Báo cáo khoa học: Enhanced peptide secretion by gene disruption of CYM1, a novel protease in Saccharomyces cerevisiae doc

Enhanced peptide secretion by gene disruption of
CYM1
, a novel
protease in
Saccharomyces cerevisiae
Lars Jønson, Jens F. Rehfeld and Anders H. Johnsen
Department of Clinical Biochemistry, Rigshospitalet, Copenhagen, Denmark
Saccharomyces cerevisiae is a widely u sed host in the pro-
duction of therapeutic peptides and proteins. Here we report
the identification of a novel endoprotease in S. cerev isiae.
It is encoded by the CYM1 gene and is specific for the
C-terminus of basic residues of heterologously expressed
peptides. Gene disruption of CYM1 not only reduced the
intracellular proteolysis, but also enhanced the secretion of
heterologously expressed peptides such as growth hormone,
pro-B-type natriuretic peptide and pro-cholecystokinin.
Cym1p resembles metalloendoproteases of the pitrilysin
family with the HXXEH(X)E(71–77) catalytic domain as
seen in insulysin, nardilysin a nd human metalloprotease 1. It
is a nuclear encoded protease that l ocalizes to mitochondria
without a hydrophobic N-terminal signal sequence or a
C-terminal tail-anchor. The protease does not require post-
translational processing prior to activation and it contains
cytosolic activity that processes peptides designated for the
secretory pathway prio r to t ranslocation into the endo-
plasmic reticulum.
Keywords: cholecystokinin; growth hormone; metallopro-
tease; proBNP; yeast.
Saccharomyces cerevisiae is often used for industrial
production of recombinant p eptides [1]. Secretion of
heterologously expressed proteins i s obtained by expression

of fusion proteins in which t he protein o f interest is f used to
the S. cerevisiae a-factor prepropeptide to direct secretion
through the secretory pathway. To enrich the secretion of
full length peptides, the host strains used are o ften made
protease deficient by inactivation of the vacuolar protease
Pep4p and the glycosyl-phophatidylinositol (GPI)-anchored
aspartyl protease, Yps1p [2]. Yps1p has been shown to be
responsible for both Arg and Lys specific cleavage in the
maturation of heterologous peptides such as prosomato-
statin, human b-amyloid precursor protein and the elas tase-
specific inhibitor, pre-elafin [3–5]. Protease activity within
the s ecretory pathway i s d ominated by the serine proteases,
Kex2 and K ex1 and the aspartic protease, Yps1. H owever,
outside this pathway another group of proteases, the
metallopeptidase family, seems to dominate. Among these
are Ste24p a nd Axl1p involved in the maturation of the
a-mating factor [6,7] and the mitochondrial signal peptid-
ases, Mas1 and Mas2 [8].
We used human cholecystokinin (CCK) as a substrate to
identify additional proteases of budding yeast with Lys-
specific activity using a processing-independent assay [9].
Peptide hormones and neuropeptides are s ynthesized as
larger precursors that are p roteolytically cleaved i n t he
trans-Golgi network or in the immature secreto ry vesicles.
The excision of mature peptides f rom prohormones/pro-
proteins is a general event in eukaryotic organisms in w hich
members of t he subtilisin family of proteo lytic e nzymes are
responsible. Members of this family are Ca
2+
dependent

serine proteases. Kex2p from y east was the first member of
this family to be discovered and is essential i n the activation
of the a-factor pheromone. In mammals, members of the
prohormone/proprotein c onvertase family are responsible
for the proteolytic cleavage after basic residues [10,11].
CCK is a vertebrate neuroendocrine peptide that is
expressed in both the gut and the brain (reviewed in
[12,13]). The post- translat ional maturation o f proCCK
requires tyrosine sulfation, endoproteolytic cleavages,
exoproteolytic trimmings and carboxyterminal amidation.
The m ajor bioactive products are C CK-58, -33, -22, and
-8 [12], of which CCK-22 is a predominant circulating
form in mammals. Most of the CCK peptides are
synthesized after cleavage at a single Arg residue, but
CCK-22 requires processing after a single Lys residue. In
an earlier study of CCK expression in yeast we found
that Kex2p cleaves after the Arg residues A rg75 and
Arg-Arg(85–86) (residue numbers refer to t he prohor-
mone), releasing CCK-8-Gly. However, th e biosynthesis
of CCK-22 was unaltered in a kex2 strain as well as in
an yps1 yps2 mutant [14].
In the present study we have identified a novel e ndopro-
tease, Cym1p, a member of the pitrilysin family with
similarity to insulysin, which c leaves after the sin gle Lys
residue to produce C CK-22. Cym1p a cts o n heterologously
expressed proCCK reducing the amount of proCCK that
enters the secretory pathway. Gene disruption of CYM1
thereby resulted in an increase in t he amount of secr eted
proCCK. Interestingly, gene disruption of CYM1 also
resulted in increased secretion of other hormones tested

such as growth horm one and p ro-B-type natriuretic peptide
(proBNP).
Correspondence to A. H. Johnsen, Department of Clinical Biochem-
istry, KB3014, Rigshospitalet, DK-2100, Copenhagen, Denmark.
Fax: +45 3545 4640, Tel.: +45 3545 3007, E - mail:
Abbreviations:APP,b-amyloid precursor protein; CCK, cholecy-
stokinin; proBNP, pro-B-type natriuretic peptide; GH1, growth hor-
mone; ER, endoplasmatic reticulum; GFP, green fluorescent protein;
TACE, tumor necrosis factor-a converting enzyme.
(Received 1 1 August 2 004, revised 13 October 20 04,
accepted 18 October 2004)
Eur. J. Biochem. 271, 4788–4797 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04443.x
Throughout the paper, t he nomenclature for n umbering
of the amino acid residues refers to proCCK, of which the
release of Gly-extended and C-terminal extended CCK-22
requires cleavage a t th e C-terminus of Lys61 (equals Lys81
of preproCCK). Unless specified otherwise, CCK-22 is used
to describe the free N-terminus of both Gly-extended
CCK-22 as well as C-terminal extended CCK-22.
Materials and methods
Yeast strains and growth conditions
The background strain used in this study is the vacuole
protease deficient strain, B J2168 [15]. Additional yeast
strains used in the screening assay are listed in Table 1.
Construction of strains was carried out using the PCR-
based gene disruption technique [16]. Media were purchased
from Difco (Kansas City, MO, USA), amino acids and
other supplements from Sigma-Aldrich ( Copenhagen, Den-
mark). Yeast cells were grown at 30 °CinYPD(1%yeast
extract, 2% peptone and 2% dextrose) or synthetic

