InteractiveFly: GeneBrief

MICAL: Biological Overview | References

The overall size and structure of a synaptic terminal is an important determinant of its function. In a large-scale mutagenesis screen, designed to identify Drosophila mutants with abnormally structured neuromuscular junctions (NMJs), mutations were discovered in Drosophila mical, a conserved gene encoding a multi-domain protein with a N-terminal monooxygenase domain. In mical mutants, synaptic boutons do not sprout normally over the muscle surface and tend to form clusters along synaptic branches and at nerve entry sites. Consistent with high expression of MICAL in somatic muscles, immunohistochemical stainings reveal that the subcellular localization and architecture of contractile muscle filaments are dramatically disturbed in mical mutants. Instead of being integrated into a regular sarcomeric pattern, actin and myosin filaments are disorganized and accumulate beneath the plasmamembrane. Whereas contractile elements are strongly deranged, the proposed organizer of sarcomeric structure, D-Titin, is much less affected. Transgenic expression of interfering RNA molecules demonstrates that MICAL is required in muscles for the higher order arrangement of myofilaments. Ultrastructural analysis confirms that myosin-rich thick filaments enter submembranous regions and interfere with synaptic development, indicating that the disorganized myofilaments may cause the synaptic growth phenotype. As a model, it is suggested that the filamentous network around synaptic boutons restrains the spreading of synaptic branches (Beuchle, 2007).

Members of the semaphorin family of secreted and transmembrane proteins utilize plexins as neuronal receptors to signal repulsive axon guidance. It remains unknown how plexin proteins are directly linked to the regulation of cytoskeletal dynamics. Drosophila MICAL, a large, multidomain, cytosolic protein expressed in axons, interacts with the neuronal plexin A (PlexA) receptor and is required for Semaphorin 1a (Sema-1a)-PlexA-mediated repulsive axon guidance. In addition to containing several domains known to interact with cytoskeletal components, MICAL has a flavoprotein monooxygenase domain, the integrity of which is required for Sema-1a-PlexA repulsive axon guidance. Vertebrate orthologs of Drosophila MICAL are neuronally expressed and also interact with vertebrate plexins, and monooxygenase inhibitors abrogate semaphorin-mediated axonal repulsion. These results suggest a novel role for oxidoreductases in repulsive neuronal guidance (Terman, 2002).

To identify mediators of semaphorin-dependent repulsive axonal guidance, the terminal highly conserved 'C2' portion of the PlexA cytoplasmic domain was used to search for interacting proteins encoded by a Drosophila embryonic (0-24 hr) yeast two-hybrid cDNA library. The strongest interactor has been called MICAL, covers >41 kb of genomic sequence and has at least 25 exons. Based on analysis of isolated cDNAs and Western analysis, there are at least three MICAL isoforms ('long,' 'medium,' and 'short' variants) (Beuchle, 2007).

Drosophila MICAL is named for its recently characterized vertebrate ortholog, MICAL-1 (for molecule interacting with CasL), which has been shown to associate with CasL (see Drosophila p130CAS) and vimentin in nonneuronal cells (Suzuki, 2002). Within the plexin-interacting region in the C terminus identified in the screen described in this paper, there is a predicted heptad-repeat, coiled-coil structure. Interestingly, this region of MICAL shares amino acid similarity with several other coiled-coil domain-containing proteins, including a portion of the alpha domain found in the Ezrin, Radixin, and Moesin (ERM) proteins (~22% identity). N-terminal to this domain there is a region rich in prolines, and the last four amino acids of MICAL (ESII) are a PDZ protein binding motif. There are two regions of varying length, with no significant similarity to other proteins, which appear to determine the size of the different MICAL proteins. MICAL has a single LIM domain, a protein-protein interaction module found in a variety of proteins involved in signal transduction cascades and in cytoskeletal organization, and also a single calponin homology (CH) domain, a domain also found in cytoskeletal and signal transduction proteins and known to be involved in actin filament binding. The MICAL N-terminal ~500 aa is highly conserved among MICAL-related proteins but is unique over its entire length in comparison to other proteins (Terman, 2002).

