Low Water Potential and At14a-Like1 (AFL1) Effects on Endocytosis and Actin Filament Organization^,

At14a-Like1 (AFL1) counteracts low water potential inhibition of endocytosis, alters actin filament organization, and partially colocalizes with actin filaments.

Abstract

At14a-Like1 (AFL1) is a stress-induced protein of unknown function that promotes growth during low water potential stress and drought. Previous analysis indicated that AFL1 may have functions related to endocytosis and regulation of actin filament organization, processes for which the effects of low water potential are little known. We found that low water potential led to a decrease in endocytosis, as measured by uptake of the membrane-impermeable dye FM4-64. Ectopic expression of AFL1 reversed the decrease in FM4-64 uptake seen in wild type, while reduced AFL1 expression led to further inhibition of FM4-64 uptake. Increased AFL1 also made FM4-64 uptake less sensitive to the actin filament disruptor Latrunculin B (LatB). LatB decreased AFL1-Clathrin Light Chain colocalization, further indicating that effects of AFL1 on endocytosis may be related to actin filament organization or stability. Consistent with this hypothesis, ectopic AFL1 expression made actin filaments less sensitive to disruption by LatB or Cytochalasin D and led to increased actin filament skewness and decreased occupancy, indicative of more bundled actin filaments. This latter effect could be partially mimicked by the actin filament stabilizer Jasplakinolide (JASP). However, AFL1 did not substantially inhibit actin filament dynamics, indicating that AFL1 acts via a different mechanism than JASP-induced stabilization. AFL1 partially colocalized with actin filaments but not with microtubules, further indicating actin-filament–related function of AFL1. These data provide insight into endocytosis and actin filament responses to low water potential stress and demonstrate an involvement of AFL1 in these key cellular processes.

Even a moderate severity of water limitation during drought (moderate decline in water potential [ψ_w]) alters plant development and causes a wide range of cellular changes. These responses to moderate severity low ψ_w are distinct from mechanisms involved in survival of severe low ψ_w and dehydration (Skirycz and Inzé, 2010; Clauw et al., 2016). The plasma membrane and cell wall are sites of many processes related to growth and stress resistance. Thus, trafficking mechanisms that control the composition of the plasma membrane can impact abiotic stress response. Examples of such mechanisms include regulated endocytosis of aquaporins to control membrane water permeability (Luu et al., 2012; Hachez et al., 2014; Chevalier and Chaumont, 2015) and endocytosis of abscisic acid (ABA) transporters and ABA receptors to control their abundance on the plasma membrane (Belda-Palazon et al., 2016; Park et al., 2016; Yu et al., 2016). Outside of these examples there is relatively little data on how drought stress, especially longer-term moderate severity stress, affects endocytosis, and little is known of stress-responsive proteins that regulate endocytosis.

Similarly, actin filaments are important for growth and morphogenesis (Szymanski and Staiger, 2018) and have been proposed to act as sensors or transducers of external signals along the plasma membrane (Staiger et al., 2009). Thus, actin filaments also are likely to have roles in responses to drought and other environmental stresses. Consistent with this idea, actin filament stabilization could increase survival of severe salt stress (Wang et al., 2010), and actin filament binding proteins have been shown to affect stomatal regulation (Liu and Luan, 1998; Zhao et al., 2011), cell swelling in hypo-osmotic media (Liu et al., 2013), and pathogen responses (Henty-Ridilla et al., 2013; Li et al., 2015). Also, pharmacological disruption of actin filaments altered the abundance of ABA-associated proteins (Takáč et al., 2017). Conversely, actin filament organization may be affected by leaf dehydration (Śniegowska-Świerk et al., 2016). As with endocytosis, there is relatively little data on how actin filament organization and dynamics are affected by longer-term moderate low ψ_w stress where plants have time to adjust cellular processes and acclimate to low ψ_w. Questions about stress effects on endocytosis and actin filaments are likely to intersect each other as actin filament disruption by Latrunculin B (LatB) or other pharmacological agents impairs endocytosis in plants (Šamaj et al., 2004). In other organisms, the connection of actin filaments to endocytosis is relatively well known; however, in plants, this is unclear as a number of endocytosis or actin filament-related proteins are not present or have differing function (Šamaj et al., 2004).

Previous work in our laboratory found that ectopic expression of the stress-induced protein At14a-Like1 (AFL1) led to enhanced growth maintenance and increased accumulation of the compatible solute Pro during low ψ_w (Kumar et al., 2015). Another study published at the same time indicated that overexpression of At14a, which is nearly identical to AFL1, could increase osmotic stress tolerance of suspension-cultured cells (Wang et al., 2015). At14a was also associated with susceptibility to Agrobacterium tumefaciens-mediated transformation (Sardesai et al., 2013).

At14a was first identified by immunoscreening an Arabidopsis (Arabidopsis thaliana) expression library using b-integrin-specific antisera (Nagpal and Quatrano, 1999). However, the integrin similarity of At14a and AFL1 is limited to a small domain (now annotated as “Domain of Unknown Function 677”). Although some studies have suggested At14a (and AFL1) to be membrane-spanning proteins similar to mammalian integrins (Lü et al., 2012; Sardesai et al., 2013; Langhans et al., 2017), we found that AFL1 is a peripheral membrane protein associated with the plasma membrane and endoplasmic reticulum (ER; Kumar et al., 2015). AFL1 interacted with AP2-2a, an adaptor protein involved in cargo selection and clathrin-coated vesicle formation (Kumar et al., 2015). Interestingly, structural predictions found similarity of AFL1 to amphiphysin, one of several types of BAR domain protein that can induce membrane curvature (Shen et al., 2012), as well as microfilament binding proteins actinin and spectrin. Mammalian amphiphysin, along with other BAR domain proteins, may connect clathrin-dependent endocytosis to actin microfilaments (Mooren et al., 2012). Amphiphysin, actinin, and spectrin have no clear orthologs in plants; conversely, AFL1 has no clear metazoan ortholog. Similarly, others have predicted that the C-terminal region of At14a has some similarity to human Myosin XVIIIB, while the At14a N-terminal domain is weakly related to human dynein H chain 7 and laminin (Langhans et al., 2017). In addition to our previous work, one other study suggested that At14a affected both microtubule and actin filament organization in protoplasts (Lü et al., 2012). Despite these efforts, whether AFL1 affects endocytosis and cytoskeleton organization is unclear.

We investigated the effect of low ψ_w and AFL1 on endocytosis and cytoskeleton organization using methods well established in our laboratory to reproducibly impose moderate severity low ψ_w stress over extended periods of time. This was combined with quantification of bulk endocytosis (FM4-64 uptake) and cytoskeleton organization. We also quantified AFL1 colocalization with actin filaments, microtubules, and clathrin light chain (CLC). Moderate severity low ψ_w decreased FM4-64 uptake in wild type. This effect was counteracted by ectopic expression of AFL1 and exacerbated by decreased AFL1 expression. AFL1 made FM4-64 uptake more resistant to disruption of actin filaments by LatB, altered actin filament organization under low ψ_w, and partially colocalized with actin filaments. These data clarify the effects of low ψ_w on these key cellular processes and give new evidence that AFL1 is involved in endocytosis and actin filament organization.

