ABB
Archives of Biochemistry and Biophysics 405 (2002) 38–43
www.academicpress.com
ATP synthase of yeast: structural insight into the different
inhibitory potencies of two regulatory peptides and identification of
a new potential regulator
Sangjin Hong and Peter L. Pedersen*
a
Department of Biological Chemistry, Johns Hopkins University School of Medicine, 725 N. Wolfe Street, Baltimore, MD 21205-2185, USA
Received 10 May 2002, and in revised form 14 June 2002
Abstract
Mitochondrial ATP synthases, the major producers of ATP in higher eukaryotic cells, are known to be regulated by a peptide
designated IF1 . In contrast, in yeast three such peptides have been identified, IF1 and STF1 , which inhibit the reverse ATPase reaction,
and STF2 , a modulator of the action of these inhibitors. Despite significant homology to IF1 , STF1 exhibits less than half (40%) its
inhibitory potency. The two-fold purpose of this bioinformatic study was to gain structural insight into the different inhibitory potencies
of IF1 and STF1 and to determine to what extent yeast are unique in employing multiple peptides to regulate the ATP synthase. Sequence and secondary structural analyses and comparison with the known structure of bovine IF1 predicted a dimeric structure for yeast
STF1 in which the C-terminal regions form a coiled-coil. Moreover, sequence comparisons showed that within this C-terminal region a
conserved acidic residue (Asp 59) in yeast IF1 is replaced by Asn in STF1 . In the known structure of bovine IF1 , predicted to be very
similar to that of yeast IF1 , the residue Glu 68 corresponding to Asp 59 participates in the formation of a four-residue conserved acidic
cluster in the middle of the coiled-coil in the C-terminal region. It is deduced here that this acidic cluster is likely to be important in the
regulation of IF1 ’s inhibitory capacity and that replacement of conserved Asp 59 by Asn in STF1 may reduce its potency. Although
other homologs to the inhibitors IF1 and STF1 were not found in searches of available eukaryotic genomes, including human, a new
homolog, named STF3 , with 65% identity to the modulator STF2 , was discovered within the yeast genome and identified to be expressed
by searching the yeast EST database. Thus, yeast appears unique in regulating the ATP synthase by involving multiple peptides (IF1 ,
STF1 , STF2 , and perhaps STF3 ). Ó 2002 Elsevier Science (USA). All rights reserved.
Keywords: ATP synthase (F0 F1 ); IF1 ; Peptide regulators; Yeast; STF2 ; STF3 ; Bioinformatics
Mitochondrial F0 F1 -ATP synthase is an intricate
multisubunit membrane protein complex that is composed of at least 16 different subunit types in mammals
and 20 in yeast [1,2]. Most of the subunits of ATP synthases form two ‘‘motor’’ units F1 and F0 that couple an
electrochemical gradient of protons to ATP synthesis
(Fig. 1A). However, distinct from the functional subunits,
mammalian ATP synthases have a small regulatory peptide ‘‘subunit’’ designated IF1 [3–9]. Significantly, mammalian IF1 inhibits the ATPase activity catalyzed by
mitochondrial ATP synthase under conditions (e.g., hypoxia) where the membrane potential collapses, giving
*
Corresponding author. Fax: +1-410-614-1944.
E-mail address: [email protected] (P.L. Pedersen).
rise to a low internal pH. Under these more acidic conditions, IF1 dimers form in which the two monomers are
arranged antiparallel to one another with the C-terminal
regions forming a coiled-coil. The inhibitory activity is
located in the N-terminal half of each monomer [residues
14–47 (bovine)] such that one IF1 dimer inhibits the ATPase activity of two ATP synthase molecules. In contrast,
the pH-dependent regulatory information that permits or
prevents inhibition resides in the remaining part of IF1
which includes in part the C-terminal region that participates in coiled-coil formation in the active dimer at low
pH (below 6.5). This region is believed to act as a pHsensitive switch that at more alkaline pH (above 6.5) facilitates the N-terminal region of individual IF1 molecules
to also dimerize, masking the inhibitory sites and forming
inactive tetramers.
