Scheffler Immo E. Mitochondria

Подождите немного. Документ загружается.

Post - transcriptional mechanisms as well as transcriptional regulation must

be considered in yeast when the regulation of respiratory capacity and assem-

bly of the oxidative phosphorylation system in yeast mitochondria is under

discussion. One can speculate whether such a multitude of mechanisms is

required for ﬁ ne - tuning, for establishing the overall range of the activities

involved, or for a rapid response to rapidly changing environmental conditions.

It will be important to keep these considerations in mind when similar ques-

tions are raised about various mammalian cell types in intact organisms.

4.2.2.3 Regulation of Nuclear Respiratory Genes in Mammalian Cells The

general considerations and the observations made with yeast have encouraged

hypotheses and experiments aimed at similar goals for mammalian cells: (a)

characterization of transcription factors and their targets on the relevant

promoters, which might explain the coordinate expression of the large number

of nuclear genes encoding mitochondrial proteins, and (b) the nuclear –

mitochondrial interactions, which are assumed to take place to make nuclear

gene expression match mitochondrial gene expression or vice versa. A genetic

approach is technically and logistically difﬁ cult, if not impossible, with diploid

mammalian cells, and no conditions are known in tissue culture which would

induce or repress the biogenesis of mitochondria to permit willful manipula-

tion of the experimental system. A priori one would not expect mammalian

cells in culture to be able to respond like yeast to changing carbon sources,

since they have become adapted in evolution to a very stable environment in

vivo and hence have no need regulate their mitochondria over a short term.

Experiments with Chinese hamster ﬁ broblasts have shown that a large fraction

> 50% of the energy needs of these cells is satisﬁ ed by glycolysis (115) ; and

even when glucose is deliberately limited in culture, there is no induction of

transcripts of complex II, for example (80) , and only a modest increase in res-

piration. On the other hand, it has become apparent that different tissues and

cell types may have widely differing energy needs and hence may have to

adjust their level of mitochondrial respiration and oxidative phosphorylation

to these needs as part of the process of differentiation. From the study of

mitochondrial diseases and the discovery of the differential sensitivity of

different tissues to mutations in mtDNA, neurons and muscle cells were

deduced to be the most active tissues, as already anticipated from their physi-

ological functions. Cardiac muscle is a speciﬁ c example of a tissue deriving

most of its ATP from respiration; and not surprisingly, cardiac myopathies are

prominent among the clinical manifestations associated with mutations in

mtDNA.

To repeat, one may not expect to ﬁ nd rapid, short - term ﬂ uctuations in

mitochondrial biogenesis in a given cell type, but one can anticipate that the

expression of a large number of genes encoding respiratory and other mito-

chondrial proteins may be expressed at a higher level in some cell types com-

pared to others, and thus the existence of common regulatory molecules may

be the simplest mechanism for a coordinate expression.

NUCLEAR GENES ENCODING MITOCHONDRIAL PROTEINS 91

92 BIOGENESIS OF MITOCHONDRIA

The number of mammalian genes investigated fully is still small. For practi-

cal reasons, cDNA clones are the ﬁ rst to be isolated; and the step to the full

genomic clone, or at least the promoter region, is nowadays made possible by

the availability of whole genome sequences. From the examples investigated,

one may tentatively conclude that there is no single universal cis - acting

sequence element with its associated transcriptional activator which is common

to all such genes. However, at least two factors have been characterized which

play key roles in the regulation of entire subsets of genes. Which genes belong

to the set, and why, is not yet clear.

