Detailed structure-function correlations of bacillus subtilis acetolactate synthase

Article abstract of DOI:10.1002/cbic.201402541

Isobutanol is deemed to be a next-generation biofuel and a renewable platform chemical.^[1] Non-natural biosynthetic pathways for isobutanol production have been implemented in cell-based and in vitro systems with Bacillus subtilis acetolactate synthase (AlsS) as key biocatalyst.^[2-6] AlsS catalyzes the condensation of two pyruvate molecules to acetolactate with thiamine diphosphate and Mg²⁺ as cofactors. AlsS also catalyzes the conversion of 2-ketoisovalerate into isobutyraldehyde, the immediate precursor of isobutanol. Our phylogenetic analysis suggests that the ALS enzyme family forms a distinct subgroup of ThDP-dependent enzymes. To unravel catalytically relevant structure-function relationships, we solved the AlsS crystal structure at 2.3 ? in the presence of ThDP, Mg²⁺ in a transition state with a 2-lactyl moiety bound to ThDP. We supplemented our structural data by point mutations in the active site to identify catalytically important residues.

Full text of DOI:10.1002/cbic.201402541

CHEMBIOCHEM
FULL PAPERS
DOI: 10.1002/cbic.201402541
Detailed Structure–Function Correlations of Bacillus subtilis
Acetolactate Synthase
Bettina Sommer,[a] Holger von Moeller,[b, c] Martina Haack,[a] Farah Qoura,[a]
Clemens Langner,[b] Gleb Bourenkov,[d] Daniel Garbe,[a] Bernhard Loll,*[b] and
Thomas Brꢀck*[a]
Isobutanol is deemed to be a next-generation biofuel and a
renewable platform chemical.[1] Non-natural biosynthetic path-
ways for isobutanol production have been implemented in
cell-based and in vitro systems with Bacillus subtilis aceto-
lactate synthase (AlsS) as key biocatalyst.[2–6] AlsS catalyzes the
condensation of two pyruvate molecules to acetolactate with
thiamine diphosphate and Mg2+ as cofactors. AlsS also catalyz-
es the conversion of 2-ketoisovalerate into isobutyraldehyde,
the immediate precursor of isobutanol. Our phylogenetic anal-
ysis suggests that the ALS enzyme family forms a distinct sub-
group of ThDP-dependent enzymes. To unravel catalytically
relevant structure-function relationships, we solved the AlsS
crystal structure at 2.3 ꢁ in the presence of ThDP, Mg2+ and in
a transition state with a 2-lactyl moiety bound to ThDP. We
supplemented our structural data by point mutations in the
active site to identify catalytically important residues.
Introduction
The imminent end of economically relevant petrochemical re-
sources is driving the development of sustainable, biomanufac-
turing processes to generate biofuels and commodity platform
chemicals.1 Because of its hydrophobic properties and high
caloric value isobutanol is noted as a novel renewable platform
chemical and biofuel component. Both in vivo and in vitro iso-
butanol production processes have been reported.[1–5,7–9] The
in vitro approach allows elimination of undesired metabolic
side-reactions that reduce molecular efficiency. Cell-based iso-
butanol yields are currently just 2% (v/v) due to its toxic ef-
fects.[1,3] By contrast, in vitro isobutanol synthesis potentially
offers a yield of 12% (v/v), and allows spontaneous phase sep-
aration and simplified product recovery procedures.[10] An es-
sential biocatalyst of isobutanol biosynthesis is thiamine di-
phosphate (ThDP)-dependent acetolactate synthase (ALS; EC
2.2.1.6), which is involved in natural butane-2,3-dione biosyn-
thesis.[11,12] ALS requires the cofactors ThDP and Mg2+ in order
to catalyze the decarboxylation of one molecule pyruvate and
transfer of the resulting cofactor-bound hydroxyethyl group
onto a second pyruvate molecule to make acetolactate (Fig-
ure 1A). In a side reaction, the ALS of Bacillus (AlsS) catalyzes
the decarboxylation of 2-ketoisovalerate (KIV) to isobutyralde-
hyde (Figure 1A).[2] Pyruvate is an upstream intermediate in
artificial isobutanol biosynthesis, and KIV is a key downstream
intermediate.[6] Generally, the reaction with pyruvate is favored
by wild-type AlsS, which shows poor conversion rates with KIV.
These factors may contribute to the observation that, in cell-
based isobutanol production attempts, wild-type AlsS did not
show any 2-ketoacid decarboxylase (KDC) activity.[13] Because
of its central role in synthetic isobutanol biosynthesis, enhanc-
ing the reactivity of AlsS towards KIV would potentially simplify
the cell-free reaction cascade because a specific KDC enzyme
activity could be eliminated.[6] Acetohydroxyacid synthase
(AHAS) is a related enzyme of the family; it catalyzes the same
reaction as ALS but in a different biochemical context. AHAS
catalyzes the first step in the branched-chain amino acid bio-
synthesis pathways of valine, leucine, and isoleucine, and it re-
quires flavin adenine dinucleotide (FAD).[14] However, FAD did
not appear to be involved in catalysis in Salmonella typhimuri-
um.[15] Prokaryotic AHASs and some eukaryotic AHASs have an
additional regulatory subunit.[16] This implies a different reac-
tion mechanism and structure–function relationship compared
to ALS enzymes. Crystal structures are only available for the
AHAS catalytic subunit from Saccharomyces cerevisiae (PDB ID:
1T9A)[17] and Arabidopsis thaliana (PDB ID: 1YHZ)
1
Chemicals produced from renewables that can serve as starting points for the
production of other functional compounds.
[a] Dr. B. Sommer, M. Haack, Dr. F. Qoura, Dr. D. Garbe, Prof. Dr. T. Brꢀck
Fachgebiet Industrielle Biokatalyse, Technische Universitꢁt Mꢀnchen
Lichtenbergstrasse 4, 85748 Garching (Germany)
E-mail: brueck@tum.de
[b] Dr. H. von Moeller, C. Langner, Dr. B. Loll
Institut fꢀr Chemie und Biochemie, Abteilung Strukturbiochemie
Freie Universitꢁt Berlin
Takustrasse 6, 14195 Berlin (Germany)
E-mail: loll@chemie.fu-berlin.de
[c] Dr. H. von Moeller
moloX GmbH
Takustrasse 6, 14195 Berlin (Germany)
[d] Dr. G. Bourenkov
European Molecular Biology Laboratory
Notkestrasse 85, 22603 Hamburg (Germany)
Our in-depth phylogenetic sequence analysis revealed that
the ALS enzymes represent a distinct subclass of the ThDP-de-
pendent enzyme family, of which the carboxylases are arche-
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cbic.201402541.
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 10
&
1
&
ÞÞ
These are not the final page numbers!

