Genetic Analysis of Functional Domains Within the Drosophila LARK RNA-Binding Protein

Corresponding author: F. Rob Jackson, Department of Neuroscience, Tufts University School of Medicine, 136 Harrison Ave., Boston, MA 02111., rob.jackson{at}tufts.edu (E-mail)

Communicating editor: T. W. CLINE

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

LARK is an essential Drosophila RNA-binding protein of the RNA recognition motif (RRM) class that functions during embryonic development and for the circadian regulation of adult eclosion. LARK protein contains three consensus RNA-binding domains: two RRM domains and a retroviral-type zinc finger (RTZF). To show that these three structural domains are required for function, we performed a site-directed mutagenesis of the protein. The analysis of various mutations, in vivo, indicates that the RRM domains and the RTZF are required for wild-type LARK functions. RRM1 and RRM2 are essential for viability, although interestingly either domain can suffice for this function. Remarkably, mutation of either RRM2 or the RTZF results in the same spectrum of phenotypes: mutants exhibit reduced viability, abnormal wing and mechanosensory bristle morphology, female sterility, and flightlessness. The severity of these phenotypes is similar in single mutants and double RRM2; RTZF mutants, indicating a lack of additivity for the mutations and suggesting that RRM2 and the RTZF act together, in vivo, to determine LARK function. Finally, we show that mutations in RRM1, RRM2, or the RTZF do not affect the circadian regulation of eclosion, and we discuss possible interpretations of these results. This genetic analysis demonstrates that each of the LARK structural domains functions in vivo and indicates a pleiotropic requirement for both the LARK RRM2 and RTZF domains.

RNA recognition motif (RRM) proteins constitute a large and functionally diverse class of RNA-binding proteins that have roles in RNA processing and transport, regulation of RNA stability, and translational control (BURD and DREYFUSS 1994 ; SIOMI and DREYFUSS 1997 ). Although sequence-specific RNA-binding activities have been documented for many RRM-type proteins and in vitro assays have been reported for some of these proteins, most RRM proteins have not been functionally characterized in vivo. In part, this is due to the limited capacity for genetic analysis in most organisms. Interestingly, however, the functions of at least two Drosophila RRM proteins have been investigated in vivo, using the approach of site-directed mutagenesis coupled with the generation of transgenic strains expressing altered proteins. One of these proteins, NONA, is an essential product that also functions in male courtship and vision (RENDAHL et al. 1996 ; STANEWSKY et al. 1996 ). The other well-characterized RRM protein is ELAV, a vital splicing factor that is required for the generation of several neuron-specific mRNAs (LISBIN et al. 2000 ).

We are interested in the biological functions of a vital Drosophila gene known as lark, which encodes a protein containing two consensus RRM domains and a C₂HC motif known as a retroviral-type zinc finger (RTZF). The lark gene is broadly expressed during embryogenesis, and mRNA can be detected in the developing nervous system and in nonneural tissues (NEWBY and JACKSON 1993 , NEWBY and JACKSON 1996 ). Both maternal and zygotic lark expression are required for normal embryogenesis. Embryos lacking zygotic expression arrest in development at about the time of germ band retraction (stage 12; NEWBY and JACKSON 1993 ), whereas zygotes lacking maternally inherited lark mRNA (i.e., the maternal component) arrest during early stages of embryogenesis (MCNEIL et al. 1999 ).

Several lines of evidence indicate that LARK functions in the circadian regulation of adult eclosion (the adult ecdysis) in addition to serving a vital role during embryogenesis. Perhaps most important, LARK abundance changes in a circadian manner in extracts prepared from pharate adults (MCNEIL et al. 1998 ). In addition, LARK can be detected within neurons containing a modulatory neuropeptide known as crustacean cardioactive peptide (CCAP; MCNEIL et al. 1998 ), which has an important role in the regulation of ecdysis in Drosophila and other insects (EWER and TRUMAN 1996 ; GAMMIE and TRUMAN 1999 ). Within the CCAP cells, LARK has a unique cytoplasmic localization (in all other neurons, it is nuclear), and there are dramatic circadian changes in LARK immunoreactivity within this neuronal population (ZHANG et al. 2000 ). We have postulated that the clock regulation of LARK in CCAP neurons contributes to the circadian control of adult eclosion.

We want to provide phenotypic evidence that the three consensus RNA-binding motifs of LARK are functional in vivo. RRM domains are known to mediate RNA-binding functions in a large number of different proteins (BURD and DREYFUSS 1994 ). In contrast, the RTZF (a type of "Zn knuckle") is not as well characterized in eukaryotic proteins, but its function has been extensively investigated for the retroviral nucleocapsid protein (COPELAND et al. 1983 ; SUMMERS 1991 ). The nucleocapsid protein contains one or two RTZF domains, depending on the viral species, which bind to and mediate packaging of the RNA genome of the viruses. In addition to the nucleocapsid protein of retroviruses, there are several eukaryotic proteins, including Drosophila Nanos and the 30K subunit of cleavage and polyadenylation specificity factor (CPSF), that contain retroviral-type zinc fingers or derivatives thereof. CPSF 30K is a component of a multimeric complex found in both Drosophila and mammals, which is required for 3' pre-mRNA processing. In Drosophila, it has been shown that the CPSF 30K subunit contains two C-terminal RTZF domains that confer specificity for binding to G- and/or C-rich RNA sequences (BAI and TOLIAS 1998 ). In Nanos protein, sequences similar but not identical to the RTZF motif are known to be critical for translational repression of bicoid and hunchback mRNAs, and mutations affecting the C or H residues of these sequences abolish the activity of Nanos, perhaps by interfering with RNA target recognition (ARRIZABALAGA and LEHMANN 1999 ). Neither CPSF 30K nor Nanos contain RRM domains, and, to our knowledge, there have been no published studies that explore, in vivo, the functional relationship between RRM and Zn-knuckle domains.

