Selective Trafficking of Non-Cell-Autonomous Proteins Mediated by NtNCAPP1

See allHide authors and affiliations

Science  17 Jan 2003:
Vol. 299, Issue 5605, pp. 392-396
DOI: 10.1126/science.1077813


In plants, cell-to-cell communication is mediated by plasmodesmata and involves the trafficking of non–cell-autonomous proteins (NCAPs). A component in this pathway, Nicotiana tabacumNON-CELL-AUTONOMOUS PATHWAY PROTEIN1 (NtNCAPP1), was affinity purified and cloned. Protein overlay assays and in vivo studies showed that NtNCAPP1 is located on the endoplasmic reticulum at the cell periphery and displays specificity in its interaction with NCAPs. Deletion of the NtNCAPP1 amino-terminal transmembrane domain produced a dominant-negative mutant that blocked the trafficking of specific NCAPs. Transgenic tobacco plants expressing this mutant form of NtNCAPP1 and plants in which the NtNCAPP1 gene was silenced were compromised in their ability to regulate leaf and floral development. These results support a model in which NCAP delivery to plasmodesmata is both selective and regulated.

In plants, the trafficking of NCAPs that are involved in the regulation of plant development is thought to occur through plasmodesmata (1–6). However, little information is available concerning the manner in which such NCAPs enter this cell-to-cell translocation pathway (5,6). To identify potential components in this pathway, we used the NCAP CmPP16 (7) as bait for the affinity purification of interaction partners contained within a plasmodesmal-enriched cell wall protein (PECP) fraction (8–10) prepared with tobacco BY-2 cells [fig. S1, A and B (11)]. A resultant highly enriched 40-kD protein was identified (Fig. 1, fig. S1C), cloned, and named NtNCAPP1 (GenBank accession number AF307094; hereafter called NCAPP1) [fig. S2 (11)].

Figure 1

Affinity purification of NtNCAPP1, a component in the plant NCAP translocation pathway. A BY-2 cell PECP fraction was first passed through a glutathione S-transferase (GST)–crosslinked sepharose column, and the flow-through fraction was then loaded onto a GST-CmPP16 affinity column (11). Coomassie blue–stained SDS-PAGE gel shows the profile for the loading (20 μg), flow-through (20 μg), and elution fractions. A major 40-kD protein (dart; ∼2 μg), representing a 107-fold enrichment of NCAPP1, was microsequenced (fig. S2A).

The specificity of the interaction between NCAPP1 and CmPP16 was tested using a protein overlay approach (11). Native NCAPP1 (contained within the PECP preparation) interacted with only a very small subset of the proteins present in the PECP fraction (Fig. 2A). Furthermore, fractions enriched for cytoplasmic proteins exhibited only minimal interaction with NCAPP1 (Fig. 2B). A reciprocal experiment in which the PECP fractions were probed with the CmPP16 bait confirmed the specificity of the interaction between native NCAPP1 and CmPP16 (Fig. 2C). As the CmPP16 is an endogenous NCAP located within the phloem sap (7), we next used fractionated phloem sap (11) in an overlay with PECP and, as anticipated, detected a strong signal in the region corresponding to the CmPP16 (Fig. 2D). A range of other phloem proteins also interacted positively with NCAPP1, consistent with observations that various phloem components can traffic through plasmodesmata (12). These results confirmed that the NCAPP1 enrichment achieved in our affinity chromatography experiments (Fig. 1) was due to its specific interaction with the CmPP16 bait. The presence of a range of NCAPP1-interacting proteins in the phloem sap suggests that NCAPP1 (and other isoforms) may be central to NCAP trafficking in general.

Figure 2

Specificity of the interaction between native NCAPP1 and CmPP16. (A and B) Proteins from fast performance liquid chromatography (FPLC)–fractionated PECP (A) and a cytoplasmically enriched preparation (B) (Coomassie blue–stained SDS-PAGE, left) were overlaid with unfractionated PECP (200 μg/ml) (middle) or bovine serum albumin (BSA) buffer control (right) and then immunoblotted with NCAPP1 α-Cpep antibodies (11) to detect proteins interacting with NCAPP1. Proteins in the 55- to 65-kD size range (circled) were also detected at low abundance in affinity chromatography experiments (Fig. 1). Control panel in (A) (right) shows the presence of five NCAPP1 isoforms, or derivatives, consistent with results from the screening of a BY-2 cDNA library (11). In (B), both the overlay and the control blots exhibited nonspecific immunologically reactive bands. Boxed bands identify potential NCAPP1 interaction partners. (C) Equivalent experiments performed by overlaying PECP with CmPP16 (5 μg/ml) (middle) or BSA buffer control (right) before immunoblotting with α-CmPP16 antibodies (7). Western analysis performed on this blot confirmed the identity of this single major band as being NCAPP1, further confirming the specific interaction between CmPP16 and NCAPP1 (Fig. 1). (D) Overlay assays performed on FPLC-fractionated phloem sap (11) as described for (A). Location of the CmPP16 isoforms is indicated by asterisks. In contrast to the control (right), where only one immunologically reactive 40-kD protein was observed, the PECP-overlaid blot (middle) exhibited a significant number of NCAPP1 (and isoform) interaction partners in addition to CmPP16.

