Mechanism of microtubule plus-end tracking by the plant-specific SPR1 protein and its development as a versatile plus-end marker

Abstract

Microtubules are cytoskeletal polymers that perform diverse cellular functions. The plus ends of microtubules promote polymer assembly and disassembly and connect the microtubule tips to other cellular structures. The dynamics and functions of microtubule plus ends are governed by microtubule plus end–tracking proteins (+TIPs). Here we report that the Arabidopsis thaliana SPIRAL1 (SPR1) protein, which regulates directional cell expansion, is an autonomous +TIP. Using in vitro reconstitution experiments and total internal reflection fluorescence microscopy, we demonstrate that the conserved N-terminal region of SPR1 and its GGG motif are necessary for +TIP activity whereas the conserved C-terminal region and its PGGG motif are not. We further show that the N- and C-terminal regions, either separated or when fused in tandem (NC), are sufficient for +TIP activity and do not significantly perturb microtubule plus-end dynamics compared with full-length SPR1. We also found that exogenously expressed SPR1-GFP and NC-GFP label microtubule plus ends in plant and animal cells. These results establish SPR1 as a new type of intrinsic +TIP and reveal the utility of NC-GFP as a versatile microtubule plus-end marker.

Introduction

Microtubules are dynamic polymers of αβ-tubulin heterodimers and form different arrays to orchestrate critical cellular activities in eukaryotes. The plus ends of microtubules are of particular significance because they drive the assembly and disassembly of microtubules and connect microtubule tips to various cellular structures. Specialized proteins called microtubule plus end–tracking proteins (+TIPs)² specifically accumulate at growing microtubule plus ends and control their behavior and interactions. +TIPs can directly or indirectly target growing microtubule plus ends (1). For example, the evolutionarily conserved end binding (EB) proteins autonomously target growing microtubule plus ends by recognizing GDP/P_i-tubulin structures (2–7). In turn, EBs recruit a host of other proteins as part of a dynamically changing microtubule plus-end complex (8, 9).

Land plants contain a unique set of +TIPs (10–12), perhaps because they assemble morphologically and functionally distinct microtubule arrays compared with animals (13). One of the plant-specific +TIPs is the Arabidopsis thaliana SPIRAL1 (SPR1) protein. SPR1 was identified through a genetic screen for skewed root growth (14) and encodes a 12-kDa protein that localizes to the growing plus end of cortical microtubules and contributes to anisotropic cell expansion (11, 12, 15). SPR1 genetically and physically interacts with the A. thaliana END BINDING 1b (EB1b) protein (16), which tracks growing microtubule plus ends, similar to animal and yeast EB proteins (17, 18). However, SPR1 is also able to bind to free tubulin and the microtubule lattice independently of EB1b in vitro (16), raising the question of whether SPR1 needs EB1b for microtubule plus-end localization. Here we report that SPR1 is an autonomous +TIP and identify the regions and critical residues that are required for this activity. Finally, we show that a fusion protein consisting of the conserved N- and C-terminal regions of SPR1 fused to GFP labels growing microtubule plus ends in plant and animal cells, making it a versatile microtubule plus-end marker.

Results

SPR1 is an autonomous microtubule plus end–tracking protein

To study the mechanism of SPR1's +TIP activity, we took an in vitro reconstitution approach with purified recombinant SPR1-GFP protein (Fig. 1A) and dynamic microtubules using porcine tubulin. We found that SPR1-GFP on its own labels microtubule plus ends exclusively during the growth phase (Fig. 1, B and C, and Movie S1). We observed that SPR1-GFP also weakly labels growing microtubule minus ends (Fig. 1C and Movie S1), similar to yeast and mammalian EB proteins (2, 3). Furthermore, SPR1-GFP expressed under its native promoter labels microtubule plus ends in the eb1a;eb1b double mutant (Fig. 1D), demonstrating that SPR1 does not require EB1a and EB1b to localize to microtubule plus ends in vivo.

Figure 1.

