ABSTRACT
Ro ribonucleoproteins are composed of Y RNAs and the Ro 60 kDa protein. While
the Ro 60 kDa protein is implicated in an RNA discard pathway that recognizes 3
'
-extended 5S rRNAs, the function of Y RNAs remains unknown [O'Brien,C.A.
and Wolin,S.L. (1995)
Genes Dev.
8, 2891-2903]. Y5 RNA occupies a large fraction of Ro 60 kDa protein in human Ro
RNPs, contains an atypical 3
'
-extension not found on other Y RNAs, and constitutes an RNA antigen in
certain autoimmune patients [Boulanger
et al
. (1995)
Clin. Exp. Immunol.
99, 29-36]. An overabundance of Y RNA retroposed pseudogenes has previously
complicated the isolation of mammalian Y RNA genes. The source gene for Y5 RNA
was isolated from human DNA as well as from
Galago senegalis
DNA. Authenticity of the hY5 RNA gene was demonstrated
in vivo
and its activity was compared with the hY4 RNA gene that also uses a type 3
promoter for RNA polymerase III. The hY5 RNA gene was subsequently found to
reside within a few hundred thousand base pairs of other Y RNA genes and the
linear order of the four human Y RNA genes on chromosome 7q36 was determined.
Phylogenetic comparative analyses of promoter and RNA structure indicate that
the Y5 RNA gene has been subjected to positive selection during primate evolution. Consistent with the proposal of O'Brien and Harley [O'Brian,C.A. and Wolin,S.L. (1992)
Gene
116, 285-289], analysis of flanking sequences suggest that the hY5 RNA gene may
have originated as a retroposon.
Use of autoimmune sera from patients with Sjogren's syndrome and systemic lupus
erythematosus led to the discovery of Ro RNPs and their subsequent characterization in mammals,
Xenopus
,
Iguana
and
Caenorhabditis elegans
, although their function remains unknown (
1
-
6
). Y RNAs range in size from 70-115 nucleotides (nt), exhibit a highly conserved secondary structure
motif that is recognized by the Ro 60 kDa autoantigenic protein, and accumulate
in vivo
to moderate levels in the form of Ro ribonucleoproteins (RNPs) (
5
-
9
). The 5' and 3' termini of mammalian Y RNAs as well as their [alpha]-amanatin sensitivity indicate that they are
transcribed by RNA polymerase (pol) III (
8
,
10
-
13
). Although TATA-like sequences upstream of some of the candidate Y RNA genes isolated from
several organisms are consistent with transcription by pol III, only a few have
been shown to contain a consensus proximal sequence element (PSE) motif that
comprises the core of the type 3 promoter for pol III (
5
,
6
,
8
,
9
,
11
,
12
). While mammalian and iguana Y RNAs end in 3' uridylates, a characteristic terminus indicative of transcription by pol
III and a binding site for the transcription termination factor La,
Xenopus
and
C.elegans
Y RNAs do not (
5
,
6
,
8
,
9
,
12
,
13
).
The Ro 60 kDa component of Ro RNPs has recently been shown to associate
specifically with non-functional 5S rRNAs as well as Y RNAs (
14
). These errant 5S transcripts carry internal mutations as well as 3'-extensions, the latter of which is the result of ineffective
termination by pol III at its usual 5S rDNA transcription termination site.
