Neurons in a part of the eye called the retina detect light and convert it into electrical signals that are sent to the brain. Different types of neurons in the retina are arranged vertically into layers and horizontally in a mosaic pattern so that two neurons of the same type are not next to each other. To establish this highly organized pattern, neurons in the developing retina must be able to recognize other neurons of the same type and avoid moving towards them – a process referred to as self-avoidance.

A group of proteins called the Dscams are found on the surface of neurons and play key roles in positioning them in the retina. Dscams promote self-avoidance, help to establish connections between certain neurons and kill any excess neurons that are not needed. However, the mechanisms by which Dscams serve these three roles were not known.

Scaffolding proteins in the cell interior interact with Dscams to hold them in place on the cell surface. Garrett et al. investigated whether Dscams need to interact with the scaffolding proteins in order to carry out any of their activities.

The experiments used mice that had been genetically engineered to produce mutant Dscam proteins that cannot bind to the scaffolding proteins. Garrett et al. hypothesized that this would affect the activities of Dscams in all of the different types of neurons in the retina. However, the experiments show that the mutant Dscam proteins had different effects on the neurons. Some types of neurons developed normally, while others experienced disruptions in all three of the processes that Dscams are normally involved in. Some other neurons were affected to a moderate extent. This indicates that Dscams use different mechanisms in different types of neurons to carry out the same activities. The next step is to find out what other proteins Dscams need to interact with in different types of neurons.

The vertebrate retina provides an advantageous model to study how specific neuronal cell types organize themselves during development to form functional circuits. The ~100 neuronal cell types of the retina vertically organize into layers, with light-transducing photoreceptors in the outermost cellular layer (ONL, or outer nuclear layer) and retinal ganglion cells (RGCs) – responsible for the sole output from the retina – in the innermost layer (RGL, or retinal ganglion cell layer). Between these, both functionally and physically, are the interneuron cell types (horizontal, amacrine, and bipolar cells) in the inner nuclear layer (INL), which process and relay visual information from photoreceptors to RGCs. These three cellular layers are separated by two synaptic plexiform layers, where the neurites from these cell types mingle and form synaptic connections with specific partners in specific substrata. Additionally, many cell types are horizontally spaced in a mosaic pattern, such that there is a low probability of finding two neurons of the same subtype (i.e., homotypic) in close proximity. This pattern ensures that the information processing provided by each subtype is distributed across the retina (Masland, 2012).

Establishing this pattern requires cells to be able to recognize other cells of the same type and avoid them, while also stably interacting with appropriate synaptic partners (Garrett and Burgess, 2011). Critical parts of this recognition code include Down syndrome cell adhesion molecule (DSCAM) and the highly similar Dscam-like1 (DSCAML1), collectively referred to as Dscams. Dscam and Dscaml1 encode homophilic members of the Ig-superfamily of cell adhesion molecules, and are expressed in non-overlapping neuronal subtypes in the retina (Agarwala et al., 2001; Fuerst et al., 2009; Yamagata and Sanes, 2008). The Dscams promote self-avoidance at the cell type level: When either gene is mutated, the cell types that normally express the Dscam lose their mosaic spacing and often form clusters (Fuerst et al., 2009, 2008). The neurites fail to evenly cover their receptive fields, and instead form fascicles with neighboring homotypic cells. This clustering and fasciculation is, with few exceptions, homotypic – cells of one subtype rarely cluster with the cells of another subtype. Self-avoidance requires homophilic Dscam interactions between cells, demonstrated in mosaic experiments where neurons lacking Dscam fasciculate with homotypic neurons with intact Dscam (Fuerst et al., 2012). This self-avoidance function is consistent with studies in Drosophila, which have four Dscam genes. Most notably, Dscam1 promotes self-avoidance at the individual cell level by using alternative splicing to produce 19,008 distinctly homophilic isoforms, allowing each neuron to recognize and avoid 'self' while still interacting with 'non-self' during processes like dendrite arborization and axon branching (reviewed in [Zipursky and Grueber, 2013]).

Dscams are also required for normal developmental cell death. In the mutant mice, there is an overabundance of each affected cell type, resulting in a severe expansion of the retina through the cellular and plexiform layers. The extent of cell number expansion varies with cell type. Some cell types are expanded beyond even that seen in mutants for the pro-apoptotic Bax gene, while others are more modestly expanded compared to Bax mutants (Fuerst et al., 2009, 2008; Keeley et al., 2012).

