Eliminating positively charged lysine ε-NH3+ groups on the surface of carbonic anhydrase has no significant influence on its folding from sodium dodecyl sulfate

Article abstract of DOI:10.1021/ja043804d

This study compares the folding of two polypeptides-bovine carbonic anhydrase (BCA) and peracetylated BCA (BCA-Ac₁₈)-having the same sequence of amino acids but differing by 18 formal units of charge, from a solution containing denaturing conce

Full text of DOI:10.1021/ja043804d

Published on Web 03/09/2005
+
Eliminating Positively Charged Lysine E-NH3 Groups on the
Surface of Carbonic Anhydrase Has No Significant Influence
on Its Folding from Sodium Dodecyl Sulfate
Katherine L. Gudiksen, Irina Gitlin, Jerry Yang, Adam R. Urbach,
Demetri T. Moustakas, and George M. Whitesides*
Contribution from the Department of Chemistry and Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138
Received October 11, 2004; E-mail: gwhitesides@gmwgroup.harvard.edu
Abstract: This study compares the folding of two polypeptidessbovine carbonic anhydrase (BCA) and
peracetylated BCA (BCA-Ac18)shaving the same sequence of amino acids but differing by 18 formal units
of charge, from a solution containing denaturing concentrations of sodium dodecyl sulfate (SDS). Acetylation
+
of BCA with acetic anhydride converts all 18 lysine-ꢀ-NH3 groups to lysine-ꢀ-NHCOCH3 groups and
generates BCA-Ac18. Both BCA and BCA-Ac18 are catalytically active, and circular dichroism spectroscopy
(CD) suggests that they have similar secondary and tertiary structures. SDS at concentrations above ∼10
mM denatured both proteins. When the SDS was removed by dialysis, both proteins were regenerated in
native form. This study suggests that large differences in the net charge of the polypeptide have no significant
influence on the structure, the ability to refold, or the rate of refolding of this protein from solutions containing
SDS. This study reinforces the idea that charged residues on the surface of BCA do not guide protein
folding and raises the broader question of why proteins have charged residues on their surface, outside of
the region of the active site.
surface of a protein is variable, with reports of both significant8-10
and negligible effects.11 Local electrostatic interactions may shift
Introduction
The mechanism of folding of polypeptide chains into
functional proteins is an important and still unresolved problem.1-4
While burial of hydrophobic residues is considered the major
source of stability in a folded protein,5 the contribution of
electrostatic interactions is unclear and may differ from protein
to protein.6 In this study, we drastically increased the net charge
on the surface of a model proteinsbovine carbonic anhydrase
(BCA, EC 4.2.1.1)sin order to probe the effects of surface
charge on the ability of a protein to fold. Large differences (a
factor of approximately 5) in the net charge of BCA had no
significant influence on its refolded structure and only margin-
ally observable effects on the rate at which it folds.
The interactions between charged groups in a protein can be
classified as local (e.g., in salt bridges), where the two oppositely
charged groups are in proximity (e4 Å), or long-ranged, where
the charged residues interact with multiple other charged
residues over larger distances (>4 Å). Salt bridges buried in
the low dielectric interior of a protein can contribute as much
as 2-5 kcal/mol to protein stability.7 The magnitude of the
contribution to stability of salt bridges that are exposed on the
the equilibrium of protein folding toward the native state12,13
and constrain backbone flexibility in the native state.14,15 The
effects of long-ranged, nonspecific electrostatic interactions are
usually more subtle than local interactions. Many groups have
used site-directed mutagenesis of single charged amino acids
to uncharged analogues to show that reducing the overall charge
on the protein has a small (<0.5 kcal/mol), but usually
stabilizing, effect.16-19
Since it is difficult to modify large numbers of residues using
site-directed mutagenesis, the only studies that have made large
changes in the net charge of the protein have involved chemical
modifications. Hollecker and Creighton chemically modified a
(8) Perl, D.; Mueller, U.; Heinemann, U.; Schmid, F. X. Nature Struct. Biol.
2000, 7, 380-383.
(9) Spector, S.; Wang, M.; Carp, S. A.; Robblee, J.; Hendsch, Z. S.; Fairman,
R.; Tidor, B.; Raleigh, D. P. Biochemistry 2000, 39, 872-879.
(10) Horovitz, A.; Fersht, A. R. J. Mol. Biol. 1992, 224, 733-740.
(11) Serrano, L.; Horovitz, A.; Avron, B.; Bycroft, M.; Fersht, A. R. Biochemistry
1990, 29, 9343-9352.
(12) Kumar, S.; Nussinov, R. ChemBioChem 2002, 3, 604-617.
(13) Tsai, C.-J.; Ma, B.; Sham, Y. Y.; Kumar, S.; Nussinov, R. Proteins: Struct.,
Funct., Genet. 2001, 44, 418-427.
(14) Kumar, S.; Nussinov, R. Biophys. J. 2002, 83, 1595-1612.
(15) Sahal, D.; Balaram, P. Biochemistry 1986, 25, 6004-6013.
(16) Schwehm, J. M.; Fitch, C. A.; Dang, B. N.; Garcia-Moreno, E. B.; Stites,
W. E. Biochemistry 2003, 42, 1118-1128.
(1) Fersht, A. R.; Daggett, V. Cell 2002, 108, 573-582.
(2) Dobson, C. M. Nature 2003, 426, 884-890.
(3) Baldwin, R. L. Science 2002, 295, 1657-1658.
(4) Schymkowitz, J. W. H.; Rousseau, F.; Serrano, L. Proc. Natl. Acad. Sci.
