letters

COPA mutations impair ER-Golgi transport and cause

hereditary autoimmune-mediated lung disease and arthritis
Levi B Watkin1,2,16, Birthe Jessen3,16, Wojciech Wiszniewski4,16, Timothy J Vece1, Max Jan3, Youbao Sha5,
Maike Thamsen3, Regie L P Santos-Cortez6, Kwanghyuk Lee6, Tomasz Gambin4, Lisa R Forbes1,2,
Christopher S Law3, Asbjørg Stray-Pedersen2,4, Mickie H Cheng3, Emily M Mace1,2, Mark S Anderson3,
Dongfang Liu1,2, Ling Fung Tang7, Sarah K Nicholas1,2, Karen Nahmod1,2, George Makedonas1,2, Debra L Canter1,2,
Pui-Yan Kwok7,8, John Hicks9, Kirk D Jones10, Samantha Penney4, Shalini N Jhangiani11, Michael D Rosenblum8,
Sharon D Dell12, Michael R Waterfield13, Feroz R Papa3, Donna M Muzny11, Noah Zaitlen3, Suzanne M Leal6,
© 2015 Nature America, Inc. All rights reserved.

Claudia Gonzaga-Jauregui4, Baylor-Hopkins Center for Mendelian Genomics14, Eric Boerwinkle11,15,

N Tony Eissa5, Richard A Gibbs4,11, James R Lupski1,4,11,17, Jordan S Orange1,2,17 & Anthony K Shum3,7,17

Unbiased genetic studies have uncovered surprising molecular might also be anticipated to cause autoimmunity, as disruptions in
mechanisms in human cellular immunity and autoimmunity1. protein trafficking lead to endoplasmic reticulum (ER) stress and
We performed whole-exome sequencing and targeted activation of the unfolded protein response (UPR), both of which
sequencing in five families with an apparent mendelian have been implicated in autoimmune disease10.
syndrome of autoimmunity characterized by high-titer We identified five families with a previously undescribed mendelian
autoantibodies, inflammatory arthritis and interstitial lung syndrome of autoimmunity manifested by high-titer autoantibodies,
disease. We identified four unique deleterious variants in interstitial lung disease and inflammatory arthritis (Fig. 1a–d, Table 1
the COPA gene (encoding coatomer subunit a) affecting the and Supplementary Table 1). The average age of presentation was
same functional domain. Hypothesizing that mutant COPA 3.5 years, with a range of 6 months to 22 years. Several patients pre-
leads to defective intracellular transport via coat protein sented with pulmonary hemorrhage requiring immunosuppression,
complex I (COPI)2–4, we show that COPA variants impair and all patients have lung disease (Table 1 and Supplementary Table 1).
binding to proteins targeted for retrograde Golgi-to-ER A comparison of lung biopsies from unrelated patients identified
transport. Additionally, expression of mutant COPA results lymphocytic interstitial infiltration with germinal center formation
npg

in ER stress and the upregulation of cytokines priming for a (Fig. 1b,c), findings consistent with the interstitial lung disease occur-
T helper type 17 (TH17) response. Patient-derived CD4+ T cells ring in systemic autoimmune syndromes11. Immunohistochemical
also demonstrate significant skewing toward a TH17 phenotype staining of lungs identified CD20 + B cells within the germinal
that is implicated in autoimmunity5,6. Our findings uncover centers and substantial numbers of lung-infiltrating CD4+ T cells
an unexpected molecular link between a vesicular transport (Fig. 1b,c). Autoantibodies were detected in 86% of the affected
protein and a syndrome of autoimmunity manifested by patients, including anti-nuclear antibodies (ANAs), anti-neutrophil
lung and joint disease. cytoplasmic antibodies (ANCAs) and rheumatoid factor (RF) (Table 1
and Supplementary Table 2). Immunoglobulin levels, absolute
Monogenic disorders have proven powerful in elucidating biological lymphocyte counts and CD4/CD8 cell ratios were largely normal
mechanisms underlying autoimmunity7,8 by showing that autoimmu- (Supplementary Table 3). Ninety-five percent of the patients have
nity can arise from perturbations in several non-classical pathways7. arthritis, and some initially presented with joint pain. Four unrelated
Defects in the intracellular trafficking mechanisms of immune cells9 patients underwent renal biopsies that demonstrated immune-mediated

1Department of Pediatrics, Baylor College of Medicine, Houston, Texas, USA. 2Texas Children’s Hospital Center for Human Immuno-Biology, Houston, Texas, USA.
3Department of Medicine, University of California, San Francisco, San Francisco, California, USA. 4Department of Molecular and Human Genetics, Baylor College
of Medicine, Houston, Texas, USA. 5Department of Medicine, Baylor College of Medicine, Houston, Texas, USA. 6Center for Statistical Genetics, Baylor College
of Medicine, Houston, Texas, USA. 7Cardiovascular Research Institute, University of California, San Francisco, San Francisco, California, USA. 8Department of
Dermatology, University of California, San Francisco, San Francisco, California, USA. 9Department of Pathology, Texas Children’s Hospital, Houston, Texas, USA.
10Department of Pathology, University of California, San Francisco, San Francisco, California, USA. 11Human Genome Sequencing Center, Baylor College of Medicine,

Houston, Texas, USA. 12Division of Respiratory Medicine, Hospital for Sick Children, Toronto, Ontario, Canada. 13Department of Pediatrics, University of California,
San Francisco, San Francisco, California, USA. 14Houston, Texas, USA. 15Human Genetics Center and Institute of Molecular Medicine, University of Texas–Houston
Health Science Center, Houston, Texas, USA. 16These authors contributed equally to this work. 17These authors jointly supervised this work. Correspondence should
be addressed to A.K.S. (anthony.shum@ucsf.edu) or J.S.O. (orange@bcm.edu).

Received 16 October 2014; accepted 19 March 2015; published online 20 April 2015; doi:10.1038/ng.3279

654 VOLUME 47 | NUMBER 6 | JUNE 2015  Nature Genetics

 letters

a d I
1 2

1 2 3 4 5 6 7 8 9 10 11
R II

Family A
+/m +/+ +/m +/+
1 2 3 4 5 6 * 7 8 9 10
III
+/+ +/m +/m +/m +/m
1 * 2

b IV
+/m

1 2
I
+/m +/+

Family B
1 2 3 4 5 6 7 8
II
+/+ +/m +/+ +/m +/+ +/+
1 * 2 3 4 5 6 7 8
H&E CD20
III
+/m +/m +/m +/m +/+ +/+ +/+ +/+

1 2
I

1 2 3 4 5 6
II
© 2015 Nature America, Inc. All rights reserved.

1 2 * 3 4 5 6 7 8 9 10 * 11 12 13 14 15 16 17 18
CD8 CD4
III
c +/+ +/+ +/+
Family C

+/m +/m +/+

1* 2* 3 4 5 6 7 8 9 10 11 12 13 14 15 * 16 * 17 18 19
IV
+/m +/m +/m +/m
1 * 2 3 4 5 6 7 8 * 9 *
V
+/m +/+ +/m
1
H&E CD20 VI

1 2 *
I 1 2
+/m I

Family E
1* 2 3 4 5*
Family D

* * +/m
II 1 2
+/+ +/m +/m +/+ II
1* 2 * 3* 4 * 5* 6*
+/m +/+
II
CD8 CD4 +/m +/+ +/m +/+ +/m +/+

Figure 1  Genetic segregation of interstitial lung disease and arthritis with COPA mutations. (a) Left, a chest computed tomography (CT) scan of patient
npg

B.III.1 shows basilar ground glass opacities (circle), reticulations (red arrow) and cystic formations (black arrow). Right, a bilateral knee radiograph
from A.III.6 before restorative surgery. The radiograph shows generalized osteopenia with pathologic fracture of the distal femur. There is bilateral
avascular necrosis of the metaphysis and diaphysis of the distal femur and proximal tibia with slackening of the condyles and articular cartilage
destruction. (b,c) Lung biopsy sections from patients A.III.6 (b) and B.III.1 (c) stained with hematoxylin and eosin (H&E) show cellular interstitial
infiltrates and germinal center formation; immunohistochemical analysis demonstrates the presence of CD20 + B cells and CD4+ and CD8+ T cells.
Scale bars, 100 µm. (d) The pedigrees for families A–E are shown with the presence of wild-type (+) or mutant (m) COPA alleles. Family A (c.698G>A)
shows incomplete penetrance, with one unaffected carrier over four generations. Family B (c.728A>G) shows incomplete penetrance of severe disease
in females affected over three generations. Family C (c.721G>A) shows incomplete penetrance of severe disease and segregation of the mutation
within affected patients over six generations. In family D (c.690G>T), there is an unaffected mutation carrier who is 1 year old, below the age of onset.
Family E (c. 698G>A) shows complete penetrance over two generations. Square, male; circle, female; black filled, affected with severe disease;
unfilled, unaffected; slash, deceased; asterisk, submitted for whole-exome sequencing; vertical line, unaffected carrier. Roman numerals represent
the generation within the family, and Arabic numerals designate individuals within the generation.

