snoRNAs: Biogenesis,
Structure and Function
Francois Dragon, Universite du Quebec a` Montreal, Quebec, Canada
Vincent Lemay, Universite du Quebec a` Montreal, Quebec, Canada
Christian Trahan, Universite du Quebec a` Montreal, Quebec, Canada
The small nucleolar (sno) ribonucleic acids (RNAs) are mainly involved in the modification
and processing of ribosomal RNA precursors, however, some are used to modify small
nuclear RNAs (snRNAs), messenger RNAs (mRNAs) and transfer RNAs (tRNAs). They
function in the form of ribonucleoproteins (snoRNPs).
Introduction
In eukaryotes, ribosome biogenesis takes place in the nu-
cleolus, a prominent compartment of the nucleus. The nu-
cleolus contains a large population of small ribonucleic
acid (RNA) molecules called small nucleolar (sno) RNAs,
which are involved in the modication and processing of
pre-ribosomal RNAs (pre-rRNAs). The snoRNAs func-
tion in association with specic proteins and thus form
ribonucleoproteins (snoRNPs). To date, about 80 dierent
snoRNAs have been found in the yeast Saccharomyces
cerevisiae, and almost twice as many in human cells. The
copy number of each snoRNA species varies between or-
ganisms, yeast having less than vertebrates (
1000 versus

10 000). Some snoRNAs are much more abundant than
others, for instance there is about 200 000 copies of U3 in
human cells. See also: Eukaryotic ribosomes: assembly;
Ribosomal RNA; Nucleolus
Primary and Secondary Structures and
Conserved Elements of Small Nucleolar
RNAs
The snoRNAs are relatively short molecules ranging be-
tween 60 and 600 nucleotides, but most of them are in the
range of 70200 nucleotides. They can be grouped into two
major families called C/D and H/ACA; this classication is
based on conserved sequence motifs (Figure 1).
Members of the box C/D family share the sequence mo-
tifs PuUGAUGA (box C) and CUGA (box D) located
near their 5 and 3 ends, respectively. The 5 and 3 ends are
usually base-paired, thus favouring the exposure of boxes
C and D; this basic structure is also known as the C/D
core motif. Base pairing of boxes C and D forms a K-turn,
a motif known to bind a specic protein (see below).
Intact motifs C and D are required for snoRNA stability
and accumulation in vivo. Most C/D snoRNAs contain
ENCYCLOPEDIA OF LIFE SCIENCES & 2006, John Wiley & Sons, Ltd. www.els.net
Introductory article
. Introduction
Article Contents
. Primary and Secondary Structures and Conserved
Elements of Small Nucleolar RNAs
. Role of snoRNAs in Pre-rRNA Processing and
Modification
. snoRNA-associated Proteins
. Expression of snoRNAs
. snoRNA Maturation Pathways
. Small Cajal Body-specific RNAs
. Possible Implications of snoRNAs-modified Targets
. Origin and Evolution of snoRNAs
doi: 10.1038/npg.els.0003813
additional C- and D-like motifs termed boxes C and D,
which are separated by a short spacer sequence; in some
cases the spacer can form a small helix thus reproducing the
general C/D core motif. The C/D snoRNAs usually pos-
sess one or two long sequences that are complementary to
evolutionarily conserved sequences of rRNA; these se-
quences of 1021 nucleotides are located just upstream
from box D or D (Figure 1a).
Members of the box H/ACA group are characterized by
the sequence motif ANANNA (box H), and by a highly
conserved ACA triplet (box ACA). Each motif is impor-
tant for snoRNA stability and accumulation in vivo. The
majority of H/ACA snoRNAs fold into a hairpinhinge
hairpintail structure; the hinge and tail are single-
stranded regions that contain the H and ACA motifs, re-
spectively. Note that box ACA is always positioned three
nucleotides upstream from the 3 end (Figure 1b). The hair-
pins are generally interrupted by an internal loop located
close to their base; the internal loops usually contain short
sequences that are complementary to rRNA sequences that
undergo base modication (see below). Interestingly, the
RNA component of vertebrate telomerase, the ribonuc-
leoprotein enzyme that synthesizes the ends of chromo-
somes in eukaryotic cells, contains a domain that
structurally resembles H/ACA snoRNAs; this H/ACA-
like domain is located at the 3 end of the molecule, and it is
essential for stability and accumulation of telomerase
RNA in vivo. See also: Telomerase: structure and function
The 7-2/MRP RNA (the RNA component of RNase
MRP) is another snoRNA, but it does not belong to either
C/D or H/ACA-families. This highly structured molecule
has no characteristic primary sequence motif, but its cage-
shape structure and few sequence patches make it very sim-
ilar to the RNA component of RNase P, the endonuclease
that cleaves the 5 extension of transfer RNA (tRNA) pre-
cursors. Interestingly, a mutation in yeast RNase P RNA
1
 snoRNAs: Biogenesis, Structure and Function

