Characterization of the first naturally thermostable terpene synthases and development of strategies to improve thermostability in this family of enzymes

Article abstract of DOI:10.1111/febs.14072

The terpenoid family of natural products is being targeted for heterologous microbial production as a cheaper and more reliable alternative to extraction from plants. The key enzyme responsible for diversification of terpene structure is the class-I terpene synthase (TS), and these often require engineering to improve properties such as thermostability, robustness and catalytic activity before they are suitable for industrial use. Improving thermostability typically relies on screening a large number of mutants, as there are no naturally thermostable TSs described upon which to base rational design decisions. We have characterized the first examples of natural TSs exhibiting thermostability, which catalyse the formation of the sesquiterpene τ-muurolol at temperatures up to 78 °C. We also report an enzyme with a k_cat value of 0.95 s^?1 at 65 °C, the highest k_cat recorded for a bacterial sesquiterpene synthase. In turn, these thermostable enzymes were used as a model to inform the rational engineering of another TS, with the same specificity but low sequence identity to the model. The newly engineered variant displayed increased thermostability and turnover. Given the high structural homology of the class-I TS domain, this approach could be generally applicable to improving the properties of other enzymes in this class. Database: Model data are available in the PMDB database under the accession number PM0080780.

Full text of DOI:10.1111/febs.14072

Characterization of the First Naturally Thermostable Terpene Synthases and Development of Strategies to
Improve Thermostability in This Family of Enzymes3
Article type : Original Articles
Matthew Q. Styles1, Edward A. Nesbitt1, Scott Marr1, Marc Hutchby2, David J. Leak1
1Department of Biology and Biochemistry, University of Bath, Bath, UK, BA2 7AY
2Department of Chemistry, University of Bath, Bath, UK, BA2 7AY
Running title: Thermostable terpene synthases
Corresponding author: Prof. David J. Leak, Department of Biology and Biochemistry, University of Bath,
Bath, UK, BA2 7AY, Telephone: +44 (0) 1225 384309; E-mail: d.j.leak@bath.ac.uk
Abbreviations
CAPS, N-cyclohexyl-3-aminopropanesulfonic acid; DCM, dichloromethane; DSC, differential scanning
calorimetry; FID, flame ionization detector; FPP, farnesyl pyrophosphate; IPTG, Isopropyl β-D-1-
thiogalactopyranoside; LB, Luria-Bertani medium; MES, 2-(N-morpholino)ethanesulfonic acid; MWCO,
molecular weight cut-off; OD600, optical density at a wavelength of 600 nm; PIPES piperazine-N,N′-
bis(2-ethanesulfonic acid); RMSD, root-mean-square deviation; TBA-FPP, tetrabutylammonium farnesyl
pyrophosphate; Tm, melt temperature; Tris-HCl, tris(hydroxymethyl)aminomethane hydrochloride salt.
Database: Model data are available in the PMDB database under the accession number PM0080780
Keywords: Kinetics, thermostability, terpenoid, structural model, thermophile
Conflict of interest: The authors declare they have no conflicts of interest with the contents of this
article.
Abstract
The terpenoid family of natural products are being targeted for heterologous microbial production
as a cheaper and more reliable alternative to extraction from plants. The key enzyme responsible for
diversification of terpene structure is the class-I terpene synthase, and these often require engineering to
improve properties such as thermostability, robustness and catalytic activity before they are suitable for
industrial use. Improving thermostability typically relies on screening a large number of mutants, as there
are no naturally thermostable terpene synthases described upon which to base rational design decisions.
We have characterized the first examples of natural terpene synthases exhibiting thermostability, which
catalyse the formation of the sesquiterpene τ-muurolol at temperatures up to 78 ˚C. We also report an
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process, which may lead to
differences between this version and the Version of Record. Please cite this article as doi:
10.1111/febs.14072
This article is protected by copyright. All rights reserved.

enzyme with a kcat value of 0.95 s-1 at 65 °C, the highest kcat recorded for a bacterial sesquiterpene
synthase. In turn these thermostable enzymes were used as a model to inform the rational engineering of
another terpene synthase, with the same specificity but low sequence identity to the model. The newly
engineered variant displayed increased thermostability and turnover. Given the high structural homology
of the class-I terpene synthase domain, this approach could be generally applicable to improving the
properties of other enzymes in this class.
Introduction
Enzyme stability is often the limiting factor in the development of commercial enzymes [1, 2]. In
particular, thermal stability is essential in a wide range of industrial applications, including the household
care, food, textiles and bioenergy industries, where enzymes are required to be robust under a range of
temperature conditions. Additionally, thermostable enzymes are desirable as starting points for directed
evolution of new functions, as subsequent mutants can better tolerate changes that would otherwise have
led to destabilization [3]. Class-I terpene synthases typically have low activity and stability, and so are
attractive targets for the engineering of increased thermostability and thermoactivity [4]. However, no
naturally thermostable examples have been described, and so there is no model for thermostability in this
class of enzymes.
Class-I terpene synthases produce terpenoids, the single largest class of natural products, with
over 70000 members in the natural products database, including examples in pharmaceuticals, fragrances,
flavours, polymers and biofuels [5-9]. The extraordinary diversity of terpenoids is partly driven by class-I
terpene synthases, which convert C10, C15 and C20 isoprenyl pyrophosphates into monoterpenes,
sesquiterpenes and diterpenes respectively. For example, the C15 sesquiterpene synthases use farnesyl
pyrophosphate (FPP) as a substrate to generate over 300 carbon skeletons [10, 11]. In turn these
sesquiterpenes can be processed further either enzymatically or chemically to generate the huge range of
sesquiterpenoid natural products [12, 13]. Traditionally, valuable terpenes have always been extracted
directly from the plants that natively produced them, but in the last decade, there have been substantial
efforts put into obtaining high value sesquiterpenes via microbial fermentation. Such microbial platforms
require terpene synthases that are expressed and active at fermentation temperatures. Currently most
known sesquiterpene synthases are found in plants [14, 15], with increasing numbers more recently
discovered in bacteria [16]; to this date none of these have been characterised as thermostable, and many
are optimally expressed at temperatures below 30 °C [13, 17].
In order to make terpene synthases useful industrially, protein engineering is required to improve
enzyme properties, including thermostability, without impacting product specificity. There have been no
examples of naturally thermostable class-I terpene synthases described but recently there have been
several efforts to create thermostable variants using generically applicable techniques [18, 19]. Diaz et al.
were able to generate a thermostable variant of tobacco epi-aristolochene synthase using computational
methods, with activity observed up to 65 ˚C. However, the thermostable mutant was expressed in
insoluble form and had lost much of its activity and specificity compared to the wild-type [18]. Another
approach developed a presilphiperfolan-8β-ol synthase towards thermostability using a directed evolution
approach, leading to an enzyme that retained full activity up to 50 ˚C (with a T50 value of 54 ˚C) [19].
Both approaches for generating thermostable terpene synthases could be informed by model examples
from nature that retain activity up to high temperatures.
This article is protected by copyright. All rights reserved.

