Microsoft Word - MTCH2_full_paper_v3.7.docx

Supplementary Materials for

MTCH2 is a mitochondrial outer membrane protein insertase

Alina Guna*, Taylor A. Stevens*, Alison J. Inglis*, Joseph M. Replogle, Theodore K. Esantsi,

Gayathri Muthukumar, Kelly C.L. Shaffer, Maxine L. Wang, Angela N. Pog

son, Jeff J. Jones,

Brett Lomenick, Tsui

Fen Chou, Jonathan S. Weissman

†

, and Rebecca M. Voorhees

†

correspondence to:

weissman@wi.mit.edu

(J.S.W);

voorhees@caltech.edu

(R.M.V)

This PDF file includes:

Materials and Methods

Figs. S1 to S17

Other supplementary materials for this manuscript

Tables S1 to S3

Materials and Methods

Plasmids and antibodies

Endogenous sequences used in this study for in vitro and in vivo analysis were sourced from

UniProtKB/Swiss

Prot

and included: squalene synthase isoform 1 (SQS/FDFT1;

Q6IAX1

synaptojanin

2 binding protein

(OMP25/SYNJBP;

P57105

), mitochondrial antiviral

signaling

protein (MAVS;

Q7Z434

mitochondrial import inner membrane translocase subunit Tim9

(TI

M9;

Q9Y5J7

), vesicle associated membrane protein 2 (VAMP;

P51809

FUN14 domain

containing protein 1

(FUNDC1;

Q8IVP5

1),

mitochondrial import receptor subunit TOM5

homolog

(TO

M5;

Q8N4H5

mitochondrial import receptor subunit TOM22 homolog

(TO

M22;

Q9NS69

biquitin carboxyl

terminal hydrolase 30

(USP30;

Q70CQ3

apoptosis

regulator BAX

(BAX;

Q07812

Bcl

2 homologous antagonist/killer

(BAK1;

Q16611

Bcl

like protein 1

(BCL2L1;

Q07817

Cytochrome b5 type

(CYB5B;

O43169

peptidyl

prolyl cis

ans isomerase FKBP8

(FKBP8;

Q14318

mitochondrial fission factor

(MFF;

Q9GZY8

1),

inactive hydroxysteroid dehydrogenase

like protein 1

(HSDL1;

Q3SXM5

amine oxidase [flavin

containing] A

(MAOA;

P21397

amine oxidase [flavin

containing] B

(MAOB;

27338

mitochondrial Rho GTPase 1

(RHOT1;

Q8IXI2

mitochondrial Rho

GTPase 2

(RHOT2;

Q8IXI1

), and

mitochondrial carrier homolog 2

(MTCH2;

Q9Y6C9

1).

Constructs for expression in rabbit reticulocyte lysate (RRL) used the SP64 vector as a backbone

(Promega). For in vitro insertion reactions, an N

terminal 3xFLAG tag and a C

terminal 6xHis

tag were appended to the respective terminus for affinity purifica

tion (see fig. S8A). Where

noted, the transmembrane domain (TMD) of TA proteins along with N

and C

terminal flanking

sequences were instead conjugated to maltose binding protein (MBP)

(

)

or villin headpiece

(VHP) domains

as illustrated in fig. S8B. When monitoring insertion into the endoplasmic

reticulum (fig. S4 and S16), constructs were modified to replace the 6xHis tag with a C

terminal

opsin tag containing a

glycosylation acceptor site which can be used as a proxy for

insertion

(

)

. TOM dependent mitochondrial import substrates were designed by fusing dihydrofolate

reductase (DHFR) to mitochondrial transit sequences from either

N. crassa

ATP synthase

subunit 9 (residues 1

69) or

S. cerevisiae

cytochrome b

(residue 1

167) which direct import to

the mitochondrial matrix or inner membrane space (IMS), respectively.

For expression in K562 cells, the basis for all constructs was a mammalian expression lenti

viral

backbone containing a UCOE

1α

promoter and a 3’ WPRE element (

(

)

; Addgene

#135448). The exception was the dual fluorescent reporter used for the CRISPRi screen (RFP

P2A

OMP25

GFP11) which was integrated into an

SFFV

tet3G backbone

(

)

. The dual color

reporter system used for in vivo experiments has been previously described

(

)

(

)

. Note that

the mCherry variant of RFP was used in all instances, but the simpler nomenclature of RFP is

used in the text and figures. For complementation with the GFP1

10 system, the GFP11 tag

(RDHMVLHEYVNAAGIT) was appended to the appropriate terminus of

the indicated protein

as determined by predicted topology (see Fig. 2C). In order to express GFP1

10 in the ER lumen,

the human calreticulin signal sequence was appended preceding GFP1

KDEL

(

)

(

)

. For

targeting to the inner membrane space of the mitochondria, either the targeting signal from

MICU1 (aa.1

60)

(

)

or the targeting sequence from LACTB (aa 1

68)

(

)

was appended to

the N terminus of

GFP1

10. Expression of MTCH2 or indicated mutants was from a BFP

P2A

(MTCH2) cassette, allowing us to gate and sort for expressing cells.

The single sgRNA for MTCH2

(GACGGAGCCACCAAGCG

ACC)

was generated

by annealed

oligo cloning of top and bottom oligonucleotides (Integrated DNA Technologies, Coralville, IA)

into a lentiviral pU6

sgRNA EF

1α

Puro

T2A

BFP vector digested with BstXI/BlpI (Addgene,

cat# 84832). BFP was excised in certain s

gRNA variants when the color interfered. Though most

experiments were done with a single guide that gave robust knock

down, key results were

verified with an additional guide (GGGCTCACCGGGTCGCTTGG) to exclude possible off

target effects. In many instances,

we used a programmed dual sgRNA guide vector (

(

)

;

Addgene #140096) to allow for multiple genes to be depleted at once or to increase efficiency of

knock

down. Dual guide pairs included MTCH2

ATP13A1 (GACGGAGCCA

CCAAGCGACC,

GGGTAAAGCAGCCCGGCGAA), MTCH1

MTCH1 (GCGGCACCGCCGCGAGCCCA,

GAGCCCAGGGCGCCACTTCC), and MTCH2

EMC2 (GACGGAGCCACCAAGCGACC,

GGAGTACGCGTCCGGGCCAA). Transient knock

out of MTCH1 was achieved by modifying

a pLentiCRISPR backbone (

(

)

, Addgene #102315) to express the following guides:

GGACAACGCCCCGACCACTG and

CTGCATCATCATCTCGTAGG.

