Hindawi Publishing Corporation

BioMed Research International

Volume 2015, Article ID 305716, 10 pages
http://dx.doi.org/10.1155/2015/305716

Research Article
Efficient Mitochondrial Genome Editing by CRISPR/Cas9

Areum Jo,1 Sangwoo Ham,1 Gum Hwa Lee,2 Yun-Il Lee,3 SangSeong Kim,4 Yun-Song Lee,1
Joo-Ho Shin,1 and Yunjong Lee1
1
Division of Pharmacology, Department of Molecular Cell Biology, Samsung Biomedical Research Institute,
Sungkyunkwan University School of Medicine, Suwon, Gyeonggi-do 440-746, Republic of Korea
2
College of Pharmacy, Chosun University, Gwangju 501-759, Republic of Korea
3
Well Aging Research Center, Samsung Advanced Institute of Technology (SAIT), Yongin-si 446-712, Republic of Korea
4
Department of Pharmacy, Hanyang University, ERICA Campus, 55 Hanyangdaehak-ro, Sannok-gu, Ansan,
Gyeonggi-do 426-791, Republic of Korea

Correspondence should be addressed to Yunjong Lee; ylee69@skku.edu

Received 25 May 2015; Revised 30 July 2015; Accepted 6 August 2015

Academic Editor: Janusz Blasiak

Copyright © 2015 Areum Jo et al. This is an open access article distributed under the Creative Commons Attribution License, which
permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR)/Cas9 system has been widely used for nuclear DNA
editing to generate mutations or correct specific disease alleles. Despite its flexible application, it has not been determined if
CRISPR/Cas9, originally identified as a bacterial defense system against virus, can be targeted to mitochondria for mtDNA editing.
Here, we show that regular FLAG-Cas9 can localize to mitochondria to edit mitochondrial DNA with sgRNAs targeting specific
loci of the mitochondrial genome. Expression of FLAG-Cas9 together with gRNA targeting Cox1 and Cox3 leads to cleavage of
the specific mtDNA loci. In addition, we observed disruption of mitochondrial protein homeostasis following mtDNA truncation
or cleavage by CRISPR/Cas9. To overcome nonspecific distribution of FLAG-Cas9, we also created a mitochondria-targeted Cas9
(mitoCas9). This new version of Cas9 localizes only to mitochondria; together with expression of gRNA targeting mtDNA, there
is specific cleavage of mtDNA. MitoCas9-induced reduction of mtDNA and its transcription leads to mitochondrial membrane
potential disruption and cell growth inhibition. This mitoCas9 could be applied to edit mtDNA together with gRNA expression
vectors without affecting genomic DNA. In this brief study, we demonstrate that mtDNA editing is possible using CRISPR/Cas9.
Moreover, our development of mitoCas9 with specific localization to the mitochondria should facilitate its application for
mitochondrial genome editing.

1. Introduction defense system in aging or disease processes [1]. Damage to

mtDNA, such as point mutations or deletions, contributes to
Mitochondria play roles in many important cellular functions or predisposes individuals to a variety of human diseases. Par-
including ATP production via oxidative phosphorylation, ticularly, dysfunctional mitochondria have been implicated
calcium storage, and regulation of cell death in various cell in several neurodegenerative diseases including Parkinson’s
types [1–3]. Mitochondria contain their own genome, which disease [5, 6]. Indeed, several Parkinson’s disease-associated
encodes 13 proteins that are subunits of respiratory chain genes exhibit pathological interaction with mitochondria [7].
complexes, as well as two rRNAs and 22 mitochondrial In this respect, mitochondrial function is known to
tRNAs [4]. Due to the critical roles of genes encoded by be dysregulated in some disease processes in cells or
mtDNA, maintenance of mitochondrial genome integrity is animals. For example, carbonyl cyanide m-chloro-
quite important for normal cellular functions. Mitochondrial phenyl hydrazone (CCCP), 1-methyl-4-phenyl-1,2,3,6-tetra
DNA are, however, constantly under mutational pressure hydropyridine (MPTP), and rotenone have been widely used
due to oxidative stress imposed by radicals generated by to impair the electron transport chain in order to produce
oxidative phosphorylation or an imbalance in the antioxidant mitophagy or oxidative stress-induced neurodegeneration
2 BioMed Research International

