Organization of the Caenorhabditis elegans small
non-coding transcriptome: Genomic features,
biogenesis, and expression
Wei Deng,1Xiaopeng Zhu,1,5Geir Skogerbø,2Yi Zhao,2Zhuo Fu,1Yudong Wang,1
Housheng He,1,5Lun Cai,2,5Hong Sun,1,5Changning Liu,2,5Biao Li,1,3,5Baoyan Bai,1,5
Jie Wang,1,5Dong Jia,1Shiwei Sun,2,5Hang He,1,5Yan Cui,4Yu Wang,6Dongbo Bu,2
and Runsheng Chen1,2,3,7
1Bioinformatics Laboratory, Institute of Biophysics, Chinese Academy of Sciences, Beijing 100101, China;2Bioinformatics Research
Group, Key Laboratory of Intelligent Information Processing, Institute of Computing Technology, Chinese Academy of Science,
Beijing 100080, China;3Chinese National Human Genome Center, Beijing 100176, China;4Department of Molecular
Sciences/Center of Genomics and Bioinformatics, University of Tennessee Health Science Center, Memphis, Tennessee 38163,
USA;5Graduate School of the Chinese Academy of Science, Beijing 100080, China;6School of Oncology of Peking University,
Beijing Cancer Hospital, Beijing 100036, China
Recent evidence points to considerable transcription occurring in non-protein-coding regions of eukaryote genomes.
However, their lack of conservation and demonstrated function have created controversy over whether these
transcripts are functional. Applying a novel cloning strategy, we have cloned 100 novel and 61 known or predicted
Caenorhabditis elegans full-length ncRNAs. Studying the genomic environment and transcriptional characteristics have
shown that two-thirds of all ncRNAs, including many intronic snoRNAs, are independently transcribed under the
control of ncRNA-specific upstream promoter elements. Furthermore, the transcription levels of at least 60% of the
ncRNAs vary with developmental stages. We identified two new classes of ncRNAs, stem–bulge RNAs (sbRNAs) and
snRNA-like RNAs (snlRNAs), both featuring distinct internal motifs, secondary structures, upstream elements, and
high and developmentally variable expression. Most of the novel ncRNAs are conserved in Caenorhabditis briggsae, but
only one homolog was found outside the nematodes. Preliminary estimates indicate that the C. elegans transcriptome
contains ∼2700 small non-coding RNAs, potentially acting as regulatory elements in nematode development.
[Supplemental material is available online at www.genome.org. The sequence data from this study have been
submitted to GenBank under accession nos. AY948555–AY948719.]
Small non-protein-coding RNAs (ncRNAs) such as microRNAs
(miRNAs) have received increasing attention as regulatory fac-
tors shaping cellular and organismal life. A catalog of reported
miRNAs from various eukaryotes now covers 400 different RNA
species (Liu et al. 2005). In Caenorhabditis elegans, 120 miRNAs
have been identified experimentally and computationally. How-
ever, the remainder of the C. elegans small non-coding transcrip-
tome has received far less attention (Stricklin et al. 2005). So far
the C. elegans spliceosomal snRNAs and a small number of box
C/D snoRNAs (Thomas et al. 1990; Higa et al. 2002; Wachi et al.
2004) have been experimentally verified, whereas other common
eukaryotic ncRNAs (RNase P RNA, Y RNA, SRP RNAs) have only
been identified as putative loci in the genomic sequence (Harris
et al. 2003; Stricklin et al. 2005). Investigations into the small
non-coding transcriptome of several eukaryote model organisms
have invariably revealed several ncRNAs that could not be as-
signed to any known functional class (Huttenhofer et al. 2001;
Marker et al. 2002; Yuan et al. 2003). A recent study of the protist
Disctyostelium discoideum found 16 sequences (out of a total of 36
novel ncRNAs) that, based on conserved internal and upstream
motifs, could be assigned to one of two novel classes of small
non-coding RNAs (Aspegren et al. 2004). Together, these studies
clearly indicate the presence of a still unexplored segment of the
eukaryote small non-coding transcriptome.
One particularly intriguing aspect of the non-coding tran-
scriptome is its potential to fill the regulatory gap created by the
surprisingly low number of protein-coding genes in higher or-
ganisms. Between one-celled yeast, thousand-celled nematodes,
and trillion-celled mammals, there is a difference of a mere 6000
to 19,000 to 25,000 in protein-coding gene numbers; regulation
by non-coding RNA has been suggested to account for this dis-
crepancy (Mattick 2003, 2004). The recent census of the known
and verified ncRNA population stands at around 5300 different
species (Liu et al. 2005), and may be as high as 20,000 if mRNA-
like transcripts are included (Pang et al. 2005), of which a con-
siderable fraction has been ascribed regulatory roles. High-
density microarray analyses indicate a large quantity of transcrip-
tionally active DNA outside annotated or predicted protein-
coding regions (Okazaki et al. 2002; Yamada et al. 2003; Bertone
et al. 2004; Cawley et al. 2004; Kampa et al. 2004; Stolc et al.
2005). This would bring the number of non-coding transcripts
into several tens of thousands, if a substantial fraction of the
E-mail email@example.com; fax 86-10-64889892.
Article published online ahead of print. Article and publication date are at
16:20–29 ©2006 by Cold Spring Harbor Laboratory Press; ISSN 1088-9051/06; www.genome.org
transcriptionally active DNA actually codes for functionally ac-
tive transcripts. The significance of these mostly non-conserved
transcripts, however, still remains highly contentious (Wang et
al. 2004; Huttenhofer et al. 2005).
We have attempted to map out part of the small non-coding
transcriptome of C. elegans. Except for miRNAs, we have retrieved
transcripts of nearly all predicted and previously verified small
ncRNAs in C. elegans in addition to 100 novel ncRNAs. Of these
novel ncRNAs, one-third do not belong to any known functional
class of ncRNAs, and at least two groups of transcripts display
conserved features indicative of two novel functional classes. The
genomic organization of the C. elegans small ncRNAs is also pe-
culiar in two aspects. Contrary to other metazoans, several of the
snoRNA genes appear to be transcribed from independent pro-
moters, of which at least one type bears little sequence resem-
blance to previously described ncRNA promoters (Thomas et al.
1990). Moreover, a considerable fraction of apparently indepen-
dently transcribed ncRNA genes (of various functional classes) is
located in introns of protein-coding genes, a type of genomic
organization seldom observed in other organisms.
We used a new strategy to construct an ncRNA-specific full-
length library of C. elegans from mixed-stage worms and eggs (see
Methods). Our procedure ensured a substantial fraction of full-
length clones with defined 5?- and 3?-termini, while allowing us
to distinguish between 5?-capped and uncapped transcripts. The
number of fragments of unwanted RNAs (rRNAs, tRNAs, and
mRNAs) was also significantly reduced compared to previous ef-
forts (see Discussion). Full-length cDNA libraries of both capped
and uncapped transcripts were established, and altogether 2178
clones were sequenced (Fig. 1A). Thirty-six percent of the clones
in our library represented 100 novel transcripts with confirmed
expression (Northern blot). Another 34% contained 61 different
representatives of all known and predicted families of short
ncRNAs in C. elegans, indicating that the cloning strategy had
very high sensitivity. Only four potential novel RNAs were dis-
carded because of lack of detectable expression or defined 5?- and
3?-ends. Northern blot analyses of different developmental stages
and environmental condition revealed substantial variation in
expression of at least 60% of the novel transcripts.
