip resulted in an increase of laterals. Ablation of root cap cells inhibits primary root growth while lateral root development is stimulated. This implies an involvement of the root cap cells in the signaling system that alters root system architecture (Tsugeki & Fedoroff 1999).

The lpi1 and lpi2 (low phosphorus insensitive) mutants fail to inhibit primary root growth under P starvation. Root hair density is lower, whereas root hair elongation and anthocyanin accumulation are unaffected. Expression of AtPT1, AtPT2, PAP1, ACP5, and IPS1 is reduced. lpi2 fails to increase the number of laterals in low P conditions, whereas the reaction in lpi1 is normal (Sánchez-Calderón et al. 2006).

QTL analysis of low P grown Arabidopsis RILS (Bay-0 x Shahdara) revealed three loci involved in reduction of primary root growth. The locus LPR1 is not involved in increasing root hair number or length, anthocyanin accumulation, acid phosphatase excretion, or Pht1;4 induction (Reymond et al. 2006).

1.2.7 Involvement of hormones in the regulation of phosphate homeostasis


Cytokinin favours shoot growth while inhibiting lateral root formation. A decrease in the content of cytokinin has been associated with P deficiency (Salama & Wareing 1979, Horgan & Wareing 1980). Exogenously supplied cytokinin counteracts the root growth stimulation induced by low P (Kuiper 1988, Kuiper et al. 1988). The phosphate starvation inducible genes AtIPS1, At4, AtACP5, and AtPT1 are downregulated after cytokinin treatment, and anthocyanin accumulation is decreased. However, the increase in the root hair number and length was unaffected by the hormone (Martín et al. 2000, Franco-Zorrilla et al. 2002). The cre1 (cytokinin response1) mutant displays reduced sensitivity of AtIPS1 reporter gene activity to cytokinin repression. Also the cytokinin-dependent repression of At4, AtACP5, and AtPT1 and of anthocyanin accumulation is reduced in P-deficient cre1, and lateral root formation is not inhibited (Franco-Zorrilla et al. 2002). CRE1 is allelic to the cytokinin receptor kinase WOL/AHK4 (WOODEN LEG/ARABIDOPSIS HISTIDINE KINASE4). CRE1 is downregulated by P starvation and induced by cytokinins (Mähönen et al. 2000, Inoue et al. 2001, Suzuki et al. 2001, Franco-Zorrilla et al. 2002).

Auxin treatment inhibits primary root growth and promotes lateral root initiation by activation of the cell cycle in xylem pericycle cells, which is blocked by inhibition of auxin transport (Evans et al. 1994, Casimiro et al. 2001, Himanen et al. 2002). Auxin-resistant mutants show decreased primary root growth inhibition in response to auxin and reveal a reduced number of laterals (Evans et al. 1994, Hobbie & Estelle 1995). Inhibition of acropetal auxin transport or removing the shoot also inhibits lateral root formation (Reed et al. 1998). The involvement of auxin in remodelling root system architecture adapted to P starvation is controversial. Williamson et al. (2001) suggested auxin is not involved in primary root growth reduction and increased lateral root formation of P-deficient plants, because auxin-resistant mutants reacted in the same manner as the wildtype. However, iaa28 did not respond to the stimulatory effect of low P on lateral root and root hair formation. Treatments of high and low P-grown wildtype plants with auxin and auxin antagonists resulted in an altered sensitivity of P-starved plants to the hormone with respect to primary root growth and lateral root development (López-Bucio et al. 2002). Nacry et al. (2005) measured a significantly increased auxin concentration in the primary root and short laterals of P-starved plants along with increased activity of the auxin-responsive reporter DR5-GUS. The authors concluded auxin redistribution is altered in the root of P-deficient plants. Cluster root formation in P-sufficient lupin was increased by foliar application of auxin and reduced in P-deficient plants by auxin transport inhibitors (Gilbert et al. 2000).

