Biochem. J. (2008) 413, 417–427 (Printed in Great Britain)
417
doi:10.1042/BJ20071641
Insights into the substrate specificity of plant peptide deformylase,
an essential enzyme with potential for the development of novel
biotechnology applications in agriculture
Lynnette M. A. DIRK*, Jack J. SCHMIDT†, Yiying CAI†, Jonathan C. BARNES‡, Katherine M. HANGER§, Nihar R. NAYAK*,
Mark A. WILLIAMS*, Robert B. GROSSMAN‡, Robert L. HOUTZ* and David W. RODGERS†1
*Plant Physiology/Biochemistry/Molecular Biology Program, Department of Horticulture, University of Kentucky, 441 Plant Science Building, Lexington, KY 40546-0312, U.S.A.,
†Department of Molecular and Cellular Biochemistry and Center for Structural Biology, Biomedical Biological Sciences Research Building, Rm 269, University of Kentucky, 741 South
Limestone, Lexington, KY 40536, U.S.A., ‡Department of Chemistry, University of Kentucky, 339 Chemistry-Physics Building, Lexington, KY 40506-0055, U.S.A., and §College of
Pharmacy, University of Kentucky, 725 Rose Street, Lexington, KY 40536, U.S.A.
The crystal structure of AtPDF1B [Arabidopsis thaliana PDF
(peptide deformylase) 1B; EC 3.5.1.88], a plant specific
deformylase, has been determined at a resolution of 2.4 Å
(1 Å = 0.1 nm). The overall fold of AtPDF1B is similar to other
peptide deformylases that have been reported. Evidence from
the crystal structure and gel filtration chromatography indicates
that AtPDF1B exists as a symmetric dimer. PDF1B is essential
in plants and has a preferred substrate specificity towards the
PS II (photosystem II) D1 polypeptide. Comparative analysis
of AtPDF1B, AtPDF1A, and the type 1B deformylase from
Escherichia coli, identifies a number of differences in substrate
binding subsites that might account for variations in sequence
preference. A model of the N-terminal five amino acids from
the D1 polypeptide bound in the active site of AtPDF1B suggests
an influence of Tyr178 as a structural determinant for polypeptide
substrate specificity through hydrogen bonding with Thr2 in
the D1 sequence. Kinetic analyses using a polypeptide mimic
of the D1 N-terminus was performed on AtPDF1B mutated at
Tyr178 to alanine, phenylalanine or arginine (equivalent residue
in AtPDF1A). The results suggest that, whereas Tyr178 can
influence catalytic activity, other residues contribute to the overall
preference for the D1 polypeptide.
INTRODUCTION
abundant chloroplast-encoded polypeptides, reveal a distinct
preference by PDF1B for the PSII (photosystem II) D1
polypeptide, which has the N-terminal sequence f-MTAIL (f-MetThr-Ala-Ile-Leu) [14]. Treatment of intact plants with actinonin,
a potent PDF inhibitor, results in a rapid reduction in PSII activity,
D1 processing and assembly into PSII, detectable within 2 h [15].
Overexpression of chloroplast-localized forms of either PDF1A
or PDF1B results in nearly complete resistance to actinonin, but
overexpression of mitochondrial-localized forms of PDF1A does
not convey resistance [12]. Given the widespread occurrence
of PDF1A and PDF1B in plants, the combined action of PDF
inhibitors and resistance conveyed by overexpression creates an
attractive combination of technologies for the control of plant
growth as well as tools for use as selectable markers in plant transformation vectors.
Over 300 natural and synthetic PDF inhibitors have been
described, most evaluated according to their potential as broadspectrum antibiotics and many synthesized according to the abundant and detailed structural information available for eubacterial
PDFs [16]. Crystal structures have been reported for a number
of eubacterial PDFs, including the PDF1B from Escherichia
coli, EcPDF1B [17–20], the eukaryote Plasmodium falciparum
type 1B deformylase, Pf PDF1B [21], and the plant type 1A
deformylase from Arabidopsis thaliana, AtPDF1A [4]. Detailed
structural information is not available for plant PDF1Bs, however,
PDF (peptide deformylase; EC 3.5.1.88) catalyses the removal
of the N-formyl group from the initiating f-Met at the Ntermini of nascent polypeptides synthesized in eubacteria [1],
chloroplasts [2,3] and mitochondria [4–7], although the latter is
controversial [8,9]. Most PDFs are monomeric hydrolases, and
all contain three signature sequence motifs associated with
cationic metal and substrate binding [10,11]. PDF activity is
indispensable in prokaryotes and photosynthetic organisms, and
has become a desirable target for the development of broadspectrum antibiotics, herbicides and selectable markers in plant
transformation vectors [12]. Human mitochondrial forms of PDF
have also been identified, and emerged as targets for the development of anticancer compounds with antiproliferative properties
[5]. In plants there are two different types of PDFs (PDF1A,
previously designated DEF1, and PDF1B, previously designated
DEF2), both localized to mitochondria and chloroplasts [2,3].
Insertional mutants in PDF1A do not have a visible phenotype,
whereas PDF1B mutants exhibit stunted growth, chlorophyll
bleaching and reduced seedling vigour [13]. Substantial differences exist between PDF1A and PDF1B in polypeptide substrate
specificity and catalytic efficiency in vitro [14]. In particular,
comparative analyses of polypeptide substrate specificity between
PDF1A and PDF1B (kcat /K m ) using peptide mimics of the most
Key words: Arabidopsis, actinonin, chloroplast, co-translational,
peptide deformylase, protein processing.
Abbreviations used: At PDF, Arabidopsis thaliana peptide deformylase; At PDF1Bet, engineered truncated At PDF1B; At PDF1Bt, truncated At PDF1B;
Ec PDF, Escherichia coli peptide deformylase; f-ML-ρ-NA, N -formyl-Met-Leu-ρ-nitroanilide; PDF, peptide deformylase; Pf PDF, Plasmodium falciparum
peptide deformylase; PSII, photosystem II; r.m.s.d., relative mean square deviation.
The atomic co-ordinates and structure factors (code 3CPM) for the plant peptide deformylase PDF1B crystal structure have been deposited in the
Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ, U.S.A. (http://www.rcsb.org).
1
To whom correspondence should be addressed (email [email protected]).
c The Authors Journal compilation c 2008 Biochemical Society
418
L. M. A. Dirk and others
and this information will be crucial for the development of
plant-specific inhibitors. Here we report on the structure of a plant
PDF1B and demonstrate the existence of unique changes relative
to other PDFs which may be responsible for the polypeptide
specificity of PDF1B as well as its essentiality in plants.
EXPERIMENTAL
Cloning and protein purification
A partial cDNA representing a predicted chloroplast length
protein was cloned into pET23b (Novagen) [2]. AtPDF1B represents a hexahistidyl-tagged protein purified from an induction
as previously described [2]. Production and purification of
AtPDF1Bet (engineered truncated AtPDF1B), which encodes
residues Lys65 to Arg264 of the full-length protein (with additional
Gly-His-Gly N-terminal amino acids due to cloning), was similar
to AtPDF1B [2] but the conditions for bacterial growth differed.
The Luria–Bertani broth medium was supplemented with 0.2 %
glucose and cultures were grown to an absorbance (600 nm) of 1.0
prior to a 1.0 mM IPTG (isopropyl β-D-thiogalactoside) induction
for 1 h at 37 ◦C.