Genomic location of the novel ncRNAs
The genetic loci of the novel transcripts show no particular chro-
mosomal distribution, but a major fraction (55%) is oriented in
the sense direction within the transcribed region of a known or
predicted gene, intronic loci constituting the larger part of these.
Among functionally annotated genes hosting one or more in-
tronic ncRNA loci, 30% code for a ribosomal protein, and a con-
siderable fraction (24% of all novel loci) is also located within a
C. elegans operon (see Supplemental material). Three loci are
found in UTRs of coding genes, and one locus (cen42) overlaps
the junction between a coding exon and the following intron.
Ten loci (10%) are located antisense to introns of protein-coding
genes, and 36% of the novel loci have intergenic positions.
The data reveal two aspects of the genomic organization of
small ncRNA genes specific to C. elegans. The first is that the large
fraction of intronic ncRNA loci does not only include snoRNAs,
but also snRNA, SL2 RNA, and SRP RNA loci. Secondly, a consid-
erable fraction of the snoRNAs detected has intergenic loci. Of
the 89 snoRNA loci in our material, 30 are intergenic, and with a
few exceptions, these 30 loci also display conserved upstream
motifs (see below).
Functional distribution of the ncRNAs
Of the 161 known and novel ncRNAs detected, roughly one-half
correspond to snoRNA-like transcripts, one-fourth have other
known ncRNA functions, and the remaining one-fourth have no
obvious cellular function (Fig. 1B). The 100 novel ncRNAs fall in
two major categories, novel snoRNA-like transcripts (69) and
transcripts not belonging to any previously known functional
class (31). Among the snoRNA-like transcripts, 42 have conserved
sequence elements and secondary structure (data not shown) re-
sembling the box H/ACA subclass, whereas 27 transcripts have
been classified as box C/D snoRNAs. One box H/ACA snoRNA
was found to contain conserved elements common to the verte-
brate snR30 snoRNA involved in rRNA processing (Atzorn et al.
C. elegans ncRNAs are conserved in Caenorhabditis briggsae
Of our 100 novel ncRNAs, 70 have recognizable counterparts in
C. briggsae with >40% sequence identity. The average identity of
quenced library clones on different RNA species and categories. The E. coli
RNAs are contaminants from food bacteria ingested by C. elegans.
(tRNAscan) tRNAs detected by tRNAscan (Lowe and Eddy 1997);
(mtRNA) mitochondrial RNA. (B) Functional distribution of all novel and
known ncRNAs detected in this study. Sectors representing RNAs of un
defined and novel functional classes are hatched or gridded. misc RNA
includes RNase P RNA and Y RNA.
Clonal and functional distributions. (A) Distribution of se-
The non-coding transcriptome of C. elegans
the transcribed novel ncRNAs sequences (60%) is around thrice
that of the immediate flanking 5? and 3? sequences (16.1% and
21.5%, respectively). Nearly one-third of the novel, functionally
unknown ncRNAs (10 out of 31), along with 22 novel snoRNA-
like transcripts, are >85% identical in C. briggsae, a conservation
level on par with that of the spliceosomal snRNAs. One transcript
(CeN96) was identified by two short conserved sequence motifs
as a C. elegans homolog of the yeast/vertebrate snR30/U17
snoRNA required for processing of 18S rRNA (Atzorn et al. 2004).
Aside from CeN96, no sequence homologs of the novel ncRNAs
could be found in other species.
A majority of the C. elegans ncRNAs show developmentally
Using Northern blots, we observed transcription levels of 20
known or predicted and 86 novel ncRNA families (comprising 44
and 87 different transcripts, respectively). Total RNA was ex-
tracted at 12 different developmental stages from egg to mature
adult, in dauer worms and after heat-shock treatment. Hybrid-
ization signals from Northern blots were recorded to obtain
quantitative estimates of the ncRNA expression levels. Sixty of
the ncRNA families (covering 61 ncRNA species) showed varia-
tions in expression exceeding two standard deviations at certain
developmental stages or environmental conditions. An addi-
tional 40 families (57 species) showed similar tendencies of regu-
lated expression at a lower level of variation. Only six ncRNA
families were devoid of variation in expression over all condi-
tions tested. Among these were four of the five spliceosomal
snRNA families tested (the fifth, snRNA U5, displayed a slight
[I = 1.08] tendency to elevated transcription in dauer worms) (see
Supplemental material for details).
Clustering the ncRNAs according to similarities in expres-
sion produced six distinct profiles comprising altogether 72 tran-
scripts (Supplemental Fig. S-1). Eight ncRNA families, the major-
ity of which are novel ncRNAs of unknown function, displayed
significantly elevated expression toward the later worm stages.
While only one ncRNA showed significantly elevated expression
at early development stages, at least 11 others showed slight in-
creases. Several ncRNAs, mostly snoRNA-like transcripts ex-
pressed from the UM2 containing loci, showed highest expres-
sion during middle stages of development (see below for details).
The dauer worm cluster was particularly important because
28 different snoRNAs showed significantly elevated expression at
this stage. Conversely, all snRNAs showed unaltered expression
in dauer, indicating that the data reflect physiologically relevant
ncRNA levels. Because most of the dauer-expressed transcripts are
snoRNA genes, probably transcribed with and processed from
their host gene transcripts, it is also possible that the elevated
expression levels of the ncRNAs in the dauer state represent al-
tered activation of their host genes, or differential processing of
pre-mRNAs to mRNAs or to snoRNAs (de Turris et al. 2004). Six
ncRNAs with significantly elevated expression in starved L1 (L1s)
larvae also showed high expression in dauer worms, possibly in-
dicating a relation to hunger stress. Heat-shock treatment (30°C
for 3 h) increased expression above twofold in three ncRNAs, two
of which also displayed high dauer expression, suggesting a role
in stress-related regulatory processes.
Two novel functional classes, sbRNAs and snlRNAs
Analysis of the 31 novel ncRNAs that could not be assigned to
any known functional class identified two new functional classes
of ncRNAs. According to secondary structure features and inter-
nal motifs, we labeled the respective groups stem–bulge RNAs
(sbRNAs) and snRNA-like RNAs (snlRNAs).
The stem–bulge RNA class (sbRNAs), consisting of nine
novel transcripts, is characterized by two distinct motifs (IM1
and IM2) located at the 5?- and 3?-termini (Fig. 2A,B). These con-
served motifs have the potential of forming a double-stranded
helix featuring a bulge with a conserved sequence (AACUU) sepa-
rated by a single-stranded RNA loop of varying length (Fig. 2C).
