Functional Dissection of YA, an Essential, Developmentally Regulated Nuclear Lamina Protein in Drosophila melanogaster

Abstract

The Drosophila YA protein is a nuclear lamina component whose function is essential to initiate embryonic development. To identify regions of YA required for its action in its normal cellular context, we made targeted mutations in the YA protein and tested their consequences in flies and embryos in vivo. We found that critical amino acids are distributed along the length of the YA molecule, with functionally important regions including the N- and the C-terminal ends, the cysteine residues in YA’s two potential zinc fingers, a serine/threonine-rich region, and a potential maturation-promoting factor or mitogen-activated protein kinase phosphorylation target site, ITPIR. In addition, several Ya mutations showed intragenic complementation, with N-terminal mutations complementing C-terminal mutations, suggesting that YA proteins interact with one another. In support of this interaction, we demonstrated by immunoprecipitation that YA molecules are present in complexes with each other. Finally, we showed that the C-terminal 179 amino acids of YA are necessary to target, or retain, YA in the nuclear envelope.

The nuclear lamina is a proteinaceous network underlying the nucleoplasmic side of the inner nuclear membrane (reviewed in references 9, 10, and 41). It is hypothesized to provide the structural framework for the nuclear envelope as well as anchoring sites for interphase chromosomes, thus organizing the structure of the nucleus and chromatin (reviewed in references 9, 10, and 41). The major components of the lamina are lamins, a family of intermediate filament-like proteins (see references 8, 35, and 38 for reviews of lamins). The lamina also contains nonlamin components. Examples are the Drosophila melanogaster Young Arrest (YA) protein (26, 30) and the vertebrate MAN antigens (44).

To better understand the molecular roles of nuclear lamina proteins, it is important to identify functionally important regions of these proteins. Several approaches have been taken to this question (11, 12, 17, 33, 54, 56, 60–62). However, these previous studies were carried out by necessity either in vitro or in yeast, which might not accurately reflect in vivo conditions in the metazoan lamina, or by ectopic expression in tissue culture cells, in which endogenous protein could affect the results seen with the ectopically expressed protein variants. Since the YA protein plays a developmentally essential role, and mutations that eliminate its function therefore arrest development, YA provides a unique example in which the functional domains of a nuclear lamina protein can be dissected in its normal in vivo environment and in the absence of any residual function of the protein.

YA is a maternally provided protein whose action is required at the start of Drosophila embryonic development (26, 29). Loss of YA function causes female sterility, with zygotes from homozygous mutant mothers arresting very early in development (26, 29). YA is present in developing female germ cells, postmeiotic oocytes, and 0- to 2-h (cleavage-stage) embryos (26, 28–30, 52). YA’s localization in cleavage nuclei of embryos is cell cycle dependent: YA is in the nuclear lamina from interphase through metaphase but dissociates from the nuclear periphery at anaphase and telophase (26). Phenotypic analysis suggests that YA function is required for the transition from meiosis to mitosis and that YA may play a role in organizing chromatin structure and coordinating the nuclear activities required for the first mitotic division (29). Upon ectopic expression, YA associates with polytene chromosomes (32), further supporting a role of YA in organizing chromosomes within the nucleus. In transgenic flies, a wild-type Ya cDNA can fully rescue the null Ya² mutation (30). Therefore, using transgenic flies carrying different mutant forms of YA, classic complementation assays can be used to identify functionally important regions of YA. We have taken this approach to target potential functional domains of YA and to determine their roles in vivo.

The regions that we chose for targeted mutagenesis were largely based on examining the predicted YA sequence. Though YA protein shows no significant similarity to other known proteins in the database, it contains several motifs that might be of functional relevance (26, 29): YA has two putative nuclear localization signals, consistent with its being in the nuclear lamina. It also contains two potential Cys₂-His₂-type zinc fingers and an SPKK motif, which are known DNA binding motifs (1, 4, 5, 14, 55) and thus might mediate YA’s function in organizing chromosomes within the nucleus by binding to DNA. YA also contains a glutamine (Q)-rich region, a serine/threonine (S/T)-rich region, and a highly positively charged C terminus, which are potential regions for protein-protein interaction (18, 43) and thus could mediate interactions between YA and other proteins in the nuclear envelope or in chromosomes. Finally, there are many potential phosphorylation sites in the YA protein, including the S/T-rich region and two sites (IT⁴⁴³PIR and FS⁵¹¹PKK) that match consensus for maturation-promoting factor (MPF) phosphorylation target sites (45); ITPIR also matches consensus for phosphorylation by mitogen-activated protein kinase (MAPK) (13). Since YA’s nuclear envelope localization is cell cycle dependent, FSPKK and ITPIR could be potential targets for kinases that regulate the cell cycle-dependent activity of YA.

The results that we report here on transgenic flies carrying deletion and substitution mutations in various regions of YA identify several regions important for YA function in vivo, including YA’s N and C termini, the cysteine residues in the two potential zinc fingers, and one potential MPF and MAPK phosphorylation site (ITPIR). We also developed an assay for biochemically probing YA-YA interactions in vivo and showed that YA proteins can interact with one another. These results confirmed and extended our previous (29) and present findings of genetic interactions between different Ya alleles. Finally, we showed that the C-terminal 179 amino acids of YA are needed for targeting YA to, or retaining YA in, the nuclear envelope.

MATERIALS AND METHODS

Construction of epitope-tagged and targeted mutant Ya constructs.

All epitope-tagged and targeted mutant constructs were made by using the wild-type Ya cDNA described by Lopez et al. (30). This cDNA was cloned into either pUC19 or M13 vectors for further mutagenesis. A combination of standard site-directed mutagenesis, PCR, and cloning techniques was used to generate the epitope-tagged or targeted deletion or substitution constructs. Details of how each mutant was generated, and primer sequences, are described in reference 27 and are available upon request. The constructs are summarized in Fig. 1. All tagged constructs and targeted mutant Ya constructs were verified by sequencing both before and after being cloned into the EcoRI site of the P-element vector pW5g26 (30) for fly germ line transformation. This cloning places the mutated Ya cDNAs under the control of the Drosophila hsp26 promoter.