Limited proteolysis
Proteolysis, using 2 mg · ml−1 of AtPDF1B, was performed with
1 % (w/w) trypsin [TPCK (tosylphenylalanylchloromethane)treated, Sigma] as a stock in 2 mM sulphuric acid for 1, 5, 10, 20,
40, 60 or 120 min at room temperature (22 ◦C) in 20 mM sodium
phosphate, pH 7.4, and 500 mM sodium chloride. Digestion was
terminated by addition of 2 μg of protein to SDS sample buffer
[62.5 mM Tris, pH 6.8, 2 % (w/v) SDS, 5 % (v/v) β-mercaptoethanol, 10 % (w/v) sucrose and 0.002 % (w/v) Bromophenol
Blue] and boiling for 3 min. This same procedure was used for
studying the inhibitor binding effects with the inclusion of a
3 min preincubation of AtPDF1B with 1 mM actinonin prior to
the addition of 1 % (w/w) trypsin.
Tryptic proteolysis was also performed on AtPDF1B as the enzyme was processing substrate. To ensure that catalytic conditions
were maintained for 30 min during limited trypsin proteolysis,
concentrations of AtPDF1B were reduced to 8.5 μg · ml−1 .
Digestion was terminated after 1, 5, 10, 20 or 30 min by freezing
with liquid nitrogen followed by freeze-drying. After resuspension in SDS sample buffer and boiling, the sample was desalted
using a spin column (Pierce), and the sample loaded on to SDS/
12.5 % PAGE after addition of ∼ 10 % (v/v) glycerol and boiling
again. The same trypsinolysis was performed with and without
substrate (127 mM f-Met-Ala-Ser).
In preparing truncated AtPDF1B for kinetic studies, proteolysis
was terminated after 20 min at ∼ 22 ◦C by incubation with
agarose-immobilized soybean trypsin inhibitor (Sigma) for
20 min with rocking at ∼ 22 ◦C. This truncated AtPDF1B,
AtPDF1Bt, which lost its hexahistidyl sequence, was subsequently
separated from undigested AtPDF1B by loading the digested
sample on to a HiTrap® affinity column and concentrating
the eluate. Undigested AtPDF1B remained bound to the
column. AtPDF1B (∼ 7.5 μg) prior to and after trypsinolysis
was electrophoresed in a precast Tris/HCl linear gradient gel
(8–16 %) to estimate molecular mass changes (results not
shown). The molecular mass of proteolysed AtPDF1B was also
determined with a Sephadex® G-75 column (0.9 · 28.5 cm, 40–
120 μ particle size) equilibrated with 100 mM Tris/HCl and
50 mM sodium chloride, pH 8. The column was eluted at a
flow rate of 0.4 ml · min−1 and ∼ 1 ml fractions were collected.
The column was calibrated using 100 μl loads (500–600 μg)
c The Authors Journal compilation c 2008 Biochemical Society
of bovine serum albumin (66.0 kDa), ovalbumin (43.0 kDa),
β-lactoglobulin (35.0 kDa) or lysozyme (14.3 kDa). Protein
concentrations were determined by Coomassie Plus – The Better
BradfordTM assay (Pierce). AtPDF1Bt (200 μg) was tracked by
activity with the f-ML-ρ-NA (N-formyl-Met-Leu-ρ-nitroanilide)
and Aeromonas proteolytica (Sigma) aminopeptidase-linked
assay [22]. The elution volume (V e ) was plotted versus the log of
the molecular mass and a linear regression generated in Sigma Plot
(version 10.0).
Immunological determination of possible trypsin-sensitive sites
A 25 μg aliquot of AtPDF1B was treated with 1 % (w/w) trypsin
and a 100-fold excess immobilized inhibitor prior to being mixed
with SDS sample buffer and boiled. The sample was loaded on
a SDS/12.5 % PAGE and electro-transferred to ImmobilonTM -P
membrane. This blot was then probed with a 1:2000 dilution
of anti-penta-His antibody (Qiagen, Valencia, CA, U.S.A.) and
1:1000 dilution of goat anti-mouse IgG alkaline phosphatase
conjugated secondary antibody, using washes and detection as
described previously [2].
Kinetic parameters for At PDF1B, proteolysed-, and engineered
truncated-At PDF1B
f-ML-ρ-NA substrate coupled with Aerom. proteolytica aminopeptidase was used to assay PDF activity spectrophotometrically
[22]. Undigested (7.5 μg; [2]), proteolysed (31.6 μg; AtPDF1Bt),
and AtPDF1Bet (7.5 μg) data were collected and fitted to
the Michaelis–Menten equation V = V max [S]/(K m + [S]) using
SigmaPlot (Version 8.0).
Crystallization, data collection and structure determination
Crystals of AtPDF1Bt were grown by hanging drop vapour
diffusion in 24 well plates with well solutions containing 15–18 %
polyethylene glycol monomethyl ether 550, 28–70 mM ZnSO4
and 70 mM Mes, pH 6.5. Protein solution at approx. 5 mg· ml−1
was mixed 1:1 with well solution to a final volume of 2–5 μl for
the crystallization drops. Crystals formed in several days to several
weeks. To prepare the crystals for data collection, they were briefly
placed into a solution containing the same components as the
crystallization well in addition to 20 % glycerol, mounted in
nylon loops, and flash cooled by plunging into liquid nitrogen
[23]. The space group of the AtPDF1Bt crystals is P41 21 2 with
unit cell dimensions of a = 50.9 Å (1 Å = 0.1 nm), b = 50.9 Å, c =
144.8 Å.
X-ray diffraction data were collected on a CCD (chargecoupled-device) detector at the Advanced Photon Source
beamline 22-ID (Southeast Regional Collaborative Access Team),
Argonne National Laboratory, Argonne, IL, U.S.A. Data were
processed with HKL2000 [24], and the structure was determined
by molecular replacement with CNS (Crystallography & NMR
System) [25] using the crystal structure of Pf PDF (Protein Data
Bank entry 1RQC) as a search object. Model building and analysis
were performed with the program O [26], and structures were
refined with CNS. Illustrations were made using the programs
PyMOL [W.L. DeLano, The PyMOL Molecular Graphics System
(2002), http://www.pymol.org] and GRASP [27].
Model of the At PDF1Bt complex with a pentapeptide
A pentapeptide with the sequence MTAIL (Met-Thr-Ala-IleLeu) was modelled into the AtPDF1Bt active site using the
co-ordinates of the Met-Ala-Ser peptide in complex with
Structure of plant peptide deformylase
419
the EcPDF (PDB entry 1BS8) as a guide. This model
was subjected to a molecular dynamics simulation procedure with AMBER8 [28]. The structure was first prepared with
an energy minimization procedure and then allowed to evolve in
a dynamics simulation with implicit solvent for 2 ns at a constant
temperature of 300 K and with a time step of 0.001 ps. The
dynamics trajectory was viewed with the program VMD (Visual
Molecular Dynamics) [29].
