ABSTRACT
The retrotransposon Ty4 is found in different yeast strains at only one to three
copies per haploid genome. In the present study, we aimed at relating the
apparent low transpositional activity of Ty4 to transcriptional features of
this element. RT-PCR revealed that Ty4 is transcribed at a very low level,
being comparable with that of
GAL4
. Contrary to other Ty elements, the transcriptional rate of Ty4 is not affected
in a
sin4
background nor by treatment of cells with
[alpha]
factor. From experiments measuring the expression levels in
lac
Z fusion constructs, we conclude that Ty4 transcription is repressed by a
negative regulating element residing within the LTR, whereas positive
cis
-acting elements, like those that have been found to mediate expression of
Ty1/2 and Ty3, are absent from Ty4. Analysing Ty4 transcript termini by the
RACE-PCR method, we found several distinct transcriptional initiation sites.
But surprisingly, the majority of the polyadenylated Ty4 transcripts terminate
shortly upstream from the 3'
LTR boundary, so that these transcripts do not contain a U3-R sequence, which is normally required for obligate strand transfer during
DNA synthesis. Thus, the extremely low transcription rate of Ty4 and imperfect
Ty4 transcripts are the reason for the low transpositional activity of this
element.
Yeast transposable elements are a family of genetically mobile, repetitive DNA
sequences. Two different classes have been characterized: the copia-like family, including the Ty1, Ty2, Ty4 and Ty5 elements in
Saccharomyces
cerevisiae
, and the gypsy-like family, including Ty3 in
S.cerevisiae
(
1
,
2
) as well as Tf1 and Tf2 in
Schizosaccharomyces
pombe
(
3
). Ty elements share structural similarities with and resemble the proviral form
of retroviruses (
4
) in that they transpose through an RNA intermediate (
5
). Whereas Ty1 and Ty2 occur in some 30 copies throughout the yeast genome, Ty3
and Ty4 are low copy number elements (
6
,
7
). The different abundances of the elements appears to be due to the fact that
their transcriptional and transpositional activities are differently regulated
(
8
). In accordance with the high transcriptional activity of Ty1, Ty1 RNA is found
as one of the most abundant poly(A)
+
RNAs in many haploid strains of
S.cerevisiae
(
9
)
.
By contrast, little Ty3 RNA is found in haploid cells under normal conditions.
Transcription of Ty3, however, can be induced in cells by exposure to the
mating pheromone [alpha] factor (
6
). Responsible for this effect is the pheromone-responding element (PRE) located within the Ty3 LTR,
sigma
(
10
).
Transcription of Ty1 and Ty3 has been shown to initiate within the 5' LTR and to terminate in the 3' LTR, resulting in polyadenylated transcripts which are shorter
than the proviral DNA. In both cases, the transcripts contain an R-U5 sequence at the 5'-end and an U3-R sequence at the 3'-end (
6
,
11
). During reverse transcription, full-length DNA with complete LTRs is regenerated (
12
).
Ty4 is 6.2 kb in length and is composed of 371 bp terminal repeats (
tau
), flanking an internal domain of 5.4 kb. Sequence comparisons at the amino acid
level show that the domains in Ty4 diverge considerably from those of the other
yeast retrotransposons. Although a number of
tau
elements are found dispersed throughout the yeast genome, indicating that Ty4
is able to transpose, attempts to obtain direct proof for transposition of this
element have failed so far (
13
,
14
). For example, the analysis of transformants in which an intact Ty4 copy was
tagged by the insertion of the
neo
gene cassette and overexpressed under the control of the inducible
GAL1
promoter, analogous to the constructs used to monitor Ty1 transposition (
15
), did not reveal any novel integration sites of Ty4 (
13
). These experiments suggested to us that the transcription rate of Ty4 even under
these conditions must be very low and, as transposition is intimately
correlated to transcription, that the low rate of Ty4 transcription
per se
might be the cause for the low transposition frequency of Ty4.