Dscams also contribute to the vertical organization of the retina. In chick, Dscams label-specific sublaminae of the IPL and can instruct neurite targeting to these layers (Yamagata and Sanes, 2008). DSCAM protein localization is punctate throughout the IPL in mouse, and is not confined to specific sublaminae (de Andrade et al., 2014). Despite this, some neuronal types do have disorganized neurite stratification in Dscam mutants, although the disorganization varies with genetic background (Fuerst et al., 2010). There are also indications that the synaptic connections that form do not mature normally. For instance, Dscaml1 is expressed both in rod bipolar cells and AII amacrine cells, which connect at dyad ribbon synapses in the IPL. These synapses can still be found in Dscaml1 mutants, but are morphologically abnormal, with indistinct, detached presynaptic ribbons, and functionally abnormal, with much slower decay of the synaptic current (Fuerst et al., 2009).

The mechanisms by which the Dscams mediate their developmental functions are unknown. One attractive hypothesis is that different functions, including cell death, self-avoidance, and synapse maturation, are mediated by different molecular interactions in the cytosol. Signaling molecules have been found in complex with Dscam1 in Drosophila (e.g., Dock/Pak, Ableson, tubulin binding cofactor D [Okumura et al., 2015; Schmucker et al., 2000; Sterne et al., 2015]) and DSCAM in vertebrates (e.g., PAK1, FAK, Fyn [Purohit et al., 2012]). Both DSCAM and DSCAML1 also have canonical PDZ-interacting motifs at their C-termini by which they interact with scaffolding proteins in the MAGI (membrane-associated guanylate kinase with inverted orientation) and PSD95 families (Yamagata and Sanes, 2010). Because this motif is common to both Dscams, we chose to test the functional significance of these C-terminal interactions, with the initial hypothesis that this interaction would be required for a specific subset of Dscams’ functions.

We engineered mouse mutations in which the sequences encoding the final 10 amino acids of DSCAM and DSCAML1, which interact with PDZ domain-containing proteins, were replaced with epitope tags. Contrary to our initial hypothesis, we found that rather than distinguishing phenotypes based on specific molecular mechanisms, these mutations distinguished cell types. Some cell types were essentially like controls in cell number, spacing, and stratification, whereas others were nearly as disrupted as null mutants. Still other cell types were intermediate in severity. Together, these results demonstrate that different cell types have different dependencies on the PDZ-interacting C-termini of Dscams for function, indicating multiple intracellular molecular mechanisms are involved.

We tested the functional significance of the Dscams’ PDZ-interacting domains in vivoby replacing the C-terminal ten amino acids with epitope tags (Dscam^∆C and Dscaml1^∆C). The Dscams have similar functions in the different cell types in which they are expressed: (1) They promote developmental cell death, (2) they promote self-avoidance at the cell type level both between cell bodies and neurites, preventing cell body clustering and fasciculation of processes, and (3) they can promote the laminar specificity of neurite stratification and synapse maturation (Fuerst et al., 2009, 2008). We hypothesized that interactions with PDZ domain-containing proteins via the C-termini would mediate a subset of these functions. Instead, we found that a subset of cell types required the PDZ interaction for all of these functions; while in other cell types these processes could proceed relatively normally without the Dscams’ C-termini. The phenotypes found in ∆C mutants compared to null mutants are summarized by cell type in Table 1.

