Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
Current address: Universidade Federal do Rio Grande do Sul, Departmento de Plantas de Lavoura, cp 15100, Porto Alegre, RS, Brazil.
These two authors share first authorship.
Plants have evolved complex strategies to maintain phosphate (Pi) homeostasis and to maximize Pi acquisition when the macronutrient is limiting. Adjustment of root system architecture via changes in meristem initiation and activity is integral to the acclimation process. However, the mechanisms that monitor external Pi status and interpret the nutritional signal remain to be elucidated. Here, we present evidence that the Pi deficiency response, pdr2, mutation disrupts local Pi sensing. The sensitivity and amplitude of metabolic Pi-starvation responses, such as Pi-responsive gene expression or accumulation of anthocyanins and starch, are enhanced in pdr2 seedlings. However, the most conspicuous alteration of pdr2 is a conditional short-root phenotype that is specific for Pi deficiency and caused by selective inhibition of root cell division followed by cell death below a threshold concentration of about 0.1 mm external Pi. Measurements of general Pi uptake and of total phosphorus (P) in root tips exclude a defect in high-affinity Pi acquisition. Rescue of root meristem activity in Pi-starved pdr2 by phosphite (Phi), a non-metabolizable Pi analog, and divided-root experiments suggest that pdr2 disrupts sensing of low external Pi availability. Thus, PDR2 is proposed to function at a Pi-sensitive checkpoint in root development, which monitors environmental Pi status, maintains and fine-tunes meristematic activity, and finally adjusts root system architecture to maximize Pi acquisition.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Inorganic phosphate (Pi), the fully oxidized and assimilated form of phosphorus (P), plays a pivotal structural and regulatory role in general metabolism and at the nexus of photosynthesis and carbon allocation. Therefore, environmental Pi availability has a direct impact on plant function and productivity. Although P is abundant in the lithosphere, the soil chemistry of Pi renders the element the second least available macronutrient, after nitrogen, in many natural and agricultural ecosystems (Schachtman et al., 1998). Plants have evolved sophisticated metabolic and developmental strategies to conserve Pi and to maximize its acquisition from the rhizosphere when Pi is limiting. Examples of biochemical adjustments include bypass reactions of adenylate- and Pi-dependent steps in respiratory pathways (Plaxton and Carswell, 1999), changes of membrane lipid composition (Härtel and Benning, 2000; Yu et al., 2002), enhanced expression of high-affinity Pi transporters as well as secretion of enzymes and organic acids to increase Pi accessibility from insoluble complexes and organophosphates in recalcitrant soil matter (Jones, 1998; Raghothama, 1999; Schachtman et al., 1998). Remodeling of root system architecture, increased root hair formation, and associations with symbiotic mycorrhizal fungi to accelerate soil exploration are typical developmental responses to low Pi (Harrison, 1999; López-Bucio et al., 2003; Ma et al., 2001; Williamson et al., 2001). Despite numerous physiological studies on acclimatory responses to Pi limitation, little is known about the underlying molecular processes or regulatory genes that are involved in the Pi-starvation response of plants.
In bacteria and fungi, phosphate (pho) regulons comprising more than 20 genes with functions in Pi acquisition and recycling are activated by low Pi availability (Ogawa et al., 2000; Wanner, 1996). Several components of an analogous plant pho system have been isolated and functionally characterized (for review, see Abel et al., 2002), and global gene expression studies in Arabidopsis thaliana have expanded the catalog of Pi-responsive genes (Hammond et al., 2003; Wu et al., 2003). However, the Pi sensory mechanisms and the components of signal transduction pathway(s) that integrate plant responses to Pi deficiency remain to be identified. Early genetic screens in Arabidopsis for aberrant Pi accumulation identified the pho1 and pho2 mutants. Pi translocation from root epidermal and cortical cells to the xylem is affected in pho1 (Poirier et al., 1991), which was shown to affect a novel protein involved in ion transport (Hamburger et al., 2002). Conversely, the pho2 mutant accumulates Pi to toxic levels in the shoot, suggesting inactivation of a gene product that may sense and regulate shoot Pi status (Dong et al., 1998). In an attempt to dissect Pi-starvation responses, phosphatase-underproducer (pup)1 and pho3 were isolated based on reduced acid phosphatase activity in roots of Pi-starved plants. While pup1 affects an acid phosphatase isoenzyme (Trull and Deikman, 1998), the characteristics of pho3 plants suggest the lack of a signaling component or of a Pi transporter (Zakhleniuk et al., 2001). Using an Arabidopsis line transgenic for a Pi starvation-inducible β-glucuronidase (GUS) reporter gene, Rubio et al. (2001) identified the phr1 mutant that showed reduced responsiveness of the reporter gene and several other Pi-regulated genes to Pi limitation. phosphate starvation response (PHR)1 encodes a transcription factor with homology to phosphorous starvation response (PSR)1 from Chlamydomonas reinhardtii (Wykoff et al., 1999) and regulates a subset of Pi-starvation responses (Rubio et al., 2001). A similar approach identified a cytokinin receptor, CYTOKININ RESISTANT 1 (CRE1) as a necessary component for the repression of Pi starvation-inducible genes by cytokinin (Franco-Zorrilla et al., 2002).