complete media based on yeast nitrogen base with ammo-
nium sulphate, succinic acid, NaOH and appropriate amino
acids. Tr ansformations with either linear DNA or plasmids
were performed using the modified lithium acetate proce-
dure as described [17]. Analysis of expressed peptides/
proteins was performed from yeast growing in the expo-
nential phase due to the consistency in CCK-22 b iosynthe-
sis. Secretion of proCCK, proBNP and growth hormone
(GH1) were analysed from the media of 25 A
600
units of cells
in 5 mL synthetic complete media which were inoculated
for 4 h. Cell growth was followed by the absorbance at
600 nm.
DNA extraction and amplification
Yeast genomic DNA was isolated as described [18]. PCR
was performed using either Pwo polymerase or the enzyme
cocktail based on Taq, Pwo and Pfu polymerase ( Expand
long range PCR kit, XL-PCR) both from Roche Diag-
nostics, Mannheim, Germany. All PCR products were
visualized by agarose gel electrophoresis and P CR products
were either purified from the gel using the gel-extraction kit
(Qiagen, Hilden, Germany) or from the reaction mixture by
PCR purification spin columns (GENOMED, Lo
¨
hne,
Germany). PCR-based, one step gene disruption was
performed using 50 ng of plasmid from the pRS400 series
[16] as template. Amplification of the marker was performed
with oligonucleotides having 20 nucleotides towards t he

plasmid and an additional 50 nucleotides flanking the target
gene. All other DNA manipulations were carried out
according to standard procedures [19].
Table 1. S. cerevisiae stra ins used i n this s tudy. Null mutants of putative metalloproteases a re named b y the ORF in the genotype.
Strain Genotype Source
BY4705 MATa ade2D::hisG his3D200 leu2D0 lys2D0 ura3D0 Euroscarf [16]
BJ2168 MATa prc1-407 prb1-1122 pep4-3 leu2 trp1 ura3-52 gal2 ATCC [15]
LJY122 MATa ape1::KANMX ape2::LYS2 his3D0 leu2D0 lys2D0 ura3D0 This study
LJY123 MATa ape1::KANMX ape2::LYS2 ape3::LEU2 his3D0 leu2D0 lys2D0 ura3D0 This study
LJY201 MATa prc1-407 prb1-1122 pep4-3 leu2 trp1 ura3-52 gal2 axl1::LEU2 This study
LJY202 MATa prc1-407 prb1-1122 pep4-3 leu2 trp1 ura3-52 gal2 ste24::LEU2 This study
LJY203 MATa prc1-407 prb1-1122 pep4-3 leu2 trp1 ura3-52 gal2 prd1:LEU2 This study
LJY204 MATa prc1-407 prb1-1122 pep4-3 leu2 trp1 ura3-52 gal2 yil108w::LEU2 This study
Y15298 MATa his3D1 leu2D0 lys2D0 ura3D0 ste23::KANMX Euroscarf [16]
Y11874 MATa his3D1 leu2D0 lys2D0 ura3D0 aap1::KANMX Euroscarf [16]
Y10148
a
MATa his3D1 leu2D0 lys2D0 ura3D0 afg3::KANMX Euroscarf [16]
Y14953 MATa his3D1 leu2D0 lys2D0 ura3D0 ape1::KANMX Euroscarf [16]
Y16224
a
MATa his3D1 leu2D0 lys2D0 ura3D0 rca1::KANMX Euroscarf [16]
Y14984
a
MATa his3D1 leu2D0 lys2D0 ura3D0 mip1::KANMX Euroscarf [16]
Y17144 MATa his3D1 leu2D0 lys2D0 ura3D0 yme1::KANMX Euroscarf [16]
Y13211 MATa his3D1 leu2D0 lys2D0 ura3D0 ybr074w::KANMX Euroscarf [16]
Y13801 MATa his3D1 leu2D0 lys2D0 ura3D0 ydl104c::KANMX Euroscarf [16]
Y11941 MATa his3D1 leu2D0 lys2D0 ura3D0 yhr113w::KANMX Euroscarf [16]
Y11960 MATa his3D1 leu2D0 lys2D0 ura3D0 yhr132c::KANMX Euroscarf [16]

Y12296 MATa his3D1 leu2D0 lys2D0 ura3D0 yil137c::KANMX Euroscarf [16]
Y15370 MATa his3D1 leu2D0 lys2D0 ura3D0 ynl045w::KANMX Euroscarf [16]
10864B MATa ura3-D851 leu2-D1 his3D200 lys2D202 ykr035c-ykr038c::URA3 Euroscarf [16]
Y11749 MATa his3D1 leu2D0 lys2D0 ura3D0 yol057w::KANMX Euroscarf [16]
Y16248 MATa his3D1 leu2D0 lys2D0 ura3D0 yol098c::KANMX Euroscarf [16]
10231B MATa his3D200 leu2D1 trp1D63 ura3-52 yol154w(4,744)::KANMX Euroscarf [16]
Y14266 MATa his3D1 leu2D0 lys2D0 ura3D0 ydr430c::KANMX Euroscarf [16]
LJY430 MATa prc1-407 prb1-1122 pep4-3 leu2 trp1 ura3-52 gal2 ydr430c::LEU2 This study
LJY431 MATa prc1-407 prb1-1122 pep4-3 leu2 trp1 ura3-52 gal2 yps1::TRP1 This study
LJY432 MATa prc1-407 prb1-1122 pep4-3 leu2 trp1 ura3-52 gal2 yps1::TRP1 ydr430c::LEU2 This study
a
Known mitochondrial proteases.
Ó FEBS 2004 Cym1p, a novel metalloprotease in S. cerevisiae (Eur. J. Biochem. 271) 4789
Plasmid constructs
Human proCCK, GH1 and proBNP were all expressed
as a fusion protein to the prepro leader sequence of yeast
a-mating factor (preproMfa1p). All constructs were
cloned in the EcoRI/ XbaI sites of pLJY15 ( Supplementary
material). The cDNA encoding preproGH1 was ampli-
fied from the IMAGE-clone, NIH MGC 179 (The
I.M.A.G.E. Consortium, Babraham, UK), whereas the
ORF encoding proBNP was amplified from first strand
cDNA synthesized from m RNA isolated from human
heart. The fusion constructs were expressed on multicopy
plasmids, with constitutive gene transcription from the
phosphoglycerate kinase promoter. CYM1-green fluores-
cent protein (GFP) constructs were cloned in p UG35; a
kind gift from J. H. Hegemann, Du
¨
sseldorf, Germany.

CYM1 expression was driven b y its own promoter on a
multicopy plasmid. Plasmid construction and oligonucleo-
tides used are listed in Supplementary material and Table
S1, respectively.
Enzymatic treatment
Trypsin treatment was performed using 1 mgÆmL
)1
trypsin (Worthington Biochemical Corporation, Lake-
wood, NJ, USA) in a 50 m
M
sodium phosphate buffer
(pH 7 .5) for 30 min at room temperature and terminated
by immersion into boiling water for 10 min. Carboxy-
peptidase B (Roche Diagnostics) treatment with a final
concentration o f 4 lgÆmL
)1
was p erformed in 0.1 m
M
sodium phosphate buffer ( pH 7.5) at room temperature
for 3 0 min. The reaction was terminated by immersion
into boiling water for 10 min.
Radioimmunoassay (RIA)
Two different antisera were used to determine the amount
of processed CCK. Ab 89009 [ 9] is specific for the
N-terminus of CCK-22 and Ab 7270 [20] is specific for
Gly-extended C CK. The fraction of CCK processed to
CCK-22 is calculated by division of the immunoreactivity
measured with Ab 89009, with the amount measured with
the same a ntibody after t he sample was treated with trypsin
to measure the total amount of N-terminal extended CCK-

