ABSTRACT
The macronuclear genome of
Euplotes crassus
contains two different genes,
EFA1
and
EFA2
, encoding EF-1
[alpha]
proteins. They are derived from micronuclear precursors in the course of a sexual process termed conjugation. We have found that two
apparent micronuclear precursors exist for
EFA1
. They differ in their potential coding sequences and in the internal sequences
interrupting the genes, which are normally removed during the processing from
micronuclear into macronuclear genes. One of these genes is not processed into a macronuclear gene and has accumulated C
->
T transitions in a limited section of the coding region. The gene obviously constitutes a recent duplication which has lost its
ability to be processed into a macronuclear gene and has therefore become a
micronuclear pseudogene. The true
EFA1
precursor harbours a novel type of internal sequence in addition to a classical
AT-rich IES. As usual, only one micronuclear
EFA2
precursor gene was found. Its coding sequence is interrupted by a 79 bp TelIES.
Ciliated protozoa are characterized by nuclear dimorphism (
1
,
2
). The cells contain generative diploid micronuclei and vegetative macronuclei.
This situation can be viewed as a parallel to the germ line-soma differentiation of higher eukaryotes, since only macronuclei are
physiologically active. The DNA contents and sequence complexities of the two nuclear types differ. Macronuclei have lower sequence complexity but contain more DNA. This is due to the mode of
their generation. During a sexual phase cells of compatible mating types form
pairs and exchange meiosis products of their micronuclei. They fuse with the
non-transferred haploid nuclei of the respective partner cell to form
synkarya. The new diploid nucleus in each of the partner cells divides at least
once to form one or more micronuclei, depending on the organism, and the
macronuclear precursor, termed the anlage. The anlagen chromosomes undergo
rearrangements, which are most pronounced in the hypotrichous ciliates (
3
-
6
) to which
Euplotes crassus
belongs. First, the anlagen chromosomes undergo polytenization. Different types
of sequences which interrupt the macronucleus-destined sequences of the micronuclear genes are then removed in several steps (
7
,
8
) and the chromosomes are subsequently fragmented into gene sized molecules. These nascent macronuclear genes are differentially amplified and telomeres are attached (
9
-
11
) before a further, general amplification follows to complete the development of the new macronuclear chromosomes. The macronucleus-destined sequences, i.e. the precursor sequences of the macronuclear
genes, are clustered in the micronuclear chromosomes (
12
). Since in addition to the removed intragenic sequences the long intergenic
regions are lost during formation of the macronuclear chromosomes, the sequence complexity is reduced by >90%, while the total DNA content increases.
Three classes of sequences that can interrupt the coding regions of macronucleus
gene precursors have been described. Internal eliminated sequences (IES) are AT-rich non-coding short sequences (
13
), transposon-like elements of
E.crassus
(Tec) are very frequent long sequences with open reading frames and inverted
repeats (
14
-
16
), which probably enable them to transpose (
17
). TelIES elements are very short relatively GC-rich sequences containing C
4
A
4
motifs reminiscent of macronucleus chromosome telomeres (
18
). All these elements are flanked by TA repeats. A common excision mechanism has been proposed for the Tec and IES sequences
involving circular intermediates which are most likely not formed in the course
of removal of the very short TelIES (
19
,
20
).
We have been interested in the organization and expression of genes encoding EF-1[alpha] translation factors in hypotrichous ciliates.
Euplotes crassus
contains two different such genes,
EFA1
and
EFA2
, which are both transcribed (
21
). They differ in copy number and codon usage, pointing to different expression
levels, which have indeed been found. Since we have indications that the
macronuclear gene copy number might be correlated with the timing of excision
from the polytene anlagen chromosome (Dönhoff and Klein, unpublished results), we were interested to characterize
the micronuclear counterparts of the two
EFA
genes. The micronuclear
EFA1
gene is interrupted by an AT-rich IES sequence and in addition contains a long novel interrupting
sequence with 12 bp terminal repeats. In addition, we have discovered an
EFA1
pseudoprecursor which is not processed into a macronuclear gene. It contains
the same IES element but lacks the novel intervening sequence found in the
functional precursor and shows point mutations in a narrow range of its open
reading frame. Only one
EFA2
precursor was found, which harbours a short TelIES element.