Impaired accumulation of these errant 5S rRNAs led to the idea that Ro 60 kDa
functions in an RNA discard pathway that assures quality control of 5S rRNA (
14
). At present it is unknown if a role in 5S rRNA metabolism represents a primary
function for Ro 60 kDa or if Ro-Y RNPs exhibit distinct activities related to the metabolism of other as
yet unidentified RNAs. It was suggested that Y RNAs may have evolved from
faulty 5S transcripts, perhaps to adopt a regulatory role in 5S rRNA metabolism
(
14
). Therefore, it might be informative to examine Y RNA phylogeny. The
invertebrate
C.elegans
expresses only one Y RNA, a homologue of hY3 which appears to be the most
conserved of the Y RNAs (
4
-
6
,
9
). Ro RNPs are heterogeneous in higher vertebrates which express two to four Y
RNAs (
4
,
5
,
9
). The situation in mammals is noteworthy since rodents express only Y1 and Y3
RNAs, while other mammals express these plus either or both of the smaller Y
RNAs, Y4 and Y5 (
4
,
9
,
13
). Although each of the human Y RNAs are synthesized from a single copy gene (
8
,
12
), a great excess of Y pseudogenes are found juxtaposed with human
Alu
repetitive elements (
15
,
16
). The propensity of Y sequences to generate pseudogenes by an RNA-mediated process known as retroposition prompted O'Brien and Harley to
suggest that the smaller Y RNA genes originated as retroposons from the larger Y RNAs, a proposal that reconciles the evolutionary heterogeneity of Y RNAs (
4
,
12
,
15
-
18
). In any case, although the functionality of Y RNAs remains undetermined, their
retroposition as well as involvement in autoimmunity nonetheless indicate that
Y RNAs have had significant impact on human biology.
All four of the human Y RNAs, hY1, hY3, hY4 and hY5 are precipitated by
autoimmune sera by virtue of their association with the autoantigens Ro 60 kDa
and La. However, hY5 RNA appears to be unique in that hY5-specific autoantibodies are directed to the RNA component of Y5 Ro RNPs (
19
). Human Y5 RNA can be further distinguished by an atypical 3' sequence motif which is not found on other Y RNAs, as well as
biochemical properties that are distinct from the other Ro RNPs (
8
,
18
-
20
). Unlike most pol III transcripts, hY RNAs and perhaps especially hY5, maintain
stable rather than transient association with the La antigen transcription
factor (
13
,
21
-
27
). Also, hY5 RNA is significantly over-represented on Ro RNPs relative to hY1 and hY3 although the mechanism
responsible for this pattern of expression is unknown (
4
,
13
,
21
,
28
). Yet, although the sequence of Y5 RNA has been known for sometime (
8
,
29
) and many pseudogenes of Y5 are apparent in the human genome (
9
), the gene encoding hY5 RNA had remained elusive. In order to explore the above
mentioned aspects of hY5 RNA metabolism and to obtain clues to its biology and
evolution, we chose to isolate the gene that encodes hY5 RNA.
The hY1 and hY3 genes are adjacent on a 4 kb fragment of human DNA although the
significance of this close linkage is unknown (
8
). Attempts to isolate additional mammalian Y RNA genes have been impeded by the
abundance of Y RNA pseudogenes (
4
,
16
,
17
). Localization of the hY4 RNA gene to chromosome 7 allowed a targeted approach
to its cloning (
12
). It was demonstrated that the hY1, hY3 and hY4 genes reside on a 200 kb yeast
artificial chromosome (YAC) that maps to chromosome 7. The hY5 RNA gene was
also localized to chromosome 7 by a functional approach but it was not found
associated with the other hY genes and remained to be isolated (
12
). We therefore employed an exhaustive screening of human chromosome 7-enriched DNA to isolate the hY5 RNA gene. Once isolated, physical linkage
between the hY5 gene and the other hY RNA genes was established.
The yWSS2977 clone was isolated from the CEPH mega-YAC library (position 803G01) and therefore was not colony purified upon
screening; the yWSS2977 clone used here was a colony-purified isolate of yWSS2977 (designated yWSS2977.3). Clone yWSS4352 was
also isolated from the CEPH library (position 742G08). Clones yWSS1020,
yWSS1476 and yWSS756 were isolated from a library made from somatic cell hybrid
GM10791 DNA that retains chromosome 7 as its only human chromosome; these
clones represented colony pure isolates, their sizes are relatively small (<300 kb), and the frequency of chimerism of YACs isolated from this library is ~15% (
30
).