22. The amount of proBNP was measured using an antibody
specific to the N-terminus of proBNP [21]. G rowth
hormone expression was analysed with t he AutoDELFIA
hGH fluoroimmonoassay (PerkinElmer Life Sciences,
Boston, MA, USA).
Yeast extract and protease assay
Ten A
600
units of yeast cells growing in the exponential
phase were sedimented by centrifugation at 3000 g fo r
5 min, washed once i n 2 5 m L H
2
Oandtransferredtoa
2 m L Eppendorf tube. An equal amount of acid washed
glass beads (Sigma-Aldrich) was added, followed by
200 lLof0.1
M
NaH
2
PO
4
(pH 4.5) including various
inhibitors (1 m
M
bestatin, 30 l
M
E-64, 10 l
M
leupeptin,
1 l

M
pepstatin A, 1 m
M
phenylmethylsulfonyl fluoride,
1m
M
EDTA and 1 m
M
1,10-orthophenanthroline or 1
tablet complete inhibitor with o r without EDTA per
2.5 mL 0.1
M
NaH
2
PO
4
(Boehringer M annheim). The cells
were broken by vortexing 3 · 20 s and the extracts were
clarified by centrifugation a t 15 000 g for 10 min. All step s
were carried out at 4 °C. The protease assay was
performed using 20 pmol synthetic amidated CCK-33
(Peninsula Laboratorie Europe, Exeter, UK) or
Ac-CCK-33-Gly (Cambridge Research Biochemicals,
Stockton, UK) as substrate, 20 lL yeast extract, various
inhibitors and activators in a total volume of 30 lL. The
mixture was incubated at 30 °C f or 2.5 min to 1 h and t he
reaction terminated by adding 500 lL VBA buffer ( 20 m
M
barbital buffer, 0.11% bovine serum albumin and 0.6 m
M

thiomersal) followed by i mmediate immersion into a
boiling water bath for 10 min.
Size exclusion chromatography
Yeast transformants grown to late exponential phase were
centrifuged at 15 000 g to collect the cells and 500 lLofthe
medium was loaded directly onto a Sephadex G-50
superfine (Amersham Bioscience, Uppsala, Sweden) column
(1 · 100 cm) at 4 °C. The sample w as eluted in VBA buffer
at a flow rate of 3.5 mLÆh
)1
and fractions were collec ted
every 17 m in. Calibrations were performed by including
125
I-labelled albumin ( V
0
)and
22
NaCl (V
t
). The elution
constants K
d
, of peaks eluting at V
e
are calculated a s K
d
¼
(V
e
) V

0
)/(V
t
) V
0
).
Fluorescence microscopy
The distribution of Cym1-GFP and Cym1D30-GFP were
analysed from cells grown i n synthetic complete media-
ura-Met. When mitochondria were to be stained, 50 n
M
Mitotracker (Molecular Probes, Eugene, OR, USA) was
added t o the medium of living cells for 30 min. Cells were
washed three times in water and a 3 lL a liquot placed on a
microscope slide under a cover slip and sealed with n ail
polish. Fluorescence microscopy in living cells was exam-
ined with a Zeiss LSM 510 confocal laser scanning
microscope.
Statistical analysis
Statistical calculations were performed using an unpaired
students t-test to analyse whether the change in peptide
expression between the background strain and the isogenic
cym1D0 can b e considered to be statistically significant.
P-values below 0.05 are considered significant.
Results
Intracellular Lys-specific activity of
S. cerevisiae
An in vitro assay was established using crude preparation s
from of BJ2168 (Table 1) and synthetic CCK-33 as substrate
(Fig. 1A). The nature of the protease performing the

cleavage of synthetic human CCK-33 to CCK-22 was
analysed by inclusion of a number o f inhibitors with the
extract from B J2168. Only the a ddition of a metal chelating
agent inhibited proteolysis of CCK-33 to CCK-22 (Fig. 1B).
The m etal dependency o f the protease was t ested in vitro,
after the activity was inhibited initially by addition o f 1 m
M
4790 L. Jønson et al.(Eur. J. Biochem. 271) Ó FEBS 2004
EDTA. Reconstitution of the activity leading to maturation
of CCK-22 was tested by addition of different d ivalent
cations in 0.2 m
M
surplus. Addition of Zn
2+
,Co
2+
and
Mn
2+
could re-establish the protease activity, whereas
Ca
2+
,Cu
2+
or Mg
2+
had no effect (Fig. 1C) in accordance
with the properties of known metalloproteases, which are
only activated by Zn
2+

,Co
2+
and Mn
2+
. Reactivation
using i ncreasing Zn
2+
concentration s howed a biphasic
pattern, with Zn
2+
acting in an inhibitory manner at
concentrations above 5 m
M
(data not shown).
CYM1
encodes a novel endoprotease
Computer analysis of the S. cerevisiae genome shows that it
contains 25 ORFs encoding metalloproteases. Strains
deficient in all t he enzymes ( except the two essential
enzymes, Mas1 and Mas2, were prepared. Cell extracts
were prepared from e ach and t ested in t he in vitro protease
assay against cell extracts of unmodified B J2168 (Fig. 2A).
In this assay Ac-CCK-33-Gly was used as substrate to
minimize the degradation by aminopeptidases, and 1 m
M
Mn
2+
and 1 m
M
bestatin were included to r educe the

aminopeptidase activity on the product, CCK-22. There
was no difference in processing activity between the two
background strains, BJ2168 and BY4705 (data not shown).
Deletion of CYM1 abolished t he protease activity (Table 2,
Fig. 2B), whereas none of the other metalloprotease defi-
cient s trains showed a significant change in th e biosynthes is
of CCK-22 (Table 2).
Expression of
CYM1
on a multicopy plasmid
To determine whether the amount of synthesized CCK-
22 correlates with the amount of Cym1p in vitro, Cym1p
was expressed on a multicopy plasmid and the fraction
of synthesized CCK-22 a nalysed over time. Cell extract
from BJ2168, BJ2168 cym1D0 transformed with pRS425
CYM1 and the control transformed with the empty
pRS425 vector were used in the in vitro proteas e assay
with 1 m
M
Mn
2+
and 1 m
M
bestatin in which the
reactions were terminated after 15, 30, 45 a nd 6 0 min.
The CCK-22 immunoreactivity was measured with Ab
89009 and the remaining Ac-CCK-33-Gly was m easured
with the same antibody after tryptic cleavage (Fig. 2C).
Expression of CYM1 on a multicopy plasmid enhanced
the rate o f CCK-22 p roduction several-fold. When the

