Development and validation of an HPLC-based screening method to acquire

polyhydroxyalkanoate synthase mutants with altered substrate specificity

Yoriko Watanabe,
1
Yousuke Ichinomiya,
1
Daisuke Shimada,
1
Azusa Saika,
1
Hideki Abe,
2
Seiichi Taguchi,
3
and Takeharu Tsuge
1,

Department of Innovative and Engineered Materials, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama 2268502, Japan,
1
Bioplastic
Research Team, RIKEN Biomass Engineering Program, 21 Hirosawa, Wako-shi, Saitama 3510198, Japan,
2
and Division of Biotechnology and
Macromolecular Chemistry, Graduate School of Engineering, Hokkaido University, N13W8, Kita-ku, Sapporo 0608628, Japan
3
Received 19 June 2011; accepted 16 October 2011
Available online 15 November 2011
A rapid and convenient method for the compositional analysis of polyhydroxyalkanoate (PHA) was developed using high-
performance liquid chromatography (HPLC) and alkaline sample pretreatment in a 96-well plate format. The reliability of this
system was confirmed by the fact that a mutant with a D171G mutation of Aeromonas caviae PHA synthase (PhaC
Ac
), which
gained higher reactivity toward 3-hydroxyhexanoate (3HHx), was selected from the D171X mutant library. Together with
D171G mutant, several single mutants showing high reactivity toward 3HHx were isolated by the HPLC assay. These new
mutants and double mutants combined with an N149S mutation were used to synthesize P(3-hydroxybutyrate-co-3HHx) in
Ralstonia eutropha PHB

4 from soybean oil as carbon source, achieving higher levels of 3HHx fraction than the wild-type
enzyme. Based on these results, the high-throughput screening system will serve as a powerful tool for exploring new and
beneficial mutations responsible for regulating copolymer composition of PHA.
2011, The Society for Biotechnology, Japan. All rights reserved.
[Key words: Polyhydroxyalkanoates; High-throughput screening; High-performance liquid chromatography (HPLC); Alkaline pretreatment; PHA
synthase; Substrate specificity]
Polyhydroxyalkanoates (PHAs) are biological polyesters pro-
duced by a wide variety of microorganisms as an intracellular
storage material for carbon and energy. PHAs have attracted
industrial attention for use as biodegradable and biocompatible
thermoplastics (1,2). PHAs are synthesized by PHA synthases
(PhaCs) which catalyze the polymerization reaction of 3-hydro-
xyalkanoates (3HAs) as monomer substrates. Therefore, the sub-
strate specificity of PhaC significantly influences the monomer
composition of synthesized PHA.
Poly(3-hydroxybutyrate) [P(3HB)] is the most common PHA
synthesized by bacteria in nature. P(3HB) has high rigidity but is
brittle with low elasticity. Therefore, flexible 3HB-based copolymers
such as P(3HB-co-3-hydroxyvalerate) [P(3HB-co-3HV)], P(3HB-co-3-
hydroxyhexanoate) [P(3HB-co-3HHx)], P(3HB-co-3-hydroxy-4-
methylvalerate) [P(3HB-co-3H4MV)], and P(3HB-co-medium-chain-
length-3-hydroxyalkanoate) [P(3HB-co-mcl-3HA)] are recognized as
more suitable polymers for practical use (35). These monomer
structures are shown in Fig. 1.
Aeromonas caviae is capable of synthesizing P(3HB-co-3HHx)
random copolymer from vegetable oils as the carbon source (6),
because this bacterium possesses the PHA synthase (PhaC
Ac
) that has
the unique ability to polymerize 3HB and 3HHx units. P(3HB-co-
3HHx) is highly desired by industry as a bio-based plastic, but this
bacterium has poor ability to produce and accumulate it (less than
about 30 wt.% of the cells). On the other hand, Ralstonia eutropha is a
PHA over-producer (greater than 80 wt.% of the cells), while the type
of PHA synthesized by this strain is limited to P(3HB) with vegetable
oils as the carbon source. Thus, the higher production of P(3HB-co-
3HHx) was achieved from vegetable oils in recombinant R. eutropha
PHB

4 transformed with a vector-borne PhaC

Ac
gene (7,8). The
resultant host-vector system, however, suffered from weak incorpo-
ration of the 3HHx unit into P(3HB-co-3HHx), limiting the 3HHx
fraction to 34 mol% in cultivation on soybean oil. This phenomenon
was mainly ascribed to the substrate specificity of PhaC
Ac
(7). We
applied directed evolution to create mutant enzymes that gain higher
reactivity toward 3HHx. In the initial stage, two beneficial single
mutants, N149S (asparagine 149serine) and D171G (aspartic acid
171glycine), were obtained (9). Subsequent mutation studies with
PhaC
Ac
allowed us to regulate the 3HHx fraction from 0 to 5.2 mol%
(10,11). The highest 3HHx fraction (5.2 mol%) was obtained using the
PhaC
Ac
double mutant with N149S and D171G mutations (NSDG
mutant) (10).
To extend the regulation range of the 3HHx fraction, PhaC
Ac
mutants with further increased reactivity toward 3HHx were
required. For this, a much more efficient high-throughput screening
system was necessary. Kichise et al. developed an initial assay
Journal of Bioscience and Bioengineering
VOL. 113 No. 3, 286292, 2012
www.elsevier.com/locate/jbiosc

