The atypical mammalian ligand Delta-like homologue 1 (Dlk1) can regulate Notch signalling in Drosophila
- 6.6k Downloads
Mammalian Delta-like 1 (Dlk-1) protein shares homology with Notch ligands but lacks a critical receptor-binding domain. Thus it is unclear whether it is able to interact with Notch in vivo. Unlike mammals, Drosophila have a single Notch receptor allowing a simple in vivo assay for mammalian Dlk1 function.
Here we show that membrane-bound DLK1 can regulate Notch leading to altered cellular distribution of Notch itself and inhibiting expression of Notch target genes. The resulting adult phenotypes are indicative of reduced Notch function and are enhanced by Notch mutations, confirming that DLK1 action is antagonistic. In addition, cells expressing an alternative Dlk1 isoform exhibit alterations in cell size, functions previously not attributed to Notch suggesting that DLK1 might also act via an alternative target.
Our results demonstrate that DLK1 can regulate the Notch receptor despite its atypical structure.
KeywordsNotch Signalling Notch Receptor Notch Ligand Notch Activity Wing Margin
Notch-ligands are transmembrane proteins that are characterised by a series of EGF-repeats in their extracellular domain, and an N-terminal domain, referred to as the DSL domain (Fig. 1; [6, 7, 8]). The latter appears to be critical for the interactions between the ligands and the Notch receptor [6, 9]. Although DLK1 contains 6 EGF-repeats that are closely related to those found in Delta and Serrate/Jagged, and can bind to Notch EGF repeats in a yeast two-hybrid assay , it lacks the N-terminal DSL-domain which is required for receptor binding and activity of known ligands (Fig. 1); it also differs from characterised Notch ligands because its intracellular domain is considerably shorter. Truncation of the intracellular domain renders Drosophila or Xenopus Delta proteins incapable of activating the receptor [11, 12, 13]. Both its lack of a DSL-domain and its short intracellular domain suggest therefore that DLK1 would not be able to function as a Notch ligand. However, as an inverse correlation between Dlk1 levels and Notch activity have been observed in cultured cells, it is possible that DLK1 could be a receptor antagonist .
The Notch pathway is highly conserved throughout the animal kingdom and has been most extensively studied in Drosophila where it was first identified [7, 15]. Unlike mammals, which have 4 different Notch receptors, Drosophila has a single receptor and provides a simple system to test whether DLK1 has the capability to regulate Notch in vivo and whether it acts positively or negatively. This approach has successfully been used to investigate functions of bone fide mammalian Delta ligands, two of which were effective in activating Drosophila Notch . We therefore generated transgenic flies that express Dlk1 isoforms under the control of the Gal4/UAS system  and tested for phenotypes indicative of alterations in Notch function.
Results and Discussion
Phenotypes from expressing Dlk1 in the Drosophila wing
Three different Dlk1 variants were constructed that mimic the different isoforms detected in vivo: S-Dlk, a secreted form produced by proteolytic cleavage in the extracellular domain, M-Dlk, transmembrane tethered form that lacks the cleavage site, SM-Dlk retains the cleavage site and thus has the capacity to be both membrane tethered and cleaved (Fig. 1). We used a series of different Gal4 driver lines to express these proteins in the developing wing and assessed the adults for phenotypes that would be indicative of effects on Notch activity.
Notch activity is required at many different stages in wing/notum development. Characteristic phenotypes of Notch gain of function are wing over-growth, ectopic wing margin structures, loss of veins and bristle loss [18, 19, 20, 21]. Phenotypes caused by Notch loss of function are notching of the wing margin, vein thickening and bristle duplication/tufting or bristle loss [22, 23]. The apparently contradictory effects on bristles occur because Notch is involved both in the selection of the sensory organ precursor (SOP) cells and in the fates within each bristle lineage . Expression of Drosophila ligands in the wing produces characteristics of both Notch activation and inhibition. The latter occurs in the cells expressing highest levels of ligands, and is due to an autonomous inhibitory effect on the receptor referred to as cis-inhibition [9, 24]. Expression of mammalian ligands in similar assays either resulted primarily in Notch activation (Dll1 and Dll4) or had no effect (Dll3), none had robust cis-inhibitory effects .