In situ hybridization analysis using RNA probes corresponding to the N or C terminus of MICAL shows that MICAL and PlexA have similar patterns of embryonic mRNA expression. During early Drosophila development (stages 7-8), both are expressed in the ventral neurogenic region and in many nonneuronal tissues (including developing mesoderm, cells surrounding the cephalic furrow and amnioproctodeal invagination, and in gut primordia). This nonneuronal expression is also seen later in embryonic development (stages 11-17), where both are present within the anterior and posterior midgut primordia, the visceral musculature, and weakly in somatic musculature. During axonal pathfinding (stage 13 onward), both are expressed within the developing brain and ventral nerve cord in most, if not all, CNS neurons, but MICAL, like Sema1a and PlexA, is not highly expressed in peripheral sensory neurons (Terman, 2002).

Western blot analysis using a polyclonal antibody directed against the MICAL C terminus (MICAL-CT) has revealed prominent bands at 530 kDa, 330 kDa, 300 kDa, 200 kDa, and 125 kDa in lysates from wild-type embryos that increase in intensity in lysates from embryos harboring a chromosomal duplication that includes the MICAL locus. The three largest protein bands correspond to the predicted molecular weights of the three MICAL cDNAs (Terman, 2002).

MICAL protein is present in neuronal cell bodies, along axons, and in growth cones. MICAL immunostaining first appears in the nervous system at stage 13 and labels motor and CNS projections, and at later embryonic stages, it is present on axons that make up all motor axon pathways: the intersegmental nerve (ISN), the intersegmental nerves b and d (ISNb and ISNd), and the segmental nerves a and c (SNa and SNc). MICAL immunostaining is also present in segment boundaries at the position of muscle attachment sites and at low levels in the lateral cluster of chordotonal organs (Terman, 2002).

MICAL proteins have conserved protein domains with identical organization in all family members and a high degree of amino acid identity among these domains in different MICALs. There is one MICAL in Drosophila and three mammalian MICALs. The MICALs appear unique with respect to containing both calponin homology (CH) and LIM domains, in addition to their conserved N- and C-terminal regions. There is a family of MICAL-like (MICAL-L) proteins, members of which have a similar organization to MICALs but lack the region N-terminal to the CH domain. There is one MICAL-L protein in Drosophila (MICAL-like) and at least two family members in humans. D-MICAL-L cDNA and genomic DNA sequence information suggest that D-MICAL-L begins just N-terminal to the CH domain. Analysis of publicly available mammalian cDNA and genomic sequences suggests that human MICAL-L1 and MICAL-L2 are similar in overall domain organization to D-MICAL-L and do not contain the highly conserved ~500 aa MICAL N-terminal domain (Terman, 2002).

The high degree of conservation of the MICAL N terminus among family members (up to 62% identical between flies and humans) suggests that this domain is functionally important. Upon closer examination of the 500 aa conserved N-terminal region, a consensus dinucleotide binding sequence, GXGXXG was found, which is distinct from the sequence present in classical mononucleotide binding motifs. Further, this region contains three separate motifs found in flavoprotein monooxygenases (also called hydroxylases), a subclass of oxidoreductases. The amino acid sequence surrounding the GXGXXG motif matches perfectly the consensus sequence for the ADP binding region of flavin adenine dinucleotide (FAD) binding proteins (Rossmann fold or FAD fingerprint 1), and distinguishes this region from consensus NAD, or NADP binding folds. MICALs also have a well-conserved GD motif (FAD fingerprint 2) C-terminal to the FAD fingerprint 1 region, which is important for binding the ribose moiety of FAD. Finally, MICALs have the conserved DG motif between the FAD fingerprint 1 and 2 motifs that has been reported to be involved in binding the pyrophosphate moiety of FAD. Proteins with these consensus FAD binding regions use FAD in the catalysis of oxidation-reduction reactions. Flavoprotein monooxygenases are oxidoreductases (enzymes that catalyze oxidation and reduction reactions) that catalyze the insertion of one atom of molecular oxygen into their substrate using nucleotides as electron donors. These monooxygenases are also defined by their use of FAD as a coenzyme. Apart from these three consensus regions, monooxygenases vary significantly, reflecting the wide range of enzymes in this family and their variable substrate binding pockets also encompassed within this domain. However, MICALs and other monooxygenases show significant similarity within these three FAD binding regions and also similar spacing of these regions within the monooxygenase domain (Terman, 2002).