(a)
U
Box D UG

A
G

5 C
A
U Box C
G
A

m m

rRNA 3 N N N N NN 5
3
3
snoRNA 5 N N N N NN CUGA 3
5-16
A m Box
G
U D or D
Box C A
G
U
C
U 5
U
R G
N Box D

A
5 3

(b)
snoRNA

N N 14-16 nt Box H or ACA

N N A N AN N A 3
N N 5 N
N ACA NNN OH
N N N
3-10 4-10

5
3 rRNA 5
3

Box H Box ACA

5 A N AN N A A C A N N N 3

Figure 1 Structure of guide snoRNAs and mode of action. The schematic secondary structure of snoRNAs are presented on the left; the blue line
represents the target RNA that gets modified. The conserved sequences are boxed and indicated in colour. Models for the selection of specific position to be
modified are shown on the right. In (a), box C/D snoRNAs are characterized by conserved motifs C and D (green) that form a kink-turn (K-turn) or alternate
boxes C and D (orange) that could also form a K-turn; 2-O-ribose methylation is performed on the rRNA residue that is base-paired to the fifth position
upstream from box D (or D). In (b), H/ACA snoRNAs adopt a hairpin-hinge-hairpin-tail structure where box H (red) is found in the hinge region and box
ACA (red) is found three nucleotides upstream of the 3- end; each hairpin usually contains an internal loop called pseudouridylation pocket where C
formation in rRNA occurs on the first unpaired U residue upstream from box H or ACA (usually at a distance of 14 to 16 nucleotides from the box).

(RPR1) can aect pre-rRNA processing, suggesting that
RNase P could have a function in the nucleolus. Because of
their striking similarity, it has been proposed that RNase P
and RNase MRP RNAs evolved from a common ancestor.
See also: Cellular RNAs: varied roles; Ribonucleases; RNA
structure; Transfer RNA synthesis and regulation
Role of snoRNAs in Pre-rRNA Processing
and Modification
The small- and large-subunit rRNAs are transcribed as
part of a large precursor that is subsequently cleaved and
trimmed to release mature rRNA products, the 18S rRNA,
which is part of the small subunit (SSU), and the 5.8S and
2528S rRNAs, which are found in the large subunit
(LSU) of the ribosome. In contrast to pre-mRNAs, which
begin to be processed (spliced) before completion of syn-
thesis, pre-rRNAs are generally fully transcribed before
cleavages occur. Post-transcriptional modications are
made on the precursors, not the mature products, and
some modications could be prerequisites to processing
events (Figure 2). The great majority of snoRNAs are
involved in modication rather than processing of pre-
rRNAs: box C/D snoRNAs direct the 2-O-ribose methyl-
ation of specic residues, whereas box H/ACA snoRNAs
target specic uridines that are converted into pseudouri-
dines (C), an isomerization process called pseudouridylat-
ion. Only a fraction of snoRNAs are required for the
cleavage reactions. The box C/D U3, U13, U14 and U22
snoRNAs are involved in 18S rRNA production, and U8
participates in processing of the 5.8S and 28S rRNAs.
Processing reactions involving U3, U8, U13 and U14 like
those participating in modication require specic base
pairing interactions between the pre-rRNA and the
snoRNPs. The box H/ACA snoRNAs snR10, snR30/
U17 and E3 participate in 18S rRNA synthesis but their
role is not well dened. It has been shown in yeast that
RNase MRP is involved in the cleavage reaction upstream
of 5.8S rRNA, and RNase P could be involved in cleaving
downstream of the 5.8S rRNA (see Figure 2). Processing
complexes form early after transcription initiation; they
can be visualized under the electron microscope and ap-
pear as granules (or terminal knobs) at the edge of the
so-called Christmas trees in chromatin spreads of rRNA

