Developmental Neuroscience · Unit 6: Axons find their way

Choosing a Partner: Layers, Self-Avoidance and Smell

A gradient can bring an axon to the right neighbourhood of its target, but it cannot tell the axon which layer to stop in, which cell to touch, or which of its own branches to stay away from. This lesson covers three systems that answer those questions with molecular labels: adhesion proteins that sort retinal arbors into layers, surface identities that make a neuron's branches avoid each other, and odorant receptors that help sort smell axons into glomeruli. Afterwards you can predict what happens when one of these labels is removed, swapped or given to the wrong cell.

Watch first One Match, Two Opposite Outcomes · 2:48

A single fly gene, Dscam1, can be spliced into 38,016 different proteins. A neuron uses them to recognise its own branches.

That number answers a question you have not asked yet. The previous lesson ended with axons arriving at roughly the right place in a target, sorted by gradients. Arriving is not the end of the trip. A retinal axon reaching the tectum, or a bipolar cell process reaching the inner retina, still has to stop in the right layer, find the right partner, and avoid tangling with itself. This lesson is about the labels that do that work.

Where in the target?

Targeting can be broken into four stages: reach the target region, slow down and branch, find the right place, and choose the right partner. Growth cones that enter a target slow down, pause and branch, and branches tend to form where the growth cone paused.

“How does a neuron know its position and know where to go?” There are three answers. Position can be “determined by individual labels”, “by gradients”, or “by neighbors”. The topographic maps lesson was about gradients. A gradient is good at continuous order: neighbouring points in the eye map to neighbouring points in the tectum. It is poor at discrete choices. A thin layer is either the right one or not, and a glomerulus either belongs to one receptor channel or does not. Discrete choices call for discrete labels.

Think of a conference with name tags. A gradient is like telling people to sit in rows by age. A label is like seating people at the table whose sign matches their tag.

The idea: layers are chosen by like binding like

The inner plexiform layer (IPL) of the retina is a thin band where bipolar cells, amacrine cells and retinal ganglion cells make synapses. It is divided into thinner sublaminae, and each cell type sends its processes into one or a few of them. A cell that ends in the wrong sublamina meets the wrong partners.

In chick retina, Masahito Yamagata, Joshua Weiner and Joshua Sanes found in 2002 that two related proteins, Sidekick-1 and Sidekick-2, are made by largely separate groups of retinal neurons, and that the processes of each group end in different sublaminae. Both are members of the immunoglobulin (Ig) superfamily, cell-surface proteins built from repeated Ig-like domains. Each binds its own kind, which is called homophilic binding. Cells carrying Sidekick-1 stick to other cells carrying Sidekick-1.

The key test was misexpression: making a protein in a cell that normally lacks it. When retinal neurons were made to produce a Sidekick they did not normally make, their processes shifted toward the sublamina where that Sidekick was already concentrated. In 2008 Yamagata and Sanes extended the set to four proteins, adding Dscam and DscamL, each marking a different group of cells and sublaminae. Reducing their levels disturbed the order. The short version: misexpressing one of these adhesion molecules shifts the sublamina an arbor occupies.

Keep the scope in view. The four-protein code was described in chick. Later work in mouse retina from the same laboratory tied Sidekick-2 to one particular circuit between an amacrine cell type and a ganglion cell type, but mouse sublaminae are not simply assigned by these four proteins. These molecules bias a choice; they are one part of how layers form.

Protein isoforms for neuronal individuality

The problem this heading names is the opposite of the layer problem. Instead of joining like with like, a neuron must keep like away from like, where “like” means itself.

Many neurons spread their dendrites so that they cover their territory evenly, without their own branches crossing or bundling. This is self-avoidance. It requires each branch to recognise sister branches from the same cell and move away from them, while still overlapping the branches of other neurons. To do that, each neuron needs a surface identity that differs from its neighbours’.

