Structural and functional insights into asymmetric enzymatic dehydration of alkenols

Article abstract of DOI:10.1038/nchembio.2271

The asymmetric dehydration of alcohols is an important process for the direct synthesis of alkenes. We report the structure and substrate specificity of the bifunctional linalool dehydratase isomerase (LinD) from the bacterium Castellaniella defragrans that catalyzes in nature the hydration of β-myrcene to linalool and the subsequent isomerization to geraniol. Enzymatic kinetic resolutions of truncated and elongated aromatic and aliphatic tertiary alcohols (C5-C15) that contain a specific signature motif demonstrate the broad substrate specificity of LinD. The three-dimensional structure of LinD from Castellaniella defragrans revealed a pentamer with active sites at the protomer interfaces. Furthermore, the structure of LinD in complex with the product geraniol provides initial mechanistic insights into this bifunctional enzyme. Site-directed mutagenesis confirmed active site amino acid residues essential for its dehydration and isomerization activity. These structural and mechanistic insights facilitate the development of hydrating catalysts, enriching the toolbox for novel bond-forming biocatalysis.

Full text of DOI:10.1038/nchembio.2271

article
puBLIsHed OnLIne: 9 januarY 2017 | dOI: 10.1038/nCHeMBIO.2271
sꢀꢁꢂcꢀꢂꢁꢃl ꢃꢄꢅ fꢂꢄcꢀioꢄꢃl iꢄꢆighꢀꢆ iꢄꢀo ꢃꢆymmꢇꢀꢁic
ꢇꢄzymꢃꢀic ꢅꢇhyꢅꢁꢃꢀioꢄ of ꢃlkꢇꢄolꢆ
1
1
2
3
3
Bꢇꢀꢀiꢄꢃ M nꢇꢆꢀl , Chꢁiꢆꢀoꢈhꢇꢁ Gꢇiꢄiꢀz , sꢀꢇꢈhꢃꢄiꢇ poꢈꢃ , sꢃꢁi rizꢇk , robꢇꢁꢀ j Hꢃꢆꢇlbꢇck ,
3
3
1
3
2
2
roꢆꢃꢁy sꢀꢇꢈhꢇꢄ , Michꢃꢇl a noblꢇ , Mꢃx-philiꢈꢈ Fiꢆchꢇꢁ , eꢁik C rꢃlꢈh , Hoi tiꢄg Hꢃꢂ , Hꢇꢄꢁy Mꢃꢄ ,
2
2
3
3
2
Mꢂhiꢃꢅiꢄ Omꢃꢁ , johꢃꢄ p tꢂꢁkꢇꢄbꢂꢁg , sꢀꢇꢈhꢇꢄ vꢃꢄ diꢇꢄ , sꢀꢇꢈhꢃꢄiꢇ j Cꢂllꢇꢁ , Giꢅꢇoꢄ Gꢁogꢃꢄ &
1
Bꢇꢁꢄhꢃꢁꢅ Hꢃꢂꢇꢁ *
The asymmetric dehydration of alcohols is an important process for the direct synthesis of alkenes. We report the structure
and substrate specificity of the bifunctional linalool dehydratase isomerase (LinD) from the bacterium Castellaniella defragrans
that catalyzes in nature the hydration of b-myrcene to linalool and the subsequent isomerization to geraniol. Enzymatic kinetic
resolutions of truncated and elongated aromatic and aliphatic tertiary alcohols (C5–C15) that contain a specific signature motif
demonstrate the broad substrate specificity of LinD. The three-dimensional structure of LinD from Castellaniella defragrans
revealed a pentamer with active sites at the protomer interfaces. Furthermore, the structure of LinD in complex with the prod-
uct geraniol provides initial mechanistic insights into this bifunctional enzyme. Site-directed mutagenesis confirmed active site
amino acid residues essential for its dehydration and isomerization activity. These structural and mechanistic insights facilitate
the development of hydrating catalysts, enriching the toolbox for novel bond-forming biocatalysis.
ehydrations of tertiary alcohols are very valuable functional
Tofurtherexpandtheportfolioofavailabledehydratases, thelina-
group interconversions for the preparation of olefins. lool dehydratase isomerase (LinD, EC 4.2.1.127) from Castellaniella
However, most of the chemical reagents applied are highly defragrans strain 65Phen appears to be a good candidate for the
D
1
9–22
acidic or basic and therefore not practical for the dehydration of reversible, selective addition of water to small alkenes . LinD
1–7
alcohols containing acid- or base-sensitive functionalities . The catalyzes the selective hydration of the monoterpenoid β-myrcene
need for milder and more efficient methods for converting hydroxyl to the tertiary alcohol (S) linalool and its isomerization to geraniol
groups into new functional alkenes has led to the development of (Fig. 1). Metabolite studies on the anaerobic degradation pathway
asymmetric enzymatic synthesis approaches, for example, using in C. defragrans elucidate the degradation of β-myrcene via hydra-
dehydratases. Lately, dehydratase catalysts that are involved in the tion to linalool and isomerization to geraniol, followed by oxidation
23
selective addition or elimination of water have received consider- to geranial and then to geranic acid . Hydration and dehydration
8,9
able attention . The nucleophilic addition of water to C=C double experiments using purified protein with β-myrcene or linalool
bonds is a challenging area. This usually acid-catalyzed reaction is demonstrate that the hydration of myrcene is slower than the dehy-
often unsuitable for asymmetric synthesis because of the occur- dration of linalool. Rates for the isomerization and dehydration for
−
1
−1
rence of side reactions, applied reaction conditions and the reaction LinD are in the range of 0.0023 U mg and 0.0068 U mg pro-
equilibrium. Nature offers hydro-lyase enzymes capable of transfer- tein, respectively. Further experimental and structural data reveal
ring water to C=C double bonds (addition or removal of water). similarity with a linalool isomerase from Thauera linaloolentis
24
These biocatalysts are involved in primary metabolism, energy stor- 47Lol and the absence of prosthetic groups, metal ions and addi-
10
25
age and release systems of living organisms . In most cases, enzy- tional cofactors .
matic hydration involves either the asymmetric addition of water
In this work we report on the substrate specificity of native
to isolated C=C double bonds or the Michael addition of water LinD as well as crystal structures of the substrate-free LinD enzyme
11
to α,β-unsaturated carbonyls . Industrial dehydration reactions (‘LinD-free’), its selenomethionine derivative (‘SeMet-LinD’) and
are currently strictly limited to small molecules like isobutanol, that of SeMet-LinD in complex with the natural substrate geraniol.
1
,4-butanediol, ethylene glycol and 1-phenylethanol that are valu- More notably, the structure of the SeMet-LinD–geraniol complex
able platform chemicals for important polymers, plastics, drugs and along with site-directed mutagenesis studies of amino acids in the
solvents. Recently, Genencor International and DuPont constructed active site allowed us to propose unusual mechanisms for enzymatic
a recombinant Escherichia coli strain for the large-scale production dehydration and isomerization of tertiary alcohols, and provided
of 1,3-propanediol from glucose using coenzyme-B -dependent a new platform for engineering bifunctional dehydratases for the
1
2
1
2,13
dehydratase enzymes . The activity of glycerol dehydratase is preparative dehydration of industrially interesting substrates.
the limiting step in production, especially at high concentrations
of glycerol, and thus its improvement is highly desirable for the RESULTS
1
4,15
biosynthesis of 1,3-propanediol . In this context, the cofactor- Substrate specificity using wild-type LinD
16
17
independent oleate hydratase , kievitone hydratase and promis- To assess the substrate specificity of LinD, we evaluated a range of
18
cuous phenolic acid decarboxylase represent possible alternatives. sterically demanding tertiary alcohols similar to the natural sub-
However, the restricted substrate specificities of these enzymes have strate. We also tested two monoterpenoids for isomerization activity.
so far limited their applicability.
Preliminary activity tests with soluble cell-free extracts of the
1
2
iꢀsꢁꢆꢁꢂꢁꢃ ꢅf tꢃꢄhꢀꢆꢄaꢇ bꢆꢅꢄhꢃmꢆsꢁry, uꢀꢆꢉꢃrsꢆꢁaꢃꢁ Sꢁꢂꢁꢁgarꢁ, Sꢁꢂꢁꢁgarꢁ, Gꢃrmaꢀy. Yꢅrk Sꢁrꢂꢄꢁꢂraꢇ bꢆꢅꢇꢅgy laꢈꢅraꢁꢅry, dꢃꢊarꢁmꢃꢀꢁ ꢅf chꢃmꢆsꢁry, uꢀꢆꢉꢃrsꢆꢁy ꢅf Yꢅrk,
3
Hꢃsꢇꢆꢀgꢁꢅꢀ, Yꢅrk, uK. Gꢃꢀꢅmaꢁꢆꢄa, iꢀꢄ., Saꢀ dꢆꢃgꢅ, caꢇꢆfꢅrꢀꢆa, uSA. *ꢃ-maꢆꢇ: ꢈꢃrꢀharꢋ.haꢂꢃr@ꢆꢁꢈ.ꢂꢀꢆ-sꢁꢂꢁꢁgarꢁ.ꢋꢃ
nature CHeMICaL BIOLOGY | AdvAnce online publicAtion | www.ꢀaꢁꢂrꢃ.ꢄꢅm/ꢀaꢁꢂrꢃꢄhꢃmꢆꢄaꢇꢈꢆꢅꢇꢅgy
1

NꢀTURE CꢁEꢂꢃCꢀL ꢄꢃꢅLꢅꢆꢇ dOI: 10.1038/nCHeMBIO.2271
article
Interestingly, LinD converted the ether analog 3-methoxy-3,7-di-
methylocta-1,6-diene (20) (Table 1). The formation of the elimi-
OH
nation product β-myrcene (3) in small amounts (Supplementary
Results, Supplementary Fig. 2) supported the view that despite
the additional methyl group, the protonation of the oxygen was
facilitated. However, LinD showed no activity toward substrates
2
OH
Linalool dehydratase
isomerase
2
,6-dimethylhept-5-en-2-ol (23), 3,7-dimethyloct-6-en-3-ol (24)
1
and (E)-3-methyloct-4-en-3-ol (25), which lack the terminal vinyl
group (Supplementary Fig. 3). We assumed that the double bond
functions as a stabilizing anchor for a potential carbocation inter-
mediate. In addition to that, the hydroxyl group might be more dif-
3
ficult to protonate as a result of a decreased electron density and pK
a
value. This view was further strengthened by the non-conversion
Figure 1 | Natural reaction and primary structure of LinD. lꢆꢀd-ꢄaꢁaꢇyzꢃꢋ
ꢆsꢅmꢃrꢆzaꢁꢆꢅꢀ ꢅf ꢇꢆꢀaꢇꢅꢅꢇ (1) ꢁꢅ gꢃraꢀꢆꢅꢇ (2) ꢅr ꢋꢃhyꢋraꢁꢆꢅꢀ ꢁꢅ myrꢄꢃꢀꢃ (3).