2
 snoRNAs: Biogenesis, Structure and Function

5ETS ITS1 ITS2 3ETS

18 S 5.8S 2528 S

Guide snoRNPs
( 2-O-ribose methylation,
pseudouridylation)

Me Me Me

Processing snoRNPs
(processome)

U3 U14 snR10 snR30

/U17

E3 U22 U8

Me Me Me

U13 MRP RPR1

Me Me Me

18 S 5.8S 2528 S

Figure 2 Pre-rRNA modification and processing. The large pre-rRNAs encode the small ribosomal subunit rRNA (18S) and the large ribosomal subunit
rRNAs (5.8S and 2528S). The coding sequences are separated by spacer sequences named 5ETS (5 external transcribed spacer), ITS1 (the internal
transcribed spacer 1), ITS2 (internal transcribed spacer 2), and 3ETS (3 external transcribed spacer). The guide snoRNPs are involved in modification
reactions; C/D snoRNPs direct 2-O-ribose methylation (Me), and H/ACA snoRNPs direct pseudouridylation (C). The processing snoRNPs (represented by
spheres) are involved in cleavage reactions. Cleavages are likely orchestrated by a large complex containing many snoRNPs, which is referred to as the
processome (by analogy with the spliceosome). The transcribed spacers are removed by a series of endo- and exonucleolytic reactions, and mature rRNAs
are liberated and packaged into ribosomal subunits.

transcription units. Terminal knob formation and subse-
quent processing requires the U3 snoRNA and all of its
associated proteins; this very large U3 complex (
80S)
contains over 30 dierent proteins and it has been coined
the SSU processome. Because of its large size and function
in early processing events, it is very likely that the SSU
processome corresponds to the terminal knob seen at the 5
end of nascent pre-rRNA transcripts. See also: mRNA
splicing: regulated and dierential; Nucleotides: uncom-
mon, modied and synthetic; rRNA structure
The box C/D snoRNAs generally contain long se-
quences (1021 nucleotides) that are complementary to
rRNA. These sequences are located upstream from box D
(or D) and specify the site of 2-O-methylation in rRNA:
the position complementary to the fth nucleotide from
box D gets methylated (Figure 1a). The box H/ACA sno-
RNAs select the target rRNA sequences by forming two
short helical structures (310 base pairs) that are separated
by two unpaired rRNA nucleotides. The rst unpaired
base in the rRNA sequence (in a 5 to 3 orientation) is a U

3
 snoRNAs: Biogenesis, Structure and Function
residue that gets isomerized into C. The C is always po-
sitioned 1416 nucleotides from box H or ACA (Figure 1b).
snoRNA-associated Proteins
The snoRNAs function in the form of ribonucleoproteins
(RNPs). Box C/D snoRNAs share four common proteins:
Fibrillarin (named Nop1p in yeast), NOP5/NOP58,
NOP56 and 15.5 K (Snu13p in yeast). Fibrillarin is an
abundant nucleolar protein, and an autoimmune antigen
in humans. It is also well conserved among eukaryotes, and
even has a homologue in archaebacteria. Nop1p is essential
for growth in yeast, and mutations in Nop1p can prevent
pre-rRNA processing and 2-O-methylation. A methylt-
ransferase-like domain has been identied in Nop1 from
the archaebacterium Methanococcus jannaschii, and it is
assumed that this protein is responsible for the methylase
activity of C/D snoRNPs. The yeast Nop5/Nop58p and
Nop56p proteins are essential for growth, and depletion of
Nop5/Nop58p leads to decreased levels of C/D snoRNAs
and defects in pre-rRNA processing. Snu13p (15.5 K) has
been shown to bind the K-turn motif formed by boxes C
and D, and this protein plays a nucleating role in RNP
assembly.
Box H/ACA snoRNAs, including the H/ACA-like do-
main of vertebrate telomerase RNA (see above), also
share four common proteins: dyskerin (named Cbf5p in
yeast), GAR1, NHP2 and NOP10. These four proteins
can form a complex before binding H/ACA snoRNAs.
Cbf5p has strong homology to the Escherichia coli pro-
tein truB, a pseudouridine synthase that converts U55
into C55 in most tRNAs. Thus, Cbf5p could be the
enzyme responsible for the U to C isomerization in
pre-rRNAs. Apart from the C synthase signature,
dyskerin and Cbf5p contain a nuclear localization signal
(NLS) and an acidic/lysine-rich domain (KKE/D mo-
tifs), which is also found in other nucleolar proteins. In-
terestingly, mutations in dyskerin cause a genetic disease
called dyskeratosis congenita, which is characterized by
a progressive bone marrow failure, abnormal skin pig-
mentation, nail dystrophy and mucosal leucoplakias;
note that mutations in the RNA component of human
telomerase have been linked to a dominant, autosomal
form of the disease. GAR1 is composed of a highly con-
served core domain, which contains all the elements
required for its subnuclear localization and function.
The core region is anked by glycine- and arginine-rich
(GAR) domains; these domains (as well as the GAR
domain of brillarin) interact with survival of motor
neuron (SMN), a protein that plays a critical role in
the biogenesis of various nuclear RNPs. NOP10 is a
small protein of about 60 amino acids with no known
motif. NHP2 is very similar to 15.5 K (Snu13p), and
these proteins could have evolved from a common an-
cestor (see below). In contrast to 15.5 K, NHP2 does not
4
bind specically to RNA in vitro, but gains specicity
when associated with the other H/ACA proteins. All four
H/ACA proteins are essential for cell viability and pre-
rRNA maturation. Except for GAR1, these proteins
are also necessary for H/ACA snoRNA accumulation
in vivo.
It has been known for sometime that human RNase
MRP and RNase P share the Th/To antigen, a protein of