A collection of yeast artificial chromosomes (YACs) highly enriched for human
chromosome 7 DNA (
30
-
32
), previously used to isolate the hY4 RNA gene (
12
), were screened by PCR using the primers 5'-AGTTGGTCCGAGTGTT-3' and 5'-GCAAGCTAGTCAAGCG-3', designated hY5-16S and hY5-16AS,
respectively. The expected product represents the first 78 bp of the 84 bp hY5
sequence. DNA from the positive clone yWSS1476 was used as template to isolate
the hY5 RNA gene in two phases similar to that described previously for hY4 (
12
). Phase one consisted of two independent `hemispecific PCR' methods developed
in this lab and described in detail previously (
12
), to obtain 5' and 3' flanking sequence information which was used to design primers to
isolate the hY5 gene by conventional PCR. A 5' sense primer: 5'-CTGAGCCCTCGGCGTCCGCA-3' designated HY55PR20, and a 3' antisense primer: 5'-CGTGTAAATTTTCTTCTCAGGCATTTTGGAGGTTAATACTT-3' designated
HY53P40 readily amplified the expected 0.8 kb fragment from yWSS1476 DNA. This
was cloned into pCRII (Invitrogen) and the recombinant designated phY5. The
plasmid p5'[Delta]hY5 was derived from phY5 by PCR-mediated deletion and subcloned into pCRII. The 5' sense primer: 5'-AGAGACTCACAGGATAACACAGTTGGTCCGA-3' was used with the HY53P40
antisense primer to delete the 5' flanking sequence up to position minus 21 generating p5'[Delta]hY5. phY5 and p5'[Delta]hY5 were verified by sequencing.
Transient expression of hY5 RNA in NIH 3T3 cells was achieved by transfection
with Transfectamine (BRL) as described (
12
). Equal amounts of experimental and control plasmids were co-transfected. Forty-eight hours after transfection, total RNA was isolated and analyzed
by Northern blot (
12
). RNA quantity and integrity were verified by polyacrylamide gel
electrophoresis and staining (not shown).
Total cellular RNA was electrophoresed on 8 M urea/6% polyacrylamide gels, and
transferred to nylon membrane as described (
12
). Probes were labeled either by incorporation of [[alpha]-
32
P]dCTP, into a 78 bp hY5 DNA, or a 78 bp gY5 DNA, or by
32
P-end labeling of oligoDNA that is complementary to positions 20-45 of hY3 RNA. Hybridizations were done in 6* SSC and blots were washed with 2* SSC, 0.1% SDS for 10-15 min at the hybridization temperature (hY5
and gY5 at 60oC, Y3 at 57oC) (
12
). Southern blotting of DNA purified from somatic cell hybrids and other cell
lines was previously described (
12
). The probe was the [[alpha]-
32
P]dCTP-containing 800 bp hY5 fragment shown in Figure
1
A. The final wash was with 0.5* SSPE at 60oC. Cell lines and DNAs thereof used here were previously described
by Chang
et al
. (
33
).
We used primers based on the hY5 RNA sequence (
8
) to screen a human chromosome 7-enriched YAC library by PCR (
12
,
30
,
31
). Three positive clones (yWSS756, yWSS1476 and yWSS2977) were identified. Two
different `hemispecific' PCR-based methods developed in this lab were then used to identify sequences
flanking the hY5 RNA gene (
12
). Primers complementary to both flanking regions amplified a single 800 bp
fragment containing the hY5 RNA gene and its flanking sequences (Fig.
1
A). The coding region is identical to the sequence of hY5 RNA (
8
) and ends in four dT residues, a termination signal for pol III. In addition,
the 5' flanking region contains motifs with homology to a TATA box at positions
-36 to -26, a proximal sequence element (PSE) at -64 to -47, and two potential octamer enhancer motifs at
positions -242 to -234 and -216 to -209. The sequence and arrangement of these elements
are homologous to ones in the other hY RNA genes as well as the upstream
promoters of a variety of human pol III snRNA genes (Fig.