U.S.A. 2002, 99, 15846-15848.
(5) Fersht, A. R. Structure and Mechanism in Protein Science; W. H.
Freeman: New York, 1999.
(6) Dill, K. A. Biochemistry 1990, 29, 7133-7155.
(7) Tissot, A. C.; Vuilleumier, S.; Fersht, A. R. Biochemistry 1996, 35, 6786-
6794.
(17) Sun, D. P.; Soederlind, E.; Baase, W. A.; Wozniak, J. A.; Sauer, U.;
Matthews, B. W. J. Mol. Biol. 1991, 221, 873-887.
(18) Meeker, A. K.; Garcia-Moreno E, B.; Shortle, D. Biochemistry 1996, 35,
6443-6449.
(19) Grimsley, G. R.; Shaw, K. L.; Fee, L. R.; Alston, R. W.; Huyghues-
Despointes, B. M. P.; Thurlkill, R. L.; Scholtz, J. M.; Pace, C. N. Protein
Sci. 1999, 8, 1843-1849.
9
10.1021/ja043804d CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 4707-4714
4707

A R T I C L E S
Gudiksen et al.
number of proteins using succinic anhydride to convert the
positive charge on the lysine residues to a negative charge.20
They found that the stability toward denaturation of the modi-
fied protein by urea was similar to that of the unmodified
proteins. Carbeck and co-workers used acetylation of lysine
groups to increase the charge on R-lactalbumin by ∼8 charges
(from ∼ -5.6 to ∼ -13).21 This change had a minimal effect
on the stability of the protein toward thermal denaturation
(∆∆Gunfolded-folded = 2.5 kcal/mol at 25 °C). In an attempt to
remove all electrostatic interactions in ubiquitin, Loladze and
Makhatadze used a combination of site-directed mutagenesis
and chemical modification.22 They first replaced all of the
arginine residues with lysine residues by site-directed mutagen-
tions.26-29 BCA is stable, has a single polypeptide chain, and
is readily available commercially. It has no cysteines and,
therefore, no disulfide bonds to complicate denaturation experi-
ments. Ten â-strands dominate the structure of BCA; they span
the width of the folded protein.30 The central â-strands are
composed primarily of hydrophobic residues and, even at 6.2
M guanidine (GuHCl), are not fully accessible to solvent.31,32
BCA has a stable molten-globule31,33 intermediate in its folding
pathway; the molten globule has a compact structure, significant
exposure of hydrophobic residues to solvent, and fluctuating
tertiary structure.34 The rate-limiting steps in the folding of BCA
are thought to be the isomerizations of two proline residues:
35,36 BCA has 18 proline residues, two of which are in the less
stable cis-conformation in the native structure.30
+
esis, carbomylated all of the lysine-ꢀ-NH3 groups, and then
examined denaturation at pH 2; at this value of pH, all of the
carboxylic acid residues are protonated and the net charge of
the protein is close to zero. They found that the chemically
modified ubiquitin (with all Arg mutated to Lys) was more
stable than unmodified ubiquitin toward denaturation with urea
(∆∆G ) 2.6 kcal/mol) and concluded that surface charge-
charge interactions are not essential for protein folding and
stability.
We eliminated 66% (18 out of 27) of the positively charged
residues on BCA by acetylating the 18 Lys-ꢀ-NH3+ side chains
on its surface;37 the nine arginine groups are unchanged. Each
modification neutralized the charge of a lysine residue, while
changing its size only minimally. We chose chemical modifica-
tion instead of site-directed mutagenesis in order to avoid the
potential of other changes in the structure of the protein, as well
as the technical difficulties of multiple rounds of mutation, and
of poor expression of mutants or of formation of inclusion
bodies.
Experimental Design. We wished to look at the effects of
large perturbations to the long-range interactions of surface
charges on the ability of a protein to fold into an active form
by increasing the charge on the surface of a model protein, BCA.
In general, we might expect that increasing the net charge on a
protein should contribute unfavorably to its stability, due to
increased electrostatic repulsion, and favorably to its solubility.23
Figure 1 shows the electrostatic potential at the surface of
BCA and BCA-Ac18 (see Supporting Information for details of
the calculation). The chemical modifications do not alter the
potential near the active site because there are no lysine groups
in or near the active site. The backside of the enzyme shows a
much larger change in potential than the front of the enzyme
and demonstrates that BCA is a good model system for studying
the effects of long-range electrostatic interactions in proteins
without large changes in the activity. The large amount of
negative (red) surface in BCA-Ac18 is a reflection of the large
net negative charge on this protein.
The increase in net charge due to chemical modification is
analogous to the increase in net charge upon changing the pH
+
to a value above the pKa of the Lys-ꢀ-NH3 groups. By either
changing the pH or adding charge to the surface of a protein
using chemical modifications, we expect a decrease in stability
due to the electrostatic repulsion between surface-exposed
ionized groups; this repulsion is minimized in the unfolded state.
We hypothesized that by substantially increasing the net charge
on the protein, without changing the pH, we might be able to
explore the influence of electrostatic repulsion on stability.
Of course, the removal of the charge on the lysine side chains
not only perturbs the network of electrostatic interactions in
the protein but also increases the hydrophobicity of these
residues. (The Hansch π parameter,38,39 that is, the hydropho-
bicity constant, for NH3 is -2.12,40 and for NHCOCH3 is
+
We chose to focus on comparing the refolding of BCA and
BCA-Ac18 after denaturation in SDS for two reasons. First,
SDS-PAGE is one of the most widely used techniques in
protein chemistry.24,25 The operation of SDS in this technique
has been extensively studied, but never fully understood. We
thus, wished to study SDS for its relevance to proteomics.
