RESEARCH ARTICLE TECHNIQUES AND RESOURCES

Second-generation Notch1 activity-trap mouse line (N1IP::CreHI)provides a more comprehensive map of cells experiencing Notch1activityZhenyi Liu1,2, Eric Brunskill3, Scott Boyle2, Shuang Chen2,3, Mustafa Turkoz2,3, Yuxuan Guo2,4, Rachel Grant2

and Raphael Kopan2,3,*

ABSTRACTWe have previously described the creation and analysis of a Notch1activity-trap mouse line, Notch1 intramembrane proteolysis-Cre6MT orN1IP::CreLO, that marked cells experiencing relatively high levels ofNotch1 activation. Here, we report and characterize a second line withimprovedsensitivity (N1IP::CreHI) tomark cells experiencing lower levelsof Notch1 activation. This improvement was achieved by increasingtranscript stability and by restoring the native carboxy terminus of Cre,resulting ina five- to tenfold increase inCreactivity. Themagnitudeof thiseffect probably impacts Cre activity in strains with carboxy-terminal Ert2fusion. These two trap lines and the related line N1IP::CreERT2 form acomplementary mapping tool kit to identify changes in Notch1 activationpatterns in vivo as the consequence of genetic or pharmaceuticalintervention,and illustrate thevariation inNotch1signal strength fromonetissue to the next and across developmental time.

KEY WORDS: Activity, Cre, Neuron, Notch1, Kidney

INTRODUCTIONThe Notch signaling pathway is an evolutionarily conserved, short-range communication mechanism utilized throughout life in allmetazoan species (Artavanis-Tsakonas et al., 1999). The binding of aligand to a Notch receptor triggers the unfolding of a juxtamembranenegative regulatory region (NRR),which exposes it to cleavage by theADAM10 protease (a disintegrin and metalloprotease) at anextracellular sequence termed S2 (Groot et al., 2013; Mumm et al.,2000; van Tetering et al., 2009). S2 cleavage leads to the shedding ofthe extracellular domain and allows γ-secretase to cleaveNotchwithinits transmembrane domain at S3, releasing the intracellular domain ofNotch (NICD) from themembrane. Three nuclear localization signalswithin NICD provide for nuclear translocation, where it complexeswith RBPjk, recruits the MAML adaptor and the transcriptionalactivator p300 (Ep300 – Mouse Genome Informatics) to thepromoters of target genes (Kopan and Ilagan, 2009).Whereas only one receptor (Notch) and two ligands (Delta and

Serrate) are present in Drosophila, mammals have four receptorparalogs (Notch1-4) and at least five canonical ligands (Dll1, Dll3,Dll4, Jag1 and Jag2). These receptors and ligands exhibit complexexpression patterns and play essential roles both duringdevelopment and in the maintenance of adult tissue homeostasis,

which is evident from various Notch-related congenital diseasesand tumors in human patients (Koch and Radtke, 2010; Pentonet al., 2012) and in mice lacking Notch receptors or ligands invarious tissues (Conlon et al., 1995; Demehri et al., 2008; Galeet al., 2004; Hamada et al., 1999).

Several methods have been developed and used to report Notchactivity, each with specific strengths and weaknesses [reviewed byIlagan et al. (2011)]. Taking advantage of the fact that the intracellulardomain of Notch (NICD) requires ligand-induced proteolysis to bereleased from the cell membrane, we have previously described thecreation and analysis of a Notch1 activity-trap mouse line, Notchintramembrane proteolysis-Cre or N1IP::Cre (Vooijs et al., 2007)(Fig. 1A,E). In this knock-in line, the coding sequence for phagerecombinase Cre, which was followed by six myc tags (6MT) andpreceded by a nuclear localization signal (NLS), was targeted into oneNotch1 allele within exon 28, encoding the transmembrane domain ofNotch1 (Fig. 1E). This targeting strategy resulted in the expression ofa Notch1-NLSCre6MT fusion protein from this locus, in which theintracellular domainofNotch1 is completely replacedbyNLSCre6MT.The binding of the ligand to cells expressing Notch1 and Notch1-NLSCre6MT fusion protein triggers the release of NICD and theNLSCre6MT protein, respectively.Whereas the NICD permits normaldevelopment,NLSCre6MT, once inside the nucleus, could catalyze theremoval of any floxed allele in the genome. If a reporter is present, Creactivitywill ‘trap’ the cell and permanentlymark its descendent lineageindependent of any further Notch1 activation (Fig. 1A). We have usedthis line to show that this activity-trap strategy could identify lineagesexperiencing Notch1 activation (Vooijs et al., 2007).

Despite our success, this NIP-Cre reporter line is relativelyinsensitive and seemed to predominantly trap cells in which high orrepeated activation cycles are known to occur (e.g. endothelium),while missing cells in many tissues known to rely on Notch1activity [e.g. the hematopoietic system, (Vooijs et al., 2007)]. Wespeculated that this could be caused by less-efficient transcriptionfrom the manipulated allele and reduced activity by the insertion ofthe 6MT (Vooijs et al., 2007). It was also hypothesized that, in sometissues, sequences required for proper trafficking of Notch to themembrane were lost when NICD was replaced with Cre.

Here, we report that removing the 6MT and adding an additionalpolyadenylation signal significantly improved sensitivity, allowingus to trap cells experiencing lower Notch activation thresholds.When compared with the original N1IP-Cre reporter (nowdesignated as N1IP::CreLowActivity or N1IP::CreLO), this new line(N1IP::CreHighActivity or N1IP::CreHI allele) traps Notch1 activationin most tissues known to rely on Notch1, and identifies novel cellsutilizing Notch1 during their development and/or maintenance,some of which will be reported elsewhere. Importantly, biochemicalanalysis indicates that this new allele is cleaved as efficiently asReceived 6 November 2014; Accepted 23 January 2015

1SAGE Labs, St Louis, MO 63146, USA. 2Department of Developmental Biology,Washington University, St Louis, MO 63110, USA. 3Division of DevelopmentalBiology, Children’sHospital Medical Center, Cincinnati, OH 45229, USA. 4CarnegieInstitution for Science, Department of Embryology, Baltimore, MD 21218, USA.

*Author for correspondence (Raphael.Kopan@cchmc.org)

1193

© 2015. Published by The Company of Biologists Ltd | Development (2015) 142, 1193-1202 doi:10.1242/dev.119529

DEVELO

PM

ENT

Notch1, supporting our previous observation that Notch traffickingrelies on signals coded in the extracellular domain (Liu et al., 2013).This trap line, N1IP::CreLO and the related sibling N1IP::CreERT2

(Liu et al., 2009; Pellegrinet et al., 2011) line, are powerful newtools for mapping alteration in Notch1 activation in vivo as aconsequence of genetic or pharmaceutical intervention.