Evans and Scarpulla (116, 117) were the ﬁ rst to deﬁ ne a factor, designated

as NRF - 1 (nuclear respiratory factor 1), from their analysis of the promoters

of the mammalian (rat) cytochrome c and cytochrome oxidase subunit Vb

(COXVb). Sequence analysis of the cytochrome c promoter revealed the pres-

ence of several consensus sites for well - known transcription factors such as

Sp1, CCAAT, and the cAMP response element binding protein CREB. In

addition, a new cis - acting element was deﬁ ned as a palindromic motif consist-

ing of tandemly repeated GC sequences over a single helical turn that interacts

with the NRF - 1. The case was strengthened when over the years several other

nuclear genes encoding mitochondrial proteins were found to have the

same element (COXVb, COXVIIaL, COXVIc, mtTFA, a ubiquinone - binding

protein, a mtRNA processing enzyme, aminolevulinate synthase) (118) ,

prompting Scarpulla and his colleagues to purify the human protein to near

homogeneity from cultured cells. Partial peptide sequences were used to syn-

thesize degenerate primers, which in turn led to the isolation of full - length

cDNA for NRF - 1 and the production of authenticated recombinant protein

in bacteria (119, 120) . NRF - 1 is a single peptide with a novel DNA - binding

domain (related to DNA - binding domains in two developmental regulatory

proteins found in Drosophila and the sea urchin) and a transactivating domain.

The ﬁ nding of an NRF - 1 site in the promoter for the mitochondrial transcrip-

tional activator (mtTFA) is of special interest. Since mtTFA is required for

both replication and transcription of mtDNA (see Section 4.4.1 ), one can begin

to build a case for a role of NRF - 1 in the nuclear control of mtDNA copy

number and gene expression. Aminolevulinate synthase is also not an arbi-

trary, irrelevant enzyme for this discussion. It is the ﬁ rst and rate - limiting

enzyme in the pathway of heme synthesis in the matrix. Although a role for

heme in the control of expression of mammalian cytochromes is not yet as

well - deﬁ ned as in yeast, the connection is tantalizing.

When several other promoters of respiratory genes (e.g., COXIV) either

failed to show an NRF - 1 - binding site or could not be footprinted with pure

NRF - 1, the search for additional recognition sites and factors by Virbasius and

Scarpulla succeeded in the identiﬁ cation of a second factor termed NRF - 2

(121, 122) . The core GGAA motif was somewhat familiar from the previous

characterization of the ETS - domain family of transcription factors, but when

the NRF - 2 site from COXIV was used in an afﬁ nity column, a novel DNA

binding activity could be puriﬁ ed (123) . The DNA - binding activity co - puriﬁ ed

with ﬁ ve distinct polypeptides ( α , β

, β

, γ

). The α subunit was subsequently

identiﬁ ed as the DNA - binding peptide, but it is found as a heteromeric complex

with one of the other peptides ( αβ

, αβ

, αγ

). These complexes differ in

their afﬁ nity for the NRF - 2 site and to tandem NRF - 2 sites, a characteristic

that had already been observed for the related ETS - domain transcription

factor GABP of the mouse. Further cloning and sequencing of all ﬁ ve human

NRF - 2 subunits established NRF - 2 as the homologue of the mouse GABP,

a widely expressed transcription factor ﬁ rst deﬁ ned by the study of viral

promoters. Consensus sequences for the NRFs are the following: NRF - 1: . . .

YGCGCA Y GCGC R . . . ; NRF - 2: . . . ACC GGAA GAG. . . .

Reminiscent of the Hap2/3/4 complex in yeast, NRF - 2 has DNA - binding

domains and transactivation domains on separate peptides. In addition, the

afﬁ nity of some of the heterodimers for a promoter can be further enhanced

by cooperative binding to tandem NRF - 2 motifs, and overall a mechanism is

created to modulate promoter strength for different genes.