RESULTS

FM4-64 Uptake Decreases during Low ψ_w Acclimation

Normalized FM4-64 dye uptake (Bashline et al., 2013) was used to access the overall effect of low ψ_w on endocytosis. FM4-64 uptake was assayed under unstressed conditions or after short- (6 h) and longer-term (96 h) exposure to a moderate-severity low ψ_w (−0.7 MPa). Low ψ_w treatments were imposed using high M_r polyethylene glycol (PEG), which cannot penetrate the cell wall and thus causes cytorrhysis, as typically observed for plants in drying soil, and not plasmolysis, as seen in solutions of low M_r osmoticum (Verslues et al., 2006). Over the 96-h period used in this study, Arabidopsis seedlings can acclimate to −0.7 MPa to maintain water content and have growth at ∼30% of the level seen in unstressed plants (Verslues, 2010; Kumar et al., 2015; Bhaskara et al., 2017).

In wild type, we found that FM4-64 uptake decreased by >40% after 96 h at low ψ_w (Fig. 1A). It was interesting to note that the greatest reduction in FM4-64 uptake occurred not at 6 h after transfer to low ψ_w, where the plants may still be experiencing reduced turgor, but rather at 96 h, where water content and turgor are similar to that in the unstressed control (Verslues, 2010). These data contrast with other studies of short-term dehydration or salt stress where increased endocytosis was suggested to be a way to decrease membrane area during cell shrinkage (Zwiewka et al., 2015). Our data instead indicated that endocytosis may be regulated in response to low ψ_w rather than being solely affected by physical factors such as changes in turgor or cell volume.

Figure 1.

AFL1 expression level influences the rate of endocytosis as measured by FM4-64 uptake and counteracts the inhibitory effect of low ψ_w on endocytosis. A, Effect of low ψ_w on FM4-64 uptake in root cells of wild type (W.T.). FM4-64 uptake was measured in four independent experiments and the low ψ_w (−0.7 MPa for 6 or 96 h) data were normalized to the unstressed control in each experiment. In each experiment, normalized FM4-64 uptake was measured for 10–20 cells. Data shown are the means ± se (n = 4) of the normalization results from each experiment. Asterisks (*) indicate significant difference from the unstressed control by one-sample Student’s t-test (P ≤ 0.05). B, Effect of 35S:AFL1 and 100 nm LatB treatment on FM4-64 uptake in the unstressed control and after 6 or 96 h of exposure to −0.7 MPa. Data were analyzed as described in (A) except that the 35S:AFL1 and LatB data were normalized versus wild type for each time point. Data shown are means ± se (n = 3–4) of the results from three to four independent experiments with FM4-64 uptake measured for 10–20 cells in each experiment. Asterisks (*) indicate significant difference versus the wild type by one-sample Student’s t-test (P ≤ 0.05). C, Representative images of FM4-64 uptake in root cells of unstressed or stress-treated (−0.7 MPa, 96 h) seedlings with or without 1-h treatment with 100 nm LatB. Scale bars = 20 μm. D, FM4-64 uptake in root cells of AFL1 DEX-inducible RNAi lines (AFL1 K.D.) and empty vector (E.V.) control. Two independent RNAi lines were used, and combined data from both lines are shown. Data normalization and replication are as described in (B), where two independent experiments were performed; except that in this case, each individual measurement was normalized versus the mean FM4-64 uptake of the E.V. at that time point (n = 40–75). Asterisks (*) indicate significant difference versus the wild type by one-sample Student’s t-test (P ≤ 0.05). The E.V. + DEX versus AFL1 K.D. + DEX and AFL1 K.D. with or without DEX (significant difference indicated by yellow asterisk) were also compared by Student’s t-test. E, Effect of low ψ_w on FM4-64 uptake in root cells of the E.V. control line. The E.V. line was tested both with and without DEX treatment. The 6- and 96-h low ψ_w treatments were normalized versus the unstressed control (time 0) data. Data shown are the means ± se of individual cells (n = 45–60) combined from two independent experiments. Asterisks (*) indicate significant difference from the unstressed control by one-sample Student’s t-test (P ≤ 0.05). FM4-64 uptake was not significantly affected by DEX at either 6 or 96 h (Student’s t-test, P ≤ 0.05).

AFL1 Promotes FM4-64 Uptake at Low ψ_w and Makes it More Resistant to Inhibition by LatB

To investigate the effect of AFL1 on FM4-64 uptake, we first assayed two transgenic lines that express 35S:AFL1-FLAG (hereafter referred to as “35S:AFL1”) and have severalfold increase in AFL1 protein level (Supplemental Fig. S1A; Kumar et al., 2015). We found that 35S:AFL1 expression could counteract the decrease in FM4-64 uptake caused by low ψ_w. At 6 and 96 h of low ψ_w treatment, 35S:AFL1 increased FM4-64 uptake by 25% to 35% (Fig. 1, B and C). Because the 35S:AFL1 data were normalized to wild type within each time point, the relative increase in FM4-64 uptake in 35S:AFL1 in the 96-h low ψ_w treatment was essentially a reversal of the reduction in FM4-64 uptake seen in wild type (compare Fig. 1, A and B). Conversely, dexamethasone (DEX)-inducible RNA interference (RNAi) knockdown of AFL1 (hereafter referred to as “AFL1 K.D.”) decreased FM4-64 uptake in the unstressed control as well as at 6 and 96 h after transfer to low ψ_w (Fig. 1D). The knockdown lines had only a partial decrease in endocytosis, consistent with the fact that AFL1 protein level was only moderately decreased (Supplemental Fig. S1B). DEX treatment of the empty vector line had a small but significant effect on endocytosis (Fig. 1D). However, the DEX-treated AFL1 K.D. line had a significantly larger effect under all conditions tested (Fig. 1D). When the mock- or DEX-treated empty vector lines’ FM4-64 uptake at 6 or 96 h of stress was normalized to the unstressed control, both had an essentially identical effect of stress that was similar to that seen in wild type (Fig. 1E, compare to Fig. 1A). Both the 35S:AFL1 and AFL1 K.D. data were consistent with AFL1 acting to promote endocytosis, as quantified by FM4-64 uptake.

LatB was used to see if the low ψ_w- and AFL1-induced changes in FM4-64 uptake were affected by actin filaments. LatB binds to globular actin and inhibits its incorporation into actin filaments (Holzinger and Blaas, 2016). In wild type, LatB treatment reduced FM4-64 uptake by 30% to 40% in both the control and low ψ_w treatments (Fig. 1, B and C). Interestingly, the same LatB treatment caused no decrease in FM4-64 uptake of 35S:AFL1 under either control or low ψ_w stress (Fig. 1, B and C).

