1. Cell Biology
  2. Developmental Biology

Dzip1 and Fam92 form a ciliary transition zone complex with cell type specific roles in Drosophila

  1. Jean-André Lapart
  2. Marco Gottardo
  3. Elisabeth Cortier
  4. Jean-Luc Duteyrat
  5. Céline Augière
  6. Alain Mangé
  7. Julie Jerber
  8. Jérôme Solassol
  9. Jay Gopalakrishnan
  10. Joëlle Thomas  Is a corresponding author
  11. Bénédicte Durand  Is a corresponding author
  1. Institut NeuroMyoGène, CNRS UMR 5310, INSERM U1217, Université Claude Bernard Lyon 1, France
  2. Institute of Human Genetics, Universitätsklinikum Düsseldorf, Heinrich-Heine-Universität Düsseldorf, Germany
  3. IRCM, INSERM, Université de Montpellier, ICM, France
https://doi.org/10.7554/eLife.49307
Share this article
Cite this article
  1. Jean-André Lapart
  2. Marco Gottardo
  3. Elisabeth Cortier
  4. Jean-Luc Duteyrat
  5. Céline Augière
  6. Alain Mangé
  7. Julie Jerber
  8. Jérôme Solassol
  9. Jay Gopalakrishnan
  10. Joëlle Thomas
  11. Bénédicte Durand
(2019)
Dzip1 and Fam92 form a ciliary transition zone complex with cell type specific roles in Drosophila
eLife 8:e49307.
https://doi.org/10.7554/eLife.49307
  2. Download BibTeX
  3. Download .RIS

Abstract

Cilia and flagella are conserved eukaryotic organelles essential for cellular signaling and motility. Cilia dysfunctions cause life-threatening ciliopathies, many of which are due to defects in the transition zone (TZ), a complex structure of the ciliary base. Therefore, understanding TZ assembly, which relies on ordered interactions of multiprotein modules, is of critical importance. Here, we show that Drosophila Dzip1 and Fam92 form a functional module which constrains the conserved core TZ protein, Cep290, to the ciliary base. We identify cell type specific roles of this functional module in two different tissues. While it is required for TZ assembly in all Drosophila ciliated cells, it also regulates basal-body growth and docking to the plasma membrane during spermatogenesis. We therefore demonstrate a novel regulatory role for Dzip1 and Fam92 in mediating membrane/basal-body interactions and show that these interactions exhibit cell type specific functions in basal-body maturation and TZ organization.

eLife digest

Many animal cells have hair-like structures called cilia on their surface, which help them to sense and interact with their surroundings. The cilia are supported by protein filaments and must assemble correctly because faulty cilia can lead to several life-threatening diseases. Problems in an area at the base of the cilia, known as the ‘transition zone’, account for the most severe forms of these diseases in humans.

The transition zone is responsible for selecting which proteins are allowed in and out of the cilia. The transition zone itself is made up of many proteins that work together to determine the cilia composition. But not all of these proteins are known, and it is unclear how those that are known affect cilia structure.

One protein found in transition zones of several animals, including fruit flies and mice, is called Cby. Lapart et al. set out to understand which other proteins interact with Cby in fruit flies to better understand what this protein does in the transition zone. A series of experiments showed that Cby interacts with two proteins called Dzip1 and Fam92 to regulate the assembly of transition zones. Together these three proteins constrain a core component of the transition zone, a fourth protein called Cep290, to the base of the cilia.

Fruit flies only have cilia on cells in their sensory organs and testes and, in both types of tissue, cilia could only form properly when Dzip1 and Fam92 were present. Lapart et al. also showed that, in the fruit fly testes, Dzip1 and Fam92 helped to anchor the newly forming cilia to the cell surface. This anchoring role was particularly important for the fruit flies’ sperm to grow their characteristic whip-like tails, which are a specialized type of cilia that allow sperm cells to move.

Overall, the findings show how some transition zone proteins work together and that they can have different effects in different tissues. Understanding the mechanisms behind healthy cilia assembly will likely be key to tackling cilia-related diseases.

Introduction

Cilia and flagella are highly conserved organelles present at the surface of eukaryotic cells. They play major physiological roles in animals such as cell or fluid mobility, signaling during development and cellular homeostasis. The importance of cilia and flagella is highlighted by the discovery of human diseases, classified as ciliopathies, that are associated with defects in cilia structure and/or function (Badano et al., 2006; Baker and Beales, 2009; Brown and Witman, 2014). Cilia and flagella are templated from the basal body, derived from the mother centriole, from which the microtubule based axoneme is assembled.

At the base of the cilium, a specific compartment, the transition zone (TZ) plays a critical role in cilium assembly and function. Many genes responsible for cilia associated diseases such as the Meckel syndrome (MKS), Joubert syndrome or nephronophthisis (NPHP) are caused by defects in proteins of the TZ (Czarnecki and Shah, 2012). The TZ functions as a ciliary gate by sorting selected components in and out of the cilium, thus controlling the specific composition of the ciliary compartment (Garcia and Reiter, 2016; Reiter et al., 2012; Gonçalves and Pelletier, 2017). TZ assembly starts during the first steps of cilia formation, when the mother centriole associates with cytoplasmic vesicles before docking to the plasma membrane. Assembly of TZ proteins is spatiotemporally controlled and requires, in most organisms, at least three different protein modules namely MKS, NPHP and CEP290. Extensive genetic, biochemical studies and super resolution microscopy analysis helped to establish a hierarchy of these components and a structural view of the TZ architecture (Williams et al., 2011; Garcia-Gonzalo et al., 2011; Sang et al., 2011; Chih et al., 2012; Gupta et al., 2015; Garcia and Reiter, 2016; Reiter et al., 2012; Gonçalves and Pelletier, 2017).

Although largely conserved from worms to mammals, all TZ proteins are not conserved in all ciliated species and variations exist between model organisms. For example, the NPHP module, present in mammals, worms and protozoa, is not conserved in flies, whereas CEP290 and several members of the MKS module are conserved in most organisms (Basiri et al., 2014; Pratt et al., 2016; Vieillard et al., 2016). In addition to the core TZ components, several others have been identified but their precise relationships with the core known TZ components are not understood. Among these other proteins, Chibby (Cby), a conserved TZ component in vertebrates and flies, is required for cilia function both in mammals and Drosophila (Burke et al., 2014; Voronina et al., 2009; Enjolras et al., 2012; Vieillard et al., 2016). In vertebrates, CBY has been shown to interact with several basal body (BB) associated proteins, such as ODF2 or CEP164 (Lee et al., 2014; Siller et al., 2017; Burke et al., 2014; Steere et al., 2012; Chang et al., 2013) and more recently DZIP1L, DZIP1 and FAM92a or b proteins (Wang et al., 2018; Li et al., 2016b; Breslow et al., 2017). However, the functional integration of CBY and these interactors in TZ assembly is unclear and some of those, such as Cep164 cannot likely sustain Cby function in Drosophila, as Cep164 does not seem to be expressed in Drosophila testes (Flybase).

We show here that the unique Drosophila orthologs Dzip1 (CG13617) and Fam92 (CG6405) of respectively, vertebrate DZIP1 or DZIP1L and FAM92a or b, interact and cooperate with Cby in flies. We demonstrate that all three proteins form a strictly ordered functional module, and cooperate in building the TZ in the two Drosophila ciliated tissues, with Dzip1 acting upstream of Fam92 and Cby. While our observations establish that Dzip1 and Fam92 localization at the TZ relies on Cep290, they reveal that Dzip1 and Fam92 exert a negative regulatory feedback loop by restraining Cep290 localization to the ciliary base.

Last, our work reveals remarkable differences in the role of Dzip1 and Fam92 in regulating basal bodies (BB) docking between the two Drosophila ciliated tissues. Whereas, loss of Dzip1 or Fam92 does not affect basal body docking in sensory cilia, it impairs BB-membrane growth and attachment in spermatocytes. As a consequence, we observed aberrant and premature elongation of the axoneme before completion of meiosis, highlighting a primary role of the BB-membrane associated compartment for regulating axonemal microtubule elongation in Drosophila spermatocytes. These aberrant elongations mostly affect the daughter centrioles, revealing functional differences of the mother and daughter centrioles in Drosophila spermatocytes.

Results

Dzip1/Fam92/Cby form a complex at the ciliary transition zone in Drosophila

We initially identified mouse CBY1 interactors by LAP-Tag affinity purification using IMCD3 cells (murine Inner Medullary Collecting Duct cells) stably expressing CBY1 fused to the EGFP-TEV-S peptide tag (see Materials and methods section). This EGFP-TEV-S tag in N-terminus allows to detect the protein in cells and its purification by a two-step affinity procedure (Torres et al., 2009). Analysis of LAP-CBY1 complex revealed the presence of DZIP1 and FAM92 among 22 identified proteins (see full list in Supplementary file 1), in agreement with recently published data (Li et al., 2016b; Breslow et al., 2018; Wang et al., 2018). DZIP1/1L and FAM92a/b each show a unique ortholog gene in Drosophila, CG13617 and CG6405 respectively (Figure 1—figure supplement 1A), but are absent, like CBY1, from the C. elegans genome. Hereafter, we name CG13617 and CG6405 as Dzip1 and Fam92, respectively. By co-immunoprecipitating Cby-GFP and HA-Dzip1 or HA-Fam92, we demonstrate that Drosophila Dzip1 and Fam92, each interact with Cby (Figure 1—figure supplement 1B–C). Dzip1 or Fam92 do not apparently interact with each other in these co-IP experiments (Figure 1—figure supplement 1D). However, when all three proteins are expressed together, immunoprecipitation of GFP-Dzip1 pulls down both Cby and Fam92, suggesting that all three proteins are present in one complex when co-expressed in mammalian cells (Figure 1—figure supplement 1D).

To identify the subcellular localization of Drosophila Dzip1 and Fam92, we generated transgenic flies expressing Dzip1-GFP or Fam92-GFP under the control of their respective promoters (Figure 1—figure supplement 2). We determined that both dzip1 and fam92 are exclusively expressed in the two kinds of Drosophila ciliated cell types, namely type I sensory neurons and male germ cells (Figures 1 and 2). Type I sensory neurons comprise chordotonal (Ch) and external sensory (ES) neurons, which harbor motile and immotile cilia respectively (Figure 1A) (Gogendeau and Basto, 2010; Jana et al., 2016). Each sensory neuron is enclosed in several support cells forming the sensory organ or scolopidia. Dzip1 and Fam92 decorate the base of the cilia at the tip of the sensory dendrites (labeled with 22C10) (Figure 1B, arrows). By performing, super-resolution 3D structured-illumination microscopy (3D-SIM), we confirmed that Dzip1 and Fam92 co-localize with Cby in sensory neurons (Figure 1C), demonstrating their restricted localization at the ciliary transition zone.

Figure 1 with 2 supplements see all
Download asset Open asset
Drosophila Dzip1 and Fam92 are expressed in ciliated sensory neurons and locate at the ciliary transition zone.

(A) Scheme of the two types of sensory organs (or scolopidia) present in the antennae of the Drosophila head. Chordotonal scolopidia are proprioceptors present in the second antennal segment of the Drosophila head and respond to sound vibrations and gravity. Each chordotonal scolopidia is composed of several support cells (gray) and comprise two or three ciliated neurons (red). External sensory organs (or scolopidia) are present under each sensory sensilla of the third antennal segment and react to olfactory and chemical stimuli. In other body parts, ES organs also respond to mechanical stimuli. External sensory scolopidia comprise support cells (gray) and only one ciliated neuron (red). In ciliated sensory neurons, the two centrioles (or proximal and distal basal bodies) stand above each other at the tip of the dendrites. (B) Whole-mount staining of the second and third antennal segment showing neuronal cell bodies and dendrites (22C10), Dzip1-GFP and Fam92-GFP. Dzip1-GFP and Fam92-GFP are present at the tip of the dendrites in each type of neurons. Arrows point both to Dzip1-GFP and Fam92-GFP localization at the tip of the dendrites. (C) 3D-SIM imaging of Ch and ES neurons. Both Drosophila Dzip1 and Fam92 overlap with Cby at the transition zone. Bars = 10 μm for (B); = 0.5 μm for (C).