It was also ﬁ rst found by Scarpulla and his co - workers that some genes

have both NRF - 1 and NRF - 2 motifs that contribute to the control of expres-

sion of these genes (124) . Another example was found in the promoter for

the iron – sulfur subunit of succinate dehydrogenase (and complex II) (125,

126) . In this promoter the two sites are within 200 nucleotides upstream from

the transcription start site, and mutagenesis of either site reduced the expres-

sion of a reporter gene from this promoter 5 - to 10 - fold. Both NRF - 1 and

NRF - 2 were shown to footprint the respective sites independently. Other

promoter elements were found further upstream (Sp1, CCAAT), but deleting

them had no signiﬁ cant effect on the expression of a reporter gene in ﬁ bro-

blasts, myoblasts, or differentiated myotubes. For a more detailed and expert

review the reader should consult (127, 128) . A highly reduced summary of

many studies in mammals is that environmental conditions inﬂ uence the

expression of PGC - 1 family coactivators (PGC - 1alpha, PGC - 1beta, and PRC),

which, in turn, target speciﬁ c transcription factors (NRF - 1, NRF - 2, and ERR

alpha) in the expression of respiratory genes (128a) . The notion of having a

limited, select number of transcription factors and promoter sequences exclu-

sively devoted to the transcription of respiratory genes and mitochondrial

proteins is apparently too simple. Several such genes have been found to lack

NRF - 1 or NRF - 2 sites altogether (based on sequence inspection), while other

genes unrelated to mitochondrial function contain such sites in their promoter

region.

Mammalian cytochrome oxidase (COX, complex IV) consists of 13 sub-

units, 10 of which are encoded in the nucleus. It has been intriguing that three

of the nuclear - encoded subunits (VIa, VIIa, and VIII) exist as multiple iso-

forms that are differentially expressed in different tissues. The L - form is

named after its identiﬁ cation in liver, and this form is also found in many

other tissues. The H - form is found predominantly in skeletal and cardiac

muscle (heart). A detailed analysis of the COXVIaH promoter region with a

reporter construct in transgenic mice as well as in tissue culture has veriﬁ ed

NUCLEAR GENES ENCODING MITOCHONDRIAL PROTEINS 93

94 BIOGENESIS OF MITOCHONDRIA

the highly speciﬁ c elevation of expression in heart and muscle, with some

activity in the brain but virtually none in other tissues. Expression of this

gene was suppressed in myoblasts, but dramatically induced when these myo-

blasts were induced to differentiate into myotubes. A 300 - bp promoter

segment was entirely sufﬁ cient to mimic the expression of this gene and its

dependence on muscle differentiation. No NRF - 1, NRF - 2, or other element

previously associated with respiratory genes (see above) was found in this

region. Instead, an MEF2 consensus sequence and an E - box element were

shown to be essential by mutational analysis (129) . The muscle enhancer

factor (MEF) comprises a family of four genes induced by myogenic HLH

proteins (myo - D, myogenin, Myf5, and MRF4), and MEF2 has been demon-

strated to be signiﬁ cantly involved in muscle differentiation (see references

130 and 131 for summaries and examples). Thus, the activation of the

COXVIaH gene during muscle cell differentiation is understandable from a

consideration of its promoter, but the promoter does not betray that the

protein functions in mitochondria.

A very similar analysis was made with the SDH2 promoter (iron – sulfur

subunit of complex II) (126) . Transcripts of this gene are also signiﬁ cantly

elevated in adult skeletal and cardiac muscle. In contrast to COXVIaH,

however, the expression of the SDH2 gene was not up - regulated upon myotube

formation, consistent with the absence of promoter elements responding to

myogenic transcription factors. If there is no general, coordinate induction of

respiratory chain peptides, one may wonder about the purpose of inducing

COXVIaH mRNA by at least an order of magnitude in myotubes compared

to myoblasts. It has been proposed from experiments with the homologous

subunit in yeast that this subunit is responsible for sensing ATP and phosphate

levels — that is, for the modulation of enzymatic activity of complex IV under

varying metabolic conditions (132) . For this factor to participate in the regula-

tion of cytochrome oxidase activity requires one of two possible mechanisms:

Either the subunit VIaH replaces the preexisting subunit VIaL directly, or

there is turnover of the entire complex, with newly made complex including

the muscle - speciﬁ c VIaH subunit.

Thus, at ﬁ rst sight, the well - controlled and already thoroughly investigated

rat or mouse myoblast/myotube system in tissue culture may serve as a good

model system in which to study mitochondrial biogenesis and the control of

respiratory gene expression. Much previous attention has been focused on

muscle speciﬁ c proteins such as creatine kinase and the various cytoskeletal

proteins of the sarcomere. On the other hand, the preliminary lessons are that

more subtle mechanisms for modulating mitochondrial respiration may be at

work, requiring only a limited number of changes in the composition of the

electron transport complexes. A second and very interesting preliminary con-

clusion is that myotube differentiation in culture is incomplete, in the sense

that maturation, stimulation, and contractile activity of a myoﬁ bril may be

required for further changes in respiratory gene expression and mitochondrial

capacity.

