INFORMATION TO USERS

This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer.

The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction.

In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion.

Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book.

Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order.

University Microfilms International A Bell & Howell Information Company 300 North Zeeb Road. Ann Arbor, Ml 48106-1346 USA 313/761-4700 800/521-0600

Order Number 9427817

Analysis of the in vivo role of regulatory light chain phosphory lat ion

Tohtong, Rutaiwan, Ph.D.

The Ohio State University, 1994

UMI 300 N. ZeebRd. Ann Arbor, MI 48106

ANALYSIS OF THE IN VIVO ROLE OF MYOSIN REGULATORY LIGHT CHAIN PHOSPHORYLATION

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

By

Rutaiwan Tohtong, B.Sc.

*****

The Ohio State University

1994

Dissertation Commitee: Approved by Berl Oakley Amanda Simcox Caroline Breitenberger Richard Swenson

Department of Molecular Grenetics To my parents ACKNOWLEDGEMENTS

I would like to thank my parents for their constant moral support through out these years. I thank Dr. Berl Oakley and Dr. Amanda Simcox for their guidance over the last three years. I am especially grateful to Dr. Simcox for being both a friend and an adviser. I would like to thank my committee members for their valuable advice and time. To all my friends in Molecular Genetics, thank you for having made these years a good memory. And lastly, I would like to express my gratitude to Dr. Scott Falkenthal who helped initiate this project. VITA

April 18, 1962 Bom - Bangkok, Thailand 1982 ...... B.Sc. Microbiology, Chulalongkom University, Bangkok, Thailand 1986-present Graduate Research and Teaching Associate Department of Molecular Genetics The Ohio State University Columbus, Ohio

PUBLICATIONS

Tohtong, R., M. Graham, R. Schaaf, J. Hurley, A. Simcox, and D. Maughan. Analysis of the Role of Myosin Regulatory Light Chain (MRLC) Phosphorylation in Drosophila melanogaster. 1993. Biophys. J. 64(2): 8a.

Yamashita, H., R. Tohtong, A. Simcox. J. Vigoreaux, J. Haeberle, C. Hyatt, S. Brown, and D. Maughan. 1994. Assessment of the Role of Myosin Regulatory Light Chain (MRLC) Phosphorylation by In Vivo Mutagenesis in Drosophila. Biophys. J. 66(2): A123.

FIELDS OF STUDY

Major Field: Molecular Genetics iv TABLE OF CONTENTS

ACKNOWLEDGEMENTS...... iii VITA...... iv LIST OF TABLES...... xi LIST OF FIGURES...... xii CHAPTER I. INTRODUCTION ...... 1

Introduction ...... 1 Muscle structure...... 2 ...... 2 ...... 8 Cardiac...... 9 Composition of ...... 9 The thick filaments...... 9 Myosin heavy chain ...... 10 Myosin regulatory light chain ...... 11 Myosin essential light chain ...... 15 The hinge region ...... 16 The thin filaments ...... 17 ...... 17 ...... 18 The complex ...... 18 Other accessory proteins ...... 19 a-actinin ...... 19 ...... 19 ...... 20 The sliding filament model of muscle contraction 20

v Regulation of muscle contraction ...... 21 Thin filament control ...... 21 Thick filament control ...... 22 Regulation by phosphorylation of myosin light chain-2 ...... 23 Drosophila musculature...... 24 Supercontractile muscle ...... 24 Tubular muscle ...... 27 Fibrillar muscle ...... 28 Muscle proteins of Drosophila...... 29 Myosin heavy chain ...... 29 Myosin light chain 2 ...... 30 Myosin alkali light chain ...... 34 Actin ...... 34 Tropomyosin ...... 35 Troponin ...... 36 Param yosin ...... 36 Alpha actinin and other Z disc associated proteins ...... 37 Objectives ...... 39

II. MATERIALS AND METHODS...... 41

Fly stocks and culture conditions ...... 41 Construction of P element transformation vector and generation of transformants by germline transformation ...... 44 Mobilization of the P[Mlc2S66 67A] insert to multiple chromosome sites ...... 47 Rescue of the dominant flightless phenotype and the recessive lethality of Mlc2 null mutation ...... 48 Flight test ...... 53

vi Muscle protein preparation ...... 53 Protein electrophoresis ...... 54 In vivo labeling of muscle proteins ...... 55 Expression of Drosophila MLC-2 fusion protein in E. coll...... 56 Purification of the Drosophila MLC-2 fusion protein ...... 56 Production of MLC-2 specific antibodies and Western Hybridization ...... 59 Peptide sequencing...... 60 Electron microscopy...... 60 Preparation of genomic DNA from adult Drosophila...... 61 PCR amplification of the genomic DNA for cloning of the MLCK gene ...... 62 Isolation of Drosophila MLCK cDNA clones ...... 63 Southern Hybridization ...... 64 DNA sequencing ...... 64 Total RNA preparation ...... 64 Synthesis of RNA probes for RNAase protection assays ...... 65 RNAase protection assays ...... 65

III. MULTIPLE PHOSPHORYLATED VARIANTS OF DROSOPHILA MYOSIN LIGHT CHAIN 2 PROTEIN ACCUMULATE M VIVO...... 67

Introduction ...... 67 Results and Discussion ...... 74 Expression of Drosophila MLC-2 fusion protein in E. colt...... 74

vii Multiple variants of the Drosophila MLC-2 protein accumulates in vivo ...... 77 Reduction in MLC-2 protein accumulation in Mlc2 m u tan ts ...... 78 In vivo labeling of IFM proteins with 32P ...... 79 D iscussion ...... 80 Multiple IFM protein spots are recognized by an MLC-2 specific antibody ...... 80 Post translational modification of MLC-2 ...... 81

IV. THE EFFECTS OF DISRUPTION OF MYOSIN LIGHT CHAIN-2 PHOSPHORYLATION SITES ON FLIGHT. VIABILITY, AND INDIRECT FLIGHT MUSCLE ULTRASTRUCTURE...... 85

Introduction ...... 85 R esults...... 88 Disruption of MLC-2 phosphorylation by in vivo site-directed mutagenesis and construction of transformation vectors carrying the mutant Mlc2 genes ...... 88 Generation of a series of lines expressing mutant MLC-2 protein with disrupted phosphorylation sites ...... 89 Rescue of the dominant flightless phenotype of Mlc2E38 heterozygote with a transformed copy of the mutant Mlc2 gene ...... 92 Rescue of the recessive lethality of Mlc2E38 homozygotes with a transformed copy of the mutant MLc2 gene ...... 93

viii Rescue of the flightless phenotype of the Mlc2E38 homozygotes with a transformed copy of the mutant Mlc2 gene ...... 96 Examination of the indirect flight muscle ultrastructure of transformants expressing one copy of the mutant Mlc2 genes ...... 97 Rescue of flight ability of Mlc2E38 homozygotes with two transformed copies of the mutant Mlc2 gene ...... 107 Examination of the indirect flight muscle ultrastructure of transformants expressing two copies of the mutant Mlc2 genes ...... 110 Examination of the MLC-2 protein pattern in transformants carrying the mutant Mlc2 genes ...... 119 Wingbeat analyses of transformants expressing m utant MLC-2 protein ...... 122 Mechanical analyses of skinned single IFM fibers expressing m utant MLC-2 protein ...... 124 Discussion ...... 128

V. CLONING OF A DROSOPHILA MYOSIN LIGHT CHAIN KINASE GENE...... 133

Introduction ...... 133 Results...... 144 Amplification of Drosophila genomic DNA by Polymerase Chain Reaction ...... 144 Isolation of MLCK cDNA clones using the PCR product as a hybridization probe ...... 147

ix Analysis of the nucleotide and protein sequences of the Drosophila MLCK clones ...... 159 Analysis of the Drosophila MLCK functional domain ...... 162 Catalytic domain ...... 162 Regulatory domain ...... 162 Preliminary characterization of the MLCK genomic organization ...... 170 Expression of the MLCK transcripts...... 171 Discussion ...... 174

REFERENCES...... 177

x LIST OF TABLES

TABLES PAGE

1. Myosin light chain nomenclature ...... 68

2. Rescue of the dominant flightless phenotype of Mlc2E38 heterozygotes with a transformed copy of the mutant Mlc2 gene ...... 91

3. Rescue of the dominant flightless phenotype and viability of Mlc2E38 homozygotes with a transformed copy of the mutant Mlc2 gene ...... 95

4. Rescue of the dominant flightless phenotype and viability of Mlc2E38 homozygotes with two transformed copies of the mutant Mlc2 gene ...... 109

5. The effects of a mutant MLC-2 on wingbeat frequencies ...... 123

xl LIST OF FIGURES

FIGURES PAGE

1. Structure of a myofiber ...... 5

2. Diagramatic representation of a of vertebrate skeletal muscle ...... 7

3. Diagramatic representation of the sarcomere during contraction ...... 13

4. Drosophila skeletal muscle system ...... 26

5. Protein sequence alignment of the Drosophila MLC-2 with MLC-2 sequences of other species ...... 32

6. Construction of the w; Mlc2E38,e/TM3 stock by recombination ...... 43

7. Restriction endonuclease maps of the Mlc2 transformation vectors pCasJWl and P[Mlc2S66.67A]...... 46

xii 8. Crosses to rescue the dominant flightless phenotype and the recessive lethality of the Mlc2E38 allele ...... 50

9. Mobilization of the transformed Mlc2S66<67A gene and selection of non-third chromosome inserts ...... 52

10. Purification of the Drosophila MLC-2 fusion protein 58

11. Accumulation pattern of MLC-2 variants in vivo 70

12. Expression of Drosophila MLC-2 in E. colt...... 76

13. Electron micrographs of the IFM of transformants carrying one copy of the wild type Mlc2 gene (P[Mic2+]/+; Mlc2E3S/MIc2e38, line JW1[63.1])...... 99

14. Electron micrographs of the IFM of transformants carrying one copy of the wild type Mlc2 gene (P[MIc2+]/+; Mlc2E38/Mlc2E38, line J W1[40.1J)...... 101

15. Electron micrographs of the IFM of transformants carrying one copy of p[Mlc2S66<67A] {P[Mlc2S66,67A]/P[Mlc2S66,67A] . Mlc2E38/Mlc2E88, line R3T1)...... 103

16. Electron micrographs of the IFM of transformants carrying one copy of p( Mic2S66’67A] (P[ MZc2 S66.67AJ /p( MIc2 S66,67A^ . Mlc2E38/Mlc2E38, line R3T4)...... 105

xiii 17. Electron micrographs of the IFM of transformants, carrying two copies of p[Mlc2S66'67A] (P[Mlc2S66.67A]/P[Mlc2S66,67A\ . Mlc2E38/Mlc2E38, line R3T4)...... 112

18. Electron micrographs of the IFM of transformants carrying two copies of the transformed wild type Mlc2 genes (P[Mlc2+]/P[Mlc2+); Mlc2E88fMlc2E38 line JW 1[40.1])...... 114

19. Electron micrographs of the IFM of transformants carrying two copies of p[Mlc2S66A) (P[Mlc2888A)/P[Mlc2888A] ; Mlc2 E88/M lc 2888 line R1T3)...... 116

20. Electron micrographs of the IFM of transformants carrying two copies of p[Mlc2S67A] {P[Mlc2887A]/P[Mlc2887A) ; Mlc2E38/Mlc2E88 line R2T2)...... 118

21. Pattern of MLC-2 accumulation in transformants Mlc2+ homozygotes and Mlc2S66’67A homozygotes ...... 121

22. Sinusoidal analysis of single IFM fibers from transformant homozygotes ...... 126

23. General scheme for myosin phosphorylation in skeletal and smooth muscles ...... 136

xtv 24. Hypothetical model for regulation of myosin light chain kinase activity by pseudosubstrate sequence and Ca2+/ ...... 139

25. Organization of the structural domain of myosin light chain kinase ...... 141

26. Amplification of a Drosophila MLCK genomic sequence by Polymerase Chain Reaction (PCR) 146

27. Diagram of sequence organization of the Drosophila MLCK cDNA clones ...... 149

28. Nucleotide sequence alignment of the Drosophila MLCK cDNA clones ...... 151

29. The deduced protein sequence alignment of the Drosophila MLCK with MLCK of other species ...... 164

30. Developmental expression pattern of Drosophila MLCK transcripts ...... 173

xv CHAPTER I

INTRODUCTION

Motility is an important characteristic of eukaryotes. It includes intracellular movements ranging from cytoplasmic organelle movement, cell division, and the amoeboid movement of unicellular organisms to the muscle contraction of multicellular organisms. These processes are mediated by the same basic mechanisms, utilizing similar contractile machinery. All forms of motility occur by transformation of chemical energy into mechanical energy, mediated by motor proteins. To date, three major classes of molecular motors have been identified; myosin, dynein, and kinesin (Vale and Goldstein, 1990). Myosin has both structural and enzymatic roles in muscle contraction and intracellular motility. The role of myosin in movement has been most extensively studied in vertebrate skeletal muscles. In muscles, myosin and actin form two distinct sets of contractile filaments (thick and thin filaments) that slide relative to each other during contraction (Huxley, 1969). Contraction in all muscles is activated by Ca2+, which acts through either thick or thin filament-associated proteins. In some muscle types, Ca2+ mediates contraction through both thick and thin filament- associated proteins (Lehman and Szent-Gyorgyi, 1975). Although much effort has been invested in studying muscle , precisely how myosin and actin interact to produce contraction, and how this process is regulated is not fully understood.

1 Molecular-genetic techniques are being used to study the regulation of muscle contraction. In organisms with powerful genetic systems such as Drosophila melanogaster, mutations affecting muscle function can be analyzed, and the role of the affected myofibrillar proteins elucidated. Moreover, genes can be inserted back into the fly genome using P element mediated germline transformation, permitting one to assess the consequence of expressing altered geiie products in vivo. One Drosophila muscle type, the indirect flight muscle (IFM), has been used extensively to assess the effects of mutations in muscle genes. The IFM is dispensable for viability and fertility of the fly but is required for flight. Thus, mutations that effect the IFM often give a flightless phenotype. These muscles also have the advantage that they are large, have a highly ordered structure and can be analyzed mechanically in vitro. My research focuses on the mechanism of contraction using Drosophila IFM as a model system and a combination of genetic, biochemical, cytological, and molecular techniques. Specifically, I am interested in determining the role of myosin regulatory light chain phosphorylation in regulating muscle contraction.

Muscle structure

Three principle types of muscle have been classified based on the vertebrate muscle systems: skeletal muscle, smooth muscle, and . The three muscle types differ from one another in their morphology and physiology and these correlate with the specialized contractile function of each muscle type.

Skeletal muscle Skeletal muscle, or striated muscle, is under voluntary control, and capable of rapid contraction. Skeletal muscles are composed of bundles of myofibers, which are multi-nucleated cells formed by fusion of many myoblasts during development (Figure 1). Each myofiber is surrounded by the . The nuclei lie beneath the sarcolemma. Two-thirds of the myofiber dry mass is derived from the which are the contractile elements of the myofiber and run in parallel along the length of myofiber. Two membranous systems, the T tubules (transverse tubules) and the sarcoplasmic recticulum, surround each and function to relay the signal for contraction. After a nerve impulse, Ca2+ is released from the sarcoplasmic recticulum, activating muscle contraction. When observed by phase contrast light microscopy, individual myofibrils show an alternate light and dark banding pattern, which repeats along the entire length of the fibrils (Figure 1). The banding pattern of adjacent myofibrils that run in parallel along the length of the myofibers are aligned in register, giving the muscle a characteristic "striated" appearance. When examined by electron microscopy, the myofibrils are divided into repeating units called the (~ 2.3-2.8 pm in vertebrates, and up to >10 pm in some invertebrates) (Figure 2), each of which is delimited by Z bands or lines (Harrington, 1979). Within the sarcomeres, there are two distinct sets of myofilaments running in parallel to the longitudinal axis of the myofibrils: thick filaments and thin filaments (Huxley, 1953). These two sets of filaments alternate and partially overlap one another along the entire length of the myofibrils, giving rise to the alternate dark and light banding pattern of striated muscle. Thick filaments extend across the dark region (A-band, -1.6 pm in vertebrates to 6-8 pm in some invertebrate muscles), whereas thin filaments extend across the light region (I-band -0.7-1.2 pm in vertebrates, and as short as 0.1-0.2 pm in some indirect flight muscles), and partially into the A-band (Figure 2). The thin filaments terminate within Figure 1. Structure of a myofiber. Each myofiber is a multinucleated cell, which is composed mainly of myofibrils, the contractile component of muscles. Within the myofiber, T tubules and sarcoplasmic recticulum are the two membranous systems that wrap around each myofibril in order to relay the contraction signals from the nervous systems to the myofibrils. Each myofibrils contains repeating contractile units, the sarcomeres, which are made up of overlapping arrays of thick and thin filaments. Contraction occurs when these two groups of filaments slide pass each other, resulting in shortening of the sarcomeres (taken from Villee et al., 1985).

4 5

•Nucleus of muscie ecu

Schwann ceil of axon sneam

Nerve ceil e n a m a

■ucteus a t Schw ann ceil

Schwann con of motor cna onto

Mvoneurai (unctions Hillock of motor «na otate

N jcieus of muscie ceil

Thick filaments*

T tubule

Sarcooiasmc rvtm lum Mitocnonanon

Nucleus of muse* ceil-

Figure 1. Figure 2. Diagramatic representation of a sarcomere of vertebrate skeletal muscle. A sarcomere is delimited at both ends by Z lines. Thick filaments run the length of the A band; thin filaments run through the I band and overlap the thick filaments within the A band. The H band, central to the A band, is the area containing thick filaments only (taken from Villee et al., 1985).

6 7

Sarcom ere

A band I band

H zone A Z line>

\ Thin filamentj (actinl C ro n bridges Thick filaments (myosin)

)))))))))»«

Figure 2. the A-band at the edge of the H-zone, a region of ~3800 °A located central to the A-band, which appears lighter. Thick filaments of opposite polarity are joined together in the middle of the H-zone at the M-line (Huddart, 1975). Mitochondria are located on both sides of the myofibrils along the entire length and provide ATP uniformly along the myofibrils (Amos, 1985).

Smooth muscle Smooth muscles are under involuntary control, and are responsible for the slow and sustained contractions such as those that occur in the stomach, uterus, artery walls, and ducts of secretory glands. These muscles are composed of elongated spindle-shaped cells, each with a single large nucleus. Smooth muscle cells are irregularly organized; thus, no striation is observed. Within each cell, thick and thin filaments are found, but they are not organized into an ordered pattern and do not form myofibrils. Instead, they form loosely organized contractile filaments which are oriented obliquely to the longitudinal axis of the cell via attachment to dense bodies in the plasma membrane. Unlike in skeletal muscles, the sarcoplasmic recticulum and the T tubules are poorly developed or absent. This feature of smooth muscle correlates with the finding that a larger portion of the Ca2+ required for activating contraction comes from the extracellular medium, rather than the sarcoplasmic recticulum as in skeletal muscle (Daniel et al., 1985). Contraction of smooth muscle (as well as heart muscle) can be modified by a number of hormones that change the intracellular concentration of either cyclic nucleotides or Ca2+, via phosphorylation of the sarcoplasmic recticulum or the sarcolemma proteins by cyclic nucleotide or Ca2+-activated protein kinases. Phosphorylation of proteins associated with the sarcoplasmic recticulum or the sarcolemma presumably results in an alteration of ion fluxes into and out of the muscles.

Cardiac muscle Cardiac muscle has features of both skeletal and smooth muscle. It is under involuntary control like smooth muscle. However, like striated muscle, thick and thin filaments organize to form myofibrils, but do not give a striated pattern because the myofibrils often branch and adjacent myofibrils do not align in register. Heart muscle cells contain a single nucleus located centrally. Each cell joins with others through a special structure called the , which functions in synchronizing contraction in heart muscle. Heart muscle is also stimulated by Ca2+ which is released from the sarcoplasmic recticulum, although a certain amount comes from the extracellular medium (Price and Sanger, 1983).

Composition of myofilaments

The thick filaments The major component of thick filaments is myosin, a hexameric protein that is composed of one pair of myosin heavy chains (MHC, ~200 kD) and two pairs of each of the chemically distinct light chains. One class of light chain is termed myosin essential light chain (ELC) or myosin alkali light chain (see Table 1). The ELC is, in fact, composed of two proteins of -25 kD and -16 kD, which are termed myosin light chain-1 (MLC-1) and myosin light chain-3 (MLC-3), respectively (Chapter III, Table 1). The other class of light chain is -20 kD and is termed myosin regulatory light chain (RLC) or myosin light chain-2 (MLC-2) (Chapter III, Table 1). The C terminal part of two MHC molecules forms an a-helical coiled-coil rod that is responsible for the assembly of myosin into thick filaments (Figure 3). The myosin rods aggregate in an anti-parallel fashion, giving rise to a bipolar structure with a central bare region. The N-terminal part of each MHC forms a globular head that contains the actin binding site and the ATP hydrolysis site. The light chains bind non covalently to the MHC head.

Myosin heavy chain (MHC) Myosin heavy chain (MHC) is a highly asymmetric molecule. The N-terminal half of MHC forms a globular head, whereas the C- terminal half forms a long a-helical tail (Figure 3). Proteolytic enzymes divide myosin heavy chain (MHC) into two main fragments: the N-terminal HMM (heavy meromyosin, -150 kD) and the C-terminal LMM (light meromyosin, -70 kD) (Harrington, 1979; Harrington and Rodgers, 1984; Warrick and Spudich, 1987). The HMM fragment can be further subdivided into a SI subfragment (-95 kD), corresponding to the myosin head, and a S2 subfragment (55 kD), corresponding to the N-terminal portion of the tail. The LMM has been shown to contain a specific region necessary for filament formation (the assembly subdomain), as well as sites of phosphorylation shown to be important for regulating filament assembly in nonmuscle cells (Tan et al., 1992). The myosin rods aggregate in an anti-parallel fashion, giving rise to a bipolar structure with a central bare region. The globular head (SI) of MHC contains the actin binding site and the ATP hydrolysis site. The SI subfragment has been shown to be sufficient to support movement of actin filaments on a nitrocellulose-coated surface in vitro (Toyoshima et al., 1987), demonstrating that the minimum requirements for producing motility are contained within the myosin head. 11

Myosin regulatory light chain (MLC-2) The myosin regulatory light chain (RLC; hereafter called myosin light chain 2, MLC-2) was found to be involved in regulation in various muscle types. MLC-2 is a Ca2+-binding protein, which shares a common ancestor with myosin essential light chain (ELC), , parvalbumin, and calmodulin. It has been proposed that the prototype protein contained four Ca2+- binding sites, which arose by duplication and reduplication of a gene encoding a protein with a single Ca2+-binding site (Collins, 1991). The four Ca2+-binding sites are designated 1 through IV. Some genes that are related to the ancestral prototype encode proteins that have lost their Ca2+-binding ability at some or all of these sites. However, the sites in these proteins have maintained a similar primary and tertiary structure (Collins, 1976). MLC-2 appears to have retained one functional Ca2+-binding site in region I, which is a low-affinity Ca2+-Mg2+ site (Bagshaw and Kendrick- Jones, 1979). Whether direct Ca2+-binding to MLC-2 occurs in vivo, and whether this binding is involved in regulation of contraction in most muscle types is not known. In scallop muscles, direct binding of Ca2+ to myosin does occur. In the absence of Ca2+, MLC-2 represses the actomyosin interaction, but this inhibition is released in the presence of Ca2+ (Szent-Gyorgyi et al., 1973). When MLC-2 is selectively removed, the myosin ATPase activity is constitutively activated without a requirement for Ca2+. It has been shown that MLC-2 regulates the Ca2+-dependent activation of the actomyosin ATPase activity by forming a "regulatory domain" with MPIC and ELC, and inducing a specific Ca2+-binding site on the ELC. Only when this site is bound by Ca2+ does MLC-2 release the inhibition of myosin MgATPase activity (Kwon et al., 1992). Therefore, the role of MLC- 2 in regulating the actomyosin ATPase is to induce a specific Ca2+- binding site on the ELC, rather than binding Ca2+ directly. Figure 3. Diagramatic representation of the sarcomere during contraction. (a) The sarcomere is a functional unit of muscle, consisting of two types of contractile filaments: the thick filament and the thin filament. The thick and thin filaments slide past each other during corssbridge cycling, resulting in the thin filaments being pulled towards the center of the sarcomere. The thin filament (upper right) is composed of actin, tropomyosin and the troponin complex. The thick filament is composed primarily of myosin. (b) Diagram of myosin structure. Myosin is a hexamer, consisting of a pair of myosin heavy chains and two pairs of myosin light chains (myosin regulatory light chain and myosin essential light chain). Myosin heavy chain consists of two proteolytic fragments, HMM (heavy meromyosin) and LMM (light meromyosin). The HMM can be further subdivided into two fragments: SI (subfragment 1) and S2 (subfragment 2). The SI fragment (corresponding to the globular head of the myosin heavy chain) contains the binding sites for actin, ATP and myosin light chains (taken from Sellers and Adelstein, 1987).

12 13

Tropomyosin A ctin

\ / \ £ T. / \ Troponin Complex 1C ' \ / \ / T X 3 X U -U •in m m m m Xs-» T h in F ila m e n t Thick Filament u z L in e

n ■le »!< > -t - C T x x x x x n "J

•LMM- -HMM-

C a — filament formation ATPase and actin binding

Figure 3. 14

Although MLC-2 from most other species is phosphorylated by the myosin light chain kinase (MLCK), the scallop MLC-2 is not phosphorylated. Although Ser 6 of the scallop protein can be aligned with the MLCK phosphorylation sites in other species, it lacks clusters of basic residues N-terminal to the putative phosphoiylatable serine shown to be required for MLCK phosphorylation (see review by Stull et al., 1986). In contrast to molluscan , vertebrate skeletal muscles are under thin filament control, and MLC-2 appears to play a minor role in regulating skeletal muscle function. Removal of MLC-2 does not abolish the regulation of the myosin MgATPase activity. Nevertheless, it has been observed that there is a decrease in the interaction of myosin with actin, and reduced sensitivity of troponin C to Ca2+ when myosin was depleted of MLC-2 (Pemrick, 1977). It appears that MLC-2 stabilizes a particular conformation of myosin which enhances actin interaction by increasing the Ca2+ sensitivity of troponin C. Skeletal muscle MLC-2 is phosphorylated by MLCK at specific serine residues (see review by Stull et al., 1986). However, MLC-2 phosphorylation is not required for skeletal muscle contraction, although it has been shown to modify the steady state kinetics of the actomyosin ATPase under certain conditions (Pemrick, 1980). The role of skeletal muscle MLC-2 in regulating myosin ATPase activity has been tested using a scallop/vertebrate myosin hybrid system, in which the scallop MLC-2 is replaced with vertebrate skeletal muscle MLC-2. The vertebrate skeletal muscle MLC-2 was shown to be incapable of inducing a specific Ca2+- binding site in the scallop myosin, resulting in depression of the myosin ATPase activity whether or not Ca2+ was present (Sellers et al., 1980). Therefore, it seems that if Ca2+ binds to the thick filament component(s) of vertebrate skeletal muscle in vivo, this binding does not play a role in regulating contraction. 15

In smooth muscle and nonmuscle cells, phosphorylation of MLC-2 is required for the activation of contraction (see review by England, 1980). It was shown that activation of smooth muscle myosin MgATPase activity in vitro was proportional to the degree of MLC-2 phosphorylation (Small and Sobieszek, 1977), and that the activation is Ca2+ dependent. MLC-2 phosphorylation has also been correlated with contraction in vivo (Driska and Murphy, 1979). Frozen muscles either in the contracted or relaxed state were analyzed by two-dimensional gel electrophoresis, and the amount of phosphorylated MLC-2 in the two tissue samples was compared. Muscles in contracted state appeared to have more phosphorylated MLC-2 than relaxed muscles. In studies using scallop/vertebrate smooth muscle hybrid myosin, smooth muscle MLC-2, like the native (scallop) MLC-2, is capable of inducing a specific Ca2+-binding site on scallop myosin (in combination with MHC and ELC). The hybrid myosin exhibits a similar Ca2+ sensitivity to that of the native scallop myosin (Kwon et al., 1992). However, it is unlikely that activation by Ca2+ binding to thick filaments occurs in vivo in vertebrates since it has been shown that gizzard myosin cannot be activated by Ca2+, or bind Ca2+, due to the inability of gizzard ELC to form a specific Ca2+-binding site.

Myosin essential light chain Myosin essential light chain (ELC) is also a member of troponin C superfamily, and contains one functional Ca2+-binding site in region III. The name "the essential light chain" derives from the finding that alkaline conditions used for the initial isolation of ELC resulted in a loss of the myosin ATPase activity. However, ELC was later shown not to be essential for myosin function, since enzymatically active myosin devoid of ELC was obtained with a different isolation procedure (Wagner and Giniger, 1981). It has 16

also been called "the alkali light chain" because alkaline conditions were initially used for its extraction (see review by Harrington, 1979). Although the role of ELC has not been well characterized in most muscle systems, varying isoforms of ELC have been correlated with different maximal velocities of shortening and ATPase activates in skeletal and smooth muscles (see review in Sweeny et al., 1993). The ELC of Physamm has been shown to be important for regulating the myosin ATPase activity. Unlike other ELCs, the Physam m protein contains two functional Ca2+-binding sites. When Physam m ELC is bound by Ca2+, activation of the myosin ATPase activity is inhibited (Koyama and Kendrick-Jones, 1986). This inhibition is released when Ca2+ is released from the ELC.

The hinge region The junction between the SI and the S2 fragments is thought to be a flexible area ("the hinge region"), which allows rotation of the myosin head against the actin filament during muscle contraction (Figure 3). It was thought that the two myosin light chains bind close to this region (see review by Harrington, 1979), and this has been confirmed recently by analyzing the three dimensional structure of the myosin subfragments (Rayment et al., 1993a; Xie et al., 1994). The three dimensional structure of the SI subfragment from chicken skeletal muscle myosin, and the regulatory domain (corresponding to the C-terminal end of SI) from scallop myosin has been determined (Rayment et al., 1993a, Xie et al., 1994). Although the two myosins are regulated by different mechanisms (chicken skeletal muscle myosin is regulated by phosphorylation of the MLC-2, whereas scallop myosin is regulated by Ca2+ binding to the regulatory domain, formed by MHC, MLC-2 and ELC), they share a number of features with regard to light chain binding. In both myosins, the MLC-2 is located at the C-terminal end of the SI head, close to the hinge region. The MLC-2 assumes a "dumbbell-shape", and wraps around the MHC through the N- and C-terminal globular domains. It has been proposed that a conformational change of the MLC-2, induced either by phosphorylation of the MLC-2 (chicken skeletal muscle) or Ca2+ binding to the regulatory domain (scallop), is transduced through the hinge region (as well as other flexible regions in the myosin head) to the motor domain (SI), resulting in activation of MgATPase activity. This region could also function as a lever arm to amplify the conformational changes in the nucleotide- and actin-binding sites of the SI head during crossbridge cycling.

The thin filam ents Thin filaments of vertebrate skeletal muscles are composed of actin, tropomyosin, and the troponin complex (composed of troponin C, , and ) at a molar ratio of 7:1:1 (Figure 3). Smooth muscles and nonmuscle cells appear to lack troponin C.

Actin Actin is the main constituent of the thin filament (up to ~ 60%). The actin monomer (G-actin) is a globular protein of -42 kD, with one Mg2+/Ca2+ binding site and one ATP binding site. When ionic strength is increased, G-actin monomers polymerize to form filamentous actin (F-actin), concomitant with hydrolysis of an ATP molecule. Although ATP hydrolysis is not necessary for actin polymerization, it is important for the dynamic behavior of actin filaments. Two actin filaments twist around each other in a tight helix to form a thin filament backbone. Tropomyosin Tropomyosin is a -410 °A-long rod-shaped molecule with a MW of -70 kD that lies in the grooves between two actin strands, covering about seven actin monomers. Two tropomyosin molecules form an a-helical coiled-coil structure, and adjacent tropomyosin molecules polymerize by overlapping in a head-to-tail fashion along the actin filament backbone. Models have been proposed to explain the role of tropomyosin in regulation of muscle contraction. One model is called the steric blocking model. According to this model, tropomyosin inhibits actin-myosin interaction in the absence of Ca2+ by steric hindrance. In the presence of Ca2+, tropomyosin shifts its position relative to the actin filament, allowing the myosin head to interact with actin. This model predicts that in the absence of tropomyosin, the muscles would be constitutively activated. However, recent genetic and biophysical analysis of tropomyosin in the Drosophila IFM is not consistent with this idea (Kreuz, 1993), and suggests that tropomyosin is more likely involved in the control of crossbridge kinetics rather than steric blocking of crossbridge formation.