40 kDa that is recognized by autoimmune sera of pa-
tients suering from scleroderma. However, most of the
known protein constituents of these RNases have been
discovered in the yeast S. cerevisiae. The rst characterized
protein subunit of yeast RNase MRP was Snm1p: this
protein is MRP-specic, it is essential for growth, and it
contains a zinc-cluster domain, a leucine zipper motif and a
serine/lysine-rich tail. Yeast RNase MRP shares many
common proteins with RNase P, namely Pop1p, Pop3p,
Pop4p, Pop5p, Pop6p, Pop7p (Rpp20 in humans), Pop8p
and Rpp1p (Rpp30 in humans); each one of them is
essential for viability. Rpr2p is RNase P-specic and es-
sential for its activity. Human proteins Rpp30 and Rpp38
cross-react with autoimmune patient sera, and it has been
proposed that Rpp38 (
38 kDa) could be the Th/To
antigen. See also: Autoimmune disease; Autoimmune
disease: genetics
Expression of snoRNAs
Eukaryotic cells have developed dierent strategies to ex-
press snoRNAs. Indeed, snoRNAs are not only transcribed
from independent genes, but they can also be encoded
within introns of pre-messenger RNAs (pre-mRNAs), as
gene clusters or intronic gene clusters (Figure 3). See also:
Messenger RNA in eukaryotes
snoRNAs encoded by independent genes
Most snoRNAs are encoded by independent genes in yeast
(S. cerevisiae). However, there are seven intronic genes and
ve polycistronic snoRNA gene clusters in this organism.
Most of the snoRNAs produced from independent genes
are transcribed by RNA polymerase II and thus contain a
7-methylguanosine (m7G) cap at their 5 end. The m7G
gets further methylated to form a 2,2,7-methyl guanosine
(m3G) cap, also called TMG (trimethylguanosine) cap.
This particular cap structure is also typical of spliceosomal
snRNAs and telomerase RNA. There are some exceptions
however, like the 7-2/MRP RNA and plant U3 snoRNA,
which are produced by RNA polymerase III and therefore
have no conventional cap structure. In humans and yeast,
U3 is transcribed by RNA polymerase II. Note that yeast
MRP RNA may be transcribed by RNA polymerase II.
See also: Caps on eukaryotic mRNAs; mRNA splicing:
role of snRNAs; RNA polymerase II holoenzyme and
transcription factors; Spliceosomal machinery

