Lethal kinesin mutations reveal amino acids important for ATPase activation and structural coupling

Abstract

To study the relationship between conventional kinesin's structure and function, we identified 13 lethal mutations in the Drosophila kinesin heavy chain motor domain and tested a subset for effects on mechanochemistry. S246F is a moderate mutation that occurs in loop 11 between the ATP- and microtubule-binding sites. While ATP and microtubule binding appear normal, there is a 3-fold decrease in the rate of ATP turnover. This is consistent with the hypothesis that loop 11 provides a structural link that is important for the activation of ATP turnover by microtubule binding. T291M is a severe mutation that occurs in alpha-helix 5 near the center of the microtubule-binding surface. It impairs the microtubule-kinesin interaction and directly effects the ATP-binding pocket, allowing an increase in ATP turnover in the absence of microtubules. The T291M mutation may mimic the structure of a microtubule-bound, partially activated state. E164K is a moderate mutation that occurs at the beta-sheet 5a/loop 8b junction, remote from the ATP pocket. Surprisingly, it causes both tighter ATP-binding and a 2-fold decrease in ATP turnover. We propose that E164 forms an ionic bridge with alpha-helix 5 and speculate that it helps coordinate the alternating site catalysis of dimerized kinesin heavy chain motor domains.

FIG. 1. A model for dimeric kinesin ATPase mechanism

The ATPase cycle begins as one motor domain binds to the microtubule, leading to rapid release of ADP (step 1). ATP binding (step 2) stimulates binding of the second motor domain to the microtubule and release of its ADP (step 3). Subsequently, ATP hydrolysis (step 4) occurs forming an intermediate in which both motor domains are bound to the microtubule but the nucleotide state of each is different. Step 5 is the slowest in the pathway, encompassing both detachment of the motor domain from the microtubule and phosphate release. Step 6 positions the detached motor domain toward the microtubule plus end. Rebinding of the detached head to the next available microtubule-binding site is rapid (step 2 and 3) and advances the center of mass of the dimer by 8 nm, thereby completing one step cycle. The cycle repeats for the second head, resulting in 2 steps driven by the hydrolysis of 2 ATPs. 0, no nucleotide at active site. k₁ = 300 s⁻¹; k₂ = 2 µm⁻¹ s⁻¹; k₋₂ = 70 s⁻¹; k₃ = 200 s⁻¹; k₄ = 100 s⁻¹; k₅ = 50 s⁻¹; k_cat = 20 s⁻¹ per kinesin active site, 40 s⁻¹ per dimer.

FIG. 2. Stability of mutant KHC proteins in vivo

A Western blot of cytosol preparations from adult flies of the indicated genotypes was probed with an antibody that binds to the motor domain of Drosophila KHC (36). Genotypes: +, wild-type chromosomes; Df, a deletion that removes Khc from chromosome 2; m-Khc, a fusion gene on chromosome 3 that expresses a Myc-tagged KHC at high levels; 4, 17, 23, 27, and 37, the respective mutant Khc alleles. The upper band corresponds to transgenic Myc·KHC whereas the lower band corresponds to endogenous wild-type or mutant KHC. The absence of detectable protein suggests that the nonsense allele Khc²⁷ is a protein null. The 4 missense alleles that were selected for mechanochemical characterization (Khc 4, 17, 23, 37) all produce stable mutant KHCs.

FIG. 3. Determination of the severities of KHC mutations in vivo

A graphic representation of the time course of lethality caused by mutant Khc alleles. The x axis shows successive stages in the development of Drosophila. The y axis shows the percent of live animals remaining at the beginning of each stage. Each line is labeled with an abbreviated genotype: numbers represent the different Khc alleles and Df represents a deletion of the Khc locus. The number of second instar larvae tested for each genotype were: Khc²⁷/Df = 25, Khc⁴/Df = 122, Khc²³/Df = 86, Khc¹⁷/Df = 86, Khc³⁷/Df = 96, and Khc²³/Khc⁴ = 209. Note that the effect of Khc⁴ is similar to both the null mutation Khc²⁷ (compare first and second lines) and the deletion (compare third and fourth lines). Therefore Khc⁴ causes a near complete loss of function and is classified as “severe” in Table I. Khc³⁷ allows the survival of some animals to adulthood and thus is classified as “mild.” Khc²³ and Khc¹⁷ are less severe than a null but do not allow survival of any adults and thus are classified as “moderate.”