The construct could serve as a binding site for common protein/
protein complexes. Searches for combinations of the IM1 and
IM2 sequence revealed four additional high confidence hits in
the C. elegans genome, and 11 in the C. briggsae genome (data
not shown). All sbRNAs share a common upstream motif (UM3,
see below) including a TATA-box. A few of the sbRNA genes
(e.g., cen73 and cbp9) are obvious sequence and structural ho-
mologs. But for most of the C. briggsae sbRNAs, only IM1 and IM2
are conserved, whereas the intervening loop shows little se-
quence conservation. No sequence homologs of sbRNAs were
found in any other organism; however, the recently discovered
Class I ncRNAs in Dictyostelium shows a similar pattern of inter-
nal stem-forming conserved 5?- and 3?-end motifs (Aspegren et
al. 2004), possibly indicating some relationship to the nematode
sbRNAs. An additional aspect of the sbRNA loci is that they fre-
quently occur in clusters of two or three closely located genes,
reflecting perhaps a dependence on a common DNA environ-
ment. The expression levels of most sbRNAs vary considerably
during the course of C. elegans development, with a tendency
toward higher expression in later stages of worm development
The second novel class, snlRNAs, consisting of eight novel
transcripts, is characterized by the presence of internal motif 3
(IM3) (Fig. 3A; Supplemental Table S-2). The IM3 sequence in-
cludes the conserved sequence AARUUUUGGA reported as the
Sm protein-binding site in spliceosomal snRNAs (Riedel et al.
1987), but is at least 20 nt longer than a traditional Sm protein-
binding site, and may thus contain binding sites for additional
proteins (Fig. 3D). The IM3 sequence is mostly located in the
3?-terminal half of the transcripts (Fig. 3B), and is also found in
all SL2 snRNAs and the predicted U3 snoRNAs. The presence of
the Sm binding site within IM3 possibly indicates that snlRNAs
have a role in splicing or other Sm related functions. Similar to
snRNAs, several of the snlRNAs appear as multi-copy gene fami-
lies consisting of two or more ncRNAs with minor sequence dif-
ferences. With a few exceptions, the snlRNA loci share the same
upstream motif (UM1, see below) with C. elegans spliceosomal
Like spliceosomal snRNAs, most snlRNAs appear to have
high expression levels, which tend to increase even more toward
later stages of worm development. The snlRNA CeN31 is particu-
larly intriguing, showing high expression both in the egg/
embryo stage and in the mature adult, whereas intervening
stages all have low expression (Fig. 3C). This class of ncRNAs
correlates strongly with several mRNAs in C. elegans (Wang and
Kim 2003), including two of the seven cyclin genes (Fig. 3C),
suggesting involvement of these ncRNAs in cell division.
ncRNA-specific upstream motifs
Analysis of the 100-bp 5?-end flanking sequences of the novel
ncRNAs using MEME (Bailey and Elkan 1995) yielded three dis-
tinct 50-bp upstream motifs (UM1–3) (Fig. 4). The first of these,
Deng et al.
22 Genome Research
upstream motif 1 (UM1), is found at 84 loci, including all but two
spliceosomal snRNA loci, and contains a highly conserved core
sequence that overlaps the snRNA proximal sequence element
(PSE) previously identified in C. elegans (Thomas et al. 1990).
UM1 was also found at six new box C/D snoRNAs loci and at a
majority of the spliced leader (SL) RNA loci.
Upstream motif 2 (UM2) was found at the loci of 47 ncRNAs,
of which 40 (31 novel) had snoRNA-like features. The remaining
seven transcripts could not be assigned to any known class of
ncRNAs. UM2 is on average located 20 bp closer to the transcript
5?-end than UM1, and differs substantially from UM1 in base
composition (Fig. 4D). UM2 contains no highly conserved core
motif. Rather, the most conserved bases are concentrated toward
both ends of the motif. An interesting feature of these two con-
served ends is that they show considerable resemblance to the
box A and box B components of the internal tRNA promoter.
There are no tRNA annotations upstream of any of the 198 loci of
our data set. However, four pseudo-tRNA annotations occur up-
stream of four UM2-containing loci (see Supplemental material
Upstream motif 3 (UM3) is shared by all C. elegans verified
and predicted sbRNA loci, and was also found at all predicted C.
briggsae sbRNA loci. The motif is composed of a highly conserved
central core sequence followed invariably by a canonical TATA-
box located at approximately ?30 bp (Figs. 4C and 5D). The
UM3 core shares a small submotif (TGTCNG) with the UM1 core
sequence, but otherwise the two motifs show little sequence
similarity. Based on their upstream motifs, their genomic loca-
tions, and several other factors, we suggest that the 161 known
and novel ncRNAs detected in our study can be divided into
several different biogenesis groups (Fig. 5 and Supplemental ma-
ncRNA cap structure correlates to upstream motifs
Our ncRNA library construction strategy was designed to dis-
criminate between 5?-end capped and non-capped transcripts. Of
the 161 known and novel ncRNAs detected, 52 had more than a
95% probability of carrying a 5?-end cap, and an additional 11
were found to have more than an 80% probability of being
capped (Supplemental Table S-1). Most capped transcripts origi-
nated from the TATA-less UM1 loci, or from intergenic loci with-
out a discernible upstream motif.
The 50 TATA-less UM1 ncRNAs had either positive (42) or
undetermined (8) cap status, suggesting a strong relationship be-
tween cap structure and this upstream motif. In our data, U6
snRNA also showed strong indications of a 5?-end cap, probably
implying post-transcriptional processing of a ?-monomethyl-
GTP cap previously found on human U6 snRNA (Gupta et al.
1990). Ninety-one transcripts had more than an 80% probability
(62 had a >95% probability) of not carrying a cap. These were
derived mainly from intronic loci without upstream motifs, or
from both intergenic and intronic loci with upstream motifs
other than UM1.
motifs of the sbRNAs (see Supplemental material for details on the E-value). (B) Relative positions of IM1 and IM2 within each of the verified sbRNAs.
(C) Predicted (Mfold) secondary structure of sbRNAs CeN73-1 and CeN76. (D) Relative expression of seven of the sbRNAs at heatshock (HS) and different
developmental stages (Egg through Dauer; see complete list of abbreviations in Supplemental material).
The stem–bulge RNAs of C. elegans. (A) Sequential composition of the 5?-end (IM1; E = 1.2 ? 10?19) and 3?-end (IM2; E = 1.0 ? 10?20)
The non-coding transcriptome of C. elegans
Intronic loci show signs of independent transcription
A feature of the C. elegans small non-coding transcriptome not
found in other organisms (Dominski et al. 2003) is the high fre-
quency of conserved motif elements upstream of intronic ncRNA
loci. Of the 198 different chromosomal loci corresponding to 161
known and novel ncRNAs, 88 are located in the sense direction
within an intron of a verified or predicted protein-coding gene.