Constructs for recombinant bacterial protein expression were all cloned in the pQE plasmid

(Qiagen). Su9

DHFR and CaM

Alfa

Sec61

β(2

60)

OMP25(112

145)F128Amber were

cloned downstream of a His

SUMO tag. A GFP Nb fusion protein used in this manuscript

was modified from a previously established construct (

(

)

; Addgene ID #149336) to excl

ude

the SUMO

Eu1

tag upstream of the GFP Nb. Constructs for the expression of SENP

EuB

(

)

;

Addgene ID #149333),

SENP1 (

(

)

; Gift from Dirk Görlich;

Addgene ID #104962), and

BirA (

(

)

; Gift from Dirk Görlich, Addgene ID #149334), and for site

specific incorporation of

BpA at the Amber codon position (

(

)

; Addgene #3119

0) have all been previously described.

Constructs for generating human stable cell lines for recombinant protein expression used either

the pHAGE2 plasmid (Gift from Magnus A. Hoffmann and Pamela Bjorkman) for lentiviral

integration into the Expi293F cell

line or the pcDNA5/FRT/TO plasmid (Thermo Scientific Cat.

#V652020) for recombinase

mediated integration into Flp

In 293 T

REx cell line. MTCH2 was

terminally fused with a GFP

SUMO

Eu1

tag and downstream of a doxycycline

inducible CMV

promoter. The EMC3

GFP

P2A

RFP expression vector was previously described (

(

)

The following antibodies were used in this study: MTCH2 (ab113707, Abcam, UK); FUNDC1

(OAAB12808, Aviva systems biology, USA); CYB5B (HPA007893, Atla

s antibodies, USA);

MIRO2 (RHOT2) (ab224089, Abcam, UK); CYC1 (4272, Cell signaling technology, USA

);

SYNJ2BP (OMP25) (15666

AP, Proteintech, USA); EMC3 (67205, Proteintech, USA);

VDAC1 (sc

390996, Santa Cruz Biotech, USA); mitofilin (ab110329, Abcam, UK

); SAMM50

(ab133709, Abcam, UK); ATP13A1 (16244

AP, Proteintech, USA); tubulin (T9026, Sigma

Aldrich, USA). The ALFA tag was detected by coupling HRP to an ALFA nanobody

(

)

. The

Sec61

antibody was a gift from

Ramanujan Hegde. Secondary antibodies used for Western

blotting were: Goat anti

mouse

and anti

rabbit

HRP (#172

1011 and #170

6515, Bio

Rad,

USA).

Cell culture and cell line generation

K562 cells were grown in RPMI

1640 with 25 mM HEPES, 2.0 g/L

NaHCO

, and 0.3 g/L L

glutamine supplemented with 10% FBS (or Tet System Approved FBS), 2 mM glutamine, 100

units/mL penicillin, and 100 μg/mL streptomycin. Cells were maintained at a confluency

between 0.25

1 x10

cells/mL. HEK293T cells were grown in DM

EM supplemented with 10%

FBS, 100 units/ml penicillin and 100 μg/ml streptomycin.

All cell lines were grown at 37°C.

For

he expression of EMC3

GFP, Flp

In 293 T

Rex cells were purchased from

ThermoFisher

Scientific (USA) (RRID: CVCL_U427) and grown in DME

M supplemented with 2 mM

glutamine, 10% FBS, 15 μg/mL Blasticidine S, and 100 μg/mL Zeocin. The open reading frame

to be integrated into the genomic FRT site was cloned into the pcDNA5/FRT/TO vector

backbone and cell lines were generated according to the m

anufacturer’s protocol. To allow for

large scale growth, these cells were adapted to grow in suspension. Briefly, over the course of 10

days the FBS

supplemented DMEM was serially diluted with FreeStyle 293 Expression Medium

(ThermoFisher Scientific, USA).

Once growing in 100% FreeStyle Medium, the cells were

transferred to 1

2 L roller bottles (Celltreat, USA) and grown in a shaking incubator operating at

8% CO

and rotating at 125 rpm. For the expression of MTCH2, lenti

viral infected inducible

Expi293F s

uspension cells were grown in Expi293 expression media

(ThermoFisher Scientific,

USA) at

37°C

, 8% CO

and 125 rpm shaking in 1 L roller bottles with vented caps (Celltreat,

USA)

Three K562 cell lines were used a basis for cell lines generated in this stud

y: CRISPRi K562

cells expressing dCas9

BFP

KRAB (KOX1

derived)

(

)

, K562

dCas9

BFP

KRAB Tet

cells

(

)

, and CRISPRi cells generated by stably expressing ZIM3 KRAB

dCas9

P2A

BFP

from a UCOE

SFFV promoter

(

)

. To generate cell lines with GFP1

10 in the mitochondrial

IMS or ER lumen, virus was made from the respect

ive constructs: LACTB(GFP1

10),

MICU1(GFP1

10) or CalR(GFP1

10)

KDEL. CRISPRi K562 cells were infected with lenti

virus

and sorted into 96

well plates as single cell clones using a Sony Cell Sorter (SH800S). After

expansion, correct cell lines were confirm

ed by successful complementation with a construct

targeted to the respective compartment, appended to a GFP11. To generate the cell line used for

screening, lenti

virus containing MICU1(GFP1

10) and RFP

P2A

OMP25

GFP11 under a tet

inducible promoter were c

infected to one copy per cell in CRISPRi K562 Tet

On cells, single

cell sorted, and verified by induction with doxycycline (100 ng/uL) and microscopy in

conjunction with MitoTracker staining to confirm correct localization.

To generate MTCH2 knock

out c

ell lines, K562 CRISPRi cells with or without MICU1(GFP1

10) were nucleofected with a MTCH2 targeting guide in the

pX458 backbone (Addgene plasmid

# 48138)

using the Lonza SF Cell Line 96

well Nucleofector Kit (V4SC

2096). The pX458

backbone was adapted to

express two sgRNAs targeting MTCH2

[AGCCGACATGTCTCTAGTGG], [ GGCTTTGCGAGTCTGAACGT]. Two days following

nucleofection, GFP

positive cells were single cell sorted into 96

well plates. After colonies from

single cells grew out, loss of MTCH2 was confirmed by

Western blotting. CRISPR

Cas9

induced genome edits were identified using the computational pipeline described in

(

)

Lentivirus

Lentivirus was generated by co

transfecting HEK293T cells with two packaging pl

asmids

(pCMV

VSV

G and delta8.9, Addgene #8454) and the desired transfer plasmid using

TransIT

293 transfection reagent (Mirus). 48 hours after transfection, the supernatant was collected and

flash frozen. In all instances, virus was rapidly thawed prior t

o transfection. Virus for the

genome

wide CRISPRi screen was also generated using this method.

