Genetic interactions between the DBL-1/BMP-like pathway and dpy body size-associated genes in Caenorhabditis elegans.

Bone morphogenetic protein (BMP) signaling pathways control many developmental and homeostatic processes, including cell size and extracellular matrix remodeling. An understanding of how this pathway itself is controlled remains incomplete. To identify novel regulators of BMP signaling, we performed a forward genetic screen in Caenorhabditis elegans for genes involved in body size regulation, a trait under the control of BMP member DBL-1. We isolated mutations that suppress the long phenotype of lon-2, a gene that encodes a negative regulator that sequesters DBL-1. This screen was effective because we isolated alleles of several core components of the DBL-1 pathway, demonstrating the efficacy of the screen. We found additional alleles of previously identified but uncloned body size genes. Our screen also identified widespread involvement of extracellular matrix proteins in DBL-1 regulation of body size. We characterized interactions between the DBL-1 pathway and extracellular matrix and other genes that affect body morphology. We discovered that loss of some of these genes affects the DBL-1 pathway, and we provide evidence that DBL-1 signaling affects many molecular and cellular processes associated with body size. We propose a model in which multiple body size factors are controlled by signaling through the DBL-1 pathway and by DBL-1-independent processes.

INTRODUCTION

The bone morphogenetic protein (BMP) family of signaling ligands, a group within the transforming growth factor-β (TGF-β) superfamily, are used by animal cells to convey developmental and homeostatic messages. Target cells receive the secreted BMP ligands and convert the cues into cell-specific transcriptional responses. How BMP pathways control and are controlled by cells is complex and incompletely clarified. In the roundworm Caenorhabditis elegans, the BMP member DBL-1 has a well-defined pathway that includes the core components and conserved regulators, such as the negative regulator LON-2/glypican. There are clear, dose-dependent phenotypes associated with this pathway, including body size. Loss of DBL-1 signaling leads to small body size, while increased signaling results in long animals (Brenner, 1974; Savage ; Krishna ; Morita ). Genetic screens and powerful molecular techniques available in C. elegans have been used to isolate, identify, and characterize DBL-1 and associated conserved BMP pathway players (Brenner, 1974; Savage ; Krishna ; Suzuki ; Savage-Dunn ; Gumienny , 2010; Liang ; Gleason ).

Previous screens for body size mutants have been done by mutagenizing otherwise wild-type animals. To increase the likelihood of identifying new modulators of the pathway, we performed a forward genetic screen using a sensitized background, lon-2(e678). Because this genotype increases DBL-1 pathway signaling and causes a long-body phenotype, we expected that we could isolate alleles that cause mild changes in body length that may not have been identified in the wild-type background used in previous screens (Brenner, 1974; Savage-Dunn ). Suppressors were expected either to act downstream of LON-2 in the DBL-1 signaling pathway or to render animals smaller or shorter independent of DBL-1 signaling. In this screen, we identified additional genetic interactions between the BMP-like DBL-1 pathway and cellular processes and components affecting body size. Here, we analyze genetic interactions between the DBL-1 pathway and other processes that affect body size.

RESULTS AND DISCUSSION

Isolation of alleles of known DBL-1 pathway signaling components

LON-2 is a conserved glypican, a heparan sulfate proteoglycan that inhibits DBL-1 from activating its receptors (Gumienny ). Loss of LON-2 results in increased DBL-1 pathway signaling and a long-body phenotype (Brenner, 1974). We performed an EMS mutagenesis F2 screen for animals with recessive mutations that reduce the body length of lon-2(e678) animals. We isolated and characterized 46 alleles from a screen of ∼9000 mutagenized genomes, which is less than saturation (Table 1). As expected, multiple alleles of core DBL-1 signaling pathway components were isolated (Table 1A). Two alleles of the dbl-1 gene encoding the ligand, wk91 and wk92, were confirmed by complementation testing with dbl-1(wk70). Four alleles of the gene encoding the sma-6 Type I receptor, wk87, wk103, wk129, and wk134, were confirmed by genetic testing with sma-6(e1482). Two alleles of the R-Smad sma-2 gene, wk83 and wk84, were identified by noncomplementation with sma-2(e502). The co-Smads sma-3 and sma-4 are represented in this screen by sma-3(wk107) and sma-4(wk85), confirmed by complementation testing with sma-3(e491) and sma-4(e729), respectively. Two alleles of the Smad regulator Schnurri homologue sma-9, wk97 and wk131, were shown to be allelic to sma-9(wk55). Three alleles were isolated for sma-10, a pathway regulator that interacts physically with the DBL-1 receptor complex and promotes DBL-1 pathway signaling. wk88, wk89, and wk90 did not complement sma-10(wk66) (Gumienny ). sma-11/kin-29 is a serine–threonine kinase gene that acts genetically between dbl-1 and lon-1, a DBL-1 pathway target gene (Maduzia ). wk117 fails to complement kin-29(wk26). Because we selected against alleles that were dauer constitutive, we did not isolate daf-4 Type I receptor mutations. These results show that this screening approach was successful for isolating DBL-1 pathway–associated alleles.

