Optimization of Enzyme Mechanism along the Evolutionary Trajectory of a Computationally Designed (Retro-)Aldolase

Article abstract of DOI:10.1021/jacs.7b05796

De novo biocatalysts have been successfully generated by computational design and subsequent experimental optimization. Here, we examined the evolutionary history of the computationally designed (retro-)aldolase RA95. The modest activity of the starting enzyme was previously improved 10⁵-fold over many rounds of mutagenesis and screening to afford a proficient biocatalyst for enantioselective cleavage and synthesis of β-hydroxyketones. Using a set of representative RA95 variants, we probed individual steps in the multistep reaction pathway to determine which processes limit steady-state turnover and how mutations that accumulated along the evolutionary trajectory influenced the kinetic mechanism. We found that the overall rate-limiting step for aldol cleavage shifted from C-C bond scission (or an earlier step in the pathway) for the computational design to product release for the evolved enzymes. Specifically, interconversion of Schiff base and enamine intermediates, formed covalently between acetone and the catalytic lysine residue, was found to be the slowest step for the most active variants. A complex hydrogen bond network of four active site residues, which was installed in the late stages of laboratory evolution, apparently enhances lysine reactivity and facilitates efficient proton shuffling. This catalytic tetrad accounts for the tremendous rate acceleration observed for all steps of the mechanism, most notably Schiff base formation and hydrolysis. Comparison of our results with kinetic and structural studies on natural aldolases provides valuable feedback for computational enzyme design and laboratory evolution approaches alike.

Full text of DOI:10.1021/jacs.7b05796

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES
Article
Optimization of Enzyme Mechanism along the Evolutionary
Trajectory of a Computationally Designed (Retro-)Aldolase
Cathleen Zeymer, Reinhard Zschoche, and Donald Hilvert
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b05796 • Publication Date (Web): 07 Aug 2017
Downloaded from http://pubs.acs.org on August 7, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Optimization of Enzyme Mechanism along the Evolutionary
Trajectory of a Computationally Designed (Retro-)Aldolase
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Cathleen Zeymer, Reinhard Zschoche, and Donald Hilvert*
Laboratory of Organic Chemistry, ETH Zꢀrich, 8093 Zꢀrich, Switzerland
ABSTRACT: De novo biocatalysts have been successfully generated by computational design and subsequent experimental
optimization. Here, we examined the evolutionary history of the computationally designed (retro-)aldolase RA95. The modest
activity of the starting enzyme was previously improved 105 fold over many rounds of mutagenesis and screening to afford a
proficient biocatalyst for enantioselective cleavage and synthesis of β-hydroxyketones. Using a set of representative RA95 variants,
we probed individual steps in the multistep reaction pathway to determine which processes limit steady-state turnover and how
mutations that accumulated along the evolutionary trajectory influenced the kinetic mechanism. We found that the overall rate-
limiting step for aldol cleavage shifted from C-C bond scission (or an earlier step in the pathway) for the computational design to
product release for the evolved enzymes. Specifically, interconversion of Schiff base and enamine intermediates, formed covalently
between acetone and the catalytic lysine residue, was found to be the slowest step for the most active variants. A complex hydrogen
bond network of four active site residues, which was installed in the late stages of laboratory evolution, apparently enhances lysine
reactivity and facilitates efficient proton shuffling. This catalytic tetrad accounts for the tremendous rate acceleration observed for
all steps of the mechanism, most notably Schiff base formation and hydrolysis. Comparison of our results with kinetic and
structural studies on natural aldolases provides valuable feedback for computational enzyme design and laboratory evolution
approaches alike.
8F (kcat/KM = 34,000 M-1 s-1).16 This highly active enzyme
INTRODUCTION
possesses an active site with a sophisticated hydrogen bond
network of catalytic residues, reminiscent of the functional
group arrays in natural class I aldolases.16,17 RA95.5-8F not
only exhibits a >109-fold rate acceleration over uncatalyzed
methodol cleavage, it is also an efficient and highly enantiose-
lective catalyst for the synthesis of aldol products.16 As such, it
may be useful for the stereospecific formation of C-C bonds in
diverse high-value compounds.18
Combining computational enzyme design with laboratory
evolution has proven to be a successful strategy for developing
novel biocatalysts with non-natural function,1,2 such as Kemp
eliminases3-6 and Diels-Alderases.7,8 Monitoring the
evolutionary history of such systems has the potential to
provide valuable insights into both enzyme mechanism and the
basic principles of molecular evolution. One class of de novo
protein catalysts, namely the (retro-)aldolases, is particularly
well suited for this kind of retrospective study, as successful
designs are available in multiple protein scaffolds.9-11 Because
the aldol reaction proceeds via a multistep mechanism
involving several covalent intermediates, its analysis can shed
light on how complex reaction manifolds adapt to selective
evolutionary pressures.
Retro-aldolase RA95, one of the best-studied de novo
enzymes, was computationally designed to promote carbon-
carbon bond cleavage of β-hydroxyketones via amine
catalysis. The starting catalyst, RA95.0, which was equipped
with a reactive lysine adjacent to a hydrophobic binding
pocket (Figure 1), cleaves the abiological aldol substrate
methodol with a ca. 104-fold rate acceleration over background
(kcat/KM = 0.17 M-1 s-1).9,10 It was subsequently improved
>9000 fold over many rounds of mutagenesis and medium-
throughput screening in microtiter plates to give variant
RA95.5-8 (kcat/KM = 1,600 M-1 s-1).10,12 Optimization entailed
substantial structural rearrangements at the active site, includ-
ing complete replacement of the originally designed catalytic
apparatus. Further evolution using fluorescence-activated
Figure 1. The computationally designed retro-aldolase RA95
before (RA95.0, green, PDB code: 4A29) and after laboratory
evolution (RA95.5-8F, orange, PDB code: 5AN7) in complex
with a mechanism-based diketone inhibitor. Although the overall
structure of the TIM barrel scaffold did not change, active site and
substrate binding pocket were remodeled substantially, including
a change in the position of the catalytic lysine residue.12,16
droplet sorting (FADS),
a microfluidic-based ultrahigh-
throughput screening technique,13-15 yielded variant RA95.5-
ACS Paragon Plus Environment
1