SPR1 is an autonomous microtubule plus end–tracking protein. A, Coomassie-stained gel of recombinant SPR1-GFP protein. The arrowhead indicates the expected size (40 kDa) of the protein. Numbers on the left indicate the size of the protein bands in the ladder. B, representative image of SPR1-GFP (green) and dynamic rhodamine-labeled microtubules (red). Scale bar = 5 μm. C, representative kymograph of SPR1-GFP (green) and a dynamic rhodamine-labeled microtubule (red). x axis scale bar = 5 μm, y axis scale bar = 100 s. D, representative kymograph of SPR1-GFP (green) and RFP-TUB6 (red) in the eb1a;eb1b double mutant. x axis scale bar = 5 μm, y axis scale bar = 100 s. E, graph of SPR1-GFP signal intensity at the plus end of growing microtubules in vitro. Values represent mean ± S.D. n = 52, 89, and 133 for 0.1 μm, 0.5 μm, and 1 μm SPR1-GFP, respectively. Inset, the length of the comet tail as a function of SPR1-GFP concentration. Comet tail lengths were obtained by fitting the right halves of the signal intensity curves to single exponentials. a.u., arbitrary units. F, representative image of SPR1-GFP (green) bound to GTPγS microtubules grown from GMPCPP microtubule seeds (red). The montage on the right shows an example of a growing GTPγS microtubule being bound by SPR1-GFP. Scale bars = 5 μm. G–J, dot plots of microtubule growth rate (G), shortening rate (H), catastrophe frequency (I), and rescue frequency (J) at the indicated SPR1-GFP concentrations. Statistical analyses were conducted using one-way ANOVA. ****, p < 0.0001; ***, p < 0.001; **, p < 0.01; *, p < 0.05.

The SPR1-GFP signal is concentrated at the tips of growing microtubules and decays exponentially along the microtubule lattice in a comet-like pattern (Fig. 1, B and E). Increasing the concentration of SPR1-GFP protein enhanced the accumulation of SPR1 at the tips of growing microtubule plus ends and increased the length of the comet tails along the microtubule lattice (Fig. 1E). Human EB1 has been reported to result in either an increase (3) or decrease (6) in comet size with increasing EB1 concentration. To determine whether the plus end–tracking activity of SPR1 involves recognition of the nucleotide state of tubulin, we examined the binding of SPR1 to microtubules polymerized in the presence of the GTP analog guanosine 5′-3-O-(thio)triphosphate (GTPγS), which mimics the structure of growing microtubule ends (4–7). We found that SPR1-GFP preferentially binds to the GTPγS microtubule lattice compared with the guanosine-5′-[(αβ)-methyleno] triphosphate (GMPCPP) microtubule seeds (Fig. 1F). These data suggest that the SPR1 and EB proteins share a similar mechanism for recognizing growing microtubule ends.

To determine whether SPR1-GFP affects the behavior of microtubule plus-ends in vitro, we measured the dynamic instability parameters at various SPR1-GFP concentrations. We observed that SPR1-GFP inhibited the growth rate of microtubule plus ends and enhanced their catastrophe frequency in a concentration-dependent manner (Fig. 1, G and I). In contrast, SPR1-GFP did not significantly affect the shortening rate of microtubule plus ends and only modestly increased their rescue frequency at the highest SPR1-GFP concentration tested (Fig. 1, H and J). These observations (summarized in Table S1) are largely consistent with the reported changes in microtubule plus-end dynamics in spr1 loss-of-function mutants compared with the WT (19).

The conserved N-terminal region is the primary determinant of SPR1's +TIP activity

SPR1 is the smallest known +TIP; therefore, we wanted to identify the regions that confer microtubule plus-end localization. The conserved N- and C-terminal regions of SPR1 (Fig. 2A) have been proposed to contribute to microtubule binding because they contain a GGG and a PGGG motif, respectively, which are part of the microtubule-binding repeats of the MAP2/tau family of microtubule-associated proteins (11, 12).

Figure 2.

The N-terminal region of SPR1 is required for microtubule plus-end tracking. A, SPR1 amino acid sequence. The conserved N-terminal and C-terminal regions are shown in red and orange, respectively. The central variable region is shown in gray. Gray boxes indicate the conserved GGG and PGGG motifs in the N-terminal and C-terminal regions, respectively. B, schematic of the full-length SPR1-GFP fusion protein and mutant SPR1 proteins used in this study. The color code is the same as in A. C, Coomassie-stained gel of the truncated and chimeric versions of SPR1 protein used in this study. White dots indicate the expected protein bands. D, representative kymographs of SPR1(Δ1–18)-GFP, SPR1(Δ87–119)-GFP, N-GFP-C, and NC-GFP proteins. x axis scale bar = 5 μm, y axis scale bar = 100 s. E, representative kymographs of SPR1^GGG-AAA-GFP and SPR1^PGGG-AAAA-GFP proteins. A kymograph of WT SPR1-GFP is shown for comparison. x axis scale bar = 5 μm, y axis scale bar = 100 s. F, representative kymographs from single-molecule imaging experiments of SPR1-GFP, N-GFP-C, and NC-GFP proteins. Images at the top and bottom of the kymographs show the microtubule at the beginning and at the end of streaming-mode image acquisition. x axis scale bar = 5 μm, y axis scale bar = 10 s. G, table of the k_on, k_off, and apparent K_D of SPR1-GFP, N-GFP-C, and NC-GFP proteins. Values are mean ± S.D. (n = number of molecules).