The class-I terpene synthase active site is contained within an α-helical domain, containing two
conserved Mg2+ binding motifs coordinating three Mg2+ ions. In turn the Mg2+ ions coordinate the
diphosphate head group of the substrate, whilst the isoprenyl tail positions within the binding pocket [20].
A cascade of intramolecular reactions, often including numerous cyclizations and rearrangements, can
result in remarkable specificity for a single product [21]. However, most terpene synthases have
promiscuous activity and will produce a range of products, and preference for a particular product can be
tipped by relatively few active-site mutations [21, 22]. Whilst those terpene synthases for which a
structure is known share a high structural homology, outside the conserved motifs there is generally low
sequence identity, especially in bacterial examples [16, 23]. There is a possibility therefore that some of
the structural features that affect thermostability and those that affect product specificity can be
considered separately, and that by using a thermostable terpene synthase as a model, increased
thermostability can be engineered into a variant from a mesophile.
In this work two enzymes were characterized as the first examples of thermostable sesquiterpene
synthases. The enzymes RoseRS_3509 and Rcas_0622 had previously been functionally identified in vivo
as τ-muurolol synthases by Yamada et. al. [16]. We demonstrate that one of the enzymes, RoseRS_3509,
has a higher turnover rate than any other bacterial terpene synthase, and that these two enzymes can be
used as the first model for thermostability in this class of enzymes.
Results
Identification of thermostable terpene synthases
Potential thermostable terpene synthases were identified using the HMMER web server, which uses
profile hidden Markov models to return potential sequences of interest [24]. In this case, the sequence for
the bacterial τ-muurolol synthase, SSCG_03688 (Accession Number: EDY50541.1), was used as an
initial entry, and the output genes were searched for thermophilic native organisms. Gene sequences were
codon optimised and expressed in E.coli, then screened for terpene synthase activity. Two terpene
synthases were identified, RoseRS_3509 from Roseiflexus sp.RS-1 (AN: WP_011958209.1) and
Rcas_0622 from Roseiflexus castenholzii DSM 13941. The two Roseiflexus protein sequences, recently
identified as τ-muurolol synthases using a similar search strategy [16], shared just 32% identity with
SSCG_03688 with 94% and 89% coverage respectively. For context, a blast of the Rosieflexus protein
sequences search reveals over 1000 putative bacterial terpene synthase proteins with similar identity
(27%-35% identity, >90% coverage). Production of τ-muurolol as the major product was confirmed by
GC-MS.
Preliminary screening using tris(tetrabutylammonium) farnesyl pyrophosphate
In previous reports describing in vitro terpene synthase activity the ammonium salt of FPP has
been used [17, 21, 25], which is complicated to prepare and expensive to buy. We discovered that the
tetrabutylammonium salt of FPP (TBA-FPP), which could be prepared readily and economically from
farnesol, was also an effective substrate for each of the sesquiterpene synthases we tested. For preliminary
screening of terpene synthases it was feasible to use TBA-FPP assays in place of the ammonium salt,
allowing the determination of suitable assay parameters including assay time, enzyme concentration and
This article is protected by copyright. All rights reserved.

temperature range for each enzyme. However, in order to provide consistency with previous literature, all
experimental data reported in this manuscript used the commercial ammonium salt of FPP as substrate.
Characterization of terpene synthase thermoactivity
Both enzymes were found to be sesquiterpene synthases, converting farnesyl pyrophosphate
(FPP) into τ-muurolol as the major product in vitro. The thermoactivity of each enzyme was assessed over
1 min and the optimum temperature for activity was determined to be 61 ˚C for RoseRS_3509 and 69 ˚C
for Rcas_0622 (Fig 1A). For RoseRS_3509, activity was still detected at 78 ˚C, which to our knowledge
represents the highest temperature for which activity in sesquiterpene synthases has ever been reported.
Moreover, given the notoriously loose product profile of such enzymes, it is significant that the product
profile remained unchanged at higher temperatures, as determined by both GC-MS and the flame
ionization detector (FID) peak area.
Characterization of terpene synthase thermostability
The melting profile for the two enzymes were determined by monitoring the characteristic α-helix
circular dichroism (CD) signal at 222 nm, from 20 ˚C to 94 ˚C, as a surrogate for enzyme thermostability.
The CD signal at 222 nm was fitted using a simple two-state transition equation, adapted from Catici et
al. [26], with R2 > 0.99 in both cases. Fitting the data gave a melt temperature (Tm) of 64.8 0.1 ˚C for
RoseRS_3509 and 59.6 0.1 ˚C for Rcas_0622 (Fig 1B), where the error represent the deviation from the
fit.
Kinetics of thermostable terpene synthases
Both thermostable terpene synthases were found to follow single-substrate Michaelis-Menten
kinetics at 65 ˚C. Kinetic parameters were measured at 65 ˚C by carrying out 1 min incubations over a
range of FPP concentrations from 1.2 μM to 92 μM giving the kinetic values shown in Table 1. The
product was quantified using GC-FID, and the kcat value of 0.95 s-1 for RoseRS_03509 is the highest kcat
value reported for a bacterial sesquiterpene synthase at any temperature.
Sequence alignment and modelling
The amino acid sequences of the two thermostable terpene synthases investigated, RoseRS_3509
and Rcas_0622, were aligned with the mesophile τ-muurolol synthase, SSCG_03688. This showed that
overall there was good similarity between the three throughout most of the sequence, with two notable
exceptions. Firstly the thermostable variants are significantly truncated at the C-terminus relative to
SSCG_03688, where the C-terminus extends about 100 residues further. As a result SSCG_03688 is 418
amino acids in length, whilst RoseRS_3509 has just 326 amino acids. Secondly the region directly
following the first magnesium binding site motif (amino acids 89-118 in SSCG_03688) has low similarity
with the two thermostable sequences. In order to observe the relationship of the low homology regions to
the overall enzyme structure, both sequences were modelled using the I-TASSER server [27, 28] and the
models overlaid using PyMOL (Fig 2). The model for RoseRS_3509 has an estimated RMSD = 3.7
This article is protected by copyright. All rights reserved.