[8, 9]. Since these chemical treatments lack genetic specificity, and lentiCRISPR-sgRNA-eGFP#2 constructs were purchased
direct manipulation of mitochondrial DNA could have from Addgene. Oligonucleotides were ordered from and
potential to generate genetic cell or animal models synthesized by Cosmo Genetech.
mimicking mitochondrial dysfunction. Moreover, once
designed to target mutant mtDNA, selective correction of 2.2. Construction of Mitochondria-Targeting Cas9 and gRNAs
damaged mitochondrial DNA could provide an opportunity Expressing DNA. sgRNA targeting mtDNA was cloned into
for therapeutic application. Indeed, TALEN has been the lentiCRISPR construct according to the instructions
reengineered to localize to mitochondria and specifically posted on the Addgene website (deposited by Dr. Feng
remove truncated dysfunctional mtDNA [10]. Zhang Lab). The oligonucleotides used for construction
Despite the huge potential of mitoTALEN-mediated of each sgRNA are listed in Table 1. NCBI and Ensembl
mtDNA editing, more user-friendly and efficient alternative genome browser database were used to retrieve mitochon-
methods are necessary to overcome difficulties in mtDNA drial genome sequence. Sequence integrity was verified by
modification either for correction of dysfunctional mtDNA sequencing. LentiCRISPR-EGFP sgRNA 2 construct was
or for producing dysfunctional mtDNA in order to create reengineered to mitoCas9. Briefly, a FLAG tag and NLS in
mitochondria-associated disease models. the N-terminus of Cas9 were removed and replaced with
Here we report a novel approach to generate mtDNA the following sequences in order to achieve a mitochondrial
dysfunction with the CRISPR/Cas9 system. Cas9, widely used targeting sequence and HA tag: ATGTCCGTCCTGACG-
for genome editing, showed distribution to mitochondria as CCGCTGCTGCTGCGGGGCTTGACAGGCTCGGCC-
well as the nucleus. Expression of FLAG-Cas9 with gRNAs CGGCGG CTCCCAGTGCCGCGCGCCAAGATCCAT-
designed to target mtDNA resulted in cleavage of mtDNA TCGTTG (MTS), which encodes a mitochondrial targeting
and alterations in mitochondrial integrity as determined by sequence (MTS) from subunit VIII of human cytochrome c
Western blots for some mitochondrial proteins. Moreover, oxidase, and peptides MSVLTPLLLRGLTGSARRLPVPRA-
regular FLAG-Cas9 was modified to contain mitochondrial KIHSL and TACCCATACGACGTCCCAGACTACGCT,
targeting sequence instead of nuclear localization sequence which encode HA peptide YPYDVPDYA. To obtain DNA
(NLS) in order to localize it to mitochondria (namely, expressing gRNA targeting Cox1 region without expression
mitoCas9). MitoCas9 robustly localized to mitochondria; of FLAG-NLS-Cas9, U6 promoter, gRNA targeting sequence,
together with gRNA targeting of mtDNA, specific cleavage of scaffold, and termination signal were PCR-amplified with
mtDNA was observed, demonstrating its functional applica- the following primers using lentiCRISPR-sgRNA-Cox1 con-
tion for mtDNA editing. These results together demonstrate structs as PCR templates. LentiCRISPR-sgRNA-eGFP#2
the successful application of CRISPR/Cas9 in mitochondrial was used as a template for control gRNA PCR. Forward
genome editing and suggest the possibility for in vitro and in primer was as follows: CAGAGAGGGCCTATTTCCCATG.
vivo manipulation of mtDNA in a site-specific manner. Reverse primer was as follows: CTAGAATTCAAAAAAG-
CACCGACTC. PCR products were separated in an agarose
gel and purified using the Qiagen gel extraction kit. Purified
2. Materials and Methods PCR products were used for transient transfection together
with mitoCas9-expressing plasmid.
2.1. Reagent, Plasmids, and Antibodies. For primary anti-
bodies, we used mouse antibody against PARP1 (cat# 2.3. Cell Culture and Transfection. Human embryonic kidney
556494, 1 : 1,000, BD Bioscience), mitochondrial marker (HEK-293T) cells were grown in DMEM containing 10% FBS
antibody sampler kit (cat# 8674, 1 : 3,000, Cell Signal- (vol/vol) and antibiotics in a humidified 5% CO2 /95% air
ing, rabbit antibody to CoxIV, Cytochrome C, HSP60, atmosphere at 37∘ C. For transient transfection, cells were
PHB1, pyruvate dehydrogenase, SDHA, and mouse anti- transfected with indicated target vectors using X-tremeGENE
body to SOD1), mouse antibody against GAPDH (GT239, reagent (Roche) according to the manufacturer’s instructions.
1 : 5,000, GeneTex), mouse antibody against FLAG (clone To assess cell growth, HEK-293T cells were transfected with
M2, Sigma), rabbit antibody against PINK1 (cat# 100- mitoCas9 and gRNA constructs targeting Cox1 and Cox3 or
494, 1 : 2,000, Novus Biologicals), and mouse antibody eGFP as control. 5 days following transfection, cells were
against UQCRC2 (cat# ab14745, 1 : 3,000, Abcam). For passed onto 24-well plates and maintained in pyruvate-
secondary antibodies, we used horseradish peroxidase- deficient complete media. Cell counts were determined by
(HRP-) conjugated mouse antibody against FLAG (cat# trypan blue staining followed by unbiased cell counting via
A8592, 1 : 5,000, Sigma-Aldrich), HRP-conjugated antibody EVE automatic cell counter (NanoEnTeK) at the indicated
against HA (1 : 3,000, Roche), HRP-conjugated mouse anti- time points.
body against 𝛽-actin (AC15, 1 : 10,000, Sigma-Aldrich),
HRP-conjugated sheep antibody against mouse IgG (cat# 2.4. Subcellular Fractionation. HEK-293T cells were subcel-
RPN4301, 1 : 5,000, GE Healthcare), HRP-conjugated donkey lular fractionated into cytosol, mitochondria, and nucleus
antibody against rabbit IgG (cat# RPN4101, 1 : 5,000, GE using the Qproteome Mitochondria Isolation Kit (Qiagen),
Healthcare), Alexa Fluor 488-conjugated donkey antibody following the instructions in the manual. Cytosolic fractions
against mouse IgG (H+L) (cat# A21202, 1 : 1,000, Invitrogen), were further concentrated with acetone precipitation. The
and Alexa Fluor 568-conjugated donkey antibody against purity of each fraction was validated with Western blots using
rabbit IgG (cat# A10042, 1 : 1,000, Invitrogen). LentiCRISPR antibodies to marker proteins for cytosolic (GAPDH, 2% of
BioMed Research International 3

Table 1: Oligonucleotides used for gRNA construction.

mtDNA Forward Reverse

Cox1 CACCGGGCCCAGCTCGGCTCGAATA AAACTATTCGAGCCGAGCTGGGCCC
Cox2 CACCGTATGAGGGCGTGATCATGAA AAACTTCATGATCACGCCCTCATAC
Cox3 CACCGGCCTAGTATGAGGAGCGTTA AAACTAACGCTCCTCATACTAGGCC
ATP8/6 CACCGTCGTCCTTTAGTGTTGTGTA AAACTACACAACACTAAAGGACGAC

Table 2: Real-time PCR primers.