TABLE 1:

Alleles isolated in an EMS screen that suppress the lon-2(e678) long body size.

Gene	Known allele	Function or homology (Drosophila)	Allele isolated
A. Alleles of DBL-1 pathway genes
dbl-1	wk70	TGF-β-like ligand	wk91, wk92
sma-6	e1482	Type I receptor	wk87, wk103, wk129, wk134
sma-2	e502	Smad	wk83, wk84
sma-3	e491	Smad	wk107
sma-4	e729	Smad	wk85
sma-9	wk55	Transcription regulator (Schnurri)	wk97, wk131
sma-10	wk66	LRIG (binds receptors to facilitate signaling)	wk88, wk89, wk90
kin-29/sma-11	wk26	Kinase	wk117
B. Alleles of uncloned sma genes
sma-12	wk69	Positive body size effector	wk119
sma-13	wk52	Positive body size effector	wk109
—	-	Positive body size effector	wk94
—	-	Positive body size effector	wk98
—	-	Positive body size effector	wk123
C. Alleles of ECM-associated genes
dpy-1	e1	Mucin-like	wk111, wk142, wk143
dpy-2	e8	Collagen (Viking)	wk104, wk122
dpy-3	e27	Collagen (Col4a1)	wk105
dpy-4	e1166	Collagen (Col4a1)	wk101, wk102
dpy-5	e61	Collagen (Col4a1)	wk137
dpy-6	e14	Mucin-like	wk113, wk118, wk127, wk132
dpy-7	e88	SFTPD (collagen-containing C-type lectin) (Col4a1)	wk106
dpy-8	e130	Collagen (Viking)	wk120
dpy-9	e12	Collagen (Col4a1)	wk124
dpy-13	e184	Collagen (Viking)	wk95, wk135, wk145
dpy-14	e188	Collagen (Viking)	wk116, wk128
dpy-17	e164	Collagen (Viking)	wk140
D. Alleles of other body size-associated genes
dyf-11	wk57	Intraflagellar transport particle B component (IFT54)	wk136, wk139
dpy-20	e1282	BED zinc finger protein	wk141

Isolation and mapping of alleles of unknown loci

This screen also provided additional alleles of uncloned genes (Table 1B). wk119 is allelic to sma-12(wk69) V, and wk109 is allelic to sma-13(wk52) I (Savage-Dunn ). Three loci remain unnamed. Another allele, wk94, was mapped to chromosome II and complements sma-6, dpy-2, and dpy-10. It is slow-growing. wk123 also maps to Chromosome II and complements sma-6, dpy-2, dpy-10, and wk94. wk98 maps to Chromosome III and is not allelic to daf-4, sma-2, sma-3, sma-4, or sma-17(wk59). wk98 males have a normal male tail and mate well, unlike most alleles of DBL-1 core-pathway genes.

Isolation of alleles of extracellular matrix–associated genes

Many gene products contribute to the C. elegans cuticle, a specialized extracellular matrix (ECM) that protects animals. Loss of some cuticle-associated genes produces a reduced body size (Page and Johnstone, 2007). Alleles of several collagen genes were isolated in the lon-2 suppressor screen (Table 1C). These loci were identified by their failure to complement published dpy alleles (see Table 1 for alleles tested). wk116 and wk128 are alleles of collagen type III alpha 1-like gene dpy-14. Alleles of genes encoding proteins with collagen domains were isolated: dpy-2 (wk104 and wk122), dpy-3/dpy-12 (wk105), dpy-4 (wk101 and wk102), dpy-5 (wk137), dpy-7 (wk106), dpy-8 (wk120), dpy-9/col-100 (wk124), dpy-13 (wk95, wk135, and wk145), and dpy-17 (wk140).

The screen also produced alleles of two mucin-like genes, dpy-1 (wk111, wk142, and wk142) and dpy-6 (wk113, wk118, wk127, and wk132; Table 1C; S. Eimer, personal communication, July 18, 2019; Shen ). Mucins are glycosylphosphatidylinositol (GPI)-anchored cell surface glycoproteins that are components of the C. elegans cuticle’s surface coat. In other systems, mucins are transcriptionally regulated by the TGF-β pathway (Jonckheere ,b). Human mucin 1 is required for TGF-β expression, secretion, and activity (Li ,b; Grover ).