The troponin complex The troponin complex is composed of three noncovalently linked subunits: troponin C (Ca2+-binding), troponin T (tropomyosin-binding), and troponin I (inhibitory) (Figure 3). The troponin complex has an elongated shape, with troponin C and troponin I forming a globular head, and troponin T forming a long tail. The tail of troponin T binds to tropomyosin and is thought to position the complex on the thin filament. Troponin I binds to actin, and functions in concert with tropomyosin and troponin T to inhibit actin-myosin interaction. Troponin C confers Ca2+ sensitivity to the system. It has four Ca2+-binding sites, two high affinity sites and two low affinity sites. Ca2+-sensitivity is possibly regulated by the low-affinity sites, and full actomyosin ATPase is only reached when all four Ca2+-binding sites are occupied. It is thought that Ca2+ binding to troponin C results in a structural change that causes a stronger troponin subunit interaction and a weaker troponin I-actin link. As a consequence, tropomyosin moves laterally toward the center of the actin groove, exposing the myosin binding site on the actin filament (Harrington, 1979).

Other accessory proteins Many sarcomeric proteins do not participate directly in the contractile activity, but provide structural integrity and elasticity to the sarcomere to facilitate contraction.

Alpha-actinin The sarcomeres of vertebrate skeletal muscles are demarcated by Z lines which serve as links for the transmission of tension along the myofibril and also act as structural anchors to maintain the actin filaments in longitudinal and lateral order. The major component of the Z line is a-actinin, which is also present in dense bodies of smooth muscles and adherence-type junctions in nonmuscle cells, as well as other intracellular structures where actin is anchored (see review by Blanchard et al., 1989). a-actinin is a homodimer of two antiparallel subunits, which bind actin. It is thought that a-actinin is involved in anchoring the actin filaments in the Z line, but precisely how this occurs is not known.

Titin A giant protein, "titin" or "connectin'1, has been found in vertebrate skeletal muscle, connecting the thick filaments from the M line to the Z bands (Maruyama, 1986). Titin is very elastic, and is believed to be responsible for passive tension generation and for positioning of the thick filaments in the center of the sarcomere between the Z lines.

Muscle contraction

The sliding filament model of muscle contraction H.E. Huxley and Hanson (1954) and A.F. Huxley and Niedergerke (1954) independently proposed that muscle contraction occurs by the sliding of the thick and thin filaments relative to each other, resulting in the thin filaments being pulled toward the center of the thick filament and, thus, in the shortening of the sarcomeres. This model is based on the observation that during contraction, the sarcomeres shortened, reducing the length of the I-band and the H-zone without changing the width of the A-band and the distance between the Z line and the edge of the H zone (Figure 2). This observation suggests that the lengths of the thick and thin filaments do not change during contraction, but the filaments slide past each other resulting in a greater overlap. Subsequently, Huxley (1957) observed that the thick filament had numerous projections that interact with adjacent thin filaments, leading him to propose that these projections (or crossbridges) interact cyclically with the thin filaments, causing the thick and thin filaments to slide past each other. Although this model was developed based on studies in skeletal muscle, it is believed to be the basis of contraction in smooth muscle and in many cellular activities in nonmuscle cells. Recently, the three-dimensional structure of the SI myosin subfragment and the SI-actin complex has been resolved at 2.8 °A (Rayment et al., 1993a; Rayment et al., 1993b; Xie et al., 1994). A model describing the mechanism of crossbridge cycling has been proposed which is based on the sliding filament model and incorporates information from the recently determined 3-D structure and other structural, biochemical and biophysical data. In this model, the ATP binds in an ATP binding pocket, which is located on the opposite side from the actin binding site of the myosin head. ATP binding causes the actin binding site to open, releasing the myosin head from the actin filament. The ATP binding occurs in two steps, initial "weak binding" which results in closure of the ATP binding pocket to produce "strong binding" which results in a conformational change in myosin and a movement of the MHC C-terminus at least 50 °A relative to the actin binding site. The myosin head then reattaches to the actin filament by a multi-step process, concomitant with ATP hydrolysis. Axtin catalyses the release of the hydrolysis product, Pi, and the myosin undergoes another conformational change which initiates the power stroke. This conformational change of myosin causes reopening of the ATP binding pocket, releasing the ADP, and returning the myosin to the original state.

Regulation of m uscle contraction

Contraction in all muscles is initiated by a rise in the intracellular Ca2+ concentration from ~10'7 to ~10'5 M, resulting in an interaction between actin and the myosin head. Myosins, which possess an actin-activated MgATPase activity, hydrolyze ATP molecules to provide the energy for the cyclical interaction of the myosin heads with the actin filaments. Regulation by Ca2+ is classified into two types based on the proteins that Ca2+ interacts with: 1) thin filament-linked regulation, 2) thick filament-linked regulation.

Thin filament control The steric blocking model has been proposed to explain thin filament control of muscle contraction (Haselgrove and Huxley, 22

1973). According to the model, in the absence of Ca2+ (<10-7 M) tropomyosin physically blocks the actin-myosin interaction. Upon an intracellular rise of Ca2+ concentration (~10~5 M), troponin C of the thin filament binds four Ca2+ ions, resulting in a conformational change which leads to the shift of tropomyosin on the actin filaments. This exposes the myosin binding site on the actin filaments, allowing actin to interact with the myosin head. Although the steric blocking model has been widely accepted, recent biochemical and biophysical evidence indicates this model may be too simple to explain how muscle contraction is regulated. Recent genetic, biophysical and biochemical data suggests that contraction is regulated not only by physical blocking/deblocking of the actin-myosin interaction by tropomyosin, but also by processes involving kinetics of actin-myosin interaction, which is influenced by binding of ATP and the intermediates of ATP hydrolysis (El Saleh et al., 1986; Kreuz, 1993). Therefore, regulation of muscle contraction appears to be influenced not only by steric effects, but also by kinetic factors.

Thick filament control Thick filament regulation can occur in two ways: 1) direct binding of Ca2+ to myosin, 2) activation of the myosin light chain kinase by a Ca2+/calmodulin complex, leading to phosphorylation of MLC-2. In both types of regulation MLC-2 represses actin- myosin interactions in the absence of Ca2+. In the scallop adductor muscle, Ca2+ activates contraction by binding directly to a specific Ca2+ binding site on the myosin head, formed by MHC, ELC, and MLC-2 (Kwon et al., 1992). The binding results in a conformational change of MLC-2 and a release of the inhibition of the actin-myosin interaction. However, in vertebrate smooth muscle and nonmuscle cells Ca2+ mediates contraction indirectly by forming a complex of four Ca2+ ions with calmodulin, which 23 activates the MLCK to phosphorylate MLC-2 (see review by Kamm and Stull, 1985). Phosphorylation of MLC-2 occurs at specific serine or threonine residues, and has been shown to be a requirement for the initiation of contraction in smooth and nonmuscle cells. Regulation by both thick and thin filaments can operate in a single muscle type. Although it is known that vertebrate skeletal muscle is under thin filament-linked regulation, phosphorylation of MLC-2 has been shown to modulate the kinetics of contraction under certain conditions (Pemrick, 1980). Similarly, other regulatory mechanisms besides MLC-2 phosphorylation by MLCK have been proposed to occur in smooth muscles. These include thin filament-linked regulation associated with leiotonin (Nomura and Ebashi, 1980) and caldesmon (Ngai and Walsh, 1984), and direct Ca2+ binding to myosin (Chacko et al., 1977; Rees and Frederiksen, 1981). Invertebrate muscles, particularly those of the , have been shown to possess Ca2+-sensitive regulation associated with both thick and thin filaments (Lehman and Szent- Gyorgyi, 1975).

Regulation by phosphorylation of myosin light chain 2 The role of MLC-2 phosphorylation in regulating muscle contraction has been most extensively studied in vertebrate smooth and nonmuscle cells, in which phosphorylation of MLC-2 is required for contraction (see review by Sellers and Adelstein, 1987; Kamm and Stull, 1985). Phosphorylation of smooth muscle MLC-2 was first discovered by Frearson et al. (1976), and was subsequently shown to be associated with an increase in the actin- activated MgATPase of myosin (Sobieszek, 1977). Phosphorylation of MLC-2 results in a dramatic (50-100 fold) increase in the actin- activated ATPase activity of smooth muscle, whereas only a two­ fold increase is observed in skeletal muscle. Phosphorylation of MLC-2 occurs in several steps, and is catalyzed by myosin light chain kinase (MLCK). Upon an intracellular rise of Ca2+ concentration, four Ca2+ ions form a complex with calmodulin, which activates the MLCK to phosphoiylate the MLC-2 at specific serine and/or threonine residues. This results in a conformational change of the myosin molecule, allowing interaction between actin and myosin and activation of myosin MgATPase activity to occur (Sweeney et al., 1993). Phosphorylation of MLC-2 has been correlated with myosin MgATPase activity in vitro, and contraction both in vitro and in vivo (see review by Sellers and Adelstein, 1987; Kamm and Stull, 1985). Relaxation occurs when the sarcoplasmic Ca2+ concentration returns to resting levels, causing Ca2+ to dissociate from calmodulin, resulting in an inactivation of the MLCK.

Drosophila musculature

I have used Drosophila musculature as a model for studying the role of MLC-2 phosphorylation in muscle contraction. It is a useful model system because the morphology and the protein components of Drosophila muscles resemble those of other animals, including vertebrates. All Drosophila muscles are striated but are divided into three distinct classes: 1) the supercontractile muscles of the larva and the adult, 2) the tubular muscles of the adult and 3) the fibrillar muscles of the adult.

Supercontractile muscle The supercontractile muscles are found in the larval body wall, adult abdomen and viscera, including the heart. These muscles are able to contract up to 80% of their resting length due to perforation of the Z lines that allows thick and thin filaments from adjacent sarcomeres to slide pass one another during Figure 4. Drosophila skeletal muscle system. (A) Sagittal section of an adult Drosophila female. The dorsal longitudinal muscles (DLMs) of the indirect flight muscle (IFM) are numbered 45 a-f. (B) Sagittal section of adult Drosophila thorax showing the location of the dorsal ventral muscles (DVMs) of the IFM and the tergal depressor of the trochanter muscle (TDT). The DVMs are numbered 46 a-b, 47 a-c and 48 a-b. The TDT is numbered 66. (C) Sagittal section of the adult Drosophila thorax showing the direct flight muscles (numbered 49-58). The direct flight muscles are attached to the base of the wings and serve to control the angle and position of the wings (taken from Miller, 1950).

25 26

PIA, M f i |

Figure 4. contraction (Crossley, 1978). These muscles contain a few large peripheral nuclei. The sarcomere length varies from 2.0-8.0 pm depending on the state of contraction. The ultrastructural organization of these muscles is rather poor compared to the other two muscle types. In cross sections, 8-12 thin filaments surround each thick filament, and in longitudinal sections the muscle structure appears diffuse due to poorly organized alignment of thick and thin filaments.

Tubular muscle Tubular muscles have a characteristic tube-like structure with axial nuclei. These muscles are found in the head, leg and direct flight muscles. The largest tubular muscle is the tergal depressor of the trochanter (TDT) which is responsible for jumping at the initiation of flight. Tubular muscles are not highly organized, with 10-12 thin filaments surrounding each thick filament. However, the Z lines of adjacent myofibrils align in register, giving a striated pattern. At rest, the I band occupies -50% of the sarcomere length, and during contraction, the muscle is capable of shortening up to 50% of the resting length. Tubular muscles are classified as "synchronous" muscles, i.e., muscles which contract in response to each nerve impulse. The sarcoplasmic recticulum is well developed in this muscle type, because contraction of the tubular muscle is strictly dependent on Ca2+ release from the sarcoplasmic recticulum. However, the mitochondria are not abundant, correlating with a low energy requirement in this muscle type which contracts at a lower frequency compared to the fibrillar muscle. In addition, delayed tension rise in response to stretch (stretch-activation) of tubular muscles is of smaller magnitude and is less well maintained compared to fibrillar muscle, resulting in a lower power output of the muscle fiber (Peckham et al., 1990). 28

Fibrillar muscles (IFM) The fibrillar muscles are only found in the indirect flight muscles (IFM) of the thorax (Figure 4). The IFM is termed "indirect" because it powers flight indirectly by causing a deformation of the thoracic exoskeleton to which the wings are attached, rather than by moving the wings directly. The muscle is capable of generating wingbeat frequencies up to 200 Hz or greater. The IFM constitutes up to 80% of the thoracic mass, and is composed of two groups of 26 muscle fibers lying perpendicular to each other; 12 dorsal longitudinal fibers and 14 dorsal ventral fibers. These muscles do not appear striated because Z lines of adjacent myofibrils are not in register, but the structure of each individual myofibril is remarkably well organized. In cross section, six thin filaments surround each thick filament in a rigid double hexagonal array. In longitudinal sections, the thick and thin filaments overlap nearly completely, resulting in a veiy short I band (0.1-0.2 pm). During contraction the sarcomeres shorten by only 1-2% of their resting length, facilitating a rapid oscillatory contraction. Numerous large mitochondria are found along both sides of the myofibril to provide energy for contraction. The sarcoplasmic recticulum is poorly developed in this muscle type, correlating with its "asynchronous" contractile property, i.e., the frequency of contraction is independent of the frequency of neural stimulation (Pringle, 1949). This is because the IFM is "stretch-activated". The muscle is partially activated by Ca2+, but full activation is achieved only after the muscle is stretched (Pringle, 1978). Once activated, one group of IFM fibers "stretch-activates" perpendicular fibers, leading to oscillatory contraction in which the two groups of muscles contract in response to stretch generated by the other. Thus, multiple contractions are produced as a result of a single neural impulse. Stretch activation is found not only in insect 29 fibrillar flight muscle, but is also a general mechanism of muscular contraction (Pringle, 1978). The effect of stretch-activation is more pronounced and stable in insect fibrillar flight muscles and cardiac muscles of vertebrates, but is quite small and transient in vertebrate skeletal muscles.

Muscle proteins of Drosophila

Drosophila muscle contains homologs of most muscle proteins, and because Drosophila is amenable to genetic analysis the characterization of phenotypes caused by mutations in muscle protein genes has made a substantial contribution to the understanding of muscle structure and function (Bernstein et al., 1993). Contraction is thought to be conferred in part by muscle proteins specific to each muscle type. For example, muscles with unique structural and physiological properties such as the IFM accumulate protein isoforms that are expressed exclusively in the IFM, including those of the myosin heavy chain, myosin alkali light chain, actin, and tropomyosin (see review by Bernstein et al., 1993). Besides contractile proteins, the IFM contains structural proteins such as Z disc proteins and proteins of the connecting filaments, as well as unique proteins such as Arthrin and Flightin whose functions are still unknown.

Myosin heavy chain The sarcomeric myosin heavy chain gene is encoded by a single gene at chromosome position 36B1.2. The gene contains 29 exons, 6 of which are alternately utilized to produce multiple MHC transcripts (George et al., 1989). These alternatively spliced exons encode tissue specific isoforms, which are thought to contribute to the specialized contractile activities of each muscle type. These correspond to parts of the SI region encoding the actin and myosin light chain binding sites and the ATPase activity. In addition, alternative exons also encode the S2 hinge region, which may be important in force generation. At least 10 MHC protein isoforms are produced during Drosophila muscle development. While some muscle types accumulate multiple isoforms, the IFM appears to accumulate only one isoform. The nonmuscle MHC gene is also a single copy gene and it is localized to chromosome position 60EF. At least two transcripts are produced by alternative splicing and accumulate in a stage specific manner (Kiehart et al., 1989).

Myosin light chain-2 The sarcomeric myosin light chain 2 (MLC-2 or myosin regulatory light chain, RLC) is encoded by a single copy gene at chromosome position 99E1-3, and is composed of 3 exons (Parker et al., 1985; Toffenetti et al., 1987). At least two different transcripts are generated by alternate transcription start sites and polyadenylation sites but there is no evidence for alternative splicing. The Mlc2 gene encodes a single protein with a predicted molecular weight of 24,000 kD and a pi of ~5, although the protein has an apparent molecular weight of 30 kD as judged by SDS- PAGE. Multiple MLC-2 protein variants accumulate in vivo, suggesting that they are products of post-translational modification of a single MLC-2 protein (Chapter III). Drosophila MLC-2 is similar to the MLC-2 of other species, except it contains an extra -50 amino acids at the N-terminus. The protein contains conserved myosin light chain kinase (MLCK) phosphorylation sites, which correspond to serine 66 and serine 67 (Figure 5). In addition, a single functional Ca2+ binding site has been identified, and shown to bind Ca2+ in vitro (Tanaka et al., 1988). A single null mutation, Mlc2E38, has been identified for the Mlc2 gene (Warmke et al., 1992). Mutant heterozygotes are Figure 5. Protein sequence alignment of the Drosophila MLC-2 with MLC-2 sequences of other species. The Drosophila MLC-2 sequence is aligned with the rabbit skeletal muscle MLC-2 (RSK) (Collins, 1976), the chicken smooth muscle MLC-2 (CSM) (Maita et al., 1981) and the chicken cardiac muscle MLC-2 (CCA) (Matsuda et al., 1981). Residues conserved in two or more sequences are shown in black boxes. Besides the -50 amino acid N-terminal extension that the vertebrate proteins lack, the Drosophila MLC-2 is conserved with the vertebrate counterparts throughout the entire protein sequence. Asterisks denote Ser66 and Ser67 which are conserved with the myosin light chain kinase (MLCK) phosphorylation sites of the vertebrates. Ser66, which is phosphorylated to a greater extent in vitro (Graham, 1992), is aligned with the major phosphorylation site of the vertebrate proteins.

31 . . . . 5 ...... 15 ...... 25 ...... 35 ...... 4 5 . . . DMRLC MADEKKKVKK KKTKEEGGTS ETASEAASEA ATPAPAATPA PAASATGjgKR RSK CSM -gSK CCA 65**r DMRLC ASGGSRGSRI RSK -PKKA0RR| CSM RKRPQRAK__ CCA P0KAK

DMRLC LG-AN RMATSBf RSK CSM H! - - n B CCA ROTPAA

DMRLC

DMRLC D| QI09TAA I EML RSK CSM K EF TRI GAKD CCA

Figure 5. 33 flightless. Mlc2E38 has an A to T transversion, resulting in an amber nonsense mutation at the tenth codon of the protein coding sequence. Thus, this gene encodes a severely truncated protein. This mutation also appears to effect RNA stability as Mlc2 mRNA accumulation is reduced by ~50% in the Mlc2E38 heterozygous flies (Warmke et al., 1992). The flightless phenotype of Mlc2E38 heterozygous mutants is due to abnormal IFM structure, which is probably the result of low MLC-2 accumulation. Interestingly, the structure of the TDT muscle (a tubular muscle) is only marginally affected in Mlc2E88 heterozygotes, and they have a normal jump function. The Mlc2E38 heterozygotes are viable and are otherwise normal, indicating that reduction of the MLC-2 protein by half presumably has no effects in most or all other muscle types of the fly. Mlc2E38 heterozygous flies exhibit an ~30% reduction in wing beat frequency, and in single fiber mechanical tests they show altered contractile kinetics (Warmke et al., 1992). Both of these factors contribute to the flightless behavior. Mlc2E38 homozygotes are embryonic lethal, demonstrating that MLC-2 is essential for viability and development of the fly. The cytoplasmic myosin light chain 2 gene is also encoded by a single gene which is located at chromosome position 5D5.6-5E. The gene is composed of 3 exons, but there is no evidence of alternative splicing. A single 1.2 kb message is transcribed and produces a single protein of 174 amino acids (Mr 19,954). spaghetti-squash (sqh) is a mutation in the cytoplasmic MLC-2 gene (Karess et al., 1991). sqh m utants produce only ~ 10% of the wild type level of MLC-2 transcripts, resulting in failure of cytokinesis in developing larvae. 34

Myosin alkali light chain The sarcomeric myosin alkali light chain (MLC-ALK or myosin essential light chain, ELC) is encoded by a single gene located at chromosome position 98B (Falkenthal et al., 1984). The gene is composed of 6 exons, which give rise to two classes of transcripts by inclusion or omission of exon 5. The transcript lacking exon 5 encodes a protein isoform which accumulates only in the adult IFM (Falkenthal et al., 1987). The transcript which includes exon 5 encodes the protein isoform found in all other developmental stages and tissues but it is not found in the adult IFM (Falkenthal et al., 1987). The Drosophila MLC-ALK protein has a single Ca2+-binding domain which has been shown to bind Ca2+ in vitro (Tanaka et al., 1988). But the role of the protein in muscle structure and/or function is not known.

Ac tin Drosophila actin is encoded by six genes, the sequences of which are highly conserved (Fyrberg et al., 1980). The six genes are dispersed in the genome, and named by their polytene location: Act5C, Act42A, Act57B, Act79B, Act87E, and Act88F. Three actin protein isoforms have been identified: actin I (pi 5.70), actin II (pi 5.77), and actin III (pi 5.84), all of which have a Mr of ~42 kD. Horovitch et al. (1979) have shown that actin I accumulates in supercontractile muscles, whereas actin III is present in thoracic flight muscles. Actin II is not present in muscles, and is most probably a cytoplasmic actin. The Act5C and Act42A encode cytoplasmic , whereas Act57A and Act87E are expressed predominantly in the embryonic and larval muscles. It is likely th at Act57A and Act87E are expressed in the adult abdominal muscle as well. Act79B expression is predominant in the adult tubular muscles, and Act88F is the only actin isoform accumulating in the IFM. 35

All actins are modified by N-terminal acetylation, which has been shown to be essential for actin polymerization (Henessey et al., 1991). In addition, actin is methylated at His 73, the role of which is still unknown. Act88F undergoes stable post- translational ubiquitination to give rise to arthrin, which accumulates exclusively in the IFM. Arthrin and actin are present at a ratio of 1:6, suggesting that there is one arthrin molecule per each thin filament cooperative unit (composed of seven actin monomers). Given that arthrin is specific to the IFM, and that arthrin also accumulates in Lethocerus flight muscles which are stretch-activated, it has been postulated that arthrin might play a role in stretch-activation of the IFM in Drosophila.

Tropomyosin Drosophila tropomyosin is encoded by two tandemly arranged genes (TM1 and TM2) located in the 88F region of the third chromosome (Karlik and Fyrberg, 1986; Hanke and Storti, 1988). The two genes produce a total of six tropomyosin isoforms. The TM1 gene comprises five exons, which produce two muscle-specific isoforms by alternative splicing of the transcripts. TM2 gene is composed of 17 exons, which produce 4 tropomyosin isoforms through a combination of alternative promoter utilization and transcript splicing. Two of the TM2 isoforms, originally called "troponin H" due to their higher molecular weight than other (Bullard et al., 1988), are restricted to the IFM of Drosophila. These proteins consist of the complete 257 amino acids of the tropomyosin sequence fused with either 246 or 260 residues of a hydrophobic, alanine-proline-rich C-terminal sequence. The Ala-Pro peptide restricts the flexibility of the protein backbone, favoring an extended structure (Bullard et al., 1988). Similar Ala- Pro-rich sequence is also observed in the N-terminal portion of the Drosophila MLC-2. The Ala-Pro-rich tail of Lethocerus troponin H 36 has been proposed to form projections at ~39 nm intervals seen on the Lethocerus thin filaments, which are not present in the vertebrate striated muscles. This projection may form a relatively inflexible link with some other muscle proteins (possibly in the thick filament), providing the mechanical basis for stretch activation in the IFM.

Troponin Troponin T of Drosophila is a 53-55 kD protein which is larger than the vertebrate proteins (30-37 kD), and has an unusual, highly acidic C-terminus (Bullard et al., 1988). The protein sequence shares similarity to MLC-2 in the N-terminal half and to tropomyosin in the C-terminal half of the conserved a- helical region that is thought be the tropomyosin binding domain. The troponin T gene is located within polytene chromosome band 12A3 of the X-chromosome (Fyrberg et al., 1990a). The gene is composed of nine exons, and three distinct classes of mRNA are produced as a result of alternative splicing. The troponin I gene of Drosophila is localized to position 17A on the X chromosome (Barbas et al., 1991; Beall and Fyrberg, 1991). The gene encodes a family of 10 isoforms resulting from the differential splicing of 13 exons. Troponin C is a protein of approximately 18 kD, and appears to be encoded by a gene family. Two troponin C genes are located at 73F, one of these is expressed in larval muscles and the other in pupal muscles (Fyrberg, personal communication). The third gene is located at 41C, and is expressed in developing pupae. A complete molecular characterization has not been done for these two genes.

Paramyosin Drosophila paramyosin is encoded by a single gene at 66E1 on the third chromosome (Becker et al., 1992). Differential 37 promoter utilization and alternative splicing of transcripts from the paramyosin gene generates two protein isoforms, the ~105 kD standard paramyosin and the ~55 kD miniparamyosin. Paramyosin is a dimer which forms an a-helical coiled-coil rod at the core of thick filaments in invertebrates (Bullard et al, 1973). Paramyosin is found in all invertebrates, but has no known vertebrate homolog. Standard paramyosin is expressed in all larval and adult tissues, whereas miniparamyosin is expressed in all adult muscles except the IFM (Becker et al., 1992). The expression of the standard paramyosin is low in the IFM, but expression of both the standard paramyosin and miniparamyosin is high in the TDT muscles. ' It has been observed that the amount of paramyosin, which makes up the core of insect thick filaments, is variable among insects. Large insects with low wing beat frequencies possess a higher content of paramyosin than small insects with high wing beat frequencies (Bullard et al, 1973). It has been postulated that the amount of paramyosin determines the mechanical properties of the thick filament, and low paramyosin content appears to correlate with rapid contractile activity of the muscle.

Alpha-actinin and other Z discs associated proteins Drosophila a-actinin has been shown to be a component of Z-discs. The a-actinin gene is a single copy gene localized at chromosome band 2C3 of the X chromosome (Saide et al., 1989; Fyrberg et al., 1990b). Three transcripts are produced which encode three protein isoforms; cytoplasmic a-actinin, an a-actinin isoform expressed in larval intersegmental muscles and muscles of the adult head and abdomen, and an isoform expressed in adult thoracic muscles, head muscles, and abdominal muscles. Other Z-disc associated proteins have been identified in Drosophila. These include Z(210), a 210 kD protein, which is 38 localized exclusively to the Z-discs of the IFM and the 28 large cells of the TDT, and Z400/600 which is found in the Z bands of all larval and adult muscles (Saide et al., 1989, Vigoreaux et al., 1991). Z400/600 is restricted to the lateral borders of the Z-band along the Z/I junction. Projectin is a ~60Q kD protein which is a homolog of vertebrate titin and C. elegans twitchin (Ayme-Southgate et al., 1991). Titin, which extends from the Z band to the M line, is thought to play a role in positioning the thick filaments and providing resting tension in the sarcomere. Twitchin shares some structural similarity to titin, but contains a myosin light chain kinase domain which is thought to confer a regulatory function to the protein. In addition, twitchin is localized to the A band, supporting the idea that it may interact with the myosin filaments. The presence of connecting filaments in asynchronous muscles is thought to be a morphological adaptation for stretch- activation. Surprisingly, projectin was found not only in asynchronous muscles, but also in synchronous muscles, although the localization is different in each muscle type (Vigoreaux et al., 1991). In asynchronous muscles (IFM), projectin is localized to the I-band, whereas it is present in the A-band in synchronous muscles. Despite a high degree of similarity, the projectin in synchronous and asynchronous muscles is distinct. It has been suggested that different projectin variants may have distinct functional roles in these muscle types. It is possible that, like titin, projectin has a structural role in asynchronous muscles, whereas in synchronous muscles, it has a regulatory role like that of twitchin. 39

Objectives

A number of mechanisms have been shown to regulate contraction in various muscle types. One mechanism, phosphorylation of MLC-2, has been shown to occur in almost all muscle types. My research focuses on the role of MLC-2 phosphorylation in regulating muscle contraction in Drosophila. The goal of my research is threefold: 1) to determine if Drosophila MLC-2 is phosphoiylated in vivo ; 2) to determine the role of MLC-2 phosphorylation in muscle function by assessing the consequence of disrupting phosphorylation sites; and 3) to clone the gene encoding the myosin light chain kinase (MLCK), an enzyme which phosphorylates MLC-2. In this thesis I show: 1) Phosphorylation of Drosophila MLC-2 does occur in vivo, possibly at multiple sites, resulting in the accumulation of multiple phosphorylated variants. 2) Disruption of phosphorylation at the conserved MLCK phosphorylation sites (Ser66 and Ser67) by substitution with unphosphorylatable residues (Ala) results in an impairment in indirect flight muscle (IFM) function. However, other muscle types are apparently unaffected. Ser66 appears to be functionally more important than Ser67. The substitutions have no effect on the MLC-2 structure, as the mutant proteins are capable of assembly, giving rise to normal muscle structure. The requirement for MLC-2 phosphorylation is apparently unique to the IFM, and control of MLC-2 phosphorylation likely regulates the IFM contractile activity. 3) I have cloned cDNAs encoding a Drosophila MLCK. The cDNA sequences show a high degree of conservation to the catalytic and regulatory domains of the vertebrate smooth muscle MLCK, suggesting that the Drosophila protein shares similar 40 enzymatic properties and may have a similar role in MLC-2 phosphorylation. CHAPTER II

MATERIALS AND METHODS

Fly stocks and culture conditions Marker alleles and balancer chromosomes used In these experiments are described in Lindsley and Zimm (1992). Mlc2E38, is a null allele of Mlc2 and is carried on a chromosome marked with claret (ca), designated Mlc2E38'13 (Warmke, 1990). A new chromosome was made carrying the body color marker ebony (e) which was used to follow the Mlc2E38 allele (Figure 6). In the course of these experiments it was found this chromosome carried a recessive lethal. Thus, Mlc2E38 homozygotes were generated by combining the Mlc2E38' 13 and Mlc2E38, e chromosomes (Figure 8). JW1163.1], JW1[40.1], JW1[18.1], and JW1[16.2] are transformed lines carrying the wild type Mlc2 gene on the second chromosome (Warmke, 1990). These transformed genes have been shown to rescue the dominant flightless phenotype (Warmke, 1990) and were used here as wild type transformation controls for comparison with the transformed mutant Mlc2 genes. Fly stocks were maintained at 17°C on cornmeal-molasses- agar based media. Crosses were performed at 25°C.

41 Figure 6. Construction of the w; Mlc2E38,e/TM3,e stock by recombination. rupP ePr ca/TMl,Me males were crossed to w /w ; Mlc2E38 l3,ca/TM6B,e virgin females. FI females that were heterozygous for the Mlc2E3813,ca and rupP eP rca chromosomes were mated with w/y; Mlc2E3813,ca/TM6B,e males, and recombinant progeny of the genotype w /w or y; Mlc2E38~13, e/TM6B, e were selected by body color (ebony) and dominant flightless phenotype. These recombinants were balanced with a TM3,e balancer chromosome (w; Mlc2E38,e/TM3,e).