CRISPRi KD Screen

The genome

scale CRISPRi screen was performed in duplicate as previously described

(

)

(

)

The hCRISPRi

v2 compact library (5 sgRNAs per gene, Addgene pooled

library #83969) was

transduced in duplicate into 330 million K562

CRISPRi

Tet

((MICU1)

GFP1

10)

(tet

RFP

P2A

OMP25

GFP11) cells at multiplicity of infection (MOI) < 1 (percentage of transduced cells

48 hours after infection as measured by BFP positive c

ells: 30

35%). Cells were grown in 1 L of

media in 1 L spinner flasks (Bellco, SKU: 1965

61010). 48 hours after spinfection with the

genome

wide library, guide positive cells were selected with 1

g/mL puromycin for three days.

Following a 36 hour recovery

, cells were induced with 100 ng/mL doxycycline for 36 hours and

sorted on a FACS AriaII Fusion Cell Sorter. To ensure that the culture was maintained at an

average coverage of more than 1000 cells per sgRNA, cells were diluted daily to 0.5x10

cells/mL.

During sorting, cells were gated for BFP (indicating a guide

positive cell), as well as GFP and

RFP signal (successfully induced). Cells were sorted based on the GFP:RFP ratio of this final

gated population. Roughly 40 million cells with either the highest

(30%) or lowest (30%)

RFP:GFP ratio were collected, pelleted and flash

frozen. Genomic DNA was purified using the

Nucleospin Blood XL kit (Takara Bio, #740950.10) and amplified by index PCR with barcoded

primers. The resulting guide library (~264 bp) was

purified using SPRIbeads (SPRIselect

Beckman Coulter #B23318). Sequencing was performed using an Illumina HiSeq2500 high

throughput sequencer. Sequencing reads were aligned to the CRISPRi v2 library sequences,

counted and quantified

(

)

. Generation of negative control genes and calculation of phenotype

scores and Mann

Whitney p

values was performed as described previously

(

)

(

)

. Gene

level

phenotypes and counts are available in Supplementary Table 1.

Protein expression and purification

Su9

DHFR and CaM

Alfa

Sec61β

OMP25(BpA)

The BL21(DE3) expression strain was used to express Su9

DHFR and CaM

Alfa

Sec61β

OMP25(BpA) in

LB media

. Su9

DHFR expressing cultures were induced with 1 mM IPTG after

an optical density of 0.6 was reached. After induction, Su9

DHFR was expressed at 37°C for 3

hours. CaM

Alfa

Sec61β

OMP25(BpA) was co

transformed with pEVOL

BpF and grown to

an op

tical density of 0.2 followed by induction with 1% arabinose and the addition of 1 mM BpA

(Bachem). Cells were then grown to an optical density of 0.6 followed by induction with 1 mM

IPTG and expression at 25°C for 6 hours. Cells were pelleted by centrifug

ation and resuspended

in a lysis buffer containing 500 mM NaCl, 50 mM Tris pH 7.5, 10 mM imidazole, 5 mM β

ME.

For purification,

E. coli

resuspensions were supplement with EDTA

free protease inhibitor

tablets (Roche) and lysozyme prior to lysis via sonic

ation. Lysate was clarified by centrifugation

at 18,000 rpm for 30 minutes in an SS

34 rotor. Clarified lysate was incubated with NiNTA resin

for 30 minutes while rolling at 4°C. NiNTA resin was washed extensively with resuspension

buffer, and then equilib

rated with a SENP elution buffer containing 150 mM NaCl, 50 mM Tris

pH 7.5, 10 mM imidazole, 5 mM β

ME, 10% glycerol. NiNTA resin was then incubated with

SENP1 for 2 hours at 4°C to release SUMO

cleaved protein from the resin.

For CaM

Alfa

Sec61β

P25(BpA) the resuspension buffer included 1 mM CaCl

, and

SENP elution buffer contained 100 mM NaCl, 50 mM Tris pH 7.5, 10 mM imidazole, 5 mM β

ME, 1 mM CaCl

. Additionally, protein was cleaved overnight with 3C protease following

SENP1 elution. Cleaved pr

otein was then concentrated to 250 μL and injected onto a Superdex

200 increase 10/300 GL size exclusion column equilibrated in a buffer containing 150 mM

KoAc

, 50 mM HEPES pH 7.4, 2 mM

MgoAc

, 1 mM CaCl

, 1 mM DTT, 10% glycerol. Protein

containing fractions

were pooled, concentrated, aliquoted and flash frozen.

Biotinylated GFP

Nb and Alfa

Expression and purification of all GFP and Alfa nanobody constructs as well as

SENP1, BirA,

and SENP

EuB

generally proceeds as follows: the NEB Express I

expression strain was used with

TB medium. Cultures were induced with 0.2 mM IPTG after an optical density of 2.0 was

reached. Protein was expressed at 18°C for 18

20 hours. Cells were pelleted by centrifu

gation

and resuspended in a lysis buffer containing 50 mM Tris pH 7.5, 300 mM NaCl, 20 mM

imidazole, 1 mM DTT, 1 mM PMSF. Cells were lysed by sonication and lysate was clarified by

centrifugation at 18,000 rpm for 30 minutes in an SS

34 rotor. Clarified ly

sate was incubated

with NiNTA resin for 1 hour while rolling at 4°C. NiNTA resin was washed extensively with

resuspension buffer and the eluted with a buffer containing 50 mM Tris pH 7.5, 300 mM NaCl,

500 mM imidazole, 250 mM sucrose. The imidazole was rem

oved using a PD

10 desalting

column (GE Healthcare, USA). The following protein

specific modifications were applied to the

above protocol: the Rosetta

gami 2(DE3) expression strain was used to express His

Avi

SUMOstar

AlfaNb and expression time was limit

ed to 6 hours at 18°C. SENP

EuB

was limited to

an expression time of 6 hours at 18°C.

