An Interscholastic Network To Generate LexA Enhancer Trap Lines in Drosophila.

Binary expression systems like the LexA-LexAop system provide a powerful experimental tool kit to study gene and tissue function in developmental biology, neurobiology, and physiology. However, the number of well-defined LexA enhancer trap insertions remains limited. In this study, we present the molecular characterization and initial tissue expression analysis of nearly 100 novel StanEx LexA enhancer traps, derived from the StanEx1 index line. This includes 76 insertions into novel, distinct gene loci not previously associated with enhancer traps or targeted LexA constructs. Additionally, our studies revealed evidence for selective transposase-dependent replacement of a previously-undetected KP element on chromosome III within the StanEx1 genetic background during hybrid dysgenesis, suggesting a molecular basis for the over-representation of LexA insertions at the NK7.1 locus in our screen. Production and characterization of novel fly lines were performed by students and teachers in experiment-based genetics classes within a geographically diverse network of public and independent high schools. Thus, unique partnerships between secondary schools and university-based programs have produced and characterized novel genetic and molecular resources in Drosophila for open-source distribution, and provide paradigms for development of science education through experience-based pedagogy.

Binary gene expression systems are an important foundation for investigating and manipulating Drosophila gene expression with temporal and cellular specificity. Generation of a yeast GAL4-based transactivator to induce expression of target genes fused to GAL4-responsive upstream activating sequences (UAS) has established a widely-used binary gene expression system in Drosophila (Brand & Perrimon 1993; Hayashi ; Gohl ). Common examples of GAL4-based gene fusions include transgenes with relatively short enhancer elements that direct GAL4 expression, and endogenous enhancer-directed GAL4 expression following random genome insertion by transposons encoding GAL4 (‘enhancer trapping’; O’Kane & Gehring 1987).

Studies of many biological problems benefit from simultaneous manipulation of two or more independent cell populations or genes (Rajan & Perrimon 2011). In prior studies, parallel use of two binary expression systems allowed insightful clonal analysis of multiple cell populations (Lai & Lee 2006; Bosch ), powerful studies of epistasis between different tissues (Yagi ; Shim ), and discovery of specific cell-cell contacts (Gordon & Scott 2009; Bosch ; Macpherson ). This multiplex approach requires an additional expression system that functions independently of the UAS-Gal4 system, such as the LexA system derived from a bacterial DNA binding domain (Szüts & Bienz 2000; Lai & Lee 2006; Pfeiffer ; Knapp ; Gnerer ). In that system, fusion of the LexA DNA binding domain to a transactivator domain generates a protein that regulates expression of transgenes linked to a LexA operator-promoter (LexAop). Unfortunately, the number and quality of fly lines expressing a LexA transactivator remains small compared to the thousands of comparable GAL4-based lines. From a scholastic network including Stanford University and secondary school science classes in New Hampshire, New York, New Jersey and Illinois we generated novel LexA-based enhancer trap drivers for fly biology. The addition of multiple secondary schools producing novel LexA-based enhancer traps represents a significant expansion of the prior effort piloting this approach (Kockel ).

Material and Methods

Immuno-histochemistry (IHC)

L3 larvae from cross of StanEx with a line harboring a LexA operator- GFP reporter transgene (LexAop2-CD8::GFP; Pfeiffer ) were dissected in PBS and fixed in 4% Formaldehyde/PBS for 30 min, permeabilized in 0.2% Triton X-100/PBS for 4 hr, and blocked in 3% BSA Fraction V/PBS for 1 hr. Incubation of primary and secondary antibodies were O/N in 3%BSA/PBS at 4° Celsius using a platform rocker. All specimens were rinsed (1 min) and washed (20 min) three times with PBS after antibody incubations. Primary Antibody: Goat anti-GFP 1:3000 (Rockland 600-101-215). Secondary antibody: Donkey anti-Goat Alexa488 (Life Technologies, A11055). All samples were mounted in SlowFade Gold mounting medium with DAPI (Life Technologies, S36938).

Microscopy

Microscopy was performed on a Zeiss AxioImager M2 with Zeiss filter sets 49 (DAPI) and 38HE (Alexa488) using the extended focus function.

Fly husbandry and fly strains

A standard cornmeal-molasses diet was used to maintain all fly strains (http://flystocks.bio.indiana.edu/Fly_Work/media-recipes/molassesfood.htm). The following strains were used: w[*]; ry[506] Sb[1] P{ry[+t7.2]=Δ2-3}99B/TM2, Ubx (Bloomington 1798), w[*]; P{y[+t7.7] w[+mC]=26XLexAop2-mCD8::GFP}attP2 (Bloomington 32207), y[1],w[1118]; P{w[mC]=LHG]StanEx[1]} (Bloomington 66673), w[*]; L[*]/CyO; ftz[*] e[*]/TM6,Tb[*].