To determine the relative contribution of the N- and C-terminal regions of SPR1, we created truncated versions of SPR1 that lacked either the N-terminal or the C-terminal regions (Fig. 2, B and C). We found that deleting the N-terminal 18 amino acids (SPR1Δ1–18) abolished +TIP activity in vitro (Fig. 2D), indicating that the N-terminal region of SPR1 is essential for recognizing the microtubule plus end. Deletion of the C-terminal 33 amino acids (SPR1Δ87–119) diminished the signal at growing microtubule plus ends but did not eliminate it (Fig. 2D), indicating that the C-terminal region enhances the +TIP activity of SPR1 but is not necessary for it. Consistent with these findings and with previous in vivo work (11), the chimeric N-GFP-C protein clearly localized to growing microtubule plus ends in vitro (Fig. 2D and Movie S2). To test whether the spacing of the N- and C-terminal regions is important for +TIP activity, we fused these regions in tandem to GFP (Fig. 2, B and C). The resulting NC-GFP fusion protein also localized to growing microtubule plus ends in vitro (Fig. 2D and Movie S3).

During these experiments, we noticed that the microtubule plus-end signals of N-GFP-C and NC-GFP were less than the SPR1-GFP signal. To determine whether the N-GFP-C and NC-GFP proteins have a lower affinity for the microtubule plus end compared with SPR1-GFP, we conducted in vitro dynamic microtubule assays under single-molecule imaging conditions followed by kymograph analysis (Fig. 2F). Photobleaching experiments revealed that the majority of SPR1-GFP particles were monomeric (97.7% monomers, 2.3% dimers, n = 175). Similarly, the majority of N-GFP-C and NC-GFP particles were also monomeric (92% monomers, 8% dimers, n = 88 for N-GFP-C; 91.2% monomers, 8.8% dimers, n = 158 for NC-GFP). We measured the binding rate and dwell time of individual particles to calculate the binding rate constant (k_on) and unbinding rate constant (k_off) of SPR1-GFP, N-GFP-C, and NC-GFP for the growing microtubule plus end. Our findings showed that N-GFP-C and NC-GFP bind more weakly to microtubule plus ends compared with SPR1-GFP (Fig. 2G).

Next we tested the significance of the N-terminal GGG and C-terminal PGGG motifs by mutating these amino acids to alanine within the context of full-length SPR1 (Fig. 2, B and C). Although mutation of PGGG to AAAA (SPR1^PGGG-AAAA) had no effect on the +TIP activity of SPR1, mutation of GGG to AAA (SPR1^GGG-AAA) abolished this activity (Fig. 2E). These findings underscore the essential role of the SPR1 N-terminal region and indicate that the GGG motif in the N-terminal region is critical for +TIP activity.

The N-GFP-C and NC-GFP markers do not significantly alter microtubule dynamics

The autonomous +TIP activity of N-GFP-C and NC-GFP makes them attractive microtubule plus-end markers. To determine whether they affect microtubule plus-end dynamics similar to the WT SPR1 protein, we tested the effect of N-GFP-C and NC-GFP on microtubule dynamics in vitro. We found that N-GFP-C did not significantly alter the microtubule growth rate (Fig. 3A), catastrophe frequency (Fig. 3C), or rescue frequency (Fig. 3D) and only modestly increased the shortening rate (Fig. 3B). NC-GFP did not significantly alter any of these microtubule dynamic instability parameters (Fig. 3, E–H, and Table S1). Furthermore, we compared the number of growth-to-shortening and shortening-to-growth transition events exhibited by individual microtubules at different concentrations of SPR1-GFP, N-GFP-C, and NC-GFP. We found that SPR1-GFP boosted microtubule transitions in a concentration-dependent manner (Fig. 3I), consistent with its enhancement of catastrophe and rescue frequencies (Fig. 1, I and J). In contrast, neither N-GFP-C nor NC-GFP significantly altered the number of microtubule transitions (Fig. 3, J and K). Taken together, our findings demonstrate that N-GFP-C and NC-GFP localize to growing microtubule plus ends without significantly perturbing their dynamics, unlike WT SPR1.

