Supplementary Figure
1
Capsid library fragment generation and Cre
-
dependent capsid sequence recovery
.
(
a
) Schematic shows PCR products (yellow bar) with 7AA of randomized sequence (represented by the full spectrum vertical bar) i
nserted
after amino acid 588. The primers used to generate the library are indicated by name and half arrows. The PCR template was
modified
to eliminate a naturally occurring EarI restriction site within the capsid gene fragment (xE) (See
Online
Methods for details). (
b
) The
schematic shows the rAAV
-
Cap
-
in
-
cis genome and the primers used to quantify vector genomes (left) and recover s
equences that have
transduced Cre expressing cells (right). The PCR
-
based recovery is performed in two steps. Step 1 (blue) provides target cell
-
specific
sequence recovery by selectively amplifying
Cap
sequences from genomes that have undergone Cre
-
depende
nt inversion of the
downstream polyadenylation (pA) sequence. For step 1, 9CapF functions as a forward primer and the CDF primer functions as the
reverse
primer on templates recombined by Cre. Step 2 (magenta) uses primers XF and AR to generate the PCR pro
duct that is cloned into
rAAV
-
Δ
Cap
-
in
-
cis plasmid (library regeneration) or to clone into an AAV2/9 rep
-
cap trans plasmid (variant characterization).
(
c
)
The
table
p
rovides the sequences for the primers shown in
a
and
b
.
Nature
Biotechnology:
doi:10.1038/nbt.3440
Supplementary Figure
2
The most enriched variants recovered from
in vivo
selections in GFAP
-
Cre mice
.
(
a
) The table
pro
vides the 7
-
mer AA and nucleic acid sequences, percentage enrichment (% of total variants sequenced), capsid
characteri
stics, and production efficiencies of the three most enriched variants from each selection.
(
b
)
Images of representative sagittal
brain sections from mice assessed 2 weeks after injection of 3.3x10
10
vg/mouse of ssAAV
-
CAG
-
mNeonGreen
-
farnesylated (mNeGreen
-
f) packaged into AAV
-
PHP.B or the second or third most enriched variants, AAV
-
PHP.B2 and AAV
-
PHP.B3. Data are representative of
2 (AAV
-
PHP.B) and 3 (AAV
-
PHP.B2 and AAV
-
PHP.B3) mice per group.
(
c
)
DNase
-
resistant vg obtained from preparations of the
indivi
dual variants recovered from GFAP
-
Cre selections. Yields are given as the number of purified vector genome
(vg)
copies per 150
mm dish of producer cells; mean ± s.d. *p<0.05, one
-
way ANOVA and Tukey
’s
multiple comparison test. The number of independent
pre
parations for each capsid is shown within the bar.
Nature
Biotechnology:
doi:10.1038/nbt.3440
Supplementary Figure
3
AAV
-
PHP.B transduces several interneuron cell types and endothelial cells but does not appear to transduce microglia
.
(
a-d
)
Adult
mice were
injected with 1x10
12
vg of AAV
-PHP.B:CAG-
GFP and assessed for GFP expression 3 weeks later. Representative
images show IHC for GFP (
a
-
c
) or native GFP fluorescence (
d
) in green together with IHC for the indicated antigen (magenta) and brain
region. (
e
) Adu
lt mice were injected with 3.3x10
10
vg of ssAAV
-
PHP.B:CAG
-
mNeGrn
-
f and assessed at 2 weeks post injection. Native
fluorescence from mNeGrn
-
f co
-
localizes with some endothelial cells expressing CD31. (
f
,
g
) Adult mice were injected with 2x10
12
vg of
ssAAV
-
P
HP.B:CAG
-
NLS
-
GFP and assessed at 3 weeks post injection. Images show native NLS
-
GFP expression along with
Iba1
.
A
sterisks indicate cells that express the indicated antigen, but not detectable GFP. Parvalbumin (PV), Calbindin (Calb) and Ca
lretinin
(CR). Sca
le bars: 20 μm (
a
-
d
)
; 50 μm (
e
,
f
) and
500 μm (
g
)
.
Nature
Biotechnology:
doi:10.1038/nbt.3440
Supplementary Figure
4
Long
-
term eGFP expression in the brain following gene transfer with AAV
-
PHP.B.
Adult mice were intravenously injected with the indicated dose of
ssAAV
-
CAG
-
GFP packaged into
AAV9 or AAV
-
PHP.B and were
assessed for native eGFP fluorescence 377 Days later.
N
=1 per vector/dose.
Nature
Biotechnology:
doi:10.1038/nbt.3440
Supplementary Figure
5
Representative images of native GFP fluorescence and IHC for several neuron and glial cell types following transduction by
AAV
-
PHP.B:CAG
-
NLS
-
GFP.