Figure 2
Download asset Open asset
Drosophila Dzip1 and Fam92 are located at the transition zone during spermatogenesis.

(A) Scheme illustrating centriole/basal bodies (BB) behavior during spermatogenesis in male germ cells. In early spermatocytes, two pairs of centrioles (gray) are present in each cell. During spermatocyte maturation, centrioles convert to BB and dock to the plasma membrane while extending a primary like cilium which is entirely decorated with TZ proteins (green and red). In late spermatocytes, the TZ, also described as the primary like cilium, reaches approximatively 500 nm and protrudes at the cell surface. During meiosis, BB with primary like cilia/TZ are internalized, thus retaining the ciliary membrane (yellow) connected to the plasma membrane, hence creating a membrane cap in spermatids. In round spermatid, BB are apposed to the nuclear membrane and the ciliary cap (yellow) connected to the plasma membrane is extended. The ring centriole (green) marks the base of the ciliary cap which is decorated by TZ proteins. At the onset of axoneme elongation, axonemal microtubule nucleation inside the cap extends the ciliary cap/TZ, and as the axoneme grows (black), the ring centriole is pushed away from the basal body. (B–C) Confocal imaging of whole-mount testes showing Dzip1-GFP (B), Fam92-GFP (C), Cby-Tom and Asterless (centrioles). Dzip1 and Fam92 appear together with Cby in early spermatocytes at the tip of centrioles. In elongating spermatids, Drosophila Dzip1 and Fam92 mark the ring centriole (arrows) separating from the BB (asterless, gray). (D) 3D-SIM imaging of male germ cells. Drosophila Dzip1 and Fam92 overlap with Cby in spermatocytes. Plots of the intensity profile of the centrioles along the dotted lines illustrate the overlay between Cby-Tom and Dzip-GFP or Fam92-GFP. Bars = 1 μm.

In the male germline, Dzip1 and Fam92 appear first in spermatocytes (Figure 2) at the distal end of centrioles. In spermatocytes, centrioles have a specific behavior as both mother and daughter centrioles of each pair elongate and dock to the plasma membrane. All four basal bodies (BB) extend a TZ, also described as a primary cilium-like structure (Figure 2A) (Pasmans and Tates, 1971; Riparbelli et al., 2012; Gottardo et al., 2013; Vieillard et al., 2016). Subsequently, during meiosis, all four basal bodies are engulfed inside the cytoplasm, together with the primary like cilium which hence creates a ciliary cap that ensheaths each basal body distal end. After meiosis and at the onset of axoneme elongation, the ciliary cap is remodeled and a distinct domain, the ring centriole, appears at its base (Vieillard et al., 2016; Basiri et al., 2014; Fabian and Brill, 2012). The axoneme grows inside the ciliary cap, which concomitantly migrates, extruding the nascent axoneme out in the cytoplasm (Figure 2A) (Pasmans and Tates, 1971; Riparbelli et al., 2013; Gottardo et al., 2013). We observed that Dzip1, Fam92 and Cby all localize at the centriolar distal ends in early spermatocytes (Figure 2B–C). During centriolar/BB maturation, as the ciliary cap grows, Dzip1 and Fam92 localizations are extended and overlap with Cby, as revealed by 3D-SIM observations (Figure 2D). In spermatids, when the TZ migrates away from the basal body, we observed that Dzip1 and Fam92 strongly accumulate with Cby at the ring centriole (Figure 2B–C, arrows).

Together, these results strongly indicate that Dzip1/Fam92/Cby interact at the ciliary transition zone.

Dzip1 and Fam92 are required for cilia and flagella formation in Drosophila

To determine the roles of the Dzip1/Fam92 module in TZ assembly or function, we generated two deletion alleles of dzip1 or fam92 (Figure 1—figure supplement 2 and Materials and method section). In the dzip11 allele, the 65 N-ter codons (including start codon) were deleted and replaced by a 3 kb insertion including the mini-white gene (Figure 1—figure supplement 2). In fam921, exon two and part of exon three were removed (294 bp, Figure 1—figure supplement 2), leading to the deletion of 81 aa and a frameshift with several stop codons in the remaining downstream sequence, leaving only 47 aa of the WT protein and 39 amino acids of the −1 frameshifted open reading frame.

dzip11 and fam921 flies are viable but show typical behavioral defects associated with ciliary dysfunction in Drosophila. Fly geotaxis response (Figure 3A) was monitored by bang assay (Enjolras et al., 2012). Whereas most of the control flies reach and stay at the top of the tube 30 s after the bang, no dzip11 flies reached the top of the tube (Video 1). This phenotype was worsened in dzip11/Df as the flies not only fail to climb, but are completely immotile with held up wings (Video 2), indicating that dzip11 is likely not a complete null allele. Both dzip11 and dzip11/Df behavioral phenotypes were fully rescued by one copy of dzip1::GFP (not shown). In contrast, a few percentage (17%) of fam921 flies could reach the top of the tube (Figure 3A) and no differences could be observed when comparing fam921 flies and fam921 over its cognate deficiency (fam921/Df), indicating that fam921 is likely a null allele. The fam92::GFP rescue construct partially restores the climbing behavior of the fam921 mutant flies. These observations reveal a graded requirement for Fam92 and Dzip1 in Drosophila geotaxis response and indicate a regulatory role of this complex in sensory cilia assembly or function.

Figure 3
Download asset Open asset
Dzip1 and Fam92 are required for ciliogenesis in Drosophila sensory cilia.

(A) Quantification by bang assay of geotaxis response of fam921 flies compared to control or rescued flies. The percentage of flies that reach a defined level in the tube (0–2, 2–4, 4–6, 6–8) is represented. fam921 flies are unable to climb along the tube compared to control flies. This defect is partially rescued by adding two copies of fam92::GFP transgene (fam921/+ n = 61, fam921/fam921 n = 63, rescue n = 68). (B) EM analysis of cross sections of antennal Chordotonal (Ch) neurons. Whereas two neurons/cilia can be observed in each control scolopidia, cilia are almost completely absent in dzip11 scolopidia (arrowheads). In fam921, reduced number of cilia are observed on several Ch neurons (arrows). (C) Ch neurons cilia ultrastructure of fam921 antennae, showing reduced number of microtubule doublets (asterisks) and/or accumulation of dense material (arrows) underneath the ciliary membrane and deformation of the membrane (dot). Similar defects are also observed on cross sections of the TZ (lower panels). Note that the linkers connecting the axoneme to the membrane are still present (arrowheads). (D) Serial sections of the basal body-transition zone region of Ch neurons in dzip11 compared to control, from the basal body (proximal, lower panels) to the TZ (distal, upper panels). Whereas doublet microtubules are present and symmetrically organized at the basal body level (arrows, lower panels), they fail to elongate along the transition zone which is incompletely assembled. (E) Longitudinal sections of dzip11 showing basal body with aberrant TZ compared to control (arrows). Membrane bulges (arrowhead) along the aberrant dzip11 TZ can also be detected. (F) Quantifications of the cilium defects observed in scolopidia.

Video 1
Download asset
Real time imaging of dzip11adult flies.
Video 2
Download asset
Real time imaging of dzip11/Df adult flies.

To determine the impact of Dzip1 or Fam92 loss on cilia assembly, we performed ultra-structural analysis by transmission electron microscopy (EM). Strikingly, we observed that cilia were essentially absent in dzip11 antennal chordotonal organs, but were still present in fam921 antennae (Figure 3B and F), in agreement with their milder geotaxis defect behavior. In fam921 flies, 25% of scolopidia (total observed n = 54) showed reduced cilia number (one or no cilia). In addition, defects in axonemal structure were occasionally observed, such as lack of microtubule doublets (Figure 3C, asterisks, 6%), excess accumulation of material beneath the ciliary membrane (arrows) and deformation of the membrane (Figure 3C, upper panel, dot, 4%). Most TZ sections showed normal ultrastructure, but a few (6%) showed incomplete radial symmetry and accumulating material as observed for cilia. Linkers connecting the axoneme to the membrane could still be observed (Figure 3C, arrowhead, lower panel). Transition zones were completely disorganized in dzip11 antennae as revealed by serial-cross and longitudinal sections (Figure 3D–E) and no more axoneme-to-membrane linkers could be observed. Basal bodies were normally present at the dendrite distal tip (Figure 3D, lower panel), but we observed a rapid disorganization of the axoneme, with its complete abrogation a few microns above the basal body (Figure 3D–E).

Hence, these observations demonstrate that loss of Dzip1 or Fam92 affects TZ assembly required for sensory cilia formation.

To address the function of Dzip1 and Fam92 in sperm flagellum assembly, we first investigated male fertility. We observed a strong reduction of the fertility of fam921 males compared to controls. This defect is restored by expressing Fam92-GFP (Figure 4A). Because dzip11 mutant flies are severely uncoordinated, their fertility could not be tested. However, dzip11 testes showed a marked dispersion of the nuclei along the cysts and, as a consequence, altered migration of sperm individualization complexes (Figure 4B and Figure 4—figure supplement 1A). Nuclear dispersion is also observed, to a lesser extent, in fam921 testes (Figure 4B and Figure 4—figure supplement 1A). One possible origin of nuclei dispersion is defective axonemal elongation (Vieillard et al., 2016; Soulavie et al., 2014) and EM analysis confirmed that axonemes were affected in fam921 cells (Figure 4D, 125/448 broken axonemes, 25/448 missing central pair, 52/448 missing axoneme). As well, in dzip11 or dzip11/Df mutant testes, marked axonemal defects could be observed. In dzip11 testes, we observed 6/126 broken axonemes, 1/126 missing central pair, 1/126 missing axonemes (Figure 4C). More severe defects were observed in very young dzip11/Df spermatid cysts, with up to 60% of spermatids with missing or broken axonemes (Figure 4—figure supplement 1B–C). However, in older cysts, the number of spermatids/cyst was much reduced compared to control (mean = 44 spermatids/cyst compared to control = 63.4), but almost all remaining elongated spermatids showed normal axonemes (Figure 4—figure supplement 1B). These observations suggest that young spermatids with severe axonemal defects fail to elongate and are not observed in cross sections of older spermatid cysts.

Figure 4 with 1 supplement see all
Download asset Open asset
Dzip1 and Fam92 are required for flagella formation.

(A) Fertility assays of fam921 males showing severe reduction of fertility in fam921 or fam921/Df compared to control or rescued flies (fam921/+ n = 18; fam921/fam921 n = 10; rescue n = 15; fam921/Df n = 15; Df/+ n = 16) ****, p<0.0001; ns, p>0,05. (B) Confocal analysis of whole mount testes of control, dzip11, dzip11/Df, fam921 and rescued genotypes stained for nucleus (Hoechst, cyan) and actin cones (phalloidin, red). Nuclei are dispersed in mutant cysts compared to controls, which could reflect axonemal elongation defects. As a consequence, actin cones required for spermatid individualization are mislocalized. Axoneme elongation and spermatid individualization are normal in rescued flies. Bars = 10 μm. (C) EM analysis of cross sections of axoneme of round spermatid. In dzip11, flagella ultrastructure is altered with missing microtubule doublets (arrows) or central pair (asterisks). (D) In fam921 mutant spermatids, most axonemes show altered ultrastructure, with missing microtubule central pairs (asterisks) or doublets (arrows), broken symmetry with each part of the axoneme being relocated along the mitochondria (M, arrowheads).

