The Drosophila melanogaster attP40 docking site and derivatives are insertion mutations of msp-300

The ɸC31 integrase system is widely used in Drosophila melanogaster to allow transgene targeting to specific loci. Over the years, flies bearing any of more than 100 attP docking sites have been constructed. One popular docking site, termed attP40, is located close to the Nesprin-1 orthologue msp-300 and lies upstream of certain msp-300 isoforms and within the first intron of others. Here we show that attP40 causes larval muscle nuclear clustering, which is a phenotype also conferred by msp-300 mutations. We also show that flies bearing insertions within attP40 can exhibit decreased msp-300 transcript levels in third instar larvae. Finally, chromosomes carrying certain “transgenic RNAi project” (TRiP) insertions into attP40 can confer pupal or adult inviability or infertility, or dominant nuclear clustering effects in certain genetic backgrounds. These phenotypes do not require transcription from the insertions within attP40. These results demonstrate that attP40 and insertion derivatives act as msp-300 insertional mutations. These findings should be considered when interpreting data from attP40-bearing flies.


Introduction
For several years, Drosophila melanogaster investigators have used a genome integration method based on the site-specific ϕC31 integrase [1] to target transgenes to specific loci [2]. With this method, ϕC31 integrase catalyzes sequence-directed recombination between a phage attachment site (attP, present within each of >100 attP "docking sites" in Drosophila melanogaster) and a bacterial attachment site (attB, present within the integrating plasmid) [3][4][5][6][7]. By allowing transgene insertion into specific, defined, docking site sequences, the ϕC31 integrase method increases the reproducibility and decreases the variability of transgene expression observed with the random transgene integration utilized by P elements.
Two docking sites, attP2, located at position 68A4 on chromosome III and attP40, located at position 25C on chromosome II, are widely used docking sites for LexA drivers and Gal4-driven Transgenic RNAi Project (TRiP) insertions [8]. These two attP docking sites are favorable because they express inserted transgenes at high levels while maintaining low basal expression [9]. In fact, the Drosophila melanogaster stock center at Bloomington, IN, reports possessing 16,503 Drosophila melanogaster lines carrying attP40 and 14,970 lines carrying attP2; most lines carry TRiP insertions or the activation domains or DNA binding domains from the Janelia split-Gal4 collection (Annette Parks, personal communication). Although originally reported to be located in an intergenic region, between CG14035 and msp-300 [10], attP40 lies within the first intron of certain msp-300 isoforms ( [11] FlyBase FB2022_02). This observation raises the possibility that attP40 might act as an insertional mutation for msp-300. Indeed, it was previously reported that certain insertions into attP40 could cause spreading of the H3K27me3 mark over the large msp-300 exon [12]. Thus, transgenes inserted within the attP40 docking site might affect expression of at least a subset of msp-300 isoforms. MSP-300 (Muscle-specific protein 300 kDa) is a nuclear-associated Nesprin1 orthologue and a component of the Linker of Nucleoskeleton and Cytoskeleton (LINC) complex [13,14]. The C-terminal domain contains a Klarsicht/Anc1/Syne Homology (KASH) domain that interacts with Sad1p/UNC-84 (SUN)-domain-containing proteins, connecting the outer and inner nuclear membranes [15,16]. In Drosophila melanogaster larvae, msp-300 transcription has been reported in muscle (Volk, 1992) and fat body [17]. In larval muscle, msp-300 forms striated F-actin-based filaments that lie between muscle nuclei and postsynaptic sites at the neuromuscular junction. MSP-300 also wraps around immature boutons in response to electrical activity and is required for postsynaptic RNA localization and synaptic maturation [18]. MSP-300 is also required for normal nuclear localization in muscle cells and for integrity of muscle cell insertion sites into the cuticle [14,19,20]. MSP-300 isoforms lacking the KASH domain confer deficits in larval locomotion, localization of the excitatory neurotransmitter receptor GluRIIA at the neuromuscular junction (NMJ), and proper NMJ functioning, independently of its role in muscle nuclear positioning [21]. Non-muscle deficits conferred by msp-300 mutations include defects in oocyte development and female fertility [22]. In humans, mutations in the Nesprin family are associated with several musculoskeletal disorders, including bipolar disorder, autosomal recessive cerebellar ataxia type 1 (ARCA1), X-linked Emery-Dreifuss muscular dystrophy (EDMD) and are risk factors for schizophrenia and autism [23].
Here, we show that flies carrying attP40 exhibits a nuclear clustering phenotype in larval muscle, which suggests that attP40 is an msp-300 insertional mutation. Further, we show that inserting transgenes into attP40 can increase severity of this phenotype. We use quantitative RT-PCR (Q-PCR) to show that insertions within attP40 decrease msp-300 transcript levels in 3 rd instar larvae. Finally, we show that chromosomes carrying certain TRiP insertions constructed from the Valium 20 vector [9] into attP40 confer recessive lethality or sterility. Because of the variable effects of different transgene insertions into attP40, investigators should use caution in interpreting data collected from Drosophila melanogaster carrying these insertions.

