Jour

nal o

f Cel

l Sci

ence

RESEARCH ARTICLE

Papp-a2 modulates development of cranial cartilage andangiogenesis in zebrafish embryos

Kasper Kjaer-Sorensen, Ditte H. Engholm*, Malene R. Jepsen*, Maria G. Morch, Kathrin Weyer,Louise L. Hefting, Louise L. Skov, Lisbeth S. Laursen and Claus Oxvig`

ABSTRACT

Pregnancy-associated plasma protein A2 (PAPP-A2, also known as

pappalysin-2) is a large metalloproteinase that is known to be

required for normal postnatal growth and bone development in

mice. We here report the detection of zebrafish papp-a2 mRNA in

the chordamesoderm, notochord and lower jaw of zebrafish (Danio

rerio) embryos, and that papp-a2-knockdown embryos display

broadened axial mesoderm, notochord bends and severely reduced

cranial cartilages. Genetic data link these phenotypes to insulin-

like growth factor (Igf)-binding protein-3 (Igfbp-3) and bone

morphogenetic protein (Bmp) signaling, and biochemical analysis

show specific Igfbp-3 proteolysis by Papp-a2, implicating Papp-a2

in the modulation of Bmp signaling by Igfbp-3 proteolysis.

Knockdown of papp-a2 additionally resulted in angiogenesis

defects, strikingly similar to previous observations in embryos with

mutations in components of the Notch system. Accordingly, we find

that Notch signaling is modulated by Papp-a2 in vivo, and,

furthermore, that human PAPP-A2 is capable of modulating Notch

signaling independently of its proteolytic activity in cell culture.

Based on these results, we conclude that Papp-a2 modulates Bmp

and Notch signaling by independent mechanisms in zebrafish

embryos. In conclusion, these data link pappalysin function in

zebrafish to two different signaling pathways outside the IGF

system.

KEY WORDS: Zebrafish, Papp-a2, Pappa2, Igfbp-3, Bmp, Notch

INTRODUCTIONThe insulin-like growth factors I and II (IGF1 and IGF2), are

potent mitogenic polypeptides with important regulatory roles

in cellular processes, including mitosis, cell migration and

apoptosis, which are crucial for embryonic development (De

Meyts et al., 2004; Jones and Clemmons, 1995). In the

extracellular environment, the vast majority of IGF is found in

1:1 complexes with one of six IGF-binding proteins (IGFBPs),

which prevent bound IGF from activating the IGF receptor

(IGF1R). Tight regulation of IGF bioavailability for IGF1R

activation is obtained by specific proteolytic cleavage of IGFBPs

resulting in cleavage fragments of reduced IGF affinity (Forbes

et al., 2012). The metalloproteinase pregnancy-associated plasma

protein-A (PAPP-A, also known as pappalysin-1) is known to

cleave IGFBP-4 and -5 in vitro (Gyrup and Oxvig, 2007; Laursen

et al., 2001; Lawrence et al., 1999), and several lines of evidence

suggest that PAPP-A and IGFBP-4 function as an interdependent

pair to increase IGF1R activation (Conover et al., 2004; Laursen

et al., 2007; Ning et al., 2008). Limited information is available

on PAPP-A2, the only homolog of PAPP-A in the vertebrate

genome. However, biochemical evidence indicates that PAPP-A2

modulates IGF bioavailability by specific proteolytic cleavage of

IGFBP-3 and -5 (Overgaard et al., 2001). Accordingly, PAPP-A2

was recently identified as the protease responsible for IGFBP-5

cleavage in human pregnancy serum (Yan et al., 2010).

The IGF system is well conserved between vertebrate species

(Duan and Xu, 2005). In the zebrafish, orthologs of the IGFs, the

IGF receptors, and IGFBPs have been identified, and the system

has been characterized in several publications (e.g. Dai et al.,

2010; Li et al., 2005; Maures et al., 2002; White et al., 2009; Zou

et al., 2009). It is of particular importance for the current study

that a zebrafish ortholog of PAPP-A has been shown to cleave

human IGFBP-4 and IGFBP-5 (Kjaer-Sorensen et al., 2013), and

that the zebrafish ortholog of human bone morphogenetic protein-

1 (BMP1), an IGFBP-3 protease, is capable of cleaving zebrafish

Igfbp-3. Furthermore, overexpression of Bmp1 reverses the

phenotype of Igfbp-3 overexpression in zebrafish embryos (Kim

et al., 2011). These findings suggest that IGFBP proteolysis as a

regulatory mechanism is conserved in zebrafish.

PAPP-A2 is a large homodimeric multi-domain protein

encompassing a proteolytic domain, three Lin12-Notch repeats,

five complement control protein modules, and a laminin-G-like

domain (Overgaard et al., 2001). The proteolytic domain displays

the characteristics of the metzincin metalloproteinases, including

the elongated zinc-binding consensus sequence and a Met-turn,

and together with ulilysin, PAPP-A and PAPP-A2 constitute the

pappalysin family, within the metzincin superfamily (Boldt et al.,

2001; Tallant et al., 2006). In PAPP-A, the Lin12-Notch repeats

are known to regulate proteolytic specificity by substrate

interaction (Boldt et al., 2004; Mikkelsen et al., 2008).

PAPP-A2-knockout mice show no significant difference in

body mass pre- and perinatally but exhibit postnatal growth

deficiency beginning at the age of 3 weeks. The knockout mice

also display disproportionally reduced dimensions of specific

bones, including skull and mandible, with size reductions being

more severe than would be expected from the decrease in total

body mass and length (Christians et al., 2013; Conover et al.,

2011). Whereas similar phenotypes of the skeleton were not

reported in IGFBP-3- and IGFBP-5-knockout, or transgenic mice

overexpressing IGFBP-3 or IGFBP-5 (Ning et al., 2006; Salih

et al., 2005; Silha et al., 2003), addition of purified IGFBP-5

to metatarsal bones in culture reduced longitudinal growth

Department of Molecular Biology and Genetics, Aarhus University, Gustav WiedsVej 10C, DK-8000 Aarhus C, Denmark.*These authors contributed equally to this work

`Author for correspondence (co@mb.au.dk)

Received 4 March 2014; Accepted 8 September 2014

� 2014. Published by The Company of Biologists Ltd | Journal of Cell Science (2014) 127, 5027–5037 doi:10.1242/jcs.152587

5027

Jour

nal o

f Cel

l Sci

ence

(Mukherjee and Rotwein, 2008). Furthermore, knockdown ofigfbp-3 in zebrafish larvae has been reported to cause reduction in

the cartilage scaffolds of the jaw and neurocranium (Li et al.,2005).

To explore the function of Papp-a2 in zebrafish, we havecharacterized embryonic and adult papp-a2 expression, determined

the proteolytic specificity of zebrafish Papp-a2, and performedtargeted knockdown in zebrafish embryos. Expression patterns andbiochemical properties are conserved between human PAPP-A2

and zebrafish Papp-a2, suggesting that the zebrafish is a suitablemodel for the study of Papp-a2 function in vertebrate development.We find that Papp-a2 is required for normal development of cranial

cartilages, for normal development of the notochord and for normalangiogenesis. Further analyses suggest that Papp-a2 exerts separatefunctions by specific Igfbp proteolysis and by non-proteolytic

modulation of Notch signaling.

RESULTSSequence analysisA homology search of the zebrafish genome using human PAPP-A2 (UniProt entry Q9BXP8) as query identified a single papp-a2

gene located on chromosome 2, and synteny analysis suggests

orthology to human PAPP-A2 (Fig. 1A). The prepro form ofzebrafish Papp-a2 showed 50% and 43% overall sequence identityto human PAPP-A2 and PAPP-A, respectively. Protein modules

that have known function in pappalysins, including the proteolyticdomain and the C-terminus, showed higher percentages of identity(Fig. 1B,C), indicating functional conservation (Boldt et al., 2004;

Boldt et al., 2001; Weyer et al., 2007).

Expression analysisIn organs and tissues dissected from adult zebrafish, high levels

of papp-a2 expression were found in brain and gills, loweramounts were detected in bone (skull and vertebrae), skeletalmuscle, eye, gastrointestinal tract, and ovary, and small amounts

were found in kidney and testis (Fig. 2A). No papp-a2 transcriptwas detected in heart or liver.