Second, SDS is a denaturant that is negatively charged, and we
thought that the chances of seeing an interaction between SDS
and proteins with different net charges were higher than with
urea.
(26) Urbach, A. R.; et al. Manuscript in preparation.
(27) Colton, I. J.; Carbeck, J. D.; Rao, J.; Whitesides, G. M. Electrophoresis
1998, 19, 367-382.
(28) Christianson, D. W.; Cox, J. D. Annu. ReV. Biochem. 1999, 68, 33-57.
(29) Christianson, D. W.; Fierke, C. A. Acc. Chem. Res. 1996, 29, 331-339.
(30) Eriksson, A. E.; Kylsten, P. M.; Jones, T. A.; Liljas, A. Proteins: Struct.,
Funct., Genet. 1988, 4, 283-293.
(31) Henkens, R. W.; Kitchell, B. B.; Lottich, S. C.; Stein, P. J.; Williams, T.
J. Biochemistry 1982, 21, 5918-5923.
(32) Svensson, M.; Jonasson, P.; Freskgaard, P.-O.; Jonsson, B.-H.; Lindgren,
M.; Maartensson, L.-G.; Gentile, M.; Boren, K.; Carlsson, U. Biochemistry
1995, 34, 8606-8620.
(33) Dolgikh, D. A.; Kolomiets, A. P.; Bolotina, I. A.; Ptitsyn, O. B. FEBS
Lett. 1984, 165, 88-92.
We chose BCA as the model protein for these studies because
it is a particularly well-characterized globular protein and is
often used as a model protein for physical-organic and
biophysical studies of proteins and protein-ligand interac-
(34) Ptitsyn, O. B.; Pain, R. H.; Semisotnov, G. V.; Zerovnik, E.; Razgulyaev,
O. I. FEBS Lett. 1990, 262, 20-24.
(35) Semisotnov, G. V.; Uverskii, V. N.; Sokolovskii, I. V.; Gutin, A. M.;
Razgulyaev, O. I.; Rodionova, N. A. J. Mol. Biol. 1990, 213, 561-568.
(36) Fransson, C.; Freskgaard, P. O.; Herbertsson, H.; Johansson, A.; Jonasson,
P.; Martensson, L. G.; Svensson, M.; Jonsson, B. H.; Carlsson, U. FEBS
Lett. 1992, 296, 90-94.
(20) Hollecker, M.; Creighton, T. E. Biochim. Biophys. Acta 1982, 701, 395-
404.
(21) Negin, R. S.; Carbeck, J. D. J. Am. Chem. Soc. 2002, 124, 2911-2916.
(22) Loladze, V. V.; Makhatadze, G. I. Protein Sci. 2002, 11, 174-177.
(23) Creighton, T. A. Proteins: Structure and Molecular Properties, 2nd ed.;
1984.
(24) Lilley, K. S.; Razzaq, A.; Dupree, P. Curr. Opin. Chem. Biol. 2002, 6,
46-50.
(25) Zhou, M.; Yu, L.-R. AdV. Protein Chem. 2003, 65, 57-84.
(37) The N-terminus of BCA is acetylated post-translationally and is, therefore,
not modified by acetic anhydride.
(38) The hydrophobicity constant is a measure of the change in energy to transfer
a compound from octanol to water with the change of functionality (e.g.
from NH3+ to NHCOCH3) to a parent molecule. A large number indicates
that the additional functionality requires more energy to transfer the
molecule to water.
(39) Fujita, T.; Iwasa, J.; Hansch, C. J. Am. Chem. Soc. 1964, 86, 5175-5180.
9
4708 J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005

+
Influence of Lys
ꢀ
-NH3 Groups on Folding of BCA
A R T I C L E S
on native BCA is -2.9 at this pH.46 The net change in charge
is thus approximately -16.
Briefly, the Linderstrom-Lang model assumes that a protein
is a sphere with uniform electrostatic potential on its surface.
This potential determines the local concentration of hydrogen
ions near the surface via a Boltzmann distribution. The
ionization state of every titratable group can be estimated using
the local proton concentration, rather than the bulk value, and
the ionization constant, pKa,i, of that group.47 The net charge
on the proteinsthe summation of all negative and positive
chargesswas in turn used to recalculate the surface potential
via the Debye-Huckel equation. This process was repeated until
the values become constant; the net charge was thus calculated
iteratively.
Denaturation and Renaturation of BCA and BCA-Ac18.
We denatured BCA and BCA-Ac18 in a buffered (25 mM Tris-
192 mM Gly, pH 8.3) solution of 10 mM SDS. Using
microcalorimetry,48,49 we measured the critical micelle concen-
tration (cmc) to be 4.3 mM in this medium (Supporting
Information). The solutions of denatured proteins were diluted
5-fold to reduce the SDS concentration below the cmc. The
diluted solutions were then dialyzed against Tris-Gly buffer
containing 10-µM ZnSO4 for 48 h to remove SDS; during this
dialysis, the proteins refolded.50
We used capillary electrophoresis (CE) to monitor the
denaturation of BCA and BCA-Ac18. As the protein associated
with negatively charged SDS molecules, its electrophoretic
mobility increased (Figure 2). Upon renaturation, the mobility
returned to its original value.
