PXGS: a Poly-Transgene Expression System based on Mutually Exclusive Splicing of Dscam

  1. Renee Yin Yu
  2. Alyeri Bucio-Mendez
  3. Brian E Chen

  • Curated by eLife

    eLife logo

    eLife Assessment

    This valuable study describes PXGS, a poly-transgene expression system that exploits the mutually exclusive splicing of Dscam variable exon 4 to enable conditional, simultaneous expression of up to 12 transgenes in Drosophila, addressing a longstanding limitation in which conditional co-expression has been restricted to a handful of genes. The approach is conceptually elegant and technically accessible, with potential applications spanning neuroscience, synthetic biology, and biomanufacturing across arthropod species. The evidence that Dscam exon 4 splicing is preserved in a UAS vector and that individual alternates can be replaced with functional transgenes is solid, and the in vivo axonal re-wiring application provides a convincing proof of principle. Quantitative characterization of expression levels, a direct demonstration of expression across all twelve positions, and additional imaging controls would further substantiate the system's utility and scope.

Reviewed by

Read the full article See related articles

Discuss this preprint

Start a discussion What are Sciety discussions?

Listed in

Log in to save this article

Abstract

Biologists often need to investigate multiple genes simultaneously in an organism. However, it is currently not possible to express more than a few transgenes in an animal under conditional control. Here, we developed a technique based on the mutually exclusive splicing of the Down Syndrome Cell Adhesion Molecule1 (Dscam1) gene in Drosophila melanogaster to achieve simultaneous transgene expression of 12 genes at a time. We show that the hypervariable Dscam1 exon 4 region maintains its alternative splicing when placed in a UAS expression vector. Each of the twelve exon 4 alternates can be replaced with an exogenous gene of at least 10 kilobases and will express properly in vivo all under conditional genetic control. We demonstrate the expression of four different fluorophores placed in different exon 4 alternate positions in neural and non-neural cells in vivo. We validated the technique by rewiring Drosophila sensory neuron axons in vivo by simultaneously expressing several cell surface receptors within the neuron. This technology will also enable Drosophila melanogaster as a model system for synthetic biology research.

Article activity feed

  1. eLife

    eLife Assessment

    This valuable study describes PXGS, a poly-transgene expression system that exploits the mutually exclusive splicing of Dscam variable exon 4 to enable conditional, simultaneous expression of up to 12 transgenes in Drosophila, addressing a longstanding limitation in which conditional co-expression has been restricted to a handful of genes. The approach is conceptually elegant and technically accessible, with potential applications spanning neuroscience, synthetic biology, and biomanufacturing across arthropod species. The evidence that Dscam exon 4 splicing is preserved in a UAS vector and that individual alternates can be replaced with functional transgenes is solid, and the in vivo axonal re-wiring application provides a convincing proof of principle. Quantitative characterization of expression levels, a direct demonstration of expression across all twelve positions, and additional imaging controls would further substantiate the system's utility and scope.

  2. eLife

    Reviewer #1 (Public review):

    Summary:

    This manuscript describes the development of an expression system enabling up to 12 transgenes using the alternatively spliced fourth exon of *Drosophila* *Dscam* gene under the control of a UAS element. This will be a useful tool if expression is needed in *Drosophila* cells (in culture or in vivo). where *Dscam* splicing machinery is active, which limits its use.

    Strengths:

    The tool developed is based on a well-established genomic element. The underlying idea is relatively simple yet effective.

    Weaknesses:

    The authors describe the weaknesses of their system well, most importantly, depending on the presence of adequate levels of Dscam splicing factors in targeted cells. This likely limits effective use of the methodology to some cell lines (e.g., S2) and certain tissues (nervous system and innate immune system). The manuscript could do a better job in showing protein expression levels more quantitatively, either in comparison to other methods or as absolute values (transcript numbers, protein molarity, etc.).

  3. eLife

    Reviewer #2 (Public review):

    Summary:

    In this manuscript, Yu et al seek to develop a Drosophila genetic tool to simultaneously co-express up to 12 transgenes. They leverage the native Dscam exon 4 alternative splicing to generate a UAS to enable cell- and temporal-specific expression of transgenes. This tool is called the poly-transgene expression system (PXGS). Previous approaches to co-express transgenes have been limited to four to five genes, so PXGS would be a significant advancement, especially when examining processes that require robust expression of many genes to confer function. The authors showed that PXGS can drive expression of multiple (1) fluorescent reporters and (2) cell surface receptors in different cell types (neurons, glia, and muscles). However, there are major proof-of-principle experiments missing to demonstrate the utility of PXGS and its potential limitations. Additionally, some of the data is just not interpretable, and experimental rigor is significantly lacking.

    Strengths:

    Developing a genetic tool to co-express transgenes beyond what is currently available would be significant.