In this study, we carried out experiments to determine the level of Ty4
transcription and to characterize the nature of Ty4 transcripts. By using
RT-PCR technology, we show that Ty4 is expressed at a very low level and
comparable with that of
GAL4
. Contrary to the expression of Ty1 or
GAL4,
Ty4 expression is not repressed in a
sin4
-
background. Measurement of [beta]-galactosidase activity from a variety of Ty4-
lacZ
fusion constructs revealed that the expression of Ty4 at a low level might be
due to the presence of negative regulatory elements within the leftmost LTR and
the absence of activating
cis
-regulatory elements within the translated region. Furthermore, an analysis
of Ty4 transcript initiation and termination sites using RACE-PCR methodology
indicates that Ty4 transcripts initiate at different sites within the left LTR,
but that the majority of the transcripts terminate shortly upstream from the
rightward LTR. Thus these Ty4 transcripts lack the U3-R sequence generally required for strand transfer during DNA synthesis in
retroelements.
Plasmids were cultured in
Escherichia coli
strain 490A (
r
k
,
m
k
,
met
,
thr
,
leu
,
recA
), except pCR-Script SK(+) plasmids (Stratagene, Heidelberg, Germany), which were grown
in
E.coli
strain XL1-Blue MRF' Kan (Stratagene). The yeast shuttle plasmids YEp353-358 were cultured in yeast strain YM4127 (MAT
a
ura3-52
his3-200
ade2-101
lys2-801
trp1-903
leu2-3
leu2-112
tyr1-501
) (M. Johnston).
Yeast strain MC45-5/A (MAT[alpha],
trp1-259
,
leu3-3
,
117
,
ura3-52
,
suc2
) (M. Ciriacy) and YM4127 were the source of poly(A)
+
RNA. The wild-type strain YS18 (MAT[alpha]
his3-11
his3-15
leu2-3
leu2-112
can
R
ura3
[Delta]5) (
16
) and the
sin4
mutant strain YS84 (MAT[alpha]
his3-11
his3-15
leu2-3
leu2-112
can
R
ura3
[Delta]5
SIN4
::
URA3
) (D. J. Stillman) was used in the analysis of Ty4 transcript levels. Yeast
cells were cultured as described by Sherman
et al
. (
17
).
All constructs used in the [beta]-galactosidase assays are based on the multi-copy
E.coli
-yeast shuttle vectors YEp353-358 carrying a truncated
lacZ
gene and
URA3
as a marker (
18
). To construct pTy1-
Pvu
II, the entire Ty1-
gag
region was amplified using primers annealing to the LTR start site (Ty1-1-
Hin
dIII) and to the region of the frameshift at position 1600 (Ty1-1600-
Bgl
II), thereby introducing artificial
Hin
dIII and
Bgl
II restriction sites (cf. Table
1
). The Ty1 fragment was subcloned into Yep356, cut out with
Hin
dIII and
Pvu
II (position 475) and religated into the
Hin
dIII/
Sma
I sites of YEp357. pTy4-
Eco
RV was constructed by inserting a 1200 bp
Hin
dIII-
Eco
RV fragment containing the Ty4 element (positions 1-907, flanked by genomic sequence) into YEp358.
Table 1
pTy4-
Bgl
II was constructed by amplifying the entire Ty4-
gag
region using primers annealing to the LTR start site (Ty1-1-
Hin
dIII) and to the region of the frameshift at position 1428 (Ty4-1428-
Bgl
II), thereby introducing
Hin
dIII and
Bgl
II restriction sites. The PCR product was cleaved with
Hin
dIII and
Bgl
II and ligated into YEp356.