Isolation of alleles of other body size–associated genes

Loss of some intraflagellar transport (IFT) components is associated with a reduced body size (Starich ; Kobayashi ). We discovered that wk136 and wk139 are allelic to dyf-11(mn392), which produces “slightly short” animals and affects a protein related to TRAF3IP1 (TRAF3 interacting protein 1)/MIP-T3 (microtubule-interacting protein that interacts with TRAF3, the tumor necrosis factor receptor–associated factor 3). MIP-T3 is associated with intraflagellar transport particle B, a complex that transports cargo to the tip of flagella and cilia (Starich ; Bacaj ). wk139 and wk136 are also allelic to previously uncloned sma-16(wk57), which has a similar mild small body size phenotype (Table 1D; Savage-Dunn ). This result suggests that the small body size associated with loss of IFT suppresses DBL-1 pathway-mediated body size. Future work will determine whether the mechanism of this body size suppression is independent of DBL-1 signaling.

dpy-20 encodes a novel BED zinc-finger protein required for normal body morphology, movement, male mating (but not ray morphology), and germ cell proliferation (Hodgkin, 1983; Baird and Emmons, 1990; Clark ; Schertel and Conradt, 2007; Yemini ). A role as a transcription regulator, a common (but not exclusive) role for other BED zinc-finger-containing proteins, has been suggested for DPY-20 (Aravind, 2000). Transcriptional regulation of cuticle genes by DPY-20 is supported by observed changes in cuticle structure in animals with absent or reduced levels of DPY-20 (Clark ). An allele of dpy-20, wk141, was isolated in the lon-2 suppressor screen (Table 1D). In addition to the dumpy body shape, wk141 animals also displayed a round head characteristic of dpy-20 (unpublished data; Clark ).

Loss of dpy gene function reduces body size of animals overexpressing DBL-1

LON-2 is a glypican that normally sequesters DBL-1 ligand and prevents DBL-1 from activating receptors that mediate body size gene expression (Gumienny ). Loss of LON-2 results in long animals (Brenner, 1974). Overexpressing DBL-1 also results in long animals (Morita ; Schultz ). To generate a more comprehensive understanding of the interaction of DBL-1 signaling with other factors that have strong effects on body size, we asked whether loss of the function of a panel of genes affecting body size in C. elegans can suppress the long body size defect exhibited by animals overexpressing DBL-1. This panel includes all dpy genes identified in the screen and dpy-10, dpy-11, sqt-3/dpy-15, dpy-18, dpy-19, dpy-21, dpy-23, and blmp-1/dpy-24, for which alleles were not isolated. Regardless of gene function, most dpy(RNAi)s and all dpy mutants tested in green fluorescent protein (GFP)-tagged DBL-1 overexpressing backgrounds are significantly shorter than the controls (Supplemental Table S2). Thus, all dpy mutants tested suppress the long phenotype associated with overexpressed GFP-tagged DBL-1.

Some extracellular matrix–associated genes interact genetically with the DBL-1 pathway

In C. elegans, signaling by the DBL-1 pathway controls components that form the cuticle, a specialized ECM that not only provides a barrier between the animal and its environment, but also affects the size of the organism within it (Mochii ; Liang ; Luo ; Roberts ; Madaan ). Possibly because the C. elegans cuticle is dynamic, with different constituents composing the cuticle of each of its developmental stages, different cuticle component genes were identified in these studies. Altering the cuticle content by reduced DBL-1 signaling may underlie dbl-1 loss-of-function phenotypes besides smaller body length, including male tail ray fusions, increased cuticular permeability, and “worm-star” formation by clusters of nematodes becoming entangled by their tails (Savage ; Morita ; Suzuki ; Schultz ). However, loss of single cuticular proteins can also alter nematode body length, possibly independent of the DBL-1 pathway (Brenner, 1974; Nystrom ; Page and Johnstone, 2007; Fernando ).

We tested for interaction of selected genes with the DBL-1 pathway in two ways. First, we used a functional GFP-tagged DBL-1 (GFP::DBL-1) translational reporter (dbl-1p::gfp::dbl-1) to determine whether dpy gene products affect the DBL-1 ligand. Second, we used a transcriptional reporter for DBL-1 pathway activity, GFP expressed from the spp-9 promoter (spp-9p::gfp). spp-9 is a negatively regulated transcriptional target of the DBL-1 pathway that encodes a saposin-like protein, an innate immune response factor (Roberts ).

We tested these two reporters against dpy genes that encode structural components of the extracellular matrix. RNA interference (RNAi) of these genes in the GFP::DBL-1 background suggested that many affect DBL-1 ligand (Supplemental Table S3). We crossed these reporter transgenes into the panel of dpy mutants and examined their effect on reporter expression in 24-h adult hermaphrodites. Interestingly, all tested ECM mutants that significantly reduced GFP::DBL-1 fluorescence levels also affected spp-9p::gfp expression (Table 2B). This result confirms spp-9p::gfp reporter activity is strongly linked to the DBL-1 pathway.

