Selective mobility and sensitivity to SNAREs is exhibited by the Arabidopsis KAT1 K+ channel at the plasma membrane

Abstract

Recent findings indicate that proteins in the SNARE superfamily are essential for cell signaling, in addition to facilitating vesicle traffic in plant cell homeostasis, growth, and development. We previously identified SNAREs SYP121/Syr1 from tobacco (Nicotiana tabacum) and the Arabidopsis thaliana homolog SYP121 associated with abscisic acid and drought stress. Disrupting tobacco SYP121 function by expressing a dominant-negative Sp2 fragment had severe effects on growth, development, and traffic to the plasma membrane, and it blocked K(+) and Cl(-) channel responses to abscisic acid in guard cells. These observations raise questions about SNARE control in exocytosis and endocytosis of ion channel proteins and their organization within the plane of the membrane. We have used a dual, in vivo tagging strategy with a photoactivatable green fluorescent protein and externally exposed hemagglutinin epitopes to monitor the distribution and trafficking dynamics of the KAT1 K(+) channel transiently expressed in tobacco leaves. KAT1 is localized to the plasma membrane within positionally stable microdomains of approximately 0.5 microm in diameter; delivery of the K(+) channel, but not of the PMA2 H(+)-ATPase, to the plasma membrane is suppressed by Sp2 fragments of tobacco and Arabidopsis SYP121, and Sp2 expression leads to profound changes in KAT1 distribution and mobility within the plane of the plasma membrane. These results offer direct evidence for SNARE-mediated traffic of the K(+) channel and a role in its distribution within subdomains of the plasma membrane, and they implicate a role for SNAREs in positional anchoring of the K(+) channel protein.

Figure 1.

The KAT1 K⁺ Channel Coding Sequence Yields a Functional Channel Protein as a Fusion Construct with Photoactivatable GFP and Tagged with HA Epitopes. (A) Schematic of the haKAT1:GFP (top) and haKAT1:paGFP (bottom) constructs with HA epitopes, spacer sequences, and insertion points as indicated. Underlined amino acid residues and flanking residue numbers relate to the KAT1 channel sequence; residues in boldface compose the HA epitopes. (B) Protein gel blot analysis of haKAT1:paGFP and KAT1 after expression in Xenopus oocytes. SDS-PAGE of solubilized microsomal membrane fractions (8 μg total protein/lane) from oocytes 3 d after injections probed with monoclonal antibody to GFP (αGFP), then stripped and reprobed with monoclonal antibody to HA (αHA). The antibodies register a single band near 110 kD, consistent with the molecular mass of 105.6 kD for the fusion construct. (C) Steady state current voltage curves for oocytes expressing haKAT1:paGFP (closed circles) and control (water-injected) oocytes (open circles) recorded in 10 mM KCl. Data are means ± se of three recordings and show the characteristic appearance of inward current at voltages negative from −80 mV associated with the KAT1 K⁺ channel. Inset shows current relaxations from one control-injected oocyte and one oocyte expressing haKAT1:paGFP recorded under voltage clamp showing activation of the haKAT1:paGFP current. Voltage clamp cycles (data not shown): holding voltage, −50 mV; test voltage steps (eight), −90 to −160 mV. (D) and (E) Cross-sectional confocal fluorescence images of one Xenopus oocyte expressing haKAT1:paGFP before (D) and after (E) photoactivation of paGFP with 351/364-nm light from the UV laser. Excitation, 488 nm; emission, 505 to 530 nm.

Figure 2.

Tobacco Leaves Transfected with haKAT1:GFP and haKAT1:paGFP Yield a Peripheral Distribution of Punctate GFP Fluorescence. Confocal images of tobacco leaf epidermis expressing haKAT1:GFP at 72 h after transfection at low ([A] and [B]) and high ([C] and [D]) magnification with ([A] and [C]) and without ([B] and [D]) the corresponding bright-field overlay. Corresponding confocal images of tobacco leaf epidermis expressing haKAT1:paGFP before ([E] and [F]) and after ([G] to [I]) 4-μs photoactivation of paGFP within the boxed region indicated in (F). (E) and (H) are bright field only, and (H) and (I) are high-magnification images of the top portion of the photoactivated region in (G). Again, note the punctate, peripheral distribution of the fluorescence. Relative fluorescence activation time course (J) as a function of total pixel dwell time with 351/364-nm light indicated a 4- to 5-μs irradiation for maximum efficiency in photoactivation and a 30- to 50-fold increase in fluorescence yield when excited with 488-nm light. Data are means ± se of three experiments. Bars = 50 μm in ([A], [B], and [E] to [G]) and 5 μm in ([C], [D], [H], and [I]).

