The Crystal Structure of Netrin-1 in Complex with DCC Reveals the Bifunctionality of Netrin-1 As a Guidance Cue

Please cite this article in press as: Finci et al., The Crystal Structure of Netrin-1 in Complex with DCC Reveals the Bifunctionality of Netrin-1 As a Guid-ance Cue, Neuron (2014), http://dx.doi.org/10.1016/j.neuron.2014.07.010

Neuron

Article

The Crystal Structure of Netrin-1 in Complexwith DCC Reveals the Bifunctionalityof Netrin-1 As a Guidance CueLorenzo I. Finci,1,2,5 Nina Kruger,3,5 Xiaqin Sun,1,5 Jie Zhang,1 Magda Chegkazi,3 Yu Wu,1 Gundolf Schenk,3

Haydyn D.T. Mertens,3 Dmitri I. Svergun,3 Yan Zhang,1,4,* Jia-huai Wang,1,2,* and Rob Meijers3,*1State Key Laboratory of Biomembrane and Membrane Biotechnology, College of Life Sciences, Peking University, Beijing 100871, China2Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA 02215, USA3European Molecular Biology Laboratory (EMBL), Hamburg Outstation, Notkestrasse 85, 22607 Hamburg, Germany4PKU-IDG/McGovern Institute for Brain Research, Peking University, Beijing 100871, China5Co-first authors

*Correspondence: [email protected] (Y.Z.), [email protected] (J.-h.W.), [email protected] (R.M.)http://dx.doi.org/10.1016/j.neuron.2014.07.010

SUMMARY

Netrin-1 is a guidance cue that can trigger eitherattraction or repulsion effects on migrating axons ofneurons, depending on the repertoire of receptorsavailable on the growth cone. How a single chemo-tropic molecule can act in such contradictory wayshas long been a puzzle at the molecular level. Herewepresent thecrystal structureof netrin-1 in complexwith the Deleted in Colorectal Cancer (DCC) receptor.We show that one netrin-1 molecule can simulta-neously bind to two DCC molecules through a DCC-specific site and through a unique generic receptorbinding site, where sulfate ions staple together posi-tively charged patches on both DCC and netrin-1.Furthermore, we demonstrate that UNC5A canreplace DCC on the generic receptor binding site toswitch the response from attraction to repulsion. Wepropose that the modularity of binding allows for theassociation of other netrin receptors at the genericbinding site, eliciting alternative turning responses.

INTRODUCTION

Axon guidance is mediated by a combination of attractive and

repulsive guidance cues that interact with their corresponding re-

ceptors gathered on the surface of the growth cone of an axon

(Kolodkin and Tessier-Lavigne, 2011). Diffusible chemoattrac-

tants attract axons to their targets, whereas repulsive guidance

cues generate exclusion zones that axons avoid. This strategic

process dictates a defined trajectory that the axon will navigate

among many possible routes for precise neuronal wiring (Dick-

son, 2002). Netrins were the first axon guidance cue family mem-

bers identified in both invertebrates and vertebrates. They consist

of a laminin VI domain, a V domain containing three EGF repeats,

and aC-terminal netrin-like domain (Kennedy et al., 1994; Serafini

et al., 1994). Netrin-1 can act either as an attractive or a repulsive

cue. The bifunctionality of netrin-1 has been linked to the reper-

toire of receptors presented on the growth cone (Hong et al.,

1999). A receptor termed Deleted in Colorectal Cancer (DCC)

constitutively expresses on the axonal surface. Netrin-1 binding

to DCC induces chemoattraction (Keino-Masu et al., 1996), but

this response is turned into repulsion if another receptor, uncoor-

dinated-5 (UNC5) coexists with DCC (Colamarino and Tessier-

Lavigne, 1995). Other known netrin-1 receptors that assist in

axon guidance are DSCAM (Ly et al., 2008; Andrews et al.,

2008; Liu et al., 2009) and Neogenin (Wilson and Key, 2006).

More receptors are expected to be involved including mem-

brane-bound proteoglycans (Rajasekharan and Kennedy, 2009).

DCC belongs to the immunoglobulin superfamily. The protein

consists of four N-terminal Ig-like domains, in a horseshoe-

shaped conformation (Chen et al., 2013), followed by six fibro-

nectin type III domains (FN), a single transmembrane segment,

and a large cytosolic portion that contains three highly conserved

sequence motifs termed P1, P2, and P3 (Keino-Masu et al.,

1996). DCCwas initially discovered as aprognostic tumormarker

(Fearon et al., 1990) whose absence on colon cells implied a pro-

gressed and malignant state in colon cancer. Recently, the dele-

tion of DCC has been linked to the propensity of cancerous cells

to metastasize (Krimpenfort et al., 2012), and it has been pro-

posed that the deletion of DCCmakes the tumor cells insensitive

to a netrin-1-dependent apoptotic pathway (Mazelin et al., 2004).

According to this theory, netrin-1 acts as a tropic factor, andDCC

is a dependence receptor that depends on netrin-1 to maintain a

netrin/DCC signaling complex (Castets et al., 2012). In the

absence of netrin-1, the signaling complex disintegrates, leading

to apoptosis. Tumor cells switch off this apoptotic pathway by

deleting DCC. The plasticity of the netrin/DCC complex also

seems to play a key role in the regulation of angiogenesis (Wilson

et al., 2006) and tissue development (Lai Wing Sun et al., 2011).

Taken together, these data demonstrate that netrins exert very

complex biological functions in addition to axon guidance.

The formation of a netrin-1/DCC signaling complex is initiated

when netrin-1 associates with the extracellular portion of DCC,

resulting in the homodimerization of DCC via its cytoplasmic

P3motif (Stein et al., 2001). This recruits an intracellular signaling

complex that leads to the release of calcium (Hong et al., 2000),

kinase activation (Liu et al., 2004), and a rearrangement of the

cytoskeleton (Li et al., 2008). In the presence of the chemorepel-

lent UNC5, a ternary complex is formed between netrin-1 and the

Neuron 83, 1–11, August 20, 2014 ª2014 Elsevier Inc. 1

mailto:[email protected]



http://dx.doi.org/10.1016/j.neuron.2014.07.010

Figure 1. Domain Architecture and Crystal Structure of the Netrin-1/

DCC Complex

(A) Domain diagram of DCC and netrin-1, with the domains present in the

crystal structure colored red.

(B) Ribbon diagram of the NetrinVIV/DCCFN56 complex is depicted showing two

DCC fragments (in green) bound to the V domain (in salmon red) of one netrin-1

molecule. Glycosylation sites on the VI domain of netrin-1 (in orange) are

shown as sticks.

Neuron

Molecular Basis for the Netrin-1 Receptor Switch


extra-cellular portions of UNC5 and DCC (Hong et al., 1999). The

heterodimerization, also transpiring through a cytoplasmic inter-

action, is between the P1 motif of DCC and the DCC-binding

motif of UNC5 and gives rise to an alternate signaling complex.

Therefore, netrin-1 appears to present a scaffold both for sym-

metric clustering of a singular receptor and asymmetric clus-

tering of a pair of different receptors with remarkable versatility.

Despite its key role in so many cellular processes, the structural

basis of netrin-1-induced clustering of its receptors is not known.