The fly found one in a single gene. In 2000 Dietmar Schmucker, James Clemens, Huidy Shu, Lawrence Zipursky and colleagues described Dscam in Drosophila (now called Dscam1). It has four clusters of alternative exons, with 12, 48, 33 and 2 members, 95 alternative exons in all. Each messenger RNA keeps one exon from each cluster:

12 × 48 × 33 = 19,008 outer domains; × 2 membrane-spanning choices = 38,016 isoforms

In 2004 Woj Wojtowicz, Clemens, Zipursky and colleagues showed that Dscam1 isoforms bind mainly to identical isoforms. Binding is homophilic, as with Sidekick, but here the outcome is repulsion: when two branches carrying the same isoforms touch, they pull apart. In 2007 three groups (Matthews, Grueber and colleagues; Hughes, Schmucker and colleagues; Soba, the Jans and colleagues) showed that fly sensory neurons lacking Dscam1 have dendrites that cross and bundle with their own branches.

Then came the population test. Daisuke Hattori, Ebru Demir and colleagues in 2007 made flies whose Dscam1 could produce only one isoform. Circuits in these flies were severely disorganised. The experiment also showed that neighbouring neurons must carry different isoforms, while which isoform any one neuron carries matters little: neighbours sharing one isoform cannot tell themselves from each other. In 2009 Hattori and colleagues found that thousands of isoforms are needed for normal wiring. Each neuron appears to express a small, partly random subset, so a large menu keeps accidental matches rare.

Watch a trap here. Vertebrate DSCAM and DSCAML1 are not spliced this way and cannot supply this diversity. In mouse, the job falls largely to the clustered protocadherins, about 58 genes in three clusters (alpha, beta and gamma), expressed in different combinations in different neurons. In 2012 Julie Lefebvre, Sanes and colleagues showed that starburst amacrine cells in mouse retina need the gamma protocadherins for their dendrites to avoid each other. When every starburst cell was made to express the same single isoform, each cell’s branches separated, but neighbouring cells also began to avoid each other. Same problem, different molecules.

Map of spatial organization, map of qualities: smell

Put two kinds of map side by side: a “Map of spatial organization” and a “Map of qualities”. The retina-to-tectum map preserves where things are. The olfactory map sorts what things are.

In 1991 Linda Buck and Richard Axel identified a large family of genes in rat that encode odorant receptors. They are G protein-coupled receptors, and in rodents the family runs to about a thousand genes. How binding an odour leads to an electrical signal is covered in the senses lesson. Here the question is wiring.

Each olfactory sensory neuron in the nose expresses one odorant receptor gene, with rare exceptions. Neurons expressing a given receptor are scattered within a zone of the epithelium, yet in situ hybridisation by Vassar, Axel and colleagues and by Ressler, Sullivan and Buck in 1994 showed that their axons gather in a few glomeruli in the olfactory bulb. A glomerulus is a ball of neuropil where many sensory axons meet the dendrites of bulb neurons. In 1996 Peter Mombaerts and colleagues in Axel’s laboratory engineered mice in which neurons using the P2 receptor promoter also made a marker that fills the axon. The P2 axons converged on a pair of glomeruli in each bulb, one on its medial side and one on its lateral side. The rule is one receptor, one neuron, and one receptor channel per glomerulus.

Something surprising sits underneath. The receptor protein is found on the axon terminals as well as the dendrites, and it helps choose the target. Receptor-made cAMP is also thought to set how much of some guidance receptors the axon makes (see the cAMP lesson), though that link has been worked out in only a few receptor lines.

Design problem

You replace the coding sequence of the P2 odorant receptor in mouse with a marker, so neurons that use the P2 promoter make the marker but no P2 receptor. Predict where their axons go. Then predict what happens if, instead, you put a different receptor’s coding sequence in place of P2’s.

One way to do it

Start from what each experiment keeps and what it removes. Both keep the P2 promoter, so the same neurons are labelled, in the same zone of the nose. Only the receptor protein changes.

If the promoter or the neuron’s position decided the target, the receptor-less axons would still converge on the P2 glomerulus. They do not. In the deletion experiments reported by Feng Wang, Axel and colleagues in 1998, the labelled axons failed to form a P2 glomerulus; they wandered instead of converging on a specific glomerulus. So the receptor is needed.