2.5Å, whilst the model for SSCG_03688 has RMSD = 5.0 3.2Å. It is clear that the first region of low
similarity occurs in a loop immediately following one of the magnesium binding sites, and is unstructured
in SSCG_03688, but appears to have more α-helical structure in RoseRS_3509. The C-terminus region in
SSCG_03688 that is not present in RoseRS_3509 appears to wrap 180˚ around the barrel of the protein,
and has some β-sheet structure (Fig 2).
Truncation of SSCG_03688
In the directed evolution study by Lauchli et al it was found that the most thermostable terpene
synthase generated had an early stop codon, resulting in a 3 amino acid truncation of the C-terminus [19].
Given that the thermostable examples also had shortened C-terminal regions than SSCG_03688, the effect
of C-terminal truncation on thermostability was investigated. A set of mutants of SSCG_03688 was
created in which the C-terminus was sequentially shortened by removing blocks of 7-15 amino acids
(R2-R9), with the final mutant having a C-terminus of a comparable length to RoseRS_3509 (Fig 3A).
The mutants were expressed, purified, and melting curves were determined by circular dichroism. To
improve the accuracy of the comparison, melting temperatures were determined by fitting the data using
Eq. 1 in the same way as described for the thermostable terpene synthases, where each curve was fit with
R2 > 0.99. The CD data shows that removing blocks of amino acids from the C-terminus does indeed lead
to a sequential increase in melting temperature, with the Tm increasing by up to 1.4˚C with the R6 mutant
(Fig 3B). However, further truncation of the C-terminus beyond R6 (i.e. R7-R9) affected protein folding
in E. coli, resulting in heterologous expression of insoluble protein aggregates.
Differential scanning calorimetry (DSC)
The Tm was also determined for the terpene synthases investigated using DSC in native activity
buffers. In particular, the Tm was determined in the presence and absence of Mg2+ and pyrophosphate,
which have been shown to induce the closed and open state respectively of other terpene synthases [29].
Changing the buffer to include both Mg2+ and pyrophosphate resulted in a small but significant increase
in Tm for RoseRS_3509 and Rcas_0622, by 2.0 °C and 4.9 °C respectively, with no change in peak shape
(Fig 4). DSC carried out on SSCG_03688 and the R6 mutant demonstrated the same trend of Tm increase
for the mutant as determined by CD, with Tm increasing from 44.9 °C to 45.8 °C. The introduction of
Mg2+ and pyrophosphate again appeared to result in a Tm shift to higher temperatures for both
SSCG_03688 variants, but also led to a substantially altered peak shape and more convoluted baseline,
making firm conclusions impossible for the R6 mutant (Fig 4).
Kinetics of SSCG_03688 R6 truncated mutant
Kinetic parameters were determined at 30 °C for both the wild-type enzyme and the most
truncated mutant expressed (R6, Table 2). The kcat value reported here for the wild-type is higher than
previously reported [17], but consistent with the findings that GC-FID vial assays give relatively higher
values than radioactivity assays for terpene synthases [30].
This article is protected by copyright. All rights reserved.

Thermoactivity of SSCG_03688 R6 truncated mutant
The thermoactivity of the R6 truncated mutant was compared with that of the wild-type between
27 °C and 47 °C using the malachite green assay [30]. The new mutant had consistently higher activity at
all temperatures. These results followed the same trend when the experiments were repeated using the
malachite green assay protocol described by Vardakou et al.[30] (Fig 5), which appeared to give reduced
variability between replicates compared to the GC-MS protocol. The malachite green assay protocol also
gave consistently higher figures for activity compared to the GC-MS protocol. However, GC-MS data
were required to determine the sesquiterpene product profile, which remained unchanged at each
temperature for both the wild-type and the mutants (Fig 6).
Discussion
Roseiflexus species are natural thermophiles that have been isolated from phototrophic mats in
hot springs, growing at 60 °C [31]. Genome sequences are available for Roseiflexus sp. RS-1 and
Roseiflexus castenholzii DSM 13941, and both contain genes encoding terpene synthases. Coincidentally,
both enzymes have previously been characterized as τ-muurolol synthases as part of a wider screen
demonstrating that bacterial terpene synthases are much more widespread than previously appreciated,
but no investigation was made into their thermostability [16]. In this work, RoseRS_3509 from
Roseiflexus sp. RS-1 and Rcas_0622 from R. castenholzii DSM 13941 have been characterized as
naturally thermostable terpene synthases and demonstrated to have activity in vitro up to 78 ˚C, which
represents the highest temperature for which activity in this class of enzyme have been observed
(including a terpene synthase computationally designed for improved thermostability) [18].
Bacterial terpene synthases have a simpler structure than their plant counterparts, lacking any N-
terminal domains. However, they retain the typical α-helical fold of the catalytic components of terpene
synthases, despite having low primary sequence identity. This simpler structure is attractive from the
prospect of protein engineering but the sluggish kinetics of mesophilic bacterial terpene synthases (kcat
typically between 10-2 – 10-3 s-1) restricts their application [32]. The larger plant terpene synthases
typically have kcat values between 10-1 – 10-2 s-1 [32]. Due to their low turnover, terpene synthases are rate
limiting in industrial terpene producing organisms and they are typically artificially overexpressed
relative to the precursor pathway [33]. We report kcat values in a bacterial τ-muurolol synthase at 65 °C
which are similar, if not superior, to most plant enzymes. The mesophile τ-muurolol synthase
SSCG_03688, which converts the same substrate into the same product as our thermostable examples,
presumably by the same reaction cascade, has a reported kcat value two orders of magnitude lower [17].
This is interesting as, whilst increased temperature typically accelerates enzyme catalysis up to the
stability limits of the enzyme, it does not follow that thermostable enzymes have higher catalytic rates
than their less thermostable counterparts. Indeed, it has been shown that thermostability is typically
achieved from greater structural rigidity, such that the appropriate mobility for good catalysis is only
achieved at higher temperatures [34]. The dramatic enhancement in kcat going from an enzyme which
works optimally at 40 °C to one that works optimally at >60 °C, but with broadly similar tertiary structure
suggests that the factors which provide thermostability do not significantly affect catalysis (i.e. the active
site is not compromised by increased structural rigidity), so that providing a more thermostable
framework does allow the benefits of temperature on accelerating catalysis to be realized in this class of
enzymes. Given the evidence that the α-helical fold structure of the catalytic component of class-I terpene
This article is protected by copyright. All rights reserved.

