Download A unique pattern of intrastrand anomalies in base

© 2000 Oxford University Press
Nucleic Acids Research, 2000, Vol. 28, No. 23 4679–4688
A unique pattern of intrastrand anomalies in base
composition of the DNA in hypotrichs
David M. Prescott* and Sarah J. Dizick
University of Colorado, Department of Molecular, Cellular and Developmental Biology, Boulder, CO 80309-0347, USA
Received August 16, 2000; Revised and Accepted October 4, 2000
ABSTRACT
The 50 non-coding bases immediately internal to the
telomeric repeats in the two 5′ ends of macronuclear
DNA molecules of a group of hypotrichous ciliates
are anomalous in composition, consisting of 61%
purines and 39% pyrimidines, A>T (ratio of 44:32),
and G>C (ratio of 17:7). These ratio imbalances violate
parity rule 2, according to which A should equal T and
G should equal C within a DNA strand and therefore
pyrimidines should equal purines. The purine-rich
and base ratio imbalances are in marked contrast to
the rest of the non-coding parts of the molecules,
which have the theoretically expected purine content
of 50%, with A = T and G = C. The ORFs contain an
average of 52% purines as a result of bias in codon
usage. The 50 bases that flank the 5′ ends of macronuclear sequences in micronuclear DNA (12 cases)
consist of ∼50% purines. Thus, the 50 bases in the 5′
ends of macronuclear sequences in micronuclear
DNA are islands of purine richness in which A>T and
G>C. These islands may serve as signals for the
excision of macronuclear molecules during macronuclear development. We have found no published
reports of coding or non-coding native DNA with
such anomalous base composition.
INTRODUCTION
The DNA in the macronuclei of hypotrichs is organized in
short molecules with lengths ranging from ∼400 to ∼15 000 bp
with averages of ~2000 bp. With rare exceptions each molecule
encodes a single gene. These short DNA molecules are derived
from very high molecular weight DNA of micronuclear
chromosomes after cell mating through a complex series of
DNA processing steps that produce a new macronucleus. The
first step in conversion of a micronucleus to a macronucleus is
polytenization of the micronuclear chromosomes. In the polytene chromosome stage, thousands of short, non-coding
segments called internal eliminated segments, or IESs, are
eliminated from the genes that subsequently become the short
macronuclear DNA molecules (1). The segments within a gene
that are separated by IESs are called macronuclear-destined
segments, or MDSs; these are ligated when IESs are removed.
In some micronuclear precursors of macronuclear molecules
the MDSs are in scrambled configuration. These MDSs
become unscrambled and ligated in the orthodox order in
conjunction with IES removal from polytene chromosomes.
The polytene chromosomes are then fragmented, and ∼95% or
more of the original micronuclear DNA sequence complexity
is eliminated. These eliminated sequences are the spacers in
micronuclear DNA between the successive genes. The ∼5% of
sequence complexity that remains forms the gene-size molecules
of the macronuclear genome. Telomeric repeats are added to
the ends of the thousands of gene-size molecules as they are
released from micronuclear DNA. Finally, the short molecules
are replication-amplified to one to several thousand copies
each, depending on the species.
The macronuclear, gene-size DNA molecules created by
these processing steps generally consist of a single open
reading frame (ORF) and 5′ and 3′ non-coding segments, as
illustrated in Figure 1a. The ends of the molecules are capped with
telomeric repeats of G4T4/C4A4 with a 16-base, single-stranded 3′
overhang of 5′-T4G4T4G4-3′ (except Euplotes species, which
lack the last two Gs at the 3′ ends of both telomeres). The 5′
non-translated leaders and 3′ non-translated trailers are AT-rich in
contrast to the lower AT content of ORFs. The base compositions
of 5′ leaders, ORFs and 3′ trailers are here documented quantitatively in the analysis of a group of 72 macronuclear DNA
molecules.
The main point of this paper is a previously unreported
structural feature of macronuclear DNA molecules: regular
intrastrand anomalies in base composition of non-coding
regions at the ends of the molecules. This anomalous base
composition at the ends of macronuclear molecules raises three
questions: (i) how are such anomalies in base composition
created and maintained; (ii) how are the anomalies restricted to
particular segments of macronuclear DNA molecules and their
micronuclear precursors; and (iii) what is the structural/
functional significance of the anomalous base composition?
MATERIALS AND METHODS
Forty-six macronuclear DNA molecules with known coding
functions and 34 molecules with unidentified coding functions
studied in this report are listed in Table 1 with the organisms of
origin, molecule lengths, GenBank accession numbers and
references. Some of the molecules listed as unpublished were
sequenced directly from PCR products. For others, PCR
products were cloned in plasmid Puc 19 and sequenced. PCR
products that included terminal, telomere repeats were generated
with telomere primers of 5′-C4A4C4A4C4NN-3′, where NN
stands for one of 16 possible dinucleotides. Telomere primers
were paired with primers at internal locations in molecules.
*To whom correspondence should be addressed. Tel: +1 303 492 8381; Fax: +1 303 492 7744; Email: [email protected]
4680 Nucleic Acids Research, 2000, Vol. 28, No. 23
Figure 1. (a) The average structure of the 46 macronuclear DNA molecules in Table 1 with identified coding functions. 5′ non-coding leader = 290 bp; ORF =
1410 bp; 3′ non-coding trailer = 201 bp. (b) The AT percentage along the length of the macronuclear DNA molecule that encodes βTP in O.trifallax. Averages
were determined for 50-base segments that successively overlapped by 25 bases from point to point. Data are from DuBois and Prescott (25).
Some of the unpublished sequences in Table 1 were determined
on randomly picked, macronuclear molecules of Oxytricha
nova cloned in plasmid Puc 19 using M13 and M13R primers
for sequencing. Molecules in Table 1 listed as unpublished
were sequenced by the departmental sequencing facility.
Sequences of macronuclear molecules from Euplotes species
(Table 2) were all obtained from GenBank. Molecules were
analyzed with MacVector sequence analysis software.