Figure 1. Calculations of electrostatic potential on the surface of BCA
and BCA-Ac18. Negative potentials at or below -8 kbT/e are colored red;
positive potentials at or above +8 kbT/e are colored blue (8 kbT/e ≈ 220
mV). Regions that are colored white have a potential of 0kbT/e. The active
site is circled in both proteins. The potential in the region around the active
site is similar in both BCA and BCA-Ac18; the back of the enzymes shows
significant differences, with BCA-Ac18 having a large, negative potential.
-1.21.5,41 Both groups are thus hydrophilic, but the NHCOCH3
is less so.) We anticipated that this increase in hydrophobicity
might allow the lysine-ꢀ-NHCOCH3 residues to partition into
the hydrophobic core of the protein, and thus, perhaps, to hinder
the folding of modified protein.
Measurement of Recovered Protein Activity and Struc-
ture. The amount of recovered protein after renaturation was
measured by comparing the absorbance at 280 nm before
denaturation and after renaturation. Recovery of total protein
was quantitative (>99.5%, as measured by absorbance at 214
nm) for both BCA and BCA-Ac18.
We assayed the amount of protein that was recovered and
that had the original electrophoretic mobilitysand hence charge
and sizesby measuring the area of the peak with the mobility
of the starting material in the electropherograms. The recovery
of protein with the original mobility was experimentally
indistinguishable for BCA (83 ( 5%) and BCA-Ac18 (78 (
5%).51 We did not observe the other 20% of the protein by CE.
We attribute this loss of protein to aggregation, or perhaps, to
refolding into a range of conformations with different mobilities;
aggregates and/or misfolded protein might appear in CE as broad
peaks indistinguishable from noise or baseline.
Results and Discussion
Preparation of BCA-Ac18. We peracetylated the ꢀ-amino
lysine groups of BCA to >90% conversion to the peracetylated
form (BCA-Ac18).37,42 The remaining 10% of the protein in the
sample was BCA-Ac17, with 17 out of the 18 lysine groups
acetylated, as determined by peak areas in CE separation.
Additional equivalents of acetic anhydride did not significantly
increase the amount of perfunctionalized species. We believe
that this conversion is sufficient to follow the behavior of the
majority species, BCA-Ac18, clearly.
BCA has 18 lysine groups, but the change in net charge from
BCA to BCA-Ac18 is smaller than 18 units due to charge
regulationsthe adjustment of protonation states of other ioniz-
able residues in a manner that reduces the total change in
charge.43,44 Using the Linderstrom-Lang model of cooperativity
in proton binding,45 we calculated the charge on BCA-Ac18 to
be -19 at pH 8.4 and an ionic strength of 10 mM; the charge
To determine if the proteins had refolded to the same state
as before denaturation, we measured both the activities (binding
of inhibitors and esterase activity) of the refolded proteins and
(46) This calculation was done using the linearized Poisson-Boltzmann equation
and is valid for potentials below 25 mV. The potential around the BCA-
Ac18 exceeds 25 mV and, therefore, this calculation is only an approxima-
tion.
(47) The pKa values for the titratable groups used were 12.5 for arginine, 10.3
for tyrosine and lysine, 7.0 for the hydroxide bound to the Zn(II) cofactor,
6.2 for histidine, 4.5 for glutamic acid, 3.2 for the C-terminus, and 2.5 for
aspartic acid.
(40) Pliska, V.; Schmidt, M.; Fauchere, J. L. J. Chromatogr. 1981, 216, 79-
92.
(41) Hansch, C.; Coats, E. J. Pharm. Sci. 1970, 59, 731-743.
(42) Yang, J.; Gitlin, I.; Krishnamurthy, V. M.; Vazquez, J. A.; Costello, C. E.;
Whitesides, G. M. J. Am. Chem. Soc. 2003, 125, 12392-12393.
(43) Menon, M. K.; Zydney, A. L. Anal. Chem. 2000, 72, 5714-5717.
(44) Gitlin, I.; Mayer, M.; Whitesides, G. M. J. Phys. Chem. B 2003, 107, 1466-
1472.
(48) Bai, G.; Santos, L. M. N. B. F.; Nichifor, M.; Lopes, A.; Bastos, M. J.
Phys. Chem. B 2004, 108, 405-413.
(49) Wang, G.; Olofsson, G. J. Phys. Chem. 1995, 99, 5588-5596.
(50) The addition of the Zn(II) ion, the cofactor of BCA, to the refolding buffer
increases the yield of refolded protein and causes all of the refolded protein
to be in the active, Zn(II)-containing form.
(45) Linderstrom-Lang, K. Compt. Rend. TraV. Lab. Carlsberg 1924, 15, 29
(51) The error reported in the amount of recovered protein is the difference in
the largest and smallest values measured from four replicate measurements.
pp.
9
J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005 4709

A R T I C L E S
Gudiksen et al.
Figure 2. Electropherograms showing BCA (B) and BCA-Ac18 (E) before
denaturation, after denaturation in 10 mM SDS (C, F), and after refolding
(D, G). An electropherogram of the charge ladder of BCA (Colton et al. J.
Am. Chem. Soc. 1997, 119, 12701-12709) (A) provides a reference for
the change of mobility of BCA as a function of the charge of BCA. The
mobility of the renatured peak matches that of the protein before denaturation
for both BCA and BCA-Ac18. For both BCA and BCA-Ac18, the aggregate
of protein with SDS has higher mobility than the protein in the absence of
SDS. Dimethyl formamide (DMF) served as an electrically neutral marker
to measure the rate of electroosmotic flow. An additional peak in D at µ )
0.6 cm2 V-1 s-1 corresponds to an active form of BCA with approximately
one additional negative charge (Gudiksen et al. Anal. Chem. 2004, 76,
7151-7161); we hypothesize that this peak corresponds to BCA in which
an asparagine, probably either Asp23 or -61, has been deamidated to an
aspartic acid.
