|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
RIBOZYMES TARGETED TO THE MITOCHONDRIA
USING THE 5S RIBOSOMAL RNA
JENNIFER ANN BONGORNO
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
To my grandmother, Hazel Traster Miller, whose interest in genealogy sparked my
interest in genetics, and without whose mitochondria I would not be here
I would like to thank all the members of the Lewin lab; especially my mentor, Al
Lewin. Al was always there for me with suggestions and keeping me motivated. He and
the other members of the lab were like my second family; I would not have had an
enjoyable experience without them. Diana Levinson and Elizabeth Bongomo worked
with me on the fourth and third mouse transfections respectively. Joe Hartwich and Al
Lewin tested some of the ribozymes in vitro and cloned some of the constructs I used.
James Thomas also helped with cloning and was an invaluable lab manager. Verline
Justilien worked on a related project and was a productive person with whom to bounce
ideas back and forth. Lourdes Andino taught me how to use the new phosphorimager for
my SYBR Green-stained gels. Alan White was there through it all, like the older brother
I never had. Mary Ann Checkley was with me even longer than Alan, since we both
came to Florida from Ohio Wesleyan, although she did manage to graduate before me.
Jia Liu and Frederic Manfredsson were there when I needed a beer. Marina Gorbatyuk
was there when I just had to learn some Russian.
From other labs, Dr. John Guy allowed me to use his equipment and taught me how
to assay for ATP synthesis. Xioping Qi helped me section mouse eyes and, although
technically part of the Guy lab, was part of the Lewin lab family in my eyes. Dr. Adrian
Timmers skillfully injected my mouse eyes. Ian Elder taught me how to use the ERG
machine. Vince Chiodo packaged my ribozymes in AAV. Sue Moyer, Tom Rowe and
Nick Muzyczka gave continued support and valuable advice as my committee. The entire
Molecular Genetics and Microbiology department and the IDP program were a pleasure
to be a part of. I am sure I will miss everyone.
TABLE OF CONTENTS
A C K N O W L E D G M E N T S ................................................................................................. iv
LIST O F TA B LE S ...... .. .. .. ........................................ .. .. .... .............. viii
LIST OF FIGURES ......... ......................... ...... ........ ............ ix
LIST OF ABREVIATIONS .. ..................................................... ............... xi
ABSTRACT ........ .............. ............. ...... ...................... xiv
1 INTRODUCTION TO MITOCHONDRIA ................................... ............... 1
M itochondrial C om ponents ............................................................... ..................... 1
G en o m e ........................................................ 1
R replication ............................................ 5
Transcription and Translation.......................................... ........... ............... 7
Protein Im port........................................................ 9
R N A Im p ort ................................................................................ 1 1
M itochondrial F unction s................................................. ...................................... 19
C itric A cid C y cle ................................................................................. 19
Respiratory Chain .................. ........................... .. ...... ................. 20
A T P Sy nth esis ................................................... ................ 2 2
Reactive Oxygen Species ............................................................................. 25
M mitochondrial D diseases ...................................... .............................25
M itochondrial DN A M stations ........................................ ....................... 26
NARP and Leigh Syndrome................................................... ...................27
M itochondrial M anipulations .............................................. ........................... 31
Creating Heteroplasmy in Cells and Mice.........................................................31
M anipulating the M itochondrial Genome ................................. ............... 33
R N A Im port .................................................................................... 35
A ltering P rotein s........................................................................ ............... 35
R ib o z y m e s ..................................................................................................... 3 8
A deno-A associated V irus......... ........................... ................... ............... 39
2 R E S U L T S .......................................................................... 4 1
Ribozyme Design and In Vitro Analysis ....................................... ............... 41
M house Cell Culture Phenotypes ........................................ ........................... 47
Four Transfection Attem pts................................................... .. ................... .. 48
Analysis of Clonal Isolates ......... .. .............. ................... 53
Localization of Ribozymes to the Mitochondria ....................... .... ...............61
H um an Cell Culture Phenotypes ........................................ .......................... 65
H eLa Cells and a Toxic 5S Transcript ..................................... ............... ..65
N o Phenotype in 293 C ells ...................................................................... .. .... 67
Injection of M ice .................................................................. ................... ..................69
3 M ATERIALS AND M ETHOD S ........................................ ......................... 75
In Vitro R ibozym e A analysis ............................................... ............................ 75
C lo n in g ...............................................................................7 6
C e ll C u ltu re ........................................................................................................... 7 6
T ran section s ................................................................................................77
RNA Isolation ........................................................................................ ...... .... ......... 78
R T -P C R ..................................................................7 8
G row th in G alacto se ............................................................................................. 8 0
Cyanide R resistance A ssay .............................................................. 81
A TP Synthesis A ssay .....................................................82
M itochondrial RNA Isolation .............................................................. .............83
M ou se Subretinal Injections ................................................................................. 84
Electroretinograms .................... ......... .......... 85
Sectioning of Mouse Eyes ....... .........................................85
4 D ISC U S SIO N ............................................................................... 87
In Vitro Ribozyme Assays........................................ .......... 87
Studies on Total Transfected Mouse Cells ...... .............................................88
Cell Culture Studies on Mouse Clonal Isolates .................... ................. ..... 90
Measurements by RT-PCR.................. ............ ........90
Mitochondrial RNA Levels and Growth in Galactose .................................. 91
ATP Synthesis and Cyanide Resistance .......................................................92
Human Cell Transfection Studies .............................................. ................. 94
Transfection Toxicity .................................. .......................................94
5S-A TP8/6 Toxicity ......................... .... ....... ... ............ .. ............. 95
Localization of Ribozymes to the Mitochondria .......................... ............... 96
Injection of M ice ............... ................. ............... ........ ...... .............. 96
Other M means to an End ............................................................... .....97
L IST O F R E FE R E N C E S ............... ....... ............. .. ........................... ....................... 98
BIOGRAPHICAL SKETCH .......................... ............................................... 129
LIST OF TABLES
2-1. Kinetic parameters for mouse ATP6 ribozymes............................................ 45
2-2. Rates of succinate-dependent ATP synthesis in total selected cells........................49
3-1. RNA oligonucleotides used for in vitro ribozyme assays ......................................75
3-2. DNA oligonucleotides used for cloning mouse and human ATP6 ribozymes..........76
3-3. DNA oligonucleotides used for RT-PCR............................ .... ............... 79
LIST OF FIGURES
1-1. Mitochondrial genome comparison of different species ..........................................2
1-2. Genetic map of the human mitochondrial DNA .....................................................3
1-3 A T P synthesis pathw ay s................................................................ .....................23
2-1. Annotated mouse ATP6 mRNA sequence ............................ ..... ............42
2-2. Depiction of ribozymes annealed to their targets..................................................43
2-3. In vitro time course of ribozyme cleavage .......... .................. ..... ..........44
2-4. Lineweaver-Burke plots of multiple turnover reactions for ribozymes ................45
2-5. In vitro ribozyme cleavage of total mitochondrial RNA.......................................46
2 -6 5 S an d U F con stru cts ................................................................. .....................4 7
2-7. Total G418-selected cells grown in galactose with 0.4 ng/ml oligomycin ..............50
2-8. Total selected cell growth in 1 mM cyanide ................................. ..................... 52
2-9. Total selected cell grow th in galactose ........................................ .....................53
2-10. Ribozyme and reverse transcript levels in clonal isolates by RT-PCR .................54
2-11. ATP6 levels from the first set of colonies as measured by RT-PCR ....................56
2-12. Levels of four mitochondrial mRNAs as measured by RT-PCR ..........................56
2-13. Growth of colonies in glucose and galactose ............... ............ .....................57
2-14. Growth in galactose of all fourth transfection colonies................ ..... .......... 58
2-15. Growth of colonies in cyanide............................ ............. ............... 59
2-16. Relative rates of complex I- and complex II-dependent ATP synthesis .................61
2-17. Detection of ribozyme in the mitochondria of NIH3T3 cells..............................63
2-18. Detection of ribozyme in the mitochondria of 293 cells .............. ...................64
2-19. Absence of the 5S-ATP8/6 transcript in HeLa cells .......................................... 67
2-20. Growth of 293T cells for two days in 20 mM cyanide............... ............... 68
2-21. Growth of HeLa and 293 cells in 20 mM cyanide .................. ............... 69
2-22. Levels of ATP6 and COX2 in transiently transfected 293T cells........................70
2-23. Electroretinogram results from mice injected with AAV5-ATP6m252rz...............71
2-24. Electroretinograms of mouse K at 4 months post injection...............................72
2-25. Sections of retinas from mice J and K ............... ............ .............................. 73
LIST OF ABREVIATIONS
3' is the three prime end of nucleic acid.
5' is the five prime end of nucleic acid.
5S rRNA is the ribosomal RNA with a sedimentation constant of 5 svedberg units.
AAV is Adeno-associated virus.
AAV5 is Adeno-associated virus serotype 5.
Ala is alanine, an amino acid.
ADP is adenosine 5'-diphosphate
ANT1 is the adenine nucleotide translocater.
Arg is arginine, an amino acid.
Asp is aspartic acid, an amino acid.
ATP is adenosine 5'-triphosphate.
ATP6 is the mitochondrial gene for subunit 6 of ATP synthase.
ATP6hl 14rz is the ribozyme designed to cleave after nucleotide 114 of human ATP6.
ATP6h206rz is the ribozyme designed to cleave after nucleotide 206 of human ATP6.
ATP6m252rev is the reverse orientation of ATP6m252rz.
ATP6m252rz is the ribozyme designed to cleave after nucleotide 252 of mouse ATP6.
ATP6m69rev is the reverse orientation of ATP6m69rz.
ATP6m69rz is the ribozyme designed to cleave after nucleotide 69 of mouse ATP6.
ATP8/6 is the overlapping mouse ATP8 and ATP6 sequence attached to the 5S rRNA.
bp is base pairs.
BSA is bovine serum albumin.
CBA is chicken 3-actin enhancer plus the CMV promoter.
COX2h24rz is the ribozyme designed to cleave after nucleotide 24 of human COX2.
COX2 is the mitochondrial gene for subunit 2 of cytochrome c oxidase.
COX3 is the mitochondrial gene for subunit 3 of cytochrome c oxidase.
CMV is cytomegalovirus.
D-loop is the displacement loop, the mitochondrial major non-coding region.
DEPC is diethyl pyrocarbonate.
DMEM is Dulbecco's modified Eagle's media.
DNA is deoxyribonucleic acid.
DTT is dithiothreitol.
EDTA is ethylenediaminetetraacetic acid.
EGTA is ethylene glycol-tetraacetic acid.
ES cells are embryonic stem cells.
EtBr is ethidium bromide.
FAD is flavin-adenine dinucleotide.
FBS is fetal bovine serum.
FI is the rotating portion of ATP synthase.
Fo is the membrane-anchored portion of ATP synthase.
G418 is geneticin, a neomycin analog.
GFP is green fluorescent protein.
Gln is glutamine, an amino acid.
GPxl is glutathione peroxidase.
H-strand is the mitochondrial heavy strand.
Kb is kilobase pairs.
KCN is potassium cyanide.
Kcat is the reaction rate with saturating substrate.
HIV is the human immunodeficiency virus.
Km is the Michaelis-Menten affinity constant.
KRS is yeast cytoplasmic lysine amino-acyl tRNA synthetase.
L-strand is the mitochondrial light strand.
LHON is Leber's Hereditary Optic Neuropathy.
Lys is lysine, an amino acid.
LysRS is the human lysine amino-acyl tRNA synthetase.
m252rev is the reverse orientation of m252rz.
m252rz is the ribozyme designed to cleave after nucleotide 252 of mouse ATP6.
m252rzl is colony number 1 from m252rz cells from the first transfection.
m69rev is the reverse orientation of m69rz.
m69rz is the ribozyme designed to cleave after nucleotide 69 of mouse ATP6.
m69rzI is colony number I from m69rz cells from the fourth transfection.
MERRF is myoclonic epilepsy and ragged-red fiber disease.
Met is methionine, an amino acid.
MgC12 is magnesium chloride.
mRNA is messenger RNA.
MRP RNA is mitochondrial RNA processing RNA.
MSK is yeast mitochondrial lysine amino-acyl tRNA synthetase.
mtDNA is mitochondrial DNA.
mtTFA is the same as Tfam, the mitochondrial transcription factor alpha.
N-terminal is amino terminal.
NADH is reduced nicotinamide adenine dinucleotide.
NARP is Neuropathy, Ataxia and Retinitis Pigmentosa.
NCS is newborn calf serum.
ND4 is NADH dehydrogenase (complex I) subunit 4.
PBS is phosphate buffered saline.
PCR is polymerase chain reaction.
PMSF is phenylmethylsulfonyl fluoride.
PolgA is the mitochondrial DNA polymerase-y.
p cells are cells completely lacking mtDNA.
Rb is the retinoblastoma gene.
RG6 is the mutagen rhodamine-6-G.
RNA is ribonucleic acid.
RNase is a ribonuclease.
RNase MRP is the ribonuclease for mitochondrial RNA processing.
ROS are reactive oxygen species.
rRNA is ribosomal RNA.
RT is reverse transcription.
rz is a ribozyme.
SOD is superoxide dismutase.
TCA cycle is the tricarboxylic acid cycle or citric acid cycle.
Tfam is the same as mtTFA, the mitochondrial transcription factor alpha.
TIM is the translocase of the mitochondrial inner membrane.
tK1 is yeast tRNALyScuu.
tK2 is yeast tRNALySsuu.
tK3 is yeast tRNALySuuu.
TOM is the translocase of the mitochondrial outer membrane.
TR is a terminal repeat, from AAV.
tRNA is transfer RNA.
UF 11 is the plasmid with AAV-TRs, GFP and neomycin resistance.
Ucpl is an uncoupler protein.
Vmax is maximum velocity.
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
RIBOZYMES TARGETED TO THE MITOCHONDRIA
USING THE 5S RIBOSOMAL RNA
Jennifer Ann Bongorno
Chair: Alfred S. Lewin
Major Department: Molecular Genetics and Microbiology
The 5S ribosomal RNA has been shown to be naturally imported into mitochondria.
To reduce the expression of mitochondrial genes, ribozymes were targeted to the
mitochondria by attaching them to the 5S ribosomal RNA. Ribozymes were designed to
target mouse ATP6, human ATP6 and human COX2. Down regulation of ATP6 could
provide a model for the degenerative disease Neuropathy, Ataxia and Retinitis
Ribozymes were tested in vitro on short RNA targets. The active ribozyme
sequences were cloned at the 3' end of the 5S rRNA sequence, and these constructs were
transfected into mouse NIH3T3 cells or human 293 cells. Clonal isolates of the NIH3T3
cells expressing high levels of ribozyme were used for most of the mouse experiments,
whereas the human cells were analyzed during transient transfections. RT-PCR was used
to detect the presence of the ribozyme in RNA isolated from RNase-treated mitochondria.
Target RNA levels were measured by semi-quantitative RT-PCR of total cell RNA.
Phenotypic assays included growth in galactose, resistance to cyanide, and ATP synthesis
in permeabilized cells.
Two ribozymes against the mouse ATP6 sequence were shown to be active in vitro.
The 5S rRNA-ribozyme transcripts were detected in the mitochondria, but in very small
amounts. No decrease in target RNA levels was detected, and no phenotype was seen
that correlated with the presence of a ribozyme. A phenotype consistent with a
mitochondrial defect was seen in some of the mouse clonal isolates, but this was seen in
some of the control colonies as well as ribozyme colonies.
Since only about 1% of the 5S ribosomal RNA is normally imported, import of
ribozymes into the mitochondria by this method may not be sufficient to affect the
comparatively high mitochondrial RNA levels. Alternatively, these ribozymes may not
be active within mitochondria.
INTRODUCTION TO MITOCHONDRIA
Mitochondria are energy-producing organelles found in most eukaryotic cells.
They are thought to be the result of an endosymbiotic relationship, in which a
proteobacterium provided an energy benefit to a host cell (Margulis and Bermudes, 1985;
Gray et al., 1999). The mitochondria still retain a rudimentary genome with similarities
to an ca-proteobacterium (Yang et al., 1985).
Today's mitochondrial DNA (mtDNA) contains the genes for some of the RNAs
and less than 1% of the proteins used in the organelle. Therefore, not only is the
mitochondrial DNA replicated and inherited in a different manner from nuclear
chromosomes, but also over 1,000 proteins and some RNAs must be imported into the
The size of mitochondrial genomes ranges from less than six kilobase pairs (kb) in
the human malaria parasite, Plasmodiumfalciparum (Gray et al., 1999); to 16 kb in
humans (Anderson et al., 1981); to more than 490 kb in rice, Oryza sativa (Notsu et al.,
2002) (Fig. 1-1A). Mitochondrial genomes encode anywhere from three proteins in P.
falciparum (Gray et al., 1999); to 13 in humans (Anderson et al., 1981) (Fig. 1-2); to 62
in the flagellated protozoon, Reclinomonas americana (Lang et al., 1997) (Fig. 1-1B).
With such a large number of genes, little non-coding sequence, and a standard genetic
code, the mtDNA of R. americana is thought to be similar to the ancestral mitochondrial
genome (Gray et al., 1999). The bacteria whose genome is most like any mitochondrial
genome is Rickettsiaprowazekii, with a genome of 1,112 kb encoding 834 proteins
(Andersson et al., 1998). R. prowazekii, an obligate intracellular parasite, already has a
reduced genome as compared with its free-living bacterial relatives (Andersson et al.,
a-protobacteria A ccmA B C F B
Rickeltsia rpl I 10 I 14 18 19 20 27 31 32 34
Oryza yeWi, nadsio,ii( secY
Plasmodlum rpl2,5,6 16 dh3,4 s atA
o Homo rps i,2,47,8, 10,1, 12,13, 14, 19 rpoA-D
C. rmnhaurdi yejR yejU yejV cox? 11
Marchantia o S.pombe I aC rps alp3,4
0 C.eugameos ap I mrp ms ufA
S Jakoba 0 Chondrus n5 coxl )
Seanormonas 0 PhyoMhora a- cO2 RNaseP
ch s nad4L atp6 atp \
0 A~ c, '22 RNAs rtf2
O Pots0 Acanthamoeb 2 t Sc
Tetrahymenw Homo +2 RNA 2 3 4
+1 RNA 0.0e-e
Fig. 1-1. Mitochondrial genome comparison of different species. A) Genome size
including the bacterial genome of Rickettsia. For the larger genomes, red
represents known coding sequence; blue represents unidentified open reading
frames, introns and intergenic sequences. B) Gene content of select
mitochondrial genomes. Adapted (Gray et al., 1999).