Drosophila melanogaster stocks
All fly stocks were maintained on standard cornmeal/agar Drosophila melanogaster media: 69.1 g/l corn syrup, 9.6 g/l soy flour, 16.7 g/l yeast, 5.6 g/l agar, 70.4 g/l cornmeal, 4.6 ml/l propionic acid and 3.3 gm/l nipagin. Flies carrying attP2 and attP40 lacking insertions were retrieved as white-eyed progeny from transgene insertions carried out at GenetiVision  [24,25]. All experiments were performed on Drosophila melanogaster that had been reared and maintained at room temperature (22˚C) with a 12h: 12h light:dark cycle unless otherwise indicated.

Immunocytochemistry
All larvae were dissected in HL3.1 [26] in a magnetic chamber, and fixed in 4% paraformaldehyde for 10 minutes, then washed in PBS with 0.3% Triton-X (PBS-T) and blocked for 30 minutes in PBS-T with 1% BSA. Samples were incubated overnight at 4˚C with primary antibody, washed thoroughly with PBS-T and then incubated for 2.5 hours at room temperature with secondary antibody. Samples were then washed with PBS-T and mounted in VectaShield Antifade Mounting Medium containing DAPI (Vector laboratories; H-1200-10).
Alexa Fluor1 647 phalloidin (1:200) was used to visualize F-actin. All images were acquired on a Zeiss LSM800 with an Airyscan confocal microscope.

Nuclear clustering analysis
Third instar larval body wall muscle 6 was chosen for all nuclear clustering analysis. Images were opened in ImageJ and nuclei clusters in muscle 6 were counted. We defined a "cluster" as two or more nuclei in which the distance between two nuclear borders was less than five microns. We analyzed six larvae from hemisegments A2-A4 (18 hemisegments total) for each genotype.
Microsoft excel was used to import all nuclear clustering data. "Normal" muscles contained no clusters. For muscles that contained cluster(s), we determined cluster number and nuclei number per cluster. Cluster sizes for each genotype were plotted on a frequency histogram. The percentage of normal hemisegments for each genotype was plotted on a column graph. For each hemisegment we calculated the percentage of nuclei in clusters using the following calculation: number of nuclei within clusters divided by total number of nuclei in muscle 6 multiplied by 100. Then, the average of three hemisegments per larva was calculated and plotted on a column graph. Data was visualized and analyzed using GraphPad Prism v9.3.1 or Synergy Software KaleidaGraph v4.5.2.

Quantitative RT-PCR (Q-PCR)
Primers were designed with PrimerBLAST software according to the published sequence of msp-300. We amplified and analyzed two msp-300 regions: first, the 3' end region, which amplifies transcripts RD, RG, RH, RI, RJ, RK, RL and RM (transcripts that contain the KASH domain), and second, the middle region, which amplifies transcripts RB and RF (B/F) (transcripts that lack the KASH domain) (Fig 1). For the 3' end region, we used the forward primer sequence 5'-TCAACCTCTTCCAATGCAGGC-3', and the reverse primer sequence 5'-CGC CAGAACCGTGGTATTGA-3'. The B/F forward primer sequence was 5'-CACGTACTTGCC GCACGAT-3', and B/F reverse primer sequence was 5'-ATTTTTGACACGTTCCCGGC-3'. For amplification of Rp49, chosen as the housekeeping gene, the forward primer sequence was 5'-TGTCCTTCCAGCTTCAAGATGACCATC-3' and the reverse primer sequence was 5'-CTTGGGCTTGCGCCATTTGTG-3'. Total RNA (500 ng) was extracted from frozen whole larvae and larval fillets with Direct-zol TM RNA MicroPrep (Zymo Research) according to manufacturer's protocol. The yield of RNA was estimated with the Nanodrop2000 (ThermoFisher). The A 260 :A 280 ratio was between 1.8-2.1. Superscript TM III First-Strand Synthesis System (ThermoFisher) was used to generate cDNA according to manufacturer's protocol. Reverse transcribed cDNA was then amplified in a 20 μl PCR reaction by the ABI Prism 7000 system (Applied Biosystems) with the universal conditions: 50˚C for 2 min, 95˚C for 10 min, and 40 cycles (15s at 95˚C, 1 min at 60˚C). Each sample contained 10 whole larvae. Three separate biological samples were collected from each genotype, and triplicate measures of each sample were conducted for amplification consistency. Data were analyzed with the relative 2 -ΔΔCt method to ensure consistency [27].

