Kinetic characterization of the sole nonmuscle myosin-2 from the model organism Drosophila melanogaster

Abstract

Nonmuscle myosin-2 is the primary enzyme complex powering contractility of the F-actin cytoskeleton in the model organism Drosophila. Despite myosin’s essential function in fly development and homeostasis, its kinetic features remain elusive. The purpose of this in vitro study is a detailed steady-state and presteady-state kinetic characterization of the Drosophila nonmuscle myosin-2 motor domain. Kinetic features are a slow steady-state ATPase activity, high affinities for F-actin and ADP, and a low duty ratio. Comparative analysis of the overall enzymatic signatures across the nonmuscle myosin-2 complement from model organisms indicates that the Drosophila protein resembles nonmuscle myosin-2s from metazoa rather than protozoa, though modulatory aspects of myosin motor function are distinct. Drosophila nonmuscle myosin-2 is uniquely insensitive toward blebbistatin, a commonly used myosin-2 inhibitor. An in silico modeling approach together with kinetic studies indicate that the nonconsensus amino acid Met466 in the Drosophila nonmuscle myosin-2 active-site loop switch-2 acts as blebbistatin desensitizer. Introduction of the M466I mutation sensitized the protein for blebbistatin, resulting in a half-maximal inhibitory concentration of 36.3 ± 4.1 µM. Together, these data show that Drosophila nonmuscle myosin-2 is a bona fide molecular motor and establish an important link between switch-2 and blebbistatin sensitivity.—Heissler, S. M., Chinthalapudi, K., Sellers, J. R. Kinetic characterization of the sole nonmuscle myosin-2 from the model organism Drosophila melanogaster.

The sole nonmuscle myosin-2 in the fruit fly Drosophila is encoded by the zipper (zip) gene, named after an extensive defect in cytokinesis that leads to failure in dorsal closure and embryonic lethality (1). Structurally, the nonmuscle myosin-2 heavy chain follows a prototypic architecture, including an N-terminal catalytic motor domain, a light chain binding neck region, followed by a coiled-coil-forming domain that terminates in a short nonhelical tailpiece (2). The quaternary structure of the myosin holoenzyme is hexameric, composed of 2 heavy chains, which associate with a set of essential [cytoplasmic myosin light chain (mlc-c)] and regulatory [spaghetti squash (sqh)] light chains. Under physiologic conditions, multiple myosins oligomerize into defined structures of bipolar minifilaments, which interact with F-actin in an ATP-dependent manner and power cytoplasmic contractility (2).

Nonmuscle myosin-2 plays pivotal roles in cell shape changes and tissue dynamics in morphogenic processes such as organogenesis, myogenesis, and the establishment of left-right asymmetry during Drosophila development (3–5). Furthermore, essential roles of nonmuscle myosin-2 in tension transmission and contraction in cytokinesis, cell proliferation, adhesion, tissue patterning, border cell migration in the egg chamber, and nurse cell dumping are established (1, 6–11). The functional spectrum of Drosophila nonmuscle myosin-2 is similar to the mechanisms underlying developmental and physiologic functions of mammalian nonmuscle myosin-2 paralogs (12). Therefore, it is not surprising that Drosophila has been successfully used as a model system to study human nonmuscle myosin-2 (MYH9)-related diseases (13). This raises the general question if human disease phenotypes that are manifested in mutations within the catalytically active myosin motor domain could be predicted from Drosophila as a model system. To address this question, we performed a detailed kinetic characterization of a Drosophila nonmuscle myosin-2 subfragment 1 (S1)–like construct. Comparative analysis of kinetic signatures with selected nonmuscle myosin-2 paralogs from model organisms suggests a similar overall kinetic mechanism between Drosophila and mammalian nonmuscle myosin-2s, underlining the evolutionary closer relationship among metazoa compared with protozoa.

Blebbistatin is a widely used myosin-2-specific inhibitor whose potency has some variation among different myosin-2 isoforms (14, 15). Drosophila nonmuscle myosin-2 is insensitive toward blebbistatin, as shown by the failure of blebbistatin to inhibit cytokinesis in Drosophila S2 cells (15). We show that blebbistatin in the concentration range up to 200 µM does not inhibit the actin-activated ATPase activity of Drosophila nonmuscle myosin-2. In contrast, a switch-2 mutant in which Met466 is replaced with isoleucine, hereafter referred to as M466I, is partially inhibited by blebbistatin with a half-maximal inhibitory concentration (IC₅₀) of 36.3 ± 4.1 µM. This finding indicates that switch-2 plays a pivotal role in the establishment of blebbistatin sensitivity and the allosteric communication between nucleotide and inhibitor binding site.

MATERIALS AND METHODS

Chemicals and reagents

Standard chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA). Mant-fluorophores were from BIOLOG Life Science Institute (Bremen, Germany). Racemic blebbistatin was purchased from Calbiochem (San Diego, CA, USA). Rabbit skeletal muscle acetone powder was from Pel-Freez Biologicals (Rogers, AR, USA).

Cloning, recombinant overproduction, and protein purification

The cDNA of a myosin S1-like motor domain construct corresponding to amino acids 1–841 of Drosophila nonmuscle myosin-2 (isoform C, accession number NP_001014553.1) was amplified and inserted into a modified pFastBac1 vector, encoding a C-terminal Flag-tag. The switch-2 mutation M466I was introduced with standard cloning techniques. A polycistronic vector comprising the cDNAs for Sqh (19.1 kDa, accession number NP_511057.1) and Mlc-c (16.6 kDa, accession number NP_511049.1) under the control of the polyhedrin and P10 promoter, respectively, was a generous gift from Adam C. Martin (Massachusetts Institute of Technology, Cambridge, MA, USA). Transposition, production of recombinant baculoviruses, and Spodoptera frugiperda (Sf9) insect cell infection were performed as recommended by the manufacturer (Invitrogen, Carlsbad, CA, USA). The myosin heavy chain constructs were overproduced along with the respective light chains and purified to electrophoretic homogeneity as described earlier (16). F-actin was prepared according to Lehrer and Kerwar (17) and labeled with N-(1-pyrene)iodoacetamide (pyrene) as described in Criddle et al. (18).