synthases essentially provides a cage within which a carbocation can be generated and rearranged, it is
possible to envisage that increased rigidity of the cage (providing thermostability) is independent of the
mobility of catalytically functional residues. If this is true, then understanding how the α-helical fold
structure is stabilized could provide strategies for thermostabilization of all single domain class-I terpene
synthases. Indeed, the DSC experiments demonstrated that the putative closed conformation of these
terpene synthases, induced by the presence of Mg2+ and pyrophosphate, resulted in an increase in Tm in all
cases.
By modelling both the thermostable terpene synthases and SSCG_03688 on a closely related
structure, several differences could be observed, the most obvious being the extended C-terminal tail of
the mesophile enzyme. Previous work has shown that even a small truncation of a terpene synthase C-
terminus can result in an increase in thermostability [19]. Sequentially truncating the C-terminus of
SSCG_03688 resulted in a moderate but consistent increase in thermostability, as determined by CD (Fig
3B). The trend in Tm was also observed by DSC for SSCG_03688 and the R6 mutant, and the Tm values
obtained for the open enzyme conformations (in the absence of Mg2+ and pyrophosphate) were
comparable between the two techniques. Moreover, by truncating the C-terminus of SSCG_03688, the
difference between the Tm of the closed and open states was reduced, implying that the truncation leads to
a reduction in flexibility around the active-site lid without impacting negatively on the kinetics [29].
Although we were not able to remove the C-terminal tail entirely without affecting protein folding, we
were able to truncate the tail by 57 amino acids and still obtain a functional and soluble enzyme, which
additionally had a significantly higher kcat value (Table 2). Referring to the model overlay, it can be
observed that further truncation would remove a short helical region in the tail, which appears to be
essential for soluble expression of the protein (Fig 2). In experiments with the most truncated mutant, it
was observed that there was an overall trend towards higher thermoactivity (Fig 5), demonstrating that the
C-terminal tail in the wild-type has a detrimental effect on thermostability.
Whilst removing 57 amino acids only resulted in a moderate change in thermostability, it did not
negatively affect either the kinetics or product profile of the enzyme. This is the second example (along
with the directed evolution work from Lauchli et. al.[19]) of C-terminal truncation coinciding with
improved thermostability of a terpene synthase, suggesting that a more minimal (truncated) enzyme is a
reasonable starting place for thermostability improvement. It is interesting that such a substantial
truncation does not result in a loss of catalytic efficiency, and implies that the unstructured C-terminal tail
has some other function, such as influencing protein localization through some kind of protein-protein
interaction.
It is interesting that thermophilic bacteria have enzymes for making volatile terpenes. Both
thermostable terpene synthase gene sequences are found in operons containing genes coding for a putative
type 11 methyl transferase and an anti-sigma-factor antagonist, and are flanked by a genes for a
transmembrane protein and a beta-lactamase. Roseiflexus species are typically found in the diverse
communities of microbial mats in hot springs [31], so it is possible that a diffusible terpene might be used
as a signalling molecule.
Given that these naturally thermostable terpene synthases share a comparable level of sequence
identity to over a thousand bacterial putative sesquiterpene synthases as they do to SSCG_03688, it is
clear they can be used as a model for improvement of class-I terpene synthase thermostability, and that
further structural studies might provide fundamental insights into terpene synthase engineering.
This article is protected by copyright. All rights reserved.

Experimental procedures
Chemicals
The pure (E,E)-FPP standard was purchased from Isoprenoids (Tampa, Florida, USA) and used
for all terpene synthase assays reported. Gene sequences were optimized for codon usage in E. coli and
purchased from GeneArt (ThermoFisher Scientific, Paisley, UK). DNA modifying enzymes were
purchased from Life Technologies (Paisley, UK). Bradford Protein Assay dye reagent was obtained from
Bio-Rad (Hertfordshire, UK). All other chemicals were purchased from Sigma-Aldrich except where
indicated otherwise.
Strains, plasmids and growth conditions
All cloning and plasmid propagation was carried out in E. coli BIOBlue chemically competent
cells (Bioline, London, UK). All protein expression was carried using E. coli BL21(DE3). The three
terpene synthase genes, roseRS_3509, rcas_0622 and sscg_03688 were inserted between the NheI and
EcoRI restriction sites of pET28a, for the expression of N-terminal His6-tagged proteins. E. coli cultures
were grown at 37˚C and 250 rpm in Luria-Bertani (LB) medium supplemented with 50 μg mL-1 of
kanamycin for selection of the plasmid except when specified otherwise.
Protein production and purification
For the production of RoseRS3509 and Rcas_0622, 1 mL of overnight E.coli BL21(DE3) culture
transformed with the appropriate plasmid was used to inoculate 1 L of LB. Expression of genes controlled
by the T7-promoter was induced by adding 0.1 mM IPTG to cultures once they had reached an optical
density (OD600) of 0.5. Cultures were harvested 20 hrs post-induction and the cells pelleted by
centrifugation (4000 rpm, 4˚C, 10 min). The supernatant was discarded, and cell pellets resuspended in
Buffer A (20 mM imidazole, 20 mM Tris, 300 mM NaCl, pH 8), before being lysed by sonication (3 x 15
s pulses, samples kept on ice). Cell debris was separated by centrifugation (14000 rpm, 10 min), and the
supernatant taken as the soluble protein fraction. The soluble fraction was loaded onto a column
containing pre-equilibrated TALON® metal affinity resin (Clontech Laboratories, Saint-Germain-en-
Laye, France) as per the manufacturer’s instructions. The column was washed with 20 column volumes of
Buffer A, and eluted with Buffer B (500 mM imidazole, 20 mM Tris, 300 mM NaCl, pH 8). Protein size
and purity was checked using SDS-PAGE (both proteins expressed well under these conditions and
resulted in a single major band at ~40 kDa with >95% purity), and buffer was exchanged into terpene
synthase buffer 1 (25 mM MES, 50 mM Tris, 25 mM CAPS, 100 mM NaCl, 10 mM MgCl2) using
Amicon® Ultra 15 mL spin concentrators (10 kDa MWCO, Watford, UK). Protein was then used
immediately for assays, or 10% glycerol was added and samples aliquoted into tubes for storage at -80˚C.
Production of SSCG_03688 was adapted from the method used by Hu et al. [17]. Briefly, the conditions
were the same as above with the following exceptions; immediately post-induction cultures were grown at
18˚C for the 20 hrs protein expression. For the lysis and purification, Buffer C (50 mM Tris-HCl, 0.5 M
NaCl, 20 mM imidazole, 5 mM 2-mercaptoethanol, pH 7.5) was used in place of Buffer A, and Buffer D
(50 mM Tris-HCl, 0.5 M NaCl, 500 mM imidazole, 5 mM 2-mercaptoethanol, pH 7.5) was used in place
of Buffer B for elution. Protein size and purity was checked using SDS-PAGE (the wild-type and all
mutant bands had >95% purity; the wild-type band was at ~48kDa and truncated mutant proteins were
This article is protected by copyright. All rights reserved.

