Figure 1. Complete Lefty loss causes severe patterning defects.

(A) Nodal activates itself, mesendodermal genes, and the secreted feedback inhibitor Lefty by inducing phosphorylation and nuclear translocation of the signal transducer Smad2. (B) Nodal (blue) and Lefty (brown) are expressed in overlapping domains at the margin (black line) in zebrafish embryos, generating a signaling gradient of phosphorylated Smad2 (red). (C) A 13-base-pair deletion at the 5’ end of lft1^a145 removes the translational start site (TSS) and part of the signal sequence. An alternative TSS 36 bp downstream of the deletion could produce an in-frame protein product, but the first 16 amino acids, and therefore most of the predicted 20-aa signal sequence, would be missing. (D) An 11-base-pair deletion at the 5’ end of lft2^a146 removes part of the predicted 19-aa signal sequence, causing a frame shift after 36 bp resulting in a stop codon 18 bp later. (E–J’’) Testing the activity of lft mutant mRNA. All images were obtained at 24 hr post-fertilization (hpf). Wild type embryos at the one-cell stage were injected with 1, 10, and 100 pg wild type lft1 (G–G’’) or lft1^a145 (H–H’’) mRNA as indicated. Embryos expressing lft1 mRNA exhibit Nodal loss-of-function phenotypes, similar to maternal-zygotic mutants for the zebrafish EGF-CFC co-receptor one-eyed pinhead (oep) (E), which are insensitive to Nodal signaling (Gritsman et al., 1999). Embryos expressing mutant mRNA do not exhibit Nodal loss-of-function phenotypes. Similar results were obtained for wild type lft2 (I–I’’) and lft2^a146 (J–J’’) mRNA. (F) Uninjected wild type embryo. (K–S) lft mutant phenotypes. All images were obtained at 24–29 hpf. A single functional lft allele is sufficient for grossly normal patterning (Q,R). (S) lft1^-/-;lft2^-/- double homozygous mutants have severe patterning defects and lack eyes, heart, and full length tails, and often exhibit excess tissue along the posterior trunk. (T) Percentage of embryos with normal gross morphology at 1 day post-fertilization (dpf). Number of normal/total embryos: wild type = 50/50, lft1^+/- = 49/50, lft2^+/- = 46/46, lft1^-/- = 50/50, lft2^-/- = 46/46, lft1^+/-;lft2^+/- = 50/50, lft1^+/-;lft2^-/- = 43/43, lft1^-/-;lft2^+/- = 24/24 lft1^-/-;lft2^-/- = 0/55.

Figure 1—figure supplement 1. lefty1^-/- mutants exhibit partially penetrant heart laterality defects.

(A) In situ hybridization using a probe against the heart muscle marker myl7 in 24 hr post-fertilization (hpf) wild type and lft mutant embryos (dorsal views). Most wild type and lft2^-/- mutants exhibit myl7 expression on the left side. In contrast, some lft1^-/- mutants exhibit bilateral myl7 expression, and lft1^-/-;lft2^-/- mutants express very little myl7 and typically fail to generate hearts (dorsal views). (B) Quantification of heart laterality defects in lft1^-/- and lft2^-/- mutants (ventral views). Live embryos were scored at 30 hpf. Whereas lft2^-/- mutants typically exhibit normal heart laterality, lft1^-/- mutants frequently have misplaced hearts. Despite this heart laterality defect, lft1^-/- mutants are viable.

Figure 1—figure supplement 2. lefty double mutants and morphants have distinct phenotypes.

About 97% of double lft morphants die by 24 hr post-fertilization (hpf) and undergo strongly disrupted gastrulation (Agathon et al., 2001; Feldman et al., 2002). In contrast, most lft1^-/-;lft2^-/- mutants underwent gastrulation and survived past 24 hpf. Axis bifurcation occurred with variable penetrance in lft1^-/-;lft2^-/- mutants, but has not been reported in double lft morphants. Double morphants exhibited expansion of sox32 and tb at 30% epiboly (Agathon et al., 2001; Chen and Schier, 2002), whereas expansion was only evident in double mutants starting at gastrulation (Figure 3, Figure 3—figure supplement 1). In addition, lft transcripts are present at low or undetectable levels prior to sphere stage (Pauli et al., 2012), and lft1^-/-;lft2^-/- adult females crossed to wild type males produce viable offspring (data not shown), suggesting that differences in morphant and mutant phenotypes are unlikely to be due to inhibition of maternally contributed lft. We also note that loss of sqt, but not cyc, reduced mesendodermal expansion and gastrulation defects in double morphants (Chen and Schier, 2002; Feldman et al., 2002), but loss of either sqt or cyc partially rescued morphology in double mutants (Figure 4). These differences suggest that, as other work has shown (Kok et al., 2015; Schulte-Merker and Stainier, 2014), off target morpholino effects may confound studies of gene function. (A–N) Wild type (A–D, I–K) and lft1^-/-;lft2^-/- mutant (E–H, L–N) embryos were injected with 7 ng lft1 morpholino (B,F,J,M), 7 ng lft2 morpholino (C,G,K,N), or 7 ng of each morpholino together (D,H). At 12 hpf (A–H), wild type and mutant embryos injected with both morpholinos exhibited severe gastrulation defects not observed in uninjected embryos. At 1 day post-fertilization (dpf) (I–N), all embryos injected with both morpholinos were dead. (O) Percentage of embryos alive at 1 dpf. Number of living/total embryos at 1 dpf: uninjected wild type = 38/39, wild type +lft1 MO = 25/26, wild type + lft2 MO = 18/19, wild type + lft1/2 MOs = 0/22, uninjected lft1^-/-;lft2^-/-= 47/52, lft1^-/-;lft2^-/- + lft1 MO: 7/24, lft1^-/-;lft2^-/- + lft2 MO = 1/27, lft1^-/-;lft2^-/- + lft1/2 MOs = 0/22. lft1 MO sequence: 5’-cgcggactgaagtcatctttcaag-3'. lft2 MO sequence: 5’-agctggatgaacagagccatgct-3'.