Kinetic assays

Steady-state ATPase assays were performed as described previously at 25°C in buffer containing 10 mM 3-(N-morpholino) propanesulfonic acid (MOPS) (pH 7.0), 2 mM MgCl₂, 0.15 mM EGTA, 40 U/ml l-lactic dehydrogenase, 200 U/ml pyruvate kinase, 200 µM β-NAD, 1 mM phosphoenolpyruvate, 50 mM NaCl, and 2 mM ATP (16). The myosin concentration was between 0.2 and 0.68 µM. To assay the effect of blebbistatin on myosin’s steady-state ATPase activity, dilutions of the compound were made in 90% DMSO. The reaction mixture was incubated with blebbistatin for 10 min at room temperature prior to the assay. To minimize the impact of the solvent on the assay, the DMSO concentration was kept constant at 1.8%.

Transient kinetic assays were carried out in an SF-61DX2 stopped-flow system (Hi-Tech Scientific, Bradford-on-Avon, United Kingdom) at 20°C in a buffer containing 25 mM MOPS (pH 7.0), 100 mM KCl, 5 mM MgCl₂, and 0.1 mM EGTA, unless stated otherwise. Fluorescence excitation/emission setups were as follows: tryptophan, 297/320 nm cutoff filter; pyrene and mant-nucleotides, 365/390 nm long-pass filter; and Förster transfer from tryptophan to the mant moiety, 297/390 nm long-pass filter. Light-scattering assays were performed at a wavelength of 320 nm, and the signal was detected in a 90° angle to the incident beam. For F-actin binding and release assays, the KCl concentration was reduced to 2.5 mM. In F-actin binding assays, the [myosin]/[F-actin] ratio was kept between 1:5 and 1:7.5 and the reaction mixture treated with 0.1 U/ml apyrase prior to the assays for apo conditions, when required. Concentrations stated throughout the text are postmixing (1:1). Data storage and initial data analysis were accomplished with Kinetic Studio 2.20 (Hi-Tech Scientific). Secondary plots were made with Origin 7 (OriginLab, Northampton, MA, USA). Kinetic rate and equilibrium constants were interpreted as described earlier based on the proposed kinetic features established for skeletal muscle myosin-2 (Fig. 1A; where bold indicates rate and equilibrium constants in the lower pathway and in pathways perpendicular to the lower that are involved in actin-bound states) (19).

Figure 1.

A) Minimum kinetic scheme of the myosin and actomyosin ATPase cycle. K_A represents the affinity of myosin for F-actin, K_D is the affinity of ADP for myosin, K_AD is the affinity of ADP for the actomyosin complex, and K_DA is the affinity of F-actin for myosin in the presence of saturating [ADP]. M refers to myosin. The asterisk represents different conformational states of the myosin motor domain detected by the intensity of the intrinsic tryptophan fluorescence. Kinetic parameters in the presence and absence of F-actin are in boldface (k₊₁ and K₁) vs. normal (k₊₁ and K₁) notation. Subscripts A and D refer to actin (K_A) and ADP (K_D), respectively. Dissociation equilibrium constants are calculated according to K_x = k_-x/k_+x. B) The overview figure shows the homology model of the Drosophila nonmuscle myosin-2 motor domain in cartoon representation. The nucleotide is in stick representation; switch-1, switch-2, and P-loop are in teal color. Sequence analysis is shown for the conserved structural elements P-loop, switch-1, and switch-2 in the nucleotide binding pocket. Dark gray highlights conserved amino acids. The nonconsensus amino acid Met466 in Drosophila nonmuscle myosin-2 is shown in teal. Dm, Drosophila melanogaster; Dd, Dictyostelium discoideum; Ac, Acanthamoeba castellanii; Sc, Saccharomyces cerevisiae; Ce, Caenorhabditis elegans; Rn, Rattus norvegicus; Mm, Mus musculus; Hs, Homo sapiens. C) Actin-activated steady-state ATPase activity. The dependence of the steady-state ATPase activity on [F-actin] was measured in the concentration range up to 200 µM. The data are corrected for the k_basal. Fitting of the data with the Michaelis-Menten equation yields the parameters k_cat = 0.54 ± 0.06 s⁻¹ and K_app = 152 ± 39 µM.

Homology modeling

The structure of the Drosophila nonmuscle myosin-2 motor domain (amino acids 1–780) was modeled using the structure of the human nonmuscle myosin-2C motor domain [Protein Data Bank identification (PDB ID) 2YCU] in the prepower stroke state as a template. Both proteins share 72% sequence identity and 82% homology. The model was built using Modeler 9.14 (20). For the blebbistatin binding site analysis, the Drosophila nonmuscle myosin-2 homology model was superimposed onto the Dictyostelium nonmuscle myosin-2 motor domain in complex with blebbistatin (PDB ID 1YV3). In silico mutagenesis was performed to mutate Met466 to Ile466, and the residues in the binding site were minimized for structural analysis. Side-chain conformations in the blebbistatin binding sites were optimized using backbone-dependent rotamer libraries (21) included in the University of California (San Francisco, CA, USA) Chimera molecular visualization package (22).