Figure 3. Characterization of BCA and BCA-Ac18 before denaturation and
after renaturation. Binding of DNSA to (A) BCA and (B) BCA-Ac18,
measured by fluorescence spectroscopy.
constant for this reaction, kcat/Km, was determined by measuring
the rate of reaction as a function of the substrate concentration,
using a protein concentration of 1.5 µM estimated by the
intensity of UV absorbance at 280 nm (ꢀ280 ) 57 000 M-1
cm-1).52 BCA has a kcat/Km of 500 ( 50 M-1 s-1; BCA-Ac18
has a kcat/Km of 430 ( 40 M-1 s-1. We also determined the
observed values of kcat/Km for the refolded proteins: 400 ( 100
56
M-1 s-1 for BCA, and 390 ( 50 M-1 s-1 for BCA-Ac18 on
their tertiary structure (by circular dichroism spectroscopy, CD).
We measured the binding activity using a dansyl amide (DNSA)
binding assay.52 DNSA, when free in solution, has negligible
fluorescence (quantum yield 0.055, emission wavelength 580
nm). When bound to BCA, the quantum yield is enhanced by
a factor of 15 (quantum yield 0.84, emission wavelength 468
nm).52 To monitor binding, we measured the intensity of
fluorescence as a function of the concentration of DNSA. The
modification of the lysine groups on BCA does not significantly
change the activity of the protein (Figure 3): the dissociation
constant (Kd) for the BCA‚DNSA complex was 0.15 ( 0.03
µM (Figure 3A); Kd for the BCA-Ac18‚DNSA complex was 0.34
( 0.03 µM (Figure 3B).53 The dissociation constants for the
refolded proteins (and thus, we infer, their active site structures)
for both BCA (Kd ) 0.19 ( 0.02 µM) and BCA-Ac18 (Kd )
0.32 ( 0.03 µM) are restored upon renaturation.
the basis of the total protein concentration in solution. The
apparent values of kcat/Km for the refolded material were 81%
of that of the starting material for BCA and 91% of that of the
starting material for BCA-Ac18; these values of rates are
consistent with our estimate by CE that ∼80% of the refolded
protein was in the active form.
Further evidence that the refolded proteins had regained their
original fold comes from CD. The CD spectrum is a sensitive
measure of both the secondary and tertiary structure of a protein.
CD indicated that refolded BCA and refolded BCA-Ac18 had
structures very similar to the proteins from which they were
derived (Figure 4).
Because both polypeptides refolded to structures having
activities and inhibitor-binding properties indistinguishable from
those before denaturation, we conclude that, for this protein,
the charges distributed on the surface do not influence the ability
to fold into the original state.
Rates of Refolding. We also investigated the kinetics of
refolding of BCA and BCA-Ac18 from SDS solution by
monitoring the quenching of tryptophan residues by acrylamide
following rapid (<30 s) dilution. A solution containing 30 µM
denatured protein and 10 mM SDS was diluted 100-fold into a
In addition to the binding of inhibitors, we assayed the
undenatured and refolded proteins for their enzymatic activity.
BCA catalyses the hydrolysis of p-nitrophenylacetate to p-
nitrophenoxide and acetate.54,55 The apparent second-order rate
(52) Chen, R. F.; Kernohan, J. C. J. Biol. Chem. 1967, 242, 5813-5823.
(53) The error reported in the dissociation constants of DNSA from BCA is the
difference in the largest and smallest values measured from three replicate
measurements.
(54) Pocker, Y.; Stone, J. T. J. Am. Chem. Soc. 1965, 87, 5497-5498.
(55) Pocker, Y.; Stone, J. T. Biochemistry 1967, 6, 668-678.
(56) The error reported in the activity assay is the difference in the largest and
smallest values measured from three replicate measurements.
9
4710 J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005

+
Influence of Lys
ꢀ
-NH3 Groups on Folding of BCA
A R T I C L E S
Figure 4. CD spectra of BCA in the (A) far and (B) near UV regions and BCA-Ac18 in the (C) far and (D) near UV regions.
phobic (acrylamide excluded) environment, with the tryptophans
in interaction with SDS molecules, to a hydrophilic environment,
with the tryptophans exposed to solvent, and, finally, to an
internally folded state of the protein with the tryptophans buried
(and thus also excluded from quenching by acrylamide). The
time-scale of this decrease did not change when we changed
the acrylamide concentration from 0.2 to 0.04 M (data not
shown). The two observed time-scales in the refolding are t1/2
) 1.2 ( 0.4 min, for the decrease in fluorescence, and t1/2
)
11 ( 1 min, for the subsequent increase in fluorescence for
58
BCA, and t1/2 ) 1.0 ( 0.3 and 21 ( 2 min, for BCA-Ac18.
The shorter time-scale was indistinguishable for BCA and BCA-
Ac18; the longer of the two time-scales was similar for BCA
and BCA-AC18, but may differ by as much as a factor of 2.
Because of the mixing time for the dilution of the protein, we
were unable to measure any processes occurring on a time-
scale of less than ∼20 s.