0003-9861/02/$ - see front matter Ó 2002 Elsevier Science (USA). All rights reserved.
PII: S 0 0 0 3 - 9 8 6 1 ( 0 2 ) 0 0 3 0 3 - X
S. Hong, P.L. Pedersen / Archives of Biochemistry and Biophysics 405 (2002) 38–43
39
Fig. 1. (A) Simplified overview of the role of the ATP synthase located on the inner mitochondrial membrane in coupling an electrochemical proton
gradient, generated by the electron transport chain, to ATP synthesis. (B) Sequence alignment of IF1 and STF1 (see Materials and methods).
Conserved residues are in bold face. Full species names are Bovine, Bos taurus; C. elegans, Caenorhabditis elegans; Mouse, Mus musculus; Rat, Rattus
norvegicus; Pig, Sus scrofa; Human, Homo sapiens; Rice, Oriza sativa; Potato, Solanum tuberosum; S. pombe, Schizosaccharomyces pombe; P. jadinii,
Pichia jadinii; Yeast, Saccharomyces cerevisiae.
40
S. Hong, P.L. Pedersen / Archives of Biochemistry and Biophysics 405 (2002) 38–43
Interestingly, the ATP synthase regulatory system of
yeast appears from biochemical data obtained to date to
be much more complex than that found in mammalian
systems. Here, three regulatory peptides, two inhibitors,
IF1 and STF1 (both 9 kDa) and a modulator STF2
(15 kDa) [10], have been identified. IF1 and STF1 , when
added separately, bind to the F1 unit, inhibiting its ATP
hydrolyzing activity 100 and about 40%, respectively. In
contrast, STF2 has no capacity to inhibit the enzyme.
Rather, it facilitates simultaneous binding of IF1 and
STF1 to the F1 unit. Although the binding of the two inhibitory peptides is competitive, neither one is displaced
by the other in the absence of STF2 . From these observations, it has been suggested that STF2 participates in the
regulation of the yeast ATP synthase by modulating the
binding of its two inhibitory peptides [11].
Although the sequence of STF1 is very similar to that
of IF1 with 49% identity and 77% similarity (Fig. 1B),
we found it of interest that STF1 exhibits less than half
the inhibitory potency of IF1 [10]. To understand the
structural basis of this difference, the sequences of IF1
from various species were aligned, and the aligned sequences were compared with that of STF1 . Also, to
determine whether regulation of ATP synthases from
eukaryotic organisms other than yeasts also involves
multiple peptides, we scanned several other available
genomes using BLAST. The results obtained not only
add new insight into how all eukaryotic ATP synthases
may be regulated by IF1 , but also indicate that yeasts
are unique by involving peptides other than IF1 , among
which may be the newly discovered candidate (STF3 )
described here.
Materials and methods
Protein sequences exhibiting high homology to IF1
and to STF2 were obtained from the yeast genome database and the nonredundant (nr) protein database using PSI-BLAST database search [12,13]. The default
parameters were BLOSUM62 matrix, 0.005 as an Evalue threshold, and no low complexity filtering in the
query sequence. The homologous sequences were
aligned using ClustalW [14] and Macaw [15] followed by
manual adjustment. The sequence analysis for the prediction of secondary structure, transmembrane helices,
and structural motifs were performed using pSAAM
(http://www.life.uiuc.edu/crofts/pSAAM/PSAAM_guest),
Peptool [16], SOSUI [17], TMHMM [18], and Coiledcoil prediction server [19]. WebLab ViewerPro was used
for the construction of structural models and the representation of protein structures.
Results and discussion
Our attention focused first on determining whether
important differences exist between IF1 and STF1 of
yeasts that might explain why IF1 is the better inhibitor.
Such information could be of general value in understanding the mechanism of action of this class of ATPase inhibitors. For these reasons, we first aligned the
sequences of IF1 from various species including metazoans, plants, and fungi. This revealed that the highly
conserved residues of IF1 are located in three regions:
N-terminal (residues 14–28 in yeast), central (residues
Fig. 2. Three conserved regions of IF1 (pdb1gmj). Residues 1–19 and residues 80–84 are not shown in the structure.