42 Mlc2E3S13 ZC □ &a 0 rew^^r^vvvvv\i X TMl,Me TM6B,e

f Mlc2B3813 Mlc2E3813 4 “ FI r^ w v ^ - ^ ^ J g X I?T TM6B,e P r , n o n T b 9 f e Mlc2E3S13 ddd,e -Q- E sS“ F 2 tfVVWVWVWMWH X TM6Bte TM3,e

e , n o n P r i r Mlc2E38'13 5 ic a L fezS F 3 TM3,e

e ,S b

Figure 6. Construction of P Element Transformation Vector and Generation of Transformants by Germline Transformation pCaSJWl was constructed by inserting a 3.4 Kb genomic sequence encoding the entire Mlc2 transcription unit plus 700 bp of 5’ flanking sequence and 650 bp of 3' flanking sequence into a P element transformation vector, pCaSpeR (Warmke et al., 1992). pCaSpeR contains a white+ gene as a selectable marker and P element sequences for insertion (Pirrotta, unpublished). This plasmid was introduced into flies and independent transformed lines were generated (Warmke et al., 1992). Four of these lines with second chromosome inserts were used in my studies {JW1 [63.11 JW1 [40.11, JW1 [18.1] and JW1 [16.2]). For site directed mutagenesis, the 130 bp XhoI/BamHI fragment of the Mlc2 gene was subcloned from pCaSJWl into pBlueScript, and the nucleotide sequences encoding Ser66 and/or Ser67 were changed to Ala using standard procedures (Graham, 1992). The mutations were verified by sequencing (Graham, 1992). P element transformation vectors containing the mutant Mlc2 sequences [P[Mlc2s66A], P[Mlc2S67A\ and P[Mtc2S66>67A)) were constructed by replacing the 130 bp XhoI/BamHI fragment of pCaSJWl with a corresponding fragment encoding the Ala66 and/or Ala67 substitutions. A detailed description of the construction of these vectors is given in Figure 7. Embryos of the genotype y w&7c (Lindsley and Zimm, 1992) were co-injected with 300 |ig/ml of the transformation vector {P[Mlc2S66A], P[MIc2S67a] or P [Mlc2S66’67A]) and 100 jig/ml of 'wings clipped' helper plasmid (Karess and Rubin, 1984) in injection buffer (5 mM KC1, 0.1 mM phosphate buffer, pH 7.8). 'Wings clipped' encodes the transposase required for insertion of the co­ injected pCaSpeR vector. Surviving adults were mated singly to y w 67c males or females, and the transformants were identified by Figure 7. Restriction endonuclease maps of the Mhc2 transformation vectors pCasJWl and P[Mlc2S66-67A], P[Mlc2S66'67A\ was constructed by replacing a BamHI/XhoI fragment of pCasJWl with a corresponding fragment with coding sequences for Ser66 and Ser67 changed to alanines. P[Mlc2s66A] and P [Mlc2S67A] were constructed the same way using the BamHI/XhoI fragments with Ser66 and Ser67 changed to alanines, respectively. Abbreviations for restriction endonucleases are E, EcoRI, H, Hindlll, B, BamHI, and X, Xhol. Symbols used are (■«■), p element sequences; (i-—i), Mlc2 exons; ( CZ3), Mlc2 introns and flanking sequences; (E33), white minigene; ( ------), Carnegie 4 vector; (■ !), the mutant Mlc2 fragment; ( 0 ), polyadenylation sites; (->#-), transcription start site. The Mlc2 exons are numbered below the corresponding sequences (1,2,3).

45 46

H E H I l SEE 3 2 1 pCasJWl

H Q 0 XB EE H ' y » ynferil > 3 2 1

PIMIc2 S66»67A1

Figure 7. 47

scoring for white* eyes in the FI progeny. Lines homo 2ygous for the P element inserts were established. By following the segregation pattern of the white* m arker gene present in pCaSpeR, the inserts were mapped to a chromosome. As Mlc2E3S is a third chromosome gene, only inserts on the other chromosomes could be used in rescue experiments. Three non-chromosome 3 inserts were recovered for P[Mlc2S66A\ and 6 were recovered for P[Mlc2S67A], However, only a single non­ chromosome 3 line carrying P[Mlc2S66<67A] was recovered. Due to potential position effects on expression, more lines with different chromosomal insertion sites were generated by remobilizing the element in a dysgenic cross (see following section).

Mobilization of the PlMIc2S66*67A] insert to multiple chromosome sites The transformed line carrying P[Mlc2S66’67A] on the third chromosome (line RT11) was used as a parental line for mobilization of the insert to new locations. The strategy for mobilization is outlined in Figure 9. Flies from the transformed line (RT11) carrying P[Mlc2S66>67A] on chromosome 3 were crossed to flies carrying a source of transposase, the A23 gene (w ; Dp/TM3, Sb P[ry+ A23]). In Fi heterozygotes with both the A23 chromosome and the P[Mlc2S66-67A] chromosome, transposition of the element occurs at low frequency in the germline. Fi males of this genotype were mated to w /w ; +/+ ; ML2E3813/Tb virgin females, and Sb, Tb male progeny with white+ eyes were selected. These represent cases where the P[Mlc2s66'67A\ has jumped to a new site. These F2 males were mated singly to the parental A23/Dp, e stock to replace the A23 chromosome with the Dp, e balancer chromosome. A23 was removed from the stock to prevent further mobilization of the insert and the flies were crossed into an MZc 2 ^ S background. 48

Rescue of the Dominant Flightless Phenotype and the Recessive Lethality of Mlc2 null mutation The scheme for the rescue of the dominant flightless phenotype and the recessive lethality of the Mlc2 null mutation by either the wild type ( Mlc2+) or the mutant Mlc2 genes {Mlc2S66A, Mlc2S67A, MIc2S66’67A] is shown in Figure 8. In the crosses the transformed gene is followed by the white+ marker gene present in pCaSpeR. To test if a single copy of the transformed Mlc2 gene can rescue the dominant flightless phenotype of the Mlc2 null heterozygote {Mlc2E38/+), w/Y; P[Mlc2]/+ ; +/+ males carrying a transformed copy of the Mlc2 gene were crossed to w /w ; +/+ ; Mtc2E38, e/TM3, e virgin females. An example of a transformed line carrying an insert on the second chromosome is illustrated in Figure 8. Non -Stubble (Sb) FI progeny {w/w or Y; P[Mlc2]/+ ; Mlc2E38, e/+) were collected for flight testing. To test if a single copy of the transformed Mlc2 gene can rescue the recessive lethality of the Mlc2E38 homozygotes, w /y ; P[Mtc2]/+ ; Mlc2E38, e/+ FI progeny were mated with w /w ; +/+ ; Mlc2E38~13 / TM6B, e virgin females in order to balance the Mlc2E38,e chromosome with a marked balancer chromosome ( TM6B, e). Tubby (Tb) F2 progeny carrying a copy of the transformed gene [w/Y; P[Mlc2]/+ ; Mlc2E38, e/TM6B, e) were collected and mated with w /w ; +/+ ; Mlc2E38~13/TM6B, e virgin females, and F3 progeny with wild type body color and carrying a copy of the transformed gene {w/w or Y; P[Mlc2]/+ ; Mlc2E38, e/Mlc2E38~13) were scored for viability and survivors were flight tested. To determine the ability of two copies of the transgenes to rescue the flightless phenotype of the Mlc2E38 homozygotes, two classes of Tb F3 progeny {w/w or Y; P[Mlc2]/+; Mlc2E38, e/TM6B, e and w /w or Y; P[Mlc2]/+; Mlc2E38'13/TM6B, e) were sib mated, and F4 progeny with wild type body and carrying two copies of the Figure 8. Crosses to rescue the dominant flightless phenotype and the recessive lethality of the Mlc2E38 allele. P[Mlc2] represents either the wild type Mlc2 gene (P[Mlc2+) or the mutant Mlc2 genes [P[Mlc2S66A], P[MIc2S67A], or P [M lc^66-67A]). An example of a second chromosome insert is illustrated in this scheme. Virgin females of the genotype w /w ; +/+; Mlc2E38,e/TM3,e were crossed with males carrying the transformed gene ( w /y ; P[Mlc2]/+ ; +/+). Non-Stubble FI progeny heterozygous for Mlc2E38 and carrying the transformed gene were flight tested. FI males of this genotype were then crossed with w /w ; +/+ ; Mlc2E38~13/TM6B,e virgin females in order to balance the Mlc2E38,e allele with a marked chromosome, and ebony, Tubby F2 males carrying the transformed gene were then mated with w /w ; +/+ ; Mlc2E3813/TM 6B,ev irgin females. Non-Tubby F3 progeny carrying one copy of the transformed gene were scored for viability and flight tested. The other two Tubby F3 classes carrying a single copy of the transformed gene were sib mated to produce F4 progeny carrying two copies of the transformed genes. Non-Tubby flies with bright red eyes were then flight tested. These two Tubby F3 classes were also used to establish balanced stocks homozygous for the transformed genes.

49 3 3

V

+ 1+

— N —M 6 I? 6* ♦ 1+ + 1+ s t * Z 9>

8 Figure 9. Mobilization of the transformed Mlc2S66-67A gene and selection of non-third chromosome inserts. The A23, Sb chromosome was introduced into a transformed line ( RT11) carrying a P element insert P[Mlc2S66’67A] on the third chromosome in order to provide transposase for P element mobilization. The A23,Sb chromosome is carried in the w; Dp/TM3,Sb P{ ry+ A23 } 99B stock. In the FI progeny which cany the transformed gene and the A23,Sb chromosome, transposase produced from the A23,Sb chromosome causes the P element insert to mobilize at a low frequency. Males of this FI progeny class were mated to w/w; +/+ ; MJc2E3813/Tb,e virgin females, therefore, most of their Sb,Tb offsprings are expected to be white-eyed except for a few in which the P element has moved out from chromosome 3, yielding colored eyed Sb,Tb flies. These flies were then mated singly to the A23,Sb/Dp,e stock to replace the A23,Sb chromosome with a different balancer chromosome ( Dp,e), since A23 may cause instability of the insert.

51 w . + . A 23,Sb X w • + V 0 Y ' + ’ Dp,e . w ’ + ’ + ^ _ 1 (line RT11)

F I u? • + • ’XT’ X ± • Mlc2E38'13 Y * + * A23,Sh , itf' + ' Tb,e f F2 10 *^7* A 23,Sh X w . + • A 2 3 ,S b 7 * + ’ Tb,e . w ’ + ’ Dp,e

f F3 . \Z* Tb.e 7 ' + ’ Dp.e

V ’ represents the Mlc2s66’67A gene carried on pCaSpeR (P[MIc2S66-67A]) as described in text.

Figure 9. 53

transformed genes (indicated by darker eye color than that produced by a single copy of the gene) ( w /w or Y; P[Mlc2]/P[Mlc2]; Mlc2E38, e/Mlc2 ) were flight tested.

Plight Test Flight testing was performed using a graduated glass cylinder coated internally with mineral oil as described by Benzer (Benzer, 1973). Graduations were marked on the cylinder from 8 at the top to 0 at the bottom. Flies aged 2-5 days old were dropped in groups of 2-3 from the top of the cylinder and scored according to the position they landed inside the cylinder. Good fliers landed near the top and scored high values. Flight impaired flies tended to land closer to the bottom and scored lower values.

Muscle protein preparation Total protein from the thorax was prepared by dissecting thoraces away from the head, abdomen, legs and wings with fine forceps. Isolated thoraces were homogenized in either isoelectric focusing (IEF) lysis buffer (for use in isoelectric focusing gels) or Laemmli SDS sample buffer (for use in SDS-PAGE). The composition of IEF lysis buffer is 9M urea, 4% acrylamide, 2.0% NP-40, 1% ampholytes pH 4-6, 1% ampholytes pH 5-7, 0.01% ammonium persulfate, 0.1% TEMED. The composition of laemmli SDS sample buffer is 0.0625 M Tris-HCl, pH 6.8, 2.3% (w/v) SDS, 5.0% (v/v) p-mercaptoethanol, 10% glycerol, 0.05% (w/v) bromophenol blue. For IFM protein preparation, isolated thoraces were dissected in TB-1 buffer (80 mM potassium chloride, 16 mM sodium chloride, 5 mM magnesium chloride, 15 mM potassium phosphate, pH 7.0, 1% [w/v] polyethylene glycol 6000. The dissected IFM samples were homogenized in the same electrophoresis buffers. Insoluble debris was removed by centrifugation at 10,000 rpm for 2 minutes at room temperature. 54

Protein electrophoresis SDS gel electrophoresis was performed according to Laemmli (Laemmli, 1970). The acrylamide composition of separating gels used in the studies varied from 12.5-16%. The composition of a 12.5% gel, was, for example, 12.5% aciylamide:his-aciylamide (29:1), 375 mM Tris-HCl pH 8.8, 0.1% (w/v) sodium dodecyl sulfate, 0.02% (w/v) ammonium persulfate, and 0.04% (v/v) TEMED. The stacking gel composition was 3% aciylamiderbis- aciylamide (29:1), 125 mM Tris-HCl, pH 6.8, 0.1% (w/v) sodium dodecyl sulfate, 0.1% (w/v) ammonium persulfate, and 0.1% (w/v) TEMED. Protein samples in laemmli SDS sample buffer were heated at 100°C for 5 minutes prior to loading. Electrophoresis was performed at 200 volts in SDS-PAGE buffer pH 8.3 (0.025 M Tris base, 0.192 M glycine, 0.1% SDS). Two dimensional gel electrophoresis was performed according to the method of O'Farrell (1975) as described in the Bio-Rad MINI-PROTEANr II instruction booklet. The pH gradient used in the isoelectric focusing gel was 4-7. The composition of the isoelectric focusing gel for the pH gradient of 4-7 was 9M urea, 4% aciylamide:bis-acrylamide 29:1, 2% (v/v) Nonidet P-40, 2.5% (v/v) ampholytes pH 5-7, 2.5% (v/v) ampholytes pH 4-6, 0.01% (w/v) ammonium persulfate, 0.7% (v/v) TEMED. Protein samples in IEF lysis buffer were loaded on the top of the gel and overlaid with sample overlay buffer (9.5 M urea, 0.5% ampholytes pH 4-6, 0.5% ampholytes pH 5-7, 0.0025% (w/v) bromophenol blue. Electrophoresis buffer for the upper chamber was 0.01 M histidine, and for the lower chamber it was 0.01 M DL-glutamic acid. The gels were pre-run at 500 V for 10 minutes, then the voltage was increased to 750 V for 3.5 hours. The IEF gels were directly placed on the top of SDS gels, and the second dimension was run as described above. Protein gels were stained with Coomassie Brilliant Blue R or with silver. For Coomassie Blue staining, the 55 gels were fixed with 25% (v/v) isopropanol and 10% (v/v) acetic acid for 30 minutes, followed by staining with 25% (v/v) isopropanol, 10% (v/v) acetic acid, and 0.1% (w/v) Coomassie Brilliant Blue R for 1 hour. After staining the gels were destained in 10% (v/v) acetic acid overnight. Silver staining was performed according to Oakley et al. (1980) with some modifications. The gels were fixed in 40% (v/v) methanol and 10% (v/v) acetic acid for one hour, followed by an incubation in 50% (v/v) methanol overnight. The gels were rinsed with demineralized double glass distilled water for 2 hours prior to fixation with 5% glutaraidehyde for 30 minutes. The gels were then rinsed extensively for at least 2 hours with frequent changes of double distilled water, followed by staining with 6 gel volumes of ammoniacal silver solution (21 mM ammonium hydroxide, 0.076% [w/v] sodium hydroxide, 0.78% [w/v)]silver nitrate) with vigorous shaking for 20 minutes. Silver grains were allowed to form in 20 gel volumes of developing solution (0.005% [w/v] citrate, 0.19% [v/v] formaldehyde) until color was visible. The reaction was terminated by quickly rinsing the gels in double distilled water several times, then transferring the gels to 5% (v/v) acetic acid.

In vivo labeling of muscle proteins Two day old adult flies were stuck by their wings to double­ sided tape and 10 (iCi of 32p-0rthophosphoric acid (Dupont/NEN) was injected into the thorax of each fly with a fine needle. Flies that survived post injection for 1 hour at room temperature were dissected and total thoraces were homogenized in IEF lysis buffer. 2D gel electrophoresis and staining were as described above. The stained gels were dried and exposed to X-ray film for autoradiography. 56

Expression of Drosophila MLC-2 fusion protein in E. coli To obtain a large quantity of protein for in vitro biochemical analyses and antibody production, the Drosophila MLC-2 protein was expressed as a fusion protein from E. coli using the pET3a expression system (Rosenberg et al., 1987). Hie N-terminus of the fusion protein contains amino acids 1 to 12 of the T7 gene 10 product, followed by amino acids 54 to 222 of the Drosophila MLC-2 protein (Graham, 1992). The bacterial culture containing the expression plasmid, pMLC14.i, was grown in 10 ml of NZ broth (1% casein hydrolysate, 0.5% NaCl, 0.5% yeast extract, 0.4% glucose, 30 |ig/ml chloramphenicol, 200 |ig/ml ampicillin) overnight before diluting into 1 litre of fresh NZ media. The culture was grown at 37 °C with vigorous shaking until OD600 = 0.3. The protein production was induced by addition of IPTG to a final concentration of 1 mM. The culture was grown for an additional 2 hours before harvesting.

Purification of the Drosophila MLC-2 fusion protein The bacterial culture expressing the fusion protein was centrifuged at 4,000 rpm for 15 minutes at 4 °C. The cell pellet was washed in 1/10 volume of TEB (10 mM Tris, pH 8.0, 1 mM EDTA, 1 mM 2-mercaptoethanol) containing protease inhibitors (1 mM PMSF, 2 (ig/ml leupeptin, 1 (ig/ml betastin, 2 (ig/ml aprotinin, 10 mM benzamidine, and 1 mM pepstatin A) once before resuspension in 1/20 volume of TEB. The cells were lysed by repeated freezing (-70 °C) and thawing (37 °C), and the cell debris was pelleted by centrifugation at 18,000 rpm for 2 hours at 4 °C. Polymin P was added drop wise to a final concentration of 0.3% to precipitate nucleic acids, which were subsequently removed by Figure 10. Purification of the Drosophila MLC-2 fusion protein. The Drosophila MLC-2 fusion protein was purified from the bacterial lysates by Fast Protein Liquid Chromatography (FPLC) using a mono Q column. A linear NaCl gradient of 50-300 mM was applied to the column, and fractions were collected. An aliquot from each fraction was analysed on a 16% SDS-PAGE. Lane 1, molecular weight marker; lane 2, total lysate of the bacterial culture expressing the fusion protein; lanes 3-10, fractions collected from the mono Q column. The MLC-2 fusion protein (19 kD, as indicated by an arrow) was present in lanes 6-9, which correspond to NaCl concentrations of 187.5-200 mM.

57 58

2 6 7 8 9 10

Figure 10. centrifugation at 10,000 rpm for 30 minutes at 4 °C. Solid ammonium sulfate was slowly added to 80% saturation and the protein solution was allowed to precipitate overnight at 4 °C with gentle mixing. The precipitate was collected by centrifugation at 10,000 rpm for 30 minutes at 4 °C, and was resuspended in TEBP (10 mM Tris, pH 8.0, 1 mM EDTA, 1 mM 2-mercaptoethanol, 1 mM PMSF). The protein solution was filtered through a 0.22 pM cellulose acetate filter to remove insoluble particles before loading on a mono Q column. A linear NaCl gradient from 50 to 300 mM was applied to the column, and 60 fractions (1 ml each) were collected. Samples from each fraction were analyzed on a 16% SDS-PAGE to identify the fusion protein containing fractions. Fractions containing the fusion proteins were pooled and concentrated in a Centricon 10 microconcentrator (Amicon).

Production of MLC-2-Specific Antibodies and Western Hybridization Purified MLC-2 fusion protein synthesized in E. coli was used for immunization by mixing 100 pg of protein in a 1:1 ratio with complete Freund’s adjuvant and injected intramuscularly into Sprague-Dawley rats. After three weeks a booster dose of 50 pg protein mixed 1:1 with incomplete Freund’s adjuvant was injected intramuscularly. After one week serum was collected, and the titer was determined by dot blots using HRP-conjugated goat anti-rat IgG. To detect the MLC-2 protein, 50 pg of IFM proteins were separated on either a 12.5% SDS-PAGE or 2-D PAGE gels, and transferred onto a nitrocellulose membrane in transblotting buffer (25 mM Tris, 192 mM glycine, 20% [v/v] methanol, 0.1% SDS) at 100 V for 2 hours. The membrane was blocked in 5% Blotto (5% dried milk powder in PBS buffer) for 1 hour at room temperature, followed by a 2 hour incubation with the MLC-2 antisera (1:100) diluted in 5% Blotto. The membrane was reacted with a secondary antibody (1:200 dilution of HRP-conjugated goat anti-rat IgG) in 5% Blotto for one hour. The color reaction was elicited by the addition of 0.6 mg/ml 4-chloro-1 -napthol in methanol and 0.1% (v/v) hydrogen peroxide in TBS (50 mM Tris, pH 7.5, 200 mM NaCl).

Peptide sequencing To demonstrate that the overproduced proteins detected in bacterial cultures containing pMLC14.1 were the Drosophila MLC-2 fusion proteins, specific protein bands of interest were subjected to peptide sequencing. Lysates of induced bacterial culture containing pMLC14.1 were separated by 16% SDS-PAGE, and the proteins were electroblotted onto PVDF membrane (Millipore, Bedford, Massachusettes) in Caps buffer (10 mM 3-[cyclohexylamino]-l-propanesulfonic acid, pH 11.0, 10% methanol). The electroblotting was carried out at 300 mA for 40 minutes at room temperature. The PVDF membrane was then rinsed in deionized water for 5 minutes, stained with 0.2% (w/v) Coomassie Blue R-250 in 45% methanol, 10% acetic acid for 5 minutes, and destained in 90% methanol, 7% acetic acid for 5-10 minutes. The membrane was rinsed in deionized water for 5-10 minutes, air dried, and stored at -20 °C. Protein bands of interests were excised from the PVDF membrane with a clean razor blade and subjected to 15 cycles of sequencing (Protein sequencing was performed at the Biochemical Instrument Center by automated Edman degradation using an Applied Biosystems Model 470A Gas Phase Protein Sequencer).

Electron microscopy Thoraces were dissected from 2-3 day old adult flies by removing the head, legs, abdomen and wings with fine point 61

forceps. Care was taken to preserve the condition of the thorax and the myofibers inside it. Dissected thoraces were quickly transferred to fixative (3% paraformaldehyde, 3% glutar aldehyde, 100 mM sucrose, 100 mM sodium phosphate, pH 7.2, 2 mM EGTA) and were incubated overnight at room temperature with slow agitation. The following day, the samples were rinsed in wash buffer (100 mM sucrose, 100 mM sodium phosphate, pH 7.2) three times for 10 minutes each, before a post-fixation in 1% osmium tetroxide, 100 mM sodium phosphate, pH 7.2 for 1.5 hours at room temperature. After fixation with osmium, the samples were washed three times as previously described, then they were dehydrated with graded acetone washes (10 minutes each in 30%, 50%, 70%, 80%, 85%, 2 changes of 10 minutes each in 95% and 2 changes of 10 m inutes each in 100%). Following dehydration, the samples were incubated in 50% acetone:50% Spurr's embedding media (Electron Microscopy Sciences) overnight at room temperature with gentle agitation. The samples were incubated in additional 2 changes of 100% Spurr's (2 hours each) before they were transferred to embedding molds and were allowed to polymerize in 100% Spurr's overnight at 60 °C. Embedded samples were sent to the Campus Microscopy and Imaging Facility for sectioning and staining. Sections were viewed and photographed in a Phillips 300 transmission electron microscope.

Preparation of genomic DNA from adult Drosophila Approximately 20 flies were placed in an eppendorf tube and frozen at -70 °C for 10 minutes. The tube was thawed and the flies were ground in 200 gl of Buffer A (10 mM Tris-HCl, pH 7.5, 60 mM NaCl, 10 mM EDTA, 150 mM spermine, 150 mM spermidine, 5% sucrose) with a micro-pestle. Then 200 gl of Buffer B (0.3 M Tris- HCl, pH 9.0, 0.1 M EDTA, 5% sucrose, 1.25% SDS) and 1.5 gl diethyl pyrocarbonate was added, and the tube was mixed gently by 62 inversion. The tube was incubated at 65 °C for 40 minutes. After cooling to room temperature, 30 pi of 8M potassium acetate was added, and the tube was placed on ice for 15 minutes. The cellular debris and protein precipitate were removed by microcentrifugation at 12,000 rpm at room temperature for 30 minutes. The supernatant was removed and the DNA was precipitated with two volumes of absolute ethanol at room temperature twice. The DNA was resuspended in ~ 50 pi of ddlHkO.

PCR amplification of the genomic DNA for cloning of the MLCK gene Amplification was performed following a standard protocol of Perkin Elmer Cetus. The PCR reaction mix was in a reaction buffer supplied by the manufacturer (10 mM Tris-HCl, pH 8.3, 50 mM KC1, 1.5 mM MgCl2 , 0.01% (w/v) gelatin), with added 200 pM each dNTP (dATP, dCTP, dGTP, dTTP), 1 pM each of the 5' and 3’ primers, 0.1 pg genomic DNA (prepared as described above) and 2.5 units Taq polymerase (Perkin Elmer Cetus). The reaction mix was overlaid with 50 pi of light mineral oil. The amplification was performed for 30 cycles, each consisted of 3 segments: denaturation at 94 °C for 1 minute, annealing at 50 °C for 2 minutes and extension at 72 °C for 2 minutes. The oligonucleotide primers are degenerate primers. The underlined nucleotides correspond to the tagged restriction sites, and the nucleotides in bold are the MLCK sequences. Nucleotides in parentheses are the degenerate nucleotides. The sequence of the 5' primer corresponds to amino acid sequence "ELFERI", and the nucleotide sequence is: 5-GCGGATCCGAfA.G)fT.CmA.G.C.T)TTfT.C)GAfA-GlfA.C) G(A,G,C,T)AT -3'. Two 3' primers were used. The first 3' primer corresponds to amino acid sequence "IDFGLA", and the nucleotide sequence is: 63

5'-GGAATTCGCfA.G.C.TlAfA.G)fA.G.C.T)CCfA.GlAAfA.G)TCfA.G.T) AT-3'. The second 3' primer corresponds to amino acid sequence "DDAKDF", and the nucleotide sequence is: 5'-GGMTTCAA(A,G)TC(C,T)TT(A,G,CfT)GC(AfG,C,T)TC(AfG) TC-3'.

Isolation of Drosophila MLCK cDNA clones An imaginal disc cDNA library (constructed by Dr. A. Cowman and provided by Dr. G. Rubin) was screened with a 200 bp PCR product labeled with 32P by random priming. Approximately 180,000 plaques were screened using a standard protocol (Maniatis et al., 1982). Briefly, a nitrocellulose membrane is overlaid on a confluent lawn of plaques for 1 minute before processing by floating the membrane on denaturing solution (1.5 M NaCl, 0.5 M NaOH) for 1 minute and immersing in the same solution for 1 minute. The membrane was neutralized in 1.5 M NaCl, 1.5 M Tris- HCl, pH 8.0 for 5 minutes, rinsed with 20X SSC (3M NaCl, 0.3 M Na Citrate, pH 7.0), then UV cross-linked. The membrane was prehybridized in 5X Denhardts solution, 6 X SSC, 0.5% SDS and 100 |ig/ml denatured salmon sperm DNA at 65 °C for >1 hour. Random primed probe (prepared following the protocol supplied by the manufacturer, United States Biochemical) was boiled for 10 minutes before adding to the membranes in prehybridization solution. Hybridization was carried out at 65 °C overnight. Washing was performed at 65 °C by transferring the membranes into two changes each of 2X SSC, 0.1% SDS for 15 minutes, and 0. IX SSC for 15 minutes. The membranes were partially air dried and exposed to X-ray film. 64

Southern Hybridization Southern hybridization was performed using a standard protocol as described by Maniatis et al. (1982). Digested DNA was separated on 0.8-1.0% agarose gels using the TAE buffer system. The DNA was denatured in two changes of 1.0 M NaCl, 0.5 M NaOH for 15 minutes, followed by two changes of neutralization solution (3 M NaCl, 0.5 M Tris-HCl, pH 7.5), 15 minutes each. The gel was then assembled for blotting by capillary transfer using 20X SSC. The transfer was allowed to proceed overnight, and the membrane was UV cross-linked. Prehybridization, hybridization and washing were as described for plaque hybridization. r DNA sequencing DNA sequencing was essentially as described in the protocol supplied with the Sequenase Version 2.0 sequencing kit (United States Biochemical). Both double stranded and single stranded DNA were used as templates for sequencing.

Total RNA preparation Flies from various developmental stages were frozen in liquid nitrogen and ground to powder with mortar and pestle. The powder was added to a mix of phenol (saturated with 0.2 M Na acetate pH 5.0): 0.2 M Na acetate pH 5.0: 10% SDS at a ratio of 4:3: 0.8. The tube was heated at 65 °C for 20 minutes and vortexed while the solution was still warm. After cooling to room temperature, 0.5 volume of chloroform was added and the mixture was vortexed thoroughly. The mixture was then spun at 5,000 rpm at 4 °C for 5 minutes, and the aqueous phase was transferred to a fresh tube. One third volume each of phenol and chloroform was added, and the extraction was repeated once. The aqueous phase was extracted once with equal volume of chloroform, and the RNA was precipitated with 2.5 volume of absolute ethanol at -70 °C for 65

30 minutes. The RNA was suspended in sterile ddH 2 0 and stored at -70 °C.

Synthesis of RNA probes for RNase Protection Assays Antisense RNA probes were synthesized from linear DNA templates by mixing 2 pi 5X transcription buffer (supplied with T3 or T7 RNA polymerase from GIBCO BRL; 0.2 M Tris-HCl (pH8.0), 40 mM MgCl2 , 10 mM spermidine-(HC1) 3 , 125 mM NaCl), 1 pi DTT, 0.5 pi RNase inhibitors, 0.5 pi rNTPmix (10 mM rATP, 10 mM rCTP, 10 mM rGTP), 1.5 pi 100 pM rUTP, 0.1-1 pg DNA template, 2.5 pi sterile dd H2O, 0.5 pi a 32P-rUTP (800 Ci/mmol) and 25 units of T3 or T7 RNA polymerase. The reaction was incubated at 37 °C for 1 hour before 20 units of DNase I (RNase free) was added. The reaction was allowed to proceed for an additional 30 minutes and the unincorporated nucleotides were removed by passing the reaction through a NucTrap column (Stratagene).

RNase Protection Assays Total RNA (200 pg) was co-precipitated with an RNA probe (prepared as described above) with one-tenth volume of 3M Na acetate pH 5.2 and 2.5 volume of absolute ethanol. The precipitate was resuspended in 20 pi of 5X hybridization buffer (0.2 M PIPES, pH 6.4, 2 M NaCl, 5 mM EDTA), 80 pi deionized formamide. After heating at 95 °C for 5 minutes, the tube was quickly transferred to 45 °C and the hybridization was allowed to proceed overnight. Digestion with RNases was initiated by adding 300 pi of RNase digestion buffer (0.3 M NaCl, 10 mM Tris-HCl, pH 7.0, 5 mM EDTA), 2 pi 10 mg/ml RNase A, 2 pi 100 units/pi RNase Tl, and the reaction mix was incubated at 30 °C for 30 minutes. The digest was terminated by addition of 20 gl 10% SDS and 5 pi 10 mg/ml protease K. After a 15 minute incubation at 37 °C, the reaction was extracted once with phenol/chloroform, and the RNA 66 was precipitated with 1/10 volume of 3M Na acetate and 2.5 volume of absolute ethanol. The digestion products were resuspended in 1 0 pi formamide loading buffer (80% form amide, 10 mM EDTA, 1 mg/ml xylene cyanol, 1 mg/ml bromophenol blue), and separated on a 6 % urea-acrylamide sequencing gel. The position of the protected RNA fragments was visualized by autoradiography. CHAPTER HI

MULTIPLE PHOSPHORYLATED VARIANTS OF DROSOPHILA MYOSIN LIGHT CHAIN 2 PROTEIN ACCUMULATE IN VIVO. Introduction

Muscle contraction is regulated by Ca2+ which exerts control through either thick or thin filament associated proteins (Sellers and Adelstein, 1987). In striated muscles, Ca2+ binds to either troponin (vertebrates) or myosin (molluscs), whereas in smooth muscles, Ca2+ activates myosin light chain kinase (MLCK) to phosphorylate myosin regulatory light chain (RLC, or myosin light chain 2, MLC-2) or binds directly to myosin (molluscs). Certain types of muscles, particularly those of invertebrates, possess both thick and thin filament-linked regulation (Lehman and Szent- Gyorgyi, 1975). Drosophila muscles contain protein homologs of both vertebrate skeletal and smooth muscle contractile proteins (see review by Bernstein et al., 1993). Although it is not known at present which mode of regulation operates in various Drosophila muscle types, it has been suggested that phosphorylation of MLC-2 has an important role in regulating the indirect flight muscle (IFM). Three types of myosin light chain proteins have been identified in Drosophila, MLC-1 (34 kD), MLC-2 (30 kD) and MLC-3 (20 kD). Initially it was thought that these correspond to the vertebrate MLC-1 (25 kD), MLC-2 (20 kD) and MLC-3 (16 kD),

67 TABLE 1 MYOSIN LIGHT CHAIN NOMENCLATURE

DROSOPHILA PROTEIN VERTEBRATEDROSOPHILA PROTEIN NUMBERING (Mogaml et al., 1982) IFM TUBULAR

myosin alkali light chain MLC-1

138 144 myosin regulatoiy light chain MLC-2 MLC-1 MLC-2 148,149 148,149

myosin alkali light chain MLC-3 MLC-3 185 186 Figure 11. Accumulation pattern of MLC-2 variants in vivo. (a) Pattern of MLC-2 accumulation in wild type (Canton S) adult fly. Total thoracic proteins from adult flies were separated by 2-D i PAGE and silver-stained. The pH gradient of the first dimension is 4-7, and the polyacrylamide composition of the second dimension is 12.5%. The arrow indicates the position of spot 148. Spots 149, 138 and 158-160 (flightin) are also indicated. (b) Pattern of MLC-2 variants illustrated by Western Hybridization. IFM proteins separated by 2-D PAGE were electroblotted onto a nitrocellulose membrane, and reacted with anti -Drosophila MLC-2 antisera. Protein spots 148, 149, as well as 138 were recognized by the antiserum. (c) Accumulation pattern of MLC-2 phosphoiylated variants shown by in vivo labeling. 32p was injected into thoraces of adult flies, and total thoracic proteins were displayed by 2-D PAGE. The gel was stained with Coomassie Brilliant Blue, dried, and exposed to x-ray film. Three sets of proteins incorporated 32p; SpQts 149, 138, and flightin (158-160). The arrow indicates the position of spot 148 which is not phosphoiylated. The position of spot 148 was established by aligning the autoradiogram with the stained gel (not shown).