Over 99% of the genes for mitochondrial proteins are found in the nucleus, and
sequences of the mitochondrial genes have also been found in nuclear chromosomes
either as pseudogenes or fragments (Jacobs et al., 1983; Farrelly and Butow, 1983;
Woischnik and Moraes, 2002). It appears that mitochondrial genes have moved to the
nucleus many times over the course of evolution and are still translocating. The size of
the metazoan mitochondrial genome (the common ancestor of all animals) seems to have
stabilized about 800 million years ago (Saccone et al., 2002), which may partly be due to
the acquisition of an altered genetic code. Sequences still seem to move to the nucleus,
but they do not become functional and are not lost from the mtDNA (Woischnik and
Moraes, 2002). Transfer of mitochondrial genes to the nucleus in flowering plants seems
to be an ongoing process (Palmer et al., 2000).
S, ,,.,',Human mtDNA "
I ..I.." --f -" \ :::
I[ '*.:" G|
'Kk \ p 1 'I 1i
Si" I ,
FG ma of th Human mtDNA I c re
DA outer e p ( ei
NO 20 "L" V 0 .. ,) I. I
divide and share mitochondrial genomes (Hayashi et al., 1994). When mitochondria are
',,^ T*. s- .. ,)/,, .* ', ,*..
Mitochondria can exist as individual organelles or as a reticulum; they can fuse and
divide and share mitochondrial genomes (Hayashi et al., 1994). When mitochondria are
distinct organelles, there are one to 20 mtDNA molecules per mitochondrion (King and
Attardi, 1989; Wiesner et al., 1992) and at least hundreds of mitochondria per animal cell
(Pica-Mattoccia and Attardi, 1971; Posakony et al., 1975; Posakony et al., 1977). In
total, there are many thousands of copies of mtDNA per cell. Normally, all of the copies
of mtDNA in one cell are identical, and such a cell is termed homoplasmic. If two
different types of mtDNA exist in the same cell, it is termed heteroplasmic (Hauswirth
and Laipis, 1982). In a situation with both mutant and wild type mtDNA, there is
inter-mitochondrial complementation such that all mitochondria within the cell still
function alike; either all are normal or all are have the mutant phenotype, because they
can fuse and share mitochondrial components (Nakada et al., 2001). A threshold level of
mutant genomes (usually greater than 60%) must be exceeded before the mutant
phenotype is seen (Zeviani and Antozzi, 1997; Hayashi et al., 1991).
In animals, mitochondria are inherited maternally, so a homoplasmic state usually
remains homoplasmic in later generations unless mutations occur. Occasionally,
mitochondria can be inherited from both parents (Schwartz and Vissing, 2003).
Normally, any mitochondria from the sperm that enter the egg are eliminated by a
ubiquitin-related method (Sutovsky et al., 1999; Sutovsky and Schatten, 2000). Sperm
only have about 100 mtDNA molecules (Hecht et al., 1984) as compared to oocytes,
which have up to 100,000 copies (Piko and Matsumoto, 1976; Michaels et al., 1982).
These abundant mitochondrial genomes are derived from a small subset of mtDNA
during oogenesis, creating a bottleneck that can drastically skew any existing
heteroplasmy from one generation to the next (Marchington et al., 1998).
There are two models of mtDNA replication (Clayton, 1982; Holt et al., 2000).
The original model is strand asymmetric with a different origin of replication for each
strand, which can be distinguished as heavy (H) and light (L) strands based on buoyant
density (Clayton, 1982). Replication starts at the L-strand transcriptional promoter and
uses the L-strand transcripts to prime the synthesis of H-strand DNA (Gillum and
Clayton, 1979; Chang and Clayton, 1985; Fish et al., 2004). This creates the
displacement or D-loop and includes the original single-stranded H-strand, the L-strand
template and the new H-strand of about 500 to 700 nucleotides. The whole D-loop
regulatory region is about 1 kb long (Attardi and Schatz, 1988). When RNase MRP
cleaves the L-strand transcript at specific locations between 75 and 165 nucleotides
downstream from the transcriptional initiation site, replication continues extending the
D-loop (Chang and Clayton, 1987b; Lee and Clayton, 1997). Replication proceeds using
the L-strand as a template. When the replication fork reaches the origin on the H-strand,
replication of the L-strand can begin (Hixson et al., 1986; Wong and Clayton, 1985).
Thus the new H-strand is complete well before the new L-strand is (Chang and Clayton,
1987a; Chang and Clayton, 1987b; Clayton, 1982). Finally, the two strands must be
decatenated, RNA primer gaps must be filled in and the DNA-binding proteins must be
added to create the proper tertiary structure (Lecrenier and Foury, 2000).
Recently, a second mechanism of mtDNA replication was described in mammals
and is thought to be the primary mechanism (Holt et al., 2000; Kajander et al., 2001;
Yang et al., 2002). Synthesis of the mtDNA can start anywhere in the genome, although
there seem to be preferences for a 4 to 6 kb region 3' of the D-loop (Bowmaker et al.,
2003). In birds, there is also a preference for initiation 3' of the D-loop region, but
sometimes synthesis also starts 5'of the D-loop region (Reyes et al., 2005). Replication
is bi-directional and strand-coupled with a leading and lagging strand in each direction,
until it reaches the D-loop region (Bowmaker et al., 2003). At this point, one replication
fork continues away from the D-loop, such that the H-strand is the leading strand, and the
L-strand is the lagging strand (Holt et al., 2000; Yang et al., 2002; Bowmaker et al.,
2003). During synthesis, the L-strand includes large stretches of RNA; some regions
contain more RNA than other regions (Yang et al., 2002). Once replication is complete,
most of the RNA is replaced by DNA, but ribonucleotides still exist randomly throughout
both strands: usually about every 500 nucleotides (Yang et al., 2002; Grossman et al.,
1973). Most of the partially single-stranded replication intermediates described as
evidence for strand asymmetric replication may actually have been artifacts caused by the
degradation of the RNA regions during preparations (Yang et al., 2002).
Replication of mtDNA is stochastic. Some genomes are replicated multiple times
and others may not be replicated at all (Clayton, 1982). Replication generally stops when
the number of genomes has doubled. In the case of cells with a depleted number of
mitochondrial genomes, replication more than doubles the number of genomes per cell
division until the normal number of genomes exists (King and Attardi, 1989). In HeLa
cells, an increase in mtDNA replication has been seen in late S and G2 phases of the cell
cycle (Pica-Mattoccia and Attardi, 1972). Segregation into daughter cells is also
arbitrary, such that daughters of heteroplasmic cells can be homoplasmic or
heteroplasmic; and because of random drift, future daughter cells will eventually reach
homoplasmy (Birky, Jr., 2001). Replication is not always random, however. Genomes
with a more efficient origin of replication or a large deletion may have a replicative
advantage. Intercellular selection is also possible, but due to inter-mitochondrial
complementation; only cells with high percentages of mutant genomes are likely to be
Transcription and Translation
The mammalian mitochondrial genome is transcribed as two genome-length
transcripts originating from the 5' end of the D-loop (Fernandez-Silva et al., 2003).
Additionally, a second H-strand transcript containing only the two ribosomal RNAs is
about 20 times more abundant than the genome-length transcript (Montoya et al., 1983;
Montoya et al., 1982). There are no introns and almost no non-coding sequences among
the 13 protein coding and 24 RNA genes in mammalian mtDNA (Montoya et al., 1981;
Ojala et al., 1981; Anderson et al., 1981; Ojala et al., 1980). This is not true of all
species; for example, the 366 kb mitochondrial genome ofArabidopsis thaliana contains
more than 80% non-coding sequence and still encodes only 32 proteins (Unseld et al.,
In mammals, excision of the tRNAs (fairly evenly interspersed among the
protein-coding genes) separates most of the mRNAs (Ojala et al., 1981; Montoya et al.,
1981; Ojala et al., 1980). A few processing events not mediated by tRNA processing,
include the separation of ATP6 from COX3 (Nardelli et al., 1994). The H-strand,
containing most of the mRNAs and 14 of the tRNAs, seems to be processed quickly,
since full-length transcripts cannot be detected (Attardi et al., 1990). In contrast,
full-length L-strand transcripts can be detected, suggesting processing after completion of
transcription. The L-strand contains one mRNA (ND6) and eight tRNAs (Attardi and
Schatz, 1988). It is known to be synthesized at a much higher rate (Attardi et al., 1990)
and has a much shorter half-life than the H-strand transcripts (Aloni and Attardi, 1971;
Cantatore and Attardi, 1980). Once the tRNAs are processed, the mRNAs are
polyadenylated with tails of only about 55 nucleotides: much shorter than nuclear mRNA
tails (Ojala and Attardi, 1974; Ojala et al., 1981; Montoya et al., 1981; Hirsch and
Penman, 1974). Ribosomal RNA transcripts have poly-A tails of only one to 10 residues
(Dubin et al., 1982).
Most mitochondria use an imported bacteriophage T7/T3-like RNA polymerase for
transcription (Tiranti et al., 1997). The primary transcription factor is Tfam (aka mtTFA),
which is able to wrap, bend and unwind DNA in vitro (Fisher et al., 1992; Fisher and
Clayton, 1988; Parisi and Clayton, 1991). Tfam binds both H- and L-strand promoters
with a low degree of sequence specificity (Parisi and Clayton, 1991). Tfam (along with
TFB1M or TFB2M) are used for initiation (Falkenberg et al., 2002; McCulloch et al.,
2002). Termination of transcription is mediated by mTERF (Daga et al., 1993;
Fernandez-Silva et al., 1997).
The mammalian mitochondrial genome contains only 22 tRNA genes that are
sufficient to recognize all codons because of the slightly altered genetic code used in the
mitochondria and the simplified codon-anticodon pairing system (Attardi and Schatz,
1988). The standard stop codon, UGA, codes for tryptophan in the mitochondria, and
there are two additional stop codons, AGA and AGG, that normally code for arginine.
There is also one additional methionine codon, AUA, which is an isoleucine codon in the
cytoplasm. For half of the codon families, the amino acid is specified by the first two
bases alone. The mammalian tRNAs that recognize these codons all contain an
unmodified uracil in the wobble position (Fox, 1987). In the cytosol, two tRNAs are
necessary for each of these codon families, and often inosine is used to recognize
adenine, cytosine and uracil (Topal and Fresco, 1976). There are two methionine codons,
which can both be used for translation initiation; and in certain species, isoleucine also
initiates translation (Fox, 1987). For termination, some human mitochondrial genes do
not contain stop codons at the DNA level, but they end in U or UA; such that when
poly-adenylated, the RNA has the UAA stop codon (Anderson et al., 1981). The
mitochondria of many other species also use non-standard genetic codes, some with
different sets of deviations from the standard.
With only 13 proteins expressed in the mitochondria, over 1,000 other proteins
must be imported. Proteins synthesized in the cytoplasm and destined for the
mitochondria contain import signals of two major varieties that have been reviewed
(Pfanner and Wiedemann, 2002). Proteins targeted to the matrix typically have an
amino-terminal (N-terminal) presequence of 25 to 50 amino acids that is then removed
once the protein has reached its destination. Some inner membrane proteins (mostly with
a single membrane-spanning domain) and intermembrane-space proteins also contain
presequences (Schnell and Hebert 2003). These presequences are not identical, but are
all positively charged and can form amphiphilic alpha helices (Roise et al., 1986; Roise,
1988; von Heijne, 1986). Not all proteins targeted to the mitochondria have N-terminal
targeting sequences. Many hydrophobic proteins headed for the mitochondrial inner
membrane have multiple internal import signals that are never removed. There are also
other types of proteins with special signals for the intermembrane space and outer
Protein import begins with passage through the translocase of the outer membrane
(TOM) and then proceeds through one of two inner membrane complexes (TIM23 or
TIM22)(Wiedemann et al., 2004). The TOM complex consists of at least seven proteins.
Tom20 recognizes cleavable presequences (Abe et al., 2000; Brix et al., 1997; Schleiff et
al., 1997), whereas Tom70 recognizes proteins with internal targeting signals
(Wiedemann et al., 2001; Brix et al., 1997; Sollner et al., 1990; Steger et al., 1990). Both
interact with Tom22, which is considered the gate; because without it the channel is
constitutively open (van Wilpe et al., 1999). Both types of proteins are then passed
through channels formed by Tom40, the core of the general insertion pore (Vestweber et
al., 1989; Hill et al., 1998). Precursor proteins are threaded through as linear chains, but
inner-membrane-destined proteins are folded; such that some internal region of the
protein goes in first (Wiedemann et al., 2001). No energy is required for transport
through the TOM complex (Pfanner and Wiedemann, 2002). Tom5 is one of three
smaller Toms, which is involved in passing precursor proteins from Tom22 to Tom40
(Dietmeier et al., 1997). Tom6 and Tom7 are involved in assembly of the TOM itself
(Dekker et al., 1998; Honlinger et al., 1996; Model et al., 2001).
ATP and an electrochemical gradient are required for import of proteins into the
mitochondrial matrix through the Tim23 complex (Schleyer et al., 1982; Martin et al.,
1991; Neupert et al., 1990). Although the majority of Tim23 is in the inner membrane,
its N-terminus is anchored in the outer membrane, which could keep it in close proximity
to the TOM complex (Donzeau et al., 2000). Tim23 and Timl7 form the core of the
Tim23 complex; both are essential for cell viability (Dekker et al., 1997; Ryan et al.,
1998). Once a protein has entered the matrix, its N-terminal pre-sequence is cleaved off
by the heterodimeric mitochondrial processing peptidase (Gavel and von Heijne, 1990).
Transport into the inner mitochondrial membrane via Tim22 still requires the membrane
potential, but does not need ATP (Wachter et al., 1992; Kovermann et al., 2002). Tim22
is also necessary for cell viability (Kovermann et al., 2002).
Most of the proteins imported into the mitochondria from nuclear-encoded genes
are not also used in the nucleus or the cytoplasm, although in many cases proteins with
the same function are needed in both compartments. Since the proteins targeted to the
mitochondria must contain a targeting signal sequence, they must be either encoded on a
separate gene or be transcribed or translated from alternative start sites from their nuclear
or cytoplasmic counterparts. For example, mitochondrial and nuclear versions of
alanyl-tRNA synthetase in Arabidopsis thaliana are transcribed and translated from
alternative initiation sites in the same gene (Mireau et al., 1996). Another example is the
tRNA-Lys amino-acyl tRNA synthetase gene in humans that codes for the cytoplasmic as
well as the mitochondrial protein (Tolkunova et al., 2000).
There are four types of nuclear encoded RNA that are imported into mitochondria:
tRNAs, the 5S ribosomal RNA, and the RNA components of RNase P and RNase MRP.
These are not all imported in all species, but in most cases they are either encoded in the
mitochondrial DNA or they are imported. The only RNAs that are always found in the
mitochondrial genome are the 12S and 16S ribosomal RNAs.
RNase P is a site-specific endoribonuclease involved in nuclear and mitochondrial
tRNA processing. The 340-nucleotide RNA component of RNase P confers the catalytic
activity required to process 5' tRNA ends, while the protein component plays a structural
role (Bartkiewicz et al., 1989; Guerrier-Takada et al., 1986). The RNase P RNA is
encoded in the mtDNA of some non-mammalian species (Entelis et al., 2001b) including
yeast (Hollingsworth and Martin, 1986). In mammals this RNA must be imported into
The import of RNase P into mitochondria in mammals was first proposed by
Doersen et al., when its activity was detected in HeLa cell mitochondria (Doersen et al.,
1985). The mitochondrial RNase P activity was found to be nuclease sensitive,
suggesting that an RNA component is imported (Doersen et al., 1985; Puranam and
Attardi, 2001). Looking directly at the presence of the RNase P RNA in the
mitochondrial fraction, about 33 to 175 molecules were found per HeLa cell, which
corresponds to 0.1 to 0.5% of the nuclear pool (Puranam and Attardi, 2001). This should
be enough RNase P for two or three molecules per transcriptionally active
mitochondrion, sufficient for all the tRNA processing required (Puranam and Attardi,
RNase MRP (mitochondrial RNA processing) was first isolated from the
mitochondrial fraction of mouse cells and shown to be a site-specific endoribonuclease
that can cleave the RNA primer involved in mammalian mitochondrial DNA heavy
strand replication (Chang and Clayton, 1987a). This activity is dependent on a region of
complementarity between the MRP and primer RNAs (Bennett and Clayton, 1990). The
MRP RNA is encoded in the nucleus by a single copy gene (Chang and Clayton, 1987a;
Chang and Clayton, 1989; Yuan et al., 1989; Gold et al., 1989). The protein components
of this enzyme are also encoded in the nucleus (Chang and Clayton, 1987b; Clayton,
1994; Kiss et al., 1992). Most of the MRP RNA is actually located in the nucleolus in
ribonucleoprotein particles distinct from RNase MRP and is involved in ribosomal RNA
processing (Chang and Clayton, 1987b; Gold et al., 1989; Kiss et al., 1992; Li et al.,
1994; Lygerou et al., 1996; Topper et al., 1992; Hashimoto and Steitz, 1983; Reddy et al.,
1983; Reimer et al., 1988). The amount of MRP RNA in the mitochondria is close to the
limits of detection. Puranam and Attardi (Puranam and Attardi, 2001) detected 6 to 15
molecules per HeLa cell, (which is equivalent to 0.02 to 0.05% of the total cell RNA).
Little is known about how much enzyme is actually needed, but only about 8 primer
cleavage events are required per minute in HeLa cells (Puranam and Attardi, 2001).
The mouse, human and yeast RNase MRPs are capable of cleaving an RNA primer
with either mouse or human sequences (Bennett and Clayton, 1990; Chang and Clayton,
1987b; Stohl and Clayton, 1992). RNase MRP contains many of the same protein
subunits as RNase P (Gold et al., 1989; Chamberlain et al., 1998) and their RNA
components also share sequence and structural similarities (Lindahl et al., 2000; Forster
and Altman, 1990). By transfecting mutant MRP RNA genes into mouse C2C12
myogenic cells it was shown that when nucleotides 118 through 175 were deleted, the
RNA was not imported, whereas the 5' and 3' regions were dispensable (Li et al., 1994).