RESULTS AND DISCUSSIONBiochemical analyses of N1IP::Cre constructs demonstratethat a 6MT C-terminal fusion reduces Cre activity five- totenfoldN1IP::Cre mice contain a 6MT fused to the C-terminus of the Crerecombinase protein (Fig. 1A,E). Similar Cre fusions describedelsewhere include CreErt2, in which a tamoxifen-responsivedomain controls nuclear entry of Cre (Feil et al., 1996; Zhang

et al., 1996). Cre activity is reliant on Tyr 324, located 20 aminoacids from the C-terminus (Guo et al., 1997). Given the low labelingindex of N1IP::Cre in some tissues (Vooijs et al., 2007), we testedwhether this fusion might reflect a negative interference withCre-recombinase enzymatic activity. To test this hypothesis, wedeveloped an in vitro system to quantify Cre activity in primarycells. First, we prepared reporters for Cre activity by isolating mouseembryonic fibroblasts (MEFs) from embryos heterozygous for theRosaR26R allele (Soriano, 1999). In these cells, β-galactosidase(lacZ) will be expressed upon Cre-mediated recombination of asingle locus (Fig. 1C). Second, we constructed plasmids encodingCre and Cre-6MT containing a nuclear localization sequence(NLS), followed by an internal ribosomal entry site (IRES) andfirefly luciferase (Luc, Fig. 1B). In a given transfected cellpopulation, lacZ activity will be a measurable surrogate for

Fig. 1. Creation of N1IP::CreHI. (A) Principle of cleavage-dependent Notch reporter. The replacement of Notch intracellular domain with Cre recombinase leadsto the release of Cre, but not of NICD, from the cell membrane after ligand-induced S2 and S3 cleavage. Cre will activate reporters, such as Rosa lacZ orRosa EYFP, which will remain active in all descendants. (B-D) Comparison of the enzymatic activity of Cre and Cre6mt fusion proteins. (B) Structure of theplasmids used in the analysis. (C) Schema of MEF preparation and testing. (D) Comparison of β-gal activity in R26R-MEF cells expressing similar levels ofluciferase (note that the data are plotted on a log scale). (E) Comparison of N1IP::CreLO and N1IP::CreHI targeted locus. ANK, ankyrin repeats; LNR, Lin-Notchrepeat; TM, transmembrane domain; PEST, proline/glutamic acid/serine/threonine-rich motif; 2pA and 3pA, 2× and 3× polyadenylation signal. (F) Comparison ofmRNA levels among Notch1, N1IP::CreLO and N1IP::CreHI in 9.5-day embryos. (G) Comparison of released Cre levels between N1IP::CreLO and N1IP::CreHI innewborn kidneys by western blot using anti-V1744 antibody. (H) Quantification of released Cre relative to the released N1ICD in the same sample.

1194

RESEARCH ARTICLE Development (2015) 142, 1193-1202 doi:10.1242/dev.119529

DEVELO

PM

ENT

recombinase activity within the population, whereas luciferaseactivity will report relative output from the Cre-expressing plasmids(transfection efficiency), independent of their recombinase activity.WhenMEFswere transfectedwith increasingDNAconcentrations,

luciferase activity scaled linearly throughout the entire range (from7.8to 500 µg, supplementary material Fig. S1), demonstrating that Creand Cre-6MT were expressed at similar levels. However, β-Galactivity in extracts prepared from transfected cells indicated thatNLSCre6MT recombinase activity was significantly diminishedrelative to Cre at every concentration (supplementary materialFig. S1). Notably, measurable β-Gal activity could be detected atthe lowest transfected Cre concentration and continued to increase in arelatively linear fashion,whereasNLSCre6MTactivity increased onlyabove 60 ng (supplementary material Fig. S1). To compare Creactivity at similar protein concentrations, we transfected the MEFswith plasmid concentrations yielding identical photon flux andmeasured β-Gal activity (Fig. 1D). Even at this high plasmidconcentration a sevenfold difference was noted between Cre andCre6MT (Fig. 1D). These results indicate that our original strain,N1IP::Cre mice (Vooijs et al., 2007) and, most likely, the related lineN1IP::CreERT2 (Pellegrinet et al., 2011), under-report Notch activityin vivo. This conclusion applies to all CreErt2 lines driven by weakpromoters, and could explain the differences in labeling observedwhen different Cre fusion proteins are expressed from the same locusas in Grisanti et al. (2013).

The creation of N1IP::CreHI reporter miceAnunfused activity-trap reporter might be able to trapNotch1 activityat lower thresholds missed with the Cre6MT protein. To generatemice with improved Notch-trap ability, we modified our targetingconstructs to create a new line of knock-in reportermouse,N1IP::creHI

(Fig. 1E; supplementary material Fig. S2). First, we removed the6Myc tag from the targeting construct to restore the originalC-terminus. Second, we added an additional polyadenylation signal(AAUAAA) to the SV40 early polyadenylation sequence following

the Cre ORF, thus matching the total number of polyadenylationsignals (three) found in the wild-type Notch1 transcript. Finally, wesubstituted a cytidine for an adenine at position 5178 of exon 28,corresponding to a silent mutation at the third position for L1726 inNotch1 (Fig. 1E; supplementary material Fig. S2). This offers asimple, fast and non-radioactivemethod to pre-screen embryonic stemcells (ESCs) for homologous recombination with pyrosequencing(Liu et al., 2009), and enabled us to compare the abundance ofmRNAproduced by each allele (Liu et al., 2013, 2011). We expected thesemodifications to increase the relative abundance of the N1IP::CreHI

mRNA and to fully restore Cre activity.To directly compare the mRNA level of N1IP::CreHI and

N1IP::CreLO, we mated Notch1+/CreHI (N1IP::CreHI heterozygote)to Notch1+/CreLO (N1IP::CreLO heterozygote) mice and collectedE9.5 Notch1CreHI/CreLO embryos. These animals have no functionalNICD and do not survive beyond E10.5. Notch1+/CreHI embryosserved as a control. From these, we purified total RNA for RT-PCRand performed pyrosequencing to compute the ratio of adenine(N1IP::CreHI) to cytidine (wild-type Notch1 in Notch1+/CreHI

embryos or N1IP::CreLO in Notch1CreHI/CreLO embryos) atnucleotide 5178 in exon 28. The results show that the additionof the extra polyadenylation signal increases the stability ofN1IP::CreHI mRNA ∼twofold, to a level comparable to that ofthe endogenous Notch1 [Fig. 1F; (Liu et al., 2011)]. To compare theamount of the Cre protein that is released from the cell membraneafter ligand binding from the N1IP::CreHI and N1IP::CreLO

proteins, we performed western blot on wild type, N1IP::CreHI

heterozygote and N1IP::CreLO heterozygote newborn kidneys withan antibody against cleaved Notch1 (V1744), thus normalizing thelevel of Cre to the NICD produced from the cleaved endogenousNotch1. Consistently, we found a twofold increase in the abundanceof cleaved Cre relative to Cre6MT (Fig. 1G,H).

Importantly, because N1IP alleles are cleaved as effectively asthe endogenous Notch1 allele in the kidney (Fig. 1G,H), theyestablish unequivocally that sequences in the intracellular domain

Fig. 2. Comparison of the labelingpattern between N1IP::CreHI andN1IP::CreLO at different embryonicstages. N1IP::CreLO; RosaR26R/R26R

(A-D) and N1IP::CreHI; RosaR26R/R26R

(E-M) embryos at different stages stainedwith X-gal. White arrowheads, notochord;yellow arrows, intersegmental arteries;black arrowhead in F, ventral half of thedeveloping CNS. (I) Sagittal section of anE15.5 N1IP::CreHI; RosaR26R/R26R embryostained with X-gal. J and K show coronalsections of E9.5 and E10.5 embryos,respectively. Black arrowheads, notochord.(L) Enlarged view of an E10.5 embryo.Black arrows, intersegmental arteries.(M) Stained sagittal section of the head of anewborn pup. Thyroid gland is labeled witha white arrow.

