Phenotypic and molecular characterization of GAL4/UAS-mediated LARK expression

The lark gene was initially identified in a behavioral screen for flies that displayed circadian rhythm defects (Newby and Jackson, 1993, 1996). It encodes an RNAbinding protein of the RNA Recognition Motif (RRM) class. Whereas lark heterozygotes exhibit a rhythm defect, homozygotes die during embryonic development because of a zygotic requirement for the gene product. Subsequently, it was shown that there is also a maternal requirement for lark function during oogenesis and early embryonic development (McNeil et al., 1999). Currently, several fly labs are pursuing studies of LARK function as related to oogenesis and embryonic development. LARK protein contains three potential RNA binding domains, two RNA recognition motifs, and a single retroviral type zinc finger, that have been shown to be important for LARK function during development but appear to play functionally redundant roles in the circadian regulation of eclosion (Newby and Jackson, 1996; McNeil et al., 2000). The LARK protein is widely expressed in most if not all tissues and throughout development (Zhang et al., 2000). In the majority of cells that express LARK, the protein has a nuclear localization. A notable exception occurs in a subset of neurosecretory cells that express the neuropeptide crustacean cardioactive peptide (CCAP), in which LARK is cytoplasmic. Of interest for the rhythm phenotype, CCAP cells exhibit circadian changes in LARK abundance (Zhang et al., 2000). In order to further characterize LARK protein function we generated pP{UAS-lark T:Ivir }(UAS-lark-3HA, Fig. 1, panel a). To distinguish GAL4 driven expression of the UAS-lark transgene from endogenous LARK expression, sequences encoding a triple hemagglutinin (HA) epitope were incorporated into the construct immediately prior to the LARK stop codon. Seven independent UAS-lark3HA transgenic lines were generated and here we report the preliminary characterization of three of these lines. We employed one line carrying a second chromosome insertion (w; P{w mC UAS-lark }94A) and two lines with independent third chromosome insertions (w; P{w mC UAS-lark }9A and (w; P{w mC UAS-lark }23A). All three lines are homozygous viable and fertile. To examine the effect of LARK overexpression, we first used the strong ubiquitous Act5C-GAL4 driver to drive expression of UAS-lark-3HA transgenes. Flies carrying an Act5C-GAL4 third chromosome insert (balanced over TM6b, Tb) were crossed to flies homozygous for a UAS-lark-3HA transgene. Adults carrying both the Act5C-GAL4 driver and any of the three UAS-lark-3HA transgenes were never observed; i.e., the only adult progeny that resulted from these crosses carried the TM6b, Tb chromosome (Table 1). In the case of two of the UAS-lark-3HA lines, 94A and 9A, some of the LARK overexpressing flies were able to develop to the early pupal stage (as indicated by a few Tb pupae in the vials; data not shown). However, these individuals ceased to develop soon after pupal formation and no brown or black Tb pupae were observed in these crosses. In the case of crosses with the 23A UAS-lark-3HA line, Tb pupae were never observed, suggesting that ubiquitous overexpression of LARK from this insertion is lethal prior to pupation. To determine if the observed lethality was caused by increased LARK protein expression in the nervous system, we used an elav-GAL4 driver (c155) to express UAS-lark-3HA transgenes in all differentiated neurons. Such a pan-neural overexpression of LARK caused lethality, the severity of which was dependent on the particular UAS-lark-3HA line employed in crosses. For example, crosses of males bearing the X-linked elav-GAL4 driver to females homozygous for the 23A UAS-lark-3HA transgene failed to produce any female offspring (Table 1). In similar crosses with elav-GAL4 and either the 94A or 9A UAS-lark-3HA transgene, GAL4/UAS flies were observed to eclose at a reduced frequency compared to UAS sibling controls (Table 1). Of the GAL4/UAS flies that do eclose, all display a cuticle-tanning defect characterized by a significant delay in cuticle hardening and pigmentation (not seen in control flies). In addition, the GAL4/UAS flies do not show normal wing expansion (Fig. 1, panel b). With the 94A UAS-lark-3HA transgene, none of the GAL4/UAS flies ever expanded their wings. Approximately 2% of the GAL4/UAS flies displayed nor