RESULTS
Overall characteristics of molecules
The two groups of macronuclear molecules listed in Table 1
were analyzed. One group consists of 46 molecules from
Oxytricha, Stylonychia, Histriculus and Halteria species with
identified coding functions, determined in various laboratories
(see accession and reference numbers in Table 1). The second
group consists of 26 molecules from O.nova with, as yet,
unidentified coding functions, sequenced by us. The group of
46 molecules with known coding functions has well identified
ORFs, and thus, the lengths of the non-coding 5′ leaders and 3′
trailers are well defined. The lengths of the 5′ leaders and
3′ trailers in the 26 molecules with unidentified coding functions have not been established. For this reason, the two groups
of molecules were initially analyzed separately for base
compositions at their ends.
Halteria is traditionally considered to be an oligotrich, not a
hypotrich. However, its macronuclear DNA is organized into
the short, gene-size molecules characteristic of hypotrichs (56;
D.M.Prescott, unpublished results), and it replicates this DNA
by means of replication bands (D.M.Prescott, unpublished
results). The sequence of the ssrDNA in our Halteria strain
indicates that it is closely related to the group including
Oxytricha/Stylonychia, etc.
The 46 molecules with identified coding function range in
length from 821 to 5007 bp (excluding telomere sequences),
with an average of 1901 bp (Fig. 1a). 5′ non-coding leaders
range from 72 to 2230 bp, average 290 bp; ORFs range from
248 to 4542 bp, average 1410 bp; and 3′ non-coding trailers
range from 91 to 994 bp, average 201 bp. The base composition
along the lengths of the molecules varies in a regular pattern.
The AT bp content is high in the 5′ leaders, ranging from 59 to
84%, averaging 72 ± 6%. The AT bp content decreases to an
average of 52 ± 7% in the ORFs, and rises again to 70 ± 5% in
the 3′ trailers. This is illustrated in Figure 1b for the molecule
encoding β telomere binding protein (βTP gene) in Oxytricha
trifallax. The 5′ leader is 78% AT, the ORF is 51% AT and the
3′ trailer is 73% AT. The βTP gene contains an intron of 86 bp,
which is evident in Figure 1b by its high AT content (79%); a
high AT content is characteristic of introns in hypotrich genes.
The 26 molecules from O.nova with unidentified coding
function (Table 1) range in length from 292 to 6004 bp, with an
average of 1900 bp. ORFs that occupy >50% of the lengths of
many of these molecules were identified by start and stop
codons, using the NCBI ORF finder. However, these ORFs
have not been confirmed by other means, e.g., detection of
mRNAs, cDNAs or encoded proteins, or by the known bias in
codon usage in hypotrichs, and are not recorded in this report.
None of these putative ORFs encode amino acid sequences
with significant identity with amino acid sequences in
GenBank (BLAST search). Thus, the 26 molecules of
unknown coding function could not be used with confidence to
define the total length of 5′ leaders and 3′ trailers. However,
these molecules were useful in analyzing the composition of the
Nucleic Acids Research, 2000, Vol. 28, No. 23 4681
Table 1. Forty-six sequenced macronuclear DNA molecules with known coding functions from Oxytrichia, Stylonychia, Histriculus and Halteria species and 26
sequenced DNA molecules with unidentified coding functions from Oxytrichia nova
Organism
Gene
Length (bp)
Accession no.
Reference
Oxytricha trifallax
Actin I
1558
U18940
(2)
Oxytricha nova
Actin I
1604
M22480
(3)
Stylonychia lemnae
α-tubulin
1829
X01746
(4)
Oxytricha granulifera
α-tubulin
1672
Z11763
(5)
Oxytricha trifallax
αTP
2166
AF067831
(6)
Stylonychia mytilus
αTP
2141
X61749
(7)
Oxytricha nova
DNA pol α
4978
U02001
(8)
Oxytricha trifallax
DNA pol α
5007
U59426
(9)
Stylonychia lemnae
Calmodulin
821
M76407
(10)
(11)
Stylonychia lemnae
β2-tubulin
1831
X06874
Stylonychia lemnae
β1-tubulin
1846
X06653
(11)
Stylonychia lemnae
α2-tubulin
1731
X12365
(12)
X07905
Oxytricha granulifera
CCT-γ
2036
Y11967
(13)
Oxytricha nova
hsp-70
2654
U37280
(14)
Oxytricha nova
Histone H4
1659
M24411
(15)
Oxytricha nova
Actin II
1428
U06071
(16)
Oxytricha trifallax
Telomerase
3642
AF060230
(17)
Stylonychia mytilus
βTP
1738
X61748
(7)
Stylonychia lemnae
EF 1α
1846
X57926
(18)
S82230
Oxytricha nova
βTP
1790
M31310
(19)
Oxytricha nova
αTP
2217
M68930
(20)
Stylonychia lemnae
MA52
3791
X73879
(21)
Stylonychia lemnae
MA68
1892
X73880
(21)
Oxytricha fallax
Actin gene
1507
J01163
(22)
Histriculus cavicola
Actin I
1585
Y12047
(23)
Oxytricha nova
Actin III
1448
AF134156
(24)
Stylonychia mytilus
β-tubulin
1817
AF188162
V.Ku and D.M.Prescotta
Stylonychia lemnae
βTP
1837
AF190703
V.Ku and D.M.Prescotta
Halteria grandinella
Actin I
1545
AF188161
L.McCollester and D.M.Prescotta
Oxytricha trifallax
Ubiquitin
1115
AF188158
E.Quirk and D.M.Prescotta
Oxytricha trifallax
Asparagine-rich (1.7)
1724
AF188163
A.F.Greslin and D.M.Prescotta
Oxytricha trifallax
βTP
1858
U63565
(25)
Oxytricha nova
ATPase
3290
AF188144
J.D.Prescott, S.J.Dizick and D.M.Prescotta
Oxytricha nova
Chaperonin-β
1893
AF188130
″
Oxytricha nova
CGI-128
836
AF188131
″
Oxytricha nova
Euk. RF-1
1835
AF188150
″
Oxytricha nova
Polytene protein
1302
AF188141
″
Oxytricha trifallax
Histone H4
1629
AF192970
P.P.Kneeland and D.M.Prescotta
Unknown F
Actin I
1518
AF188159
K.E.Orr and D.M.Prescotta
Oxytricha nova
C2
778
K02624
(26)
Oxytricha nova
AS2
491
M57403
(27)
Stylonychia lemnae
pob4
1217
X16613
(28)
Stylonychia lemnae
U-44 ORF 1
1209
M75100
(29)
Stylonychia lemnae
1.3
1325
X72956
(30)
Stylonychia lemnae
1.1
1086
X72955
(30)
Oxytricha nova
AS1
513
M57402
(27)
Oxytricha nova
Unknown function
539
AF188128
J.D.Prescott, S.J.Dizick and D.M.Prescotta
4682 Nucleic Acids Research, 2000, Vol. 28, No. 23
Table 1. Continued
Organism
Gene
Length (bp)
Accession no.
