GTP hydrolysis of cell division protein FtsZ: evidence that the active site is formed by the association of monomers.

Abstract

The essential prokaryotic cell division protein FtsZ is a tubulin homologue that forms a ring at the division site. FtsZ forms polymers in a GTP-dependent manner. Recent biochemical evidence has shown that FtsZ forms multimeric structures in vitro and in vivo and functions as a self-activating GTPase. Structural analysis of FtsZ points to an important role for the highly conserved tubulin-like loop 7 (T7-loop) in the self-activation of GTP hydrolysis. The T7-loop was postulated to form the active site together with the nucleotide-binding site on an adjacent FtsZ monomer. To characterize the role of the T7-loop of Escherichia coli FtsZ, we have mutagenized residues M206, N207, D209, D212, and R214. All the mutant proteins, except the R214 mutant, are severely affected in polymerization and GTP hydrolysis. Charged residues D209 and D212 cannot be substituted with a glutamate residue. All mutants interact with wild-type FtsZ in vitro, indicating that the T7-loop mutations do not abolish FtsZ self-association. Strikingly, in mixtures of wild-type and mutant proteins, most mutants are capable of inhibiting wild-type GTP hydrolysis. We conclude that the T7-loop is part of the active site for GTP hydrolysis, formed by the association of two FtsZ monomers.

Full text

Page 1

42

Chapter 4

GTP-hydrolysis of cell division protein FtsZ:
evidence that the active site is formed by two monomers




Dirk-Jan Scheffers, Janny G. de Wit, Tanneke den Blaauwen and Arnold J.M. Driessen

Submitted for publication






Summary

The essential prokaryotic cell division protein FtsZ is a tubulin homologue that forms a ring at
the division site. FtsZ forms polymers in a GTP dependent manner. Recent biochemical evidence
has shown that FtsZ forms multimeric structures in vitro and in vivo, and functions as a self-
activating GTPase. Structural analysis of FtsZ points to an important role for the highly conserved
tubulin-like loop 7 (T7-loop) in the self-activation of GTP hydrolysis. The T7-loop was postulated
to form the active site together with the nucleotide-binding site on an adjacent FtsZ monomer. To
characterize the role of the T7-loop of Escherichia coli FtsZ, we have mutagenized residues M206,
N207, D209, D212 and R214. All the mutant proteins, except the R214 mutant, are severely
affected in polymerization and GTP hydrolysis. Charged residues D209 and D212 cannot be
substituted with a glutamate residue. All mutants interact with wt FtsZ in vitro, indicating that the
T7-loop mutations do not abolish FtsZ self-association. Strikingly, in mixtures of wildtype and
mutant proteins, most mutants are capable of inhibiting wildtype GTP hydrolysis. We conclude that
the T7-loop is part of the active site for GTP hydrolysis, formed by two FtsZ monomers.

Page 2

Role of the FtsZ T7-loop

43
Introduction

FtsZ is a key prokaryotic cell division
protein, forming a structural element known as
the Z-ring at the site of cell division (for
reviews see (18, 83, 85)). FtsZ has been
identified in most prokaryotic species studied to
date (reviewed in (92)) and is conserved across
the evolutionary divide, being essential for
division of chloroplasts and mitochondria in
some eukaryotes (7, 111). In Escherichia coli,
the Z-ring is critical for the localization of all
other protein components of the cell division
machinery (92). The Z-ring is likely to consist
of polymers similar to the FtsZ polymers that
can be formed in vitro in a GTP-dependent
manner (19, 51, 101). In E. coli, the Z-ring is
thought to be tethered to the cytoplasmic
membrane by ZipA (57).
FtsZ is a 40.3 kDa protein, that binds and
hydrolyzes GTP (34, 100, 117). It shows a
striking structural similarity to α- and β-
tubulins, both as a monomer and when modeled
onto protofilaments (Fig. 1A) (77, 78, 107,
108). It shares with tubulins a unique GTP-
binding motif, the G-box [GGGTG(ST)G] (34).
Recent in vivo data suggest that FtsZ is present
in the cytosol as multimeric structures (69).
Analytical ultracentrifugation has shown that
FtsZ associates into multimeric structures, and
that Mg2+ induces
oligomerization of FtsZ in the presence of GDP,
which is enhanced in the presence of crowder
proteins (120, 120a, 128). Though the formation
of FtsZ polymer protofilaments may be non-
cooperative (120-121), both the GTPase activity
and the formation of large FtsZ polymers of
FtsZ are dependent on the FtsZ concentration
(81, 102, 128). Also, FtsZ polymers consist
mainly of GDP bound FtsZ (chapter 3). These
findings indicate the GTPase activity of FtsZ is
self-activated, i.e. that association of FtsZ
monomers triggers the GTPase activity of FtsZ.
Structural analysis of the organization of
FtsZ subunits in polymers, as well as described
mutants, point at an important role for the T7-
indefinite linear
loop region (“tubulin-loop” no. 7) (106) of FtsZ
in self-association of monomers (the loop is
also known as HC1-loop, since it includes the
first short α-helical structure of the C-terminal
part of FtsZ, H8 (77)). Modeling of the crystal
structure of Methanococcus jannaschii FtsZ1
onto electron microscopical
protofilaments of FtsZ1, suggests that the active
site for GTP hydrolysis may be shared by two
FtsZ subunits, with the GTP-binding domain
located on one monomer and modulation of
hydrolysis by the T7-loop of the other domain
(Fig. 1A) (78). This model shows some striking
similarities to the organization of tubulin in
filaments (107). The T7 loop is highly
conserved among FtsZ proteins from different
species (Fig. 1B). The region shows homology
to the Mg2+-binding loop of the Walker B
consensus sequence for ATPases (118, 140).
Mutations in E. coli FtsZ affecting GTP
hydrolysis occur either in the T7-loop or in
domains involved in nucleotide binding as
judged from the crystal structure (77). A strictly
conserved Glu254 in the α-tubulin T7-loop is
thought to reach into the active site of β-tubulin
to activate the GTPase activity of β-tubulin,
whereas the strictly conserved Lys254 of β-
tubulin renders α-tubulin inactive as a GTPase
(residue numbers corresponding to bovine brain
tubulin, in yeast: E255 [α] and K252 [β], Fig.
1) (106). The Glu254 residue of α-tubulin
corresponds to Asp212 in E. coli FtsZ. The E.
coli SulA resistant ftsZ2 mutant carries a
D212G mutation, which abolishes the GTPase
activity of FtsZ but not GTP binding (31, 137).
Engineered N207D and D209N mutations bind
GTP, but display a severely reduced GTPase
activity (145), while R214A and R214K
mutations do not alter GTPase activity (128).
Recently, we have shown that replacement of
the D212 residue with cysteine or asparagine,
but not with glutamate, allows purification of
GTP-containing mutant protein. The D212C
and D212N mutants could be polymerized by
the addition of cations, indicating that the
nucleotide binding site in the polymer contains
images of