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
family, by using a combined structural and biochemical ap-
proach. For the first time, we were able to elucidate the crystal
structure of AlsS in the resting state and in a catalytically rele-
vant transition state by X-ray crystallography. Structural data
were used for targeted mutagenesis. Although the reaction
mechanism and structure–function relationships of the arche-
typical ThDP-dependent enzyme family of carboxylases has
been studied extensively, we demonstrate that the ALS sub-
family differs significantly in structure and possibly in catalytic
mechanism.[20] Certainly, our data suggest that the structure–
function properties of AlsS not only differ from those of car-
boxylases but also show differences from the phylogenetically
related AlsK.[19]
Results and Discussion
Phylogenetic analysis of ThDP-dependent enzymes
AlsS is an important biocatalyst in isobutanol biosynthesis,
with dual catalytic activity: condensation of two pyruvate mol-
ecules to acetolactate, and conversion of KIV into isobutyralde-
hyde. This ability to catalyze both reactions is unique in the
ALS family. This suggests that AlsS might have a different
active-site architecture from that of other ThDP-dependent en-
zymes. We initially performed amino acid sequence alignments
of diverse ThDP-dependent enzymes, and obtained informa-
tion for a direct comparison of AlsS to the well-characterized
ThDP-dependent decarboxylases (Figure S1 in the Supporting
Information). For the latter family, a two-histidine motif (“HH-
motif”), a fully conserved aspartate, as well as glutamate have
been described.[21,22] These structural features are important for
the correct orientation of the pyruvate molecule (Fig-
ure S1),[21,22] and establish a catalytic triad involved in the pro-
tonation of the enamine intermediate.[21–24] Our sequence
alignment revealed that this motif is absent in ALS enzymes.
Next, we performed a phylogenetic analysis of the ThDP-de-
pendent enzymes. To our surprise, the enzymes clustered into
three distinct sequence groups: acetolactate synthases, aceto-
hydroxyacid synthases, and carboxylases (Figure S2). The align-
ment indicated that ALS active-site residues differ significantly
from those of the well-studied carboxylase subgroup. It is also
interesting to note that AlsS and AlsK are in different branches
within the acetolactate synthase subgroup. Even though ALS
and AHAS catalyze the same reaction, this likely reflects their
different cofactors and domain structure: AHAS family enzymes
have both catalytic and regulatory subunits. To shed light on
the molecular architecture of AlsS, we set out to crystallize the
protein and to determine its molecular architecture.
Figure 1. Reactions and structure of AlsS. A) The main reaction is the con-
version of two molecules of pyruvate to acetolactate; the side reaction pro-
duces isobutyraldehyde. B) Overall structure of tetrameric AlsSwt (ribbon
presentation) with Mg2+ (orange spheres) and ThDP (stick representation).
C) Solvent-accessible channels in monomers C and D are drawn in surface
representation (beige). These point to the C2’ atom of the ThDP cofactor
and can be detected for all four monomers.
Crystallization of AlsS
typical representatives. To date, structural information on ALS
enzymes is limited to the structure of ALS from Klebsiella pneu-
moniae (AlsK, PDB IDs: 1OZF, 1OZG, 1OZH).[19] However struc-
ture-guided mutagenesis of AlsS could be a powerful tool to
improve the catalytic properties of AlsS towards its native sub-
strate pyruvate as well as KIV. In this study, we obtained valua-
ble insights into the reaction mechanism of the ALS enzyme
Wild-type AlsS (AlsSwt) and a C-terminal his-tagged variant
(AlsSC-His) were overexpressed in Escherichia coli and purified to
homogeneity for crystallization experiments. Prior to crystalliza-
tion, the proteins were incubated with an excess of ThDP and
Mg2+. Initially crystals of AlsSC-His were obtained in the tetrago-
nal space group P41212 (Table 1), and the crystal structure was
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 10
&
2
&
ÞÞ
These are not the final page numbers!

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
by the non-crystallographic symmetry (NCS) on a general sym-
metry position, and another four are at the intersections of the
NCS axis and the crystallographic twofold axes. The structure
of ALSwt was refined to R/Rfree of 0.172/0.217 with excellent
geometry (Table 1). The intensity modulation can be plausibly
explained by the small deviations between the unit cell direc-
tions and the directions of twofold axes in mixed NCS crystallo-
graphic tetramers. Structure refinement using the data inte-
grated in a double unit cell (a=111.3 ꢁ, b=340.0 ꢁ, c=
340.0 ꢁ; not shown) improved neither the refinement statistics
nor the quality of the electron density maps. A detailed de-
scription of this crystal will be published elsewhere.
Table 1. Data collection and refinement statistics.
Data collection
AlsSC-His-ThDP
AlsSwt-ThDP
AlsSwt-LThDP
PDB ID
4RJI
4RJJ
4RJK
P21221
0.9763
111.5, 170.8,
342.6
space group
wavelength [ꢁ]
unit cell a, b, c [ꢁ] 141.7, 141.7,
P41212
0.9184
P21221
0.9184
111.3, 170.0,
340.0
239.0
a, b, g [8]
90.0, 90.0, 90.0 90.0, 90.0, 90.0 90.0, 90.0, 90.0
resolution [ꢁ][a]
30.00–3.20
(3.39–3.20)
40142 (6254)
99.5 (98.9)
11.3 (2.5)
30.00–2.34
(2.48–2.34)
268024 (41131) 225545 (16544)
98.4 (94.3)
11.5 (3.0)
30.00–2.50
(2.56–2.50)
unique reflections
completeness[a]
99.8 (100.0)
15.4 (3.2)
0.133 (0.979)
99.8 (86.4)
13.3 (14.2)
[a]
hI/s(I)i
[a,b]
Rmeas
0.206 (0.879)
99.4 (81.5)
7.2 (7.1)
0.108 (0.490)
99.6 (80.6)
4.5 (4.1)
[a]
CC1/2
redundancy[a]
Refinement
non-H atoms
AlsS architecture
The apparent biological unit of AlsS is a homo-tetramer
formed by dimers of dimers (Figure 1B) as reported previously
for AlsK[19] and other ThDP-dependent enzymes such as pyru-
vate oxidase.[25] For a structural description, we shall focus on
the highest-resolution structure of the complete tetramer in
the asymmetric unit of the small unit cell of AlsSwt. Each mono-
mer has 571 amino acids. The 13 N-terminal residues and five
C-terminal residues could not be modeled due to the lack of
interpretable electron density. The data suggest that these resi-
dues are highly flexible and do not form ordered structures.
Each monomer is composed of three domains (Figure S3). The
a-domain (up to N181) is connected by a random coil to the
central b-domain (P195 to A346). The C-terminal g-domain
(from H376) is connected to the central b-domain by an a-
helix and a random coil linker. Overall, the monomers of the
tetramer are practically indistinguishable (RMSD<0.3 ꢁ). The
ALS enzymes of B. subtilis and K. pneumoniae share 51% se-
quence identity. For 512 Ca atoms, the structures show
RMSD=1.8 ꢁ (Figure S4).
16994
36644
35466
[a,c]
Rwork
Rfree
0.187 (0.252)
0.253 (0.329)
4
0.172 (0.233)
0.217 (0.297)
8
0.166 (0.204)
0.217 (0.280)
8
[a,d]
no. of protein
chains
Average B-factor [ꢁ2]
protein residues
ThDP molecules
LThDP molecule
Mg2+ cations
ligand molecules
water molecules
RMSD[e]
2208/64.5
4/74.1
–
4/59.6
1/84.4
23/26.1
4416/24.2
8/24.4
–
8/28.3
33/37.7
1955/24.9
4416/50.2
7/52.4
1/57.8
8/54.4
47/63.3
837/45.3
bond length [ꢁ]
bond angles [8]
Ramachandran
outliers [%]
0.005
0.890
0.009
1.177
0.010
1.220
0.1
0.1
0.4
97.0
favored [%]
97.8
98.0
[a] Values in parentheses refer to the highest-resolution shell. [b] Rmeas
=
Sh[n/(nꢀ1)]1/2 Si jIhꢀIh,i j/ShSiIh,i.where Ih is the mean intensity of symme-
try-equivalent reflections and n is the redundancy. [c] Rwork =Sh jFoꢀFc j
/SFo (working set, no s cut-off applied). [d] Rfree is the same as Rcryst, but
calculated on 5% of the data excluded from refinement. [e] RMSD from
target geometries.