69 70

1 4 8 \

Figure 11 respectively. However, results from other groups, and this work suggests that Drosophila MLC-1 and MLC-2 are encoded by the same gene and correspond to the regulatory light chain (MLC-2 of vertebrates) (Table 1). In Drosophila, MLC-3 is the alkali light chain and thus corresponds to vertebrate MLC-1 and MLC-3 which are isoforms of the alkali light chain (Table 1). According to the protein numbering system of Mogami et al. (1982), IFM-specific MLC-1, MLC-2 and MLC-3 correspond to spots 138, 148 and 149, and 185 (Figure 11; Table 1). Tubular muscle- specific MLC-1, MLC-2 and MLC-3 correspond to spots 144, 148 and 149, and 186 (Table 1). Each of the spots 138 and 149 is in fact a series of spots with similar molecular weight and decreasing isoelectric focusing (pi) value. I show here, using an MLC-2 specific antibody, and others have shown by phosphatase treatment, that spots 138 and 149 are probably generated by phosphorylation of MLC-2 (Takano-Ohmuro et al., 1990). Phosphatase treatment resulted in gradual elimination of spots 138 and 149 and the only remaining spot was 148, which corresponds to unphosphorylated MLC-2. The identity of these spots was further analyzed by proteolytic mapping with V 8 protease. The mapping data showed that spot 138 contains some common peptides with spots 148 and 149. However, some peptides are not common between spots 138 and 148/149. This difference could be because 138 and 148/149 are related but distinct proteins, or because they are post translationally modified variants of the same protein. Taken together with the above evidence the latter seems more likely. It has been observed that spots 138 and 149 are not present in pupal protein preparations, but start to accumulate soon after eclosion of adult flies concomitant with the onset of flight ability (Takahashi and Maruyama, 1987). Phosphorylation of MLC-2 has also been correlated with activation of myosin MgATPase activity in 72 vitro (Takahashi and Maruyama, 1987, Takahashi et al., 1990a). Thus, it was postulated that phosphorylation of MLC-2, which results in accumulation of spots 138 and 149, is responsible for activation of myosin MgATPase activity which, in turn, correlates with the acquisition of flight. The molecular characterization of the Drosophila muscle- specific MLC-2 gene indicates that it is a single copy gene and there are no other related genes in the Drosophila genome (Parker et al., 1985, Toffenetti et al., 1987). The MLC-2 gene is composed of three exons. At least two classes of transcripts (1.1 and 1.4 kb) accumulate as a result of alternative transcription stat sites and polyadenylation sites. However, the gene appears to encode a single polypeptide with no evidence of alternative splicing to yield additional protein products. A single null allele of the sarcomeric Mlc2 gene (Mlc2E38) has been identified (Warmke et al., 1992). The mutation produces a stop codon at amino acid 10 of the protein coding sequence. The Mlc2E38 heterozygote produces ~50% less Mlc2 mRNA and, as I show here, MLC-2 protein (spots 148 and 149) are reduced in comparison to the wild type. In addition, I have observed that there is a proportional reduction of spot 138, concomitant with reduction of spots 148 and 149. This observation suggests that spot 138 may be a variant encoded by the Mlc2 gene th at is post- translationally modified to produce a protein with different electrophoretic mobility. The Drosophila MLC-2 protein contains a num ber of phosphoiylatable amino acids which are potential substrates for various kinases. Among these, Ser 6 6 and Ser67 are conserved with the known myosin light chain kinase (MLCK) phosphorylation sites in other species (Stull et al., 1986) (Figure 5). Ser 6 6 in the Drosophila MLC-2 is preceded by an arginine residue at position -3 and a lysine at position - 6 . A basic residue at position -3 has been shown to be required for efficient phosphorylation at the analogous residue, Serl9, which is the major phosphorylation site of smooth muscle MLC-2 (Kemp and Pearson, 1985). Shifting the basic residue to position -4 favors phosphorylation at Thrl 8 instead of Serl9, and the distance between basic residues at position -3 and -6 has been shown to affect kinetic properties and the site of phosphorylation by smooth muscle MLCK (Kemp and Pearson, 1985). In addition, it has been shown that 6 -8 and 10-11 basic residues N-terminal to the phosphorylation sites are important for determining kinetics of phosphorylation by smooth and skeletal muscle MLCK, respectively (Kemp et al., 1983; Michnoff et al., 1986). There are also additional basic residues N- terminal to Ser 6 6 and Ser67 of the Drosophila MLC-2 , some of which may be MLCK substrate determinants analogous to those of the vertebrate counterparts. A Drosophila MLC-2 fusion protein expressed in E. coli was shown to be phosphoiylated by a rabbit skeletal muscle MLCK in vitro (Graham, 1992). The phosphorylation sites were mapped to Ser6 6 and Ser67, with Ser 6 6 phosphoiylated to a greater extent than Ser67. Ser 6 6 aligns with residue Seri9, the major phosphorylation site of the chicken smooth muscle MLC-2. Ser 6 6 is in a similar context and is preceeded by basic residues at positions -3 and -6 (Kemp and Pearson, 1985). However, phosphorylation at Ser 6 6 and/or Ser67 alone cannot account for the num ber of MLC-2 variants observed in vivo (Takano-Ohmuro et al., 1990; this work). It is likely that the MLC-2 protein is further modified in vivo, either by additional phosphorylation or other modifications. As discussed above, a number of questions remain about the relationship between spots 148, 149 and 138. To resolve this issue, I have generated a polyclonal antibody against a purified Drosophila MLC-2 fusion protein expressed in E. coll The antibody was used to identify MLC-2 variants by Western Hybridization. I show here, that spot 148 is an unphosphorylated MLC-2 variant, whereas spots 149 and 138 are phosphoiylated variants of MLC-2.

Results and Discussion

Expression of Drosophila MLC-2 fusion protein in E. coli A Drosophila MLC-2 fusion protein was expressed in E. coli using the pET3a expression vector (Rosenberg et al., 1987). The expression plasmid, pMLC14.1 (Graham, 1992), contains the coding information of the Drosophila MLC-2 cDNA from amino acids 54 to 2 2 2 (C-terminus) fused in frame with the coding sequence for amino acids 1 to 12 of the phage T7 gene-10 product (Figure 12). pMLC14.1 was transformed into an E. coli strain BL21 (DE3)pLysS which contains a chromosomal copy of the T7 polymerase gene under the control of the lacUV5 promoter and an extrachromosomal plasmid (pLysS) that constitutivefy expresses lysozyme. Under non-induced conditions the low level of T7 polymerase produced from the lacUV5 promoter is titrated out by the lysozyme (Moffatt and Studier, 1987), thus preventing expression of the Mlc2 fusion gene which is under the control of a T7 promoter. Addition of IPTG into the culture medium relieves the repression of the lacUV5 promoter, resulting in a production of a large quantity of T7 polymerase which, in turn, initiates transcription of the Mlc2 fusion gene. Two proteins of molecular weight 21 kD and 19 kD were induced in the lysates of bacteria harboring pMLC14.1, but not in the control lysates containing pET3a plasmid without insert (Figure 12). After induction, the bacterial cells were lysed and analyzed for expressed proteins by SDS-PAGE. Newly expressed 21 kD and the 19 kD proteins were found to be associated with both the supernatant and the pellet fractions (Figure 12, lanes 1 and 4). Figure 12. Expression of Drosophila MLC-2 in E. coll (A) pMLC14.1 used for the expression of Drosophila MLC-2 in E. coli and the parental plasmid, pET3a. <(>10, S10, Tty represent the T7 promoter, the N-terminal coding region of T7 phage gene 10 product, and T7 transcription termination region, respectively (Rosenberg et al., 1987). (B) Lysates of bacteria transformed with various plasmids were separated on a 12.5% SDS gel. Lanes 1-3, supernatants; lanes 4-6, pellets. Lanes 1 and 4, pET3a (no insert); lanes 2 and 5, MLC13.1 (MLC-2 insert in the anti-sense orientation); lanes 3-6, MLC14.1 (MLC-2 insert in the sense orientation). The arrowheads indicate the position corresponding to the expressed proteins from MLC14.1 plasmid (21 kD and 19 kD, respectively). (C) N-terminal amino acid sequence of the induced 21 and 19 kD proteins which is identical in the N-terminus. This sequence is shown aligned with the sequence of the lull length MLC-2 protein (Parker et al., 1985). Identities are indicated by dots. The fusion protein has 6 amino acids derived from the phage 10 gene at the N-terminus.

75 76

A.

J3 I0 SIO fe pET3a

X

c

M L C -2 1 MADEKKKVKKKKTKEEGGTSETASEAASEAATPAPAATPA 4 0

M L C -2 41 P A A 5 A ? G S K P A S G G S P. G S P K S K P A G 3 S V F S V F S Q K Q I A F. F 8 0

.* 2 FUSION 1 m g;>c.m3P 3 8 PROTEIN

Figure 12 Twenty one kD was the size expected for the fusion protein, but the expression of the 19 kD polypeptide was not predicted. To determine the identity of these proteins, fractionated bacterial lysates were transblotted onto a PVDF membrane, stained, and the protein bands in question were excised for sequencing. The sequences of the N-terminal 15 amino acids for both proteins are identical and are as expected for the fusion protein sequence (Figure 12). Therefore, it was concluded that the 21 kD protein is indeed the Drosophila MLC-2 fusion protein and the 19 kD protein is produced by cleavage of the 21 kD species in the C-terminal region.

Multiple variants of the Drosophila MLC-2 protein accumulate in vivo To determine the accumulation pattern of the MLC-2 variants in vivo, I purified the Drosophila MLC-2 fusion protein from the bacterial lysates to near homogeneity for antiserum production (see Chapter II). The purified proteins were used for immunization of rats, and the serum was collected and tested against fly homogenates, separated on an SDS polyacrylamide gel, followed by Western Hybridization. The antiserum showed specific binding to a 30 kD protein, the expected size of Drosophila MLC-2. When the antiserum was reacted with IFM proteins separated by 2-D PAGE, spots 148, 149 and 138 were shown to label. Labeling of spots 148 and 149 was expected since they are known to correspond to MLC-2 based on their copurification with myosin (Takano-Ohmuro et al., 1983). Protein 148 migrates as a single spot on a 2-D PAGE, whereas protein 149 is composed of multiple protein spots with similar molecular weight and decreasing pi values (Figure 11). The size of proteins 148 and 149 were determined by Mogami, et al. (1982) to be 30 kD and their pi values were 5.24 and 5.08 respectively. In contrast to the result of 78

Mogami’s group, I observed many more protein variants corresponding to spot 149 (up to 10 spots) (Figure 11). This discrepancy may be due to differences in preparing the muscle protein samples. I found variability in the number of protein variants detected between experiments. In addition, the pi range used for the isoelectric focusing gels in my experiments is narrower (4-7) than th at used by Mogami's group (3-10). Spot 138 has an apparent molecular weight of 34 kD and a pi of 4.50 (Mogami et al., 1982). I also observed many more variants of spot 138 than Mogami's group (Figure 11). The MLC-2 antibody reacted with spot 138, which is consistent with its being a variant of MLC-2. This is supported by the results of phosphatase treatment which showed spot 138 is a phosphorylated variant of MLC-2 (Takano-Ohmuro et al., 1990). Thus although 138 was thought to be a homclog of vertebrate MLC-1 based on its size (Emerson and Bernstein, 1987) my data from antibody analysis is consistent with the idea that 138 is a variant of MLC-2. Although the evidence suggests that spots 149 and 138 are produced by phosphorylation of spot 148, the multiplicity of spots and major difference in apparent molecular weight between 149 and 138 suggest that multiple phosphorylations and perhaps other modifications of MLC-2 occur.

Reduction in MLC-2 protein accumulation in Mlc2 m utants The Mlc2 null allele, Mlc2E38, has a stop codon at amino acid 10 of the protein coding sequence and results in a -50% reduction of Mlc2 RNA accumulation in heterozygotes (Mlc2E38/+) (Warmke et al., 1992). To test if the mutation causes a concomitant reduction in the MLC-2 protein, total thoracic proteins from Mlc2E38/+ flies were separated by 2-D PAGE, and the gels were silver stained. The 2-D PAGE gels were scanned at two Optical Densities (O.D. 2.00 and O.D. 0.75) and the density of spots 148, 149, 138 and a 79

control protein were measured. The control protein is not a known muscle protein. The densities of spots 148 and the 149 series of spots were added and normalized to the control protein in wild type and Mlc2E38/+ samples. The same was done for the 138 series of spots. At O.D. 2.00 the mutant 148/149 and 138 densities were 0.31 and 0.21 (respectively) of the wild type value, At O.D. 0.75 the mutant 148/149 and 138 densities were 0.26 and 0.42 (respectively) of the wild type value, These results show that MLC-2 accumulation is reduced in Mlc2E38/+. The concomitant reduction of spot 138 with spotsl48/149, to about 30% of the wild type level in Mtc2E38/+, suggests that these proteins are encoded by the same gene (Mtc2).

In vivo labeling of 1PM proteins with 32P A likely post-translational modification for MLC-2 is phosphorylation since addition of phosphate groups to proteins results in an increase in acidity, consistent with the observed decrease in pi of the MLC-2 variants. As described, Takano- Ohmuro et al. (1990) have shown that these multiple protein spots can be eliminated by phosphatase treatment, strongly suggesting that they are phosphoiylated variants of MLC-2 (Takano-Ohmuro et al., 1990). To further characterize MLC-2 modification, I performed in vivo labeling of thoracic myofibrillar proteins with 32p. Total proteins from isolated thoraces were displayed by 2-D PAGE and the labeled proteins were detected by autoradiography (Figure 11). Only three species of proteins incorporated 32p to a detectable level: protein 138, protein 149 and flightin (composed of protein spots 158, 159 and 160) (Vigoreaux et al., 1992). Therefore, proteins 149 and 138, that react with the anti-MLC-2 antisera, appear to be phosphoiylated variants of MLC-2. Spot 148, which is the most basic variant, is unphosphoiylated. 8 0

Discussion

Multiple IFM protein spots are recognized by an MLC-2 specific antibody I generated MLC-2 specific antibodies and used these to characterize the protein products of the Mlc2 gene. The antibody crossreacts with a large number of IFM protein spots that have been designated 138, 148 and 149 by Mogami et al., (1982). Spots 148 and 149 were known to be MLC-2 proteins (Table 1; Takano- Ohmuro et al. 1983). However, the IFM spot 138 was originally thought to be an alkali light chain isoform based on its similar molecular weight'to a vertebrate alkali light chain isoform, MLC-1 (Emerson and Bernstein, 1987). The following results, from this work and others support the idea that 138 is a product of post- translational modification of MLC-2 : 1. Protein 138 reacts strongly with the anti-Drosophila MLC-2 antibody which was produced against a highly purified Drosophila MLC-2 fusion protein. It is very unlikely that cross reactivity to protein 138 is due to a contamination by another Drosophila protein as the MLC-2 preparation used for immunization was bacterially produced. 2. In Mlc2E38 heterozygotes only a single copy of the MLC-2 gene is functional as the mutation causes RNA instability and has a stop codon at amino acid 10. I observed that proteins 148,149 and 138 accumulate to ~30% of the wild type level in Mlc2/+ flies. The concomitant reduction of 138 with 148 and 149 suggests that they are proteins encoded by the same gene (Mlc-2). 3. Phosphafase treatments result in a gradual shift of proteins 138 and 149 to the position corresponding to protein 148, suggesting that they are phosphorylated variants of protein 148 (Takano- Ohmuro et al., 1990). 81

It is still a formal possibility that protein 138 is a product of a different gene which is related to MLC-2 gene. Indeed the MLC-2 genes of vertebrates comprise a gene family (Emerson and Bernstein, 1987). A number of groups have analyzed the Drosophila genome for Mlc2 genes, b ut only a single sarcomeric MLC-2 gene and a single cytoplasmic MLC-2 gene have been identified (Parker et al., 1985; Toffenetti et al., 1987; Karess et al., 1991). Neither the muscle-specific nor the cytoplasmic MLC-2 gene shows cross hybridization to the other or localizes to additional chromosomal regions in the Drosophila genome as judged by in situ hybridization to salivary gland chromosomes (Parker et al., 1985; Toffenetti et al., 1987; Karess et al., 1991). At the nucleotide level, there is a maximum of 60.6% identity over a 198 bp region of the gene coding sequences (the transcription unit of the muscle-specific Mlc2 and the cytoplasmic Mlc2 are ~ 3.4 Kb and ~ 2.2 Kb, respectively). At the protein level, there is a maximum of 67.3% similarity or a 39.8% identity over a stretch of 173 amino acids (the muscle- specific MLC-2 is 222 amino acids in length whereas the cytoplasmic MLC-2 is 174 residues). It is possible that protein 138 is encoded by an MLC-2-related gene that has not yet been discovered by nucleic acid hybridization, and there is sufficient degree of homology at the protein level to cause cross-reaction with the anti-MLC-2 antisera. But given the above evidence this seems unlikely.

Post translational modification of MLC-2 MLC-2 from a number of animals is known to be phosphoiylated by MLCK at specific serine or threonine residues. In Drosophila MLC-2, these residues would correspond to Ser 6 6 and Ser67, phosphorylation of which would generate a maximum number of 3 protein variants: unphosphorylated, mono- and di- phosphorylated variants. However, there appear to be at least 10 82 phosphorylated variants comprising proteins 149 and 138 as shown by in vivo 32p labeling (Figure 11), suggesting that phosphorylation sites other than Ser 6 6 and Ser67 exist. Alternatively, the MLC-2 protein may be modified by other post- translational modifications, resulting in an increase of negative charges (and mass, in the case of protein 138). MLC-2 has been shown to be a substrate for several other kinases besides MLCK. These include Protein Kinase C (PKC), cAMP-dependent protein kinase (cAK), Ca2+/calmodulin-dependent protein kinase type II (CaMKII), protease-activated protein kinase I (PAPKI), cell cycle-dependent protein kinase (pp 3 4 c^c2 ) my0sin I heavy chain kinase (MIHCK), the epidermal growth factor receptor, the insulin receptor, and casein kinase I and II (see reviews in Tan et al., 1992, Sellers and Adelstein, 1986). Most of these enzymes have been shown to phosphorylate purified MLC-2 in vitro. But in most cases phosphorylation of MLC-2 by these enzymes has not been examined in vivo. It has been shown that Protein Kinase C phosphorylates thymus myosin and platelet myosin in vivo at Seri, Ser2 and Thr9 (Carroll and Wagner, 1989; Ikebe and Reardon, 1990). The effect of PKC phosphorylation has been shown to be variable. Phosphorylation by PKC inhibits MgATPase activity, b ut not myosin conformation or the filament formation of platelet myosin prephosphoiylated by MLCK. In contrast, neither MgATPase activity nor filament stability of thymus myosin is affected by PKC phosphorylation. Phosphorylation by cAK, CaMKII, PAPKI has been shown to occur at the same sites as those of MLCK, whereas phosphorylation by pp 3 4 cdc2 has been shown to occur at Seri or Ser2. The Drosophila MLC-2 protein has an N-terminal extension of approximately 50 amino acids which the vertebrate counterparts lack. This region of the protein contains 9 serine and 6 threonine 83 residues, some of which may be substrates for other kinases. For example, the context of Ser52 fits with the consensus for recognition by cAK and PKC, (R-R-X-S, K-R-X-X-S or R-X-X-R-X- X-S). cAK and PKC do not have very stringent substrate recognition sequence, and may phosphoiylate additional serine or threonine residues. The size of protein 138 (34 kD) is considerably larger than that of 148 and 149, (30 kD), suggesting that phosphorylation is unlikely to be the sole mechanism of MLC-2 modification, and additional modification may be involved, causing a mass increase. An example of a myofibrillar protein that is post-translationally modified, resulting in an increase in molecular weight is arthrin (Ball et al., 1987). Arthrin is a stable form of ubiquitinated actin III, encoded by the Act88F gene. Ubiquitination, shifts the molecular weight of actin III from 43 kD to 55 kD (Ball et al., 1987) and the pi from 5.88 to 6.0 (Mogami et al., 1982). Arthrin cross reacts strongly with an anti-actin antibody and its accumulation pattern is altered in parallel with that of actin 8 8 F protein in several Act88F m utant alleles (Ball et al., 1987). Since ubiquitination causes a molecular weight increase of ~10 kD and a shift of protein to a more basic pi, it is unlikely to be responsible for modifying the MLC-2 138 variant. Modifications that increase mass of the proteins besides ubiquitination are, for example, glycosylation and lipation. However, these mechanisms occur in secreted proteins and membrane-bound proteins, and modification of muscle proteins by these mechanisms have not been documented. Other modifications that increase acidity of proteins besides phosphorylation are, for example, sulfation, acetylation, acylation (Wold, 1981). Sulfation of tyrosine residues occurs in many eukaryotic secretoiy proteins and a few membrane proteins (Baeuerle, 1985); therefore, it is unlikely a candidate for MLC-2 modification. N-terminal acetylation of myofibrillar proteins, for example, actin (Hennessey et al., 1991) and MLC-2 (Toffenetti et al., 1987), has been documented. It cannot, however, be excluded that phosphorylation alone is responsible for a drastic mobility shift of MLC-2 protein in an SDS-PAGE gel. Phosphorylation at multiple sites has been shown to cause a conformational change in Physarum histone HI, resulting in a retardation of mobility on an SDS-PAGE gel (Jerzmanowski and Maleszewski, 1985). It was found that the number of conformational variants was relatively small compared to the number of phosphorylation sites, indicating that multiple phosphorylations are needed to bring about a unit conformational transition of histone HI. By analogy, a number of phosphorylations may be required to induce a change in MLC-2 conformation, to yield protein 138 with a slower migration on an SDS-PAGE gel. CHAPTER IV

THE EFFECT OF DISRUPTION OF MYOSIN LIGHT CHAIN-2 PHOSPHORYLATION SITES ON FLIGHT, VIABILITY AND INDIRECT FLIGHT MUSCLE ULTRASTRUCTURE

Introduction

I By Western Hybridization and in vivo labeling, I have shown Drosophila MLC-2 is phosphoiylated in vivo, possibly at multiple sites, resulting in a complex accumulation pattern of phosphorylated variants (Figure 11, Chapter III). In this chapter I address the role of MLC-2 phosphorylation in muscle function. Phosphorylation of MLC-2 has been shown to be correlated with tension development in smooth muscle fibers in vivo, suggesting that MLC-2 phosphorylation is important for contraction (Silver and Stull, 1982; Gerthoffer and Murphy, 1983). The requirement for MLC-2 phosphorylation is further supported by the block in contraction produced by treatment of smooth muscles with calmodulin agonists which prevent MLC-2 phosphorylation by inhibiting MLCK which is Ca2+/calmodulin- dependent (Hidaka et al., 1979; Sparrow et al., 1981). In addition, phosphorylation of MLC-2 has been correlated with activation of myosin MgATPase activity in vitro, suggesting that muscle contracts as a result of an increase in myosin MgATPase activity activated by MLC-2 phosphorylation (Sobieszek, 1977; Persechini et al., 1981). In contrast to smooth muscles, MLC-2 phosphorylation is not required for activation of skeletal muscle myosin MgATPase activity. Nevertheless, it has been shown to modulate kinetics of

85 contraction at suboptimal Ca2+ concentrations (Pemrick, 1980). Thus far, the role of MLC-2 phosphorylation in Drosophila muscle has not been determined. However, it has been noted that there is a change in the pattern of MLC-2 accumulation which correlates with acquisition of flight in young adult flies. Proteins 138 and 149, which I have shown to be phosphoiylated variants of MLC-2 (Figure 11, Chapter III), are absent in pupal muscle protein preparations, but accumulate within a few hours after eclosion of adult flies, in parallel with the gain of flight ability. Phosphorylation of Drosophila MLC-2 has also been correlated with an increase in myosin ATPase activity in vitro (Takahashi and Maruyama, 1987). Together, these results strongly suggest that MLC-2 phosphorylation is important for flight. However, this suggestion is only based on indirect evidence. Here, I address the roles of MLC-2 phosphorylation directly by expressing mutant, unphosphorylatable, MLC-2 proteins in vivo, and analyzing the effects of the mutations on muscle structure and function. In Drosophila MLC-2, serine 6 6 (Ser6 6 ) and serine 67 (Ser67) correspond with the vertebrate MLCK phosphorylation sites, and a bacterially produced MLC-2 protein has been shown to be phosphorylated in vitro by a rabbit skeletal muscle MLCK at these sites (Graham, 1992). The Mlc2 sequences encoding Ser 6 6 and/or Ser67 were changed to alanine (Ala) by in vitro site-directed mutagenesis (Graham, 1992). Substitution with a small, neutral amino acid such as alanine (Ala) is expected to produce the least perturbation of protein structure and thus the least effect on protein function as a result of structural perturbation. My results show that indeed muscle structure is unaffected, as judged by electron microscopy, and thus I have been able to assess the requirem ent of MLC-2 phosphorylation on muscle function. IFM structure is veiy sensitive to the amount and form of muscle proteins and there are many hypomorphic mutations, including missense mutations, that cause a disruption of IFM structure and consequently a flightless phenotype (see review by Bernstein et al., 1993). Missense mutations include the up101 and int3 alleles of troponin-T, and the heldup2 allele of troponin-I. In these mutants normal flight muscles form but degenerate soon after eclosion as they start to be utilized (Fyrberg et al., 1990 a; Beall and Fyrberg, 1991). Thus, in these cases, the effect of the missense mutation on the IFM function cannot be tested because the mutation causes perturbation of muscle structure. However, two flightless mutations of the Act88F gene, which encodes the IFM-specific isoform, have been described (Drummond et al., 1991). Each of these mutations, E316K and G368E, has a single amino acid change (Glu316 to Lys and Glu368 to Lys, respectively) in the Act8 8 F protein, and was shown to have near normal IFM structure Here I describe the in vivo role of MLC-2 phosphorylation in Drosophila muscle function. Phosphorylation of MLC-2 at Ser 6 6 and Ser67 is essential for normal function of the IFM, but not for other muscle types. Ser 6 6 appears to be functionally more important than Ser67. Substitution of either or both serines with alanines has no effect on the IFM structure, clearly demonstrating that flight impairment of flies expressing the mutant proteins is not due to abnormal muscle structure. Rather, it is due to lack of MLC-2 phosphorylation at Ser 6 6 and/or Ser67. 88

Results

Disruption of MLC-2 phosphorylation by in vitro site-directed mutagenesis and construction of transformation vectors carrying the m u ta n t Mlc2 genes The Drosophila MLC-2 protein is phosphorylated at Ser 6 6 and Ser67 by a heterologous myosin light chain kinase (MLCK) in vitro (Graham, 1992). To determine the role of phosphorylation at these residues, they were rendered unphosphoiylatable by in vitro site- directed mutagenesis, either or both serines were replaced with unphosphoiylatable alanines (Graham, 1992). The side chain (CH2 -H) group of alanine is neutral and comparable in size to that of serine (CH 2-OH), thus, it is expected to cause minimal perturbation of the MLC-2 protein structure or conformation (Graham, 1992). Three mutant Mlc2 genes with either or both serines replaced with alanines were generated by in vitro site-directed mutagenesis. They are designated Mlc2S66A<, Mlc2S67A, or Mlc2S66<67A to represent Mlc2 gene with the coding sequence for Ser 6 6 and/or Ser67 changed to that of Ala. P element transformation vectors carrying mutant Mlc2 genes were constructed based on pCaSJWl (Warmke et al., 1992) in which the entire transcription unit of the wild type Mlc2 gene including ~700 bp of 5' flanking sequence and ~650 bp of 3' flanking sequence are inserted in pCaSpeR (Pirrotta, unpublished) (see Figure 7, Chapter II). This P element transformation vector contains the white+ gene as a selectable marker, and the P element inverted repeats required for insertion. A 130 bp XhoI/BamHI fragment of pCaSJWl, which contains Ser6 6 and Ser67 of the wild type Mlc2 sequence, was replaced with the corresponding fragments containing Ala 6 6 and/or Ala67 substitutions. The P element transformation vectors carrying m utant Mlc2 genes are designated P [Mlc2S66A], P[Mlc2S67A] and 89

PMlc2S66'67A] and correspond to alanine substitutions at Ser66, Ser67 and both serines respectively.

Generation of a series of lines expressing mutant MLC-2 protein with disrupted phosphorylation sites Mlc2E38 is a null allele of the sarcomeric Mlc2 gene, and does not produce MIX)-2 protein (Warmke et al., 1992). Mlc2E38 heterozygotes exhibit a dominant flightless phenotype and disrupted IFM structure due to a >50% reduction of MLC-2 (Chapter III). Mlc2E38homozygotes are embryonic lethal. The dominant flightless phenotype of the Mlc2E38 heterozygotes and the recessive lethality of the Mlc2E38 homozygotes can be rescued by a transformed copy of the wild type Mlc2 gene (Warmke et al., 1992). In order to assess the role of MLC-2 phosphorylation, m utant Mlc2 genes with coding sequences for Ser66 and/or Ser67 changed to alanines were introduced into the Mlc2E38 mutant flies by P element mediated germline transformation and genetic crosses (Chapter II). The Mlc2E38 allele has a stop codon at amino acid 10 of the protein coding sequence (Warmke et al., 1992). Thus, in the transformants, the only full-length form of the MLC-2 protein produced is encoded by the introduced gene. The ability of the mutant genes to rescue the Mlc2E38 phenotype was compared to the rescue ability of the wild type Mlc2 gene. P element insertion occurs at multiple sites in the genome and expression of the transformed gene can be subject to position effects, resulting in variable gene expression depending on the genomic insertion site. The effect of insert position can be assessed by analysis of a number of lines with independent insertion sites (Spradling, 1986). Therefore, a number of lines were generated for each of the transformed mutant Mlc2 genes: 3 lines for P[Mlc2S66A], 6 lines for Table 2. Rescue of the dominant flightless phenotype of Mlc2E38 heterozygotes with a transformed copy of the mutant Mlc2 genes. a Canton S, wild type; w ; M ^ 838'6 / TM3,e and w ; Mlc2E3813 / TM6B,Tb, heterozygotes with two different chromosomes carrying Mlc2 null allele. The MIC2838 heterozygotes are flightless because only one copy of the wild type Mlc2 gene is functional (Warmke et al., 1992). The JW1 series are Mlc2E38/+ transformants carrying the wild type copy of the Mlc2 gene described previously by Warmke et a l (1992). The R1T, R2T, and R3T series are transformants carrying mutant Mlc2 genes (P[Mlc2S66A], P[Mlc2S67A], and P[Mlc2866>67Afi, respectively. b Mean flight index with 95% confidence interval. Flight index was determined for both Mlc2E38/+ flies and Mlc2E38/+ flies with a transformed copy of either wild type or mutant Mlc2 gene The flight index was determined using a Benzer cylinder (Benzer 1973). Higher flight index values indicated that the flies are flight competent. Flight impaired flies scored lower values, and flightless flies dropped to the bottom of the cylinder. At least 100 flies of each genotype were flight tested.

90 TABLE 2

Rescue of the dominant flightless phenotype of Mlc2&38 heterozygote with a transformed copy of the mutant Mlc2 gene

Flight Flight Linea Unea Indexb Index*5

Canton S 6.5+ 0.3 R2T1 6.3 ± 0.2 R2T2 6.5 ±0.3 w ;Mlc2E38.e 1.4 ±0.4 R2T3 6.7 ± 0.2 TM3,e R2T4 6.5 ±0.4 w ;Mlc2E38-13 1.9 ±0.6 R2T5 6.9 ±0.2 TM6B,e R2T6 6.5 ± 0.2

JW1[63.U 6.0 ±0.5 R3T1 5.2 ± 0.5 JW1 [40.1) 6.4 ± 0.4 R3T2 6.6 ± 0.3 JW1[18.1J 6.4 ± 0.4 R3T3 6.5 ±0.8 JW1[16.2] 6.1±0.4 R3T4 5.8 ± 0.3 R3T5 5.5 ± 0.4 R3T6 5.5 ± 0.4 R1T1 6.1 ±0.3 R3T7 5.6 ± 0.4 R1T2 6.5 ±0.2 R3T8 5.6 ± 0.4 R1T3 6.5 ±0.2 92

P[Mlc2S67A] and 8 lines for P[Mlc2S66’67A] (see Chapter II for the germline transformation procedure).