Journal of the American Chemical Society
Page 2 of 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. The multistep (retro-)aldolase mechanism. The colored circles indicate the individual steps of the mechanism that were probed
by different enzyme kinetics experiments. Red: C-C bond cleavage and formation were monitored by 6-MNA absorbance and fluorescence
under steady state and single turnover conditions. Green: Schiff base formation and hydrolysis were monitored by oxygen isotope
exchange. Blue: Enamine formation was monitored by hydrogen/deuterium exchange.
Exploiting the broad scope of amine catalysis, RA95 vari-
ants have also been applied to a range of other transformations
that proceed via reactive enamine and iminium ion intermedi-
ates, including Michael additions, Knoevenagel condensations
and Henry reactions.19-21
constants with the steady-state turnover numbers (kcat), which
report on the slowest step(s) in the aldol cleavage mechanism,
allow us to assess which microscopic processes limit these
catalysts. Such experiments also provide valuable insights into
how mutations that arose during evolutionary optimization
influenced the kinetic mechanism.
Single-Turnover Measurements. We first investigated
the rate of the C-C bond cleavage step. High enzyme
concentrations, well in excess over substrate and, if possible,
significantly higher than the KD value for methodol, were used
to guarantee only a single catalytic turnover per active site.
The measurements were performed by rapid mixing of enzyme
and methodol in a stopped flow apparatus, monitoring the
increase in fluorescence due to formation of the product
6-methoxy-2-naphthaldehyde (6-MNA) as a function of
increasing catalyst concentration.
The single turnover rate constant (kST) for the original com-
putational design, RA95.0, is 1.1 x 10-4 s-1 (Figure S1). This
value is the same as the kcat for (S)-methodol cleavage by this
enzyme, indicating that C-C bond scission or an earlier step in
the aldol cleavage mechanism must be rate limiting for the
starting catalyst. Brønsted values determined for non-
enzymatic retro-aldol cleavage of methodol catalyzed by small
organic amines have suggested that cleavage of the C-C bond
rather than iminium ion formation likely limits the reaction
rates for these simple systems.22 Although mechanistic
differences are conceivable, a common rate-determining step
for the enzymatic and amine-catalyzed transformations seems
plausible considering the highly solvated RA95.0 active site.12
The kST values determined for RA95.5-8 and RA95.5-8F are
respectively >6 x 103 and 3 x 105 times higher than the kST for
RA95.0, highlighting the extent to which C-C bond scission
was accelerated over the course of experimental evolution.
Because these rate constants also exceed the turnover numbers
of these enzymes by roughly a factor of 5 (Figure 3A and
Table 1), cleavage of the C-C bond (and all previous steps in
the pathway) cannot be rate determining overall. A later pro-
cess associated with product release must therefore limit catal-
ysis by the evolved variants. Screening of mutant libraries for
more efficient catalysts evidently shifted the kinetic bottleneck
for retro-aldol cleavage to a later step in the reaction pathway.
The current study was motivated by the belief that
understanding the factors that promote high catalytic
efficiency in these de novo biocatalysts could help to improve
computational design protocols as well as refine experimental
optimization strategies. We thus focused on dissecting what
happened mechanistically along the RA95 evolutionary
trajectory. What was the rate-limiting step in the aldol
cleavage mechanism initially and did it change as the enzyme
was optimized? Did elementary steps in the pathway respond
uniformly or differentially to mutation? Which modifications
enabled effective catalysis of aldol synthesis? To address these
questions, we probed specific steps of the aldolase reaction
pathway kinetically, as depicted in Figure 2, and compared our
findings directly with mechanistic studies of natural
aldolases. Our results contribute to a deeper understanding of
the aldolase mechanism and reveal how directed evolution
gradually reshaped the complex reaction profile to optimize
catalytic efficiency.
RESULTS
Preparation and Steady-State Characterization of
RA95 Variants. We chose the starting computational design
RA95.0 and the evolved variants RA95.5-8 and RA95.5-8F
for detailed mechanistic investigation. These proteins, which
have been kinetically and structurally characterized,10,12,16 were
produced and purified as previously described, and the
reported steady-state parameters for methodol cleavage were
confirmed with minor batch-to-batch variation (Table 1).
RA95.0, which exhibits modest (S)-selectivity, was assayed
with enantiopure (S)-methodol, whereas enantiopure (R)-
methodol was used for all measurements with the (R)-selective
RA95.5-8 and RA95.5-8F variants. As outlined below, we
utilized pre-steady-state kinetic methods and isotope exchange
experiments to dissect the key steps in the aldol reaction
pathway (Figure 2). Comparison of the collected rate
ACS Paragon Plus Environment
2

Page 3 of 17
Journal of the American Chemical Society
buffer, and the rates of 18O-acetone accumulation were
1
2
3
4
5
6
7
8
measured with a GC-MS instrument and fitted to the
Michaelis-Menten equation (Figure 4A and Table 1). Under
saturating conditions, the exchange rate constants (kex) were
2.7 s-1 for RA95.5-8 and 330 s-1 for RA95.5-8F, which are
substantially (8- and 45-fold) larger than the respective kcat
values for methodol cleavage. Thus, hydrolysis of the Schiff
base during acetone release cannot be rate limiting for either
retro-aldolase.
The 300-fold increase in kex/KM for RA95.5-8F compared to
RA95.5-8 is particularly notable in this context, underscoring
the importance of mutations introduced in the last stages of
optimization for enhancing the efficiency of Schiff base
formation and hydrolysis (via a carbinolamine intermediate).
Conversion of the Enamine and Schiff Base Interme-
diates Is Slow. Having eliminated all plausible alternatives,
the only remaining step in the retro-aldol reaction pathway,
namely protonation of the enamine to give the Schiff base
between acetone and the catalytic lysine, must be rate limiting
overall. Because interconversion of the enamine and Schiff
9
10
11
12
13
14
15
16
17
18
Figure 3. Stopped flow kinetics. (R)-Methodol cleavage under
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
single turnover conditions (A) and 6-MNA binding kinetics
(B) were measured with the same stopped flow setup (C). The
kinetic parameters obtained from these experiments are listed in
Table 1.
6-MNA Binding and Dissociation Are Fast Processes.
Methodol cleavage yields two products, acetone and 6-MNA.
Whereas acetone release from the enzyme occurs over several
steps via distinct covalent intermediates (enamine to Schiff
base to carbinolamine), 6-MNA dissociates in a single step.
We measured the binding kinetics of 6-MNA in a stopped-
flow experiment (Figure 3B and Table 1). The change in
6-MNA fluorescence upon binding to the enzyme was
monitored at different concentrations of 6-MNA. Single
exponential kinetic traces were obtained, and the fitted rate
constants show a linear dependence on 6-MNA concentration,
allowing estimation of the dissociation rate constant (koff) from
the y-axis intercept. Because the binding transitions are very
fast, measurements were performed at low temperature (5 °C).
The rates at 29 °C, the temperature used for all other kinetic
measurements, are expected to be about four times higher.
Although the rate constants were determined in the absence of
acetone, which might influence binding to some extent, and
formation of some covalent 6-MNA-enzyme adducts10,22
cannot be ruled out, the data show that 6-MNA association is
quite rapid (kon,5°C > 105 M-1 s-1), and that dissociation (koff,5°C
=
6 and 23 s-1 for RA95.5-8 and RA95.5-8F, respectively) is
considerably faster than steady-state methodol cleavage. Thus,
6-MNA release can be excluded as the rate-determining step
for the retro-aldol reaction under steady-state conditions, indi-
cating that one of the steps associated with acetone release
must limit overall turnover of the improved enzymes. Com-
paratively rapid 6-MNA release can be rationalized by the
structures of RA95.5-8 and RA95.5-8F, which show that the
acetone binding site is more deeply buried than the apolar
pocket that accommodates 6-MNA.12,16
Figure 4. Isotope exchange kinetics. (A) 16O/18O isotope
exchange kinetics measurement probing the formation and
hydrolysis of the Schiff base intermediate formed from acetone.
For both RA95.5-8 and RA95.5-8F, the Schiff base for-
mation/hydrolysis is significantly faster than steady-state (R)-
methodol cleavage (kcat is indicated by the dotted line). (B) Hy-
drogen/deuterium exchange kinetics measurement probing the
formation of the enamine intermediate formed from acetone. The
exchange rate constant kex for both RA95.5-8 and RA95.5-8F is
30-40 % lower than kcat for methodol cleavage (dotted line). The
kinetic parameters obtained from these experiments are listed in
Table 1.
Directed Evolution Greatly Accelerates Schiff Base
Formation and Hydrolysis. We used an oxygen isotope
exchange experiment to probe formation and hydrolysis of the
covalent Schiff base intermediate formed between acetone and
the catalytic lysine. The evolved RA95 variants were
incubated with various concentrations of 16O-acetone in H218O
ACS Paragon Plus Environment
3