Analysis of papp-a2 expression during embryonic and earlylarval stages of zebrafish development showed a steep increase inthe amount of transcript from 6 to 12 hpf, followed by a slowdecrease to a lower, steady level between 18 and 48 hpf, and a

second increase between 48 and 96 hpf (Fig. 2B). Expression wasdetected in chordamesoderm at 12 hpf, and in caudal notochordand anterior central nervous system at 18 hpf (Fig. 2C,D). At 24

hpf, expression was prominent in the mid- and hind-brain and thecaudal-most tip of the notochord, and by 48 hpf, papp-a2

transcript was detected in the lower jaw and in several distinct

neural tissues including the retinal ganglion cell layer,cerebellum, and lateral clusters identified as ganglia of thehindbrain (Fig. 2E–G).

Papp-a2 knockdown disrupts normal development ofnotochord and cranial cartilageMorpholino-mediated knockdown of papp-a2 was performed to

assess its function in the zebrafish embryo. A morpholinotargeting the exon-1–intron-1 boundary (e1i1) of the papp-a2 pre-mRNA resulted in a dose-dependent reduction of the wild-type

transcript and in formation of two aberrant splice products(Fig. 3A,B). Sequencing and in silico translation indicated thatpremature stop codons were introduced in both e1i1-induced

splice variants, causing truncations within the N-terminal LamG-like domain. To confirm specificity of the observed phenotypes, atranslation-inhibiting morpholino (tMO, labeled ATG in figures)

annealing to the 59 untranslated region (UTR) and start codon,and a splice-site-targeted morpholino (i3e4) (Fig. 3C) were alsoused. Necrosis of the central nervous system was observed fromboth e1i1 and tMO, indicating non-specific p53-activating

activity (Ekker and Larson, 2001; Robu et al., 2007). Co-injection with p53-targeting morpholino (Robu et al., 2007)specifically diminished the observed necrosis, whereas the other

Fig. 1. Sequence analysis of Papp-a2.(A) Schematic representation showing genesof the syntenic group within 33.80–35.60 Mbon zebrafish chromosome 2 and 173.55–177.45 Mb on human chromosome 1.(B) Schematic representation of maturezebrafish Papp-a2 primary structure indicatingoverall conservation relative to human PAPP-A2 (UniProt entry Q9BXP8.4). The positions ofdomains and their sequence identities tohuman PAPP-A2 are shown, and functionalmotifs, modules, and amino acid residuenumbers are indicated. LamG-like, laminin G-like. (C) ClustalW2 alignments of regions ofparticularly high identity between humanPAPP-A2 and zebrafish Papp-a2. Upper panel,partial proteolytic domain sequence includingthe elongated zinc-binding site and the Met-turn. Lower panel, LNR-3 and C-terminus.Motifs and modules are boxed and labeled.Asterisks indicate identical residues; colonsand dots indicate similar residues.

RESEARCH ARTICLE Journal of Cell Science (2014) 127, 5027–5037 doi:10.1242/jcs.152587

5028

Jour

nal o

f Cel

l Sci

ence

phenotypes remained unaffected. Co-injection of p53-targeting

morpholino was used in all subsequent knockdown experiments.We find that papp-a2-knockdown embryos display ventral

curvature and notochord undulations at 24 hpf (Fig. 3D), andan underdevelopment of the jaw is readily visible at 72 hpf

(Fig. 4A). Given the expression of papp-a2 in thechordamesoderm and the observation of notochord defects inknockdown embryos, we speculated that Papp-a2 might be

required for normal development of the chordamesoderm. Indeed,whole-mount in situ hybridization for the axial mesoderm markernotochord homeobox (noto) revealed a marked broadening in 9

hpf embryos, which could be rescued by co-injection with in-

vitro-transcribed Papp-a2-encoding mRNA (Fig. 3E). Takentogether, these data suggest that Papp-a2 is involved in early

stages of notochord development.The reduced jaw in 72 hpf papp-a2-knockdown embryos

(Fig. 4A) was investigated by Alcian Blue staining of cartilage in3.5 days post fertilization larvae (Fig. 4B). The staining revealed

a severe reduction or absence of Meckel’s cartilage and thepalatoquadrate of the mandibular arch, and of the ceratohyal andhyosymplectic cartilages of the hyoid arch, which together form

the lower jaw. Similarly, the ethmoid plate and trabeculae of theneurocranium, and the five caudal branchial arches were reducedor absent. In knockdown embryos with elements of the

mandibular and hyoid arches present, the angles between theright and left aspects of the arches were visibly shallower,explaining the reduced protrusion of the lower jaw (Fig. 4A).Expression of the cranial neural crest marker dlx2a, which is

required for the development of the pharyngeal cartilages, wasdetected in a reduced number of the pharyngeal arches in papp-

a2-knockdown embryos (Fig. 4C), suggesting that a reduction in

cranial neural crest cells is a cause of the reduced cartilages.

Papp-a2 specifically cleaves Igfbp-3 and Igfbp-5bSimilar to papp-a2, knockdown of igfbp-3 has previously beenreported to cause reduction of cranial cartilage associated withreduced dlx2a expression in the pharyngeal arches (Li et al.,

2005). In addition, broadening of the noto-expression domain has

been reported in zebrafish embryos overexpressing igfbp-3

(Zhong et al., 2011), indicating a functional link between Papp-a2 and Igfbp-3. For biochemical characterization, we transfectedHEK293T cells with zebrafish papp-a2 cDNA and confirmed its

expression by western blotting (Fig. 5A). Edman degradation ofpurified recombinant zebrafish Papp-a2 yielded the sequenceSer(173)-Ile-Ser-Glu-Phe, demonstrating that prepro-Papp-a2 is

composed of a 172-residue prepro-peptide and a 1559-residuemature polypeptide with 52% sequence identity to mature humanPAPP-A2. We then assessed the proteolytic activity of Papp-a2

towards human IGFBP-1 through -6. Whereas no proteolyticactivity of zebrafish Papp-a2 was detected towards IGFBP-1, -2,-4 and -6, IGFBP-3 and -5 were cleaved, each apparently at a

single site (Fig. 5B). This suggests conservation of proteolyticspecificity between the human and zebrafish orthologs. We thentested the activity of Papp-a2 towards the zebrafish orthologs ofthe identified human substrates. We found that Papp-a2 is able to

cleave zebrafish Igfbp-3 and Igfbp-5b (Fig. 5C), thus showingsimilar activity towards endogenous substrates.

Papp-a2 knockdown disrupts vascular development andperturbs normal vascular Notch signalingAbnormal angiogenesis in zebrafish embryos has previously been

reported to result from human IGFBP-3 overexpression (Liuet al., 2007), and from combined igf-2a and igf-2b knockdown(Hartnett et al., 2010). To test whether Papp-a2 is also requiredfor normal angiogenesis, we carried out the papp-a2-knockdown

experiment in Tg(fli1a:eGFP) (Lawson and Weinstein, 2002)embryos, which express eGFP in endothelial cells. Thisexperiment revealed an abnormal branching of intersegmental

vessels (ISVs) extending over the somites, and a reduced growthof ISVs, which was clearly visible at 26 hpf, with the ISVsreaching their full-length by 48 hpf (Fig. 6A,B). Although similar

to the disorganized ISVs associated with igf-2 knockdown, ourspecific observation of ISVs extending over the somites withaberrant branching has, intriguingly, also been reported in

Fig. 2. Expression analysis of papp-a2 inzebrafish embryos, larvae and adults.(A) Quantitative RT-PCR analysis of papp-a2mRNAin adult tissues. Analysis was performed on pooledorgans from three females, with the exception oftestes, which were dissected and pooled from twomales. ND, not detected. (B) Quantitative RT-PCRanalysis of papp-a2 expression levels during earlydevelopment (30–50 embryos per group). papp-a2expression was normalized relative to that of beta-2-microglobulin and fold change was calculatedrelative to bone (A) or 96 hpf (B), respectively. Plottedvalues are mean6s.d. of triplicate measurements.(C–G) Whole mount in situ hybridization showingpapp-a2 expression at 12 (C), 18 (D), 24 (E), and 48hpf (F,G). G is a magnified section of a stained 48 hpfembryo showing expression in the lower jaw (*).Anterior to the left, dorsal to the top, except 12 hpf,which is dorsal view, animal pole to the top. cm,chordamesoderm; nc, notochord; mb, midbrain; hb,hindbrain; cb, cerebellum; rgc, retinal ganglion celllayer; arrowheads, hindbrain ganglia. Identicalstaining patterns were observed when using twoother non-overlapping probes, and no signal wasobserved using sense probe at any developmentalstage (not shown).

