Developmental Biology 450 (2019) 1–8Contents lists available at ScienceDirect
Developmental Biology
journal homepage: www.elsevier.com/locate/developmentalbiology
Short Communication
Multiplex clonal analysis in the chick embryo using retrovirally-mediated
combinatorial labeling
Weiyi Tang, Yuwei Li, Shashank Gandhi, Marianne E. Bronner *
Division of Biology and Biological Engineering, California Institute of Technology, Pasadena, CA 91125, USA
A R T I C L E I N F O
A B S T R A C T
Keywords:
Replication incompetent avian retrovirus
Multiplexed clonal analysis
Combinatorial labelling
Neural tube
Trunk neural crest
Multipotency
Lineage analysis plays a central role in exploring the developmental potential of stem and progenitor cell populations. In higher vertebrates, a variety of techniques have been used to label individual cells or cell populations,
including interspecies grafting, intracellular microinjection, and Cre-mediated recombination. However, these
approaches often suffer from difficulties in progenitor cell targeting, low cellular resolution and/or ectopic labeling. To circumvent these issues, here we utilize replication incompetent avian (RIA) retroviruses to deliver
combinations of fluorescent proteins into distinct cellular compartments in chick embryos. In particular, RIAmediated lineage tracing is optimal for long term mapping of dispersing cell populations like the neural crest.
Using this tool, we confirm that trunk neural crest cells are multipotent. Furthermore, our RIA vector is engineered to be fully adaptable for other purposes such as cell fate analysis, gene perturbation studies and time-lapse
imaging. Taken together, we present a novel approach of multiplex lineage analysis that can be applied to normal
and perturbed development of diverse cell populations in avian embryos.
1. Introduction
A fundamental challenge in developmental biology is to determine
the range of cell types that can arise from single embryonic cells. Whereas
the fertilized egg or embryonic stem cells are totipotent, other embryonic
cell types may be multipotent, bipotent or unipotent. Distinguishing
between these possibilities requires methods that enable analysis of a
cell’s developmental history. In organisms like C. elegans, cell lineage
analysis has been elegantly done by visually mapping the progeny of each
cell division (Sulston, 1983). While this was a technical tour de force in a
simple organism, it is not feasible in more complex and/or opaque organisms. Analysis of cell lineage is particularly challenging in vertebrate
embryos.
In the past decades, various techniques have been developed to follow
and analyze cell fates that greatly expanded our understanding of both
embryogenesis and organogenesis. One approach is to directly inject
single cells with fluorescent dyes that are large and thus cannot pass
through gap junctions (Bronner-Fraser and Fraser, 1988). A second
approach involves virally-mediated lineage analysis using β-galactosidase (Frank and Sanes, 1991; Sanes, 1989) or barcoded libraries (Gerrits
et al., 2010). Third, quail-chick chimeric grafts have been a
well-established tool for understanding the contribution of neural crest in
avian embryos (Ayer-Le Lievre and Le Douarin, 1982). Finally, transgenic
Confetti and Brainbow technology in mice and zebrafish, respectively,
have been used to express fluorescent proteins at specific time points,
enabling clonal analysis using high resolution imaging (Cai et al., 2013;
Livet et al., 2007). A complication of the latter, however, is that it is
difficult to establish clonality as it involves sophisticated statistical
analysis of rare color combinations (Baggiolini et al., 2015). Moreover,
this technology is not easily applicable to non-genetic organisms and is
particularly challenging for cell types that are migratory and thus
disperse widely in the embryo.
One example of such a cell type is the neural crest. Neural crest cells
originate within the closing neural tube but then migrate to diverse locations in the embryo and form diverse cell types (Ayer-Le Lievre and Le
Douarin, 1982). Despite the broad range of the neural crest derivatives at
a population level, there are conflicting views in the literature regarding
the degree of multipotency of individual neural crest cells. Single cell
lineage analysis using microinjection of fluorescent dyes suggested that
trunk neural crest cells in avian embryos are multipotent at both premigratory (Bronner-Fraser and Fraser, 1988) and migratory (Bronner-Fraser and Fraser, 1989) stages. In the mouse, Sommer and colleagues
(Baggiolini et al., 2015) found similar results using R26R-Confetti technology. However, other studies have suggested that neural crest cells are
* Corresponding author.