The effect of LatB on AFL1-CLC colocalization was also tested as an indication of whether AFL1 effects on endocytosis could be related to actin filaments. LatB treatment decreased AFL1-CLC colocalization in the unstressed control and blocked the stress-induced increase in AFL1-CLC colocalization seen at 24 or 96 h after transfer to low ψ_w (Fig. 2A). At 6 h after transfer, AFL1-CLC colocalization was variable between cells and the effect of LatB was marginally nonsignificant. In the unstressed control, LatB clearly decreased the amount of CLC present along the along the plasma membrane (Fig. 2B). Foci of AFL1-CLC colocalization along the plasma membrane could be observed (some prominent examples of such foci are indicated by arrows in Fig. 2B), consistent with our previous results (Kumar et al., 2015). Whether stress or LatB treatments affected the frequency of such foci could not be reliably determined, as there was, in many cases, a nearly continuous band of AFL1-CLC colocalization along the cell periphery. This made it difficult to reproducibly define individual foci of colocalization. Despite this uncertainty, the LatB data together indicated that AFL1 effects on endocytosis may be related to actin filaments.

Figure 2.

Effect of LatB on AFL1-CLC colocalization. A, PCC analysis of root cells exposed to the indicated duration of low ψ_w (−0.7 MPa) stress and treated with mock solution or 100 nm LatB for 1 h. In each plot, the box contains the 25th–75th percentiles of data points, whiskers indicate the 10th–90th percentiles, and outlying data points are shown as gray circles. Gray line in each box indicates the mean, whereas the black line indicates the median (in some cases the black median line is obscured by the mean line). The P value for the comparison of mock to LatB treatment is shown in each graph. Data are means ± se (n = 32–35 cells for unstressed mock and LatB treatments, 20–40 cells for stress mock and LatB treatments) combined from two independent experiments. B, Representative images of AFL1-CLC colocalization. For each treatment, the top image shows the YFP-AFL1 localization, middle image shows CLC-mOrange, and the lower image (labeled “Co” for colocalization) is the merged YFP-AFL1 and CLC-mOrange images to show colocalization. Areas of AFL1-CLC colocalization are indicated by blue and white color. Yellow arrows indicate examples of AFL1-CLC colocalized foci along the plasma membrane. Scale bars = 20 μm.

Ectopic Expression of AFL1 Leads to More Aggregated Actin Filaments at Low ψ_w

To more directly determine the effects of low ψ_w and AFL1 on actin filaments, 35S:AFL1 (line 8-1; Supplemental Fig. S1) was crossed with a line expressing green fluorescent protein (GFP)-tagged Fimbrin Actin Binding Domain 2 (fABD2; Sheahan et al., 2004). Note that in experiments including actin filament visualization, the GFP-fABD2 line is referred to as “wild type” and GFP-fABD2 with AFL1 ectopic expression is referred to as “35S:AFL1.” Actin filament organization was quantified based on the degree of skewness, which indicates actin filament bundling or aggregation, and occupancy, which measures the dispersion of the GFP-fABD2 signal (Higaki et al., 2010; Henty et al., 2011).

Both low ψ_w and 35S:AFL1 affected actin filament organization. In unstressed plants (Fig. 3A, time 0), leaf cells had low skewness and high occupancy, indicative of a relatively dispersed actin filament array, while root cells had a substantially higher skewness and hypocotyl cells were intermediate. Low ψ_w increased skewness and decreased occupancy in all cell types (Fig. 3A), although the timing and extent of the effect differed between cell types. Interestingly, 35S:AFL1 increased skewness and decreased occupancy at low ψ_w but had no significant effect in the unstressed control (Fig. 3A). A similar trend was also observed in basal hypocotyl cells (Supplemental Fig. S2). Thick actin cables were more prevalent in 35S:AFL1 compared to wild type (Fig. 3B). These data indicated that increased expression of AFL1 led to more bundled actin filament arrays and this effect was most prominent during low ψ_w stress. The skewness and occupancy values we observed in unstressed wild type were overall similar to some previous reports (Higaki et al., 2010; Cai et al., 2014), but lower than others (for example Henty et al., 2011; Li et al., 2012), possibly because of our use of light-grown seedlings rather than dark-grown seedlings, use of media without added Suc, or imaging of young growing leaves rather than mature leaves.

Figure 3.

Ectopic expression of AFL1 (35S:AFL1) leads to more aggregated actin filaments at low ψ_w. A, Quantitative analysis of actin filament skewness and occupancy in the indicated cell types for unstressed seedlings (time 0) or seedlings exposed to moderate severity low ψ_w (−0.7 MPa) for 6 and 96 h. Data are means ± se (n = 10–15) combined from two independent experiments. Black asterisks (*) inside the wild-type bars indicate a significant difference compared to unstressed wild type (ANOVA, P ≤ 0.05). Red asterisks above the 35S:AFL1 bars indicate a significant difference between wild type and 35S:AFL1 at that time point. B, Representative images of cells used in the skewness and occupancy measurements. Images shown are from unstressed plants and plants exposed to −0.7 MPa for 96 h. Images show the increased prevalence of thick actin cables in 35S:AFL1 at low ψ_w, particularly in leaf tissue, consistent with the increased skewness and decreased occupancy shown in (A). W.T. = wild type. Scale bars = 20 μm. C, Quantitative analysis of JASP effect on actin filament skewness and occupancy in apical hypocotyl cells of unstressed seedlings (time 0) or seedlings exposed to moderate severity low ψ_w (−0.7 MPa) for 6 and 96 h. Seedlings were treated with 100 nm JASP for 1 h or mock treatment. Data are means ± se (n = 12–16) combined from three independent experiments. Letters on top of each bar indicate significantly different groups (ANOVA, P ≤ 0.05) within each time point. For the wild-type (W.T.) mock and 35S:AFL1 mock data, asterisks within the bars indicated significant difference compared to time 0 (ANOVA, P ≤ 0.05). D, Representative images of cells from unstressed control (time 0) or 96-h stress treatment for the data shown in (C). W.T. = wild type. Scale bars = 20 μm.

These differences in actin filament organization caused by 35S:AFL1 could indicate a specific effect of AFL1 or a generally increased actin filament stability, which then indirectly leads to more extensive bundling. We tested this possibility by treating wild type and 35S:AFL1 with a low concentration (100 nM) of Jasplakinolide (JASP). JASP promotes actin polymerization and filament stability (Holzinger and Blaas, 2016). In unstressed wild type, JASP treatment led to increased skewness and decreased occupancy similar to the effect of 35S:AFL1 at low ψ_w (Fig. 3C). In 35S:AFL1, JASP treatment increased skewness and decreased occupancy in the unstressed control but had no additional effect at low ψ_w (Fig. 3C). Inspection of individual images (representative examples shown in Fig. 3D) confirmed that both JASP and 35S:AFL1 increased the prevalence of large actin structures. However, the patterns were not identical in that 35S:AFL1 cells often had long, mostly longitudinal, actin bundles, whereas JASP treatment produced many types of actin aggregates, including large actin foci with many branches (Fig. 3D). This was consistent with a previous report that treatment of hypocotyl cells with higher concentration of JASP (5 μM) produced short, thick actin filament aggregates that were less dynamic than untreated actin filaments (Sampathkumar et al., 2011).