FIG. 1.

Schematic diagrams and summary of complementation results for all mutant constructs. YA proteins encoded by each mutant construct discussed in this report are shown schematically under a diagram of wild-type YA with motifs indicated. Numbers of independent lines obtained for each mutant construct are also shown, and all lines were analyzed. The position of a mutation is indicated by an asterisk; the letter above the asterisk is the amino acid change introduced. V, a delection; A… .A, two alanine residues with several amino acids in between. Results of complementation of the apparent null mutation Ya² or the leaky alleles Ya⁷⁷ (an N-terminal lesion [29]) and Ya⁷⁰ (a C-terminal lesion [29]) by each mutant protein are also summarized. Complementation was tested either without heat shock (NHS) or with heat shock (HS) as described in Materials and Methods. Levels of complementation were scaled from − to +++++. −, no complementation (no progeny were produced); +, very weak rescue (only pupae or at most one or two adult progeny were produced); +++++, complete rescue, as indicated both by production of progeny numbers comparable to that of sense Ya cDNA as well as by completely wild-type-like embryonic phenotype of embryos from rescued females. ++, +++, and ++++ represent intermediate amounts of complementation (approximately 10 to 30%, 30 to 60%, and 60 to 80%, respectively, of the level for complete rescue). ΔQ4.28.1.1 is a mutation which results in a truncation of the C-terminal 179 amino acids of the YA protein in addition to the Q-rich region deletion.

SY160, a serine-to-tyrosine change at amino acid 160, is a mutation generated fortuitously in the process of making the C2 mutant construct. Both C2-SY160 and SY160 were also included in all the analyses.

ΔQ4.28.1.1 was a mutant line obtained when analyzing transgenic lines carrying the ΔQ mutation. This line produces a truncated YA protein (see below). To identify mutation(s) in the Ya transgene from ΔQ4.28.1.1, the Ya transgene from the ΔQ4.28.1.1 line was cloned by using PCR. For both Northern blot and 3′ RACE (rapid amplification of cDNA ends) analyses, total RNAs from heat-shocked ΔQ and ΔQ4.28.1.1 transgenic flies were prepared by using an RNeasy RNA isolation kit (Qiagen). Northern blots were then probed with a Ya cDNA probe labeled with ³²P. The 3′ RACE system from GIBCO-BRL was used to amplify the C-terminal end of Ya cDNA from ΔQ4.28.1.1. PCR products were then cloned into pBluescript for sequencing analysis by the Cornell Biotechnology Program Automated Sequencing Facility.

Fly germ line transformation.

The N(HA)1, N(HA)3, C(HA)3, ΔITPIR, ΔFSPKK, and ΔITPIR-ΔFSPKK constructs were injected into w; Δ2-3 Sb/TM6 embryos (48). All other constructs were injected into w; Δ2-3(99B) embryos, using the protocol of Park and Lim (42) with small modifications. G₀ flies were crossed to z¹w^11E4 flies for several generations to generate stable transformant lines carrying each construct. Independent lines for each construct were verified by genomic Southern blotting, and at least two independent single-insert lines per construct were tested unless otherwise noted (Fig. 1).

Complementation assays.

Male flies carrying each mutant construct were crossed to one of the following three kinds of virgin females: X^X Ya² [X^X, y² Ya² w^bf spl sn³/Y, y⁺ Ya⁺ w⁺ B^s], X^X Ya⁷⁷ [X^X, y² Ya⁷⁷ cv v f/Y, y⁺ Ya⁺ w⁺ B^s], and X^X Ya⁷⁰ [X^X, y² Ya⁷⁰ cv/Y, y⁺ Ya⁺ w⁺ B^s]. Ya² is an apparent null allele (26), whereas Ya⁷⁰ and Ya⁷⁷ are two leaky but complementing Ya alleles (29, 37). Ten female progeny from each of these crosses were checked for fertility as specified by Lopez et al. (30), but either without heat shock or with heat shock. For heat shock treatment, the progeny from the above crosses were heat shocked either three times a day for 3 days or once a day for 3 days, each time at 37°C for 1 h. For any fertile crosses, the resulting progeny were examined to verify that they were of the right phenotype. Male flies carrying the sense or antisense Ya cDNA construct (30) were used as either positive or negative controls, respectively. Since flies carrying the N(HA)1 construct are homozygous, and behave exactly like flies carrying the sense Ya cDNA construct based on both female fertility and embryonic phenotypes (see Results), these flies were also used as wild-type controls for some of the experiments described below. For each mutant construct, all independent lines were tested a minimum of three times in each of the three complementation assays. Ovaries from the females used for complementation of Ya² were also dissected and subjected to Western blot analysis in order to correlate levels of transgene expression with levels of complementation.

To test complementation of Ya² by N(HA)3/C(HA)3, orange-eyed flies carrying one copy of N(HA)3 were crossed with those carrying one copy of C(HA)3. Red-eyed male progeny, which presumably contain both N(HA)3 and C(HA)3, were crossed with X^X Ya² virgin females. Their red-eyed female progeny were checked for fertility as described above.

Immunofluorescence.

Embryos from females carrying a given mutant construct in the Ya² background were collected over a 0- to 2-h period, fixed, and stained with 4′,6-diamidino-2-phenylindole (DAPI) and anti-YA antibodies as described by Lopez et al. (30).

For accessory gland, third-instar larval brain, or polytene squashes, transgenic animals carrying the mutant constructs were heat shocked at 37°C for 30 min to 1 h and allowed to recover at room temperature for 1 h. Accessory glands were dissected and fixed using the procedure for testis squashes described by Pisano et al. (46) and Williams et al. (58). Larval brain dissection and fixation were carried out according to Williams et al. (58). Polytene squashes were prepared as specified by Lopez and Wolfner (32). The tissues were then stained with affinity-purified anti-YA antibodies or, for double staining, both anti-YA and anti-lamin antibodies. Monoclonal anti-lamin antibody T40 (47) and affinity-purified polyclonal anti-lamin antibodies (51) were the kind gifts of H. Saumweber and P. Fisher.

