Effects of the G376E and G157D mutations on the stability
of yeast enolase – a model for human muscle enolase
deficiency
Songping Zhao*, Bonny S. F. Choy* and Mary J. Kornblatt
Department of Chemistry and Biochemistry, Concordia University, Montreal, Canada
Enolase (EC 4.2.1.11), an essential enzyme of glyco-
lysis and gluconeogenesis, catalyses the intercon-
version of 2-phosphoglyceric acid (PGA) and
phosphoenolpyruvate. Enolases from most species are
dimeric, with subunit molecular masses of 40 000–
50 000 Da. Mammals have three genes for enolase,
coding for the a, b and c subunits; the subunits asso-
ciate to form both homo- and heterodimers. The
a gene is expressed in many tissues, c primarily in
neurones and b in muscle. In 2001, the first human
case of enolase deficiency was reported [1]. The
affected individual showed reduced levels of enolase
activity in the muscles. Western blot analysis showed
the presence of normal levels of aa-enolase, but no
detectable bb-enolase. This individual was heterozy-
gous for the gene for b-enolase, and carried two mis-
sense mutations, one inherited from each parent. His
muscle cells synthesized two forms of b-enolase, each
carrying a different mutation. These mutations chan-
ged glycine at position 374 to glutamate (G374E) and
glycine at position 156 to aspartate (G156D). In
order to study the effects of each of these mutations
on the structure and function of enolase, we have
made the corresponding changes, G376E and G157D,
in yeast (Saccharomyces cerevisiae) enolase. We chose
to work with yeast enolase, not bb-enolase, as yeast

values for subunit dissociation. At 37 °C, in the presence of
salt, both are partially dissociated and are extensively cleaved by trypsin.
Under the same conditions, wild-type enolase is fully dimeric and is only
slightly cleaved by trypsin. However, wild-type enolase is also extensively
cleaved if it is partially dissociated. The identification of the cleavage sites
and spectral studies of enolase have revealed some of the structural differ-
ences between the dimeric and monomeric forms of this enzyme.
Abbreviations
AUC, analytical ultracentrifugation; MES, 2-(N-morpholino)ethanesulfonic acid; PGA, 2-phosphoglyceric acid; PhAH,
phosphonoacetohydroxamate; Q-TOF, quadrupole time-of-flight; s
20,w
, sedimentation coefficient at 20 °C in pure water;
TLCK, N-a-tosyl-
L-lysine chloromethyl ketone.
FEBS Journal 275 (2008) 97–106 ª 2007 The Authors Journal compilation ª 2007 FEBS 97
yeast enolase can be overexpressed and purified in
quantity [4,5]. The basic three-dimensional structure
of the monomeric unit is the same in all enolases
crystallized to date [6–10]. Yeast and bb-enolase share
79% sequence similarity. All residues that have been
described as being involved in subunit interactions
(salt bridges and hydrogen bonds) [6], or contributing
to the active site, are conserved [2–5]. These two gly-
cines are conserved and are in highly conserved
regions of the protein: G376E is in a totally con-
served sequence of 25 amino acids, whereas G157D is
in a loop in which 11 of the 15 residues are con-
served. In view of the structural similarities between
yeast and mammalian enolases and the high degree of
sequence conservation, we believe that the effects of

G157D and 100 mg for G376E. All enolases were
highly pure, as judged by SDS-PAGE (not shown).
The specific activities of the variants, relative to wild-
type enolase, were 0.1% (G157D) and 0.01% (G376E).
MS confirmed that the desired mutations were present
(data not shown).
Secondary and tertiary structure
CD was used to examine the structure of these pro-
teins. In the peptide bond region, there were no signifi-
cant differences between wild-type enolase and the
variants, indicating that the variants were folded cor-
rectly. However, there were significant differences in
the aromatic region. The spectrum of the G157D vari-
ant was very similar to that of the wild-type protein
(Fig. 2A) in the region 280–300 nm; however, there
were differences in intensity below 280 nm. The CD
spectrum in the aromatic region of the G376E variant
was markedly different from that of the wild-type eno-
lase (Fig. 2A).
Quaternary structure and K
d
value for
dissociation
Analytical ultracentrifugation (AUC) was used to
examine the quaternary structure of the variants.
Wild-type enolase and the G157D variant had the
same sedimentation coefficient at 20 °C in pure water
(s
20,w
) at both 10.6 and 1.06 lm, indicating that they

