She1 affects dynein through direct interactions with the microtubule and the dynein microtubule-binding domain

Cytoplasmic dynein is an enormous minus end-directed microtubule motor. Rather than existing as bare tracks, microtubules are bound by numerous microtubule-associated proteins (MAPs) that have the capacity to affect various cellular functions, including motor-mediated transport. One such MAP is She1, a dynein effector that polarizes dynein-mediated spindle movements in budding yeast. Here, we characterize the molecular basis by which She1 affects dynein, providing the first such insight into which a MAP can modulate motor motility. We find that She1 affects the ATPase rate, microtubule-binding affinity, and stepping behavior of dynein, and that microtubule binding by She1 is required for its effects on dynein motility. Moreover, we find that She1 directly contacts the microtubule-binding domain of dynein, and that their interaction is sensitive to the nucleotide-bound state of the motor. Our data support a model in which simultaneous interactions between the microtubule and dynein enables She1 to directly affect dynein motility.

between individual steps) for motor domain labeled (via Qdot 525 ) GST-dynein 331 in the absence or presence of She1, and with either 1 µM or 1 mM ATP, as indicated. The histograms were fit to a convolution of two exponential functions [tk 2 exp(-kt)] with equal decay constants, k, which reflects the number of steps taken per second 1,2 (k ± standard error of the fit is shown).
Supplementary Figure 2. Cumulative probability functions used for determination of mean run length and dwell time values of dynein with and without She1 on control and subtilisin-treated microtubules. Raw data (circles) and fits (dashed lines) are shown for run length (top) and dwell time (bottom) in the absence (green) and presence of She1 (red) on control (left) or subtilisin-treated microtubules (right; n ≥ 199 individual motors for each condition). Data were fit as previously described 2 .
Supplementary Figure 3. Characterization of recombinant fragments used in the recruitment assays, and two-hybrid data. (a) Mean fluorescence intensity values (along with standard deviations) of microtubule-bound monomeric GFPdynein 331 in the absence (magenta) or presence of ATP and vanadate (green). (b) Recombinant protein fragments used in the recruitment assays. With the exception of the dynein motor domain fragments, which were purified from yeast, all proteins were purified from E. coli (see Methods). (c) Cartoon representation and reconstructed structural model of the SRS CC -dynein MTBD fusion. Image was generated from a yeast model of the DYN1 MTBD (threaded into 3ERR 3 ) and 1SRY 4 . (d) Mean fluorescence intensity values (along with standard deviations) of microtubule-bound GFP-SRS CC -dynein MTBD (red) and GFP-dynein CC+MTBD (green; n ≥ 19 microtubules, and ≥ 151 µm of MT length for each condition) along with fits and resulting dissociation constants (K D ). (e) Two-hybrid assay demonstrating an interaction between dynein MTBD and She1 (see Methods). (f) Schematic representation of the experimental setup used for panel g. (g) Representative images depicting the inability of microtubule-bound She1 to recruit GFP-dynein 331 ∆ MTBD to microtubules. Images were acquired prior and subsequent to washing the chamber with motility buffer (see Methods; scale bar, 1 µm).
Supplementary Figure 4. Comparison of various dynein MTBDs. (a) Cartoon and homology models of the yeast dynein MTBD bound to α and β-tubulin in the high (grey) and low (tan) microtubule affinity states. The models were generated using one-to-one threading of the yeast DYN1 sequence into 3J1T 5 (high affinity) and 3J1U (low affinity). CC1 and H1, which exhibit the largest differences between the two structures, are depicted as follows: CC1, red and pink, for high and low affinity states; H1, blue and cyan, for high and low affinity states, respectively. (b, left) Crystal structure of human dynein-2 (4RH7) docked onto microtubules (from 3J1T). (right) Homology model of the yeast MTBD (colored)along with a short region of the