Pupal size measurements
We measured length and width in pupae homozygous for each of six TRiP lines and the attP40 parent chromosome. Pupal length was determined from top to bottom, excluding the anterior and posterior spiracles. Pupal width was determined by measurement at the midpoint along the pupal anterior/posterior axis.

Viability measurements
Each of six TRiP lines were placed in combination with balancer "CyO-TbA" [28], which carries the Tb 1 dominant "Tubby" marker and enabled us to unambiguously genotype balancercontaining from balancer-lacking larvae, pupae and adults. Each of these six TRiP lines were then brother-sister mated and F1 progeny reared in uncrowded vials. The number of Tubby and non-Tubby pupae were counted. Non-Tubby pupae were removed and placed into fresh vials. After seven days, the number of successful eclosions was counted. To monitor viability in the control attP40 line, pupae reared in uncrowded vials were collected and after seven days the number of successful eclosions was counted.

Fertility measurements
We measured male and female fertility in adults homozygous for each of six TRiP lines and the attP40 parent chromosome. To measure male fertility, single males were placed in vials with two phenotypically wildtype and fertile virgin females, and ability to generate larval progeny was measured. To measure female fertility, single females were placed in vials with two phenotypically wildtype and fertile males, and ability to generate larval progeny was measured.

Construction of Aop-atlRNAi
We chemically synthesized the reported sequence of the TRiP atl short hairpin with Xho1 and Xba1 sites added to the left and right ends, respectively. (TCGAGAGTCTGGTATAGGTCATT AGTTTAtagttatattcaagcataTAAACTAATGACCTATACCAGGCT-lower case letters indicate the loop sequence). This construct was introduced into the Xho1-Xba1 sites of pJFRC19-13XlexAop2 (Addgene, Inc) and introduced into embryos at the attP40 site using with ϕC31-mediated recombination (GenetiVision, Houston TX). Insert-containing flies were recognized by eye color.

Statistical analysis
For percentage of nuclei in clusters analysis, a One-way ANOVA with Tukey post hoc test was performed. For pupa length and width analysis, a One-way ANOVA with Bunnett post hoc test was performed. All statistical tests were performed in GraphPad Prism v9.3.1.

Results
The attP40 docking site is located within or upstream of msp-300, depending on the isoform attP2 and attP40 are two widely used attP docking sites in Drosophila melanogaster for insertions of LexA drivers, Gal4-driven TRiP insertions, and other constructs. Both docking sites provide high levels of induced transgene expression while maintaining low basal expression. As of February 2022, the Bloomington Drosophila Stock Center (Bloomington, IN) provides 16,503 stocks carrying attP40 and 14,941 stocks carrying attP2 (Annette Parks, personal communication). The attP40 chromosomal location is closest to msp-300, which expresses 11 different isoforms. attP40 lies within intron 1 for transcripts RH, RI, RJ and RK and upstream of transcripts RB, RD, RE, RF, RG, RL and RM ( Fig 1A). These results raise the possibility that attP40 could affect msp-300 transcript levels. Indeed, De et al. [12] reported that certain transgene insertions into attP40 alter the H3K27me3 epigenetic mark over at least a part of msp-300. Similarly, in a Ph.D. thesis, Cypranowska showed that attP40 decreased msp-300 transcription in certain genetic backgrounds and conferred synaptic phenotypes at the larval neuromuscular junction consistent with decreased msp-300 activity [29].
Given previous reports that msp-300 variants alter muscle myonuclear spacing [14,30], we hypothesized that effects of attP40 on msp-300 expression might alter larval muscle nuclear clustering. To address this hypothesis, we first defined a "cluster" as two or more nuclei in which the distance between two nuclear borders was less than 5 μm (Fig 1B). Then, for every nucleus in a larval body wall muscle, we measured the distance to its nearest neighbor, using larval body wall muscle 6 as assay platform. Using this approach, we observed abnormalities in nuclear positioning in larvae homozygous for attP40 vs. control larvae (homozygous for attP2) or larvae heterozygous for attP40 and attP2 (Fig 2A-2C). In particular, larvae homozygous for attP40 ( Fig 2D, middle panel) exhibited more nuclear clusters and clusters of greater size than control ( Fig 2D, top panel) or heterozygous larvae ( Fig 2D, bottom panel), and a significantly greater number of nuclei within clusters (Fig 2E). As a result, larvae homozygous for attP40 exhibited fewer hemisegments with a normal nuclear distribution than control or heterozygous larvae (Fig 2F). We conclude that attP40 causes a recessive muscle nuclear clustering phenotype.