We previously demonstrated that maternal lark function is essential for development and that the RTZF of LARK is critical for this maternal function (MCNEIL et al. 1999 ). Females carrying RTZF mutations can survive and mate, but the resulting zygotes undergo developmental arrest at early embryonic stages (MCNEIL et al. 1999 ). In this study, we further characterize RTZF mutants as well as analyze the RRM domains of LARK with regard to maternal, embryonic, and postembryonic functions. To determine if the consensus RRM elements are functional in vivo, we performed a site-directed mutagenesis of LARK in vitro and then created transgenic strains that express each of the mutant proteins in a lark-null genetic background. These studies demonstrate that the RRM domains contribute to both the maternal and embryonic functions of LARK and, in addition, reveal roles for the protein in functions as diverse as mechanosensory bristle formation, wing morphogenesis, and flight. Interestingly, this genetic analysis indicates that the RRM2 and RTZF domains of LARK act together in vivo to determine protein function.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

Construction of an epitope-tagged lark transgene:
To create a transgene encoding an epitope-tagged LARK protein, a sequence encoding the hemagglutinin (HA) epitope was ligated in frame at the 3' terminus of the lark-gene coding region immediately prior to the stop codon. PCR cloning was performed using the lark rescue vector p-W8-larkP1 (MCNEIL et al. 1999 ) to produce pW8-larkHA.7. The pW8-larkHA.7 vector contains a wild-type copy of the lark gene including the normal promoter sequence (a 3.2-kb genomic DNA fragment including 1 kb of upstream sequence). It is capable of restoring wild-type function and completely rescuing embryonic development in lark¹ homozygotes (MCNEIL et al. 1999 ).

Site-directed mutagenesis and production of mutant transgenic strains:
Site-directed mutations in the RRM domains were created using a 0.5-kb EcoRV-SalI lark DNA fragment containing both the RRM1 and RRM2 domains. This DNA fragment was subcloned into M13, mp19 and used as a template for mutagenesis with the Sculptor in vitro oligonucleotide-directed mutagenesis system (Amersham, Arlington Heights, IL). A single amino acid substitution (F > A) was introduced into the RNP2 motif of RRM1 or RRM2 (see Fig 1) using the following oligonucleotide primers: RRM1-RNP2, 5'-TCAAGATTCCCGATGGCTAACTTGAACGTGC-3'; and RRM2-RNP2, 5'-GTTAGATTTCCCCGGCGATTTTCGTGGTCG-3'. The underlined residues of each sequence correspond to the site-directed changes. To produce a transgene carrying RNP2 mutations within both RRM domains, site-directed RNP2 mutations were sequentially introduced into the RRM2 and RRM1 coding domains of the 0.5-kb EcoRV-SalI lark DNA fragment. Direct sequencing of M13, mp19 DNA templates was employed to identify mutants. Mutant DNA fragments were cloned into the pW8-larkHA.7 transformation vector to replace the corresponding wild-type sequence.

View larger version (29K): In this window In a new window Download PPT slide   Figure 1. Schematic representation of LARK protein showing the amino acid substitutions present in single and double mutants. The locations of the two RRM domains and the RTZF are indicated on the shaded horizontal box representing LARK. The asterisk indicates residue 178, which marks the C terminus of a truncated LARK protein. The sequences of the two RNP2 motifs and the RTZF are shown below the schematic of LARK. The underlined residues of the RNP2 and RTZF sequences represent the amino acid substitutions present in the various mutants. The positions and types of substitutions are indicated at the bottom.

RTZF mutations were previously created using a SalI-PstI lark DNA fragment (MCNEIL et al. 1999 ). For these mutations, the first two Cys (C) residues of the RTZF were changed to Tyr (Y) residues (Fig 1). RRM; RTZF double mutants were generated by inserting a SalI-PstI lark DNA fragment with an RTZF mutation into a lark P-element construct carrying an RRM-RNP2 mutation.

Transgenic strains expressing mutant or wild-type LARK proteins were generated by injection of nondechorionated embryos using standard P-element transformation procedures (SPRADLING and RUBIN 1982 ) and the P-element helper plasmid 2-3. Multiple independent transgenic strains were established for each mutant transgene.

Drosophila strains, growth conditions, and crosses:
A standard Drosophila culture medium, used in the lab for many years, was employed for stock maintenance and genetic crosses (NEWBY et al. 1991 ). Fly stocks were maintained and crosses were performed at 25° in a 12-hr light:12-hr dark (LD 12:12) cycle. The lark¹ mutation is a P-element insertion into the 5' untranslated region of lark that causes embryonic lethality (NEWBY and JACKSON 1993 ). The mutation behaves as a null or severely hypomorphic allele on the basis of two observations. First, lark¹/+ flies have about one-half the normal amount of LARK protein (NEWBY and JACKSON 1996 ). Second, the lark¹ mutation is indistinguishable from an excision-generated allele known as lark^R3, which removes lark function and an adjacent vital gene and thus is assumed to be a chromosomal deletion of the region (NEWBY and JACKSON 1993 ). All other Drosophila strains are described in LINDSLEY and ZIMM 1992 . The w¹¹¹⁸ mutation is a null allele of the white (w) gene. TM6 is a third chromosome balancer.

Crosses were performed to generate lark¹ homozygotes carrying a wild-type or mutant transgene (tg-lark). In this way, the viability of site-directed lark mutants was assessed in the complete absence of endogenous LARK. For X chromosome tg-lark insertions, w¹¹¹⁸/Y; lark¹/TM6 males were crossed to w¹¹¹⁸ tg-lark (w+)/w¹¹¹⁸; lark¹/TM6 females. For second chromosome insertions, w¹¹¹⁸/Y; tg-lark (w+)/+; lark¹/TM6 males were mated to w¹¹¹⁸/w¹¹¹⁸; tg-lark (w+)/+; lark¹/TM6 females. Adult progeny were counted to determine the viability of lark¹/ lark¹ tg-lark flies.

Immunoblotting procedures:
Head protein samples (100 µg), representing the different transgenic strains, were denatured, run into 10% SDS-PAGE gels, and blotted to Immobilon-P membrane (Millipore, Bedford, MA) as described (MCNEIL et al. 1999 ). HA-tagged LARK was detected using an anti-HA antibody (Santa Cruz Biochemicals, Santa Cruz, CA) at a dilution of 1:200. To examine gel loading, the blots were reprobed with a mitogen-activated protein (MAP) kinase antibody (Sigma, St. Louis, no. M5670) at a dilution of 1:20,000.