RESULTS

Sequence analysis and comparison with other nonmuscle myosin-2s

The Drosophila nonmuscle myosin-2 motor domain core shares around 71 and 86% sequence identity and similarity, respectively, with vertebrate nonmuscle myosin-2 isoforms from human, rat, zebrafish, and Xenopus at the protein level. Those numbers/similarities reduce to 45 and 71% for unicellular nonmuscle myosin-2s from Dictyostelium, Acanthamoeba, and Saccharomyces, underlining the phylogenetic closer relationship to higher eukaryotes (Supplemental Fig. 1S). Sequence analysis further indicates that the nucleotide binding pocket comprises the conserved structural elements P-loop (GESGAGKT) and switch-1 (NXSSRFGK). Switch-2, the third active-site loop involved in nucleotide coordination, deviates from the consensus sequence (DIXGFE) and comprises the amino acids DMAGFE (Fig. 1B).

Protein production and steady-state ATPase activity

An S1-like fragment of Drosophila nonmuscle myosin-2 was overproduced along with the respective light chains in Sf9 insect cells. Approximately 1–2 mg protein was obtained per liter of cells and purified to electrophoretic homogeneity.

The steady-state ATP turnover of the recombinant protein is activated ∼33-fold by [F-actin] from a basal activity of 0.016 s⁻¹ to a maximum calculated catalytic activity (k_cat) of 0.54 ± 0.06 s⁻¹ (Fig. 1C). The K_app of 152 ± 39 µM is very high, resulting in a low coupling efficiency k_cat /K_app of ∼0.004 µM⁻¹ s⁻¹. A low k_cat /K_app is a characteristic of all kinetically described nonmuscle myosin-2 paralogs (Table 1).

TABLE 1.

Steady-state ATPase parameters obtained in the present study for Drosophila nonmuscle myosin S1

Parameter	Signal	Drosophila nonmuscle myosin-2	Acanthamoeba nonmuscle myosin-2 (23)	Dictyostelium nonmuscle myosin-2 (24)	Human nonmuscle myosin-2A (25)
kbasal (s−1)	NADH assay	0.016	0.05	0.08 ± 0.02	0.013 ± 0.004
kcat (s−1)a	NADH assay	0.54 ± 0.06	4.68 ± 0.73	2.6 ± 1	0.17 ± 0.005
Kapp (µM)a	NADH assay	152 ± 39	428 ± 97	102 ± 20	72 ± 4
kcat /Kapp (µM−1 s−1)	NADH assay	0.004	0.01b	0.025c	0.0025

For comparison, selected kinetic parameters from nonmuscle myosin-2s are listed. ^aMichaelis-Menten parameters obtained from a hyperbolic fit to the data set. ^bMaximum steady-state ATPase at 270 µM F-actin. ^cFrom the initial slope of the steady-state ATPase vs. [F-actin] plot.

ATP interaction

ATP binding to the myosin motor fragment results in a single-exponential transient increase in tryptophan fluorescence (Fig. 2A, inset). The concentration dependence of the observed rate constant (k_obs) on [ATP] can be fitted with a hyperbolic function, yielding a maximum value k₊₃ + k₋₃ of 11.65 ± 0.56 s⁻¹ (Fig. 2A). Half-saturation (1/K₁) is observed at [ATP] = 3.73 ± 1.32 µM (Fig. 2A). A linear fit of the data at low [ATP] yields a second-order rate constant for ATP binding, K₁k₊₂, of 1.16 ± 0.06 µM⁻¹ s⁻¹(Fig. 2B). A single-exponential time evolution of the fluorescence signal is also observed upon mixing 2′-deoxy-3′-O-(N′-methylanthraniloyl) (d-mant) ATP with myosin under pseudo first-order conditions. A representative transient is shown in the inset of Fig. 2B. A secondary plot of the k_obs versus [d-mantATP] results in a linear dependence describing the second-order rate constant K₁k₊₂ to 0.94 ± 0.02 µM⁻¹ s⁻¹ (Fig. 2B).

Figure 2.

Interaction between myosin and actomyosin and ATP. A) Increasing [ATP] was mixed under pseudo first-order conditions with 0.25 µM myosin in a stopped-flow spectrophotometer. The dependence of k_obs on [ATP] is hyperbolic, yielding the parameter k₊₃ + k_-3 of 11.65 ± 0.56 s⁻¹ at saturation. 1/K₁ is observed at [ATP] = 3.73 ± 1.32 µM. The inset shows a representative trace of the transient increase in tryptophan fluorescence after mixing 7.5 µM ATP with 0.25 µM myosin. Single-exponential fit to the data yields an amplitude of 14.2% and a k_obs of 6.6 s⁻¹. B) Comparison of the second-order rate constants for ATP and dmantATP binding to myosin at low [nucleotide]. The dependence of k_obs on [nucleotide] is linear and results in a K₁k₊₂ of 1.16 ± 0.06 µM⁻¹ s⁻¹ (ATP) and K₁k₊₂ to 0.94 ± 0.02 µM⁻¹ s⁻¹ (d-mantATP). A fluorescence transient of the bimolecular interaction between 0.15 µM myosin and 3 µM d-mantATP is shown in the inset of (B). Single-exponential fit to the data set yields an amplitude of 2.8% and a k_obs of 3.06 s⁻¹. C) ATP-induced dissociation of the actomyosin complex. Fluorescence signals were obtained after mixing pyrene-actomyosin under pseudo first-order conditions with increasing [ATP] in a stopped-flow spectrophotometer. Hyperbolic fit to the data yields a k₊₂ of 482.8 ± 7.85 s⁻¹ at saturating [ATP]. 1/K₁ is observed at 1224.37 ± 88.39 µM [ATP]. Inset shows a representative transient of the interaction between 0.5 µM pyrene-actomyosin and 7.5 µM ATP. Single-exponential fit to the data set yields an amplitude of 69% and a k_obs of 3.18 s⁻¹. D) Comparison of the k_obs upon the ATP-induced dissociation of the actomyosin complex. Monitored is either the postmixing pyrene or the light-scattering signal. Linear fits to the data sets up to 30 µM [ATP] define the second-order rate constants K₁k₊₂ of 0.39 ± 0.01 µM⁻¹ s⁻¹ (pyrene), and K₁k₊₂ of 0.34 ± 0.01 µM⁻¹ s⁻¹ (light-scattering).