Page 3

Chapter 4
44
Figure 1: Model of the nucleotide binding site of FtsZ in a protofilament and alignment of the T7-loop of
FtsZ region. A) The crystal structure of M. jannaschii FtsZ1 (PDB accession code 1FSZ, (77)) was oriented
according to the αβ-tubulin heterodimer structure (PDB accession code 1TUB, (108)) as described by Löwe and
Amos (78). Residues N233, D235, D238 and K240 of M. jannaschii FtsZ1 are the equivalent of E. coli FtsZ N207,
D209, D212 and R214, respectively.
B) Alignment of FtsZ from different bacterial and eukaryotic species, and yeast α- and β-tubulin. The T7-loop
region and the following helix 8 are indicated in gray and black at the top. Amino acid changes described in this
paper are indicated at the corresponding positions above the E. coli FtsZ sequence. Fully conserved residues are
highlighted; other conserved groups are indicated using ClustalX terminology (135). A BLAST search was
performed with E. coli FtsZ, and an evolutionary tree was drawn with proteins with less than 60% positives. FtsZs
from representative branches were selected and aligned using ClustalX 1.8 (135). E. coli FtsZ (SWISSPROT:
FTSZ_ECOLI), Rickettsia prowazekii FtsZ (SWISSPROT: FTSZ_RICPR), Bacillus subtilis FtsZ (SWISSPROT:
FTSZ_BACSU), Borrelia burgdorferii FtsZ (GENBANK: C70137), Pyrococcus horikoshii FtsZ1 (SWISSPROT:
FTZ1_PYRHO), Methanococcus jannaschii FtsZ1 (SWISSPROT: FTZ1_METJA), Thermotoga maritima FtsZ
(GENBANK: AAC24604), Helicobacter pylori FtsZ (GENBANK: D71873), Mycoplasma pneumoniae FtsZ
(SWISSPROT: FTSZ_MYCPN), Arabidopsis thalania chloroplast FtsZ (GENBANK: AAC35987), Mallomonas
splendens mitochondrial FtsZ (GENBANK: AAF35433), Saccharomyces cerevisiae α- and β-tubulin
(GENBANK: AAA35181 and CAA24603).

Page 4

Role of the FtsZ T7-loop

45
a cation that is coordinated by D212 (see
chapter 5).
Although the T7-loop seems to be involved
in GTP hydrolysis, it is unlikely that it is critical
in the formation of FtsZ dimers. A study of ten
randomly generated mutations that abolished
the FtsZ-FtsZ interaction in a λ-hybrid assay
revealed only one mutant with two amino acid
substitutions of which one mapped in the T7-
loop, F210A (39).
In this paper we present a mutational
analysis of the T7-loop. Various T7-loop
residues were mutated on the basis of their
position in the interaction domain (Fig. 1A) and
their possible role in GTP hydrolysis. Cysteine
mutants of residues M206, N207, D209, D212
and R214 were tested for polymerization, GTP-
binding and hydrolysis. Polymerization was
studied with and without Ca2+, which induces
FtsZ-polymer bundling and stabilizes the
polymers (104, 160). D209 and D212 were also
replaced with asparagine and glutamate to
evaluate the role of the negative charges. All
mutants except R214C displayed different
polymerization and GTP hydrolysis behavior
compared to wildtype FtsZ. Interaction studies
of the mutant proteins with wildtype FtsZ
revealed that most mutants suppress the GTP
hydrolysis activity of the wild type FtsZ. These
data provide the first conclusive biochemical
evidence that the T7-loop is part of the active
site, formed by two FtsZ monomers, for GTP
hydrolysis of FtsZ.

