Greater duckweed (Spirodela polyrhiza) is a monocotyledonous aquatic plant in the family Lemnaceae (Landolt, 1986). Individual duckweed plants are small with simplified body morphology and grow vigorously, forming colonies of single or mixed species ( Figure 1 ). Duckweeds comprise 36 species in five genera (Bog et al., 2019) and are a surprisingly diverse group of plants that have applications in basic research to explore genetic, physiological, and biochemical pathways, as well as various practical applications (Cao et al., 2018; Fourounjian et al., 2020; Acosta et al., 2021; Oláh et al., 2023). For example, duckweed biomass is rich in proteins, carbohydrates, crude fiber, and minerals, making it an attractive source of feed for animals (particularly aquaculture) and humans (Appenroth et al., 2017). Additionally, duckweeds have been widely used for wastewater treatment (Zhou et al., 2018), biosensing (Ziegler et al., 2019), and phytoremediation (Ekperusi et al., 2019; Zhou et al., 2023).

S. polyrhiza has the largest body size among duckweeds, with individual fronds of about 7–10 mm in diameter ( Figure 1B ). It is the most cosmopolitan duckweed species, inhabiting five continents (Tippery and Les, 2020), and mostly propagates vegetatively (Landolt, 1986). S. polyrhiza has a relatively small genome of ~160 Mb distributed among 20 chromosomes (Hoang et al., 2019). Whole-genome sequencing of S. polyrhiza ecotypes from the United States (Wang et al., 2014) and Europe (Michael et al., 2017), followed by parallel genome examination of more than 100 additional S. polyrhiza ecotypes sampled around the world (Ho et al., 2019; Xu et al., 2019) showed exceptionally low variation rates.

The ribosomal DNA (rDNA) has become a favorite molecular target in studies related to plant systematics, evolution, and biodiversity and is used as a genome-specific marker in allopolyploids and hybrids (Borisjuk et al., 1988; Stadler et al., 1995; Mahelka et al., 2017; Volkov et al., 2017; Stepanenko et al., 2022). The 35S and 5S rDNA loci are organized in clusters of tandemly repeated units composed of conserved sequences that are transcribed to produce rRNAs; these are separated by diverse intergenic spacers (NTSs), which are not transcribed (Volkov et al., 2003; Hemleben et al., 2021). Its abundance, counted in hundreds to thousands of tandemly repeated copies per genome (Prokopowich et al., 2003; Goffová and Fajkus, 2021), and the diversity of the NTSs, make the rDNA a useful target for genomic studies. However, despite substantial progress in whole-genome sequencing techniques (Michael and VanBuren, 2015; 2020), resolving repetitive sequences, such as the tandemly arranged 5S and 35S rDNA repeats (Lopez et al., 2023), remains a challenge with many repeat sequences missing from annotated plant genomes.

The relatively short, tandemly arranged 5S rDNA repeats composed of conserved sequences encoding the 5S rRNA and highly variable NTSs are especially useful for differentiating closely related species and varieties, and for evaluating genomic diversity and genome molecular evolution in numerous plant taxa (Zanke et al., 1995; Matyácek et al., 2002; Simon et al., 2018; Piredda et al., 2021; Tynkevich et al., 2022; Andreev et al., 2023), including duckweeds (Chen et al., 2021). Genomes of S. polyrhiza and S. intermedia, the only two representatives of the ancient duckweed genus Spirodela, have fewer copies of rDNA genes compared to other plants (Michael et al., 2017; Hoang et al., 2018, 2020). Chromosome characterization by in situ hybridization showed that both species have two 5S rDNA loci, a major and minor one (Hoang et al., 2019). Consistent with the low variability of nuclear DNA sequences in Spirodela, the rDNA producing the 5S rRNA is also highly conserved, as revealed by conventional sequencing of 5S rDNA units of four ecotypes of S. polyrhiza (Borisjuk et al., 2018) and S. intermedia (Hoang et al., 2020). Sequence analysis of the 5S rDNA repeats in two ecotypes of S. intermedia demonstrated that the two loci have slightly different repeat organization (Hoang et al., 2020). Information on the molecular structure of 5S rRNA genes in S. polyrhiza is limited to direct sequencing of specific PCR fragments for four ecotypes isolated in eastern China, showing a single repeat type composed of a conserved 119-bp 5S rRNA gene and a 400-bp NTS with almost no variation in the nucleotide sequence (Borisjuk et al., 2018). Therefore, further studies are needed to shed light on the organization of the 5S rDNA in diverse S. polyrhiza ecotypes.

Our recent survey of duckweed biodiversity in Ukraine and China (Chen et al., 2022) revealed S. polyrhiza as the dominant duckweed species in both countries. Here, we present data on comparative molecular organization of 5S rDNA repeats in six ecotypes of S. polyrhiza isolated in different regions of Ukraine and China. Sequencing of more than hundred cloned PCR fragments covering at least twelve whole 5S rRNA gene units for each ecotype complemented with estimation of the gene copy number, revealed intra- and inter-genome variability and the contrasting evolutional dynamics of two types of 5S rDNA repeat units in the greater duckweed, S. polyrhiza.