Reference
Oxytricha nova
″
6004
AF188129
J.D.Prescott, S.J.Dizick and D.M.Prescotta
Oxytricha nova
″
949
AF188132
″
Oxytricha nova
″
2791
AF188133
″
Oxytricha nova
″
3311
AF188134
″
Oxytricha nova
″
1368
AF188135
″
Oxytricha nova
″
1911
AF188136
″
Oxytricha nova
″
292
AF188137
″
Oxytricha nova
″
1996
AF188138
″
Oxytricha nova
″
1494
AF188139
″
Oxytricha nova
″
964
AF188140
″
Oxytricha nova
″
3262
AF190702
″
Oxytricha nova
″
483
AF188142
″
Oxytricha nova
″
2516
AF188143
″
Oxytricha nova
″
2312
AF188145
″
Oxytricha nova
″
3874
AF188146
″
Oxytricha nova
″
1106
AF188147
″
Oxytricha nova
″
3137
AF188148
″
Oxytricha nova
″
1445
AF188149
″
Oxytricha nova
″
946
AF188151
″
Oxytricha nova
″
1295
AF188152
″
Oxytricha nova
″
555
AF188153
″
Oxytricha nova
″
2771
AF188154
″
Oxytricha nova
″
1424
AF188155
″
Oxytricha nova
″
1015
AF188156
″
Oxytricha nova
″
1016
AF188157
″
aUnpublished.
50 bases in non-coding regions at the very ends of molecules,
immediately adjacent to the telomeres, since it can be safely
assumed the non-coding 5′ leaders and 3′ trailers are longer
than 50 bp. For example, for the 16 genes of known coding
function in O.nova in Table 1 the leaders and trailers have
average lengths of 299 bp (range, 82 to 1153) and 228 bp
(range, 91 to 446), respectively.
Forty-one molecules from Euplotes species (Table 2) serve
as a comparison group. Although Euplotes is a hypotrich, it is
very distantly related to the hypotrich group (including
Halteria) in Table 1 according to the sequence of the ssrDNA
gene (57). As in other hypotrich species the AT base pair
content in leaders and trailers is high, averaging 78 and 74%,
respectively. The AT content of ORFs averages 59%.
Base composition at the ends of macronuclear molecules
The average percentage of A+G in the first 80 bp in the 5′ end
of the sense strand of the non-coding leader, and the 5′ end of
the antisense strand in the 46 molecules of known coding
function listed in Table 1 is shown in 10-base segments in
Figure 2a and b. The average A+G content is high in the first
60–70 bases in the 5′ end of the sense strand and the first
∼50 bases of the 5′ end of the antisense strand. The A+G
percentage then declines to 53–54% by 80 bases and falls
further to ∼50% in the rest of the non-coding leader and trailer
(Table 3). Thus, the 5′ ends of both strands are similarly purine
rich.
The corresponding values for the combined 5′ ends of the
two strands in the 26 molecules with unidentified coding
functions are similar to the values for the 46 molecules with
identified coding functions (Fig. 2c). The data for the two
strands of the 26 molecules are combined because the orientation
of their ORFs are unknown, and therefore leader and trailer
ends are not distinguishable. The A+G content in the 5′ ends
ranges from 58 to 63% in the 10-base segments up to 50 bases
and declines to 54–55% between 50 and 80 bases.
The average A+G percentages in 10-base segments for all
144 5′ ends of the two strands in the 72 molecules are
combined in Figure 2d. Average A+G percentages range from
58 to 63% in the 50 bases at 5′ ends and decline to 54–56%
between 50 and 80 bases. Although the averages for both
groups of molecules in Table 1 show a dip in A+G content at
∼15 bases (Fig. 2a, b and c), the dip is probably not significant
because some individual molecules in both groups do not show
the dip, although these molecules have the same overall high
percentage of A+G in their 5′ ends as all other molecules.
The data for the 10-base segments are of limited use because
averages for such small short segments have high standard
deviations, but these averages define roughly the extent,
i.e. ∼50 bases, of a consistent purine richness in the 5′ ends.
Nucleic Acids Research, 2000, Vol. 28, No. 23 4683
Table 2. Forty-one sequenced macronuclear DNA molecules from Euplotes species
Organism
Gene
Length (bp)
Accession no.