Corresponding author. Tel.: +81 45 924 5420; fax: +81 45 924 5426.
E-mail address: tsuge.t.aa@m.titech.ac.jp (T. Tsuge).
1389-1723/$ - see front matter 2011, The Society for Biotechnology, Japan. All rights reserved.
doi:10.1016/j.jbiosc.2011.10.015
method to measure cellular P(3HB) content using high-performance
liquid chromatography (HPLC) (12). To prepare samples for this
method, P(3HB)-accumulating cells are treated with sulfuric acid at
100C to convert P(3HB) to crotonic acid (trans-2-butenoic acid).
Subsequently, the treated samples are subjected to HPLC with an
ultraviolet (UV) detector to measure absorption at 210 nm due to
unsaturated crotonic acid bonds (13). This method is very
convenient to measure P(3HB) concentration; however, it is not
applicable to other PHAs because PHA monomers longer than 3HB
cannot be converted to the corresponding unsaturated fatty acids. In
a previous study (12), PhaC
Ac
mutants showing higher reactivity
toward 3HHx were obtained as a result of screening for high-
polymerization activity mutants based on P(3HB) accumulation
levels in host Escherichia coli. Thus, establishing a sample prepara-
tion method applicable to longer PHA monomers including 3HHx
would allow direct screening for substrate-specificity-altered
synthases based on PHA copolymer composition using HPLC,
improving screening efficiency. On the other hand, gas chromatog-
raphy (GC) analysis is widely used to determine PHA composition
regardless of 3HA carbon chain length. It requires derivatization
process of polymer samples, which is difficult to simplify for a high-
throughput assay, prior to GC analysis.
In this study, a new HPLC-based screening method capable of
analyzing PHA copolymer composition was developed by applying
alkaline (sodium hydroxide) pretreatment instead of acid pretreatment
to sample preparation for HPLC analysis. In addition, a high-throughput
protocol was established by introducing a 96-well plate format for
cultivation of the PHA producing host and for sample preparation. The
new method was used to isolate PhaC
Ac
mutants with high reactivity
toward 3HHx from a D171 random point mutation (D171X) library to
determine whether the D171G mutation is the most effective for
increasing PhaC
Ac
reactivity toward 3HHx.
MATERIALS AND METHODS
Bacterial strains and plasmids E. coli JM109 was used as the host strain for
screening PhaC
Ac
mutants and for P(3HB-co-3HHx) accumulation from dodecanoate,
while R. eutropha PHB

4 (PHA-negative mutant, DSM541) was used for PHA copolymer

production from soybean oil, octanoate, or 4-methylvalerate. Plasmid pBBR1phaPCJ-
Ac
AB
Re
was constructed by introducing a 6.4-kb XbaIHindIII fragment of pBSEE32ph-
bAB (12) into the same sites of a broad-host-range vector pBBR1MCS-2 (14). The
resultant plasmid carries PHA polycistronic genes (accession no. D88825) for PhaP
Ac
(granule-associated protein), PhaC
Ac
, and PhaJ
Ac
(R-specific enoyl-CoA hydratase) with
a promoter derived from A. caviae FA440 and the genes for the (R)-3HB-CoA monomer
supplying enzymes PhaA
Re
(3-ketothiolase) and PhaB
Re
(NADPH-dependent acetoa-
cetyl-CoA reductase) from R. eutropha H16 (accession no. J04987).
Random point mutagenesis at position 171 of PhaC
Ac
Random point
mutagenesis at position 171 of PhaC
Ac
(Fig. 2A) was performed using an inverse
polymerase chain reaction (PCR) method described by Imai et al. (15). The PCR primers
used in this study were designed in inverted tail-to-tail directions to amplify
pBBR1phaPCJ
Ac
AB
Re
with the target sequence for amino acid substitution as follows:
for D171X, 5-CCT GGA GTC CNN NGG CCA GAA CCT G-3 (the underlined sequence
indicates a mutation site, and N represents a random nucleotide) as the sense primer
and 5-GTC AGC TTG AGC AGC TCG GGG TTG G-3 as the antisense primer. After PCR
amplification with the primer set, the amplified linear DNA was phosphorylated and
self-ligated using a BKL kit (Takara Bio Inc., Otsu). Subsequently, the self-ligated PCR
products were transformed into E. coli JM109 to prepare a PhaC
Ac
mutant library.
HPLC-based screening of PhaC
Ac
mutants Fig. 2B shows a schematic diagram
of the HPLC-based screening system developed in this study. Briefly, it is consisted of
site-specific random mutagenesis, preparation of a mutant library, primary assay of
P(3HB) accumulation in E. coli JM109 using Nile red (9-(diethylamino)-5H-benzo[]
phenoxazin-5-one) dye on an agar plate, liquid cultivation in M9 medium plus
dodecanoate (C12) using a 96-well plate for P(3HB-co-3HHx) accumulation, alkaline
sample pretreatment, HPLC assay, and nucleotide sequence determination. A PhaC
Ac
D171X mutant library constructed with E. coli JM109 was spread on LuriaBertani (LB)
agar plates (Bacto-Tryptone 10 g, Bacto-yeast extract 5 g, NaCl 10 g, and agar 15 g per
liter of distilled water) supplemented with 20 gL
1
glucose, 0.5 mg/L
1
Nile red, and
50 mgL
1
kanamycin and cultured at 37C overnight. The polymerization ability of
PhaC
Ac
mutants was judged based on the intensity of the pinkish pigmentation of the
cells caused by Nile red staining (16). Next, single pinkish pigmented colonies were
inoculated in M9 medium (0.6 mL) containing 1.0 gL
1
sodium dodecanoate, 1.0 gL
1
Bacto-yeast extract, 0.4 vol% Brij35, and 50 mgL
1
kanamycin in each 1.2-mL well of a
96-deep well culture plate (BMEquipment Co., Ltd., Tokyo). After sealing the plate with
an air-permeable film(4titude, Ltd., Surrey, UK), the cells were cultured at 37C for 72 h
in a reciprocal shaker (1035 rpm, Bio Shaker, Taitec Co., Ltd., Saitama). After the grown
cells were replicated on LB agar plates, the cultured 96-well plates were centrifuged
using a Hitachi R6S swing rotor at 3000 rpm(1500 g) for 10 min, and then the culture
supernatants were discarded. The 96-well plates with cell pellets remaining at the
bottom of each well were dried at 55C for 3 days.
For alkaline hydrolysis of the dried cells, 200 L of 1 NNaOHwas added to each well of
the 96-well plates using a handheld multichannel pipettor. After heat-sealing with a
polypropylene/aluminumfilm(4titude) using a microplate heat sealer (ABgene Ltd., Surrey,
UK), the 96-well plates were heated at 100C for 3 h on a hot plate. The cell hydrolysates
wereneutralizedwith200 L of 1 NHCl andthenfiltratedwith96-well filter plates (0.45 m
pore size polytetrafluoroethylene (PTFE) membrane, Pall Co., NY, USA) by centrifuging at
3000 rpm(1500g) for 30 min. The filtrates werecollectedwitha new96-well assay plate,
sealed with a cover film to prevent evaporation, and subjected to HPLC analysis.
HPLC analysis was performed using a Shimadzu LC-10Avp system with an auto-
sample injector applicable to 96-well plates. The samples were separated on two types
of ion-exclusion columns, Aminex HPX-87H (300 mm7.8 mm I.D., Bio-Rad, CA, USA)
and Fast Acid Analysis (100 mm7.8 mm I.D., Bio-Rad), at 60C using 0.014 N H
2
SO
4
with or without 20% CH
3
CN as a mobile phase at a flow rate of 0.7 mL/min. The
chromatograms were recorded at 210 nm using a UV detector.
Site-specific mutation of PhaC
Ac
Aspartic acid (D) at position 171 of the PhaC
Ac
N149Smutant was replacedbyalanine(A), leucine(L), or histidine(H) as asecondmutation
to yield a doubly mutated gene (phaC
Ac
NSDA, NSDL, or NSDH) using a QuickChange Multi
Site-directed Mutagenesis Kit (Stratagene Co., CA, USA) or a similar method. The primer was
designed as 5-CTG GAG TCC NNNGGC CAG AAC CTG G-3. The underlined NNNsequences
in the primer were GCC, CTG, and CAC for alanine (A), leucine (L), and histidine (H)
replacement, respectively, and were designed based on the codon usage of R. eutropha.
Recombination of R. eutropha and PHA analysis To express the phaC
Ac
gene
in R. eutropha PHB