Rescue of the dominant flightless phenotype of Mlc2E3a heterozygote with a transformed copy of the mutant Mlc2 gene The mutant Mlc2 genes with residues Ser66 and/or Ser67 substituted with alanines [P[Mlc2S66A], PMlc2S67A] and P[Mlc2S66’67A]) were introduced into Mlc2E38 heterozygotes by genetic crosses, and the ability of the mutant Mlc2 genes to rescue the dominant flightless phenotype of the Mlc2E38 heterozygous flies was determined (Figure 8, Chapter II). The results of the rescue experiments are summarized in Table 2. The Mhc2 null heterozygotes {w ; Mlc2E38,e/TM3, e and w ; Mlc2E38~13/TM6B, e) are flightless, and score low in a flight test (see Chapter II for a description of the flight test). In the Mlc2E38 heterozygotes, only one copy of the wild type Mlc2 gene is functional, resulting in insufficient expression of the MLC-2 protein to allow normal IFM structure (Warmke et al., 1992; this study). It had been shown previously that transformation with a wild type copy of the Mlc2 gene can restore the IFM structure and the dominant flightless phenotype of the Mlc2E38 heterozygotes to wild type level (Warmke et al., 1992). I retested the flight ability of 4 independent lines transformed with wild type Mlc2 genes (generated by Warmke, 1990). These served as positive controls in my experiments as all the flies flew with apparent wild type ability (Table 2). All three m utant Mlc2 genes (P [Mlc2S66A], P[Mlc2S67AJ, P[Mlc2S66’67A]) also restored the flight ability of the Mlc2E38 heterozygotes to a wild type level. All lines with the same transformed gene exhibited essentially the same flight indices, indicating that the effect of chromosomal position was insignificant (Table 2). 93

A single copy of the wild type Mlc2 gene rescued the flightless phenotype of the Mlc2E38 heterozygotes, and restored the IFM structure (Warmke et al., 1992). The mutant genes with alanines substituted for Ser66 and/or Ser67 also rescued the flightless phenotype of the Mlc2E38 heterozygotes. The ability of the mutant MLC-2 proteins to rescue the flightless phenotype of the Mlc2E38 heterozygote also implies that normal IFM structure was formed in these flies, and both the mutant and the wild type MLC-2 proteins contribute to the muscle of the transformants. However, as only the wild type MLC-2 protein can be phosphorylated, these results suggest either this level of phosphorylation is sufficient for flight or that phosphorylation of MLC-2 is not required for flight. To distinguish between these possibilities, I analyzed Mlc2E38 homozygous flies with transformed mutant Mlc2 genes. In these flies no endogenous (wild type) MLC-2 protein is produced, allowing the effect of the mutation on muscle function to be assessed.

Rescue of the recessive lethality Mlc2E38 homozygotes with a transformed copy of the mutant Mlc2 genes Mlc2E38 homozygotes die at the late embryonic stage. The lethality can be rescued by introduction of a transformed copy of the wild type Mlc2 gene (Warmke et al., 1992). I determined if the m utant Mlc2 genes [P[Mlc2S66A], P[Mlc2S67A], P[Mlc2S66’67AJ) also rescued the embryonic lethality of the Mlc2E38 homozygotes. The ability of the transformed genes to rescue the recessive lethality of the Mlc2E38 homozygotes is summarized in Table 3. All four lines transformed with the wild type Mlc2 gene were able to rescue the recessive lethality of the Mlc2E38 homozygotes, albeit with varying efficiencies (Table 3). Similarly, all three classes of the mutant Mlc2 genes [P[Mlc2S66A], P[Mlc2S67A], P[Mlc2s66>67A]) also Table 3. Rescue of the dominant flightless phenotype and viability of Mlc2E38 homo 2ygotes with a transformed copy of the mutant Mlc2 genes. a The JW1 series are Mlc2E38/Mlc2E38 transformants carrying the wild type copy of the Mlc2 gene described previously (Warmke et al., 1992). The R1T, R2T, and R3T series are Mlc2E88/Mlc2E38 transformants carrying mutant Mlc2 genes (P[Mtc2S66A], P[Mlc2S67A], and P [Mlc2S66'67A])f respectively.

Mean flight index with 95% confidence interval. The flight index was determined for Mlc2E38/Mlc2E38 flies with a transformed copy of either wild type or mutant Mbc2 gene. The flight index was determined using a Benzer cylinder (Benzer 1973). Higher flight index values indicated that the flies are flight competent. Flight impaired flies scored lower values, and flightless flies dropped to the bottom of the cylinder. c Percentage of expected number of w ; P [Mlc2]/+ ; Mlc2E38'13/Mlc2E38’e progeny (see cross in Chapter II, Figure 8). d Flight test. Flight indices were not quantitated, but by observation, these flies are clearly flight impaired. There is variability in the degree of flight impairment among the transformed lines. However, none is able to gain altitude, although some lines are capable of gliding over a short distance. Depending on viability, 20-100 flies of each genotype were flight tested.

94 TABLE 3 Rescue of the dominant flightless phenotype and viability of Mlc2E3S homozygotes with a transformed copy of the mutant Mlc2 gene

Linea Flight Indexb viability (%)c

JW1[63.1] 6.7 ± 0.3 146.5 JW1[40.1] d 97.3 JW1[18.U 83.8 JW1[16.2] - 51.1

R1T1 - 42.9 R1T2 - 25.0 R1T3 - 19.0

R2T1 . 44.0 R2T2 - 67.5 R2T3 - 48.8 R2T4 - 77.6 R2T5 20.4 R2T6 98.8

R3T1 —, 38.1

R3T2 - 9.7 R3T3 - 10.4 R3T4 - 3.6 R3T5 - 150.8 R3T6 _ 7.7 R3T7 _ 138.0 R3T8 - 0 96

rescued the recessive lethality of the Mlc2E38 homozygotes to varying degrees (except line R3T8 which failed to rescue, Table 3).

Rescue of the flightless phenotype of the Mlc2E38 homozygotes with a transformed copy of the mutant Mlc2 gene The Mlc2E38 homozygotes rescued with a copy of the transformed genes were tested for their flight ability and the results are shown in Table 3. None of the lines carrying either wild type or mutant Mbc2 transformed genes was capable of flight except for one line, JW1[63.1], which carried a wild type Mlc2 gene (Table 3). This linp is flight conipetent with a flight index comparable to that of wild type (6.7 + 1.3, compared to the Canton S flight index of 6.5 + 1.9, Table 2). This result suggests that in this exceptional line, the single copy of the transformed gene is overexpressed and sufficient MLC-2 accumulates to restore IFM structure (this was confirmed by EM analysis, see below). The other three lines expressing the wild type Mlc2 gene [JW1[40.1], JW1[18.1], and JW1[16.2]) are incapable of flight. In these cases the transformed gene is presumed to be expressed at a normal level (or lower than normal level) resulting in a disruption of IFM structure, due to low accumulation of MLC-2, from the one functional copy of the gene present in the genome. These flies are functionally analogous to Mlc2E38 heterozygotes (Warmke et al., 1992). Therefore, it is also likely that failure of the mutant Mlc2 genes to rescue flight ability of Mlc2E38 homozygotes is likewise due to insufficient accumulation of MLC-2 for normal IFM structure. To further analyze the flies and test these hypotheses the IFM ultrastructure of the transformants was examined by electron microscopy. 97

Examination of the in direct flight muscle ultrastructure of transformants expressing one copy of the mutant Mlc2 genes Two lines expressing one copy of the wild type Mlc2 gene (P[Mlc2+]) were analyzed: JWi [63.1 ] and JW1[40.1 ]. In line JW1[63.1], one copy of P[Mlc2+] rescued flight of Mlc2E38 homozygotes. This line was postulated to overexpress the wild type MLC-2 protein, and restore the IFM structure to normal. As predicted, the IFM ultrastructure of this line appears normal and indistinguishable from that of wild type (Figure 13). As seen in both longitudinal and cross sections, the myofibrils are discrete and well organized. The rigidity of double hexagonal lattice and interdigitating arrays of thick and thin filaments are apparent. However, in line JW I [40.1], one copy of P[Mlc2+] fails to rescue flight of the Mlc2E38 homozygotes. This line has disrupted IFM structure (Figure 14). The IFM phenotype of this line is very similar to that of Mlc2E38 heterozygotes, in which expression from a single wild type copy of the Mlc2 gene results in disrupted IFM structure and flight ability (Warmke et al., 1992). Longitudinal sections show that the myofibrils do not maintain a constant width through out the entire length as seen in wild type. Branching at the periphery and gaps between filaments within the myofibrils are common. The filaments are not compacted due to the altered spatial arrangement of thick and thin filaments, resulting in loss of the highly ordered interdigitating arrays of the myofilaments. Z bands appear wavy and some do not span the width of the myofibrils. Broken or branched Z bands are also apparent. The M lines are also wavy and are often broken. In cross sections, the cylindrical aspect of the myofibrils is lost and they have irregular circumferences. The rigid double hexagonal packing of thick and thin filaments is maintained in the core of the myofibrils but is lost around the periphery as these filaments become more disorganized. Filament packing is not as tight as Figure 13. Electron micrographs of the IFM of transformants carrying one copy of the wild type Mlc2 gene (P[Mlc2+]/+ ; Mlc2E38/Mlc2E38, line JW I [63.1]). (A) Longitudinal section. Myofibrils are intact and run parallel to the longitudinal axis of the myoflber. Each myofibril is made up of well-organized interdigitating thick and thin filaments, and is demarcated by Z lines into sarcomeres with a constant length. Z lines and M lines appear straight and span the entire width of each myofibril. (B) Cross section. Myofibrils appear cylindrical with similar diameters. With in each myofibril, six thin filaments surround each thick filament, giving rise to a rigid double hexagonal pattern. Bars: (A) 1.0 (im, (B) 0.5 Jim.

98 99

V 1 J

l -v % , ’V Vrt1 t i!u^ R > An

t t -taia^ $V ' <111

Figure 13. Figure 14. Electron micrographs of the IFM of transformants carrying one copy of the wild type Mlc2 gene {P[Mlc2+]/+ ; Mlc2E38/Mlc2E38t line JW I[40.1]}. (A) Longitudinal section. Myofibrils do not maintain a constant width through out the entire length, as the spatial arrangement of thick and thin filaments within the myofiber is altered and not as well compacted as seen in wild type. Filaments at the periphery often peel away from the myofibrillar core (arrow), and gaps are occationally seen within the myofiber (arrowhead). Z lines and M lines are wavy, and sometimes are broken or do not span the entire width of the fiber. (B) Cross section. The circumference of the myofibers appear irregular as the organization of thick and thin filaments at the periphery is lost. However, the core region of the fiber is still intact and the double hexagonal organization is maintained. Fusion between myofibers are occasionally seen (arrowhead), and gaps or breaks within the fiber are common (arrow). Bars: (A) 2.0 pm, (B) 0.5 pm.

100 t 101

Figure 14. Figure 15. Electron micrographs of the IFM of transformants carrying one copy of P [Mlc2S66>67A\ {P[Mtc2866’67A]/P[Mlc2S66’67A\ ; Mlc2E38/Mlc2E38, line R3T1). (A) Longitudinal section. Myofibrils are intact and run parallel to the longitudinal axis of the myofiber. Each myofibril is made up of well-organized interdigitating thick and thin filaments, and is demarcated by Z lines into sarcomeres with a constant length. Z lines and M lines appear straight and span the entire width of each myofibril. (B) Cross section. Myofibrils appear cylindrical with similar diameters. With in each myofibril, six thin filaments surround each thick filament, giving rise to a rigid double hexagonal pattern. Bars: (A) 1.0 |im, (B) 0.5 pm.

102 103

'taw**’”

nr ; r:'.v /•'. V , ,

Figure 15. Figure 16. Electron micrographs of the IFM of transformants carrying one copy P [Mlc2S66'67A] (P[JVflc2S66*67A]/+ ; Mlc2E38/Mlc2E38, line R3T4). (A) Longitudinal section. Myofibrils do not maintain a constant width through out the entire length, as the spatial arrangement of thick and thin filaments within the myofiber is altered and not as well compacted as seen in wild type. Filaments at the periphery often peel away from the myofibrillar core (arrow), and gaps are occationally seen within the myofiber (arrowhead). Z lines and M lines are wavy, and sometimes are broken or do not span the entire width of the fiber. (B) Cross section. The circumference of the myofibers appear irregular as the organization of thick and thin filaments at the periphery is lost. However, the core region of the fiber is still intact and the double hexagonal organization is maintained. Fusion between myofibers are occasionally seen (arrowhead), and gaps or breaks within the fiber are common (arrow). Bars: (A) 1.0 pm, (B) 1.0 pm.

104 105

mtr&h.

n

Figure 16. 106 that seen in wild type because filament loss and gaps are occasionally seen within the myofibrils. In addition, fusion between myofibrils is also apparent. The IFM ultrastructure of two lines carrying one copy of P[Mlc2S66>67A] were examined {R3T1 and R3T4, Figure 15 and 16, respectively). Both lines fail to rescue the flightless behavior of the Mlc2E38 homozygotes, but only one has abnormal IFM structure. The IFM ultrastructure of R3T1 appears normal, suggesting that the transformed mutant gene is over expressed, producing sufficient MLC-2 protein to allow normal muscle structure. This shows that the mutant MLC-2 protein is capable of assembly into muscles that lack any wild type MLC-2. However, these flies are flightless. The impaired flight is most likely due to lack of MLC-2 phosphorylation at Ser66 and Ser67. This is analyzed in detail below in a number of lines expressing 2 copies of the mutant genes. R3T4 exhibits abnormal muscle structure similar to that of flies expressing one copy of P[Mlc2+] (line JW I [40.1], Figure 14). The failure of the mutant gene {P[Mlc2S66-67A]) in line R3T4 to rescue the flightless behavior of the Mlc2E38 must in part be due to insufficient expression of the MLC-2. In summary, in most cases introducing one copy of a transformed wild type or mutant {P[Mlc2S66'67A]) Mlc2 gene into MIc2E3& homozygotes failed to rescue flight (see above and Table 3). As expected the IFM structure is disrupted in these flies (two lines tested). The underlying cause is presumed to be that insufficient MLC-2 is expressed from a single gene to restore the IFM structure. However, IFM structure can be restored in Mlc2 homozygotes, provided the transformed gene is expressed at higher levels than the endogenous Mlc2 gene. However, in these cases only the wild type gene can restore flight suggesting that phosphorylation at Ser66 and/or Ser67 is also important for flight. 107

Rescue of flight ability of Mlc2E38 homozygotes with two transformed copies of the mutant Mlc2 gene Since the flightless phenotype of Mlc2E38 homozygotes could not be rescued with one copy of the P[MIc2+] gene in most of the lines described above, it is clear that gene dose is important for normal muscle structure and flight. To test if IFM structure and flight can be restored by increasing the gene dose, two copies of either wild type [P[Mlc2+]) or mutant Mlc2 genes [P[Mlc2S66A], P[M1c2S67A], P[MIc2S66<67A]) were introduced into Mlc2E38 homozygotes. Mlc2E38 homozygotes expressing 2 copies of P[Mlc2+] (lines JW 1[40.1], JW1[18.1], and JW1[16.2]) flew as well as wild type flies (Table 4). However, the flight ability of Mlc2E38 homozygotes expressing 2 copies of the mutant Mlc2 genes varied depending on which mutant gene they expressed. Flies expressing two copies of P[Mlc2S67A] were flight competent with flight indices comparable to that of the wild type (Table 4). Transformants expressing two copies of P[Mlc2S66A] exhibited significantly reduced flight ability, and those expressing two copies of P[Mlc2S66'67A] were essentially flightless (Table 4). All lines within each strain exhibit essentially the same flight indices, suggesting that the position effects due to insertion site are minimal. The flightless phenotype of Mlc2E38 heterozygotes and the requirement for two copies of the wild type Mlc2 gene (P [Mlc2+fl to confer flight to Mlc2E38 homozygotes (except in one line in which the transformed gene is overexpressed) indicate that gene dose is im portant and a certain level of MLC-2 is necessary for the formation of normal IFM structure and flight. Two copies of the m utant P [Mlc2S67A] gene were sufficient to rescue the flightless phenotype of the Mlc2E38 homozygotes to wild type or near wild type levels (Table 4). However, none of the lines expressing the m utant genes, P [Mlc2S66A] and P[Mlc2S66’67A], restored the flightless Table 4. Rescue of the dominant flightless phenotype of Mlc2E38 homozygotes with two transformed copies of the mutant Mlc2 genes. a The JWI series are Mlc2E38/Mlc2E38 transformants carrying two wild type copies of the Mlc2 genes described previously (Warmke et a l, 1992). The R1T, R2T, and R 3Tseries are Mlc2E38/Mlc2E88 transformants carrying mutant Mlc2 genes (P [Mlc2S66A], P[Mlc2S67A], and P [Mhc2S66<67AJ\, respectively. b Mean flight index with 95% confidence interval. The flight index was determined for Mlc2E38/Mlc2E38 flies with two transformed copies of either wild type or mutant Mlc2 genes. The flight index was determined using a Benzer cylinder (Benzer 1973). Higher flight index values indicated that the flies are flight competent. Flight impaired flies scored lower values, and flightless flies dropped to the bottom of the cylinder. Depending on viability, 20- 100 flies of each genotype were tested. c Less than 20 flies were tested due to low viability.

108 TABLE 4 Rescue of the dominant flightless phenotype of Mlc2E38 homozygotes with two transformed copies of the mutant Mlc2 gene

Llnea Flight Indexb

JW1163.1] 5.6 + 0.5 JW1[40.1] 5.1 ± 1.4 JW 1[18.1] 6.9 ± 0.3 JW1J16.2] 5.9 ± 0.5

R1T1 3.3 ± 0.7 R1T2 4.4 ± 0.8 R1T3 3.2 + 0.5

R2T1 5.7 ± 0.8 R2T2 6.6 ± 0.3 R2T3 5.9 ± 0.6 R2T4 5.4 ± 0.6 R2T5 1.0 ± 1.5 R2T6 6.8 ± 0.7

R3T1 0.2 ± 0.2 R3T2 0.2 + 0.1 R3T3 0.1 ±0.1 R3T4 1.0+ 0.3 R3T5 0.4 ± 0.2 R3T6 0.0±0.0C R3T7 0.7 + 0.2 R3T8 0.1 ±0.2 110 phenotype of the MIC2838 homozygotes to near wild type (Table 4). It is unlikely, but cannot be excluded, that none of these lines produced wild type levels of MLC-2, since at least 3 lines from each strain were examined (3 lines for P[Mlc2S66A] and 8 lines for P[Mlc2S66’67A]). The inability of individuals from these lines to fly suggests either that their IFM structure is abnormal because they accumulate insufficient MLC-2 or that they are flightless because the mutant MLC-2 cannot be phosphorylated. To distinguish these possibilities, the IFM of transformants was examined by electron microscopy.

Examination of the indirect flight muscle ultrastracture of transformants expressing two copies of the mutant Mlc2 genes Two Mlc2E38 homozygous lines expressing either two copies of P[Mlc2S66’67A] (lines R3T4 and R3T7 ; Figure 17) or two copies of P[Mlc2+] (lines JW1[40.1] and JW1[18.1]\ Figure 18) were analyzed. The IFM ultrastructure of transformants carrying two copies of P[Mlc2s66'67A] (line R3T4; Figure 17) appears normal and indistinguishable from that of transformants carrying two copies of P[Mlc2+] [JW1[40.1JI (Figure 18). In longitudinal sections, myofibrils are intact and run parallel along the longitudinal axis of the myofiber. Large mitochondria and collections of glycogen granules occupy much of the space between myofibrils, each of which is composed of well organized arrays of interdigitating thick and thin filaments. The sarcomeres are well demarcated by Z lines and appear regular with an average length of 3.0 ±0.1 pm, which is not significantly different from that of transformants expressing P[Mlc2+] ( line JW1[40.1], 3.2 ± 0.6 pm) or that of Canton S (3.5 ± 0.1) (Warmke, 1992). Both the Z lines and M lines are straight and well-defined, spanning the entire width of the myofibril. Cross sections show the well-defined cylindrical shape of the myofibers which have similar diameters. Within each myofibril, thick and Figure 17. Electron micrographs of the IFM of transformants carrying two copies of P Mlc2S66’67A] {P[Mlc2S66’67A]/P[Mlc2S66’67A]', Mlc2E38/Mlc2 E38, line R3T4). (A) Longitudinal section. Myofibrils are intact and run parallel to the longitudinal axis of the myofiber. Each myofibril is made up of well-organized interdigitating thick and thin filaments, and is demarcated by Z lines into sarcomeres with a constant length. Z lines and M lines appear straight and span the entire width of each myofibril. (B) Cross section. Myofibrils appear cylindrical with similar diameters. With in each myofibril, six thin filaments surround each thick filament, giving rise to a rigid double hexagonal pattern. Bars: (A) 1.0 pm, (B) 0.5 pm.

111 112

Figure 17. Figure 18. Electron micrographs of the IFM of transformants carrying two copies of the transformed wild type Mlc2 genes (P[Mlc2+]/P[Mlc2+] ; M \c2^IM \c2^, line JW1[40.1]). (A) Longitudinal section. Myofibrils are intact and run parallel to the longitudinal axis of the myofiber. Each myofibril is made up of well-organized interdigitating thick and thin filaments, and is demarcated by Z lines into sarcomeres with a constant length. Z lines and M lines appear straight and span the entire width of each myofibril. (B) Cross section. Myofibrils appear cylindrical with similar diameters. Within each myofibril, six thin filaments surround each thick filament, giving rise to a rigid double hexagonal pattern. Bars: (A) 5.0 pm, (B) 5.0 pm.

113 114

m m *

■'••S'-

Figure 18. Figure 19. Electron micrographs of the IFM of transformants carrying two copies of P[Mlc2S66A].{P[Mlc2S66A]/P[Mlc2S66A] ; Mlc2E38/Mlc2E38, line R1T3). (A) Longitudinal section. Myofibrils are intact and run parallel to the longitudinal axis of the myofiber. Each myofibril is made up of well-organized interdigitating thick and thin filaments, and is demarcated by Z lines into sarcomeres with a constant length. Z lines and M lines appear straight and span the entire width of each myofibril. (B) Cross section. Myofibrils appear cylindrical with similar diameters. With in each myofibril, six thin filaments surround each thick filament, giving rise to a rigid double hexagonal pattern. Bars: (A) 1.0 pm, (B) 1.0 pm.

115 116

t ii t }4 tir , hi, \ i j '' i >i * (Ji) /*U % 'V *V- <»« ■* "v *Jls //• /• Wv< &■** ^MKP® |# X <* i v\* * > #x Y *>*i . n* - f ’ H.’V f „__ V F •‘■'*- ** =-. X i^ii^.* H V f t n T.

"Sent"*

# f ' X v

-v « <■ '» r ^ o a r3f1^fK Ay^> v' f-'i'f,,-'. V . T. VV/ % ‘ v". 8% j< ^ j' r - , V f» I Sw§g^.\-v.*.: \ :r ' *K! .

Figure 19. Figure 20. Electron micrographs of the IFM of transformants carrying two copies of P [Mlc2S67A] {P[Mlc2S67A]/P[Mlc2S67A] ; Mlc2E38/Mlc2E38, line R2T2). (A) Longitudinal section. Myofibrils are intact and run parallel to the longitudinal axis of the myofiber. Each myofibril is made up of well-organized interdigitating thick and thin filaments, and is demarcated by Z lines into sarcomeres with a constant length. Z lines and M lines appear straight and span the entire width of each myofibril. (B) Cross section. Myofibrils appear cylindrical with similar diameters. With in each myofibril, six thin filaments surround each thick filament, giving rise to a rigid double hexagonal pattern. Bars: (A) 1.0 pm, (B) 1.0 pm.

117 118

&PK' '* tu? * ^'cn & i

^#t|lfS|i■i ', ;«-

Figure 20. 119 thin filaments are arranged in a double hexagonal lattice with six thin filaments surrounding each thick filament. Transformants with P [Mlc2S66-67A] (R3T4) have -33 thick filaments across the diameter of the myofibrils, whereas myofibrils of transformants with P[Mlc2+] JW1 [40.1] and Canton S measure -34 and 34-36 filaments, respectively. One representative line from each of the other two mutant strains (line R1T3, expressing P [Mlc2S66A], and line R2T2, expressing P[Mlc2S67AJ) were also analyzed. The IFM ultrastructure of these lines is also normal (Figure 19 and 20). The average sarcomere length of lines expressing P[Mlc2S66A] (line R1T3) and P[Mlc2S67A] (line R2T.2) is 2.8 + 0.1 and 2.8 + 0.1 pm, respectively, whereas the average filament counts across the myofibrils are 31 and 37, respectively. In summary, expression from two gene copies, either wild type or mutant, apparently yields sufficient MLC-2 protein for muscle assembly. Substituting Ser66 and/or Ser67 for Ala appears to have no effect on the MLC-2 protein structure, as the mutant proteins assemble and give rise to normal muscle structure. Despite normal IFM structure, flies expressing mutant MLC-2 proteins are flight impaired. The flight impairment is most likely to be due to lack of MLC-2 phosphorylation at Ser66 and/or Ser67.

Examination of the MLC-2 protein pattern in transformants carrying the mutant Mlc2 gen es As described in Chapter III, multiple phosphorylated variants of the MLC-2 protein accumulate in vivo. These variants appear to differ in the number of phosphate groups incorporated. Disruption of two phosphorylation sites is expected to cause a reduction in the num ber of MLC-2 phosphorylated variants (by two). To determine if this predicted reduction in the number of variants can be detected, three independent lines expressing two copies of Figure 21. Pattern of MLC-2 accumulation in transformants; Mlc2+ homozygotes (P[Aflc2+]/P[Mlc2+] ; M\c2^3/Mlc2^38j (a) and MIc2S66,67A homozygotes (P[Mlc2S66>37A]/p[Mic2S66,67A] . Mlc2E38/ Mlc2E38) (h). Total thoracic proteins from transformants carrying two copies of the wild type Mlc2+ genes (a), or the mutant Mlc2 genes (b) were separated by 2-D PAGE and silver-stained. The pH gradient of the first dimension is 4-7, and the polyacrylamide composition of the second dimension is 12.5%. Spot 148 is the unphosphorylated MLC-2 variant. Spots 149 and 138 are the phosphorylated MLC-2 variants (Chapter III). Note that the unphosphorylated MLC-2 (arrow) accumulates to a higher level in Mlc2$66,67A than in Mlc2+ flies.

120 121

Figure 21. 122

P[Mlc2S66’67A] were analyzed for their MLC-2 accumulation pattern (lines R3T1, jR3T4, and R3T7 ; one shown in Figure 21). Transformants expressing two copies of P[Mlc2+]. (line JW1[63.1]) were analyzed as controls (Figure 21). Total thoracic homogenates were separated by 2-D PAGE, followed by silver-staining. In addition to unphosphorylated MLC-2 (spot 148 ), there appear to be at least 18 spots corresponding to MLC-2 phosphorylated variants (10 spots corresponding to the protein 149 series, and 8 spots corresponding to the protein 138 series). Surprisingly, the number of MLC-2 phosphorylated variants in lines expressing the m utant (P[Mlc2s66>67A]) or the wild type gene ( P[Mlc2+]) appears to be the same but minor phospho-variants not detected in the gels could be specifically deleted. However, there is at least a three-fold increase in the accumulation of the unphosphorylated MLC-2 (protein 148) in the mutant lines compared to the wild type lines (Figure 21). The lack of evidence for discrete loss of specific phospho- variants is surprising but may be explained as follows. If there are multiple phosphorylation sites in Drosophila MLC-2, and this seems likely (Chapter III), individual molecules could be phosphorylated at different numbers of sites at any one time. Phosphorylation at specific sites other than the two conserved serines (Ser 66 and Ser67) may not be critical for function. In the double mutants, the population of MLC-2 molecules shows a reduced overall level of phosphorylation, because Ser66 and Ser67 cannot be phosphorylated.

W ingheat analyses of transformants expressing mutant MLC-2 p rotein The flightless behavior of many muscle mutants has been shown to correlate with reduced wingbeat frequencies [fwb)- Flies heterozygous for the Mlc2 null allele (Mlc2E38, Warmke et al., Table 5 The effects of a mutant MLC-2 on wing beat frequencies

Genotype Wing beat frequencies (Hz)

Canton 8 (wild type) 218+19

M c2B3 S/M lc2E 3 8 ND

Ptmc2+l/P[Mic2*l; MLc2e 3 8 /MLc2E 3 8 234+10 line JW1 (60.1)

P[BSlc2S8 6 Al/lM c288eA] . Mlc2E3 8 /Mlc2E3B 179+16 line R2T3

PfM c28 67A]^Mic2S87A] . M Ic 2 ^ 8 /Mlc2E 3 8 203±12 line R3T2 p[Ml4*?rS6 6 t 67Aj/inff(.yrS66,67Aj . Mlc2E3 8 /Mlc2E 8 8 169+22 line R3T1 124

1992), or the IFM-specific TM1 null allele ( Ifm(3)3, Molloy et al., 1992), have disrupted IFM structure, resulting in a flightless phenotype coupled with a ~30% reduction of fw b compared to wild type. However, two missense mutations of the Act88F gene, G368E and E316K (Drummond et al., 1990), exhibit minimal structural perturbation of the IFM, but these flies produced a ~20% reduction of fw b (G368E) or no wingbeat ( E316K). To determine if flight impairment exhibited by transformants expressing the mutant MLC-2 proteins with disrupted phosphoiylation sites correlates with a reduction of fwb> one representative line from each of the three mutant strains was examined (lines R1T3, R2T2, and R3T1 carrying two copies of P[Mlc2s66A], P[MIc2S67a], and P[Mlc2S66’67A], respectively), and compared to a line carrying two copies of P [Mlc2+] (line JW1[63.1Jj. Wing beat analysis and single fiber mechanics (shown below) were performed by Dr. David Maughan's group at The University of Vermont. As shown in Table 5, fw b appears to correlate well with flight ability in these transformants. Flies expressing P[Mlc2S66’67A] (line R3T1), which are effectively flightless (flight index 0.2 + 0.8), exhibit a fw b of 173 ± 11 Hz, a ~20% reduction compared to flies expressing P[Mlc2+] (line JW1[63.1J). Flies carrying P[Mlc2S66A] (line R1T3), which fly, albeit with reduced ability (flight index 3.2 + 2.9), show an improved fw b of 192 + 7 Hz (a ~10% reduction). Flies expressing P [Mlc2S67A] (line R2T2) exhibit a flight index of 6.8 ±1.7 and a fw b of 208 ± 16 Hz (a ~3% reduction), which is not significantly different than flies expressing P[Mlc2+] (line JW1[63.1] flight index 5.6 ± 2.6; fw b 215 ± 10 Hz).

Mechanical analyses of skinned single IFM fibers expressing m utant MLC-2 protein Mechanical analysis of isolated muscle fibers serves as a tool to understand how muscles function, and furthermore, how a Figure 22. Sinusoidal analysis of single IFM fibers from transformants Mlc2+, Mlc2 S6 6 At Mlc2^^7A, and Mlc2S66,67A homozygotes. These flies have a Mlc2^^/Mlc2E38 genetic background. IFM from transformants carrying two copies of Mlc2+ (a), Mlc2S67A (b)t Mlc2S66A (c) and ^jiC2S66f67A ((j)> were analysed for their active stretch-responses. Sinusoidal analysis was conducted by applying length oscillations of 0.25% fiber length and 1-1000 Hz. The active response of the fibers are presented as a Nyquist plot. Note that there is a graded reduction of the amplitude of active response in the mutant fibers (b,c,d) compared to wild type (a).