Gene Forward Reverse

Cox1 CGCCGACCGTTGACTATTCT GGGGGCACCGATTATTAGGG
Cox2 TTCATGATCACGCCCTCATA TAAAGGATGCGTAGGGATGG
Cox3 CAGCCCATGACCCCTAACAG TGTGGTGGCCTTGGTATGTG
ND1 TCTCACCATCGCTCTTCTAC TTGGTCTCTGCTAGTGTGGA
GAPDH CATGTGCAAGGCCGGCTTCG CTGGGTCATCTTCTCGCGGT
GAPDH for cDNA AAACCCATCACCATCTTCCAG AGGGGCCATCCACAGTCTTCT
Forward and reverse primers are designed to anneal upstream and downstream regions flanking gRNA target sites for each mtDNA locus or specific genes.

total cytosolic fraction was used), mitochondrial (SDHA, to nitrocellulose membrane for immunoblot experiments.
20% of total mitochondria fraction was used), and nuclear The blotted nitrocellulose membrane was Ponceau (Sigma)
(PARP1, 5% of total nuclear fraction was used) subcellular stained to visualize even transfer of proteins. Immunoblot-
fractions. ting was performed with an antibody of interest with
chemiluminescence visualization (Pierce). The densitometric
2.5. Real-Time Quantitative PCR. Total genomic DNA and analyses of the bands were performed using ImageJ (NIH,
mitochondrial DNA were extracted using DirectPCR tail lysis http://rsb.info.nih.gov/ij/).
buffer (Viagen) supplemented with proteinase K (Roche)
using protocols slightly modified from the manufacturer’s 2.7. Immunofluorescence. HEK-293T cells, 24 hrs after tran-
instructions. Two days or five days after transient transfec- sient transfection with the indicated constructs, fixed in
tion, HEK-293T cells were briefly washed with PBS and lysed 4% paraformaldehyde/PBS (pH 7.4) were blocked with 10%
in 100 𝜇L DirectPCR lysis buffer supplemented with 3 𝜇L donkey serum (Sigma-Aldrich)/PBS plus 0.3% Triton X-100
proteinase K per each well (12-well culture). Proteinase K was and incubated with antibodies against FLAG or HA and
inactivated by incubating samples at 80∘ C for 1 hr following CoxIV. For mitotracker Red labeling of functional mitochon-
incubation at 55∘ C for 2 hrs. DNA extract was diluted 100-fold dria, cells were incubated with mitotracker Red CMXRos
and directly used for real-time quantitative PCR (Rotor-Gene (200 nM, 10 min, Molecular Probes, Invitrogen) before PFA
Q, Qiagen). For cDNA synthesis, five days after transient fixation. After brief washes with PBS, cells on glass plates
transfection of mitoCas9 and corresponding gRNA con- were incubated with corresponding secondary antibodies
structs, total RNA was extracted using Qiagen RNeasy Mini conjugated with fluorescent dyes. After DAPI nuclear stain-
Kit following manufacturer’s instructions. Using total RNA as ing, samples were mounted with mounting solution (Aqua-
a template, cDNA was synthesized (iScript cDNA synthesis Mount, Thermoscientific). Images were obtained using a
kit, Bio-Rad) for downstream real-time PCR application. fluorescent microscope (Zeiss Axiovert 200 M).
Using the primers indicated in Table 2, mtDNA regions
comprising each gene or mitochondria-encoded genes of 2.8. Statistics. Quantitative data are presented as mean ±
interest were amplified. Primers for Cox1 and Cox3 were s.e.m. Statistical significance was assessed via an unpaired
annealed to mtDNA sequences flanking the cleavage sites two-tailed Student’s 𝑡-test for comparison of two groups (con-
targeted by sgRNA-Cox1 and sgRNA-Cox3, respectively, in trol and test) or an ANOVA test and Student-Newman-Keuls
order to monitor specific cleavage of mtDNA. GAPDH was post hoc analysis for comparison among multiple groups
used as an internal loading control to amplify genomic DNA. of more than three as indicated in each figure legend. 𝑃
RTQ PCR was performed using SYBR green master mix values lower than 0.05 were considered to indicate significant
(QuantiFast SYBR Green PCR Kit, Qiagen) according to the difference among groups.
manufacturer’s instructions.
3. Results
2.6. Western Blot. Total protein lysates were prepared
by directly adding 2x sample buffer supplemented with 3.1. CRISPR/Cas9 Distribution to Mitochondria. The Clus-
𝛽-mercaptoethanol (Bio-Rad) to HEK-293T cells briefly tered Regularly Interspaced Short Palindromic Repeats
washed with ice-cold PBS. After boiling the samples for 5 (CRISPR) defense system in bacteria has been successfully
minutes, they were separated on SDS-PAGE and transferred applied to edit specific genomic loci in many species [11].
4 BioMed Research International