Taken together, these results demonstrate that Fam92 and Dzip1 are necessary at the ciliary TZ for cilia formation both in sensory neurons and male germ cells.

Dzip1 and Fam92 form a strictly ordered functional module with Cby

To determine the functional hierarchy between Dzip1, Fam92 and Cby, we analyzed their respective localization in dzip11 or fam921 or previously described cby1 mutants (Figure 5 and Figure 5—figure supplement 1).

Figure 5 with 1 supplement see all
Download asset Open asset
Dzip1, Fam92 and Cby work coordinately at the TZ.

(A) Cby-Tom is present at the tip of the centrioles in control spermatocytes but strongly reduced in dzip11/Df (arrows) and can even be lost at one centriole of the pair (arrowhead), but in a few situations (2.3% see D quantifications below), Cby-Tom domain is extended (asterisk). In fam921 mutant, Cby-Tom is completely lost at the tip of the spermatocyte centrioles (arrowheads). (B) Fam92-GFP is reduced at both centrioles of the pair in dzip11/Df (arrows) compared to control and completely lost at the tip of the centrioles (arrowhead) in cby1 mutants. (C) Dzip1-GFP is strikingly expanded in a few cby1 spermatocyte centrioles (arrow, 4.4% see quantifications in D) and is reduced at the tip of both centrioles in fam921 testes (arrows). Bars = 5 μm. (D) Quantifications of the distribution of Dzip1-GFP, Fam92-GFP and Cby-Tom in the different mutant contexts. ****, p<0.0001; ***, p<0.001; ns, p>0,05.

Cby-Tom is almost completely lost from ciliary TZ in dzip11/Df sensory neurons (Figure 5—figure supplement 1A). In spermatocytes, Cby-Tom signal is strongly reduced at the tip of centrioles (Figure 5A,D) but, in rare occasions, expanded at one centriole of the pair (asterisk, 2%). The amount of Fam92-GFP is severely reduced in dzip11/Df sensory cilia base (Figure 5—figure supplement 1B) and in spermatocytes (Figure 5B,D, arrows). Altogether, these observations indicate that Dzip1 is required at the basal body/TZ to recruit or stabilize both Cby and Fam92.

In fam921 mutant flies, Dzip1-GFP is slightly reduced at the distal tip of basal bodies in sensory cilia (Figure 5—figure supplement 1C). In fam921 spermatocytes, Dzip1-GFP domain is sometimes expanded (8.3% of cases, Figure 5D), but the overall Dzip1-GFP expression is also slightly reduced (Figure 5C, arrows, 5D). Cby-Tom is completely lost in both fam921 ciliated tissues (Figure 5A, arrowheads and Figure 5—figure supplement 1A). Hence, Fam92 is required to recruit Cby, but not Dzip1, at the TZ. As well, in cby1 tissues, Fam92-GFP completely disappears from basal bodies (Figure 5B, arrowhead and Figure 5—figure supplement 1B), indicating that Cby and Fam92 stabilize each other at the basal body. In cby1 mutant spermatocytes, an extended domain of Dzip1-GFP is occasionally observed (Figure 5C, arrows, 4.5%, 5D). However, no overall difference in Dzip1-GFP intensity was observed in cby1 mutant testes (Figure 5C–D) or antennae (not shown) compared to controls.

These observations establish a functional hierarchy for Dzip1, Fam92 and Cby: Dzip1 is required to recruit or stabilize both Fam92 and Cby at centrioles, but does not depend on Fam92 or Cby for its targeting to centrioles, the latter only restricting Dzip1 to the proximal TZ. Conversely, Fam92 and Cby mutually depend on each other to localize at the TZ.

Dzip1 and Fam92 restrict Cep290 to the proximal TZ

To understand how Dzip1 and Fam92 organize the ciliary base, we investigated their functional relationships with other core TZ components. We first investigated the behavior of Mks1, a component of the core conserved MKS complex (Weatherbee et al., 2009). In Drosophila, Mks1 is required to assemble the MKS complex, but removal of MKS components leads to only very mild defects of cilia assembly (Vieillard et al., 2016; Pratt et al., 2016). In dzip11/Df and fam921 mutant flies, Mks1-GFP is lost or reduced at the TZ in both ciliated tissues (Figure 6A, arrows, and Figure 6—figure supplement 1A). Dzip1 and Fam92 are hence important for the recruitment or stabilization of the MKS module at the TZ.

Figure 6 with 2 supplements see all
Download asset Open asset
Dzip1 and Fam92 organize the transition zone downstream of Cep290.

(A) In spermatocytes, Mks1-GFP is severely reduced at the ciliary cap or lost in dzip11/Df and fam921 (arrows, quantifications on right graph). Cep290-GFP domain is expanded in dzip11/Df (arrows), but less affected in fam921 (quantifications on right graph). (B) Cep290-GFP is expanded in chordotonal cilia of dzip11/Df or fam921 antennae (arrows). (C) Fam92-GFP and Dzip1-GFP are strongly reduced (arrows) or lost (arrowheads) at the tip of the centrioles in cep2900153-G4 spermatocytes (quantifications are illustrated on the graphs). ****, p<0.0001; *, p<0.01; ns, p>0.05. Bars = 5 μm.

Cep290 is a core conserved TZ component which plays a prominent role in TZ assembly in many organisms, including Drosophila (Craige et al., 2010; Rachel et al., 2015; Li et al., 2016a; Basiri et al., 2014). In flies, Cep290 is located at the base of the TZ in sensory neurons and is a critical component of the migrating ring centriole during spermatogenesis (Basiri et al., 2014). In the absence of Dzip1, besides a small but significant reduction of Cep290-GFP on spermatocyte centrioles, we observed striking Cep290-GFP expanded domains, both in spermatocytes and in Ch neurons (Figure 6A–B, arrows). Around 13% of centrioles in spermatocytes and a majority in antennae showed an expanded Cep290-GFP domain (Figure 6A–B). In fam921 mutants, we also observed a few Cep290-GFP expanded domains in spermatocytes and antennae, but with no significant difference in overall Cep290-GFP intensity (Figure 6A–B, arrows).

To further understand the relationships between Cep290, Dzip1 and Fam92, we took advantage of a strong cep290 hypomorphic mutant, cep2900153-G4. This mutant shows severe uncoordination and completely disorganized spermatid cysts with dispersed nuclei (Figure 6—figure supplement 2A), which is a consequence of the severe axonemal elongation defects observed in this mutant (not shown). The phenotypes were completely rescued by two copies of cep290::GFP (Figure 6—figure supplement 2A–B). We observed that in cep2900153-G4 flies, Dzip1 and Fam92 are strongly reduced or lost at the TZ, both in spermatocytes (Figure 6C) and in sensory neurons (Figure 6—figure supplement 1B).

These observations demonstrate that Cep290 is required during TZ assembly to recruit Dzip1 and Fam92, which in turn restrict Cep290 scaffolding to the proximal part of the TZ.

Dzip1 and Fam92 facilitate basal body docking to the plasma membrane in spermatocytes

Defects in BB anchoring to the plasma membrane in spermatocytes lead to aberrant growth of axonemal microtubules (Vieillard et al., 2016). We used the specific axonemal marker CG6652-GFP that only labels axonemal microtubules in flies (Figure 6—figure supplement 2C–D). With this marker, we observed aberrant growth of axonemal microtubules in spermatocytes, with graded severities increasing from fam921, dzip11 to cep2900153-G4 mutant flies (Figure 7A–B). 30% of centrioles showed CG6652 extensions in fam921 testes, 34% in dzip11/Df and 76% in cep2900153-G4. More strikingly, whereas 68% of the centriole pairs present extensions from both centrioles in cep2900153-G4 spermatocytes, we observed that in fam921 testes, 96% of the centriole pairs present microtubule extensions on only one centriole. This asymmetric penetrance of the phenotype is related to centriole age. Indeed, among 24 centriole pairs for which mother and daughter identities could be unequivocally assigned, 19 daughter centrioles (79%) and only five mother centrioles (21%) were affected. As well, in dzip11/Df testes, among 40 centriole pairs, we found 30 daughter centrioles (75%) and 10 mother centrioles (25%) affected by the absence of Dzip1 (Figure 7A–B).

Figure 7
Download asset Open asset
Dzip1 and Fam92 are required for both centriolar docking and proper axonemal formation during Drosophila spermatogenesis.

(A) Confocal imaging of whole-mount testes showing aberrant extensions of axonemes labeled by CG6652-GFP in fam921, dzip11 and cep2900153-G4 spermatocytes. Asterisks point to the centriole in each pair that can be unambiguously assigned as the daughter. Bar = 5 μm (B) Quantifications of the penetrance of CG6652-GFP labeled aberrant axonemes. Schemes representing the different classes that were quantified in the table: (i) two centrioles of the pair are affected; (ii) only one centriole of the pair is affected, either the daughter* or the mother**. (C) EM analysis of cross and longitudinal sections of centrioles in spermatocytes. In dzip11, centriole docking is impaired with partial docking of the centriole (arrowhead) or docking to the plasma membrane of only one centriole of the pair (arrow). Centrioles present an altered/irregular cap, compared with control (asterisk). In fam921, undocked centrioles are also observed in spermatocytes and show irregular distal end (asterisk) or microtubules extending from the tip (arrow). (D) Scheme of the consequences of dzip1 and fam92 loss of function on basal body and cilium assembly in Drosophila ciliated tissues. Whereas basal body anchoring and TZ assembly is affected in male germ cells, only TZ and cilium assembly is affected in sensory cilia. In addition, Cep290-GFP is systematically expanded in dzip1 mutant and occasionally expanded in fam92 mutant chordotonal neurons. The scheme on the right summarizes the functional interactions observed between Dzip1, Cby, Fam92 and Cep290. Red arrows are inhibitory interactions and green arrows define positive interactions.

We analyzed the ultrastructure of the centrioles and primary like cilium in the initial stages of elongation (Figure 7C). We observed irregular distal end (asterisk) of centrioles in dzip11 testes at young spermatocytes stages before docking (upper panel). Whereas centrioles dock in polar spermatocytes before reaching their full size in control testes, undocked centrioles were observed in dzip11. Either both centrioles of the pair were undocked or partially docked (n = 2, lower panel) (Figure 7C, arrowhead) or only one centriole of the pair, the mother, was docked (n = 3) (Figure 7C, arrow). These results show that Dzip1 is required to cap the centriole distal end and foster its anchoring to the plasma membrane. As well, we observed undocked centrioles in fam921 mutant spermatocytes, with irregular distal end (asterisk) or with microtubules extending from the distal end (Figure 7C, arrow), illustrating the role of Fam92 in controlling centriolar distal growth and docking to the plasma membrane.

Altogether, these observations demonstrate that Dzip1 and Fam92 are required for the proper distal elongation of basal bodies and their membrane anchoring in Drosophila spermatocytes. This docking is required to regulate TZ elongation (Figure 7D). In addition, our results show that in male germ cells, although all centrioles have the capacity to generate a ciliary cap, a functional asymmetry of the mother and the daughter centrioles is revealed by their ability to dock to the plasma membrane in absence of key TZ proteins.

Discussion

Our work establishes the functional hierarchy of Dzip1, Fam92 and Cby, which thus define an intrinsic transition zone module for the initiation of ciliogenesis in Drosophila. Dzip1 is required to recruit Fam92 and Cby at the basal body distal end, allowing proper TZ assembly. Dzip1 and Fam92 act downstream of the core TZ protein Cep290 but also regulate its accumulation at the distal basal body. Our work sheds light on tissue specific variations in the initiation of ciliogenesis in the two Drosophila ciliated tissues. In testes, basal body docking strictly depends on the integrity of Dzip1 and Fam92. However, in sensory neurons alterations of the latter complex affect only TZ assembly and cilia formation but not BB docking. Last, our work demonstrates that Dzip1 and Fam92 control the distal elongation of basal bodies and their docking to the plasma membrane, which thus regulate the onset and proper elongation of axonemal microtubules in Drosophila male germ cells.