Phrased genetically, the Dscam^-/- and Dscaml1^-/- alleles represent a complete loss of function. The functions we have defined, promoting developmental cell death, promoting self-avoidance in both processes and cell bodies to enable mosaic spacing and coverage factor, and promoting laminar specificity and synapse maturation in at least some cell types and genetic backgrounds, are based on the loss of function mutant phenotype. The deletion of the PDZ-binding C-termini of DSCAM and DSCAML1 had several possible outcomes in this context. It could have resulted in an unstable or mis-trafficked protein, causing a complete loss of function. Our data do not support this, as protein is found at normal levels by both immunoprecipitation and immunohistochemistry, is on the cell surface by live cell staining and surface biotinylation, and is in the correct topological orientation, although more subtle or specific defects may still have been missed by these analyses. The PDZ-interacting C-termini of DSCAM and DSCAML1 could also have been totally superfluous for its function. Consistent with previous results (Yamagata and Sanes, 2010), this also is not the case. A partial loss of function could have resulted in a subset of the phenotypically defined functions being present, and others being lost. Given the interaction with synaptic scaffolding proteins, a phenotype of impaired synapse maturation with normal developmental cell death and self-avoidance was a reasonable anticipated outcome. Interestingly, we instead found a partial loss of function at the level of cell types. Some cell types appear to almost completely require the C-termini, whereas in other cell types they are almost completely dispensable. In yet other cell types, the phenotype was intermediate, between wild type and the null, but all phenotypically defined functions changed roughly in parallel. An additional caveat is that the epitope tags used to replace the PDZ-binding motifs may contribute to these phenotypes. While this is formally possible, we believe that it is unlikely to be the case, as we used two different epitope tags (Myc and HA) with two different genes and found a similar range of phenotypes between different cell types. Thus, our main conclusion from these studies is that the requirement for a PDZ-interacting C-terminus in DSCAM or DSCAML1 is variable and depends on the cell type.

How cell adhesion molecules function to prevent adhesion and promote self-avoidance remains and interesting question. The other well described mediators of self-avoidance function by generating thousands of distinctly homophilic recognition units (Zipursky and Grueber, 2013). Dscam1 in Drosophila uses three banks of alternatively spliced exons to produce 19,008 isoforms with distinct extracellular domains (Schmucker et al., 2000). Each of these three exons encode regions of the protein essential for homophilic binding, resulting in 19,008 potential molecules that preferentially recognize other copies of the same isoform (Sawaya et al., 2008; Wojtowicz et al., 2004, 2007; Wu et al., 2012). By biased stochastic exon choice, a given neuron expresses an estimated 10–50 isoforms, most of which will differ from neighboring neurons (Miura et al., 2013; Neves et al., 2004; Zhan et al., 2004). This gives each neuron a distinct fingerprint allowing it to recognize and avoid 'self' through repulsion while still interacting with its neighbors, a process called self/non-self discrimination (Hughes et al., 2007; Matthews et al., 2007; Soba et al., 2007). As might be expected with such a mechanism in which each cell is uniquely identified, vast isoform diversity is required for Dscam1 to function normally (Hattori et al., 2009, 2007).

Similarly, the vertebrate gamma protocadherin cluster (Pcdhg) can also produce diverse recognition units. There are only 22 Pcdhg isoforms, specified by alternative promoter choice within the gene cluster instead of alternative splicing, but the resulting proteins form cis multimers with trans homophilic binding specificity at the multimer level (Schreiner and Weiner, 2010; Thu et al., 2014; Wu and Maniatis, 1999). This incorporation of isoforms into multimers can generate thousands of distinctly homophilic interactors. Intriguingly, the multimer compositions are regulated not only by promoter choice, but also by the relative expression levels of each isoform. PCDHG promotes self-avoidance in cerebellar Purkinje cells and retinal starburst amacrine cells (SACs) (Lefebvre et al., 2012). Here again, diversity is required for normal function. When only one isoform was expressed in SACs, self-avoidance at the single cell level was preserved, but interactions between neighboring SACs were aberrant, resulting in circuit level dysfunction (Kostadinov and Sanes, 2015).

Both Dscam1 and Pcdhg promote self-avoidance by conferring self/non-self discrimination. Without extensive isoform diversity, vertebrate DSCAMs are not able to provide this fine level of self-recognition. Our conclusion that DSCAMs function through at least two molecular mechanisms, one that requires a C-terminal PDZ-interacting motif, and one that does not, fits a model in which DSCAMs interact with cell-type-specific adhesion mechanisms to serve their function. This is most easily discussed for self-avoidance, where the deletion of DSCAM or DSCAML1 results in the clumping and fasciculation of homotypic cells – phenotypes that could be described as excessive adhesion. Under this model, the DSCAMs serve a generic role, and the specificity and complexity of the system is conferred by the cell-type-specific adhesion mechanisms. Thus, in cell types in which the PDZ-interacting C-terminus is required, the predominant cell adhesion system may also interact with PDZ-scaffolded complexes. These systems could include Ig-superfamily CAMs such as L1-CAM or NRCAM, or possibly neurexins or neuroligins. Alternatively, the CAMs could interact with PDZ domains indirectly through mediator proteins. In cell types that do not require the C-terminus of DSCAM or DSCAML1, the predominant cell adhesion systems may not involve PDZ interactions, such as cadherins or protocadherins. In cell types with intermediate phenotypes, a mix of PDZ-dependent and –independent mechanisms may be involved. This is not surprising, as each cell type is expected to express more than one class of cell adhesion molecule.