Hybrid dysgenesis

F0: Females of donor stock y w StanEx[1] were mated to males w[*]; ry[506] Sb[1] Δ2-3/TM2,Ubx. F1: y w StanEx[1]; ry[506] Sb[1] Δ2-3/+ males were crossed to w[*]; L[*]/CyO; ftz[*] e[*]/TM6,Tb Hu females. F2: w+ males were mated to w[*]; L[*]/CyO; ftz[*] e[*]/TM6,Tb Hu. F3: The insertion line was stably balanced deploying a brother-sister cross.

This P-element vector also enables subsequent recombinase-mediated cassette exchange (RMCE; Gohl ).

Insertion site cloning

We followed an inverse PCR approach (Kockel , http://www.fruitfly.org/about/methods/inverse.pcr.html), to molecularly clone the insertion sites of StanEx P-elements. The molecular insertion coordinate is defined as the first nucleotide 3′ to the insertion site on the genomic scaffold, independent of the direction of the P-element insertion. Inverse PCR fragments were mapped to the genome of Drosophila melanogaster, Release 6 (Hoskins ). DNA restriction enzymes used: Sau3AI (NEB R0169) and HpaII, (NEB R0171). Ligase used: T4 DNA Ligase (NEB M0202). All primer sequences are displayed 5′->3′. 5′ end cloning: Inverse PCR primer “Plac1” CAC CCA AGG CTC TGC TCC CAC AAT and “Plac4” ACT GTG CGT TAG GTC CTG TTC ATT GTT. 3′ end cloning: Primer pair “Kurt” TGT CCG TGG GGT TTG AAT TAA C and “Ulf” AAT ACT ATT CCT TTC ACT CGC ACT. Sequencing primer 5′ end: “Sp1” ACA CAA CCT TTC CTC TCA ACA. Sequencing primer 3′ end: “Ulf” AAT ACT ATT CCT TTC ACT CGC ACT or “Berta” AAG TGG ATG TCT CTT GCC GA. For insertions where the sequence of one end only could be determined by inverse PCR, we pursued a gene-specific PCR approach (Ballinger and Benzer 1989) using P-element and gene-specific primers. 5′ end specific P-element primer “Chris”: GCA CAC AAC CTT TCC TCT CAA C, sequencing primer 5′ end: “Sp1”. 3′ end specific P-element primer “Dove”: CCA CGG ACA TGC TAA GGG TTA A, sequencing primer 3′ end: “Dove”. Sequence of gene-specific primers are available upon request.

Generation of Sequence Logos and position frequency matrices (PFMs)

The generation of sequence logos was performed as described (Crooks ) using the web tool http://weblogo.threeplusone.com/. The input sequence motif data are listed in Suppl. Table 2. The insertion site sequence is displayed and utilized with the genomic scaffold co-directionally oriented to inserted P-element (Linheiro and Bergman 2008). If P-elements are inserted 5′->3′, the strand of insertion was called +, and unprocessed genomic scaffold sequences were used to extract the insertion site sequences. If P-elements are inserted 3′->5′, the strand of insertion is called -, and the reverse complement of the genomic scaffold sequences were used to extract these insertion site sequences.

Position frequency matrices were generated using data from (Linheiro and Bergman 2008), (Gohl ) and (Kockel ). For GT P-element PFMs, quality issues of the insertion site data were noted (Linheiro and Bergman 2008), and only the insertions on the + strand were used (Linheiro and Bergman 2008). As a result of only 12 InSITE P-elements insertion sites sufficiently mapped (Gohl ), the PFM was not testable (0% C and 0% G in position 1; 0% G at position 13). All data are displayed in Suppl. Table 3. Chi-square testing was performed in MS Excel (2007) using the CHITEST function.

Sequencing the genomic site in NK7.1 / Heatr2 around hot spot at 3R:14,356,561

The genomic site was amplified in 5 fragments of overlapping segments of approx. 500bp called A, B, C, D, and E Fragments. C Fragment contained the hot spot 3R:14,356,561. Co-ordinates A-Fragment: 3R:14,355,462 - 3R:14,355,962. Primer sequences for A-Fragment amplification: NK7.1-GS_A.FOR: AGTGGAAACGAGCGAAGCTG; NK7.1-GS_A.REV: AAAGGTCAAATGTGATGCAGCGAG. Co-ordinates B-Fragment 3R:14,355,862 - 3R:14,356,362. Primer sequences for B-Fragment amplification: NK7.1-GS_B.FOR: TCTGTGCAGATAGGAAATTACTCATT and NK7.1-GS_B.REV: CTCTTGCCACTTTCTGTGAGCTT. Co-ordinates C-Fragment: 3R:14,356,190 - 3R:14,356,709. Primer sequences for C-Fragment amplification: NK7.1-GS_C.FOR2: CTGGGCCAGTCAAGTGTGTA and NK7.1-GS_C.REV2: AGAGCTACGAACCTGGCC. Co-ordinates D-Fragment: 3R:14,356,662 - 3R:14,357,162. Primer sequences for D-Fragment amplification: NK7.1-GS_D.FOR: GCGATGAGGATGAAGTTGTCGG and NK7.1-GS_D.REV: GACTCTCTTCATCGCCAGCC. Co-ordinates E-Fragment: 3R:14,357,124 - 3R:14,357,638. Primer sequences for E-Fragment amplification: NK7.1-GS_E.FOR: CCTGGCCATAGAGATCCAAG and NK7.1-GS_E.REV: TGCGAAGCTGCAAAGTAAAA. According to the published genomic sequence, the expected size of the C-Fragment is 520 bp. PCR in genomic StanEx1 DNA amplified a DNA fragment of 1682bp, consisting of C-Fragment sequence containing KP element sequence.