TABLE 2:

Effects of body size–associated genes on GFP::DBL-1 and DBL-1 pathway reporter spp-9p::GFP fluorescence.

Gene	Genotype	dbl-1::gfp ­fluorescence % control ± 95% CI	p value	Genotype	spp-9p::GFP fluorescence % control ± 95% CI	p value
A. Controls
Control	texIs100	100 ± 15.57		texIs127	100 ± 11.47
Control	texIs101	100 ± 25.95		wkIs40	100 ± 13.20
B. ECM-associated genes
dpy-1	dpy-1(e1); texIs100	96.02 ± 40.89	0.8335	dpy-1(e1); texIs127	100.80 ± 16.97	0.9234
dpy-2	dpy-2(e8); texIs100	115 ± 52.15	0.5080	dpy-2(e8); texIs127	107.04 ± 12.20	0.2344
dpy-3	texIs100; dpy-3(e27)	43.45 ± 21.63	0.0012	wkIs40; dpy-3(e27)	156.26 ± 16.06	0.0001
dpy-4	dpy-4(e1166); texIs101	102.24 ± 28.06	0.9043	dpy-4(e1166); texIs127	125 ± 25.16	0.0192
dpy-5	dpy-5(e61); texIs100	54.13 ± 26.47	0.0127	dpy-5(e61); texIs127	155.23 ± 25.16	0.0001
dpy-6	texIs101; dpy-6(e14)	59.96 ± 17.12	0.0009	wkIs40; dpy-6(e14)	179.04 ± 19.77	0.0001
dpy-7	texIs100; dpy-7(e88)	79.88 ± 34.85	0.2689	wkIs40; dpy-7(e88)	104.89 ± 12.95	0.4299
dpy-8	texIs100; dpy-8(e130)	79.64 ± 26.74	0.1828	wkIs40; dpy-8(e130)	92.68 ± 11.12	0.2061
dpy-9	dpy-9(e12); texIs101	95.02 ± 29.01	0.7248	dpy-9(e12); texIs127	100.29 ± 10.24	0.9533
dpy-10	dpy-10(e128); texIs100	61.46 ± 2 4.99	0.0172	dpy-10(e128); texIs127	115.50 ± 11.56	0.0054
dpy-11	texIs100; dpy-11(e224)	92.10 ± 28.95	0.6094	dpy-11(e224); texIs127	128.88 ± 18.06	0.0007
dpy-13	dpy-13(e184); texIs101	91.98 ± 30.42	0.5829	dpy-13(e184); texIs127	156.65 ± 24.39	0.0001
dpy-14	dpy-14(e188); texIs100	64.97 ± 2 6.13	0.0289	dpy-14(e188); texIs127	142.81 ± 22.04	0.0001
dpy-15	texIs100; dpy-15(e24)	99.14 ± 43.40	0.9654	dpy-15(e24); texIs127	86.33 ± 19.32	0.1875
dpy-17	dpy-17(e164); texIs100	80.99 ± 29.16	0.1969	dpy-17(e164); texIs127	121.07 ± 17.84	0.0116
dpy-18	dpy-18(e364); texIs100	96.43 ± 36.45	0.8543	dpy-18(e364); texIs127	110.60 ± 15.13	0.1336
C. Other body size–associated genes
dpy-19	dpy-19(e1259); texIs100	109.22 ± 47.09	0.6514	dpy-19(e1259); texIs127	132.90 ± 29.35	0.0107
dpy-20	dpy-20(e1282); texIs101	91.56 ± 54.19	0.7181	dpy-20(e1282); texIs127	109.77 ± 15.57	0.1818
dpy-21	texIs100; dpy-21(e428)	110.37 ± 44.89	0.6011	dpy-21(e428); texIs127	116.86 ± 11.93	0.0033
dpy-23	texIs100;dpy-23(e840)	39.87 ± 14.82	0.0005	wkIs40;dpy-23(e840)	120.47 ± 13.55	0.0015
dpy-24	dpy-24(s71); texIs100	59.89 ± 22.45	0.0182	dpy-24(s71); texIs127	135.42 ± 19.64	0.0001

Some extracellular matrix–associated genes had no effect on either GFP::DBL-1 or spp-9p::gfp expression: dpy-1/mucin and dpy-2, dpy-7, dpy-8, dpy-9, and sqt-3/dpy-15 (Figure 1 and Table 2B). These results are consistent with these gene products having no effect on DBL-1 signaling. dpy-3, dpy-5, dpy-6, dpy-10, and dpy-14 decreased GFP::DBL-1 fluorescence and increased spp-9p::gfp expression, consistent with loss of these gene products having a negative impact on DBL-1 signaling, suggesting feedback. Interestingly, other specific collagens that are DBL-1–regulated also show this feedback on DBL-1 signaling (Madaan ).