The C-termini of both DSCAM and DSCAML1 interact with at least six different multi-PDZ domain-containing proteins (Yamagata and Sanes, 2010) and (Figure 1—figure supplement 1). de Andrade and colleagues analyzed the co-localization of DSCAM with seven different multi-PDZ proteins, including MAGI-2 and MAGI-3, and found little co-localization in the adult IPL. Interestingly, there was increased co-localization during development, indicating that these interactions are dynamic and transient (de Andrade et al., 2014). None of the analyzed PDZ proteins were clearly localized to specific laminae, but were punctate throughout the IPL. Thus, DSCAM and DSCAML1 could interact with different PDZ proteins in different cell types throughout development.

As mentioned, we have thus far been generally unable to genetically separate functions such as developmental cell death and self-avoidance. Cell types with any phenotype in ∆C mutants had multiple phenotypes, which changed roughly in parallel. For example, DA cells were more numerous with disrupted spacing and neurite fasciculation, while bNOS-positive amacrine cells and Cdh3-GFP RGCs both had a relatively normal density, spacing, and neurite arborization (Figures 2,3). In ipRGCs, cell density was variably increased and did not reach statistical significance, and cell spacing and dendrite fasciculation were both intermediate in severity between controls and null mutants (Figure 3). This covariation could suggest that there is one primary function and the other phenotypes are secondary to this. Keeley and colleagues tested if cell death was this primary function in DA cells and ipRGCs, and found that while increasing cell density by inhibiting developmental cell death could contribute to spacing defects, it was not sufficient to explain the loss of self-avoidance in Dscam^-/- mutants (Keeley et al., 2012). We confirmed those findings here, and extended them in three ways. (1) We added a third cell type, VGLUT3-positive amacrine cells, and found that their spacing was more disrupted in Dscaml1^∆C/∆C mutants than Bax^-/- mutants despite having a lower cell density (Figure 5I–L). (2) In ipRGCs, cell clustering was more severe in Dscam^∆C/∆Cmutants than in Bax^-/- mutants despite comparable cell densities (Figure 5A–D). ipRGC density in Dscam^-/- mutants is much higher than in Bax^-/- retinas, complicating the correlation of density and spacing defects. (3) We quantified neurite fasciculation independently from the cell body spacing, finding that cell density can contribute to fasciculation in ipRGCs, but does not significantly drive fasciculation in DA cells or bNOS cells (Figures 6, Figure 6—figure supplement 1). Thus, promotion of cell death is not the sole primary function of the Dscams. This is consistent with our previous finding that ipRGC clustering and fasciculation occurs in Dscam^-/-mutants even when cell density is severely reduced by Pou4f2 double mutation (Fuerst et al., 2012).

As we have not been able to induce clustering and fasciculation without an increase in cell density, it remains possible that the self-avoidance function of Dscams is primary and the clustering and fascicle formation is protective against cell death. It is also possible that the process through which the Dscams prevent fasciculation also promotes proper neurite stratification. We have now identified two examples of aberrant interaction between different cell types. DA cells co-fasciculate with M1 ipRGCs in Dscam^∆C/∆C retinas (Figure 7E) and VGLUT3-positive amacrine cells become directly adjacent to AII amacrine cells in Dscaml1^∆C/∆C mutants (Figure 7P–Q). As DA and M1 ipRGCs normally co-stratify, this interaction does not result in any misprojection. VGLUT3-positve amacrine neurites, however, are normally confined in the ON-OFF strata between the ChAT layers, not in the ON region proximal to the RGL where AII projections reside. Improper interactions with AII amacrine cells could provide a mechanism for VGLUT3 misprojection (Figure 7O–Q).

In summary, DSCAM and DSCAML1 in different cell types function through different molecular mechanisms. We hypothesize that this mechanistic diversity is based on the diverse adhesion systems masked by the DSCAMs. Determining how DSCAMs interact with other systems to mask their adhesivity for self-avoidance, and whether this is the primary function and developmental cell death and synapse maturation are secondary, will require additional studies. However, what these results clearly demonstrate is that the molecular mechanisms through which DSCAMs function will need to be defined on a cell type by cell type basis, and that a single molecular complex or pathway will not account for DSCAMs’ function(s) in all cell types.