Yet there are major structural differences in ALS between
B. subtilis and K. pneumoniae. A loop in the a-domain of AlsS
(K120 to S125) adopts a strikingly different conformation from
that in AlsK (Figure S5). Interestingly, the loop is two amino
acids longer in AlsS, and Q124 points to the active site. Nota-
bly, for AlsK the equivalent loop region (V118–Q120) is mod-
eled in the “resting state”,[19] not in the catalytically relevant
transition state. In contrast, for the AlsK structure, we could
completely model the linker between the a- and b-domain as
a random coil (Figure S4).
solved at medium resolution by molecular replacement with
a poly-alanine model derived from K. pneumoniae ALS.[19] The
asymmetric unit of AlsSC-His contains one homo-tetramer with
bound cofactors ThDP and Mg2+. Attempts to improve the dif-
fraction quality of these crystals failed. Therefore, we crystal-
lized untagged AlsSwt and incubated it with ThDP and Mg2+
under similar crystallization conditions. These crystals had a dif-
ferent crystal morphology compared to AlsSC-His, and diffracted
to 2.34 ꢁ resolution. Indexing of the diffraction data turned
out to be very difficult, with various possibilities in monoclinic,
orthorhombic, and tetragonal space groups. Finally, the diffrac-
tion patterns were indexed in the orthorhombic unit cell (a=
111.3 ꢁ, b=170.0 ꢁ, c=340.0 ꢁ). This indexing left out system-
atic weak reflections at positions h, k+1/2, 2L. The ratio of in-
tensities of neglected satellite reflections to regular reflections
was 0.033:1 on average in the resolution shell 30.0–2.5 ꢁ.
Molecular replacement using the coordinates of one proto-
mer of AlsSC-His derived from our medium-resolution data set as
search model in space group P21221 was successful. The unit
cell contains eight ALSwt tetramers, four of which are formed
AlsS and AlsK show additional differences in the linker
region connecting the a- and b-domains. These differences are
particularly obvious between Q365 and D373 (AlsS) and in
a loop region (R409 to T416) within the g-domain (Figure S4).
These differences result in distinct main-chain conformations,
but these are distant from the active site and hence most
likely have no functional implications. A comparison of the
amino acid sequences of the less-conserved C termini of AlsS
and AlsK reveals seven additional residues in AlsS. The 12 C-
terminal resolved residues of AlsS (D556–K567) fold into
a short a-helix (Figure S4) that is not present in AlsK. This a-
helix packs onto the b-domain and is solvent exposed on the
opposing site. This a-helix is an extension of the structurally
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 10
&
3
&
ÞÞ
These are not the final page numbers!

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
conserved a-helix that shields the ThDP and Mg2+ cofactors
from the solvent.
The active site
A substrate entrance channel extends 13 ꢁ from the solvent-
exposed surface to the active site (solvent-exposed diameter:
10–12 ꢁ, narrowing towards the active site; Figure 1C). The
wall of the channel comprises amino acid side chains of all
three domains, as well as from both monomers of a functional
dimer. The active site includes an Mg2+ ion and the ThDP co-
factor (Figure 2A and Figure S6). The octahedral coordination
of Mg2+ is established by two hydroxyl functions (a- and b-
phosphate of ThDP), the carboxylates of D451 and D478, the
backbone carbonyl of T480, and one water molecule (Fig-
ure 2A).
In other ThDP-dependent enzymes, Mg2+ is coordinated by
an asparagine at the equivalent position to D478.[26] Amino
acid sequence analysis revealed that aspartate is strictly con-
served in all known ALS sequences, and therefore seems to be
characteristic for this subclass of ThDP-dependent enzyme.
ThDP and Mg2+ were bound to all eight monomers, and
hence we conclude that the structure represents the resting
state of the enzyme. The four active sites of AlsS are at the
two dimer interfaces (Figures 1B, C and 2). The Mg2+ cation is
coordinated to residues of a single monomer, whereas the thi-
amine function of ThDP additionally interacts with residues of
the adjacent monomer (Figure 2A). The ThDP is bound in a “V”
conformation, maintained by the hydrophobic side chain of
AlsS L426’ from the neighboring subunit. The V conformation
provides close proximity (3.1 ꢁ) between the 4’-amino nitrogen
and the C2’ atom of ThDP. The pyrimidine function is stabilized
by two hydrogen bonds. The first is to the carboxylate of E61’,
which is part of the adjacent monomer (Figure 2A). This inter-
action has been suggested to assist in the formation of the 4’-
imino form of ThDP (Figure 2A).[27,28] The second hydrogen
bond is established by the backbone carbonyl of Q424, there-
by stabilizing the 4’-amino/imino group, which is then oriented
for deprotonation of the adjacent C2’ atom. This results in the
reactive species (Figure 2A), which initiates the catalytic cycle.
The phosphates of ThDP are stabilized by hydrogen bonds to
protein side chains and backbone amides. Moreover, the N-ter-
minal ends of three a-helices (with their partial positive
charge) point to the phosphates, thus conferring additional
stabilization. Surprisingly there are no water molecules closer
than 3.7 ꢁ to the thiamine function of ThDP.
Figure 2. A) Active site of AlsS with bound ThDP and Mg2+. Coordination of
Mg2+ and interaction of ThDP with amino acids from two monomers lining
the active site. Mg2+ and water molecules are drawn as spheres in orange
and red, respectively. Grey lines indicate the coordination sphere of Mg2+
.
Hydrogen bonds are drawn as dashed red lines (distance cut off 3.5 ꢁ). B) As
above but with bound LThDP and Mg2+. C) Amino acid residues at the ac-
tive site that have been subjected to site-directed mutagenesis.
ThDP. There are no direct interactions to the protein backbone,
but rather indirect contacts (via water molecules). The hydroxy
group of T84 interacts via a water molecule with the terminal
hydroxy group of the PEG molecule (3.2 ꢁ). Given the proximi-
ty of PEG to ThDP and the side chain of T84, we investigated
the potential importance of this residue for correct positioning
of the second pyruvate molecule.
A poly(ethylene glycol) (PEG) molecule in the vicinity of the
active site
Interestingly, in two monomers of AlsSwt we observed elongat-
ed electron densities pointing to the C2’ atom of the ThDP
thiazole ring (Figure S7). We interpreted the electron density as
PEG 200 molecules from the crystallization solution. It is some-
how surprising to observe PEG molecules in an extended con-
formation, as they tend to adopt a horseshoe-like shape. The
terminal hydroxyl of PEG 200 is 3.5 ꢁ from the thiazole ring of
Catalytic aspects of structural features
To obtain structural information relevant to the catalytic cycle
of AlsS, we soaked AlsSwt crystals in solutions of pyruvate. We
observed the evolution of gas bubbles at the surface of the
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 10
&
4
&
ÞÞ
These are not the final page numbers!