Journal of the American Chemical Society
Page 4 of 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Kinetic isotope effects (KIEs). (A) (R)-Methodol cleavage under steady state conditions. Deuterated (red) and non-deuterated
(blue) (R)-methodol were used as substrates for RA95.5-8 and RA95.5-8F in order to determine secondary KIEs. An inverse secondary
KIE was found for kcat. (B) Methodol synthesis from 6-MNA and fully deuterated (red) or non-deuterated (blue) acetone under steady state
conditions. The concentration of acetone was kept constant at 1.0 M, whereas c(6-MNA) was varied. Here, KIEs are 2.6 and 1.5 for
RA95.5-8 and RA95.5-8F, respectively. Because RA95.5-8 exhibits strong substrate inhibition and cannot be saturated with acetone, the
kcat for methodol synthesis is substantially underestimated in this case. The kinetic parameters obtained from these experiments are summa-
rized in Table 1.
base intermediates occurs via acid/base-catalyzed protonation/
deprotonation, we used an NMR-based hydrogen/deuterium
exchange experiment to probe this step. RA95.5-8 and
RA95.5-8F were incubated with h6-acetone in D2O buffer and
the formation of mono-deuterated h5,d1-acetone was measured
acetone release dictates the rate of methodol cleavage
catalyzed by RA95.5-8 and RA95.5-8F. An ordered water
molecule, bound by polar residues proximal to the enamine
intermediate,16 is a possible source of the respective proton.
Aldol Synthesis. From a synthetic point of view, the
formation of enantiopure β-hydroxyketones is more valuable
than their cleavage. However, methodol synthesis from
6-MNA and acetone was not possible with the original
computationally designed aldolases, including RA95.0.
Because the synthetic reaction is a bimolecular process that is
thermodynamically unfavorable, high acetone concentrations
are required to push the equilibrium toward the aldol product.
Unfortunately, the original designs do not tolerate large
amounts of organic co-solvent. Even more problematic,
6-MNA is a better electrophile than acetone and inhibits the
enzyme by forming a covalent Schiff base adduct with the
catalytic lysine10,22, thus blocking the active site.
1
by H-NMR spectroscopy (Figure 4B and Table 1). In com-
plementary experiments, the enzymes were incubated with
fully deuterated d6-acetone in H2O buffer and 2H-NMR
spectroscopy was used to monitor formation of h1,d5-acetone.
In both cases, the H/D exchange rate constants kex were
30-40% lower than kcat for methodol cleavage for both
RA95.5-8 and RA95.5-8F. This result, together with the fact
that all other steps were found to be faster than kcat and there-
fore not rate-limiting, supports the inference that protonation
of the enamine during acetone release is the step in the
cleavage pathway that limits the turnover efficiency of both
catalysts.
To provide additional evidence for this hypothesis, we
determined the secondary kinetic isotope effect (KIE) for
methodol cleavage under steady-state conditions using
non-deuterated and partially deuterated methodol as substrates
(Figure 5A and Table 1). The requisite d5-(R)-methodol was
synthesized enzymatically from d6-acetone and 6-MNA in
D2O buffer using RA95.5-8F as a catalyst. Steady-state
measurements yielded an inverse secondary deuterium KIE on
kcat of ~0.8 for both RA95.5-8 and RA95.5-8F, implicating a
rate-limiting process in which an sp2-hybridized carbon atom
is converted into an sp3 center. Protonation of the enamine
intermediate is the only such transformation in the cleavage
pathway (Figure 2). If C-C bond scission of methodol to form
the enamine were rate limiting, a normal KIE > 1 would have
been expected, as this step is associated with an sp3 to sp2
hybridization change. Thus, we can conclude that conversion
of the enamine into the Schiff base intermediate during
Both obstacles were gradually overcome by directed evolu-
tion. The intermediate RA95.5-8 variant, for example, catalyz-
es the formation of significant amounts of methodol from 6-
MNA and acetone. Nevertheless, its affinity for acetone is too
low (KD ≈ 1 M) to saturate the enzyme under practical experi-
mental conditions, and substantial substrate inhibition by 6-
MNA is still observed (Figure 5B). These features prevent
RA95.5-8 from being a synthetically useful catalyst for aldol
synthesis. They also complicate kinetic analysis of the synthet-
ic reaction, thwarting efforts to determine the true kcat for
methodol formation at saturating acetone concentrations.
In contrast, structural and mechanistic modifications intro-
duced in the late stages of evolution make RA95.5-8F an
excellent aldolase. The 3-fold increase in apparent affinity for
acetone allows saturation of the enzyme without co-solvent-
induced losses in activity, and steady-state kinetics measured
at 1 M acetone show no sign of substrate inhibition by 6-MNA
ACS Paragon Plus Environment
4

