ER-resident proteins PDR2 and LPR1 mediate the developmental response of root meristems to phosphate availability
- Carla A. Ticconia,
- Rocco D. Luceroa,
- Siriwat Sakhonwaseea,
- Aaron W. Adamsona,
- Audrey Creffb,
- Laurent Nussaumeb,
- Thierry Desnosb and
- Steffen Abela,c,1
- aDepartment of Plant Sciences, University of California, One Shields Avenue, Davis, CA 95616;
- bLaboratoire de Biologie du Développement des Plantes, Commisariat a l'Energie Atomique Cadarache, 13108 St Paul-lez-Durance, Cedex, France; and
- cLeibniz Institute of Plant Biochemistry, Weinberg 3, D-06120 Halle, Germany
Abstract
Inadequate availability of inorganic phosphate (Pi) in the rhizosphere is a common challenge to plants, which activate metabolic and developmental responses to maximize Pi acquisition. The sensory mechanisms that monitor environmental Pi status and regulate root growth via altered meristem activity are unknown. Here, we show that phosphate deficiency response 2 (PDR2) encodes the single P5-type ATPase of Arabidopsis thaliana. PDR2 functions in the endoplasmic reticulum (ER) and is required for proper expression of scarecrow (SCR), a key regulator of root patterning, and for stem-cell maintenance in Pi-deprived roots. We further show that the multicopper oxidase encoded by low phosphate root 1 (LPR1) is targeted to the ER and that LPR1 and PDR2 interact genetically. Because the expression domains of both genes overlap in the stem-cell niche and distal root meristem, we propose that PDR2 and LPR1 function together in an ER-resident pathway that adjusts root meristem activity to external Pi. Our data indicate that the Pi-conditional root phenotype of pdr2 is not caused by increased Fe availability in low Pi; however, Fe homeostasis modifies the developmental response of root meristems to Pi availability.
Plant productivity critically depends on postembryonic development of the root system. Dynamic remodeling of root architecture to fluctuating soil conditions is accomplished by adjustment of primary root growth and lateral root formation. Whereas the sensory mechanisms that monitor nutrient availability and optimize root development remain to be elucidated, major players of the intrinsic pathways controlling root patterning and meristem activity have been identified (1–3). The Arabidopsis root comprises 3 concentric cell layers (epidermis, cortex, endodermis) surrounding the vascular cylinder (stele). A distal organizer, the quiescent center (QC), maintains stem-cell status of its adjoining undifferentiated cells (initials) whose proximal descendents give rise to differentiated tissue layers after passage through the cell division and elongation zone. The position and functionality of the root stem-cell niche depends on the combinatorial, and in part noncell-autonomous, action of key transcription factors such as short root (SHR), scarecrow (SCR), and plethora (PLT1/2) (3, 4). Several plant hormones have been implicated in the maintenance of root meristem activity and could mediate growth responses to edaphic cues (4).
Phosphate (Pi) constitutes a major nexus in metabolism and its bioavailability directly impacts plant performance (5, 6). To cope with Pi shortage, often a result of complex soil chemistries (7), plants redesign root system architecture to accelerate soil exploration. In Arabidopsis, Pi limitation attenuates primary root growth and stimulates lateral root formation, which is thought to maximize Pi acquisition in the topsoil (8, 9). A genetic approach to dissect Pi sensing identified Arabidopsis mutants (10, 11) and accessions (12, 13) with altered sensitivity to the inhibitory effect of Pi deprivation on primary root growth. Characterization of a major quantitative trait locus for root length, low phosphate root 1 (lpr1), revealed a role for multicopper oxidases (LPR1, LPR2) in Pi sensing at the root tip (13, 14). We previously described phosphate deficiency response 2 (pdr2) that displays a hypersensitive response of root meristems to low Pi (10). Here, we show that PDR2 encodes the P5-type ATPase and is required for maintaining SCR protein during Pi deprivation, revealing a link between root patterning and the adjustment of meristem activity to Pi availability. Interestingly, PDR2 and LPR1 interact genetically, and both proteins are expressed in overlapping cell types of the root tip and are localized to the endoplasmic reticulum (ER). Our analyses of pdr2 and lpr1 growth responses further uncover previously described complex interactions between Pi and Fe homeostasis (14–17). We propose that PDR2 and LPR1 are components of an ER-resident pathway mediating root growth responses to Pi and Fe availability.