same experiment was performed at pH 6.0 and pH 7.5,
A
B
C
Fig. 1. In vitro maturation of CCK-22 depends on metal-ions. (A)
Synthetic Gly-extended CCK-33 used in the in vitro protease assay.
ThebarsrepresentthetwoGly-extended CCK p roducts, CCK-22 a nd
-8 resu lting after endoproteolytic cleavage at the C-terminus of
monobasic residues. (Nomenclature discussed in Introduction.) (B)
In vitro p rote ase assay including different inhibitors. (C) Protease
reactivation by addition of 1.2 m
M
divalent metal ions to extracts
where the activity had been i nhibited with 1 m
M
EDTA. The fraction
of CCK-22 was calculated from the immunoreactivity using Ab 89009
divided by th e total amount of mature an d N -terminal exten ded
CCK-22 measured w ith Ab 89009 after trypsin t reatment. The data
represents mean ± S D of three independent experiments. PMSF,
phenylmethylsulfonyl fluoride.
AB C
Fig. 2. In vitro maturation of CCK-22 depends on Cym1p. Yeast cell
extract in which bestatin was included prior to the additio n of acet-
ylated CCK-33-Gly. The assay was terminated by boiling a t 2.5, 15, 30,
45 and 60 m in after the addition of the s ubstrate. (A) Cell extract from
BJ2168 transformed w ith t he e mpty vector, pRS425. ( B) Cell extract of
BJ2168 cym1D0 transformed with the empty vector, pRS425. (C) Cell
extract made from BJ2168 cym1D0 transformed w ith pRS425 CYM1
where transcription of CYM1 is un de r contro l of its own promoter and

terminator (note that the sca le is larger). The fraction of CCK-22 was
calculated from the immunoreactivity u sing Ab 89009 divided b y the
total a mount of mature and N-terminal extended CCK-22 measured
with Ab 89009 after trypsin treatment. The data represents
mean ± S D of t hree in dependent experiments.
Ó FEBS 2004 Cym1p, a novel metalloprotease in S. cerevisiae (Eur. J. Biochem. 271) 4791
there was an increased degradation and a fter 30 min
incubation the CCK immunoreactivity was undetectable
at pH 6.0 (data not shown). These results s how the Lys-
specific cleavage in CCK-22 maturation in vitro is
dependent on the amount of Cym1p.
Enhanced CCK secretion in a
cym1D0
strain
To elucidate the role of Cym1p in t he biosynthesis of CCK-
22 in vivo, both the intrace llular and extracellular proCCK
and CCK-22 content was analysed from transformants of
BJ2168 and BJ2168 cym1D0. Deletion of CY M1 resulted in
an  40% increase in the total amount of proCCK within
the cells (Fig. 3 C) accompanied by a similar decrease in
CCK-22 (Fig. 3B). Also, the secreted amount of total CCK
in the cym1D0 strain in creased with more than 50%
(Fig. 3 C), but unlike t he fraction al decrease in intrac ellular
CCK-22 there was an increase in t he extracellular f ractions
of CCK-22 compared to a vacuole p rotease d eficient strain
(Fig. 3 D).
Cym1p only process proCCK at Lys61
To investigate whether the Lys61 of proCCK is the only
processing site within the expression c onstruct, we c on-
structed a proCCK mutant in which Lys61 was exchanged

with an Ala residue (K61A). The secreted forms of CCK
were subjected to size exclusion chromatography and the
Gly-extended CCK measured by RIA using Ab 7270.
The maturation o f CCK-22-Gly was abolished in t he
CCK(K61A) mutant which is s een by the disappearance of
thepeakoftheCCK-22peakelutingatK
d
0.8 and the
appearance corresponding to CCK-39 (Fig. 4A,B).
The i ntracellular amount of CCK and the CCK(K61A)
mutant in the vacuole protease deficient s train, BJ2168 and
the cym1 D0 strain were analys ed using Ab 7270 before and
after trypsin/carboxypeptidase B t reatment. In w ild-type
yeast there is an increase in the intracellular CCK immu-
noreactivity for the CCK(K61A) mutant compared to
expression of wild-type CCK (Fig. 4C,D). These results
indicate that approximately one-th ird of the wild-type CCK
is process ed a t Lys61 and subsequently degraded. The total
amount o f CCK(K61A) is the same in both wild-type yeast
and the cy m1 mutant showing that Lys61 is the only target
for Cym1p in proCCK. Substitution of the Lys residue with
an Arg showed the same processing pattern and total
amounts of CCK(K61A) c ompared t o the wild-type CCK,
showing that the specificity of Cym1p is not limited to the
processing at single Lys residues (data not shown). As
expected, the amount of Gly-extended CCK, which requires
the action of Kex2p and Kex1p in the Golgi apparatus is
unchanged in all four expression constructs as a result of the
limited number of these serine proteases present in the cell
(Fig. 4 C,D).

Table 2. Metalloproteases in Saccharomyces ce revisiae. Search performed in SwissProt Sequence Retrieval System (SRS; />Protease assay performed in two independent assays (A and B ) using extracts from the metalloprotease deficient strains. T he a mount o f CCK -22 is
measured with Ab 89009 and t he total a mount of CCK is m easured after tryptic cleavage w ith Ab 8 9009. ND, not determined.
Gene SwissProt ORF
CCK-22 Total CCK [nM] Fraction
A
1
B
1
A
2
B
2
A
1
/A
2
B
1
/B
2
AAP1 P37898 YHR047c 3.2 3.2 36 32 0.09 0.10
AFG3 P39925 YER017c 2.8 2.4 23 24 0.12 0.10
APE1
a
P14904 YKL103c 4.4 3.7 34 28 0.13 0.13
APE2
a
P32454 YKL157w 4.4 3.7 34 28 0.13 0.13
APE3
a

P37302 YBR286w 4.4 3.7 34 28 0.13 0.13
DPP3 Q08225 YOL057w 4.3 4.6 31 35 0.14 0.13
LTA4 Q10740 YNL045w 3.9 4.0 32 29 0.12 0.13
MIP1 P35999 YKL134c 3.4 3.3 34 33 0.10 0.10
PRD1 P25375 YCL057w 2.4 2.8 28 26 0.09 0.11
QRI7
b
P43122 YDL104c 2.8 2.9 23 24 0.12 0.12
RCA1 P40341 YMR089c 4.2 3.5 31 27 0.14 0.13
STE23 Q06010 YLR389c 2.6 2.3 25 20 0.10 0.12
STE24 P47154 YJR117w 3.4 2.8 19 17 0.18 0.16
YBS4
b
P38244 YBR074w 2.6 2.7 27 29 0.10 0.09
YHR3
b
P38821 YHR113w 2.5 2.4 26 26 0.10 0.09
YHT2
b
P38836 YHR132c 3.8 3.3 34 32 0.11 0.10
YIK8
b
P40483 YIL108w 5.7 5.2 43 39 0.13 0.13
YIN7
b
P40462 YIL137c 3.5 2.8 23 20 0.15 0.14
YK18
b
P36132 YKR038c 2.9 2.1 23 20 0.13 0.11
YME1 P32795 YPR024w 2.3 2.7 25 25 0.09 0.11

