A Secretion Trap Screen in Yeast Identifies Protease Inhibitor 16 as a Novel

Antihypertrophic Protein Secreted From the Heart

Robert J.A. Frost and Stefan Engelhardt

Circulation. 2007;116:1768-1775; originally published online October 1, 2007;

doi: 10.1161/CIRCULATIONAHA.107.696468
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2007 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539

The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circ.ahajournals.org/content/116/16/1768

Data Supplement (unedited) at:

http://circ.ahajournals.org/content/suppl/2007/10/01/CIRCULATIONAHA.107.696468.DC1.html

Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial
Office. Once the online version of the published article for which permission is being requested is located,
click Request Permissions in the middle column of the Web page under Services. Further information about
this process is available in the Permissions and Rights Question and Answer document.

Reprints: Information about reprints can be found online at:

http://www.lww.com/reprints

Subscriptions: Information about subscribing to Circulation is online at:

http://circ.ahajournals.org//subscriptions/

Downloaded from http://circ.ahajournals.org/ by guest on June 2, 2014

 Heart Failure

A Secretion Trap Screen in Yeast Identifies Protease

Inhibitor 16 as a Novel Antihypertrophic Protein Secreted
From the Heart
Robert J.A. Frost, MD; Stefan Engelhardt, MD, PhD

Background—Cardiomyocyte hypertrophy is of central importance in the development of congestive heart failure.

Whether proteins secreted from the myocardium itself contribute to myocardial hypertrophy is largely unknown.
Methods and Results—We performed a genetic yeast secretion trap screen using a murine cardiac cDNA library and
identified 54 cardiac proteins that contained a secretion signal. When determining their mRNA expression in the
myocardium of failing hearts, we found protease inhibitor 16 (PI16) to be strongly upregulated in hypertrophic and
failing myocardium. PI16, a 489 –amino acid protein with an unknown function, also displayed enhanced expression on
the protein level after serum stimulation of primary cardiomyocytes and in failing myocardium. We found PI16 to be
secreted rapidly by primary cardiomyocytes into the culture medium, where it inhibited cardiomyocyte growth. RNA
interference–mediated suppression of endogenous PI16 in primary cardiomyocytes significantly enhanced cardiomyo-
cyte size. Transgenic mice overexpressing PI16 in a cardiomyocyte-specific manner showed normal cardiac function but
had smaller hearts with hypotrophic cardiomyocytes.
Conclusions—Taken together, we identified 54 putatively secreted cardiac proteins. PI16, a novel protein secreted from
the heart, is strongly upregulated early in heart failure and inhibits growth of cardiomyocytes both in vitro and in vivo.
PI16 might represent a novel therapeutic target in heart failure. (Circulation. 2007;116:1768-1775.)
Key Words: growth substances 䡲 heart failure 䡲 hypertrophy 䡲 proteins 䡲 remodeling

H eart failure remains one of the most frequent causes of

death in industrialized countries despite significant
progress in pharmacological treatment.1 Remodeling of the
myocardium, characterized by hypertrophy of cardiomyo-
cytes and interstitial fibrosis, is of central importance for the
development and progressive clinical course of heart failure.
Several lines of evidence indicate that communication be-
tween cardiac cells via secreted factors may contribute to
myocardial remodeling. For example, conditioned medium
from cultured fibroblasts induces hypertrophy of cardiomyo-
cytes, and conditioned medium from cardiomyocytes pro-
motes proliferation of fibroblasts.2 Furthermore, evidence is
growing that stem cell therapy after myocardial infarction
prevents cardiac remodeling through paracrine factors.3,4
Clinical Perspective p 1775
Secreted and transmembrane proteins show promise as
therapeutic targets or possibly even as therapeutic agents.
They are accessible to various drug-delivery mechanisms, as
they are present on the cell surface or within the extracellular
space.
Paracrine factors that are known to contribute to cardio-
myocyte hypertrophy and myocardial remodeling include
angiotensin II, transforming growth factor-␤, endothelin,
catecholamines, and insulin-like growth factor-1.5,6 The num-
ber of identified proteins secreted from the heart is small,
possibly owing to their low expression level. In contrast,
bioinformatic analyses imply very high numbers of secreted
proteins, comprising up to 2000 proteins for the mouse
secretome.7,8 These approaches assume that the presence of a
secretion signal sequence per se is sufficient to mediate
effective secretion of a particular protein; however, no strictly
defined consensus sequence exists for a functionally active
signal sequence. Signal sequences are typically characterized
by hydrophobic amino acids followed by a signal peptidase
cleavage site; but not all sequences with these characteristics
function as signal sequences, for example, because the signal
sequence is not accessible after folding of the tertiary struc-
ture. In addition, identification of the genes themselves in
genomic data is still difficult, and estimates of the total
number of human genes vary widely.9 Expressed sequence
tags, on the other hand, are often truncated and miss the

Received February 16, 2007; accepted August 10, 2007.

From Rudolf Virchow Center, DFG Research Center for Experimental Biomedicine and Institute of Pharmacology and Toxicology, University of
Wuerzburg, Wuerzburg, Germany.
The online-only Data Supplement, consisting of an expanded Methods section, figures, and tables, is available with this article at
http://circ.ahajournals.org/cgi/content/full/CIRCULATIONAHA.107.696468/DC1.
Correspondence to Stefan Engelhardt, MD, PhD, Rudolf Virchow Center/DFG Research Center for Experimental Biomedicine, University of
Wuerzburg, Versbacher Strasse 9, 97078 Wuerzburg, Germany. E-mail Stefan.engelhardt@virchow.uni-wuerzburg.de
© 2007 American Heart Association, Inc.
Circulation is available at http://circ.ahajournals.org DOI: 10.1161/CIRCULATIONAHA.107.696468

1768
Downloaded from http://circ.ahajournals.org/ by guest on June 2, 2014
 Frost and Engelhardt
N-terminal signal sequence. Furthermore, tissue-specific ex-
pression data for the majority of the identified proteins are
insufficient, which makes the identification of tissue-specific
secretomes difficult. Lastly, many of the putative secreted
genes were identified by homology screening, but homologs
do not need to have the same subcellular localization. Thus,
computational screens may miss secreted proteins and may
yield a high rate of false-positive results.
In the present study, we performed a genetic yeast secre-
tion trap screen to systematically search for genes encoding
proteins that are secreted by the heart. Using a murine cardiac
cDNA library, we identified 54 cardiac proteins that con-
tained a secretion signal. Many of these proteins have not
been studied previously or even discovered in the heart, and
the function of several is completely unknown. Among the
latter is protease inhibitor 16 (PI16), a protein we found to be
secreted by cardiomyocytes and that inhibits cardiomyocyte
hypertrophy both in vitro and in vivo.
Methods
Secretion Trap Screen
The secretion trap screen was performed as described previously10,11
with some modifications. Briefly, the host yeast strain, O66-2 lacks
invertase, an enzyme that must be secreted for the strain to grow on
sucrose. A cDNA library was prepared (SuperScript Plasmid System,
Invitrogen, Carlsbad, Calif) from left heart ventricles of 4- to
5-month-old wild-type FVB mice by use of random hexamer
primers. The cDNA library was then cloned into a yeast expression
vector encoding a mutant invertase (pSuc2dMSP), which lacked the
start codon for methionine and the signal sequence necessary for
secretion. After transformation of the host strain with the library,
colonies could form on sucrose only in those cases in which a clone
from the cDNA library provided both a start codon and a functional
signal sequence in frame with the invertase. cDNA inserts of positive
clones were amplified by polymerase chain reaction (PCR) with
flanking primers, sequenced, and analyzed by a BLAST search
(Basic Local Alignment and Search Tool). To avoid sequencing of
redundant clones, we performed an iterative cross-hybridization of
the PCR products of newly identified clones with 32P-labeled probes
directed against repeatedly found cDNAs.
RNA Isolation and Real-Time Quantitative PCR
Total RNA was extracted from frozen tissue with the RNeasy Mini
Kit (Qiagen, Hilden, Germany), including DNase digestion accord-
ing to the manufacturer’s instructions. Real-time PCR was per-
formed with SYBR Green as fluorescent dye, and data were
calculated by the 2-⌬⌬CT method.12 Please refer to the online-only
Data Supplement for details.
Human Heart Tissue
Please refer to the online-only Data Supplement for a detailed
description. The present study was approved by the Ethics Commit-
tee of the Medical Faculty of the University of Wuerzburg.
Generation of a PI16-Specific Antibody and
Western Blot Analysis
Polyclonal antibodies directed against PI16 were generated and
affinity purified as detailed in the online-only Data Supplement.
Isolation of Primary Cardiac Cells
Neonatal rat cardiomyocytes were isolated as described
previously.13,14
Downloaded from http://circ.ahajournals.org/ by guest on June 2, 2014
PI16 in Heart Failure 1769
Generation of PI16 Adenovirus,
Immunofluorescent Staining, Isoleucine
Incorporation, RNA Interference Transfection,
Histological Analyses, and In Vivo
Hemodynamic Analysis
Please refer to the online-only Data Supplement for a description.
Generation of Transgenic Mice
A detailed description is presented in the online-only Data Supple-
ment. All animal experiments were approved by the responsible
authorities.
Statistical Analysis
For comparison of 2 individual groups, unpaired 2-tailed t tests were
performed with the Prism software package (GraphPad, San Diego,
Calif). For comparison of 1 factor in ⬎2 groups, ANOVA followed
by Bonferroni multiple-comparison posttest was performed. To test
for differences in means of 2 factors, 2-factor ANOVA was per-
formed. If the interaction was significant, the t test was used instead.
Data are presented as mean⫾SEM. Differences were considered
significant when P⬍0.05.
The authors had full access to and take full responsibility for the
integrity of the data. All authors have read and agree to the
manuscript as written.
Results
Identification of Proteins Secreted in the Heart by
a Secretion Trap Screen
We generated a murine cardiac cDNA library with a com-
plexity of 2⫻107 independent clones, which means an aver-
age gene transcript was represented ⬎100 times. The cDNA
library also contained transcripts from weakly expressed
genes, as tested by PCR amplification of a variety of genes
(data not shown). We then successively plated 1.7⫻107 yeast
transformants (approximately one the cDNA library) and
isolated 1900 clones that grew on selection media containing
2% sucrose. Only yeast clones carrying cDNA fragments
encoding for secreted proteins grew under selection condi-
tions (Figure 1A). Using an iterative cross-hybridization
strategy, we could reduce the number of clones that were
sequenced to 347 and thereby identified 54 nonredundant
cardiac cDNAs encoding for putatively secreted proteins
(online-only Data Supplement Table II). Currently, the func-
tion of approximately one third of these genes is unknown.
We then determined the mRNA expression of 25 of the
identified genes in a well-characterized murine heart failure
model (␤1-adrenergic receptor transgenic mice15; Figure 1B).
Compared with healthy wild-type littermates, we found
upregulation of PI16 (420⫾130%, P⬍0.001) and atrial natri-
uretic peptide (620⫾170%, P⬍0.001). We therefore aimed to
further characterize PI16.
Cloning of PI16 and Identification of
a Splice Variant
The murine PI16 gene comprises 7 exons (exon 7 encodes for
the 3⬘-UTR) spanning ⬇10 200 bp and resides on chromo-
some 17.16 We cloned murine PI16 using primers at the 5⬘
and 3⬘ ends of the putative coding sequence and confirmed
the predicted sequence by sequencing (1470 bp). In the heart,
we thereby identified a more weakly expressed splice variant
comprising 684 bp that lacks exon 5 (online-only Data
Supplement, Figure IA). The coding sequence of the human
1770 Circulation October 16, 2007

