Figure 1.
Mating in Oxytricha trifallax leads to production of a class of 27 nt RNAs.
Total RNA was purified from vegetative Oxytricha trifallax (lanes 1–4) or at various time points after mixing together complementary mating strains (lanes 5–11). In addition, total RNA was purified from strain ALXC9 treated under identical conditions as a mating for 24 hrs (lane 12 - Mock 24hr). Total RNA was phosphatase treated followed by 5′ end labeling with 32P, separated on a 15% polyacrylamide denaturing gel and visualized using a PhosphorImager. Sizes from Decade RNA 10 nt Ladder (Ambion) are indicated at left. Positions of small RNAs of interest are indicated at right. Lanes 1–11 directly correspond to RNA preparations used to prepare libraries for Illumina sequencing listed in the same order in Table 1.
Figure 2.
The mating-specific 27 nt RNAs in Oxytricha trifallax are not modified at their 3′ ends.
RNAs were tested with the beta elimination assay in order to determine if there is a modification at the 3′ end. RNAs tested are 32P 5′ end labeled total Oxytricha trifallax RNA (lanes 2–11) or 5′ end labeled synthetic positive control RNAs containing the sequence of the C. elegans lin-4 (21 nt) or mir-90 (23 nt) microRNAs which were subsequently mixed with 1.5 µg of unlabeled Oxytricha total RNA prior to beta elimination (lanes 13–16). Control untreated samples (-) or samples subjected to the beta elimination reaction (+) are indicated. RNAs were isolated from vegetative ALXC2 (lanes 2 and 3) or from a mating between ALXC2 and ALXC9 that were harvested at the indicated timepoint after mixing (lanes 4–11). Beta elimination will remove the terminal ribose if both a free 2′ OH and 3′ OH are present, and will leave a 3′ cyclic phosphate. This loss of a nucleotide and the presence of the extra phosphate group will result in the treated RNAs running almost two bases faster than their untreated counterparts.
Figure 3.
Flowchart of sequence analysis.
Table 1.
Oxytricha trifallax small RNA sequencing mapping statistics.
Figure 4.
Vegetative small RNA libraries contain mostly 20–22 nt RNAs and mating libraries contain mostly 27 nt RNAs.
The length distributions of sequencing reads in the small RNA range (18–30 nt) are plotted for five of the 11 libraries (comprising one replicate from each cell type or timepoint). The histograms are for distinct reads; only one occurrence of a sequence that appears more than once in a library is counted. The five representative libraries shown can be cross-referenced to table 1; VegALXC2 - veg02_01, VegALXC9 - veg09_04, Mat24hr - mat24_06, Mat48hr - mat48_08, Mat72hr - mat72_11. Black bars represent reads that mapped to the macronuclear sequence assembly. White portions of the bars represent reads that did not map to the macronuclear assembly. The vegetative strain VegALXC2 is capable of self-mating, leading to a minor peak at 27 nt.
Figure 5.
Nucleotide position bias in the different classes of sequenced small RNAs.
These charts show the nucleotide frequency at each position for different small RNA size classes. The distinct sequences in the veg09_03 library and the mat24_05 library were filtered against non-coding RNAs. The distinct RNAs that made it through the filter were selected by size and the nucleotide composition of each position for the indicated size class was determined. A. Distinct 20 mers from the veg09_03 library (222,272 sequences). B. Distinct 21 mers from the veg09_03 library (416,2227 sequences). C. Distinct 22 mers from the veg09_03 library (327,422 sequences). D. Distinct 27 mers from the mat24_05 library (2,919,225 sequences). E. Abundant 27 mers from mat24_05 library comprising distinct sequences that were detected 10 or more times in the library (17,981 sequences).
Table 2.
Venn Diagram analysis of 26–28 nt small RNA alignment to macronuclear/micronuclear sequence pairs - evidence for a macronuclear origin for the mating-specific small RNAs.
Figure 6.
27 nt RNA reads from different mating libraries align to both strands of a nanochromosome.
This screen shot from the Oxytricha trifallax macronuclear genome build of the UCSC Genome Browser shows the macronuclear nanochromosome corresponding to the alpha telomere binding protein gene indicated as a red bar (Nano42874.1) with the telomere sequences at the beginning and end of the nanochromosome noted as small black boxes in the track above that. The top track shows the locations of groups of MDS sequences or individual MDSs, as aligned using BLAT of the micronuclear sequence against the macronuclear genome. The alignments of 26–28 nt small RNAs from 7 different mating small RNA libraries (mat24_05, mat24_06, mat30_07, mat48_08, mat48_09, mat55_10 and mat72_11 - see Table 1) are shown with alignments to the plus strand of the nanochromosome indicated in blue and alignments to the minus strand of the nanochromosome indicated in orange. Numbers at the left indicate the total number of 26, 27 and 28 nt RNA sequences from each of the seven libraries that mapped to the macronuclear genome.
Figure 7.
Histogram of unadjusted p-values for changes in relative expression level for 27 mers between early and late mating libraries.