FIGURE 1:

Effects of ECM-associated gene mutations on GFP::DBL-1 and DBL-1 pathway reporter spp-9p::GFP fluorescence. Arrows point to GFP::DBL-1 fluorescent punctae in A–R. Representative images show adult hermaphrodite expression of GFP::DBL-1 (A–R) and spp- 9p::GFP (A′–R′) in the following mutant backgrounds: A and B, wild type; C, dpy-1; D, dpy-2; E, dpy-3; F, dpy-4; G, dpy-5; H, dpy-6; I, dpy-7; J, dpy-8; K, dpy-9; L, dpy-10; M, dpy-11; N, dpy-13; O, dpy-14; P, sqt-3/dpy-15; Q, dpy-17; R, dpy-18; dpy mutants that reduced GFP-tagged DBL-1 and increased spp-9p::GFP activity are shown in E, G, H, L, and O. dpy mutants that increased spp-9p::GFP reporter activity without affecting GFP-tagged DBL-1 are shown in F, M, N, and Q. dpy mutants that had no effect on either of the reporters are shown in C, D, I, J, K, P, and R. Bars in A and B represent 10 and 100 μm, respectively.

Notably, we discovered that dpy-4, dpy-13, and dpy-17 significantly up-regulate spp-9p::gfp expression without affecting GFP::DBL-1 fluorescence (Figure 1; Table 2B). C. elegans DPY-17 is predicted to act downstream of DBL-1 secretion (Fotopoulos ). In other systems, basement membrane collagens bind BMPs and enhance their interaction with receptors, promoting downstream signaling (Paralkar , 1992; Wang ). However, DPY-4, DPY-13, and DPY-17 are predicted cuticle collagens and therefore probably do not act between the secreting and receiving cells (see Figure 3 later in the paper). It is possible that these collagens indirectly regulate DBL-1 signaling targets by affecting the cuticle. Also, while loss of some collagen genes, including dpy-2 and dpy-7, induced a robust stress response in C. elegans, loss of dpy-4 and dpy-13 failed to do so (Dodd ). The significant increase of spp-9p::gfp expression in dpy-4 and dpy-13, but not in dpy-2 or dpy-7, also supports the conclusion that spp-9p::gfp is strongly responsive to DBL-1 signaling and not to stress by cuticle alteration. Nonetheless, we show specificity of collagen regulation of DBL-1 signaling, because some known cuticle collagens have no effect on DBL-1 signaling in one-day adults (DPY-2, DPY-7, DPY-9, and SQT-3/DPY-15). It should be noted, though, that some collagens are known to be expressed at certain stages in C. elegans, so some ECM-associated dpy genes may affect DBL-1 signaling at other stages, which we would not have identified in this study (Celniker ; Jackson ; Madaan ).

FIGURE 3:

Proposed model of body size regulation by the DBL-1 pathway and dpy body size–associated genes in C. elegans. Within the sending cell, DBL-1 may be modified by DPY-19/C-mannosyltransferase and transported by the AP2 complex, which includes the DPY-23/AP2 mu subunit. Signaling through the receptor complex may also be controlled by DPY-23. DBL-1 signaling directly controls the expression of cuticle components, which affects body size (arrow 1). DPY-11/disulfide oxidoreductase may process DBL-1-regulated cuticle components before secretion. DBL-1–regulated gene expression, in part through BLMP-1, may modulate signaling within the receiving cell (arrow 2) or between the sending and receiving cells (arrow 3), or may feed back on the sending cell (arrow 4). Cues from the receiving cell or cuticle may be received (indirectly) by the sending cell and affect DBL-1 expression or secretion (arrow 5) or affect signaling downstream of DBL-1 secretion (starred arrowhead in arrow 5 and starred DPY collagens). Other cell-signaling pathways act independent of DBL-1 to control body size (arrow 6). Dashed lines represent potential indirect regulation. dpy gene products that had no effect on DBL-1 signaling are not included in this model.

In addition, we tested these two reporters against two dpy genes that encode ECM-processing enzymes. DPY-18 is a prolyl-4-hydroxylase alpha catalytic subunit, which in other organisms is required for collagen secretion as well as triple-helix assembly and stability (Mazzorana ; Brodsky and Ramshaw, 1997; Walmsley ; Hill ). dpy-11 encodes a hypodermally expressed disulfide oxidoreductase that is predicted to modify extracellular matrix constituents (Nishiwaki and Miwa, 1998; Ko and Chow, 2002). Based on their phenotypes, both of these gene products may affect both cuticle and basement membrane (Nishiwaki and Miwa, 1998; Hill ; Ko and Chow, 2002; Torpe and Pocock, 2014). dpy-18 loss had no effect on either DBL-1 pathway reporter, but dpy-11 loss affected spp-9p::GFP without affecting the ligand within the secreting cell (Figure 1; Table 2B).