Ty1/Ty4 chimeric plasmids were designed using the PCR technique. To construct
pTy1-Ty4-
lacZ
, the
delta
region and the Ty4 internal region were amplified separately, using the primer
pairs Ty1-1-
Hin
dIII/Ty1-LTR-Ty4 and Ty4-ATG-Ty1/Ty4-925-
Eco
RI respectively (the backward primer Ty1-LTR-Ty4 contains extra nucleotides overlapping Ty4; the forward primer
Ty4-ATG-Ty1 contains extra nucleotides overlapping Ty1). The resulting
products were subjected to PCR for five cycles without primers, followed by 25
cycles with the addition of the `outer' primers Ty1-1-
Hin
dIII and Ty4-925-
Eco
RI. The final PCR product was cleaved with
Hin
dIII and
Eco
RI and ligated into YEp358. pTy4-Ty1-
lacZ
was constructed in a similar manner except that Ty4-LTR-Ty1 and Ty1-ATG-Ty4 were used as `overlapping' and Ty4-1-
Hin
dIII and Ty1-850-
Eco
RI as outer primers. The resulting PCR product was subcloned in Yep358.
pTy1-Ty4-
Eco
RV was constructed by replacing the 5'-part of
tau
up to the
Hpa
II site with the corresponding
delta
region of Ty1. Ty4-230-
Xho
I and Ty4-910-
Eco
RV were used in a PCR reaction. The resulting product was subcloned in the
Srf
I-digested pCR-Script(+) vector, which allows selection for recombinants by
including
Srf
I in the ligation reaction. A clone containing the insert's
Xho
I site next to the vector's
Hin
dIII site was selected, digested by
Hin
dIII/
Sac
I and the resulting fragment recloned in pUC18. Finally, the 260 bp Ty1
Hin
dIII-
Xho
I fragment was inserted into this intermediate and the chimeric Ty1/Ty4
Hin
dIII-
Eco
RV fragment transferred into YEp358. The absence of PCR artefacts was verified
by sequencing all constructs between
lacZ
and the respective LTR.
Standard PCR was carried out by the use of a GenAmp PCR System 2400 thermocycler
(Perkin Elmer, Weierstadt, Germany) in a total volume of 50 [mu]l [10 mM Tris-HCl, 10 mM MgCl
2
, 50 mM KCl, pH 8.3, 200 [mu]M each dNTP, 1 [mu]M each primer and 2.5 U
Taq
DNA polymerase (Boehringer, Mannheim, Germany)] using thin-walled 200 [mu]l tubes without mineral oil overlay. Cycle conditions were: hold (94oC/5 min), three-step cycle (96oC/1 min, 55oC/40 s, 72oC/1.5 min, 25-35 cycles), hold (72oC/5 min, then 4oC).
PCR products were isolated from agarose gels by the glassmilk method or directly
after PCR using QIAquick spin columns (Qiagen, Hilden, Germany). Purified PCR
fragments were cloned into the pCR-Script SK(+) vector (Stratagene) according to the manufacturer's protocol.
Poly(A)
+
RNA samples were separated in 0.8% agarose gels (20 mM MOPS, 5 mM sodium
acetate, 1 mM EDTA, pH 5.5-7.9, and 2% formaldehyde) and transferred directly to PALL B filters
(Pall, Dreieich, Germany).
Standard procedures were used for labelling DNA probes and in Southern or
Northern hybridizations (
19
).
Total RNA was isolated from yeast cells as described (
20
). After treatment with DNase I (20 U/ml DNase I, 25 mM Tris-HCl, pH 7.5, 5 mM MgCl
2
, 160 U/ml RNasin; 37oC/60 min), poly(A)
+
RNA was enriched using the Oligotex-dT poly(A)
+
RNA Mini Kit (Qiagen). mRNA (100 ng) was retrotranscribed in a total volume of
50 [mu]l using oligo(dT) primers and RNase H
-
reverse transcriptase as recommended by the manufacturer (StrataScript;
Stratagene). Aliquots of 2-8 [mu]l of the reaction mixture were used for RT-PCR. The reaction was
carried out in a total volume of 50 [mu]l (10 mM Tris-HCl, pH 8.3, 10 mM MgCl
2
, 50 mM KCl, 200 [mu]M each dNTP, 1 [mu]M each primer and 2.5 U
Taq
polymerase (Boehringer). Cycle conditions were: hold (94oC/5 min), three-step cycle (94oC/60 s, 55oC/40 s, 72oC/90 s, 25 cycles), hold (72oC/5 min, then 4oC).