Auxin and ethylene promote root hair elongation, as auxin- or ethylene-resistant mutants have no or shorter root hairs; treatment of these mutants or wildtype plants with auxin or an ethylene precursor significantly increases root hair length (Wilson et al. 1990, Pitts et al. 1998). As with the role of auxin in remodelling root architecture, the role of auxin and ethylene in root hair development during adaption to the P supply is also controversial. An involvement of auxin in the root hair elongation under P deficiency is suggested by Bates and Lynch (1996), because this process was inhibited by blocking auxin transport. Schmidt and Schikora (2001) found that the auxin-resistant mutants axr1 and axr2, and the ethylene-insensitive mutant ein2 are not able to produce root hairs under sufficient nutrient supply or Fe deprivation, whereas root hairs of P-deficient plants develop normally, which is in support of inhibitor studies. These authors suggest that root hair formation in adaption to P limitation is regulated independently from root hair development in sufficient or Festarved plants and apparently does not require auxin or ethylene. Also Ma et al. (2001a) observed no significant effect of ethylene inhibitors on root hair development of P-starved plants. In contrast, Zhang et al. (2003) and He et al. (2005) observed an inhibition of P stress-induced root hairs and a reduced response of ethylene-insensitive mutants to P limitation assuming an interacting effect of ethylene and phosphate signaling on root hair development.

1.3 Root hair development


Root hairs are tubular outgrowths of root epidermal cells that serve in water and nutrient uptake. They are the primary site of infection of rhizobia. Root hairs evolved 400 million years ago within the lineage of tracheophyta, indicating an important role for root hairs in the adaptation of land plants (Peterson 1992). Different stages of root hair development are defined. These stages are root hair specification, initiation, bulge formation, and tip growth (Schiefelbein & Somerville 1990, Dolan et al. 1994).

1.3.1  Root hair specification

In the rhizodermis of crucifers like Arabidopsis, root hair cells are formed in a position dependent pattern. Only those epidermal cells located over the clefs of two underlying cortical cells develop a root hair, which is called the hair position (H position). Epidermal cells with contact to only one cortical cell become a non-hair cell, which is typically the non-hair position (N position). If the epidermis is separated from the cortex, root hairs develop in nearly all rhizodermal cells (Bünning 1951, Dolan et al. 1994). An arrangement of H and N cell files is also observed in other Brassicales as well as in Caryophyllales, Malpighiales, Rosales, Myrtales, Cornales, Ericales, Solanales, Lamiales, and Boraginaceae (Dolan 2006, Kim et al. 2006). First differences between H and N cells are visible in the late meristematic region, where H cells show a more intense cytoplasmic staining, shorter cells that undergo a higher rate of cell division, and a delay in vacuolization relative to cell elongation (Dolan et al. 1994, Galway et al. 1994, Berger et al. 1998a). Laser ablation experiments have shown that it is positional information not cell lineage that defines cell fate (Berger et al. 1998b).

Two opposing pathways determine the development of H and N cells in the late meristematic region and are assumed to act by lateral inhibition with feedback (Lee & Schiefelbein 1999, Bernhardt et al. 2005, Schiefelbein & Lee 2006). In the first pathway, a transcription factor complex specifies the non-hair cell fate in the N position. This complex includes the MYB class transcriptional regulator WER (WEREWOLF), which possesses a DNA binding as well as a putative transcriptional activation domain, two bHLH transcription factors GL3 (GLABRA3) and EGL3 (ENHANCER OF GLABRA3) that act in a partially redundant manner, and TTG1 (TRANSPARENT TESTA GLABRA1) containing a WD40 domain involved in protein-protein interaction (Galway et al. 1994, Walker et al. 1999, Lee & Schiefelbein 1999, Bernhardt et al. 2003). This complex induces the expression of the homeodomain protein GL2 (GLABRA2) in the N position as WER recognizes a MYB binding site within the GL2 promoter (Hung et al. 1998, Lee & Schiefelbein 1999, Koshino-Kimura et al. 2005). GL2 inhibits root hair development since the gl2 mutant develops root hairs in the N position (Masucci et al. 1996). The gene CAPRICE (CPC) also contains a MYB binding site and is likewise transcribed in the N position through positive regulation by the WER/TTG/GL3/EGL3 complex. CPC encodes a small MYB protein without transcriptional activation domain that is able to move from the N cells to the neighboring H cells, probably via plasmodesmata (Wada et al. 1997, 2002, Lee & Schiefelbein 2002, Kurata et al. 2005). In the H position, CPC acts as a negative regulator of WER and its own expression, and, in addition, it replaces WER in the transcription factor complex described above by binding to the bHLH proteins (Lee & Schiefelbein 1999, Koshino-Kimura et al. 2005). Because CPC lacks the transcriptional activation domain, this alternatively composed complex inhibits GL2 expression in the H position, thus specifying the hair fate. As a feedback loop, CPC induces GL3/EGL3 in the H position. The proteins are transported towards the N position, where their expression is blocked by WER. In addition, GL3 and EGL3 also inhibit their own transcription in the N position (Bernhardt et al. 2003, 2005).