Together, these results suggest that the DBL-1 pathway and spp-9 target gene expression are sensitive to cuticle-related factors that indirectly impact DBL-1 signaling.

Other body size genes interact genetically with the DBL-1 pathway

dpy-19 encodes a C-mannosyltransferase (Buettner ). dpy-19(e1259) resulted in significantly higher spp-9p::gfp reporter activity but did not alter GFP-tagged DBL-1 fluorescence levels (Figure 2; Table 2C). This enzyme family recognizes W-x-x-W motifs and adds α-mannose to the tryptophans in these motifs. C. elegans DBL-1 has a conserved W-x-x-W motif and may be a direct target of DPY-19. However, ADAMTS protease ADT-2/SMA-21 regulates DBL-1 signaling positively, and ADAMTS proteases are also targets of this enzyme family (Fernando ; Ihara ). ADT-2/SMA-21 isoforms have up to seven W-x-x-W or W-x-x-W-x-x-W motifs (Fernando ). Therefore, DPY-19 may regulate DBL-1 signaling directly through DBL-1′s W-x-x-W motif, which may not affect DBL-1 localization within secreting cells but may affect DBL-1′s ability to activate its receptors. Alternatively, DPY-19 may regulate DBL-1 signaling indirectly through modification of an ADAMTS protease.

FIGURE 2:

Effects of other body size–associated gene mutations on GFP::DBL-1 and DBL-1 pathway reporter spp-9p::GFP fluorescence. Arrows point to GFP::DBL-1 fluorescent punctae in A–E. Representative images show adult hermaphrodite expression of GFP::DBL-1 (A–E) and spp-9p::GFP (A′–E′) in the following mutant backgrounds: A, dpy-19; B, dpy-20; C, dpy-21; D, dpy-23; E, dpy-24. Compare to controls in Figure 1. dpy mutants that reduced GFP-tagged DBL-1 and increased spp-9p::GFP activity are shown in D and E. dpy mutants that increased spp-9p::GFP reporter activity without affecting GFP-tagged DBL-1 are shown in A and C. A dpy mutant that had no effect on either of the reporters is shown in B. Bars in A and B represent 10 and 100 μm, respectively.

Some sex and dosage compensation mutants are also shorter, perhaps from overexpression of X-linked body size genes (Meyer, 2005). We tested one of these, dpy-21, a dosage compensation–complex subunit (Yonker and Meyer, 2003). Loss of dpy-21 increased DBL-1 pathway reporter expression but did not affect the GFP-tagged DBL-1 ligand (Figure 2; Table 2C).

dpy-23/apm-2 encodes the adaptor protein complex 2 (AP2) mu subunit. dpy-23(e840), a 100-kb deletion that removes 18 open reading frames, significantly reduces GFP-tagged DBL-1 fluorescence and increases reporter activity (Figure 2; Table 2C; Gu ). DPY-23 is involved in protein trafficking, including the trafficking of the DBL-1 receptor SMA-6, which affects DBL-1 signaling (Gleason ). Feedback caused by loss of the DBL-1 receptor SMA-6 at the membrane could decrease ligand expression. However, DPY-23 is also expressed in the same neurons as DBL-1 and may play a direct role in the secretion of DBL-1 (Shim and Lee, 2000). While all phenotypes associated with e840 deletion are fully rescued by expression of wild-type dpy-23, it is possible that the DBL-1–associated phenotypes identified here are caused by a different part of the e840 deletion (Gu ).

Two putative transcription factor genes also were tested. Loss of the BED zinc-finger protein gene dpy-20 had no effect on GFP-tagged DBL-1 fluorescence or spp-9 transcriptional reporter expression (Figure 2; Table 2C). However, loss of the SET domain and the zinc-finger protein gene blmp-1/dpy-24 significantly reduced GFP-tagged DBL-1 fluorescence and increased spp-9 transcriptional reporter expression (Figure 2; Table 2C). This result supports a genetic interaction between BLMP-1 and the DBL-1 pathway. Another C. elegans superfamily member related to DBL-1, daf-7, requires blmp-1 for proper signaling (Hyun ). In mammals, BLIMP1 is highly regulated by the TGF-β pathway (Salehi ; Telerman ).