Reference
Euplotes aediculatus
γ-tubulin
1610
X85233
C.Weiligmanna
Euplotes aediculatus
Telo p123
3227
U95964
(31)
Euplotes crassus
51 kDa Telo binding protein 1546
M96818.1
(32)
Euplotes crassus
α-phosphoinisotide
524
M63336
(33)
Euplotes crassus
Actin
1247
J04533
(34)
Euplotes crassus
β-tubulin
1468
J04534
(34)
Euplotes crassus
ConF5
1096
AF063084
(35,36)
Euplotes crassus
ConZA7
867
U65976
(35)
Euplotes crassus
ConZA8
1632
AF061334
(35,36)
Euplotes crassus
γ-tubulin gene 1
1571
X85234
(37)
Euplotes crassus
γ-tubulin gene 2
1581
X85235
(37)
Euplotes crassus
Histone H1-1
1229
AF127331
(38)
Euplotes crassus
Histone H1-2
633
AF127332
(38)
Euplotes crassus
Histone H3
607
U65646
(35)
Euplotes crassus
Histone H4
1818
U75430
(39)
Euplotes crassus
J2
1620
AF072707
C.L.Jahn and C.M.Tebeaua
Euplotes crassus
Unknown gene
1014
AF072706
C.L.Jahn and C.M.Tebeaua
Euplotes crassus
ORF 1/ORF2
761
M73025
(40)
Euplotes crassus
PGK
1372
U97355
R.E.Pearlmana
Euplotes crassus
Protein kinase
1751
U47679
(41–43)
Euplotes crassus
rpl29
553
U13207
(44)
Euplotes crassus
Telo binding protein
1545
M96819
(32)
Euplotes crassus
Telo RNA Component
609
M33461
(45)
Euplotes eurystomas
Histone H1
1254
L15293
(46)
Euplotes eurystomas
5sRNA
846
X13718
(47)
Euplotes eurystomas
hsp 11
2294
L15291
L.J.Hauser, A.L.Herrmann and D.E.Olinsa
Euplotes eurystomas
hsp 1a
2149
L15292
L.J.Hauser, A.L.Herrmann and D.E.Olinsa
Euplotes eurystomas
Polyubiquitin
898
M57231
(48)
Euplotes octocarinatus
α-tubulin
1531
X69466
(49)
Euplotes octocarinatus
β-tubulin
1468
X69467
(49)
Euplotes octocarinatus
Centrin
634
Y18899
A.Lianga
Euplotes octocarinatus
γ-tubulin
1577
X71353
(50)
Euplotes octocarinatus
γ-tubulin-2
1577
Y09553
M.Tan, A.Liang and K.Heckmanna
Euplotes octocarinatus
pher4-MAC
1622
X58838
(51)
Euplotes octocarinatus
Phermone 1
1600
Y15316
(52)
Euplotes octocarinatus
Phermone 2
1702
Y15318
(52)
Euplotes octocarinatus
Phermone 3
1694
Y15317
(52)
Euplotes octocarinatus
Phermone 5
1560
Y17505
(53)
Euplotes octocarinatus
RPA2
3879
Euplotes octocarinatus
Euplotes vannus
RPB2
α-tubulin
3715
1505
X66451
(54)
S50742
(55)
X66453
(54)
S50850
(55)
Z11769
(5)
aUnpublished.
Therefore, the average percentage of A+G in the entire 50 base
segment in the 5′ ends was determined and compared to the
A+G percentage in other parts of the molecules. This was done
separately for the two groups of molecules in Table 1 because
the molecules with established ORFs have defined 5′ leaders
and 3′ trailers. The average percentages of A, T, G, C and A+G
for the various segments of the molecules with established
ORFs are given in Table 3. The first 50 bases in the 5′ end of the
sense strand contain an average of 61% A+G (range, 54–80%).
Allowing for a transition from higher to lower purine content
4684 Nucleic Acids Research, 2000, Vol. 28, No. 23
Table 3. Average percentage base composition of the sense strand (5′→3′) of
46 macronuclear molecules with identified coding functions
T
G
First 50 bases of the 5′ leader 44 ± 5
A
32 ± 6
17 ± 5
7±3
61 ± 6
5′ Leader minus first 60 bases 38 ± 5
34 ± 4
13 ± 4
15 ± 4
51 ± 4
52 ± 6
ORF
C
A+G
30 ± 6
24 ± 7
22 ± 5
24 ± 8
3′ Trailer minus last 60 bases 35 ± 6
35 ± 4
14 ± 5
16 ± 4
49 ± 6
Terminal 50 bases of 3′ trailer 30 ± 7
45 ± 6
8±4
17 ± 6
38 ± 7
Results of similar analysis of the 26 molecules of O.nova
with undefined ORFs are given in Table 4. In this case the 5′
end of the sense strand cannot be distinguished from the 5′ end
of the antisense strand because the ORFs are unidentified. The
base compositions of the 50 bases at the 5′ end of one strand
and the 50 bases at the 3′ end of the same strand from each of
the 26 molecules were averaged. The 50 bases at the 5′ end are
high in A+G (61%), whereas at the 3′ end they are low in A+G
(40%). Thus, the 5′ ends of both strands are high in A+G,
similar to the 5′ ends of the 46 molecules in Table 3. The
combined data for all 72 molecules (144 5′ ends) showing the
high average content of A+G in the 5′ ends are given in
Table 5.
Table 4. Average percentage base composition of the 50 bases in the 5′ and 3′
ends of one strand of each of the 26 macronuclear molecules with unidentified
coding functions from O.nova
A
T
G
Fifty bases at 5′ end
of one strand
44 ± 6
32 ± 7
17 ± 5
C
7±4
A+G
61 ± 7
Fifty bases at 3′ end
of the same strand
33 ± 7
44 ± 6
7±4
16 ± 6
40 ± 6
Table 5. Average percentage base composition of the 144 5′ ends of strands in
the 72 macronuclear molecules listed in Table 1
Fifty bases at 5′ ends
Figure 2. Average percentage content of A+G in 10-base segments for the 80 bases
(a) in the 5′ end of the sense strand and (b) at the 5′ end of the antisense strand of
46 macronuclear molecules with known coding function. (c) Average percentage
content of A+G in 10-base segments for the 80 bases in the 5′ ends of both
strands combined for 26 macronuclear molecules with unidentified coding
functions from O.nova. (d) Average percentage content of A+G for 5′ ends in
(a), (b) and (c) (144 ends).
between 50 and 60 bases, the average content of A+G of the 5′
strand in the leaders from 60 bases to the start of the ORF is
51%. The average percentage of A+G then rises slightly to
52% in the ORFs, declines to 49% in the 3′ trailer, excluding
the last 60 bases, and is 38% in the last 50 bases before the
telomere. The pattern of A+G percentages is reciprocal, beginning
at the 5′ end of the complementary strand with corresponding
numbers 62, 51, 48, 49 and 39%. Thus, the 50 bases at the 5′
end of the sense strand and the 50 bases at the end of the antisense strand have the same purine richness, i.e. the ends of the
molecule are symmetrical.