Figure 5. Kinetics of refolding of BCA and BCA-Ac18 upon rapid (<30
s) dilution in the presence 0.2 M acrylamide. Tryptophan fluorescence was
excited at 280 nm and the emission was measured at 340 nm. The inset
shows the data plotted on a logarithmic scale to demonstrate the two time-
scales, an initial decrease followed by an increase in fluorescence, in the
renaturation. Only the data in the dotted box (before 35 min) was used for
the inset.
Many groups have observed three time-scales in the refolding
of BCA after denaturation with GuHCl.33,35,59-61 Semisotnov
et al. showed that the fastest of the three stages (t1/2 = 0.04 s)
is a compaction of the protein chain and the formation of a
molten-globule intermediate.61 Jonasson and co-workers found
that the hydrophobic core of BCA forms on a time-scale of a
few milliseconds.62 In this fast collapse, the hydrophobic clusters
solution containing Tris-Gly buffer with 10 µM ZnSO4 and 0.2
M acrylamide. Acrylamide is an effective quencher of tryp-
tophan fluorescence. Since quenching is strongly dependent on
distance between fluorophores, the tryptophan fluorescence
would be quenched when the protein is denatured and the
tryptophan residues are exposed to solvent. As BCA or BCA-
Ac18 folds and the tryptophan residues are buried, the distance
between quencher and fluorophore increases and the observed
fluorescence increases. We observed an initial decrease in
fluorescence and a subsequent increase (Figure 5) during the
refolding of both BCA and BCA-Ac18. We speculate that the
initial decrease in tryptophan fluorescence and subsequent
increase could reflect the transition of moving from a hydro-
(57) Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 2nd ed.; Kluwer
Academic/Plenum Publishers: New York, 1999; p 238.
(58) The error reported in the refolding kinetics is the difference in the largest
and smallest values measured from five replicate measurements.
(59) Carlsson, U., and Jonsson, B.-H. In The Carbonic Anhydrases: New
Horizons; Chegwidden, W. R., Carter, N. D., Edwards, Y., H., Eds.;
Birkhaeuser Verlag: Basel, 2000, pp 241-259.
(60) Wong, K.-P.; Tanford, C. J. Biol. Chem. 1973, 248, 8518-8523.
(61) Semisotnov, G. V.; Rodionova, N. A.; Kutyshenko, V. P.; Ebert, B.; Blanck,
J.; Ptitsyn, O. B. FEBS Lett. 1987, 224, 9-13.
(62) Jonasson, P.; Aronsson, G.; Carlsson, U.; Jonsson, B. H. Biochemistry 1997,
36, 5142-5148.
9
J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005 4711

A R T I C L E S
Gudiksen et al.
Table 1. Charges and Charged Residues on Soluble CA Isoforms
enzyme
accession
number
molecular
mass (Da)
positive
residuesc
negative
residuesd
total
label
isoform
organism
chargea
pIb
residues
A
B
C
D
E
F
G
H
I
CA 1
CA 2
CA 1
HCA XIII
CA 2
CA 2
HCA VI
CA 1
HCA I
CA 1
CA 2
flaveria
bovine
mouse
human
mouse
chicken
human
horse
human
gorilla
flaveria
macaque
human
human
mouse
mouse
human
rhesus
human
opossum
horse
-2.3
-1.5
-1.4
-1.4
-1.2
-1.1
-0.8
-0.8
-0.7
-0.7
-0.1
0.0
0.2
0.2
0.2
0.3
0.4
0.9
1.0
1.2
5.70
6.58
6.47
6.45
6.52
6.55
6.65
6.59
6.63
6.63
6.21
6.88
6.86
6.94
6.86
6.97
6.92
7.34
7.18
7.27
7.84
35 906
28 980
28 482
29 734
29 228
29 156
35 601
28 613
29 019
29 030
20 801
29 087
29 394
29 767
36 977
29 574
29 961
29 085
35 059
29 442
29 707
34.5
28.1
22.2
28.2
27.4
26.5
27.8
27.0
26.1
26.1
19.5
27.0
32.2
31.2
38.2
31.3
27.4
27.0
28.1
29.4
33.1
37.0
29.0
24.0
30.0
29.0
28.0
29.0
28.0
27.0
27.0
19.0
27.0
32.0
31.0
38.0
31.0
27.0
26.0
27.0
28.0
31.0
330
261
260
262
259
259
308
255
260
260
190
260
259
259
317
259
264
260
305
262
259
P46512
P00921
P13634
Q8N1Q1
P00920
P07630
P23280
P00917
P00915
Q7M316
P46513
P35217
P00918
P07451
Q9QZA0
P16015
P43166
P00916
P35218
Q8HY33
P07450
J
K
L
M
N
O
P
Q
R
S
CA 1
HCA II
HCA III
CA V
CA 3
HCA VII
CA 1
HCA V
CA 1
CA 3
T
U
1.8
a The charge on the protein was calculated at the pH of the organism (pH 6.5 for flaveria, pH 7.4 for all others) using a Linderstrom-Lang model. The
details of the calculation are provided in the text. b The isoelectric point (pI) was calculated using the EXPASY isoelectric point calculator: http://us.expasy.org/
tools/pi_tool.html. c The number of positively charged residues is the number of lysine + arginine + Xi × histidine residues, where Xi is the fractional
charge of the histidine residues at the pH of the calculation (Xi ) 0.5 at pH 6.5 and 0.1 at pH 7.4). d The number of negatively charged residues is the
number of aspartic acid + glutamic acid residues.
desolvate and a nativelike hydrophobic core forms. The other
two stages (t1/2 = 2 and 10 min) have been attributed to
isomerizations of two proline groups.35, 36
stable. We emphasize that we do not directly observe the
hydrophobic collapse in our measurements of the kinetics of
refolding because it is faster than the time resolution of our
instrument. The lysine residues, therefore, might, in principle,
affect the kinetics of the hydrophobic collapse; we simply are
unable to examine this process. Nonetheless, the acetylation of
lysine residues does not prevent the folding of BCA, and we
conclude that exclusion of charges from the interior of the
protein plays little or no role in the ability of this particular
protein to fold.