SENP1 was expressed as a His

TEV

fusion and was

cleaved with TEV protease overnight after buffer exchange and then ran over NiNTA resin to

remove the cleaved tag. The

expression and purification of His

Avi

SUMO

Eu1

anti GFP

nanobody,

SENP1, BirA, and SENP

EuB

have all

(

)

been previously described

(

)

(

)

All nanobody constructs used in this publication were biotinylated in a buffer containing 50 mM

Tris pH 7.5, 100 mM NaCl, 12.5 mM MgCl

, 10 mM ATP, 10 mM biotin and BirA at a 1:50

molar ratio to the nanobody substrate. Biotinylat

ion reactions were incubated at 25°C for 3 hours

and then buffer exchanged in a PD

10 column equilibrated with a buffer containing 50 mM Tris

pH 7.5, 200 mM NaCl, 1 mM DTT, 250 mM sucrose. Biotinylated protein was aliquoted and

flash frozen.

MTCH2 and EMC

We isolated MTCH2 and EMC (via EMC3

GFP) under native conditions from detergent

solubilized cells using a biotinylated anti

GFP nanobody, expressed and purified as previously

described

(

)

. Cells were grown in

1 L roller bottles. For the MTCH2 lines, expression was

induced for at least 48 hours with 1 μg/mL doxycycline. Cells were then harvested by

centrifugation, washed with 1x PBS, and then pellets were weighed.

Purification generally proceeded as follows: ce

ll pellets were resuspended at a ratio of 1 g to 10

mL hypotonic lysis buffer containing 10 mM HEPES pH 7.5, 10 mM

KoAc

, 0.15 mM

MgoAc

0.5 mM DTT, supplemented with EDTA

free protease inhibitor tablets (Roche). The cell

resuspension was incubated on ice fo

r 10 minutes to allow cells to swell, and then lysed in a

Dounce homogenizer with 10x strokes. The NaCl concentration was adjusted to 180 mM

immediately after Dounce homogenization. Cell membranes were pelleted by centrifugation at

18,000 rcf in an SS

34 (

28020TS, Thermo Fisher) rotor for 10 minutes. Supernatant was

discarded and cell membranes were washed by resuspending and pelleting 2x in membrane wash

buffer containing 10 mM HEPES/KOH pH 7.5, 200 mM NaCl, 0.15 mM

MgoAc

, 0.5 mM DTT.

The resulting pellet

was resuspended at a ratio of 1 g (original cell pellet weight) to 6.8 mL

solubilization buffer containing 50 mM HEPES pH 7.5, 200 mM NaCl, 2 mM

MgoAc

, 1%

deoxy

BigCHAP (DBC; Anatrace Cat # B310), 1 mM DTT, supplemented with EDTA

free

protease inhibitor ta

blets (Roche). After 30 minutes of head

over

tail incubation with

solubilization buffer, the lysate was cleared by centrifugation for 30 minutes at 4°C and 18,000

rpm in a SS

34 rotor. The supernatant was then added to pre

equilibrated magnetic Streptavidi

resin (ThermoFisher) bound to biotinylated anti

GFP nanobody and blocked with free biotin.

After 1 hour of binding while rolling at 4°C, the resin was washed four times with wash buffer

(solubilization buffer with 0.2% DBC). Resin was then incubated in w

ash buffer + 600 nM

SENP

EuB

on ice for 2 hours to release SUMO

cleaved MTCH2 from the resin. Eluted samples

were analyzed via SDS

PAGE with Sypro Ruby stain (Bio

Rad).

For EMC3

GFP, the whole cell pellet was solubilized in 1% DBC

containing buffer, withou

t a

hypotonic lysis and membrane washing step, and the GFP Nb was fused to a cleavable SUMO

Eu1

module.

Mitochondrial isolation and semi

permeabilized cells

Mitochondrial isolation for K562 cells was adapted from an established protocol

(

)

. K562 cells

were centrifuged at 220

for 5 minutes. Pellets were washed once in PBS and pelleted by

spinning at 500

for 5 minutes. Pellets were resuspended in a homogenization buffer containing

210 mM mannitol, 70

mM sucrose, 5 mM HEPES pH 7.4, 10 mM EDTA, 1 mM PMSF, and 2

mg/mL BSA. After incubating on ice for 10 minutes, cells were then lysed with a glass Dounce

homogenizer with a tight

fitting pestle, or a Potter

Elvehjem homogenizer motor

driven at 1600

rpm for

large scale purifications. Homogenized cells were pelleted at 1300

for 5 minutes to

remove nuclei and unbroken cells, then the supernatant was transferred to a clean tube. This step

was repeated twice. Nuclei

free homogenized cells were then centrifuged a

t 11,000

for 10

minutes. The supernatant was removed and the mitochondria

containing pellet was then

resuspended in an isolation buffer containing 210 mM D

mannitol, 70 mM sucrose, 5 mM

HEPES pH 7.4, 10 mM EDTA. Mitochondria were then pelleted and resuspe

nded in fresh

isolation buffer to wash away BSA and cytoplasmic proteins. After a final pelleting step,

mitochondria were resuspended in a small volume (5

50 μL) of isolation buffer. To normalize

mitochondrial samples, the protein concentration was measure

d using a Bradford assay.

For experiments using trypsin

treated mitochondria, the final 2 was steps performed after

pelleting mitochondria used import buffer contain

ing 250 mM sucrose, 5 mM Mg

Ac2, 80 mM

, and 20 mM HEPES. Mitocho

dria were then mixed with 0 or 50 μg/ml trypsin (Sigma

Aldrich Cat# T1426) dissolved in import buffer and incubated on ice for 30 minutes. 50 μg/ml

trypsin inhibitor (Sigma

Aldrich Cat# T91

28) and 1 mM PMSF were added to quench the

reaction. Mitochondria were pelleted, washed once in import buffer + 5 μg/ml trypsin inhibitor,

and then resuspended in a concentrated volume prior to use in import experiments

To further enrich mitochondrial samples for certain experiments, a percoll gradient was used.

Isolation buffer density gradients were formed in 3 layers with 40%, 26%, and 12% percoll.