A Cardiac cDNA-fragments
B
Signal sequence

Not I Xho I
pSuc2∆MSP
ADH- Invertase mut
promoter

(fold induction failing vs. nonfailing)

Glucose 9 ***
Invertase-/- Invertase-/- 8
Wild-type + Invertase mut + cDNA-Invertase mut
7

mRNA expression
Intracellular ***
6
5
Energy Energy 4
3
Sucrose
Fructose Fructose 2
+ +
Glucose Glucose Secreted 1
proteins
Non-secreted 0
Sucrose Sucrose proteins

GCSH

Prohibitin
Neuropilin
2700059D21
9530068E07

Crisp2

PI16
CD14
CD164
Cxcl14
Psap
Art1
Nov
9130403P13
9130213B05
Manba
Lrg1
MPG1
4930534E15
1500004A08
Tde1l
ANP
0610038D11

Tsfm

TAPbp
Sucrose
Extracellular
Growth No growth Growth of rare colonies

Figure 1. Identification of proteins secreted from the heart by a secretion trap screen. A, Design of the cardiac secretion trap screen
(lower left). Wild-type yeast is able to grow on sucrose medium by secreting invertase, which metabolizes sucrose and thereby pro-
vides glucose as an energy source. An invertase-deficient strain is not able to grow on sucrose medium. We introduced a cardiac
cDNA library in front of a mutated (mut) invertase gene that lacks the N-terminal signal sequence for secretion (top) and transformed
the invertase-deficient yeast strain with these constructs. Only clones carrying a cDNA fragment encoding for a secreted protein and
whose sequences are in frame are able to secrete invertase, which enables them to grow on sucrose-containing selection medium.
Clones on the upper row containing 4 different cDNA fragments encoding for secreted proteins grew on glucose and on sucrose-
containing selection medium (right). In contrast, the 2 clones in the lower row, which carry cDNA fragments encoding for intracellular
nonsecreted proteins, grew only on glucose medium. B, Differential mRNA expression of genes identified by the yeast secretion trap
screen in a murine heart failure model. RNA was isolated from the left ventricular myocardium of 6-month-old ␤1-adrenergic receptor
transgenic mice; gene expression levels were determined by real-time PCR and compared with wild-type littermates (nⱖ4). Expression
of PI16 was nearly as strongly upregulated as that of atrial natriuretic peptide. ***P⬍0.001 for wild-type vs ␤1-adrenergic receptor trans-
genic mice by ANOVA with Bonferroni posttest. RIKEN cDNAs are abbreviated with the corresponding cDNA number.

full-length homolog is shorter (1392 bp), but due to a smaller
exon 5, the putative splice variant is slightly longer (714 bp).
Bioinformatic analysis with SMART (Simple Modular Ar-
chitecture Research Tool)17 revealed that PI16 contains a
sperm-coating glycoprotein domain (online-only Data Sup-
plement, Figure IB). There are homologs to PI16 in other
mammals, such as opossum (ENSMODP00000016841; aa
identity 55.9%) and chicken (ENSGALP00000000778; aa
identity 47.0%; online-only Data Supplement, Figure IC).
The evolutionarily most distant homolog that we could
identify was in fish (ENSDARP00000067664; aa identity
32.9%), which suggests that PI16 evolved in vertebrates.
Protein Expression of PI16
The open reading frame of full-length PI16 encodes a protein
of 489 amino acids. We generated a polyclonal antibody
directed specifically against the full-length PI16 protein
(PI16-FL antibody). To examine the organ expression pattern
of PI16, we isolated diverse organs from 3-month-old wild-
type FVB mice and determined expression of PI16 by
Western blotting. PI16 protein expression was strongest in
aorta and skin and weaker in adipose tissue (online-only Data
Supplement, Figure IIA). In healthy hearts, PI16 protein
expression was low. Surprisingly, the polyclonal PI16-FL
antibody detected 3 specific PI16 bands (74, 100, and 108
kDa) in tissue lysates under both reducing and nonreducing
conditions. We generated 2 additional antibodies directed
against exon 5 (excluding the sperm-coating glycoprotein
domain, which is localized in exons 1 to 4) and a small
Downloaded from http://circ.ahajournals.org/ by guest on June 2, 2014
peptide in exon 2 of PI16, respectively. Both our original
antibody directed against full-length PI16 and the 2 new
antibodies detected a virtually identical pattern of PI16 bands
when used for Western blotting analysis in tissue lysates
(online-only Data Supplement, Figure IIB). Bioinformatic
analysis indicated potential glycosylation sites. Therefore, we
deglycosylated myocardial lysates from PI16 transgenic mice
using PNGaseF. Enzymatic deglycosylation led to a signifi-
cant shift of the PI16 bands to lower molecular weights,
which suggests glycosylation of the PI16 protein (online-only
Data Supplement, Figure IIC). However, a different glyco-
sylation pattern cannot explain the existence of more than 1
PI16 band. The theoretical molecular weight of PI16 is 52.7
kDa; however, recombinant PI16 from Escherichia coli fused
to a GST tag (27 kDa) and, lacking its signal sequence (2
kDa), ran at ⬇95 kDa, which suggests a molecular weight of
unmodified PI16 of ⬇70 kDa (data not shown). This is in line
with the deglycosylation experiment. The higher-molecular-
weight bands of PI16 may represent a further modification or
a covalent linkage to another protein.
PI16 Is a Secreted Protein
The identification of PI16 in the secretion trap screen and the
presence of a bioinformatically predicted N-terminal signal
peptide (online-only Data Supplement, Table II) both suggest
secretion of PI16. To determine whether PI16 is secreted
from mammalian cells, we transfected neonatal rat cardio-
myocytes with a recombinant adenovirus expressing murine
PI16 (Adv-PI16). Interestingly, we detected predominantly