Immunohistochemistry:
Immunocytochemical detection of HA-tagged LARK in paraffin sections of pharate adults was achieved using standard laboratory procedures (ZHANG et al. 2000 ). An anti-HA antibody was employed at a dilution of 1/40 in these experiments.

Flight tests:
Individual flies were tested for flight ability using a modified Sparrow test (DRUMMOND et al. 1991 ; NELSON et al. 1997 ). Flies were collected by aspiration and placed individually in microfuge tubes to recover. To assess flight ability, a tube containing a fly was placed above a small hole in a plastic top covering a 1-liter graduated cylinder. The fly was gently tapped from the tube and the position of first contact with the graduated cylinder was recorded. Contact at the 1-liter mark or higher was scored as a value of 10.0, whereas first contact with the bottom of the cylinder was scored as 0. Contacts at points between the 1-liter mark and the bottom of the cylinder were assigned intermediate scores, proportional to the distance traveled down the cylinder. Each individual was assayed four times and 10 flies were assayed per genotype (five males and five females). Table 2 shows average values for the 40 trials.

Analysis of eclosion rhythms:
Flies were reared at 25° in bottles for 4 days in LD 12:12 and were then transferred to 18° for the remainder of development. Emerging adults were collected every 2 hr over a 48-hr period in an LD cycle. Since only a small percentage of the emerging flies are lark¹ homozygotes, 20–40 bottles were used for each strain analyzed. After 2 days of LD collections, the bottles were transferred to constant darkness (DD) for 3 days. On day 3 of DD, newly emerged adults were collected every 2 hr for one full 24-hr cycle. Several independent transgenic strains were tested for each site-directed lark mutant, and each strain was analyzed two or more times.

	RESULTS AND DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

Production and expression of LARK mutant proteins:
Within the 80-amino-acid consensus RRM domain, there are two highly conserved regions of eight (RNP1) and six (RNP2) residues. A number of studies have shown that these short sequences are critical for protein:RNA interactions (KENAN et al. 1991 ; OUBRIDGE et al. 1994 ), and amino acid substitutions within either RNP1 or RNP2 severely perturb RRM function in vivo (RENDAHL et al. 1996 ; STANEWSKY et al. 1996 ; LISBIN et al. 2000 ). For example, the substitution of a single tyrosine (Y) residue at position 2 of the hexapeptide RRM1-RNP2 sequence of the Drosophila NONA protein results in a mutant phenotype that is as extreme as that produced by a nonA null mutation (STANEWSKY et al. 1996 ). That is, this single Y residue is crucial for RRM function. Structural studies of the splicing factors U1A and Drosophila Sxl indicate that the homologous Y residue participates in ring-stacking interactions with RNA ligands (KENAN et al. 1991 ; OUBRIDGE et al. 1994 ; HANDA et al. 1999 ).

Invariably, there is a Phe (F) or Y residue present at position 2 of the RNP2 hexamer, and, in LARK, this residue is F. To perturb LARK function, we replaced the F residue at position 2 of RRM1-RNP2 or RRM2-RNP2 with an Ala (A) residue (Fig 1). A single residue of RNP2 was altered in an attempt to perturb RNA recognition without destroying the structure of the RRM domain. While such a mutation might not completely abrogate RNA-binding activity, it is known that a similar mutation in U1A substantially reduces binding activity and decreases the affinity of specific RNA binding by 25-fold (JESSEN et al. 1991 ). Thus, we assumed that this single residue replacement would significantly reduce or abolish the binding activity of an RRM. In addition to such single mutants, we created double mutants in which the consensus F residue was changed in both RRM1 and RRM2.

Mutant LARK proteins were expressed in vivo, making use of a transgene under control of the normal lark-gene promoter (see MATERIALS AND METHODS). Importantly, flies carrying a wild-type lark⁺ transgene expressed an HA-tagged LARK protein of the predicted size that could be detected by Western blotting using anti-HA antibody (Fig 2). Several independent transgenic strains were established for each lark mutant transgene and expression of LARK in vivo was confirmed by immunoblot analysis of adult head protein extracts using an anti-HA antibody. As shown in Fig 2, LARK-HA protein could be detected in head extracts of all transgenic strains that were examined. Furthermore, endogenous LARK is normally detected as a protein doublet of 38–40 kD (NEWBY and JACKSON 1996 ), and HA-tagged proteins corresponding to the two isoforms were detected in all of the characterized transgenic strains.

View larger version (38K): In this window In a new window Download PPT slide   Figure 2. Western analysis of head protein samples from transgenic strains expressing wild-type and mutant forms of LARK. Samples were prepared from lark+ flies that were homozygous for a given transgene insertion. (A) Western analysis. The types of mutant strains and the strain numbers are indicated above the data. All blots were sequentially probed with anti-HA and anti-MAP kinase (MAPK) antibodies (see MATERIALS AND METHODS). The asterisk indicates a nonspecific band seen in some experiments. (B) Relative level of LARK-HA protein present in the different transgenic strains, as a ratio of LARK-HA to MAPK. For two strains listed in Table 1, 64-3 and 8-1A, Western analysis was not performed.

RRM domains and viability:
We showed previously that an RTZF mutant transgene could rescue the zygotic lethality of lark¹ homozygotes, albeit at a reduced frequency relative to a wild-type transgene, and that rescued females exhibited a maternal phenotype (see below). To determine if the consensus RRM motifs of LARK represent functional domains, we tested transgenes expressing mutant LARK proteins (RRM1, RRM2, or both RNA-binding motifs) for their ability to restore viability to lark¹ mutants. In these experiments, we characterized the viability of flies expressing a mutant protein as their only source of LARK. In crosses involving X-linked transgene insertions, rescued progeny carried a single copy of the lark transgene (see Table 1). For autosomal insertions, rescued progeny carried either one or two copies of the transgene, and it was not possible to distinguish one-dose and two-dose progeny on the basis of eye color, because a single copy of the mini-w⁺ marker in these transgenes confers red eye color.