ATP binding to pyrene-actomyosin dissociates the complex and results in a transient increase in pyrene fluorescence, as shown in the inset of Fig. 2C. A k_obs versus [ATP] plot is hyperbolic (Fig. 2C). A maximum value of 482.8 ± 7.85 s⁻¹, representing the isomerization constant k₊₂, can be extrapolated from the data set. Half-saturation is observed at 1222.3 ± 60.29 μM ATP. The data are linear at low [ATP] up to 30 µM, yielding the second-order rate constant K₁k₊₂ of 0.39 ± 0.01 µM⁻¹ s⁻¹ (Fig. 2D). In agreement, the light-scattering signal upon the ATP-induced dissociation of the actomyosin complex results in the second-order rate constant K₁k₊₂ of 0.34 ± 0.01 µM⁻¹ s⁻¹ (Fig. 2D). Table 2 lists all kinetic constants for the interaction between myosin and actomyosin with ATP.

TABLE 2.

Transient kinetic parameters obtained in the present study for Drosophila nonmuscle myosin-2 S1

Parameter	Signal or calculation	Drosophila nonmuscle myosin-2	Acanthamoeba nonmuscle myosin-2 (23)	Dictyostelium nonmuscle myosin-2 (24, 26–28)	Human nonmuscle myosin-2A (25)
ATP interaction
K1k+2 (µM−1 s−1)	Tryptophan	1.16 ± 0.06	1.33 ± 0.04	0.86	0.56 ± 0.01
K1k+2 (µM−1 s−1)	d-mantATP	0.94 ± 0.02	1.93 ± 0.02	1.13	1.03 ± 0.14
1/K1 (µM)	Tryptophan	3.73 ± 1.32	5.19 ± 0.58	25.8	n.d.
k3 + k-3 (s−1)	Tryptophan	11.65 ± 0.56	19.0 ± 0.5	30	14.1 ± 0.5
K1k+2 (µM−1 s−1)	Pyrene-actin	0.39 ± 0.01	2.11 ± 0.03	0.25	0.21 ± 0.04
K1k+2 (µM−1 s−1)	Light scattering	0.34 ± 0.01	0.73 ± 0.02	n.d.	0.14 ± 0.003
1/K1 (µM)	Pyrene-actin	1222.3 ± 60.29	651.3 ± 68.4	509	∼900
k+2 (s−1)	Pyrene-actin	482.80 ± 7.85	779.6 ± 24.9	380	∼190
ADP interaction
k+D (µM−1 s−1)	d-mantADP	0.52 ± 0.04	0.50 ± 0.03	1.02	0.55 ± 0.06
k−D (s−1)	d-mantADPa	2.42 ± 0.13	2.29 ± 0.20	n.d.	0.54 ± 0.23
k−D (s−1)	d-mantADPb	2.92 ± 0.03	1.53 ± 0.01	1.6	1.12 ± 0.13
k−D (s−1)	Tryptophanb	2.35 ± 0.01	1.68 ± 0.02	n.d.	n.d.
KD (µM)	k−D/k+D	4.65 ± 0.44	4.58 ± 0.49	1.6	1.5 ± 0.4
k+AD (µM−1 s−1)	d-mantADP	3.46 ± 0.11	∼4.95c	0.72	2.72 ± 0.16
k−AD (s−1)	d-mantADP a	5.19 ± 0.55	n.d.	n.d.	n.d.
k−AD (s−1)	Pyrene-actinb	5.53 ± 0.08	28.9 ± 0.07	15	n.d.
k−AD (s−1)	Light scatteringb	4.27 ± 0.05	29.67 ± 0.15	n.d.	n.d.
k−AD (s−1)	d-mantADPb	5.06 ± 0.05	31.43 ± 0.49	n.d.	2.68 ± 0.30
KAD (µM)	k−AD/k+AD	1.5 ± 0.17	5.99 ± 0.88	58	1.4 ± 0.4
Coupling	KAD/KD	0.32 ± 0.05	1.30 ± 0.24	∼36	0.7
Coupling	k−AD/k−D	2.14 ± 0.25	∼17.6	∼9.4	2.8
Actin interaction
k+A (µM−1 s−1)	Light scattering	1.14 ± 0.03	10.88 ± 0.22	1.34	0.73 ± 0.03
k−A (s−1)	Pyrene-actin	∼0.07	∼0.001	0.0068	<0.007
KA (nM)	k−A/k+A	∼60	∼0.1	∼5.1	<10
k+DA (µM−1 s−1)	Light scattering	0.26 ± 0.01	1.17 ± 0.04	n.d.	0.19 ± 0.02
k−DA (s−1)	Pyrene-actin	∼0.01	∼0.014	n.d.	∼0.004
KDA (nM)	k−DA/k+DA	∼38	∼12	n.d.	∼20

For comparison, the kinetic parameters from selected nonmuscle myosin-2s are listed. n.d., not determined. ^aFrom y-intercept. ^bFrom chasing experiment. ^cCalculation from k_−AD/K_AD.

ADP interaction

The bimolecular interaction between myosin and the fluorescent nucleotide analog d-mantADP results in a single-exponential transient fluorescence increase, indicating a 1-step binding mechanism. Plotting of k_obs vs. [d-mantADP] results in a linear dependence (Fig. 3A). Fitting the data to a straight line yields a second-order binding rate constant for ADP, k_+D, of 0.52 ± 0.04 µM⁻¹ s⁻¹. Extrapolating the data to [d-mantADP] = 0 yields the ADP release rate constant k_−D of 2.42 ± 0.13 s⁻¹. For comparison, direct measurement of the respective constant in ATP-chasing experiments yields a k_−D of 2.92 ± 0.03 s⁻¹ or 2.35 ± 0.01 s⁻¹ with the extrinsic (d-mantADP) or intrinsic (tryptophan) fluorescence signal (Fig. 3B). Calculation of the dissociation equilibrium constant, based on the mathematical relationship k_−D/k_+D, results in a value of 4.65 ± 0.44 µM.