Biochemical and genetic analyses strongly suggest that MICALs contain functional FAD binding monooxygenase domains required for mediating plexin signaling. In support of this idea, inhibition of flavoprotein monooxygenase enzymatic activity dramatically attenuates semaphorin-mediated axon repulsion and growth cone collapse. However, though the inhibitor EGCG has a high degree of selectivity for flavoprotein monooxygenases, similar concentrations of EGCG inhibit other enzymes including steroid 5alpha-reductase, NADPH-cytochrome P450 reductase, telomerase, matrix metalloproteinases MMP-2 and MMP-9, and phenol sulfotransferase. Although most of these enzymes are unlikely to be expressed in the growth cones of DRG axons, potential nonspecific effects of these inhibitors cannot be ruled out despite their demonstrated selectivity for monooxygenases. Taken together with the in vivo Drosophila experiments showing a requirement for the MICAL FAD binding region in Sema-1a-mediated axon repulsion, these data suggest redox signaling plays an important role in vertebrate semaphorin-mediated axonal repulsion (Terman, 2002).

Flavoprotein monooxygenases specifically catalyze the oxidation of a number of substrates, and in some contexts they can function as oxidases and generate reactive oxygen species. These results suggest that MICALs are flavoproteins most similar to the flavoprotein monooxygenase family of oxidoreductases, but a complete understanding of the chemical nature of the reactions catalyzed by MICALs awaits future study and identification of substrates. The redox regulation of amino acid residues within signaling proteins (including kinases, phosphatases, small GTPases, guanylate cyclases, and adaptor proteins) and cytoskeletal proteins (including actin, actin binding proteins, intermediate filament proteins, and GAP-43) has been shown to modulate their function. In addition, oxidation of actin leads to disassembly of actin filaments, instability and collapse of actin networks, reduced ability of actin to interact with actin crosslinking proteins, and a decrease in the ability of actin monomers to form polymers. Finally, it is also interesting that MICALs have a putative actin filament binding domain (CH domain) and that MICAL-1 interacts with vimentin, an intermediate filament protein (Terman, 2002).

The proline-rich region of vertebrate MICAL-1 interacts with the SH3 domain of the adaptor protein CasL (HEF1) in nonneuronal cells. CasL, along with the related proteins p130Cas and Efs (Sin), make up the Cas family of proteins that assembles and transduces intracellular signals that stimulate cell migration and axon outgrowth. These proteins have numerous protein-protein interaction domains, including a Src-homology 3 (SH3) domain, multiple SH2 binding sites in their substrate domain, several proline-rich motifs, and a C-terminal dimerization module. This structure suggests a role for Cas family proteins as docking molecules, and numerous interacting proteins have been identified, including kinases (e.g., FAK, Src, and Abl), phosphatases (e.g., PTP-1B and SHP2), GEFs (e.g., C3G), and adaptor proteins (e.g., Nck, Crk, Grb2, and 14-3-3). Studies indicate that Cas proteins localize mainly to focal adhesions and stress fibers and that they are required in integrin-dependent cell migration and actin filament assembly. Cas proteins, therefore, may play an important role in plexin-mediated repulsive and attractive guidance events (Terman, 2002).

Drosophila MICAL regulates myofilament organization and synaptic structure

In a postembryonic screen designed to identify mutations that affect the structure, maintenance and remodeling of neuromuscular synapses, a complementation group was isolated that harbors point mutations in Drosophila mical. Synaptic boutons in mical mutant NMJs have the tendency to cluster around initial nerve-muscle contact sites and along synaptic branches and therefore fail to spread properly along the muscle fiber. The dense bouton accumulations in mical mutants develop postembryonically, starting in first instar larvae. Embryonic NMJs were not visibly altered suggesting that mical mutations do not affect synapse formation per se but synaptic remodeling during the period of rapid larval growth (Beuchle, 2007).

