Identification of genes involved in the ciliary trafficking of C. elegans PKD-2


  • Young-Kyung Bae,

    1. Department of Genetics, Rutgers University, The State University of New Jersey, Piscataway, New Jersey
    2. Laboratory of Genetics, University of Wisconsin, Madison, Wisconsin
    Search for more papers by this author
    • Young-Kyung Bae and Jamie Lyman-Gingerich contributed equally to this work.

  • Jamie Lyman-Gingerich,

    1. Division of Pharmaceutical Sciences, University of Wisconsin, Madison, Wisconsin
    Current affiliation:
    1. Department of Biology, University of Wisconsin-Eau Claire, Eau Claire, WI 54702
    Search for more papers by this author
    • Young-Kyung Bae and Jamie Lyman-Gingerich contributed equally to this work.

  • Maureen M. Barr,

    Corresponding author
    1. Department of Genetics, Rutgers University, The State University of New Jersey, Piscataway, New Jersey
    2. Division of Pharmaceutical Sciences, University of Wisconsin, Madison, Wisconsin
    • Rutgers University, The State University of New Jersey, Department of Genetics, 145 Bevier Road Piscataway, NJ 08854
    Search for more papers by this author
  • Karla M. Knobel

    1. Division of Pharmaceutical Sciences, University of Wisconsin, Madison, Wisconsin
    Current affiliation:
    1. Department of Biochemistry, University of Wisconsin-Madison, 433 Babcock Drive, Madison, WI 53706
    Search for more papers by this author


Ciliary membrane proteins are important extracellular sensors, and defects in their localization may have profound developmental and physiological consequences. To determine how sensory receptors localize to cilia, we performed a forward genetic screen and identified 11 mutants with defects in the ciliary localization (cil) of C. elegans PKD-2, a transient receptor potential polycystin (TRPP) channel. Class A cil mutants exhibit defects in PKD-2::GFP somatodendritic localization while Class B cil mutants abnormally accumulate PKD-2::GFP in cilia. Further characterization reveals that some genes mutated in cil mutants act in a tissue-specific manner while others are likely to play more general roles in such processes as intraflagellar transport (IFT). To this end, we identified a Class B mutation that disrupts the function of the cytoplasmic dynein light intermediate chain gene xbx-1. Identification of the remaining mutations will reveal novel molecular pathways required for ciliary receptor localization and provide further insight into mechanisms of ciliary signaling. Developmental Dynamics 237:2021–2029, 2008. © 2008 Wiley-Liss, Inc.


Cells have developed extraordinary measures to sense the external environment, including the formation of specialized sensory organelles called cilia. Sensory receptors and ion channels localize to and function within the ciliary membrane. In humans, defects in cilia formation or function cause the “ciliopathies,” a class of severe genetic diseases that affect millions (Christensen et al.,2007). While a fair amount is known regarding how cilia form, relatively little is understood regarding ciliary targeting of membrane proteins.

The simple and transparent anatomy of the nematode Caenorhabditis elegans enables visualization of ciliogenesis and ciliary transport in a living organism (Scholey et al.,2004). In addition to the 60 core ciliated sensory neurons, the male possesses an additional 48 ciliated neurons. All of these cilia are located on dendritic endings, exhibit diverse morphologies, and express distinct sensory receptors on their surface (Sulston et al.,1980; Perkins et al.,1986; Bargmann,2006). PKD-2/TRPP2 and the polycystin-1 (PC1) receptor LOV-1 are expressed in the male-specific CEM neurons of the head and the ray RnB and hook HOB neurons of the tail, which are required for chemotaxis to mates (Chasnov et al.,2007), response to mate contact, and location of the mate's vulva (Liu and Sternberg,1995; Barr and Sternberg,1999; Barr et al.,2001). In humans, defects in PC1 or TRPP2 result in autosomal dominant polycystic kidney disease (ADPKD) (Igarashi and Somlo,2002). Since our initial discovery that LOV-1 and PKD-2 function in male-specific sensory cilia (Barr and Sternberg,1999), the mammalian polycystins were found on primary cilia of renal epithelial cells and shown to act as flow sensors (Pazour et al.,2002; Yoder et al.,2002; Nauli et al.,2003). The evolutionarily conserved ADPKD genetic pathway, polycystin ciliary localization, and polycystin sensory function make the nematode an attractive model to study both PKD-2 ciliary trafficking and ADPKD, the most common human monogenic disease (Igarashi and Somlo,2002).