CRISPR-associated (Cas) genes together with a single guide 3.3. Alteration of Mitochondrial Proteins Induced by CRISPR/
RNA (sgRNA) enable sequence-specific cleavage and editing Cas9-Mediated mtDNA Cleavage. Next, we examined alter-
of DNA. The currently available CRISPR/Cas9 system uses ations in mitochondria-associated proteins as an indication
nuclear localization sequence- (NLS-) tagged SpCas9 to facil- of disturbance of mitochondria that could be induced by
itate nuclear transfer of SpCas9, where genome editing should CRISPR/Cas9-mediated mtDNA cleavage. When mtDNA
occur [12]. However, the precise subcellular distribution was truncated at the Cox1 and Cox3 loci by CRISPR/Cas9,
has not been examined. To determine in which subcellular transcription of mtDNA heavy strands and thus expression
compartment overexpressed FLAG-tagged SpCas9 is present, of downstream genes from Cox1 locus could be affected
we performed subcellular fractionation into cytoplasmic, (Figure 1(b)). When we monitored several mitochondrial
mitochondrial, and nuclear fractions for HEK-293T cells marker proteins, there were alterations in several proteins:
transiently transfected with lentiCRISPR-sgRNA-eGFP#2 the levels of SDHA, heat shock protein 60 (HSP60), and pro-
construct, which should express FLAG-tagged Cas9 and hibitin 1 (PHB1) decreased, whereas pyruvate dehydrogenase
sgRNA against eGFP. As expected, FLAG-Cas9 was detected (PDH) and superoxide dismutase 1 (SOD1) levels increased
in the nuclear fraction (Figure 1(a)). However, overexpression (Figures 2(a) and 2(b)). However, no significant changes were
of FLAG-Cas9 showed substantial distribution to the other observed in CoxIV or Cytochrome C (CytC) (Figures 2(a)
compartments, cytoplasm and mitochondria (Figure 1(a)). and 2(b)).
Consistent with the results from subcellular fractionation, Due to the distribution of tRNA and peptide-encoding
immunofluorescence for overexpressed FLAG-Cas9 demon- regions and bidirectional transcriptions of mtDNA, cleavage
strates its colocalization with mitotracker red-labeled mito- at different loci of mtDNA could exert differential outcomes
chondria and DAPI-stained nucleus (Figure 1(a)). (Figure 2(c)). In this regard, we also examined whether
cleavage of different mtDNA single loci led to differen-
tial regulation of mitochondria-associated proteins. HEK-
3.2. CRISPR/Cas9-Mediated mtDNA Cleavage. Based upon 293T cells were transfected with sgCox1, sgCox2, sgCox3,
this result showing FLAG-Cas9 distribution to mitochon- or sgATPase8/6. Each sgRNA transfection exerted differ-
dria, we hypothesized that the CRISPR/Cas9 system could ential effects on mitochondria-associated proteins. Cleav-
be implemented to edit mitochondrial DNA (mtDNA) age of Cox1 or Cox3 or ATP8/6 regions led to slight
when combined with sgRNA targeting specific loci of reduction in ubiquinol-cytochrome c reductase core pro-
mtDNA. To test this possibility, we created several sgRNA tein 2 (UQCRC2) and PHB1, whereas PDH level increased
targeting mtDNA regions encoding peptides (Cox1 and slightly (Figure 2(d)). Interestingly, mitochondria-associated
Cox3) that function with oxidative phosphorylation com- PD gene PTEN-induced putative kinase 1 (PINK1) level
plexes (Figure 1(c)). Oligonucleotides designed to target decreased when Cox1 or Cox3 loci were cleaved, while
each mtDNA region (Table 1) were cloned into lentiCRISPR cleavage of Cox2 regions resulted in an increase in PINK1
plasmid, which can express sgRNA under the U6 pro- level (Figure 2(d)). Mitofusin-2 (Mfn2) for mitochondrial
moter and SpCas9 under the EFS promoter. To exam- fission decreased when Cox1, Cox3, or ATP8/6 loci were
ine whether the CRISPR/Cas9 system can cleave mtDNA cleaved by the CRISPR/Cas9 system (Figure 2(d)). Together,
specifically as it does for genomic DNA, we transfected these results indicate that CRISPR/Cas9-mediated mtDNA
HEK-293T cells with both lentiCRISPR-sgRNA-Cox1 and cleavage results in an alteration of mitochondrial protein
lentiCRISPR-sgRNA-Cox3. These constructs should cleave homeostasis. Moreover, depending on the loci of mtDNA
mtDNA and generate mtDNA fragments. LentiCRISPR- cleavage, the pattern of dysregulation of mitochondrial pro-
sgRNA-eGFP#2 was used as a control. Specific cleavage of teins in early response to mtDNA cleavage differs. Changes in
Mfn2 or PINK1 suggest a potential compensatory response
mtDNA was evaluated using primers flanking the cleavage
for mitochondria quality control in cells following mtDNA
sites (primers for Cox1 and Cox3; please refer to Figure 1(b)
damage.
and Table 2 for detailed information on primers and anneal-
ing loci). According to real-time PCR, there was an almost
3.4. Mitochondria-Targeted Cas9. Although we demon-
90% reduction in intact mtDNA for the Cox1 locus in strated that FLAG-Cas9 with NLS peptide can effectively
HEK-293T cells transfected with both lentiCRISPR-sgRNA- localize to mitochondria for functional editing of mtDNA,
Cox1 and lentiCRISPR-sgRNA-Cox3 compared to control Cas9, which can specifically localize to mitochondria, is
cells transfected with lentiCRISPR-sgRNA-eGFP#2 plasmid required for safe application for mtDNA editing without
(Figure 1(d)). We also observed 80% cleavage for Cox3 locus affecting genomic DNA. Therefore, we aimed at creating
with lentiCRISPR-sgRNA-Cox1 and lentiCRISPR-sgRNA- a new version of Cas9 that can specifically target mtDNA.
Cox3 transfection (Figure 1(d)). The noncleaved site (ND1 By removing FLAG and NLS sequences from lentiCRISPR-
locus) with our lentiCRISPR constructs was also monitored sgRNA-eGFP#2 and adding mitochondrial targeting
using specific primers (Table 2). Interestingly, the mtDNA sequence (MTS) and HA tag in the N-terminus of SpCas9,
copy number as determined by ND1 site RTQ PCR was we synthesized mitochondria-targeting Cas9 (mitoCas9)
increased by approximately 50% (Figure 1(d)). These results (Figure 3(a)). MTS of cytochrome C (mitochondria matrix
indicate that FLAG-Cas9 can actually move into the matrix protein) was used to direct mitoCas9 into the matrix of
of mitochondria, and, when combined with sgRNA targeting mitochondria so that it can interact with mtDNA. MitoCas9
mtDNA, it cleaves mtDNA in a site-specific manner. was expressed and detected with HA-specific antibody at
BioMed Research International 5