Page 5 of 17
Journal of the American Chemical Society
(Figure 5B). Both of these properties are byproducts of ultra- acetone Schiff base into the corresponding enamine (Figure 2),
1
2
3
4
5
6
7
8
high-throughput screening for mutations that maximized the
amount of product formed in a given time interval, thus
simultaneously optimizing total turnover number and initial
rates of product formation. As a consequence, RA95.5-8F
catalyzes the addition of acetone to a broad range of function-
alized aromatic and aliphatic aldehydes, affording useful
conversions and good-to-excellent enantioselectivities.16
Although the affinity for acetone increased over the course
of evolution, the KD value for this substrate is still high,
precluding true single turnover conditions even with the most
active variant. As a result, an unambiguous assignment of the
overall rate-limiting step for the synthetic reaction is not pos-
sible. Nevertheless, we determined steady-state KIEs for
methodol synthesis using 1 M deuterated and non-deuterated
acetone as substrates (Figure 5B and Table 1). Primary KIEs
on kcat for RA95.5-8 and RA95.5-8F are 2.6 and 1.5, respec-
tively, indicating that cleavage of a C-H/D bond is at least
partially rate limiting under these conditions. The only depro-
tonation step in the synthetic direction is the conversion of the
i.e. the reverse of the step that limits enzyme-catalyzed aldol
cleavage. However, because the H/D exchange rate constants
for enamine formation are still higher than the kcat values
measured for methodol synthesis, a later step, either C-C bond
formation or methodol release, must partially limit aldol syn-
thesis
DISCUSSION
Natural class I aldolases like fructose-1,6-diphosphate (FDP)
aldolase, a key enzyme in glucose metabolism, employ a
multistep pathway involving amine catalysis and enzyme-
bound Schiff base intermediates to convert substrates to
products.17 In addition to a catalytic lysine, their active sites
contain extensive arrays of functional groups that mediate
substrate binding, transition state stabilization, and essential
proton transfers during catalysis. Mechanistic studies on FDP
aldolase have shown that neither Schiff base formation/
hydrolysis nor C-C bond formation/cleavage is rate limiting.
Instead, the slowest steps for both FDP cleavage and synthesis
are associated with product release.23-27
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In an attempt to mimic the properties of such enzymes,
computational methods have been used to install reactive
lysines in a variety of natural protein scaffolds.9-11 Enzymes
that accelerate the retro-aldol cleavage of methodol have been
successfully produced in this way, but their activities are low
(kcat/KM < 1 M−1 s−1) compared to natural aldolases (kcat/KM =
105 to 106 M−1 s−1). Investigations of RA61, one of the most
active original designs, have suggested that most of the
enzyme’s 105-fold rate acceleration can be attributed to the
positioning of the substrate in a hydrophobic binding pocket
proximal to a lysine with a reduced pKa.22 Although most
designs also included a catalytic water molecule oriented by
polar residues to facilitate proton transfers, mutagenesis
studies have shown that this element generally does not
contribute to catalytic efficiency.22
Like other computationally designed retro-aldolases, RA95.0
was equipped with a simple catalytic motif consisting of a
reactive lysine (Lys210) to activate the substrate by Schiff
base formation and a glutamate (Glu53) to position an ordered
water molecule for proton transfer. As for RA61, though,
targeted mutagenesis showed that the latter feature does not
play a role in catalysis.12 Analysis of the subsequent evolu-
tionary trajectory of this catalyst suggests that its comparative-
ly low starting activity is due to the absence of additional
functional groups in the vicinity of Lys210 that could assist
formation and breakdown of the first carbinolamine interme-
diate or deprotonation of the ensuing Schiff base intermediate
to initiate C-C bond scission. Unsurprisingly, turnover is
therefore limited by C-C bond cleavage or one of the steps
leading up to it, rather than product release as in FDP aldolase.
As the activity of RA95.0 was improved by directed
evolution, its active site underwent substantial remodeling.
The associated molecular changes enabled increasingly precise
molecular recognition, as evidenced by a steady increase in
enantiospecificity (the specificity for (R) over (S)-configured
methodol increased from 0.4 for the starting design to 14 and
480 for RA95.5-8 and RA95.5-8F, respectively).16 The crystal
structure of RA95.5-8F in complex with a mechanism-based
inhibitor revealed that the enzyme, like its natural counter-
parts, exploits an extensive range of specific polar and apolar
interactions to orient the substrate for reaction and discrimi-
nate between the competing diasteromeric transition states.16
Table 1. Enzyme kinetic parametersa
RA95.5-8
RA95.5-8F
(R)-Methodol cleavage
- Steady state
kcat = 0.36 ± 0.01 s-1
KM = 346 ± 20 μM
kcat = 7.5 ± 0.2 s-1
KM = 153 ± 11 μM
Non-deuterated
kcat = 0.42 ± 0.02 s-1
KM = 418 ± 40 μM
kcat = 9.7 ± 0.4 s-1
KM = 240 ± 23 μM
Deuterated
- KIE
KIE(kcat) = 0.85
kST >> 0.6 s-1
KIE(kcat) = 0.78
kST = 35 ± 4 s-1
- Single turnover
6-MNA binding (at 5°C)
kon = (1.4 ± 0.1)
x 105 M-1 s-1
koff = 6.0 ± 0.3 s-1
kon = (2.2 ± 1.2)
x 105 M-1 s-1
koff = 23 ± 6 s-1
Schiff base formation (16O/18O isotope exchange)
kex = 2.7 ± 0.6 s-1
16O-acetone in H218O
kex = 330 ± 50 s-1
KM = 320 ± 90 mM
KM = 1200 ± 340 mM
Enamine formation (H/D isotope exchange)
kex = 0.21 ± 0.05 s-1
h6-acetone in D2O
kex = 5.2 ± 0.5 s-1
KM = 270 ± 60 mM
KM = 280 ± 112 mM
kex = 0.23 ± 0.06 s-1
d6-acetone in H2O
kex = 7.6 ± 1.8 s-1
KM = 670 ± 230 mM
KM = 575 ± 240 mM
Methodol synthesisb
- Steady state
kcat = 0.038 ± 0.001 s-1
kcat = 1.3 ± 0.03 s-1
KM = 87 ± 7 μM
Non-deuterated
acetone
KM = Ki = 128 ± 6 μM
Deuterated
acetone
kcat = 0.015 ± 0.001 s-1
KM = Ki = 140 ± 11 μM KM = 68 ± 6 μM
kcat = 0.84 ± 0.02 s-1
- KIE
KIE(kcat) = 2.6 KIE(kcat) = 1.5
a
If not indicated otherwise, all measurements were performed at 29 °C in
25 mM HEPES buffer (pH 7.5) containing 150 mM NaCl.
b Steady-state data measured at constant c(acetone) = 1 M. Km and Ki refer
to 6-MNA, which was varied in concentration.
ACS Paragon Plus Environment
5