NaClO
4
. For wild-type enolase, the K
d
value, extrapo-
lated to 0 m NaClO
4
(Fig. 3), was (1.5 ± 0.3) · 10
)8
.
The K
d
value for the G157D variant, determined in
a similar experiment, was increased by a factor of 10
relative to the wild-type (Table 1). Based on the AUC
data, the mutation at position 376 had a major effect
on the quaternary structure (Table 1). The s
20,w
value
for the G376E variant was measured at four protein
concentrations and the K
d
value for this variant was
calculated; K
d
was increased by a factor of 10
3
(Table 1).
Thermal denaturation
Temperature stability was studied by monitoring the

)8b
55.4
G367E 4.51 ± 0.14 (1.4 ± 0.3) · 10
)5c
51.2
G157D 5.65 ± 0.02 (1.8 ± 0.4) · 10
)7d
49.9
a
Average and standard deviations of two (G376E), three (G157D)
or four (wild-type) determinations; the standard deviation for individ-
ual deviations was < 0.01.
b
From perchlorate dissociation experi-
ment using AUC and CD data (Fig. 4).
c
Determined from AUC data
at four protein concentrations.
d
From perchlorate dissociation
experiment using CD data.
e
Mid-point of curves shown in Fig. 4.
Fig. 3. K
d
value for dissociation of wild-type enolase by NaClO
4
.
Enolase was incubated in varying concentrations of NaClO
4

Under the same conditions, there was partial cleavage
at one site by chymotrypsin, but no cleavage by pep-
sin, endoproteinase Glu-c or elastase (data not shown).
The cleaved samples were analysed by quadrupole
time-of-flight (Q-TOF) MS and the cleavage sites were
identified (Table 2). At 15 °C, there was no cleavage of
the wild-type enolase or of G157D by either trypsin or
chymotrypsin.
Does enolase become susceptible to cleavage by
trypsin when it is partially dissociated? Table 3 sum-
marizes the results of several experiments performed to
determine whether there is a correlation between disso-
ciation and susceptibility to cleavage by trypsin. For
all three forms of enolase, cleavage by trypsin occurs
when there is measurable dissociation. Shifting the
equilibrium towards monomers (37 °C and KCl for
G157D, NaClO
4
for the wild-type) promotes cleavage.
Shifting the equilibrium towards dimers (G376E plus
0.5 mm PhAH) provides substantial protection against
proteolysis. MS confirmed that the fragments produced
Fig. 4. Thermal denaturation of wild-type enolase and its variants,
as monitored by the CD signal at 222 nm. Full line, G376E at
10.6 l
M; long broken lines, wild-type at 10.6 lM; short broken lines,
G157D at 5.3 l
M; broken lines–dots, wild-type at 5.3 lM.
Fig. 5. SDS-PAGE analysis of tryptic digests. Samples were incu-
bated at 37 °C in the presence (+) or absence of trypsin for

4
0 Yes
G376E, 37 °C, TME, 0.15
M KCl 50 Yes
G157D, 37 °C, TME, 0.15
M KCl 88–90 Yes
Wild-type, 37 °C, TME, 0.15
M KCl 100 Very slight
a
Based on s
20,w
.
The G376E and G157D mutations in yeast enolase S. Zhao et al.
100 FEBS Journal 275 (2008) 97–106 ª 2007 The Authors Journal compilation ª 2007 FEBS
during the trypsin digest of wild-type monomers were
the same as those produced from the G376E variant.
Is monomeric enolase cleaved because it is partially
unfolded? Although the CD spectra of monomeric and
dimeric wild-type enolase appear to be identical in the
region 205–240 nm, differences become apparent when
spectra are recorded at lower wavelengths (Fig. 6).
These spectra were analysed using dichroweb [11,12],
with the variable selection method (cdsstr) [13].
According to this analysis, the percentage of unordered
structure in enolase increases from 17% in the dimeric
protein to 21% in the monomeric form. Three pairs of
monomeric and dimeric enolases, including enzyme
from two separate purifications, were examined. All
three showed a 4% increase in unordered structure on
dissociation.