renal disease (Supplementary Fig. 1 and Supplementary Table 1),
providing further evidence of autoimmunity. All patients have
required long-term immunosuppression.
We enrolled patients for whole-exome sequencing to identify
genetic variants potentially associated with disease. Family pedi-
grees suggested autosomal dominant inheritance with incomplete
penetrance (Fig. 1d). We obtained whole-exome sequencing data for
patients A.III.6, A.IV.1 and B.III.1 and carried out analysis to identify
shared variants. There was only one gene, COPA, observed to have
variants in both families, with the two individuals in family A having
a c.698G>A (p.Arg233His) variant and individual B.III.1 from family
B having a c.728A>G (p.Asp243Gly) variant (NM_004371.3). These
variants had not been identified in several large databases12. The
variants cosegregated with the disease phenotype, and we identified
them in all other affected family members (Supplementary
Fig. 2). For family C, we performed whole-exome sequencing analysis
and direct sequencing for five affected individuals (C.IV.1, C.IV.2,
C.IV.15, C.V.1 and C.V.9) and four unaffected pedigree members
(C.III.2, C.III.10, C.IV.16 and C.V.8) and identified a c.721G>A
(p.Glu241Lys) COPA variant on chromosome 1 to be the only rare,

Nature Genetics  VOLUME 47 | NUMBER 6 | JUNE 2015 655

 letters

Table 1  Demographic and clinical characteristics of affected four variants of the COPA gene in five affected families, that the rare
individuals with COPA mutations variants cosegregated with the disease and that the variation at this
Demographic and clinical characteristics Patients n (%) locus is linked to the autoimmunity phenotype, we conclude that it
Total 21 is highly likely that pathological variants in COPA are the underlying
Age at presentation <5 years 16 (76) cause of this syndrome.
Sex COPA encodes the COPA subunit of COPI. COPI and COPII are
  Male 8 (38) carrier complexes required for membrane trafficking between the ER
  Female 13 (62) and the Golgi. COPII facilitates the anterograde movement of newly
Symptom at initial presentation created proteins or lipids from the ER toward the cell surface, whereas
  Tachypnea, cough, hemoptysis 14 (67) COPI is engaged in the retrograde movement of proteins from the
  Joint pain 5 (24) Golgi to the ER14. To investigate potential mechanisms by which the
Arthritis 20 (95) COPA variants might lead to disease, we measured the levels of COPA
Pulmonary manifestations transcript and COPA protein and found no difference in expression
  Hemorrhage or interstitial lung disease 21 (100)
levels between patient and control B-lymphoblastoid cell lines (BLCLs)
Autoantibodies 18 (86)
(Supplementary Fig. 3a–c). Imaging analysis demonstrated that the
  ANAs 14 (67)
COPA variants did not cause aggregation or mislocalization of COPA
  ANCAs 15 (71)
protein thereby impairing its function4 (Supplementary Fig. 3d).
  RF 9 (43)
Given that COPA was expressed at normal levels in patients, we
Response to immunosuppression 21 (100)
sought to determine a functional consequence of the COPA variants.
Notably, point mutations mapping to the COPA WD40 domain in
nonsynonymous variant that segregated with the disease phenotype yeast, analogous to the disease-associated damaging alleles identified
© 2015 Nature America, Inc. All rights reserved.

(Table 2 and Supplementary Fig. 2). In family D, whole-exome
sequencing of subjects D.I.2, D.II.1–D.II.5 and D.III.1–D.III.6 iden-
tified another COPA variant, c.690G>T (p.Lys230Asn), that cosegre-
gated with the disease phenotype. In family E, given that the clinical
phenotype was nearly identical to that for the other patients with
COPA alterations, we performed targeted Sanger sequencing of
COPA exons 8 and 9 and identified the c.698G>A (p.Arg233His)
missense variant that we also discovered in family A (Table 2 and
Supplementary Fig. 2). Interestingly, genotype analysis of families
A and E identified the same ancestral haplotype in the COPA locus
spanning approximately 140 kb and 29 SNPs. The presence of a com-
mon haplotype provides a potential explanation for the co-occurrence
of the COPA c.698G>A (p.Arg233His) variant in these two apparently
unrelated families (Supplementary Table 4). Overall, we employed
whole-exome sequencing and targeted sequencing at two institutions
and independently identified four unique variants in five affected
families. Within these 5 families, 21 of 30 COPA mutation carriers
were affected by disease whereas 9 were unaffected clinically, consist-
npg
in the patients, impair binding of dilysine-tagged proteins and cause a
defect in retrograde transport2,3,15. To test whether binding of dilysine
proteins to the COPA mutants identified was similarly impaired, we
generated a synthetic peptide corresponding to the cytoplasmic tail of
yeast Wbp1 (Wbp1p peptide), which bears a dilysine retrieval signal
that binds to mammalian COPA2,16. We incubated recombinant COPA
protein with Wbp1p peptide and found that Glu241Lys mutant COPA
had impaired binding to Wbp1p peptide in comparison to wild-type
COPA (Fig. 2b). We next transfected cell lines stably expressing
Glu241Lys mutant COPA with a construct for a reporter protein bear-
ing a dilysine retrieval signal. In the subsequent immunoprecipitation
assays, we also found impaired binding of the dilysine reporter to
Glu241Lys mutant COPA (Fig. 2c). We obtained comparable results in
experiments using Lys230Asn mutant COPA (Supplementary Fig. 4).
Taken together, these results demonstrate that mutant COPA bearing
a WD40 variant identified in patients behaves in an anomalous way,
analogous to the altered activity observed for the point mutations in
yeast shown to cause a defect both in the binding of dilysine-tagged

ent with incomplete penetrance. Notably, all variants identified were proteins and retrograde transport2,3,15.
predicted to be deleterious and mapped to a 14-residue region in the We next sought to identify physiological effects of mutant COPA
highly conserved, functionally important WD40 domain of the COPA that might be linked mechanistically to autoimmune disease. Defects
protein3 (Fig. 2a and Table 2).
To provide statistical evidence that variants Table 2  Bioinformatic evaluation of segregating missense variants in COPA
in COPA are involved in disease etiology, we
Family ID A and E B C D
performed an affected-only parametric link-
hg19 physical position 160,293,229 160,283,894 160,283,901 160,293,237
age analysis on the four families that were
phyloP (100wayall) 7.12 7.53 7.26 2.81
informative, analyzing the COPA missense
13 PhastCons 1 1 1 1
variants in these families . No accurate
GERP 4.73 5.62 5.62 2.87
estimate of the penetrance for this rare pheno- cDNA change c.698G>A c.728A>G c.721G>A c.690G>T
type exists. We assumed that the disease allele Amino acid change p.Arg233His p.Asp243Gly p.Glu241Lys p.Lys230Asn
frequency and variant allele frequency were PolyPhen-2 HVAR Probably damaging Probably damaging Probably damaging Probably damaging
both 0.001. The four informative pedigrees SIFT Damaging Damaging Damaging Damaging
provided the following logarithm of odds LRT Damaging Damaging Damaging Damaging
(LOD) scores at θ (recombination fraction) = 0 MutationAssessor Low Neutral Low Medium
for the identified COPA variants: family A, MutationTaster Disease causing Disease causing Disease causing Disease causing
1.5; family B, 0.90; family C, 2.7; family D, CADD (scaled) 32 26.7 35 19.81
0.60. The combined statistically significant phyloP, phastCons and GERP scores for nucleotide conservation were derived from the UCSC Genome Browser.
LOD score of 5.7 for these families provides PolyPhen-2, SIFT, LRT, MutationAssessor and MutationTaster provided a functional prediction for each missense
variant. For Combined Annotation-Dependent Depletion (CADD), a scaled C score ≥10 places a variant within the
statistical evidence that COPA is involved top 10% of likely deleterious variants, whereas a C score ≥20 indicates that the variant is within the top 1% of likely
in disease etiology. Given that we observed deleterious variants across the genome.