of compounds (E)-3,7-dimethyl-1,4,6-trien-3-ol (26) and (E)-3,
7
-dimethylocta-4,6-dien-3-ol (27; Supplementary Fig. 3). In com-
parison to 1, the enhanced resonance stability of the conjugated
codon-optimized LinD using racemic linalool (1) and geraniol (2) π-system should lead to an increased stabilization of possible car-
showed functional expression. In accordance with previous reports, bocation intermediates of substrates 26 and 27. From a structural
we identified conversion of the trans-isomer geraniol using both point of view, their conjugated π-systems caused a highly increased
19
geraniol and nerol . We added DTT as a reducing agent to the inflexibility and thus an increased demand in space. This is further
reaction in order to suppress oxygen-dependent inactivation and underlined by comparative analysis of the rotamers of the substrates
to diminish further oxidation of the isomerization products. We 1, 16, 8 and 26 (Fig. 2a). We selected only those rotamers demon-
verified the absence of competing dehydratase as well as isomerase strating the greatest difference between those with an internal C=C
activities of E. coli host cells bearing an empty vector by separate con- double bond and those without. Because of the internal single bond,
trol experiments. In control biotransformations, little background substrates 1 and 16 were more likely to undergo structural changes
activity (<0.5%) was observed, and was measured by subtracting the than 8 and 26. We hypothesized that the overall rigidity of the sub-
uncatalyzedconversionfromtheLinD-catalyzedconversion. Table 1 strate and the additional methyl group of substrate 26 would inter-
shows the conversion of substrates applied in biotransformations fere with the active site, preventing a productive binding in contrast
after 20 h reaction time. We tested alternative substrates like race- to substrate 8, which was accepted by LinD.
mic nerolidol (18) and farnesol (22) that differ only by one isoprene
Furthermore, as previously indicated, the dehydration of 1
moiety from 1 and 2, respectively. We observed conversions of both revealed that the deprotonation takes place at the methyl group adja-
tested substrates, and could detect farnesene (19) as a product using cent to the hydroxyl group. In order to determine the importance
reference material. During the conversions of prenol (21) and farne- of this methyl group for dehydration, we synthesized the second-
sol (22), we obtained only small amounts of isomerization products ary alcohol 28 lacking the geminal methyl group (Supplementary
methyl-1-penten-3-ol (10) and nerolidol (18); yet the equilibrium Fig. 3); however, we obtained no conversion. The methyl hydrogen
here is also far on the side of the dehydrated products isoprene (11) was essential for deprotonation caused by the non-accessibility for
and farnesene (19). Moreover, LinD also accepted various lina- a hydrogen atom to the alkyl chain adjacent to the methyl group. In
lool analogs including racemic 3-ethyloct-1-en-3-ol (6), (E)-4,8- addition, we tested the novel linalyl amine (29) and the alkyne sub-
dimethylnona-2,7-dien-4-ol (4), (E)-3-methylocta-1,4-dien-3-ol (8) strate 3,7-dimethyloct-6-en-1-yn-3-ol (30; Supplementary Fig. 3).
and the non-chiral 10. Inspired by these results, we then explored The amino group in its protonated form is a slightly poorer leaving
the chain-length selectivity in LinD-catalyzed dehydration reac- group than water, owing to the higher nucleophilicity of ammonia.
tions by studying linalool derivatives. Remarkably, LinD displayed However, the biotransformation showed no elimination of ammonia.
excellent activity toward truncated substrates 12 and 14 and the We also found that 29 inhibited LinD. For example, the simultaneous
aromatic derivative 3-methyl-5-phenylpent-1-en-3-ol (16). In our utilization of 10 mM 1 and varying concentrations (0.01–10 mM) of
investigations, we performed dehydration reactions within a range amine 29 in biotransformations resulted in almost complete inacti-
of pH values 6–8. Within our experimental error, the regioselec- vation (>98%) of LinD even when the amine concentration was low
tivity remained unchanged in this pH range. Excellent selectivities (0.01 mM amine; Supplementary Table 1). Also, the alkyne substrate
demonstrated exclusive conformational control in dehydratase- 30 did not show any conversion. This can be attributed to the rigid-
catalyzed kinetic resolutions of the natural linalool 1 and tertiary ity of the non-flexible alkyne function, which prevents the substrate
alcohols 4–20 (Table 1). We performed initial biotransformations from binding in an active conformation into the active site.
with LinD using enantiopure (R) and (S) linalool and observed
no dehydration activity for (R) linalool. This result is consistent ꢅverall structure of LinD
with previous studies on the stereochemistry of LinD, according LinD is an unusual enzyme and possesses only short sections of
to which (S) linalool was formed in the hydration of β-myrcene sequence homology to proteins of known structures in the Protein
(
3) (ref. 26). Transformations of 1 and its analogs resulted in Data Bank (PDB) (Supplementary Fig. 4). We crystallized LinD and
highly enantioselective kinetic resolutions of tertiary alcohols with obtained a 1.8-Å data set on native crystals; however, none of the
selectivities exceeding 200.
attempted solutions using molecular replacement were successful.
Solution of the structure therefore required the preparation of a sele-
nomethionine derivative (SeMet-LinD), which was expressed and
Structural requirements for substrates
In order to study and specify the requirement of an essential purified with equal facility to the native protein (Supplementary
α-methylallyl alcohol signature motif of substrates, we synthesized Fig. 1). Comparative mass-spectrometric analyses of native and
a panel of linalool derivatives by either removing the methyl and SeMet-LinD revealed incorporation of ten selenium atoms within
vinyl groups or inserting double bonds. We also included alterna- the SeMet-LinD, as expected from the amino acid sequence. SeMet-
tive substrates such as ethers, as we assumed that the protonation LinD was then crystallized, and data collected on those crystals was
of the ether oxygen should be accomplished in equal measure as the used to solve the structure. We determined structures of native and
protonation of the hydroxyl group because of similar pK values. SeMet-LinD as well as a structure of the SeMet-LinD in complex
a
2
nature CHeMICaL BIOLOGY | AdvAnce online publicAtion | www.ꢀaꢁꢂrꢃ.ꢄꢅm/ꢀaꢁꢂrꢃꢄhꢃmꢆꢄaꢇꢈꢆꢅꢇꢅgy

NꢀTURE CꢁEꢂꢃCꢀL ꢄꢃꢅLꢅꢆꢇ dOI: 10.1038/nCHeMBIO.2271
article
with geraniol (Supplementary Tables 2 and 3). In the asymmetric solution, we noticed that the five molecules A to E associate forming
unit we observed five molecules in each crystal structure. In contrast a toroidal pentamer (Fig. 2b) with internal and external diameters
19
to earlier reports , in which LinD was suggested to be a tetramer in of approximately 32 Å and 108 Å, respectively.
Table 1 | LinD-catalyzed dehydration and isomerization of substrates (10 mꢂ) after 20 h reaction times
dꢇhyꢅꢁꢃꢀioꢄ
sꢂbꢆꢀꢁꢃꢀꢇ
pꢁoꢅꢂcꢀ
Coꢄvꢇꢁꢆioꢄ (%)
sꢇlꢇcꢀiviꢀy fꢃcꢀoꢁ
47.6 ± 2.8
>200
2
1.8 ± 3.0
6a
5a
13.4 ± 1.5
4
7.1 ± 0.5
>200
–
67.5 ± 1.2
46.6 ± 0.7
98
52.5 ± 0.4
>200
4
7.6 ± 0.5
1.2 ± 0.6
45
2
11
3.0 ± 0.1
nd
ꢃsomerization
9
.1 ± 1.9
nA
nA
21.5 ± 3.6
52.2 ± 13.1
nA
anꢅ ꢈasꢃꢇꢆꢀꢃ sꢃꢊaraꢁꢆꢅꢀ ꢅf sꢂꢈsꢁraꢁꢃ ꢃꢀaꢀꢁꢆꢅmꢃrs. nd, ꢀꢅꢁ ꢋꢃꢁꢃrmꢆꢀꢃꢋ; nA, ꢀꢅꢁ aꢊꢊꢇꢆꢄaꢈꢇꢃ; −, ꢄꢅmꢊꢅꢂꢀꢋ ꢀꢅꢁ ꢄhꢆraꢇ.
daꢁa arꢃ rꢃꢊrꢃsꢃꢀꢁꢃꢋ as mꢃaꢀ ± s.ꢋ. exꢊꢃrꢆmꢃꢀꢁs wꢃrꢃ ꢊꢃrfꢅrmꢃꢋ as ꢁrꢆꢊꢇꢆꢄaꢁꢃs (n = 3).