In C. elegans male-specific neurons, endogenous and GFP-tagged PKD-2 localize to the cell body endoplasmic reticulum (ER), dendritic puncta, and the cilium (Barr et al.,2001; Bae et al.,2006). Two subcellular sorting steps regulate PKD-2 localization (Bae et al.,2006). First, the somatodendritic step directs PKD-2 in the cell body ER to vesicles that are transported by an unidentified dendritic motor to the ciliary base. UNC-101, the mu1 subunit of clathrin adaptor protein complex 1 (AP-1), is required for directing PKD-2 and other sensory receptors to dendrites of diverse sensory neurons (Dwyer et al.,2001; Bae et al.,2006). LOV-1, the C. elegans polycystin-1 homolog, is required for efficient somatodendritic targeting of PKD-2::GFP (Bae et al.,2006). In addition to UNC-101 and LOV-1, unidentified cell type–specific factors in pkd-2-expressing neurons act in the somatodendritic sorting step (Bae et al.,2006). Next, the ciliary sorting step, which involves KLP-6 (Peden and Barr,2005), intraflagellar transport (IFT) (Bae et al.,2006), calcineurin (Hu et al.,2006), and the Hrs-STAM ubiquitin-sorting complex (Hu et al.,2007), regulates PKD-2 access to and abundance in the cilium. Here the N- and C-tails of PKD-2 regulate ciliary abundance, but are not essential for PKD-2 ciliary targeting (Knobel et al.,2008). This contrasts with the ciliary targeting of mammalian TRPP2, which requires an RVxP region in the cytosolic N-terminus (Geng et al.,2006). Although much is known about PKD-2 targeting from previous studies, many questions remain. What molecules regulate the vesicular packaging and polarized targeting of PKD-2 to the dendrite? What motors drive PKD-2 along the sensory dendrite? How is PKD-2 insertion at the ciliary base and PKD-2 abundance in the ciliary membrane regulated? To identify new genes involved in ciliary membrane protein trafficking, we performed a genetic screen looking for mutants with PKD-2::GFP ciliary localization (Cil) defects. We isolated 11 mutants with defects in either the somatodendritic sorting step (Class A, abnormal dendritic PKD-2 localization) or the ciliary sorting step (Class B, abnormal PKD-2 ciliary accumulation) in CEM neurons. Ongoing characterization and cloning of these cil genes will provide important insight into how sensory receptors are delivered to cilia.


Identification of Ciliary Localization (Cil) Defective Mutants

In wild-type neurons, PKD-2::GFP is localized to somatodendritic (ER in the cell body and puncta along the dendrite) and ciliary regions of 4 CEM, 16 RnB, and the HOB male-specific sensory neurons (Figs. 1A, 2A, 3A). We performed a forward genetic screen for mutants with PKD-2::GFP ciliary localization (Cil) defects (Fig. 1B) and identified 11 mutants that were grouped with respect to their PKD-2::GFP localization patterns in the head CEM neurons (Table 1). Six Class A mutants exhibited abnormal distribution of PKD-2::GFP in the CEM dendrites (cil-1(my15), cil-2(my2), my9, cil-3(my11), my12, my22; Fig. 2). Five Class B mutants accumulated PKD-2::GFP in CEM ciliary regions (cil-5(my13), my14, my16, my17, my21; Fig. 2).

Figure 1.

Identification of new genes that regulate PKD-2::GFP ciliary localization in sensory neurons. A: Schematic of a C. elegans sensory neuron expressing PKD-2::GFP. PKD-2::GFP localizes to the cell body ER and cilium (including the cilium proper and ciliary base) and small moving dendritic puncta. Arrows in the dendrite indicate that these puncta move in both anterograde and retrograde directions. B: F1 and F2 clonal screens were performed with EMS mutagenized myIs1[Ppkd-2::PKD-2::GFP; Pcc::GFP] pkd-2(sy606)IV; him-5(e1490)V hermaphrodites. F2 or F3 adult males were scored for PKD-2::GFP localization defects in head CEM and tail ray neurons. In total, 11 mutants were identified from 1,032 haploid genomes. C: xbx-1 genomic region, mutations, and the XBX-1 protein, which is comprised of a single dynein light intermediate chain (DLIC) domain.

Figure 2.