The majority of the 5S rRNA is incorporated into cytoplasmic ribosomes, but in
humans about 1% is imported into the mitochondria (Entelis et al., 2001a). The 5S rRNA
is also known to be imported in other mammals. Its presence in the mitochondria was
first discovered in cows, rabbits and chickens (Yoshionari et al., 1994). Three small
ribosomal RNAs (5S, 9S and 12S) are imported in trypanosomatids (Mahapatra et al.,
1994). The 5S rRNA is not imported in yeast (Entelis et al., 2001a). In other species
such as land plants, some types of algae and the flagellate protist Reclinomonas
Americana, the 5S rRNA is transcribed from the mitochondrial genome and its sequence
differs from that of the nuclear and chloroplast 5S rRNA (Lang et al., 1996).
In plants the mitochondrial 5S rRNA has been shown to be incorporated into
mitochondrial ribosomes (Lang et al., 1996), whereas it is not thought to be part of
mitochondrial ribosomes in fungi or mammals (Curgy, 1985). Mammalian mitochondrial
ribosomes, while being more similar to bacterial ribosomes than cytoplasmic ribosomes,
differ from both other ribosome types in the number of subunits. The mitochondrial
ribosomes are missing the subunit (O'Brien, 2002) that has been shown to interact with
the 5S rRNA in cytoplasmic ribosomes (Home and Erdmann, 1972). Interestingly,
Entelis et al. quantitated the number of 5S rRNA molecules in the mitochondrial fraction
to be about 3.2 x 104 per cell (Entelis et al., 2001a), which is similar to the number of
mitochondrial ribosomes per cell (Ojala and Attardi, 1972; Posakony et al., 1977). Since
no function has been shown for the 5S rRNA in mammalian mitochondria, it is likely to
be an evolutionary remnant from a time when it was actually imported and incorporated
into mitochondrial ribosomes. This would predict the presence of an evolutionary
intermediate where the mitochondrial ribosomes still possessed the L5 subunit allowing
the incorporation of the 5S rRNA.
Import of the 5S rRNA requires ATP, the mitochondrial membrane potential, and
the protein import machinery in isolated human mitochondria (Entelis et al., 2001a). This
import is dependent on soluble factors similar to but distinct from those required for
tRNA import (Entelis et al., 2001a). The yeast 5S rRNA, which is not thought to be
imported naturally in yeast, was imported into human mitochondria when human extracts
were used, but not with yeast extracts (Entelis et al., 2001a). A 5S rRNA transgene
tagged with a point mutation was capable of producing an importable 5S rRNA transcript
(Magalhaes et al., 1998).
The import of tRNA into the mitochondria is highly variable between species. In
mammals, a sufficient set of tRNAs is contained in the mtDNA and no import is
necessary (Anderson et al., 1981). In trypanosomatids no tRNAs are present in the
mitochondrial (kinetoplast) genome, and some of the nuclear encoded tRNAs function in
both the cytosol and mitochondria (Hancock and Hajduk, 1990; Lye et al., 1993; Simpson
et al., 1989; Mottram et al., 1991). Another protist, Tetrahymenapyriformis imports 26
of its 36 mitochondrial tRNAs (Chiu et al., 1975; Chiu et al., 1974; Suyama, 1967).
Yeast import only one tRNA (Martin et al., 1979; Martin et al., 1977) in spite of the fact
that they already contain a sufficient number of tRNAs in the mtDNA (Entelis et al.,
2001b). Plant mitochondria have between 14 and 27 tRNAs in their mtDNA and
anywhere between two and 11 are imported (Chiu et al., 1975; Entelis et al., 2001b;
Kumar et al., 1996; Marechal-Drouard et al., 1988; Oda et al., 1992; Suyama, 1986;
Sugiyama et al., 2005). Various plants, fungi and protests known to import some or all of
their tRNAs are listed in Entelis et al. (Entelis et al., 2001b).
RNAs are imported by different mechanisms in different species. Yeast use the
protein import channels (Tarassov et al., 1995a) plus soluble factors to import their
tRNALyscuu. Specifically, two amino-acyl tRNA synthetases are required (Tarassov et
al., 1995b) and at least one other factor (Entelis et al., 2001a). Trypanosomatids have
receptors that are used for tRNA and rRNA import that are distinct from those for protein
import (Mahapatra et al., 1994; Nabholz et al., 1999; Bhattacharyya et al., 2000;
Bhattacharyya et al., 2003), and no additional soluble factors are required (Mahapatra et
al., 1994). Similarities between yeast and trypanosomatids include the requirement for
ATP (Mahapatra et al., 1994; Tarasov et al., 1993; Tarassov and Entelis, 1992), some sort
of protein receptor (Mahapatra et al., 1994; Tarassov et al., 1995a) and the mitochondrial
membrane potential (Yermovsky-Kammerer and Hajduk, 1999; Adhya et al., 1997;
Tarassov et al., 1995a).
The basic tRNA structure (Hauser and Schneider, 1995; Entelis et al., 1998) and
certain sequences seem to be required for import in both trypanosomatids and yeast. In
trypanosomatids the D-loop and 5' precursor sequences seem to be most important
(Adhya et al., 1997; Mahapatra et al., 1998; Hancock et al., 1992; Rubio et al., 2000;
Lima and Simpson, 1996), whereas in yeast the anticodon loop and the aminoacceptor
stem determine the selectivity of import, probably because the tRNA must be
aminoacylated before it is imported (Entelis et al., 1998; Kazakova et al., 1999).
In yeast between 2 and 5% of tRNALyScuu is imported into the mitochondria
(Entelis et al., 1996; Entelis et al., 1998), and it has been shown to be funtional
(Kolesnikova et al., 2000). In Tetrahymena thermophila 10 to 20% oftRNAG nUUG is
localized to the mitochondria (Rusconi and Cech, 1996). In trypanosomatids different
tRNAs are imported at different levels (Adhya et al., 1997). In most cases imported
tRNAs have also been shown to function in the cytoplasm (Rusconi and Cech, 1996;
Simpson et al., 1989; Martin et al., 1979).
Import of artificial tRNA has been shown in plants, protests, yeast and humans.
Small et al (Small et al., 1992) have shown the importability of a mutant tRNA carrying
extra nucleotides in the anticodon loop as an approach for altering mitochondrial gene
expression in potatoes. Plant tRNAs have also been imported into other plant
mitochondrion that do not normally import them (Schneider and Marechal-Drouard,
2000). Yeast, tRNAs have been imported into trypanosome mitochondria (Schneider and
Marechal-Drouard, 2000). The addition of extra nucleotides into the D-loop did not
prevent import in Tetrahymena thermophila (Rusconi and Cech, 1996). In Leishmania
tRNAs with various unspliced introns can be imported (Sbicego et al., 1998). Also, small
RNAs (17 bases or less) were non-specifically imported in Leishmania tarentolae (Rubio
et al., 2000).
In yeast imported exogenous tRNAs were shown to be functional (Kolesnikova et
al., 2000). Yeast have three lysine tRNAs: tK1 (tRNALyScuu), which functions in the
cytoplasm and the mitochondria; tK2 (tRNALySsuu), which is only cytoplasmic in yeast;
and tK3 (tRNALySuuu), which is present in the mitochondrial DNA (Martin et al., 1979).
Changing the anticodon of tK1 from CUU to CAU resulted in the aminoacylation of the
tRNA with methionine instead of lysine. This modified tRNA was imported into isolated
mitochondria and was capable of participating in mitochondrial translation (Kolesnikova
et al., 2000). In whole cells, another mutant tRNA, tK2AlaCUA was shown to be functional
by its ability to suppress a cox2 nonsense mutation (Kolesnikova et al., 2000). This
mutant tRNA contains the anticodon base C34 that has been shown to allow tK2 to be
imported (Entelis et al., 1998). It also has a mutation in an acceptor stem base pair
(G3:U70) that causes it to be aminoacylated by alanyl-tRNA synthetase (AlaRS) (Hou
and Schimmel, 1988). Some artificial tRNAs require aminoacylation for import while
others do not (Entelis et al., 1998). In both cases the importable tRNAs retained the
ability to bind the mitochondrial lysine amino-acyl tRNA synthetase (pre-MSK) (Entelis
et al., 1998). For those that required aminoacylation, the identity of the amino acid did
not seem to matter (Kolesnikova et al., 2000).
Although no tRNAs are imported naturally in mammalian cells (Lynch and Attardi,
1976; Amalric et al., 1978), certain exogenous tRNAs are capable of being imported in
cells. Yeast tK1 (tRNALyScuu), mutant tK2 (tRNALySuuu), tK3 (tRNALySuuu), and human
mitochondrially encoded tRNALySuuu can all be imported in isolated human mitochondria
(Entelis et al., 2001a). The mutant yeast tKlMetCAU was also imported into human
mitochondria and was able to participate in mitochondrial translation (Kolesnikova et al.,
2000). Importable versions of yeast tK2 and tK3 were able to partially rescue the human
A8344G MERRF mutation that affects tRNALys (Kolesnikova et al., 2004). This was
shown in two different cybrid cell lines homoplasmic for the MERRF mutation. The
yeast MSK1 gene was transfected along with tK2 or tK3, and a correlation was found
between the amount of tRNA imported and the increased respiration rate (Kolesnikova et
The import conditions and sequence requirements were shown to be the same for
human as for yeast including the human or yeast mitochondrially targeted amino-acyl
tRNA synthetases (pre-LysRSmt and pre-MSKlp) (Kolesnikova et al., 2000; Entelis et
al., 2001a). The tRNA was imported more efficiently if it was already aminoacylated or
if additional yeast or human aminoacyl tRNA synthetases were added (yeast cytoplasmic
= KRS; human cytoplasmic = aaRS), suggesting again that aminoacylation is important
for import (Entelis et al., 2001a). Yeast have distinct amino-acyl tRNA synthetase genes
for cytoplasmic and mitochondrial tRNA-Lys. Humans have only one, which codes for
the cytoplasmic as well as the mitochondrial protein (Tolkunova et al., 2000). Sequence
alignments show that the N-terminal region of LysRS (human) is more similar to
pre-MSK (yeast mitochondrial), and the C-terminal domain is more similar to KRS (yeast
cytoplasmic), which would indicate that the N-terminal region is involved in import
(Entelis et al., 2001a). Interestingly, the N-terminal domain is not required for
aminoacylation (Shiba et al., 1997).
All of the proteins synthesized in the mitochondria are involved in the generation of
ATP. The mitochondria can generate up to 95% of the energy in a cell, and they do so
mainly by oxidative phosphorylation. Glycolysis takes place in the cytoplasm, but its
final product, pyruvate used in the mitochondria for the citric acid cycle. Electrons from
citric acid cycle are transported down the respiratory chain to pump protons into the
intermembrane space. This electrochemical gradient is used to drive the production of
ATP through ATP synthase. Other functions of the mitochondria include apoptosis,
calcium regulation, heat generation, fatty acid oxidation, waste elimination, and amino
acid and heme biosynthesis. Mitochondria are also involved in scavenging the free
radicals often generated by the respiratory chain and are influential in the aging process.
Citric Acid Cycle
Pyruvate dehydrogenase converts pyruvate from the cytoplasm to acetyl-coenzyme
A and the reduced form of nicotinamide adenine dinucleotide (NADH). This is the first
of four reactions that take place in the mitochondria to generate NADH.
Acetyl-coenzyme A enters the citric acid cycle, also know as the Krebs cycle or
tricarboxylic acid cycle, where the other three reactions occur. Isocitrate, ca-ketoglutarate
and malate are the other substrates directly involved in the generation of NADH. NADH
is used to carry electrons to the respiratory chain. Additionally, succinate in the citric
acid cycle is also a substrate for the reduction of flavin-adenine dinucleotide (FAD) to
FADH2 by succinate dehydrogenase. This enzyme, also known as complex II of the
respiratory chain, passes the electrons from FADH2 to ubiquinone in the electron
The electron transport chain consists of three protein complexes that pass electrons
from NADH to more than 15 different electron carriers each with an increased affinity for
electrons (Fig. 1-3A). The energy generated from passing electrons to carriers with a
higher affinity for electrons is used to pump protons from the mitochondrial matrix to the
intermembrane space generating an electrochemical gradient (AY). The final electron
acceptor is oxygen, so NADH is oxidized to generate, NAD+, water and ATP. The
complexes involved in this process are NADH dehydrogenase, cytochrome b-c1 complex
and cytochrome oxidase. Through an alternative pathway an additional complex,
succinate dehydrogenase accepts electrons from FADH2 and bypasses NADH
dehydrogenase in the transport of electrons. Finally, the energy from the proton gradient
is used by ATP synthase to drive the formation of ATP from ADP and inorganic
Nicotinamide adenine dinucleotide:ubiquinone oxidoreductase or NADH
dehydrogenase, also known as complex I, is the first and largest of the complexes in the
respiratory chain. This complex of at least 46 subunits contains seven or eight iron-sulfur
centers and a flavin that accept and pass electrons releasing the energy to pump protons
into the intermembrane space (Hirst et al., 2003; Skehel et al., 1991; Walker, 1992;
Skehel et al., 1998; Brandt, 1997). The complex is L-shaped, composed of a matrix arm
and a membrane arm separated by a thin collar (Grigorieff, 1998; Guenebaut et al., 1998).
The matrix domain contains all but one of the electron acceptors as well as the binding
site for NADH (Finel et al., 1992). Seven of its 14 core subunits are encoded in the
metazoan mitochondrial genome (Hirst et al., 2003; Anderson et al., 1981). NADH
dehydrogenase can be inhibited with rotenone (Davey and Clark, 1996).
Succinate dehydrogenase, also known as complex II of the respiratory chain, is also
one of the enzymes involved in the citric acid cycle. It converts FAD to FADH2 and
passes two electrons to the electron transport chain. This complex contains 3 iron-sulfur
centers and one flavoprotein covalently bound to FAD (Robinson et al., 1994), but here
the transfer of electrons is not coupled to the pumping of protons (Hagerhall, 1997). All
four the subunits of succinate dehydrogenase are encoded in the nucleus in metazoans
(Anderson et al., 1981).
Both NADH dehydrogenase and succinate dehydrogenase pass their electrons to
ubiquinone, also called coenzyme Q (CoQ). Ubiquinone is a small hydrophobic
molecule that can reside in the lipid bilayer, and pass electrons to complex III. It consists
of a benzoquinone headgroup and a polyisoprenyl tail (Nohl et al., 2003; von Jagow et
al., 1986). It can carry up to two electrons at the same time, now called ubisemiquinone
(CoQH) or ubiquinol (CoQH2), and pass them to complex III.
Cytochrome b-c1 complex, ubiquinol:cytochrome c oxidoreductase, or complex III
contains eleven (Wallace, 2001) different proteins and acts as a dimer. It contains one
iron-sulfur center and three hemes with iron atoms that carry electrons (van Loon et al.,
1983). Cytochrome b contains two of the hemes, and cytochrome ci contains the third.
Two protons are pumped for each electron that passes through this complex (Saraste,
1999). Cytochrome b-c1 complex passes its electrons to cytochrome c, a water-soluble
hemoprotein in the intermembrane space. Cytochrome c transports the electrons to
cytochrome oxidase. Its crystal structure helped determine its proton pumping activity
(Iwata, 1998; Xia et al., 1997; Zhang et al., 1998). Only one of its subunits is encoded in
mammalian mitochondria DNA (Anderson et al., 1981). Complex III can be inhibited by
myxothiazol or antimycin A (Villani and Attardi, 1997; Taylor et al., 1994).
Cytochrome oxidase or complex IV is composed of 13 different proteins
(Yoshikawa et al., 1998; Tsukihara et al., 1996) and also functions as a dimer. The three
major subunits are encoded in the mitochondrial DNA (Anderson et al., 1981). Subunit I
contains one copper atom and two cytochromes each with one heme. Subunit II contains
a copper center, which acts as the first electron acceptor (Malmstrom and Aasa, 1993;
Lappalainen et al., 1993). The copper atom in subunit I works together with the iron
atom from one of the hemes to pass electrons to the final acceptor, oxygen. Oxygen
needs a total of four electrons to produce two water molecules. Since copper and iron
can only carry one electron each, oxygen stays bound to the bimetallic center after
receiving the first two electrons. Once a third electron is passed, one water molecule can
be released. The fourth electron completes the cycle (Babcock and Wikstrom, 1992). Its
mechanisms of proton pumping have been partially described through crystallographic
studies (Yoshikawa et al., 1998; Ostermeier et al., 1997; Tsukihara et al., 1996; Tsukihara
et al., 1995; Iwata et al., 1995). The activity of cytochrome oxidase can be inhibited with
cyanide (Letellier et al., 1994).
ATP synthase, F1Fo ATPase or complex V contains 15 subunits in mammals
(Devenish et al., 2000; Lutter et al., 1993), some of which are present in more than one
copy per complex. It is composed of an anchored membrane portion, Fo of up to twelve
subunits depending on the species (Boyer, 1997; Devenish et al., 2000), and a rotating
portion, Fi of five different subunits (Karrasch and Walker, 1999; Wilkens and Capaldi,
1998; von Meyenburg et al., 1982) (Fig. 1-3B). The two mitochondrially encoded
subunits, ATP6 and ATP8, are part of the Fo portion.
Mitochondrial inner membrane B
synthase; adapted (Schon et al., 2001).
N ADH 010.
Fig01. 1 s ATP synthesis pathways. A) Tricarboxylic acid (TCA) cycle and oxidative
phosphorylation; adapted (Maechler and Wollheim, 2001). B) Bacterial ATP
synthase; adapted (Schon et al., 2001).
FjFo-ATPase couples the flow of protons down the electrochemical gradient to the
production of ATP from ADP and inorganic phosphate (Pi) (Elston et al., 1998; Noji and
Yoshida, 2001). Protons flow through a channel in the ATP6 subunit of Fo and cause the
neighboring ring of ATP9 subunits to rotate (Rastogi and Girvin, 1999; Hutcheon et al.,
2001). This rotates the F, portion, which catalyzes the phosphorylation of ADP (Duncan
et al., 1995; Fillingame, 1997; Duncan et al., 1995; Noji et al., 1997). The "stator"
portion of the structure consists primarily of subunit 6, two copies of subunit b that create
a stalk and 6 of Fi; the "rotator" consists of the 12 copies of subunit 9 and the rest of the
FI portion, U3P3ys (Elston et al., 1998).