Pharmacological inhibition of histone deacetylase affects the position-dependent expression of WER, CPC, GL2, and root hair pattern. Mutation of the histone deacetylase HDA18, which is expressed in all root tissues, leads to a randomized patterning of WER expression and root hair occurrence suggesting an involvement in transcriptional regulation of cell specification genes after the positional cues have been perceived (Xu et al. 2005).

Another regulation level for the expression of GL2 in the N position is chromatin remodelling. Around the GL2 locus in N cells, chromatin is in an 'open' conformation, and it is 'closed' in H cells. CPC is required to establish the 'closed' state in the H position, but neither GL2 nor cell fate specification is required for the 'open' conformation. The chromatin state is reset at mitosis and is respecified during the following G1 phase according to the underlying positional information (Costa & Shaw 2006).

The erh1 and erh3 mutants (ectopic root hair) have additional root hairs and show a denser cytoplasmic staining also in N cells (Schneider et al. 1997). In erh3, the orientation of cell walls is abnormally oblique in all tissues of the root tip. Microtubules reorganization is delayed in the erh3 allele fra2 (fragile fibre2) and cell wall biosynthesis disturbed. The activity of ERH3 is required for both, hair and non-hair cells. The expression pattern of GL2 is altered in erh3 as well as the occurrence of lateral root cap, endodermis, and cortical markers. ERH3 codes for a kataninp60 protein expressed throughout the whole plant. As kataninp60 proteins sever microtubules; ERH3 may act either directly by microtubule disruption or by a katanin-dependent cell wall biosynthetic process that incorporates molecules conferring the positional information spatially into the cell wall (Burk et al. 2001, Webb et al. 2002).


The positional information is likely to be perceived by SCM (SCRAMBLED), a receptor-like kinase expressed in all tissues of the developing root except the root cap. SCM is necessary for the position-dependent pattern of WER, CPC, and GL2. In scm mutants, root hairs are distributed randomly. The cell division rates in H and N position are altered indicating that early cell characteristics are affected. SCM is predicted to possess a signal sequence for secretion, an extracellular leucine-rich repeat typically participating in protein-protein interactions, and an intracellular kinase domain (Kwak et al. 2005).

The root hairless mutants rhl1, rhl2, and rhl3 do not reveal a denser cytoplasmic staining or delay in vacuolization in any epidermal cell in the late meristematic region. In addition, the rhl mutants are dwarf and their nucleolus is deformed. RHL1, RHL2, and RHL3 are subunits of the topoisomerase VI complex important for ploidy-dependent cell growth. Thus, endoreduplication is important for root hair development. According to double mutant analysis, this process is independent of GL2 action (Schneider et al. 1997, 1998, Sugimoto-Shirasu et al. 2005).