To further investigate the consequences of expressing Dlk1 we focussed on the ptc::Gal4 driver which is expressed in a stripe in the central region of the wing and in the scutellum (Fig. 2A,B). We used two approaches to vary the levels of expression. We tested effects at 25°C and 30°C, the higher temperature increasing the effectiveness of the Gal4, and we compared phenotypes from one and two copies of the transgenes. Expression from one copy of M-Dlk at 25°C resulted in highly penetrant phenotypes of wing notches and multiple sense-organs, indicative of reduced Notch activity (Fig. 2E,F,M,N). Both defects were strongly enhanced at 30°C. In contrast, minor defects of occasional bristle duplications were produced using one copy of SM-Dlk (data not shown). The bristle defects were more penetrant with two copies of SM-Dlk and occasional wings had mild-notching (Fig. 2N). These effects were considerably enhanced at 30°C (Fig. 1D,N). Thus both M-Dlk and SM-Dlk produce defects consistent with an antagonistic effect on Notch. The transgene producing the membrane tethered M-Dlk has a much more potent effect.
Cell numbers (+/- s.d.) in two regions of wings from wild type flies and flies expressing Dlk isoforms.
N/+; SM 30°C
N/+; SM 25°C
Some Dlk1 expression phenotypes are enhanced by Notch mutations
To confirm that phenotypes produced by M-Dlk and SM-Dlk are due to effects on Notch, we asked whether they could be enhanced if the levels of Notch were reduced in flies heterozygous for a loss-of-function Notch allele (N55e11). Heterozygosity for Notch alone produces mild wing nicking, slight broadening of the veins at the tips (deltas) and occasional duplication of the scutellar macrochaetae (Fig. 2G and data not shown). Even taking into account the effects of Notch mutation on the wing margin, the frequency and the extent of wing notching from both M-Dlk and SM-Dlk was enhanced in N/+ females (Fig. 2H,N). Furthermore, the effects of bristle duplication were also significantly enhanced (Fig. 2I–N). Thus by two different criteria, the phenotypes produced by DLK1 are enhanced by reducing the amount of Notch protein present, indicating that DLK1 is antagonising the activity of the Drosophila Notch receptor.
Although most of the effects of Dlk1 expression are enhanced by Notch mutations, the wings from Notch/+; ptc::Gal4/UAS::SM-Dlk have a similar size and number of cells in the domain of expression to those with wild-type dose of Notch. This is consistent with the hypothesis that the effects on cell size are independent of an action on Notch raising the possibility that there is an additional target of DLK1.
Effects of Dlk1 expression on Notch targets in the wing imaginal disk
The second target used was the Enhancer of split mβ gene (E(spl)mβ), which is expressed in response to Notch in a more widespread pattern in the wing (Fig. 3E; [32, 33]. We have previously mapped the regulatory sequences of this gene and shown that its expression is dependent on Notch activity . Using a reporter construct where the regulatory sequences drive expression of a heterologous protein, the membrane protein CD2  we tested effects of Dlk1 expression. M-Dlk strongly down-regulated E(spl)mβ::CD2 throughout the domain of misexpression (Fig. 3F; occasionally this was accompanied by slight upregulation at the boundary) SM-Dlk had weaker but still significant inhibitory effects (data not shown). The inhibition of both E(spl)mβ and cut expression confirms that expression of Dlk1 is antagonising Notch activity. As this inhibition occurs within the domain of ligand expression, it most likely represents cis-inhibition, i.e. effects on the receptor in the same cell. However we cannot rule out the possibility that there are slight non-autonomous inhibitory effects with M-Dlk, as the epithelium often becomes deformed around the domain of expression.