Our recent study of duckweed biodiversity in Ukraine and China identified S. polyrhiza as the most represented duckweed species in both countries (Chen et al., 2022). In our previous analysis of four Chinese S. polyrhiza strains isolated in central-eastern Jiangsu province, direct sequencing of PCR fragments revealed a single unit of 5S rDNA repeats with nearly uniform structure of 119-bp long 5S rRNA genes separated by the 400-bp long NTS (Borisjuk et al., 2018). Genome studies revealed that S. polyrhiza and S. intermedia contain two 5S rDNA loci located on separate chromosomes (Hoang et al., 2019). Previous research in different plant species suggested that different loci usually contain 5S rDNA repeats with different molecular architecture (Schneeberger et al., 1989; Kellogg and Appels, 1995; Simon et al., 2018; Wang et al., 2023). This was confirmed for two ecotypes of S. intermedia where the loci on chromosomes SiChr14 and SiChr15 contain 5S rDNA units composed of identical sequences of 5S rRNA genes separated by NTS sequences, specific for each locus (Hoang et al., 2020). Indeed, despite preferential amplification of a single variant of the 5S rDNA fragments (the short one of about 500 bp for two Chinese ecotypes and one from Ukraine, and the other one of about 1100 bp for the two ecotypes from Ukraine and one from China ( Figure 2B ) using the standard PCR protocol applied earlier (Borisjuk et al., 2018), optimizing the reaction parameters resulted in amplification of both types of 5S rDNA repeats using genomic DNA of each ecotype as template ( Figure 2C ).

The strategy of sequencing multiple individual clones containing 5S rDNA units for each genotype allowed us to gain insight into the intra-genomic variability of gene organization in the analyzed ecotypes. Nucleotide alignments demonstrated high conservation of the 119-bp gene encoding 5S rRNA, with 74 out of 92 obtained gene sequences belonging to a single ribotype. These results confirm our earlier data obtained for two other duckweed species, S. intermedia (Hoang et al., 2020) and Landoltia punctata (Chen et al., 2021), and the more distantly related aquatic plant Pistia stratiotes (Stepanenko et al., 2022), and agree with the general conservation of the 5S rDNA sequence in plants (Park et al., 2000; Hemleben et al., 2021; Tynkevich et al., 2022). However, the 2-D secondary structure model built for the S. polyrhiza, following a stem–loop architecture that is typical for plants, showed a characteristic difference in the β domain ( Figure 3 ). This difference in the β domain was previously observed in Landoltia and Pistia (Chen et al., 2021; Stepanenko et al., 2022) and in species from the Asteraceae family (Garcia et al., 2010). In particular, in addition to a smaller loop C compared to the majority of plants, these species have an additional minor loop between loops B and C.

Regardless of this general conservation, we detected several site-specific SNPs represented by C↔T or G↔A transitions in eighteen 5S rRNA genes, grouping the gene sequences into five minor ribotypes in addition to the dominant one. Potentially, the minor ribotypes could be expressed in a tissue- or developmental stage- specific manner, as has been reported for Arabidopsis (Arabidopsis thaliana) (Cloix et al., 2002; Layat et al., 2012; Simon et al., 2018). Moreover, the SNPs represented in ribotypes 2 and 4 and located in the regulatory A-box ( Figure 3 ) could potentially impact the expression of the genes.

Except for those SNPs, we detected no other variation among the 5S rRNA gene sequences of the S. polyrhiza ecotypes, unlike the large number of 5S rRNA pseudogenes reported in the much bigger genomes of the grass Thinopyrum intermedium (Mahelka et al, 2013), hexaploid wheat (Triticum aestivum) (Sergeeva et al., 2017), and some species in the Solanaceae family (Volkov et al., 2017; Tynkevich et al., 2022). We observed a similar high sequence homogeneity among shorter-length Sp-5S-S variants of S. polyrhiza 5S rDNA repeats: only one out of 57 analyzed NTSs of Sp-5S-S with uniform length of 400 bp differed by a 7-bp deletion (clone Sp634-3, Supplementary Figure 2 ). However, in addition to few random nucleotide changes, these NTS sequences contain a number of ecotype-/region-specific SNPs.

Compared to the Sp-5S-S, the longer 5S rDNA units of Sp-5S-L demonstrate significant variability with NTS lengths ranging between 1056 and 1084 bp caused by indels in certain DNA motifs and SNPs in generally conserved core sequences. A significant portion of these changes appeared to be ecotype- or region-specific ( Supplementary Figure 3 ). In particular, the highest number of ecotype-specific SNPs was detected in DW78, represented by six nucleotide transitions and three transversions. The long stretch of repeated GA dinucleotides, representing the binding sites for GAGA-binding proteins (GBPs), which play a role in regulating multiple genome functions (Berger and Dubreucq, 2012; Gong et al., 2018; Sahu et al., 2023), is another characteristic feature revealed in the Sp-5S-L NTS sequences of the duckweed 5S rDNA. The DW78, Sp5548, Sp2014, and SpYa1 ecotypes share an enlarged version of the GA_n element consisting of 11–16 GA dinucleotides, contrasting with the GA_n element in ecotypes DW30 and DW100 containing 6–8 GA repeats.