For all stainings, flies carrying the sense Ya cDNA (30) or the N(HA)1 construct were used as positive controls, and flies carrying the antisense Ya cDNA construct were used as negative controls. Expression of each YA construct in the animals whose accessory glands or brains were stained was confirmed by Western blot analysis of the fly or larval carcasses from which the tissues had been dissected.

Immunoprecipitation and Western blot analysis.

Transgenic flies carrying one copy of either the N(HA)3 or the C(HA)3 construct (orange-colored eyes) were crossed to transgenic flies carrying one copy of the sense Ya cDNA or each mutant Ya construct (also orange-colored eyes). Red-eyed male flies, which contained both the hemagglutinin (HA)-tagged construct and the sense Ya cDNA or the mutant construct, were used as sources of proteins for immunoprecipitation assays. These red-eyed male flies were heat shocked at 37°C for 1 h and allowed to recover at room temperature for 1 h. Then they were homogenized in 1× lysis buffer (0.5% Triton X-100, 0.5% sodium deoxycholate, 20 mM Tris-HCl [pH 7.5], 50 mM NaCl), using 10 μl of buffer per fly. The homogenates were centrifuged at 4°C for 5 min. The supernatant was then incubated on ice for 1 to 2 h with either anti-YA antibodies or monoclonal anti-HA antibody (Boehringer Mannheim) followed by incubation with Pansorbin beads (Calbiochem) for 1 to 2 h on ice. The beads were spun down, washed eight times, each time with 1 ml of lysis buffer, boiled in sodium dodecyl sulfate (SDS) sample buffer (26), and loaded on an SDS–7.5% polyacrylamide gel (6 fly-equivalents per lane). Proteins were transferred to nitrocellulose filters and probed with affinity-purified anti-YA antibodies (26). For detection, the enhanced chemiluminescence Western blotting system (Amersham Corp.) was used as described in the supplier’s instruction manual.

Tissues or whole flies used for Western blot analysis were homogenized in SDS sample buffer. SDS-polyacrylamide gel electrophoresis and Western blotting were performed as described above.

Assays for position effect variegation (PEV) and viability.

Homozygous males (except in the case of the C1-C2 and C1-C2-FAPKK mutations, when heterozygous males were used) carrying each mutant construct were crossed to In(1)l^v231/FM6 virgin females (6). Eggs produced in a 48-h period were heat shocked once a day at 37°C for 1 h, for the subsequent 10 days of fly development. Progeny were counted. The absolute male viability (AMV) was calculated as numbers of In(1)l^v231/Y males over the number of total females. The relative male viability was calculated as AMV for each construct over AMV for either the antisense Ya cDNA construct or the N(HA)1 construct. Calculations were done according to Dimitri and Pisano (6).

For viability assays, homozygous flies carrying each of the mutant constructs were allowed to lay eggs in vials for either a 24-h period or a 2-h period. Flies carrying the N(HA)1 or the antisense Ya cDNA construct were tested in parallel as controls. The parental flies were then discarded, and the eggs in each vial were counted. The vials were then subjected to heat shock three times a day for 10 days, each time at 37°C for 1 h. The pupae and flies eclosed from each vial were then counted, and the percentage of flies eclosed was determined.

RESULTS

We made all of the targeted mutations in the wild-type, untagged Ya cDNA. These constructs are summarized in Fig. 1. They include mutations in potential regions for DNA binding (C1, C2, C1-C2, ΔFSPKK, FAPKK, C2-FAPKK, and C1-C2-FAPKK), protein-protein interaction (ΔQ, ΔS/T, and ΔQ-ΔS/T) and phosphorylation (ΔITPIR, ΔFSPKK, ΔITPIR-ΔFSPKK, IAPIR, FAPKK, IAPIR-FAPKK, FEPKK, ΔS/T, and SY160). We used the untagged Ya cDNA to make the mutant constructs because all the terminally HA-tagged Ya constructs which we initially made for functional dissection analysis could not be used for our purposes. Those initial tagged constructs contained one or three copies of the influenza virus nine-amino-acid HA tag (22, 59) at either end of the YA protein [N(HA)1, N(HA)3 and C(HA)3]. YA proteins produced by N(HA)3 or C(HA)3 were nonfunctional, and the YA protein produced by the N(HA)1 construct, although fully capable of rescue of Ya mutations, was not detectable by anti-HA antibodies.

All of the epitope-tagged and targeted mutant constructs were introduced into flies through germ line transformation (as described in Materials and Methods), and stable transformant lines were obtained for all constructs (Fig. 1). In all cases but one, there were no effects of the constructs on viability and fertility of Ya⁺ flies carrying the transgene, in agreement with results reported for the wild-type Ya cDNA (sense Ya cDNA [30]). The one exception was that mutations in the ITPIR motif caused dominant lethality upon ectopic expression in early z¹w^11E4 pupae; however, in other genetic backgrounds, ectopic expression of this mutant YA had no effect on viability or fertility of the transgenic animals (27). To further test for effects of ectopic expression of all the constructs shown in Fig. 1, we implemented the assay of PEV, which is very sensitive to the organization of chromosomes (reviewed in references 53 and 57). Lopez (31) had shown previously that ectopic expression of wild-type YA has no effect on the level of male lethality associated with the variegating lethal In(1)l^v231 (6). We tested whether the lethality of In(1)l^v231 was suppressed or enhanced by the ectopic addition of the mutant YA proteins. We observed that ectopic expression of any mutant YA proteins similarly had no effect on PEV (data not shown). We then went on to characterize the function of all the mutant YA proteins.

Targeted mutations identified regions of the YA protein important for its function.