Conservation of TZ assembly pathways in eukaryotes

In vertebrates, two orthologs have been described for Dzip1 and Fam92 and three for Cby. DZIP1, DZIP1L, FAM92a and b also interact with CBY1 in vertebrates (Wang et al., 2018; Breslow et al., 2017; Ye et al., 2014) indicating a conservation of the interacting capacities of the family members during evolution. The precise hierarchy between members of the complex has not been established in mammals, but depletion of CBY1 prevents the recruitment of FAM92a/b to the centriole (Li et al., 2016b) and DZIP1L was shown to act upstream of CBY1 (Wang et al., 2018; Keller et al., 2009).

In vertebrates, mutations in members of this module lead to ciliogenesis defects of various severities and with different phenotypic outcomes (Wolff et al., 2004; Tay et al., 2010; Glazer et al., 2010; Li et al., 2016b; Breslow et al., 2017) Dzip1 or Dzip1L KO mice die during embryogenesis (Wang et al., 2018; Wang et al., 2013), whereas Fam92a-/- mice show skeletal defects (Schrauwen et al., 2019) and Cby1 KO mice exhibit differentiation defects of motile ciliated epithelia (Burke et al., 2014; Voronina et al., 2009). This phenotypic variability is not expected if all three proteins only act together in a functional module at the TZ, as demonstrated by our work in Drosophila. This suggests that the different mouse paralogs of Cby and Fam92 may have acquired specialized ciliogenic functions in mouse. However, we also observed that dzip11 and fam921 mutant phenotypes show small differences indicating specific functions of each proteins. For instance, in both ciliated tissues, dzip11 hypomorphic mutant phenotype is more severe than fam921, suggesting that Dzip1 has additional functions that are not solely mediated by Fam92.

Despite the conserved role of Dzip1, Cby and Fam92 in TZ and cilia assembly from Drosophila to humans, no homologs could be detected in the genomes of C. elegans. This could be linked to the diversification of cilia function, with both motile and immotile cilia being present from Drosophila to humans, in contrast to C. elegans, where only one subtype of cilia that are immotile is found. Another possible explanation could be that DZIP1/FAM92/CBY are associated with specific signaling or developmental functions in animals that still need to be understood.

Tissue specific functional properties of the TZ proteins in BB docking

This work emphasizes the essential role of Dzip1 and Fam92 in building the ciliary transition zone in the two types of ciliated tissues of Drosophila. Strikingly, it also reveals tissue specific function of these proteins in priming basal body/membrane docking in Drosophila testes. This reveals intrinsic differences in the mechanisms that link basal body to membranes in Drosophila ciliated tissues. In mammals, basal body docking requires transition fibers built from the distal appendages on the mother centriole prior to docking (Wei et al., 2015). In Drosophila, distal appendages have not been observed on centrioles, but structures similar to transition fibers are described at the base of sensory cilia (Ma and Jarman, 2011; Vieillard et al., 2016), whereas only scarce links could be observed in male germ cells (Gottardo et al., 2018; Roque et al., 2018). These differences could explain why destabilization of the TZ leads to basal body anchoring defects in spermatocytes, but not in sensory neurons. This structural characteristic of the spermatocyte TZ is likely to be related to its specific functional properties. Indeed, whereas the TZ is stably built at the ciliary base in sensory neurons, it shows a dynamic behavior during sperm flagella extension, separating from the basal body and migrating along the growing end of the axoneme (Basiri et al., 2014; Avidor-Reiss et al., 2017; Baker et al., 2004; Wei et al., 2008). In addition, in spermatocytes, basal bodies have a dynamic behavior, being first docked at the plasma membrane and next internalized during meiosis (Pasmans and Tates, 1971; Fabian and Brill, 2012). This could induce mechanical constraints on basal bodies that would increase their sensitivity to TZ disruption. In agreement with this hypothesis, when TZ maturation was challenged by modulating membrane phospholipids (Gupta et al., 2018), BB were released from the plasma membrane during meiosis, but their initial docking was not impaired. However, we observed by EM, that BB fail to initially dock in significant occurrences (8 among 13) in dzip11 and fam921 mutant spermatocytes, indicating that Dzip1 and Fam92 are at least involved in the initial steps of BB docking. Previous observations of another strong hypomorphic cep290mecH allele showed docked basal bodies in spermatocytes and spermatids (Basiri et al., 2014). In cep2900153-G4 mutant, we did not quantify the number of docked versus undocked basal bodies in spermatocytes, but we observed up to 76% of aberrant axonemal growth, suggesting that basal body to membrane attachment is compromised in this mutant.

Differences in the organization of the ciliary base associated with variations in the distribution and function of several centriolar and TZ proteins have been documented in Drosophila ciliated cells (Jana et al., 2018). However, none of these identified differences help to explain the behavioral properties of BB docking and TZ dynamics that we have identified in the two ciliated Drosophila tissues. Hence, additional screens for specific factors of basal-body docking or TZ assembly either in sensory neurons or male germ cells need to be designed.

Mother and daughter centriole asymmetry in ciliogenesis

In all our observations, there is a striking phenotypic difference between the mother and daughter centrioles in spermatocytes. In all mutants examined (i.e. dzip1, fam92 and cby) we observed a more penetrant defect on the daughter centriole than the mother. Thus, although the 2 centrioles of each pair are able to form cilia, the daughter centriole appears more sensitive to transition zone perturbations. We do not have molecular explanations for these intrinsic differences of mother and daughter centrioles in spermatocytes. In Drosophila sensory neurons, centrobin plays a critical role in maintaining the daughter centriole fate, precluding its capacity to build a cilium (Gottardo et al., 2015). In sperm cells, centrobin is required for the formation of the C tubule (Reina et al., 2018), which also plays a critical role in TZ assembly (Gottardo et al., 2018). However, in spermatocytes centrobin is equally distributed at the base of both the mother and daughter centrioles and does not show a functional asymmetry. Recently, a transient microtubule based structure that anchors the base of the mother centriole on one of the centriole pair at the onset of meiosis was identified, but its function is unknown (Riparbelli et al., 2018). This structure could stabilize the mother centriole and favor its attachment to the membrane, but the function of this microtubule rootlet in Drosophila spermatocytes needs further investigations. The intrinsic difference of the mother versus daughter centrioles could also be related to the timing of centriole docking during spermatogenesis (Gottardo et al., 2018), where the mother centriole was shown to dock first. This timing difference would be sufficient to better stabilize the TZ of the mother centriole, and hence explain the phenotypic differences observed in our study. There are other situations in the animal kingdom were both mother and daughter centrioles build a cilium. Among the best studied are the bi-flagellated Chlamydomonas and the peculiar case of multiple ciliated epithelia, where numerous de novo centrioles are assembled just at the onset of ciliogenesis. Interestingly, in mammals, CBY1 was shown to play a critical role in multiple ciliated cells to allow proper docking of the multiple basal bodies to the plasma membrane (Burke et al., 2014). It is tempting to speculate that the increased susceptibility of multiple ciliated cells to CBY1 loss, is related to a particular status of these de novo centrioles, as we observed for daughter centrioles in male germ cells. More work will be needed to understand the molecular determinants of the mother versus daughter centriolar functional asymmetry in Drosophila male germ cells.

Specific regulatory role of the ciliary cap in Drosophila male germ cells

We observed that defects of TZ assembly and/or basal body docking lead to aberrant elongation of axonemal microtubules as revealed by the specific Drosophila axonemal marker CG6652-GFP. These abnormal elongation defects appear only in late G2 phase, just at the onset of meiosis, indicating that specific signals, yet to be identified, enable centrioles to start axonemal elongation at the onset of meiosis. The membrane cap restricts this capacity by inhibiting axonemal growth until the round spermatid stage, where a second signal turns on axonemal elongation and TZ migration. Among the candidate proteins recruited by the membrane cap which may coordinate axonemal assembly are microtubule depolymerizing kinesins as previously proposed (Vieillard et al., 2016). This membrane cap could also be involved in stabilizing centriolar capping proteins such as CP110 or CEP97, which in mammals need to be removed from centrioles to allow ciliary assembly (Spektor et al., 2007). However, there are no clear evidence in Drosophila that these proteins need to or are specifically removed before axonemal elongation (Galletta et al., 2016; Franz et al., 2013; Delgehyr et al., 2012). Our observations indicate that regulation of axonemal assembly and cell cycle regulation are tightly linked in these dividing cells, but the effectors of this control still need to be identified.

In conclusion, our work demonstrates the critical role of the conserved Dzip1/Fam92/Cby module downstream of Cep290 in initiating the assembly of the ciliary transition zone in flies. It also reveals key tissue specific differences in basal body docking pathways in Drosophila.

Materials and methods

Key resources table
Reagent type
(species)
or resource		DesignationSource or referenceIdentifiersAdditional
information
Gene (Drosophila melanogaster)cbyEnjolras et al., 2012FLYB: FBgn0067317
Gene (Drosophila melanogaster)fam92/CG6405This studyFLYB: FBgn0032428
Gene (Drosophila melanogaster)dzip1/CG13617This studyFLYB: FBgn0039201
Strain, strain background (Escherichia coli)DH5alphaThermo Fisher Scientific18265017
Genetic reagent (D. melanogaster)Cep2900153-G4
mutant strain		Bloomington Drosophila Stock CenterBDSC: 62671;
FLYB: FBst0062671
RRID:BDSC_62671		FlyBase symbol: w1118; PBac{IT.GAL4}cep2900153-G4
Genetic reagent (D. melanogaster)Df(3R)Exel8178Bloomington Drosophila
Stock Center		BDSC : 7993
FBab0038335
RRID:BDSC_7993
Genetic reagent (D. melanogaster)Df(2L)Exel6033Bloomington Drosophila
Stock Center
BDSC : 7516
FLYB : FBab0037871 RRID:BDSC_7516
Genetic reagent (D. melanogaster)sas4s2214
mutant strain		Basto et al., 2006BDSC : 12119
FLYB : FBal0196943
RRID:BDSC_12119
Genetic reagent (D. melanogaster)cby1 mutant strainEnjolras et al., 2012FLYB : FBal0270281
Genetic reagent (D. melanogaster)dzip11 mutant strainThis studySection Materials and methods “Generation of dzip11 and fam921 alleles”
Genetic reagent (D. melanogaster)fam921 mutant strainThis studySection Materials and methods “Generation of dzip11 and fam921 alleles”
Genetic reagent (D. melanogaster)Unc::mkateVieillard et al., 2016FLYB : FBal0324713
Genetic reagent (D. melanogaster)Cby::TomVieillard et al., 2016FLYB : FBal0270280
Genetic reagent (D. melanogaster)Cep290::GFPBasiri et al., 2014FLYB : FBal0301636
Genetic reagent (D. melanogaster)CG6652::GFPVieillard et al., 2016FlyB: FBal0324714
Genetic reagent (D. melanogaster)mks1::GFPVieillard et al., 2016FLYB : FBal0324710
Genetic reagent (D. melanogaster)fam92::GFPThis studySection Materials and methods “Plasmids and Drosophila reporter gene constructs”
Genetic reagent (D. melanogaster)dzip1::GFPThis studySection Materials and methods “Plasmids and Drosophila reporter gene constructs”
Cell line (Mus musculus)IMCD3ATCCCRL-2123
Cell line (Homo-sapiens)HEK293ATCCCRL-1573
Cell line (Cercopithecus aethiops)COS-7ATCCCRL-1651
AntibodyMouse monoclonal anti-Futsch/22C10DHSBAB_528403IF (1:250)
AntibodyRabbit polyclonal anti-GFPAbcamAB6556IF(1:1000)
WB (1:10000)
AntibodyGuinea pig polyclonal anti-AsterlessKlebba et al., 2013IF (1:50000)
AntibodyRat polyclonal anti-AsterlessMcLamarrah et al., 2018IF (1:50000)
AntibodyRat monoclonal anti-HA (clone 3F10)Roche11867423001WB (1:5000)
AntibodyGoat polyclonal anti-rabbit Alexa 488InvitrogenA11008IF (1:1000)
AntibodyGoat polyclonal anti-rabbit Alexa 647InvitrogenA21244IF (1:1000)
AntibodyGoat polyclonal anti-guinea pig Alexa 488InvitrogenA11073IF (1:1000)
AntibodyGoat polyclonal anti-guinea pig Alexa 594InvitrogenA11076IF (1:1000)
AntibodyDonkey polyclonal anti-guinea pig Alexa 647Invitrogen706-605-148IF (1:1000)
AntibodyGoat polyclonal anti-rat Alexa 488InvitrogenA11006IF (1:1000)
AntibodyGoat polyclonal anti-rat Alexa 555InvitrogenA21434IF (1:1000)
AntibodyGoat polyclonal anti rat Alexa 647InvitrogenA21247IF (1:1000)
AntibodyMouse monoclonal anti-V5InvitrogenR960-25WB (1:5000)
AntibodyGoat polyclonal anti-rabbit-HRPBiorad170–6515WB (1:10000)
AntibodyGoat polyclonal anti-mouse-HRPBiorad170–6516WB (1:3000)
AntibodyGoat polyclonal anti-rat-HRPSigmaA5795-1MLWB (1:20000)
AntibodyMouse monoclonal anti-γ-TubulinSigmaGTU88IF (1:500)
Chemical compound, drugphalloidin FluoProbes 547InterchimFP-AZ0330IF (1:200)
Chemical compound, drugphalloidin FluoProbes 505InterchimFP-AZ0130IF (1:200)
Commercial assay or kitGFP-TRAPChromotekgta-100
Commercial assay or kitMouse monoclonal Anti-HA-agarose antibodySigmaA2095
Commercial assay or kitS-protein agaroseMerck69704
Commercial assay or kitGibson Assembly Master MixNew England Biolabs, IncE5510S
Recombinant DNA reagentpBFv-U6.2Kondo and Ueda, 2013
Nig-Fly		FLYB: FBmc0003127
Recombinant DNA reagentpBFv-U6.2BKondo and Ueda, 2013
Nig-Fly		FLYB: FBmc0003128
Recombinant DNA reagentpG-LAP3 vectorTorres et al., 2009Addgene#79704
Recombinant DNA reagentpEGFP-N1ClontechCat #6085–1

