falciparum enolase dimer interface with respect to mammalian enolases could be exploited to selectively dissociate the dimeric parasite enzyme into its catalytically inefficient, thermall
Trang 1versus yeast and mammalian enolases to dissociation into active monomers
Ipsita Pal-Bhowmick, Sadagopan Krishnan and Gotam K Jarori
Department of Biological Sciences, Tata Institute of Fundamental Research, Colaba, Mumbai, India
Enolase (EC 4.2.1.11) is a glycolytic enzyme that
cata-lyzes the interconversion between
2-phospho-d-glycer-ate (PGA) and phosphoenolpyruv2-phospho-d-glycer-ate Enolases from
most organisms exist as homodimers of subunit mass
40–50 kDa [1], the exceptions being homo-octameric
enolases from thermophilic bacteria [2–4] A variety of
physical data indicate that each dimer contains two
active sites The active site in each subunit is
com-pletely independent [5,6] As the active site is fully
contained in each subunit, attempts to dissociate the
dimeric form into an active monomeric form have been made using genetic [7], chemical and physical methods, but without much success Although dissociation of the dimer could be achieved, the monomers formed were found to be inactive [8–11] The formation of act-ive monomers was inferred under conditions of high temperature (40–45C) and low protein concentration (in the nanomolar range) [12,13] However, at such low concentrations, neither the kinetic nor the struc-tural characterization of the monomer could be carried
Keywords
enolase; monomers; Plasmodium
falciparum; rabbit muscle; yeast
Correspondence
G K Jarori, Department of Biological
Sciences, TIFR, Homi Bhabha Road, Colaba,
Mumbai 400 005, India
Fax: +91 22 22804610
Tel: +91 22 22782000
E-mail: gkj@tifr.res.in
(Received 7 December 2006, revised 16
January 2007, accepted 12 February 2007)
doi:10.1111/j.1742-4658.2007.05738.x
In the past, several unsuccessful attempts have been made to dissociate homodimeric enolases into their active monomeric forms The main objective
of these studies had been to understand whether intersubunit interactions are essential for the catalytic and structural stability of enolases Further motivation to investigate the properties of monomeric enolase has arisen from several recent reports on the involvement of enolase in diverse non-glycolytic (moonlighting) functions, where it may occur in monomeric form Here, we report successful dissociation of dimeric enolases from Plas-modium falciparum, yeast and rabbit muscle into active and isolatable monomers Dimeric enolases could be dissociated into monomers by high concentrations ( 250 mm) of imidazole and ⁄ or hydrogen ions Two forms were separated using Superdex-75 gel filtration chromatography A detailed comparison of the kinetic and structural properties of monomeric and dimeric forms of recombinant P falciparum enolase showed differences in specific activity, salt-induced inhibition and inactivation, thermal stability, etc Furthermore, we found that enolases from the three species differ in their dimer dissociation profiles Specifically, on challenge with imidazole, Mg(II) protected the enolases of yeast and rabbit muscle but not of
P falciparum from dissociation The observed differential stability of the
P falciparum enolase dimer interface with respect to mammalian enolases could be exploited to selectively dissociate the dimeric parasite enzyme into its catalytically inefficient, thermally unstable monomeric form Thus eno-lase could be a novel therapeutic target for malaria
Abbreviations
M im , imidazole-generated monomer; M nt , native monomer; M pH , pH-generated monomer; 2-PGA, 2-phospho- D -glycerate; RMen, rabbit muscle enolase; r-Pfen, recombinant Plasmodium falciparum enolase; Yen, yeast enolase.