Results

Construction of T7 loop mutants
To investigate the role of the T7 loop in
FtsZ function, we
mutagenesis to replace methionine 206,
asparagine 207, aspartate 209, phenylalanine
210, aspartate 212 and arginine 214 with a
cysteine residue (Fig. 1). Aspartate residues 209
and 212 were also replaced with asparagine and
glutamate residues. All mutant proteins, except
F210C, were overexpressed to levels similar to
wild type (wt) FtsZ and purified (not shown).
used site directed
For expression of the D212E mutant protein E.
coli strain BL21(DE3)pLysS was used since the
expression plasmid was toxic to strain
BL21(DE3), probably due to promoter leakage.

Mutant polymerization activities
To characterize the in vitro properties of the
T7-loop mutants we first analyzed whether the
mutants were capable of polymerization with
GTP using light scattering and electron
microscopy. Mutant R214C showed a behavior
similar to wt FtsZ, with GTP-dependent
polymerization and depolymerization (Figs. 2A,
3C) and increased bundling in the presence of
10 mM Ca2+ (Figs. 2B, 3D). M206C, N207C,
D209E and D209C were incapable of
polymerization as studied by light scattering
and electron microscopy both with and without
Ca2+ (not shown, Table 1). D209N and D212E
showed polymerization in the presence of Ca2+
(Figs. 2C, D and 3F, G) in a protein
concentration dependent fashion. The gradual
increase observed in the light scattering
represents slow polymer bundling as evidenced
by electron microscopy (not shown). Electron
microscopy revealed the polymerization of
D209N in the absence of calcium, although the
amount of filaments was very low (Fig. 3E) but
increased in time (not shown). The D212N and
D212C mutants showed polymerization with
divalent cations in the absence of added
nucleotide (chapter 5). The results show that
R214C polymerizes to a lesser extent as wt
FtsZ, but acts similar to the wild type with
instant polymerization upon GTP addition
independent of Ca2+. The threshold protein
concentration for the calcium dependent
polymerization of D209N and D212E is
increased.

GTPase activities
Since the polymerization activity of wt FtsZ
is coupled to the GTP hydrolysis rate (81, 102,
128) we tested if the polymerization behavior of
the mutants was coupled to the GTP hydrolysis
activity. First, binding of GTP was assessed
using TLC. All T7 loop mutants bound GTP,

Page 5

Chapter 4
46
although
differences in affinity (Fig. 4A). M206C,
N207C and D209E bound less GTP that wt
FtsZ, whereas D209C and D209N bound more
GTP than wt FtsZ. This shows that GTP-
binding is affected by alterations in the T7-loop.
The inclusion of MgCl2 converts the assay to a
“single turnover” GTPase assay. Only wt FtsZ,
M206C, D212E and R214C showed significant
hydrolysis as evidenced by the presence of GDP
(Fig. 4B). “Multiple
measurements revealed
exhibited, albeit reduced, Mg2+-dependent
GTPase activity (Fig. 5A, Table 1). Out of the
mutants capable of polymerization, R214C
retained about half of the wt GTPase activity
while D209N and D212E were dramatically
impaired in the GTPase activity. Even at
to various extents indicating
turnover”
that
GTPase
mutants all
concentrations up to 0.8 mg/ml, D209N and
D212E showed no self-activation of the GTPase
activity (not shown). The reduced activity of
D209E and D212E indicates that the aspartate
residues in the T7 loop cannot be substituted by
another negatively charged residue.
The T7 loop harbors a Walker B-like amino
acid sequence motif (118) and is presumably
located in the vicinity of the GTP binding site
on a contacting FtsZ monomer (78). Our results
with the D212 mutants show that D212 is
implicated in cation coordination (chapter 5).
To further characterize cation coordination by
the T7-loop region, we studied the GTPase
activity of the mutants at various concentrations
of Mg2+ and Ca2+. R214C responded similar as
wt FtsZ to increased concentrations of these
divalent cations (Figs. 5A and 5B). In the
Figure 2: Polymerization of wild type and mutant FtsZ. Polymerization of wt FtsZ and R214C in the absence (A)
and presence (B) of calcium, and of D209N (C) and D212E (D) in the presence of calcium and at a protein
concentration of 0.25; 0.5; 0.75 and 1 mg/ml. Protein (0.5 mg/ml, unless indicated otherwise) was incubated at 30 °C
in buffer A supplemented with 10 mM MgCl2 (A) or 5 mM MgCl2 and 10 mM CaCl2 (B-D) for 1 min before the
addition of 0.5 mM GTP. Polymer formation was monitored by 90° light scattering. For A the photomultiplier was set
to 350 V; for B, C, and D to 275 V; therefore the signal amplitude cannot be compared between panel A and panels B,
C and D.