tion of the subunit was increased by approximately 10
3
for the G376E variant: at 1 mgÆmL
)1
and 15 °C, the
protein was partially dissociated. The K
d
value for the
G157D variant was also increased, but by a smaller
amount, such that the protein was dimeric under our
standard conditions of 1 mgÆmL
)1
and 15 °C. Physio-
logical conditions of ionic strength and temperature
promoted the dissociation of both variants. This is not
surprising, as it has been reported that both salt [15]
and increasing temperature [16,17] favour dissociation
of the wild-type enolase. Conditions which promoted
dissociation also promoted proteolysis by trypsin.
The initial observation that we are trying to under-
stand is the lack of any bb-enolase in the muscle of
the patient [1]. We recognize that yeast enolase is not
identical to bb-enolase and that the cytoplasm of
mammalian cells does not contain trypsin or chymo-
trypsin. However, the results with the G376E and
G157D variants of yeast enolase show that these muta-
tions destabilize the protein and result in partial disso-
ciation. If, in muscle cells, the monomer is recognized
as abnormal and is degraded, the proteolysis of the
monomer would continually shift the dimer–monomer

Replacing glycine by glutamate or aspartate is a very
nonconservative change, although one that occurs in
nature as it requires only a single base change. Are the
effects that we have observed a result of a change in
the size or charge of the amino acid? Variants with
alanine at these positions were also studied. For tem-
perature denaturation, any change at these positions
was destabilizing. G376A and G376E had identical T
m
values. At position 157, alanine had a smaller effect
than aspartate, but even alanine decreased the T
m
value by 4 °C. There was no correlation between the
degree of dissociation and T
m
. Under the conditions
used for thermal denaturation, the G376E variant was
30% monomeric, whereas the G376A variant was
100% dimeric; however, their T
m
values were identical
and about 4 °C lower than that of the wild-type. A
different picture emerged when dissociation was stud-
ied, at least for G376E. In this case, alanine had little
or no effect on the K
d
value; G376A, at 1 mgÆmL
)1
,
had the same s

During the course of this study, it was observed that
both variants showed reduced enzymatic activity rela-
tive to the wild-type enolase. This is not surprising,
considering the location of these changes. Glycine 157
is in one of the loops that moves on binding of sub-
strate and divalent cation. This loop contains histidine
159, which is essential for catalysis. Nearby residues
that contribute to the stabilization of one of the transi-
tion states include 152, 155 and 168 [20]. Glycine 376
is close to residues 373 and 374, which are also impor-
tant for the reaction [20].
How does dissociation into monomers promote pro-
teolysis? Studies on these variants have revealed some
interesting differences between the monomeric and
dimeric forms of enolase. Trypsin cleaves at arginine
49. This residue is in a long loop (residues 36–60),
most of which is on the surface of the protein. How-
ever, this residue points into the protein, is surrounded
by other amino acids and is not accessible to trypsin.
On dissociation, there must be significant changes in
the conformation of this loop, leading to the exposure
of arginine 49. The chymotrypsin cleavage site,
between residues 56 and 57, is also in this loop. Tryp-
tophan 56 is surrounded by residues from both mono-
mers, and the backbone amide at position 56 of one
subunit is hydrogen bonded to the side chain of gluta-
mate 188 of the other subunit. Therefore, it is not sur-
prising that it is not accessible in the dimer.
The identification of the 56–57 bond as a site that is
hidden in the dimer, but accessible to chymotrypsin in

of this same loop. In a study of the pressure dissocia-
tion [26], we demonstrated that pressure dissociation
and inactivation of yeast enolase is a multistep process.
The first step, dissociation of the dimer into mono-
mers, is accompanied by small changes in the UV spec-
trum of the protein, changes which were attributed to
changes in the environment of several tyrosine resi-
dues. This is followed by conformational changes in
the monomer, which are reflected in further spectral
changes (both absorbance and fluorescence) and a loss
of activity. Based on our current data, we now propose
that the transition between the initial active monomers
formed by pressure and the subsequent inactive mono-
mers is a result of changes in the conformation of loop
36–60, changes similar to those observed in the current
experiments.
The other bond cleaved by trypsin is between resi-
dues 329 and 330. This bond is located in the last turn
of a small a-helix and far from the subunit interface.
We have no idea why it becomes susceptible to cleav-
age. We do not know whether the small increase in
disordered structure, observed in the CD spectrum,
affects this part of the protein, or whether there is
transient unfolding of the end of this helix. There are
examples of helices in proteins that undergo transient
unfolding, unfolding that is not apparent from the
crystal structure [27]. However, in neither case is it
obvious why this region of the enolase monomer
would be affected.
Comi et al. [1] suggested that the lack of bb-protein