RESULTS AND DISCUSSION
Characterization of trypsinolysed At PDF1B
Limited proteolysis of AtPDF1B with trypsin resulted in the
relatively slow formation of a lower-molecular-mass form,
AtPDF1Bt, reduced by approx. 3 kDa, indicative of N- and/or
C-terminal cut sites (Figure 1A). Although small reductions in
molecular mass were observed after prolonged incubation (>24 h;
results not shown), the proteolysed AtPDF1B generated was
largely resistant to further digestion. Recombinant AtPDF1B has a
requirement for 0.5 M NaCl for solubility at concentrations as low
as 0.5 mg · ml−1 , as reflected by limited solubility after desalting
or exhaustive dialysis (<10 %). However, after proteolysis with
trypsin, AtPDF1B lost this characteristic and was soluble without
NaCl at concentrations as great as ∼ 30 mg · ml−1 . We identified
an N-terminal trypsin-sensitive site at Lys65 using direct Nterminal sequencing that accounted for a mass loss of 0.944 kDa.
Therefore, the additional ∼ 2 kDa loss, to explain the final
monomeric molecular mass of the proteolysed form of AtPDF1B,
probably originated from the C-terminus. Because AtPDF1B had
a carboxyl-His-tag, immunological loss of reactivity against an
anti-His-tag antibody (Figure 1B) verified a C-terminal tryptic
site. Within a margin of error of 0.9 kDa, residues Arg261 , Lys263 ,
Arg264 and Lys265 would be reasonable candidates for C-terminal
sensitive sites. To identify the C-terminal site, we attempted
MALDI-TOF (matrix-assisted laser-desorption ionization–timeof flight) MS, but the measured molecular mass of AtPDF1Bt of
22 kDa was insufficiently accurate to unambiguously identify the
C-terminal trypsin-sensitive site.
The possible C-terminal cleavage sites and the N-terminal
trypsin-sensitive site at Lys65 are not located in the catalytic core of
the enzyme. However, removal of the N- and C-terminal segments
produced a ∼ 4-fold decrease in kcat (Figure 2), suggesting that
one or both of these segments may play a role in catalysis.
K m values were not affected by the removal of the segments.
Therefore, we explored the possibility that these trypsin-sensitive
sites in AtPDF1B might be located in regions involved in dynamic
conformational changes during catalysis or inhibitor binding.
Under catalytic conditions, the cleavage profile of AtPDF1B was
unchanged with time (Figure 1C compared with 1D); similarly,
binding of the potent inhibitor actinonin to AtPDF1B did not affect
trypsin
sensitivity (Figure 1E compared with 1A).
‘
Figure 1 Limited tryptic [1 % (w/w)] proteolysis of At PDF1B as a function
of time (A), concentration (C), catalytic conditions (D) and inhibitor binding
(E) analysed by SDS/(12.5 %) PAGE; and C-terminal sensitivity analysed by
immunoanalysis (B)
(A) Time course of molecular mass changes, indicative of final ∼ 3 kDa mass loss. Each lane
represented 2 μg of At PDF1B (2 mg · ml−1 ) after treatment with trypsin for the indicated times
(min). (B) Loss in immunoreactivity of C-terminal His-tagged At PDF1B to an anti-penta-His
antibody after proteolysis with trypsin (At PDF1Bt). (C) Proteolysis with concentrations of
At PDF1B (8.5 μg · ml−1 ) permitting a ∼ 10-fold excess of substrate at the end of 30 min in the
absence of catalysis. (D) Same conditions as in (C), but with the addition of 127 mM substrate
(f-Met-Ala-Ser). (E) Proteolysis as in (A), but with a 3 min preincubation of At PDF1B with 1 mM
actinonin prior to the addition of 1 % (w/w) trypsin. MM, molecular mass markers.
Overview of the At PDF1Bt structure
The crystal structure of AtPDF1Bt was determined at 2.4 Å
resolution (Table 1) by molecular replacement. One molecule
of the enzyme is present in the asymmetric unit, and residues
74–257 are defined by the electron density. Residues 65–73 are
therefore disordered at the N-terminus, and given the likely tryptic
cleavage sites, between four and eight residues are disordered at
the C-terminal end of the molecule. A total of 69 ordered water
molecules were located along with four sulphate ions. In addition
to the active site metal, two other zinc ions were placed, both
Figure 2
Kinetic analyses of the three forms of At PDF1B
Between 20 and 300 μM f-ML-ρ-NA was used as substrate for 7.5 μg of At PDF1B (filled
triangles), 31.6 μg of At PDF1Bt (filled circles) and 7.5 μg of At PDF1Bet (filled squares).
Collected data were fitted to the Michaelis–Menten equation V =V max [S]/(K m +[S]) using Sigma
Plot (Windows Version 4.0; SPSS, Chicago, IL, U.S.A.). At PDF1B (from [2]): V max = 20 +
− 0.3
0.1 sec−1 ; apparent K m = 130 +
4 μM;
μmoles · min−1 per mg of protein; k cat = 8.6 +
−
−
−1
−1
+
At PDF1Bt: V max = 4.6 +
− 0.5 μmoles · min per mg of protein; k cat−1= 1.7 − 0.2 sec ;
+ 2 μmoles · min per mg of protein;
31
μM;
At
PDF1Bet:
V
=
59
apparent K m = 118 +
max
−−1
−
+
k cat = 22.6 +
− 0.8 sec ; apparent K m = 78 − 8 μM.
c The Authors Journal compilation c 2008 Biochemical Society
420
Table 1
L. M. A. Dirk and others
Summary of crystallographic data and refinement for At PDF1Bt
Crystallographic data
Wavelength (Å)
Resolution (Å)
Last shell (Å)
Average redundancy (last shell)
R sym (last shell)
I/σ I (last shell)
Completeness (last shell) (%)
Refinement
Resolution (Å)
Number of reflections included in refinement
R work /R free
r.m.s.d. bond lengths (Å)
r.m.s.d. bond angles (◦)
r.m.s.d. improper angles (◦)
r.m.s.d. dihedral angles (◦)
B r.m.s.d. bonded atoms (main/side)
Average B for all protein atoms (Å2 )
Average B for ordered solvent (Å2 )
Number of solvent molecules
Number of metal ions
1.0
50–2.4
2.49–2.40
8.8 (9.1)
0.124 (0.394)
19.6 (7.1)
99 (99)
50–2.4
8016
0.235/0.296
0.007
1.3
1.2
23.9
1.4/2.0
23.1
20.3
60
4
The other structurally characterized eukaryotic type 1B PDF,
Pf PDF, forms two relatively strong interactions in the crystal
lattice and tends to aggregate when purified [21]. However,
neither contact resembles the proposed dimerization interface of
AtPDF1Bt, although the less extensive of the two also involves
the β5–6 hairpin that forms part of the interface in AtPDF1Bt.
Two other PDFs, AtPDF1A and the unusual deformylase from
Leptospira interrogans [31], have been shown to dimerize, and
their dimer interfaces are highly similar. Again, though, the
elements involved, the C-terminal helix and the small N-terminal
310 helix with surrounding coil regions, are different from those
in AtPDF1Bt. The observation of dimerization in different PDFs
suggests that it has been driven by a conferred functional advantage, though the nature of this advantage has not been uncovered.
In all cases, access to the active site is unlikely to be affected
by dimerization, and there have been no reports of co-operativity
in the activity of PDFs known to dimerize. Other properties
such as stability, lifetime and localization to nascent polypeptide
chains may be altered, and further investigation will be required
to determine if dimerization has functional significance.