Figure 3.

haKAT1:paGFP Is Nonmobile at the Cell Periphery. Bright-field (top left frame) and GFP confocal fluorescence images from one experiment with tobacco epidermal cells expressing the fusion construct. Images are sections within the epidermal cell layer taken parallel to the leaf surface. Images were collected at intervals before and after 4-μs photoactivation with 351/364-nm light within the area circled in the first two frames. Time in seconds relative to photoactivation is as indicated for each frame. See Supplemental Movie 1 online for the full set of images. Note the spatial stasis of the punctate fluorescent pattern around the lateral surface of the adjoining epidermal cells as viewed in this optical section and the apparent lack of any lateral movement or dispersion of the fluorophore within the plane of the cell surface beyond the photoactivation boundary. See Figure 5C for quantitative analysis. Bar = 5 μm.

Figure 4.

Comparative Analysis of Fluorophore Mobilities by FRAP with H⁺-ATPase PMA2:GFP, Which Labels the Plasma Membrane, and GFP:HDEL, Which Labels the Endoplasmic Reticulum. (A) Plasma membrane labeled by H⁺-ATPase PMA2:GFP. (B) Endoplasmic reticulum labeled by GFP:HDEL. In each case, bright-field (top left frames) and GFP confocal fluorescence images from one experiment with tobacco epidermal cells expressing the fusion constructs are shown. Images were collected at intervals before and after photobleaching with 458/488-nm light within the areas circled in the first two frames. Images are cross sections through the epidermal cell layer taken parallel to the leaf surface. Time in seconds relative to photoactivation is as indicated for each frame. See Supplemental Movies 2 and 3 online for the full sets of images. Note the absence of any significant fluorescence recovery in (A) and its rapid recovery within the photobleached areas in (B). The nuclear ring, characteristic of labeling in the endoplasmic reticulum, is clearly visible in the bottom part of each frame in (B) and is seen to migrate slowly to the left. See Figures 5A and 5B for quantitative analysis. Bars = 5 μm.

Figure 5.

haKAT1:paGFP Is Stationary at the Cell Periphery. Kymographic analysis of data from Figures 3 and 4 for the H⁺-ATPase PMA2:GFP (A), GFP:HDEL (B), and haKAT1:paGFP (C). In each case, the fluorescence signal was taken over the time course of the experiment from pixels along a line traced around the edge of the cell, averaging over a width of 10 pixels (1.4 μm), as shown in the bright-field images (dotted lines, insets). Position along the line determines the horizontal axis, time progression determines the vertical axis, and fluorescence intensity is color-coded (see inset in [A]). Photobleached (b) and photoactivated (p) regions are indicated above each kymograph and are color-coded for the two lines in (B). Times of photobleaching/photoactivation are indicated at the left (lollipop) in each case. Photobleaching of the H⁺-ATPase (A) led to a virtually complete loss of fluorescence within the two bleached regions that showed no evidence of filling or lateral movement near the edges of the two regions during the time course of the experiment. By contrast, photobleaching of GFP:HDEL (B) was followed by a rapid filling and recovery of fluorescence and was associated with diagonal patterns of fluorescence intensity, indicative of lateral movement within the cell. Kymographic analysis of data from Figure 3 for haKAT1:paGFP (C) gave an effective image-inverse of the results for the H⁺-ATPase, as expected of the complementarity of photoactivation with photobleaching and confirming the absence of any measurable displacement of the fluorescence laterally over the surface of the cell. Arrows (color-coded with the bright-field inset) mark two of the brighter puncta; vertical lines of fluorescence intensity indicate the absence of local mobility.

