The embryonic transcriptome of Arabidopsis thaliana

Background

In flowering plants, the fertilization of egg and sperm cell forms the zygote which upon a series of cell divisions produces an embryonic tissue with a characteristic body plan, often referred to as embryonic patterning. This body plan later then lays the foundation of shoot and root meristem, epidermal and ground vascular tissues. Each of these processes is marked by changes at the gene expression level and a significant amount of research has been done to understand how a zygote develops into an embryo [1-2]. However, technical challenges in isolating small embryos have precluded successful transcriptomic analysis of the different stages of development. Embryos are deeply embedded within a seed which also makes RNA extraction prone to contamination from surrounding nourishing tissues of endosperm. A lack of sensitive and low input RNA-seq protocols also hampered the study of this vital tissue.

In this preprint, the authors have compared three different low input RNA-seq protocols to study the transcriptome profile of 8 different stages (Fig. 1) of embryo development. The authors compared the publicly available transcriptome datasets of 27 different tissues from 31 developmental time points to gene expression in embryos. Based on these results, they were able to conclude that the embryo transcriptome is highly distinct and specialized compared to the other tissues.

Fig 1: Different stages of embryonic development in Arabidopsis (Hofmann et al., 2018, bioRxiv 479584)

Significant findings

The authors compared low-input mRNA seq library preparation methods – Ovation PicoSL WTA System V2, SMARTer Ultra Low Input RNA Kit, and the non-commercial Smart-seq2 method. According to this study, in terms of the cost per library, quality, and reproducibility, a Smart-seq2 protocol using off-the-shelf reagent is the clear winner.

The time-series RNA–seq data from the 8 different stages of embryo development were free from contamination by RNAs of the surrounding maternal tissues, a problem associated with most of the publicly available datasets from this development stage [3].

The authors compared the transcriptome of the embryo from this study with 27 different tissues using the publicly available datasets. Using model-based clustering analysis, the authors concluded that Arabidopsis embryonic gene expression dynamics are tightly regulated and undergoes large-scale changes which make it a highly specialized and unique transcriptome.

The authors identified 792 specific marker transcripts during embryonic development (Fig. 2). Other than the already known transcription factor markers such as Wox2, Wox8, and DRN, they also identified new markers which call for further research to understand their function. In addition to transcription factors, the authors also found potential transcriptional co-activators such as putative ubiquitin ligase and ubiquitin-like proteins.

Fig 2: : Identification of different stage-specific marker from different phases of embryo development (Hofmann et al., 2018, bioRxiv 479584)

However, the most interesting aspect of this study was the comparison of somatic and zygotic embryos in terms of their gene expression profile. This is particularly important because somatic embryos are often considered as a suitable model to study zygotic embryogenesis, but this study proves otherwise. In terms of the transcriptome, the study revealed significant differences between the somatic and zygotic embryos. They concluded that the somatic embryos are more similar to germinating tissues than the embryogenesis process of zygotic embryos.

What I liked about this preprint and why I think this work is important

This study presents a high quality, contamination-free dataset using a low input RNA-seq protocol to study the gene expression profiles during embryo development. While previous studies have shown the transcriptome of either early or late embryonic development, the transcriptome profile of the developmental time series from preglobular to mature stage was still lacking. The authors have tried to bridge this gap by studying the gene expression pattern of intermediate stages as well. The dataset from the embryonic development will be useful for the community to explore this nascent tissue, in which the dynamics of development is still not clear.

Future studies

The authors have identified several potential transcripts from different stages. These transcripts might be a regulator of embryonic development and most of these are yet not fully characterized. It would be important to understand the function of some of these regulator genes during embryonic development.

Questions for the author

While I think the experiments were well conducted and all relevant information provided for the readers, given the nature of this pre-print (genomics paper), there are some minor things I would wish authors could have included.

For example, although the differences between different RNA-seq protocols were pointed out clearly, I didn’t find much discussion regarding the possible reasons for such differences given the same kind of starting material. This extra information would be useful for the people interested in trying one of these methods for other low input RNA-seq applications.

The authors have compared the somatic and zygotic embryo and, in a way, tried to bust a myth that somatic embryo can be proxy for studying embryogenesis. They have argued that culturing conditions rich in auxin may influence the transcriptome. Given that some of the recent publication highlights the role of auxin in embryo development, did authors find any possible mechanism to explain how auxin can differentially influence the transcriptome of somatic embryos and the zygotic embryos?

Did the authors studied the function characterization of any of the development marker gene they identified in this pre-print?

References

1. Slane, D., Kong, J., Berendzen, K.W., Kilian, J., Henschen, A., Kolb, M., Schmid, M., Harter, K., Mayer, U., De Smet, I. and Bayer, M. (2014). Cell type-specific transcriptome analysis in the early Arabidopsis thaliana embryo. Development, pp.dev-116459.
2. Palovaara, J., Saiga, S., Wendrich, J.R., van‘t Wout Hofland, N., van Schayck, J.P., Hater, F., Mutte, S., Sjollema, J., Boekschoten, M., Hooiveld, G.J. and Weijers, D. (2017). Transcriptome dynamics revealed by a gene expression atlas of the early Arabidopsis embryo. Nature plants, 3(11), p.894.
3. Schon, M.A. and Nodine, M.D. (2017). Widespread contamination of Arabidopsis embryo and endosperm transcriptome data sets. The Plant Cell, 29(4), pp.608-617.

Note: The article has been accepted and will appear in an upcoming issue of Plant Reproduction.

 

Tags: arabidopsis, embryo, transcriptome

Posted on: 17 December 2018 , updated on: 19 December 2018

doi: https://doi.org/10.1242/prelights.6352

Read preprint (No Ratings Yet)