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
crystals. Most likely the gas bubbles were CO2 from the decar-
boxylation step of the reaction cycle. We determined the crys-
tal structure of pyruvate-soaked crystals to 2.50 ꢁ and refined
them to R/Rfree of 0.166/0.217. In simulated annealing with
2FoꢀFc and FoꢀFc electron-density maps, we observed addi-
tional electron density in close vicinity to the C2’ atom of
ThDP. The shape and level of the electron density differed be-
tween the monomers. However, for one monomer it was di-
rectly connected to the C2’ atom of the thiazolium ring. The
connecting electron density perpendicular to the thiazolium
ring of ThDP was unambiguously identified as a 2-lactyl moiety
bound to ThDP (LThDP) with an sp3-hybridized Ca atom (Fig-
ure 2B). Our structure of AlsSwt with LThDP represents a catalyti-
cally relevant transition state. This is the first time that the
structure of any catalytically relevant reaction intermediate has
been reported for any member of the ALS enzyme family, al-
though it has been reported for other ThDP-depending en-
zymes.[21,24,29] Examination of AlsSwt with bound LThDP indicat-
ed that the 2-lactyl moiety bound to ThDP is stabilized by sev-
eral interactions to protein side chains and to the core ThDP
moiety. Interestingly, Q124 (pointing to the active site cavity)
forms a bifurcated hydrogen bond to the carboxylate and
hydroxyl of LThDP (Figure 2B). Furthermore, the hydroxyl of
LThDP forms a short intermolecular hydrogen bond to the 4’-
amino group of the pyrimidine (Figure 2B). The pyruvate-de-
rived methyl points towards M483. This overall arrangement is
very similar to that of pyruvate oxidase.[29] Kinetic studies of
other ThDP-dependent enzymes revealed that the 4’-amino
group of the pyrimidine moiety acts as the proton donor for
pyruvate.[30] The short hydrogen bond observed in our struc-
ture is consistent with this function of the 4’-amino group.
Wille et al. proposed that a distortion of the aromatic thiazoli-
um ring might assists in the decarboxylation reaction.[29] Be-
cause of the limited resolution we could not detect this ring
distortion in our crystal structure. Notably, no major conforma-
tional rearrangements occurred to amino acid side chains in
the active side or to the bound cofactor upon binding of pyru-
vate to ThDP. We could not detect any pyruvate molecule in
close proximity to the active site or in the channel. Instead,
two molecules of pyruvate were identified on the protein sur-
face.
tate. The catalytic activity of LDH can be measured photomet-
rically (340 nm) by following NADH depletion. Hence, the
amount of consumed pyruvate by AlsS can be calculated. The
assay conditions for both AlsS and LDH are flexible and can be
readily adapted; however, the LDH assay temperature must be
maintained at 258C. We performed a detailed characterization
of AlsSC-His to determine its temperature optimum, its half-life
at 508C, its solvent stability at pH 7.0, and its pH tolerance at
508C. Incubation at 508C and pH 7.0 was chosen for the subse-
quent assays, as these conditions were well tolerated by all
enzyme components of the cell-free isobutanol biosynthesis
pathway.[6]
Activity of AlsS variants towards the alternative substrate
KIV
The activities of AlsS towards both pyruvate and KIV are rele-
vant for reducing the complexity of cell-free isobutanol pro-
duction (Figure 1A). Therefore, increasing the activity towards
both was an objective of our study. Thus, information on
amino acid residues critical for the decarboxylation step with
KIV is required. Atsumi et al. reported 5.5 Umgꢀ1 for the KDC
activity of AlsSwt.[2] In this study we could not reproduce these
results under equivalent experimental conditions (37–408C);
we observed 1000-fold lower activity (maximum 53 mUmgꢀ1).
Subsequently, we prepared a set of mutants that are described
in detail below. The lower KM of AlsSQ487S (154ꢁ21 mm) com-
pared to that of AlsSC-His (300ꢁ35 mm) did not result in a
higher activity towards KIV (Table 2). However AlsSK40I showed
a slightly improved activity (59 mUmgꢀ1) towards KIV. AlsS cat-
alyses the conversion of pyruvate to acetolactate as well as the
decarboxylation of KIV to isobutyraldehyde, albeit at a much
lower rate (Figure 1A). As both reactions are required for the
designed isobutanol synthesis, increasing the AlsS activity to-
wards KIV could significantly lower the overall reaction com-
plexity. To determine KIV specific activity, a coupled assay of
AlsS with alcohol dehydrogenase (ADH) was applied. In this
assay, KIV is converted to isobutyraldehyde by AlsS, then isobu-
tyraldehyde is converted to isobutanol by ADH. This reaction
can be measured photometrically at 340 nm because ADH
Table 2. Activities and half-lives of all AlsSC-His variants at 508C with pyru-
vate as substrate. Activities of all AlsSC-His variants towards KIV are also re-
ported.
Biochemical characterization of AlsSC-His
In baseline measurements of ALS variants, we verified that
AlsSwt and AlsSC-His have identical activities. AlsS is a key
enzyme in a recently published cell-free isobutanol biosynthe-
sis pathway.[6] To assess the performance of these AlsS variants
in isobutanol production, the stability and activity of each
enzyme in the presence of process-relevant alcohols (isobuta-
nol, butanol, and ethanol) were measured. To determine the
catalytic activity of AlsS towards pyruvate we applied a previ-
ously reported coupled assay with lactate dehydrogenase
(LDH).[31] In brief, the reactions take place in two steps. In the
first step AlsS converts pyruvate to acetolactate. At intervals,
samples are taken and immediately added to the second step
mixture, in which the remaining pyruvate is converted to lac-
Protein
A [Umgꢀ1] at 508C t1/2 [h] at 508C AKIV [Umgꢀ1] at 408C
AlsSHisC
24ꢁ4.00
4ꢁ0.20
3ꢁ0.10
7ꢁ0.60
1ꢁ0.08
18ꢁ2.54
6ꢁ0.50
27ꢁ5.00
2ꢁ2.00
1ꢁ0.06
–
81
44
89
110
2.5
33
42
104
94
19
–
0.053ꢁ0.0020
0.045ꢁ0.0080
0.059ꢁ0.0020
0.031ꢁ0.0020
0.021ꢁ0.0020
0.020ꢁ0.0030
0.014ꢁ0.0003
0.028ꢁ0.0020
0.016ꢁ0.0010
0.004ꢁ0.0010
0.009ꢁ0.0020
0.046ꢁ0.0030
AlsSK40H
AlsSK40I
AlsSK40Y
AlsST84V
AlsSP87A
AlsSQ124S
AlsSQ424S
AlsSQ424S/Q487S
AlsSY481A
AlsSM483N
AlsSQ487S
35ꢁ2.00
22
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 10
&
5
&
ÞÞ
These are not the final page numbers!

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
concomitantly oxidizes NADH. Because of the moderate ther-
mostability of horse liver ADH used in our assay, all KIV assays
were carried out at 408C. Under these conditions AlsSC-His
showed 53 mUmgꢀ1.[32]
Determination of pH and temperature optima for AlsSC-His
At pH 4.0 (and 508C) no activity was observed, but an activity
profile was obtained over pH 5.0–6.0, with a plateau at pH 7.0–
8.0; the activity reduced over pH 9.0–11.0 and at pH 11 no ac-
tivity was detected (Figure 3A). Thus, AlsSC-His showed excellent
Figure 4. Solvent tolerance of AlsSC-His at 508C and pH 7.0 for A) 0–6% (v/v)
&
&
n-butanol ( ) and isobutanol ( ) and B) 0–20% (v/v) ethanol.
of 3% (v/v) isobutanol and n-butanol, and of up to ~13% (v/v)
ethanol (Figure 4B).
