ABSTRACT
Although mitochondria and chloroplasts are considered to be descendants of
eubacteria-like endosymbionts, the mitochondrial RNA polymerase of yeast is a nucleus-encoded, single-subunit enzyme homologous to bacteriophage T3 and T7 RNA
polymerases, rather than a multi-component, eubacterial-type
[alpha]
2
[beta][beta]'
enzyme, as encoded in chloroplast DNA. To broaden our knowledge of the
mitochondrial transcriptional apparatus, we have used a polymerase chain
reaction (PCR) approach designed to amplify an internal portion of phage T3/T7-like RNA polymerase genes. Using this strategy, we have recovered
sequences homologous to yeast mitochondrial and phage T3/T7 RNA polymerases
from a phylogenetically broad range of multicellular and unicellular
eukaryotes. These organisms display diverse patterns of mitochondrial genome organization and expression, and include species that separated from the main eukaryotic line early in the
evolution of this lineage. In certain cases, we can deduce that PCR-amplified sequences, some of which contain small introns, are localized in
nuclear DNA. We infer that the T3/T7-like RNA polymerase sequences reported here are likely derived from genes
encoding the mitochondrial RNA polymerase in the organisms in which they occur,
suggesting that a phage T3/T7-like RNA polymerase was recruited to act in transcription in the
mitochondrion at an early stage in the evolution of this organelle.
The evolutionary origin of the mitochondrial (mt) transcription system is
puzzling. It is generally accepted that both mitochondria and chloroplasts
arose from eubacteria-like endosymbionts, closely related to contemporary [alpha]-Proteobacteria and Cyanobacteria, respectively (
1
-
3
). Consistent with this view, chloroplast DNA is known to encode components of a
eubacteria-like [alpha]
2
[beta][beta]' RNA polymerase (RNAP) (
4
). No such genes have been found in any of the mitochondrial genomes sequenced
to date; instead, all of the genes for the mitochondrial transcriptional
machinery appear to be encoded by nuclear DNA, with the protein products being
imported into the organelle (
5
).
In yeast (
Saccharomyces cerevisiae
), the mitochondrial RNA polymerase (mtRNAP) is encoded by a nuclear gene,
RPO41
(
6
,
7
), and is a homolog of the single-polypeptide RNAPs of bacteriophages T3 and T7 (
8
). In view of the similar endosymbiotic origins of mitochondria and chloroplasts
and the discovery of eubacteria-like RNAP genes in chloroplast DNA, this is a surprising finding.
Moreover, this observation raises questions about the nature of the
mitochondrial transcription system in other eukaryotes, and the evolutionary
origin(s) of this system.
The identification of a second T3/T7-like mtRNAP sequence in
Neurospora crassa
(
9
), coupled with the recent appearance of homologous sequences in various
expressed sequence tag (EST) databases (human, rice,
Caenorhabditis elegans
; Fig.
1
), prompted us to devise a polymerase chain reaction (PCR) amplification
strategy to search more widely within the eukaryotic lineage for phage T3/T7-like RNAP sequences. Information about the types and phylogenetic
distribution of mtRNAPs, and their structural similarity to one another, is
necessary to determine whether the mtRNAPs of different eukaryotes all arose
from a single common ancestor and, if so, what the evolutionary source of this
enzyme might have been. In particular, because the multicellular eukaryotes
(animals, fungi, plants) represent relatively late radiations in the eukaryotic
lineage (
10
), it is important to explore a phylogenetically broad range of earlier
diverging unicellular eukaryotes (protists), in order to address the question
of whether a T3/T7-like mtRNAP was acquired early or at a relatively late stage in the
evolution of the mitochondrial transcription system. Comparative information
about mtRNAPs and other transcriptional components in different eukaryotes is
also essential for defining and understanding species-specific peculiarities in the biochemical mechanism of expression of
mitochondrial genomes that can vary tremendously in size, base composition and
organization.