Resuspended mitochondria were layered on

top of the gradient. Gradients were centrifuged at

45,000

for 45 minutes in a TLS

55 rotor (Beckman Coulter). Pure mitochondria were retrieved

from the 40%

26% percoll interface (see fig. S1B). Mitochondria were diluted 5

fold in isolation

buffer, and the

n pelleted and washed in isolation buffer twice more.

HEK293T cells, either wild

type or an EMC5 knock

out background

(

)

were semi

permeabilized using standard methods

(

)

. Briefly, 3x10

cells were collected and washed once

with ice cold wash buffer containing 25 mM HEPES pH 7.4, 100 mM

KoAc

, 2 mM Mg

Ac2.

The cells were resuspended in 1 mL SP buffer containing 25 mM Hepes pH 7.4, 100 mM K

Ac,

2 mM Mg

Ac2, 50

g/mL digitonin)

and incubated on ice for 5 minutes. The semi

permeabilized cells were collected by centrifugation at 500

for 5 minutes at 4°C, with the

digitonin removed by washing three times. Finally, the cells were pelleted at 12,000

for 15

seconds and resuspended i

n 10

L wash buffer. To test the integrity of the outer mitochondria

membrane as in fig. S1A, SP cells and purified mitochondria (as described above) were

incubated with the indicated amount of proteinase K (PK). Following quenching, the resulting

reaction

was subjected to blotting against mitofilin, an inner membrane localized protein which

should not be accessible by PK when the outer mitochondrial membrane is intact.

In vitro translation and insertion

In vitro translations were carried out in rabbit

reticulocyte lysate (RRL) as previously described

(

)

. Constructs for in vitro translation reactions were based on the SP64 vector (Promega,

USA). Templates for transcription were generated by PCR, with primers bind

ing and upstream

of the SP64 promoter and roughly 200 bp downstream of the stop codon

(

)

. Following

transcription at 37°C for 1.5 hours, reactions were used directly in a translation reaction.

Substrates were tra

nslated for 15

30 minutes at 32°C in the presence of radioactive

methionine. Prior to addition of mitochondria or semi

permeabilized cells, 1 mM puromycin was

added to prevent further synthesis.

Mitochondrial insertion reactions used isolated mitocho

ndria, prepared as described above.

Insertion reactions were performed by diluting 4

L of a puromycin treated translation reaction

in 50

L of import buffer (250 mM sucrose, 5 mM

MgoAc

2, 80 mM

KoAc

, 20 mM HEPES pH

7.4, 2.5 mM ATP, 15 mM succinate) with 15

g of purified mitochondria and further incubating

at 32°C for 30 minutes. For competition experiments in Fig. 1A, insertion reactions were

carried

out in the presence of 1, 2, or 5

M recombinant Su9

DHFR, and 5

M methotrexate

(BP266510, Fisher Chemicals, USA).

Protease digestions were initiated by the addition of proteinase K at 0.25 mg/mL, and reactions

were then incubated on ice for 1 ho

ur. Reactions were quenched by the addition of 5 mM PMSF

in DMSO, followed by transfer to boiling 1% SDS (final concentration) in 0.1 M Tris/HCl pH

8.0. His

tagged protected fragments were enriched by incubating with NiNTA resin in IP buffer

(50 mM HEPES p

H 7.5, 500 mM NaCl, 10 mM imidazole, 1% Triton). Proteinase K digested

reactions were diluted to 1 mL and mixed with 10

L resin, then incubated with end

over

end

mixing for 1.5 hours at 4°C. The resin was further washed with 3x1 mL IP buffer, and the

prod

ucts eluted from the resin in sample buffer containing 50 mM EDTA pH 8.0.

Insertion reactions with semi

permeabilized (SP) cells used a ratio of 1

L cells per 10

translation reaction. To verify insertion of substrates into the ER as indicated by glyc

osylation, a

tripeptide competitor of glycosylation (Asn

Tyr

Thr) was added at 50

M when indicated.

Mass spectrometry

For the crosslinking

IP samples, TCA

precipitated pellets were resuspended in a buffer

containing 8 M Urea and 100 mM Tris pH 8.5. T

he sample was reduced by incubation with 3

mM TCEP for 20 minutes, then alkylated by incubation with 10 mM iodoacetamide for 15

minutes, all at room temperature. The sample was then digested with 2 ng/μL LysC for 4 hours

at room temperature, diluted 4

fold

with 100 mM Tris pH 8.5 and CaCl

was added to 1 mM.

The sample was then digested with 4 ng/μL Trypsin overnight at room temperature. Samples

were acidified by adding trifluoroacetic acid to 0.5%, desalted using Pierce C18 spin columns

(Pierce), lyophili

zed, and then resuspended in 2% acetonitrile, 0.2% formic acid.

For the mitochondrial proteomics experiments, the S

trap sample preparation kit (ProtiFi) was

used according to the manufacturer’s instructions. Sample was digested on the S

trap column

with

1 μg Trypsin per 10 μg protein overnight at 37°C. In addition to the provided S

trap sample

preparation protocol, a final elution step with 70% acetonitrile, 1% formic acid was added.

Eluted peptides were lyophilized, and then resuspended in 2% acetonitril

e, 0.2% formic acid.

The peptide concentration was determined with a Quantitative Fluorometric Peptide Assay

(Pierce) kit.

MS/MS analysis for the crosslinking

IP experiment was performed with an EASY

nLC 1200

(ThermoFisher Scientific, San Jose, CA) cou

pled to a Q Exactive HF hybrid quadrupole

Orbitrap

mass spectrometer (ThermoFisher Scientific, San Jose, CA). Peptides were separated on an

Aurora UHPLC Column (25 cm × 75 μm, 1.6 μm C18, AUR2

25075C18A, Ion Opticks) with a

flow rate of 0.35 μL/min for a t

otal duration of 75 min and ionized at 1.8 kV in the positive ion

mode. The gradient was composed of 2

6% solvent B (3.5 min), 6

25% B (42 min), 25

40% B

(14.5 min), and 40

–

98% B (15 min); solvent A: 2% ACN and 0.2% formic acid in water; solvent

B: 80% ACN

and 0.2% formic acid. MS1 scans were acquired at the resolution of 60,000 from

375 to 1500 m/z, AGC target 3e6, and maximum injection time 15 ms. The 12 most abundant

ions were targeted for MS2 scans acquired at a resolution of 30,000, AGC target 1e5, max

imum

injection time of 60 ms, and normalized collision energy of 28. Dynamic exclusion was set to 30

s and ions with charge +1, +7, +8 and >+8 were excluded. The temperature of ion transfer tube

was 275°C and the S

lens RF level was set to 60.