The Fo portion in Escherichia coli is composed of subunits a (analogous to the
mitochondrial subunit 6), b (two copies per complex), and c (12 copies per complex,
analogous to mitochondrial subunit 9) (Jones and Fillingame, 1998). All are required for
proton translocation (Schneider and Altendorf, 1985). Subunit a, a very hydrophobic
(probably five-transmembrane helix) protein, has been shown to interact closely with
subunit c suggesting a joint catalytic role in proton translocation (Jiang and Fillingame,
1998). Specifically, Arg210 (mitochondrial Argl59) in helix-4 of subunit a interacts
with Asp61 (mitochondrial Glu58) in subunit c (Jiang and Fillingame, 1998; Cox et al.,
1986; Cain and Simoni, 1989; Cain and Simoni, 1988). The antibiotic oligomycin
inhibits ATP synthase through interaction with the Fo portion (Breen et al., 1986; John
and Nagley, 1986).
The F1 sector is the catalytic component (Boyer et al., 1975). It is composed of five
subunits (Ua33y6E). Although the a and 3 subunits are similar, only 3 has catalytic
activity. Energy is required for binding of ADP and release of ATP, but not for the
synthesis of ATP (Boyer, 1993; Boyer, 1997). The Fi portion not only has ATP synthesis
activity when coupled with the Fo sector, but it has ATP hydrolysis activity on its own
whether tethered to a membrane or free-floating (Jiang and Fillingame, 1998). Its crystal
structure (Abrahams et al., 1994) confirmed the idea that the a and 3 subunits rotate
Reactive Oxygen Species
Mitochondria are the primary endogenous source of reactive oxygen species
(Harman, 1972). When the electron transport chain is inhibited, electrons accumulate at
complex I and CoQ and can easily be transferred to oxygen to create superoxide (02).
02' can be converted to hydrogen peroxide (H202) by magnesium superoxide dismutase
(MnSOD) in the mitochondria or copper/zinc superoxide dismutase (Cu/ZuSOD) in the
cytosol. H202 can be converted to H20 by glutathione peroxidase (GPxl) in the
mitochondria or the cytoplasm and by catalase in the cytoplasm. Reactive oxygen
species such as 02-, H202 and the hydroxyl radical (OH), formed from the interaction of
H202 with Fe2 can damage DNA, proteins and lipid membranes (Snyder, 1990). ROS
are an important source of mitochondrial DNA mutations, up to ten times more than
nuclear DNA mutations (Richter, 1988; Ames et al., 1993). They can also further inhibit
ATP synthesis by inactivating iron-sulfur electron carriers in the respiratory chain and
aconitase in the citric acid cycle (Wallace, 1997; Brazzolotto et al., 1999; Hentze, 1996).
Mitochondria are involved in many degenerative diseases, aging and cancer. Each
mitochondrial disease occurs in only about 1 in 10,000 people (Larsson and Clayton,
1995; Wallace, 2001; Man et al., 2003), which can be considered rare, but when all
mitochondiral diseases are considered together, their numbers are much more impressive
(Schaefer et al., 2004). Most of these are due to defects in nuclear genes associated with
mitochondrial functions (Orth and Schapira, 2001; Triepels et al., 2001) and some are due
to mutations in the mitochondrial DNA itself (Holt et al., 1988; Wallace et al., 1988;
Shoffner et al., 1990; Goto et al., 1990). In some cases, mutations in nuclear genes lead
to mtDNA damage (Suomalainen and Kaukonen, 2001). They can be maternally
inherited or sporadic meaning that the mutations can be inherited or can accumulate over
time (DiMauro and Schon, 2001).
Mitochondrial DNA Mutations
Over 50 point mutations and hundreds of rearrangements in mitochondrial DNA
have been associated with disease (Kogelnik et al., 1998). Some mtDNA mutations
involve tRNA (Enriquez et al., 1995; Flierl et al., 1997; Hao and Moraes, 1996) or rRNA
genes (Entelis et al., 2001b; Moraes et al., 1991) and thus affect all mitochondrially
encoded proteins. About one-third of point mutations affect only a single protein-coding
gene (Manfredi et al., 2002). Mutations that affect all mitochondrially synthesized
proteins are usually heteroplasmic, such that there are enough good copies of mtDNA to
provide some function. However, some tRNA point mutations have a milder effect and
can be homoplasmic (Wallace, 1999). Likewise, there are some point mutations in
protein-coding genes that are only found in the heteroplasmic state implying lethality if
homoplasmic and other point mutations that do not result in a disease phenotype unless
homoplasmic. Mutations associated with Leber's Hereditary Optic Neuropathy do not
always result in disease even when homoplasmic suggesting the presence of modifier
genes or a role for environmental factors (Carelli et al., 2003). In all cases, at least 60%
of mitochondrial genomes must have the mutation in order for the disease to manifest
itself (Hayashi et al., 1991; Zeviani and Antozzi, 1997; Moraes et al., 1993; Tatuch et al.,
1992; Larsson et al., 1995).
Since all of the genes in the mitochondrial DNA are involved in energy production,
tissues having high energy requirements are likely to be affected. Mitochondrial
diseases commonly affect the brain, central nervous system, skeletal muscle, eye, heart,
and renal and endocrine systems (Wallace, 1999). Diseases associated with mutations in
mitochondrial DNA typically have variable onset and symptoms even within the same
family. This may be due to the amount of the mutant DNA relative to wild type within a
given tissue or individual cells (Schon et al., 1997; Petruzzella et al., 1994; Sciacco et al.,
1994). Additionally, most mitochondrial diseases are progressive, which is probably due
to the accumulation of mtDNA mutations over time (Zhang et al., 1992; Linnane et al.,
1990; Corral-Debrinski et al., 1991; Cortopassi and Arnheim, 1990; Simonetti et al.,
1992; Cooper et al., 1992).
Leber's Hereditary Optic Neuropathy (LHON), the first disease associated with a
mutation in mtDNA (Wallace et al., 1988), has a very noticeable variation in penetrance.
It is caused by point mutations in genes for subunits ofNADH dehydrogenase. It causes
sudden midlife blindness associated with degeneration of the optic nerve (Carelli et al.,
2002; Kerrison, 2005). The most common mutations are usually homoplasmic including
ND4:G11778A, ND1:G3460A, and ND6:T14484C (Carelli et al., 2003). LHON is more
prevalent in males within a homoplasmic family (Carelli et al., 2002; Qu et al., 2005;
Howell and Mackey, 1998). An X-linked recessive locus may contribute to LHON (Bu
and Rotter, 1991). This could explain the general variability in penetrance as well as the
increased incidence in males, but no gene has yet been identified (Carelli et al., 2003).
Environmental factors including tobacco and alcohol have been shown to increase the
penetrance of LHON, which may also contribute to the increased occurrence in males, if
they are more likely to smoke and drink (Carelli et al., 2002).
NARP and Leigh Syndrome
The original goal of this dissertation was to create a mouse model of a disease
caused by mutations in the mitochondrial gene ATP6 that codes one of the subunits of
ATP synthase. Mutations in this gene can lead to two different diseases, Neuropathy,
Ataxia, and Retinitis Pigmentosa (NARP) or Leigh Syndrome. The percent of the
mitochondrial DNA that carries the mutation roughly determines whether the individual
will be healthy or develop one of the two diseases (Holt et al., 1990; Santorelli et al.,
1993; Tatuch et al., 1992). NARP is associated with 70 to 90% mutant mtDNA (White et
al., 1999), while Leigh Syndrome is associated with 90 to 95% mutant mtDNA (Shoffner
and Wallace, 1992; Ortiz et al., 1993; Tatuch et al., 1992).
NARP affects the brain, central nervous system, peripheral nervous system and
eye. Symptoms include developmental delay, dementia, seizures, ataxia, corticospinal
tract atrophy, muscle weakness, sensory neuropathy, salt and pepper retinopathy, retinitis
pigmentosa, sluggish pupils, lazy eye and blindness (Hamosh et al., 2002). Onset often
occurs in childhood and usually coincides with an infection. Symptoms at onset are
highly variable and tend to be progressive (Sciacco et al., 2003). Affected individuals
tend to live into their thirties (White et al., 1999). In mild cases, symptoms may not
appear until later in life. Leigh Syndrome is characterized clinically by ataxia, hypotonia,
spasticity, developmental delay, optic atrophy, ophthalmoplegia (paralysis of the
extra-ocular eye muscles), degeneration of the basal ganglia and vascular proliferation
(Shoffner et al., 1995). Onset is during infancy, not at birth, and affected individuals
usually die within a few years.
A thymine to guanine transversion at nucleotide 8993 of the mitochondrial genome
results in an arginine instead of a leucine at amino acid 156 of the ATP6 protein. This
T8993G mutation has been associated with both NARP and Leigh syndrome depending
on state of heteroplasmy, and is the most common mutation associated with NARP (Holt
et al., 1990; Santorelli et al., 1993; Tatuch et al., 1992). Other mutations in ATP6 have
also been correlated with both diseases including T8993C (Hamosh et al., 2005;
Santorelli et al., 1996), T9176G (Carrozzo et al., 2004), T9176C (Thyagarajan et al.,
1995) and A8527G (Dubot et al., 2004). The T-G transversions are more severe than
the T--C transitions for both the 8993 and 9176 mutation sites (Schon et al., 2001).
Leigh syndrome can also be caused by mutations in nuclear encoded subunits of the
respiratory chain and pyruvate dehydrogenase. It can be X-linked or autosomal recessive
in addition to being maternally inherited. The maternally inherited Leigh syndrome
(MILS) mutations account for 18% of all Leigh syndrome percentages (Dahl, 1998), of
which the T8993G mutation in ATP6 is the most common (Schon et al., 1997).
Studies on the cellular defects associated with specific mutations in ATP6 have
used E. coli, patient cells and human cytoplasmic hybrids (cybrids). Cybrids are cells
containing the nuclear DNA from one cell type, usually an established cell line, and the
mitochondria from another cell type, in this case the mitochondria from patients with the
disease (Ziegler and Davidson, 1981). In this way, the mitochondrial mutations can be
studied with the same nuclear genetic background to ensure that the effects are due to the
Theoretically, mutations in the ATP6 subunit could prevent assembly of the Fo
portion of ATP synthase (Houstek et al., 1995), inhibit proton flow through the subunit
(Hartzog and Cain, 1993), or prevent the proper coupling of proton flow with the rotation
of the c subunits. (Garcia et al., 2000b) have shown proper assembly of ATP synthase in
patient cells with high levels of the T8993G mutation. In NARP mutant cybrids,
however, some cells were able to assemble only one-third of the F1Fo-ATPase with the
ATP6 subunit and others had no defect in assembly (Nijtmans et al., 2001). Cells devoid
of mitochondrial DNA, and therefore having no ATP6 or ATP8, were able to assemble
half of the Fi with the c subunits (Nijtmans et al., 2001). In E. coli the mutation
analogous to the human T9176G mutation showed defects in proton pumping and ATP
synthesis (Hartzog and Cain, 1993; Carrozzo et al., 2004).
Many studies have shown a defect in ATP synthesis in cybrid cells and patient
fibroblasts containing the most common NARP mutation, T8993G. Studies on
homoplasmic T8993G mutant cells show 50 to 80% reduction in ATP synthesis (Tatuch
and Robinson, 1993; Uziel et al., 1997; Manfredi et al., 1999; Mattiazzi et al., 2004;
Nijtmans et al., 2001; Vazquez-Memije et al., 1996; Baracca et al., 2000; Garcia et al.,
2000a). Platelets from unaffected family members with -35% T8993G mutant mtDNA
showed a 65% decrease in ATP synthesis (Carelli et al., 2002). Complex I- and complex
II-dependent ATP synthesis was detectably decreased even in clones having only 10%
mutant mtDNA (Mattiazzi et al., 2004).
Two studies showed a decrease in respiration even in an uncoupled state (Mattiazzi
et al., 2004; Trounce et al., 1994). Oxygen consumption in uncoupled cells is
independent of ATP synthase, so these data suggest a defect in other complexes of the
respiratory chain. Direct assays on complexes I, II and IV confirmed a defect in the
activities of each of these three complexes that correlated with the amount of mutant
mtDNA (Mattiazzi et al., 2004). Additionally, the activity of aconitase in the citric acid
cycle was reduced in mutant cybrids (Mattiazzi et al., 2004). Aconitase contains
iron-sulfur centers similar to complexes I, II and III. This implicates reactive oxygen
species (ROS) as the causative agent of the decrease in activity of these complexes, since
ROS act on iron-sulfur centers. Consistent with this implication, the activity of MnSOD
was increased (Mattiazzi et al., 2004; Geromel et al., 2001) and levels of H202 were
increased (Mattiazzi et al., 2004). Although respiration rates were decreased, there was a
slight increase in membrane potential (AY) in the mutant cybrids, and the pH of the
matrix in 100% mutant cybrids was increased relative to wild type (Mattiazzi et al.,
2004). This increase in matrix pH is consistent with reduced flow through the proton
channel of ATP synthase as seen in E. coli studies (Hartzog and Cain, 1993; Ogilvie and
Capaldi, 1999; Carrozzo et al., 2000) and human studies (Schon et al., 2001). This
increase is also consistent with studies correlating ROS with a high membrane potential
(Mattiazzi et al., 2004; Korshunov et al., 1997). Mattiazzi et al. also showed a galactose
growth defect in cybrids with 30 to 100% mutant mitochondrial DNA (Mattiazzi et al.,
2004). Treatment with antioxidants partially rescued the defect seen in NARP mutant
cybrids (Mattiazzi et al., 2004).
The ability to alter the components of the mitochondria is useful for creating animal
models of mitochondrial defects as well as for treating mitochondrial disease patients.
Unfortunately, the techniques used for creating or correcting nuclear DNA mutations
cannot be directly applied to mammalian mitochondria, and new methods had to be
developed to manipulate the mitochondria.
Creating Heteroplasmy in Cells and Mice
As mentioned above with respect to NARP, cybrids or cytoplasmic hybrids are
cells that contain foreign mitochondria. If the recipient cell mitochondrial DNA was
removed before introducing the other mitochondria, the resulting cybrid will contain the
nuclear DNA from the recipient cell and the mitochondria from the donor cell. The
mitochondria from the parental cell line are removed by growing the cells in the presence
of a mutagen such as ethidium bromide or rhodamine-6-G (RG6), which select against
the mitochondrial DNA. Cells lacking mitochondrial DNA, also called po cells, can be
grown in media containing pyruvate and uridine. The cells containing the donor
mitochondria are enucleated with cytochalasin B and then the two cell types are then
fused (King and Attardi, 1989; Trounce and Wallace, 1996; Ziegler and Davidson, 1981).
Cybrids can be used to study mutations from non-dividing cells by putting them in a cell
line that can be easily manipulated. It requires the existence of the mutation in the
species of interest.
The first heteroplasmic mice were created by fusing oocytes from different mouse
strains (Jenuth et al., 1996; Jenuth et al., 1997; Meirelles and Smith, 1997; Meirelles and
Smith, 1998). While not harboring deleterious mutations, these mice showed skewed
heteroplasmy in blood, spleen, liver and kidney (Jenuth et al., 1997). The first mutant
heteroplasmic mouse was generated by fusing embryonic stem (ES) cells (Marchington et
al., 1999). Levy et al. (Levy et al., 1999) used ES cells previously treated with R6G to
remove wild type mtDNA and fused them with enucleated cells containing mutant
mitochondrial DNA. Resultant mice were nearly homoplasmic for the mutant (Levy et
al., 1999). The first heteroplasmic mouse with germline transmission was created by
fusion of zygotes with enucleated cybrids bearing mtDNA with large deletions (Inoue et
al., 2000). Although the wild-type mtDNA was not removed from the zygotes, and the
first generation had less than 20% mutant mtDNA, the third generation had up to 80%
mutant mtDNA (Inoue et al., 2000). Another heteroplasmic mutant mouse with germ line
transmission was produced by fusion of RG6-treated ES cells with enucleated cell
cytoplast containing mtDNA with point mutations (Sligh et al., 2000).
Manipulating the Mitochondrial Genome
Altering existing mitochondrial DNA
Restriction enzymes can be targeted to the mitochondria by adding a mitochondrial
signal sequence to their amino-termini. In a heteroplasmic state where a difference
between the two haplotypes includes a unique restriction site, the corresponding
restriction enzyme can eliminate one of the haplotypes. Mitochondrially targeted PstI has
been shown to shift the heteroplasmic state of rodent cells where only one haplotype
contained PstI sites (Srivastava and Moraes, 2001). The T8399G mutation that causes
NARP and MILS creates a Smal restriction site that is not found in wild type
mitochondrial genomes. Restriction enzymes could be therapeutic for diseases with such
Different ways of introducing small segments of DNA have been explored.
Polynucleotides cross-linked to a protein with a mitochondrial-targeting signal can be
imported into isolated mitochondria (Vestweber and Schatz, 1989; Seibel et al., 1995).
Peptide nucleic acids were designed to be complementary to point mutations and deletion
breakpoints and were able to selectively inhibit replication of the mutant mtDNA in
isolated mitochondria (Taylor et al., 1997). They have been imported into mitochondria
in cultured cells by fusion to a mitochondrial targeting signal (Taylor et al., 1997;
Muratovska et al., 2001), but have not been shown to be functional in intact cells (Taylor
et al., 2000; Muratovska et al., 2001).
Even if DNA can be delivered to the mitochondria, there is no reason for it to stay
there. It is believed that homologous recombination does not occur in mammalian
mitochondria. Crossover products and excision repair activity have not been found in
mammalian mitochondria (Clayton et al., 1975). Zuckerman et al. (Zuckerman et al.,
1984) were not able to detect recombination of mitochondrial DNA in somatic cell
hybrids. On the other hand, there is some evidence for homologous recombination in the
mitochondria. A RecA-like activity necessary for homologous recombination has been
found in the mitochondria (Thyagarajan et al., 1996). Two population studies claim
sequence comparisons contain evidence for homologous recombination in humans (Eyre-
Walker et al., 1999; Awadalla et al., 1999). High frequency homologous recombination
occurs in mussels, plants and fungi (Ladoukakis and Zouros, 2001). It seems likely that
recombination occurs in mammalian mitochondria, but not frequently enough to facilitate
targeted alterations of the mitochondrial genome (Eyre-Walker, 2000).