1.3.2 Epidermal cell polarity and root hair initiation

After specification, the rhizodermal cells elongate, and then root hairs are initiated near the apical end of the cell towards the root tip. Root hair initiation is, therefore, linked with epidermal cell polarity (Schiefelbein & Somerville 1990, Dolan et al. 1994, Grebe 2004). Establishment of cell polarity bases on vesicle trafficking. An important mediator of this process in Arabidopsis is the ArfGTPase ARF1 (ADP-RIBOSYLATION FACTOR1), which is localized in Golgi membranes and endocytic vesicles (Donaldson et al. 1992a, Xu & Scheres 2005). ArfGTPases are involved in membrane trafficking. They recruit cytosolic coat proteins to the sites of vesicle budding (Vernoud et al. 2003). In ARF1 mutants, root hairs develop more basally and additional root hairs derive from one trichoblast (Xu & Scheres 2005). Treatment with the vesicle transport inhibitor brefeldin A (BFA) causes basal shifting of root hair initiation (Grebe et al. 2002, Xu & Scheres 2005). BFA inhibits Arf GEF (ARF-guanosine exchange factor), which regulates ArfGTPase by catalyzing the guanine nucleotide exchange (Donaldson et al. 1992b, Helms & Rothman 1992). Weak mutant alleles of the Arf GEF GNOM also display a basal shift of root hair emergence (Steinmann et al. 1999, Fischer et al. 2006).


Pharmacological and mutant analyses suggest an involvement of auxin and ethylene in polar root hair initiation (Masucci & Schiefelbein 1994, Grebe et al. 2002). Auxin may, thereby, induce root hair initiation through an increased ethylene production (Cho & Cosgrove 2002). Auxin and ethylene act downstream of GL2 expression (Masucci & Schiefelbein 1996).

The auxin influx carrier AUX1 (AUXIN-RESISTANT1) is localized at the apical and basal ends of epidermal cells. This localization is abolished by BFA and, conversely, AUX1 is required for BFA-sensitive vesicle trafficking. The aux1 mutation causes a basal shift of root hair initiation (Grebe et al. 2002). Polar root hair localization is, moreover, mediated by combined activity of AUX1, EIN2, and GNOM. In the triple mutant, the auxin gradient is abolished and the root hair initiation site is displaced, namely stronger than in the single mutants. Locally applied auxin can coordinate root hair positioning (Fischer et al. 2006). Also the localization of PIN2 (PIN-FORMED2), an auxin efflux carrier, is a potential target of ARF1 action in the rhizodermis (Xu & Scheres 2005). PIN2 localizes in the late meristematic and elongation zone at the basal rhizodermal plasma membranes (Müller et al. 1998). However, pin2, pin2pin1, or pin2pin4pin7 mutants do not show alterations in root hair polarity indicating that the epidermal cell polarity does not rely on PIN function (Fischer et al. 2006).

A further component important for polar root hair initiation is RHD6. The rhd6 mutant reveals very few root hairs that emerge more basally. When treated with auxin or the ethylene precursor ACC, root hair number is increased in rhd6 (Masucci & Schiefelbein 1994).

1.3.3 Root hair initiation and bulge formation


Polar localization of RopGTPases specifies the site of root hair outgrowth. Rops are plant specific RhoGTPases that are key regulators of the actin cytoskeleton. Expression of constitutively active Rop2, Rop4, and Rop6 abolishes polar root hair formation. Rops localize at the sites of root hair emergence before bulge formation and remain at the tips of elongating root hairs until growth ceases. This localization is sensitive to BFA treatment (Zheng & Yang 2000, Molendijk et al. 2001, Jones et al. 2002, Vernoud et al. 2003). Rop2 recruitment depends on ARF1 action and on the combined activity of AUX1, EIN2, and GNOM (Xu & Scheres 2005, Fischer et al. 2006). As targets of Rop action in developing root hairs, the plasma membrane-bound NADPH oxidase complex and the actin binding protein profilin involved in actin organization have been suggested (Molendijk et al. 2001, Jones et al. 2002, Vernoud et al. 2003).

The RhoGTPase GDP dissociation inhibitor (RhoGDI) SCN1 (SUPERCENTIPEDE1) is involved in regulating the localization of Rop2. The scn1 mutant has multiple tip-growing sites that do not elongate. RhoGDIs regulate RhoGTPases by sequestering them in the cytosol and inhibiting the dissociation of GDP from the GTPases. SCN1/AtRhoGDI1 interacts with Rop4 and Rop6. This interaction is weakened in scn1 and Rop2 mislocalized (Carol et al. 2005, DerMardirossian & Bokoch 2005, Bischoff et al. 2000, Vernoud et al. 2003).