Independent P P
Intronic P Exon Exon
Gene cluster P Plants
Intronic
gene cluster P Exon
Figure 3 Expression of snoRNAs. Various types of snoRNA production are illustrated: synthesis from independent genes, pre-mRNA introns, gene clusters
and intronic gene clusters. The promoter regions (P) are shown as white boxes, exons (E) as grey boxes, snoRNA coding sequences as coloured boxes and
snoRNAs as coloured arrows.
snoRNAs encoded within pre-mRNA introns
In contrast to yeast snoRNAs, the great majority of verte-
brate snoRNAs are encoded within introns of pre-mRNA.
Of all known human snoRNAs, only four (U3, U8, U13 and
7-2/MRP) are encoded by independent genes. A pre-mRNA
can harbour many snoRNAs (generally one snoRNA per
intron but this can vary; see below), which allows the pro-
duction of a large number of snoRNAs from a relatively
small number of genes. The genes containing intronic sno-
RNAs usually encode proteins involved in ribosome syn-
thesis or function, i.e. nucleolar proteins, ribosomal proteins
or translation factors. However, some snoRNA host genes
do not encode proteins; their spliced poly(A)+ RNA has no
protein-coding potential, thus the functional product of
these genes is contained in the introns only. Interestingly, the
primary transcripts of genes hosting snoRNAs usually start
with a C residue followed by a stretch of pyrimidines, which
is characteristic of the 5-terminal oligopyrimidine (5 TOP)
mRNAs coding for ribosomal and some house-keeping
proteins. It is known that the 5 TOP mRNAs are transi-
tionally regulated, but since some snoRNA hosts do not
code for proteins, it is possible that the 5 oligopyrimidine
tracts at the transcription start site also function as tran-
scriptional signals in order to regulate the production of
snoRNAs with that of proteins of the translational appa-
ratus. The observation that dierent organisms can encode a
particular snoRNA from a dierent gene strongly suggests
that intronic snoRNAs are mobile. See also: mRNA
snoRNAs: Biogenesis, Structure and Function
Yeast
Metazoa
Plants
Animals
Exon Exon Exon Yeast
Plants
Yeast
Trypanosomes
Exon Plants
splicing: regulated and dierential; Protein-coding seg-
ments: evolution of exonintron gene structure
Polycistronic snoRNAs
In plants, most snoRNAs are produced from polycistrons,
thus more than one snoRNA is found in a transcription
unit. In yeast for example, snR190 and U14 are cotran-
scribed; endonucleolytic cleavage of the polycistronic
RNA by Rnt1p (the orthologue of bacterial RNase III)
generates pre-snR190 and pre-U14 RNAs, which get fur-
ther trimmed at their 5 and 3 ends to produce mature
snoRNAs. The most striking examples of polycistronic
snoRNAs are found in trypanosomes: up to nine dierent
snoRNAs have been found in a single polycistron (which is
repeated multiple times in tandem), and since no exons are
present in the inter-snoRNA regions, it is quite certain that
endonucleolytic cleavages, not splicing, are responsible for
the production of individual snoRNAs. Exceptionally in
plants, some snoRNAs are found as intronic gene clusters.
In contrast to vertebrate and yeast, there is more than one
snoRNA per intron. See also: Gene expression in plants;
Gene expression in yeast
snoRNA Maturation Pathways
Rapid and ecient production of large quantities of sno-
RNAs may be crucial for cell survival. The snoRNAs
5
 snoRNAs: Biogenesis, Structure and Function
expressed from independent gene units have strong pro-
moters, and the intronic snoRNAs are usually found in
highly expressed genes. Upon splicing, the excised intron is
released in the form of a lariat. After debranching of the
intron lariat, the intronic RNA is degraded from both 5
and 3 ends by exonucleases acting in the 5!3 and in the
3!5 directions. The exonucleolytic trimming of introns
stops at the borders of the snoRNA, most likely due to the
snoRNA sequence being packaged into an snoRNP.
Processing of plant intronic snoRNA gene clusters re-
quires endonucleolitic activity, with or without splicing of
the intron. Specic signals are required for faithful process-
ing of intronic snoRNAs, namely boxes C and D, and the
terminal helix for C/D snoRNAs, and boxes H and ACA,
and the adjacent stems for H/ACA snoRNAs. These sig-
nals may serve as binding sites for snoRNP proteins. The
general mechanism for the production of intronic snoRNA
has some exceptions. There are examples where snoRNAs
located in poorly spliced pre-mRNAs are excised from the