To elucidate Pi starvation-signaling pathway(s) in plants, we developed complementary procedures for isolating insensitive and constitutive Pi starvation-response mutants (Chen et al., 2000; Ticconi et al., 2001). Our general approach is based on the facultative ability of Arabidopsis to utilize organophosphates such as nucleic acids in Pi-limiting conditions. We previously isolated several conditional mutants that vigorously grow on high-Pi medium, but which show a slow-growth phenotype on medium containing DNA or RNA as the only source of P (Chen et al., 2000). Our screen scores a simple growth response to Pi limitation rather than a specific Pi starvation-inducible enzyme or reporter gene activity and is thus more inclusive and expected to identify molecular components involved in mediating both biochemical and developmental responses to low Pi availability. Here, we present a detailed characterization of a conditional phosphate deficiency response, pdr2, mutant that displays hypersensitive responses to Pi limitation, most notably Pi concentration-dependent inhibition of cell division in developing root meristems. Thus, the pdr2 mutation is likely to disrupt a Pi-sensing pathway necessary for the adaptation of root system architecture to Pi limitation.
Genetic analysis and mapping of pdr2
Genetic analysis of the pdr2 mutation revealed that the short-root phenotype on +DNA/−Pi medium segregated as a single recessive Mendelian trait. All F1 progeny of the backcross showed the wild-type (long root) phenotype on +DNA/−Pi medium. While the long-root phenotype was observed for 311 F2 progeny, 100 F2 plants developed short roots (segregation ratio of 3.1 : 1; χ2 = 0.098). The PDR2 locus was genetically mapped to a 2.8-cM interval on the upper arm of chromosome 5 between markers CIW8 and CER456575. This region does not contain experimentally characterized genes involved in Pi-starvation responses.
Altered root system architecture of pdr2 in low Pi
Remodeling of root system architecture is a typical developmental response to Pi limitation (López-Bucio et al., 2003). We therefore examined growth parameters of wild-type and pdr2 seedlings in media with different Pi supply (Figure 1). When compared with +Pi seedlings, total FW of the wild type was reduced by 55% on −Pi agar but only by 30% on +DNA/−Pi medium, suggesting that wild-type plants use external DNA as a source of P. Relative to the wild type, the total FW of pdr2 seedlings was 40% higher on +Pi medium but not significantly different on −Pi agar. However, on +DNA/−Pi medium, the FW of pdr2 was 45% less than that of the wild type (Figure 1a). The shoot-to-root ratio of wild-type seedlings grown on −Pi or +DNA/−Pi medium was reduced by 70% relative to Pi-sufficient plants, whereas a less pronounced reduction (40%) was observed for pdr2 in identical conditions (Figure 1b). Primary roots of wild-type plants were 65 and 30% shorter in −Pi and +DNA/−Pi conditions, but developed about four- and twofold more visible lateral roots, respectively, than in +Pi medium. Total wild-type root length was only reduced by 15% in −Pi and +DNA/−Pi medium when compared with the +Pi condition (Figure 1c–e). The pdr2 mutant differed strikingly from the wild type. The primary root axis of pdr2 was 50 and 70% shorter in −Pi and +DNA/−Pi conditions, respectively. The changes in root development were also reflected by total root length, which was significantly reduced for pdr2 on −Pi (by 80%) and +DNA/−Pi medium (by 60%). Although the number of visible lateral roots on pdr2 was unaffected by Pi status, the density of lateral root meristems, as determined by histochemical cycB1::GUS transgene expression, was increased by more than four- and twofold in −Pi and +DNA/−Pi conditions, respectively. These data demonstrate an altered response of pdr2 root development to Pi limitation (Figure 1c–f).
pdr2 is conditionally P deficient
Accumulation of transitory starch and anthocyanins are typical biochemical responses to Pi deficiency. We visualized starch accumulation in whole seedlings and observed intense iodine staining in Pi-starved wild-type and pdr2 seedlings (Figure S1). Starch accumulation in the wild type was only detectable at low external Pi (0.05 mm). The pdr2 mutant was hypersensitive and showed appreciable staining at 0.1 mm Pi. On +DNA/−Pi medium, wild-type plants did not stain for starch whereas a strong staining was evident for pdr2, suggesting conditional Pi deficiency. Measurement of extractable anthocyanins in shoots of +Pi-cultivated wild-type and pdr2 plants revealed threefold higher concentrations in the mutant seedlings (Table S1). Both genotypes dramatically increased anthocyanin levels in −Pi. In +DNA/−Pi medium, pdr2 seedlings contained approximately 10-fold more anthocyanins than the wild-type control, again suggesting conditional Pi deficiency.