Two other characterizations of genes had once raised expectations about

ﬁ nding more universal promoter elements that might be unique or speciﬁ c for

mitochondrial proteins. The analysis of the promoters for the ATP/ADP trans-

locator and for the β subunit of the F1 - ATP synthase had identiﬁ ed so - called

OXBOX and REBOX elements that were postulated to respond to nuclear

factors present at variable levels in myoblasts, myotubes, and HeLa cells (133,

134) . A generalization of these conclusions awaits the characterization of more

genes in this category, but at this time one can state that many other relevant

genes already analyzed do not contain these particular elements as a recogniz-

able consensus sequence.

4.2.2.4 Co-evolution of Nuclear and Mitochondrial Genomes From the

above description it is clear that several dozen nuclear - coded proteins interact

either with the mitochondrial genome (replication, transcription, etc.) or

directly with proteins encoded by the mt genome. The mt genome suffers

sequence substitutions at a signiﬁ cantly faster rate than the nuclear genome.

If altered proteins result from mtDNA substitutions, compensatory changes

may have to occur in the nuclear genome to assure proper strong, functional

interactions between the gene products, in forming a complex of the electron

transport chain, for example. Thus the two genomes are expected to co - evolve,

and strong evidence for this concept has been deduced from studies with

interspeciﬁ c hybrids and cybrids of mammalian cells. The ﬁ rst experiments

addressing this issue were reported by Clayton et al. (135) from their analysis

of mtDNAs in human - mouse interspecies hybrids. Further pioneering studies

of Wallace and Eisenstadt (136) indicated quite convincingly that the nuclear

and mitochondrial genome must be compatible; that is, there is species speciﬁ c-

ity, and a human mtDNA cannot coexist for long in a rodent cell, or vice versa.

More speciﬁ cally, in the Schefﬂ er laboratory, it was found that nuclear muta-

tions in a hamster cell mutant causing a defect in complex I could be comple-

mented by a gene on a hamster or mouse X chromosome, but not by a human

X - linked gene (137) . This result was also interpreted in terms of the incompat-

ibility of heterologous gene products.

A very informative test of genome compatibility was recently reported by

Kenyon and Moraes (138) . By fusing enucleated cells from the hominoid apes

(chimpanzee, pigmy chimpanzee, gorilla, and orangutan) with ρ

(mtDNA -

less) human cells these authors created “ xenomitochondrial cybrids ” with ape

mtDNA and human nuclear DNA. Only the combinations of human nuclei

with mitochondria from the most closely related species (chipanzee, gorilla)

yielded cells capable of oxidative phosphorylation. Orangutan mitochondria

in such cells did not express a functional OXPHOS system. An even more

speciﬁ c example of incompatibility between species was reported by Yadava

et al. (139) . The MWFE subunit of complex I is highly conserved in mamma-

lian species, and yet a human protein fails completely to complement a deﬁ -

ciency in a hamster cell. A change of two amino acids (out of 70) in the human

protein restores the function of this protein in hamster cells.

NUCLEAR GENES ENCODING MITOCHONDRIAL PROTEINS 95

96 BIOGENESIS OF MITOCHONDRIA

4.3 REPLICATION AND MAINTENANCE OF

MITOCHONDRIAL DNA

4.3.1 DNA Replication in Mammalian Mitochondria

Transcription and replication of mtDNA are basic processes which are closely

related to both the function of the organelle and its biogenesis (75) . The

present section will deal with mtDNA replication, with an emphasis on animal

mitochondria. When mtDNAs were discovered almost 50 years ago and the

ﬁ rst estimates of their size became available, it also became clear that the rela-

tively small size of animal mtDNAs made them attractive candidates for study.

Expectations were not disappointed, and our understanding of DNA replica-

tion and transcription in animal mitochondria is relatively far advanced.

Insights from such studies will be presented ﬁ rst. An update on studies in other

organisms will summarize the progress, provide some contrasting or unique

characteristics, and emphasize some of the issues for which answers still have

to be sought.