Comparison with other peptide deformylases in the PDF1B group
interacting with side chain carboxylate groups from the enzyme
at intermolecular contacts in the crystal lattice.
AtPDF1Bt adopts an α+β fold (Figures 3A and 3B), similar to
other PDFs for which crystal structures have been reported. The
model consists of three α helices, three small 310 helices, and eight
β strands organized into two three-stranded antiparallel sheets and
a β hairpin with an extended turn. Strands β3 and β4 have been
considered to be one continuous strand in some PDF models
(creating a single five-stranded sheet), but the absence of sheet
contacts at Pro155 and the bend of about 90◦ in the strand near that
point suggests it is better assigned as two secondary elements in
AtDEF1Bt. The central helix (α2) contains a HEXXH sequence
motif (213 HEYDH217 in AtPDF1B) conserved among PDFs and
many zinc metallopeptidases. The two histidine residues from this
motif co-ordinate the metal ion cofactor (Figure 3C), which was
modelled as a zinc ion rather than the native iron, because zinc was
present at high concentration in the crystallization buffer. A third
metal ligand, Cys171 , arises from the extended turn connecting β
strands 5 and 6, and a water molecule, which acts as the attacking
nucleophile in deformylation, also co-ordinates the metal ion
cofactor.
Three intermolecular contacts in the crystal lattice, each
burying between 1300 and 1400 Å2 of accessible surface area,
are sufficiently extensive to be of interest. Two of the interactions
are asymmetric and repeat head to tail throughout the crystal
lattice to form filaments. The other lattice contact is symmetric,
occurring across a crystallographic two-fold axis (Figures 4A and
4B). This contact includes an intermolecular sheet-like interaction
between the N-terminal residues (162–164) of strand β5 with the
corresponding residues from the second molecule. The interface
also buries two hydrophobic residues (Leu164 and Val182 ) and
includes a number of residues involved in polar contacts (Tyr160 ,
Tyr215 , Thr228 and Asp229 ). Most peptide deformylases in the
PDF1B subgroup (see reference [30]) are thought to function as
monomers. However, gel filtration chromatography (Figure 4C)
indicates that AtPDF1Bt is primarily a dimer in solution, and there
is no indication of higher order multimer formation. It seems
likely, therefore, that the symmetric interaction mediates dimer
formation in solution, and that the other two interactions are
not sufficiently favourable to stabilize oligomerization under the
conditions used for testing.
c The Authors Journal compilation c 2008 Biochemical Society
As noted, AtPDF1Bt adopts a fold common to all reported
PDFs. Within the PDF1B subgroup, AtPDF1B shares a substantial
degree of sequence identity (45 % for the AtPDF1Bt construct;
see Figure 3B) with Pf PDF, which is the only other member
from a eukaryote with a known structure [21]. As expected from
this sequence relationship, the three-dimensional models of the
two enzymes are also very similar (Figure 5A), with an r.m.s.d.
(relative mean square deviation) on Cα positions (165 of 184 in
the AtPDF1Bt structure) of only 0.74 Å. Similarly, the EcPDF, as
a representative of the bacterial PDF1B enzymes with available
structures, shares somewhat less but still substantial sequence
identity (35 %). This sequence relationship translates into a high
degree of structural similarity (Figure 5B), with an r.m.s.d. of
only 0.98 Å on 145 Cα positions. The most prominent structural
difference among these three enzymes occurs at the C-terminal
helix. In EcPDF, this helix is shorter and rotated compared with
the other two enzymes, a difference noted by Kumar et al. [21]
in reporting the crystal structure of Pf PDF. This altered position
creates a more extensive contact with the β5-β6 hairpin turn,
which together with strand β5 forms one side of the active site
pocket. Somewhat unexpectedly, the thermal factors in the EcPDF
turn and strand β6 are higher relative to the core of the protein
than in the other two enzymes. This apparent increased mobility is
accompanied by higher thermal factors in the EcPDF C-terminal
helix, suggesting that is not as constrained in the altered position.
One other region, the loop between strands β2 and β3, adopts a
significantly different backbone conformation in each of the three
enzymes (Figure 5). This loop, which is often termed the “CDloop”, juts out from the surface of each enzyme on one side
of the active site pocket, effectively increasing its depth. Both the
AtPDF1Bt and Pf PDF loops have residues inserted relative to
the EcPDF loop, with two and three additional residues present
respectively. With these insertions, the larger AtPDF1Bt and
Pf PDF loops extend closer to the active site, more extensively
burying hydrophobic and aromatic residues that arise from the
C-terminus of helix α1. To some extent, the side chain of Arg66
in EcPDF buries the corresponding residues in that structure
(Figure 5C), but its coverage is less complete. The CD-loops in
Pf PDF and EcPDF are involved in crystal contacts, potentially influencing their conformations. This does not appear to be the case,
however, since for both there are multiple molecules in the crystal
Structure of plant peptide deformylase
Figure 3
421
Overview of the At PDF1Bt structure
(A) Stereo ribbons view of the At PDF1Bt crystal structure. Helices are coloured in blue and sheets in yellow. Secondary elements are labelled, and the metal cofactor (modelled as Zn2+ ) is in cyan.
(B) Sequence alignment of At PDF1B, Pf PDF and Ec PDF. Conserved and partially conserved residues are highlighted in red or shown in red text respectively. Conserved positions are also boxed in
blue. Secondary elements for the At PDF1Bt crystal structure are shown above the sequences, and the green arrows indicate the first and last residues found in the structure. The residue numbers
below the sequences are for At PDF1B. Numbers for the first residue of each line are given on the left. The green boxes indicated sequence regions widely conserved in PDFs. (C) Stereo view of
electron density for the active site of At PDF1Bt. The density was calculated using density modified phases with the initial phase start from the model, and it is displayed at 1.25 times the r.m.s.d. of the
map. The metal cofactor is in cyan.
asymmetric unit, and the different environments seen by the CDloops do not substantially affect their conformation. The binding
mode of short peptides suggests that the differences in the CDloop most likely affect interactions with residues distant from
the N-terminus, starting around position P4 . Although sensitivity
to substrate sequence has not been extensively studied for these
more distant positions, there are reports of changes in activity of
type 1B PDFs in response to changes in this region [11,14,32].
Loop swap experiments might be informative in establishing the
role of the CD-loop, and they should certainly be a part of future
work in this area.
Despite the overall high degree of structural similarity,
electrostatic potentials at the surfaces surrounding the active
sites differ markedly for the three enzymes (Figure 6). On
c The Authors Journal compilation c 2008 Biochemical Society
422
L. M. A. Dirk and others
be an adaptation to that specialized compartment. The relatively
electronegative surfaces of EcPDF and, particularly, AtPDF1Bt,
are likely to account for the low activity when negatively charged
residues are present in substrate sequences [14,32].
Substrate subsites and specificity
Figure 4
Dimerization of At PDF1Bt
(A and B) Two views of possible structure for an At PDF1Bt dimer. Two interacting molecules
(grey and green) from the crystal lattice are shown along with the side chains of residues
involved in the contact. Residues from the grey molecule are labelled. The two monomers in
the dimer are related by a crystallographic two-fold axis, and the two panels show views of the
dimer related by a 90◦ rotation about the horizontal axis. (C) Profile of At PDF1Bt eluted from a
gel filtration column, which was calibrated and run as described in the Experimental section.
that molecular face, AtPDF1Bt is decidedly electronegative.