If charge outside of the active site does not influence the
secondary and tertiary structure of BCA (and by extension, of
at least some other proteins) and does not determine the rate or
product of folding, what is the role of charged residues on its
surface (and on the surface of other proteins)? We can, at this
point, only speculate. Charges on the surface of a protein may
be important for preventing aggregation63 or for increasing
solubility.64 Charges can affect substrate binding and reactiv-
ity,65,66 protein-protein interactions,67 and secretion68 or local-
ization of specific proteins. In addition, the osmotic pressure
of a cell is affected by the sum of the charges of all of its
proteins.
Because we cannot observe processes faster than ∼20 s, it is
possible that BCA and BCA-Ac18 have different rates of collapse
into a molten-globule intermediate. The two time-scales that
we observe match those reported in the literature when BCA is
refolded after denaturation with GuHCl.33,35,36,59-61 After rena-
turation from GuHCl, the formation of the molten globule is
orders of magnitude fastershundreds of millisecondssthan the
time-scale we measure. We, therefore, think it is unlikely that
the rate-limiting step measured here is the exclusion of water
from the hydrophobic core of the molecule. We thus attribute
the two time-scales we measured to isomerizations of two
proline groups. The modification of the lysine residues of BCA
appears to have only a minimal effect on the rate of proline
isomerizations. Since we cannot observe SDS directly, it is
possible that the dissociation of SDS from a protein‚SDS
complex also occurs on this time-scale.
Analysis and Conclusions
This work demonstrates that a change in charge of ∼ -16
units on the surface of BCA neither prevents refolding into a
catalytically active state nor significantly changes the kinetics
of refolding after denaturation with SDS. Changing the charge
also does not significantly alter the activity of the protein, as
measured by the esterase activity and the binding of inhibitor.
Acetylated lysinesswhich have higher hydrophobicity than the
charged, unmodified lysinessare not incorporated into the core
upon folding. This result may imply that the hydrophobic
collapse is due to a few specific residues in the primary sequence
and lysines, acetylated or not, do not participate. We, however,
do not wish to draw inferences about hydrophobic collapse from
this result, or to generalize it to other proteins. CA is only one
protein, and it may be a special case, since it is particularly
To address the question of whether the net charge of the
carbonic anhydrase (CA) isoforms is conserved, we analyzed
20 isoforms of CA that are catalytically active and not
membrane bound (Table 1). We calculated the charges on these
proteins using the charge regulation calculation described in the
(63) Chiti, F.; Stefani, M.; Taddei, N.; Ramponi, G.; Dobson, C. M. Nature
2003, 424, 805-808.
(64) Pace, C. N.; Alston, R. W.; Shaw, K. L. Protein Sci. 2000, 9, 1395-1398.
(65) Garcia-Viloca, M.; Gao, J.; Karplus, M.; Truhlar, D. G. Science 2004, 303,
186-195.
(66) Caravella, J. A.; Carbeck, J. D.; Duffy, D. C.; Whitesides, G. M.; Tidor,
B. J. Am. Chem. Soc. 1999, 121, 4340-4347.
(67) Davis, S. J.; Davies, E. A.; Tucknott, M. G.; Jones, E. Y.; van der Merwe,
P. A. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 5490-5494.
(68) Stephenson, K.; Jensen, C. L.; Jorgensen, S. T.; Lakey, J. H.; Harwood:
C. R. Biochem. J. 2000, 350, 31-39.
9
4712 J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005

+
Influence of Lys
ꢀ
-NH3 Groups on Folding of BCA
A R T I C L E S
Figure 6. Histogram of calculated charges on CA isoforms.
Figure 7. A plot of the number of negative amino acids divided by the
total number of amino acids versus the number of positive amino acids
divided by the total number of amino acids for the 20 CA isoforms listed
in Table 1, identified by the label to the right of each data point. The line
of best fit has a slope of 0.81 and a correlation coefficient of 0.92.
Results section that we used to calculate the charge on BCA
and BCA-Ac18. We performed this calculation at the pH of the
organism or, for mammals, the pH of blood. Figure 6 shows a
histogram of the charges found on the CA isoforms. The charge
on these CA isoforms varied from +1.8 to -2.3, with 90% of
them between +1.5 and -1.5. The charge on BCAII in blood
(pH 7.4) is nearly a full charge less than that in Tris-Gly buffer
(pH 8.4) because the eight titratable histidine residues and the
hydroxide bound to the Zn(II) cofactor are protonated to a larger
extent at the lower pH.
Figure 8. Depictions of calculated electrostatic potentials on the surface
of the four CA isoforms in the Protein Data Bank. The colors and scales
match those of Figure 1. Protein data bank id numbers are 1ca2 for HCA
II, 1v9e for BCA II, 1crm for HCA I, and 1urt for mouse CA V. The active
site is circled in each isoform for reference. The backbone has an extremely
similar conformation in all four isoforms. The blue, positive patch on the
side of the enzymes, shown with an arrow, appears to be conserved across
all four isoforms.