(
a-d
) Adult mice were injected with 2x10
12
vg of ssAAV
-PHP.B:CAG-
NLS -GFP and assessed at 3 weeks post
injection. Images show
native NLS
-
GFP expression along with IHC for the indicated antigen in the indicated brain region. In all panels, arrows indicate co
-
localization of GFP expression with IHC for the indicated antigen.
(
b, c
)
S
ingle
-
plane confocal imag
es;
(
a, d
)
MIP. Corpus
c
allosum (cc),
substantia nigra pars compacta (SNc), ventral tegmental area (VTA). Scale bars: 50 μm.
Nature
Biotechnology:
doi:10.1038/nbt.3440
Supplementary Figure
6
AAV9, AAV
-
PHP.A and AAV
-
PHP.B transduce human iPSC
-
derived cortical
neurons and astrocytes in dissociated cultures and
intact 3D cortical cultures
.
(
a
) AAV -PHP.B provides
higher transduction of human neurons and astrocytes in dissociated monolayer cultures. Representative images
show GFP expression at five days after vira
l transduction (ss
AAV
-
CAG
-
NLS
-
GFP packaged in
AAV9, AAV
-
PHP.A or AAV
-
PHP.B; 1x10
9
vg/well) of dissociated iPSC
-
derived human cortical spheroids differentiated
in vitro
. GFP expressing cells (green) co
-
localize with
astrocytes immunostained for GFAP (cyan)
or neurons immunostained for MAP2 (magenta) as indicated by white arrows. (
b
)
Quantification of the percentage of GFAP
+
or MAP2
+
cells
transduced
by AAV9, AAV
-
PHP.A or AAV
-
PHP.B (n=3 differentiations into
cortical spheroids of two human iPSC
lines derived from two subjects; two
-
way ANOVA, Tukey
’s
multiple comparison test; mean ± s.d).
(
c
)
AAV9, AAV
-
PHP.A and AAV
-
PHP.B transduce intact human 3D cortical cultures (cortical spheroids
differentiated from human iPSCs
).
Images of human iPSC
-
derived
cortical spheroid cryosections (day 205 of
in vitro
differentiation) transduced with ssAAV
-
CAG
-
NLS
-
GFP
packaged in AAV9, AAV
-
PHP.A or AAV
-
PHP.B show native GFP fluorescence together with immunostaining for GFAP and MAP2. Insets
show co
-
labeling of GFP flu
orescence with GFAP
+
astrocytes (cyan) or MAP2
+
(magenta) neurons.
Scale bars: 40 μm (
a
); 100 μm (
c
).
Nature
Biotechnology:
doi:10.1038/nbt.3440
Supplementary Figure
7
AAV-PHP.B and AAV
-PHP.A capsids localize to the brain vasculature after intravenous injection
and transduce cells along the
vasculature by 24 hours post
-
administration
.
Adult mice were injected with 1x10
11
vg of ssAAV
-CAG- NLS -GFP packaged into AAV9, AAV
-PHP.A or AAV-
PHP.B as indicated. (
a
,
b
)
Representative images of capsid immunostaining (green)
using the B1 anti
-
AAV VP3 antibody that recognizes a shared internal epitope
in the cerebellum (
a
) or striatum (
b
) in the brains of mice injected intravenously one hour prior to fixation by cardiac perfusion. Cell nuclei
are labeled with DAPI (magenta). L
ipofuscin autofluorescence (yellow) can be distinguished from capsid staining by its presence in both
green and red channels. The inset (right) shows a 3D MIP image of the area highlighted in the AAV
-
PHP.B image. Arrows highlight capsid
IHC signal; asteris
ks indicate vascular lumens. Data are representative of 2 (no virus and AAV
-
PHP.A) or 3 (AAV9 and AAV
-
PHP.B) mice
per group. (
c
) Representative image of GFP expression (green) with D
API
(white) and CD31 (magenta) 24 hours post
-
administration of
AAV
-
PHP.B.
Arrows
highlight GFP
-
expressing cells.
(
d
) Quantification of the number of GFP expressing cells present along the
vasculature in the indicated brain regions.
N
=3 per group
;
mean ± s.d;
AAV
-
PHP.B vs AAV9 and AAV
-
PHP.A, ***
p
<0.001 for all regions;
AAV9 vs
AAV
-
PHP.A, not significant; two
-
way ANOVA.
Scale bars: 200 μm (
a
)
; 50 μm (
b, c
);
Major tick marks are 50
μm in the high
magnification inset (
a
).
Nature
Biotechnology:
doi:10.1038/nbt.3440