Three acidic residues in the CD-loop (see Figure 5C) just
discussed contribute to this feature, but residues arising from
the β5-β6 hairpin, the N-terminal segment and the C-terminal
helix (α3) are also important. In contrast, this face of Pf PDF
carries an overall positive potential, although there are prominent
acidic residues (Glu127 and Glu163 ) on either side of the active
site. EcPDF differs from both, having a strong electrostatic
gradient across the surface. Acidic residues arise from the two
β sheets, the β5-β6 hairpin, and the N-terminal helix, whereas
the C-terminal elements are largely basic. These differences
in orthologous enzymes are surprising and may contribute to
variations in substrate preferences or possible differences in
subcellular localization. The unusual surface in Pf PDF has been
noted previously, and the relatively low activity of Pf PDF in
comparison with other type 1B PDFs has been ascribed in part to
this difference [21]. It should be mentioned that Pf PDF functions
in a specialized organelle of the Plasmodium parasite known as
the apicoplast [33], and the change in surface potential may in part
c The Authors Journal compilation c 2008 Biochemical Society
A number of studies have shown that AtPDF1B and other
PDFs exhibit marked substrate specificity despite their role in
deformylating a range of proteins in vivo [2,11,14,32,34]. The first
three N-terminal residues (P1 –P3 ) have the greatest effect
on substrate affinity and catalytic rate, although effects out to
positions P5 or P6 have been reported. For AtPDF1B, catalytic
activity with the peptide f-MTAIL, the best substrate sequence
so far identified, is maximal when only the first three residues
are present [14]. The tripeptide is also effectively the minimal
substrate, because activity on the dipeptide f-MT is less than 10 %
of the maximum, and no activity is detected with a single f-Met
residue.
Structurally, the first three subsites (S1 –S3 ) of the enzymes
have been most extensively studied, and EcPDF [17] and
AtPDF1A [4] in complex with the standard substrate f-MAS serve
as useful comparisons for AtPDF1Bt (Figure 7). Subsite S1 ,
which receives the N-formylmethionine residue, is a sterically
well defined, roughly elliptical pocket in all three enzymes
(Figure 7A). However, the detailed shape and electrostatic nature
of this pocket vary between the enzymes. In AtPDF1A, the
presence of Trp146 shortens one side of the pocket relative to
AtPDF1Bt, which has a leucine (Leu206 ) at this position. The
bulky tryptophan causes Phe107 to adopt a different rotamer from
the corresponding Phe167 in AtPDF1Bt, flipping it out of the site.
This change elongates the other side of the pocket in AtPDF1A. In
EcPDF, the corresponding two residues are a leucine (Leu125 ), as
in AtPDF1Bt, and an isoleucine (Ile86 ). The presence of the smaller
Ile86 relative to Phe167 in AtPDF1Bt gives a more symmetric shape
to the pocket in EcPDF, and the differences in pocket shape
cause the methionine side chain to adopt a different conformation
relative to the AtPDF1A complex. The EcPDF pocket also differs
from both AtPDF1Bt and AtPDF1A electrostatically. As noted, it
lacks both the tryptophan of AtPDF1A and the phenylalanine of
AtPDF1Bt, so the site has no aromatic character. It also has an
isoleucine at position 128 instead of the arginine present in the
other two enzymes (Arg209 in APDF1Bt and Arg148 in AtPDF1A).
In general, then, the EcPDF S1 site is more hydrophobic and
less featured sterically and electronically than the S1 sites of
the other two enzymes. These structural differences probably
explain the greater effect of substitutions at P1 on K m values
for AtPDF1A compared with EcPDF [7,32,35]. The structure of
the AtPDF1Bt subsite suggests that this enzyme would also be
relatively sensitive to changes at P1 , and kinetic results support
that prediction, although it does not appear to be as sensitive as
AtPDF1A [7].
The S2 subsite of the enzymes consists of an initially narrow
surface channel that extends for at least 15 Å to bulk solvent
(Figure 7B). The alanine Cβ from the bound f-Met-Ala-Ser peptides in the AtPDF1A and EcPDF complexes only just penetrates
the opening of this channel, suggesting that residues with longer
side chains might easily be accommodated. As with S1 , the shapes
of the S2 sites vary in detail between the enzymes, with the
EcPDF channel in particular opening up more than the other
two after a common initial restriction. At that initial restriction point, however, there are sequence differences between the
enzymes that could potentially lead to distinct P2 preferences.
The major difference is that a tyrosine residue (Tyr178 ) narrows
Structure of plant peptide deformylase
Figure 5
423
Comparison of the At PDF1Bt, Pf PDF, and Ec PDF crystal structures
(A) Ribbons stereo view of Pf PDF (yellow, PDB code 1RQC, [21]) superimposed on At PDF1Bt (grey). (B) Ribbons stereo view of Ec PDF (green, PDB code 1BS8, [17]) superimposed on At PDF1B
(grey). In both panels, the metal cofactor is shown in cyan. (C) Worm and stick view of the β2-β3 loops from At PDF1Bt (left), Pf PDF (middle), and Ec PDF (right). N- and C-terminal residues of the
loop are labelled, as is Arg66 from the Ec PDF structure.
Figure 6
Surface electrostatic potential of At PDF1Bt, Pf PDF and Ec PDF
Molecular surface views of the At PDF1Bt (A), Pf PDF (B) and Ec PDF (C). Strength of the electrostatic potential at the surface is indicated by the colour saturation (blue for positive, red for negative).
The molecules are shown in equivalent orientations with their active site pockets approximately at the centre of the visible faces.
c The Authors Journal compilation c 2008 Biochemical Society
424
Figure 7
L. M. A. Dirk and others
Comparison of substrate-binding subsites
Enzyme subsites that interact with the P1 (A), P2 (B) and P3 (C) substrate residues are compared for At PDF1A (left Figure in each panel, salmon colour, PDB code 1ZY1, [4]), At PDF1Bt (middle
Figure in each panel, grey colour), and Ec PDF (right Figure in each panel, lime colour, PDB code 1BS8, [17]). The enzymes are represented in ribbon and stick views, with selected side chains shown
and labelled. The molecular surfaces are drawn in semi-transparent grey. Metal ion cofactors are in cyan when visible. Residues from bound Met-Ala-Ser tripeptides are shown in stick representation
with carbons coloured in cyan for the At PDF1A and Ec PDF Figures. In panels (A) and (B) only, the first or first two ligand residues are shown respectively for clarity.