nature CHeMICaL BIOLOGY | AdvAnce online publicAtion | www.ꢀaꢁꢂrꢃ.ꢄꢅm/ꢀaꢁꢂrꢃꢄhꢃmꢆꢄaꢇꢈꢆꢅꢇꢅgy
3

NꢀTURE CꢁEꢂꢃCꢀL ꢄꢃꢅLꢅꢆꢇ dOI: 10.1038/nCHeMBIO.2271
a b
article
Each monomer of LinD comprised a α/α barrel in which a core of
6
E
six α-helices surrounded a cavity, which, as identified by the geran-
iol complex, constitutes the active site of the enzyme. The struc-
ture of the monomer LinD is shown in Figure 2b (Supplementary
Fig. 4 depicts the relationship between the sequence and secondary
structure of LinD monomer). An intramolecular disulfide bridge,
as suggested by mass-spectrometry analysis on native LinD, was
formed between Cys49 and Cys102 in each monomer, with a dis-
A
D
27
tance between Cβ atoms of 3.8 Å. A DALI search of structurally
related monomers revealed that LinD presents secondary structural
elements comparable to those of the rhamnogalacturonyl hydro-
lase (PDB code 2GH4) (ref. 28) classified in glycoside hydrolase
family 105 from Bacillus subtilis and cellobiose 2-epimerase (PDB
code 3VW5) (ref. 29) from Ruminococcus albus. Both hydrolase
and sugar epimerase have a low sequence identity with LinD (8%
and 6% sequence identity, respectively) and calculated r.m.s. devia-
tions between 3.2 Å over 278 C-alphas (Z-score 22) or 3.4 Å over
B
C
c
d
2
95 C-alphas (Z-score 21.5). A superimposition of the structure of
LinD with the rhamnogalacturonyl hydrolase illustrates the homol-
ogy between these enzymes. Interestingly, the reaction catalyzed by
rhamnogalacturonyl hydrolase features aspects of (de)hydration
3
7–42
28
Figure 2 | Structure of C. defragrans linalool dehydratase isomerase (LinD,
PDꢄ code 5ꢆ1W). (a) pꢅꢁꢃꢀꢁꢆaꢇ sꢂꢈsꢁraꢁꢃ rꢅꢁamꢃrs ꢅf (ꢇꢃfꢁ) 3-mꢃꢁhyꢇ-5-
ꢊhꢃꢀyꢇꢊꢃꢀꢁ-1-ꢃꢀ-3-ꢅꢇ 16 (ꢁwꢅ rꢅꢁamꢃrs ꢆꢀ ꢈꢇꢂꢃ), (E)-3-mꢃꢁhyꢇꢅꢄꢁa-1,
-ꢋꢆꢃꢀ-3-ꢅꢇ 8 (ꢁhrꢃꢃ rꢅꢁamꢃrs ꢆꢀ ꢄyaꢀ) rꢃꢇaꢁꢆꢉꢃ ꢁꢅ ꢁhꢃ sꢂꢈsꢁraꢁꢃ ꢇꢆꢀaꢇꢅꢅꢇ 1
ꢁhrꢃꢃ rꢅꢁamꢃrs hꢆghꢇꢆghꢁꢃꢋ ꢆꢀ grꢃꢃꢀ), aꢀꢋ (rꢆghꢁ) ꢇꢆꢀaꢇꢅꢅꢇ aꢀaꢇꢅg 8 (ꢁhrꢃꢃ
rꢅꢁamꢃrs ꢆꢀ ꢄyaꢀ) rꢃꢇaꢁꢆꢉꢃ ꢁꢅ (E)-3,7-ꢋꢆmꢃꢁhyꢇꢅꢄꢁa-1,4,6-ꢁrꢆꢃꢀ-3-ꢅꢇ 26
fꢅꢂr rꢅꢁamꢃrs ꢆꢀ ꢊꢆꢀk). (b) Sꢁrꢂꢄꢁꢂrꢃ ꢅf a lꢆꢀd ꢊꢃꢀꢁamꢃr shꢅwꢀ ꢆꢀ rꢆꢈꢈꢅꢀ
fꢅrmaꢁ. (c) Sꢁrꢂꢄꢁꢂrꢃ ꢅf a lꢆꢀd mꢅꢀꢅmꢃr ꢆꢀ rꢆꢈꢈꢅꢀ fꢅrmaꢁ wꢆꢁh hꢃꢇꢆꢄꢃs ꢆꢀ
ꢇꢆghꢁ ꢈꢇꢂꢃ aꢀꢋ sꢁraꢀꢋs ꢆꢀ grꢃꢃꢀ. Sꢃꢇꢃꢄꢁꢃꢋ hꢃꢇꢆꢄꢃs aꢀꢋ ꢁhꢃ ꢇꢅꢅꢊ ꢄꢅmꢊrꢆsꢆꢀg
rꢃsꢆꢋꢂꢃs 37–42, ꢊarꢁ ꢅf ꢁhꢃ aꢄꢁꢆꢉꢃ sꢆꢁꢃ wꢆꢁh a ꢀꢃꢆghꢈꢅrꢆꢀg mꢅꢀꢅmꢃr, arꢃ
ꢇaꢈꢃꢇꢃꢋ. (d) Aꢄꢁꢆꢉꢃ sꢆꢁꢃ ꢅf lꢆꢀd ꢆꢀ ꢄꢅmꢊꢇꢃx wꢆꢁh ꢁhꢃ ꢀaꢁꢂraꢇ sꢂꢈsꢁraꢁꢃ
gꢃraꢀꢆꢅꢇ shꢅwꢀ ꢆꢀ gray sꢁꢆꢄks. baꢄkꢈꢅꢀꢃ aꢀꢋ sꢆꢋꢃ ꢄhaꢆꢀ aꢁꢅms ꢅf sꢂꢈꢂꢀꢆꢁs
A aꢀꢋ b arꢃ shꢅwꢀ ꢆꢀ ꢈꢇꢂꢃ aꢀꢋ gꢅꢇꢋ sꢁꢆꢄks, rꢃsꢊꢃꢄꢁꢆꢉꢃꢇy. Sꢃꢇꢃꢄꢁꢃꢋ hyꢋrꢅgꢃꢀ
ꢈꢅꢀꢋs arꢃ shꢅwꢀ ꢈy ꢋashꢃꢋ ꢇꢆꢀꢃs. eꢇꢃꢄꢁrꢅꢀ ꢋꢃꢀsꢆꢁy maꢊs ꢆꢀ ꢈꢇꢂꢃ aꢀꢋ grꢃꢃꢀ
ꢄꢅrrꢃsꢊꢅꢀꢋ ꢁꢅ ꢁhꢃ 2F − F rꢃfꢆꢀꢃꢋ aꢀꢋ F − F ꢅmꢆꢁ maꢊs ꢄꢅꢀꢁꢅꢂrꢃꢋ aꢁ ꢇꢃꢉꢃꢇs
chemistry as part of its mechanism . From the perspective of ter-
pene biochemistry, the overall α/α barrel of LinD is also somewhat
6
similar to farnesyl transferase (for example, farnesyl transferase PDB
code 1FT2, 12% sequence identity, r.m.s. deviations 3.79 Å), estab-
lishing an evolutionary relationship with the β-domain of class II
4
(
30
terpenoid synthases .
(
There were significant differences between family 105 glyco-
side hydrolases, terpene synthases and LinD, however, as might be
expected from their relative dissimilarity in sequence. The first two
exist as catalytically competent monomers, in which substrate binds
within one active site. There was little or no conservation of active
site residues between these enzymes. Protein interfaces, surfaces
31
and assemblies (PISA) analysis of the LinD structure indicated
the pentamer as the native biological assembly in this enzyme. The
interfacial area between monomers calculated by PISA was between
approximately 1,120 and 1,155 Å, with ΔG values for solvation free-
energy gain upon formation of the interface of between −10.7 and
ꢅ
ꢄ
ꢅ
ꢄ
ꢅf 1 σ aꢀꢋ 2.5 σ, rꢃsꢊꢃꢄꢁꢆꢉꢃꢇy. thꢃ ꢇaꢁꢁꢃr was ꢅꢈꢁaꢆꢀꢃꢋ ꢈy rꢃfꢆꢀꢃmꢃꢀꢁ ꢆꢀ ꢁhꢃ
aꢈsꢃꢀꢄꢃ ꢅf ꢇꢆgaꢀꢋ aꢁꢅms, whꢆꢄh haꢉꢃ ꢈꢃꢃꢀ aꢋꢋꢃꢋ sꢂꢈsꢃqꢂꢃꢀꢁꢇy fꢅr ꢄꢇarꢆꢁy.
−1
1
2.5 kJ mol . Additionally, oligomerization resulted in the contri-
bution of residues from two monomers to the putative active site allowed the first identification of residues that are involved in catal-
at the interface of two monomers. A loop between helices α2 and ysis. LinD is a bifunctional enzyme catalyzing two quite different
α3, which extended between residues 37 and 42 (shown in Fig. 2c), but sequential reactions at a single active site. Comparison with the
appeared to form a lid for the active site in the neighboring mono- apo structure also revealed a considerable amount of movement
mer, and both Asp39 and Phe40, in addition to Tyr45, projected into within the active site. The position of geraniol identified a number
the cavity from the loop (Fig. 2d). The active site is made up of amino of amino acid residues in proximity to the ligand essential for dehy-
acid residues Asp39 and Tyr45 from chain B and Met125, Cys171, dration and isomerization: residues Cys171, Cys180, Met125 and
Cys180, His129, Gln179 and Tyr420 from chain A. Furthermore, His129 from chain A and the two residues from chain B Asp39 and
the side chain of (B)Asp39 formed a hydrogen bond with the phe- Tyr45. In addition to the formation of the disulfide bridge between
nolic hydroxyl of (A)Tyr240, which in turn made a hydrogen bond Cys171 and Cys180, the key observations were: (a) the methyl group
with the side chain of Gln179 (Fig. 2d). (B)Tyr45 formed a hydro- of geraniol (methyl at the stereocenter of linalool) was placed near
gen bond with the side chain of (A)Glu172. In addition to these the side chains of (B)Asp39 and (B)Tyr45; (b) the geraniol hydroxyl
residues, this region of the active site was bordered by (A)Phe177, formed a hydrogen bond contact with the side chain of His129; and
(
A)Tyr66, (A)Met125 and two cysteine residues, (A)Cys171 and (A) (c) the side chain of Met125 in the apo structure blocked contact
Cys180, which, with Cβ atoms 5.5 Å apart, did not form a disulfide between His129 and Cys171, but in the geraniol complex Met125
bridge in the apo structure but were themselves bridged in some moved to accommodate the geraniol hydroxyl. The prenyl group of
subunits by spherical density consistent with the presence of a water geraniolprojectedintoahydrophobicpocketoftheactivesite,formed
molecule. However, in the structure of the LinD–geraniol complex, by residues including (A)Trp244, (A)Val293, (A)Leu295, (A)L341,
a disulfide bridge was formed between the two cysteine residues (A)F299 and (A)F70, any of which may be important in dictating
(
Fig. 2d). It is not clear at this stage if the formation of the active site substrate specificity in LinD.
disulfide was part of the catalytic cycle.