1195

RESEARCH ARTICLE Development (2015) 142, 1193-1202 doi:10.1242/dev.119529

DEVELO

PM

ENT

beyond the cleavage site do not play a significant role in regulatingthe trafficking of Notch proteins to the proper apical membranedomain. This extends the conclusion derived from kidney epithelia(Liu et al., 2013) to the kidney endothelial cells, the maincontributor ofNotch1 ICD towhole kidney extracts.Whether this isthe case in all metazoans or in all mammalian cell types remains tobe determined.

Comparing N1IP::CreHI and N1IP::CreLO in vivoN1IP::CreHI; RosaR26R/R26R sires were mated with RosaR26R/R26R

dams (Soriano, 1999) to trap cells experiencing Notch1activation. Consistent with the data showing increased Creprotein level and activity, we found that, throughout earlyembryonic development, N1IP::CreHI embryos were moreintensely labeled by whole-mount β-Gal staining than N1IP::CreLO embryos (4°C overnight, Fig. 2). Consistent with what weobserved in vitro, N1IP::CreHI-trapped cells experienced Notch1activity earlier than N1IP::CreLO at all embryonic stages (Fig. 2).Under the same X-gal development conditions, N1IP::CreHI;RosaR26R/R26R embryos contained many labeled cells in the heart(Fig. 2E, asterisk) and the notochord (Fig. 2E, white arrowhead;Fig. 2J, black arrowhead) by E9.5. By contrast, no signal wasdetected in N1IP::CreLO; RosaR26R/R26R embryos before E10.5,when the only labeled cells were found in the heart [Fig. 2A,B;(Vooijs et al., 2007)]. At E10.5, more cells were labeled inN1IP::CreHI hearts (Fig. 2F, asterisk) and most, if not all, inter-segmental arteries (Fig. 2F, yellow arrows; Fig. 2L, black arrows).This indicates that Notch1 is activated earlier than E10.5 inarterial endothelial cells, consistent with genetic observationsand phenotypes (Gridley, 2010). By contrast, the inter-segmentalarteries were sparsely labeled in the N1IP::CreLO mice at E11.5(Fig. 2C). Accordingly, N1IP:CreHI/flox trans-heterozygotes, inwhich the activated N1IP::Cre excises the floxed Notch1 allele,suffer embryonic lethality at E10.5 due to the developmentalfailure of the vascular system in the yolk sac and the embryo (Liuet al., 2011), similar to the result of early endothelial cell-specificconditional knockout of Notch1 (Limbourg et al., 2005) witheither one of two available Tie2-Cre strains (Kisanuki et al., 2001;

Koni et al., 2001). By contrast, all N1IP:CreLO/flox trans-heterozygotes survive without discernible phenotype to P50(Liu et al., 2011). Together, these data suggest that N1IP::CreHI

reports the activation of Notch1 in the endothelium with highfidelity, trapping nearly all cells, whereas N1IP::CreLO reportedactivation only in cells experiencing substantially higher Notchactivation levels (Vooijs et al., 2007). Interestingly, removal ofADAM10 with Tie2-Cre (Kisanuki et al., 2001) was not lethal(Zhao et al., 2014), perhaps allowing sufficient activity to persistthrough an early, crucial developmental window. The use of theselines to elucidate the role of signal strength in endothelial andhematopoietic lineage separation from hemogenic endotheliumwill be described elsewhere.

At E15.5, staining of sagittal sections showed widespreadlabeling in the neural tube, brain, retina, major blood vessels,intestine and vertebrae (Fig. 2I), correlating well with the reportedrole of Notch1 in these tissues. At P0, a majority of the brain islabeled and nerve fibers in the head region can be clearly identifiedwith β-Gal staining (Fig. 2M). Recently, it has been suggested thatNotch signaling regulates thyrocyte differentiation (Carre et al.,2011; Ferretti et al., 2008). Accordingly, we observed widespreadlabeling in the thyroid gland at P0 (Fig. 2M, white arrow).

To assess the behavior of N1IP::CreHI with a different reporterstrain, we crossed N1IP::CreHI and N1IP::CreLO with RosaeYFP/eYFP

mice (Srinivas et al., 2001). We collected kidneys at post-natal day 28(p28) and compared the labeling patterns in the kidney cortex.The entire CD31+ vascular tree within glomeruli is labeled inboth N1IP::Cre lines [Fig. 3, arrows; see also (Liu et al., 2013)],consistent with the robust activation of Notch1 during thedevelopment and maintenance of endothelial cells (Gridley,2010). Despite the activation of Notch1 in the proximal tubuleepithelium during kidney development and its known contributionto the maturation of these cells (Surendran et al., 2010a,b), very littlelabeling is detected within this compartment in N1IP::CreLO;Rosa+/eYFP mice. By contrast, a majority of proximal tubuleepithelia are Notch1-lineage positive in N1IP::CreHI; Rosa+/eYFP

mice (Fig. 3, arrowheads), consistent with its role in thedevelopment and in regulation of tubule diameter (Liu et al.,

Fig. 3. 6Myc-Tag compromises Notch1-Cre activity in vivo. Kidneys were isolated from P28 N1IP::CreLO; Rosa+/eYFP (A-E) and N1IP::CreHI; Rosa+/eYFP

(F-J), and examined for lineage-positive cells in the kidney cortex. Both N1IP::CreLO and N1IP::CreHI label CD31+ endothelial cells within glomeruli(white arrows) and in vessels (asterisks in A-C, not shown in F-H). In N1IP::CreLO mice, the associated tubular epithelium (Aqp1 positive; D,E) surroundingglomeruli infrequently contains labeled cells (white arrowheads). By contrast, a majority of tubular epithelial cells in N1IP::CreHI mice are labeled (I,J). Scale bars:100 µm.

1196

RESEARCH ARTICLE Development (2015) 142, 1193-1202 doi:10.1242/dev.119529

DEVELO

PM

ENT

2013; Surendran et al., 2010a). These data demonstrate that removalof the 6MT and addition of a third polyA site result in higherrecombinase Cre activity in vivo, consistent with our in vitrofindings.Although these patterns alleviate the concern that N1IP::

CreHI might over-report Notch1 activity, it is possible that areporter that recombines at lower Cre concentrations willgenerate a false-positive signature of Notch1 activity. Toevaluate the potential for false positives, we examined thelabeling patterns in the most sensitive reporter line we haveavailable, Ai3 (Madisen et al., 2010). The Ai3 reporterrecombination and expression are highly efficient and servedto determine the false-positive labeling rates in the distaltubules of the kidney, where Notch1 expression is too low to bedetected (Liu et al., 2013). In the Ai3 kidneys, a significantnumber of distal tubule cells experienced N1 activation, butvery few connecting segment cells were labeled (Fig. 4). Thismight indicate a physiological reality in which signal strengthwas insufficient to divert some cells from assuming a distal fate,or the result of background proteolytic release of Cre. Thedecline in labeling frequency from the proximal tubules(∼100%) to the thick ascending limb (62.29±5.2%) to theconnecting segment (11.18±1.8%) is more consistent with theformer possibility; the intense labeling of distal tubule isconsistent with the first.These data demonstrate that the N1IP::CreLO allele only labels the

subset of cells that experience high and/or persistentNotch1 activation.Removing the 6MT increased labeling efficiency in many tissueswhere Notch1 has a known function. It is worth noting that, althoughN1IP::CreHI exhibits vastly improved sensitivity over N1IP::CreLO,not all knownNotch-mediated decisions are captured. For instance, wedid not observe labeling in the somite at E9.5 and E10.5 (Fig. 2E,F,J,K), where Notch1 plays an important role at very low activation levels(Conlon et al., 1995; Huppert et al., 2005; Lewis, 2003).