Figure 6.

The KAT1 K⁺ Channel Resides in Clusters at the Plasma Membrane Surface. Confocal fluorescence images from dual-labeling experiments using protoplasts from tobacco leaf tissue previously transfected with haKAT1:GFP and bound with Alexa594-αHA. Bright-field composite (A), GFP (B), and chloroplast (C) fluorescence from one protoplast in tangential surface view without prior Alexa594-αHA treatment. The images show characteristic fine punctate fluorescence of the fusion constructs and an appreciable intrinsic fluorescence from the chloroplasts frequently observed, especially in the GFP channel (see [Q] to [T]). Analysis of one point near the center of this image set ([D]; analysis area boxed in inset at left) shows the fluorescence intensity pseudo-color-coded (bottom) and in three-dimensional surface representation (top). Fluorescence intensity was well-fitted to a two-dimensional Gaussian distribution function as , where A is the maxiumum fluorescence amplitude, x and y are the image coordinates, x_o and y_o are the corresponding coordinates of the fluorescence maximum, and b_x and b_y reflect the corresponding dimensional spread of the fluorescence signal away from the maximum. As shown, the analysis yielded visually satisfactory and statistically best fittings with an isodiametric surface and FWHM peak height of 528 ± 13 nm (gray surface overlay). Dual-labeling after 5-min exposure to Alexa594-αHA of protoplasts expressing haKAT1:GFP in tangential surface view ([E] to [J]) and in equatorial cross section ([K] to [P]) as composites with ([E] and [K]) and without ([F] and [L]) bright-field overlay shows colocalization ([G] and [M]) of the Alexa594 ([H] and [N]) and GFP ([I] and [O]) fluorescence signals distinct from that of the chloroplasts ([J] and [P]). Colocalization analysis ([G] and [M]) was determined as the relative fluorescence intensities of Alexa594 and GFP labeling along the dotted lines (tail to head = position left to right) shown in (F) and (L). Bright-field composite (Q), Alexa594 (R), GFP (S), and chlorophyll (T) fluorescence images of a protoplast from an untransfected tobacco leaf after Alexa594-αHA treatment are shown. This tangential surface view shows the absence of Alexa594-labeled surface structure, although the wide-spectrum fluorescence of the chloroplasts is evident, especially in the GFP channel. Fluorescence excitation: 543 nm (Alexa594-αHA), 488 nm (GFP, chloroplasts); emission: 585 to 615 nm (Alexa594-αHA), 505 to 530 nm (GFP), 560-nm long-pass filter (chloroplasts). Bars = 5 μm.

Figure 7.

KAT1 K⁺ Channels Distribute in Clusters with Prevalent Diameters of 0.5 to 0.6 μm. Relative fluorescence intensities were fitted to a two-dimensional Gaussian distribution function as in Figure 6, and the FWHM peak height was averaged over the two dimensions used as a measure of the cluster diameter. (A) Data comparing the frequency distribution of cluster diameters measured from protoplasts treated without (GFP) and with Alexa594-αHA (HA), with dimensions determined using the corresponding fluorescence signals. Diameter frequencies in each case were well fitted to a single, log-normal distribution function (solid curve) with a maximum of 532 ± 22 nm and peak spread coefficient of 0.329 ± 0.003. (B) Data, including those in (A), comparing the frequency distribution of haKAT1:GFP and haKAT1:paGFP cluster diameters in intact tobacco epidermal cells (see Figures 2, 3, and 8) with haKAT1:GFP cluster diameters in protoplasts as in (A). Diameter frequencies for expression in epidermal cells were well fitted to single, log-normal distribution function (solid curve) with a maximum of 509 ± 18 nm and peak spread coefficient of 0.359 ± 0.003.

Figure 8.