If the receptor were the whole address, a swapped receptor would send the axons to the donor receptor’s own glomerulus. That is not quite what happened. In the 1996 swap, with M12 in place of P2, the axons converged near the P2 glomerulus and far from M12’s own. In a 1998 swap, with P3 in place of P2, they converged on a glomerulus touching the P3 glomerulus. So the receptor moved the axons, but not to a fixed donor address. So the receptor helps choose, and other features of the neuron also count. The receptor is part of the label, not all of it.

Labels and gradients together

Put the cases side by side. Gradients give continuous position across a sheet. Labels give discrete membership: a layer, a glomerulus, or self versus other. The same chemistry can serve opposite ends. Homophilic binding of Sidekick draws matching processes together, while homophilic binding of Dscam1 drives matching branches apart. What the cell does after the bond forms decides which.

None of this is final. Target selection ends with “Refinement later in life”: maps and connections set up by these molecules are adjusted by activity, a subject for the lessons on spontaneous activity and synapses.

How we measure it

Gene-targeted knock-in reporters

A marker such as tau-lacZ is placed into a gene in mouse so that every neuron using that gene's promoter also makes the marker, which fills the axon. Staining then shows where those axons end. Replacing the gene's coding sequence at the same time shows what the protein itself contributes.

Misexpression and knockdown in developing retina

An adhesion protein is made in cells that normally lack it, or reduced in cells that normally have it, while the cells are still growing. Comparing where their processes end in the inner plexiform layer shows whether the protein biases sublaminar choice.

Single-neuron labelling

Only one or a few neurons in a field are made to express a fluorescent protein, so that one cell's whole arbor can be traced. This is how self-avoidance is scored: you count how often a cell's own branches cross or bundle, separately from its overlap with neighbours.

Homophilic binding assays

Purified outer domains or whole cells carrying one isoform are tested for binding to the same or a different isoform, by aggregation or by binding of soluble protein. A grid of such tests shows how strictly each form binds only its own kind.

Finish line

Test yourself

8 questions

Practice and MCAT-style, with explained answers.

Next · Lesson 28

Too Many Neurons: Naturally Occurring Cell Death

A chick embryo makes about 20,000 motor neurons for its legs, and about 40% of them die within a week. Why?

Continue →

← Lesson 26 · Topographic Maps: Sperry's Eye and Eph Gradients

Fly Dscam1, Ig1 to Ig4 — one isoform paired with itself
Fly Dscam1, Ig1 to Ig4 — one isoform paired with itself ● LIVE · RCSB PDB 2V5M
open on RCSB PDB ↗
Loading live structure 2V5M from RCSB PDB…

Two copies of the same isoform lie face to face, the variable surfaces of Ig2 and Ig3 meeting their twins on the partner. Change the residues at those surfaces and the pairing fails, which is why one isoform mostly binds only itself.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Human odorant receptor OR51E2 with propionate — an odorant receptor in atomic detail
Human odorant receptor OR51E2 with propionate — an odorant receptor in atomic detail ● LIVE · RCSB PDB 8F76
open on RCSB PDB ↗
Loading live structure 8F76 from RCSB PDB…

Seven helices cross the membrane, with the small odorant held in a pocket among them. On the inner face the receptor holds a G protein, which in olfactory neurons leads to cAMP.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Propionic acid — a small odorant that OR51E2 detects
Propionic acid — a small odorant that OR51E2 detects ● LIVE · PubChem CID 1032
open on PubChem ↗
Loading live structure 1032 from PubChem…

A three-carbon acid with a sharp, sour smell. Its small size shows how little an odorant needs to be to fit a receptor's pocket.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Joshua R. Sanes

Joshua R. Sanes

With Masahito Yamagata, showed in chick retina that four related adhesion proteins, Sidekick-1, Sidekick-2, Dscam and DscamL, mark separate sets of neurons and separate sublaminae, and that changing their expression shifts where arbors end; later, with Julie Lefebvre and colleagues, showed that mouse clustered protocadherins are needed for dendritic self-avoidance.