1
B) (
8
,
10
,
12
). The hY5 PSE (and the gY5 PSE, see below) contains two non-consensus residues (underlined in Fig.
1
B) where invariant A residues reside in the PSEs of the other genes (
10
). This suggests that the hY5 PSE might be a suboptimal promoter. The PSE and
TATA of hY5 bear no more overall nucleotide identity with other hY PSEs or
TATAs than they do with the other human class 3 promoters (not shown).
Comparison with the flanking sequences of the other hY RNA genes reveals that
homology is limited to the TATA and PSE motifs (
8
,
10
,
12
) (not shown).
Because many Y pseudogenes exist in human DNA it was important to demonstrate
that the isolated hY5 sequence was functional for RNA synthesis. The hY5 gene
was cotransfected into mouse NIH 3T3 cells with a pol III-dependent VA1 RNA gene as a control. The VA1 RNA gene does not contain a
PSE (or TATA), and therefore does not compete for the factors that recognize
this element (
34
,
35
). Transfections were done in duplicate and RNA was purified 48 h later and
examined by Northern blot (Fig.
2
). HY5 RNA was expressed after transfection of the intact gene (Fig.
2
, lanes 1 and 2) but not after transfection of a 5' deletion mutant (lanes 3 and 4). In this 5' deletion mutant the entire region upstream of position -21 was replaced with vector DNA. Reprobing the blot for VA1
RNA revealed uniform transfection efficiency (lanes 1-4). We conclude that the hY5 RNA gene we isolated represents the
authentic gene, and that the upstream promoter is required for hY5 RNA
expression
in vivo
.
Each of the hY5-containing YACs we identified exhibited the same restriction fragment
length pattern by Southern blot analysis using multiple restriction
endonucleases suggesting that they represented a single hY5 locus on chromosome
7 (not shown). Since chromosome 7 was found to be the only human chromosome
that expressed hY5 RNA (
12
), this further suggested that these clones represented the authentic hY5 RNA
gene and that it is single copy in the human genome. Southern blot analysis of
total human DNA as well as somatic cell hybrids including the cell line GM10791
which contains chromosome 7 as its only human chromosome (
12
), confirmed this (Fig.
3
A and data not shown). Although it is possible that another hY5 RNA gene might
exist in human DNA, the cumulative evidence argues against this.
In an attempt to identify additional hY-containing YACs we screened the library for other hY RNA genes, as well as
additional chromosome 7-enriched clones. Our initial screening for Y4- homologous sequences detected several YACs that contained hY4
pseudogenes (
12
). Therefore, the entire collection of YAC clones was rescreened with a PCR
assay that was specific for the 5' flanking region of the authentic hY4 RNA gene (not shown). Four positive
clones were obtained, yWSS1020 as expected (
12
), yWSS4352, yWSS3230 and yWSS2977, the latter of which was independently
positive for hY5 (above). Clones yWSS756 and yWSS2977 were previously found to
contain the genetic marker D7S688 (
32
). Clones yWSS756 and yWSS1476 were previously found to overlap and were mapped
to chromosome 7q36; this independently co-localizes the hY5 gene to this region (
32
; E.D.G., unpublished data). Matera and colleagues mapped a cosmid clone that
contained the hY1/hY3 locus to 7q36 (
36
). Fluorescence
in situ
hybridization localized yWSS1020 to human chromosome 7q36 with no evidence of
chimerism extending these earlier results (R.J.M. and A.L.S., data not shown).
Since yWSS1020 contains hY4 as well as hY1/hY3, this YAC together with yWSS2977
and/or yWSS4352 establish a contiguous region estimated to be 300-600 kb that contains all four hY RNA genes. These data represent the
first report that the hY5 RNA gene maps to 7q36.