The in vivo functions of all constructs were tested by their ability to rescue progeny production by the putative null Ya² mutation. Except for N(HA)1, YA proteins from all mutant constructs were present at levels comparable to those from control flies carrying the sense Ya cDNA construct (examples are shown in Fig. 2). Complementation tests were carried out under conditions either without heat shock or with heat shock (as described in Materials and Methods). Heat shock treatment increased the level of expression of the Ya transgene in ovaries above the non-heat shock basal level driven by the hsp26 promoter in the constructs (data not shown). For some partially functional mutants, the higher level of expression improves their ability to complement Ya mutations (Fig. 1). Flies carrying the fully functional sense Ya cDNA construct (30) were used as positive controls. Results from these analyses, summarized in Fig. 1, have allowed the identification of regions important for YA function, as described below.

FIG. 2.

Western blots of male transgenic flies carrying certain targeted mutations. Extracts from heat-shocked transgenic male flies were subjected to electrophoresis on an SDS–7.5% polyacrylamide gel. Bradford assays were used to ensure that the same amount of total protein (75 μg) was loaded in each lane. A Western blot of the gel was probed with anti-YA antibodies as described in Materials and Methods. Proteins from flies carrying the antisense Ya cDNA, the sense Ya cDNA, ΔQ, ΔQ4.28.1.1, ΔS/T, and ΔQ-ΔS/T (two different lines) were probed. While full-length YA from transgenic males has an apparent molecular mass of 97 kDa, and ΔQ, ΔS/T, and ΔQ-ΔS/T have apparent molecular masses of 96, 87, and 86 kDa, as predicted, ΔQ4.28.1.1 produces a YA protein of only 70 kDa.

N-terminal mutations.

N(HA)1, which has only nine amino acids of the HA tag inserted right after the ATG, fully complemented Ya². However, when 48 amino acids (three copies of the HA tag and associated linker sequences) were inserted at the same position, in N(HA)3, the mutant protein could rescue Ya² only very weakly. This finding suggests that the N terminus of YA protein is very sensitive to perturbations.

C1 and C2, which are alanine substitution mutants of the two cysteines in YA’s first (C1) or second (C2) potential Cys₂-His₂-type zinc finger, failed to complement Ya². YA carrying both the C1 and C2 mutations (C1-C2) also failed to complement Ya². Thus, the four cysteine residues in the two potential zinc fingers are essential for YA function. Since the two putative Cys₂-His₂ zinc fingers are potential DNA binding sites, and FSPKK is another potential DNA binding motif (4, 55), we also made and tested C2-FAPKK and C1-C2-FAPKK mutations. Both failed to complement Ya², similar to the C2 mutation alone.

We also tested the SY160 mutation since serine 160 could be a potential phosphorylation target site. SY160 fully complemented Ya². A C2-SY160 mutation behaved like the C2 mutation alone. These results suggest that serine 160 is not essential for YA function.

Therefore, these complementation results suggest that the N terminus of YA is very sensitive to perturbations and is important for YA function. Previously, we had shown that a single ethyl methanesulfonate-induced alteration of the YA N terminus, Ya⁷⁷, affected YA function (29). The present data provide further evidence strongly implicating the N terminus in YA function.

Q-rich and S/T-rich regions.

Q-rich and S/T-rich regions are potential regions for protein-protein interaction (18, 43). Their importance in YA function was thus tested by deletions of each region. ΔQ, which has a deletion of 12 amino acids of the Q-rich region, fully complemented Ya², suggesting that the Q-rich region is not essential for YA function. In contrast, ΔS/T, which has a deletion of 99 amino acids of the S/T-rich region, can complement Ya² only very weakly. Therefore, the S/T-rich region is required for YA function. Deletions of both the Q-rich region and the S/T-rich region (ΔQ-ΔS/T) behaved similarly to the ΔS/T mutation by itself, giving very weak complementation of Ya², consistent with the conclusion that the Q-rich region is not required for YA function.

Potential MPF phosphorylation sites (ITPIR and FSPKK).

Two sets of mutants were made to test the importance of the two motifs that match the consensus of MPF phosphorylation sites in YA. We deleted each or both of the motifs (ΔITPIR, ΔFSPKK, and ΔITPIR-ΔFSPKK) and made alanine substitution mutations at the serine or threonine residue (IAPIR, FAPKK, and IAPIR-FAPKK) to further test their roles in phosphorylation. Since FSPKK is part of YA’s potential nuclear localization signal, we also made a glutamic acid substitution of the serine residue in the FSPKK motif (FEPKK) to test its involvement in regulating nuclear entry of YA. Dephosphorylation of the Xenopus transcription factor NF7 (nuclear factor 7) in its nuclear localization signal allows the protein’s entry to the nucleus, whereas changing the phosphorylation site to a glutamic acid to mimic the negative charge of phosphorylation blocks entry of the protein to the nucleus (26).

As shown in Fig. 1, mutations in the FSPKK motif, including ΔFSPKK, FAPKK, and FEPKK, fully complemented Ya². Thus, the FSPKK motif is not essential for YA function. In contrast, mutations at the ITPIR motif, ΔITPIR and IAPIR, affected YA function. The IAPIR mutation completely abolished YA function: alone or in combination with FAPKK, IAPIR-YA cannot rescue Ya² at all. This lack of complementation is not due to a lower level of expression of the IAPIR protein in the transgenic lines (data not shown). Therefore, the ITPIR motif is required for YA function. However, a deletion of ITPIR, ΔITPIR, still retains partial YA function: ΔITPIR weakly complemented Ya². We hypothesize that this partial function might be due to a structural change in the ΔITPIR mutant protein. Deleting the FSPKK motif from ΔITPIR (ΔITPIR-ΔFSPKK) abolished the protein’s ability to complement Ya².

C-terminal mutations.

The C terminus of the YA protein also is sensitive to perturbations and is important for YA function: C(HA)3, which has three copies of the nine-amino-acid HA tag (plus linker sequences) inserted immediately before the stop codon, almost completely abolishes YA function (Fig. 1).