    Weaknesses:

    (1) While the authors stated that each PXGS construct can express 12 transgenes, this was not directly tested - the largest number of genes tested was in the PXGS_fluorophores, which has 4 genes inserted in 10 alternates (and therefore it can be determined if genes from all the alternates are spliced in at a meaningful level).

    a. First, the authors state that they tested the expression of the fluorophores in S2 cells using RT-PCR before generating the fly line. However, this data is not shown. Also, it is possible to test expression and localization of UAS transgenes in S2 cells with a ubiquitous GAL4, similar to what they did for GFP expression in Supplemental Figure 1.

    b. In the corresponding figure for this experiment (Figure 2), there are some concerning expression patterns and potential channel bleed-through/crosstalk. The nSyb-Gal4 is a pan-neuronal driver, yet expression of three fluorophores was extremely minimal. This could potentially be explained by the deterministic vs random alternative splicing. However, it is more concerning that in the GFP, RFP, and iRFP channels, the exact same tiny cluster of neurons is observed, suggesting potential bleed-through of the channels. Appropriate controls are required, including expression of a traditional fluorescent reporter with the nSyb-Gal4 they are using. And replicates would help with the experimental rigor.

    c. Additionally, it would be helpful to know if genes inserted at each alternate exon are expressed at a similar efficiency (vs. some alternate exons have higher levels of expression)

    (2) PXGS expression in non-neuronal cells: The authors attempt to show that fluorophores targeted to different cellular compartments can be expressed in neurons and non-neuronal cells (glia and ubiquitously).

    a. In Supplemental Figure 2A, they first use S2 cells to confirm expression, which does confirm. However, they use no markers to show that the fluorophores localize to the corresponding compartments (e.g., mitochondria and nucleus). In Supplementary Figure 2B, with that magnification and resolution, it is impossible to determine if the fluorophores localize properly.

    b. In Figure 3, it is impossible to know if there is any glial expression based on those images. They state that "subcellular localization of fluorophores was observed in the flight muscle", but again, that cannot be concluded from the images. Also, the schematic of the construct is the same one used in Supplementary Figure 2. Why show it again?

    (3) In the functional expression of PXGS transgenes section, while the authors used RT-PCR to show that each receptor gene is transcribed in S2 cells, it was not tested if they are correctly expressed, translated, and localized in the fly. The functional outcome observed (Figure 4 b-c) could be the result of misexpression of one or multiple genes.

    a. Supp Figure 3: Why does the Sli lane have so many bands?

    b. Why were these genes chosen for misexpression? Is there any evidence that they are required for the wiring of the mechanosensory neuron? Does the co-expression lead to an additive effect?

    c. The RT-PCR result (Supplemental Figure 3) showed variations (e.g., kek and kir being significantly dimmer than tutl; multiple products for Sli). Is there any explanation behind this, and could this be the outcome of some alternate exons being more efficiently spliced than others?

    d. Figure 4: These images seem to be taken with a widefield scope and only one plane. Is it possible that some of the pSC neurons are in a different Z plane, and they are not being captured here? There definitely is part of the axon terminal out of focus in some of the images. Also, most of the figure graph axes (e.g., 4b) are extremely difficult to read. And the figure overall is not easy to interpret.

    (4) Supplemental Figure 5: This figure is quickly mentioned in the Discussion without much explanation. First, this must be in the Results section since it is an experiment. Second, this needs more context because, as is, it seems like it was just thrown into the manuscript.

    (5) The authors mentioned that the size of the inserted genes could be a limitation for this technique and tested cell surface receptors of different sizes. However, there was no explicit discussion in the main text.

  4. eLife

    Reviewer #3 (Public review):

    This paper, by Brian Chen and collaborators, adapts the highly alternatively spliced Dscam1 gene locus for use in a system for simultaneous multi-transgene expression in a variety of insect species. Specifically, they show that the hypervariable Dscam1 exon 4 region maintains its alternative splicing when placed in a UAS expression vector, and that each of the twelve exon 4 alternates can be replaced with an exogenous gene such that co-expression of up to twelve proteins can be achieved. Since the co-expression of more than a few proteins simultaneously is difficult, this represents a significant advance with multiple use cases. The authors validated the technique by assessing expression in vitro and in vivo, and by rewiring Drosophila sensory neuron axons by simultaneously expressing several cell surface receptors within the neuron. Overall, this is a clearly written paper that describes a potentially important new system. I have no major criticisms.

  5. eLife

    Reviewer #4 (Public review):

    From the Reviewing Editor:

    All three reviewers recognized the conceptual originality of PXGS and its value to the Drosophila community and the broader multi-gene expression field. The core demonstration - that Dscam exon 4 mutually exclusive splicing is maintained in an exogenous UAS vector, and that individual exon alternates can be replaced with genes of interest for conditional in vivo expression - was viewed as solid and creative. The in vivo application re-wiring pSc axonal arbors using PXGS constructs loaded with cell surface receptors was noted as an encouraging functional validation.