was used for the expression of the protein.
Mutagenesis was performed using the QuickChange
method (Stratagene, La Jolla, CA, USA). The primer
sequences were as follows: 5¢-GG GGT GTT ATG GTT
TCC CATCGA TCT GAA GAA A CT GAA GAC (G376E)
and 5¢-CCA TTC TTG AAC GTT TTA AAC GGT GAT
TCC CAC GCT GGT GG (G157D). Each sequence differs
from that of the wild-type in two ways (the bases changed
are indicated in italic type): (a) a glycine codon was changed
to either a glutamate or aspartate; and (b) silent mutations
were introduced that produced new restriction sites. These
sites, BglII for G376E and AhaIII for G157D, were used for
screening purposes following mutagenesis of the gene. Ala-
nine was introduced at these positions using the same strat-
egy. DNA sequencing was performed by BioS&T, Inc.
(Lachine, Canada).
The expression of enolase was performed as described
previously [4]. The cell paste was either used immediately
or stored at )20 °C. Cell paste from 4 L of cells was sus-
pended in 60 mL of TME buffer [50 mm Tris ⁄ HCl, pH 7.4,
1mm Mg(OAc)
2
and 0.1 mm EDTA] containing 1 mm
phosphonoacetic acid and about 3 mg each of DNase and
RNase. The suspension was sonicated, on ice, using six 30 s
bursts per 10 g of cell paste. The suspension was cooled on
ice for 30–60 s between bursts. The pH was adjusted to 7.4
using 1 m Tris base and the sonicated cell suspension was
centrifuged at 24 000 g for 30 min at 4 °C. The supernatant
was decanted and recentrifuged at the same speed for

and 0.2 mm
EDTA. It was then applied to a CM-Sepharose Fast Flow
column equilibrated in MES buffer. Enolase was eluted by
a gradient of 0–0.25 m KCl in the same buffer. The purified
enolase was precipitated and stored as described above.
Purification of the G157 variants was identical to that of
the wild-type enzyme, except that the initial (NH
4
)
2
SO
4
cut
was 60–95%. For the G376E mutant, the (NH
4
)
2
SO
4
cut was 50–85% and the order of the chromatography steps
was reversed. The enzyme was first applied to the CM-
Sepharose column and eluted with the KCl gradient. Fol-
lowing precipitation of the pooled fractions by 4.3 m
(NH
4
)
2
SO
4
and dialysis against TME buffer containing

unless stated otherwise, and monitored at either 280 or
230 nm, depending on the protein concentration. Data
were analysed using dcdt+, version 1.15 or 2.02 (J. Philo,
www.jphilo.mailway.com); the viscosity and density of this
buffer were determined by sednterp, version 1.07 (D. B.
Hayes, T. Laue and J. Philo, available at www.bbri.org/
RASMB/rasmb.html). In order to determine the sedimen-
tation coefficients of dimeric and monomeric enolase,
measurements were made over a range of protein concen-
trations. In the presence of 0.14 m Na(OAc), enolase is
fully dimeric; s
20,w
, over a range of 2.1–0.056 mgÆmL
)1
,
was constant, with an average value of 5.49 ± 0.16. Simi-
larly, in the presence of 0.3 m NaClO
4
and protein con-
centrations ranging from 1.22 to 0.122 mgÆmL
)1
, s
20,w
was
also constant, with an average value of 3.35 ± 0.13. For
enolases that were partially dissociated, the concentration
of dimeric enzyme was calculated from the total protein
concentration and the s
20,w
value [31]:

, with a 1 nm bandwidth and a 1 s
response time. A minimum of four scans was averaged;
baseline subtraction and smoothing were performed using
jasco software. For temperature denaturation studies, the
sample was monitored at 222 nm. The temperature was
increased at a rate of 15 °C per hour. The CD signal was
used to calculate the fraction unfolded:
f
U
¼ðy
F
À yÞ=ðy
F
À y
U
Þð2Þ
where y
F
and y
U
are the CD signals at 222 nm for the ini-
tial and final forms of the protein, respectively [32] and y is
the signal of the sample. Samples for all CD experiments
were in TME buffer, unless otherwise stated. The protein
concentration was either 0.5 or 1.0 mgÆmL
)1
; in any given
experiment, mutant and wild-type enolases were at the same
concentration.
The K