Mitochondrial DNA replication can be investigated conveniently in tissue

culture, where intracellular mitochondrial proliferation is loosely coupled to

cellular division and multiplication. The total number of mtDNAs has to

double in each cell cycle. Replication intermediates can be found readily.

Before entering into the details of this process, one should be reminded that

the situation may be more complicated in an intact animal, and several aspects

of the process need to be discussed separately. It is necessary to distinguish

between the number of mitochondria per cell and the number of mtDNA

molecules per cell (estimated to be a few thousand). A priori it makes sense

for mitochondrial DNA replication to be controlled to maintain a roughly

constant number of mitochondria per cell, and thus there has to be coordina-

tion with DNA replication in the nucleus. However, during embryonic devel-

opment and differentiation in mammals the number of mitochondria per cell

may be altered, depending on the energy requirements of particular tissues. It

is possible that in adult tissue such as muscle and brain there is also turnover

of mitochondria and hence mtDNA replication which is not linked to nuclear

DNA replication and cell division. A particularly fascinating situation may

occur during oogenesis, where only a small subset of the total population of

mtDNA molecules may be involved in mtDNA replication to generate the

large number of mitochondria found in the mammalian egg. While there is as

yet no proof of this hypothesis, such a “ bottleneck ” has been invoked to

explain observations on mitochondrial inheritance and the phenotypic expres-

sion of mitochondrial mutations in heteroplasmic cells (140) . Alternatively, a

bottleneck can also be created if there is no replication of mtDNA during the

ﬁ rst 10 – 15 zygotic divisions following fertilization. A more complete discussion

of this topic will be found in Chapter 7 .

In cells in tissue culture, mtDNA replication seems to be loosely coupled to

nuclear DNA replication, with the precise mechanism of the control of copy

numbers yet to be elucidated. On the other hand, the process of replication

itself is quite well understood (141 – 143) . Mammalian mtDNAs, and probably

most vertebrate mtDNAs, have two origins of DNA replication, one responsi-

ble for each strand. The asymmetric distribution of nucleotides (G + C) in

mammalian mtDNA allows one to distinguish a “ heavy ” and “ light ” strand on

alkaline CsCl gradients, and hence one can also speak of O

as the origin of

H - strand synthesis and O

as the origin of L - strand synthesis. O

is located

within the region devoid of genes referred to as the displacement loop, or D -

loop region, and at other times as the control region. Initiation and elongation

of the H - strand is the ﬁ rst event, and it proceeds a considerable distance around

the circle; only after the second origin (O

) is displaced as a single - stranded

template will DNA replication of the other strand and in the opposite direction

begin. O

is therefore dominant, and O

is not an independent origin, since

DNA replication beginning at this origin without the prior activation of O

has

never been observed. It is signiﬁ cant that O

is located within a cluster of ﬁ ve

tRNA genes. When this region has been displaced and is found in single -

stranded form, it is likely that a secondary structure is formed which contrib-

utes to the functioning of the origin. Support for such a hypothesis comes from

the observation that in chickens the O

sequence itself (found in mammals) is

absent, but a similar cluster of ﬁ ve tRNAs is still found at this location (144) .

Strand elongation by a mitochondrial DNA polymerase (polymerase γ )

proceeds in the usual 5 ′ – 3 ′ direction. The nature of this polymerase and its

relationship to the various bacterial and eukaryotic DNA polymerases is an

interesting one (see reference 51 for a more explicit discussion). Its low abun-

dance has made biochemical studies a challenge, but the enzyme has now been

characterized from several organisms including humans (63, 64, 145 – 148) . Pol

γ prefers ribohomopolymer templates, and therefore it was initially thought to

represent a cellular reverse transcriptase resembling the reverse transcriptase

of tumor viruses. Later it was recognized that nuclear preparations had been

contaminated with mitochondria. It is antigenically completely different from

the viral enzymes, and it is further distinguished by its inability to use natural

RNAs as templates. Other distinguishing characteristics include stimulation

by salt, resistance to aphidocolin, inhibition by N - ethylmaleimide, and by

dideoxynucleotide triphosphates (51) .