The inner plexiform layer of the retina is a thin band where bipolar, amacrine and ganglion cells meet, and it is divided into sublaminae, each holding the synapses of a few cell types. Sanes and Masahito Yamagata asked whether molecules on the cells could explain that order. In 2002, with Joshua Weiner, they reported that two related adhesion proteins, Sidekick-1 and Sidekick-2, are made by largely separate groups of chick retinal neurons whose processes end in different sublaminae. Each protein binds its own kind. When they made a cell produce a Sidekick it normally lacked, its processes moved toward the sublamina rich in that protein. In 2008 Yamagata and Sanes added Dscam and DscamL, giving four proteins for four sets of sublaminae, and showed that reducing them disturbed the order. In 2012, with Julie Lefebvre, Dimitar Kostadinov, Weisheng Chen and Tom Maniatis, the laboratory turned to a second question in mouse retina: how a neuron keeps its own dendrites apart. The work made it possible to treat a layer as an address written in adhesion molecules.

✓ verbatim Yamagata, Weiner & Sanes, Cell 110:649 (2002); Yamagata & Sanes, Nature 451:465 (2008); Lefebvre et al., Nature 488:517 (2012).
S. Lawrence Zipursky

S. Lawrence Zipursky

His laboratory found that fly Dscam can be spliced into 38,016 forms, showed with James Clemens and Woj Wojtowicz that the forms bind mainly to identical forms, and with Daisuke Hattori and colleagues showed that thousands of forms are needed for neurons to tell their own branches from those of other cells.

In 2000 Dietmar Schmucker, James Clemens, Huidy Shu and colleagues in Zipursky's laboratory reported a fly protein that binds the adaptor Dock and acts in axon guidance. It resembled a human protein, DSCAM, named for the region of chromosome 21 linked to Down syndrome. The fly gene was strange. Four of its exons came as clusters of alternatives, 12, 48, 33 and 2, and each messenger RNA kept one from each cluster. Multiplied out, that is 38,016 possible proteins. In 2004 Woj Wojtowicz, Clemens and colleagues showed that the outer parts of these isoforms bind mainly to identical isoforms. In 2007 Daisuke Hattori, Ebru Demir and colleagues, with Barry Dickson, replaced the variable region with a single choice. Circuits in these flies were badly disorganised, and the work showed that neighbouring neurons must carry different isoforms. In 2009 Hattori and colleagues showed that the fly needs thousands of isoforms to wire normally. The work explained what so much splicing was for: telling self from other.

✓ verbatim Schmucker et al., Cell 101:671 (2000); Wojtowicz et al., Cell 118:619 (2004); Hattori et al., Nature 449:223 (2007) and Nature 461:644 (2009).
Linda B. Buck

Linda B. Buck

With Richard Axel, identified a large family of genes for G protein-coupled receptors expressed in the rat olfactory epithelium, the odorant receptors; her own laboratory then mapped how neurons expressing each receptor are arranged in the nose and the bulb.

In the late 1980s Buck was a postdoctoral fellow in Richard Axel's laboratory at Columbia University, looking for the receptors that detect odours. She set three expectations for them, as her Nobel lecture describes: they should probably be G protein-coupled receptors, since odorants were known to act through G proteins in olfactory neurons; they should belong to a large family of related genes, to cover the many odours an animal can tell apart; and they should be expressed in the olfactory epithelium. She used the polymerase chain reaction with primers matched to conserved stretches of known G protein-coupled receptors, then kept only the products that were made in olfactory tissue and that formed a family. The 1991 paper with Axel described that family in rat and found it extremely large, with more than a hundred members; later work put it near a thousand genes in mouse. In 1994 her own laboratory, with Kerry Ressler and Susan Sullivan, reported that neurons expressing one receptor send axons to a few glomeruli in the bulb. The receptors gave researchers a way to label each smell channel.