MAS2 P11914 YHR024c ND, Essential proteins
MAS1 P10507 YLR163c ND, Essential proteins
AXL1 P40851 YPR122w 3.5 3.4 31 30 0.11 0.11
CYM1
b
P32898 YDR430c 0.6 0.4 42 39 0.01 0.01
YOJ8
b
Q12496 YOL098c 3.6 3.8 35 34 0.10 0.12
a
These three genes were disrupted in one strain.
b
Putative metalloproteases.
4792 L. Jønson et al.(Eur. J. Biochem. 271) Ó FEBS 2004
Cym1p localizes to the mitochondria without the
hypothetical signal peptide
Cym1p has been shown previously, by usage of GFP
tagging, to l ocalize to the mitochondria [22,23]. I n a
study using mass spectrometry on highly pure yeast
mitochondria Cym1p was also found in this compart-
ment, however , it could not be distinguished w hether
Cym1p is a true mitochondrial protein or associated with
the mitochondria [24]. Computer analysis using the
MITOP
program predicts that Cym1p localizes to the
mitochondria via an N -terminal signal peptide [25]. To
investigate whether the intracellular localization of
Cym1p is due to residues 1–30 of the N-terminus, GFP
was fused to the C-terminus of Cym1p in which residues
1–30 of Cym1p were deleted and residue number 31

changed t o a Met [ Cym1p(D1–30)-GFP]. A s a control,
GFP was fused to the C-terminus of intact Cym1p.
Expression of the construct in BJ2168 cym1D0 shows
mitochondrial localization as seen by double staining with
Mitotracker (Fig. 5C). That the Cym1p-GFP fusion is
active as a protease was shown by an in vitro assay as
described a bove. The activity of the Cym1-GFP fusion
(Fig. 5 D) is comparable with the native Cym1p (Fig. 2A).
The Cym1p(D1-30)-GFP construct also localizes to the
mitochondria, but in an inactive form (Fig. 5H). The
mitochondria seem to aggregate i n the bud site of BJ2168
cym1D0 expressing the inactive Cym1p(D1-30)-GFP
(Fig. 5D–G). Unlike other mutants in w hich mitochondrial
fusion and fission have been interrupted, further investiga-
tions need to be carried out in order to determine wh ether
this deletion of CYM1 results in l oss o f m itochondrial
DNA.
Expression of other peptides
To investigate whether the effect of a cym1 deletion on
peptide expression is a general p henomenon we cloned and
expressed three peptides of different length, requiring
different post-translational modifications and which contain
a number o f basic residues. Expression of the precursor of
brain n atriuretic peptide (proBNP) and growth hormone
(GH1) w ere all performed as f usions to the preprosequence
of the a-mating factor. Both constr ucts as well as proCCK
were expressed in yps1D0 mutant as well as the yps1D0
cym1D0 mutant. Analysis of the secreted peptides by mass
spectrometry showed that all three peptides are secreted.
However, we were unable t o detect intact GH1 and

proBNP, but were able to identify peptide fragments of
the two precursors (data n ot shown). T he amounts of
secreted proCCK and proBNP were quantified by RIA
after tryptic cleavage. GH1 content were measured by
A
BD
C
Fig. 4. Intracellular degradation of CCK d epends on Cym1p c leavage to
CCK-22. Expression of wild-type CCK and t he CCK(K61A) mutant
in BJ2168 and the cym1D0 mutant. (A) Gly-extended CCK forms
secreted to the media from BJ 2168 transfor med w ith wild-type CCK
and (B) G ly-extended CCK form s secreted f rom BJ2168 transformed
with the CCK(K61A) mutant. One millilitre of media w as subjected t o
size e xclusion chromatography. The immunoreactivity was measured
with Ab 7270 , w hich is specific for G ly-extended f orms of CCK. (C)
Intracellular amounts of Gly-extended CCK and (D) amou nts of t he
Gly-extended CCK(K61A) mutant. White bars show the amount of
Gly-extended CCK and grey bars the total amount of CCK measured
after trypsin/carboxypeptidase B treatment. The data in (C) and (D )
are given a s mean ± S D (n ¼ 3).
A
BD
C
Fig. 3. Effect of CYM1 deletion on proCCK secretion and CCK-22
maturation. BJ2168 and cym1D0 transformed with the proCCK
expression construct were harvested during exponential phase and the
media collected. (A) The intracellular and (C) extracellular amount of
total CCK was measured with A b 89009 after trypsin treatment. (B)
Fraction of intracellular and (D) s ecreted CCK-22 w as calculated as
the immunoreactivity measured with Ab 89009 before tryptic cleavage

divided with the total amount of CCK measured in (A) and (C),
respectively. The data are given as mean ± SD ( n ¼ 4). Statistics w ere
performed using unpaired t-test (***P<0.001 and **P<0.01).
Ó FEBS 2004 Cym1p, a novel metalloprotease in S. cerevisiae (Eur. J. Biochem. 271) 4793
immunoassay directly from the media. T he secretion of
proCCK, proBNP and GH1 were significantly enhanced
when expressed i n the yps1 cym 1 double m utant c ompared
to th e yps1 background (Fig. 6). The increases in immuno-
reactivity in the double mutant w ere 130% for GH1, 1 38%
for proBNP and 45% for proCCK.
Discussion
The type of protease responsible for the intracellular
maturation of CCK-22 was investigated in an in vitro
protease assay using a c rude extract of S. cerevisiae to
analyse the processing of synthetic human CCK-33 to
CCK-22 in the presence of different inhibitors. By
avoiding detergents in the extraction of protease activity,
we were able to exclude activity from Kex2p a s well as the
GPI-anchored yapsins [26–28]. Of the inhibitors tested,
the p roteolysis was inhibited only by EDTA a nd 1,10-
orthophenanthroline, and the activity could be restored by
addition of the divalent cations Zn
2+
,Co
2+
and Mn
2+
.
This suggests that a metalloprotease participates in the
maturation of CCK-22. Endoproteolytic cleavage after a

single Lys residue has previously been attributed to a
metalloprotease. T umor necrosis factor-a converting
enzyme (TACE) was initially identified as the metallopro-
tease shedding the extracellular domain of the cell-surface
protein, to release soluble tumor-necrosis factor-a [29].
A
B
C
DH
G
F
E
Fig. 5. Localization of Cym1p-GFP and the Cym1p(D1–30)-GFP
mutant. Confocal imaging of BJ2168 cy m1 D0 transformed with the
Cym1p-GFP construct (A–D) and the same construct lacking the first
30 amino acid residues Cym1p(D1–30)-GFP (E–H). The cells were
stained with Mitot racker as de scribed i n Mate rials and m ethods.
(A and E) Shows the fl uorescence of GFP in the two constructs,
(B and F ) the mitochondria labelled with Mitotracker and ( C a nd G) a
Nomarski picture of t he cells in w hi ch the GFP and Mitotracker
images have been included. (D and H) In v itro assay using cell extract
in which b estatin and Mn
2+
was included prior to the addition of
acetylated CCK-33-Gly. The assay was terminated 2.5, 15, 30, 45 and
60 min after the addition of the s ubstrate by immers ion i nto boiling
water for 10 min. The data a re given as mean ± SD (n ¼ 3).
A
B
C