We also wanted to verify that the effects of Dlk1 on bristles reflect a reduction in Notch activity, because some of the adult phenotypes (most notably bristle loss) can also be brought about by increased Notch activity. We therefore looked at a marker of sensory-organ precursors, Senseless . If DLK1 is antagonising Notch activity we would expect an increase the number of sensory organ precursors in a given cluster. This is what was observed in the clusters that will form the chordotonal sensory organs of the dorsal and ventral radius (Fig. 3G,H). Both M-Dlk and SM-Dlk increase the number of Senseless-expressing precursors in these clusters, with M-Dlk having the stronger effect (Fig 3H and data not shown).
M-Dlk and SM-Dlk show different cellular locations and effects on Notch protein
Co-staining with anti-Notch antibody revealed that M-Dlk strongly affects the distribution of the endogenous Notch protein (Fig. 4A',B',C'). The levels of Notch are elevated in M-Dlk expressing cells, and there is more protein accumulating at or close to the membrane of these cells. Both these effects suggest that M-Dlk has an action on Notch that stabilises the protein. Notch distribution is much less affected by SM-Dlk, despite the fact that this isoform consistently accumulates to much higher levels than M-Dlk (Fig. 4D'). Nevertheless in some cells we also observed redistribution of Notch to co-localise with the SM-Dlk (Fig. 4D',E'). We take this co-variance in Notch localisation with the expressed Dlk isoforms to indicate that they are able to interact, since this resembles the effects seen with expression of known ligands with functional DSL domains . In contrast, in these previous studies, ligands with a mutated DSL domain were not able to influence Notch distribution .
Dlk1 can regulate Notch signalling despite its atypical structure
Our data indicate that DLK1 has the capability to regulate Notch receptors in spite of the fact that it lacks an N-terminal DSL domain. This in vivo functional demonstration that DLK1 can affect Drosophila Notch signalling and is consistent with yeast two-hybrid data showing interaction of DLK1 with Notch1 EGF repeats . That this interaction has likely functional consequences in mammals in vivo, is supported by evidence suggesting that DLK1 can modulate levels of Notch signalling in adipogenic cells in culture . In our assay the effects of DLK1 appear to be largely inhibitory. This is in contrast to the effects seen when the typical mammalian Delta ligands Dll1, Dll3 and Dll4 were expressed in flies . In those experiments, both DLL1 and DLL4 were able to activate Drosophila Notch, and DLL3 had no effect. Neither DLL3 not the other two ligands produced inhibitory effects similar to those seen here with DLK1. Thus, these effects are likely to reflect specific characteristics of DLK1 itself.
Notch ligands have the unusual characteristic that they inhibit the Notch receptors present on the same cell, as well as activating receptors on adjacent cells [24, 36]. It is the autonomous cis-inhibitory activity that we detect with DLK1, with little or no evidence for activation. Furthermore, the effects are most profound when we use a membrane-tethered form of DLK1 (M-Dlk). However, our results with DLK1 do not a priori indicate that DLK1 acts solely as an inhibitor at the membrane. The ability of ligands to activate the Notch receptors is affected by glycosylation  and it is possible that the glycosylation state of the Drosophila Notch is incompatible with DLK1 activation. It is also possible that species-specific differences in endocytosis, in interactions with adaptor proteins, or in ligand specificities might also influence the outcome in our experiments. Furthermore, mammalian cells have 4 Notch receptors, and while two of these (Notch 1 and Notch 2) contain the same number of extracellular EGF repeats as Drosophila Notch, the others (Notch 3 and 4) have fewer and may have different interactions/ligand pairings. Thus the outcome of DLK1 actions could differ according to which specific Notch receptor or ligand-receptor pairing is affected. Nevertheless our results do show that DLK1 has the ability to regulate Notch signalling particularly when membrane tethered; hence it is likely to be an important factor acting on this pathway in mammalian development.