Nuclear DNA from
Pycnococcus
provasolii
,
Thraustochytrium aureum
and
Isochrysis
sp. Tahiti was isolated by B.F. Lang. The organisms were cultured axenically,
with
Isochrysis
and
Pycnococcus
grown in sterilized seawater with additions (F/2 medium; for further details,
consult WWW site URL http://megasun.bch.
umontreal.ca/People/lang/FMGP/methods.html). Cells were broken mechanically by
shaking with glass beads (
11
), DNA was solubilized in 0.5% SDS, and proteins were hydrolyzed in the presence
of 100 [mu]g/ml proteinase K. After removal of detergent by salt precipitation (1 M
NaCl, 1 h on ice), the isolated total cellular DNA was further purified by
equilibrium centrifugation in CsCl density gradients (1.1 g/ml CsCl, 10 [mu]g/ml Hoechst dye 33258 (Serva); 40 000 r.p.m., 48 h). The main, lower band
in the gradient, representing the nuclear DNA fraction in all three species,
was used.
Nuclear DNA from
Acanthamoeba castellanii
(strain Neff; axenic culture) was prepared by K.M. Lonergan from isolated
nuclei and further purified by two rounds of CsCl density-gradient centrifugation after RNase A treatment (
12
).
Cells from an axenic culture of
Cryptomonas
[Phi] were supplied by M.A. Ragan, and DNA and cellular RNA were isolated from
these by D.F. Spencer. Cells were lysed in 1% SDS containing 1 M sodium
perchlorate, following which the lysate was extracted with chloroform/isoamyl
alcohol and nucleic acids precipitated with isopropanol. The pellet was
redissolved and extracted several times with phenol-cresol, after which the nucleic acids were twice precipitated from
ethanol before further purification by CsCl density-gradient centrifugation in the presence of Hoechst dye 33258. The main-band (nuclear) DNA fraction and RNA pellet were recovered.
DNA was prepared from isolated nuclei of an axenic culture of
Naegleria fowleri
(strain LEE) and further purified by CsCl density-gradient centrifugation (
13
). The
N.fowleri
DNA sample was obtained from R. N. Band via A. Roger and P. Keeling (Department
of Biochemistry, Dalhousie University).
Cells of
Tetrahymena pyriformis
(axenic culture), from which total cellular DNA was prepared by phenol
extraction followed by RNase treatment, were grown by J. Edqvist.
A wheat (
Triticum aestivum
) cDNA clone bank was graciously provided by B.G. Lane. The library was
constructed with mRNA that had been isolated as described (
14
), using [lambda]gt11 as vector in a Y1090 host (B.G. Lane, personal communication).
Total cellular RNA from rice (
Oryza sativa
cv. Lacassine) was kindly provided by P. Gros. The RNA sample was prepared from
leaves homogenized in the presence of guanidinium hydrochloride, followed by
phenol extraction.
Degenerate primers used in this study are listed in Figure
1
. Primers specific for [lambda]gt11 DNA were used to pre-amplify cDNA inserts, as follows: forward, 5'-ACTCCTGGAGCCCGTCAGTA-3'; reverse, 5'-CAGACCAACTGGTAATGGTA-3'. For amplification of
rice cDNA, the following were used: primer 1, 5'-AACAACCGGAAGACCAACG-3'; primer 2, 5'-CGTCCATTTGACAGGGTGG-3'.
With
Pycnococcus
,
Thraustochytrium
and
Isochrysis
DNAs, PCR amplification was performed in a total volume of 50 [mu]l containing PCR reaction buffer (Pharmacia) and 100 ng DNA, 0.2 mM each
dNTP, 1 [mu]M each primer (R-3.2 and R-8.1) and 100 [mu]g/ml BSA. Reactions were carried out in a Perkin Elmer
GeneAmp PCR System 9600, according to the following protocol: denaturation at
95oC for 4 min; addition of 1 U
Taq
DNA polymerase at 75oC; 3 cycles of 95oC for 1 min, 40oC for 1 min, slow ramp to 72oC for 1 min, and hold at 72oC for 3 min; 30 cycles of 95oC for 1 min, 50oC for 1 min, slow ramp to 72oC for 1 min, hold at 72oC for 3 min; hold at 72oC for 4 min. PCR products were
resolved by agarose gel electrophoresis and extracted using QIAquick columns
(QIAGEN).