We confirmed the presence of hY sequences in these YACs by Southern blot
analysis including yWSS1020 as a positive control (
12
) (Fig.
3
B). Yeast DNAs were digested with
Taq
I, transferred to nylon membranes and probed for hY5 and hY4 (Fig.
3
B, upper and lower panels respectively). Of these DNAs, only yWSS1020 reacted
with a hY1/hY3 probe by Southern analysis (not shown). Since yWSS1020 contains
hY4 and hY1/hY3, but does not contain hY5 whereas yWSS2977 and yWSS4352 contain
hY4 and hY5, it can be concluded that hY5 resides to one side of hY4 while
hY1/hY3 lie to the other side of hY4. YACs yWSS2977 and yWSS4352 were each
positive for both hY4 and hY5 genes. No hY-homologous sequence was detected in yWSS3230 by Southern blotting and this YAC was presumed to have been falsely negative by the
PCR screening assay (not shown). A tentative structure of the hY RNA gene
family based on these results is summarized in Figure
3
C. It is noteworthy that although YACs yWSS1476 and yWSS756 are distinguishable
by their sizes and genetic marker content, e.g. yWSS756 contains marker DS7688
while yWSS1476 does not (
32
), they each contained a single Y sequence, hY5 (
32
; E.D.G. and R.J.M., unpublished data). These data suggest that the distance
between the hY5 gene and the other hY RNA genes is large. The hY4-to-hY1/hY3 distance may also be large since we detected two distinct
YACs (yWSS4352 and yWSS2977) that contain hY4 but not hY1/hY3. We conclude that
unlike the hY1 and hY3 genes which are adjacent, the hY4 and hY5 genes appear
not to be tightly linked with each other or with the hY1/hY3 locus.
Because the suggestion that hY5 RNA may have arisen as a retroposon implies non-functionality, we wanted to examine the conservation of the Y5 RNA gene in
primates (
16
). The prosimian galago (bush baby) branched off from the primate lineage,
before the emergence of the lineage that led to monkeys, apes and humans 65-80 million years ago. Moreover, galago DNA has mutated at a rate that is
comparable with rodents, nearly five times the substitution rate in the
monkey/ape/human lineage, a characteristic that makes galago attractive for
comparisons with higher primates (
37
,
38
), especially in the case of Y RNAs since rodents do not have an active Y5 RNA
gene and therefore can not be used for this purpose (
13
). We were able under low stringency conditions to amplify the Y5 RNA gene
sequence from DNA from
Galago senegalis
but only with one of four primer pairs that flank the hY5 RNA gene.
Recombinants containing the amplified fragment were sequenced and a consensus was obtained confirming that the locus orthologous to the human Y5 RNA gene was isolated. Sequences
corresponding to the PSE, T/A box, and coding region of galago and human Y5 RNA
genes exhibited 94, 91 and 88% identity respectively (Fig.
1
B and below). By contrast, the sequences that reside between the PSE and T/A
box, and the T/A box and the start site of transcription, exhibited only 62 and
59% identity, respectively, while the 72 bp of obtainable sequence downstream
of the Y5 coding region revealed only 45% homology, consistent with the high
rate of human/galago divergence (not shown) (
37
).
In order to compare the relatedness of the galago (g)Y5 RNA gene with the gY5
RNA that is actually expressed in galago cells, we examined Y5 RNA expression
using human and galago Y5-derived probes. Figure
4
A is a Northern blot that revealed that while Y5 RNA is absent in rodents, as
expected (
13
), a slightly shorter and less intense RNA signal is detectable in galago (lane Ga) as compared with modern primates including human (lane Hu), when probed with
DNA corresponding to hY5 (upper panel). The faster mobility of gY5 RNA relative
to hY5 was expected since the transcribed region of the gY5 gene predicted that
gY5 RNA would be 4 nt shorter than hY5 (below). After stripping of the blot, a
probe corresponding to gY5 detected substantially more gY5 than hY5 RNA (middle
panel). Comparable amounts of Y3 RNA were detected in all species examined
(lower panel). The striking difference in relative intensities of human and
galago Y5 RNAs using human versus galago Y5-derived probes further confirms that the sequence isolated from galago
indeed represents the gY5 RNA gene and that the nucleotide divergence between
the two species' Y5 RNAs is significant.
