Models for ASE neuron development and the establishment of left/right asymmetric gene expression profiles. (A) Regulatory network for ASE neuron specification. Green boxes indicate genes whose asymmetric regulation is addressed in this paper. (B) Three models are depicted that may explain the regulatory logic of L/R asymmetric gene expression programs in the ASE neurons. For simplicity, examples are limited to the regulation of ASEL-specific genes.

The ASE motif from ASEL- and ASER-specific genes directs bilateral expression in the ASE neurons. (A) 24 bp encompassing the ASE motifs from ASEL-specific (lsy-6, gcy-7 and lim-6) and ASER-specific (gcy-5) genes are fused via PCR into the multiple cloning site of the gfp expression vector pPD95.75, which contains an uncharacterized basal promoter before the gfp start codon. To exclude the possibility that the basal promoter has any ASE-regulatory information, a 21 bp sequence with no obvious match to the ASE motif was also tested as control and found to not be expressed in ASE; empty vector also produces no expression (data not shown). ASE expression of the gcy-5 ASE-motif construct has been described previously with no mention of the symmetry of its expression (Etchberger et al., 2007). (B) Representative images of the lsy-6 ASE motif driving expression of gfp in ASEL/R. Multiple lines show similar expression for all constructs listed in A. In all cases, expression in the ASE neurons was assessed with a ceh-36^prom::DsRed2 (otIs151) transgene contained in the background. Arrows indicate ASEL and ASER.

Altering ASE motifs can disrupt L/R symmetric gene expression. (A) CHE-1 binds with higher affinity to the ASE motif of the ceh-36 promoter versus that of the gcy-5 promoter. The schematic on the left indicates the structure and consensus sequence of each C2H2 zinc finger with its DNA-contacting residues indicated in red. DNA contacts were abolished by mutating one of the DNA contacting residues in each zinc finger to an alanine. The right panel shows that wild-type CHE-1 shifts less gcy-5 versus ceh-36 probe. DNA binding of mutated CHA-1 protein (mutated as shown in left panel) demonstrates that three zinc fingers (2, 3 and 4) impact on CHE-1 binding to the ASE motif of ceh-36, whereas only two zinc fingers (3 and 4) impact on CHE-1 binding to the ASE motif of gcy-5. See Fig. S2 in the supplementary material for competitive gel shift that further solidifies the differential motif affinity. (B,C) Swapping the high-affinity ASE motif from the ceh-36 promoter into the low-affinity motifs from gcy-5 (B) or gcy-14 (C) results in de-repression of gfp expression in the contralateral neuron. Multiple independent transgenic lines were scored (numbers shown in parentheses) and a representative line is shown. (D) Relative binding affinities of CHE-1 for various ASE motifs. The left panel shows results of quantitative EMSAs where binding to the ASE motif of a labeled ceh-36 probe was competed with unlabeled ASE motifs from the genes indicated [sequences are the same as used by Etchberger et al. (Etchberger et al., 2007)]. In the right panel, the data are transformed into a graphical representation of binding affinities relative to the ceh-36 probe (see Materials and methods). The lsy-mut probe mimics the lsy-6(ot150) mutation in ASE motif (Sarin et al., 2007). ceh-36mut abolishes expression of ceh-36^prom::gfp in vivo (data not shown).

CHE-1 acts in the context of many gene regulatory network motifs. CHE-1 autoregulates its own expression via an ASE motif (Etchberger et al., 2007) and co-regulates many bilaterally expressed target genes by what has been termed a `single input motif' (Alon, 2007). Further diversification programs downstream of CHE-1, which result in differential gene expression in ASEL versus ASER, involve a more complex version of a feed-forward loop (FFL) motif (see text). Depending on the signs of the regulatory interactions, these motifs can be coherent or incoherent (Alon, 2007). The bracket indicates that CHE-1 directly regulates the expression of multiple bistable feedback loop components (Fig. 1A) via ASE motifs. Multiple additional network motifs are also embedded here, as discussed previously (Hobert, 2006). For example, die-1 affects gcy-7 expression via fozi-1 in what appears to be an embedded FFL (Johnston et al., 2006). Note that the ASE motif in the asymmetrically expressed genes is stippled to indicate that, at least in some cases, intrinsic features of the ASE motif (i.e. its affinity to ASE) play a role in determining the laterality of gene expression. The model shown here applies to the ASEL neuron (ASEL genes activated; ASER genes repressed); the opposite holds for the ASER neuron.