656 VOLUME 47 | NUMBER 6 | JUNE 2015  Nature Genetics

 letters

Figure 2  The binding of mutant COPA to dilysine motifs is impaired. a p.R233H

p.E241K
(a) A schematic of the COPA protein and its domains is shown.
p.K230N p.D243G
The amino acid positions corresponding to the mutations identified and
the conserved sequence homology at these sites in distantly related WD40 repeat Coatomer WD40-associated region COOH
eukaryotic organisms are indicated. (b) Binding of mutant COPA to a 1 1,224
dilysine motif was examined by incubating Flag-tagged wild-type (WT) or Amino acid position 228 230 233 241 243 257
mutant Glu241Lys COPA with Wbp1p peptide–coupled agarose beads. Homo sapiens
Gallus gallus
Protein-bead complexes were analyzed by immunoblot (IB) using an Bos taurus
antibody to Flag. The number below each lane is the relative protein Tursiops truncatus
Mus musculus
amount expressed as the percentage of wild-type COPA levels. Statistical Drosophila melanogaster
analysis of six independent experiments is shown. ***P < 0.001. Danio rerio
(c) Binding of mutant COPA to a dilysine motif was also assessed Saccharomyces cerevisiae

within T-REx 293 cells stably expressing Flag-tagged wild-type or mutant

COPA by transiently transfecting cells with an expression construct for
b 125 Wbp1p

COPA-Flag (% of WT)
100
GFP-CD8-KKTN. Cell lysates were precipitated with antibody to CD8, Wbp1p beads Input
and immunoprecipitates were assayed for COPA-Flag expression and COPA-Flag COPA-Flag
75
GFP-CD8 enrichment. The number below each lane is the relative ***
IB WT E241K WT E241K 50
protein amount expressed as the percentage of wild-type COPA or
Flag 25
GFP-CD8 levels in cells expressing wild-type COPA. Statistical analysis
of five independent experiments is shown. ***P < 0.001. 100 58 100 115 0
WT E241K

c IP: CD8 Input

125
IP: CD8
in the COPI pathway, including ones caused by WD40 variants in

COPA-Flag (% of WT)
COPA-Flag COPA-Flag 100
COPA2,3,15, have been shown to affect retrograde trafficking and to IB WT E241K WT E241K
© 2015 Nature America, Inc. All rights reserved.

75
also indirectly affect COPII-mediated anterograde trafficking from Flag
***
the ER to the Golgi4. Thus, we hypothesized that the COPA variants 50
100 55 100 90
in patients likewise impair intracellular protein trafficking and, as a GFP 25
result, lead to ER stress and activation of the UPR4,17. Notably, ER (CD8)
0
stress and activation of the UPR have been linked to autoimmunity and 100 100 100 107
WT E241K
lung disease10,18,19. To determine whether patients exhibited increased
ER stress, we examined lung biopsies for expression of the molecular in this pathway. Fluorescent imaging of cells showed an increase in
chaperone binding immunoglobulin protein (BiP) 20. We found autophagosome and endolysosome size (Supplementary Figs. 6a–f
that BiP levels were markedly increased in the lung epithelium and and 7a–d), and patient-derived BLCLs demonstrated a decrease in the
alveolar macrophages of patients in comparison to controls (Fig. 3a levels of the autophagy substrate p62 and torin-induced p62 catabo-
and Supplementary Fig. 5a,b), consistent with our hypothesis. lism (Supplementary Fig. 8a,b). This constellation of findings sug-
We next performed dynamic functional assessments of patient- gests a baseline defect in autophagic function that might be due to
derived cells for evidence of enhanced ER stress caused by mutant impaired early endosomal function in patients24. Thus, it is possible
COPA. We treated BLCLs from patients with low levels of thapsigargin, that impaired autophagy exacerbates the increase in ER stress caused
an inhibitor of the ER Ca2+ ATPase used experimentally to induce ER by mutant COPA25. In aggregate, our findings demonstrate that the
stress. Cells from patients with COPA variants had significantly elevated expression of mutant forms of COPA in patients is sufficient for the
levels of transcript for HSPA5 (which encodes BiP) (Fig. 3b and induction of ER stress.
Supplementary Fig. 5c) in comparison to control cells from related Finally, to investigate potential immunological links between
npg

and unrelated individuals (P < 0.05). Interestingly, the clinically asymp-
tomatic COPA mutation carriers (with apparent non-penetrance)
demonstrated a weaker increase in ER stress in comparison to those
severely affected (Fig. 3b), suggesting that the mutation’s effect on
healthy carriers is below a threshold required for clinical disease.
To establish the specific role of mutant COPA in the induction of
ER stress, we performed small interfering RNA (siRNA)-mediated
knockdown experiments in HEK (human embryonic kidney) cells.
Indeed, we found that reducing COPA expression resulted in higher
levels of ER stress, as indicated by a marked increase in HSPA5 (BiP)
expression (Fig. 3c). To demonstrate that the increased ER stress
in patient-derived cells was specifically due to mutant COPA, we over-
expressed wild-type or mutant COPA in HEK cells (Supplementary
Fig. 5e,f). Consistent with a dominant-negative effect of the muta-
tions, presence of any of the four COPA mutations identified in
patients resulted in elevated HSPA5 (BiP) levels in comparison to
cells transfected with wild-type COPA (Fig. 3d and Supplementary
Fig. 5d). The levels of additional ER stress markers such as ATF4 and
DDIT3 (CHOP) were also significantly increased when COPA expres-
sion was reduced or mutant COPA was present (Supplementary
Fig. 5g,h). Given the role of autophagy in alleviating ER stress21 and
regulating immune responses22,23, we sought to evaluate alterations
mutant COPA and the clinical phenotype, we examined CD4+ T cells,
given their high numbers in diseased lungs (Fig. 1b,c) and their critical
function in autoimmune disease26. Interestingly, recent studies have
shown that ER stress has a role in the generation of TH17 cells27–30,
a CD4+ effector T cell population implicated in autoimmunity5,6,
by causing antigen-presenting cells (APCs) to secrete cytokines
necessary for the induction and expansion of TH17 populations.
Thus, we hypothesized that mutant COPA might lead to the production
of TH17 cells through a similar mechanism.
Given that B cells are professional APCs, we tested our BLCLs for
the expression of cytokines involved in the generation of the CD4 +
TH1, TH2 and TH17 helper T cell subsets. Patient-derived BLCLs with
mutant COPA showed a significant elevation in the levels of tran-
scripts encoding interleukin (IL)-1β, IL-6 and IL-23 (p19 and p40
subunits) (Fig. 4a), cytokines that are critical for the priming and
expansion of TH17 populations31, and this effect was enhanced by
ER stress. In contrast, transcript levels for the TH2-priming cytokine
IL-4 in patients were similar to those in controls, whereas transcript
levels for the TH1-priming cytokine IL-12 (p35 subunit) were very
low in both groups (Fig. 4a). Transcript levels for IL-10 and trans-
forming growth factor (TGF)-β were not different between patients
and controls, and interferon (IFN)-α (IFNA2) levels in patients were

Nature Genetics  VOLUME 47 | NUMBER 6 | JUNE 2015 657

 letters

a
Control E241K D243G R233H

b Healthy controls E241K D243G R233H

c d
20 1.5 8 6
**

Fold increase HSPA5/GAPDH

Fold increase HSPA5/GAPDH

Fold increase HSPA5/GAPDH

Fold increase COPA/GAPDH

*** ***
15 6 **
1.0 4
*
10 ** * 4
*
* 0.5 2
5 2

***
0 0 0 0
TG – + – + – + – + – + – + – + – + – + – + – + CTRL COPA CTRL COPA WT E241K D243G R233H
CO1 CO2 CO3 CO4 HC1 PA1 PA2 HC2 PA4 PA5 PA6 siRNA siRNA siRNA siRNA

Figure 3  Mutant COPA leads to an increase in ER stress. (a) Immunohistochemistry for the ER stress component BiP in lung biopsies from patients with
© 2015 Nature America, Inc. All rights reserved.