PKD-2::GFP localization in the CEM neurons of wild type (A, G), Class A (B, C, H, I, J), and Class B (D, E, F, K, L, M, N) mutants. In all panels, double arrowheads indicate the cilium proper (▴), arrowheads indicate the ciliary base (≪), and arrows indicate the dendrite. A–F: The distribution of PKD-2::GFP throughout the entire neuron (projected confocal images, scale bar = 10 μm). Images in A–F are projected confocal images. G–N: Generated from projected epifluorescent images after 3D devolution overlayed on Normarski pictures. A: In wild-type males, PKD-2::GFP is localized in the ER of cell bodies (double-headed arrow), the cilium proper (≪), and the ciliary base (▴), which corresponds to the distal most dendritic ending and ciliary transition zone. B: In cil-2(my2) mutants, PKD-2::GFP is mislocalized throughout the dendrite and small GFP particles are visible outside the dendrite (near arrow). PKD-2::GFP accumulates at the ciliary base. C: my12 mutants exhibit punctate PKD-2::GFP localization along the dendrite (near arrow). PKD-2::GFP accumulates in the cilium proper and within the ciliary base. D: my14 mutants exhibit increased PKD-2::GFP in the cilium proper and the ciliary base. E: In my16 males, PKD-2::GFP accumulates at the ciliary base. F: PKD-2::GFP accumulates in the ciliary base of xbx-1(my17) mutants. G–N: The localization of PKD-2::GFP in the ciliary region (compare with Table 2; scale bar = 5 μm). A dotted line indicates the outline of the buccal cavity. The ciliary base is located midway between the anterior and posterior ends of the cavity. G: In wild-type CEM neurons, PKD-2::GFP localizes to the ciliary base and cilium proper. H: In a my9 mutant, PKD-2::GFP accumulates at the ciliary base and in the cilium proper. A double-arrow points out PKD-2::GFP localization in the distal dendrite (underlying the buccal cavity) and more proximal dendrite. I: In a my11 mutant, PKD-2::GFP accumulates in the cilium proper. Both the cilium proper and the base exhibit morphological defects such as a hook-like projection (open arrowhead). PKD-2::GFP is visible in the dendrite (arrow). J: In a my22 mutant, PKD-2::GFP accumulates in the cilium proper. There are increased numbers of PKD-2::GFP puncta in the cilium proper, ciliary base, and dendrite. K: In a cil-5(my13) mutant, the cilium bends inward towards the buccal cavity, much like a kap-1 or klp-11 mutant (Bae et al.,2006). PKD-2::GFP puncta accumulate at the ciliary base. L: In a my14 mutant, PKD-2::GFP levels are increased in the cilium proper, which is often bent inward like cil-5(my13). The morphology of the ciliary base is also abnormal: hook-like structures are often visible (open arrowhead). M: In a my16 mutant, PKD-2::GFP accumulates in small puncta at the ciliary base. PKD-2::GFP labeling is also visible in the distal but not proximal dendrites (double arrows). N: In an xbx-1(my17) mutant, PKD-2::GFP aggregates at the ciliary base and PKD-2::GFP localization extends into the distal (not proximal) dendrite (double arrows). When visible, the xbx-1(my17) cilium proper appears stunted.

Figure 3.

PKD-2::GFP localization in ray neurons. A: In wild-type, PKD-2::GFP is found in the cell bodies and ciliary regions (arrowheads) at the ray tips. BD: In my22, cil-5(my13), and my16 males, PKD-2::GFP accumulates along the distal part of ray dendrites (two arrows). These images are projected epifluorescent images after 3D deconvolution. Scale bar = 5 μm.

Table 1. Phenotypic and Genetic Characterization of cil Mutants
ClassGene (allele)PKD-2::GFP Localization PhenotypeBehavior PhenotypeDye filling (Dil)LG
CEMsRaysResponse efficiencybMating efficiencyc
DendritesCiliary regiona
  • a

    Ciliary region includes the cilium proper and ciliary base. ↑: increased; ↓: reduced; WT: wild-type; LG: linkage group; sem: standard error of mean; NA; not applicable; ND: not determined, *: significantly different from the control, p < 0.05 (t-test). Behavioral Assays were performed on outcrossed mutants, in some cases the original marked chromosome was removed (myls 1 [PKD-2::GFP; Pcc::GFP1]pkd-2). All data were statistically compared to the appropriate control.

  • b

    Response Efficiency (RE): each trial consists of 10–25 males. % Response ± sem (number of trials).

  • c

    Mating Efficiency (ME): for each trial, six males and two dpy-17 (e904) hermaphrodites are placed on a plate. % ME = sem (number of trials).

  • d

    Control: him-5(e1490) V, the cil mutants were in the him-5(e1490) V background.

  • e

    Control: myls 1[PKD-2::GFP; cc::GFP] pkd-2 (sy606) IV; him-5(e1490) V, the experimental group carried myls 1pkd-2 (sy606) IV; him-5(e1490) V.

  • f

    Occasionally PKD-2::GFP is visible in the distal dendrite below the ciliary base.

  • g

    Amphids weakly fill with Dil, but phasmids are completely dye-filling defective (Dyf).

Acil-2(my2)↑ in/outsideWT79.2 ± 5.1 (5)d64.3 ± 5.3 (11)dWTX
 my9↑ in/outsideWT87.5 ± 1.0 (6)e34.2 ± 9.0 (9)eWTX
 cil-3(my11)WT87.1 ± 2.6 (5)e31.9 ± 10.2 (7)eWTI
 my12↑ in dendrite & cilia16.3 ± 7.5 (4)*1.3 ± 1.3 (8)*WTND
 cil-1(my15)WTEven distribution83.4 ± 3.9 (5)d1.9 ± 0.8 (6)*WTIII
 my22↑ in dendrite & cilia26.0 ± 8.6 (3)*2.8 ± 2.5 (7)*WTI
Bcil-5(my13)WT↑ in dendrite & cilia72.5 ± 1.9 (3)*33.0 ± 10.9 (9)egIII
 my14WTWT82.5 ± 17.5 (2)e40.5 ± 10.8 (7)eWTX
 my16WTf↑ in dendrite & cilia49.2 ± 5.6 (3)*11.3 ± 4.7 (4)*WTX
 xbx-1(my17)WTf↑ in dendrite & cilia19.4 ± 0.3 (3)*0.0 ± 0.0 (6)*V
 my21WT↑ in dendrite & cilia30.6 ± 12.2 (9)*0.0 ± 0.0 (6)*WTND
Controlshim-5NANANA94.4 ± 1.0 (17)60.1 ± 4.7 (9)WTNA
 myIs1 pkd-2; him-5WT (in cilia and cell body ER)88.9 ± 2.1 (10)43.1 ± 8.9 (9)WTNA
 pkd-2; him-5NANANA22.2 ± 5.1 (12)12.9 ± 7.8 (7)WTNA
Table 2. Characterization of PKD-2::GFP in CEM Neuronsa
ClassGene (allele)CEM dendritesbCEM ciliac
% Ab (no. of CEM)PhenotypeCilium properCiliary base
% presentdPhenotype% Ab (no. of CEM)Phenotype
  • a

    Ab: Abnormal; WT: wild-type.