All primer sequences are described in Supplementary file 2.

Plasmids and Drosophila reporter gene constructs

Request a detailed protocol

CBY1 coding sequence was PCR amplified from mouse ependymal primary cell cDNA with primers F-CBY1 and R-CBY1. PCR product was cloned into pDONR223 (Invitrogen) and then subsequently Gateway cloned into the pG-LAP3 vector (Torres et al., 2009; gift from P. Jackson, Addgene#79704). This vector contains the double EGFP-TEV-S peptide tag in N- terminus allowing a two step affinity purification.

Coding sequences of Drosophila cby, fam92/CG6405 and dzip1/CG13617 were obtained by PCR on cDNA from testis. Cby cDNA was cloned in pEGFP-N1 (Clontech) in frame with GFP (primers F- CbyGFP/R-CbyGFP) and in pCDNA3.1-HA (gift from S. Khochbin, Institut Albert Bonniot, Grenoble ; primers F-HACby/R-HACby). cDNA of fam92/CG6405 and dzip1/CG13617 were cloned in pCDNA3.1-HA with primers F-Fam92HA/R-Fam92HA and F-Dzip1HA/R-Dzip1HA, respectively. pCDNA3.1-GFP, GFP-Fam92 and GFP-Dzip1 were obtained by replacement of the HA tag with EGFP. V5-Fam92 was obtained by replacement of the HA tag with V5 tag.

Drosophila reporter gene construct of dzip1/CG13617 was obtained by cloning 1.6 kb upstream regulatory sequences and the entire coding sequence (primers F-Dzip1GFPrep/R-Dzip1GFPrep) in frame with GFP of pJT61, a pattB plasmid with an extra EGFP-6xMycTag-SV40polyA cassette (Vieillard et al., 2016). The 4 kb fragment of fam92/CG6405 including 1.4 kb of upstream regulatory sequences, the entire coding sequence fused to the multipurpose tag cassette including GFP and the 3'UTR was obtained by PCR with primer F-Fam92GFPrep/R-Fam92GFPrep on fosmid from the FlyFos library (Sarov et al., 2016). The resulting constructs were integrated in the 53B2 VK00018 landing site on the second chromosome by PhiC31 integrase (BestGene). All transgenic lines were obtained from BestGene Inc.

Generation of dzip11 and fam921 alleles

Request a detailed protocol

The dzip11 allele (CG13617) was generated by CRISPR/Cas9 induced homologous directed repair (Gratz et al., 2015). Two gRNA were selected using the http://crispor.tefor.net/crispor.py website: 5’- CCCGTTTCACGGACCATCTG CGG −3’ and 5’-GTTTCCAGCACTGTGCCCAG TGG −3’ (protospacer adjacent motifs are underlined). Oligos were phosphorylated by T4PNK (New England Biolabs, Inc) and annealed. Double-stranded 5’ gRNA and 3’ gRNA were cloned in the BbsI site of pBFv-U6.2 and pBFv-U6.2B vectors, respectively (Kondo and Ueda, 2013). 5’ gRNA was further subcloned in the EcoRI–NotI sites of pBFv-U6.2B to express the two gRNAs from one vector.

The 5’ and 3’ homology arms (1.7 kb and 2.3 kb respectively) were amplified by PCR (primers F-5’armDzip1/R-5’armDzip1 and F-3’armDzip1/R-3’armDzip1) and cloned, respectively, into the pJT38 plasmid (pRK2 plasmid [Huang et al., 2008], with an attB cassette) using Gibson Assembly Master Mix (New England Biolabs, Inc). The two vectors (gRNAs and homology arms) were injected into vasa::Cas9 embryos. Flies were crossed to w; TM2, e Ubx/TM6B, e Hu Tb virgin females and the offspring were screened for red-eyed-flies. Homologous recombination was checked by PCR.

The fam921 allele (CG6405) was generated by CRISPR/Cas9 induced deletion. Two gRNA were selected as before: 5’- CATAAGACCTTGCAGATATC GGG −3’ and 5’- GGCTGTCATAGCGCGGGATA AGG −3’ (protospacer adjacent motifs are underlined). Double-stranded phosphorylated 5’ gRNA and 3’ gRNA were cloned into the pBFv-U6.2B plasmid. The vector was integrated into the 89E11 VK00027 landing site on the third chromosome by PhiC31 integrase (BestGene). Surviving G0 males were individually crossed to y2 v1 virgins. A single male transformant from each cross was mated to y2 cho2 v1; Pr Dr/TM6C, Sb Tb virgins.

Males carrying a U6-double gRNA transgene were crossed to nos-Cas9 females to obtain male founder animals. Each male founder was crossed to three virgin females y2 cho2 v1 ; Sco/CyO. Deletion in the fam92 locus was characterized by PCR with primers F-Fam92KO/R-Fam92KO on genomic DNA from the resultant offspring. Flies showing a 294 bp deletion in the fam92 locus were selected for further studies.

Cell culture and transfection

Request a detailed protocol

All reagents for cell culture were purchased from Thermo Fisher Scientific. IMCD3 cells (murine Inner Medullary Collecting Duct cells, ATCC CRL-2123 a gift from A. Benmerah, Institut Imagine, Paris) were cultured in DMEM/HAM'S F12 medium supplemented with 10% FBS, 1X non-essential amino acids, 100 U/ml penicillin-streptomycin and 100 µg/ml hygromycin. COS-7 cells (ATCC CRL-1651) or HEK293 (ATCC CRL-1573) were maintained in DMEM medium containing 10% FBS, 1X non-essential amino acids and 100 U/ml penicillin-streptomycin. Cells were tested negative for mycoplasma.

Transfections were performed using jetPRIME (Polyplus transfection) according to the manufacturer’s instructions.

Stable cell lines

Request a detailed protocol

IMCD3 cell lines expressing LAP-CBY1 or control LAP-GFP were created using the Flp-In system kit (Invitrogen) and established method (Torres et al., 2009).

Tandem affinity purification and mass spectrometry of CBY1-interacting proteins

Request a detailed protocol

Two rounds of purification of CBY1-protein complexes were performed. Cells were seeded in 35 15 cm dishes and grown to confluence before serum starvation for 24 hr to induce primary cilia formation. Cells were washed once with ice-cold PBS1X and purification was performed as described previously (Nachury, 2008). Briefly, cells were collected by scraping and resuspended in 11 ml lysis buffer containing HEPES pH7.5 50 mM, EGTA pH8 1 mM, MgCl2 1 mM, KCl 300 mM, Glycerol 10%, NP-40 0.31% and 1X protease inhibitor cocktail (Roche). CBY1-complexes were purified using GFP-TRAP (Chromotek). S-Tag-CBY1 was cleaved off GFP beads by TEV cleavage and the eluate was further purified on S-protein agarose (Merck) and eluted in 100 μl 2X Laemmli Sample Buffer.

For protein identification, samples were separated by 10% SDS-PAGE. Gels were stained with Coomassie and cut into four slices. Each gel slice was washed, reduced with 10 mM DTT, alkylated with 55 mM iodoacetamide, and subjected to in-gel trypsin digestion overnight (Trypsin Protease, MS Grade Pierce, Thermo Fisher Scientific). The extracted tryptic peptides were cleaned up using OMIX C18 100 µl pipette tips (Agilent), and lyophilized before being reconstituted for the LC-MS/MS analysis.

The peptides were separated using an Eksigent Ultra nano-LC HPLC system coupled with an AB Sciex Triple TOF 5600 mass spectrometer. The LC separations were performed using a Discovery Bio Wide Pore HPLC column (C18, 3 µm, 100 × 5 mm). The mobile phases used were 0.1% formic acid in water (A) and 100% acetonitrile with 0.1% formic acid. The gradient elution steps were performed with a flow rate of 5 µl/min as follows: 0–40% B for 106 min, 40–80% for 5 min, and then 80% B for an additional 5 min. All data were acquired using Analyst software (AB Sciex) in the data dependent mode. Peptide profiling was performed using a mass range of 350–1600 Da, followed by a MS/MS product ion scan from 100 to 1500 Da. For each survey MS1 scan (accumulation time of 250 msec), MS/MS spectra (accumulation time of 75 msec per MS/MS) were obtained for the 30 most abundant precursor ions. The protein identification was performed with the ProteinPilot Software 5.0 (AB Sciex). The MS/MS spectra obtained were searched against the mouse UniProt Proteome database (release 2015_09 with 46470 proteins). The search parameters for tryptic cleavage and accuracy are built-in functions of the software. The other data analysis parameters were as follows: sample type: identification; Cys-alkylation: Iodoacetamide; Digestion: Trypsin; Instrument: TripleTOF 5600; Special factor: gel based ID, and biological modifications; Species: Homo sapiens; Search effort: Thorough ID. Proteins comprising one or more peptides with a high confidence score (95%) and a low false discovery rate (FDR) (estimated local FDR of 1%) were considered positively identified.