To our knowledge this is the first account of isobutanol tol-
erance by any member of the ALS family. Our ethanol toler-
ance results for AlsS mirror those obtained for AlsE (~50% ac-
tivity up 15% (v/v)).[35]
Figure 3. Determination of pH stability and temperature optimum of
AlsSC-His. A) 50 mm sodium acetate (pH 4.0 and 5.5), 50 mm sodium phos-
phate (pH 6.0 to 6.5), 50 mm Tris (pH 7.0–9.0), or 50 mm CAPS (pH 10 and
11); all measurements at 508C. B) Enzyme assays for temperature optimum
determination were performed from 20 to 658C at pH 7.0. Data are mean of
three replicates; 100% activity is the highest measured activity in the experi-
ment.
Structure-guided mutagenesis strategy to generate AlsS
variants
An examination of the AlsS active-site structure indicates that
only a few residues are in proximity of the bound ThDP cofac-
tor or the reaction intermediate LThDP (Figure S8). Interesting-
ly, the AlsS residues are substantially different from their equiv-
alents in ThDP-dependent enzymes involved in the decarboxy-
lation of a-ketoacids; these have been studied in great
detail.[24] The degree of conservation within ALS enzymes sug-
gests that ALS enzymes constitute a distinct subclass of ThDP-
dependent enzymes. Based on structural assessment we select-
ed eight amino acid residues in the vicinity of the ThDP bind-
ing site, for targeted single- or double-mutation replacement
(Figure 2C). Even though AlsS and AlsK share only 55% amino
acid identity, all our mutated amino acids are conserved in
AlsK. A primary target was Q124, which is within hydrogen-
bonding distance of LThDP (Figure 2A and B). Interestingly, the
side chain of Q424 points directly towards LThDP; this indi-
cates an involvement in the positioning the second pyruvate
moiety in the active site to facilitate the second condensation
reaction that yields acetolactate. Other mutagenesis targets
were M483 and Q487 because their side chains are on an a-
helix and separated by one a-helical turn (Figure S8). Hence,
both side chains point iin the same direction, towards the co-
factor. Another target was Y481, on the opposite site of the
catalytically active C2’ atom of ThDP. Therefore Y481 might be
important for the accurate positioning of ThDP. Additionally,
we selected K40, whose homologue in AlsK (K36) is thought to
be involved in catalysis.[19] Amino acid residue replacement
was achieved by site-specific mutagenesis. All AlsS variants
were expressed with a C-terminal His6-tag and characterized
for activity towards pyruvate at 508C and pH 7.0 (HEPES
buffer), as this pH is the most compatible with the enzyme sys-
tems involved in our artificial, cell-free isobutanol cascade.[6]
activity over pH 5.5–6.5 with an optimum at pH 6.0. The activi-
ty at pH 6.0 was six times higher than at pH 7.0. In the context
of cell-free isobutanol biosynthesis, the pH optimum at pH 6.0
is a prerequisite for subsequent decarboxylation of acetolac-
tate by action of an acetolactate decarboxylase, or by a low
pH in the cell-based 2,3-butanediol production pathway.[11] This
AlsSC-His pH profile is consistent with data observed for other
members of ALS enzyme family.[33,34]
At the optimum temperature (508C) the activity of AlsSC-His
(24ꢁ4 Umgꢀ1 at pH 7.0) was almost three times higher than
at 20, 30, and 658C (Figure 3B). At 508C and pH 7.0, we mea-
sured an AlsSC-His half-life of 81 h, and it was 16 h at 608C. The
stability at 608C is rather surprising, as B. subtilis is a mesophilic
bacterium; the optimal temperature for ALS derived from the
mesophilic bacterium Enterococcus faecalis (AlsE) has been re-
ported to be 378C.[32] In contrast, to AlsS, AlsE shows no tem-
perature stability at temperatures above 408C.[32] The extended
temperature stability observed for AlsS might be due to its
specific architecture.
Determination of AlsSC-His solvent tolerance
The production of solvents like isobutanol implicitly requires
the enzyme to tolerate the end-products. In cell-based sys-
tems, 1–2% (v/v) isobutanol induced toxic effects in the host.[6]
Therefore, in our assay, AlsSC-His activity was determined at 0–
6% (v/v) isobutanol and n-butanol, and 0–20% (v/v) ethanol
(Figure 4A). The enzyme retained 50% activity in the presence
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 10
&
6
&
ÞÞ
These are not the final page numbers!

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
structure of AlsS bound to LThDP, the 2-lactyl moiety is stabi-
lized by the side chain of Q124 (Figure 2B). The importance of
this for catalysis was confirmed by the drastically decreased
activity of AlsSQ124S (Figure 1C and Table 2).
Conclusion
The substrate promiscuity of AlsS towards pyruvate and KIV
makes it potentially ideal for a minimized enzyme reaction cas-
cade. Initially, the activities of wild-type AlsS towards pyruvate
and KIV were determined: the activity towards KIV
In the AlsK reaction model, K36 has been proposed to be
important for recognition and correct positioning of the
second pyruvate molecule.[19] The corresponding residue in
AlsS (K40) has an identical side-chain orientation. In other
ThDP-dependent carboxylases, tyrosine or phenylalanine are
frequently at this sequence position. Replacement of K40 (Fig-
ure 1C) by the hydrophobic amino acid isoleucine (AlsSK40I),
basic histidine (AlsSK40H), or tyrosine (AlsSK40Y) greatly reduced
activity (Table 2).
(53 mUmgꢀ1
)
was
a
side reaction compared to the
2400 mUmgꢀ1 measured for pyruvate. AlsSwt had a pH opti-
mum at pH 6.0 and exhibited maximum activity at 508C. In
addition, AlsSwt tolerated higher product concentrations (50%
activity in the presence of 3% (v/v) isobutanol/n-butanol, or
13% (v/v) ethanol) than did microbial hosts.
We also obtained detailed structural information on AlsS
and its catalytic intermediate. The crystal structures allowed us
to identify catalytically important residues in the active site of
AlsS. Based on this information we generated AlsS variants,
which were characterized for activity towards pyruvate and KIV
(Table 2).
As proposed for other ThDP-dependent enzymes,[36] the cat-
alytic cycle is initiated by proton abstraction from the 1’,4’-imi-
nopyrmidine tautomer of the ThDP cofactor by the conserved
aspartate (E61 in AlsS) to yield the reactive ThDP. For other
ThDP-dependent enzymes the HH-motif, a conserved aspar-
tate, and a glutamate have been described as important for
the correct positioning of the pyruvate molecule (Fig-
ure S1).[21,22] It had been proposed that the signature motif
forms a catalytic triad involved in the protonation of the enam-
ine intermediate of Zymomonas mobilis pyruvate decarboxy-
lase (Figure 5 and Figure S1).[21,23,24]
The proximity of a PEG 200 molecule to ThDP and the hy-
droxyl function of T84 inspired variant AlsST84V in order to in-
vestigate whether this residue is involved in substrate position-
ing. AlsST84V cannot establish a hydrogen bond to PEG 200. In
our enzymatic assays, the enzymatic activity was almost com-
pletely abolished, thus highlighting the importance of this resi-
due.
AlsSY481A showed almost complete loss of activity (Table 2).
This residue not only lines the active site, but, with its hydroxyl
group, it is hydrogen bonded to the carbonyl backbone of
A58, thereby clamping the b- and a-domains to each other.
Disruption of this interaction does not only locally affect the
cofactor binding site but most likely the overall stability of
AlsS. The pyrimidine function of ThDP faces P87 at the N termi-
nus of an a-helix (Figure 1C). Mutation to alanine reduced the
activity by about one third (Table 2), likely caused by reorienta-
tion of ThDP. Q424 and Q487 are on opposing sides of the
tunnel leading to the active site. The side chains of these resi-
dues are oriented towards the C2’ of ThDP or the 2-lactyl
moiety of LThDP, and are potentially important for recognition
and positioning of the second incoming pyruvate molecule.