E-mail address: mbronner@caltech.edu (M.E. Bronner).
https://doi.org/10.1016/j.ydbio.2019.03.007
Received 6 December 2018; Received in revised form 11 March 2019; Accepted 12 March 2019
Available online 15 March 2019
0012-1606/© 2019 Elsevier Inc. All rights reserved.
W. Tang et al.
Developmental Biology 450 (2019) 1–8
HH11 chicken embryos. The embryos were sealed with surgical tape,
incubated at 37 C for 2 days. Embryos were harvested at HH21 (n ¼ 16),
dissected and a 500 μm thick transverse slice at forelimb region was cut
for imaging analysis (detailed procedure in Li et al., 2019). For immunohistochemistry, embryos were fixed in 4% PFA in PBS for 30 min at
4 C, embedded in Tissue-Tek O.C.T compound (Sakura #4583) and
sectioned (Microm HM550 cryostat).
restricted in their fate, even before emigration from the neural tube
(Krispin et al., 2010; Weston and Thiery, 2015). Thus, further analysis of
neural crest lineage is warranted, particularly since little is known about
their developmental potential at axial levels other than the trunk region.
The chick embryo has been an excellent system for neural crest
biology. As amniotes, chick embryos develop similar to humans at stages
of neural crest migration, but are cost-effective compared with mice.
Moreover, their ability to develop outside of the mother facilitates not
only genetic and surgical manipulations but also dynamic imaging
(Benazeraf et al., 2010). Recently, we have applied RIA-mediated lineage
analysis to chick cartilage to understand how sister cells rearrange within
a constrained region surrounded by extracellular matrix (Li et al., 2017).
Here, we expand the repertoire of colors and subcellular localizations of
these reagents and demonstrate their utility for defining clonal relationship of other cell types, including those that disperse widely in the
embryo, like neural crest cells. As proof of principle, we confirm the
multipotent property of trunk neural crest cells. Thus, our study enriches
the toolkit for cell lineage analysis in a non-genetic model organism.
2.4. Immunohistochemistry and imaging analysis
Frozen tissue sections were incubated in 1xPBS at room temperature
to remove O.C.T, permeabilized with 0.3% vol/vol Triton-X100 in
1xPBS. Primary and secondary antibodies were diluted in blocking buffer
(1xPBS with: 5% vol/vol normal donkey serum, 0.3% vol/vol TritonX100). Sections were stained with primary antibody at 4 C overnight
(primary antibody dilutions: 1:500 Mouse anti-smooth muscle actin,
Sigma-Cat# F1840-200μG; 1:500 Mouse antiHuC/D IgG2b, InvitrogenCat#A21271; 1:10 E1.9 primary sensory and motor neuron marker,
DSHB). The next day, sections were washed with 1xPBS, and treated with
secondary antibody for 1 h at room temperature. The following secondary antibodies were used: 1:1000 donkey anti-mouse IgG 647, 1:1000
goat anti-mouse IgM 647, 1:1000 goat anti mouse IgG2b 647, Molecular
Probes. Sections were imaged using a Zeiss AxioImager.M2 with Apotome.2 and Zeiss LSM 800 confocal microscope. The number of infected
cells displaying single color was quantified with thresholding followed
by automated particle analysis using FIJI software. The number of doubly
and triply infected cells was counted manually. For Feret’s angle analysis
of clonal orientation, a cell cluster with identical double-or triple-color
was identified. Subsequently, a polygon was drawn along the boundary
of the clone and the angle of the polygon was automatically measured by
FIJI.
2. Materials and methods
2.1. Molecular cloning and virus preparation
The original RIA vector (Chen et al., 1999) was modified by inserting
AscI, SpeI, NotI upstream of existing ClaI digestion site. H2B-YFP
(#96893), H2B-RFP (#92398), Mito-CFP (#36208), Membrane-YFP
(#56558) and Utrophin-Scarlet (#26739) were obtained from Addgene, and were subsequently subcloned into the modified RIA vector.