All primer and genomic reference sequences are also deposited in Suppl. Table 4, Workbook “Genomic Seq Flybase”.

Sequencing KP element at 88B4-6

The KP element at 88B4-6 of the StanEx1 strain was amplified using the C-Fragment primer (see above) NK7.1-GS_C.FOR2: CTGGGCCAGTCAAGTGTGTA and NK7.1-GS_C.REV2: AGAGCTACGAACCTGGCC. Due to the A:T-rich region and the resulting A:T stutter during sequencing 5′ of the insertion site, all sequencing was performed on the lagging strand, 3′->5′ relative to the genomic scaffold in three replicates. Sequencing primers: NK7.1-GS_C.REV2 (see above), NK7.1-KP_BPS_REV1: TAGGTACGGCATCTGCGTTG, KP_BPS_REV2: CAGCCTTCCACTGCGAATCA KP_BPS_REV3: CAAGGCTCTGCTCCCACAAT. All individual sequence and primer data are shown in Suppl. Table 4, Workbooks: “Seq StanEx KP Element” and “Primer Sequences”.

Testing Strains for KP element insertions

Genomic DNA was extracted using standard methods and presence of KP element(s) was tested using Primers NK7.1-KP_3_REV2, ATGCCCAGGATGAATTGAAA and NK7.1-KP_3_FOR2, TCAGATGTGGAAACGTCGAT. The following strains were tested for the presence of KP element(s) in their genome: Oregon R (Bloomington Stock #5); bw (Bloomington Stock #245); cn (Bloomington Stock #263); y (Bloomington Stock #1495); w (Bloomington Stock #2390); w (Bloomington Stock #5905); w iso II, iso III (Bloomington Stock #6326); Oregon-R-SNPiso3 (Bloomington Stock #6363); Canton-S-SNPiso3 (Bloomington Stock #6366); Canton-S (Bloomington Stock #64349). Only w (Bloomington Stock #2390) was found KP positive.

GenBank accession for KP element

The GenBank accession number for the KP element sequence at 88B4-6 characterized in this study is MK510925. The 5′->3′ annotated KP element sequence is also displayed in Suppl. Table 4, Workbook “Reconstituted StanEx-KP seq 5-3”.

Probability calculation of StanEx P-element insertion site hot spot at 3R:14,356,561, 88B4-6

We calculated the probability that a single genomic P-element insertion site would be selected at least 9 times in a series of 188 (total number of StanEx insertions generated so far) random insertions. The number of confirmed and non-identical individual P-element insertion sites present on the autosomes of the genome of Drosophila melanogaster was determined conservatively by counting confirmed and unique EPgy2, GT1, SUPor-P, GawB(+) and XP insertions (Linheiro and Bergman 2008). Hence, we tested the null hypothesis that the transposable element is equally likely to insert itself at any of 8,161 target sites. Mathematically, , where is given by a multinomial distribution with all event probabilities equal to and 188 trials. Since the computation of the exact p-Value is computationally expensive, we approximated the p-Value by using the simplification that . The value of can be calculated using a binomial distribution with success probability and 188 trials. Thus, we calculated the value of using the R command: 8161 * pbinom(q = 8, size = 188, P = 1/8161, lower.tail = FALSE).

The probability of obtaining nine insertions into an identical genomic site by chance is estimated to be small (P < 3.32 × 10−17). For the purpose of comparison, the chances to win the California Power Ball Lottery in October 2018 were 2.9 × 10−8. To corroborate this result empirically, we ran a computer simulation of the stochastic process in which a transposable element was inserted into an array of 8,161 equally probable target sites 188 times. After 100,000 runs of the simulation, we did not detect choice of the same insertion site 9 times.