The backbone carbonyl of Q424 is hydrogen bonded to ThDP
(Figure 2A and B), but there is no direct interaction between
Q487 and ThDP. The side chains of Q424 and Q487 are 4.6 and
3.8 ꢁ from the carboxylate function of LThDP. Therefore both
amino acid residues were subjected to mutagenesis. Surpris-
ingly, we detected increased activities towards pyruvate when
Q424 or Q487 was mutated to serine (Table 2), although the
double mutant AlsSQ424S/Q487S had only residual activity (Table 2),
thus supporting the hypothesis that at least one residue is
important for positioning the second pyruvate molecule.
Our structural findings are in agreement with our phyloge-
netic analysis of ThDP-dependent enzymes (Figure S2). The
AlsS enzyme clusters in a distinct subgroup, with different
amino acids at its active site from those in other ThDP-depen-
dent enzymes. Our data show for the first time that the ALS
subfamily of ThDP-dependent enzymes differ in both structure
and function compared to the archetypical carboxylases and
the AHAS enzyme family. Whether or not these structure–func-
tion differences have a significant impact on the general cata-
lytic mechanism of ThDP-dependent enzymes, models for
which have been developed based on the carboxylase, remains
to be elucidated.
The active site residues are different in AlsS compared to
other ThDP-dependent enzymes. AlsS has A39 at the equiva-
lent position of the conserved aspartate (E27) of Z. mobilis de-
carboxylase (Figure 5); this is conserved across other ALS en-
zymes (Figure S1). The HH-motif is not present in ALS enzymes.
In place of the second histidine, AlsS has Q124, again con-
served within the ALS family (Figure S1), and the flanking re-
gions of the HH-motif in ALS enzymes (Figure 5) are longer in
other ThDP-dependent enzymes. A loop region in our AlsS
structure (L92 to S125) differs from that in AlsK and allows
Q124 to approach the active site (Figure S5). In the crystal
Figure 5. Superposition of active site residues of AlsS (green and violet) and
pyruvate decarboxylase (PDZ, gray) of Z. mobilis (PDB ID: 3OE1),[21] with
bound LThDP cofactor (AlsS: sand; PDZ: gray). The HH-motif of Z. mobilis
PDZ is formed by H113 and H114.
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 10
&
7
&
ÞÞ
These are not the final page numbers!

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
with a reservoir solution supplemented with pyruvate (2 mm).
After 3 s the crystals were flash frozen in liquid nitrogen.
Experimental Section
Reagents and bacterial strains: Restriction enzymes, T4 ligase,
desoxynucleotides, phusion polymerase, DNA and protein-stand-
ards were purchased from Thermo Scientific; Klenow fragment and
T4 kinase were from New England Biolabs. DNA sequencing and
oligonucleotides synthesis were performed by Eurofins MWG
Operon (Ebersberg, Germany). The acetolactate synthase gene
from B. subtilis was synthesized by Geneart (Life Technology, Re-
gensburg, Germany) with optimized Escherichia coli codon usage.
Horse liver ADH was obtained from evocatal (Dꢀsseldorf, Germany).
Porcine heart LDH was from Serva (Heidelberg, Germany). 3-
Methyl-2-oxobutanoic acid, thiamine diphosphate, and isobutyral-
dehyde were purchased from Sigma–Aldrich. Other chemicals were
purchased from Carl Roth or from AppliChem (Darmstadt, Germa-
ny). pET28a was purchased from Merck Millipore. E. coli
HMS174(DE3) was purchased from Merck Millipore, E. coli XL1-Blue
was from Stratagene (Agilent Technologies).
X-ray data collection, structure determination, and refinement:
Synchrotron diffraction data were collected at MX Beamline 14.2 at
BESSY (Berlin, Germany) or Beamlines P13 and P14 of Petra III
(Deutsches Elektronen Synchrotron, Hamburg, Germany). X-ray
data collection was performed at 100 K. Diffraction data were pro-
cessed with the XDS package (Table 1).[38,39] For calculation of the
free R-factor, a randomly generated set of 5% of the reflections
from the diffraction data set was used and excluded from the re-
finement. Initial phases of the AlsSC-His were determined by molecu-
lar replacement with PHASER[40] and a search model of one mono-
mer derived from the coordinates of AlsK (PDB ID: 1OZF)[19] that
had been side-chain and cofactor depleted. An initial model was
built with the program BUCCANEER.[41] The structure was initially
refined by applying a simulated annealing protocol, and in subse-
quent refinement cycles maximum-likelihood restrained refinement
in PHENIX[42] was used, followed by iterative, manual model build-
ing with COOT.[43] AlsSwt diffraction data were analyzed with PHE-
NIX.XTRIAGE.[44] The structure of AlsSwt in the small unit cell was
solved by molecular replacement with one protomer of the initially
built AlsSC-His with the program PHASER.[40] The structure of AlsSwt
was refined with PHENIX.REFINE.[42] Water molecules were located
with COOT[43] and manually inspected. Model quality was evaluated
with PROCHECK[45] and MolProbity.[46] Figures were prepared using
PyMOL.[47] The active-site tunnel was calculated with the PyMOL
plug-in CAVER.[48] The atomic coordinates and structure factor am-
plitudes have been deposited in the Protein Data Bank (IDs: 4RJI,
4RJJ, and 4RJK).
Cloning: B. subtilis alsS was cloned into the vectors pCBRHisC
(AlsSC-His) and pCBRHisNo (AlsSwt) at BfuaI and BsaI restriction
sites.[6] The exchanges were done using the megaprimer PCR
method (Table S1).[37] PCR products were digested with XhoI and
NcoI and ligated into pET28a.
Protein expression and purification for crystallization: Plasmids
coding for AlsSwt or AlsSC-His were transformed into E. coli BL21(DE3)
pLys Express cells (New England Biolabs). AlsS was cultured in 2ꢃ
YT medium at 378C until an OD600 ~1.3 was reached, and subse-
quently cooled to 188C. Protein expression was induced by addi-
tion of isopropyl-b-d-thiogalactopyranoside (IPTG, 0.5 mm). Cells
were grown overnight and were harvested by centrifugation
(9000g, 10 min, 48C). The pellet was resuspended in Tris·HCl
(50 mm pH 8.0) containing NaCl (500 mm) and MgCl2 (1 mm). The
cells were lysed by sonication at 48C, and the supernatant was
cleared by centrifugation (48000g, 1 h, 48C). Ni2+-NTA (~1 mL; GE
Healthcare) was equilibrated with Tris·HCl (20 mm, pH 8.0) contain-
ing NaCl (500 mm), MgCl2 (1 mm), imidazole (10 mm), and DTT
(2 mm). AlsSC-His was loaded on the column and washed with equi-
libration buffer (3 cv). AlsSC-His was eluted in Tris·HCl (20 mm,
pH 8.0) containing NaCl (500 mm), MgCl2 (1 mm), DTT (2 mm), and
imidazole (500 mm). The cleared AlsSwt supernatant was diluted
(1:10) with Tris·HCl (50 mm, pH 8.0) containing MgCl2 (1 mm), and
loaded on a 1 mL HiTrap Q XL column (GE Healthcare). AlsSwt was
eluted in a linear NaCl gradient (100–1000 mm). The pooled frac-
tions were diluted (1:10) in Tris·HCl (50 mm, pH 8.0) with MgCl2
(1 mm) and loaded on a MonoQ 10/100 column (GE Healthcare).