Effects of inserting transgenes into attP40 on phenotypic severity
To determine if insertions into attP40 could affect nuclear clustering, we used the functionally neutral LexAop2-IVS-myr::GFP reporter introduced into attP40 and crossed these flies with attP2, attP40, lexA::Mef2 or lexA::nSyb. Representative images of nuclei clustering in body wall muscle 6 for each genotype is shown in Fig 3A-3D. First, we found that LexAop2-IVS-myr:: GFP/+ larvae (only one copy of attP40) showed only 6.18% of nuclei in clusters and a mostly normal nuclear distribution (Fig 3E top panel, 3F and 3G), indicating that LexAop2-IVS-myr:: Frequency distribution of the number of nuclei within each cluster within muscle 6 for each genotype shown in upper right. Lines in upper right depict the genotype of each homologue; green pentagons indicate the attP2 docking site, blue pentagon depicts attP40 docking site E) Percentage of nuclei found within clusters for each genotype. Means +/-SEMs are shown. The following calculation was used: ((number of nuclei within clusters / total number of nuclei) x 100). F) Percentage of hemisegments with normal nuclear distribution within muscle 6. attP2 is represented in green, attP40 in blue and attP40/+; attP2/+ in half green/half blue for panels D-F. For each genotype 3 hemisegments per larvae for 6 larvae total were analyzed. One way ANOVA with Tukey post-hoc test was used for statistical analyses. Scale bar = 50 μm.
https://doi.org/10.1371/journal.pone.0278598.g003 GFP, like "empty" attP40 (attP40 lacking an insertion), is recessive. However, when LexAo-p2-IVS-myr::GFP was in combination with empty attP40, we found extremely large nuclear clusters (containing up to 12 nuclei; Fig 3E, lower panel in blue), which were not observed in larvae homozygous for attP40 (Fig 2D). In addition, 46.4% of nuclei were in clusters compared to 36.7% in larvae homozygous for attP40 homozygous (Figs 3F and 2E). Thus, transgene insertion into one attP40 site increases nuclear clustering severity. To determine if inserting transgenes into both attP40 sites would affect nuclear clustering, we examined nuclear clustering in LexAop2-IVS-myr::GFP/LexA::nSyb larvae and found nuclear clustering similar to larvae homozygous for empty attP40 (Figs 3E bottom panel and 2D). These results suggest that inserting transgenes into a second attP40 site has little additional effect on nuclear clustering.
In the experiments described above, the insertions in attP40 were not transcribed in muscle. Given the previous report [12] that certain transgene insertions into attP40 could generate epigenetic marks that spread into msp-300, potentially altering msp-300 expression, we determined if expressing the insertions in muscle would affect nuclear clustering. Thus, we created LexAop2-IVS-myr::GFP/LexA::Mef2 larvae, in which the insertions in each attP40 site would be expressed in muscle. We found that these larvae exhibited a more severe nuclear clustering phenotype even than LexAop2-IVS-myr::GFP/empty attP40 (Fig 3E lower panel in cyan, 3F and 3G), and was the only genotype tested in which every muscle exhibited some nuclear clustering ( Fig 3G). These observations suggest that transcribing insertions in attP40 might further increase the severity of nuclear clustering phenotype.
To determine if Gal4-regulated insertions, like LexA-regulated insertions, would also induce nuclear clustering, we investigated properties of one Gal4-driven shRNA construct (targeting spatacsin (spat) (CG13531)), created by the transgenic RNAi project (TRiP) [9], as our example. We chose a TRiP insertion for this study because these insertions are widely used in the community and TRiP insertions represent more than half (8,884) of attP40 insertions maintained at the Drosophila stock center in Bloomington, IN. Representative images are shown in Fig 4A-4C. We found that TRiP spat/+;attP2/+ larvae exhibited wildtype nuclear positioning (Fig 4D, top panel, 4E and 4F), but both TRiP spat/empty attP40 or homozygous TRiP spat exhibited a strong nuclear clustering phenotype (Fig 4D middle and bottom panel, 4E and 4F)in fact, larvae homozygous for TRiP spat exhibited clusters up to 15 nuclei in size (Fig 4D). Thus, TRiP spat confers a recessive clustering phenotype, which indicates that this Gal4-driven transgene insertion behaves in a manner similar to the LexA insertions described above.