Table 1 shows the viable progeny classes that arose from these rescue crosses; column 5 of this table shows the rescued lark¹/lark¹; tg-lark progeny class. The lethality of the lark¹ mutation is fully penetrant, and homozygous lark¹ individuals lacking the transgene were never observed in these crosses. In the case of the RRM1, RNP2 single mutant (F > A; see Fig 1), most transgene insertions (tg-lark) were able to rescue lark¹ homozygotes at a frequency close to that expected for the wild-type transgene (tg-lark⁺). Rescued flies were fertile and displayed no apparent mutant phenotypes, suggesting that the RRM1 motif of LARK might be dispensable for embryonic development. We note that the tg-lark 1-1 transgene was exceptional and did not rescue lark¹/lark¹ mutants to adulthood. A lack of rescue for this transgene may be a consequence of reduced LARK expression relative to other similar insertions (Fig 2), perhaps because of a genomic position effect. We also recognize that some of the variability in percentage rescue could be due to gene dosage effects since some of the progeny carry only a single copy of the transgene, whereas other progeny are expected to carry two copies of the transgene. However, for crosses involving X-linked insertions, all of the rescued progeny carry only a single copy of the transgene, and in these crosses rescue was essentially complete for both wild-type and mutant (RRM1) transgenes (Table 1).

Transgene insertions carrying an RNP2 mutation in RRM2 were also able to rescue the zygotic lethality associated with the lark¹ mutation (Table 1). However, the degree of rescue was reduced relative to either the wild-type or RRM1 mutant transgenes, and only 20–69% of the expected lark¹ homozygotes were observed in the rescue crosses. Nonetheless, the observation that both RRM1 and RRM2 mutant transgenes can rescue viability suggests that the two domains might be partially redundant for this function. To address this issue, a transgene that carried RNP2 mutations in both RRM1 and RRM2 was generated (Fig 1), and this double mutant was tested for the ability to rescue viability. As shown in Table 1, the doubly mutant transgene completely failed to rescue the lethality of lark¹ homozygotes for several independent genomic insertions that were tested. Lack of rescue was probably not due to a low level of transgene expression, because these double mutants produced LARK at levels equal to or nearly equal to that seen in strains carrying a wild-type or singly mutated lark transgene (Fig 2). Furthermore, the doubly mutant protein was observed throughout the nervous system and other LARK-containing tissues with an appropriate cellular and subcellular pattern of localization (see later section). These results indicate that the two RRM domains of LARK are at least partially redundant, with regard to viability, and that either one of them is sufficient to provide function.

To our knowledge, this is the first example, in vivo, of functionally redundant RRM domains within a single RRM-type protein. Previous analyses of such domains, in vivo, have indicated that individual RRM elements specify vital functions (STANEWSKY et al. 1996 ; LISBIN et al. 2000 ). In the case of ELAV, for example, mutation of any one of the three individual RRM domains results in lethality (LISBIN et al. 2000 ). As postulated by LISBIN et al. 2000 , single ELAV mutants might be lethal because the RRM domains of the protein (in particular RRM1 and 2) cooperate to create a single RNA recognition site. In agreement with that idea, the results of certain RNA-binding studies, conducted in vitro, have shown cooperative interactions between RRM domains (BURD et al. 1991 ; CACERES and KRAINER 1993 ; KANAAR et al. 1995 ; MAYEDA et al. 1998 ; PARK et al. 2000 ). It is known, for example, that multiple RRM domains within both Sxl protein and poly(A)-binding protein (PABP) are required for specific and high-affinity binding to RNA targets (BURD et al. 1991 ; KANAAR et al. 1995 ). A similar result was obtained with the Human D (HuD) protein (an ELAV homolog), for which high-affinity binding depends on the presence of all three RRM domains (PARK et al. 2000 ).

Obviously, however, the two RRMs of LARK do not provide completely redundant functions. RRM1 mutants show better survival than RRM2 mutants (Table 1), and, more important, both RRM2 and RTZF mutants exhibit a number of other phenotypes not observed in RRM1 mutants. The analysis of these additional phenotypes suggests that the RRM2 and RTZF domains function together in vivo (see below). Consistent with a partial divergence of function, the two LARK RRMs are 63% identical to one another within a 65-amino-acid conserved interval that includes RNP1 and RNP2 (NEWBY and JACKSON 1996 ).

A truncated LARK protein, lacking C-terminal sequences, does not rescue viability:
As all three potential RNA-binding domains are present in the N-terminal half of LARK, we wondered whether a truncated version of the protein might rescue viability. We constructed a lark-HA transgene (C-LARK) that encodes a LARK protein containing residues 1–178 (see Fig 1). This truncated protein is predicted to contain both RRM domains, the RTZF, and 15 residues C-terminal to the RTZF (it lacks residues 179–352). We tested the ability of this transgene to rescue the lark¹ lethal phenotype as described above. No rescue of lark¹ homozygotes was observed in five independent transgenic strains expressing the truncated LARK protein (Table 1). The lack of rescue was not due to absence of LARK expression, because a protein of the predicted size (24 kD) could be detected in head extracts of all five transgenic strains (Fig 2). Thus, the inability of the truncated protein to rescue viability is a consequence of its abnormal structure, activity, or intracellular localization (see later section).

RRM2 and RTZF mutants exhibit identical novel phenotypes:
Although lark transgenes with single mutations in either RRM2 or the RTZF were able to rescue embryonic lethality, the percentage rescue was significantly reduced relative to wild-type transgenes (Table 1). In addition to the reduced viability, bristle and wing phenotypes were observed in these two types of mutants (Fig 3). Although Fig 3 shows representative phenotypes for RTZF mutants, the defects observed for RRM2 mutants were the same. A bristle defect, for example, was observed in both RRM2 and RTZF mutants and was restricted to the four large mechanosensory bristles of the scutellum (Fig 3); all other bristles, both large and small, were normal in distribution and morphology. Defects in the mutant flies included the presence of ectopic scutellar bristles (Fig 3B, thin arrow) or the absence of one or more scutellar bristles (Fig 3B, thick arrow) or malformed (i.e., forked) bristles (Fig 3B, arrowhead). One or more of these bristle abnormalities were seen in the majority of mutants; penetrance of the phenotype was 50–100% for RRM2 mutants (three independent insertions) and 66–99% for RTZF mutants (four independent insertions, three of which exhibited 99% penetrance for this phenotype). We note that this bristle phenotype was never observed in lark¹ homozygotes carrying wild-type or RRM1 mutant transgenes, nor was it observed in control siblings heterozygous for lark¹ (with or without the mutant transgene), although infrequently such control flies (and a Canton-S wild-type population in the lab) had a duplication of one of the anterior scutellar bristles (asterisk in Fig 3A). Thus, we conclude that the mutant phenotype is a specific consequence of altering the RRM2 or RTZF domains and that it is caused by reduced lark function, rather than a dominant-negative effect of transgene-encoded LARK. As already mentioned, the effects of lark mutations seem to be specific for the large scutellar bristles, and this is different from mutations of another Drosophila RRM gene known as musashi (msi), which have more global effects on mechanosensory cell development (NAKAMURA et al. 1994 ).