Regulation of body size–associated genes by DBL-1

TGF-β pathways regulate ECM genes in many systems, both directly and indirectly (Morikawa ; Kim ; Madaan ). In C. elegans, microarray analyses showed that some ECM genes are highly regulated by this pathway (Mochii ; Liang ; Roberts ). Furthermore, ChIP-seq analyses showed that the DBL-1 pathway signal transducer SMA-3 binds the regulatory region of the cuticle collagen genes col-141 and col-142 (Madaan ). To better understand the interplay between the DBL-1 pathway and other body size players, we performed RNA-seq analyses on wild-type and dbl-1(nk3) strains. Like previous microarray results, we identified several cuticle-associated genes that were down-­regulated in the dbl-1(nk3) strain as compared with the wild type (Supplemental Table S4). However, our panel of regulated genes is overlapping but more extensive. In addition, we identified SMA-3 binding sites upstream of body size–associated genes by interrogating the SMA-3 MODENCODE data set (Gerstein ). We also determined that blmp-1 is highly regulated by DBL-1. Notably, SMA-3 binds within 1000 bp upstream of blmp-1, suggesting that C. elegans BLMP-1 is transcriptionally regulated by DBL-1/BMP, as seen by TGF-β regulation of BLIMP1 in other systems (Gerstein ; Salehi ; Telerman ). BLMP-1 binds upstream of most of the ECM-associated genes that are differentially regulated by DBL-1, suggesting that DBL-1 signaling regulates these ECM genes through transcriptional control of BLMP-1 (Supplemental Table S4) (Gerstein ). It remains to be experimentally determined whether this correlation is physiologically relevant.

Concluding remarks

In this work, we have presented the analysis of a lon-2(-) suppressor screen and a related panel of body size–associated dpy genes. We have identified complex interactions between DBL-1 signaling and other body size regulators, which have previously been presumed to act independent of DBL-1 signaling. This work supports the model that the DBL-1 pathway is a master regulator of several processes that together affect the body size trait, and some body size–associated genes also feed back on the DBL-1 pathway (Figure 3; Mochii ; Roberts ; Liang ; Madaan , 2019). Some DBL-1 pathway suppressors and other body size–associated genes may be regulated directly by the DBL-1 pathway Smads. However, expression of other body size–associated genes may be controlled by DBL-1–regulated transcription factors such as BLMP-1/DPY-24. We are poised to define the molecular and cellular relevance of the genetic interactions between DBL-1 signaling and other body size regulators, including the uncloned loci, DPY-6/mucin-like protein, intraflagellar transport proteins, DPY-19/C-mannosyltransferase, and BLMP-1/DPY-24/transcription regulator.

MATERIALS AND METHODS

Strains and maintenance

C. elegans strains used in these studies were derived from the wild-type variety Bristol strain N2 and were cultured on modified EZ nematode growth media plates as previously described (Madhu ). All strains were cultured on Escherichia coli strain OP50 at 20°C. Strains generated and used are listed in Supplemental Table S1. Double mutant lines were generated using standard genetic crosses.

The transgenic alleles texIs100 [dbl-1::dbl-1:gfp; ttx-3p::rfp] IV and texIs101 [dbl-1::dbl-1:gfp; ttx-3p::rfp] V were created by UV/TMP mutagenesis of an extrachromosomal array (Beifuss and Gumienny, 2012). wkEx52 (spp-9p::gfp) was integrated using an established UV/TMP mutagenesis procedure to yield texIs127 X (Yandell ). These integrated transgenes were backcrossed five times and mapped. wkEx52(spp-9p::gfp) was integrated into N2 animals using an x-ray source to yield wkIs40. This strain was backcrossed four times to the N2 wild-type strain. The resulting strain was named LT998.

Double mutants with texIs100, texIs101, texIs127, and wkIs40, and the triple mutant strain TLG281 rrf-3(pk1426) II; texIs100 IV; dbl-1(nk3) V were constructed by standard methods (Supplemental Table S1) (Brenner, 1974). A complete list of strains used in this work is provided in Table 1 and Supplemental Table S1.

Forward genetic screen for suppressors of the lon-2(e678) long phenotype

To identify factors involved in regulating DBL-1 pathway-mediated body length, lon-2(e678) hermaphrodites were mutagenized with 50 mM ethyl methanesulfonate (EMS) as previously described (Brenner, 1974; Savage-Dunn ). Mutagenized animals were transferred to plates to self-propagate. F1 animals were transferred two to a plate to self. F2 populations were scored for a reduced body length in about one-eighth of the population, assuming a recessive mutation. Only one animal from each plate containing a candidate mutant was maintained, to confirm that each candidate was independently obtained. About 9000 mutagenized genomes were scored (less than saturation), and lines with high body size–phenotype penetrance and expressivity, and with good health, were selected for characterization. Animals that showed other phenotypes, including rolling movement (Rol) and reduced viability or sterility, were excluded. Mutants were outcrossed five times to N2 before further analyses. Outcrossing also identified whether alleles resided on the X chromosome or an autosome. Mutations were complementation-tested for allelism with known DBL-1 pathway genes. Alleles of novel loci were then tested for linkage to STS markers or to unc-15(e73) I, unc-104(e1265) II, unc-32(e189) III, unc-5(e53) IV, and unc-23(e25) V (Williams ). Next, mutations mapping to the same linkage group were complementation-tested to each other and candidate gene mutant alleles. Some alleles were further mapped using standard two-factor, three-factor, and SNP mapping techniques (Brenner, 1974; Wicks ).