MS2 fragment

ation spectra were searched with

SEQUEST running within Proteome Discoverer (version 2.5, Thermo Scientific) against the

UniProt human reference proteome comprised of 79,052 proteins covering 20,577 genes

(UP000005640)

. The maximum missed cleavages was set

to 2. Dynamic modifications were set

to oxidation on methionine (M, +15.995 Da), deamidation on asparagine and glutamine (N and

Q, +0.984 Da), phosphorylation of serine and threonine (S and T, +79.966 Da), protein N

terminal acetylation (+42.011 Da),

and

protein N

terminal Met

loss (

131.040 Da)

Carbamidomethylation on cysteine residues (C, +57.021 Da) was set as a fixed modification. The

maximum parental mass error was set to 10 ppm, and the MS2 mass tolerance was set to 0.03

Da. The false discovery thre

shold was set strictly to 0.01 using the Percolator Node validated by

value. The relative abundance of parental peptides was calculated by integration of the area

under the curve of the MS1 peaks using the Minora LFQ node.

To identify enriched proteins,

proteins that were detected in both samples were ranked by iBAQ

intensity within each sample, and enrichment was assessed based on the difference between the

+UV and the

UV iBAQ rank. The final results of this analysis are listed in Supplementary Data

MS/MS analysis to assess differences in protein content between percoll gradient

enriched

mitochondria derived from K562 wild

type or MTCH2 depleted cells was performed with an

EASY

nLC 1200 (ThermoFisher Scientific, San Jose, CA) coupled to an Orbit

rap Eclipse

Tribrid mass spectrometer (ThermoFisher Scientific, San Jose, CA). Peptides were separated on

an Aurora UHPLC Column (25 cm × 75 μm, 1.6 μm C18, AUR2

25075C18A, Ion Opticks) with

a flow rate of 0.35 μL/min for a total duration of 75 min and ion

ized at 1.8 kV in the positive ion

mode. The gradient was composed of 2

6% solvent B (3.5 min), 6

25% B (42 min), 25

40% B

(14.5 min), and 40

–

98% B (15 min); solvent A: 2% ACN and 0.2% formic acid in water; solvent

B: 80% ACN and 0.2% formic acid. MS1 scan

s were acquired at the resolution of 120,000 from

350 to 1,600 m/z, AGC target 1e6, and maximum injection time of 50 ms. MS2 scans were

acquired in the ion trap using fast scan rate on precursors with 2

7 charge states and quadrupole

isolation mode (isolat

ion window: 0.7 m/z) with higher

energy collisional dissociation (HCD,

30%) activation type. Dynamic exclusion was set to 30 s. The temperature of ion transfer tube

was 300°C and the S

lens RF level was set to 30. MS2 fragmentation spectra were searched wi

SEQUEST running within Proteome Discoverer (version 2.5, Thermo Scientific) against the

reviewed sequences from the UniProt human reference proteome comprised of 20,361 proteins

(UP000005640). The maximum missed cleavages were set to 2. Dynamic modifi

cations were set

to oxidation on methionine (M, +15.995 Da), deamidation (N and Q, +0.984 Da), protein N

terminal acetylation (+42.011 Da) and protein N

terminal Met

loss (

131.040 Da).

Carbamidomethylation on cysteine residues (C, +57.021 Da) was set as a

fixed modification. The

maximum parental mass error was set to 10 ppm, and the MS2 mass tolerance was set to 0.6 Da.

The false discovery threshold was set strictly to 0.01 using the Percolator Node validated by q

value. The relative abundance of parental

peptides was calculated by integration of the area

under the curve of the MS1 peaks using the Minora LFQ node.

LFQ was then performed with the Minora feature detector, feature mapper, and precursor ions

quantifier nodes. Retention time alignment was perfo

rmed with maximum RT shift of 5 min and

a minimum S/N threshold of 10. Quantified peptides included unique + razor, protein groups

were considered for peptide uniqueness, shared Quan results were not used, Quan results with

missing values were not rejected

, and precursor abundance was based on extracted ion intensity.

Imputation was performed using a low abundance resampling method at the peptide level. The

Quantitative proteomics data was exported from ProteomeDiscoverer as an excel file and the

MitoCoP da

tabase

(

)

was used to remove all non

mitochondrial proteins from the dataset prior

to statistical analysis using the R statistical computing environment (R version 4.0.2). Protein

abundances were normalized b

etween samples with a random forest (R randomForest version

4.6

14) regression

(

)

, then modeled for expression differences (R limma version 3.44.3) using

the linear model fit, with fold changes calculated as a log2 diff

erence. Proteins and peptides

identified are listed in Supplementary Table 2.

Volcano plot figures were generated in Python using the matplotlib package. Figures that specify

submitochondrial localization are based on annotations from Mitocarta 3.0

(

)

, with the

following modifications: subce

llular localizations for TDRKH, HSDL1, and ARMC1 were

added manually based on published data

(

)

(

)

, TMEM11 is annotated as IM in Mitocarta

3.0; however, a recent preprint pro

vides evidence that TMEM11 is largely in the OM, so this

gene was included in analysis of OM proteins in our proteomic data

(

)

Photo

crosslinking of recombinant substrate with isolated

mitochondria

An insertion reaction was prepared in a buffer containing 100 mM

KoAc

, 50 mM HEPES pH

7.4, 2 mM

MgoAc

, 0.5 mM CaCl

. Percoll

gradient enriched mitochondria were added to a final

protein concentration of 0.2 mg/mL and CaM

Alfa

Sec61β

OMP25(B

pA) was added to a

final concentration of 1 μg/mL. Insertion was initiated by the addition of 2 mM EGTA. The

insertion reaction was split into

UV and +UV samples. The +UV sample was transferred to a

prechilled 6

well plate and left on top of an ice

cooled

aluminum block ~8 cm under a UVP B

100 lamp for 10 minutes.