Trang 2out Hydrostatic pressure, which is viewed as a gentle
and reversible perturbation, has also been employed in
attempts to obtain active monomers from dimeric
eno-lase [8,11,14–17] Although the dissociation of enoeno-lase
dimers into active monomers was inferred in some of
the above studies, it has never been unequivocally
demonstrated As none of these attempts had resulted
in the isolation of active monomers and their
charac-terization, it was not clear whether the monomeric
form is intrinsically inactive or the means of
dissoci-ation selectively inactivated it Suggestions have also
been made that intersubunit interactions may be
essen-tial for completion of the catalytic cycle In recent
years, enolase has also been shown to participate in a
variety of nonglycolytic (moonlighting) biological
func-tions [18] The oligomeric state of enolase recruited for
the moonlighting functions is not known
Our interest in the enolase from the malarial
para-site (Plasmodium falciparum) originates from the fact
that the intraerythrocytic stages of the parasite lack
active mitochondria and hence rely solely on glycolysis
for their energy needs [19,20] The level of glycolytic
flux in parasite-infected cells is 50–100-fold greater
than that in uninfected red blood cells, and the activity
of many glycolytic enzymes is upregulated, enolase
being one of them [21–23] P falciparum enolase (Pfen)
could be a potential drug target, as there is only one
gene for this enzyme, and it shows greater resemblance
to plant enolases than to mammalian enolases In
order to examine such possibilities and explore
whe-ther it has any moonlighting functions, we have
recently cloned the P falciparum enolase gene,
over-expressed it, and obtained pure protein [24] We have
also raised polyclonal and monoclonal antibodies
against recombinant r-Pfen for subcellular localization
studies [25] Our observations have shown a diverse
subcellular localization (enolase is associated with
plasma membrane, cytosol, cytoskeletal elements and
nucleus; unpublished results) for enolase, indicating
that it may be recruited for certain other nonglycolytic
functions
One of the conventional approaches in rational drug
design has been to make active site-specific inhibitors
that can differentiate between host and parasite
pro-teins and bring about selective inhibition of the
para-site enzyme The major limitation of this strategy is
that active sites are evolutionarily highly conserved,
and structural differences may be rather subtle or even
nonexistent As there are numerous protein–protein
interactions that operate in living systems [26–28], and
most proteins exist as oligomers [29] with a
predomin-ance of homodimers [30], another approach could
be to target protein–protein interfaces for perturbing
protein and cell functions [31] In cases where oligo-meric structure is essential for biological activity, selective disruption of such an interface in a protein of parasite origin can have therapeutic effects Rationally designed peptides that can compete for the interaction between monomeric subunits (peptidomimetics) have yielded encouraging results [32–35] Attempts have also been made to find nonpeptide small molecule inhibi-tors that effectively interfere with protein–protein interactions [36–38] Here, we have examined the possi-bility of dissociating malarial parasite enolase into monomers using small molecules, and characterized the properties of the monomeric state We report the successful dissociation of r-Pfen, and yeast and rabbit muscle enolases, into isolatable active monomers Thus
we demonstrate that the dimeric structure is not essen-tial for catalysis However, comparative kinetic and structural studies on the monomeric and dimeric forms
of r-Pfen showed several interesting differences The monomeric form has low specific activity, is more ther-molabile, and is more prone to lose activity in the presence of salts as compared to the dimer Our experi-ments have also identified conditions under which selective dissociation of the parasite enolase may be accomplished Although the concentrations of dissoci-ating ligand used in this study are rather high and unrealistic for therapeutic applications, the possibility remains that such differences may be exploitable for targeting this enzyme for therapeutic purposes