Human MICAL-1 was first discovered by Far Western screening using the SH3-domain of human CasL to probe an expression library. Cas family proteins generally serve as docking molecules that assemble intracellular complexes to transduce extracellular signals, especially at focal adhesions where they regulate the anchorage of integrins to actin-rich stress fibers. MICAL-1 was therefore implicated in the regulation of the cytoskeleton (Suzuki, 2002). Human MICAL-1 has also been shown to interact with Rab1, a small GTPase that plays a role in vesicle trafficking (Weide, 2003), and to assemble into a filamentous network when transfected into tissue culture cells (Suzuki, 2002; Fischer, 2005). In Drosophila, MICAL has been shown to regulate embryonic motor axon guidance by binding to Plexin A, a co-receptor for Semaphorin 1A (Terman, 2002). Consequently, MICAL was suggested to control cytoskeletal changes in migrating axons in response to repulsive environmental cues. Embryonic motor axon guidance errors can persist into larval stages, with NMJs forming at unusual positions or being entirely absent at a high percentage of larval muscles. In mical mutant alleles, however, NMJs generally formed at their wild-type positions, apart from occasional errors on muscles innervated by the SNa pathway. Therefore motor axon trajectories were examined in mutant embryos; relatively mild but statistically significant projection errors were observed. When the original mical deficiency allele was examined, however, motor axon guidance and innervation defects were more pronounced. A possible explanation for this discrepancy could be that Df(3R)swp2^mical deletes 12 annotated genes, whereas the currently isolated alleles carry single point mutations (Beuchle, 2007).

Deleterious point mutations were identified in the mical gene in all seven EMS-induced alleles, the most interesting of which are single, non-truncating missense mutations in mical^I1367 and mical^G158. The mutation in mical^I1367 (G134R) alters a highly conserved glycine residue in the monooxygenase domain that is necessary for binding the co-factor FAD (Nadella, 2005; Siebold, 2005). Loss of the ability to interact with FAD likely disrupts the function of the monooxygenase. In fact, point mutations in this motif have been shown to abolish enzymatic activity in related monooxygenases. FAD and the monooxygenase domain are therefore critically required for MICAL function. The mutation in mical^G158 (G433S) disrupts a glycine residue located in FAD fingerprint 2 that interacts with the ribose moiety of FAD. This rather subtle change alters the subcellular localization of MICAL in muscles and neurons. MICAL accumulates in the cytoplasm of these cells and is not transported to peripheral locations, e.g. along axons. In addition to binding to FAD, this motif may therefore serve as a checkpoint to control for fully folded and functional MICAL proteins. The remaining mutations are stop codons that truncate the protein prior to the LIM domain, suggesting that the LIM-domain and/or other sequences further downstream are equally important for MICAL function, and that the monooxygenase domain by itself is not sufficient. Due to the nature of its protein–protein interaction domains, MICAL has been suggested to act as a molecular platform that interacts with substrate proteins to bring them in close proximity of its monooxygenase domain. The enzymatic modification of specific substrates may therefore be crucial for MICAL function (Beuchle, 2007).

During embryogenesis MICAL is broadly and dynamically expressed in a variety of tissues. With respect to neuromuscular development, MICAL is initially expressed in muscles and, starting with stage 15, in neurons of the central nervous system (CNS). MICAL accumulates preferentially in axons of the CNS and in peripheral regions of muscle fibers, including muscle attachment sites and the cleft between neighboring muscles. MICAL is present in muscles at least until the end of embryogenesis but is not markedly detected in muscles of third instar larvae. This is also consistent with punctate MICAL accumulations in embryonic but not in larval muscles of mical^G158 mutants. At NMJs of third instar larvae, MICAL seems therefore to be expressed predominantly at presynaptic sites. This conclusion is additionally supported by the results of the immunostainings, which detected MICAL predominantly in presynaptic terminals. However, the possibility cannot be excluded that presynaptic MICAL obscures postsynaptically localized proteins. If MICAL has a slow turnover rate, for example, it is quite possible that low levels of embryonic proteins remain stable at postsynaptic sides of NMJs or in muscles until late larval stages. To find further evidence on which side of the NMJ MICAL is required and to determine if the short isoform of MICAL is able to rescue the mutant phenotype, it was expressed in neurons or muscles in heteroallelic mical mutant animals. Neither pre- nor postsynaptic expression could substantially rescue the synaptic or flightless phenotype. One reason could be that the expression level of exogenous MICAL was quite low and higher protein levels would have been required to improve the rescue capability. Since at least three isoforms of MICAL are expressed in embryos, it is also possible that one or both of the other isoforms are functionally required. Indeed, mutations in mical^I666 and mical^G56 are located in exon 9 and 10, respectively, which should be expressed only in the middle and long isoform, indicating that the short isoform cannot compensate for all MICAL functions. As a third possibility, MICAL function may be required simultaneously on both sides of NMJs to coordinate neuromuscular development (Beuchle, 2007).