  • b

    PKD-2::GFP localization phenotype in CEM dendrites. SmD: continuous ‘smooth’ localization in dendrite; PuD: punctate localization dendrite; EC: extracellular puncta; DD: labeling in the distal dendrite.

  • c

    PKD-2::GFP localization phenotype in CEM cilia. ↑: Increased localization of PKD-2::GFP relative to wild-type; B: cilia are bent inward; H: hook-like fingers extend from ciliary base.

  • d

    Percentage of neurons with visible PKD-2::GFP localization in cilium proper. The number of CEM neurons scored for this column is identical to those for the ciliary base column.

Acil-2(my2)21 (326)SmD, EC puncta37WT25 (300)↑, H
 my923 (104)SmD, EC puncta53WT89 (93)↑, H
 cil-3(my11)46 (92)SmD, PuD85↑↑7.1 (92)↑.H
 my1292 (116)SmD, PuD6270 (115)
 cil-1(my15)100 (67)SmD, Axonal41WT6 (66)WT
 my2249 (102)SmD, PuD88↑↑79 (102)
Bcil-5(my13)2 (61)WT63↑, B26 (61)
 my143 (86)WT90↑↑, B41 (86)↑, H
 my1624 (59)DD6436 (58)
 xbx-1(myl7)15 (67)DD4939 (67)
 my217 (95)WT81↑↑78 (94)↑, H
Controlmyls1pkd-2;him-57.5 (80)WT56WT14 (79)WT

Each mutant was outcrossed and tested for genetic complementation when possible. At least four mutations mapped to LG (linkage group) X (Table 1). As males are XO, scoring of the Cil phenotype is impossible without introducing mutations that result in transgendered animals, confounding interpretation. One X-linked mutation, cil-2(my2), has a hermaphrodite temperature-sensitive (TS) sterility (Ste) phenotype that cosegregates with the Cil phenotype. In XX hermaphrodites, we determined that the X-linked mutations my9, my14, and my16 complement the cil-2(my2) TS Ste phenotype, indicating that there are at least two complementation groups on LGX with respect to the cil-2(my2) TS phenotype.

Class A cil Mutants Accumulate PKD-2::GFP in CEM Dendrites

Six cil mutants exhibit abnormally high levels of PKD-2::GFP in the dendrite (Class A, Table 1). cil-1(my15) resembles the unc-101(m1) mu1 AP-1 subunit mutant, where PKD-2::GFP is evenly distributed throughout the cell body, axon, dendrite, and cilium (data not shown; Bae et al.,2006). In the remaining five Class A mutants, cil-2(my2), my9, cil-3(my11), my12, and my22, PKD-2::GFP is appropriately targeted to the somatodendritic compartment but is abnormally distributed in CEM dendrites (Fig. 2B,C). In a subset of Class A mutants (cil-3(my11), my12, and my22), PKD-2::GFP puncta are easily visible in dendrites (Fig. 2C, arrow). In wild-type males, dendritic punta are only visible with extended fluorescence exposure (Bae et al.,2006).

cil-2(my2) and my9 are characterized by PKD-2::GFP puncta located outside the dendrite (Fig. 2B, arrow) as determined by co-expression with a cytosolic fluorescent reporter that labels the intact dendrite (Ppkd-2::DsRed2, data not shown). “Extracellular” PKD-2::GFP puncta do not colocalize with the dendritic marker (data not shown). The nature of these ectopic PKD-2::GFP puncta is not known. This unique phenotype has not been observed in over 100 mutants examined for PKD-2::GFP localization (Bae et al.,2006; Barr lab, unpublished data) and may reflect mutations in genes required for vesicular trafficking.

With the exception of cil-1(my15), all Class A mutants also accumulate PKD-2::GFP within the CEM ciliary region (Table 2 and Fig. 2B,C,H–J, arrowheads ▴ and ≪). In wild-type, PKD-2::GFP localizes to the ciliary base (corresponding to the distal most dendrite and ciliary transition zone) in all neurons and in 56% of the cilium proper of neurons examined (Table 2, Fig. 2G, Knobel et al.,2008). Atypical accumulation in the cilium proper occurs in my11, my12, and my22 (Fig. 2H–J, Table 2). In cil-2(my2), my9, cil-3(my11), my12, and my22, PKD-2::GFP accumulates in the enlarged and misshapen ciliary base and/or bright puncta (Fig. 2B,C,H–J arrowheads ▴, Table 2). In sum, these data suggest that the cil-2(my2), my9, my11, my12, and my22 mutants affect ciliary as well as somatodendritic targeting.