The 5' RACE protocol of Frohmann
et al
. (
21
) was used with minor modifications. poly(A)
+
RNA (YS4127) (2.4 [mu]g) was retrotranscribed as above except using a Ty4-specific primer (Ty4-480). The RT product was passed through a QIAquick spin column
and polyadenylated by the use of terminal deoxynucleotidyl transferase (TdT)
(Boehringer). Excess dATP was removed by the use of QIAquick spin columns. For
amplification, 1-5 [mu]l from the tailing reaction, 0.05 [mu]M RACE-N (`hook') primer, 1 [mu]M Ty4-480 primer and 1 [mu]M RACE-O primer in a 50 [mu]l PCR reaction mixture [1*
Taq
polymerase buffer (Boehringer), 200 [mu]M each of the four dNTPs] were denatured (95oC/5 min) and cooled to 72oC.
Taq
polymerase (2.5 U) was added and the mixture incubated for 5 min at 95oC, 5 min at 42oC and 30 min at 62oC in a thermocycler (GeneAmp PCR System 2400; Perkin Elmer). A
second round of amplification was carried out using a step program (96oC/1 min, 55oC/40 s, 72oC/90 min, 25 cycles) followed by a 7 min final extension at 72oC, without adding fresh
Taq
polymerase. Then a nested PCR reaction was carried out using RACE-I and Ty4-428 primers. RACE-PCR products >100 bp were isolated from a 2%
agarose gel and cloned. The bacterial colonies were screened by hybridization
with a Ty4-
gag
probe.
3'-End amplification of Ty4 cDNA (3' RACE) was performed as described above except omitting the
tailing reaction. RACE-N primer was used for reverse transcription; RACE-O, RACE-I and Ty4-specific primers were used in the subsequent
amplification reaction.
Yeast cells were grown to an optical density of 1.0-1.5 at 600 nm in synthetic medium lacking uracil. Cells pelleted from 2
ml cultures were resuspended in 1 ml Z buffer (60 mM Na
2
HPO
4
, 10 mM NaH
2
PO
4
, 10 mM KCl, 1 mM MgSO
4
, 50 mM [beta]-mercaptoethanol, pH 7.0). Toluene-permeabilized cells were assayed for
lacZ
activity as described (
22
), using the OD
600
value of the culture and the assay time for normalization. Where noted, cells
were treated with 4 [mu]M [alpha] factor (Sigma) for 30 min before measurement of [beta]-galactosidase activity.
In order to investigate Ty4 transcription, poly(A)
+
RNA was isolated from MC45-5/A (MAT[alpha] type) or YM4127 (MAT
a
type) cells, fractionated by electrophoresis in formaldehyde-agarose gels and transferred to nylon membranes. The filter-bound RNA was hybridized to a Ty4-
gag
-specific probe and a Ty1-
gag
-specific probe for comparison. An actin gene-specific probe was used for internal standardization. Even at
prolonged exposure of the autoradiograms, no Ty4-specific signals could be detected (Fig.
1
A). In contrast, major Ty1 transcripts (5.7, 5.0 and 2.2 kb) (
23
) were visualized by autoradiography after 24 h (Fig.
1
B). Therefore, we employed the highly sensitive RT-PCR method (
24
) to detect Ty4 transcripts. In brief, total yeast RNA was exhaustively digested
with RNase-free DNase and enriched for poly(A)
+
RNA. After reverse transcription using oligo(dT) primers, a subsequent PCR
reaction was carried out using Ty4-specific primers.