We determined free Pi and total P content of roots and shoots. No significant differences were observed between wild-type and pdr2 seedlings on +Pi medium. However, in −Pi medium, pdr2 seedlings had significantly higher total P concentrations than wild-type plants; free Pi was slightly but not significantly elevated in pdr2 (Table S1). On +DNA/−Pi medium, roots and shoots of pdr2 contained 30–50% lower concentrations of free Pi and total P than the wild-type controls. These data demonstrate limited Pi acquisition by pdr2 in the +DNA/−Pi condition.
pdr2 does not inhibit Pi starvation-inducible gene expression
The slow-root growth phenotype and insufficient Pi status of pdr2 seedlings on nucleic acid-containing media suggested an inability to effectively metabolize the DNA or RNA supplement. We therefore examined expression of selected Pi starvation-inducible genes. As expected for the wild type, Pi limitation considerably increased steady-state transcript levels of At4, ribonuclease (RNS)2, At acid phosphatase (ACP)5, and At phosphate transporter (PT)2 in roots and, to a lesser extent, in shoots, relative to +Pi plants. Surprisingly, expression of these genes was not inhibited but was modestly increased (1.3–2.0-fold) in Pi-starved pdr2 seedlings (Figure 2). We also compared specific ribonuclease and acid phosphatase activities in protein extracts prepared from shoots, roots, and the hydroponic culture medium of wild-type and pdr2 plants grown in +Pi and −Pi. Consistent with the above data, we observed a similar increase of both specific enzyme activities by Pi limitation (two- to threefold) in wild-type and pdr2 plants (data not shown).
Altered sensitivity of pdr2 to external Pi
We next examined the nutrient specificity of the pdr2 root phenotype. Primary roots of wild-type seedlings grown on phosphate-deficient (−Pi), sulfur-deficient (−S), or iron-deficient (−Fe) medium were 40–60% shorter than those grown on complete medium, while no difference was observed between wild-type seedlings germinated in +N or nitrogen-deficient (−N) conditions (Figure 3a). Primary root growth of pdr2 seedlings was similar to those for the wild type on −N, −S, and −Fe medium, but was reduced by more than 85% in −Pi, indicating specificity of the pdr2 phenotype for Pi deficiency.
As the conditional root phenotype of pdr2 is independent of external nucleic acid hydrolysis but specific for Pi deficiency, we measured primary root extension rates in relation to increasing external Pi concentrations. We observed a sigmoidal dose–response curve for wild-type plants, which reached a plateau at 1 mm Pi (Figure 3b). Interestingly, the Pi dose–response curve of pdr2 was strikingly biphasic. Root growth was largely inhibited at and below 0.1 mm Pi, but attained rates similar to the wild type above this threshold.
As root growth of pdr2 is strongly inhibited by low Pi concentrations that induce high-affinity Pi transporters (Raghothama, 1999), we measured Pi uptake rates in low (0.01 mm) and high (1 mm) Pi conditions. The Pi uptake rates of wild-type and pdr2 roots were similar in both media, suggesting that pdr2 plants are not defective in Pi acquisition (Figure 3c).
pdr2 reduces root cell elongation in low Pi
The Pi-dependent root growth phenotype of pdr2 suggested impaired cell elongation or cell division (or both) of the primary root axis in response to low Pi. Relative to +Pi, the length of differentiated epidermal root cells of wild-type seedlings was reduced by about 25% in the −Pi condition and, to a lesser extent, in 0.1 mm Pi and +DNA/−Pi medium (Table S2). When compared with wild-type roots in −Pi and +DNA/−Pi conditions, root cell length of pdr2 seedlings was reduced by about 50%, which indicates that impaired root cell elongation contributes, at least partially, to the pdr2 phenotype in response to low Pi.
pdr2 inhibits root cell division in low Pi
To evaluate the contribution of cell division, we crossed pdr2 with a transgenic line that expresses a labile GUS reporter enzyme under control of the cyclin B1 promoter. Only dividing cells at the G2/M transition of the cell cycle accumulate the colored GUS reaction product (Colón-Carmona et al., 1999). For transgenic cycB1::GUS wild-type seedlings, single cells of the primary and lateral root meristems stained for GUS activity when plants were cultivated on +Pi, −Pi, and +DNA/−Pi medium. Interestingly, while single cells of the primary and lateral root meristems of pdr2 plants expressed GUS activity in the +Pi condition, only the lateral root meristems of pdr2 seedlings revealed presence of dividing cells on −Pi and +DNA/−Pi medium (Figure 4a,b). GUS staining in primary root meristems of pdr2 seedlings was very similar to wild-type seedlings when grown at or above a threshold concentration of 0.1 mm external Pi. However, at lower Pi concentrations, GUS staining was either not detectable (0.025 mm Pi) or restricted to only a few cells (0.05 mm Pi) in the primary root tip, which appeared to be less vigorous than the main root tip of wild-type plants (Figure 4b). Absence of dividing cells in the primary root meristem of pdr2 seedlings was only observed in Pi-deficient but not in N-, S-, or Fe-deficient media (Figure 4b). GUS expression in the shoot apical meristem of pdr2 was comparable to the wild type in all Pi conditions (data not shown).