Co-IP and western blotting

Request a detailed protocol

Co-IP assays were performed using transfected COS-7 cells or HEK293 cells, harvested in ice-cold PBS1X. Cell pellet was resuspended in either lysis buffer (NaCl 150 mM, Tris-HCl pH 7.2 50 mM, NP-40 1%, Desoxycholate Na 1%) or milder lysis buffer (NaCl 150 mM, Tris-HCl pH 7.2 50 mM, NP-40 0.5%, glycerol 10%) with 1X protease inhibitor cocktail (Roche) and incubated for 1 hr at 4°C under agitation. Cell lysates were cleared by centrifugation at 15 000 g for 20 min at 4°C. Supernatants were incubated with either GFP-TRAP (Chromotek) or HA-coupled beads (Sigma) as indicated for 1 or 2 hr at 4°C. The beads were collected and washed five times with 1 ml of ice-cold lysis buffer before SDS-PAGE. Eluates were loaded on 10 or 12% SDS-PAGE. After migration, proteins were transferred onto a P 0.45 µm PVDF or 0.45 µm nitrocellulose Amersham Hybond membranes (GE Healthcare) and immunoblotted with according antibodies. IP were repeated three times with anti-GFP for Cby-GFP/HA-Fam92 and once each (anti-HA or anti-GFP) for Cby-GFP/HA-Dzip1 and IP was performed once with anti-GFP for GFP-Dzip1/V5-Fam92/HA-Cby.

Primary antibodies used for western blotting were the following: rat anti-HA (1:5000; Roche), rabbit anti-GFP (1:10000; Abcam) and mouse anti-V5 (1:5000; Invitrogen). HRP-conjugated secondary antibodies were the following: goat anti-rabbit (1:10000; Biorad), goat anti-mouse (1:3000; Biorad) and goat anti-rat (1:20000; Sigma). Membranes were visualized using ECL prime from GE Healthcare.

Fly stocks and maintenance

Request a detailed protocol

The cby1 mutant and cby::Tomato transgene were previously described (Enjolras et al., 2012). The cep290::GFP transgene was a gift from T. Avidor- Reiss (University of Toledo, USA; Blachon et al., 2009; Basiri et al., 2014). CG6652::GFP, mks1::GFP and unc::mKate transgenes have been described in Vieillard et al. (2016). sas4s2214 mutant flies were kindly provided by R. Basto (Basto et al., 2006). Flies were raised on standard media between 21°C and 25°C. Df(3R)Exel8178 (dzip1 deficiency), Df(2L)Exel6033 (fam92 deficiency) were obtained from the Bloomington Stock Center and uncover the 95F8-−96A6 and 33E4-−33F2 cytological interval respectively. cep2900153-G4 (Bloomington Drosophila Stock Center) harbors a piggy-bac insertion in the beginning of exon 11.

Immunofluorescence and confocal microscopy

Request a detailed protocol

Testes from young adult flies or pupae were dissected in PBS1X, fixed 15 min in PBS1X/PFA 4% and either whole mounted or squashed between coverslip and slide and frozen in liquid nitrogen. Coverslip was removed and slides were soaked 2 min in ethanol 100% at −20°C. Testes were treated 15 min in PBS1X/Triton 0.1% (PBT) and blocked 2 hr in PBT/BSA 3%/NGS 5%.

Primary antibodies were incubated in blocking buffer overnight at 4°C. Samples were then washed in PBS1X and incubated 2 hr in secondary antibodies diluted in PBS1X. Slides were washed in PBS1X and rinsed in ultrapure water. Slides were mounted using Vectashield containing Hoescht 1:1000.

Antennae were processed as previously described (Vieillard et al., 2015). Briefly, Drosophila heads from 38 to 45 hr pupae were dissected in PBS1X, fixed 1 hr in PBS1X/PFA 4% and washed in PBS1X. Antennae were permeabilized 1 hr in PBS1X/Triton 0.3%, blocked 1 hr in PBS1X/Triton 0.3%/BSA 3%/NGS 5% and incubated in primary antibodies diluted in blocking solution for 48 hr at 4°C. Samples were washed three times in PBS1X and incubated in secondary antibodies diluted in PBS1X for 48 hr at 4°C. Antennae were washed three times in PBS1X and mounted in Vectashield.

For the whole mount staining of antennae, Drosophila heads from 38 to 45 hr pupae were dissected in PBS1X/0.3% Triton X-100 and fixed in in PBS1X/0.3% Triton X-100/4% PFA for 1 hr at RT. After a few rinses, samples were incubated 1 hr at RT in PBS1X/0.3% Triton X-100/BSA 0.1% and then in phalloidin FluoProbes 505 or 547 (Interchim) diluted 1:200 in PBS1X/0.3% Triton X-100/BSA 0.1% at 4°C for 48 hr. The samples were washed 3 times for 15 min in PBS1X and mounted in Vectashield. All fluorescent observations were performed on at least five different pairs of testes or antennae. Quantifications were performed on at least three independent experiments.

Most slides were imaged using either the IX83 microscope from Olympus equipped with an iXon Ultra 888 EMCDD camera from Andor and the IQ3 software from Andor. The PlanApo N Apochromat 60 × 1.42 NA objective from Olympus was used for all acquisitions. Some slides were imaged using an SP5X confocal laser scanning microscope (Leica Biosystems) equipped with the Application Suite software (Leica Biosystems). An HCX Plan-Apochromat CS 63 × 1.4 NA objective (Leica Biosystems) was used for all acquisitions.

All images were processed with ImageJ. Figures were created with Adobe photoshop. Only contrasts and offset were adjusted.

3D-SIM microscopy

Request a detailed protocol

Testes or antennae were squashed on 12 diameter round coverslip with a 44 × 60 mm overlaying coverslip. Immunofluorescence protocols were the same as above using PFA. Images were acquired using the Elyra PS.1 system from Zeiss (Carl Zeiss, AG, Jena) equipped with a PCO edge 5.5 camera and the ZEN 2012 SP2 software (black edition). The objective used for all acquisitions is a Plan-apochromat 63 × 1.4 NA.

Antibodies

The antibodies used were the following: mouse anti-Futsch/22C10 (1:250; DSHB = 22C10), mouse anti-γ-Tubulin (1:500; Sigma), rabbit anti-GFP (1:1000; Abcam), guinea pig and rat anti-Asterless (1:50000; gift from C. Rogers, University of Arizona, Tucson, USA; Klebba et al., 2013; McLamarrah et al., 2018).

The following secondary antibodies were used (all at 1:1000 dilution): goat anti-mouse Alexa 594, goat anti-rabbit Alexa 488 or Alexa 647, goat anti-guinea pig Alexa 488 or Alexa 594, goat anti-rat Alexa 488 or Alexa 555 or Alexa 647 (Invitrogen) and donkey anti-guinea pig Alexa 647 (Jackson Immuno Research).

Quantifications

Request a detailed protocol

Bang assays were performed as previously described (Enjolras et al., 2012). Approximately 20 staged female flies were placed in graduated tubes (8 cm) and banged on the table at t = 0. The % of flies that reached each quarter of the tube was counted after 30 s. three different batchs were quantified for each genotype.

Quantifications of mother versus daughter centrioles were based on the principle that the daughter centriole, being nucleated from the mother centriole, is orthogonally positioned on the lateral wall of the mother centriole. Quantifications of TZ protein intensities was performed using ImageJ and by measuring the sum of pixel intensity in a defined region encompassing the centrioles. Background intensity was measured by measuring the sum of pixel intensity in an area close to and devoid of centrioles and then subtracted to TZ protein intensities.

Statistics

Results of fluorescence intensity quantifications are represented as scatter plots with the mean and SD on all figures. Statistical significance was determined by two-tailed unpaired parametric Student’s t test (Figure 4A, fertility assay; Figure 5D, Cby-Tom; Figure 6A, Cep290-GFP in fam921) or non-parametric Mann-Withney’s test when variances were not comparable (all other intensity quantifications of Figure 5D and Figure 6A–C). Results of phenotypic proportions are represented as contingency bar graph and statistical significance was determined by two-tailed Chi-square (Figures 5D and 6A). (Prism six software; ns, p>0.05; *, p≤0.05; **, p≤0.01; ***, p≤0.001; ****, p≤0.0001).

Electron microscopy

Request a detailed protocol

Samples were processed as previously described (Enjolras et al., 2012; Vieillard et al., 2015). Observations were performed on at least two independent tissue samples.

Data availability

All data generated or analysed during this study are included in the manuscript and supporting files.

References

    1. Baker K
    2. Beales PL
    (2009) Making sense of cilia in disease: the human ciliopathies
    American Journal of Medical Genetics Part C: Seminars in Medical Genetics 151C:281–295.
    https://doi.org/10.1002/ajmg.c.30231
  44. Book
    1. Pasmans JH
    2. Tates A
    (1971)
    Cytodifferentiation During Spermatogenesis in Drosophila Melanogaster: An Electron Microscope Study
    The University of Michigan.

Decision letter

  1. Piali Sengupta
    Reviewing Editor; Brandeis University, United States
  2. Anna Akhmanova
    Senior Editor; Utrecht University, Netherlands

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Thank you for submitting your article "Dzip1 and Fam92 form a ciliary transition zone module with cell type specific roles in Drosophila" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Anna Akhmanova as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

I have left the original reviews below. But as you will note, the majority of the comments ask you to please revise the text so that the conclusions more accurately reflect the data and to avoid over-interpretation. Additional comments require re-analysis of the data and presentation of additional images. However, in the post-review discussion, all reviewers concur that as currently presented, there is insufficient experimental proof that Dzip1/Fam92/Cby are in a 'biochemical module', and ask that you perform additional experiments to support this statement.

Reviewer #1:

The ciliary transition zone (TZ) is crucial for ciliogenesis, and mutations in genes encoding TZ proteins lead to ciliopathies. Two pairs of recently discovered ciliary TZ proteins in mammalian cells are DZIP1/1L and FAM92A/B, both of which interact with the previously identified TZ protein CBY1. In previous studies, these proteins were shown to have roles in ciliogenesis and hedgehog signaling. However, their specific functions in these processes and their relationship to each other are poorly understood.

In this manuscript, the authors provide the first analysis of the Drosophila orthologs of these proteins, Dzip1 and Fam92. In both chordotonal (Ch) neurons and developing male germ cells, these proteins are needed to restrict localization of the TZ protein Cep290 to the TZ at the base of the cilium. However, Dzip1 and Fam92 otherwise play distinct roles in the two types of ciliated cells. In Ch neurons, they are needed for normal ciliogenesis, with dzip1 mutants lacking cilia entirely and fam92 mutants having reduced numbers of cilia. Basal bodies (BB) are still associated with the tips of dendrites in these cells. However, Dzip1 and Fam92 are needed for docking of BB at the plasma membrane in primary spermatocytes. Interestingly, in dzip1 and fam92 mutants, docking of daughter centrioles is affected more severely than docking of mother centrioles, and the undocked centrioles form elongated cytoplasmic structures marked by the Drosophila lebercilin ortholog CG6652. Electron micrographs are consistent with the idea that Dzip1 and Fam92 are needed for formation of a cap that prevents this aberrant elongation. In addition, the authors demonstrate that 1) Cep290 acts upstream to ensure proper localization of Dzip1 and Fam92, 2) Dzip1 itself is needed for robust localization of Cby, Fam92 and the Meckel Syndrome protein Mks1, and 3) Cby and Fam92 are reciprocally required for each other's localization at the TZ.

Overall, the authors' results nicely lay out the cell type-specific requirements, localization and regulatory interactions of this group of TZ proteins. However, there are several caveats to the authors' claims that will need to be resolved before the manuscript is suitable for publication.

1) The authors claim "that Dzip1/Fam92/Cby form a biochemical module at the ciliary transition zone", and that they "all interact to form a biochemical complex". However, their biochemical data are limited to showing coIPs between Cby and Fam92 or Cby and Dzip1, but not between Fam92 and Dzip1. To support their conclusions, Fam92 and Dzip1 should be tested for their ability to coIP. Even if they do not coIP when co-expressed, they might form a tripartite complex with Cby, an idea that could be tested by co-expressing all three proteins and pulling on them sequentially, using different antibodies, to determine whether they interact indirectly as part of a multiprotein complex containing Cby.