Journal of the American Chemical Society
Page 6 of 17
The observed remodeling was not restricted to refining
enzyme-substrate complementarity. Complete replacement of
the original catalytic apparatus with more effective
a
suitably positioned active site phenylalanine with
a
tyrosine.32,33 A double mutant containing an additional
glutamate-to-glutamine substitution (Figure 6B) was even
slightly more active than an authentic F6P aldolase (kcat = 1.3
s-1 and kcat/KM = 108 M-1 s-1).33 Structural and mechanistic
studies of the latter variant showed that tyrosine-mediated
protonation of the enamine intermediate formed upon F6P
cleavage was the rate-limiting step. Thus, the analogy between
F6P aldolase and RA95.5-8F extends to kinetic mechanism.
Optimization of the computational design over many rounds of
laboratory evolution apparently converged on a solution to the
problem of catalyzing aldol cleavage and synthesis already
known in nature.
1
2
3
4
5
6
7
8
a
constellation of functional groups attests to even more
fundamental alterations.12,16 For example, an alternative lysine
residue (Lys83) emerged early in evolution and gradually
supplanted the originally designed Lys210. This mechanistic
substitution was made possible by the stepwise installation of
three supporting polar residues in the vicinity of the new ly-
sine. In the most evolved variant, Tyr51, Asn110 and Tyr180
form a tight hydrogen bond network with Lys83 (Figure 6A),
lowering the pKa of the catalytic lysine to 6.2 and facilitating
efficient proton shuffling during catalysis. Surrounding muta-
tions, identified by extensive ultrahigh-throughput screening
of large gene libraries in the final rounds of optimization,
likely further fine-tuned this catalytic device.
Comparisons of RA95.5-8 and RA95.5-8F show that
installation of the Lys-Tyr-Asn-Tyr tetrad improved overall
performance substantially, although the elementary steps
along the reaction coordinate responded differentially to this
evolutionary innovation (Table 1). The >120-fold acceleration
of water-acetone oxygen exchange indicates that Schiff base
formation and hydrolysis were enhanced to the greatest extent,
likely because the transition states for formation and
breakdown of the intermediate carbinolamine adduct are
particularly well stabilized by this hydrogen bonding network.
Similar effects may also be responsible for the improvements
seen for the early steps of methodol cleavage, up to and
including C-C bond scission (which is likely promoted by
either Tyr51 or Tyr180 functioning as a base), rationalizing
the shift we observed in overall rate-determining step from
aldol cleavage for the computational design to product release
for the evolved catalysts.
Judging from the 25-fold increase in the rate of H/D-
exchange with acetone catalyzed by RA95.5-8F versus
RA95.5-8, the evolved hydrogen bond network also enhanced
interconversion of the enamine and Schiff base intermediates,
which (partially) limits turnover efficiency in both the cleav-
age and synthetic directions. Similar effects have been
observed previously for both natural aldolases17 as well as
catalytic antibodies.28 The structure of RA95.5-8F suggests
that the catalytic tetrad could position an ordered wa-
ter/hydroxide molecule that could serve as the acid/base in this
step.16 Rapid formation of the enamine from acetone, of
course, is essential for efficient production of aldol products
from acetone and aldehydes. In conjunction with the proximal
hydrophobic pocket for binding 6-MNA, which was intro-
duced by design and preserved over the entire evolutionary
trajectory,16 efficient deprotonation of the acetone Schiff base
intermediate, either directly by one of the polar supporting
residues or by a bound hydroxide, accounts for RA95.5-8F’s
utility as a biocatalyst.
The catalytic tetrad in RA95.5-8F resembles the catalytic
machineries found in natural class I aldolases. The structural
resemblance with fructose-6-phosphate (F6P) aldolase is
particularly striking.29-31 In addition to the canonical catalytic
lysine, the F6P aldolase active site contains a tyrosine and a
glutamine—analogous to Tyr51 and Asn110 in the evolved
RA95 variants—that position a catalytic water near the cova-
lently bound substrate. The tyrosine in F6P aldolase is also
believed to play a multifunctional role as an acid/base in
catalysis. In support of this view, homologous transaldolases
have been converted into efficient F6P aldolases by replacing
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Structural comparison of the catalytic machinery in
active aldolases. (A) Structure of RA95.5-8F in complex with a
diketone inhibitor, bound covalently as a Schiff base via the
catalytic lysine residue. Residues Lys83, Tyr180, Tyr51 and
Asn110 form a catalytic tetrad essential for activity. The sophisti-
cated hydrogen bond network between these residues and the
substrate is proposed to facilitate efficient protonation/
deprotonation. PDB code: 5AN7. (B) Structure of a mutated trans-
aldolase from Thermoplasma acidophilum. Mutations in key
catalytic residues (E60Q and F132Y) led to a switch from
transaldolase to fructose-6-phosphate (F6P) aldolase activity.33
The protein was co-crystallized with F6P, which was converted to
glyceraldehyde-3-phosphate (G3P) and dihydroxy-acetone
(DHA), the latter being bound covalently as a Schiff base via the
catalytic lysine. The active site arrangement shown here, compris-
ing a tyrosine and a glutamine residue, is proposed to efficiently
catalyze the rate-limiting step in aldol cleavage, namely protona-
tion of the carbanion-enamine intermediate, in F6P aldolase. 31,33
PDB code: 4XZ9.
ACS Paragon Plus Environment
6