S. Hong, P.L. Pedersen / Archives of Biochemistry and Biophysics 405 (2002) 38–43
34–41), and C-terminal (residues 49–60) (Figs. 1B and
2). Recent site-directed mutagenesis studies show that
the residues 17–44 are involved in the inhibitory activity
of yeast IF1 , and among these the residues 17–28 are
essential [20]. The crystal structure of bovine IF1 showed
that residues 32–44, which correspond to residues 27–39
in yeast sequences, are associated with formation of an
inactive tetramer and a higher oligomer [8]. Thus, the Nterminal and the central conserved regions identified
here (Fig. 1B) appear to be involved directly in IF1 ’s
capacity to inhibit the ATPase activity of the ATP
synthase, with the central conserved region being involved also in IF1 ’s self inactivation via tetramer formation. From careful examination of the alignments in
these regions, we derived no obvious clues that would
account for the markedly decreased inhibitory potency
of yeast STF1 relative to IF1 .
Next we examined the C-terminal conserved region
that is located in the middle of the region that forms a
coiled-coil in the active dimeric form of bovine IF1 (Fig.
2). Significantly, this region contains four highly conserved residues including two completely conserved
acidic residues (Glu 61 and Glu 68 in bovine IF1 or Asp
52 and Asp 59 in yeast IF1 ) that have not been alluded
to in previous studies of IF1 from various sources. Of
particular interest, is the finding that in STF1 , Asp 59 is
replaced with an asparagine residue (Asn), the only
major difference among the 17 relatively conserved
amino acids identified by the sequence alignments.
Therefore, it seems likely that this change is related to
the marked differences in the capacities of IF1 and STF1
of yeast to inhibit the ATPase activity of the ATP synthase. The alternative argument that differences in inhibitory capacities may result because of structural
differences between IF1 and STF1 seems unlikely as
secondary structure predictions for the two peptides
and their propensity for coiled-coil formation, together
with other predicted properties, reveal little difference
(Fig. 3).
To gain further insight into where the conserved
acidic residues may be located within the structure of
yeast IF1 , we focused on the coiled-coil regions within
the known structure of bovine IF1 (Fig. 4). This seemed
fully justified both because of the high homology between yeast IF1 and bovine IF1 and because yeast IF1
has been reported to inhibit bovine F1 -ATPase and vice
versa [21]. It will be noted that in the active dimer of
bovine IF1 , the two C-terminal conserved regions are
arranged antiparallel to form a coiled-coil, and the
conserved acidic residues Glu 61 (Glu 52 in yeast) and
Glu 68 (Asp 59 in yeast) are arranged closely together to
face the same side in the middle of the coiled-coil region,
forming a four-residue conserved acidic cluster (Fig. 4).
The arrangement and the distances between the side
chains of the conserved acidic residues appear significantly ordered. As the C-terminal region of IF1 is be-
41
Fig. 3. Prediction of secondary structures of IF1 and STF1 in yeast (see
Materials and methods). For the prediction of coiled-coil domain, used
is the algorithm of Lupas et al. [19] which is based on the relative
frequency of occurrence of amino acids at each position of the coiledcoil heptad repeat. In the hydrophobic moment plot, the algorithm of
Eisenberg et al. [27,28] was used with a-helical amphiphilicity
(d ¼ 100°). The sequence of STF1 predicts an a-helical-soluble protein
with a coiled-coil conformation at its C terminus. The structure of
STF1 is believed to be very similar to that of IF1 .
Fig. 4. The conserved acidic cluster (blue) in the coiled-coil region of a
dimer of bovine IF1 (pdb1hf9). The conserved acidic residues in the
cluster are arranged structured and oriented to the one side of the
dimeric IF1 .
42
S. Hong, P.L. Pedersen / Archives of Biochemistry and Biophysics 405 (2002) 38–43
lieved to operate as a pH-sensitive switch that controls
the capacity of the N-terminal region to bind to and
inhibit the ATPase activity of ATP synthase, we suggest
here that Asp 59 in yeast IF1 is part of this switch and
that when this acidic residue is replaced by an asparagine in STF1 it reduces the capacity of this IF1 homolog
to bind to and inhibit the ATPase activity of yeast ATP
synthase. This is a novel concept as almost all attention
to date to account for the pH sensitivity of IF1 preparations in inhibiting the ATPase activity of ATP synthases [5,9,22] has focused on several histidine residues
present in higher eukaryotes but not conserved in plants
or yeasts (Fig. 1B).