N1IP::CreHI labeling patterns in the reproductive organsThe Notch signaling pathway plays essential roles in mediating theinteractions between the stem cells and their niche in various organs,including the reproductive organs in Drosophila (Kitadate andKobayashi, 2010; Song et al., 2007; Ward et al., 2006) andC. elegans (Hubbard, 2007). Whereas in the Drosophila ovary theNotch receptor is expressed in the niche, the receptor is required inthe germ line stem cells in C. elegans. In rodent ovaries, Notch2 isexpressed in granulosa cells and the Jag1 ligand is expressed in thegerm cells (Johnson et al., 2001; Vanorny et al., 2014; Xu andGridley, 2013). The ovary of sexually mature virgin N1IP::CreHI;Rosa+/R26R females showed ubiquitous labeling in granulosa cells(Fig. 5F), suggesting that Notch1 acts with Notch2 in the ovary butis not activated in the oocyte (Vanorny et al., 2014). Accordingly,we did not observe germline inheritance of an active reporter inoffspring of N1IP::CreHI; RosaR26R/R26R females.

The role of Notch in the rodent testis remains controversial. Somereports suggest a possible role for Notch1 in the niche duringspermatogenesis (Dirami et al., 2001; Hayashi et al., 2001; Murtaet al., 2013), but functional studies could not demonstrate anessential function for Notch1 (Batista et al., 2012; Hasegawaet al., 2012). If N1IP::CreHI is activated in germ cells duringspermatogenesis, it will result in germline inheritance of an activereporter, manifest in uniform lacZ expression (stained blue byX-gal) in offspring of N1IP::CreHI; RosaR26R/R26R fathers.However, we only saw three blue embryos out of more than 100embryos that we stained, suggesting that Notch1 activation in germcells during spermatogenesis is an infrequent event or too low to bedetected by this line. Examination of the adult testis revealedextensive N1IP::CreHI activation within Sertoli cells (as marked bySOX9 and TUBB3; Fig. 5I), consistent with previous studiesshowing transgenic Notch Reporter GFP (TNR-GFP) activation inSertoli cells during multiple developmental stages (Garcia et al.,2013; Hasegawa et al., 2012). There was also N1IP::CreHI

Fig. 4. Notch1-Cre activity in the nephron of the adult kidney.N1IP::CreHI activity in the distal nephron was detected by theoverlap of EYFP with Tamm–Horsfall staining (left panel, the thickascending limb of the loop of Henle); arrows indicate co-labeling ofEYFP and Tamm–Horsfall glycoprotein. Clcnkb staining (middlepanel) shows distal convoluted tubules (same pattern seen withCdh1+; CytoK8−). Only a few cells were positive for both Calbindinand EYFP (distal tubule and connecting segment, right panel).Scale bar: 50 µm. 20× magnification.

1197

RESEARCH ARTICLE Development (2015) 142, 1193-1202 doi:10.1242/dev.119529

DEVELO

PM

ENT

activation within the interstitial compartment of the testis, probablywithin vascular endothelial cells or perivascular cells, both of whichexpress Notch receptors and activate TNR-GFP during adulthood(DeFalco et al., 2013).By contrast, widespread labeling was detected as early as at E12.5

in the mesonephric tubules, the precursors of the epididymis(Fig. 5A), persisting to the maturing epididymis (Fig. 5B,C).The Wolffian duct in female embryos was labeled before itdegenerates into Gartner’s duct (Fig. 5E). Vasculature at the gonad-mesonephros junction also was labeled (Fig. 5E). Sections ofnewborn and adult epididymis showed that epithelial cells aroundthe lumen of all major segments, including the initial segment,caput, corpus and tails are labeled, albeit at different frequencies(Fig. 5G,H). Anti-K14 and EYFP double-staining of sections fromN1IP::CreHI; Rosa+/eYFPmice showed that all labeled cells are K14-negative principal cells (Fig. 5J). Stained particles in the lumen ofepididymis (Fig. 5K) were negative for DAPI staining, suggestingthat they are not sperm but rather epididymosomes, membranous

vesicles that are derived from principal cells of the epididymis (Saezet al., 2003). This is consistent with the idea that Notch1 activationmight not occur during spermatogenesis, but instead in theepididymis, which is essential for the maturation of the sperm.However, assignment of functional significance will necessitatefurther studies.

N1IP::CreHI labeling pattern in the neural tissueNotch activity is dispensable for establishing the neuroepitheliumas all Notch pathway mutants complete this process (de la Pompaet al., 1997; Oka et al., 1995). We noted faint labeling in theventral half of the developing CNS (black arrow, Fig. 2F). Toidentify the labeled cells we sectioned E10.5 embryos that hadbeen stained for β-Gal activity, and used a stereoscope tophotograph the slices (Fig. 6A-H). Two discrete progenitordomains adjacent to the floor plate were strongly labeled in thediencephalon (Fig. 6A), midbrain (Fig. 6B), hindbrain (Fig. 6C,D)and spinal cord (Fig. 6E-H). Labeling of subpopulations of early-

Fig. 5. Notch1 labeling pattern in thereproductive system. (A-E) Whole-mountstaining of gonads frommale (A-C) and female(D-E) embryos of indicated age and genotype.(F) X-gal staining of an ovary from P53 virginR26R, N1IP::CreHI female. (G,H) X-gal-stained section through epididymis fromimmature (G) and mature (H) R26R, N1IP::CreHI male. (I) Immunofluorescence stainingof adult N1IP::CreHI; Rosa+/eYFP testis showshigh levels of Cre activity in Sertoli cellsidentified with the Sertoli-specific markersSox9 and TUBB3. Note the colocalization ofEYFP with Sox9 or TUBB3. Scale bar:100 µm. (J) Co-staining of K14 and EYFP indifferent segments of the epididymis fromN1IP::CreHI; Rosa+/eYFP male identifieslabeled cells as principal cells. (K) GFP+

epididymosome particles in the lumen ofepididymis fromN1IP::CreHI; Rosa+/eYFPmaleare negative for DAPI. Ep, epididymis; Gd,Gartner’s duct; Kd, kidney; Ms, metanephros;Ov, ovary; Te, testis; Wd, Wolffian duct.

1198

RESEARCH ARTICLE Development (2015) 142, 1193-1202 doi:10.1242/dev.119529

DEVELO

PM

ENT

born neurons in the dorsal aspect of the midbrain and hindbrainwas also detectable (Fig. 6B-D). This stripe pattern had beenobserved previously with Dll1 and Jagged1 expression in thechick hindbrain and neural tube (Marklund et al., 2010; Myatet al., 1996). To align Notch activity stripes with its ligand wecrossed Dll1+/lacZ (Hrabe de Angelis et al., 1997) mice withN1IP::CreHI; RosaeYFP/eYFP mice to generate E10.5 embryosheterozygous for Dll1, Notch1 and Rosa-eYFP. Staining for lacZ(Dll1 expression, Fig. 6I), EYFP (Cre activity) and Jagged1confirmed that the cells experiencing Notch activation overlappedwith the ligand Dll1 (Fig. 6I). To identify the cells thatexperienced high levels of Notch1 activity we stained a sectionof the neural tube from N1IP::CreHI; RosaeYFP/eYFP mice withOlig2 and Nkx2.2, which confirmed that progenitors in the pV2and pV3 domains were labeled much more intensely than those inthe adjacent pMN and pV1 domains (Fig. 6J). We and others haveshown previously that Notch1 signaling was necessary to specifyneuronal subtypes, in particular V2 neurons and motor neurons inthe spinal chord (Del Barrio et al., 2007; Misra et al., 2014; Yanget al., 2006). It would be interesting to investigate whether V2neuron specification requires higher Notch1 signal strength thanother neuronal cell types.In the CNS, neural progenitors undergo multiple rounds of