Subcellular localization of NCAPP1 was examined by expression of fluorescently tagged NCAPP1 in BY-2 cells (Fig. 3). In contrast to the fluorescence pattern observed with free EGFP (enhanced green fluorescent protein) (Fig. 3A), NCAPP1-EGFP accumulated at the cell periphery (Fig. 3B). Similarly, fluorescence associated with CmPP16-RFP was highest at the periphery of BY-2 cells (Fig. 3C). A role for the predicted NH2-terminal transmembrane domain (residues 1 to 22) in the intracellular targeting of NCAPP1-EGFP to the endoplasmic reticulum (ER) was studied using a truncation mutant (NCAPP1Δ1–22). A shift in fluorescence from the periphery to the nucleus (Fig. 3D) is consistent with this prediction.

Figure 3

Subcellular localization of NCAPP1 in tobacco BY-2 cells. (A to D) Transient production of EGFP (A), NCAPP1-EGFP (B), CmPP16-RFP (C), and NCAPP1Δ1–22-EGFP (D) within individual cells. (Eand F) Accumulation of NCAPP1-EGFP in transgenic cells (TBY-2) along the developing cell plate (E, arrow) (fig. S5) and distribution of ER imaged with ER-targeted EGFP (F); images represent single frames from time-lapse series (11) (movies S1 to S3). (G and H) Immunodetection of NCAPP1 (fig. S5) using α-Cpep in wild-type (G) and TBY-2 (H) cells, presented in false-blue color. (I and J) Immunogold detection of NCAPP1-EGFP on the cytoplasmic face of the ER that is associated with the cell plate (CP) (I) and on the ER near a plasmodesma (PD) (J). Presence of gold particles is indicated by arrows. (K) Histogram illustrating the distribution of gold particles within subcellular compartments of TBY-2 cells (fig. S6). (L) Colocalization of NCAPP1-EGFP and CmPP16-RFP in TBY-2 cells. (M) A cell double-labeled with NCAPP1Δ1–22-EGFP (left) and CmPP16-RFP (middle) shows colocalization of probes in the cytoplasm (combined image at right). (N) The same cell shown in (M) supported on filter paper and imaged by transmitted light. Note position of the adjoining cell walls (arrows). Scale bars, 10 μm in (A) to (H), (L) to (N); 0.2 μm in (I) and (J).

A strong NCAPP1-EGFP signal was observed within the plane of the newly forming cell plate (Fig. 3E), a pattern that was in contrast with the broader signal obtained using ER-targeted EGFP (Fig. 3F). Time-lapse video imaging of BY-2 cells illustrated the dynamics of NCAPP1-EGFP targeting to and close association with the developing of the cell plate. Immunofluorescence analysis detected a pattern consistent with that observed for NCAPP1-EGFP; however, this immunological signal likely represents contributions from multiple isoforms of NCAPP1 (Fig. 3, G and H).

Examination of NCAPP1-EGFP distribution, using α-GFP as the immunological probe, revealed its presence at the forming cell plate and, in particular, on the outer surface of the cortical ER membrane (Fig. 3I, fig. S6). Furthermore, NCAPP1-EGFP was detected near plasmodesmata (Fig. 3J), but, although label was detected near the orifice, it was never found in association with the appressed form of the ER within the plasmodesmata. A detailed analysis of gold particle density supported the ER as the location of NCAPP1 (Fig. 3K).

Evidence consistent with the colocalization of NCAPP1 and CmPP16 was gained by experiments in which CmPP16-RFP was transiently expressed in transformed BY-2 cells carrying NCAPP1 (Fig. 3L). In experiments in which NCAPP1Δ1–22-EGFP and CmPP16-RFP were coexpressed, overlapping fluorescent signals were detected in the cytoplasm but not in the nucleus (Fig. 3M). These results are consistent with, but do not prove, an in vivo interaction between NCAPP1 and CmPP16.