Introduction of entire mitochondrial genomes
If the entire mitochondrial genome is introduced into mitochondria, it should be
able to replicate and remain there. One group cloned the entire mitochondrial genome in
E. coli and was able to electroporate the mtDNA into isolated mitohcondria (Yoon and
Koob, 2003). RNA was detected that had been synthesized from the new mtDNA
template in these isolated mitochondria from both wild type and po cells (Yoon and
Another group cloned the mitochondrial genome in E. coli and completed it with
the mitochondrial transcription factor, Tfam, before introducing it into cells (Khan and
Bennett, Jr., 2004). The Tfam they used was modified with a protein transduction
domain for import into cells as well as a mitochondrial localization signal. The protein
transduction domain used to allow Tfam to efficiently cross cell membranes was from the
Tat protein from the human immunodeficiency virus (HIV) (Vives et al., 1997). About
one thousand molecules of Tfam are bound to each endogenous mtDNA molecule (Alam
et al., 2003), so the exogenous mtDNA bound the modified Tfam will resemble the
endogenous one. After entry into the mitochondria, the mitochondrial localization
presequence should be cleaved from Tfam, which will also remove the protein
transduction domain at the amino terminus. This technique, termed protofection, proved
to be an efficient means of introducing DNA into the mitochondria (Khan and Bennett,
There are more than 70 different pathogenic tRNA mutations and most are
heteroplasmic (Turnbull and Lightowlers, 2002; Wallace, 1999). It may be possible to
treat these diseases by importation of different tRNAs (Entelis et al., 2001b) as has been
shown for tRNALys discussed above (Kolesnikova et al., 2004). Additionally, tRNA
suppressor mutants may be able to ameliorate the effects of point mutations in protein
coding genes of which more that 30 mutations are known (Entelis et al., 2001b). More
molecules of the 5S rRNA are imported than tRNAs (Kolesnikova et al., 2004), but a
smaller percentage (Entelis et al., 2001a). Either the 5S rRNA or tRNAs could be used as
vectors to deliver RNAs with therapeutic activity, such as interfering with replication of
mutant mtDNA (Kolesnikova et al., 2004).
Nuclear-encoded mitochondrial proteins
The vast majority of mitochondrial proteins is encoded in the nucleus and is
imported into mitochondria following translation on cytoplasmic ribosomes. The genes
for these proteins can be mutated or knocked out by transgenic methods. Knockouts have
been made for many nuclear-encoded mitochondrial proteins including MnSOD
(Lebovitz et al., 1996; Li et al., 1995; Melov et al., 1999), GPxl (Esposito et al., 2000),
uncoupler protein (Ucpl) (Enerback et al., 1997), and the adenine nucleotide translocater
(ANT1) (Graham et al., 1997). Of particular interest to the study of mitochondrial DNA
mutations are the Tfam knockout and PolgA knock-in.
Tfam is the major mitochondrial transcription factor and is also involved in
generating the primer for mitochondrial DNA replication. The mouse knockout of Tfam
is lethal between embryonic day 8.5 and 10.5 (Larsson et al., 1998). Little to no mtDNA
is detectable at this stage. The heterozygote is viable and has 34% fewer copies of
mtDNA (Larsson et al., 1998).
PolgA is the mitochondrial DNA polymerase-y. A mouse with a mutation in
PolgA, affecting its proof-reading function, was shown to have an increase in mtDNA
mutations and showed many signs of early ageing, supporting the notion that mtDNA
mutations are a cause of ageing and not just another symptom (Trifunovic et al., 2004).
This mouse could be a model for progressive external ophthalmoplegia, caused by a
mutation in human PolgA, although the symptoms are not identical (Agostino et al.,
Allotopic protein import
Allotopic protein import refers to the import of proteins into the mitochondria that
are normally synthesized there. In order to manipulate mitochondrially-encoded proteins,
they can be made in the cytoplasm with a mitochondrial import signal sequence. In order
to do this, the gene sequence must be altered to account for the differences in the
mitochondrial genetic code. At least the four codons must be changed that code for
different amino acids or stop codons. Most important is UGA that codes for trypophan in
the mitochondria and would create a truncated protein in the cytoplasm. Additionally,
codon usage frequencies differ between the mitochondria and the cytoplasm and could
affect the efficiency of translation. Although import of a wild type protein does not
involve the removal of the mutant protein, mitochondrial diseases do not typically affect
individuals with less than 70% mutant mitochondrial DNA, so the presence of the mutant
protein does not seem to be a problem (Zeviani and Antozzi, 1997; Hayashi et al., 1991).
Import of allotopic proteins is not always successful due to the hydrophobicity of
the proteins. Allotopic protein expression was first studied in yeast (Farrell et al., 1988;
Law et al., 1988; Gearing and Nagley, 1986). The S. cerevisiae ATPase 8 protein was
imported and restored respiration in a mutant strain (Farrell et al., 1988). ATPase 9 could
only be imported when modified with amino acids from the N. crassa sequence that
reduced its hydrophobicity and did not restore respiration (Law et al., 1988). Since only
COX1 and apocytochrome b (cyt b) are encoded in the mitochondrial genomes of all
species, importable versions of most of the mitochondrial proteins exist in nature (Oca-
Cossio et al., 2003). The problem still remains that the importable version from one
species may not be functional in another species (Law et al., 1988).
In human cells ATP6, ATP8 and ND4 have been imported allotopically (Manfredi
et al., 2002; Guy et al., 2002; Oca-Cossio et al., 2003). Import of ATP6 resulted in
incorporation of the allotopic protein into ATP synthase (Manfredi et al., 2002). The
allotopic ATP6 was able to rescue defects in galactose growth with oligomycin and ATP
synthesis in NARP cybrid cells (Manfredi et al., 2002). Import of ND4 was able to
rescue complex I-dependent ATP synthesis in LHON cybrid cells (Guy et al., 2002).
Another group was unable to see import of ND4 or cyt b, but did achieve import of ATP6
and ATP8 (Oca-Cossio et al., 2003).
Ribozymes are useful tools for reducing the levels of a specific RNA, for example,
the RNA for one of the nuclear-encoded subunits of complex I (Qi et al., 2003b).
Hammerhead ribozymes are the most common type used for this purpose (Lewin and
Hauswirth, 2001; Gorbatyuk et al., 2005; Ruffner et al., 1990; Birikh et al., 1997). They
were first described as cleaving in cis as part of a plant viroid's replication cycle (Forster
and Symons, 1987). They can be manipulated to cleave in trans by dividing the
ribozyme into two parts (Uhlenbeck, 1987). Hammerhead ribozymes are composed of
three helices and a core of 13 nucleotides that impart the catalytic activity (Symons,
1992). Helix I and helix III can be used as target arms by altering the nucleotide
sequence of one strand to be complementary to the RNA that should be cleaved. The
target RNA becomes the other strand of helix I and helix III.
The cleavage substrate of the hammerhead ribozyme must contain a UX, where X
is any nucleotide besides G (Shimayama et al., 1995; Ruffner et al., 1990). Hammerhead
ribozymes are capable of cleaving any site with a UX, but GUC is cleaved most
efficiently (Shimayama et al., 1995). Ribozymes are generally designed to base pair with
five to seven nucleotides on each side of the UX (Fritz et al., 2002b). Ribozymes with
UA following the GUC seem to have better activity, but not all combinations have been
tested, and this enrichment may not applicable to cleavage triplets other than GUC
(Clouet-d'Orval and Uhlenbeck, 1997).
Ribozymes are usually designed to target 11 to 14 nucleotides, a sequence short
enough to allow the ribozyme to dissociate after cleavage, so that it can cleave multiple
targets (Fritz et al., 2002a). Assuming random nucleotide sequences, a 13-nucleotide
sequence would statistically be found in the human or mouse genome about 200 times.
Fortunately, most genomic DNA is not transcribed, so it is likely that a specific target
sequence will only be present in one other transcript, if at all. Even if there are other
matches, those mRNAs may not be expressed in the cells being used. It is advisable to
check for other potential targets using BLAST (Gaeta, 2000).
Adeno-associated virus (AAV) is a non-pathogenic virus that has been used to
deliver a wide variety of genes into animals including an allotopic protein and a ribozyme
both mentioned above (Guy et al., 2002; Qi et al., 2003b). It has also been used to
deliver the Sod2 gene and a ribozyme targeting the Sod2 RNA (Qi et al., 2003a; Qi et al.,
2004), as well as GFP with a mitochondrial localization signal (Owen, IV et al., 2000).
AAV is a single-stranded DNA parvovirus of about 4.6 kb containing only two
genes, Rep and Cap, that encode four rep proteins and three capsid proteins. AAV can
efficiently transduce many cell types including non-dividing cells (Flotte and Ferkol,
1997). However, AAV requires coinfection with a helper virus, typically Adenovirus, for
a productive infection. Therefore, a recombinant virus, lacking the AAV genes, requires
both wild type AAV and a helper virus in order to replicate. This and its preferential
integration into human chromosome 19 (Balague et al., 1997) make AAV attractive for
To package recombinant AAV, the only cis requirements are the two terminal
repeats (TRs) between which the genes to be packaged can be cloned. The wild type
AAV genes and necessary Adenoviral genes can be supplied in trans without TRs, so that
only the genes of interests are packaged (Zolotukhin et al., 2002). Many serotypes of
AAV have been discovered in humans and other primate species (Gao et al., 2004; Gao et
al., 2005). The most widely studied serotype is AAV2, though other serotypes such as
AAV1 and AAV5 have been found to be superior to AAV2 in transducing specific
tissues. DNA with the AAV2 TRs can be packaged in the capsid proteins of other
serotypes (Rabinowitz et al., 2002). This practice is termed "pseudotyping." AAV2
DNA packaged in AAV5 capsids has been shown to more efficiently transduce
photoreceptors and other cells of the retina than with AAV2 or AAV1 capsids (Auricchio
et al., 2001; Surace et al., 2003).
Ribozymes were designed to target the mitochondrial ATP6 messenger RNA. The
ribozymes were shown to be capable of cleaving their target RNA in vitro. When tested
in cell culture, a phenotype resembling a mitochondrial defect was seen, but it did not
correlate with the presence of the ribozyme. Ribozymes were shown to be imported into
the mitochondria in cultured mouse and human cells, but the amount of import achieved
does not appear to be sufficient to noticeably affect the target mRNA levels. Sub-retinal
injection of mice with one of the ribozymes in Adeno-associated virus was not capable of
inducing detectable damage.
Ribozyme Design and In Vitro Analysis
Two ribozymes were designed to cleave the mouse mitochondrial ATP6 messenger
RNA. The first 43 nucleotides of ATP6 overlap the ATP8 gene, so ribozyme target sites
were sought following this region. The most common mutation in the ATP6 gene
associated with human disease is T8993G, which results in an arginine instead of a
leucine at amino acid 156 (Holt et al., 1990; Santorelli et al., 1993). This is the second
highlighted codon in Fig. 2-1. Ribozymes were designed to target upstream of this
codon. If the 5' end of the cleaved mRNA were still translated, the truncated protein
would not contain this apparently important region. Hammerhead ribozymes are known
to cleave more efficiently after certain triplet sequences (Shimayama et al., 1995), so
GUC, CUC, AUC, AUA and GUA sites were examined. Possible ribozyme targets were
eliminated if they or their ribozymes were predicted to have inhibitory secondary
structure according to M-fold (Zuker, 2003) or if more than 12 nucleotides were an exact
match to genes other than ATP6 as determined by BLAST (McGinnis and Madden,
2004) searching. Three ribozymes target sites were chosen, and the ribozymes were
named for the nucleotide of the coding sequence after which they should cleave:
ATP6m69rz, ATP6ml71rz and ATP6m252rz, where "m" indicates the mouse gene.
Ribozymes were designed similarly for the human ATP6 gene including ATP6h206rz
(designed by Alfred Lewin), ATP6h208rz and ATP6h 114rz, where "h" indicates the
human gene. A ribozyme was also designed by Alfred Lewin to target human COX2,
aug aac gaa aau cua uuu gcc uca uuc auu ace cca ac a aua aua gga uuc cca auc
guu gUA GCC AUC AUU AUA uuu ccu uca auc cua uuc cca ucc uca aaa
cgc cua auc aac aac cgt cuc cau ucu uuc caa cac uga cua guu aaa cuu auu auc
aAA CAA AUA AUG CUa auc cac aca cca aaa gga cga aca uga ace cua aua
auu guu ucc cua auc aua uuu auu gga uca aCA AAU CUC CUA GGc cuu uua
cca cau aca uuu aca ccu acu acc c aa cua ucc aua aau cua agu aua gcc auu cca
cua UGA gcu gga gcc gua auu aca ggc uuc cga cac aaa cua aaa age uca cuu
gcc cac uuc cuu cca caa gga acu cca auu uca cua auu cca aua cuu auu auu auu
gaa aca auu age cua uuu auu can ca aug gca UUA gca guc CGG cuu aca gcu
aac auu acu gca gga cac uua uua aua cac cua auc gga gga gcu acu cua gua
uua aua aau auu age cca cca aca gcu ace auu aca uuu auu auu uua cuu cua cuc
aca auu cua gaa uuu gca gua gca uua auu caa gcc uac gua uuc ace cuc CUA
gua age CUA uau cua cau gau aau aca uaa
Fig. 2-1. Annotated mouse ATP6 mRNA sequence. The longer capitalized sequences
are the targets of ribozymes m69, m171 and m252 respectively. Single
capitalized codons are the sites of mutations associated with mitochondrial
Ribozymes were tested for their ability to cleave their short target RNA molecules
in vitro. These targets were 13- or 14-nucleotide ribonucleotides that were purchased
from Dharmacon, Inc. Ribozymes and radioactively end-labeled targets were incubated
with magnesium for various lengths of time. The reactions were run on acrylamide gels
and visualized with a phosphorimager. The uncut target and 5' cleavage products were
quantitated to determine the percent cleavage at each time point.
5'UAGCCAUCAUUAUA3'5'AACAAAUAAUG C U3' 5'CAAAUCUCCUAGG3'
3'AUCGGUA UAAUAU5'3 'UIGIUA UA1 A A 5'3'GIU UU AIAGA GAU CC5'
C C C
A U A U A U
G G G
A A A A A A
U U U
G G G G G G
A A A
m69rz C G l71rz C G 252z C G
G-C G-C G-C
C-G C-G C-G
G-C G-C G-C
G U G U G U
CU CU CU
Fig. 2-2. Depiction of ribozymes annealed to their targets. Arrows indicate the
At 20 mM MgC12, ribozymes m69 and m252 cleaved 25% of their targets in one
and three minutes respectively, and both reached 70% cleavage in only 30 minutes (Fig.
2-3 A and C). Ribozyme m171 could only cleave 12% of its target in twelve hours, so
this ribozyme was discarded (Fig. 2-3B). Time courses for m69 and m252 ribozymes
were repeated at 5 mM MgC12 to slow the reaction and find an ideal time point for
saturation kinetics. At 5 mM MgC12 ribozyme m252 cleaved slightly faster than
ribozyme m69. They were able to cleave 20% of their targets in about 25 and 30 minutes
respectively (Fig. 2-3 D and E). The time courses of these two ribozymes are comparable
to other ribozymes that have been shown to work in vivo (Drenser et al., 1998).
Multiple turnover reactions were performed to calculate the kinetics of each
ribozyme (Fig. 2-4). Ribozyme m69 reactions were stopped at 25 minutes, and m252
reactions were stopped after 15 minutes. These intervals were chosen so that less than
15% of the substrate had been cleaved, and therefore the observed reaction rate could be
1.00 A 1.00 D
S0.40 0 0.40
0.00 0 I 1 0.00 I I I i
0 90 180 270 360 450 540 630 0 90 180 270 360 450 540 630 720
0 90 180 270 360 450 543 630 720
1.00 1.00 E
0.80 \ 0.80 -
0 90 180 270 360 450 540 630 0 90 180 270 360 450 540 630 720
Fig. 2-3. In vitro time course of ribozyme cleavage. A) ATP6m69rz at 20 mM MgCl2
with a target-to-ribozyme ratio of 9:1. B) ATP6ml71rz at 20 mM MgCl2 with
a target-to-ribozyme ratio of 10:1. C) ATP6m252rz at 20 mM MgC12 with a
target-to-ribozyme ratio of 10.5:1. D) ATP6m69rz at 5mM MgC12 with a
target-to-ribozyme ratio of 10:1. E) ATP6m252rz at 5mM MgC12 with a
target-to-ribozyme ratio of 10:1.
used to estimate the initial velocity. Each reaction had 15 mM ribozyme with an
increasing concentration of target. Maximum velocity (Vmax), the Michaelis-Menten
constant for affinity (Km), and a reaction rate under saturating substrate concentrations
(kcat) were calculated for both ribozymes (Table 1). Although these parameters were
measured at 5 mM MgC12, the kcat of both of these ribozymes is above the minimal
acceptable value of 0.1 min- at 20 mM MgC12 (Fritz et al., 2002a) and similar to the kcat
of a published ribozyme that worked in vivo in a rat model of retinitis pigmentosa
(Drenser et al., 1998). The kcat of ribozymes m69 and m252 would be even higher if
measured at 20 mM MgC12, because the observed initial rates were higher (Fig. 2-3).
Human ribozymes ATP6h208, ATP6h1 14 and COX2h24rz were tested in vitro on
short target RNA molecules (Alfred Lewin and Joe Hartwich, unpublished). ATP6h1 14
and COX2h24rz were able to cleave their short RNA targets efficiently, but ATP6h208
was not effective. ATP6h206rz was not tested in vitro.
S00 .o A 0.50 -
S 0. 0.40
Sy=18x+0.1 y= 158x+0.12
o 0.20 R2= 098 = 0 0.10- 0
S0.98R2 = 0.95
,1 0.00 .-. ,..V, 0.00,,
-0.001 0.000 0.001 0.002 0.003 0.004 -0.001 0.000 0.001 0.002
1/substrate concentration (1/nMA) 1/substrate concentration (1ll/M)
Fig 2-4. Lineweaver-Burke plots of multiple turnover reactions for ribozymes.