More than 50% of these have conserved upstream sequences
(mainly UM1 and UM2) (Fig. 5F). Comparing these intronic “mo-
tif loci” to intronic ncRNA loci without a discernible upstream
motif (“non-motif loci”), we found several striking differences
that further suggest independent transcription of intronic motif
loci from host genes. Most transcripts from intronic UM1 loci
appear to carry 5?-end caps, indicative of independent RNA poly-
merase II transcription. The intervening sequence between the
ncRNA 5?-end and the preceding exon is also generally more
A/T-rich and its median size much shorter at non-motif loci (35
nt) than at UM1 (253 nt) and UM2 (152 nt) loci (Supplemen-
tal Fig. S-2). More than 60% of the intronic non-motif loci (all
snoRNAs) reside in ribosomal or other translational related
genes (Supplemental Table S-1). In other organisms, such genes
are known to have a high frequency of the TOP-type pro-
moters, known to control the processing of cotranscribed, in-
tronic snoRNAs (de Turris et al. 2004). Only 5% of the host genes
with intronic UM2 loci belong to this category of genes. Also,
when comparing the cellular levels of ncRNAs (represented by
library clones) and their host gene mRNAs (represented by public
EST data), we observe a correlation between expression levels of
motif-free ncRNAs and host gene mRNAs. A similar trend was not
found for motif-containing loci, suggesting that ncRNA expres-
sion is unrelated to their host genes’ transcription (Supplemental
High frequency of independently transcribed snoRNA loci
The data also indicate a high degree of independently transcribed
snoRNAs in C. elegans. Whereas independent transcription of
sno/scaRNAs is common in plants (Brown et al. 2003), most
metazoan sno/scaRNA genes are intronic, and only a few inter-
genic loci have been reported to exhibit independent transcrip-
tion (Tycowski et al. 2004). The snoRNAs in our material map to
89 loci, of which 30 are intergenic and (with two exceptions)
associated with either UM1 or UM2. Of the remaining 59 in-
tronic snoRNA loci, four contain UM1 and 19 contain UM2, sug-
gesting that close to 60% of all snoRNAs in C. elegans are, in fact,
Estimates of 2700 ncRNA genes
Given established and novel ncRNA data, we tried estimating the
size of the C. elegans non-coding transcriptome in three ways.
The first estimate was based on the finding that, with respect to
conservation in C. briggsae, the ncRNA-containing introns show
a radically different distribution from other introns (see Supple-
mental material for details). Assuming that the distribution of
known ncRNA-containing introns is representative of all introns
hosting an ncRNA locus, we took introns ranging from 50 to 130
bp to represent introns without an ncRNA locus (no ncRNA in C.
elegans was found in an intron shorter than 135 bp). Using linear
regression analysis, we inferred the number of ncRNA-containing
Relative expression of five snlRNAs and two C. elegans cyclin mRNAs (T06E6.2 and F43D2.1; Wang and Kim 2003). (D) Comparison of the secondary
structures of snlRNA CeN25-1 and snRNA U1. In addition to the Sm-binding site, both RNAs show a similar 3?-tail stem–loop structure, but the remainder
of the IM3 motif is absent in U1 snRNA.
The snlRNAs of C. elegans. (A) Sequence of the IM3 motif of snlRNAs (E = 1.8 ? 10?37). (B) Internal position of IM3 in the snlRNAs. (C)
Deng et al.
introns to be between 0.9% and 2.3% of the total intron popu-
lation, somewhat depending on the set of introns used for the
non-ncRNA-containing set. When extrapolated to accommodate
also for possible intergenic ncRNAs, a C. elegans non-coding tran-
scriptome of 1600–4100 different species is obtained, with 2385
as the most likely estimate (Table 1).
A second approach used the observation that a considerable
fraction of both known and novel ncRNA loci harbors conserved
upstream elements. Performing Meta-MEME (Grundy et al. 1997)
searches for the conserved small ncRNA upstream motifs across
the entire C. elegans genome yielded 1404, 527, and 65 sites (cut-
off: E < 0.1) for UM1, UM2, and UM3, respectively. However,
because a relatively large fraction of both UM2 and UM3 “ge-
nomic hits” overlapped with exons of protein-coding genes, and
many additional UM2 sites overlapped with a considerable num-
ber of tRNA and pseudo-tRNA genes (see above and Supplemen-
tal material), we based the estimate on the occurrences of UM1
only. Accounting for the fraction of ncRNA genes in our library
not having UM1, we obtained an estimate of 2757 different
ncRNA species in the C. elegans small non-coding transcriptome
A third approach, used the correlation between the number
of sequenced clones of each ncRNA and their cellular concentra-
tions as observed from the Northern blots to establish a multi-
nominal model (see Supplemental materials for details), and ar-
rived at an estimate of 2936 different small ncRNAs. Taken to-
gether, these three estimates all point to an ncRNA transcriptome
of a few thousand species, with a figure close to 2700 as the most
likely estimate (Table 1).
Our data show that the eukaryotic non-coding transcriptome still
harbors plenty of novelty. Using a novel cloning strategy, we
have identified 100 novel non-coding transcripts with verified
expression. Three elements of our cloning approach contributed
significantly to our result. Fractioning of total RNA on an anion
resin instead of through PAGE reduced contamination from frag-
mented high-molecular weight RNAs (Eddy 2001; Huttenhofer et
al. 2001). Specific targeting of the most important RNA contami-
nants through oligo-coated magnetic beads prior to cloning en-
abled us to reach a pre-screen ncRNA detection efficiency of 36%
on novel ncRNAs (70% for all small ncRNAs) (Fig. 1A), compared
to the 3%∼7% obtained in previous studies (Huttenhofer et al.
2001). Finally, ligating adaptors to both 3?- and 5?-ends protected
terminal nucleotides from exonuclease degradation and guaran-
teed that incompletely reverse-transcribed RNAs would not be
cloned. This yielded full-length sequences and allowed determi-
nation of the transcript 5?-end structure, which, in turn, enabled
us to establish correlations between upstream motifs, 5?-cap sta-
tus, and 3?-end termination signals indicative of the ncRNA
mode of biogenesis.
A large contingent of snoRNA-like transcripts was to be ex-
pected, as this fraction of the C. elegans transcriptome has been
less studied than miRNAs (Higa et al. 2002; Wachi et al. 2004;
Stricklin et al. 2005). However, our data also include 31 tran-
scripts for which no sequential or structural homolog outside the
nematodes could be found. Several of these carry unique motifs
both upstream and internally, and may well represent novel
functional classes of ncRNAs. RNomics efforts in other model
organisms have all revealed 25%–50% clones that could not be
ascribed to any known functional category of ncRNAs (Hutten-
hofer et al. 2001; Marker et al. 2002; Yuan et al. 2003; Wang et al.
2004), but only in Dictyostelium could these transcripts be
grouped into potentially novel classes of ncRNAs based on com-
mon internal and upstream motifs (Aspegren et al. 2004). Of the
31 novel clones in our data, we could discern two groups with
common features, comprising 17 different transcripts. Among
these, nine stem–bulge RNAs (sbRNAs) display characteristics
slightly resembling Dictyostelium Class I RNAs (Aspegren et al.
2004), in that the defining sequence motifs are found at the RNA
5?- and 3?-termini, and have the potential to form a stem–loop
structure containing a bulge with a conserved sequence. The loci
of both classes also contain upstream motifs common to all genes
within each group. However, the length and composition of the
sequence motifs defining the respective classes are not conserved
between Caenorhabditis and Dictyostelium, and there is no further
evidence to support any functional relationship between the two
ncRNA categories. The snRNA-like RNAs (snlRNAs) clearly share
motif 1 (UM1; E = 4.0 ? 10?521). (B) Upstream motif 2 (UM2;
E = 7.3 ? 10?179). (C) Upstream motif 3 (UM3; E = 1.1 ? 10?38). (For
explanation of E, see Fig. 2.) (D) Distribution of distances from motif
position 1 of UM1, UM2, and UM3, respectively, to 5?-end of the ncRNA
transcripts. UM1 and UM3 have defined distances from start of the motif
to the 5?-end of transcript. The two peaks for UM1 represent distances for
loci with (smaller peak) and without (larger peak) an additional TATA-
box. The distances between UM2 and transcript 5?-ends are more vari-
able, possibly indicative of post-transcriptional 5?-end processing for this
group of ncRNAs.