1. 1. Agarwala KL
   2. Ganesh S
   3. Tsutsumi Y
   4. Suzuki T
   5. Amano K
   6. Yamakawa K

   (2001) Cloning and functional characterization of DSCAML1, a novel DSCAM-like cell adhesion molecule that mediates homophilic intercellular adhesion

   Biochemical and Biophysical Research Communications 285:760–772.

   https://doi.org/10.1006/bbrc.2001.5214

   • Google Scholar
2. 1. Amano K
   2. Fujii M
   3. Arata S
   4. Tojima T
   5. Ogawa M
   6. Morita N
   7. Shimohata A
   8. Furuichi T
   9. Itohara S
   10. Kamiguchi H
   11. Korenberg JR
   12. Arata A
   13. Yamakawa K

   (2009) DSCAM deficiency causes loss of pre-inspiratory neuron synchroneity and perinatal death

   Journal of Neuroscience 29:2984–2996.

   https://doi.org/10.1523/JNEUROSCI.3624-08.2009

   • Google Scholar
3. 1. de Andrade GB
   2. Kunzelman L
   3. Merrill MM
   4. Fuerst PG

   (2014)

   Developmentally dynamic colocalization patterns of DSCAM with adhesion and synaptic proteins in the mouse retina

   Molecular Vision 20:1422–1433.

   • Google Scholar
4. 1. Doyle DA
   2. Lee A
   3. Lewis J
   4. Kim E
   5. Sheng M
   6. MacKinnon R

   (1996) Crystal structures of a complexed and peptide-free membrane protein-binding domain: molecular basis of peptide recognition by PDZ

   Cell 85:1067–1076.

   https://doi.org/10.1016/S0092-8674(00)81307-0

   • Google Scholar
5. Book

   1. Elo AE

   (1978)

   The Rating of Chessplayers: Past and Present

   New York: Arco Publishing.

   • Google Scholar
6. 1. Fuerst PG
   2. Bruce F
   3. Rounds RP
   4. Erskine L
   5. Burgess RW

   (2012) Cell autonomy of DSCAM function in retinal development

   Developmental Biology 361:326–337.

   https://doi.org/10.1016/j.ydbio.2011.10.028

   • Google Scholar
7. 1. Fuerst PG
   2. Bruce F
   3. Tian M
   4. Wei W
   5. Elstrott J
   6. Feller MB
   7. Erskine L
   8. Singer JH
   9. Burgess RW

   (2009) DSCAM and DSCAML1 function in self-avoidance in multiple cell types in the developing mouse retina

   Neuron 64:484–497.

   https://doi.org/10.1016/j.neuron.2009.09.027

   • Google Scholar
8. 1. Fuerst PG
   2. Harris BS
   3. Johnson KR
   4. Burgess RW

   (2010) A novel null allele of mouse DSCAM survives to adulthood on an inbred C3H background with reduced phenotypic variability

   Genesis 48:578–584.

   https://doi.org/10.1002/dvg.20662

   • Google Scholar
9. 1. Fuerst PG
   2. Koizumi A
   3. Masland RH
   4. Burgess RW

   (2008) Neurite arborization and mosaic spacing in the mouse retina require DSCAM

   Nature 451:470–474.

   https://doi.org/10.1038/nature06514

   • Google Scholar
10. 1. Garrett AM
    2. Burgess RW

    (2011) Candidate molecular mechanisms for establishing cell identity in the developing retina

    Developmental Neurobiology 71:1258–1272.

    https://doi.org/10.1002/dneu.20926

    • Google Scholar
11. 1. Hattori D
    2. Chen Y
    3. Matthews BJ
    4. Salwinski L
    5. Sabatti C
    6. Grueber WB
    7. Zipursky SL

    (2009) Robust discrimination between self and non-self neurites requires thousands of Dscam1 isoforms

    Nature 461:644–648.

    https://doi.org/10.1038/nature08431

    • Google Scholar
12. 1. Hattori D
    2. Demir E
    3. Kim HW
    4. Viragh E
    5. Zipursky SL
    6. Dickson BJ