We further examined GUS expression in meristems of emerging primary roots during seed germination in low Pi (Figure 4c). Wild-type primary root meristems revealed dividing cells at every stage examined (2–11 days after stratification (d.a.s.)), whereas GUS expression in the main root tips of pdr2 was detectable only until day 3. At later stages, the primary root tip swelled, sometimes curled, and lost GUS expression. Subsequent transfer to +Pi medium did not recover primary root growth in pdr2 seedlings but stimulated development of secondary roots. Return of pdr2 plants to −Pi medium caused lateral root tip curling and loss of GUS expression, but promoted formation of tertiary roots, which again expressed the GUS reporter (Figure 4c). The irreversible loss of GUS expression in primary roots of Pi-starved pdr2 seedlings was accompanied by a loss of root cell viability (Figure 4d). After 11 days of continuous growth on −Pi medium, pdr2 seedlings developed a highly branched but truncated root system, which was in striking contrast to the wild type. Interestingly, cycB1::GUS expression was only observed in the most recently developed (quaternary) root meristems, but not in the primary root tip or secondary and tertiary root meristems (Figure 4e). These data suggest the existence of a switch in late root development when roots begin to sense environmental Pi availability and a role for PDR2 in sensing local rather than systemic Pi status.
Additional evidence for such a local response of root cell division was provided by a split-root experiment. Wild-type and pdr2 seeds were first germinated for 5 days on +Pi medium and then transferred to plates containing a +Pi and a −Pi patch such that the upper or lower half of the main root axis was exposed to high or low external Pi availability (Figure 5). After growth for an additional 6 days, wild-type seedlings showed increased lateral root proliferation on the main root portion that was exposed to −Pi, but reduced lateral root formation on the part of the root system that was in contact with the +Pi patch (Figure 5). The root system architecture of pdr2 seedlings, however, was conspicuously different. Lateral root formation increased on the +Pi patches, whereas primary and lateral root growth was appreciably inhibited in the −Pi condition. As expected, root growth inhibition in the −Pi patch was accompanied by loss of cycB1::GUS expression in pdr2 (Figure 5) but not in wild-type roots (data not shown).
pdr2 is defective in Pi sensing
We examined the relationship between cycB1::GUS expression, cell viability, and total P content in primary root meristems after transfer from high to low Pi (Figure 6). Wild-type and pdr2 seeds homozygous for the cycB1::GUS transgene were germinated for 5 days on +Pi agar, transferred to −Pi medium, and grown for up to five additional days. Roots were histochemically stained for GUS expression, and GUS-positive cells were counted. Replicate plant samples were analyzed for root cell viability and processed for measurement of total P content of primary root tips (0.5-mm sections) by inductively coupled plasma (ICP)-MS. The number of GUS-positive cells in primary meristems of wild-type roots gradually decreased to 85% (day 3) and 40% (day 5) after transfer to −Pi (Figure 6a,b). Viability of primary root tips of wild-type seedlings was not affected by the −Pi condition (Figure 6c). On the contrary, the number of GUS-positive cells in pdr2 primary root meristems declined rapidly to less than 5% on day 3 after transfer. A significant loss of GUS-expressing cells was already evident on day 1 (65% remaining), and no GUS expression was detectable by day 4. Root tips of pdr2 seedlings remained metabolically active for about 2 days after transfer to −Pi but lost viability between day 2 and day 3, indicating that reduction of GUS-positive cells precedes the loss of overall root tip vitality. Interestingly, secondary meristems, which emerged on pdr2 roots after the primary meristem had died (see day 3), lost viability subsequently (see day 5). ICP-MS measurement of total P in primary root tips did not reveal significant changes in total P content or in the P/magnesium (Mg) ratio between the wild type and pdr2 for up to 3 days after transfer (Figure 6d). This data suggests that loss of GUS expression and of primary root meristem viability is not caused by local nutritional P deficiency but rather by defective Pi sensing.
As cultivation of pdr2 on −Pi medium caused accelerated development of lateral root meristems, we followed the initiation of secondary and tertiary roots by histochemical analysis of cycB1::GUS expression. On day 2 after transfer to −Pi, pdr2 roots initiated about two times as many secondary meristems as wild-type roots (Figure S2a). Initiation of tertiary root meristems was negligible on day 2, but increased more than 10-fold for pdr2 on the next day, while no such an increase was observed for wild-type roots (Figure S2b). In summary, these data suggest a pattern of development and subsequent abortion of primary and lateral roots of pdr2 in low Pi condition, which is consistent with a developmental switch that leads to competency for monitoring environmental Pi availability.