AlsSwt was eluted in a linear NaCl gradient (100–1000 mm). Size-ex-
clusion chromatography of AlsSwt and AlsSC-His was performed with
a HighLoad Superdex S200 16/60 column (GE Healthcare) equili-
brated with Tris·HCl (50 mm, pH 8.0) containing NaCl (100 mm),
MgCl2 (1 mm), and DTT (2 mm). Pooled fractions were concentrat-
ed with Amicon-Ultra 30000 filters to 35 mgmLꢀ1 as determined
by absorbance at 280 nm.
Heterologous expression and enzyme purification for enzymatic
assays: AlsSC-His and all AlsSC-His variant expression was performed
in E. coli HMS174(DE3) cells in TB medium supplemented with
kanamycin (50 mgmLꢀ1). After inoculation cells were grown to
OD600 =0.5–0.7 at 378C and subsequently induced with IPTG
(1 mm). The temperature was decreased to 168C for further cultur-
ing (48 h). Cell lysates were prepared with and Emulsiflex-B15 cell
disruptor (Avestin, Mannheim, Germany), and cell debris was re-
moved by centrifugation (25000g, 20 min, 48C). Protein purifica-
tion was performed with Ni2+-NTA. Desalting of the proteins was
performed with PD-10 columns (GE Healthcare). All variants were
purified as described for AlsSC-His. AlsSC-His and its variants were
stored at ꢀ808C with the buffer supplemented with glycerol 10%
(v/v). Protein concentration was measured at 280 nm with unfold-
ed protein in 8m urea.[49] The extinction coefficient was calculated
using the ExPASY ProtParam tool.[50]
Enzyme assays: Pyruvate and KIV assays were performed by UV/
Vis spectrophotometry in 96-well microtiter plates with an En-
Spire 2 plate reader (PerkinElmer). The buffer pH was adjusted to
the corresponding temperature according to Stoll and Blan-
chard.[51] One unit of enzyme activity is defined as the amount of
enzyme necessary to convert 1 mmol substrate per minute. For the
pyruvate assay, reaction mixtures in HEPES (50 mm, pH 7), ThDP
(0.1 mm), MgCl2 (10 mm), and sodium pyruvate (15 mm) were incu-
bated at RT. The activity of AlsS/variant was measured by assaying
the pyruvate consumption at 508C (in a water bath), with standard
LDH assay triplicates.[31] Samples (4 mL) of each assay were taken
every five minutes. The remaining pyruvate in each sample was im-
mediately converted to lactate by LDH. The aliquot (4 mL, =0.3 mm
sodium pyruvate in the control (without enzyme)) was transferred
to the LDH assay mixture at RT (HEPES (25 mm, pH 7.4), NADH
(0.3 mm), sodium pyruvate (0–0.3 mm)). NADH absorbance was de-
termined at 340 nm. For thermal stability assays, AlsS was stored
Crystallization and crystal cooling: For cofactor incorporation,
AlsSwt or AlsSC-His (35 mgmLꢀ1) was incubated (3.5 h, RT) with ThDP
(1 mm) and Mg2+ (2 mm), then used for crystallization experiments
by the sitting drop method with a reservoir solution of polyethy-
lene glycol 300 (30%, v/v), CaAc (200 mm), and sodium cacodylate
(100 mm, pH 6.5). Crystals of AlsSC-His appeared after one day. AlsSwt
was crystallized with PEG 200 (45%, v/v), Tris·HCl (100 mm, pH 7.0),
and Li2SO4 (50 mm). The crystals were flash cooled in liquid nitro-
gen without cryo-protectant. Soaking experiments were performed
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 10
&
8
&
ÞÞ
These are not the final page numbers!

CHEMBIOCHEM
FULL PAPERS
www.chembiochem.org
[22] F. H. Andrews, M. J. McLeish, Bioorg. Chem. 2012, 43, 26–36.
[23] A. Schꢀtz, R. Golbik, S. Kçnig, G. Hꢀbner, K. Tittmann, Biochemistry
2005, 44, 6164–6179.
[24] R. Kluger, K. Tittmann, Chem. Rev. 2008, 108, 1797–1833.
[25] Y. A. Muller, G. E. Schulz, Science 1993, 259, 965–967.
[26] J. M. Candy, R. G. Duggleby, Biochim. Biophys. Acta Protein Struct. Mol.
Enzymol. 1998, 1385, 323–338.
[27] Y. Lindqvist, G. Schneider, U. Ermler, M. Sundstrçm, EMBO J. 1992, 11,
2373–2379.
over several days at 508C in a water bath. For solvent stability
assays, the previously standardized reaction mixture contained the
tested amount of 1-butanol, ethanol, or isobutanol. All assays were
performed at 508C. Determination of optimum pH was performed
with either sodium acetate (pH 4.0–5.5), sodium phosphate (pH 6.0
and 6.5), Tris (pH 7.0–9.0), or CAPS (pH 10.0 and 11.0) at 508C. For
the KIV assay, the reaction mixtures contained HEPES (50 mm,
pH 7.0), ThDP (0.1 mm), MgCl2 (2.5 mm), KIV (30 mm), NADH
(0.3 mm), and ADH (0.25 U/mL). The activity of AlsS towards KIV
was assayed at 408C with detection at 340 nm.
[28] R. Breslow, J. Am. Chem. Soc. 1957, 79, 1762–1763.
[29] G. Wille, D. Meyer, A. Steinmetz, E. Hinze, R. Golbik, K. Tittmann, Nat.
Chem. Biol. 2006, 2, 324–328.
[30] K. Tittmann, R. Golbik, K. Uhlemann, L. Khailova, G. Schneider, M. Patel,
F. Jordan, D. M. Chipman, R. G. Duggleby, G. Hꢀbner, Biochemistry 2003,
42, 7885–7891.
Acknowledgements
[31] S. Wolterink-van Loo, A. van Eerde, M. A. J. Siemerink, J. Akerboom,
B. W. Dijkstra, J. van der Oost, Biochem. J. 2007, 403, 421–430.
[32] S.-C. Lee, J. Kim, I.-J. La, S.-K. Kim, M.-Y. Yoon, Enzyme Microb. Technol.
2013, 52, 54–59.
We are grateful to Claudia Alings for excellent technical support.
We thank Thomas Schneider for access and support at beamline
P14 of PETRA III (Deutsches Elektronen Synchrotron, Hamburg,
Germany). We accessed beamlines of the BESSY II (Berliner Elek-
tronenspeicherring-Gesellschaft fꢀr Synchrotronstrahlung II) stor-
age ring (Berlin, Germany) via the Joint Berlin MX-Laboratory
sponsored by the Helmholtz Zentrum Berlin fꢀr Materialien und
Energie, the Freie Universitꢁt Berlin, the Humboldt-Universitꢁt zu
Berlin, the Max-Delbrꢀck Centrum, and the Leibniz-Institut fꢀr
Molekulare Pharmakologie.
[33] V. Gedi, M.-Y. Yoon, FEBS J. 2012, 279, 946–963.
[34] W. D. Holtzclaw, L. F. Chapman, J. Bacteriol. 1975, 121, 917–922.
[35] C. L. Young, Z. T. Britton, A. S. Robinson, Biotechnol. J. 2012, 7, 620–634.
[36] R. L. Schowen, Thiamin-Dependent Enzymes, Academic Press, London,
1998.