Page 7 of 17
Journal of the American Chemical Society
chromatography. The reaction yielded 11.8 mg d5-(R)-
CONCLUSION
1
methodol (47% isolated yield). Chiral HPLC analysis was
performed as described previously16 using a Reprosil 100
Chiral-NR column (8 μm, 250 x 4,6 mm) and isocratic
hexane/EtOAc = 3:7. The (R):(S) ratio for the isolated product
was 99.3 : 0.7. The deuteration pattern was confirmed by
NMR and mass spectrometry by direct comparison with non-
deuterated (R)-methodol (Figure S2). 1H NMR (300 MHz,
CDCl3) δ 7.65 (m, 3H), 7.35 (dd, 1H), 7.07 (m, 2H), 5.20 (d,
1H), 3.84 (s, 3H), 3.25 (d, 1H). 13C NMR (101 MHz, CDCl3) δ
209.42, 157.75, 137.81, 134.12, 129.45, 128.74, 127.21,
124.28, 119.06, 105.67, 77.20, 69.93, 55.31. HRMS (ESI):
calculated for [C15H11D5O3]+ : m/z = 249.1413, found: m/z =
249.1409.
Steady-State Enzyme Activity Assays. (R)-Methodol
cleavage was measured at 29 °C in buffer A supplemented
with 2.7 % acetonitrile as described previously.12,16 Briefly, the
formation of 6-MNA was monitored over time by absorption
at 350 nm (ε350 = 5970 M-1 cm-1). The enzyme concentration
used was 1 μM for RA95.5-8 and 20 nM for RA95.5-8F. (R)-
Methodol was synthesized as described previously,12 and its
concentration in the assay buffer (25 to 700 μM) was deter-
mined by absorption at 330 nm (ε330 = 1390 M-1 cm-1) prior to
initiating the reaction by addition of enzyme.
In combination with previously available structural data, the
current kinetic study has deepened our understanding of the
evolutionary transformations that enabled conversion of a
mediocre de novo aldolase into a proficient enzyme. The
mechanistic issues identified by our analysis are likely to be
relevant for future efforts to design and evolve protein cata-
lysts with true enzyme-like activities and selectivities.
2
3
4
5
6
7
8
Looking forward, improved computational and experimental
methods to incorporate functional group arrays are clearly
needed. The catalytic motifs in first-generation computational
designs are relatively simplistic and were installed with
insufficient precision to achieve high activity. Observed rate
enhancements typically fall in the range 102 to 104 fold over
background,9,10 far below those observed for natural aldolases.
As a consequence, the starting designs require extensive or, as
in the case of the RA95 aldolase, complete experimental re-
design to achieve practically useful activities. Laboratory
evolution has proved to be a particularly powerful tool for
augmenting the efficiency of these catalysts, often boosting
the activity by an additional three to five orders of magnitude.
Although the prospect of directly designing highly active
enzymes is still some way off, recent progress in modeling
atomically accurate polar networks34 is cause for optimism. By
installing hydrogen-bond donors and acceptors in the vicinity
of key catalytic residues, more effective starting catalysts can
be expected. These, in turn, may be easier to optimize by
directed evolution, taking advantage of smart libraries35-37 and
ultrahigh-throughput screening modalities13-15,38-40 to navigate
sequence space more effectively.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Methodol synthesis from acetone and 6-MNA was
measured at 29 °C in buffer A by monitoring the consumption
of 6-MNA spectroscopically (ε350 = 5970 M-1 cm-1 or ε370
=
1029 M-1 cm-1). The enzyme concentration used was 5 μM for
RA95.5-8 and 100 nM for RA95.5-8F. The concentration of
acetone was kept constant at c = 1.0 M, whereas c(6-MNA)
was varied (10 to 600 μM).
All measurements were performed in triplicate. The steady-
state parameters kcat and KM were determined by fitting the
data to the Michaelis-Menten equation. For the RA95.5-8-
catalyzed synthesis of (R)-methodol, a modified equation was
used to account for substrate inhibition by 6-MNA:
Materials. All chemicals were purchased from Sigma-
Aldrich, Merck or Fluka and used without further purification.
Isotopically labeled compounds (D2O, H218O, NaOD, CDCl3,
d6-acetone, d3-acetonitrile) were purchased from Cambridge
Isotope Laboratories Inc.
Protein Production and Purification. RA95.0, RA95.5-8
and RA95.5-8F were produced in E. coli BL21 Gold (DE3)
and purified by nickel affinity chromatography as previously
described.12,16 The enzymes were stored in buffer A (25 mM
HEPES pH 7.5 and 150 mM NaCl), which was also used for
all kinetic assays. The same buffer was prepared in either D2O
(note: pD = pH + 0.4) or H218O for the respective isotope
exchange experiments described below.
푣0
푘cat ∙ [S]
=
[S]
퐾
푖
[E]
퐾M + [S] ∙ (1 +
)
Kinetic Isotope Effects. KIEs were determined for both
the cleavage and synthesis of (R)-methodol under steady-state
conditions as described above by directly comparing the
turnover of deuterated and non-deuterated substrates using the
same enzyme batch. Deuterated d5-(R)-methodol was used as
the substrate for the cleavage experiment, whereas fully
deuterated d6-acetone and non-deuterated 6-MNA were used
to determine the KIE for methodol synthesis. The KIE was
Synthesis of Deuterated d5-(R)-Methodol. The enzy-
matic synthesis of (R)-methodol from acetone and 6-MNA
using RA95.5-8F as a biocatalyst was reported previously.16
Here, we used this approach with fully deuterated acetone as
the substrate:
calculated by the equation KIE
= kcat(H)/kcat(D). All
measurements were performed in triplicate.
Single Turnover Enzyme Activity Assays. (R)-Methodol
cleavage under single turnover conditions was measured at 29
°C in buffer A supplemented with 2.7 % acetonitrile using an
Applied Photophysics SX18 stopped-flow spectrometer
equipped with a xenon arc lamp. The formation of 6-MNA
was monitored by fluorescence upon excitation at 330 nm and
detection using a 400 nm cut-off filter. (R)-Methodol and
enzyme (RA95.5-8 or RA95.5-8F) were mixed in a 1:5 ratio to
give final concentrations of 150 μM (R)-methodol and 200 to
800 μM enzyme. An excess of enzyme over substrate and
concentrations above KD(substrate) were crucial to achieve
single turnover conditions. The kinetic traces were collected in
The reaction mixture contained 2 mM 6-MNA, 2 M d6-
acetone and 0.1 μM RA95.5-8F in 50 ml buffer A prepared in
D2O. The mixture was gently shaken for 3 hours at 29 °C, and
subsequently saturated with 15 g solid sodium chloride and
extracted with 3 x 50 ml ethyl acetate. The organic phase was
dried over sodium sulfate before removing the solvent under
vacuum. The crude material was purified by flash
ACS Paragon Plus Environment
7