A) ATP6m69rz. B) ATP6m252rz.
Table 2-1. Kinetic parameters for mouse ATP6 ribozymes measured at 5 mM MgC12
Vmax (nM/min) kcat (min-') Km (nM) kcat/Km (min-'tM-1)
ATP6m69rz 9.2 0.62 1700 0.37
ATP6m252rz 8.3 0.55 1300 0.42
The mouse ATP6m69 and ATP6m252 ribozymes were also tested in vitro on total
mitochondrial RNA. Mitochondrial RNA was isolated from NIH3T3 cells as described
for detecting the localization of the ribozyme (below and Materials and Methods). Each
ribozyme was incubated with an equal amount of mitochondrial RNA and 20 mM MgC12
for either 27 or 91 minutes at 370C. Reverse transcription and polymerase chain reaction
(RT-PCR) was performed to compare the remaining ATP6 mRNA to reactions incubated
without ribozyme (Fig. 2-5). The RNA incubated with the m252rz had less remaining
ATP6 mRNA than RNA incubated without ribozyme. The m69rz reactions had less
ATP6 than the ATP6m252rz reactions at each time point. Interestingly, the reactions
incubated for 27 minutes had less remaining ATP6 than those incubated for 91 minutes.
The m69rz reactions had 40% and 20% reductions in ATP6 compared with the no
ribozyme reactions at 27 and 91 minutes respectively. The m252rz reactions had 20%
and 10% reductions in ATP6 compared with the no ribozyme reactions at 27 and 91
minutes respectively. If maximum cleavage was reached by 27 minutes, as was seen with
the short targets, the difference in the time points could be simply the variation between
duplicate samples due to pipetting error; the difference may not be dependent on the
incubation time. Although this experiment was not repeated, and no internal control was
used, it appears as though at least the m69rz is capable of cleaving its full-length target in
27 minutes 91 minutes
m69rz m252rz no rz m69rz m252rz no rz
Fig. 2-5. In vitro ribozyme cleavage of total mitochondrial RNA. The ATP6 mRNA
was amplified by RT-PCR.
The in vitro experiments above used ribozyme molecules without the upstream 5S
rRNA sequence. To test whether the ribozyme is active when attached to the 5S rRNA,
the 5S-ribozyme was synthesized by in vitro transcription and tested on its short target at
various time points (Alfred Lewin, unpublished). The 5S-ATP6m69rz and
5S-COX2h24rz were able to cleave their respective target RNAs in vitro.
Mouse Cell Culture Phenotypes
Both ATP6m69rz and ATP6m252rz were cloned into the HindIII site at the end of
the 5S rRNA in forward and reverse orientation (Fig. 2-6 and Materials and Methods).
The reverse orientation constructs, ATP6m69rev and ATP6m252rev were used as
negative controls along with the 5S rRNA with no ribozyme. UF11 contains the green
fluorescent protein (GFP) gene driven by the cytomegalovirus (CMV) promoter and
chicken 3-actin enhancer (CBA). GFP expression was used for visualizing transfection
efficiencies and as a control that did not contain the 5S rRNA. All of these constructs
include a neomycin resistance cassette for selection purposes and adeno-associated virus
(AAV) terminal repeats (TRs) for packaging into AAV. The AAV packaged ribozymes
were only used for injection into mice, because they do not infect mouse cells in culture
very efficiently (Hansen et al., 2000).
Sfl(+) origin TR
fl(+) origin 5S Promoter and RNA 1C M M ie enhancer
iI i) 0 Ch iken n-actin promoter
_Hnd T IIIy A Exon1
HSV4k ApR Intron
SColEI odri GFPh
r0 \SV40 poly(A)
ColE1 ori hGH polyA TR PYF441 enhancer
SbGH poly(A) HSV-tk
Fig. 2-6. 5S and UF 11 constructs. Ribozymes were cloned in both orientations into the
HindIII site at the end of the 5 S rRNA.
Mouse NIH3T3 cells were transfected with both ATP6 ribozymes four separate
times. No reproducible phenotype was seen that could be correlated with the levels of
ribozyme or target mRNA. A phenotype was seen in some of the clonal isolates that
seemed to reflect a mitochondrial defect, but did not correlate with levels of ribozyme or
ATP6 mRNA. Each of the four transfections was analyzed at different stages and by
different assays. First the four transfection attempts will be reviewed, and then the two
sets of clonal isolates with the sporadic phenotype will be discussed.
Four Transfection Attempts
The first transfection included ATP6m69rz, ATP6m69rev, ATP6m252rz,
ATP6m252rev, UF 11 (GFP) and mock transfection with Lipofectamine and the Plus
reagent. Cells were plated two hours before transfection and only 3 |tg DNA was used,
resulting in about 5% transfection efficiency for UF 11. UF 11 cells were not allowed to
grow, but cells transfected with m69rz, m69rev, m252rz and m252rev were selected with
G418 while the mock-transfected cells were grown without G418.
After 12 days under selection and 2 days without G418, an ATP synthesis assay
was done on permeabilized cells using succinate as a substrate (Table 2-2). Using
succinate as a substrate measures complex II-dependent ATP synthesis; complex
I-dependent ATP synthesis was not measured in these cells. Cells were trypsinized and
permeabilized with digitonin so that the substrates could enter. ADP and succinate were
used as substrates. Luciferin and luciferase were added in order to detect the production
of ATP. The flash of light produced by the luciferase reaction was detected with a
luminometer immediately after adding the substrates. Three reactions were performed
from each plate of cells. There was some decrease in activity between the first and third
reactions, which could account for some of the variation within samples. A gradual
decrease in luciferase activity over the course of the assay has also been seen, which
could account for some of the variation between samples. The order of the samples was
as listed in Table 2-2, so the mock-transfected cells may have had the lowest rate of ATP
synthesis. None of the transfected cells had a decreased rate of ATP synthesis compared
to the mock-transfected cells.
Table 2-2. Rates of succinate-dependent ATP synthesis in total selected mouse cells.
The average rate for three consecutive reactions is given with the standard
deviation in parentheses.
Mock m69rz m69rev m252rz m252rev
nmolATP/min/[tg protein 2.7 (1.6) 4.3 (2.0) 5.1 (3.3) 2.8(1.2) 2.3 (0.9)
After at least 2 weeks under G418 selection, the growth of these cells was
compared in glucose, galactose, galactose with 0.4 ng/ml oligomycin (Fig. 2-6) and
galactose with 0.8 ng/ml oligomycin. When grown in glucose, cultured cells rely on
glycolysis for their energy source (Reitzer et al., 1979), but in galactose media they are
more dependent on mitochondrial oxidative phosphorylation (Robinson, 1996).
Therefore, cells with a defect in the oxidative phosphorylation pathway should not be
able to grow in galactose. Oligomycin is an inhibitor of ATP synthase, so addition of
small amounts of this reagent to the galactose media should make this assay more
sensitive by inhibiting some of the ATP synthesis (Breen et al., 1986; John and Nagley,
1986; Manfredi et al., 1999).
For the transiently transfected cells tested here, there was no difference between
ribozyme, reverse or mock-transfected cells in their ability to grow in galactose. They all
grew about half as fast in galactose as in glucose. When grown in galactose and the
lower amount of oligomycin, all of the cells grew even slower for the first four days, but
then the ribozyme and reverse transfected cells started to grow faster while the
mock-transfected cells continued to grow at the same slow rate (Fig. 2-7). This could
possibly be explained by the fact that the mock-transfected cells did not go through the
stress of being selected with G418 and were therefore more sensitive. Alternatively,
transfection with any of the vectors could have made the cells less sensitive to
oligomycin or more respiratory competent. When the cells were grown in galactose with
the larger amount of oligomycin, they were only grown for 4 days with galactose and 0.8
ng/ml oligomycin and then were allowed to recover in glucose media. None of the cells
grew while the oligomycin was present, and they all recovered equally well in glucose,
including the mock-transfected cells. This suggests that the ribozyme and reverse
constructs had no effect on oxidative phosphorylation or oligomycin resistance. Since no
phenotype was detectable in these total selected cells, clonal isolates were derived to
Days in Galactose and Oligomycin
2 4 6 8 10
Days in Galactose and Oligolnycin
Fig. 2-7. Total G418-selected cells grown in galactose with 0.4 ng/ml oligomycin.
Mock-transfected cells were not pre-selected with G418.
analyze homogenous populations of cells with various levels of ribozyme expression
(section below on clonal isolates). These colonies were designated with Arabic numerals.
The second transfection attempt included ATP6m69rz, ATP6m252rz,
ATP6m252rev and UF11 (GFP) each transfected in duplicate. Although about 20% of
the cells were green on the UF 11 transfected plates, the transfection efficiencies of the
others may have been much lower, since selection with G418 resulted in distinct colonies
for both m69rz and m252rev. The efficiency of m252rz transfection was apparently
better, as it did not lead to distinct colonies. The total selected m252rz cells, m69rz
colonies, and m252rev colonies were tested for their ability to grow in galactose. They
all appeared to grow equally well in glucose and galactose. In conclusion, none of the
cells had an obvious defect in oxidative phosphorylation as indicated by their ability to
grow in galactose, but this could be attributed to the low transfection efficiencies.
The third transfection was done in duplicate with ATP6m69rz, ATP6m252rz,
ATP6m252rev and UF 11 (GFP). The transfection parameters were slightly different for
each of the duplicates (see materials and methods) such that one resulted in 15% and the
other in 30% of the cells transfected as determined by the number of green cells from the
UF 11-transfected cells. UF11 cells were not grown further. Two days post-transfection,
cells were selected with G418 for either two or three days before splitting them into
either continued G418 selection or 0.5mM potassium cyanide (KCN).
Cyanide kills cells through a cytochrome oxidase (complex IV) dependent
pathway, so cells with a decrease in cytochrome oxidase are more resistant to cyanide
(Villani and Attardi, 2000). The cyanide resistance assay was done in response to a
phenotype seen in one of the first transfection colonies discussed below, although it is not
the primary phenotype expected. By visual inspection after four days in cyanide, the
ribozyme-transfected cells were surviving better than the m252rev-transfected cells in
both transfection sets. After five days, the m252rz cells looked more resistant to
cyanide-induced cell death than the m69rz cells for one of the sets, based on cells per
field, while they appeared the same for the other set. Based on their growth in G418, the
m69rz cells were more efficiently transfected, so they may be less resistant to cyanide
than ribozyme m252 in both cases.
Cells that were fully selected with G418 were grown in ImM cyanide and counted
in triplicate every two days (Fig. 2-8). Neither of the ribozymes was more resistant to
cyanide than m252rev. The mock-transfected cells (not selected with G418) were more
resistant to cyanide than any of the transfected cells. When this assay was repeated with
only 250 [LM cyanide, no significant difference was seen for any of the samples. Upon
visual inspection, no difference in growth in galactose was seen compared to glucose in
these total selected cells. No colonies were isolated from the third transfection.
-2 5 1 Mock
Days in Cyanide
Fig. 2-8. Total selected cell growth in 1 mM cyanide. Mock-transfected cells were not
pre-selected with G418. Error bars are the standard deviation of three
The fourth transfection included ATP6m252rz, ATP6m69rz, ATP6m69rev, 5S (no
ribozyme), UF 11 (GFP) and a mock transfection. Only 1% of UF 1-transfected cells
were green one day post transfection. Mock-transfected cells were not allowed to grow
further. The transfected cells were split into regular glucose media, galactose media and
media with G418 one day post transfection. No difference was visually detectable in
galactose compared to glucose media. Total G418-selected cells were tested for their
ability to grow in galactose (Fig. 2-9). The m69rz-transfected cells grew slower in
galactose than in glucose, but not significantly slower than the 5S and UF 11 controls in
galactose. Colonies were isolated from the selected cells and were designated with
*m252Rz O m69Rz O Dm69rev M5S M UF11
Fig. 2-9. Total selected cell growth in galactose. Values are the average doublings per
day between Days 1 and 3 in galactose and glucose. Error bars are the
standard deviation for three replicates in glucose and four replicates in
Analysis of Clonal Isolates
Ribozyme and target mRNA levels
The clonal isolates from the first transfection, designated with Arabic numerals,
were first screened for their levels of ribozyme expression by semi-quantitative RT-PCR
(Fig. 2-10). Total cell RNA was reverse transcribed, and products were amplified by
PCR at different numbers of cycles to ensure measurements were in the linear range of
amplification. Products were run on agarose gels and detected with ethidium bromide
(Materials and Methods). Of the ATP6m252rz colonies, m252rzl and m252rz6 had more
than ten-fold the levels of ribozyme expression than the average selected cell (T).
Colony m252rev3 had the highest 5S-reverse transcript levels of the nine reverse colonies
tested (only two shown), but it was still less than half of the levels of the two high
expressing ribozyme colonies. None of the ATP6m69 ribozyme colonies had higher
levels of ribozyme than the total selected cell average (data not shown), so they were not
pursued further. Only three ribozyme colonies (1, 3 and 6) and two reverse colonies (3
and 10) were further analyzed along with two mock-transfected colonies.
ED 252nT *252rzl 1 252rz i*252rr6 0252rev3 *252rev10 i*252rr 69TZ D 69rc
T 1 2 3 4 5 6 7 8 9 10 3 10 III VVIIVIIIX I IV V VII VIII IX X I II III IV V
m252rz m252rev m252rz m69z m69rev
first set of colonies last set of colonies
Fig. 2-10. Ribozyme and reverse transcript levels in clonal isolates by RT-PCR. The
first sample, T, was the total selected ATP6m252rz cells before isolation of
the first set colonies. Arabic numerals were used to refer to the first set of
colonies, while Roman numerals were used for the last set of colonies. These
two sets of colonies were assayed separately.
The clonal isolates from the fourth transfection, designated by Roman numerals,
were first screened for their ability to grow in galactose (Fig. 2-14). Then the ribozyme
expression levels from most of the clones were measured as was done for the colonies
from the earlier transfection (Fig. 2-10). None of the five m252rz colonies tested had
high levels of ribozyme expression compared to each other. Only one of the m69rz
colonies, m69rzI, had significantly higher levels of ribozyme expression than the other
six tested. Four of the five m69rev colonies had higher levels of reverse transcript than
the average ribozyme level, and one of those colonies, m69revII, had very high levels.
The levels of target ATP6 mRNA were measured by semi-quantitative RT-PCR for
some of the colonies from both sets. Ribozyme colony m252rzl was the only one with
significantly lower levels of ATP6 (Fig. 2-11). This was one of the two colonies with the
highest levels of ribozyme expression. The other colony with high levels of ribozyme,
m252rz6, did not have a reduction in ATP6 mRNA. The m252rzl colony not only had a
decrease in ATP6, it also had reduced levels of all other mitochondrial mRNAs tested
including ND6, the only mRNA encoded on the L-strand of mitochondrial DNA (Fig.
2-12). The mRNA levels of ATP6 and COX3 in colonies m252rzl, m252rev3 and
Mock5 were also analyzed by northern blot (data not shown). The relative mRNA levels
among the colonies were the same by northern as by RT-PCR. The mitochondria DNA
copy number was also measured by semi-quantitative PCR and did not correlate with
mRNA levels. Both m252rzl and Mock5 had fewer copies of mtDNA than m252rev3
(data not shown). None of the colonies from the fourth transfection had reduced levels of
ATP6 mRNA (data not shown), so other mRNA and mtDNA levels were not measured.
1 3 6 3 10
Fig. 2-11. ATP6 levels from the first set of colonies as measured by RT-PCR. Error bars
are the standard deviation of two to five replicates.
m252rzl m252rev3 Mock5
m252rzl m252rev3 Mock5
Fig. 2-12. Levels of four mitochondrial mRNAs as measured by RT-PCR. A) COX3.
B) ND4. C) ND6. D) ND2.
Growth in galactose
To determine if the ATP6m252rzl or other colonies had a defect in oxidative
phosphorylation, their rate of growth in galactose was measured. An equal number of
cells were plated on multiple dishes. Each day a subset of these dishes were counted.
For the first transfection (Fig. 2-13), the mock-transfected colonies, m252rz3 and
m252rev3 grew just as well in galactose as in glucose. Those were four colonies with no
decrease in ATP6 mRNA levels. For the colony with the most reduction in ATP6 levels,
m252rzl, more than half of the cells died after two days in galactose. Those cells that
survived grew such that there was more cell growth and division than death between Day
2 and Day 3. Colony m252rev10, that may have had a slight reduction in ATP6 levels,
had an intermediate galactose growth phenotype. On average, the cells from this colony
grew slower. Their growth slowed even more between Day 2 and Day 3 than between
Day 1 and Day 2. This could be due to slower growth of all the cells or some growth and
some death. The growth of colony m252rz6 in galactose was not quantitated, but no
difference was seen when compared visually with growth in glucose for three days. This
Glucose Galactose Day 1 2 Galactose Day 2 3
Fig. 2-13. Growth of colonies in glucose and galactose shown as population doublings
colony had high levels of ribozyme expression (Fig. 2-10), but no decrease in ATP6
mRNA (Fig. 2-11).
For the Roman numeral colonies, the galactose growth assay was the first used to
screen the colonies (Fig. 2-14). This initial screen was a comparison of three days of
growth in galactose to glucose and was not done in duplicate. The assay was repeated up
to three more times on nine of the colonies, which are indicated by error bars (Fig. 2-14).
Three of the m69rz colonies, two of the m69rev, one of the 5S (empty vector) and two of
the UF 11 (GFP) colonies did not grow as well in galactose as they did in glucose. All of
the m252rz colonies grew well in glucose and galactose. The colonies with the highest
levels of ribozyme and reverse transcript, m69rzI and m69revII (Fig. 2-10) were
2 1.5 1-
252Iibozlmne 691ibozyme 69rev 5S UF11
Fig. 2-14. Growth in galactose of all fourth transfection colonies. First all colonies were
screened; only those with error bars were repeated. The error bars are the
standard deviation for two to four replicates.
two of the colonies with the galactose growth defect, but the other defective colonies had
very low levels of transcript. The existence of 5S and UF11 colonies with a galactose
growth defect further supports the idea that this phenotype is not caused by the ribozyme
or 5S-reverse transcript.