Phosphite activates cyclinB1::GUS expression
We previously demonstrated that phosphite (Phi) can substitute for Pi in selectively repressing Pi-starvation responses of Arabidopsis, but fails to substitute for Pi as a nutritional source of P (Ticconi et al., 2001). As the metabolically inaccessible Phi is likely to interfere with Pi signaling (Carswell et al., 1997), we compared the effect of increasing Pi and Phi concentrations on cycB1::GUS expression in wild-type and pdr2 roots (Figure 7). Although Phi did not support seedling growth in −Pi (data not shown; see Ticconi et al., 2001), we observed a similar dose–response curve to increasing Pi and Phi concentrations of cycB1::GUS expression in wild-type and pdr2 primary roots. Interestingly, in Pi-starved pdr2 seedlings, activation of cycB1::GUS by increasing Phi correlated with restoration of meristem function. The main root rapidly resumed growth after transfer to +Pi, provided the initial −Pi medium contained at least 0.1 mm Phi (data not shown). Thus, the ability of Phi to activate cell division and to suppress root meristem death, which is best revealed in the pdr2 mutant background, strongly suggests that a Pi-sensing mechanism and not metabolic Pi availability alone regulates both processes.
Hormonal responses of pdr2 root growth
We examined whether the application of auxin, cytokinin, gibberellic and abscisic acid, and of the ethylene precursor, 1-amino-cyclopropane-1-carboxylic acid (ACC), could rescue or mimic the pdr2 phenotype on +DNA/−Pi or +Pi medium, respectively. None of the plant hormones tested rescued pdr2 seedlings on +DNA/−Pi medium, and growth of the primary root axis of wild-type and pdr2 plants showed the same sensitivity to increasing hormone concentrations (10−8 to 10−4m) on +Pi medium (data not shown). We also observed similar hormone dose–response curves for wild-type and pdr2 seedlings when inhibition of cell division in primary root meristems of transgenic cycB1::GUS seedlings were examined (data not shown).
In contrast to our understanding of Pi-starvation responses and their regulation in unicellular organisms (Lenburg and O'Shea, 1996; Wanner, 1996), little is known about the mechanisms that integrate the more sophisticated adjustments of vascular plants to fluctuating Pi supply (Abel et al., 2002; Plaxton and Carswell, 1999; Raghothama, 2000). Developmental plasticity of the root system is essential for effective soil exploration. As the physico-chemical properties of Pi often render the element the limiting macronutrient, it is not surprising that Pi availability profoundly affects root system architecture. Although a comprehensive understanding of the genetic and environmental factors that determine root development is of obvious agronomic importance, the developmental mechanisms by which plants control initiation and maintenance of root meristem activity in response to environmental signals remain to be elucidated. In an attempt to dissect external Pi sensing in vascular plants, we previously developed a conditional genetic screen that favors isolation of signaling mutants (Chen et al., 2000). In this study, we present a detailed characterization of the pdr2 mutation, which alters sensing of local Pi availability.
pdr2 enhances the sensitivity of Pi-starvation responses
Analysis of Pi-starvation responses indicates that the pdr2 mutation causes hypersensitivity to low Pi availability. Typical biochemical responses include increased synthesis of starch and anthocyanins, presumably to adjust chloroplastidic Pi supply and photosynthetic light reactions to the Pi-dependent Calvin cycle. The pdr2 mutant accumulates more starch and anthocyanins than the wild-type control in Pi-limiting conditions (Figure S1; Table S1). Likewise, steady-state mRNA levels of Pi starvation-inducible genes, such as nucleases and high-affinity Pi transporters, are moderately elevated in Pi-starved pdr2 plants, which correlate with higher total P content of pdr2 seedlings relative to the wild type (Figure 2;Table S1).
Developmental responses of root system architecture to Pi availability have recently been studied in Arabidopsis. Pi limitation favors lateral over primary root growth by attenuating primary root proliferation via decreased cell division and elongation, and by promoting secondary root development via increased lateral root density and length (Linkohr et al., 2002; López-Bucio et al., 2002; Williamson et al., 2001). Such an adjustment of root system architecture is believed to maximize Pi acquisition as the macronutrient becomes more limiting with increasing soil depth. Growth of the primary root axis is dramatically reduced below 0.1 mm Pi for pdr2 but is indistinguishable from that of the wild type above this threshold. The short-root phenotype of pdr2 is specific for Pi deficiency (Figure 3a,b). As pdr2 plants are not defective in Pi starvation-inducible expression and secretion of nucleolytic enzymes, the short-root phenotype on nucleic acid-containing media can be explained by a low external steady-state concentration of Pi, which is well below 0.1 mm when Arabidopsis is cultivated in liquid +DNA/−Pi medium (Chen et al., 2000). Interestingly, the conditional phenotype of pdr2 is caused by inhibition of primary root cell division and subsequent loss of meristematic activity, which also is accompanied by reduced cell elongation in the differentiation zone (Figure 4; Table S2). In addition to an augmented sensitivity of primary root growth inhibition by low Pi, a higher frequency of lateral root initiation is evident for pdr2 seedlings in limiting Pi. However, as pdr2 plants fail to maintain meristem function after lateral root emergence in low Pi, the number of visible lateral roots and total root length are dramatically reduced. Consequently, the shoot-to-root ratio of pdr2 is elevated relative to the wild type (Figure 1). Thus, the altered root system architecture of pdr2 seedlings in low Pi can be explained by a hypersensitive and unrestrained inhibition of root cell division in response to Pi limitation.