Results
Loss of PDR2 Alters Pi Sensitivity of Root Growth Independent of Fe Availability.
Attenuation of primary root growth is a developmental response of wild type to Pi limitation, which is exaggerated in pdr2 plants challenged by low (<100 μM) Pi (10). Because root growth inhibition in −Pi may be a consequence of increased Fe availability and its presumed toxicity (15, 16), we studied the effect of decreasing Fe on Pi-limited root growth and observed that pdr2 seedlings continue to display the characteristic short-root phenotype at Fe supplementation as low as 1 μM [see Fig. S1A]. Further Fe restriction and chelation of residual Fe in the −Pi medium partially rescues the pdr2 root phenotype, which prompted us to compare Pi dose responses under conditions where a given concentration of Fe2+ is bioavailable at each Pi concentration tested (0–2.5 mM Pi). We used the Visual Minteq program (15, 17) to calculate the free Fe2+ concentration corresponding to 10 μM FeSO4 in −Pi medium (≈8.8 μM Fe2+) and the appropriate FeSO4 adjustments required to maintain this free Fe2+ concentration at different Pi levels. In this controlled Fe condition, primary root growth of pdr2 shows the reported biphasic Pi dose response (10): A strong inhibition below 0.1 mM Pi and wild type-like growth above this threshold (Fig. S1B). Thus, the pdr2 mutation sensitizes primary root growth to the inhibitory effect of Pi deprivation, which is essentially independent of external Fe availability.
PDR2 Maintains Stem-Cell Fate in Pi-Deprived Roots.
After 5 days of transfer to −Pi medium, wild-type root tips show a slight decrease in pCYCB1;1::GUS expression, a reporter for cell division (18), and a moderate increase of pACP5::GUS activity, a marker for Pi starvation and cell differentiation (19). In contrast, only 2–3 days after transfer, root tips of pdr2 display a sharp decline in pCYCB1;1::GUS expressing cells and an expanded pACP5::GUS domain (Fig. S2 A and B). To explore the cause of meristem exhaustion in pdr2, we examined the integrity of the stem-cell niche upon transfer to −Pi by monitoring the expression of QC marker genes as well as the identity of columella initials. In pdr2, QC25 expression is detected for over 3 days, although a significant reduction in meristem size is already evident at day 2 (Fig. S2C). Similar results were obtained for QC184, suggesting that meristem failure in pdr2 is at least initially independent of QC identity. Iodine staining revealed amyloplast formation in the layer of columella initials in pdr2 between 1–2 days after transfer, indicating stem-cell differentiation (Fig. 1A). Optical sections further indicated loss of cortex/endodermal initials in pdr2 (Fig. 1B). Thus, PDR2 is required for root stem-cell maintenance when external Pi is limiting.
Loss of stem-cell fate in Pi-deprived pdr2 roots. Stem-cell identity in primary roots after transfer (2 days) to media containing 2 mM Pi (+Pi) or no Pi supplement (−Pi). (A) Double staining of QC (QC25) and columella cells (amyloplasts). White arrows point to the position of columella initials, which show signs of differentiation for pdr2 after transfer to −Pi (red arrow). (B) Optical sections of a root meristem after transfer to −Pi, revealing the absence in pdr2 roots of cortical/endodermal initials (blue dots) between the QC (white dots) and the endodermis (yellow dots)/cortex (green dots).
PDR2 Encodes the P5-Type ATPase.