A B
β-actin
kDa
181
115
82 PI16
64
49
LacZ PI16 LacZ PI16
NRCM Medium
C Left ventricular myocardium
PI16 PI16, α-actinin, nuclei
Figure 2. PI16 is a secreted protein. A, Detection of PI16 by
Western blotting in lysates from neonatal rat cardiomyocytes
(NRCM) and in conditioned medium from NRCMs 48 hours after
transfection with a PI16-expressing or a lacZ-expressing control
adenovirus (Adv, MOI 0.1, upper panel). The 74- and 108-kDa
bands were predominately detected in lysates from cells trans-
fected with PI16. In contrast, the 100-kDa band predominated in
the culture medium. Detection of ␤-actin was used to monitor
potential contamination of the culture medium with cellular
material derived from lysed cells. B, Immunofluorescence of
NRCMs 48 hours after transfection with PI16-expressing adeno-
virus (MOI 0.1). Nuclei are stained with Hoechst 33258 (blue);
cardiomyocytes are stained with an antibody directed against
␣-actinin (green). Transfected cells show vesicular structures
(red) of PI16 in the cytoplasm. Nontransfected cells show only
low amounts of endogenous PI16 protein. C, Immunofluorescent
detection of PI16 in cryosections of a wild-type mouse heart.
PI16 (red) accumulates extracellularly around cardiomyocytes
(left panel, staining of PI16; middle panel, costaining of ␣-actinin
[green] and nuclei [blue]; right panel, no signal is detectable
after omission of the PI16-specific first antibody).
the 74- and 108-kDa bands in lysates from cells transfected
with PI16 (Figure 2A). In contrast, the 100-kDa band pre-
dominated in the culture medium. This indicates that the
A B
***
**
PI16 mRNA expression
6 kDa
115
(fold induction)
5 ventricular myocardium (nⱖ4 each per age
4 82 PI16
* 64
3 64
2 mice vs wild-type mice of the same age,
1 Wild-type β1AR-transgene Serum - +
0 with pressure overload–induced cardiac
2 6 12
months Human
β1AR-transgene TAC heart failure
lysates from left ventricular myocardium from wild-type and ␤1-adrenergic receptor transgenic mice. Mice were studied at 12 months of
age (n⫽3 per genotype). C, Induction of PI16 in neonatal rat cardiomyocytes after serum stimulation (5%, 24 hours). Western blot
shown is representative of 3 independent experiments.
Downloaded from http://circ.ahajournals.org/ by guest on June 2, 2014
Frost and Engelhardt PI16 in Heart Failure 1771
N-terminal signal sequence of PI16 is functional and that
PI16 is secreted from cardiomyocytes. The 2 cellular PI16
bands and the soluble extracellular PI16 together represent
the 3 bands detected in tissue lysates that are a mixture of
cells and extracellular material (online-only Data Supple-
ment, Figure IIA).
Classically secreted proteins pass the endoplasmic reticu-
lum and the Golgi apparatus and are ultimately transported to
the plasma membrane in secretory vesicles. To elucidate the
intracellular localization of PI16, we transfected neonatal rat
PI16, α-actinin, nuclei
cardiomyocytes with Adv-PI16 (multiplicity of infection
Cardiomyocytes [MOI] 0.1). Immunofluorescent staining with the affinity-
Adv-PI16 (MOI 0.1)
purified antibody directed against PI16 (PI16-FL antibody)
revealed a distribution of PI16 in small cytoplasmatic vesi-
cles, as is typical for secreted proteins (Figure 2B). Endoge-
nous PI16 protein in nontransfected cardiomyocytes was only
scarcely detected. This could be due to the low PI16 expres-
sion in nonstimulated cardiomyocytes or to a rapid and
effective secretion of PI16.
We therefore made cryosections of murine wild-type hearts
and stained PI16 by immunofluorescence (Figure 2C). Again,
sec. antibody, α-act., nuc.
only small amounts of PI16 were visible intracellularly;
however, PI16 was readily detectable in the intercellular
space, which suggests extracellular accumulation and possi-
bly membrane association of PI16 in the heart after rapid
secretion.
Expression of PI16 Is Upregulated
in Heart Failure
In ␤ 1 -adrenergic receptor transgenic mice, a well-
characterized heart failure model that uses cardiac-specific
overexpression of the ␤1-adrenergic receptor, we found pro-
found upregulation (up to 470⫾55%, P⬍0.001) of PI16
full-length mRNA (Figure 3A). PI16 upregulation in this
model starts in very young animals (2 months old) and
precedes the macroscopically visible myocardial remodeling,
which suggests a possible role in the pathogenesis of heart
failure. Messenger RNA expression of PI16 was also strongly
upregulated (200⫾70%, P⬍0.05) in mice after 5 weeks of
pressure overload of the left ventricle by thoracic aortic
banding. Furthermore, PI16 mRNA appeared to be upregu-
lated (323⫾160%, P⫽0.11) in human heart failure. In accor-
dance with these findings, we also found PI16 protein
Figure 3. Expression of PI16 is strongly
C
upregulated early in heart failure. A, Deter-
mination of PI16 mRNA expression in car-
diomyopathic myocardium by real-time
kDa PCR. Total RNA was isolated from left
115
82 PI16 group) of a mouse model of cardiac failure
(␤1-adrenergic receptor [AR] transgenic
**P⬍0.01, ***P⬍0.001, t test) and of mice
hypertrophy (after transverse aortic con-
striction, TAC, n⫽6; *P⬍0.05 TAC vs con-
trol, t test) and from human failing myocar-
dium (n⫽6). B, Western blotting of heart
1772 Circulation October 16, 2007

Control
A B Adv-lacZ

Isoleucine incorporation
Adv-PI16
Adv-lacZ Control Adv-lacZ Iso/PE Adv-PI16 Iso/PE 14000
* **
10000

(cpm)
Adv-lacZ Iso/PE 6000

α-actinin, nuclei 2000

Iso/PE - + + + + + + +
MOI
0.1 0.4 1.6
C

Cardiomyocyte size
Adv-lacZ 2500
Adv-PI16

(arbitrary units)
2000
(fold induction)

(fold induction)
2 ** ***
ANP mRNA

BNP mRNA ***

6 1500
4 ** 1000
1
2 500
0 0 0
Iso/PE - - + + Iso/PE - - + + Iso/PE - + + + + + + +
MOI
0.1 0.4 1.6
D E Control

Isoleucine incorporation
Control scr-RNAi PI16-RNAi scr-RNAi
3000 PI16-RNAi
*

(cpm)
2000

α-actinin, nuclei 1000

kDa 0
Iso/PE - - - + + +
115
PI16
82
64
scr-RNAi PI16-RNAi
Figure 4. PI16 inhibits hypertrophy of cardiomyocytes in vitro. A, Immunofluorescent staining of neonatal rat cardiomyocytes (NRCMs).
Cells transfected with lacZ-expressing (Adv-lacZ) or PI16-expressing (Adv-PI16) adenovirus were stimulated with isoproterenol and
phenylephrine (Iso/PE, 5/50 ␮mol/L) or left untreated. Green indicates ␣-actinin; nuclei are stained with Hoechst (blue). B, Protein
synthesis and cell size of NRCMs infected with Adv-lacZ or Adv-PI16 and stimulated with Iso/PE. Top, Protein synthesis of isolated
NRCMs was ascertained by determination of [3H]-isoleucine incorporation. Data are means from 3 experiments. Bottom, For morpho-
metric analysis of cardiomyocyte surface areas, ⬎40 individual ␣-actinin–stained cells per group were analyzed by digitizing the images
and using computerized pixel counting. cpm indicates counts per minute. *P⬍0.05, **P⬍0.01, ***P⬍0.01, Adv-PI16 vs Adv-lacZ by
ANOVA with Bonferroni posttest. C, Expression of atrial natriuretic peptide (ANP) and brain natriuretic peptide (BNP) mRNA as indica-
tors of a prohypertrophic gene expression program. Real-time PCR was performed on total mRNA preparations from NRCMs treated
with isoproterenol/phenylephrine (Iso/PE, 5/50 ␮mol/L). Data are means of 3 independent experiments with 3 replicates each. Adv-PI16
or Adv-lacZ was used at an MOI of 1.6. **P⬍0.01 for Adv-PI16 vs Adv-lacZ Iso/PE-stimulated cells, t test. D, RNA interference (RNAi)–
mediated suppression of endogenous PI16 in NRCM. Top, Scrambled-RNAi (scr-RNAi) or PI16-RNAi (40 nmol/L)–transfected NRCMs or
mock-transfected NRCMs (control) after 40 hours’ starving without serum. Bottom, Western blotting of PI16 in lysates derived from
PI16-RNAi and control RNAi-treated NRCMs infected with Adv-PI16 (MOI 0.1). E, Determination of protein synthesis in NRCMs by
[3H]-isoleucine incorporation. Data are means from 4 independent experiments, each comprising 3 replicates per treatment condition.
P⬍0.01 for PI16-RNAi vs scr-RNAi by ANOVA, *P⬍0.05 for PI16-RNAi vs scr-RNAi transfected, unstimulated cells, by Bonferroni post-
test. Iso/PE indicates isoproterenol/phenylephrine.

expression to be markedly upregulated in heart failure (Figure
3B). Similar to the adult murine heart, only weak PI16
expression occurred in isolated neonatal rat cardiomyocytes
under basal conditions; however, PI16 expression was
strongly induced in cardiomyocytes after serum stimulation
(Figure 3C).
PI16 Inhibits Hypertrophy of Cardiomyocytes
To determine the effect of PI16 on cardiomyocyte growth, we
infected neonatal rat cardiomyocytes with Adv-PI16. Al-
though PI16 overexpression had no effect on cardiomyocyte
Downloaded from http://circ.ahajournals.org/ by guest on June 2, 2014
size under normal conditions (data not shown), it inhibited
isoproterenol/phenylephrine-induced cardiomyocyte hyper-
trophy as assessed by morphometric analysis of cell size and
by isoleucine incorporation (Figure 4A and 4B). Interest-
ingly, immunofluorescent staining demonstrated that the
average cardiomyocyte size was already strongly reduced by
overexpression of PI16 in only ⬇10% of the cells (MOI 0.1,
immunofluorescent detection of PI16 expression; data not
shown), consistent with secretion of PI16 (Adv-lacZ versus
Adv-PI16, P⬍0.01; Figure 4B). At MOI 0.4, ⬇30% of the
cells expressed PI16, and at MOI 1.6, ⬇90% expressed PI16.
 Frost and Engelhardt PI16 in Heart Failure 1773

kDa
115
A 82 PI16 Figure 5. Inhibition of cardiomyocyte
growth in PI16 transgenic mice. A, Genera-
αMHC promoter PI16 64 tion of transgenic mice overexpressing PI16
β α1 α2 α3 SV40 t intron pA
WT PI16-TG PI16-TG (PI16-TG) under control of the ␣-myosin
1:20 1:5 heavy chain (␣MHC) promoter. To assess
B 12000 the expression level of transgenic PI16 rela-

dP/dtmax (mmHg/s)
tive to endogenous PI16, myocardial lysates
10000 from PI16-TG mice were diluted stepwise

pressure (mmHg)
Heart rate (bpm)