    (2007) Dscam diversity is essential for neuronal wiring and self-recognition

    Nature 449:223–227.

    https://doi.org/10.1038/nature06099

    • Google Scholar
13. 1. Hughes ME
    2. Bortnick R
    3. Tsubouchi A
    4. Bäumer P
    5. Kondo M
    6. Uemura T
    7. Schmucker D

    (2007) Homophilic Dscam interactions control complex dendrite morphogenesis

    Neuron 54:417–427.

    https://doi.org/10.1016/j.neuron.2007.04.013

    • Google Scholar
14. 1. Keeley PW
    2. Reese BE

    (2014) The patterning of retinal horizontal cells: normalizing the regularity index enhances the detection of genomic linkage

    Frontiers in Neuroanatomy 8:113.

    https://doi.org/10.3389/fnana.2014.00113

    • Google Scholar
15. 1. Keeley PW
    2. Sliff BJ
    3. Lee SC
    4. Fuerst PG
    5. Burgess RW
    6. Eglen SJ
    7. Reese BE

    (2012) Neuronal clustering and fasciculation phenotype in Dscam- and Bax-deficient mouse retinas

    Journal of Comparative Neurology 520:1349–1364.

    https://doi.org/10.1002/cne.23033

    • Google Scholar
16. 1. Khiripet N
    2. Khantuwan W
    3. Jungck JR

    (2012) Ka-me: a Voronoi image analyzer

    Bioinformatics 28:1802–1804.

    https://doi.org/10.1093/bioinformatics/bts253

    • Google Scholar
17. 1. Knudson CM
    2. Tung KS
    3. Tourtellotte WG
    4. Brown GA
    5. Korsmeyer SJ

    (1995) Bax-deficient mice with lymphoid hyperplasia and male germ cell death

    Science 270:96–99.

    https://doi.org/10.1126/science.270.5233.96

    • Google Scholar
18. 1. Kostadinov D
    2. Sanes JR

    (2015) Protocadherin-dependent dendritic self-avoidance regulates neural connectivity and circuit function

    eLife 4:08964.

    https://doi.org/10.7554/eLife.08964

    • Google Scholar
19. 1. Lefebvre JL
    2. Kostadinov D
    3. Chen WV
    4. Maniatis T
    5. Sanes JR

    (2012) Protocadherins mediate dendritic self-avoidance in the mammalian nervous system

    Nature 488:517–521.

    https://doi.org/10.1038/nature11305

    • Google Scholar
20. 1. Masland RH

    (2012) The neuronal organization of the retina

    Neuron 76:266–280.

    https://doi.org/10.1016/j.neuron.2012.10.002

    • Google Scholar
21. 1. Matthews BJ
    2. Kim ME
    3. Flanagan JJ
    4. Hattori D
    5. Clemens JC
    6. Zipursky SL
    7. Grueber WB

    (2007) Dendrite self-avoidance is controlled by Dscam

    Cell 129:593–604.

    https://doi.org/10.1016/j.cell.2007.04.013

    • Google Scholar
22. 1. Miura SK
    2. Martins A
    3. Zhang KX
    4. Graveley BR
    5. Zipursky SL

    (2013) Probabilistic splicing of Dscam1 establishes identity at the level of single neurons

    Cell 155:1166–1177.

    https://doi.org/10.1016/j.cell.2013.10.018

    • Google Scholar
23. 1. Neves G
    2. Zucker J
    3. Daly M
    4. Chess A

    (2004) Stochastic yet biased expression of multiple Dscam splice variants by individual cells

    Nature Genetics 36:240–246.

    https://doi.org/10.1038/ng1299

    • Google Scholar
24. 1. Okumura M
    2. Sakuma C
    3. Miura M
    4. Chihara T

    (2015) Linking cell surface receptors to microtubules: tubulin folding cofactor D mediates Dscam functions during neuronal morphogenesis

    Journal of Neuroscience 35:1979–1990.

    https://doi.org/10.1523/JNEUROSCI.0973-14.2015

    • Google Scholar
25. 1. Osterhout JA
    2. Josten N
    3. Yamada J
    4. Pan F
    5. Wu SW
    6. Nguyen PL
    7. Panagiotakos G
    8. Inoue YU
    9. Egusa SF
    10. Volgyi B
    11. Inoue T
    12. Bloomfield SA
    13. Barres BA
    14. Berson DM
    15. Feldheim DA
    16. Huberman AD