Pearson Correlation Coefficient (PCC) analysis over a time series of images was used to determine how low ψ_w and 35S:AFL1 affected the dynamic rearrangement of actin filaments. In wild-type apical hypocotyl cells, there was a decrease in actin dynamics at 6 h of low ψ_w treatment (indicated by slower decline of PCC; Fig. 4A); however, by 96 h, the PCC profile returned to that of unstressed plants. The profiles of PCC decline we observed in wild type were of similar magnitude and pattern to those observed in previous studies (Vidali et al., 2010; Cai et al., 2014; Li et al., 2015). In the unstressed control, 35S:AFL1 had significantly faster PCC decline than wild type (Fig. 4B), indicating more dynamic rearrangement of actin filaments. After 96 h at low ψ_w, 35S:AFL1 had the opposite (but less dramatic) effect to decrease actin filament dynamics (Fig. 4B). Together, the data indicated that 35S:AFL1 could have large effects of actin filament organization while not substantially suppressing actin filament dynamics (or even increasing dynamics in the unstressed control). In contrast, JASP treatment has been shown to largely block the decline in PCC over a similar time course (Vidali et al., 2010). This, along with the different appearance of actin filaments in JASP-treated versus 35S:AFL1 plants, indicated that the mechanism by which 35S:AFL1 promotes a more bundled and less dispersed actin filament organization is likely to be different, or more specific, than that of JASP.

Figure 4.

Actin filament dynamics in wild type and 35S:AFL1 under control or low ψ_w treatments. A, Quantification of actin filament dynamics by PCC analysis in apical hypocotyl cells of wild-type seedlings in the unstressed control or after 6- and 96-h exposure to moderate severity low ψ_w (−0.7 MPa). Data are means ± se (n = 17–22) combined from three independent experiments with two seedlings and 3–4 cells per seedling analyzed in each experiment. Asterisk indicates significant difference compared to the control (ANOVA, P ≤ 0.05). B, Comparison of PCC profiles for wild type and 35S:AFL1 in the unstressed control and −0.7 MPa low ψ_w stress for 6 or 96 h. Data are from cells in the apical region of the hypocotyl. Wild-type data are the same as in (A) and are replotted here for clarity of presentation. Data are means ± se (n = 17−22) combined from three independent experiments as described for (A). Red asterisks indicate significant differences between wild type and 35S:AFL1 (ANOVA, P ≤ 0.05).

35S:AFL1 Plants Are More Resistant to the Effects of LatB and Cytochalasin D on Actin Filaments, But Are Less Effected in Microtubule Stability

We also observed the effects of two actin filament disruptors, LatB and Cytochalasin D (CytD), on actin filaments in wild type and 35S:AFL1. While LatB inhibits polymerization by binding actin monomers, CytD inhibits polymerization by binding to the barbed end of actin filaments. Both compounds caused extensive disruption of actin filaments in wild type (Fig. 5, A and B). In contrast, 35S:AFL1 maintained more extensive actin filament arrays in the presence of either LatB or CytD (Fig. 5, A and B). This was consistent with the results in Figure 1B showing that FM4-64 uptake in 35S:AFL1 was not affected by LatB. As our attempt to cross AFL1 K.D. lines with the fABD marker line were unsuccessful (AFL1 knockdown could no longer be observed after crossing), Phalloidin staining was used to investigate actin filament organization 35S:AFL1 and AFL1 K.D. Phalloidin staining of 35S:AFL1 indicated more intact actin filaments after LatB treatment (Supplemental Fig. S3A), consistent with the results using the fABD visualization of actin filaments (Fig. 5). AFL1 K.D. did not differ from wild type without DEX induction (Supplemental Fig. S3B), but had less extensive actin filaments than the empty vector control after treatment with DEX and 50-nm LatB (Supplemental Fig. S3C). Also consistent with these results, 35S:AFL1 plants had increased resistance to LatB inhibition of root elongation (Fig. 5C).

Figure 5.

Ectopic AFL1 expression (35S:AFL1) alters actin filament response to LatB and CytD. A, Representative images of actin filaments in leaf cells of wild type and 35S:AFL1 seedlings (actin filaments visualized by expression of GFP-fABD2) after mock treatment or treatment with 100 nm LatB for 1 h. Scale bars = 20 μm. B, Representative images of actin filaments in leaf cells of wild type and 35S:AFL1 seedlings after mock treatment or treatment with 100 nm CytD for 1 h. Scale bars = 20 μm. C, Response of wild-type and 35S:AFL1 root elongation to LatB or oryzalin. Four-d-old seedlings were transferred to plates containing the indicated concentrations of LatB or oryzalin and root elongation over the subsequent eight days measured. Data are means ± se (n = 16) combined from two independent experiments. Asterisks (*) indicate significant differences (P ≤ 0.05) of 35S:AFL1 compared to wild type (no significant differences were observed for the oryzalin treatments). D, Representative images of microtubule organization in leaf cells from wild type and 35S:YFP-AFL1 plants expressing the mCherry-MTUB microtubule marker. Seven-d-old seedlings were sprayed with 20 μm oryzalin or a mock control 1 h before imaging. Scale bars = 20 μm.

In contrast, there was no difference in root elongation sensitivity to oryzalin, which blocks tubulin polymerization and causes loss of microtubule organization (Fig. 5C). We also found that plants expressing 35S:yellow fluorescent protien (YFP)-AFL1 and mCherry-tagged MAP4 Microtubule Binding Domain (mCherry-MTUB) had a similar loss of microtubule organization upon oryzalin treatment as plants expressing mCherry-MTUB in the wild-type background (Fig. 5D). The combined data of these experiments showed that AFL1 clearly affected actin filament stability but did not have a similar effect on microtubules.

AFL1 Partially Colocalizes with Actin Filaments But Not with Microtubules

The above results raise the question of whether AFL1 directly regulates actin organization by binding to actin filaments. Cosedimentation assays conducted with purified recombinant AFL1 and in vitro-polymerized actin filaments found no evidence that AFL1 could directly interact with actin filaments (Supplemental Fig. S4). However, the results were complicated by the fact that AFL1 was only partially soluble under conditions that allow actin filament polymerization. While these results suggested that AFL1 does not directly bind actin filaments, we cannot rule out more specific AFL1 binding, such as binding to filament ends, which would be difficult to detect in this type of assay. We also emphasize that these data do not exclude other possibilities, such as a requirement for additional proteins or posttranslational modification of AFL1 to mediate AFL1 association with actin filaments.