Class B cil Mutants Display Abnormal PKD-2 Ciliary Accumulation

The ciliary base is a dynamic region where ciliary receptor insertion, turnover, and degradation occur. Ciliary sorting defects may result in too much or too little PKD-2 within the cilium. The five Class B mutants (cil-5(my13), my14, my16, my17, and my21) are defined by the abnormal accumulation of PKD-2::GFP in the ciliary regions (including the cilium proper and base; Fig. 2, Table 2) with normal dendritic distribution in CEM neurons. In most Class B mutants (cil-5(my13), my14, my16, and my21), PKD-2::GFP is more likely to be observed in the cilium proper than in wild-type (Fig. 2K–M, and Table 2). In contrast to wild-type CEM cilia that curve outward, cil-5(my13) and my14 CEM cilia bend inward (Fig. 2G vs. K,L), a morphological defect reminiscent of klp-11 and kap-1 IFT kinesin-II mutants (Bae et al,2006). In my16 and my17 mutants, PKD-2::GFP accumulations extend to the dendrite just below the ciliary base (Fig. 2M,N, Table 2). The abnormal accumulation of PKD-2::GFP in CEM ciliary regions suggests that Class B genes may regulate PKD-2 insertion, distribution, down-regulation, or recycling.

Reproductive Consequences of Cil Defects

pkd-2 is required for three steps in C. elegans male mating behavior: sexual attraction toward a Caenorhabditis remanei female (Chasnov et al.,2007), response to a mate, and location of vulva (Barr and Sternberg,1999). To determine if abnormal PKD-2 localization alters male reproductive abilities, we measured male response behavior efficiency (RE) and mating efficiency (ME) of the cil mutants (Table 1). A response efficiency (RE) assay measures in minutes the male's behavioral ability to respond to hermaphrodite contact (Liu and Sternberg,1995; Barr et al.,2001; Peden and Barr,2005). Alternatively, a ME assay measures in days the male's fertility (Hodgkin,1983). It should be noted that RE assays directly test PKD-2 function in rays (Liu and Sternberg,1995; Barr and Sternberg,1999). Accordingly, we examined the PKD-2::GFP distribution phenotype of the ray neurons in the cil mutants (Fig. 3, Table 1). In wild-type ray neurons, PKD-2::GFP localizes to cell bodies and ciliary regions (Fig. 3A, arrowheads).

pkd-2 males have a low RE (22%) and ME (11%) relative to wild-type males (94% and 60%, respectively; Table 1; Barr et al.,2001). Four mutants (Class A: cil-2(my2), my9, cil-3(my11); Class B: my14) display normal RE, ME, and PKD-2::GFP localization patterns in the rays (data not shown). These data suggest that the PKD-2::GFP mislocalization in CEM neurons alone does not interfere with response behavior. In contrast, five mutants (Class A: my12, my22; Class B: my16, my17, my21) have both RE and ME defects as well as abnormal accumulation of PKD-2::GFP in ray distal dendrites and cilia (Fig. 3B–D, arrows). Here, mislocalization of PKD-2 in the rays may have functional consequences. Alternatively, these mutants may have additional defects in ciliogenesis, copulation, or fertility.

Two mutants escape categorization. cil-1(my15) mutants display wild-type response behavior (normal RE) but fail to produce cross progeny (ME defective, Table 1), indicating additional defects that affect fertility. Conversely, cil-5(my13) males have RE but not ME defects (Table 1). In cil-1(my15) rays, PKD-2::GFP is evenly distributed in axons and dendrites, indicating that PKD-2 mislocalization in cil-1(my15) does not preclude PKD-2 function. In cil-5(my13) mutants, PKD-2::GFP accumulates in CEM ciliary regions (Fig. 2K, Table 2) and ray distal dendrites and ciliary regions (Figure 3C), which may account for the RE defect. The cil-5(my13) low RE may be overcome by the ample time provided to the male during ME assays, resulting in normal ME.

We conclude that, in general, PKD-2::GFP mislocalization in rays correlates with a defect in response behavior. Similarly, klp-6 and IFT mutant males exhibit similar accumulation of PKD-2::GFP in ray dendrites and response defects (Peden and Barr,2005; Bae et al.,2006). The PKD-2::GFP marker is a powerful tool to simultaneously study polycystin localization and sensory functions.

General Ciliogenesis in cil Mutants

PKD-2 mislocalization may be an indirect result of ciliogenesis defects. General ciliogenesis requires intraflagellar transport (IFT) (Rosenbaum and Witman,2002). To determine if any cil mutants have general ciliogenesis defects, we used the lipophilic fluorescent dye DiI to assess the integrity of amphid channel and phasmid cilia (Perkins et al.,1986). DiI does not label the pkd-2-expressing neurons. Most IFT mutants are dye filling defective (Dyf) due to stunted cilia (Perkins et al.,1986; Starich et al.,1995). Only two Class B mutants (cil-5(my13) and my17) are Dyf (Table 1). In cil-5(my13) animals, amphid neurons fill weakly while phasmids never fill with DiI, suggesting cil-5 is required for cell type–specific ciliogenesis. my17 is completely Dyf, indicating general ciliogenesis is disrupted.