View larger version (63K): In this window In a new window Download PPT slide   Figure 3. Scutellar bristle and wing phenotype of the RRM2 and RTZF mutants. (A) Photo of the notum and scutellum of a wild-type Canton-S fly. The asterisk indicates a bristle duplication for one of the anterior scutellar bristles. (B) An RTZF mutant (strain 12-1) that is missing a posterior scutellar bristle (thick arrow) as well as possessing an ectopic bristle (thin arrow) and a malformed posterior scutellar bristle (arrowhead). (C) The "held out" wing phenotype of RRM2 and RTZF mutants. The photo depicts an RTZF mutant (strain 12-1), but the wing phenotype of RRM2 single mutants was the same.

RTZF and RRM2 mutants exhibited abnormal wing morphology in addition to the scutellar bristle phenotype (Fig 3C). In the case of either mutant, the wings were abnormally curved (decurved) and extended from the thorax in an abnormal orientation (i.e., held out, similar to Dichaete mutants). The penetrance for this phenotype was 100% or 48–100% for RTZF or RRM2 mutants, respectively, and wing abnormalities were never observed in lark¹ homozygotes carrying wild-type or RRM1 mutant transgenes. Flies carrying either an RTZF or RRM2 mutation also frequently exhibited extra (ectopic) wing veins (not shown).

Perhaps not surprisingly, RRM2 and RTZF mutants appeared to be flightless as well. They moved slowly and never flew when released from the culture vials. To assay flight behavior in a quantitative way, we used a modification of the Sparrow box (DRUMMOND et al. 1991 ; NELSON et al. 1997 ; see MATERIALS AND METHODS). As shown in Table 2, lark¹ homozygotes carrying either an RRM2 or RTZF mutant transgene were flightless, with the exception of four flies in strain 253-3 (an RRM2 mutant), which were able to fly. The remaining six flies of this strain flew poorly or not at all. In contrast, sibling flies heterozygous for the lark¹ mutation (with or without the mutant transgene) and lark¹ homozygotes carrying a wild-type or RRM1 mutant transgene had normal flight behavior.

The abnormal flight of RRM2 and RTZF mutants might simply be a consequence of misshapen wings or alternatively be a symptom of an underlying neural or neuromuscular defect affecting the flight muscle machinery. Mutations in Drosophila calmodulin (CAM), for example, result in flight defects that are likely due to abnormal neuromuscular synapse function (NELSON et al. 1997 ). LARK protein can be detected throughout the developing nervous system (MCNEIL et al. 1999 ; ZHANG et al. 2000 ), and mutations might compromise neural or neuromuscular development.

Both RRM2 and the RTZF are required for female fertility:
We previously demonstrated that the RTZF mutant transgene could rescue zygotic lark function (i.e., the lark¹ embryonic lethal phenotype). However, RTZF mutant females were completely sterile, even though mutant males exhibited normal fertility (MCNEIL et al. 1999 ). These mutant females mated with wild-type males and deposited eggs, but no evidence of larval hatching was ever observed (MCNEIL et al. 1999 ). We interpreted that result to mean that such females have a defect in oogenesis and/or early development. Indeed, it is known from that previous study that embryos completely lacking the lark maternal component become arrested in development at early stages of embryogenesis (MCNEIL et al. 1999 ). We were interested in determining if mutations in either RRM resulted in a similar maternal phenotype. This was clearly not the case for the RRM1 mutant, for which males and females exhibited completely normal fertility (Table 3). However, RRM2 mutant females were not fertile (Table 3). Females from the 253-3 mutant strain, for example, mated normally with wild-type males and deposited eggs (albeit a low number), but, as with RTZF mutant females, there was no evidence of larval hatching (Table 3). Inspection of these eggs revealed that some had not developed at all, whereas others had arrested in development at early stages of embryogenesis. Importantly, RRM2 mutant males had normal fertility (Table 3). Thus, it would appear that the maternal function of LARK requires intact RTZF and RRM2 domains. These domains might be important for RNA transport, localization, or translational control functions that are critical for oogenesis and/or early development.

RRM2-RTZF double mutants are equivalent in phenotype to RTZF single mutants:
Because RRM2 and RTZF single mutants exhibited identical phenotypes, we were interested in whether the two domains interacted to determine LARK function. To address this issue, we examined phenotypes in RRM2-RTZF double mutants (see Fig 1). If the RTZF functions obligatorily with RRM2 in vivo, then the phenotype of a double mutant is expected to be the same as that of either single mutant. In contrast, if both domains function independently to determine phenotype, then the double mutant might have a significantly more extreme phenotype than either single mutant. As shown in Table 1, the viability of RRM2-RTZF double mutants was low and similar (although perhaps not identical) to that of RRM2 and RTZF single mutants. One of the two RRM2-RTZF mutant transgenes seemed to rescue poorly (only 0.4% rescue), but the other one rescued about as well as RTZF mutant transgenes. The scutellar bristle and wing phenotypes of the RRM2-RTZF double mutant were identical to those seen in either single mutant, and the penetrance for both phenotypes was similar to that observed in single mutants (e.g., for strain 47-1A, penetrance values were 94 and 97%, respectively, for the wing and bristle phenotypes).