SCN1 belongs to a mechanism that focuses the production of reactive oxygen species (ROS) catalyzed by the NADPH oxidase RHD2 (ROOT HAIR DEFECTIVE2) to root hair tips (Carol et al. 2005, Foreman et al. 2003). ROS are essential for root hair elongation as they stimulate hyperpolarization-activated Ca2+ channels in the apex of growing root hairs. The resulting influx of extracellular Ca2+ leading to a tip-focussed Ca2+ gradient is required for root hair tip growth (Foreman et al. 2003, Véry & Davies 2000, Wymer et al. 1997, Schiefelbein et al. 1992). The local influx of Ca2+ does not precede bulge formation suggesting Ca2+ does not trigger root hair initiation. In the rhd2 mutant, which initiates root hair bulges that do not elongate, no ROS are produced and no tip-focussed Ca2+ gradient is established (Schiefelbein & Somerville 1990, Foreman et al. 2003, Wymer et al. 1997). Differences in the transcriptomes of rhd2 and the wildtype comprise 606 genes that are higher expressed in the wildtype and 313 genes that are higher in rhd2 (Jones et al. 2006).


ACTIN2 and is essential for bulge site selection and tip growth. The der1 (deformed root hairs1) mutation in the ACTIN2 gene causes a basal shift of root hair outgrowth and short or deformed root hairs (Ringli et al. 2002).

Polar root hair initiation is possibly regulated by phospholipid signaling, since ectopic expression of PLDζ1 leads to mislocalization of root hair initiation sites. Root hairs are deformed and are also formed in non-hair cells. PLDζ1 is repressed by GL2, a negative regulator of root hair development (Ohashi et al. 2003).

Sterol biosynthesis is required for establishment of epidermal cell polarity. Mutation of enzymes belonging to the sterol synthesis pathway, like smt1 (sterol methyltransferase1) or hydra2 defective in sterol C14 reductase, leads to a randomized root hair initiation over the apical-basal axis of trichoblasts and to the development of multiple and branched root hairs. Sterols can act in lipid rafts, which are sterol-rich regions of plasma membrane important for polarity determination (Willemsen et al. 2003, Souter et al. 2002).


The S-acyl transferase TIP1 is important for root hair initiation and tip growth (Hemsley et al. 2005, Parker et al. 2000). The tip1 mutant has short root hairs that are often branched (Schiefelbein et al. 1993, Ryan et al. 1998). S-acylation is a reversible protein modification that promotes association with membranes (Yalovsky et al. 1999). Acylation has been implicated in protein sorting into lipid rafts (Bagnat & Simons 2002). Yeast S-acyl transferase localizes to the Golgi apparatus and the late endosome (Harada et al. 2003). Potential targets of TIP1 are Rops, G proteins, phospholipases, or calcium-dependent protein kinases (CDPKs). TIP1 could regulate vesicle traffic by directing proteins to a discrete area of the membrane (Hemsley et al. 2005).

Local acidification of the cell wall is required for root hair initiation. Its prevention stops the initiation process. The acidification is present from the first morphological indications of bulge formation and maintains until tip growth begins (Bibikova et al. 1998). Expansins accumulate in the developing root hair bulges (Baluška et al. 2000). Expansins are extracellular proteins that regulate plant cell enlargement by inducing relaxation and extension of cells at an acidic pH optimum (McQueen-Mason et al. 1992, Cosgrove 2000). The expansin genes EXP7 and EXP18 are expressed in root hair cells during the initiation and elongation state and are localized at the emerging root hair tip. In rhd6, EXP7 and EXP18 expression is blocked suggesting RHD6 is a positive regulator of the two expansin genes (Cho & Cosgrove 2002). EXP7 and EXP18 expression is also lower in rhd2 (Jones et al. 2006).

Root hair initiation is accompanied by an increase in xyloglucan endotransglycosylase (XET) activity at the site of the future bulge formation, which is lowered in rhd2. As XETs cleave and rejoin xyloglucan chains, XET could locally loosen the cell wall leading to turgor-mediated bulge formation. XET localization is independent of the actin cytoskeleton (Vissenberg et al. 2001, Jones et al. 2006).