One of the remarkable characteristics of LinD is its ability to
switch between elimination and isomerization activity. However, the
mechanism by which the bifunctional LinD promotes the reverse
ꢂechanism of LinD
To date, to our knowledge, nothing is known about the mechanism isomerization and dehydration or hydration had not yet been inves-
of dehydration and isomerization in LinD. Thus, the structural data tigated experimentally. Our observations permitted the formulation
of LinD were used to inform discussion of possible reaction mecha- of two mechanistic hypotheses in which LinD promotes the isomer-
nisms. As a result of this work, the complex of LinD with geraniol ization and dehydration of linalool through either a covalent or a
4
nature CHeMICaL BIOLOGY | AdvAnce online publicAtion | www.ꢀaꢁꢂrꢃ.ꢄꢅm/ꢀaꢁꢂrꢃꢄhꢃmꢆꢄaꢇꢈꢆꢅꢇꢅgy

NꢀTURE CꢁEꢂꢃCꢀL ꢄꢃꢅLꢅꢆꢇ dOI: 10.1038/nCHeMBIO.2271
article
Tyr240
Tyr240
O
O
H
H
Asp39
Asp39
O
O
O–
O–
H
Tyr45
H
Tyr45
O
H+
O
N
N
NH
NH
HO
SH
Cys171
SH
Cys171
S–
Cys180
S–
Cys180
His129
His129
Met125
S
S
Met125
Isomerization to geraniol
Dehydration to myrcene
Tyr240
Tyr240
Tyr240
Tyr240
O
H
O
H
O
O
H
Asp39
H
Asp39
Asp39
Asp39
O
O
O
O
O–
H
HO
O–
H
HO
HO
Tyr45
Tyr45
H
Tyr45
H
HO
O–
O–
H
Tyr45
O
H
S
N
H
O
N
O
NH
H
O
N
NH
H
S
N
NH
NH
O
H
S
H
S
O
H
S
S–
Cys171
S–
H
Cys171
H
S–
Cys171
His129
Met125
Cys180
S–
His129
Met125
S–
His129
Met125
His129
Cys171
Cys180
S
Cys180
S
Cys180
Met125
Figure 3 | ꢂechanistic hypotheses for the dehydration and isomerization reaction catalyzed by LinD. iꢀ ꢁhꢃ ꢄꢅꢉaꢇꢃꢀꢁ mꢃꢄhaꢀꢆsm (ꢇꢃfꢁ), cys171 ꢄaꢁaꢇyzꢃs
ꢁ
ꢁ
ꢁ
hꢃ ꢊrꢅꢁꢅꢀaꢁꢆꢅꢀ ꢅf ꢁhꢃ (S) aꢇꢄꢅhꢅꢇ, aꢇꢇꢅwꢆꢀg ꢆꢁ ꢁꢅ ꢇꢃaꢉꢃ as waꢁꢃr, whꢆꢇꢃ cys180 aꢄꢁs as a ꢀꢂꢄꢇꢃꢅꢊhꢆꢇꢃ ꢁꢅ aꢁꢁaꢄk ꢇꢆꢀaꢇꢅꢅꢇ aꢁ ꢁhꢃ ꢁꢃrmꢆꢀaꢇ aꢇkꢃꢀꢃ. thꢆs rꢃsꢂꢇꢁs ꢆꢀ
hꢃ fꢅrmaꢁꢆꢅꢀ ꢅf a ꢄꢅꢉaꢇꢃꢀꢁ ꢁhꢆꢅꢁꢃrꢊꢃꢀꢃ ꢆꢀꢁꢃrmꢃꢋꢆaꢁꢃ ꢈꢃꢁwꢃꢃꢀ cys180 aꢀꢋ ꢁhꢃ ꢇꢆꢀaꢇꢅꢅꢇ rꢃsꢆꢋꢂꢃ. thꢃ sꢁrꢆꢄꢁ rꢃqꢂꢆrꢃmꢃꢀꢁ fꢅr cys180 sꢂꢊꢊꢅrꢁs a mꢃꢄhaꢀꢆsm
haꢁ ꢆꢀꢉꢅꢇꢉꢃs ꢀꢂꢄꢇꢃꢅꢊhꢆꢇꢆꢄ aꢁꢁaꢄk ꢈy a ꢁhꢆꢅꢇaꢁꢃ ꢁꢅ gꢆꢉꢃ ꢁhꢆs ꢆꢀꢁꢃrmꢃꢋꢆaꢁꢃ. Fꢅr ꢆsꢅmꢃrꢆzaꢁꢆꢅꢀ ꢁꢅ gꢃraꢀꢆꢅꢇ (grꢃꢃꢀ), ꢁhꢃ rꢃꢇꢅꢄaꢁꢆꢅꢀ ꢅf ꢁhꢃ Mꢃꢁ125 sꢆꢋꢃ ꢄhaꢆꢀ
aꢇꢇꢅws ꢁhꢃ ꢈꢆꢀꢋꢆꢀg aꢀꢋ aꢄꢁꢆꢉaꢁꢆꢅꢀ ꢅf waꢁꢃr ꢈy ꢁhꢃ ꢀꢃarꢈy Hꢆs129. Hꢆs129 ꢁhꢃꢀ aꢄꢁꢆꢉaꢁꢃs waꢁꢃr fꢅr hyꢋrꢅꢇysꢆs ꢅf ꢁhꢃ ꢄꢅꢉaꢇꢃꢀꢁ ꢆꢀꢁꢃrmꢃꢋꢆaꢁꢃ ꢁꢅ fꢅrm gꢃraꢀꢆꢅꢇ.
cꢅꢀꢉꢃrsꢆꢅꢀ ꢅf ꢁhꢃ ꢆꢀꢁꢃrmꢃꢋꢆaꢁꢃ ꢁꢅ myrꢄꢃꢀꢃ (ꢈꢇꢂꢃ) ꢆs sꢂggꢃsꢁꢃꢋ ꢁꢅ ꢅꢄꢄꢂr ꢉꢆa ꢁhꢃ aꢈsꢁraꢄꢁꢆꢅꢀ ꢅf ꢁhꢃ ꢊrꢅꢁꢅꢀ frꢅm ꢁhꢃ ꢇꢆꢀaꢇꢅꢅꢇ mꢃꢁhyꢇ grꢅꢂꢊ ꢈy tyr45,
ꢋꢃꢊrꢅꢁꢅꢀaꢁꢃꢋ ꢈy a ꢀꢃarꢈy Asꢊ39. thꢆs rꢃsꢂꢇꢁs ꢆꢀ ꢁhꢃ fꢅrmaꢁꢆꢅꢀ ꢅf β-myrꢄꢃꢀꢃ aꢀꢋ ꢆꢁs rꢃꢇꢃasꢃ frꢅm ꢁhꢃ aꢄꢁꢆꢉꢃ sꢆꢁꢃ. iꢀ aꢀ aꢄꢆꢋ–ꢈasꢃ mꢃꢄhaꢀꢆsm (rꢆghꢁ), ꢁhꢃ
ꢆsꢅmꢃrꢆzaꢁꢆꢅꢀ (grꢃꢃꢀ) aꢀꢋ ꢋꢃhyꢋraꢁꢆꢅꢀ (ꢈꢇꢂꢃ) ꢅf ꢇꢆꢀaꢇꢅꢅꢇ ꢅꢄꢄꢂr ꢁhrꢅꢂgh ꢁhꢃ fꢅrmaꢁꢆꢅꢀ ꢅf a ꢄarꢈꢅꢄaꢁꢆꢅꢀ ꢆꢀꢁꢃrmꢃꢋꢆaꢁꢃ, gꢃꢀꢃraꢁꢃꢋ agaꢆꢀ ꢁhrꢅꢂgh ꢊrꢅꢁꢅꢀaꢁꢆꢅꢀ
ꢅf ꢁhꢃ (S) aꢇꢄꢅhꢅꢇ aꢀꢋ rꢃꢇꢃasꢃ ꢅf waꢁꢃr. dꢃꢊrꢅꢁꢅꢀaꢁꢆꢅꢀ ꢅf ꢁhꢃ mꢃꢁhyꢇ grꢅꢂꢊ ꢈy tyr45 qꢂꢃꢀꢄhꢃs ꢁhꢃ ꢄarꢈꢅꢄaꢁꢆꢅꢀ ꢆꢀ ꢁhꢃ ꢋꢃhyꢋraꢁꢆꢅꢀ rꢃaꢄꢁꢆꢅꢀ ꢁꢅ β-myrꢄꢃꢀꢃ;
aꢋꢋꢆꢁꢆꢅꢀ ꢅf waꢁꢃr, aꢄꢁꢆꢉaꢁꢃꢋ ꢈy cys180 ꢅr Hꢆs129, qꢂꢃꢀꢄhꢃs ꢁhꢃ ꢄarꢈꢅꢄaꢁꢆꢅꢀ ꢆꢀ ꢁhꢃ ꢆsꢅmꢃrꢆzaꢁꢆꢅꢀ rꢃaꢄꢁꢆꢅꢀ ꢁꢅ gꢃraꢀꢆꢅꢇ. thꢃ hyꢋrꢅꢊhꢅꢈꢆꢄ ꢃꢀꢉꢆrꢅꢀmꢃꢀꢁ ꢅf ꢁhꢃ
aꢄꢁꢆꢉꢃ sꢆꢁꢃ mꢆghꢁ sꢂꢊꢊꢅrꢁ ꢁhꢃ sꢁaꢈꢆꢇꢆzaꢁꢆꢅꢀ ꢅf ꢁhꢃ ꢄarꢈꢅꢄaꢁꢆꢅꢀ ꢆꢀꢁꢃrmꢃꢋꢆaꢁꢃ fꢅrmꢃꢋ.