RESEARCH ARTICLE Journal of Cell Science (2014) 127, 5027–5037 doi:10.1242/jcs.152587

5029

Jour

nal o

f Cel

l Sci

ence

after-eight, beamter and deadly-seven mutant zebrafish embryos.These mutant fish lines harbor missense mutations in deltaD,

deltaC and notch1a, respectively, all of which are components ofthe Notch signaling system (Therapontos and Vargesson, 2010).This prompted us to investigate whether Papp-a2 might be a

modulator of Notch signaling activity by knocking down papp-a2

in the transgenic Notch reporter line Tg(Tp1bglob:eGFP)um13

(Parsons et al., 2009). Indeed, we found that knockdown embryosdisplayed increased numbers of ISVs positive for Notch signaling

activity (Fig. 6C,D) without affecting the total number of ISVs,as determined by knockdown in Tg(fli1a:eGFP) embryos(Fig. 6E). A profound reduction of Notch signaling was evident

in the midbrain and hindbrain of knockdown embryos at 48 hpf,without dramatic decreases in the size of the brain structures. Inthe epiphysis and telencephalon of the forebrain, Notch signaling

appeared less affected by the knockdown (Fig. 7A,B), in

accordance with the pattern of papp-a2 expression (Fig. 2E,F).

In agreement with these results, papp-a2 knockdown specificallyreduced the expression of the Notch responsive gene her6 inneural tissues, including the dorsal midbrain and hindbrain, withno apparent reduction in the somites (Fig. 7C). This further

substantiates a regulatory effect of papp-a2 on Notch signalingactivity in vivo.

PAPP-A2 promotes Notch activity in vitroTo further assess the capability of PAPP-A2 to modulate Notchsignaling activity, we made use of a previously developed cell-

culture-based Notch reporter assay (Hansson et al., 2006). Theassay is based on a reporter construct containing 12 CBF1–Su(H)–Lag-1 (CSL)-binding motifs upstream of sequence

encoding a red fluorescent reporter (DsRed). Notch receptoractivation results in proteolytic release and translocation of theintracellular domain of the receptor to the nucleus, where itassociates with the CSL transcription factor complex inducing

transcriptional activation of the fluorescent reporter. To mimicthe cell–cell interaction in ligand-induced Notch receptoractivation, two populations of HEK293T cells were prepared

for each experiment: one population was transiently transfectedwith cDNAs encoding Notch1 and the DsRed reporter construct,and one population of non-transfected cells or cells stably

overexpressing Jagged1. The co-culturing of such cellpopulations allows the ligand-induced signaling to occur. Forsignal quantification, we used flow cytometry.

As expected, overexpression of Notch and co-culturing with

Jagged1-overexpressing cells induced Notch activity (Fig. 8A,B).Lower levels of activity were detected in co-cultures of cells notoverexpressing either ligand, or ligand and receptor, revealing a

base-line level of endogenous Notch signaling in the cell line.Addition of the c-secretase inhibitor DAPT, which prevents thefinal cleavage (S3) of the Notch receptor, efficiently abolished

the induced signaling. Co-transfecting PAPP-A2 with Notch1caused a significant increase in Notch signaling in a DAPTsensitive manner (Fig. 8C), suggesting that PAPP-A2 promotes

Notch receptor activation. We finally carried out a similar

Fig. 3. papp-a2 knockdown causes ventral curvature and notochorddefects. (A,B) RT-PCR analysis of the effect of papp-a2-targeted morpholinoei1i on papp-a2 pre-mRNA splicing showing a fast-migrating band (LMW)(A) resulting from the use of a cryptic splice donor site in exon 1, and a slow-migrating band (HMW) (B) resulting from the retention of intron 1 asdetermined by sequencing. These mis-spliced transcripts encode amino acidresidue 1–242 or 1–275 of the prepro-protein plus eight or 30 residues,respectively, encoded by intron 1 followed by a stop codon. Values for e1i1are the amount of morpholino in ng per embryo. WT, wild type. (C) RT-PCRanalysis of the effect of papp-a2-targeted morpholino i3e4 on papp-a2 pre-mRNA splicing. Values for i3e4 are the amount of morpholino in ng perembryo. (D) 24 hpf embryos injected with 5.0 ng control morpholino (cMO),5.0 ng papp-a2-targeted splice inhibiting morpholino (e1i1), 2.5 ngtranslation inhibiting morpholino (ATG), or 2.0 ng splice-inhibiting morpholino(i3e4) with or without 5 ng p53-targeted morpholino (p53). Necrosis of thecentral nervous system was observed from both e1i1 and ATG, indicatingnon-specific p53-activating activity (Ekker and Larson, 2001; Robu et al.,2007). Co-injection with p53-targeting morpholino (Robu et al., 2007)specifically diminished the observed necrosis, whereas the other phenotypesremained unaffected. p53-targeting morpholino was included in allsubsequent knockdown experiments. Insets show a magnified in view fromthe area indicated by the arrow. Notochord outlines are indicated with dashedlines. (E) noto whole mount in situ hybridization of 9 hpf embryos. e1i1+P2indicates co-injection of e1i1 morpholino with 500 pg Papp-a2-encodingmRNA. Numbers indicate fractions of embryos displaying the depictedphenotype. Animal pole view, dorsal to the top.

RESEARCH ARTICLE Journal of Cell Science (2014) 127, 5027–5037 doi:10.1242/jcs.152587

5030

Jour

nal o

f Cel

l Sci

ence

experiment using a PAPP-A2 mutant that had been renderedproteolytically inactive by a single amino acid substitution(E734Q) in the active site (Overgaard et al., 2001). The ability

of this mutant to potentiate Notch signaling did not differ fromwild-type PAPP-A2 (P,0.05) (Fig. 8D). We conclude, therefore,that PAPP-A2 is capable of potentiating Jagged1-induced Notchsignaling independently of its proteolytic activity.

DISCUSSIONWe previously analyzed the phylogenetic relationship between

members of the pappalysin family, including zebrafish Papp-aand Papp-a2 (Kjaer-Sorensen et al., 2013). We here present datashowing that Papp-a2 is expressed during early development and

in adult tissues of the zebrafish, and that adult expression patternsare conserved between mice (Conover et al., 2011) and zebrafish.In addition, biochemical analysis revealed that zebrafish Papp-a2,

like human PAPP-A2, specifically cleaves human IGFBP-3 andIGFBP-5. Furthermore, zebrafish Papp-a2 cleaves zebrafishIgfbp-3 and Igfbp-5b, suggesting that the zebrafish is a suitablein vivo model for studying PAPP-A2 function. Gene targeting

revealed that Papp-a2 is involved in the development of thenotochord and cranial cartilages, and in the development ofintersegmental vessels. Interestingly, our data further indicate

that these effects of Papp-a2 are mediated by two differentmechanisms: proteolytic cleavage of Igfbp-3 and non-proteolyticmodulation of Notch signaling.

In papp-a2-knockdown embryos, we observed reductions ofcranial cartilages and of the cranial neural crest progenitors ofthese tissues. Interestingly, this phenotype closely resembles the

phenotype of igfbp-3 knockdown (Li et al., 2005). The fact thatthe loss of Igfbp-3 and the loss of Papp-a2 have similardevelopmental consequences immediately suggests that Papp-a2affects the development of cranial cartilages through Igfbp-3

proteolysis. We therefore propose that Papp-a2 and Igfbp-3constitute an interdependent pair of proteins, similar to PAPP-Aand IGFBP-4 in mammals (Laursen et al., 2007; Ning et al.,

2008). Although igfbp-4 does not appear to be present in thezebrafish genome (Daza et al., 2011), it is unlikely that the newroles of the Papp-a2/Igfbp-3 pair described in this study can

substitute for the role of PAPP-A/IGFBP-4 in the regulation of

IGF bioactivity. In support of this, the strikingly similarembryonic expression patterns of zebrafish papp-a and mousePapp-A (Conover et al., 2004; Kjaer-Sorensen et al., 2013) differ

remarkably from the embryonic zebrafish papp-a2 expressionpattern.

We also observed broadening of the axial mesoderm uponpapp-a2 knockdown, a phenotype classically associated with

reduced BMP signaling. Broadening of the axial mesodermhas previously been observed in zebrafish upon Igfbp-3overexpression (Zhong et al., 2011), indicating that Papp-a2

might modulate BMP signaling by proteolytic cleavage of Igfbp-3. In accordance with such mechanism, human IGFBP-3 wasrecently reported to bind BMP2 and BMP4 thereby antagonizing

their signaling activity (Kim et al., 2011; Zhong et al., 2011).Furthermore, zebrafish Bmp1 cleaves Igfbp-3, and proteolyticcleavage of human IGFBP-3 by BMP1 is known to reduce its

affinity for BMP4 (Kim et al., 2011). These events are analogousto the regulatory cleavage of the BMP antagonist chordin byTolloid proteins (Blader et al., 1997; Marques et al., 1997;Piccolo et al., 1997), and suggest that IGFBP-3 proteolysis is a

means of modulation of BMP bioavailability. In accordancewith the temporal and spatial expression of papp-a2 mRNA,zebrafish Bmp4 activity is specifically required for the normal

development and differentiation of the chordamesoderm duringlate gastrulation and segmentation (Esterberg et al., 2008).Taken together, our data thus suggest that Papp-a2 is a local

proteolytic modulator of Igfbp-3-mediated Bmp antagonismduring chordamesoderm development.

