Development and scaling-up of the fragrance compound 4-ethylguaiacol synthesis via a two-step chemo-enzymatic reaction sequence

Article abstract of DOI:10.1021/acs.oprd.6b00362

The transformation of (abundant) oxygenated biomass-derived building blocks via chemo-enzymatic methods is a valuable concept for accessing useful compounds, as it combines the high selectivity of enzymes and the versatility of chemical catalysts. In this work, we demonstrate a straightforward combination of a phenolic acid decarboxylase (PAD) and palladium on charcoal (Pd/C) that affords the flavor compound 4-ethylguaiacol from ferulic acid. The use of a two-phase system proved to be advantageous in terms of enzyme activity, stability, and volumetric productivity and allows us to carry out the hydrogenation step directly in the organic layer containing exclusively the intermediate, vinylguaiacol. The enzymatic decarboxylation step in the biphasic system afforded 89% conversion of 100 mM (19 g L^-1) ferulic acid with an isolated yield of 75%. By extracting 4- vinylguaiacol continuously into the organic phase, conversion was enhanced to 92% using 170 mM (33 g L^-1) ferulic acid, which was only possible in the continuous extraction and distillation setup developed. The reaction cascade (PAD-Pd/C) is demonstrated at gram scale, affording the target product 4-ethylguaiacol (1.1 g) in 70% isolated yield in a two-step two-pot process. The enzymatic step was characterized in detail to overcome major constraints, and the process favorably compares in terms of the environmental impact with traditional approaches.

Full text of DOI:10.1021/acs.oprd.6b00362

Subscriber access provided by RUTGERS UNIVERSITY
Full Paper
Development and Scaling-Up of the Fragrance Compound 4-Ethylguaiacol
Synthesis via a Two-Step Chemo-Enzymatic Reaction Sequence
Lorenzo Pesci, Maik Baydar, Silvia Glueck, Kurt Faber, Andreas Liese, and Selin Kara
Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.6b00362 • Publication Date (Web): 02 Dec 2016
Downloaded from http://pubs.acs.org on December 6, 2016
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.
Organic Process Research & Development 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 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Development and ScalingꢀUp of the Fragrance
Compound 4ꢀEthylguaiacol Synthesis via a Twoꢀ
Step ChemoꢀEnzymatic Reaction Sequence
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
Lorenzo Pesci,† Maik Baydar, † Silvia Glueck, ‡,§ Kurt Faber, ‡
Andreas Liese *,† and Selin Kara *,†
† Institute of Technical Biocatalysis, Hamburg University of Technology, Denickestr. 15, 21073
Hamburg (Germany)
‡ACIB GmbH, Petersgasse 14, 8010 Graz (Austria)
§Department of Chemistry, Organic & Bioorganic Chemistry, University of Graz,
Heinrichstrasse 28, 8010 Graz (Austria)
ACS Paragon Plus Environment
1

Organic Process Research & Development
Page 2 of 33
1
2
3
4
ABSTRACT
5
6
7
8
9
The transformation of (abundant) oxygenated biomassꢀderived building blocks via chemoꢀ
enzymatic methods is a valuable concept for accessing useful compounds, as it combines the
high selectivity of enzymes and the versatility of chemical catalysts. In this work, we
demonstrate a straightforward combination of a phenolic acid decarboxylase (PAD) and
palladium on charcoal (Pd/C) that affords the flavor compound 4ꢀethylguaiacol from ferulic acid.
The use of a twoꢀphase system proved advantageous in terms of enzyme activity, stability, and
volumetric productivity, and allows to carry out the hydrogenation step directly in the organic
layer containing exclusively the intermediate, 4ꢀvinylguiacol. The enzymatic decarboxylation
step in the biphasic system afforded 89% conversion of 100 mM (19 g L–1) ferulic acid with an
isolated yield of 75%. By extracting 4ꢀvinylguiacol continuously into the organic phase,
conversion was enhanced to 92% using 170 mM (33 g L–1) ferulic acid, which was only possible
in the continuous extraction and distillation setꢀup developed. The reaction cascade (PAD–Pd/C)
is demonstrated at gram scale, affording the target product 4ꢀethylguaiacol (1.1 g) in 70%
isolated yield in a twoꢀsteps twoꢀpots process. The enzymatic step was characterized in detail to
overcome major constraints and the process favorably compares in terms of the environmental
impact with traditional approaches.
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
KEYWORDS: Biocatalysis, Decarboxylation, Chemical Reduction, ChemoꢀEnzymatic
Cascade, Biphasic System
ACS Paragon Plus Environment
2

Page 3 of 33
Organic Process Research & Development
1
2
3
4
INTRODUCTION
5
6
7
8
9
The combination of enzymatic and – especially heterogeneous – chemical catalysis has been
addressed as a promising concept for the realization of sequential conversions of renewable
feedstocks.1 This is in part due to the synergic effect arising from the high selectivity offered by
homogeneously solubilized enzymes and the robustness of heterogeneous chemical catalysts. In
this context, Pdꢀcatalyzed reductions were employed in cascade reactions partnered by enzymes
in either “oneꢀpot” or “twoꢀpots” fashion.
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
Recent examples include (i) the in situ generation of H2O2 for chloroperoxidaseꢀcatalyzed
oxidation of anisole,2 (ii) the diastereoselective reduction of ꢁ1ꢀ2,6ꢀdisubstituted piperidines
obtained from a transaminase reaction,3 (iii) the alcohol dehydrogenaseꢀPd catalyzed carbonyl
and azide reductions to access βꢀamino alcohols,4 (iv) the routes to (R)ꢀ2ꢀacetoxyꢀ1ꢀindanol via
regioselective chemical hydrogenation coupled with chemoenzymatic dynamic kinetic
resolution5 and (v) to (S)ꢀγꢀhydroxymethylꢀα,βꢀbutenolide via chemical hydrogenation and
lipaseꢀmediated Baeyer–Villiger oxidation starting from celluloseꢀderived levoglucosenone.6
Alkyl and vinyl phenols are aroma compounds present in different types of foods and beverages.
For example, 4ꢀethylguaiacol (4ꢀEG) is present in cooked asparagus, raw Arabica coffee beans,
scotch whiskey and tequila as well as in soy sauce, while 4ꢀvinylguaiacol (4ꢀVG) contributes to
the aroma of beer, tortilla chips, rice bran and raw Arabica coffee beans as well.7 As these
compounds are FDAꢀapproved flavouring agents,8 their synthesis is of a high industrial interest.
Scheme 1 depicts a retrosynthetic analysis involving chemical and enzymatic methods to access
4ꢀVG and 4ꢀEG starting from the guaiacol scaffold 2ꢀmethoxyphenol. Route 1 to 4ꢀEG via
Friedel–Crafts alkylation is probably the least desirable, as recent studies using acid catalysts
ACS Paragon Plus Environment
3