Upstream motifs discovered at ncRNA loci. (A) Upstream
The non-coding transcriptome of C. elegans
an Sm protein-binding site with both spliceosomal snRNAs and
SL RNAs (Zeiner et al. 2004). However, the conserved sequence
element in this group extends beyond the Sm-binding site, and
may allow for binding of additional proteins not participating in
the general spliceosomal processes. With a few exceptions, the
snlRNAs also share the UM1/PSE upstream motif with both
snRNAs and several other ncRNAs, possibly also pointing to in-
volvement in similar processes. The C.
elegans protein-coding genes contain a
large fraction of very short introns that
may require additional RNA factors for
correct splicing, and a preliminary sug-
gestion would be that the snlRNAs are
somehow involved in splicing related ac-
The lack of conservation beyond C.
briggsae for all but one of the novel tran-
scripts is conspicuous, but not entirely
unprecedented. Previous attempts to de-
tect small ncRNAs (Huttenhofer et al.
2001; Marker et al. 2002; Tang et al.
2002; Yuan et al. 2003) have resulted in
the identification of a group of small
non-coding transcripts not assignable to
any known class of small ncRNAs. Fur-
thermore, finding a conserved homolog
for these transcripts even in related spe-
cies has been notoriously difficult. More-
over, as we have regarded as “known” or
“predicted” all ncRNAs that have been
annotated in the C. elegans genome
based on sequence homology in other
species, it is not altogether unreasonable
that our novel ncRNAs should lack non-
nematode homologs. The recent obser-
vation that even ultraconserved verte-
brate non-coding sequences are not
found in nematodes or flies (Bejerano et
al. 2004; Woolfe et al. 2005) may indi-
cate a general lack of conservation of
non-coding regulatory elements be-
tween vertebrates and non-vertebrates.
Developmentally variable expres-
sion of the C. elegans non-coding tran-
scriptome is also intriguing, particularly
considering that the constitutively ex-
pressed spliceosomal snRNAs displayed
very stable expression. We therefore as-
sume these variations in expression are
not artifactual but physiologically rel-
evant. Few small ncRNAs are likely to
themselves have catalytic or purely
structural functions; instead, develop-
mentally regulated ncRNAs may regulate
the activity of other gene products. Also,
snoRNAs show a remarkably high fre-
quency of variably expressed transcripts,
possibly indicating that methylation
and ?-uridylation of rRNA may actually
modify ribosomal properties. Several
mammalian snoRNAs have been shown
to have tissue-specific expression (Ca-
vaille et al. 2000). In Archaea, rRNA methylation has been
shown to depend on culturing temperature (Noon et al. 1998).
Adaptive rRNA modifications in response to varying environ-
ment conditions have also been suggested as an explanation to
the extended repertoire of snoRNAs in plants (Brown et al. 2003).
As most of the variably expressed snoRNAs in our data showed
increased levels at the dauer stage, this might indicate a role for
(Legend on next page)
Deng et al.
differential modification of rRNA nucleotides between normal
and hunger-stressed worms.
Transcriptional analysis of the data gave three major results.
The first was an extension of previous observations (Hernandez
2001; Ohler et al. 2004) that small ncRNAs have unique promoter
structures. Particularly interesting is the strong resemblance be-
tween UM2 and the tRNA internal promoter, perhaps suggesting
that TFIIIC (or a TFIIIC-like factor) binds UM2 and recruits RNA
polymerase III. In Arabidopsis an arrangement of a tRNAGlygene
preceding a snoRNA43 gene has been found at several loci, and a
similar tRNAMet-snoRNA locus was also observed in rice (Kruszka
et al. 2003). One of the Arabidopsis loci was demonstrated to give
rise to a dicistronic primary transcript that was cleaved by the
tRNA 3?-end-processing enzyme RNase Z, releasing the snoRNA
from the tRNA. We have no evidence for a dicistronic primary
transcript, possibly because it is very short-lived. An alternative
possibility is that an originally internal tRNA promoter has been
transformed to act as an (non-transcribed) upstream core promoter.
Whatever the case, this arrangement of tRNA (or tRNA-like) and
snoRNA genes in both C. elegans and plants possibly points to a
very old evolutionary solution to the transcription of snoRNA
genes that predates the divergence of plants and animals.
The second major result from the transcriptional analysis
was the high number of snoRNA-like transcripts with apparently
independent transcription. Whereas independent transcription
of sno/scaRNAs is common in plants (Brown et al. 2003), it has
been reported for very few sno/scaRNAs in animals (Tycowski
et al. 2004). In C. elegans, 52 of 88 known and novel snoRNA-
like transcripts appear to have an upstream promoter element
and/or an intergenic location. The third result was the extent
to which intronic C. elegans ncRNAs are independently tran-
scribed. Whereas independent transcription from intronic loci
has been reported (Dominski et al. 2003), in C. elegans >50% of
all intronic ncRNA loci show strong signs of independent tran-
Recent studies have indicated that the total number of non-
protein-coding transcripts in both human (Bertone et al. 2004;
Cawley et al. 2004; Kampa et al. 2004) and mouse (Okazaki et al.
2002) genomes may be far higher than previous estimates. As-
suming that ncRNAs play a major role in specifying eukaryotic
multicellular complexity (Mattick 2003, 2004), the complexity
and number of cell fates in a nematode (of the order 103) should
imply a higher number of non-coding transcripts than in a uni-
cellular organism, but lower than in more complex organisms
like insects or vertebrates. Present estimates for ncRNAs in Esch-
erichia coli stand at some hundreds (Vogel et al. 2003; Zhang et al.
2004; Saetrom et al. 2005), whereas data on human Chromo-
somes 21 and 22 (Cawley et al. 2004; Kampa et al. 2004; Cheng
et al. 2005) would indicate a number at least two orders of mag-
nitude higher for mammals. An estimate of ∼2700 small non-
coding RNAs in C. elegans compares well with the 3000–4000
predicted ncRNA loci reported in the most recent computational
survey of this organism (Missal et al. 2005). A number of ncRNAs
in the lower thousands would amount to a class of possible regu-
latory molecules in C. elegans as large as the complete set of
protein signal transduction factors (Chervitz et al. 1998), and
might fit the bill of this complex an organism.
Total RNA was isolated from mixed-stage worms and eggs accord-
ing to the Trizol (Invitrogen) protocol, then loaded on a Qiagen-
tip (Qiagen) and eluted with a 0.6 ∼ 1.0 M NaCl gradient of
mental data. (A) TATA-less loci with UM1. This type of locus is characterized by the Upstream Motif 1 and is found both intergenically and intronically.