As expected, all RRM2-RTZF double mutants were flightless (Table 2) and doubly mutant females were not fertile when mated to wild-type males (Table 3). Thus, in every way, the double mutants resembled single RRM2 or RTZF mutants, and there was little or no evidence of an additive effect of mutations. Together with the phenotypic similarity of single mutants, this result indicates that the RRM2 domain of LARK acts in concert with the RTZF; perhaps the two domains interact to form a single RNA-binding surface. Alternatively, the RTZF might modulate the affinity with which RRM2 binds RNA targets. It is thought, for example, that a Zn knuckle present in another RRM protein, the human splicing factor h9G8, influences RNA target recognition (CAVALOC et al. 1999 ), although this conclusion is based on binding preferences that were determined from in vitro RNA selection analyses. Our results are evidence of a functional relationship, in vivo, between the RRM2 domain and the RTZF of LARK.

Intracellular localization of transgene-encoded LARK proteins:
One possible explanation for the abnormal phenotypes associated with certain mutant transgenes is that the encoded proteins do not have a normal intracellular localization. It is known, for example, that LARK has a nuclear pattern of localization within most cells (including most neurons) but a cytoplasmic localization within a peptidergic (CCAP) neuronal population (ZHANG et al. 2000 ). Thus, we examined LARK distribution in wild-type and mutant transgenic strains, using immunocytochemical procedures. We carried out these studies using paraffin sections of late-stage pupae (pharate adults) and an anti-HA antibody to detect HA-tagged LARK protein. In previous studies, we performed a detailed examination of LARK distribution in pharate adult animals (ZHANG et al. 2000 ).

As shown in Fig 4, the expected pan-nuclear pattern of staining was observed in transgenic flies expressing wild-type (Fig 4A) or RRM1-RRM2 double mutant (Fig 4C) proteins. This result demonstrates that the lack of rescue seen with RRM1-RRM2 mutant transgenes is not due to an altered intracellular localization of LARK. We note that LARK could not be examined in the CCAP neuronal population in these experiments because the anti-HA antibody does not reliably detect LARK within these cells (data not shown).

View larger version (117K): In this window In a new window Download PPT slide   Figure 4. Spatial distribution of LARK-HA protein in transgenic strains. All flies carried two copies of a lark transgene (wild type or mutant) in a lark+ genetic background. (A, C, and E) Distribution of LARK-HA protein in the different strains. (B, D, and F) DAPI staining of the same sections. (Insets) Magnified views of the indicated regions. (A and B) Wild-type LARK-HA protein (strain 3-38). (C and D) double RRM mutant LARK-HA protein (strain 74-2). (E and F) C-LARK-HA protein (strain 42-1).

The viability analysis described previously showed that a truncated LARK protein (C-LARK), lacking residues 179–352, cannot restore function (Table 1). The C-terminal half of LARK, downstream of the RNA-binding domains, is highly enriched in proline (Pro) residues and contains at least three different Pro-rich elements that match the consensus for a Src homology 3 (SH3) ligand (NEWBY and JACKSON 1996 ). In addition, the C terminus of LARK contains Arg/Ser-rich regions, and such sequences are capable of acting as a nuclear or subnuclear localization signal in certain other RNA-binding proteins (LI and BINGHAM 1991 ). Therefore, it was of interest to determine if sequences within the C terminus were important for the intracellular localization of the protein.

Strikingly, C-LARK was not detected within the nucleus and was observed to be exclusively localized to the cytoplasm of all neurons (Fig 4E). Such a mislocalization of C-LARK was observed in two independent transgenic strains that were characterized. Furthermore, a cytoplasmic localization of the protein was observed in other tissues in which LARK normally has a nuclear localization (data not shown). These results indicate that sequences within the C terminus of the protein are necessary for nuclear localization. In theory, a mislocalization of the C-LARK protein could be due to either a failure in nuclear translocation or defective protein:RNA target interactions in the nucleus with a consequent effect on localization. In either case, an abnormal intracellular localization of the C-LARK protein would presumably preclude the ability to rescue viability.

Eclosion behavior is normal in LARK mutants:
Because lark was originally identified on the basis of a circadian eclosion phenotype (NEWBY and JACKSON 1993 ), it was of interest to determine if any of the site-directed mutations affected the clock regulation of this behavior. We surmised that a complete absence of LARK function might lead to arrhythmic eclosion behavior if the protein were essential for circadian control. In addition, more recent results indicate that pan-neural LARK expression inhibits adult eclosion (animals die as pharate adults), whereas overexpression of LARK within the CCAP neuronal population leads to a posteclosion phenotype (the animals eclose but then do not inflate their wings properly; A. SCHROEDER and F. R. JACKSON, unpublished results). The latter phenotype might reflect an abnormality in peptidergic signaling involving CCAP and a neurosecretory hormone known as Bursicon, which is involved with hardening and tanning of the cuticle (NIJHOUT 1994 ). Thus, we were also interested in determining if site-directed LARK mutants exhibited additional types of eclosion or posteclosion phenotypes.

As described in a previous section, RRM1 mutants are morphologically normal. Although viability is reduced by RRM2 or RTZF mutations, we observed that adults of the mutant populations emerge normally and undergo normal cuticle hardening and wing inflation (notwithstanding the abnormal wing shape and orientation of the mutants; Fig 3). To examine the clock control of eclosion in mutant populations, developing pupae were entrained to a LD 12:12 cycle at 18° and emerging flies were collected and counted every 2 hr over a 2-day period. As depicted in Fig 5A, lark¹ homozygotes carrying either an RRM1 or RRM2 mutant transgene exhibited normal eclosion rhythms in LD 12:12. Rhythmicity persisted in DD in these mutants, indicating that the clock control of eclosion was normal. Thus, single mutations in one RRM or the other do not affect the clock control of eclosion. It might be the case that the two RRM domains of LARK function redundantly for this physiological process, as is the case for embryonic development, or, alternatively, the protein may not be essential for clock control. Theoretically, it is also possible that different biological processes require different levels of LARK activity and that viability, fertility, and certain developmental processes are more sensitive to reductions in LARK activity than is the regulation of eclosion.