Fig. 6. Peptide synthesis in an yps1 mutant and the yps 1 cym1 double
mutant. Cells w ere i nocu lated t o l ate e xpo nential ph ase w he re m edia
were collect ed for peptide analysis. ( A) To tal a m ount o f p roCCK in
the media measured with Ab 89009 after trypsin treatment. (B)
proBNP immunoreactivity me asur ed with an tibody specific to the
N-terminus of proBNP. (C) GH 1 content measured wit h Auto-
DELFIA hGH fluoroimmonoassay. Statistics we re performed using
unpaired t-test (***P<0.001 a nd **P<0.01).
4794 L. Jønson et al.(Eur. J. Biochem. 271) Ó FEBS 2004
Later TACE was identified as the enzyme responsible for
formation of the b-amyloid peptide associated with
Alzheimer’s disease by a-secretase specific cleavage of
human b-amyloid precursor protein (APP) [ 30]. TACE
has also b een shown to b e involved i n the
L
-selectin
shedding by the ectodomain due to processing after a
single Lys [31]. Another metalloprotease, ADAM10, has
been shown to contain a-secretase activity that cleaves
APP, where it participates in the constitutive as well as the
regulated s ecretion of t he N-terminal fragment of
APP [32]. H owever, after introduction into yeas t, APP
undergoes a-secretase-type cleavage, but Yps1p and
Yps2p were identified as the active enzymes [5]. Our data
indicates that yeast contain a metalloprotease, which
could be a novel a-secretase, in addition to the two
aspartic proteases, Yps1 and Yps2.
None of the yeast metalloproteases contains an obvious
signal peptide to direct the protein into endoplasmic
reticulum (ER). Therefore, we investigated strains deficient

in each of the 23 putative metalloproteases. The in vitro
protease assay i dentified Cym1p a s an endoprotease
responsible for the post-Lys cleavage of CCK-33. T hat
Cym1p can cleave at Ly s61 in proCCK was v erified by
overexpression studies, showing a several-fold increase in
enzyme activity.
The open reading frame of CYM1 encodes a 989
amino acid residue protein with no obvious signal
peptide a s predicted by the
SIGNALP
program [ 33]. It
has previously been shown t o localize w ith the mito-
chondria [22–24] and contain a hypothetical presequence
directing it i nto this organelle based on predictions using
the
MITOP
database [25]. T he protein shows similarity to
mammalian members of the pitrilysin family, namely the
human metalloprotease 1 [34], insulysin [35] and nardi-
lysin [36], to two additional proteases f rom S . cerevisiae,
Axl1p [6] and the putative protein encoded by YOL098c,
and a s well to the bacterial pitrilysin [ 37]. Both hMP1
andnardilysinhavebeenshowntoprocessatthe
N-terminal side of Arg [ 34,38], whereas cleavage
C-terminal to basic residues has no t b een re ported.
Cym1p contains the HXXEH(X)E(71–77) motif at the
N-terminus, which is conserved among members of the
pitrilysin family, where the two His residues and the Glu
located 76 residues further downstream are crucial for
metal b inding [39]. Insulin-degrading enzyme can b e

inactivated by EDTA and 1,10-orthophenanthroline and
be reactivated by Zn
2+
,Mn
2+
,Co
2+
and Ca
2+
,where
Zn
2+
is the major endogenous associate. The Zn
2+
reactivation curve shows a biphasic pattern with low
concentrations activating the enzyme and higher concen-
trations inhibiting [ 40,41]. These c haracteristics are c on-
served for Cym1p with the exception that Ca
2+
is not a
reactivator. We could not distinguish Zn
2+
from Mn
2+
as activator of Cym1p b ecause they displayed similar
potency in the reactivation of cleavage of CCK-33 to
CCK-22 (data not shown).
The mitochondrial l ocalization o f t he Cym1-GFP fusion
protein was expected based on previous reports [22–24].
Cym1p is unlike the matrix-targeted preproteins with

cleavable presequence also directed to the mitochondria.
Notably the Cym1D30-GFP also locate to the mitochondria
but in an inactive form. Other proteins located at the
mitochondrial outer membrane are devoid of a typical N-
terminal presequen ce and instead the targeting sequence is
contained in the protein sequence itself. One c lass of proteins
are attached to t he mitochodrial outer membrane through a
single transmembrane s egment at their N-terminus as seen
for Tom20, Tom70 and OM45 [42] and the yeast deubiquit-
inating enzyme Ubp16, a metalloprotease belonging to the
large f amily of deubiquitinating enzymes, mainly consisting
of cysteine proteases [43]. Another class of proteins are the
tail-anchored proteins, which include the translocases of the
mitochondrial outer membrane, Tom5, Tom6 and Tom22
[44]. Cym1p is neither translocated into the mitochondria
through a typical N-terminal presequence or attached
through an N -terminal transmembrane se gment nor th rough
a C -terminal a nchor. Thus, deletion o f the first 30 amino a cid
residues of the N-terminus and a C-terminal GFP fusion still
localize to the mitochondria, although in a n inactive form. It
suggests that Cym1p contains an internal signal, which is
responsible for the mitochondrial location. Mitochondria in
wild-type y east show tubular str uctures around the cell
cortex. By contrast the mitochondria of cym1 D0 cells app ear
as aggregated masses in the bud-sites. The mitochondrial
phenotype s een in cym1D0 cells as well with the inactive
Cym1D30-GFP m utant is s imilar to t he observations seen
with the r homboid-type protease Pcp1 mutant, a serine
protease that p rocess c ytochrome c peroxidase and a
dinamin-like GTPase, and is required for maintenance of

mitochondrial morphology and DNA [45,46]. Whether
disruption of CYM1 influences the ability to maintain
mitochondrial DNA has not yet been investigated.
Intracellular synthesis of CCK-22 was decreased in a
cym1D0 strain accompanied by a n increased concentra-
tion of proCCK. In contrast, the fraction of extracellular
CCK-22 was increased compared to wild-type yeast with
a parallel increase in the total CCK concentration. These
findings are in accordance with a cytosolic activity of
Cym1p like m ost insulin-degrading e nzymes [ 47]. The
prepro-a-mating factor enters the ER post-translationally,
suggesting that the fusion constructs could be a target for
Cym1p prior to translocation. T hus, the pretransloca-
tional degradation of proCCK will be decreased by
CYM1 disruption and the total productio n increased.
Thesefindingssuggestthatitiseithertheactivityfrom
newly synthesized Cym1p which is responsible for the
Lys-specific processing of proCCK, or that the protease
is attached to the outer membrane of the mitochondria
with the N-terminus exposed to the cytosol.
The precursors of GH1 and proBNP contain potential
processing sites f or Cym1p. Expression of these p roteins as
well as proCCK in the yps1 cym1 double mutant showed an
increase in the a mount of s ecreted protein, indic ating that
Cym1p processes a ll of these proteins prior to translocation
into the ER.
In conclusion, we have identified a novel metalloendo-
protease, Cym1p of S. cerevisiae, w hich is the yeast
homologue to insulin-degrading enzyme. It contains Lys-
specific proteolytic activity and is a key player in the