125 126 ;o r* O 3] O) 2 I" o CO s .+ Q> _ / s 8 8 ______8

8 X} O) CD r~ O Co O O) O) O CO O) r- co > .2

8 \ \

JK o ;= i 8 8

£ m o 69 0 Q. C 7T C 0 z

£ JC o o g 0 a c c M Z 2

8 Viscous Modulus (kN/m*) Viscous Modulus (kN/m*)

k Figure 22. I 127 mutation of muscle protein effects muscle function. To extend our understanding of how the substitution of Ser66 and/or Ser67 with alanines effects flight and wing beat frequency, the kinetics of contraction of the mutant muscle fibers were determined. The IFM of Drosophila is asynchronous, i.e., each nerve impulse results in multiple cycles of contraction. Asynchronous muscles are only partially activated by Ca2+. To be fully activated they need to be stretched by the order of ~3% above their resting length (stretch-activation). After a brief delay, there is a drastic increase in both tension and ATPase activity which is maintained (Pringle, 1978). The rate constant for the delayed rise in tension has been shown to be directly proportional to the wing beat frequency of Drosophila (Molloy et al., 1987). These muscles can also be fully activated when the length of the activated fibers is sinusoidally oscillated over a restricted frequency range (Steiger & Ruegg, 1969; Lund et al., 1988). Recently mechanical analyses of small muscles (~ 1 mm long) such as Drosophila IFM have been developed (Molloy et al., 1987). Dissected skinned single fibers are mounted in an apparatus and oscillated over a range of frequencies, and the force production of the fiber in response to the oscillatory length change is recorded. Sinusoidal tension response is expressed in terms of two components; one is in phase with the length change (elastic stiffness modulus), and the other is 90 0 out of phase (viscous stiffness modulus). A useful format for presenting responses of the fiber over a broad range of applied oscillatory frequencies is the Nyquist plot (Kawai and Brandt, 1980) (Figure 22). In the Nyquist plot, the elastic stiffness modulus, plotted on the x axis, reflects the number of attached crossbridges or muscle stiffness. The viscous stiffness modulus, plotted on the y axis, corresponds to the level of active crossbridge cycling. The viscous stiffness modulus is negative when the fiber performs work on the 128 apparatus, and the frequency at which the largest negative viscous stiffness modulus (bottom frequency) occurs is the frequency at which maximum work is performed. It is at this frequency that the process responsible for driving the wing beat occurs in vivo. Fibers from transformanted lines expressing two copies of P[Mlc2S66A] (line R1T3), P[Mlc2S67A] (iine R2T2), and P[Mlc2S66-67A] (line R3T1) showed a graded reduction of the sizes of the Nyquist plots and -Ev, corresponding to a graded reduction in the amplitude of stretch-activation response and the work output generated by the fibers (Figure 22). The largest reduction of active response was observed for the M lc2S66<67A fibers and the least for the Mlc2S67A fibers. These resiilts agree with those of the flight and wing beat analyses, and suggest that the reduced active stretch-response of the mutant fibers is the primary cause of a reduction in f Wb and flight ability. The graded reduction in the amplitude of response of the mutant fibers is also consistent with in vitro phosphorylation site mapping data, which suggest that phosphorylation at Ser66 is functionally more important than phosphorylation at Ser67.

D iscussion

Combined genetic, cytological and biophysical evidence demonstrates that phosphorylation of the Drosophila MLC-2 protein at Ser66 and Ser67 is required for normal IFM function. Substitution of either or both serines with alanines results in graded flight impairment without an effect on normal muscle structure, suggesting that the flight defect observed is not due to disrupted muscle structure but is due to lack of phosphorylation at these sites. Substitution at Ser66 produces a more severe flight defect than substitution at Ser67, and dual substitution results in an essentially flightless phenotype. The results of single fiber 129 mechanics parallel those of flight testing, and show that flight impairment correlates with reduced wingbeat frequencies and power output of the IFM fibers. These results suggest that phosphorylation at Ser66 is functionally more important than phosphorylation at Ser67, and phosphorylation at both residues is required for maximal flight ability. The role of MLC-2 phosphorylation in muscle contraction demonstrated in this study can be related to a model of the myosin subfragment-1 (SI) three dimensional structure (Rayment et al., 1993a; Xie et al., 1994). The x-ray crystal structure of the myosin subfragment-1 (SI) has been determined at 2.8 °A resolution for both chicken skeletal muscle myosin (Rayment et al., 1993a) and scallop myosin (Xie et al., 1994). MLC-2 of both the chicken skeletal and scallop muscle was found to wrap around the C- terminal a-helical tail of the SI heavy chain, with an opposite polarity to that of the heavy chain. Interaction with SI occurs through both the N- and C-terminal domains of the MLC-2 molecule, and is stabilized by hydrophobic interactions. The scallop regulatory domain (which contains the specific Ca2+ binding site) and the N-terminus of the chicken skeletal muscle MLC-2 (which contains the phosphorylation sites) are positioned close to the flexible "hinge" region thought to allow myosin head to undergo bending to interact with actin during crossbridge cycling (Harrington and Rodgers, 1984). Thus, a conformational change of MLC-2 caused either by Ca2+ binding to the regulatory domain (scallop) or phosphorylation of MLC-2 (chicken skeletal muscle) possibly induces movements of proteins domains in the myosin head, through the hinge region, resulting in activation of MgATPase activity in the motor domain. According to Rayment et al. (1993a) and Xie et al. (1994), there is a ~40 ° bend in the middle portion of the long SI heavy chain a-helix, and a sharper bend of -60 ° near the C-terminus. The 60 ° bend positions all 130 hydrophobic residues in this segment of MHC in the hydrophobic core formed by the MLC-2. The 40 ° bend is thought to be involved in transducing the conformational change of the MLC-2 to the motor domain in the myosin head. In addition, other flexible regions may be involved in the internal domain movements of the SI myosin required for activation of the ATPase activity. Phosphorylation of MLC-2 may be involved in activation of the myosin ATPase activity by inducing a conformational change of the MLC-2, which, in turn, alters the myosin head conformation. Alternatively, phosphorylation may cause a neutralization of the electrostatic interaction between the myosin head (on which the MLC-2 binds) and the thick filament backbone, causing the myosin head to be released from the backbone. Biochemical and physical evidence from studies of vertebrate skeletal muscles suggest that the myosin head is held on the thick filament surface when MLC-2 is in the nonphosphorylated state (Sweeney et al., 1993). Phosphorylation of MLC-2 results in neutralization of electrostatic interaction between the MLC-2 and the thick filament, resulting in a release of the myosin head from the thick filament backbone. Therefore, the primary effect of MLC-2 phosphorylation appears to be to increase the effective number of crossbridges, possibly by allowing the myosin head to assume a conformation "primed" for actin binding. The effect of disrupting the Drosophila MLC-2 phosphorylation sites can be explained using either of the two models described above. Lack of phosphorylation possibly prevents MLC-2 from undergoing a conformational change, which, in turn prevents internal movement of the SI domain necessary for activation of the ATPase activity. Alternatively, without MLC-2 phosphorylation, the SI domain may be permanently held to the thick filament backbone, reducing the ability of the myosin head to bind to actin. 131

Except for a N-terminal extension of -50 residues, the Drosophila MLC-2 protein is conserved with MLC-2 of other species through out its entire length, and thus, presumed to interact with the SI heavy chain in a similar fashion to that of the chicken skeletal muscle myosin. Although the conserved MLCK phosphorylation sites of the Drosophila MLC-2 (Ser66 and Ser67) are more distant from the N-terminus of the protein than that of the chicken skeletal muscle (Seri3), it is possible that these residues are similarly positioned relative to the myosin heavy chain, and thus phosphorylating these residues could produce similar effect in regulating the myosin MgATPase activity. It is striking that only the IFM, and not other Drosophila muscle types, is affected by the lack of MLC-2 phosphorylation at Ser66 and Ser67. As Mlc2 is a single copy gene with no evidence of alternative splicing, it is unlikely that another MLC-2 isoform is active in these other muscle types. The differential requirement for MLC-2 phosphorylation appears to correlate with functional differences among these muscle types. Phosphorylation of MLC-2 may be required to enhance the high rate of actin-myosin interaction necessaiy for rapid contraction of the IFM. Although phosphorylation of MLC-2 at Ser66 and Ser67 is necessaiy for IFM function, phosphorylation at these two residues alone may not be sufficient for normal IFM function in the absence of other MLC-2 modifications. Phosphorylation at additional sites by MLCK, or by other kinases, as well as other post-translational modifications, may be required for full-activation of IFM contraction. The Drosophila IFM appears to be an excellent system in which to study the role of muscle protein or protein domains in muscle structure and/or function. The IFM structure is very sensitive to the amount and form of muscle proteins, thus the intrinsic effect of a mutation on muscle function can be distinguished from the effect of the mutation on muscle structure. 132

Flight mutants described here are the first to exhibit functional defects without the associated effect of disrupted muscle structure. The informative results I generated altering these two sites suggest a functional analysis scanning the entire MLC-2 protein using similar methods would provide insight into other open questions of how different MLC-2 domains contribute to the regulation of IFM function. CHAPTER V

CLONING OF A DROSOPHILA MTOSIN LIGHT CHAIN KINASE GENE

Introduction

I have shown Drosophila MLC-2 is phosphoiylated to produce multiple phospho-variants (Chapter III). Two sites (Ser66 and Ser67), which align with the conserved myosin light chain kinase (MLCK) phosphoiylation sites from other species, are known to be phosphoiylated by a heterologous MLCK in vitro (Graham, 1992). Both these sites appear to be required for normal function of the IFM because m utant MLC-2 with alanines at these positions causes a flightless phenotype (Chapter IV). Ser66 is phosphoiylated to a greater extent in vitro and appears to have a more significant role in vivo (Graham, 1992; Chapter IV). In this chapter I describe the cloning of Drosophila MLCK cDNAs as a first step towards testing whether these conserved serines in MLC-2 are phosphoiylated by a Drosophila MLCK. Identification of a Drosophila Mick will also allow a genetic analysis of the gene which will be important for understanding its role in muscle function. MLCKs are characterized by their high degree of substrate specificity and dependence on Ca2+/calmodulin for function. Except for the Dictyostelium, and possibly the Physarum MLCKs, all MLCKs characterized to date are activated by Ca2+/calmodulin (see

133 134 review by Tan et al., 1992). MLCK catalyzes phosphorylation of MLC-2, resulting in initiation of contraction in smooth muscle, or potentiation of contraction in skeletal muscle (Figure 23). MLCK does not phosphoiylate any other protein substrates to a significant level. MLCKs from different species and tissues display a large difference in molecular weight, primary structure, substrate specificity, and kinetic properties. Generally, the MLCKs fall into two classes: skeletal muscle MLCK and smooth muscle MLCK. The sizes of skeletal muscle MLCKs range from ~68 kD (human) to ~150 kD (chicken), whereas those of smooth muscles range from ~130 kD (hog) to -155 kD (steer) (see review by Stull et al., 1986). Several invertebrate MLCKs have been purified, and have been shown to be smaller than the vertebrate counterparts ( Limidus , 37 and 39 kD; Dictyostelium, 30 kD). Several cDNAs encoding MLCK have been cloned and sequenced. The deduced protein sequences show that there is limited homology between the skeletal muscle and smooth muscle MLCKs. The cDNA sequence of the rabbit skeletal muscle MLCK is -70% identical to the rat skeletal muscle MLCK, with -90% identity in the region corresponding to the catalytic core and the calmodulin-binding domain (Roush et al., 1988). However, the coding sequence for the catalytic domain of the chicken gizzard smooth muscle MLCK cDNA shares only 54% identity with those of rat and rabbit skeletal muscle MLCKs (Olson et al., 1990), and no significant homology was detected elsewhere. The invertebrate MLCKs appear to be less well conserved with the vertebrate counterparts, since only a -35% homology is observed between the sequences encoding the catalytic domains of the Dictyostelium MLCK and the chicken gizzard MLCK (Tan and Spudich, 1990). MLCK is activated by calmodulin, a ubiquitous protein of 16.7 kD that binds up to 4 Ca2+ ions. In the absence of Figure 23. General scheme for myosin phosphorylation in skeletal and smooth muscles. Myosin light chain kinase (MLCK) is activated by the binding of a Ca2+/calmodulin complex. The activated MLCK phosphorylates the regulatory light chain of myosin. The reverse reaction is catalysed by protein phosphatases. Phosphorylation of smooth muscle regulatory light chain initiates contraction, whereas phosphorylation of skeletal muscle protein potentiates contractile force. MLCK, myosin light chain kinase; CaM, calmodulin (Kemp and Stull, 1990).

135 4C a2* - y ' Calmodulin

« 2* ^ Myosin Light Ca 4 *CaM - ' y - Chain Kinase | Smooth Ca2*A * CaM • MLCK Muscle * ► Contraction Myosin Myosin © — ^ Twitch Phosphatase Striated Pot0ntiation Muscle

Figure 23. Ca2+/calmodulin, the MLCK is inactive, but it becomes activated when it is bound by Ca2+/calmodulin. The number and the spatial arrangement of basic residues of a sequence, located C-terminal to the MLCK catalytic core, and overlapping with the calmodulin binding domain, is similar to that of the MLCK recognition sequence on MLC-2 itself (Figure 24 and Figure 25; Kemp et al., 1987). This region of MLCK has been termed "the pseudosubstrate region" because of the similarity to its substrate sequence (in MLC- 2). It has been proposed that MLCK activity is regulated by autoinhibition mediated by the pseudosubstrate sequence (Kemp et al., 1987, Pearson et al., 1988, Pearson et al., 1991, and Ito et al., 1991). According to the model, the pseudosubstrate sequence occupies the catalytic domain of MLCK in the absence of Ca2+/calmodulin. When Ca2+/calmodulin binds the calmodulin binding domain, adjacent to the pseudosubstrate sequence, the catalytic domain is exposed for interaction with other substrates (MLC-2) (Figure 24). Modeling the structure of MLCK based on the x ray structure of cAMP-dependent protein kinase (cAK), with which it shares 30 % amino acid identity in the catalytic region, suggests the "intrasteric model" for regulation of MLCK is structurally feasible (Knighton et al., 1992). The model demonstrates that a series of electrostatic interactions occur between the acidic residues of MLCK and basic residues of the substrate. The results of previous kinetic and biochemical studies suggested these regions interact and the interaction is required for enzyme activity (Kemp et al., 1983). In addition, the model shows that there are a large number of possible electrostatic interactions between the catalytic core and a connecting peptide. Interaction of these regions would bring the catalytic core and the pseudosubstrate sequence into proximity, explaining the unusually tight binding of the pseudosubstrate sequence to the catalytic domain. Figure 24. Hypothetical model for regulation of myosin light chain kinase activity by pseudosubstrate sequence and Ca2+/calmodulin. (a) Amino acid sequence alignment of substrate (MIX-2) and pseudosubstrate sequences for smooth muscle myosin light chain kinase (sm MLCK). The alignment is based on the substrate phosphorylation site (designated P). Each residue is assigned a position relative to the phosphorylation site. Note the similarity in the spatial arrangement of basic residues in the substrate and pseudosubstrate sequences. Basic residues involved in recognition by MLCK kinase are shaded, and the interacting acidic residues in MLCK are indicated above each basic residue (Knighton et al., 1992). (b) A model for regulation of myosin light chain kinase (MLCK) activity by pseudosubstrate sequence and Ca2+/calmodulin. In the absence of Ca2+/calmodulin, the MLCK activity is depressed by the binding of pseudosubstrate sequence to the catalytic domain. Binding of Ca2+/calmodulin to the calmodulin binding site, which overlaps with the pseudosubstrate sequence, results in a structural change that releases the pseudosubstrate from the catalytic site enabling the catalytic site to interact with the substrate (Kemp et al., 1987).

138 139

Asp7' 9 Enzyme contacts PO,

M IX (1-22) (SabAratt)

MLCK (787 -808 ) (Pmdorabarate) sXKXD'jlB(MYKYKXYYMYAlOGM3rwYQttKirTYGYHYAYvYR 18 -11 -16 ■ \S -U .13 -12 .10 -9 4 -1 -6 •5 4 •1 •3 1 I ■i *i P P P P P P PP P P P P P P p P PP

b

UytilA LlQht Ch«l« (■ SSKRAKAKTTKKRPOAATSNVPA -4

HO-C

CaU

HO-C

Figure 24. Figure 25. Organization of structural domain of myosin light chain kinase. The large box represents the conserved catalytic domain with the ATP binding domain (hatched). The regulatory domain includes the pseudosubstrate and the calmodulin binding sequences. Note the overlap between the pseudosubstrate and the calmodulin binding sequences. Residue numbering is according to chicken smooth muscle MLCK (Knighton et al., 1992).

140 Connecting peptide CalmocUtn bindng

L^‘^KDTKm NMEAKKLS787 KDRMKKYMARRKWQKTGHAV^Cf7RAIGRL stnMLCK Pseudosufaaaate

5 7 3 7 7 4

Figure 25. 142

Studies using synthetic peptides have provided information about substrate sequence requirements for MLCK (Figure 24, Kemp and Stull, 1990), Clusters of basic residues, N-terminal of the MLC-2 phosphoiylation site, have been shown to be important for recognition by MLCK (six to eight basic residues for the smooth muscle MLCK, and ten to eleven residues for the skeletal muscle MLCK). In particular, a basic residue at position -3 and -6 relative to the phosphorylation site, and the spacing between these basic residues, has been shown to influence the specificity and the kinetics of phosphorylation by the MLCK (Kemp et al., 1983; Kemp and Pearson, 1985). MLCK itself is a substrate for phosphorylation by a number of kinases in vitro (see review by Sellers and Adelstein, 1987). Only phosphorylation by cAMP dependent protein kinase (cAK) has been demonstrated in vivo (Kamm and Stull, 1985), and has been shown to regulate the activity of smooth muscle MLCK, but not of skeletal muscle MLCK. In the absence of calmodulin, cAK catalyzes incorporation of two phosphate groups into MLCK, resulting in a concomitant decrease in the affinity for calmodulin and the kinase activity. The in vivo role of MLCK phosphorylation by cAK is controversial. Phosphorylation of MLCK by cAK has been proposed to be the mechanism that causes relaxation of smooth muscles. However, other evidence suggests that cAK phosphorylation of MLCK might only play a minor role in smooth muscle relaxation, and that the primary effects of cAK phosphorylation involve regulation of sarcoplasmic Ca2+ concentration through phosphorylation of membrane proteins that control ion fluxes (Stull et al., 1987). MLCK also undergoes a slow autophosphorylation, but the significance of this reaction is unclear. Phosphorylation of MLC-2 is regulated not only by MLCK, b u t also by phosphatases (PPs), which dephosphorylate MLC-2. 143

There are 4 classes of serine/threonine-specific protein phosphatases, all of which have broad and overlapping substrate specificities in vitro (Cohen, 1989; Ballou and Fischer, 1987), There are two major forms of a type I phosphatase in skeletal muscle (PP- 1); PP-1G (glycogen-bound form) and PP-1M (myofibrillar form), the latter likely accounts for most of the MLC-2 phosphatase activity in vivo. PP-1M is the only phosphatase that is found complexed with the actomyosin, and it dephosphorylates MLC-2 in the native myosin much more efficiently than PP-1G and the other three classes of protein phosphatases. Two protein phosphatases (PP-III and PP-IV), which dephosphorylate MLC-2 in native myosin, have also been isolated from turkey gizzard smooth muscle (Cohen, 1989; Ballou and Fischer, 1987). PP-IV appears to be the smooth muscle equivalent of PP- 1M. PP-1 is composed of two subunits: a catalytic subunit and a regulatory subunit. The two forms of skeletal muscle PP-1 (PP-1G and PP-1M) share a common catalytic subunit, but the regulatory subunits are distinct. The regulatory subunit is important for localization of the enzyme and may, in part, confer substrate specificity (Cohen, 1989; Ballou and Fischer, 1987; Li, 1982). The activity of PP-1 is inhibited by heat- and acid-stable proteins, termed inhibitor-1 (1-1) and inhibitor-2 (1-2). The inhibitory activity of 1-1 is regulated by cAMP and calcium. Regulation of 1-1 activity through these secondary messengers may be a mechanism that controls the PP-1 activity in vivo. PP-1, PP-2A, and PP-2B have been identified in Drosophila, and have been shown to have similar biological and biochemical properties to the vertebrate counterparts (Orgad et al., 1987; Dombradi et al., 1989). A Drosophila cDNA encoding the catalytic subunit of PP-1 (PP-1 a) has been isolated, and the deduced amino acid sequence shows 92% identity with skeletal muscle PP-1 (Dombradi et al., 1989). 144

The final goal of this work was to identify a Drosophila Mick and I have cloned 4 Drosophila cDNAs encoding predicted proteins that are conserved with known MLCKs. The deduced amino acid sequences of all four cDNAs share a common region corresponding to the conserved catalytic domain of MLCK (Hanks et al., 1988). The cDNAs are related, but distinct from one another. Analysis of the cDNAs and a preliminary characterization of their genomic organization suggests that they are either products of alternative splicing of a single gene or derive from two adjacent genes. Transcripts corresponding to the cDNAs are expressed throughout development at low levels .

Results

Amplification of Drosophila genomic DNA by Polymerase Chain R eaction. The coding sequence for Drosophila MLCK was amplified from genomic DNA by the Polymerase Chain Reaction (PCR), using degenerate oligonucleotide primers (see Chapter II for methods). The primers were designed to correspond to MLCK conserved sequences, which are not well conserved with other kinases (Hanks et al., 1988). Sequences ELFERI, DDAKDF, and IDFGLA were chosen for primer synthesis (Figure 26). To maximize the specificity of the amplification, two rounds of amplification were performed. Primers corresponding to ELFERI and DDAKDF were used as the 5' and 3' primers in the first round of amplification; and primers ELFERI and IDFGLA as 5' and 3' primers for the second round. Products of 439 and 208 base pairs were expected from the first and second round of amplification, respectively (assuming that there is no intron in the genomic sequence flanked by the primers). As predicted the final amplification product was approximately 200 base pairs. This DNA fragment was subcloned Figure 26. Amplification of Drosophila MLCK genomic sequence by Polymerase Chain Reaction (PCR). (a) Nucleotide sequence of the PCR product. Drosophila genomic DNA was amplified by PCR using degenerate oligonucleotide primers. The sequence of the final 206 bp PCR product is flanked by the 5* and 3’ primer sequences (shaded boxes). The deduced amino acid sequence of the PCR product is shown beneath the nucleotide sequence. The direction of amplification from each primer is indicated by an arrow. The boxed sequence “DLKPEN” is highly conserved among Ca2+/calmodulin dependent kinases. (b) Sequences of degenerate oligonucleotide primers used for PCR amplification of Drosophila genomic DNA. The 5’ and 3’ primers used for the final round of amplification are shown. All possible combinations of codons for each amino acid were used for primer synthesis. Each primer is tagged with a restriction site (shaded box) to facilitate cloning of the PCR product.

145 146

PCR SEQUENCE

BamHI .....■> tgga Cte: gag fctg tt:t: gaa agg a^g gtg get gac gac ttc acc ttg ace gaa atg gac tgc ELFERMVADDFTLTEMDC

ate ctg ttc ctg cga cag gtt tgc gat ggc gtg gee tac atg cac ggc cag agt gtg gtg ILPJjRQVCDGVAYMHGQSVV

cat etc gat ctg aag ceg gag aac ata' atg tgc cat acg ege act age cac cag ate aag H L |p L K P B Hi I MCHTRTSHQIK atc |ata:agac:;:::1^t:';;gggiit^ai?gcg;::aafcv:-tc| I I D F G L A ECORI

b 5 ' PCR PRIMER

BamHI E L K E R 5‘ gc Jgga tccI GAA TTA TTT GAA AGA AT 3 1 G C T C G C T G G C C

3 ' PCR PRIMER

G ECORI TAA CTA AAA CCT AAT CGfc TTA AG G G G C G C T A A G G

Figure 26. 147 into pBlueScript for sequencing. The predicted protein sequence of the PCR product is highly conserved with known MLCKs, it shows ~61% amino acid identity with both csm MLCK and rsk MLCK (Figure 27). The sequence is less well conserved with the Dictyostelium MLCK (~42% identity). Homology with other kinases is significantly less; ~39%, 23%, and 17% amino acid identity with vertebrate Ca2+/calmodulin dependent protein kinase type II (CaM II), Drosophila cAMP dependent protein kinase (DCO), and Drosophila protein kinase C (PKC53e), respectively. Therefore, it is likely the PCR product corresponds to a Drosophila MLCK. The PCR product was used as a probe to isolate cDNAs.

Isolation of MLCK cDNA clones using the PCR product as a hybridization probe. The 200 bp PCR product was used as a probe to screen a num ber of cDNA libraries . The 4 longest cDNA clones were isolated from an imaginal disc cDNA library constructed in A,gtlO (Dr. A. Cowman provided by Dr. Gerald Rubin). The phage inserts (D3, D4, D5, and D6) were subcloned into pBlueScript and each insert was sequenced entirely on both strands. The nucleotide sequence shows that the cDNAs are related, but distinct from one another. The nucleotide sequence alignment of the four cDNA clones is shown in Figure 28. Each clone contains a region homologous to the catalytic domain of the MLCK of other species (Figure 27 and Figure 29). A GenBank search with the cDNA sequences showed they were most conserved with chicken smooth muscle MLCK, and rabbit skeletal muscle MLCK. The region of conservation is confined to the MLCK catalytic and regulatory domains, and no significant homology is detected in other regions of the proteins. The alignment of the deduced amino acid sequences with MLCKs of other species is shown in Figure 29. Figure 27. Diagram of sequence organization of the Drosophila MLCK cDNA clones. The sequences of the four cDNA clones were divided into 5 regions: I, II, III, IV, and V. These regions divide the cDNAs from the 5' to 3' end. There are two alternate sequences (a or b) for regions II, III, and V that are highly conserved. Region I, which corresponds to the 5’ end of D3 and D6, is completely different in the two cDNAs. There is only one form of region IV. Each cDNA contains different combinations of these regions. The putative translation start codon (ATG) in D3 is indicated, as well as the corresponding codons in D4 and D6. The translation termination codon ( ^ ), and the poly(A) tail are also shown. The relative location of the coding sequences for the catalytic domain for each cDNA is shown in the bottom of the figure. The relative location of the coding sequences for the pseudosubstrate ( T ) and the calmodulin binding sequence ( CaM) are shown on the top of the figure.

148 149

D3 : 3260 bp ATG

A)21 1 3 1 4 j**“ la ECa ■ ID aw — Va mw»w>gpMJ D5 2164 bp

5 4 7 1 69 6 3 9 j*#“— ^nxa IIV Va D4 : 2774 bp

(A )8 0 1 3 0 0 547 |169| 674 I ■ nb' TTTh -W£Vp^~ Vb

D6 : 2358 bp ATG f c ,10 0 116 1 3 0 2 001 lb H&— ■ nb ~ -Vb—®*j ^tl^fllc dom«ln

Figure 27. Figure 28. Nucleotide sequence alignment of the Drosophila MLCK cDNA clones. Nucleotide sequences of the four cDNAs are aligned. Gaps are introduced to maximize the alignment. Nucleotides conserved in at least two sequences are shaded. Blocks I, II, III, IV, and V are divided by vertical lines. The PCR amplified sequence (PCR), the sequence encoding the catalytic domain (cat), the putative translational start (Met 1), and the putative stop codons (stop) are indicated.

150 151

.25, .35. .45, a immspgq i« » rcT ^i 50 D3 50 D5 50 D4 MB- • - - - g h i wrnrnmr 50 D6

5 5 ...... _____.65, ______.75. ______.85 ...... _ . _ . . _ .95 _...... _ lTO>qQQa*i AmmiM.m Gwmmzm tqttsmgm 100 D3 100 D5 100 D4 w sm m m a 100 D6 ifil m . 1M_ JiL H ttqa^ta aittggaim a«bmkk$gg A & ^ A m c m 150 D3 150 D5 150 D4 150 D6 155 ML ML m . M L ^ mimAiMm pcM&meM mfmrzim cmcmzMM cogmimm 20 0 D3 20 0 D5 200 D4 1 Id 200 D6

_m _ _ _ML _ _m _ _ ML S0PAPCOJA-H CICO^m^TA GBfK&OMXra TWBfiOmi GSqCAXma 250 D3 25 0 D5 2 50 D4 - -- M M 2 50 D6

ML I, ML m ML »• I..*..!..,. !..,!...■■» ??•! I . .1 A., ACOAS&KKXQ GAPimiCGCT CCACqGTGAC TGTGTCACSA 300 D3 300 D5 300 D4 mmmmm 300 D6 J9L ML ML ML ML TBASmOPfli ajm iG tW G CIXQGQTITC ojSftGCCMTC TGGGTTGQTP 350 D3 350 D5 350 D4 -m g i m t i a m - 350 Dti 355 ML ML ML ML mrmxmM cmzmmm mmzczAG 40 0 D3 40 0 D5 40 0 D4 ■z& s?----- 4 00 D6

acg a a g g ttg ac .tgcgtc .ga 4 50 D3 4 50 DS 450 D4 --I q EI t - 4 5 0 D6

Figure 28. Figure 28 (continued).

5 0 0 D3 5 0 0 D5 5 0 0 D4 * gg H 5 0 0 D6

T77rrrrr, 5 5 0 D3 5 5 0 d s 5 5 0 D4 5 5 0 D6 -^*-11 555 m J 2 i j&l JUL QJSTGATCCQ AJJMJm&m GCJCpAATOX GACGATGQ< 3® A«JSA<3 0 A« 3 6 0 0 D3 6 0 0 D5 6 0 0 D4 81 WMStiBSI 6 0 0 Dfi m . 6 5 0 D3 6 5 0 d s 6 5 0 D4 6 5 0 D6

7 0 0 D3 7 0 0 D5 OQACTATGGO GTGGAGSCAG CCGCTGCCT0 7 0 0 D4 Gmmmzktt 7 0 0 Dfi 705 JU­ JU. J2L JZ45_ 7 5 0 D3 7 5 0 d s clsfacc^cq aagt S c^ a g c c c a g to ta ! 7 5 0 D4 7 5 0 Dfi

8 00 D3 800 d s GGTCC ACCCTAGGA 8 0 0 D4 800 D6

iTteWSaSsjlKlnitnt.W (**!5 850 D3 850 d s 850 D4 850 Dfi

K*!3SRsfi:«®lifc

152 iue 8 (continued). 28 Figure

sa~-i -4£- ~tsj-s? -4 £3 ;:C* I I n t f 3*S* -4 ***, :♦*# \; o O h i SwSSM^iwI Ifc&&5i2£3fcH oMimssii a o > <> s': 05cn s t 1 * 1

c S»'3>yfl*c c #3U,'i|o o « *r.S /■W ; >}>**= S^Mgy ii *>&& , o o ? ^£ug&2 <>5 S3 S' o o $ $ r o« § *9ta3£&£&&i 5* ^ > o o u> lo lo lo WWWW to to to N) N> N) tO S3 vo vo vo VO LA LA LA LA LA LA LA LA LA LA LALA LA LA LA LA LA LA LA LA o o o o 8888 OOOO 8888 o o 8888 222 2 222 8 222 2 2 5 s 2 2 2 2 1*) ^ ^ n in ^ « Q ©ss a o SS SSSS a a 3 S a a 3 2 o fi a a a a a £ S 0 s s 2 8 3 2 o o o o o o o o o o o o o o o o to «o in »o 8888 _ _ _ . *0 10*0^ !88 m to >n »n 8 8 8 8 •n in in in 8 8 8 8 to v> to to to to to to SO vO sO VO r~ r-~ t-~ r-~ r-- r~ r- t-~ 00 00 00 (50

feSBSSpgS*® j!T C g l g t f e r q r a n a j . 0 0 0 0 i o o> -» j o o j ? c . 'S g ^ e r ; ® TO | S m W'l-* I I** f<* ■" 1 in Ki^ =~ **•?' m ? K K 8 V"i fr** ♦*** •■"* V** irijO O O O 5- i- m r f l e o o m S in §<§ <§ $ - g * s l M S ? <3 Qjm § <3 §0 O (3 o e* III® ^ p O O s> a w v» s ^ 8 'If! ] ** <. < ^ i i i i o;J . 0 0 0 0 j < '£

»<* o ^

T3 o t? i ® to “ «'333 to 00 LA o ' 5***t H g g g o SO-tf-CU o a.o p e 2S P P i 5 p a r c ao « £> w f -5 ?*T*5 > <3553 vO cn 3 8 8 8 LA cn §§?-§ sUtesiSfe i33f ■ 4 W . l l l i i lit 3 9 8 8 a a . <> 6 o o to oo u>O “ &<$ggrvD W m m ■la ^ ^ k/> M &• - » o " H© 0/0 0 gipOGi ;f> a o o >OOfi > >■, OOOfl ? 7t rf to *•4 H **•$ *•*!: jgaoo!» > ^ > Mimi ^ £ & £ £ ! 3_< 8~j 3 8 ^ :> > S> •» to to to to to to to to fO tO N) tO to to to to to to to to to to to to VO VO VO VO \£> VO 00 00 00 00 o o t tt to to to to to to toto OOOO LA LA LA LA LA LA LA LA OOOOW W lO iU » 8888 OOOOla La la la OOOOLA LA LA LA 8888 OOOO S 2 a S »a o £ a Ul o u 8 2 8 S 8 Figure 28 (continued).