intron by endonucleases, and the remaining extra nucleo-
tides anking the snoRNA are then removed by exonuc-
leases. Since the nal steps of maturation are the result of
exonucleolytic trimming, intronic and polycistronic sno-
RNAs have a monophosphate at their 5 end and a free
hydroxyl group at their 3 end. See also: Genes: denition
and structure
Small Cajal Body-specific RNAs
Cajal bodies (CBs) are spherical nuclear organelles that
are often seen juxtaposed or near the nucleolus. CBs ap-
pear to be sites of biogenesis or recycling of various
RNPs. These organelles harbour a specic subset of
small RNAs called scaRNAs (small Cajal body-specic
RNAs) that are structurally and functionally indistin-
guishable from snoRNAs. Some of them contain one
C/D or one H/ACA domain, like U90 and U91 or U92,
respectively. There is also one scaRNA harbouring two
H/ACA domains in tandem (U93). However, most
scaRNAs are hybrids since they are composed of one
H/ACA and C/D core domain, such as U85, U87, U88
and U89. The scaRNAs are believed to contain all of the
RNP proteins known to be associated to C/D and H/
ACA snoRNAs, although this assumption has only been
veried for brillarin and GAR1 proteins.
Most scaRNAs are implicated in post-transcriptional
modications of polII-transcribed spliceosomal snRNAs
(U1, U2, U4 and U5). These modication events take place
in the CBs, with the exception of the polIII-transcribed U6
snRNA, which is most probably modied in the nucleolus
by snoRNAs. The H/ACA scaRNAs possess two CB lo-
calization signals (CAB boxes) that are located in the ter-
minal loop of the 5 and 3 hairpins. The consensus
sequence of the CAB boxes is ugAG, where the rst ug
nucleotides are variable but the last AG nucleotides
6
are highly conserved. These motifs have been found in
humans, Drosophila and Arabidopsis, although the impor-
tance of the CAB boxes has not been conrmed experi-
mentally in the last two organisms.
It has been observed that box C/D snoRNAs assemble
or mature in CBs before transiting to the nucleolus. Some
data tend to show that it could be the same for H/ACA
snoRNAs. For example, over expressed U64 snoRNA in
transiently transfected HeLa cells appears rst in CBs and
later in nucleoli, and human telomerase RNA, which con-
tains an H/ACA-like domain, transits to CBs before ac-
cumulating in the nucleoli of transfected Xenopus oocytes.
For H/ACA scaRNAs, which reside and meet their targets
in the CBs, the CAB box could serve as a retention signal.
No CAB box has been identied in scaRNAs of the C/D
class, however, they most probably contain a specic el-
ement that allow their retention in CBs, whereas genuine
C/D snoRNAs are able to accumulate in the nucleolus.
Possible Implications of snoRNAs-
modified Targets
As seen above, the majority of snoRNAs act as guides for
the site-specic modication of rRNAs. The function of
modied nucleosides in rRNAs is not known; however, the
modied residues are clustered in evolutionarily conserved
and functional regions of rRNAs, and it has been hypoth-
esized that they could be required for folding or some other
aspects of ribosome function. Likewise, many modied
nucleosides are found in functionally important regions of
spliceosomal snRNAs, supporting the idea that modica-
tions confer important folding or other functional advan-
tages. Addition of methyl groups generates more
hydrophobic surfaces and usage of C residues increases
the hydrogen bonding potential (three H bonds instead of
two). These properties might contribute to rRNA folding
or interaction with proteins. In the same order of idea,
processing snoRNAs could serve as chaperons necessary
for proper rRNA folding which might be essential for
cleavage reactions. One can also imagine that modica-
tions such as 2-O-methylation are made to protect crucial
regions of rRNA from endonucleolytic degradation. Al-
ternatively, the use of snoRNAs that base pair throughout
the molecule could ensure proofreading and proper folding
of critical rRNA sequences, and modications would only
constitute check marks. Eukaryotic rRNAs are bigger and
contain many more modications then prokaryotic
rRNAs; larger (eukaryotic) rRNAs may require more
quality control checkpoints. It is worth noting that deletion
of individual or even multiple modication guide sno-
RNAs is not lethal in yeast, suggesting that modications
are not critical for function; however it is not known what
would happen if tens of snoRNAs were missing. See also:
RNA editing; RNA synthesis