Figure 3.

N-GFP-C and NC-GFP do not significantly affect microtubule dynamics in vitro. A–D, dot plots of microtubule growth rate (A), shortening rate (B), catastrophe frequency (C), and rescue frequency (D) at the indicated N-GFP-C protein concentrations. *, p < 0.05; one-way ANOVA. E–H, dot plots of microtubule growth rate (E), shortening rate (F), catastrophe frequency (G), and rescue frequency (H) at the indicated NC-GFP protein concentrations. I–K, the percentage of microtubule transition events (i.e. switch from growth to shortening or from shortening to growth) at increasing concentrations of SPR1-GFP (I), N-GFP-C (J), and NC-GFP (K). Numbers are the percentages of each category of microtubule transitions during the observation period.

NC-GFP labels growing microtubule plus ends in plants and partially complements the spr1 mutant phenotype

To determine whether NC-GFP is functional in vivo, we expressed it under the constitutive A. thaliana UBIQUITIN10 (pUBQ10) promoter in spr1–6 mutant plants. pUBQ10:SPR1-GFP and pUBQ10:N-GFP-C constructs were included as controls. All three constructs were also expressed in WT plants expressing an RFP-TUB6 microtubule marker to determine their microtubule localization in plants.

We found that SPR1-GFP fully complements the hypocotyl cell file skewing phenotype of the spr1–6 mutant, whereas the NC-GFP and N-GFP-C constructs only partially complemented this phenotype (Fig. 4, A and B). Our data for SPR1-GFP and N-GFP-C are consistent with previous work that expressed these constructs under a cauliflower mosaic virus 35S promoter and reported complete and partial complementation of root skewing in spr1–6, respectively (11). Notably, although the N-GFP-C signal was barely above our detection threshold, NC-GFP localized to growing microtubule plus ends in vivo (Fig. 4C). These observations indicate that NC-GFP can be used as a microtubule plus-end marker in plants. We also found that NC-GFP preferentially binds to GTPγS microtubules, similar to SPR1-GFP (Fig. S1), indicating that NC-GFP also recognizes GDP/P_i-tubulin at growing microtubule plus ends.

Figure 4.

NC-GFP labels growing microtubule plus ends in plants. A, bright-field images of dark-grown hypocotyls of Col-0, spr1–6 mutant, and spr1–6 mutant expressing either SPR1-GFP, NC-GFP, or N-GFP-C. Scale bar = 100 μm. B, dot plot of the orientation of cells with respect to the hypocotyl growth axis. One-way ANOVA; ****, p < 0.0001; n = 26, 37, 67, 123, and 53 cells for Col-0, spr1–6, SPR1-GFP, NC-GFP, and N-GFP-C plants, respectively. C, representative kymographs of SPR1-GFP and NC-GFP expressed in WT Arabidopsis plants expressing an RFP-TUB6 microtubule marker. x axis scale bar = 5 μm, y axis scale bar = 100 s.

SPR1-GFP and NC-GFP label growing microtubule plus ends in animal cells

Because SPR1-GFP, N-GFP-C, and NC-GFP labeled the plus ends of microtubules assembled from porcine tubulin in vitro, we reasoned that they might label microtubule plus ends in animal cells. To test the suitability of these proteins as microtubule plus-end markers in animal cells, we introduced them in HEK293 cells and imaged the transfected cells using total internal reflection fluorescence microscopy. GFP-labeled human tubulin and EB3 were included as controls in these experiments. As expected, GFP-tubulin labeled the entire microtubule lattice, whereas EB3-GFP localized to growing microtubule plus ends (Fig. 5, A and B). Interestingly, both SPR1-GFP and NC-GFP clearly labeled growing microtubule plus ends and the microtubule shaft to a lesser extent (Fig. 5, C and D, and Movies S4 and S5). To assess whether SPR1-GFP and NC-GFP affect microtubule dynamics in HEK293 cells, we quantified microtubule growth rates in cells expressing these proteins (Fig. 5E). Cells expressing SPR1-GFP had slightly higher microtubule growth rates (0.23 ± 0.07 μm/s, n = 105 microtubules) compared with cells expressing GFP-tubulin (0.17 ± 0.06 μm/s, n = 114 microtubules) or EB3-GFP (0.17 ± 0.07 μm/s, n = 105 microtubules). Expression of NC-GFP in HEK293 cells had a less pronounced effect on the microtubule growth rate (0.19 ± 0.04 μm/s, n = 99 microtubules) compared with SPR1-GFP. Importantly, SPR1-GFP and NC-GFP did not visibly impair the viability and morphology of HEK293 cells (Fig. S2), indicating that these proteins are not toxic to animal cells. Although N-GFP-C showed an extensive cytoplasmic signal in HEK293 cells, we were unable to detect unambiguous microtubule plus-end labeling.