the Glu241Lys, Asp243Gly and Arg233His COPA variants is shown. A non-carrier is shown as a control. Scale bars, 40 µm. Enlarged insets emphasize
BiP staining in lung epithelial cells. (b) Quantitative PCR (qPCR) of HSPA5 (BiP) expression in untreated (white; −) and thapsigargin (TG)-treated
(striped, gray and black; +) BLCLs. The leftmost graph shows the results from two unrelated controls (CO1, CO2) and two related non-carrier controls
(CO3, CO4). Results were normalized to values for untreated CO1. Data shown are from healthy carriers HC1 and HC2 (C.III.2, B.III.4) and patients
PA1, PA2 and PA4–PA6 (C.IV.15, C.V.1, B.III.3, A.II.5, A.III.6). One-way ANOVA followed by Dunnett’s multiple-comparison test was used to compare
thapsigargin-treated experimental groups with thapsigargin-treated CO1 as a reference. Means and s.e.m. of data pooled from 2–4 independent
experiments are shown. *P < 0.05, **P < 0.01, ***P < 0.001. (c) COPA knockdown in HEK293 cells. The left graph demonstrates >90% reduction
in COPA mRNA levels after treatment with COPA siRNA. The right graph shows induction of ER stress after COPA knockdown as measured by HSPA5
(BiP) levels. Means and s.e.m. of data pooled from 3 independent experiments each with n = 3–4 (pooled n = 11) are shown. Statistical comparisons
were made using Student’s unpaired t tests with Welch’s correction. **P < 0.01, ***P < 0.001. CTRL, control. (d) HEK293 cells were transfected
with vectors for wild-type COPA or mutant Glu241Lys, Asp243Gly or Arg233His COPA, and HSPA5 (BiP) mRNA levels were measured 48 h after
transfection. Means and s.e.m. are shown. One-way ANOVA followed by Dunnett’s multiple-comparison test was used to compare experimental groups
with wild-type control. Results are shown for data pooled from three independent experiments. *P < 0.05, **P < 0.01, ***P < 0.001.

increased and decreased in comparison to controls (Supplementary
Fig. 9). To determine whether CD4+ T cells in patients demonstrated
an immunological phenotype consistent with this pattern of cytokine
expression, we profiled peripheral blood mononuclear cells (PBMCs).
Intracellular cytokine staining demonstrated equal frequencies of
npg
disorders in which alterations to widely expressed proteins lead
to distinct clinical phenotypes32,33. For instance, four mendelian
syndromes with disparate clinical manifestations result from
mutations affecting the COPII trafficking machinery34. We can only
speculate that, in patients with COPA mutations, there are tissue-

IL-13–secreting TH2 cells among patient and control cells (Fig. 4b,c).
By contrast, patient CD4+ T cells demonstrated a marked skewing
away from IFN-γ–secreting TH1 cells and a significant increase
in the frequency of IL-17A–secreting TH17 cells (Fig. 4b,c and
Supplementary Fig. 10). Thus, mutant COPA leads both to higher
levels of ER stress in patient-derived BLCLs and the production of
cytokines that promote the expansion of TH17 populations.
Through genetic studies in five unrelated families, we provide
evidence that point mutations affecting a 14-residue region of the
WD40 domain of COPA cause a systemic autoimmune syndrome of
lung and joint disease. Functional analyses demonstrate that mutant
COPA is unable to bind proteins with dilysine-based motifs and pro-
vide evidence of defective intracellular trafficking. Specifically, we
show that expression of mutant COPA in cells induces higher levels of
ER stress, which may be further exacerbated in patients by impaired
autophagy25. Finally, we establish a putative immunological mecha-
nism for COPA mutations in disease by showing that mutant COPA
via ER stress leads to a cytokine milieu that promotes the generation
of TH17 cells, a cell population shown to mediate autoimmunity5,6.
It may seem unusual that mutation of a ubiquitously expressed
protein in an essential trafficking pathway leads to an autoimmune
syndrome restricted to the lung and joints. However, there are several
specific cells central to the disease that are sensitive to alterations in
COPA function. In addition, given the reduced penetrance of disease,
there may be differences attributable to sex or other genetic and envi-
ronmental factors that dictate whether a carrier of a mutant COPA
allele develops clinical disease.
Future studies may elucidate how defects in COPA lead to ER
stress and autoimmunity. ER stress and the UPR have been strongly
associated with the pathogenesis of inflammatory and autoimmune
disorders, in part through the intersection of key pathways that lead
to activation of nuclear factor (NF)-κB or alterations in antigen pres-
entation10,18. Furthermore, because the COPA mutations identified
are predicted to impair retrograde transport, proteins destined for
return to the ER may instead escape the cell or be presented in tissue
sites to prime autoreactive immune responses35. Interestingly, both
ER stress and TH17 cells have previously been implicated in patients
with interstitial lung disease19,36,37 and inflammatory arthritis10,38,39,
including rheumatoid arthritis5. It remains unknown whether TH17
cells in COPA-mutant patients are pathogenic or arise as a result of
inflammation. The role of regulatory T cells also needs to be deter-
mined. In conclusion, the discovery of a molecular link between a
defect in the COPI vesicular trafficking machinery and a new syn-
drome of autoimmunity defines a starting point for understanding

658 VOLUME 47 | NUMBER 6 | JUNE 2015  Nature Genetics

 letters

a 100 250 40 10 10
IL23A *** IL12B IL1B 105 IL6 IL12A IL4
50
200 *** *** 10 4
*** 8 8
30
cytokine/GAPDH

8
Fold increase

150 *** 10 3 6 6
6 20
100 102 4 4
4
*** 10
2 50 101 2 2

*
0 0 0 100 0 0
TG – + – + – + – + –+ –+ –+ –+ –+ –+ –+ –+ –+ –+ –+ –+ –+ –+ –+ –+ –+ –+ –+ –+
CO1 CO4 PA3 PA2 CO1 CO4 PA3 PA2 CO1 CO4 PA3 PA2 CO1 CO4 PA3 PA2 CO1 CO4 PA3 PA2 CO1 CO4 PA3 PA2

Figure 4  Patients with b TH17 TH1 TH2 TH17 TH1 c TH2

mutant COPA demonstrate

IL-17A+ proportion of CD4+ T cells (%)

IL-17A

IFN-γ

IL-13

IFN-γ proportion of CD4 T cells (%)

0 16 0 17 0 0.4

IL-13 proportion of CD4 T cells (%)

an increase in TH17-priming 50 20 5
** *
Control

cytokine levels in BLCLs and 40 4

an increase in TH17 cell 15

+
numbers. (a) Cytokine 0 84 0 83 0 99.6 30 3
expression was measured 10
0 37 0 3 0 0.7 20 2
in BLCLs from CO1, CO4
(C.III.10), PA2 (C.V.1) and 5
Patient

10 1
PA3 (C.IV.2) either untreated

+
(white bars; −) or treated 0 0 0
(gray and black bars; +) with 0 63 0 97 0 99.3

ls

ls

ls

s
© 2015 Nature America, Inc. All rights reserved.

nt

nt

nt
tro

tro

tro
tie

tie

tie
thapsigargin. Samples were CD4

on

on

on
Pa

Pa

Pa
C

C
analyzed via qPCR for mRNA
levels of IL-23 p19 (IL23A), IL-12 p40 (IL12B), IL-12 p35 (IL12A), IL-4 (IL4), IL-1β (IL1B) and IL-6 (IL6). Shown are mRNA levels normalized against
the average value for healthy control 1 (CO1) in the absence of stimulation. Error bars indicate s.e.m. for three technical replicates in the assay. Results
are representative of three independent experiments. *P < 0.05, ***P < 0.001. (b,c) CD4+ T cells from three COPA-mutant patients with stable disease
(C.IV.1 and C.IV.2 were on mycophenolate; C.IV.15 was on no treatment) and three age-matched controls were analyzed for intracellular levels of the
IL-17A, IFN-γ and IL-13 cytokines after stimulation with PMA (phorbol 12-myristate 13-acetate) and ionomycin. (b) Representative FACS plots for CD4+
T cells producing IL-17A, IFN-γ and IL-13 are shown. Plots are gated on live CD3+CD4+ T cells. (c) Graphs demonstrating the frequency of TH1, TH2 and
TH17 cells in patients and controls. More than 350,000 cells in each of 3 patients and 3 controls were analyzed in a single experiment. The frequency of
IL-17A–secreting TH17 cells is significantly increased in patients in comparison to controls (patient mean 39% versus control mean 19%; **P < 0.01;
n = 3), whereas patients show a significant decrease in IFN-γ–producing TH1 cells (patient mean 4% versus control mean 13%; *P < 0.05; n = 3).
No significant difference was found in the frequency of IL-13–secreting T H2 cells. Student’s unpaired t tests were used for statistical evaluation.

the role of intracellular transport in the induction of human autoim- (National Human Genome Research Institute) grant U54HG003273 (R.A.G.),
munity and as a potential therapeutic target in disease. US NIH grant K23NS078056 (W.W.) and US NIH grant AI053831 (L.B.W.).