As an alternative approach to investigate possible AFL1-cytoskeleton association, we constructed lines expressing YFP-AFL1 and mCherry-MAP4 or mCherry-tagged fABD2 (fABD2-mCherry) to assay AFL1-actin filament colocalization or YFP-AFL1 and mCherry-MTUB for AFL1-microtubule colocalization. Projected images of Z stacks across the plasma membrane, cell cortex, and cytoplasm in leaf cells showed foci of AFL1 colocalization with actin filaments along with extensive colocalization of AFL1 with thick actin cables (Fig. 6A; examples of AFL1-fABD2 colocalization foci are indicated by arrows). AFL1 did not colocalize with fine actin filaments except for the aforementioned foci. Cells where thick actin cables were visible had more extensive AFL1-fABD2 colocalization than cells where only dispersed actin filaments were visible (Fig. 6A). In the Z-stack images, the colocalization observed may have included both plasma membrane and endomembrane (most likely ER)-associated AFL1. We also examined single optical slices through the cell interior to more specifically examine AFL1-actin filament colocalization along the cell periphery, possibly associated with the plasma membrane (Supplemental Fig. S5). In these images also, we found extensive AFL-actin colocalization and many foci of colocalization along the cell periphery.

Figure 6.

Colocalization of AFL1 with actin filaments and microtubules in leaf cells. A, Representative images of YFP-AFL1 and actin microfilaments (visualized by fABD2-mCherry) in leaf cells of unstressed seedlings or 96 h after transfer to −0.7 MPa stress treatment. In the merged images (“Co-localization”), areas of AFL1-fABD2 colocalization are indicated by blue and white color. Yellow arrows indicate examples of AFL1-fABD2 colocalized foci. Quantification of the extent of colocalization by PCC is indicated in each merged image. Green boxes indicate the area of interest used to calculate PCC. Scale bars = 20 μm. B, Representative images of YFP-AFL1 and microtubules (visualized by mCherry-MTUB) in leaf cells of unstressed seedlings or 96 h after transfer to −0.7 MPa stress treatment. Data presentation are as described in (A).

To further quantify AFL1-actin filament colocalization along the cell periphery, and to check colocalization in the same cell type used for FM4-64 uptake and CLC colocalization assays, we imaged single optical slices through root cells and used PCC analysis to quantify the extent of colocalization. AFL1-fABD2 colocalization PCC values were relatively high in the unstressed control, decreased at 6 h after transfer to low ψ_w, and then partially recovered at 96 h (Fig. 7A). Some foci of AFL1-actin filament colocalization could be observed along the periphery of cells (see control images in Fig. 7B), although the diffuse band of actin filaments around the cell periphery often obscured such individual foci. The AFL1-fABD2 PCC values observed in root cells were overall similar to those observed in leaf cells (compare Fig. 7A to Fig. 6A and Supplemental Fig. S5A). Interestingly, the decreased AFL1-actin filament colocalization at low ψ_w differed from AFL1-CLC colocalization, which increased under low ψ_w (compare Fig. 7A to Fig. 2).

Figure 7.

Quantitative analysis of AFL1 colocalization with actin filaments and microtubules in root cells. A, PCC analysis of root cells in the unstressed control or after exposure to the indicated duration of low ψ_w (−0.7 MPa) stress. In each plot, the box contains the 25th–75th percentiles of data points, whiskers indicate the 10th–90th percentiles, and outlying data points are shown as circles. Red line in each box indicates the mean, whereas the black line indicates the median (in some cases the black median line is obscured by the mean line). Red asterisks indicated significant difference (P ≤ 0.05) compared to the time-0 unstressed control. Data are means ± se (n = 40–67 for AFL1-fABD2 colocalization and 34-45 for AFL1-MTUB colocalization) combined from three independent experiments. B, Representative images of YFP-AFL1 and actin microfilaments (visualized by fABD2-mCherry) and microtubules (visualized by mCherry-MTUB) in root cells of unstressed seedlings or 96 h after transfer to −0.7 MPa stress treatment. In the merged images (“Co-localization”), areas of colocalization are indicated by blue and white color. Yellow arrows indicate examples of AFL1-fABD2 colocalized foci. Quantification of the extent of colocalization by PCC is indicated in each merged image. Green boxes indicate the area of interest used to calculate PCC. Scale bars = 20 μm.

In contrast, there was less colocalization between AFL1 and microtubules. Average PCC values for AFL1-MTUB colocalization were near 0 (indicating only random overlap of the YFP-AFL1 and mCherry-MTUB signals) in the unstressed control for both leaf and root cells. (Figs. 6B and 7; Supplemental Fig. S5B). In root cells, the AFL1-MTUB colocalization increased slightly 6 h after transfer to low ψ_w (Fig. 7A). However, the colocalization observed at this time was not along microtubule strands, and at this time after transfer to low ψ_w, it is expected that microtubules are at least partially disorganized. We also did not observe foci of AFL1-MTUB colocalization comparable to the foci of AFL1-fABD2 colocalization. In cases of relatively high AFL1-MTUB PCC values, the overlapping signal occurred in aggregates, which were more prevalent at 6 h after transfer to low ψ_w, and not along intact microtubules (examples of such aggregates can be seen in Fig. 6B and Supplemental Fig. S5B). These microtubule observations were consistent with the oryzalin data, indicating that AFL1 had little effect on microtubules. The AFL1-MTUB data also provide an important comparison to the AFL1-fABD2 data, showing that the pattern of AFL1 actin filament colocalization is specific and likely to indicate a functionally important association. At the same time, the AFL1-fABD2 colocalization patterns, more extensive colocalization of AFL1 with CLC than with actin filaments, and lack of detectable AFL1 actin filament binding in vitro all suggest that AFL1 does not indiscriminately associate with actin filaments but rather has a specific pattern of colocalization that may depend on the presence of other actin-associated proteins.

DISCUSSION

There is little information on how drought acclimation influences endocytosis and actin filament organization or of the proteins involved in the effects of low ψ_w stress on these key cellular processes. Plant-specific proteins are likely to be involved in regulating these processes, but are little known. Our demonstration that plants exposed to moderate-severity low ψ_w for an extended period of time have decreased endocytosis (as measured by FM4-64 uptake) and altered actin filament organization show new cellular aspects of low ψ_w and drought acclimation. AFL1 promoted endocytosis and influenced actin filament organization under low ψ_w and also had a specific pattern of colocalization with actin filaments. Together with previous reports of increased AFL1 protein abundance under low ψ_w, enhanced growth maintenance of 35S:AFL1 lines under low ψ_w, and interaction of AFL1 with endocytosis-related proteins (Kumar et al., 2015), the data presented here indicate that AFL1 has a role in coordinating endocytic trafficking and actin filament organization with environmental signals such as low ψ_w.