The most highly puriﬁ ed preparations suggested that it consists of a holo-

enzyme of ∼ 125 – 140 kDa (POL

A) that has both catalytic and exonuclease

activity, and it is associated with a smaller subunit of 35 – 54 kDa (POL

B) that

functions as an essential processivity factor. When pol γ genes were cloned

(e.g., MIP1 from S. cerevisiae ), a 140 - kDa polypeptide was found to include

both polymerase and exonuclease domains with recognizable homology to

prokaryotic, A - type DNA polymerases — for example, E. coli DNA polymerase

I. The cloning of X. laevis and D. melanogaster pol γ genes conﬁ rmed the con-

clusion that the polymerization and 3 ′ to 5 ′ exonuclease functions are com-

bined in one polypeptide (143) . The reconstitution of a minimal mtDNA

replisome in vitro has recently been achieved in the laboratory of M.

REPLICATION AND MAINTENANCE OF MITOCHONDRIAL DNA 97

98 BIOGENESIS OF MITOCHONDRIA

Falkenberg (149a) . Three proteins were deﬁ ned to act in concert at the replica-

tion fork by a mechanism that resembles T4 or T7 replisomes: Polymerase γ ,

mtSSB (a single - strand binding protein), and TWINKLE, a mitochondrial

DNA helicase with 5 ′ to 3 ′ directionality. Single - stranded DNA molecules

could be generated at a rate of 180 bp/min, close to the rate reported for in

vivo DNA replication.

While there are no novel or unusual aspects of the mechanism of DNA

strand elongation in mitochondria, the same cannot be said about priming and

initiation of DNA replication. It became clear some time ago that DNA rep-

lication and transcription of mtDNA were directly and intimately linked. The

ﬁ rst indication was that the two known promoters for transcription are both

located in the control region. They are ∼ 150 bp apart in mouse and human

mtDNA and control transcription in opposite directions; hence they can be

referred to as the light - strand promoter (LSP) and the heavy - strand promoter

(HSP). Their precise identiﬁ cation became possible with the establishment of

a faithful in vitro system for transcription which will be described in further

detail in the section on transcription. Three key ﬁ ndings were made when

detailed characterizations were made of the major nascent DNA chains at the

earliest stages in DNA replications, and of all the transcripts from the LSP.

First, all newly initiated DNA chains had a 5 ′ end that corresponded to the

sequence ∼ 200 nt downstream from the LSP. Second, LSP transcripts were of

two types: (a) the expected long transcripts that would later be processed to

yield eight tRNAs and one mRNA and (b) short transcripts with 3 ′ termini

precisely or closely adjacent to the 5 ′ termini of the nascent DNA chains. It

was a very plausible jump to the conclusion that the RNA transcripts might

serve as primers for the start of DNA replication from the O

origin. A satisfy-

ing conﬁ rmation of this idea was provided by the third ﬁ nding: LSP transcripts

still covalently attached to the nascent DNA (in the mouse). Transcriptional

activation is therefore closely linked to, and a prerequisite for, leading - strand

DNA replication. A current model incorporating many ﬁ ndings is presented

in Figures 4.5 and 4.6 (143) .

Figure 4.5 shows the control region of the vertebrate mt genome between

the Phe - tRNA on the left and the Thr - tRNA on the right. Initiation of tran-

scription at the L - strand promoter generates an RNA that remains ﬁ rmly

associated with the DNA in the region deﬁ ned by the conserved sequence

blocks (CSBs), displacing the DNA strand to form a loop. Subsequent process-

ing on either side releases a long downstream transcript and a short upstream

fragment. The RNA segment remaining in the R - loop can then serve as primer

for DNA synthesis. The D - loop is formed when the extension of the newly

synthesized H - strand is arrested (see below). Initiation of mtDNA replication

in other vertebrates is likely to be quite similar, with some variations in the

size of the control region, the nature of the CSB elements (with CSB1 appar-

ently the most essential), and the spacing between them (143) .

Recent studies in vitro have provided additional support to this model

(150) . RNA of deﬁ ned length made in vitro was capable of annealing to

Figure 4.5 The D - loop region of mammalian mitochondrial DNA and the ﬂ anking

genes. HSP and LSP are the heavy - strand and light - strand promoters. A processed

transcript from the LSP serves as the primer for DNA replication from the O

origin.