To determine the nature of the my17 ciliogenesis defect, we cloned the gene. my17 was mapped to the region of LGV between marker pkP5085 (6.62cM) and unc-76 (7.31 cM), which includes xbx-1, a dynein light intermediate chain gene required for retrograde IFT. Like other IFT components, XBX-1::YFP localizes to the transition zones and axonemes of amphid cilia (Schafer et al.,2003). Similar to the my17 mutant, the null xbx-1 (ok279) deletion mutant is Dyf and ME defective (Schafer et al.,2003). Sequencing of the xbx-1 locus in my17 animals revealed a nonsense mutation in exon 6 that changes a tryptophan (W316) to a termination codon (Fig. 1C), which is likely degraded by nonsense-mediated mRNA decay (NMD) (Chang et al.,2007). The my17 allele failed to complement the Dyf defects of the xbx-1(ok279) null deletion allele (Schafer et al.,2003). Finally, a fluorescent-tagged wild-type xbx-1 coding region fused to the osm-5 promoter (and thus expressed in all ciliated neurons; Posm-5::XBX-1:YFP), rescued the my17 Dyf phenotype. We conclude that my17 is a mutation in xbx-1. As expected in a mutant deficient in a retrograde IFT motor, xbx-1(my17) mutants have bulbous cilia in addition to accumulation of PKD-2 in ciliary base in CEM (Fig. 2F,N, Bae et al.,2006).


Classical genetic screens provide a powerful and unbiased approach to study questions of biological importance (Jorgensen and Mango,2002). Here, we use forward genetics to define mechanisms of PKD-2 ciliary receptor targeting in C. elegans. We previously proposed that two sorting steps regulate PKD-2 localization (Bae et al.,2006), and this study has identified genes acting in both of these PKD-2 sorting steps. We also isolated mutants with novel phenotypes suggesting that unidentified targeting mechanisms exist. In wild-type worms, TRPP2/PKD-2 localizes to the cell body endoplasmic reticulum (ER) and is selectively targeted to dendritic particles that move at characteristic anterograde and retrograde rates (Bae et al.,2006). These particles presumably fuse with the plasma membrane at the ciliary base releasing PKD-2 into the ciliary membrane (Reiter and Mostov,2006), as demonstrated for rab8-docked, rhodopsin-bearing vesicles in photoreceptors (Moritz et al.,2001). Mutations in Class A genes disrupt the distribution of PKD-2 between the dendrite and cilium of a CEM neuron. Class B mutants affect PKD-2 levels at the CEM ciliary base and in the cilium proper.

Class A mutants exhibit PKD-2::GFP dendritic mislocalization in CEM neurons. We predict that Class A cil genes act at the somatodendritic sorting step, which regulates PKD-2 trafficking from the cell body ER, to the dendrite, and ultimately to the ciliary base. Some genes may act early in the PKD-2 somatodendritic targeting process. For example, the AP-1 mu1 clathrin adaptor UNC-101 is involved in polarized dendritic transport of PKD-2 and other ciliary membrane receptors (Dwyer et al.,2001; Bae et al.,2006). With the exception of cil-1(my15), all class A mutants display accumulation of PKD-2::GFP in the ciliary regions as well as in dendrites (Table 1), indicating crosstalk between the somatodendritic and ciliary sorting steps. The class A mutants cil-3(my11), my12, and my22 exhibit an increase in PKD-2::GFP dendritic puncta and an accumulation of PKD-2::GFP in the cilium. These mutations may reflect a disruption in vesicular transport, the failure to retain PKD-2 in the cell body ER, or an inability to regulate PKD-2::GFP ciliary abundance. We observed the mislocalization of PKD-2::GFP throughout the dendrite and outside the distal dendrite in two Class A mutants cil-2(my2) and my9, suggesting a defect in vesicular transport or secretion. Although each of the Class A mutants shares a common phenotype (dendritic accumulation of PKD-2::GFP in CEM neurons), each likely represents a mutant that affects a distinct aspect of PKD-2 localization.