A map-based strategy identified PDR2 to encode the single P5-type ATPase (At5g23630). The EMS-induced pdr2–1 mutation alters an invariant Thr to an Ile residue in a conserved motif involved in ATP binding (Fig. S3). Two T-DNA insertion lines (pdr2–2; pdr2–3) show the characteristic root phenotype as well as reduced fertility; the latter phenotype was reported for T-DNA knockouts designated as male gametogenesis impaired anthers (mia) alleles (20). All phenotypes of pdr2–1 are complemented with a 10-kb fragment of the wild-type At5g23630 locus (Fig. S4 A–E). Overexpression of PDR2 (p35S::PDR2) increased primary root length in −Pi by ≈35% relative to wild type, rendering root growth essentially insensitive to Pi limitation (Fig. S4 F and G). In situ hybridization revealed high expression in the central meristematic region as well as weaker signals in the distal meristem and transition zone, which was not observed in pdr2–2 roots. Analysis of pPDR2::GUS expression confirmed these results (Fig. 2A). Transverse cross-sections revealed PDR2 promoter activity in all cell types of the root meristem. GUS expression was also detected in pavement cells of trichomes, stipules, stamens, and pollen grains (Fig. 2B). The PDR2 expression patterns are consistent with the morphological phenotypes of pdr2/mia mutants.
Expression of PDR2 and its function in the ER. (A) Whole-mount in situ hybridization of PDR2 in WT and pdr2–2 primary root tips (antisense probe) and histochemical staining of pPDR2::GUS expression in a primary root tip (Lower Right) and a transverse cross section (Upper Right). (B) pPDR2::GUS expression in anthers (Left), mature pollen (Right) and shoot apex (Lower). (C) Expression of markers for the ER (p35S::sGFP∼HDEL), Golgi (p35S::NAG∼GFP) and the plasma membrane (p35S::PHT1;1∼GFP) in root tips after germination (5 days) on +Pi medium and subsequent transfer to −Pi (0–2 days). Loss of GFP∼HDEL fluorescence is evident in pdr2 roots after 2 days in −Pi. (D) qRT-PCR analysis of PDR2 and UPR gene expression in roots of wild-type (black bars) and pdr2 (gray bars) after transfer to +Pi or −Pi media in the presence or absence of 5 μM TM for 6 h.
PDR2 Functions in the ER.
Because the Arabidopsis P5-type ATPase resides in the ER (20, 21), we tested whether loss of PDR2 impairs the secretory pathway by monitoring markers for the ER [GFP∼HDEL (22)], Golgi [NAG∼GFP (23)] and plasma membrane [PHT1;1∼GFP (24)] (Fig. 2C). The subcellular localization of PHT1;1∼GFP and NAG∼GFP was similar for wild type and pdr2 root meristems after transfer to −Pi, suggesting that PDR2 inactivation does not disrupt general protein secretion. However, exposure to −Pi caused within 2 days a substantial decline in GFP∼HDEL fluorescence in pdr2 but not wild-type root meristems. The GFP∼HDEL reporter was used to study ER-associated protein degradation (ERAD) in plants, which is 1 mechanism to control protein quality in the secretory pathway (25). Loss of GFP∼HDEL expression in Pi-limited pdr2 roots suggests sensitization of ER quality control mechanisms, which also include activation of unfolded response genes (UPR) genes. We examined expression of PDR2 and select UPR genes in response to altered Pi availability and treatment with tunicamycin (TM), a potent UPR inducer (Fig. 2D). PDR2 mRNA expression did not appreciably change in response to −Pi or +TM treatment. Whereas expression of CNX1, PDIL, or BiP1 was ≈2-fold higher in Pi-sufficient pdr2 roots relative to wild type, transfer to −Pi marginally increased (<2-fold) transcript levels in either genotype. As expected, a strong induction (≈10-fold) was evident for wild type during transfer from +Pi to +Pi/+TM medium. Interestingly, expression of CNX1 and PDIL mRNA was notably higher for pdr2 roots (>20-fold). Transfer from +Pi to −Pi/+TM medium did not further elevate UPR gene expression (Fig. 2D). Collectively, our data indicate that loss of PDR2 sensitizes a subset of ER quality control responses.