HY5 RNA secondary structure has been determined (
39
). Prediction of gY5 RNA secondary structure yielded a structure that was
overall very similar to hY5 RNA (not shown) (
40
). For Figure
4
B, nucleotide differences between gY5 and hY5 RNAs were superimposed onto the
hY5 RNA secondary structure (
39
). The gY5 coding sequence is 4 nt shorter than hY5 RNA due to the absence of
two dinucleotides (encircled, with arrow). Remarkably, most of the
substitutions were found to be clustered on opposing single-stranded regions of the Y5 RNAs between two highly conserved stems (boxed
in Fig.
4
B) (
5
,
6
,
8
,
9
,
18
,
39
). This comparison of gY5 and hY5 RNAs revealed that Y5 RNA evolution has been
restrained by a high degree of conservation of structural motifs important for
recognition by Ro 60 kDa (
7
,
23
).
Because it had been proposed that hY5 RNA might be the product of a mutated Y
RNA retroposon that became transcriptionally active we were obliged to examine
the hY5 gene locus diligently for fossil remnants of retroposition (
16
). A common characteristic of retroposons is that they are flanked by short direct repeats (DRs) of the
insertion site. The sequence flanking the 3'-end of the hY5 gene is also found in the hY5 upstream flank.
However, the sequence in the 5' flank is split into two parts by what appears to be an internal
expansion. Figure
5
A shows the sequence of the uninterrupted 3' DR (line `3' DR'), below which this sequence is separated into two parts (line
3') aligned with the upstream flank of the hY5 gene (line 5'). The first part of the 3' DR bears strong homology to the upstream part of the
putative TATA-box at positions -35 to -27 while the distal part of the 3' DR bears strong homology to the region immediately 5' to the hY5 sequence. The 5' copy of the putative DR appears as if
it underwent expansion of its A-rich central region and presently contains ATAA, GAGA and CACA sequences.
It is noteworthy here that in a recent phylogenetic study, Arcot
et al
. showed that the sequence ATAA [see
Alu
A62 in Arcot
et al.
(
41
)] was present at the preintegration site of an
Alu
retroposon and was asymmetrically expanded in the 3' DR upon
Alu
insertion. This is similar to the ATAA sequence present in the hY5 3' DR that presumably expanded to three ATAA repeats in the 5' DR (Fig.
5
A). Thus, the hY5 sequence is flanked by imperfect direct repeats whose
composition is consistent with a hY5 retroinsertion-mediated event that occurred a long time ago. Additional support for the
retroposition model of hY5 origin is provided by the fact that the last 3 nt of
the hY5 DR, AGT, is found as the first 3 nt of the hY5 coding sequence (Fig.
5
A) (
42
,
43
).
As mentioned above, the presence of a sequence motif that extends the 3'-end of hY5 RNA is a feature that distinguishes hY5 from the other Y
RNAs (
5
-
8
,
18
,
39
). A 3' extension is limited to hY5 and although it may be found on the Y5-homologous RNAs of other mammals, a similar motif is not found on
Xenopus
Y5 RNA (
5
). The lineage of the
Xenopus
and human Y5 genes may be different even though their RNAs may appear related
due to constraints of Ro 60 kDa and internal pyrimidine richness superimposed
on their small size (
5
,
9
). With this in mind, it is not unreasonable to suspect that the unique and
atypical 3'-end of hY5 RNA may provide a clue to its lineage especially in
those genomes in which the mutation rate has been relatively low, i.e. human (
37
).