Studies with yeast mutants have shown that the expression of several genes,
including Ty1 elements, requires pleiotropic factors of the Swi/Snf class for
transcriptional activation. Recent evidence suggests that the
SWI
/
SNF
genes, together with the
SPT
/
SIN
genes, mediate transcriptional control via effects on chromatin (
26
,
27
). Jiang and Stillman (
28
) found that
sin4
mutants display defects both in negative and positive regulation of
transcription: the expression of the Ty1
delta
element is repressed in these mutants.
To investigate whether Ty4 expression is similarly affected in
sin4
mutants, we compared the Ty4 transcript level in a wild-type (YS18) to that in a
sin4
mutant (YS84) yeast strain using competitive RT-PCR (Fig.
3
). As above, actin gene primers and
GAL4
-specific primers were used for comparison. Since our PCR conditions were
chosen such that no plateau effect occurs, it was also feasible to compare the
amount of products in both strains. An equal amount of the competitor
Mlucomp
was added to each of the respective Ty4 RT samples as an internal standard
(Fig.
3
, lanes 4 and 5 and 21 and 22 respectively). The PCR products of co-amplification were digested with either
Hpa
II (lanes 4 and 21) or
Mlu
I (lanes 5 and 22) before loading the samples onto the gel. Direct PCR without
RT yielded no product in both strains. PCR after RT resulted in equal amounts
of product of the expected size in both strains (Fig.
3
, compare lanes 3-5 with lanes 20-22), confirming that the RT-PCR products were derived from RNA.
However, no differences between the two strains were observed. Thus, our
conclusion is that Ty4 is not repressed in
sin4
mutants. In contrast, the
sin4
mutation leads to an increase in
GAL4
transcription (compare lane 32 with lane 15); the larger bands seen in lanes 25
and 30 in strain YS84 most probably represent products derived from non-specific priming.
Expression of Ty1/2 and Ty3 is regulated by a variety of cellular factors and Ty-specific
cis
-acting elements (
1
). The principal regulatory elements of Ty1 are located within the internal
translated region rather than upstream of the transcription start site in the
leftmost LTR (
29
-
31
). Regulation of Ty3 transcription differs in several respects from Ty1 and Ty2
regulation and appears to be largely independent of activation by the internal
domain (
8
).
To address the question whether Ty4 transcription is regulated via potential
cis
-acting elements, we used the following approach. Two Ty4-
lacZ
fusions containing different 5'-portions of the element were constructed, transformed into YM4127
(MAT
a
) and [beta]-galactosidase activity was determined in cellular extracts. Ty1-
lacZ
fusions were employed for comparison. As seen in Figure
4
, 216 U [beta]-galactosidase activity was obtained from a construct (pTy1-
Pvu
II, no. 1) containing the 5'-portion of Ty1 including the internal domain up to the `critical'
Pvu
II site (cf.
