Menu

Close

CLADES: a programmable sequence of reporters for lineage analysis

Jorge Garcia-Marques, Ching-Po Yang, Isabel Espinosa-Medina, Minoru Koyama, Tzumin Lee

Preprint posted on May 30, 2019 https://www.biorxiv.org/content/10.1101/655308v1

CLADES - An innovative strategy to resolve lineage progression in an organism through multicolor labeling.

Selected by Ying-Tsen Tung

 

Background:

Lineage tracing identifies all the descendants generated from a single ancestor cell, which provides important information about how cell fates are progressively acquired during tissue development. Several revolutionary techniques have been developed for lineage tracing [1]. Imaging-based systems utilize genetic approaches to induce the expression of randomly combined fluorescent reporters in different lineages. These methods reveal cell clones in situ but cannot trace lineage progression. On the other hand, DNA sequencing based strategies rely on the CRISPR/Cas9 system to stochastically generate DNA barcodes that progressively accumulate over cell divisions, therefore allowing to deduce the birth order of lineages at large-scale. However, both spatial and morphological information is lost with this approach. In this preprint, the authors developed CLADES (Cell Lineage Access Driven by an Edition Sequence) as a programmable imaging based tool to trace cell clones and resolve their lineage progression by pre-established color order.

 

Key findings:

CLADES utilizes programmable genetic switches to induce or inactivate fluorescent reporters sequentially. The combination of different reporters generates a pre-determined color code to reveal the birth order of cells in a lineage.

The genetic switch builds upon a gRNA cascade where each gRNA is sequentially activated via DNA editing mediated by its corresponding pre-activated gRNA. There are two keys. One is to make the editing outcome predictable, and the other is to express the gRNA structure in an inactive form and make it functional after the predictable DNA editing. For this purpose, the authors utilized CaSSA, a CRISPR based approach they developed in a recent preprint [2] (highlighted by a preLight post [3]), which is designed to allow single strand annealing (SSA) to mediate precise recombination upon Cas9 editing. Here, the authors optimized a gRNA structure with a CaSSA switch in the scaffold region to achieve minimal leakiness without the trigger gRNA in Drosophila and zebrafish.

Another critical step of the CLADES design is to synchronize the cascade of gRNAs and reporters. Ideally, preactivated gRNA#1 directs Cas9 to simultaneously activate the reporter#1 (ex. GFP) and gRNA#2, which in turn switches on the reporter #2 (ex.RFP). Yet the authors found a failure of reporter progression when the reporter and gRNA loci were independently positioned in Drosophila. After several rounds of optimization, they alternatively embed gRNA cassettes within the targeted reporter which is actively expressed as a truncated non-functional protein. The reporter contains two gRNA switching cassettes, one for making the reporter in-frame upon editing (ON cassette) and one for placing the same reporter out-of-frame upon editing (OFF cassette) (Fig. 1). This approach ensures the sequential regulation of reporters by corresponding gRNA and maximizes the lineage tracing capacity.

Figure 1. The design of coupled gRNA- and reporter-cascades. (Figure adapted from the preprint by Garcia-Marques et al.)

 

Based on the elegant design, the authors developed the advanced CLADES 2.0 system with 3 reporters and 5 colors in Drosophila. It combined a serial ON and OFF cassettes to generate gRNA cascades for the sequential expression of YFP → YFP+RFP → RFP → RFP+CPF → CFP. The authors also incorporated the GAL4/UAS system to drive reporters that only label lineages with specific cell identities. When testing CLADES 2.0 in all neuroblasts, the larva brain can be labeled by all five predicted colors. Furthermore, the progression of the 5 color cascade can also reflect the birth order in the well-characterized ALad1 lineage (Fig. 2).

Figure 2. CLADES 2.0: a five-color cascade with GAL4 induction (a) Schemes illustrating the cascade progression. (b) A larval brain with all blastocysts labeled by CLADES 2.0 + Dpn-GAL4. (c) The ALad1 lineage labeled by CLADES 2.0 + Acj6-Gal4. Note the yellow color might be due to perdurance Cas9 activity in GMC. (Figure adapted from the preprint by Garcia-Marques et al.)

 

In addition to lineage tracing, CLADES has great potential to reveal any sequential biological event. By using a germline-specific driver for Cas9 expression in CLADES 1.0, the color of the dominant fly population can refer generations.