4 under our basal conditions, a 0.6-kb PstIScaI fragment of

pBBR1phaPCJ
Ac
AB
Re
was introduced into the same pBBREE32d13dPB sites (10,11) to
yield pBBREE32d13dPB D171X carrying only the mutated phaC
Ac
gene downstream of
the pha promoter from A. caviae. These plasmids were introduced by transconjugation
from E. coli S17-1 into R. eutropha PHB

4 (17).
The recombinant R. eutropha strain PHB

4 was cultivated at 30C for 72 h on a

reciprocal shaker (130 strokes/min) in500-mL flasks containing 100 mL of nitrogen-limited
mineral salt (MS) mediumsupplemented witha carbonsource (20 gL
1
soybeanoil, 5 gL
1
sodium octanoate, or 2.5 gL
1
4-methylvalerate). The composition of the MS medium was
as follows (per liter of distilled water): 9 g of Na
2
HPO
4
12H
2
O, 1.5 g of KH
2
PO
4
, 0.5 g of
NH
4
Cl, 0.2 gof MgSO
4
7H
2
O, and1 mL of trace element solution(18). ThepHof the medium
was adjusted to 7.0. Kanamycin (50 mgL
1
) was added to the medium to maintain the
expression plasmid. The PHA content in dry cells was determined by GC after methanolysis
of the lyophilized cells in the presence of 15% sulfuric acid (18). GC analysis was carried out
by using Shimadzu GC-14B system with a non-polar capillary column (InertCap 1,
30 m0.25 mm, GL Sciences Inc., Tokyo) and a flame ionization detector.
The polymers accumulated in the cells were extracted with chloroform for 72 h at
roomtemperature and purified via precipitation with methanol. Molecular weight data
were obtained by gel permeation chromatography (GPC) at 40C using a Shimadzu 10A
GPC system and a 10 A refractive index detector with Shodex K806M and K802
columns. Chloroform was used as the eluent at a flow rate of 0.8 mL/min, and sample
concentrations of 1.0 mg/mL were applied. Polystyrene standards with a low
polydispersity were used to make a calibration curve.
RESULTS
Pretreatment by acid or alkaline for HPLC analysis Karr et al.
established an HPLC technique for rapid analysis of P(3HB) with
O
O
*
*
3H4MV
O
O
*
*
3HV
O
O
*
*
O
O
*
*
3HHx 3HO
O
O
*
*
3HB
FIG. 1. Structure of monomer units in PHA synthesized in this study. 3HB, 3-
hydroxybutyrate; 3HV, 3-hydroxyvalerate; 3H4MV, 3-hydroxy-4-metylvalerate;
3HHx, 3-hydroxyhexanoate; 3HO, 3-hydroxyoctanoate.
HPLC ASSAY FOR PHA SYNTHASE MUTANTS 287 VOL. 113, 2012
sulfuric acid pretreatment, which is now widely used (19). Thus, we
treated the cells accumulating P(3HB-co-19 mol% 3HHx) with sulfuric
acid and performed HPLC analysis. The 3HB-derived peak (crotonic
acid) at 12.5 min was detected (Fig. 3A), while the 3HHx-derived peak
(hexenoic acid) did not appear. Acid pretreatment was not applicable
to the compositional analysis of P(3HB-co-3HHx).
According to Del Don et al., the composition of P(3HB-co-3HV) was
successfully analyzed by HPLC with alkaline (sodium hydroxide) pretreat-
ment (20). To verify alkaline pretreatment in this study, the cells were
treated with a sodium hydroxide solution and subjected to HPLC analysis.
Fig. 3Bshows a chromatogramof the cells accumulating P(3HB-co-19 mol%
3HHx). Unlike acid pretreatment, alkaline pretreatment allowed the
detection of both 3HB- and 3HHx-derived peaks. Thus, alkaline pretreat-
ment is useful in copolymer compositional analysis. Furthermore, to verify
the efficacy of alkaline pretreatment, cells accumulating a different type of
PHA copolymer (21), P(3HB-co-21 mol% 3HA) containing 7 mol% 3HHx,
8 mol% 3HO, 5 mol% 3HD, and 1 mol% 3HDD as 3HA units, were treated in
the same manner. The HPLC chromatogram is shown in Fig. 4. All
components of the copolymer other than 3HHD were detected by this
method. The 3HDD fraction was too low to be detected clearly. The
retention time of each 3HA-derived peak was consistent with that of the
corresponding 2-alkenoic acid used as a standard reference material.
Alkaline pretreatment is thus applicable to PHA consisting of a 3HB unit as
well as medium-chain-length 3HA units. However, detection sensitivity of
3HA units decreased with increasing 3HA alkyl side-chain length.
HPLC conditions for rapid analysis To apply HPLC analysis to a
high-throughput assay for PHA composition, we tried to shorten the
HPLC analysis time by modifying the HPLC conditions. The basal HPLC
condition, which used a Bio-Rad Aminex HPX-87H ion-exchange
column (column length 300 mm) and 0.014 N H