visibly and sequentially smaller on the gel). Eluted SSCG_03688 protein was exchanged into terpene
synthase buffer 2 (TS2; 50 mM PIPES, 10 mM MgCl2, 100 mM NaCl, 5 mM 2-mercaptoethanol, pH 7.5).
Finally aliquots were stored at -80˚C in 20% glycerol. All protein stocks were quantified using the
Bradford Protein Assay.
Terpene synthase activity assay
Assays were adapted from the GC vial assay described by Garrett et al.[25]. Stock enzyme
solution was diluted to 50 μM concentration by addition of TS buffer. Final 200 μL reaction volumes
were prepared in 500 μL tubes by addition of 4 μL of 50 μM enzyme solution to 196 μL TS buffer, and
the reaction started by adding the appropriate amount of FPP stock (1 mg mL-1, 7:3 methanol:10 mM
NH4OH). For the GC vial assays, reactions were stopped by snap freezing in liquid nitrogen, and 150 μL
of hexane spiked with 10 μg mL-1 caryophyllene internal standard was added to the frozen sample. Tubes
were then vortexed until the reaction solution was completely melted (~1 min), and then centrifuged
(4000 rpm, 2 min) and the organic layer transferred to a 200 μL vial insert for analysis by GC-MS. For
assays carried out at higher temperatures, tubes containing buffer were incubated in a thermal cycler
block at the desired temperature for at least 10 minutes prior to the start of the reaction. All terpene
synthase assays were carried out in triplicate, and errors reported as standard error of the mean.
Thermoactivity of RoseRS3509 and Rcas_0622
Assays were carried out with an enzyme concentration of 1 μM over 1 min at a range of
temperatures 50 °C – 84 °C, with a substrate concentration of 58 µM. Reaction buffer was pre-warmed at
each temperature in a thermocycler for at least 10 min, the enzyme added, and after a 15 s equilibration
time the reaction was started by the addition of FPP. The samples were then processed as before using the
GC vial assay protocol. The data were fit to a simple Gaussian function to aid visualization (Fig 1A). The
errors are reported as standard error of the mean.
Thermoactivity of SSCG_03688 variants using the malachite green assay
Thermoactivity of SSCG_03688 wild-type and mutant were both determined using the recently developed
malachite green protocol for terpene synthase activity in order to confirm the trends observed using the
GC-MS protocol [30]. Assays were carried out with an enzyme concentration of 1 μM over 8 min at a
range of temperatures 30 °C – 47 °C, with a substrate concentration of 58µM. Reaction buffer was pre-
warmed at each temperature in a thermocycler for at least 10 min, the enzyme added, and after a 15 s
equilibration time the reaction was started by the addition of FPP. The data were fit to a simple Gaussian
function to aid visualization (Fig 5). Assays were carried out in triplicate, and the errors are reported as
standard error of the mean.
GC-MS and GC-FID Analysis
Analysis of terpene products was carried out on an Agilent 7890B GC coupled to a 5977A mass
spectrometer (Agilent Technologies, Stockport, UK), and separation was achieved using a DB-FFAP
capillary GC column (30 m x 250 μm x 0.25 μm). 1 μL of sample was injected onto the column, which
was held at 40˚C for 1 min. The temperature was then ramped 20˚C min-1 to a final temperature of 250˚C,
which was held for a further 8 min. Terpene compounds typically eluted between 5 min and 13 min, and
the total method time was 19.5 min, and were monitored on both the MS and FID detector. The product
peak area from the FID chromatogram was divided by the internal standard peak area, then multiplied by
the caryophyllene standard concentration to generate a ‘caryophyllene equivalents’ value for product
This article is protected by copyright. All rights reserved.

concentration in the organic extract. Due to the unit carbon response of the FID detector, and in the
absence of a commercially available τ-muurolol standard, caryophyllene is suitable for the relative
quantification of sesquiterpene products [35].
Steady state kinetics
Assays were carried out with an enzyme concentration of 1 μM and substrate concentrations of 1.2 μM,
2.3 μM, 4.6 μM, 12 μM, 23 μM, 46 μM and 92 μM using the GC-FID vial assay protocol described
above. For the thermostable terpene synthases, reactions were carried out in triplicate at 65 ˚C and were
stopped after 1 minute. For SSCG_03688 and the truncated mutant variant, assays were carried out in
triplicate at 30 ˚C and were stopped after 5 minutes. The data were fit to the Michaelis-Menten equation
and kinetic values derived using SigmaPlot Version 12.3 (Systat Software, San Jose, CA).
Circular dichroism (CD) protein melts
Fresh enzyme was purified and the elution buffer exchanged with a buffer compatible with the CD
instrument (100 mM NaF, 10 mM phosphate buffer, pH 8.0). Buffer exchange was carried out by dialysis
(10 kDa MWCO) overnight, and 250 μL of sample (enzyme concentration 5 μM) was pipetted into a 1
mm path length quartz cuvette. CD was monitored on a Chirascan CD Spectrometer (Applied
Photophysics Ltd, Surrey, UK) at 222 nm for 5 s at 1˚C intervals between 10˚C and 95˚C.
Tm determination – The data was fit using Equation 1 for a simple two-state transition,
ꢄꢅ + ꢆꢅꢇ + (ꢄꢈ + ꢆꢈꢇ)ꢉꢅ
ꢀꢁꢁꢁ ꢂꢃ
=
(ꢊꢋ. 1)
1 + ꢉꢅ
where
ꢇ
ꢇꢃ
∆ꢍ ꢎ1 −
ꢐꢇ
ꢏ
ꢉꢅ = exp ꢌ
ꢑ
(ꢊꢋ. 2)
where b and m are the slope and intercept of the unfolded (u) and folded (f) baseline respectively; Tm is
the melting temperature of the protein; and ΔH is the van ‘t Hoff enthalpy of unfolding at Tm. Data fitting
was carried out using SciDAVis data analysis software.
Differential scanning calorimetry
Analysis of protein by DSC was conducted on a µSC (Setaram Instrumentation, Caluire, France). It was
determined that >800 µg of protein was required in order to observe a consistent signal for protein
melting experiments. Sufficient stock protein was diluted to 500 µL (final volume) using either terpene
synthase buffer 2 (25 mM MES, 50 mM Tris, 25 mM CAPS, 100 mM NaCl) or terpene synthase buffer 3
(terpene synthase buffer 2 with 10 mM MgCl2 and 1 mM Na4P2O7). The buffers were pH 8.0 for
RoseRS_3509 and Rcas_0622 and pH 7.5 for the SSCG_03688 variants. The 500 µL protein samples
were sealed inside a measurement cell, and the heatflow determined against a reference cell, which
contained an equivalent volume of the appropriate buffer. The instrument was programmed to hold the
samples at 20 °C for 5 minutes, before it increased the temperature by 1 °C min-1 up to 95 °C (75 °C for
SSCG_03688 experiments). Data were baseline corrected using the Calisto Processing software (v 1.41,
Setaram Instrumentation).
This article is protected by copyright. All rights reserved.