PKD-2 ciliary abundance is tightly regulated via post-translational modifications (Hu et al.,2006), other polycystins (Bae et al.,2006), the kinesin-3 KLP-6 (Peden and Barr,2005), intraflagellar transport (IFT) (Qin et al.,2005), and the STAM-Hrs ubiquitin degradation machinery (Hu et al.,2007). Class B mutants (cil-5(my13), my14, my16, xbx-1(my17), my21) accumulate PKD-2 at the ciliary base and occasionally in the cilium proper, suggesting that these genes regulate PKD-2 ciliary abundance. We tested this prediction by cloning a Class B gene and discovered that my17 is a nonsense mutation in the xbx-1 gene, a dynein light intermediate chain protein required for retrograde IFT (Schafer et al.,2003). As expected in an IFT mutant, my17 mutants are Dyf (dye filling defective), indicative of general ciliogenic defects. cil-5(my13) is another cil mutant that is also Dyf, although not as severely defective as an IFT mutant. One possibility is that cil-5 is involved in cell-specific ciliogenesis or that cil-5 modulates but is not essential for IFT. cil-5 maps to LGIII, where there are at least four gene mutations that affect cilia formation or function: dyf-2, osm-10, osm-12/bbs-7, and kap-1. cil-5(my13) does share some phenotypes with the kinesin-II kap-1 mutant. We do not believe my13 is a mutation in osm-10 or osm-12/bbs-7, based on dissimilar phenotypes. osm-10 is not described as Dyf on and osm-12/bbs-7 mutant males exhibit normal PKD-2::GFP localization in CEMs and ray neurons (Bae et al2006). Like my13 animals, dyf-2 amphid neurons occasionally fill with lipophilic fluorescent dyes (FITC and DiO; Starich et al.,1995). The remaining Class A and B mutants are not Dyf and do not affect general ciliogenesis, and are thus likely to perturb PKD-2 ciliary levels via other mechanisms. For example, these cil genes may specifically regulate loading/unloading or degradation of PKD-2 at the ciliary base. Alternatively, these cil mutants may have defects in male-specific cilia that would not be revealed by DiI.

In this study, we focused on response behavior, one of the male sexual behaviors mediated by pkd-2 expression in the ray neurons (Barr and Sternberg,1999). PKD-2 localization and function in rays is important for male response behavior. Response defects correlate with PKD-2::GFP localization defects in rays (Class A: my12, my22, Class B: cil-5(my13), my16, xbx-1(my17), my21). A recent report found that CEM neurons contribute to male chemotaxis to Caenorhabdis remanei female- or C. elegans hermaphrodite-derived cues (Chasnov et al.,2007; White et al.,2007). The requirement of pkd-2 for sexual attraction depends on the nature of the pheromone and assay method (Chasnov et al.,2007; White et al.,2007). It will be interesting to determine if cil mutants are defective in sexual chemotaxis.

Cloning and characterization of the cil genes will provide great insight to the mechanisms regulating ciliary receptor localization. For example, Class A genes likely regulate vesicular targeting and transport of PKD-2 between the cell body and cilium, while class B genes may be required for general or cell type–specific ciliogenesis (xbx-1 or cil-5, respectively) or receptor recycling or degradation. This approach reveals the intricacy of pathways regulating PKD-2 localization and function, yet provides us with an efficient means to identify new molecular players in this process.


C. elegans Maintenance and Mapping

Nematodes were raised using standard conditions (Brenner,1974). Strain PT572 myIs1[PKD-2::GFP; Pcoelomocyte(cc)::GFP] pkd-2(sy606) IV; him-5(e1490) V (Bae et al.,2006) was mutagenized. General strains: CB1490, him-5(e1490) V; PT496, bli-6(sc16) IV; myIs1 pkd-2(sy606) IV; dpy-11(e224) him-5(e1490) V; PT575, dpy-5(e61) I; rol-6(e189) II; myIs1 pkd-2(sy606) IV; him-5(e1490) V; PT576, lon-1(e1870) III; myIs1 pkd-2(sy606) IV; him-5(e1490) V; PT621, him-5(e1490) myIs4[PKD-2::GFP; Pcc::GFP] V; PS622, dpy-17(e164) III; him-5(e1490) V; CB169, unc-31(e169) IV; JT1109, xbx-1(ok279) V; PT1260, myIs1 pkd-2(sy606) IV; him-5(e1490) my17 V; YH140, xbx-1(ok279) V Ex[Posm-5::XBX-1::YFP; pRF4], PT1003, myIs1 pkd-2(sy606) IV; him-5(e1490) V; my2 myEx392[Ppkd-2::dsRED2; Pcc::dsRED2], and PT1621, dpy-11(e224) xbx-1(my17) him-5(e1490) V. The following mutant strains were characterized: Class A: PT1201, myIs1 pkd-2(sy606) IV; him-5(e1490) V; cil-2(my2) X (outcrossed 4x); PT1636, myIs1 pkd-2(sy606) IV; him-5(e1490) V; my9 X (4x); PT1637, cil-3(my11) I; myIs1 pkd-2(sy606) IV; him-5(e1490) V (3x); PT1640, my12; myIs1 pkd-2(sy606) IV; him-5(e1490) V (3x); PT1444, cil-1(my15) III; him-5(e1490) myIs4 V (6x); PT1645, my22 myIs1 pkd-2(sy606) IV; him-5(e1490) V (1x); Class B: PT1271, cil-5(my13) III; myIs1 pkd-2(sy606) IV; him-5(e1490) V (6x); PT1279, myIs1 pkd-2(sy606) IV; him-5(e1490) V; my14 X (6x); PT1646, myIs1 pkd-2(sy606) IV; him-5(e1490) V; my16 X (6x); PT1260, xbx-1(my17) myIs4 him-5(e1490) V (6x); and PT1644, my21; myIs1 pkd-2(sy606) IV; him-5(e1490) V (2x).