PDR2 Restricts SHR Movement in Limiting Pi.
Because PDR2 is necessary for stem-cell maintenance in low Pi, we determined transcript levels of 6 root patterning genes in wild-type and pdr2 roots for up to 2 days after transfer to −Pi, but observed only changes smaller than 2-fold (Fig. S5A). A function of PDR2 in the secretory pathway prompted us to examine whether SHR protein expression and localization is altered in pdr2 root meristems because radial root patterning depends on SHR trafficking from the stele into the endodermis (26). As expected, SHR promoter activity (pSHR::YFPNLS) was not affected in Pi-limited wild type and pdr2 roots (Fig. S5B). However, although SHR∼GFP protein expression (pSHR::SHR∼GFP) did not change in wild type root meristems during 2 days of Pi shortage, SHR∼GFP fluorescence initially declined in the endodermal cell layer of pdr2 roots at day 1 and noticeably decreased in the stele at day 2 after transfer (Fig. 3). Loss of SHR∼GFP could be the result of accelerated protein degradation; however, we detected ectopic SHR∼GFP fluorescence in cortex and epidermal cells, suggesting unrestricted SHR movement into adjacent cell layers of Pi-deprived pdr2 roots (Fig. 3, insets), as was previously reported for lines with reduced SCR expression (27).
PDR2 is necessary for maintaining SHR distribution in low Pi. Transgenic WT and pdr2–1 seeds were germinated on +Pi medium (5 days) and pSHR::SHR∼GFP expression was monitored after transfer to −Pi (0–2 days). The arrow points to the endodermal cell layer. Lower insets show enlarged images of framed areas. The upper inset shows a newly emerged lateral root.
PDR2 Maintains SCR Level in Limiting Pi.
Because SCR directly interacts with SHR in endodermal cell nuclei and restricts further SHR movement (27), we monitored GFP∼SCR expression and used complemented scr-4 (pSCR::GFP∼SCR) to introgress the reporter into pdr2. Whereas GFP∼SCR expression and root meristem organization did not change in the wild type (PDR2+/+scr-4−/−GFP∼SCR+/+) for at least 2 days after transfer to −Pi, both traits were strikingly altered for pdr2 in a SCR dose-dependent manner (Fig. 4A). We identified homozygous pdr2 lines harboring 1–4 SCR alleles, with at least 1 GFP∼SCR copy. When grown in +Pi, GFP∼SCR expression in pdr2 was similar to the wild type. In pdr2 (1xSCR), GFP∼SCR expression was largely abolished within 1 day after transfer to −Pi. A single cell layer of ground tissue was often observed at day 2 (32 of 41 seedlings), which is characteristic of scr and shr roots (28). GFP∼SCR fluorescence in newly emerged lateral root meristems was still detectable, as was J0571 expression, a GFP marker for cortex and endodermis (Fig. S5C). In pdr2 (2× SCR), GFP∼SCR expression was greatly reduced at day 1 and only residual fluorescence was detected in the primary meristem at day 2, which showed signs of disorganization. As expected, immunoblot analysis revealed reduction of authentic SCR protein in Pi-deprived pdr2 roots, which was not observed for wild-type or Pi-sufficient seedlings (Fig. 4B). In pdr2 (3× SCR), GFP∼SCR expression was less reduced even 2 days after transfer. GFP∼SCR fluorescence appeared also in the cortex and ectopic cell divisions in the ground tissue were frequently observed (48 of 67 seedlings). Surprisingly, GFP∼SCR expression was maintained in pdr2 (4× SCR) lines for 4–5 days after transfer. Although loss of the stem-cell niche was initially prevented, the proximal meristem showed signs of cell differentiation, and these roots ultimately displayed a short root phenotype in −Pi.