asymmetric division. Notch signaling oscillates to keep progenitorscycling between committed progenitor state (Hes target geneexpression detected, proneural genes repressed) and pre-neuron[proneural gene expression detected, Hes1 repressed; (Imayoshiet al., 2013; Kobayashi and Kageyama, 2010; Shimojo et al., 2008,2011)]. After each division, the neuron daughter cell is thought toactivate Notch in its sister to reinforce the progenitor fate. A perfectactivity trap would label progenitors after the first Notch-dependentdecision, trapping all of their subsequent descendants (Fig. 7A).Analysis of cortical layers from N1IP::CreHI; Rosa+/Ai3 (Fig. 7B;supplementary material Fig. S3) shows co-labeling of EYFP(N1IP::CreHI activity) and Foxp1 (a neuronal marker for late-bornupper layer neurons: Foxp1 labels neurons in layers II-IVa) in E15.5

cortical neurons (Fig. 7B,B′; supplementary material Fig. S3).EYFP labeling was also detected in FoxP1–, earlier born neuronsand in stem/progenitor cells in the ventricular and subventricularzones (VZ/SVZ) (supplementary material Fig. S4; arrows). Bycontrast, we detected no EYFP labeling in Foxp1+ cortical neuronsof N1IP::CreLO; Rosa+/Ai3 embryos (Fig. 7C; supplementarymaterial Fig. S3). In addition, striatal projection neurons werelabeled very efficiently in N1IP::CreHI; Rosa+/Ai3 but onlymarginally in N1IP::CreLO; Rosa+/Ai3(Fig. 7; supplementarymaterial Fig. S4). These data underscore the differences inactivation threshold between Cre and Cre6MT (and, by inference,CreERT2).

The images in supplementary material Fig. S3 suggest that EYFPlabeling is denser in upper layer neurons than in lower layer neuronsand in progenitors located in the neocortex VZ/SVZ. BecauseNotch1 signaling is thought to be activated in every asymmetricprogenitor division (de la Pompa et al., 1997; Imayoshi et al., 2010),two possible scenarios could fit this pattern. First, a commonprogenitor pool generates both lower layer (early) and upper layer(late) neurons, and late-stage progenitors receive stronger Notch1signals. Alternatively, the sensitive trap allele is still not sensitiveenough, and significant labeling requires gradual accumulation ofCre protein over time, and thus favors late-stage progenitors thatproduce upper layer neurons. We cannot rule out a third possibilitythat separate progenitor pools contribute to lower and upper layerneurons, with stronger signaling through the Notch1 receptor in thelatter.

SummaryHere, we demonstrate that Cre activity is compromised by fusion ofadditional amino acids to the C-terminus. This observation impliesthat all Ert2 fusion proteins are less active than Cre even when equalnuclear concentrations are achieved, as shown in Grisanti et al.(2013). The comparison between the two N1IP::Cre lines revealsthat Notch1 proteolysis, and thus signal strength, vary with timewithin the same tissue (e.g. early- versus late-born cortical neurons,

Fig. 6. Notch1-Cre activity in the developing CNS and neural tube defines a subset of neurons. (A-H) Whole-mount stained E10.5 N1IP::CreHI,Rosa+/R26R

male embryos were cut into slices showing the diencephalon (A), midbrain (B), hindbrain (C,D) and spinal cord (E-H), and were photographed on a stereoscope.(I) History of Notch activity detected with EYFP, Dll1 distribution with lacZ and Jagged1 expression with antibody staining in N1IP::CreHI; Rosa+/eYFP; Dll+/lacZ.(J) EYFP, Olig2 and Nkx2.2 staining identifies EYFP+ progenitors as V2 neurons.

1199

RESEARCH ARTICLE Development (2015) 142, 1193-1202 doi:10.1242/dev.119529

DEVELO

PM

ENT

or anterior versus posterior stem cells along the intestine). Whetherthe increase in signal strength is a consequence of unrelated biologyor an important requirement for the Notch signaling pathwayremains to be determined.

MATERIALS AND METHODSPreparation of MEF cellsMEF cells were prepared under sterile conditions using a modified protocolbased on Takahashi et al. (2007) and Xu (2005). Briefly, RosaR26R/R26R damswere crossed to a CD1 sire and embryos (Rosa+/R26R) were collected at E13.5.Head, visceral organs and theurogenital systemwere removed.The remainingtissues were pooled, washed three times with PBS and minced into ∼2 mmpieces, using a sterile razor blade in a 100 mmdiameter dish containing 20 mlofwarmTrypsin-EDTA (∼25 cuts per embryo). The cellmixturewas pipettedrepeatedly with a 25 ml pipette and incubated at 37°C for 15 min. Afterincubation, 10 ml of fresh Trypsin-EDTA was added, the tissue was furtherdispersed using a 10 ml pipette and placed at 37°C for an additional 15 min.The final cell mixture was transferred to a 50 ml conical tube, 20 ml of MEFmedium (DMEM, 10% FBS, 2 mM glutamine, 1000 U/ml penicillin and1000 µg/ml streptomycin)was added to total volumeof 50 ml, and spun downat 1500 g for 5 min. The supernatant was discarded and the cell pelletresuspended in 1 ml of MEF medium/embryo equivalent. One milliliter ofcell supernatant was plated in T75 flasks with 9 ml of MEF medium andincubated overnight at 37°C in 5%CO2. The next day, cellswerewashed oncewith PBS to remove dead/non-adherent cells and fresh medium was added.Cells were frozen at passage 1. All experiments were performed on cellsexpanded to passage 3 (P3).

Transfection of MEF cellsFor transfection, 3.75×104 cells/well P3 MEFs were plated in theafternoon before transfection in 24-well tissue culture plates with 500 µlof MEF medium (without penicillin/streptomycin). Transfection wasperformed using the Lipofectamine LTX system (Invitrogen) according tothe manufacturer’s protocol. For each well, 0.5 µg total DNA (with theempty pCS2 vector added to reach the final concentration), 3.25 μl of LTXreagent and PLUS reagent at a 1:1 ratio with DNA were used. Thetransfection mixture was incubated with the cells at 37°C for 18 h andreplaced with fresh MEF medium. In control experiments using GFPplasmids, ∼70% transfection efficiency was observed. Firefly luciferaseand β-gal activity were measured 24 h later, as previously described(Ilagan et al., 2011).

Generation of N1IP::CreHI knock-in miceIn brief, the early SV40 polyadenylation sequence with one extraartificially introduced polyadenylation signal (AATAAA) was cloned infront of the FRT-flanked-neomycin selective cassette. The silent mutationin L1726 was introduced with a site-directed mutagenesis kit (Stratagene)into the SKB-L genomic clone that encompasses mouse Notch1 exons26-32 (Vooijs et al., 2007). The SgrAI/SalI-digested NLScre fragmentand the SalI/BglII-digested SV40 polyadenylation sequence-FRT-neocassette were then cloned into SKB-L through the SgrAI/BamHI site toreplace the last 105 bases of exon 28 and to delete exons 29 and 30(please note that BamHI and BglII share a compatible cohesive end, andafter ligation the BamHI site is lost) (supplementary material Fig. S2).This plasmid was then sequenced and linearized with EcoRI for ESCelectroporation.