Figure 3.

Interaction between myosin and actomyosin with ADP. A) [d-mantADP] dependence of the k_obs for myosin (closed circles) or actomyosin (open circles) binding to the fluorescent ADP analog. Linear fits to the data sets yield the second-order binding rate constants k_+D of 0.52 ± 0.04 µM⁻¹ s⁻¹ and k_+AD of 3.46 ± 0.1 s⁻¹. Extrapolating the data to [d-mantADP] = 0 yields the ADP release rate constant k_−D of 2.42 ± 0.13 s⁻¹ and k_−AD of 5.19 ± 0.55 s⁻¹, in the absence and presence of F-actin, respectively. B) Representative transients for the ADP release from myosin and actomyosin (inset). Assay conditions were as follows in the absence of F-actin: 1.5 mM ATP versus 0.25 µM myosin and 15 µM ADP. The fluorescence was exited at a wavelength of 297 nm and the fluorescence emission detected after the passage of a 320 nm cutoff filter. The obtained transient was best described by a single exponential function yielding a k_obs of 2.35 s⁻¹. The inset shows a light-scattering signal upon the dissociation of a 0.25 µM pyrene-actomyosin complex in the presence of 10 µM ADP with 4 mM ATP. Single-exponential fit of the data yields a k_obs of 4.27 s⁻¹.

Mixing myosin with increasing [d-mantADP] in the presence of F-actin results in a 7-fold acceleration of the second-order ADP binding rate constant to k_+AD = 3.46 ± 0.11 µM⁻¹ s⁻¹ (Fig. 3A). Extrapolating the linear fit to [d-mantADP] = 0 yields the apparent ADP release rate constant, k_−AD, of 5.19 ± 0.55 s⁻¹. The ADP release rate from actomyosin was independently measured by chasing d-mantADP or ADP from the actomyosin complex with excess [ATP]. Fluorescence emission transients of d-mantADP, pyrene, or a change in the light-scattering signal result in observed dissociation constants of k_−AD = 5.06 ± 0.05 s⁻¹_, k_−AD = 5.53 ± 0.08 s⁻¹, or k_−AD = 4.27 ± 0.05 s⁻¹, respectively (Fig. 3B, inset). Calculation of the dissociation equilibrium constant K_AD (K_AD = k_−AD/k_+AD) gives 1.5 ± 0.17 µM. Interestingly, the actin-activated ADP release rate is not significantly altered in the presence of varying concentrations of free [Mg²⁺] in the physiologic range (0.2–2 mM; data not shown). All kinetic constants describing the interaction between myosin and actomyosin with ADP are listed in Table 2.

F-actin interaction

The bimolecular interaction between myosin and F-actin was probed in light-scattering assays in the presence and absence of saturating [ADP]. Mixing of both proteins results in an increase of the light-scattering signal (Fig. 4B). A single exponential fit to the data set indicates a 1-step binding event. Plotting of k_obs as a function of [F-actin] results in a linear dependence, yielding the second-order rate constants for F-actin binding k_+A of 1.14 ± 0.03 µM⁻¹ s⁻¹ and k_+DA of 0.26 ± 0.01 µM⁻¹ s⁻¹ in the absence or presence of saturating [ADP] (Fig. 4A). The corresponding release rate constants k_−A ∼0.07 s⁻¹ and k_−DA ∼0.01 s⁻¹ were obtained from chasing experiments (Fig. 4B, inset). From the ratio of the release and binding rate constants (k_−A/k_+A and k_−DA/k_+DA), the dissociation equilibrium constants K_A and K_AD were calculated to ∼60 and ∼38 nM. The kinetic constants describing the interaction between F-actin and Drosophila nonmuscle myosin-2 are listed in Table 2.

Figure 4.

Interaction between myosin and F-actin. A) [F-actin] dependence of k_obs for the reaction between myosin and filamentous actin. The second-order binding rate constants k_+A of 1.14 ± 0.03 µM⁻¹ s⁻¹ (open circle) and k_+DA of 0.26 ± 0.01 µM⁻¹ s⁻¹ (closed circle) in the absence or presence of 100 µM [ADP] were obtained from the linear fits to the data. B) Representative light-scattering signal of the interaction between 0.3 µM myosin and 1.5 µM F-actin. Single-exponential fit of the data yields an amplitude of 1.89% and a k_obs of 2.65 s⁻¹. The inset shows an observed F-actin release rate constant k_obs of 0.07 s⁻¹ that was obtained after fitting the transient, obtained from chasing 0.25 µM pyrene-actomyosin with 10 µM F-actin, to a single exponential function.