AUTHOR CONTRIBUTIONS
URLs. BCM-HGSC protocol, https://hgsc.bcm.edu/sites/default/
L.B.W., B.J. and W.W. designed and performed experiments, analyzed data and
files/documents/Illumina_Barcoded_Paired-End_Capture_Library_ wrote the manuscript. T.J.V., L.R.F., D.L.C., S.K.N., S.D.D. and M.R.W. provided
Preparation.pdf; Burrows-Wheeler aligner (BWA), http://bio-bwa.
npg

sourceforge.net/; Genome Analysis Toolkit (GATK), http://www.
broadinstitute.org/gatk/; GRCh37 human reference genome, http://
www.ncbi.nlm.nih.gov/projects/genome/assembly/grc/human/;
HGSC Mercury analysis pipeline, https://www.hgsc.bcm.edu/
software/mercury.
clinical care and analyzed clinical data. J.H. and K.D.J. performed pathology
services. A.S.-P., K.L., T.G., R.L.P.S.-C., C.-G.J., N.Z. and L.F.T. analyzed data.
Y.S., E.M.M., D.L., K.N., M.T., M.J. and C.S.L. performed experiments and analyzed
data. G.M., N.T.E., F.R.P. and M.D.R. designed experiments and analyzed data.
S.N.J., S.P., D.M.M., E.B., R.A.G., M.S.A. and P.-Y.K. organized studies. S.M.L.,
R.L.P.S.-C. and M.H.C. analyzed data and assisted with writing the manuscript.
J.S.O., J.R.L. and A.K.S. supervised the study and wrote the manuscript.

Methods COMPETING FINANCIAL INTERESTS

Methods and any associated references are available in the online The authors declare competing financial interests: details are available in the online
version of the paper.
version of the paper.
Note: Any Supplementary Information and Source Data files are available in the Reprints and permissions information is available online at http://www.nature.com/
online version of the paper. reprints/index.html.

Acknowledgments
We thank the patients and their families for their involvement in these studies.
The group at the University of California San Francisco thanks R. Schekman
for useful comments, S. McClymont, O. Estrada, R. Wadia and M. Newasker for
assistance with subject recruitment and the University of California San Francisco
Genomics Core. Funding was provided by US National Institutes of Health (NIH)
grant R01AI067946 (J.S.O.), the Jeffrey Modell Foundation (J.S.O.), US NIH grant
K08HL095659 (A.K.S.), the Foundation of the American Thoracic Society (A.K.S.),
the Pulmonary Fibrosis Foundation (A.K.S.), the Nina Ireland Lung Disease
Program (A.K.S.), US NIH (National Human Genome Research Institute/National
Heart, Lung, and Blood Institute) grant U54HG006542 and (National Institute
of Neurological Disorders and Stroke) grant 2RO1NS058529 (J.R.L.), US NIH
1. Zenewicz, L.A., Abraham, C., Flavell, R.A. & Cho, J.H. Unraveling the genetics of
autoimmunity. Cell 140, 791–797 (2010).
2. Letourneur, F. et al. Coatomer is essential for retrieval of dilysine-tagged proteins
to the endoplasmic reticulum. Cell 79, 1199–1207 (1994).
3. Eugster, A., Frigerio, G., Dale, M. & Duden, R. COP I domains required for coatomer
integrity, and novel interactions with ARF and ARF-GAP. EMBO J. 19, 3905–3917
(2000).
4. Popoff, V., Adolf, F., Brügger, B. & Wieland, F. COPI budding within the Golgi stack.
Cold Spring Harb. Perspect. Biol. 3, a005231 (2011).
5. Leipe, J. et al. Role of Th17 cells in human autoimmune arthritis. Arthritis Rheum.
62, 2876–2885 (2010).
6. Miossec, P., Korn, T. & Kuchroo, V.K. Interleukin-17 and type 17 helper T cells.
N. Engl. J. Med. 361, 888–898 (2009).