secretion of h eterologously expressed proteins and peptides
in yeast as C -terminal fusions to the p reprosequence of t he
a-mating factor. Removal o f Cym1p f rom S. cerevisiae will
increase the yield of useful proteins and peptides from yeast.
Whether the enhanced peptide secretion persists when the
Ó FEBS 2004 Cym1p, a novel metalloprotease in S. cerevisiae (Eur. J. Biochem. 271) 4795
system is transferred f rom laboratory conditions to indus-
trial large scale needs to be further investigated.
Acknowledgements
This study was supported by The Danish Medical R esearch Council
andComlicA/S.MitchiEgel-Mitani, Knud Vad, Ivan Diers, Jakob R.
Winther, Jens P. Goetze and Jens R. Bundgaard have contributed with
valuable suggestions. We also t hank Lis S. N ielsen, Ria Pedersen,
Birgitte Bunke, Nina Ilsøe and Allan Kastrup for skilful technical
assistance.
References
1. Walsh, G. (2003) Biopharmaceutical benchmarks – 2003. Nat.
Biotechnol. 21 , 865–870.
2. Egel-Mitani, M., Andersen, A.S., Diers, I.I., Hach, M., Thim, L.,
Hastrup, S. & Vad, K. (2000) Yield improvement of heterologous
peptides expressed in yps1-disrupted Saccharomyces cerevisiae
strains. Enzyme Microb. Technol. 26, 6 71–677.
3. Bourbonnais,Y.,Germain,D.,Ash,J.&Thomas,D.Y.(1994)
Cleavage of prosom atostatins by the yeast Yap3 and K ex2
endoprotease. Biochimie 76, 226–233.
4. Bourbonnais, Y., Larouche, C. & Tremblay, G.M. (2000) Pro-
duction of full-length human pre-elafin, an elastase specific
inhibitor, from yeast requires the absence of a functional yapsin 1
(Yps1p) endoprotease. Pro tein Expr. Purif. 20, 485–491.
5. Komano, H., Seeger, M., Gandy, S., Wang, G.T., Krafft, G.A. &

Fuller, R.S. ( 1998) Involvement of cell surface glycosyl-
phosphatidylinositol-linked aspartyl proteases in a lpha-secretase-
type cleavage and ectodomain solubilization of human Alzheimer
beta-amyloid precursor protein in yeast. J. Biol. Chem. 273,
31648–31651.
6. Adames, N., B lundell, K., A shby, M.N. & Boone, C. ( 1995)
Role o f yeast insulin-degrading enzyme homologs in pro-
pheromone processing and bud site selection. Science 270, 464–
467.
7. Schmidt, W.K., Tam, A. & Michaelis, S. (2000) Reconstitution of
the Ste24p-dependent N -terminalproteolyticstepinyeasta-factor
biogenesis. J. Bio l . Chem. 275, 6 227–6233.
8. Yang, M.J., Geli, V., Opp lige r, W., Sud a, K., Jame s, P. & Schatz,
G. (1991) The M AS-encoded pr ocessing protease of yeast
mitochondria. Interaction of the purified enzyme with signal
peptides and a purified precursor protein. J. Biol. Chem. 266,
6416–6423.
9. Paloheimo, L.I. & R ehfeld, J.F. (1994) A processing-independent
assay for human procholecystokin in and its products. Clin. Chim.
Acta 229, 49–65.
10. Rockwell, N.C., K rysan, D.J., Komiyama, T. & Fuller, R.S.
(2002) P recu rso r p ro cessi ng b y kex2/fu rin protea ses. Ch em. Rev.
102, 4 525–4548.
11. Steiner, D.F. (1998) The proprotein convertases. Curr. Opin.
Chem. Biol. 2, 31–39.
12. Rehfeld, J.F. (1999) The cholecystokinin-gastrin family of
peptides and their receptors. Results Probl. Cell Differ. 26, 293–
321.
13. Liddle, R.A. ( 1997) Cholecystokinin c ells. Annu. Rev. Physiol. 59,
221–242.

14. Rourke, I.J., Johnsen, A.H., Din, N., Petersen, J.G. & Rehfeld,
J.F. (1997) Heterologous expression of human cholecystokinin
in Saccharomyces cerevisiae. Evidence for a lysine-specific endo-
peptidase i n t he yeast secretory pathway. J. Biol . C hem. 272, 9720–
9727.
15. Jones, E.W. (1991) Tackling the protease problem in
Saccharomyces cerevisiae. Methods Enzymol. 194 , 428–453.
16. Brachmann, C.B., Davies, A., Cost, G.J., Caputo, E., Li, J.,
Hieter, P. & B oeke, J. D. (1998) De signer d eletion strains derived
from Saccharomyces cerevisiae S28 8C: a useful s et of strains a nd
plasmids for P CR-med iated g en e disruption and other applica-
tions. Yeast 14 , 115–132.
17. Gietz, R.D., Schiestl, R.H., Willems, A.R. & Woods, R.A. (1995)
Studies on t he transformation of intac t yeast cells by the LiAc/
SS-DNA/PEG procedure. Yeast 11 , 355–360.
18. Philippsen, P., Stotz, A. & Scherf, C. (1991) DNA of
Saccharomyces cerevisiae. Methods E nzymol. 194, 169–182.
19. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular
Cloning: A Laboratory Manual, 2nd edn. Cold Spring Harbor
Laboratory, Cold Spring Harbo r, NY.
20. Hilsted, L. & Rehfeld, J.F. (1986) Measurement of precursors for
alpha-amidated hormones by radioimmunoassay of glycine-
extended peptides after trypsin-carboxypeptidase B cleavage.
Anal. Biochem. 152, 119–126.
21. Goetze, J.P., Kastrup, J., Pedersen, F. & Rehfeld, J.F. (2002)
Quantification o f pro-B-typ e n atriure tic peptide and its p rodu cts
in human plasma by use o f an a nalysis independent of prec ursor
processing. Clin. C hem. 48, 1035–1042.
22. Huh, W.K., Falvo, J .V., Gerke, L.C., Carroll, A.S., Howson,
R.W., Weissman, J.S. & O’Shea, E .K. (2003) Global analy sis of

protein localization in budd ing yeast. Nature 425 , 686–691.
23. Kumar, A., Agarwal, S., Heyman, J.A., Matson, S., Heidtman,
M., Piccirillo, S., Umansky, L., Drawid, A., Jansen, R., L iu, Y .,
Cheung, K.H., Miller, P., G erstein, M., R oeder, G.S. & Snyder,
M. (2002) Subcellular localization of the yeast proteome. Genes
Dev. 16 , 707–719.
24. Sickmann, A ., Reinders, J., Wagner, Y., J oppich, C., Zahedi, R.,
Meyer,H.E.,Schonfisch,B.,Perschil,I.,Chacinska,A.,Guiard,
B., Rehling, P., Pfanner, N. & Meisinger, C. (2003) The proteome
of Saccharomyces cerevisiae mitochondria. Proc. Natl Acad. Sci.
USA 100, 13207–13212.
25. Scharfe, C., Zaccaria, P., Hoertnagel, K., Jaksch, M., K lopstoc k,
T., Dembowski, M., Lill, R., P rokisch, H., G erbitz, K.D.,
Neupert, W., M ewes, H.W. & Meitinger, T. (2000) MITOP, the
mitochondrial proteome database: 2000 update. Nucleic Acids
Res. 28, 155–158.
26. Azaryan, A.V., Wong, M., Friedman, T.C., Cawley, N.X.,
Estivariz, F.E., Chen, H.C. & L oh, Y.P. ( 1993) Purification and
characterization of a paired basic residue-specific yeast aspartic
protease encoded b y the YAP3 g ene. Similarit y t o the m ammalian
pro-opiomelanocortin-conve rting enzyme. J. Biol. Chem. 268,
11968–11975.
27. Fuller, R.S., Brake, A. & Thorner, J. (1989) Yeast prohormone
processing enzyme (KEX2 gene prod uct) is a Ca2+-dep endent
serine protease. Pr oc. Natl A cad. Sci. USA 86, 1434–1438.
28. Komano, H ., Rockwell , N., Wang, G.T., Krafft, G.A. & Fuller,
R.S. (1999) Purification and characterization of the yeast glyco-
sylphosphatidylinositol-anchored, monobasic-specific aspartyl
protease yapsin 2 (Mkc7p). J. Biol. Chem. 274, 24431–24437.
29. Black, R.A., Rauch, C.T., Kozlosky, C.J., Peschon, J.J., S lack,