Results
Imidazole-induced dissociation of r-Pfen into active monomers
Like most enolases, r-Pfen is a homodimer of two
50 kDa subunits [24] Purified dimeric r-Pfen was dia-lyzed against increasing concentrations (0–250 mm) of imidazole and then subjected to gel filtration chroma-tography on a Superdex-75 column Figure 1A shows gel filtration chromatograms obtained at several differ-ent concdiffer-entrations of imidazole at pH 6.0 A dimer peak was observed in the absence of imidazole How-ever, with increasing concentrations of imidazole, the monomeric fraction increased, and at 250 mm imidazole, > 95% of r-Pfen was dissociated The effect
of pH on imidazole-induced dissociation was examined
by dialyzing the enzyme against 50 mm sodium phos-phate at different pH values (6.0, 7.0 and 8.0) with increasing concentrations of imidazole The percentage
of monomer present in each of these samples was determined from gel filtration chromatograms The results presented in Fig 1B show that lower pH
Trang 3favored the dissociation of r-Pfen Figure 1C shows a
gel filtration chromatogram of a sample dialyzed
against 250 mm imidazole in 50 mm sodium phosphate
(pH 6.0) The enzyme activity of each fraction was
measured, and the specific activity was plotted along
with protein concentration (A280) as a function of
elu-tion volume The results showed that both forms
(monomeric and dimeric) were catalytically active, with
the dimer having 3-fold greater specific activity than
the monomer Experiments were also performed in
which the effects of adding NaCl to the dissociation
buffer were examined The results showed that
the presence or absence of salt (300 mm NaCl) did
not have any effect on monomer–dimer equilibrium
Imidazole-generated monomers of r-Pfen (Mim)
( 10 lm), when extensively dialyzed against
imida-zole-free buffer, could reassociate to form active
dimers To the best of our knowledge, this is the first
demonstration of obtaining active monomers of
enolase that could be separated from dimers
The ability of imidazole to dissociate enolases from
other species was also examined Imidazole at 250 mm
and pH 6.0 failed to dissociate yeast enolase (Yen)
However, inclusion of 300 mm NaCl along with
250 mm imidazole at pH 6.0 resulted in almost
com-plete dissociation of Yen (Fig 2A) In the case of
rab-bit muscle enolase (RMen), 250 mm imidazole did not
dissociate the enzyme As commercial preparations of
RMen contain Mg(II), we included 5 mm EDTA along
with imidazole This resulted in partial dissociation of
0
75
[Imidazole]
50 m M
250 m M
150 m M monomer
0 50
Ve (ml)
0 m M dimer
0 2 0 50
A r-Pfen
Ve (ml)
50 55 60 65 70 75 80 0
40 80
0 2 4
6 ) g
B
0 25 50 75 100
0 50 100 150 200 250 [Imidazole](m M )
pH 6.0
pH 7.0
pH 8.0
C
Fig 1 Imidazole-induced dissociation of r-Pfen (A) Superdex-75 gel filtration chromatograms obtained at different concentrations of imida-zole Enzyme (0.5 mg) was dialyzed ( 14 h) against 50 m M sodium phosphate (pH 6.0) containing different amounts (0–250 m M ) of imida-zole The column was pre-equilibrated with appropriate buffer, and chromatography was performed at room temperature (20 ± 1 C) (B) Effect of pH on imidazole-induced dissociation of r-Pfen Each data point is an average of two chromatographic runs (C) Gel filtration chro-matogram of r-Pfen in 50 m M sodium phosphate containing 250 m M imidazole (pH 6.0) A 280 nm and specific activity (r—r) for each fraction are shown.
Yen
0 3
(a)
monomer
Ve(ml)
0 10
dimer
dimer
(b) 0 2
(c)
RMen
B A
0 35
0
15 (c) (b) (a)
Ve(ml)
0 2
dimer monomer
dimer
dimer
Fig 2 Imidazole-induced dissociation of Yen and RMen Samples were in 50 m M sodium phosphate (pH 6.0), and 0.2–0.6 mg of protein was used for each chromatographic run on a Superdex-75 column (A) Yen: (a) no imidazole; (b) 250 m M imidazole; and (c) 250 m M imidazole + 300 m M NaCl (B) RMen: (a) no imidazole; (b) 250 m M imidazole; and (c) 250 m M imidazole + 5 m M EDTA.
Trang 4RMen (Fig 2B) However, further inclusion of
300 mm NaCl did not lead to complete dissociation of
RMen In this respect, RMen differs from Yen
As Mg(II) is an essential cofactor for stabilization of
the active conformation and catalysis, extensive kinetic
and direct metal ion-binding studies have been