Figure 5.

NC-GFP labels growing microtubule plus ends in animal cells. A and B, representative images of HEK293 cells expressing GFP-tubulin (A) or EB3-GFP (B). Scale bar = 5 μm. C and D, representative images of HEK293 cells expressing SPR1-GFP (C) or NC-GFP (D). Scale bar = 5 μm. Numbers in the images indicate time in seconds. Arrowheads point to examples of growing microtubule ends. E, dot plot of the microtubule growth rates in HEK293 cells expressing either GFP-tubulin, EB3-GFP, SPR1-GFP, or NC-GFP. ***, p < 0.001.

Discussion

We demonstrate that the 12-kDa plant-specific Arabidopsis SPR1 protein is an autonomous microtubule plus end–tracking protein. Using truncations and mutations in SPR1, we reveal that the conserved N-terminal region and its GGG motif are critical for plus end–tracking activity, whereas the conserved C-terminal region and its PGGG motif are dispensable. Our GTPγS microtubule binding experiments showed that SPR1 targets growing microtubule ends by recognizing the GDP/P_i state of tubulin, similar to EB proteins. However, SPR1 does not share any sequence similarity with EBs or other known +TIPs. Therefore, our findings establish a new category of intrinsic +TIPs unique to plants.

Because both SPR1 and EB1 proteins recognize the GDP/P_i tubulin structure, they might compete for binding sites at growing microtubule plus ends. Competitive binding of SPR1 and Arabidopsis EB1b proteins to the microtubule lattice has been observed in microtubule cosedimentation assays (16). In addition, mutant analysis suggests that SPR1 and EB1b have overlapping and sometimes antagonistic functions (16), consistent with SPR1 and EB1 competing for binding sites. SPR1 has been found to directly interact with EB1b (16); however, our data show that this interaction is not required for the plus end–tracking activity of SPR1. It remains to be determined whether binding of SPR1 and EB1b modulates their microtubule binding affinity and/or their ability to recruit other proteins to growing microtubule plus ends.

Previous work showed that the N- and C-terminal regions of SPR1 are individually not sufficient to bind to microtubules in vivo and do not rescue spr1 mutants (11). However, N-GFP-C was partially functional and localized to microtubule plus ends in vivo (11). We found that both N-GFP-C and NC-GFP localize to growing microtubule plus ends in vitro, demonstrating that the N- and C-terminal regions, either separated or in tandem, are sufficient to target growing microtubule plus ends. These results also show that the spacing between the N- and C-terminal regions is not a critical determinant of SPR1's plus end–tracking activity, consistent with the high level of variability in amino acid sequence and length of the central region of SPR1 (11, 12).

Despite localizing to growing microtubule plus ends, both N-GFP-C and NC-GFP only partially complement the spr1–6 mutant. The microtubule plus end binding affinity is about 2-fold lower for N-GFP-C and NC-GFP compared with full-length SPR1. In addition, although full-length SPR1 promotes microtubule dynamics, N-GFP-C and NC-GFP do not significantly affect microtubule dynamics. Together, these differences might explain why N-GFP-C and NC-GFP do not fully recapitulate SPR1 function in vivo.

The N-terminal GGG and C-terminal PGGG motifs of SPR1 have been proposed to mediate microtubule binding based on their similarity to the microtubule binding repeats of the MAP2/tau family of microtubule-associated proteins (16). However, MAP2/tau are not microtubule plus end–tracking proteins, and these repeat motifs contribute to their binding to the microtubule side walls, which primarily consist of GDP-state tubulin. We found that the N-terminal GGG motif is necessary for the +TIP activity of SPR1 and that the C-terminal PGGG motif enhances this activity. Elucidating how these motifs contribute to the binding of SPR1 to GDP/P_i-tubulin at growing microtubule plus ends will likely require structural information.