With
Acanthamoeba
,
Cryptomonas
,
Naegleria
,
Tetrahymena
and
Triticum
(wheat) DNAs, PCR was performed in a total volume of 50 [mu]l containing reaction buffer [25 mM Tricine (pH 8.5), 16 mM (NH
4
)
2
SO
4
, 2 mM MgCl
2
], 100 ng DNA, 0.1 mM each dNTP, 2 [mu]M each primer (R-3.2 or R-1 and R-8.1), 1 U
Taq
DNA polymerase (Gibco-BRL) and 0.01 U
Pfu
DNA polymerase (Stratagene). Reactions were performed in a Perkin Elmer GeneAmp
PCR System 2400, according to the following protocol: denaturation at 94oC for 3 min; 35 cycles of 94oC for 30 s, 50oC for 1 min, 55oC for 20 s, 60oC for 10 s and 72oC for 2-3 min, followed by an extension at 72oC for 10-15 min. PCR products were
isolated after electrophoresis in low-melting-point agarose gels, as above. All PCR fragments were cloned into
pT7Blue (Novagen) and sequenced using Sequenase version 2.0 (USB).
From a wheat cDNA library, inserts were pre-amplified in a total volume of 50 [mu]l containing reaction buffer [25 mM Tricine (pH 8.5), 16 mM (NH
4
)
2
SO
4
, 2 mM MgCl
2
], 100 ng of DNA, 0.1 mM each dNTP, 0.2 [mu]M each primer specific for [lambda]gt11 DNA (forward and reverse), 1 U
Taq
DNA polymerase and 0.01 U
Pfu
DNA polymerase. The reaction was performed according to the following protocol:
denaturation at 94oC for 3 min; 25 cycles of 94oC for 30 s, 55oC for 30 s and 72oC for 2 min, followed by an extension at 72oC for 5 min. A 100-fold dilution of amplified cDNAs was used for the
PCR reaction following the same method as used for
Acanthamoeba
,
Cryptomonas
,
Naegleria
and
Tetrahymena
genomic DNAs.
Additional rice sequence was obtained by rapid amplification of cDNA ends (RACE)
with construction of specific primers based on a rice EST (accession no.
D23514). Rice total cellular RNA (5 [mu]g) was used as template for reverse transcription by MuMLV (NEB) in the
presence of 1 mM dNTP (Pharmacia), 15 U RNAguard (Pharmacia) and 50 ng of
primer 1 in 20 [mu]l containing 50 mM Tris-HCl (pH 8.3), 8 mM MgCl
2
, 10 mM dithiothreitol for 1 h at 37oC, 30 min at 42oC and 15 min at 52oC. The resulting cDNAs were separated from primer using a
Centricon-30 concentrator (Amicon) and extended at the 3'-end by terminal deoxynucleotidyl transferase (Gibco-BRL) in the presence of dATP. A 200-fold dilution of polyadenylated cDNAs was taken
for two nested PCR runs, using primer 1 in the first run and primer 2 in the
second. The conditions of the two runs and the sequences of the non-specific primers used were as described (
15
).
From available protein sequences, we designed a set of degenerate
oligonucleotide primers targeted to DNA sequences encoding regions highly
conserved among known or putative mtRNAPs (all homologous to T3 and T7 RNAPs).
The latter include rice, human and
C.elegans
expressed sequence tags (ESTs) that have recently appeared in public domain
databases (Fig.
1
). Primer combinations were tested in PCR amplification experiments with total
cellular DNA preparations from a phylogenetically broad range of eukaryotes.
The characteristics of positive PCR amplification products are summarized in
Table
1
. In a number of cases (including all red algae tested and several early
diverging, amitochondriate eukaryotes), negative results were obtained with
cellular DNA samples: i.e. either no discrete amplification products were
obtained, or ones that were generated proved to be unrelated to T3/T7 RNAPs.
Aside from the possibility that the gene in question may actually be absent in
these negative cases, failure to generate a positive PCR product could result
from any number of factors, including sub-optimal concentration of the target sequence in selected DNA preparations
(high genomic complexity), spurious presence of competing sequences that
sequester PCR primers, presence of introns that interfere with PCR
amplification, and sequence divergence at primer binding sites.