ESC targeting and screeningWe targeted the Notch1 locus using 129X1Sv/J-derived SCC10 ESCs (forfurther information regarding this line see http://escore.im.wustl.edu/SubMenu_celllines/wtEScell_linesSCC10.html) and selected for G418-resistant ESC colonies. Cells were first pre-screened with pyroscreening,utilizing the C5178A polymorphism (Liu et al., 2009), and those with a C/Aratio between 40 and 60% at this position were further screened by both long-range PCR and Southern blot to confirm bona fide homologous recombinants[HR, (Liu et al., 2009); supplementary material Fig. S2)]. Positive HRs werekaryotyped, and one line (ES190) was injected into C57B6 blastocysts. Theresulting chimera male micewere mated towild-type C57B6 female mice. F1males were mated to FRT deleter mice (Rodriguez et al., 2000) to remove theFRT-flanked neomycin cassette.

Western blotWestern blot was performed as described in Liu et al. (2013). Rabbitmonoclonal anti-cleaved Notch1 (anti-V1744, Cell Signaling) antibodieswere used at 1:1000 dilution.

X-gal stainingX-gal staining was performed as described in Vooijs et al. (2007). Briefly,embryos or freshly dissected tissues were fixed in 4% PFA in 1× PBSsupplemented with 2 mMMgCl2 on ice for 1-2 h and washed with 1× PBS+2 mM MgCl2. For whole-mount staining, embryos were developed in X-gal staining solution at 4°C overnight; for frozen section staining, tissueswere immersed in 30% sucrose in 1× PBS+2 mM MgCl2 overnight and

Fig. 7. Notch1-Cre activity in neuronal precursors. Coronal sections through the telencephalic region at E15.5 detected EYFP staining as an indication ofNotch1-Cre activity in Rosa+/Ai3 developing neurons. (A) Schematic illustrating the likelihood that a precursor will be labeled by N1IP::CreLO or N1IP::CreHI. Theprobability is depicted as a gradient. G, glia; ND, not detected. (B,C) The neuronal marker FoxP1 labels upper cortical layers (boxed regions in B′,C′) andstriatal projection neurons (boxed regions in B″,C″) in the N1IP::CreHI (B) and N1IP::CreLO (C). See supplementary material Fig. S3 for higher magnification viewof the boxed regions.

1200

RESEARCH ARTICLE Development (2015) 142, 1193-1202 doi:10.1242/dev.119529

DEVELO

PM

ENT

embedded in optimal cutting medium (OCT, Tissue-Tek) and stained in X-gal staining solution at 37°C overnight or until signal was detected.

ImmunohistochemistryImmunohistochemistry was performed as described in Liu et al. (2013).Briefly, frozen sections were blocked with normal donkey serum and thenincubated with primary antibodies at 4°C overnight. Following three timeswashing with 1× PBS, the sections were incubated with fluorescence-conjugated secondary antibodies, washed with 1× PBS and mounted forimaging. The following primary antibodies were used: chicken anti-GFP(1:500; Aves Lab, GFP-1020) or rabbit anti-GFP (1:1000; Invitrogen,A11122), rabbit anti-Foxp1 (1:2000; Abcam, ab16645), mouse anti-Calbindin (1:500; Sigma-Aldrich, C9848), rabbit anti-Tamm–Horsfallprotein (THP) (1:250; Biomedical Technologies, BT-590), rabbit anti-Clcnkb (1:200; GeneTex, GTX47709), rabbit anti-olig2 (1:500; Millipore,AB9610), goat anti-Nkx2.2 (1:50; Santa Cruz Biotech, SC-15015), chickenanti-K14 (1:200; a generous gift from Dr Julie Segre, National HumanGenome Research Institute, National Institutes of Health), goat anti-Jag1(1:100; Santa Cruz Biotech, SC-6011), rabbit anti-β-galactosidase (1:5000;Cappel Labs/MP Biomedicals, 08559761), rat anti-CD31 (1:100; BDBioscience, 550274), goat anti-aquaporin-2 (1:200; Santa Cruz Biotech,SC-9882), rabbit anti-Sox9 (1:4000; Millipore, #AB5535) and mouse anti-TUBB3 (TUJ1) (1:1000; Covance, #MMS-435P). Images were acquiredwith either ApoTome2 (Zeiss) or a Nikon 90i upright wide-fieldmicroscope. Images were processed using Nikon Elements software,Photoshop or Canvas.

AcknowledgementsWe thank members of the Kopan laboratory for many fruitful discussions and MsMary Fulbright for expert technical assistance. Special thanks to Drs Tony De Falco(reproductive system), Kenny Campbell and Masato Nakafuku (central nervoussystem) for expert advice and for critical reading of the manuscript.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsZ.L. and R.K. conceived the project and wrote the manuscript; Z.L., E.B., S.B., S.C.designed and performed experiments; M.T., Y.G. and R.G. performed experiments.

FundingY.G. was supported fully, and R.K., Z.L., S.C. and M.T. were supported in part by theNational Institutes of Health (NIH) [R01GM55479 awarded to R.K.]. S.B. and R.G.were supported by the NIH [R01DK066408 awarded to R.K.]. R.K. and E.B. weresupported in part by theWilliamK. Schubert Chair for Pediatric Research. Depositedin PMC for release after 12 months.

Supplementary materialSupplementary material available online athttp://dev.biologists.org/lookup/suppl/doi:10.1242/dev.119529/-/DC1

ReferencesArtavanis-Tsakonas, S., Rand, M. D. and Lake, R. J. (1999). Notch signaling: cellfate control and signal integration in development. Science 284, 770-776.

Batista, F., Lu, L., Williams, S. A. and Stanley, P. (2012). Complex N-glycans areessential, but core 1 and 2 mucin O-glycans, O-fucose glycans, and NOTCH1 aredispensable, for mammalian spermatogenesis. Biol. Reprod. 86, 179.

Carre, A., Rachdi, L., Tron, E., Richard, B., Castanet, M., Schlumberger, M.,Bidart, J.-M., Szinnai, G. and Polak, M. (2011). Hes1 is required for appropriatemorphogenesis and differentiation during mouse thyroid gland development.PLoS ONE 6, e16752.

Conlon, R. A., Reaume, A. G. and Rossant, J. (1995). Notch1 is required for thecoordinate segmentation of somites. Development 121, 1533-1545.

de la Pompa, J. L., Wakeham, A., Correia, K. M., Samper, E., Brown, S.,Aguilera, R. J., Nakano, T., Honjo, T., Mak, T. W., Rossant, J. et al. (1997).Conservation of the Notch signalling pathway in mammalian neurogenesis.Development 124, 1139-1148.

DeFalco, T., Saraswathula, A., Briot, A., Iruela-Arispe, M. L. and Capel, B.(2013). Testosterone levels influence mouse fetal Leydig cell progenitors throughnotch signaling. Biol. Reprod. 88, 91.

Del Barrio, M. G., Taveira-Marques, R., Muroyama, Y., Yuk, D.-I., Li, S., Wines-Samuelson, M., Shen, J., Smith, H. K., Xiang, M., Rowitch, D. et al. (2007).A regulatory network involving Foxn4, Mash1 and delta-like 4/Notch1 generatesV2a and V2b spinal interneurons from a common progenitor pool. Development134, 3427-3436.

