Two long non-coding RNAs generated from subtelomeric regions accumulate in a novel perinuclear compartment in Plasmodium falciparum

Abstract

Chromosome ends have been implicated in the default silencing of clonally variant gene families in the human malaria parasite Plasmodium falciparum. These chromosome regions are organized into heterochromatin, as defined by the presence of a repressive histone H3 lysine 9 trimethylated marker and heterochromatin protein 1. Here, we show that the non-coding subtelomeric region adjacent to virulence genes forms facultative heterochromatin in a cell cycle-dependent manner. We demonstrate that telomere-associated repeat elements (TAREs) and telomeres are transcribed as long non-coding RNAs (lncRNAs) during schizogony. Northern blot assays revealed two classes of lncRNAs: a ∼4-kb transcript composed of telomere sequences and a TARE-3 element, and a >6-kb transcript composed of 21-bp repeats from TARE-6. These lncRNAs are transcribed by RNA polymerase II as single-stranded molecules. RNA-FISH analysis showed that these lncRNAs form several nuclear foci during the schizont stage, whereas in the ring stage, they are located in a single perinuclear compartment that does not co-localize with any known nuclear subcompartment. Furthermore, the TARE-6 lncRNA is predicted to form a stable and repetitive hairpin structure that is able to bind histones. Consequently, the characterization of the molecular interactions of these lncRNAs with nuclear proteins may reveal novel modes of gene regulation and nuclear function in P. falciparum.

Introduction

In recent years, heterochromatin at chromosome ends has been the focus of many studies because it plays an important role in regulating gene expression in both yeast and humans [1], [2]. Similarly, in the protozoan Plasmodium falciparum, which is the most virulent human malaria parasite, subtelomeres exert a silencing effect on the clonally variant virulence gene families. The chromosome ends in this pathogen are organized into heterochromatin, as defined by the presence of high levels of the repressive histone H3 lysine 9 tri-methylation (H3K9me3) [3], [4] and heterochromatin protein 1 (PfHP-1) [5], [6]. This heterochromatin is associated with a 1.2-kb non-coding telomere repeat region (GGGTTT/CA) and an adjacent non-coding region that together form a mosaic of six different blocks of repetitive sequences known as telomere-associated repetitive elements (TAREs) 1–6 [7]. Adjacent to TARE-6 are members of several gene families that encode variant surface proteins, such as var, rif, stevor and pfmc-2tm. These gene families are, by default, epigenetically repressed and undergo phenotypic variation through in vitro switches. Expression in a single cell occurs either by mono-allelic expression (var gene members) or by co-transcription of several members (stevor, rif and pfmc-2tm) [7], [8]. The overall chromatin and DNA organization of the chromosome ends is highly conserved among all the chromosomes of the malaria parasite [3], [7], [8]. In addition to the identified heterochromatin factors H3K9me3 and PfHP-1, a number of histone-modifying enzymes and other factors with unknown functions, such as PfSir2, PfOrc1 and PfKMT1, are recruited to subtelomeric regions [3], [6], [9]. Experimental evidence suggests that interactions between these proteins and the telomeric and subtelomeric regions generate a repressive epigenetic center [3], [10]. This repressive center may extend to adjacent virulence gene families to create a molecular platform for the phenotypic variation of virulence factors.

Specialized nuclear compartments are crucial for the function of the Plasmodium nucleus [11], and they are particularly relevant for the mutually exclusive expression of var genes when an active var gene is placed in the var expression site. However, the model that describes this expression site as the limiting factor for the expression of a single var gene needs to be reconsidered because this expression site can harbor more than one transcriptionally active var gene [12]. Finally, fluorescence in situ hybridization assays with DNA (DNA-FISH) have shown that an active episomal rifin promoter colocalizes with a transcriptionally active var promoter, indicating that these gene families utilize the same subnuclear expression site [13]. Therefore, specialized nuclear compartments are important for the regulation of expression of several multi-copy gene families and for rDNA in distinct perinuclear compartments [3], [10], [14], [15].

Although multiple epigenetic factors contribute to the coordinated expression of virulence genes, the role of non-coding RNAs (ncRNAs) remains, for the most part, unexplored in P. falciparum. Initial evidence indicating a role for ncRNAs in malaria parasites was provided by Kyes et al. in 2003, who described var-associated ncRNAs [16]. Later, Deitsch and colleagues identified non-coding transcripts in P. falciparum from centromeric repeats [17]. In addition, the same group showed that an intron-derived ncRNA is enriched in the chromatin of var genes, but no functional data were presented [18]. More recently, a genome-wide study in P. falciparum revealed the presence of short transcripts derived from non-coding subtelomeric regions [19], [20], suggesting that the heterochromatic chromosome ends can be transcribed. A more recent study using microarray analysis suggested that long ncRNA (lncRNA) are produced from TARE-3 repeats [21].

In the present study, we used northern blot, FISH and run-on analyses to demonstrate that P. falciparum telomeres and non-coding subtelomeric repeat regions are transcribed as heterogeneous long ncRNAs (lncRNAs) by RNA polymerase II. We found that only the sense strand is transcribed in asexual blood stages. Transcription peaks occur during the schizont stage and produce two lncRNAs: a ∼4 kb lncRNA that contains TARE-3 and telomere sequences (lncTARE-3-Telomere) and a >6 kb lncRNA that consists of 21-bp repeats from TARE-6 (lncTARE-6). After transcription, both lncRNAs relocate to a single novel nuclear compartment that is distinct from the chromosome ends. Furthermore, the data from gel shift, super gel shift and northwestern blot assays suggest that this lncRNA may form large networks of interlinked RNA–protein complexes whose functions are unknown.