A
T
G
C
A+G
44 ± 5
32 ± 7
17 ± 5
7±3
61 ± 6
The average intrastrand percentages of the 4 bases in
50 bases in the 5′ ends for the two groups of molecules both
individually and for the combined groups are also given in
Tables 3–5. The average percentages of A, T, G and C in each
table are remarkably consistent, with values of 44, 32, 17 and
7%, respectively, in every data set. Thus, the purine richness in
the 5′ ends is primarily due to the high content of A residues.
In addition, the ratios of A:T and G:C are anomalously high,
violating parity rule 2 (58,59). According to this rule, A should
equal T and G should equal C within a strand. Parity rule 2 may
be violated for DNA subjected to strand-specific selection
pressure, e.g., codon or mutational bias; for example, differences in mutation rates between leading and lagging strand
templates during replication (60). The observation A≠T and
G≠C in the 5′ ends is in sharp contrast with the approximate,
theoretically expected average A:T ratio of 36:35 and average
G:C ratio of 14:15 for the combined leaders and trailers of the
sense strand, minus the 60 bases at each end in the case of the
Nucleic Acids Research, 2000, Vol. 28, No. 23 4685
40 molecules with identified ORFs in Table 3. Both A and G
contribute to the purine enrichment; A is increased 22% and
G is increased 17% (Table 3; both 5′ ends combined).
The average percentage ratio of A:T and G:C for the ORFs is
30:24 and 22:24, respectively, and the ORFs contain an
average of 52% purines (Table 3). The ratio of G:C is essentially
1:1, but the ratio of A:T is somewhat higher, i.e. 30:24. The
slightly higher purine percentage of 52% and the attendant
higher ratio of A:T reflects both amino acid content of the
encoded proteins and bias in codon usage for particular amino
acids. An extreme example is the ORF encoding a glutamine/
asparagine-rich protein in O.trifallax (AF188/63 in Table 1).
Ninety-six out of 320 amino acids in this protein are glutamine,
encoded predominantly by CAA and TAA, and asparagine,
encoded predominantly by AAT and AAC. Thirty-nine percent
of the bases in this ORF are A and 18% are G, resulting in an
unusually high purine content of 57%. Bias in codon usage is
pronounced in hypotrichs (61). Among the 31 most commonly
used codons (used for ≥30% of the occurrences of a particular
amino acid), the average ratio of A:T and G:C is 33:26 and
18:23, respectively, which account at least in part for the
greater occurrence of A in the 40 ORFs (Table 3).
A similar analysis was done on the 41 macronuclear
molecules for the Euplotes species listed in Table 2. The
Euplotes molecules are on average shorter (1557 bp) than
molecules from the other hypotrichs (∼1900 bp). The 5′ leaders
and 3′ trailers in Euplotes are also shorter; averages of 147 and
128 bp, respectively. Nineteen of the 41 molecules have 5′
leaders of 60 bp or less (range, 32–1262 bp), and 10 of the 41
have 3′ trailers of 60 bp or less (range, 26–477 bp). Therefore,
the analysis was done using the 30 bases (rather than 50 bases)
at the 5′ ends of the molecules so that all molecules could be
included without invading ORFs. ORFs range from 324 to
3585 bp, with an average of 1161 bp. The purine content of the
terminal 30 bases is 50 ± 7%, significantly below the 61%
found in the other hypotrich group. The intrastrand ratios of
A:T and G:C for these 30 bases are exactly 1:1 for the two 5′
ends combined (percentages of A, T, G and C of 38, 38, 12 and
12 respectively, Table 6). ORFs contain an average of 55%
purines and average percentages for A, T, G and C of 34, 25,
21 and 20 respectively. A is higher than T, and total purine
content is high; this is the result, at least in part, of codon bias.
Thus, Euplotes molecules do not have the anomalous base
composition in the 5′ ends seen in the other hypotrichs.
Sequences flanking genes in the micronuclear DNA
Within the micronuclear chromosomes, macronuclear sequences
lack telomere repeats and are flanked by micronuclear-specific
Table 6. Average percentage base composition of the 30 bases in the 5′ ends
of 41 macronuclear DNA molecules (82 total ends) from Euplotes species
listed in Table 2
T
G
C
Thirty bases at 5′ ends 38 ± 6
A
38 ± 7
12 ± 5
12 ± 6
50 ± 8
ORFs
25 ± 6
21 ± 7
20 ± 6
55 ± 7
34 ± 6
A+G
DNA that is eliminated during macronuclear development.
The sequences of seven micronuclear-flanking DNA segments
at the two ends of the macronuclear sequences encoding the
scrambled α telomere protein (αTP) (62), the non-scrambled
gene encoding hsp 70 (K.R.Lindauer, R.C.Anderson and
D.M.Prescott, unpublished results), and the non-scrambled C2
gene (63) plus the sequence of the DNA flanking the 3′ end of
the scrambled DNA polymerase α (DNA pol α) gene (64) are
available in O.nova. In the 14 MDSs of the scrambled αTP
gene, MDSs 1 and 14 are in terminal positions, so that the first
50 bases of the 5′ leader and the last 50 bases of the 3′ trailer
are in the orthodox positions in the micronuclear version of the
gene. The same is true for the 3′ end of the scrambled micronuclear DNA pol α gene, but the 5′ end of this gene (MDS 1)
is internal in the gene because of an inversion.
In the 50-base segment that flanks the 5′ end of the sense or antisense strand (seven cases) the average percentages of A, T, G and
C are 36, 41, 11 and 12 (Table 7). This composition differs markedly from that in the immediately adjacent 50 bases in the 5′ ends
of the macronuclear sequences (Table 3). For example, in the 50base flanking segments the intrastrand A:T and G:C ratios are
close to 1:1, and the average purine content is 47 ± 17% versus 61
± 6% in the first 50 bases in the macronuclear sequence. Clearly,
a transition takes place from orthodox A:T and G:C ratios (1:1)
and orthodox percentage of A+G (47%) in the flanking strand to
anomalous A:T and G:C ratios (44:32 and 17:7, respectively) and
an anomalously high A+G percentage (61%) in the first 50 bases
of the macronuclear sequence.
According to the model for unscrambling of the DNA pol α
gene in O.nova during macronuclear development (65,66),
unscrambling leaves the gene still integrated in its micronuclear chromosome, and displaces IESs 13, 16, 12, 17 and 11
(total, 48 bases), arranged in that order, to the flank of MDS 1.