(Modiﬁ ed after D. Clayton.)

cytochrome b

LSP

12s rRNA

Thr

Phe

Leu

5’

3’

HSP

Pro

CSBs

Figure 4.6 Detailed schematic representation of the generation of the RNA primer

for DNA replication. RNA polymerase and factor(s) generate an RNA transcript. This

transcript is modiﬁ ed by a mitochondrial RNA processing enzyme (MRP), leaving a

short RNA tightly annealed to the DNA and creating an RNA loop. Priming from this

RNA of DNA replication leads to the generation of the larger D - loop. (Modiﬁ ed after

D. Clayton.)

5’

R-loop

nascent H-strand (DNA)

D-loop

RNA pol

RNA primer

MRP

REPLICATION AND MAINTENANCE OF MITOCHONDRIAL DNA 99

100 BIOGENESIS OF MITOCHONDRIA

supercoiled mouse mtDNA under deﬁ ned conditions, creating an “ R - loop. ”

This R - loop had been previously observed in in vitro transcription experi-

ments with puriﬁ ed mitochondrial RNA polymerase activity, and it had been

shown to consist of a “ persistent ” RNA – DNA hybrid. It was also possible to

create the R - loop by in vitro transcription with SP6 polymerase from an SP6

promoter inserted into mouse mtDNA control region. Its formation was not

dependent on the presence of a particular RNA polymerase or additional

protein, but did require negative superhelical turns in the mtDNA.

The R - loop with the RNA – DNA hybrid must have unusual stability, since

it is stable to manipulation and puriﬁ cation without branch migration, and

even linear restriction fragments containing the R - loop are stable for consider-

able time at 37 ° C without loss of one of the polynucleotide strands. The RNA

in the R - loop is not base paired with the complementary DNA strand along

its entire length. From a detailed examination and comparison of RNase H

and RNaseT1 cleavage sites in the R - loop with those of a double - stranded

hybrid made from the same RNA and its cDNA, it was concluded that a

unique structure must be formed. A model of the R - loop has emerged in which

intramolecular RNA – RNA interactions stabilize and protect distinct regions

against single - strand - speciﬁ c nucleases (stem - loops?), while other segments

interact with unique sites on the DNA. An examination of the relevant DNA

sequence has revealed three conserved sequence blocks (CSB I, CSB II, CSB

III), and both CSB II and CSB III contain poly G tracts and are absolutely

required for R - loop formation. From the results of the probing with single -

strand - speciﬁ c nucleases and RNase H, the whole structure cannot be repre-

sented by conventional Watson – Crick basepairing schemes. The CSBs

sequences attracted attention as potential cis - acting regulatory regions as early

as 1981, because they were short, highly conserved regions within an otherwise

highly divergent region (141, 143) .

Endonucleolytic processing of the RNA in this hybrid structure is postu-

lated to create a primer which is positioned in the correct position for exten-

sion by deoxyribonucleotides with DNA polymerase. A candidate enzyme

performing this function was isolated by Clayton ’ s laboratory in 1987 (142,

151, 152) . It was shown to be a site - speciﬁ c endonuclease capable of cleaving

RNA transcripts from the D - loop region between CSB II and CSB III; and

most intriguingly, this RNase MRP (mitochondrial RNA processing) contains

an RNA component encoded by nuclear genes in mouse and humans. Base-

pairing interactions between this RNA and the substrate RNA for this nucle-

ase are therefore likely to guide substrate selection and cleavage site selection.

Undoubtedly, the secondary structure of the LSP transcript within the R - loop

is also a strong determinant of speciﬁ city and activity. Identical ribonucleopro-

teins have been isolated from several mammals, the frog Xenopus laevis , and

even yeast. Cloning of these genes as well as the genes encoding the RNA

component has conﬁ rmed a high degree of evolutionary conservation (143) .

As discussed in the review by Shadel and Clayton (143) , it was a challenge to

show that the MRP is indeed found in mitochondria, since a very similar activ-