2
SO
4
as a mobile
phase, required 40 min for analysis (Figs. 3 and 4). Using a shorter
column (Bio-Rad Fast Acid Analysis, column length 100 mm) and a
a a 4 9 5 6 1 5 0 7 2 3 9 1
/ Hydrolase
fold region
C319 (Active site)
Non-
conserved
region
PhaCAc
N149
D171
A
B
Inoculate in M9 + C
12
medium
on 96-well plate
Mutation
phaCAc
Mutagenic
primer
HPLC assay
Alkaline treatment
on 96-well plate
PCR
E. coli
JM109
Mutant library
Plasmid extraction
from duplicated cells,
DNA sequencing
Mutagenesis
Selection
(min)
(min)
C4
C6
Cultivation
37 C, 72 h
Nile red
selection
Neutralization,
Filtration
FIG. 2. (A) The amino-acid-substituted position (D171) in PHA synthase of A. caviae (PhaC
Ac
). (B) Schematic flowdiagramof the HPLC-based assay. Site-specific randommutagenesis,
preparation of a mutant library, primary assay of P(3HB) accumulation in E. coli JM109 using Nile red dye on an agar plate, liquid cultivation in M9 medium plus dodecanoate (C12)
using a 96-well plate for P(3HB-co-3HHx) accumulation, alkaline sample pretreatment, HPLC assay, and nucleotide sequence determination were carried out.
0 10 20 30 40
(min)
R
e
l
a
t
i
v
e

i
n
t
e
n
s
i
t
y

(
A
2
1
0
)
R
e
l
a
t
i
v
e

i
n
t
e
n
s
i
t
y

(
A
2
1
0
)
0 10 20 30 40
(min)
A
B
C4
C4
C6
FIG. 3. HPLC analysis of P(3HB-co-19 mol% 3HHx)-accumulating cells [R. eutropha
PHB

4/phaC
Ac
(10)] treated with (A) 1 N H
2
SO
4
and (B) 1 N NaOH. C4, 3HB-derived
peak (crotonic acid); C6, 3HHx-derived peak (hexenoic acid). The column was a Bio-
Rad Aminex HPX-87H (column length 300 mm).
288 WATANABE ET AL. J. BIOSCI. BIOENG.,
20% CH
3
CN-containig mobile phase, the analysis time was shortened
to 10 min. The chromatogram of P(3HB-co-19 mol% 3HHx) analyzed
using the modified conditions is shown in Fig. 5. The 3HB- and 3HHx-
derived peaks were well separated, even within the shortened
analysis time. The area ratio of 3HB- to 3HHx-derived peak in the
HPLC analysis was 87:13, whereas the compositional molar ratio of
3HB to 3HHx in the polymer was 81:19. Therefore, the correction
coefficients for the 3HB and 3HHx units to calculate copolymer
composition from HPLC peak areas are determined to be 1.00 and
1.58, respectively.
Screening of PhaC
Ac
mutants with an HPLC assay Previously,
PhaC
Ac
mutants with a D171G mutation were isolated by evolutionary
engineering, which aimed to acquire a high-polymerization activity
enzyme (12). Because substitution by glycine (G) at position 171 is
known to increase reactivity toward 3HHx, the PhaC
Ac
D171X mutant
library was suitable for validation of the HPLC-based assay using the
D171G mutant as a beneficial control. Until now, the mutational effect
at position 171 had not been examined by replacement with amino
acids other than glycine (G). Thus, another objective of this study was
to verify whether the D171Gmutation is the most effective to increase
PhaC
Ac
reactivity toward 3HHx.
The D171X mutant library was constructed by introducing site-
specifically mutagenized pBBR1phaPCJ
Ac
AB
Re
into E. coli JM109. To
performa high-throughput assay, a 96-well plate format was introduced
for cell cultivation and sample pretreatment, as described in Materials
and methods. The 343 clones able to accumulate PHA, identified by Nile
red staining, were subjected to cultivation for accumulation of P(3HB-co-
3HHx) followed by alkaline pretreatment and an HPLC assay. Of these,
111 clones accumulated P(3HB-co-3HHx) copolymers, whereas other
clones accumulated P(3HB) homopolymers. E. coli harboring non-
mutated phaC
Ac
genes (wild type) synthesized P(3HB-co-3HHx) with a
3HHx fraction in the range of 713 mol%. From the HPLC assay, 22
mutants showed significantly or slightly higher 3HHx fractions than the
wild type. By repeated cultivation, we confirmed that particular mutants
listed in Table 1 showed good reproducibility in terms of high 3HHx
fractions. Nucleotide sequencing revealed these mutants were replaced
by glycine (G), alanine (A), valine (V), leucine (L), methionine (M),
glutamine (Q), and histidine (H), with the diversity of the codon at
position 171.
P(3HB-co-3HHx) synthesis by PhaC
Ac
mutants R. eutropha is
the best prospective workhorse for industrial production of PHA. To
examine the ability of PhaC
Ac
mutants to produce PHA under more
practical conditions, R. eutropha PHB