per-formed in the past These studies have shown that each
enolase subunit has three binding sites for the divalent
cation, namely a conformational site (site I), a catalytic
site (site II) and an inhibitory site (site III) [39–42] As
binding of divalent cation induces large
conforma-tional changes in enolase, we examined the effect of
Mg(II) on imidazole-induced dissociation of r-Pfen,
Yen and RMen in the presence of different
com-pounds The results are summarized in Table 1
Inclu-sion of EDTA [to chelate residual Mg(II) in the
protein sample] and NaCl in the dissociation buffer
did not have any effect on the dimeric state However,
the presence of Mg(II) affected the imidazole-induced
(or imidazole + NaCl-induced) dissociation of RMen
and Yen It is interesting to note that in the presence
of 1.5 mm MgCl2, imidazole could dissociate r-Pfen
but had no effect on Yen and RMen
pH-induced dissociation of enolases The effect of pH on the oligomeric state of enolases was examined by performing gel filtration chromatog-raphy on the enzyme, dialyzed against 50 mm sodium phosphate of the desired pH Figure 3A presents the chromatographic profiles of r-Pfen in the pH range 4.5–8.0 Figure 3B shows that 50% dissociation of r-Pfen occurs around pH 5.5 Monomeric r-Pfen gener-ated by low pH (MpH) was also found to be enzymati-cally active (see below) Low pH could also dissociate Yen and RMen The half-dissociation point for Yen was around pH 5 (Fig 3C,D), whereas for RMen it was around pH 5.5 (Fig 3E,F)
Activity and reassociation of MpHand Mim Measurements of enzyme activity in monomeric and dimeric fractions showed that Mim had 3-fold less specific activity than dimers (Fig 1C) As the assay solution did not contain imidazole, it is possible that the observed activity of Mim may have arisen from reassociation to dimers Similarly, MpH could also reassociate at the assay solution pH of 7.4 Such reassociation of monomers into dimers during an enzyme assay would result in an increase in the slope
of the reaction progress curve (as the dimer is 3-fold more active than monomers) From these considera-tions, we can state that during an enzyme assay using
MpHor Mim: (a) if there is reassociation of monomers into dimer on the timescale of the enzyme assay, the reaction progress curve will exhibit a time-dependent increase in slope; and (b) as the amount of dimer formed increases as the square of the monomer con-centration, a plot of monomer concentration versus activity would have an upward curvature if there was
a significant amount of dimer formed during the enzyme assay
Figure 4A shows the reaction progress curves at three different concentrations of MpH (0.04, 0.17 and 0.43 lm) The observed increase in slope with time sug-gests reassociation of MpHinto dimers under our assay conditions A similar experiment performed using
Mim (Fig 4B) did not show any change in slope with time We measured the concentration dependence of specific activity for MpH at time 0.0 min as well as
at time > 1 min (Fig 4C) The first slope reflects the monomer activity, whereas the slope determined after > 1 min reflects the dimer activity With the increasing concentration of MpH, initially we observed
a very low specific activity ([MpH] > 0.5 lm) However,
at higher protein concentrations, the rate of formation
of dimer increased rapidly, and only a single slope,
Table 1 Effect of Mg(II) on monomer–dimer equilibrium in enolase.
Enolase, 10 l M (0.5 mgÆmL)1), was dialyzed against 50 m M
sodium phosphate (pH 6.0) containing different compounds as
indi-cated below Concentrations of these compounds when used
were: [imidazole] ¼ 250 m M ; [EDTA] ¼ 5 m M ; [NaCl] ¼ 300 m M ;
[MgCl 2 ] ¼ 1.5 m M The Superdex-75 column was pre-equilibrated
with the same buffer as used for dialysis.
Enolase
Dimeric–monomeric states of enolasesa
P falciparum (r-Pfen)
Yeast (Yen)
Rabbit muscle (RMen)
+ Imidazole + EDTA Monomer Dimer Partial
monomer + Imidazole + NaCl Monomer Monomer Partial
monomer + MgCl2+
imidazole + NaCl
+ MgCl 2 +
imidazole + EDTA
monomer + MgCl2+
imidazole +
EDTA + NaCl
Monomer Monomer Partial
monomer
a
When the amount of dimer or monomer is ‡ 90%, it is stated as
‘dimer’ or ‘monomer’ When the amount is ‡ 40%, it is stated as
‘partial monomer’.