Endocytosis is a major mechanism to control plasma membrane organization and protein content (Baisa et al., 2013; Baral et al., 2015; Fan et al., 2015). Because of this, it also influences many sensing and signaling pathways at the plasma membrane. Our observation of decreased bulk endocytosis rates under low ψ_w may at first seem to contradict previous reports of increased endocytosis in response to salt or osmotic stress (Zwiewka et al., 2015; Xia et al., 2016). However, these studies only examined the first few hours of osmotic stress where turgor is reduced and used low M_r solutes that cause plasmolysis (shrinkage of the plasma membrane from the cell wall). In contrast, we found significant reduction of FM4-64 uptake at 96 h after transfer to low ψ_w. At this time point, and given the moderate stress severity used (−0.7 MPa), turgor pressure is positive and has largely recovered for initial reduction of turgor that occurs in the first hours after transfer to low ψ_w (Verslues, 2010). Thus, the reduced endocytosis we observed is unlikely to be a direct result of change in turgor or cell shrinkage but more likely to result from specific regulatory mechanisms controlling endocytosis. From our observations and relevant literature, we can hypothesize that in plants there may be both a direct effect of cell shrinkage to stimulate endocytosis as well as additional regulatory mechanisms that can alter endocytosis independently of turgor or cell volume change. This hypothesis applies to bulk endocytosis and does not exclude that individual proteins have different responses depending on their function.

These data also demonstrate that AFL1 effects endocytosis. Increased AFL1 expression promoted FM4-64 uptake during low ψ_w stress, while decreased AFL1 expression inhibited FM4-64 uptake in all conditions tested. Whether this involves AFL1 interaction with AP-2a or interactions with other plasma membrane proteins or lipids is under investigation. Also, given the substantial effect of AFL1 on FM4-64 uptake, it will be of interest to use lines with increased or reduced AFL1 expression to identify proteins whose plasma membrane abundance is controlled by AFL1 and low ψ_w-dependent mechanisms. This could reveal further stress-responsive plasma membrane proteins involved in low ψ_w response.

For AFL1, it is tempting to speculate that its effects on endocytosis and actin filaments are linked. The ability of AFL1 to stabilize FM4-64 uptake against LatB inhibition and the foci of AFL1-actin filament colocalization seen along the cell periphery all seem consistent with this idea. Although compelling, such a hypothesis requires further investigation as our data also indicate that AFL1 may not interact directly with actin filaments but rather needs additional proteins or posttranslational modifications to mediate its effect on actin filament organization. The colocalization of AFL1 and actin exhibited two distinct patterns. While the small puncta of colocalization along the plasma membrane suggest an endocytosis or trafficking-related function, the colocalization along thick actin filaments is reminiscent of proteins that link ER to actin filaments (see Cao et al., 2016). We previously observed ER-like patterns of AFL1 localization and found AFL1 in both plasma membrane and endomembrane fractions (Kumar et al., 2015). However, it was unclear whether AFL1 was inside the ER or associated with the cytoplasmic side of the ER membrane. Given the pattern of AFL1-actin filament colocalization, it seems likely that AFL1 is also present on the cytoplasmic side of the ER membrane. Interestingly, there is recent evidence that ER-associated proteins can directly participate in endocytosis (Stefano et al., 2018). Whether AFL1 is involved in such a mechanism, as well as the mechanism connecting AFL1 to actin filaments, is under investigation in our laboratory.

As our interest in AFL1 stems, in part, from its ability to promote growth at low ψ_w (Kumar et al., 2015), the question of whether AFL1 promotes growth via its effect on endocytosis (which could alter the plasma membrane protein profile), or via its effect on actin filament organization (which could alter a range of trafficking and organelle positioning functions), is also pertinent for further study. These possibilities are not mutually exclusive, and the possibility that more-indirect mechanisms are involved should be kept in mind. For example, AFL1 may affect cytokinin signaling or response (Sardesai et al., 2013; Kumar et al., 2015), and it has been recently reported that cytokinins promote actin bundling associated with rapid cell elongation in roots (Takatsuka et al., 2018). Conversely, AFL1 could be involved in mediating the effect of cytokinin on actin filament bundling.

More broadly, our observations on AFL1 illustrate how endocytosis and actin filament organization are less understood in plants than in yeast or metazoans. In yeasts, which have turgor pressure, endocytosis is dependent on microfilaments (Aghamohammadzadeh and Ayscough, 2009) and patches of actin polymerization generate force to drive membrane invagination against the outward force of turgor pressure (Carlsson and Bayly, 2014; Lewellyn et al., 2015). In metazoan cells, actin filaments are not always required for endocytosis but are required when the membrane is under tension (Aghamohammadzadeh and Ayscough, 2009; Boulant et al., 2011; Mooren et al., 2012). Plant cells have high turgor pressure and thus, similar to yeast, could be hypothesized to require actin patches and actin filament polymerization to drive membrane invagination (Mooren et al., 2012); however, such actin patches have not been reported in plants, and the mechanisms by which actin filaments are connected to endocytosis in plants are little known. Mammalian actin-related protein (ARP)2 and ARP3 coordinate actin polymerization at endocytosis sites; however, in plants, these proteins have different localization patterns and are not clearly related to endocytosis (Konopka et al., 2008; Zhang et al., 2013). Plant-specific proteins, such as AFL1, which have roles in these basic cellular processes remain largely uncharacterized. Further characterization of AFL1 and the mechanisms by which it associates with actin filaments and influences endocytosis may help reveal plant-specific aspects of these basic cellular processes while also revealing important aspects of drought resistance.

MATERIALS AND METHODS

Plant Materials and Growth Conditions

In a previous study, we demonstrated that 35S:YFP-AFL1 and 35S:AFL1-FLAG plants have essentially identical phenotypes (Kumar et al., 2015). FM4-64 uptake experiments were conducted using lines with high expression of 35S:AFL1-FLAG (lines 4-2 and 8-1, Supplemental Fig. S1A; Kumar et al., 2015), and line 8-1 was crossed with 35S:GFP-fABD2 (kindly provided by Christopher Staiger, Purdue University) to visualize actin filaments. Colocalization of AFL1 and CLC was assayed using a 35S:YFP-AFL1/CLC-mOrange line previously generated by Kumar et al. (2015). Lines with DEX-inducible RNAi knockdown of AFL1 were described in Kumar et al. (2015). Extraction and immunoblot detection of AFL1 in transgenic plants was conducted as described in Kumar et al. (2015). Transgenic lines used for AFL1 colocalization with actin filaments or microtubules are described below in the section "AFL1 Colocalization with Actin Filaments and Microtubules".