FLAG-NLS-Cas9
Cyt Mit Nu

FLAG
−150

FLAG-NLS-Cas9
FLAG Mitotracker
PARP1
−100

SDHA −75

−37
GAPDH

DAPI Merge

Ponceau

25 𝜇m

(a) (b)
Heavy strands
D Loop 12S rRNA
16S rRNA
Light strands
∗∗∗
bt
Cy

1.6
ND
6
ND

1.2
Relative copy number
ND2

Human mtDNA
ND5

16,569 bp
0.8

sgRNA
Co
4
ND

0.4
∗∗∗
x1

DS
∗∗∗
R
4L

Co
K
D

x2 0
N

3 G Cox1 Cox3 ND1

ND ATP
/6 8
Cox3
sgRNA control
sgRNA sgRNA-Cox1 + Cox3
(c) (d)
Figure 1: FLAG-NLS-Cas9 localizes to mitochondria. (a) Subcellular localization of FLAG-Cas9 assessed in the cytosolic (Cyt), mitochondrial
(Mit), and nuclear (Nu) fractions of HEK-293T cells transfected with lentiCRISPR-sgRNA-eGFP#2 and monitored by Western blot.
Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as a cytosolic marker, poly(ADP-ribose) polymerase 1 (PARP1) served as a
nuclear marker, and succinate dehydrogenase complex subunit A (SDHA) served as a mitochondrial marker. Ponceau staining of the blotted
nitrocellulose membrane was presented in the bottom lane to visualize relative protein loading amounts. (b) Immunofluorescence images of
FLAG-NLS-Cas9 demonstrating its localization to nucleus, cytoplasm, and mitochondria. HEK-293T cells were immunostained with mouse
antibody to FLAG (green) 24 hours following transient transfection with FLAG-NLS-Cas9 construct. Mitochondria or nucleus was stained
with mitotracker Red or DAPI, respectively. (c) Illustration of human mitochondrial DNA (mtDNA). Simplified view of mtDNA is presented
to depict regions encoding peptides or tRNA for indicated amino acids. Filled triangles indicate sites targeted by gRNAs against Cox1 or Cox3.
Arrows indicate primer annealing sites used for real-time PCR of Cox1 or Cox3 regions. (d) Quantification of copy numbers for Cox1, Cox3,
and ND1 regions of mtDNA extracted from HEK-293T cells transiently transfected with lentiCRISPR-sgRNA-Cox1 and lentiCRISPR-sgRNA-
Cox3 or lentiCRISPR-sgRNA-eGFP#2 as a control, determined by real-time quantitative PCR using primers listed in Table 2. GAPDH was
used as an internal loading control (𝑛 = 3 per group). Quantified data (d) are expressed as mean ± s.e.m., ∗∗∗ 𝑃 < 0.001, unpaired two-tailed
Student’s t-test.
6 BioMed Research International

sgRNA-Cox1, Cox3
sgRNA-eGFP
1 2 3 1 2 3
FLAG −150

SDHA −75

HSP60 −60
3.5
∗∗
PDH −37 3

Relative protein levels

2.5
PHB1 −30
2 ∗∗
1.5
SOD −30 ∗∗∗ ∗ ∗
1
CoxIV −15 0.5
0
CytC −12 FLAG SDHA HSP60 PDH PHB1 SOD CoxIV CytC

sgRNA control
𝛽-actin −37
sgRNA-Cox1 + Cox3
(a) (b)

ATP 8/6
eGFP

Cox1

Cox2

Cox3
Heavy strands sgRNA
D Loop
12S rRNA
Mfn2 −80
Light strands 16S rRNA
bt
Cy

PINK1
ND

−65
6

1
ND

ND2

UQCRC2 −48
ND5

sgRNA PHB1 −30

Co
4
ND

x1

DS
R −37
4L

Co PDH
K
D

x2 sgRNA
N

3 G ATP
ND /6 8
Cox3 sgRNA 𝛽-actin
−42
sgRNA
(c) (d)

Figure 2: Alterations in mitochondria-associated proteins following CRISPR/Cas9-mediated mtDNA editing. (a) Mitochondrial proteins in
HEK-293T cells after CRISPR/Cas9-mediated cleavage of mtDNA at Cox1 and Cox3 loci as determined by Western blots using indicated
antibodies. 𝛽-actin was used as a loading control. (b) Quantification of mitochondria proteins in HEK-293T cells transfected with
lentiCRISPR-sgRNA-Cox1 + Cox3 or lentiCRISPR-sgRNA-eGFP#2 control as shown in panel (a) normalized to 𝛽-actin. (c) Illustration
of human mtDNA. Specific loci targeted by lentiCRISPR-sgRNAs (Cox1, Cox2, Cox3, and ATP8/6) are indicated with filled triangles. (d)
Representative Western blots showing differential alteration of mitochondrial proteins following cleavage of specific mtDNA loci mediated
by indicated sgRNAs in HEK-293T cells. Quantified data (b) are expressed as mean ± s.e.m., ∗ 𝑃 < 0.05, ∗∗ 𝑃 < 0.01, and ∗∗∗ 𝑃 < 0.001,
unpaired two-tailed Student’s t-test.
BioMed Research International 7