The potential for interaction between NCAPP1Δ1–22and CmPP16 was investigated by microinjection studies. Control experiments confirmed that fluorescein isothiocyanate (FITC)–labeled CmPP16 has the capacity to move from cell to cell (Table 1). This movement capacity was blocked when CmPP16-FITC was coinjected with NCAPP1Δ1–22 (at a molar ratio of 1:1). Coinjection with a control protein [42-kDa maltose binding protein (MBP) at an equivalent molar ratio of 1:1] had no effect on the capacity of CmPP16 to mediate its own cell-to-cell transport, supporting the notion that the observed inhibition reflects an interaction between CmPP16 and NCAPP1Δ1–22.

Table 1

NCAPP1Δ1–22 prevents modulation of plasmodesmal SEL by CmPP16 and TMV MP.

View this table:

This block of NCAPP1Δ1–22 on CmPP16 function could result from a simple sequestering of this NCAP within the cytoplasm, such that it can no longer access the plasmodesmal translocation machinery. Alternatively, the NCAPP1Δ1–22 could form a dysfunctional complex, anywhere along the translocation pathway, by competing with endogenous NCAPP1. A series of microinjection experiments using FITC-dextran both as the reporter for size exclusion limit (SEL) increase (2, 9, 10) and as a proxy for cell-to-cell movement of the test protein (2,9, 10, 13–15) demonstrated that NCAPP1Δ1–22 prevented CmPP16 from inducing an increase in plasmodesmal SEL (Table 1).

Further support for the formation of a dysfunctional complex was gained from experiments in which the molar ratio of NCAPP1Δ1–22 to CmPP16 was serially decreased by holding CmPP16 constant and reducing the amount of NCAPP1Δ1–22(Table 1). Cell-to-cell movement of the FITC-dextran remained fully blocked even when the CmPP16:NCAPP1Δ1–22 molar ratio was adjusted from 1:1 to 1:0.005, a condition in which excess, unsequestered CmPP16 would be present in the cytoplasm. A further decrease in the molar ratio by a factor of 10, to 1:0.0005, permitted only limited cell-to-cell movement of FITC-dextran; full restoration of CmPP16 movement capacity was obtained at a molar ratio of 1:0.0002.

To explore whether the capacity of NCAPP1Δ1–22 to block NCAP movement was specific to CmPP16, we conducted parallel experiments with the well-characterized (MP) of Tobacco mosaic virus (TMV) (16–18). Coinjection of TMV MP along with NCAPP1Δ1–22 (molar ratio of 1:1) also completely inhibited the capacity of this MP to mediate an increase in plasmodesmal SEL. These results support the hypothesis that NCAPP1Δ1–22 can interact with CmPP16 and TMV MP in a dysfunctional complex that blocks the translocation pathway. These microinjection results suggested that NCAPP1Δ1–22 could act as a dominant-negative mutant when expressed in transgenic tobacco plants.

We next investigated the role played by NCAPP1 in mediating the process of NCAP translocation with the use of transgenic plants that ectopically expressed NCAPP1 or NCAPP1Δ1–22, or through the reduction of endogenous NCAPP1 expression by gene silencing (11). Although phenotypic changes were not observed for the NCAPP1 overexpression transgenic lines, the presence of the dysfunctional NCAPP1Δ1–22 compromised the ability of these plants to regulate organ development (Fig. 4). These effects were manifested through alterations in developmental programming, as reflected by retarded growth rates and a severe perturbation to terminal inflorescence architecture and leaf patterning. Phenotypic variation ranged from mild (Fig. 4, A to C) to intermediate (Fig. 4D) to severe (Fig. 4E), and in the latter case terminal flowers aborted in mid-development. The main features associated with these plants were a loss of organ symmetry, interorgan fusion, and an inability to separate whorl boundaries in the flowers. Silencing of endogenous NCAPP1 yielded plants with equivalent phenotypes (Fig. 4F), supporting the hypothesis thatNCAPP1Δ1–22 acts as a dominant-negative mutant.