Organic Process Research & Development
Page 4 of 33
1
2
3
4
5
6
7
8
9
reveal competition between Cꢀ and Oꢀalkylation and incomplete regioselectivity in the Cꢀ
alkylation due to the presence of the phenol and methoxy groups.9 Route 2 is selective, but
involves the use of amalgamated Zn in stoichiometric amounts.10 A recently developed
alternative is the Cuꢀdoped porous metal oxideꢀcatalyzed hydrogenolysis in methanol at 0.3
mol% catalyst loading at H2 pressure (40 bar) and 180°C.11 Mild hydrogenation conditions can
be applied using palladium on charcoal (Pd/C) catalysis (route 6). 4ꢀVG can be synthesized
starting from the corresponding aldehyde (route 3) in a Knoevenagel–Doebner condensation
using malonic acid and piperidine as organocatalyst12 in pyridine,13 or by Wittig olefination.14
Heck crossꢀcoupling using aryl bromides (route 4) at low ethylene pressure can only be used if
the phenolic hydroxyl group is protected.15 Alternatives are the Pdꢀcatalyzed crossꢀcoupling
using vinyl Grignard reagents16 or Stille coupling.17
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
4

Page 5 of 33
Organic Process Research & Development
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
Scheme 1. Retrosynthetic analysis for 4ꢀvinylguaiacol (4ꢀVG) and 4ꢀethylguaiacol (4ꢀEG).
Alternatively, 4ꢀVG can be obtained by decarboxylation of ferulic acid (FA) (route 5). This
reaction can be performed chemically by treating FA with base at elevated temperature
(inorganic base in refluxing DMF,18 DBU or NaHCO3/[Hmim]Br under microwave
irradiation.19,20 A milder method of obtaining 4ꢀVG is the biocatalytic variant using phenolic acid
decarboxylases (PADs, EC 4.1.1.–). FA is abundantly available from lignocellulosic biomass21
and constitutes an ideal platform for the synthesis of functionalized aromatics, such as vanillin.22
Although fermentation protocols using either wildꢀtype fungal strains (Candida and
Aspergillus)23,24,25 or engineered E. coli26 yielding 4ꢀVG and 4ꢀEG from FA were established
ACS Paragon Plus Environment
5

Organic Process Research & Development
Page 6 of 33
1
2
3
4
5
6
7
8
9
(route 7), these in vivo strategy yields in general low productivities due to limited substrate (FA)
concentrations. Alternatively, 4ꢀVG can also be reduced enzymatically by NADHꢀdependent
vinyl phenol reductases.27,22
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
Our study focused on routes 5 and 6 by combination of biocatalytic decarboxylation of 4ꢀ
hydroxycinnamic acids by PADs yielding 4ꢀhydroxystyrenes, followed by Pdꢀcatalyzed
hydrogenation of the vinyl bond without intermediate isolation. Regarding the first step, PADs,
first characterized in the 1990s from Bacillus and Lactobacillus28,29 sources, have received
increasing attention during the last years in view of the (industrial) production of bioꢀbased
styrene polymers.30,31,32 Twoꢀphase systems are used to run biotransformations at high substrate
concentrations33, as also shown for this reaction.33a
Scheme 2. Reaction sequence of PADꢀcatalyzed decarboxylation of FA and subsequent
hydrogenation catalyzed by Pd/C in the organic phase. The dashed line indicates that the
hydrogenation reaction in organic phase takes place subsequently in a second reactor.
ACS Paragon Plus Environment
6

Page 7 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
In this work, we carried out a detailed investigation of the biocatalytic decarboxylation of FA
catalyzed by PAD from Mycobacterium colombiense (McPAD, not studied for this purpose so
far) in oneꢀ and twoꢀliquid phase (2LPS) systems, including a mathematical modelling of the
reaction. Under optimized conditions in a 2LPS, the organic layer containing the intermediate 4ꢀ
VG could be easily transferred to the hydrogenation reactor yielding 4ꢀEG (Scheme 2).
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
RESULTS AND DISCUSSION
Enzyme Characterization, Kinetics and Simulation
Our initial experiments on the characterization of the Mycobacterium colombiense PAD
(McPAD) regarding pH and temperature showed that maximum values in terms of activity and
conversion are at pH 6–7 (potassium phosphate (KPi) buffer, ESI Figure S1), which is similar to
that of PADs from bacteria and from yeast. 33ꢀ36 Regarding the temperature, the activity increased
between 20 and 50°C (Figure S2), however, quantitative conversions were achieved only
between 20 and 37°C due limited enzyme stability (t1/2 ~1 h at 50°C). For further studies, we
chose KPi buffer at pH 7.0 and 37°C (40% residual activity after 24 h, ESI Figure S2). With
respect to the kinetics, a typical saturation curve was fitted using the singleꢀsubstrate Michaelis–
Menten equation (ESI Figure S3 and S4). Kinetic parameters with standard deviation for
McPADꢀcatalyzed decarboxylation of ferulic acid (FA) were found as: ꢀꢁ,ꢂꢃ 2.4 ± 0.1 mM,
ꢄ
ꢅꢆꢇ
90 ± 1 U mg–1, and ꢀ
0.14 ± 0.04 mM.
ꢈ,ꢉꢊꢋꢌ
In comparison to other PADs, the ꢀꢁ values are all in the same range (usually between 0.8 and
2 mM, in general <10 mM). For example, ꢀꢁ values for PADs from Candida guillermondi,
Bacillus subtilis, Enterobacter sp., Lactobacillus brevis and Saccharomyces cerevisiae are
5.3 mM,37 1.1 mM,38 2.4 mM,39 0.96 mM,35 and 0.79 mM36 for their phenolic substrates,
ACS Paragon Plus Environment
7