ENV-A plasmid encoding the Envelope A protein was a gift from Dr.
Connie Cepko and colleagues (Chen et al., 1999). ENV-A plasmid was
co-transfected with recombinant RIA plasmids into chick DF1 cells to
pseudotype RIA virus (ATCC, Manassas, VA; #CRL-12203, Lot number
62712171, Certificate of Analysis with negative mycoplasma testing
available at ATCC website) in 15 cm culture dishes using standard
transfection protocols. 24 h post-transfection, the cell culture medium
was harvested twice per day for four days, and concentrated at 26,
000 rpm for 1.5 h. The pellet was dissolved in 20–30 μl of DMEM to
achieve the titer of 1 107 pfu/ml. The viral aliquots were stored in
80 C until injection.
3. Results and discussion
3.1. Optimization and in vitro validation of replication-incompetent avian
(RIA) retroviruses for multiplex clonal analysis
RIA retrovirus has been employed widely for lineage tracing in chick
embryos (Chen et al., 1999). After stably integrating into host genome,
RIA DNA is faithfully and equally passaged to daughter cells during cell
division. In contrast to replication-competent avian (RCAS) virus
(Fig. 1A), RIA does not synthesize its own envelope (ENV) protein and
thus cannot horizontally transfer into neighboring cells. As a result, all
descendants of an RIA infected cell are permanently labeled with a
unique common signature, representing an isogenic clone. This feature
offers several advantages over other approaches for lineage analysis: first,
it is less invasive compared to quail-chick grafts, marking endogenous
host cells without the necessity for surgery or combining tissue from
different species; second, compared with progressively diluting vital
dyes, the integrated RIA transgene enables indelible marking of host cells
and their progeny (Li et al., 2018); finally, compared to Confetti mouse
and zebrafish labeling, RIA ensures specific targeting of progenitor cells
through focal injection, avoiding the general challenges faced by the
Cre-lox systems that are dependent upon the choice of promoters (Lewis
et al., 2013).
Previous studies have used RIA viruses to resolve clonal boundaries in
cartilage, tendon, perichondrium and retina (Li et al., 2017; Pearse et al.,
2007). One common characteristic of all these tissues is that they are
comprised of coherent clones with regular morphology. To extend this
analysis to broader types of tissues, we have optimized this approach and
expanded its application to cells that migrate extensively and actively
rearrange.
To this end, we modified the previous version of the RIA vector (Chen
et al., 1999) by adding three unique restriction enzyme sites (AscI, SpeI,
NotI) in the upstream of the common cloning site ClaI. The AscI and SpeI
sites were preserved for expanding the application of this viral vector for
2.2. Virus titering and clonal analysis in DF1 cells
Concentrated virus was serially diluted at 1:104, 1:105, 1:106, and
1:107 respectively, in 200 μl of DMEM. The diluted viral solution was
added to chick DF1 cells in 24-well-plates. Cells were incubated at 38 C
for two hours to permit viral infection. Another 800 μl DMEM was subsequently added and infected cells were incubated for 72 h to allow for
the expression of fluorescent proteins, which was used as a readout for
functional recombinant virus. The number of cell clusters was quantified
under an epifluorescence microscope. Typically, within a well of cells
infected with the virus at 1:107 dilution, one positive cluster was
observed, meaning the viral titer is around 1 107 pfu/ml. For clonal
analysis in DF1 cells, a mixture of H2B-RFP, Mem-YFP and Mito-CFP
viruses was serially diluted and infected DF1 cells as above. The number of cell clusters (clones) with distinct color(s) in a given field of view
(n ¼ 3) was quantified using image software FIJI.