Live imaging of microtubule plus ends in animal cells is commonly conducted by exogenously expressing fluorescent protein–tagged EB1 or EB3 (20–24). However, because EB proteins form the core of microtubule plus-end complexes (25), their ectopic expression can potentially perturb endogenous microtubule plus-end complexes. Our finding that SPR1-GFP and NC-GFP label growing microtubule plus ends in HEK 293 cells demonstrates their utility as alternate +TIP markers in animal cells. Expression of SPR1-GFP in HEK293 cells slightly increased microtubule growth rates compared with control cells. In contrast, microtubules labeled by NC-GFP in HEK293 cells had only a minor increase in growth rate compared with control cells. Because NC-GFP minimally alters microtubule growth and is about half the size of SPR1 (∼ 6 kDa), we propose NC-GFP as a versatile microtubule plus-end marker.

Experimental procedures

Constructs

For recombinant protein expression, a commercially synthesized SPR1-GFP construct was cloned into the pTEV plasmid between NcoI and XhoI restriction sites to obtain a C-terminal His₆ tag. The N-GFP-C construct was assembled using recombinant PCR. For this, EGFP was PCR-amplified with primers extending into the NSPR1 (5′-CAAAGCTCATTGGATTATCTCTTTGGTGGTGACGCTCCTATGGTGAGCAAGGGCGAG-3′) and CSPR1 regions (5′-TCCTTCAGCTCTGGCATAGTTGTTCTTGTACAGCTCGTCCATG-3′). Then CSPR1 was PCR-amplified with a forward primer containing 21 bp of the EGFP sequence (5′-GGCATGGACGAGCTGTACAAGAACAACTATGCCAGAGCTG-3′) and a reverse primer containing the XhoI restriction site (5′-TATCTCGAGCTTGCCACCAGTGAAGAGATAATCCAAGGATGATCCTCCTC-3′). These PCR fragments were then mixed at a 1:1 ratio and used as a template for a third PCR reaction with a forward primer consisting of the SPR1 N-terminal sequence and NcoI restriction site (5′-TATCCATGGGTCGTGGAAACAGCTGTGGTGGAGGT CAAAGCTCATTGGATTATC-3′) and a reverse primer containing the XhoI restriction site (5′-TATCTCGAGCTTGCCACCAGTGAAGAGATAATCCAAGGATGATCCTCCTC-3′). The resulting PCR product was inserted into the NcoI and XhoI sites of the pTEV plasmid. The NC-GFP construct was obtained by fusing the N- and C-terminal regions of SPR1 in a two-step PCR. The first PCR used primers 5′-CAAAGCTCATTGGATTATCTCTTTGGTGGTGACGCTCCTAACAACTATGCCAGAGCTGAAG-3′ and 5′-TATAGTCGACCTTGCCACCAGTGAAGAG-3′, and the second PCR used primers 5′-TATCCATGGGTCGTGGAAACAGCTGTGGTGGAGGTCAAAGCTCATTGGATTATC-3′ and 5′-TATA GTCGACCTTGCCACCAGTGAAGAG-3′, with the first PCR product serving as a template. The resulting PCR product was inserted upstream of EGFP in the pTEV plasmid at NcoI and SalI sites. GGG-to-AAA and PGGG-to-AAAA mutant versions of SPR1 were generated using site-directed mutagenesis. The SPR1(Δ1–18)-GFP and SPR1(Δ87–119)-GFP truncation constructs were generated by replacing full-length SPR1 with PCR-amplified deletion fragments in the SPR1-GFP-His₆ construct.

To express SPR1 under its native regulatory elements, we cloned a SPR1 genomic fragment that included 896 bp upstream of the SPR1 start codon using primers SPR1-attB1 (GGGGACAAGTTTGTACAAAAAAGCAGGCTGCGCCATTACACATGAATGGCCC) and SPR1-attB5r (GGGGACAACTTTTGTATACAAAGTTGTCTTGCCACCAGTGAAGAGATAATC). This construct also included 700 bp downstream of the SPR1 stop codon that was amplified from genomic DNA using primers SPR1-attB4r (GGGGACAACTTTTCTATACAAAGTTGCTTAAAATAATTGCAAAGACCTTTATCTATCC) and SPR1-attB3r (GGGGACAACTTTATTATACAAAGTTGT AACACATTTGGAGCTGCATATGC). The EGFP coding sequence was inserted before the SPR1 stop codon. To express SPR1 under the UBIQUITIN10 (pUBQ10) promoter, SPR1-GFP-His₆, N-GFP-C-His₆, and NC-GFP-His₆ were PCR-amplified using the respective bacterial expression vectors described above as templates and primers SPR1-attB5 (GGGGACAACTTTGTATACAAAAGTTGCCGTCGACATGGGTCGTGGAAACAGCTG) and His₆-attB4 (GGGGACAACTTTGTATAGAAAAGTTGGGTGGCGGCCGCTCAGTGGTGGTGGTGGTGGTG). All plant constructs were introduced in a pGAG binary vector containing a gentamycin-resistant cassette using gateway recombination. All constructs created using PCR were verified by sequencing.

