ABSTRACT
The compact genome of
Fugu rubripes
, with its very small introns, appears to be particularly suitable to study
intron-encoded functions. We have analyzed the
Fugu
gene for ribosomal protein S7 (formerly S8, see Note), whose
Xenopus
homolog contains in its introns the coding sequences for the small nucleolar
RNA U17. Except for intron length, the organization of the
Fugu
S7
gene is very similar to that of the
Xenopus
counterpart. The total length of the
Fugu S7
gene is 3930 bp, compared with 12691 bp for
Xenopus
. This length difference is uniquely due to smaller introns. Although short, the six introns are longer than the
~
100 bp size of most
Fugu
introns, as they host U17 RNA coding sequences. While four of the six U17 sequences are `canonical', the remaining two represent diverged U17 pseudocopies. In fact, microinjection in
Xenopus
oocytes of
in vitro
synthesized
Fugu
transcripts containing the `canonical' U17f sequence results in efficient production of mature U17 RNA, while injection of a transcript
containing the U17
[psi]b sequence does not.
The coding sequences for most small nucleolar RNAs (snoRNAs) are localized in the introns coding for ribosomal proteins (r-proteins) or for other housekeeping proteins involved in the production
and function of the ribosome (for reviews see
1
-
3
). These intron-encoded snoRNAs are not produced by independent transcription but by
processing of the host gene pre-mRNA (for references see
4
-
9
). One of these snoRNAs, U17 RNA, has been found to be encoded in two introns of
the human gene for RCC1 (
7
) and in each of the six introns of the
Xenopus
gene for r-protein S7 (formerly S8, see Note;
8
). It seemed of interest to study the organization of the U17 RNA coding
sequence in the genome of the fish
Fugu rubripes
, a powerful model system for studying intron-encoded functions. In fact, it has a particularly small genome, ~400 Mb in length, about 7.5 and 10 times smaller than the human and
Xenopus
genomes respectively (
10
). This genome compactness reflects short gene distances and, more relevant for
our purposes, short gene size due to particularly small introns (
10
,
11
). This being the rule, some notable exceptions have been found of long introns
comparable with human ones (
12
-
14
). It has been suggested that these exceptions are due to specific structural or
functional properties of some intron sequences.
We present here an analysis of the
S7/U17
gene organization and show that the compact
Fugu
genome with its short introns can be particularly useful for identifying and
studying intron-nested snoRNA coding sequences.
Unless stated otherwise, all techniques used for preparation, analysis and manipulation of DNA and RNA were performed according to standard
laboratory manuals (
15
). Sequencing was performed by the dideoxy chain termination method (
16
) on both strands of overlapping fragments.
A cosmid library (
11
) has been probed with
Xenopus
S7
cDNA (
17
) and snoRNA U17 (
8
) fragments. The selected cosmid 85F12 was digested by
Eco
RI to generate a 3526 bp fragment containing the central portion of the gene and
by
Dra
I to generate an overlapping ~2500 bp fragment containing the 5'-region of the gene. These fragments have been cloned in the
pEMBL18 and Bluescript KS(+) vectors respectively (clones pF-S7.1 and pF-S7.2). The 3'-region of the gene has been obtained by inverse PCR
using two oligonucleotide primers corresponding to nt 3850-3868 and 3882-3899 (Fig.
1
); as template we used cosmid 85F12 DNA, previously digested with
Bam
HI and circularized. The ~450 bp fragment obtained was cloned in the PCRII vector (Invitrogene) to
generate the pF-S7.3 clone. Plasmid pF-S7.11, containing U17c sequence, has been previously described (
18
). Plasmid pFS7.10, containing the U17[psi]b sequence, was obtained by cloning a DNA fragment encompassing the second
intron, including the U17[psi]b sequence and the two flanking exons of the
Fugu
S7
gene, into the PCRII vector. This fragment was generated by PCR amplification
of the region between nt 392 and 1313 of the
S7
gene (see Fig.
1
). For sequencing, plasmids pF-S7.1 and pF-S7.2 were digested with various restriction enzymes to generate
overlapping fragments and subcloned in Bluescript KS(+). A fragment of the rRNA
gene encompassing the 18S rRNA region complementary to U17 RNA was PCR
amplified using two primers corresponding to regions well conserved among
vertebrates, nt 568-589 and 1092-1112 according to the
Xenopus
18S rRNA gene sequence (
19
).
