Johnston's Organ development from the perspective of the dendritic cap
We focused on JO development after pupal head eversion, which occurs approximately 14 hrs after puparium formation (APF) at 23°C. Dendritic caps were labeled with endogenously expressed GFP-NompA (). At the time of head eversion, wild type JO is organized as a wide circle; most caps are globular, with few appearing elongated (, 14 hrs APF and inset). As development progresses, the caps elongate and approach one another () as the a2/a3 joint constricts. Caps are closely juxtaposed to the space between a2 and a3 as early as 16 hrs APF (), suggesting that apical connection is already established.
ck/MyoVIIA is necessary for JO organization during development.
These data indicate that NompA, possibly along with other, unknown cap components, is at least partly produced by the time of head eversion (). As pupation continues, the distance between scolopidia and the future a2/a3 joint increases through the transition of JO from a flat, 2 dimensional disc to a globular, three dimensional organ. During this process, the a2/a3 joint constricts greatly, causing caps to elongate from tension ().
At 14 hrs APF in wild type animals we observe fewer caps than we do at later stages (e.g. 24 hrs APF, ). This may indicate that not all cells have deposited NompA early in development. Alternatively, scolopidial precursor cell proliferation may be still in progress. Neuronal staining with the monoclonal antibody 22C10 suggested that neuronal numbers increase until approximately 26 hrs APF 
. Therefore increased cap numbers may indicate further cellular proliferation. A third alternative is that there may be asynchronous activation of GFP-NompA expression.
ck/MyoVIIA is necessary for JO organization during development
We compared developing JO organization in control and ck13
/MyoVIIA-null allele 
) homozygous flies. As evident in , at the time of head eversion, mutant pupae present a GFP-NompA pattern similar to controls: the caps are aligned in a circle and have a globular profile. But, unlike in controls (, 14 hrs inset), we saw no caps with an elongated profile in mutant flies. As development continues, caps in ck
/MyoVIIA-null antennae are in disarray (). Additionally, while control caps are tightly juxtaposed with, or extending into the future a2/a3 joint in controls (, 16–20 hrs APF), in mutant pupae, caps retain their globular profile and are distanced from the a2 perimeter (, 18 and 22 hrs APF). This indicates either that scolopidial apical attachment is not formed in mutants, or that it immediately fails to be maintained.
The data in are supported by findings from cultured antennal discs. Similar to what we observed with developing pupae, ck/MyoVIIA is necessary for JO organization from early stages of antennal development in cultured antennal discs. ck/MyoVIIA-null antennae have globular, disarrayed dendritic caps, unlike controls (). We also used electron microscopy (EM) to gain a more detailed perspective on JO development in the presence or absence of ck/MyoVIIA (). At approximately 26 hrs APF, control JOs show dendritic caps clearly extending into the a2/a3 joint (). Wild-type caps are elongated and slender, compared to mutant JO where we have observed no caps extending into the a2/a3 joint by 28 hrs APF. These data mirror our confocal image analysis (), together indicating that ck/MyoVIIA is necessary for early JO organization during development, at least through an involvement at scolopidial apical attachment.
Scolopidia lacking ck/MyoVIIA are not attached to the a2/a3 joint during development.
Finally, we labeled control and mutant, cultured antennal discs with an anti-DE-cadherin antibody to investigate ck/MyoVIIA effects on JO cell-cell junctions during development (). Cellular junctions, as labeled by this antibody, are present in both control and mutant discs at all developmental stages analyzed. Additionally, we noticed that cell junction “tracts” are visible in JO even though scolopidia are detached in the absence of ck/MyoVIIA (). Together with our observations that another cellular junction marker (β-catenin) localization is unaffected in ck/MyoVIIA-null flies (not shown), our data indicate that ck/MyoVIIA does not affect the presence of cell-cell junctions in JO.