View larger version (67K): In this window In a new window Download PPT slide   Figure 5. Adult eclosion profiles for wild-type and mutant LARK-HA strains. All flies were homozygous for the lark1 mutation and carried one or two copies of the indicated transgene. (A) Eclosion profiles for a wild-type transgene-bearing strain (3-38) and for RRM1, RNP2 (89-1) and RRM2, RNP2 (84-2) mutant strains. Shown to the right of each strain is eclosion on day 3 of DD. (B) Light:dark entrainment and free-running rhythmicity for an RTZF mutant strain (103-1). (C) Free-running eclosion rhythms for RRM1; RTZF (10-1A) and RRM2; RTZF (47-1A) double mutants. Zeitgeber time is the time of day relative to the entraining light:dark cycle (time 0 means lights on). Circadian time represents the subjective time of day after transfer of strains to constant darkness. Times of light and darkness are indicated by the open and solid rectangles, respectively, shown at the top (A and B). The solid and shaded rectangles (A and B, top right; C, top) indicate the previous photoperiodic regime to which strains were entrained, prior to transfer to constant darkness. All strains showed normal entrainment to LD 12:12 and normal free-running eclosion rhythms.

A lack of effect on rhythmicity was also observed with RTZF single mutants (Fig 5B), indicating that this domain may not be required for the clock control of behavior. As with other phenotypes, we examined rhythmic behavior in RRM-RTZF double mutants. Similar to the single mutants, however, double RRM (1 or 2)-RTZF mutants exhibited normal free-running eclosion rhythms (Fig 5C).

Conclusions:
In this study, we explored the functions of LARK protein, in vivo, by performing a systematic site-directed mutagenesis of the protein. This structure/function analysis demonstrated that all three potential RNA-binding domains (the two RRMs and the RTZF) are functional in vivo. The identification of RNA targets that bind to these domains in vivo will provide important insights about the function of LARK in processes as diverse as embryonic development, wing and bristle morphogenesis, and female fertility.

Interestingly, certain of the targeted lark mutations differentiate the functions of the RRM and RTZF domains (see Fig 6). For example, single mutations of RRM1 do not have discernable phenotypic effects, whereas RRM2 or RTZF mutants exhibit morphological and female fertility defects. Indeed, the strikingly similar phenotypes of RRM2 and RTZF single mutants (effects on female fertility, wing morphology, flight, and scutellar bristles) suggest that the two domains might act in concert to mediate the several different functions. Perhaps these two domains physically interact to specify a single RNA target-binding site. Notwithstanding the lack of phenotype in single RRM1 mutants, our studies indicate that this domain is functional in vivo. Whereas single RRM mutants are viable, double RRM1-RRM2 mutants fail to complete development. That result indicates that the two domains are, at least in part, functionally interchangeable and that either one is sufficient for viability. Our results also show that daily eclosion rhythms are normal in RRM1, RRM2, and RTZF mutants. Presumably, the RTZF is not critical for circadian function. The observation that single RRM mutants have normal rhythmicity can be interpreted in several alternative ways. It is possible that the RRM domains are not essential for clock control. Alternatively, the RRM mutations may not entirely abolish RNA-binding activity, and the regulation of eclosion might be less sensitive to reductions in LARK activity than other phenotypes. Finally, the two RRM domains may function redundantly for this process, as is the case for viability.

View larger version (25K): In this window In a new window Download PPT slide   Figure 6. Schematic of LARK showing how lark mutant phenotypes and functions map to the different protein domains.

	ACKNOWLEDGMENTS

We thank Mike Berne (Tufts University Core Facility) for help with DNA sequencing and Kazuhiko Kume and Bikem Akten for assistance with the characterization of eclosion rhythms. This work was supported by National Institutes of Health (NIH) grant HL59873 (F.R.J.) and NIH NRSA grants to G.P.M. (F32 NS11017) and A.J.S. (F32 MH12283).

Manuscript received March 19, 2001; Accepted for publication June 26, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

ARRIZABALAGA, G. and R. LEHMANN, 1999  A selective screen reveals discrete functional domains in Drosophila Nanos. Genetics 153:1825-1838[Abstract/Free Full Text].

BAI, C. and P. P. TOLIAS, 1998  Drosophila clipper/CPSF 30K is a post-transcriptionally regulated nuclear protein that binds RNA containing GC clusters. Nucleic Acids Res. 26:1597-1604[Abstract/Free Full Text].

BURD, C. G. and G. DREYFUSS, 1994  Conserved structures and diversity of functions of RNA-binding proteins. Science 265:615-621[Abstract/Free Full Text].

BURD, C. G., E. L. MATUNIS, and G. DREYFUSS, 1991  The multiple RNA-binding domains of the mRNA poly(A)-binding protein have different RNA-binding activities. Mol. Cell. Biol. 11:3419-3424[Abstract/Free Full Text].

CACERES, J. F. and A. R. KRAINER, 1993  Functional analysis of pre-mRNA splicing factor SF2/ASF structural domains. EMBO J. 12:4715-4726[Medline].

CAVALOC, Y., C. F. BOURGEOIS, L. KISTER, and J. STEVENIN, 1999  The splicing factors 9G8 and SRp20 transactivate splicing through different and specific enhancers. RNA 5:468-483[Abstract].

COPELAND, T. D., M. A. MORGAN, and S. OROSZLAN, 1983  Complete amino acid sequence of the nucleic acid-binding protein of bovine leukemia virus. FEBS Lett. 156:37-40[Medline].

DRUMMOND, D. R., E. S. HENNESSEY, and J. C. SPARROW, 1991  Characterization of missense mutations in the Act88F gene of Drosophila melanogaster. Mol. Gen. Genet. 226:70-80[Medline].

EWER, J. and J. W. TRUMAN, 1996  Increases in cyclic 3',5'-guanosine monophosphate (cGMP) occur at ecdysis in an evolutionarily conserved crustacean cardioactive peptide-immunoreactive insect neuronal network. J. Comp. Neurol. 370:330-341[Medline].

GAMMIE, S. C. and J. W. TRUMAN, 1999  Eclosion hormone provides a link between ecdysis-triggering hormone and crustacean cardioactive peptide in the neuroendocrine cascade that controls ecdysis behavior. J. Exp. Biol. 202:343-352[Abstract].