carbocation intermediate (Fig. 3). In the first proposal, linalool binds configuration, and a covalent LinD–terpene complex is formed as
into the active site, and the terminal vinyl group of the substrate is thehydroxyliseliminatedaswater. Thecovalentintermediateunder-
attacked by Cys180, with the thiolate being in close proximity to the goes either hydrolysis, by a water molecule activated by His129, to
substrate. The involvement of a cysteine-based covalent intermedi- give geraniol in an isomerization reaction, or base-catalyzed elimi-
ate in a hydration reaction has precedent in the mechanism of iso- nation, most likely by Asp39 or Tyr45 at the methyl group, to give
32
cyanide dehydratase from Pseudomonas fluorescens . A thioterpene myrcene in a dehydration process. In the second proposal, involv-
bond through cysteine also has precedent in the well-studied CaaX ing a carbocation intermediate, Cys180 does not interact directly
box of proteins that are the target of farnesyl or geranyl synthases, with the substrate terminal C=C double bond. Dehydration of (S)
33
and through which the prenylation of proteins is effected . Cys171 linalool, catalyzed by Cys171, results in a carbocation intermediate
is well positioned to protonate the linalool hydroxyl of the (S) that undergoes either rehydration, catalyzed by His129 or Cys180
Table 2 | Specific activity of LinD point mutantsa
rꢇlꢃꢀivꢇ ꢃcꢀiviꢀyb
Coꢄvꢇꢁꢆioꢄ of gꢇꢁꢃꢄiol
ꢀo myꢁcꢇꢄꢇ
dꢇhyꢅꢁꢃꢀioꢄ of liꢄꢃlool
Iꢆomꢇꢁizꢃꢀioꢄ of gꢇꢁꢃꢄiol
Iꢆomꢇꢁizꢃꢀioꢄ of liꢄꢃlool
b
b
b
Liꢄd vꢃꢁiꢃꢄꢀ
Wꢆꢇꢋ ꢁyꢊꢃ
tyr45phꢃ
ꢀo myꢁcꢇꢄꢇ
ꢀo liꢄꢃlool
ꢀo gꢇꢁꢃꢄiol
100%
22%
100%
0.5%
<0.5%
6%
100%
<0.5%
16%
100%
<0.5%
<0.5%
7%
Mꢃꢁ125Aꢇa
Hꢆs129Aꢇa
cys171Aꢇa
2%
23%
2%
<0.5%
<0.5%
<0.5%
<0.5%
<0.5%
<0.5%
<0.5%
cys180Aꢇa
a
<0.5%
ꢈ
Raꢁꢃs rꢃꢊꢅrꢁꢃꢋ ꢆꢀ ꢂꢀꢆꢁs ꢅf μmꢅꢇ ꢊrꢅꢋꢂꢄꢁ fꢅrmꢃꢋ ꢊꢃr h ꢊꢃr mg ꢅf ꢇysaꢁꢃ. vaꢇꢂꢃs arꢃ ꢀꢅrmaꢇꢆzꢃꢋ ꢁꢅ ꢁhꢃ wꢆꢇꢋ-ꢁyꢊꢃ raꢁꢃ fꢅr ꢁhꢃ gꢆꢉꢃꢀ rꢃaꢄꢁꢆꢅꢀ. nd, ꢀꢅꢁ ꢋꢃꢁꢃrmꢆꢀꢃꢋ.
daꢁa arꢃ rꢃꢊrꢃsꢃꢀꢁꢃꢋ as mꢃaꢀ ± s.ꢋ. exꢊꢃrꢆmꢃꢀꢁs wꢃrꢃ ꢊꢃrfꢅrmꢃꢋ as ꢁrꢆꢊꢇꢆꢄaꢁꢃs (n = 3).
nature CHeMICaL BIOLOGY | AdvAnce online publicAtion | www.ꢀaꢁꢂrꢃ.ꢄꢅm/ꢀaꢁꢂrꢃꢄhꢃmꢆꢄaꢇꢈꢆꢅꢇꢅgy
5

NꢀTURE CꢁEꢂꢃCꢀL ꢄꢃꢅLꢅꢆꢇ dOI: 10.1038/nCHeMBIO.2271
article
to form geraniol, or base-catalyzed deprotonation to form myrcene derivatives thereof, and the selective dehydration of a broad set of
by Tyr45 and Asp39. tertiary alcohols. Kinetic resolutions of linalool and synthetic ana-
Insupportofthecatalyticroleoftheaminoacidsimplicatedinthese logs by LinD-catalyzed dehydration provided chiral products with
mechanisms, we demonstrated that site-directed mutant proteins— selectivity factors exceeding 200. Because of the sterically demand-
LinD-C171A and LinD-C180A—showed no activity toward geran- ing structure of tertiary alcohols, the synthesis of optically pure
iol and linalool (<0.5% activity; Table 2). Moreover, the substitution tertiary alcohols is still challenging and is achieved via hydrolase-
of methionine at position 125 by alanine (LinD-M125A) yielded catalyzed kinetic resolution of these compounds. In fact, biotransfor-
dramatic loss of activity (2% activity in the conversion of geraniol mations with linalool enantiomers revealed that the (R) enantiomer
to myrcene) and was consistent with the methionine serving as an was not converted by LinD.
2
7,28
antioxidant for the catalytic cysteine residues . It appears that
The isomerization activity of LinD was slightly disfavored com-
Cys171 and Cys180 must be kept in a reduced state for all LinD reac- pared to the dehydration activity. We assumed the catalytically
tions, as evidenced by the requirement for the addition of DTT in all important cysteine residues in the active site and their oxidation
cases for full activity, suggesting that the disulfide bridge observed state would be essential for the bifunctionality of the enzyme. The
34,35
in the geraniol complex is relevant for the catalytic cycle
.
results from our experiments showed that LinD converted various
The structural data supported the role of His129 as a catalytic resi- linalool analogs and derivatives thereof, reflecting its relatively broad
due in activation for the formation of geraniol. However, the activity substrate specificity. Furthermore, through our biocatalytic inves-
of the mutant H129A, which displayed 23% activity of the wild type tigations, we were the first to demonstrate that substrates accepted
for the geraniol–myrcene conversion, suggested that this residue by LinD required a specific α-methylallyl alcohol signature motif.
may not play an essential catalytic role. The ligand conformation in However, the addition of methyl groups onto the linalool methyl and
which the geraniol hydroxyl and His129 are juxtaposed was found allyl substituent of the signature motif resulted in two- and three-
in an active site in which a disulfide bridge was formed between the fold reduced conversions, respectively. Our studies on the substrate
side chains of Cys171 and Cys180, which may have had an effect specificity of LinD indicated that linalool analogs with variations in
on ligand binding. If not involved in a covalent mechanism, the free the nature of the substrate (double bond) substantially influenced the
side chain thiol of Cys180 may be responsible for water activation in dehydration activity. Interestingly, the product of the C5 substrate,
the non-covalent mechanism shown in Figure 3, consistent with the isoprene, is currently used in the industrial production of synthetic
inactivity of the C180A mutant. The presumed location of the methyl rubber. Furthermore, we showed that LinD also accepted aromatic
group of linalool, as indicated by the geraniol complex, showed that substituents and ether analogs of linalool. The significant decrease in
the two amino acid residues best placed to act as a catalytic base for activity with linalool analogs modified at the specific signature motif
the deprotonation of this methyl group in the dehydration of linalool suggested important contributions from active site amino acids to
to myrcene were Tyr45 and Asp39. The relative activity of 0.5% of hydrogen bonding and intermediate stabilization. We assumed that
the Y45F mutant compared to the wild type supports the role for either substrates interfered with the active site, preventing a pro-
Tyr45 in the dehydration of linalool to myrcene, although 22% activ- ductive binding, or reasonably well-bound substrates were difficult
ity was retained for the geraniol–myrcene conversion.
to protonate as a result of decreased electron density. The attempt
The structure of LinD also allowed previous mutational studies, to exchange the hydroxyl moiety of linalool with an amine moiety
36
reported in a recent patent , to be put into a structural context. From resulted in complete inactivation of LinD.
a screen of a library of point mutants, ‘A sp199’ (Asp174 in current
Crystallization of native LinD and SeMet-LinD allowed a more
structure), when mutated to Glu, was reported to have a 6.25-fold detailed structural analysis of this catalyst. The architecture of the
improved activity for the dehydration of prenyl alcohol to isoprene, putative active site allowed us to propose a mechanism for the dehy-
and Val195 (Val170 in our structure) a 6.40-fold improvement in dration and isomerization of monoterpenes linalool and geraniol. We
activity when mutated to Phe. In the first case, Asp174 occurred at obtained insights into a unique active site that harbours two distinct
the subunit interface, close to, but not within, bonding distance of activities, each using the same amino acids to dehydrate or isomer-
(
B)His91 and (B)Arg47, but was somewhat distant from the active ize monoterpene substrates. The elucidated LinD crystal structures
site. Val170 was adjacent in sequence to Cys171, suggested as a cata- together with site-directed mutagenesis of active site residues and
lytic residue in the mechanisms above, but pointed away from the functional characterization allowed us to draw conclusions regard-
active site to the outer periphery of the pentamer. In the absence of ing some features of the reaction mechanism of this unique enzyme:
structural information that would shed light on changes in struc- (a) Cys171, Tyr45 and Asp39 acting as a general acid and base for the
ture consequent on these variants, it is difficult to speculate on the protonation of the leaving hydroxyl group of linalool and dehydra-
basis for improved activity, although these results do suggest that tion at the chiral carbon, (b) activation of water by His129 or Cys180
improvements in LinD activity may be affected through random as and addition to the covalent or carbocation intermediate species
well as structure-guided mutagenesis.
and (c) formation of either a carbocation intermediate or covalent
intermediate between Cys180 and the diene structure. The diene was
deprotonated to form myrcene or hydrolyzed, triggering the forma-
DꢃSCUSSꢃꢅN
Enzymes are currently of increasing interest in the chemical indus- tion of geraniol. Future structural studies will allow the determina-
try. Water-adding and -removing enzymes such as oleate hydratase, tion of the basis for the unique substrate spectrum of LinD. On the
fumarase and enoyl-CoA hydratase with their special proper- basis of these findings, LinD provides an exciting opportunity for
ties demonstrate their potential value in biocatalysis. Dehydratase structure-guided enzyme engineering, particularly in the realization
enzymes that catalyze the regio- and stereoselective dehydration of new biosynthetic pathways in order to gain access to novel indus-
reactions of interest are needed when seeking to exploit the potential trially interesting diene and tertiary alcohol products.
for the production of industrially important conjugated alkenes, for
example, butadiene or isoprene, as well as tertiary alcohols. Previous rꢀꢁꢀꢂvꢀd 6 apꢃꢂꢄ 2016; ꢅꢁꢁꢀpꢆꢀd 3 Novꢀmbꢀꢃ 2016;
pubꢄꢂshꢀd onꢄꢂnꢀ 9 Jꢅnuꢅꢃy 2017
investigations that addressed the role of LinD in the anaerobic deg-
radation of monoterpenes evaluated its ability to selectively hydrate
β-myrcene and to interconvert produced linalool into geraniol. ꢂETꢁꢅDS
We employed an unbiased approach to evaluate the substrate speci- Methods, including statements of data availability and any associated
ficity of the linalool dehydratase isomerase LinD. We confirmed accession codes and references, are available in the online version
the bifunctional role of LinD in the isomerization of geraniol and of the paper.