Interaction with blebbistatin

The interaction between myosin and the commonly used myosin-2 inhibitor blebbistatin was tested in the actin-activated ATPase assay. At constant [F-actin] of 30 µM, increasing [blebbistatin] up to 200 µM does not significantly inhibit the steady-state ATPase activity of Drosophila nonmuscle myosin-2 (Fig. 5A). Structural analysis of the Dictyostelium nonmuscle myosin-2/blebbistatin complex shows that the inhibitor binds to a pocket at the apex of the 50 kDa cleft, at a distance of ∼15 Å from the nucleotide binding pocket (29). Superimposition with the Drosophila nonmuscle myosin-2 motor domain homology model indicates that all key myosin/blebbistatin binding site residues, as defined by Allingham et al. (29), are conserved (Table 3). A notable feature of the Drosophila nonmuscle myosin-2 blebbistatin binding site is that amino acid Met466 substitutes for a consensus isoleucine in switch-2, as outlined in greater detail above. Switch-2 is not only an essential part of the nucleotide binding region but also flanks the rear of the blebbistatin binding site. Structural analysis of the Drosophila nonmuscle myosin-2 motor domain homology model in the prepower stroke state shows that the large side chain of Met466 occupies the blebbistatin binding site and restricts the free conformational space around the switch-2 loop. Out of the 13 possible rotamers for Met466 backbone-dependent side-chain conformations, 2 of the major rotamers, representing the energetically most favorable (17 and 13%) and the most likely side-chain conformations, occlude the inhibitor binding site (Fig. 5C). In contrast, the consensus isoleucine side-chain rotamer is positioned normally and allows free conformational space for blebbistatin to bind properly (Fig. 5D). We hypothesize that Met466 sterically hinders blebbistatin binding and interferes with the communication between the nucleotide and the allosteric blebbistatin binding site and thus could partially explain the insensitive nature of Drosophila nonmuscle myosin-2 toward blebbistatin.

Figure 5.

Interaction between myosin and blebbistatin. A) Normalized (Norm.) actin-activated steady-state ATPase activity of Drosophila nonmuscle myosin-2 (open circles) and M466I (closed circles) as a function of [blebbistatin]. [F-actin] was kept constant at 30 µM, and the data were normalized to the starting value of the ATPase activity in the absence of inhibitor. Single-exponential fit of the data set gives an IC₅₀ of 36.3 ± 4.1 µM for M466I, whereas no inhibition is observed for the native protein [wild-type (WT)]. B) Blebbistatin-mediated decrease of the steady-state actin-activated ATPase activity. Increasing [blebbistatin] leads to a concentration-dependent decrease in the actin-activated ATP turnover. At 90 µM [F-actin], the ATPase activity is reduced >2-fold. For comparison, the uninhibited control was supplemented with 1.8% DMSO to account for the solvent effect. C, D) Close-up views of the blebbistatin binding site in the Drosophila nonmuscle myosin-2 and M466I motor domain homology models in prepower stroke state (ADP·P_i). Key residues are shown as sticks in teal, blebbistatin in mango color. There are 2 major side-chain rotamers [17% (1), and 13% (2)] of Met466 in the active-site loop switch-2 shown for nonmuscle myosin-2 (C). In silico substitution of switch-2 residue methionine 466 by “isoleucine” increases the conformational space available for blebbistatin binding in mutant M466I (D).

TABLE 3.

Comparison of IC₅₀ values and blebbistatin binding site residues of nonmuscle myosin-2s from selected model organisms

	Drosophila nonmuscle myosin-2	Drosophila nonmuscle myosin-2 M466I	Acanthamoeba nonmuscle myosin-2	Dictyostelium nonmuscle myosin-2	Human nonmuscle myosin-2A	Blebbistatin interaction
IC50 (µM)	>>200	36.3	83 (14)	4.9 (14)	5.1 (14)
Residue	Leu271	Leu271	Leu262	Leu262	Leu258	Benzyl ring
Residue	Phe477	Phe477	Phe468	Phe466	Phe464	Benzyl ring
Residue	Glu478	Glu478	Glu469	Glu467	Glu465	Benzyl ring
Residue	Val651	Val651	Val652	Val630	Val646	Benzyl ring
Residuea	Ala467	Ala467	Ser458	Ser456	Ala454	Tetrahydropyrrolo ring
Residue	Ile482	Ile482	Ile473	Ile471	Ile469	Tetrahydropyrrolo ring
Residue	Thr485	Thr485	Thr476	Thr474	Thr472	Methylquinolinone
Residue	Tyr655	Tyr655	Tyr656	Tyr634	Tyr650	Methylquinolinone
Residue	Gln658	Gln658	Gln659	Gln637	Gln653	Methylquinolinone
Residue	Leu662	Leu662	Leu663	Leu641	Leu657	Methylquinolinone
Residuea	Met466	Ile466	Ile459	Ile457	Ile455	None

Reference binding site residues were previously described for the Dictyostelium nonmuscle myosin-2/blebbistatin complex by Allingham et al. (29). Residues corresponding to Drosophila nonmuscle myosin-2 Met466 are listed for completeness. ^aDeviations in the amino acid sequence.

To test for the potential impact of the nonconsensus switch-2 residue Met466 on the blebbistatin insensitivity of Drosophila nonmuscle myosin-2, we overproduced the switch-2 mutant M466I, in which methionine Met466 is replaced by a consensus isoleucine. M466I has a basal ATPase activity (k_basal) of 0.021 s⁻¹ under steady-state conditions. The k_cat of 0.58 ± 0.1 s⁻¹ is similar to the respective value of Drosophila nonmuscle myosin-2, whereas the K_app of 299 ± 88 µM is higher (data not shown). Supporting the prediction from the in silico modeling approach, M466I is blebbistatin sensitive. Blebbistatin inhibits the steady-state actin-activated ATPase of M466I by a maximum of 50% and an IC₅₀ of 36.3 ± 4.1 at 30 µM F-actin (Fig. 5A). The effect of blebbistatin on the steady-state ATPase turnover of M466I in the concentration range up to 90 µM [F-actin] was examined (Fig. 5B). It is not possible to measure the K_app of M466I with great accuracy given its very high value and the need for extremely high actin concentrations and, thus, is not possible to draw a firm conclusion whether K_app, k_cat, or both parameters are altered by [blebbistatin]. However, direct comparison of the ATPase activities at 90 µM [F-actin] indicates a 50% inhibition between 0 and 200 µM. The same trend is also observed for the k_basal (Fig. 5B).