To meet the second objective of this study, i.e., to
determine to what extent yeasts are unique in employing
multiple peptides to regulate the ATPase activity of the
ATP synthase, a Blast search was conducted on other
available eukaryotic genomes (human, Drosophila melanogaster, and Arabidopsis thaliana). In all three cases,
no protein other than IF1 with significant homology to
STF1 and STF2 of yeast was found. However, a Blast
search of the yeast genome database identified an open
reading frame (YLR327C; NCBI Accession No.
NP_013431) homologous to STF2 and referred to here
as STF3 . It is an unassigned gene 261 nucleotides long
that encompasses nucleotides 783,386–783,126 of chromosome XII. STF3 is predicted to encode an 86-residue
protein with a molecular mass of 9834 Da and to be
about 65% identical and 84% similar to STF2 in its
amino acid sequence (Fig. 5A). The expression of the
STF3 in yeast was identified by searching the yeast EST
database (GenBank Accession No. T37299). The sequence analysis of STF3 predicts a soluble basic protein
similar to STF2 . The disruption of the STF3 studied by
genomewide phenotypic analysis in which each of the
functionally unknown open reading frames of yeast was
deleted systematically [23] also showed that the protein
is not essential for the viability of the cell like STF2 .
In summary, we have shown that within the C-terminal region of all known inhibitor peptides of mitochondrial ATP synthases, except yeast STF1 , there are
two highly conserved acidic residues that are likely involved in the pH regulation of these inhibitors in suppressing the ATPase activity of ATP synthases. In yeast
STF1 , we have shown that one of these two conserved
acid groups (Asp 59) is replaced with an asparagine
residue, which may account for its weaker capacity to
inhibit the ATP synthase. We have shown also that, in
addition to the three known peptide regulators
(IF1 ; STF1 , and STF2 ) of the yeast ATP synthase, a
fourth regulator (STF3 ) reported here for the first time
may be involved and that the use of multiple peptides to
inhibit the ATPase activity of the ATP synthase in yeast
may be unique to this organism and other fungi. Finally,
as IF1 has been reported to bind to the ATP synthase
near the C-terminal region of the b subunit [24] located
Fig. 5. (A) Sequence alignment of STF2 and STF3 . (B) Model of the regulation of ATP synthase in yeast (see Materials and methods). Two types of
inhibitors (IF1 and STF1 ) and two types of modulators (STF2 and STF3 ) regulate the ATP synthase. The site of action is believed to reside, at least in
part, within the C-terminal region of the b subunit [23]. The model for the ATP synthase was generated by molecular modeling using the lowresolution crystal structure of F0 F1 ATP synthase from yeast (1qo1) as a template. Each subunit of the template structure was replaced by superimposition with the coordinates from high-resolution crystal and NMR structures. Structures of some subunits of the ATP synthase whose
coordinates are not available were generated manually based on the secondary structure prediction and biochemical data [1].
S. Hong, P.L. Pedersen / Archives of Biochemistry and Biophysics 405 (2002) 38–43
near the bottom of the F1 unit [25,26], it seems likely
that the inhibitory events involving the four regulatory
peptides in yeast occur, at least in part, at this locus (Fig.
5B). Thus, these studies employing a bioinformatic approach open the door to a number of experimentally
testable hypotheses about the regulation of ATP synthases both in yeast and in other biological systems.
Acknowledgments
We thank Dr. Young Ko for helpful discussions and
the NIH (Grant CA 10951) for support.
References
[1] P.L. Pedersen, Y.H. Ko, S. Hong, J. Bioenerg. Biomembr. 32
(2000) 325–332.
[2] J. Velours, G. Arselin, J. Bioenerg. Biomembr. 32 (2000) 383–
390.
[3] M.E. Pullman, G.C. Monroy, J. Biol. Chem. 238 (1963) 3762–
3769.