Genetic Screens (Fig.1B). Synchronized L4 stage hermaphrodites were mutagenized with 47 mM ethyl methane sulphonate (EMS) for 4 hr and allowed to self-fertilize at 15°C (Anderson,1995). F1 clonal screen. 96 F1 progeny were cloned as L4s to individual plates and raised at 20°C. Three mutants were isolated. 96 F1s were raised at 15°C and 4 mutants were isolated. F2 clonal screen. Mutagenized hermaphrodites were raised at 15°C, 324 F1s collected on large plates in groups of 5–10, and raised at 20°C. Five to ten F2 progeny from each pooled plate were cloned and raised at 25°C (F2-25C), 20°C (F2-20C), or 15°C (F2-15C). 44 F3 clones were scored from the F2-25C plates (producing no mutants), 114 from the F2-20C plates (2 mutants), and 82 from the F2-15C plates (2 mutants). In total, 1,032 haploid genomes were screened for PKD-2::GFP localization defects.

Genetic Characterization of New Mutants

Mutants were outcrossed at least two times to remove additional mutations acquired during mutagenesis and assigned an allelic designation. Mutants were mapped to a linkage group (LG) using 2-point mapping techniques (Brenner,1974, Where possible, mutations were three-factor- and SNP-mapped and tested for genetic complementation (

Epifluorescence or Confocal Microscopic Analysis of Mutant Worms

Staged L4 animals were collected and raised overnight at 20°C. Adult males were mounted on agarose pads set on microscope slides and anesthetized using 5 mM Levamisole or 50 mM NaN3. Fluorescence analysis was performed on a Zeiss Axioskop II using a 63× (NA 1.4) oil objective. CEM and ray epifluorescent analysis involved the observation of cell body, dendritic and ciliary morphology as defined by PKD-2::GFP expression. Adult worms were individually scored for the localization of PKD-2::GFP in the cell body, along the dendrite, and in the cilium of the four CEM neurons and the 16 RnB neurons. Presence of a visible cilium proper was noted as was the morphology of the ciliary base outlined by PKD-2::GFP. Mutant and wild-type worms were scored blindly. Confocal images (Fig. 2A–F) were collected on a 63× (NA 1.4) objective on a Zeiss LSM510 Meta laser-scanning microscope (AIM™ software). One-micrometer optical sections were collected and projected as Z-series that were stored as TIFF files and manipulated using Adobe PhotoShop™. Epifluorescence and Normarski images (Fig. 2G–N and Fig. 3) were taken with a 100× (NA 1.4) oil objective on a Zeiss Axioscope II equipped with a Cascade 512B CCD camera (Photometrics). Z-sections were collected with an interval of 1 μm using the MetaMorph software (Molecular Devices), and deconvoluted with the AutoQuantX (MediaCybernetics). To visualize the outline of the animals, 3D deconvoluted epifluorescent Z-stacks were projected and overlaid with Normarski images taken in the same plane.

Dye Filling Assays

Dye filling was performed as described (Perkins et al.,1986).

Mating Analysis

Mating and response efficiency assays were performed as described ( with the exception that males were given 4 min instead of 10 in response assays. Pair-wise comparisons were made using two-sided t-tests.

XBX-1 Molecular Analysis

my17 and xbx-1(ok279) V were tested for complementation by mating him-5(e1490) myIs4[PKD-2::GFP; Pcc::GFP] V males with xbx-1(ok279) hermaphrodites. F1 cross progeny hermaphrodites were mated with dpy-11(e224) my17 him-5(e1490) V males. NonDpy cross progeny were scored for the presence of the myIs4 array and tested for Dye filling. To clone the gene, myIs1 pkd-2(sy606) IV; him-5(e1490) my17 hermaphrodites were mated with him-5(e1490) males. myIs1/+ IV, him-5(e1490) my17/him-5(e1490) V males were mated to hermaphrodites carrying Posm-5::XBX-1::YFP and a dominant marker that causes transgenic worms to roll (Rol) (YH140). In the F1 generation, cross progeny (Rol) were cloned and allowed to self-fertilize. F2s were tested by dye-filling to identify my17 homozygotes, and these were observed for the roller phenotype. Evidence that Posm-5::XBX-1::YFP rescued the my17 mutation came in the form of Rol F2s that were rescued for dye filling (all nonRol worms were Dyf). These animals were singled out and tested in the F3 generation.


We thank Douglas R. Braun, Dr. Dana Byrd, Min-Chen Sun, and Anjan S. Shah for experimental assistance; Dr. Judith Kimble for generous use of her confocal microscope; Dr. Bradley Yoder for the Posm-5::XBX-1::YFP construct; Dr. Erik Jorgensen, the National Bioresource Project for the Nematode C. elegans; and the Caenorhabditis Genetics Center for strains. We are grateful for the thoughtful and constructive comments made by our three reviewers. This work was funded by the PKD Foundation (to M.M.B. and K.M.K.), American Heart Association (to J.L.G.), and NIH (to M.M.B).