Effects of attP40 and derivatives on msp-300 transcript levels
We hypothesized that nuclei clustering in larvae homozygous for attP40 and derivatives reflected altered expression of at least some of the eleven msp-300 isoforms (Fig 1). To test this hypothesis, we prepared RNA from whole third instar larvae and performed quantitative RT-PCR (Q-PCR) using primers from two regions of the msp-300 transcription unit; first, the far 3' end of msp-300, which includes the KASH domain and accounts for eight of eleven annotated transcripts [16], see Fig 5A, and second, an internal region that is predicted to amplify the RB and RF transcripts, which lack the KASH domain. Thus, these primers enable analysis of 10 of 11 annotated transcripts (Fig 5A).
We found using the 3' end primers that msp-300 transcript levels were decreased about two-fold in larvae homozygous for attP40 and that contained an insertion in at least one of the attP40 sites (Fig 5B). These results support the possibility that attP40 derivatives cause nuclei clustering by decreasing msp-300 transcript levels. These effects on transcript levels, like effects on nuclear clustering, were recessive, as msp-300 transcript levels in LexAop2-IVS-myr::GFP/+ or TRiP spat/+ larvae were not distinguishable from those from the attP2 control.
We were surprised to find that larvae homozygous for empty attP40 exhibited wildtype msp-300 transcript levels, despite exhibiting a strong nuclear clustering phenotype. However we note that msp-300 is transcribed in larval fat bodies as well as muscle [17]; thus our whole larva RNA preparations might not capture attP40 transcriptional effects specifically in muscle. To address this possibility, we performed Q-PCR on RNA prepared from larval fillets, in which the fat body as well as all other internal organs were removed, leaving only the body wall muscles, underlying epidermis, and cuticle. We found that with this semi-purified muscle tissue as RNA source, empty attP40 fillets as well as LexAop2-IVS-myr::GFP/LexA::Mef2 fillets exhibited a~two-fold decrease in msp-300 transcript levels (Fig 5C), similar to what we observed in whole larvae carrying attP40 insertions (Fig 5B). Thus, muscle msp-300 transcript levels match muscle msp-300 phenotypes in attP40 and derivatives. These results also raise the possibility that empty attP40 might increase levels of certain msp-300 isoforms in non-muscle tissues.
The primers used in Fig 5B enable amplification of KASH domain isoforms. To analyze transcript levels of the two annotated non-KASH domain isoforms, transcripts RB and RF, we used the internal primers described above to amplify RNA from whole larvae. We found that