MTS HA
MSVLTPLLLRGLTGSARRLPVPRAKIHSL YPYDVPDYA

𝛽-gal mitoCas9

sgRNA-eGFP SpCas9 HA −150

EFS

U6
mitoCas9 𝛽-actin
−42

(a) (b)
mitoCas9
mitoCas9 HA CoxIV
Cyt Mit Nu

HA
−150

PARP1
−100
DAPI Merge

−75
SDHA

−37
GAPDH

(c) (d)
sgRNA-eGFP 1.4
U6 sgRNA-Cox1
1.2
sgRNA scaffold ∗∗∗
U6-sgRNA-eGFP

U6-sgRNA-Cox1

Relative copy number

1
No primer

0.8
Ladder

0.6

0.4

0.2

0
Cox1 Cox3 ND1
500
mitoCas9 + U6-sgEGFP
(bp)

300 mitoCas9
U6-sgEGFP mitoCas9 + U6-sgCox1
100
U6-sgCox1
(e) (f)
Figure 3: Construction of mitochondria-targeted MTS-HA-Cas9. (a) Schematic illustration of mitochondria-targeting Cas9 (mitoCas9).
Mitochondria-targeting sequence (MTS) and HA tag information are presented. (b) Expression of mitoCas9 in HEK-293T cells transiently
transfected with MTS-HA-Cas9 construct determined by Western blots using HA antibody. 𝛽-actin was used as a loading control. (c)
Subcellular localization of MTS-HA-Cas9 assessed in the cytosolic (Cyt), mitochondrial (Mit), and nuclear (Nu) fractions of HEK-293T
cells transfected with lentiCRISPR-sgRNA-eGFP#2 and monitored by Western blot. GAPDH served as a cytosolic marker, PARP1 served as
a nuclear marker, and SDHA served as a mitochondrial marker. (d) Representative immunofluorescence microscopic image of MTS-HA-
Cas9 (HA, green) and CoxIV (red) subcellular distributions in HEK-293T cells transfected with MTS-HA-Cas9 construct. The merged image
(yellow, right panel) shows colocalization of mitoCas9 and CoxIV. (e) Representative gel images of the PCR product of hU6-sgRMA-eGFP
and hU6-sgRNA-Cox1 which were purified by gel extraction (bottom panel). Schematics of primers and lentiCRISPR-sgRNA templates used
to amplify U6 promoter and respective sgRNA components for transfection (upper panel). (f) Quantification of copy numbers for Cox1, Cox3,
and ND1 regions of mtDNA extracted from HEK-293T cells transiently transfected with indicated constructs (mitoCas9 is a plasmid, while
U6-sgRNAs are PCR product.), determined by real-time quantitative PCR using primers listed in Table 2. GAPDH was used as an internal
loading control (𝑛 = 3 per group). Quantified data (b) are expressed as mean ± s.e.m., ∗∗∗ 𝑃 < 0.001, analysis of variance (ANOVA) test
followed by Student-Newman-Keuls post hoc analysis.
8 BioMed Research International

the expected molecular weight (Figure 3(b)), and subcellular Cox3 exhibit decelerated growth rate compared to control cell
fractionation showed that mitoCas9 mainly localized to line (Figure 4(d)).
mitochondria (Figure 3(c)). In addition, immunofluores- Taken together, our results demonstrate that mitoCas9-
cence confirmed colocalization of HA-tagged mitoCas9 with induced mtDNA damage leads to functional defects of mito-
mitochondria marker protein CoxIV (Figure 3(d)). chondria and ultimately causes reduction in cell proliferation
capacity.
3.5. Mitochondrial DNA Editing by mitoCas9. To determine
whether this new version of Cas9 is functional for mtDNA 4. Discussion
editing, we transfected HEK-293T cells with both mitoCas9
We discovered that CRISPR/Cas9-mediated genome editing
and DNA expressing sgRNA for Cox1. For this application, can be successfully employed to manipulate the mitochon-
we PCR-amplified hU6 promoter and sgRNA-Cox1 regions drial genome. This approach still needs further study to
from lentiCRISPR-sgRNA-Cox1 plasmid (Figure 3(e)). U6- understand how gRNA can be translocated into the mito-
sgRNA-eGFP was also PCR-amplified to be used as control chondria matrix together with mitochondria-localizing Cas9.
(Figure 3(e)). As expected, mitoCas9 and hU6-sgRNA-Cox1 It is interesting to note their tropism to mitochondria even
robustly cleaved the targeted region of the mtDNA as deter- with artificially added nuclear localizing sequence. Given the
mined by real-time quantitative PCR only at Cox1 targeted wide application of CRISPR/Cas9 in genome manipulation,
locus (Figure 3(f)). Taken together, our results demonstrate researchers should be careful when studying mitochon-
that mitochondria-targeting mitoCas9 can be applied to edit dria function because Cas9 can localize to mitochondria
mtDNA in conjunction with custom-designed sgRNAs. for unwanted mtDNA editing depending on the sgRNA
sequences used for the study. Pseudogenes for genomic Cox1
3.6. Mitochondrial Dysfunction Induced by mitoCas9-Medi- and Cox3 are present and they may have been affected
ated mtDNA Damage. Mitochondria targeting of certain with expression of FLAG-NLS-Cas9 targeting mtCox1 and
restriction enzymes such as EcoRI or PstI has been employed mtCox3 regions.
to destroy mitochondrial DNA [13, 14] together with chemical One group has reported generation and characterization
induction of mtDNA elimination. To determine whether of knock-in mice expressing Cas9 [15]. This transgenic line
mitoCas9-mediated site-specific mtDNA cleavage leads to was successfully used to knockout genes of interests in
functional disruption of mitochondria, we first examined brains, lungs, and other tissues depending on the injection
mtDNA copy number and several mitochondrial gene sites of gRNA-expressing viral vectors. Although further
expressions. 5 days following transient transfection of HEK- characterization of Cas9 subcellular localization is required
293T cells with mitoCas9 and sgRNA targeting Cox1 and in vivo, our data suggest that Cas9-expressing transgenic mice
Cox3 regions, real-time PCR quantification of mtDNA copy can be manipulated to affect the mitochondrial genome with
number was performed using ND1 primers. Cell division for viral injection of gRNA targeting specific mtDNA loci.
5 days with site-specific mitoCas9 cleavage at Cox1 and Cox3 Mitochondrial sequence-specific cleavage including
regions leads to approximate 30% reduction of mtDNA copy truncation is possible with the CRISPR/Cas9 system. This
numbers (Figure 4(a)). Consistent with the fact that mtDNA approach could be applied to study functional roles of
was cleaved and steady state mtDNA copy number was mitochondria DNA damage repair protein complexes.
reduced by CRISPR/Cas9 system, the amounts of messenger DNA damage repair machinery has been reported in
RNAs encoding mitochondrial ND1, Cox1, and Cox2 genes mitochondria although its repair efficiency is known to
were reduced by about 50% compared to those of sgRNA be limited compared to that of nuclear genomic DNA
control HEK-293T cells (Figure 4(b)). repair [16]. Despite identification of several mitochondrially
localized repair enzymes, their actual roles in maintenance of
Studies have shown that mitochondrial dysfunction
mitochondrial DNA remain to be elucidated. In this respect,
induced by mtDNA damage results in alterations on mito-
our discovery of the role of CRISPR/Cas9 in mtDNA editing
chondrial membrane potential and cell growth capacity. To provides a tool to determine how mtDNA cleavage can be
assess whether mitoCas9 medicated mtDNA cleavage and handled and repaired. According to our results, once cleavage
mtDNA reduction affect mitochondrial and cellular physi- on specific loci of mtDNA and truncation is achieved by
ology, we measured mitochondria membrane potential by the CRISPR/Cas9 system, there seems to be compensatory
using mitotracker Red which stains functional mitochondria amplification of intact circular mtDNA as determined by
in a membrane potential sensitive manner. Consistent with an increase in the copy number of uncut loci. During 2-3
the notion that mitochondrial gene expression and proteomic days of CRISPR/Cas9-mediated cleavage, the cleaved linear
homeostasis are disturbed by mitoCas9-mediated mtDNA mtDNA appears to be present. To understand the repair
damage, mitochondrial membrane potential was impaired process or clearance of cleaved mtDNA, longer time points
for HEK-293T cells transfected with mitoCas9 and sgRNA after expression of CRISPR/Cas9 should be examined.
targeting Cox1 and Cox3 compared to that of control cells Mitochondria-targeting Cas9 combined with mtDNA-
(Figure 4(c)). Cell proliferation rates were also examined for specific sgRNA provides an opportunity to generate mutant
these mitoCas9 induced HEK-293T cells of mitochondrial mtDNA. Design of a conventional method to study mutant
dysfunction. When monitored for four consecutive days after mtDNA in cells has been challenging [17]. By controlling the
seeding, HEK-293T cells transfected with sgRNA to Cox1 and expression of mitochondrial Cas9 and selecting single cell
BioMed Research International 9