While MICAL was identified initially through a neuromuscular phenotype, and taking into account the relatively mild guidance defects, the strongest and most prevailing defects were observed in striation and sarcomeric patterning of muscle fibers. MICAL’s primary role may therefore reside in the spatial organization of myofilaments, or more specifically, in the assembly of myofilaments into a sarcomeric pattern. First, the earliest detectable abnormal filaments in mutant first instar larvae resemble embryonic filaments at stage 15–16, which form long needle-like structures prior to the formation of sarcomeres. Second, known mutations in muscle-specific isoforms of Myosin, Actin or Troponin-T that affect sarcomere assembly, lead to loose and scattered myofilaments in the sarcoplasm of indirect flight muscles. Third, MICAL is expressed in muscles at the time of sarcomere assembly but it could not be detected in considerable amounts in muscles of third instar larvae when it should be expressed if it has a function in the maintenance of sarcomeres. With respect to its subcellular localization in muscles, it is worth noting that MICAL is not homogenously distributed in the sarcoplasm but preferentially accumulates at the plasmamembrane and muscle attachment sites. It has been proposed that myofibrillogenesis starts beneath the plasmamembrane. MICAL would therefore be localized correctly to be involved in myofibril assembly. For these reasons, it is more likely that MICAL is required for the proper assembly of sarcomeres rather than for their maintenance or remodeling. However, since the filament phenotype worsens over time, the possibility cannot be excluded that MICAL plays a role in the maintenance of sarcomeres or in the proteolytic destruction of non-assembled muscle filaments. It is interesting to emphasize that non-functional MICAL affects various sarcomeric proteins differently. While myosin and actin filaments were strongly disorganized, the sarcomeric protein D-Titin showed an almost wild-type distribution. Titins are enormous proteins that span the entire half-sarcomere, with their N-terminus being inserted in Z-discs and their C-terminus in M-lines. Titins may therefore function as a molecular scaffold that directs sarcomere assembly. Since the Titin-based sarcomeric backbone seems to be relatively intact in mical mutants, mutations in mical may uncouple sequential steps in the assembly process of sarcomeres (Beuchle, 2007).

For myofilament organization, MICAL appears to be required in muscles only. Expression of mical double-stranded RNA in specific tissues to downregulate endogenous MICAL revealed that only muscle-specific knock-down could mimic the myofilament phenotype and flight defects of mical mutants. It was not possible, however, to reproduce the neuromuscular phenotype in these experiments, regardless of whether MICAL expression was inhibited pre- or postsynaptically. Possible explanations for this observation could be that the myofilament defects were too weak to interfere with synaptic growth or that the inhibition of MICAL expression was incomplete in neurons despite the use of two different neuronal Gal4-lines. Variability in the efficiency of MICAL downregulation was also observed with muscle-specific Gal4-lines. Whereas Mef2-Gal4 reproducibly disrupted muscle striation, the activity of 24B-Gal4 was much less potent. A requirement of MICAL in muscles is further supported by the notion that the assembly of myofilaments is generally considered to be a cell-autonomous process and independent of motoneuronal input, since it occurs in isolated myocytes, cardiomyocytes and myoblast cell lines that are cultured in the absence of any neurons. Furthermore, if MICAL would be required presynaptically for myofilament organization, muscles that lack NMJs should display myofilament defects. In Drosophila sidestep mutants, a high percentage of ventral muscles permanently lack NMJs due to embryonic motor axon bypass phenotypes. However, sarcomeric organization and muscle striation was normal in non-innervated muscles of sidestep mutants, providing further evidence that MICAL is required in muscles to regulate the higher order assembly of myofilaments. The disturbances in the architecture of contractile filaments in mical mutants do not inhibit muscle contraction completely. They seem to interfere, however, with the speed and vigor of contraction cycles, which may be less problematic for larval muscles but may drastically affect metabolically active muscles, such as the highly structured indirect flight muscles, which would explain why mical mutant adults are unable to fly (Beuchle, 2007).