DISCUSSION

Actomyosin ATPase cycle

Drosophila nonmuscle myosin-2 has a nonconsensus amino acid (Met466) in switch-2 instead of isoleucine, which is normally present in myosins. Isoleucine substitutions at this site are extremely rare in the myosin superfamily but have been reported for mammalian myosin-18A and both myosin-18 and myosin-20 from Drosophila. Both myosin classes have additional nonconsensus amino acids in the nucleotide binding region and are enzymatically inactive pseudoenzymes (30–32). However, our studies demonstrate that Drosophila nonmuscle myosin-2 is a bona fide motor as would be predicted by genetic studies in Drosophila and tissue culture models (11, 33).

Signatures include a slow steady-state ATPase activity and a low coupling ratio k_cat/K_app (Table 1). M466I has a similar k_cat but a 50% reduced coupling efficiency due to an increased K_app. The k_basal is slightly increased in M466I compared with the wild-type motor domain. Overall, the observed effects on the steady-state ATPase activity resemble alanine scanning results at the respective site in Dictyostelium nonmuscle myosin-2, indicating that an isoleucine to alanine substitution does not impact the overall in vitro and in vivo motor function but results in a slightly higher k_basal (33). Drosophila nonmuscle myosin-2 has a very high affinity for ADP, and the ADP release rate is only marginally activated by F-actin (k_−AD/k_−D, ∼2). The actin-activated ADP release k_−AD (∼5 s⁻¹) is around 10 times faster than the k_cat, making it unlikely to rate limit the catalytic cycle under physiologic conditions. In agreement with other reports on class-2 myosins, the release of inorganic phosphate (P_i) is the kinetic step limiting the actomyosin ATPase cycle (23, 25). The fast ADP release rate compared with the k_cat leads to a low duty ratio of 0.1, calculated from

Contractile forces exerted by nonmuscle myosin-2 derive from bipolar minifilamentous structures in which the myosin motor domains are radially distributed at opposing filament ends. As described by Kiehart and Feghali (34), native Drosophila nonmuscle myosin-2 bipolar filaments contain on average 16 myosins, resulting in 16 heads on each side of the filament. Although Drosophila nonmuscle myosin-2 would not be processive as a single molecule, such a filament composed of multiple myosin-2 molecules could have a high effective duty ratio and therefore act as a processive unit. Based on the filament geometry, the effective duty ratio of a Drosophila nonmuscle myosin-2 filament equals 0.81 using the following equation:

with n equaling the number of heads per filament side. The high effective duty ratio would promote a continuous, processive interaction between a Drosophila nonmuscle myosin-2 filament and neighboring F-actin, as it has been reported for human nonmuscle myosin-2B (35). The duty ratio of Drosophila nonmuscle myosin-2 is comparable to the duty ratios for isoforms of vertebrate nonmuscle myosins-2B and -2C (16, 36), whereas amoeboid nonmuscle myosins from unicellular eukaryotes have slightly lower duty ratios (23).

Kinetic comparison of nonmuscle myosin-2 paralogs from model organisms

In comparison to previously characterized nonmuscle myosin-2s from model organisms, the overall kinetic signatures of Drosophila nonmuscle myosin-2 resemble those of the mammalian paralogs, underlining the closer phylogenetic relationship (Table 2). Its steady-state ATPase activity and the actin-activated ADP release rate constants are similar to those reported for human class-2 nonmuscle myosins (Table 2) (16, 25, 36). In comparison, the reported values for class-2 myosins from Acanthamoeba and Dictyostelium (15–31 s⁻¹) are significantly higher (Table 2) (23, 26). We speculate that alternatively spliced variants of fly nonmuscle myosin-2 with insertions of either 8 or 40 amino acids in loop-1 might have faster actin-activated ADP release rates, as it has been described for human nonmuscle myosin-2C and smooth muscle myosin-2 splice variants (36–39). Possible functional consequences include alterations in duty ratio and mechanoenzymatic properties. Transient kinetic signatures of nonmuscle myosin-2s from amoeba to humans include reduced second-rate binding constants for ATP in the presence of F-actin and fast second-order binding rate constants for ADP in the presence of F-actin (k_+AD/K₁k₊₂). The resulting product inhibition impacts the duty ratio and the in vitro sliding velocity of the proteins (12). For the cytoplasmic myosin-2 from Drosophila, the k_+AD/K₁k₊₂ ratio of ∼10 is identical to the number calculated for human nonmuscle myosin-2A and slightly higher when compared to Acanthamoeba (k_+AD/K₁k₊₂, ∼7) and Dictyostelium nonmuscle myosin-2 (k_+AD/K₁k₊₂, ∼3) (Table 2). Kinetic (k_−AD/k_−D) and thermodynamic (K_AD/K_D) coupling ratios of ∼2 and ∼0.3 are low for the Drosophila paralog and similar to those reported for nonmuscle myosin-2s from human. The respective amoeboid myosins have higher kinetic coupling ratios between 9.4 and 17.6, indicating a significant increase in the ADP release rate by F-actin. The thermodynamic coupling ratio for Dictyostelium is around 10-fold higher than the one reported for Drosophila nonmuscle myosin-2.

Although nonmuscle myosin-2s from Drosophila and human are kinetically similar, the actin-activated ADP release of Drosophila cytoplasmic myosin-2 is magnesium insensitive in the physiologic range of free [Mg²⁺]. [Mg²⁺] sensitivity has been described as a modulatory mechanism for the ADP release rate constants, ADP affinities, and duty ratios of mammalian nonmuscle myosin-2s (36). This behavior suggests a different allosteric communication within the Drosophila nonmuscle myosin-2 motor domain, which resembles more the pattern of amoeboid myosins or rabbit skeletal muscle myosin-2.