Organic Process Research & Development
Page 8 of 33
1
2
3
4
respectively. On the other hand, ꢄ
values are more diverse: the PAD from yeast has a value of
ꢅꢆꢇ
5
6
7
8
6.8 × 10–3 U mg–1), the one from C. guillermondi 378 U mg–1 and the enzyme from L. brevis 10
U mg–1.
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
Next, we determined product inhibition for McPAD by measuring initial rates in the presence of
increasing product (4ꢀVG) concentrations (ESI Figure S5). Assuming a mechanism involving
competitive product inhibition (i.e. the product binds reversibly to the enzyme forming an
enzymeꢀproduct complex which impedes substrate binding), we fitted the progresses of batch
reactions to Equation 1 (Eqn. 1) by regression of the inhibition constant ꢀꢈ,ꢉꢊꢋꢌ. The equation
represents the proposed catalytic mechanism, where [PAD] is the enzyme concentration, [FA]
and [4ꢀVG] are the concentrations of substrate and product.
ꢋ
[ꢂꢃ]
ꢎꢏꢐ
[
ꢛ
ꢍ =
(Eqn. 1)
]
ꢗꢘꢙꢚ
[ ]
)ꢖ ꢂꢃ
ꢑ
ꢒ,ꢓꢔ
(ꢕꢖ
ꢜ,ꢗꢘꢙꢚ
The inhibition constant (ꢀꢈ,ꢉꢊꢋꢌ) was determined to be 0.14 mM which is in the same order of
magnitude as the one reported by Jung et al.30 Since a biotransformation characterized by a
ꢀ /ꢀꢁ ratio <1 would not proceed efficiently to complete conversion,40 product inhibition by 4ꢀ
ꢈ
VG constitutes a major problem for our approach (ꢀꢈ,ꢉꢊꢋꢌ/ꢀꢁ,ꢂꢃ of 0.06). To validate the
kinetic model, we compared the experimental data with the numerical solutions of the mass
balance equations Eqn. 2.1 and 2.2 (Figure 1 and ESI Figure S6). ꢝꢞ is the deactivation constant
for the enzyme (ꢝꢞ = 6.3×10–4 ± 1×10–4 min–1, see details below).
[
]
ꢞ ꢂꢃ
ꢊꢤ ꢟ
ꢥ
[
]
−
= ꢠꢡꢢ ꢣ
ꢍ
(Eqn. 2.1)
(Eqn. 2.2)
ꢞꢟ
[
]
ꢞ ꢋꢌ
ꢊꢤ ꢟ
ꢥ
[
]
= ꢠꢡꢢ ꢣ
ꢍ
ꢞꢟ
ACS Paragon Plus Environment
8

Page 9 of 33
Organic Process Research & Development
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. Comparison of experimental data (black diamonds: FA; grey squares: 4ꢀVG) with the
simulation of the kinetic model (continuous lines). Dotted lines represent the kinetic model
without competitive product inhibition. Reaction conditions: 5 mM FA, 3.8 U mL–1 PAD (59 µg
mL–1) McPAD in KPi buffer (0.1 M, pH 7.0) at 37°C at 700 rpm.
Figure 1 shows good agreement of the experimental with simulated data (black and grey
continuous lines) and the significant effect of the product inhibition (black and grey dotted lines).
Next, we analyzed the effect of substrate concentration on the maximum conversion (Figure 2)
and the deviation from the theoretical data derived by the kinetic model.
ACS Paragon Plus Environment
9

Organic Process Research & Development
Page 10 of 33
1
2
3
4
5
6
7
8
9
Figure 2. Influence of the substrate concentration on maximum conversion. Black diamonds:
maximum conversion; white diamonds: theoretical maximum conversion according to Eqn.1 at
24 h. Reaction conditions: 1–30 mM FA, 0.5 U mL–1 (8 µg mL–1) McPAD, in KPi buffer (0.1 M,
pH 7.0) at 37°C at 700 rpm. Standard deviations in duplicates 0.2–12%.
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
Such a trend does not derive from excess substrate inhibition (i.e. additional substrate binds to
the enzymeꢀsubstrate complex), but is most likely due to enzyme deactivation at high product
concentrations. In that respect, the maximal conversion should be dependent on the catalyst
amount as observed in our results (ESI Figure S7). Data supporting this observation were also
reported by Leisch and coꢀworkers.32
To further study the influence of enzyme/substrate ratio on the maximal conversion, we
evaluated the enzyme deactivation by comparing reaction progress curves at different biocatalyst
concentrations. In the absence of deactivation, conversion versus time × PAD concentration data
points should fall on the same line.41,42 Our experiments show that this holds true at 5 mM FA
but not at 10 mM, where different enzyme concentrations generate different progress curves at
the late stage of the reaction (ESI Figure S8). Hence, the Eqn. 1 for this reaction system seems to
be valid at high ferulic acid concentrations only when the enzyme concentrations are high
enough [PAD (µg mL–1)/FA (mM) ratio ≈ 6]. Inhibition or deactivation of PADs by the reaction
products have not been described in detail so far. Our data show an influence of 4ꢀVG
concentration on the enzyme, which could be exerted either by (irreversible) deactivation or by
the establishment of a different inhibition type. From a practical point of view, to achieve high
conversions in this system, as a rule of thumb the concentration of the product should be
maintained below 10 mM when PAD concentration is <1.3 U mL–1.
ACS Paragon Plus Environment
10