✓ verbatim Buck & Axel, Cell 65:175 (1991); Ressler, Sullivan & Buck, Cell 79:1245 (1994); Buck, Nobel Lecture, 8 December 2004.
Richard Axel

Richard Axel

Shared the discovery of the odorant receptor family with Linda Buck, and his laboratory used gene targeting in mouse to label the axons of neurons that express one receptor and to test what happens when that receptor is deleted or exchanged.

Once the receptor genes were known, Axel's laboratory at Columbia University could ask where each kind of neuron sends its axon. In 1994 Robert Vassar and colleagues in the laboratory used in situ hybridisation, a method that detects one messenger RNA in tissue, and found that axons from neurons expressing a given receptor, which are scattered within a zone of the nose, gather in a few glomeruli of the olfactory bulb. In 1996 Peter Mombaerts and colleagues engineered mice in which neurons that make the P2 receptor also make a blue- staining marker that fills the axon, and watched those axons converge. The same paper swapped in a second receptor, M12, and the axons converged near the P2 glomerulus, far from the M12 one. A 1998 study led by Feng Wang deleted P2, and the axons wandered instead of converging; with P3 in its place they went to a glomerulus touching the P3 glomerulus. Buck and Axel shared the 2004 Nobel Prize in Physiology or Medicine. The experiments made the receptor a known part of the wiring instructions.

✓ verbatim Mombaerts et al., Cell 87:675 (1996); Wang, Nemes, Mendelsohn & Axel, Cell 93:47 (1998); Nobel Prize in Physiology or Medicine 2004.
Test yourself
Question 1 of 8 Score 0 / 0

In the fly, where does the large variety of Dscam1 proteins come from?
Sources & Integrity

Every structure, slide, simulation, and clip on this page is listed below with the source it came from. None of the 3D structures is drawn by hand or by AI: each one loads by accession ID from RCSB PDB or PubChem. This panel is generated from the lesson's own data file, so what it lists is what the page loads.

Molecular structures live data

  • Fly Dscam1, Ig1 to Ig4 — one isoform paired with itself — RCSB PDB 2V5M · coordinates CC0 · Drosophila Dscam1 N-terminal four Ig domains, X-ray crystal structure (Meijers et al., Nature, 2007)
  • Human odorant receptor OR51E2 with propionate — an odorant receptor in atomic detail — RCSB PDB 8F76 · coordinates CC0 · Human OR51E2 bound to propionate and coupled to a G protein, cryo-EM structure (Billesbølle et al., Nature, 2023)
  • Propionic acid — a small odorant that OR51E2 detects — PubChem CID 1032 · PubChem compound record for propionic acid

Scientist profiles & quotes

  • Joshua R. Sanes — Yamagata, Weiner & Sanes, Cell 110:649 (2002); Yamagata & Sanes, Nature 451:465 (2008); Lefebvre et al., Nature 488:517 (2012).
  • S. Lawrence Zipursky — Schmucker et al., Cell 101:671 (2000); Wojtowicz et al., Cell 118:619 (2004); Hattori et al., Nature 449:223 (2007) and Nature 461:644 (2009).
  • Linda B. Buck — Buck & Axel, Cell 65:175 (1991); Ressler, Sullivan & Buck, Cell 79:1245 (1994); Buck, Nobel Lecture, 8 December 2004.
  • Richard Axel — Mombaerts et al., Cell 87:675 (1996); Wang, Nemes, Mendelsohn & Axel, Cell 93:47 (1998); Nobel Prize in Physiology or Medicine 2004.

Quotes are verbatim from the cited source. Profile text is original to this site and was checked against the sources cited.

Animated explainer

One Match, Two Opposite Outcomes · narration: an AI narrator voice (Gemini, Charon) reading narration written from this lesson's text · animation: AI-generated in code (Manim) by Claude, a schematic with no drawn molecular structures, plots computed from their equations · captions from the narration script.

Images & figures

This lesson carries no raster figures. Every structure is rendered live from RCSB PDB or PubChem, and every plot is drawn from its equation in the browser.

integrity Machines check the mechanics. The build confirms that every local file exists, that every question's answer key points at a real choice, and that any AI-generated scientific figure names the person who checked it; a weekly job re-resolves every accession ID against its database. Questions are original; MCAT-style items are written for this site and are not real AAMC questions. Banned sources (the Tymoczko textbook, publisher decks, copyrighted figures) are not used.