31
), whereas in accordance with earlier observations (
31
) [beta]-galactosidase activity from a construct in which
lacZ
was fused to the
delta
sequence alone was near background. Similarly, pTy4-
tau
by itself fused to
lacZ
did not reveal any [beta]-galactosidase activity (not shown). However, [beta]-galactosidase activities from pTy4-
lacZ
fusion constructs (pTy4-
Eco
RV and pTy4-
Bgl
II, nos 2a and 2b) equivalent to construct pTy1-
Pvu
II were also below the detection level. In view of the low transcriptional
activity of Ty4 that could only be monitored by RT-PCR it is not so surprising
that the expression of intact Ty4 is too low to be detected by measuring [beta]-galactosidase activity from pTy4-
lacZ
fusion constructs. The low expression level of Ty4 may be explained by either
the absence of activating elements or the presence of negative regulatory
elements within the element. To exploit these possibilities, chimeric
lacZ
fusion constructs were made by combining relevant portions of Ty1 and Ty4. pTy4-Ty1-
lacZ
(no. 3), in which the region downstream from the Ty4 translational start site
was replaced by a Ty1 portion containing the Ty1 translational start site and
sequences up to position 850, exhibited weak [beta]-galactosidase activity (9 U) compared with pTy1-
Pvu
II. This result shows that the internal activating elements located within the
Ty1-
gag
region are still capable of stimulating expression in the presence of the
upstream Ty4 promoter region, albeit to a much lower degree than in the cognate
situation. Normally, arbitrary sequences replacing Ty1 sequences up to the
internal activation region do not interfere with Ty1 transcription (
30
). Thus a possible explanation for our finding is that a negative regulatory
element is located within the 5'-portion of the Ty4 LTR. This interpretation is further
substantiated by measuring Ty4 expression from construct pTy1-Ty4-
Eco
RV (no. 4), in which only the 5'-part of
tau
up to position 238 had been replaced by a 5'-portion of Ty1
delta
; in this case, potential Ty4 transcription initiation sites were included. The
level of [beta]-galactosidase activity produced from this fusion amounted to 13 U,
suggesting that a putative negative regulatory element located within the Ty4
LTR, positions 1-238, had been removed. Finally, the comparably low level of [beta]-galactosidase activity (5 U) obtained from construct pTy1-Ty4-
lacZ
(no. 5), containing the transcriptional start site of Ty1 but the translational
start site of Ty4, suggests that sequences between the transcriptional and
translational start sites of Ty4 are required for expression. Furthermore,
internal Ty4 elements necessary to initiate transcription might exist:
otherwise no expression should be detectable with constructs nos 4 and 5, as it
is anticipated that the contribution of the Ty1 LTR portions is nil.
The low copy number element Ty4 from yeast shares several characteristics with
other retrotransposons. Although some 15-25 copies of Ty4 LTRs dispersed throughout the yeast genome (
1
) indicate that Ty4 is able to transpose, no direct proof of transposition of
this element has been obtained (
14
), possibly due to the fact that Ty4 expression occurs at a low level. In this
study, we have attempted to exploit features of Ty4 transcription and to
determine the nature of Ty4 transcripts.
In a first set of experiments we found that Ty4 mRNA is far less abundant in
yeast than is Ty1 mRNA. In fact, Northern blot analysis turned out to be too
insensitive to detect Ty4 transcripts. Only when we applied reverse
transcription together with the PCR technique (RT-PCR), which provides a
powerful tool for the detection of rare mRNAs (
24
), were we able to visualize Ty4 transcripts. Using
GAL4
as a standard (
25
), we estimate that Ty4 is transcribed at a comparable level, i.e. Ty4 mRNA
represents only 1.6 * 10
-5
of total mRNA (~0.1 molecules mRNA per cell). Care was taken in these experiments to
exclude any DNA contamination of the mRNA preparations: to ensure that genomic
DNA was not serving as the template in the PCR reaction after RT, the RNA
samples from MC45-5/A and YM4127 yeast cells were exhaustively digested with RNase-free DNase. Additionally, we confirmed that the RT-PCR products were
derived from RNA by using actin gene primers as an internal control and the
accuracy of all PCR products was verified by diagnostic restriction enzyme
analysis.
Studies with yeast mutants have shown that the expression of Ty1 is influenced
by pleiotropic factors of the Swi/Snf class, which, in conjunction with the
SPT
/
SIN
genes, mediate transcriptional control via effects on chromatin (
26
,
27
). Especially,
sin4
mutants have been shown to repress transcription from a Ty1
delta
element (
28
,
34
). To investigate whether Ty4 expression is similarly affected in
sin4
mutants, we compared the Ty4 transcript level in a wild-type (YS18) to that in a mutant yeast strain (YS84) using competitive
RT-PCR. Our results clearly demonstrate that Ty4 expression is not influenced
by the pleiotropic factor Sin4p. Similar experiments in other mutant
spt
or
sin
backgrounds will be necessary to see any influence of members of the Spt/Sin
family on Ty4 expression.