Here, we present the crystal structure of a human netrin-1/

DCC complex, revealing the basic architecture of the netrin/

receptor scaffold. The structure exhibits a unique and highly

conserved anion-dependent receptor binding site, next to a

DCC-specific receptor binding site. Our data provide insight

into the assembly of the netrin/DCC signaling complex in solu-

tion and in vivo and suggest how the environmentally tuned

selection of alternative receptor pairs can dynamically alter ne-

trin-induced cellular responses.

RESULTS

Structure of the NetrinVIV/DCCFN56 Complex RevealsTwo Receptor Binding SitesHuman netrin-1 is a soluble, glycosylated protein that is secreted

by specialized cells and can affect cell signaling on a single

molecule level (Pinato et al., 2012). To ensure production of

2 Neuron 83, 1–11, August 20, 2014 ª2014 Elsevier Inc.

active netrin-1 in large quantities, we replaced the native secre-

tion signal sequence with the pregnancy-specific glycoprotein-1

secretion signal and incorporated codon-optimized constructs

into the pXLG vector for transient expression in HEK293T cells

(Backliwal et al., 2008). We tested several constructs covering

the laminin-like (VI) domain alone and in combination with one

or all three of the EGF repeats of the V domain, and we found

that a construct containing the full VI and V domains immediately

followed by a His tag could be purified in milligram quantities

(Figure 1A). Moreover, a similar construct that contains an Fc

tag was shown previously to be sufficient to act as a chemoat-

tractant for DCC-mediated axon guidance (Keino-Masu et al.,

1996). Further optimization using thermal shift assays (Boivin

et al., 2013) revealed that the presence of sulfate ions and

MES buffer at pH 6 substantially stabilized the recombinant

NetrinVIV protein.

The netrin-1 binding site on DCC has been delineated to the

membrane-proximal portion of the receptor (Figure 1A), encom-

passing a region in the FN domains (Geisbrecht et al., 2003; Mille

et al., 2009). Several constructs covering domains FN4, FN5, and

FN6 were expressed in E. coli for structural studies, since these

domains are not glycosylated and do not contain disulphide

bonds. Eventually, diffraction quality cocrystals were obtained

with an FN5-FN6 fragment (DCCFN56) in complex with NetrinVIVby mixing the two components in a 1:1 stoichiometry in the crys-

tallization drop.

The crystal structure of the NetrinVIV/DCCFN56 complex was

determined to 3.1 A resolution by molecular replacement using

laminin gamma1 LN-LE1-2 (Carafoli et al., 2012) as a search

model (Table S1 available online). The VI domain of NetrinVIVhas a typical laminin fold that is identical to laminin gamma1

and netrinG1 and netrinG2 (Brasch et al., 2011; Seiradake

et al., 2011), except for some variations in the loops connecting

the b strands, and it includes three glycosylation sites aswell as a

calcium-binding site. The V region consists of three EGF do-

mains (EGF-1, EGF-2, and EGF-3). The EGF-2 domain contains

an extended a helix that is not present in the other EGF domains.

The FN domains of DCC adopt a common b sandwich structure

with the ABE strands on one side and the C0CFG strands on the

other.

The predominant feature of the NetrinVIV/DCCFN56 complex

structure is that, despite the one-to-one molecular ratio used in

cocrystallization, each NetrinVIV molecule associates with two

DCCFN56 molecules (Figure 1B). The closest distance between

the two bound DCC molecules is approximately 20 A, excluding

any direct contacts between the DCC molecules in the netrin-1

binding region. The two DCCFN56 molecules sit on the same

side of the V domain of netrin-1, tilted at a 90� angle when viewed

along the main axis of the V domain of NetrinVIV. One DCCFN56

receptor fragment sits at the very tip of the EGF-3 domain of

NetrinVIV, designated as binding site 1. The other DCCFN56 re-

ceptor fragment sits at the center of the NetrinVIV molecule, inter-

acting with the EGF-1 and EGF-2 domains in what is defined as

binding site 2. There is no direct contact between the laminin-like

VI domain of netrin-1 and the DCCFN56 molecules, and the gly-

cosylation sites on the VI domain do not point at the receptor

fragments either, so they are unlikely to directly affect DCCFN56

binding.

Neuron



The FN5 domain of DCC plays a fundamental role by using

different regions on the domain to interact with the two distinct

netrin binding sites. Moreover, the C termini of the two DCC frag-

ments point in the same direction toward the plasma membrane

(Figure 1B). The side-by-side dockingmode of the twoDCCmol-

ecules on one netrin-1 molecule both using the FN5 domain is

ideal to cluster two receptors in close proximity to the mem-

brane. The FN6 domain abuts FN5 to form a rigid, rod-like struc-

ture, and the N-terminal end of FN6 is also indispensable for

binding site 2. Following FN6 is a flexible 55-residue linker that

can adjust the orientation of the two DCC molecules for binding.

Each DCC molecule can associate with two netrin molecules

without sterical clashes.

A DCC-Specific Receptor Binding Site on Netrin-1The DCCFN56 molecule at binding site 1 exclusively uses the

FN5 domain for interactions. The buried surface area for binding

site 1 covers 1220 A2. It involves the EGF-3 domain of netrin-1

and the edge of the b sandwich of the FN5 domain of DCCFN56

(Figure 1B). The interface manifests the classical features of a

hotspot (Clackson and Wells, 1995) centered at residue

Met933 and the nearby Val848 on DCC (Figure 2A). Met933 is

at the start of the parallel G strand on the GFC b sheet, whereas

Val848 is located just before the beginning of the A strand on

the ABD b sheet. These two residues bulge out of the side of

the FN5 domain, poking into the EGF-3 domain of netrin-1.

The key residue to facilitate the formation of the hotspot on

the netrin-1 side is Gln443, which stacks against the side chain

of Met933 of DCC. The hydrophilic tip of the side chain of

Gln443 forms hydrogen bonds with the main chain of Met933

and Val850 of DCC. This stretches the side chain of Gln443

and orients its aliphatic portion to make hydrophobic interac-

tions with the Met933 side chain, which is the central feature

of binding site 1 (Figure 2B). In addition, there are more hydro-

phobic contacts around, involving netrin-1 residues Val409,

Val429, Pro447, and Pro450. This hydrophobic hotspot is sur-

rounded by a network of directed side chain hydrogen bonds

from three residues (Gln851, Asn867, and Thr934) of the

DCC and two residues (His407 and Gln442) of netrin-1, ensuring

specificity. There are three sulfate ions at the periphery of site 1,

but only one makes a bridge between residues of NetrinVIV and

DCCFN56. The other two sulfate ions engage positively charged

residues on NetrinVIV, potentially to open up the binding site

for DCC.

To test these structural observations, we performedmutagen-

esis on residues that contribute to the hotspot to see whether

this would affect binding of netrin-1 to COS-7 cells expressing

DCC (Figures 2C and S1). On the DCC side, mutants Val848Arg

or Met933Arg did abolish binding, but mutants Val848Ala and

Met933Ala only caused a minor reduction in netrin-1 binding.

Apparently, it is necessary to introduce a bulky and charged res-

idue at the center of the hotspot to disrupt the interaction, indi-

cating that the hotspot consists of several components that

contribute to the binding. By contrast, a mutation of Gln443Ala

on NetrinVIV is sufficient to abolish binding completely, indi-

cating that this residue makes a major contribution to the hot-

spot. To further investigate the coexistence of the two binding

sites, we examined whether full-length DCC undergoes multi-

merization when exposed to netrin-1 (Figure 2D). COS-7 cells

were cotransfected with DCC constructs containing an HA

and a His tag, respectively. Wild-type DCC tagged with an HA

epitope can coimmunoprecipitate His-tagged DCC in the pres-

ence of netrin-1 or NetrinVIV and vice versa. When we carried out

this assay for the hotspot mutants, the multimerization of the re-

ceptors is abolished even for the DCC mutants Val848Ala and

M933Ala that did retain some netrin-1 binding in the cellular

assay. Finally, we employed axon guidance assays to test

whether the hotspot mutants could recover chemoattraction of

netrin-1-coated beads by using neurons harvested from the

dorsal horn of the spinal cord of DCC�/� E15 mice that were

injected with vector DNA containing either wild-type or muta-

ted full-length DCC (Figures 2E and S2). The DCC mutants

Val848Ala andMet933Ala behave similar to wild-type DCC, sug-

gesting that the weakening of the hotspot by introducing a

single ‘‘mild’’ mutation is not sufficient to disrupt its function.

However, the more drastic mutants Val848Arg and Met933Arg

show a severe reduction in chemoattraction, as does the Netrin-

VIV mutant Gln443Ala. The effect of mutants on chemoattraction

correlates well with the effect seen in the cell binding assays,

showing that binding to site 1 is crucial for the netrin-1/DCC

signaling complex.

Evolutionary Conservation of Binding Site 1Since the netrin/DCC signaling complex is conserved among the

animal kingdom, it is interesting to verify the conservation of the

site 1 interface among species. The tip of the EGF-3 domain of

netrin-1 is almost entirely conserved from Caenorhabditis ele-

gans to Homo sapiens, including all residues involved in shaping

the binding pocket and providing hydrogen bond partners, as

well as the crucial residue Gln443 (Figure S3). The sequence

conservation for the FN5 domain of DCC is much lower (20%

sequence identity if Homo sapiens, Mus musculus, Xenopus

laevis,Danio rerio,Drosophila melanogaster, andCaenorhabditis

elegans are compared), whichmakes it more remarkable that the

two hydrophobic residues Val848 and Met933 at the core of the

hotspot are conserved among these species, whereas the resi-

dues providing hydrogen bond partners are at most conserva-

tively substituted (Figures S4A and S4B). The conservation of

the most important residues involved in binding site 1 on ne-

trin-1 and DCC suggests that this site is biologically important

among all species.

Since the discovery of netrin-1, several other soluble netrins

have been identified with overlapping functions and receptor

repertoires (Lai Wing Sun et al., 2011). We compared the

amino acid sequences of these netrins to see whether binding

site 1 is present, since it is known, for instance, for netrin-3 that

it also binds DCC (Wang et al., 1999). We find that the residues

at the tip of the EGF-3 domain involved in DCC-specific bind-

ing are conserved for human and murine netrin-3 and netrin-5

(Figure S5). Netrin-4 is structurally diverse from the other ne-

trins discovered so far (Koch et al., 2000), and direct interac-

tion with DCC has yet to be established. The most prominent

difference between netrin-1 and both netrin-3 and netrin-5 is

the probable removal of one peripheral sulfate ion binding

site, because residue Lys439 on netrin-1 is replaced by a

proline.


Figure 2. The DCC-Specific Binding Site on Netrin-1

(A) Close-up view of binding site 1 with DCCFN56 represented in green as a ribbon and the EGF-3 domain of NetrinVIV in salmon represented as a surface.

(B) Close-up view of the hotspot interface on site 1 with DCC in green and NetrinVIV colored salmon, with several tentative hydrogen bonds drawn as dotted lines.

(C) Cell binding assays for full-length wild-type (wt) DCC and mutants Val848, Met933, and the NetrinVIV Gln443 mutant showing the percentage of DCC

presenting COS-7 cells that bind netrin-1 or NetrinVIV. Data represent mean ±SE (n = 100 for each group). Two-way ANOVA, followed by post hoc Scheffe’s

test. **: p<0.01 compared with WT.

(D) DCC multimerization assays for DCC Val848 and Met933 mutants showing enrichment of oppositely tagged receptors.

(E) Axon guidance assays for DCC Val848, Met933, and NetrinVIV Gln443 mutants showing percentage of beads attracting axons. Data represent mean ±SE (n =

300 for each group). Two-way ANOVA, followed by post hoc Scheffe’s test. **: p<0.01 compared with WT.

Neuron



A Unique Generic Netrin-1/Receptor Binding Site IsMediated by Structurally Clustered AnionsThe DCCFN56 receptor fragment situated at binding site 2 buries

a surface area of 950 A2. It involves portions of both the EGF-1

and EGF-2 domains of NetrinVIV and a very different region of

FN5 than for binding site 1. It uses the face of the ABE b sheet

of the FN5 domain of DCCFN56, as well as the very tip of the A


and F strand of FN6. There is a high accumulation of positively

charged residues in binding site 2 on both netrin-1 and DCC.

The positively charged patches are bridged and neutralized by

a cluster of anions at the interface (Figure 3A). Four sulfate ions

cluster near the extended helix on the EGF-2 domain of NetrinVIV(Figure 3B). These sulfate ions are coordinated by five arginine

residues (Arg317, Arg348, Arg349, Arg351, and Arg372) from

Figure 3. Detailed Molecular View of Generic Receptor Binding

Site 2 on Netrin-1

(A) An open book view of the regions on netrin-1 and DCC involved in binding

site 2 showing the electrostatic surface potential to emphasize the positively

charged regions within the binding site interface.

(B) View of the sulfate cluster at binding site 2, with sulfate ions shown as sticks

and the surface of the netrin-1 molecule shown with its electrostatic surface

potential. The FN5 and FN6 domains of DCC are shown as a ribbon diagram in

green.

(C) Identification of a chloride ion (green ball) at the interface of binding site 2,

between the EGF-1 domain of netrin-1 colored salmon and the FN5 domain of

DCC colored light green. An Fo-Fc omit map was calculated using the CCP4

suite (Winn et al., 2011) for a model lacking the chloride ion, and the electron

density is displayed at a contour level of 5s.

Neuron



NetrinVIV and two lysines (Lys896 and Lys911) and one arginine

(Arg861) from DCCFN56. Two arginines on netrin-1 (Arg 349 and

Arg351) are involved in direct hydrogen bonding interactions

with DCC, whereas the other three indirectly interact with the re-

ceptor through a sulfate ion. Previously, the FN5 domain of DCC

was identified to both bind to heparin or heparan sulfates, as well

as a netrin-1-blocking monoclonal antibody (Bennett et al.,

1997). Further analysis by domain swapping and immunoprecip-

itation assays indicated the formation of a ternary DCC/heparin/

netrin-1 complex in this region (Geisbrecht et al., 2003). The

presence of a structured cluster of sulfate ions also correlates

with a preference for certain proteoglycans to bind netrin-1

(Shipp and Hsieh-Wilson, 2007). The presence of multiple argi-

nines near the extended EGF-2 helix of netrin-1 is evolutionarily

conserved (Figure S3), as is the positively charged patch on DCC

(Figure S4A). All these sulfate-ion-associated residues are also

present in netrin-3 (Figure S5), whereas netrin-5 lacks one argi-

nine (Arg317 is substituted by a serine).

It is striking that there are only a few direct protein-protein

contacts between netrin-1 and DCC at site 2 (Figure 3C). These

contacts center around Pro320 on EGF-1 of netrin-1 whose pyr-

role ring is surrounded by the hydrophobic portion of residues

Thr856, His857, and Asp858, positioned on the protruding AB

loop of the FN5 domain of DCC. We also identified a chloride

ion situated nearby (Figure 3C) that is coordinated by four resi-

dues from NetrinVIV (Arg317, Cys318, Pro320, and Tyr323) and

two residues from DCCFN56 (namely Thr856 and Asp858) in an

octahedral geometry. The relatively long coordination bond dis-

tances and the presence of an arginine (Arg317) as a protein

ligand led us to conclude that it is a negatively charged chloride

ion that helps to neutralize the positively charged interface. A

chloride ion was also observed close-by in an anomalous map

calculated for an unliganded netrin-G2 crystal structure (Brasch

et al., 2011). A sequence comparison with other species shows

that the protein ligands involved in coordinating the chloride

ion on netrin-1 are conserved among vertebrae but not in insects

or nematodes (Figure S3). We also predict that chloride does not

bind to netrin-3 (Figure S5), because Tyr323 in netrin-1 is re-

placed by a cysteine in netrin-3. This may explain why netrin-3

binds DCC with lower affinity (Wang et al., 1999), given that the

other residues involved in DCC binding are conserved between

netrin-1 and netrin-3 for both binding sites.

We verified the role of the residues His857 and Asp858 on the

AB loop of the FN5 domain of DCC (Figure S4A) in binding to ne-

trin-1 on site 2 by expressing mutant DCC on COS-7 cells (Fig-

ures 4A and S1). The Asp858Ala and Asp858Arg mutants reduce

netrin-1 binding significantly, and for the His857Ala mutant,

binding is abolished. The positive patch on NetrinVIV was modi-

fied by mutating two adjacent arginines (Arg349 and Arg351)

that interact with sulfate ions as well as with residues of DCC.

The Arg349Ala/Arg351Ala mutant only showed a moderate

reduction in binding, but a mutant that introduces a charge

reversal (Arg349Asp/Arg351Asp) abolished binding completely

when applied to COS-7 cells that express wild-type DCC. The

ability to rescue chemoattraction of neurons harvested from

E15 DCC�/� mice by microinjection with a vector containing

DCC was also checked (Figures 4C and S2). The mutants

affecting the AB loop on the FN5 domain of DCC showed severe

reduction in chemoattraction compared to wild-type DCC.When

mutant NetrinVIV was applied to neurons injected with wild-type

DCC, chemoattraction was reduced the most for the charge

reversal mutant Arg349Asp/Arg351Asp. Together, these data

show that mutating residues involved in anion coordination

affect binding site 2, as well as axon guidance.

Multimerization of the Membrane-Proximal DCCFN56

Fragment Is NetrinVIV Dependent in SolutionIt is assumed that isolated DCC is monomeric, and netrin-1

driven dimerization of DCC leads to a rearrangement of the

two receptors that triggers a signal across the cell membrane

(Stein et al., 2001). To determine the oligomeric state of the

membrane proximal DCCFN56 fragment alone used for crystal-

lization of the NetrinVIV/DCCFN56 complex presented here,

small-angle X-ray scattering (SAXS) experiments were per-

formed (Table S2). A concentration series was measured for re-

combinant DCCFN56, yielding scattering curves suitable for the

calculation of the low-resolution molecular envelope ab initio

and for direct comparison with the coordinates of DCCFN56


Figure 4. Mapping Residues on the Generic

Receptor Binding Site 2 byMutagenesis that

Contribute to DCC Binding, Multimerization,

and Axon Guidance

(A) Cell binding assays for DCCHis857 and Asp858

and NetrinVIV Arg349/Arg351 mutants showing

the percentage of DCC presenting COS-7 cells

that bind netrin-1 or NetrinVIV. Data represent

mean ±SE (n = 100 for each group). Two-way

ANOVA, followed by post hoc Scheffe’s test. **:

p<0.01 compared with WT.

(B) Receptor multimerization assays for DCC

His857 and Asp858 and NetrinVIV Arg349/Arg351

mutants showing enrichment of oppositely tagged

receptors.

(C) Axon guidance assays for DCC His857 and

Asp858 and NetrinVIV Arg349/Arg351 mutants

showing percentage of beads attracting axons.

Data represent mean ±SE (n = 300 for each group).

Two-way ANOVA, followed by post hoc Scheffe’s

test. **: p<0.01 compared with WT.

Neuron



derived from the crystal structure (Figures 5A and S6). The

experimental scattering of DCCFN56 in solution fits well with the

scattering curve of monomeric DCCFN56 computed from atomic

coordinates by CRYSOL (Svergun et al., 1995), and a good

agreement between the reconstructed SAXS envelope and the

monomeric DCC56 fragment is observed. SAXS experiments

performed on NetrinVIV at different concentrations (Table S2)

gave a good fit to the experimental curve (Figure S6) for a mono-

meric NetrinVIV molecule with extended glycosylation (Figures

5B and S6; Experimental Procedures). This is in good agreement

with previous experiments on the related netrinG1 and netrinG2

proteins, which are also monomeric in solution in an unliganded

state (Brasch et al., 2011).

To monitor the process of DCC binding on the NetrinVIV mole-

cule at sites 1 and 2 in solution, SAXS studies of the NetrinVIV:

DCC mixtures were performed by varying the stoichiometry

from 1:1 till 1:10. This titration scattering data can only be

adequately represented by allowing for contributions from the

free NetrinVIV and DCCFN56 molecules, as well as complexes

with DCC bound either to site 1, to site 2, or to both binding sites.


The volume fraction of each component

was calculated with OLIGOMER (Konarev

et al., 2003) using the theoretical scat-

tering curves of the components. At a

stoichiometry of 1:1, there was no fully

occupied NetrinVIV/DCCFN56 complex

present. The available DCCFN56 mole-

cules bind to site 1 with two times higher

occupancy than to site 2 (Figure 5C),

and there is still a significant amount of

free NetrinVIV in solution. To see if we

could saturate binding of DCCFN56 to Ne-

trinVIV we increased the relative molar ra-

tio between NetrinVIV and DCCFN56 to

1:5. At this ratio, all free NetrinVIV has dis-

appeared and instead is associated with

DCCFN56 bound either at site 1 or site 2,

but not to both sites at the same time (Figure 5C). It was neces-

sary to increase the NetrinVIV/DCCFN56 ratio to 1:10 before the

ternary complex with two DCCFN56 molecules could be

observed. It can be inferred that DCC has a stronger affinity for

site 1, the DCC-specific site, and the full 1:2 NetrinVIV/DCCFN56

complex is observed in solution only when DCCFN56 is present

in abundance. SAXS performed on a DCCFN56 mutant that com-

promises the hotspot on binding site 1 (Met933Arg) shows that

modification of binding site 1 changes the composition of theNe-

trinVIV/DCCFN56 complexes in solution. The occupancy of bind-

ing site 1 is lowered substantially, but the occupancy of site 2

is not affected. This indicates that there is no cooperativity

between the binding sites for this part of DCC (Figure 5D).

UNC5A Occupies the Tolerant Receptor Binding Site onNetrin-1 to Switch to ChemorepulsionNetrin-1 is capable of switching from a chemoattractant to a che-

morepulsive guidance cue when UNC5 homologs are present on

the cell surface of the growth cone, in combination with DCC

(Hong et al., 1999). An axon guidance assay was designed

Figure 5. Molecular Envelopes of NetrinVIV

and DCCFN56 and SAXS Analysis of NetrinVIV/

DCCFN56 Complex Formation in Solution

(A and B) Ab initio molecular envelopes (gray

spheres) calculated from the SAXS data overlaid

onto the crystal structures (ribbons) of the individ-

ual (A) DCCFN56 and (B) NetrinVIV molecules with the

glycosylation sites on netrin-1 depicted (red and

black balls).

(C and D) (C) Bar diagrams of the volume fractions

of components contributing to the overall scat-

tering of a NetrinVIV/DCCFN56 and (D) a mutant

NetrinVIV/DCCFN56-M933R mixture in solution. At a

molar NetrinVIV concentration of 40 mM, DCCFN56 or

DCCFN56-M933R was mixed to achieve final molar

stoichiometries of 1:1, 1:5, and 1:10. Error bars are

based on the propagation of experimental errors

from the SAXS data through the non-negative least

squares fitting procedure.

Neuron



with fixed netrin-1 or NetrinVIV-coated beads to asses both che-

moattraction and repulsion by DCC�/� neurons that were in-

jected with UNC5A and/or DCC (Figures 6 and S7). In this assay,

neurons injected with both UNC5A and DCC wild-type mani-

fested a distribution of axons predominantly growing away

from the netrin source. Neurons injected with UNC5A alone

gave an even distribution of axons in all directions, indicating

these neurons had no sensitivity for netrin-1. In contrast, neurons

injected with DCC alone have a narrow distribution in axon direc-

tionality toward the netrin source.

To assess whether DCC is still occupying the DCC-specific

binding site 1 on netrin-1 in the presence of UNC5A, we coin-

Neuron 83, 1

jected DCC mutants V848R and M933R

that affect binding to site 1. Indeed, neu-

rons injected with site 1 mutants together

with UNC5A do not show chemorepul-

sion, indicating that the presence of

DCC at site 1 is essential for both chemo-

attraction and repulsion. In contrast, DCC

mutants that affect binding to netrin-1

site 2 that did abolish chemoattraction still

cause chemorepulsion in the presence of

UNC5A. It can be concluded that DCC is

not binding to the generic receptor binding

site 2 on netrin when UNC5A is present.

However, when the chemorepulsion re-

sponse is tested with a netrin mutant

(R349D/R351D) that affects the generic

binding site 2, no chemorepulsion is

observed (Figure S7). This indicates that

UNC5A binding involves the positive

patch on netrin-1 that associates with sul-

fate ions to mediate receptor binding.

Taken together, these results clearly

demonstrate that the binding of DCC to

both sites gives rise to chemoattraction

and that UNC5A outcompetes DCC for

binding to the generic receptor binding

site 2 on netrin-1, switching the netrin-1 response from chemo-

attraction to repulsion.

DISCUSSION

In this paper, we present the structure of a human netrin-1/DCC

complex that provides the molecular mechanism of netrin-1

signaling through DCC. The structure reveals two binding sites

for DCCon the Vdomain of netrin-1, which is in good accordance

with previous functional studies (Lim andWadsworth, 2002). The

1:2 netrin/DCC stoichiometry observed in the crystal structure

ties two DCC receptors together, which enables the dimerization

–11, August 20, 2014 ª2014 Elsevier Inc. 7

Figure 6. Repulsion/Attraction Assays Involving UNC5A and DCC

Radar plots showing the distribution of axon growth around the nucleus, with

the netrin-1-coated beads at the origin. Repulsion/attraction assays were

performed with DCC�/� neurons injected with DNA of UNC5A, DCC, and DCC

mutants showing that DCC is required for UNC5A-dependent repulsion.

DCC mutants affecting site 1 (V858R and M933R) abolish repulsion, whereas

DCC mutants affecting site 2 (H857A and D858R) still manifest repulsion,

indicating UNC5A has outcompeted DCC on site 2.

Neuron




of the cytosolic domains of DCC. There are no direct contacts

observed in the crystal structure between the two DCCFN56 frag-

ments situated at these two netrin-1 binding sites. Solution

studies indicate that both sites are occupied, and saturation

with DCCFN56 or a mutant for site 1 (DCCFN56-M933R) show that

the sites have distinct kinetics. These data suggest a modular

binding mode for the two DCCFN56 binding sites on netrin-1.

Binding site 2 on the EGF-1 and EGF-2 subdomains is quite

unique. The largest binding proponents are positively charged

patches on both netrin-1 and DCC, which would prohibit direct

binding due to electrostatic repulsion if there were no negatively

charged ions to neutralize the interface. The binding is facilitated

by several sulfate ions and a chloride ion that are embedded onto

the netrin-1 surface. There are clear structural elements on the

netrin-1 molecule that keep these ions in place. The chloride

ion is coordinated by four residues from the netrin-1 molecule,

and the clustered sulfate ions are all within acceptable distance

to interact with one of five arginine residues that are located

closely together on netrin-1. The side chains of these five argi-

nines cover a large conformational space, offering adaptability

for netrin-1 to accommodate other receptor molecules with a

similarly configured positively charged surface. The sulfate ions

will play a key role as binding intermediaries. The embedded

ion binding sitesmay provide the necessary flexibility to incorpo-

rate not just discreet ions but probably larger single structures

such as linear heparan sulfates. In particular, a sulfated glycan

unit can easily replace the cluster of four sulfate ions that is

kept together by Arg348, Arg349, Arg351, and Arg372 of ne-

trin-1. Positively charged patches on proteins in the extracellular

matrix have been identified as heparan sulfate binding sites

(Capila and Linhardt, 2002). The presence of an arginine/lysine/

sulfate cluster has previously been related to heparan sulfate

binding in the N-terminal domain of the extracellular matrix pro-

tein thrombospondin-1 (Tan et al., 2008). Similarly, heparan

sulfate may mediate receptor binding in netrin-1 so that distinct

heparan sulfates may favor binding of a particular receptor.

The adaption to bind anions is less stringent on the DCC side.

The residues contributing to the hotspot on DCC-specific bind-

ing site 1 are conserved, but the residues contributing to the

anion-dependent generic binding site 2 are evolutionarily more

divergent (Figures S4A and S4B). This may explain why the V

domain on netrin-1 is very conserved, whereas different organ-

isms have evolved different and yet DCC-like receptors that all

interact with netrin-1, such as the Frazzled receptor inDrosophila

and neogenin in vertebrate.

The recently reported crystal structure of netrin-1 in complex

with the FN4 and FN5 domains of DCC agrees with our DCC-

specific binding site on netrin-1 (Xu et al., 2014) but lacks the

generic receptor binding site, because that requires interactions

from the FN6 domain that is lacking in their DCC construct. The

Xu et al. structure shows that FN5 binds at the tip of the EGF-3

domain of netrin-1, and a similar binding mode is observed in

their NetrinVIV/neogenin complex. A superimposition of NetrinVIVand the FN5 domain of DCC from the Xu et al. structure onto our

structure gives an RMSD of 1.58 A for all Ca atoms, indicating

that these binding modes are very similar. In addition, they

have observed that the FN4 domain of DCC interacts with an

adjacent netrin-1 molecule at the laminin-like VI domain. A

Figure 7. Model of Heparan-Sulfate-Depen-

dent Formation of the DCC/DCC and DCC/

UNC5 Complex with Netrin-1

(A) Composite model of an extended netrin-1/DCC

cluster based on the superimposition of the crystal

structure presented here with the crystal structure

of NetrinVIV in complex with the FN4 and FN5 do-

mains of DCC (Xu et al., 2014). The NetrinVIVmolecules are colored in cyan, the DCC molecules

(FN4-FN5-FN6) occupying binding site 1 are

colored green, and the DCCmolecules (shown only

FN5-FN6) occupying site 2 are colored purple.

(B) DCC binds specifically to the EGF-3 domain of

netrin-1 (binding site 1). Netrin-1 associates with a

heparan sulfate molecule that is selective for DCC

(HP1) or UNC5 (HP2). When heparan sulfate HP1 is

bound to netrin-1, a second DCC molecule is re-

cruited to binding site 2, and the cytosolic P3 do-

mains of the two DCC molecules associate to form

a signaling complex, leading to attraction of the

growth cone. When heparan sulfate HP2 binds to

netrin-1, UNC5A is recruited to the EGF-1/EGF-2

domains, and the cytosolic P1 domain of DCC and

the region between the ZU5 and DB domain of

UNC5A form a signaling complex, leading to

repulsion of the growth cone.

Neuron



complementary model can be constructed by combining both

structures (Figure 7A), where one netrin-1 molecule associates

with two DCC molecules along the V domain to form a signaling

unit, and the DCC molecule that occupies the DCC-specific

binding site 1 on netrin-1 engages another netrin-1 molecule

through the FN4 domain to stitch different netrin/DCC signaling

units together. The DCC specific binding site 1 acts as an an-

chor, whereas the generic receptor binding site is occupied by

DCC, UNC5, or another receptor for distinct signaling.

One of themost intriguing questions regarding netrin biology is

its bifunctionality. It can act either as a chemoattractant or repel-

lent, depending on the type of receptors present on the growth

cone surface. It is conceivable that the generic binding site 2 is

the key factor in selecting and combining different receptors

for differential biological outcomes. It is known that receptors

UNC5 (Hong et al., 1999) and DSCAM (Ly et al., 2008) bind ne-

trin-1 through immunoglobulin domains, which have a fold that

is similar to that of the fibronectin domains used by DCC. It

was previously observed that the two immunoglobulin domains

of UNC5 receptors bind netrin-1 through heparin (Geisbrecht

et al., 2003), and these domains contain positively charged

patches as well. We have shown that in the presence of

UNC5A, DCC is replaced by UNC5A on the generic receptor

binding site. We also show that the presence of DCC at binding

site 1 on netrin-1 is necessary. We propose that the ternary ne-

trin/DCC/UNC5A complex initiates interactions between the

cytoplasmic P1 domain of DCC and the intracellular DCC-bind-

ing domain (DB) of UNC5 (Wang et al., 2009) (Figure 7B), leading

to a repulsive response by the axon. This outcome contrasts with

the netrin-induced formation of the DCC homodimer through its

cytoplasmic P3 domains, which triggers an attractive response.

Unlike many other cell surface receptors, here the dimerization

does not appear to be a result of direct ecto-domain contacts.

Rather, netrin-1 binds the ecto-domains of two receptors

(DCC/DCC or DCC/UNC5) and brings two receptors close

enough to initiate interactions between cytoplasmic domains

of different receptors. The crystal structure presented here

shows the netrin-1 signaling module and suggests that corecep-

tor selection is tuned by heparan sulfates and other environ-

mental factors, leading to alternate cues, and we anticipate

that this is fundamental to the netrin-based guidance code.

EXPERIMENTAL PROCEDURES

Protein Production and Crystallization

Recombinant human NetrinVIV was expressed in HEK293T cells, and recombi-

nant human DCCFN56 was expressed in E. coli BL21-DE3. Both proteins were

affinity purified, followed by size exclusion chromatography. Crystals were

obtained by vapor diffusion at 298 K by mixing NetrinVIV and DCCFN56 in 1:1

stoichiometry in 0.1 M MES (pH 6.0), 0.15 M ammonium sulfate, and 15%

(w/v) PEG 4000.

Crystal Structure Determination

X-ray diffraction data were collected on the EMBL P14 beamline at the

PETRA3 synchrotron using a PILATUS 6M pixel detector (DECTRIS). A single,

cryo-cooled crystal of the NetrinVIV/DCCFN56 complex diffracted to 3.1 A. The

structure was solved by molecular replacement and was then refined to an

Rfactor of 23.8% (Rfree = 28.5%) (Table S1).

SAXS Experiments

X-ray solution scattering data were collected on the EMBL P12 beamline at the

PETRA3 synchrotron using a PILATUS 2M pixel detector (DECTRIS). Data

were processed with a standard pipeline (Franke et al., 2012), and analyzed

by ATSAS programs (Petoukhov et al., 2012). The low-resolution shape enve-

lopes were determined using DAMMIF (Franke and Svergun, 2009). The scat-

tering properties of the mixtures of NetrinVIV and DCCFN56 were analyzed with

OLIGOMER (Konarev et al., 2003).


Neuron



Cell Binding and Receptor Multimerization Assays

Netrin-1-dependent DCC multimerization assays (Stein et al., 2001) and cell

binding assays (Keino-Masu et al., 1996) were performed as described previ-

ously using COS-7 cells transfected with wild-type ormutant DCC, which were

then incubated in media containing 10 mg/ml netrin-1 or mutant NetrinVIV.

Detection of binding and multimerzation is described in the Supplemental

Information.

Axon Guidance Assays

Mouse primary neuronswere cultured from the dorsal horn of the spinal cord of

E15 embryos of wild-type or DCC�/� mice (Rigato et al., 2011) following the

regulations of the Peking University Animal Care and User Committee. Axon

guidance assays were performed as previously described (Chen et al.,

2013), except that the beads were fixated by dipping them into 0.1% agrose

gel. This prevented the movement of beads during experimental manipulation.

ACCESSION NUMBERS

The coordinates of the NetrinVIV/DCCFN56 crystal structure have been depos-

ited in the Protein Data Bank with entry code 4URT. The experimental SAXS

data, as well as the models, have been uploaded to the SASBDB server at

http://www.sasbdb.org/ with codes SASDAZ5 (NetrinVIV alone), SASDA26

(DCCFN56 alone), SASDA76 (complex of NetrinVIV and DCCFN56) and SASDA86

(complex of NetrinVIVMet933Arg and DCCFN56).

SUPPLEMENTAL INFORMATION

Supplemental Information includes seven figures, two tables, and Supple-

mental Experimental Procedures and can be found with this article online at

http://dx.doi.org/10.1016/j.neuron.2014.07.010.

AUTHOR CONTRIBUTIONS

J-h.W. and R.M. conceived and coordinated the work. L.I.F., J.Z., and N.K.

performed the expression and purification of protein constructs. N.K. and

M.C. prepared the netrin-1/DCC complex. R.M., L.I.F., and J.-h.W. performed

structural analysis. SAXS data was collected and analyzed by G.S, H.M., and

D.S. Axon guidance assays, DCC-netrin-1 binding, and multimerization

assays were performed and analyzed by X.S, Y.W., L.I.F., and Y.Z. The manu-

script was written by J.-h.W. and R.M.

ACKNOWLEDGMENTS

We thank Yi Rao, Matthias Wilmanns, and Steve Harrison for critical reading of

the manuscript; the staff of the EMBL beamlines P12 and P14; and the sample

preparation and characterization (SPC) facility of EMBL at PETRA3 (DESY,

Hamburg) for assistance. We also thank Jonathan Rapley and Liya Xu for

early exploration of the project and Rebecca Spoerl and Alexandra Koetter

for technical support. Work was funded by the Ministry of Education of China

to J.-H.W. and Y.Z., the National Science Foundation of China (NSFC) Major

Research Grant (91132718) and the Beijing Natural Science Foundation

(7142085) to Y.Z., an NIH grant (HL48675) and funds from the Peking-Tsinghua

Center for Life Sciences to J.-H.W., and a short-term EMBO postdoctoral

fellowship to L.I.F. The authors G.S., H.D.T.M, and D.S. acknowledge support

by the European Commission (the 7th Framework Program) projects Bio-

Struct-X (contract number 283570) and WeNMR (contract number 261572).

Accepted: July 7, 2014

Published: August 7, 2014

REFERENCES

Andrews, G.L., Tanglao, S., Farmer, W.T., Morin, S., Brotman, S., Berberoglu,

M.A., Price, H., Fernandez, G.C., Mastick, G.S., Charron, F., and Kidd, T.

(2008). Dscam guides embryonic axons by Netrin-dependent and -indepen-

dent functions. Development 135, 3839–3848.


Backliwal, G., Hildinger, M., Chenuet, S., Wulhfard, S., De Jesus, M., and

Wurm, F.M. (2008). Rational vector design and multi-pathway modulation of

HEK 293E cells yield recombinant antibody titers exceeding 1 g/l by transient

transfection under serum-free conditions. Nucleic Acids Res. 36, e96.

Bennett, K.L., Bradshaw, J., Youngman, T., Rodgers, J., Greenfield, B., Aruffo,

A., and Linsley, P.S. (1997). Deleted in colorectal carcinoma (DCC) binds hep-

arin via its fifth fibronectin type III domain. J. Biol. Chem. 272, 26940–26946.

Boivin, S., Kozak, S., and Meijers, R. (2013). Optimization of protein purifica-

tion and characterization using Thermofluor screens. Protein Expr. Purif. 91,

192–206.

Brasch, J., Harrison, O.J., Ahlsen, G., Liu, Q., and Shapiro, L. (2011). Crystal

structure of the ligand binding domain of netrin G2. J. Mol. Biol. 414, 723–734.

Capila, I., and Linhardt, R.J. (2002). Heparin-protein interactions. Angew.

Chem. Int. Ed. Engl. 41, 391–412.

Carafoli, F., Hussain, S.-A., and Hohenester, E. (2012). Crystal structures of the

network-forming short-arm tips of the laminin b1 and g1 chains. PLoS ONE 7,

e42473.

Castets, M., Broutier, L., Molin, Y., Brevet, M., Chazot, G., Gadot, N., Paquet,

A., Mazelin, L., Jarrosson-Wuilleme, L., Scoazec, J.-Y., et al. (2012). DCC con-

strains tumour progression via its dependence receptor activity. Nature 482,

534–537.

Chen, Q., Sun, X., Zhou, X.H., Liu, J.H., Wu, J., Zhang, Y., and Wang, J.H.

(2013). N-terminal horseshoe conformation of DCC is functionally required

for axon guidance and might be shared by other neural receptors. J. Cell

Sci. 126, 186–195.

Clackson, T., andWells, J.A. (1995). A hot spot of binding energy in a hormone-

receptor interface. Science 267, 383–386.

Colamarino, S.A., and Tessier-Lavigne,M. (1995). The axonal chemoattractant

netrin-1 is also a chemorepellent for trochlear motor axons. Cell 81, 621–629.

Dickson, B.J. (2002). Molecular mechanisms of axon guidance. Science 298,

1959–1964.

Fearon, E.R., Cho, K.R., Nigro, J.M., Kern, S.E., Simons, J.W., Ruppert, J.M.,

Hamilton, S.R., Preisinger, A.C., Thomas, G., Kinzler, K.W., et al. (1990).

Identification of a chromosome 18q gene that is altered in colorectal cancers.

Science 247, 49–56.

Franke, D., and Svergun, D.I. (2009). DAMMIF, a program for rapid ab-initio

shape determination in small-angle scattering. J. Appl. Cryst. 42, 342–346.

Franke, D., Kikhney, A.G., and Svergun, D.I. (2012). Automated acquisition and

analysis of small angle X-ray scattering data. Nucleic Instrum. Methods Phys.

Res. Sect. Accel. Spectrometers Detect. Assoc. Equip. 689, 52–59.

Geisbrecht, B.V., Dowd, K.A., Barfield, R.W., Longo, P.A., and Leahy, D.J.

(2003). Netrin binds discrete subdomains of DCC and UNC5 and mediates in-

teractions between DCC and heparin. J. Biol. Chem. 278, 32561–32568.

Hong, K., Hinck, L., Nishiyama, M., Poo, M.M., Tessier-Lavigne, M., and Stein,

E. (1999). A ligand-gated association between cytoplasmic domains of UNC5

and DCC family receptors converts netrin-induced growth cone attraction to

repulsion. Cell 97, 927–941.

Hong, K., Nishiyama, M., Henley, J., Tessier-Lavigne, M., and Poo, M. (2000).

Calcium signalling in the guidance of nerve growth by netrin-1. Nature 403,

93–98.

Keino-Masu, K., Masu, M., Hinck, L., Leonardo, E.D., Chan, S.S., Culotti, J.G.,

and Tessier-Lavigne, M. (1996). Deleted in Colorectal Cancer (DCC) encodes a

netrin receptor. Cell 87, 175–185.

Kennedy, T.E., Serafini, T., de la Torre, J.R., and Tessier-Lavigne, M. (1994).

Netrins are diffusible chemotropic factors for commissural axons in the embry-

onic spinal cord. Cell 78, 425–435.

Koch, M., Murrell, J.R., Hunter, D.D., Olson, P.F., Jin, W., Keene, D.R.,

Brunken, W.J., and Burgeson, R.E. (2000). A novel member of the netrin family,

beta-netrin, shares homology with the beta chain of laminin: identification,

expression, and functional characterization. J. Cell Biol. 151, 221–234.

http://www.sasbdb.org/

http://dx.doi.org/10.1016/j.neuron.2014.07.010

Neuron



Kolodkin, A.L., and Tessier-Lavigne, M. (2011). Mechanisms and molecules of

neuronal wiring: a primer. Cold Spring Harb. Perspect. Biol. 3, http://dx.doi.

org/10.1101/cshperspect.a001727.

Konarev, P.V., Volkov, V.V., Sokolova, A.V., Koch, M.H.J., and Svergun, D.I.

(2003). PRIMUS: a Windows PC-based system for small-angle scattering

data analysis. J. Appl. Cryst. 36, 1277–1282.

Krimpenfort, P., Song, J.-Y., Proost, N., Zevenhoven, J., Jonkers, J., and

Berns, A. (2012). Deleted in colorectal carcinoma suppresses metastasis in

p53-deficient mammary tumours. Nature 482, 538–541.

Lai Wing Sun, K., Correia, J.P., and Kennedy, T.E. (2011). Netrins: versatile

extracellular cues with diverse functions. Development 138, 2153–2169.

Li, X., Gao, X., Liu, G., Xiong,W.,Wu, J., and Rao, Y. (2008). Netrin signal trans-

duction and the guanine nucleotide exchange factor DOCK180 in attractive

signaling. Nat. Neurosci. 11, 28–35.

Lim, Y.S., and Wadsworth, W.G. (2002). Identification of domains of netrin

UNC-6 that mediate attractive and repulsive guidance and responses from

cells and growth cones. J. Neurosci. 22, 7080–7087.

Liu, G., Beggs, H., Jurgensen, C., Park, H.-T., Tang, H., Gorski, J., Jones, K.R.,

Reichardt, L.F., Wu, J., and Rao, Y. (2004). Netrin requires focal adhesion

kinase and Src family kinases for axon outgrowth and attraction. Nat.

Neurosci. 7, 1222–1232.

Liu, G., Li, W., Wang, L., Kar, A., Guan, K.-L., Rao, Y., and Wu, J.Y. (2009).

DSCAM functions as a netrin receptor in commissural axon pathfinding.

Proc. Natl. Acad. Sci. USA 106, 2951–2956.

Ly, A., Nikolaev, A., Suresh, G., Zheng, Y., Tessier-Lavigne, M., and Stein, E.

(2008). DSCAM is a netrin receptor that collaborates with DCC in mediating

turning responses to netrin-1. Cell 133, 1241–1254.

Mazelin, L., Bernet, A., Bonod-Bidaud, C., Pays, L., Arnaud, S., Gespach, C.,

Bredesen, D.E., Scoazec, J.-Y., and Mehlen, P. (2004). Netrin-1 controls colo-

rectal tumorigenesis by regulating apoptosis. Nature 431, 80–84.

Mille, F., Llambi, F., Guix, C., Delloye-Bourgeois, C., Guenebeaud, C., Castro-

Obregon, S., Bredesen, D.E., Thibert, C., and Mehlen, P. (2009). Interfering

with multimerization of netrin-1 receptors triggers tumor cell death. Cell

Death Differ. 16, 1344–1351.

Petoukhov, M.V., Franke, D., Shkumatov, A.V., Tria, G., Kikhney, A.G., Gajda,

M., Gorba, C., Mertens, H.D.T., Konarev, P.V., and Svergun, D.I. (2012). New

developments in the ATSAS program package for small-angle scattering data

analysis. J. Appl. Cryst. 45, 342–350.

Pinato, G., Cojoc, D., Lien, L.T., Ansuini, A., Ban, J., D’Este, E., and Torre, V.

(2012). Less than 5 Netrin-1 molecules initiate attraction but 200 Sema3Amol-

ecules are necessary for repulsion. Sci Rep 2, 675.

Rajasekharan, S., and Kennedy, T.E. (2009). The netrin protein family. Genome

Biol. 10, 239.

Rigato, C., Buckinx, R., Le-Corronc, H., Rigo, J.M., and Legendre, P. (2011).

Pattern of invasion of the embryonic mouse spinal cord by microglial cells at

the time of the onset of functional neuronal networks. Glia 59, 675–695.

Seiradake, E., Coles, C.H., Perestenko, P.V., Harlos, K., McIlhinney, R.A.J.,

Aricescu, A.R., and Jones, E.Y. (2011). Structural basis for cell surface

patterning through NetrinG-NGL interactions. EMBO J. 30, 4479–4488.

Serafini, T., Kennedy, T.E., Galko, M.J., Mirzayan, C., Jessell, T.M., and

Tessier-Lavigne, M. (1994). The netrins define a family of axon outgrowth-pro-

moting proteins homologous to C. elegans UNC-6. Cell 78, 409–424.

Shipp, E.L., and Hsieh-Wilson, L.C. (2007). Profiling the sulfation specificities

of glycosaminoglycan interactions with growth factors and chemotactic pro-

teins using microarrays. Chem. Biol. 14, 195–208.

Stein, E., Zou, Y., Poo, M., and Tessier-Lavigne, M. (2001). Binding of DCC by

netrin-1 to mediate axon guidance independent of adenosine A2B receptor

activation. Science 291, 1976–1982.

Svergun, D., Barberato, C., and Koch, M.H.J. (1995). CRYSOL – a Program to

Evaluate X-ray Solution Scattering of Biological Macromolecules from Atomic

Coordinates. J. Appl. Cryst. 28, 768–773.

Tan, K., Duquette, M., Liu, J.H., Shanmugasundaram, K., Joachimiak, A.,

Gallagher, J.T., Rigby, A.C., Wang, J.H., and Lawler, J. (2008). Heparin-

induced cis- and trans-dimerization modes of the thrombospondin-1

N-terminal domain. J. Biol. Chem. 283, 3932–3941.

Wang, H., Copeland, N.G., Gilbert, D.J., Jenkins, N.A., and Tessier-Lavigne,

M. (1999). Netrin-3, a mouse homolog of human NTN2L, is highly expressed

in sensory ganglia and shows differential binding to netrin receptors.

J. Neurosci. 19, 4938–4947.

Wang, R., Wei, Z., Jin, H., Wu, H., Yu, C., Wen, W., Chan, L.-N., Wen, Z., and

Zhang, M. (2009). Autoinhibition of UNC5b revealed by the cytoplasmic

domain structure of the receptor. Mol. Cell 33, 692–703.

Wilson, N.H., and Key, B. (2006). Neogenin interacts with RGMa and netrin-1 to

guide axons within the embryonic vertebrate forebrain. Dev. Biol. 296,

485–498.

Wilson, B.D., Ii, M., Park, K.W., Suli, A., Sorensen, L.K., Larrieu-Lahargue, F.,

Urness, L.D., Suh, W., Asai, J., Kock, G.A.H., et al. (2006). Netrins promote

developmental and therapeutic angiogenesis. Science 313, 640–644.

Winn, M.D., Ballard, C.C., Cowtan, K.D., Dodson, E.J., Emsley, P., Evans,

P.R., Keegan, R.M., Krissinel, E.B., Leslie, A.G.W., McCoy, A., et al. (2011).

Overview of the CCP4 suite and current developments. Acta Crystallogr. D

Biol. Crystallogr. 67, 235–242.

Xu, K., Wu, Z., Renier, N., Antipenko, A., Tzvetkova-Robev, D., Xu, Y.,

Minchenko, M., Nardi-Dei, V., Rajashankar, K.R., Himanen, J., et al. (2014).

Neural migration. Structures of netrin-1 bound to two receptors provide insight

into its axon guidance mechanism. Science 344, 1275–1279.


http://dx.doi.org/10.1101/cshperspect.a001727

http://dx.doi.org/10.1101/cshperspect.a001727

The Crystal Structure of Netrin-1 in Complex with DCC Reveals the Bifunctionality of Netrin-1 As a Guidance Cue

Documents