ΔQ4.28.1.1 was obtained fortuitously through analysis of transgenic lines carrying the ΔQ mutation. YA from flies of this transgenic line had a smaller apparent molecular mass (70 kDa) than YA from flies carrying only the ΔQ mutation (96 kDa [Fig. 2]). The ΔQ4.28.1.1 mutant protein failed to complement Ya², suggesting that the missing amino acids in this mutant are essential for YA function. Sequence analysis of the Ya transgene from ΔQ4.28.1.1 showed that the only mutation present in its Ya transgene was the original deletion of the Q-rich region. However, Northern blot analysis indicated the presence of an aberrant transcript recognized by the Ya cDNA probe in this transgenic line: the Ya transcript from controls was 2.9 kb, and that from ΔQ4.28.1.1 was 4.6 kb (data not shown). This result suggested that the mutant YA protein in ΔQ4.28.1.1 is likely a result of the Ya transgene integrated in a region where the C-terminal part of Ya is deleted through a novel splicing event as can occur at sites of P-element insertion (e.g., reference 20). To test this, the 3′ end of the Ya cDNA from this transgenic line was amplified by 3′ RACE. Sequence analysis of the 3′ RACE product indicated that the Ya sequence ends at nucleotide 1612 followed by sequences from a previously unknown region of the genome (GenBank accession no. U96751). This RNA encodes a predicted fusion protein containing 517 amino acids of the YA protein followed by 9 novel amino acids before a stop codon, consistent with the apparent molecular weight of YA from ΔQ4.28.1.1 on an SDS-polyacrylamide gel (Fig. 2). Therefore, the YA protein in ΔQ4.28.1.1 is missing the last 179 amino acids of the wild-type YA protein, and this mutant protein completely lacks Ya function (Fig. 1).

These results further support the conclusion based on our analysis of two existing ethyl methanesulfonate-induced mutations (29) that the C-terminal part of YA is essential for YA function.

Targeted mutations defined separable domains of the YA protein.

Although complementation of Ya² is the most stringent test for function of a mutant protein, the existence of interallelic complementation of Ya alleles (29, 37) permits tests for whether portions of the protein have functionality even if the protein as a whole is nonfunctional. Therefore, we tested each mutant protein for its ability to complement the Ya⁷⁷ and Ya⁷⁰ mutations, which are leaky mutations at the N and C termini, respectively, of YA (29). As for complementation of Ya², we also used both non-heat shock and heat shock conditions. The results are summarized in Fig. 1. As expected, all mutant proteins that can fully complement Ya² also complemented Ya⁷⁰ and Ya⁷⁷ (Fig. 1). These include N(HA)1, ΔQ, SY160, ΔFSPKK, FAPKK, and FEPKK. Therefore, we will concentrate only on the other mutants here.

N-terminal mutations.

All of the N-terminal mutations either complemented very weakly [N(HA)3] or failed to complement (C1 and C2) the N-terminal mutation Ya⁷⁷. However, they all complemented the C-terminal mutation Ya⁷⁰ significantly better. For example, C1, which did not complement Ya⁷⁷, fully complemented Ya⁷⁰. Similar results were obtained for C2, though a lower level of complementation of Ya⁷⁰ was seen. This result suggests that mutation of C2 has a more dramatic effect on YA function than that of C1. The double mutation C2-SY160 or C1-C2 behaved like the C2 mutation alone, suggesting that SY160 is not important for YA function; furthermore, simultaneous mutations of the two potential zinc fingers did not have a detectable synergistic effect. The fact that these N-terminal mutations complement the C-terminal Ya⁷⁰ mutation indicates that these mutations retained a functional C-terminal domain.

The C2-FAPKK mutation, however, failed to complement both Ya⁷⁷ and Ya⁷⁰. Since the C2 mutation alone still weakly complemented Ya⁷⁰, while FAPKK by itself was fully functional, the complete loss of function by C2-FAPKK suggests that these two motifs might have functional redundancy or may interact. Similar to the C2-FAPKK mutation, the triple mutant C1-C2-FAPKK also failed to complement both Ya⁷⁷ and Ya⁷⁰.

S/T-rich regions.

Both ΔS/T and ΔQ-ΔS/T only weakly complemented Ya⁷⁰, but they fully complemented the Ya⁷⁷ mutation. Therefore, these mutant proteins might have a more defective C-terminal domain. This suggests that the C-terminal domain of YA extends from amino acid 336 (beginning of the S/T-rich deletion) to the C-terminal end.

Mutations at the ITPIR motif.

Both ΔITPIR and IAPIR mutations complemented Ya⁷⁰ and Ya⁷⁷, in contrast to their weak or absent complementation of Ya². In addition, the ΔITPIR-ΔFSPKK or IAPIR-FAPKK mutation behaved similarly to the ΔITPIR or IAPIR mutation, respectively, in this assay. These results suggest that mutations in the ITPIR motif still retain function of both the N and C domains.

C-terminal mutations.

C(HA)3 complemented Ya⁷⁰ and Ya² poorly but complemented Ya⁷⁷ much better. This is the reverse of the situation with N(HA)3, which rescued Ya⁷⁰ better than Ya⁷⁷. These results indicate that C(HA)3 still retains partial N-terminal function.

Mutant YA proteins from ΔQ4.28.1.1 failed to complement any of the Ya alleles (Fig. 1), suggesting that the 179 amino acids that are truncated in ΔQ4.28.1.1 are essential for YA function and that the ΔQ4.28.1.1 mutation is likely a null Ya mutation.

Heterozygous combinations of N(HA)3 and C(HA)3 complemented Ya².

Since all of the N-terminal mutations complemented the C-terminal mutations, we also tested whether N(HA)3 and C(HA)3 in heterozygous combination was functional by testing whether this combination could rescue the Ya² mutation. As shown in Fig. 1, neither tagged protein by itself could complement the Ya² mutation without heat shock. However, flies carrying both N(HA)3 and C(HA)3, even without heat shock treatment, could rescue the sterility of Ya² (data not shown).