According to the model, the gene is then excised from the
chromosome by cutting between the flanking DNA formed by
IESs 13, 16, 12, 17 and 11 and the end of MDS 1. Similarly, in
the DNA pol α gene of O.trifallax the 5′ flanking sequence is
formed by IESs 15 and 19 as a result of unscrambling,
according to the recombination model of unscrambling.
Similarly, the flanking DNA at the 5′ end of the sense strand of
Table 7. Average percentage base composition in the 50 bases in micronuclear DNA immediately flanking ends of 5′ strands of seven macronuclear sequences
(with telomeres) and five reconstructed flanking sequences for scrambled genes. The right column with the average percentage A+G for macronuclear sequences
is included for comparison
A
T
G
C
A+G
A+G in the first 50 bases of the
macro-nuclear sequences
50 bases flanking 5′Æ3′ strands of orthodox macronuclear
sequences in micronuclear DNA
36 ± 10
41 ± 9
11 ± 3
12 ± 3
47 ± 10
68 ± 6
Reconstructed 50-base flanking sequence for scrambled
macronuclear sequences in micronuclear DNA
36 ± 3
36 ± 10
12 ± 5
16 ± 5
48 ± 11
62 ± 17
Combined averages
36 ± 8
39 ± 9
12 ± 5
13 ± 4
48 ± 14
65 ± 19
4686 Nucleic Acids Research, 2000, Vol. 28, No. 23
Figure 3. Composite summary of the average A+G content and A:T and G:C ratios of 50 bases of micronuclear flanking sequences and of various parts of the
macronuclear sequences.
the actin I gene in O.nova is formed by IESs 7, 8 and 5 as a
result of unscrambling. The flanking DNA at the 5′ end of the
actin I gene in O.trifallax is formed by IESs 8, 9 and 5. The
flanking sequence at the 3′ end of the sense strand of the actin I
gene in O.nova is formed by IES 6. The average base composition
for these five theoretically reconstituted flanking sequences is
A = 36, T = 36, G = 12 and C = 16 (Table 7). In contrast, the
first 50 bases in the 5′ ends of the relevant macronuclear
sequences contain an average 62 ± 17% A+G (Table 7). Thus,
reconstructed flanking sequences for scrambled genes show
the same transition at their junction with macronuclear
sequences, after unscrambling, as do the orthodox flanking
sequences. The combined averages for the seven orthodox and
five reconstructed flanking sequences are 36:39 for A:T and 12:13
for G:C. The A+G percentages are 47 and 48%, respectively.
The picture that emerges for the micronuclear version of a
macronuclear molecule is summarized in Figure 3. In flanking
DNA reading 5′→3′ in the sense strand, A+G = 47% and A ≅ T
and G ≅ C. In the first 50 bases after the transition into the
macronuclear sequence (leader or trailer), A+G = 61% and
A>T and G>C. After the first 50 bases A+G is again ∼50% and
A ≅ T and G ≅ C in the rest of the leader and trailer. In the ORF
A+G = 52% (A>T and G = C). Thus the average purine content
changes from ∼50% in flanking DNA to 61% in the first 50 bases
to an average of 50% in the remainder of the leader and trailer
and 52% in the ORF.
DISCUSSION
Approximately 50 bp at the ends of macronuclear molecules
from the Oxytricha/Stylonychia/Histriculus/Halteria group of
ciliates is anomalous in its base composition. In the 5′ ends of
strands, A>T (ratio of 44:32) and G>C (ratio of 17:7), resulting
in purine richness (purine:pyrimidine ratio of 61:39), primarily
because of the high percentage of A residues (44%). An identical,
average, anomalous composition in 5′ ends (A = 44%, T = 32%,
G = 17% and C = 7%) was observed for a group of 46 molecules
with identified coding functions and for a group of 26
molecules of unidentified coding functions. The anomalous
composition is in contrast to the rest of the non-coding leaders
and trailers in the 46 molecules with identified coding functions,
in which A ≅ T and G ≅ C, and A+G = 50%. The high purine
content of the 50 bases in the ends of 5′ strands also contrasts
with the average composition of the 5′ strands in ORFs, in
which A is a little higher than T, and G ≅ C, and the total purine
content is 52%. The slight average richness of A in ORFs is due to
bias towards the use of A-rich codons in these ciliates and the
amino acid composition of the proteins encoded by the 46 genes.
The anomalous base composition observed in the 5′ ends of
macronuclear molecules is present in the micronuclear DNA
and is transmitted to macronuclear molecules during genome
processing after cell mating. This suggests that the significance
of the anomalous base composition resides in some function or
activity in the micronuclear genome. Since the anomalous
composition occurs in both 5′ ends of macronuclear precursors
in micronuclear DNA and, in some cases, is separated from
ORFs by hundreds of base pairs, it is unlikely that significance
of the anomalous composition is in any way connected to
coding functions of genes.
Two possibilities are that the anomalous base composition
derived from differential mutation rates in the two strands
during DNA replication or originated by selection as a signal to
mark the ends of micronuclear precursors of macronuclear
molecules. For example, anomalous base composition in
bacterial DNA has been suggested to arise from preferential
deamination of bases in lagging strand templates, converting
C→T and A→G, so that T>A and G>C in that strand (67).
Lagging strand templates are transiently single-stranded, and
single-stranded DNA is deaminated in vitro more than 100 times
more rapidly than double-stranded DNA (68,69). How such a
replication effect could be restricted to ∼50 bases at the ends of
micronuclear precursors of macronuclear molecules is problematic. In any case, the deamination idea would presumably create
duplex DNA in which T>A and G>C in one strand (A>T and
C>G in the complementary strand) but in macronuclear ends A>T
and G>C in one strand, and T>A and C>G in the complement.
Thus, deamination appears unlikely to be responsible for the
anomalous base composition of macronuclear molecular ends.