The papp-a2-knockdown phenotype, ranging from reduced to

lacking cranial cartilages, appears not to be fully accounted for bythe observed partial reduction of cranial neural crest, indicatingthat Papp-a2 is also involved in subsequent processes of cranialchondrogenesis. Accordingly, the expression of papp-a2 in the

lower jaw at 48 hpf suggests that Papp-a2 might be required fornormal jaw development at such later stages. Precise regulation ofBMP signaling is crucial for cranial neural crest development

(Schumacher et al., 2011), and, interestingly, BMP signaling isrequired during the second day post fertilization for the normaldevelopment of cranial cartilages (Dalcq et al., 2012). In addition,

knockdown of runx3, an upstream transcriptional repressor of the

Fig. 4. Papp-a2 is required for normal development of cranialcartilages. (A) Standard light microscopy images of papp-a2-knockdown and control embryos showing severely reduced jawdevelopment. Lateral views, anterior to the left. (B) staining ofpapp-a2 knockdown and control embryos showing reduction orabsence of the ethmoid (e), trabecula cranii (t), Meckel’s cartilage(m), palatoquadrate (pq), ceratohyal (ch), hyosymplectic (hs), andthe five caudal branchial arches (*). Nc, notochord. (C) Flat-mounted 24 hpf zebrafish embryos whole mount in situ stained fordlx2a. Numbers indicate the fraction of embryos displaying thedepicted phenotype. Dorsal view, anterior to the left. Arrowheads,clusters of dlx2a-positive postmigratory cranial neural cells.

RESEARCH ARTICLE Journal of Cell Science (2014) 127, 5027–5037 doi:10.1242/jcs.152587

5031

Jour

nal o

f Cel

l Sci

ence

BMP antagonist Follistatin A, results in loss of cranial cartilagessimilar to papp-a2 knockdown (Dalcq et al., 2012; Flores et al.,

2006). Taken together with our biochemical analysis of Papp-a2,these data support a role of Papp-a2 as a modulator ofBmp signaling, and Papp-a2 thus presents itself as a plausiblecandidate for modulating the development of cartilages of the

cranium through its proteolytic activity.papp-a2-knockdown embryos additionally displayed

angiogenesis defects with ISVs that branch abnormally and

extend over the somites. By comparison to publishedoverexpression, knockdown, and mutant phenotypes, the

consequence of papp-a2 knockdown on angiogenesis could bedue to either perturbation of Igf signaling through impaired Igfbp

proteolysis or modulation of Notch signaling. BMP signaling, bycontrast, is not required for normal ISV development (Wileyet al., 2011) and is therefore an unlikely candidate for theobserved vascular phenotypes. Igfbp-3 and Igf-2 have both been

implicated in angiogenesis of the ISVs in zebrafish inoverexpression and knockdown experiments, respectively(Hartnett et al., 2010; Liu et al., 2007). In particular, the igf-2-

knockdown phenotype is similar to our observations in papp-a2-knockdown embryos. Interestingly, however, ISVs extendingover the somites with an aberrant branching pattern have been

specifically described in zebrafish carrying mutations in differentcomponents of the Notch signaling system. These include the

Fig. 5. Recombinant zebrafish Papp-a2 specifically cleaves human andzebrafish IGFBP-3 and IGFBP-5. (A) Western blot detection of reducedrecombinant zebrafish Papp-a2 and human PAPP-A2 in conditioned mediumfrom transiently transfected HEK293T cells. Primary antibody: 9E10 anti-Mycantibody. (B) Proteolytic assays for the detection of possible Papp-a2 activitytowards the six human IGFBPs in the absence or presence of IGF1 or IGF2.The absence or presence of zebrafish Papp-a2 is indicated with 2 or +.(C) Proteolytic assays for the detection of possible Papp-a2 activity towardsrecombinant zebrafish Igfbp-3 and Igfbp-5b. Detection of intact IGFBPs andC-terminal fragments was by blotting with 9E10 anti-Myc antibody.

Fig. 6. papp-a2 knockdown causes angiogenesis defects andmodulates vascular Notch signaling activity. (A,B) Representative z-stackprojections of fluorescence microscopy images of control and papp-a2-knockdown Tg(fli1a:eGFP) zebrafish embryos at 26 hpf (A) and 48 hpf(B). Arrows, ISVs extending across somites; asterisks, abnormal ISVbranching. (C) Representative z-stack projections of fluorescencemicroscopy images of control and papp-a2-knockdown embryos at 48 hpf inNotch activity reporter zebrafish embryos [Tg(Tp1bglob:eGFP)um13].Knockdown was performed using 1.25 ng ATG, 5.0 ng e1i1, or 2 ng i3e4papp-a2-targeted morpholino per embryo. As a negative control, 5 ng cMOwas injected. All pictures are lateral views, anterior to the left and dorsal tothe top. (D) Number of Notch-signaling positive (the mean6s.d. is indicated)ISVs in the tails of 48 hpf control and papp-a2-knockdown embryos.*P,0.05, **P,0.01; n511 (cMO), 10 (e1i1), 8 (ATG), 8 (i3e4). (E) Totalnumber of ISVs (the mean 6 s.d. is indicated) in the tails of 48 hpf controland papp-a2 knockdown Tg(fli1a:eGFP) embryos. n56 per treatment.

RESEARCH ARTICLE Journal of Cell Science (2014) 127, 5027–5037 doi:10.1242/jcs.152587

5032

Jour

nal o

f Cel

l Sci

ence

after-eight, beamter and deadly-seven mutant zebrafish, in which

deltaD, deltaC and notch1a, respectively, are mutated(Therapontos and Vargesson, 2010). We therefore hypothesizedthat Papp-a2 might be required to modulate Notch signaling in

vivo. In support of such hypothesis, papp-a2 knockdown indeedresulted in aberrant vascular Notch signaling in Notch reporterzebrafish embryos. Specifically, the number of ISVs displayingNotch activity was increased in the knockdown embryos.

Furthermore, reduced Notch signaling activity and specificreduction of Notch responsive her6 expression was observed inthe midbrain and hindbrain of papp-a2-knockdown embryos.

Our cell-culture-based assessment of human PAPP-A2 as apotential modulator of Notch signaling substantiated suchfunction of PAPP-A2. Activation of the Notch receptor requires

a proteolytic cleavage event (S2 cleavage), and identified S2proteases include the adamalysins ADAM10/kuzbanian, andADAM17/TACE (Brou et al., 2000; Lieber et al., 2002).

Because these adamalysins and PAPP-A2 all belong to themetzincin superfamily of metalloproteinases, it was tempting tospeculate that PAPP-A2 might promote Notch activity similarly,that is, by S2 cleavage. However, wild-type and proteolytically

inactive PAPP-A2 were equally efficient in promoting Notchactivity, demonstrating that PAPP-A2 does not modulate Notchreceptor activation through its proteolytic activity. PAPP-A2 does

therefore not promote Notch activity through S2 cleavage.Importantly, the lack of proteolytic dependence further suggeststhat the observed effect of PAPP-A2 on Notch signaling is not an

indirect effect of IGFBP proteolysis.Notochord-derived sonic hedgehog and the downstream

upregulation of vascular endothelial growth factor (Vegf)signaling modulate arterial endothelial differentiation through

Notch signaling (Lawson et al., 2002). It is therefore possible that

the developmental defects of the notochord in papp-a2-knockdown embryos indirectly modulate the vascular Notchactivity. Furthermore, the mind bomb zebrafish mutant with

defective Delta- and Jagged-dependent Notch signaling showsnotochord defects (Yamamoto et al., 2010), and papp-a2 mRNAexpression in the notochord appears to overlap with publishedjagged1a and jagged1b expression patterns (Yamamoto et al.,

2010). We speculate that modulation of Notch signaling by Papp-a2 might therefore also be involved in notochord developmentand functionality. We were, however, unable to detect changes in

Notch signaling activity in the notochord resulting from papp-a2

knockdown.Biochemical data suggest that PAPP-A2 acts as an upstream

modulator of AKT signaling by regulating IGF bioavailability.Notch receptor activation has been shown to modulate AKTsignaling through two distinct pathways: transcriptional

upregulation of the human ortholog of zebrafish Her6, HES1,causing inhibition of phosphoinositide 3-kinase (PI3K) activityby PTEN inhibition (Hales et al., 2013; Wong et al., 2012), ortranscriptional upregulation of IGF1R (Medyouf et al., 2011).