Effect of postdevelopmental che-1 RNAi on the execution of ASE terminal fate. (A) Representative examples for RNAi effects on the ASEL fate marker lim-6::gfp, expressed in ASEL (arrows) and on the excretory gland cells (asterisks), which serve as internal controls, as they are not affected by control or che-1 RNAi, but are affected by gfp RNAi. Control RNAi refers to RNAi with empty vector L4440. Note the differences in gfp intensities for che-1 RNAi; as quantified in B, either a reduction of gfp levels (as shown in A) or a complete loss was observed. (B) A quantification of effects of che-1 RNAi on ASEL (lim-6::gfp) and ASER marker (gcy-5::gfp) results are shown.

Cis-regulatory analysis of L/R asymmetrically expressed genes. In all panels, gfp expression was scored in ASEL and ASER. Derepression of gfp expression in one neuron to levels that do not approach normal gfp levels in the contralateral neuron are indicated by `ASEL>ASER' or `ASEL<ASER'; similar expression levels by `ASEL=ASER'. The incomplete nature of derepression in the contralateral neuron may relate to the multicopy structure of extrachromosomal arrays. Multiple independent transgenic lines were scored for each construct. See Fig. S1 for primary gfp data, Fig. S2 for quantification of data (i.e. penetrance of effects in a representative transgenic line) and Fig. S3 for nucleotide sequences (supplementary material). Deletion constructs for the gcy-7, gcy-5 and lim-6 promoters have been described before in the context of identifying the ASE motif (blue arrow, blue box) (Etchberger et al., 2007), but have not been reported for the effect on L/R asymmetry. (A)Cis-regulatory analysis of the gcy-7 locus. All constructs are subcloned reporters except ASE<sup>gcy-7#2</sup>, which were generated by PCR fusion in order to minimize potential effects of vector backbone sequence. Lower panel: model of regulation for asymmetric expression of gcy-7, based on: (1) the partial derepression of gfp upon removal of del2,5,7, which argues for the presence of some other asymmetry input; (2) the requirement for del4a activator motif to drive expression in ASEL in the presence of the del2,5,7 motifs; (3) the lack of a requirement for del4a activator motif to drive expression in ASEL in the presence of the del2,5,7 motifs; and (4) the ability of the CEH-36 binding site del4 to introduce a left/right bias to the ASE motif. In ASEL, expression is induced by che-1 and ceh-36. In ASER, multiple distinct repressive motifs and the absence of an activation mechanism result in the inhibition of che-1-mediated gene expression. Additionally, a motif in the del6 region (del6a) was deleted owing to the high level of sequence conservation and resulted in derepression of gfp in ASER (see Figs S2 and S3 in the supplementary material). The original scanning deletion, del6, did not result in derepression of gfp. (B)Cis-regulatory analysis of the gcy-5 locus. The ASE motif mutated constructs (`mut1') has been described previously (Etchberger et al., 2007) and is shown for comparison only. ¹These constructs were also examined as linear PCR fragments; the wild-type promoter yields a `ASER>ASEL' pattern and the del3,7,11 construct yields a `ASER=ASEL' pattern. Lower panel: model of regulation for asymmetric expression of gcy-5, based on: (1) the observed derepression of gfp in ASEL from the del3,7,11 construct; and (2) the sole requirement of the ASE motif for expression in ASE. (C)Cis-regulatory analysis of the lim-6 locus. All constructs were injected as subcloned circular DNA. Lower panel: model of regulation for asymmetric expression of lim-6, based on: (1) the complete loss of expression of gfp observed in ASEL from the del6 and del7 constructs; (2) the loss of maintained gfp expression in ASEL from the del4a/d constructs, which suggests an asymmetric activation factor is required for the maintained asymmetric expression of lim-6.(D)Cis-regulatory analysis of the lsy-6 locus. All constructs were generated by PCR fusion as subcloned reporter constructs yielded only very weak gfp expression (Sarin et al., 2007). Lower panel: model of regulation for asymmetric expression of lsy-6, based on: (1) the loss of gfp expression following the deletion of the E-box (del5); and (2) the observed ectopic expression from del4.

Characterization of che-1(ot101). (A) L/R asymmetric, but not bilateral ASE fate is affected in che-1(ot101) mutants. See Materials and methods for transgenes used. Pairs of green circles indicate ASE neuron pairs and relative gfp expression levels. (B) Molecular identity of class II che-1 alleles. Alignment of the zinc fingers of che-1 orthologs in other species (flies and sea urchin). The glycine in the linker position that is mutated in ot101 is conserved in other linker regions within the Glass family, but also in many other zinc-finger proteins, including TFIIIA, Krueppel and others (data not shown). Note that zinc finger 3 and 4 touch the invariant part of part of the ASE motif and are absolutely crucial for DNA binding (Fig. 5) (Etchberger et al., 2007), whereas the class II alleles that affect zinc finger 2, which contact the less well conserved parts of the ASE motif, modulate only DNA binding, as shown in C. (C) Class II che-1 alleles display reduced DNA-binding affinity. A representative gel shift experiment is shown.