Both samples were mixed with 10x volumes of IP buffer containing 100 mM

KoAc

, 50 mM

HEPES pH 7.4, 2 mM

MgoAc

, and 1% Triton X

100. After incubating on ice for 10 minutes,

samples were centrifuged

at 18,000 rpm. Supernatant was mixed with 10 μL packed streptavidin

agarose resin (Thermo Scientific Cat. #20357) which had been functionalized with biotinylated

Alfa

Nb, blocked with free biotin, and equilibrated in IP buffer. After 1 hour of binding hea

over

tail at 4°C, the unbound fraction was removed, and the resin was then washed 2x with 1 mL

IP wash buffer, 2x with 1 mL IP wash buffer + 0.5 M NaCl, and 2x with 1 mL IP wash buffer.

Resin was then incubated with 10 μL IP wash buffer + 300 nM SUMOstar

protease (LifeSensors

Cat # SP4110) for 30 minutes on ice. 1 mL IP wash buffer was added to protease treated resin

and incubated for 5 minutes on ice. Eluted protein was then removed from resin and TCA

precipitated by adding 1:10 volume 100% TCA and incub

ating on ice for 10 minutes, followed

by centrifuging at max speed in a chilled benchtop centrifuge for 10 minutes. The pellet was then

washed 2x with ice

cold acetone and prepared for mass spectrometry as described below.

Samples used for western blottin

g were prepared with the following differences: isolated

mitochondria were not purified on a percoll gradient; pierce magnetic streptavidin resin was used

(Thermo Scientific Cat. #88817), an IP buffer containing 200 mM NaCl, 50 mM HEPES pH 7.4,

1% Triton X

100 was used for solubilization; 4x 1 mL wash steps; and an elution buffer

containing 200 mM NaCl, 50 mM HEPES pH 7.4, and 0.05% Triton X

100 was used for 2x

wash steps and then elution was performed in the same buffer + 300 nM SUMOstar in a 10 μL

volume.

Proteoliposome reconstitutions and insertions

Reconstitutions of protein into liposomes were similar to previously described methods

(

)

(

)

. The following phospholip

ids were obtained from Avanti Polar Lipids: phosphatidyl

choline

(PC) and phosphatidyl

ethanolamine (PE) from egg, and synthetic 1,2

dioleoyl

glycero

phosphoethanolamine

lissamine rhodamine B (Rh

PE). The liposome mixture contained

PC:PE:Rh

PE at a

mass ratio of 8:1.9:0.1. Rh

PE was used to monitor recovery throughout the

reconstitutions and for quantification. Lipids were mixed at the indicated ratio as chloroform

stocks, adjusted to 10 mM DTT and dried by centrifugation under vacuum overnight. The

resulting lipid film was rehydrated to a final concentration of 20 mg/mL in lipid buffer (15%

glycerol, 50 mM HEPES pH 7.4) and mixed for 8 hours at 25°C until a homogenous mixture was

achieved. Lipids were then diluted with more lipid buffer and supplemen

ted with DBC to

produce a lipid/DBC mixture containing 2% DBC and 10 mg/mL lipids. BioBeads

SM2 (Bio

Rad) were prepared by activation with methanol, washing thoroughly with distilled water, and

then were resuspended in water into a final slurry where they

occupied 50% of the final volume.

For reconstitutions, excess liquid was removed from BioBeads by aspiration just before use.

Reconstitutions used purified EMC and MTCH2 in 0.25% DBC were obtained as described

above. In initial experiments, we determined t

he relative concentration of purified MTCH2

compared to the amount in isolated mitochondria from K562 cells. Different dilutions of purified

MTCH2 were mixed with constant amounts of lipids and adjusted to a final buffer concentration

of 100 mM NaCl, 25 mM

HEPES pH 7.4, 2 mM MgCl

, 0.8% DBC. Liposomes were made

using the same buffer and detergent conditions. A standard 100 μL reaction contained 10

40 μL

purified MTCH2, 20 μL of the 10 mg/mL lipid/DBC mixture, and the remaining volume made

up with buffer, sa

lts, and detergent. This protein/lipid/detergent mixture was added to 120 μL

BioBeads in 1.5 mL Eppendorf tubes. The slurry was mixed in a thermomixer for 18 hours at

4°C. The fluid phase was removed and to another 1.5 mL Eppendorf tube containing 120 μL

iobeads, and mixed in a thermomixer for 2 hours at 23°C. The fluid phase was then separated

and diluted with ten volumes of ice

cold water. The proteoliposomes were sedimented in a

TLA120.2 rotor at 70,000 rpm for 30 minutes, and resuspended in 16.7 μL lip

osome

resuspension buffer (100 mM

KoAc

, 50 mM HEPES pH 7.4, 2 mM

MgoAc

2, 250 mM sucrose,

1 mM DTT). Substrates for insertions were prepared by translating in RRL for 15 min, then the

reactions were treated with 1 mM puromycin to prevent further synthesis. In

sertion reactions

consisted of 8 μL translation in RRL, 1 mM EGTA and 2 μL buffer, liposomes,

proteoliposomes, or isolated mitochondria. The reactions were incubated at 32°C for 30 min,

before being treated with 0.5 mg/mL PK for 1 hour on ice. The reaction

s were quenched and the

protected fragments enriched with NiNTA resin as described above.

Flow cytometry

For all reporter experiments, respective K562 cell lines were spinfected with lenti

virus for

indicated constructs and analyzed by flow cytometry af

ter 48

72 hours. All reporter experiments

were performed at least twice. For the apoptosis experiment, wildtype K562 cells expressing

either MTCH2

P2A

BFP or BFP

alone were treated with 5

M imatinib mesylate (461080010,

Fisher Scientific, USA) for 72 hour

s. Cells were harvested, washed once with ice cold PBS, then

resuspended in 100

L staining buffer (10 mM HEPES, 140 mM NaCl, 2.5 mM CaCl2 pH 7.4,

5% Annexin

FITC (Invitrogen, A13199), 50

g/mL propidium iodide (P1304MP, Invitrogen,

USA) before analysis by

flow cytometry. For treatment with the GPAT inhibitor

FSG67, K562

cells were treated for 16 hours with 75

M FSG67 (Cedarlane Laboratories Cat. #10

4577) as

previously described

(

)

prior to analysis by flow cytometry.

All samples were run on either an

NXT Flow Cytometer (ThermoFisher) or a MACSQuant VYB (Mitenyi Biotec). Flow cytometry

data was analyzed either in FlowJo

v10.8 Software (BD Life Sciences)

or Python using the

FlowCytom

etryTools package.