Since the levels of COX3 mRNA were reduced in the m252rzl colony, its
sensitivity to cyanide was measured (Fig. 2-15). Cyanide is an inhibitor of cytochrome
oxidase (complex IV) (Letellier et al., 1994). When cells are grown in glucose, the
integrity of the respiratory chain is not important, but cyanide can also induce apoptosis
or necrosis (Villani and Attardi, 2000). Therefore, cells with decreased levels of complex
IV may have increased resistance to cyanide-induced cell death.
1 ? 0 -|------------------------
(ic *in 12 4 Ii
@-100 Oizn:2 5t* 3
Days in Cyanide
Fig. 2-15. Growth of colonies in cyanide. The plaid bars represent cells grown in
250 ptM KCN. The solid bars represent cells grown in 1 mM KCN.
The growth of the cells in cyanide was measured as was done for galactose growth.
Multiple plates were set up on Day 0 with two different concentrations of cyanide, and a
subset of these plates was counted for each time point. The m252rzl colony that showed
a decrease in COX3 and other mitochondrial mRNAs was more resistant to cyanide.
These cells eventually showed susceptibility to the higher concentration of cyanide by
Day 6 (Fig. 2-15) and the lower concentration by Day 8 (data not shown). The m252rev3
and Mock5 colonies were starting to die as early as Day 2 at both concentrations of
cyanide. About 50 to 80% of the cells had died by Day 4. By Day 6, in the lower
concentration of cyanide, there was more growth than death. At the higher concentration
of cyanide, the cells from colonies m252rev and Mock5 cells were still dying at Day 6.
The ribozymes are targeted to ATP6, a subunit of ATP synthase. To more directly
measure the activity of this complex, the rate of ATP synthesis was measured in
permeabilized cells (Fig. 2-16). Cultured cells were trypsinized and given ADP and
either succinate or pyruvate and malate as substrates. Succinate is the substrate for
complex II-dependent ATP synthesis; malate and pyruvate are substrates for complex
I-dependent ATP synthesis (Fig. 1-3). The production of ATP was monitored with
luciferase, which uses luciferin and ATP to produce a flash of light that can be detected
with a luminometer. Cells were permeabilized with digitonin to start the reaction.
Luminometer readings were taken at 20 sec intervals, and the rate of ATP synthesis was
No difference was seen in the rate of ATP synthesis for any samples when
succinate was used as a substrate. When pyruvate and malate were used as substrates,
however, all of the colonies that died in galactose were unable to synthesize ATP (Fig.
2-16). This included colonies from both sets of stably transfected cells, and the nature of
the DNA with which the cells were originally transfected did not make a difference.
I m252rzl Bm252rev3 IMock5 lDmi69rzl DLn69revl sl II UFII .r iiri,uife,,edl
First Colonies Fourth Colonies
Fig. 2-16. Relative rates of complex I- and complex II-dependent ATP synthesis.
Complex II-dependent rates were the same for all, so this is a comparison of
the complex I-dependent rates while correcting for loss of luciferase activity
over the course of the assay.
Since pyruvate and malate are substrates for complex I, these results suggests that the
affected colonies have a complex I defect. The ability of the cells to synthesize ATP with
the complex II substrate, succinate, suggests there is no defect in ATP synthase or any
other respiratory chain complexes. The presence of control colonies with the defect and
the fact that it resembles a complex I defect both imply that the ribozyme did not cause
Localization of Ribozymes to the Mitochondria
To determine if the 5S rRNA is effective at transporting the ribozyme into the
mitochondria, RT-PCR was done on RNA extracted from purified mitochondria from
colonies ATP6m69rzI and UF 11I (Fig. 2-17). Cells were homogenized, and the
mitochondria were collected by differential centrifugation. The mitochondria were
treated with RNase to remove any cytoplasmic RNA stuck to the outside of the
mitochondria. As a control, some of the mitochondria were not treated with RNase.
Some mitochondria were also treated with digitonin before the RNase step. Digitonin
permeabilizes the outer mitochondrial membrane, but not the inner mitochondrial
membrane, so RNA in the matrix should still be protected. As a final control, some of
the mitochondria were lysed with Triton X-100 before RNase treatment to ensure that all
unprotected RNA was degraded with the amount of RNase used. The treated
mitochondria were washed three times and DEPC-treated to remove the RNase before
The RNA extractions were treated with DNase before performing RT-PCR. The 5'
primer was identical to a portion of the 5S rRNA and the 3' primer was complementary
to the ribozyme (see Materials and Methods). In addition to the ribozyme, RT-PCR
reactions were carried out for P-actin and ND4 to monitor the presence of cytoplasmic
RNA and mitochondrial RNA respectively. The RNA extracted from the mitochondria
was not enough to quantitate by spectrophotometric means, so the samples may have had
different amounts of RNA.
The two RNase-treated ribozyme samples (R and d) were devoid of any detectable
cytoplasmic RNA, but did contain detectable ribozyme. No ribozyme was seen in cells
that were not transfected with ribozyme expressing plasmid (UF11) or in the lysed
mitochondria sample (X). This shows that the 5S rRNA with the ribozyme attached was
capable of being imported into the mitochondria. The endogenous mitochondrial RNA
was detected in these samples with only 22 amplification cycles, whereas the ribozyme
required 35 cycles. Although not very precise, this is an indication of how little of the
ribozyme is present in the mitochondria. When compared with the limits of detection in
total cell RNA, the percent of the ribozyme imported could be similar to the 0.9% of 5S
T R d X T n R d X
.. ":i,::... .. q ~i.. .:N,:: .. -,'. ...... :I..:.
P.0 ::. E .^^ ..:.:. :. ...
m 6 rz
..I .Q .. .. .. ..'I
Fig. 2-17. Detection of ribozyme in the mitochondria of NIH3T3 cells. T, total RNA; n,
mitochondria with no RNase; R, RNased mitochondria; d, digitonin-treated
and RNased mitochondria; X, Triton X-100-lysed and RNased mitochondria.
rRNA naturally imported (Entelis et al., 2001a). Again, this is not a very quantitative
method, and the estimated range is between 0.3% and 30% of the ribozyme being
The presence of the ribozyme in the mitochondria of human cells was also analyzed
(Fig. 2-18). The ribozyme used here was one designed to cleave the human COX2
mRNA. As this is an assay for the localization of the ribozyme and not for function, the
identity of the ribozyme is not significant, and an earlier experiment by Dr. John Guy
suggested that this ribozyme might be imported into mitochondria using the 5S import
sequence (personal communication). Transfections of human 293 cells were more
sequence (personal communication). Transfections of human 293 cells were more
efficient than transfections of mouse NIH3T3 cells. About 85% of the cells were
transfected as visualized by GFP fluorescence in the UF 11-transfected cells, so the cells
were analyzed during transient transfection, six days post-transfection. The isolation of
mitochondrial RNA was the same as for the mouse cells described above. Unlike the
mouse cells, enough RNA was extracted for spectrophotometric measurement, and about
0.5 |tg RNA was used for each RT reaction. As seen in the mouse cells, the ribozyme
was detectable in the mitochondrial RNA fractions R (total mitochondrial RNA) and d
(mitochondrial matrix RNA) (Fig. 2-18). Thirty-three PCR cycles were required to
visualize the ribozyme, and only 19 cycles were required to detect ATP6 in the
mitochondrial RNA fractions. Again, this is a rough indication of how little of the
T n R d X T n R d X
ATP6 6-- 0
Fig. 2-18. Detection of ribozyme in the mitochondria of 293 cells. T, total RNA; n,
mitochondria with no RNase; R, RNased mitochondria; d, digitonin-treated
and RNased mitochondria; X, Triton X-100-lysed and RNased mitochondria.
ribozyme is imported. This is consistent with the amount seen in mouse cells and for the
endogenous 5S rRNA.
Human Cell Culture Phenotypes
HeLa Cells and a Toxic 5S Transcript
Two human ATP6 ribozymes (h206rz and h114rz) and the COX2h24 ribozyme
were tested in HeLa and 293 cells for their ability to alter the phenotype of the cells.
Another construct that was tested contained the 5S rRNA followed by the entire
overlapping mouse ATP8 and ATP6 genes cloned by Alfred Lewin (see Materials and
Methods). By testing this mouse sequence in human cells, it can be distinguished from
the endogenous ATP6. Controls included the mouse ATP6m69 ribozyme, the empty 5S
construct, UF11 (GPF), mock transfection and untransfected cells.
HeLa cells were transfected three times. The first transfection resulted in 50%
GFP-positive cells of those transfected with UF 11. The growth of these transiently
transfected cells in galactose medium was compared to their growth in glucose between
one and four days post-transfection, and no difference was apparent based on cell density.
The levels of ATP6 and COX2 were measured by RT-PCR on RNA extracted four days
post-transfection, and no difference was detected. Meanwhile, some of the cells were
selected with G418, and noticeably fewer of the cells transfected with the ATP8/6
construct survived the selection process. When the second HeLa cell transfection was
selected with G418, the ATP8/6-transfected cells again survived very poorly compared
with the cells transfected with the other five constructs.
The third HeLa cell transfection was done to test whether the ATP8/6 construct was
toxic to cells. Each transfection was done in duplicate with two different preparations of
DNA to determine if transfection efficiency was the cause of the lack of resistance to
G418. The COX2 ribozyme, the empty 5S construct, UF11 and mock transfection were
used as controls. The two UF 11 transfections resulted in 1% and 25% GFP-positive
cells. When selected with G418, almost no cells remained from one of the ATP8/6
transfections. The DNA for this transfection was made at the same time as the UF11 that
resulted in only 1% green cells. The other ATP8/6 transfection still had fewer remaining
cells than both the 1%- and the 25%-green UF 11 cells. Although transfection efficiency
may have played a role, the ATP8/6 construct appeared to be toxic to HeLa cells. This
could be at the RNA level due to the 5S-ATP8/6 transcripts. Alternatively, if the RNA is
imported into the mitochondria and is translated, the mouse protein may not be
compatible with human cells.
To address this question further, RNA from the ATP8/6-transfected cells was
analyzed for the presence of the 5S-ATP8/6 transcripts (Fig. 2-19). No mouse ATP8/6
RNA was detectable one day post-transfection in HeLa cells, but it was detectable in
293T cells five days post transfection. Although the RNA had been DNased, the no RT
controls were still positive. This shows that the transfected DNA was indeed in the cells.
Since only one-fifth of the RT reaction was used for the PCR, the DNA was apparently
below the levels of detection for the ATP8/6 RT-PCR. The greater intensity of the
COXh24rz band in the first RT-PCR compared to the no RT control showed that
transcripts could be detected one day post transfection. Interestingly no COX2h24rz was
detected in the second sample. The ATP8/6 RNA was also undetectable in the first HeLa
cell transfection four days post transfection (data not shown). When regular 293 cells
were transfected and selected with G418, the ATP8/6-transfected cells survived the
selection process as well as the other transfected cells. This is further evidence that the
lack of G418-resistant cells was not due to the transfection efficiency, but rather a
HeLa-cell specific toxicity.
293T HeLa 3T3 HeLa HeLa
86 M H 86 86 + 86 CX CX CX
86 = ATP8/6
M = mock
H = no RNA
CX = COX2h24rz
Fig. 2-19. Absence of the 5S-ATP8/6 transcript in HeLa cells. Duplicate HeLa cell
samples are from duplicate transfections. No RT controls were done for the
first of the duplicates.
No Phenotype in 293 Cells
Human 293T cells were transfected twice. This cell line is already resistant to
G418, so the transfected cells could not be selected for. The first transfection tested the
three human ribozymes; ATP6h206, ATP6h 14 and COX2h24; compared to UF 11- and
mock-transfected controls. UF 11-transfected cells were more than 50% green one day
post transfection. No difference was seen between growth in galactose and glucose
between two and five days after transfection.
The second 293T cell transfection included all nine of the ribozymes and controls
mentioned above, and each was done in duplicate with two different preparations of
DNA. The two UF 11 transfections resulted in 40% and 70% GFP-positive cells. Three
days post transfection, the growth of the cells was compared with and without cyanide
(Fig, 2-20). The COX2 ribozyme could lead to cyanide resistance, if it reduces the levels
of cytochrome oxidase. One of the COX2h24rz transfections was more resistant to
cyanide than control cells, but both of the ATP8/6-transfected cells were more resistant.
The ATP8/6 gene was not expected to cause cyanide resistance.
Fig. 2-20. Growth of 293T cells for two days in 20 mM cyanide. Error bars are the range
of the duplicate transfections.
The cyanide resistance assay was repeated with HeLa and 293 cells transfected
with all combinations of COX2h24rz, UF 11 or mock transfection using Lipofectamine
2000 or Mirus LT1 (Fig. 2-21). No difference was seen between COX2h24rz and UF 11,
but there was a difference between cell types, transfection reagents and mock-transfected
cells. There was a correlation between the transfection efficiency and sensitivity to
cyanide. The 293 cells transfected with Lipofectamine 2000 resulted in 20%
GFP-positive UF11-transfected cells and had the most cell death after two days in
cyanide. The 293 cells transfected with Mirus LT1 resulted in only 1% green cells, and
both HeLa cell transfections were less than 1% green. Mock-transfected cells, by
definition, have the lowest transfection efficiency and were the most resistant to cyanide
in each set. This was the same trend seen when cells were grown without cyanide for the
first day of this experiment, so the differences between cells is apparently due to
transfection toxicity, which correlates with transfection efficiency. Although the
previous cyanide resistance assay was done three days post transfection as opposed to the
one day done here, it could be an explanation for the perceived resistance of the
ATP8/6-transfected cells to cyanide.
Lipofec t2000 Miius LT1 Li pofect2000 Miris LT1
U o.-o f i
E 0,20 -
e 1 .00 ---------------------------------
The levels of both ATP6 and COX2 mRNA were measured in RNA extracted five
days post transfection from 293T cells not grown in cyanide. There was no significant
difference in ATP6 or COX2 levels (Fig. 2-22). All of the transfected cells may have less
ATP6 and COX2 than the untransfected cells, but the untransfected-cell values were
highly variable. Again, compared to the other transfected cells, ATP8/6 more closely
resembled the untransfected cells possibly due to low transfection efficiency.
Injection of Mice
If the ribozyme were capable of creating the disease phenotype, it would cause a
loss of photoreceptors and a loss of vision (Chapter 1, Mitochondrial Disease section).
5 ,0 0 -i-----------------------------
0 o0 -. R
Fig. 2-22. Levels of ATP6 and COX2 in transiently transfected 293T cells. Error bars
are the standard deviation of 7-10 PCR reactions from two or three RT
Although the levels of ribozymes delivered to cultured cells may have been insufficient,
it was possible that viral mediated delivery to mouse photoreceptors could result in
higher levels of ribozyme expression. Ribozyme ATP6m252 and its reverse were
higher levels of ribozyme expression. Ribozyme ATP6m252 and its reverse were
packaged in Adeno-associated virus pseudotype 5 (AAV5). Pseudotype AAV5 has
AAV2 terminal repeats (TRs) and AAV5 capsid proteins. AAV5 has been shown to
more efficiently transduce photoreceptors than other AAV serotypes (Auricchio et al.,
2001; Surace et al., 2003). Two sets of mice were injected subretinally with 2 x 1011
AAV genome copies of the ATP6m252rz in the right eye and the m252rev control in the
The ability of the mice to see was monitored by electroretinogram (ERG) every
month for about four months (Fig. 2-23). The ERG measures the electrical response of
the eye to flashes of light. The response includes a negative a-wave and a positive
0 2 months
0 3 months
A D E I J K
M N Q U V V
Fig. 2-23. Electroretinogram results from mice injected with AAV5-ATP6m252rz.
Values are the ratios of the right to left eye A waves at 5 Candela/mm2
Fig. 2-24. Electroretinograms of mouse K at 4 months post injection. A) Left eye
injected with ATP6m252rev. B) Right eye injected with ATP6m252rz.
b-wave (Fig. 2-24). The a-wave represents the response of the photoreceptors (Hood and
Birch, 1990). The b-wave represents the response of the bipolar and Muller cells that are
post-synaptic to the photoreceptors (Witkovsky et al., 1975). If the ribozyme caused loss
of photoreceptors, a decrease in the amplitudes of the a-wave and the b-wave would be
Of the 12 mice (6 in each set), eight retained vision in both eyes (A, D, E, I, N, U, V
and W). Two of these had slightly reduced vision in the left eye compared to the right
eye (N and U), but U was not consistently different. One mouse had reduced vision in
both eyes (M). One mouse completely lost vision in the right eye by one month (J),
andthis was likely due to injection damage. Three mice had reduced vision in the right
eye (K, Q and W). Two of these had a decrease in the right eye by one month and no
change after that (Q and W), which could have been due to injection damage. The
decrease in W was small and only detectable in the a-wave (b-wave not shown). Only
one mouse had a gradual loss of vision in the right eye (K), which is what was expected.
Its ERG wave forms are shown for four months post injection (Fig. 2-24)
After four and one half months, the first set of animals was sacrificed, and eyes
were removed and sectioned. While dissecting the eyes, it was noted that the right eye of
mouse K had a cataract. Although cataracts do not always interfere with the ERG
response, this could be the reason for the decreased vision seen in this eye. Hematoxylin
and Eosin staining of sections revealed multiple areas of this retina with a loss of
photoreceptors (Fig. 2-25) that could also have caused the decrease in ERG response.
The right eye of mouse J that had no ERG response had a total loss of photoreceptors.
J le J right
J left J right
K left A. -K
Fig. 2-25. Sections of retinas from mice J and K. The left eye was injected with
AAV5-ATP6m252rev. The right eye was injected with AAV5-ATP6m252rz.
The photoreceptors can be seen in three layers, the outer segments (OS), inner
segments (IS), and outer nuclear layer (ON). The inner nuclear layer (IN) is
made up of other retinal cells.
The second set of mice was sacrificed after five months for sectioning of their eyes. No
significant thinning of the retina was seen in any of these animals.
MATERIALS AND METHODS
In Vitro Ribozyme Analysis
Synthetic RNA oligonucleotides were ordered for both ribozymes and targets from
Dharmacon, Inc. (Lafayette, CO) and de-protected as described by the manufacturer.
These were used for time courses and multiple-turnover kinetic reactions as described
(Fritz et al., 2002a).