Euplotes crassus
strains Por-3 and Liv-1 (
22
), as well as the feeding alga
Dunaliella tertiolecta
, were kindly supplied by Prof. P. Luporini (University of Camerino, Italy). The
ciliates and the algae were grown at 22oC in artificial seawater as described previously.
General DNA preparation, electrophoresis, restriction analysis and hybridization techniques have been previously described (
23
-
25
).
For the preparation of total cellular DNA
E.crassus
was grown to a density of 2-3 * 10
3
/ml. Feeding was then stopped and the organisms left without food for 3 days.
The cells were harvested by two subsequent filtration steps through 30 and 10 [mu]m nylon gauze, rinsed off with seawater and collected by low speed centrifugation. They were resuspended in 3 vol lysis buffer (10 mM Tris-HCl, pH 7.5, 10 mM EDTA, 250 mM NaCl, 0.5% SDS) (
8
) and lysed at 65oC for 15 min. Proteins were digested at 50oC overnight after addition of 200 [mu]g/ml proteinase K. The solution was extracted with phenol and the
DNA dissolved in TE (10 mM Tris-HCl, pH 7.5, 1 mM EDTA) after ethanol precipitation.
The cells were collected as described above. They were resuspended in 10 ml TE/ml packed cells. After addition of 0.1% Triton X-100 and 1 [mu]g/ml trypsin the suspension was incubated at room temperature. Lysis
of the cells and the macronuclei was followed microscopically. As soon as the
macronuclei had been destroyed the trypsin treatment was stopped by
transferring the lysate to ice and addition of 2 [mu]g/ml trypsin inhibitor. The nuclei were spun out of the lysate at low speed
(500
g
) and 4oC, then washed twice with TE to remove most of the macronuclear DNA. After
the addition of 1 ml lysis buffer, the micronuclear DNA was prepared by
proteinase K digestion, phenol extraction and ethanol precipitation as
described above. The DNA was applied to a preparative 1% agarose gel and separated from residual macronuclear DNA. The visible high molecular weight DNA was cut out of the gel and
electrophoresed through a similar gel. The DNA was removed from the agarose by
electroelution and precipitated by isopropanol in the presence of 1% glycogen.
DNA amplification by polymerase chain reactions (PCR) and inverse PCR were
performed as described (
26
,
27
). The products were directly sequenced according to established methods (
28
,
29
). The primers used are listed in Table
1
. Their positions are shown in Figure
1
. Sequencing of cloned DNA was performed according to the dideoxynucleotide
termination method (
30
).
Table 1
The method of Hongyo
et al.
(
31
) was modified as follows. PCR products were separated on agarose gels to
separate them from primer DNA and eluted from the gel. Aliquots of 10 ng each
of the DNA samples to be compared were mixed in 15 [mu]l H
2
O. Then 1 [mu]l sample buffer (15% Ficoll, 0.25% bromophenol blue, 0.25% xylene cyanol FF)
and 1 [mu]l 1 N NaOH were added. The DNA was denatured at 90oC for 4 min and immediately chilled on ice. The samples were analysed
on 120 * 155 * 1 mm 10% polyacrylamide gels containing 5% glycerol in 0.5* TBE (4.5 mM Tris-borate, pH 8.3, 0.25 M EDTA). The gels were run at
120 V for 16-18 h at 4 and 20oC in parallel, since the single strand conformation critically
depends on temperature. The DNA was visualized by silver staining (
32
).
Southern hybridization of a restriction digest of micronuclear DNA showed that
two different micronuclear
EFA1
genes exist. The DNA was digested with the restriction endonuclease
Stu
I, which does not cut within the sequence of the macronuclear gene nor in the
IES sequence described above (compare Fig.