Sequence alignment and modelling
The protein sequences for the two thermostable τ-muurolol synthases (RoseRS_3509 Accession
Number: WP_011958209.1, Rcas_0622 Accession Number: WP_012119179.1) were aligned with that of
SSCG_03688 (Accession Number EDY50541.1) using the T-Coffee web server alignment tool [36, 37].
The sequences of RoseRS_3509 and SSCG_03688 were both initially modelled against the template
crystal structure of selinadiene synthase (PDB entry 4OKM.4.A) using the SWISS-MODEL web server
[38]. The sequences were then modelled using iterative threading assembly refinement (I-TASSER
server) [27, 28]. The resultant .pdb files from the I-TASSER modelling were visualized (Fig 2) using
PyMOL (The PyMOL Molecular Graphics System, Version 1.8 Schrödinger, LLC.).
Preparation of tris(tetra-n-butylammonium) hydrogen phosphate
A solution of disodium dihydrogen pyrophosphate (3.33 g, 15 mmol) in 15 mL of 10% (v/v)
aqueous ammonium hydroxide was passed through a column (3 x 9 cm) of Dowex 50WX8 (100-200
mesh) cation exchange resin and subsequently flushed with deionized water (110 mL). The resulting
solution (pH = 1) was immediately titrated to a pH of 7.3 with 40% (w/w) aqueous tetra-n-
butylammonium hydroxide followed by lyophilization of the entire reaction mixture to leave a
hygroscopic white solid (12.7 g, 96%). All analytical data matches that of the original report [39]. This
crude reaction material was used in the following steps with no further purification.
Preparation of farnesyl chloride
In a flame dried, two necked 250 mL round-bottomed flask under a blanket of nitrogen was
charged N-chlorosuccinimide (292 mg, 2.2 mmol) dissolved in anhydrous dichloromethane (DCM, 100
mL). The reaction was cooled to -30 °C followed by the dropwise addition of dimethyl sulfide (149 mg,
2.4 mmol) under vigorous stirring of the heterogeneous mixture. The reaction was warmed to 0 °C for 5
mins before cooling back down to -30 °C. Farnesol (444 mg, 2 mmol) was dissolved in anhydrous DCM
(10 mL) and was added dropwise to the milky white suspension. The reaction was slowly warmed to 0 °C
over 1 h and stirred at this temperature for a further 1 h. After warming to room temperature and stirring
for an additional 15 min, the now clear solution was extracted with a cold saturated brine solution (50
mL) with further extraction of the aqueous layer (DCM, 2 x 50 mL). The combined organic extracts were
further washed with a saturated brine solution (50 mL) and dried over magnesium sulfate before the
removal of the solvent in vacuo. The crude colourless oil (450 mg, 94%) matched the analytical data
originally reported [39] and was used without further purification.
Preparation of FPP tetra-n-butylammonium salt
In a flame dried, two-necked 25 mL round-bottomed flask, under a blanket of nitrogen was added a
solution of the crude tris(tetra-n-butylammonium) hydrogen pyrophosphate (1.8 g, 2 eq.) dissolved in
anhydrous acetonitrile (2 mL). To this was added a solution of the crude farnesyl chloride (240 mg, 1 eq.)
in anhydrous acetonitrile (2 mL) and the reaction was stirred at room temperature for 2 h. The solvent was
then removed in vacuo to yield the crude material as an orange oil (2.2 g).
This article is protected by copyright. All rights reserved.