An alignment of T3/T7-like RNAP sequences amplified from eight eukaryotes (seven of them
protists) is shown in Figure
2
. Also included are known bacteriophage (T7, K11, SP6) and fungal mitochondrial
(yeast,
Neurospora crassa
) RNAP sequences, as well as human, rice and
C.elegans
EST homologs. The alignment excludes several blocks (the sizes of which are
indicated in Fig.
2
) that display pronounced sequence and length variation. Within the remaining
alignable blocks (I-III), a high degree of amino acid sequence identity is evident, with a
number of positions universally conserved or almost so among the eukaryotic
sequences, and to a lesser extent between the eukaryotic and phage ones.
Despite the phylogenetic breadth of the source organisms, the extent and degree
of positional identity leaves little doubt that all of these sequences are
related by descent from a common ancestral sequence (i.e. are homologous).
The amplified region (encompassing T7 positions 637-813) comprises part of the palm and fingers domain of the recently
determined, hand-shaped T7 RNAP crystal structure (
16
). The most highly conserved portions of the alignment correspond mainly to
regions facing the template-binding cleft of the enzyme. All of the sequences shown in Figure
2
contain the invariant and catalytically essential aspartate residue equivalent
to D812 in T7 RNAP (
17
), as well as the adjacent and catalytically significant H811 (
17
). It should be noted that these two functionally important residues, as well as
the equivalents of T7 RNAP residues Y639 and G440, are also found in DNA
polymerases (
18
,
19
); however, overall sequence similarity clearly identifies the non-phage sequences in Figure
2
as RNAP, not DNAP, homologs.
Table 1
In several cases, PCR amplification generated positive products that were
substantially larger than expected (compare Fig.
1
and Table
1
). In these instances (
Acanthamoeba castellanii
,
Cryptomonas
[Phi] and
Tetrahymena pyriformis
), we infer the presence of small intron sequences in the PCR products (Fig.
3
and Table
2
). Exon/intron boundaries were assigned based on consideration of optimal amino
acid sequence alignment and maintenance of open reading frames, as well as the
assumption that introns begin with GT and end in AG. The deduced intron
sequences display distinctive base compositions relative to their flanking
exons, being particularly C+T-rich (61-69%) in
Acanthamoeba
and A+T-rich (81-85%) in
Tetrahymena
(Table
2
); the latter feature is characteristic of spliceosomal-type nuclear introns in this organism (
21
). Intron junction sequences also correspond closely to known consensus
sequences at intron splice sites in
Acanthamoeba
and
Tetrahymena
nuclear genes (Table
3
). A well-defined consensus sequence is also evident at
Cryptomonas
intron junctions; in this case, however, no published nuclear intron sequences
are available for comparison.
The T3/T7-like sequences reported here, which are also homologous at the amino acid
level to the yeast mtRNAP sequence, represent a phylogenetically broad sampling
within the eukaryotic lineage. In several cases (e.g.
Cryptomonas
, wheat), we have direct evidence that the PCR-amplified gene is expressed. In the remaining instances, additional work
will be required to verify that the gene we have identified is expressed, and
that it encodes a functional mitochondrial RNAP. Whether or not all of these
sequences ultimately prove to be functional, their widespread occurrence
throughout the eukaryotic lineage raises intriguing questions about their
evolutionary origin, and their present or former function.
Table 3
The species from which positive PCR products were obtained include members of a
number of major protist phyla, representing all three mitochondrial cristal
types (discoidal, tubular, flattened) (Table
1
). The organisms branch widely within a eukaryotic phylogenetic tree [see e.g.
Cavalier-Smith (
22
)]. Of particular interest is our finding of a mtRNAP homolog in the protozoon,
Naegleria fowleri
.
Naegleria
is thought to be one of the earliest diverging genera among mitochondria-containing eukaryotes
(
22
). It is also striking that sequences homologous to yeast mtRNAP could be
isolated from organisms displaying very diverse patterns of mitochondrial gene
organization and expression, such variability being a hallmark of mitochondrial
genomes (
23
). Allowing that the PCR-derived T3/T7-like sequences described here likely encode a portion of the mtRNAP
in these organisms, this would suggest that transcriptional mechanisms and the
transcriptional machinery itself (
24
) may have more features in common than the diversity of mitochondrial
transcriptional patterns might suggest.