4 were used as a production

host using soybean oil, one of the most abundant vegetable oils in the
world. The mutated region of these phaC
Ac
genes was transferred to
the pBBREE32d13dPB and expressed in R. eutropha. Cultivation results
are listed in Table 2. All PhaC
Ac
mutants synthesized P(3HB-co-3HHx)
with higher 3HHx fractions than the wild-type enzyme, which was in
good agreement with the results of the HPLC-based assay. Of the
examined mutants, the highest 3HHx fraction was found in an N149S
mutant. Newly found mutants D171L/H/A showed equal or higher
3HHx fractions than the previously isolated mutant D171G. We
previously reported the synergistic effect of the combination of N149S
and D171G mutations in PhaC
Ac
, which increased reactivity toward
3HHx (10). Thus, to investigate the combinational effect of N149S and
D171A/L/H mutations, doubly mutated phaC
Ac
genes were con-
structed and expressed in R. eutropha. The newly generated double
mutants (NSDH, NSDA, and NADL) synthesized PHA with a compo-
sition similar to the NSDG mutant, suggesting a synergistic effect of
double mutation on substrate specificity. A very small amount of a 3-
hydroxyoctanoate (3HO) unit was detected in the polymers as well as
in the NSDG mutant.
Furthermore, P(3HB-co-3HHx) production was performed from
octanoate, which allows 3HHx production from sources other than
soybean oil (10). As a result, all strains accumulated PHA in the range
of 7080 wt.%. The 3HHx fractions of PHA are shown in Fig. 6A. 3HHx
fractions higher than the wild type were observed for all mutants. The
newdouble mutants exhibited slightly higher 3HHx fractions than the
NSDG mutant. Also, 3HO units were detected up to 0.4 mol% in PHA
synthesized by N149S and the double mutants.
P(3HB-co-3HV-co-3H4MV) synthesis by PhaC
Ac
mutants Re-
cently, our group discovered a newPHA component, 3H4MV, which is
able to be polymerized by PhaC
Ac
(4,5). The isolated mutants and their
double mutants were subjected to polymerization by P(3HB-co-3HV-
co-3H4MV) in R. eutropha PHB

4 by feeding 4-methylvalerate to
compare 3H4MV incorporation. These cells accumulated 1540 wt.%
PHA. The 3H4MV and 3HV fractions are shown in Fig. 6B. The double
0 10 20 30 40
(min)
R
e
l
a
t
i
v
e

i
n
t
e
n
s
i
t
y

(
A
2
1
0
)
C4
C6
C8
C10
C12
FIG. 4. HPLC analysis of P(3HB-co-21 mol% 3HA)-accumulating cells [R. eutropha PHB

4
harboring Pseudomonas sp. 613 PHA synthase gene (21)] treated with 1 NNaOH(solid
line) and the corresponding 2-alkenoic acid used as a standard reference material
(dotted line). P(3HB-co-21 mol% 3HA) consisted of 79 mol% 3HB (C4), 7 mol% 3HHx
(C6), 8 mol% 3HO (C8), 5 mol% 3HD (C10), and 1 mol% 3HDD (C12). The column was a
Bio-Rad Aminex HPX-87H (column length 300 mm).
0 2.5 5 7.5 10
(min)
R
e
l
a
t
i
v
e

i
n
t
e
n
s
i
t
y

(
A
2
1
0
)
C4
C6
FIG. 5. HPLC analysis of P(3HB-co-19 mol% 3HHx)-accumulating cells [R. eutropha
PHB