Our investigations of scolopidial detachment in adult flies 
and in the developing JO indicate that ck
/MyoVIIA is important for JO arrangement at a level other than gross cellular morphological juxtapositions. Indeed, our previous data strongly implied dendritic cap anomalies in apical detachment observed in adult flies 
. We do not know exactly how the dendritic cap is produced, nor the identity of other components besides NompA 
. Nonetheless, based on the evidence presented here, we can propose a model for the role of ck
/MyoVIIA in JO organization during development: the dendritic cap is deposited around the dendrites and extends between epithelial cells into the future a2/a3 joint upon cuticle secretion by epidermal cells (). As scolopidium distance from the a2/a3 joint increases during development, the dendritic cap, which has already established its apical attachment, elongates under tension. Scolopale rods grow from assembled actin and may also play a role in this distancing process and cap elongation (in fact, we do not clearly observe scolopale rods using actin labeling until approximately 24 hrs APF, although defects in GFP-NompA distribution are already detected at 18 hrs APF ()). In ck
/MyoVIIA mutants, as scolopidial distancing from the a2/a3 joint begins, the cap detaches from an inadequate apical anchor, possibly because of dendritic cap morphological anomalies 
due to faulty transport/deposition of some component, while cell-cell junctions are present among scolopidial components ( and data not shown). This is also supported by our previous investigations of the a2/a3 joint in adult ck
/MyoVIIA-null JO, where we noticed that a2/a3 epithelial cells remain attached to the joint, and that scolopidial cell-cell junction integrity is intact 
. Two non-mutually exclusive mechanisms can account for this model. In the first possibility, ck
/MyoVIIA directly facilitates extension of the dendritic cap through the epithelial layer. Failure of the cap to penetrate through this epithelium prevents the anchoring which is necessary for its sufficient elongation to reach the luminal space for anchoring in the cuticle. In the second possibility, elongation of the cap is purely a passive secondary consequence of tension that requires ck
/MyoVIIA-mediated anchoring. Thus these two mechanisms differ in whether ck
/MyoVIIA directly mediates the cap extension or directly mediates the anchoring. It is possible that both mechanisms are operating, but our current data cannot distinguish them. In conclusion, our explorations indicate that ck
/MyoVIIA functions in JO development in its organization during development, from the perspective of NompA. Initially, the absence of this motor protein leads to faulty apical anchoring, which in turn, leads to scolopidial detachment.
ck/MyoVIIA maintains adult JO organization
Next, we investigated ck
/MyoVIIA involvement in maintaining adult JO organization. To reduce ck
/MyoVIIA levels in adult flies we used an ubiquitously-expressed, temperature-sensitive tubulin-Gal80ts
to inhibit protein expression using the Gal4-UAS
system. Active Gal80ts
(hereon referred to as Gal80) binds Gal4 and prevents it from activating UAS sequences. Flies maintained at the Gal80 restrictive temperature, on the other hand, express Gal4 target genes. We generated ck
/MyoVIIA-null adult flies that were fully rescued by a ubiquitously expressed UAS-GPF-ck
) rescue construct 
, which were also homozygous for Gal80
. These flies were reared at 29°C (Gal80 restrictive temperature; GFP-ck
/MyoVIIA expressed) until eclosion, at which time they were transferred to 18°C (Gal80 permissive temperature) for 0–20 days. Under these conditions further expression of GFP-ck
/MyoVIIA is blocked and the level of GFP-ck
/MyoVIIA decreases (). JO is fully mature and functional at eclosion.
ck/MyoVIIA is important for maintaining adult JO organization.
We had previously shown that there is a correlation between degree of detachment and degree of deafness associated with ck
/MyoVIIA mutations when comparing antennae with no, partial, or complete JO organization 
. To investigate effects of reduced ck
/MyoVIIA levels () on JO transduction, and to gain some insight into JO disorganization through this technique, we performed electrophysiological recordings from the antennal nerve. Recordings from flies maintained at the Gal80 permissive temperature for a period of 0, 8 or 20 days showed no statistically significant difference, indicating that lower ck
/MyoVIIA levels in these adult flies have no effect on JO transduction ().
Nevertheless, we used confocal laser microscopy to investigate the extent of detachment present in JO. It is impossible for us to accurately quantify the number of scolopidia detached in a particular JO. However, we could easily provide a semi-quantitative score of no, some or large/full detachment. We compared antennae from flies maintained for 0 days at the Gal80-permissive temperature to ones which were maintained at the Gal80-permissive temperature for up to 20 days. As shown in , lower ck/MyoVIIA levels for increasing periods of time lead to higher numbers of antennae with detached scolopidia. Since we know that aging does not lead to scolopidial detachment (ESL & DFE, unpublished observations) we conclude that ck/MyoVIIA is important for maintaining adult JO organization.