one side of the channel entrance in AtPDF1Bt, while that role
is played by an arginine side chain in the other two enzymes
(Arg118 in AtPDF1A, and Arg97 in EcPDF). The arginine residue
in EcPDF interacts with the main chain carbonyl oxygen at
P2 and the terminal carboxylate group, but in AtPDF1A the
arginine and bound substrate adopt different conformations that
prevent this interaction. Tyr178 is well conserved among plant
c The Authors Journal compilation c 2008 Biochemical Society
type 1B PDFs, being present in 29 out of 40 sequences examined
(see Supplementary Figure S1 at http://www.BiochemJ.org/bj/
413/bj4130417add.htm), and its possible role in recognition is
explored in the next section. On the other side of the channel
entrance, AtPDF1Bt has an aspartate residue (Asp120 ) that defines
the opening, and this position is also a negatively-charged residue,
Glu42 , in EcPDF. AtPDF1A, on the other hand, has a non-polar
Structure of plant peptide deformylase
residue, Pro146 , at that position. EcPDF also has an additional
negatively-charged residue at the adjacent position (Glu41 ), while
AtPDF1Bt has a threonine at the equivalent site (Thr119 ). AtPDF1A
has another non-polar residue (Ala145 ) to go with the proline
residue at position 146. These differences result in much more
charged/polar surfaces in AtPDF1B and EcPDF relative to
AtPDF1A. Interestingly, all three enzymes have relatively low
K m values for substrates with positively-charged residues at P2
[34–36]. In AtPDF1Bt and EcPDF, it is possible that Asp120
or Glu41 /Glu42 respectively play a role in this preference. The
more distant Asp115 may interact with positively-charged side
chains in AtPDF1A, because the negatively-charged residue at the
channel entrance is absent. AtPDF1B and EcPDF have proline
residues at the position equivalent to Asp115 , but EcPDF does
have a glutamate (Glu95 ) at the adjacent position. This extra
negative charge may account for the particularly low K m values
for P2 arginine or lysine in EcPDF [36]. Hu et al. [32] report
that EcPDF does not accommodate negatively-charged residues
at P2 , possibly consistent with the presence of Glu41 /Glu42 and
Glu95 in the site. However, other studies only show a modest
increase in K m for aspartate or glutamate at this position [34–
36]. AtPDF1B does seem to select against an aspartate residue at
P2 , at least in the context of a particular tetrapeptide sequence
[14], and this selectivity probably results from the presence
of Asp120 , since there are no other negative charges nearby.
Overall, the existing reports indicate modest levels of selectivity
and selectivity differences among the orthologues at P2 that
seem at odds with the relatively featured subsite and sequence
changes between the enzymes. Additional studies with systematic
variation of P2 substrate sequences in different contexts, coupled
with structures of peptide complexes, would permit a more
thorough analysis of selectivity at this position.
The S3 subsite is located on the opposite side of the active
site pocket from the S2 site (Figure 7C). In contrast with the
first two subsites, S3 is much broader and less sterically featured.
The major difference between the three enzymes is that additional
structural elements in AtPDF1A close off the back end of the site,
limiting its depth relative to the other two enzymes (see reference
[4]). These additional elements also add charged or polar residues
to the site (particularly Glu66 , Tyr77 and Arg84 ). Trp146 on one
side and Pro46 on the other act as side-to-side boundaries for the
site. In contrast, the S3 sites in AtPDF1Bt and EcPDF are
broader and deeper, open to solvent at the back. They are also
much less featured electronically, being composed almost entirely
of hydrophobic side chains. Only Asp120 (AtPDF1Bt) or Glu42
(EcPDF) add a charge in the vicinity, although there are polar
main chain groups and charged side chains (Glu143 and Glu146 in
AtPDF1Bt) contributed by the CD-loop at the distant back of the
site (not shown in Figure 7). As noted by Fieulaine et al. [4],
the positions of the P3 serine residues in the AtPDF1A and
EcPDF complexes differ due to the variations in the subsites of the
two enzymes. Less is known about P3 preferences than for the first
two substrate positions. It has been shown that EcPDF has
higher activity on peptides with tyrosine, phenylalanine and to a
lesser extent histidine at this position [32]. The aromatic residues
probably interact well with both the hydrophobic surface and,
since the site is exposed, solvent. Given the similarity of the sites,
the same preferences would be predicted for AtPDF1B. The P3
alanine in the favoured AtPDF1B substrate f-MTAI also probably
interacts with the surface well without a great cost associated
with solvent exposure. Interestingly, AtPDF1B does not appear to
accommodate a proline well at P3 [14]. The reason for bias against
proline is not clear, as the structure would predict a reasonably
good interaction with the S3 surface. It may be that the backbone
constraints from the proline force the P4 residue of the tetra-
425
peptide substrates into a position that partially clashes with
the enzyme surface. AtPDF1A also accommodates the aromatic
phenylalanine well at P3 , and substituting ρ-nitroanilide at that
position gives an even tighter K m . In contrast, placing a glutamate
at P3 gives a much higher K m , which is perhaps surprising given
the potential for a salt bridge with Arg84 .
Most of the 79 genes encoded by the Arabidopsis thaliana
chloroplast genome are N-terminally processed, but a significant
number are not [14,37], presumably due to specificity preferences
of the plant PDFs [14]. The number of chloroplast proteins for
which the lack of N-terminal processing produces a detrimental
phenotype is probably small [13,37], suggesting that there would
be selective pressure to maintain high activity for only a limited
number of N-terminal sequences. Thus, considerable specificity
variations may have arisen among the eukaryotic and prokaryotic
type 1B PDFs. An example of this selectivity may be the high
activity of AtPDF1B on the N-terminal sequence of the D1 polypeptide that forms part of PSII. This protein is rapidly translated in
the chloroplast under high light conditions [38], so considerable
activity may be required to process the large amounts during these
periods.
Possible role of Tyr178 in P2 recognition
Given the biological importance of PSII D1 polypeptide
processing by plant PDF1Bs [15], the preference of AtPDF1B
for the N-terminal sequence of D1, f-MTAIL [14], is of some
interest. Full activity is achieved with the f-MTA polypeptide, and
the threonine at P2 probably plays a large role in determining the
susceptibility of this substrate [14]. As previously noted, Tyr178 is a
conserved feature of plant PDF1B S2 subsites. On the other hand,
plant PDF1As, including AtPDF1A, for which f-MTAIL is not a
particularly good substrate, have a highly conserved arginine at the
equivalent position (see Supplementary Figure S1). It is therefore
possible that Tyr178 and the P2 threonine directly interact, perhaps
through hydrogen bond formation between the Oη of the tyrosine
and Oγ of the substrate threonine.
To explore this possibility, we modelled a f-MTAIL peptide
into the active site of AtPDF1B using the structure of the EcPDFMAS complex [17] as a guide. We then allowed the complex
to evolve in a molecular dynamics simulation, following Tyr178
interactions. The complex adjusted within about 50 ps to form a
relatively stable structure in which the side chain of Tyr178 formed
a hydrogen bond with the O of Thr-P2 of the bound peptide
(Figure 8A). Because the dynamics simulation showed that the
hydrogen bond interaction was structurally feasible, mutagenesis
was used to test this possibility experimentally (Figures 8B and
8C). The Y178A mutant shows an increase in K m for the f-MTA
substrate as expected if a direct binding interaction occurs, but
this increase is small, only just above the level of error (Table 2).
In addition, both the Y178F and, particularly, the Y178R mutants
show a decrease in K m (as well as an increase in kcat ), unexpectedly
indicating tighter binding of the substrate. Taken together, these
results suggest that Tyr178 in AtPDF1B does not directly hydrogen
bond to the side chain of Thr-P2 . Also, the presence of a tyrosine
at that position over the arginine present in AtPDF1A does not
mediate the relatively high affinity for the f-MTAIL sequence.