1.3.4 Root hair tip growth 


Auxin is also involved in root hair elongation. Mutation in the K+ transporter TRH1 prevents root hair tip growth, which can be rescued by the application of auxin. Also multiple initiation sites are observed in trh1. Although TRH1 has K+ transport activity, this phenotype is independent of the external K+ concentration (Rigas et al. 2001, Desbrosses et al. 2003). TRH1 accelerates auxin efflux in Arabidopsis root segments and yeast. TRH1 is expressed in the root cap and the lateral root cap. It is suggested that TRH1 assists in basipetal auxin transport in the columella providing an auxin concentration in the rhizodermis adequate for root hair development (Vincente-Agullo et al. 2004). Ectopic expression of the auxin effluxer PIN3 or of PID (PINOID), a kinase positively activating auxin efflux, in root hair cells inhibits root hair elongation underlining the importance of auxin for root hair tip growth (Lee & Cho 2006).

During root hair tip growth, vesicles containing cell wall material are secreted by exocytosis at the apical zone of the growing tip. The resulting membrane surplus is compensated by endocytosis in the subapical zone. The mechanism regulating the directional secretion to the tip must involve a continuous re-localization towards the advancing tip. This process is regulated by a network of small GTPases, Ca2+, phospholipids, ROS, protein kinases, and the cytoskeleton. Multiple feedback loops allow the growing tip to maintain itself (Hepler et al. 2001, Dolan & Davies 2004, Šamaj et al. 2004, 2006, Cole & Fowler 2006).

The microtubules cytoskeleton is important for regulating the direction of root hair tip growth. By placing proteins that serve as spatial cues to the expanding tip, microtubules may determine the site where exocytosis takes place (Bibikova et al. 1999, Sieberer et al. 2005). The actin cytoskeleton is fundamental for polarized growth during root hair development. It is responsible for the motility of vesicles and organelles (Hepler et al. 2001).


ARF1 is involved in the regulation of membrane trafficking during root hair tip growth. Mutation in the ArfGAP (ArfGTPase-activating protein) RPA (ROOT AND POLLEN ARFGAP), which regulates ARF1 by catalyzing GTP hydrolysis, leads to short and deformed root hairs. RPA is localized at the Golgi system (Song et al. 2006, Vernoud et al. 2003).

Mutation in the small G protein RHD3 causes short wavy root hairs and a reduced cell expansion also other tissues (Wang et al. 1997, Schiefelbein & Somerville 1990, Hu et al. 2003). rhd3 reveals a striking reduction in vacuole size and an abnormally high number of secretory vesicles concentrated in the subapical region of the root hairs (Galway et al. 1997). RHD3 is required for vesicle trafficking from the ER to the Golgi, ER organisation, and membrane transport from the plasma membrane to the vacuole (Zheng et al. 2004). RHD3 is essential for actin organization and cell wall biosynthesis; cell wall thickness is dramatically reduced in rhd3 (Hu et al. 2003). RHD3 acts downstream of auxin, ethylene, and RHD2 (Wang et al. 1997, Schiefelbein & Somerville 1990).

The exocyst is an oligomeric protein crucial for the specification of vesicle docking and fusion during exocytosis (Eliás et al. 2003). In the maize rth1 (roothairless1) mutant, which is defective in a homolog of the SEC3 exocyst subunit, root hair bulges fail to elongate (Wen et al. 2005). Mutation in the Arabidopsis EXO70A1 gene, a family member of another putative exocyst subunit, also stops root hair development after bulge formation. In analogy with yeast and mammals, an interaction of EXO70A1 with Rabs or Rops, e. g. Rop2, has been suggested (Synek et al. 2006).


Localization of the RabGTPase RabA4b at the tips of growing root hairs is correlated with tip growth. RabA4b is associated with vesicles distinct from the trans-Golgi network. RabA4b localization depends on the actin cytoskeleton and on RHD2, RHD3, and RHD4 action (Preuss et al. 2004). RabA4b interacts with the phosphatidylinositol-4-kinase PI-4Kβ1. PI-4Kβ1 interacts with the Ca2+ sensor AtCBL1. Thus, RabA4b recruitment of PI-4Kβ1 could result in Ca2+-dependent generation of phosphatidylinositol-4-phosphate (PIP) at a tip-localized membrane compartment (Preuss et al. 2006). The actin binding protein profilin and its ligand phosphatidylinositol-4,5-bisphosphate (PIP2) accumulate at the tips of bulges and growing root hairs (Braun et al. 1999). Installation of the actin-based tip growth machinery requires profilin action and occurs after expansin-associated bulge formation (Baluška et al. 2000).