Transcripts from TATA-less UM1 loci generally carry a 5?-end cap, most likely transcribed by RNA polymerase II, and make up biogenesis group I-A, which
comprises most spliceosomal snRNAs, a fraction of the SL RNAs, most snlRNAs, and a few C/D snoRNAs along with some unclassified transcripts. (B) Loci
with UM1 and a TATA-box. This type of locus combines the UM1 with a TATA-box, and most often a tract of four or more Ts is found within 10 bp of
the transcript 3?-terminus. Known RNA polymerase III transcripts like U6 snRNA and RNase P RNA are found at this type of locus. The transcripts may
have a single methyl group added at the ?-phosphate post-transcriptionally, as is commonly found in U6 and 7SK snRNAs (Gupta et al. 1990). (C) Loci
with UM2. This type of locus comprises a number of both intergenic and intronic snoRNA-like transcripts, along with a few uncharacterized ncRNAs,
and makes up biogenesis group II. Transcripts are generally uncapped, and an oligo-T tract is found close to the 3?-terminus, indicating transcription
by RNA polymerase III. FB (Front Box) and TB (Tail Box) are the most conserved 15-bp motifs within the 100-bp upstream sequence of these loci, and
show strong resemblance to Box A and Box B of the tRNA promoter. A “possible tRNA transcription” initiation site has been indicated to account for
the possibility that UM2 is transcribed as a part of the primary transcript (see Supplemental material for details). (D) Loci with UM3. This type of locus
has only been found in sbRNAs, and is characterized by UM3, which contains a TATA-box preceded by a strongly conserved G residue. The loci are
terminated by an oligo-T tract, and most transcripts are uncapped, suggesting transcription by RNA polymerase III. (E) SRP RNA loci. The C. elegans SRP
RNA loci are characterized by a rudimentary TATA-box and a Box A element at ∼10–20 bp downstream of the transcription start, and are terminated
by an oligo-T tract. (F) Independently transcribed intronic loci. This type of locus represents subgroups of locus types A–E, in which both the transcribed
sequence and the corresponding control elements (promoter, terminator) are found within the intron of a protein-coding gene. This type of locus is
found for all the above promoter elements, but is most common for UM1 and UM2 type loci. (G) Motif-less intronic loci. These loci are exclusively made
up of snoRNA-like genes, and are often found within an intron of a ribosomal gene. The distance between the ncRNA locus and the preceding exon
is generally short (<50 bp) and AT-rich. Transcription is initiated from the host gene promoter, and the snoRNA is processed either directly from the
pre-mRNA, or from a spliced intron lariat.
Arrangements of transcriptional elements and genomic locations of small non-coding ncRNA loci, as inferred from genomic and experi-
Estimates of the C. elegans ncRNA transcriptome
ModelEstimated no. of ncRNA species
1. Intron conservation
2. Conserved upstream motifs
3. Clone no. versus expression level
Model 1 is based on the difference between conservation between
ncRNA-containing introns and the total intron population when compar-
ing Caenorhabditis elegans to Caenorhabditis briggsae. Basically, the in-
trons shorter than 130 bp are taken to represent non-ncRNA introns, and
the fraction of ncRNA-containing introns in the total intron population is
inferred by linear regression analysis. Correcting for the fraction of inter-
genic ncRNA loci yields the estimated number of ncRNAs. Model 2 is
based on the occurrence of upstream motif 1 (UM1) in the C. elegans
genome, corrected for one highly repetitive sequence, and adjusted for
nonmotif loci. Model 3 is a statistical calculation based on the correlation
between the frequency of identical clones in the ncRNA library, and the
concentration (expression level) of the corresponding ncRNA as esti-
mated from Northern blots. (All three models are explained in detail in
the Supplemental material.)
The non-coding transcriptome of C. elegans
QRW2 buffer (Qiagen RNA/DNA Handbook) at 50°C. Fractions
corresponding to tRNAs, small RNAs (80–500 nt), and high-
molecular-weight mRNA/rRNA were collected. The Ambion
MicrobExpress kit was adapted to remove remaining mRNAs and
rRNAs: The small RNA fraction was hybridized to a mixture of
specifically designed oligonucleotides in binding buffer, then un-
wanted RNA molecules were targeted and removed from ncRNA
by a magnetic beads-based process as per protocol (Ambion). The
enriched ncRNA pool was cloned using an adaptor-mediated li-
brary construction protocol (modified from Elbashir et al. 2001):
RNAs were dephosphorylated with calf intestine alkaline phos-
phatase (Fermentas), then ligated to the 3?-adaptor oligonucleo-
tide by T4 RNA ligase (Fermentas). The ligation product was split
into two aliquots; one was treated with PolyNucleotide Kinase
(PNK; Fermentas) to phosphorylate uncapped RNA, and the
other was treated with Tobacco Acid Pyrophosphatase (TAP; Epi-
centre) to remove 5?-end methyl-guanosine caps from capped
RNA. Thereafter, both were ligated to the 5?-adaptor oligonucleo-
tide. Small molecules and excessive adaptors were removed from
the ligation products with the mirVana miRNA isolation kit (Am-
bion) and reverse-transcribed (RT) with Thermoscript (Invitro-
gen) at 50°C, using oligo 3RT as the RT primer. Library normal-
ization (when applied) was carried out by adding mRNA and
rRNA from the above removal procedure to the RT products, then
denaturing at 98°C in hybridization buffer, followed by reanneal-
ing at 70°C and treatment with duplex-specific nuclease (DSN;
Evrogen) for 25 min. The cDNA was PCR-amplified by using
Platinum Taq (Invitrogen) with 3RT and 5CD primers for 15
cycles, digested with PaeI and SacI (Fermentas) and cloned in
pGEM-4Z (Promega; see Supplemental material for the oligo-
nucleotide sequences used in this study).
5?- and 3?-RACE
RACE was performed by PCR amplification of the RT products
(see above), with one primer designed specific to the ncRNA se-
quence and another primer being either 5CD or 3RT for 5?- and
RNA probes were synthesized and labeled by in vitro transcrip-
tion of plasmids with T7 RNA polymerase (Fermentas) and Dig-
11-UTP (Roche). Total RNA extracted from 12 different develop-
mental stages and two environmental conditions was analyzed.
Northern blotting was performed per standard and manufactur-
ers’ protocols. Blots were hybridized in ULTRAhyb (Ambion) at
68°C overnight, then treated with Blocking and Washing Buffer
(Roche) and detected by CDP-star (Roche). Chemiluminescent
signals were recorded in an image system ChemiCapt 3000 (Vil-
ber). See Supplemental material for definitions of developmental
stages and details of the Northern blot analysis.
Determination of ncRNA sequences and 5?-structure
The 161 ncRNA sequences were either determined as a consensus
of aligned sequencing reads belonging to the same ncRNA
species, when these agreed with size estimates from Northern
blots (most cases), by joining two adjacent ncRNAs inferred
from genomic and Northern blots’ information (four cases), or
from elongation of consensus sequences based on RACE and
Northern data (10 cases). The probability of an ncRNA being
5?-capped or not was determined from a statistic model based on
the distribution of known capped and uncapped small RNAs in
TAP or PNK libraries (see Supplemental material, “Capping Prob-
C. elegans genome annotation and sequence data, and the C.
briggsae genome data, were downloaded from WormBase (version
WS123) (Harris et al. 2003). The 161 ncRNA sequences were
mapped to 198 C. elegans genome loci by BLASTN. These loci, as
well as their equal-sized 5?- and 3?-flanking sequences, were
searched by BLASTN for homologs in the C. briggsae genome,
with default parameters except for having the low-complexity
filter switched off. The conservation score of a sequence used
here is defined as the identical residues count, divided by the
length of the sequence, in the alignment of the best High-Scoring
The MEME motif discovery tool (version 3.0.13) (Bailey and
Elkan 1995) was used to search for conserved motifs in all the
ncRNA sequences and 100-bp upstream sequences of all ncRNA
gene loci, respectively. This discovered three clearly discernible
upstream motifs: UM1, UM2, and UM3; and three internal mo-
tifs: IM1, IM2, and IM3. The Meta-MEME software (Grundy et al.