It should be noted that CLADES is still at a suboptimal status and can be further improved. The progression efficiency gradually reduced over each step, mainly due to the occurrence of non-SSA mediated recombination events (~40%). Therefore, a portion of the progeny loses the tracers, especially for the late-born cells. Also, the occasional perdurance of Cas9 activity might create an unwanted color progression in progenies (Fig. 2c). Therefore, at least 50 fly brains are required to achieve accurate reconstruction. Nevertheless, it is still much more efficient than other imaging-based methods. The authors also proposed to enhance SSA effectiveness by incorporating more repeats for SSA and to speed up cascade progressions by a more robust control of Cas9 expression.

 

Why I liked the preprint :

CLADES is the first imagining based method to trace temporal events of a cell lineage, including changes in morphology and distribution. This elegant design takes the lineage tracing research another step forward. I am particularly impressed by their persistence and rigorous tests to make the CLADES progression works! Besides, CLADES can be customized to study various biological events of interest, as the capacity (number of colors) and resolution (the speed of color progression) can be programmed. Given that the optimized gRNA structure works in zebrafish as well, I look forward to seeing the application of CLADES in vertebrate model systems.

 

Questions for the authors:

Interestingly, the progression of the CLADES cascade is much faster in Dpn-GAL4 (fig 4B) than GH146-GAL4 (Fig4D) lines (the same for comparing GH146-GAL4 to Acj-GAL4). Does driver activity of the reporter also affect the expression level of embedded gRNAs, therefore influencing the progression rate of the cascade? Is it possible to use a stronger promoter for gRNA expression to speed up reporter progression?

 

References:

  1. Ma, J., et al., Neural lineage tracing in the mammalian brain. Curr Opin Neurobiol, 2018. 50: p. 7-16.
  2. J. Garcia-Marques et al., Unlimited genetic switches for cell-type specific manipulation. bioRxiv (2018), available at (https://www.biorxiv.org/content/10.1101/470443v1)
  3. Almeida, R. CaSSA: a new versatile system for intersectional gene expression enabling cell-type specific manipulations in vivo. Nov 14, 2018, available at  (https://prelights.biologists.com/highlights/unlimited-genetic-switches-for-cell-type-specific-manipulation/ )

 

 

Posted on: 7th July 2019 , updated on: 9th July 2019

Read preprint (No Ratings Yet)




  • Author's response

    Jorge Garcia-Marques shared

    While we think additional experiments will be required to truly prove that different drivers influence the speed of the cascade, this is a very reasonable argument. Depending on the driver, transcription could occur at different levels or not occur at all. This transcription could interfere with the U6 promoters and therefore reduce the cascade speed. Similar mechanisms of transcriptional interference have been previously described. On the other hand, some U6 genes are naturally located within introns, similar to what we designed in our construct. Moreover, U6 promoters are quite strong and we do not think this is the limiting factor in the cascade. Increasing the Cas9 concentration specifically in neuroblasts or using more accessible loci to integrate the CLADES constructs will probably result in a faster cascade.

    2 comments

    1 month

    Jorge Garcia-Marques

    Thanks Ying-Tsen Tung for selecting our work for your highlight. While we think additional experiments will be required to truly prove that different drivers influence the speed of the cascade, this is a very reasonable argument. Depending on the driver, transcription could occur at different levels or not occur at all. This transcription could interfere with the U6 promoters and therefore reduce the cascade speed. Similar mechanisms of transcriptional interference have been previously described. On the other hand, some U6 genes are naturally located within introns, similar to what we designed in our construct. Moreover, U6 promoters are quite strong and we do not think this is the limiting factor in the cascade. Increasing the Cas9 concentration specifically in neuroblasts or using more accessible loci to integrate the CLADES constructs will probably result in a faster cascade.

    1

    1 month

    Ying-Tsen Tung

    Thank you for your explanation. I’ve incorporated your response into the prelight post. I look forward to advanced versions of CLADES for further applications.

    Have your say

    Your email address will not be published. Required fields are marked *

    This site uses Akismet to reduce spam. Learn how your comment data is processed.

    Sign up to customise the site to your preferences and to receive alerts

    Register here
    Close