Effects of TRiP insertions into attP40 on viability and fertility
Many transgenes introduced into attP40 are RNAi short hairpin sequences from the "TRiP" project cloned into the Valium20 vector. We noticed some unexpected viability phenotypes, even in the absence of expression, when working with some of these insertions, so we wanted to characterize TRiP insertion viability systematically. However, monitoring TRiP insertions for recessive viability and fertility was problematic for two reasons. First, many TRiP insertions into attP40 are balanced with CyO; this is problematic because the Cy 1 "Curly" dominant marker on the CyO balancer is unreliable. The Curly wing phenotype is easily suppressed by genetic modifiers as well as the environmental conditions of low temperature and larval crowding [31]; indeed, the FlyBase allele report for Cy 1 states that Cy 1 "frequently overlaps To address these difficulties, we placed several TRiP insertions in combination with a modified CyO balancer upon which the dominant Tb 1 "Tubby" transgene had been introduced by P-element mediated transformation [28]. This CyO-TbA balancer also carries the dominant Star marker, which confers "Rough eyes" [32]. Both Tb 1 and Star are completely penetrant. Fig 5. Effect of attP40 and derivatives on msp-300 transcript levels. A) Schematic overview of the primer set used in this analysis (green arrowheads for mid region primers, red arrowheads for 3' primers). Grey inset enhances the mid to 3'end region of transcripts RB and RD. The (light color) inset shows the primer locations used to analyze transcripts RB and RF. The (dark color) inset show the primer locations used to analyze transcripts RD, RG, RH, RI, RJ, RK, RL and RM (KASH-containing transcripts). Annotated transcripts as shown in Fig 1. B) Quantitative RT-PCR (Q-PCR) was used to measure msp-300 KASH-containing transcript levels, normalized to attP2, from whole third instar larvae of the indicated genotypes. C) Q-PCR was used to measure msp-300 KASH-containing transcript levels, normalized to attP2, from 3 rd instar larval fillets of the indicated genotypes. D) Q-PCR was used to measure msp-300 transcripts B and F, normalized to attP2, from whole third instar larvae of the indicated genotypes. Red bars represent transcript levels measured with msp-300 KASH-containing primers and green bars represent transcript levels measured with transcripts B and F primers. Mean +/-SEM was determined from at least three separate biological samples collected from each genotype, and triplicate measures of each sample. The 2 -ΔΔCt was employed for this measurement. Each sample contained a mix of 10 whole larvae for (B) and (D) or a mix of 5 larval fillets (C).
https://doi.org/10.1371/journal.pone.0278598.g005 Thus, the use of CyO-TbA to balance TRiP insertions enabled us to unambiguously distinguish flies heterozygous from homozygous for each TRiP insertion at the larval, pupal, and adult stages. From six TRiP insertions tested when homozygous, we found a wide variety of viability and fertility deficits (Tables 1 and 2). The mostly strongly affected insertions, TRiP atl and TRiP spat, permitted no homozygous viable adults, although a few pupae homozygous for TRiP spat were observed. Flies bearing the TRiP rop and TRiP Mcu insertions were less strongly affected. Pupae homozygous for either insertion were observed, albeit at a lower frequency than expected, and most (~80%) failed to eclose. Escaper adults exhibited greatly decreased fertility ( Table 2). In particular, none of the males homozygous for TRiP rop, and only 2 of 10 males from TRiP Mcu, were fertile. Likewise, most females homozygous for TRiP rop and TRiP Mcu were infertile, and the rare fertile females produced very few progeny. Flies bearing the TRiP JNK insertion were even less strongly affected. Adults homozygous for TRiP JNK were plentiful (Table 1), and the females displayed wildtype fertility (Table 2). However, males homozygous for TRiP JNK appeared to be completely sterile ( Table 2). All phenotypes of flies homozygous for TRiP spartin appeared similar to wildtype. TRiP spartin was the only one of the six TRiP lines tested for which we were able to construct and maintain a homozygous stock.
Pupae homozygous for TRiP insertions showed body size defects as well as viability or fertility defects. We did not notice any visible defects in the pupa case and posterior/anterior spiracles in four of our TRiP lines (Fig 6A). Pupae homozygous for spat usually failed to evert their anterior spiracles (Fig 6B), a phenotype previously observed in mutants defective in ecdysone production [33] or larval muscle function [34]. Although pupal length in these homozygous pupae was not significantly different from the control (pupae homozygous for empty attP40), pupal width in these homozygous pupae was significantly decreased (Fig 6C and 6D). The TRiP Mcu insertion had the strongest effect on pupal width.
In an attempt to separate the recessive lethality of the TRiP atl chromosome from the attP40 insertion, we performed two rounds of free recombination between TRiP atl and another stock with no lethal mutations on chromosome II. However, we were unable to separate the lethality from the TRiP atl insertion, indicating that the lethal allele of TRiP atl is at or close to the insertion site.
We also tested the viability and fertility of the flies carrying the identical hairpin sequence found in TRiP atl but introduced into attP40 via a different vector. In particular, we introduced the TRiP atl hairpin sequence into pJFRC19-13XlexAop2, and then introduced this plasmid into attP40 with ϕC31-mediated recombination. We found that flies bearing this Aop- atlRNAi insertion were fully viable and fertile. Therefore, the phenotype conferred by the hairpin sequence from TRiP atl is affected by genetic context.