Plants were routinely propagated for seed production in a growth room at 23°C and 16-h light period. For plate experiments, seeds were sterilized and plated on agar plates followed by stratification for 3–4 d at 4°C and then placed vertically in the growth chamber at 23°C and continuous light (70–100 μmol photons m⁻² s⁻¹) as previously described for experiments in our laboratory (Kumar et al., 2015; Bhaskara et al., 2017). The standard (unstressed) growth media consisted of half-strength Murashige and Skoog (MS) salts, 2 mm 2-(n-morpholino)ethanesulfonic acid buffer (pH 5.7), and 1.5% (w/v) agar with no sugar added.

Low ψ_w Stress, Pharmacological Treatments, and Growth Assays

Low ψ_w stress was imposed by transferring 7-d-old seedlings (for actin filament organization, FM4-64, and colocalization experiments) to −0.7 MPa PEG-8000 infused agar plates prepared using an established protocol (Verslues et al., 2006). To assay root elongation response to LatB or oryzalin, 4-d-old seedlings were transferred to plates containing the indicated concentrations of LatB or oryzalin, and root elongation was measured over the subsequent 7 d. Stocks of LatB (Sigma-Aldrich) and oryzalin (Sigma-Aldrich) were made in ethanol and stored at −20°C. For actin filament observations and colocalization experiments, 100 nm LatB was sprayed onto seedlings 1 h before experimental observations were conducted (Staiger et al., 2009). Mock control plates were sprayed with 0.02% ethanol solution. For CytD (Sigma-Aldrich) and JASP (Sigma-Aldrich), stocks were made in dimethyl sulfoxide (DMSO) and stored at −20°C. For observing their effect on actin organization, 100 nm CytD or JASP (or 0.5% DMSO as a mock control) was sprayed onto seedlings 1 h before observation. For observing the effect of oryzalin on microtubules, 20 μm oryzalin (or same concentration of ethanol without oryzalin as a mock control) was sprayed onto seedlings 1 h before microtubule observation.

For dexamethasone (DEX) treatment, 4-d-old seedlings were transferred to plates containing 10 μM DEX and grown for another 3 d. On d 7, seedlings were transferred to −0.7 MPa PEG-infused plates containing 10 μM DEX or the fresh high ψ_w control plates containing DEX. At the time of transfer and every day thereafter, 30 μM DEX (or mock solution for the minus DEX control) was sprayed on the seedlings to maintain induction of the RNAi construct.

Analysis of FM4-64 Uptake, AFL1-CLC Colocalization, and Actin Filament Organization and Dynamics

An LSM-510 Meta confocal microscope (Zeiss) was used to observe actin filament organization in the hypocotyl, root elongation zone, and leaves. GFP-fABD2 was excited at 488 nm and emission detected at 500–550 nm. For colocalization experiments, EYFP-AFL1 was imaged using excitation at 488 nm and emission at 560–575 nm, while CLC-mOrange was imaged using excitation at 514 nm and emission detected at 530–590 nm. Previous analysis showed that there was no bleed-through in signal between the YFP-AFL1 and CLC-mOrange images (Kumar et al., 2015). To quantify the extent of colocalization, areas of interest encompassing 1–2 cells (in root) or one cell or portion of cell (leaf) were selected and PCC calculated using the Zeiss LSM-510 analysis software in Expert Mode. For quantification of actin filament skewness and occupancy, cells were imaged using a fixed exposure time and gain settings set for all images from each treatment and genotype and all time points of the experiment. A maximum intensity projection of 15–20 image stacks was subjected for image processing and analysis in the software ImageJ (National Institutes of Health) performed as described by Higaki et al. (2010). For each genotype and treatment, 10–15 seedlings were analyzed over two or more independent experiments.

For analysis of actin filament dynamics, imaging was conducted using an LSM-880 (Zeiss), which allowed imaging at lower laser power to avoid bleaching. Individual cells from the apical hypocotyl region were imaged every 2 s over a 60-s time course. To quantify actin filament dynamics, 18–20 cells (combined from two to three independent experiments) from the apical hypocotyl region were used to calculate actin filament dynamics following the protocols in Li et al. (2015) and Vidali et al. (2010). FIJI ImageJ with the Coloc2 Plug-in was used to calculate a PCC between the time-0 image versus each subsequent image over the 60-s time course. The acquisition parameters such as laser power and gain settings were standardized and the same settings used for all actin dynamics experiments.

For assays of FM4-64 uptake, a stock solution of FM4-64 (Merck) was prepared in DMSO and aliquots stored at −20°C. At the time of analysis, the stock solution was diluted to 2 μM with half-strength MS medium for analysis of unstressed seedlings or with a −0.7 MPa solution of PEG-8000 in half-strength MS for analysis of seedlings growing on low ψ_w agar plates. Intact seedlings were removed from agar plates and incubated in 2-μM FM4-64 solution in a microfuge tube for 3 min on ice. After incubation, seedlings were washed with media lacking FM4-64, mounted on a glass slide and further incubated for 12 min before imaging. Fully expanded root epidermal cells (20–40 mm from the root tip) were imaged. Imaging of FM4-64 uptake used excitation at 488 nm and emission at 575–610 nm. Normalized FM4-64 internalization was quantified using ImageJ according to Bashline et al. (2013).

Actin filaments in leaves were stained with Alexa Fluor phalloidin (Thermo Fisher Scientific) as described in Panteris et al. (2006) and Yang et al. (2011), but with slight modification. Seven-d-old seedlings of AFL1 K.D. and empty vector control line were treated with DEX or mock control as described above and then treated with 50 nm LatB or mock control for 1 h. 35S:AFL1 lines were treated with 100 nm LatB or mock control in the same manner as for fABD imaging. This was followed by incubation with 300 μM m-maleimidobenzoyl-n-hydroxysuccinimide ester in 50 mm piperazine-n,n′-bis(2-ethanesulfonic acid), 5 mm magnesium sulfate, and 5 mm EGTA, pH 6.8 (PME buffer) plus 0.1% (v/v) Triton X-100 and 2% (v/v) DMSO for 30 min at room temperature in darkness. The samples were rinsed once with PME buffer and fixed with 2% paraformaldehyde in PME buffer for 1 h. Samples were then rinsed with PME buffer and incubated with 200-nm Alexa-488 phalloidin in PME buffer at 4°C overnight in darkness. Images were collected using an LSM-510 Meta confocal microscope (Zeiss) with filters and other settings as described in Panteris et al. (2006) and Yang et al. (2011).