    The reviewers differed in their overall enthusiasm. Reviewer 3 found the experiments straightforward, the results clear, and the system a significant advance with broad use cases, with no major criticisms. Reviewer 1 viewed the evidence as broadly solid, with the principal limitation being a lack of quantitative expression data and a dependence on adequate Dscam splicing-factor levels that constrain the system's applicable cell types - a limitation the authors themselves describe well. Reviewer 2 was the most critical, finding the underlying concept significant but the supporting data insufficiently rigorous to conclusively establish the tool's utility. The points below reflect the areas where reviewers - principally Reviewers 1 and 2 - felt the manuscript could be strengthened, should the authors choose to revise.

    (1) Quantitative characterization of expression. Reviewers 1 and 2 both noted the absence of a quantitative comparison of PXGS-driven expression - in absolute terms (transcript numbers, protein amounts) or relative to standard UAS constructs. Given that signal is inherently divided across 12 alternates per transcription event, characterizing expression efficiency and whether all alternates are spliced and expressed at comparable levels would substantially strengthen the manuscript. Reviewer 2 specifically asked whether genes at different exon 4 positions are expressed with similar efficiency.

    (2) Direct demonstration of 12-transgene expression. Reviewer 2 noted that, although the manuscript claims expression of up to 12 transgenes, this was not directly tested - the largest construct placed 4 distinct genes across 10 alternates, and the largest functional test used 3 genes per construct. Either a direct demonstration with more positions occupied or a more carefully bounded claim supported by the probabilistic framework would address this.

    (3) Controls and interpretability of fluorophore expression. Reviewer 2 raised concerns about Figure 2, where expression of three fluorophores under nSyb-Gal4 was minimal, and the same small neuronal cluster appeared across the GFP, RFP, and iRFP channels - raising the possibility of channel bleed-through. Appropriate controls (including a conventional single UAS-fluorophore driven by the same nSyb-Gal4) and replicates were requested. Reviewer 2 also noted that the S2 cell validation data for the fluorophore constructs, described as having been performed prior to fly line generation, are not shown.

    (4) Non-neuronal expression and subcellular localization. Reviewer 2 noted that the compartment-specific localization claims (mitochondria, nucleus) in Supplemental Figure 2 and Figure 3 are not supported by co-markers, and that the magnification and resolution in key panels are insufficient to confirm proper localization, particularly in glia.

    (5) Functional expression of receptor constructs. Reviewer 2 noted that, while RT-PCR confirms transcription of each receptor in S2 cells, correct translation and localization in the fly were not directly tested, so the observed phenotypes could reflect mis-expression of one or a subset of the genes. Clarification of why the specific genes were chosen, whether co-expression produces additive effects, and the cause of the variable RT-PCR band patterns (e.g., the multiple Sli products, dimmer kek and kir signals) was requested.

    (6) Supplemental Figure 5 (synthetic biology / RNAi). Reviewer 2 noted that this figure is mentioned only briefly in the Discussion despite representing an experiment, and recommended moving it to the Results with appropriate context. Reviewer 1 separately queried the meaning of the two white boxes in this figure and whether the RT-PCR convincingly supports expression of all genes shown.

    (7) Figure quality and labeling. Both reviewers flagged that the labels in Figure 4 (particularly panel d) and the axes in Figure 4b are too small to read. Additional labeling points were noted (alignment of "Repo-GAL4" and "brain" in Figure 3; unlabeled images in Supplemental Figures 2 and 4; clarification of whether the two lanes per group in Figure 1c are replicates or use different primers). Reviewer 1 also suggested that some figure legends describe conclusions rather than what is shown, and recommended that legends describe the data with interpretation kept to the text.

    (8) Additional points. Reviewer 1 suggested showing more of the gel in Figure 1 to demonstrate the absence of non-spliced fragments; clarifying the 1/12 probability argument (or moving it to the Discussion with transcript-number context); providing sequences for the fluorophore variants and fusion tags; and minor prose corrections ("Regardless if" → "Regardless of whether"; "dependent on three things" → "dependent on three factors"). Reviewer 2 noted that the gene-size limitation, though tested, is not explicitly discussed in the main text.

  6. eLife

    Author response:

    We are pleased that the reviewers viewed the core demonstration (that Dscam mutually exclusive splicing is preserved in a vector and that exon alternates can be replaced with genes of interest) as a solid foundation for the system. We agree that the manuscript would be strengthened by clearer quantitative characterization of expression, additional controls for fluorophore imaging, improved presentation of the figures, and more precise wording about the current scope of evidence. In a revised manuscript, we plan to address these points by adding or clarifying quantitative expression analyses, including S2 cell validation data, adding appropriate imaging controls where available, revising claims about 12-transgene expression to distinguish design capacity from direct experimental demonstration, and improving figure labels and legends throughout.

    We also plan to expand the discussion of PXGS limitations, including cell-type dependence on Dscam splicing machinery, possible position effects, and gene size considerations. Finally, we will improve Methods reporting by adding resource identifiers, cell culture quality control information, statistical design details, and data/code availability statements where appropriate.

    We appreciate the opportunity to revise the manuscript and believe these changes will make the strengths and limitations of PXGS clearer to readers.

  7. Version published to 10.1101/2024.10.27.620485 on bioRxiv