Nature Genetics  VOLUME 47 | NUMBER 6 | JUNE 2015 659

 letters
7. Cheng, M.H. & Anderson, M.S. Monogenic autoimmunity. Annu. Rev. Immunol. 30,
393–427 (2012).
8. Rieux-Laucat, F. & Casanova, J.L. Autoimmunity by haploinsufficiency. Science 345,
1560–1561 (2014).
9. Garg, S. et al. Lysosomal trafficking, antigen presentation, and microbial
killing are controlled by the Arf-like GTPase Arl8b. Immunity 35, 182–193
(2011).
10. Todd, D.J., Lee, A.-H. & Glimcher, L.H. The endoplasmic reticulum stress response
in immunity and autoimmunity. Nat. Rev. Immunol. 8, 663–674 (2008).
11. Strange, C. & Highland, K.B. Interstitial lung disease in the patient who has
connective tissue disease. Clin. Chest Med. 25, 549–559 (2004).
12. Bamshad, M.J. et al. The Centers for Mendelian Genomics: a new large-scale
initiative to identify the genes underlying rare Mendelian conditions. Am. J. Med.
Genet. A. 158A, 1523–1525 (2012).
13. Fishelson, M. & Geiger, D. Exact genetic linkage computations for general pedigrees.
Bioinformatics 18 (suppl. 1), S189–S198 (2002).
14. Brandizzi, F. & Barlowe, C. Organization of the ER-Golgi interface for membrane
traffic control. Nat. Rev. Mol. Cell Biol. 14, 382–392 (2013).
15. Schröder-Köhne, S., Letourneur, F. & Riezman, H. α-COP can discriminate between
distinct, functional di-lysine signals in vitro and regulates access into retrograde
transport. J. Cell Sci. 111, 3459–3470 (1998).
16. Goldberg, J. Decoding of sorting signals by coatomer through a GTPase switch in
the COPI coat complex. Cell 100, 671–679 (2000).
17. Claerhout, S. et al. Abortive autophagy induces endoplasmic reticulum stress and
cell death in cancer cells. PLoS ONE 7, e39400 (2012).
18. Hasnain, S.Z., Lourie, R., Das, I., Chen, A.C.-H. & McGuckin, M.A. The interplay
between endoplasmic reticulum stress and inflammation. Immunol. Cell Biol. 90,
260–270 (2012).
19. Tanjore, H., Blackwell, T.S. & Lawson, W.E. Emerging evidence for endoplasmic
© 2015 Nature America, Inc. All rights reserved.
reticulum stress in the pathogenesis of idiopathic pulmonary fibrosis. Am. J. Physiol.
Lung Cell. Mol. Physiol. 302, L721–L729 (2012).
20. Samali, A., Fitzgerald, U., Deegan, S. & Gupta, S. Methods for monitoring
endoplasmic reticulum stress and the unfolded protein response. Int. J. Cell Biol.
2010, 830307 (2010).
21. Ogata, M. et al. Autophagy is activated for cell survival after endoplasmic reticulum
stress. Mol. Cell. Biol. 26, 9220–9231 (2006).
22. Deretic, V., Saitoh, T. & Akira, S. Autophagy in infection, inflammation and
immunity. Nat. Rev. Immunol. 13, 722–737 (2013).
23. Bhattacharya, A. & Eissa, N.T. Autophagy and autoimmunity crosstalks.
Front. Immunol. 4, 88 (2013).
npg
660 VOLUME 47 | NUMBER 6 | JUNE 2015  Nature Genetics
24. Razi, M., Chan, E.Y.W. & Tooze, S.A. Early endosomes and endosomal coatomer
are required for autophagy. J. Cell Biol. 185, 305–321 (2009).
25. Adolph, T.E. et al. Paneth cells as a site of origin for intestinal inflammation.
Nature 503, 272–276 (2013).
26. Palmer, M.T. & Weaver, C.T. Autoimmunity: increasing suspects in the CD4+
T cell lineup. Nat. Immunol. 11, 36–40 (2010).
27. Wheeler, M.C. et al. KDEL-retained antigen in B lymphocytes induces a
proinflammatory response: a possible role for endoplasmic reticulum stress in
adaptive T cell immunity. J. Immunol. 181, 256–264 (2008).
28. Glatigny, S. et al. Proinflammatory Th17 cells are expanded and induced by
dendritic cells in spondylarthritis-prone HLA-B27–transgenic rats. Arthritis Rheum.
64, 110–120 (2012).
29. DeLay, M.L. et al. HLA-B27 misfolding and the unfolded protein response augment
interleukin-23 production and are associated with Th17 activation in transgenic
rats. Arthritis Rheum. 60, 2633–2643 (2009).
30. Goodall, J.C. et al. Endoplasmic reticulum stress-induced transcription factor,
CHOP, is crucial for dendritic cell IL-23 expression. Proc. Natl. Acad. Sci. USA
107, 17698–17703 (2010).
31. Gaffen, S.L., Jain, R., Garg, A.V. & Cua, D.J. The IL-23–IL-17 immune axis: from
mechanisms to therapeutic testing. Nat. Rev. Immunol. 14, 585–600 (2014).
32. Bolze, A. et al. Ribosomal protein SA haploinsufficiency in humans with isolated
congenital asplenia. Science 340, 976–978 (2013).
33. Gissen, P. & Maher, E.R. Cargos and genes: insights into vesicular transport from
inherited human disease. J. Med. Genet. 44, 545–555 (2007).
34. Russo, R., Esposito, M.R. & Iolascon, A. Inherited hematological disorders due to
defects in coat protein (COP)II complex. Am. J. Hematol. 88, 135–140 (2013).
35. Weber, C.K. et al. Antibodies to the endoplasmic reticulum–resident chaperones
calnexin, BiP and Grp94 in patients with rheumatoid arthritis and systemic lupus
erythematosus. Rheumatology 49, 2255–2263 (2010).
36. Olewicz-Gawlik, A. et al. Interleukin-17 and interleukin-23: importance in the
pathogenesis of lung impairment in patients with systemic sclerosis. Int. J. Rheum.
Dis. 17, 664–670 (2014).
37. Weaver, C.T., Elson, C.O., Fouser, L.A. & Kolls, J.K. The Th17 pathway and
inflammatory diseases of the intestines, lungs, and skin. Annu. Rev. Pathol. 8,
477–512 (2013).
38. Nistala, K. et al. Interleukin-17–producing T cells are enriched in the joints of
children with arthritis, but have a reciprocal relationship to regulatory T cell
numbers. Arthritis Rheum. 58, 875–887 (2008).
39. Benham, H. et al. Th17 and Th22 cells in psoriatic arthritis and psoriasis. Arthritis
Res. Ther. 15, R136 (2013).
 ONLINE METHODS
Study subjects. Subjects were selected on the basis of their unusual mende-
lian phenotype and, with their written consent, were studied via protocols
approved by the Research Ethics Board of The Hospital of Sick Children and
the institutional review boards for the protection of human subjects of Baylor
College of Medicine or the University of California San Francisco.
Genomic DNA preparation. DNA was isolated from clotted whole blood
by using the Clotspin Baskets and the Gentra PureGene Blood kit (Qiagen)
according to the manufacturer’s instructions.
Exome sequencing. Library preparation. DNA samples were constructed into
Illumina paired-end pre-capture libraries according to the manufacturer’s
protocol (Illumina Multiplexing_SamplePrep_Guide_1005361_D) with
modifications as described in the BCM-HGSC protocol (see URLs). Libraries
were prepared using Beckman robotic workstations (Biomek NXp and FXp
models). Briefly, 1 µg of DNA was sheared into fragments of approximately
300-400 bp with the Covaris E210 system and was end repaired, A tailed
and ligated into Illumina multiplexing paired-end adaptors. Pre-capture
ligation-mediated PCR (LM-PCR) was performed for 6–8 cycles of amplifi-
cation using 2× SOLiD Library High-Fidelity Amplification Mix (a custom
product manufactured by Invitrogen). Purification was performed with
Agencourt AMPure XP beads after enzymatic reactions, and, after the final
© 2015 Nature America, Inc. All rights reserved.
purification, the amount and size distribution of the pre-capture LM-PCR
product was determined using the LabChip GX electrophoresis system
(PerkinElmer).
Capture enrichment. The pre-capture libraries were pooled as a 4-plex
(approximately 500 ng/sample, 1 µg per pool) and hybridized in solution
to the HGSC CORE design40 (52 Mb; NimbleGen) according to the manu-
facturer’s protocol (NimbleGen SeqCap EZ Exome Library SR User’s Guide)
with minor revisions. Human Cot-1 DNA and full-length Illumina adaptor-
specific blocking oligonucleotides were added into the hybridization to block
repetitive genomic sequences and the adaptor sequences, and post-capture
LM-PCR amplification was performed using 2× SOLiD Library High-Fidelity
Amplification Mix with 14 cycles of amplification. After the final SPRI bead
purification step, the quantity and size of the capture library was analyzed
using an Agilent BioAnalyzer 2100 DNA Chip 7500. The efficiency of the
capture was evaluated by performing a qPCR-based quality assay on the four
standard NimbleGen internal controls. Successful enrichment of the capture
libraries was estimated to range from a ∆Ct value of 6 to 9 over the non-
enriched samples.
Sequencing. Library templates were prepared for sequencing using Illumina’s
npg
cBot cluster-generation system with TruSeq PE Cluster Generation kits. Briefly,
these libraries were denatured and diluted in hybridization buffer to achieve
a load density of ~800,000 clusters/mm40. The library pool was loaded on a
single lane of a HiSeq flow cell, with 2% phiX control library spiked into the
lane for run quality control. The sample libraries then underwent bridge ampli-
fication to form clonal clusters, followed by hybridization with the sequencing
primer. The sequencing run was performed in paired-end mode using the
Illumina HiSeq 2000 platform. Using TruSeq SBS kits, sequencing-by-synthesis
reactions were extended for 101 cycles from each end, with an additional
7 cycles for the index read. With the sequencing run yielding an average of
~8.4 Gb per sample, samples achieved an average of 89% of the targeted exome
bases covered to a depth of 20× or greater.
Primary data analysis for families A and B. Initial sequence analysis was
performed using the HGSC Mercury analysis pipeline41 (see URLs). In sum-
mary, the primary analysis software on the instrument produces .bcl files that