[37] S. Barik in Methods in Molecular Biology, Vol. 57: Site-Directed Mutagene-
sis in Vitro by Megaprimer PCR (Ed.: M. K. Trower), Humana, Totowa,
1996, pp. 203–215.
[38] W. Kabsch, Acta Crystallogr. Sect. D Biol. Crystallogr. 2010, 66, 125–132.
[39] M. Krug, M. S. Weiss, U. Heinemann, U. Mueller, J. Appl. Crystallogr.
2012, 45, 568–572.
Keywords: biofuels · biosynthesis · biotechnology · enzyme
catalysis · isobutanol
[40] A. J. McCoy, R. W. Grosse-Kunstleve, P. D. Adams, M. D. Winn, L. C. Storo-
ni, R. J. Read, J. Appl. Crystallogr. 2007, 40, 658–674.
[41] K. Cowtan, Acta Crystallogr. Sect. D Biol. Crystallogr. 2008, 64, 83–89.
[42] P. V. Afonine, R. W. Grosse-Kunstleve, N. Echols, J. J. Headd, N. W. Moriar-
ty, M. Mustyakimov, T. C. Terwilliger, A. Urzhumtsev, P. H. Zwart, P. D.
Adams, Acta Crystallogr. Sect. D Biol. Crystallogr. 2012, 68, 352–367.
[43] P. Emsley, B. Lohkamp, W. G. Scott, K. Cowtan, Acta Crystallogr. Sect. D
Biol. Crystallogr. 2010, 66, 486–501.
[44] P. D. Adams, P. V. Afonine, G. Bunkꢄczi, V. B. Chen, I. W. Davis, N. Echols,
J. J. Headd, L.-W. Hung, G. J. Kapral, R. W. Grosse-Kunstleve, A. J. McCoy,
N. W. Moriarty, R. Oeffner, R. J. Read, D. C. Richardson, J. S. Richardson,
T. C. Terwilliger, P. H. Zwart, Acta Crystallogr. Sect. D Biol. Crystallogr.
2010, 66, 213–221.
[45] R. A. Laskowski, M. W. McArthur, D. S. Moss, J. M. Thornton, J. Appl. Crys-
tallogr. 1993, 26, 283–291.
[46] V. B. Chen, W. B. Arendall III, J. J. Headd, D. A. Keedy, R. M. Immormino,
G. J. Kapral, L. W. Murray, J. S. Richardson, D. C. Richardson, Acta Crystal-
logr. Sect. D Biol. Crystallogr. 2010, 66, 12–21.
[1] S. Atsumi, T. Hanai, J. C. Liao, Nature 2008, 451, 86–89.
[2] S. Atsumi, Z. Li, J. C. Liao, Appl. Environ. Microbiol. 2009, 75, 6306–6311.
[3] S. Atsumi, T.-Y. Wu, E.-M. Eckl, S. D. Hawkins, T. Buelter, J. C. Liao, Appl.
Microbiol. Biotechnol. 2010, 85, 651–657.
[4] E. A. Savrasova, A. D. Kivero, R. S. Shakulov, N. V. Stoynova, J. Ind. Micro-
biol. Biotechnol. 2011, 38, 1287–1294.
[5] A. Baez, K.-M. Cho, J. C. Liao, Appl. Microbiol. Biotechnol. 2011, 90,
1681–1690.
[6] J.-K. Guterl, D. Garbe, J. Carsten, F. Steffler, B. Sommer, S. Reiße, A. Phil-
ipp, M. Haack, B. Rꢀhmann, A. Koltermann, U. Kettling, T. Brꢀck, V.
Sieber, ChemSusChem 2012, 5, 2165–2172.
[7] A. F. Cann, J. C. Liao, Appl. Microbiol. Biotechnol. 2008, 81, 89–98.
[8] C. R. Shen, J. C. Liao, Metab. Eng. 2008, 10, 312–320.
[9] M. R. Connor, J. C. Liao, Curr. Opin. Biotechnol. 2009, 20, 307–315.
[10] GESTIS—Substance Database, Sankt Augustin, (Germany), 2011.
[11] M. C. Renna, N. Najimudin, L. R. Winik, S. A. Zahler, J. Bacteriol. 1993,
175, 3863–3875.
[12] O. Maestri, F. Joset, Mol. Microbiol. 2000, 37, 828–838.
[13] S. Li, J. Wen, X. Jia, Appl Microbiol. Biotechnol. 2011, 91, 577–589.
[14] J. A. McCourt, R. G. Duggleby, Amino Acids 2006, 31, 173–210.
[15] J. V. Schloss, D. E. Van Dyk, Methods Enzymol. 1988, 166, 445–454.
[16] R. G. Duggleby, S. S. Pang, J. Biochem. Mol. Biol. 2000, 33, 1–36.
[17] S. S. Pang, R. G. Duggleby, L. W. Guddat, J. Mol. Biol. 2002, 317, 249–
262.
[47] http://www.pymol.org.
[48] E. Chovancova, A. Pavelka, P. Benes, O. Strnad, J. Brezovsky, B. Kozlikova,
A. Gora, V. Sustr, M. Klvana, P. Medek, L. Biedermannova, J. Sochor, J.
Damborsky, PLoS Comput. Biol. 2012, 8, e1002708.
[49] C. N. Pace, F. Vajdos, L. Fee, G. Grimsley, T. Gray, Protein Sci. 1995, 4,
2411–2423.
[50] “Protein Identification and Analysis Tools on the ExPASy Server”, E. Gas-
teiger, C. Hoogland, A. Gattiker, S. Duvaud, M. R. Wilkins, R. D. Appel, A.
Bairoch in The Proteomics Protocols Handbook (Ed.: J. M. Walker),
Humana, Totowa, 2005, pp. 571–607.
[18] J.-G. Wang, P. K.-M. Lee, Y.-H. Dong, S. S. Pang, R. G. Duggleby, Z.-M. Li,
L. W. Guddat, FEBS J. 2009, 276, 1282–1290.
[51] V. S. Stoll, J. S. Blanchard, Methods Enzymol. 2009, 463, 43–56.
[19] S. S. Pang, R. G. Duggleby, R. L. Schowen, L. W. Guddat, J. Biol. Chem.
2004, 279, 2242–2253.
[20] S. Zhang, M. Liu, Y. Yan, Z. Zhang, F. Jordan, J. Biol. Chem. 2004, 279,
54312–54318.
Received: September 17, 2014
[21] D. Meyer, P. Neumann, C. Parthier, R. Friedemann, N. Nemeria, F. Jordan,
K. Tittmann, Biochemistry 2010, 49, 8197–8212.
Published online on && &&, 0000
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 10
&
9
&
ÞÞ
These are not the final page numbers!

FULL PAPERS
B. Sommer, H. von Moeller, M. Haack,
F. Qoura, C. Langner, G. Bourenkov,
D. Garbe, B. Loll,* T. Brꢀck*
Acetolactate synthase (ALS) is an es-
sential enzyme in designed isobutanol
biosynthesis pathways. Because of its ef-
ficient catalytic conversion of pyruvate
to acetolactate and its reactivity to-
wards the alternative substrate ketoiso-
valerate (KIV), the ALS of B. subtilis is of
particular interest. This study provides
new insights into the structure–function
relationships of this catalytic mecha-
nism.
&& – &&
Detailed Structure–Function
Correlations of Bacillus subtilis
Acetolactate Synthase
ꢂ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemBioChem 0000, 00, 1 – 10
&10
&
ÞÞ
These are not the final page numbers!