PDR2 is necessary for maintaining SCR protein level in low Pi. (A) pSCR::GFP∼SCR expression in root meristems of WT and pdr2–1 harboring 1–4 alleles of SCR (1–4×), with at least 1 copy of the GFP∼SCR reporter. Transgenic WT and pdr2 seeds were germinated on +Pi medium (5 days) and transferred to −Pi (0–2 days). The white bracket for pdr2 (1× SCR) outlines a single layer of ground tissue characteristic of shr and scr mutants. White brackets for pdr2 (3× SCR) indicate supernumerary endodermal cells displaying GFP∼SCR fluorescence similar to SCR RNAi lines (27). Insets show newly emerged lateral roots. The white bar for pdr2 (4× SCR) denotes the proximal meristem border. (B) Endogenous SCR level in root extracts of WT and pdr2 after transfer to +Pi or −Pi (2 days). Left shows anti-SCR immunoblot of SCR-IP. Right shows anti-BiP immunoblot of 10% SCR-IP input.
The rescue of distal meristem activity in Pi-deprived pdr2 (4× SCR) roots and ensuing proximal meristem reduction points to a dual function of PDR2 in Pi-dependent root meristem maintenance, which is also indicated by the analysis of the pdr2–1 scr-3 double mutant (Fig. S6). Primary roots of pdr2 scr seedlings display the pdr2 phenotype on −Pi and respond like pdr2 to decreasing Pi availability. However, meristem organization of the double mutant is indistinguishable from scr on −Pi, revealing the characteristic single ground cell layer, which is reminiscent of Pi-starved pdr2 (1× SCR) roots (Fig. 4A). Thus, PDR2 is required for maintaining SCR expression in the stem-cell niche as well as for proximal meristem activity under Pi limitation.
The ER Participates in Root Meristem Response to Pi Deficiency.
It is not clear how ER-resident PDR2 maintains nuclear SCR level when Pi is limiting. However, we observed an uncharacteristic GFP∼SCR localization in a substantial fraction of endodermal cells of wild-type roots harboring extra (3 or 4) SCR copies (Fig. 5A). In addition to its expected nuclear localization, GFP∼SCR fluorescence was detected throughout the cell in a pattern consistent with the endomembrane system. ER localization of GFP∼SCR was confirmed by counterstaining with ER-Tracker and further corroborated by its sensitivity to TM, which caused rapid loss of GFP∼SCR fluorescence only in wild-type roots with extra SCR copies (Fig. 5A). This observation raises the possibility that nuclear SCR level depends on ER-associated activities, which is further suggested by the effect of brefeldin A (BFA) on pdr2 root meri stems. BFA reversibly inhibits vesicle trafficking and induces redistribution of Golgi proteins into the ER (29). When compared to wild type on +Pi, primary root growth of pdr2 is strongly inhibited by 5 μM BFA. BFA also inhibits root growth of wild type on −Pi, with no additional effect on pdr2 (Fig. S7). Reduction of pdr2 root meristem size by BFA treatment in +Pi is illustrated by loss of GFP∼SCR and pCYCB1;1::GUS expression (Fig. 5B), which is reminiscent of Pi-deprived pdr2 root meristems (Fig. 4A and Fig. S2A). Thus, BFA treatment of Pi-sufficient pdr2 roots mimics the effect of Pi deficiency.
A role of the secretory pathway for root meristem maintenance. (A) Increased SCR dosage results in localization of GFP∼SCR to the ER in wild type. Optical sections of WT root tips harboring 3 SCR copies (a–d). Inset in (a) indicates endodermal cells at higher magnification. GFP∼SCR (b) co-localizes with ER tracker DPX fluorescence (c and d). GFP∼SCR fluorescence in wild type with 2 (e and f) or 3 SCR copies (g and h) before (e and g) and after (f and h) tunicamycin treatment (5 μM, 4 h). (B) BFA mimics the effects of −Pi on pdr2 root meristems in Pi sufficiency. (Upper), pSCR::GFP∼SCR expression in roots following transfer to +Pi media with or without 25 μM BFA (1 day). The white bracket outlines meristem size reduction. (Lower), pCYCB1::GUS expression in 14-day-old primary root tips of WT and pdr2 seedlings grown in +Pi media in the absence (−) or presence (+) of 10 μM BFA.