Phylogenetic trees (not shown) constructed using the aligned amino acid
sequences in Figure
2
did not display robust branching patterns, a consequence of the limited
information content in the sequence data currently available. Not surprisingly,
the wheat and rice protein sequences are highly similar (142 identical residues
over 153 positions), and associate strongly in the trees. The same is true of
the
Neurospora
and yeast protein sequences, although branch lengths are much longer in this
case, reflecting a greater degree of sequence divergence. Other than these two
affiliations, the only other notable feature of this analysis is that, as a
group, the putative mtRNAP sequences are more similar to one another than to
the phage sequences, consistent with the idea that the mtRNAPs diverged from a
more recent common ancestor than they shared with the phage RNAPs. The fact
that none of the amplified sequences branches with the yeast-
Neurospora
clade makes fungal contamination of the DNA preparations used for PCR analysis
an unlikely possibility.
Because PCR products were generated using total cellular DNA or partially
purified nuclear DNA fractions as target, the genomic localization of each of
the amplified sequences remains to be definitively established. In the case of
A.castellanii
and
T.pyriformis
, no T3/T7-like RNAP sequences are present in the completely sequenced mitochondrial
genomes of these organisms (
25
; Burger,G., Zhu,Y., Littlejohn,T., Greenwood,S.J., Schnare,M.N. and Gray,M.W.,
in preparation), whereas the presence of small, splicesosomal-like intron sequences in the respective PCR products supports a nuclear
location for the mtRNAP homologs. Moreover, in translating the
Tetrahymena
exon sequences for the alignment, it was necessary to use the modified genetic
code (UAA and UAG decoded as Gln) that is known to be employed for nucleus-encoded mRNAs in this organism (
26
).
Introns in the
Cryptomonas
[Phi] PCR product also suggest that these RNAP sequences are localized in the
nuclear genome. However, a complicating factor in this case is the additional
presence of a nucleomorph genome that is the evolutionary remnant of the
nuclear genome of an endosymbiotic alga (
27
-
29
). In Southern hybridization experiments with total DNA resolved by pulse-field gel electrophoresis, the
Cryptomonas
PCR product specifically hybridized with nuclear DNA (data not shown). In the
case of wheat, the mtRNAP-homologous sequence could not be amplified from purified mtDNA; moreover,
no sequences homologous to T3/T7 RNAPs have been found in several completely
sequenced chloroplast DNAs, including those from rice (a monocotyledon closely
related to wheat) (
30
) and tobacco (a dicotyledon) (
31
). These observations indicate that the wheat and rice homologs are also nuclear
genes.
Fungal and plant mitochondria commonly contain linear plasmids encoding single-subunit RNAPs (
32
). These plasmid-encoded RNAP sequences form a distinct group that is only distantly
related to the clade of phage T3 and T7 RNAPs and to the nucleus-encoded yeast mtRNAP (
32
), as well as to the PCR-derived RNAP sequences listed in Figure
2
. In pairwise comparisons, the non-phage RNAP sequences listed in Figure
2
are clearly more closely related to one another than to either their phage- or plasmid-encoded RNAP homologs (data not shown). Moreover, as a consequence
of sequence divergence at the binding sites against which PCR primers were
constructed (Fig.
1
), we would not expect that the equivalent region in mitochondrial plasmid-encoded RNAP genes would be amplified under the conditions employed in the
current study (particularly with the R-8.1/R-3.2 primer combination used to recover all but the
N.fowleri
sequence; Table
1
). For these reasons, we are confident that that sequences shown in Figure
2
do not originate from linear plasmids of the type characterized in fungal and
plant mitochondria, the evolutionary origin of which is also obscure.
In photosynthetic organisms, there is evidence for the existence of a second
chloroplast RNAP activity, not encoded in chloroplast DNA (
33
-
35
). In spinach chloroplasts, RNAP activity has been associated with a 110 kDa
polypeptide that has some phage RNAP-like properties (
36
). Thus, it is possible that there are separate nuclear genes encoding distinct
mitochondrial and chloroplast enzymes, each homologous to T3/T7 RNAPs, and that
in certain cases, the gene for the chloroplast enzyme might be amplified
preferentially over the gene for the mitochondrial enzyme under our conditions.