Trang 5characteristic of dimer-specific activity, could be
observed (Fig 4C) As in the low-concentration range
of MpHthe observed activity is rather small, it is likely
that either MpHis inactive and the measured activity is
a reflection of the formation of tiny amounts of active
dimer, or the MpH has intrinsically very low activity
Irrespective of these two possibilities, the observed
sec-ond slope reflects the activity for the dimer formed
during the assay As expected, the specific activity
computed using the second slope did not exhibit the
monomer concentration-dependent variation (Fig 4C)
These results indicate that MpH rapidly reassociates to
form dimer However, similar measurements of
con-centration dependence of activity for Mimgave a linear
increase in activity, suggesting that r-Pfen Mimdid not
associate rapidly (with respect to assay timescale) to
form more active dimers (Fig 4D) A replot of these
data as specific activity versus Mim concentration (Fig 4D) showed that the observed specific activity was low ( 1.8–1.9 unitsÆmg)1, a characteristic of Mim) and was concentration invariant on enzyme assay time-scales These observations support the view that Mim associates slowly to form dimers, whereas MpH associ-ates rapidly Thus MpHand Mimdiffer in their ability
to reassociate, indicating that they represent two differ-ent conformational states of the monomeric form of the enolase A schematic representation of the dissoci-ation and associdissoci-ation of r-Pfen is presented in Scheme 1, where Mnt is the native monomer, which is assumed to be the only monomeric form that can dimerize As Mim mostly remained in the monomeric state under our assay conditions, kinetic characteriza-tion of the monomeric forms of enolase was performed using Mimonly
D
A
Fig 3 pH-induced dissociation of enolases from different organisms Each enzyme (0.2–0.6 mg) was dialyzed against 50 m M sodium phos-phate (pH 4.5–8.0) containing 150 m M NaCl and subjected to gel filtration chromatography (A) Gel filtration profile of r-Pfen at different pH values (B) pH versus % monomer (data are from two different experiments) for r-Pfen The r-Pfen used here was prepared by eluting the protein from Ni–nitrilotriacetic acid resin at low pH (C) Gel filtration chromatograms for Yen (D) pH versus% monomer for Yen (E) Gel filtra-tion chromatograms for RMen (F) pH versus % monomer for RMen.
Trang 6MpH from Yen and RMen also showed nonlinear
reaction progress curves, very similar to those observed
for r-Pfen (data not shown) In enzyme assays, Mim
from Yen gave nonlinear reaction progress curves
dis-playing an increase in slope with time This is likely to
be due to Mg(II)-induced rapid dimerization of yeast
Mim As Mg(II) does not affect the dimerization of
r-Pfen monomers, such a change of slope was not
observed for Mim prepared from r-Pfen (Fig 4B)
Observed low activities of monomeric (Mimand MpH)
enolases from all three species would imply that
quater-nary interactions between two subunits stabilize
cata-lytically more active conformations of each subunit
Comparison of kinetic properties of monomeric
(Mim) and dimeric forms of r-Pfen
As Mim did not reassociate rapidly into dimers, this
form of the enzyme was used for comparative kinetic
studies Figure 5A shows the variation of enzyme activ-ity as a function of phosphoenolpyruvate concentration Data were fitted to the Michaelis–Menten equation using sigmaplot software The nonlinear fit of data gave Vmax¼ 13.5 ± 0.5 UÆmg)1, and Km (phospho-enolpyruvate)¼ 0.28 ± 0.03 mm for dimers, and
Vmax¼ 3.7 ± 0.3 UÆmg)1, and Km (phosphoenolpyru-vate)¼ 0.38 ± 0.07 mm for monomers, respectively Thus disruption of subunit–subunit interactions resulted
in a significant decline in enzyme activity ( 3-fold), but did not have much effect on Km We also compared the thermal stability of monomeric (Mim) and dimeric forms Equal amounts of protein were incubated at three different temperatures (4C, 37 C and 50 C), and activity was assayed at different time intervals The dimeric form was stable for a prolonged (£ 250 min) duration at 37C, and showed 20–25% inactivation
at 50C In comparison, the monomeric form was
80% inactivated at 37 C (£ 250 min) and completely
0.00 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.50 1.379
1.39 1.40 1.41 1.42 1.43 1.44 1.45 1.464
B Imidazole monomer (Mim):
reaction progress curves
Time (minutes)
0.13
0.33
0.6 0.00 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.50
1.255
1.28
1.30
1.32
1.34
1.354
A pH monomer (MpH):
reaction progress curves
Time (minutes)
0.04
0.17
0.43
( µ M )
( µ M )
D Imidazole monomer (Mim):
concentration dependence of activity
C pH monomer (MpH):
concentration dependence of activity
0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40
0.6 0.8 1.0 1.2 1.4 1.6 1.8
[Monomer] ( µ M )
–1 )
[Monomer] ( µ M )
0.0
0.5
1.0
1.5
2.0
Fig 4 Reaction progress curves for the monomeric r-Pfen-catalyzed reaction Conversion of phosphoenolpyruvate to 2-PGA was monitored
at 240 nm (A) MpH (B) Mim Note the time-dependent change in progress curve slopes in (A) (C) Variation in specific activity as a function
of [M pH ], calculated from first (time 0.0 min) (d) and second (time > 1 min) (j) slopes of reaction progress curves Note the change in specific activity computed from the time 0.0 min slope, reflecting the rapid formation of dimer at higher concentrations of M pH (D) Vari-ation in activity (and specific activity) as a function of [Mim] Activity showed a linear increase (o), with specific activity remaining constant (d), suggesting that M im did not associate to form high-activity dimer on the enzyme assay timescale.