The finding that papp-a2 knockdown attenuates Notchsignaling activity, in turn reducing her6 expression inembryonic midbrain and hindbrain, might thus indicate

additional levels of PAPP-A2 regulation of AKT signalingthrough the Notch signaling pathway. Further work is requiredto delineate the details of PAPP-A2 function within this

signaling network.In conclusion, we present in vivo data suggesting

interdependence of Papp-a2 and Igfbp-3 in the development ofcranial cartilages, and proposing that Papp-a2 is a local modulator

Fig. 7. papp-a2-knockdown reducesNotch signaling activity in the CNS.(A) Representative images of control andpapp-a2-knockdown Notch activity reporterembryos [Tg(Tp1bglob:eGFP)um13]showing the range of Notch reporterfluorescence levels. Embryos weregrouped according to fluorescence levelsranging from 1 (very low Notch signalingactivity) to 4 (maximum Notch signalingactivity). (B) Plot showing the percentageof embryos at each fluorescence level incontrol and knockdown embryos. n520(cMO), 24 (e1i1), 30 (ATG), 14 (i3e4).(C) her6 whole-mount in situ hybridizationof 24 hpf embryos. Numbers indicatefractions of embryos displaying thedepicted phenotype. hb, hindbrain;mb, midbrain.

RESEARCH ARTICLE Journal of Cell Science (2014) 127, 5027–5037 doi:10.1242/jcs.152587

5033

Jour

nal o

f Cel

l Sci

ence

of Igfbp-3 ligand signaling through specific Igfbp-3 proteolysis.Furthermore, our data suggest a conserved function of PAPP-A2

in the development of specific cartilages across the vertebratephylum. We additionally characterize abnormalities in thedevelopment of the notochord and the vascular system not

previously associated with PAPP-A2. Importantly, our data forthe first time implicate Papp-a2 in the modulation of Bmpsignaling by specific Igfbp-3 proteolysis and in the modulation of

Notch signaling through a non-proteolytic mechanism. Futurework is required to determine in which physiological systems theproteolytic activity of Papp-a2 promotes Igf and Bmp signaling,respectively, and to further delineate the exact mechanism of

Notch signaling modulation by Papp-a2 in vivo.

MATERIALS AND METHODSThe nucleotide sequence for the Danio rerio partial papp-a2 gene

reported in this paper has been submitted to the European Nucleotide

Archive with accession number HG934471.

AnimalsZebrafish were fed four times daily and maintained on a 14-h-light–10-h-

dark cycle on recirculating housing systems at 28 C. Embryos were

obtained by natural crosses, reared in E3 buffer [5 mM NaCl, 0.17 mM

KCl, 0.33 mM CaCl2, 0.33 mM MgSO4, 1025% (w/w) methylene blue,

2 mM HEPES pH 7.0] at 28.5 C, and staged according to Kimmel et al.

(Kimmel et al., 1995). Embryos were sedated in tricaine (150 ng/ml)

(Aldrich) in E3 buffer for phenotypic analysis when required. Phenotypic

analyses were performed in embryos and larvae of the AB strain, and in

transgenic embryos of the Tg(fli1a:eGFP) (Lawson and Weinstein,

2002) and Tg(Tp1bglob:eGFP)um13 (Parsons et al., 2009) strains. All

experiments involving zebrafish embryos and larvae were carried out

according to Danish legislation.

Sequence analysisZebrafish papp-a2 was previously identified using BLAST (Kjaer-Sorensen

et al., 2013). Synteny analysis was performed using the Synteny Database

(Catchen et al., 2009) on the Ens61 dataset using sliding window size 25

and exclusion of peripheral genes. Alignments with human PAPP-A2

(UniProt entry Q9BXP8.4) were performed using ClustalW2, and gaps

were excluded in the calculations of sequence identity percentages.

Cloning and sequencingpapp-a2 cDNA was synthesized from total RNA extracted from a pair of

adult Tubingen zebrafish. Briefly, the fish were killed by tricaine

overdose and homogenized, and total RNA was extracted using TRI

reagent (Sigma-Aldrich). Random hexamer-primed cDNA was reverse-

transcribed using the Thermoscript RT-PCR System for First-Strand

cDNA Synthesis (Invitrogen). papp-a2 cDNA was amplified by nested

PCR using KOD Hot Start DNA Polymerase (Novagen) and primer sets

59-AATTAGTCATCTTGCGCAGGC-39 and 59-AATGGTGTAGCTAC-

GCAAACGGGC-39 (outer PCR); and 59-AAAAAAGGTACCAATGT-

CCACGCATTGAAGGCTGCA-39 and 59-TTTGATCAGCCTGGCT-

CTCTC-39 (inner PCR). The PCR product was blunt-end cloned into

the pJet1/blunt vector using the GeneJet PCR Cloning Kit (Fermentas).

For transfer into the pcDNA3.1/myc-His(+)A expression vector,

Fig. 8. PAPP-A2 promotes Notch signaling independently of its proteolytic activity. The reporter construct 126CSL-DsRedExpressDR (DsRed) contains12 CSL-binding motifs upstream of a sequence encoding a fluorescent reporter. The intracellular domain of the Notch receptor translocates to the nucleus uponreceptor activation. It associates with the CBF1–Su(H)–Lag-1 transcription factor complex (CSL) in the nucleus inducing transcriptional activation from the CSL-binding motifs. Notch receptor activation thus induces expression of the fluorophore in cells transfected with the reporter construct (Hansson et al., 2006).(A) Flow cytometry quantification of red fluorescence from the Notch reporter construct. DsRed, cells transfected with 126CSL-DsRedExpressDR co-culturedwith mock transfected cells; DsRed+N1, cells co-transfected with 126CSL-DsRedExpressDR and human Notch1 co-cultured with mock-transfected cells;DsRed+N1/hJ1, cells co-transfected with 126CSL-DsRedExpressDR and human Notch1 co-cultured with cells stably overexpressing human Jagged1. Insetsshow representative fluorescence microscopy images of cell culture plates (confluency .100%). (B–D) Relative mean fluorescence of transfected, co-culturedcells with or without DAPT measured by flow cytometry with subtracted mock values and normalized to DsRed+N1 (B) or DsRed+N1/J1 (C,D). P2, humanPAPP-A2; Control, irrelevant cDNA (angiotensinogen); E734Q, proteolytically inactive human PAPP-A2. Plotted values are mean6s.d. of three to fivemeasurements. **P,0.01; ***P,0.001; ns, not significant.

RESEARCH ARTICLE Journal of Cell Science (2014) 127, 5027–5037 doi:10.1242/jcs.152587

5034

Jour

nal o

f Cel

l Sci

ence

the coding sequence was amplified by the primer set 59- AAA-

AAAGGTACCAATGTCCACGCATTGAAGGCTGCA-39 and 59-GCT-

CTCTCTCTCTCTGTTCTAGACGGAGCTC-39 followed by restriction

cleavage by KpnI and XbaI and ligation. All constructs were confirmed

by sequencing.

Expression analysisFor quantitative real-time PCR (qRT-PCR), total RNA was purified and

DNase-treated using an RNeasy Mini Kit (Qiagen) and RNase-Free

DNase Set (Qiagen), respectively, according to the manufacturer’s

recommendations. The qRT-PCR reaction was performed using the

Brilliant III Ultra Fast SYBR Green QRT-PCR Master Mix (Agilent

Technologies) on a Stratagene Mx3005P (Agilent Technologies) with the

MxPro QPCR Software (Agilent Technologies). Cycling conditions

were: 10 min at 50 C and 3 min at 95 C followed by 40 cycles of 20 s at

95 C and 20 s at 58 C. Primer sets used were 59-GCCTTCA-

CCCCAGAGAAAGG-39 and 59-CCGGTTTGGATTTACATGTTG-39

for beta-2-microglobulin, and 59-ACACACCGTACAACAACTACA-39

and 59-TCATATCACCCTCCCTGTGG-39 for papp-a2. Whole mount in

situ hybridization was performed according to Thisse et al. (Thisse and

Thisse, 2008) with the addition of 5% (w/v) dextran sulfate

(Mr.500,000, Sigma) to hybridization mix for improved signal:noise

ratio. Two non-overlapping papp-a2 probes were generated from cDNA:

within the coding sequence (nucleotides 75–575 of the coding sequence

in European Nucleotide Archive accession number HG934471) by single

PCR using primer set 59-CCATATCCAGAAGTATAAGCGGACACTG-

39 and 59-GTGTGGTGGTTATAGGATTTTTCTGGGATG-39 and

within the 59-UTR (nucleotide 526183–526581 of GenBank NW_

003334160.1) by nested PCR using outer primer set 59-GCCGGT-

GAAAAGGTGGCAA-39 and 59-GCGTGGACATTTGTTTGTTTTG-39,

and inner primer set 59-CCTTTATCCAGCGTTACATCTC-39 and 59-

GAAATGCACTTGCTAAGGGA-39. The noto probe was generated

from cDNA template by nested PCR using outer primer set 59-

CTATGCCACCTCTTCTTCAAAACC-39 and 59-GAGAGGAGTAAC-

AATCTGGGGAAA-39 and inner primer set 59-GCGGAGATGAGA-

GAACGAACAAA-39/59-TAATACGACTGAGTATACCCTCTTCTGT-

GAAATCCCTCTCCTC-39 incorporating the T7 promoter (nucleotide

162–706 of GenBank NM_131055.1). The her6 probe template was

generated from random hexamer primed cDNA from whole adult

zebrafish by single PCR using primer set 59-CAACGAAAGCTT-

GGGTCAGC-39 and 59-TAATACGACTCACTATAGGGATAACAG-

GGCCGTTTGGAGC-39 incorporating the T7 promoter (nucleotide

373–939 of GenBank NM_131079.1).