Our alignment of hY RNA sequences mostly agrees with previous ones except for
details at the 3'-ends. Therefore, for the purpose of this analysis we limit
comparisons to the 3'-end regions of hY RNAs as depicted in Figure
5
B. The hY5-specific, atypical terminal sequence GCUGUUUU (underlined in Fig.
5
B) appears as if it might have been added to the 3'-end of a pre-existing Y RNA. Upon close examination, this hY5 3' RNA sequence motif bears significant similarity to a
short DNA sequence which lies immediately 3' to the hY3 RNA gene that currently resides in the human genome (depicted
in Fig.
5
B as
hY3 3
'
DNA
) (
8
). This provides limited evidence to suggest that the hY5 RNA gene might have
been derived from an ancestral Y3 allele. This homology ends in a run of four
dT residues just downstream of the hY3 gene and the hY3 and hY5 DNAs are
unrelated beyond this (not shown) (
8
). The fact that the homology ends in four dT residues which corresponds to a
hY3 downstream terminator for pol III further suggests that the hY5 gene might
have been derived from a 3'-extended, i.e. pol III readthrough, hY3 transcript.
The source gene for the smallest of the human Y RNA genes, hY5 was isolated,
completing the cloning of all four of the hY RNA genes. Previous results
demonstrated that chromosome 7 is the only human chromosome that expresses hY5
RNA (
12
). In the present study, we screened chromosome 7-enriched DNA for hY5 sequences and detected three isolates of the same
gene. Cumulatively, the results indicate that the hY5 RNA gene is single copy
in human DNA. The physical presence of this hY5 RNA gene on chromosome 7 was
conclusively demonstrated (Fig.
3
A). Others have localized the hY1 and hY3 RNA genes to the q36 region of
chromosome 7 (
36
). In the present report we have independently mapped each of the hY4 and hY5
RNA genes to chromosome 7q36. After isolating the hY5 RNA gene, we examined the
hY5-containing YACs for other hY RNA genes and then derived a physical map of
the four hY RNA genes (Fig.
3
C).
The hY5 RNA gene contains a type 3 promoter for pol III and a consensus pol III
terminator, features of functional hY RNA genes that distinguish them from
pseudogenes (
16
). HY5 RNA gene expression was demonstrated
in vivo
to be dependent on the upstream promoter. One question raised by this work is
the role of differential promoter strength versus RNA stability in hY RNA
expression. Partial characterization of the hY5 RNA gene promoter, performed
here to establish the authenticity of the hY5 RNA gene, suggests that it is
less active than the hY4 RNA gene promoter. Availability of functional hY RNA
gene transcription units such as described here and elsewhere will allow
examination of this issue in the future (
12
).
Primary evidence of a retroposon origin for hY5 RNA was derived from flanking
DNA sequences whose composition and characteristics suggest a retroposition
event: (i) enrichment of adenosine residues in the DNA strand corresponding to
the RNA sequence (
44
,
45
), (ii) di and trinucleotide microsatellite repeats associated with the DRs (
41
,
43
), (iii) a microsatellite-like sub-sequence asymmetrically represented in the 5' DR (
41
,
43
,
45
-
47
) and (iv) identity of 3 nt at the 3'-end of the DR and the 5'-end of hY5 (
43
,
48
).
The above features are entirely consistent with, but in no way prove, that the
hY5 RNA gene was derived by retroposition, especially in light of findings that
may appear to be unexpected of retroposons. Specifically, juxtaposition of hY
RNA genes suggests that DNA-mediated duplication played a role in hY RNA gene evolution. However, at
7q36, hY5 resides adjacent to the telomeric region of chromosome 7 and
retroposon elements are known in some cases to target telomeres (
49
). In addition, the large, i.e. >100 kb, distance between the solitary hY5 RNA
gene and the other hY RNA genes as well as lack of homology of their flanking
sequences argue against a simple duplication event as the source of the hY5
gene. Another inconsistency might appear to be the lack of an identifiable self-primer for reverse transcriptase, necessary for conversion of Y RNA into
cDNA. However, recent results indicate that nicked genomic DNA can serve
directly as the primer for some classes of retrotransposons (
50
).