Page 11 of 33
Organic Process Research & Development
1
2
3
4
TwoꢀLiquid Phase System (2LPS)
5
6
7
8
9
In order to increase the starting FA concentration, we investigated the possibility to use a 2LPS,
where the organic phase would continuously remove the product from the aqueous phase,
thereby keeping it under its inhibition threshold. Since the stability of PADs is an issue in the
presence of organic solvents,43,33,30 a tradeꢀoff between partition coefficients33b of the product
and enzyme activity/stability needs to be made. Figure 3 shows the partition coefficients
ꢀꢦ (expressed as logarithm) and the reaction rate for extraction of 5 mM 4ꢀVG while using
solvents with ꢧꢨꢩꢠ (1ꢀoctanol/water partition coefficient) values of 0.7–4.
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. Extraction rates and partition coefficients of 4ꢀVG in buffer:organic solvent systems.
The ꢧꢨꢩꢠ of each solvent is indicated in brackets. 5 mM 4ꢀVG, 1:1 solvent ratio (v/v).
Equilibrium was reached in all cases after approx. 1 h. DCM: dichloromethane, EtOAc: ethyl
acetate. Standard deviations in duplicates 0.1–12% (ꢧꢨꢩꢀꢦ) and 3–12% (extraction rate).
ACS Paragon Plus Environment
11

Organic Process Research & Development
Page 12 of 33
1
2
3
4
5
6
7
8
9
The highest partition was observed for toluene and EtOAc, while the differences in extraction
rates are not as significant. Biotransformations were first tested in small scale using 5 mM FA to
verify whether the McPAD was compatible (Table 1).
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 1. McPADꢀcatalyzed conversion of FA in twoꢀliquid phase system using different organic
solvents.
Solvent
Hexane
Toluene
DCM
Solubility in watera
Conversion (%)
ꢧꢨꢩꢠ
<0.1
<0.1
1.3
4
99
99
99
5
2.7
1.2
0.7
EtOAc
8.7
a
g in 100 mL, 20°C. Reaction conditions: 5 mM FA, 0.78 U mL–1 (12 µg mL–1) McPAD, in a
2LPS [aqueous: organic, 1:1 (v/v)], aqueous 0.1 M KPi buffer at pH 7 and 37°C, 700 rpm. In the
case of DCM, the organic phase is the lower phase, whereas in all other 2LPSs the organic phase
is the upper phase.
The reaction runs to completion in hexane, toluene and DCM but not in EtOAc, due to its higher
polarity and water solubility, leading to enzyme deactivation. We then selected hexane, toluene
and DCM for further studies. Experiments performed with 20 mM FA and 8 µg mL–1 enzyme
concentration using either hexane or toluene as the second phase showed already a consistent
ACS Paragon Plus Environment
12

Page 13 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
improvement with respect to the reaction performed in buffer, as quantitative conversion was
obtained after <10 h. The 4ꢀVG concentration in buffer remained <4 mM (ESI Figure S9). The
reaction rate in the 2LPS was 69% of the rate observed in the oneꢀphase system. The organic
phases contained only the target product allowing straightforward workꢀup to afford 4ꢀVG.
However, in the case of toluene, the product was isolated as a yellow oil (ESI Figure S10),
indicating low purity. This fact has also been observed by Leisch et al.32 Using hexane, 4ꢀVG
could be isolated in <70% yield as transparent oil. The maximum conversion in the presence of
DCM was ~75%, probably due to enzyme deactivation by longer exposure to the organic phase.
In addition, due to its higher density than water (DCM appears as lower phase) and low boiling
point (40 °C), DCM is not an appropriate solvent in a 2LPS. Hence, hexane was chosen as a
model solvent for further optimization studies. Enzyme stability in the presence of organic
solvents is a critical factor for solvent choice and we therefore studied the stability of McPAD in
a buffer/hexane system. Surprisingly, an increased stability of the enzyme (>3ꢀfold) in the
presence of hexane was detected (Figure 4).
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
13

Organic Process Research & Development
Page 14 of 33
1
2
3
4
Figure 4. Stability of McPAD in buffer and buffer:hexane. Continuous black and grey lines
5
6
represent an exponential fit to the experimental data. Reaction conditions: 5 mM FA, 0.39 U mL–
7
8
9
1
(6 µg mL–1) McPAD; in a 2LPS (aqueous:organic, 1:2 (v/v)), aqueous medium 0.1 M KPi
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
buffer at pH 7 and 37°C, 700 rpm. t1/2 of McPAD in buffer = 15 h; t1/2 of McPAD in
buffer:hexane = 53 h. Standard deviations in duplicates 0.6–20% in buffer; 3–9% in
buffer:hexane.
Decreased activity and increased stability in a 2LPS suggest the establishment of a rigid McPAD
conformation. More flexible conformations facilitate substrateꢀenzyme interactions but at the
same time decrease the thermostability.44,45 Such flexibility changes may occur either at the
interphase or by direct action of solubilized solvent molecules. In fact, even though the solubility
of hexane in water is very low, this would result in a hexane/PAD concentration ratio of approx.
20–30 (depending on the enzyme loading).
Although halfꢀlife times in 2LPSs have not been reported so far, the stability of McPAD (t1/2 of
53 h) seems to be superior to other PADs in a 2LPS.30 It is worth mentioning that the PAD from
(solvent tolerant) B. licheniformis CGMCC 7172 showed a remarkable stability towards high
ꢧꢨꢩꢠ organic solvents after 12 h of incubation.33
The construction of a kinetic model would allow a better understanding of the reaction and the in
silico planning of further improvements of the reaction for technicalꢀscale applications.
Assuming no changes in the catalytic mechanism, we introduced in the mass balance the
reversible extraction of the product in the organic phase (Eqn. 3.1 and 3.2).
[
]
[
]
+ ꢝ ꢊꢕ 4 − ꢄꢪ
ꢬ
ꢍꢕ = ꢍ − ꢝ ꢕ 4 − ꢄꢪ
(Eqn. 3.1)
ꢆꢫ
ACS Paragon Plus Environment
14