Origin and Evolution of snoRNAs
snoRNAs are highly conserved throughout evolution.
They are found in mammals, amphibians, shes, plants,
yeast, trypanosomes and even archaebacteria. Archaeal
box C/D and H/ACA small RNPs (sRNP) share the RNA-
binding protein L7Ae, the homologue of yeast Snu13p and
Nhp2p that are associated with C/D and H/ACA sno-
RNAs, respectively. L7Ae binds the K-turn motif, which is
present in both C/D and H/ACA sRNAs. Snu13p an
Nhp2p possibly evolved from L7Ae but Nhp2p lost its
specicity for the K-turn motif and regained specicity for
H/ACA snoRNAs by associating with the other H/ACA
proteins. The archaeal H/ACA sRNAs also have a dier-
ent structure than eukaryotic snoRNAs: they usually
present one or three hairpins generally without box H.
H/ACA sRNAs that contain three hairpins harbour two
ACA motifs or one ACA box with a degenerate AGA box.
In trypanosomes, which are eukaryotes of ancient origin,
snoRNAs are similar to sRNAs from archaebacteria. In-
deed, known H/ACA snoRNAs from trypanosomes har-
bour only one hairpin, no box H and a degenerate AGA
box. Interestingly, a trypanosome-specic H/ACA sno-
RNA (SLA1) mediates pseudouridylation of the spliced
leader (SL) RNA, which is the donor of the leader sequence
found in all trypanosome mRNAs. The SL RNA is added
to the pre-mRNAs via trans-splicing.
Another fascinating feature of the snoRNA world is the
presence of brain-specic snoRNAs in mice and humans.
Very little is known about the function of those snoRNAs.
Mouse snoRNAs MBI-36, MBII-48, MBII-52 and MBII-
85 are all expressed in the hippocampus and the amygdala.
These two regions are crucial for contextual memory con-
solidation. Moreover, the mouse and human H/ACA
snoRNAs (MBI-36 and HBI-36, respectively) are intron-
encoded in the brain-specic serotonin 2C receptor gene.
The three human C/D box snoRNAs (HBII-13, HBII-52
and HBII-85) map to a chromosomal region implicated in
snoRNAs: Biogenesis, Structure and Function
the PraderWilli syndrome, a neurobehavioural disorder.
Moreover, HBII-52 has an 18-nucleotide complementarity
to the serotonin 2C receptor implicated in memory con-
solidation. These ndings tend to link these brain-specic
snoRNAs to learning and memory.
The very large number of snoRNAs, the diversity of
their structure and biological roles (modication of
rRNAs, tRNAs, mRNAs and snRNAs) in addition to
the fact that they are highly conserved throughout evolu-
tion may reect the ancient RNA world, where biochem-
ical reactions were carried out by RNA molecules.
Further Reading
Bachellerie J-P, Cavaille J and Huttenhofer A (2002) The expanding
snoRNA world. Biochimie 84: 775790.
Brown JW, Echeverria M and Qu LH (2003) Plant snoRNAs: functional
evolution and new modes of gene expression. Trends in Plant Science 8:
4249.
Decatur WA and Fournier MJ (2003) RNA-guided nucleotide modi-
cation of ribosomal and other RNAs. Journal of Biological Chemistry
278: 695698.
Filipowicz W and Pogacic V (2002) Biogenesis of small nucleolar rib-
onucleoproteins. Current Opinion in Cell Biology 14: 319327.
Gerbi SA, Borovjagin AV and Lange TS (2003) The nucleolus: a site of
ribonucleoprotein maturation. Current Opinion in Cell Biology 15:
318325.
Granneman S and Baserga SJ (2004) Ribosome biogenesis: of knobs and
RNA processing. Experimental Cell Research 296: 4350.
Kiss T (2004) Biogenesis of small nuclear RNPs. Journal of Cell Science
117: 59495951.
Rogelj B and Giese KP (2004) Expression and function of brain specic
small RNAs. Reviews in the Neurosciences 15: 185198.
Tran E, Brown J and Maxwell ES (2004) Evolutionary origins of the
RNA-guided nucleotide-modication complexes: from the primitive
translation apparatus? Trends in Biochemical Sciences 29: 343350.
Uliel S, Liang XH, Unger R and Michaeli S (2004) Small nucleolar
RNAs that guide modication in trypanosomatids: repertoire, targets,
genome organisation, and unique functions. International Journal for
Parasitology 34: 445454.