The anomalous base composition in hypotrichs could
possibly act as a signal for excision of macronuclear molecules
from micronuclear DNA during macronuclear development after
cell mating. One test of this hypothesis is the base composition of
DNA that flanks macronuclear precursors in micronuclear
DNA. In 50-base segments flanking the 5′ end of the sense
strand of three genes and 50-base segments flanking the
complement to the 3′ end of the sense strand of four genes, the
average base A:T and G:C base ratio is 36:41 and 11:12,
respectively, and the A+G average is 47% (Table 7). These are
orthodox values and contrast with the anomalous base percentages
Nucleic Acids Research, 2000, Vol. 28, No. 23 4687
in the 50 bases in the 5′ ends of the immediately adjacent
macronuclear sequences.
Five additional flanking sequences reconstructed by
unscrambling the micronuclear precursors of actin I and DNA
pol α molecules, using the recombination model, also have, on
average, orthodox base compositions (Table 7). Thus, the
50 bp in the ends of micronuclear precursors are islands of
anomalous base composition embedded in sequence of
orthodox composition, i.e. A = T, G = C and A+G = ∼50%.
This is consistent with the idea that the anomalous base
composition marks the ends of micronuclear precursors of
macronuclear genes and might serve as identifying targets for
excision of macronuclear molecules in these hypotrichs.
In contrast, the ends of 41 macronuclear molecules in
Euplotes species do not have anomalous compositions.
However, Euplotes molecules contain a 10-bp consensus
sequence 17 bp downstream or upstream of apparent excision
points that acts to identify those excision points (36). No
consensus sequences have been detected in the 50-bp segments
either in micronuclear precursors or in micronuclear flanking
sequence in the group including Oxytricha/Stylonychia, etc. Both
poly(A) and poly(T) tracts up to 6 bases long are present, but they
occur at the expected frequency for AT-rich DNA and occupy
random positions in the 50-base segments at 5′ ends of strands.
Anomalous base compositions could possibly act as a target
for proteins responsible for excision of macronuclear molecules.
Such a target signal would presumably be imprecise since it is
diffuse, and there is no discernible specific pattern in the
sequences with the anomalous base composition. This imprecision
is consistent with the earlier finding by Baird and Klobutcher
(70) that in O.nova the telomere addition site (excision site) for
a particular macronuclear molecule can vary over a range of
several base pairs. This has been confirmed in several hypotrichs in the Oxytricha/Stylonychia group (K.E.Croft, K.E.Orr
and D.M.Prescott, unpublished results). In contrast, the excision
point in Euplotes, which has an apparent specific signal
sequence, is precise to the base pair (70).
What structural property of the anomalous 50-bp segments
might be specifically recognized by excision machinery is not
known. Anomalous base composition might, however, affect
DNA structure to produce a specific, recognizable target,
perhaps by creating a differential chromatin structure along
micronuclear DNA. Finally, we have been unable to find any
published reports of native, non-coding DNA sequences with
the extreme anomaly in base composition described here for 5′
ends of macronuclear DNA.
ACKNOWLEDGEMENTS
We thank Gayle Prescott for manuscript preparation and David
Hoffman for technical assistance with cloning of some DNA
molecules. This work is supported by NSF grant MCB9974680 to D.M.P.
REFERENCES
1. Tausta,S.L. and Klobutcher,L.A. (1990) Nucleic Acids Res., 18, 845–853.
2. DuBois,M. and Prescott,D.M. (1995) Proc. Natl Acad. Sci. USA, 92,
3888–3892.
3. Greslin,A.F., Loukin,S.H., Oka,Y. and Prescott,D.M. (1988) DNA, 7,
529–536.
4. Helftenbein,E. (1985) Nucleic Acids Res., 13, 415–433.
5. Gaunitz,F. (1990) Thesis. Abteilung Zellbiologie, Universität Tübingen,
Germany.
6. Prescott,J.D., DuBois,M.L. and Prescott,D.M. (1998) Chromosoma, 107,
293–303.
7. Fang,G.W. and Cech,T.R. (1991) Nucleic Acids Res., 19, 5515–5518.
8. Mansour,S.J., Hoffman,D.C. and Prescott,D.M. (1994) Gene, 144, 155–161.
9. Hoffman,D.C. and Prescott,D.M. (1997) J. Mol. Evol., 45, 301–310.
10. Gaunitz,C., Witte,H. and Gaunitz,F. (1992) Gene, 119, 191–198.
11. Conzelmann,K.K. and Helftenbein,E.J. (1987) J. Mol. Biol., 198, 643–653.
12. Helftenbein,E. and Muller,E.W. (1988) Curr. Genet., 13, 425–432.
13. Palmedo,G. and Ammermann,D. (1997) Eur. J. Biochem., 247, 877–883.
14. Anderson,R.C., Lindauer,K.R. and Prescott,D.M. (1996) Gene, 168, 103–107.
15. Harper,D.S. and Jahn,C.L. (1989) Gene, 75, 93–107.
16. Mitcham,J.L. and Prescott,D.M. (1994) Gene, 144, 119–122.
17. Bryan,T.M., Sperger,J.M., Chapman,K.B. and Cech,T.R. (1998)
Proc. Natl Acad. Sci. USA, 95, 8479–8484.
18. Bierbaum,P., Donhoff,T. and Klein,A. (1991) Mol. Microbiol., 5, 1567–1575.
19. Hicke,B.J., Celander,D.W., Macdonald,G.H., Price,C.M. and Cech,T.R.
(1990) Proc. Natl Acad. Sci. USA, 87, 1481–1485.
20. Gray,J.T., Celander,D.W., Price,C.M. and Cech,T.R. (1991) Cell, 67, 807–814.
21. Maercker,C. and Lipps,H.J. (1994) Gene, 141, 147–148.
22. Kaine,B.P. and Spear,B.B. (1982) Nature, 295, 430–432.
23. Pérez-Romero,P., Villalobo,E., Díaz-Ramos,C., Calvo,P. and Torres,A.
(1999) J. Euk. Microbiol., 45, 469–472.