Trang 7inactivated at 50C in £ 150 min (Fig 5B) Thus,
subunit–subunit interface interactions confer higher
thermal stability to the protein
As enolases from different organisms are known to
differ in their response to different salts [43], we
exam-ined the effect of various salts on the catalytic activity of
the monomeric and dimeric forms To assess the effect
of a salt, we assayed the enzyme with assay
mix-tures containing different concentrations of a salt
Figure 6A–C shows the effects of NaCl, KCl and KBr
on the activity of the dimeric and monomeric (Mim)
forms of r-Pfen NaCl inhibited both forms of the
enzyme, but the inhibition was stronger for the
mono-meric form (Fig 6A) In the case of KCl and KBr, the
dimeric form was mildly activated ( 10–20%), whereas
the monomeric form was strongly inhibited (Fig 6B,C)
For assessing the activating⁄ inactivating effect of the
salts (NaCl, KCl and KBr), the enzyme was incubated in
buffer containing several different concentrations of salt
for 24 h at 20 ± 1C, and then assayed in the absence
of the salt The results are presented in Fig 6D–F At
median concentrations ( 300 mm), NaCl and KCl had
an activating effect on the dimeric form of the enzyme,
whereas all three salts had a concentration-dependent
inactivating effect on the monomeric form
Comparison of CD and fluorescence spectra
of monomeric and dimeric enolase
The effect of the loss of subunit–subunit interface
interactions on the secondary and⁄ or tertiary structure
of enolase was probed by recording CD and fluores-cence spectra of monomeric and dimeric forms of the protein As monomer preparations made by imidazole treatment contain about 250 mm imidazole, which interferes with fluorescence and CD spectra, it was not possible to record meaningful spectra for Mim Instead,
we recorded CD and fluorescence spectra for r-Pfen and Yen MpH, and compared them with the spectra from dimers (Fig 7) CD spectra of monomeric and dimeric forms (r-Pfen and Yen) were very similar (Fig 7A,C) Analysis of r-Pfen CD spectra using the
[PEP] (m M )
1-)
0 2 4 6 8 10
B A
Time (min)
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Fig 5 (A) Comparison of monomer and dimer activity with varying concentrations of phosphoenolpyruvate For activity measurements,
50 lL of 2 l M r-Pfen (monomeric or dimeric) was added to 450 lL
of assay mixture As monomeric protein was in 250 m M imidazole, the final concentration of imidazole in the assay mixture was
25 m M The presence of 25 m M imidazole in the assay mixture had
no effect on dimer activity Data were fitted to the Michaelis– Menten equation The best-fit parameters are K m (phos-phoenolpyruvate) ¼ 0.38 ± 0.07 m M and Vmax (specific activity) ¼ 3.7 ± 0.3 UÆmg)1 for monomer (j), and K m (phosphoenolpyru-vate) ¼ 0.28 ± 0.03 m M and V max ¼ 13.5 ± 0.5 UÆmg)1 for dimer (h) (B) Temperature dependence of stability of dimeric (open sym-bols) and monomeric (Mim) (filled symbols) forms of r-Pfen Enzyme was incubated at 4 C (s or d), 37 C (n or m) and 50 C (h or
j ) The activity was assayed at different time intervals.
+H + (slow) -H + (fast)
pH monomer
Imidazole
monomer
Native monomer
Native
+ imidazole
(slow)
-imidazole
(slow)
Scheme 1 Schematic representation of dissociation of the dimeric
(D) form of enolase to the imidazole-induced and pH-induced
mono-meric (Mimor MpH) forms The dimer is shown to dissociate into
the native form (M nt ), which has an enzyme activity similar to that
of the dimer Lowering of the pH or addition of imidazole (or both)
stabilize different conformations of monomers Mimand MpHdiffer
in activity and stability from the dimer and M nt M pH is rapidly
con-verted into M nt on raising of the pH, whereas M im is slow to
con-vert to Mnt, a form that is competent to form dimers.