To obtain the transcripts to be used as processing substrates in microinjected
oocytes, 1 [mu]g pF-S7.11 DNA was digested with
Eco
RI and transcribed with T7 RNA polymerase, while 1 [mu]g pF-S7.10 was digested with
Xba
I and transcribed with SP6 RNA polymerase, in the presence of 50 [mu]Ci [[alpha]-
32
P]UTP as described (
20
). Transcripts of 1048 nt (IIi) and 592 nt (IIIi) were obtained respectively
from plasmid pF-S7.10 and pF-S7.11. After transcription and DNase digestion, the RNAs were
purified by phenol/chloroform/isoamyl alcohol (50:50:1) extraction and ethanol
precipitation and resuspended in H
2
O for microinjection.
Isolation of
Xenopus
stage V-VI oocytes and microinjection of RNA into the germinal vesicle were
carried out essentially as previously described (
21
). Oocytes were injected with 40 nl H
2
O containing ~80 000 c.p.m. (corresponding to 10-40 ng) of the
in vitro
transcribed RNAs and incubated for increasing time intervals at 22oC. After incubation, nuclei from pools of 10 oocytes were manually prepared (
21
) and then lysed in 300 [mu]l 100 mM Tris, pH 7.5, 300 mM NaCl, 10 mM EDTA, 2% SDS, containing 1 mg/ml
proteinase K (
22
). RNA was extracted and analyzed by 6% polyacrylamide-8 M urea gel electrophoresis according to standard procedures.
A
F.rubripes
cosmid library (
11
) was probed at low stringency with a
Xenopus
cDNA specific for r-protein S7 (
17
) and a
Xenopus
genomic fragment containing a U17 snoRNA gene copy (
8
). The finding of clones hybridizing to both probes suggested that the U17 RNA
coding sequence is hosted in the
S7
gene of
Fugu
, as occurs in
Xenopus
. The selected cosmid 85F12 was digested with
Eco
RI and analyzed by Southern blot with the same probes as above. The positive
Eco
RI 3.5 kb fragment, corresponding to most of the gene, was subcloned and
sequenced. Completion of the 5'- and 3'-regions was obtained as described in Materials and
Methods. The sequence of the entire
Fugu
rp-S7
gene with its flanking sequences is shown in Figure
1
.
Sequence comparison with the
Xenopus
r-protein
S7
cDNA (
17
) and gene (
23
) allowed precise determination of the exon-intron boundaries in the
Fugu
gene. The 5'-end of the first exon (transcription start site) and the 3'-end of the last exon have not been experimentally determined, however, the very good
sequence conservation in the regions surrounding the two sites allows their tentative
identification in
Fugu
by comparison with
Xenopus
. The
Fugu S7
gene is made up of seven exons and six introns, as is its
Xenopus
homolog, and the positions of the six introns are also perfectly conserved
between the two species. The overall size of the
Fugu S7/U17
gene is 3930 bp, compared with the 12691 bp
Xenopus
counterpart. Exon size ranges from 27 to 151 bp, as in
Xenopus
, while introns range from 339 to 920 bp, shorter than the 1057-4645 bp observed in
Xenopus
. Thus, the more compact
Fugu
gene organization is uniquely due to the smaller size of the introns, while the
coding regions for the r-protein S7 are identical in the two species. Comparison of the coding
sequences (not shown) indicates a very high homology between
Xenopus
and
Fugu
at the protein level (95.9%) and a somewhat lower homology at the nucleotide
level (82.4%), due to several silent nucleotide substitutions, mainly in the
third position of codons.
All vertebrate r-protein genes analyzed up to now have their transcription start site
located within a 12-25 nt pyrimidine sequence, so as to transcribe mRNAs always starting at
their 5'-end with a 6-12 pyrimidine sequence, which has been implicated in the
translational regulation of this class of mRNAs (for references see
24
,
25
). The
Fugu
S7
gene also follows this rule: the presence of two G residues `contaminating' the
pyrimidine sequence is not unusual, as a similar situation has also been
described for other r-protein genes.