Cytosolic (Dominant-Negative) Sp2 Fragment of Tobacco SYP121 Selectively Suppresses KAT1 K⁺ Channel Traffic and Affects Its Distribution at the Plasma Membrane. (A) Three-dimensional reconstructions from confocal fluorescence images of tobacco epidermal cells expressing haKAT1:GFP ([a] to [c]) and the H⁺-ATPase PMA2:GFP ([d] to [f]) fusion constructs. GFP fluorescence is in green, and chloroplast fluorescence is overlayed in red. Excitation, 488 nm; emission, 505 to 530 nm (GFP) and 560-nm long-pass filter (chloroplasts). Expression of the fusion constructs only ([a] and [d]), together with the Sp2 fragment of NtSyp121 ([b] and [e]), and with the dominant-negative Rab1b-N121I mutant ([c] and [f]) to block export from the endoplasmic reticulum to the Golgi apparatus is shown. Note the punctate distribution of haKAT1:GFP (a) compared with the essentially uniform distribution of the PMA2:GFP marker (d) over the cell surfaces. A reticulate network, cytoplasmic strands, and a nuclear ring were evident in coexpression with the Rab1b-N121I mutant in both cases ([c] and [f]) and for the K⁺ channel in coexpression with the Sp2 fragment (b). For the H⁺-ATPase, coexpression with the Sp2 fragment had no visible effect on its distribution to the cell periphery (e). Bars = 20 μm. (B) Kymographic analysis of data from Figure 9 for haKAT1:paGFP. The fluorescence signal was taken over the time course of the experiment from pixels along a line traced around the edge of the cell, averaging over a width of 10 pixels (1.4 μm), as shown in the bright-field image (dotted line, inset). Position along the line determines the horizontal axis, time progression determines the vertical axis, and fluorescence intensity is color-coded (scale inset). Photoactivated (P) region is indicated above the kymograph, and time of photoactivation is indicated at the left (lollipop). The rapid lateral dispersal of the fluorescence signal after photoactivation doubled the predominant fluorescence spread over a period of 40 to 50 s and gives rise to the diagonal pattern of fluorescence intensities seen here (cf. Figure 5C). (C) and (D) Confocal dual-labeling experiments using protoplasts from tobacco leaf tissue previously transfected with haKAT1:GFP together with either the Sp2 fragment of tobacco SYP121 (C) or the Rab1b-N121I mutant (D). Protoplasts in each case were bound with Alexa594-αHA for 5 min prior to imaging and are shown here in tangential surface views as (left to right): bright-field composite, Alexa594, GFP, and chloroplast fluorescence. Imaging parameters were as in Figure 6. Colocalization analysis ([C], bottom inset) was determined as the relative fluorescence intensities of Alexa594 and GFP labeling along the dotted lines (tail to head = position left to right) as shown. Note the absence of Alexa594-αHA labeling in coexpression with the Rab1b-N121I mutant (D) and the presence of both Alexa594 and GFP signals with Sp2 fragment coexpression (C). With Sp2 coexpression, Alexa594-αHA and GFP labeling do not colocalize but do overlap, as expected with retention of a portion of the K⁺ channel protein in the endoplasmic reticulum. Bars = 5 μm.

Figure 9.

haKAT1:paGFP Is Mobile at the Epidermal Cell Periphery When Coexpressed with the Sp2 Fragment of NtSyp121. Bright-field (top left frame) and GFP confocal fluorescence images from one experiment with tobacco epidermal cells expressing the fusion construct. Images were collected at intervals before and after photoactivation with 351/364-nm light within the boxed area in the first two frames. Images are sections within the epidermal cell layer taken parallel to the leaf surface. Time in seconds relative to photoactivation is as indicated for each frame. Note that the nuclear ring, characteristic of labeling in the endoplasmic reticulum, was clearly visible after photoactivation along with two chloroplasts, just visible in the bright-field image, that enter the fluorescence focal plane later in the frame series. See Figure 8C for quantitative kymographic analysis and Supplemental Movie 4 online for the full sets of images. Imaging parameters were as in Figures 3 and 4. Bar = 5 μm.