6
nature CHeMICaL BIOLOGY | AdvAnce online publicAtion | www.ꢀaꢁꢂrꢃ.ꢄꢅm/ꢀaꢁꢂrꢃꢄhꢃmꢆꢄaꢇꢈꢆꢅꢇꢅgy

NꢀTURE CꢁEꢂꢃCꢀL ꢄꢃꢅLꢅꢆꢇ dOI: 10.1038/nCHeMBIO.2271
article
rꢇfꢇꢁꢇꢄcꢇꢆ
23. Heyen, U. & Harder, J. Geranic acid formation, an initial reaction of
1
2
3
4
.
.
.
.
Frija, L.M.T. & Afonso, C.A.M. Amberlyst-15: a reusable heterogeneous
anaerobic monoterpene metabolism in denitrifying Alcaligenes defragrans.
Appl. Environ. Microbiol. 66, 3004–3009 (2000).
catalyst for the dehydration of tertiary alcohols. Tetrahedron 68, 7414–7421
(
2012).
24. Marmulla, ꢀ., Šafaric, B., Marꢂert, S., Schweder, T. & Harder, J. Linalool
isomerase, a membrane-anchored enzyme in the anaerobic monoterpene
degradation in auera linaloolentis 47Lol. BMC Biochem. 17, 1–11 (2016).
25. Weidenweber, S., Marmulla, ꢀ., Ermler, U. & Harder, J. X-ray structure of
linalool dehydratase/isomerase from Castellaniella defragrans reveals
enzymatic alꢂene synthesis. FEBS Lett. 590, 1375–1383 (2016).
26. Lüddeꢂe, F. & Harder, J. Enantiospecific (S)-(+)-linalool formation from
β-myrcene by linalool dehydratase-isomerase. Z. Naturforsch. C 66, 409–412
(2011).
Alvarez-Manzaneda, E.J. et al. Triphenylphosphine-iodine: an efficient reagent
for the regioselective dehydration of tertiary alcohols. Tetrahedr. Lett. 45,
4
453–4455 (2004).
ꢀaju, S., Moret, M.E. & ꢁlein Gebbinꢂ, ꢀ.J.M. ꢀhenium-catalyzed
dehydration and deoxydehydration of alcohols and polyols: opportunities for
the formation of olefins from biomass. ACS Catal. 5, 281–300 (2015).
ꢁantam, M.L., Santhi, P.ꢁ. & Siddiqui, M.F. Montmorillonite-catalyzed
dehydration of tertiary alcohols to olefins. Tetrahedr. Lett. 34, 1185–1186
(
1993).
27. Holm, L. & ꢀosenström, P. Dali server: conservation mapping in 3D.
Nucleic Acids ꢁes. 38, W545–W549 (2010).
5
6
.
.
ꢁantam, M.L., Prasad, A.D. & Santhi, P.L. Molybdenum-catalyzed
dehydration of tertiary alcohols to olefins. Synth. Commun. 23, 45–48 (1993).
Posner, G.H. et al. Boron trifluoride etherate promotes fast, mild, clean and
regioselective dehydration of tertiary alcohols. Tetrahedr. Lett. 32, 6489–6492
28. Itoh, T., Ochiai, A., Miꢂami, B., Hashimoto, W. & Murata, ꢁ. Structure of
unsaturated rhamnogalacturonyl hydrolase complexed with substrate.
Biochem. Biophys. ꢁes. Commun. 347, 1021–1029 (2006).
(
1991).
29. Fujiwara, T. et al. Crystal structure of ꢁuminococcus albus cellobiose
2-epimerase: structural insights into epimerization of unmodified sugar.
FEBS Lett. 587, 840–846 (2013).
7
8
.
.
Zhang, X. et al. Low temperature dehydrations of non-activated alcohols via
halide catalysis. Org. Chem. Front. 3, 701–708 (2016).
Manabe, ꢁ., Iimura, S., Sun, X.-M. & ꢁobayashi, S. Dehydration reactions in
water. Brønsted acid-surfactant-combined catalyst for ester, ether, thioether,
and dithioacetal formation in water. J. Am. Chem. Soc. 124, 11971–11978
30. Oldfield, E. & Lin, F.Y. Terpene biosynthesis: modularity rules. Angew. Chem.
Int. Ed. Engl. 51, 1124–1137 (2012).
31. ꢁrissinel, E. & Henricꢂ, ꢁ. Inference of macromolecular assemblies from
crystalline state. J. Mol. Biol. 372, 774–797 (2007).
(
2002).
9
1
1
1
.
Harmer, M.A. & Sun, Q. Solid acid catalysis using ion-exchange resins.
32. Laꢂshminarasimhan, M., Madzelan, P., Nan, ꢀ., Milꢂovic, N.M. & Wilson, M.A.
Evolution of new enzymatic function by structural modulation of cysteine
reactivity in Pseudomonas fluorescens isocyanide hydratase.
J. Biol. Chem. 285, 29651–29661 (2010).
Appl. Catal. A Gen. 221, 45–62 (2001).
0. ꢀesch, V. & Hanefeld, U. e selective addition of water. Catal. Sci. Technol.
5
, 1385–1399 (2015).
1. Jin, J. & Hanefeld, U. e selective addition of water to C=C bonds; enzymes
are the best chemists. Chem. Commun. (Camb.) 47, 2502–2510 (2011).
2. Tong, I.T., Liao, H.H. & Cameron, D.C. 1,3-Propanediol production by
Escherichia coli expressing genes from the ꢀlebsiella pneumoniae dha regulon.
Appl. Environ. Microbiol. 57, 3541–3546 (1991).
33. Gao, J., Liao, J. & Yang, G.Y. CAAX-box protein, prenylation process and
carcinogenesis. Am. J. Transl. ꢁes. 1, 312–325 (2009).
34. Levine, ꢀ.L., Mosoni, L., Berlett, B.S. & Stadtman, E.ꢀ. Methionine residues
as endogenous antioxidants in proteins. Proc. Natl. Acad. Sci. USA 93,
15036–15040 (1996).
1
1
1
3. Biebl, H., Menzel, ꢁ., Zeng, A.-P. & Decꢂwer, W.-D. Microbial production of
35. Stadtman, E.ꢀ., Mosꢂovitz, J., Berlett, B.S. & Levine, ꢀ.L. Cyclic oxidation
and reduction of protein methionine residues is an important antioxidant
mechanism. Mol. Cell. Biochem. 234-235, 3–9 (2002).
1,3-propanediol. Appl. Microbiol. Biotechnol. 52, 289–297 (1999).
4. Jiang, W., Wang, S., Wang, Y. & Fang, B. ꢁey enzymes catalyzing glycerol to
1
,3-propanediol. Biotechnol. Biofuels 9, 57 (2016).
36. Marliere, P., Delcourt, M. & Mazaleyrat, S. Alꢂenol dehydratase variants.
WO patent 2014184345 A1 (2014).
5. Liu, J.Z., Xu, W., Chistoserdov, A. & Bajpai, ꢀ.ꢁ. Glycerol dehydratases:
biochemical structures, catalytic mechanisms, and industrial applications in
1,3-propanediol production by naturally occurring and genetically engineered
bacterial strains. Appl. Biochem. Biotechnol. 179, 1073–1100 (2016).
6. Volꢂov, A. et al. Myosin cross-reactive antigen of Streptococcus pyogenes M49
encodes a fatty acid double bond hydratase that plays a role in oleic acid
detoxification and bacterial virulence. J. Biol. Chem. 285, 10353–10361
1
1
ackꢄowlꢇꢅgmꢇꢄꢀꢆ
This project has received funding from the European Union’s Horizon 2020 research
and innovation programme (EmPowerPutida) under grant agreement no. 635536
(to B.H.). This work was supported by the Diamond Light Source for access to
beamlines I02 and I04 under proposed number mx-9948 (to G.G.).
(
2010).
7. Turbeꢂ, C.S., Smith, D.A. & Schardl, C.L. An extracellular enzyme from
Fusarium solani f. sp. phaseoli which catalyses hydration of the isoflavonoid
phytoalexin, phaseollidin. FEMS Microbiol. Lett. 73, 187–190 (1992).
8. Wuensch, C. et al. Asymmetric enzymatic hydration of hydroxystyrene
derivatives. Angew. Chem. Int. Ed. Engl. 52, 2293–2297 (2013).
9. Brodꢂorb, D., Gottschall, M., Marmulla, ꢀ., Lüddeꢂe, F. & Harder, J. Linalool
dehydratase-isomerase, a bifunctional enzyme in the anaerobic degradation of
monoterpenes. J. Biol. Chem. 285, 30436–30442 (2010).
aꢂꢀhoꢁ coꢄꢀꢁibꢂꢀioꢄꢆ
1
1
S.P., H.T.H., H.M., M.O., J.P.T. and G.G. conceived and designed the crystallization
experiments and performed 3D structure determination. S.R., R.J.H., R.S., M.A.N.,
E.C.R. S.V.D. and S.J.C. created and characterized all of the LinD mutants used. C.G.,
M.-P.F. and B.M.N. performed biotransformations and contributed to data analysis.
B.M.N., G.G. and B.H. wrote the manuscript. All authors revised and contributed to
the manuscript.
2
2
2
0. Lüddeꢂe, F. et al. Geraniol and geranial dehydrogenases induced in anaerobic
monoterpene degradation by Castellaniella defragrans. Appl. Environ.
Microbiol. 78, 2128–2136 (2012).