Demehri, S., Liu, Z., Lee, J., Lin, M.-H., Crosby, S. D., Roberts, C. J., Grigsby,P. W., Miner, J. H., Farr, A. G. and Kopan, R. (2008). Notch-deficient skininduces a lethal systemic B-lymphoproliferative disorder by secreting TSLP, asentinel for epidermal integrity. PLoS Biol. 6, e123.

Dirami, G., Ravindranath, N., Achi, M. V. andDym,M. (2001). Expression of Notchpathway components in spermatogonia and Sertoli cells of neonatal mice.J. Androl. 22, 944-952.

Feil, R., Brocard, J., Mascrez, B., LeMeur, M., Metzger, D. and Chambon, P.(1996). Ligand-activated site-specific recombination in mice. Proc. Natl. Acad.Sci. USA 93, 10887-10890.

Ferretti, E., Tosi, E., Po, A., Scipioni, A., Morisi, R., Espinola, M. S., Russo, D.,Durante, C., Schlumberger, M., Screpanti, I. et al. (2008). Notch signaling isinvolved in expression of thyrocyte differentiation markers and is down-regulatedin thyroid tumors. J. Clin. Endocrinol. Metab. 93, 4080-4087.

Gale, N. W., Dominguez, M. G., Noguera, I., Pan, L., Hughes, V., Valenzuela,D. M., Murphy, A. J., Adams, N. C., Lin, H. C., Holash, J. et al. (2004).Haploinsufficiency of delta-like 4 ligand results in embryonic lethality due to majordefects in arterial and vascular development. Proc. Natl. Acad. Sci. USA 101,15949-15954.

Garcia, T. X., DeFalco, T., Capel, B. and Hofmann, M.-C. (2013). Constitutiveactivation of NOTCH1 signaling in Sertoli cells causes gonocyte exit fromquiescence. Dev. Biol. 377, 188-201.

Gridley, T. (2010). Notch signaling in the vasculature. Curr. Top. Dev. Biol. 92,277-309.

Grisanti, L., Clavel, C., Cai, X., Rezza, A., Tsai, S.-Y., Sennett, R., Mumau, M.,Cai, C.-L. and Rendl, M. (2013). Tbx18 targets dermal condensates for labeling,isolation, and gene ablation during embryonic hair follicle formation. J. Invest.Dermatol. 133, 344-353.

Groot, A. J., Cobzaru,C.,Weber, S., Saftig, P., Blobel, C. P., Kopan,R., Vooijs,M.and Franzke, C.-W. (2013). Epidermal ADAM17 is dispensable for notchactivation. J. Invest. Dermatol. 133, 2286-2288.

Guo, F., Gopaul, D. N. and van Duyne, G. D. (1997). Structure of Cre recombinasecomplexed with DNA in a site-specific recombination synapse.Nature 389, 40-46.

Hamada,Y.,Kadokawa,Y.,Okabe,M., Ikawa,M.,Coleman,J.R. andTsujimoto,Y.(1999). Mutation in ankyrin repeats of the mouse Notch2 gene induces earlyembryonic lethality. Development 126, 3415-3424.

Hasegawa, K., Okamura, Y. and Saga, Y. (2012). Notch signaling in Sertoli cellsregulates cyclical gene expression of Hes1 but is dispensable for mousespermatogenesis. Mol. Cell. Biol. 32, 206-215.

Hayashi, T., Kageyama, Y., Ishizaka, K., Xia, G., Kihara, K. and Oshima, H.(2001). Requirement of Notch 1 and its ligand jagged 2 expressions forspermatogenesis in rat and human testes. J. Androl. 22, 999-1011.

Hrabe de Angelis, M., Mclntyre, J.II and Gossler, A. (1997). Maintenance ofsomite borders in mice requires the Delta homologue Dll1. Nature 386, 717-721.

Hubbard, E. J. A. (2007). Caenorhabditis elegans germ line: a model for stem cellbiology. Dev. Dyn. 236, 3343-3357.

Huppert, S. S., Ilagan, M. X. G., De Strooper, B. andKopan, R. (2005). Analysis ofNotch function in presomitic mesoderm suggests a γ-secretase-independent rolefor presenilins in somite differentiation. Dev. Cell 8, 677-688.

Ilagan, M. X. G., Lim, S., Fulbright, M., Piwnica-Worms, D. and Kopan, R. (2011).Real-time imaging of notch activation with a luciferase complementation-basedreporter. Sci. Signal. 4, rs7.

Imayoshi, I., Sakamoto, M., Yamaguchi, M., Mori, K. and Kageyama, R. (2010).Essential roles of Notch signaling in maintenance of neural stem cells indeveloping and adult brains. J. Neurosci. 30, 3489-3498.

Imayoshi, I., Shimojo, H., Sakamoto, M., Ohtsuka, T. and Kageyama, R. (2013).Genetic visualization of notch signaling in mammalian neurogenesis. Cell. Mol.Life Sci. 70, 2045-2057.

Johnson, J., Espinoza, T., McGaughey, R. W., Rawls, A. and Wilson-Rawls, J.(2001). Notch pathway genes are expressed in mammalian ovarian follicles.Mech. Dev. 109, 355-361.

Kisanuki, Y. Y., Hammer, R. E., Miyazaki, J.-i., Williams, S. C., Richardson, J. A.and Yanagisawa, M. (2001). Tie2-Cre transgenic mice: a new model forendothelial cell-lineage analysis in vivo. Dev. Biol. 230, 230-242.

Kitadate, Y. and Kobayashi, S. (2010). Notch and Egfr signaling actantagonistically to regulate germ-line stem cell niche formation in Drosophilamale embryonic gonads. Proc. Natl. Acad. Sci. USA 107, 14241-14246.

Kobayashi, T. andKageyama, R. (2010). Hes1 oscillation: making variable choicesfor stem cell differentiation. Cell Cycle 9, 207-208.

Koch, U. and Radtke, F. (2010). Notch signaling in solid tumors. Curr. Top. Dev.Biol. 92, 411-455.

Koni, P. A., Joshi, S. K., Temann, U.-A., Olson, D., Burkly, L. and Flavell, R. A.(2001). Conditional vascular cell adhesion molecule 1 deletion in mice: impairedlymphocyte migration to bone marrow. J. Exp. Med. 193, 741-754.

1201

RESEARCH ARTICLE Development (2015) 142, 1193-1202 doi:10.1242/dev.119529

DEVELO

PM

ENT

Kopan, R. and Ilagan, M. X. G. (2009). The canonical Notch signaling pathway:unfolding the activation mechanism. Cell 137, 216-233.

Lewis, J. (2003). Autoinhibition with transcriptional delay: a simple mechanism forthe zebrafish somitogenesis oscillator. Curr. Biol. 13, 1398-1408.

Limbourg, F. P., Takeshita, K., Radtke, F., Bronson, R. T., Chin, M. T. and Liao,J. K. (2005). Essential role of endothelial Notch1 in angiogenesis.Circulation 111,1826-1832.

Liu, Z., Obenauf, A. C., Speicher, M. R. and Kopan, R. (2009). Rapid identificationof homologous recombinants and determination of gene copy number withreference/query pyrosequencing (RQPS). Genome Res. 19, 2081-2089.

Liu, Z., Turkoz, A., Jackson, E. N., Corbo, J. C., Engelbach, J. A., Garbow, J. R.,Piwnica-Worms, D. R. and Kopan, R. (2011). Notch1 loss of heterozygositycauses vascular tumors and lethal hemorrhage in mice. J. Clin. Invest. 121,800-808.