Trang 8cdnn program for secondary structure analysis [44]
gave helix 36.6%, beta 14.1%, turn 19.3% and
ran-dom 29.9% for the dimeric form, and helix 31.5%,
beta 18.3%, turn 18.4% and random 31.8% for the
monomeric form Thus, it appears that dissociation of
dimers to form monomers lead to a slight decrease in
helical content with a concomitant increase in the beta
sheet content
Fluorescence emission spectra of the monomic,
dimeric and denatured states of r-Pfen and Yen are
shown in Fig 7B,D For both of these enzymes,
disso-ciation of enolase into monomers led to a decrease in
emission intensity In the case of r-Pfen, the emission
maximum was blue-shifted upon dissociation (Fig 7B;
compare traces a and b), whereas for Yen, a slight red
shift was observed (Fig 7D) Such a red shift may
arise because the dimer interface of Yen has Trp56
(which becomes solvent-exposed upon dissociation),
whereas the analogous position in Pfen is occupied by
Tyr59
Discussion
Alignment of enolase sequences from several different
organisms have shown that it is a highly conserved
protein [18] Furthermore, the comparison of known
three-dimensional structures showed complete
posi-tional conservation of active site residues across species [5,6,41,45–49] The enolase polypeptide chain folds into two domains, with the small domain having a mixture
of a-helices and b-sheets, and the large domain having
an a⁄ b-barrel structure (Fig 8D) The binding of sub-strate brings these two domains together to constitute
an active site As most enolases are homodimeric, the intersubunit interface is formed by contacts between the small domain of one subunit and the large domain
of the other subunit If the subunit–subunit interface is large, the dissociated monomers are generally unstable, due to exposure of large hydrophobic surfaces to the solvent [50] The intersubunit interface in enolases is rather small ( 11–13% buried surface) and quite hydrophilic, suggesting that it may be possible to dis-sociate the dimer into active monomers
In the past, several attempts have been made to dis-sociate the dimeric enolase into active monomers [7–9,11,14–17] However, the active monomeric form could never be obtained in isolation from the dimeric form Failure to obtain active monomers of enolase strengthened the belief that enolases are active only in the dimeric state [51] and that the quaternary structure
of enolase is necessary for catalytic activity Thus, any attempt to dissociate the dimer is accompanied by changes in the tertiary and secondary structures that are detrimental to enzyme activity [8] However, the
20
40
60
80
100
120
140
20 40 60 80 100 120 140
20 40 60 80 100 120 140
40
60
80
100
120
140
40 60 80 100 120 140
40 60 80 100 120 140
Fig 6 Effect of NaCl, KCl and KBr on the activity of monomeric (Mim) (j) and dimeric (h) r-Pfen Upper panel (a, b, c): Enzyme samples were assayed in the presence of different concentrations of salts Lower panel (d, e, f): Enzyme samples (200 lL of 2 l M r-Pfen) containing different concentrations of salts were incubated for 24 h at 20 C, and assayed in the absence of the salt For each assay, 50 lL of the enzyme was added to 450 lL of the assay mixture Activity is expressed as percentage of control where no salt was added to the enzyme sample.
Trang 9results presented here unequivocally establish that both
H+and imidazole are able to dissociate enolase dimers
into active monomers that could be isolated from the
dimeric form (Figs 1–3) We observed that at any
given concentration, the ability of imidazole to
dissoci-ate the dimer was enhanced by low pH (Fig 1B) In
the absence of imidazole, a change in pH (in the range
6–8) did not have any significant effect on the
dissoci-ation of the enzyme (Fig 1B), suggesting that the
increased dissociation induced by imidazole at low pH
is mostly due to pH-dependent ionization of imidazole
Imidazole has a pKaof 6.9, and hence at pH 6.0 it will
exist predominantly as an imidazolium ion Thus, the
data presented in Fig 1B are commensurate with the
idea that it is the imidazolium ion that is effective in
dissociating the r-Pfen dimer into monomers (Mim)
The pH-dependent dissociation of various enolases
indicate that it is the protonation of group(s) at the