To coordinate expression with the yeast life cycle, transcription of Ty1/2 and
Ty3 is controlled by a variety of cellular factors in conjunction with specific
Ty internal
cis
-acting elements (
1
). For example, Yu and Elder (
31
) showed that a region internal to the Ty1 coding sequence is essential for
transcription, while the
delta
region by itself had little or no ability to promote transcription. In Ty3,
major regulatory elements reside within the 5' LTR that can be viewed as a compact, highly regulated, mobile promoter
responsive to cell type and mating (
8
). In these cases, detailed analyses could largely be based on sequence
homology, which revealed the presence of particular elements, and on direct
measurement of RNA levels and/or expression levels of reporter genes in
appropriate constructs. In Ty4, these approaches are complicated by two facts:
(i) inspection of the Ty4 sequence did not reveal elements similar to those
found in the other yeast retrotransposons or yeast gene promoters, except two
PRE consensus sequences detected at positions 391 and 1258 respectively; (ii)
Northern blot analysis was hampered by the low level of Ty4 transcription. As
an alternative, we therefore attempted to determine the influence of particular
regions from Ty4 on the expression of
lac
Z in reporter constructs. These constructs were also tested for their ability to
promote expression under [alpha] pheromone induction. In Ty4 constructs (Fig.
4
, nos 2a and 2b) that are similar to that of Ty1 (Fig.
4
, no. 1), used for comparison, expression is virtually undetectable. Only when
sequences from the Ty4 5' LTR are removed (and replaced by `inert' Ty1 sequences; Fig.
4
, nos 4 and 5) does expression occur at a low level. The comparison of
constructs nos 4 and 5 suggests that internal Ty4 sequences are necessary for
expression, but that a negative regulatory element located within the first 238
bp of Ty4 has been abolished. The existence of such a negative element is also
suggested by comparing constructs nos 1 and 3: the internal control region of
Ty1 is still capable of initiating expression when fused to the Ty4 5' LTR, although the expression level is reduced more than 20 times as
compared with the cognate Ty1 construct. In accordance with this interpretation
is the finding of Bilanchone
et al
. (
10
) that a negative regulatory element is present in the Ty3 LTR which shows
matches (7 of 9 bp) to sequences in the Ty4 LTR that are located in a similar
position with respect to the 5'-end of the LTR. Nevertheless, all of these interpretations have to
be taken with precautions, because mutual (positive or negative) interactions
conferred by the sequences in the above constructs cannot be excluded and
sequence similarities might also be misleading. From our experiments, for
example, we have to conclude that despite a reasonable match to the PRE element
(
35
), the PRE consensus sequences present in Ty4 (
14
) have no relevance for [alpha] pheromone induction. Altogether, more detailed experiments are necessary
to clarify this point.
A requirement for active transposition of retrotransposons is the synthesis of
RNA intermediates that are capable of serving as templates to regenerate full-length DNA with complete LTRs during reverse transcription and replication
(
12
,
36
). In all cases known to date, the transcripts contain an R-U5 sequence at the 5'-end and an U3-R sequence at the 3'-end (
6
,
37
). Since primer extension analysis was not feasible, we examined the nature of
Ty4 mRNA by taking advantage of the RACE technique (
21
) to determine the 5'- and 3'-termini respectively of the transcripts. To ensure the
reliability of this method, we first applied it to Ty3, confirming the data of
Clark
et al
. (
6
). Four different Ty4 RNA start sites were mapped within the 5' LTR (Fig.