Comꢈꢇꢀiꢄg fiꢄꢃꢄciꢃl iꢄꢀꢇꢁꢇꢆꢀꢆ
1. Lüddeꢂe, F., Diꢂfidan, A. & Harder, J. Physiology of deletion mutants in
the anaerobic β-myrcene degradation pathway in Castellaniella defragrans.
BMC Microbiol. 12, 192 (2012).
The authors declare no competing financial interests.
aꢅꢅiꢀioꢄꢃl iꢄfoꢁmꢃꢀioꢄ
2. Foss, S., Heyen, U. & Harder, J. Alcaligenes defragrans sp. nov., description of
four strains isolated on alꢂenoic monoterpenes ((+)-menthene, alpha-pinene,
Any supplementary information, chemical compound information and source data
are available in the online version of the paper. Reprints and permissions information
is available online at http://www.nature.com/reprints/index.html. Correspondence and
requests for materials should be addressed to B.H.
2
2
-carene, and alpha-phellandrene) and nitrate. Syst. Appl. Microbiol. 21,
37–244 (1998).
nature CHeMICaL BIOLOGY | AdvAnce online publicAtion | www.ꢀaꢁꢂrꢃ.ꢄꢅm/ꢀaꢁꢂrꢃꢄhꢃmꢆꢄaꢇꢈꢆꢅꢇꢅgy
7

ꢅNLꢃNE ꢂETꢁꢅDS
LB broth containing 50 μg/ml kanamycin and grown overnight with shaking
Chemicals. All chemicals and solvents were purchased from Sigma-Aldrich at 37 °C. These starter cultures were used to inoculate 100 ml each of Spirulina
(
(
(
Steinheim, Germany), Alfa Aesar (Karlsruhe, Germany) and Carl Roth Simple Medium 5 (SSM5, Algae-Lab) containing 50 μg/ml kanamycin. These
Karlsruhe, Germany): linalool (Sigma-Aldrich; L2602, 97%), (R) linalool cultures were shaken at 250 r.p.m. and induced with 1 mM IPTG at an OD600 of
Sigma-Aldrich; 62139, >95%), myrcene (Sigma-Aldrich; W276200, >95%), far- 0.6–0.75. After induction, cultures were incubated with shaking at 16 °C for 24 h,
nesene (Sigma-Aldrich; W383902), nerolidol (Sigma-Aldrich; H59605, 98%),
final OD600 values measured, and then harvested by centrifugation at 5,000
farnesol (Sigma-Aldrich; F203, 95%), 3-methyl-2-buten-1-ol (prenol; Sigma- r.p.m. at 4 °C for 15 min. Pellets were thawed on ice and lysed using a lysis
Aldrich; W364703, >98%) and geraniol (Alfa Aesar; A13736, 97%).
solution of SoluLyse Tris–HCl buffer pH 7.4 (Genlantis, San Diego, CA, USA)
containing 1× EDTA-free Halt Protease Inhibitor Cocktail (ThermoScientific),
5 mM DTT, 5 μl of rLysozyme Solution (Novagen), and 5 μl of Benzonase
General analytics. The H and 13C NMR spectra were recorded on an Avance
1
5
00 spectrometer (Bruker) operating at 500.15 MHz and 125.76 MHz, respec- Nuclease HC (Novagen). 5 mM EDTA was added to the lysates, then insoluble
tively. All spectra were recorded at room temperature in CDCl . Chemical debris removed by centrifugation at 15,000 r.p.m. Total protein concentration
3
shifts (δ) are expressed in p.p.m. and referenced to tetramethylsilane (TMS; of the clarified lysates was measured by Bradford assay. 200 μL of the appropri-
δ = 0 p.p.m.).
ate lysate (2,530 mg/ml total protein) was added into an amber 2 ml screw cap
GC–MS vial (Phenomenex Verex Certified 9 mm 2 ml Amber 51 vials). 7.5 μl
Gas chromatography (GC). GC analyses were performed on a Shimadzu of geraniol was added to each lysate, for a final concentration of 200 mM, and
GC-2010 equipped with a flame ionization detector using a DB-5 capillary col- then the vial was immediately sealed. The bottles were incubated at 30 °C at
umn (Agilent; 30 m × 250 μm × 0.25 μm) and hydrogen as carrier gas (linear 250 r.p.m. At selected time points, reactions were quenched by placing them
velocity 30 cm/s). Injections were performed in split mode (split ratio 10:1). in an 80 °C water bath for 10 min. Products (geraniol, linalool, myrcene) from
All substrates except nerolidol (18) and farnesol (22) were measured with a the reaction mixtures were extracted with n-hexane (2 × 200 μl) and analyzed
temperature program starting at 80 °C for 3 min, followed by an increase of
0 °C/min to 160 °C, then 40 °C/min to 310 °C, which was held for 2 min.
Nerolidol (18) and farnesol (22) were analyzed using a temperature program
starting at 100 °C for 2 min, then raised 20 °C/min to 130 °C, 10 °C/min to 240 °C
by gas chromatography.
1
Biotransformations. Biotransformations were typically performed with 10 mM
substrate, 80 mM Tris–HCl buffer pH 6.0 and 200 μl of cell lysate (45 mg/ml
and finally 40 °C/min to 310 °C, which was held for 2 min. Chiral GC analy- total protein) in an amber 2 ml screw cap GC–MS vial (Phenomenex Verex
sis was performed as follows: Hydrodex β-6TBDM capillary column (Agilent;
5 m × 250 μm × 0.25 μm) and hydrogen as carrier gas (linear velocity 30 cm/s).
Certified 9 mm 2 ml Amber 51 vials). For the reaction, 500 μl of substrate
DMSO stock solution (20 mM stock solution in DMSO 10% v/v) was mixed
2
Injections were performed in split mode (split ratio 10:1). The temperature pro- with 200 μl of LinD cell lysate and diluted with 300 μl buffer (80 mM
gram started at 80 °C, then was raised 1 °C/min to 95 °C, 0.5 °C/min to 100 °C, Tris–HCl pH 6.0). Substrates were converted for 24 h at 35 °C. Reactions in
1
5 °C/min to 130 °C and finally 40 °C/min to 230 °C.
buffer were used as negative controls. After the conversion, the reaction mix-
ture was extracted with cyclohexane (2 × 800 μl), the combined organic phase
Chemical synthesis of substrates. Substrates (E)-4,8-dimethylnona-2,7- was dried with sodium sulfate and analyzed by gas chromatography.
dien-4-ol (4), 3-ethyloct-1-en-3-ol (6), (E)-3-methylocta-1,4-dien-3-ol
(
5
2
8), 3-methylhex-1-en-3-ol (12), 3-methylhept-1-en-3-ol (14), 3-methyl- Purification of native LinD. In order to purify LinD, competent E. coli BL21
-phenylpent-1-en-3-ol (16), 3-methoxy-3,7-dimethylocta-1,6-diene (20), (DE3) cells were transformed with the plasmid containing the LinD gene,
,6-dimethylhept-5-en-2-ol (23), 3,7-dimethyloct-6-en-3-ol (24), (E)-3- and colonies from plates grown overnight were used to inoculate 5 ml starter
methyloct-4-en-3-ol (25), (E)-3,7-dimethylocta-1,4,6-trien-3-ol (26), (E)-3,7- cultures of LB broth containing kanamycin (30 μg/ml). These starter cultures
dimethylocta-4,6-dien-3-ol (27), 7-methylocta-1,6-dien-3-ol (28), 3,7-dimeth- were incubated at 37 °C overnight with shaking and were then used to inocu-
ylocta-1,6-dien-3-amine (linalyl amine 29), and 3,7-dimethyloct-6-en-1-yn-3
late 500 ml each of LB medium containing kanamycin (30 μg/ml). The 500 ml
cultures were incubated at 37 °C with shaking at 180 r.p.m. and induced with
-
ol (30) were synthesized according to literature (Supplementary Note)37–39
.
1
mM IPTG at an OD600 of 0.6. After induction, cultures were incubated at 16 °C
Cloning and site-directed mutagenesis. The codon-optimized gene encod- overnight. Cells were then harvested by centrifugation at 5,000 r.p.m. for
ing for LinD from Castellaniella defragrans was cloned into a pET-28a(+)-
vector system (Merck, Darmstadt, Germany). NcoI and HindIII were used as
restriction sites. The plasmid was transformed into Escherichia coli DH5α, and
15 min using a Sorvall RC5B centrifuge and resuspended in 30 ml of 50 mM
Tris buffer (pH 8.0) containing 500 mM NaCl (henceforth referred to as
‘buffer’) with 20 mM imidazole. Ultrasonication was used to disrupt cells at an
successful cloning was verified by DNA sequencing of the entire gene. The amplitude of 14,000 microns for 8 × 30 s, after which the soluble lysate was sep-
mutations were introduced by site-directed mutagenesis using the standard arated from the crude cell extract by centrifugation at 14,000 r.p.m. for 25 min.
Stratagene QuikChange protocol. The mutated plasmids were transformed The lysate was then loaded onto a 5 ml HisTrap FF crude column charged with
into E. coli DH5α using heat shock (30 s at 42 °C). Successful mutagenesis was
verified by DNA sequencing of the coding regions of the constructs.