4/phaC
Ac
(10)] treated with 1 N NaOH. The column was a Bio-Rad Fast Acid
Analysis (column length 100 mm).
TABLE 1. Codons at position 171 of PhaC
Ac
mutants showing higher 3HHx fractions
than the wild-type enzyme using an HPLC assay.
At position 171 of PhaC
Ac
Clone name
Amino acid Codon
Gly(G) GGG 11-H5, 11-H7
GGT 11-F11, 15-D12
GGC 15-C4
Leu(L) TTA 11-B2, 15-D8
CTT 11-B7
TTG 15-B1
Ala(A) GCC 11-G4
GCT 11-C8
Val(V) GTT 11-H3
Met(M) ATG 11-E4
Gln(Q) CAA 11-A10
His(H) CAC 11-G8
HPLC ASSAY FOR PHA SYNTHASE MUTANTS 289 VOL. 113, 2012
mutants exhibited higher 3H4MV incorporation than the wild-type
and the single mutants. The trend of substrate specificities of the
mutants was similar regardless of the carbon source (Fig. 6C).
DISCUSSION
This study demonstrated that alkaline pretreatment of PHA-
accumulating cells enabled rapid and convenient analysis of PHA
composition by HPLC. Due to the ease of sample preparation, a high-
throughput HPLC assay based on PHA copolymer composition was
possible. In contrast, conventional sample pretreatment for an HPLC
assay, which uses acid instead of alkaline, was unable to analyze 3HA
units other than 3HB. The reason why 3HA was detected using
alkaline pretreatment is related to the reaction mechanism during
sample pretreatment. For sensitive detection of constituent mono-
mers of PHA, it is necessary to generate 2-alkenoic acid, which has
strong UV absorption due to unsaturated bonds, by eliminating the -
proton of 3HA unit (2224). The nucleophiles in this reaction are
water molecules and hydroxide ions in acidic and basic conditions,
respectively. The electron density at the -carbon of 3HA increases
with an increasing number of alkyl side chains. Therefore, proton
elimination from longer 3HA units by nucleophilic attack becomes
less efficient. Thus, water molecules, having weak nucleophilicity
for -protons, may not act as a nucleophile, but hydroxide ions
maintain nucleophilicity toward 3HA with longer alkyl side chains.
This may result in detection of mcl-PHA monomers with alkaline
pretreatment.
GC and nuclear magnetic resonance (NMR) analyses are gener-
alized methods for determination of PHA composition. As a
pretreatment for GC analysis, PHA samples are subjected to acidic
methanolysis and then extracted using organic solvents (25). Solvent
extraction is a time-consuming step of sample pretreatment for GC
analysis. For NMR analysis, PHAextraction fromcells and purification
are required. Therefore, these methods are preferable for precise
quantitative analysis of PHA composition but are not well suited for a
high-throughput assay. The sample pretreatment for the HPLC assay
established in this study consisted of simple procedures: NaOH
addition, heating at 100C, neutralization with HCl, and sample
filtration. Thus, all procedures can be completed on a 96-well plate. In
addition, by optimizing HPLC conditions, analysis of P(3HB-co-
3HHx) was shortened to 10 min. The analysis time will be further
shortened using the latest equipment, such as ultra high-perfor-
mance liquid chromatography (UHPLC). Although the HPLC assay is
rapid and convenient, the overlap of PHA- and cell-derived peaks
should be carefully considered. Therefore, the HPLC-based method is
TABLE 2. PHA production from soybean oil by recombinant R. eutropha PHB

4
expressing PhaC
Ac
mutants.
PhaC
Ac
Dry
cell
weight
(g/L)
PHA
content
(wt.%)
PHA composition
a
Molecular
weight
3HB
(mol%)
3HHx
(mol%)
3HO
(mol%)
M
n
(10
4
)
M
w
/
M
n
Wild-type 1.9
0.1
753 96.7 3.3 0 34 2.5
N149S
b
2.0
0.3
792 94.9 5.1 0 21 2.4
D171V 2.3
0.1
801 96.1 3.9 0 34 2.5
D171M 2.2
0.1
771 96.0 4.0 0 28 2.5
D171Q 2.4
0.1
841 95.8 4.2 0 25 2.9
D171G
b
2.3
0.2
811 95.6 4.4 0 24 2.5
D171H 2.2
0.1
821 95.4 4.6 0 33 2.5
D171A 2.1
0.1
831 95.2 4.8 0 29 3.1
D171L 2.3
0.1
843 95.2 4.8 0 40 2.6
N149S / D17 1G
(NSDG)
b
2.3
0.2
711 94.6 5.2 0.2 33 2.9
N149S / D171H
(NSDH)
2.8
0.1
761 94.9 5.0 0.1 28 3.6
N1 4 9S / D1 7 1 A
(NSDA)
2.6
0.1
772 94.7 5.2 0.1 35 3.5
N1 4 9 S / D1 7 1 L
(NSDL)
2.6
0.2
811 94.6 5.3 0.1 35 3.5
a
PHA composition was determined by GC.
b
PhaC
Ac
mutants that were generated in
previous studies (1012).
M
n
: number-average molecular weight. M
w
/M
n
: polydispersity index.
Results are the averages standard deviations from three separate experiments
(standard deviations of PHA composition are less than 5% of the mean).
3
H
H
x

+

3
H
O

(
m
o
l
%
)
0
5
10
15
20
25
3HO
3HHx
A
0
20
40
60
3HV
3H4MV
3
H
4
M
V

+

3
H
V

(
m
o
l
%
)
B
20
25
30
35
40
45
50
55
10 12 14 16 18 20 22
3
H
4
M
V

+

3
H
V

(
m
o
l
%
)
3HHx + 3HO (mol%)
C
WT
S
V
G
A
L
H
Q
NSDL
NSDH
NSDA
NSDG
M
FIG. 6. (A) 3HHx fraction of P(3HB-co-3HHx) copolymers produced fromoctanoate, and
(B) 3H4MV and 3HV fractions of P(3HB-co-3HV-co-3H4MV) copolymers produced
from 4-methylvaleric acid by R. eutropha PHB