5
). The finding of multiple transcriptional start sites is not unusual and has
been observed, for example, to exist in Ty3 (
6
). Unfortunately, the RACE protocol does not allow determination of a major
transcriptional start site. However, by several criteria, position 276 is
considered to act as a major transcriptional initiation site that in
conjunction with an appropriate termination site would result in a mRNA the
retrotranscript of which would fulfil the basic requirement for a
retrotranspostion intermediate (
12
,
33
). As we have demonstrated, the majority of the Ty4 transcripts lack the U3-R sequence which is required for obligate strand transfer during DNA
synthesis (
36
). Only one of the minor transcripts (d in Fig.
6
) does extend into the U3-R region of the 3' LTR and ends some 20 nt beyond the position equivalent to the
transcriptional start site located at position 276 (in the 5' LTR). Thus, a minor retrotranscript extending from position 276 through
position 6139 would give an overlap of some 20 bp, which would be sufficient
for strand transfer (cf. Fig.
5
). These positions in Ty4 would be homologous to the transcriptional start site
and the site of poly(A) addition respectively, as found for Ty1 and Ty3 (
6
,
11
,
37
). Furthermore, a putative `efficiency element' (TACATATA) found in some yeast
genes (
38
,
39
) is present at a canonical distance (~100 bp) upstream of this polyadenation site in Ty4. However, no
`positioning element' (
38
) is present in the interspace. It should be noted that inspection of the Ty4
LTR sequence has not revealed any further reasonable array of the three
polyadenylation signals as described for yeast genes (
38
,
40
,
41
) nor any one of the signal signatures described for Ty1 or Ty3 (
6
,
42
). Because of these notions, we have thus far discounted position 154 (Fig.
5
) as a major initiation site. Of course, we have to consider the possibility
that yet unknown signals operate in Ty4 giving rise to appropriate `full-length' transcripts. In any case, however, we can rule out that these are
major Ty4 transcripts.
From the above results we arrive at a two-fold explanation for the anticipated low transpositional activity of the
Ty4 element. Firstly, we have shown that transcription of this element occurs at an extremely low
level. From the experiments measuring the expression levels in
lac
Z fusion constructs, we favour the idea that active repression of Ty4
transcription is the major reason for this effect. This is supported by
preliminary experiments in our laboratory indicating that a protein(s)
specifically binds to the 5'-part of the
tau
element, which presumably contains a repressor binding element. It will be of
interest to identify this protein(s) which might influence Ty4 transcription.
Secondly, the generation of truncated Ty4 transcripts over `full-length' transcripts significantly reduces the predisposition for
transpositional activity. Altogether, these findings suggest that Ty4
constitutes a type of retroelement whose behaviour drastically deviates from
that of other retrotransposons or retroviruses described thus far.
We are grateful to Dr S. B. Sandmeyer (University of California, Irvine, CA) for
providing the pDF445 plasmid. We would like to thank H.-G. Klobeck for helpful technical advice with the PCR experiments and
discussions. This work was supported by the Deutsche Forschungsgemeinschaft
(Sonderforschungsbereich 190). A.M.H. is a recipient of a fellowship from the Bavarian State.
5
'
-Termini.
In order to characterize the Ty4 promoter in more detail, we determined
transcript initiation sites by the RACE-PCR protocol (
21
). The first strand Ty4 cDNA was made using a 21mer reverse transcription primer
(Ty4-480) complementary to positions 461-480 in Ty4. After tailing, the 48mer RACE-N primer was annealed and a PCR reaction carried out using
the forward primer RACE-O together with the backward primer Ty4-480; a nested PCR followed (RACE-I and Ty4-428). Amplimers were cloned and nine of these clones
subjected to sequence analysis. Four different Ty4 transcript 5'-termini were monitored by this approach and are shown in Figure
5
. In contrast to the primer extension mapping method, the RACE technique does
not allow quantification, so that we cannot specify a major transcriptional
start site. To prove the reliability of the RACE method, we used Ty3 in control
experiments and found virtually the same three start sites as determined by
Clark
et al
. (
6
) using reverse transcription extension (data not shown).
REFERENCES
Return