Left ventricular
with lysates from wild-type myocardium. B,
120 8000 In vivo cardiac hemodynamics were ana-
600
400 100 lyzed in anesthetized mice by left ventricular
80 6000
200 WT catheterization. Dobutamine was adminis-
60 TG tered via the left jugular vein (n⫽10 to 15).
0 4000
WT TG WT TG WT TG WT TG Heart rate and left ventricular pressure are
-- - -
0 3 10 30 100
Dobutamine + + Dobutamine + + Dobutamine (µg/kg/min) depicted under basal conditions and during
maximal dobutamine stimulation (125 ␮g ·
kg⫺1 · min⫺1). WT indicates wild-type. C,
C Wild-type PI16-TG Top, Histological analysis of left ventricular
myocardium from 3-month-old wild-type
and PI16-TG mice. Five-micrometer sec-
tions were cut from paraffin-embedded tis-
sues and stained with hematoxylin and
eosin. Bottom left, Ventricular weight to
body weight ratio of wild-type and PI16-TG
mice. Age, 3 months; n⫽7. Bottom right,
Morphometric analysis of myocyte cross-
body weight (mg/g)

Cardiomyocyte size
Ventricular weight /

4.5 3000 sectional areas. A total of 40 to 60 individual

(arbitrary units)

4.0 cells per animal were analyzed by digitiza-

3.5 2500
tion of the images and computerized pixel
3.0 ** 2000 ** counting (5 to 6 animals per group).
2.5
2.0 1500 **P⬍0.01, t test.
1.5 1000
WT PI16-TG WT PI16-TG

Use of higher MOIs resulted in accumulation of PI16 in the PI16 transgenic mice developed normally, had normal
endoplasmic reticulum and Golgi apparatus, probably due to cardiac function (Figure 5B), and showed a normal myocar-
time-consuming posttranslational modifications. We next an- dial structure without any signs of interstitial fibrosis (Sirius
alyzed the effect of PI16 expression on changes in the gene red staining, data not shown); however, PI16 transgenic mice
expression program typically associated with cardiomyocyte had significantly smaller hearts than wild-type mice (Figure
hypertrophy. Treatment of neonatal rat cardiomyocytes with 5C). In line with these findings, individual cardiomyocytes
isoproterenol/phenylephrine led to a marked induction of from PI16 transgenic mice were significantly smaller (⫺24%,
atrial natriuretic peptide and brain natriuretic peptide mRNA P⬍0.01; Figure 5C).
expression, which was blunted in Adv-PI16 – expressing cells
(Figure 4C). We then sought to determine whether endoge-
nous PI16 is required for cardiomyocyte growth control. We
Discussion
To the best of our knowledge, the present study describes the
transfected synthetic interfering RNA directed against PI16
first systematic search for secreted proteins in the heart using a
into neonatal rat cardiomyocytes. This led to a marked
suppression of PI16 protein levels (Figure 4D). Subsequent biological screen. We identified 54 cardiac cDNAs that contain
analysis of neonatal rat cardiomyocyte cell size by immuno- a secretion signal by a secretion trap screen in yeast. Among
fluorescent staining (Figure 4D) revealed a marked increase them are well-known genes such as atrial natriuretic peptide, but
in cardiomyocyte cell size and reorganization of the sarco- they also include genes with unknown cardiac expression and
meres. These characteristics of cardiomyocyte hypertrophy genes with unknown function and expression. Thirty of the
were accompanied by a significant increase in cardiomyocyte identified proteins are putatively secreted. Another 15 proteins
protein synthesis through RNA interference–mediated sup- have secretion signals and 1 membrane domain but might be
pression of PI16 (Figure 4E). secreted after being shed from the plasma membrane.18
To determine the role of PI16 in the intact mammalian Secreted proteins play a major role in intercellular com-
heart, we generated transgenic mice that overexpress PI16 in munication. Several secreted peptides and proteins have been
a cardiomyocyte-specific manner (Figure 5A). To assess described as being involved in cardiac remodeling.6 How-
expression of the transgenic protein relative to endogenous ever, many proteins known to be secreted in the heart, such as
PI16 levels, we performed stepwise dilutions of myocardial extracellular matrix proteins, proteases, and autocrine/para-
lysates from PI16 transgenic mice with lysates from wild- crine factors such as transforming growth factor-␤, tumor
type littermates (Figure 5A). We thereby determined the level necrosis factor-␣, fibroblast growth factor, platelet-derived
of transgene expression to be ⬇20-fold higher than endoge- growth factor, and interleukin-6, have been shown to be
nous PI16 expression in nonfailing myocardium. secreted by activated cardiac fibroblasts.6 Little is known
Downloaded from http://circ.ahajournals.org/ by guest on June 2, 2014
1774 Circulation October 16, 2007
about the secretome of cardiomyocytes, although proteins
secreted from cardiomyocytes might play an important role in
myocardial remodeling. Cardiomyocyte-specific overexpres-
sion of the ␤1-adrenergic receptor, for example, is accompa-
nied by activation of fibroblasts and interstitial fibrosis,
although the initial stimulus for remodeling was set in
cardiomyocytes.15 To avoid enrichment of mRNAs encoding
for secreted proteins from noncardiomyocytes, we generated
a cardiac cDNA library from young, healthy wild-type mice
that showed no signs of fibroblast activation.
We then studied regulation of the identified genes during the
development of heart failure. The mRNA expression of PI16, a
thus far largely uncharacterized protein, was strongly upregu-
lated early in heart failure. Here, we demonstrate that PI16 is
expressed and secreted in the mammalian heart and subse-
quently accumulates extracellularly. Although there are putative
PI16 homologs in several species, the function of PI16 is
unknown. Bioinformatic analysis revealed that PI16 contains a
sperm-coating glycoprotein domain. This evolutionary highly
conserved protein domain can be found in ⬎600 mainly extra-
cellular eukaryotic proteins of many different species, including
bacteria.19 It was recently shown that PSP94 binding protein, the
human homolog of PI16, may be partially bound in the serum to
the prostate secretory protein of 94 amino acids (PSP94).20
PSP94 is a protein that is upregulated in prostate cancer and may
have growth-regulating properties in this disease,21 but the
signaling pathway and the receptor are unknown. However, we
could not detect any PSP94 mRNA or protein expression in
either the healthy or the failing heart, which suggests a different
signaling pathway in this setting. In the gene ontology classifi-
cation, PI16 was inferred as protease by electronic annotation.
Our own sequence analysis (ProtFun 2.222) supports an enzy-
matic function (probability 0.48; odds 1.67) and negates a
function as a structural protein (probability 0.008; odds 0.28).
Proteases have several important functions in cardiac disease; for
example, matrix metalloproteases and their inhibitors (tissue
inhibitors of matrix metalloproteinase) play a crucial role in
extracellular matrix modification in heart failure.23
Expression of PI16 in primary cardiomyocytes revealed a
potent function of PI16 as an endogenous repressor of
cardiomyocyte growth. Interestingly, PI16 exerted its antihy-
pertrophic effect even when only a small fraction of cardio-
myocytes were infected with Adv-PI16, which supports a
paracrine function of PI16. When we expressed PI16 in the
hearts of transgenic mice, we found a profound inhibition of
cardiac growth in the absence of any impairment of cardiac
structure or function. In line with the experiments conducted
in vitro, we found cardiomyocytes from PI16 transgenic
hearts to be significantly smaller than wild-type cardiomyo-
cytes. In vitro, PI16 did not influence cardiomyocyte cell size
under resting conditions, which indicates that the ability of
PI16 to control cardiomyocyte size critically depends on the
cellular environment that accompanies cell growth.
The rapid and strong induction of PI16 and its growth-in-
hibitory effect under conditions of cardiac stress are reminis-
cent of atrial natriuretic peptide and brain natriuretic pep-
tide24,25 and the cytokine GDF15 (growth-differentiation
factor 15).26 All 3 secreted proteins are induced in cardiac
Downloaded from http://circ.ahajournals.org/ by guest on June 2, 2014
disease and serve as endogenous feedback mechanisms to
control excessive growth-promoting stimuli.
When PI16 was overexpressed to an extent that is comparable
to conditions of cardiac disease, we observed a significant
suppression of physiological cardiac growth. This finding sets
PI16 apart from other antihypertrophic molecules such as
GDF1526 and NAB1,27 which exert their antihypertrophic prop-
erties only in cardiac disease. This pivotal difference suggests
alternative downstream growth control pathways that are af-
fected by PI16. Secretion of PI16 may serve a similar role as an
endogenous cardioprotective signaling pathway and might be
used as a therapeutic strategy in disease states such as cardiac
failure and hypertrophic cardiomyopathy.
In summary, using a genetic yeast screen of a heart cDNA
library, we identified 54 cardiac cDNAs encoding for proteins
that contain a secretion signal. Among them are many genes
with unknown cardiac expression and function. PI16, a newly
identified protein secreted by the heart, is strongly upregu-
lated early in the development of heart failure and inhibits
cardiomyocyte hypertrophy in vivo and in vitro.
Acknowledgment
We would like to thank M.J. Lohse (Institute of Pharmacology,
University of Wuerzburg) for many helpful discussions during this
study. We thank A. Spang (Friedrich Miescher laboratory of the
Max-Planck Institute (MPI), Tuebingen, Germany) for advice on
yeast work; C. Weitz (Department of Neurobiology, Harvard Med-
ical School, Boston, Mass) for kindly providing the host yeast strain
066-2 and the mutant invertase plasmid pSuc2dMSP; A. Kramer
(Humboldt University, Berlin, Germany) for support with the yeast
screen; and A. Ahles for help with real-time PCR analysis, E.
Schmitteckert for pronuclear injection of the transgene construct, and
M. Babl for support with ventricular catheterization (Rudolf Vir-
chow Center, University of Wuerzburg). We gratefully acknowledge
the help of A. Kopp-Schneider (Department of Biostatistics of the
German Cancer Research Center, Heidelberg, Germany) with statis-
tical analysis, J. Schultz and coworkers (Department of Bioinformat-
ics, University of Wuerzburg) with bioinformatic analysis, and A.
Sickmann (Rudolf Virchow Center, University of Wuerzburg) with
protein biochemistry.
Sources of Funding
These studies were supported by grants from the Deutsche For-
schungsgemeinschaft (DFG), the German Cardiac Society (Hengst-
berger award), the Rudolf Virchow Center/DFG Research Center for
Experimental Biomedicine, the Bavarian Ministry of Technology,
Trigen, and Sanofi-Aventis. Dr Frost is a fellow of the MD/PhD
program of the University of Wuerzburg funded by the Interdisci-
plinary Center for Clinical Research.
Disclosures
Drs Frost and Engelhardt have filed a patent application for the use of
PI16 in cardiovascular disease.
References
1. Towbin JA, Bowles NE. The failing heart. Nature. 2002;415:227–233.
2. Fredj S, Bescond J, Louault C, Potreau D. Interactions between cardiac
cells enhance cardiomyocyte hypertrophy and increase fibroblast prolif-
eration. J Cell Physiol. 2005;202:891– 899.
3. Gnecchi M, He H, Liang OD, Milo LG, Morello F, Mu H, Noiseux N,
Zhang L, Pratt RE, Ingwall JS, Dzau VJ. Paracrine action accounts for
marked protection of ischemic heart by Akt-modified mesenchymal stem
cells. Nat Med. 2005;11:367–368.
4. Uemura R, Xu M, Ahmad N, Ashraf M. Bone marrow stem cells prevent
left ventricular remodeling of ischemic heart through paracrine signaling.
Circ Res. 2006;98:1414 –1421.
 Frost and Engelhardt PI16 in Heart Failure 1775