LPR1 Interacts Genetically with PDR2 and Localizes to the ER.
Overexpression of PDR2 alleviates the inhibitory effect of Pi deprivation on primary root growth (Fig. S4 F and G), a phenotype similar to recessive lpr1 and lpr2 mutants that develop longer primary roots in −Pi than the wild type (13, 14). To study the relationship between PDR2 and LPR genes, we compared the growth response of wild type, pdr2, lpr1lpr2, and pdr2lpr1lpr2 roots to Pi shortage. In −Pi, primary root length and meristem size of lpr1lpr2 and pdr2lpr1lpr2 lines were similar and both parameters exceeded those of the wild type (Fig. 6A). The nearly full epistasis indicates that loss of LPR1/LPR2 cannot be bypassed by PDR2 inactivation. Because both predicted LPR multicopper oxidases contain a putative N-terminal signal peptide (SP), we determined the subcellular localization of p35S::SP∼eGFP∼LPR1 expression (Fig. 6B). GFP∼LPR1 fluorescence was detected in transgenic Arabidopsis root cells in a pattern consistent with a reticular network. Transient co-expression with an ER marker (p35S::DsRed2∼KDEL) revealed co-localization of both reporter proteins in N. benthamiana leaf cells, including the perinuclear intermembrane space (Fig. 6B). Thus, the ER-resident LPR enzymes function together with the P5-type ATPase in a common pathway that adjusts root meristem activity to Pi availability.
LPR genes functionally interact with PDR2 and encode ER proteins. (A) Loss of LPR1 and LPR2 rescues the pdr2 root phenotype in low Pi. The mean primary root length (mm) of seedlings grown for 8 day in +Pi or −Pi medium (± SD, n = 10–15) is given below. Insets show cleared root tips of seedlings grown in −Pi. The approximate size of primary meristems is indicated by yellow bars. (B) Subcellular localization of LPR1. Expression of p35S::SP∼eGFP∼LPR1 in stably transformed Arabidopsis root tips: (a) GFP∼LPR1 fluorescence, (b) propidium iodide staining, and (c) overlay. Transient co-expression of p35S::SP∼eGFP∼LPR1 and p35S::DsRed2∼KDEL in epidermal leaf cells of N. benthamiana: (d, g) GFP∼LPR1 fluorescence, (e and h) DsRed2∼KDEL fluorescence, (f and i) overlays. Insets (g–i) show close-up of the cell nucleus.
Discussion
We show that PDR2 encodes the single P5-type ATPase (At5g23630) previously named MIA (20). The Arabidopsis family of P-type ATPases consists of 46 members that can be assorted into 5 groups according to their transport ions (heavy metals, Ca2+, H+, and aminophospholipids) (30). The specificity and precise role of any P5 pump is unknown (31). Cod1p, the PDR2/MIA ortholog in yeast (S. cerevisiae), is located in the ER and nuclear envelope (32). Properties of cod1Δ strains point to its function in protein biogenesis and ER quality control, including protein processing, control of protein insertion orientation, ERAD, and cation homeostasis (33–36). Several observations support a role for PDR2/MIA in the ER: (i) its C-terminal ER retrieval signal of the KKXX-type suggests retrograde Golgi-to-ER transport (37); (ii) a proteomics approach identified the At5g23630 protein in ER microsomes (21); and (iii) Jakobsen et al. (20) detected MIA by immunogold labeling in the ER and small vesicles in tapetal cells and demonstrated its ability to complement cod1Δ. Disruption of MIA deregulates gene expression related to protein folding and secretion in anthers, suggesting enhanced protein quality control and reduced ER protein load, whereas changes in transcript levels of solute transporter genes likely cause cation imbalances in mia leaves (20). We show that loss of PDR2/MIA selectively sensitizes UPR gene expression in roots (Fig. 2D). General secretion is not severely compromised in pdr2 root meristems (Fig. 2C), which is supported by our previous report that Pi-starved pdr2 roots secrete higher activities of phosphohydrolases than the wild type, likely as a consequence of elevated expression of Pi starvation-inducible genes (10) (Fig. S5A).