Therefore, mutations at the N terminus can complement mutations at the C terminus. This result confirms the findings of Liu et al. (29), who reported complementation of one C-terminal mutation (Ya⁷⁰) and one N-terminal mutation (Ya⁷⁷). It extends those findings significantly since the interallelic complementation of and by several N- and C-terminal mutations rules out that the phenomenon reflects aberrant properties of two specific alleles. Rather, it suggests that YA protein is organized in a domain-like structure, with N and C termini of the protein being separable domains.

YA proteins coimmunoprecipitate.

One model to explain the above intragenic complementation is that YA proteins form complexes with each other and that the combination of a functional N terminus and a functional C terminus in trans can allow YA function. To test this, we developed a biochemical assay to directly determine whether YA associates with itself in vivo.

We generated transgenic flies that carry two Ya transgenes: the sense Ya cDNA construct and an HA-tagged construct, either N(HA)3 or C(HA)3. Proteins were then extracted from adult males, which express both constructs but have no endogenous YA protein, and precipitated with anti-HA or, as a control, anti-YA antibodies. On SDS-polyacrylamide gels, the YA protein encoded by the sense Ya cDNA can be distinguished from N(HA)3 or C(HA)3 due to the addition of about 5 kDa of amino acids in the HA-tagged proteins. Controls showed that anti-HA antibody can precipitate N(HA)3 and C(HA)3 proteins but not untagged YA (Fig. 3A). However, when flies containing both N(HA)3 and the sense Ya cDNA were used for immunoprecipitation with an anti-HA antibody, both tagged and untagged proteins were present in the pellet (Fig. 3A). This precipitation is specific, since tubulin, which is not expected to interact with YA, was not detected in the pellets of immunoprecipitates (data not shown).

FIG. 3.

Immunoprecipitation of untagged YA or mutant YA associated with N(HA)3 or C(HA)3. (A) YA and N(HA)3. Transgenic male flies carrying either the sense Ya cDNA or N(HA)3 or both were extracted, and their proteins were subjected to immunoprecipitation. Western blots were probed with anti-YA antibodies as described in Materials and Methods. Lanes 1, 4, and 7 show YA proteins in extracts without immunoprecipitation. Lanes 3, 6, and 9 show proteins in immunoprecipitates with anti-HA antibody (+anti-HA). Note that untagged YA does not precipitate with this antibody (lane 3) except in the presence of N(HA)3 (lane 6). N(HA)3 is precipitated by anti-HA antibody in the presence or absence of untagged YA (lanes 6 and 9). Lanes 2, 5, and 8 show immunoprecipitation of YA (untagged or tagged) from these extracts to demonstrate that untagged YA is immunoprecipitable with anti-YA antibodies (+anti-YA). All lanes have the same amount of starting material. Results with YA and C(HA)3 are essentially the same (data not shown). (B) IAPIR-YA and N(HA)3. Transgenic male flies carrying one or both of IAPIR-YA and N(HA)3 were used for immunoprecipitations as described above. Results with IAPIR-YA and C(HA)3 are essentially the same (data not shown).

We then used the immunoprecipitation assay to test each targeted-mutant YA protein for its ability to interact with other YA molecules. Both N(HA)3 and C(HA)3 were used for all tests, just as described above. All mutant proteins coimmunoprecipitated with N(HA)3 and C(HA)3 (with results for IAPIR shown in Fig. 3B), consistent with the genetic complementation results described above. Moreover, since C2-FAPKK, C1-C2-FAPKK, and ΔQ4.28.1.1 can still coimmunoprecipitate with N(HA)3 and C(HA)3, the failure of these mutant proteins to rescue any Ya alleles is not due to their inability to interact with other YA molecules but rather is due to the complete loss of function of these mutant proteins. Since none of the targeted mutations affect YA-YA interaction, regions involved in YA-YA interaction must have been maintained in all mutant constructs. It is possible that the regions involved in YA-YA interaction were not targeted or that redundant regions are involved.

Thus, we observe complexes containing at least two YA molecules in Drosophila extracts, supporting our molecular explanation for the intragenic complementation described above.

The C-terminal 179 amino acids are essential for the proper nuclear envelope localization of YA.

YA function appears to correlate with its nuclear envelope localization (29). Therefore, failure of a mutant protein to be functional could be a result of its not being able to be properly localized to the nuclear envelope. To address this question, we tested all mutant proteins for the ability to localize to the nuclear envelope.

For those mutant constructs that rescue Ya², localization of the encoded proteins was tested by staining 0- to 2-h embryos collected from females carrying each construct in the Ya² background. As predicted based on complementation analysis, all of these mutant YA proteins localized to the nuclear envelope in a cell cycle-dependent manner, similar to wild-type YA protein (Fig. 4). Their nuclear envelope localization was also confirmed by staining accessory glands from males that ectopically express each construct (data not shown).

FIG. 4.

Localization of mutant YA proteins in embryos. Zero- to 2-h embryos were collected from females carrying the sense Ya cDNA (A to F) or FAPKK (G to L) construct in the Ya² background and stained with DAPI (A, C, E, G, I, and K) and anti-YA antibodies (B, D, F, H, J, and L). FAPKK has a cell cycle-dependent nuclear envelope localization like that of the YA produced by the Ya sense cDNA construct. The same staining pattern was seen for all other mutant proteins that were tested in this assay. The bar represents 10 μm and applies to all panels.

For those mutant constructs that failed to rescue Ya², their nuclear envelope localization had to be assayed upon postembryonic ectopic expression in Ya⁺ animals, since Ya² mutant embryos have abnormal nuclear envelopes (29). Male accessory glands and larval brains were dissected from heat-shocked transgenic animals carrying each mutant construct and stained with anti-YA antibodies, as previously done for sense Ya cDNA (30). Except for ΔQ4.28.1.1, all mutant YA proteins retained their ability to enter the nuclear envelopes of accessory gland nuclei and retained their cell cycle dependence as seen in larval brains (examples are shown in Fig. 5). These results suggest that the failure of these mutant proteins to complement Ya² is not due to their inability to be localized to the nuclear envelope.