3
). The DNA was then electrophoretically separated and a Southern blot was
hybridized with the macronuclear
EFA1
sequence as a probe. The autoradiogram presented in Figure
6
shows a relatively strong signal caused by a macronuclear impurity and two
additional ones. One of these corresponds to the
Stu
I fragment containing the mic
EFA1
A gene. The second one is due to a restriction fragment containing most of a
second micronuclear
EFA1
gene, mic
EFA1
B, as will be shown below. The fact that the signal strength of this smaller
micronuclear fragment is lower than that of the larger one indicated that one
of the
Stu
I sites was located in an intervening sequence interrupting the coding sequence
of mic
EFA1
B. This could indeed be shown by inverse PCR. Micronuclear DNA was digested with
Afl
II or
Eae
I restriction endonucleases, which fulfil the same criteria as
Stu
I, and subsequently religated to obtain DNA circles. These were suitable
substrates for amplification with the primer pairs 1R2 and 1F3 or 1R6 and 1F7
respectively. The amplification products were re-amplified with primer pairs 1R1 and 1F4 or 1R5 and 1F8 respectively, to
make sure that specific products had been obtained. Sequence analyses yielded
the nucleotide sequences of the ends of an intervening sequence between
nucleotides 1134 and 1135 of the
EFA1
sequence. Strikingly, a 12 bp sequence is duplicated at the ends of this
intervening sequence, in contrast to the TA duplication found in IES and Tec
sequences (Fig.
7
). Attempts to amplify the missing part of the intervening sequence with appropriate primer combinations, derived from the sequences obtained after the inverse PCR reactions, failed. This indicates that it may be
quite long, because >2 kb amplification products with other primer-template combinations were obtained under the same conditions. The
amplification product obtained with the primer pair 1R1 and 1F4 also includes
the region containing the 347 bp IES in the mic
EFA1
A gene. Sequence analysis of the resulting product did show the known 347 bp IES
in the mic
EFA1
B gene in the same location. If this gene is indeed the precursor of the
EFA1
gene, identical coding sequences are expected in both genes. Sequence
determination of the portion of the mic
EFA1
B gene known to have deviations from the macronuclear sequence in the mic
EFA1
A gene yielded the same sequence as in the macronuclear gene. This supports the
view that mic
EFA1
B is the precursor of
EFA1
.
PCR reactions with micronuclear DNA as template and three sets of primers
derived from the sequence of the macronuclear
EFA2
gene (compare Fig.
1
and Table
1
) yielded overlapping products, one of which (obtained with the primer pair 2F1
and 2R1) was slightly longer than the corresponding PCR product obtained with a macronuclear
EFA2
template. Sequence analysis showed a 79 bp TelIES sequence characterized by sequences reminiscent of the telomeric repeats of
E.crassus
macronuclear chromosomes (Fig.
8
). It ends in a 5 bp repeat.
Comparative sequence analysis of the two micronuclear
EFA1
genes has shown that they are very similar. This concerns not only the coding
sequences but also the 347 bp IES, which is identical in both genes, as well as
the flanking sequences up to positions -220 and +130 at their 5'- and 3'-ends (data not shown). Since no selective
pressure is imposed on the micronuclear genes except for the conservation of
their capability to be processed into macronuclear genes, one would expect an
accumulation of mutations in both the coding and non-coding sequences in an inactive gene. However, only a few mutations in a
limited area of the coding sequence of the pseudoprecursor have occurred so
far. This indicates that we are looking at a recent duplication.
Analysis of the flanking sequences of the micronuclear
EFA1
genes (Fig.
9
) did not provide any clue as to why any of them should be processed more or
less efficiently. A consensus sequence for chromosome fragmentation has been
established for
E.crassus
(
33
). One of the potential fragmentation sequences detected in either the non-coding regions of the potential macronucleus-destined sequences or the neighbouring regions in the micronuclear sequences shows significant sequence similarity with the consensus sequence in each case, while the sequence
similarity of the other one is at most moderate. One fragmentation sequence
close to each end of the genes would suffice for excision of the macronucleus-destined sequences according to the current model of the excision process
(
33
). Of course, other determinants in the intergenic regions may be necessary,
which might be more distant than the sequences we have analysed adjacent to the
two
EFA1
genes. Such fragmentation signals would therefore have escaped our attention.