Figure 10.

haKAT1:GFP Is Mobile at the Plasma Membrane When Coexpressed with the Sp2 Fragment of NtSyp121. (A) Bright-field (top left frame) and Alexa594 confocal fluorescence images from one experiment with tobacco protoplasts expressing the fusion construct. Protoplasts were bound with Alexa594-αHA for 5 min prior to imaging and are shown here in tangential surface view. Images were collected at intervals before and after photobleaching with 543-nm light within the circled area (a) in the second frame. Time in seconds relative to photobleaching is as indicated for each frame. Bar = 5 μm. (B) Fluorescence recovery analysis shows fluorescence intensities in areas (cross-referenced to circles [a], [b], and [c] in [A]) fitted by nonlinear least squares (Marquardt, 1963) to single exponential functions (solid line) for time points after photobleaching (t > 0 s). Note the recovery in the photobleached area (a) is accompanied by a loss in fluorescence in the neighboring area (b), but no appreciable change is seen in the more outlying area (c). Fitted time constants were as follows: (a), 6.5 ± 0.4 s; (b), 6.8 ± 0.3 s. See Supplemental Movie 5 and Figure1 online for the full image sequence, composite, and overlay color images. Imaging parameters were as in Figure 6.

Figure 11.

Dominant-Negative SNARE Fragments Affect Surface Clustering of the KAT1 K⁺ Channel as Well as Its Trafficking to the Plasma Membrane. (A) Frequency distributions of the K⁺ channel taken from ×20 (objective) magnification images at random from all 211 experiments expressing haKAT1:GFP or haKAT1:paGFP alone (Control), together with the Sp2 fragments of tobacco or Arabidopsis SYP121 (+Sp2), and together with the Rab1b-N121I mutant (+Rab1b-N121I). Localizations scored as peripheral only, reticulate, and reticulate with nuclear ring. Difficulties in identifying peripheral labeling at low magnification mean that a peripheral distribution cannot be excluded from the latter two categories. (B) Frequency distributions of cluster diameters for the KAT1 K⁺ channel from protoplasts of tobacco leaves expressing haKAT1:GFP alone (-Sp2) or together with the Sp2 fragments of tobacco or Arabidopsis SYP121 (+Sp2). Cluster diameters were determined as in Figure 6 using the Alexa594-αHA fluorescence signal. Data were fitted to log-normal functions (solid lines) as in Figure 7. Fitted parameters were as follows: diameter at maximum frequency, 532 ± 22 nm (-Sp2) and 3080 ± 70 nm (+Sp2); peak spread coefficient, 0.329 ± 0.003 (-Sp2) and 0.40 ± 0.02 (+Sp2). (C) Frequency distributions of cluster diameters for the KAT1 K⁺ channel from intact epidermal cells of tobacco leaves expressing haKAT1:GFP alone (-Sp2) or together with the Sp2 fragments of tobacco or Arabidopsis SYP121 (+Sp2). Cluster diameters were determined as in Figure 6 using the GFP fluorescence signal. Data were fitted to a log-normal function (solid lines) as in Figure 7. Fitted parameters were as follows: diameter at maximum frequency, 509 ± 18 nm (-Sp2) and 4300 ± 100 nm (+Sp2); peak spread coefficient, 0.359 ± 0.003 (-Sp2) and 0.43 ± 0.02 (+Sp2). Note that the inability to distinguish internal labeling from true localization to the plasma membrane in this case almost certainly skews the results with the Sp2 fragments to larger diameter values.

Figure 12.

Sp2 Fragment of Tobacco SYP121 Selectively Suppresses KAT1 K⁺ Channel Distribution to the Plasma Membrane. Protein gel blot analysis of the H⁺-ATPase (PMA2:GFP) and KAT1 K⁺ channel (haKAT1:GFP) localization after expression in tobacco leaves alone (Control), with the Sp2 fragment of tobacco SYP121 (+Sp2), and with the mutant Rab1b-N121I. SDS-PAGE (20 μg total protein/lane) of solubilized microsomal membranes after separation by two-phase partitioning to isolate plasma membrane (PM) and endomembrane (IM) fractions. After blotting, nitrocellulose filters were probed with polyclonal antibody to GFP and visualized by chemiluminescence (H⁺-ATPase) and radiotracer phosphor imaging (K⁺ channel).