Plant growth conditions

Plant constructs were transformed into the spr1–6 mutant and an RFP-TUB6 microtubule marker line using an Agrobacterium-mediated floral dip method. To check the localization of SPR1-GFP in the eb1a;eb1b double mutant, pSPR1:gSPR1-GFP was transformed into an eb1a;eb1b line expressing the RFP-TUB6 marker. T1 seeds were surface-sterilized in 25% (v/v) bleach for 8 min, washed four to five times with sterile water, and suspended in 0.1% (w/v) agarose solution. Seeds were then plated on ½ MS plates containing 100 mg/liter gentamycin and stratified at 4 °C in dark for 3 days. To facilitate selection, stratified seeds were exposed to light for 6 h and then grown in the dark for 48 h at 23 °C, after which they were grown under long-day conditions (16 h light and 8 h dark) for another 48 h at 23 °C. For complementation experiments, at least 8 independent T1 lines were used to analyze the hypocotyl-skewing phenotype. Col-0 and spr1–6 seedlings were used as controls and grown on ½ MS without gentamycin.

Protein purification

Escherichia coli Rosetta (DE3) cells were used for protein expression. Cells were grown to 0.6–0.8 A₆₀₀ at 30 °C and then induced using 0.15 mm isopropyl 1-thio-β-d-galactopyranoside at 16 °C for 16–20 h. His₆-tagged proteins were isolated using nickel-nitrilotriacetic acid–agarose beads using lysis buffer (50 mm Tris (pH 8.0), 20 mm imidazole, 500 mm NaCl, 10% glycerol, 1% Tween 20, and protease inhibitors). Lysis buffer without Tween 20 and protease inhibitors was used for washing off unbound proteins, and wash buffer with 150 mm imidazole was used to elute protein from the nickel-nitrilotriacetic acid–agarose column. Purified proteins were desalted and exchanged into BRB80 buffer (80 mm piperazine-1,4-bis(2-ethanesulfonic acid), 1 mm MgCl₂, and 1 mm EGTA (pH 6.8)) supplemented with 10 mm DTT and 50 mm NaCl using PD-10 columns (GE Healthcare). Protein concentration was measured using the Bradford method. Proteins were aliquoted, snap-frozen in liquid nitrogen, and stored at −80 °C. All purified proteins used in this study had a His₆ tag attached at the C terminus of the protein.

Expression in animal cells

SPR1-GFP-His₆, N-GFP-C-His₆, and NC-GFP-His₆ were PCR-amplified using primers SPR1-SalI-EcoRI (TATAGTCGACGAATTCATGGGTCGTGGAAACAGCTG) and His₆-SmaI (TATACCCGGGTCAGTGGTGGTGGTGGTGGTG) and the respective bacterial expression constructs described above as templates. The resulting PCR products were cloned into the EcoRI-SmaI restriction sites in the lentiviral vector pLVX-puromycin. The GFP-tubulin and EB3-GFP constructs were obtained from Weiping Han (Addgene plasmid 64060) (26) and the Washington University viral core facility, respectively. HEK293 cell transfection was conducted using 500 ng of DNA, and 2000 ng of polyethyleneimine HCI MAX (Polysciences, Inc.) was added to 50 μl of 300 mm NaCl. This solution was mixed, incubated for 10 min at room temperature, and then added to the plated cells in culture medium containing 10% fetal bovine serum in DMEM plus 1% penicillin-streptomycin. Cells were imaged 24–48 h after transfection. Expression in HEK293 cells was repeated four times. Quantification of microtubule growth rate was repeated twice (n = 20 cells for SPR1-GFP, n = 32 cells for NC-GFP, n = at least 10 cells each for EB3-GFP and GFP-tubulin).