Since the mic
EFA1
B gene apparently contains an additional internal sequence, it cannot be ruled
out that it could be a prerequisite for the processing of mic
EFA1
B into the macronuclear
EFA1
gene. This intervening DNA sequence is most likely >2 kb long. It will
therefore be interesting to obtain its entire sequence and look for potential
open reading frames. Such analyses are presently underway.
The mutations encountered in the mic
EFA1
A gene are all C -> T transitions. This situation is reminiscent of repeat induced point
mutations (RIP) in fungi, in which duplicated genes are mutated in the same way
(
35
-
37
). In this case the mutations are correlated with methylation of cytosines.
There are indications that cytosine methylation can cause elevated rates of
deamination, leading to the observed mutation type (
38
). If this was true, the pseudogene could be expected to be more highly
methylated. This might also indicate that duplication has led to its
positioning in a special region of the chromosome, most likely outside a gene
cluster, since high mutation rates in macronucleus-destined sequences have not been observed in authentic micronuclear
precursor genes so far. This interpretation is in line with the recent finding
that chromosomal regions exist into which Tec elements can be positioned outside gene clusters. These Tec elements are amplified to a lesser degree in the polytene stage of the anlagen chromosomes at a late
stage of the processing events. They are probably lost during chromosome
fragmentation (S. Frels, C. M. Tebeau, S. Z. Doktor and C. L. Jahn, personal communication).
The 12 bp repeats flanking the intervening sequence in the mic
EFA1
B gene (Fig.
7
) suggest that the rearrangement of the coding sequence leading to the coherent
macronuclear gene sequence involves a site-specific recombination event. This is reminiscent of the events occurring
during formation of several genes in
Oxytricha
, where macronuclear genes have to be formed from patches of coding sequences in their micronuclear precursors, involving intricate recombination processes between short repeated sequences separating the various macronucleus-destined sequences (reviewed in
4
). In the cyanobacterium
Anabaena
removal of an intervening sequence by site-specific recombination in the course of the activation of genes involved
in nitrogen fixation is another example of the transformation of a precursor
into a functional gene by site-specific recombination (
39
). We are presently looking for a potential ring intermediate, which is to be
expected if our hypothesis invoking an intrachromosomal element that is removed
by a simple recombination step is correct.
We thank Helga Bestgen for photographic work and Sophie Curtenaz for critical
reading of the manuscript. This work was supported by the Deutsche
Forschungsgemeinschaft.
Name
Sequence
Position
a
1F1
GATAAATTTTCCGAGTTGAGCG
3
1R1
CAATAACGACAAGATTGAGATG
139
1F2
GAGTCTGCTGAAATGGGTAAGGC
232
1R2
AGCCTTGAGCTTATCGAGTACCC
294
1F3
GAAGCTGGTATCTCCAAGGAAGG
469
1F4
CATACAGTGAAGGACAGATACG
581
1R3
CTGGCTTGTATCCGACCTTAGCG
660
1F5
GCCGGTATGGTCATCACCTTCGC
883
1R4
GTTTCCTGGAGCCGCTTCTGGGAC
981
1F6
ATTGTCACACTGCCCATATTGCC
1148
1R5
CTTGGTGAGGAGTTCTTCG
1197
1R6
TGATGAGTCCAGCATCTCCTGCC
1270
1F7
GCTGTCAGAGATATGAGACA
1336
1F8
CTGTCGCAGTCGGTGTCATC
1357
1R7
ATGATGACTATACTATAGAAGAG
1488
1iF
GCCATGGTATAGGCCCTATC
723
1iR
GTTGGGCCCTTGTATAATCT
743
2F1
CCATTGGAGATGACAAATTTATCGC
45
2R1
CCAGATGGTCAGAGACGTTCTCTAC
685
2F2
GAGATTCAATGAGATCGTAGAG
644
2R2
CTTGTTGTGATGAGCTTCTAAAG
1004
2F3
GAAGTTCAGTTCACCACTGGAG
942
2R3
CCTCTGTGTAATCAAGAGG
1495
REFERENCES
Return