Like I'm five

A nerve cell growing into the brain has to find the right street, then the right house, then the right person. Some cells wear name tags that stick only to the same name, so they end up beside friends with matching tags. A cell also has to keep its own arms from tangling, so each cell wears its own special badge and its arms push away from anything with that badge. In the nose, each smell cell carries one smell detector, and that detector also helps it find its meeting place.
High school

After an axon reaches the right region of a target, it still has to choose a layer and a partner. In the chick retina, cells that should connect make the same adhesion protein, a protein that sticks to copies of itself, so they gather in the same thin layer. Neurons also spread their branches out evenly by making their own branches repel each other. In flies a single gene, Dscam1, can be cut and joined into tens of thousands of versions, so each neuron has its own version and recognises its own branches. Smell works with a different label: each smell neuron makes one odorant receptor, and all neurons with the same receptor send their axons to the same one or few spots in the olfactory bulb.
College

Targeting proceeds in stages: reach the region, slow and branch, find the place, choose the partner. Gradients such as Eph and ephrin set position within a continuous map, but laminae and glomeruli are discrete and need discrete labels. In chick retina the immunoglobulin superfamily proteins Sidekick-1, Sidekick-2, Dscam and DscamL are expressed by largely separate neuron subsets, bind homophilically, and bias arbors to particular sublaminae of the inner plexiform layer. Self-avoidance uses homophilic recognition with a repulsive outcome. Drosophila Dscam1 has four mutually exclusive exon clusters (12, 48, 33 and 2 alternatives), giving 19,008 ectodomains and 38,016 isoforms; each neuron expresses a probabilistic subset, and isoform-specific binding between sister branches triggers their separation. In mouse, clustered protocadherins serve the same function. Odorant receptors are G protein-coupled receptors; each olfactory sensory neuron expresses one, and receptor deletion and swap experiments show that the receptor protein participates in choosing the glomerulus.
Graduate

Dscam1 isoforms bind through matching variable surfaces in Ig2, Ig3 and Ig7, and binding grids found that most variable domains bind preferentially to themselves. Single-isoform alleles show that one isoform suffices for self-recognition within an isolated neuron, while repertoire-reduction alleles show that population diversity in the thousands is needed so that neighbours rarely share isoforms by chance. In mouse, roughly 58 clustered protocadherin genes in the alpha, beta and gamma clusters are expressed in stochastic combinations, and trans recognition is thought to require a match across the whole cis assembly, which multiplies the number of distinguishable identities. In olfaction, receptor-derived cAMP levels are thought to set anterior-posterior position through guidance receptors such as neuropilin-1, while activity-dependent adhesion molecules sort axons into separate glomeruli. Open question: what intracellular pathway turns a homophilic binding event at the membrane, normally an adhesive act, into retraction of the touching branches?
Ninety-five exons for one job

The fly Dscam gene carries 95 alternative exons in four clusters: 12, 48, 33 and 2. A messenger RNA keeps one exon from each cluster, so the gene can in principle encode 38,016 proteins, more than the number of genes in the fly genome. When the paper appeared, a reasonable guess was that each isoform might guide a different axon to a different place. The later work pointed elsewhere. The isoforms seem to matter mostly for being different from each other, so that any two neurons are unlikely to match.

— Schmucker et al., Cell 101:671 (2000).
Three expectations for a smell receptor

Buck did not search for an odorant receptor by purifying one. She decided what such a receptor should look like. It should be a G protein-coupled receptor, because odorants were known to act through G proteins in olfactory neurons. It should belong to a big family of related genes. And it should be made in the olfactory epithelium. Amplifying DNA with primers aimed at shared stretches of known receptors and then filtering by those expectations turned up the family. It proved to be among the largest gene families in mammals.

— Buck, Nobel Lecture, 8 December 2004; Buck & Axel, Cell 65:175 (1991).

 ↑ Top