5. Bouzegrhane F, Thibault G. Is angiotensin II a proliferative factor of
cardiac fibroblasts? Cardiovasc Res. 2002;53:304 –312.
6. Manabe I, Shindo T, Nagai R. Gene expression in fibroblasts and fibrosis:
involvement in cardiac hypertrophy. Circ Res. 2002;91:1103–1113.
7. Clark HF, Gurney AL, Abaya E, Baker K, Baldwin D, Brush J, Chen J,
Chow B, Chui C, Crowley C, Currell B, Deuel B, Dowd P, Eaton D, Foster
J, Grimaldi C, Gu Q, Hass PE, Heldens S, Huang A, Kim HS, Klimowski L,
Jin Y, Johnson S, Lee J, Lewis L, Liao D, Mark M, Robbie E, Sanchez C,
Schoenfeld J, Seshagiri S, Simmons L, Singh J, Smith V, Stinson J, Vagts A,
Vandlen R, Watanabe C, Wieand D, Woods K, Xie MH, Yansura D, Yi S,
Yu G, Yuan J, Zhang M, Zhang Z, Goddard A, Wood WI, Godowski P, Gray
A. The secreted protein discovery initiative (SPDI), a large-scale effort to
identify novel human secreted and transmembrane proteins: a bioinformatics
assessment. Genome Res. 2003;13:2265–2270.
8. Grimmond SM, Miranda KC, Yuan Z, Davis MJ, Hume DA, Yagi K,
Tominaga N, Bono H, Hayashizaki Y, Okazaki Y, Teasdale RD. The
mouse secretome: functional classification of the proteins secreted into
the extracellular environment. Genome Res. 2003;13:1350 –1359.
9. Brent MR. Genome annotation past, present, and future: how to define an
ORF at each locus. Genome Res. 2005;15:1777–1786.
10. Klein RD, Gu Q, Goddard A, Rosenthal A. Selection for genes encoding
secreted proteins and receptors. Proc Natl Acad Sci U S A. 1996;93:
7108–7113.
11. Kramer A, Yang FC, Kraves S, Weitz CJ. A screen for secreted factors of
the suprachiasmatic nucleus. Methods Enzymol. 2005;393:645– 663.
12. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using
real-time quantitative PCR and the 2(-Delta Delta C(T)) method.
Methods. 2001;25:402– 408.
13. Rochais F, Vilardaga JP, Nikolaev VO, Bunemann M, Lohse MJ,
Engelhardt S. Real-time optical recording of ␤1-adrenergic receptor acti-
vation reveals supersensitivity of the Arg389 variant to carvedilol. J Clin
Invest. 2007;117:229 –235.
14. Merkle S, Franz S, Schoen MP, Bauersachs J, Buitrago M, Frost RJ,
Schmitteckert E, Lohse MJ, Engelhardt S. A role for caspase-1 in heart
failure. Circ Res. 2007;100:645– 653.
15. Engelhardt S, Hein L, Wiesmann F, Lohse MJ. Progressive hypertrophy
and heart failure in ␤1-adrenergic receptor transgenic mice. Proc Natl
Acad Sci U S A. 1999;96:7059 –7064.
16. Hubbard T, Andrews D, Caccamo M, Cameron G, Chen Y, Clamp M,
Clarke L, Coates G, Cox T, Cunningham F, Curwen V, Cutts T, Down T,
Durbin R, Fernandez-Suarez XM, Gilbert J, Hammond M, Herrero J,
Hotz H, Howe K, Iyer V, Jekosch K, Kahari A, Kasprzyk A, Keefe D,
Keenan S, Kokocinsci F, London D, Longden I, McVicker G, Melsopp C,
Meidl P, Potter S, Proctor G, Rae M, Rios D, Schuster M, Searle S,
Severin J, Slater G, Smedley D, Smith J, Spooner W, Stabenau A, Stalker
J, Storey R, Trevanion S, Ureta-Vidal A, Vogel J, White S, Woodwark C,
Birney E. Ensembl 2005. Nucleic Acids Res. 2005;33:D447–D453.
17. Schultz J, Milpetz F, Bork P, Ponting CP. SMART, a simple modular
architecture research tool: identification of signaling domains. Proc Natl
Acad Sci U S A. 1998;95:5857–5864.
18. Tschumperlin DJ, Dai G, Maly IV, Kikuchi T, Laiho LH, McVittie AK,
Haley KJ, Lilly CM, So PT, Lauffenburger DA, Kamm RD, Drazen JM.
Mechanotransduction through growth-factor shedding into the extra-
cellular space. Nature. 2004;429:83– 86.
19. Mulder NJ, Apweiler R, Attwood TK, Bairoch A, Bateman A, Binns D,
Bradley P, Bork P, Bucher P, Cerutti L, Copley R, Courcelle E, Das U,
Durbin R, Fleischmann W, Gough J, Haft D, Harte N, Hulo N, Kahn D,
Kanapin A, Krestyaninova M, Lonsdale D, Lopez R, Letunic I, Madera
M, Maslen J, McDowall J, Mitchell A, Nikolskaya AN, Orchard S, Pagni
M, Ponting CP, Quevillon E, Selengut J, Sigrist CJ, Silventoinen V,
Studholme DJ, Vaughan R, Wu CH. InterPro, progress and status in 2005.
Nucleic Acids Res. 2005;33:D201–D205.
20. Reeves JR, Xuan JW, Arfanis K, Morin C, Garde SV, Ruiz MT, Wisniewski
J, Panchal C, Tanner JE. Identification, purification and characterization of a
novel human blood protein with binding affinity for prostate secretory protein
of 94 amino acids. Biochem J. 2005;385:105–114.
21. Garde SV, Basrur VS, Li L, Finkelman MA, Krishan A, Wellham L,
Ben-Josef E, Haddad M, Taylor JD, Porter AT, Tang DG. Prostate
secretory protein (PSP94) suppresses the growth of androgen-
independent prostate cancer cell line (PC3) and xenografts by inducing
apoptosis. Prostate. 1999;38:118 –125.
22. Jensen LJ, Gupta R, Blom N, Devos D, Tamames J, Kesmir C, Nielsen H,
Staerfeldt HH, Rapacki K, Workman C, Andersen CA, Knudsen S, Krogh
A, Valencia A, Brunak S. Prediction of human protein function from
post-translational modifications and localization features. J Mol Biol.
2002;319:1257–1265.
23. Spinale FG. Matrix metalloproteinases: regulation and dysregulation in
the failing heart. Circ Res. 2002;90:520 –530.
24. Knowles JW, Esposito G, Mao L, Hagaman JR, Fox JE, Smithies O,
Rockman HA, Maeda N. Pressure-independent enhancement of cardiac
hypertrophy in natriuretic peptide receptor A-deficient mice. J Clin
Invest. 2001;107:975–984.
25. Holtwick R, van Eickels M, Skryabin BV, Baba HA, Bubikat A, Begrow
F, Schneider MD, Garbers DL, Kuhn M. Pressure-independent cardiac
hypertrophy in mice with cardiomyocyte-restricted inactivation of the
atrial natriuretic peptide receptor guanylyl cyclase-A. J Clin Invest. 2003;
111:1399 –1407.
26. Xu J, Kimball TR, Lorenz JN, Brown DA, Bauskin AR, Klevitsky R,
Hewett TE, Breit SN, Molkentin JD. GDF15/MIC-1 functions as a pro-
tective and antihypertrophic factor released from the myocardium in
association with SMAD protein activation. Circ Res. 2006;98:342–350.
27. Buitrago M, Lorenz K, Maass AH, Oberdorf-Maass S, Keller U, Schmit-
teckert EM, Ivashchenko Y, Lohse MJ, Engelhardt S. The transcriptional
repressor Nab1 is a specific regulator of pathological cardiac hypertrophy.
Nat Med. 2005;11:837– 844.