FIG. 4. Amino acid conservation and location of motor domain mutations selected for mechanochemical characterization

A, the amino acid changes caused by the four Drosophila motor domain alleles selected for mechanochemical characterization are placed above partial sequence alignments of Drosophila (Dm), human (Hs), and rat (Rn) KHCs, which are all members of the KHC subfamily, and Drosophila NCD, which is a member of the COOH-terminal motor subfamily (, –61). The corresponding amino acid numbers for the different motor proteins are included to facilitate location of the affected amino acids in published motor domain crystal structures. B, location of the affected amino acids in a crystal structure model of the rat brain KHC motor domain (27). This model was rotated to display the locations of four wild-type amino acids that are changed in the motor domain alleles and the positions of their side chains. The helices and loops thought to be most directly involved in microtubule binding are on the left, and the nucleotide-binding pocket is at the rear near the top. The ADP in the active site is shaded yellow. The side chains of the wild-type residues altered in the motor domain mutations are green except for Khc¹⁷. The location of the residue affected by the Khc¹⁷ mutation cannot be shown accurately because loop 11 has not been resolved in KHC crystal structures. In the linear sequence of loop 11, the affected amino acid is 3 residues from the switch II region and 17 residues from the beginning of α4.

FIG. 5. The K401-BIO proteins used for motility assays

Various KHC motor domains fused to a portion of an E. coli protein that is post-translationally biotinylated were purified by sedimentation with and release from microtubules using Mg/AMP-PNP and MgATP, respectively. Equal volumes of the ATP release supernatant of K401-BIO wild-type (WT), 4, 17, 23, and 37 were run on a 7.5% SDS-polyacrylamide gel and blotted to nitrocellulose. The blot shown was probed with a monoclonal anti-Drosophila KHC antibody (FLYK2) (36). The variability in the amount of KHC in each lane may be due to the effects of the mutations on microtubule-motor binding. The rates of microtubule gliding generated by these K401-BIO protein preparations are reported in Table II.

FIG. 6. Expression and purification of K401 proteins for kinetic analysis

The left panel shows a Coomassie Blue R-250 stained 8% acrylamide, 2 m urea gel of various fractions collected during the purification of the K401 proteins that were used for ATPase assays. The lanes contain from left to right: molecular weight standards, preinduced cell lysate, induced cell lysate, low speed supernatant of induced lysate (LS-S), low speed pellet (LS-P), high speed supernatant (HS-S), S-Sepharose column eluant, and DEAE-Sephacel column eluant. The right panel shows a Coomassie-stained gel of the concentrated, purified protein preparations that were used in ATPase assays. The lanes contain from left to right: wild-type K401 and the mutant proteins K401-4, K401-17, K401-23, and K401-37. The lanes were loaded with equal volumes from equivalent preparations.

FIG. 7. Steady-state ATPase activity of wild-type and mutant K401 motors

A, the rate of [α-³²P]ATP hydrolysis was determined for MT·K401 complexes (0.5 µm K401) in the presence of saturating MgATP (2 mm) as a function of taxol-stabilized microtubule concentration (0.25–20 µm tubulin). The data were fit to the quadratic equation as described under “Experimental Procedures.” B, the rate of hydrolysis was determined in the presence of saturating microtubules (20 µm tubulin) as a function of MgATP concentration (0.01–2 mm). Each line is the fit of the data to a hyperbola with the kinetic parameters for wild-type and mutant kinesins reported in Table III. ●, K401-WT; ○, K401-17; ▲, K401-4; △, K401-37; □, K401-23.

FIG. 8. Two ionic interactions that may influence dimeric kinesin structure and function

Shown is part of the crystal structure model of dimeric rat brain KHC (24). One subunit is on the left (A) and one is on the right (B). They are joined by their α7 helices, which project vertically. A potential ionic contact described by Kozielski et al. (24) is shown between the Lys¹⁶⁰ side chain (orange) in L8b of subunit A and the Glu²¹⁹ side chain (yellow) in L10 of subunit B. The side chains of Glu¹⁶⁴ (green) and Arg²⁹² (magenta) are highlighted to point out their proximity and their probable ionic interaction which we propose as an important link between L8 and α5. Thr²⁹¹ (green), the amino acid mutated in Khc⁴, is also shown to illustrate that changes in the Glu¹⁶⁴-Arg²⁹² link could shift the positions of a side chain known to influence both nucleotide binding and microtubule binding.