    (2011) Cadherin-6 mediates axon-target matching in a non-image-forming visual circuit

    Neuron 71:632–639.

    https://doi.org/10.1016/j.neuron.2011.07.006

    • Google Scholar
26. 1. Purohit AA
    2. Li W
    3. Qu C
    4. Dwyer T
    5. Shao Q
    6. Guan KL
    7. Liu G

    (2012) Down syndrome cell adhesion molecule (DSCAM) associates with uncoordinated-5C (UNC5C) in netrin-1-mediated growth cone collapse

    Journal of Biological Chemistry 287:27126–27138.

    https://doi.org/10.1074/jbc.M112.340174

    • Google Scholar
27. 1. Rodieck RW

    (1991) The density recovery profile: a method for the analysis of points in the plane applicable to retinal studies

    Visual Neuroscience 6:95–111.

    https://doi.org/10.1017/S095252380001049X

    • Google Scholar
28. 1. Sawaya MR
    2. Wojtowicz WM
    3. Andre I
    4. Qian B
    5. Wu W
    6. Baker D
    7. Eisenberg D
    8. Zipursky SL

    (2008) A double S shape provides the structural basis for the extraordinary binding specificity of Dscam isoforms

    Cell 134:1007–1018.

    https://doi.org/10.1016/j.cell.2008.07.042

    • Google Scholar
29. 1. Schmidt TM
    2. Chen SK
    3. Hattar S

    (2011) Intrinsically photosensitive retinal ganglion cells: many subtypes, diverse functions

    Trends in Neurosciences 34:572–580.

    https://doi.org/10.1016/j.tins.2011.07.001

    • Google Scholar
30. 1. Schmucker D
    2. Clemens JC
    3. Shu H
    4. Worby CA
    5. Xiao J
    6. Muda M
    7. Dixon JE
    8. Zipursky SL

    (2000) Drosophila Dscam is an axon guidance receptor exhibiting extraordinary molecular diversity

    Cell 101:671–684.

    https://doi.org/10.1016/S0092-8674(00)80878-8

    • Google Scholar
31. 1. Schramm RD
    2. Li S
    3. Harris BS
    4. Rounds RP
    5. Burgess RW
    6. Ytreberg FM
    7. Fuerst PG

    (2012) A novel mouse Dscam mutation inhibits localization and shedding of DSCAM

    PLoS One 7:e52652.

    https://doi.org/10.1371/journal.pone.0052652

    • Google Scholar
32. 1. Schreiner D
    2. Weiner JA

    (2010) Combinatorial homophilic interaction between gamma-protocadherin multimers greatly expands the molecular diversity of cell adhesion

    PNAS 107:.

    https://doi.org/10.1073/pnas.1004526107

    • Google Scholar
33. 1. Soba P
    2. Zhu S
    3. Emoto K
    4. Younger S
    5. Yang SJ
    6. Yu HH
    7. Lee T
    8. Jan LY
    9. Jan YN

    (2007) Drosophila sensory neurons require Dscam for dendritic self-avoidance and proper dendritic field organization

    Neuron 54:403–416.

    https://doi.org/10.1016/j.neuron.2007.03.029

    • Google Scholar
34. 1. Sterne GR
    2. Kim JH
    3. Ye B

    (2015) Dysregulated Dscam levels act through Abelson tyrosine kinase to enlarge presynaptic arbors

    eLife 4:e05196.

    https://doi.org/10.7554/eLife.05196

    • Google Scholar
35. 1. Thu CA
    2. Chen WV
    3. Rubinstein R
    4. Chevee M
    5. Wolcott HN
    6. Felsovalyi KO
    7. Tapia JC
    8. Shapiro L
    9. Honig B
    10. Maniatis T

    (2014) Single-cell identity generated by combinatorial homophilic interactions between α, β, and γ protocadherins

    Cell 158:1045–1059.

    https://doi.org/10.1016/j.cell.2014.07.012

    • Google Scholar
36. 1. Warming S
    2. Costantino N
    3. Court DL
    4. Jenkins NA
    5. Copeland NG

    (2005) Simple and highly efficient BAC recombineering using galK selection

    Nucleic Acids Research 33:e36.