The oil was dissolved in methanol to a stock concentration of 0.5 g mL-1, and stored at -20 °C. The stock
was diluted in methanol:10 mM aqueous NH4OH (7:3) in preparation for assays. FPP content was
determined by comparison of activity compared to a pure standard, and was typically 15-20 % without
any further work-up steps.
Acknowledgements: We thank Dr. Christopher Pudney for his advice on fitting the CD melt data. This
work was supported in whole or part by the Engineering and Physical Sciences Research Council grant
(EP/K014889/1).
Author contributions: MQS designed and conducted most of the experiments, analyzed the results and
wrote the paper. EAN conducted the experiments on the SSCG_03688 enzyme and its mutants. SM
optimized the expression of Rcas_0622. MH synthesized the FPP tetra-n-butylammonium salt. DJL
conceived the idea for the project, and helped write the paper.
References
1. Asial, I., Cheng, Y. X., Engman, H., Dollhopf, M., Wu, B., Nordlund, P. & Cornvik, T. (2013) Engineering
protein thermostability using a generic activity-independent biophysical screen inside the cell, Nature
Communications. 4. DOI: 10.1038/ncomms3901
2. Eijsink, V. G. H., Gaseidnes, S., Borchert, T. V. & van den Burg, B. (2005) Directed evolution of enzyme
stability, Biomolecular Engineering. 22, 21-30.
3. Bommarius, A. S. & Paye, M. F. (2013) Stabilizing biocatalysts, Chemical Society Reviews. 42, 6534-
6565.
4. Siddiqui, K. S. (2015) Some like it hot, some like it cold: Temperature dependent biotechnological
applications and improvements in extremophilic enzymes, Biotechnology Advances. 33, 1912-1922.
5. Paduch, R., Kandefer-Szerszen, M., Trytek, M. & Fiedurek, J. (2007) Terpenes: substances useful in
human healthcare, Archivum Immunologiae Et Therapiae Experimentalis. 55, 315-327.
6. Bohlmann, J. & Keeling, C. I. (2008) Terpenoid biomaterials, Plant Journal. 54, 656-669.
7. Wilbon, P. A., Chu, F. X. & Tang, C. B. (2013) Progress in Renewable Polymers from Natural Terpenes,
Terpenoids, and Rosin, Macromolecular Rapid Communications. 34, 8-37.
8. Sarria, S., Wong, B., Martin, H. G., Keasling, J. D. & Peralta-Yahya, P. (2014) Microbial Synthesis of
Pinene, Acs Synthetic Biology. 3, 466-475.
9. George, K. W., Alonso-Gutierrez, J., Keasling, J. D. & Lee, T. S. (2015) Isoprenoid Drugs, Biofuels, and
Chemicals-Artemisinin, Farnesene, and Beyond, Biotechnology of Isoprenoids. 148, 355-389.
10. Rynkiewicz, M. J., Cane, D. E. & Christianson, D. W. (2001) Structure of trichodiene synthase from
Fusarium sporotrichioides provides mechanistic inferences on the terpene cyclization cascade,
Proceedings of the National Academy of Sciences of the United States of America. 98, 13543-13548.
11. Cane, D. E. (1990) Enzymatic formation of sesquiterpenes, Chemical Reviews. 90, 1089-1103.
12. Paddon, C. J., Westfall, P. J., Pitera, D. J., Benjamin, K., Fisher, K., McPhee, D., Leavell, M. D., Tai, A.,
Main, A., Eng, D., Polichuk, D. R., Teoh, K. H., Reed, D. W., Treynor, T., Lenihan, J., Fleck, M., Bajad, S.,
Dang, G., Dengrove, D., Diola, D., Dorin, G., Ellens, K. W., Fickes, S., Galazzo, J., Gaucher, S. P.,
Geistlinger, T., Henry, R., Hepp, M., Horning, T., Iqbal, T., Jiang, H., Kizer, L., Lieu, B., Melis, D., Moss, N.,
This article is protected by copyright. All rights reserved.

Regentin, R., Secrest, S., Tsuruta, H., Vazquez, R., Westblade, L. F., Xu, L., Yu, M., Zhang, Y., Zhao, L.,
Lievense, J., Covello, P. S., Keasling, J. D., Reiling, K. K., Renninger, N. S. & Newman, J. D. (2013) High-
level semi-synthetic production of the potent antimalarial artemisinin, Nature. 496, 528-532.
13. Beekwilder, J., van Houwelingen, A., Cankar, K., van Dijk, A. D. J., de Jong, R. M., Stoopen, G.,
Bouwmeester, H., Achkar, J., Sonke, T. & Bosch, D. (2014) Valencene synthase from the heartwood of
Nootka cypress (Callitropsis nootkatensis) for biotechnological production of valencene, Plant
Biotechnology Journal. 12, 174-182.
14. Chen, F., Tholl, D., Bohlmann, J. & Pichersky, E. (2011) The family of terpene synthases in plants: a
mid-size family of genes for specialized metabolism that is highly diversified throughout the kingdom,
Plant Journal. 66, 212-229.
15. Degenhardt, J., Koellner, T. G. & Gershenzon, J. (2009) Monoterpene and sesquiterpene synthases
and the origin of terpene skeletal diversity in plants, Phytochemistry. 70, 1621-1637.
16. Yamada, Y., Kuzuyama, T., Komatsu, M., Shin-ya, K., Omura, S., Cane, D. E. & Ikeda, H. (2015)
Terpene synthases are widely distributed in bacteria, Proceedings of the National Academy of Sciences
of the United States of America. 112, 857-862.
17. Hu, Y., Chou, W. K. W., Hopson, R. & Cane, D. E. (2011) Genome Mining in Streptomyces
clavuligerus: Expression and Biochemical Characterization of Two New Cryptic Sesquiterpene Synthases,
Chemistry & Biology. 18, 32-37.
18. Diaz, J. E., Lin, C.-S., Kunishiro, K., Feld, B. K., Avrantinis, S. K., Bronson, J., Greaves, J., Saven, J. G. &
Weiss, G. A. (2011) Computational design and selections for an engineered, thermostable terpene
synthase, Protein Science. 20, 1597-1606.
19. Lauchli, R., Rabe, K. S., Kalbarczyk, K. Z., Tata, A., Heel, T., Kitto, R. Z. & Arnold, F. H. (2013) High-
Throughput Screening for Terpene-Synthase-Cyclization Activity and Directed Evolution of a Terpene
Synthase, Angewandte Chemie-International Edition. 52, 5571-5574.
20. Baer, P., Rabe, P., Fischer, K., Citron, C. A., Klapschinski, T. A., Groll, M. & Dickschat, J. S. (2014)
Induced-Fit Mechanism in Class I Terpene Cyclases, Angewandte Chemie-International Edition. 53, 7652-
7656.
21. Salmon, M., Laurendon, C., Vardakou, M., Cheema, J., Defernez, M., Green, S., Faraldos, J. A. &
O'Maille, P. E. (2015) Emergence of terpene cyclization in Artemisia annua, Nature Communications. 6.
DOI: 10.1038/ncomms7143
22. Yoshikuni, Y., Ferrin, T. E. & Keasling, J. D. (2006) Designed divergent evolution of enzyme function,
Nature. 440, 1078-1082.
23. Greenhagen, B. & Chappell, J. (2001) Molecular scaffolds for chemical wizardry: Learning nature's
rules for terpene cyclases, Proceedings of the National Academy of Sciences of the United States of
America. 98, 13479-13481.
24. Finn, R. D., Clements, J., Arndt, W., Miller, B. L., Wheeler, T. J., Schreiber, F., Bateman, A. & Eddy, S.
R. (2015) HMMER web server: 2015 update, Nucleic Acids Research. 43, W30-W38.
25. Garrett, S. R., Morris, R. J. & O'Maille, P. E. (2012) Steady-State Kinetic Characterization of
Sesquiterpene Synthases by Gas Chromatography-Mass Spectroscopy, Natural Product Biosynthesis by
Microorganisms and Plant, Pt A. 515, 3-19.
26. Catici, D. A. M., Horne, J. E., Cooper, G. E. & Pudney, C. R. (2015) Polyubiquitin Drives the Molecular
Interactions of the NF-kappa B Essential Modulator (NEMO) by Allosteric Regulation, Journal of
Biological Chemistry. 290, 14130-14139.
27. Zhang, Y. (2008) I-TASSER server for protein 3D structure prediction, Bmc Bioinformatics. 9. DOI:
10.1186/1471-2105-9-40
28. Roy, A., Kucukural, A. & Zhang, Y. (2010) I-TASSER: a unified platform for automated protein
structure and function prediction, Nature Protocols. 5, 725-738.
This article is protected by copyright. All rights reserved.