Table 3-1. RNA oligonucleotides used for in vitro ribozyme assays
ATP6m 171rz agcaucugaugagcgcuucggcgcgaaauuuguu
Time course reactions included 50 to 300 nM ribozyme (denatured at 700C for
2 min), 500 nM to 3 [tM target spiked with y-32P (ICN; Costa Mesa, CA) end-labeled
target (such that the target to ribozyme ratio is about 10:1), 40 mM Tris HC1 pH 7.5,
20 mM or 5 mM MgC12 as indicated, 9 mM DTT and 1% RNasin from Promega
(Madison, WI). One 100-[tl reaction was set up at 370C and 10 [tl was removed at each
time point (3n min, where n = 0-6), and the reaction was stopped with an equal volume
formamide dye mix (90% formamide, 50 mM EDTA, 0.05% bromophenol blue and
0.05% xylene cyanol). A separate reaction was set up without ribozyme for negative
controls at the shortest and longest time points.
A time point at which about 15% of the target was cleaved was used for saturation
kinetics. Reactions included 15 nM ribozyme, 300 nM to 1.92 [tM cold target plus 0.12
[tCi y-32P end-labeled target, 40 mM Tris HC1 pH 7.5, 5 mM MgCl2, 4.5 mM DTT and
0.5% Promega (Madison, WI) RNasin. Each reaction was run in a separate tube and
done in duplicate. Again, each reaction was stopped with an equal volume of formamide
dye containing 50 mM EDTA.
The 5S ribosomal RNA with a point mutation creating a Smal site at nucleotide
105 of the RNA sequence was a gift from the laboratory of Dr. Eric Schon at Columbia
University (Magalhaes et al., 1998). The 648 bp EcoRI fragment containing 272 bp
upstream and 250 bp downstream of the 121 bp 5S rRNA sequence was cloned into the
EcoRI and MfeI sites of an AAV2-TR vector. (EcoRI and MfeI have compatible
overhanging ends.) A neomycin resistance cassette was cloned into the SalI site 54 bp
downstream of the EcoRI/MfeI site. The ribozymes were cloned in forward and reverse
orientation into the HindIII site at the very end of the 5S rRNA sequence. The ribozymes
were constructed with overlapping DNA oligonucleotides that have HindIII-compatible
ends when annealed.
Table 3-2. DNA oligonucleotides used for cloning mouse and human ATP6 ribozymes
ATP6h 114rzsense AGCTTGTTGTTCTGATGAGCGCTTCGGCGCGAAATGAGA
ATP6h 114rzantisense AGCTTCTCATTTCGCGCCGAAGCGCTCATCAGAACAACA
ATP6b (h206rzsense) AGCTTATAAGACTGATGAGCCGTTCGCGGCGAAATCAGGA
ATP6brev (h206rzanti) AGCTTCCTGATTTCGCCGCGAACGGCTCATCAGTCTTATA
Mouse NIH3T3 cells were grown in Dulbecco's Modified Eagle's Media (DMEM)
from Mediatech, Inc (Herndon, VA). with 4.5 g/L glucose; 10% Newborn Calf Serum
(NCS); IX antibiotic/antimycotic from Mediatech, Inc (Herndon, VA); 1 mM pyruvate,
unless otherwise stated, and, in later experiments, 50 mg/L uridine. For glucose-free
media DMEM base from Sigma (St. Louis, MO) was used with 3.7 g/L sodium
bicarbonate, 2 mM L-glutamine, either 4.5 g/L glucose or 5 mM galactose, 10% NCS,
antibiotic/antimycotic, pyruvate and sometimes 50 mg/L uridine. To detach cells, they
were treated with 0.5% trypsin EDTA from Mediatech, Inc. (Herndon, VA) in phosphate
buffered saline (PBS). PBS is 130 mM NaCl and 10 mM sodium phosphate, monobasic
and dibasic mixed to pH 7.3. Human cells included 293, 293T and HeLa cells. They
were grown in DMEM media as described for mouse cells except 10% Fetal Bovine
Serum (FBS) was substituted in place of NCS.
All transfections of NIH3T3 cells were on 6-well plates and used serum-free media.
The first transfection used 4 x 105 cells, 3 |tg DNA, 14 [tl Lipofectamine, 6 [tl PLUS
reagent from Invitrogen (Carlsbad, CA). The second transfection was done in two sets
and used 2 x 105 or 3 x 105 cells, 10 or 16 [tl Lipofectamine respectively, 15 |tg DNA and
6[tl PLUS reagent. The third transfection consisted of two different sets. One set used
2 x 105 cells, 12 |tg DNA, 10 [tl Lipofectamine and 3 [tl PLUS reagent and was carried
out by Elizabeth Bongorno resulting in 30% GFP-positive cells the next day. The other
set used 2.5 x 105 cells, 15 |tg DNA, 11 tl Lipofectamine and 2 [tl PLUS reagent and
resulted in 15% GFP-positive cells the next day. The fourth transfection used 2 x 105
cells, 9 |tg DNA, 14 [tl Lipofectamine and 4 [tl PLUS reagent and was carried out by
Diana Levinson. Uridine was added to the media after transfection and for all
experiments on these cells.
Human cell transfections on 6-well plates used 8[tl Lipofectamine 2000 from
Invitrogen (Carlsbad, CA) and serum-free media. The three transfections of HeLa cells
used 5 x 105, 6 x 10 and 4 x 10 cells with 8 |tg, 5 |tg and 5 |tg DNA respectively.
Transfections of 293 cells on 6-well plates used 6 x 105 cells and 5 |tg DNA.
Transfections of 293 cells to look for the presence of the ribozyme in the mitochondria
were done on 6cm dishes with 1 x 106 cells, 10 |tg DNA and 16 [tl Lipofectamine 2000.
RNA was isolated either by TRIzol extraction from Invitrogen (Carlsbad, CA) or
with the Sigma (St. Louis, MO) GenElute Mammalian Total RNA Miniprep Kit. The
Sigma kit was used for all of the assays for the first set (Arabic numerals) of colonies,
except for the measurements of ND2 and ND6 mRNA levels. RNA was extracted from a
10 cm dish of cells and eluted in 90 [tl of water. RT-PCR of ND2 and ND6, all assays on
the Roman numeral colonies and assays on human cells used TRIzol extracted RNA. For
TRIzol extraction, a 10 cm dish of cells was trypsinized as described above and cells
were resuspended in 100 [tl PBS before adding 1 ml TRIzol as Invitrogen recommends.
The remainder of the Invitrogen protocol was followed except that no second
precipitation step was done. If less than a 10 cm dish of cells was harvested, the amount
of TRIzol was usually decreased to 500 [tl or 250 [tl and the other reagents were
Reverse transcription (RT) reactions for detecting 5S transcripts from the first set
of clonal isolates were performed using the Ambion (Austin, TX) Retroscript kit in 20 [tl
reaction volumes, and all other RT reactions used the Amersham (Piscataway, NJ)
First-Strand cDNA Synthesis Kit in 10 [tl reaction volumes (2.5 [tl bulk mix and 0.5 [tl
200 mM DTT). Before adding the RT reaction mix, 50 pmol of a random decamer and
Ipmol of the appropriate antisense primer for the ribozyme were annealed to the RNA
(usually 1 to 2 [tg) at 700C for 5 min.
Polymerase chain reactions (PCR) contained 3 pmol each of the appropriate
primers, 0.2 mM dNTPs, 1X Promega Buffer, 2.5 mM MgCl2, 2U Promega Taq
Polymerase and usually 2 [tl of the RT reaction in a 25 [tl volume. Cycles consist of 30 s
each at 95, 50 and 72C in most cases. Duplicate PCR reactions were carried out for
different numbers of cycles to insure they were still within the linear range.
Table 3-3. DNA oligonucleotides used for RT-PCR
HH antisense 04
mouse beta-actin 5'
mouse beta-actin 3'
The ribozyme transcript levels for the first (Arabic) clonal isolates were detected at
20 to 26 cycles with an annealing temperature of 52C. The ribozyme levels in the
Roman numeral colonies were assayed at 24 and 26 cycles with an annealing temperature
of 50C. For ATP6 mRNA measurements, 12 to 16 cycles were used with an annealing
temperature of 48C. For COX3 and ND4, 14 to 16 cycles and an annealing temperature
of 48C were used. ND2 and ND6 were assayed with 14 to 18 cycles and an annealing
temperature of 50C. Bands were detected on agarose gels stained with ethidum bromide
(EtBr) in most cases. A comparison with radioactively labeled PCR produced ribozyme
to Rb retinoblastomaa) ratios that were half that of EtBr-detected bands, but samples were
consistent relative to each other.
Growth in Galactose
Total selected cells from the first transfection were plated at 3 x 105 cells per 6 cm
dish. Five plates each were set up in 5 mM galactose and 0.4 ng/ml oligomycin. One
plate was trypsinized as described above and counted on a hemacytometer counted every
two days, and the media was replaced on the remaining plates.
Total selected cells from the fourth transfection were plated at 1 x 105 cells per
6 cm dish in two separate experiments. In the first experiment, two galactose and one
glucose plate were counted one day after plating, and another two galactose and one
glucose plates were counted three days after plating. In the second experiment, two
galactose and one glucose plate were counted one day after plating, and two galactose
and two glucose plates were counted three days after plating. Standard deviations were
calculated for the doubling rates between one and three days of growth for both
experiments combined, four replicates in galactose and three in glucose.
The first set of clonal isolates grown in galactose (6 colonies designated with
Arabic numerals) was plated at 2 x 105 cells per 6 cm dish. Each colony was plated on 12
dishes in 5 mM galactose. Each day three plates were counted for each colony. Medium
was replaced every two days. After three days, increased confluence caused a significant
decrease in growth rate, so the numbers for Day four are not shown. Standard deviations
were calculated for the three replicates.
The colonies designated with Roman numerals from the fourth transfection were
screened for their ability to grow in galactose. Once colonies had grown enough, they
were split in half to two wells of a 12-well plate. One well was grown in glucose and one
in galactose for three days. Both wells were counted on the third day and the number in
galactose was compared to the number in glucose.
Galactose growth measurements were repeated for those Roman numeral colonies
that appeared to have a defect in galactose growth. They were plated at 1 x 105 cells per
6 cm dish. Each colony was plated on five dishes in 5 mM galactose and three in
glucose. On Day one, two galactose plates and one glucose plate were counted for each
colony. On Day three, the remaining three galactose plates and two glucose plates were
counted. For some of the colonies, the cells were only counted on Day three, two in
galactose and two in glucose. Error bars are the standard deviation of two to four
Cyanide Resistance Assay
The assay for cyanide (KCN) resistance is essentially the same as the galactose
growth assay, except for the contents of the media and the expected outcome. Cells with
a decrease in cytochrome oxidase should grow better in cyanide. Cells were plated at
1 x 105 cells per 6 cm dish with 250 [tM or 1 mM KCN. For the third transfection cells,
12 plates were set up for each sample. For the first transfection colonies, 18 plates were
set up for each colony, 9 at each cyanide concentration. Three plates were counted every
two days, and the medium was replaced on the remaining plates.
ATP Synthesis Assay
The rate of ATP synthesis was measured in permeabilized cells as described
(Manfredi et al., 2001). A 15 cm dish of cells were trypsinized as described above and
resuspended at 1 x 107 cells/ml in buffer A. Buffer A is composed of 150 mM KC1, 25
mM Tris HC1, 2 mM EDTA, 0.1% BSA, 10 mM K-phosphate, 0.1 mM MgCl2, pH 7.4.
Each reaction consisted of 1.6 x 106 cells, 1 mM ADP, 1.5 mM adenosine pyrophosphate,
10 mM pyruvate, 10 mM malate, 5% buffer B (containing luciferin and luciferase), and
100 [tg/ml digitonin. Buffer B is 800 [tM luciferin from Promega Life Science (Madison,
WI) and 20 [tg/ml luciferase from Roche Applied Sciences (Indianapolis, IN) in 500 mM
Tris Acetate, pH 7.75. Digitonin was added last to start the reaction, and the other
components were added to the cells immediately before the digitonin. Using luciferin
and ATP, luciferase catalyses a reaction that produces a flash of yellow-green light
(DeLuca et al., 1979). The ATP-dependent production of light was detected immediately
in a luminometer reading every 15 to 20 s for 3 to 5 min. Pyruvate and malate are
substrates for complex I. Succinate (50 mM) was used instead of pyruvate and malate to
measure complex II-dependent ATP synthesis. As a negative control, oligomycin was
added to the reactions at 2 [tg/ml. Relative light units detected with the luminometer
were converted to moles of ATP by reading luciferase activity with ATP standards.
Succinate-dependent rates were approximately the same for all, but there was variability
due to the loss of luciferase activity over time. Since the reactions with the different
substrates were done immediately after one another for each sample, using the ratio of the
two rates was able to correct for this difference in luciferase activity.
Mitochondrial RNA Isolation
Mitochondria were isolated from four 15 cm dishes ofNIH3T3 cells or one 15 cm
dish of 293 cells as described (Gaines, III, 1996). Briefly, cells were harvested and
washed twice with six pellet volumes (about 600 [il) of cell wash solution (1 mM Tris pH
7.0, 130 mM NaC1, 5 mM KC1 and 7.5 mM MgCl2). Cells were centrifuged for 5 min at
1500 gav. Washed cells were resuspended in 50 [tl 10X spin solution (35 mM Tris pH
7.7, 20 mM NaC1, 5 mM MgCl2, and 1 mM phenylmethylsulfonyl fluoride, PMSF).
Cells were homogenized with 15 strokes of a Teflon coated pestle. The homogenate was
transferred to a 1.5ml Eppendorftube with 800 [tl 1X spin solution with EDTA (3.5 mM
Tris pH 7.7, 2 mM NaC1, 0.5 mM MgCl2, 0.5% BSA, 20 mM EDTA and 1 mM PMSF).
A low speed spin, 1500 gay for 10 min, was used to pellet unbroken cells, nuclei and other
debris. The supernatant was transferred to a new tube and spun at 1500 gay for 5 min two
The supernatant containing the mitochondria and other cytoplasmic components
was divided in order to subject the mitochondria to different treatments. One fraction,
"n" or no RNase, was spun at 18,000 gay for 5 min to pellet mitochondria and washed
with final wash solution (10 mM Tris pH 7.5, 1 mM EDTA, 6 mM CaC12, 320 mM
sucrose and ImM PMSF) multiple times. The "R" (RNased), "d" (digitonin), and "X"
(Triton X-100) fractions were incubated for 15 min at room temperature with 32.5 |tg
RNase A before pelleting mitochondria. The R mitochondrial pellet was resuspended in
50 [tl final wash solution and treated with 65 |tg RNase A and 20 units of micrococcal
nuclease for 15 min more at room temperature. The micrococcal nuclease was
inactivated by adding 20 mM EGTA to chelate the calcium. The "d" or digitonin fraction
was resuspended in 30 [tl final wash solution with 1 mg/ml digitonin and left to shake at
four degrees Celsius for 15 min. The mitochondria were pelleted for 10 min at 18,000 gav
and then treated with nuclease as described for the R fraction. The two RNase-treated
mitochondria fractions were washed three times with final wash solution containing 20
mM EDTA. The X fraction mitochondrial pellet was resuspended in 50 [tl final wash
solution with 1% Triton X-100. The sample was mixed for a few seconds and then
RNase and micrococcal nuclease were added and incubated for 15 min at room
temperature. EGTA, 0.05% diethylpyrocarbonate (DEPC) and TRIzol were added
directly to extract RNA. (The mitochondria could not be pelleted again, because they
should have been lysed.) All other fractions (n, R and d) were resuspended in 20 [tl final
wash solution with 20 mM EDTA, 0.05% DEPC and 0.1 M sodium acetate, and RNA
was extracted with 250 [tl TRIzol (Invitrogen, Inc.)
The RNA was treated with DNase before performing RT-PCR. To determine if
any cytoplasmic RNA remained, beta-actin primers were used in the RT-PCR mix.
Mouse ND4 primers or human ATP6 primers were used to monitor mitochondrial
transcripts. RT-PCR products were separated on 5% acrylamide, 6 M urea gels. The gels
were stained with 0.01% SYBR Green I from Invitrogen (Carlsbad, CA) and visualized
with a Storm Phosphorimager (GE Healthcare).
Mouse Subretinal Injections
For sub-retinal injections, mouse eyes were dilated with atropine (Bausch & Lomb
Pharmaceuticals, Inc, Tampa, FL) about one hour before and then phenylephrine
(AK-DILATE from Akorn; Buffalo Grove, IL) a few minutes before injection. Injections
were done by Adrian Timmers. The mice in the first set were 36 days old, and those in
the second set were 21 days old. Mice were anesthetized with a 1.2:0.3:3.5 ratio of
ketamine:xylazine:PBS (100 mg/ml ketamine, Fort Dodge Animal Health, Fort Dodge,
IA; 100 mg/ml xylazine from Phoenix Pharmaceutical, Inc., St. Joseph, MO; IX PBS
see tissue culture section above). Both eyes were injected with 0.5 [tl full AAV particles
at 4 x 1011 genomes per til. The right eye received ATP6m252rz and the left eye got
For electroretinograms (ERGs)(Timmers et al., 2001) mice were dark-adapted
overnight and anesthetized with a 1:1:5 ratio of ketamine:xylazine:PBS. Eyes were
dilated with phenylephrine and given a topical anesthetic, proparacaine (Akorn; Buffalo
Grove, IL). Mice were positioned in a dome with a ground electrode in their hind thigh,
reference electrode in the back of the neck and contact electrodes kept in contact with
each eye with hydroxypropyl methylcellulose (Gonak from Eye Supply USA, Inc.;
Tampa, FL). Flashes of light were given with increasing intensity from 0.01 up to 10,000
mCandela/mm2 (TONNIES). The electrical responses were averaged from ten flashes for
each intensity of light. After removing electrodes, mouse eyes were moistened with the
antibacterial, vetropolycin (Pharmaderm, Melville, NY).
Sectioning of Mouse Eyes
For sectioning of eyes, mice were euthenized with an overdose of carbon dioxide or
ketamine and xylazine followed by cervical dislocation. Eyes were removed from the
mice and fixed with 4% glutaraldehyde. The corneas and lenses were removed. The