RNA interference

RNAi by bacterial feeding was performed as previously described (Timmons and Fire, 1998). The C. elegans strain TLG281 rrf-3(pk1426) II; texIs100 IV; dbl-1(nk3) V was used for this experiment (Beifuss and Gumienny, 2012). Bacteria from the Vidal and Ahringer cDNA libraries (Open Biosystems and Source BioScience, respectively) were used (Kamath and Ahringer, 2003; Rual ). C06C3.5, a predicted pseudogene, was used as the negative control. bli-4(RNAi) was used as the positive control (Thacker ). Bacteria from single colonies were grown overnight in Luria–Bertani broth (Sigma, St. Louis, MO) containing 50 µg/ml carbenicillin (Gold Biotechnology, St. Louis, MO) and induced to express dsRNA using 1 µg/ml isopropyl β-d-1-thiogalactopyranoside (IPTG) for 4 h. After induction, bacterial broth was plated on nematode growth medium containing 50 µg/ml carbenicillin and 1 µg/ml IPTG and dried. Embryos staged by bleaching were transferred to these plates and grown to the young adult stage for observation as previously described (Kamath and Ahringer, 2003; Rual ; Beifuss and Gumienny, 2012).

Imaging

C. elegans adults 24 h post-L4 were anesthetized using 1 mM levamisole hydrochloride (Sigma, St. Louis, MO) for fluorescence and body size imaging.

Fluorescence imaging

Fluorescence imaging for RNAi and mutant studies was performed using a Nikon A1 confocal system (Nikon Instruments, Melville, NY). A 60× objective was used to capture GFP::DBL-1 intensities. Fluorescence intensities of spp-9p::gfp were captured using a 10× objective. Microscope settings were standardized to obtain fluorescence intensity values in the measurable dynamic range for both control and experimental samples, and these image-capture conditions were kept constant throughout the experiments. Mean fluorescence intensities were measured using Nikon NIS Elements AR-5.02 software.

Body size imaging

For RNAi studies, animals were imaged using the same confocal system with a 10× objective to measure their body length. Body size was quantified using NIS Elements AR-5.02 imaging software.

For mutant studies, animals were imaged to measure body size using a Nikon SMZ1500 dissecting microscope (Nikon, Melville, NY). Body size was measured using the length measurement image tool within iVision-Mac (BioVision Technologies, Exton, PA).

RNA-seq

We used two biological replicates of wild-type and dbl-1(nk3) populations for RNA sequencing with the Illumina MiSeq system. Animals were staged by bleaching and harvested for RNA extraction at the 48 h post-L4 stage. RNA extraction was performed using the freeze-cracking method as previously described (Portman, 2006). A total of 2 µg of total RNA of each sample was used for TruSeq Stranded mRNA sample preparation using the low sample protocol per manufacturer’s instructions (Illumina). RNA and cDNA concentrations were measured by a Qubit fluorometer (Molecular Probes, Life Technologies). The quality of RNA and cDNA was assessed by an Experion automated electrophoresis system (BioRad). The libraries were sequenced with ∼40–50 million total reads and FASTQ quality scores of the raw data from sequencing were generated. The reads were mapped against the C. elegans genome ce10 (UCSC RefSeq gene annotation) using the STAR aligner (RNA-Seq Alignment, Illumina). On the average, 97.9% of reads were aligned. Cufflinks 2 was used for the FPKM estimation of reference genes and transcripts. Differential gene expression analysis was performed with Cuffdiff2 using the Cufflinks Assembly and DE application (Illumina). Candidates with a q value <0.05 were classified as differentially expressed genes.

Statistical analysis

We used at least seven animals to quantify GFP::DBL-1 fluorescence and at least 10 animals to measure body size for RNAi studies. For the mutant studies, we used at least 10 animals to quantify GFP::DBL-1 fluorescence intensities and at least 15 animals to measure spp-9p::gfp fluorescence intensities. We used at least 30 animals to determine body size for the mutant studies. Mean fluorescence intensities and body length values of test populations were calculated as percentages of their respective controls with 95% confidence intervals. Statistical analyses were performed using an independent t test to compare test and control sample means.

Data and reagent availability

Strains listed in Table 1 and Supplemental Table S1 are available upon request.

Click here for additional data file.