FIG. 5.

Localization of mutant YA proteins in nuclei of male accessory glands and larval brain squashes. (A) Male accessory glands from transgenic flies carrying the sense Ya cDNA (A and B) or C2-FAPKK (C and D) construct were stained with DAPI (A and C) and anti-YA antibodies (B and D). The bar represents 10 μm and applies to all panels. (B) Larval brains from transgenic flies carrying the sense Ya cDNA (A to F) or C2-FAPKK (G to L) construct were stained with DAPI (A, C, E, G, I, and K) and anti-YA antibodies (B, D, F, H, J, and L). The bar represents 10 μm and applies to all panels. YA protein is visible in nuclear envelopes of accessory gland nuclei, as well as interphase and prophase nuclei in larval brain squashes. Staining of anaphase nuclei also shows weak nuclear envelope staining, but it appears more diffuse. The staining pattern shown here is the same as for flies carrying the sense Ya cDNA (30). A similar staining pattern was seen for all the mutant constructs listed in Fig. 1 except ΔQ4.28.1.1. (C) Accessory glands (A and B) and larval brains (C and D) were dissected from ΔQ4.28.1.1 heat-shocked males or larvae and stained with DAPI (A and C) and anti-YA antibodies (B and D). Stainings of brain squashes of larvae carrying the antisense Ya cDNA (E and F) are shown as a negative control. Photographs for panels C to F were taken using the same exposure time. Notice the nucleoplasm staining of YA in both tissues of ΔQ4.28.1.1 and that the staining is excluded from the nucleolus. The bar represents 10 μm and applies to all panels.

YA proteins from ΔQ4.28.1.1, however, failed to localize to the nuclear envelope. Staining of accessory glands and larval brains from heat-shocked ΔQ4.28.1.1 animals showed that the YA protein from ΔQ4.28.1.1 is not localized to the nuclear envelope (Fig. 5). Instead, the ΔQ4.28.1.1 YA protein showed nucleoplasmic staining significantly above the background level seen in flies carrying the antisense Ya cDNA. The localization of the YA protein in ΔQ4.28.1.1 flies suggests that the lesion in this protein does not interfere with YA’s entry into the nucleus but does prevent its entry into or retention by the nuclear envelope. Therefore, the C-terminal 179 amino acids are essential for the proper nuclear envelope localization of YA.

All of the mutant proteins retain their ability to bind polytene chromosomes upon ectopic expression.

Phenotypic analyses of Ya mutant eggs and embryos suggest that YA may be needed to organize chromosomes within the nucleus (29). YA protein is capable of associating with polytene chromosomes upon ectopic expression (using transgenic flies carrying the sense Ya cDNA [30]). Therefore, the loss of function of some mutant proteins could be due to their failure to interact with chromatin. We tested all of the nonfunctional mutant proteins for the ability to bind polytene chromosomes upon ectopic expression. As shown in Fig. 6, all mutant proteins, including those with single, double, or triple mutations of the potential DNA binding motifs, associate with polytene chromosomes. There are no obvious differences between the intensity or the banding pattern of staining of the control and any of the mutant proteins. Therefore, none of the mutant proteins, including those altering the potential DNA binding regions, is essential for polytene chromosome binding.

FIG. 6.

Localization of mutant YA proteins to polytene chromosomes of larval salivary glands. Polytene chromosomes from salivary glands of transgenic larvae carrying the sense Ya cDNA (A and B, as controls) or the transgene in ΔQ4.28.1.1 line (C and D), as an example, were stained with DAPI (A and C) and anti-YA antibodies (B and D). The staining pattern seen for the sense Ya cDNA is as reported by Lopez and Wolfner (32); the staining pattern of ΔQ4.28.1.1 is the same. For all mutant constructs listed in Fig. 1, a staining pattern similar to the one shown was observed. The bar denotes 20 μm and applies to all panels.

DISCUSSION

We used a genetic assay to define important regions of the Drosophila YA nuclear envelope protein in its normal in vivo context. Our studies were made possible by the existence of nonfunctional and partially functional alleles of ya, and by the absolute requirement for YA function to allow embryos to pass through a critical developmental transition. By introducing epitope-tagged and targeted Ya mutant constructs into flies and testing their abilities to complement the maternal lethality of existing Ya mutations, we identified regions important for YA function in vivo. To our knowledge, this is the first study of this type of a metazoan nuclear envelope protein. As mutations are identified in other metazoan nuclear envelope proteins (e.g., a partially functional Drosophila lamin Dm₀ mutant was reported recently [24]), this type of analysis can be extended to other nuclear envelope components.

Our results showed YA’s function requires both its N- and C-terminal ends, the cysteine residues in its two potential zinc fingers, its S/T-rich region, and an ITPIR motif. Critical amino acids are distributed along the whole length of the YA molecule. Previous genetic analysis showed that an N-terminal Ya mutation can complement a C-terminal Ya mutation (29, 37). Here we used a set of mutations at either end of the YA protein to show that the complementation reported in the previous studies is not simply aberrant properties of two alleles; rather, YA has separable domains. In addition to noting the N- and C-terminal domains and distribution of essential elements in YA, we can draw several conclusions, as discussed below.

YA is present in complexes with other YA molecules.

Ya mutant alleles show intragenic complementation, with all lesions at the N terminus complementing all alleles with C-terminal lesions (references 29 and 37 and this study). These results suggest that individual YA proteins have separable domains and that YA protein might interact with itself. Here, we provided biochemical evidence, in the form of coimmunoprecipitation, in support of this interaction in vivo (Fig. 3). Thus, both genetic and biochemical evidence suggest that YA proteins are in complexes with each other. At this point, we cannot distinguish if this interaction is direct or indirect.