Comparing our electrophysiology data with the confocal images, it is likely that a large number of scolopidia need to be detached in order for an electrophysiological difference to be detected with our extracellular recording. Therefore, confocal laser microscopy may be more informative on some finer details of JO organization than is extracellular electrophysiology.
Our data strongly suggest that ck
/MyoVIIA participates in JO organization through a role in scolopidial apical attachment. Interestingly, similar to myoVIIA
-deficient mice where stereocilia are initially formed but they become progressively more disorganized 
pupae show overall JO organization akin to controls early in development (; 14 hrs APF). Later, disorganization becomes increasingly evident in the developing JO (), reminiscent of the large disarray in stereocilia observed in developing vertebrate ears 
. This supports the idea that MyoVIIA plays evolutionarily conserved functions in the organization of developing auditory organs, although the exact mechanisms through which it acts in organizing them may be different. In JO, ck
/MyoVIIA is important for proper dendritic cap morphology 
. It is likely that ck
/MyoVIIA-related cap malformations are responsible for apical detachment during development and in adult flies. Recent reports from vertebrate studies presented evidence that MyoVIIA is important for proper localization of a number of membrane-associated proteins involved in trans-stereocilia links, including the ankle-link protein Vlgr1, the transmembrane proteins Usherin and Protocadherin 15, and PDZ domain-containing protein Whirlin 
. These mislocalizations may explain the disorganization phenotype observed in MyoVIIA-deficient mice 
. The identity of proteins for whose transportation ck
/MyoVIIA is required in JO remains unknown. Whether the transmembrane and submembrane proteins putatively transported by vertebrate MyoVIIA have orthologues in Drosophila
and whether they could be implicated in JO organization needs to be investigated. Future work employing proteomic approaches, the versatility of Drosophila
genetics and confocal imaging alongside additional JO cellular markers, will provide further details into the function of this protein necessary for auditory transduction in vertebrates and flies.
ck/MyoVIIA function in JO is modulated by myosin-associated proteins
To understand mechanisms in which ck
/MyoVIIA participates in JO, we focused on a pathway important to Planar Cell Polarity (PCP; the two-dimensional polarity established within an epithelium). Manifestations of PCP are very obvious in the arrangement and orientation of the stereocilia in the hair cells of the vertebrate inner ear sensory epithelium, as well as in bristle organ and even non-sensory hair organization in insects 
. While there is no obvious morphological manifestation of PCP associated with JO, it is conceivable that the molecular mechanisms of PCP may operate and could underlie the defects we see in ck
/MyoVIIA-deficient scolopidia. To address this possibility, we studied a genetic pathway with which ck
/MyoVIIA interacts in the PCP of the fly wing 
. Here, each epidermal cell constructs a single prehair, consisting of bundled actin filaments that mature into a single hair that faithfully projects from the distal vertex of each cell 
. Many studies have examined the mechanisms that govern the molecular and morphological orientation of hairs in these cells.
Frizzled (Fz) a transmembrane receptor of Wnt signals, and Dishevelled (Dsh), an adaptor protein downstream of Fz, are evolutionarily conserved components of PCP 
. Ligand-dependent activation of Fz leads to Dsh activation, which triggers Drok (Drosophila
Rho-kinase) (); in turn, Drok phosphorylates Spaghetti-Squash (Sqh; the nonmuscle myosin II regulatory light chain protein) 
. Regulatory light chain phosphorylation causes a conformational change in non-muscle myosin II (zip
/MyoII) that activates zip
/MyoII motility in in vitro
motility assays 
. Throughout phylogeny, myosin II molecules are heterohexamers that consist of a pair of myosin II heavy chains (in flies encoded by zipper
/MyoII), a pair of regulatory light chains (in flies encoded by spaghetti squash
) and a second pair of so called essential light chains (in flies, encoded by Mlc-c
, see 
and references therein). In all systems studied to date, binding of the light chains to the heavy chain is so tight that dissociation constants have not been measured. In the wing, malfunction of any one of the elements in this pathway disturbs the assembly of F-actin prehairs and leads to the aberrant generation of multiple hairs on each cell (), which are usually interpreted as a defect in PCP or its output.
ck/MyoVIIA function is modulated by Sqh and DMBS in adult JO.