Knockdown and rescueMorpholino sequences were as follows: standard control morpholino

(cMO), 59-CCTCTTACCTCAGTTACAATTTATA-39 (Gene Tools,

LLC); standard p53 morpholino (p53): 59-GCGCCATTGCTTTGCA-

AGAATTG-39 (Robu et al., 2007); and Papp-a2-targeted morpholinos

e1i1, 59-GAATTTACAGACACACCCACCTGCG-39, tMO (labeled

ATG in the figures), 59-CATCATTTTATTGAGGAGAAGCTGC-39

and i3e4, 59- TTCCACCTGAAAAATGACACAAAGT-39. Capped

mRNA was in vitro transcribed using the mMessage mMachine T7

ULTRA Kit (Ambion, Inc.) with a gel-purified NaeI/MluI-linearized

Papp-a2 expression plasmid as template. Microinjection volumes were

calibrated by performing 10 injections into a 0.5 ml microcapillary tube

(Drummond Microcaps), measuring the amount of liquid using a ruler,

and calculating the volume per injection. Volumes of 5 nl were

microinjected into the center of the yolk of zebrafish zygotes.

Recombinant protein expression and purificationHEK293T cells (293tsA1609neo) were maintained in high-glucose

Dulbecco’s modified Eagle’s medium (DMEM) supplemented with

10% fetal bovine serum, 2 mM glutamine, non-essential amino acids and

gentamicin (Invitrogen). Cells were transiently transfected by calcium

phosphate co-precipitation (Pear et al., 1993). Conditioned medium was

harvested after 48 h and cleared by centrifugation. For N-terminal

sequence analysis, recombinant Myc–His-tagged zebrafish Papp-a2 was

partially purified by nickel affinity on chelating Sepharose Fast Flow

(Amersham Biosciences), followed by immunoprecipitation on protein-

G–Sepharose 4 Fast Flow (Amersham Biosciences) saturated with

anti-Myc (9E10) monoclonal antibody. Recombinant IGFBPs

(IGFBP1–IGFBP6) were expressed, purified and radiolabeled as

described previously (Boldt et al., 2001).

Biochemical analysisWestern blotting was performed as previously described (Boldt et al., 2001)

using purified 9E10 anti-Myc antibody (1 mg/ml) diluted 1:1000 and

horseradish peroxidase (HRP)-conjugated rabbit anti-mouse (PO260,

DAKO) diluted 1:1000, and visualized using an Image Quant LAS4000

instrument (GE). Edman degradation was performed to determine the N-

terminal sequence of purified mature zebrafish Papp-a2. Briefly, cleavage

products were separated by SDS-PAGE and blotted onto PVDF

membranes. The membranes were stained with Coomassie Blue, and

bands of interest were excised and analyzed (1.1–4.8 pmol) on an Applied

Biosystems 491 Protein Sequencer. Proteolytic assays for the detection of

possible Papp-a2 activity towards the human IGFBPs were conducted as

described previously (Boldt et al., 2001). Briefly, medium harvested from

transiently transfected HEK293T cells was added to 125I-labelled substrate

pre-incubated with or without molar excess of human IGF1 or IGF2

(Bachem). Reactions were incubated at 28 C and stopped by the addition of

25 mM EDTA, followed by separation by SDS-PAGE and visualization by

autoradiography. Proteolytic assays for the detection of possible Papp-a2

activity towards zebrafish Igfbps were conducted similarly by adding

medium harvested from HEK293T cells transiently transfected with Papp-

a2 or empty vector to medium harvested from HEK293T cells transiently

transfected with Myc-tagged Igfbp-3 (Li et al., 2005) or Igfbp-5b (Dai et al.,

2010). Intact recombinant Igfbp and C-terminal fragments were detected by

western blotting using purified 9E10 anti-Myc antibody as described above.

Notch reporter in vitro assayHEK293T cells were maintained as above. Selection medium for

HEK293T:J1 cells stably expressing human Jagged1 and carrying the

puromycin resistance gene, pCAG-J1-IRES-Puro (kindly provided by

Urban Lendahl) was additionally supplemented with 1 mg/ml puromycin

(Sigma) (Hansson et al., 2006). 2.56106 HEK293T cells were seeded

onto 6-cm culture dishes and transiently transfected 20 h later by calcium

phosphate co-precipitation. Cells were co-transfected with a total of 10–

15 mg plasmid DNA. Plasmid DNA mixtures for transfection contained

2.5 mg 126CSL-dsRedExpressDR reporter plasmid (Hansson et al.,

2006), 7.5 mg human Notch1 cDNA or empty plasmid, and, for PAPP-A2

experiments, 5 mg cDNA encoding human PAPP-A2 or active-site

mutated PAPP-A2 (E734Q) (Overgaard et al., 2001), irrelevant cDNA

(human angiotensinogen) (Kløverpris et al., 2013) or empty plasmid as

controls. Plasmid DNA was prepared with a GenElute HP Plasmid

Miniprep Kit (Sigma) or Plasmid Giga Kit (Qiagen), and eluted in 5 mM

Tris-HCl pH 7.4. At 10 h post-transfection, ,2.56106 HEK293T:hJ1 or

HEK293T control cells were detached by trypsin treatment and added to

each plate of transiently transfected cells for co-culturing. At 48 h post-

transfection, the co-cultures were analyzed for Notch-activated reporter

expression by fluorescence microscopy or flow cytometry. For flow

cytometry, cells were detached by trypsin treatment, resuspended in PBS

and analyzed on a Cytomics FC500 MPL flow cytometer (Beckman

Coulter). 100,000 cells of each sample were analyzed. Mean-

fluorescence-subtracted mock values were calculated as specific

measures of Notch activity. Pharmacological inhibition of Notch

signaling was performed by the addition of 10 mM N-[N-(3,5-

Difluorophenacetyl-L-alanyl)]-S-phenylglycine t-butyl ester (DAPT)

dissolved in DMSO or DMSO alone as vehicle control. The inhibitor

was added to the cells at 10 h post-transfection.

Microscopy and imagingLive embryos were mounted in 1.5% hydroxypropyl methyl cellulose

(Mr5 86,000; Sigma) in E3 buffer and fixed embryos were mounted in

100% glycerol. Embryos were observed and documented on a Zeiss

AxioObserver.Z1 equipped with colibri.2 lightsource, AxioCam HRm, and

RESEARCH ARTICLE Journal of Cell Science (2014) 127, 5027–5037 doi:10.1242/jcs.152587

5035

Jour

nal o

f Cel

l Sci

ence

Apotome.2, or on an Olympus SZX16 stereomicroscope with an Altra20

camera and a Photonic LED transmitted light stage. Light and contrast of

images was adjusted and images were cropped using Adobe Photoshop

CS5. Optic sections were stacked and z-projected using ImageJ.

StatisticsData were subjected to the extreme studentized deviate method for

detection of outliers (P,0.05), and to one-way analysis of variance with

Newman–Keuls post-test.

AcknowledgementsWe thank Urban Lendahl (Karolinska Institute, Sweden) for kindly providing us withthe 126CSL-DsRedExpressDR plasmid and the pCAG-J1-IRES-Puro cell line. Wethank Cunming Duan (University of Michigan, MI) for kindly providing us withexpression plasmids encoding Myc-tagged zebrafish Igfbp-3 and Igfbp-5b. We alsothank Michael J. Parsons (Johns Hopkins University, MD) for kindly providing uswith the Tg(Tp1bglob:eGFP)um13 Notch activity reporter zebrafish strain.