Page 15 of 33
Organic Process Research & Development
1
2
3
4
[
]
[
]
− ꢝꢊꢕ 4 − ꢄꢪ
ꢬ
ꢍꢊꢕ = ꢝ ꢕ 4 − ꢄꢪ
(Eqn. 3.2)
ꢆꢫ
5
6
7
8
9
Where [4ꢀVG]aq and [4ꢀVG]o are the product concentrations in aqueous and organic phase, ꢍꢕ is
the velocity for 4ꢀVG formation in the aqueous phase, ꢍ is the velocity in Eqn.1, ꢝꢕ is the rate
constant for the extraction into the organic phase and ꢝꢊꢕ is the rate constant for the reverse
direction; ꢍꢊꢕ is the velocity of 4ꢀVG extraction in the organic phase. The two constants ꢝꢕ and
ꢝꢊꢕ were estimated measuring extraction velocities in 5 mL thermostated vessels using 1 mL
buffer and 2 mL hexane at two different 4ꢀVG concentrations. ꢝꢕ = 0.097 min–1 and ꢝꢊꢕ=
0.0078 min–1 (calculated dividing ꢝꢕ by the partition coefficient measured using an
aqueous:organic ratio of 1:2 (v/v) (ꢧꢨꢩꢀꢦ = 1.082). Since the catalytic constants of the enzyme
would change in the presence of the second phase, we realized a fit of the proposed mass balance
(using the deactivation constant for the enzyme ꢝꢞ =1.6 × 10ꢀ4 ± 7.9 × 10ꢀ5 min–1) to the reaction
progress. The results show that this is a realistic model for the reaction (Figure 5 and ESI Figure
S11). Table S1 shows the determined parameters.
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
15

Organic Process Research & Development
Page 16 of 33
1
2
3
4
5
6
7
8
9
Figure 5. Comparison of experimental data with the kinetic model (continuous lines). Black
diamonds: FA; grey squares: 4ꢀVG in buffer, light grey triangles: 4ꢀVG in hexane. Reaction
conditions: 20 mM FA, 0.58 U mL–1 (9 µg mL–1) McPAD, in a 2LPS [aqueous:organic, 1:2
(v/v)], aqueous medium 0.1 M KPi buffer at pH 7 and 37°C, 700 rpm.
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 shows a typical reaction progress in the 2LPS, where the product accumulates in the
aqueous medium reaching a maximum and the extraction into the organic phase continues up to
an equilibrium point. Initial attempts to run the reaction at 50–100 mM failed because of the pH
increase to 8.0 during the course of the reaction. In order to solubilize FA in KPi buffer (0.1 M,
pH 7.0) KOH was added (final concentration 70 mM), and as the acid substrate is converted, the
pH would become increasingly alkaline and hence buffer capacity (especially at high FA
concentrations) plays a major role in the productivity of the enzyme. Table 2 shows that the
increase in the buffer concentration significantly improved the conversions.
Table 2. Influence of the ferulic acid concentration and buffer strength on the maximum
conversion.
Ferulic
Acid
(FA) KPi Buffer
Max. Conversion
Concentration (mM)
Concentration (M)
(%)
41
100
50
0.1
0.1
66
ACS Paragon Plus Environment
16

Page 17 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
50
0.5
0.5
0.5
0.1 a
0.5
99
95
30
60
87
100
200
200
120
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
a Reaction operated with pH control and autotitrator. Reaction conditions: 50–200 mM FA, 0.84
U mL–1 (13 µg mL–1) McPAD, in a 2LPS (aqueous:organic, 1:2 (v/v)), aqueous medium KPi
buffer at pH 7 and 37°C, 700 rpm.
The results show that with sufficient buffer capacity, the reaction can proceed to high
conversions (87%) up to 120 mM. Beyond 200 mM, even with continuous pH titration,
inhibitory product concentrations in the aqueous phase are dominant. Table 3 shows the process
parameters for the optimized reaction at 100 mM FA.
Table 3. Process parameters for the optimized reaction conditions at 100 mM (19 g Lꢀ1) ferulic
acid (FA).
FA
PAD
X
Y
TON
EC
ACS Paragon Plus Environment
17

Organic Process Research & Development
Page 18 of 33
1
2
3
4
5
6
7
8
(g Lꢀ1)
19
(mg mLꢀ1)/(U mLꢀ1)
0.09/5.8
(%)
95
(%)
79
(mol molꢀ1)
11734
(kg gꢀ1)
89
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
X = Conversion; Y = Isolated yield; TON = Turnover number (maximum conversion at 23 h);
EC = Enzyme consumption (kg product/g of enzyme). Optimized reaction conditions: 100 mM
(19 g L–1) FA, 5.8 U mL–1 (0.09 mg mL–1) McPAD, in a 2LPS (aqueous:hexane, 1:2 (v/v),
10 mL:20 mL), aqueous medium 500 mM KPi buffer at pH 7 and 37°C, 700 rpm.
At this concentration, the reaction progress can be simulated in good agreement only during the
first 2 h of the reaction (Figure S11ꢀC), afterwards it proceeded slower than expected. A possible
reason can be the presence of product concentrations >10 mM in the aqueous phase, causing
enzyme deactivation or product inhibition as mentioned above. The reaction was demonstrated
also at gram scale (1.9 g FA) in 100 mL reaction volume, which afforded 1 g of 4ꢀVG with 89%
conversion and 75% isolated yield.
A limitation of this system is caused by the partition coefficient of 4ꢀVG, which generates a
steadyꢀstate product concentration in the aqueous phase of about 2 mM and which limits the use
of higher FA concentrations. These problems could be circumvented by using higher organic
solvent volumes. We addressed these issues by designing a recycling system that allowed to
reduce the amount of hexane, while at the same time facilitated product isolation. The system is
presented in Scheme 3 and a photograph of the setꢀup is shown in the ESI, Figure S12.
ACS Paragon Plus Environment
18

Page 19 of 33
Organic Process Research & Development
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
Scheme 3. Flow scheme of the extraction/distillation reactor setꢀup. 1: twoꢀphase reactor; 2:
distillation condenser (4°C); 3: vacuum pump connected with the distillation system (400–450
mbar); 4: bubble condenser (4°C); 5: thermostat (50–55°C); 6: vessel collecting product
solution; 7: tube pump (0.25 mL min–1); 8: thermostat (37°C).
In this reactor setꢀup, the organic phase containing the product is continuously pumped (7) into a
second reactor (6) where hexane is distilled under vacuum to reꢀenter the twoꢀphase reactor (1),
while 4ꢀVG (Bp. 220°C) accumulates in the distillation vessel (6). Figure 6 compares the
reaction in the standard setꢀup and in the reactor setꢀup depicted in Scheme 3. Using the
integrated recycling system, quantitative conversions were achieved in shorter time with
complete product extraction from the aqueous phase. This setꢀup was also successfully used at
170 mM substrate concentration (with 92% conversion). Moreover, since at such high FA
concentrations a continuous autotitration system (3 M HCl) is necessary, the reaction was
performed in unbuffered distilled water, thereby reducing the input material (Table 4).
ACS Paragon Plus Environment
19