Since mutations affecting the N terminus complement those at the C terminus, and since the ΔQ4.28.1.1 deletion does not abolish ability to interact with YA, it is likely that region(s) involved in YA-YA interaction are located more centrally in the YA protein, such as the Q-rich region and the S/T-rich region, which are similar to regions identified as mediating protein-protein interactions (18, 43). Accordingly, we tested the effects of altering these regions on YA function and YA-YA interaction. Deletion of the Q-rich region does not affect the function of YA or its ability to interact with other YA molecules. Thus, either this region is nonessential or its deletion is compensated for by redundant functional domain(s) elsewhere in YA. Deletion of the S/T-rich region abolishes YA function, but both intragenic complementation and coimmunoprecipitation assays suggest that this loss of YA function is not due to the lack of YA-YA interaction. Therefore, the S/T-rich region is required for YA function but not on its own for YA-YA interaction. It is still possible that this region is involved in interactions between YA and other proteins that are required for YA function.

The C terminus of YA regulates YA’s nuclear envelope localization.

The mutant YA protein in ΔQ4.28.1.1 flies fails to be localized to the nuclear envelope, and this protein is nonfunctional in complementation rescue assays. These findings suggest that nuclear envelope localization is essential for YA function, consistent with previous studies showing that YA function is correlated with its nuclear envelope localization (29). The ΔQ4.28.1.1 mutation causes a C-terminal 179-amino-acid truncation of the YA protein. This protein still contains one of the potential nuclear localization signals in the YA protein, and it is capable of entering the nucleus (Fig. 5). Therefore, the 179 amino acids missing in ΔQ4.28.1.1 are essential for targeting YA to, or retaining YA in, the nuclear envelope.

So far, except for the CaaX motif in lamins (7, 16, 19, 21, 23, 34) and a chromatin-binding sequence in Xenopus lamin A that targets this protein to the nuclear periphery (49, 56), no signals have been identified for nuclear lamina targeting. The ΔQ4.28.1.1 mutation thus provides us with an entry point to identify functionally important sequences that allow YA to be localized to the nuclear lamina, which will be very informative in understanding how proteins are targeted to or retained in the nuclear lamina.

ITPIR motif and phosphorylation.

Two regions of the YA protein, ITPIR and FSPKK, match the consensus for MPF phosphorylation target sites (26, 45); ITPIR also matches consensus for a MAPK target site (13). Our mutational analysis showed that the ITPIR motif is required for YA function, though the FSPKK motif by itself is not. Deletions of the ITPIR motif, ΔITPIR and ΔITPIR-ΔFSPKK, as well as substitution mutations of the threonine residues in this motif, IAPIR and IAPIR-FAPKK, all adversely affected the function of YA. In addition, mutations of the ITPIR motif cause dominant lethal effects upon ectopic expression during early pupal development in some genetic backgrounds (27). These results suggest that the threonine residue of ITPIR is important for YA function. Given that this site matches consensus for important developmental or cell cycle-regulated kinases, the threonine of ITPIR may be a site for regulated phosphorylation needed for YA’s function or one of multiple phosphorylation sites that regulate YA’s nuclear entry as in lamin (15, 45), Xenopus NF7 (25, 36, 50), and yeast SWI5 proteins (39, 40). Consistent with this suggestion, we have recently observed that YA is a phosphoprotein and that IAPIR-YA protein is less phosphorylated than wild-type YA (60a).

DNA binding and chromatin association.

Phenotypic analysis of Ya mutant eggs and embryos suggests a role of YA in organizing chromatin structure (29). Consistent with this, ectopically expressed YA associates with polytene chromosomes (32). Since YA protein contains two Cys₂-His₂-type zinc fingers and an SPKK motif, which are potential DNA binding motifs (1, 4, 5, 14, 55), we initially expected that YA functioned by binding to DNA directly through these regions. Indeed, when the cysteine residues were substituted by alanines in either or both the potential zinc fingers, the protein lost its function. This result indicates that these cysteine residues are essential for YA function.

However, when these mutants were tested for polytene binding, all still bound to polytene chromosomes, with the same intensity and banding pattern as seen with wild-type YA. This finding is hard to reconcile with the functional analysis described above. There are several possibilities to explain these results. First, it is possible that YA protein does not directly bind to DNA but binds to chromosomes through interactions with other proteins. The cysteine residues in the potential zinc finger regions might be essential for proper folding of the N-terminal region which is required for proper YA function independent of polytene binding. Second, the zinc finger regions in the YA protein may be involved in protein-protein interaction instead of protein-DNA interaction. Examples are the cysteine-rich zinc fingers which have been shown to function in protein-protein interactions in the presence of zinc (1). Since YA’s second zinc finger is followed by a half zinc finger, which includes two cysteine residues and has similarity to Krox-20 (2, 3, 27), it is possible that a zinc cluster structure can be formed by the four cysteine residues and two histidine residues present in the second and the half zinc fingers. A zinc cluster structure coordinated by six cysteine residues has been found in the yeast transcription factor GAL4 and serves as a DNA binding motif (14). Though without precedence, it is possible that the zinc cluster in YA can serve as a motif involved in protein-protein interaction. Finally, it is possible that the two potential zinc fingers in YA are indeed involved in binding to DNA in embryos, but the presence of YA protein on polytene chromosomes upon ectopic expression is not a proper test for this. Therefore, further experiments are needed to test whether YA can directly bind to DNA in embryos and whether the zinc finger regions are involved in this binding.

In summary, we identified regions that are important for the function or localization of YA, an essential nuclear lamina protein in Drosophila. Regions important for YA function include both its N and C termini, the cysteine residues in its two potential zinc fingers, its ITPIR motif and its S/T-rich region. Sequences needed to target YA to the nuclear lamina lie within its C-terminal 179 amino acids. Interaction between YA proteins can occur, and might mediate the complementation we observe between lesions at the N terminus of the protein and lesions at the C terminus. Our results will guide the focus of further studies of YA’s developmentally essential function to the specific molecular roles and interactions of these crucial regions.