The organization of nuclear ribosomal DNA in gnetophytes – physically separate and physically linked arrangements of 35S and 5S genes. A commentary on: ‘Remarkable variation of ribosomal DNA organization and copy number in gnetophytes, a distinct lineage of gymnosperms’

Abstract

This article comments on:

Wencai Wang, Tao Wan, Hannes Becher, Alena Kuderova, Ilia J. Leitch, Sonia Garcia, Andrew R. Leitch and Aleš Kovařík. 2019. Remarkable variation of ribosomal DNA organization and copy number in gnetophytes, a distinct lineage of gymnosperms. Annals of Botany 123(5): 767–781.

Eukaryotic cells contain as many as 10 million ribosomes, and the DNA loci coding for ribosomal RNA (rRNA) are correspondingly abundant. These loci are organized in repeat units occurring on specific locations on the chromosomes. Because of their abundance and highly conserved sequences, rDNA loci can be analysed by microscopy, once they have been suitably stained. This is achieved either by silver staining (conventional technique) or by fluorescence in situ hybridisation (FISH) in which signals emitted by probes that bind to the rDNA are captured by a camera coupled to a fluorescence microscope and analysed with a digital imaging system. Relevant for plants are 35S rDNA loci (encoding 18S-5.8S-26S rRNA genes) and 5S loci (encoding 5S rRNA), which are usually physically separated from each other. This is called the Separate or S-type arrangement. Physical linking of 35S and 5S rDNA, the so-called Linked or L-type arrangement, is rare (Garcia et al., 2017). Wang et al. (2019) now report that of three species of gymnosperms in the genera Gnetum, Welwitschia, and Ephedra that they investigated, the former two have the S-type, while the third has the L-type. What is the relevance of this?

Firstly, gymnosperms are under-represented among the so far 1791 plant species in which the organization of rDNA has been analysed with FISH (Garcia et al., 2017), and since the three genera make up the phylogenetically isolated clade Gnetales, one of five main seed plant clades (Ickert-Bond and Renner, 2016), Wang et al. (2019)’s work adds critical empirical data on the occurrence of the physical rDNA organization in land plants. Secondly, their finding weakens the hypothesis (Wicke et al., 2011) that the L-type organisation was the ancestral condition in land plants, since the new data add yet another instance of either independent gain or loss of this arrangement of rDNA loci. But how did they infer the physical rDNA arrangement (and other karyotype traits) of their study species?

For this, Wencai Wang and her colleagues in the labs of Andrew Leitch and Ales Kovarik used low-coverage high-throughput (Illumina) sequencing to first identify and characterize the rDNA sequences of the relevant species, and secondly to develop FISH probes. To analyse the resulting pools of millions of short sequences, they used the pipeline RepeatExplorer (GalaxyProject.org), which produces clusters of similar sequences (Fig. 1). that allow to identify and visualize 5S and 35S rDNA sequence clusters and to quickly decide if these genes are part of a single cluster (L-type) or not (S-type). Smooth lines in the graphical displays indicate little sequence variation between multiple rDNA repeats, while diffuse output graphs are indicative of more divergent sequences. The advantage of such in silico analyses is that they enhance one’s chance of recovering the real copy numbers of a gene in a species, since mutated copies and pseudogenes are also recovered. Wang et al. (2019) then verified their estimates by southern blot hybridization with 5S and 35S probes. As a third approach, they carried out FISH with 5S and 18S rDNA probes.

Fig. 1.

Steps used for reconstructing rDNA units of Ephedra altissima, Gnetum montanum, and Welwitschia mirabilis by Wang et al. (2019). Following low-coverage Illumina HiSeq sequencing, the authors used the RepeatExplorer pipeline to cluster rDNA sequences by similarity. The resulting graphic displays reveal whether 5S and 35S rDNA sequences are arranged separately or are linked, i.e., had the S- or L-type arrangement of rDNA. The authors also calculated the lengths of the rRNA units and verified their estimates by conventional slot-blot hybridization. Lastly, they carried out FISH.

The repeated gains and losses of S- and L-type arrangements of rDNA loci in land plants begs the question of their possible functional differences. One idea is that the physical proximity of 35S and 5S operons might have to do with the stoichiometric balance of their transcripts. Perhaps balance can be more readily achieved with a linked arrangement (L-type) of 5S and 35S units, since array homogenisation probably acts to maintain functionality and copy numbers (Garcia et al., 2017). The L-type arrangement’s scattered phylogenetic distribution, however, throws doubts on stoichiometric balance being a strong constraint on the localisation of rDNA loci (Garcia et al., 2017). Alternatively, it is not the L-type that is under strong conserving selection, but instead the S-type arrangement. This might be because each rDNA unit type (35S vs. 5S) is transcribed by its own polymerase (RNA polymerase I for 35S and RNA polymerase III for 5S) in a different nuclear compartment, inside or outside the nucleolus (Garcia et al., 2017). For this, a separate physical location might be advantageous.

Of the three Gnetales genera investigated by Wang et al. (2019), Ephedra is unusual among gymnosperms in having many polyploid species (Ickert-Bond and Renner 2016); for Gnetum polyploidy is also suspected albeit not proven (Won and Renner, 2005). With FISH analyses, using 35S and 5S probes, one can compare how ploidy levels affect rDNA loci (Fig. 1B). Such a comparison reveals that E. monosperma with 2n = 28, the same chromosome number as the Ephedra species investigated by Wang et al., (2019), the number of 35S rDNA loci is the same as that in E. californica with twice the number of chromosomes (Fig. 2), confirming Wang et al. (2019)’s findings when comparing their three species from three genera. Similarly, the number of 5S and 35S rDNA loci in Welwitschia mirabilis, with 2n = ca. 42, is not higher than that in the Ephedra species with 28 chromosomes (Wang et al. (2019); Fig. 2). While comparison of genera that have evolved separately since the Cretaceous, may not be meaningful—because of the millions of years available for genome diploidization—our comparison of species within the genus Ephedra (Fig. 2) supports Wang et al. (2019)’s assessment that in Gnetales there is no increase of rDNA loci with a doubling of chromosome number.

Fig. 2.

FISH analyses using standard rDNA probes from Arabidopsis thaliana (35S, green) and Beta vulgaris (5S, red) support Wang et al. (2019)’s conclusion that rDNA locus number is not correlated with chromosome number/ploidy. Welwitschia mirabilis, with 2n = ca. 42, has just two 5S loci (red arrowheads) and two separate 35S loci (green arrowheads). Ephedra foeminea with 2n = 14 has six 35S loci, E. monosperma with 2n = 28 has sixteen 35S loci, and E. californica with 2n = 42 has at least 15 loci, including four small 45S rDNA loci that are labelled with ‘?’. Bars correspond to 10 μm.

In the end, while we are still in the dark about the function of the different types of arrangements of ribosomal DNA in plants, the study of Wang et al. (2019) importantly reveals that the S- and the L-type both exist in a monophyletic lineage of gymnosperms.