Journal of the American Chemical Society
triplicate, averaged and fitted to single exponential functions.
Page 8 of 17
1
2
The resulting single turnover rate constants kST were plotted
against enzyme concentration.
Supporting Information
3
4
5
6
7
8
9
The Supporting Information is available free of charge on the
ACS Publications website at DOI:
6-MNA Binding Kinetics. 6-MNA binding to RA95.5-8
and RA95.5-8F was measured at 5 °C in buffer A supple-
mented with 2.7 % acetonitrile using an Applied Photophysics
SX18 stopped-flow spectrometer equipped with a xenon arc
lamp. Binding was monitored by a change in 6-MNA
fluorescence with excitation at 330 nm and detection using a
400 nm cut-off filter. Enzyme (RA95.5-8 or RA95.5-8F) and
6-MNA were mixed in a 1:1 ratio to generate final concentra-
tions of 5 μM enzyme and 15 to 70 μM 6-MNA. The kinetic
traces were collected in triplicate, averaged and fitted to
single exponential functions. The observed rate constants were
linearly dependent on c(6-MNA), indicating a simple one-step
binding mechanism. Thus, the association and dissociation
rate constants, kon and koff, could be extracted from slope and
y-axis intercept, respectively. Note that binding and release of
6-MNA are expected to be about 4-fold faster at 29 °C, the
temperature used for all other kinetics experiments.
16O/18O Isotope Exchange Kinetics by GC-MS.
Oxygen isotope exchange in acetone was monitored using a
Thermo Finnigan TRACE GC-MS instrument equipped with
an Rtx-Wax column (30 m, RESTEK). Helium was used as
the carrier gas (1.2 ml/min); the injection temperature was 250
°C; analysis temperature was constant at 100 °C; and the run
time was 3 min. The enzyme was incubated with 16O-acetone
at 29 °C in buffer A containing 96 % H218O. Samples (V =
1 μl) taken 1 min, 5 min and 9 min after starting the reaction
were injected directly into the GC-MS, and the ratio of
16O-acetone (m/z = 58) and 18O-acetone (m/z = 60) was
quantified. The measurements were performed at acetone
concentrations of 10 to 500 mM and an enzyme concentration
of 50 μM for RA95.5-8 and 500 nM for RA95.5-8F, respec-
tively. The rate of background exchange, which was measured
in the absence of enzyme, was subtracted before fitting the
data to the Michaelis-Menten equation to determine kex and
KM(acetone) for Schiff base formation. All measurements
were performed in triplicate.
Additional kinetic, NMR and mass spectrometry data (PDF)
AUTHOR INFORMATION
Corresponding Author
*hilvert@org.chem.ethz.ch
Notes
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The authors declare no competing financial interest.
ACKNOWLEDGMENT
We thank Professor Peter Chen and Armin Limacher for access
to and technical support with the GC-MS. We are grateful to the
NMR Service of the Laboratory of Organic Chemistry at ETH
Zurich and Dr. Marc-Olivier Ebert and René Arnold in particular
for performing the hydrogen/deuterium exchange experiments.
We thank Professor Bernhard Jaun, Dr. Xavier Garrabou and
Adrian Bunzel for helpful discussions and Duncan Macdonald for
help with chiral HPLC. This work was generously supported by
ETH Zurich and the Swiss National Science Foundation. C.Z. is
recipient of a Marie Skłodowska-Curie Individual Fellowship
(TIMEnzyme). R.Z. is grateful for a scholarship from the
Stipendienfonds der Schweizerischen Chemischen Industrie
(SSCI).
REFERENCES
(1) Kiss, G.; Celebi-Olcum, N.; Moretti, R.; Baker, D.; Houk, K.
N. Angew. Chem., Int. Ed. 2013, 52, 5700-5725.
(2) Hilvert, D. Annu. Rev. Biochem. 2013, 82, 447-470.
(3) Rothlisberger, D.; Khersonsky, O.; Wollacott, A. M.; Jiang, L.;
DeChancie, J.; Betker, J.; Gallaher, J. L.; Althoff, E. A.; Zanghellini,
A.; Dym, O.; Albeck, S.; Houk, K. N.; Tawfik, D. S.; Baker, D.
Nature 2008, 453, 190-195.
(4) Khersonsky, O.; Rothlisberger, D.; Dym, O.; Albeck, S.;
Jackson, C. J.; Baker, D.; Tawfik, D. S. J. Mol. Biol. 2010, 396, 1025-
1042.
(5) Khersonsky, O.; Kiss, G.; Rothlisberger, D.; Dym, O.; Albeck,
S.; Houk, K. N.; Baker, D.; Tawfik, D. S. Proc. Natl. Acad. Sci. U. S.
A. 2012, 109, 10358-10363.
(6) Blomberg, R.; Kries, H.; Pinkas, D. M.; Mittl, P. R.; Grutter,
M. G.; Privett, H. K.; Mayo, S. L.; Hilvert, D. Nature 2013, 503, 418-
421.
(7) Siegel, J. B.; Zanghellini, A.; Lovick, H. M.; Kiss, G.; Lambert,
A. R.; St Clair, J. L.; Gallaher, J. L.; Hilvert, D.; Gelb, M. H.;
Stoddard, B. L.; Houk, K. N.; Michael, F. E.; Baker, D. Science 2010,
329, 309-313.
(8) Preiswerk, N.; Beck, T.; Schulz, J. D.; Milovnik, P.; Mayer, C.;
Siegel, J. B.; Baker, D.; Hilvert, D. Proc. Natl. Acad. Sci. U. S. A.
2014, 111, 8013-8018.
(9) Jiang, L.; Althoff, E. A.; Clemente, F. R.; Doyle, L.; Rothlis-
berger, D.; Zanghellini, A.; Gallaher, J. L.; Betker, J. L.; Tanaka, F.;
Barbas, C. F., 3rd; Hilvert, D.; Houk, K. N.; Stoddard, B. L.; Baker,
D. Science 2008, 319, 1387-1391.
(10) Althoff, E. A.; Wang, L.; Jiang, L.; Giger, L.; Lassila, J. K.;
Wang, Z.; Smith, M.; Hari, S.; Kast, P.; Herschlag, D.; Hilvert, D.;
Baker, D. Protein Sci 2012, 21, 717-726.
Hydrogen/Deuterium Exchange Kinetics by NMR
Spectroscopy. Hydrogen/deuterium exchange in acetone was
monitored at 29 °C in buffer A using a 600 MHz Bruker
Avance (UltraShield) NMR spectrometer equipped with a cryo
probe. Two sets of experiments were performed: (i) The
enzyme was incubated with h6-acetone in D2O buffer and the
formation of mono-deuterated h5,d1-acetone was measured in a
1H-NMR experiment; (ii) The enzyme was incubated with
d6-acetone in H2O buffer and the formation of mono-
2
protonated h1,d5-acetone was measured in a H-NMR experi-
ment. D2O buffer was prepared by dissolving 358 mg HEPES
and 526 mg NaCl in a final volume of 60 ml D2O and
adjusting to pD 7.5 using NaOD (note: pD = pH + 0.4). For
the experiment performed in D2O buffer, the enzyme was
buffer-exchanged prior to measurement. The measurements
were performed with acetone concentrations ranging from
25 to 500 mM. The enzyme concentration was 100 μM for
RA95.5-8 and 3 μM for RA95.5-8F. 1H/2H spectra were
collected every 10 min over the course of 2 hours (Figure
S3/S4). Acetonitrile (200 mM) served as an internal standard
for the integration of acetone peaks in all H/D exchange
experiments. The kinetic data were fitted to the Michaelis-
Menten equation to determine kex and KM(acetone) for
enamine formation.
(11) Bjelic, S.; Kipnis, Y.; Wang, L.; Pianowski, Z.; Vorobiev, S.;
Su, M.; Seetharaman, J.; Xiao, R.; Kornhaber, G.; Hunt, J. F.; Tong,
L.; Hilvert, D.; Baker, D. J. Mol. Biol. 2014, 426, 256-271.
(12) Giger, L.; Caner, S.; Obexer, R.; Kast, P.; Baker, D.; Ban, N.;
Hilvert, D. Nat. Chem. Biol. 2013, 9, 494-498.
ACS Paragon Plus Environment
8