2) It is unclear why the authors refer to CG6652-GFP as an axonemal marker (subsection “Dzip1 and Fam92 cap the basal body and facilitate its docking to the plasma membrane in spermatocytes”, first paragraph). In wild-type spermatocytes, CG6652-GFP is restricted to the distal tips of the BB, whereas in fam92, dzip1 and cep290 mutants, CG6652-GFP localizes to an extended structure that could either be an axoneme or an elongated TZ. I could not find evidence that CG6652-GFP localizes to axonemes in differentiating spermatids, either in the current manuscript or in the authors' previously published paper (Vieillard et al., 2016). A supplementary micrograph illustrating CG6652-GFP localization in whole-mount wild-type testes would be informative in this regard.

3) Given the authors' observation that BB are defective in docking in 8/13 Dzip1 and Fam92 mutant spermatocytes (subsection “Tissue specific functional properties of the TZ proteins in BB docking”, first paragraph) and their conclusion "that Dzip1 and Fam92 are at least involved in the initial steps of BB docking", it seems contradictory to claim that Dzip1 and Fam92 are involved in "maintaining basal body/membrane docking in Drosophila testes". The latter statement should be reworded to make clear that Dzip1 and Fam92 are needed for docking to initiate/occur (rather than be maintained).

4) Based on the data provided, it seems unclear whether "Dzip1 and Fam92 form a functional module which caps centrioles" or if their role in this process is indirect. It would be necessary to perform immuno-EM to determine if these proteins are physically associated with the centriolar cap. Alternatively, the authors should soften their conclusion regarding the function of Dzip1 and Fam92 in this process.

5) Supplementary file 1 and 2 are missing, so it is not possible evaluate the CBY1-associated proteins identified by mass spectrometry (subsection “Dzip1/Fam92/Cby form a complex module at the ciliary transition zone in Drosophila”, first paragraph).

Reviewer #2:

This is an interesting paper that requires some modification to remove several statements that over-interpret the data.

The authors show by that Cby, Dzip1 and Fam82 co-IP and from this claim the proteins form a biochemical complex. However, there is no evidence presented showing how these proteins might physically interact – if indeed they do physically interact. The authors go on to say that the three proteins perfectly co-localise – referring to Figure 1C. This reader is not able to see this perfect co-localisation. Thus, the conclusion reached at the end of this sections – that "Dzip1, Fam82 and Cby form a biochemical module" has not been demonstrated.

The difference in the strength of the mutant phenotypes for the dzip1 and fam82 also seems to make the notion that their gene products form a biochemical entity appear unlikely. The authors report that a hypomorphic allele of dzip1 has a strong uncoordinated phenotype whereas what appears to be a null allele of fam92 exhibits a weaker phenotype. It is difficult to relate these phenotypes to any biochemical module. Reflecting the severity of the phenotype, dzip1 mutants lack neurosensory cilia whereas fam82 flies could have with zero or one cilium in the scolopidia. This is a curious phenotype that perhaps suggests some mother – daughter difference in one of the cell generations leading to the mature scolopidium.

The authors then present convincing evidence that Dzip1 is needed to recruit both Cby and Fam92… and that Fam82 is required to recruit Cby but not Dzip1. They also see that in the absence of Dzip1, the domain of Cep290 staining is expanded. As this reader understands the text (clarification required), there "a few" (please define) examples of expanded Cep 290. Finally, in a strong Cep290 mutants both Dzip1 and Fam92 are reduced. This is interpreted to mean that Cep290 is required to recruit Dzip1 and Fam92 and these feed back to regulate localisation of Cep290.

Finally, the authors examine mutant defects in spermatocytes. My reading of this – although the text could be clarified – is that in both fam92 and dzip1 mutants, many centriole pairs show axonemal extensions from just one of the two centrioles. This is a particular interesting part of the phenotype that could be further characterised – or at least discussed in relation to the potential mother-daughter differences seen in the sensory cilia.

Reviewer #3:

The manuscript by Lapart et al. identifies Dzip1 and Fam92 as Chibby (Cby) interactors and convincingly show that Dzip1 and Fam92 localize and function at the transition zone and basal body is essential for ciliogenesis. In spermatocytes, authors make the exciting and novel discovery that fam92 and dzip1 mutants display an asymmetric phenotypic difference between mother and daughter centriole, with the daughter being more sensitive to TZ perturbation. This manuscript is supported by strong data, high quality imaging, and eloquently writing and will appeal to the readers of eLife who are interested in cilia, cell biology, germ cells, and human genetic diseases.

https://doi.org/10.7554/eLife.49307.sa1

Author response

Reviewer #1:

[…] Overall, the authors' results nicely lay out the cell type-specific requirements, localization and regulatory interactions of this group of TZ proteins. However, there are several caveats to the authors' claims that will need to be resolved before the manuscript is suitable for publication.

1) The authors claim "that Dzip1/Fam92/Cby form a biochemical module at the ciliary transition zone", and that they "all interact to form a biochemical complex". However, their biochemical data are limited to showing coIPs between Cby and Fam92 or Cby and Dzip1, but not between Fam92 and Dzip1. To support their conclusions, Fam92 and Dzip1 should be tested for their ability to coIP. Even if they do not coIP when co-expressed, they might form a tripartite complex with Cby, an idea that could be tested by co-expressing all three proteins and pulling on them sequentially, using different antibodies, to determine whether they interact indirectly as part of a multiprotein complex containing Cby.

We agree with the reviewer that our first results did not demonstrate that all three proteins interact together in a same complex, but only that Cby interacts with either Fam92 or Dzip1. We have performed supplementary co-IPs and observed that Dzip1 and Fam92 do not apparently interact together, but when expressed with Cby, immunoprecipitation of Dzip1 pulls down both Cby and Fam92, suggesting that all three are present in one complex. These data are now included in Figure 1—figure supplement 1D and described in the first paragraph of the Results section.

Nevertheless, we agree that the term biochemical module might be too strong as our functional data also indicate that specific functions can be attributed to Dzip1 compared to Fam92 or Cby (see comments to reviewer 2 below). We have thus modified the text accordingly.

2) It is unclear why the authors refer to CG6652-GFP as an axonemal marker (subsection “Dzip1 and Fam92 cap the basal body and facilitate its docking to the plasma membrane in spermatocytes”, first paragraph). In wild-type spermatocytes, CG6652-GFP is restricted to the distal tips of the BB, whereas in fam92, dzip1 and cep290 mutants, CG6652-GFP localizes to an extended structure that could either be an axoneme or an elongated TZ. I could not find evidence that CG6652-GFP localizes to axonemes in differentiating spermatids, either in the current manuscript or in the authors' previously published paper (Vieillard et al., 2016). A supplementary micrograph illustrating CG6652-GFP localization in whole-mount wild-type testes would be informative in this regard.

Indeed, we never published the full description of CG6652-GFP distribution. We have included additional data on Figure 6—figure supplement 2 showing the expression of CG6652 in whole mount testes, illustrating the distribution of CG6652 along the sperm tail. In support of this axonemal association, CG6652-GFP is lost in sas4 mutant elongated spermatids and we have added these observations on Figure 6—figure supplement 2 and in the first paragraph of the subsection “Dzip1 and Fam92 facilitate basal body docking to the plasma membrane in spermatocytes”.

3) Given the authors' observation that BB are defective in docking in 8/13 Dzip1 and Fam92 mutant spermatocytes (subsection “Tissue specific functional properties of the TZ proteins in BB docking”, first paragraph) and their conclusion "that Dzip1 and Fam92 are at least involved in the initial steps of BB docking", it seems contradictory to claim that Dzip1 and Fam92 are involved in "maintaining basal body/membrane docking in Drosophila testes". The latter statement should be reworded to make clear that Dzip1 and Fam92 are needed for docking to initiate/occur (rather than be maintained).

Indeed, the majority of basal bodies fail to initially dock and our writing was misleading. We have modified the statement as follows: “it also reveals tissue specific function of these proteins in priming basal body/membrane docking in Drosophila testes”.

4) Based on the data provided, it seems unclear whether "Dzip1 and Fam92 form a functional module which caps centrioles" or if their role in this process is indirect. It would be necessary to perform immuno-EM to determine if these proteins are physically associated with the centriolar cap. Alternatively, the authors should soften their conclusion regarding the function of Dzip1 and Fam92 in this process.

We agree with the reviewer that we do not prove that Dzip1 and Fam92 directly cap the centrioles. We have thus modified the text accordingly at several occurrences.

5) Supplementary file 1 and 2 are missing, so it is not possible evaluate the CBY1-associated proteins identified by mass spectrometry (subsection “Dzip1/Fam92/Cby form a complex module at the ciliary transition zone in Drosophila”, first paragraph).

We apologize for this problem and have provided the missing files.

Reviewer #2:

This is an interesting paper that requires some modification to remove several statements that over-interpret the data.

The authors show by that Cby, Dzip1 and Fam82 co-IP and from this claim the proteins form a biochemical complex. However, there is no evidence presented showing how these proteins might physically interact – if indeed they do physically interact. The authors go on to say that the three proteins perfectly co-localise – referring to Figure 1C. This reader is not able to see this perfect co-localisation. Thus, the conclusion reached at the end of this sections – that "Dzip1, Fam82 and Cby form a biochemical module" has not been demonstrated.

We agree with the reviewer that we did not demonstrate that all three physically interact together in a same complex but only that Cby and Fam92 or Dzip1 can be co-immuno-precipitated. We have performed supplementary co-IPs and observed that Dzip1 and Fam92 do not reside together in a complex stable enough to be immuno-precipitated. However, when co-expressed with Cby, immunoprecipitation of Dzip1 pulls down both Cby and Fam92, suggesting that all three are indeed present in one complex. These data are now included in Figure 1—figure supplement 1D and described in the first paragraph of the Results section.

We nevertheless modified the text to clearly state that we do not demonstrate physical interactions but only co-purification properties.

The difference in the strength of the mutant phenotypes for the dzip1 and fam82 also seems to make the notion that their gene products form a biochemical entity appear unlikely. The authors report that a hypomorphic allele of dzip1 has a strong uncoordinated phenotype whereas what appears to be a null allele of fam92 exhibits a weaker phenotype. It is difficult to relate these phenotypes to any biochemical module. Reflecting the severity of the phenotype, dzip1 mutants lack neurosensory cilia whereas fam82 flies could have with zero or one cilium in the scolopidia. This is a curious phenotype that perhaps suggests some mother – daughter difference in one of the cell generations leading to the mature scolopidium.

We agree with the reviewer that the difference in the severity of the phenotypes of dzip1 and fam92 mutant flies suggest that at least Dzip1 has additional functions than only recruiting Fam92 and Cby. We anticipate that Dzip1 interacts also with other components, yet to be identified, that are also required to organize the TZ. We agree that if Dzip1, Fam92 and Cby likely interact in a ternary complex based on CoIP experiments, their range of action is not limited to these interactions. This is now more clearly stated in the Discussion section.

We thank the reviewer for the suggested hypothesis of mother-daughter difference in cell generation leading to the scolopidium, but we think that testing this hypothesis goes beyond the scope of this manuscript.

The authors then present convincing evidence that Dzip1 is needed to recruit both Cby and Fam92… and that Fam82 is required to recruit Cby but not Dzip1. They also see that in the absence of Dzip1, the domain of Cep290 staining is expanded. As this reader understands the text (clarification required), there "a few" (please define) examples of expanded Cep 290. Finally, in a strong Cep290 mutants both Dzip1 and Fam92 are reduced. This is interpreted to mean that Cep290 is required to recruit Dzip1 and Fam92 and these feed back to regulate localisation of Cep290.

We have clarified the text as requested by the reviewer.