If there were evolutionary pressure to maintain the charge
on CA between +1 and -2, changes in the number of positively
charged amino acids in the protein would be balanced by
corresponding changes in the number of negatively charged
amino acids. In particular, a mutation that increased or decreased
the charge on CA would be more likely to be followed by a
mutation that restores the charge to the original value than one
that left the new charge unaltered. We wanted to see whether
an increase or decrease in the number of negatiVely charged
amino acids in a CA isoform correlated with an increase or
decrease in the number of positiVely charged amino acids.
Specifically, we hypothesized that if there were pressure to
maintain the net charge on CA, there might exist isoforms with
a high percentage of charged amino acids and isoforms with a
low percentage of charged amino acids, but both would have a
similar net charge.
To test this hypothesis, we normalized the number of
positively and negatively charged amino acids by the total
number of amino acids for each protein in Table 1. Figure 7
shows a plot of the normalized number of negative residues
versus the normalized number of positive residues. A protein
with a high proportion of amino acids that are negatively charged
also has a high proportion of amino acids that are positively
charged.69 The slope of the data is less than unity, so each
positive charge is balanced by only 0.8 negative charges.
This correlation implies that there is some evolutionary
pressure to maintain the charge on CA. The net charge is
(69) A more detailed analysis than presented here, on a larger data set, would
be needed to ascertain whether a mutation that increases or decreases the
charge on CA is more likely to be followed by a mutation that restores the
charge to the original value than one that leaves the new charge unaltered.
9
J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005 4713

A R T I C L E S
Gudiksen et al.
preserved, while the total number of charges varies. Even though
the total number of charges (positive + negative) ranges from
36 (for CA 2 from flaveria) to 75 (for CA V from mouse), the
net charge only varies by ∼4.5 charges, between 1.5 and -2.9.
The reason for this small difference in net charge remains
unclear. The net charge, thus, may be more important from an
evolutionary perspective to carbonic anhydrases than the details
of the distribution of electrostatic potential on the surface.
We further examined the location of the charges on the
surface of the CA isoforms to determine if the location of the
charges was conserved. We analyzed the four isoforms of CA
that have known crystal structures: human carbonic anhydrases
I and II, bovine carbonic anhydrase II, and mouse carbonic
anhydrase V. Because these isoforms have a net charges of -0.2,
-1.9, -1.1, and -0.2 at pH 7.4 and a total number of charges
of 63, 56, 52, and 75, respectively, they are representative of
the isoforms detailed in Table 1.
We used the MOE program70 to calculate the electrostatic
potential on the surface of the proteins (Figure 8). There are
regions on the surface of the proteins that appear to be roughly
conserved between isoforms (e.g. the blue, positive patch on
the side of each isoform marked with an arrow), but the overall
distribution of charges appears to be significantly different for
each isoform. We do not know if this patch of positive charges
on the side of the CA isoforms has a function; we know only
that it appears in all four isoforms with known crystal structures.
It appears that there is not an overall distribution of charges
that has been conserved to optimize CA; instead, only the net
charge is conserved between isoforms.
osmotic pressure than if it were required to balance a large
number of highly charged proteins with a large number of small
ions. A net charge on a protein near zero, however, may induce
aggregation, so a balance between these two factors may cause
many proteins in a cell to have a small charge.
We find that the net charge on CA isoforms is more conserved
than either the number of charges or the distribution of the
charges on the surface. Because CA is not known to interact
with other proteins, and because its substrates are relatively
simple (CO2, H2O, and HCO3-), it seems unlikely that there is
a need to orient reactants or products using electrostatics. It is,
however, unclear if the conservation of charge is specific to
CA or if the net charge on all proteins is maintained near zero.
We have demonstrated that, for BCA II, the charges are not
important for folding or activity in vitro.
BCA is not a perfectly representative protein (probably no
protein is perfectly representative): it is exceptionally stable,
has a high percentage of amino acids in â-sheets, and has a
catalytically active His3-Zn(II)-OH site. Still, the results in this
paper demonstrate that there is at least one, not atypical, globular
proteinsBCAsfor which charge is not important in folding.
This observation is almost certainly not true for all proteins,
and further experiments will be necessary to determine the
generality of these results.
Acknowledgment. We thank Prof. Carol Fierke from the
University of Michigan for thoughtful discussions. We acknowl-
edge the National Institutes of Health for research support
(GM51559) and for a postdoctoral fellowship to A.R.U.
(AI057076). I.G had a graduate research fellowship from the
National Science Foundation.
Sear calculated the charge on all of the proteins in the
Escherichia coli genome71 and found that the distribution of
charges on proteins was roughly symmetric and centered near
zero, with a standard deviation of 8.72 The soluble isoforms of
CA fall in the center of this distribution. Perhaps if the charge
on proteins is near zero, the cell can more easily regulate its
Supporting Information Available: Experimental protocols
for the acetylation of BCA, the calculations of the surface
charges, the DNSA binding and enzymatic assays, the CE and
CD procedures, and the microcalorimetry data for the cmc
determination of SDS in TG buffer. This material is free of
charge via the Internet at http://pubs.acs.org.
(70) MOE (The Molecular Operating Environment), v 2003, Chemical Comput-
ing Group Inc.
(71) Sear, R. P. J. Chem. Phys. 2003, 118, 5157-5161.
(72) Sear neglected histidine residues in his calculations. If the histidines were
included, the charges on proteins would be more positive than the charges
he calculated, but the magnitude of this effect is unknown.
JA043804D
9
4714 J. AM. CHEM. SOC. VOL. 127, NO. 13, 2005