Figure 4

Ectopic expression of NCAPP1Δ1–22and silencing of endogenous NCAPP1 in tobacco perturbs developmental programming (fig. S7). (A toE) Phenotypic analyses performed on seven independent lines (T1 through T3) showed segregation into mild (A to C), intermediate (D), and severe (E) classes (11). (A and B) Early and mid stages in terminal inflorescence development in NCAPP1Δ1–22 plants. (C) Lack of organ symmetry and whorl separation, plus the enlarged condition of NCAPP1Δ1–22 transgenic terminal flower (left, unsectioned) and organs (gynoecium, inset) contrasted with a control flower (right, longitudinal section). (D) Intermediate and severe phenotypes exhibited a loss of apical dominance; leaves were epinastic, thickened, and lanceolate to highly asymmetric in shape, and flowers were infertile. (E) Plants with severe phenotype were much reduced in size. Inset: Equivalent aged transgenic (left) and wild-type (right) plants. (F) Silencing of NCAPP1 caused developmental perturbations equivalent to those observed inNCAPP1Δ1–22 transgenic lines. (Gto I) Anatomical analyses. (G) Transverse sections of paraffin-embedded tissue (11) from basal regions of the leaves subtending the terminal inflorescence of control (top) andNCAPP1Δ1–22 transgenic (bottom) plants shown in insets to (H) and (I). Note the level of disorganization in the cell layers within the lamina of NCAPP1Δ1–22 plants. (H) Scanning electron microscopic images of the abaxial surfaces of control (left) and NCAPP1Δ1–22 (right) leaves shown in insets. Note that NCAPP1Δ1–22 leaves have areas with elevated guard cell density (fig. S7, F and G). Scale bars: 50 μm in (G) to (I); 1 cm in insets.

At the cellular level, NCAPP1Δ1–22 appeared to perturb the underlying program responsible for cell fate in leaves (Fig. 4, G to I). Anatomical analyses performed on developing leaves revealed the presence of sectors exhibiting abnormal patterning within the epidermis as well as the distortion of cell layers within the mesophyll, as compared with wild-type plants.

The dominant-negative effects of NCAPP1Δ1–22 over the capacity of CmPP16 and TMV MP (Table 1) to modulate plasmodesmal SEL were further investigated using these transgenic plants. Introduction of CmPP16 and TMV MP into cells of wild-type tobacco leaves resulted in the expected increase in plasmodesmal SEL, whereas in NCAPP1Δ1–22 tissues the movement of FITC-dextran was completely blocked (Table 2). Thus, NCAPP1Δ1–22 has an equivalent effect on plasmodesmal function, whether it is introduced by microinjection of a recombinant protein or is ectopically expressed in the plant. The fact that other NCAPs, such as KNOTTED1 (KN1) (2) and cucumber mosaic virus (CMV) MP (19, 20), were unaffected in their ability to mediate an increase in plasmodesmal SEL when microinjected into NCAPP1Δ1–22 transgenic plants (Table 2) established that a level of selectivity is associated with plasmodesmal function by this NCAPP1 pathway. Collectively, these findings are consistent with the hypothesis that NCAPP1 acts in the pathway that mediates the translocation through plasmodesmata of some, but not all, NCAPs.

Table 2

Transgenic tobacco plants expressingNCAPP1Δ1-22 exhibit a selective block on the modulation of plasmodesmal SEL. FITC-labeled dextran (9.3 kD) was used as a fluorescent reporter to test the plasmodesmal SEL in response to injected proteins.

View this table:

In a number of cases, flowers with malformed organs result from changes in the expression patterns for genes known to act non–cell-autonomously. The presence of NCAPP1Δ1–22 could mimic these changes through its selective influence over the capacity for NCAP movement in the meristem. When the floral transcription factor LEAFY, known to act non–cell-autonomously (3), or its tobacco homolog, NFL, was ectopically expressed in tobacco (21), terminal flowers displayed phenotypes similar to those observed in our NCAPP1Δ1–22 transgenic plants. Consistent with this notion, our immunological experiments revealed that, in NCAPP1Δ1–22 plants, the pattern of NFL was altered, whereas the cellular distribution of the tobacco homolog of KN1 remained unaffected (fig. S8).

Our findings offer support for the concept that cell-to-cell communication, via plasmodesmata, occurs in a regulated manner. A potential role for NCAPP1, within the context of the NCAP translocation pathway, is that it functions to shuttle some NCAPs to the plasmodesmal microchannel (fig. S9). As in studies on the nuclear pore complex (22), a combination of proteomic and genomic approaches may prove effective in identifying the components that constitute and/or regulate NCAP movement on this pathway and the function(s) of the NCAPP1 gene family in orchestrating plant development.

Supporting Online Material

Materials and Methods

Figs. S1 to S9

Movies S1 to S3


  • * To whom correspondence should be addressed. E-mail: wjlucas{at}


View Abstract

Stay Connected to Science

Navigate This Article