Secondary school class descriptions

We formed partnerships between Stanford University investigators and classes at 4 high schools in the U.S. The class at Phillips Exeter Academy (NH) was taught as two 11-week courses “BIO586” and “BIO670” in the winter and spring terms. One year classes called “Research 11/12” ran at Commack High School (NY), “BIO670” at Pritzker College Prep (Chicago, IL), and “Sci574” 2/3 class + 1/3 elective extension at The Lawrenceville School (NJ). Pre-requisites for admission to these class were determined by individual schools, and included no pre-requisites (Commack High School and Pritzker College Prep), advanced placement (AP) biology or one term of a genetics elective (Phillips Exeter Academy, NH), or one term of advanced science (The Lawrenceville School). Partnering secondary school science teachers and their schools were self-selecting and initiated contact independently with S.K.K. to inquire about joining this network of experiment-based science instruction. Teachers trained in the S.K. Kim laboratory in Developmental Biology at Stanford, typically for two week-long sessions. The first week was dedicated to acquiring practical skills in fly biology and breeding, transmission genetics, larval dissection, transposable elements and hybrid dysgenesis. The second week focused on training in molecular methods, including DNA isolation, inverse PCR, cloning and sequence analysis. Instructors learned immuno-histochemistry and microscopy with an epi-fluorescent compound microscope in additional visits to Stanford. Students were instructed in fluorescence immunomicroscopy in summertime research internships in the Kim lab (see below), and during on-site visits of Stanford instructors at collaborating secondary schools.

Bio670 / Res 11/12 / Sci574 were scheduled for three to four 50 min periods / week, and up to 5-6 unscheduled hours per week. The instruction manuals, additional manuals for teachers, schedules and related problem sets are available on request.

At Exeter, students spent about nine weeks executing the hybrid dysgenesis crosses including mapping and balancing of novel strains. Further characterization of the insertion site was performed by polymerase-chain reaction and DNA sequencing using standard genomic DNA recovery. Crosses with reporter strains (lexAop2-CD8::GFP) were performed during the final 3 weeks of class, permitting instruction in larval dissection and microscopy to document tissue expression patterns of candidate enhancer traps. This schedule was modified to fit the year-long schedule at other secondary schools. Based on class performance and teachers’ evaluations, selected students were invited to continue studies in the Kim group at Stanford University School of Medicine during summer internships lasting about 6 weeks. These studies included further molecular mapping of transposon insertion sites, and verification of tissue patterns of enhancer trap expression. Students returning in the fall term helped instructors to run the subsequent iteration of Bio670, and also pursued independent projects.

Data and reagent availability

All StanEx derivatives and associated data are available at the Bloomington stock center. All molecular and image data are additionally available at http://stanex.stanford.edu/about/. Supplemental material available at FigShare: https://doi.org/10.25387/g3.8056007.

Results

Generating novel LexA-based enhancer trap strains

To generate LexA-based enhancer trap fly lines, we mobilized the P-element in a previously characterized StanEx line (Kockel ). The StanEx strain contains a single X-linked derivative of the InSITE P-element (Gohl ) harboring a weak P-promoter driven cDNA encoding a LexA DNA-binding domain fused to the hinge-transactivation domain of Gal4 (LexA::HG, Yagi ). We mobilized this X-linked StanEx-element to the autosomes using transposase Δ2-3 at 99B (Robertson ), to generate LexA P-element enhancer trap lines using standard methods (Methods; Suppl. Table 1; O’Kane and Gehring 1987). Our goal was to permit interaction of the weak promoter in the mobilized StanEx P-element with the local enhancer environment of the insertion site, and thereby generate spatial and temporal expression specificity of each LexA::HG insertion (O’Kane and Gehring 1987).

Mapping StanEx P-element insertion sites

We next used established inverse PCR-based molecular methods to map the chromosomal insertion position of the StanEx-elements to the molecular coordinates of the genomic scaffold (Figure 1; http://stanex.stanford.edu/about/). The 93 novel insertions of this study were equally distributed across autosomes II and III, and their chromosomal arms (2L, 24 insertions: 2R, 24 insertions: 3L, 20 insertions: 3R, 25 insertions). In this collection, we included 2 lines (SE133 and SE174) that inserted into repetitive DNA, and whose insertion site we could not map molecularly. We excluded 7 lines that were inserted into the identical location in /Heatr2 at 3R:14,356,561, as this precise location was tagged by prior StanEx insertions (see below, lines RJ-3 and EH-4; Kockel ). As observed previously (Bellen ), the majority of molecularly mapped novel insertions (81/91 or 89%) integrated into 5′ gene elements, including promoters, and first exons or introns. Of the 93 novel insertions presented here, we observed an even distribution of insertional direction by the StanEx P-element into genomic DNA. Using the 5′ and 3′ end of the P-element as coordinates, we found 46/93 insertions were oriented 5′->3′, and 44/93 insertions were oriented 3′->5′. In three cases we were unable to determine the direction of P-element insertion. In two of these three cases, the StanEx-element inserted into repetitive DNA (see above). In one of these three cases (SE444) the StanEx P-element inserted into the Hsp70 locus at 87A2 that consists of a 1:1 mirror-image arrangement of two nearly-identical copies of the hsp70 promoter and coding region (Hsp70Aa, FBgn0013275 and Hsp70Ab, FBgn0013276); sequence analysis of 5′ and 3′ adjacent genomic DNA in this line failed to resolve unambiguously the orientation of the SE444 P-element (Suppl. Table 1).

Figure 1

Insertion sites and genes tagged by LexA::HG enhancer traps on autosomes II and III. Associated molecular data are detailed in Suppl. Table 1.