Microscopy

To visualize localization to the mitochondria, K562 cells were stained with 1

M Mitotracker

Deep Red FM (

M22426

, ThermoFisher) for 30 minutes. Cells were collected, spun down,

washed in fresh media and plated onto a 96

ll glass bottomed plate (

160376

, Thermo Fisher).

K562 cells were then briefly spun down by centrifugation at 100

for 5 minutes and imaged on a

Zeiss LSM710 NLO Laser Scanning confocal microscope.

Sequence alignments

An alignment of individual SLC25 rep

eats from various human transporter was generated using

the hmmalign tool from the HMMER3.3 package

(

)

and a precomputed hmm profile of the

SLC25 family from pfam (pf00153)

(

)

ig S1. In vitro protease protection assays into human mitochondria

(A)

The integrity of the

outer membrane was tested using both semi

permeabilized K562 cells and mitochondrially

enriched membrane fractions by treating each with decreasing amounts of proteinase K (PK). The

resulting reactions were analyzed by western blot fo

r the inner mitochondrial membrane protein

mitofilin, which should be inaccessible to PK when the outer membrane is intact.

(B)

Further

purification of mitochondria from K562 cells using a percoll gradient and density centrifugation.

Blots for a mitochondr

ial marker (SAM50) and an ER marker (SEC61β) demonstrate relative

proportion of mitochondria and ER in the ‘mitochondrial enriched membranes’ that are used for

most in vitro experiments. Proteomics and crosslinking

mass spectrometry experiments were

perfor

med with percoll

gradient enriched mitochondria and used fractions 8

10 (Fig. 2A, Fig.

3C).

(C)

Schematic of the basic

in vitro

insertion assay, using protease protection as a readout for

insertion. In all insertion assays, substrates were translated using

an in vitro translation reaction

(IVT) in rabbit reticulocyte lysate and

methionine labelled, permitting detection by

autoradiography. Translation was terminated by addition of puromycin, followed by incubation

with membranes enriched in mitochondria derived from human K562 cells. After incubating with

membranes, reactions we

re treated with PK. For TA proteins, the resulting protease protected

band was immunoprecipitated via a tag on its C

terminus, ensuring insertion into the outer

membrane in the correct orientation, and analyzed by SDS

PAGE and autoradiography.

(D)

The

IMS

localized TOM substrate composed of the cytochrome b targeting sequence fused to the inert

protein DHFR (cyt. b

DHFR) was translated and incubated with purified mitochondria to

confirm their activity. Insertion was detected by protease protection as descr

ibed.

ig S2. A CRISPRi screening platform to identify factors involved in mitochondrial TA

protein biogenesis in human cells

(A)

Microscopy showing that the IMS

targeting sequence

from MICU1 conjugated to full length GFP results in its localization to

the mitochondria in

mammalian K562 CRISPRi cells.

(B)

The endogenous sequences of two TA proteins: SQS,

which is localized to the ER, and OMP25, which under these conditions is dual

localized to both

the outer membrane and ER, were appended to a C

termina

l GFP11 in a backbone containing a

translational control (RFP) separated by a viral 2A sequence (see Fig. 1

. These constructs were

independently introduced into cells expressing IMS localized GFP1

10 and analyzed by flow

cytometry.

(C)

Histograms of (B)

comparing GFP fluorescence for OMP25 and SQS compared

to a mock transduced control. The marked increase in GFP fluorescence suggest that OMP25,

but not SQS can successfully conjugate with GFP1

10 localized to the IMS.

(D)

Workflow of the

FACS

based CRISPRi

screen. A K562 CRISPRi reporter cell line was constructed that

constitutively expressed GFP1

10 in the IMS and the OMP25

GFP11 reporter under an inducible

promoter. For the screen, these cells were transduced with a genome

scale CRISPRi sgRNA

library and

then the OMP25

GFP11 reporter was induced with doxycycline for 24 hours prior to

cell sorting. Cells were sorted based on ratiometric changes in GFP relative to RFP, and sgRNAs

expressed in the isolated cells were identified using deep sequencing.

S3. UBQLN1 is a quality control factor for mitochondrial TAs

. K562 CRISPRi cells

expressing IMS GFP1

10 were depleted of UBQLN1 using two different sgRNAs. Reporters

were introduced for either two mitochondrial TAs (OMP25 and MAVS) or an IMS localized

con

trol (TI

M9A) and cells were analyzed by flow cytometry. Lack of UBQLN results in a

ratiometric increase in GFP:RFP fluorescence for mitochondrial TAs, consistent with its

previously reported role in targeting mislocalized mitochondrial TAs for degradation

by the

ubiquitin proteasome pathway

(

)

ig S4. The EMC is required for insertion of mislocalized mitochondrial TAs to the ER

(A)

Volcano plot of the genome

wide CRISPRi screen with EMC subunits shown

in black.

(B)

panel of ER (SQS and VAMP) and mitochondrial (OMP25 and MAVS) TA proteins were

conjugated to a C

terminal opsin epitope. The opsin epitope contains a consensus glycosylation

sequence that is modified upon insertion into the ER lumen. These

constructs were then

translated in rabbit reticulocyte lysate in the presence of

methionine. The reactions were

puromycin treated and incubated with either wild

type (WT) or EMC knockout (KO) semi

permeabilized (SP) cells. Insertion into the ER, as mo

nitored by appearance of a glycosylated

band (‘glyc’), was dependent on the EMC for its canonical substrate SQS, and both

mitochondrial TAs. In contrast, VAMP’s insertion was unaffected by EMC knockout, consistent

with its previously reported dependence on

the GET pathway for insertion

(

)

ig S5

Assessing the effects of lipid biogenesis defects on mitochondrial TAs.

(A)

Flow

cytometry analysis as in Fig.

D but with an alternative IMS targeting sequence derived from

LACTB appended to the GFP1

(

)

. (

) Volcano plot of the genome

wide CRISPRi screen

indicating MTCH2 (in light blu

e) and factors previously implicated in the regulation of outer

membrane fatty acid synthesis or transport (in black).

(C)

K562 IMS GFP1

10 expressing cells

were treated with the pan GPAT inhibitor FSG67 for 16 hours (75

M) or a vehicle. MTCH2

dependent m

itochondrial fusion has been shown to require Glycerol 3

phosphate acyltransferases

(GPATs) catalysed LPA synthesis. A reporter expressing either a mitochondrial TA (OMP25) or

an IMS localized control (TI

M9A) were expressed in GPATi cells and analyzed by

flow

cytometry.