Interestingly, in the wing there is a strong genetic interaction of ck
/MyoVIIA with the fz
pathway, where ck
/MyoVIIA mutations have the opposite effect of zip
/MyoII heavy chain mutations 
(). The molecular basis of this interaction is not known. Analysis of nonmuscle myosin II subunits indicates that in embryos ck
/MyoVIIA does not bind to Sqh 
/MyoII function is also inhibited by myosin phosphatase which dephosphorylates Sqh at the Drok-dependent sites () 
We explored genetic interactions between ck
/MyoVIIA and Drok
/MyoII and the myosin binding subunit of myosin phosphatase (DMBS
) in JO by reducing the copy number of each putative interacting gene in a sensitized ck
), and by using electrophysiology to determine the effects such reductions had on already compromised auditory transduction 
. We first focused on Drok
/MyoII genetic interactions with ck
/MyoVIIA in JO. As evident from histograms in , neither Drok1
(lethal recessive) nor zip1
(lethal recessive) heterozygosity altered transduction ability of ck
/MyoVIIA hypomorphs significantly (P>0.5). However, sqhAX39
(lethal recessive) significantly reduced sound-evoked potentials (; P<0.01), suggesting that Sqh normally promotes ck
/MyoVIIA function in JO scolopidia. Genetic interaction with sqh
but not Drok
/MyoII was unexpected based on the current understanding of the wing PCP pathway (). Therefore, we explored whether ck
/MyoVIIA genetically interacts with DMBS. We generated flies heterozygous for DMBS03802
(lethal recessive), again in a ck13
background. In the presence of only one functional gene copy of DMBS
flies showed a significantly enhanced ability to transduce the pulse song (P<0.001; ) suggesting that DMBS normally suppresses ck
/MyoVIIA activity in JO. This DMBS interaction is predictably opposite to the interaction with sqh
, but the direction of both of these interactions is unexpected if their effects are mediated through the same underlying mechanisms as PCP in the wing or embryonic dorsal closure.
These differences could be attributable to two main possibilities. The first is that scolopidial detachment in ck/MyoVIIA is due to defects in the PCP pathway outlined in the wing, yet this pathway is skewed in JO, either because: a) some players in the pathway are absent, b) new players are present, or c) levels of players are shifted sufficiently in a tissue-dependent manner. The second possibility is that scolopidial detachment is due to functions of ck/MyoVIIA independent of a role in PCP. While we currently cannot distinguish formally between these two possibilities, the overall profile of genetic interactions presented here may favor the PCP-independent mechanisms. Interestingly, the role of MyoVIIA in vertebrate hearing has not been directly associated with a core role in PCP, despite the highly oriented arrangement of hair cells that clearly rely on PCP mechanisms for their precision.
In summary, we have demonstrated that ck/MyoVIIA is acutely required early in the apical attachment of scolopidia, either by directly facilitating dendritic cap extension through the epithelial layer, or anchoring the cap during tensioning and elongation. Furthermore, we have shown a continuing requirement for ck/MyoVIIA to maintain attachments throughout adulthood. Finally, we have presented evidence that in the JO, ck/MyoVIIA genetically interacts with the non-muscle myosin light chain regulatory protein (Sqh) and the myosin binding subunit of myosin phosphatase (DMBS). Our data suggest that the mechanism of ck/MyoVIIA in JO differs from a previously characterized PCP pathway (), at least at the level of the components we examined (Drok, zip/MyoII; ). Instead, novel genetic interactions between ck/MyoVIIA and myosin regulatory proteins appear to operate in JO. Whether these tissue-specific interactions reflect deployment of alternative forms of PCP or whether the JO function of ck/MyoVIIA is unrelated to PCP functions remains to be established. Our data are consistent with the idea that similar genetic pathways may affect auditory transduction in vertebrates, and draw attention to the complexity of myosin regulatory mechanisms in different cell types that still require dissection in Drosophila and vertebrate models of deafness.