The lack of Y5 RNA in rodents and certain other mammals [see refs (
4
,
9
)] suggests that an active Y5 RNA gene arose in a lineage following an early
branch point in the mammalian radiation, 65-100 million years ago. An hY5-homologous sequence is detectable in rabbit and mouse cell DNA even
though hY5-homologous RNA is not detectable in these cells with the same probe (
9
). This suggests that a hY5-homologous sequence resides in these species but is transcriptionally
inactive. Also, the number of hY5-homologous pseudogenes increased dramatically during primate evolution (
9
). These data are reconciled by a simple model in which, after retroposition,
the hY5 sequence was dormant until it acquired transcriptional competence, at
which time it could produce Y5 RNA and simultaneously establish itself as a
source gene for hY5 retroposons (
9
,
16
). If expansion of the 5' DR was indeed the source of the sequence downstream of the 5' TATA box, this would have generated the appropriate spacing
between the TATA and the start site of hY5 required for accurate transcription
(
10
). The proposal that hY5 originated from Y3 is consistent with the idea that Y3
is the oldest of the Y RNAs and therefore that it would have been a likely
progenitor of the Y RNA genes that appeared later in evolution (
6
).
The previously established association between Ro 60 kDa and 3'-extended 5S transcripts compelled us to consider the homology
between the 3' extension of hY5 RNA and the downstream region of hY3 DNA to be
relevant. According to the retroposition model of the origin of the hY5 RNA
gene proposed here, pol III would have transcribed through the normal 3' terminus of an ancestral Y3 allele to the next run of four dT residues
which occurs ~10 bp downstream of the current hY3 gene, generating an `errant' transcript
similar to the Ro-associated, 3'-extended 5S RNAs. This model, based on sequence homology is
supported by the unusual 3'-extension on hY5 RNA that has not otherwise been accounted for. In
any case, this extension may be responsble for, either directly or as a result
of association with La, for some of the properties mentioned in the
Introduction that distinguish hY5 from the other hY RNAs.
Regardless of the exact mechanism by which the hY5 RNA gene arose, it is clear
that this gene, which is the current source of Y5 RNA in humans, indeed became
fixed in our ancestral genome. The observed sequence conservation in the
upstream promoter and coding regions of the galago and human Y5 RNA genes
argues that the capacity for expression of this gene is beneficial to its
hosts. Nucleotide substitutions in the coding regions of these Y5 genes were
clustered in a region of human and galago Y5 RNAs adjacent to conserved motifs
that are important for Ro 60 kDa binding. Presumably, mutations occurred
throughout the coding sequence of the Y5 gene during human and galago
divergence yet it appears that only those alleles that preserved the ability of
the RNA to fold properly co-evolved with these species. This argues that the hY5 RNA gene alleles that
retained their transcripts' ability to form the conserved stem structures (and
presumably associate with Ro 60 kDa) were selected for during primate
evolution, while alleles carrying mutations that disrupted this folding were
not. Although this phylogenetic comparison suggests that the ability of Y5 RNA
to bind Ro 60 kDa is beneficial to the species, it is possible that selection
was driven simply by the ability of Y5 RNA to occupy Ro 60 kDa. Whether Y5 RNA
serves a distinct function other than this remains unknown. Finally, we wish to
note that by whatever mechanism that led to activation of the hY5 RNA gene in
primates and certain other mammals, its product may represent an evolving
autogen (
19
,
51
).
We thank J. Keene for sharing unpublished Y3 gene sequence, N. Sasaki-Tozawa for technical assistance early in this project and to G. Boire, J.
Craft, A. Furano, S. Wolin and our colleagues in the LMGR for comments and
discussions. We are grateful to B. Howard and the NICHD for support for E.B.
REFERENCES
Return