D
Þð3Þ
where f
M
and f
D
are the fractions of monomeric and
dimeric enzyme, respectively. A plot of K
d
versus [NaClO
4
]
gives K
d
at 0 m NaClO
4
. The K
d
value for wild-type enolase
was also determined, using the same experimental design,
The G376E and G157D mutations in yeast enolase S. Zhao et al.
104 FEBS Journal 275 (2008) 97–106 ª 2007 The Authors Journal compilation ª 2007 FEBS
but measuring the s
20,w
value of each sample, in addition to
recording the CD spectrum. The data were then analysed
as described above.
Samples of enolase (1 mgÆmL
)1
) were incubated with

phosphonoacetohydroxamate at 2.1-A
˚
resolution.
Biochemistry 33, 9333–9342.
3 Lebioda L & Stec B (1991) Mechanism of enolase: the
crystal structure of enolase-Mg
2+
-2-phosphoglycer-
ate ⁄ phosphoenolpyruvate complex at 2.2-A
˚
resolution.
Biochemistry 30, 2817–2822.
4 Poyner RR, Laughlin LT, Sowa GA & Reed GH
(1996) Toward identification of acid ⁄ base catalysts in
the active site of enolase: comparison of the properties
of K345A, E168Q, and E211Q variants. Biochemistry
35, 1692–1699.
5 Vinarov DA & Nowak T (1999) Role of His159 in yeast
enolase catalysis. Biochemistry 38, 12138–12149.
6 Stec B & Lebioda L (1990) Refined structure of yeast
apo-enolase at 2.25-A
˚
resolution. J Mol Biol 211,
235–248.
7 da Silva Giotto MT, Hannaert V, Vertommen D, de
AS Navarro MV, Rider MH, Michels PA, Garratt RC
& Rigden DJ (2003) The crystal structure of Trypano-
soma brucei enolase: visualisation of the inhibitory
metal binding site III and potential as target for selec-
tive, irreversible inhibition. J Mol Biol 331, 653–665.

yeast enolase into active monomers. Biochemistry 10,
2506–2508.
17 Holleman WH (1973) The use of absorption optics to
measure dissociation of yeast enolase into enzymati-
cally active monomers. Biochim Biophys Acta 327,
176–185.
18 Brewer JM, Robson RL, Glover CV, Holland MJ &
Lebioda L (1993) Preparation and characterization of
the E168Q site-directed mutant of yeast enolase 1.
Proteins 17, 426–434.
19 Brewer JM, Glover CV, Holland MJ & Lebioda L
(1997) Effect of site-directed mutagenesis of His373 of
yeast enolase on some of its physical and enzymatic
properties. Biochim Biophys Acta 1340, 88–96.
20 Liu HY, Zhang YK & Yang WT (2000) How is the
active site of enolase organized to catalyze two different
reaction steps? Journal of the American Chemical
Society 122, 6560–6570.
21 Paladini AA & Weber G (1981) Pressure-induced
reversible dissociation of yeast enolase. Biochemistry 20,
2587–2593.
22 Brewer JM, Bastiaens P & Lee J (1987) Investigation of
conformational changes in yeast enolase using dynamic
S. Zhao et al. The G376E and G157D mutations in yeast enolase
FEBS Journal 275 (2008) 97–106 ª 2007 The Authors Journal compilation ª 2007 FEBS 105
fluorescence and steady-state quenching measurements.
Biochem Biophys Res Commun 147, 329–334.
23 Kornblatt MJ, Al-Ghanim A & Kornblatt JA (1996)
The effects of sodium perchlorate on rabbit muscle eno-
lase – spectral characterization of the monomer. Eur J

virus protease dimerization by analytical centrifugation.
Biochemistry 35, 15601–15610.
32 Pace CN & Scholtz JM (1997) Measuring the confor-
mational stability of a protein. In Protein Structure
(Creighton TE, ed.), pp 299–321. IRL Press, Oxford.
The G376E and G157D mutations in yeast enolase S. Zhao et al.
106 FEBS Journal 275 (2008) 97–106 ª 2007 The Authors Journal compilation ª 2007 FEBS