[4] K. Schwerzmann, P.L. Pedersen, Arch. Biochem. Biophys. 250
(1986) 1–18.
[5] M.S. Lebowitz, P.L. Pedersen, Biochem. Biophys. 330 (1996) 342–
354.
[6] E. Cabez
on, P. Jonathan, G. Butler, M.J. Runswick, J.E. Walker,
J. Biol. Chem. 275 (2000) 28353–28355.
[7] E. Cabez
on, I. Arechaga, P. Jonathan, G. Butler, J.E. Walker, J.
Biol. Chem. 275 (2000) 28353–28355.
[8] E. Cabez
on, M.J. Runswick, A.G. Leslie, J.E. Walker, EMBO J.
20 (2001) 6990–6996.
[9] D.J. Gordon-Smith, R.J. Carbajo, J.-C. Yang, H. Videler, J.
Runswick, J.E. Walker, E. Neuhaus, J. Mol. Biol. 308 (2001) 325–
339.
43
[10] T. Hashimoto, Y. Yoshida, K. Tagawa, J. Bioenerg. Biomembr.
22 (1990) 27–38.
[11] Y. Yoshida, T. Sato, T. Hashimoto, N. Ichikawa, S. Nakai, H.
Yoshikawa, F. Imamoto, K. Tagawa, Eur. J. Biochem. 192 (1990)
49–53.
[12] S.F. Altschul, T.L. Madden, A.A. Schaffer, J. Zhang, Z. Zhang,
W. Miller, D.J. Lipman, Nucleic Acids Res. 25 (1997) 3389–3402.
[13] S.F. Altschul, E.V. Konini, Trends Biochem. Sci. 23 (1998) 444–
447.
[14] J.D. Thompson, Nucleic Acids Res. 22 (1994) 4673–4680.
[15] G.D. Schuler, S.F. Altschul, D.J. Lipman, Proteins 9 (1991) 180–
190.
[16] D.S. Wishart, P. Stothard, G.H. van Domselaar, Methods Mol.
Biol. 132 (2000) 93–113.
[17] T. Hirokawa, S. Boon-Chieng, S. Mitaku, Bioinformatics 14
(1998) 378–379.
[18] A. Krogh, B. Larsson, G. von Heijne, E.L.L. Sonnhammer, J.
Mol. Biol. 305 (2001) 567–580.
[19] A. Lupas, M. Van Dyke, J. Stock, Science 252 (1991) 1162–
1164.
[20] N. Ichikawa, A. Karaki, M. Kawabata, S. Ushida, M. Mizushima, T. Hashimoto, J. Biochem. 130 (2001) 687–693.
[21] G. Klein, M. Satre, P. Vignais, FEBS Lett. 84 (1977) 129–134.
[22] R. Schnizer, G. van Heeke, D. Amaturo, S.M. Schuster, Biochim.
Biophys. Acta 1292 (1996) 241–248.
[23] E.A. Winzeler, D.D. Shoemaker, A. Astromoff, H. Liang, K.
Anderson, B. Andre, R. Bangham, R. Benito, J.D. Boeke, H.
Bussey, A.M. Chu, C. Connelly, K. Davis, F. Dietrich, S.W. Dow,
M. El Bakkoury, F. Foury, S.H. Friend, E. Gentalen, G. Giaever,
J.H. Hegemann, T. Jones, M. Laub, H. Liao, R.W. Davis, et al.,
Science 285 (1999) 901–906.
[24] P.J. Jackson, D.A. Harris, FEBS Lett. 229 (1988) 224–228.
[25] J.P. Abrahams, A.G. Leslie, R. Lutter, J.E. Walker, Nature 370
(1994) 621–628.
[26] M.A. Bianchet, J. Hullihen, P.L. Pedersen, L.M. Amzel, Proc.
Natl. Acad. Sci. USA 95 (1998) 11065–11070.
[27] D. Eisenberg, R.M. Weiss, T.C. Terwilliger, Nature 299 (1982)
371–374.
[28] D. Eisenberg, Annu. Rev. Biochem. 53 (1984) 595–623.