4 expressing PhaC
Ac
mutants. PHA
composition was determined by GC. A3HOunit was detected as a minor component of
P(3HB-co-3HHx). Results are averages standard deviations from three separate
experiments. (C) Correlation between 3HHx+3HO fractions (shown in panel A) and
3H4MV+3HV fractions (shown in panel B) incorporated into PHA.
290 WATANABE ET AL. J. BIOSCI. BIOENG.,
preferable for a high-throughput assay but not well suited for precise
quantitative analysis.
To validate the HPLC assay, a D171X random point mutation
library of PhaC
Ac
was screened. Isolation of D171G mutants validated
the accuracy of the HPLC assay because the mutants showed higher
reactivity toward 3HHx than the wild-type enzyme, as previously
demonstrated (12). Additionally, this study aimed to verify whether
the D171G mutant is the most effective to increase PhaC
Ac
reactivity
toward 3HHx in the D171X library. The screening proved that the
HPLC assay functioned well because of successful isolation of the
D171G mutant and other new mutants showing equal or slightly
higher reactivity toward 3HHx than the D171G mutant. These isolated
mutants had a variety of codons at position 171 (Table 1), supporting
the diversity of the D171X library. In particular, D171G/L/A mutants
were repeatedly isolated with different codons. This implies that the
replacement by glycine (G), leucine (L), and alanine (A) leads to
increase in the reactivity toward 3HHx. On the other hand, the PHA
composition might be differentiated by difference in the expression
levels of PhaC
Ac
mutants. Previous study demonstrated that D171G
mutant was expressed in E. coli at the same level of wild-type enzyme,
as revealed by western blot analysis (12). Thus, it might be assumed
that expression levels among D171G/L/A mutants are not significantly
different.
The amino acids at position 171 of the isolated mutants were more
hydrophobic than aspartic acid (D) of the wild-type enzyme. Histidine
(H) can be either positively charged or uncharged at neutral pH; it can
be uncharged in the D171H mutant based on the screening results. It
is presumed that hydrophobicity at position 171 is related to the
substrate specificity of PhaC
Ac
. However, hydrophobicity was not
the only factor determining specificity because some hydrophobic
amino acids, such as isoleucine (I) and phenylalanine (F), were not
isolated. The structure of the amino acids seems to be a factor affecting
substrate specificity.
To our knowledge, the N149S mutation in PhaC
Ac
is the most
effective single mutation increasing reactivity toward 3HHx, as seen
from Table 2 and Fig. 6 (1012). We examined the amino acids at
position 149, other than serine (S), in terms of substrate specificity by
site-specific mutagenesis and preliminary HPLC-based screening of
the N149X mutant library. However, no mutant showed higher
reactivity toward 3HHx than the N149S mutant (data not shown).
From this result, we concluded that the most effective 149th amino
acid for increasing reactivity toward 3HHx is serine (S). Consequently,
the combined effect of the N149S and D171H/A/L mutations was
investigated in this study. These double mutations also exhibited
synergistic effects on substrate specificity as well as the NSDG
mutation.
In P(3HB-co-3H4MV) synthesis, an increase in the 3H4MV
fraction is also desired (26). Because 3H4MV is a structural isomer
of 3HHx, there is interest in studying these PhaC
Ac
mutants to assess
their ability to polymerize 3H4MV. As a result of this study, the
mutants were found to have an increased ability to polymerize
3H4MV in proportion to 3HHx (Fig. 6C). This observation provides
important insights into the mutational effect on the substrate
binding pocket of PhaC
Ac
. These mutations make the substrate
binding pocket deeper and wider due to the structural difference
between 3H4MV and 3HHx. Because there is a correlation between
the polymerization of both units, the screening approach used in
this study is applicable to mutants with an increased ability to
polymerize 3H4MV.
In conclusion, a new HPLC screening method based on PHA
copolymer composition was developed by applying alkaline sample
pretreatment and introducing a 96-well plate format. The new
method was used to isolate PhaC
Ac
mutants with increased reactivity
toward 3HHx from a D171X mutant library. For the 171st amino acid,
histidine (H), alanine (A), and leucine (L) were effective as well as
glycine (G). The double mutants, NSDH, NSDA, and NSDL, also
exhibited a synergistic effect on the alteration of substrate specificity,
as previously demonstrated by the NSDG mutant. The acquired
mutants have an increased ability to polymerize not only 3HHx but
also 3H4MV.
ACKNOWLEDGMENT
This work was supported by a Grant-in-Aid for Industrial
Technology Research Grant Program from the New Energy and
Industrial Technology Development Organization (NEDO) of Japan.
References
1. Sudesh, K., Abe, H., and Doi, Y.: Synthesis, structure and properties of polyhydroxyalk-
anoates: biological polyesters, Prog. Polym. Sci., 25, 15031555 (2000).
2. Tsuge, T.: Metabolic improvements and use of inexpensive carbon sources in
microbial production of polyhydroxyalkanoates, J. Biosci. Bioeng., 94, 579584
(2002).
3. Noda, I., Green, P. R., Satkowski, M. M., and Schechtman, L. A.: Preparation and