PDR2 is required for maintaining nuclear SCR protein when Pi is limiting (Fig. 4A), a failure of which causes stem-cell differentiation (Fig. 1) and meristem reduction (Fig. S2). SCR and SHR are members of the GRAS family of transcription factors and key regulators of radial root patterning. SCR is normally expressed in the QC, cortex/endodermal initials and endodermis, and at low level in the cortex (38, 39). SHR transcription is restricted to the stele, but the SHR protein moves into the adjacent cell layer where it activates SCR transcription and controls endodermis specification (26). SCR delimits SHR movement by direct protein interaction and SCR/SHR-dependent autoactivation of SCR transcription. Interestingly, RNAi-mediated reduction of SCR expression allows unsequestered SHR to pass into the presumptive cortex where it activates SCR and ectopic endodermis specification (27). Our data suggest that PDR2 inactivation causes rapid reduction of SCR protein in low Pi (within 1 day), which can be compensated by increasing SCR gene dosage. In Pi-starved pdr2 roots harboring only 1 or 2 SCR copies, SCR level is possibly too low to confine SHR in the endodermis and to trigger SCR/SHR-dependent SCR up-regulation (Fig. 4 A and B). Thus, the observed loss of SHR∼GFP fluorescence in the endodermis in −Pi (Fig. 3) is likely a consequence of unhindered SHR∼GFP movement into peripheral cell layers. In pdr2 (3× SCR) roots, SCR may approach a level closer to wild-type threshold, which still permits continued SHR movement but is already sufficient to activate SCR/SHR-dependent SCR transcription in recipient cell files. Threshold is restored or exceeded in pdr2 (4× SCR) roots that show wild type-like GFP∼SCR expression in the distal meristem (Fig. 4A). Thus, the pdr2 mutation strikingly mimics in −Pi the effect of RNAi-mediated SCR reduction (27). However, PDR2 inactivation likely affects nuclear SCR protein level posttranscriptionally because steady-state SCR and SHR mRNA levels in roots do not respond to Pi deprivation (Fig. S5).
Several observations suggest that endosomal compartments participate in regulating nuclear SCR level: (i) increased SCR gene dosage in wild type can cause atypical GFP∼SCR localization to the ER (Fig. 5A); (ii) BFA treatment of Pi-sufficient pdr2 roots mimics the effect of low Pi on GFP∼SCR expression (Fig. 5B); (iii) SCR physically interacts with a SEC14 family member of phospholipid transfer proteins (40), which regulate membrane trafficking in yeast (41); and (iv) the GRAS domain of SCR has the capacity to mediate cell-to-cell movement (39). The unknown biochemical activity of P5-type ATPases and the diverse molecular phenotypes of Arabidopsis and yeast knockouts leave open the question how PDR2 maintains SCR in low Pi. Although the mechanism is most likely indirect, for example, by maintaining ER ion homeostasis, specific PDR2-dependent activities of the early secretory pathway may regulate SCR level. For example, the key enzyme of the mevalonate pathway, HMG-CoA reductase, is feedback-regulated via ERAD in yeast, and loss of Cod1p causes its constitutive degradation without promoting removal of other ER proteins (32, 33). Because SCR down-regulates RBR (retinoblastoma-related) in the stem-cell niche (42), PDR2-dependent growth response to Pi status likely impinges on the SCR-RBR pathway to adjust the balance of cell division and differentiation in the root meristem.