Liu, Z., Chen, S., Boyle, S., Zhu, Y., Zhang, A., Piwnica-Worms, D. R., Ilagan,M. X. G. and Kopan, R. (2013). The extracellular domain of Notch2 increases itscell-surface abundance and ligand responsiveness during kidney. Dev. Cell 25,585-598.

Madisen, L., Zwingman, T. A., Sunkin, S. M., Oh, S. W., Zariwala, H. A., Gu, H.,Ng, L. L., Palmiter, R. D., Hawrylycz, M. J., Jones, A. R. et al. (2010). A robustand high-throughput Cre reporting and characterization system for the wholemouse brain. Nat. Neurosci. 13, 133-140.

Marklund, U., Hansson, E. M., Sundstrom, E., de Angelis, M. H., Przemeck,G. K. H., Lendahl, U., Muhr, J. and Ericson, J. (2010). Domain-specific control ofneurogenesis achieved through patterned regulation of Notch ligand expression.Development 137, 437-445.

Misra, K., Luo, H., Li, S., Matise, M. and Xiang, M. (2014). Asymmetric activation ofDll4-Notch signaling by Foxn4 and proneural factors activates BMP/TGFbetasignaling to specify V2b interneurons in the spinal cord. Development 141,187-198.

Mumm, J. S., Schroeter, E. H., Saxena, M. T., Griesemer, A., Tian, X., Pan, D. J.,Ray, W. J. and Kopan, R. (2000). A ligand-induced extracellular cleavageregulates gamma-secretase-like proteolytic activation of Notch1. Mol. Cell 5,197-206.

Murta, D., Batista, M., Silva, E., Trindade, A., Henrique, D., Duarte, A. andLopes-da-Costa, L. (2013). Dynamics of Notch pathway expression duringmouse testis post-natal development and along the spermatogenic cycle. PLoSONE 8, e72767.

Myat, A., Henrique, D., Ish-Horowicz, D. and Lewis, J. (1996). A chick homologueof serrate and its relationship with Notch and delta homologues during centralneurogenesis. Dev. Biol. 174, 233-247.

Oka, C., Nakano, T., Wakeham, A., de la Pompa, J. A., Mori, C., Sakai, T.,Okazaki, S., Kawaichi, M., Shiota, K., Mak, T. W. and Honjo, T. (1995).Disruption of the mouse RBP-Jkappa gene results in early embryonic death.Development 121, 3291-3301.

Pellegrinet, L., Rodilla, V., Liu, Z., Chen, S., Koch, U., Espinosa, L., Kaestner,K. H., Kopan, R., Lewis, J. and Radtke, F. (2011). Dll1- and Dll4-mediated Notchsignaling are required for homeostasis of intestinal stem cells. Gastroenterology140, 1230-1240.e7.

Penton, A. L., Leonard, L. D. and Spinner, N. B. (2012). Notch signaling in humandevelopment and disease. Semin. Cell Dev. Biol. 23, 450-457.

Rodriguez, C. I., Buchholz, F., Galloway, J., Sequerra, R., Kasper, J., Ayala, R.,Stewart, A. F. and Dymecki, S. M. (2000). High-efficiency deleter mice show thatFLPe is an alternative to Cre-loxP. Nat. Genet. 25, 139-140.

Saez, F., Frenette, G. and Sullivan, R. (2003). Epididymosomes and prostasomes:their roles in posttesticular maturation of the sperm cells. J. Androl. 24, 149-154.

Shimojo, H., Ohtsuka, T. and Kageyama, R. (2008). Oscillations in notch signalingregulate maintenance of neural progenitors. Neuron 58, 52-64.

Shimojo, H., Ohtsuka, T. and Kageyama, R. (2011). Dynamic expression of notchsignaling genes in neural stem/progenitor cells. Front. Neurosci. 5, 78.

Song, X., Call, G. B., Kirilly, D. and Xie, T. (2007). Notch signaling controlsgermline stem cell niche formation in the Drosophila ovary. Development 134,1071-1080.

Soriano, P. (1999). Generalized lacZ expression with the ROSA26 Cre reporterstrain. Nat. Genet. 21, 70-71.

Srinivas, S., Watanabe, T., Lin, C.-S., William, C. M., Tanabe, Y., Jessell, T. M.and Costantini, F. (2001). Cre reporter strains produced by targeted insertion ofEYFP and ECFP into the ROSA26 locus. BMC Dev. Biol. 1, 4.

Surendran, K., Boyle, S., Barak, H., Kim, M., Stomberski, C., McCright, B. andKopan, R. (2010a). The contribution of Notch1 to nephron segmentation in thedeveloping kidney is revealed in a sensitized Notch2 background and can beaugmented by reducing Mint dosage. Dev. Biol. 337, 386-395.

Surendran, K., Selassie, M., Liapis, H., Krigman, H. and Kopan, R. (2010b).Reduced Notch signaling leads to renal cysts and papillary microadenomas.J. Am. Soc. Nephrol. 21, 819-832.

Takahashi, K., Okita, K., Nakagawa, M. and Yamanaka, S. (2007). Induction ofpluripotent stem cells from fibroblast cultures. Nat. Protoc. 2, 3081-3089.

van Tetering, G., van Diest, P., Verlaan, I., van der Wall, E., Kopan, R. andVooijs, M. (2009). Metalloprotease ADAM10 is required for Notch1 site 2cleavage. J. Biol. Chem. 284, 31018-31027.

Vanorny, D. A., Prasasya, R. D., Chalpe, A. J., Kilen, S. M. and Mayo, K. E.(2014). Notch signaling regulates ovarian follicle formation and coordinatesfollicular growth. Mol. Endocrinol. 28, 499-511.

Vooijs, M., Ong, C.-T., Hadland, B., Huppert, S., Liu, Z., Korving, J., van denBorn, M., Stappenbeck, T., Wu, Y., Clevers, H. et al. (2007). Mapping theconsequence of Notch1 proteolysis in vivo with NIP-CRE. Development 134,535-544.

Ward, E. J., Shcherbata, H. R., Reynolds, S. H., Fischer, K. A., Hatfield, S. D. andRuohola-Baker, H. (2006). Stem cells signal to the niche through the Notchpathway in the Drosophila ovary. Curr. Biol. 16, 2352-2358.

Xu, J. (2005). Preparation, culture, and immortalization of mouse embryonicfibroblasts. Curr. Protoc. Mol. Biol. Chapter 28, unit 28.1.

Xu, J. and Gridley, T. (2013). Notch2 is required in somatic cells for breakdown ofovarian germ-cell nests and formation of primordial follicles. BMC Biol. 11, 13.

Yang, X., Tomita, T., Wines-Samuelson, M., Beglopoulos, V., Tansey, M. G.,Kopan, R. and Shen, J. (2006). Notch1 signaling influences V2 interneuron andmotor neuron development in the spinal cord. Dev. Neurosci. 28, 102-117.

Zhang, Y., Riesterer, C., Ayrall, A.-M., Sablitzky, F., Littlewood, T. D. and Reth,M. (1996). Inducible site-directed recombination in mouse embryonic stem cells.Nucleic Acids Res. 24, 543-548.

Zhao, R., Wang, A., Hall, K. C., Otero, M., Weskamp, G., Zhao, B., Hill, D.,Goldring, M. B., Glomski, K. and Blobel, C. P. (2014). Lack of ADAM10 inendothelial cells affects osteoclasts at the chondro-osseus junction. J. Orthop.Res. 32, 224-230.

1202

RESEARCH ARTICLE Development (2015) 142, 1193-1202 doi:10.1242/dev.119529

DEVELO

PM

ENT