29. van der Kamp, M. W., Sirirak, J., Zurek, J., Allemann, R. K. & Mulholland, A. J. (2013) Conformational
Change and Ligand Binding in the Aristolochene Synthase Catalytic Cycle, Biochemistry. 52, 8094-8105.
30. Vardakou, M., Salmon, M., Faraldos, J. A. & O'Maille, P. E. (2014) Comparative analysis and
validation of the malachite green assay for the high throughput biochemical characterization of terpene
synthases, MethodsX. 1, 187-96.
31. van der Meer, M. T. J., Klatt, C. G., Wood, J., Bryant, D. A., Bateson, M. M., Lammerts, L., Schouten,
S., Damste, J. S. S., Madigan, M. T. & Ward, D. M. (2010) Cultivation and Genomic, Nutritional, and Lipid
Biomarker Characterization of Roseiflexus Strains Closely Related to Predominant In Situ Populations
Inhabiting Yellowstone Hot Spring Microbial Mats, Journal of Bacteriology. 192, 3033-3042.
32. Schomburg, I., Chang, A. & Schomburg, D. (2002) BRENDA, enzyme data and metabolic information,
Nucleic Acids Research. 30, 47-49.
33. Pitera, D. J., Paddon, C. J., Newman, J. D. & Keasling, J. D. (2007) Balancing a heterologous
mevalonate pathway for improved isoprenoid production in Escherichia coli, Metabolic Engineering. 9,
193-207.
34. Danson, M. J., Hough, D. W., Russell, R. J. M., Taylor, G. L. & Pearl, L. (1996) Enzyme thermostability
and thermoactivity, Protein Engineering. 9, 629-630.
35. Rodriguez, S., Kirby, J., Denby, C. M. & Keasling, J. D. (2014) Production and quantification of
sesquiterpenes in Saccharomyces cerevisiae, including extraction, detection and quantification of
terpene products and key related metabolites, Nature Protocols. 9, 1980-1996.
36. Notredame, C., Higgins, D. G. & Heringa, J. (2000) T-coffee: a novel method for fast and accurate
multiple sequence alignment1, Journal of molecular biology. 302, 205-217.
37. Di Tommaso, P., Moretti, S., Xenarios, I., Orobitg, M., Montanyola, A., Chang, J.-M., Taly, J.-F. &
Notredame, C. (2011) T-Coffee: a web server for the multiple sequence alignment of protein and RNA
sequences using structural information and homology extension, Nucleic Acids Research. 39, W13-W17.
38. Guex, N., Peitsch, M. C. & Schwede, T. (2009) Automated comparative protein structure modeling
with SWISS-MODEL and Swiss-PdbViewer: A historical perspective, Electrophoresis. 30, S162-S173.
39. Davisson, V. J., Woodside, A. B., Neal, T. R., Stremler, K. E., Muehlbacher, M. & Poulter, C. D. (1986)
Phosphorylation of isoprenoid alcohols, Journal of Organic Chemistry. 51, 4768-4779.
Table 1
Kinetic parameters of thermostable terpene synthases at 65 °C
Enzyme
RoseRS_3509
Rcas_0622
kcat (s-1)
0.95 ± 0.29
0.090 ± 0.008
KM (µM)
87 ± 31
5.9 ± 2.0
Table 2
Kinetic parameters of SSCG_03688 and truncated mutant.
Enzyme
SSCG_03688 wild-type
SSCG_03688 R6
kcat (s-1)
0.018 ± 0.001
0.083 ± 0.006
KM (µM)
4.5 ± 1.0
16.3 ± 4.0
This article is protected by copyright. All rights reserved.

Figure Legends
Figure 1. (A) Thermoactivity profile of the two thermostable sesquiterpene synthases, RoseRS_3509 and
Rcas_0622. The error bars represent the standard error of mean for three replicates. Inset; the structure of
the sesquiterpene product, τ-muurolol. (B) CD absorption at 222 nm for the two thermostable
sesquiterpene synthases.The lines show the curves fit in order to determine the Tm of the proteins.
This article is protected by copyright. All rights reserved.

Figure 2. Overlaid models for RoseRS_3509 (orange, estimated RMSD = 3.7 2.5Å) and SSCG_03688
(cyan, estimated RMSD = 5.0 3.2Å) demonstrate the predicted structural homology between the two
enzymes, despite the low sequence identity. The mesophile enzyme has an extended C-terminus
compared to the thermostable variants; the SSCG_03688 R6 mutant tail is shown in blue, and the full
length wild-type tail is shown in magenta. Two of the Mg2+ binding residues are at the entry to the active
site and are shown in red.
This article is protected by copyright. All rights reserved.

Figure 3. (A) The sequence of the SSCG_03688 protein, with the termini of the truncated mutants created
in this study marked. (B) The Tm of the wild-type SSCG_03688 and truncated mutants, as derived from
their respective CD melt profiles at 222 nm. The error bars represent the deviation of the Tm variable from
the fit described in Eq. 1.
Figure 4. DSC traces for each protein. Red traces were recorded in buffer containing Mg2+ and
pyrophosphate. Blue traces were recorded in the absence of Mg2+ and pyrophosphate. The Tm for each
trace is shown the legend.
This article is protected by copyright. All rights reserved.

Figure 5. Thermoactivity profile of the two SSCG_03688 sesquiterpene synthases, SSCG_03688 wild-
type and the most truncated mutant SSCG_03688 R6. The error bars represent the standard error of mean
for three replicates.
This article is protected by copyright. All rights reserved.

Figure 6. Normalized MS chromatograms showing the GC-MS product profile of the SSCG_03688 R6
mutant at different temperatures. The major peak is τ-muurolol.
This article is protected by copyright. All rights reserved.