1997) was used to search for further UM1, UM2, and UM3 sites in
the C. elegans genome, using Hidden Markov Models (HMM) for
each upstream motif that had been generated from MEME-
produced weight matrices. To avoid the high scores of tandem
repeat regions when using a Meta-MEME scan on a whole-
genome scale, repeats were masked prior to the HMM search. An
E-value threshold (0.1) was chosen such that most (>90%) of the
upstream motifs associated with ncRNAs in our material could be
We thank Haitao Guo, Jinzhao Li, and Jun Liu for early experi-
ment discussion; Zhihua Zhang for Perl programming; and Mu-
hammad Nauman Aftab and Jimmy Ye for careful reading of the
manuscript. The C. elegans strain N2 used in this work was pro-
vided by the Caenorhabditis Genetics Center, which is funded by
the NIH National Center for Research Resources. This work was
supported by the Chinese Academy of Sciences Grant Nos.
KSCX2-2, KJCX1-08, and KSCXZ-SW-223; National Sciences
Foundation of China Grant No. 39890070; the National High
Technology Development Program of China under Grant No.
2002AA231031; National Key Basic Research & Development
Program 973 under Grant Nos. 2002CB713805 and
2003CB715900; and Beijing Science and Technology Commis-
sion Grant Nos. H020220030130 and H010210010113.
Aspegren, A., Hinas, A., Larsson, P., Larsson, A., and Soderbom, F. 2004.
Novel non-coding RNAs in Dictyostelium discoideum and their
expression during development. Nucleic Acids Res. 32: 4646–4656.
Atzorn, V., Fragapane, P., and Kiss, T. 2004. U17/snR30 is a ubiquitous
snoRNA with two conserved sequence motifs essential for 18S rRNA
production. Mol. Cell. Biol. 24: 1769–1778.
Bailey, T.L. and Elkan, C. 1995. The value of prior knowledge in
discovering motifs with MEME. Proc. Int. Conf. Intell. Syst. Mol. Biol.
Bejerano, G., Pheasant, M., Makunin, I., Stephen, S., Kent, W.J., Mattick,
J.S., and Haussler, D. 2004. Ultraconserved elements in the human
genome. Science 304: 1321–1325.
Bertone, P., Stolc, V., Royce, T.E., Rozowsky, J.S., Urban, A.E., Zhu, X.,
Rinn, J.L., Tongprasit, W., Samanta, M., Weissman, S., et al. 2004.
Global identification of human transcribed sequences with genome
tiling arrays. Science 306: 2242–2246.
Brown, J.W.S., Echeverria, M., and Qu, L.-H. 2003. Plant snoRNAs:
Functional evolution and new modes of gene expression. Trends
Plant Sci. 8: 42–49.
Cavaille, J., Buiting, K., Kiefmann, M., Lalande, M., Brannan, C.I.,
Deng et al.
28 Genome Research
Horsthemke, B., Bachellerie, J.P., Brosius, J., and Huttenhofer, A. Download full-text
2000. Identification of brain-specific and imprinted small nucleolar
RNA genes exhibiting an unusual genomic organization. Proc. Natl.
Acad. Sci. 97: 14311–14316.
Cawley, S., Bekiranov, S., Ng, H.H., Kapranov, P., Sekinger, E.A., Kampa,
D., Piccolboni, A., Sementchenko, V., Cheng, J., Williams, A.J., et al.
2004. Unbiased mapping of transcription factor binding sites along
human chromosomes 21 and 22 points to widespread regulation of
noncoding RNAs. Cell 116: 499–509.
Cheng, J., Kapranov, P., Drenkow, J., Dike, S., Brubaker, S., Patel, S.,
Long, J., Stern, D., Tammana, H., Helt, G., et al. 2005.
Transcriptional maps of 10 human chromosomes at 5-nucleotide
resolution. Science 308: 1149–1154.
Chervitz, S.A., Aravind, L., Sherlock, G., Ball, C.A., Koonin, E.V.,
Dwight, S.S., Harris, M.A., Dolinski, K., Mohr, S., Smith, T., et al.
1998. Comparison of the complete protein sets of worm and yeast:
Orthology and divergence. Science 282: 2022–2028.
de Turris, V., Di Leva, G., Caldarola, S., Loreni, F., Amaldi, F., and
Bozzoni, I. 2004. TOP promoter elements control the relative ratio of
intron-encoded snoRNA versus spliced mRNA biosynthesis. J. Mol.
Biol. 344: 383–394.
Dominski, Z., Yang, X.-C., Purdy, M., and Marzluff, W.F. 2003. Cloning
and characterization of the Drosophila U7 small nuclear RNA. Proc.
Natl. Acad. Sci. 100: 9422–9427.
Eddy, S.R. 2001. Non-coding RNA genes and the modern RNA world.
Nat. Rev. Genet. 2: 919–929.
Elbashir, S.M., Lendeckel, W., and Tuschl, T. 2001. RNA interference is
mediated by 21- and 22-nucleotide RNAs. Genes & Dev. 15: 188–200.
Grundy, W.N., Bailey, T.L., Elkan, C.P., and Baker, M.E. 1997.
Meta-MEME: Motif-based hidden Markov models of protein families.
Comput. Appl. Biosci. 13: 397–406.
Gupta, S., Busch, R., Singh, R., and Reddy, R. 1990. Characterization of
U6 small nuclear RNA cap-specific antibodies. Identification of
?-monomethyl-GTP cap structure in 7SK and several other human
small RNAs. J. Biol. Chem. 265: 19137–19142.
Harris, T.W., Lee, R., Schwarz, E., Bradnam, K., Lawson, D., Chen, W.,
Blasier, D., Kenny, E., Cunningham, F., Kishore, R., et al. 2003.
WormBase: A cross-species database for comparative genomics.
Nucleic Acids Res. 31: 133–137.
Hernandez, N. 2001. Small nuclear RNA genes: A model system to study
fundamental mechanisms of transcription. J. Biol. Chem.
Higa, S., Maeda, N., Kenmochi, N., and Tanaka, T. 2002. Location of
2?-O-methyl nucleotides in 26S rRNA and methylation guide
snoRNAs in Caenorhabditis elegans. Biochem. Biophys. Res. Comm.
Huttenhofer, A., Kiefmann, M., Meier-Ewert, S., O’Brien, J., Lehrach, H.,
Bachellerie, J.P., and Brosius, J. 2001. RNomics: An experimental
approach that identifies 201 candidates for novel, small,
non-messenger RNAs in mouse. EMBO J. 20: 2943–2953.
Huttenhofer, A., Schattner, P., and Polacek, N. 2005. Non-coding RNAs:
Hope or hype? Trends Genet. 21: 289–297.
Kampa, D., Cheng, J., Kapranov, P., Yamanaka, M., Brubaker, S., Cawley,
S., Drenkow, J., Piccolboni, A., Bekiranov, S., Helt, G., et al. 2004.