2.3. Viral injection and chick embryology
Equal amounts of RIA viruses encoding H2B-YFP, H2B-RFP, MitoCFP, Membrane-YFP and Utrophin-Scarlet were mixed. A working solution was made at 1:2 dilution of the viral mixture with Ringer’s solution
(0.9% NaCl, 0.042%KCl, 0.016%CaCl2 2H2O wt/vol, pH7.0), and
supplemented with 0.3 μl of 2% food dye (Spectral Colors, Food Blue
002, C.A.S# 3844-45-9) as an indicator. About 1 μl of working solution
was injected into the lumen of the neural tube posterior to somite 3 in
2
W. Tang et al.
Developmental Biology 450 (2019) 1–8
Fig. 1. Validating RIA-mediated multicolor clonal analysis in cell culture. (A) Construct map of RCAS, RIA, and modified RIA with restriction enzyme sites (blue). (B)
Transgenes encoding fluorescent proteins targeted to specific subcellular regions were cloned into the RIA vector. Note the maps in (A) and (B) are not to scale. (C–F)
Equal mixture of three viruses (H2B-RFP, Mem-YFP, Mito-CFP) was serially diluted and infected into DF1 cells. (G, H) The number of double and triple-infected clones
showed an inverse relationship to viral titer. (I–L) High magnification imaging on infected cells with 63 objective lens confirmed specific subcellular targeting of
fluorescent proteins. Scale bars: C-F 100 μm, I-L 10 μm.
improvements in the complexity of viral infection enable spatial resolution of individual clones.
We performed a careful characterization of this novel cell-tagging
reagent in vitro before applying it to in vivo systems. First, we showed
these viruses can label clones in cell culture by introducing serial dilutions of an equal mixture of three viruses (H2B-RFP, Mem-YFP, MitoCFP) into DF1 cells. As dilution factor increased (Fig. 1C–F), the numbers
of double and triple-colored cell clusters decreased logarithmically
(Fig. 1G and H); such a pattern was expected if infection with a viral
particle is an independent event (Turner and Cepko, 1987). Second, we
performed high magnification imaging to confirm the fidelity of the
future research, such as, creating fusion proteins and bicistronic systems
to visualize protein dynamics and perform gene functional study,
respectively. For the purpose of lineage tracing in wild-type chick embryos in this study, we generated five distinct viruses by inserting
H2B-YFP,
H2B-RFP,
Mito-CFP,
Membrane-YFP
(Mem-YFP),
Utrophin-Scarlet (Utr-Scarlet) between NotI and ClaI sites (Fig. 1B).
Clonality assessment was based on three criteria: a complex color combination created by multiple (two or more) infections of a single progenitor cell; similar signal intensity, based on numbers of copies
integrated and/or insertion location in the host genome; distinct subcellular localizations of fluorescent proteins. Taken together, these
3
W. Tang et al.
Developmental Biology 450 (2019) 1–8
subcellular localization of distinct fluorescent proteins and excluded the
possible effects of optical bleed-through (Fig. 1I–L).
rule out the possibility that clonal assessment could be biased if viral
aggregates infected adjacent progenitor cells in neural tube, we calculated the error rate of the most frequent doubly infected clone expressing
both Mem-YFP and Mito-CFP with the formula previously described
(Fields-Berry et al., 1992):
3.2. Clonal relationships within the neural tube
To extend our analysis in vivo, we used our recombinant virus to
confirm the spatial arrangement of clones in the neural tube. A previous
study using time-lapse imaging revealed that these cells undergo oriented
division, and in turn, form laterally distributed clusters (Tawk et al.,
2007), providing a simple in vivo system to validate our reagents. As such,
we injected a mixture containing equal amounts of five distinct RIA viruses into the lumen of the neural tube at Hamburger and Hamilton (HH)
stage 11 (n ¼ 6), and harvested the embryos 48 h-post infection for imaging thick slices in transverse orientation (Fig. 2A). In a single orthogonal slice, we identified six clones according to the criteria established
above; all the clones appeared to align orthogonal to the elongation axis
of the neural tube (Fig. 2B and C, Movie S1). We segmented these clones
and enclosed them with a polygon using FIJI to measure their Feret’s
angle, the angle between the maximum diameter of the polygon and the
lateral axis of the neural tube (see details in Methods) (Fig. 2D and E).
This endpoint analysis revealed an average Feret’s angle of about 10 ,
consistent with the model of lateral clonal expansion suggested by previous longitudinal analyses (Fig. 2F) (Tawk et al., 2007). Both double and
triple infection events were rare, with probabilities p
Order an Essay Now & Get These Features For Free:
Turnitin Report
Formatting
Title Page
Citation
Outline