Finally, the authors examine mutant defects in spermatocytes. My reading of this – although the text could be clarified – is that in both fam92 and dzip1 mutants, many centriole pairs show axonemal extensions from just one of the two centrioles. This is a particular interesting part of the phenotype that could be further characterised – or at least discussed in relation to the potential mother-daughter differences seen in the sensory cilia.

Yes indeed, we observed extensions more often on the daughter centriole as presented in the Results section and discussed. We added one sentence at the end of the Introduction and modified the sentence in the Results section to better emphasize on this result.

https://doi.org/10.7554/eLife.49307.sa2

Article and author information

Author details

  1. Jean-André Lapart

    Institut NeuroMyoGène, CNRS UMR 5310, INSERM U1217, Université Claude Bernard Lyon 1, Lyon, France
    Contribution
    Validation, Investigation, Visualization, Writing—original draft, Writing—review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-9167-8391

  2. Marco Gottardo

    Institute of Human Genetics, Universitätsklinikum Düsseldorf, Heinrich-Heine-Universität Düsseldorf, Düsseldorf, Germany
    Contribution
    Validation, Investigation, Visualization, Writing—review and editing
    Competing interests
    No competing interests declared

  3. Elisabeth Cortier

    Institut NeuroMyoGène, CNRS UMR 5310, INSERM U1217, Université Claude Bernard Lyon 1, Lyon, France
    Contribution
    Validation, Investigation, Visualization, Methodology
    Competing interests
    No competing interests declared

  4. Jean-Luc Duteyrat

    Institut NeuroMyoGène, CNRS UMR 5310, INSERM U1217, Université Claude Bernard Lyon 1, Lyon, France
    Contribution
    Validation, Investigation, Visualization, Methodology
    Competing interests
    No competing interests declared

  5. Céline Augière

    Institut NeuroMyoGène, CNRS UMR 5310, INSERM U1217, Université Claude Bernard Lyon 1, Lyon, France
    Present address
    Centre de recherche du CHU de Québec (CHUL), Québec, Canada
    Contribution
    Validation, Investigation
    Competing interests
    No competing interests declared

  6. Alain Mangé

    IRCM, INSERM, Université de Montpellier, ICM, Montpellier, France
    Contribution
    Formal analysis, Validation, Investigation, Writing—original draft
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-1566-9407

  7. Julie Jerber

    Institut NeuroMyoGène, CNRS UMR 5310, INSERM U1217, Université Claude Bernard Lyon 1, Lyon, France
    Present address
    Open Targets, Wellcome Genome Campus, Cambridge, United Kingdom
    Contribution
    Validation, Investigation, Methodology
    Competing interests
    No competing interests declared

  8. Jérôme Solassol

    IRCM, INSERM, Université de Montpellier, ICM, Montpellier, France
    Contribution
    Resources, Validation, Project administration
    Competing interests
    No competing interests declared

  9. Jay Gopalakrishnan

    Institute of Human Genetics, Universitätsklinikum Düsseldorf, Heinrich-Heine-Universität Düsseldorf, Düsseldorf, Germany
    Contribution
    Resources, Validation, Writing—review and editing
    Competing interests
    No competing interests declared

  10. Joëlle Thomas

    Institut NeuroMyoGène, CNRS UMR 5310, INSERM U1217, Université Claude Bernard Lyon 1, Lyon, France
    Contribution
    Conceptualization, Supervision, Validation, Investigation, Writing—original draft, Writing—review and editing
    Contributed equally with
    Bénédicte Durand
    For correspondence
    joelle.thomas@univ-lyon1.fr
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-0461-6131

  11. Bénédicte Durand

    Institut NeuroMyoGène, CNRS UMR 5310, INSERM U1217, Université Claude Bernard Lyon 1, Lyon, France
    Contribution
    Conceptualization, Supervision, Funding acquisition, Validation, Investigation, Writing—original draft, Project administration, Writing—review and editing
    Contributed equally with
    Joëlle Thomas
    For correspondence
    durand-b@univ-lyon1.fr
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-8530-0613

Funding

Agence Nationale de la Recherche (DIVERCIL)

  • Jean-André Lapart
  • Elisabeth Cortier
  • Jean-Luc Duteyrat
  • Céline Augière
  • Julie Jerber
  • Joëlle Thomas
  • Bénédicte Durand

Fondation pour la Recherche Médicale (FRM DEQ20131029168)

  • Jean-André Lapart
  • Elisabeth Cortier
  • Jean-Luc Duteyrat
  • Céline Augière
  • Julie Jerber
  • Joëlle Thomas
  • Bénédicte Durand

Université Claude Bernard Lyon 1 (PhD Fellowship)

  • Céline Augière

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank T Avidor-Reiss for the cep290::GFP Drosophila stock and C Rogers for sharing the rat and guinea pig anti-Asterless. Many thanks to E Goillot for helpful discussions and M NGassa for technical assistance. This project was supported by a grant from the Fondation pour la Recherche Médicale, FRM DEQ20131029168 and the Agence Nationale de la Recherche (ANR Divercil). JAL was supported by the FRM DEQ20131029168. CA was supported by a PhD fellowship from the University Claude Bernard Lyon-1. We thank the PLATIM, CiQLE and CTµ platforms of the LyMIC for confocal and EM microscopy. We are grateful to Elodie Chatre for helpful discussions.

Mass spectrometry experiments were carried out using facilities of the Plaform ‘Proteomic Imaging and molecular Interactions’ (PP2I) of Montpellier.

Senior Editor

  1. Anna Akhmanova, Utrecht University, Netherlands

Reviewing Editor

  1. Piali Sengupta, Brandeis University, United States

Publication history

  1. Received: June 13, 2019
  2. Accepted: November 12, 2019
  3. Version of Record published: December 10, 2019 (version 1)

Copyright

© 2019, Lapart et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 1,496
    Page views
  • 182
    Downloads
  • 10
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Jean-André Lapart
  2. Marco Gottardo
  3. Elisabeth Cortier
  4. Jean-Luc Duteyrat
  5. Céline Augière
  6. Alain Mangé
  7. Julie Jerber
  8. Jérôme Solassol
  9. Jay Gopalakrishnan
  10. Joëlle Thomas
  11. Bénédicte Durand
(2019)
Dzip1 and Fam92 form a ciliary transition zone complex with cell type specific roles in Drosophila
eLife 8:e49307.
https://doi.org/10.7554/eLife.49307

Categories and tags

Research organism

Further reading

    1. Cell Biology
    2. Neuroscience

    APOE expression and secretion are modulated by mitochondrial dysfunction

    Meghan E Wynne, Oluwaseun Ogunbona ... Victor Faundez
    Research Article Updated

    Mitochondria influence cellular function through both cell-autonomous and non-cell autonomous mechanisms, such as production of paracrine and endocrine factors. Here, we demonstrate that mitochondrial regulation of the secretome is more extensive than previously appreciated, as both genetic and pharmacological disruption of the electron transport chain caused upregulation of the Alzheimer’s disease risk factor apolipoprotein E (APOE) and other secretome components. Indirect disruption of the electron transport chain by gene editing of SLC25A mitochondrial membrane transporters as well as direct genetic and pharmacological disruption of either complexes I, III, or the copper-containing complex IV of the electron transport chain elicited upregulation of APOE transcript, protein, and secretion, up to 49-fold. These APOE phenotypes were robustly expressed in diverse cell types and iPSC-derived human astrocytes as part of an inflammatory gene expression program. Moreover, age- and genotype-dependent decline in brain levels of respiratory complex I preceded an increase in APOE in the 5xFAD mouse model. We propose that mitochondria act as novel upstream regulators of APOE-dependent cellular processes in health and disease.

    1. Cell Biology

    The Uso1 globular head interacts with SNAREs to maintain viability even in the absence of the coiled-coil domain

    Ignacio Bravo-Plaza, Victor G Tagua ... Miguel A Peñalva
    Research Article

    Uso1/p115 and RAB1 tether ER-derived vesicles to the Golgi. Uso1/p115 contains a globular-head-domain (GHD), a coiled-coil (CC) mediating dimerization/tethering and a C-terminal region (CTR) interacting with golgins. Uso1/p115 is recruited to vesicles by RAB1. Genetic studies placed Uso1 paradoxically acting upstream of, or in conjunction with RAB1 (Sapperstein et al., 1996). We selected two missense mutations in uso1 resulting in E6K and G540S in the GHD that rescued lethality of rab1-deficient Aspergillus nidulans. The mutations are phenotypically additive, their combination suppressing the complete absence of RAB1, which emphasizes the key physiological role of the GHD. In living hyphae Uso1 recurs on puncta (60 sec half-life) colocalizing partially with the Golgi markers RAB1, Sed5 and GeaA/Gea1/Gea2, and totally with the retrograde cargo receptor Rer1, consistent with Uso1 dwelling in a very early Golgi compartment from which ER residents reaching the Golgi recycled back to the ER. Localization of Uso1, but not of Uso1E6K/G540S, to puncta is abolished by compromising RAB1 function, indicating that E6K/G540S creates interactions bypassing RAB1. That Uso1 delocalization correlates with a decrease in the number of Gea1 cisternae supports that Uso1-and-Rer1-containing puncta are where the protein exerts its physiological role. In S-tag-coprecipitation experiments Uso1 is an associate of the Sed5/Bos1/Bet1/Sec22 SNARE complex zippering vesicles with the Golgi, with Uso1E6K/G540S showing stronger association. Using purified proteins, we show that Bos1 and Bet1 bind the Uso1 GHD directly. However, Bet1 is a strong E6K/G540S-independent binder, whereas Bos1 is weaker but becomes as strong as Bet1 when the GHD carries E6K/G540S. G540S alone markedly increases GHD binding to Bos1, whereas E6K causes a weaker effect, correlating with their phenotypic contributions. AlphaFold2 predicts that G540S increases binding of the GHD to the Bos1 Habc domain. In contrast, E6K lies in an N-terminal, potentially alpha-helical, region that sensitive genetic tests indicate as required for full Uso1 function. Remarkably, this region is at the end of the GHD basket opposite to the end predicted to interact with Bos1. We show that unlike dimeric full-length and CTR∆ Uso1 proteins, the GHD lacking the CC/CTR dimerization domain, whether originating from bacteria or Aspergillus extracts and irrespective of whether it carries or not E6K/G540S, would appear to be monomeric. With the finding that overexpression of E6K/G540S and wild-type GHD complement uso1∆, our data indicate that the GHD monomer is capable of providing, at least partially, the essential Uso1 functions, and that long-range tethering activity is dispensable. Rather, these findings strongly suggest that the essential role of Uso1 involves the regulation of SNAREs.

    1. Cell Biology

    Global analysis of contact-dependent human-to-mouse intercellular mRNA and lncRNA transfer in cell culture

    Sandipan Dasgupta, Daniella Y Dayagi ... Jeffrey E Gerst
    Research Article

    Full-length mRNAs transfer between adjacent mammalian cells via direct cell-to-cell connections called tunneling nanotubes (TNTs). However, the extent of mRNA transfer at the transcriptome-wide level (the 'transferome') is unknown. Here, we analyzed the transferome in an in vitro human-mouse cell co-culture model using RNA-sequencing. We found that mRNA transfer is non-selective, prevalent across the human transcriptome, and that the amount of transfer to mouse embryonic fibroblasts (MEFs) strongly correlates with the endogenous level of gene expression in donor human breast cancer cells. Typically, <1% of endogenous mRNAs undergo transfer. Non-selective, expression-dependent RNA transfer was further validated using synthetic reporters. RNA transfer appears contact-dependent via TNTs, as exemplified for several mRNAs. Notably, significant differential changes in the native MEF transcriptome were observed in response to co-culture, including the upregulation of multiple cancer and cancer-associated fibroblast-related genes and pathways. Together, these results lead us to suggest that TNT-mediated RNA transfer could be a phenomenon of physiological importance under both normal and pathogenic conditions.