properties of a novel class of polyhydroxyalkanoate copolymers, Biomacromole-
cules, 6, 580586 (2005).
4. Tanadchangsaeng, N., Kitagawa, A., Yamamoto, T., Abe, H., and Tsuge, T.:
Identification, biosynthesis, and characterization of polyhydroxyalkanoate copol-
ymer consisting of 3-hydroxybutyrate and 3-hydroxy-4-methylvalerate, Bioma-
cromolecules, 10, 28662874 (2009).
5. Tanadchangsaeng, N., Tsuge, T., and Abe, H.: Comonomer compositional
distribution, physical properties, and enzymatic degradability of bacterial
poly(3-hydroxybutyrate-co-3-hydroxy-4-methylvalerate) copolyesters, Bioma-
cromolecules, 11, 16151622 (2010).
6. Doi, Y., Kitamura, S., and Abe, H.: Microbial synthesis and characterization of
poly(3-hydroxybutyrate-co-3-hydroxyhexanoate), Macromolecules, 28, 48224828
(1995).
7. Fukui, T. and Doi, Y.: Cloning and analysis of the poly(3-hydroxybutyrate-co-
3-hydroxyhexanoate) biosynthesis genes of Aeromonas caviae, J. Bacteriol., 179,
48214830 (1997).
8. Fukui, T. and Doi, Y.: Efficient production of polyhydroxyalkanoates from plant
oils by Alcaligenes eutrophus and its recombinant strain, Appl. Microbiol.
Biotechnol., 49, 333336 (1998).
9. Taguchi, S. and Doi, Y.: Evolution of polyhydroxyalkanoate (PHA) production
system by enzyme evolution: successful case studies of directed evolution,
Macromol. Biosci., 4, 146156 (2004).
10. Tsuge, T., Watanabe, S., Shimada, D., Abe, H., Doi, Y., and Taguchi, S.:
Combination of N149S and D171G mutations in Aeromonas caviae polyhydrox-
yalkanoate synthase and impact on polyhydroxyalkanoate biosynthesis, FEMS
Microbiol. Lett., 277, 217222 (2007).
11. Tsuge, T., Saito, Y., Kikkawa, Y., Hiraishi, T., and Doi, Y.: Biosynthesis and
compositional regulation of poly[(3-hydroxybutyrate)-co-(3-hydroxyhexanoate)]
in recombinant Ralstonia eutropha expressing mutated polyhydroxyalkanoate
synthase genes, Macromol. Biosci., 4, 238242 (2004).
12. Kichise, T., Taguchi, S., and Doi, Y.: Enhanced accumulation and changed
monomer composition in polyhydroxyalkanoate (PHA) copolyester by in vitro
evolution of Aeromonas caviae PHA synthase, Appl. Environ. Microbiol., 68,
24112419 (2002).
13. Taguchi, S., Maehara, A., Takase, K., Nakahara, M., Nakamura, H., and Doi, Y.:
Analysis of mutational effects of a polyhydroxybutyrate (PHB) polymerase on
bacterial PHB accumulation using an in vivo assay system, FEMS Microbiol. Lett.,
198, 6571 (2001).
14. Kovach, M. E., Elzer, P. H., Hill, D. S., Robertson, G. T., Farris, M. A., Roop, R. M.,
II, and Peterson, K. M.: Four new derivatives of the broad-host-range cloning
vector pBBR1MCS, carrying different antibiotic-resistance cassettes, Gene, 166,
175176 (1995).
15. Imai, Y., Matsushima, Y., Sugimura, T., and Terada, M.: A simple and rapid
method for generating a deletion by PCR, Nucleic Acids Res., 19, 2785 (1991).
16. Spiekermann, P., Rehm, B. H. A., Kalscheuer, R., Baumeister, D., and
Steinbchel, A.: A sensitive, viable-colony staining method using Nile red for
direct screening of bacteria that accumulate polyhydroxyalkanoic acids and other
lipid storage compounds, Arch. Microbiol., 171, 7380 (1999).
17. Friedrich, B., Hogrefe, C., and Schlegel, H. G.: Naturally occurring genetic transfer
of hydrogen-oxidizing ability between strains of Alcaligenes eutrophus, J. Bacteriol.,
147, 198205 (1981).
18. Kato, M., Bao, H. J., Kang, C. K., Fukui, T., and Doi, Y.: Production of a novel
copolyester of 3-hydroxybutyric acid and mediumchain length 3-hydroxyalkanoic
acids by Pseudomonas sp. 613 from sugars, Appl. Microbiol. Biotechnol., 45,
363370 (1996).
HPLC ASSAY FOR PHA SYNTHASE MUTANTS 291 VOL. 113, 2012
19. Karr, D. B., Waters, J. K., and Emerich, D. W.: Analysis of poly--hydroxybutyrate in
Rhizobium japonicum bacteroids by ion-exclusion high-pressure liquid chromatogra-
phy and UV detection, Appl. Environ. Microbiol., 46, 13391344 (1983).
20. Del Don, C., Hanselmann, K. W., Peduzzi, R., and Bachofen, R.: Biomass
composition and methods for the determination of metabolic reserve polymers
in phototrophic sulfur bacteria, Aquat. Sci., 56, 115 (1994).
21. Tsuge, T., Yamamoto, T., Yano, K., Abe, H., Doi, Y., and Taguchi, S.: Evaluating the
ability of polyhydroxyalkanoate synthase mutants to produce P(3HB-co-3HA)
from soybean oil, Macromol. Biosci., 9, 7178 (2009).
22. Law, J. H. and Slepecky, R. A.: Assay of poly--hydroxybutyric acid, J. Bacteriol., 82,
3336 (1961).
23. Slepecky, R. A. and Law, J. H.: Synthesis and degradation of poly--hydroxybutyric acid
in connection with sporulation of Bacillus megaterium, J. Bacteriol., 82, 3742 (1961).
24. Yu, J., Plackett, D., and Chen, L. X. L.: Kinetics and mechanism of the monomeric
products from abiotic hydrolysis of poly[(R)-3-hydroxybutyrate] under acidic
and alkaline conditions, Polym. Degrad. Stab., 89, 289299 (2005).
25. Braunegg, G., Sonnleitner, B., and Lafferty, R. M.: A rapid gas chromatographic
method for the determination of poly--hydroxybutyric acid in microbial biomass,
Eur. J. Appl. Microbiol. Biotechnol., 6, 2937 (1978).
26. Saika, A., Watanabe, Y., Sudesh, K., Abe, H., and Tsuge, T.: Enhanced incorporation
of 3-hydroxy-4-methylvalerate unit into biosynthetic polyhydroxyalkanoate using
leucine as a precursor, AMB Express, 1, 6 (2011).
292 WATANABE ET AL. J. BIOSCI. BIOENG.,