pdr2 indicates surveillance of environmental Pi availability
Our data suggest that inhibition of mitotic activity in pdr2 root meristems by low Pi is caused by a defect in external Pi sensing rather than by metabolic Pi limitation. Although we cannot entirely rule out the possibility that specific cells of the root meristem are Pi deficient and thus responsible for the pdr2 phenotype, it is unlikely that PDR2 encodes a component required for high-affinity Pi acquisition: (i) the Pi content and uptake rate of Pi-starved pdr2 roots are at least as high as those measured for wild-type controls (Table S1; Figure 3c); (ii) ICP-MS measurements indicated no significant changes in total P content of primary root meristems of wild-type and pdr2 seedlings for at least 3 days after transfer from +Pi to −Pi; however, a significant decrease in the number of mitotic cells was evident for pdr2 as early as 1 day after transfer (Figure 6); (iii) loss-of-function mutations of genes on chromosome 5 coding for high-affinity Pi transporters (Pht1; 1–3) do not cause slow growth on +DNA/−Pi medium (C.A. Ticconi, S. Abel, unpublished results); and (iv) the tissue-specific expression domains of these Pi transporters are either considerably broader than the primary root tip or exclude this region (Karthikeyan et al., 2002; Mudge et al., 2002). Furthermore, meristem activity of pdr2 roots developing in low Pi can be maintained if the −Pi medium is supplemented with Phi (>0.1 mm; Figure 7), a non-metabolizable Pi analog (Plaxton and Carswell, 1999). As Phi mimics molecular responses to Pi (Carswell et al., 1997; Ticconi et al., 2001; Varadarajan et al., 2002), this observation suggests operation of a Pi-sensing pathway in root development that monitors local Pi availability. Previous studies with wild-type Arabidopsis plants and the Pi-overaccumulating pho2 mutant suggested that primary root meristems are sensitive to shoot Pi status and possibly external Pi availability (Linkohr et al., 2002; Williamson et al., 2001). However, our studies of the pdr2 phenotype provide compelling evidence for the hypothesis that root meristematic activity is highly sensitive to local Pi availability rather than to systemic Pi status (Figures 4 and 5).
Local Pi availability regulates root meristem maintenance
The pdr2 mutation unmasks a Pi-sensitive checkpoint in late root development, which monitors environmental Pi status and fine-tunes meristem activity to adjust root system architecture. In the wild type, such a checkpoint is suggested by the progressive reduction of cycB1::GUS-expressing cells in the primary root meristem after transfer of seedlings from +Pi to −Pi (Figure 6a,b). Prolonged cultivation in −Pi does not lead to a complete loss of meristem activity because primary wild-type roots rapidly accelerate growth upon transfer to +Pi (Chen et al., 2000). The ability to adjust meristem activity and to eventually maintain a basal level of meristem function in low Pi is abolished in primary and lateral roots of pdr2 plants, suggesting that PDR2 is necessary for maintenance of root meristem function when external Pi is limiting. Early lateral root development, which may recapitulate embryonic and primary root formation, can be separated into five major stages: initiation, establishment, and emergence of lateral root primordia, followed by meristem activation and maintenance (Malamy and Benfey, 1997). Experimental evidence supports a role for auxin during most of these processes, as well as for abscisic acid, which regulates root developmental checkpoints in response to nitrate and water availability (Casimiro et al., 2003; López-Molina et al., 2001). The pdr2 phenotype in low Pi is strikingly similar to the auxin-conditional alf3 root phenotype (Celenza et al., 1995). The aberrant lateral root formation (alf3) mutation causes primary root growth arrest and death, followed by increased formation of lateral root primordia, which fail to mature and die subsequently. The alf3 mutant can be rescued by auxin; however, lateral roots develop de novo, which arrest growth upon return to auxin-free medium. Thus, alf3 reveals an essential role for auxin during the late stage of lateral root development when continued maintenance of meristem function is required (Celenza et al., 1995). A role for auxin in adjusting root system architecture in response to low Pi availability remains unclear. While Pi limitation was shown to increase the sensitivity of roots to auxin with respect to primary root growth inhibition and enhanced lateral root formation (López-Bucio et al., 2002), analysis of root system architecture in several auxin-signaling mutants revealed normal responses to changes in Pi availability (López-Bucio et al., 2002; Williamson et al., 2001). As auxin fails to rescue or mimic the pdr2 root phenotype, and auxin sensitivity of primary root growth is not altered in pdr2 seedlings, it is unlikely that auxin is directly involved in the response of root meristems to low external Pi availability. The evidence rather suggests that a Pi-specific nutrient-sensing pathway reports local Pi status and modulates mitotic activity downstream of auxin action in meristem maintenance.