1.4 1.4
Relative ND1 copy number

1.2 1.2

Relative mRNA levels

1 1
0.8 ∗ 0.8
0.6 ∗ ∗∗∗
0.6 ∗∗∗
0.4 0.4
0.2 0.2
0 0
2 5 ND1 Cox1 Cox2
(d)
gRNA control
gRNA control gRNA-Cox1, Cox3
gRNA-Cox1, Cox3
(a) (b)
gRNA control gRNA-Cox1, Cox3
Mitotracker red

25

20
Cell counts (×100,000)

60
15
50
Mitotracker intensity

40 10
∗∗∗
30
∗∗∗
20 5
∗∗
10

0 0
Day 1 Day 2 Day 3 Day 4
gRNA control
gRNA control
gRNA-Cox1, Cox3
gRNA-Cox1, Cox3
(c) (d)

Figure 4: MitoCas9-induced mtDNA damage leads to mitochondria dysfunction. (a) Quantification of copy numbers for ND1 region
of mtDNA extracted from HEK-293T cells transiently transfected with indicated constructs (gRNA control, mitoCas9 and U6-sgRNA to
eGFP; gRNA Cox1, Cox3, mitoCas9, and U6-sgRNA to Cox1 and Cox3), determined by real-time quantitative PCR using primers listed
in Table 2 at the indicated time points (2 days and 5 days after transfection). GAPDH was used as an internal loading control (𝑛 = 3 per
group). (b) Quantification of messenger RNAs for ND1, Cox1, and Cox2 genes in HEK-293T cells 5 days following transient transfection
with lentiCRISPR-sgRNA-Cox1 + Cox3 or lentiCRISPR-sgRNA-eGFP#2 control determined by real-time PCR and normalized to GAPDH
(𝑛 = 3 per group). (c) Representative images of mitotracker Red staining for functional mitochondria in HEK-293T cells transfected with the
indicated constructs (upper panel). Quantification of relative mitotracker Red staining intensities in two groups was shown in the bottom
panel (𝑛 = 50 mitochondria from three independent cultures per group). (d) Cell counts demonstrating proliferation rate of the HEK-293T
cells following cleavage of specific mtDNA loci (𝑛 = 6 individual measurements per group). Days indicate the time intervals proceeded after
cell seeding following 5 days of transient transfectin. Quantified data are expressed as mean ± s.e.m., ∗ 𝑃 < 0.05, ∗∗ 𝑃 < 0.01, and ∗∗∗ 𝑃 < 0.001,
unpaired two-tailed Student’s t-test.