We then analyzed the number of novel loci tagged by our StanEx LexA enhancer trap. To search for genes tagged by multiple constructs or insertions, we surveyed publically-available lines cataloged in FlyBase, including previous StanEx releases, and lines generated by the FlyLight project (Pfeiffer ). 76 of 91 mapped lines presented here mark novel loci, among them eiger (SE315), Star (SE316), (SE433), kismet (SE306 and SE435), (SE245), and branchless (SE 119 and SE232). Of note, we obtained insertions into several long non-coding RNA loci (lncRNA:CR43626, SE250 and SE312; lncRNA:CR43651, SE307; lncRNA:CR44120, SE425; lncRNA:CR44206, SE427; lncRNA:CR44320, SE336; lncRNA:CR45433, SE214; and lncRNA:CR46006, SE400). Apart from transcriptional mapping (Graveley ), these non-coding RNAs have not been functionally characterized.

Of the fifteen genes previously tagged by LexA, four were tagged in FlyLight lines (meng-po, SE441, bunched, SE226, G-coupled Receptor Kinase 2, SE216, pointed, SE118), ten were identified by prior StanEx insertions (RapGap1 SE422, split ends SE407, SE312 and SE250, bicoid interacting protein 3 SE234, α-Esterase-10 SE208, Hsrω SE140, bacchus SE133, escargot SE100, and the , vasa, locus SE102), and one was tagged in both FlyLight and StanEx studies (SE120, spire; see Suppl. Table 1). In summary, our approach generated multiple novel LexA-based autosomal enhancer traps.

Selected tissue expression of LexA in the StanEx collection

To verify the use of the StanEx P-element for enhancer trapping, we intercrossed this line with flies harboring a transgene encoding a LexA operator linked to a cDNA encoding a membrane-GFP reporter (lexAop2-CD8::GFP; Pfeiffer ) and confirmed membrane-associated GFP expression in several tissues, including larval and adult tissues (data not shown). Next we used this strategy to assess the tissue expression patterns of novel insertion lines. 3rd instar larvae of bi-transgenic offspring were analyzed by immuno-histochemical (IHC) staining for GFP expression, and simultaneous counter-staining for cell nuclei (DAPI). Image data from selected LexA enhancer trap lines were collected and tissue expression catalogued (Suppl. Table 1). Within the collection, we detected GFP expression in nearly all tissues of the L3 larva, including a variety of neuronal cell types in the Central Nervous System (CNS), Ventral Nerve Cord (VNC: Figure 2) and Peripheral Nervous System (PNS), imaginal discs, and a wide range of other somatic tissues like fat body, malphigian tubules and trachea (Suppl. Table 1). In the cases of lexAop-CD8::GFP expression directed by LexA from an insertion in the solo/vasa/vig locus (SE102), (SE204), a-Esterase 10 (SE208), OatP30B (SE218), (SE229) and cornetto (SE134), we observed distinct patterns of cell labeling in the CNS, VNC and ring gland (Figure 2A-F). In Diap1-LexA; lexAop-CD8::GFP larva we noted strong staining of neurons in the pars intercerebralis of the CNS, corresponding to neuroendocrine insulin-producing cells (IPCs: Figure 2B).

Figure 2

Expression pattern of six selected StanEx enhancer trap insertions in central nervous system (CNS), ventral nerve chord (VNC) and ring gland (RG) complexes of third instar larvae. Enhancer traps of (a) SE102 (insertion in solo/vasa/vig), (b) SE204 (insertion in Diap1), (c) SE208 (insertion in α-Est-10), (d) SE218 (insertion in oatp30B), (e) SE229 (insertion in NK7.1) and (f) SE134 (insertion in corn) crossed to lexAop-CD8::GFP are shown. Green: anti-GFP, Blue: DAPI. Scale bar = 100 μm.

To facilitate accessibility of the molecular and image data (Suppl. Table 1), we uploaded these to the searchable StanEx website (http://stanex.stanford.edu/about/; Kockel ), a database searchable by expression pattern, cytology and specific genes. This includes supplementary image analysis, data from immunostaining and molecular features of StanEx insertion loci, and is freely accessible to the scientific community.

An unrecognized KP Element in the StanEx1 line

During the generation of 188 individual StanEx enhancer trap P-element lines (from this work and Kockel ), we observed nine independent insertions into the Nk7.1 / Heatr2 locus at 88B4-6 (Figure 3). Molecular characterization of these nine insertions revealed 3R:14,356,561 as the common insertion coordinate. This particular insertion hot spot was not reported in prior hybrid dysgenesis efforts, which included use of a variety of distinct P-element constructs (Bellen , Linheiro and Bergman 2008). The probability of obtaining 9/188 insertions into an identical genomic site by chance is small (estimated P < 3.32 × 10−17: see Methods).