are transferred off the instrument into the HGSC analysis infrastructure by
the HiSeq Real-Time Analysis module. Once the run is complete and all .bcl
files are transferred, Mercury runs the vendor’s primary analysis software
(CASAVA), which demultiplexes pooled samples and generates sequence
reads and base call confidence values (qualities). The next step is the map-
ping of reads to the GRCh37 human reference genome (see URLs) using
the Burrows-Wheeler aligner (BWA41; see URLs) and producing a BAM42
(binary alignment/map) file. The third step involves quality recalibration using
the Genome Analysis Toolkit (GATK; see URLs) and, where necessary, the
doi:10.1038/ng.3279
merging of separate sequence-event BAMs into a single sample-level BAM.
BAM sorting, duplicate read marking and realignment to improve indel dis-
covery all occur at this step.
Confirmation of sequencing results. The candidate disease-associated muta-
tions identified by whole-exome sequencing were independently confirmed
by Sanger sequencing and analyzed for familial segregation.
Primary data analysis of families C and D. Nine samples in family C (C.III.2,
C.III.10, C.IV.1, C.IV.2, C.IV.15, C.IV.16, C.V.1, C.V.8 and C.V.9) and 11 sam-
ples in family D (D.I.2, D.II.1, D.II.3–D.II.5 and D.III.1–D.III.6) were sent to
the University of California San Francisco Genomics Core for whole-exome
sequencing. Library preparation, capture enrichment and sequencing were
performed similar to the methods used for families A and B. Sequencing
reads were aligned to the hg19 reference genome using BWA (v0.5.9) with
the default parameter that allows for two mismatches. Indexing, realign-
ment and duplicate removal were then performed using Picard (v1.56) and
SAMtools (0.1.18). Variants were called and recalibrated using GATK42,43,
version 1.4-25-g23e7f1b, and hard filtered with the recommendations listed in
Best-Practice Variant Detection in GATK v3 with ‘QD < 2.0’, ‘MQ < 40.0’, ‘FS >
60.0’, ‘HaplotypeScore 13.0’, ‘MQRankSum < 12.5’, 1 ‘ReadPosRankSum < -8.0’
for SNPs. The filtered list was then annotated with snpEff v2.0.5 in GRCh37.64
with information from public databases to determine the significance of new
sequence alterations, including the National Heart, Lung, and Blood Institute
(NHLBI) Exome Sequencing Project, 1000 Genomes Project database and
dbSNP132. A filtering analysis of the exome variants in the nine subjects in
family C was performed in VarSifter44. The filtering strategy was designed
to find variants shared among 6 carriers that were not seen in 3 related
controls or in 20 other whole-exome sequences generated for unrelated
control subjects. We also excluded variants with a minor allele frequency
(MAF) >0.001. This filtering analysis yielded a single rare variant in COPA
(c.721G>A) that was shared by affected subjects and not present in any control.
The filtering strategy in family D was performed similarly to that employed
for family C to identify a variant in COPA (c.690G>T) that cosegregated with
the disease phenotype.
Plasmids and stably transfected cell lines. pCMV6-Entry vectors encoding
Myc-DDK–tagged or untagged human COPA (Origene) were used as the
parental vectors for mutant plasmid generation. Mutant plasmids were created
via site-directed mutagenesis using the QuikChange Lightning kit (Agilent
Technologies) according to the manufacturer’s instructions or a third-party
service (Epoch Life Sciences). CD8A sequence spanning residues 1 through 207
was subcloned from human cDNA into the pCMV6-Entry vector. A sequence
encoding mTagGFP at the N terminus was added to the CD8α sequence,
and a sequence encoding residues KYKSRRSFIDEKKTN including the
dilysine motif KKTN at the C terminus was added. Cell lines stably expressing
wild-type or Glu241Lys mutant COPA-Flag fusion protein in the presence
of doxycycline were generated with the Flp-In T-REx system (Invitrogen).
Full-length sequences were amplified from patient cDNA and subcloned into
pcDNA5/FRT/TO with a sequence encoding a C-terminal Flag tag. Complete
vectors were then transfected into Flp-In T-REx 293 cells (Invitrogen) and
selected according to the manufacturer’s protocol. T-REx 293 cells are certified
by the manufacturer to be free of mycoplasma.
Cell transfection. For gene expression studies, HEK293 cells were transiently
transfected with 5 µg of plasmid DNA using TransIT-293 Transfection Reagent
(Mirus) according to the manufacturer’s instructions. For COPA knockdown
experiments, HEK293 cells were transfected with either COPA-targeting
siRNA or control scrambled siRNA (Thermo Scientific Dharmacon) using
Lipofectamine RNAiMAX transfection reagent (Life Technologies) following
the manufacturer’s instructions. After 48 h, cells were collected and processed
for RNA isolation. HEK293FT cells (Invitrogen) are certified by the manufac-
turer to be free of mycoplasma.
For microscopic analysis, HEK293T cells stably expressing GFP-LC3 were
electroporated using the Amaxa 4D nucleofector system according to the
manufacturer’s protocol. Briefly, HEK293T cells were grown to 80% conflu-
ence. Then, 1 × 106 cells were nucleofected with 2 µg of plasmid DNA encoding
the COPA mutants. After 24 h, cells were collected and imaged.
Nature Genetics
 Generation and stimulation of B-lymphoblastoid cell lines. Exponentially
growing B95-8 cells were seeded at 1 × 106 cells/ml and incubated for 3 d at
37 °C in 5% CO2. B95-8 supernatant was collected by centrifugation at 300g
at 4 °C and passed through a 0.45-µm filter. PBMCs were resuspended at
1-5 × 106 cells/ml in RPMI-1640 medium supplemented with 20% FBS
followed by the addition of 0.5–1 volumes of B95-8 supernatant with or
without 10 µg of a human TLR9 ligand (ODN 2006, Class B CpG oligonucle-
otide, Invivogen). Cells were incubated overnight at 37 °C in the presence of
5% CO2 and 0.5 µg/ml cyclosporine A, and an equal volume of medium was
added the next day. Cells were cultured until they formed macroscopically
visible clusters and then maintained at concentrations of 0.5–1 × 106 cells/ml. For
ER stress induction, Epstein-Barr virus (EBV)-transformed cells were seeded
at 1 × 106 cells/ml and incubated with 100 nM thapsigargin (Sigma) for 6 h.
Cells were collected, washed in PBS and processed for RNA isolation.
RNA isolation and quantitative PCR analysis. RNA was isolated with the
RNeasy Mini kit (Qiagen). SuperScript III Reverse Transcriptase (Invitrogen)
was used to synthesize cDNA. qPCR was performed with TaqMan Gene
Expression assays from Life Technologies (COPA, Hs00189232_m1; HSPA5
(BiP), Hs00607129_gH; GAPDH, Hs02758991_g1; IL23p19, Hs00900828_g1;
IL12p35, Hs01073447_m1; IL12p40, Hs01011519_m1; IL6, Hs00985639_m1;
IL1B, Hs01555410_m1; IL4, Hs00174122_m1) or by using Qiagen SYBR green
for ACTB, DDIT3 or ATF4 (sequences in Supplementary Table 5).
© 2015 Nature America, Inc. All rights reserved.
Routine and immunohistochemical staining of lung biopsy tissues. Lung
biopsies from the patients were fixed in 10% buffered formalin and routinely
processed for paraffin embedding. The resulting formalin-fixed, paraffin-
embedded lung biopsy tissues were sectioned, resulting in glass slides with
5-µm tissue sections that underwent routine hematoxylin and eosin staining
and immunohistochemical staining. Tissue section staining was performed
by the University of California San Francisco and/or Texas Children’s Hospital
Pathology Laboratories certified by the College of American Pathologists,
employing procedures certified by the College of American Pathologists, with
appropriate positive and negative controls and using automated hematoxy-
lin and eosin (Leica Microsystems) and immunohistochemical (Leica Bond
III Automated Immunohistochemical and In Situ Hybridization Biosystem,
Leica Microsystems) staining systems. Epitope antigen retrieval was performed
using a proprietary automated system employing citrate-based buffer and sur-
factant proprietary kits (Novocastra Bond Epitope Retrieval Solutions 1, Leica
Microsystems). Proprietary antibody kits (Leica Microsystems) for CD20, CD4
and CD8 were used with appropriate positive and negative control tissues.
Immune infiltrates of organs were confirmed by an independent reading of
npg
the slides by a blinded observer.
For BiP staining, microwave antigen retrieval was performed on depar-
affinized and rehydrated lung sections with sodium citrate (pH 5.5). After
peroxidase blocking (3% H2O2) and blocking with 1% BSA and 3% goat serum,
sections were stained with antibody to BiP (Abcam, ab21685) overnight. BiP
staining was visualized with biotinylated goat anti-rabbit antibody (Jackson
ImmunoResearch Laboratories) and Elite ABC reagent followed by DAB stain-
ing (Vector Laboratories). Slides were counterstained with Mayer’s hematoxylin.
Images were obtained using a standard histology microscope and captured
with an AxioCam using AxioVision software (Carl Zeiss MicroImaging).
Binding assay and immunoblot. The Flp-In T-REx 293 cells described above
were cultured for 2 d in the presence of doxycycline to induce wild-type or
mutant COPA-Flag overexpression. Cells were then lysed in 500 µl of lysis
buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% NP-40
and 5% glycerol). A synthetic peptide (Genscript) representing the cytoplas-
mic tail of yeast Wbp1 (Wbp1p peptide; CKKLETFKKTN) was covalently