0.1 M nickel sulfate. The LinD protein was then eluted using a 50–500 mM
imidazole gradient over 15 column volumes. The pooled peak fractions were
concentrated, loaded onto a HiLoad 16/60 Superdex 75 Prep Grade gel filtra-
Expression of LinD. For protein expression each vector construct was trans- tion column and eluted using buffer. After the final purification step, the con-
formed in E. coli BL21 (DE3) and glycerol stocks were prepared. The individual centrated enzyme was stored at 4 °C for crystallization experiments.
glycerol stock was used to inoculate a 5 ml LB medium pre-culture (30 μg/ml
kanamycin), which was incubated overnight at 37 °C and 180 r.p.m. These cul- Statistics. Data are represented as mean ± s.d. Experiments were performed as
tures were used to inoculate 800 ml of LB medium (10 g/L tryptone, 5 g/L yeast triplicates (n = 3).
extract, 10 g/L NaCl, pH 7.0) in 2 L Erlenmeyer flasks. 0.5 mM IPTG was used
for induction at an optical density at 600 nm (OD600) of 0.4–0.6 followed by Expression and purification of the selenomethionine derivative of LinD.
incubation for 4 h at 37 °C (180 r.p.m.). Cells were harvested (16,900 × g, 15 min, Expression of SeMet-LinD was performed using the same plasmid as was used
4
°C), frozen in liquid nitrogen and stored at −80 °C.
for the native protein and using the ‘inhibition method’ in E. coli BL21 (DE3)
as the host strain. A single colony of cells containing the recombinant plasmid
Expression and screening of LinD variants. Gibson assembly with alternative was used to inoculate 300 ml of sterile LB medium, and the culture incubates
gene fragments was also used to create plasmid variations that each include res- at 37 °C overnight in an orbital shaker at 180 r.p.m. The cells were harvested by
idue changes Tyr45Phe, Met125Ala, His129Ala, Cys171Ala and Cys180Ala in centrifugation at 5,000 r.p.m. for 15 min and the cell pellet washed twice with
the LinD gene. LinD variant plasmids were transformed into competent E. coli
deionized water (2 × 100 ml) to minimize the methionine content in the sub-
BL21 (DE3). Colonies from overnight growth at 35 °C were picked into 5 ml sequent expression culture. The cells were divided into twelve equal aliquots,
nature CHeMICaL BIOLOGY
doi:10.1038/nchembio.2271

each of which was used to inoculate 500 ml of growth medium in 2 L
Erlenmeyer flasks. 500 ml of growth medium contained, per liter: 100 ml of
M9 salts (containing 13.7 g Na HPO ; 6 g KH PO ; 1 g NaCl; 2g NH Cl per
data were collected at the absorption edge of Se. The self-rotation function
for the SeMet data suggested the presence of five-fold non-crystallographic
symmetry. SeMet-LinD contains 9–10 SeMet residues per molecule. The Se
2
4
2
4
4
liter); 12.5 ml 20% w/v glucose; 1 ml 1M MgSO ; 50 μl 1M CaCl ; 1 ml of a vita- heavy atom sites were readily identified using SHELXC/D43 as implemented in
4
2
min solution (containing 1 mg/ml each of riboflavin, niacinamide, pyridoxine
monohydrochloride, thiamine and biotin) and 5 ml 50 mM FeSO The 500 ml
cultures were grown at 37 °C to an OD600 of 0.6. At this point, 50 mg each of
powdered L-lysine, L-phenylalanine, L-threonine and L-selenomethionine, and
CCP4, and 44 Se were found. Density modification including five-fold aver-
aging in PARROT44 resulted in an electron density map of sufficient quality
4
.
45
for automated model building in BUCCANEER with intermittent refinement
46
in REFMAC . The resulting model had crystallographic R/Rfree of 0.26/0.30
and consisted of five essentially complete molecules. The solutions of native,
SeMet and geraniol complex data sets each contained five molecules in the
2
5 mg L-isoleucine, L-leucine and L-valine were added, along with IPTG to a
final concentration of 1 mM. The cultures were then incubated at 16 °C for 18 h
at 215 r.p.m. in an orbital shaker. From this point, cells were harvested and the asymmetric unit, each of which constituted a toroidal pentamer. The struc-
tures were built and refined using iterative cycles using Coot47 and REFMAC ,
46
SeMet-LinD purified in the same way as for the native enzyme. The presence
of selenium in the sample was confirmed by electrospray MS analysis. For the the latter employing local NCS restraints. The final structures of native,
native sample, the measured mass was 44,155.04 Da (expected 44,157.06 Da). SeMet and geraniol complex LinD exhibited Rcryst and Rfree values of 16.9/19.5,
The difference of 2 Da was attributed to the possible presence of a single
disulfide bond. For the SeMet sample the mass measured was 44,634.72 Da
16.8/20.3 and 16.1/19.8, respectively. All structures were finally validated
using PROCHECK . Refinement statistics for all structures are presented in
48
(
expected 44,636.50 Da), which corresponded to the mass of native LinD with Supplementary Table 2. The Ramachandran plot for the native LinD showed
the incorporation of 10 Se per monomer, as expected from the sequence of
the protein.
98.0% of residues to be situated in the most favored regions, 1.0% in additional
allowed and 1.0% residues in outlier regions. For the SeMet-LinD, the corre-
sponding values were 97.0%, 2.0% and 1.0%. For the geraniol complex, the
Crystallization. Purified native, or selenomethionine-containing LinD, were values were 97.2%, 2.1% and 0.7%.
concentrated to 10 mg/ml using an Amicon Ultra-15 with a 10 kDa molec-
ular weight cut-off membrane in a centrifuge at 4,000 r.p.m. Screening was
Data availability. Structural data have been deposited in the Protein Data
carried out using a Mosquito LCP (TTP Labtech) robot and commercial 96 Bank (PDB) with coordinate accession numbers 5G1W (native LinD), 5G1V
well-plate screens in which 0.15 μl of protein solution was mixed with 0.15 μl (SeMet-LinD) and 5G1U (SeMet-LinD in complex with geraniol). All other
of the precipitant solution. Crystals of native LinD were obtained in 0.1 M
Bis–Tris propane buffer pH 6.5 with 20% w/v PEG 3350 and 0.2 M sodium
data generated or analyzed during this study are included in this published
article (and its supplementary information files) or are available from the cor-
malonate. Crystals were optimized using the hanging drop method in 24 well- responding author on reasonable request. The alignment of tertiary alcohol
plate Linbro dishes, and using drops of 1 μl protein and 1 μl precipitant solu- substrates was produced using PyMol (PyMol Molecular Graphics System ver-
tion. The best crystals were obtained in conditions of 0.1 M Bis–Tris propane
sion 1.3 Schrödinger, LLC) and the rotamers were created with Ex:OMEGA
buffer pH 6.5 with 25% w/v PEG 3350, 10% w/v methylpentane diol and 0.1 M version 2.4.6 (OpenEye Scientific Software, Santa Fe, NM).
sodium malonate. Crystals were flash cooled in liquid nitrogen in a solution of
the mother liquor plus 20% w/v ethylene glycol as cryoprotectant, before dif-
fraction testing. For the SeMet-LinD, the best crystals were obtained in 1.1 M
sodium tartrate at pH 7.0. On scale-up, best crystals were obtained again in
3
7. Overman, L.E. A general method for the synthesis of amines by the
rearrangement of allylic trichloroacetimidates. 1,3-Transposition of alcohol
and amine functions. J. Am. Chem. Soc. 98, 2901–2910 (1976).
8. Willimann, L. & Schinz, H. Über 7-Methyl-octadien-(2,6)-ol-(1), ein
‘A pogeraniol ’. Helv. Chim. Acta 35, 2401–2406 (1952).
1
.1 M sodium tartrate at pH 7.0 with 1% w/v glycerol. In this case, crystals
3
3
were flash cooled in a solution of the mother liquor plus 20% w/v glycerol as
cryoprotectant before diffraction testing. For SeMet-LinD in complex with
geraniol, crystals were grown as for the SeMet derivative. Prior to flash cooling
in liquid nitrogen, the crystals were soaked in a solution of the mother liquor
containing 20% w/v glycerol as cryoprotectant plus 2% v/v racemic linalool for
9. Schwartz, B.D. et al. Towards the total synthesis of vibsanin E, 15-O-
methylcyclovibsanin B, 3-hydroxyvibsanin E, furanovibsanin A, and
3
-O-methylfuranovibsanin A. Eur. J. Org. Chem. 2006, 3181–3192 (2006).
4
0. ꢁabsch, W. XDS. Acta Crystallogr. D Biol. Crystallogr. 66, 125–132 (2010).
41. Evans, P.ꢀ. & Murshudov, G.N. How good are my data and what is the
3
h, before diffraction testing. All crystals were tested for diffraction in-house
resolution? Acta Crystallogr. D Biol. Crystallogr. 69, 1204–1214 (2013).
4
4
2. Winter, G. xia2: An expert system for macromolecular crystallography data
reduction. J. Appl. Cryst. 43, 186–190 (2010).
using a Rigaku Micromax-007HF fitted with Osmic multilayer optics and a
Marresearch MAR345 imaging plate detector. Those crystals that diffracted to
a resolution of equal to, or better than, 3-Å resolution were retained for data set
collection at the Diamond Light Source synchrotron.
3. Sheldricꢂ, G.M. Experimental phasing with SHELXC/D/E: combining chain
tracing with density modification. Acta Crystallogr. D Biol. Crystallogr. 66,
479–485 (2010).
4
4
4
4. Cowtan, ꢁ. ꢀecent developments in classical density modification. Acta
Crystallogr. D Biol. Crystallogr. 66, 470–478 (2010).
Data collection and refinement. Complete data sets described in this paper
were collected at the Diamond Light Source, Didcot, Oxfordshire, UK. The
temperature of data collections was 120 K in each case. The native LinD and
SeMet-LinD data sets were collected on beamline I04, with a wavelength of
5. Cowtan, ꢁ. Fitting molecular fragments into electron density. Acta
Crystallogr. D Biol. Crystallogr. 64, 83–89 (2008).
6. Murshudov, G.N., Vagin, A.A. & Dodson, E.J. ꢀefinement of macromolecular
structures by the maximum-liꢂelihood method. Acta Crystallogr. D Biol.
Crystallogr. 53, 240–255 (1997).
0
0
.97949 Å and the SeMet-LinD data set on beamline I02, with a wavelength of
40
47. Emsley, P. & Cowtan, ꢁ. Coot: model-building tools for molecular graphics.
.97902 Å. Data were processed and integrated using XDS and scaled using
Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132 (2004).
41
42
AIMLESS as part of the Xia2 processing system . Data collection statistics
are given in Supplementary Table 2. The crystals of native LinD and SeMet-
LinD were in space groups I2 and P2 2 2 , respectively. For the SeMet-LinD
4
8. Lasꢂowsꢂi, ꢀ.A., MacArthur, M.W., Moss, D.S. & ornton, J.M.
PꢀOCHECꢁ: a program to checꢂ the stereochemical quality of protein
structures. J. Appl. Crystallogr. 26, 283–291 (1993).
1
1 1 1
doi:10.1038/nchembio.2271
nature CHeMICaL BIOLOGY