Alignment of the NTS type-L sequences using phylogenetic tools separated the analyzed ecotypes into two groups ( Figure 5 ), one containing NTSs of Ukrainian ecotype DW78 and Chinese ecotype Sp5548, and the other containing the sequences of ecotypes DW30, DW100, Sp2014, and SpYa1. Such a grouping assumes that the ancestors of each group separated prior to their geographic distribution, most probably due to migrating birds. Following their distribution to different geographic locations, the rDNAs in these groups underwent individual clone evolution and occasional hybridization/recombination, which led to the observed complex pattern of the type II NTSs. In that respect, the simultaneous presence of Sp-5S-S NTS variants with SNPs specific for DW30 and the rest of the analyzed ecotypes ( Figure 4 ), hints at a possible hybridization event.

Although this study was designed to have a balanced representation of Sp-5S-S and Sp-5S-L 5S rDNA sequences, our estimation of copy number of each type of repeat revealed a highly asymmetric distribution of Sp-5S-S and Sp-5S-L rDNA units among the ecotypes. Thus, based on the ratio of Sp-5S-S/Sp-5S-L units ( Table 1 ), the six analyzed ecotypes are clearly divided into two groups with the dominant Sp-5S-S repeats in one group (DW78, Sp2014, and Sp5548) and the Sp-5S-L repeats in another (DW30, DW100, and SpYa1), which agrees with the results of amplification of the 5S rDNA fragments using the non-optimized PCR method ( Figure 2B ). The ratio between the 5S rDNA types among the studied ecotypes widely varied from about 3-fold dominance of Sp-5S-L in ecotype DW30 to more than 6-fold dominance of Sp-5S-S in ecotype DW78, with Sp-5S-L being more prevalent in the Ukrainian ecotypes and Sp-5S-S more common in ecotypes from China. Our data also confirmed the previous estimation (Hoang et al., 2019) that S. polyrhiza has one of the lowest copy numbers of 5S rDNA genes among investigated plants, ranging from 46 to 220 copies per genome of Sp5548 and Sp2014 ecotypes, respectively.

We can also assume that, similar to other plants like Arabidopsis (Tutois et al., 2002; Simon et al., 2018), tobacco (Nicotiana tabacum) (Fulnecek et al., 2002), fir (Abies alba) (Besendorfer et al., 2005), and the duckweed S. intermedia (Hoang et al., 2020), the 5S rDNA units of the same type are to certain extent homogenized and arrayed together at one of the two 5S rDNA loci reported for the genome of S. polyrhiza (Hoang et al., 2018; 2019). Therefore, the different levels of polymorphism between the 5S rDNA Sp-5S-S and Sp-5S-L arrays are rather locus-specific features, as has been previously noticed for Arabidopsis (Simon et al., 2018) and wheat (Sergeeva et al., 2017). Moreover, the two 5S rDNA loci might be differentially regulated, based on the profound NTS differences of the 5S rDNA arrays at different loci, including regions just upstream of the 5S rRNA genes containing TATA-like boxes, which are implicated in transcription (Cloix et al., 2003).

In summary, our data demonstrate a contrasting speed of molecular evolution between the 5S rDNA units depending on the type of their NTS sequences. Our finding of almost no length variability and low SNP frequency in the short-type NTSs is in agreement with the general low mutation rate in S. polyrhiza revealed by whole-genome sequencing of multiple geographically distinct genotypes (Ho et al., 2019; Xu et al., 2019), and the concept of repeated sequence homogenization through the process of concerted evolution (Nei and Rooney, 2005). On other hand, more profound variability of the NTSs of longer Sp-5S-L rDNA units is in contrast with the greater duckweed genome conservation and suggests the influence of some factors suppressing sequence homogenization within the array, as comprehensively discussed in a recent review by Wang et al. (2023). The domination of vegetative propagation might be one of these factors in S. polyrhiza, since repeated meiotic crossovers were proposed as the major driving force of DNA repeat homogenization (Smith, 1974; Dover, 1982), with evidence related to rDNA mounting for different eukaryotic species including plants (Pringle et al., 2000; Fehrer et al., 2021; Macháčková et al., 2022). In addition to revealing that S. polyrhiza has a very low 5S rDNA copy number compared with most eukaryotes, our data demonstrated intriguing copy number dynamics within the arrays composed of different length variants of 5S rDNA units. Combined, the unique features of greater duckweed’s 5S rDNA that we unveiled in this study represent a promising system for studying basic mechanisms of genome organization, function, and evolution. Much novel information can be gained by reliable long-read sequencing (Hotaling et al., 2023) through the 5S rDNA loci and application of modern genomics for functional characterization of the rDNA variants.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/ Supplementary Material .

GC: Validation, Methodology, Investigation, Writing – review & editing. AS: Writing – original draft, Visualization, Software, Formal analysis, Conceptualization, Validation, Methodology. NB: Writing – review & editing, Supervision, Resources, Project administration, Data curation, Writing – original draft, Conceptualization.