Figure 3

Properties of the StanEx enhancer trap P-element. (A) Position frequency matrix (PFM) of the 14 bp P-element insertion site (see text for details, Suppl. Table 2). Base position 1-14 in 5′->3′ direction on X-axis. The 8 bp sequence that will give rise to the directed repeat after insertion is highlighted in orange. Base composition (in %) on Y-axis as indicated. (B) Sequence logo derived from (A), X-axis as in (A), Y-axis in bits (Crooks ). (C) Insertion site of 9 StanEx lines within the locus of NK7.1/Heatr2 at 3R:14,356,562 88B4-6 on chromosome III. Transcription site start arrows mark alternative first exons of NK7.1. SE-number identifiers of StanEx enhancer traps and the year of their derivation are shown along their insertion direction (Suppl. Table 4).

The presence of this observed P-element insertion hotspot suggested a non-classical mechanism of targeting or insertion of the StanEx element. Preferential target site selection of P-elements based on the presence of DNA homology between the P-element and its target site has been described (Taillebourg and Dura 1999). To investigate a possible homology-based mechanism targeting the insertion into the hotspot at 88B4-6, we searched for apparent DNA homology between the StanEx enhancer trap P-element and the published D. melanogaster genomic sequences present in Release 6 of Flybase (Hoskins ) at that site, and did not detect significant homologies. Hence, we explored the possibility of an insertion site bias inherent to our StanEx P-element construct. We generated a position frequency matrix (PFM, Figure 3A) and sequence logo of the 14 bp P-element insertion site motif (Majumdar and Rio 2014) using 128 unique StanEx insertions that were characterized by insertion site sequencing with single nucleotide precision on the 5′ and 3′ ends (Crooks ; Figure 3B). This StanEx PFM was individually compared to PFMs of the P elements EPgy2 (n = 3112), GT1 (n = 465), SUPor-P (n = 2009), GawB(+) (n = 1072) and XP (n = 4126) (Suppl. Table 3; Linheiro and Bergman 2008). We also compared the StanEx PFM to a ‘summary’ consensus motif based on available P-element insertion site data (Suppl. Table 3). Using χ2 testing, we were unable to detect a significant difference of the StanEx insertion site motif and any of the other individual or agglomerated P-element insertion site motifs (Suppl. Table 3). We conclude that StanEx and other P-elements have similar insertion site preferences.

Thus, we assessed the possibility that the specific 3R:14,356,561 target site harbored by the StanEx starter strain might contain sequences not annotated in the published D. melanogaster genome (Version 6), and that might contribute to the increased targeting frequency at this locus. Sequencing the StanEx-element recipient hotspot at 3R:14,356,561 in our StanEx donor strain, prior to hybrid dysgenesis, revealed the presence of an unrecognized 1.1 kb KP element at 3R:14,356,561 (Figure 3C, Suppl. Table 4, GenBank Accession MK510925). This KP element was flanked by the 8 bp target site direct duplication GCCCAACC. A KP element is a non-autonomous P-element with intact inverted repeats, whose transposase-encoding exons 2-4 contain deletions. These deletions produce an (ORF) frame that encodes a type II repressor instead of functional transposase (Black ; Rio 1990; Majumdar and Rio 2014; Kelleher 2016), permitting KP element mobilization only when transposase is provided in trans. Thus, an interaction of a KP-element with the StanEx P-element in the StanEx starter strain could underlie the observed repeated integration of StanEx P-element into 88B4-6.

To investigate whether the KP element was deleted upon StanEx insertion at 3R:14,356,561, we analyzed DNA sequence generated by inverse and conventional PCR covering the breakpoint between the StanEx P-element and adjacent genomic sequences in these 9 StanEx lines. We also attempted to amplify KP specific sequences using KP-specific primers. In the nine insertions of StanEx into 3R:14,356,561, no KP element sequences were detected (data not shown). Thus, in the process of the hybrid dysgeneses that gave rise to StanEx insertions at 3R:14,356,561, the KP element was concurrently deleted at that site. Additional molecular analysis (Figure 3A-C, Suppl. Table 4) revealed that in 7/9 cases, this KP replacement by the StanEx P-element conserved the direction of the original KP-element, and in 2/9 cases the P-element replacement led to small genomic DNA deletions adjacent and 5′ to the integration site (Suppl. Results 1 and Suppl. Table 4). Together these findings suggest that the StanEx-element replaces the KP element at the site 3R:14,356,561 in all cases.

Creating an interscholastic network to generate resources for Drosophila genetics

In a prior study, we produced and characterized novel fly enhancer trap lines through a unique course partnering students and instructors in a genetics class (Bio670) at an independent New Hampshire secondary school (Phillips Exeter Academy) with university-based researchers (Kockel ). To test if this paradigm could be expanded to include additional classes, we developed a second molecular biology class at Exeter that mapped StanEx P-element genomic insertions with inverse PCR-based molecular methods (Bio586). The Bio586 class was taught by teachers who also led Bio670. This expansion of the curriculum to include experimental molecular biology relieved a bottleneck that arose due to the single term duration (11 weeks) of Bio670, that left >50% of newly-generated StanEx insertion lines unmapped. Thus the two classes, Bio586 and Bio670, integrated and enhanced the longitudinal quality of genetic experiments presented here.