A role of the ER in root growth response to external Pi status is further supported by the subcellular localization of LPR1 and its genetic interaction with PDR2 (Fig. 6). The expression domains of LPR1 (14) and PDR2 (Fig. 2A) overlap in the stem-cell niche and distal root meristem, suggesting that both proteins function together in an ER-resident pathway that adjusts meristem activity to external Pi status. Because the recessive lpr1lpr2 and pdr2 mutations result in opposite phenotypes in low Pi and lpr1lpr2 is epistatic to pdr2, the P5-type ATPase likely restricts LPR output, either by negatively regulating LPR biogenesis or activity, or by removing products generated by LPR multicopper oxidases (Fig. S8). It has recently been argued that root growth inhibition by limiting Pi is rather an indirect result of elevated Fe availability and its presumed toxicity than a consequence of Pi shortage (15). We observed that pdr2 root growth is hypersensitive to change in external Pi even under controlled Fe conditions and is essentially unaffected by reduction of the Fe supplement to as low as 1 μM; however, further Fe restriction or chelation of residual Fe in the −Pi medium substantially rescues the pdr2 root phenotype (Fig. S1A). Thus, the pdr2 mutation sensitizes root growth response to both external Pi and Fe. Because Pi deprivation causes elevated Fe tissue content and regulates genes affecting Fe homeostasis (16, 17), our data on PDR2 and LPR1 are consistent with a model in which Pi availability interacts with Fe homeostasis to adjust root meristem activity via proteins of the ER and cell nucleus (Fig. S8).
Materials and Methods
Plant Material and Growth Conditions.
Arabidopsis lines were grown on 0.8% phytagar containing 0.5% sucrose and 0.5× MS salts with (+Pi) or without (−Pi) 0.625 mM KH2PO4, pH 5.6. For media with altered Pi and/or Fe content, the previously described MS-based salt composition was used (10). Nutrient bioavailability was calculated by using the Visual Minteq program (15). See SI Text for details.
Identification of PDR2.
The PDR2 locus was mapped to chromosome 5 between markers CER 460453 and CER 456861. This region was sequenced and the pdr2–1 point mutation verified by molecular complementation (SI Text).
Microscopy.
Roots were cleared and imaged with DIC optics (Zeiss Axioskop). For confocal microscopy, roots were counterstained in 10 μM propidium iodide and imaged (Olympus FV-1000 or Leica SP2 AOBS). Transient assays of p35S::SP∼eGFP∼LPR1 and p35S::DsRed2∼KDEL co-expression were performed as described (SI Text).
qRT-PCR.
Gene expression analysis was performed by qRT-PCR by using the 7300 Real-Time PCR System (Applied Biosystems), SYBR Green I PCR master mix, and the amplimers listed in Tables S1 and S2. Data were analyzed by the ΔΔ-CT method (SI Text).
Acknowledgments
We thank B. Scheres and R. Heidstra (University of Utrecht, The Netherlands), P. Benfey (Duke University, Durham, NC), I. Moore (University of Oxford, Oxford), and J. Paz-Ares (Centro Nacional de Biotecnologia, Madrid, Spain) for transgenic lines; P. Benfey for SCR antibody; V. Rubio (Centro Nacional de Biotecnologia, Madrid, Spain) and R. Bhat (Commisariat al'Energie Atomique Cadarache) for plasmids; and C. Arnaud (Commisariat a l'Energie Atomique Cadarache) for initiating construction of pdr2lpr1lpr2 lines. We further thank B. Scheres for discussion and J. Harada for critical comments. This work was supported by U.S. Department of Energy Grant DE-FG0203ER15447 (to S.A.). A.C., L.N., and T.D. were supported by the Commissariat à l'Energie Atomique.
Footnotes
- 1To whom correspondence should be addressed. E-mail: sabel{at}ucdavis.edu
-
Edited by Richard M. Amasino, University of Wisconsin, Madison, WI, and approved June 30, 2009
-
Author contributions: C.A.T., L.N., T.D., and S.A. designed research; C.A.T., R.D.L., S.S., A.W.A., A.C., T.D., and S.A. performed research; C.A.T., R.D.L., S.S., A.W.A., A.C., T.D., and S.A. analyzed data; and C.A.T. and S.A. wrote the paper.
-
The authors declare no conflict of interest.
-
This article is a PNAS Direct Submission.
-
This article contains supporting information online at www.pnas.org/cgi/content/full/0901778106/DCSupplemental.
-
Freely available online through the PNAS open access option.