Other features of the S2 site of AtPDF1B must establish the
preference for a threonine.
Basis for specific inhibition
Inhibitors of plant PDFs hold great promise as broad-spectrum
herbicides and selectable markers in plant transformation vectors.
The general PDF inhibitor, actinonin, is known to arrest growth
c The Authors Journal compilation c 2008 Biochemical Society
426
L. M. A. Dirk and others
Table 2 Kinetic parameters of three site-directed mutants at Y178 in
At PDF1B
At PDF1B activity was measured using a synthetic polypeptide substrate mimic of the
unprocessed D1 N-termini (f-MTA) and the data were plotted and analysed using Sigma Plot
(version 10.0). Data are shown as means +
− S.E.M., n = 2.
Y178
Y178F
Y178R
Y178A
Figure 9
k cat (s−1 )
K m (mM)
k cat /K m (s−1 · mM−1 )
88.5 +
− 4.7
112.2 +
− 15.7
109.9 +
− 7.3
111.2 +
− 23.3
9.2 +
− 0.8
7.5 +
− 1.8
4.9 +
− 0.6
14.4 +
− 4.2
9.6
15.0
22.6
7.7
Potential for development of specific inhibitors
Residues in At PDF1Bt that are conserved among plant PDF1Bs (> 50 %) and not prevalent
(< 20 %) in bacterial PDF1Bs were identified. Sequences of 40 plant and 100 bacterial PDFs
were examined. Those residues meeting the criteria are shown as stick figures with a ribbons
representation of At PDF1Bt. Only residues from three known conserved regions that surround
the active site of the enzyme were included in the comparison. Those regions are indicated by
the green, yellow, and blue backbone colours.
Figure 8
Role of Tyr178 in recognition of substrate P2 threonine
(A) Snapshot from a molecular dynamics simulation (AMBER [28]) of At PDF1Bt with modelled
bound MTAIL peptide that represents the N-terminal sequence of the substrate D1 peptide from
PSII. The snapshot shows the model at 126 ps of a 2 ns long simulation. The enzyme is shown
in a ribbons representation with the Tyr178 side chain drawn in a stick representation. The MTAIL
peptide is also shown in stick form with residues labelled, and the metal ion cofactor is in cyan.
A potential hydrogen bond between Tyr178 and Thr2 of the ligand is indicated by the dashed
line. (B) SDS/PAGE of the site-directed mutants at Tyr178 in At PDF1B, which were constructed
and the proteins purified from E . coli as described in the Experimental section. The residues
substituted in the mutant for Tyr178 are indicated at the top of each lane. (C) At PDF1B activity
was measured using a synthetic polypeptide substrate mimic of the unprocessed D1 N-termini
(f-MTA) and the data were plotted and analysed using Sigma Plot (version 10.0). Curves are
for wild-type At PDF1B (filled circles) and the mutants Y178A (open triangles), Y178F (open
circles) and Y178R (filled triangles).
and development in all plant species that have been tested [2], and
an important goal is to develop new inhibitors that both resist
metabolism and do not damage critical micro-organisms that
establish the soil ecology. An ideal inhibitor, then, would be
specific to plant PDF1Bs, and would not, for example, be effective
against the bacterial PDFs. With bacteria, it is also important to
prevent the development of resistance to PDF inhibitors, since
their potential use as antibiotics is a particularly active area of
biomedical research. Because the nearly identical backbone folds
of type 1B PDFs themselves provide little basis for developing
c The Authors Journal compilation c 2008 Biochemical Society
selective inhibitors, conserved sequence differences within this
group may afford the best opportunity for obtaining specificity.
There are three regions of high sequence conservation among
groups of PDFs (see Figure 3B). Examining these regions for
residues that are conserved (> 50 %) in plant PDF1B orthologues
and not prevalent (< 20 %) in bacterial PDFs identified ten
positions that meet these criteria (Figure 9; see also Figure S1 in
the Supplementary online data). Of these ten positions, however,
only four are located close enough to the active site pocket to be
of interest for development of competitive inhibitors. These residues are Thr119 , Asp120 , Phe167 and Pro207 . The most common
residues at these positions in bacteria are alanine (43 %), arginine
(26 %), glycine (19 %) and leucine (44 %) respectively, all nonconservative changes. Thr119 and Asp120 are located in subsite S2
and seem particularly good candidates for developing specific
interactions. Phe167 , which forms part of the S1 and S3 subsites,
would again be a good candidate, although variability among the
bacterial sequences at this position may limit its usefulness.
Finally, Pro207 is located in the back of the S3 subsite, a little further from the active site than the other positions. It is near the CDloop, which is the exception to the rule of backbone similarity
among the PDF1Bs. An effective strategy might therefore be
to target both the proline and the CD-loop for specific inhibitor
development.
Structure of plant peptide deformylase
Funding for this project was provided by a National Science Foundation Award (no.
MCB-0240165) to L. M. A. D., R. L. H., M. A. W. and Anne-Frances Miller. Support for
D. W. R., J. J. S. and core facility use was provided by United States Public Health Service
Grants NS38041 and P20 RR20171. X-ray data were collected at the SER-CAT (Southeast
Regional Collaborative Access Team) 22-ID beamline at the Advanced Photon Source,
Argonne National Laboratory. We thank the staff of SER-CAT for their help with data
collection.
REFERENCES
1 Mazel, D., Pochet, S. and Marliere, P. (1994) Genetic characterization of polypeptide
deformylase, a distinctive enzyme of eubacterial translation. EMBO J. 13, 914–923
2 Dirk, L. M. A., Williams, M. A. and Houtz, R. L. (2001) Eukaryotic peptide deformylases.
Nuclear-encoded and chloroplast-targeted enzymes in Arabidopsis . Plant Physiol. 127,
97–107
3 Giglione, C., Serero, A., Pierre, M., Boisson, B. and Meinnel, T. (2000) Identification of
eukaryotic peptide deformylases reveals universality of N-terminal protein processing
mechanisms. EMBO J. 19, 5916–5929
4 Fieulaine, S., Juillan-Binard, C., Serero, A., Dardel, F., Giglione, C., Meinnel, T. and
Ferrer, J. L. (2005) The crystal structure of mitochondrial (Type 1A) peptide deformylase
provides clear guidelines for the design of inhibitors specific for the bacterial forms.
J. Biol. Chem. 280, 42315–42324
5 Lee, M. D., She, Y., Soskis, M. J., Borella, C. P., Gardner, J. R., Hayes, P. A., Dy, B. M.,
Heaney, M. L., Philips, M. R., Bornmann, W. G. et al. (2004) Human mitochondrial
peptide deformylase, a new anticancer target of actinonin-based antibiotics.
J. Clin. Invest. 114, 1107–1116
6 Leeds, J. A. and Dean, C. R. (2006) Peptide deformylase as an antibacterial target: a
critical assessment. Curr. Opin. Pharmacol. 6, 445–452
7 Serero, A., Giglione, C., Sardini, A., Martinez-Sanz, J. and Meinnel, T. (2003) An unusual
peptide deformylase features in the human mitochondrial N-terminal methionine excision
pathway. J. Biol. Chem. 278, 52953–52963
8 Lee, M. D., Antczak, C., Li, Y., Sirotnak, F. M., Bornmann, W. G. and Scheinberg, D. A.
(2003) A new human peptide deformylase inhibitable by actinonin. Biochem. Biophys.