24. Dizick,S.J. and Prescott,D.M. (1999) Eur. J. Protistol., 35, 375–377.
25. DuBois,M.L. and Prescott,D.M. (1997) Mol. Cell. Biol., 17, 326–337.
26. Klobutcher,L.A., Jahn,C.L. and Prescott,D.M. (1984) Cell, 36, 1045–1055.
27. Harper,D.S., Song,K. and Jahn,C.L. (1991) Gene, 99, 55–61.
28. Wegner,M., Helftenbein,E., Muller,F., Meinecke,M., Muller,S. and
Grummt,F. (1989) Nucleic Acids Res., 17, 8783–8802.
29. Kreyenberg,H. (1993) Thesis. Abteilung Zellbiologie, Universität
Tübingen, Germany.
30. Eder,C., Maercker,C., Meyer,J. and Lipps,H.J. (1993) Int. J. Dev. Biol.,
37, 473–477.
31. Lingner,J., Hughes,T.R., Shevchenko,A., Mann,M., Lundblad,V. and
Cech,T.R. (1997) Science, 276, 561–567.
32. Wang,W., Skopp,R., Scofield,M. and Price,C.M. (1992) Nucleic Acids Res.,
20, 6621–6629.
33. Klobutcher,L.A., Turner,L.R. and Peralta,M.E. (1991) J. Protozool., 38, 425–427.
34. Harper,D.S. and Jahn,C.L. (1989) Proc. Natl Acad. Sci. USA, 86, 3252–3256.
35. Ling,Z., Ghosh,S., Jacobs,M.E. and Klobutcher,L.A. (1997) J. Euk. Microbiol.,
44, 1–11.
36. Klobutcher,L.A., Gygax,S.E., Podoloff,J.D., Vermeesch,J.R., Price,C.M.,
Tebeau,C.M. and Jahn,C.L. (1998) Nucleic Acids Res., 26, 4230–4240.
37. Tan,M. and Heckmann,K. (1998) Gene, 210, 53–60.
38. Ray,S., Jahn,C., Tebeau,C.M., Larson,M. and Price,C. (1999) Gene, 231, 15–20.
39. Jahn,C.L., Ling,Z., Tebeau,C.M. and Klobutcher,L.A. (1997) Proc. Natl
Acad. Sci. USA, 18, 1332–1337.
40. Krikau,M.F. and Jahn,C.L. (1991) Mol. Cell. Biol., 11, 4751–4769.
41. Hale,C.A., Jacobs,M.E., Estes,H.G., Ghosh,S. and Klobutcher,L.A.
(1996) J. Euk. Microbiol., 43, 389–392.
42. Baird,S.E., Fino,G.M., Tausta,S.L. and Klobutcher,L.A. (1989) Mol. Cell.
Biol., 9, 3793–3807.
43. Ghosh,S., Jaraczewski,J.W., Klobutcher,L.A. and Jahn,C.L. (1994)
Nucleic Acids Res., 22, 214–221.
44. Jahn,C.L., Erbeznik,M., Jaraczewski,J.W., Melek,M. and Shippen,D.E.
(1994) Gene, 151, 231–235.
45. Shippen-Lentz,D. and Blackburn,E.H. (1990) Science, 247, 546–552.
46. Hauser,L.J., Treat,M. and Olins,D.E. (1993) Nucleic Acids Res., 21, 3586.
47. Roberson,A.E., Wolffe,A.P., Hauser,L.J. and Olins,D.E. (1989)
Nucleic Acids Res., 17, 4699–4712.
48. Hauser,L.J., Roberson,A.E. and Olins,D.E. (1991) Chromosoma, 100, 386–394.
49. Liang,A., Schmidt,H.J. and Heckmann,K. (1994) J. Euk. Microbiol., 41,
163–169.
50. Liang,A. and Heckmann,K. (1993) Gene, 136, 319–322.
51. Meyer,F., Schmidt,H.J. and Heckmann,K. (1992) Dev. Genet., 13, 16–25.
52. Bruenen-Nieweler,C., Weiligmann,J.C., Hansen,B., Kuhlmann,H.W.,
Moellenbeck,M. and Heckmann,K. (1998) Eur. J. Protistol., 34, 124–132.
53. Moellenbeck,M. and Heckmann,K. (1999) Eur. J. Protistol., 35, 225–230.
54. Kaufmann,J. and Klein,A. (1992) Nucleic Acids Res., 20, 4445–4450.
55. Kaufmann,J., Florian,V. and Klein,A. (1992) Nucleic Acids Res., 20,
5985–5989.
56. Metenier,G. and Hufschmid,J.-D. (1988) J. Protozool., 35, 71–73.
57. Lynn,D.H. and Sogin,M.L. (1988) BioSystems, 2, 249–254.
4688 Nucleic Acids Research, 2000, Vol. 28, No. 23
58.
59.
60.
61.
Sueoka,N. (1995) J. Mol. Evol., 40, 318–325; Erratum 42, 323.
Lobry,J.R. (1995) J. Mol. Evol., 40, 326–330; Erratum 41, 680.
Morton,B.R. (1999) Proc. Natl Acad. Sci. USA, 96, 5123–5128.
Hoffman,D.C., Anderson,R.C., DuBois,M.L. and Prescott,D.M. (1995)
Nucleic Acids Res., 23, 1279–1283.
62. Mitcham,J.L., Lynn,A.J. and Prescott,D.M. (1992) Genes Dev., 6, 788–800.
63. Klobutcher,L.A., Jahn,C.L. and Prescott,D.M. (1984) Cell, 36, 1045–1055.
64. Hoffman,D.C. and Prescott,D.M. (1996) Nucleic Acids Res., 24, 3337–3340.
65.
66.
67.
68.
Prescott,D.M. and Greslin,A.F. (1992) Dev. Genet., 13, 66–74.
Hoffman,D.C. and Prescott,D.M. (1997) Nucleic Acids Res., 25, 1883–1889.
Lobry,J.R. (1996) Mol. Biol. Evol., 13, 660–665.
Frederico,L.A., Kunkel,T.A. and Shaw,B.R. (1990) Biochemistry, 29,
2532–2537.
69. Beletskii,A. and Bhagwat,A.S. (1996) Proc. Natl Acad. Sci. USA, 93,
13919–13924.
70. Baird,S.E. and Klobutcher,L.A. (1989) Genes Dev., 3, 585–597.