Page 9 of 17
Journal of the American Chemical Society
(13) Agresti, J. J.; Antipov, E.; Abate, A. R.; Ahn, K.; Rowat, A.
C.; Baret, J. C.; Marquez, M.; Klibanov, A. M.; Griffiths, A. D.;
Weitz, D. A. Proc. Natl. Acad. Sci. U. S. A. 2010, 107, 4004-4009.
(14) Baret, J. C.; Miller, O. J.; Taly, V.; Ryckelynck, M.; El-
Harrak, A.; Frenz, L.; Rick, C.; Samuels, M. L.; Hutchison, J. B.;
Agresti, J. J.; Link, D. R.; Weitz, D. A.; Griffiths, A. D. Lab Chip
2009, 9, 1850-1858.
1
2
3
4
5
(15) Kintses, B.; Hein, C.; Mohamed, M. F.; Fischlechner, M.;
Courtois, F.; Laine, C.; Hollfelder, F. Chem. Biol. 2012, 19, 1001-
1009.
(16) Obexer, R.; Godina, A.; Garrabou, X.; Mittl, P. R.; Baker, D.;
Griffiths, A. D.; Hilvert, D. Nat. Chem. 2017, 9, 50-56.
(17) Gefflaut, T.; Blonski, C.; Perie, J.; Willson, M. Prog.
Biophys. Mol. Biol. 1995, 63, 301-340.
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(18) Windle, C. L.; Muller, M.; Nelson, A.; Berry, A. Curr. Opin.
Chem. Biol. 2014, 19, 25-33.
(19) Garrabou, X.; Beck, T.; Hilvert, D. Angew. Chem., Int. Ed.
2015, 54, 5609-5612.
(20) Garrabou, X.; Wicky, B. I.; Hilvert, D. J. Am. Chem. Soc.
2016, 138, 6972-6974.
(21) Garrabou, X.; Macdonald, D. S.; Hilvert, D. Chemistry 2017,
23, 6001-6003.
(22) Lassila, J. K.; Baker, D.; Herschlag, D. Proc. Natl. Acad. Sci.
U. S. A. 2010, 107, 4937-4942.
(23) Model, P.; Ponticorvo, L.; Rittenberg, D. Biochemistry 1968,
7, 1339-1347.
(24) Biellmann, J. F.; O'Connell, E. L.; Rose, I. A. J. Am. Chem.
Soc. 1969, 91, 6484-6488.
(25) Pratt, R. F. Biochemistry 1977, 16, 3988-3994.
(26) Rose, I. A.; Warms, J. V. Biochemistry 1985, 24, 3952-3957.
(27) Rose, I. A.; Warms, J. V.; Kuo, D. J. J. Biol. Chem. 1987,
262, 692-701.
(28) Shulman, A.; Sitry, D.; Shulman, H.; Keinan, E. Chemistry
2002, 8, 229-239.
(29) Schurmann, M.; Sprenger, G. A. J. Biol. Chem. 2001, 276,
11055-11061.
(30) Thorell, S.; Schurmann, M.; Sprenger, G. A.; Schneider, G.
J. Mol. Biol. 2002, 319, 161-171.
(31) Tittmann, K. Bioorg. Chem. 2014, 57, 263-280.
(32) Schneider, S.; Sandalova, T.; Schneider, G.; Sprenger, G. A.;
Samland, A. K. J. Biol. Chem. 2008, 283, 30064-30072.
(33) Sautner, V.; Friedrich, M. M.; Lehwess-Litzmann, A.;
Tittmann, K. Biochemistry 2015, 54, 4475-4486.
(34) Boyken, S. E.; Chen, Z.; Groves, B.; Langan, R. A.;
Oberdorfer, G.; Ford, A.; Gilmore, J. M.; Xu, C.; DiMaio, F.; Pereira,
J. H.; Sankaran, B.; Seelig, G.; Zwart, P. H.; Baker, D. Science 2016,
352, 680-687.
(35) Patrick, W. M.; Firth, A. E. Biomol. Eng. 2005, 22, 105-112.
(36) Kazlauskas, R. J.; Bornscheuer, U. T. Nat. Chem. Biol. 2009,
5, 526-529.
(37) Kille, S.; Acevedo-Rocha, C. G.; Parra, L. P.; Zhang, Z. G.;
Opperman, D. J.; Reetz, M. T.; Acevedo, J. P. ACS Synth. Biol. 2013,
2, 83-92.
(38) Fallah-Araghi, A.; Baret, J. C.; Ryckelynck, M.; Griffiths, A.
D. Lab Chip 2012, 12, 882-891.
(39) Fischlechner, M.; Schaerli, Y.; Mohamed, M. F.; Patil, S.;
Abell, C.; Hollfelder, F. Nat. Chem. 2014, 6, 791-796.
(40) Chen, B.; Lim, S.; Kannan, A.; Alford, S. C.; Sunden, F.;
Herschlag, D.; Dimov, I. K.; Baer, T. M.; Cochran, J. R. Nat. Chem.
Biol. 2016, 12, 76-81.
ACS Paragon Plus Environment
9

Journal of the American Chemical Society
Page 10 of 17
1
2
3
Table of Contents artwork
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Page 11 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. The computationally designed retro-aldolase RA95 before (RA95.0, green, PDB code: 4A29) and
after laboratory evolution (RA95.5-8F, orange, PDB code: 5AN7) in complex with a mechanism-based
diketone inhibitor. Although the overall structure of the TIM barrel scaffold did not change, active site and
substrate binding pocket were remodeled substantially, including a change in the position of the catalytic
lysine residue.
55x58mm (300 x 300 DPI)
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. The multistep (retro-)aldolase mechanism. The colored circles indicate the individual steps of the
mechanism that were probed by different enzyme kinetics experiments. Red: C-C bond cleavage and
formation were monitored by 6-MNA absorbance and fluorescence under steady state and single turnover
conditions. Green: Schiff base formation and hydrolysis were monitored by oxygen isotope exchange. Blue:
Enamine formation was monitored by hydrogen/deuterium exchange.
137x63mm (300 x 300 DPI)
ACS Paragon Plus Environment

Page 13 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. Stopped flow kinetics. (R)-Methodol cleavage under single turnover conditions (A) and 6-MNA
binding kinetics (B) were measured with the same stopped flow setup (C). The kinetic parameters obtained
from these experiments are listed in Table 1.
83x71mm (300 x 300 DPI)
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 14 of 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 4. Isotope exchange kinetics. (A) 16O/18O isotope exchange kinetics measurement probing the
formation and hydrolysis of the Schiff base intermediate formed from acetone. For both RA95.5-8 and
RA95.5-8F, the Schiff base formation/hydrolysis is significantly faster than steady-state (R)-methodol
cleavage (kcat is indicated by the dotted line). (B) Hydro-gen/deuterium exchange kinetics measurement
probing the formation of the enamine intermediate formed from acetone. The exchange rate constant kex
for both RA95.5-8 and RA95.5-8F is 30-40 % lower than kcat for methodol cleavage (dotted line). The
kinetic parameters obtained from these experiments are listed in Table 1.
84x113mm (300 x 300 DPI)
ACS Paragon Plus Environment

Page 15 of 17
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. Kinetic isotope effects (KIEs). (A) (R)-Methodol cleavage under steady state conditions. Deuterated
(red) and non-deuterated (blue) (R)-methodol were used as substrates for RA95.5-8 and RA95.5-8F in order
to determine secondary KIEs. An inverse secondary KIE was found for kcat. (B) Methodol synthesis from 6-
MNA and fully deuterated (red) or non-deuterated (blue) acetone under steady state conditions. The
concentration of acetone was kept constant at 1.0 M, whereas c(6-MNA) was varied. Here, KIEs are 2.6 and
1.5 for RA95.5-8 and RA95.5-8F, respectively. Because RA95.5-8 exhibits strong substrate inhibition and
cannot be saturated with acetone, the kcat for methodol synthesis is substantially underestimated in this
case. The kinetic parameters obtained from these experiments are summarized in Table 1.
177x77mm (300 x 300 DPI)
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 16 of 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Structural comparison of the catalytic machinery in active aldolases. (A) Structure of RA95.5-8F in
complex with a diketone inhibitor, bound covalently as a Schiff base via the catalytic lysine residue.
Residues Lys83, Tyr180, Tyr51 and Asn110 form a catalytic tetrad essential for activity. The sophisti-cated
hydrogen bond network between these residues and the sub-strate is proposed to facilitate efficient
protonation/ deprotonation. PDB code: 5AN7. (B) Structure of a mutated trans-aldolase from Thermoplasma
acidophilum. Mutations in key cata-lytic residues (E60Q and F132Y) led to a switch from transaldolase to
fructose-6-phosphate (F6P) aldolase activity. The protein was co-crystallized with F6P, which was converted
to glyceraldehyde-3-phosphate (G3P) and dihydroxy-acetone (DHA), the latter being bound covalently as a
Schiff base via the catalytic lysine. The active site arrangement shown here, comprising a tyro-sine and a
glutamine residue, is proposed to efficiently catalyze the rate-limiting step in aldol cleavage, namely
protonation of the car-banion-enamine intermediate, in F6P aldolase. PDB code: 4XZ9.
ACS Paragon Plus Environment

Page 17 of 17
Journal of the American Chemical Society
1
2
3
4
55x111mm (300 x 300 DPI)
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 18 of 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table of Contents artwork
81x36mm (300 x 300 DPI)
ACS Paragon Plus Environment