Res. Commun. 312, 309–315
9 Nguyen, K. T., Hu, X., Colton, C., Chakrabarti, R., Zhu, M. X. and Pei, D. (2003)
Characterization of a human peptide deformylase: implications for antibacterial drug
design. Biochemistry 42, 9952–9958
10 Mazel, D., Coic, E., Blanchard, S., Saurin, W. and Marliere, P. (1997) A survey of
polypeptide deformylase function throughout the eubacterial lineage. J. Mol. Biol. 266,
939–949
11 Rajagopalan, P. T., Datta, A. and Pei, D. (1997) Purification, characterization and
inhibition of peptide deformylase from Escherichia coli . Biochemistry 36, 13910–13918
12 Hou, C. X., Dirk, L. M. A., Pattanaik, S., Das, N. C., Maiti, I. B., Houtz, R. L. and Williams,
M. A. (2007) Plant peptide deformylase: a novel selectable marker and herbicide target
based on essential cotranslational chloroplast protein processing. Plant Biotechnol. J. 5,
275–281
13 Giglione, C., Vallon, O. and Meinnel, T. (2003) Control of protein life-span by N-terminal
methionine excision. EMBO J. 22, 13–23
14 Dirk, L. M. A., Williams, M. A. and Houtz, R. L. (2002) Specificity of chloroplast-localized
peptide deformylases as determined with peptide analogs of chloroplast-translated
proteins. Arch. Biochem. Biophys. 406, 135–141
15 Hou, C. X., Dirk, L. M. A. and Williams, M. A. (2004) Inhibition of peptide deformylase in
Nicotiana tabacum leads to decreased D1 protein accumulation, ultimately resulting in a
reduction of photosystem II complexes. Am. J. Bot. 91, 1304–1311
16 Yuan, Z. and White, R. J. (2006) The evolution of peptide deformylase as a target:
contribution of biochemistry, genetics and genomics. Biochem. Pharmacol. 71,
1042–1047
427
17 Becker, A., Schlichting, I., Kabsch, W., Groche, D., Schultz, S. and Wagner, A. F. (1998)
Iron center, substrate recognition and mechanism of peptide deformylase.
Nat. Struct. Biol. 5, 1053–1058
18 Becker, A., Schlichting, I., Kabsch, W., Schultz, S. and Wagner, A. F. (1998) Structure of
peptide deformylase and identification of the substrate binding site. J. Biol. Chem. 273,
11413–11416
19 Chan, M. K., Gong, W., Rajagopalan, P. T., Hao, B., Tsai, C. M. and Pei, D. (1997) Crystal
structure of the Escherichia coli peptide deformylase. Biochemistry 36, 13904–13909
20 Meinnel, T., Blanquet, S. and Dardel, F. (1996) A new subclass of the zinc
metalloproteases superfamily revealed by the solution structure of peptide deformylase.
J. Mol. Biol. 262, 375–386
21 Kumar, A., Nguyen, K. T., Srivathsan, S., Ornstein, B., Turley, S., Hirsh, I., Pei, D. and Hol,
W. G. (2002) Crystals of peptide deformylase from Plasmodium falciparum reveal critical
characteristics of the active site for drug design. Structure 10, 357–367
22 Wei, Y. and Pei, D. (1997) Continuous spectrophotometric assay of peptide deformylase.
Anal. Biochem. 250, 29–34
23 Rodgers, D. W. (1997) Practical cryocrystallography. Methods Enzymol. 276, 183–203
24 Otwinowski, Z. and Minor, W. (1997) Processing of X-ray diffraction data collected in
oscillation mode. Methods Enzymol. 276, 307–326
25 Brunger, A. T., Adams, P. D., Clore, G. M., DeLano, W. L., Gros, P., Grosse-Kunstleve,
R. W., Jiang, J. S., Kuszewski, J., Nilges, M., Pannu, N. S. et al. (1998) Crystallography &
NMR system: a new software suite for macromolecular structure determination. Acta
Crystallogr. D Biol. Crystallogr. 54, 905–921
26 Jones, T. A., Zou, J. Y., Cowan, S. W. and Kjeldgaard, M. (1991) Improved methods for
building protein models in electron density maps and the location of errors in these
models. Acta Crystallogr. A 47, 110–119
27 Nicholls, A., Sharp, K. A. and Honig, B. (1991) Protein folding and association: insights
from the interfacial and thermodynamic properties of hydrocarbons. Proteins: Struct.
Funct. Genet. 11, 281–296
28 Case, D. A., Cheatham, 3rd, T. E., Darden, T., Gohlke, H., Luo, R., Merz, Jr, K. M.,
Onufriev, A., Simmerling, C., Wang, B. and Woods, R. J. (2005) The Amber biomolecular
simulation programs. J. Comput. Chem. 26, 1668–1688
29 Humphrey, W., Dalke, A. and Schulten, K. (1996) VMD: Visual Molecular Dynamics.
J. Mol. Graph. 14, 33–38
30 Giglione, C., Boularot, A. and Meinnel, T. (2004) Protein N-terminal methionine excision.
Cell. Mol. Life Sci. 61, 1455–1474
31 Zhou, Z., Song, X. and Gong, W. (2005) Novel conformational states of peptide
deformylase from pathogenic bacterium Leptospira interrogans : implications for
population shift. J. Biol. Chem. 280, 42391–42396
32 Hu, Y. J., Wei, Y., Zhou, Y., Rajagopalan, P. T. and Pei, D. (1999) Determination of
substrate specificity for peptide deformylase through the screening of a combinatorial
peptide library. Biochemistry 38, 643–650
33 Waller, R. F. and McFadden, G. I. (2005) The apicoplast: a review of the derived plastid of
apicomplexan parasites. Curr. Issues Mol. Biol. 7, 57–79
34 Serero, A., Giglione, C. and Meinnel, T. (2001) Distinctive features of the two classes of
eukaryotic peptide deformylases. J. Mol. Biol. 314, 695–708
35 Meinnel, T., Patiny, L., Ragusa, S. and Blanquet, S. (1999) Design and synthesis of
substrate analogue inhibitors of peptide deformylase. Biochemistry 38, 4287–4295
36 Ragusa, S., Mouchet, P., Lazennec, C., Dive, V. and Meinnel, T. (1999) Substrate
recognition and selectivity of peptide deformylase. Similarities and differences with
metzincins and thermolysin. J. Mol. Biol. 289, 1445–1457
37 Giglione, C. and Meinnel, T. (2001) Organellar peptide deformylases: universality of the
N-terminal methionine cleavage mechanism. Trends Plant Sci. 6, 566–572
38 Kettunen, R., Tyystjarvi, E. and Aro, E. M. (1996) Degradation pattern of photosystem II
reaction center protein D1 in intact leaves. The major photoinhibition-induced cleavage
site in D1 polypeptide is located amino terminally of the DE loop. Plant Physiol. 111,
1183–1190
Received 5 December 2007/2 April 2008; accepted 16 April 2008
Published as BJ Immediate Publication 16 April 2008, doi:10.1042/BJ20071641
c The Authors Journal compilation c 2008 Biochemical Society