CAGTG 2 3 0 0 D3 GTGGAPCC GXSAGACGCTQ 23 0 0 DS 2 3 0 0 D4 2300 D6 *3321 2325 . . .2335 CTACTOTCOC TGCSGAAATIJ GhAaCTCGaR CC'-Xi \CCi APGCPAGGAi 2350 D3 cfrtO top {,' c . ' . - A4^0m®m 2350 ds ' kmG£&&&:Yd cM mimM ---- ...... 23 5 0 D4 23 5 0 06

*3352* m , ,.*33,8?^ , ,, ^ * 3 1 1 ™ OTTOCCCACG GTCGTGCACG ATACO.GAGCY ACCCGTTAG0 CTGGAGCTGG 24 0 0 D3 CnOCCCACO QTGQ-mc/.GQ ATACCMACCA W tX m ’t& M CTGOAOCTGb 2400 ds t a w ; CSBBSMSISti 2400 D4 ...... 24 0 0 06

*3111 Ji42.1 m *3441 Q A A G O C A ^ ;TQC9C£CeB^ fcTJSCACCAAG AOCATCAAO* 2 4 5 0 D3 g*m»6 £k * T«0Goroc«4 bttmooMd accato^m 2450 ds M£mm gigggmipA 'mmimMi 2450 D4 2450 D6

,345? 3461 . . .2475. . . . .,* ,■ , 3 1 1 1 . * ,3 4 2 1 qacaotgct ^ Pacgggagaa ACCCAGGTCP 2500 D3 Gjkswjmti cacggsagaa ACP.CAGGtCG 25 0 0 ds jM&R&OKBti 2500 D4 ...... 2500 D6

*3521 *311 * 3 5 2 ’ .2535.^ I 8 . . .2545. AAACCCCAGQ TCAAOC0GCH W rn ■ 2550 D3 j^C60C*6g TOAAG85PQCT p t t g i mm fMAWcfm I 25 5 0 ds TCMGPQSPS TccAAtm m ammmmi 2550 d4 ...... 2550 D6

*211 .2565. *3321 .2 5 8 5 . .2595...... 2600 D3 mxmmm. 2m os M 2 6 0 0 D4 ...... 2600 D6

.2605. .2615. *3631 .2 6 3 5 . .2645...... 2650 D3 53SSATCCGC5 i'^pOATCCil c 'T P c c f1^ miMmmm WBMMj., cRmamM 2650 D4 ...... 2650 D6

.2655. *3661 *3221 .2695. pAGASS AOCOOPGAAG 27 0 0 D3 a«ppa1ca<^ ACUAPAOA AP

156 Figure 28 (continued).

. ■ .2745 . 2 7 5 0 0 3 TCACCCCAfed 2750 D5 M KG A Q O 2750 D4 fiM iliM l 2750 D6

2785. . . . ,2795. . . COCrAOUTC cc< ort*t XI’C'UTCOOC "iACCC.ro. C's 2 8 0 0 D3 GCCCAGfeYC wCeO'jiv.x .fefefeA Ti.CCCCGACt. 2 8 0 0 D5 GCCCftftC’fC* CCCCTO-'lH r.ucccOA.u 2800 D4 OGCCicr r r V ' • t ‘M 1 1 / f - ^ r # ‘ I 2 8 0 0 D«

. . 2835. . . ..2845. CQQCT«ATGA aiWAnCCra'.. C"C.r'!' tji 2 8 5 0 D3 COCCiCATC, • ■ GccfeASOfercj nr< 2 8 5 0 d s fe&’ccAGCcnjq jrrai 2 8 5 0 D4 feesMaA&x&a y, 2 8 5 0 D6

r»r.G'TT-'"TTr QferCAAtCOfe ACCC'-i 2 9 0 0 D3 fefeACAfeGCtfe C ( .S-i^V 'C - /.IMUfeWrt 2900 d s * YOACSC C#«\^rS *0 t r % I ** ' Oi r f / GfefeOAAYfefed \ a c r < ' vC r<3 2 9 0 0 D4 in W K M fi* m m zm i 2900 ds

^2905™ , ^291^ ...... 2 9 2 5 . . . . : ...2935. ..: bcact< 3«»oa/. ?ri.-..AArCO W , - A,' 7«C1.»oTaCfcs 2 9 5 0 D4 feOAAtfefeCAA 7$r?^.rr'JCs fe&AMfefeOM tfefe£Mlfefe$2i 2950 d s

qcgc«ac®t<| wmrikmci 3000 d s SfeACCfeCQfe AOACOMTfefe f e fe j & f f e ^ M GCfeCGACGTfe YCATCAAfefefe 3 0 0 0 d s ^CACCGCGG AGACGAATCfe <^|$CA©fe«| GCGOGACGtd TCATCAA^pfe 3000 D4 mm&p&zm tmmsmid omsstoma TCATCMfeafe 3000 d« , .3005...... 3015. JQ4L A'TCAATAGC GAAa CQCAC ATATCACOTC T C G Q T T A n A t 3 0 5 0 D3 ATTOAATAGC GAAACfeCAC A'ATCAfefeTC. TCOS7TAAA1 3 0 5 0 d s A7TCAA CAGC GAAACGCAC .ATCAfeorq 7AG7TT 7CCGTT.W . 3 0 5 0 D4 S in B A M U flfi Gf.AAfefeCAC gg“ ‘" ' 3 0 5 0 d s Figure 28 (continued).

. .3155...... 31.65. . . . . ■ JU75. . . . ■ ■ .31,85...... 3195, , , , ATGAATOSCGA CT&CQGATCA GCATCCCACC CACCCAGAAT CCTGGATCQC 3 2 0 0 D3 ATSAATQC6A CTGCGOAtCA GCATCCCACC CACCCAGa AT. CCTGGATGGG 3 2 0 0 DS AtGAATGCOA CTGCGQATCA QCATCCCACC CACCCA...... C 3 2 0 0 D4 AtSAATGCGA CJSGSOkKttGQAXZCQACQ CAPGCA C. 3 2 0 0 D6

. . .3205...... 32.15...... 3,225. . . . ,.„32 3 ?______, , AMS ______AGACACACAC AOAGACACAq ACACCCaTO^ ACCCAOTSAG AGTAGGATAT 3250 D3 mcACAc&d mim’Mm mmmArn 3250 D5 A&AOACAG GTCfi ACCOACTCAOi AGTAG8ATA| 3250 D4 A&k&kcAc mm mmmmd mmmMM 3250 m

.3255...... 3265...... 3,275...... 3285...... 3295. , , , A£GTATGTA1 GTOTGCATTG AAGCGGAJGC ACTCCCTCCG JTjDAGAGGGA 33 0 0 D3 feT A tG T A l G*GTGGA?TG AACH^GATOd AG1GCGTCG4 tfMGAGGG^ 3300 d s AW TA TGTA li GTCTGCATTC AAGCGGATGd ACTCCGTCCG TTCAGAGGGA 33 0 0 D4 A & r A m m t d m B ttA i& g i m m m s M 3300 d«

3j30g., ^ 1 ,^^3,1^^. u*u, A ftG T A C T T A GGCTAGAGGA A m t 3 3 2 4 D3 m f m A m r q ss&tagagga agg? 3324 d s ABT0TAO.TTA ©SCTAGAGGA ASO^ 3324 D4 m m m m ii m&xm&mk *m% 3324 d «

158 159

Because of this sequence similarity these cDNAs will be referred to as Drosophila MLCK.

Analysis of the nucleotide and protein sequences of the Drosophila MLCK clones. Comparison of nucleotide sequence suggests the four cDNA clones are alternatively spliced products of a single gene (or possibly 2 genes, see below for analysis of genomic organization). The nucleotide sequences of the four cDNA clones can be divided into 5 blocks: I, II, III, IV, and V (Figure 27). These blocks divide the cDNAs from the 5' to the 3' end and each is described in detail below. There are two alternate sequences for blocks I, II, III and V (called a and b) that are nevertheless highly related, except in block I where the sequences are completely different. Only one form of block IV exists in the clones. The blocks are probably defined by intron/exon boundaries but this has not been shown by analyzing genomic DNA. Sequence blocks could consist of more than one exon. Each cDNA clone contains different combinations of the sequence blocks. D3 contains I, II, III, and V; D6 contains I, II, and V and D4 and D5 contain all five blocks. D3 is the only cDNA that is apparently complete, it encodes a predicted 660 amino acid protein. Nucleotide sequence comparison of the four cDNA clones is shown in Figure 28. B lock I: This is the most 5’ region and is found in D3 and D6. The sequence of block I in these 2 cDNAs is completely different. In D3 the region includes multiple stops and a putative TATA box. In D6 the region is open. B lock IS: This region includes a putative translational start for D3. The TATA box (in D3-block I) is ~630 bp 5' to this ATG. The nucleotide sequence preceding the ATG is not conserved with the consensus translational start sequence for Drosophila (Cavener, 1987). The translation start consensus sequence is (C/A)AA(A/C) ATG. whereas the putative translation start sequence in D3 is GGTG ATG. However, Drosophila genes have been characterized that have non-consensus related start sites. This ATG is also present in D6 but as the region 5' to this (block I) is open in D6 it is unlikely this is also the translation start site for D6. For clarity this methionine residue is referred to as residue # 1 in the deduced protein sequences of all 4 clones. Block II spans 1429 nucleotides in D3 and D6 (it is shorter in D4 and D5 which are not complete at the 5’ end). The deduced protein sequence suggests that there is a open reading frame (ORF) throughout the region. The deduced protein sequence of a C- terminal region, spanning 238 residues, is highly homologous to the catalytic domain of MLCK. The protein sequence of this entire stretch of 238 residues is identical in all four clones, and the nucleotide sequences are 97.6% identical in Ila and lib (Figure 28). The 3' end of block II appears to encode the "pseudosubstrate sequence" and the calmodulin binding sequence of MLCK (discussed below) . The pseudosubstrate sequence terminates at the 3' end of block II, whereas the calmodulin binding sequence continues into block III. In D6, block II is fused directly to block V, skipping blocks III and IV. As will be discussed in the next section, lack of block III in D6 appears to result in a disruption of the calmodulin binding sequence of D6. When the entire Ila and lib regions are compared, there is a 98.7% identity at the nucleotide level and 99.3 % amino acid identity. Variation in the sequences include nucleotide substitutions throughout and there is an insertion of 12 nucleotides in Ila that is not present in lib. It is unlikely that these are errors in reverse transcription induced during cDNA library construction as the same differences are found in more than one cDNA. 161

B lock HI: As mentioned above block III includes part of the calmodulin binding domain. C-terminal to the calmodulin binding this region does not appear to contain a significant homology to the vertebrate proteins. Blocks Ilia and Illb share 98.9% identity at the nucleotide level and 97.3% at the amino acid level. Block IV: Block IV exists in one form and is found in clones D4 and D5 (but not D3 and D6). It encodes an inframe stop codon thus the differential inclusion of this block produces proteins with different C-terminal regions. Block V: Block V is the 3’ end of three of the cDNAs as each has a poly A tail (D3, D4 and D6). D5 lacks a poly A tail. There is a stop codon 120 nt into block V. This is presumed to be the termination signal in D3 and D6. The other cDNAs terminate earlier as they include block IV which has a stop codon. D3, D4 and D6 have poly A tails of 21, 80, and 56 nucleotides, respectively (Figure 28). No conserved polyadenylation signal (AATAAA) could be identified upstream of the poly A tail. The conserved polyadenylation signal generally occurs 10-35 nucleotides 5' of the poly A tail and is usually required for poly adenylation of mRNAs. However, some mRNAs do not contain a conserved AATAAA element and most have been shown, to be associated with alternative polyadenylation sites (Wahle and Keller, 1992). The MLCK cDNAs contain a sequence AATTAAA located -15 nucleotides upstream of the poly A tail and alternate poly A addition sites were observed. Except for a single nucleotide difference and an insertion of 31 nucleotides, the sequence of Va is collinear with that of Vb (95% nucleotide sequence homology). 162

Analysis of the Drosophila MLCK functional domains. Myosin light chain kinase is a single polypeptide chain, consisting of two major functional domains: the catalytic domain and the regulatory domain. Catalytic domain: In the catalytic domain the amino acid sequence of the 4 Drosophila cDNAs is identical (Figure 29) and homologous to the MLCK catalytic domain in a number of other animals. Hanks et al. (1988) identified 15 conserved amino acids contained within the catalytic domain of many kinases. The Drosophila MLCK has the 15 residues identified by Hanks et al. (1988) (Figure 29). Comparison of the deduced amino acid sequences of the Drosophila protein with that of chicken smooth muscle MLCK (csm MLCK) reveals a 55.3% identity and 66.7% homology over a stretch of 238 amino acids corresponding to catalytic domain (Olson et al., 1990). Comparison with the rabbit skeletal muscle MLCK (rsk) shows a 51.5% identity and 62.4% homology (Herring et al., 1990). The Drosophila MLCK is as like these other proteins as they are to each other; csm MLCK and rsk MLCK are 54% identical. The Drosophila MLCK also contains a motif Gly-X-Gly-X-X- Gly Lys at the N-terminus of the catalytic domain which is known to be important for ATP binding. However, unlike the csm MLCK, the Drosophila MLCK appears to have only one conserved phosphorylation site for cAK at Ser 839 (csm MLCK has two of these sites at Ser815 and Ser828). Regulatory domain: C-terminal to the catalytic domain of vertebrate MLCKs is the regulatory domain, which consists of a pseudosubstrate sequence and a calmodulin binding sequence, arranged in tandem (Kemp et al., 1987, Figure 25). Each of the two sequences spans -20 amino acid residues, with an -10 residue overlap (Figure 25). Drosophila MLCK cDNAs have a region conserved with the vertebrate proteins at the corresponding Figure 29. The deduced protein sequence alignment of the Drosophila MLCK with MLCK of other species. The deduced protein sequences of the four cDNA clones (D3, D4, D5, and D6) are aligned with those of chicken smooth muscle MLCK (csm), rabbit skeletal muscle MLCK (rsk), and Dictyostelium MLCK (dicty). The Drosophila protein sequences used in the alignment start at the putative translation start codon (Met 1, see text) in D3, D4, and D5, and end at stop codons of each clone. D3, D4, D5, and D6 are 660, 619, 458, and 474 residues, respectively, csm, rsk, and dicty MLCK are 972, 607, and 301 residues, respectively. The catalytic domain (cat), the ATP binding region (ATP), the pseudosubstrate sequence (pseu), and the calmodulin binding sequence (CaM) are indicated. Gaps are introduced in order to maximize the alignment.

163 164

.15. .25. .35. .45. 5 0 D3 5 0 D4 5 0 d s 5 0 D6 NDFRANLQRQ VKPKTLSEEE RKVHAPQQVD FRSViAKKGT P$TPLPEKVP 5 0 csm ...... MATENGAVEL GIQSpTDEA SiGAASg--- 5 0 rsk 5 0 dicty

.55. .65. .75. .85. .95. 1 0 0 D3 1 0 0 D4 1 0 0 D5 1 0 0 D6 PPKPAVTDFR SV plK iK lP AENGSASTfA iNARAfiSEAQ NATlNSEAKA 1 0 0 csm ...... ESSAiEioiA P ...... | d iEKGPiPSDT KQd|DPST^- 1 0 0 rsk 1 0 0 dicty JLSL ill. i l l m . l i l . 1 5 0 D3 1 5 0 D4 1 5 0 DS 1 5 0 D6 PKPVVKSEEK N DRKCEHGCA VVDGGI IGKK AENKPAASKP f P P , P S K Q____ T A P 1 5 0 csm I l D A N ...... - ...... TiAfrEl«|DV- 1 5 0 rsk 1 5 0 dicty

165 1 7 5 1M . 2 0 0 D3 2 0 0 D4 2 0 0 DS 2 0 0 D6 SFTEKLQDAK VADGEKLVLQ CRI SSDPPAS ViWfLDSKAI KSSKSI V ISQ 200 - -VPAQgSAG GSj...... 2 0 0 rsk ...... 200 dicty

205 jm . J3L .2 4 5 . 2 5 0 D3 ...... 25 0 D4 ...... 25 0d s ...... 25 0 D6 EGTLCSLTIE KVMPEK 1EY KCIM, *AM AECACKVyVE D j|s S T K A A K P 2 5 0 csm QGPAGES I q v e a p II P A A |P Q O f A T ...... 2 5 0 rsk - ...... 250 dicty

2 5 5 2 6 5 2 7 5 JSL 2 9 5 3 0 0 D3 3 0 0 D4 3 0 0 d s 3 0 0 D6 f i n k t k SP k t t l p I v l s t e ? siUflKiRpi p k t p p k § | t p p q i t q i p e d r 3 0 0 csm IgA SEK&&EA EKOiSGMQOP GKP^GKISVi EGQ- - -IlA R RGSP a | l HSP 3 0 0 rsk 3 0 0 dicty

Figure 29. Figure 29 (continued).

325 345 ML 315 ML 350 D3 350 D4 350 d s 350 D6 KVRAGESVEL rAKVVGTAP I TCTWMKFRKQ IQENEY I KI E NAENiSSKLTl! 350 csm 350 rsk 350 dicty

355 365 385 395 . M Ngrava p p s p p s a q p s VSU0P0RVAV1 400 D3 msssMmSi 400 D4 400 d s

- - A A t e & t & m PEBPfcMQ&S 400 D6 ssI kqehcgc Y TLVVEN KLG SBQftQVN LTV VDKRiDPPAGT = CASDI FjSSS 400 csm asieklpaqk P LSEASE LIF EG...... V iAT^GPTEfg...... 400 rsk ...... 400 dictv

405 ML ML ML 445 mmpmma CMKTSFi

.455 465 ML ... 485 ... . 495 vstNLas»aRC^ RFRVPAEW StH GRSAPGQASS tyQR-|STR<^ -ssoAsq- 500 D3 r fr v r a e n m GRSAPGQASS km im cE P i 500 D4 500 d s R i s i t t i m 500 D6 ADREY KFBVR AANWYGI SEP SQESEVVKVG EKQEEELKEE EAELSODEGK 500 csm ADQAKVQGDT SRGIE FQAVP SERPRPE-VG QALCLPAREE DCFQILDDCP 500 rsk 500 dlctv - > » c a t a t p ___ . 505 ...... 515 ...... 525 ... . 5 3 5 f . ■ ■ ML ------R53SQAT VSVQSGGOF^ S H F E tlE ^ -8 GKGRFGtVY^ VSJERCQPEGd 550 D3 — s e m i V$VO5QG0F« m i G K o a e a m K VQERCQPEQli 550 D4 VQEBCOPEQli 550 DS -•OB G K G E tity X E yaERCQEEay 550 D6 e t e v n y q t v t ;- GSGKFGQ^FR LVRKKTGKVW 550 csm PPPAPFPHRI \tELRTQNVSS EFSMNSKEAti g g g k p ^ a v c t CTEKSTGLK- 550 rsk ...... MTEV EKIYEFK£ e | G R G A E S k^L GENKQTKQRY 550 dicty

595 ■ ■■$«■■■■ ■ ■375 _ § M - K$O0 ftQKVf.LB ElSfMAiGfl PKCU&AASR ESPRSlVtf-V 600 D3 tAAKVfs-C^ KSODRQKVU^ s^ mbaVoh S p i& u » y & sif ESRB^yM-V 600 D4 l£ ak * i k -c $ KSODRQKVC^ PSti'CQC&ASR ESPREIVJd-Sl 600 D5 D6 K m t m m m PKlUUA&fiR E£gB&l3rM-VI 600 AGKFFK*AYS AKEKE-NIRD £USi^Cl3HH PEliVQCVDAR EEKANtVMVL 600 csm

m r n m p - r PKDKEMVMli- £IEVWQtNl| RNqidilYiAl ETEHEMLF- 600 rsk -iiiaaNisE LGK$YE|SNFK MEVDILKKVN hPnI 1alKEL FDTPEKLYLi 600 dicty

165 3 m n 5 V) $ ft q a 2 It ssss ill ft ft ft ft Iff OOOOOOO O O O O O O O 1/1o oin o

m m m m m m ® ^ O T E K G I i 33s:GK m a o • U .i* «4,U!.U , a, a i.l u. a. l ic to: ft* o x x r ' pttM » 0* Z $cr> - 1 SCKSCdXini;, _l _J'™. ■■■ ‘ 3 H S ->’> i|0f at a t i d u f'u O T D O Q O D O a. iZtSZiZsaE 5c aim m m - m - “ '■ DC Z - m i l if W 05 to W z V z tinsrei*uui urw), ©LOC'EC ifilioui aDQQQH t r n cr * > *3 < u» tu O m 1 3 5 «| ll| < z uj UI Ul u tn U. It. U. t « _l -Jftli vd l i l l p i l i p f s h h h l * F > ? - «*tj « J : m| ’ «m* hu »«£ « W iz m >;>!>: 51 -I ~i 2 ut ut tu uj ui ui m m i m i um It It u a t?* a. d . a ai to * > > > > > Z h l "III?I a r . z . S ooa gSlSSf? O LU wroursax 11 tn cl a •;mS S 35 23 Sc > z LU z o © (n 'x. 5 *C mm . ta! toJ I88S w "iS v''wy$Q , j o p - o a « < * ®ftf& jU J 'U ~± t t t J **<| mJ H*Jt *mI mJ M*|; fQ m % m % > m m m vCJiiCi O W C5 2 - U ‘ W 0T'';ittf:::n? Q Ui Q ...., PCX uj lu ^ /& S ^ « 5 £ 'J 8 ,^ <: % o;o,'i)ual-5- ° v 2 . s ’O >>>>>> ^ > > > - - ! - ■ggggggl a u. u, v . a . it. u. u :ja^as*p:pg fjrijs ui or©’©'05 © w m m m & m m +3 * «* fiC (C DC KS ^ P iC a iK|3<'S~ Q Z Z ® W W m m t Si o ' " > * ui w © U rU .U .11 CJ «-> aocotu^ui SC;;Sfi E S. C & > a * s e .sd X ' a:WQM-VL(n Q «tr»us m S 5 £ 3 -

• ■■ 9 15 . . .925 .935 . . 945 . . ^ s e R s ^ l a v m o m c $ m ceSESla! sE S i^ S l tuEivwmS 950 d3 Y*E

m * m , & Ti^oA a^i aimwe&wRGj p m m m w i 1000 dj QgyaL^Lft V RPSCfcftTLBf TPSSC&iSHWI RTAtHGlW G PSPRSSRSGV 1000 D4 WBu m b b k ^Jfessmsdv 1000 ds ...... Q 1 0 0 0 D6 KA:|NSLGEAT CTAELLVETM GKEGEGEGEG EEDEEEEEE...... 1 0 0 0 csm ...... - ...... 1 0 0 0 rsk ...... 1000 dicty

ii s S l M ! mmmmmrn 1 0 3 9 D3 s i s ...... 1 0 3 9 D4 sls- 1 0 3 9 d s 1 0 3 9 D6 1 0 3 9 csm 1 0 3 9 rsk 1 0 3 9 dicty

167 168 position. Protein sequences of three of the Drosophila cDNA clones (D3, D4, and D5) are identical in the regulatory domain. (D6 shares the first 15 residues with the other three clones; but because it lacks the coding sequence for block III, the calmodulin binding sequence is disrupted. The significance of this is not known. The pseudosubstrate regions of csm MLCK and rsk MLCK share 57.1 % amino acid sequence identity. The pseudosubstrate sequence of the Drosophila MLCK is 57.1% identical to csm MLCK and 42.9 % identical to rsk MLCK. The calmodulin binding region is located at the C-terminus of the pseudosubstrate sequence. The rsk MLCK contains a calmodulin binding region which is 13 residues from the C- terminus of the protein, whereas the corresponding sequence of the csm MLCK is located 156 residues from the C-terminus. Like csm MLCK, the calmodulin binding region of the Drosophila MLCK is also distant from the C-terminus (174 residues in D3, 211 residues in D4 and D5). The calmodulin binding domains of csm MLCK and rsk MLCK share ~ 40 % identical amino acids. The Drosophila MLCK is 77.8 % identical to the csm MLCK and 44.4 % identical to the rsk MLCK. The structure of calmodulin complexed with the target peptide has been determined in solution by multidimensional nuclear magnetic resonance (NMR) spectroscopy, and the crystal structure at a 2.4 °A resolution (Ikura et al., 1992; Meador et al., 1992). The calmodulin crystal structure has also been determined, and shown to be a dumbbell-shaped molecule with two structurally similar globular domains separated by an a-helix (Babu et al., 1988). Calmodulin has been shown to interact with the target peptide (corresponding to the calmodulin binding sequences of rsk MLCK and csm MLCK) by bending of the N- and C- terminal halves 169 to form a hydrophobic tunnel, engulfing the helical target peptide through hydrophobic interactions. Although the calmodulin binding region of various proteins shows considerable primary sequence diversity, most are predicted to form amphipathic helices, with a pair of aromatic or long-chain aliphatic residues separated by a stretch of 12 residues. The structure of the calmodulin-target peptide complex suggests that the aromatic or long-chain aliphatic residues are involved in anchoring the target peptide to calmodulin. These residues correspond to Trp 800 and Leu 813 of csm MLCK, and Trp 580 and Phe 593 of rsk MLCK. The Drosophila MLCK contains a tryptophan and a methionine at the corresponding positions of the vertebrate proteins (Figure 29). Between these two residues there are several other hydrophobic residues thought to be involved in the hydrophobic interactions. In csm MLCK and rsk MLCK, these residues correspond to Val 807 and Val 587, respectively; whereas in the Drosophila protein there is an lie at the homologous position. Besides hydrophobic interactions, electrostatic interactions have also been inferred from the calmodulin-target peptide structure between basic residues of the MLCKs and glutamate residues of calmodulin. The Drosophila MLCK contains 8 basic residues within the calmodulin binding region, all of which are conserved with the csm MLCK. Therefore, it is likely that the Drosophila MLCK binds calmodulin. (It is unlikely that the protein encoded by the D6 cDNA, which is interrupted in the calmodulin binding region, is capable of binding calmodulin because it encodes a Glu and a Pro in this region. The presence of Glu potentially results in electrostatic repulsion with other Glu residues on the calmodulin surface and Pro is known to be a helix-breaker.) 170

Preliminary characterization of th e MLCK genom ic organization The organization and the relationships of the cDNA sequences suggest that they are products of alternate splicing. The inclusion or exclusion of "cassettes" of sequences in different cDNA clones is characteristic of alternately spliced products of a single gene. In addition, the presence of more than one form of each cassette (the related a or b forms of blocks II, III, and V) has a precedent in Drosophila ; the myosin heavy chain gene (Mhc) (George et al., 1989). The.Mhc gene is a single copy gene, encoded by 29 exons. These 29 exons include five sets of between 2-5 copies of related exons that are alternately spliced to produce at least 10 different transcripts. Preliminary characterization of the MLCK genomic organization by Southern hybridization suggests that the MLCK gene is a single copy gene. Small probes (-300 bp and 700 bp), corresponding to the MLCK catalytic region which are highly conserved among the four cDNA clones were hybridized to samples of genomic DNA, cut with 4 different enzymes. In all cases the probe hybridized to a single band, suggesting the probes derive from one genomic restriction fragment. In a second experiment probes corresponding to the unique 5' end of cDNAs, D3 and D6, were hybridized to DNA cut with 4 different enzymes. In all cases the probes hybridized to the same size fragments. In two cases the D3 derived probe hybridized to an additional fragment suggesting it extended further. If the cDNAs derive from two genes they must be very close (within 1.5 kb of each other) and expressed divergently. In addition, I have shown by in situ hybridization to salivary gland chromosomes using cDNA D4 as a probe, that the hybridization signal was limited to a single band on the right arm of the second chromosome at bands 52 C-D. The Drosophila actin genes comprise a family of 6 genes which are dispersed throughout 171 the genome (Fyrberg et al., 1980). The nucleotide and the protein sequences of these genes are highly homologous (~85-95% homology). In situ hybridization using one actin gene probe shows hybridization to all 6 chromosomal positions (Fyrberg et al., 1980; Tobin et al., 1980). Finding only one site of hybridization with the D4 clone, which is highly related to the other cDNAs (>98%), suggests they derive from one gene or two closely linked genes. Many related genes are adjacent to one another in the genome. For example, the two Drosophila protein kinase C genes, are co­ localized within 50 kb at position 53E on the second chromosome (Schaeffer et al., 1989). If the Drosophila MLCK cDNAs are encoded by more than one gene that are less than 50 kb apart in the genome, they may not be distinguishable by in situ hybridization. Characterization of genomic sequences will be required to show whether these MLCK cDNAs are encoded by a single gene. The data presented here cannot exclude the possibility they are encoded by 2 genes that are very close together and divergently expressed.

E xpression o f th e MLCK transcripts MLCK transcripts are apparently expressed at a low level as they cannot be detected by Northern hybridization. However, they are detected in an RNase Protection assay, which is about 10 times more sensitive than Northern hybridization. A 370 bp Xhol/Kpnl probe corresponding to the highly conserved MLCK catalytic core was used in these assays. This probe was derived from the Ila sequence, and differs by 3 nucleotides with the lib sequence. A -350 nt fragment of the Ila derived transcripts is expected to be protected by the probe. Because of the sequence difference between the probe and the lib related transcripts the protected region will be spilt into 4 fragments of 220, 102, 24 and 8 nucleotides. The two smallest fragments were not detected on the gels. To estimate Figure 30. Developmental expression pattern of Drosophila MLCK transcripts. The expression of Drosophila MLCK transcripts was examined by RNase Protection. Total RNA from a series of developmental stages was hybridized with a 370 nucleotide antisense RNA probe derived from region Ila (a), which encodes part of the catalytic domain and is highly conserved in the four cDNAs. There are three nucleotide differences between Ila and lib in this region. Three protected fragments (~350, 220, and 100 bp) were detected as predicted (a longer exposure was required to see the 220 and 100 nucleotide fragments). The much stronger intensity of the 350 bp protected fragment suggests either that the Ila-derived transcripts are expressed at a much higher level than the lib-derived transcripts, or that the single base mismatches between the Ila-derived probe and the lib-related transcripts are not efficiently cleaved. A control probe derived from a rRNA gene (RP49) was used to normalize the RNA samples (b). The stage from which the RNA sample was obtained is shown above each lane. The positions of the three protected fragments are indicated by arrows. The postion of the probe is indicated by an arrowhead.

172 cr m 7 probe m probe

0 3 a t 0 1 % O)

* sj t 0 3 to

o

tRNA tRNA

probe probe 173 174 the amount of RNA in each sample a control RNase protection assay using a housekeeping gene probe (rRNA gene, RP49) was done. As expected, an intense band of ~35Q nt was protected (Figure 30). This band is derived from the Ila class of transcripts. In addition, two much less intense bands of -220 and ~ 120 bp were also protected. These derive from the Hb transcript class. The lower intensity of these smaller protected bands indicates the lib related transcripts are expressed at a lower level, or the single base mismatches between the Hb transcripts and the Ila-derived probe were not effeiciently cleaved by the RNases. Protected fragments were found in aU developmental stages examined. The intensity of the 350 nucleotide protected fragments was quantitated and standardized with the RP49 control. Apparently the MLCK message level varies during development. Embryos have the lowest level (5% of control RNA) and there are peaks of expression at the end of larval life (17%), the day before eclosion (21%) and in one day old adults (28%).