DESeq [35] was used to compare relative expression levels of distinct 27 mers whose sequence was found 10 or more times in any of the mating libraries. The relative expression of each of 57,140 abundant 27 mers between early mating time points (mat24_05, mat24_06 and mat30_07) and later mating time points (mat55_10 and mat72_11) was compared. Unadjusted p-values for changes in expression for each of the 27 mers were plotted on the histogram.
Figure 8.
27 nt small RNAs are macronuclear in origin and do not require intron removal for their generation.
A. Schematic diagram of internally eliminated sequence (IES) removal and macronuclear destined sequence (MDS) joining during macronuclear development. Pointer sequences are direct repeats found in the micronucleus at MDS borders. Only one copy of the pointer is found in the macronucleus. B. Oxytricha trifallax Macronuclear Genome Browser screen shot of a region of the alpha telomere binding protein gene showing the junctions and overlapping pointer sequences of MDSs 9, 10 and 11. Only the mat24_06 library 26–28 nt small RNA track is shown. Note that the 26–28 nt small RNAs from the mat24_06 library from both strands overlap the MDS junction and pointer sequences. This is consistent with these small RNAs having originated from the mature parental macronucleus. C. Screen shot of a region of the alpha telomere binding protein gene containing its intron. 26–28 nt small RNAs from all seven mating libraries are shown. Note that sense strand small RNAs are found within the intron, indicating that they could be processed from introns. Also, small RNAs in the sense strand are found crossing the downstream intron/exon border, indicating that these are made from an RNA that did not undergo intron processing.
Figure 9.
Screen shots of four different macronuclear genes from the Oxytricha macronuclear genome browser.
In each panel, black boxes on top are a BLAT alignment of individual MDS sequences to show MDS junction location. Short match below that shows telomeric repeat locations. The red bar below that shows the nanochromosome contig extent. Blue and orange bars indicate the alignment location of 26–28 nt long small RNAs from the mat24_06 library. Plus strand alignments are in blue and minus strand alignments are in orange. A. DNA Polymerase Alpha. Note that this nanochromosome is incompletely assembled and is spread across three partially assembled contigs on chr2. B. Beta Telomere Binding Protein. C. Actin-I. D. CCCH Zinc Finger protein.
Figure 10.
Analysis of the distribution of 27 nt RNAs on macronuclear nanochromosomes.
A. 27 nt RNA small RNAs are produced in equal numbers from both strands of the nanochromosome. For each full length nanochromosome with at least ten 26–28 nt small RNAs aligning to each strand, the total number of 26–28 nt small RNAs that map to each strand was plotted (left graph) or the number of distinct 26–28 nt small RNAs that map to each strand was plotted (right graph). Pearson R correlation values of 0.91 for all reads and 0.94 for distinct reads were obtained. This indicates a strong correlation of small RNA production from one strand of the nanochromosome with production of small RNAs from the other strand of the nanochromosome. Red curved lines represent 2 standard deviations from the mean; 95% of points would be expected to fall within these regions if there is a one-to-one correlation between the number of 26–28 nt small RNAs aligning to each strand of a nanochromosome (thin black line along the main diagonal). The blue dotted line indicates the actual lines within which 95% of the data points fall. B. There is a non-uniform distribution in the positioning of small RNAs on the nanochromosomes. For each position in each complete nanochromosome, the number of 26–28 nt small RNAs that start at that position in the mat24_06 sequencing library were determined. Then we determined the mean and standard deviation of coverage density on the nanochromosome. The coefficient of variation (standard deviation of coverage divided by the mean coverage) is plotted for each nanochromosome in the histogram. The peak of coefficient of variation at ∼6.0 indicates that the standard deviation is 6.0 times greater than the mean. This is highly indicative of a non-uniform distribution of small RNA coverage on the nanochromosomes.
Figure 11.
The first 30 nt of the nanochromosomes proximal to the telomeres have >8-fold lower coverage of 26–28 nt small RNAs relative to the rest of the nanochromosome.
Graphic at the top shows a nanochromosome (rectangle) with 20 nt telomere sequence in gray shadow. 27 nt small RNAs for the plus strand of the nanochromosome are indicated as arrows above the nanochromosome. The average density of the location of 5′ ends of plus strand-aligning 26–28 nt RNAs was determined over 500 positions (in 20 bins of 25 bp each, shown by tick marks) relative to the 5′ (right side, positive values) and 3′ (left side, negative values) ends of 10,002 complete nanochromosomes. These densities were plotted relative to a uniform distribution of the aligned 26–28 nt RNAs over the same sets of 500 positions on the same nanochromosomes. A dotted line at 20 nt from either end is included on both sides of the zero point to indicate the average telomere length of 20 nt on the nanochromosome sequences to which these data were plotted. The shaded gray area, 27 nt wide, is included on the plot from the nanochromosome 3′ end because the plus strand-aligning reads would have 27 nt of sequence between their plotted 5′ end and the 3′ end of the nanochromosome. When minus strand-aligning reads were analyzed by this same method, a mirror image plot was obtained that is otherwise identical to the graph for plus end reads (not shown).