Competing interestsThe authors declare no competing interests.

Author contributionsK.K.S., L.S.L. and C.O. conceived and designed the experiments. K.K.S., D.H.E.,M.R.J., M.G.M. and L.L.H. performed the experiments. K.W. and L.L.S. performedpreliminary experiments. K.K.S., D.H.E., M.R.J., M.G.M., L.S.L. and C.O.analyzed the data. K.K.S. and C.O. wrote the paper with input from M.R.J.

FundingThis study was supported by the Lundbeck Foundation; the Novo NordiskFoundation; the Carlsberg Foundation; and the Danish Council for IndependentResearch.

ReferencesBlader, P., Rastegar, S., Fischer, N. and Strahle, U. (1997). Cleavage of theBMP-4 antagonist chordin by zebrafish tolloid. Science 278, 1937-1940.

Boldt, H. B., Overgaard, M. T., Laursen, L. S., Weyer, K., Sottrup-Jensen, L.and Oxvig, C. (2001). Mutational analysis of the proteolytic domain ofpregnancy-associated plasma protein-A (PAPP-A): classification as ametzincin. Biochem. J. 358, 359-367.

Boldt, H. B., Kjaer-Sorensen, K., Overgaard, M. T., Weyer, K., Poulsen, C. B.,Sottrup-Jensen, L., Conover, C. A., Giudice, L. C. and Oxvig, C. (2004). TheLin12-notch repeats of pregnancy-associated plasma protein-A bind calciumand determine its proteolytic specificity. J. Biol. Chem. 279, 38525-38531.

Brou, C., Logeat, F., Gupta, N., Bessia, C., LeBail, O., Doedens, J. R.,Cumano, A., Roux, P., Black, R. A. and Israel, A. (2000). A novel proteolyticcleavage involved in Notch signaling: the role of the disintegrin-metalloproteaseTACE. Mol. Cell 5, 207-216.

Catchen, J. M., Conery, J. S. and Postlethwait, J. H. (2009). Automatedidentification of conserved synteny after whole-genome duplication. GenomeRes. 19, 1497-1505.

Christians, J. K., de Zwaan, D. R. and Fung, S. H. (2013). Pregnancy associatedplasma protein A2 (PAPP-A2) affects bone size and shape and contributes tonatural variation in postnatal growth in mice. PLoS ONE 8, e56260.

Conover, C. A., Bale, L. K., Overgaard, M. T., Johnstone, E. W., Laursen, U. H.,Fuchtbauer, E. M., Oxvig, C. and van Deursen, J. (2004). Metalloproteinasepregnancy-associated plasma protein A is a critical growth regulatory factorduring fetal development. Development 131, 1187-1194.

Conover, C. A., Boldt, H. B., Bale, L. K., Clifton, K. B., Grell, J. A., Mader, J. R.,Mason, E. J. and Powell, D. R. (2011). Pregnancy-associated plasma protein-A2 (PAPP-A2): tissue expression and biological consequences of geneknockout in mice. Endocrinology 152, 2837-2844.

Dai, W., Kamei, H., Zhao, Y., Ding, J., Du, Z. and Duan, C. (2010). Duplicatedzebrafish insulin-like growth factor binding protein-5 genes with split functionaldomains: evidence for evolutionarily conserved IGF binding, nuclearlocalization, and transactivation activity. FASEB J. 24, 2020-2029.

Dalcq, J., Pasque, V., Ghaye, A., Larbuisson, A., Motte, P., Martial, J. A. andMuller, M. (2012). RUNX3, EGR1 and SOX9B form a regulatory cascaderequired to modulate BMP-signaling during cranial cartilage development inzebrafish. PLoS ONE 7, e50140.

Daza, D. O., Sundstrom, G., Bergqvist, C. A., Duan, C. and Larhammar, D.(2011). Evolution of the insulin-like growth factor binding protein (IGFBP) family.Endocrinology 152, 2278-2289.

De Meyts, P., Palsgaard, J., Sajid, W., Theede, A. M. and Aladdin, H. (2004).Structural biology of insulin and IGF-1 receptors. Novartis Found Symp. 262,160-171; discussion 171-176, 265-268.

Duan, C. and Xu, Q. (2005). Roles of insulin-like growth factor (IGF) bindingproteins in regulating IGF actions. Gen. Comp. Endocrinol. 142, 44-52.

Ekker, S. C. and Larson, J. D. (2001). Morphant technology in modeldevelopmental systems. Genesis 30, 89-93.

Esterberg, R., Delalande, J. M. and Fritz, A. (2008). Tailbud-derived Bmp4 drivesproliferation and inhibits maturation of zebrafish chordamesoderm.Development 135, 3891-3901.

Flores, M. V., Lam, E. Y., Crosier, P. and Crosier, K. (2006). A hierarchy of Runxtranscription factors modulate the onset of chondrogenesis in craniofacialendochondral bones in zebrafish. Dev. Dyn. 235, 3166-3176.

Forbes, B. E., McCarthy, P. and Norton, R. S. (2012). Insulin-like growth factorbinding proteins: a structural perspective. Front. Endocrinol. (Lausanne) 3,38.

Gyrup, C. and Oxvig, C. (2007). Quantitative analysis of insulin-like growth factor-modulated proteolysis of insulin-like growth factor binding protein-4 and -5 bypregnancy-associated plasma protein-A. Biochemistry 46, 1972-1980.

Hales, E. C., Orr, S. M., Larson Gedman, A., Taub, J. W. and Matherly, L. H.(2013). Notch1 receptor regulates AKT protein activation loop (Thr308)dephosphorylation through modulation of the PP2A phosphatase inphosphatase and tensin homolog (PTEN)-null T-cell acute lymphoblasticleukemia cells. J. Biol. Chem. 288, 22836-22848.

Hansson, E. M., Teixeira, A. I., Gustafsson, M. V., Dohda, T., Chapman, G.,Meletis, K., Muhr, J. and Lendahl, U. (2006). Recording Notch signaling in realtime. Dev. Neurosci. 28, 118-127.

Hartnett, L., Glynn, C., Nolan, C. M., Grealy, M. and Byrnes, L. (2010). Insulin-like growth factor-2 regulates early neural and cardiovascular systemdevelopment in zebrafish embryos. Int. J. Dev. Biol. 54, 573-583.

Jones, J. I. and Clemmons, D. R. (1995). Insulin-like growth factors and theirbinding proteins: biological actions. Endocr. Rev. 16, 3-34.

Kim, B., Huang, G., Ho, W. B. and Greenspan, D. S. (2011). Bonemorphogenetic protein-1 processes insulin-like growth factor-binding protein 3.J. Biol. Chem. 286, 29014-29025.

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.

Kjaer-Sorensen, K., Engholm, D. H., Kamei, H., Morch, M. G., Kristensen,A. O., Zhou, J., Conover, C. A., Duan, C. and Oxvig, C. (2013). Pregnancy-associated plasma protein A (PAPP-A) modulates the early developmental ratein zebrafish independently of its proteolytic activity. J. Biol. Chem. 288, 9982-9992.

Kløverpris, S., Skov, L. L., Glerup, S., Pihl, K., Christiansen, M. and Oxvig, C.(2013). Formation of high-molecular-weight angiotensinogen during pregnancyis a result of competing redox reactions with the proform of eosinophil majorbasic protein. Biochem. J. 449, 209-217.

Laursen, L. S., Overgaard, M. T., Søe, R., Boldt, H. B., Sottrup-Jensen, L.,Giudice, L. C., Conover, C. A. and Oxvig, C. (2001). Pregnancy-associatedplasma protein-A (PAPP-A) cleaves insulin-like growth factor binding protein(IGFBP)-5 independent of IGF: implications for the mechanism of IGFBP-4proteolysis by PAPP-A. FEBS Lett. 504, 36-40.

Laursen, L. S., Kjaer-Sorensen, K., Andersen, M. H. and Oxvig, C.(2007). Regulation of insulin-like growth factor (IGF) bioactivity by sequentialproteolytic cleavage of IGF binding protein-4 and -5. Mol. Endocrinol. 21, 1246-1257.

Lawrence, J. B., Oxvig, C., Overgaard, M. T., Sottrup-Jensen, L., Gleich, G. J.,Hays, L. G., Yates, J. R., III and Conover, C. A. (1999). The insulin-like growthfactor (IGF)-dependent IGF binding protein-4 protease secreted by humanfibroblasts is pregnancy-associated plasma protein-A. Proc. Natl. Acad. Sci.USA 96, 3149-3153.