D iscussion I have cloned and characterized cDNA clones which encode predicted proteins that are highly conserved with MLCKs from other animals. The four cDNA clones are similar to one another but not identical and thus have the potential to encode different protein isoforms. These protein isoforms may have specialized functions. There are many examples of structural proteins and enzymes such as kinases that have tissue-specific isoforms with specialized functions. For example, in Drosophila the myosin heavy chain (MHC) gene encodes multiple isoforms which are tissue specific. Interestingly, ectopic expression of the larval-specific MHC isoform in the adult IFM causes a disruption of muscle structure (Well, S., personal communication). This study shows 175 that the IFM-specific isoform of the MHC protein is specifically required for the IFM assembly. Unique regions of the two MHC isoforms likely specify the different requirements for assembly of the two muscle types. In addition to producing multiple isoforms from a single gene, multiple protein isoforms can be encoded by a gene family. A number of Ser/Thr kinase coding genes have been cloned and characterized in Drosophila. At least three of these kinases (protein kinase C, cAMP-dependent protein kinase, and cGMP-dependent protein kinase) have been shown to be encoded by a gene family (Schaeffer et al., 1989; Kalderon and Rubin, 1988; Kalderon and Rubin, 1989), producing multiple enzyme isoforms thought to have a specialized functions or substrate specificities. Production of multiple forms of MLCK may be a way of controlling muscle activity, through regulating substrate specificity and kinetics of phosphorylation of the MLCK. I have shown that phosphorylation of MLC-2 at Ser66 and Ser67, which are phosphorylated in vitro, is required for IFM function, but apparently not for other muscle types (Chapter III). A simple prediction from these experiments suggests there may be an IFM specific isoform or a change in the level of expression of MLCK at the onset of flight. The RNase protection assays only detect transcripts that contain region II which encodes the conserved catalytic domain, and is most likely to be contained in all transcripts. Thus, putative tissue-specific transcripts would not be detected in these assays. However, the transcript accumulation varied during development and the highest level of expression was seen in adults. This may correlate with the onset of flight and a correspondingly higher requirement for MLCK. There is clearly a complex pattern of transcripts from the region, as the 4 clones analyzed here are all different, and it would be interesting to characterize the the transcription pattern by in situ 176 analysis to whole or sectioned animals throughout development. Transcript specific probes could be used, such as the unique 5' ends of D3 and D6 and the region IV of D4 and D5, for hybridization. A similar analysis has been done for MHC. Regulation of MLCK activity may also occur at the translation level, or post-translation level. The activity of a number of enzymes has been shown to be regulated by post- translational modification, particularly phosphorylation (Cohen et al., 1983). The activity of MLCK has been shown to be depressed by cAK phosphorylation. It is possible that Drosophila MLCK activity is additionally modified in vivo by other mechanisms, such as autophosphorylation or modifications by other enzymes. This is the first description of the isolation of MLCK cDNAs from Drosophila. The genomic organization of these cDNAs suggests they derive from a single gene but a detailed analysis of the genomic region is required to confirm this. I have mapped one of the cDNAs to the 52CD region. There are no obvious known candidate mutants in the region but it would be feasible to isolate putative MLCK mutants by screening over deletions of the region. Having a mutant in hand would allow a functional characterization of the gene. To date there is no known mutant in m lck in any species. REFERENCES

Amos, L.A. 1985. Structure of muscle filaments studied by electron microscopy. Ann. Rev. Biophys. Biophys. Cherrt 14: 291- 313.

Ayme-Southgate, J. Vigoreaux, B. Benian, and M.L. Pardue. 1991. Drosophila has a twitchin/titin-related gene that appears to encode projectin. Proc. Nall. Acad. Set 88: 7973-7977.

Babu, Y.S., C.E. Bugg, and W.J. Cook. 1988. Structure of calmodulin refined at 2.2 °A resolution. J. Mol. Biol. 204: 191-204.

Baeuerle, P.A., and W.B. Huttner. 1985. Tyrosin sulfation of yolk proteins 1,2,3 in Drosophila melanogaster. J. Biol Chem. 260(10): 6434-6439.

Bagshaw, C.R., and J. Kendrick-Jones. 1979. Characterization of homologous divalent metal ion binding sites of vertebrate and molluscan myosins using electron paramagnetic resonance spectroscopy. J. Mol Biol 130: 317-330.

Ball, E., C.C. Karlik, C.J. Beall, D.L. Saville, J.C. Sparrow, B. Bullard, and E.A. Fyrberg. 1987. Arthrin, a myofibrillar protein of muscle, is an actin-ubiquitin conjugate. Cell. 51: 221-228.

177 178

Ballou, L.M., and E.H. Fischer. 1986. Phosphoprotein phosphatases. In The Enzymes, Vol. 27, pp. 311-361. Academic Press, New York.

Barbas, J.A., J. Calceran, I. Krah-Jantgens, J.L. de la Pompa, I. Canal, O. Pongs, and A. Ferias. 1991. Troponin I is encoded in the haplo-lethal region of the shaker gene complex of Drosophila. Genes & Dev. 5: 132-140.

Beall, C.J., and E.A. Fyrberg. 1991. Muscle abnormalities in Drosophila melanogaster heldup mutants are caused by missing or aberrant troponin-I isoforms. J.Ceil B io l 144(5): 941-951.

Becker, K.D., P.T. O'Donnell, J.M. Heitz, M. Vito, and S.I. Bernstein. 1992. Analysis of Drosophila paramyosin: identification of a novel isoform which is restricted to a subset of adult muscles. J. Mus. Res. Ceil MotiL 116(3): 669-681.

Benian, G.M., J.E. Kiff, N. Neckelmann, D.G. Moeman, and R.H. Waterston. 1989. Sequence of an unusually large protein implicated in regulation of myosin activity in C. elegans. Nature 342: 45-50.

Benzer, S. 1973. Genetic dissection of behaviour. S e t Am . 229: 24-37.

Bernstein, S.I., P.T. O’Donnell, and R.M. Cripps. 1993. Molecular genetic analysis of muscle development, structure and function in Drosophila. International Review of Cytology, A Survey of Cell Biology, pp. 63-152. Academic Press, New York. 179

Blanchard, A., V. Ohanian, and D. Critchley. 1989. The structure and function of a-actinin. J. Mus. Res. Cell MotiL 10: 280-289.

Blumenthal, D.K., K. Takio, A.M. Edelman, H. Charbonneau, K. Titani, K.A. Walsh, and E.G. Krebs. 1985. Identification of the calmodulin binding domain of skeletal muscle myosin light chain kinase. Proc. Natl Acad. Set 82: 3187-3191.

Bullard, B., B. Luke, and L. Winkelman. 1973. The paramyosin of insect flight muscle. J. Mol. Biol 75: 359-367.

Bullard, B., K. Leonard, A. Larkins, G. Butcher, C. Karlik, and E.A. Fyrberg. 1988. Troponin of asynchronous flight muscle. J. Mol. B io l 204: 621-637.

Bullard, B. 1983. Contractile proteins of insect flight muscle. Trends Biochem. Set 8(2): 68-70.

Carroll, A.G., and P.D. Wagner. 1989. Protein kinase C phosphorylation of thymus myosin. J. Mus. Res. Cell MotiL 10: 379-384.

Cavener, D.R. 1987. Comparison of the consensus sequence flanking translational start sites in Drosophila and vertebrates. Nucleic Acids Res. 15(4): 1353-1361.

Chacko, S., M.A. Conti, and R.S. Adelstein. 1977. Effect of phosphorylation of smooth muscle myosin on actin activation and calcium regulation. Proc. Nall Acad. Set U.S.A. 74: 129-133. 180

Cheng, H.C., S.M. Van Patten, A.J. Smith, D.A. Walsh. 1985. An active twenty-amino acid residue peptide derived from the inhibitor protein of the cAMP-dependent protein kinase. B iochem J. 231: 655-661.

Cohen, P. 1989. The structure and regulation of protein phosphatases. Arm. Rev. Biochem. 58: 453-508.

Cohen, P., A. Aitken, Z. Damuni, B.A. Hemmings, T.S. Ingebritsen, P.J. Parker, C. Picton, T. Resink, A.A. Stewart, N.K. Tonks, and J. Woodgett. 1983. Protein phosphorylation and the neural and hormonal control of enzyme activity. In Post-translational covalent modification of proteins, (edited by B. Connor Johnson), pp. 19-38. Academic Press, New York.

Collins, J.H. 1976. Homology of myosin DTNB light chain with alkali light chains, troponin C and parvalbumin. Nature. 259: 699-700.

Crossley, A.C. 1978. The morphology and development of the Drosophila system. In The Genetics and Biology of Drosophila, (edited by M.Ashbumer and T.R.F. Wright), Vol. 2B, pp. 499-560. Academic Press, New York.

Daniel, E.E., A.K. Grover, and C.Y. Kwan. 1983. Calcium. In Biochemistry of smooth muscle (edited by N.L. Stephens), Vol. 3, pp. 1-88. CRC Press, Boca Raton, Florida.

Driska, S.P., and R.A. Murphy. 1979. Phosphorylation of myosin light chain on stimulation of intact smooth muscle. Biophys. J. 25: 73a. Dombradi, V., J.M. Axton, D.M. Glover, and P.T.W. Cohen. 1989. Cloning and chromosomal localization of Drosophila cDNA encoding the catalytic subunit of protein phosphatase la. Eur. J. Biochem. 183: 603-610.

Drummond, D.R., E.S. Hennessey, and J. Sparrow. 1991. Characterization of missense mutations in the A ct88F gene of Drosophila melanogaster. Mol Gen. Genet. 226: 70-80.

El-Saleh, C.S., K.D. Warber, and J.D. Potter. 1986. The role of tropomyosin- troponin in the regulation of skeletal muscle contraction. J. Mus. Res. Cell Motil. 7: 387-804.

Emerson, C.P., and S.I. Bernstein. 1987. Molecular genetics of myosin. Ann. Rev. Biochem. 56: 695-726.

England, P.J. 1980. Regulation by phosphorylation and dephosphoiylation. In The Enzymology o f Post-Translational Modification o f Protein, (edited by R.B. Freedman and H.C. Hawkins), Vol. 1, pp. 291-344. Academic Press, New York.

Falkenthal, S., M.G. Graham, and J. Wilkinson. 1987. The indirect flight muscle of Drosophila accumulates a unique myosin alkali light chain isoform. Dev. Biol. 121: 263-272.

Falkenthal, S., V.P. Parker, W.W. Mattox, and N. Davidson. 1984. Drosophila melanogaster has only one myosin alkali light chain gene which encodes a protein with considerable amino acid sequence homology to chicken myosin alkali light chains. Mol. Cell. Biol. 4(5): 956-965. 182

Frearson, N., B.W.W. Focant., S.V. Perry. 1976. Phosphorylation of a light chain component of myosin from smooth muscle. FEBS. le tt 63: 27-32.

Fyrberg, E.A., K.L. Kindle, and N. Davidson. 1980. The actin genes of Drosophila: a dispersed multigene family. C ell 19: 365- 378.

Fyrberg, E.A., C.C. Fyrberg, C. Beall, and D.L. Saville. 1990a. Drosophila melanogaster troponin-T mutations engenders three distinct syndromes of myofibrillar abnormalities. J. Mol Biol 216: 657-675.

Fyrberg, E.A., M. Kelly, E. Ball, C. Fyrberg, and M.C. Reedy. 1990b. Molecular genetics of Drosophila alpha-actinin: mutant alleles disrupt Z disc integrity and muscle insertions. J. Cell Biol. 110: 1999-2011.

Gazith, J., S. Himmelfarb, and W.F. Harrington. 1970. Studies on the subunit structure of myosin. J.Biol. Chem. 245: 15-22.

George E.L., M.B. Ober, and C.P. Emerson, Jr. 1989. Functional domains of the Drosophila melanogaster muscle myosin heavy-chain gene are encoded by alternatively spliced exons. Mol Cell Biol 9(7): 2957-2974.

Gerthoffer, W.T., and R.A. Murphy. 1983. Myosin phosphorylation and regulation of the cross-bridge cycle in trachael smooth muscle. Am. J. Physiol. 244 (Cell Physiol. 13):C182-187. 183

Graham, M. 1992. In vitro phosphorylation of the recombinant Drosophila myosin light chain-2. Masters Thesis. The Ohio State University.

Guerriero, V.Jr., M.A. Russo, N.J. Olson, J.A. Putkey, A.R. Means. 1986. Domain organization of chicken gizzard myosin light chain kinase deduced from a cloned cDNA. Biochemistry. 25: 8372-8381.

Hanke, P., and R.V. Storti. 1988. The Drosophila melanogaster tropomyosin II gene produces multiple proteins by use of alternative tissue-specific promoters and alternative splicing. Mol. C ell B iol 8(9): 3591-3602.

Hanks, S., A.M. Quinn, and T. Hunter. 1988. The protein kinase family: conserved features and deduced phylogeney of the catalytic domains. Science. 241: 42-52.

Harrington, W. and Rodgers, M. 1984. Myosin. A nn. Rev. Biochem. 53: 35-73.

Harrington, W.F. 1979. Contractile proteins of muscle. In The Proteins, Vol. 4, pp. 245-409. Academic Press, New York.

Hennessey, E.S., D.R. Drummond, and J.C. Sparrow. 1991. Post- translational processing of the amino terminus affects actin function. Eur. J. Biochem. 197: 345-352.

Herring, B.P., J.T. Stull, and P.J. Gallagher. 1990. Domain characterization of rabbit skeletal muscle myosin light chain kinase. J. Biol Chem. 265(3): 1724-1730. 184

Hidaka, H., M. Haka, T. Yamaki. 1979. Effect of novel specific myosin light chain kinase inhibitors on Ca2+-activated Mg2+- ATPase of chicken gizzard actomyosin. Biochem. Biophys. Res. Commun. 90: 694-699.

Horovitch, S.J., R.V. Storti, A. Rich, and M.L. Pardue. 1979. Multiple actins in Drosophila melanogaster. J. Cell. Biol 82: 86-92.

House, C., and B.E. Kemp. 1987. Protein kinase C contains a pseudosubstrate prototype in its regulatory domain. Science 238: 1726.

Huddart, H. 1975. The mechanics of muscles. In The Comparative Structure and Function of Muscle. Pergamon Press, Oxford, pp. 296- 318.

Huxley, H.E. 1969. The mechanism of muscular contraction. Science 164: 1356-1366.

Huxley, H.E. 1957. J. Biophys. Biochem. Cytol 3: 631

Huxley, H.E. 1953. Electron microscopic studies of the organization of the filaments in striated muscle. Biochim. biophys. Acta. 12: 387-394.

Huxley, A.F., and R. Niedergerke. 1954. Interference microscopy of living muscle fibers. Nature. 173: 971-973.

Huxley, H.E., and J. Hanson. 1954. Changes in the cross- striations of muscle during contraction and stretch and their structural interpretation. Nature. 173: 973-976. 185

Ikebe, M., and S. Reardon. 1990. Phosphorylation of bovine platelet myosin by protein kinase C. Biochemistry. 29: 2713-2720.

Ikura, M., G.M. Clore, A.M. Gronenebom, G. Zhu, C.B. Klee, and A. Bax. 1992. Solution structure of a calmodulin-target peptide complex by multidimensional NMR. Science. 256: 632-638.

Ito, M., V. Guerriero, X. Chen, and D.J. Hartshome. 1991. Definition of the inhibitory domain of smooth muscle myosin light chain kinase by site-directed mutagenesis. Biochemistry 30: 3498- 3503.

Jerzmanowski, A., and M. Maleszewski. 1985. Phosphorylation and methylation of P hysarum histone HI during mitotic cycle. Biochemistry 24: 2360-2367.

Kalderon, D., and G.M. Rubin. 1988. Isolation and characterization of Drosophila cAMP-dependent protein kinase genes., Genes & Dev. 2: 1539-1556.

Kalderon, D., and G.M. Rubin. 1989. cGMP-dependent protein kinase genes in Drosophila. J. Biol Chem. 264(18): 10738-10748.

Karlik, C.C., and E.A. Fyrberg. 1986. Two Drosophila melanogaster tropomyosin genes: structural and functional aspects. Mol. Cell. B iol 6: 1965-1973.

Kamm, K.E., and J.T. Stull. 1985. The function of myosin and myosin light chain kinase phosphorylation in smooth muscle. Ann. Rev. Pharmacol. Toxicol 25: 593-620. 186

Karess, R.E., and G.M. Rubin. 1984. Analysis of P transposable element function in Drosophila. Cell 38: 135-146.

Karess, R.E., X. Chang, K.A. Edwards, S. Kulkami, I. Aguilera, and D.P. Kiehart. 1991. The regulatory light chain of nonmuscle myosin is encoded by spaghetti-squash, a gene required for cytokinesis in Drosophila. Cell 65: 1177-1189.

Kawai, M., and P.W. Brandt. 1980. Sinusoidal analysis: a high resolution method for correlating biochemical reactions with physiological processes in activated skeletal muscles of rabbit, frog and crayfish. J. Mus. Res. Cell Motil. 1: 279-303.

Kemp, B.E., R.B. Pearson, and C. House. 1983. Role of basic residues in the phosphorylation of synthetic peptides by myosin light chain kinase. Proc. Natl. Acad. Set 80: 7471-7475.

Kemp, B.E., and R.B., Pearson. 1985. Spatial requirements for location of basic residues in peptide substrates for smooth muscle myosin light chain kinase. J. Biol. Chem. 260(6): 3344-3349.

Kemp, B.E., R.B. Pearson, V. Guerriero, I.C. Bagchi, and A.R. Means. 1987. The calmodulin binding domain of chicken smooth muscle myosin light chain kinase contains a pseudosubstrate sequence. J. Biol. Chem. 262(6): 2542-2548.

Kemp, B.E., and J.T. Stull. 1990. Myosin light chain kinases. In Peptide and Protein Phosphorylation, (edited by B.E. Kemp) CRC Press, Boca Raton, FL. pp. 115-133. 187

Kemp, B.E., R.B. Pearson, V. Guerriero, I.C. Bagchi, and A.R. Means. 1987. The calmodulin binding domain of chicken smooth muscle myosin light chain kinase contains a pseudosubstrate sequence. J. Biol. Chem. 262(6): 2542-2548.

Kennelly, P.J., A.M. Edelman, A.M. Blumenthal, and E.G. Krebs. 1987. Rabbit skeletal muscle myosin light chain kinase, the calmodulin binding domain as a potential active site-directed inhibitory domain. J. Biol Chem. 262: 11958-11963.

Kiehart, D.P., M.S. Lutz, D. Chan, A.S. Ketchum, R.A. Laymon, B. i Nguyen, and L.S.B. Goldstein. 1989. Identification of the gene for fly non-muscle myosin heavy chain: Drosophila myosin heavy chains are encoded by a gene family. EMBO 8(3): 913-922.

Knighton, D.R., R.B. Pearson, J.M. Sowdski, A.R. Means, L.F. Ten Eyck, S.S. Taylor, and B.E. Kemp. 1992. Structural basis of the intrasteric regulation of myosin light chain kinases. Science. 258: 130-135.

Koyama, K., and J. Kendrick-Jones. 1986. The inhibitory Ca2+- regulation of the actin-activated MgATPase activity of myosin from Physarum polycephalum plasmodia. J. Biochem. 99: 1433-1446.

Kreuz, A. 1993. PhD. Thesis. The Ohio State University.

Kwon, H., F.D. Melandri, and A.G. Szent-gyorgyi. 1992. Role of gizzard myosin light chains in calcium binding. J. Mus. Res. Cell MotiL 13: 315-320. 188

Laemmli, U.K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London). 270: 680-685.

Lehman, W., and A.G., Szent-Gyorgyi. 1975. Regulation of muscular contraction; distribution of actin and myosin control in the animal kingdom. J. Gen. Physiol. 66: 1-30.

Li, H.C. 1982. Phosphoprotein phosphatases. In Current Topics In Cellular Regulation. Vol. 21, pp. 129-173. Academic Press, New York.

Lindsley, D.L., and G.G. Zimm. 1992. In The genome o f Drosophila melanogaster. Academic Press, California.

Lund, J., M.R. Webb, and D.C.S. White. 1988. Changes in the ATPase activity of insect fibrillar flight muscle during sinusoidal length oscillation probed by phosphate-water oxygen exhange. J. B io l Chem. 263: 5505-5511.

Malta, T., J. Chen, and G. Matsuda. 1981. Amino-acid sequence of the 20,000 molecular-weight light chain of chicken gizzard- muscle myosin. Eur. J. Biochem. 117:417-424.

Maniatis, T., E.F. Fritsch, and J. Sambrook. 1982. In Molecular Cloning. A laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York.

Maruyama, K. 1986. Connectin, an elastic filamentous protein of striated muscle. Int. Rev. Cytol 104: 81-114. 189

Matsuda, G., T. Maita, Y. Kato, J. Chen, and T. Umegame. 1981. Amino acid sequence of the cardiac L-2A, L-2B, and gizzard 17,000 M light chains of chicken muscle myosin. FEBS {eft. 135: 232-236.

Meador, W.E., A.R. Means, and F.A. Quiocho. 1992. Target enzyme recognition by calmodulin: 2.4 °A structure of a calmodulin-peptide complex. Science. 257: 1251-1254. <

Michnoff, C.H., B.E. Kemp, and J.T. Stull. 1986. Phosphorylation of synthetic peptides by skeletal muscle myosin light chain kinase. J. Biol Chem. 260(18): 8320-8326.

Miller, A. 1950. The internal anatomy and histology of the imago of Drosophila melanogaster. In Biology of Drosophila, (edited by M. Demerc), pp. 420-534. John Wiley & sons, Inc., New York.

Moffat, B.A., and F.W. Studier. 1987. T7 lysozyme inhibits transcription by T7 RNA polymerase. C ell 49: 221-227.

Mogami, K., S.C. Fujita, and Y. Hotta. 1982. Identification of Drosophila indirect flight muscle myofibrillar proteins by means of two-dimensional electrophoresis. J. Biochem. 91: 643-650.

Molloy, J.E., V. Kyrtatus, J.C. Sparrow, and D.C.S. White. 1987. Kinetics of flight muscles from insects with different wing beat frequencies. Nature (London) 328: 449-451.

Ngai, P.K., and M.P. Walsh. 1984. Inhibition of smooth muscle actin-activated myosin Mg2+-ATPase activity by caldesmon. J. B io l Chem. 259(22): 13656-13659. 190

Nomura, Y., and S. Ebashi. 1980. Calcium regulatory mechanism in vertebrate smooth muscle. Biomed. Res . 1: 1-14.

Oakley, B.R., D.R. Kirsch, and R.N. Morris. 1980. A simplified ultrasensitive silver stain for protecting proteins in polyacrylamide gels. Anal Biochem. 1 0 5 : 361-363.

O'Farrell, P.H. 1975. High resolution two-dimensional electrophoresis of proteins. J. Biol Chem. 250: 4007-4021.

Olson, N.J., R.B. Pearson, D.S. Needleman, M.Y. Horwitz, B.E. Kemp, and A.R. Means. 1990. Regulatory and structural motifs of chicken gizzard myosin light chain kinase. Proc. Natl Acad. Set 8 7 : 2284-2288.

Parker, V.P., S. Falkenthal, and N. Davidson. 1985. Characterization of the myosin light chain-2 gene of Drosophila melanogaster. Mol Cell Biol 5: 3058-3068.

Pearson, R.B., R.E.H. Wettenhall, A.R. Means, D.J. Hartshome, and B.E. Kemp. 1988. Autoregulation of enzymes by pseudosubstrate prototypes: myosin light chain kinase. Science. 241: 970-973.

Pearson, R.B., M. Ito, N.A. Morrice, A.J. Smith, R. Condron, R.E.H. Wettenhall, B.E. Kemp, and D.J. Harshome. 1991. Proteolytic cleavage sties in smooth muscle myosin light chain kinase and their relation to structural and regulatory domain. Eur. J. Biochem. 200: 723-730. Peckham, M., J.E. Molloy, J.C. Sparrow, and D.C.S. White. 1990. Physiological properties of the dorsal longitudinal flight muscle and the tergal depressor of the trochanter muscle of Drosophila melanogaster. J. Mus. Res. Cell MotiL 11: 203-215.

Pemrick, S. 1977. Comparison of the calcium sensitivity of actomyosin from native and ^-deficient myosin. Biochemistry 16(18): 4047-4054.

Pemrick, S. 1980. The phosphoiylated L 2 light chain of skeletal myosin is a modifier of the actomyosin ATPase. J. Biol. Chem. 255(18): 8836-8841.

Perrie, W.T., L.B. Smillie, and S.V. Perry. 1973. A phosphorylated light-chain component of myosin from skeletal muscle. Biochem J. 135: 151-156.

Persecchini, A., U. Mrwa, D. Hartshome. 1981. Effect of phosphorylation on the actin-activated ATPase activity of myosin. Biochem. Biophys. Res. Common. 98: 800-805.

Price, M.G., and J.W. Sanger. 1983. Intermediate filaments in striated muscle: A review of structure studies in embryonic and adult skeletal and cardiac muscle. In Cell and Muscle Motility, (edited by R.M. Dowben and J.W. Shay), Vol. 3, pp. 1-40. Plenum Press, New York.

Pringle, J.W.S. 1949. The excitation and contraction of the flight muscles of insects. J. Physiol., Lond. 108: 226-232.

Pringle, J.W.S. 1978. Stretch activation of muscle: function and mechanism. Proc. R. Soc. Lond. 201: 107-130. Rayments, I., W. Rypniewski, K. Schmidt-Base, R. Smith, D.R. Tomchick, M.M. Benning, D.A. Winklemann, G. Wesenberg, H.M. Holden. 1993a. Three-dimensional structure of myosin subfragment-1. A molecular motor. Science. 261: 50-58.

Rayments, I., H.M. Holden, M. Whittaker, C.B. Yohn, M. Lorenz, K.C. Holmes, and R.A. Milligan. 1993b. Structure of the actin- myosin complex and its implications for muscle contraction. Science. 261: 58-65.

Rees, D.D., and D.W. Frederiksen. 1981. Calcium regulation of porcine aortic myosin. J. Biol. Chem. 259: 357-364.

Rosenberg, A.H., B.B. Lade, D.S.Chui, S.W. Lin, J.J. Dunn, and F.W. Studier. 1987. Vectors for selective expression of cloned DNA's by T7 polymerase. Gene 56: 125-135.

Roush, C.L., P.J. Kennelly, M.B. Glaccum, D.M. Helfman, J.D. Scott, and E.G. Krebs. 1988. Isolation of the cDNA encoding rat skeletal muscle MLCK, sequence and tissue distribution. J. Biol. Chem. 263(21): 10510-10516.

Saide, J.D., S. Chin-Bow, J. Hogan-Sheldon, L. Busquets-Tumer, J.O. Vigoreaux, K. Valgeirsdottir, and M.L. Pardue. 1989. J. Cell B iol 109: 2157-2167.

Schaeffer, E., D. Smith, G. Mardon, W. Quinn, and C. Zuker. 1989. Isolation and characterization of two new protein kinase C genes, including one specifically expressed in photoreceptor cell. Cell. 57: 403-412. 193

Sellers, J.R., P.D. Chantler, and A.G. Szent-gyorgyi. 1980. Hybrid formation between scallop myofibrils and foreign regulatory light- chains. J. Mol Biol. 144: 223-245.

Sellers, J.R., and R.S. Adelstein. 1987. Regulation of contractile activity. In The Enzymes, Vol. 13, pp. 381-418. Academic Press, New York.

Silver, P.J., and J.T. Stull. 1982. Regulation of myosin light chains and phosphoiylase phosphorylation in trachael smooth muscle. J. Biol. Chem. 257: 6145-6150.

Small, J.V., and K.A. Sobieszek. 1977. Calcium regulation of mammalian smooth muscle actomyosin via a kinase-phosphatase- dependent phosphorylation and dephosphorylation of the 20,000- Mr light chain of myosin. Eur. J. Biochem. 76: 521-530.

Sobieszek, A. 1977. Vertebrate smooth muscle myosin. Enzymatic and structural properties. In The Biochemistry of Smooth Muscles, (edited by N.L. Stephens), pp. 413-443. University Park, Baltimore.

Sparrow, M.P., U. Mrwa, F. Hofmann, and J.C. Ruegg. 1981. Calmodulin is essential for smooth muscle contraction. FEBS lett. 125: 141-145.

Spradling, A.C. 1986. P element-mediated transformation. In Drosophila: A Practical Approach, (edited by D.B. Roberts), pp. 175- 197. ORL Press, Oxford. 194

Steiger B.J., and J.C. Ruegg. 1969. Energetics and efficiency in the isolated contractile machinery of an insect fibrillar muscle at various frequencies of oscillation. PJlugers Arch. 307(1): 1-21.

Stull, J.T., M.H. Nunnally, R.L. Moore, and D.K. Blumenthal. 1985. Myosin light chain kinases and myosin phosphorylation in skeletal muscle. In Advance in Enzyme Regulation. Vol 23, pp. 123- 140. Pergamon Press, New York.

Stull, J.T., M.H. Nunnally, and C.H. Michnoff. 1986. Calmodulin-dependent protein kinases. In The Enzymes, Vol. 17, pp. 113-165. Academic Press, New York.

Sweeney, H.L., B.F. Bowman, and J.T. Stull. 1993. Myosin light chain phosphorylation in vertebrate striated muscle: regulation and function. Am. J. Physiol 264:C 1085-1095.

Szent-Gyorgyi, A.G., E.M. Szentkiralyi, and J. Kendrick-Jones. 1973. The light chains of scallop myosin as regulatory subunits. J. Mol Biol 74: 179-203.

Takahashi, S., and K. Maruyama. 1987. Activity changes in myosin ATPase during metamorphosis of fruitfly. Zoological science. 4: 833-838.

Takahashi, S., H. Takano-Ohmuro, and K. Maruyama. 1990a. Regulation of Drosophila myosin ATPase activity by phosphorylation of myosin light chain-I. wild-type fly. Comp. Biochem. Physiol 95B(1): 179-181. 195

Takahashi, S., H. Takano-Ohmuro, K. Maruyama, and Y. Hotta. 1990b. Regulation of Drosophila myosin ATPase activity by phosphorylation of myosin light chain-II. flightless mfd- fly. Comp. Biochem. Physiol 95B(1): 183-185.

Takano-Ohmuro, H., S. Takahashi, G. Hirose, and K. Maruyama. 1990. Phosphorylated and dephosphoiylated myosin light chains of Drosophila fly and larva. Comp. Biochem. Physiol. 95B(1): 171- 177.

Takano-Ohmuro, H., G. Hirose, and T. Mikawa. 1983. Separation and identification of Drosophila myosin light chains. J. Biochem. 94: 967-974.

Tan, J.L., S. Ravid, and J.A. Spudich. 1992. Control of nonmuscle myosins by phosphorylation. Ann. Rev. Biochem. 61: 721-759.

Tan, J.L., and J.A. Spudich. 1990. Dictyosteliiun myosin light chain kinase: purification and characterization. J. Biol. Chem. 265(23): 13818-13824.

Tanaka, Y., K. Maruyama, T. Mikawa, and Y. Hotta. 1988. Identification of calcium binding proteins in two-dimensional gel electrophoretic pattern of Drosophila thorax and their distribution in two types of muscles. J. Biochem. 104: 489-491.

Tobin, S.L., E. Zulauf, F. Sanchez, E.A. Craig, and B.J McCarthy. 1980. Multiple actin-related sequences in the Drosophila melanogaster genome. Cell 19: 121-131. 196

Toffenetti, J., D. Mischke, and M.L. Pardue. 1987. Isolation and characterization of the gene for myosin light chain two of Drosophila melanogaster. J. CeU. Biol 104: 19-28.

Toyoshima, Y.Y., S.J. Kron, E.M. McNally, K.R., Niebling, C. Toyoshima, and J.A. Spudich. 1987. Myosin subfragment-1 is sufficient to move actin filaments in vitro. Nature. 328: 536-539.

Vale, R.D., and L.S.B. Goldstein. 1990. One motor, many tails: an expanding repertoire of force-generating enzymes. Cell 60: 883- 885.

Vigoreaux, J.O., Saide, J.D., and M.L. Pardue. 1991. Structurally different Drosophila stiated muscles utilize distinct variants of Z- band-associated proteins. J. Mus. Res. Cell MotiL 12: 340-334.

Vigoreaux, J.O., J.D. Saide, K. Valgeirsdottir, and M.L. Pardue. 1992. Flightin, a novel myofibrillar protein of Drosophila stretch- activated muscles. J. Cell Biol. 121(3): 587-598.

Villee, C.A., E.P. Solomon, and P.W. Davis. 1985. In Biology. CBS College Publishing, Philadelphia.

Wagner, P.D., and E. Giniger. 1981. Hydrolysis of ATP and reversible binding of F-actin by myosin heavy chains free of all light chains. Nature. 292: 560-562.

Wahle, E., and W. Keller. 1992. The Biochemistry of 3'-end cleavage and polyadenylation of messenger RNA precursors. Arm. Rev. Biochem, 61: 419-440. 197

Walsh, M.P. 1985. Calcium regulation of smooth muscle contraction. In Calcium and Cell Physiology, (edited by D. Marme), pp. 170-204. Springer-Verlag.

Wold, F. 1981. In vivo chemical modification of proteins (post- translational modification). Ann. Rev. Biochem. 50: 783-814.

Warmke, J. 1990. Ph.D. Thesis. The Ohio State University.

Warmke, J., M. Yamakawa, J. Molloy, S. Falkenthal, and D. Maughan. 1992. A myosin light chain-2 mutation affects flight, wing beat frequency and indirect flight muscle contraction kinetics in Drosophila. J. Cell Biol 119(6): 1523-1539.

Warrick, H.M., and J.A. Spudich. 1987. Myosin structure and function in cell motility. Ann. Rev. Cell Biol 3: 379-421.