lines with desired mutations, it is possible to create a cell We also showed that the CRISPR/Cas9 system-mediated
model that contains normal and dysfunctional mitochondria cleavage of different loci of mtDNA initiates differential alter-
in order to study the role of specific mtDNA mutations in ations of mitochondria-associated proteins. Mitochondrial
related human diseases. transcription involves polycistronic expression of peptides,
10 BioMed Research International

tRNA, and rRNA from several long RNAs, either heavy References
strands or light strands. In this regard, it is not surprising
that there is differential disruption of the mitochondrial [1] I. Bratic and A. Trifunovic, “Mitochondrial energy metabolism
and ageing,” Biochimica et Biophysica Acta, vol. 1797, no. 6-7, pp.
transcriptome depending on the cleavage loci. The alterations
961–967, 2010.
in mitochondria-associated proteins assessed in this study
may reflect complex characteristics of the mitochondrial pro- [2] E. Carafoli, “Historical review: mitochondria and calcium: ups
and downs of an unusual relationship,” Trends in Biochemical
teome and response to mtDNA damage. Time-course mon-
Sciences, vol. 28, no. 4, pp. 175–181, 2003.
itoring of the mitochondrial proteome after CRISPR/Cas9-
[3] S. W. G. Tait and D. R. Green, “Mitochondria and cell death:
mediated mtDNA cleavage may provide some clues on
outer membrane permeabilization and beyond,” Nature Reviews
how mitochondrial function is regulated against mtDNA Molecular Cell Biology, vol. 11, no. 9, pp. 621–632, 2010.
damage.
[4] J.-W. Taanman, “The mitochondrial genome: structure, tran-
Successful demonstration of mtDNA editing by the scription, translation and replication,” Biochimica et Biophysica
CRISPR/Cas9 system suggests its therapeutic potential. Acta—Bioenergetics, vol. 1410, no. 2, pp. 103–123, 1999.
Mitochondria-targeting TALEN (mitoTALEN) has been [5] M. Filosto, M. Scarpelli, M. S. Cotelli et al., “The role of mito-
shown to selectively remove mutation-harboring mtDNA chondria in neurodegenerative diseases,” Journal of Neurology,
[10]. In a similar fashion, with corresponding sgRNA design vol. 258, no. 10, pp. 1763–1774, 2011.
targeting frequent mtDNA mutations, CRISPR/Cas9 could be [6] A. Johri and M. F. Beal, “Mitochondrial dysfunction in neu-
used to treat cells harboring mutant mtDNAs. Particularly, rodegenerative diseases,” The Journal of Pharmacology and
mitoCas9 can reduce nonspecific targeting of genomic DNA, Experimental Therapeutics, vol. 342, no. 3, pp. 619–630, 2012.
thus alleviating any potential side effects during therapeutic [7] D. J. Moore, A. B. West, V. L. Dawson, and T. M. Dawson,
application compared to that with regular Cas9, which “Molecular pathophysiology of Parkinson’s disease,” Annual
demonstrates nonspecific subcellular distribution. Review of Neuroscience, vol. 28, pp. 57–87, 2005.
[8] E. Deas, N. W. Wood, and H. Plun-Favreau, “Mitophagy
and Parkinson’s disease: the PINK1-parkin link,” Biochimica et
5. Conclusions Biophysica Acta, vol. 1813, no. 4, pp. 623–633, 2011.
[9] T. N. Martinez and J. T. Greenamyre, “Toxin models of mito-
In this brief report, we first establish a proof of concept model chondrial dysfunction in Parkinson’s disease,” Antioxidants and
which demonstrates that CRISPR/Cas9 can be employed to Redox Signaling, vol. 16, no. 9, pp. 920–934, 2012.
target specific loci of mitochondrial genome. We showed [10] S. R. Bacman, S. L. Williams, M. Pinto, S. Peralta, and C.
regular Cas9 nuclear enzyme has a tendency to localize T. Moraes, “Specific elimination of mutant mitochondrial
to mitochondria and somehow gRNA can be transported genomes in patient-derived cells by mitoTALENs,” Nature
together with mitochondrial matrix for its function. Given Medicine, vol. 19, no. 9, pp. 1111–1113, 2013.
the vast amount of CRISPR/Cas9 tools available, our discov- [11] P. D. Hsu, E. S. Lander, and F. Zhang, “Development and
ery of Cas9 editing of mitochondrial genome will provide applications of CRISPR-Cas9 for genome engineering,” Cell, vol.
a new avenue of its application to mitochondria study. 157, no. 6, pp. 1262–1278, 2014.
Additionally, we also provided reengineering of the regular [12] L. Cong, F. A. Ran, D. Cox et al., “Multiplex genome engineering
Cas9 to restrict its localization to mitochondria matrix. This using CRISPR/Cas systems,” Science, vol. 339, no. 6121, pp. 819–
mitochondria-targeting Cas9 (mitoCas9) is functional and 823, 2013.
cleaves mtDNA at the site guided by specific sgRNA. Viral [13] A. Kukat, C. Kukat, J. Brocher et al., “Generation of 𝜌0 cells
production or animal model generation of mitoCas9 would utilizing a mitochondrially targeted restriction endonuclease
be beneficial to expand these findings to in vivo studies of and comparative analyses,” Nucleic Acids Research, vol. 36, no.
mitochondria. 7, article e44, 2008.
[14] S. Srivastava and C. T. Moraes, “Double-strand breaks of mouse
muscle mtDNA promote large deletions similar to multiple
Conflict of Interests mtDNA deletions in humans,” Human Molecular Genetics, vol.
14, no. 7, pp. 893–902, 2005.
The authors declare that there is no conflict of interests [15] R. J. Platt, S. Chen, Y. Zhou et al., “CRISPR-Cas9 knockin mice
regarding the publication of this paper. for genome editing and cancer modeling,” Cell, vol. 159, no. 2,
pp. 440–455, 2014.
[16] N. B. Larsen, M. Rasmussen, and L. J. Rasmussen, “Nuclear and
Authors’ Contribution mitochondrial DNA repair: similar pathways?” Mitochondrion,
vol. 5, no. 2, pp. 89–108, 2005.
Joo-Ho Shin and Yunjong Lee contributed equally to this [17] N. S. Chandel and P. T. Schumacker, “Cells depleted of mito-
paper. chondrial DNA (𝜌0) yield insight into physiological mecha-
nisms,” FEBS Letters, vol. 454, no. 3, pp. 173–176, 1999.

Acknowledgment
This work was supported by grants from Samsung Biomedical
Research Institute (SBRI, SMX1151191).