Organic Process Research & Development
Page 20 of 33
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. Decarboxylation of FA in 2LPS (FA: white triangles; 4ꢀVG: white squares) and in the
extractionꢀdistillation apparatus (FA: black triangles; 4ꢀVG: black squares). Reaction conditions:
100 mM FA, 3.9 U mL–1 (60 µg mL–1) McPAD, in a 2LPS [aqueous:hexane, 1:2 (v/v)], aqueous
medium 0.1 M KPi buffer at pH 7 and 37°C, 700 rpm.
Table 4. Comparison of reactor setꢀups involving distillation.
FA Conc. (mM)
FA Conc. (g L–1)
Conditions
Conversion (%)
100
200a
170
19.4
38.8
33
Extract./distill.
0.1 M KPi Buffer, pH 7
Extract./distill.b
>99
60
92
a Data from Table 2 inserted as comparison.
ACS Paragon Plus Environment
20

Page 21 of 33
Organic Process Research & Development
1
2
3
4
b Reactions operated in distilled water with pH control and autotitrator.
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
Chemical Reduction of 4ꢀVG to 4ꢀEG
The hydrogenation of 4ꢀVG was attempted at 2 mL scale using the homogeneous Wilkinson
catalyst [RhCl(PPh3)3] and heterogeneous Pd on C. The use of 10 mol% Wilkinson catalyst
afforded low conversion (<10%) after 1 h at room temperature and the reaction did not proceed
further. Temperature increase up to 40°C and solvent change (toluene, methanol,
dichloromethane) did not improve the reaction performance. These results indicate catalyst
inhibition/deactivation by the substrate, possibly via coordination of the phenolic group. Pd on C,
on the other hand, was very efficient (Scheme 4).
Scheme 4. Hydrogenation of 4ꢀVG to 4ꢀEG using Pd/C as catalyst.
The reaction afforded the expected product 4ꢀEG with complete selectivity in 83% isolated yield
after a simple filtration of the catalyst and solvent evaporation.
PAD–Pd/C Reaction Sequence to 4ꢀEG
The complete twoꢀstep PADꢀPd/C sequence was performed on a gram scale (1.94 g FA, 100 mL
volume), where after the biocatalytic reaction the organic layer was withdrawn with a syringe
ACS Paragon Plus Environment
21

Organic Process Research & Development
Page 22 of 33
1
2
3
4
5
6
7
8
9
and transferred into the reactor containing the metal catalyst. An isolated yield of 70% for the
final product 4ꢀEG (1.1 g) was achieved. The identity of both products was verified by TLC and
HPLC by comparison with commercial reference materials, and by GCꢀMS (ESI Figure S13).
The performances of both steps in terms of reaction rates and conversion show that the
bottleneck of the whole system is the enzymatic step, which needed further optimization in view
of substrate concentration, biocatalyst loading and reaction conditions.
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
Use of Alternative Organic Solvents
Hexane was chosen as solvent because of the satisfactory performance in view of conversion,
enzyme stability, product purity and low boiling point allowing faster product isolation.
However, it has recently been classified as a nonꢀdesirable solvent.46 For this reason we
investigated alternative solvents (ESI Table S2), such as MTBE (methyl tertꢀbutyl ether), 2ꢀ
methylꢀTHF (2ꢀMeTHF) and heptane. The reaction proceeded smoothly to 96% conversion at 5
mM FA using MTBE as the second phase but only to 45% at 100 mM. 2ꢀMeTHF caused
instantaneous enzyme deactivation, comparable to EtOAc (ESI Table S2). Heptane seems to be
the best alternative, as the reaction performed equally well as in hexane. In general, solvents with
ꢧꢨꢩꢠ below ~1 and water solubility >2 g in 100 mL were found to be detrimental to the enzyme.
This behavior can be related to the stripping of essential water molecules by hydrophilic
solvents.47
Analysis of the Environmental Impact
The Eꢀfactor (kgwaste/kgproduct) is a simple indicative estimation of the environmental impact of a
synthetic procedure,48 which allows to quantitatively compare methodologies at the early stages
of development and to pinpoint improvements. According to Scheme 1, environmental analysis
ACS Paragon Plus Environment
22

Page 23 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
for 4ꢀVG synthesis was realized for routes 3, 4 and 5 (baseꢀcatalyzed decarboxylation). Eꢀ
Factors were calculated separately for the reaction (ESI Figure S14) and for product isolation
(ESI Figure S15ꢀA), solvents were addressed separately and expressed as solvent demand (mL
solvent/g of product, ESI Figure S15ꢀB).
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 chemical (AcOHꢀcatalyzed at ∆T) and biocatalytic routes from FA appear comparable, the
advantage of the enzymatic route becomes clear when considering the Eꢀfactor for workꢀup and
isolation (ESI Figure S15ꢀA), which causes a 150ꢀfold increase in the heat/AcOH
decarboxylation. However, the reaction’s solvent demand for the enzymatic route is highest (ESI
Figure S15ꢀB). Subꢀmolar reactant concentrations in enzymatic reactions are a preponderant
cause of high Eꢀfactors.49 However, considering both reaction and workup together, the
biocatalytic route has the second lowest value among the seven routes (ESI Figure S15ꢀB).
The synthesis of 4ꢀEG in the twoꢀstep sequence was compared with route 2 and 3 (via
stoichiometric Clemmensen reduction versus the catalytic variant and the Wittig/hydrogenation
sequence). The alkylation route is difficult to compare because of unavailability of data,
especially its selectivity, and because these reactions are normally conducted in the gas phase in
continuous flow. Figure S16 shows the results in terms of Eꢀfactor for the reaction. The
superiority of the catalytic routes is evident.
Due to the relatively low concentration of the starting material in the enzymatic reaction, the Eꢀ
factor for workꢀup/isolation and the solvent demand are quite high. Such values are in fact direct
consequences of the large volumes used (ESI Figure S17). Solvent recycling, as shown in Figure
4 is clearly an appealing solution to decrease the overall solvent load substantially. In addition to
that, high substrate concentrations would lower the Eꢀfactor as well.
ACS Paragon Plus Environment
23