CC (grey)bound to α and β-tubulin in the high microtubule affinity state. The residues are colored to reflect the degree of conservation between yeast and mouse primary sequence (see legend). (c) Mean fluorescence intensity values (along with standard deviations) of microtubule-bound GFP-dynein 331 (green) and GFP-dynein 331 mMTBD (red; n ≥ 15 microtubules, and ≥ 68 µm of MT length for each condition) along with fits and resulting dissociation constants (K D ). Note the differences in apparent B max values (4645 ± 763 A.U. for wild-type, and 2452 ± 517 A.U. for mMTBD; ± SE of fit) are likely a consequence of microtubule unbinding during the chamber washes (see Methods), and likely differences in microtubule dissociation rates between the two motor domains.
Supplementary  Figure 6. Dynein relocalization to astral microtubules upon She1 overexpression requires the dynein MTBD, but not Pac1. (a) Cartoon representation of the two possible models to account for dynein relocalization upon She1 overexpression. The model on the left depicts a mechanism whereby the entire plus end targeting complex (composed of Dyn1, Pac1, Bik1 and Bim1; note that dynactin is not an obligate component of this complex 6 ) is required for the relocalization. Given the dispensable nature for the MTBD in plus end targeting 7 , this would indicate an MTBD-independent mechanism. The model on the right depicts a mechanism whereby dynein microtubule binding activity via the MTBD is required. (b and c) Representative images of GAL1p:SHE1 cells expressing mTurquoise2-Tub1 (b), or mRuby2-Tub1 (c), and either Dyn1 ∆ MTBD -3YFP (b) or Dyn1-3GFP (c), the latter of which is deleted for PAC1. Cells were grown to mid-log phase in SD media supplemented with raffinose (uninduced; "galactose") or galactose plus raffinose (induced for 3.5 hours; "+ galactose") and then mounted on agarose pads for confocal fluorescence microscopy. Foci were identified in two-color movies and scored accordingly (see Methods; blue arrows, plus end foci; blue arrowheads, cortical foci; red arrowhead, dynamic cytoplasmic foci not associated with microtubules or spindle poles). Note the accumulation of Dyn1 near the spindle poles in pac1∆ GAL1p:SHE1 cells grown in galactose-containing media (the same was observed in GAL1p:SHE1 PAC1 cells; not shown). Movies reveal these spots exhibit dynamic movements in a manner that is consistent with them localizing to short astral microtubules, and not the spindle poles themselves. Note that our data support the MTBDdependent model, depicted in panel a, right.
Supplementary Figure 7. In vivo assessment of dynein mMTBD mutant function.
(a and b) The percentage of cells with the indicated spindle orientation phenotype (green, normal; blue, aligned along mother-bud axis, but not through the neck; red, improperly aligned) is plotted for the indicated yeast strains (WT, wild-type). Anaphase spindles were visualized using mRuby2-Tub1 (α-tubulin). Strains were imaged after growth to mid-log phase in SD media supplemented with either (a) 2% galactose, or (b) 2% glucose, the former of which induces overexpression of She1 in GAL1p:SHE1 cells (scale bars, 2 µm; error bars, standard error of proportion; n ≥ 17 and n ≥ 21 anaphase spindles for each strain in panels a and b, respectively). Note the higher prevalence of misoriented spindles in She1-overexpressing cells (GAL1p:SHE1) than in cells lacking dynein (dyn1∆; *, p ≤ 0.015). This suggests that She1-overexpression disrupts other non-dynein-mediated spindle orientation processes (e.g., Kar9 pathway 8 ). (c) Representative fluorescence images of kar9∆, hydroxyurea (HU)-arrested GFP-Tub1 (α-tubulin) expressing cells with the indicated SHE1 and DYN1 alleles (scale bars, 1 µm), along with kymographs depicting spindle movements over time (horizontal scale bars, 1 µm; vertical scale bars, 1 min). Dashed lines indicate the position of the bud neck in each example. Note the frequency with which the spindle traverses the bud neck in wild-type, but not mutant cells (green arrows; see Fig. 7g for quantitation). P-values were calculated using a two-tailed unpaired t test.

YEF473
This study