Extrinsic regulation

Blebbistatin is a class-2–specific, allosteric, noncompetitive myosin inhibitor, which was identified as a small molecule compound that inhibits cytokinesis in various cells types from model organisms, with the exception of Drosophila (15). Previous in vitro studies demonstrated that blebbistatin inhibited the actin-activated ATPase activity of vertebrate and amoeboid nonmuscle myosin-2s by >80%. The IC₅₀ values, reflecting the half-maximal inhibitory blebbistatin concentration, range from 1.8 to 5.1 µM (14, 15). The Acanthamoeba paralog shows only 60% inhibition and a high IC₅₀ with 83 µM (14). Consistent with the findings of the in vivo studies, the in vitro actin-activated steady-state ATPase activity of Drosophila nonmuscle myosin-2 is blebbistatin insensitive up to an inhibitor concentration of 200 µM (Fig. 5A) (15). This does not exclude a blebbistatin-mediated inhibition of the enzymatic activity at higher inhibitor concentration but suggests an IC₅₀ outside the experimentally accessible concentration for in vitro and in vivo studies.

Mechanistically, switch-2 undergoes a conformational change during the ATP hydrolysis cycle and alternates between an open and a closed conformation including the main-chain rotations at its consensus isoleucine (33, 40). Blebbistatin decreases the rate-limiting phosphate release step of the catalytic cycle and increases the lifetime of the weakly bound ADP·P_i intermediate by stabilizing the closed switch-2 conformation (41–43). In the Drosophila nonmuscle myosin-2 homology model in the prepower stroke state, Met466 rotamers occupy the free conformational space of the blebbistatin binding site, thereby interfering with compound binding. We further predict that an open switch-2 conformation results in an even more pronounced decrease of the inhibitor binding pocket. The switch-2 mutant M466I is blebbistatin sensitive with an IC₅₀ value of 36.3 ± 4.1 µM, which lies between those reported for Acanthamoeba nonmuscle myosin-2 and the paralogs from Dictyostelium and mammals (Table 3) (14, 44). As reported previously for Acanthamoeba nonmuscle myosin-2, blebbistatin only leads to a partial inhibition of the steady-state ATPase activity of M466I. Despite the substantial increase in blebbistatin sensitivity for M466I when compared to the wild-type protein, the solely 2-fold inhibition and the high IC₅₀ make the mutant protein not suitable for in vivo cell biologic applications. However, it underlines the proposed link between switch-2 and blebbistatin inhibition, which contributes to the understanding of the inhibitor-mediated myosin inhibition. Introducing further mutations at switch-2 in Drosophila nonmuscle myosin-2 in addition to isoleucine could possibly increase the conformational space at this region and increase the potency of blebbistatin. It may also lead to a new generation of blebbistatin derivatives that increase the inhibitory potency. Furthermore, the in vitro blebbistatin insensitivity of Drosophila nonmuscle indicates that the myosin/inhibitor interaction is more than an induced-fit mechanism. Nonbinding site residues within the myosin motor domain are less conserved in the class-2 myosins, allowing different degrees of protein motion and conformers, which could favor or constrain compound binding (45). Furthermore, the myosin/inhibitor complex formation must be thermodynamically favorable and long lived enough in terms of complex stability so that the induced fit can take place within a reasonable time. A third explanation for the reduced in vitro and in vivo sensitivity might be a different allosteric communication pathway that operates between the blebbistatin binding site and the active site. Blebbistatin has been occasionally used to probe nonmuscle myosin-2 motor function in Drosophila cells and tissues, and some small effects were observed (46). This would not be predicted by the in vitro results of the present study and the initial in vivo study (15), suggesting that there might be off-target effects as were noted in Dictyostelium myosin-2 null cells where blebbistatin inhibits cell streaming and other cellular processes (44).

CONCLUSIONS

Drosophila cytoplasmic myosin-2 is a low duty ratio molecular motor with kinetic signatures that resemble those of mammalian rather than amoeboid nonmuscle myosin-2s. This is consistent with the diverse roles that nonmuscle myosin-2s play in multicellular organisms. The nonconsensus amino acid Met466 in switch-2 is not crucial for Drosophila nonmuscle myosin-2 motor function but serves as blebbistatin desensitizer. It is unfortunate for those studying the function of nonmuscle myosin-2 in Drosophila that blebbistatin cannot be used as a pharmacologic tool. Our demonstration that the M466I mutation partially sensitizes nonmuscle myosin-2 to blebbistatin is the first step toward engineering a Drosophila nonmuscle myosin-2 that would be fully inhibited by this reagent. If such a mutant myosin could be engineered into the Drosophila genome using state-of-the-art gene-editing methodology, replacing the endogenous nonmuscle myosin-2 with a blebbistatin-sensitive version would greatly facilitate studies on the role of this myosin in fly development and homeostasis.

The overall kinetic similarities between Drosophila and mammalian nonmuscle myosin-2s suggest that Drosophila may be a good model system for testing the function of MYH9 disease mutations. However, we do not recommend this approach. Mutations responsible for most human genetic diseases involving myosin-2s lie in regions that are not directly associated with nucleotide or F-actin binding and are sometimes associated with subtle kinetic phenotypes (47, 48). The overall sequence identity between the Drosophila and mammalian nonmuscle myosin-2s motor domain is around 72%, which means that there will be many other amino acid substitutions, in addition to the residue that is intentionally mutated, which could alter the impact of the chosen mutation.

Glossary

d-mant

2′-deoxy-3′-O-(N′-methylanthraniloyl)

IC₅₀

half-maximal inhibitory concentration

k_basal

basal ATPase activity

k_cat

maximum catalytic activity

k_obs

observed rate constant

MOPS

3-(N-morpholino) propanesulfonic acid

PDB ID

Protein Data Bank identification

P_i

inorganic phosphate

S1

subfragment 1

Sf9

Spodoptera frugiperda