Organic Process Research & Development
Page 24 of 33
1
2
3
4
CONCLUSIONS
5
6
7
8
9
A twoꢀstep synthetic sequence to the fragrance compound 4ꢀethylguaiacol (4ꢀEG) starting from
ferulic acid was investigated and established. The route involved the decarboxylation of ferulic
acid catalyzed by a phenolic acid decarboxylase from Mycobacterium colombiense, yielding 4ꢀ
vinylguaiacol (4ꢀVG) followed by hydrogenation to 4ꢀEG by Pd/C, without intermediate
isolation. Careful optimization was required regarding the enzymatic step to overcome
constraints in terms of product inhibition and deactivation.
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
Without continuous product removal, the PADꢀcatalyzed decarboxylation of FA was
demonstrated at gram scale (100 mM, 19 g L–1) resulting in 89% conversion and 75% isolated
yield. However, a reactor strategy integrating product isolation afforded 92% conversion of 170
mM (33 g L–1) FA to 4ꢀVG. Drawbacks of this reactor setꢀup are, however, the higher
complexity and energy input. The complete sequence (PAD–Pd/C) was demonstrated at gram
scale, resulting in 4ꢀEG (1.1 g) with 70% overall isolated yield in a two stepsꢀtwo pots process.
The environmental feasibility of these synthetic steps was estimated, which favourably compared
with other approaches. Enzyme screening/optimization in terms of activity and stability in
connection with reaction engineering makes the whole system appealing for technical
applications. Expanding of the scope of the chemoꢀenzymatic approach by using alternative
carboxylic acids (e.g. coumaric acid, sinapic acid, caffeic acid) is ongoing in our laboratories.
EXPERIMENTAL SECTION
Materials. Ferulic acid was purchased from Carl Roth (Germany), all the other chemicals were
purchased from Sigma Aldrich (Buchs, Switzerland). For preparative experiments, hexane was
degassed using nitrogen prior to use. Data were acquired using HPLC with UV detection using
ACS Paragon Plus Environment
24

Page 25 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
calibration curves obtained with authentic reference materials. Mathematical regressions to
algebraic equations were performed using Microsoft Excel® (Excel 2010, Microsoft).
Mathematical regressions to differential equations and numerical solutions of differential
equations were performed using the ODE45 solver implemented in Matlab® (Matlab 2010,
Mathworks). Docking simulations were realized with Autodock 4.0 and the estimation of the
environmental impact was partly performed using EATOS (Environmental Assessment Tool for
Organic Syntheses).
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
Enzyme Production and Purification. E. coli BL21 (DE3) cells were transformed with the
plasmid pET21a(+) containing the Hisꢀtagged McPAD decarboxylase gene and heterologous
expression was performed as previously reported.50 To prepare cellꢀfree extracts (CFEs), wet
cells (4 g mLꢀ1) were suspended in potassium phosphate (KPi) buffer (0.05 M, pH 7.0) and the
suspension was sonicated (5 mm sonication tip, 70 % power, 50% duty cycle, 3 min) on ice for
three times and centrifuged at 24000 rpm (69673 rcf) for 20 min at 4oC.
Enzyme purification was carried out using a NickelꢀNTAꢀagarose packed column (9 cm high,
conical 0.8 x 4 cm). After washing the column with an equilibration buffer (Tris 20 mM, NaCl
300 mM, imidazole 10 mM, pH 7.4) with 3 column volumes, the CFE was loaded by gravity
flow using a ratio of 0.4 mL of CFE mLꢀ1 of the chromatographic matrix (corresponding to 10
mg total protein/mL of matrix). Unspecifically adsorbed proteins were removed by washing with
5 column volumes of washing buffer (Tris 20 mM, NaCl 300 mM, imidazole 20 mM, pH 7.4)
and McPAD was eluted using 2.5 column volumes of elution buffer (Tris 20 mM, NaCl 300 mM,
imidazole 250 mM, pH 7.4). Protein adsorption was carried out at 4°C, all the other steps were
carried out at room temperature. Imidazole was removed by centrifugation (5000 rpm (3850 rcf)
ACS Paragon Plus Environment
25

Organic Process Research & Development
Page 26 of 33
1
2
3
4
5
6
for 10 min at 4°C) using membraneꢀtubes (10 KDa cutꢀoff) using an exchange buffer (KPi
buffer, 50 mM, pH 7.0) for storage. Obtained CFEs were stored at ꢀ20°C.
7
8
9
Activity Assays. Activity data were acquired via direct initial rate measurements using HPLC
equipped with an UV detection system. Unless otherwise stated, for screening experiments and
kinetic constant determinations, assays were performed in KPi buffer 0.1 M, pH 7.0 in glass vials
(1 mL volume) shaken at 700 rpm. After addition of the appropriate buffer and substrate, the
reaction was started by addition of the enzyme. Samples (50 ꢂL) were withdrawn and diluted
with a water:acetonitrile solution (20:80 + 3% v/v trifluoroacetic acid) to stop the reaction. The
samples were centrifuged (5 min, 13000 rpm/15700 rcf) and analyzed by HPLC. Using the
fermentation protocol and the CFE preparation method described above, 10–20 µL of extract
(corresponding to 0.2–0.5 mg mL–1 total protein in the assay) are appropriate to detect the linear
range (<10 % conversion) within a 3 min. interval. U mLꢀ1 values reported for each experiment
refer to a reference assay performed in 1 mL volume (HPLC glass vials) using KPi buffer 0.1 M
pH 7.0, 5 mM FA, shaking with 700 rpm at 37°C. Apart from the kinetic measurements,
reactions were performed using CFEs. Enzyme concentrations given in the manuscript represent
the target enzyme in the CFE.
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
Determination of Partition Coefficients and Extraction Rates. Partition coefficients were
calculated following extraction progress curves of 5 mM 4ꢀVG in buffer/organic solvents
mixtures (1:1) in 1 mL volume in glass HPLC vials at 37 °C and 700 rpm. The ratios of the
concentration in organic/aqueous phase were calculated at equilibrium, which was reached after
~1 hour. Extraction rates were calculated by fitting the concentration increase in the organic
phase to an exponential function.
ACS Paragon Plus Environment
26