Comparison of the 5'-region of the
Fugu S7
gene with the promoter regions of other vertebrate r-protein genes has revealed the presence of at least two relevant
sequences, indicated in Figure
1
. At position -72 the sequence 5'-ACTTCCTGCG is present, also found in the promoters of other
vertebrate r-protein genes and shown to be responsible for binding of a transcription
factor, called [beta] in the mouse (
26
,
27
) or XrpFI in
Xenopus
(
28
). Moreover, the sequence 5'-GGCCGTCGTT at +11 shows high homology to an element located at the
same position in the
Xenopus
S7
gene and described in the promoters of some mouse r-proteins genes, where it has been shown to be responsible for binding of
the [delta] transcription factor (
26
,
27
).
As already suggested by the hybridization data (see above), the introns of the
S7
gene host the coding sequences for U17 RNA. Figure
1
shows, underlined, four easily identifiable U17 sequence copies, one in each of
the last four introns (copies c-f, named according to their intron localization). The four sequences are
compared in Figure
2
. Their homology is very high (~98% using copy f as reference) and somewhat lower (~73%) in comparison with the f copy of
Xenopus
U17 sequence.
The computer-derived secondary structure of
Fugu
U17 RNA, shown in Figure
3
, is in excellent agreement with that proposed for
Xenopus
U17 RNA (
8
). Two of the very few nucleotide differences among the four
Fugu
sequences represent compensatory substitutions that leave the secondary structure unaltered. In fact, as
indicated in Figure
3
, in U17c nucleotide 2 is a U, matching with an A at position 101, while in the
other three U17 copies there is a C base paired with a G. The other few
substitutions occurred in unpaired regions. The comparison between the
Fugu
and
Xenopus
sequences shows many more differences; in this case also they either involve
nucleotides not implicated in base pairing or are compensatory nucleotide changes.
Several snoRNAs contain regions of complementarity to rRNA, probably implicated
in interactions at these sites. In particular, it was proposed that
Xenopus
U17 snoRNA has two boxes complementary to regions of the 18S rRNA and of the
ETS respectively (
8
). These sequences are conserved in the four copies of
Fugu
U17 sequence and are boxed in Figure
2
(rRCSI and rRCSII, rRNA complementary sequences I and II). We have also cloned
and sequenced a fragment of the
Fugu
18S rRNA gene encompassing the rRCSI complementary sequence (not shown) that is
also conserved between
Xenopus
and
Fugu
. Thus the complementarity between U17 RNA and 18S rRNA is maintained.
Careful inspection of the first two introns revealed the presence of two
degenerate U17 sequences: copy [psi]a in the first intron and [psi]b in the second. The sequences are dashed underlined in Figure
1
and compared with the canonical U17 sequences in Figure
2
. The relation to the U17 sequence is indicated by the presence of some
conserved sequence blocks: copy [psi]b presents somewhat better matches than copy [psi]a. Notice that some blocks are more conserved between the two
pseudocopies than between these and the other copies. Attempts to generate computer-derived secondary structures for these two U17 pseudocopies comparable with the canonical
one have failed. The absence or poor conservation of the rRCS elements also strongly suggests that these two sequences represent pseudogene copies.
The correct and efficient processing of various intron-encoded snoRNAs by heterologous systems indicates that the processing
mechanism is evolutionarily very well conserved among all vertebrates (
2
). In particular, we have shown that U17 RNA production by processing of precursor transcripts, mainly due to 5' and 3' exonucleolytic activities, is conserved between fish and
amphibians (
18
). In fact, a radioactive RNA precursor containing a
Fugu
U17 sequence is correctly and efficiently processed to yield mature U17 RNA
when microinjected into the germinal vesicle of
Xenopus
oocytes. Now, to verify that the two degenerate U17[psi]a and [psi]b sequences, present in the first two introns, are indeed pseudogenes,
we microinjected the following radioactive RNAs: (i) a 1048 bp transcript (IIi)
containing the entire intron 2, including the less degenerate U17[psi]b copy sequence and part of the flanking exons; (ii) a 592 bp transcript
(IIIi) containing the entire intron 3, including the canonical U17c copy
sequence and part of the flanking exons. Figure
5
A shows that the injected transcript IIIi is correctly processed to produce
mature U17 RNA; in contrast, transcript IIi does not produce any stable RNA and
is completely degraded.