intersubunit interface with a pKa 5–5.5 that is
responsible for the dissociation (Fig 3)
An examination of the enolase intersubunit interface
showed that it is stabilized by two salt bridges, several
hydrogen bonds, and p–cation (Tyr or Trp-Arg) and hydrophobic interactions A relatively low contribution
of the salt bridge in stabilizing the dimer interface was evident from the fact that replacement of an interface glutamate residue with a leucine (E414L) in RMen did not result into the dissociation of dimer DGo for dissociation for the mutant enzyme decreased from 49.7 kJÆmol)1to 42.3 kJÆmol)1[7] One may ask the fol-lowing question: which are the subunit interface inter-actions that imidazolium and⁄ or hydrogen ions can possibly disrupt? An examination of the intersubunit interface suggests the following two possibilities (a) There is a hydrogen bond between His191 NE2 and the carbonyl oxygen of Gly15 (or Arg14) where His191 NE2 acts as a proton donor (Fig 8A,B) Imidazole, being an analog of histidine, can compete for this hydrogen bond The importance of this interaction in stabilizing the dimer is evident from the observed chan-ges in the intersubunit hydrogen bond pattern on Mg(II) binding In one Mg(II)-bound form of Yen (Protein Data Bank: 2ONE, 1EBH), a hydrogen bond forms between His191 and Arg14 (Fig 8B) with a bond
-14000 -12000 -10000 -8000 -6000 -4000 -2000 0 2000 4000
1-Wavelength (nm)
-14000
-12000
-10000
-8000
-6000
-4000
-2000
0
2000
4000
1-)
Wavelength (nm)
Wavelength (nm)
0
1e+8
2e+8
3e+8
4e+8
5e+8
6e+8
7e+8
a b
c
Wavelength (nm)
0.0 5.0e+7 1.0e+8 1.5e+8 2.0e+8 2.5e+8 3.0e+8
3.5e+8
a b c
Yen r-Pfen
a
Yen
a b
r-Pfen
b
Fig 7 Comparison of CD and fluorescence spectra of monomeric (MpH) and dimeric forms of r-Pfen and Yen (A) CD spectra of (a) dimer and (b) monomer of r-Pfen (B) Fluorescence spectra of (a) dimer, (b) monomer and (c) urea-denatured r-Pfen (C) Yen CD spectra: (a) dimer, (b) monomer (D) Fluorescence spectra of (a) dimer, (b) monomer and (c) urea-denatured Yen Fluorescence emission intensities were nor-malized for protein concentration.
Trang 10length (as measured between two non-hydrogen atoms
in the hydrogen bond) of 3.2–3.3 A˚ In two
Mg(II)-bound forms (Protein Data Bank: 1ONE, 1EBG), the
hydrogen bond is between His191 and Gly15 (Fig 8A)
with a bond length of 2.9–3.0 A˚ Thus, the loss of
Mg(II) may favor weaker hydrogen bonds and
dissoci-ation, whereas addition of Mg(II) may reverse the
pro-cess, leading to subunit–subunit association It is
interesting to note that in the neuronal enolase crystal
structure (Protein Data Bank: 1TE6), where one
sub-unit is bound to one Mg(II) ion and the other subsub-unit
has two Mg(II) ions, both types of hydrogen bond
(His–Gly and His–Arg) are observed (Fig 8C) Our observations that Mg(II) favors dimer formation are in agreement with earlier reports [15], and suggest that interactions with His191 are critical for dimer stability (b) The imidazolium cation can compete with cationic side chains if a p–cation interaction is present at the dimer interface Recent analysis of dimeric interfaces has shown that p–cation interactions can make signifi-cant contributions to the binding energy for protein– protein complex formation, and it is suggested that such interactions be included in the list of criteria for characterizing protein interfaces [52] We examined the
Fig 8 Intersubunit interface hydrogen bond involving His191 in enolase crystal structures (A) Yen with two Mg(II) ions bound per subunit (Protein Data Bank: 1EBG) showing hydrogen bond between His191 and Gly15 (B) Yen with one Mg(II) ion bound per subunit (Protein Data Bank: 2ONE) has a hydrogen bond between His191 and Arg14 (C) Neuronal enolase in which two subunits are asymmetrically bound to one and two Mg(II) ions exhibits both types of hydrogen bond The subunit bound to one Mg(II) ion has a hydrogen bond between His189 and Arg14, and the other subunit, which is bound to two Mg(II) ions (Protein Data Bank: 1TE6), has a His189–Gly15 hydrogen bond (D) Rib-bon diagram of dimeric Yen (Protein Data Bank: 1EBG), in which yellow and blue represent two subunits Interface residues (Trp56 and Arg184) involved in p–cation interaction are shown in a stick representation.