In wild-type larvae, synaptic boutons are located on the surface of the muscle fiber and are fully wrapped by an insulating layer of subsynaptic reticulum (SSR). The SSR begins to form in first instar larvae. Initial membrane invaginations of the postsynaptic membrane develop into a multiply folded stack of membrane cisternae during larval life. In mical mutants, the SSR appears almost completely absent. The remaining invaginations are flattened to sac-like cavities and widely dispersed in the postsynaptic cytoplasm. The SSR has no assigned function but apart from an exchange of metabolites between the extracellular space and the muscle fiber it may represent a specialized region that facilitates synaptic growth. Repetitive imaging of the same set of NMJs during consecutive stages of development has revealed that sprouting buds of dividing boutons grow out and into the surrounding SSR. Newly forming boutons therefore have to displace the membrane stacks of the SSR in order to gain new synaptic territory. Remnants of the SSR will eventually be adopted by the new bouton. In this view, and due to its flexible and dilatable structure, one of the functions of the SSR could be to facilitate bouton outgrowth. Bouton division and outgrowth seems therefore to be facilitated by two factors: superficial location and SSR envelopment. In mical mutants, synaptic boutons are not superficially located and not wrapped by a soft layer of SSR. They are firmly embedded in the muscle tissue. Actin and myosin filaments accumulate in postsynaptic regions and beneath the sarcolemma. As one possibility, it is proposed that the detached filaments corrupt the development of the SSR. The network of interdigitated cytoskeletal elements around synaptic boutons would then make it difficult for budding boutons to push out into the muscle tissue and would thus hinder the spreading of synaptic boutons across the muscle surface during larval growth. Although this model favors postsynaptic functions of MICAL, at present, the possibility that MICAL has also presynaptic functions, which control synaptic structure, cannot be excluded. Under this scenario, the myofilament phenotype is likely to be independent of the synaptic phenotype. Further studies including high efficient tissue-specific inhibition of MICAL expression, will hopefully allow determination of which of these possibilities is correct. Regardless of whether MICAL functions pre- or postsynaptically, the data presented here show that MICAL is a critical regulator of myofilament architecture and synaptic structure during postembryonic development (Beuchle, 2007).

Beuchle, D., Schwarz, H., Langegger, M., Koch, I. and Aberle, H. (2007). Drosophila MICAL regulates myofilament organization and synaptic structure. Mech. Dev. 124(5): 390-406. Medline abstract: 17350233

Fischer, J., Weide, T. and Barnekow, A. (2005). The MICAL proteins and rab1: a possible link to the cytoskeleton? Biochem. Biophys. Res. Commun. 328: 415-423. Medline abstract: 15694364

Nadella, M., et al. (2005). Structure and activity of the axon guidance protein MICAL, Proc. Natl. Acad. Sci. 102: 16830-16835. Medline abstract: 16275926

Siebold, C., et al. (2005). High-resolution structure of the catalytic region of MICAL (molecule interacting with CasL), a multidomain flavoenzyme-signaling molecule, Proc. Natl. Acad. Sci. 102: 16836-16841. Medline abstract: 16275925

Suzuki, T., Nakamoto, T., Ogawa, S., Seo, S., Matsumura, T., Tachibana, K., Morimoto, C. and Hirai, H. (2002). MICAL, a novel CasL interacting molecule, associates with vimentin. J. Biol. Chem. 277: 14933-14941. Medline abstract: 11827972

Terman, J. R., et al. (2002). MICALs, a family of conserved flavoprotein oxidoreductases, function in Plexin-mediated axonal repulsion. Cell 109: 887-900. Medline abstract: 12110185

Weide, T., et al. (2003). MICAL-1 isoforms, novel rab1 interacting proteins, Biochem. Biophys. Res. Commun. 306: 79-86. Medline abstract: 12788069

date revised: 2 December 2007

Home page: The Interactive Fly © 2007 Thomas Brody, Ph.D.