Dominance and recessiveness of attP40 and derivatives
Data shown in Figs 2-5 indicate that attP40 and derivatives are recessive for nuclear clustering and msp-300 transcription phenotypes. However, in certain genetic backgrounds, we have found that attP40 and derivatives can have dominant effects on clustering. In particular, we introduced the pan-neuronal Gal4 driver nSyb-Gal4 into flies heterozygous for either empty attP40, or the TRiP atl insertion into attP40. Note that nSyb-Gal4 is not inserted into attP40 and the stock does not contain attP40. We found that the nSyb-Gal4 background did not have a significant effect on larval muscle nuclear clustering in empty attP40/+ (Fig 7A, 7D top panel, 7E and 7F); however, TRiP atl/+ larvae carrying nSyb-Gal4 displayed nuclear clustering that was significantly increased compared to both empty attP40/+ and empty attP40/+; nSyb-Gal4 (Fig 7B, 7D middle panel, 7E and 7F). This phenotype is not a consequence of atl knockdown in neurons; when we used nSyb-Gal4 to drive expression of an atlRNAi that was not inserted into attP40, we observed no increase in nuclear clustering (Fig 7C, 7D bottom panel, 7E and 7F). Therefore, attP40 derivatives can confer dominant effects in certain genetic backgrounds. In addition, these results support the conclusion that the TRiP atl insertion into attP40 confers a stronger MSP-300 phenotype than empty attP40.

Effects of the attP40 docking site and derivatives on msp-300-mediated phenotypes
The attP40 docking site for Drosophila melanogaster transgene integration is widely used for ϕC31-mediated transgene integration. attP40 lies within the transcription unit of msp-300, which encodes the Drosophila melanogaster orthologue of Nesprin-1; this observation raises the possibility that attP40 is an msp-300 insertional mutation. Here we show that attP40, either alone or containing any of several specific transgene insertions, causes phenotypes similar to msp-300 mutations, such as larval muscle nuclear clustering and viability deficits. In addition, . F) Percentage of hemisegments with normal nuclear distribution within muscle 6. Each genotype is represented with a different color in panels D-F with attP40/+; attP2/+ (grey), attP40/+; nSyb-Gal4/+ (blue), TRiP atl/+; nSyb-Gal4/+ (red) and nSyb>Dcr, atlRNAi (green). For each genotype 3 hemisegments per larvae for 6 larvae total were analyzed. ANOVA with Tukey was used for statistical analyses. Scale bar = 50 μm. https://doi.org/10.1371/journal.pone.0278598.g007 attP40-containing larvae exhibit an approximately two-fold decrease in msp-300 transcript levels of certain isoforms. The effect on nuclear clustering and viability varies depending on the precise transgene introduced into attP40. We conclude that attP40 is an insertional mutation for msp-300.

Additional msp-300-dependent and independent phenotypes conferred by attP40
Reports from other investigators are now appearing in which phenotypic consequences of attP40-bearing flies are described. In a Ph.D. thesis, Cypranowska (2020) reported that attP40 decreases msp-300 transcription and also confers abnormal synapse function at the larval neuromuscular junction [29]. Many of these phenotypes are similar to those conferred by loss of the muscle glutamate receptor gluRIIA; this observation is significant because it was previously reported that gluRIIA localization is affected by msp-300 [21]. These observations raise the possibility that attP40 regulates synaptic transmission via msp-300-dependent effects on gluRIIA.
Other investigators have reported other attP40-dependent phenotypes. For example, Groen et al. reported that attP40 decreases transcription of ND-13A, the gene immediately centromere-distal from attP40 and which encodes a component of mitochondrial complex I [35]. Further, they suggest that this decreased ND-13A transcription might mediate the resistance to the chemotherapy agent cisplatin observed in attP40 flies [35]. More recently, Duan et al. reported that attP40 alters neuronal architecture of the olfactory glomerulus, a phenotype that appears to be msp-300-independent [36]. Taken together, these results indicate that attP40 can confer a variety of phenotypes in a variety of tissues by affecting transcription of at least two genes.