In this regard, it may be significant that among the amplified sequences shown
in Figure
2
, the most divergent is the one from the prasinophyte
Pycnococcus
provasolii
, a primitive green alga. On the other hand, in no case was more than one T3/T7
RNAP-homologous PCR product obtained from any of the DNA samples analyzed. In
addition, a Southern hybridization experiment not only verified that the
Pycnococcus
amplification product comes from
Pycnococcus
DNA, but also indicated that the corresponding gene is single copy (data not
shown). If a chloroplast homolog does exist, it is formally possible that both
it and its mitochondrial counterpart are encoded by the same nuclear gene, with
the protein products being targeted to the respective organelles. Precedent
exists for targeting of the protein products of a single nuclear gene to
different subcellular compartments (
37
). Additional experimentation will be required to establish the genomic location
of each of the T3/T7-like RNAP genes identified in this study and the subcellular location and
function of their encoded protein products.
The observations reported here support the thesis that a T3/T7-like RNAP was recruited to act as a mtRNAP at an early stage in the
evolution of the mitochondrion. So far, however, we have few clues as to the
evolutionary origin of the gene encoding this enzyme. In addition to negative
results with amitochondriate eukaryotes, we were unable to amplify homologous
sequences from DNAs of eubacteria or archaebacteria. Moreover, a search of the
recently published complete genome sequences from
Haemophilus influenzae
(
38
) and
Mycoplasma genitalium
(
39
) failed to reveal any T3/T7-like RNAP sequences. Thus, it is not yet clear whether the gene for a
T3/T7-like enzyme was acquired from a eubacteria-like symbiont or was provided by the ancestor of the nucleus-containing eukaryotic host cell. Other questions that warrant
further investigation are whether mtRNAP genes of the phage T3/T7 type are
present in any mitochondrial genomes not yet characterized, and whether such an
enzyme is used as the mtRNAP in all eukaryotes. The results summarized here
provide a useful entry to the study of mtRNAP evolution and function in a broad
range of eukaryotes, particularly protists.
We are indebted to R. N. Band (Department of Zoology, Michigan State
University), E. Denovan-Wright (Department of Biology, Dalhousie University), P. Gros (Department
of Biochemistry, McGill University), B.F. Lang (Département de Biochimie, Université de Montréal), and K.M. Lonergan and D.F. Spencer (Department of Biochemistry, Dalhousie University) for
generous gifts of DNA and/or RNA; to M.A. Ragan (Institute for Marine Biosciences,
National Research Council, Halifax) and J. Edqvist (Department of Biochemistry,
Dalhousie University) for gifts of
Cryptomonas
[Phi] and
T. pyriformis
cells, respectively; to B.G. Lane (Department of Biochemistry, University of
Toronto) for provision of a wheat cDNA library; and to G.I. McFadden (School of
Botany, University of Melbourne) for a blot of
Cryptomonas
[Phi] chromosomes separated by pulse field gel electrophoresis. This work was
supported by grants from MRC Canada (to M.W.G.; MT-4124) and from NSERC Canada (to R.C.), and by an NSERC 1967 Science and
Engineering Scholarship (to N.C.). Salary and interactions support from the
Canadian Institute for Advanced Research are gratefully acknowledged by M.W.G.
and R.C., who are Fellows in the Program in Evolutionary Biology.
Organism
Abbreviation
Phylum
Type
Product (bp)
Cristal type
Introns
Triticum aestivum
Tri
Angiospermatophyta
land plant
500
flattened
?
b
Pycnococcus
provasolii
Pyc
Chlorophyta
green alga (prasinophyte)
692
flattened
0
Acanthamoeba castellanii
Aca
Rhizopoda
amoeboid protozoon
820
tubular
3
Isochrysis
sp. Tahiti
Iso
Haptophyta
haptomonad alga
524
tubular
0
Thraustochytrium aureum
Thr
Heterokonta
slime net
581
tubular
0
Cryptomonas
[Phi]
Cry
Cryptista
cryptomonad alga
824
flattened
6
Tetrahymena pyriformis
Tet
Ciliophora
ciliate protozoon
728
tubular
3
Naegleria fowleri
Nae
Percolozoa
amoeboid protozoon
411
discoidal
0
REFERENCES
Return