PDR2 functions in a response pathway to low Pi
Our analysis of the conditional pdr2 phenotype evokes at least two hypotheses on PDR2 function. First, PDR2 may encode a general repressor that attenuates multiple biochemical and developmental responses to Pi starvation. However, this scenario is unlikely because PDR2 functions as part of a local rather than systemic response to Pi limitation, which is supported by the fact that shoots of pdr2 grow normally. In an alternative hypothesis, PDR2 may function as a root-specific factor required for modulating meristem activity and adjusting root system architecture to low external Pi availability. In the latter case, the arrested and truncated root system of Pi-starved pdr2 seedlings causes compensatory responses to meet the Pi demand of the proportionally larger shoot. Such secondary responses include an augmented induction of Pi starvation-inducible genes and slightly higher accumulation of Pi and total P in Pi-starved pdr2 roots relative to wild-type seedlings. Our data further suggest that maintenance of root meristem activity is regulated by at least two Pi-sensing mechanisms. A low-affinity sensor is operational in high-Pi condition and promotes nearly maximal proliferation of the primary root. In Pi-limiting condition, however, a high-affinity sensor that monitors the lower range (<0.1 mm) of external Pi availability becomes prominent. The information on insufficient environmental Pi status is transmitted to the meristematic cells to attenuate the rate of cell division and thus primary root growth, which is followed by accelerated lateral root formation. PDR2 may function in such a response pathway to Pi limitation. The mechanistic links between the perception of environmental cues and the core cell cycle machinery remain to be established (Potuschak and Doerner, 2001). Cloning and molecular characterization of PDR2 will not only provide the tools to test the above hypotheses but also holds the promise to provide fundamental insight into how local Pi availability and other external signals regulate cell cycle progression.
Wild-type A. thaliana (L.) Heynh. (accessions Columbia-0 (Col-0) and Landsberg erecta (Ler)) were obtained from the Arabidopsis Biological Resource Center (Ohio State University, Columbus, OH, USA). The pdr2 mutant was derived from ethyl methanesulfonate-mutagenized populations of A. thaliana (Col-0) and was previously referred to as line p88 (Chen et al., 2000). Seeds of transgenic cycB1::GUS Arabidopsis plants (Col-0) were obtained from Colón-Carmona et al. (1999), and double-homozygous F3 progeny of crosses with pdr2 were used for analysis of root cell division.
Seeds were surface-sterilized and germinated on 0.8% (w/v) phytagar medium (+Pi, −Pi, +DNA/−Pi, +RNA/−Pi, or +Phi/−Pi) as described previously by Chen et al. (2000) and Ticconi et al. (2001). For −Fe medium, Fe-EDTA was omitted from the nutrient solution. −N medium was prepared by substituting CaCl2 for Ca(NO3)2 and omitting KNO3. For −S medium, MgCl2, CuCl2, and Zn(NO3)2 were substituted for the respective sulfate salts. Phytagar (commercial grade) was purchased from Gibco-BRL (Gaithersburg, MD, USA) and contributed to about 25 µm total P and 130 µm total nitrogen to the final medium (Catalog no. 10695-039, Lot no. 1021483); concentrations of free Pi were below the detection limit (<5 µm). After stratification at 4°C for 1–2 days, seeds were germinated in a controlled environmental chamber at 25°C under illumination with fluorescent and incandescent light at a photon fluence rate of approximately 60 µmol m−2 sec−1 for 16 h daily. To facilitate FW determination and analysis of roots, plants were grown on vertically oriented agar plates as indicated.
Homozygous pdr2 plants were crossed with wild-type plants of accession Ler. The F1 progeny of these crosses were allowed to self-fertilize, and F2 progenies were scored for the short-root pdr2 phenotype on +DNA/−Pi medium. Plant DNA was extracted from flowers of mutant F2 plants according to Glazebrook et al. (1998). Recombination frequencies between mutant phenotype and simple sequence length polymorphism (SSLP) or cleaved-amplified polymorphisms (CAPS) were measured according to established methods (Bell and Ecker, 1994; Konieczny and Ausubel, 1993). The PDR2 locus was mapped to chromosome 5 between markers CIW8 and CER456575. The latter is a newly developed marker for Arabidopsis annotation unit MOP9 (amplification with the oligomers 5′-TTGGGAACAGAAAAAGTTGGA-3′ and 5′-TTGACACCTCGCTTTCACTG-3′ yields a 263-bp product in Col-0 and a 232-bp product in Ler).