A Northern blot analysis has been carried out on total RNA isolated from
different
Fugu
tissues (not shown). The results obtained prove that U17 RNA is expressed in
the adult fish and provide an estimate of its relative abundance that seems to
be more or less the same in all the somatic and germinal tissues examined. As
for most other snoRNAs, the presence of U17 RNA in normal tissues and its
conservation in evolution are the only indications that it has a functional
role, the specific function remaining to be determined.
It has been shown that the fish
F.rubripes
has, among vertebrates, a particularly compact genome, approximately eight times smaller than that of mammals (
10
,
11
). One of the reasons for this compactness is the small size of most introns,
which have a modal length of <100 nt. It seemed to us that this situation is particularly suitable for the
study of the structural organization of those genes which contain, nested in
their introns, the coding sequences for snoRNAs. Since the intron coding
arrangement was revealed for mouse U14 RNA, the number of intron-encoded snoRNA genes in various vertebrates has grown to >15 during the
last 3 years and is still growing fast (for references see
1
-
3
). All these snoRNA sequences are hosted in the introns of genes coding for r-proteins or for other proteins involved in the production and function of
the translation apparatus. In some cases specific snoRNAs are encoded in
different host genes in different organisms. In particular, U17 RNA coding
sequences have been found in all six introns of the gene for r-protein S7 in
Xenopus
(
8
) and in the first two introns of the human
RCC1
gene (
7
). We have shown here that in
Fugu
the same snoRNA is encoded, as in
Xenopus
, in the introns of the r-protein
S7
gene. However, only four copies (c-f, in introns 3-6) appear to be canonical on the basis of their conserved sequence
and computer-derived secondary structure. For the c copy, we have shown that the
corresponding precursor transcript, when microinjected into
Xenopus
oocytes, is correctly processed to produce mature U17 RNA (
18
). On the other hand, two other copies, [psi]a and [psi]b, located in the first two introns of the host
S7
gene, appear to be degenerate copies of the U17 sequence, [psi]b being somewhat better conserved than [psi]a. This structural divergence is accompanied by a loss of the ability
to be correctly processed; in fact, microinjection into
Xenopus
oocytes of a transcript corresponding to the second intron and containing the
U17[psi]b sequence did not result in the production of any stable RNA, but in
complete degradation of the injected precursor.
The presence of these U17 coding sequences seems to be the reason for the
relatively large size of these six introns, ranging between 339 and 920 bp;
although much smaller than in
Xenopus
, this size is larger than that reported for the majority of
Fugu
introns (
10
). Thus analysis of the gene structure in this species could facilitate the
study of intron-encoded functions, in part because of the reduced cloning and sequencing
workload, but mainly because the intron length itself, if substantially
exceeding 100 nt, could be indicative of the presence of an intron-specific function.
The results presented here also bear on the problem of evolution of genome
organization. At present it is difficult to decide if the short intron-containing genome of
Fugu
represents an ancestral situation from which longer intron-containing genomes of other vertebrates originated or it is the result of
progressive reduction of an ancient large intron-containing genome. The first view is consistent with the finding that
since the time the two U17[psi]a and [psi]b sequences started to degenerate their two host introns have
maintained a length similar to that of the four introns containing conserved
U17 RNA sequences and several times larger than the average
Fugu
intron. However, one might object that not enough time has yet elapsed for an
appreciable reduction in intron size to have occurred.
In previous papers we have used the ribosomal protein numbering system
introduced in our first study of
Xenopus
r-proteins (
29
). The large amount of sequencing data now accumulated allows us to adopt, as a
unified nomenclature, the rat system (
30
). Thus the r-protein that we previously designated S8 is now identified as S7 for both
Xenopus
and
Fugu
.
This work was supported by grants from Progetto Finalizzato Ingegneria Genetica,
CNR and from Ministero Università e Ricerca Scientifica e Tecnologica. Part of this work has been carried out in the Department of Medicine of the University of Cambridge
Clinical School, UK, by G.Cesareni, with the support of a CNR fellowship.
REFERENCES
Return