We next assessed if the curriculum of fly-based transmission genetics and hybrid dysgenesis, molecular characterization of insertion lines, and expression analysis with the lexAop-GFP reporter gene could be adapted to year-long genetics classes at other secondary schools.

Over a 4 year span (2016-2019), we expanded the curriculum originally developed with Phillips Exeter Academy, NH to three new sites: Commack High School in New York (Res11/12), Pritzker College Prep, Illinois (Bio670), and The Lawrenceville School, New Jersey (Sc574); (Figure 4). Thus, data and resources detailed here stem from secondary schools collaborating throughout the academic year with a research university (Stanford). To foster production and sharing of data and fly strains, and to achieve student learning goals, the partners in this interscholastic network benefitted from structured interactions, including (1) summer internships for students (n = 17) or instructors (n = 9) with the university research partner, (2) weekly term-time research teleconferences organized by university partners with high school instructors and classes, and (3) annual site visits of university collaborators to secondary school classes during term time (Figure 4). Multiple initially un-programmed pedagogical outcomes resulted for students and teachers at partnering schools. These included service by students, who completed these classes, as proctors or teaching assistants in the next term (n = 10), instruction of incoming teachers by students who successfully completed the course in a prior year (n = 3), and collaboration between science instructors at different schools to establish new science curriculum through direct consultation and sharing of open-source materials (n = 4). Thus, a consortium of students, teachers and leadership at multiple, geographically-unconnected secondary schools and university-based programs have formed a unique research network actively generating well-characterized fly strains suitable for investigations by the community of science.

Figure 4

Structure of scholastic network and routes of resource exchange. The name of the high school and class number specification is shown. The exchange of materials and other resources (black contiguous arrows), student, teacher and instructor visits (red arrows) and voice/video/E-mail communication (black dotted arrows) is shown.

Discussion

Here we used P-element mobilization in Drosophila melanogaster to generate 93 new fly lines tagging 76 novel loci with a LexA enhancer trap. Expression of LexAop-based reporters suggests that enhancer traps in this collection are distinct, and are expressed and active in a wide variety of tissues, including neurons within the CNS and VNC. This collection has been submitted to an international fruit fly repository (Bloomington Stock Center) and should prove useful for genetic, developmental, and other studies of cells and tissues. Assessment of an apparently biased insertion frequency at locus 88B4-6 in our screen led to discovery of a previously undetected KP element in the StanEx starter strain, and we present evidence for the StanEx P-element replacing the KP element at this locus, a finding that will influence future enhancer trap screens. Data and biological resources here were generated from partnerships connecting a research university with teachers and students at multiple secondary schools. This collaboration illustrates the feasibility of building an interscholastic network to conduct biological research that impacts the community of science, and provides an ‘experiential’ paradigm for STEM education.

Experimental approaches in biology benefit from temporal- or cell-specific control of gene expression, like that possible with binary expression strategies pioneered in the Drosophila GAL4-UAS system (Brand and Perrimon 1993). Intersectional approaches, like simultaneous use of the LexA-lexAop and GAL4-UAS systems, have also greatly enhanced experimental and interpretive power in fly biology, particularly studies of neuroscience and intercellular communication (Simpson, 2016, Martín and Alcorta 2017, Dolan ). For example, enhancer traps generated here (Suppl. Table 1) include insertions into genes regulating: (1) the Insulin Receptor-Akt-TORC1 pathway, like , Scylla (ortholog of the HIF-1 target REDD1), widerborst (ortholog of PP2A) and (4E-BP1 ortholog): (2) receptor kinase signaling elements, like branchless, eiger (TNFα ortholog), Star (EGFR ligand transporter), and pointed (ETS transcription factor ortholog): (3) regulators of chromatin and histone methylation, like toutatis (chromodomain protein, kismet, MEP-1, windei (H3K9 methyl transferase) and (lysine specific demethylase). Thus, new LexA enhancer trap lines presented here significantly expand the arsenal of available LexA expression tools (Kockel , Pfeiffer ).