The PI transfer protein (PITP) COW1 (CAN OF WORMS1) is important for root hair elongation. It is localized in epidermal cells and in the central cylinder of the meristematic, elongation, and root hair zone (Böhme et al. 2004). The COW1 homolog AtSfh1p enriches along the plasma membrane at the growing tip. Dysfunction impairs PIP2 accumulation at the tip, leads to Ca2+ influxes all along the root surface, causes defects in the actin cytoskeleton, and disorganizes microtubule networks in the root hair. AtSfh1p could generate PIP2 landmarks that couple to components of the actin cytoskeleton (Vincent et al. 2005, Hsuan & Cockcroft 2001). In yeast and humans, PITPs are involved in PLC and PLD signaling. PITP presents PI to PI-4kinase, thereby increasing PIP2. PIP2 is cleaved by PLC to generate IP3. In addition, PIP2 has been shown to potentiate PLD activity (Spiegel et al. 1996).

Expression of the IP3 phosphatase MRH3 and of the two predicted GPI-anchored proteins MRH4 and MRH5 is suppressed in rhd2. The proteins could act in lipid rafts, as is the case in mammals (Jones et al. 2006).


PLDs produce phosphatidic acid (PA). In Arabidopsis, PA activates the phosphoinosite-dependent kinase PDK1 by specific binding, which, in turn, stimulates AGC2, a cAMP/cGMP-dependent kinase C. AGC2 localizes to the root hair tip and to the nucleus. The agc2 mutant displays reduced root hair length (Anthony et al. 2004).

MtCDPK1 is important for root hair tip growth. In the aberrant root hairs of RNAi plants, ROS production at the tip is abolished and the actin cytoskeleton disorganized (Ivashuta et al. 2005).

In the Arabidopsis oxi1 (oxidative signal-inducible1) mutant, root hair length is reduced. OXI1 is a protein kinase induced by H2O2 that is required for the H2O2-dependent activation of the mitogen-activated protein kinases (MAPKs) MPK3 and MPK6 (Rentel et al. 2004). The alfalfa MPK6 homologue SIMK (stress-induced MAPK) shows an actin-dependent localization in the tips of growing root hairs (Šamaj et al. 2002).


Root hairs of rhd4 are short and deformed and have localized cell wall thickenings. rhd2 is epistatic to rhd4 (Schiefelbein & Somerville 1990, Galway et al. 1999).

1.3.5 Cell wall components involved in root hair tip growth

The cell wall is a crucial determinant for cell shape. Without it, the protoplast forms a sphere. It is a rigid but pliable structure that confers protection and cell cohesion and is important for communication between cells. The cell wall consists of a cellulose-xyloglucan network that is embedded in a matrix of pectic polysaccharides. During cell growth, cellulose is synthesized at the plasma membrane, whilst the other components are synthesized in the endomembrane system and released by exocytosis into the matrix (Carpita & Gibeaut 1993). Mutation in the cellulose-like synthase gene CSLD3/KOJAK (KJK) gene leads to inhibition of root hair tip growth and to leakage of cytoplasm at the root hair tip (Wang et al. 2001). CSLD3/KJK is expressed preferentially in root hair cells. Its location on the ER indicates the gene is required for the synthesis of noncellulose wall material. kjk is epistatic to cow1 and acts independent of tip1, rhd2, rhd3, and rhd4 (Favery et al. 2001).

The rhd1/reb1 (root hair defective1/root epidermal bulger1) mutant develops large bulges at the base of the root hairs; root hair length is normal. RHD1/REB1 encodes an UDP-D-glucose 4-epimerase that is required for the galactosylation of xyloglucan and type II arabinogalactan, which are components of the pectin matrix and of cell wall proteins (Schiefelbein & Somerville 1990, Seifert et al. 2002).