Lawson, N. D. and Weinstein, B. M. (2002). In vivo imaging of embryonicvascular development using transgenic zebrafish. Dev. Biol. 248, 307-318.

Lawson, N. D., Vogel, A. M. and Weinstein, B. M. (2002). sonic hedgehog andvascular endothelial growth factor act upstream of the Notch pathway duringarterial endothelial differentiation. Dev. Cell 3, 127-136.

Li, Y., Xiang, J. and Duan, C. (2005). Insulin-like growth factor-binding protein-3plays an important role in regulating pharyngeal skeleton and inner earformation and differentiation. J. Biol. Chem. 280, 3613-3620.

Lieber, T., Kidd, S. and Young, M. W. (2002). kuzbanian-mediated cleavage ofDrosophila Notch. Genes Dev. 16, 209-221.

Liu, B., Lee, K. W., Anzo, M., Zhang, B., Zi, X., Tao, Y., Shiry, L., Pollak, M., Lin,S. and Cohen, P. (2007). Insulin-like growth factor-binding protein-3 inhibition ofprostate cancer growth involves suppression of angiogenesis. Oncogene 26,1811-1819.

Marques, G., Musacchio, M., Shimell, M. J., Wunnenberg-Stapleton, K., Cho,K. W. and O’Connor, M. B. (1997). Production of a DPP activity gradient in theearly Drosophila embryo through the opposing actions of the SOG and TLDproteins. Cell 91, 417-426.

Maures, T., Chan, S. J., Xu, B., Sun, H., Ding, J. and Duan, C. (2002). Structural,biochemical, and expression analysis of two distinct insulin-like growth factor Ireceptors and their ligands in zebrafish. Endocrinology 143, 1858-1871.

Medyouf, H., Gusscott, S., Wang, H., Tseng, J. C., Wai, C., Nemirovsky, O.,Trumpp, A., Pflumio, F., Carboni, J., Gottardis, M. et al. (2011). High-levelIGF1R expression is required for leukemia-initiating cell activity in T-ALL and issupported by Notch signaling. J. Exp. Med. 208, 1809-1822.

Mikkelsen, J. H., Gyrup, C., Kristensen, P., Overgaard, M. T., Poulsen, C. B.,Laursen, L. S. and Oxvig, C. (2008). Inhibition of the proteolytic activity ofpregnancy-associated plasma protein-A by targeting substrate exosite binding.J. Biol. Chem. 283, 16772-16780.

RESEARCH ARTICLE Journal of Cell Science (2014) 127, 5027–5037 doi:10.1242/jcs.152587

5036

Jour

nal o

f Cel

l Sci

ence

Mukherjee, A. and Rotwein, P. (2008). Insulin-like growth factor-binding protein-5inhibits osteoblast differentiation and skeletal growth by blocking insulin-likegrowth factor actions. Mol. Endocrinol. 22, 1238-1250.

Ning, Y., Schuller, A. G., Bradshaw, S., Rotwein, P., Ludwig, T., Frystyk, J. andPintar, J. E. (2006). Diminished growth and enhanced glucose metabolism intriple knockout mice containing mutations of insulin-like growth factor bindingprotein-3, -4, and -5. Mol. Endocrinol. 20, 2173-2186.

Ning, Y., Schuller, A. G., Conover, C. A. and Pintar, J. E. (2008). Insulin-likegrowth factor (IGF) binding protein-4 is both a positive and negative regulator ofIGF activity in vivo. Mol. Endocrinol. 22, 1213-1225.

Overgaard, M. T., Boldt, H. B., Laursen, L. S., Sottrup-Jensen, L., Conover,C. A. and Oxvig, C. (2001). Pregnancy-associated plasma protein-A2 (PAPP-A2), a novel insulin-like growth factor-binding protein-5 proteinase. J. Biol.Chem. 276, 21849-21853.

Parsons, M. J., Pisharath, H., Yusuff, S., Moore, J. C., Siekmann, A. F.,Lawson, N. and Leach, S. D. (2009). Notch-responsive cells initiate thesecondary transition in larval zebrafish pancreas. Mech. Dev. 126, 898-912.

Pear, W. S., Nolan, G. P., Scott, M. L. and Baltimore, D. (1993). Production ofhigh-titer helper-free retroviruses by transient transfection. Proc. Natl. Acad. Sci.USA 90, 8392-8396.

Piccolo, S., Agius, E., Lu, B., Goodman, S., Dale, L. andDeRobertis, E.M. (1997).Cleavage of Chordin by Xolloid metalloprotease suggests a role for proteolyticprocessing in the regulation of Spemann organizer activity. Cell 91, 407-416.

Robu, M. E., Larson, J. D., Nasevicius, A., Beiraghi, S., Brenner, C., Farber,S. A. and Ekker, S. C. (2007). p53 activation by knockdown technologies. PLoSGenet. 3, e78.

Salih, D. A., Mohan, S., Kasukawa, Y., Tripathi, G., Lovett, F. A., Anderson,N. F., Carter, E. J., Wergedal, J. E., Baylink, D. J. and Pell, J. M. (2005).Insulin-like growth factor-binding protein-5 induces a gender-related decrease inbone mineral density in transgenic mice. Endocrinology 146, 931-940.

Schumacher, J. A., Hashiguchi, M., Nguyen, V. H. and Mullins, M. C. (2011). Anintermediate level of BMP signaling directly specifies cranial neural crestprogenitor cells in zebrafish. PLoS ONE 6, e27403.

Silha, J. V., Mishra, S., Rosen, C. J., Beamer, W. G., Turner, R. T., Powell, D. R. andMurphy, L. J. (2003). Perturbations in bone formation and resorption in insulin-likegrowth factor binding protein-3 transgenic mice. J. Bone Miner. Res. 18, 1834-1841.

Tallant, C., Garcıa-Castellanos, R., Seco, J., Baumann, U. and Gomis-Ruth, F. X. (2006). Molecular analysis of ulilysin, the structural prototypeof a new family of metzincin metalloproteases. J. Biol. Chem. 281, 17920-17928.

Therapontos, C. and Vargesson, N. (2010). Zebrafish notch signalling pathwaymutants exhibit trunk vessel patterning anomalies that are secondary to somitemisregulation. Dev. Dyn. 239, 2761-2768.

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

Weyer, K., Boldt, H. B., Poulsen, C. B., Kjaer-Sorensen, K., Gyrup, C. andOxvig,C. (2007). A substrate specificity-determining unit of three Lin12-Notch repeatmodules is formed in trans within the pappalysin-1 dimer and requires a sequencestretch C-terminal to the third module. J. Biol. Chem. 282, 10988-10999.

White, Y. A., Kyle, J. T. and Wood, A. W. (2009). Targeted gene knockdown inzebrafish reveals distinct intraembryonic functions for insulin-like growth factor IIsignaling. Endocrinology 150, 4366-4375.

Wiley, D. M., Kim, J. D., Hao, J., Hong, C. C., Bautch, V. L. and Jin, S. W.(2011). Distinct signalling pathways regulate sprouting angiogenesis from thedorsal aorta and the axial vein. Nat. Cell Biol. 13, 686-692.

Wong, G. W., Knowles, G. C., Mak, T. W., Ferrando, A. A. and Zuniga-Pflucker,J. C. (2012). HES1 opposes a PTEN-dependent check on survival,differentiation, and proliferation of TCRb-selected mouse thymocytes. Blood120, 1439-1448.

Yamamoto, M., Morita, R., Mizoguchi, T., Matsuo, H., Isoda, M., Ishitani, T.,Chitnis, A. B., Matsumoto, K., Crump, J. G., Hozumi, K. et al. (2010). Mib-Jag1-Notch signalling regulates patterning and structural roles of the notochordby controlling cell-fate decisions. Development 137, 2527-2537.

Yan, X., Baxter, R. C. and Firth, S. M. (2010). Involvement of pregnancy-associated plasma protein-A2 in insulin-like growth factor (IGF) binding protein-5 proteolysis during pregnancy: a potential mechanism for increasing IGFbioavailability. J. Clin. Endocrinol. Metab. 95, 1412-1420.

Zhong, Y., Lu, L., Zhou, J., Li, Y., Liu, Y., Clemmons, D. R. and Duan, C. (2011).IGF binding protein 3 exerts its ligand-independent action by antagonizing BMPin zebrafish embryos. J. Cell Sci. 124, 1925-1935.

Zou, S., Kamei, H., Modi, Z. and Duan, C. (2009). Zebrafish IGF genes: geneduplication, conservation and divergence, and novel roles in midline andnotochord development. PLoS ONE 4, e7026.

RESEARCH ARTICLE Journal of Cell Science (2014) 127, 5027–5037 doi:10.1242/jcs.152587

5037