During mammalian embryogenesis, Dlk1 is expressed in many of the key lineages known to depend on Notch signalling for appropriate development, including somite derivatives . The relative expression of Dlk1 and the more typical Notch ligands may therefore play an important role in modulating the temporal and spatial control of Notch signalling in these cells. Once the expression and function of DLK1 in situ is better characterised, this may shed light on which of the mammalian receptors is likely to be a target of DLK1, what the functional relationships between DLK1 and other Notch ligands are, and whether DLK1 acts positively as well as antagonistically on Notch receptors.
Dlk1 constructs and transgenic flies
Three constructs were generated, representing the predominant forms of Dlk1 that are generated by alternate splicing and proteolysis. Genomic fragments were generated from BAC103N10 (Accession number AJ320506) and combined with appropriate fragments from a full length cDNA clone containing the Dlk1 open reading frame (IMAGE 604466) but lacking the protease cleavage domain. Open reading frames were ligated into pUAST for injection into yellow white flies. The first construct contained the largest open reading frame harbouring the intracellular and proteolytic cleavage domains (SM-Dlk), the second lacked the cleavage domain (M-Dlk), the third contained only the extracellular domain, equivalent to the secreted form of DLK1 after cleavage (S-Dlk). This was generated from the M form by truncating 5' of the proteolytic cleavage domain (via Sac I digestion), prior to cloning into pUAST. All constructs were sequenced to confirm accurate synthesis. Several independent lines were isolated and 3 lines of each were selected for further analysis, based on expression levels (antibody staining) and strength of phenotypes. All lines for a given construct gave qualitatively similar results. Homozygous stocks were prepared containing insertions on chromosomes II and III and these were combined for experiments with 2 copies of the transgenes.
Fly stocks and analysis of Adult phenotypes
Except where otherwise stated fly stocks used are as described . Adult flies were stored in 100% ethanol and selected parts were mounted in Euparol for photography.
Dissection of larvae and immunofluorescence were as described previously  except that dissections were carried out in PBS. Primary antibodies were rabbit anti-DLK1 ; mouse anti-Senseless ; mouse anti-Cut (1/20; ); mouse anti-Notch (C17.9C6 1/20; . The latter two antibodies were obtained from Developmental Studies Hybridoma Bank. Secondary antibodies were from Jackson Immnunological. Images were collected using a Leica TCS-NT-UV scanning confocal microscope.
We thank Simao Rocha and Isabel Gutteridge for comments on the manuscript. Work was supported by grants from the CR-UK to AFS and from the Wellcome Trust and Medical Research Council to SJB.
- 10.Baladron V, Ruiz-Hidalgo MJ, Nueda ML, Diaz-Guerra MJ, Garcia-Ramirez JJ, Bonvini E, Gubina E, Laborda J: dlk acts as a negative regulator of Notch1 activation through interactions with specific EGF-like repeats. Exp Cell Res. 2005, 303: 343-59. 10.1016/j.yexcr.2004.10.001.CrossRefPubMedGoogle Scholar
- 16.Geffers I, Serth K, Chapman G, Jaekel R, Schuster-Gossler K, Cordes R, Sparrow DB, Kremmer E, Dunwoodie SL, Klein T, et al: Divergent functions and distinct localization of the Notch ligands DLL1 and DLL3 in vivo. J Cell Biol. 2007, 178: 465-76. 10.1083/jcb.200702009.CrossRefPubMedCentralPubMedGoogle Scholar
- 20.DiazBenjumea FJ, Cohen SM: Serrate signals through Notch to establish a Wingless-dependent organizer at the dorsal/ventral compartment boundary of the Drosophila wing. Development. 1995, 121: 4215-4225.Google Scholar
- 39.Flybase. [http://flybase.bio.indiana.edu/]
- 41.Carlsson C, Tornehave D, Lindberg K, Galante P, Billestarp N, Michelsen B, Larsson LI, Neilsen JH: Growth hormone and prolactin stimulate the expression of rat preadipocyte factor-1/delta-like protein in pancreatic islets: molecular cloning and expression pattern during development and growth of the endocrine pancreas. Endocrinology. 1997, 138: 3940-3948. 10.1210/en.138.9.3940.PubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.