HANDA, N., O. NUREKI, K. KURIMOTO, I. KIM, and H. SAKAMOTO et al., 1999  Structural basis for recognition of the tra mRNA precursor by the sex-lethal protein. Nature 398:579-585[Medline].

JESSEN, T.-M., C. OUBRIDGE, C. H. TEO, C. PRITCHARD, and K. NAGAI, 1991  Identification of molecular contacts between the U1 A small nuclear ribonucleoprotein and U1 RNA. EMBO J. 10:3447-3456[Medline].

KANAAR, R., A. L. LEE, D. Z. RUDNER, D. E. WEMMER, and D. C. RIO, 1995  Interaction of the sex-lethal RNA binding domains with RNA. EMBO J. 14:4530-4539[Medline].

KENAN, D. J., C. C. QUERY, and J. D. KEENE, 1991  RNA recognition: towards identifying determinants of specificity. Trends Biochem. Sci. 16:214-220[Medline].

LI, H. and P. M. BINGHAM, 1991  Arginine/serine-rich domains of the su(wa) and tra RNA processing regulators target proteins to a subnuclear compartment implicated in splicing. Cell 67:335-342[Medline].

LINDSLEY, D. L., and G. G. ZIMM, 1992 The Genome of Drosophila melanogaster. Academic Press, New York.

LISBIN, M. J., M. GORDON, Y. M. YANNONI, and K. WHITE, 2000  Function of RRM domains of Drosophila melanogaster ELAV: RNP1 mutations and RRM domain replacements with ELAV family proteins and SXL. Genetics 155:1789-1798[Abstract/Free Full Text].

MAYEDA, A., S. H. MUNROE, R. M. XU, and A. R. KRAINER, 1998  Distinct functions of the closely related tandem RNA-recognition motifs of hnRNP A1. RNA 4:1111-1123[Abstract].

MCNEIL, G. P., X. L. ZHANG, G. GENOVA, and F. R. JACKSON, 1998  A molecular rhythm mediating circadian clock output in Drosophila. Neuron 20:297-303[Medline].

MCNEIL, G. P., X. ZHANG, M. ROBERTS, and F. R. JACKSON, 1999  Maternal function of a retroviral-type zinc-finger protein is essential for Drosophila development. Dev. Genet. 25:387-396[Medline].

NAKAMURA, M., H. OKANO, J. A. BLENDY, and C. MONTELL, 1994  Musashi, a neural RNA-binding protein required for Drosophila adult external sensory organ development. Neuron 13:67-81[Medline].

NELSON, H. B., R. G. HEIMAN, C. BOLDUC, G. E. KOVALICK, and P. WHITLEY et al., 1997  Calmodulin point mutations affect Drosophila development and behavior. Genetics 147:1783-1798[Abstract].

NEWBY, L. M. and F. R. JACKSON, 1993  A new biological rhythm mutant of Drosophila melanogaster that identifies a gene with an essential embryonic function. Genetics 135:1077-1090[Abstract].

NEWBY, L. M. and F. R. JACKSON, 1996  Regulation of a specific circadian clock output pathway by lark, a putative RNA-binding protein with repressor activity. J. Neurobiol. 31:117-128[Medline].

NEWBY, L. M., L. WHITE, S. M. DIBARTOLOMEIS, B. J. WALKER, and H. B. DOWSE et al., 1991  Mutational analysis of the Drosophila miniature-dusky (m-dy) locus: effects on cell size and circadian rhythms. Genetics 128:571-582[Abstract].

NIJHOUT, H. F., 1994 Insect Hormones. Princeton University Press, Princeton, NJ.

OUBRIDGE, C., N. ITO, P. R. EVANS, C.-H. TEO, and K. NAGAI, 1994  Crystal structure at 1.92 A resolution of the RNA-binding domain of the U1A spliceosomal protein complexed with an RNA hairpin. Nature 372:432-438[Medline].

PARK, S., D. G. MYSZKA, M. YU, S. J. LITTLER, and I. A. LAIRD-OFFRINGA, 2000  HuD RNA recognition motifs play distinct roles in the formation of a stable complex with AU-rich RNA. Mol. Cell. Biol. 20:4765-4772[Abstract/Free Full Text].

RENDAHL, K. G., N. GAUKHSHTEYN, D. A. WHEELER, T. A. FRY, and J. C. HALL, 1996  Defects in courtship and vision caused by amino acid substitutions in a putative RNA-binding protein encoded by the no-on-transient A (nonA) gene of Drosophila. J. Neurosci. 16:1511-1522[Abstract/Free Full Text].

SIOMI, H. and G. DREYFUSS, 1997  RNA-binding proteins as regulators of gene expression. Curr. Opin. Genet. Dev. 7:345-353[Medline].

SPRADLING, A. C. and G. M. RUBIN, 1982  Transposition of cloned P elements into Drosophila germ line chromosomes. Science 218:341-347[Abstract/Free Full Text].

STANEWSKY, R., T. A. FRY, I. REIM, H. SAUMWEBER, and J. C. HALL, 1996  Bioassaying putative RNA-binding motifs in a protein encoded by a gene that influences courtship and visually mediated behavior in Drosophila: in vitro mutagenesis of nonA. Genetics 143:259-275[Abstract].

SUMMERS, M. F., 1991  Zinc finger motif for single-stranded nucleic acids?: investigations by nuclear magnetic resonance. J. Cell. Biochem. 45:41-48[Medline].

ZHANG, X., G. P. MCNEIL, M. J. HILDERBRAND-CHAE, T. M. FRANKLIN, and A. J. SCHROEDER et al., 2000  Circadian regulation of the lark RNA-binding protein within identifiable neurosecretory cells. J. Neurobiol. 45:14-29[Medline].

This article has been cited by other articles:

				M. W. Kankel, G. D. Hurlbut, G. Upadhyay, V. Yajnik, B. Yedvobnick, and S. Artavanis-Tsakonas Investigating the Genetic Circuitry of Mastermind in Drosophila, a Notch Signal Effector Genetics, December 1, 2007; 177(4): 2493 - 2505. [Abstract] [Full Text] [PDF]