Images of seedlings were recorded under a stereomicroscope (Zeiss Stemi SV11, Jena, Germany) equipped with a Cool Snap high-performance CCD camera and imported into NIH image software for quantitative analysis of root growth. For measurements of epidermal root cell length, seedlings were fixed in 4% (v/v) formaldehyde and observed with differential interference contrast microscopy (Zeiss Axioskop, Jena, Germany). Mean mature cell length was determined in the zone of root hair formation. Histochemical analysis of the GUS reporter enzyme activity was adapted from the study by Jefferson et al. (1987). Sample tissues were fixed in 90% cold acetone for 20 min and incubated for 12 h in reaction buffer containing 5-bromo-4-chloro-3-indolyl-β-d-glucoronic acid (X-Gluc) as the substrate. Plant pigments were destained with ethanol, and the GUS staining patterns were recorded. For cell viability analysis, seedlings were incubated according to Celenza et al. (1995) in a solution of propidium iodide (5 µg ml−1) and fluorescein diacetate (10 µg ml−1). After washing with distilled water and mounting in 50% glycerol, root tips were viewed using a Leica (Wetzlar, Germany) TCS 4D confocal laser microscope. Specimens were viewed using wavelengths specific for each stain, producing two images that were merged using Leica confocal software.
Seedlings were grown for 7 days on a 300-µm mesh nylon screen placed on the surface of high Pi (1 mm) or low Pi (10 µm) agar medium. Plants (40–50) were transferred to the respective liquid Pi media for acclimation (1 h) and were subsequently incubated, with roots immersed and shoots kept in air, in 2.5 ml of fresh high- and low-Pi media containing 1 µCi ml−1[32P]Pi. After labeling (1 h), plants were rinsed in tracer-free medium, the fresh weights were determined, and the radioactivity present in whole seedlings was measured by scintillation spectrometry (Poirier et al., 1991). Pi uptake rates were calculated as the amount of element taken up by whole plants per gram FW of roots.
Determination of P in root tips by ICP-MS analysis
Wild-type and pdr2 seeds were germinated for 5 days on filter paper placed on +Pi agar, transferred to −Pi medium, and allowed to grow for up to three additional days. Roots were placed on an acid-washed glass microscope slide, and root tip sections of 0.5 mm were cut under a magnifying device using a razor blade. Per sample, 50 root tips were dissected, which were transferred into an acid-washed glass tube containing 0.1 ml of deionized H2O. The inner side of the tube was carefully rinsed with 1.9 ml of deionized H2O to ensure collection of the entire tissue sample. After complete drying, the solid root material was digested for 2 h at 118°C with 0.2 ml of concentrated HNO3 (OmniTrace, EM Science, Gibbstown, NJ, USA) and diluted to 4 ml with deionized H2O. Samples were analyzed for Mg, P, Ca, Mn, and Zn on a Thermo Elemental PQ ExCell ICP-MS (Lahner et al., 2003), using a glass conical nebulizer drawing 1 ml min−1. Gallium (EM Science) was included in the digestion acid as an internal standard, and National Institute of Standards and Technology (NIST) traceable calibration standards (ULTRAScientific, North Kingstown, RI, USA) were used to calibrate for each element.
A diagnostic kit from Sigma (No. 670-A) was used to measure free Pi. For total P determinations, agar and tissue samples were ashed according to Ames (1966). Anthocyanins were determined according to Lange et al. (1971). For staining of starch, seedlings were incubated overnight in 96% ethanol to remove pigments, transferred to 1.5% KI/0.3% I2 for 30 min and extensively washed with water. Protein was assayed according to the method of Bradford (1976) using bovine serum albumin as the standard.
Data sets were analyzed by one-way anova followed by the Tukey's multiple comparison test. Statistical significance of differences between means of the groups was defined as a P-value < 0.05.
We thank Arthur Grossman, John Harada, and Doug Grubb for stimulating discussions, Adán Colón-Carmona for seeds of cycB1::GUS transgenic plants, Arnold Bloom and his laboratory for nitrate measurements, and Sun Chong for technical help with some of the experiments. C.A.D. was supported by a pre-doctoral fellowship from the Conselho Nacional de Desenvolvimento Científico e Tecnológico, Brazil. This work was supported by the United States Department of Energy (PR#03-00ER15057.000 to S.A.) and the National Science Foundation Plant Functional Genomics program (0077378-DBI to D.E.S.).
Wild-type and pdr2 seeds were germinated for 10 days on vertically oriented agar plates of the indicated media composition and were subsequently stained for starch.
Figure S2.Development of secondary and tertiary root meristems after transfer from high- to low-Pi condition.
Transgenic (cycB1::GUS) wild-type (closed circles) and pdr2 (open circles) seeds were grown as described in Figure 7. After histochemical GUS staining, total secondary (a) and total tertiary (b) root meristems were counted per root system. Mean values (±SD) are given (n = five seedlings).
Table S1 Accumulation of anthocyanins, total P and free Pi