Page 27 of 33
Organic Process Research & Development
1
2
3
4
5
6
7
8
9
Decarboxylation of FA in a 2LPS. The optimized reaction conditions at 100 mM FA
concentration are described here as an example: 194 mg (1 mmol) of ferulic acid were added to a
50 mL thermostated vessel equipped with a magnetic stir bar. After addition of 7 mL of an
aqueous solution composed of KPi buffer (0.5 M, pH 7.0) and 70 mM KOH, the mixture was
stirred vigorously at 37°C until FA was completely solubilized. After the addition of 20 mL of
hexane, the reaction was started by the addition of enzyme solution (3 mL, 90 µg mL–1 final
concentration) to the aqueous layer and stirred at 37°C and 500 rpm. After the reaction was
complete, the organic layer was separated, dried (MgSO4), filtered and evaporated, yielding 118
mg (79% yield) of 4ꢀVG as transparent oil. The identity of the product was assayed comparing
analytical data with the commercial reference material. TLC (ꢭꢮ = 0.7 in hexane:EtOAc 2:1);
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
+
+
GCꢀMS: m/z = 150 [M+], 135 [C8H7O2 ], 107 [C7H7O+], 77 [C6H5 ].
Synthesis of 4-EG. 100 mM 4ꢀVG in hexane was transferred via syringe and under nitrogen
flow into a twoꢀneck reactor previously charged with Pd/C (2.5 mol%, 5 % Pd basis) and
connected to a hydrogen balloon. After three vacuum (2 mbar)ꢀnitrogen cycles, the reaction was
started by the establishment of a hydrogen atmosphere inside the reactor and was carried out for
1 h at room temperature under vigorous stirring. After the reaction was complete, the mixture
was filtered through Celite® under nitrogen gas and the solvent was evaporated to yield a
transparent oil in 80% yield. In the synthesis starting from ferulic acid, the organic layer from the
decarboxylation reactor was transferred to the hydrogenation reactor following the same steps as
mentioned before. The identity of the product was assayed comparing analytical data with
commercial reference material. TLC (ꢭꢮ= 0.8 in hexane:EtOAc 2:1); GCꢀMS: m/z = 152 [M+],
+
+
+
137 [C8H9O2 ], 91 [C7H7 ], 77 [C6H5 ].
ACS Paragon Plus Environment
27

Organic Process Research & Development
Page 28 of 33
1
2
3
4
5
6
7
8
9
Analytical Methods. The reaction components FA, 4ꢀVG and 4ꢀEG were analyzed by an
Agilent LCꢀ1100 HPLC equipped with a diode array detector using a LiChroCART® 250ꢀ4
LiChrospher 100 RPꢀ18 5 ꢂm column, at 25°C. As mobile phases, water/trifluoroacetic acid (0.1
%) and acetonitrile/trifluoroacetic acid (0.1 %) with a flow rate of 1 mL minꢀ1 with a gradient (0–
2 min 15% acetonitrile; 2–10 min 15–100%; 10–12 min 100–15%) were used. FA and 4ꢀEG
were detected at 280 nm, whereas 4ꢀVG was detected at 254 nm. Typical retention times were
7.8 min for FA, 9.8 min for 4ꢀVG and 10.1 min for 4ꢀEG. Calibration curves were set up using
commercial reference materials. 4ꢀVG and 4ꢀEG were analyzed by an Agilent 7890A GC using a
HPꢀ5 (30 m x 320 µm, 0.25 µm) column coupled to an Agilent 5975C quadrupole mass
spectrometer. Method: 80°C for 1 min, 5°C min–1 to 235°C for 34 min (He 1.5 mL minꢀ1).
Retention times for 4ꢀVG and 4ꢀEG were 6.75 and 7.78 min, respectively.
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
Estimation of the Environmental Factor (E-Factor). Eꢀfactors were calculated with the aid of
the freeꢀware software EATOS (Environmental Assessment Tool for Organic Syntheses). When
not specified by the procedure, the following values were assumed: desiccant: 20 g L–1 of
solution; Celite® for filtration: 0.1 g mL–1 of solution; silica gel for chromatography: 20 g g–1 of
product; eluents for chromatography: 500 mL g–1 of product. ‡
ASSOCIATED CONTENT
Supporting Information Available
Details on the characterization of the enzyme, use of alternative organic solvents, experimental
setꢀup, Eꢀfactor analysis.
ACS Paragon Plus Environment
28

Page 29 of 33
Organic Process Research & Development
1
2
3
4
AUTHOR INFORMATION
5
6
7
Corresponding Authors
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
*Tel.: (+49) 040 42878 3018 (A. Liese), (+49) 040 42878 2890 (S. Kara), Fax: (+49) 040 42878
2127, Eꢀmail: liese@tuhh.de; selin.kara@tuhh.de
Author Contributions
The manuscript was written through contributions of all authors. All authors have given approval
to the final version of the manuscript.
Notes
‡ Additional information about the assumptions made is presented in the Supporting
Information.
The authors declare no competing financial interest.
ACKNOWLEDGMENT
The authors gratefully acknowledge Dr. Joerg H. Schrittwieser (Department of Chemistry,
University of Graz) for proofreading the manuscript.
ABBREVIATIONS
PAD, phenolic acid decarboxylases; McPAD, PAD from Mycobacterium colombiense; 4ꢀVG, 4ꢀ
vinylguaiacol; 4ꢀEG, 4ꢀethylguaiacol; Pd/C, palladium on charcoal.
ACS Paragon Plus Environment
29