Prior studies have demonstrated that P-element insertion in flies is non-random (O’Hare and Rubin 1983; Berg and Spradling, 1991, Bellen ), with a strong bias for transposition to the 5′ end of genes (Spradling ). Here and in prior work, we have found a similar preference with StanEx P-element transposition; 89% of unique insertions were located in the promoter or 5′ UTR regions of genes. However, we also detected an unexpectedly high rate (9/192; 4.7%) of transposase-dependent StanEx P-element transpositions to a defined site on chromosome III within the NK7.1/Heatr2 locus, at 3R:14,356,561. Subsequent analysis provided strong evidence for P element replacement of an undiscerned KP element at 3R:14,356,561 by the StanEx donor P-element as the basis of this finding (Figure 3). P-element ‘replacement’ as a mechanism explaining biased insertion frequency in our screens is supported by our findings of (1) concomitant KP deletion upon StanEx P-element insertion into 3R:14,356,561, (2) precise substitution into the 5′ and 3′ breakpoints defined by the prior KP element, (3) absence of 8bp direct repeats generated de novo by transposon-mediated integration of the StanEx element, (4) high rate of P-element insertion into the site occupied by the prior KP element, and (5) the not-yet-explained tendency of the donor P-element to maintain the directionality of insertion of the outgoing P-element (Williams , Heslip and Hodgetts, 1994, Gonzy-Treboul , de Navas , de Navas ). KP elements in the genome of wild strains of Drosophila melanogaster have been reported frequently (Itoh ), but the origins of this particular KP element remains enigmatic. In our search to identify the source of the KP element at 88B4-6, we screened eleven fly strains obtained from the Bloomington Stock Center and identified one strain of eleven, w[1] (stock 2390), that also harbored this KP element. Stock 2390 was added to the Bloomington Center in 1989 and might therefore not represent a true copy of the original w[1] strain (Morgan 1910; Johnson 1913).

Since the inception of P-element mutagenesis screens, hot spots for P-element insertions have been noted (Bellen ). Prior enhancer trap screens, which initially used a Δ2-3 transposase-dependent strategy, later adopted an alternative Δ2-3 transposase-independent approach (e.g., PiggyBac; Thibault , Gohl ), due to unacceptably high insertional hot spot over-representation (Bellen ) and mobilization of undetected KP elements (Dr. T. Clandinin, personal communication). In these prior screens, it remains unknown if an occult KP element also distorted insertion site selection and frequencies. Our findings suggest that KP elements can contribute to phenotypes like insertional over-representation following transposase-mediated enhancer screens. However, in practice, the StanEx insertion bias to 3R:14,356,561 did not impact the productivity or strategy of our screen. Other findings (see Supplementary Results) suggest that the presence of KP might be considered when controlling use of StanEx lines.

The mutated transposase ORF in KP elements is thought to encode an inhibitor that can suppress P-element transposition (Rio 1990; Rio 1991; Simmons ). This suppression has been suggested to be KP-dose dependent (Sameny and Locke 2011), or caused by KP elements with distinct chromosomal location and high inhibitory activity (Fukui ) . The KP-element observed in this study is located at 88B4-6 and has not been previously described (Fukui ). While the rate of StanEx transposition in the absence of the KP at 88B4-6 is unknown, the data here and our prior work (Kockel ) show that StanEx element transposition was not overly inhibited by this KP element.

The resources and outcomes described here significantly extend and develop the interscholastic partnership in experiment-based science pedagogy described in our prior study (Kockel ), which involved researchers at Stanford University and a single biology class at an independent secondary school. A curriculum based on fruit fly genetics combined with developmental and molecular biology provided an ideal framework for offering authentic research experiences for new scientists, including practical and curricular features detailed in Table 1 and prior work (Kockel ; Redfield 2012). These course features offered both students and teachers a tangible prospect of generating one or more novel fly strains, thereby promoting a sense of discovery and ownership (Hatfull ), and connection to a broader community of science, each key research and educational goals. For example, within the first 9 months of submission of StanEx lines described in our prior work to the Drosophila Bloomington Stock Center there were 153 strain requests from 63 labs in 15 countries. Use of StanEx lines, in publications and through direct requests (e.g., Babski 2018; Babski ; Cohen ; Drs. L. O’Brien, A. Baena-Lopez, personal communication) are additional indicators of practical outcomes from our work.

Table 1

Qualities of this Drosophila-based curriculum promoting adoption by partnering secondary schools. See text for additional details

Qualities promoting course adoption by partnering schools
• Relative technical simplicity of fruit fly husbandry
• Modest pre-requisites for mastery of biology and genetics concepts
• Opportunity for transitioning to complex operations like tissue dissection, histology, microscopy, and code-writing
• Compatibility with flexible class time scheduling
• Concrete achievement milestones for students and teachers, including peer-reviewed publications
• Project ownership and autonomy
• Cost feasibility

Here, we show that that our interscholastic partnerships and classroom-based research have expanded to include high schools in four states in the U.S.A. These schools encompass a spectrum of public, charter, independent and ‘high needs’ schools, with day or boarding students. This expanded framework has fostered additional curricular attributes, including (1) development of student leaders who teach peers and novice adult teachers and develop curricular innovations, (2) interscholastic collaboration and data sharing through regular video conferencing organized by university researchers, (3) additional professional development for adult teachers who train or mentor novice incoming adult teachers, and create new course content tailored to their student body, (4) infrastructure development to accommodate course expansions (like larger and/or additional courses: Figure 4), and (5) improved placement of student course graduates in competitive university- and industry-based summer research internships. In summary, this experience demonstrates how longitudinal studies involving multi-generational genetics, molecular and developmental biology, and bioinformatics developed by university researchers can build a thriving, interconnected network of secondary school teachers, students and classes that impacts science, personal growth, and professional development.