Expression and Purification of Recombinant AFL1

BL21(DE3) competent cells were freshly transformed with pET28-AFL1 plasmid (to express His-tag AFL1). The LB-recovered primary transformant was used to directly inoculate an overnight Luria-Bertani broth (Miller) preculture. We noted that overgrowth of the overnight preculture would severely impair the recombinant protein expression. A 3.5-mL overnight preculture was used to inoculate 350 mL of fresh Luria-Bertani broth supplemented with 50 mg/L of kanamycin sulfate. The culture was grown in a 37°C shaker incubator until OD600 nm reached 0.4–0.6 and was induced by addition of 1.0-mm Isopropyl-β-d-thiogalactoside. The cells were grown for an additional 4 h at the same culture condition before harvest. The cell pellet was resuspended in 6 mL of prechilled lysis buffer (50 mm potassium P at pH 7.4, 200-mm NaCl, 10% [v/v] glycerol, 5-mm β-mercaptoethanol, 1× cOmplete, EDTA-free Protease Inhibitor Cocktail [Roche]) supplemented with 1% (v/v) IGEPAL CA-630 and 10 mm Imidazole-chloride at pH 8.0, followed by sonication on ice using a 1/4-inch microtip controlled by the S-4000 ultrasonic processor (Misonix Sonicators). The cell lysates were clarified by high-speed centrifugation at 13,000g for 30 min at 4°C (Avanti J-26 XP centrifuge with JA-25.50 rotor; Beckman Coulter), followed by a 0.45-μm filtration (Millex-HP, 33 mm, polyethersulfone; Merck Millipore) before applying to a 1.5 × 10 cm gravity Econo-column (Bio-Rad Laboratories) packed with 2 mL of His-60 Ni Superflow Resin (Takara Bio USA). The lysate was incubated with resin at 4°C for 2 h using end-over-end gentle agitation. The resin was washed by 15 column volumes of lysis buffer, followed by 15 column volumes of wash buffer (lysis buffer supplemented with 49 mm Imidazole-chloride at pH 8.0) before elution. AFL1 was eluted with 20 mL of HisB elution buffer (50 mm potassium P at pH 7.4, 200 mm NaCl, 10% [v/v] glycerol, 5 mm β-mercaptoethanol, 0.1% [v/v] IGEPAL CA-630, 400 mm Imidazole-chloride at pH 8.0) and was concentrated using a 10-kD molecular-mass cut-off Amicon Ultra Filter (Merck Millipore) before liquid N snap-freezing and storing at −80°C.

All purification fractions were examined using denaturing SDS-PAGE followed by Coomassie blue R-250 staining. The protein concentration was determined using the Bradford method (Bio-Rad Protein Assay Dye Reagent Concentrate; Bio-Rad Laboratories). This expression and purification procedure would typically generate ∼3 mg AFL1 recombinant protein from 350 mL of culture.

AFL1:Filamentous Actin Cosedimentation Assay

All the components in the Actin Binding Protein Biochem Kit, Muscle Actin (BK001; Cytoskeleton) were reconstituted and handled following the manufacturer’s instructions. The frozen AFL1 stock was slow-thawed on ice before conducting buffer exchange by Zeba Spin Desalting Column (7-kD molecular-mass cut-off; Thermo Fisher Scientific), which was pre-equilibrated with the modified actin buffer 20 mm Tris-chloride at pH 8.0, 200-mm potassium chloride, 1.8 mm calcium chloride, 2.0 mm magnesium chloride, and 5 mm β-mercaptoethanol (K200 buffer) or with other samples kept in the HisB elution buffer. The lyophilized actin was freshly reconstituted on ice and polymerized into filamentous actin at room temperature before use. The bovine serum albumin negative control and AFL1 were clarified using ultracentrifugation at 150,000g for 60 min at 4°C (Optima MAX-XP ultracentrifuge with TLA-120.2 rotor, Beckman Coulter). The AFL1 stock concentration in elution buffer and K200 buffer were 8.00–10.24 and 3.83–5.39 mg/mL, respectively. Protein and buffer components were assembled following the manufacturer’s instructions, incubated at 24°C for 30 min and sedimented by ultracentrifugation at 150,000g for 90 min at 24°C. The pellets were resuspended with 1% SDS. The supernatant and pellet fractions were examined by denaturing SDS-PAGE followed by either Coomassie blue R-250 staining (3.2% of the reaction loaded) or antipoly-His immunoblot (1.6% of the reaction loaded). For immunoblot detection of AFL1, protein was blotted onto polyvinylidene difluoride membrane by wet-tank procedure at 30 V for 16 h at 4°C. Membranes were blocked using 5% nonfat milk and probed with antipoly-His primary antibody (cat. no. H1029; Sigma-Aldrich) at 1:3,000 dilution followed by secondary antibody (rabbit anti-mouse IgG H&L [HRP], AbCam ab6728) at 1:10,000 dilution. Immunoreactive bands were visualized by the enhanced chemiluminescence method (Pierce ECL Western Blotting Substrate, Thermo Fisher Scientific) on a charge-coupled device imaging system (UVP ChemiDoc-It 510, Analytik Jena).

AFL1 Colocalization with Actin Filaments and Microtubules

Visualization of actin used previously described vectors (Ivanov and Harrison, 2014) in which expression of mCherry-tagged versions of either fABD2 (fABD2-mCherry) actin filament marker (Sheahan et al., 2004) or the MAP4 microtubule binding domain (mCherry-MTUB) microtubule marker (Marc et al., 1998) is driven by the Arabidopsis (Arabidopsis thaliana) Ubiquitin10 promoter. Vectors were obtained from Addgene (pCMU-ACTFr, Addgene cat. no. 61191; pCMU-MTUBr, Addgene cat. no. 61196) and used to transform Col-0 wild type. 35S:EYFP-AFL1 (line number 442/1-1 described in Kumar et al., 2015) was crossed with the fABD2-mCherry or mCherry-MTUB transgenics and double homozygous plants isolated. An LSM-880 confocal microscope (Zeiss) was used to observe the mCherry signal from both fABD2 and MTUB using excitation/emission wavelengths of 561 nm/630–650 nm, whereas AFL1 was observed by the excitation/emission wavelengths of 488 nm/500–550 nm. To quantify colocalization in root cells, a single image plane from both channels was captured by 40× apochromatic objective of the LSM-880 microscope. Regions of interest (where the focal plane went through the interior of one or more cells) were selected and PCC calculated using the Zeiss LSM-510 analysis software in Expert Mode. For leaf cells, PCC was calculated for both single-plane images as well as maximum projections of 15–20 Z-stack images.

Statistical Analysis

All experimental data shown were collected from two to four independent biological experiments. Significant differences were determined by analysis of variance (ANOVA), implemented in the software SigmaPlot 12 (Systat Software), or by Student’s t-test.

Supplemental Data

The following supplemental materials are available.

• Supplemental Figure S1. Immunoblot assay of AFL1 protein levels in AFL1 RNAi (AFL1 K.D.) and overexpression lines.

• Supplemental Figure S2. Actin filament skewness and occupancy of basal hypocotyl cells.

• Supplemental Figure S3. Phalloidin staining of filamentous actin in AFL1 K.D. and 35S:AFL1 lines.

• Supplemental Figure S4. AFL1 has little or no direct binding to actin filaments in vitro.

• Supplemental Figure S5. AFL1-fABD2 or AFL1-MTUB colocalization in leaf cells.