Total internal reflection fluorescence microscopy

For the in vitro reconstitution assays, flow chambers were prepared by attaching silanized coverslips to a glass slide using two layers of double-sided adhesive tape. GMPCPP-stabilized (Jena Biosciences) microtubule seeds were prepared by polymerizing 50 μm porcine tubulin containing biotin-labeled (1:14.5) and rhodamine-labeled (1:14.5) porcine tubulins (Cytoskeleton, Inc.) in the presence of 1 mm GMPCPP at 37 °C for 30 min. The polymerized microtubules were then collected by centrifugation at 25,000 × g for 20 min, resuspended in 50 μl of warm BRB80 containing 1 μm GMPCPP, and fragmented by passing through a 100-μl Hamilton syringe four to five times. Flow chambers were first coated with 20% anti-biotin antibody (Sigma-Aldrich), followed by blocking with 5% Pluronic for 5 min each. Then 300 nm GMPCPP-stabilized seeds were added and allowed to bind to the anti-biotin antibody for 5 min. Microtubule polymerization was initiated by adding a mixture of 20 μm 1:25 rhodamine-labeled porcine tubulin (i.e. 4% of the tubulin is rhodamine-labeled), 1% methylcellulose (4000 cP, Sigma-Aldrich), 50 mm DTT, 2 mm GTP, an oxygen-scavenging system (250 mg/ml glucose oxidase, 35 mg/ml catalase, and 4.5 mg/ml glucose), and different concentrations of SPR1 or mutant proteins, as indicated in the text. 5-mW 488-nm and 5-mW 561-nm diode-pumped, solid-state lasers were used to excite GFP and rhodamine, respectively. Images were collected by a ×100 (numerical aperture 1.45) objective with a 2x tube lens and a back-illuminated electron-multiplying charge-coupled device camera (ImageEM, Hamamatsu) at 2.5-s intervals and for a total of about 5 min.

For single-molecule imaging experiments, we used 100 nm, 150 nm, and 250 nm of SPR1-GFP, NC-GFP, and N-GFP-C proteins, respectively. Images were captured at 10 frames/s in streaming mode. A snapshot of microtubules was taken just before and immediately after the streaming mode acquisition to document microtubule growth. For photobleaching experiments, we used 1 nm of SPR1-GFP, NC-GFP, and N-GFP-C proteins. To promote photobleaching, we exited GFP using 20 mW of laser power. Images were acquired at an exposure time of 100 ms and 200-ms intervals between frames.

For GTPγS microtubule binding experiments, we followed the method described in Maurer et al. (4). Flow chambers with GMPCPP microtubule seeds bound to anti-biotin antibody on coverslips were prepared as described above. Microtubule polymerization was initiated by adding a polymerization mixture as described above with the following modifications: 25 μm porcine tubulin (4% of which is rhodamine-labeled), 10 mm 2-mercaptoethanol, and 2 mm GTPγS (Sigma) with 500 nm SPR1-GFP or NC-GFP.

Live imaging of Arabidopsis seedlings was done using 4-day-old T1 seedlings. At least three independent T1 lines were used for these experiments. Live imaging of HEK293 cells was conducted 2–3 days after transfection. Specimens were illuminated with 3-mW 488-nm and 3-mW 561-nm lasers to excite GFP and RFP, respectively.

Data analysis

Microtubule dynamics and single-molecule kinetics were quantified from kymographs of individual microtubules generated using the Fiji ImageJ package (27). Microtubule growth and shortening rates were calculated for individual growth and shortening phases. Rescue frequency was calculated by dividing the sum of the number of transitions from shrinkage to growth by the time spent shrinking. Catastrophe frequency was calculated by dividing the sum of the number of transitions from growth to shrinkage by the time spent growing. To determine the microtubule transition frequency, a switch from growth to shortening or vice versa was considered as a transition event. When a microtubule depolymerized all the way back to the GMPCPP seed followed by a pause of 5 or more frames, subsequent microtubule growth was considered as an independent microtubule growth event and not a transition event of the previous microtubule. SPR1-GFP signal intensity plots were created using the plot profile tool in Fiji. The angles of hypocotyl cell files were measured with respect to the hypocotyl growth axis using the angle measure tool in Fiji. Dot plots and statistical analyses were conducted using GraphPad Prism. All data points were included in the statistical analyses.

Author contributions

R. B., L. F., V. C., and R. D. conceptualization; R. B., L. F., R. E., and R. D. formal analysis; R. B., L. F., E. E., and R. E. investigation; R. B., L. F., E. E., and R. E. visualization; R. B., L. F., and E. E. methodology; R. B. writing-original draft; R. B., L. F., E. E., R. E., V. C., and R. D. writing-review and editing; V. C. and R. D. supervision; V. C. and R. D. funding acquisition; V. C. and R. D. project administration.