linked to SulfoLink resin (Pierce) according to the manufacturer’s protocol.
Cell lysate containing 80 µg of protein was added to 20 µl of resin in 0.5 ml
of binding buffer (200 mM NaCl, 20 mM Tris-HCl, pH 7.4, 7.5 mM MgCl2,
5% glycerol and 0.2% Triton X-100) and gently rotated at room tempera-
ture for 5 min. Resin was subsequently washed by inverting eight times in 1
ml of binding buffer and lastly boiled in SDS-PAGE sample buffer to release
bound COPA. For immunoprecipitation, Flp-In T-REx 293 cells cultured
in the presence of doxycycline were transfected with the GFP-CD8-KKTN
Nature Genetics
construct for 36 h. Cells lysates were precleared with Sepharose beads and
then incubated with antibody to CD8 (Abcam, ab85792) covalently linked to
Sepharose beads. Protein-bead complexes were washed and boiled in SDS-
PAGE sample buffer. Proteins were subjected to SDS-PAGE and transferred
to PVDF, and membranes were blocked and probed with one of the following
antibodies: a horseradish peroxidase (HRP)-conjugated monoclonal anti-
body against Flag (Sigma, A8592), a mouse monoclonal antibody to mGFP
(Origene, TA180076), a rabbit antibody to COPA (Sigma, HPA028024) or
a mouse antibody to GAPDH (Santa Cruz Biotechnology, sc-32233). Bands
were detected by incubation with secondary HRP-coupled antibodies and
SuperSignal West Femto (Pierce) chemiluminescence.
Flow cytometry. PBMCs were suspended in FACS buffer (PBS with 2% FCS
and 1% sodium azide) at 1 × 106 cells per well of 96-well culture plates and
incubated for 4 h in the presence of brefeldin A (Sigma) with or without PMA
(70 ng/ml) and ionomycin (700 ng/ml). For evaluation of surface receptors, cells
were stained with the viability Ghost Dye violet 510 (Tonbo) and monoclonal
antibodies, including anti-CD3 clone SKY7 (BioLegend) and anti-CD4 clone
OKT4 (BioLegend). Cells were incubated with antibodies at room temperature
for 20 min and then washed with FACS buffer. For intracellular staining, cells
were permeabilized with Cytofix/Cytoperm reagent (BD) and incubated for
30 min at room temperature after incubation with anti–IFN-γ clone 4S.B3
(eBioscience), anti–IL-17A clone eBio64CAP17 (eBioscience) or anti–IL-13
clone PVM13-1 (eBioscience). Samples were acquired using a Fortessa flow
cytometer (BD Bioscience) and analyzed using FlowJo software.
Statistical analysis. Statistical analyses were performed using Prism 6.0
(GraphPad Software). Statistical comparison was made using either
Mann-Whitney U tests or the Student’s t tests as indicated in the figure legends.
Welch’s t test was applied when variances were unequal. Comparison between
more than two experimental groups was made using one-way ANOVA,
followed by Dunnett’s post-hoc test. P < 0.05 was considered statistically
significant.
Genotype analysis. We performed OMNI Express high-resolution cSNP
arrays (~720,000) for families A (IV.1, III.1 and III.2) and E (II.1, I.1
and I.2). We used the data to obtain the length of the haplotype block
around the causative COPA mutations shared by affected individuals in two
families. For each family, we phased and compared the genotype calls for
each trio to identify the haplotypes inherited by affected family members.
The haplotype analysis of cSNP data was performed using custom scripts
implemented in the R statistical programming language.
Autophagy stimulation and detection. Functional and biochemical analy-
ses were performed using the following reagents: Torin1 (EMD Millipore
Chemicals); baflomycin A1 (Sigma-Aldrich); guinea pig antibody to the
C terminus of p62 (American Research Product, 03-GP62-C); rabbit antibody
to LC3B (Novus Biologicals, NB100-2220); rabbit antibody to S6 ribosomal
protein phosphorylated at Ser235/Ser236 (Cell Signaling Technology, 4858);
mouse monoclonal antibody (54D2) to S6 ribosomal protein (Cell Signaling
Technology, 2317); mouse monoclonal antibody to GAPDH (Sigma-Aldrich,
G8797); and infrared fluorescent dye–conjugated secondary antibodies
(Li-Cor). To perform the experiments, an mTOR-specific inhibitor Torin1 was
used to induce autophagy, and a lysosome inhibitor bafilomycin A were used to
block the degradation of autophagy content by lysosome. Briefly, cells (2 × 106)
were treated with vehicle only, 250 nM Torin1 or 250 nM Torin1 together
with 200 nM bafilomycin A for 4 h. Cells were washed twice with PBS and
lysed on ice for 30 min in GFB buffer (40 mM Bis-Tris propane, 150 mM
NaCl, 1% Triton X-100, 10% glycerol) supplemented with 1× PIC (protease
inhibitor cocktail, BD Biosciences) and 1× Phosphatase Inhibitor Cocktail
(Thermo Scientific). After centrifugation at 5,000g for 1 min, protein concen-
trations in cell lysate supernatants were measured using BCA protein assay
reagents (Thermo Scientific). Cell lysates (25 µg of protein) were then heated at
95 °C for 5 min in 1× LDS sample buffer (Invitrogen) and resolved by SDS-
PAGE (Invitrogen). Proteins were transferred to nitrocellulose membranes
using a wet transfer cell system (Invitrogen). Membranes were blocked with
blocking buffer (Li-Cor) for 1 h, followed by a 2-h incubation with primary
doi:10.1038/ng.3279
 antibody and a 1-h incubation with secondary antibody in blocking buffer
supplemented with 0.1% Tween-20. Immunoreactive bands were visualized
using the Odyssey Infrared Imaging System (Li-Cor). For direct assessments
of autophagosomes, cells were allowed to adhere to coverslips coated with
poly(l)-lysine for 20 min at 37 °C in 5% CO2 and were then washed in PBS and
permeabilized with Cytofix/Cytoperm reagent. Cells were stained with rabbit
IgG to LC3B (D11, Cell Signaling Technology) and mouse IgG to CD107a
(H4A3, BD Biosciences). Cells were washed and then stained with anti-mouse
antibody conjugated to Alexa Fluor 488 and anti-rabbit antibody conjugated
to Alexa Fluor 532. GFP-LC3–expressing HEK293T cells were electroporated
as stated above. Cells were imaged as stated below. The formation of GFP
puncta was quantified to indicate autophagosome formation upon expression
of mutant or wild-type COPA protein.
Microscopy. For stimulation emission depletion (STED) microscopy, cells
were allowed to adhere to coverslips coated with poly(l)-lysine for 20 min at
37 °C in 5% CO2, and cells were then washed in PBS and permeabilized with
Cytofix/Cytoperm reagent. Cells were stained with rabbit IgG to LC3B (D11,
Cell Signaling Technology) and mouse IgG to CD107a (H4A3, BD Biosciences)
for 1 h. Cells were washed and stained with Alexa Fluor 488–conjugated goat
anti-mouse IgG and Alexa Fluor 532–conjugated goat anti-rabbit antibody
(Life Technologies) for 30 min. Slides were visualized on a Leica SP8 TCS
STED microscope (Leica Microsystems). Acquisition was performed as previ-
© 2015 Nature America, Inc. All rights reserved.
ously described24. Briefly, excitation was performed using a tunable white-light
laser, and emission was detected by HyD detectors. Settings were adjusted on
the basis of single-color controls, and independent channels were acquired
sequentially to limit fluorescence contamination. Images were acquired using
LAS AF (Leica) and exported to Volocity software (PerkinElmer) for analysis.
For confocal microscopy, GFP-LC3–expressing HEK293T cells were electro-
porated as stated above. Electroporated cells were plated on a delta-T imaging
npg
doi:10.1038/ng.3279
chamber (Bioptechs), and 24 h later cells were imaged by confocal microscopy
on a Zeiss Axio Observer Z1 with Yokogawa CSU10 spinning disk. Images
were acquired with a Hamamatsu Orca-AG camera using Volocity Software
(PerkinElmer). Excitation was performed using a 488-nm laser with optimized
exposure times. Images were analyzed using Volocity software.
For COPA staining and co-staining of the ERGIC compartment in BLCLs,
cytospins were performed. After fixing with 0.2% Triton X-100, 2% formalde-
hyde in PBS and blocking, cells were stained with antibodies to COPA (Sigma)
and ERGIC-53 (Santa Cruz Biotechnology) and with DAPI. Staining was
visualized using a Zeiss Apotome wide-field microscope.
Image analysis. Microscopy images were analyzed using Volocity software.
For confocal images, the intensity of GFP was thresholded by 2 s.d. above
the mean intensity. Objects under 1 µm3 were excluded. For STED images, a
signal-to-noise ratio of 2 was used to define positive signal. Objects smaller
than 0.1 µm2 were excluded from analysis. All values were exported to
GraphPad 6.0 (Prism Software) for graphing and statistical analysis. Presented
images were contrast enhanced uniformly.
40. Reid, J.G. et al. Launching genomics into the cloud: deployment of Mercury, a next
generation sequence analysis pipeline. BMC Bioinformatics 15, 30 (2014).
41. Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows-Wheeler
transform. Bioinformatics 25, 1754–1760 (2009).
42. DePristo, M.A. et al. A framework for variation discovery and genotyping using
next-generation DNA sequencing data. Nat. Genet. 43, 491–498 (2011).
43. McKenna, A. et al. The Genome Analysis Toolkit: a MapReduce framework for
analyzing next-generation DNA sequencing data. Genome Res. 20, 1297–1303
(2010).
44. Teer, J.K., Green, E.D., Mullikin, J.C. & Biesecker, L.G. VarSifter: visualizing and
analyzing exome-scale sequence variation data on a desktop computer. Bioinformatics
28, 599–600 (2012).
Nature Genetics