J.L., Wolfson, M.F., Castner, B.J., Stocking, K.L., Reddy, P.,
Srinivasan, S., Nelson, N., B oiani, N., S chooley, K.A., Gerhart,
M.,Davis,R.,Fitzner,J.N.,Johnson, R.S., Paxton, R.J., March,
C.J. & C erretti, D .P. (1997) A metalloproteinase disintegrin that
releases tumour-necrosis factor-alpha from c ells. Nature 385,
729–733.
30. Buxbaum, J.D., Liu, K.N., Luo, Y., Slack, J.L., Stocking, K.L.,
Peschon, J .J., J ohnson , R .S., Castner, B .J., C er retti, D .P . & Black,
R.A. (1998) Evidence that tumor necrosis factor alpha converting
enzyme is involved in regulated alpha-secretase cleavage of the
Alzheimer a myloid protein p recursor. J. Biol. Chem. 273, 27765–
27767.
4796 L. Jønson et al.(Eur. J. Biochem. 271) Ó FEBS 2004
31. Peschon, J.J., Slack, J.L., Reddy, P., Stocking, K.L., Sunnarborg,
S.W., Lee, D.C., Russell, W.E., Castner, B.J., Johnson, R.S.,
Fitzner, J.N., Boyce, R.W., Nelson, N ., Kozlosky, C.J., Wolfson,
M.F., Rauch, C .T., Cerretti, D.P., Paxton, R.J., March, C.J. &
Black, R.A. (1998) An essential role for ec todomain shedding in
mammalian development. Science 282, 1281–1284.
32. Lopez-Perez, E., Zhang, Y ., Frank, S .J., Creemers, J., Seidah, N .
& Checler, F . (2001) Constitutive alpha -secretase c leavage of the
beta-amyloid precursor protein in the furin-deficient LoVo cell
line: i nvolvement of the pro-hormone convertase 7 and the disin-
tegrin metalloprotease ADAM10. J. Neurochem. 76, 1 532–1539.
33. Nielsen, H., Engelbrecht, J., Brunak, S. & von Heijne, G. (1997)
Identification of prokaryotic and e ukaryotic signal peptides and
prediction of their cleavage sites. Protein Eng. 10 , 1–6.
34. Mzhavia, N., Berman, Y.L., Qian, Y., Yan, L . & De vi, L.A.
(1999) Cloning, expre ssion, and characterization of human
metalloprotease 1: a novel member of the pitrilysin family o f

metalloendoproteases. DNA Cell B iol. 18, 3 69–380.
35. Affholter, J.A., F ried, V.A. & R oth, R.A. ( 1988) Human i nsulin-
degrading enzyme shares structural and functional homologies
with E. coli protease III. Science 242, 1415–1418.
36. Pierotti, A.R., Prat, A., Chesneau, V., Gaudoux, F., Leseney,
A.M., Foulon, T. & Cohen, P. (1994) N-arginine dibasic con-
vertase, a metalloendopeptidase as a prototype of a class of pro-
cessing enzymes. Proc. Natl Acad. Sc i. USA 91, 6078–6082.
37. Finch, P.W., W ilson, R.E., Brown, K ., Hic kson, I.D. &
Emmerson, P.T. (1986) Complete nucleotide sequence of the
Escherichia coli ptr gene encoding protease III. Nucleic Acids Res.
14, 7695–7703.
38. Chesneau, V., Pierotti, A.R., Prat, A., Gaudoux, F., Foulon, T. &
Cohen, P. ( 1994) N- arginine dibasic convertase ( NRD c on-
vertase): a newc omer to th e family of processin g endopeptidases.
An overview. Biochimie 76, 234–240.
39. Becker, A.B. & Roth, R.A. (1993) Identification of glutamate-169
as the third zinc -binding residue in proteinase III, a member of
the family of insulin-degrading enzymes. B ioc hem . J. 292,137–
142.
40. Duckworth, W.C., Bennett, R.G. & Hamel, F.G. (1998) Insulin
degradation: progre ss and p otential. Endocr. Rev. 19, 608–624.
41. Ebrahim, A., Hamel, F.G., Bennett, R.G. & Duckworth, W.C.
(1991) Identification of the metal associated with the insulin
degrading enzyme. Biochem. Biophys. Res. Commun. 181, 1398–
1406.
42. Waizenegger, T., Stan, T., Neupert, W. & Rapaport, D. (2003)
Signal-anchor domains of proteins of the outer membrane of
mitochondria: structural and functional characteristics. J. Biol.
Chem. 278 , 42064–42071.

43. Kinner, A. & Kolling, R. (2003) The yeast deubiquitinating
enzyme Ubp16 is anchored to the outer mitochondrial membrane.
FEBS Lett. 549 , 135–140.
44. Egan,B.,Beilharz,T.,George,R.,Isenmann,S.,Gratzer,S.,
Wattenberg, B. & L ithgow, T. ( 1999) Targeting o f tail-anchored
proteins to yeast mitochondria in vivo. FEBS Lett. 451 , 243–248.
45. McQuibban, G .A., Saurya, S. & Freeman, M. (2003) Mitochon-
drial membrane remodelling r eg ulated by a conserved rhomboid
protease. Nature 423, 5 37–541.
46. Herlan,M.,Vogel,F.,Bornhovd,C.,Neupert,W.&Reichert,
A.S. (2003) Proc essing o f Mgm 1 by the rhomboid-type protea se
Pcp1 is requ ired for maintenance of mitochondrial morphology
and of m itochondrial DNA. J. Biol. Chem. 278, 27781–27788.
47. Bai, J.P., H ong, H.J., R othenberger, D .A., Wong, W.D. & Buls,
J.G. (1996) The prese nce of i nsulin-degrading en zyme in human
ileal a nd colonic m ucosal cells. J. Pharm. Pharmac ol. 48, 1180–
1184.
Supplementary material
The following material is available from http://www.
blackwellpublishing.com/products/journals/suppmat/EJB/
EJB4443/EJB4443sm.htm
Appendix S1. Further details for s train and plasmid
constructions.
Table S1. Oligonucleotides used.
Ó FEBS 2004 Cym1p, a novel metalloprotease in S. cerevisiae (Eur. J. Biochem. 271) 4797