Structural proteins form an independent structure-determining network within the extracellular matrix that assists in proper cell wall assembly (Cassab 1998). Hydroxyproline-rich glycoproteins of the extensin and arabinogalactan protein (AGP) families representing cell wall structural proteins amount a large proportion among the genes, whose expression is lowered in rhd2 (Jones et al. 2006). The two extensin-like genes Dif10 und Dif54 are specifically expressed in root hairs (Bucher et al. 1997). The proline-rich cell wall structural protein PRP3 is expressed in root hair cells of the root hair zone, which depends on correct localization of GL2 expression and on GL2 function (Bernhardt & Tierney 2000). AGP30 is induced in atrichoblasts of the late meristematic region. This pattern does not depend on localization and function of GL2. In the root hair zone, AGP30 appears in the cortex, endodermis, and vasculature. Expression of AGP30 is regulated by abscisic acid and ethylene. The complementary expression pattern of PRP3 and AGP30 indicates a role of the two proteins in determining the extracellular matrix structure specifically for hair and non-hair cells (van Hengel et al. 2004).

A possible regulator of proper cell wall assembly is the chimeric leucine-rich repeat/extensin protein LRX1 localized in the wall of root hairs. The lrx1 mutant develops root hairs that frequently abort, swell, or branch due to an aberrant cell wall structure. LRX are involved in protein-protein or ligand-protein interaction and might play a role in connecting the cell wall with the plasma membrane by anchoring target proteins (Baumberger et al. 2001, Cassab 1998). An enhancer of lrx1, enl1, is affected in the ACTIN2 gene (Diet et al. 2004). rhd2 is epistatic to lrx1 and LRX1 expression lowered in rhd2. tip1 is also epistatic to lrx1, whereas rhd3 and rhd4 act parallel (Baumberger et al. 2003, Jones et al. 2006). The repressor of the lrx1 mutation, rol1, compensates for the absence of LRX1. ROL1 is involved in rhamnose biosynthesis, a major component of pectin, suggesting structural changes in the cell wall cause suppression of the lrx1 mutant phenotype (Diet et al. 2006). EXP7, EXP18, CSLD/KJK PRP3, LRX1, and LRX2 contain a root hair-specific cis-element (RHE), which is conserved across angiosperms. This transcriptional module acts downstream of cell-fate determining pathways regulating the position-dependent root hair development like in Arabidopsis or the position–independent root hair development occurring in other species (Kim et al. 2006).

1.4 Aim of the work

Both iron and phosphate deficiency lead to an increase in the number of root hairs. First evidence for a differential regulation of iron- and phosphate-related root hair development was provided by Schmidt and Schikora (2001). In this work, an examination was conducted to identify further differences that may exist between root hairs of iron- and phosphate-deficient Arabidopsis plants. To this end, different approaches were used.


To find out, if the iron- or phosphate-sensitive root hair development was under local or systemic control, split root experiments were conducted that were combined with a sufficient or deficient shoot.

To determine, which stage of root hair development was influenced by iron and phosphate, mutants with defects in different stages of the root hair developmental pathway were investigated for their root hair patterning and root hair phenotype.

Finally, to identify potentially novel genes involved in root hair formation in adaptation to P starvation, mutants were screened that did not develop root hairs under P-deficient conditions but appeared normal, when the plants were transferred to P-sufficient medium. The phenotype of one mutant was examined in further detail and the mutated locus was isolated by map-based cloning.

© Die inhaltliche Zusammenstellung und Aufmachung dieser Publikation sowie die elektronische Verarbeitung sind urheberrechtlich geschützt. Jede Verwertung, die nicht ausdrücklich vom Urheberrechtsgesetz zugelassen ist, bedarf der vorherigen Zustimmung. Das gilt insbesondere für die Vervielfältigung, die Bearbeitung und Einspeicherung und Verarbeitung in elektronische Systeme.
DiML DTD Version 4.0Zertifizierter Dokumentenserver
der Humboldt-Universität zu Berlin
HTML generated: