amigo adhesion protein regulates development of neural circuits

Amigo Adhesion Protein Regulates Development of NeuralCircuits in Zebrafish Brain*

Received for publication, December 23, 2013, and in revised form, June 3, 2014 Published, JBC Papers in Press, June 5, 2014, DOI 10.1074/jbc.M113.545582

Xiang Zhao‡1, Juha Kuja-Panula‡, Maria Sundvik‡§, Yu-Chia Chen‡§, Vilma Aho¶, Marjaana A. Peltola‡,Tarja Porkka-Heiskanen¶, Pertti Panula‡§, and Heikki Rauvala‡2

From the ‡Neuroscience Center, §Institute of Biomedicine/Anatomy, and ¶Institute of Biomedicine/Physiology, University ofHelsinki, Helsinki FIN-00014, Finland

Background: Amigos are transmembrane proteins suggested to mediate adhesive interactions of cells.Results: Down-regulation of Amigo protein expression or inhibition of its function results in defects of fiber pathway develop-ment and disturbed locomotor functions in zebrafish.Conclusion: The neural form of Amigo is required for construction of functional neural circuitries.Significance: Amigo is identified as a novel factor that regulates formation of neuronal connections.

The Amigo protein family consists of three transmembraneproteins characterized by six leucine-rich repeat domains andone immunoglobulin-like domain in their extracellular moi-eties. Previous in vitro studies have suggested a role as homo-philic adhesion molecules in brain neurons, but the in vivo func-tions remain unknown. Here we have cloned all three zebrafishamigos and show that amigo1 is the predominant family mem-ber expressed during nervous system development in zebrafish.Knockdown of amigo1 expression using morpholino oligonu-cleotides impairs the formation of fasciculated tracts in earlyfiber scaffolds of brain. A similar defect in fiber tract develop-ment is caused by mRNA-mediated expression of the Amigo1ectodomain that inhibits adhesion mediated by the full-length protein. Analysis of differentiated neural circuitsreveals defects in the catecholaminergic system. At thebehavioral level, the disturbed formation of neural circuitry isreflected in enhanced locomotor activity and in the inabilityof the larvae to perform normal escape responses. We suggestthat Amigo1 is essential for the development of neural cir-cuits of zebrafish, where its mechanism involves homophilicinteractions within the developing fiber tracts and regulationof the Kv2.1 potassium channel to form functional neuralcircuitry that controls locomotion.

Development of functional neural circuits depends on spa-tially and temporally precise cellular interactions mediated by awide variety of secreted and membrane-bound factors. Proteinscontaining extracellular leucine-rich repeat (LRR)3 domains

have recently gained increasing interest as key factors guidingdevelopment of neuronal connectivity (for a recent review, seeRef. 1). LRR domains are protein-protein interaction motifsthat have been implicated in development and plasticity of fibertracts, recognition of axon targets, synaptogenesis, and disor-ders of the nervous system. Slit proteins (Slit 1–3), Trk recep-tors (TrkA, TrkB, and TrkC), and Nogo receptor 1 (NgR1) areamong the most studied LRR proteins in the nervous systemdevelopment (2, 3).

AMIGO (amphoterin-induced gene and open reading frame)was found by ordered differential display as a transcript up-reg-ulated in rat hippocampal neurons that extend neurites uponligation of the transmembrane receptor RAGE (receptor foradvanced glycation end products) by amphoterin (HMGB1;high mobility group box-1) or by anti-RAGE antibodies (4).Together with two homologous proteins, the three AMIGOs(AMIGO1–3) form a novel family of transmembrane LRR pro-teins. The three AMIGOs display about 50% sequence similar-ity compared with each other, and they have the same domainorganization, with six extracellular LRR motifs and one immu-noglobulin-like (Ig) domain, another protein motif that iswidely expressed in neural recognition molecules (5). In BLASTsearches using the AMIGO1 ectodomain to identify homolo-gous proteins outside of the AMIGO family, the best fits arefound for the axon-guiding Slit proteins and the Nogo-66receptor (NgR1).

AMIGO1 and AMIGO2 are highly expressed in the mamma-lian nervous system, whereas AMIGO3 displays a broad expres-sion pattern in different tissues (4, 6, 7). In vitro studies havesuggested that AMIGO1 acts as a homophilic adhesion mole-cule that induces outgrowth and fasciculation of neurites incentral neurons (4). The crystal structure of the AMIGO1dimer has recently revealed that homophilic binding ofAMIGO1 occurs through its LRR domains (8). In addition,AMIGO1 and AMIGO2 (also designated as Alivin1) have beenreported to enhance survival of neurons in culture (9, 10). Fur-thermore, AMIGO1 has recently been found to bind to the

* This work was supported by the Academy of Finland and the Sigrid JuséliusFoundation.

1 Supported by the Doctoral Program in Brain and Mind of the DoctoralSchool in Health Sciences.

2 To whom correspondence should be addressed: Neuroscience Center, P.O.Box 56 (Viikinkaari 4), FIN-00014 University of Helsinki, Finland. Tel.: 358-9-19157621; Fax: 358-9-19157620; E-mail: [email protected].

3 The abbreviations used are: LRR, leucine-rich repeat; CA, catecholaminergic;DiV, diencephalon ventricle; dpf, day(s) postfertilization; hpf, hour(s) post-fertilization; Ig domain, immunoglobulin-like domain; MBP, maltose-bind-ing protein; MLCT, medial longitudinal catecholaminergic tract; MLF,medial longitudinal fasciculus; MO, medulla oblongata; POC, postopticcommissure; TH, tyrosine hydroxylase; vcc, ventral caudal cell cluster; qRT-

PCR, quantitative RT-PCR; SLC, short latency C-start; LLC, long latencyC-start; ANOVA, analysis of variance.

THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 289, NO. 29, pp. 19958 –19975, July 18, 2014© 2014 by The American Society for Biochemistry and Molecular Biology, Inc. Published in the U.S.A.

19958 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 289 • NUMBER 29 • JULY 18, 2014

by guest on January 31, 2018http://w

ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

Kv2.1 potassium channel and to affect excitability of centralneurons via the Kv2.1 interactions (11).

Despite several in vitro findings providing clues to develop-mental roles of the AMIGOs, their in vivo functions remainunknown. In the current study, we have used the zebrafishmodel to study Amigo functions in nervous system develop-ment. We have cloned all three amigo genes of zebrafish andfocused our functional studies on amigo1, which was found tobe the major member of the gene family expressed in zebrafishduring the nervous system development. Using the morpholinoknockdown approach and expression of the Amigo1 ectodo-main as a dominant negative receptor, we show that Amigo1 isessential for the development of fiber tracts in the zebrafishbrain.

EXPERIMENTAL PROCEDURES

Animals—An outbred zebrafish (Danio rerio) strain from alocal resource, the Turku line, was used in this study for itssteady yield of embryos (12). Fish feeding, breeding, and main-tenance were done according to an established protocol (13).The experiment permits were obtained from the University ofHelsinki Committee for animal experiments and the NationalAnimal Experiment Board in agreement with the ethical guide-lines of the European Convention. Embryos of either sex werestaged according to the number of hours postfertilization (hpf)or days postfertilization (dpf). To prevent pigment formation,0.2 mM 1-phenyl-2-thiourea (Sigma) was added to the mediumof embryos after spawning.

Cloning of amigo Genes and Quantitative RT-PCR—Ze-brafish orthologs were found by BLASTing the peptidesequences of the mammalian AMIGO proteins using the UCSCgenome browser and Ensembl zebrafish ZV9 database. TotalRNA from 1– 8-dpf larvae was extracted and reverse-tran-scribed. The subsequent PCR was performed as described pre-viously (12). RNA quality and amount were analyzed spectro-photometrically by a NanoVue Plus spectrophotometer (GEHealthcare). The primers 5�-ATG CCC CCT TCC ATT AATTG-3� and 5�-CCG GTC AAA AGA TAC ACA TCC TC-3�were used for cloning the full-length coding sequence ofamigo1 (ENSDARG00000079620). The primers 5�-ATG ACCTCG ACA TCT TGC ATG GTT-3� and 5�-GAT TCA AACAAG CAG GAT TTT AAG G-3� were used for cloning thefull-length coding sequence of the second gene that displaysclear homology compared with the amigo genes (ENS-DARG00000079569; designated here as amigo3a). The primers5�-ATG CTG TGT GCT CAG GGT GTG GC-3� and 5�-TCATCT CTC TGA TTC TAT GTG CTT TCC T-3� were used forcloning the full-length coding sequence of the third homologousgene (ENSDARG00000074469; designated here as amigo3b). ThePCR products were purified with the MinEluteGel Extraction kit(Qiagen, Hilden, Germany), ligated to the pGEM-Teasy plasmid(Promega), and sequenced.

The quantitative RT-PCR (qRT-PCR) primers of amigo1were 5�-TCG CCG TGA GTG AAT ACC TC-3� and 5�-TGCCAA GCA ACC CAC CAA A-3�. The qPCR primers ofamigo3a were 5�-GGC TGT GTT GTG ACC CTT GT-3� and5�-GAT GAG ATG GCT GGA GAT GGA-3�. The qPCR prim-ers of amigo3b were 5�-ACA CTG GCT TCA CCA CAC T-3�

and 5�-AGC GAC AGG GAG TCA AGT AG-3�. The zebrafish �-actin (AF057040.1, GI:3044209), elongation factor 1�1 (zgc:109885, GI:90652818), and ribosomal protein L13a(BC047855.1, GI:28838761) were set as template quantity con-trols, and the same primers were used as described previously(12). The PCRs were processed with the Bio-Rad CFX96 real-time PCR machine using the CFX96TM real-time PCR detec-tion system. The three constructed amigo plasmids prepared indifferent dilutions (from 1 mg/ml to 0.1 �g/ml) in milli-Q waterwere used as standards for qPCR. The amigo1 expression levelnormalized to �-actin in 1 dpf larvae was set as 1.

Cloning of amigo1 and kv2.1 mRNAs—The full-length openreading frame mRNA construct encoding Amigo1 was pre-pared by RT-PCR with primers 5�-ATA AGA TCT ATG CCCCCT TCC ATT AAT TG-3� and 5�-AAG AAT TCC CGG TCAAAA GAT ACA CAT CCT C-3�. The construct includes BglIIand EcoRI restriction sites on both sides of the amigo1 firststrand cDNA in the full-length transcript. These restrictionsites were used for insertion into the pMC expression vector,which caps the inserted fragment with the sequence elementsof the �-globin 5�-UTR and 3�-UTR on both of its sides and inaddition the SV40 poly(A) signal at the 3�-tail. This elevates thetranscribed mRNA stability and activity about 100-fold inmRNA injection experiments (14). The mRNA encoding theAmigo1 ectodomain was prepared with primers 5�-ATA AGATCT ATG CCC CCT TCC ATT AAT TG-3� and 5�-ATG AATTCT CAG CCA TTG CCG TTG AGG-3�. GFP mRNA wasused as another control in mRNA rescue experiments, and itwas prepared using the pEGFP-C1 vector. All mRNAs wereprepared using the mMessagemMachine kit (Ambion, Austin,TX) according to the manufacturer’s instructions.

The full-length kv2.1 cDNA (ZDB-GENE-090831-3) wasprepared by RT-PCR with primers 5�-ATA AGC TTC CCTCGG CAG GAA TGA GTA A-3� and 5�-ATG GAT CCT CAAAGG CCC TTA TCA AAA G-3� and cloned into pMC expres-sion vector. A cDNA fragment encoding the N terminus andthe transmembrane loops of the Kv2.1 protein (432 aminoacids) was cloned into pMal-c2E vector (New England BiolabsInc.) for Kv2.1 ectodomain-maltose-binding protein (MBP)recombinant expression with primers 5�-ATG GAT CCC CCTCGG CAG GAA TGA GTA A-3� and 5�-ATA AGC TTC TTGATG GCC TTC TCT TGT-3�.

Antibodies against Amigo1 and Kv2.1—Chicken polyclonalantibodies were produced against the zebrafish Amigo1(Agrisera, Sweden) using Amigo1 ectodomain-GST (glutathi-one S-transferase) fusion protein as the antigen. In order tomake a high quality antigen, the Amigo1 extracellular codingsequence was cloned into pGEX-2TK GST fusion vector (GEHealthcare) using the primers 5�-ATA GGA TCC TGC GCCAGC AAC ATT GTC AG-3� and 5�-ATG AAT TCT CAG CCATTG CCG TTG AGG-3� to add the BamHI and EcoRI restric-tion sites on both sides of the coding sequence. The Amigo1ectodomain-MBP fusion protein was produced by cloning thecoding sequence into pMAL-c2E vector. The primers 5�-ATGAAT TCT GCG CCA GCA ACA TTG TCA G-3� and 5�-ATAGGA TCC TCA GCC ATT GCC GTT GAG G-3� were used foradding EcoRI and BamHI on both sides of the coding sequence.The plasmids were transformed into Escherichia coli XL-1 blue

Amigo in Brain Development

JULY 18, 2014 • VOLUME 289 • NUMBER 29 JOURNAL OF BIOLOGICAL CHEMISTRY 19959


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

strain by electroporation (Bio-Rad) for Amigo1 ectodomain-GST/MBP recombinant expression. The recombinant proteinwas purified using two different affinity columns (GST or MBPtag). The purified Amigo1 ectodomain-GST/MBP fusion pro-tein was analyzed on SDS-PAGE stained with Coomassie Blue(Invitrogen).

The IgY antibodies in the immunized egg yolk were first puri-fied by ammonium sulfate precipitation. In affinity purification,the first step was carried out on an affinity column loaded withAmigo1 ectodomain-GST fusion protein. The second step inaffinity purification was carried out on another affinity columnloaded with Amigo1 ectodomain-MBP fusion protein toexclude the proteins that bind nonspecifically to the GST tag.The purified zebrafish Amigo1 antibodies were diluted to 1mg/ml aliquots containing 0.01% NaN3 and stored at �80 °C.Monoclonal anti-Kv2.1 antibodies (K39/25) were obtainedfrom the NeuroMab facility (University of California, Davis/National Institutes of Health).

Western Blotting and Co-immunoprecipitation—ZebrafishAmigo1 antibodies (1:1000 corresponding to 1 �g/ml) and thepurified monoclonal Kv2.1 antibodies (2 �g/ml) were used forWestern blotting and for co-immunoprecipitation that wascarried out according to a previously described protocol (11).Dithiobis(succinimidyl propionate) cross-linking (Pierce) wasapplied as before (11). Anti-HNK-1 antibodies (C0678, 1:500;mouse monoclonal, Sigma-Aldrich) and anti-acetylated tubu-lin (6-11B-1 mouse monoclonal, 1:1000; Sigma-Aldrich) wereused to detect zebrafish axonogenesis and axonal developmentin Western blotting and whole mount immunostaining. Mousemonoclonal anti-�-actin antibody (A2228, 1:1000; Sigma-Al-drich) was used as the control of sample loading. The zebrafishtissue preparation and Western blotting were carried out asdescribed previously (12, 13).

Immunocytochemistry—Mouse monoclonal anti-tyrosinehydroxylase (TH) antibody (Diasorin, Stillwater, MN) was usedfor the whole mount staining of the catecholaminergic systemin 3 dpf larval brains as described earlier (12). The mouse anti-3A10 (1:100; Developmental Studies Hybridoma Bank),chicken anti-Amigo1 antibodies (5 �g/ml), and anti-HNK-1antibodies (1:100) were used for whole mount immunostainingof zebrafish larvae. The Amigo1 ectodomain-MBP fusion pro-tein was purified and used as a competing antigen in antibodybinding (100 �g/ml). For immunohistochemistry of the larvae,the samples were fixed in 2% paraformaldehyde for 2 h at roomtemperature and then washed and preincubated in phosphate-buffered saline containing 0.1% Tween 20 (PBS-T, pH 7.4) with1% dimethyl sulfoxide (DMSO) and 4% normal goat serum at4 °C overnight or longer. The specimens were incubated withthe primary antibodies in the preincubation solution (PBS-Twith 2% normal goat serum) for over 12 h at 4 °C under slowstirring. The samples were then washed thoroughly with PBS-Tand incubated with the Alexa�-conjugated goat anti-chicken(Alexa488 and Alexa546) or goat anti-mouse (Alexa488 andAlexa568) secondary antibodies (diluted 1:2000) in the prein-cubation solution for over 12 h at 4 °C. The samples werewashed with PBS-T twice for 30 min, once with PBS for 30 min,and once with 50% glycerol in PBS for 1 h and then infiltratedovernight in 80% glycerol in PBS before mounting.

Whole Mount in Situ Hybridization—Whole mount in situhybridization was carried out as described previously (15),using the specific probes (16) of pax2a (the ZIRC cb378), pax6a(ZIRC cb280), and krox20 (ZIRC cb427). Larvae at 28 hpf(Prim-5) and 2 dpf (Long-pec) stages were used in the experi-ments. The probe for the detection of amigo1 expressionwas obtained from the cDNA clone encoding the Amigo1ectodomain.

Detection of Apoptosis and Proliferation—A fluorometricTUNEL system (DeadEndTM, Promega) was used for stainingof apoptosis in whole mounts, and Edu staining (Click-iT EduAlexa Fluor 555, Invitrogen) was used for detection of cell pro-liferation in whole mounts. Apoptosis and proliferation wereexplored according to the protocol recommended by the man-ufacturer (12).

Morpholino Oligonucleotide and mRNA Injections—Knock-down experiments of amigo1 were carried out with translation-blocking antisense morpholino oligonucleotides (MOs) (GeneTools) targeted to the 5�-upstream sequence flanking the trans-lation start site (MO1, 5�-GGC ATT TCT GAC ACG CAGTTA AAA T-3�; MO2, 5�-TGT GGT TGT AGC ACA AGTCAT AAA C-3�) of the amigo1 transcript. To knock downamigo3b expression, the translation-blocking MO was 5�-GGCCAC ACC CTG AGC ACA CAG CAT T-3�. The 5-mispairoligonucleotide 5mis MO, GcC ATT TgT GAg ACc CAc TTAAAA T-3� (mispairs are in lowercase type), was used as an MOinjection control. Cloning of the mRNAs encoding the full-length Amigo1 and the ectodomain of Amigo1 (mRNA andEmRNA, respectively) is explained above.

Microinjections of the MOs and the mRNAs into zebrafishembryos were carried out as described previously (12). An ali-quot of 4 nl of the mixture (corresponding to 4 ng of MOs when100 M solution was used) was injected into the yolk of a 1– 4-cellembryo that was allowed to develop at 28.5 °C. For Amigo1mRNA and GFP mRNA injections, 0.1– 0.2 �g/�l mRNA wasmixed with injection solution after heating. The injection con-centration was tested for eliminating ubiquitous overexpres-sion effects (12). For Kv2.1 mRNA injections, 0.1 �g/�l mRNAwas mixed with the injection solution.

Microscopy and Image Analysis—Confocal imaging of larvaestained with anti-TH and anti-Amigo1 antibodies was carriedout as described (12). Zeiss LSM710 Pascal confocal micros-copy system was used for imaging larval samples stained withanti-3A10 antibodies and anti-HNK1 antibodies. Stacks ofimages taken at 0.6 –1.2-�m intervals were compiled to makemaximum intensity projection images. Specimens from in situhybridization were examined with inverted light microscopyusing Olympus IX 70 connected through a CCD camera to theAnalysis� software. High resolution images were obtained witha digital MicroFire S99808 camera (Optronics) attached toOlympus BX51 epifluorescence microscope (Olympus, Tokyo,Japan). The acquired images were further processed with Core-lDRAW� Graphics Suite X6 (Corel Corp.). Analysis of cellnumbers from the whole stack of images scanned through thesamples was carried out using ImageJ to count the stained cells(17).

Staining intensity of axonal tracts was measured by ZeissEfficient Navigation (ZEN 2012) software (Carl Zeiss MicroIm-




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

aging GmbH) with single maximum intensity z-stacking pro-jections containing the whole image stacks. To quantify thedegree of axonal tract development, the axonal tract/cell bodyratio was measured on each maximum intensity z-stacking pro-jection (18). For the determination of medial longitudinal fas-cicle (MLF), average pixel intensity of anti-HNK-1 immunore-active MLF descending from the ventral caudal cluster (vcc)cells was measured (50 � 2 �m) and compared with that of thevcc cell bodies (5 � 5 �m) in the same z-stacking projection. Forthe determination of the medial longitudinal catecholaminer-gic tract (MLCT), average pixel intensity of anti-TH1-immuno-reactive MLCT descending from the locus coeruleus wasmeasured (30 � 3 �m) and compared with that of the locuscoeruleus cell bodies (3 � 3 �m) in the same z-stacking projec-tion. The average value of 5mis MO larvae was used for normal-ization in the tract analyses. The background fluorescence wassubtracted from each measurement. All image acquisition andanalysis was performed blind to the experimental group.

Anatomical structures of larval brain were named and num-bered using the neuroanatomical atlas of developing zebrafishbrain (19) and the atlas based on location of TH neurons in 5 dpflarval fish (20, 21).

Behavioral Assays—Two behavioral assays were used.First, the locomotor activity of 6-dpf larvae was observedduring 10 min in 24-well plates as described previously (20,22). The analysis included 80 fish (20/group in four inde-pendent experiments).

In the second behavioral assay, we assessed the startleresponse (C-start response) of 9-dpf larvae elicited by electricalstimuli. We chose electrical stimulation because it allowed us toprovide a fixed and stable startle stimulus in each group of lar-vae during constant illumination (23). For the electrical stimu-lation experiments, recording chambers (60 ml, x � 5 cm, y � 4cm, z � 3 cm shared from the laboratory of H. A. Burgess) werefitted with stainless steel sidewalls as bipolar stimulating elec-trodes (5 V, 0.01 s). For each experiment, 20 larvae were accli-mated in the recording chamber for 3 min before testing. TheMotionPro X3 fast recording camera (Redlake Imaging) wascontrolled by a script written in Matlab (MathWorks). Record-ing was started 200 ms before the electrical stimulus was given(via the MotionPro XTM data acquisition hardware module,Redlake Imaging) and was continued for 1 s. This was repeatedthree times with the same larvae in the chamber, with a 3-mininterval set between each electrical stimulus. Sequential imagesof the C-start response were captured every 1 ms. Latency wasdefined as the time from the beginning of the stimulation to theC-bend formation. The captured images were analyzed withFlote version 2.1. Short latency C-start (SLC; initiated within 20ms after stimulus) and long latency C-start (LLC; initiated laterthan 20 ms after stimulus) share identical movement trajecto-ries and were classified according to the previous report (24).For each group, the experiment was repeated four times withfish from different injections.

Analytical Software—Student’s t test and one-way ANOVAanalyses were carried out using OriginPro version 8.0. Intensi-ties of the Western blot bands were examined with QuantityOne version 4.6.2 (Bio-Rad). Sequence alignments of the

AMIGO orthologs of different species were carried out usingGeneious Pro R6 (Biomatters Ltd.).

RESULTS

Zebrafish Amigo Proteins and Detection of Zebrafish Amigo1—According to the characteristic LRR and Ig domains, threeputative amigo transcripts are found in the zebrafish Zv9 data-base (Ensembl, search zebrafish). The full-length cDNAs werecloned by using mRNA from zebrafish larvae and primersdesigned according to the putative homologous sequencesfound in the Zv9 database. Comparison of the deduced aminoacid sequences of the zebrafish Amigos with the human,rodent, chicken, medaka, and Xenopus proteins was carried outusing the Geneious software. Within the 17 Amigo proteinsquoted, the zebrafish Amigo1 shares over 50% identity in itsamino acid sequence compared with the Amigo1 protein in allother species (Fig. 1A). The two other zebrafish sequences (des-ignated as Amigo3a and -3b) display a higher degree of similar-ity to AMIGO3 than to AMIGO2. It appears that no AMIGO2ortholog has yet been identified in teleost species.

A phylogenetic tree view of the alignment result clearlyshowed that within the Amigo family, Amigo1 has the highestdegree of homology when compared with the orthologsexpressed in different species (Fig. 1A). Zebrafish Amigo3a andAmigo3b are both similar to mammalian AMIGO3 andgrouped into a different branch compared with AMIGO2 of theother species (Fig. 1A). Zebrafish Amigo3a and -3b might beencoded by duplicate copies of the same gene.

In order to detect expression of the Amigos in larvalzebrafish, the primers for qRT-PCR of amigo1, amigo3a, andamigo3b were designed and used for expression analysis in lar-vae at the developmental stages 1– 8 dpf. The qRT-PCR resultsrevealed that all three amigo expressions start from the begin-ning of early developmental stages (Fig. 1B). Expression ofamigo1 was found to be the highest of the three amigo tran-scripts. Expression of amigo1 already started to increase after24 hpf and was increased in a logarithmic manner during thestages 1– 8 dpf. Expression of amigo3a and amigo3b started toincrease from 3 dpf and was always lower than expression ofamigo1. Expression of amigo3b was found to be the lowest ofthe three transcripts. Before 4 dpf, amigo1 expression wasfound to be over 5-fold higher than that of amigo3a or amigo3b.After 4 dpf, amigo1 expression was still over 3-fold higher thanexpression of the two other transcripts.

Because amigo1 was found to be the major amigo transcriptin zebrafish early embryos, we decided to focus on this familymember to elucidate putative amigo functions in nervous sys-tem development. Because the antibodies that we have previ-ously generated against the mouse AMIGO1 (11) did not spe-cifically detect the zebrafish Amigo1 protein in Westernblotting of SDS extracts of larvae, we generated antibodiesagainst the zebrafish Amigo1 sequence. We constructed afusion plasmid clone encoding Amigo1 ectodomain-GST andpurified the expressed protein to be used as an immunogen inchicken (4). Western blotting revealed that the affinity-purifiedantibodies detect specifically the Amigo1 ectodomain-MBPfusion protein and the endogenously expressed protein in larval




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

FIGURE 1. AMIGO proteins in different species, expression of the amigos in zebrafish, and production of antibodies to detect Amigo1 of zebrafish. A,alignment of the amino acid sequences of the zebrafish Amigos with 14 AMIGO proteins found in other species (on the left). Identity of the amino acidsequences is indicated in the table. A phylogenetic tree of 17 AMIGO proteins found in Xenopus, chicken, rat, mouse, human, zebrafish, and medaka wasgenerated by the Clustal multiple-sequence alignment program (on the right). The similarities and identities were computed by Geneious. Scale bar, number ofamino acid substitutions/site. B, expression of the three amigos during zebrafish development detected by qRT-PCR. Full-length cDNA plasmid clones ofamigo1, amigo3a, and amigo3b were constructed and used for the standardization of each qRT-PCR. Templates were normalized by three reference genes asdescribed under “Experimental Procedures.” amigo1 expression compared with �-actin in 1-dpf larvae was set as 1. amigo1 expression increases over 30-foldduring development from 1 to 8 dpf. For each qRT-PCR test, total RNA was extracted from a pool of 20 larvae. The data are based on three independentexperiments. Mean values � S.E. (error bars) are indicated. C, antibodies against zebrafish Amigo1. Specificity of the anti-Amigo1 antibodies is shown byWestern blotting on the left. The purified antibodies stain the Amigo1 ectodomain-MBP fusion protein (E-MBP) band (�75 kDa) clearly with weak background.The antibodies also detect specifically the Amigo1 band (�60 kDa) in the SDS extract of the adult zebrafish brain. Coomassie Blue staining on the right confirmshigh expression of the Amigo1 ectodomain-MBP fusion protein used for purification and binding tests of the antibodies. D, whole mount immunostaining of5-dpf larval brains with the anti-Amigo1 antibodies. Goat anti-chicken Alexa 546 (orange to white) was used as the secondary antibody for fluorescent labeling.The maximum z-stacking projection images (maximum depth, 130 �m) were scanned from both ventral and dorsal sides to show the Amigo1 expressionpattern throughout the brain (the thickness of the samples over 200 �m). Compared with the uninjected or the 5mis MO-injected larvae, immunostaining isclearly inhibited in the MO1-injected larvae or in the uninjected larvae in the presence of the competing antigen (E-MBP comp). CC, cerebellar crest; CeP,cerebellar plate; Di, diencephalon; Pr, pretectum; SR, superior raphe; Tel, telencephalon; TeO, tectum opticum. Scale bar, 100 �m. E, domain structure ofzebrafish Amigo1 based on the sequence of the transcript. Recognition sites of the designed morpholino oligonucleotides (MO1 and MO2) around the 5�-UTRof the transcript are marked by short red lines. The part of the transcript encoding the Amigo1 ectodomain was used for dominant negative expression (Am1EmRNA). IG-domain, immunoglobulin-like domain; LRRCT, leucine-rich repeat C-terminal flanking domain; LRRNT, leucine-rich repeat N-terminal flankingdomain; TM-domain, transmembrane domain.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

SDS extracts that corresponds to the molecular mass of Amigo1(Fig. 1C).

In addition to the zebrafish Amigo1 specific antibodies, wedesigned a knockdown approach to study expression and func-tions of Amigo1 in zebrafish. Two gene-specific antisense MOs(MO1 and MO2; Fig. 1E) were designed for amigo1 translationinhibition (25). Because the amigo1 transcript contains onlyone integrated exon without any introns, MO1 and MO2 weretargeted for binding to the 5�-UTR sites near the start codon(Fig. 1E). They are not expected to inhibit protein translationfrom capped amigo1 mRNA injected in rescue experiments.Uninjected larvae and larvae injected with 5mis MO (havingfive non-pairing nucleotides compared with MO1) were used ascontrols.

To verify the specificity of the Amigo1 immunostaining, weused two different approaches. First, immunostaining using theanti-Amigo1 antibodies was strongly reduced in the knock-down morphants compared with the uninjected or the 5misMO-injected controls (Figs. 1D and 2A). To restore Amigo1expression in knockdown experiments, we tested injections ofAmigo1 mRNA at the doses ranging from 0.2 to 0.8 ng. Amigo1was immunostained at an intensity comparable with the endog-enous expression in larvae injected with about 0.4 ng of Amigo1mRNA (Fig. 2A; see also Western blotting results of Fig. 3, Aand B). As an additional control, Amigo1 ectodomain-MBPfusion protein was premixed with the antibodies for antigencompetition in immunostaining. Amigo1-specific immuno-staining was blocked in the prebinding experiments in a similarmanner as in the knockdown experiments (Fig. 1D, E-MBPcomp).

Amigo1 Expression in Larval Brain—The affinity-purifiedantibodies detected Amigo1 in whole mount immunostainingalready at 28 hpf in the larval brain (Fig. 2A). Amigo1 wasintensely stained in the neuronal progenitor cell layers liningthe diencephalic ventricle (DiV) and the optic recess of the fore-brain (26, 27). Furthermore, the anterior diencephalon was alsospecifically stained on the alar plate prosomere, from which theanterior thalamic nuclei derive (28), and on the upper layer ofposterior tuberculum. Essentially no detection was observed inthe amigo1 knockdown morphants compared with the unin-jected or the 5mis MO-injected controls (Fig. 2A). Expressionof amigo1 was partially rescued by coinjection of the amigo1full-length mRNA with the MOs, as evidenced by intensive andubiquitous Amigo1 staining in DiV cells (Fig. 2A, MO1 �mRNA and MO2 � mRNA).

At 28 hpf, Amigo1 was found to be expressed in both celllayers and the early developing fiber tracts. MLF and thepostoptic commissure (POC) belong to early developing fiberscaffolds of zebrafish that were immunostained with the anti-Amigo1 and the anti-HNK-1 antibodies. Amigo1 was detectedat the leading front and along the fascicle of the developingtracts (Fig. 2, B and C).

The early embryonic brain development in zebrafish islargely accomplished at 5 dpf (29, 30). At this stage, Amigo1 iswidely distributed in most nuclei in telencephalon and com-missures in diencephalon (for the expression pattern at 5 dpf,see Fig. 1D).

Gross Morphology of amigo1 Knockdown Morphants—Ininjections with different MO doses, the death rates in larvaeinjected with MO1, MO2, or 5mis MO were essentially thesame (about 5%) at the doses of 2– 4 ng but were significantlyincreased at higher doses (data not shown). We thereforeselected 4 ng as the dose for the injections in further experi-ments because it was found to be the highest dose of the MOsallowing high proportions of live larvae.

We used Western blotting of 3 dpf larval lysates to evaluatethe effects of the MOs at a dose of 4 ng on Amigo1 proteinexpression. Amigo1 expression essentially disappeared in theMO1- and MO2-injected larvae, whereas no major change wasobserved in amigo3a MO-injected morphants or in larvaeinjected with 5mis MO. Coinjection of the amigo1 mRNA withMO1 or MO2 displayed a clear rescue effect on the protein

FIGURE 2. Immunohistochemistry of Amigo1 expression during earlydevelopment of the zebrafish brain. A, whole mount immunostaining ofAmigo1 in 28-hpf larvae. Anti-Amigo1 antibodies were used as the primaryantibodies, and goat anti-chicken Alexa 488 antibodies were used as the sec-ondary antibodies for fluorescent labeling. Control groups are the 5mis MO-injected and the uninjected larvae. amigo1 knockdown morphants aremarked as MO1 and MO2. Rescued groups are coinjected with the full-lengthmRNA (MO1 � mRNA and MO2 � mRNA). DT, dorsal thalamus; P, pallium; Pr,pretectum; Po, preoptic area. The confocal images were scanned and stackedas shown by the inset on the right. The arrow in each frame indicates the opticrecess of DiV as the landmark. Scale bar, 60 �m. B, double staining of 28-hpflarvae with chicken anti-Amigo1 and mouse anti-HNK1 antibodies; the pri-mary antibodies were detected with anti-chicken Alexa 546 (red; anti-Amigo1antibodies) and goat anti-mouse Alexa 488 (green; anti-HNK1 antibodies). B1,anti-Amigo1 antibodies showed staining on the superficial layer of the hind-brain ventricle (HbVe). The arrow indicates the direction from posterior toanterior. B2 and B3, Z-stacking confocal images scanned under higher mag-nification show that anti-Amigo1 and anti-HNK1 antibodies stain the tracts.The rectangles in B1 indicate the areas of B2 and B3. H, hindbrain; TPOC, tract ofthe postoptic commissure. Scale bars, 25 �m (B1) and 4 �m (B2 and B3). C, theascending POC in telencephalon (Tel) stained with mouse anti-HNK1 antibod-ies (left) and chicken anti-Amigo1 antibodies (middle) and overlay of the stain-ings (right). The arrow shows the leading part of the tract in the direction of itsextension. Scale bar, 5 �m.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

expression (Fig. 3A). Quantification of the Western blot resultsrevealed 85% decrease in Amigo1 expression in the MO1-injected morphants and 90% in the MO2-injected morphants(Fig. 3B). The 4-ng dose in the MO injections thus appearsoptimal to avoid nonspecific effects and to obtain efficientknockdown effects of the Amigo1 protein expression. Immu-nostaining of Amigo1 in the larvae injected with 4 ng of theMOs (Figs. 1D and 2A) was found to agree with the proteinexpression data.

When the endogenous amigo1 expression was inhibited dur-ing the development using injections of MO1 and MO2 at adose of 4 ng, the morphants did not show gross morphological

defects during development (Fig. 4A). The knockdown mor-phants did not either display any difference in expression levelsof regional transcription factors (pax2a, pax6a, and krox20)that are known to be important for brain development (Fig. 4, Band C). Closer inspection, however, suggested that the mor-phants have shorter length and width of head (Fig. 4B, pax6aand krox20 stainings), and smaller tectum opticum (Fig. 3C,TeO) and midbrain-hindbrain boundary (MHB; cell layerbetween cerebellar plate and midbrain tectum; Figs. 3C and 4B)compared with the controls.

Amigo1 knockdown might perturb brain development bycausing problems in cell proliferation and survival. We there-

FIGURE 3. Knockdown of Amigo1 protein expression and phenotypic features of the knockdown larvae. A, Western blotting of 3-dpf larvae lysates withanti-Amigo1 antibodies. No Amigo1 band is detected on the MO1 and MO2 lanes. B, quantification of Amigo1 expression by plot density analyses. In 3-dpfamigo1 knockdown morphants (MO1 and MO2), amigo1 expression is significantly inhibited compared with the 5mis MO-injected controls. The plot density ofthe bands from uninjected larval SDS extracts was set as 1 for normalization. Western blots were repeated three times in three independent experiments (n �3; ***, p 0.001 in one-way ANOVA followed by Tukey’s post hoc test). C, Edu staining (red) and fluorescent TUNEL staining (green) do not display differencesbetween the amigo1 knockdown morphants (MO1) and the uninjected control. CeP, cerebellar plate; Di, diencephalon; H, hindbrain; MHB, midbrain hindbrainboundary; Tel, telencephalon; TeO, tectum opticum. Scale bar, 50 �m in the Edu-stained and 40 �m in the TUNEL-stained panel. D, numbers of apoptotic cellsbased on TUNEL staining in prosencephalon and mesencephalon of 28-hpf larvae. The amigo1 morphants do not display a significant difference comparedwith the 5mis MO-injected controls. The data are based on three independent experiments; apoptotic cell numbers in 15 larvae of each group were counted(n � 15, one-way ANOVA followed by Tukey’s post hoc test). Error bars, S.E.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

fore carried out EdU and TUNEL stainings in whole mountpreparations of larvae but did not find any apparent changes(Fig. 3C). Counting of apoptotic cells in the area covering thewhole prosencephalon and mesencephalon (31) did not revealany differences in different experimental settings used in thestudy (Fig. 3D).

Amigo1 Regulates Expression of the Kv2.1 Potassium Channeland Binds to Kv2.1—We have recently shown that AMIGO1binds to Kv2.1 in adult mouse neurons and regulates voltage-gated potassium currents through Kv2.1 (11). We thereforedecided to study whether expression of Kv2.1 is changed in the

Amigo1 knockdown morphants and whether Amigo1 andKv2.1 interact in zebrafish neurons.

Based on the sequences in the zebrafish Zv9 database(Ensembl, search D. rerio), a putative kv2.1 transcript isexpressed in zebrafish displaying 66% identity in its sequencecompared with the mouse Kv2.1 transcript. We produced anN-terminal fragment of the zebrafish Kv.2.1 containing theextracellular domain and the transmembrane loops as an MBPfusion protein (�80 kDa). Monoclonal antibodies againstmouse Kv2.1 (K39/25) were found to recognize the recombi-nant Kv2.1 fragment (Fig. 5A). Western blotting of SDS extracts

FIGURE 4. Amigo1 morphants do not display defects in gross morphology or in expression of transcription factors that regulate neuronal develop-ment. A, light microscopy images of 3-dpf larvae. Scale bar, 1 mm. B, whole mount in situ hybridization of pax2a and pax6a in 2-dpf larvae (the pharyngulaperiod from High-pec to Long-pec) and krox20 in 24 hpf larvae (the pharyngula period from Prim 5 to Prim 15). The first lane shows a lateral view of pax2a in2-dpf larvae as a marker of the midbrain-hindbrain boundary (MHB) development. Scale bar, 200 �m. The second lane shows a dorsal view of pax6a in 2-dpflarvae as a marker of forebrain development. Scale bar, 100 �m. The last lane shows a dorsal view of krox20 in 24-hpf larvae as a marker of rhombomere3 (r3)and rhombomere5 (r5) in hindbrain. Scale bar, 80 �m. The asterisks indicate the smaller midbrain-hindbrain boundary and rhombomere3 of MO1 larvae. C,quantification of pax2a, pax6a, and krox20 expression by qRT-PCR using total RNA extracted from 2-dpf larvae. Each qRT-PCR sample was obtained from a poolof 20 larvae. Templates were normalized by three reference genes, as described under “Experimental Procedures.” The data are based on three independentexperiments. Mean values � S.E. (error bars) are indicated.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

from 3 dpf larvae showed that the Kv2.1 protein is expressed inthe larvae, and its expression is strongly decreased in theamigo1 knockdown larvae (MO1), displaying about 10% of theexpression found in control larvae (Fig. 5, A and B). Coinjectionof amigo1 mRNA with MO1 (MO1 � Am1 mRNA) restoredthe expression to 30 – 40% of that found in the controls. Coin-jection of kv2.1 mRNA with MO1 restored Kv2.1 expression to80% of the level seen in the control larvae (Fig. 5, A and B).

Double immunostaining of brain sections from adultzebrafish showed colocalization of Kv2.1 and Amigo1 through-out the brain, shown for diencephalic neurons in Fig. 5C. Coex-pression was also observed in the hindbrain ventricle cell layers(HbVe) and the edge of the cerebellar plate (CeP) in 28-hpflarvae (Fig. 5D), but the staining patterns did not suggestexpression in the same cells.

Because the immunostaining experiments suggested colo-calization of Amigo1 and Kv2.1 in adult zebrafish neurons, wecarried out co-immunoprecipitation experiments using adultbrain samples. These experiments suggested binding ofAmigo1 to Kv2.1 (Fig. 5E). The binding was more prominent indithiobis(succinimidyl propionate)-cross-linked samples com-pared with samples without any cross-linker. The apparent rea-son for this is that the transmembrane domains of the proteinsmediate binding are largely lost in detergent extraction used inthe co-immunoprecipitation experiments, which was previ-ously shown in the experiments using brain extracts from mice(11).

Amigo1 Is Required for the Development of Early AxonalScaffolds in Zebrafish Brain—At 28 hpf, the initial axon scaffoldof the zebrafish brain is simple, consisting of a few longitudinaltracts connected by commissures (32, 33). Antibodies againstthe cell surface marker HNK-1 (34) specifically label neuronsand axons in the early embryonic nervous system (35). We usedimmunocytochemistry with anti-HNK-1 antibodies to visual-ize the possible role of Amigo1 in the development of earlynervous system structures (Fig. 6, A and B).

Confocal stacking projections of the lateral side of larvaeclearly showed disturbed fiber tract development in amigo1knockdown morphants at 28 hpf (Fig. 6A). The MLF and tractof the postoptic commissure belong to the very early majortracts and only appear as thin fibers in the amigo1 MO1 mor-phants compared with the wild-type or 5mis MO-injected

embryos. Coinjection of the full-length amigo1 mRNA (Am1mRNA) with MO1 displayed a rescuing effect (Fig. 6, A and B).

We have previously shown that the AMIGO proteins displayhomophilic binding that appears important for fasciculation ofneurites in cultures of rat brain neurons. Furthermore, the ect-odomain of AMIGO1 in solution was shown to act as a domi-nant negative receptor and to inhibit AMIGO1-mediated adhe-sion and fasciculation in brain neurons in vitro (4). In order toprovide an assay that does not depend on MOs, we prepared anmRNA encoding the extracellular part of Amigo1 (Am1EmRNA, coding for the six LLR domains and one Ig domain ofAmigo1; Fig. 1E) and injected it into fertilized embryos. Expres-sion of the dominant negative protein in the injected larvaecould be distinguished by Western blotting with anti-Amigo1antibodies (Fig. 6C). As in the amigo1 MO1 morphants, devel-opment of the early tracts was severely disturbed in larvaeinjected with the Am1 EmRNA. For example, in the confocalstacking projections, MLF could hardly be discerned in theAm1 EmRNA-injected larvae (Fig. 6A). In contrast to the Am1EmRNA, larvae injected with the full-length mRNA did notdisplay such inhibition (Fig. 6A, Am1 mRNA) but even dis-played rescue of axonal growth (Fig. 6A, MO1 � Am1 mRNA).The larvae coinjected with MO1 and kv2.1 mRNA (MO1 �kv2.1 mRNA) showed MLF defects, as did the amigo1 MO1morphants and the Am1 EmRNA-injected larvae (Fig. 6A).

Statistics of the 28-hpf larvae immunostained with anti-HNK-1 antibodies showed that significantly lower numbersof the amigo1 knockdown morphants and amigo1 EmRNA-injected larvae had intact MLF than 5mis MO-injected oramigo1 mRNA-injected larvae (Fig. 6D). Intensity of the MLFstaining confirmed the defective tract development in theknockdown morphants and in the EmRNA-injected larvae (Fig.6E). A clear rescuing effect in the development of MLF wasfound in coinjection of the amigo1 mRNA with the MOs (Fig. 6,D and E), which could not be achieved by kv2.1 mRNAcoinjection.

The vcc and the hindbrain cell cluster are important for thedevelopment of early ascending and descending long tracts (35,36) and were found to be correctly located in the knockdownmorphants (Fig. 6A, vcc and hc). Under higher magnification,the MO1 morphants displayed a clearly reduced HNK-1 stain-ing of axons from the ventral caudal cell clusters (Fig. 6B, top

FIGURE 5. Amigo1 regulates Kv2.1 expression and binds to Kv2.1. A, monoclonal anti-Kv2.1 antibodies detect the ectodomain of zebrafish Kv2.1 (Kv2.1E-MBP) in Western blotting. Western blotting of 3-dpf larval SDS extracts using the anti-kv2.1 antibodies shows strong down-regulation of Kv2.1 in Amigo1knockdown morphants (MO1) compared with the 5mis MO-injected or the uninjected larvae. Western blotting suggests that Kv2.1 expression is enhanced bycoinjection of amigo1 or kv2.1 mRNA with MO1. B, quantification of Kv2.1 expression in the amigo1 MO1 morphants shows significant down-regulation of theexpression compared with the 5mis MO-injected or the uninjected larvae. In the MO1/Amigo1 mRNA-coinjected larvae, Kv2.1 expression is increased to 30%of the uninjected controls, which is still significantly less than the expression in the 5mis MO group. In the MO1/Kv2.1 mRNA-coinjected larvae, Kv2.1 expressionis increased to over 60% of the uninjected controls. Western blots were repeated four times in four independent experiments (n � 4; **, p 0.01 in one-wayANOVA followed by Tukey’s post hoc test). C, double staining of adult brain sections with anti-Amigo1 and anti-Kv2.1 antibodies shows their coexpression in thediencephalic neuronal cells in the adult (over 1-year-old) zebrafish. The primary antibodies were detected with goat anti-chicken Alexa 488 (green; anti-Amigo1antibodies) and goat anti-mouse Alexa 568 (red; anti-Kv2.1 antibodies). The light microscopy image on the right shows Amigo1 whole mount in situ hybrid-ization in a brain sagittal section. Rectangle, the Amigo1-expressing brain region that was used for the double immunostaining of Amigo1 and Kv2.1. Scale bar,10 �m. CP, central posterior thalamic nucleus; PGZ, periventricular gray zone of optic tectum; PPv, ventral periventricular pretectal nucleus; TeV, tectumventricle; TL, torus longitudinalis; TPp, periventricular nucleus of posterior tuberculum; Val, vlavula cerebelli. D, double staining of 28-hpf larvae with anti-Amigo1 and anti-Kv2.1 antibodies shows that Kv2.1 and Amigo1 are both expressed in the hindbrain ventricle area from an early development stage but arenot expressed in the same cells. The rectangular area in the light microscopy image on the right is the region shown in the fluorescent images. Scale bar, 20 �m.CeP, cerebellar plate; H, hindbrain; HbVe, hindbrain ventricle; MHB, midbrain-hindbrain boundary. E, coimmunoprecipitation (IP) of Kv2.1 and Amigo1 fromadult brain SDS extracts using the monoclonal anti-Kv2.1 and the polyclonal chicken anti-Amigo1 antibodies. Dithiobis(succinimidyl propionate) (DSP)-treatedSDS extracts show enhanced coimmunoprecipitation. Non-immune IgY and IgG were used as negative controls in Western blotting of Amigo1 and Kv2.1,respectively. In the anti-Amigo1 and anti-Kv2.1 controls, immunoprecipitation and Western blotting were carried out with the same antibodies. Error bars, S.E.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

panels). HNK-1 is also present in Mauthner neurons and thelong growing trigeminal axons (35). As the largest reticulospi-nal interneurons in zebrafish, the Mauthner neurons and theirlong axons contribute to early sensorimotor circuit formation(37). Anti-3A10 antibodies were used for specific labeling ofMauthner neurons and their axonal projections (38). Disturbedaxonal growth was also clearly observed in the Mauthner neu-rons in the hindbrain of the Amigo1 knockdown morphants(Fig. 6B, bottom panels). HNK-1 and 3A10 staining of the axons

was partially rescued by coinjection with the amigo1 mRNA butnot with the kv2.1 mRNA (Fig. 6B).

Acetylated tubulin is a component of microtubules, which arethe first cytoskeletal elements to appear in axons (39, 40). Asanother frequently used marker for labeling early developing axontracts in zebrafish, the anti-acetylated tubulin antibodies show asimilar immunostaining pattern as the anti-HNK-1 antibodies (32,41). Here we used Western blotting to quantify expression of theHNK-1 epitope and acetylated tubulin as markers of the neuronal




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

fiber tract development (Fig. 7). Western blotting of 3-dpf larvalsamples showed that expression of both HNK-1 and acetylatedtubulin is strongly reduced in the amigo1 knockdown (MO1 andMO2) morphants (Fig. 7A). The HNK-1 expression level of theamigo1 MO1 and MO2 morphants was decreased to about 20% ofthe uninjected controls (Fig. 7B). Coinjection of the full-lengthamigo1 mRNA (Am1 mRNA) with MO1 or MO2 caused a signif-icant rescue in the HNK-1 expression (Fig. 7, A and B). HNK-1expression in the Am1 mRNA-MO1-coinjected larvae was ele-vated to 50% of the uninjected control, and in the Am1mRNA-MO2-coinjected larvae, the expression was enhanced to60% of the uninjected control (Fig. 7B). In the same larval sam-ples, Western blotting of acetylated tubulin in the amigo1 MO1and MO2 morphants showed a similar decrease as the HNK-1epitope (Fig. 7A). In the mRNA-coinjected rescued groups, acety-lated tubulin expression was recovered to 70% of the uninjectedcontrol.

As in the MO-injected larvae, HNK-1 and acetylated tubulinwere also decreased in the dominant negative amigo1EmRNA-injected larvae (Fig. 7A). HNK-1 was reduced to about25% of the normal control (Fig. 7B), and acetylated tubulin wasreduced to about 40% of the normal control (Fig. 7C).

In the amigo1 mRNA-injected larvae, the expression ofHNK-1 and acetylated tubulin is 90% of the uninjected larvae.From the statistical results, there is no significant change in theexpression of the pathway markers in the mRNA-injected, the5mis MO-injected, or the GFP mRNA-injected controls (Fig. 7,B and C). In the kv2.1 mRNA and MO1- or MO2-coinjectedlarvae, Amigo1 expression was also significantly decreased as inthe amigo1 knockdown larvae, to 10 –20% of the uninjectedlarvae (Fig. 7D). In contrast to the amigo1 mRNA, the kv2.1mRNA did not restore HNK-1 and acetylated tubulin expres-sion. HNK-1 expression in the kv2.1 mRNA and MO1- orMO2-coinjected groups was about 30% of the uninjected con-trols (Fig. 7B). Acetylated tubulin expression in the kv2.1mRNA and MO1-coinjected group was 20% of the uninjectedlarvae, and in the kv2.1 mRNA and MO2-coinjected group, theexpression was about 30% (Fig. 7C).

Development of Aminergic Circuits in amigo1 KnockdownMorphants—In addition to the early fiber scaffolds, we haveanalyzed the development of differentiated circuits at laterdevelopmental stages in the amigo1 knockdown morphants. Tothis end, we chose to elucidate the development of the cat-echolaminergic (CA) system using immunostaining with THantibodies.

The CA system is the major neuromodulatory system withfar ranging projections in the brain, which is important in themodulation of circuit activities in a broad range of behaviors(37, 42). CA development in zebrafish starts quite early, andmost of the CA groups and axon tracts found in the adult braincan already be detected in 3-dpf larvae (29, 38, 43). The longi-tudinal axon tracts from dopaminergic neurons extend fromthe diencephalon toward the spinal cord in the vicinity of theMLF (see above for the role of Amigo1 in MLF formation) andlateral longitudinal fasciculus (LLF), but their formation maynot depend on these early axonal scaffolds (38, 42).

From the confocal z-stacking sections of the 3-dpf whole lar-vae, the amigo1 knockdown morphants (MO1 and MO2)showed predominant defects in the formation of the MLCT(Fig. 8, A and B). The MLCT projections ascending anddescending from the locus coeruleus displayed a disorderedpattern. Coinjection of the amigo1 mRNA with MO1 or MO2(MO1 � Am1 mRNA and MO2 � Am1 mRNA) had a promi-nent rescue effect on MLCT development, which was notobserved in the kv2.1 mRNA and MO1 coinjected larvae(MO1 � kv2.1 mRNA). Statistics of the CA development basedon the MLCT formation revealed a developmental defect in ahigh proportion of the amigo1 knockdown larvae (MO1 andMO2) at 3 dpf compared with the 5mis MO-injected larvae orthe uninjected larvae (Fig. 8C). In the MO1 and MO2 mor-phants, 70 – 80% of larvae had MLCT defects. In MO1 � Am1mRNA and MO2 � Am1 mRNA groups, the larvae with MLCTdefects decreased to 20 –30%. In the MO1 � kv2.1 mRNAgroups, over 75% of larvae displayed MLCT defects (Fig. 8C).The intensity analysis of the TH1 antibody staining confirmedthe dramatically decreased MLCT development in the MO1

FIGURE 6. Immunohistochemistry showing decreased fasciculation and fiber outgrowth in amigo1 knockdown morphants (MO1 and MO2) and inlarvae expressing the Amigo1 ectodomain (Am1 EmRNA) as an inhibitor of the endogenous full-length protein. A, whole mount anti-HNK-1 immuno-staining of 28-hpf larvae. The amigo1 knockdown morphant (MO1) shows decreased staining in MLF, indicated by the arrows. The ectopically expressedamigo1 ectodomain mRNA (Am1 EmRNA) results in a similar MLF defect. Coinjection of the amigo1 mRNA but not the kv2.1 mRNA rescues the tract defectcaused by MO1. The schematic map on the right shows the main axonal tracts stained in uninjected 28-hpf larvae. The rectangle in the schematic map indicatesthe region shown in B. hc, hindbrain cell cluster; PC, posterior commissure; TPOC, tract of the postoptic commissure. Scale bar, 100 �m. B, early axonaldevelopment at the area of the vcc shown by whole mount immunostaining using mouse anti-HNK-1 antibodies (top) and mouse anti-3A10 antibodies(bottom) in 28-hpf larvae. The amigo1 knockdown morphant (MO1), the 5mis MO control, and the mRNA rescue effects (MO1 � Am1 mRNA and MO1 � Kv2.1mRNA) are shown. Fibers in the tract of the postoptic commissure, originating from the vcc, are decreased in the MO1 and the MO1 � kv2.1 mRNA larvae.Anti-3A10 antibodies used in whole mount immunostaining show axon development of Mauthner neurons in the hindbrain of 28-hpf larvae. The schematicmap on the right shows the normal staining pattern. The asterisk indicates the defective development of axon projections from Mauthner neurons in the MO1and the MO1 � kv2.1 mRNA larvae compared with the 5mis MO and the MO1 � Am1 mRNA larvae. M, Mauthner neuron; MiD2cm, middle dorsal 2 contralateralMLF interneuron; MiD3cm, middle dorsal 3 contralateral MLF interneuron; OV, otic vesicle. Scale bars, 10 �m in the top row and 40 �m in the bottom row. C,Western blotting of 2-dpf larval lysates (10 larvae/sample) with anti-Amigo1 antibodies. Expression of the Amigo1 ectodomain in the Am1 EmRNA-injectedgroup is detected as the strong band at around 45 kDa, indicated by the arrow. D, statistics of larvae with normal MLF stained by anti-HNK-1 antibodies. Theamigo1 knockdown morphants (MO1 and MO2), the Am1 EmRNA larvae, and the MO1 � Kv2.1 mRNA larvae display significant MLF defects compared with therescued (MO1 � Am1 mRNA and MO2 � Am1 mRNA) and the 5 misMO-injected groups. For each group, 10 larvae were selected randomly from the pool of 50larvae, and four independent experiments were included. Mean values � S.E. (error bars) are indicated (n � 4; **, p � 0.001, one-way ANOVA followed byTukey’s post hoc test). E, average pixel intensity analyses of anti-HNK1 stained MLF from Z-stacking confocal images. The average pixel intensity of MLF in 5misMO larvae was used for normalization. In larvae injected with MO1, MO2, Am1 EmRNA, or MO1 � Kv2.1 mRNA, MLF staining intensities are significantly lowerthan in larvae injected with 5mis MO or amigo1 mRNA or coinjected with MO1 or MO2 and the amigo1 mRNA. The results are based on 12 larvae randomlyselected from three independent experiments (n � 12; **, p � 0.001, one-way ANOVA followed by Tukey’s post hoc test).




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

and MO2 morphants, which was rescued by the amigo1 mRNAbut not by the kv2.1 mRNA (Fig. 8D).

Behavioral Consequences of amigo1 Knockdown—Theremarkable changes observed in neuronal circuitry in theamigo1 knockdown morphants are likely to be reflected inbehavior, such as locomotion. We therefore tested motor func-tions of the amigo1 knockdown morphants using two differentbehavioral tests.

In the first test, free swimming was assayed in 6-dpf larvae todetermine the locomotor activity. Surprisingly, the amigo1 MO1and MO2 morphants showed increased locomotor activity com-pared with 5mis MO-injected control larvae (Fig. 9A). The amigo1mRNA-MO1- or mRNA-MO2-coinjected larvae did not show asignificant difference in locomotor activity compared with the5mis MO-injected control larvae (Fig. 9A). These results revealed

that coinjection of the amigo1 mRNA with the MOs rescued thehyperactivity observed in both types of amigo1 morphants andthat the increase in locomotor activity is a trait specific to theamigo1 knockdown. Interestingly, coinjection of the kv2.1 mRNAwith MO1 or MO2 also rescued the hyperactivity phenotype of themorphants (Fig. 9A).

Our immunohistochemical analysis showed that Amigo1is expressed in Mauthner neuron axons (early axonal tractsthat are crucial for the functions of the sensorimotor cir-cuits), and the MO experiments indicated that Amigo1 reg-ulates development of connections from the Mauthner neu-rons in vivo (Fig. 6B, bottom panels). Based on previousreports, we estimated that amigo1 knockdown could there-fore perturb startle motor activity (23, 24). Therefore, wedecided to study the behavioral repertoire further and inves-

FIGURE 7. Western blotting of larval lysates showing decreased development of neuronal connections in amigo1 knockdown morphants (MO1 andMO2) and in larvae expressing the Amigo1 ectodomain as an inhibitor of the endogenous protein (Am1 EmRNA). A, the left panel shows Westernblotting of 3 dpf larval lysates using anti-HNK-1 and anti-acetylated tubulin antibodies as pathway markers. Anti-�-actin antibodies were used for normalizingthe loading amount of each sample. The amigo1 knockdown morphants (MO1 and MO2) and the amigo1 EmRNA-injected larvae (Am1 EmRNA) show muchweaker bands than the rescued groups (MO1 � Am1 mRNA and MO2 � Am1 mRNA) and the control groups (the 5mis MO-injected, the GFP mRNA-injectedand the uninjected controls). The right panel shows the corresponding Western blotting analysis, where the rescue effect of the kv2.1 mRNA is compared withthat of the amigo1 mRNA. In addition, expression of Amigo1 in morpholino-injected and in morpholino/mRNA-coinjected larvae is shown. B–D, quantificationof the results shown in A. In B–D, the plot density of the bands from the uninjected larvae was set as 1. The experiments were repeated four times with larvaefrom four independent experiments. Each SDS extract contained 8 –10 larvae selected randomly from the pool of 50 larvae. The statistical significance wasconfirmed by comparison with the 5mis MO-injected larvae. All graphs show mean values � S.E. (error bars) (n � 4; **, p � 0.001, one-way ANOVA followed byTukey’s post hoc test).




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

FIGURE 8. Whole mount immunostaining with anti-TH antibodies showing disturbed development of catecholaminergic circuits in amigo1 mor-phants. A, maximum projection of confocal image stacks of 3-dpf whole larval heads. The MLCT staining and its absence in the amigo1 knockdown morphants(MO1 and MO2) is indicated by an arrow. The lateral catecholaminergic projections (lcp), originally lateral to the MLCT in the region between the locus coeruleus(LC) and medulla oblongata (MO), also show a perturbed pattern in the MO1 and MO2 morphants compared with the controls (labeled by an asterisk). Theamigo1 but not the kv2.1 mRNA displays a rescue effect on the pathway development. The schematic map shows the normal pattern of anti-TH staining in 3-dpflarvae. The rectangular area was scanned in high resolution and is shown in B. AP, area postrema; DC, diencephalic dopaminergic clusters; lcp, lateral cat-echolaminergic projections; MO, medulla oblongata; OB, olfactory bulb; Po, preoptic region; Tel, telencephalon. Scale bar, 150 �m. B, Z-stacking confocalimages scanned under higher magnification focusing on MLCT ascending and descending from the locus coeruleus in 3-dpf larvae. The amigo1 MO1-injectedmorphant shows much less staining of MLCT than the 5mis MO-injected control (indicated with an arrow). Scale bar, 40 �m. C, quantification of the MLCTdefects in 3-dpf larvae. Over 65% of the MO1 morphants and 80% of the MO2 morphants display defects in the MLCT development. In the amigo1 full-lengthmRNA-rescued larvae (MO1 � Am1 mRNA and MO2 � Am1 mRNA), the proportion of larvae with MLCT defects is decreased to only 20 –30%. In theMO1/Kv2.1-coinjected group (MO1 � Kv2.1 mRNA), around 80% of larvae show MLCT defects. For each group, 10 larvae were picked randomly from the poolof 50 larvae for whole mount immunostaining in three independent experiments. Mean values � S.E. (error bars) are indicated (n � 3; **, p 0.01, one-wayANOVA followed by Tukey’s post hoc test). D, average pixel intensity analyses of anti-TH1-stained MLCT from Z-stacking confocal images. The average pixelintensity of MLCT staining in uninjected larvae was used for normalization. The MO1, MO2, and MO1 � Kv2.1 morphants show significantly decreased stainingintensity of MLCT compared with the 5mis MO-injected control groups. The normalized average pixel intensity of MLCT in the MO1 morphants is only about40%, and in the MO2 and the MO1 � Kv2.1 mRNA morphants, the staining intensity is even lower than 20% of that found in the control groups. The data arebased on five larvae picked randomly for intensity analyses in each group. The experiment was repeated three times (n � 15; **, p 0.001, one-way ANOVAfollowed by Tukey’s post hoc test).




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

tigated the startle response of the amigo1 knockdown mor-phants elicited by electrical stimuli (10 ms, 5 V) (23, 44). Theamigo1 MO1 and MO2 morphants had significant defects inperforming both SLC and LLC responses (Fig. 9, B and C). Incontrast, the 5mis MO-injected larvae performed C-bendwithin 20 ms after the stimulation, eliciting an escaperesponse (Fig. 9C, compare the top and bottom rows). Inlarvae coinjected with amigo1 mRNA-MO1 or amigo1mRNA-MO2, the SLC and LLC were rescued to the levelfound in the 5mis MO-injected larvae (Fig. 9B). Similar res-cue effects were observed in the larvae coinjected withkv2.1 mRNA-MO1 or kv2.1 mRNA-MO2 (Fig. 9B). Takentogether, our results demonstrate that transient knockdownof amigo1 in vivo significantly affects the formation offunctionally mature neuronal circuits that are necessary fordriving essential behaviors in an intact vertebrate modelorganism.

DISCUSSION

Amigo Protein Family in Zebrafish—To study the role of theAmigo protein family in the development of functionally activeneural circuitry in vivo, we decided to use the zebrafish model inthe current study because development of the neuronal connec-tions is more readily analyzed in the zebrafish compared with themouse. Furthermore, the current results show that the parentform of the protein family (Amigo1) is the main family memberexpressed until 5 dpf, when several neural circuits, such as thosecontrolling locomotion, have already been formed in zebrafish. Inaddition, zebrafish does not have a protein that would be clearlysimilar to AMIGO2 that displays a high expression level in rodentbrain and interacts with neurons in a similar manner as AMIGO1.The risk of redundancy should therefore be lower in zebrafishcompared with mice, and we expected that knocking down justone family member might produce phenotypic changes.

FIGURE 9. Altered motor functions in amigo1 morphants. A, locomotor activity in 6-dpf larvae during 10 min. The swimming distance was calculated for eachlarva. The MO1 and MO2 morphants display significantly higher activity compared with the 5mis MO-injected or the uninjected control groups; their swimmingdistance is enhanced by more than 40% (160 mm). The amigo1 mRNA-coinjected groups (MO1 � Am1 mRNA and MO2 � Am1 mRNA) and the kv2.1mRNA-coinjected groups (MO1 � kv2.1 mRNA and MO2 � kv2.1 mRNA) do not show a significant difference from the 5mis MO-injected or the uninjectedcontrol groups. 10 larvae of each group were selected randomly for the tests, which were repeated four times in independent experiments. The significancewas confirmed by one-way ANOVA followed by Fisher’s least significant difference post hoc test and Bonferroni correction; n � 40; *, p 0.05. B, electricallyelicited C-start response test in 9-dpf larvae. Under 5 V electrical stimulus for 10 ms, over 90% of the larvae in the uninjected group show C-start response, 40%with SLC and 50% with LLC. The MO1 and MO2 morphants show significantly decreased SLC and LLC compared with the 5mis MO-injected or the uninjectedcontrol groups. The amigo1 mRNA-coinjected groups (MO1 � Am1 mRNA and MO2 � Am1 mRNA) and the kv2.1 mRNA-coinjected groups (MO1 � kv2.1mRNA and MO2 � kv2.1 mRNA) perform similarly to the control groups. There were 20 larvae in each group selected randomly for three tests, which wererepeated four times independently. The significance was confirmed by one-way ANOVA followed by Fisher’s LSD post hoc test and Bonferroni correction; n �12; **, p 0.01. C, real-time recordings of the C-start response. The 5mis MO-injected and the MO1-injected 9-dpf larvae are shown. Pictures were taken from5 ms after the voltage stimuli. The recording frames are shown every 5 ms and marked every 10 ms. The C-start response captured in the 5-ms MO larvae isdefined as the SLC. Error bars, S.E.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

To follow the expression of the Amigo1 protein, we haveproduced in the current study affinity-purified antibodies thatspecifically detect the zebrafish Amigo1 in immunohistochem-istry and Western blotting. In 28-hpf larvae, Amigo1 is mainlyexpressed in the anterior dorsal part of the forebrain area,resembling the expression in embryonic day 10 (E10) mouseembryos (6). In 5-dpf zebrafish larvae, Amigo1 expression isfound in most brain areas that already have functionally activeneural circuits. This finding also resembles the situation inrodents, in which AMIGO1 expression is increased in postnatalbrain during construction of the adult-type circuitry and isfound in practically all areas of the CNS in most neurons andtheir fibers in young adults (4, 7, 9). Such a conserved spatio-temporal expression pattern in vertebrates suggests a con-served function in the CNS development.

Morphological Phenotypes Caused by Down-regulation ofAmigo1 Expression or Inhibition of Its Function—Analysis ofknockdowns produced by two different morpholino oligonu-cleotides (MO1 and MO2) shows that the Amigo1 proteinessentially disappears, which causes striking defects in the for-mation of early fiber scaffolds. For example, MLF, which is oneof the major early fiber scaffolds, is only composed of thin fiberscompared with the robust fascicles seen in uninjected or 5misMO- injected embryos. Because MOs can sometimes causeunexpected off-target effects, we have used as an alternativeapproach, injection of the mRNA encoding the protein ectodo-main. This approach is similar to the one used in our previous invitro experiments, where the protein ectodomain was shown toact as a dominant negative receptor that inhibits AMIGO-me-diated adhesion and fasciculation of neurites in brain neurons(4). As in studies in vitro, the dominant negative approachcauses inhibition of fasciculation in zebrafish in vivo. Further-more, coinjection of the full-length amigo1 mRNA with theMOs or with the dominant negative construct causes statisti-cally significant rescue in fiber tract development.

It is noteworthy that in addition to the morphological analy-sis, defective development of neuronal connections in theamigo1 knockdown morphants is also suggested by Westernblotting of larval SDS extracts using anti-HNK1 and anti-acety-lated tubulin antibodies as fiber pathway markers. Acetylatedtubulin has been shown to be a marker of most if not all earlyaxons (32), and the anti-HNK-1 antibodies have been shown tolabel fiber pathways in a similar manner as the anti-acetylatedtubulin antibodies (41). These experiments provide further evi-dence of the role for Amigo1 in construction of fiber pathwaysduring early development.

Analysis of the amigo1 knockdown morphants at later devel-opmental stages does not show very prominent defects in spe-cific brain structures. This finding is consistent with the findingthat the knockdown morphants do not display changes inexpression of regional transcription factors, such as Pax2a,Pax6a, and Krox20, that are known to be important for braindevelopment. Analysis at the stages 3–5 dpf, however, displaysa modest reduction in head size and disordered development ofthe aminergic system. This is specifically demonstrated for thecatecholaminergic system, but reduced immunostaining of the

serotonergic system also appears obvious.4 It seems possiblethat the defects in the early fiber scaffold development generallydisturb development of the aminergic connections, but we can-not exclude specific effects of Amigo1 on the development ofneurotransmitter-specific connections.

Behavioral Phenotypes in amigo1 Knockdown Morphants—Disturbed development of neuronal connections observed inthe amigo1 knockdown morphants might cause behavioralalterations. Unexpectedly, the amigo1 knockdowns were foundto display increased locomotor activity. However, this pheno-type is specific because coinjection of the amigo1 mRNA withthe morpholinos reduces the enhanced locomotion to the levelseen in the misMO-injected and the uninjected larvae.

In addition to spontaneous locomotor activity, we haveassessed the escape behavior in the amigo1 morphants. In thisanalysis, we have used electrical stimulation that has been usedto elicit startle responses in adult and larval fish in a range ofstudies (23, 44). The amigo1 knockdowns display clearly defec-tive responses in both SLC response and LLC response. As inenhanced locomotor activity, the amigo1 mRNA displays aclear rescue effect in both SLC and LLC. It has been reportedthat the SLC response is mainly mediated by Mauthner neurons(45, 46), whereas the LLC response is not mediated by theseneurons (24, 46). Defective development of the Mauthner neu-ron decussating axons that enter the contralateral MLF (47)may therefore contribute to the defective SLC response. How-ever, a widespread escape network has been identified in thebrain stem of zebrafish, containing several descending fiberpathways (46). Maturation of this escape network to form func-tional connections from the brain to the spinal cord probablyregulates the escape responses in zebrafish and is compromisedin the amigo1 morphants.

Homophilic and Heterophilic Interactions of Amigo1 in theDevelopment of Neural Circuitry—Analysis of fiber pathwaydevelopment based on their staining intensity reveals clearlydefective tract development in the knockdown morphants andin larvae injected with the mRNA encoding the Amigo1 ectodo-main. This suggests that tract development is specifically tar-geted in the experiments. The finding that the numbers ofTUNEL-positive cells are not increased in the experimentsagrees with this interpretation and suggests that compromisedneuron survival is not the reason why defective tract develop-ment is observed.

The mechanism through which Amigo1 specifically affectstract development therefore needs to be considered. Until now,the Robo/Slit signaling pathway has been reported to be crucialfor the regulation of longitudinal axonal pathfinding (48). Thedefects in the development of the early long tracts observed inthe amigo1 knockdown morphants appear even more pro-nounced compared with those reported for manipulation ofRobo/Slit signaling. It appears that the role of Amigo1 cannotbe explained by Slit/Robo signaling because the phenotypicchanges in the amigo1 morphants are different from thosefound in Slit/Robo studies, and we have not found such changesin Slit or Robo expression that could explain our findings.4

4 X. Zhao, P. Panula, and H. Rauvala, unpublished observations.




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

We propose that homophilic binding of Amigo1 within fibertracts underlies at least partially its remarkable role in the devel-opment of long tracts observed in the current study. This infer-ence is supported by our previous studies in which homophilicbinding of the Amigo proteins has been demonstrated usingpull-down experiments of labeled proteins from cells, beadassays demonstrating aggregation through AMIGO-AMIGOinteractions (4), and identification of the LRR region mediatingdimerization in the crystal structure (8). Furthermore, inhibi-tion of fasciculation by the Amigo1 ectodomain observed in thecurrent study is strikingly similar to the fasciculation inhibitionfound in isolated neurons where homophilic interactions andheterophilic interactions within the AMIGO protein familymediate binding to brain neurons (4).

Homophilic interactions within fiber tracts are well knownfor their essential roles in growth and guidance of axons insimple systems, such as Drosophila, but their roles in complexvertebrate tracts are less clear. However, it has been shown thatthe zebrafish retinotectal system uses such “isotypic” interac-tions as an essential strategy for growth and guidance of axontracts (48, 49). It appears that this fundamental cellular mech-anism, acting through pioneer-follower interactions and peer-peer interactions (community effect), is as important as guid-ance signals outside the tracts and may be used throughout thedevelopment of complex nervous systems in vertebrates. Fur-ther work on the Amigo family members in different neuralsystems is required to characterize Amigo-mediated homo-philic interactions and cell signaling following such interac-tions. For example, Amigo-mediated homophilic interactionsmight affect development of neurotransmitter-specific pheno-types or expression of ion channels required in neural signaling.

We have previously shown that AMIGO1 displays hetero-philic interactions with the Kv2.1 potassium channel in brainneurons (11). This inference is based on the finding that the twoproteins show a striking colocalization in specific clusters at theplasma membrane. Furthermore, when the Kv2.1 clusters aredispersed at the plasma membrane upon the addition of variousstimuli, such as glutamate, AMIGO1 is dispersed with theKv2.1 channel. These findings, temporal coexpression of theproteins in developing mouse brain and co-immunoprecipita-tion experiments, have led us to propose that AMIGO1 shouldbe regarded as an auxiliary subunit of the Kv channel (11). Fromthe functional viewpoint, AMIGO1 regulates the channel activ-ity at voltage values close to those required to start actionpotentials and is therefore expected to play a role in neuronalsignaling.

We have therefore considered whether interactions ofAmigo1 with Kv2.1 might play a role in the amigo1 morphantphenotypes found in the current study. Surprisingly, theamigo1 morphants were found to lack the Kv2.1 channel almostcompletely. Expression of the channel in the amigo1 morphantscould be restored by coinjection of the kv2.1 mRNA and to alesser extent by coinjection of the amigo1 mRNA in the mor-pholino experiments. It thus appears clear that Amigo1 doesnot only bind to the channel and regulate its activity at theplasma membrane but, in addition, strongly influences Kv2.1expression in vivo. It is possible that the Kv2.1 protein is notstable in the absence of its auxiliary subunit Amigo1, but fur-

ther studies will be required to elucidate the biochemical basisof the regulation.

Because the expression of Kv2.1 is strongly inhibited in ouramigo1 morphants, we attempted to rescue the phenotypes byrestoring the Kv2.1 protein expression through coinjecting thekv2.1 mRNA with the morpholino oligonucleotides. The kv2.1mRNA did not cause any rescue in the defects of early (1–5 dpf)tract development. In contrast, clear rescue effects were foundin the behavioral phenotypes that obviously depend on func-tionally active neural circuitries. Sufficient anatomical connec-tivity therefore seems to remain in the morphants to performnormal sensorimotor functions in the case where Kv2.1 isexpressed. Interestingly, Amigo1 and Kv2.1 were found to colo-calize in mature brain but not in early developing brain,although they were found to be expressed in overlapping ana-tomical areas in both cases.

During the preparation of this manuscript, a study on Kv2.1knock-out mice appeared showing that deficiency of the Kv2.1channel leads to neuronal and behavioral hyperexcitability (50).General locomotor activity in the Kv2.1 knock-out mice was foundto be strongly enhanced, clearly resembling the enhanced activityin the amigo1 morphants that we had found in the current study.Because hyperactivity and defective startle responses in theamigo1 knockdown morphants can be rescued by the kv2.1mRNA, it appears clear that defective Kv2.1 channel functionstrongly contributes to the behavioral phenotypes caused byknockdown of Amigo 1 expression.

In conclusion, Amigo1 functions independently of Kv2.1during early fiber tract development. For the construction ofneural circuitry with adult-type normal functions, AMIGO1regulates the expression and function of the Kv2.1 potassiumchannel of central neurons.

Acknowledgments—The technical assistance of Henri Koivula, ReetaHuhtala, and Erja Huttu is gratefully acknowledged.

REFERENCES1. de Wit, J., Hong, W., Luo, L., and Ghosh, A. (2011) Role of leucine-rich

repeat proteins in the development and function of neural circuits. Annu.Rev. Cell Dev. Biol. 27, 697–729

2. Sollner, C., and Wright, G. J. (2009) A cell surface interaction network ofneural leucine-rich repeat receptors. Genome Biol. 10, R99

3. Simpson, J. H., Bland, K. S., Fetter, R. D., and Goodman, C. S. (2000)Short-range and long-range guidance by Slit and its Robo receptors: acombinatorial code of Robo receptors controls lateral position. Cell 103,1019 –1032

4. Kuja-Panula, J., Kiiltomaki, M., Yamashiro, T., Rouhiainen, A., and Rau-vala, H. (2003) AMIGO, a transmembrane protein implicated in axontract development, defines a novel protein family with leucine-rich re-peats. J. Cell Biol. 160, 963–973

5. Maness, P. F., and Schachner, M. (2007) Neural recognition molecules ofthe immunoglobulin superfamily: signaling transducers of axon guidanceand neuronal migration. Nat. Neurosci. 10, 19 –26

6. Homma, S., Shimada, T., Hikake, T., and Yaginuma, H. (2009) Expressionpattern of LRR and Ig domain-containing protein (LRRIG protein) in theearly mouse embryo. Gene Expr. Patterns 9, 1–26

7. Laeremans, A., Nys, J., Luyten, W., D’Hooge, R., Paulussen, M., and Arck-ens, L. (2013) AMIGO2 mRNA expression in hippocampal CA2 andCA3a. Brain Struct. Funct. 218, 123–130

8. Kajander, T., Kuja-Panula, J., Rauvala, H., and Goldman, A. (2011) Crystalstructure and role of glycans and dimerization in folding of neuronal leu-




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

cine-rich repeat protein AMIGO-1. J. Mol. Biol. 413, 1001–10159. Chen, Y., Hor, H. H., and Tang, B. L. (2012) AMIGO is expressed in

multiple brain cell types and may regulate dendritic growth and neuronalsurvival. J. Cell Physiol. 227, 2217–2229

10. Ono, T., Sekino-Suzuki, N., Kikkawa, Y., Yonekawa, H., and Kawashima, S.(2003) Alivin 1, a novel neuronal activity-dependent gene, inhibits apo-ptosis and promotes survival of cerebellar granule neurons. J. Neurosci. 23,5887–5896

11. Peltola, M. A., Kuja-Panula, J., Lauri, S. E., Taira, T., and Rauvala, H. (2011)AMIGO is an auxiliary subunit of the Kv2.1 potassium channel. EMBORep. 12, 1293–1299

12. Zhao, X., Kuja-Panula, J., Rouhiainen, A., Chen, Y. C., Panula, P., andRauvala, H. (2011) High mobility group box-1 (HMGB1; amphoterin)is required for zebrafish brain development. J. Biol. Chem. 286,23200 –23213

13. Westerfield, M. (2000) The Zebrafish Book: A Guide for the Laboratory Useof Zebrafish (Danio rerio), 4th Ed., University of Oregon Press, Eugene, OR

14. Fink, M., Flekna, G., Ludwig, A., Heimbucher, T., and Czerny, T. (2006)Improved translation efficiency of injected mRNA during early embryonicdevelopment. Dev. Dyn. 235, 3370 –3378

15. Thisse, C., and Thisse, B. (2008) High-resolution in situ hybridization towhole-mount zebrafish embryos. Nat. Protoc. 3, 59 – 69

16. Thisse, C., and Thisse, B. (2005) High throughput expression analysisof ZF-Models Consortium clones, ZFIN direct data submission(http://zfin.org)

17. Sallinen, V., Torkko, V., Sundvik, M., Reenila, I., Khrustalyov, D., Kaslin, J.,and Panula, P. (2009) MPTP and MPP� target specific aminergic cellpopulations in larval zebrafish. J. Neurochem. 108, 719 –731

18. Rivera, J., Chu, P. J., Lewis, T. L., Jr., and Arnold, D. B. (2007) The role ofKif5B in axonal localization of Kv1 K� channels. Eur. J. Neurosci. 25,136 –146

19. Mueller, T., and Wullimann, M. F. (2005) Atlas of Early Zebrafish BrainDevelopment: A Tool for Molecular Neurogenetics, pp. 29 –120, Elsevier,Amsterdam

20. Sallinen, V., Sundvik, M., Reenila, I., Peitsaro, N., Khrustalyov, D., Anicht-chik, O., Toleikyte, G., Kaslin, J., and Panula, P. (2009) Hyperserotonergicphenotype after monoamine oxidase inhibition in larval zebrafish. J. Neu-rochem. 109, 403– 415

21. Schweitzer, J., Lohr, H., Filippi, A., and Driever, W. (2012) Dopaminergicand noradrenergic circuit development in zebrafish. Dev. Neurobiol. 72,256 –268

22. Panula, P., Sallinen, V., Sundvik, M., Kolehmainen, J., Torkko, V., Tiittula,A., Moshnyakov, M., and Podlasz, P. (2006) Modulatory neurotransmittersystems and behavior: towards zebrafish models of neurodegenerative dis-eases. Zebrafish 3, 235–247

23. Nissanov, J., Eaton, R. C., and DiDomenico, R. (1990) The motor output ofthe Mauthner cell, a reticulospinal command neuron. Brain Res. 517,88 –98

24. Burgess, H. A., and Granato, M. (2007) Sensorimotor gating in larval ze-brafish. J. Neurosci. 27, 4984 – 4994

25. Eisen, J. S., and Smith, J. C. (2008) Controlling morpholino experiments:don’t stop making antisense. Development 135, 1735–1743

26. Wullimann, M. F., and Puelles, L. (1999) Postembryonic neural prolif-eration in the zebrafish forebrain and its relationship to prosomericdomains. Anat. Embryol. 199, 329 –348

27. Wullimann, M. F., and Mueller, T. (2002) Expression of Zash-1a in thepostembryonic zebrafish brain allows comparison to mouse Mash1 do-mains. Brain Res. Gene. Expr. Patterns 1, 187–192

28. Mueller, T. (2012) What is the thalamus in zebrafish? Front. Neurosci.6, 64

29. Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullmann, B., and Schilling,T. F. (1995) Stages of embryonic development of the zebrafish. Dev. Dyn.203, 253–310

30. McLean, D. L., and Fetcho, J. R. (2004) Ontogeny and innervation patternsof dopaminergic, noradrenergic, and serotonergic neurons in larval ze-

brafish. J. Comp. Neurol. 480, 38 –5631. Cole, L. K., and Ross, L. S. (2001) Apoptosis in the developing zebrafish

embryo. Dev. Biol. 240, 123–14232. Chitnis, A. B., and Kuwada, J. Y. (1990) Axonogenesis in the brain of

zebrafish embryos. J. Neurosci. 10, 1892–190533. Wilson, S. W., Ross, L. S., Parrett, T., and Easter, S. S., Jr. (1990) The

development of a simple scaffold of axon tracts in the brain of the embry-onic zebrafish, Brachydanio rerio. Development 108, 121–145

34. Kruse, J., Mailhammer, R., Wernecke, H., Faissner, A., Sommer, I., Goridis,C., and Schachner, M. (1984) Neural cell adhesion molecules and myelin-associated glycoprotein share a common carbohydrate moiety recognizedby monoclonal antibodies L2 and HNK-1. Nature 311, 153–155

35. Metcalfe, W. K., Myers, P. Z., Trevarrow, B., Bass, M. B., and Kimmel, C. B.(1990) Primary neurons that express the L2/HNK-1 carbohydrate duringearly development in the zebrafish. Development 110, 491–504

36. Tay, T. L., Ronneberger, O., Ryu, S., Nitschke, R., and Driever, W. (2011)Comprehensive catecholaminergic projectome analysis reveals single-neuron integration of zebrafish ascending and descending dopaminergicsystems. Nat. Commun. 2, 171

37. McLean, D. L., and Fetcho, J. R. (2004) Relationship of tyrosine hydroxy-lase and serotonin immunoreactivity to sensorimotor circuitry in larvalzebrafish. J. Comp. Neurol. 480, 57–71

38. Kastenhuber, E., Kern, U., Bonkowsky, J. L., Chien, C. B., Driever, W., andSchweitzer, J. (2009) Netrin-DCC, Robo-Slit, and heparan sulfate pro-teoglycans coordinate lateral positioning of longitudinal dopaminergic di-encephalospinal axons. J. Neurosci. 29, 8914 – 8926

39. Peters, A., and Vaughn, J. E. (1967) Microtubules and filaments in theaxons and astrocytes of early postnatal rat optic nerves. J. Cell Biol. 32,113–119

40. Piperno, G., and Fuller, M. T. (1985) Monoclonal antibodies specific for anacetylated form of �-tubulin recognize the antigen in cilia and flagellafrom a variety of organisms. J. Cell Biol. 101, 2085–2094

41. Ross, L. S., Parrett, T., and Easter, S. S., Jr. (1992) Axonogenesis and mor-phogenesis in the embryonic zebrafish brain. J. Neurosci. 12, 467– 482

42. Kastenhuber, E., Kratochwil, C. F., Ryu, S., Schweitzer, J., and Driever, W.(2010) Genetic dissection of dopaminergic and noradrenergic contribu-tions to catecholaminergic tracts in early larval zebrafish. J. Comp. Neurol.518, 439 – 458

43. Rink, E., and Wullimann, M. F. (2002) Development of the catecholamin-ergic system in the early zebrafish brain: an immunohistochemical study.Brain Res. Dev. Brain Res. 137, 89 –100

44. Liu, Y. C., Bailey, I., and Hale, M. E. (2012) Alternative startle motor pat-terns and behaviors in the larval zebrafish (Danio rerio). J. Comp. Physiol.A Neuroethol. Sens. Neural Behav. Physiol. 198, 11–24

45. Kimmel, C., Eaton, R., and Powell, S. (1980) Decreased fast-start perform-ance of zebrafish larvae lacking mauthner neurons. J. Comp. Physiol. 140,343–350

46. Gahtan, E., Sankrithi, N., Campos, J. B., and O’Malley, D. M. (2002) Evi-dence for a widespread brain stem escape network in larval zebrafish.J. Neurophysiol. 87, 608 – 614

47. Kimmel, C. B., Powell, S. L., and Metcalfe, W. K. (1982) Brain neuronswhich project to the spinal cord in young larvae of the zebrafish. J. Comp.Neurol. 205, 112–127

48. Dickson, B. J., and Gilestro, G. F. (2006) Regulation of commissural axonpathfinding by slit and its Robo receptors. Annu. Rev. Cell Dev. Biol. 22,651– 675

49. Pittman, A. J., Law, M. Y., and Chien, C. B. (2008) Pathfinding in a largevertebrate axon tract: isotypic interactions guide retinotectal axons atmultiple choice points. Development 135, 2865–2871

50. Speca, D. J., Ogata, G., Mandikian, D., Bishop, H. I., Wiler, S. W., Eum, K.,Wenzel, H. J., Doisy, E. T., Matt, L., Campi, K. L., Golub, M. S., Nerbonne,J. M., Hell, J. W., Trainor, B. C., Sack, J. T., Schwartzkroin, P. A., andTrimmer, J. S. (2014) Deletion of the Kv2.1 delayed rectifier potassiumchannel leads to neuronal and behavioral hyperexcitability. Genes BrainBehav. 13, 394 – 408




ww

.jbc.org/D

ownloaded from

http://www.jbc.org/

A. Peltola, Tarja Porkka-Heiskanen, Pertti Panula and Heikki RauvalaXiang Zhao, Juha Kuja-Panula, Maria Sundvik, Yu-Chia Chen, Vilma Aho, Marjaana

BrainAmigo Adhesion Protein Regulates Development of Neural Circuits in Zebrafish

doi: 10.1074/jbc.M113.545582 originally published online June 5, 20142014, 289:19958-19975.J. Biol. Chem.

10.1074/jbc.M113.545582Access the most updated version of this article at doi:

Alerts:

When a correction for this article is posted•

When this article is cited•

to choose from all of JBC's e-mail alertsClick here

http://www.jbc.org/content/289/29/19958.full.html#ref-list-1

This article cites 47 references, 14 of which can be accessed free at


ww

.jbc.org/D

ownloaded from

http://www.jbc.org/lookup/doi/10.1074/jbc.M113.545582

http://www.jbc.org/cgi/alerts?alertType=citedby&addAlert=cited_by&cited_by_criteria_resid=jbc;289/29/19958&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/289/29/19958

http://www.jbc.org/cgi/alerts?alertType=correction&addAlert=correction&correction_criteria_value=289/29/19958&saveAlert=no&return-type=article&return_url=http://www.jbc.org/content/289/29/19958

http://www.jbc.org/cgi/alerts/etoc

http://www.jbc.org/content/289/29/19958.full.html#ref-list-1

http://www.jbc.org/

amigo adhesion protein regulates development of neural circuits

Documents