CLINICAL PERSPECTIVE
Left ventricular hypertrophy is a strong and independent predictor of outcome in cardiac failure. Because the heart is a
largely postmitotic organ, growth of the heart in response to mechanical or neurohumoral stimulation primarily occurs
through an increase in cardiomyocyte size. Cardiomyocyte hypertrophy may initially serve a compensatory role; a classic
example of this is the partial normalization of an increase in wall stress after pressure overload of the left ventricle. With
the chronic presence of a detrimental stimulus, such as enhanced catecholamine concentrations in heart failure,
cardiomyocyte hypertrophy is believed to ultimately become maladaptive and contribute to the progression of heart failure.
Therefore, a more detailed understanding of the molecular events that mediate and regulate cardiomyocyte growth may lead
to novel therapeutic strategies in cardiac hypertrophy and heart failure. In this respect, proteins that are secreted from the
heart, such as the cardioprotective atrial natriuretic peptides, are of particular interest, but only a limited set of such proteins
are currently known. The present work sought to identify additional proteins that are secreted from the heart and that may
act as regulators of cardiomyocyte growth. To this end, a genetic yeast-based screening approach was applied to the heart
to systematically identify novel secreted proteins. Among the discovered proteins, proteinase inhibitor 16 was identified
as being strongly upregulated in diseased myocardium and as acting as a potent repressor of cardiomyocyte growth.
Although its effects are speculative at this early point, proteinase inhibitor 16 might eventually be used as a therapeutic
means to treat cardiac hypertrophy and heart failure.

Downloaded from http://circ.ahajournals.org/ by guest on June 2, 2014

Fig. S1
A
Genomic PI16-DNA
E1 E2 E3 E4 E5 E6 E7

PI16-FL
1000bp
PI16-SV
500bp

B Hu
M an
ou
m

PI16-Protein se

SS SCP-Domain
aa 24 193 489

C
PI16-homologs

SCP domain
 Fig. S2
A
kDa
115
82 PI16
64
Ovary
Muscle
Heart
Aorta
Uterus
Intestine
Colon

Spleen

Brain
Skin
Adipose tissue
Kidney
Liver

Testis

B Peptide-Ab Exon5-Ab
N- PI16 -C
Full length-Ab

kDa kDa kDa

115 115 115
82 82 82
64 64 64
48 48 48
Full length-Ab Exon5-Ab Peptide-Ab

C
kDa PI16
115 PI16 (deglycosylated)
82
64
48
PNGaseF - +
Supplementary methods

Bioinformatic analyses of identified genes

Prediction of signal peptides of the identified genes was performed by SignalP 3.0 1 and
2
prediction of non-classical secretion was carried out by SecretomeP 1.0 .

Transmembrane helices have hydrophobic properties similar to signal peptides. To

detect membrane proteins among the identified genes, transmembrane helices were

predicted by TMHMM 2.0 3.

Bioinformatic search for homologues

A BLAST search against the sequenced genomes of several species was performed

with the protein sequence of mouse PI16. The identified proteins were only annotated

as putative homologues if the mouse PI16 was found in a back search. Of the identified

proteins, many were just seeming homologues that only contained the large SCP

domain, as for example the Crisp (Cysteine rich protein)-family. To exclude such false-
4
positive homologues we calculated an evolutionary tree using CLUSTAL W with the

protein sequences of more than 30 of the most similar proteins of several species by

means of that we clearly could separate real PI16 homologues. CLUSTAL W was then

used for multiple alignment. The grade of homology was calculated taking the different

sizes of the proteins into account, e.g. the short fish protein (Danio rerio) was aligned

only with the corresponding part of mouse PI16.

RNA isolation and real-time quantitative PCR

The quality of the isolated RNA was analyzed by denaturing agarose gel

electrophoresis. First-strand cDNA was synthesized by using Superscript II reverse

 CIRCULATIONAHA/2007/696468R2 Frost et al.

transcriptase (Invitrogen, Carlsbad, CA) and oligo-dT primers. Real-time PCR was

performed with Sybr Green (Cambrex BioScience, East Rutherford, NJ) as fluorescent

and 6-carboxy-X-rhodamine (Rox, Invitrogen) as reference dye using an ABI PRISM

Sequence Detection System 7700 (Applied Biosystems, Foster City, Ca). Threshold

cycle (Ct) values were determined using the Sequence Detector 1.7 software. Data

were calculated with the 2-ΔΔCT method. After each experiment dissociation curves

were analyzed to control for specificity of the amplification product. GAPDH was used

as a reference. The primer sequences used for real-time PCR analysis are listed in

supplementary table S1. PCR conditions using heat-activatable Taq polymerase (Hot

Master Taq; Eppendorf, Hamburg, Germany) were as follows: 40 cycles of 94ºC for 20 s

(2 min initial cycle), 56ºC for 20 s and 65ºC for 35 s. Human PI16 was detected by

TaqMan gene expression assay Hs00542137_m1 according to the manufacturer’s

instructions (Applied Biosystems).

Human heart tissue

Human heart specimens were analyzed from stored samples that had originally been

obtained from patients undergoing heart transplantation 5. After removal of the hearts,

samples had been snap frozen in liquid nitrogen and stored at -80ºC. Non-failing donor

hearts that had not been transplanted for technical reasons were used as controls.

Echocardiography and past medical history of the donors had shown no indications of

heart disease.

2
 CIRCULATIONAHA/2007/696468R2 Frost et al.

Generation of PI16-specific antibodies and Western blot analysis.

The coding sequence of full-length PI16 lacking the first 23 amino acids (i.e. the signal

sequence) was amplified by PCR from a mouse cardiac cDNA library and inserted into

the pDEST15 expression vector (Invitrogen). The recombinant GST-tagged protein was

insoluble and was purified from inclusion bodies in E. coli. Exon 5 of PI16 was amplified

by PCR and inserted into the pDEST15 expression vector. The soluble exon 5 fragment

was then purified by column chromatography from bacterial cell lysates. The purified

full-length protein, the exon 5 peptide and a synthetized peptide corresponding to amino

acid positions 84-94 of PI16 were then injected into rabbits (Immunoglobe, Himmelstadt,

Germany) to generate polyclonal antibodies. The purified antigen coupled to HiTrap

NHS-activated columns (GE Healthcare, Chalfont St Gilles, UK) was used for affinity

purification of the antisera according to the manufacturer`s instructions.

For Western blotting tissue or cell lysates containing 30 µg of protein were separated by

electrophoresis on 10% SDS-PAGE gels and transferred to Immobilion-P membranes

(Millipore Corporation, Billerica, MA). After blocking the membranes with 5% milk for 2

h, incubation with primary antibody overnight at 4ºC (diluted 1:5,000), followed by

secondary anti-rabbit antibody for 1 h at room temperature (diluted 1:10,000; Dianova,

Hamburg, Germany) and chemiluminescent detection (ECL-Plus, GE Healthcare) were

performed.

Generation of PI16 adenovirus

Mouse full-length PI16 was cloned under the control of the CMV promotor into the

vector pAD/CMV/Dest according to the manufacturer’s protocol (Invitrogen). HEK293A

cells were then transfected with the PI16 vector and cultured until they rounded off as a
3
 CIRCULATIONAHA/2007/696468R2 Frost et al.

sign of high adenovirus load. After three rounds of virus amplification with increasing

amounts of HEK293A cells the virus titer was determined with a plaque assay in

agarose covered HEK293A cells.

Immunofluorescent staining

Cells or cryosections (5 µM) were fixed for 10 min using 4% paraformaldehyde. After 5

min permeabilization with 0.2% Triton X100, samples were incubated with primary

antibodies (Anti-PI16 diluted 1:1000, Anti-α-actinin diluted 1:500 (Sigma-Aldrich, St.

Louis, MO) and Hoechst for 30 min at 37ºC. Incubation with secondary antibodies (Anti-

Rabbit Alexa 555, Anti-Mouse Alexa 488, Invitrogen) was then for 30 min at 37ºC. For

determination of cell size cardiomyocytes were digitized by microscopy and cross-

sectional areas were determined in a blinded way by computerized pixel counting.