    https://doi.org/10.1093/nar/gni035

    • Google Scholar
37. 1. Wojtowicz WM
    2. Flanagan JJ
    3. Millard SS
    4. Zipursky SL
    5. Clemens JC

    (2004) Alternative splicing of Drosophila Dscam generates axon guidance receptors that exhibit isoform-specific homophilic binding

    Cell 118:619–633.

    https://doi.org/10.1016/j.cell.2004.08.021

    • Google Scholar
38. 1. Wojtowicz WM
    2. Wu W
    3. Andre I
    4. Qian B
    5. Baker D
    6. Zipursky SL

    (2007) A vast repertoire of Dscam binding specificities arises from modular interactions of variable Ig domains

    Cell 130:1134–1145.

    https://doi.org/10.1016/j.cell.2007.08.026

    • Google Scholar
39. 1. Wu Q
    2. Maniatis T

    (1999) A striking organization of a large family of human neural cadherin-like cell adhesion genes

    Cell 97:779–790.

    https://doi.org/10.1016/S0092-8674(00)80789-8

    • Google Scholar
40. 1. Wu W
    2. Ahlsen G
    3. Baker D
    4. Shapiro L
    5. Zipursky SL

    (2012) Complementary chimeric isoforms reveal Dscam1 binding specificity in vivo

    Neuron 74:261–268.

    https://doi.org/10.1016/j.neuron.2012.02.029

    • Google Scholar
41. 1. Yamagata M
    2. Sanes JR

    (2008) Dscam and Sidekick proteins direct lamina-specific synaptic connections in vertebrate retina

    Nature 451:465–469.

    https://doi.org/10.1038/nature06469

    • Google Scholar
42. 1. Yamagata M
    2. Sanes JR

    (2010) Synaptic localization and function of Sidekick recognition molecules require MAGI scaffolding proteins

    Journal of Neuroscience 30:3579–3588.

    https://doi.org/10.1523/JNEUROSCI.6319-09.2010

    • Google Scholar
43. 1. Zhan XL
    2. Clemens JC
    3. Neves G
    4. Hattori D
    5. Flanagan JJ
    6. Hummel T
    7. Vasconcelos ML
    8. Chess A
    9. Zipursky SL

    (2004) Analysis of Dscam diversity in regulating axon guidance in Drosophila mushroom bodies

    Neuron 43:673–686.

    https://doi.org/10.1016/j.neuron.2004.07.020

    • Google Scholar
44. 1. Zipursky SL
    2. Grueber WB

    (2013) The molecular basis of self-avoidance

    Annual Review of Neuroscience 36:547–568.

    https://doi.org/10.1146/annurev-neuro-062111-150414

    • Google Scholar

Author details

1. Andrew M Garrett

   The Jackson Laboratory, Bar Harbor, United States

   Contribution

   AMG, Experiments were conceived, Executed all genetic analyses, collected and analyzed data on retinal phenotypes, wrote the manuscript with input from all authors

   Competing interests

   The authors declare that no competing interests exist.

2. Abigail LD Tadenev

   The Jackson Laboratory, Bar Harbor, United States

   Contribution

   ALDT, Acquisition of data, Drafting or revising the article, Performed immunoprecipitations and western blots

   Competing interests

   The authors declare that no competing interests exist.

3. Yuna T Hammond

   The Jackson Laboratory, Bar Harbor, United States

   Contribution

   YTH, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article, Identified PDZ interactions

   Competing interests

   The authors declare that no competing interests exist.

4. Peter G Fuerst

   1. Department of Biological Sciences, University of Idaho, Moscow, United States
   2. WWAMI Medical Education Program, University of Idaho, Moscow, United States

   Contribution

   PGF, Conception and design, Acquisition of data, Identified PDZ interactions, Generated targeting vectors for C-terminal deletion of DSCAM

   Competing interests

   The authors declare that no competing interests exist.

5. Robert W Burgess

   The Jackson Laboratory, Bar Harbor, United States

   Contribution

   RWB, Conception and design, Analysis and interpretation of data, Drafting or revising the article, Experiments were conceived, wrote the manuscript with input from all authors

   For correspondence

   robert.burgess@jax.org

   Competing interests

   The authors declare that no competing interests exist.

   "This ORCID iD identifies the author of this article:" 0000-0002-9229-3407

• 1,114

  views

• 203

  downloads

• 15

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.