Novel RNAs identified from an in-depth analysis of the
transcriptome of human chromosomes 21 and 22. Genome Res.
Kruszka, K., Barneche, F., Guyot, R., Ailhas, J., Meneau, I., Schiffer, S.,
Marchfelder, A., and Echeverria, M. 2003. Plant dicistronic
tRNA-snoRNA genes: A new mode of expression of the small
nucleolar RNAs processed by RNase Z. EMBO J. 22: 621–632.
Liu, C., Bai, B., Skogerbo, G., Cai, L., Deng, W., Zhang, Y., Bu, D., Zhao,
Y., and Chen, R. 2005. NONCODE: An integrated knowledge
database of non-coding RNAs. Nucleic Acids Res. 33: D112–D115.
Lowe, T. and Eddy, S. 1997. tRNAscan-SE: A program for improved
detection of transfer RNA genes in genomic sequence. Nucleic Acids
Res. 25: 955–964.
Marker, C., Zemann, A., Terhorst, T., Kiefmann, M., Kastenmayer, J.P.,
Green, P., Bachellerie, J.P., Brosius, J., and Huttenhofer, A. 2002.
Experimental RNomics: Identification of 140 candidates for small
non-messenger RNAs in the plant Arabidopsis thaliana. Curr. Biol.
Mattick, J.S. 2003. Challenging the dogma: The hidden layer of
non-protein-coding RNAs in complex organisms. Bioessays
———. 2004. RNA regulation: A new genetics? Nat. Rev. Genet.
Missal, K., Zhu, X., Rose, D., Deng, W., Skogerboe, G., Chen, R., and
Stadler, P.F. 2005. Prediction of structured noncoding RNAs in the
genome of the nematode Caenorhabitis elegans. J. Exp. Zoolog. B Mol.
Dev. Evol. (in press).
Noon, K.R., Bruenger, E., and McCloskey, J.A. 1998. Posttranscriptional
modifications in 16S and 23S rRNAs of the Archaeal
hyperthermophile Sulfolobus solfataricus. J. Bacteriol. 180: 2883–2888.
Ohler, U., Yekta, S., Lim, L.P., Bartel, D.P., and Burge, C.B. 2004.
Patterns of flanking sequence conservation and a characteristic
upstream motif for microRNA gene identification. RNA
Okazaki, Y., Furuno, M., Kasukawa, T., Adachi, J., Bono, H., Kondo, S.,
Nikaido, I., Osato, N., Saito, R., Suzuki, H., et al. FANTOM
Consortium. RIKEN Genome Exploration Research Group Phase I &
II Team. 2002. Analysis of the mouse transcriptome based on
functional annotation of 60,770 full-length cDNAs. Nature
Pang, K.C., Stephen, S., Engstrom, P.G., Tajul-Arifin, K., Chen, W.,
Wahlestedt, C., Lenhard, B., Hayashizaki, Y., and Mattick, J.S. 2005.
RNAdb—A comprehensive mammalian noncoding RNA database.
Nucleic Acids Res. 33: D125–D130.
Riedel, N., Wolin, S., and Guthrie, C. 1987. A subset of yeast snRNAs
contains functional binding sites for the highly conserved Sm
antigen. Science 235: 328–331.
Saetrom, P., Sneve, R., Kristiansen, K.I., Snove Jr., O., Grunfeld, T.,
Rognes, T., and Seeberg, E. 2005. Predicting non-coding RNA genes
in Escherichia coli with boosted genetic programming. Nucleic Acids
Res. 33: 3263–3270.
Stolc, V., Samanta, M.P., Tongprasit, W., Sethi, H., Liang, S., Nelson,
D.C., Hegeman, A., Nelson, C., Rancour, D., Bednarek, S., et al.
2005. Identification of transcribed sequences in Arabidopsis thaliana
by using high-resolution genome tiling arrays. Proc. Natl. Acad. Sci.
Stricklin, S.L., Griffiths-Jones, S., and Eddy, S.R. 2005. C. elegans
noncoding RNA genes. In WormBook (ed. The C. elegans Research
Community), p. 7, http://www.wormbook.org.
Tang, T.H., Bachellerie, J.P., Rozhdestvensky, T., Bortolin, M.L., Huber,
H., Drungowski, M., Elge, T., Brosius, J., and Huttenhofer, A. 2002.
Identification of 86 candidates for small non-messenger RNAs from
the archaeon Archaeoglobus fulgidus. Proc. Natl. Acad. Sci.
Thomas, J., Lea, K., Zucker-Aprison, E., and Blumenthal, T. 1990. The
spliceosomal snRNAs of Caenorhabditis elegans. Nucleic Acids Res.
Tycowski, K.T., Aab, A., and Steitz, J.A. 2004. Guide RNAs with 5? caps
and novel box C/D snoRNA-like domains for modification of
snRNAs in Metazoa. Curr. Biol. 14: 1985–1995.
Vogel, J., Bartels, V., Tang, T.H., Churakov, G., Slagter-Jager, J.G.,
Huttenhofer, A., and Wagner, E.G.H. 2003. RNomics in Escherichia
coli detects new sRNA species and indicates parallel transcriptional
output in bacteria. Nucleic Acids Res. 31: 6435–6443.
Wachi, M., Ogawa, T., Yokoyama, K., Hokii, Y., Shimoyama, M., Muto,
A., and Ushida, C. 2004. Isolation of eight novel Caenorhabditis
elegans small RNAs. Gene 335: 47–56.
Wang, J. and Kim, S.K. 2003. Global analysis of dauer gene expression
in Caenorhabditis elegans. Development 130: 1621–1634.
Wang, J., Zhang, J., Zheng, H., Li, J., Liu, D., Li, H., Samudrala, R., Yu,
J., and Wong, G.K.-S. 2004. Mouse transcriptome: Neutral evolution
of ‘non-coding’ complementary DNAs. Nature 431: 758.
Woolfe, A., Goodson, M., Goode, D.K., Snell, P., McEwen, G.K., Vavouri,
T., Smith, S.F., North, P., Callaway, H., Kelly, K., et al. 2005. Highly
conserved non-coding sequences are associated with vertebrate
development. PLoS Biol. 3: e7.
Yamada, K., Lim, J., Dale, J.M., Chen, H., Shinn, P., Palm, C.J.,
Southwick, A.M., Wu, H.C., Kim, C., Nguyen, M., et al. 2003.
Empirical analysis of transcriptional activity in the Arabidopsis
genome. Science 302: 842–846.
Yuan, G., Klambt, C., Bachellerie, J.P., Brosius, J., and Huttenhofer, A.
2003. RNomics in Drosophila melanogaster: Identification of 66
candidates for novel non-messenger RNAs. Nucleic Acids Res.
Zeiner, G.M., Foldynova, S., Sturm, N.R., Lukes, J., and Campbell, D.A.
2004. SmD1 is required for spliced leader RNA biogenesis. Eukaryotic
Cell 3: 241–244.
Zhang, Y., Zhang, Z., Ling, L., Shi, B., and Chen, R. 2004. Conservation
analysis of small RNA genes in Escherichia coli. Bioinformatics
Received April 16, 2005; accepted in revised form August 22, 2005.
The non-coding transcriptome of C. elegans