Isoleucine incorporation

Neonatal rat cardiomyocytes were cultured in 24-well plates and transfected with a lacZ

control or a PI16-expressing adenovirus. After 24 h starvation without serum, cells were

stimulated for 40 h with isoproterenol/phenylephrine (5 µM/50 µM). [3H]isoleucine was

added at a final concentration of 1 nCi/ml (0.25 nCi/ml for RNAi experiments). After

precipitation with 10% trichloroacetic acid and dissolving with 0.5 N NaOH, the

incorporated radioactivity was counted in a scintillation counter.

Suppression of endogenous PI16 by RNAi

Neonatal rat cardiomyocytes (NRCM) were transfected 24 h after isolation with RNAi

using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions.

Cells were transfected using a nonhomologous scrambled RNAi (RNAi negative control

4
 CIRCULATIONAHA/2007/696468R2 Frost et al.

with a medium GC content, Invitrogen) or a RNAi directed against PI16 (5’-AUG GAU

GUU AGC UUC CUC CAC UCC C-3’, final concentration 40 nmol, Invitrogen) or only

exposed to the transfection reagent (control). After starving without serum for 24 h,

[3H]isoleucine and isoproterenol/phenylephrine (2,5 µM/25 µM) or control buffer were

added for isoleucine incorporation assays and cells were cultivated for another 40 h.

Generation of transgenic mice

PI16-transgenic mice were generated by pronuclear injection of fertilized oocytes from

FVB/N mice with a transgene construct containing the coding sequence of the mouse

PI16 full-length under the control of the murine α-myosin heavy chain promotor. All mice

were housed in a SPF facility.

In vivo hemodynamic analysis

For left ventricular catheterization, a miniaturized pressure sensing catheter (1.4 F

Micro-tip catheter, Millar instruments, Houston, TX) was introduced via the right carotid

artery under anesthesia with tribromoethanol. Increasing doses of dobutamine were

infused via the left jugular vein. Chart software (Chart v5.0.2, ADInstruments,

Spechbach, Germany) was employed for data recording (2000 Hz) and analysis of

parameters of cardiac function.

Histological analyses

Tissue sections (5 µm) of left ventricular myocardium were stained with

hematoxylin/eosin for determination of cardiomyocyte cross-sectional areas. Cell size

was analyzed by digitizing the images and computerized pixel counting. Only nucleated

5
 CIRCULATIONAHA/2007/696468R2 Frost et al.

cardiac myocytes from areas of transversely cut muscle fibres were included in the

analyses by an investigator blinded to the genotype.

6
 CIRCULATIONAHA/2007/696468R2 Frost et al.

References to Supplemental Methods

1. Bendtsen JD, Nielsen H, von Heijne G, Brunak S. Improved prediction of signal

peptides: SignalP 3.0. J Mol Biol. 2004;340:783-795.

2. Bendtsen JD, Jensen LJ, Blom N, Von Heijne G, Brunak S. Feature-based

prediction of non-classical and leaderless protein secretion. Protein Eng Des Sel.

2004;17:349-356.

3. Krogh A, Larsson B, von Heijne G, Sonnhammer EL. Predicting transmembrane

protein topology with a hidden Markov model: application to complete genomes.

J Mol Biol. 2001;305:567-580.

4. Thompson JD, Higgins DG, Gibson TJ. CLUSTAL W: improving the sensitivity of

progressive multiple sequence alignment through sequence weighting, position-

specific gap penalties and weight matrix choice. Nucleic Acids Res.

1994;22:4673-4680.

5. Buitrago M, Lorenz K, Maass AH, Oberdorf-Maass S, Keller U, Schmitteckert

EM, Ivashchenko Y, Lohse MJ, Engelhardt S. The transcriptional repressor Nab1

is a specific regulator of pathological cardiac hypertrophy. Nat Med. 2005;11:837-

844.

7
 CIRCULATIONAHA/2007/696468R2 Frost et al.

Figure legends to supplemental tables

Table S1: Primer sequences and product lengths of the mRNAs assessed by real-

time PCR

ART 1, ADP-ribosyltransferase 1; Cst3, Cystatin; Crip2, cysteine rich protein 2; Cxcl14,

chemokine (C-X-C motif) ligand 14; GAPDH, Glyceraldehyde-3-phosphate

dehydrogenase; GCSH, glycine cleavage system protein H; Lrg1, leucine-rich α-2-

glycoprotein; Manba, mannosidase β A; MPG1, melanocyte proliferating gene 1; Nppa,

natriuretic peptide precursor type A (atrial natriuretic peptide); PI16, protease inhibitor

16; Psap, prosaposin; TAPbp, TAP binding protein; Tdel 1, tumor differencially

expressed 1 like; Tsfm, Ts translation elongation factor; Nov, nephroblastoma

overexpressed gene; Nppb, natriuretic peptide precursor B (brain natriuretic peptide);

Primers to amplify the respective mouse sequence are listed, unless indicated

otherwise.

Table S2: Genes identified by the secretion trap screen

54 genes containing a secretion signal were identified by a secretion trap screen using

a cardiac cDNA library. Genes are depicted in descending order according to their

probability of secretion. Gene name, function and GeneID have been assigned

according to Entrez Gene (NCBI). IEA means that the function is inferred only from

electronic annotation. Prediction of signal sequence (SS) or signal anchor (SA) by the

neural network method and the hidden markov model of SignalP or if indicated only by

the hidden markov model (HMM). The number of transmembrane helices was predicted

by TMHMM 2.0. Prediction of non-classical secretion was carried out by SecretomP 1.

8
 CIRCULATIONAHA/2007/696468R2 Frost et al.

“W” = warning e.g. prediction of a signal sequence for classical secretion; a score >0.6

suggests secretion including the odds that the sequence in fact is secreted.

9
 CIRCULATIONAHA/2007/696468R2 Frost et al.

Figure legends to supplemental figures

Figure S1: Structure of the PI16 gene and protein

A. Identification of a novel cardiac splice variant of PI16. The genomic sequence of PI16

contains seven exons. By cloning of PI16 with primers at the 3`- and 5`-end we

identified a new splice variant of murine PI16 that lacks exon 5. White, 3`-UTR; E, exon

B. Protein structure of mouse PI16. The first 23 amino acids (aa) of PI16 comprise a

putative signal sequence (SS) for secretion, the following 169 aa represent a SCP-

domain (SCP, sperm coating glycoprotein).

C. Homologues of PI16. Partial alignment of the amino acid sequences of homologues

of PI16 by CLUSTAL W 4. The evolutionary oldest PI16 homolog is found in fish. Black

labeling of residues indicates identity within the homologues, grey labeling indicates

conservative replacement.

Figure S2. Protein expression of PI16

A. Detection of PI16 in diverse tissues of wild-type FVB mice by Western blotting.

Strong PI16 expression was found in aorta, skin and to a lower extent in adipose tissue,

PI16 expression in the heart of healthy wild-type animals was low. The antibody

directed specifically against full-length PI16 detected three specific PI16 bands in

eukaryotic tissues, migrating at 74 kDa, 100 kDa and 108 kDa.

B. Characterisation of three different polyclonal antibodies raised against PI16.

10
 CIRCULATIONAHA/2007/696468R2 Frost et al.

(upper panel) Localization of a peptide (aa 84-94 of mouse PI16) and the protein

fragment comprising exon 5 used to raise anti-PI16 polyclonal antibodies. All three

antibodies were affinity-purified using the respective antigen.

C. Enzymatic deglycosylation of PI16. After treatment of myocardial tissue of PI16

transgenic mice with PNGaseF, PI16 was detected by Western blotting.

11
 CIRCULATIONAHA/2007/696468R2 Frost et al.

Tab. S1:

Gene Primer sequences Product

size (bp)