attP40 phenotypes can be dominant or recessive depending on genetic background
We tested if attP40 and derivatives containing insertions into attP40 were dominant or recessive for the nuclear clustering and msp-300 transcription phenotypes. We found that in most cases, the effects of attP40 and derivatives were recessive. For example, the neutral reporter LexA-myr-GFP or the TRiP spat insertion within attP40 each conferred severe nuclear clustering and decreased msp-300 transcription when homozygous or in combination with empty attP40, but not when attP40 was absent, from the other homologue. These observations indicate that attP40-dependent phenotypes are recessive. However, in other genetic backgrounds, we detected dominant effects of attP40 derivatives. In particular, the TRiP atl insertion within attP40 conferred a dominant nuclear clustering phenotype in a genetic background containing the neuronal Gal4 driver nSyb-Gal4. Similarly, Cypranowska reported that attP40 conferred a dominant 2.7-fold decrease in msp-300 transcript levels in the presence of the motor neuron Gal4 driver OK6 [29]. We conclude that the phenotypes of attP40 and derivatives can be dominant or recessive depending on genetic background.

Effects of specific transgene insertions into attP40 on mutant phenotypes
We tested if the nucleotide sequence of specific transgenes inserted into attP40 would affect mutant phenotypes. First, we found that transgenes from either the LexA or the Gal4 regulatory systems were capable of enhancing the nuclear positioning phenotype and msp-300 transcript phenotypes to moderate degrees. Second, we tested if transcription of inserted transgenes in muscle would affect mutant phenotypes. In particular, we compared nuclear positioning in larvae expressing LexA-myr-GFP in neurons vs. muscle and found that muscle expression modestly increased severity of the nuclear clustering phenotype but did not affect msp-300 transcript levels. These results indicate that muscle transcription of inserted transgenes is not necessary for mutant phenotype, but we are unable to rule out the possibility muscle transcription could contribute to severity of mutant phenotype.

TRiP insertions into attP40: extremely variable degrees of recessive lethality and sterility
Of the 16,503 lines carrying attP40 maintained at the Drosophila stock center (Bloomington, IN), 8,884 contain TRiP (short hairpin sequences for RNAi) insertions, mostly in the Valium 20 vector. In studies of six of these lines, we observed a range of recessive phenotypes, from full lethality to full (or partial) sterility to~wildtype. This extreme variability in phenotypes among the TRiP lines was unexpected, as each lines contains inserts with identical sequences, with the exception of the sequence of the short hairpin itself. These results suggest either that phenotypic strength is affected unexpectedly strongly by the precise nucleotide sequence of the hairpin, or alternatively, that these TRiP lines have accumulated genetic modifiers at an unusually high rate, and that these modifiers have variable effects on phenotype. These two possibilities are not mutually exclusive.
Because of the wide variety of phenotypes exhibited by flies carrying distinct TRiP insertions, we wondered if we could predict phenotypes conferred by specific TRiP lines from information presented at the Bloomington Drosophila Stock Center (BDSC, https://bdsc. indiana.edu). We noticed that in the description of the six TRiP lines tested, five (TRiP atl, TRiP spat, TRiP rop, TRiP Mcu, and TRiP JNK) contained the phrase "May be segregating CyO", or a related phrase. These five lines are the same ones for which we were unable to maintain a homozygous line. The sixth line (TRiP spartin), for which we were able to maintain a homozygous line, lacked this phrase. This correspondence raises the possibility that researchers might be able to distinguish homozygous viable and fertile TRiP lines from others by the absence/presence of this phrase. As of the most recent report (from 2019), the BDSC reports that 31.46% of the 8884 TRiP lines in attP40 (2795 total) are listed as homozygous, with the rest as "CyO", "CyO fl" or "CyO mix".

Conclusions
We have shown attP40 and derivatives containing insertions confer msp-300 mutant phenotypes and can decrease msp-300 transcript levels. Regardless of the mechanism underlying the variety of phenotypes conferred by various TRiP insertions within attP40, investigators should be aware that these insertions might confer phenotypes that are difficult to predict and might manifest in a variety of ways. Going forward, investigators should use caution when interpreting data resulting from flies containing attP40, especially in muscle tissues.