scispace - formally typeset
Open AccessJournal ArticleDOI

Increasing the genome-targeting scope and precision of base editing with engineered Cas9-cytidine deaminase fusions

TLDR
Five C to T (or G to A) base editors that use natural and engineered Cas9 variants with different protospacer-adjacent motif (PAM) specificities to expand the number of sites that can be targeted by base editing 2.5-fold are reported.
Abstract
Base editing induces single-nucleotide changes in the DNA of living cells using a fusion protein containing a catalytically defective Streptococcus pyogenes Cas9, a cytidine deaminase, and an inhibitor of base excision repair. This genome editing approach has the advantage that it does not require formation of double-stranded DNA breaks or provision of a donor DNA template. Here we report the development of five C to T (or G to A) base editors that use natural and engineered Cas9 variants with different protospacer-adjacent motif (PAM) specificities to expand the number of sites that can be targeted by base editing 2.5-fold. Additionally, we engineered base editors containing mutated cytidine deaminase domains that narrow the width of the editing window from ∼5 nucleotides to as little as 1-2 nucleotides. We thereby enabled discrimination of neighboring C nucleotides, which would otherwise be edited with similar efficiency, and doubled the number of disease-associated target Cs able to be corrected preferentially over nearby non-target Cs.

read more

Content maybe subject to copyright    Report

Increasing the genome-targeting scope and
precision of base editing with engineered Cas9-
cytidine deaminase fusions
Citation
Kim, Y. Bill, Alexis C. Komor, Jonathan M. Levy, Michael S. Packer, Kevin T. Zhao, and David R.
Liu. 2017. “Increasing the genome-targeting scope and precision of base editing with engineered
Cas9-cytidine deaminase fusions.” Nature biotechnology 35 (4): 371-376. doi:10.1038/nbt.3803.
http://dx.doi.org/10.1038/nbt.3803.
Published Version
doi:10.1038/nbt.3803
Permanent link
http://nrs.harvard.edu/urn-3:HUL.InstRepos:34375291
Terms of Use
This article was downloaded from Harvard University’s DASH repository, and is made available
under the terms and conditions applicable to Other Posted Material, as set forth at http://
nrs.harvard.edu/urn-3:HUL.InstRepos:dash.current.terms-of-use#LAA
Share Your Story
The Harvard community has made this article openly available.
Please share how this access benefits you. Submit a story .
Accessibility

Increasing the genome-targeting scope and precision of base
editing with engineered Cas9-cytidine deaminase fusions
Y. Bill Kim
1,2
, Alexis C. Komor
1,2
, Jonathan M. Levy
1,2
, Michael S. Packer
1,2
, Kevin T.
Zhao
1,2
, and David R. Liu
1,2,3,*
1
Department of Chemistry and Chemical Biology, Harvard University, Cambridge, MA, 02138
2
Howard Hughes Medical Institute, Harvard University, Cambridge, MA, 02138
3
Broad Institute of Harvard and MIT, Cambridge, MA, 02142
Abstract
Base editing is a recently developed approach to genome editing that uses a fusion protein
containing a catalytically defective
Streptococcus pyogenes
Cas9, a cytidine deaminase, and an
inhibitor of base excision repair to induce programmable, single-nucleotide changes in the DNA of
living cells without generating double-strand DNA breaks, without requiring a donor DNA
template, and without inducing an excess of stochastic insertions and deletions
1
. Here we report
the development of five new CT (or GA) base editors that use natural and engineered Cas9
variants with different protospacer-adjacent motif (PAM) specificities to expand the number of
sites that can be targeted by base editing by 2.5-fold. Additionally, we engineered new base editors
containing mutated cytidine deaminase domains that narrow the width of the apparent editing
window from approximately 5 nucleotides to as little as 1 to 2 nucleotides, enabling the
discrimination of neighboring C nucleotides that would previously be edited with comparable
efficiency, thereby doubling the number of disease-associated target Cs that can be corrected
preferentially over nearby non-target Cs. Collectively, these developments substantially increase
the targeting scope of base editing and establish the modular nature of base editors.
Keywords
Base editing; genome editing; CRISPR; Cas9; protein engineering; genetic disease; single-
nucleotide polymorphism
Users may view, print, copy, and download text and data-mine the content in such documents, for the purposes of academic research,
subject always to the full Conditions of use: http://www.nature.com/authors/editorial_policies/license.html#termsReprints and
permissions information is available at www.nature.com/reprints.
*
Correspondence should be addressed to David R. Liu: drliu@fas.harvard.edu.
Author Contributions
Y.B.K., A.C.K., J.M.L. and K.Z. conducted the experiments. M.S.P. performed computational analyses. D.R.L supervised the research.
All authors contributed to writing the manuscript.
Author Information
Plasmids encoding SaBE3, SaKKH-BE3, VQR-BE3, EQR-BE3, VRER-BE3, YE1-BE3, EE-BE3, YE2-BE3, and YEE-BE3 are
available from Addgene (plasmids 85169, 85170, 85171, 85172, 85173, 85174, 85175, 85176, 85177). High-throughput sequencing
data will be deposited in the NCBI Sequence Read Archive.
The authors declare competing financial interests: D.R.L. is a consultant and co-founder of Editas Medicine, a company that seeks to
develop genome-editing therapeutics. Y.B.K., A.C.K., and D.R.L. have filed patent applications on base editing.
HHS Public Access
Author manuscript
Nat Biotechnol
. Author manuscript; available in PMC 2017 August 13.
Published in final edited form as:
Nat Biotechnol
. 2017 April ; 35(4): 371–376. doi:10.1038/nbt.3803.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript

CRISPR-Cas9 nucleases have been widely used to mediate targeted genome editing
2
. In
most genome editing applications, Cas9 forms a complex with a single guide RNA (sgRNA)
and induces a double-stranded DNA break (DSB) at the target site specified by the sgRNA
sequence. Cells primarily respond to this DSB through the non-homologuous end-joining
(NHEJ) repair pathway, which results in stochastic insertions or deletions (indels) that can
cause frameshift mutations that disrupt the gene. In the presence of a donor DNA template
with a high degree of homology to the sequences flanking the DSB, gene correction can be
achieved through an alternative pathway known as homology directed repair (HDR)
3,4
.
Unfortunately, under most non-perturbative conditions HDR is inefficient, dependent on cell
state and cell type, and dominated by the formation of indels
3,4
. As most of the known
genetic variations associated with human disease are point mutations
5
, methods that can
more efficiently and cleanly make precise point mutations are needed.
We recently described base editing, which enables replacement of a target base pair with a
different base pair in a programmable manner without inducing DSBs
1
. The first examples
of base editing use a fusion protein between a catalytically inactivated (dCas9) or nickase
form of
Streptococcus pyogenes
Cas9 (SpCas9), a cytidine deaminase such as APOBEC1,
and an inhibitor of base excision repair such as uracil glycosylase inhibitor (UGI) to convert
cytidines into uridines within a five-nucleotide window specified by the sgRNA
1
. Our third-
generation base editor, BE3, converts C:G base pairs to T:A base pairs, including disease-
relevant point mutations, in a variety of cell lines with higher efficiency and lower indel
frequency than what can be achieved using other genome editing methods
1
. Subsequent
studies have validated the base editing approach to genome editing in a variety of
settings
6,7,8,9,10
.
Efficient editing by BE3 requires the presence of an NGG PAM that places the target C
within a five-nucleotide window near the PAM-distal end of the protospacer (positions 4–8,
counting the PAM as positions 21–23)
1
. This PAM requirement substantially limits the
number of sites in the human genome that can be efficiently targeted by BE3, as many sites
of interest lack an NGG 13- to 17- nucleotides downstream of the target C. Moreover, the
high activity of BE3 results in conversion of all Cs within the editing window to Ts, which
can potentially introduce undesired changes to the target locus
1
. Here we report new C:G to
T:A base editors that address both of these limitations and thereby substantially expand the
targets suitable for base editing.
We hypothesized that any Cas9 homolog that binds DNA and forms an “R-loop” complex
11
containing a single-stranded DNA bubble could in principle be converted into a base editor.
These new base editors would expand the number of targetable loci by allowing non-NGG
PAM sites to be edited. The Cas9 homolog from
Staphylococcus aureus
(SaCas9) is
considerably smaller than SpCas9 (1,053 vs. 1,368 residues), can mediate efficient genome
editing in mammalian cells, and requires an NNGRRT PAM
12
. We replaced the nickase
form of SpCas9 with that of SaCas9 in BE3 to generate APOBEC1–SaCas9n–UGI (SaBE3),
and transfected HEK293T cells with plasmids encoding SaBE3 and sgRNAs targeting six
human genomic loci (Fig. 1a). After 3 d, we used high-throughput DNA sequencing (HTS)
to quantify base editing efficiency. SaBE3 enabled C to T base editing of target Cs at a
variety of genomic sites in human cells, with very high conversion efficiencies
Kim et al.
Page 2
Nat Biotechnol
. Author manuscript; available in PMC 2017 August 13.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript

(approximately 50–75% of total DNA sequences converted from C to T, without enrichment
for transfected cells) (Fig. 1b). The efficiency of SaBE3 on NNGRRT-containing target sites
in general exceeded that of BE3 on NGG-containing target sites
1
. Perhaps due greater
solvent exposure of the strand not paired with the guide RNA
13
, SaBE3 can also result in
detectable base editing at target Cs at positions outside of the canonical BE3 activity
window (Fig. 1b). In comparison, BE3 showed greatly reduced editing for the same non-
NGG target sites (0–11% editing), consistent with the known PAM requirement of SpCas9
(Supplementary Fig. 1a)
14
. These data show that SaBE3 can mediate base editing at sites not
accessible to BE3.
We sought to further expand the targeting range of base editors by applying recently
engineered Cas9 variants that expand or alter PAM specificities. Joung and coworkers
recently reported three SpCas9 mutants that accept NGA (VQR-Cas9), NGAG (EQR-Cas9),
or NGCG (VRER-Cas9) PAM sequences
15
as well as an engineered SaCas9 variant
containing three mutations (SaKKH-Cas9) that relax its PAM requirement to NNNRRT
16
.
We replaced the SpCas9 portion of BE3 with these four Cas9 variants to produce VQR-BE3,
EQR-BE3, VRER-BE3, and SaKKH-BE3, which should target NGAN, NGAG, NGCG, and
NNNRRT PAMs respectively. We transfected HEK293T cells with plasmids encoding these
constructs and sgRNAs targeting several genomic loci for each new base editor, and
measured C to T base conversions using HTS.
SaKKH-BE3 edited sites with NNNRRT PAMs with efficiencies up to 62% of treated, non-
enriched cells (Fig. 1b,c). As expected, SaBE3 was unable to efficiently edit targets
containing PAMs that were NNHRRT (where H = A, C, or T) (Fig. 1c). VQR-BE3, EQR-
BE3, and VRER-BE3 exhibited more modest, but still substantial base editing efficiencies of
up to 50% of treated, non-enriched cells at genomic loci with the expected PAM
requirements with an editing window similar to that of BE3 (Fig. 1d–f). Base editing
efficiencies of VQR-BE3, EQR-BE3, and VRER-BE3 in general closely paralleled the
reported PAM requirements of the corresponding Cas9 nucleases; for example, EQR-BE3
was unable to efficiently edit targets containing NGAH PAM sequences (Fig. 1e). Consistent
with the known PAM requirements of SpCas9
14
, BE3 was unable to efficiently edit sites
with NGA or NGCG PAMs (0–3% efficiency) (Supplementary Fig. 1b). To confirm that the
five new base editors functioned in multiple mammalian cell types, we assessed their
performance in U2OS cells and observed robust editing, albeit with slightly lower editing
and/or transfection efficiency. (Supplementary Fig. 2a). Collectively, the properties of
SaBE3, SaKKH-BE3, VQR-BE3, EQR-BE3, and VRER-BE3 establish that base editors
behave in a modular fashion that facilitates our ability to repurpose Cas9 homologs and
engineered variants for base editing.
We examined the off-target activity of the altered-PAM base editors in human cells. We
selected two on-target loci for each new editor that have been previously analyzed for Cas9
off-target cleavage and sequenced 7 off-targets for the SaBE3 constructs and 10 off-targets
for the SpBE3 constructs
15,16
(Supplementary Fig. 3a). Consistent with our previous study
1
,
we detected off-target base editing by SaBE3 and SaKKH-BE3 at a subset of known Cas9
off-target cleavage loci containing an appropriately placed target C and that conform to their
PAM requirements (Supplementary Fig. 3b). In contrast, we observed substantially less off-
Kim et al.
Page 3
Nat Biotechnol
. Author manuscript; available in PMC 2017 August 13.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript

target base editing from VQR-BE3 or EQR-BE3 at ten known off-target loci of VQR
SpCas9
15
, suggesting that these base editors may offer enhanced specificity (Supplementary
Fig. 3c).
Next, we sought to develop base editors with altered activity window widths. All Cs within
the 5-nucleotide activity window of BE3 are typically converted to Ts with comparable
efficiency
1
. The ability to modulate the width of this window is useful when it is important
to edit only a subset of Cs present in the activity window, such as cases in which the target C
is adjacent to other Cs that if changed would result in undesired mutations to the gene of
interest.
We previously noticed that the length of the linker between APOBEC1 and dCas9 modulates
the number of bases that are accessible by APOBEC1
in vitro
1
. In HEK293T cells, however,
varying the linker length did not significantly modulate the width of the editing window,
suggesting that in the complex cellular milieu, the relative orientation and flexibility of
dCas9 and the cytidine deaminase are not strongly determined by linker length
(Supplementary Fig. 4). We hypothesized that truncating the 5end of the sgRNA might
narrow the base editing window by reducing the length of single-stranded DNA accessible to
the deaminase upon formation of the RNA-DNA heteroduplex. We co-transfected HEK293T
cells with plasmids encoding BE3 and sgRNAs of different spacer lengths targeting several
loci with multiple Cs in the editing window. Although for some target loci, truncated guide
RNAs with 16- or 17-base spacers showed narrowed editing windows, we observed no
consistent changes in the editing window width when using truncated sgRNAs with 15- to
19-base spacers (Supplementary Fig. 5).
As an alternative approach, we envisioned that mutations to the deaminase domain might
narrow the width of the editing window through at least two possible mechanisms. First,
because the high activity of APOBEC1 likely contributes to the deamination of multiple Cs
per DNA binding event
1,17,18
, mutations that reduce the catalytic efficiency of the deaminase
domain of a base editor might prevent it from catalyzing successive rounds of deamination
before dissociating or being displaced from DNA. Once any C:G to T:A editing event has
taken place, the sgRNA no longer perfectly matches the target DNA sequence and re-
binding of the base editor to the target locus should be less favorable. Second, some
mutations may alter substrate binding, the conformation of bound DNA, or substrate
accessibility to the active site in ways that reduce tolerance for non-optimal presentation of a
C to the deaminase active site. We sought to test both strategies to discover new base editors
that distinguish among multiple cytidines within the original editing window.
Given the absence of an available APOBEC1 structure, we identified several mutations
previously reported to modulate the catalytic activity of APOBEC3G, a cytidine deaminase
from the same family that shares 42% sequence similarity of its active site-containing
domain to that of APOBEC1
19
. We incorporated the corresponding APOBEC1 mutations
into BE3 and evaluated their effect on base editing efficiency and editing window width in
HEK293T cells at two C-rich genomic sites containing Cs at positions 3, 4, 5, 6, 8, 9, 10, 12,
13, and 14 (site A); or containing Cs at positions 5, 6, 7, 8, 9, 10, 11, and 13 (site B). For
analysis purposes, we define the “editing window width” as the number of nucleotide
Kim et al.
Page 4
Nat Biotechnol
. Author manuscript; available in PMC 2017 August 13.
Author Manuscript Author Manuscript Author Manuscript Author Manuscript

Citations
More filters
Journal ArticleDOI

Programmable base editing of A•T to G•C in genomic DNA without DNA cleavage

TL;DR: Adenine base editors (ABEs) that mediate the conversion of A•T to G•C in genomic DNA are described and a transfer RNA adenosine deaminase is evolved to operate on DNA when fused to a catalytically impaired CRISPR–Cas9 mutant.
Journal ArticleDOI

Efficient generation of mouse models of human diseases via ABE- and BE-mediated base editing

TL;DR: In vivo generation of mouse models carrying clinically relevant mutations using C→T and A→G editors is demonstrated, making it feasible to model and potentially cure relevant genetic diseases.
Journal ArticleDOI

RNA editing with CRISPR-Cas13

TL;DR: A type VI CRISPR-Cas system containing the programmable single-effector RNA-guided ribonuclease Cas13 is profiled in order to engineer a Cas13 ortholog capable of robust knockdown and REPAIR presents a promising RNA-editing platform with broad applicability for research, therapeutics, and biotechnology.
References
More filters
Journal ArticleDOI

Multiplex Genome Engineering Using CRISPR/Cas Systems

TL;DR: The type II prokaryotic CRISPR (clustered regularly interspaced short palindromic repeats)/Cas adaptive immune system has been shown to facilitate RNA-guided site-specific DNA cleavage as discussed by the authors.

Multiplex Genome Engineering Using CRISPR/Cas Systems

TL;DR: Two different type II CRISPR/Cas systems are engineered and it is demonstrated that Cas9 nucleases can be directed by short RNAs to induce precise cleavage at endogenous genomic loci in human and mouse cells, demonstrating easy programmability and wide applicability of the RNA-guided nuclease technology.
Journal ArticleDOI

Genome engineering using the CRISPR-Cas9 system

TL;DR: A set of tools for Cas9-mediated genome editing via nonhomologous end joining (NHEJ) or homology-directed repair (HDR) in mammalian cells, as well as generation of modified cell lines for downstream functional studies are described.
Journal ArticleDOI

Programmable editing of a target base in genomic DNA without double-stranded DNA cleavage

TL;DR: E engineered fusions of CRISPR/Cas9 and a cytidine deaminase enzyme that retain the ability to be programmed with a guide RNA, do not induce dsDNA breaks, and mediate the direct conversion of cytidine to uridine, thereby effecting a C→T (or G→A) substitution.
Journal ArticleDOI

CRISPR-Cas systems for editing, regulating and targeting genomes

TL;DR: A modified version of the CRISPR-Cas9 system has been developed to recruit heterologous domains that can regulate endogenous gene expression or label specific genomic loci in living cells, which will undoubtedly transform biological research and spur the development of novel molecular therapeutics for human disease.
Related Papers (5)