ART 1 forward 5`-GGGCAAACGCAGTACATACAA-3` 81

reverse 5`-CTGAGCTACCCAGCCAGCA-3`

CD14 forward 5`-CATGGAGCGTGTGCTTGG-3` 113

reverse 5`-CGGATCTGAGAAGTTGCAGGA-3`

CD 164 forward 5`-GCAGCCCAACATCACCAC-3` 119

reverse 5`-GTTGAAGCTCGCACAGGTTT-3`

Cst3 forward 5-GAGTACAACAAGGGCAGCAAC-3` 92

reverse 5`-AATAGTTCACTCCAGCCACGA-3`

Crip2 forward 5`-CCCAGCAAAGCCTCTAGTGT-3` 108

reverse 5`-GGTCAGGGTCTTGGAGCAG-3`

Cxcl14 forward 5`-CACACTGCGAGGAGAAGATG-3` 126

reverse 5`-CCAGGCATTGTACCACTTGA-3`

GAPDH human forward 5`-GATCATCAGCAATGCCTCCT-3` 140

reverse 5`-GGGCCATCCACAGTCTTCT-3`

GAPDH mouse forward 5`-TGGCAAAGTGGAGATTGTTG-3` 119

reverse 5`-CATTATCGGCCTTGACTGTG-3`

GAPDH rat forward 5`-CTCCCTCAAGATTGTCAGCA-3` 120

reverse 5`-TGATGGCATGGACTGTGG-3`

GCSH forward 5`-TGTCAACAAATCCTGTTACGAAG-3` 120

reverse 5`-CCTCAATGGACTTCACATATTTCTC-3`

Lrg1 forward 5`-AGTCGGCTGAGGGTAGCA-3` 88

12
 CIRCULATIONAHA/2007/696468R2 Frost et al.

reverse 5`-TCCACCGACAGATGGACAG-3`

Manba forward 5`-TTGCTCATCTGGATTACCTCAC-3` 120

reverse 5`-CACCTGACCACTAACGGACTT-3`

MPG1 forward 5`-GTGTGACATCGTGGTGGATG-3` 103

reverse 5`-GGGCACAGTGAACTCATGG-3`

Nov forward 5`-TCGCCAGTGTGAGATGGTAA-3` 114

reverse 5`-ATTTCTTGGTGCGGAGACAC-3`

Nppa forward 5`-TTTCAAGAACCTGCTAGACCAC-3` 98

reverse 5`-CCCTGCTTCCTCAGTCTGCT-3`

Nppa rat forward 5`-GGGTAGGATTGACAGGATTGG-3` 85

reverse 5`-TCGAGCAGATTTGGCTGTT-3`

Nppb rat forward 5`-GCTTTGGGCAGAAGATAGACC-3` 113

reverse 5`-AGGCAGAGTCAGAAGCCAGA-3`

PI16 forward 5`-CCAGTGCCCTCTTGGCTAC-3` 97

reverse 5`-ACCTCGGTCACCCTTGGA-3`

Psap forward 5`-GTGGACCAGTATTCCGAGGT-3` 120

reverse 5`-GGGACCAGAGTCTTCATTGG-3`

Prohibitin forward 5`-AAGGGACTCATTTCCTCATCC-3` 125

reverse 5`-CGCAGTGTGATATTGACGTTCT-3`

Riken cDNA forward 5`-GATCCCCAAAGTGGAGTGG-3` 68

0610038D11
reverse 5`-GGTGCCCTCCAGTACCATCAA-3`

Riken cDNA forward 5`-AGAGACCACTTCCCAAGCAC-3` 87

1500004A08
reverse 5`-ACCTGTGCCTCTTTGTCACC-3`

Riken cDNA forward 5`-CATCAGTGGACCAGCATCAC-3` 85

2700059D21
reverse 5`-TCATTCCAACCTGCCTTAGA-3`
13
 CIRCULATIONAHA/2007/696468R2 Frost et al.

Riken cDNA forward 5`-TGGAGGACGCGATACTGG-3` 129

4930534E15
reverse 5`-CCCGAGAGGACATTGCTG-3`

Riken cDNA forward 5`-AGACTGTATCGGTGCCTGCT-3` 105

9130403P13
reverse 5`-AGCAGTTGGGCTGTCTGTCT-3`

Riken cDNA forward 5`-CAGTCATCGCCGTAGTGTTG-3` 102

9130213B05
reverse 5`-GGACCCGTAGTCATGGTCGT-3`

Riken cDNA forward 5`-TGTTACCTATCACAACAAAAGGAA-3` 99

9530068E07
reverse 5`-CCAAACGATGGTACTCCACA-3`

TAPbp forward 5`-ACAAGGCCCCCAGAGTGT-3` 100

reverse 5`-GGAAGAAGTGGGATGCAAGA-3`

Tde1l forward 5`-TTCTTGCTCGTCGGAGTATGT-3` 111

reverse 5`-CAGGGAACAACACCTTTCTCA-3`

Tsfm forward 5`-GACAGGGAAGGCTCTCTCAA-3` 117

reverse 5`-GAGCCGACATAGAACCCAGA-3`

14
 CIRCULATIONAHA/2007/696468R2 Frost et al.

Tab. S2:

function/process GeneID SignaleP TMHMM SecretomeP

(TMD) (score;
odds)
proteins with secretion signal,

without transmembrane domain

cystatin C protease inhibitor 13010 SS 0 W 0.95;7.1

apolipoprotein E lipid transport 11816 SS 0 W 0.92;6.5

nephroblastoma overexpressed gene growth factor activity 18133 SS 0 W 0.88;5.1

compl. component 1, q subcomp., complement activation 12262 SS 0 W 0.86;5.0

gamma polypeptide

follistatin-like 3 activin-inhibitor 83554 SS 0 W 0.86;4.7

lipoprotein lipase lipid metabolism 16956 SS 0 W 0.81;3.4

nephrin 1 role in glomerular 170643 SS 0 W 0.77;2.7

permeability

protease inhibitor 16 peptidase activity (IAE) 74116 SS 0 W 0.76;2.6

prosaposin sphingolipid 19156 SS 0 W 0.73;2.3

metabolism

glycine cleavage system protein H glycine catabolism 68133 SS 0 + 0.72;2.5

(IEA)

leucine-rich alpha-2-glycoprotein granulocytic 76905 SS 0 W 0.71;2.3

differentiation

CD14 antigen immune response 12475 SS 0 W 0.71;2.1

mannosidase, beta A glycoprotein 110173 SS 0 W 0.70;2.1

catabolism

melanocyte proliferating gene 1 unknown 60315 SS 0 W 0.66;1.9

Ts translation elongation factor, translation elongation 66399 SS 0 + 0.65;1.8

mitochondrial activity (IEA)

chemokine (C-X-C motif) ligand 14 chemokine, immune 57266 SS 0 W 0.59;1.6

response

15
 CIRCULATIONAHA/2007/696468R2 Frost et al.

cytochrome c oxidase, subunit VI a, electron transport (IEA) 12862 SS (HMM) 0 W 0.81;3.8

polypeptide 2

RIKEN cDNA 2700059D21 gene unknown 433693 SS (HMM) 0 + 0.76;2.7

cytochrome c oxidase, subunit Vb electron transport (IEA) 12859 SS (HMM) 0 + 0.73;3.6

ATP synthase, F1 complex, alpha hydrogen ion 11946 SS (HMM) 0 - 0.56;1.3

subunit, isoform 1 transporter (IEA)

ADP-ribosyltransferase 1 transferase activity 11870 SS 0 - 0.55;1.2

prohibitin DNA replication 18673 SS (HMM) 0 - 0.45;0.9

transforming growth factor beta 1 induced transcription factor 21807 SS (HMM) 0 - 0.29;0.6
transcript 4 activity (IEA)

protocadherin alpha 6 unknown 12937 SS 0 - 0.19;0.4

cysteine rich protein 2 cell proliferation, 68337 - 0 + 0.81;3.5

hemopoiesis

TGF-beta1-induced anti-apoptotic factor 1 apoptosis (IEA) 21842 - 0 + 0.78;3.2

RIKEN cDNA 0610038D11 gene unknown 67674 - 0 + 0.78;3.1

angiomotin-like 1 unknown (at tight- 75723 - 0 + 0.68;2.0

junction)

properdin factor, complement alternative complement 18636 - 0 + 0.64;1.6

activation (IEA)

ribosomal protein L36 protein biosynthesis 54217 - 0 + 0.61;1.7

(IEA)

proteins with secretion signal and one

transmembrane domain

UDP-N-acetyl-alpha-D-galactosamine transferase activity 108760 SS (HMM) 1 + 0.86;4.5

(IEA)

natriuretic peptide precursor type A hormone 24602 SS 1 W 0.76;3.2

similar to Atrial natriuretic factor precursor hormone 230899 SS 1 W 0.76;3.3

UDP-Gal:betaGlcNAc beta 1,4- carbohydrate 56375 SS 1 W 0.66;2.0

galactosyltransferase, polypeptide 4 metabolism (IEA)

immediate early response 3 unknown 15937 SS 1 + 0.65;1.7

Neuropilin angiogenesis, heart 18186 SS 1 W 0.60;1.4

development

16
 CIRCULATIONAHA/2007/696468R2 Frost et al.

RIKEN cDNA 9530068E07 gene unknown 213673 SS 1 W 0.60;1.5

CD164 antigen cell adhesion 53599 SS 1 - 0.47;1.2

TAP binding protein defense response 21356 SS 1 - 0.40;0.8

colony stimulating factor 2 receptor, alpha receptor activity 12982 SS 1 - 0.36;0.7

RIKEN cDNA 1500004A08 gene unknown (Kringle 216505 SS 1 - 0.34;0.6

domain)

Phosphatidylinositol Glycan, class T GPI anchor 78928 SS 1 - 0.32;0.7

biosynthesis (IEA)

MANSC domain containing 1 unknown 67729 SS 1 - 0.27;0.5

RIKEN cDNA 9130213B05 gene unknown 231440 SS 1 - 0.31;0.6

interleukin 6 signal transducer receptor of the notch 16195 SS 1 - 0.14;0.3

signalling pathway

proteins with secretion signal and

mutiple transmembrane domains

RIKEN cDNA 1110014C03 gene protein transport (IEA), 68581 SS 2 W 0.88;5.3

golgi

NADH dehydrogenase 3, mitochondrial oxidoreductase activity 17718 SS 3 W 0.82;3.9

(IEA)

dolichyl-di-phosphooligosaccharide- glycotransferase 13200 SS 2 W 0.64;1.8

protein glycotransferase activity

tumor differentially expressed 2 unknown 56442 SS 11 W 0.61;1.7

RIKEN cDNA 5630401J11 gene unknown 106489 SA 3 W 0.72;3.0

transmembrane protein 41B hydrolase activity (IEA) 233724 SA 6 - 0.53;1.9

RIKEN cDNA 0610009E20 gene unknown 66048 SA 2 - 0.37;1.0

TM2 domain containing 1 GPCR (apoptosis 94043 SS (HMM) 2 - 0.32;0.6

induction)

adiponectin receptor 1 lipid metabolism 72674 - 7 + 0.62;1.9

17